Datasets:

griffin

/

ChemSum

like 12

Cessation of biomechanical stretch model of C2C12 cells models myocyte atrophy and anaplerotic changes in metabolism using non-targeted metabolomics analysis

Studies of skeletal muscle disuse, either in patients on bed rest or experimentally in animals (immobilization), have demonstrated that decreased protein synthesis is common, with transient parallel increases in protein degradation. Muscle disuse atrophy involves a process of transition from slow to fast myosin fiber types. A shift toward glycolysis, decreased capacity for fat oxidation, and substrate accumulation in atrophied muscles have been reported, as has accommodation of the liver with an increased gluconeogenic capacity. Recent studies have modeled skeletal muscle disuse by using cyclic stretch of differentiated myotubes (C2C12), which mimics the loading pattern of mature skeletal muscle, followed by cessation of stretch. We utilized this model to determine the metabolic changes using non-targeted metabolomics analysis of the media. We identified increases in amino acids resulting from protein degradation (largely sarcomere) that occurs with muscle atrophy that are involved in feeding the Kreb\xe2\x80\x99s cycle through anaplerosis. Specifically, we identified increased alanine/proline metabolism (significantly elevated proline, alanine, glutamine, and asparagine) and increased \xce\xb1-ketoglutaric acid, the proposed Kreb\xe2\x80\x99s cycle intermediate being fed by the alanine/proline metabolic anaplerotic mechanism. Additionally, several unique pathways not clearly delineated in previous studies of muscle unloading were seen, including: 1) elevated keto-acids derived from branched chain amino aicds (i.e. 2-ketoleucine and 2-keovaline), which feed into a metabolic pathway supplying acetyl-CoA and 2-hydroxybutyrate (also significantly increased); and 2) elevated guanine, an intermediate of purine metabolism, was seen at 12 hours unloading. Given the interest in targeting different aspects of the ubiquitin proteasome system to inhibit protein degradation, this C2C12 system may allow the identification of direct and indirect alterations in metabolism due to anaplerosis or through other yet to be identified mechanisms using a non-targeted metabolomics approach.

cessation_of_biomechanical_stretch_model_of_c2c12_cells_models_myocyte_atrophy_and_anaplerotic_chang

1. Introduction<!>2.1 C2C12 plating and differentiation on FlexCell plates<!>2.2 Biomechanical stretch of differentiated C2C12 cells<!>2.3 Imaging and cell surface area determinations<!>2.4 Non-targeted metabolomics analysis<!>2.5 Metabolomic statistical analyses<!>3.0 Results<!>4.0 Discussion<!>5.0 Conclusions

<p>Recovery from injury and illness can cause extended periods of muscle disuse and/or unloading resulting in rapid skeletal muscle atrophy and loss of functional strength1–3. The extent of muscle loss that occurs during illness is an important predictor of hospitalization duration and need for rehabilitation4. The resulting decline in functional capacity and strength, decline in basal metabolic rate, and onset of insulin resistance reported to be some of the sequelae in humans5. Since the signs of muscle disuse atrophy occurs rapidly, understanding the short term alterations that occur may help us better appreciate their contribution to sarcopenia and elucidate ways in which to counteract the particularly fast and severe consequences that result.</p><p>Studies of skeletal muscle disuse either in patients on bed rest or experimentally in animals (immobilization) have demonstrated that decreased protein synthesis is common, with transient parallel increases in protein degradation (as recently reviewed5). Muscle disuse atrophy involves a process of transition from slow to fast myosin fiber types6. A shift toward glycolysis, decreased capacity for fat oxidation, and substrate accumulation in atrophied muscles have been reported, as has accommodation of the liver with an increased gluconeogenic capacity6. These alterations in glycolysis appear to regulate more than carbohydrate metabolism7.</p><p>Recent studies have modeled skeletal muscle disuse by using cyclic stretch of differentiated myotubes (C2C12), which mimics the loading pattern of mature skeletal muscle, followed by cessation of stretch8. The resulting atrophy mimicked disuse atrophy and allowed investigation of the underlying mechanisms. In the current study, we sought to more broadly identify the metabolic changes that occur acutely in differentiated muscle cells using a non-targeted metabolomics approach of both cells and media in vitro. While we identified previously reported effects on glucose utilization, a host of previously undescribed alterations in protein metabolism was found and offers insight into the mechanisms underlying acute skeletal muscle disuse atrophy that may be targeted therapeutically in the future.</p><!><p>The C2C12 mouse myoblast cell line (ATCC, CRL-1772) was plated at 30–50% confluence onto six-well BioFlex® culture plates (BF-3001, Flexcell International Corporation, Hillsborough, NC) coated with Collagen type I. Myoblasts were grown to 80% confluence in DMEM (4500 mg/L glucose) with 20% FBS, then differentiated to myotubes by changing media to DMEM with 2% horse serum supplemented with insulin (1 microM final), as previously described9–11. Cells were plated were then allowed to quiesce in DMEM (25 mM glucose) without media for 24 hours before stretch stimulus was applied to quiesce the cells, as previously described in the context of cardiomyocyte-induced hypertrophy by IGF-1 and thyroid hormone in vitro12,13.</p><!><p>Differentiated C2C12 cells were either stretched at 15% biaxial stretch for 6 hours using the Flexcell® FX-5000™ Compression System (Flexcell International Corporation, Hillsborough, NC), or used as controls (non-stretched) and incubated in parallel to experimental cells without stretch (non-stretched cells). The stretched differentiated C2C12 cells were then unloaded by stopping the stretch and allowing the cells to incubate for another 12 hours. Media was collected at baseline, or after 6 hours of stretch, and 1, 3, 6, and 12 hours after the cessation of stretch. Replicate control samples and cells undergoing cessation of stretch were maintained under static conditions with no applied cyclic strain. In parallel, second experimental group of differentiated C2C12 cells were stimulated with 10% FBS for 6 hours to induce hypertrophy (experimental controls were incubated in parallel without FBS).</p><!><p>Cultured myotubes were viewed and imaged using a Zeiss inverted microscope (Axiovert 200, Oberkochen, Germany). Three digital images per culture were captured. The images were analyzed for myotube length, diameter and area using ImageJ imaging software (NIH) as previously described8.</p><!><p>Collection of samples (spent media) from all 6 wells was performed, and samples were placed on dry ice/stored at −80°C. Samples were then analyzed by GC/MS, as previously described14. The raw, transformed, and sorted data used for each of the three comparisons in the metabolomics analyses can be found in Supplemental Table 1. Up to 1 missing value per group was imputed using lowest value in the same group, with groups missing 2 or more excluded from the analysis. All data used in this analysis has been archived in the UCSD Metabolomics Workbench (http://www.metabolomicsworkbench.org/), accession #(Pending Assignment).</p><!><p>Metaboanalyst (v3.0) run on the statistical package R (v2.14.0) used metabolite peaks areas (as representative of concentration)19,20. Data were first normalized to a pooled average sample from their control and scaled using Pareto scaling feature, and then analyzed to calculate fold change using Metaboanalyst fold change feature. Unsupervised principal component analysis (PCA) was performed next, which identified the effect of stretch cessation as the principal source of variance. The metabolites that best differentiated the groups were then individually tested using univariate analysis of individual component by t-test (Metaboanalyst v 3.0), and then the t-test and VIP significant metabolites were matched to metabolomics pathways using the Pathway Analysis and enrichment analysis features in Metaboanalyst 3.0. In addition, a Pairwise One-Way Analysis of Variance (ANOVA) and Fisher LSD post-hoc were conducted among all groups at different time points using Metaboanalyst v3.0 to delineate the changes induced by stretch and cessation. Only metabolites identified and detected in all groups were included in the One-Way ANOVA. If two or more values of each metabolite were missing in a given group, the entire metabolite was removed from the analysis. Data used in this study are available in Supplemental Table 1. Heat maps were generated using the GENE-E software (http://www.broadinstitute.org/cancer/software/GENE-E/index.html). All data is shown as mean +/− standard error of the mean (SEM), unless otherwise indicated.</p><!><p>Using differentiated C2C12 myotubes seeded at the same time with equal numbers of cells, we cyclically stretched cells (15% at 60 cycles/min) for 6 hours, then stopped the stretch and collected media at 1, 3, 6, and 12 hours post-stretch. Media from the same cells were collected just before and 6 hours post-stretch (indicated by 0 h) was analyzed using non-targeted metabolomics (Figure 1A). Compared to non-stretch controls, C2C12 cells significantly increased in cross-sectional are after 6 hours (Figure 1B). Stretch cessation resulted in a significant reduction in myocyte area within 3 hours (Figure 1B), which continued to decrease at 6 and 12 hours, respectively. To discover the alterations that occurred metabolically in these myocytes, we analyzed the media in the stretch cells and compared it to non-stretched cells in vitro, investigating each of the five times analyzed (6 hours stretch, and 1, 3, 6, and 12 hours post-stretch cessation).</p><p>Using non-targeted metabolomics analysis, we compared the metabolites found in control no-stretch cell media to cells stretched for 6 hours without serum (Figure 2). We identified eighty-two metabolites (of 110 substances peaks recorded)(Figure 2A). Of the 82 metabolites named, 12 were significantly different from non-stretch controls by t-test (Figure 2B). By category, stretch affected arginine/proline metabolism and starch/sucrose metabolism (Figure 2C). We identified 84 metabolites in serum-stimulated C2C12 cells (Supplemental Figure 1A), with 32 of them significantly different from parallel controls (Supplemental Figure 1B). Like stretch-induced hypertrophy, serum-stimulated C2C12 cells showed significant alterations in protein biosynthesis and alanine metabolic pathway as the most enriched pathways and lowest p-value (Supplemental Figure 1). Twelve (12) significantly altered metabolites were found in the stretch challenged C2C12 cells compared to thirty-two (32) in the serum-induced after 6 hours illustrate the many differences in these stimuli as well, including citrate cycle, arginine/proline metabolism, and glycine/serine/threonine metabolic pathways identified (Supplemental Figure 1). Pathway enrichment analysis of the 6 hours stretch and 6-hour serum significant metabolites illustrate the similarities between the two, including "protein biosynthesis" and "alanine metabolism" having the lowest p values, while citric acid cycle changes were unique to the serum-stimulated pathway (Supplemental Figure 2).</p><p>After identifying the metabolic alterations that occurred with stretch, we next determined the metabolic changes at 1, 3, 6, and 12 hours after (Figures 3–6). After 1 hour, three metabolites were significantly different between groups (of 83 metabolites identified from 191 substance peaks recorded)(Figure 3A and 3B), involving starch and sucrose metabolism (Figure 3C). Similarly, after 3 hours of stretch cessation, nine metabolites were significantly different between groups (from 87 metabolites identified from 197 substances peaks recorded)(Figure 4A and 4B), also involving aminoacyl-tRNA biosynthesis (Figure 4C). At 6 hours of stretch cessation, 16 metabolites were significantly different between groups (of 86 metabolites identified from 196 substances peaks recorded)(Figure 5A and 5B), involving valine/leucine/isoleucine biosynthesis, pyruvate metabolism, and aminoacyl-tRNA biosynthesis (Figure 5C). By 12 hours of stretch cessation, 30 metabolites were significantly different (of 80 metabolites identified from 193 substances peaks recorded)(Figure 6A and 6B), involving primarily intermediate metabolism (alanine/aspartate, and glutamate metabolism, citrate cycle, and nitrogen metabolism)(Figure 6C).</p><p>To summarize the dynamic metabolic alterations throughout the entire experiment, a One-Way ANOVA was performed on all the metabolites at each time point compared to the 6 hours stretch (0 hr) time point (Figure 7, Figure 8, Supplemental Figure 3, Supplemental Figure 4). The metabolic process affected most prominently by ANOVA analysis is the Kreb's Cycle (Figure 7). Since glucose uptake occurs in response to stretch, we followed media glucose levels as it was an identified metabolites found throughout. Glucose levels did not significantly change throughout the time course (Supplemental Figure 5). Significant increases in lactic acid, citric acid, α-ketoglutaric acid, and malate were identified and significantly increased by 12 hours of stretch cessation and hypertrophy reversal (see inset plots). Both ANOVA and t-tests identified decreases in pyruvate, complementing the analysis performed in Figure 6, investigating the significant alterations by t-test at the 12-hour stretch cessation compared to the 6-hour hypertrophy time point (0 Hr), indicated in blue and red. In addition to the ANOVA significant findings, fumarate was found to be decreased 3.5 fold (Figure 7).</p><p>One-way ANOVA performed on all the time points revealed alterations in alanine and proline metabolism (Figure 8). As adjacent pathways to the Kreb's cycle, increases in proline, alanine, glutamine, and asparagine were increased with the cessation of stretch by 3–12 hours (Figure 8, see inset graphs). In addition to the ANOVA significant metabolites, the 12 hours t-test identified glutamate and glycine reduced (see metabolites in blue, Figure 8). Lastly, ANOVA significant metabolites were also identified in purine metabolism (Supplemental Figure 3A), branched chain amino acid synthesis (Supplemental Figure 3B), phospholipid metabolism (Supplemental Figure 3C), and in pathways not clearly delineated, including sugar alcohols identified as pentitols and hexitols (Supplemental Figure 4).</p><!><p>Recent studies have demonstrated that cyclic stretch of differentiated C2C12 myotubes for as little as 1 hour per day (12% at 0.7 Hz) increases myocyte size in as little as two days8. To streamline this model, we found that in preliminary studies that stretching cells similarly (15% at 1.0 Hz) could induce myocyte growth in 6 hours, which became the basis for the present study. Cyclic stretch of C2C12 myotubes induces the release IGF-1 and AngII, which act as autocrine/paracrine stimulators of IGF-18 and Angiotensin II15 signaling on the C2C12 cells themselves, inducing an increase in cell size. As myocyte hypertrophy represents the increase in newly synthesized sarcomeres, the significance of the observed IGF-1 induced protein synthesis (evidenced by increased p-Akt and p-GSK3beta, respectively) is clear. Conversely, cessation of stretch reproducibly causes a loss of myocyte mass8, involving the proteasome-dependent degradation of sarcomere proteins that involves both the ubiquitin-proteasome system and calpain enzymes16,17. The myocyte-specific ubiquitin ligases MuRF1 and MAFbx mediate muscle atrophy18, recognizing and ubiquitinating sarcomere proteins, then targeted for proteasome-dependent degradation19. Atrophy in C2C12 myotubes can be inhibited by blocking the proteasome20 or by blocking ubiquitination to prevent atrophy in C2C12 myotubes16. Recent studies have identified that activating calpain enzymes serve to induce the ubiquitin-proteasome pathway, thereby acting in concerted to degrade sarcomere proteins during atrophy21.</p><p>The induction of anaplerosis in atrophy is an important compensatory mechanism by which amino acids feed the TCA cycle to generate energy. The link between atrophy and anaplerosis is logical, given the tremendous sarcomere protein degradation that occurs, freeing peptides and amino acids used for energy. In addition to the structural changes that occur with muscle unloading and atrophy, recent studies have identified an intricate link between skeletal muscle mass regulation and energy metabolism22. Conceptually, this link lies in the well-established increase in net protein degradation that produces an increase in amino acids5,7,23. During the development and progression of colorectal cancer, tumor-bearing animals experienced decreases in pyruvate kinase (PKM1) and pyruvate kinase activity but were unexpectedly able to maintain energy7. These studies revealed that an increase in autophagy supported the loss of muscle mass and increased amino acid production, including increased glutamine7. Through a well-known process of anaplerosis that occurs in skeletal muscle, glutaminase can change glutamine to glutamate, whereby glutamate dehydrogenase can create the α-ketoglutaric acid intermediate to feed the TCA cycle24,25. During the progression of skeletal muscle atrophy due to cancer, increases in glutamine were thought to be the main mechanism by which ATP levels were maintained in the presence of decreased pyruvate kinase activity7. In the present study, we expand this evidence to demonstrate increases in alanine/proline metabolism (significantly elevated proline, alanine, glutamine, and asparagine, (Figure 8) and increased α-ketoglutaric acid, the proposed intermediate (Figure 7) being fed by the alanine/proline metabolic anaplerotic mechanism. Lastly, the anaplerotic pathway shunting pyruvate to malate has been described as a potential anaplerotic mechanism in skeletal muscle through the malic enzyme26. While it has not clearly been described to our knowledge in unloading, it may be at work in the present model to describe the elevated malate seen after unloading (Figure 7).</p><p>In addition to the anaplerotic mechanism feeding α-ketoglutaric acid seen with C2C12 unloading, we identified several unique pathways not clearly delineated in previous studies of muscle unloading. The CDP-ethanolamine pathway has been described to regulate skeletal muscle diacylglycerol content and mitochondrial biogenesis27. The drug clenbuterol is an ethanolamine that has been used to treat skeletal muscle atrophy, including dexamethasone-induced muscle atrophy28, denervated skeletal muscle29, and in treating humans with spinal and bulbar muscular atrophy for a 12 month trial30. These studies indicate that the ethanolamine clenbuterol protects skeletal muscle loss by antagonizing Akt/mTOR and by increasing polyamines (putrescine, spermidine, and spermine), which are repressed in muscle atrophy31–33. The presence of elevated ethanolamine may represent novel endogenous pathways involved in the recovery of atrophy (Supplemental Figure 3C). Significantly increased concentration of the keto-acids derived from branched chain amino acids (i.e. 2-ketoleucine and 2-keovaline) were identified after unloading (Supplemental Figure 3B) and have been described as intermediates in the Ehrlich amino acid degradation pathway. These keto-acids derived from branched chain amino acids feed into a metabolic pathway supplying acetyl-CoA and 2-hydroxybutyrate (also significantly increased, Figure 7). Since the keto-amino acids are further reacted in this pathway to make higher alcohols, it is possible that the generic sugar alcohols may be related to this process (Supplemental Figure 4A). While the elevation of amino acids during skeletal muscle atrophy is well known34, the keto-acids derived from branched chain amino acids seen here have not previously been described specifically. Lastly, elevated guanine was seen at 12 hours unloading, which previously was undetectable at multiple time points (Supplemental Figure 3A). While guanine is an intermediate in purine metabolism, its role in unloading remodeling has not been described to our knowledge in the literature previously.</p><p>Metabolomic analysis of cell culture can involve extracellular (footprint) and intracellular (fingerprint) metabolic profiles. One advantage of metabolic footprinting is the ability to assay serially, as in the present study, to allow identification of changes over time in the same cell culture system. Footprinting of B-lymphocytes into Ig secreting plasma cells using supernatants identified the consumption of glucose, glutamine, and other essential amino acids35. 5′-methylthioadenosine (5′MTA) was produced and released during the activation phase, while the second phase of differentiation corresponded to deoxycytidine release35. Changes in extracellular metabolites correlated with the known biochemical pathway changes reported previously35. The secretome footprinting of cancer cells has similarly shown parallels between the metabolites secreted and the underlying biology of the cells releasing these metaoblites36. Specifically, identification of isoleucine, α-aminoadipic acid (AADA), and γ-amino-N-butyric acid (GABA) in media were accurate biomarkers of c-Myc, KLF4, and Oct1 variable expression, with modulation of these markers indicative of changes in the underlying biology of the tumor36. The reason for the parallel metabolic mirroring of the footprint with the intracellular (fingerprint) in these studies is not addressed directly in these studies.</p><p>One potential mechanism that may explain the parallel metabolomics profiles seen in extracellular media and cells alike may be exosome-like vesicles (ELV)37. Small molecule profiling from ELVs isolated from cell culture media has recently demonstrated to be much more dynamic. The rich metabolome that was found included glycerophospholipids/sphingolipids, peptides, cyclic alcohols, fatty acid esters/amides, alcohols, sugars, steroids, amino acids/conjugates, nucleotides, and organic acids in cell culture media from exosome-like vesicles found in the media37. While we did not isolate ELVs from media in the present study, this finding demonstrates that the diversity of metabolites we did identify have been reported in media ELVs as a source. The transport and movement of these diverse metabolites are not well described but are different enough that multiple mechanisms, including those outside of ELVs not previously described, may similarly be involved.</p><p>A limitation of this study is the limited fatty acids available to the differentiated C2C12 cells during the stretch phase and subsequent hypertrophy reversal. C2C12 cells have the ability to take up and oxidize fatty acids for energy production and have the ability to store fats as triglycerides as well38. Adequate supplementation of fatty acids complexed to albumin was available in the serum used to proliferate and differentiate these cells. DMEM with 10% fetal bovine serum was used to induce proliferation of C2C12 myoblasts; the media was then replaced with DMEM with 2% horse serum to induce differentiation. However, the stretch studies were then performed in the absence of serum as serum is a potent inducer of hypertrophic growth (containing IGF-1, Angiotensin II, and thyroid hormone, among other pro-hypertrophic constituents8,15; see serum-induced hypertrophic controls, Supplemental Figure 1). While fatty acids were not directly added to the media via serum for these reasons, fatty acids were identified in the media by our present analysis. Specifically, palmitic acid, alpha-Monopalmitin, and/or Beta-Monopalmitin were found at each time found (Supplemental Table 1). Recent studies have demonstrated that the essential fatty acids linoleic acid (LA) and alpha-linolenic acid (ALA) themselves can improve glucose uptake in C2C12 cells made insulin resistant using palmitic acid39, making the balance between fatty acids and glucose critically important when modeling myocytes in vitro.</p><!><p>Cessation of biaxial stretch in C2C12 myotubes in culture resulted in a reduction in cross-sectional area over 12 hours, paralleling some metabolic changes using non-targeted metabolomics analysis of the media. Using a One-Way ANOVA analysis comparing the cessation time point, increases in amino acids were identified, resulting from the protein degradation (largely sarcomere) that occurs with muscle atrophy. Increases in alanine/proline metabolism (significantly elevated proline, alanine, glutamine, and asparagine) and increased α-ketoglutaric acid, the proposed intermediate being fed by the alanine/proline metabolic anaplerotic mechanism were identified. Additionally, several unique pathways not clearly delineated in previous studies of muscle unloading were seen, including: 1) elevated keto-acids derived from branched chain amino acids (i.e. 2-ketoleucine and 2-keovaline), which feed into a metabolic pathway supplying acetyl-CoA and 2-hydroxybutyrate (also significantly increased); and 2) elevated guanine, an intermediate of purine metabolism, was seen at 12 hours unloading. Given the interest in targeting different aspects of the ubiquitin-proteasome system to inhibit protein degradation18,40,41, this C2C12 system may allow the identification of direct and indirect alterations in metabolism due to anaplerosis or through other yet to be identified mechanisms using a non-targeted metabolomics approach.</p>

PubMed Author Manuscript

Quantum Chemical Approaches in Structure-Based Virtual Screening and Lead Optimization

Today computational chemistry is a consolidated tool in drug lead discovery endeavors. Due to methodological developments and to the enormous advance in computer hardware, methods based on quantum mechanics (QM) have gained great attention in the last 10 years, and calculations on biomacromolecules are becoming increasingly explored, aiming to provide better accuracy in the description of protein-ligand interactions and the prediction of binding affinities. In principle, the QM formulation includes all contributions to the energy, accounting for terms usually missing in molecular mechanics force-fields, such as electronic polarization effects, metal coordination, and covalent binding; moreover, QM methods are systematically improvable, and provide a greater degree of transferability. In this mini-review we present recent applications of explicit QM-based methods in small-molecule docking and scoring, and in the calculation of binding free-energy in protein-ligand systems. Although the routine use of QM-based approaches in an industrial drug lead discovery setting remains a formidable challenging task, it is likely they will increasingly become active players within the drug discovery pipeline.

quantum_chemical_approaches_in_structure-based_virtual_screening_and_lead_optimization

Introduction<!>Quantum chemical approaches in protein-ligand docking<!>Calculation of ligand binding free energy using quantum mechanics-based methods<!>Conclusions and perspective<!>Author contributions<!>Conflict of interest statement

<p>The drug discovery process relied for many years on the experimental high-throughput screening of large chemical libraries to identify and optimize new drug lead compounds. In spite of efforts to improve its efficiency, this remained an expensive and time consuming process (Phatak et al., 2009). The availability of 3D structures of protein-ligand (PL) complexes has guided lead optimization for many years, paving the way to a more rational approach. Later on, theoretical developments, coupled with better computational algorithms and faster computing resources, allowed the routine use of in silico methods to model PL interaction, estimate binding affinity, and screen chemical libraries using structure-based approaches. Today, computational chemistry is a well-established and valuable tool in the drug discovery process (Cavasotto and Orry, 2007; Jorgensen, 2009).</p><p>The central quantity in PL association is the binding free energy (ΔGbinding), a property of enormous relevance in the pharmaceutical industry, and no effort is too great to accurately estimate it in a computationally efficient way. Reliable prediction of receptor-small-molecule affinities in the early-stages of the drug discovery pipeline would be instrumental to rationally design new, more potent, and safer drugs, saving precious effort, time and cost. The accurate calculation of ΔG depends on several factors: (i) the energy model of the system; (ii) the accounting for protein flexibility; (iii) the presence of water molecules within the binding site and the solvation model. The last two challenges have been thoroughly addressed in recent reviews [(Cavasotto, 2011, 2012b; Spyrakis et al., 2011), and (Spyrakis and Cavasotto, 2015), respectively]. The last 20 years have seen a remarkable advance in theoretical and algorithmic developments for the calculation of binding affinities (Gohlke and Klebe, 2002; Gilson and Zhou, 2007; Mobley and Gilson, 2017), ranging from fast estimates, to be used in high-throughput docking and scoring (Cross et al., 2009; Cavasotto, 2012a), to much slower—yet more accurate—calculations using free energy perturbation or thermodynamic integration (Mobley and Klimovich, 2012; Hansen and Van Gunsteren, 2014), well-suited to guide chemical synthesis for hit-to-lead optimization. Most of these applications have been rooted in molecular mechanics (MM) force-fields (FF), but recent years have seen the development and application of quantum mechanical (QM) methods to biomacromolecular systems in the context of drug lead discovery and design. The recent blind challenges for ligand-pose and binding affinity predictions ran by the Drug Design Data Resource (D3R) in 2015 (Gathiaka et al., 2016) and 2016 (Gaieb et al., 2018) highlight the critical relevance of method development and benchmarking in pose prediction and affinity ranking of bound ligands.</p><p>It should be highlighted that the QM formulation accounts for all contributions to the energy (including effects missing in FFs, such as electronic polarization, charge transfer, halogen bonding, and covalent-bond formation), and thus is, in principle, theoretically exact; moreover, it offers the advantage of being general across the chemical space, avoiding system-dependent parameterizations, so that all elements and interactions can be considered on equal footing. In fact, QM has been present since the early days of computer-aided drug design (cf. the pioneering work of W.G. Richards on quantum pharmacology; Richards, 1977), and it is routinely used to derive FF parameters [such as torsional potentials from high level ab initio data and partial atomic charges by fitting to electrostatic surface potentials (ESP) (Mucs and Bryce, 2013)], in QSAR methods (De Benedetti and Fanelli, 2014), to study reaction mechanisms (Blomberg et al., 2014), and small-molecule strain (Forti et al., 2012; Juárez-Jiménez et al., 2015).</p><p>The goal of this short mini-review is to highlight the growing importance of quantum chemistry (QC) in the study of PL interaction, and present the latest applications of explicit QM calculations to structure-based drug design in the context of lead identification and optimization [for a survey on the development rather than application of QM methods for ligand-binding affinity calculations the reader is referred to an excellent recent review (Ryde and Söderhjelm, 2016); the review by Korth also offers a comprehensive coverage on the development of semiempirical QM and density functional theory (DFT) methods augmented by hydrogen-bonding and dispersion corrections (Yilmazer and Korth, 2016)].</p><!><p>In silico molecular docking has been widely used to determine the binding mode (pose) of small-molecules to a binding site. However, the true potential of this technique is revealed when used in a high-throughput fashion to screen up to millions of molecules, aiming to generate a sub-library rich in potential binders, thus imposing a structural filter on a given chemical library to prioritize compounds for synthesis. In high-throughput docking (HTD), where usually the protein is considered rigid or with very few degrees of freedom, two stages could be identified: (i) the prediction of the binding modes of molecules within the binding site (docking stage); (ii) the calculation of a score which attempts to predict the likelihood that a molecule will actually bind to the target. Although docking accuracy depends on the program used, the number of ligand poses with RMSD < 2 Å compared to the native structure can reach up to 80% of the studied cases (Warren et al., 2006; Wang et al., 2016). In some docking programs the binding pose is assessed by searching the global energy minimum ("docking energy") within the potential energy surface (PES) of the protein-molecule system. Other energetic contributions should be accounted for (such as the free energy of the unbound molecule, the entropy change, and desolvation effects) in order to assign a "docking score" to molecules of a chemical library; scoring functions are classified as force-field-based, empirical, and knowledge-based (Kitchen et al., 2004). It should be highlighted that the docking energy discriminates among poses of the same molecule, while the docking score is aimed at discriminating among different molecules of the set [usually docking scores are calculated on the best pose (of few best poses) of each molecule]. In many docking programs, however, the docking score is used for both purposes.</p><p>In the last 10 years there have been continuous efforts to enhance scoring functions by incorporating some type of QM-based calculations, especially deriving system-specific charges, such as the QM-polarized ligand docking approach (Cho et al., 2005). Some degree of improvement was observed using these tailored energy functions in terms of pose prediction. However, these advances will not be addressed here, and the reader is referred to a sound review covering these issues (Mucs and Bryce, 2013).</p><p>There are fewer works describing PL interactions with explicit calculations at the QM level. One should highlight the pioneering work of Raha and Merz (Raha and Merz, 2004, 2005), who introduced QMScore, a semiempirical QM (SQM) scoring function based on the Austin Model 1 (AM1) Hamiltonian (Dewar et al., 1985), complemented with a FF dispersion term and a Poisson-Boltzmann implicit solvent model, and calculated using the linear-scaling divide and conquer method (Dixon and Merz, 1996). QMScore was able to discriminate native and decoy poses and captured essential binding affinity trends in a set of 165 PL complexes; a series of QM/MM scoring functions were also studied to discriminate native from decoy poses in six HIV-1 proteases (Fong et al., 2009), showing in some of them improvements over MM empirical potentials.</p><p>Very recently, the SQM/COSMO energy filter was introduced, aimed at discriminating native from decoy ligand docking poses (Pecina et al., 2016). The SQM/COSMO filter is a simplified version of a general binding free energy function (Raha and Merz, 2005; Lepšík et al., 2013),</p><p>(where ΔEint is the gas-phase interaction energy, ΔΔGsolv the solvation energy change upon complex formation, ΔGconf the change of conformational free energy, and –TΔS the entropy change upon binding). In this new filter, only the first two dominant terms in Equation (1) are conserved, thus avoiding expensive SQM optimizations. ΔEint is calculated at the PM6 level (Stewart, 2007) with the D3H4X correction for dispersion, hydrogen- and halogen bonding interactions (Rezáč and Hobza, 2011, 2012), and the implicit solvent model COSMO (Klamt and Schüürmann, 1993) is used to calculate the ΔΔGconf term (this filter was named PM6-D3H4X). It was shown that calculations in a small subsystem (the ligand and neighboring amino acids) do not deteriorate results compared to the whole system, with a clear benefit in terms of computational speed. The ability of this filter to discriminate binding-like poses from decoy poses was evaluated in four challenging systems [acetyl cholinesterase (AChE), TNF-α converting enzyme (TACE), aldose reductase (AR), and HIV-1 protease (HIV PR)], and compared to seven well-known empirical scoring functions and a physics-based AMBER/GB. It was shown that the SQM/COSMO filter performed best by two metrics: the number of false-positive solutions, and the maximum ligand RMSD of all poses within a given range of a normalized score. The worst performance was on the TACE metalloprotein, containing a Zn2+… S− interaction. As for the computational requirements, this filter is ~100 times slower than the traditional scoring functions, ~10 times slower than the AMBER/GB scoring scheme, but ~100 times faster than the standard SQM filter calculated using the full Equation (1). In a follow up contribution (Pecina et al., 2017), the SQM/COSMO filter was evaluated in the same four systems (AChE, TACE, AR, HIV PR) using the self-consistent charge density functional tight-binding (SCC-DFTB) (Elstner et al., 2001), complemented with the D3H4 corrections for dispersion and hydrogen-bond interactions (Rezáč and Hobza, 2012). This improved filter (named DFTB3-D3H4) retained its excellent performance in AChE, AR and HIV PR, and clearly improved the results on the TACE system at a reasonably higher computational price. To further validate the two variants of SQM filters, diverse 17 PL complexes were studied using the PM6-D3H4X and the DFTB3-D3H4X (extended in this case to account for halogen bonding), and compared to four standard docking programs (Ajani et al., 2017). The QM-based energy functions clearly outperformed the standard scoring functions in terms of the number of false positives.</p><p>Using MD simulations and QC energy evaluations, Burton and co-workers evaluated the preferred docking (binding) mode of the natural salpichrolide A and a synthetic analog with an aromatic D ring within the estrogen receptor α (ERα) binding site (Alvarez et al., 2015). The MM/QM-COSMO (Anisimov and Cavasotto, 2011; Anisimov et al., 2011) method with the PM6 Hamiltonian was used for the energy calculations. The MD simulations coupled with energy evaluations corresponding to different ligand-binding modes support the preferred inverted orientation of the steroids in the ERα binding site, in which the aromatic ring D occupies a similar position to the corresponding A ring of estradiol.</p><p>G protein-coupled receptors (GPCRs) present a challenging case for docking due to their solvent-exposed and polar binding sites (Cavasotto and Palomba, 2015). A new docking protocol was recently presented where a QM/MM + implicit solvation model was used to rescore docked ligand poses (Kim and Cho, 2016). The gas energy was calculated at a QM/MM level, considering the ligand and neighboring residues within 5 Å as the QM region, and the solvation energy was calculated using a Poisson-Boltzmann (PB) approach with partial charges derived from ESP fitting. Evaluating their protocol on 40 GPCR complexes including representatives of classes A, B, and F, the authors obtained an average RMSD of 0.78 Å, and a success rate of 40/40 for ligands with RMSD < 2 Å.</p><p>Chaskar et al. (2014) developed an on-the-fly QM/MM approach combining the EADock DSS docking algorithm (Grosdidier et al., 2007) with calculations based on the SCC-DFTB model and the CHARMM FF (Brooks et al., 2009), and evaluated it on a dataset of high-quality x-ray structures of zinc metalloproteins. Their method significantly improved the success rate compared to classical docking programs for orthosteric ligands in terms of ligand pose RMSD. Recently, a similar approach (Chaskar et al., 2017), but coupled with the Attracting Cavities docking algorithm (Zoete et al., 2016), was applied on three different sets: (i) the Astex Diverse data set of 85 common non-covalent drug/target complexes; (ii) a zinc metalloprotein data set of 281 complexes: (iii) a heme protein data set of 72 complexes, where ligand/protein interactions are dominated by covalent ligand/iron binding. On the first set the performance was similar to the standard scoring functions, but on the other two, QM/MM showed an improved performance, especially in the third set.</p><!><p>The binding process of five classical AChE inhibitors was analyzed using free energy perturbation (FEP) and QM/MM MD simulations (Nascimento et al., 2017). The QM calculations were performed at the AM1 level. The ΔGbinding was obtained as the sum of two terms, introducing two parameters into the electrostatic and van der Waals QM/MM interaction terms in the total energy (Swiderek et al., 2014). The correlation between the experimental and calculated values was in very good agreement (R2 value of 0.96 for 100 ps simulation time). Moreover, there was a qualitative agreement of the order of inhibition between theoretical and experimental values. The use of QM to describe these ligands was of great importance due to their polar nature and the high aromaticity of the enzyme binding site.</p><p>In order to analyze the efficiency of different approaches to calculate ΔGbinding at the QM/MM level employing MD simulations, Ryde and Olsson have recently compared the results of the calculation of the binding of nine small carboxylate ligands to the octa-acid deep cavity host (Olsson and Ryde, 2017), via reference-potential FEP calculations (Rod and Ryde, 2005) and full QM/MM FEP simulations. The ligand was described using a SQM PM6 Hamiltonian augmented by the DH+ empirical dispersion and hydrogen-bond corrections (Korth, 2010). The results showed that the reference-potential approach is approximately three times more effective than the direct approach, and the convergence of the MM → QM/MM perturbations is improved by the addition of QM/MM MD simulations for a number of coupling parameter values between the MM and QM/MM energies.</p><p>Grimme and co-workers presented a full QM approach to evaluate absolute ligand binding free energies as the sum of three terms: the interaction energy, the solvation contribution, and the entropic term (Ehrlich et al., 2017). Calculations were performed on a reduced system consisting of the ligand and neighboring binding site atoms (~1,000 atoms in total). For the interaction energy, two methods were used: the minimal basis Hartree-Fock HF-3c (Sure and Grimme, 2013) which includes a D3 dispersion correction (Grimme et al., 2010), and the composite hybrid PBEh-3c DFT lower computational cost method (Grimme et al., 2015); entropic contributions were calculated using a semiempirical DFTB3-D3 hessian (Gaus et al., 2011; Brandenburg and Grimme, 2014); the solvation contribution was calculated with the COSMO-RS method (Klamt, 1995, 2011). Two molecular systems were studied: the activated serine protease factor X (FXa) with 25 ligands and the non-receptor tyrosine-protein kinase 2 (TYK2) with 16 ligands. The mean absolute deviation (MAD) of the ΔGbinding using the HF-3c level was 2.8 and 2.7 kcal/mol, with a Pearson correlation coefficient 0.47 and 0.51, respectively; while a MAD of 2.1 kcal/mol was obtained on the FXa system using the PBEh-3c method, with a Pearson coefficient of 0.53. Although the results are clearly encouraging from a QC standpoint, this approach cannot be yet used in an industrial setting, and errors stemming from the structural optimization level, conformational sampling and the solvation contribution need further development.</p><p>Frush et al. performed a QM/MM-based evaluation of ΔGbinding on four diverse protein targets of pharmaceutical relevance: beta-secretase 1 (BACE1), TYK2, heat shock protein 90 α (HSP90), and protein kinase R (PKR)-like endoplasmic reticulum kinase (PERK), using 22, 16, 70, and 32 ligands, respectively (Frush et al., 2017). Binding affinities were calculated using the linear interaction energy (LIE) protocol (Aqvist et al., 1994), with α and β LIE coefficients similar to those reported elsewhere (Su et al., 2007), but modified to fit the experimental affinities of the TYK2 set. Ensemble averages were calculated through QM/MM calculations on MD trajectories, describing the ligand at the SQM level using the AM1 Hamiltonian, and the rest of the system using MM. On each of the four systems, the obtained MAE was 0.86, 0.42, 0.86, and 1.11 kcal/mol, respectively, and a correlation of 0.73, 0.71, 0.60, and 0.86, respectively. The authors concluded that their methodology reached a reasonable balance between accuracy and computational cost.</p><p>In the context of the D3R grand challenge blind test competition (Gathiaka et al., 2016), Ryde and co-workers evaluated four different approaches for predicting the binding affinities of three sets of ligands of the HSP90 protein (Misini Ignjatovic et al., 2016): (i) induced-fit docking (Sherman et al., 2006) followed by calculations with three energy functions; (ii) MM/GBSA calculations on minimized docked structures; (iii) optimization of docked structures with QM/MM calculations followed by QM-based energy evaluation of a subset of ~1,000 atoms using continuous solvent; (iv) calculations of relative binding affinities using free-energy simulations. Although the results were somehow poor, the authors were able to identify the sources of error: in one case the ligand could displace water molecules (this could be found only after the experimental data was released), and for other two, ligands might exhibit alternative binding modes that those in the crystal, or conformational changes of the system might be critical.</p><!><p>In this short review we presented the most recent applications of QM-based methods to molecular docking and ligand binding free energy prediction in the context of drug lead discovery, focusing on cases where QM is explicitly used to calculate at least some of the free energy contributions. The last 10 years have seen a remarkable interest in the development and application of QM-based methods in the field of drug discovery. This was triggered by the interest in modeling biomolecular systems in a more accurate way, and allowed by the unprecedented growth of computational power. QM methods are theoretically exact, capturing the underlying physics of the system and accounting for all contributions to the energy; thus, missing effects in FFs (such as electronic polarization, covalent-bond formation, and coupling among terms) are de facto accounted for in QM formulations, which are thus systematically improvable; being generally valid across the chemical space, they offer greater freedom to deal with non-standard molecules, avoiding the FF parameterizations.</p><p>Overall, the results obtained using QM approaches are very encouraging, but still different sources of error should be addressed in order to improve accuracy and predictability of these methods: (i) they are still system-dependent; thus, further validation and benchmarking are needed; (ii) in spite of the progress in computational speed, most QM applications to drug discovery cannot still be used in industrial settings, highlighting the need for optimized codes, especially those using GPUs; (iii) conformational sampling and protein flexibility: due to computing time, in most approaches aimed for high-throughput use, only local energy minimization is performed, or even no minimization at all; this should be integrated with the possibility of system cutout, and optimal combinations of these thoroughly validated; (iv) solvation contribution, especially in charges systems; (v) entropic considerations, usually omitted in many of this type of calculations. In spite of these limitations, it is clear that reliable QM methods for biomolecular systems would be a tremendous step forward toward predictive binding free energy calculations.</p><!><p>CNC conceived, designed, and supervised this review. All authors listed have made direct and intellectual contribution to the work, and approved it for publication.</p><!><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p>

PubMed Open Access

Fused-Ring Formation by an Intramolecular \xe2\x80\x9cCut-and-Sew\xe2\x80\x9d Reaction between Cyclobutanones and Alkynes

The development of a catalytic intramolecular \xe2\x80\x9ccut-and-sew\xe2\x80\x9d transformation between cyclobutanones and alkynes to construct cyclohexenone-fused rings is described herein. The challenge arises from the need for selective coupling at the more sterically hindered proximal position, and can be addressed by using an electron-rich, but less bulky, phosphine ligand. The control experiment and 13C-labelling study suggest that the reaction may start with cleavage of the less hindered distal C\xe2\x80\x94C bond of cyclobutanones, followed by decarbonylation and CO reinsertion to enable Rh insertion at the more hindered proximal position.

fused-ring_formation_by_an_intramolecular_\xe2\x80\x9ccut-and-sew\xe2\x80\x9d_reaction_between_cyclo

<p>Transition-metal-catalyzed C—C bond activation provides unique opportunities to develop various intriguing transformations.[1] In particular, oxidative addition of transition metals into C—C σ bonds followed by 2π insertion, namely a "cut-and-sew" process, has been demonstrated to be effective for the construction of complex ring scaffolds.[1r] Cyclobutanone derivatives are of special interest for this type of transformation because of their easy access from olefins and their high reactivity towards C—C activation.[1i,n,o,q,r] To date, significant progress has been achieved for the synthesis of bridged rings by means of intramolecular "cut-and-sew" reactions, in which cyclobutanones are coupled with an unsaturated unit tethered at the C3 position (Scheme 1a).[2] However, using such a strategy to assemble fused-ring systems is still challenging (Scheme 1b).[3]</p><p>The main difficulty associated with the fused-ring formation arises from the need for C—C cleavage and coupling at the more sterically hindered C2 (proximal) position (Scheme 2a); the selectivity typically favors the less bulky C4 (distal) position (Scheme 2b).[2g] In addition, decarbonylation of cyclobutanones to form the corresponding cyclopropane by product is always a major competing pathway.[2a,g,h] As illustrated in Scheme 2a, direct formation of rhodacycle A, the reactive intermediate for subsequent 2π insertion, is more difficult than formation of rhodacycle B. One possible solution is to enable a facile and reversible decarbonylation and reinsertion pathway,[4] in which rhodacyclopentanone B can be initially converted to rhodacyclobutane intermediate C and then to rhodacycle A by CO reinsertion. We anticipated that the choice of ligand would be critical for this transformation; the ligand should allow efficient decarbonylation and CO reinsertion without promoting further reductive elimination of C (an irreversible process to give cyclopropanes, see below, Scheme 5a), and represents the main difference from the prior benzocyclobutenone system.[5] Herein, we disclose the development of an effective catalytic system for fused-ring formation by means of an intramolecular "cut-and-sew" reaction between cyclobutanones and alkynes (Scheme 3a).[6] The transformation is enabled by the use of an electron-rich, less bulky phosphine ligand and an electron-deficient Rh precatalyst, offering rapid access to cyclohexenone-fused rings.</p><p>Notably, similar bicyclic structures could also be obtained through [3+2+1] cycloaddition reactions[4e, 7] involving C—C bond cleavage of cyclopropanes. The coupling of simple cyclopropanes, CO, and alkynes was first reported by Koga and Narasaka,[8] albeit with low catalyst turnover and limited substrate scope (Scheme 3b). The use of more reactive vinyl cyclopropanes and cyclopropanes containing a directing group were recently developed by the groups of Yu[9] and Bower,[10] respectively; both substrate types exhibited excellent reactivity and selectivity. Hence, methods that directly activate simple cyclobutanones should offer a complementary approach to the prior [3+2+1] reactions without the need for CO gas or auxiliary directing groups.</p><p>To explore the proposed "cut-and-sew" reaction, cyclobutanone 1a was employed as the initial substrate (Table 1). After careful optimization, the desired benzofused [6.5.6] tricycle product (2a) was ultimately obtained in 82% yield by using [Rh(CO)2Cl]2 and PMe2Ph as the metal–ligand combination (Table 1, entry 1). Initially, control experiments showed that both the phosphine and Rh complex played pivotal roles in this reaction (Table 1, entries 2 and 3). A range of monodentate phosphine ligands was found to be effective, and generally, higher conversion was obtained with more electron-rich ligands (Table 1, entries 4–6). Surprisingly, one important factor was the ligand/metal ratio, with 1.6:1 being optimal (for detailed optimization, see the Supporting Information). When less ligand was employed (P/Rh=1:1), the reaction still gave complete conversion albeit with more cyclopropane side product (2a′); however, increasing the P/ Rh ratio to 2:1 completely stopped the reactivity (Table 1, entries 7 and 8). The finding could be attributed to the generation of the inactive trans-Rh(CO)(L)2Cl species. We reason that the active catalytic species likely contains only one phosphine ligand, but it is relatively unstable in the absence of extra PMe2Ph. In addition, use of the more π-acidic [Rh(CO)2Cl]2 as a precatalyst is also crucial to generate the active species; in contrast, use of more electron-rich Rh–olefin complexes gave almost no conversion of cyclobutanone 1a (Table 1, entries 9 and 10). A survey of solvents revealed 1,4-dioxane to be optimal (Table 1, entries 11 and 12). At a lower temperature (115 °C), the reaction can still proceed to give 67% yield (Table 1, entry 13). Finally, the temporary directing-group strategy was not effective, likely because the bulkier proximal C—C bond is difficult to cleave (Table 1, entry 14).[2c]</p><p>With the optimized conditions in hand, the substrate scope was next investigated (Table 2). Different aryl-substituted alkynes all underwent the "cut-and-sew" sequence to give the corresponding tricycle products (2a–2e). Alkyl-substituted alkynes are also competent coupling partners; primary, secondary, and tertiary alkyl substituents are all tolerated. Unsurprisingly, increasing the bulkiness on the substituent from propyl (2g) to isopropyl (2h) to tert-butyl (2i) groups reduced the yield. It is noteworthy that the reaction conditions are both pH and redox neutral. The acidlabile tert-butyldimethylsilyl (TBS) ether is compatible and 89% yield of product 2j was isolated. In addition, cycloalkyl-substituted alkynes can be effectively coupled; the generated vinyl cyclopropane moiety (2m) remained intact. Moreover, substitution on the arene (2n) or the methylene bridge (2o) (between the arene and cyclobutanone) is tolerated. The reduced yield for product 2o is due to the increasing cyclopropane formation; it is likely that the substitution hindered the migratory insertion to a certain extent. Interestingly, the aniline linkage provided an indoline scaffold (2 p). On the other hand, the nitrogen linker was also found efficient.[11] With such a linker, coupling with aryl-, alkyl-, and even silyl-substituted alkynes has been achieved, and the corresponding 6H-isoindole products can potentially serve as valuable synthetic building blocks.[10] Finally, both α- and β-substituted cyclobutanones can be employed, albeit in moderate yields [Equations (1) and (2)], probably caused by the increased steric hindrance in the substrates.</p><p> (1)</p><p> (2)</p><p>The intriguing cyclohexanone-fused ring structures generated from this "cut-and-sew" reaction can be conveniently derivatized (Scheme 4). Excellent diastereoselectivity was obtained in most cases, possibly driven by the formation of less strained [5.6] cis-fused rings. Dissolving-metal reduction, followed by alkylation or oxidation, afforded the α-disubstituted cyclohexanone products 3 (X-ray structure obtained) and 4 (stereochemistry tentatively assigned), respectively.[12, 13] Moreover, enolate-based alkylation occurred site- and diastereoselectively at the C6 position of the cyclohexenone moiety. Pd/C-catalyzed hydrogenation took place at the syn side to the methine proton and directly gave the corresponding saturated alcohol. Treatment of product 2a with base and hydrogen peroxide unexpectedly led to a γ-hydroxylation product (7).[14] Finally, iodine/dimethyl sulfoxide (DMSO) oxidation[15] converted the tricycle into a functionalized fluorene, and a Pd-catalyzed aerobic oxidation[16] surprisingly gave 9-fluorenone 9 as the dominant product.</p><p>With regard to the plausible reaction mechanism, there are two major questions. One is whether this [4+2] cycloaddition shares the same catalytic pathway as the [3+2+1] reaction involving cyclopropane ring opening.[10] The other question is whether the reaction pathway involves the cleavage of the less hindered distal C—C bond. To address the first question, control experiments with cyclopropane side product 2a′ were conducted (Scheme 5a). Subjecting 2a– to the standard [4+2] reaction conditions in the presence of CO gas, or to the optimal conditions developed by the groups of Narasaka[8] or Bower[10] for the [3+2+1] reaction, gave no desired 2a product. This result suggests that cyclopropane formation during the [4+2] reaction is probably irreversible and 2a′ is not an intermediate on the way to product formation. This observation is also consistent with the fact that coupling of unactivated cyclopropanes in the absence of directing groups is rather difficult.[8]</p><p>To explore the second question, 13C-labelling study was conducted (Scheme 5b). We hypothesized that, if the reaction involved cleavage of the less hindered distal C—C bond, a CO deinsertion and reinsertion into the less hindered alkyl group would have to occur (see above, Scheme 2a). Thus, if this were the case, use of the Rh catalyst containing 13CO ligands would introduce a 13C-labelled carbonyl moiety into the product. Indeed, replacement of [Rh(CO)2Cl]2 with [Rh-(13CO)2Cl]2 under the standard reaction conditions afforded product 2a in 82% yield with 21% 13C incorporation. Given that only 5 mol% [Rh(13CO)2Cl]2 was used, 86% 13CO from the Rh complex has been transferred into product. When the reaction was terminated at an earlier stage, higher 13C incorporation (34%) was observed without significant 13C incorporation in the recovered starting material (for more details, see the Supporting Information). These observations suggest that 1) decarbonylation and CO reinsertion must have occurred (Scheme 5c), 2) the exchange between the coordinated CO on the Rh center and the free CO is faster than the subsequent steps, and 3) reductive elimination of the rhodacyclopentanone intermediate to give back cyclobutanone 1a is significantly slower than migratory insertion into the alkyne moiety. Hence, this observation is consistent with the hypothesis that the reaction may involve cleavage of the less hindered distal C—C bond, followed by a decarbonylation and CO reinsertion process. However, the pathway initiated from direct activation of the bulkier proximal C—C bond cannot be completely ruled out at this stage.</p><p>In summary, we have developed the first intramolecular coupling between cyclobutanones and alkynes to construct versatile fused cyclohexenone scaffolds. In this reaction, 2π insertion can selectively take place at the more sterically hindered proximal position, and significantly extends the "cut-and-sew" scope with cyclobutanones, thereby enabling access to other fused structures. Detailed mechanistic studies are ongoing in our laboratory.</p><p>Supporting information and the ORCID identification number(s) for the author(s) of this article can be found under: https://doi.org/10.1002/anie.201712487.</p><p> Conflict of interest </p><p>The authors declare no conflict of interest.</p>

PubMed Author Manuscript

Tracking the Growth of Superparamagnetic Nanoparticles with an In-Situ Magnetic Particle Spectrometer (INSPECT)

which leads to the requirement of a higher field strength. Furthermore, the Brownian and Néel relaxation time increase with increasing size 13,14 .In addition to the magnetic properties, the geometrical properties such as size and shape of the nanoparticles are of major importance for the biodistribution and biocompatibility in biomedical applications [15][16][17] . Moreover, the pH of the solution affects the susceptibility of the SPION solution as reported by Lucchini et al. 18 . These properties cannot be measured with the help of an MPS device, but they play an essential role in the imaging process as well as in therapeutic applications. It is worth mentioning that the change in temperature has a direct effect on the response of SPIONs in an MPS. This is due to the fact that the Brownian relaxation time (τ B ) and the Néel relaxation time (τ N ) are dependent on the thermal energy. Moreover, heating leads to a change in viscosity which also affects the Brownian relaxation time (τ B ) 19,20 .There are a variety of techniques for measuring the magnetic properties of SPIONs such as vibrating sample magnetometry that measures the static magnetization curve and the AC magnetometry, which measures the magnetic susceptibility. Unfortunately, none of these are able to measure the spectral magnetic moment at the desired frequency and field strength used in an MPI scanner. On the other hand, there has been an immense progress in the development of MPS devices as, for instance, one-dimensional MPS without offset and with various excitation frequencies, two-dimensional Lissajous MPS, arbitrary-waveform spectrometer and many more [21][22][23] , developed for different MPI associated applications. Moreover, there are also new approaches to characterize the SPIONs such as a frequency-mixing method based on search coils, which measure the phase lag of the magnetic moment compared to the drive field, giving more precise information regarding saturation magnetization and average hydrodynamic size 24 . The measurements from MPS devices usually consist of amplitude and phase spectra. The amplitude spectra contain both the even and odd harmonics, odd harmonics are attributed to the particle signal and even harmonics are caused due to the presence of a DC offset magnetic field. These MPS devices are able to provide information regarding physical and geometrical particle properties after synthesis completion. Therefore, the synthesis process is largely unobserved and the nanoparticles produced are characterized after the completion of the synthesis process giving no information regarding the nucleation and growth of the SPIONs. Other methods such as X-ray scattering or transmission electron microscopy are being used for real-time monitoring of nanoparticles growth 25,26 , but these devices have a very tedious setup and cannot be used in a normal chemistry laboratory. Real-time transmission electron microscopy is a very advanced tool and provides optical and structural information in the nanometer to sub-Angstrom scale, but requires special in-situ chambers for carrying out the chemical reactions [27][28][29] . X-ray diffraction techniques do not provide any optical details but also require special experimentation chambers for observing different samples at higher temperatures and pressures 30,31 . Moreover, X-ray diffraction is very useful in studying mechanochemical reactions 32,33 Contrarily, in-situ Magnetic Particle Spectrometer (INSPECT) does not require any special sample chambers or experimentation setups for conducting syntheses. Moreover, INSPECT is a low-cost device with low instrumentation complexity and does not require user expertise in comparison to X-ray diffraction and TEM. INSPECT directly gives access to the change in the magnetic moment during the synthesis, a parameter which is important for magnetic imaging (e.g. MPI) and magnetic thermo-therapeutic scenarios. However, one limitation of INSPECT is that it relies on the magnetic properties of the SPIONs and hence cannot be used with other nanostructures.Synthesis research is a time-consuming and difficult task because there has been no real-time control of the synthesis process so far. The control of the synthesis parameters has been carried out in the case of practical blindness, as there was no insight into the current synthesis situation. The aim of this research is to develop a spectrometer for the real-time monitoring of the changes in the magnetic properties of the SPIONs during synthesis. To achieve this, a novel in-situ Magnetic Particle Spectrometer (INSPECT) is introduced. INSPECT measures the magnetization response of the tracer material during a sinusoidal magnetic excitation in-situ. The INSPECT hardware is based on the same principles as typical MPS 21 , but contains less hardware components and is adapted for usage in a chemistry laboratory. In Theory, a brief description of MPS related magnetism physics and the particle nucleation and growth process are given. The dedicated hardware and software as well as the laboratory setup and basic chemical reactions are described in Materials and Methods. In Results and Discussion, initial results obtained by performing a synthesis process within the INSPECT reaction chamber are shown. A brief conclusion concludes this contribution. theoryThis section deals with the basics of field strength calculations in a solenoid and the induced signal in the receiving coil. Moreover, a brief introduction to the chemical nanoparticle synthesis process including nucleation and particle growth is given.

tracking_the_growth_of_superparamagnetic_nanoparticles_with_an_in-situ_magnetic_particle_spectromete

<!>Fundamentals of nanoparticle nucleation and growth.<!>Materials and Methods<!>AC power amplifier.<!>Housing.<!>Receive chain calibration.<!>Measurement protocol.<!>Results and Discussion<!>Reference measurement for INspeCt. INSPECT works very efficiently with a high concentration of<!>Conclusion

<p>Magnetic particle spectroscopy (Mps) is a measurement technique to determine the magnetic properties of superparamagnetic iron oxide nanoparticles (SPIONs) in an oscillating magnetic field as applied in Magnetic particle Imaging (MpI). state of the art Mps devices are solely capable of measuring the magnetization response of the spIoNs to an oscillatory magnetic excitation retrospectively, i.e. after the synthesis process. In this contribution, a novel in-situ magnetic particle spectrometer (INspeCt) is presented, which can be used to monitor the entire synthesis process from particle genesis via growth to the stable colloidal suspension of the nanoparticles in real time. the device is suitable for the use in a biochemistry environment. It has a chamber size of 72 mm such that a 100 ml reaction flask can be used for synthesis. For an alkaline-based precipitation, the change of magnetic properties of spIoNs during the nucleation and growth phase of the synthesis is demonstrated. the device is able to record the changes in the amplitude and phase spectra, and, in turn, the hysteresis. Hence, it is a powerful tool for an in-depth understanding of the nanoparticle formation dynamics during the synthesis process.</p><p>Magnetic Particle Spectroscopy (MPS) is a technique to characterize superparamagnetic iron oxide nanoparticles (SPIONs) for their possible performance in magnetic particle imaging (MPI). MPI is a tomographic imaging modality measuring the spatial distribution of the SPIONs 1,2 . It is a quantitative imaging technique, which provides high sensitivity and can provide submillimeter spatial resolution 3 . The hardware realization for MPI is possible in a number of ways and is illustrated by Buzug et al. 4 . The properties of the magnetic particles strongly influence their usage for MPI or other applications. The quality of MPI images, for example, strongly depends on the anisotropy and size distribution of the SPIONs used as tracer material 3 . The nanoparticles used for the first experiments in MPI were predominately developed for magnetic resonance imaging (MRI) and are of limited use in the excitation fields of an MPI scanner. Moreover, magnetic nanoparticles have also been used in magnetic fluid hyperthermia (MFH) 5 . In MFH the nanoparticles serve as a medium for deposition of energy to a specific area using hysteresis losses 6 . Therefore, the hysteresis of the particles plays a critical role for most applications. Furthermore, SPIONs can be used for drug delivery 7 . In general, one can state that different fields of application require individually tailored nanoparticles. Actually, there is immense research to produce SPIONs, which are specifically tailored for the use in MPI as well as for MFH-based therapy 8,9 . These new types of SPIONs should have defined magnetic properties (e.g. saturation magnetization, steepness of magnetization, anisotropy) as well as geometrical properties (core diameter, hydrodynamic diameter distribution in suspension) to show optimal results in MPI. There are a lot of commercially available particles which have been characterized for MPI and compared according to their magnetic properties with the help of an MPS 10 . For example, nanoparticles with a core diameter of 20 nm and hydrodynamic diameter of 30 nm showed the best MPI performance so far 11 .</p><p>For MFH the hysteresis curve area should be large enough to deposit the required energy to cause cell ablation. For example, the larger the size of the nanoparticles, the steeper is the rise in the magnetization curve leading to the measurement of higher harmonics 12 . Due to this, the particles require more energy to get magnetized, Magnetic field strength. The magnetic field strength produced by a solenoid can be determined with the Biot-Savart law and Ampere's law. In particular, the axial magnetic flux density near the center of the solenoid with length L, number of windings N, the applied current I, μ 0 is the permeability of vacuum and n the density of turns (turns per meter length of the solenoid) can be determined by</p><p>Induced signal. The induced voltage u(t) in a receiving coil due to a temporal change in the particle magnetization M(x, t) can be calculated by using the reciprocity principle for magnetic recording 34</p><p>where V is the sample volume, p 0 is the spatial sensitivity of the coil, and x denotes the x-direction of the magnetic field. For simplicity, only a homogeneous generated magnetic field is considered. The average coil sensitivity p 0 is given as</p><p>, where, p 0 (x) = H r (x)/i 0 and H r (x) is the magnetic field generated by the receiving coil due to current i 0 . Using the above equation, the frequency spectrum of the magnetization m f ( ) is defined as 21 ∫</p><p>T j ft 0 0 0 2 with f the frequency, T = 1/f 0 and f 0 the fundamental frequency.</p><p>particle theory. The non-linear magnetization curve of a SPION is typically modeled by the Langevin theory of paramagnetism 35</p><p>is the magnetic moment at saturation of a single particle and c is the concentration of the particles. The Langevin parameter ξ can be defined as</p><p>where k B is the Boltzmann constant and T is the absolute temperature. The Langevin model is rather simple and neglects the size distribution of particles. Therefore, there are different saturation magnetizations M s of the particles in suspension. Hence, the magnetizations M D (t) have to be averaged. The resulting bulk magnetization ∼ M t ( ) is defined as</p><p>D 0 where ρ(D) is the probability density function of the particle size distribution. Using the bulk magnetization ∼ M t ( ), the induced voltage u(t) and the magnetic moment spectrum m f ( ) of the particle ensemble can be calculated 21 . But the Langevin model just provides a static model for the magnetic response of SPIONs. The Debye model corresponds to the magnetic response of SPIONs to a weak excitation field of frequency ω. According to the model, the real part of the magnetic susceptibility χ′(ω) monotonically decreases with the increase in frequency. Therefore, by using the value of ω it is possible to ascertain the mean value of Brownian relaxation time (τ B ) more precisely [36][37][38] .</p><!><p>Nucleation and growth of nanoparticles are a very complex multi-step process. Here a simplified model for the nucleation and growth is explained. The most important factor responsible for the nucleation and growth is a change in Gibbs free energy ΔG V , which depends on the concentration of the solute and can be described as</p><p>V b b 0 where c is the concentration of the solute, c 0 is the equilibrium concentration or solubility, Ω is the atomic volume of the solid phase and σ is the supersaturation (when the solution contains higher volume of the solute which can be dissolved in a solvent) defined by (c − c 0 )/c 0 . If there is no supersaturation (σ = 0), then no nucleation will happen and hence no particle growth. For the stability of the nuclei, it is important that they reach a critical size or they would dissolve back into the solution causing a reduction of the Gibbs free energy 39 . After the attainment of a critical radius the growth process will continue without any barrier. To achieve monodispersity, it is important that no new nuclei are formed after the initiation of the growth process. Therefore, the ideal condition for synthesizing monodisperse SPIONs is the formation of the nuclei in a very short time interval subsequently leading to a coherent growth process, which is highly dependent on the kinetics and change of the surface energy and can be controlled by changing different parameters of the chemical reaction such as temperature and the amount of solutes and solvents 39 . In the case of the alkaline co-precipitation synthesis process, the nucleation is an unending phenomenon. This was observed by Baumgartner et al. with the help of real-time TEM. They reported that in such synthesis processes the nanostructures are formed due to the aggregation of the primary particles rather than atomic accretion, hence suggesting a secondary nucleation and growth mechanism 40 .</p><!><p>In this section, the hardware and software realization of INSPECT is described. The diameter of the measurement chamber is 72 mm and can fit a 100 ml reaction flask for synthesis (see Fig. 1). The complexity of the hardware modules is quite modest compared to that of recently described MPS systems [21][22][23] .</p><p>signal generation and acquisition. For the generation of the transmitted signal T x and acquisition of the received signal R x , an X3-25M (Innovative Integration, USA) is used 41 . The X3-25M features two 16-bit ADC (analog to digital converter) channels and two 16-bit DAC (digital to analog converter) channels. Moreover, the ADC has programmable input ranges and a bandwidth of 75 MHz. The DAC channels have an output of ±2V.</p><p>www.nature.com/scientificreports www.nature.com/scientificreports/ One of the DAC channel is used for generating the transmitted signal for the transmitting coil. The resulting voltage across the transmitting coil is fed back to the I/O card to control the flux density inside the measurement chamber. The X3-25M generates the sinusoidal T x signal at 23 kHz with a sampling frequency of 5.06 MSPS (mega samples per second), ensuring a high signal quality. R x is also acquired with the X3-25M at a sampling rate of 5.06 MHz. Both signals are synchronized with the help of the timing unit present on the I/O card itself. The total measurement time for one measurement is about 1 s acquiring 10 periods with 2300 averages. The same measurement parameters are used for analyzing the sensitivity of INSPECT as well as for monitoring the synthesis process described in results and discussion.</p><!><p>The T x signal generated by the I/O card has to be amplified by a power amplifier (AE Techron, USA) to provide the necessary magnetic field of up to 10 mT 42 . This amplifier has a low total harmonic distortion (THD) of less than 0.1% at f 0 = 23 kHz (fundamental frequency) with a maximum gain of 20. This reduces the need for analog filter stages, which are usually employed for the send chain in a conventional MPS setup. To ensure that the maximum power is transferred to the transmitting coil and to minimize the reflections between the AC power amplifier and the transmitting coil, a capacitive impedance matching is used. The circuit diagram for the impedance matching and the overall design is shown in Fig. 2. The feedback consists of a voltage divider and an amplifier to measure the changes in the magnetic flux density.</p><p>transmitting coil and gradiometer coil. The transmitting coil is responsible for the generation of the required field to excite the SPIONs. The gradiometer coil cancels the transmitting signal out and provides the response of the nanoparticles for further analysis. The entire design of the transmitting coil and the gradiometer coil with the housing is shown in Fig. 3.</p><p>Transmitting coil. The transmitting coil generates a magnetic field and consists of a solenoid. The outer diameter of the transmitting coil is 121 mm and the inner diameter of the coil is 92 mm, which can easily fit a glass flask of 100 ml for synthesis. The height of the transmitting coil is 93.75 mm. The transmitting coil consists of 40 × 4 windings realized with 2000 × 5 μm Litz wire. The inductance of the coil is approximately 229 μH at 23 kHz (which is the excitation frequency of INSPECT). The air cooled transmitting coil is able to produce a field of up to 10 mT.</p><p>Gradiometer coil. The receiving coil lies inside the transmitting coil, which leads to higher mutual inductances causing a direct feed through of the excitation signal to the receiving coil. This excitation signal is usually 5-10 times higher than the nanoparticle signal and therefore has to be damped in the receiving channel. Most of the state of the art MPS devices rely on a compensation coil or bandstop filters to dampen the excitation signal in the receiving channel 22 . Other techniques have been employed such as a dedicated cancellation unit 21 and perpendicular sensing 43 . A comparison between some of these techniques can be found in Graeser et al. 44 . For the MPS presented here, a one-dimensional gradiometer coil is used, which is a receiving coil including a cancellation coil (with reverse winding). Gradiometer coils have been used in a number of fields like for constructing magnetometers 45 , in geology 46 as well as in medical science 47 . For fine tuning, the gradiometer coil holder, as well as the flask holder (shown in Fig. 3), can be mechanically adjusted to achieve the maximum attenuation of the excitation signal in the receiving channel. The gradiometer coil consists of pure copper wire with a diameter of 0.50 mm and has 30 turns in the receiving coil and approximately 28 windings in the cancellation coil and has a diameter of 74 mm. The CAD drawing of the entire setup is shown in Fig. 3.</p><!><p>The housing for the transmission coil and the gradiometer coil is manufactured with a ProJet 3510 HDPlus 3D printer (3D Systems, USA). The printed material can withstand temperatures up to 80 °C. The cooling of the coils is realized by pressurized air. Through the cooling channel the air is channeled over the www.nature.com/scientificreports www.nature.com/scientificreports/ Calibration. As the signal could lose its fidelity while passing through the receiving chain or the sending chain, it is important to calibrate the system. As INSPECT has no filters or pre-amplifiers in either the sending chain or the receiving chain, the transfer functions of the chains are free of heating effects and can easily be determined.</p><p>Transmission chain calibration. For calibration of the transmitting chain, a transmitting calibration coil (u Tcal ) made up of a copper wire and a voltage divider circuit is used. This coil is mounted on a cylinder of 51 mm in diameter with 21 windings of 0.40 mm pure copper wire. The transmitting calibration coil (u Tcal ) detects the magnetic field generated by the transmitting coil. With the known geometry and the received voltage, the magnetic field can be easily calculated by using the Faraday's law of induction. The applied magnetic field is given by</p><p>Tx Tcal 0 where B Tx (t) is the generated magnetic flux density, f 0 is the fundamental frequency of 23 kHz, A is the cross-section area, and N is the number of turns in the transmitting calibration coil. Using Eq. 7 the field strength and phase of the exciting magnetic field can be determined.</p><!><p>In general, components such as filters and amplifiers add distortions to the signal detected at the receive chain. correct for these distortions, the final measurement has to be corrected by the transfer function of the receiving chain. In the current research the transfer function is obtained with a 20 mm receive-chain calibration coil made by pure copper wire of 0.5 mm thickness having 5 windings. The coil is connected to a 500 Ω series resistor to measure the current. The phase and magnitude of the measured transfer function are shown in Fig. 4. As the absence of filters and amplifiers in our receive chain implies, the transfer function is straight in a wide frequency range.</p><p>spIoN synthesis. For the synthesis of SPIONs there are a number of methods, but usually, in biomedical science, the three most investigated processes are alkaline precipitation in water, water-in-oil micro-emulsions and the thermal decomposition of organometallic iron in organic solvents 49 . For the validation of INSPECT, the alkaline co-precipitation in water is used, which is the most simple and common method, however leading to a broad core-size distribution. This process comprises of two stages, in the first stage there is a nucleation, which is followed by a slow growth in the second stage. The chemical reaction for the formation of the iron oxide can be divided into two steps: For the synthesis the iron salts, iron (III) (1.32 g of FeCl 3 • 6H 2 O, ≥99% obtained, Carl Roth GmbH Karlsruhe, Germany) and iron (II) (0.50 g of FeCl 2 • 4H 2 O, ≥99% obtained, Merck kGaA, Darmstadt, Germany) plus Dextran T70 (1.32 g, AppliChem GmbH, Darmstadt, Germany) are placed in a round bottle flask inside INSPECT and ammonia as a base is dropped into the solution. The flow rate of the base is controlled with an infusion pump (PERFUSOR secura FT, B. Brown), which is set to 20 ml/h. After the addition of the base, the mixture is slowly heated to a temperature around 80 °C under ultrasonic control 50 , where the growth of SPIONs takes place.</p><!><p>Before initiating the synthesis process, an empty measurement is taken, which is subtracted from the subsequent measurements to correct the background noise. The second measurement is taken after putting the iron salts and dextran and water in the flask. Then, 25 ml of base (ammonia (NH 3 ) for the current research) is dropped controlled by an infusion pump at 20 ml/hr. After the addition of all the required</p><!><p>In this section, initial results of synthesis monitoring with INSPECT are presented and discussed. The change in the magnetic properties during the synthesis is demonstrated. The measurements consist of both the amplitude and the phase spectra and the hysteresis curve at different stages of the synthesis. Hysteresis curves are derived by measuring the magnetization of the nanoparticles depending on the dynamic excitation field. The dynamic excitation field is monitored using the feedback signal consisting of the voltage divider as shown in Fig. 2. The minimum measurement time possible is the length of one period of the excitation field, which in this case is approximately 42.9 μs.</p><!><p>SPIONs in suspension. However, it is crucial that the device can also detect small concentrations of SPIONs, because the SPIONs have a small size in the nucleation phase of the synthesis process. In this section, results of an experiment are presented that demonstrates the lowest quantity of SPIONs detectable with INSPECT. For this particular analysis, commercially available SPIONs (Resovist, Bayer, Berlin) are used. Resovist is a multicore particle with an iron oxide core consisting of multiple single crystals of approximately 4.2 nm in diameter 51 . The spectral response of the Resovist samples are compared to an empty measurement after subtracting the background. Figure 5 shows the background and the amplitude spectrum of Resovist. The empty measurement is shown in red and the signal of a 10 μL (containing 0.28 mg of iron) Resovist sample is shown in black. INSPECT is able to detect the first 6 odd harmonics. This ensures that the device is capable of detecting small quantities of SPIONs. Furthermore, even harmonics are also visible in the amplitude spectrum, this is due to the presence of a slight DC offset in the produced field.</p><p>In-situ magnetic particle measurements during particle synthesis. Figure 6 shows the change of magnetic properties in the colloidal solution versus time of the entire synthesis process. Starting from third harmonic to ninth harmonic are plotted versus time. For simplification and without loss of generalization, the discussion of the results is limited to the temporal change of magnetic moment in the third harmonic as the third harmonic has the maximum power in comparison to subsequent harmonics.</p><p>This particular synthesis can be divided into three different intervals. The first interval consists of the nucleation process, in which, for a long period of time, there is no or negligible formation of superparamagnetic material. As stated by Baumgartner et al. secondary nucleation and growth mechanisms widely occur in the co-precipitation synthesis process 40 . Therefore, it is very difficult to estimate when the nucleation phase has completely ended and growth has started.</p><p>This interval ends with the initial particle formation, which shows a weak magnetic moment. In the second interval, a sudden growth of the nanoparticles occurs. In a span of 30 min the magnetic moment increases from 5.32 × 10 −7 Am 2 to 1.99 × 10 −5 Am 2 , which is approximately a factor of 37.4. This happens because the nuclei that had already formed in the nucleation phase reach a reasonable size to show a magnetic moment. In the last interval, there is a very slow growth and a small change by a factor of 1.45 in the magnetic moment from www.nature.com/scientificreports www.nature.com/scientificreports/ 1.99 × 10 −5 Am 2 to 2.887 × 10 −5 Am 2 . At this point, the synthesis process is stopped as no further changes in the magnetic properties occur. For this particular synthesis, 46 measurements are taken in the total time span of approximately 225 min. In Fig. 6 specific measurement points are indicated (marked with red dots) that will be further discussed in detail.</p><p>Nucleation phase. In the nucleation phase, there is no visible change in the magnetic properties of the nanoparticles as it takes some time for the solution to reach the supersaturation state to start the nucleation process. The nucleation occurs only when the concentration of the growth species increases above the equilibrium concentration, i.e. the concentration of the base increases in the solution. The base is being dropped slowly, therefore it takes around 80 min for completion. The slow addition of the base is important, because a sudden increase in the concentration of the base may lead to the formation of different sizes of nuclei and hence, leads to a broader size distribution 52 . Some of the measurements taken during this interval are shown in Fig. 7. The points at which these measurements are taken within the synthesis process are indicated in Fig. 6 (red dots with time stamps). Figure 7 shows the amplitude and phase spectra, and the magnetization curve at the different time points. The first time point comprises of two measurements in Fig. 7 at 0 min, the first one is an empty measurement (marked in blue). In the second measurement water and the flask containing all the reagents i.e. iron salts and dextran have been measured (marked in red). The diamagnetic behavior of water can be seen in the magnetization curve 53 . Followed by measurements after addition of base, which leads to the initialization of the nucleation. However, there are no significant changes in the magnetic properties observed, as supersaturation occurs locally, at the position where the base is being dropped. However, after a few seconds of ultrasonic stirring, the base is mixed into the solution. As shown at 0 min, there is a negligible change in the amplitude of the signal received, but there is a prominent change in the phase and the shape of magnetization curve for the subsequent measurements at 50 min. This process continues for measurements till 50 min and 90 min, leading to phase changes and shifting of the magnetization curve. In the measurements at 50 min and 90 min, respectively, it can be seen that the magnetization curve is changing. Moreover, the amplitude of the third harmonic increases for the measurement at the time stamp at 50 min to measurement at 90 min. In the nucleation phase, the detection of SPIONs is challenging as the magnetic moments of the particles are small. Sudden growth phase. When the required quantity of the base is added to the solution, the heating is started. This leads to a decrease in the concentration of the growth species and the change in the volume Gibbs free energy Figure 6. The magnetic moment of the third, fifth, seventh, and ninth harmonics versus time of the entire synthesis process. The synthesis process can be subdivided into three different intervals. The first interval is the nucleation phase, where only a slight change in the magnetic properties can be seen over time. This is followed by a sudden growth phase. The last interval is the growth phase, where a slow linear particle growth takes place which converges into a static behavior. The synthesis process is stopped in this phase. The measurements marked with red points will be discussed in the next figures in detail. The heating phase is also indicated, which marks the time period in which the solution is heated up to 80 °C. Before the heating phase, the base is dropped into the solution containing the iron salts and dextran.</p><p>www.nature.com/scientificreports www.nature.com/scientificreports/ decreases. Ideally, this should not lead to further nuclei formation and a coherent growth process continues until the concentration of growth species reaches the equilibrium concentration 39 . For alkaline based synthesis this is not the case. Figure 8 shows the measurements taken in this phase. These measurements are marked in Fig. 6 and are acquired at 135 min, 165 min, and 175 min, respectively. The measurement at the time stamp 135 min is acquired before the sudden growth initiates. The amplitude and phase spectra, and the magnetization curve obtained are shown in Fig. 8. The spectral magnetic moment indicates that superparamagnetic particles reached the targeted size in the solution, as the higher odd harmonics appear in the spectrum, which is a prerequisite for a good resolution for MPI 3 . The rows comprise the amplitude and phase spectra, and the magnetization curves at different time intervals. The first row, acquired at time stamp 0 min, includes two measurements, one is the empty measurement (marked in blue color) and the other one with the glass flask containing the reagents (iron salts, dextran) and water inside the chamber (marked in red color). The other measurements are performed at time stamp 50 min, 90 min, and 110 min, respectively. As all the above measurements have been acquired in the nucleation phase of the synthesis process, for the amplitude spectrum there are no significant changes till 90 min but at 110 min the fourth odd harmonic appears. There are significant changes in the phase spectra, and, in turn, in the resulting magnetization curves. Due to small amplitudes, the phase measurements are inaccurate and fluctuate. However, the area under the magnetization curve is maximum for the measurement at time stamp 110 min. (2019) 9:10538 | https://doi.org/10.1038/s41598-019-46882-6 www.nature.com/scientificreports www.nature.com/scientificreports/ The next measurement is taken at 165 min. In the amplitude spectrum, the first nine odd harmonics can be observed and the phase spectrum shows linearity. The magnetization curve shows no saturation at field strengths of 10 mT. This situation changes in the following 10 min as shown in measurement at the time stamp at 175 min. There is a sudden growth of the particles, which significantly increases the magnetic moment. The magnetization curve also shows a saturation effect indicating larger particles and the phase spectrum is almost constant.</p><p>Growth phase. The third interval is the growth phase, where the SPIONs show a linear growth behavior. The results are shown in Fig. 9. There is an increase in the area of the magnetization curve over time. The measurements have been acquired after 205 min and 225 min after the synthesis initiation. The measurement at the time stamp 225 min marks the termination of the synthesis process as no significant change in the magnetic properties over time can be observed.</p><!><p>An in-situ MPS is presented, which is capable of tracking the growth of the nanoparticles in an ongoing synthesis process. The hardware realization is compact, consists of few hardware components, and is well suited for the chemical laboratory. The whole setup consists of a housing consisting of a transmitting coil and gradiometer coil. In addition, there is an impedance matching module and a power amplifier. There are no additional band pass or band stop filters for either the transmitting signal or the receiving signal. The excitation frequency can easily be changed by switching the capacitors in the impedance module. As the gradiometer coil is vertically movable, it can be adapted to receive the maximum signal and to cancel out the excitation signal.</p><p>The measurements presented here have been acquired every 5 min, however, the sampling rate is adaptable to real-time measurements with increasing cooling effort. To demonstrate the usefulness of the device a complete 8. Sudden growth phase. The measurements comprise of the amplitude and phase spectra, and the magnetization curve, respectively. The first row is acquired when the growth is initiated at the time stamp at 135 min. In comparison to the measurements in Fig. 7, there is a significant increase in the amplitude of the odd harmonics. This directly implies that there is further growth in the particles, which are formed in the nucleation phase. The measurements at the time stamp 165 min and 175 min show harmonics up to 500 kHz in the spectral magnetic moments, which confirms the particle growth. At time stamp 175 min, the magnetization curve indicates the beginning of the saturation effect. Furthermore, with the progression of the synthesis the phase becomes much more stable. As it can be seen by comparing the phase spectra at time stamp 135 min to the time stamp at 175 min. alkaline precipitation based synthesis process of SPIONs has been monitored with INSPECT. A series of data sets have been acquired over a time span of 225 min. From the results it can be deduced that the synthesis process can be divided in three main intervals, i.e. the nucleation phase, the sudden growth phase, and the growth phase. In the nucleation phase there is no significant change in the amplitude spectrum. In the nucleation phase, there is a decrease in the coercivity for the measurement at time stamp 90 min to 55 min. As in the nucleation phase there are a lot of competing processes happening such as particle coarsening (also called Ostwald ripening) and digestive ripening 54 . Ostwald ripening leads to dissolving of the smaller particles back to solution and a redeposition on the larger particles. On the other hand, digestive ripening leads to shirking of the larger particles leading to further growth of the smaller particles. These processes can cause changes in the Brownian relaxation and in turn affect the magnetic moment of the particles. With INSPECT we will be able to study these phenomena in more detail. In the sudden-growth interval, as the SPIONs become bigger in size, the spectral magnetic moment changes significantly. This phase ends in the growth interval, where particles grow slowly and finally due to no significant changes in the spectral magnetic moment, the synthesis process is stopped. INSPECT is able to monitor these changes and can be used to optimize the synthesis process, because it allows for an insight into the growth dynamics and reveals the direct effects of synthesis parameter tuning. This device provides a platform to study the dynamics of different synthesis processes to improve SPIONs for applications in both MRI and MPI.</p>

Scientific Reports - Nature

Signaling pathways controlling the phosphorylation state of WAVE1, a regulator of actin polymerization

WAVE1 (the Wiskott-Aldrich syndrome protein (WASP)-family verprolin homologous protein 1) is a key regulator of Arp (actin-related protein) 2/3 complex-mediated actin polymerization. We have established previously that the state of phosphorylation of WAVE1 at three distinct residues controls its ability to regulate actin polymerization and spine morphology. Cyclin-dependent kinase 5 (Cdk5) phosphorylates WAVE1 at Ser310, Ser397 and Ser441 to a high basal stoichiometry, resulting in inhibition of WAVE1 activity. Our previous and current studies show that WAVE1 can be dephosphorylated at all three sites and thereby activated upon stimulation of the D1 subclass of dopamine receptors and of the NMDA subclass of glutamate receptors, acting through cAMP and Ca2+ signaling pathways, respectively. Specifically we have identified PP-2A and PP-2B as the effectors for these second messengers. These phosphatases act on different sites to mediate receptor-induced signaling pathways, which would lead to activation of WAVE1.

signaling_pathways_controlling_the_phosphorylation_state_of_wave1,_a_regulator_of_actin_polymerizati

INTRODUCTION<!>Preparation of striatal slices<!>Chemicals and reagents<!>Immunoblotting<!>Data Analysis<!>Forskolin-induced WAVE1 dephosphorylation<!>Role of PP-2A in forskolin-induced WAVE1 dephosphorylation<!>Role of PP-1 in forskolin-induced WAVE1 dephosphorylation<!>Role of PP-2B in forskolin-induced WAVE1 dephosphorylation<!>Role of PP-2A and PP-2B in NMDA-induced WAVE1 dephosphorylation<!>DISCUSSION<!>

<p>Actin plays critical roles in various membrane-cytoskeleton-coupled processes such as cell movement, vesicular trafficking events including endocytosis and exocytosis, cytokinesis and intracellular movement of pathogens (Rottner et al. 2004; Lanzetti 2007; Heasman and Ridley 2008). In neurons, reorganization of the actin cytoskeleton is required for neurite extension, axonal guidance, cycling of neurotransmitter vesicles, dendritic spine formation and synaptic plasticity (Luo 2002; Dillon and Goda 2005). Abnormal regulation of the actin cytoskeleton is associated with mental retardation (Newey et al. 2005) and cognitive deficits (Frangiskakis et al. 1996) as well as with neurodegenerative diseases (Minamide et al. 2000; Zhao et al. 2006).</p><p>Reorganization of the actin cytoskeleton is tightly controlled by various regulatory proteins that govern uncapping, severing and filament formation (Pollard et al. 2000). The WASP family proteins use their C-terminal VCA domain to stimulate the Arp2/3 complex to nucleate the de novo synthesis and branching of actin filaments (Takenawa and Suetsugu 2007). WAVE1, a member of the WASP family, is abundant in brain, where its highest levels are found in cerebral cortex, hippocampus, amygdala and striatum (Dahl et al. 2003; Soderling et al. 2003). WAVE1 is critical for the development and function of the central nervous system (CNS). WAVE1-null mice show CNS-related problems such as limb weakness, neuroanatomical malformations and behavioral abnormalities including reduced anxiety, sensorimotor retardation, and deficits in hippocampal-dependent learning and memory (Dahl et al. 2003; Soderling et al. 2003). Homozygote WAVE1 knockout mice also exhibit reduced body size and reduced viability.</p><p>In neurons, WAVE1 is localized to dendrites and dendritic spines (Pilpel and Segal 2005; Kim et al. 2006b; Soderling et al. 2007; Sung et al. 2008). WAVE1 is also localized to axonal growth cones (Nozumi et al. 2003; Soderling et al. 2007) as well as to mitochondria (Cheng et al. 2007; Sung et al. 2008). As a result, WAVE1 plays critical roles in growth cone dynamics, neurite outgrowth, dendritic spine morphogenesis and synaptic plasticity (Kim et al. 2006b; Soderling et al. 2007). WAVE1 also mediates neuronal activity-induced mitochondrial trafficking to dendritic spines and spine morphogenesis (Sung et al. 2008). Furthermore WAVE1 is localized to oligodendrocytes and plays a role in CNS myelination (Kim et al. 2006a).</p><p>Previously, we have identified WAVE1 as a novel target of p35/Cdk5 (Kim et al. 2006b). WAVE1 is phosphorylated at multiple sites by Cdk5 in vitro and in intact mouse neurons. Phosphorylation of WAVE1 by Cdk5 inhibits its ability to regulate Arp2/3 complex-dependent actin polymerization. In brain, WAVE1 is basally phosphorylated with high stoichiometry but the level of phosphorylation is reduced by stimulation of striatal slices with a dopamine D1 agonist or with forskolin, a stimulator of adenylyl cyclase, both of which elevate cAMP levels. Thus WAVE1 is largely in an inactive form under basal conditions, but can be activated by neurotransmitters such as dopamine, that increase the levels of cAMP. Previously we also observed a critical role for NMDA receptor-dependent signaling in WAVE1 dephosphorylation following repetitive depolarization of primary cortical neurons (Sung et al. 2008). Thus, phosphorylation and dephosphorylation of WAVE1 are both likely to be important mechanisms involved in the regulation of actin polymerization and, in turn, of neuronal function.</p><p>The molecular mechanisms that mediate neuronal stimulation-induced WAVE1 dephosphorylation have not been investigated. The cAMP-mediated reduction in phosphorylation of WAVE1 could be caused by inhibition of kinases or stimulation of phosphatases. However, activation of the cAMP pathway had no effect on Cdk5 activity (Kim et al. 2006b) suggesting the involvement of protein phosphatases in cAMP-mediated WAVE1 dephosphorylation. In the present study, we have investigated the effect of specific inhibitors of protein phosphatases on cAMP or NMDA receptor-induced WAVE1 dephosphorylation. We have also analyzed the phosphorylation of WAVE1 in DARPP-32 (Dopamine and adenosine 3',5'-monophosphate-regulated phosphoprotein, 32 kilodaltons) knockout mice (Fienberg et al. 1998) and in RCS (regulator of calmodulin signaling ) knockout mice (Rakhilin et al. 2004), to evaluate the role of protein phosphatase 1 (PP-1) and PP-2B, respectively, in the dephosphorylation of WAVE1. The results obtained indicate that both PP-2A and PP-2B are major WAVE1 phosphatases, which differentially mediate receptor-mediated dephosphorylation of the various sites in WAVE1.</p><!><p>Male C57BL/6 mouse (6–8 weeks old) brains were quickly removed following rapid decapitation and placed in ice-cold, oxygenated Krebs-HCO3− buffer (124 mM NaCl, 4 mM KCl, 26 mM NaHCO3, 1.5 mM CaCl2, 1.25 mM KH2PO4, 1.5 mM MgSO4 and 10 mM d-glucose, pH 7.4). Coronal slices (350 µm) were prepared using a vibrating blade microtome, VT1000S, (Leica Microsystems, Nussloch, Germany). Striata were dissected from the slices in ice-cold Krebs-HCO3− buffer. Each slice was placed in a polypropylene incubation tube with 2 mL of fresh Krebs-HCO3− buffer. The slices were pre-incubated at 30°C under constant oxygenation with 95% O2/5% CO2 for 60 min. The buffer was replaced with fresh Krebs-HCO3− buffer after 30 min of pre-incubation. Slices were treated with either forskolin (1 µM) for 30 min or NMDA (100 µM) for 10 min in the absence or presence of various drugs as specified in each experiment. DMSO (2 µL/2 mL) was used as a vehicle for forskolin, okadaic acid, cyclosporin A and tautomycetin. After drug treatment, slices were transferred to micro centrifuge tubes, frozen on dry ice, and stored at −80°C until assayed.</p><!><p>Drugs were obtained from the following sources: okadaic acid (OKA) and NMDA from Sigma (St Louis, MO, USA); cyclosporin A (CyA) from Tocris (Ellisville, MO, USA); forskolin from Alexis Biochemicals (Lausen, Switzerland), and tautomycetin from Calbiochem (San Diego, CA, USA).</p><!><p>Frozen tissue samples were sonicated in boiling 1% SDS and boiled for an additional 10 min. Small aliquots of the homogenate were retained for protein determination by the BCA protein assay method (Pierce, Rockford, IL) using bovine serum albumin as a standard. Equal amounts of protein (10 µg) were separated by SDS-PAGE (4–20% polyacrylamide gels) and transferred to nitrocellulose membranes. The membranes were immunoblotted using anti-Cdk5 (C-8, 1:2,000, Santa Cruz), anti-p35 (C-19, 1:2,000, Santa Cruz), anti-WAVE1 (C-terminus, 1:10,000) antibodies and phosphorylation state-specific antibodies for phospho-Ser310 (1:10,000), -Ser397 (1:5,000) or -Ser441 (1:1,000) (Kim et al. 2006b). Antibody binding was detected using the enhanced chemiluminescence (ECL) immunoblotting detection system or the Odyssey infrared imaging system (LI-COR, Lincoln, NE, USA). For the ECL immunoblotting detection system, membranes were incubated with horseradish peroxidase-linked goat anti-rabbit IgG antibody (1:10000) (Pierce, Rockford, IL, USA). Chemiluminescence was detected by autoradiography using Kodak autoradiography film, and phospho-WAVE1 bands were quantified by densitometry, using NIH Image 1.63 software. For the Odyssey infrared imaging system, membranes were incubated with an IRDYE 800CW-conjugated goat anti-rabbit IgG antibody (1:7500) (LI-COR, Lincoln, NE, USA). Fluorescence at infrared wavelengths was detected by the Odyssey Imager, and quantified using Odyssey software (LI-COR).</p><!><p>The data are expressed as means ± standard error (SE) of the means and were analyzed using statistical methods as described in the figure legends.</p><!><p>Our previous studies showed that, under basal conditions, WAVE1 in mouse striatal slices was highly phosphorylated at Ser310, Ser397 and Ser441 by Cdk5 (Kim et al. 2006b). Treatment of slices with a dopamine D1 receptor agonist or with forskolin for 30 min induced dephosphorylation of WAVE1, indicating a role for cAMP signaling in the dephosphorylation of WAVE1 (Kim et al. 2006b). Stimulation of striatal slices with a dopamine D2 receptor agonist did not induce dephosphorylation (not shown). In contrast, there was a trend for a slight increase in phosphorylation at Ser310 and Ser397 (-fold change after the treatment with 1 µM quinpirole for 30 min; Ser310, 1.19±0.12; Ser397, 1.29±0.27; Ser441, 1.07±0.18; n=5, P>0.05 for all sites compared to control). These results indicate the specific role of dopamine D1 receptor signaling in the dephosphorylation of WAVE1.</p><p>We examined the time-course of the effect of forskolin on WAVE1 dephosphorylation in mouse striatal slices. Stimulation of slices with forskolin (1 µM) significantly decreased WAVE1 phosphorylation at Ser310 and Ser397 after 15 min of incubation and this dephosphorylation was sustained for at least 30 min (Fig 1D,E). Forskolin treatment also resulted in dephosphorylation at Ser441 but this was not observed until 30 min of incubation (Fig.1F). No difference in total levels of WAVE1 was found (Fig.1C). Moreover, the levels of p35 and CdK5 were not altered by forskolin stimulation (Fig. 1A,B), which is consistent with our previous observation that Cdk5 activity was unaffected by treatment with forskolin (Kim et al. 2006b). These results suggest that the decrease in WAVE1 phosphorylation does not stem from a down-regulation of Cdk5/p35. Instead, protein phosphatases are likely to mediate the cAMP-dependent dephosphorylation of WAVE1. This hypothesis is supported by the relatively slower onset of the effect of forskolin on WAVE1 dephosphorylation compared to those on cAMP-dependent phosphorylation of various substrates, such as Thr34 of DARPP-32 in striatal slices. Therefore, we investigated the possible involvement of three major serine/threonine protein phosphatases in brain (PP-1, PP-2A and PP-2B) in the cAMP-induced dephosphorylation of WAVE1.</p><!><p>Okadaic acid (OKA) is a preferential inhibitor of PP-2A but can inhibit PP-1 at higher concentrations. Our previous studies in striatal slices indicated that 1 µM OKA potently inhibited the activity of PP-2A, while this concentration of OKA reduced PP-1 activity only by ~33% (Nishi et al. 1999). In order to determine the appropriate concentration of okadaic acid for experiments designed to evaluate the role of PP-2A in forskolin-induced WAVE1 dephosphorylation, mouse striatal slices were incubated with various concentrations of OKA for 60 min, and the levels of phosphorylation of the different sites in WAVE1 were measured (Fig. 2A). OKA increased the levels of WAVE1 phosphorylation at Ser310 and Ser441 in a dose-dependent manner with a maximal effect at 0.5–1 µM but had no effect on phosphorylation at Ser397 (Fig. 2A). Our previous studies showed that the basal stoichiometry of Ser397 in striatum is relatively high (~0.85) compared to the stoichiometries of Ser310 (~0.58) and Ser441 (~0.27) (Kim et al. 2006b). It seems likely that we did not see any increase in the phosphorylation at Ser397 in response to incubation with OKA because phosphorylation of this site was close to maximal under basal conditions.</p><p>We next examined the role of PP-2A in mediating the effect of forskolin on WAVE1 dephosphorylation. Pretreatment with OKA (1 µM) elevated the basal level of phosphorylation at Ser310 and Ser441 (1.7 ± 0.2 and 1.6 ± 0.29 fold compared to control, respectively), but not at Ser397 (Fig. 2B–D). Pretreatment with OKA completely blocked forskolin-induced WAVE1 dephosphorylation at Ser310 and Ser397. However, forskolin treatment still resulted in a significant reduction in phosphorylation of Ser441 when compared to treatment with OKA alone. Thus, PP-2A likely mediates cAMP-induced dephosphorylation at Ser310 and Ser397. Although PP-2A regulates the basal level of phosphorylation at Ser441, its contribution to cAMP-induced dephosphorylation at this site appears negligible.</p><!><p>To investigate whether PP-1 is at all involved in WAVE1 dephosphorylation, striatal slices were incubated with tautomycetin, which inhibits PP-1 more potently than PP-2A (Mitsuhashi et al. 2001; Mitsuhashi et al. 2003). Treatment of slices with tautomycetin (5 µM) alone for 60 min did not affect the basal level of WAVE1 phosphorylation. Tautomycetin did not affect forskolin-induced WAVE1 dephosphorylation either at Ser310 or Ser397 (Fig. 3A,B). With respect to Ser441, tautomycetin appeared to have a small inhibitory effect on forskolin-induced dephosphorylation at this site but this did not reach significance (Fig. 3C). We also studied WAVE1 phosphorylation in slices from mice in which DARPP-32, a well-characterized inhibitor of PP-1 that is enriched in striatum, was knocked out (Hemmings et al. 1984; Greengard et al. 1999). Previously the level of phosphorylation of PP-1 substrates was shown to decrease in DARPP-32 knockout mice, due to reduced inhibition of PP-1 (Fienberg et al. 1998; Snyder et al. 1998). Forskolin-induced dephosphorylation was not altered at any of the three sites in slices from DARPP-32 knockout mice compared to their wild-type controls (Fig. 3D–F). Together, these results suggest that PP-1 does not play a major role in WAVE1 dephosphorylation.</p><!><p>To examine the role of PP-2B in forskolin-induced WAVE1 dephosphorylation, slices were incubated with a specific inhibitor of PP-2B, cyclosporin A (CyA, 10 µM) for 60 min in the absence or presence of forskolin. Pretreatment with cyclosporin A did not affect the basal level of WAVE1 phosphorylation at any of the three sites but was able to block forskolin-induced dephosphorylation at Ser441 (Fig. 4C). Forskolin-induced dephosphorylation at Ser310 and Ser397 was not affected by the pretreatment with CyA (Fig. 4A,B). We also examined the phosphorylation of WAVE1 in RCS (regulator of calcium/calmodulin-dependent signaling) knockout mice to further address the role of PP-2B. Like DARPP-32, RCS is enriched in striatum where G protein-coupled receptor (GPCR)-dependent activation of protein kinase A (PKA) leads to phosphorylation of RCS at Ser55 and increases its binding to calmodulin (CaM) (Rakhilin et al. 2004). Phospho-RCS can then act as a competitive inhibitor of CaM-dependent enzymes, including PP2B. Increasing RCS phosphorylation was found to block GPCR- and PP2B-mediated suppression of L-type Ca2+ currents in striatal neurons (Rakhilin et al. 2004). Conversely, genetic deletion of RCS significantly increased this modulation. We found no difference in forskolin-induced WAVE1 dephosphorylation at Ser310 in striatal slices from wild-type and RCS knockout mice (Fig. 4D). We observed a small, but not statistically significant, increase in forskolin-induced dephosphorylation at Ser397 and Ser441 in slices from RCS knockout mice compared to their wild-type controls (Fig. 4E,F).</p><!><p>We next examined the role of striatal NMDA receptor signaling on WAVE1 dephosphorylation in mouse striatal slices. Stimulation with NMDA (100 µM) significantly decreased the level of phosphorylation at all three WAVE1 sites within 5 min, and the dephosphorylation was sustained for at least 30 min (Fig. 5D–F). NMDA treatment in striatal slices did not result in the down-regulation of p35 or CdK5 (Fig. 5A,B) and the level of total WAVE1 remained unaltered (Fig. 5C). These results indicate that protein phosphatases are likely to be involved in NMDA-induced dephosphorylation of WAVE1, as in the case with cAMP-mediated WAVE1 dephosphorylation.</p><p>To examine the mechanism by which NMDA reduced the level of WAVE1 phosphorylation, NMDA-induced WAVE1 dephosphorylation was measured in the presence or absence of OKA or CyA (Fig. 6). Pretreatment of striatal slices with OKA (1 µM) for 60 min blocked the effect of NMDA on WAVE1 dephosphorylation at Ser310 and Ser441, but not at Ser397 (Fig. 6A–C). Pretreatment of slices with CyA (10 µM) for 60 min significantly blocked the effect of NMDA on WAVE1 dephosphorylation only at Ser397 (Fig. 6D–F). Tautomycetin did not affect NMDA-induced WAVE1 dephosphorylation at any of the three sites (data not shown). These results suggest that NMDA-induced WAVE1 dephosphorylation at Ser310 and Ser441 is mediated by PP-2A, whereas WAVE1 dephosphorylation at Ser397 is mediated by PP-2B.</p><!><p>The function of WAVE1 in actin polymerization and dendritic spine formation is largely suppressed by phosphorylation (Kim et al. 2006b). Under basal conditions, WAVE1 is phosphorylated by Cdk5 to a high stoichiometry (~60 to 90% depending on the site of phosphorylation). Our current studies have revealed a complex pattern of regulation of WAVE1 in response to activation of PP-2A and PP-2B acting at the three Cdk5 sites of WAVE1 (Fig. 7). We find that protein dephosphorylation is mediated by both PP-2A and PP-2B acting at the three Cdk5 sites of WAVE1, which would lead to WAVE1 activation. In brain, neurotransmitters trigger the rearrangement of actin filaments presumably through receptor-mediated signal transduction mechanisms involving WASP/WAVE family proteins. This actin rearrangement plays pivotal roles in axonal and dendritic development, synapse formation and plasticity. The current studies have thus identified novel neurotransmitter-mediated signaling pathways that are linked to activation of WAVE1.</p><p>In striatal slices, dopamine acting upon D1 receptors increases cAMP level through activation of receptor-coupled Gsα and adenylyl cyclase. In the current studies we utilized forskolin to directly activate adenylyl cyclase to increase cAMP signaling. Forskolin led to the dephosphorylation of Ser310, Ser397 and Ser441, with PP-2A being involved in dephosphorylation of Ser310 and Ser397, while PP-2B is involved in dephosphorylation of only Ser441 (Fig. 7). PP-2A is composed of catalytic C subunit, scaffolding A subunit and regulatory B subunit (Virshup and Shenolikar 2009). There are at least 15 different B subunit isoforms that are generated from four B subunit gene families, and there are many multiple splicing variants. The B subunits are critical for controlling localization, substrate specificity and catalytic activity of the C subunit. Our previous studies have shown that one particular B subunit, B56δ, was enriched in striatum, could be phosphorylated and activated by PKA, and that the PP-2A heterotrimeric complex containing B56δ was responsible for dopamine-mediated dephosphorylation of DARPP-32 at Thr75 (Ahn et al. 2007a). Given the similar features of phosphorylation and dephosphorylation between WAVE1 (Kim et al. 2006b) and DARPP-32 at Thr75 (Bibb et al. 1999; Nishi et al. 2000), it seems likely that PP2A containing the B56δ subunit may mediate cAMP-induced WAVE1 dephosphorylation. The discovery that cAMP could lead to increased dephosphorylation of DARPP-32 was somewhat unexpected, but the results from the present study support the idea that cAMP-mediated dephosphorylation may be a common process that affects multiple proteins in neurons.</p><p>cAMP signaling also leads to activation of PP-2B, a calcium-dependent protein phosphatase, and this enzyme is involved in the dephosphorylation of WAVE1 at Ser441. The precise mechanism involved in regulation of PP-2B is not known. However, previous studies have found in various cell types that L-type calcium channels can be activated via phosphorylation by PKA. Moreover, our previous studies have shown that dopamine can activate L-type calcium channels through D1 receptor-mediated regulation of PKA (Surmeier et al. 1995). Thus an increase in intracellular calcium through L-type calcium channels may be the mechanism involved in the PP-2B-mediated dephosphorylation at Ser441.</p><p>We have also found that activation of NMDA receptors leads to dephosphorylation of WAVE1, that this also involves PP-2A and PP-2B, but that the pattern of dephosphorylation is distinct from that of cAMP-mediated dephosphorylation (Fig. 7). NMDA receptor-induced increases in intracellular calcium would be expected to lead to activation of PP-2B. Notably, however, only phospho-Ser397 was dephosphorylated by PP-2B, while NMDA-induced dephosphorylation of Ser310 and Ser441 was mediated by PP-2A. In another previous study, we have also found that the B"/PR72 PP2A subunit is enriched in striatum (Ahn et al. 2007b). B"/PR72 has EF-hand motifs, binds to calcium, and was shown to be involved in NMDA/AMPA-induced dephosphorylation of Thr75 of DARPP-32 (Ahn et al. 2007b). Thus, PP-2A containing B"/PR72 may mediate NMDA-induced dephosphorylation of WAVE1 at Ser310 and Ser441.</p><p>Our previous studies have found in primary cultured cortical and hippocampal neurons that repetitive depolarization leads to dephosphorylation of WAVE1 at Cdk5 sites (Sung et al. 2008). In these studies in cultured neurons, the dephosphorylation of WAVE1 was associated with down-regulation of p35, the regulatory subunit of Cdk5. Co-incubation with the NMDA receptor antagonist 2-amino-5-phosphonovaleric acid (APV) blocked p35 down-regulation and the decrease in WAVE1 phosphorylation, but co-incubation with an AMPA/kainate receptor antagonist, 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) had no effect. The reason for the differences observed in the mechanisms of reduced phosphorylation of WAVE1 in cultured neurons compared to striatal slices is not known. However, differential expression of PP-2A heterotrimers containing the B"/PR72 subunit may be responsible in part.</p><p>Synaptic vesicle endocytosis involves the dephosphin family of phosphoproteins, which includes amphiphysin I and dynamin I. These proteins are constitutively phosphorylated by Cdk5 and are dephosphorylated by the calcium-dependent activation of PP-2B in response to depolarization (Cousin and Robinson 2001; Tan et al. 2003; Tomizawa et al. 2003). Post-synaptically, WAVE1 is phosphorylated by Cdk5 and dephosphorylated in response to stimulation of calcium-dependent activation of PP-2A and PP-2B. Given the similarity in the regulatory mechanisms found for presynatic dephosphins and WAVE1, WAVE1 might be considered as a postsynatic dephosphin, which couples neurotransmitter-mediated signaling pathways to morphological changes in dendritic spines.</p><p>One of the interesting features of the current study is the convergence of different second messenger-regulated protein phosphatases on different sites in WAVE1. For example, Ser397 is a good substrate for cAMP-stimulated PP-2A but not for calcium-stimulated PP-2A, while Ser441 can be dephosphorylated by PP-2B when slices are stimulated with forskolin but not when stimulated with NMDA. The specificity of different PP-2A activities towards specific sites is presumably a reflection of the ability of different B subunits to influence the selective dephosphorylation of different sites. However, it is not clear what the basis is for the selective activation of PP-2B towards different sites. It is possible that there is a hierarchy in terms of the relationship of PP-2A and PP-2B: for example, if a specific B subunit of PP-2A is active toward a given site, then the sites might not be available for PP-2B, perhaps because of some structural occlusion. However, if PP-2A is not targeted to a specific site then this could make it available for another phosphatase such as PP-2B (Fig. 7).</p><p>In summary, our studies indicate that neurotransmitters such as dopamine and glutamate activate WAVE1 by reducing the level of serine phosphorylation through the involvement of PP-2A as well as PP-2B. Future studies will be required to investigate how the phosphorylation at serine residues of WAVE1 influences the other features of the regulation of WAVE1, including intermolecular interactions of the WAVE1 complex (Eden et al. 2002; Ismail et al. 2009), interaction of the WAVE1 complex with diverse upstream activators such as phosphoinositides, Rac and SH3-domain-containing proteins (Takenawa and Miki, 2001), dimerization/oligomerization of WAVE1 (Padrick et al. 2008) and phosphorylation at tyrosine residues (Ardern et al. 2006). Dopaminergic and glutamatergic transmission in striatum and morphological changes in dendritic spines in striatal neurons have been implicated in many psychiatric and neurological disorders including drug addiction (Lee et al. 2006; Kalivas 2009; Kim et al. 2009) and Parkinson's disease (Day et al. 2006; Deutch et al. 2007). The signal transduction mechanisms we have found in this study may therefore play crucial regulatory roles in striatal synaptic plasticity, and alterations in these pathways may be associated with these psychiatric and neurological disorders. Future studies of regulation of WAVE1 in mouse models of psychiatric and neurological disorders will hopefully give new insight into the synaptic pathology of these disorders.</p><!><p>Wiskott-Aldrich syndrome protein (WASP)-family verprolin homologous protein 1</p><p>actin-related protein</p><p>cyclin-dependent kinase 5</p><p>protein phosphatase 2A</p><p>protein phosphatase 2B</p><p>central nervous system</p><p>Dopamine and adenosine 3',5'-monophosphate-regulated phosphoprotein (32 kilodaltons)</p><p>regulator of calmodulin signaling</p><p>protein phosphatase 1</p><p>enhanced chemiluminescence</p><p>standard error</p><p>okadaic acid</p><p>cyclosporin A</p>

PubMed Author Manuscript

A two-photon ratiometric fluorescence probe for Cupric Ions in Live Cells and Tissues

Development of sensitive and selective probes for cupric ions (Cu 21 ) at cell and tissue level is a challenging work for progress in understanding the biological effects of Cu 21 . Here, we report a ratiometric two-photon probe for Cu 21 based on the organic-inorganic hybrids of graphene quantum dots (GQDs) and Nile Blue dye. Meanwhile, Cu-free derivative of copper-zinc superoxide dismutase (SOD) -E 2 Zn 2 SOD is designed as the unique receptor for Cu 21 and conjugated on the surface of GQDs. This probe shows a blue-to-yellow color change in repose to Cu 21 , good selectivity, low cytotoxicity, long-term photostability, and insensitivity to pH over the biologically relevant pH range. The developed probe allows the direct visualization of Cu 21 levels in live cells as well as in deep-tissues at 90-180 mm depth through the use of two-photon microscopy. Furthermore, the effect of ascorbic acid is also evaluated on intracellular Cu 21 binding to E 2 Zn 2 SOD by this probe. Copper (Cu) is an essential trace element which ranks the third in abundance in the human body, and plays a pivotal role in many metabolic processes 1 . The ability of Cu to cycle between stable oxidized Cu 21 and unstable reduced Cu 1 states is used by cuproenzymes involved in redox reactions, e.g., Cu, Zn-superoxide dismutase and cytochrome oxidase 2 . However, the Cu 21/1 redox can in certain circumstances result in the generation of reactive oxygen species, which if not detoxified efficiently, would damage susceptible cellular components 3,4 . Alterations in the cellular copper homeostasis may also cause cell death and neurodegenerative diseases including Alzheimer's disease, Wilson's disease, Parkinson's disease, and so on [5][6][7] . Because the bulk of intracellular copper is believed to be present in its monovalent oxidation state (Cu 1 ), more attention has been paid on the development of fluorescence probes for imaging and biosensing of Cu 1 8-10 . Actually, the cupric state (Cu 21 ) is found most often in biological system 11 . It is as important as Cu 1 in cellular copper homeostasis and redox cycling processes of Cu 12 .Up to now, a few efficient fluorescent sensors for Cu 21 -selective detection have been reported [13][14][15][16][17][18][19][20] . We have also developed a ratiometric strategy for intracellular sensing and imaging of Cu 21 , with high sensitivity and accuracy 18,19 . However, most of these probes have been evaluated using one-photon microscopy and require relatively short excitation wavelengths, limiting their use in deep-tissue imaging because of the shallow penetration depth (,80 mm). To determine Cu 21 deep inside living tissues, it is crucial to use two-photon microscopy (TPM), a new technique that utilizes near-IR two-photon excitation [21][22][23][24][25][26][27][28][29][30] . TPM offers several advantages such as larger imaging depth (.500 mm), minimized autofluorescence background, and less photodamage associated with the use of near infrared excitation. In the past decades, organic dyes and semiconductor quantum dots (QDs) are widely studied two-photon probes 21,25,31,32 . But the rapid photobleaching effect and limited two-photon absorption crosssection of organic dyes hamper the imaging depth. The potential hazard of heavy metals in semiconductor QDs causes concern for in vivo bioimaging, despite of their strong two-photon fluorescence. Recently, graphene quantum dots (GQDs) have emerged as promising fluorescent materials with their good biocompatibility and photostability [33][34][35][36][37] . Two-photon imaging using the GQDs has also been reported for cell labeling 38 . However, no research about applications of GQDs as two-photon probes to sense biological activities or molecules has been reported so far.In this article, we report a GQDs-based two-photon ratiometric probe for imaging and sensing of Cu 21 in live cells and tissues with high selectivity and sensitivity. Three new strategies were developed in the present work. First, as shown in Fig. 1, the GQDs emitting blue fluorescence is hybridized with Nile Blue organic dye emitting red fluorescence as dual-emission fluorophore (GQD@Nile nanohybrid), in which Nile Blue is inert to Cu 21 and only serves as reference signal for providing built-in correction to avoid environmental effects. Next, Cu-free

a_two-photon_ratiometric_fluorescence_probe_for_cupric_ions_in_live_cells_and_tissues

<!>Results<!>Discussion<!>Methods

<p>derivative of bovine liver copper-zinc superoxide dismutase (SOD) -E 2 Zn 2 SOD (E designates an empty site) is employed as the unique receptor specific for Cu 21 because E 2 Zn 2 SOD can interact with Cu 21 with high specificity to reconstitute SOD [39][40][41] . Then, the specific E 2 Zn 2 SOD is conjugated with the GQD@Nile to form the GQD@ Nile-E 2 Zn 2 SOD fluorescent probe. The organic-inorganic hybrided probe shows dual emission bands centered at ,465 and ,675 nm, respectively, upon two-photon (800 nm) excitation. The GQDs functionalized with E 2 Zn 2 SOD can selectively recognize Cu 21 , leading to blue fluorescence quenching, whereas the red fluorescence of Nile Blue stays constant. Consequently, variations of the two fluorescence intensities display clear color changes from blue to yellow upon addition of Cu 21 , resulting in a ratiometric two-photon fluorescent sensor for Cu 21 . This probe also demonstrates high selectivity for Cu 21 over other metal ions and amino acids. Finally, the remarkable analytical performance of the present probe including high sensitivity and selectivity, as well as the fascinating properties of GQDs such as excellent aqueous stability, good biocompatibility, and long-term photostability, enables the imaging and sensing of Cu 21 in live cells and tissues at depths of 90-180 mm by TPM. To the best of our knowledge, this is the first report for development of a two-photon ratiometric fluorescence probe suitable for detection of Cu 21 in live cells and tissues.</p><!><p>GQDs were prepared by tailoring the carbonization degree of citric acid according to the previous report 42 . The as-made GQDs are mono-dispersed with an average size of ,10 nm (Fig. 2a) and a topographic height of ,2 nm (Fig. 2b) 35,36 . A broad diffraction peak corresponding to (002) planes of graphite was observed at around 25u in the X-ray diffraction (XRD) pattern (Fig. 2a) of GQDs, suggesting that carbonizing citric acid would produce graphite structures [43][44][45] . The Fourier transform infrared (FT-IR) spectrum of the GQDs was given in Fig. 2c (curve I). Five peaks located at 3447 (n O-H ), 1685 (n C5O ), 1588 (n COO 2), 1386 (n COO 2), and 1119 cm 21 (n C-OH ) were observed. The presence of these hydrophilic groups including -COOH and/or -OH imparts GQDs water-solubility. Then, the two-photon ratiometric probe was developed by conjugating GQDs with amino-containing Nile Blue chloride and E 2 Zn 2 SOD in an activation reaction with 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC) and N-Hydroxysuccinimide (NHS). The modification of Nile Blue dye and E 2 Zn 2 SOD onto the surface of GQDs was confirmed by FT-IR and X-ray photoelectron spectroscopy (XPS). Four peaks were clearly observed at 3292 (n N-H ), 1645 (Amide I C5O), 1539 (Amide II -COO 2 ), and 1450 cm 21 (n C-N ) in the FT-IR spectra of GQD@Nile-E 2 Zn 2 SOD (Fig. 2c, curve IV), which suggests the successful attachment of Nile Blue and E 2 Zn 2 SOD onto the surface of GQDs. In addition, the conjugated process was also tracked by XPS. As shown in Fig. S1, after the conjugation of Nile blue onto the GQDs surface, two new N 1s peaks were found at 399.9 eV and 401.7 eV, which are attributed to amide nitrogen and quaternary ammonium nitrogen, respectively (curve II) 46 . The observation of S 2p peak (Fig. S1) of thiol group or disulfide at 163.0 eV is evident the exact modification of E 2 Zn 2 SOD on the GQD@Nile surface, because this peak cannot be obtained at the surfaces of bare GQDs or GQD@Nile.</p><p>Upon both one-photon and two-photon excitation, the aqueous solution of GQDs was found to be strongly emissive in the visible and shows an emission maximum at 465 nm. As shown in the inset of Fig. 2d, the quadratic relationship between the excitation laser power and the luminescence intensity is obvious, thus confirming that the excitation with two near-infrared photons was indeed responsible for the observed visible luminescence of GQDs. Using quinine sulfate as a standard, the fluorescence quantum yield (W F ) of GQDs was calculated to be 8.2 6 1.0% 47 . The excitation-independent emission of the GQDs (data not shown) and the well-defined absorption band at 360 nm with a narrow full width at half maximum of 66 nm (Fig. 2d) verified that both the size and the surface state of those sp 2 clusters contained in GQDs should be uniform, which may contribute to their strong fluorescence 42 . As the Nile Blue dye exhibits fluorescence maxima at ,675 nm upon 800 nm fs-laser excitation (Fig. S2), both the obtained GQD@Nile and GQD@Nile-E 2 Zn 2 SOD show well-resolved dual emission bands centered at ,465 and ,675 nm, respectively (Fig. 3 and Fig. S3). Using a femtosecond (fs) fluorescence measurement technique, the two-photon action cross section (W F s 2P ) of GQDs and Nile Blue at 800 nm were estimated to be 22.0 6 6.0 GM and 1.6 6 0.5 GM (Goeppert-Mayer unit), respectively.</p><p>The response of the dual-emission two-photon fluorescent probe to Cu 21 was then carried out to prove the working principle, as demonstrated in Fig. 3. Upon addition of Cu 21 , the blue emission from the GQDs shows continuous quenching, whereas the red emission from Nile Blue still remains constant. Furthermore, F blue /F red , the ratio of the integrated intensities at 450-600 nm (F blue ) and 620-700 nm (F red ), gradually decreases with the increasing concentration of Cu 21 and the signal ratio shows good linearity with Cu 21 concentration in the range of 2 3 10 27 -3 3 10 26 M. The detection limit was calculated to be ,100 nM (based on a signal-to-noise ratio of S/N 5 3), which is comparable to those of previously reported Cu 21 biosensors [13][14][15][16] . The interaction between E 2 Zn 2 SOD and Cu 21 was further studied by optical spectroscopy (Fig. S4-S5) and XPS (Fig. S1). The absorption spectrum of the reconstituted SOD by Cu 21 was in a good agreement with that of native SOD with l max 5 680 nm, but obviously different from that of a Cu 21 aqua ion and that of E 2 Zn 2 SOD. Meanwhile, two clear peaks ascribed to Cu 2p were observed at 932.74 and 952.74 eV in XPS obtained at the reconstituted SOD-modified surface, which were not obtained at the surfaces of bare GQDs, GQD@Nile, or GQD@Nile-E 2 Zn 2 SOD. In addition, for better understanding the quenching mechanism, the timeresolved fluorescence (TRF) signals of GQD@Nile-E 2 Zn 2 SOD were probed in the absence and presence of Cu 21 at 465 nm with an excitation at 400 nm (Fig. S6). In the absence of Cu 21 , the fluorescence decays single exponentially by 1.68 ns time constant, which is the lifetime of GQDs in the nanohybrids. However, the lifetime changed to 1.35 ns after Cu 21 was bond at the surface of GQDs by E 2 Zn 2 SOD, which may be explained for the fluorescence quenching by the excitation energy transfer from the GQDs to the copper dorbital and/or GQDs to Cu 21 charge transfer 13,48 . The clear mechanism is unclear known at the present stage, because the fluorescence process of GQDs is still unclear and the related work is underway.</p><p>The complexity of intracellular system presents a great challenge for biosensors not only in sensitivity but more importantly in selectivity. The selectivity experiments were carried out by monitoring the intensity ratio (F blue /F red ) of the probe in the presence of millimolar concentrations of Na 1 , K 1 , Ca 21 , and Mg 21 , 10 mM concentrations of Mn 21 , Fe 21 , Co 21 , Ni 21 , Zn 21 , and Cu 1 , that may coexist in the living system. Remarkably, unperturbed fluorescence response (Fig. S7) was observed for the other metal ions, compared with that obtained for Cu 21 . Meanwhile, these potential metal ion interferences showed negligible effects on the signal for Cu 21 sensing. Taking into account that amino acids in the biological system are capable of interacting with a lot of metal cations, several typical amino acids were also examined. Little effect on the intensity ratios of the probes was obtained (Fig. S7) after their exposure to 10 mM concentrations of amino acids. On the other hand, an obvious decrease in the intensity ratio was observed upon the subsequent addition of 2 mM Cu 21 . In addition, negligible effect on the fluorescence probe was observed for the other biological species such as glutathione, glucose, hydrogen peroxide, proteins including cytochrome c, hemoglobin and myglobin (Fig. S7). Furthermore, the GQD@Nile-E 2 Zn 2 SOD signal was independent of solution pH in the biologically relevant pH (Fig. S8). The combined results reveal that GQD@Nile-E 2 Zn 2 SOD can detect Cu 21 with minimum interference from pH and from other metal ions, amino acids, and potential biological species due to the specific interaction of designed molecule E 2 Zn 2 SOD with Cu 21 to reconstitute SOD.</p><p>For biological application, the long-term cellular toxicity of GQD@Nile-E 2 Zn 2 SOD toward the A549 cell lines was determined by means of a standard MTT (methyl thiazolyl tetrazolium) assay 49,50 . In the presence of the GQD@Nile-E 2 Zn 2 SOD probe with the concentration from 20 to 1000 mg mL 21 , the cellular viabilities were estimated to be greater than 90% and 87% (Table S1) after incubation for 24 and 48 h, respectively. The results indicate that the GQDsbased probe is generally low-toxic for cellular imaging. The conclusion was also supported by the result of the flow cytometry experiments. Apoptosis assay confirmed no obvious increase of cell death or apoptosis after incubating cells with GQD@Nile-E 2 Zn 2 SOD (Fig. S9). Moreover, the two-photon excited fluorescence intensity of the GQD@Nile-E 2 Zn 2 SOD-labeled A549 cells remained nearly the same after continuous irradiation by the fs-pulses for 60 min, indicating its high photostability (Fig. S10).</p><p>We next sought to assess whether GQD@Nile-E 2 Zn 2 SOD as a two-photon probe could report changes in the Cu 21 level in live cells by ratiometric fluorescence imaging. Upon two-photon excitation at 800 nm, the emission ratio images of A549 cells labeled with GQD@ Nile-E 2 Zn 2 SOD were constructed from two collection windows (Fig. S10, 450-600 nm, F blue and 620-700 nm, F red ). From Fig. 4a, we can see that the average emission ratio is 5.56 6 0.39, revealing very low levels of available Cu 21 in the normal condition. Then, GQD@Nile-E 2 Zn 2 SOD was responsive to the change in the Cu 21 concentration: the F blue /F red ratio decreased to 2.18 6 0.51 when the cells were preincubated with 100 mM CuCl 2 (Fig. 4b). Treatment of cells with an excess of the cell-permeable, high affinity copper chelator, ethylenediaminetetraacetic acid (EDTA), increases the fluorescence ratio to 4.77 6 0.20 (Fig. 4c). These initial experiments in A549 cells demonstrate that we could observe changes in fluorescence ratios in cells treated with exogenous Cu sources. We then examined whether this indicator could be used to study the physiological process of intracellular binding of Cu 21 to E 2 Zn 2 SOD. In this context, L-ascorbic acid has been reported to inhibit the intracellular copper binding to apo-enzymes 51 . Fluorescence ratio images of A549 cells labeled with GQD@Nile-E 2 Zn 2 SOD revealed that L-ascorbic acid treatment depressed the reconstitution of E 2 Zn 2 SOD with Cu 21 , showing no obvious change in fluorescence ratio (the F blue /F red ratio only decreased to 4.20 6 0.30) in cells when compared with solely Cu source treatment (Fig. 4b, d). More importantly, D-isoascorbic acid, the epimer of the ascorbic acid, had no effect on the change of fluorescence ratio in cells treated with exogenous Cu sources (the F blue /F red ratio decreased to 2.42 6 0.35) (Fig. 4e). Clearly, the GQD@Nile-E 2 Zn 2 SOD probe provides a facile and effective model and strategy for evaluating the effects of L-ascorbic acid on intracellular Cu 21 binding to E 2 Zn 2 SOD.</p><p>To further investigate the utility of this probe in deep tissue imaging, TPM images were obtained from a part of lung cancer tissue slice incubated with 0.1 mg mL 21 GQD@Nile-E 2 Zn 2 SOD for 12 h at 278 K. As the structure of the obtained lung tissue slice is known to be inhomogeneous throughout its entire depth, we accumulated 10 TPM images from the two collection windows at depths of 90-180 mm to visualize the overall Cu 21 distribution. The concentrations of Cu 21 were estimated from the F blue /F red ratios and the titration curve (see above). Moreover, the image at a higher magnification clearly shows the Cu 21 distribution in the individual cells with an average emission ratio of 5.78 6 0.56 at a depth of 120 mm (Fig. 5c). When the tissue was pretreated with 100 mM CuCl 2 , the ratio decreased to 1.92 6 0.48 (Fig. 5f). It is worth noting that the changes in the emission ratios measured deep inside the tissue are comparable to those in the cells. Furthermore, the TPM images at depths of 90, 120, 150, and 180 mm demonstrate that GQD@Nile-E 2 Zn 2 SOD is capable of detecting Cu 21 at depths of 90-180 mm in live tissues using TPM (Fig. S11).</p><!><p>we have developed a new ratiometric two-photon probe, GQD@ Nile-E 2 Zn 2 SOD, which shows a marked blue-to-yellow emission color change in response to Cu 21 , with high selectivity and accuracy. Meanwhile, the GDQ-based fluorescence probe demonstrates low cytotoxicity, insensitivity to pH over the biologically relevant pH range, long-term photostability, and good cell-permeability. As a consequence, the ratiometric probe can visualize Cu 21 levels in live cells and tissues at depths of 90-180 mm, and further evaluate the effect of L-ascorbic acid on intracellular Cu 21 binding to E 2 Zn 2 SOD. This work has opened up a way to understanding the role that Cu 21 plays in the biological and pathological systems, as well as provided a methodology to design organic-inorganic ratiometric two-photon probes for detection of metal ions and other biological species.</p><!><p>Synthesis of GQDs. Graphene quantum dots (GQDs) were synthesized following the method reported by Dong et al. 42 . Citric acid (2 g) was heated to 473 K until the color of the liquid was changed from colorless to orange in about 30 min. The orange liquid was then added to 100 mL of 10 mg mL 21 NaOH solution dropwise under vigorous stirring. The obtained graphene quantum dots solution was kept in refrigerator.</p><p>Preparation of E 2 Zn 2 SOD. The Cu-free derivative E 2 Zn 2 SOD was prepared according to the method described by Cocco 52 . Briefly, DDC was added to 0.1 mM SOD solutions, buffered with 0.1 M potassium phosphate at pH 7.4, at a final concentration of 0.5 mM. The mixture was incubated at 310 K for about 2 h until no further increase in absorbance at 450 nm was observed. Then, the yellow solution was centrifuged at 39000 g for 30 min, and the colorless supernatant was exhaustively dialysed against doubly distilled water. Finally, the E 2 Zn 2 SOD solution was lyophilized and kept in the refrigerator for further use.</p><p>Conjugation of GQD@Nile-E 2 Zn 2 SOD. For the synthesis of GQD@Nile-E 2 Zn 2 SOD, the obtained GQDs solution was mixed with Nile Blue chloride (10 24 M) and E 2 Zn 2 SOD (2 3 10 25 M) and stirred for 5 min. Then, EDC/NHS (20 mM) were added and stirred for ,2-4 h for the conjugation. Finally, the nanohybrids were separated from free EDC/NHS, unreacted Nile Blue chloride and E 2 Zn 2 SOD by three cycles of concentration/dilution (1051), using a Nanosep centrifugal device (Pall Corporation, MW cutoff of 3 kDa), and redispersed in PBS buffer (pH 5 7.4).</p><p>Linear optical properties. Linear absorption was measured with an Agilent 8453 UV-vis spectrophotometer. Fluorescence emission and excitation spectra were measured on a Hitachi F-2700 fluorescence spectrophotometer equipped with a 90 W Xenon lamp in a 1-cm cuvette. Fluorescence quantum yield (W) of GQDs was obtained from measurements at five different concentrations in HEPES buffer using the following equation:</p><p>Where W F is the quantum yield, Abs and F denote the absorbance and fluorescence intensity, respectively, and P F denotes the peak area of the fluorescence spectra, calculated by summation of the fluorescence intensity. g is the refractive index of the solvent. ST is the standard and X is the sample. Quinine sulfate (literature W F 5 0.54) was used as a standard 47 , and excitation was performed at 400 nm.</p><p>Measurements of two-photon fluorescence excitation spectra and two-photon action cross section. The two-photon action cross section (W F s 2P ) was determined by using a femtosecond (fs) fluorescence measurement technique as described 53 . The two-photon excited fluorescence spectra were measured on a spectrometer (HORIBA Model iHR 550) and the pump laser beam came from a mode-locked Ti:sapphire laser (Coherent Mira 900) with a pulse duration of 80 fs and a repetition rate of 76 MHz. The two-photon excited fluorescence spectra of the reference and the sample were determined at the same excitation wavelength. Rhodamine B (0.6 mM) in MeOH was used as a reference, whose two-photon properties have been well characterized in the literature 54 . By taking the ratio of equation 2 for sample and reference, the two-photon action cross-section (W F s 2P ) of the sample was calculated.</p><p>Here, F (t) 5 the time-averaged fluorescence photon flux, g 5 fluorescence collection efficiency, C 5 sample concentration, W F 5 fluorescence quantum efficiency, s 2P 5 non-linear two-photon absorption cross-section, g 5 degree of second order temporal coherence, n 5 index of refraction of medium lens works in, P (t) 5 instantaneous incident power, and l 5 wavelength in vacuum.</p><p>Fluorescence lifetime measurement. Fluorescence lifetime measurements were performed on an Edinburgh Instruments (FLS 920) spectrometer by the timecorrelated single photon counting (TCSPC) method. Excitation was achieved by a hydrogen-filled nanosecond flash lamp (repetition rate 40 kHz). Measurements were taken in ambient conditions, at room temperature, on solutions diluted to yield reasonable signal intensity. Excitation was performed at 400 nm for the GQD@Nile-E 2 Zn 2 SOD and the time-resolved emission data were collected at 465 nm. The TCSPC traces were analyzed by standard tail fit implemented in the software of the fluorimeter. Weighted residuals and x 2 values were used to judge the quality of the fit.</p><p>Two-photon fluorescence microscopy. Two-photon fluorescence images of dyelabeled cells and tissues were obtained with spectral confocal and multiphoton microscopes (Leica TCS SP8) with 310 dry and 363 oil objectives, numerical aperture (NA) 5 0.4 and 1.4. The two-photon fluorescence microscopy images were obtained with a DMI 6000 Microscope (Leica) by exciting the probes with a modelocked titanium-sapphire laser source (Mai Tai DeepSee, 80 MHz, ,90 fs) set at wavelength 800 nm and output power 2920 mW. To obtain images at 450-600 nm (blue) and 620-700 nm (red) range, internal PMTs were used to collect the signals in an 8 bit unsigned 1024 3 1024 pixels at 200 Hz scan speed.</p><p>Cell culture and cytotoxicity assay. Human lung cancer cells line A549 were cultured in RPMI 1640 supplemented with 10% fetal bovine serum. When in the proliferative period, A549 cells (,3 3 10 5 cell mL 21 ) were dispersed within replicate 96-well micro-liter plates to a total volume of 100 mL well 21 and maintained at 310 K in a 5% CO 2 /95% air incubator for 24 h. Then, the culture media was removed and the cells were incubated in culture medium containing the as-prepared GQD@Nile-E 2 Zn 2 SOD with different concentrations for 24 h or 48 h and washed with the culture medium. An amount of 100 mL of the new culture medium containing MTT (10 mL, 5 mg mL 21 ) was then added, followed by incubating for 4 h to allow the formation of formazan dye. After removing the medium, 150 mL DMSO was added to each well to dissolve the formazan crystals. Absorbance was measured at 490 nm in a Multiskan MK3 microplate photometer (Thermo Scientific). Cell viability values were determined (at least three times) according to the following formulae: cell viability (%) 5 mean of absorbance value of treatment group/mean absorbance value of control 3100%.</p><p>Flow cytometry. The percentage of apoptotic cells was determined by monitoring the translocation of phosphatidylserine to the cell surface using an Annexin V-FITC apoptosis detection kit (KeyGEN Biotech) according to the manufacturer's instructions. Briefly, after dye treatment, the cell culture medium was collected to retain floating cells and attached cells were dislodged using the EDTA-free trypsin.</p><p>Floating and attached cells were combined and harvested by centrifugation. The cell pellets were suspended in 500 mL binding buffer and incubated with 5 mL FITC-Annexin V and 5 mL of a propidium iodide solution for 15 min in the dark. Cells were evaluated for apoptosis using a Becton-Dickinson flow cytometer with Annexin V-FITC and PI double staining. Fluorescence was measured with an excitation wavelength of 480 nm through FL-1 (530 nm) and FL-2 filters (585 nm).</p><p>Preparation and staining of lung cancer tissue slice. Tissue slices were prepared from A549 human lung cancer cells. A total of 2 3 10 6 A549 cells diluted in 200 ml of serum-free DMEM medium were injected subcutaneously into the right flank of 6-to 8-week-old BALB/c nude mice to inoculate tumors. On day 15 after A549 inoculation, mice were sacrificed. Tumors were removed and embedded with O.C.T (Sakura Finetek, USA, Torrance, CA) for frozen sections. The tissues were cut into 250 mmthick slices using a vibrating-blade microtome. Slices were then treated with nanoprobe GQD@Nile-E 2 Zn 2 SOD for 12 h at 4uC. After washing with PBS, the slice were mounted with 10% glycerol and sealed with nail varnish on a glass substrate. The TPM images of lung cancer tissue labeled with nanoprobe were obtained at depths from 90 to 180 mm in a spectral confocal multiphoton microscope as shown in Figure S11.</p>

Scientific Reports - Nature

REACTION DYNAMICS OF DIELS-ALDER REACTIONS FROM MACHINE LEARNED POTENTIALS

Recent advances in the development of reactive machine-learned potentials (MLPs) promise to transform reaction modelling. However, such methods have remained computationally expensive and limited to experts. Here, we employ different MLP methods (ACE, NequIP, GAP), combined with automated fitting and active learning, to study the reaction dynamics of representative Diels-Alder reactions. We demonstrate that the ACE and NequIP MLPs can consistently achieve chemical accuracy (± 1 kcal mol −1 ) to the ground-truth surface with only a few hundred reference calculations. These strategies are shown to enable routine ab initio-quality classical and quantum dynamics, and obtain dynamical quantities such as product ratios and free energies from non-static methods. For ambimodal reactions, product distributions were found to be strongly dependent on the QM method and less so on the type of dynamics propagated.

reaction_dynamics_of_diels-alder_reactions_from_machine_learned_potentials

Introduction<!>MLP comparison for a Diels-Alder reaction<!>Product ratios and time gaps<!>Free energies<!>Conclusion<!>Methods<!>Conflicts of interest<!>Author contributions

<p>Simulating chemical reactions is essential to developing a fundamental understanding of their mechanism and predicting experimental outcomes. 1 Machine learned potentials (MLPs) offer an enticing approach to such simulations, enabling the efficient mapping between nuclear configurations and energies (R → E). Moreover, in contrast to classical force fields, they offer flexibility and systematic improvability. 2 In the limit of correct forces and converged sampling, such simulations should afford access to the accurate estimation of rate and equilibrium constants. However, despite the development of Gaussian Approximation Potentials (GAPs) 3,4 and high dimensional neural network potentials (NNPs) 5 more than ten years ago, MLPs are still yet to find routine use for chemical reaction simulation. 6 This slow uptake is likely due to the computational and time investment required to train reactive potentials for new systems.</p><p>Iterative training is essential to construct robust MLPs and can be summarised as follows: i) developing a training set, ii) performing the regression, iii) repeating the process until the desired accuracy is obtained. Automated approaches to construct large training sets have been previously reported; 7,8,9 however, their application remain limited to small systems. Moreover, the need to perform several thousands of energy and gradient evaluations on a reference potential energy surface (PES) makes it difficult to go beyond density functional theory (DFT) and employ accurate but computationally expensive wavefunction (WF)-based methods. 10 Exceptions are rare and limited to systems with ≲ 10 atoms. 8,11 Finally, achieving the desired accuracy when training reactive MLPs is challenging because the energy scale required for accurate dynamics is much larger than for non-reactive MLPs (where sampling around the minima is often sufficient). As such, more training data is required to reach the same accuracy. This is particularly relevant when sampling areas around the transition states (TSs), which often require WF-based methods, such as coupled-cluster (CC). 12 Here, we show that recently developed MLP methods 13,14 can be used to generate accurate potentials for modestly sized reactions (≈ 50 atoms) in an automated fashion and outline some of the enabled insights. 2 Results and Discussion</p><!><p>We previously introduced a strategy to generate reactive GAPs in an autonomous manner, requiring only hundreds to a few thousand energy and gradient evaluations. 8 Here, we extend this methodology to explore more complex processes and compare it to recently developed MLP approaches (Fig. 1). As a starting point, we considered Diels-Alder (DA) reactions because of the available theoretical and experimental data, 15,16 and their prominence in chemical and biochemical contexts. 17,18,19 Initial efforts to apply our original active learning (AL) strategy using GAP regression to model the reaction between ethene + butadiene, R1, proved promising, producing qualitatively reasonable reactive molecular dynamics (MD). However, the quality of these GAPs were not within 1 kcal mol −1 accuracy of the QM reference method, which is required for accurate rate estimation or dynamic studies (Fig. S1a, MAD = 5.3 kcal mol −1 ). We hypothesised that this was due to the exothermicity of this reaction, which requires the MLP to be accurate in a 60 kcal mol −1 energy window. Indeed, using the same strategy and hyperparameters to model a less exothermic H-abstraction reaction (</p><p>resulted in a more accurate potential (Fig. S1b, MAD = 2.1 kcal mol −1 ). A modest improvement was obtained upon hyperparameter optimisation, at moderate computational cost (≈ 500 configurations required for R1). Specifically, reducing the regularisation, increasing the quality of the radial basis, and doubling the number of atomic environments considered in the training all improved the GAP (SI §S2). However, attempts to improve GAPs by training a component-wise potential separated over covalent bonds were unsuccessful (SI §S3.3).</p><p>We then turned our attention to the recently developed Atomic Cluster Expansion (ACE 13 ) and Neural Equivariant Interatomic Potentials (NequIP 14 ) MLP methods and evaluated their performance within the same training strategy for R1 (Fig. 2). While rather different in philosophy, both ACE and NequIP outperform GAPs and provide MLPs of similar accuracy (Fig. 2a, Fig. S24). Here, accuracy is based on deviations between QM ground-truth energies (true) and predicted energies over independent DFT-MD trajectories propagated from the TS to the reactant and product states. Previously, we have shown that a prospective validation strategy in the configuration space accessible to that MLP is essential to characterising 'good' MLPs. 8 These potentials are stable by construction as we use an AL strategy (Fig. 2, bottom left), where we define a stable potential where 10 × 1 ps trajectories can be propagated without encountering a configuration that deviates > 2.3 kcal mol −1 (0.1 eV) from the true energy.</p><p>When considering the amount of data required to train an accurate MLP for R1, GAPs both with and without hyperparameter optimisation required ≈ 500 configurations. However, this is surpassed by both ACE and NequIP potentials, where n train ≈ 100 (Fig. 2b). Comparing training time across these methods is less straightforward due to their different architecture and implementation. In all cases they are found to be efficient and therefore suitable for rapidly developing bespoke MLPs. Among them, NequIP had the greatest training time (half a day using 10 CPU cores + 1 GPU), while GAP and ACE required just 5 ± 2 h of total training time on 10 CPU cores. It is important to note that this difference in training time reduces with the system size, as reference energy and force evaluations dominate the computational time for equally data efficient MLPs.</p><p>Tangent to our goal of developing accurate MLPs for DA reactions, we found that GAP regularisation could be harnessed to reduce the computational cost of developing CCSD(T)-quality potentials (SI §S4). For simple molecules, MP2 forces are accurate to within 1.1 kcal mol −1 Å −1 (0.05 eV Å −1 ) of their CCSD(T) counterparts (Fig. S16), thus within the 'expected error' of the GAP. This removes the requirement for numerical CCSD(T) gradients. This strategy, combined with CUR 20 selection, afford a 100-fold reduction in the number of required CCSD(T) energy calculations, thus greatly reducing the time required to train a CCSD(T)-quality potential (Table S3). Furthermore, re-evaluating energies and forces from AL configurations with a different reference method (i.e. 'uplifting') reduces the computational cost associated with training WF-quality MLPs. Such strategies were also applied to other electronic structure methods. For example, uplifting PBE0/DZ configurations to MP2/TZ affords an ACE MLP within ± 1 kcal mol −1 of MP2-MD-sampled configurations (Fig. 2c, DZ=double-ζ basis set, TZ=triple-ζ). These strategies enable us to attain a higher level of accuracy in the MLP at a reduced computational cost.</p><p>Comparison of the ACE-, GAP-and NequIP-predicted two-dimensional PES over the two forming C-C bonds in R1, where all other degrees of freedom are free to relax, reveals that all three potentials are accurate and smooth in the extrapolation regime (Fig. S26). While all three potentials are suitable for training this reaction, ACE offers the best overall balance in terms of data efficiency, training time and accuracy of the final potential (Fig. 2b). Moreover, even when the closest configuration in the training data is 0.3 Å away in the forming bonds, the error is only a couple of kcal mol −1 when AL is initiated at the TS (r 1 = r 2 = 2.30 Å), suggesting that ACE is accurate in the extrapolation regime (Fig. 2d). Based on these results, in the following sections we focus our discussions on ACE due to its overall efficiency and accuracy.</p><!><p>To demonstrate the applicability of the ACE framework to more complex reactions, we modelled the reaction between tropone and cyclohepatriene, R2, and monitored its product distribution. This reaction has been previously studied by Houk and co-workers using quasi-classical reaction dynamic simulations. 21 The reaction has been found to involve an ambimodal TS that leads to three distinct products: one [8+2] and two [6+4] cycloadducts (Fig. 3, top).</p><p>Initial training of an ACE potential from the TS (R2 TS1 ) generated ≈ 450 configurations using standard AL with sampling in the reactant and product regions of two of the products (R2 P1 and R2 P2 ). However, even when propagating MLP-MD trajectories at 500 K, the Cope product (R2 P3 ) was not observed in the training data. Only when the TS that leads directly to R2 P3 was included as an AL starting configuration was the training data sufficiently complete. This second TS was 4.4 kcal mol −1 higher in energy than the TS that leads to R2 P1 and R2 P2 . 21 This result highlights the importance of considering relevant points in the PES that may not be obtained using automated sampling methods when training MLPs. For this reaction, ACE training required ≈ 2 days of AL.</p><p>Having successfully trained this ACE MLP for R2, we propagated dynamics to compute product distribution. Using 100 trajectories, each of them taking only minutes, we were able to obtain converged results (Fig. 3a). This efficiency also allowed us to propagate ring polymer MD (RPMD) 22 using the developed ACE potential. As expected, given the nature of the bonds being formed during this reaction, the product ratios were similar to those obtained using classical MD. We also evaluated the effect of temperature by equilibrating the system at the TS using constrained NVT dynamics (1 eV Å −2 , ≈ 0.5 kcal mol −1 additional ZPE), prior to propagating NVE downhill dynamics. Both direct NVE downhill and equilibrated dynamics provide qualitatively similar trends. However, a more detailed analysis shows that the latter affords almost double the proportion of R2 P1 and the formation of R2 P3 as 1% of the total product distribution, which would otherwise not be observed (right most distribution in Fig. 3b). This demonstrates the importance of propagating equilibrated quantum dynamics to obtain converged distributions. 23 Training an MLP uniquely allows for the 100-fold more expensive simulations to be performed, with the effect of ZPE and tunnelling accounted for.</p><p>Finally, we evaluated the effect of QM methodology on product distributions by performing MLP uplifts to different levels of theory. Uplift corresponds to performing single point energy and force evaluations on ACE AL configurations (PBE0/def2-SVP) at different levels of theory and then retraining the ACE potential to propagate classical NVE trajectories from the respective TSs. For the product distributions in Fig. 3c, MLPs can be obtained at the desired level of theory in 100s CPUh by applying uplifts, compared with > 200,000 CPUh without uplifts. It was found that some levels of theory require more AL iterations than others, suggesting that the overlap of the configuration space within the training set may not be optimal. For example, while the product distributions are unchanged for B3LYP on additional AL, M06-2X distributions continue changing upon further AL iterations, affording ≈ 10% more product R2 P2 after 5 iterations. Product distributions vary considerably between QM methodology, with the major state varying from the reactant to R2 P2 , and the proportion of R2 P1 varying from 1% to 10% (Fig. 3c). This evidences the large influence that the level of theory has on the propagated dynamics and consequently on the product distribution. The differences are significantly more important than those observed when comparing classical MD and RPMD results.</p><p>To further investigate the influence of the level of theory and the type of dynamics, the time gap, defined as the time difference between the formation of two C-C σ-bonds, 24 is evaluated for the reaction of methyl-vinyl ketone (MVK) + cyclopentadiene (CPD), R3.ACE MLPs trained at different levels of theory led to slightly different time gaps. For example, both PBE0 and ωB97X-D generated average gaps of 20 fs, while for M06-2X it was ≈ 10 fs (Fig. 4a). The latter is consistent with the DFT-MD value of 11±6 fs in the literature 25 , indicating the reaction is dynamically concerted, as concluded by Houk and co-workers. 24,26 Furthermore, as expected, quantum effects on this reaction are minimal, (Fig. 4b), since the differences between classical MD and RPMD are within 2 fs, while the differences due to the level of theory vary by over 5 fs. Traditionally, such estimates require several hundred quasi-classical trajectories, which are computationally demanding. However, by employing our strategy, we can compute the time gap from 1000 trajectories in fewer than 6 hours.</p><!><p>We further demonstrate the applicability of the ACE MLPs by employing them to compute free energy contributions for the reaction between MVK and CPD, R3. These contributions are often calculated within the rigid-rotor-harmonicoscillator (RRHO) approximation. However, this method is known to be inaccurate for the description of low-lying vibrational modes. 27 To overcome these limitations, different static approaches have been developed. They include the method introduced by Truhlar, referred here to as shifted RRHO (sRRHO), where all low frequency harmonic modes below a threshold are shifted to that value 28 and the qRRHO method introduced by Grimme, where those modes are treated between harmonic vibrations and rigid rotors. 27 Using the ACE-MLPs trained for this reaction at the PBE0-D3BJ/def2-SVP level of theory we computed the free energy profile for R3 using umbrella sampling (MLP-US) over an averaged reaction coordinate, r = (r 1 + r 2 )/2 (Fig. 5a). This approach enabled us to fully sample the different conformations not only at the critical points but along the reaction pathway, totalling 1.5 ns of sampling, which would be impossible to reach with AIMD. Moreover, our MLP-US simulations also capture anharmonic effects arising from low vibrations modes, which are not directly accounted for using the static methods. The MLP-US also provides a piece-wise free energy, where the choice of reaction coordinate is flexible.</p><p>Using this MLP, we also computed the temperature dependence for both activation and reaction energy in order to compare the MLP-US method to the static methods (Fig. 5b and 5c). For this DA reaction, MLP-US gives activation and reaction free energies comparable to the static methods, with a linear dependence on temperature. While no experimental data is available for the free energy of activation of this reaction in the gas phase, the comparison of the different static approaches to our MLP-US values suggest an estimate error of ≈ 3 kcal mol −1 for the free energy contributions. When comparing free energies of reaction (Fig. 5c), a larger deviation and larger error bars are observed at low temperature. This is likely due to the insufficient sampling; thus this point is excluded from the linear fit.</p><p>Using the same approach as for R3, we considered the DA reaction between cyclopentadiene and acetylene (R4, Table 1), for which comparison to experimental gas-phase kinetic data is possible. 29 Performing MLP-US enabled us to obtain quantitative accuracy in both enthalphic and entropic contributions. These results show that MLPs, in combination with US, enable us to achieve accurate free energies with associated errors, and treat anharmonic effects without ad-hoc corrections. While the MLP fitting and US protocol was completely automated and achieved within a day for this reaction, the choice of the QM reference method remains a critical choice that still requires human intervention.</p><p>Comparison to a wider range of DA and other types of reactions, for which experimental data is available, will be the subject of future work. expt. 29 21.9(1) −37.3(2)</p><!><p>In this work, we compared the performance of three MLP methods, ACE, NequIP and GAP, to study different aspects of modestly complex (≈ 50 atoms) Diels-Alder reactions. We also introduce a freely available automated fitting code mlp-train to facilitate the generation of these potentials. 30 While all three methods provide reasonable potentials, ACE and NequIP emerged as more efficient methods for generating accurate MLPs within 1 kcal mol −1 accuracy to the ground-truth surface, in particular for highly exothermic reactions.</p><p>ACE MLPs were also obtained for other Diels-Alder reactions and consistently achieved chemical accuracy (± 1 kcal mol −1 ). This enabled us to perform RPMD in order to introduce nuclear quantum effects, as well as to obtain free energies employing umbrella sampling. While product distributions were strongly dependent on the DFT reference method, they were less dependent on the type of dynamics propagated (classical vs quantum). When computing free energies, MLP-US predicts barrier free energies within the static methods' ranges, which vary by ≈ 3kcal mol −1 . For reaction free energies, larger errors bars were obtained at low temperature due to insufficient sampling. Comparison to experimental values demonstrates that MLP-US provide quantitative accuracy.</p><p>The diverse range of properties studied in this work demonstrate the applicability of our strategy, delivering accurate and efficient MLPs within a day. However, advancements in reference methods are needed to obtain accurate potentials without a preceding benchmark study. We are confident this will enable the routine development of chemically accurate reactive MLPs.</p><!><p>Training strategy. All Gaussian Approximation Potentials (GAPs) were trained using the gap_fit and QUIP codes with a Smooth Overlap of Atomic Positions (SOAP) 31 kernel with hyperparameters as defined in Table S2. The relationship between the size of the training set required to obtain an accurate GAP scales with system size is detailed for linear alkanes and small organic molecules in SI §S5. GAP-MD simulations were performed with ASE 32 interfaced to QUIP with the quippy wrapper using the Langevin integrator. Initial configurations, CUR 20 selection and automated training set construction were generated with the mlp-train Python package (v. 1.0.0a). 30 The training set developed by AL was based on MLP-dynamics with the energy selection strategy 'diff', E T = 2.3 kcal mol −1 (0.1 eV). 8 The other AL selection strategies, based on SOAP similarity, are discussed in SI §S7.</p><p>A stable potential was defined by its ability to propagate 10 × 1 ps MD trajectories without finding a configuration</p><p>Training sets for initial GAPs including those that used GP variance as a selection criteria were constructed using the gap-train code (v. 1.0.0b). 33 ACE 13,34 potentials were trained using the ACE.jl code 35 (v 0.8.4, Julia v. 1.6.3) and wrapped by pyjulip (v. 0.1, using PyCall). All MD simulations were propagated using the Langevin (0.02 friction coefficient) integrator in ASE v. 3.22.0 with initial velocities sampled from a Maxwell-Boltzmann distribution then propagated using a 0.5 fs timestep. NequIP 14 potentials used the nequip 36 v. 0.3.3 Python package (pytorch v. 1.9.1, pytorch_geometric v. 1.7.2, e2nn 37 v.0.3.5) with a maximum 1000 epochs and a 90:10 training to validation data split; all other hyperparameters were retained as their defaults (4.0 Å cut off radius, see nequip module in mlp-train, commit 0ba027c). Gas phase (vacuum) simulations in periodic potentials used a large cubic box (l = 10 nm). Combinations of harmonic restraining potentials were added and incorporated into the dynamics using ASE.</p><p>RPMD simulations. RPMD were performed using I-PI 38 with 16 beads, a 0.5 fs timestep, 300 K, initiated from the TS, or first equilibrated using constrained NVT dynamics as specified (80 fs, harmonic: 4 distances, k = 1 eV Å −2 ) followed by NVE dynamics.</p><p>Ground truth calculations. DFTB calculations were performed with DFTB+ 39 using 3ob 40 parameters in a periodic box with side lengths of 50 Å to mimic a vacuum. Molecular DFT used GFN2-XTB 41 in XTB v. 6.4.0. Molecular DFT (inc. AIMD), MP2 and coupled cluster [CCSD(T)] calculations performed with ORCA 42,43 v. 4.2.1 wrapped with autodE 44 using PBE 45 , PBE0 46 , B3LYP 47 , M06-2X 48 and ωB97X-D 49 functionals, def2-SVP, def2-TZVP, def2-TZVPP basis sets. 50 DLPNO coupled cluster calculations employed TightPNO thresholds. 51 Double-hybrid DFT calculations were performed in ORCA v. 5.0.2. Transition states were located with autodE. Plots were generated with matplotlib and, where quoted, include a standard error of the mean averaged over three independent repeats.</p><!><p>There are no conflicts to declare.</p><!><p>TAY, TJW and FD conceptualised the study. TAY developed and implemented the strategy in mlp-train, and carried out the calculations. TJW implemented the free energy methodology in mlp-train and HZ carried out the time gap analysis. All authors participated in data analyses. FD supervised the study and TAY and TJW wrote the original draft. All authors contributed to writing and editing of the manuscript.</p>

ChemRxiv

Role of AQP9 in transport of monomethyselenic acid and selenite

AQP9 is an aquaglyceroporin with a very broad substrate spectrum. In addition to its orthodox nutrient substrates, AQP9 also transports multiple neutral and ionic arsenic species including arsenic trioxide, monomethylarsenous acid (MAsIII) and dimethylarsenic acid (DMAV). Here we discovered a new group of AQP9 substrates which include two clinical relevant selenium species. We showed that AQP9 efficiently transports monomethylselenic acid (MSeA) with a preference for acidic pH, which has been demonstrated in X. laevis oocyte following the overexpression of human AQP9. Specific inhibitors that dissipate transmembrane proton potential or change the transmembrane pH gradient, such as FCCP, valinomycin, and nigericin did not significantly inhibit MSeA uptake, suggesting MSeA transport is not proton coupled. AQP9 was also found to transport ionic selenite and lactate, with much less efficiency compared with MSeA transport. Selenite and lactate uptake via AQP9 is pH gradient dependent and inhibited by FCCP and nigericin, but not valinomycin. The selenite and lactate uptake via AQP9 can be inhibited by different lactate analogs, indicating that their translocation share similar mechanisms. AQP9 transport of MSeA, selenite and lactate is all inhibited by a previously identified AQP9 inhibitor, phloretin, and the AQP9 substrate AsIII. These newly identified AQP9 selenium substrates imply that AQP9 could play a significant role in MSeA uptake and possibly selenite uptake involved with cancer therapy under specific microenvironments.

role_of_aqp9_in_transport_of_monomethyselenic_acid_and_selenite

Introduction<!>Expression of AQP9 in Xenopus oocytes<!>Transport Assays of MSeA and selenite<!>Transport Assays of lactate<!>Selenium quantification<!>Statistical Analysis<!>MSeA transport by AQP9 in Xenopus oocytes<!>Lactate transport by AQP9 in Xenopus oocytes<!>Selenite transport by AQP9 in Xenopus oocytes<!>Discussion

<p>Selenium is an essential micro-nutrient for all mammals. Dietary selenium can be metabolized into different organic species and is dominantly present as a methylated species. Selenite (SeIV) and selenate (SeVI) are two common inorganic selenium species, and selenite has been approved by the USDA as a food additive for farm animals. Organic dietary selenium includes selenocysteine and selenomethione, in which elemental selenium replaces sulfur in the corresponding amino acids cysteine and methionine. All of the selenium containing amino acids can be metabolized intracellularly for the synthesis of essential selenoproteins.</p><p>Inorganic selenite has been applied in many clinically relevant studies involving cell culture and animals, mainly for cancer prevention and treatment (Brodin et al. 2015; Ganther 1999; Jackson and Combs 2008). Selenite regulates a wide range of downstream cell signals. For example, selenite affects the phosphoinositide 3-kinase(PI3K)- serine-threonine kinase Akt pathway and three major mammalian mitogen-activated protein kinase(MAPK) pathways: exctracellular signal-regulated kinase (ERK) 1/2, c-Jun NH2-terminal kinase(JNK), and p38(Jiang et al. 2002; Zou et al. 2008). Membrane transporters that facilitate the cellular permeation of selenite have been recently reported. In yeast, the major selenite transporter is characterized as the lactate transporter, Jen1p(McDermott et al. 2010b). It is predicted that selenite resembles the structure of lactate and is recognized by Jen1p by molecular mimicry. However, functional homologues of Jen1p, mammalian monocarboxylate transporters (MCT), have no detected selenite transport activity (unpublished data, our lab). In 2016, the first direct transporter for selenite permeation was identified as ZIP8, which is a member of the Zinc import family (ZIP)(McDermott et al. 2016). ZIP8 transports selenite in a zinc and bicarbonate dependent manner. Since ZIP8 is found upregulated during inflammation, its role is predicted to recruit circular anti-inflammatory selenite and zinc to combat inflammation. This is consistent with the observation of selenite application in inflammatory diseases such as severe sepsis(Yang et al. 2016). While ZIP8 may play a critical role to facilitate selenite transport into inflammatory tissues, here we studied selenite uptake via AQP9. Our results showed AQP9 transports selenite under acidic conditions, but not under physiological pH, which indicates that AQP9 may serve as a secondary transporter under specific conditions, such as in an acidic microenvironment. The AQP9 mediated transportation of lactate has been determined to be comparable with that of selenite uptake, and results showed their transportation shares a similar translocation mechanism as we observed in yeast.</p><p>Monomethylselilinic acid (MSeA) is an intracellular metabolite derived from inorganic selenium methylation. Recently, MSeA was found to exhibit promising effects in the prevention and treatment for multiple cancer types, such as pancreatic cancer (Wang et al. 2014), lung cancer (Swede et al. 2003), breast cancer (Qi et al. 2012a), prostate cancer (Jiang et al. 2002), ovarian cancer (Swede et al. 2003; Zhang and Azrak 2009), and primary effusion lymphoma (PEL) (Wang et al. 2016), when MSeA was used at lower micro-molar range. In addition, MSeA was used as an adjuvant to synergistically enhance growth-inhibitory effect of the chemotherapeutic drugs doxorubicin and paclitaxel in breast cancer cells (Hu et al. 2008; Li et al. 2007). At in vivo levels, MSeA showed a dose-dependent restriction of xenograft tumor growth (Li et al. 2008; Wu et al. 2012).</p><p>Mechanisms of MSeA function includes inhibition of specific cell signaling pathways, some growth factors or extracellular matrix proteins, as well as inducing G1 arrest, DNA fragmentation, and caspase-mediated apoptosis. For example, treatment of primary effusion lymphoma (PEL) with MSeA was found induce an anti-proliferative effect by causing endoplasmic reticulum (ER) stress and subsequent apoptosis (Shigemi et al. 2017). MSeA induces apoptosis and G1 cell cycle arrest by perturbing PI3K through Akt kinase and forkhead box O protein (FOXO) dephosphorylation (Tarrado-Castellarnau et al. 2015). In human umbilical vein endothelial cells (HUVECs), MMP2 and VEGF expression was decreased upon short-term exposure to MSeA (Jiang et al. 2000). MSeA has a higher reactivity and displays superior efficacy against human cancer than other selenium species such as selenite. It is discovered that MSeA is readily metabolized to methylselenol, a bioactive selenium metabolite for cancer chemoprevention(Ip et al. 2000; Li et al. 2008).</p><p>However, despite its high toxicity for cells and therapeutic effects, mechanisms of MSeA permeation into cell membranes have not been studied. Given the higher toxicity and efficient cellular effect, one or more transporters for MSeA is predicted to universally exist. Here for the first time, we report that AQP9 transports MSeA effectively in a wide pH range and suggest it may serve as a major transporter for MSeA cell permeation. We demonstrated that the uptake is in favor of anacidic pH. Inhibitory studies have supported a hypothesis that MSeA transport does not require a transmembrane proton gradient. Since membrane permeation of MSeA is the rate limiting step for intracellular concentration and determines its potency, identification of a MSeA transporter can aid future studies of MSeA pharmacokinetics. In addition, the selective toxicity of MSeA for cancer cells implies that the expression of an AQP9 membrane transporter may play a role in the outcome of MSeA treatment.</p><!><p>The human AQP9 were cloned into pXβG-ev1, as described previously (Liu et al. 2004; Qi et al. 2012b). Capped cRNAs were synthesized in an in vitro reaction using mMessage mMachine T3 ultra kit (Applied Biosystem) with pXβG-ev1 plasmids linearized with NotI (Liu et al. 2006a). Oocytes from Xenopus laevis were defolliculated and injected with 25 ng of cRNA or with 50 nl of water. They were then incubated in complete ND96 buffer for 3 days at 16 °C and used for uptake assays.</p><!><p>For the assay of selenite and MSeA accumulation in AQP9 expressed oocytes, oocytes with either AQP9 cRNA or water injected were incubated in 1 mM of sodium selenite (Sigma), 100 μM monomethylselilinic acid (Sigma), respectively, at room temperature for 60 min or indicated time. When necessary, oocytes were pretreated by 20 μM carbonyl cyanide 4-trifluoromethoxyphenylhydrazone (FCCP, Sigma), 10 μM phloretin (Sigma), 100 μM valinomycin (Sigma) or 100 μM nigericin (Sigma) for 30 min. When organic acid competitors, including formate, acetate, pyruvate, benzoate and succinate, were used, oocytes were pretreated with these substrates at 1mM of each for 5 minutes prior to adding the tested selenium substrates. Sodium arsenite (AsIII) is added at final concentration of 1mM (Sigma) to study the inhibitory effect. All inhibitory experiments were performed under pH 5.5. The oocytes were then collected and washed in ND96 buffer three times. Oocytes were completely digested using 70% (vol/vol) HNO3 for at least 2 hrs. The samples were then diluted with HPLC grade water for selenium quantification.</p><!><p>For assay of lactate accumulation in oocytes, oocytes were incubated in 1 mM of sodium lactate mixed with appropriate 14C labeled lactate at room temperature for 60 min. When necessary, oocytes were pretreated by FCCP, phloretin, valinomycin, or nigericin for 30 min before transport assay, as described above. When organic acid competitors, including formate, acetate, pyruvate, benzoate and succinate, were used, oocytes were pretreated with these substrates at 1mM concentration for 5 minutes at indicated concentrations prior to adding the tested substrates, as described above. All inhibitory experiments were performed under pH 5.5. After transport, the oocytes were collected and washed in ND96 buffer three times. Oocytes were completely dissolved in 10% SDS and cocktail was added. Radioactivity (CPM) is determined by a scintillation counter.</p><!><p>Total elemental selenium in each sample was determined by inductively coupled plasma mass spectroscopy (ICP-MS) (Nexion 300, PerkinElmer, Norwalk, CT).</p><!><p>All experiments contain at least two batches of oocytes from two animals; at least 3 replicates are used each time. One batch of experiments is used to present in this paper. Quantitative results are shown as means ± standard deviations. The statistical analysis was performed by Student's t test for paired data between control and treated groups. P values <0.05 were considered significant.</p><!><p>The ability of human AQP9 to conduct MSeA in oocytes was examined in a time course. The transport is time-dependent, which does not saturate within 60 minutes (Fig. 1A). AQP9 transport of MSeA was examined under different pH conditions (pH 4.5, 5.5, 6.5 and 7.5, respectively) (Fig. 1B). Results showed AQP9 has a high efficiency to facilitate MSeA with a preference of lower pH, but significant MSeA uptake is still observed at physiological pH. In order to know whether other organic acids can affect AQP9-mediated MSeA transport, we performed the transport assay in the presence of different organic acid substrates. Our results showed that these organic acids have no significant inhibition for MSeA uptake. However, the inorganic AsIII, one well characterized substrate for AQP9, can effectively inhibit MSeA transport (Fig. 1C). In order to investigate whether AQP9 mediated transport depends on transmembrane proton potential or pH gradient. Four different inhibitors including FCCP, valinomycin, pholoretin, and nigericin were applied in the transport assay., As is shown in Fig. 1D, these inhibitors has no effect on MSeA uptake via AQP9, indicating transport is independent on transmembrane potential and pH gradient. However, the AQP9 specific inhibitor phloretin exhibits substantial inhibition of MSeA uptake (Fig. 1D).</p><!><p>Human AQP9 transport of lactate in oocytes was investigated by transport assay. AQP9 transportation of lactate is time dependent and stays linear within 60 minutes, indicating human AQP9 facilitates the uptake of lactate in a time-dependent manner. (Fig. 2A). The ability of human AQP9 to increase lactate permeability at different pH levels from 4.5 to 7.5 was examined in oocytes. The results show that lactate exhibits much lower transport at lower pH, indicating the neutral forms of these compounds are substrates for AQP9 (Fig. 2B). Moreover, lactate analogs including the weak organic acids formate, acetate, pyruvate, benzoate, and succinate showed universal inhibition of lactate uptake, indicating they are competing with lactate in membrane transport via AQP9 (Fig. 2B). In addition, consistent with the result AsIII blocks MseA uptake, AsIII can effectively inhibit lactate permeation as well (Fig. 1C). In contrast, the inhibitors FCCP and nigericin showed effective inhibition of lactate permeation, which supports that AQP9 transports lactate requires a transmembrane proton and pH gradient. Phloretin expectedly exhibits substantial inhibition of lactate uptake (Fig. 2D).</p><!><p>AQP9 transport of inorganic selenite has been shown to have a much lower efficiency compared with MSeA (approximately 1000-times less). The selenite transport resembles the pattern of lactate uptake, demonstrating that selenite and lactate are transported similarly via AQP9. Selenite transport is linear within 60 minutes (Fig. 3A). The selenite transport is in favor of a lower pH, indicating the neutral form selenite is the transported species. There is no visible selenite uptake at physiological pH (pH 7.5) (Fig 3B). Organic acids including formate, acetate, pyruvate, benzoate, and succinate showed inhibition of selenite uptake similar to lactate. AsIII also inhibits selenite permeation (Fig. 3C). Inhibitors of FCCP and nigericin showed effective inhibition of selenite permeation, which supports that selenite transport also requires a transmembrane proton and pH gradient. AQP9 inhibition by phloretin effectively inhibits MSeA, lactate, and selenite uptake in a similar pattern (Fig. 3D).</p><!><p>Inorganic selenite (Se) and organic methylseleninic acid (MSeA) have clinical importance in cancer prevention and treatment. They share different valence, structures, and charges and therefore have distinct mechanisms in inducing cell responses and regulating downstream targets once entering cells.</p><p>MSeA transported by AQP9 can be reasonably explained. MSeA was reported to have a pKa of 8.5, which means it exists mostly as a non-charged neutral molecule at physiological pH. Our results showed that AQP9 serves as an effective transporter to permeate MSeA in a pH range of 4.5–7.5, with a preference for acidic pH. The fully protonated MSeA species is predicted as the substrate for translocation, similarly to that of AQP9 transport of inorganic arsenite (AsIII), which is also a neutral form at physiological pH. The AsIII substrate has significant inhibition of MSeA uptake (Fig. 1C), which supports this prediction. This hypothesis is further supported by the observation that the inhibitors of the transmembrane electron potential and proton gradient, including FCCP, valinomycin, and nigericin, do not inhibit MSeA uptake. The result of the pH preference of MSeA transportation could explain the selective toxicity of MSeA to cancer tissue. Cancer tissues have been known to have a lower pH and therefore they can take up more MSeA than normal tissues. It is a plausible prediction that expression of AQP9 controls the MSeA availability and determines MSeA toxicity in cancers, which requires more studies. AQP9 expression was detected in multiple tumor cells. For example, in human hepatocellular carcinoma, AQP9 expression was decreased compared with non-tumourigenic liver tissue (Jablonski et al. 2007). AQP9 levels in human astrocytic tumors were positively related to physiological grade(Tan et al. 2008). In addition, microarray analysis found that AQP9 mRNA was lower in the adjuvant chemotherapy nonresponse colorectal cancer patients (Dou et al. 2013). Therefore, manipulation of AQP9 expression, as well as AQP9-mediated drug sensitivity would be a promising anti-cancer therapy. This strategy has been investigated in leukemia treatment by using AsIII. Overexpression of human AQP9 in K562 leukemia cells was applied to increase AsIII sensitivity (Bhattacharjee et al. 2004).</p><p>Lactate is a weak acid with a pKa of 3.85 and exists as a monovalent anion under physiological pH. Lactate permeation by AQP9 has been previously reported (Tsukaguchi et al. 1998; Tsukaguchi et al. 1999a). Mechanisms of AQP9 anion transport have been discussed, and AQP9 is proposed to act as a channel for the protonated lactic acid form (Rambow et al. 2014). Since humans have a large family of monocarboxylate transporters (MCTs) expressed in all tissues(Halestrap 2013), the physiological role of AQP9 in lactate transport is questionable.</p><p>In our study, we compared AQP9 transportation of lactate and selenite since selenite is an analog of lactate and transported by the lactate transporter Jen1p in yeast. Our results showed that selenite is a weak substrate for AQP9 with a preference of acidic pH, with a low efficiency comparable to lactate uptake. In addition, the transport is not observed at physiologic pH, it is therefore predicted that AQP9 selenite transport does not represent a major pathway under normal physiological conditions. However, at some acidic microenvironments, such as the stomach or cancerous environments, selenite could have an improved permeation rate by AQP9. Under most situations, the direct selenite transport is believed to be mediated by ZIP8, a ZIP family member which has been identified to transport selenite under neutral conditions with cofactors Zn and bicarbonate. Roles of these two different selenite transporter systems under different conditions are illustrated in Fig. 4.</p><p>Results from inhibitory experiments by different lactate analogs showed that they inhibit both lactate and selenite permeation with a similar pattern. Organic acids including formate, acetate, pyruvate, benzoate, and succinate can compete with lactate and selenite transport via AQP9, indicating these organic acids are likely to be substrates of AQP9 as well. Lactate and selenite inhibit each other reciprocally, which supports the hypothesis that they share similar transport mechanisms and compete with each other for AQP9 transport. On the contrary, these organic acids had no effect on MSeA uptake via AQP9, indicating AQP9 transports MSeA in a different mechanism. One well characterized AQP9 substrate, AsIII, can inhibit both selenium species and lactate permeation. Likely transportation of these substrates share partial similarity in substrate translocation.</p><p>Both lactate and selenite were transported under lower pH conditions. The lower pH would alter the transmembrane gradient and membrane potential. Therefore, the dissipation of these gradients by the inhibitor FCCP significantly decreases lactate and selenite transport via AQP9. Nigericin, which disrupts pH concentrations across membranes, also blocks AQP9-mediated lactate and selenite transport, indicating the transmembrane ΔpH is a critical factor for effective transport. Phloretin, another inhibitor of AQP9 that inhibits both water and glycerol permeation (Ishibashi and Sasaki 1998; Tsukaguchi et al. 1999b), completely inhibited transport of MSeA (Fig. 1D). Instead, valinomycinhas no effect on lactate, selenite, and MSeA uptake, which indicates the electron potential is not involved in driving substrate movement. These results suggest selenite and lactate share the same translocation pathway, while MSeA, a neutral molecule, may transport similarly to that of AsIII and MAsIII.</p><p>MSeA was transported at much higher rate than that of selenite. Considering different concentration levels used in the assays (100μM versus 1mM), the uptake of MSeA is approximately 1000 times higher than that of selenite at pH 4.5–6.5. Consistent with this, MSeA is also a more toxic selenium form for liver cells, while liver cells are more resistant to selenite (Li et al. 2013; Zhao et al. 2004). The selective uptake for methylated species was also observed in arsenic substrates, which monomethylated arsenic has 5-time higher efficiency than the inorganic arsenite with same valence(Liu et al. 2006b). This shows that molecular polarity and hydrophobicity are critical factors involved in AQP9 permeation.</p><p>AQP9 transport many neutral nutrient molecules and multiple arsenic species(McDermott et al. 2010a). The current new findings additionally address the multiple roles of AQP9 in the uptake of therapeutic selenium compounds, particularly MSeA. The dependence of AQP9 in MSeA transport renders AQP9 as an important consideration for future MSeA application. More work in cultured cell lines and experimental animals with altered AQP9 expression are required to investigate the physiological and pharmacological relevance of this transport.</p>

PubMed Author Manuscript

Photo-induced carbocation-enhanced charge transport in single-molecule junctions

We report the first example of photo-induced carbocation-enhanced charge transport in triphenylmethane junctions using the scanning tunneling microscopy break junction (STM-BJ) technique. The electrical conductance of the carbocation state is enhanced by up to 1.5 orders of magnitude compared to the initial state, with stability lasting for at least 7 days. Moreover, we can achieve light-induced reversible conductance switching with a high ON-OFF ratio in carbocation-based single-molecule junctions.Theoretical calculations reveal that the conductance increase is due to a significant decrease of the HOMO-LUMO gap and also the enhanced transmission close to the Fermi levels when the carbocation forms. Our findings encourage continued research toward developing optoelectronics and carbocationbased devices at the single-molecule level.

photo-induced_carbocation-enhanced_charge_transport_in_single-molecule_junctions

Introduction<!>Results and discussion<!>Conclusions

<p>Since its inception, molecular electronics has aimed for functional electronic devices using individual molecules. 1 Understanding and investigations of charge transport through single molecules provide critical information for the design of molecular-scale devices. [2][3][4][5] Various molecular species are being employed in single-molecule devices for regulating charge transport properties in single-molecule junctions, for instance oligoynes, 6 oligo(phenylene ethynylene), 7 organic radicals, 8 DNA, 9 and peptides. 10 Although carbocations have been widely found in many chemical reactions, [11][12][13] to date, the applications of carbocation-based molecules as building blocks to fabricate stable and highly conductive molecular devices remain highly challenging due to the intrinsic instability of a majority of carbocations.</p><p>On the other hand, an external stimulus for tuning charge transport through molecular junctions also plays a key role in the fabrication of single-molecule devices. Previously several external stimuli such as light, 14 pH, 15 electric elds, 16 mechanical forces, 17 solvents, 18 and electrochemical gates 19 were used to tailor the electronic properties of single-molecule junctions. Among these stimuli, light has an advantage in terms of its remote manner, non-invasiveness, and high spatiotemporal resolution. 20 Therefore the utilization of light to tune charge transport may provide a unique strategy for creating new conceptual molecular devices.</p><p>Triphenylmethane leuco derivatives are well-known photochromic molecules, which dissociate into ion pairs under ultraviolet irradiation, producing stable carbocations in the form of triarylmethane dyes and hydroxide ions. [21][22][23] The stable carbocation dyes have re-emerged as a highly efficient Lewis acid catalysts for a variety of organic transformations. Investigations of this reaction have revealed that the dissociation processes are very fast and proceed with high quantum yields. 22 Such prominent changes in pH have received wide interests as light-induced pH-jump reactions for many applications, 22,23 and also offer potential opportunities to study carbocation-based charge transport properties at the single-molecule level.</p><p>In the present work, we selected malachite green leucohydroxide (MGOH) molecules as carbocation emitters to explore laserinduced charge transport in single-molecule junctions. Using the scanning tunneling microscopy break junction (STM-BJ) technique, 24 the single molecule conductance of MGOH and the corresponding junction elongation can be evaluated (Fig. 1a). The dimethylamino moieties at both ends of MGOH serving as the anchoring groups are used to build efficient transport junctions. In the presence of 302 nm UV light, MGOH produces, with high efficiency, malachite green carbocations (MG + ) and hydroxide ions, thus inducing a minor structural change where the orbital hybridization of the central carbon atom of initial states changes from sp 3 to sp 2 . Such a photo-triggered structural transformation shis the HOMO-LUMO gap and we demonstrate that the carbocations signicantly enhance the conductance and the conductance switching could be highly reversible. The large enhancement in single-molecule conductance suggests that the transport pathway of charges in the carbocation state is markedly different to that of initial states, and it is further supported by the DFT calculations that Breit-Wigner resonance occurs in the carbocation state, leading to a large ON-OFF ratio in the junctions.</p><!><p>To conrm the formation of the carbocations, we performed UV-Vis absorption spectroscopy of MGOH before and aer UV illumination as shown in Fig. 1b. Before illumination, MGOH exhibits a single peak at 299 nm. Aer illumination at 25 C, two prominent peak bands at 416 nm and 610 nm appeared and their intensities progressively increased with time, which correspond to the characteristic of the triphenylmethane carbocation. 16,18 In situ photo-radiation of MGOH results in the orbital hybridization of the central carbon atom changing from sp 3 to sp 2 . The formation of the carbocation states is also conrmed by 1 To ensure these features are statistically reproducible, we repeated the measurements thousands of times and compiled the traces into one-dimensional (1D) and two-dimensional (2D) conductance histograms. From the 1D conductance histograms in Fig. 2b, we observed that the molecular conductance peak for MGOH appeared at $10 À4.91 G 0 (0.9 nS), while the molecular conductance for MG + at $10 À3.40 G 0 (30.9 nS) is 34 times higher than that of MGOH. The corresponding 2D histograms for MGOH and carbocations are displayed in Fig. 2c and d. Both molecules exhibit distinct molecular plateaus and intensity clouds. The junction elongations for both molecular states show the approximate length of around 0.5 nm (insets of Fig. 2c and d), suggesting the structural change of molecular junctions while the binding conguration of the molecular junctions remained similar. We also calculated the formation probability 25 for both molecular junctions, and the results show that the junction formation probabilities are between 40% and 50% (Fig. S6 in the ESI †), indicating that the anchoring ability of both molecule species to Au electrodes is similar.</p><p>Control experiments were carried out by measuring the conductance of 4,4 0 -bis(dimethylamino)triphenylmethane leucohydroxide (LMG). It is found that photo-irradiation of a solution of LMG does not induce any detectable change of conductance within the uncertainty of the measurements (Fig. S7 and S8 in the ESI †). These results present unambiguous evidence that the prominent increase of the conductance in triphenylmethane junctions aer illumination is ascribable to the photo-induced formation of carbocations. We noted that LMG has a higher conductance compared to MGOH, and the rigorous and fully theoretical explorations on the highconductance for LMG need further study. To study the stability of the carbocations, we measured the time-dependent conductance evolution of carbocations in a low-polar solvent (THF/TMB, v/v ¼ 1 : 4, 0.1 mM) that we used and found it was stable for at least 7 days (Fig. 3a). The ESI-MS results and UV-Vis spectrum further indicate that the material is not completely photo-degraded aer 7 days (Fig. S9-S12 in the ESI †). The high stability of MG + is probably due to the fact that sulfonated triphenylmethyl carbocations and hydroxide ions form loose ion pairs in a low-polar solvent. [26][27][28] Previous studies reported that triphenylmethyl carbocations aren't stable and can only exist in water (high-polar) for dozens of minutes. 22,23 To further probe the effect of solvent polarity on carbocation stability, we dissolved MGOH molecules in water (0.1 mM), illuminated it with UV light and measured its single-molecule conductance. It is found that the high conductance states lasted for only 60 min (Fig. 3a) in aqueous solution, which is consistent with previous reports, suggesting the stability of carbocations is affected by solvent polarity. By tuning the solvent polarity, the switching could be designed to be irreversible or reversible. In aqueous solutions, the in situ light-induced reactions involving carbocations are reversible and at least three cycles of conductance switching can be achieved as shown in Fig. 3b (Fig. S13 in the ESI †). Carbocation-based junctions in low-polar environments with high stability can serve as conducting interconnects for durable electrical circuits, while carbocation-based junctions with low stability in high-polar environments can act as a light-reversible single-molecule switch. These results suggest carbocations are one of the potential multifunctional building blocks for molecular electronics. Compared with the molecular devices constructed by common molecular systems such as spiropyran 29 and azobenzene, 30 this work based on carbocation molecular species of interest broadens the scope of building blocks in molecular electronics.</p><p>To understand the charge transport properties of initial and carbocation states, we turned to density functional theory (DFT) calculations. We rst calculated the Au-N binding energies in both molecular junctions and the results show that the Au-N binding energies for MGOH and the carbocations are respectively À0.89 eV and À0.41 eV (see Fig. S14 and S15 in the ESI †), which imply that dimethylamino moieties in both molecular species can bind to gold electrodes for stable molecular junctions. Next, we calculated their transmission spectra using the non-equilibrium Greens function (NEGF) formalism with DFT. [31][32][33] Fig. 4a shows the transmission spectra for MGOH and carbocations, which represent the probability that an electron with given energy will transmit through the molecule between the electrodes (see the conformations of both molecules in Fig. S16 in the ESI †). DFT calculations indicate that the HOMO-LUMO gap changes from 4.65 eV for MGOH to 2.59 eV for MG + when the carbocation forms (see frontier molecular orbitals and their energy levels in Fig. S17 and Table S1, † and carbocationbased molecular conformations with different counter-ions and different positions and corresponding transmission spectra in Fig. S18 and S19 in the ESI †), which is also conrmed by the red-shi of the UV-Vis spectra in Fig. 1b. Moreover, it is also found that carbocations create a new resonance in the energy level close to Fermi levels (E F ), which is responsible for the boosted conductance phenomena through carbocation-based transport junctions. We also note that this resonance in the transmission spectra has a line-shape of the Breit-Wigner distribution. [34][35][36] The transmission for carbocation states at the Fermi level is more than one order of magnitude higher than that for MGOH, which is in good agreement with the experimental observation. Fig. 4b shows a schematic energy diagram of the position of molecular orbitals of carbocations respective to E F of the two electrodes before and aer light illumination. Before illumination, the energy-level between molecular orbitals of MGOH and E F is not aligned, suggesting a large energy barrier in charge transport through the HOMO; while a resonance peak is highly aligned to E F in carbocation-based junctions aer illumination, which reduces the energy barrier in charge transport and enhances the conductance through LUMO energy liing. Together, the overall experimental and theoretical studies demonstrate that the in situ light-induced approach produces carbocations, which greatly enhance charge transport in triphenylmethane single-molecule junctions.</p><!><p>In summary, we have demonstrated that carbocations induced by light greatly enhance charge transport in triphenylmethane junctions using the STM-BJ technique. The electrical conductance of the carbocation state is enhanced by up to 1.5 of magnitude compared to the initial state and the conductance switching could be reversible. The stable and conductive triphenylmethane carbocations may serve as promising building blocks for future molecular electrical devices. DFT calculations reveal that the conductance increase is due to a signicant decrease of the HOMO-LUMO gap and also the enhanced transmission close to the Fermi levels when the carbocation forms. The large conductance enhancement by a lightinduced strategy in carbocation-based molecular junctions may be useful for single-molecule optoelectronics and carbocationbased molecular devices.</p>

Royal Society of Chemistry (RSC)

Going beyond the borders: pyrrolo[3,2-b]pyrroles with deep red emission†

A two-step route to strongly absorbing and efficiently orange to deep red fluorescent, doubly B/N-doped, ladder-type pyrrolo[3,2-b]pyrroles has been developed. We synthesize and study a series of derivatives of these four-coordinate boron-containing, nominally quadrupolar materials, which mostly exhibit one-photon absorption in the 500–600 nm range with the peak molar extinction coefficients reaching 150 000, and emission in the 520–670 nm range with the fluorescence quantum yields reaching 0.90. Within the family of these ultrastable dyes even small structural changes lead to significant variations of the photophysical properties, in some cases attributed to reversal of energy ordering of alternate-parity excited electronic states. Effective preservation of ground-state inversion symmetry was evidenced by very weak two-photon absorption (2PA) at excitation wavelengths corresponding to the lowest-energy, strongly one-photon allowed purely electronic transition. π-Expanded derivatives and those possessing electron-donating groups showed the most red-shifted absorption- and emission spectra, while displaying remarkably high peak 2PA cross-section (σ2PA) values reaching ∼2400 GM at around 760 nm, corresponding to a two-photon allowed higher-energy excited state. At the same time, derivatives lacking π-expansion were found to have a relatively weak 2PA peak centered at ca. 800–900 nm with the maximum σ2PA ∼50–250 GM. Our findings are augmented by theoretical calculations performed using TD-DFT method, which reproduce the main experimental trends, including the 2PA, in a nearly quantitative manner. Electrochemical studies revealed that the HOMO of the new dyes is located at ca. −5.35 eV making them relatively electron rich in spite of the presence of two B−–N+ dative bonds. These dyes undergo a fully reversible first oxidation, located on the diphenylpyrrolo[3,2-b]pyrrole core, directly to the di(radical cation) stage.

going_beyond_the_borders:_pyrrolo[3,2-b]pyrroles_with_deep_red_emission†

Introduction<!>Design and synthesis<!>X-ray analysis<!><!>Photostability<!>Two-photon absorption<!>First-principle studies<!><!>First-principle studies<!>Electrochemical properties<!>Spectroelectrochemistry<!>Conclusions<!>Data availability<!>Author contributions<!>Conflicts of interest

<p>In recent years, boron has found new distinct roles in materials science, being incorporated in various polycyclic aromatic hydrocarbons (PAHs).1–7 It has been discovered that doping with three-coordinate, sp2-hybridized boron has the most pronounced effect on aromatic systems among main group elements, due to its low Pauling electronegativity and enhanced π-conjugation resulting from its vacant pz orbital. As a consequence, in such systems boron atoms can act simultaneously as π-electron acceptors and σ-electron donors. The fascinating properties of the materials obtained in such a way (so-called B-PAHs), such as strong fluorescence8–11 and enhanced charge-transport characteristics,12,13 has resulted in a wide range of applications in optoelectronics, including materials for organic light-emitting diodes (OLEDs)14–16 circularly polarized luminescence,17 organic photovoltaics (OPVs)18 and organic field-effect transistors (OFETs)19–21 as well as materials for electrodes in lithium batteries.22 Much attention has also been paid to the B–N/O isosteres of PAHs, i.e. compounds in which two adjacent carbons in a π-conjugated core have been replaced by one boron and one N/O atom.23–28 Of particular importance are π-conjugated systems containing a B–N covalent bond, B←N coordination bond29 and N–B←N motif, as evidenced by growing number of reports dealing with BN-embedded heteroacenes29–32 and BODIPY analogues,33–37 mostly due to their excellent performance in OLEDs38a and OPVs.38b,c They were also computed to possess inverted singlet-triplet gap.39 In such systems, the N–B–N motif not only constrains the π-conjugated skeleton in a coplanar fashion, but also considerably lowers the LUMO level, making the core of the dye a strong acceptor.</p><p>1,2,4,5-Tetraarylpyrrolo[3,2-b]pyrroles (TAPPs)40 are aza-analogues of well-known thieno[3,2-b]thiophenes.41 Recent synthetic breakthroughs,42 prompted their application in research related to studying symmetry breaking in the excited state,43 solvatofluorochromism,43b,c direct solvent probing via H-bonding interactions44 photochromic analysis of halocarbons45 organic light emitting diodes46 resistive memory devices,47 bulk heterojunction organic solar cells,48 dye-sensitized solar cells,49 aggregation-induced emission50 and MOFs.51 Furthermore, the very high reactivity at positions 3 and 6 of the pyrrolo[3,2-b]pyrrole core makes them very convenient starting materials for the construction of ladder type heterocycles. Taking advantage of this feature, many reports have been published on the synthesis and properties of TAPP-based PAHs,52 however, there has only been one successful synthesis of BN-embedded TAPPs (Fig. 1).53 The resulting dyes, containing boron atoms incorporated into six-membered cycles, exhibited very high absorption coefficients and strong fluorescence, both in solution and the solid state, with very small Stokes shifts.</p><p>Although the combination of suitable photophysical properties have made TAPPs and fused TAPP analogues popular chromophores for a range of applications, shifting their emission to the red region of the visible spectrum has proved difficult. We have sought to address this issue in this work. Here, we have considered the reaction of pyrrolo[3,2-b]pyrroles bearing 2-pyridil substituents in positions 2 and 5 (C <svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="13.200000pt" height="16.000000pt" viewBox="0 0 13.200000 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.017500,-0.017500)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z"/></g></svg> N bonds in N-heteroaromatic rings are effective in coordinating with boron moieties) with an appropriate boron source (Scheme 1). The resultant B←NC five membered chelate ring fuses the pyrrolopyrrole and pyridine substituents together and effectively fixes the π-conjugated framework in a coplanar fashion resulting in the formation of a unique chromophore, which perfectly fits into the current intense research on BN-embedded heteroacenes.</p><!><p>Our approach towards new BN-embedded heterocycles capitalizes on pyrrolo[3,2-b]pyrroles' intrinsically high reactivity at positions 3 and 6, and on the straightforward access to derivatives bearing pyridyl substituents at positions 2 and 5 (Scheme 1). The necessary TAPPs 2a–2e were obtained in good yields from appropriate formylpyridines and 4-octylaniline using our optimized multicomponent condensation (see ESI† for details).42a TAPPs 2f–2h were formed in lower, but sufficient yields from appropriate quinoline- and isoquinoline-derived aldehydes. We anticipated that the planar structure of BN-embedded heterocycles would result in extensive π-stacking, therefore long alkyl chains were installed on the N-linked aryl substituents in order to secure good solubility of the final products. In the next step, we planned to apply a well-known strategy consisting of treatment of the parent pyrrolopyrrole with BBr3 in the presence of base, however, this led to complete decomposition of the starting material. An alternative route was to use diarylchloroboranes, but this seemed difficult due to the limited availability of such reagents. Fortunately, Song and co-workers recently developed a straightforward cascade B–Cl/C–B cross-metathesis and C–H bond borylation procedure,54 which we successfully used in the synthesis of the desired four-coordinated boron-containing pyrrolo[3,2-b]pyrroles 1a–1h (BN-TAPPs) (Scheme 1). According to Song's findings, in this procedure aryltrifluoroborate reacts with SiCl4 to form aryldichloroborane in highly selective manner, which then reacts with another molecule of aryltrifluoroborate to give diphenylchloroborane. Subsequent pyridine directed electrophilic aromatic borylation leads to four-coordinate triarylborane. The yields in the last step are greatly improved when sterically hindered organic base is applied. Importantly, this simple procedure can be conveniently applied to starting materials containing both pyridyl and quinolinyl substituents, and it tolerates the presence of halogens in the starting materials. Consequently, we prepared a few BN-embedded TAPPs substituted with various halogens and submitted them to Sonogashira or Buchwald–Hartwig cross-couplings, in order to further expand the π-system. Although the outcomes of these reactions were not obvious due to the presence of reactive boron and halogen substituents, the desired compounds 1i–1l (Schemes 2 and 3) were isolated in acceptable yields.</p><!><p>The molecular structure of 1a was determined by X-ray crystallography. Single crystals of 1a suitable for analysis were obtained at ambient temperature by slowly diffusing hexanes into a tetrahydrofuran solution. Dye 1a crystallizes in the monoclinic space group P21/c, and the unit cell comprises of two molecules (Fig. 2). The pyrrolo[3,2-b]pyrrole core adopts a perfectly planar structure (180° N–C–C–N torsion angle) and the peripheral pyridine rings are both twisted by 5° from this plane.</p><p>Benzene rings attached to nitrogen atoms of the pyrrolopyrrole core are twisted by 59° and 61°, while phenyl substituents attached to boron atoms are twisted by 83° and 78°. The lengths of the newly formed B–C bonds are 1.6 Å, which is a slightly shorter value than those reported for related compounds and suggests a stronger bond,35,55 while the lengths of the B–N bonds are 1.63 Å, which is similar to the B–N bond length found in analogous compounds55 and BODIPY-type molecules.56 Van der Waals interactions govern the crystal packing, with the main structural motif being the interaction of bent C8 aliphatic chain with two phenyl substituents attached to the boron atom adjacent to the same molecule and then with the octylphenyl substituent attached to the nitrogen atom of the neighboring molecule. No obvious π-stacking interactions are observed, probably due to the orientation of phenyl substituents on the boron atom that prevents these heteroacenes with extended π systems from approaching each other.</p><p>To characterize the photophysical properties of the novel dyes we conducted a multipronged campaign. The optical properties of TAPPs 2a–2h correspond very well to those reported earlier for this class of dyes (see ESI† for details).42a The incorporation of boron into the TAPPs backbone, however, brings spectacular changes in their optical properties (Table 1 and Fig. 3). The absorption maxima for compounds 1a–1g are bathochromically shifted by 122–155 nm when compared to parent TAPPs. Only for compound 1h, a somewhat lower, but still remarkable 68 nm red-shift of the absorption maxima was observed, accompanied by an unusual broadening of the band shape and by a low intensity band between 600 and 700 nm (see ESI†). Theoretical calculations indicate that the latter is most likely not connected to the isolated 1h structure (see below), whereas preliminary measurements did not detect formation of aggregates or other related structures. Similarly, the emission maxima were red-shifted by over 100 nm, with very small Stokes shifts of less than 1000 cm−1 and large fluorescence quantum yields for most of the dyes. These results clearly indicate a rigidifying effect of the boron component thanks to the character of boron atoms. Noticeably, in the case of compounds 1k and 1l, in which the π-systems were further expanded, the barrier of 600 nm for both absorption and emission maxima was broken, which means that these are the first examples of pyrrolo[3,2-b]pyrrole derivatives absorbing and emitting in the deep-red region. For dyes 1i, 1k and 1l, exceptionally high extinction coefficients were observed, exceeding 100 000 M−1 cm−1, together with large Φfl. This combination of properties is particularly rare for deep red-emitting dyes due to additional deactivation pathways by vibrational relaxation (energy gap law). A very weak emission was detected at 670 nm for dye 1h. The excitation spectrum however, is not consistent with the absorption profile, which points at potential more complex underlying photophysics to be addressed in subsequent investigations. When compared to recently published, structurally related B/N-doped p-arylenevinylene chromophores,35 BN-TAPPs absorb and emit at much longer wavelengths, with the difference being most noticeable for the derivatives bearing similar, unsubstituted quinoline moieties (63 nm and 80 nm bathochromic shift in the absorption and emission maxima, respectively). p-Arylenevinylene analogues of BN-TAPPs exhibit higher fluorescence quantum yields which are close to being quantitative, although the latter compounds are characterized by much higher molar extinction coefficients, which overall results in a similar emissive brightness for these two classes of organic chromophores.</p><!><p>Determined with fluorescein in NaOH (0.01 M) as a standard.</p><p>Determined with cresyl violet perchlorate in MeOH as a standard.</p><p>Fluorescence of dye 1h was not investigated in 2PA due to the inconsistency of the one-photon fluorescence excitation results.</p><!><p>Photostability tests were carried out for compounds 1a and 2a as representative examples (Fig. 4). TAPPs, which generally exhibit very poor stability under irradiation, acquire remarkable photostability upon incorporation of boron, as demonstrated by comparing the properties of 1a with common dyes, such as BODIPY and diketopyrrolopyrroles. Even prolonged exposure to a strong light source did not cause any noticeable changes in the absorption spectrum of 1a.</p><!><p>Two-photon absorption studies were performed using a two-photon excited fluorescence (2PEF) method (Table 1 and Fig. 5). Similarly to previously studied symmetric TAPPs,40,43b,c the most prominent two-photon transitions are observed at higher energies, well above the lowest-energy one-photon allowed S0→S1 transition. The latter is only very weakly present in the two-photon spectra, thus following the Laporte rule for low-energy bands. At energies above the S0→S1 transition, the 2PA profiles of 1a-e show similar features comprising a double-band or vibronic progression at 800–950 nm and a distinct peak at ca. 650–670 nm. At <600 nm there is a steep increase of the cross section due to the near-resonance enhancement effect, followed by onset of linear absorption at even shorter wavelengths (not shown). When pyridine is replaced with quinoline (1f, 1g) and further with the addition of diarylamino groups, these spectral features undergo a gradual bathochromic shift. Two-photon absorption was not investigated in the case of dye 1h for two main reasons: (1) the fluorescence intensity is too weak for 2PEF method to be used; (2) the perplexing photophysics of this BN-TAPP has to be delineated before intrinsically more complex non-linear phenomena will be studied. The substantial change, however, comes only after adding strong electron-withdrawing groups at the periphery (dyes 1j and 1k).</p><p>In such case the peak σ2 becomes remarkably large, exceeding 103 GM. In the case of the dyes 1k and 1l, the peak values at 760–870 nm reach σ2 = 2500 GM and 1800 GM, respectively, whereas strongly resonance-enhanced values at higher energy become even larger. These values are significantly higher than for simpler A–D–A type TAPPs.43b,c An important feature to note is that the values are higher for the case with the peripheral diaryl-amino groups which would suggest that the new heterocyclic core is somewhat electron-withdrawing.</p><!><p>Time-Dependent Density Functional Theory (TD-DFT) has been used to explore the nature of the excited-states involved in the BN-TAPPs 1a–1k. The technical details are given in the ESI.†Table 2 summarizes the main results obtained by the calculations.</p><!><p>Nearly dark state after optimization.</p><!><p>Let us start our analysis by the one-photon absorption. In the full series (except for 1h), the S0→S1 transition is well separated from the next one, and is strongly dipole allowed with oscillator strengths ranging from 0.71 to 2.33. The S0→S1 transition undoubtedly corresponds to the observed most red-shifted absorption band. Experimentally, this absorption band is typically composed of two peaks and a shoulder, resulting from vibronic couplings. To ascertain this statement, we have performed vibrationally resolved calculations within the FC-AH approach57 for a simplified version of dye 1f (see Fig. S4†). As can be seen, the overall topology of both the absorption and emission bands is reproduced confirming the vibronic nature of these two peaks. For 1h, the three lowest excited-states are very close according to theory (Table 1), with differences below the TD-DFT accuracy, so that it is not possible to have a definitive information regarding their ordering.</p><p>While one cannot compare directly computed vertical absorption wavelengths to experimental λmaxabs, it is also obvious that the trends in the series are reproduced by calculations. Indeed, taking 1a as reference, one notices that the strongest red-shift is obtained for 1c (theory: +17 nm, experiment: +26 nm) in the 1a–1e series. For the six dyes showing the most red-shifted absorption, the experimental order is 1j ∼ 1i < 1f < 1g < 1l ∼ 1k, whereas theory yields: 1j < 1f < 1i ∼ 1g < 1l ∼ 1k.</p><p>For all investigated dyes but 1j, the second and third excited states are very close on the energy scale, but possess vastly different oscillator strengths, which is a consequence of the nearly centro-symmetric nature of the investigated dyes. This second (or third) state corresponds to the second weaker absorption band found in the 380–420 nm region experimentally (Tables 1 and 2). For dye 1a, we show density difference plots in Fig. 6. For all three states the central pyrrolopyrrole unit acts as a donor group and the side pyridine rings are accepting moieties. The quadrupolar CT character is, however, not strongly marked for the lowest transition and becomes significantly stronger in the two higher ones. As can be seen, the nature of the state is conserved when going to 1f (derived from quinoline), the contribution of the additional benzene rings being trifling for the first state and remaining limited in the two higher states. In dye 1i, the lowest excited state is partially delocalized on the ethynyl bridges, whereas in 1j, the electron-donating amino groups play their expected role (Fig. 7) explaining the observed redshifts.</p><p>As far as fluorescence is concerned, for only one dye of the series, namely, 1h, the optimization of the excited-state geometry led to a nearly dark state (f = 0.009), suggesting a very slow radiative decay rate. This is consistent with the trifling emission quantum yield measured experimentally (see Table 1). Indeed the computational studies revealed that the lowest energy allowed transition is markedly blue-shifted placing is energetically close to the dark state. The fact that this pair of opposite-parity electronic states may become almost isoenergetic corroborates the anomalous photophysical behavior of dye 1h (vide infra). Apart from this specific case, one globally finds similar substituent effects as for absorption. In Table 2, we also report the 0–0 wavelengths that offer more physically well-grounded comparisons with experiment,57b and more precisely, with the crossing point between absorption and emission. Illustratively, for 1a, 1j and 1l, we compute λ0–0 of 463, 505, and 560 nm, respectively. These values are blue-shifted by +0.26, 0.31, and 0.24 eV as compared to the measured values of 512, 577, and 629 nm. These errors are on the upper side of the typical TD-DFT errors, which can be explained by the very specific nature of these boron-containing systems: TD-DFT is known to overshoot the transition energies of such states.58</p><p>We have also computed two-photon absorption cross sections. In 1a, the experimental rather broad response around 800–900 nm is due to the S0→S2 transition that is dark in one-photon, a typical outcome in quadrupolar dyes. Theory provides a peak σ2PA of 94 GM, of the same order of magnitude as the experimental response (45 GM). This response remains of the same order of magnitude but is slightly increased in 1b, 1c, and 1d, but slightly decreased in 1e, which holds both in the experiment and in the simulations. In 1f, the σ2PA attains the respectable value of ca. 220 GM experimentally, and it is clear that it is again the second state that is responsible for this response (418 GM theoretically). The situation is totally similar in 1g. In the more extended 1i, the calculations indicate that the 2PA response comes from the S0→S3 transition, which is consistent with the experimental spectra. Again, theory gives value that closely matches the experiment (1500 GM versus 1400 GM). In 1i, the experimental 2PA spectrum shows two bands, one moderately active due to the second transition, and one much more intense due to the fifth state. In 1k, the second and third states appear at almost the same energy but have different (pseudo) symmetry, one being active in one-photon, the other in two-photon absorption, potentially explaining the shape of the experimental spectra resembling a superposition between the two phenomena. The computed response is 1600 GM for the 2PA-active transition, again with the same order of magnitude as the experimental measurement. Eventually, the analysis is more difficult in 1l, but it appears that the second excited state should be responsible for the shoulder in the 2PA experimental spectrum at ca. 475 nm, whereas the fourth transition likely yields the much more intense peaks at ca. 440 nm.</p><!><p>BN-TAPPs 1a–1l are characterized by the presence of a reversible redox pair of the first oxidation process (Fig. 8, red CVs, Table 3) which according to its potential value can be tentatively attributed to the oxidation of the N,N′-diphenylpyrrolo[3,2-b]pyrrole core. An exception is 1c where partial reversibility was found, the cause of which was further investigated during spectroelectrochemical measurements (vide infra). The lowest value of the first oxidation potential in the entire BN-TAPP series (0.11 V) was recorded for the derivative 1h with isoquinoline scaffolds. For the remaining dyes the first oxidation occurs between 0.25 and 0.45 V which corresponds to HOMO values in the range (−5.26 eV)–(−5.43 eV) (Table 3). Compared to structurally analogous dyes possessing two thiophene rings BN-TAPPs are slightly less electron-rich but they are more electron-rich than dyes π-expanded TAPPs possessing two thiophene-S,S-dioxide moieties.52b</p><p>The second oxidation curve is partially irreversible in all cases (Fig. 8, green and black CVs) and shows a dual nature more visible in differential pulse voltammetry (DPV) experiments (Fig. 8, e.g.1a and 1b – inserts). The double peak of the second oxidation state may indicate a slight difference in oxidation potential between both peripheral units, which was also registered in separate processes as green and black CV curves with the peak potential marked as E2aox and E2box, respectively. Dyes 1h and 1d possess, respectively, the lowest (0.74 V) and highest (1.02 V) potential of E2aox. Irreversibility of the second oxidation does not impact on the potential and reversibility of the first redox couple in subsequent CV cycles after polarity reversal.</p><p>DPV measurements also show that the peak areas of the first and second oxidation peaks are comparable, which indicates the full oxidation of the N,N′-diphenylpyrrolo[3,2-b]pyrrole scaffold under the potential of the first oxidation peak to dication or di(radical cation), while the second oxidation peak is associated with the oxidation of both the peripheral moieties.</p><p>One irreversible reduction process was registered for BN-TAPPs 1a–1e bearing two pyridine moieties at (−2.40)–(−2.04 V). In the case of dye 1f (bearing two quinoline moieties), we observed two reduction processes, and the first was fully reversible. The values of reduction were shifted towards positive potentials, compared with the pyridine series.</p><!><p>Changes in the UV-Vis-NIR absorption spectra during polarization within the first oxidation peak revealed a decrease in the intensity of the bands at 468, 492 nm (1a) and 495 and 522 nm (1c)59 (Fig. 9), while the shift of the absorption in the infrared direction of the output bands of 1c indicates a decrease in the energy of the π–π* transition due to a partial localization of the HOMO orbital also on the pyridine rings. Polarization within the first oxidation peak causes an increase in the bands at 530, 734 nm (1a) and 579, 773 nm (1c), characteristic for a di(radical cationic) state, as was observed in a previous spectroelectrochemical study of dyes based on the pyrrolo[3,2-b]pyrrole scaffold.52b The gradual formation of a radical cation and its transition to di(radical cation) was not observed.</p><p>ESR spectroelectrochemical measurements were performed for representative dyes 1a and 1c, where the reversible first redox process is associated with the formation of a di(radical cation) in both cases. Unpaired electrons of the di(radical cation) of 1a are located on the N,N′-diphenylpyrrolo[3,2-b]pyrrole core and are characterized by a low value of the g-factor (g = 2.0023). In turn, the ESR signal recorded under polarization of the first oxidation state of 1c gives the g-factor equal 2.0050, which may indicate the coupling of one of the electrons of the di(radical cation) with nitrogen and/or boron atoms, as proposed in Scheme S1.† Although the ESR signals are quite broad in both cases, which may indicate the cleavage of the energy levels of the unpaired electron by the nuclei coupled to it, the hyperfine structure was not recorded.</p><!><p>In summary, we have designed, synthesized and fully characterized a new family of BN-embedded heteroacenes. This was achieved through incorporation of duel cyclopenta[c][1,2]azaborole moieties with a pyrrolo[3,2-b]pyrrole core in a straightforward two-step synthetic procedure. The B−–N+ dative bond reduces the HOMO–LUMO gap of the parent dye, which results in a marked red-shift of absorption and emission and almost quantitative fluorescence quantum yield. Their excellent properties, such as superb photostability, strong absorption and intense emission in the orange to deep red region, together with large two-photon absorption cross sections and rich electrochemistry, opens the door for future applications in optoelectronics. Given the recent renaissance of interest in boron-doped PAHs and their related B–N/O isosteres, this work should inspire the future design and synthesis of pyrrolopyrrole and related frameworks with distinctive π-expanded architectures.</p><!><p>Data associated with this article, including experimental procedures, compound characterization, steady-state absorption and emission along with the two-photon absorption details, electrochemical details and computational analysis details are available in the ESI.†</p><!><p>M. T. conceived the idea and wrote the manuscript. M. T., P. K and M. P. performed all synthetic experiments including condition optimizations and exploring the scope. M. C. and P. J. performed electrochemical and spectroelectrochemical measurements. M. B. performed several TD-DFT calculations, and vibronic analysis. M. R. performed 2PA measurements. M. Ł. wrote formal electrochemical analysis and reviewed the manuscript. A. R. performed 2PA measurements, wrote formal analysis of this part of the manuscript and reviewed the manuscript. D. J. performed DFT and TD-DFT calculations, analyzed data, wrote and reviewed the manuscript. D. G. supervised the project, performed formal analysis, wrote and reviewed the manuscript. All the authors discussed the results and commented on the manuscript.</p><!><p>There are no conflicts to declare.</p>

PubMed Open Access

Real-space Imaging of the Multiple Halogen Bonds by Ultrahigh-resolution Scanning Probe Microscopy

Understanding the physical origin of STM/AFM image contrast is of significance for not only promoting surface characterization techniques, but also probing surface nanostructures with the atomic/sub-molecular resolution. Herein, we demonstrate the real-space imaging of halogen bonds acquired by non-contact atomic force microscopy (nc-AFM)/bond-resolution scanning tunneling microscopy (BR-STM) with functional CO-tip, and study the image contrast origin of halogen bonds. The presence of bright line features is associated to the specific site where halogen bond forms, which is experimentally evidenced to be contributed by both CO-tip bending and electron density exchange. Three distinct types of halogen bonds are observed, which origin from the noncovalent interactions of Br-atoms with positive potential H-atom, neutral potential Br-atom and negative potential N-atom, respectively. Our work shows that nc-AFM and BR-STM can directly image halogen bonds and can be used to unambiguously discriminate their bonding features. This work demonstrates the potential use of this technique to image other non-covalent intermolecular bonds and to understand complex supramolecular assemblies at the sub-molecular level.

real-space_imaging_of_the_multiple_halogen_bonds_by_ultrahigh-resolution_scanning_probe_microscopy

Introduction<!>Results<!>Conclusion<!>Preparation of halogen bonding molecular clusters.

<p>The chemical bond is of central interest in the chemistry discipline, and it plays an important role in the construction of both matter and life. In particular, non-covalent bonds are the fundamental basis of supramolecular self-assembly, 1 including hydrogen bonds, 2 halogen bonds, 3 coordination bonds, 4,5 dipole-dipole interactions 6 and so on. [7][8][9] The halogen bond has recently attracted much interest, 10 while the hydrogen bond has been the most studied intermolecular interaction in the past. 11,12 Halogen bonds form mainly through the mutual interaction between halogen atoms (e.g. F, Cl, Br, I) due to the coexistence of both positive and negative potential centers at the halogen atom caused by its asymmetrical electron distribution. 13,14 Recent research has discovered that halogen atoms also interact with other non-halogen elements (e.g. S and O elements) to form a halogen bond, [15][16][17][18] which further enriches the database of halogen bonds. The study of halogen bonds is useful not only to understand the internal mechanism of supramolecular self-assembly at the molecular level, but also to develop functional nanomaterials constructed via halogen bonds.</p><p>On-surface supramolecular self-assembly based on halogen bonds has also attracted much interest. 10,16 By employing halogen bonding, various ordered supramolecular structures have been reported, such as two-dimensional porous networks, 19 Sierpiński triangle Fractals. 20 Over the past decades, both scanning tunneling microscope (STM) and atomic force microscopy (AFM) have been employed to study various surface structures, molecular assembly and on-surface reactions. 21,22 However, the realization of atomic-resolution molecular structures is challenging for the conventional AFM and STM technique. The invention of qPlus AFM sensor greatly improves the resolution of AFM image and resolves the atomic structure within a single molecule. 23,24 A metallic tip modified with functional molecules (most commonly CO-tip) further improves the AFM resolution. 25 Leo et al. has demonstrated the chemical structure of a single pentacene molecule resolved using qPlus atomic force microscopy (or nc-AFM) with CO-tip. 26 The similar technique can also be applied for the internal chemical structure and bond-order discrimination. [27][28][29][30] Moreover, qPlus-AFM can also be used for direct imaging of the intermolecular hydrogen bond. 31 For example, Zhang et al. reported in 2013 a real-space image of the hydrogen bond formed in 8-hydroxyquinoline clusters using noncontact atomic force microscopy (nc-AFM) with CO-tip, although the origin of the hydrogen bond remains controversial. 32 It is worth noting that functional tip also greatly improves the resolution of the STM image. In recent years, several reports have shown that STM with CO-tip is also able to unravel the internal atomic structure of a single molecule. [33][34][35] The intermolecular force of triangular halogen bonds has also recently been studied by nc-AFM and STM, 36 and the bright lines appear at positions where the triangular halogen bond form at STM images. 13 The bright lines (or image contrast) are thought to originate from the interacting electrons of intermolecular bonds, while others propose that bright lines are caused by the bending of the CO molecule. 37,38 Currently, the origin of the bright line is still controversial, and understanding the origin of the bright line (image contrast) is not only conducive to the understanding of non-covalent bonds, but also of great significance to the improvement of this surface characterization technique. Furthermore, there are many other forms of halogen bonds other than the reported triangular halogen bond, and more details of halogen bonds need to be carefully studied by ultrahigh-resolution STM and nc-AFM. 36 In this work, we present the diversity of halogen bonds, including tetragonal Br-Br bonds, Br-N bond and Br-H bond, and directly image them using nc-AFM and BR-STM. A nitrogen-doped and bromine-terminated 2-TBQP molecule was synthesized and deposited on Au(111) substrate to obtain the self-assembled molecular clusters. At low molecular coverage, two distinct forms of 2-TBQP dimers are observed. Nc-AFM images the real-space atomic structure of the 2-TBQP molecule. Moreover, two different types of Br-N bonds (namely type-1 and type-2 Br-N halogen bonds) are observed in the two forms of 2-TBQP dimers, respectively, verified by the presence of bright lines between two corresponding atoms (Br-H and Br-N atoms). Density functional theory (DFT) calculations shows that there is obvious electron density redistribution between the adjacent Br-N and Br-H atoms, suggesting that the halogen bonds are essentially electrostatic interactions.</p><p>At high molecular coverage, we observe the formation of a 2-TBQP tetramer. Both nc-AFM and BR-STM directly image the formation of tetragonal halogen bonds on top of the 2-TBQP tetramer, appearing as tetragonal bright lines that link four Br-atoms. By contrast, no bright line is observed between two adjacent but no-bonding Br-atoms even though the tip-sample distances are close.</p><!><p>Atomic structure and electrostatic potential of a single 2-TBQP molecule. Figure 1 displays atomic structure and electrostatic potential of 2, 7, 13, 18-tetrabromodibenzo[a,c]dibenzo [5,6:7,8]-quinoxalino[2,3-i]-phenazine (2-TBQP). The 2-TBQP molecule was synthesized following previous work. 21,39 Figure 1a shows its atomic structure obtained from single crystal XRD. 2-TBQP molecule is a planar π-conjugated polycyclic hydrocarbon with four doping nitrogen atoms (red atoms) and four bromine atom terminals (blue atoms). 2-TBQP molecules were deposited on a clean Au(111) crystal substrate held at room temperature, and the as-prepared samples were then transferred in-situ into LT-STM/nc-AFM chamber for imaging at 4.2 K. Figure 1b presents an atomic resolution nc-AFM image of a single 2-TBQP molecule on Au(111) acquired with a CO-modified tip. The nc-AFM molecular image structure is experimentally reproducible. Nine benzene rings are observed with four bright spots at its terminals, corresponding to the four Br-atoms. We infer that the Br-atoms exhibit higher frequency shift (Δf) due to its larger atom size than the other element atoms, indicating stronger interaction of the Bratom to CO-tip at a given tip-sample distance. In contrast, the N-atom sites have low Δf intensity and the N-containing benzene ring becomes larger compared to atomic structure shown in Figure 1a. The distinct Δf contrasts in the nc-AFM image may be due to the variable van der Waals (vdW) interactions between CO-tip and element atoms. sites display a negative electrostatic potential, while H-atom sites have a positive electrostatic potential. Both negative and positive electrostatic potential are observed within a Br-atom due to the anisotropic distribution of halogen atom electron density. This electron anisotropic distribution leads to the coexistence of electron depleted δ-hole cap at the terminal of the C-Br bond (light blue) and encircling electron-rich belt perpendicular to the C-Br bond (yellow), in line with previous reports. 13,15 The formation of two distinct types of 2-TBQP dimers. Figure 2 displays two distinct forms of 2-TBQP dimers stabilized via two types of Br-N halogen bonds. Figure 2a shows an atomic resolution nc-AFM image of type-1 dimer. Two Br-atoms are observed to be adjacent to two Natoms, respectively. An enlarged nc-AFM image in Figure 2b shows that one of the Br-atoms is adjacent to a N-atom of another 2-TBQP molecule. A bright line (labeled by a red arrow) between the Br-N atoms appears, suggesting the formation of Br-N bond (named as type-1 Br-N halogen bond). Three bright lines (labeled by blue arrows) are also observed at the vicinity of the Br-N bright line, which connects three series of Br-H atoms, respectively. We note that the intensity of Br-N bright line is somewhat lower than that of the Br-H bright lines. Figure 2c Figure 2d shows a nc-AFM image of type-2 dimer. Like type-1 dimer, both Br-N and Br-H halogen bonds form in type-2 dimer although their molecular packing structure is different. A realspace atomic-resolution image in Figure 2e shows that two Br-atoms are adjacent to two N-atoms, respectively. Two bright lines (labeled by two red arrows) are observed to connect two couples of Br-N atoms, suggesting the formation of two Br-N bonds (named as type-2 Br-N halogen bond).</p><p>In addition, another two bright lines (labeled by blue arrows) appear at the sides of each Br-N bright line, which connect two couples of Br-H atoms, respectively. Figure 2f presents a DFToptimized molecular structure of type-2 dimer with red and black dotted-lines that connect Br-N atoms and Br-H atoms, respectively. According to the electrostatic potential of a 2-TBQP molecule in Figure 1c, we believe that the formation of a Br-N halogen bond results from the mutual interaction between an electron pair in N-atom and an electron depleted δ-hole at the terminal of the C-Br, and the formation of Br-H halogen bond arises from the interaction between the electron depleted H-atom and an encircling electron-rich belt at the Br-atom. We infer that the nc-AFMobserved bright lines origin from a synergistic effect of the exchange electron density of halogen bonds and the CO-tip bending, as further discussed in Figure 5.</p><p>To obtain more insight regarding the two types dimers, DFT calculations were carried out to determine the optimized molecular packing structure and their electron density difference. Figure 3a and 3b plot the calculated intermolecular interaction energy of type-1 and type-2 dimers as a function of Br-N distance, respectively. As expected, the interaction energy per molecule first decreases and then increases as the Br-N distances increase for both type-1 and type-2 dimers. The Br-N bond distances at interaction energy minimization are 3.54 Å for type-1 dimer and 3.58 Å for type-2 dimer, respectively, which match well with the experimental measurements (3.45 ± 0.1 Å, labeled by red pentacles). DFT calculations were also carried out to obtain the electron density difference (Δρ), which provides the intermolecular bonding features of Br-N and Br-H halogen bonds. Figure 3c and 3d show the charge density difference projected onto type-1 and type-2 dimers, respectively. Obvious electron density accumulation is observed at the connection of both Br-N atoms and Br-H atoms, suggesting the charge transfer between Br-atoms and N-atoms/H-atoms. Similar results of the charge transfer between Br-O atoms were also reported by Giovanni et al. 15 The formation of Br-N and Br-H halogen bonds is primarily electrostatic. The calculated results suggests that charge transfer between Br-N and Br-H atoms also play an important role, in line with IUPAC recommendations and literature reports. 3,15,41 We infer that the electron density between Br-N and Br-H atoms may also contribute to the bright lines feature observed at nc-AFM images in Figure 2b and 2e, and the higher intensity of electron density between Br-H atoms than that between Br-N atoms results in a brighter line of Br-H bond than that of Br-N bond shown at nc-AFM images.</p><p>More details are discussed in Figure 5.</p><p>A significant difference is found in the proportion yield of the two types of 2-TBQP dimers, with a higher yield of type-1 than that of type-2. Statistically, the yield of type-1dimer is greater than 90%, while the yield of type-2 dimer is less than 10%. To explain the observed results, we summarize in Figure 3e the characters of Br-N halogen bonds, which include formation energy, the numbers of halogen bonds, bond distance and bond angle. Type-1 dimer has lower formation/interaction energy than type-2 dimer, and Figure 3e shows that the interaction energy are -0.172 eV and -0.133 eV for type-1 and type-2 dimer, respectively. The distinct interaction energy result from a difference in the bonding number of halogen bonds formed between type-1 and type-2 dimer. As show at the third row in Figure 3e, two Br-N intermolecular bonds form for both type-1 and -2 dimers. However, the fourth row in Figure 3e shows that type-1 dimer (6 Br-H bonds) has more formed Br-H halogen bonds than that of type-2 dimer (4 Br-H bonds). We infer that the two additional Br-H bonds lead to a lower formation energy and thus a higher yield of type-1 dimer. Another distinct feature between the two dimers is the orientation of molecule with the Au[110] direction of Au substrate. Figure S2 shows that the relative orientations of molecule with respect to the Au[110] direction are 30 degree and 0 degree for type-1 and type-2 dimer, respectively. The calculated total absorption energy with substrate is -4035.95 eV and -4035.88 eV for the 30 degree case and the 0 degree case, respectively. The total energy difference is only -0.07 eV, and such a small energy difference suggests that the Au substrate does not play an important role in the formation of the dimers.</p><p>The bonding angle of the Br-N bond is also acquired. Unexpectedly, nc-AFM images in Figure 2b and 2e show that the Br atom does not align with two N-atoms of a pyrazine ring. The bond angle of the Br-N bond, measured along N-Br-C, is 166 ± 2 o and 165 ± 2 o for type-1 and type-2 dimers, respectively, which are close to but different from a theoretical bonding angle of 180 o . Similar bending halogen bond was reported by Ebeling et al. 36 We infer that the steric hindrance of two adjacent H-atoms hinders the alignment of Br-atom with two N-atoms of a pyrazine ring.</p><p>The Br-atom optimises its position to enable the formation of both Br-N and Br-H halogen bonds, suggesting the flexible bonding angle of non-covalent halogen bonds, unlike rigid covalent bonds.</p><p>We propose that the flexibility of non-covalent intermolecular bonds is one of the reasons why supramolecular structures form over a larger ordered area than the covalently linked network, besides the reversibility of the non-covalent bonds. 42 Formation of 2-TBQP tetramer. Figure 4 displays another self-assembled structure of 2-TBQP tetramer stabilized by tetragonal Br-Br halogen bonds with assistance of Br-atoms absorption. The 2-TBQP tetramer are observed at a higher molecular coverage of 0.2 ML in Figure S3. Figure 4a gives a STM image of the supramolecular structure of 2-TBQP tetramer on Au(111).</p><p>The STM image displays the H-shaped molecular skeleton of a single 2-TBQP molecule.</p><p>Individual Br-atoms (labeled by blue arrows) adsorb on the periphery of 2-TBQP molecules. The individual Br-atoms come from the partial Br-atom cleavage from 2-TBQP molecules at room temperature, and the evidence is presented in Figure S4. The formation of self-assembled 2-TBQP tetramer is observed, and four Br-atoms (labeled by blue quadrate) from four 2-TBQP molecules are adjacent and form a square geometric structure. We infer that the formation of tetragonal Br-Br halogen bonds governs the self-assembled 2-TBQP tetramer. To verify our hypothesis, nc-AFM measurements were carried out. Figure 4b Obvious electron density exchange is observed between both Br-Br atoms and Br-H atoms, suggesting the charge transfer occurs across the Br-Br atoms and the Br-H atoms, similar to the case of Br-N bond. We infer that the electron density can be probed by nc-AFM and contribute to the appearance of bright lines between Br-Br atoms and Br-H atoms.</p><p>The origin of bright line feature of halogen bonds. We note that similar bright line feature of intermolecular hydrogen bond was also imaged by nc-AFM with CO-tip by Xiaohui et al. 32 The appearance of bright line feature of hydrogen bond was explained to be contributed by the increased electron density at the hydrogen bond position, which enhances the Pauli repulsion between CO-tip and the hydrogen bond. However, Habala et al. found in their simulated calculations that a bright line feature of the intermolecular hydrogen bond could also be caused by the CO-tip bending. 38 Since no intermolecular electron density is involved in their simulated calculations, they claimed that CO-tip bending plays a major role in nc-AFM imaging of intermolecular hydrogen bond. Pavlicek et al. observed an apparent bond ridge (bright line) between two sulfur atoms of a DBTH molecule where no chemical bond forms. 43 Liljeroth et al. also observed experimentally a bright line in nc-AFM images between two N atoms that did not form an intermolecular chemical bond. 37 Therefore, their experiments suggest that CO-tip bending is a major factor in nc-AFM images. However, whether CO-tip bending or electron density plays a key role in imaging of intermolecular bond has not fully been determined.</p><p>We obtained both nc-AFM and BR-STM images of two adjacent Br-atoms with intermolecular bonding and the absence of chemical bonding, respectively. Figure 5 displays constant-height nc-AFM images (top row) and BR-STM images (bottom row) as the function of tip-sample heights (ΔZ). One can note that the Br-Br distances for the two cases are very close, and their measured Br-Br distances are 370 ± 10 pm and 345 ± 10 pm for intermolecular bonding case and nonbonding case, respectively. At ΔZ= + 0.2 Å, no bright line feature is observed for the two cases in both nc-AFM and BR-STM images. As the CO-tip approaches to the sample, a bright line feature in both nc-AFM and BR-STM images become more prominent for the intermolecular bonding case, while no bright line feature appears for the no bonding case. For example, at ΔZ < -0.2 Å, an obvious bright line feature is observed between two Br-atoms where halogen bond forms (the intermolecular bonding case) in both nc-AFM and BR-STM images. In contrast, no bright line appears between two adjacent but individual Br-atoms where no bond forms (no bonding case) even at close tip-sample distance. These experimental observations verify that the formation of intermolecular halogen bond contribute to the appearance of bright line features, suggesting that the electron density play an important role in nc-AFM and BR-STM image contrasts. Sweetman et al have also reported that the bright line originated from the electron repulsion that occurs within the tip-sample junction based on a comparison of experimental images and force spectra. Besides the contribution of electron density, we believe that the CO-tip bending also contributes to the bright line features based on the following two experimental observations. The first observation is that a gradual frequency shift peak also appears between the two non-bonding Br-atoms at close tip-sample distance in Figure S6, although no corresponding bright line is observed and its frequency shift intensity is much lower than that of two intermolecular bonding Br-atoms. The second observation is that the intensity of the bright lines is not proportional to the electron density (compared Br-Br bond with Br-H bond in Figure 4c and 4f), suggesting that not only electron density contributes to the bright line feature. At close tip-sample distance, CO-tip bending often occurs, evidenced by a disturbed line appearance and the observation of apparent sample drift in Figure S6. Considering of both our experimental observation and the other reported works by Pavlicek and Liljeroth et al, 37,43 we believe that both CO-tip bending and electron density play key roles in imaging of intermolecular bond.</p><!><p>In conclusion, we demonstrated the physical origin and the diversity of halogen bonds in selfassembled 2-TBQP molecules on Au(111). The halogen bond was directly imaged as bright line features by nc-AFM/BR-STM, and the bright line is contributed by a synergistic effect of CO-tip bending and electron density exchange, as evidenced by the experimental observation that bright lines are observed at two intermolecular bonding Br-atoms, whereas two adjacent but non-bonding Br-atoms do not display such lines. We observed three types of halogen bonds, namely the Br-H halogen bond, Br-Br halogen bond and Br-N halogen bond, suggesting the diverse class of halogen bonds results from its bonding environment; halogen atoms can interact with manifold element atoms of positive (H-atom), neutral (Br-atom) and negative (N-atom) potential. In addition, a bending Br-N halogen bond was observed with a bond angle of ~ 165 o rather than theoretical bond angle of 180 o , implies the flexibility of halogen bonds. This work demonstrates the versatility of combined nc-AFM/BR-STM in understanding halogen bonding, and further exploration of other intermolecular bonds using this technique is needed. (2-TBQP). The 2-TBTBP molecule was synthesized according to the literature. 39 To confirm its structure, 2-TBQP powder was sublimed to obtain single crystals for single-crystal XRD analysis. CCDC 1455701 (2-TBQP) contains the supplementary crystallographic data. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre.</p><!><p>Experiments were carried out under UHV conditions (10 -9 mbar). The metal single crystals were cleaned via repeated cycles of Argon sputtering and subsequent annealing to 800 K. 2-TBQP precursors were thermally deposited from a Knudsen effusion cell onto metal substrates held at room temperature. After molecular deposition, the sample was in-situ transferred to STM/AFM chamber for imaging at 4.2 K.</p><p>Characterization. BR-STM/nc-AFM measurements were carried out an integrated scanning probe system consists of ScientaOmicron low-temperature scanning tunneling microscopy (LT-STM) combined with non-contact atomic force microscopy (nc-AFM). The BR-STM images were recorded in both constant current/height mode using AFM tip with functional CO molecule, and bias voltages were applied to the sample. For nc-AFM images, the constant-height mode with AFM CO-tip was used to record the frequency shift (∆f) of the qPlus resonator (sensor frequecy f0 ≈ 27000 Hz, Q ≈ 25000). All the measurements were performed at 4.2 K under a base pressure better than 10 -11 mbar.</p><p>Computational details. All calculations were performed by using Perdew-Burke-Ernzerhof (PBE) 44 electronic interaction as implemented in Vienna ab initio package (VASP) [2]. 45,46 The projector augmented wave (PAW) 47 method was used with cut-off energy of 500 eV. A Gammapoint-only k-mesh was used for sampling Brillouin zone. All structures were fully relaxed until forces and energy differences were less than 0.05 eV/Å and 10 -5 eV, respectively. The thickness of vacuum was larger than 12 Å to minimize interactions between periodic images. Van der Waals interactions were introduced by DFT-D2 method. 48 The electrostatic potential was calculated by using Gaussian software within the framework of DFT.</p>

ChemRxiv

Direct LC-MS/MS Detection of Guanine Oxidations in Exon 7 of the p53 Tumor Suppressor Gene

Oxidation of DNA by reactive oxygen species (ROS) yields 8-oxo-7,8-dihydroguanosine (8-oxodG) as primary oxidation product, which can lead to downstream G to T transversion mutations. DNA mutations are nonrandom, and mutations at specific codons are associated with specific cancers, as widely documented for the p53 tumor suppressor gene. Here, we present the first direct LC-MS/MS study (without isotopic labeling or hydrolysis) of primary oxidation sites of p53 exon 7. We oxidized a 32 base pair (bp) double-stranded (ds) oligonucleotide representing exon 7 of the p53 gene. Oxidized oligonucleotides were cut by a restriction endonuclease to provide small strands and enable positions and amounts of 8-oxodG to be determined directly by LC-MS/MS. Oxidation sites on the oligonucleotide generated by two oxidants, catechol/Cu2+/NADPH and Fenton\xe2\x80\x99s reagent, were located and compared. Guanines in codons 243, 244, 245, and 248 were most frequently oxidized by catechol/Cu2+/NADPH with relative oxidation of 5.6, 7.2, 2.6, and 10.7%, respectively. Fenton\xe2\x80\x99s reagent oxidations were more specific for guanines in codons 243 (20.3%) and 248 (10.4%). Modeling of docking of oxidizing species on the ds-oligonucleotide were consistent with the experimental codon oxidation sites. Significantly, codons 244 and 248 are mutational \xe2\x80\x9chotspots\xe2\x80\x9d in nonsmall cell and small cell lung cancers, supporting a possible role of oxidation in p53 mutations leading to lung cancer.

direct_lc-ms/ms_detection_of_guanine_oxidations_in_exon_7_of_the_p53_tumor_suppressor_gene

<!>Chemicals and Reagents<!>Oligonucleotide Oxidation and Product Preparation<!>Removal of Enzymes and Salts<!>UPLC-QTOF MS/MS Analysis<!>Molecular Modeling<!>Method Development<!>Oxidation by Catechol/Cu2+/NADPH<!>Oxidation by Fenton\xe2\x80\x99s Reagent<!>Modeling of Oxidant Binding<!>DISCUSSION<!>CONCLUSION

<p>Reactive oxygen species (ROS) such as hydroxyl radical (•OH), singlet oxygen (1O2), peroxynitrite (ONOO–), and superoxide (O2•−) play important roles in living organisms, e.g. as messengers in cell signaling cascades. However, they can also induce cell apoptosis during oxidative stress,1 featuring unregulated accumulation of ROS from endogenous or exogenous processes. Oxidative stress likely play roles in aging, cardiovascular disease, Parkinson's disease, and cancer.2–6</p><p>One source of ROS is oxidation of transition metals by H2O2 to produce •OH, which is usually called Fenton's reaction, especially when involving ferrous ions.7–9 Another source of ROS is redox cycling of quinones and other metabolites involving Cu2+ and NADPH.10–14 When chemicals or their metabolites lead to DNA damage, they are referred to as genotoxic. Catechol, a model for quinoid metabolites, undergoes redox cycling in the presence of Cu2+ and NADPH that leads to a two-electron oxidation to o-quinone, which subsequently undergoes two one-electron reductions back to catechol by NADPH. ROS, including H2O2, O2•−, and •OH, are formed in these redox cycles (Scheme 1) and can oxidize DNA in vitro.15,16 Also, polycyclic aromatic hydrocarbon (PAH) o-quinones are implicated as causes of mutations in tumor suppressor genes and are known risk factors for lung cancer.17</p><p>Guanine is the main oxidation target in genes due to having the lowest redox potential of the DNA nucleobases (1.3 V vs NHE).18,19 The primary oxidation product of deoxyguanosine is 8-oxo-7,8-dihydrodeoxyguanosine (8-oxodG), which can lead to G to T transversion mutations during DNA replication.20,21 Cancer may result when such mutations occur at tumor suppressor genes or oncogenes.22,23</p><p>Tumor suppressor genes provide cancer protection by coding for proteins that inhibit cell proliferation and tumor development.24 TP53 (or p53) was identified as a tumor suppressor gene in the 1980s.25 It encodes p53 protein that regulates the cell cycle and inhibits tumor growth. Mutations in p53 genes are found in >50% of all human cancers.26–28 Extensive databases document mutations found at specific p53 codons from human tumors and cancer cell lines and show that many mutations are well-correlated with specific cancers. Most p53 mutations occur at exons 5–8.29 For example, highly mutated codons 157, 158, 248, and 249 appear in lung cancers; mutated codons 175, 248, and 273 are found in breast cancer; and mutated codons 175, 248, and 282 occur in liver cancer.30,31 Thus, DNA oxidation at specific sites that lead to mutations at the same sites may be possible to correlate with specific cancers.</p><p>Early work in p53 gene oxidation employed gel electrophoresis of DNA fragments. The structure of guanine oxidation products was not addressed, and most studies focused on the resulting mutations. Both guanine residues (AGG → ATG, AGG → AGT) in codon 249 mutated to thymine as a result of Fenton oxidation in human fibroblasts.32 Oxidation of human skin fibroblasts by UVB light (280–320 nm) caused G to T mutations at the third position of codon 249 (AGG → AGT).33 PAH-induced oxidation of p53 has been observed at hot spots associated with lung cancer, including codons 245 (GGC → TGC), 249 (AGG → ATG), and 273 (CGT → CTT).17</p><p>Mass spectrometry (MS) is a powerful tool to detect structurally damaged DNA.34 LC-MS/MS approaches have been developed to size and sequence DNA oligonucleotides up to 20 bp.35,36 QTOF-MS/MS was employed to detect sequence and agent specificity on the further oxidation products of 8-oxodG in DNA oligonucleotides.37,38 In that study, synthesized 8-oxodG containing oligonucleotides were oxidized to detect the 8-oxodG derived oxidation products, and standard oligonucleotides were not used. Stable isotope labeling of DNA has been widely used with mass spectrometry to examine the reactive sites and cytosine methylation on the formation of 8-oxodG and downstream oxidation products of reactions with nitrosoperoxycarbonate and riboflavin-mediated photolysis.39 This approach detects the amount of 8-oxodG and other products formed, but isotopic labeling syntheses and enzymatic oligonucleotide digestion are required.</p><p>To investigate reactive sites, the positional isomeric modified oligonucleotides must be separated. Harsch et al. reported separation of isomeric benzo[c]phenanthrene diol epoxide adducted HRPT gene sequence by using ammonium acetate as ion-pairing reagent.40 Xiong et al. identified nine isomeric (±)-anti-7r,8t-dihydroxy-9t,10-epoxy-7,8,9,10-tetrahydro-benzo[a]pyrene adducted oligonucleotides using ion pair reagent triethylammonium bicarbonate (TEAB).41 Sharma et al. used TEAB to separate oligonucleotides adducted by N-acetylaminofluorene, N-hydroxy-4-aminobiphenyl, and (±)-anti-benzo[a]pyrene diol epoxide to investigate the site selectivity.42,43 These papers provide guidance for HPLC separations of isomeric oligonucleotide products.</p><p>We previously described a restriction enzyme-assisted LC-QTOF MS/MS oligonucleotide sequencing methodology to detect covalent adduction on ds-oligonucleotides longer than 20 base pair (bp).45 We used this approach to evaluate reaction products of oligonucleotides with the diol epoxide metabolite of benzopyrene (B[a]P) and examined the effect of cytosine methylation on adduct reaction kinetics of exon 7 of the p53 gene.46 In the present paper, we tailor this approach to directly detect 8-oxodG formation in oxidized ds-DNA strands. We apply the method to the 32 bp exon 7 ds-oligonucleotide fragment of the p53 gene, representing p53 codons 242–253 (Scheme 2). Ion pairing reagent TEAB was used to separate positional isomeric oxidative oligonucleotides. The primary products of exon 7 oligonucleotide from catechol/Cu2+/NADPH and Fenton's reagent oxidation were compared. Analysis of >20 bp oligonucleotides is facilitated by a restriction enzyme that cuts the oligonucleotide into smaller fragments suitable for MS/MS sequencing. Using this approach, Fenton's reagent oxidized G at codons 243 and 248; the catechol system oxidized guanines at codons 243, 244, 245, and 248, correlating with lung cancer mutation hot spots at codons 245 and 248. The reacting guanines in codons 245 and 248 are C-phosphate-guanine (CpG) sites, and the guanine in codon 248 is also the most reactive toward SN2 reactions with B[a]P diol epoxide.46 The p53 oxidation sites have been uncovered in a purely chemical reaction, and this is of course much less complex than in vivo in humans. However, we find that the most chemically reactive p53 codons present correlations to mutation sites found in tumors.</p><!><p>Custom oligonucleotides were from Sigma-Aldrich. Sources of chemicals and method details are in the Supporting Information file.</p><!><p>Double stranded 32 bp exon 7 fragment of the p53 gene was oxidized in aqueous solution by catechol/Cu2+/NADPH or Fenton's reagent (see Supporting Information). Conditions were optimized to limit G oxidations to 8-oxodG without strand breaks and to yield LC-MS/MS results of sufficient intensity for oxidation site location and quantitation. Briefly, 100 μg of oligonucleotide was incubated in 200 μL of Tris buffer with 1 mM catechol,15 50 μM CuCl2, and NADPH regeneration system (10 mM G6P, one unit G6PDH enzyme, 0.8 mM NADP+, 1 mM MgCl2) at 37 °C for 12 h. Alternatively, 100 μg of oligonucleotide was incubated with Fenton's reagent containing 0.1 mM FeSO4 and 40 mM H2O216 at 37 °C for 4 h with constant stirring. Amicon Ultra-0.5 3K centrifugal filters from MilliporeSigma were used to remove excess catechol, NADPH system, and FeSO4 products. The reaction mixture was put into the vial and centrifuged at 13 000 rpm for 30 min. Ds-oligonucleotide fragments were retained on the filter, which was then reversed, put into a new vial, and centrifuged at 13 000 rpm for 30 min to collect the approximately 50 μL of ds-oligonucleotide. Then, 15 μL (150 units) of restriction enzyme Nla III, 20 μL of 10 × NE buffer (New England Biolabs, 1× buffer contains 50 mM potassium acetate, 20 mM tris-acetate, 10 mM magnesium acetate, 100 ug/mL BSA, pH 7.9) and 115 μL of pure water were added to the reacted oligonucleotide solution, and this solution was incubated 37 °C for 12 h to cut oligonucleotides into smaller fragments.</p><!><p>Two hundred microliters of phenol:chloroform:isoamyl alcohol (25:24:1) was added to the mixture from the above process. The mixture was shaken for 15 min and then centrifuged at room temperature for 10 min. Then, the upper aqueous solution was carefully transferred to a fresh tube for subsequent extraction. The solution was extracted with phenol:chloroform:isoamyl alcohol (25:24:1) three times and chloroform:isoamyl alcohol (24:1) twice. The resulting aqueous oligonucleotide solution was about 200 μL.</p><p>Ethanol was added to the oligonucleotide solution to precipitate oligonucleotide as a small pellet (full details in Supporting Information). The supernatant was carefully removed, and then 1 mL of 70% cold ethanol in water was used to wash the pellet, which was then dried under N2 and dissolved in 100 μL of HPLC-grade water. This solution was heated at 90 °C for 15 min and cooled rapidly to convert ds- to ss-oligonucleotides and then stored at −20 °C.</p><!><p>A Dionex Ultimate 3000 UPLC with a Gemini C-18 column (150 × 0.5 mm, 3 μm particle size) was used with mobile phases 25 mM triethylammonium bicarbonate (TEAB, solvent A) and methanol (solvent B).42 TEAB facilitated positional isomer separation of oligonucleotides. A gradient of 15% B for 2 min followed by increasing from 15% B to 20% B over 36 min, then back to 15% B for another 2 min at flow rate 8 μL/min was used. The UHPLC was interfaced to an AB Sciex QSTAR Elite mass spectrometer in negative ion mode with a −4500 V ion spray voltage, −60 V declustering potential at 300 °C. m-Nitrobenzyl alcohol (0.1%) was infused at 4 μL/min to mix with the LC flow postcolumn using a three way connector before entering the QSTAR to enhance oligonucleotide charging and signal intensity.47 The ss-oligonucleotides were analyzed by time-of-flight scan mode for sizing unoxidized and oxidized fragments. Product ion scan mode was used for sequencing oxidized fragments at −40 eV collision energy. In most cases, analysis was restricted to singly oxidized oligonucleotides.</p><!><p>The standard B-DNA form of 32 bp p53 exon 7 ds-oligonucleotide was modeled using web 3DNA48 and solvated using CHIMERA software.49 An Amber solvation model was used with appropriate box size to accommodate water molecules.46 Autodock 4.2.6 was used for docking.50,51 Ligands (Fe2+, Fe3+, H2O2, •OH, catechol, benzoquinone, and Cu(I)OOH) were imported into the model software. A Lamarckian genetic algorithm (LGA) was used in Autodock 4.2.6 to find the binding energy between the oligonucleotides and the ligands.</p><!><p>The 32 bp exon 7 p53 fragment was oxidized with catechol or Fenton's reagent, then restriction enzymes were used to cut the oxidized ds-oligonucleotide to smaller size suitable for LC-MS/MS. Restriction enzyme Nla III cuts both strands of ds-DNA after the sequence CATG. In the p53 exon 7 oligonucleotide, this results in two ds fragments of 13 and 19 bases, leaving unpaired CATG sequences at the ends (Scheme 2). The end 5′-CATG is not cut by Nla III as there is no 3′-complement to be cut. Subsequently, heat was used to convert all ds-oligonucleotides to single strands.</p><p>Our focus was to identify the primary oxidation sites on exon 7, so we developed conditions to determine 8-oxodG sites in singly oxidized fragments. Our reaction conditions yielded 8-oxodG as the only oxidation product. If a single dG is oxidized to 8-oxodG, the mass of the reacted ss-oligonucleotide increases by 15.999. The identity of each 8-oxodG-containing oligonucleotide was determined by comparing the measured m/z of the oxidized oligonucleotide to the expected m/z of the corresponding unreacted species calculated using Mongo Oligo Mass Calculator, v2.08.52 m/z values for the four unoxidized and oxidized oligonucleotide fragments are shown in Table S1. For example, m/z 1002.4 was observed for unoxidized fragment 1 (Scheme 2) with z = −4, and the 8-oxodG-containing fragment 1 had m/z = 1006.4 (z = −4). TOF MS spectra of the four oxidized fragments are in Figure S1.</p><p>Collision-induced dissociation (CID) provided information on the sequence of oxidized oligonucleotide products by fragmentation of the phosphodiester backbone (Scheme 3), forming an–bn and wn ions.53,54 The position of 8-oxodG was determined by detecting the difference in m/z values of the an–bn and wn ions of corresponding unoxidized and oxidized oligonucleotides.</p><!><p>Catechol is an industrial chemical and a major component of tobacco smoke55 and is classified as a group 2B carcinogen.56,57 In the presence of Cu2+ and an enzymatic NADPH regenerating system, redox cycling of catechol, semiquinone radical, and benzoquinone produces ROS, which oxidize dG to 8-oxodG in DNA (Scheme 1).15,16,58</p><p>Figure 1A shows an extracted ion chromatogram (XIC) of a selected ion of oxidized exon 7 fragment 1 (Scheme 2) with m/z 1006.4 representing 8-oxodG-containing fragment 1 with z = −4 (the most intense ion was used to present XIC). There are four major peaks in the chromatogram, suggesting that 4 positional isomers for singly oxidized fragment 1 were formed with retention times 13.6, 14.9, 16.2, and 17.9 min. The CID spectrum of fragment ion 1006.4 for peak 1 is shown in Figure 1B and for peak 2 in Figure 1C (total ion chromatogram (TIC) in Figure S2).</p><p>Differences in a–b and w ions between unoxidized and oxidized fragment 1 were used to locate the oxidation sites (i.e., 8-oxodG), as used previously for metabolite adduction sites.35,36,45,59 Table 1 summaries a–b and w ions for unoxidized fragment 1 and singly oxidized fragment 1 for the two peaks in Figures 1B and C. Red numbers indicate the ions increased in mass by 15.999 by oxidation of dG to 8-oxodG in the unoxidized fragment. The MS/MS spectrum for peak 1 of singly oxidized fragment 1 (Figure 1B) shows increase in mass of ions from a6–b6 to a8–b8 compared to unoxidized fragment 1. This indicates that the fifth guanine was oxidized to 8-oxodG, CATGGoxGCGGCATG (Gox = 8-oxodG), which was confirmed by the increase in m/z from w9 among the w ions (Table 1).</p><p>The MS/MS spectrum for peak 2 of singly oxidized fragment 1 (Figure 1C) shows an increase in mass of all ions from a9–b9 compared with that of the unoxidized fragment 1. This shows that the eighth guanine was oxidized to 8-oxodG, CATGGGC-GoxGCATG, and is confirmed by increases in mass of w ions from w6 ion. Similar analysis of the third and the fourth peaks revealed that peak 3 indicates oxidation of the sixth guanine (CATGGGoxCGGCATG) and peak 4 represents oxidation of the fourth guanine in CATGoxGGCGGCATG) (Figure S3, Table S2).</p><p>Singly oxidized fragment 2 gave m/z of 971.7 at z = −6 (Figure S1B). The XIC of m/z 971.7 (Figure 2A) showed only one major peak, which indicated that only one base on fragment 2 was oxidized during the reaction. MS/MS spectra for singly oxidized fragment 2, m/z 971.7 (Figure 2B), shows increase in mass of all ions from a6–b6 to a7–b7, suggesting that the fifth guanine was oxidized, AACCGoxGAGGCCCATCCTCA. This is confirmed by increase in mass of w15 and w16 for w ions. All ions below w14 gave the same m/z values as the unoxidized fragment (Table S3).</p><p>Two unknown peaks were found in oxidized fragments 1 (20.3 min) and 2 (28.7 min) after catechol/Cu2+/NADPH oxidation (Figures 1A and 2A). They are also positional isomers of oxidized fragments 1 and 2 but were not intense enough for reliable identification.</p><p>Similar analyses were done on oxidized complementary strands of fragments 3 and 4. Results identified oxidized fragment 3 with m/z 917.1, z = −3 (Figure S1C) and oxidized fragment 4 with m/z 1186.4 (Figure. S1D), z = −6 (Table S1). MS/MS spectra are shown in the Supporting Information (Figures S4 and S5 and Tables S4 and S5). In fragment 3, the third and the last guanines were oxidized, and the fourth guanine in fragment 4 was oxidized. Oxidation sites for all the fragments are summarized in Table 2.</p><p>We estimated the relative abundance of specific oxidized fragments by comparing areas under extracted ion chromatograms of oxidized fragments to those of unoxidized fragments (Figure S6), assuming that oxidized and unoxidized fragments have similar MS ionization efficiencies (details in Supporting Information).45 The relative amounts of oxidation in the 32 bp fragment were estimated (Table 2). Ratio of relative amounts of oxidation for codons 248/243 was 1.9; codon 248/244 was 1.3, and 248/245 was 4.1.</p><!><p>Deoxyguanosine in DNA is easily oxidized to 8-oxodG by Fenton's reagent. The active oxidant in Fenton's reagent is thought to be •OH,60,61 and oxidation products of 8-oxodG have also been identified.9,62,63 We found none of these overoxidation products under our conditions.</p><p>Oxidation products of p53 exon 7 by Fenton's reagent revealed several differences from oxidation by catechol/Cu2+/NADPH. Figure 3A shows there is only one major peak in the XIC of singly oxidized fragment 1, indicating only one base was oxidized. The MS/MS spectrum for singly oxidized fragment 1 (Figure 3B) shows an increase in mass of all ions from a5–b5 to a7–b7 compared to the unoxidized fragment. This indicates that the fourth guanine was oxidized to 8-oxodG, CAT-GoxGGCGGCATG, confirmed by increased mass of w ions above w10. All ions from w2 to w9 had the same m/z as that of the unoxidized fragment (Table S6).</p><p>Oxidized fragments 2 and 4 gave results similar to those of catechol/Cu2+/NADPH oxidation (Figures S7A and C). For fragment 3 (Figure S7B), there is only one major peak for the oxidized fragment 3. Results are summarized in Table 2.</p><p>The relative amounts of oxidation by Fenton's reagent were also estimated (Table 2) by comparing areas under extracted ion chromatograms. Ratio of relative amounts of oxidation for codons 248/243 was 0.5.</p><!><p>In DNA oxidation mediated by catechol, NADPH and Cu2+, molecular oxygen is reduced to O2•− and H2O2, and catechol reacts with Cu2+ yielding Cu+ and benzoquinone (Scheme 1). Also, Cu+ reacts with H2O2 to form •OH radical and Cu(I)-hydroperoxyl complex (Cu(I)OOH), which may lead to DNA strand breaks. In Fenton's reagent oxidation, Fe2+ reacts with H2O2 to form Fe3+ ions and •OH radical. To gain insight into the codon specificity of the oxidations, molecular modeling was done with AutoDock 4.2.6 by docking each possible involved species (Fe2+, Fe3+, H2O2, •OH, catechol, benzoquinone, and Cu(I)-OOH) with the B-DNA form of the exon 7 ds-oligonucleotide. Parameters were set with 25 000 000 evaluations and 50 docked conformations for an individual docking computation. Examples of preferred Fe2+ binding site on exon 7 from this analysis is shown in Figure 4, and preferred catechol sites are shown in Figure 5. The preferred binding positions and binding free energies (ΔGb) were estimated from this analysis. ΔGb values were used to calculate relative apparent binding constants from ln Kb = −ΔGb/RT, where R is the ideal gas constant and T is temperature in Kelvin (Table S7). For Fe2+, 35 out of 50 docked conformations were between the fourth and fifth guanines in fragment 1 (Figure 4), but the binding is quite weak. For catechol, preferred binding was found at multiple positions, including the 4th, 5th, 6th and 8th, and 13th guanines (Figure 5), and the binding is stronger. Results for all possible ROS are listed in the Supporting Information, Table S7.</p><p>The two unknown peaks in oxidized fragment 1 (20.3 min) and fragment 2 (28.7 min) from catechol-mediated oxidation (Figures 1A and 2A) were not intense enough for reliable identification. However, docking results (Figure 5D) are consistent with the possibility that the 13th guanine was oxidized in fragment 1. G in codon 249 was probably oxidized in fragment 2 due to the fact that codon 249 is also a "hotspot" in many cancers such as liver, lung, and hepatocellular carcinoma.30</p><!><p>Results above demonstrate successful development of LC-MS/MS methodology to directly locate primary oxidation sites in ds-oligonucleotides of >20 bp without labeling or hydrolysis. This is the first report directly identifying multiple oxidation sites in an intact ds-oligonucleotide. The restriction enzyme assisted methodology is simple, reliable, relatively rapid, and facilitates direct detection and mapping of modified sites. It does not require stable isotope labeling or DNA hydrolysis.</p><p>Results reveal that primary oxidation sites involve only guanines on the p53 exon 7 oligonucleotide. This is consistent with guanines as the most easily oxidized of the DNA nucleobases. In general, reactivity of specific guanines within the exon 7 p53 ds-fragment is influenced additionally by neighboring bases and to some extent by secondary structure of the oligonucleotide.46 The most frequently oxidized guanine by catechol/Cu2+/NADPH in the present study was in codon 248 (11%, Table 2), which correlates with codon 248 as a major mutation hot spot in many cancers.30 Guanines within codons 243, 244, and 245 were oxidized to a smaller extent. Thus, there is a quantitative correlation between the high reactive frequency of codon 248 in the p53 exon 7 oligonucleotide with its high mutation frequency in cancers using an oxidation process that mimics oxidations mediated by redox active metabolites (Scheme 1).10–14 However, in Fenton's reagent oxidation, which may or may not have a directly analogous process in humans, guanine in codon 248 gave a smaller extent of oxidation (Table 2) than codon 243.</p><p>According to the p53 database,30 the mutation frequency ratio of codons 248/243 is 24; of codons 248/244 is 7, and of codons 248/245 is 2. Our ratios of relative amounts of oxidation by catechol/Cu2+/NADPH for codons 248/243 was 1.9 and for codons 248/244 was 1.3, which are smaller than the mutation ratios. While we would not expect exact numerical correspondence in codon reactivity and mutation frequency, the lower reactivity ratios involving guanine in codon 243 may be related to it being the first guanine in our exon 7 oligonucleotide and more reactive due to its proximity to the end of a strand.45,64 Also, the guanine in codon 248 (CGG) is a CpG site. In vivo, all cytosines in CpG sites are methylated, which may increase the reactivity and mutation frequency of the neighbor guanine in tumors.39,46,65 For 248/245, the oxidation reactivity ratio was 4.1, which is larger than that for mutations. This could be related to the first guanine in codon 245 (GGC) being a CpG site, making the guanine more reactive in vivo, resulting in a smaller mutation ratio.</p><p>The ratio of relative amounts of oxidation for codons 248/243 was 0.5 from Fenton's reagent oxidation, which is much lower than that in the database. Fenton's reagent is just an oxidant, not a carcinogen, and may simply oxidize the most accessible guanine with lowest oxidation potential in the sequence. The guanine in codon 243 (ATG) is close to the end of the fragment, adjacent to the AT sequence, and may be more exposed due to a partially unwound duplex with fewer hydrogen bonds.45,64 Thus, the guanine in codon 243 may be more reactive in the fragment than the guanine in codon 248 (CGG) under in vitro conditions.</p><p>Oxidations of guanines at codon 245 (GGC), 246 (ATG), 248 (CGG), and 249 (AGG) were observed in isotopic labeled guanines in oligonucleotides.39 Two distinct sites, first guanine in codon 245 (GGC) and first guanine in codon 248 (CGG) had the highest oxidation percentage at 18.7 and 14.2% compared to our values 2.6 and 10.7%, respectively. This may because both guanine sites of highest reactivity had an identical sequence context MeCGG with a methylated cytosine on the 5' side. In our exon 7, the cytosine is nonmethylated and was less easily oxidized.</p><p>Another important observation is the single major peak for oxidized fragment 1 with Fenton's reagent (Figure 3A), as opposed to four major peaks for oxidized fragment 1 with catechol/Cu2+/NADPH (Figure 1A), suggesting Fenton's reagent is more specific toward DNA oxidative damage in our system. Several reports indicated that free hydroxyl radical caused DNA damage with no marked site specificity.66,67 But some investigations show iron(II) binding to phosphate groups and a guanine N-7 moiety first, with subsequent oxidative damage of dG, oligomers, and calf-thymus DNA in solution.68–70</p><p>Docking studies showed that 35 out of 50 conformations of exon 7 oligonucleotide had Fe2+ docked between the fourth and fifth guanines in fragment 1 (Figure 4, Table S7), somewhat consistent with Fe2+ binding weakly at a specific guanine prior to the actual oxygen transfer, leading to 8-oxodG. However, catechol (Figure 5) was found to bind more strongly at multiple positions, including the 4th, 5th, 6th, 8th, and 13th guanines, which correlates with experimental results showing multiple products (Figure 1, Table 2). Other possible ROS reactants in the catechol reaction also had multiple preferred binding positions on the exon 7 oligonucleotide (Table S7). Thus, the molecular modeling results are consistent with the broader codon specificity found in catechol-induced DNA oxidation.</p><!><p>We describe here an LC-MS/MS methodology to directly sequence and quantify the oxidation sites on ds-oligonucleotides of >20 bp. The high oxidation frequency at codon 244 and 248 with catechol/Cu2+/NADPH, a model for quinoid drug metabolites, coincides with high mutation frequency of the p53 gene in lung and other cancers. Fenton's reagent specificity in oxidation may be consistent with binding of a key oxidizing component on the oligonucleotide near the oxidation site before subsequent oxygen transfer to guanines. On the other hand, multiple binding sites of the model metabolite catechol may explain multiple exon 7 codons oxidized by catechol/Cu2+/NADPH. Our results have implications for understanding the role of oxidation of tumor suppressor genes on carcinogenesis. Future work will focus on investigating tumor suppressor gene oxidation using longer, more representative oligonucleotides and confirming links between oxidation and mutation sites for a broad range of chemicals.</p>

PubMed Author Manuscript

Effect of Natural Nanostructured Rods and Platelets on Mechanical and Water Resistance Properties of Alginate-Based Nanocomposites

A series of biopolymer-based nanocomposite films were prepared by incorporating natural one-dimensional (1D) palygorskite (PAL) nanorods, and two-dimensional (2D) montmorillonite (MMT) nanoplatelets into sodium alginate (SA) film by a simple solution casting method. The effect of different dimensions of nanoclays on the mechanical, water resistance, and light transmission properties of the SA/PAL or MMT nanocomposite films were studied. The field-emission scanning electron microscopy (FE-SEM) result showed that PAL can disperse better than MMT in the SA matrix in the case of the same addition amount. The incorporation of both PAL and MMT into the SA film can enhance the tensile strength (TS) and water resistance capability of the film. At a high content of nanoclays, the SA/PAL nanocomposite film shows relatively higher TS, and better water resistance than the SA/MMT nanocomposite film. The SA/MMT nanocomposite films have better light transmission than SA/PAL nanocomposite film at the same loading amount of nanoclays. These results demonstrated that 1D PAL nanorods are more suitable candidate of inorganic filler to improve the mechanical and water resistance properties of biopolymers/nanoclays nanocomposites.

effect_of_natural_nanostructured_rods_and_platelets_on_mechanical_and_water_resistance_properties_of

Introduction<!>Materials<!>Preparation of SA/PAL and SA/MMT Nanocomposite Films<!>Test of Mechanical Properties<!>Test of Moisture Uptake<!>Characterizations<!>XRD Analysis<!><!>XRD Analysis<!>FTIR Analysis<!><!>Micrographs of PAL, MMT, and Corresponding Composites<!><!>Micrographs of PAL, MMT, and Corresponding Composites<!><!>Micrographs of PAL, MMT, and Corresponding Composites<!><!>Moisture Uptake (MU)<!><!>Moisture Uptake (MU)<!>Mechanical Performance of SA/PAL and SA/MMT Nanocomposites<!><!>Mechanical Performance of SA/PAL and SA/MMT Nanocomposites<!>Light Transmission and Transparency<!><!>Light Transmission and Transparency<!>Conclusions<!>Author Contributions<!>Conflict of Interest Statement

<p>Over the past decades, the commodity plastics (i.e., polyethylene, polypropylene, and polyethylene terephthalate) as commonly used food packaging materials have plaid very important roles in human daily production. However, these plastic packaging materials are totally non-biodegradable, so their widespread use caused serious environmental pollution problems (Souza et al., 2017; Costa et al., 2018; Salama et al., 2018). Therefore, the development of biodegradable films using natural, non-toxic, and environment benign polymers such as polysaccharides, proteins, and lipids has drawn much more attention in both of academic and industrial areas (Mushi and Berglund, 2014; Wang and Jing, 2017; Youssef and El-Sayed, 2018). Among numerous biodegradable natural polymers, sodium alginate (SA) was especially concerned owing to its excellent biocompatibility, film-forming ability, and active functional groups (Shankar et al., 2016; Fabra et al., 2018). SA is an anionic natural biomacromolecule, which is composed of poly-b-1, 4-D-mannuronic acid (M units), and a 1, 4-L-glucuronic acid (G units) in different proportions by 1–4 linkages. It is extracted from marine algae or produced by bacteria, and so it has the advantages including abundance, renewability, non-toxicity, water-solubility, biodegradability, and biocompatibility (Wang and Wang, 2010). However, the inherent hydrophilicity and brittleness of neat SA films limited their applications in film materials (Rhim, 2004; Zhang et al., 2017).</p><p>In order to overcome the drawbacks of neat SA films, a variety of nanoscale particles such as montmorillonite (MMT) (Tunç and Duman, 2010; Zlopasa et al., 2015), graphene oxide (Liu et al., 2017), and cellulose nanocrystals (Sirvio et al., 2014) have been incorporated into the SA matrix to fabricate a nanocomposite. Abdollahi et al. (2013) developed an alginate/MMT nanocomposite by a solvent casting method, and found that the mechanical properties of the alginate/MMT composites were enhanced significantly after the addition of MMT. However, MMT forms an agglomeration in the polymer matrix when its addition amount exceeds a certain value, which leads to the decrease of the mechanical properties of the film. It has been shown that MMT is a 2:1-type layered clay mineral with a sandwiched structure composed of two 2D platelets and interlayer cations (i.e., Na+, Ca2+, Mg2+). The strong hydrogen-binding and electrostatic interaction, and van der Waals forces between two platelets make MMT difficult to be exfoliated and tend to be present in a form of agglomeration (Zhang et al., 2014; Block et al., 2015; Liu et al., 2016). In comparison, natural 1D rod-like nanoclays are easy to be dispersed as nanoscale size, and showed great potential to be used to develop polymer/nanoclays composites (Nikolic et al., 2017; Ajmal et al., 2018; Shankar et al., 2018; Zhang P. et al., 2018). It has been demonstrated that the dispersion of nanoclays in polymer matrix, and the comprehensive performance of the resultant polymer composites exhibited interesting dependence on the shape of fillers. Usually, rod-like nanoclays have a relatively smaller contact surface and weaker interaction amount rods, so that they could probably be dispersed in the polymer matrix well with less aggregation (Bilotti et al., 2009). Palygorskite (PAL) is a naturally available 1D nanorod-like silicate clay mineral (Deng et al., 2012; Zhang et al., 2018b). It consists of two double chains of the pyroxene-type (SiO3)2− like amphibole (Si4O11)6− running parallel to the fiber axis (Gard and Follett, 1968; Zhu et al., 2016; Zhang et al., 2018a). PAL is a potential filler to fabricate polymer composite due to its unique advantages, such as high aspect ratio, large specific surface area, good thermal stability, and high modulus (Huang et al., 2012; Ruiz-Hitzky et al., 2013; Ding et al., 2019). It has been confirmed that the incorporation of silylated PAL into the polyurethane matrix improved significantly the thermal stability and mechanical properties of polyurethane (Peng et al., 2011). In addition, 1D fibrous nanoclay has relatively higher density of silanol groups on its surface than 2D layered silicates, making it able to form more hydrogen bonds with hydrophilic biopolymers (Alcantara et al., 2014). So far, the studies on the comparison of 1D and 2D nanoclays in fabricating SA/nanoclays nanocomposites still received less attention.</p><p>In this paper, we have prepared a series of SA/nanoclays nanocomposite film using 1D PAL and 2D MMT as the inorganic ingredients, and studied the effect of different dimensions of nanoclays on the structure, organic/inorganic interface interaction and the mechanical, water resistance, light transmission properties of the films. The potential of 1D PAL nanorods and 2D MMT for fabricating SA-based nanocomposite film was also assessed by a systematic comparative study.</p><!><p>SA [characteristic viscosity of 1% aqueous solution at 20 °C ≥0.02 Pa●s, mannuronic (M) acid/guluronic (G) acid ratio = 65/35, molecular weight ≈120,000 g/mol] was purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Glycerol was obtained from Guangdong Guanghua Chemical Factory Co., Ltd. (Guangdong, China). MMT was purchased from Zhejiang Fenghong New Material Co., Ltd. (Zhengjiang, China). PAL was purchased from Changzhou Dingbang Clay Co., Ltd (Jiangsu, China).</p><!><p>The SA, SA/PAL, and SA/MMT films were prepared by a solution casting method. The aqueous solution containing 2 wt.% of SA were prepared by dissolving SA powder into distilled water under mechanical stirring for 24 h. Then, glycerol (40 w/w% of the mass of SA) as a plasticizer was added into the SA solution, and the mixture was stirred continuously at 600 rpm to obtain a homogeneous solution. PAL or MMT dispersions (2 wt.%) were prepared by dispersing given amounts of clays in distilled water under mechanical agitation at 600 rpm for 1 h at room temperature. Afterward, the obtained mixture was dispersed with the aid of ultrasound equipment operating at 40 kHz for 10 min. The obtained dispersion of PAL or MMT was added to the SA solutions (the amount of clays is 2, 4, 6, 8, and 10 wt.% of the mass of SA) to form a precursor solution for film-forming. Then, the resultant mixture solutions were poured into the PS dishes (9 cm diameter) and evaporated at atmosphere for 72 h. After drying, the films were peeled off from the plate surface and then kept in a conditioning cabinet with the relative humidity (RH) of 53% for further treatments. The neat SA film was prepared according to a similar procedure without the addition of PAL or MMT.</p><!><p>The tensile performance of the film with length of 80 mm and width of 10 mm was studied using an AG-IS material testing machine (Shimadzu Co., Ltd., Japan). A 200 N load cell was used, and the strain rate was 10 mm/min. Samples were studied in each test and all tests occurred at room temperature. The averages of tensile strength and elongation at break of five specimens of each film were taken and presented.</p><!><p>The moisture uptake (MU) was measured by cutting the film samples into small pieces (2 × 2 cm). The samples were dried at 70°C for 24 h. After the samples were weighed (W0), they were conditioned for 72 h at 76% RH (saturated NaCl solution) to ensure equilibrium of the moisture before reweighing (W1).</p><p>MU of the samples was calculated as follows: (1)MU=[W1-W0]/W0×100%</p><p>An average value of five replicates for each sample was taken.</p><!><p>The crystalline structure of samples was characterized with an X-ray diffractometer (Philips, X'pert PRO, The Netherland), with the Cu Ka radiation at 40 kV and 30 mA. The diffraction patterns were collected in reflection mode by scanning the 2θ range from 5° to 30°, with a scan speed of 6°/min. A FTIR Spectrophotometer (Thermo Nicolet, NEXUS 670, USA) was employed to record the Infrared spectra of films with the attenuated total reflectance (ATR) model in the range of 4,000–500 cm−1. The surface morphology of samples was observed with a JEOL JSM-6701F microscope after the samples were coated with gold. The films for test were immersed in liquid nitrogen and cryo-fractured manually. TEM images of samples were taken with a TECNAIG2-F30 Transmission electron microscope (FEI, USA). Light transmission of films was measured with a UV–vis spectrophotometer [N4, INESA (Group) Co., Ltd, China] at selected wavelengths between 200 and 600 nm.</p><!><p>As shown in Figure 1, no sharp diffraction peaks were observed in the XRD patterns of SA/glycerol films (Figure 1A), indicating that neat SA film shows no crystalline state and the polymer chains are arranged randomly. The neat PAL showed the diffraction peak at 2θ = 8.4°, ascribed to the characteristic reflection of (110) crystallographic plane (Figure 1A) (Xu et al., 2017). The diffraction peak of (110) plane of PAL (2θ = 8.4) was observed in the XRD pattern of SA/PAL nanocomposite films. The position of (110) diffraction peak has not been changed, but the peak intensity increased with increasing the loading amount of PAL in the SA/PAL film. This result indicated that SA polymer chains cannot enter the tunnel to cause the change of the crystal structure of PAL, which is in good agreement with previous reports (Peng et al., 2011; Huang et al., 2012). Different from smectite, PAL has a layer-chain structure in which the structure unit composed of continuous tetrahedrons and discontinuous octahedrons are very stabile and cannot be intercalated or exfoliated, so the crystal structure, and morphology of PAL can remain well after forming nanocomposite.</p><!><p>XRD patterns of (A) PAL, SA film, and SA/PAL nanocomposite films; and of (B) MMT and SA/MMT nanocomposite films.</p><!><p>The XRD patterns of MMT and the SA/MMT nanocomposite films with different loading amount of MMT are presented in Figure 1B. An intense diffraction peak at 2θ = 7.06° (basal spacing is 1.25 nm), ascribed to the characteristic reflection of (001) crystalline plane of MMT, was observed in the XRD pattern of MMT. The characteristic reflection of MMT has not been observed in the XRD pattern of SA/MMT when the content of MMT is 2 wt.%, indicating MMT was exfoliated after its composite with SA polymer chains at a relatively lower MMT loading amounts. When the loading amount of MMT increased to 6 and 10 wt.%, the diffraction peak of MMT shifted from 7.06° to 5.28° and 5.02°, respectively. These results indicated that some SA polymer chains were intercalated into the interlayer space of MMT (Majdzadeh-Ardakani and Nazari, 2010), forming an intercalated-exfoliated nanocomposite structure at relatively higher contents of MMT (Nagarajan et al., 2015).</p><!><p>The incorporation of PAL or MMT into SA matrix was evaluated by ATR-FTIR analysis of films, because the vibrations and shifts of the FTIR absorption bands may reflect the interactions between SA and nanoclays. The ATR-FTIR spectra of the SA and SA/PAL or SA/MMT nanocomposite films are shown in Figure 2. As shown in Figure 2A, the absorption bands of neat SA film at 3,256, 1,600 and 1,410, 1,020 cm−1 can be attributed to the O-H stretching vibration, the COO symmetric, and asymmetric stretching vibration of carboxylate groups, and the C-O-C stretching vibration (Tezcan et al., 2012). After introducing PAL or MMT into the SA matrix, the absorption band at 3,256 cm−1 shifted to low wavenumber region (Figures 2B–G), which confirmed there are strong interaction between SA and PAL or MMT, and the incorporation of PAL or MMT partially break the the hydrogen bonding between SA and glycerol, and formed new hydrogen bonding between SA chains and PAL or MMT (Huang et al., 2004; Liu M. et al., 2012). A similar tendency has also been observed in the CS/PVA/PAL composite film (Huang et al., 2012).</p><!><p>ATR-FTIR spectra of (A) neat SA film, (B) SA/PAL2, (C) SA/PAL6, (D) SA/PAL10, (E) SA/MMT2, (F) SA/MMT6, and (G) SA/MMT10.</p><!><p>Figure 3 showed the SEM and TEM micrographs of PAL and the FESEM image of MMT used in this study. It can be seen from the SEM and TEM images of PAL that the PAL shows good rod-like shape with a diameter of 30–70 nm, and a length of about 0.3–1.5 μm. The rods are slightly aggregated together to form a loose bundle. The MMT exhibits a platelet-like shape, but most of the lamellae of MMT stacked together without presence of single MMT layers.</p><!><p>SEM and TEM micrographs of PAL and SEM micrograph of MMT.</p><!><p>The dispersion of the PAL or MMT in the SA matrix was also studied by SEM observation. The SEM images (magnifications × 5,000 and × 20,000) of the cross-section of SA/PAL nanocomposites with the PAL loading of 0, 2, 6, and 10 wt.% are shown in Figure 4, respectively. The fracture surface of neat SA film showed a glossy and ordered morphology, indicating the neat SA film has homogeneous microstructure, and the SA and glycerol are compatible very well. All the SA/PAL nanocomposites showed much rougher fractured surfaces in comparison with SA film. The bright dots observed in the SEM image of the nanocomposites indicated that the PAL rods embedded within the SA matrix very well. The number of the bright dots reflecting a well-dispersion of inorganic components increased with increasing the loading amount of PAL. The PAL has a well-dispersion in the SA matrix without obvious agglomeration, indicating a good compatibility between SA matrix and PAL. The well-dispersion of PAL in the SA matrix can be ascribed to the following reasons. First, the 1D nanorod-like feature of PAL and the relatively weaker interaction among rods allowed them easy to be dispersed in polymer matrix under the action of shear forces. Secondly, PAL can interact with SA via hydrogen bonding interactions, which is also helpful to strip effectively the crystal bundles as individually dispersed nanorods. In addition, the good compatibility between PAL and SA matrix also indicated a strong matrix-filler interfacial adhesion, which facilitates to enhance the mechanical properties of the resulting SA/PAL composite films.</p><!><p>SEM micrographs of the cross-sections of SA/PAL nanocomposite films: neat SA films (A,B), SA/PAL (4 wt.%) (C,D), SA/PAL (6 wt.%) (E,F), and SA/PAL (10 wt.%) (G,H).</p><!><p>Figure 5 showed the SEM micrographs of the cross-section of SA/MMT nanocomposite films with the MMT loading amounts of 2, 6, and 10 wt.% at different magnifications (× 5,000 and × 20,000). Compared with the neat SA film, the cross-section of SA/MMT nanocomposite films is more rough and uneven in varying degrees. The roughness degree of cross-section of SA/MMT films increased with increasing the loading amount of MMT, owing to the platelet-shaped structure of MMT. It can be seen that the MMT platelets were directionally stacked and densely packed with SA matrix, forming a laminated structure (Figure 5D). When the loading amount of MMT increased to 10%, the cross-section of SA/MMT nanocomposite film becomes looser owing to the agglomeration of MMT in the SA matrix, which may have a negative impact on the performance of the composites. From Figures 4, 5 it can be seen that PAL has relatively better dispersion than MMT in the SA matrix at high loading amounts (6 and 10 wt.%). This behavior may be because that the interactions among PAL rods are relatively weaker than the interaction between the nanoplatelets of MMT (Lu et al., 2005; Liu M. X. et al., 2012).</p><!><p>SEM images of the surface of SA/MMT nanocomposite films: SA/MMT (2 wt.%) (A,B), SA/MMT (6 wt.%) (C,D), and SA/MMT (10 wt.%) (E,F).</p><!><p>MU capability of the SA/PAL and SA/MMT was measured to study the effect of the nanoclays on the water-resistant properties of the SA films. As shown in Figure 6A, The MU of neat SA film was higher than that of SA/PAL and SA/MMT nanocomposites, indicating that the introduction of PAL or MMT can inhibit the penetration of water molecules into the film, and decrease the MU of the nanocomposite films (Almasi et al., 2010). This reduction of MU is because that PAL and MMT are able to form hydrogen bonding networks with SA matrix, increase the surface roughness of film, and block the diffusion pathway of water molecules, which decreased the water sensitivity of the nanocomposites (Wu et al., 2009; Almasi et al., 2010; Bidsorkhi et al., 2014). In addition, the hydrophilicity of PAL or MMT is weaker than that of neat SA matrix, so the incorporation of PAL or MMT is favorable to reduce the MU.</p><!><p>(A) MU of the SA/PAL and SA/MMT nanocomposite films with different loadings of nanoclays and (B) Schematic illustration of the presence of PAL or MMT in the SA matrix.</p><!><p>In comparison, the PAL has a relatively better effect than MMT on reducing the MU of SA film. The total MU of SA film reduced about 12.9% after the incorporation of 10 wt.% PAL, which was higher than that of SA/MMT10 (MU only reduced 8.1%). This can be attributed to the fact that PAL has a better dispersion than MMT in the SA matrix at high loadings of nanoclays (see Figure 6B). The good dispersion of PAL nanorod in SA matrix enable it to form more stable and dense hydrogen bond networks with the hydroxyl groups of SA than MMT (little agglomeration in SA matrix) at the same loading amounts. In addition, SA/MMT nanocomposites had a relatively looser surface structure than SA/PAL nanocomposites at the sample loading amount of 10 wt.%, which enable them to absorb easily more water molecules.</p><!><p>The effect of PAL and MMT addition on mechanical properties of SA film was studied by a tensile testing experiment. As shown in Figure 7A, the addition of PAL or MMT has great influence on the tensile strength of nanocomposites. The tensile strength (TS) of SA/PAL nanocomposites increased by 84.56% (from 13.67 to 25.23 MPa) with an increase of PAL loading amount from 0 to 10 wt.%. This obvious improvement of TS of SA/PAL can be ascribed to the homogenous dispersion of PAL in SA matrix and the strong hydrogen bonding interaction between the silanol groups of PAL and the –OH or –COOH groups of SA. In addition, the high aspect ratio of PAL was also favorable to the stress transfer when the film was subject to be stretched. The similar results were observed in other PAL nanorod-reinforced polymers (Chen et al., 2012; Liu et al., 2014).</p><!><p>TS (A) and EB (B) of SA/PAL and SA/MMT nanocomposites with different contents of nanoclays.</p><!><p>The TS of SA/MMT increased initially, reached the optimal value at the MMT content of 6 wt.%, and then decreased with the increase of addition amounts of MMT. The TS of nanocomposite films increased by 82.96% in contrast to the SA films. The TS of nanocomposites decreased when the MMT loading amount is larger than 6 wt.%. This tendency is ascribed to the aggregation of MMT in the SA matrix, as confirmed by SEM and XRD analysis results. The similar behavior was also observed by previous research result that the tensile strength of gelatin/MMT films decreased after adding 5% of MMT due to a loss in the quality of MMT dispersion in the gelatin matrix (Flaker et al., 2015).</p><p>Figure 7B showed the elongation at breakage (EB) of SA film, SA/PAL, and SA/MMT nanocomposite films. The EB of nanocomposites showed a decreasing trend with the increase of the loading amounts of PAL or MMT. This reduction of EB of SA/PAL and SA/MMT nanocomposite films can be attributed to the fact that the inclusion of rigid clay restricted the motion of SA molecular chains in the matrix. Peng et al. (2011) confirmed that the EB of the waterborne polyurethane/PAL nanocomposite decreased with the increase of PAL loading amount. Slavutsky et al. (2014) also proved the EB of brea gum/MMT composites decreased with increasing the MMT content.</p><!><p>The light transmission of neat SA film, SA/PAL, and SA/MMT nanocomposite films as a function of wavelength is shown in Figure 8. The transmittance of SA film is about 87% at 600 nm (visible region), but it decreased slightly with the increase of PAL or MMT contents. This decrease was possibly related to the presence of PAL or MMT nanoparticles, because light would be absorbed partly by clay nanoparticles, leading to the decrease of energy of transmitted light. This result is similar to the previous findings that the light transmittance of pure polyimide film at 800 nm sharply decreased by incorporating 7 wt.% of PAL (An et al., 2008). Slavutsky et al. (2014) also found that the light transmission of brea gum/MMT composite films decreased after incorporation of MMT.</p><!><p>Light transmission properties of SA film, SA/PAL, and SA/MMT nanocomposite films with different contents of nanoclays.</p><!><p>In addition, the light transmittance of SA/PAL nanocomposites is slightly lower than that of SA/MMT nanocomposites at the same addition amount. The light transmittance of the nanocomposites with addition of 10 wt.% PAL and 10 wt.% MMT at 600 nm was about 67 and 75%, respectively. This result may be due to the diameter of PAL nanorods (50–100 nm) is higher than the thickness of MMT nanoplatelets (about 1 nm) (Ge et al., 2015). Thus, the PAL obstructed the transmission of light more significantly than MMT.</p><!><p>The effects of 1D PAL nanorods and 2D MMT nanoplatelets on the mechanical, water resistance, and light transmission properties of the SA/PAL and SA/MMT nanocomposites were studied by a comparative study. It was revealed that PAL can enhance the mechanical and water resistance properties of SA/PAL nanocomposites better than MMT, because the 1D PAL has better dispersion than 2D MMT in the SA matrix at the same loading amounts. As a result of it, the TS of SA film increased sharply by 84.56% (from 13.67 to 25.23 MPa) after incorporation of 10 wt.% PAL, which is better than the TS of SA/MMT nanocomposite. In addition, the effect of PAL on the reduction of MU was more significant than that of MMT. Therefore, the 1D PAL nanorods are more suitable candidate of nanofillers to fabricate biopolymer-based nanocomposite films than MMT, and this study would lay a foundation to the design and preparation of new types of environment-friendly packing film materials.</p><!><p>DH and ZZ conceived and designed the experiments; ZM performed the experiments; DH, ZM, and ZZ analyzed the data; QQ contributed reagents, materials, and analysis tools; DH wrote the paper.</p><!><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p>

PubMed Open Access

Dipole Switching by Intramolecular Electron Transfer in Single-Molecule Magnetic Complex [Mn 12 O 12 (O 2 CR) 16 (H 2 O) 4 ]

We study intramolecular electron transfer in the single-molecule magnetic complex [Mn 12 O 12 (O 2 CR) 16 (H 2 O) 4 ] for R = -H, -CH 3, -CHCl 2 , -C 6 H 5 , -C 6 H 4 F ligands as a mechanism for switching of the molecular dipole moment. Energetics are obtained using the density functional theory (DFT) with onsite Coulomb energy correction (DFT+U). Lattice distortions are found to be critical for localizing an extra electron on one of the easy sites on the outer ring in which localized states can be stabilized. We find that the lowest energy path for charge transfer is for the electron to go through the center via superexchange mediated tunneling. The energy barrier for such a path ranges from 0.4 meV to 54 meV depending on the ligands and the isomeric form of the complex. The electric field needed to move the charge from one end to the other, thus reversing the dipole moment, is 0.01-0.04 V/Å.

dipole_switching_by_intramolecular_electron_transfer_in_single-molecule_magnetic_complex_[mn_12_o_12

INTRODUCTION<!>COMPUTATIONAL APPROACH<!>RESULTS<!>Negatively charged [Mn 12 ]complex.<!>Energy barrier for dipole switching.<!>Ligand<!>Effect of electric fields.<!>CONCLUSION

<p>[Mn 12 O 12 (O 2 CR) 16 (H 2 O) 4 ] (with R = -H, -CH 3, -CHCl 2 , -C 6 H 5 , -C 6 H 4 F) or "Mn 12 " for brevity is the prototype singlemolecule magnet (SMM). First synthesized in 1980, 1 it is also the first SMM to demonstrate quantum tunneling of magnetization. 2 , 3 , 4 Compared to solid state magnetic materials, advantages of materials based upon SMMs include small size, perfect monodispersity, low cost, and wide variety of ligands. From a practical point of view, SMMs deposited on a large variety of substrates serve as a prolific platform for prototype device investigation because of their high blocking temperatures and abundant choice of ligands for use in tuning the properties of SMMs. 5,6,7 SMMs have been considered as candidates for high-density information storage due to co-existence of electric and magnetic dipoles 2, 8, 9, 10 and potentially strong magneto-electric coupling. 11 , 12 , 13 SMMs also have potential applications in high sensitivity sensors, 14 , 15 , 16 controllable molecular switches for spintronics applications, 17,18,19 and quantum information processing as qubits. 3,20 Functionality of electronic devices can be achieved by manipulating SMM electronic states by electric field, 21 , 22 gate voltage, 23 magnetic field, 24 or circularly-polarized radiation. 25 Functionality of writing and reading information from SMMs recently has been demonstrated. 22,26 Switching the dipole moment of a SMM can modify its structural and electronic configuration. 27 There are several dipole switching mechanisms in different SMMs, including ferroelectric polarization, 28 asymmetric metal electrode screening, 29 carrier trapping/detrapping, 30 , 31 and dipole−electric field interaction. 32 Intramolecular electron transfer is another way a molecule can switch its dipole moment. In metal-organic and covalent-organic frameworks, metal atoms form metal-oxide columns that work as paths for electron transfer. 33,34,35,36,37,38,39,40 A similar charge transfer pathway may also exist in Mn 12 . We suggest that a possible pathway for electron movement from one peripheral Mn atom to another in Mn 12 may be through the nearest connected oxygen. For developing electronic devices based on SMMs, it is very important to know the energy barrier for such electron transfer.</p><p>In this work, we use the density functional theory (DFT) with onsite Coulomb energy correction (DFT+U) to investigate the intramolecular charge transfer and dipole switching process. First, we show that when Mn12 is charged, the added electron is localized on one of the four peripheral Mn atoms, the so-called "easy sites", or on four core Mn atoms. Second, we find the path for moving the electron from a localized location on one easy-site Mn atom to another easy site on the opposite side of the molecule. We show that the lowest energy path is through the four Mn atoms in the center part (core) of the molecule. The energy barrier for such electron transfer is studied by the nudged elastic band method (NEB). 41,42,43 Third, we also study the impact of different ligands on the energy barrier. Finally, we estimate the electric field needed to initiate such an electron transfer process.</p><p>The rest of the paper is organized as follows. The Computational Approach section describes our methodology for calculating localized states in a Mn12 complex, the energy barrier for electron transfer from one localized state to another, and technical details of calculations. The Results section presents the geometry and electronic configurations of the molecular states studied, the calculated energy barriers for Mn12 SMM for the R= -H, -CHCl 2 , -CH 3 , -C 6 H 5 , and -C 6 H 4 F ligands, and the shift of the energy levels as a function of the electric field.</p><!><p>All calculations are based on DFT as implemented in the Vienna ab initio simulation package (VASP). 44,45 We used the Perdew-Burke-Ernzerhof (PBE) exchange-correlation energy functional 46 and projector augmented wave (PAW) pseudopentials. 47 The plane-wave energy cutoff was set to 450 eV. The energy and force convergence tolerances were set to 10 −8 eV and 0.001 eV/Å, respectively. Phonon calculations were done with the Phonopy package. 48,49 We included 10 Å vacuum between periodic SMMs and incorporated the dipole correction 50 , 51 for both energy and potential. In the case of charged systems, a uniform background charge (jellium) was assumed to avoid divergence of the total energy. 52 Brillouin zone sampling was via the single Γ point. We used spinpolarized calculations and applied the rotationally invariant DFT+U method 53,54 for Mn d-orbitals with the on-site Coulomb interaction parameter U set to 4 eV. That value produces the correct Mn magnetic moment in such systems. 55,56 The external electric field is modeled by a saw-tooth potential. 57 The energy barrier for electron transfer from one Mn to another was calculated using the NEB method 41,42,43 as implemented in VASP. To ensure that the electron remains localized on either side of the barrier along the path, the intermediate images in the NEB method start from the configurations calculated from the conventional two-state model. 58,59,60,61 In it, we consider two states for localization of an extra electron on one Mn atom (1) and on another (2), and consider the displacement of those atoms according to reaction coordinates using linear interpolation between initial and final positions, and allow all atoms except those two Mn atoms to relax. The NEB calculations then were done using preconverged configurations obtained from the conventional two-state model with fixed two Mn atoms.</p><p>We have studied several ligands, R= -H, -CHCl2, -CH 3 , -C 6 H 5 , and -C 6 H 4 F. The unit cell in the calculation contains one such [Mn 12 ]molecule (again, in short-hand notation), compensated by a uniform background charge. This allows us to focus on a single molecule instead of the real crystal structure of the [Mn 12 ]complexes, which contains close packed negatively charged molecules and positively charged counterions. 62 , 63 , 64 , 65 Such complexes are computationally much more expensive than the simplified systems considered here since unit cell contains several SMMs and counterions. The simplified systems nonetheless provide insight to key issues of charge transfer and dipole switching. Another reason to model a separate molecule instead of those close packed in the crystal form is that applying of saw-tooth potential with periodic boundary condition (PBC) may disturb the crystal.</p><!><p>We start from the electronic configuration of the neutral [Mn12] 0 molecule, then consider the electronic configuration for the [Mn 12 ]molecule at the negative charge state from having an extra electron. Structural stability is confirmed by the fact that the calculated phonon spectrum has no imaginary frequencies. The section concludes with the calculated energy barrier for electron movement inside the molecule from one Mn atom to another and the electric field needed to initiate electron transfer. These Mn atoms have three electrons in the 3d-oribitals with total spin of S = 3/2 per atom. The interaction among them is ferromagnetic, the total core spin of S = 6. 62 Eight of the Mn atoms, comprising all four core Mn atoms and the four peripheral Mn atoms on the hard sites, and their O neighbors form two parallel Mn-O planes highlighted by the grey and yellow areas in Fig. 1. The direction perpendicular to those planes is associated with an easy axis of magnetization for the molecule. 63 66 show that the magnetic interaction between the Mn 4 core and the Mn 8 ring is antiferromagnetic. Thus the ground state of the neutral [Mn 12 ] 0 complex has a total spin of S = 10. Fig. 2 shows the calculated spin distribution of a [Mn 12 -H] 0 molecule. Depending upon the various ligand configurations the four water molecules are absorbed on different Mn sites. Those configurations can be classified as water isomers according to the location of the four adsorbed water molecules. The possible isomer forms are 1:1:1:1, 2:1:1, or 2:2, depending on whether the water molecules are attached to four, three, or two Mn atoms 63 (see Fig. 3). For the 1:1:1:1 configuration, there are three inequivalent Mn atom types, as pointed out by Pederson et.al. 67 In the configuration shown in Fig. 1, the water molecules are attached to Mn atoms marked as 1-3-5-7. For the 2:2 configuration, water molecules may be attached to the 1 and 5 (or 3 and 7) Mn atoms. Thus there are four nonequivalent Mn atoms in this isomeric form.</p><p>For neutral Mn 12 , all eight peripheral Mn atoms have octahedral oxygen environment with Jahn-Teller (JT) distortion. 62 However, these eight Mn atoms are not equivalent, as shown in Fig. 1. The JT elongation axes for Mn atoms on hard sites (2, 4, 6, 8) are parallel to the easy axis of magnetization, whereas the JT elongation axes for Mn atoms on easy sites (1, 3, 5, 7) have significant nonzero angles with respect to the easy axis of magnetization. The Mn atom on an easy site, e.g., atom 1, is connected to two core Mn atoms, 9 and 12, through oxygen atoms. This is in contrast with a Mn atom on a hard site, e.g., atom 2. It is connected to a single core Mn atom, atom 9, through two oxygen bridges that form a square. The square geometry of the Mn-O-Mn connection for the hard site is more rigid than the six-sided shape for the easy site. Thus it is easier for the easy site to relax and accommodate an extra localized electron. That difference is also manifested by the fact that the water molecules only attach to the easy sites, as shown in Fig. 3.</p><!><p>When the SMM molecule charge state is -1, the additional electron localizes on one of the peripheral easy-site Mn atoms. Its charge state is changed from Mn 3+ to Mn 2+ . That change can be confirmed experimentally by measuring the elongation of Mn-O bond lengths. 64 In the [Mn 12 -H]complex in the 1:1:1:1 isomeric form (shown in Fig. 3a), one water molecule is attached to each Mn atom on the easy sites. This form is maintained when an electron is added to the molecule. The Mn-O bond lengths are increased from 1.937 and 1.929 Å for the Mn 3+ atom (Fig. 4a) and to 2.183 and 2.139 Å for the Mn 2+ atom (Fig. 4b). In Figs. 4 and 5, the significantly changed Mn-O bond lengths caused by accepting an extra electron are highlighted by bold numbers; JT elongation axes are highlighted by yellow lines. Also shown are the energy levels of d-electrons of Mn atoms in an octahedral field with and without JT distortion. In addition to the easy sites on the periphery, our calculations for single molecule [Mn12]in the unit cell with uniform background charge show that Mn atoms from the molecular core also can accept a localized electron. In fact, for an isolated [Mn 12 -H]molecule, the resulting state is predicted to be more stable than one which has an electron localizing on one of the peripheral easy sites. We calculated vibrational modes of the [Mn 12 -H] − molecule with an extra electron localized at either a peripheral easy site or a center site. All vibrational frequencies are real (except for three translational modes), confirming structural stability of both systems. For all studied structures in -1 charge states, the calculated total magnetic moments of SMMs were 21 µ B . Fig. 4d shows the core Mn atom (marked as 9 in Fig. 1) in Mn 4+ charge state, while Fig. 4e shows the core in the Mn 3+ charge state upon acceptance of the extra electron. As pointed out by Christou et.al., 63 a JT distortion is expected for the resulting high-spin Mn 3+ ion. As may be seen from Fig. 4d, the Mn 4+ atom does not have a JT elongation axis. However, after receiving an extra electron, the Mn-O bond lengthens (Fig. 4e) to form a JT elongation axis perpendicular to the parallel Mn-O planes in Fig. 1. That also introduces strain in the rigid [Mn 4 O 4 ] core cubane unit.</p><p>With one exception, for all structures studied with an electron located on the core Mn atom, the JT elongation axes of the core Mn atom are perpendicular to the parallel Mn-O planes. The exception is the -CHCl2 ligand. For it, the JT elongation axis for the core Mn atom lies parallel to the planes. For the other ligands, localizing an extra electron on one of the core Mn atoms becomes less energetically favorable or even unfavorable relative to localizing on one of the easy-sites on the peripheral Mn ring due to the strong ligand field (see Table 1). Thus, in contrast to -H and -CH3 ligands, for the -CHCl 2 ligand localization of an extra electron on the outer ring is energetically more preferable than localization on the core. This fact also confirms the experimental findings that [Mn 12 -CHCl 2 ] SMM is much more effective in attracting electrons than the [Mn 12 -CH 3 ]. The Mn atom 2 in the 3+ charge state, shown in Fig. 4c, has Mn-O bond lengths identical with those for the Mn atom 1 with the JT elongation axis parallel to easy axis of magnetization of the molecule. However, this atom cannot accept an additional electron (neither can the equivalent 4-6-8 atoms), because of the rigid square geometry discussed earlier, so an extra electron initially placed on this atom prefers to move to Mn atoms 1 or 3.</p><p>The 2:2 isomeric form shown in Fig. 3c has four water molecules attached to the 1 and 5 easy-sites in pairs, thus breaking the symmetry between the 1 and 5 sites and the 3 and 7 sites, with the 3 and 7 sites attached to the ligands instead of water molecules. Our calculations show that an extra electron placed on 1 is energetically more favorable than for placement on 3 by 144 meV. Comparison of the 1:1:1:1 and 2:2 isomeric forms shows that an electron localized on 1 in the 1:1:1:1 form is less stable by 21 meV than for localization on 1 in the 2:2 isomeric form. This means that in the 2:1:1 isomeric form (Fig. 3b) localization of an extra electron will be more stable on atom 1, where two water molecules are attached. In contrast, localization on 3 or on 5 will be 21 meV less energetically favorable, and localization on 7 will be ≈150 meV less energetically favorable.</p><!><p>We consider two pathways for electron transfer, either of which can be the first stage of an electron transfer process that leads to dipole switching. Both pathways start from the easy site 1. The first pathway involves electron transfer to the nearest possible localization on the easy-site of the outer ring (3), 1→3. The second pathway is to the nearest Mn atom in the core, 1→9. Table 1 shows the electric dipole moment when the electron is localized on sites 1, 3, 5, 7, and 9, the difference in energy for localized states relative to site 1, and the energy barrier for electron transfer, calculated for an isolated charged [Mn12]molecule in a large box with a uniform positive charge background. The dipole moment is defined as 50,51 ∫ d𝒓𝒓 (𝒓𝒓 − 𝑹𝑹 center )𝜌𝜌 ions+valence (𝒓𝒓), where 𝑹𝑹 center is the center of the [Mn 12 ]molecule and ρ ions+valence is the total charge density. As can be seen from Table 1, transferring an electron from 1 to 3 changes the dipole direction, whereas transferring an electron 1→9 changes the dipole moment magnitude. In the 1:1:1:1 isomeric form, sites 1 and 3 are equivalent for localization, but in the 2:2 isomers those two localized states differ by 144-157 meV, depending on the ligand (see Table 1). The state with an electron localized on site 3 in [Mn 12 -H]is higher in energy by 144 meV. Transferring an electron 1→3 needs to overcome an energy barrier of 146 meV. Because of this, in the 2:2 isomers the 1→3 pathway is not likely. Fig. 6 shows energy profiles for paths 1→3 and 1→9 for the [Mn 12 -H]complex in 1:1:1:1 isomeric form using the conventional two-state model and NEB discussed above. Dots on the NEB curve show several images on the pathway. Displacements 1→A and 1→B correspond to Mn atom shifts to 0.099 and 0.135 Å, correspondingly. For 1→3 the conventional two-state model gives an energy barrier of 31.1 meV, whereas the NEB method finds a pathway with an energy barrier of 5.3 meV. To transfer an electron from 1 to the core 9 requires the overcoming of an energy barrier of 0.4 meV in order to move to the lower energy state and create JT distortion on 9 (see Fig. 4e). Table 1. Electron localization energy differences (in meV) between outer state 1 and core state 9, and between 1 and 3 outer states, energy barriers (in meV) for electron transfer from 1 →3 and 1 →9, electric field for initiation of dipole switching, electric displacement 1 2 D δ , and electric dipole moment of the states in Mn 12 complexes with different ligands. For labeling of the atoms see Fig. 1. In the 1:1:1:1 isomeric form (Fig. 3a) water molecules are attached to 1, 3, 5, and 7 Mn atoms, whereas in 2:2 isomeric form (Fig. 3c) water molecules are attached to 1 and 5 Mn atoms.</p><!><p>Water</p><!><p>Applying an external electric field alters the energetically favorable site on which the extra electron can be localized. This is the basis for dipole switching. Here, we apply an electric field computationally to estimate the strength of an external electric field needed to alter the localization site. The applied electric field is in the plane of easy Mn sites and parallel to the projection of the vector from 1 to 9 onto the same plane. Let 𝐸𝐸 1 (𝐸𝐸 9 ) be the total energy of [Mn 12 ] − with the extra electron localized at 1 (9). Fig. 7 shows the energy difference 𝛿𝛿𝐸𝐸 = 𝐸𝐸 9 − 𝐸𝐸 1 versus electric field strength ℇ. 𝛿𝛿𝐸𝐸 changes linearly with ℇ, and reaches zero at ℇ = −0.034 V/Å for [Mn12-H] -. The negative sign of the electric field corresponds to the direction of the field from the center of the molecule to outside. In order to be sure that our calculation procedure with saw-tooth potential in PBC is valid, we also calculated the electric field when the molecules are separated by 20 Å. The resulting changes in the electric field value are less than 6%. Calculated values of electric field strengths to initiate dipole switching process for all the SMMs studied are in Table 1. We need to point out here, that though this field energetically aligns the two states (1 and 9), it is not enough to move an electron from one atom to another, since it doing that also requires overcoming some activation barrier in order to make displacements on atoms, thus make the JT distortion. However, these calculated values can give some estimation of the actual electric field for dipole switching. Thus, for -H ligand, where the activation barrier is only 0.4 meV, thus much less than the energy difference E9-E1 of 24 meV, the actual electric field for electron movement will be very close to 0.034 V/Å. The slope of the δE dependence over ℇ (see Fig. 7) can be associated with the electric displacement of the media according to 𝛿𝛿𝐸𝐸 = 1/2𝛿𝛿𝛿𝛿ℇ . The slope coefficients 1/2𝛿𝛿𝛿𝛿 are summarized in Table 1. These calculations consider only isolated [Mn 12 ] − molecules. In some experiments, however, the molecules are in crystalline form, with counterions in each unit cell. 62,63,64 The positively charged counterions and [Mn 12 ] − molecules of such structures impose a crystal electric field upon the [Mn 12 ] − molecules. We have calculated the crystal electric field imposed on the Mn 12 molecule in the [Mn 12 -C 6 H 4 F] crystal with lattice parameters a = 17.41 Å, b = 17.41 Å, c = 23.87 Å. 64 Both the crystal electric field in the plane z = c/2 and the orientation of the [Mn 12 ]molecule in the crystal are shown in Fig. 8. The field from the periodic counterions and from all other periodic [Mn 12 ]molecules is included in the plot, but the field from the ligands of the molecule itself is not included. As can be seen from Fig. 8, the Mn 12 molecule in the crystal is oriented such that the Mn atoms on the easy-sites are located at points (corners of the blue square) with minimum possible electric field strength. Thus, the electric field on the easy-site Mn atoms is 0.022 V/Å, whereas on the hard-site atoms it is 0.063 V/Å. The electric field strength on the core Mn atoms is 0.00025 V/Å.</p><!><p>We have studied different possible pathways for dipole switching in [Mn 12 ] -SMM complex with diverse ligands. Ligands play an important role in the energy barrier for electron transfer in the various SMM Mn 12 molecules. We find that the extra electron can be localized on the easy sites of the peripheral Mn ring or on any of the four core Mn atoms. For some ligands, localization on one of the core sites is energetically preferable for isolated molecules. Lattice distortions are found to be critical for localizing the extra electron on one of the easy sites of the peripheral ring. We find that the lowest energy path for charge transfer is for the electron to go through the center via superexchange mediated tunneling. The energy barrier for dipole switching is between 0.4 meV to 54 meV depending on the ligands and the isomeric form of the complex. An important semi-quantitative finding is the huge range in energy barriers for dipole switching, more than two orders of magnitude. We expect this finding to hold up even if the smaller barriers turn out to be strongly sensitive to the particular exchangecorrelation functional. The electric field needed to move the charge from one end to the other thus reversing the dipole moment is 0.01-0.04 V/Å. Parameters for the structures studied are available for download at M 2 QM' GitHub page. 68</p>

ChemRxiv

Bioimaging with Macromolecular Probes Incorporating Multiple BODIPY Fluorophores

Seven macromolecular constructs incorporating multiple borondipyrromethene (BODIPY) fluorophores along a common poly(methacrylate) backbone with decyl and oligo-(ethylene glycol) side chains were synthesized. The hydrophilic oligo(ethylene glycol) components impose solubility in aqueous environment on the overall assembly. The hydrophobic decyl chains effectively insulate the fluorophores from each other to prevent detrimental interchromophoric interactions and preserve their photophysical properties. As a result, the brightness of these multicomponent assemblies is approximately three times greater than that of a model BODIPY monomer. Such a high brightness level is maintained even after injection of the macromolecular probes in living nematodes, allowing their visualization with a significant improvement in signal-to-noise ratio, relative to the model monomer, and no cytotoxic or behavioral effects. The covalent scaffold of these macromolecular constructs also permits their subsequent conjugation to secondary antibodies. The covalent attachment of polymer and biomolecule does not hinder the targeting ability of the latter and the resulting bioconjugates can be exploited to stain the tubulin structure of model cells to enable their visualization with optimal signal-to-noise ratios. These results demonstrate that this particular structural design for the incorporation of multiple chromophores within the same covalent construct is a viable one to preserve the photophysical properties of the emissive species and enable the assembly of bioimaging probes with enhanced brightness.

bioimaging_with_macromolecular_probes_incorporating_multiple_bodipy_fluorophores

INTRODUCTION<!>Design, Synthesis, and Structural Characterization<!>Absorption and Emission Spectroscopies<!>Fluorescence Imaging<!>CONCLUSIONS<!>Materials and Methods<!>General Procedure for the Synthesis of 1a\xe2\x80\x93e<!>1a. GPC<!>1b. GPC<!>1c. GPC<!>1d. GPC<!>1e. GPC<!>General Procedure for the Synthesis of 2a and 2b<!>2a. GPC<!>2b. GPC<!>Synthesis of 2c<!>6<!>8<!>11<!>Nematode Imaging<!>Cell Imaging<!>

<p>Fluorescence measurements1 permit the detection of cells in biological fluids2 and the visualization of tissues in living organisms.3 Fast response, inherent sensitivity, and noninvasive character are the main reasons behind their widespread application in cytometry and imaging. Their experimental implementation, however, requires first the labeling of a given biotarget with appropriate fluorescent probes.4 Generally, a fluorescent chromophore is attached to a targeting agent, in the form of an antibody, an aptamer, a peptide, or even a relatively small ligand, capable of selectively directing the signaling unit to the target of interest. The supramolecular association of the complementary components then immobilizes the fluorophore on the target and allows the detection of the latter after collecting the emission of the former. Under these conditions, the brightness of the fluorophore ultimately dictates the intensity of the detected signal and the sensitivity of the overall protocol.2,3 This particular photophysical parameter is equal to the product of the molar absorption coefficient (ε) of the fluorophore at the excitation wavelength and its fluorescence quantum yield (ϕ).1 For most organic fluorophores, ε is lower than 102 mM−1 cm−1, and therefore, their brightness (ε × ϕ) barely approaches 102 mM−1 cm−1, even when ϕ is close to unity. For example, indocyanine green, which is one of the most common synthetic dyes for fluorescence measurements in vivo and is approved for use in humans, has a brightness of only 11 mM−1 cm−1.5</p><p>In principle, more than one fluorescent chromophore can be attached to a single targeting agent, in order to increase the values of ε and, hence, the brightness associated with each binding event. However, the clustering of independent fluorophores within a restricted volume generally leads to significant interactions in the excited state that inhibit radiative deactivation.1 Therefore, the increase in ε with the assembly of multiple chromophores into a single molecular, macromolecular, or supramolecular construct is often counteracted by a pronounced decrease, or even complete suppression, of ϕ. As a consequence of this fundamental limitation, inherent to the excitation dynamics of organic chromophores, the identification of viable designs to integrate multiple fluorophores within relatively compact synthetic structures to enhance brightness remains far from trivial.6–15</p><p>Hundreds of identical chromophores can be encapsulated within the same polymer particle to generate fluorescent assemblies with a brightness approaching 104 mM−1 cm−1.16–20 Such brightness levels are several orders of magnitude greater than those accessible with individual organic fluorophores and far exceed even those of semiconductor quantum dots.21,22 Nonetheless, the encapsulation of the fluorescent guests within the polymer particles and, often, also the structural integrity of the host matrix itself are a result of noncovalent interactions. The reversible nature of these supramolecular contacts can lead to the leakage of the guests out of the particles, as well as to the disassembly of the hosts, under the extreme dilution that can occur in biological fluids. A possible solution to this problem can be the covalent connection, rather than the noncovalent encapsulation, of many emissive chromophores to a common polymer chain11,15 or cross-linked scaffold.12 For example, polymers with BODIPY fluorophores in their side chains can be assembled with relatively simple synthetic procedures to efficiently generate covalent constructs with a brightness that can also approach 104 mM−1 cm−1.23,24 Nonetheless, the hydrodynamic diameter (dH) of these fluorescent constructs, as well as those of their noncovalent counterparts, exceeds 50 nm in most instances.16–20 Such physical dimensions are more than 2 orders of magnitude greater than those of a single chromophore and can have detrimental effects on the supramolecular association of the targeting agent, as well as complicate administration in vivo and clearance from the organism. Thus, viable structural designs to engineer fluorescent probes with bright emission together with long-term stability, compact dimensions, targeting compatibility, and lack of toxicity are still very much needed. The availability of such materials would be particularly valuable in the biomedical laboratory and could ultimately facilitate the implementation of clinical and surgical applications based on convenient fluorescence measurements.</p><p>Our laboratories developed a series of amphiphilic polymers with multiple decyl and oligo(ethylene glycol) side chains along a common poly(methacrylate) backbone.25–28 In aqueous environments, these macromolecules assemble spontaneously into particles with dH close to 20 nm, hydrophilic surface, and hydrophobic interior. In the process of assembling, they can capture multiple hydrophobic fluorophores in their interior and transfer these, otherwise insoluble, chromophores into aqueous solutions. The supramolecular shell encapsulating the fluorescent components protects them from the aqueous environment and preserves their photophysical properties. Furthermore, the hydrophobic side chains of the macromolecular components isolate the entrapped fluorophores from each other, preventing interchromophoric interactions in both the ground as well as excited states. As a result, a negligible influence on ϕ is generally observed upon encapsulation of multiple fluorophores within these supramolecular containers, offering the possibility of constructing supramolecular assemblies with relatively large brightness. Nonetheless, noncovalent contacts are solely responsible for holding fluorophores and polymers together. In fact, these supramolecular nanocarriers can discharge their fluorescent cargo into the many lipophilic domains found in biological preparations.28,29 These considerations suggested the design of similar amphiphilic macromolecular constructs with multiple fluorophores covalently connected to the polymer backbone, instead of being noncovalently encapsulated into the corresponding supramolecular nanocarriers.28 This article reports the synthesis and characterization of two families of macromolecular probes engineered around this structural design, together with the spectroscopic investigation of their photophysical properties and the assessment of their performance in model biological preparations.</p><!><p>Two structural designs for the covalent integration of BODIPY chromophores within amphiphilic macromolecular constructs were envisaged. In one, the chromophoric units are directly attached to a poly(methacrylate) backbone together with decyl and oligo(ethylene) side chains (1 in Figure 1). In the other, the chromophores are connected to the ends of the hydrophobic side chains (2 in Figure 2). Both were prepared from the random polymerization of the corresponding methacrylate monomers in tetrahydrofuran (THF), under the assistance of azobis(i-butyronitrile) (AIBN). In the first instance, the stoichiometry of the three monomers was varied systematically to produce five macromolecules (1a–e in Table 1) differing in the number (N in Table 1) of chromophores per polymer chain and in the ratio (χ in Table 1) between the hydrophobic and hydrophilic side chains. The second design was exploited to assemble macromolecules without and with (2a and 2b in Figure 2) a third side chain terminated by a carboxylic acid. The latter was treated with N-hydroxysuccinimide (NHS), N,N-dicyclohexylcarbodiimide (DCC), and 4-N,N-dimethylaminopyridine (DMAP) in dichloromethane and then coupled to a secondary antibody (9 in Figure 2) in N,N-dimethylformamide (DMF) and bicarbonate buffer (BCB). The resulting conjugate (2c in Figure 2) was isolated from any unreacted starting materials by size-exclusion chromatography.</p><p>Gel permeation chromatography (GPC) of 1a–e and 2a,b indicated their number-average molecular weight (M̅n in Table 1) to range from 14.0 to 24.9 kDa with a dispersity index (Đ in Table 1) varying from 1.42 to 2.36. Integration of the resonances associated with protons of the BODIPY chromophores, hydrophobic chains, and hydrophilic tails in the corresponding 1H NMR spectra provided estimates of N and χ (Table 1). For example, the 1H NMR spectrum (Figure 3) of 1e shows a multiplet at 7.30 ppm for the two pairs of homotopic protons on the phenylene ring of the BODIPY chromophores, a singlet at 3.36 ppm for the methoxy protons at termini of the hydrophilic chains, and a multiplet at 0.88 ppm for the methyl protons at the ends of the hydrophobic chains. These three sets of resonances integrate for 4, 5, and 18 protons, respectively. These values correspond to N of 5.3 and χ of 3.6. Similarly, the 1H NMR spectrum (Figure 3) of 2b shows a pair of doublets at 8.16 and 7.39 ppm for the phenylene protons of the BODIPY chromophores and a singlet at 3.36 ppm for the methoxy protons of the hydrophilic chains. These resonances integrate for 16 and 40 protons, respectively, and correspond to N and χ of 6.4 and 0.3, respectively. Additionally, the spectrum shows also a multiplet at 2.64 ppm for the methylene protons adjacent to the carboxylic acid, integrating for 4 protons. This value suggest that 2b incorporates an average of 1.6 chains with terminal carboxylic acids within its macromolecular backbone.</p><!><p>The absorption and emission spectra (Figure S3) of a model monomer (10 in Table 2) dissolved in THF show the characteristic bands of the BODIPY chromophore with maxima at wavelengths (λAb and λEm in Table 2) of 527 and 541 nm. The very same bands are also observed for all polymers (Figure 4) with minimal shifts in λAb and λEm. The presence of multiple BODIPY chromophores, however, enhances the molar absorption coefficient (ε in Table 2) at λAb from 58.0 mM−1 cm−1 for monomer 10 to up to 429.2 mM−1 cm−1 for polymer 2a, but has negligible influence on the fluorescence quantum yield (ϕ in Table 2). Specifically, ϕ is 0.63 for 10 and ranges from 0.45 to 0.71 for the polymers. As a result, the brightness (ε × ϕ in Table 2) of all polymers, with exception of 1c, is greater than that of the monomer and approaches 206.0 mM−1 cm−1 for polymer 1e, while it is only 36.4 mM−1 cm−1 for 10. Furthermore, this photophysical parameter increases with N for both series of polymers. The only outliner on this trend is 1d, which has more chromophores per polymer chain, but lower brightness, than 1e. Presumably, this apparent contradiction is a consequence of the high relative amount (χ in Table 1) of hydrophobic components in 1e, which ensure the effective insulation of the chromophores from each other. These observations suggest that the brightness of these macromolecules can be optimized even further with the elongation of their polymer backbone, and hence an increase of their M̅n, as long as a significant excess of hydrophobic components, relative to their hydrophilic counterparts, is maintained within the overall amphiphilic construct.</p><p>The transition from THF to phosphate buffer saline (PBS) has negligible influence on the photophysical parameters of the monomer (Table 2 and Figure S3). By contrast, it shifts bathochromically λEm and depresses ϕ for all polymers (Table 2 and Figure 4). These observations suggest that the environment around the BODIPY chromophores changes drastically on going from organic to aqueous solution only when they are attached to a macromolecular backbone. Presumably, the amphiphilic polymers alter their geometry significantly to avoid direct exposure of their hydrophobic domains to water molecules and encourage the nonradiative deactivation of the excited BODIPY chromophores. In spite of these effects, the brightness of some of the polymers remains greater than that of the monomer, which is only 22.6 mM−1 cm−1 under these conditions. Specifically, 1e maintains the largest brightness, out of all the macromolecular constructs tested, to approach 61.2 mM−1 cm−1.</p><p>The photophysical parameters of 1a–e and 2a,b were determined at a concentration of 30 µg mL−1. This value is lower than the corresponding critical micellar concentration (CMC in Table 1), which ranges from 43.4 to 104.1 µg mL−1, and was selected to avoid the aggregation of the amphiphilic polymers into supramolecular assemblies in aqueous environment. In turn, the CMC of all macromolecules was determined with the aid of a hydrophobic BODIPY chromophore (14 in Figure S4) with extended electronic conjugation. This compound is essentially insoluble in PBS and its fluorescence can be detected only in the presence of sufficient amounts of any one of the seven amphiphilic polymers. Indeed, plots (Figure S4) of the emission intensity of 14 against the polymer concentration all show a sudden fluorescence increase above a given concentration threshold, which is the corresponding CMC value. These observations suggest that at concentrations greater than the threshold values the amphiphilic macromolecules assemble into supramolecular nanocarriers capable of capturing 14 in their hydrophobic interior, transferring it into the aqueous phase, and allowing the detection of its fluorescence. Consistently, dynamic light scattering (DLS) measurements performed at concentrations greater than CMC confirm the formation of nanoscaled aggregates with hydrodynamic diameters (dH in Table 1) ranging from 8.0 to 21.0 nm.</p><p>The methyl groups in positions 1 and 7 of the BODIPY chromophores of all polymers, as well as of the model monomer, are expected to hinder the rotation of the heterocyclic platform about the [C–C] bond connecting it to the adjacent phenylene ring and preserve ϕ. Indeed, literature data30–35 demonstrate that the conformational freedom of unsubstituted BODIPY derivatives facilitates instead the nonradiative deactivation of their excited state and translates into a significant dependence of their ϕ and fluorescence lifetime (τ) on the viscosity of the surrounding environment. By contrast, the normalized absorption and emission spectra (Figure S5) of 10 in THF and glycerol are essentially identical and ϕ is 0.63 in one solvent and 0.66 in the other, respectively. In both environments, the fluorescence decays monoexponentially with similar kinetics and τ is 4.67 ns in THF and 5.99 ns in glycerol. These observations confirm that the steric hindrance engineered into the BODIPY of 10 restricts rotation, in agreement with the logic behind the structural design of these chromophores. The very same constraints are also designed into the BODIPY components of the polymers. Nonetheless, their photophysical properties appear to be affected by the viscosity of the solvent. Specifically, comparison of the normalized emission spectra (Figure S5) of 1e in THF and glycerol shows a bathochromic shift of 11 nm and considerable broadening with a viscosity increase, while the absorption spectra in both solvents are almost identical. Additionally, ϕ decreases from 0.67 in THF to only 0.14 in glycerol and τ shortens from 4.66 ns in one solvent to 1.41 ns in the other (Figure S6). These trends are in contrast to the negligible changes observed for the model monomer and the enhancements in ϕ and τ with viscosity reported for unsubstituted BODIPY derivatives.30–35 Such an apparent contradiction indicates that the influence of viscosity on the photophysical properties of 1e must be a result of the restricted conformational freedom on the amphiphilic scaffold, rather than the actual chromophoric components. Presumably, the inability of the amphiphilic polymer to reorganize in a viscous medium, upon excitation of the chromophores, is responsible for promoting the nonradiative deactivation of the BODIPY components and causing the observed decrease in ϕ and τ.</p><!><p>Images of live Caenorhabditis elegans, microinjected with either blank PBS (a in Figure 5) or a PBS solution of 10 (b in Figure 5) in the gonadal region, show the fluorescent probes to accumulate in the fat droplets in the intestinal tract of the nematode. Under identical experimental conditions, injection of 1e (c in Figure 5) resulted in bright, diffuse fluorescence throughout the rest of the nematode with the exception of the eggs. Presumably, the different distributions are a consequence of the amphiphilic character of 1e and hydrophilic nature of 10. The amphiphilic polymer can diffuse much more easily than the hydrophilic model into additional tissues of the worm, possibly via trafficking through the coelomocytes.</p><p>The emission intensity detected within the living organism for the macromolecular probe is significantly greater than that measured for the model monomer. Fluorescence profiles (e and f in Figure 5), collected along lines drawn across the imaged nematodes, reveal a 4-fold enhancement in emission for the polymer, even although the concentration (50 µM) of the probes was the same in both injected solutions. These observations are fully consistent with the brightness measured spectroscopically for the two systems in PBS, which is 61.2 mM−1 cm−1 for 1e but only 22.6 mM−1 cm−1 for 10 (Table 2), and demonstrate that the macromolecular probe retains its photophysical properties within the living organism. Furthermore, the injected polymer does not appear to have any effect on the nematode. Sequences of images (Videos S1–S3), recorded to monitor the movements of the worms in real time, do not reveal any significant difference in the behavior of three injected nematodes over the course of 60 min.</p><p>The ability of 2c to immunolabel biological preparations can be assessed with the aid of Alexa Fluor 647-conjugated affinipure goat anti-mouse IgG (H+L). This model bioconjugate incorporates the very same secondary antibody of 2c and is known to associate with anti-α-tubulin to allow the imaging of tubulin filaments in a variety of cell lines.4 Furthermore, the absorption and emission bands of the BODIPY fluorophores of 2c and those of the Alexa dyes (Figure S7) of the model system are sufficiently resolved across the visible region of the electromagnetic spectrum to permit the imaging of both in separate detection channels. As a result, cells can be labeled with the two conjugates simultaneously and their localization probed independently. Specifically, images of HeLa cells treated with both bioconjugates show the BODIPY fluorescence in one channel (a in Figure 6) and Alexa emission in the other (b in Figure 6). Both images reveal the characteristic shape of the tubulin filaments and an overlay (c in Figure 6) of the two clearly shows the colocalization of the two sets of probes. These observations demonstrate that the conjugation of the macromolecular probe to the antibody does not affect the ability of the biomolecule to associate selectively with its target and that such antibody–polymer construct can, indeed, be employed to label and visualize biological preparations.</p><!><p>Multiple BODIPY chromophores can be appended to a common poly(methacrylate) backbone together with decyl and oligo(ethylene glycol) side chains. The hydrophobic decyl components isolate the fluorophores effectively from each other, preventing interchromophoric interactions. As a result, the fluorescence quantum yield of these macromolecular constructs remains relatively high and is comparable to that of a model BODIPY monomer. In turn, the hydrophilic components ensure solubility in aqueous environment, where the photophysical properties of the emissive species are, once again, preserved by the effective insulation of the hydrophobic side chains. The presence of multiple chromophoric units, however, enhances the molar absorption coefficient significantly and translates into a 3-fold increase in brightness, relative to the model BODIPY. Such macromolecular probes can be microinjected into living nematodes, where they retain their characteristic brightness levels and allow the visualization of the organisms with signal-to-noise ratios greater than those accessible with the model monomer. Furthermore, the fluorescent polymers do not have any significant effects on the behavior of the living organisms. The architecture of these macromolecular assemblies can be modified to permit the subsequent conjugation of antibodies and enable the immunolabeling of biological preparations. Specifically, carboxylic acids can be appended to the termini of some of the hydrophilic chains and then connected to the primary amino groups of a secondary antibody. Comparison of the resulting antibody conjugates to model systems demonstrate that the biomolecules retain their ability to bind their complementary targets and allow the visualization of the tubulin filaments of model cells with optimal signal-to-noise ratios. In summary, these particular structural designs for the covalent integration of multiple fluorescent chromophores into the same construct (1) provide brightness levels greater than those accessible with conventional fluorophores, (2) allow the imaging of living organisms, and (3) enable the immunolabeling of model biological preparations. Thus, a general strategy for the assembly of macromolecular probes with optimal photophysical properties and targeting capabilities for bioimaging applications can ultimately evolve from these investigations.</p><!><p>Chemicals were purchased from commercial sources and used as received with the exception of THF, which was distilled over Na and benzophenone, and H2O, which was purified with a Barnstead International NANOpure DIamond Analytical system. Compounds 3–5, 10, 12, and 14 were prepared according to literature procedures.29,36–38 The synthesis of 6 and 8 are illustrated in Figures S1 and S2. GPC was performed with a Phenomenex Phenogel 5-µm MXM column (7.8 × 300 mm) operated with a Shimadzu Nexera X2 system in THF at a flow rate of 1.0 mL min−1. Monodisperse polystyrene standards (2.7–200.0 kDa) were employed to determine the M̅n and Đ of the polymers from the GPC traces, following a literature protocol.39 EISMS was performed with a Bruker micrOTO-Q II spectrometer. NMR spectra were recorded with a Bruker Avance 500 spectrometer. DLS measurements were performed with a Malvern ZEN1600 apparatus. The values listed for dH in Table 1 are averaged over ten independent experiments of ten runs of 10 s each. Absorption spectra were recorded with a Varian Cary 100 Bio spectrometer, using quartz cells with a path length of 1.0 cm. Emission spectra were recorded with a Varian Cary Eclipse spectrometer in aerated solutions. Fluorescence quantum yields were determined with a fluorescein standard, following a literature protocol.40 Time-correlated single-photon counting measurements were performed with an Edinburgh Analytical Instruments nF920 spectrometer. Samples were excited at 405 ± 10 nm, using a light-emitting diode with a pulse width of 1 ns, and the emission intensity was recorded at 540 nm.</p><!><p>A solution of 3 (23.2 mg, 0.05 mmol), 4 (40.9 mg, 0.18 mmol for a; 56.8 mg, 0.25 mmol for b; 40.9 mg, 0.18 mmol for c; 18.2 mg, 0.08 mmol for d; 68.1 mg, 0.3 mmol for e), 5 (540 mg, 0.27 mmol for a; 540 mg, 0.27 mmol for b; 340 mg, 0.17 mmol for c; 280 mg, 0.14 mmol for d; 300 mg, 0.15 mmol for e), and AIBN (4.9 mg, 0.03 mmol) in degassed THF (8 mL) was heated for 72 h at 75 °C under Ar in a sealed vial. After cooling down to ambient temperature, the reaction mixture was transferred to a centrifuge tube and diluted with THF to a total volume of 10 mL. Hexane was added in portions of 1 mL and the tube was shaken vigorously, after each addition, until the formation of a precipitate was clearly observed. After centrifugation, the oily layer at the bottom of the tube was separated from the supernatant and dissolved in THF (10 mL). The treatment with hexane, followed by centrifugation, was repeated 3 times and the final oily residue was dried under reduced pressure to give 1 (90 mg for a; 100 mg for b; 65 mg for c; 39 mg for d; 48 mg for e) as a red solid.</p><!><p>M̅n = 14.0 kDa, Đ = 1.42; 1H NMR (CD2Cl2): δ = 0.82–0.95 (19H, m), 0.96–1.18 (17H, m), 1.20–1.46 (58H, m), 1.54–1.74 (15H, m), 1.76–2.06 (11H, m), 2.09–2.16 (4H, m), 2.33 (4H, bs), 2.44–2.56 (6H, m), 3.36 (15H, s), 3.39–3.82 (702H, m), 3.86–4.25 (16H, m), 7.18–7.42 (4H, m).</p><!><p>M̅n = 21.8 kDa, Đ = 1.90; 1H NMR (CD2Cl2): δ = 0.81–0.96 (13H, m), 0.96–1.19 (15H, m), 1.21–1.48 (45H, m), 1.54–2.22 (34H, m), 2.34 (3H, s), 2.50 (4H, s), 3.36 (12H, s), 3.41–3.86 (551H, m), 3.88–4.35 (17H, m), 7.18–7.44 (4H, m).</p><!><p>M̅n = 17.9 kDa, Đ = 1.69; 1H NMR (CDCl3): δ = 0.75–0.90 (28H, m), 0.94–1.02 (20H, m), 1.18–1.38 (106H, m), 1.55–1.66 (16H, m), 1.68–1.92 (20H, m), 2.29 (8H, bs), 2.45–2.57 (14H, m), 3.37 (23H, s), 3.52–3.74 (1024H, m), 3.89–4.20 (23H, m), 7.16–7.24 (4H, m).</p><!><p>M̅n = 16.4 kDa, Đ = 1.53; 1H NMR (CD2Cl2): δ = 0.81–0.96 (6H, m), 0.97–1.05 (4H, m), 1.19–1.46 (23H, m), 1.65 (4H, bs), 2.14 (2H, s), 2.20–2.40 (4H, m), 2.44–2.56 6H, m), 3.36 (3H, s), 3.41–3.82 (187H, m), 3.90–4.22 (64H, m), 7.19–7.41 (4H, m).</p><!><p>M̅n = 24.9 kDa, Đ = 2.34; 1H NMR (CD2Cl2): δ = 0.81–0.95 (18H, m), 0.96–1.09 (11H, m), 1.19–1.48 (63H, m), 1.65 (9H, bs), 1.72–1.86 (6H, m), 2.23–2.42 (5H, m), 2.51 (5H, bs), 3.36 (5H, s), 3.43–3.82 (254H, m), 3.86–4.24 (11H, m), 7.22–7.38 (4H, m).</p><!><p>A solution of 6 (100 mg, 0.15 mmol), 7 (225 mg, 0.45 mmol for a; 205 mg, 0.41 mmol for b) without (for a) or with 8 (29 mg, 0.03 mmol for b), and AIBN (5 mg, 0.03 mmol) in degassed THF (8 mL) was heated for 72 h at 75 °C under Ar in a sealed vial. After cooling down to ambient temperature, the reaction mixture was transferred to a centrifuge tube and diluted with THF to a total volume of 10 mL. Hexane was added in portions of 1 mL and the tube was shaken vigorously, after each addition, until the formation of a precipitate was clearly observed. After centrifugation, the oily layer at the bottom of the tube was separated from the supernatant and dissolved in THF (10 mL). The treatment with hexane, followed by centrifugation, was repeated 3 times and the final oily residue was dried under reduced pressure to give 2 (100 mg for a; 110 mg for d) respectively as a red solid.</p><!><p>M̅n = 20.1 kDa, Đ = 2.26; 1H NMR (CDCl3): δ = 0.80–0.91 (25H, m), 0.95–1.02 (11H, m), 1.25–1.30 (19H, m), 1.32–1.40 (15H, m), 1.58–1.65 (4H, m), 1.70–1.78 (4H, m), 1.80–1.90 (8H, m), 2.50–2.56 (6H, s), 3.38 (15H, s), 3.51–3.72 (158H, m), 3.85–3.96 (3H, m), 4.04–4.14 (10H, m), 4.32–4.40 (2H, m), 7.41 (2H, d, 8 Hz), 8.18 (2H, d, 8 Hz).</p><!><p>M̅n = 15.8 kDa, Đ = 2.36; 1H NMR (CDCl3): δ = 0.75–0.90 (7H, m), 0.92–1.05 (11H, m), 1.25 (7H, s), 1.28–1.38 (12H, m), 1.40–1.1.48 (3H, m), 1.55–1.62 (3H, m), 1.75–1.90 (14H, m), 2.25–2.35 (4H, m), 2.58 (6H, s), 2.60–2.67 (1H, m), 3.36 (10H, s), 3.55–3.70 (113H, m), 3.78–3.82 (1H, m), 3.85–3.95(2H, m), 4.08 (7H, bs), 4.30–4.37 (2H, m), 7.39 (2H, d, 8 Hz), 8.16 (2H, d, 8 Hz).</p><!><p>A solution of DCC (19.3 mg, 0.09 mmol) in CH2Cl2 (2 mL) was added dropwise, over the course of 20 min, to a solution of 2b (150 mg, 0.009 mmol), DMAP (2.2 mg, 0.1 mmol) and NHS (12.9 mg, 0.11 mmol) in CH2Cl2 (6 mL) maintained at 0 °C under Ar. The reaction mixture was allowed to warm up to ambient temperature and was stirred for 24 h under these conditions. The resulting precipitate was filtered off and the solvent was distilled off under reduced pressure to afford a red oil (120 mg). 1H NMR (CDCl3): δ = 0.78–0.90 (7H, m), 0.92–1.05 (10H, m), 1.22–1.28 (10H, m), 1.30–1.40 (12H, m), 1.42–1.1.50 (3H, m), 1.55–1.65 (3H, m), 1.75–1.90 (13H, m), 2.23–2.34 (4H, m), 2.52 (6H, bs), 2.80–2.86 (2H, m), 3.36 (10H, s), 3.55–3.70 (109H, m), 3.78–3.82 (1H, m), 3.85–3.95 (2H, m), 4.04–4.15 (8H, m), 4.30–4.37 (2H, m), 7.39 (2H, d, 8 Hz), 8.16 (2H, d, 8 Hz). An aliquot (1.5 mg) of the residue was dissolved in PBS (100 µL) and combined with an aqueous solution of 9 (Jackson ImmunoResearch Inc. 115–005–003; 1 mg, 435 µL), whose pH was adjusted to 8.8 with BCB (1 M, 75 µL). The mixture was stirred for 1 h at ambient temperature and purified by column chromatography (Sephadex G25, PBS). The collected fractions (0.4 mL each) were analyzed by absorption spectroscopy and those showing the characteristic absorption of the BODIPY chromophores at 528 nm were used for further spectroscopic and imaging experiments.</p><!><p>A solution of DCC (106.6 mg, 0.52 mmol) in CH2Cl2 (2 mL) was added dropwise, over the course of 20 min, to a solution of 11 (114 mg, 0.48 mmol), DMAP (12 mg, 0.1 mmol), and 12 (200 mg, 0.48 mmol) in CH2Cl2 (5 mL) maintained at 0 °C under Ar. The reaction mixture was allowed to warm up to ambient temperature and was stirred for 24 h under these conditions. The resulting precipitate was filtered off and the solvent was distilled off under reduced pressure. The residue was purified by column chromatography [SiO2: hexane/EtOAc (4:1, v/v)] to afford 6 (120 mg, 39%) as a red oil. ESIMS: m/z = 671.3783 [M + Na]+ (m/z calcd. for C38H51BF2N2O4Na = 671.3808); 1H NMR (CDCl3): δ = 0.97 (6H, t, 8 Hz), 1.27 (6H, s), 1.30–1.43 (10H, m), 1.45–1.52 (2H, m), 1.65–1.72 (2H, m), 1.78–1.85 (2H, m), 1.95 (3H, s), 2.28–2.234 (4H, q, 8 Hz), 2.55 (6H, s), 4.13–4.17 (2H, t, 8 Hz), 4.36–4.39 (2H, t, 8 Hz), 5.55 (1H, s), 6.10 (1H, s), 7.40– 7.42 (2H, d, 8 Hz), 8.17–8.19 (2H, d, 8 Hz); 13C NMR (CDCl3): δ 11.9, 12.5, 14.5, 14.7, 17.1, 17.8, 18.3, 18.9, 26.0, 28.6, 29.3, 29.4, 64.8, 65.5, 125.2, 128.6, 128.7, 130.1, 130.3, 130.9, 133.1, 136.5, 138.1, 138.7, 140.6, 154.2, 166.2, 167.6.</p><!><p>A solution of succinic anhydride (2.28 g, 22.8 mmol) and 13 (4.0 g, 7.6 mmol) in pyridine (21 mL) was stirred for 48 h at ambient temperature. The solvent was distilled off under reduced pressure and the residue was purified by column chromatography [SiO2:hexane/EtOAc (3:2, v/v)] to afford 8 (3.2 g, 67%) as a colorless oil. ESIMS: m/z = 649.3031 [M + Na]+ (m/z calcd. for C28H50O15Na = 649.3047); 1H NMR (CDCl3): δ = 1.92 (3H, s), 2.62 (6H, bs), 3.55–3.72 (56H, m), 4.20–4.30 (5H, m), 5.55 (1H, s), 6.10 (1H, s).</p><!><p>A solution of DCC (1.3 g, 6.3 mmol) in CH2Cl2 (5 mL) was added dropwise, over the course of 20 min, to a solution of 1,10-decanediol (1.0 g, 5.7 mmol), DMAP (140 mg, 1.1 mmol), and methacrylic acid (494 mg, 5.7 mmol) in CH2Cl2 (5 mL) maintained at 0 °C under Ar. The reaction mixture was allowed to warm up to ambient temperature and stirred for 24 h under these conditions. The resulting precipitate was filtered off and the solvent was distilled off under reduced pressure. The residue was purified by column chromatography [SiO2:hexane/EtOAc (4:1, v/v)] to afford 11 (700 mg, 50%) as a colorless oil. ESIMS: m/z = 265.1780 [M + Na]+ (m/z calcd. for C14H26O3Na = 265.1780); 1H NMR (CDCl3): 1.24–1.40 (12H m), 1.50–1.70 (6H, m), 1.92 (3H, s), 3.59–3.63 (2H, t, 8 Hz), 4.09–4.13 (2H, t, 8 Hz), 5.52 (1H, s), 6.07 (1H, s); 13C NMR (CDCl3): δ = 17.8, 18.3, 18.9, 25.7, 25.9, 28.6, 29.2, 29.4, 29.5, 32.8, 63.0, 64.8, 125.1, 136.5, 141.8, 154.0, 167.6, 173.8.</p><!><p>Caenorhabditis elegans strain KG1188 lite-1(ce314) strain41 was used for fluorescence imaging. Microinjections were done using a Zeiss Axio Observer inverted microscope in age-matched adult hermaphrodites ~24 h past the L4 stage as described.42 Borosilicate glass needles were loaded with PBS or PBS solutions of 1e (1.10 mg mL−1) or 10 (0.12 mg mL−1) by capillary action and injections (~200 pL each) were performed into worm gonads. Worms were recovered in M9 buffer, mounted onto Nematode Growth Media agar chunks, and overlaid with a glass coverslip for imaging using a HC Plan-Apo 20× (0.7 NA) objective Leica SP5 laser-scanning confocal microscope, as described.43,44</p><!><p>HeLa cells were cultured in Dulbecco's modified Eagle's media supplemented with fetal bovine serum (10% v/v), penicillin (100 U mL−1), and streptomycin (0.01% v/v) at 37 °C. Cells were fixed in MeOH (0.2 mL) at −20 °C for 15 min, washed with PBS (0.1 mL) three times, and maintained in a PBS solution of bovine serum albumin (BSA, 10 mg mL−1, 0.1 mL) for 30 min at ambient temperature. Then, PBS solutions of anti-α-tubulin antibody (mouse monoclonal antibody, DM1A, Sigma; 0.8 µg mL−1, 0.1 mL) and BSA (10 mg mL−1, 0.1 mL) were added to the extracellular matrix and the resulting sample was maintained for 1 h at ambient temperature. Cells were washed with a PBS solution of BSA (10 mg mL−1, 0.1 mL) three times, incubated in a PBS solution (0.1 mL of 1% BSA) of 2c (20 µL, 2 µg mL−1) and Alexa Fluor 647-conjugated affinipure g anti-mouse IgG (H+L) (100 µL, 3 µg mL−1) for 1 h at ambient temperature, washed with a PBS solution of BSA (10 mg mL−1, 0.1 mL) a further three times, and imaged with a Leica SP5 confocal laser-scanning microscope.</p><!><p> ASSOCIATED CONTENT </p><p> Supporting Information </p><p>Synthesis of 6 and 8; Normalized absorption and emission spectra of 10; Critical concentration of 1e; Normalized absorption and emission spectra of 10 and 1e; Fluorescence decays of 10 and 1e; Absorption and emission spectra of the model Alexa-conjugated antibody.(PDF)</p><p>Video of Caenorhabditis elegans injected with blank PBS (AVI)</p><p>Video of Caenorhabditis elegans injected with 10 (AVI)</p><p>Video of Caenorhabditis elegans injected with 1e (AVI)</p><p>The authors declare no competing financial interest.</p>

PubMed Author Manuscript

A comprehensive analysis of selected medicines collected from private drug outlets of Dhaka city, Bangladesh in a simple random survey

Comprehensive data are needed to prevent substandard and falsified (SF) medicines as they pose a major risk to human health. To assess the quality of selected medicines, samples were collected from random private drug outlets of Dhaka North and South City Corporation, Bangladesh. Sample analysis included visual observation of the packaging, authenticity of the samples, legitimacy and registration verification of the manufacturer, physicochemical analysis, and price. Chemical analysis of the samples was performed using a portable Raman spectroscopy and high-performance liquid chromatography according to the pharmacopoeia. Several discrepancies were noted in the visual observation of samples. Among the 189 collected samples of esomeprazole (ESM), cefixime (CFIX), and amoxicillinclavulanic acid (CVA-AMPC), 21.2% were confirmed to be authentic, 91.3% manufacturers were confirmed legitimate, and 2.1% of all samples were unregistered. Chemical analysis of the samples revealed that 9.5% (95% CI 5.7-14.6) of samples were SFs. Falsified samples and quality variation in the same generic branded samples were both detected by Raman spectroscopic analysis. Overall, sample prices were satisfactory relative to the international reference price. This study documents the availability of poor-quality medicines, demonstrating the need for immediate attention by the national medicine regulatory authority.Assuring quality in pharmaceutical products is a major public health challenge, requiring an action plan which is capable of mitigating numerous unfavorable factors [1][2][3][4] . According to the World Health Organization (WHO), an estimated two billion people around the world do not have access to necessary medicines, vaccines, medical devices (including in vitro diagnostics), and other healthcare products, creating a vacuum that is too often filled by substandard and falsified (SF) medicines 5,6 . The WHO defines substandard medicines, also termed as out of specification, as authorized medical products that fail to meet either their quality standards or specifications, or both. In contrast, falsified medicines are those that deliberately/fraudulently misrepresent their identity, composition, or source 7 . In addition, unregistered/unlicensed medical products are those that have not undergone due evaluation and/or approval by the medicine regulatory authorities (MRAs) for the market in which they are marketed/distributed or used, but are subject to national or regional regulations and legislation 7 .This issue has been identified as an urgent health challenge for the next decade, given that more than one in ten medicines in low-and middle-income countries (LMICs) are estimated to be SF 8,9 . In recent years, chemical

a_comprehensive_analysis_of_selected_medicines_collected_from_private_drug_outlets_of_dhaka_city,_ba

Methods<!>Sample analysis.<!>Chemical analysis.<!>Sample compliance criteria.<!>Statistical analyses. Statistical analyzes was performed using IBM SPSS Statistics for Windows, version 25<!>Results<!>Results of observation. Shop observation.<!>Authenticity, legitimacy investigation, and registration verification.<!>Detection of falsified CFIX and CVA-AMPC.<!>Prices of medicines.<!>Discussion<!>Conclusion

<p>Reporting system and ethical approval. This project was a collaborative effort between Kanazawa University, Japan and University of Asia Pacific, Bangladesh represented by Professor K.K. All the methods in this study were carried out in accordance with relevant guidelines and regulations. The study was conducted and reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines, Medicine Quality Assessment Reporting Guidelines (MEDQUARG), and WHO Guidelines on the conduct of surveys of the quality of medicines [28][29][30] . The importance of good ethical practice for such studies has been discussed recently by Tabernero et al., to maintain the privacy and confidentiality of the surveyors and the surveyed; however, institutional review board approval was not required for this study at Kanazawa University as it did not involve any live vertebrates, experimental animal and/or human subject research 31 . Regulatory approval was taken in writing from the Directorate General of Drug Administration (DGDA) under the Ministry of Health & Family Welfare, Government of the People's Republic of Bangladesh on the condition that raw data sharing will be limited, and confidentiality with respect to the name of the manufacturers or their commercial brand will be maintained. The final report of this study was submitted to the DGDA in September, 2019.</p><p>Study settings and design. The entire healthcare system of Bangladesh serves an estimated population of 161.03 million, with a density of 1,077 people per square km and a population of the metropolitan area of the capital city Dhaka of approximately 17 million [32][33][34] . Target medicines were purchased from private retail drug shops (retail shops, wholesale shops selling retail medicines, and a shop in a private hospital located in Dhaka City Corporation of Dhaka District) during April, 2018, without prescriptions, using the mystery buyer approach (Supplemental Fig. S1). A list of shops was obtained from the DGDA and the outlets were classified according to area (Thanas included in the Dhaka North and South City Corporation) and included in the main database. Among the 4400 outlets in Dhaka district listed in the DGDA database, 831 outlets were excluded due to predetermined criteria (for example, Savar, Ashulia). In total, 1,885 outlets from Dhaka North City Corporation (Adabar, Badda, Banani, Gulshan, Kafrul, Khilkhet, Mirpur, Mohammadpur, Pallabi, Shaymoli, Tejgaon, and Uttara) and 1684 outlets from Dhaka South City Corporation (Dhanmondi, Jatrabari, Khilgaon, Kotowali, Motijheel, Paltan, Ramna, Rampura, Shabujbag, and Shahbag) were included in the final list of outlets. The final list of outlets was randomized using a randomization table and the samples were purchased according to the randomized list by four mystery buyer teams over 9 d, including a one-day pre-sampling to demonstrate the sample collection approaches 35 . The required sample size (188) was calculated with a 5% margin of error and 95% confidence levels assuming the sample proportion to be 15%. Target medicines were esomeprazole (ESM), cefixime (CFIX), and amoxicillin-clavulanic acid (CVA-AMPC), and they were chosen based on the frequency of use, availability, and characteristics of the medicines that are included in the essential medicine list of WHO 36 (ESM as omeprazole).</p><!><p>Collected samples were shipped within one month to the analytical laboratory at Kanazawa University for analysis. Sample analysis consisted of visual examination of the samples (packages, strips, and tablets/capsules), an investigation of the authenticity of the product by the manufacturer, verification of the legitimacy of manufacturers, registration verification of the product in Bangladesh by DGDA, Raman scattering, and pharmacopoeial analysis. Details of the packaging condition and label information were recorded carefully, according to the tabulated checklist. During observation, the packaging and labeling, physical appearance of the tablet/capsule, batch number, manufacturing date, and date of expiration, were examined according to the WHO guidelines on the conduct of surveys of the quality of medicines and the International Pharmaceutical Federation (FIP) checklist for visual inspection of medicines 37,38 .</p><p>For confirming the authenticity, a detailed questionnaire was sent to each manufacturer to confirm the authenticity of the product. Each questionnaire provided detailed information about the product, including manufacturer, batch number, date of manufacture and expiry date, dosage, and strength of the product, as recommended by the WHO 29 . Verification of the legitimacy of the manufacturers and the registration status of each product was evaluated by visual inspection of the packaging, and then by sending a questionnaire along with the package and photographs of the sample to the DGDA to confirm the legitimacy of the manufacturers and the registration of the product 39 .</p><!><p>Laboratory analysis of all the samples was carried out before the expiration date. Pharmaceutical analysis of the samples was performed according to the British and United States pharmacopoeia, with slight modifications as specified in the sample package of the respective dosage form for each of the medicines [40][41][42][43][44][45] . Some minor adjustments were made to the analytical procedure of ESM and CFX indicated in the pharmacopeia monograph. They are described in the Supplemental File S1. The pharmacopoeial quality assessment included potency (drug content), content uniformity test, and dissolution test.</p><p>The reference standards for ESM (as omeprazole), CFIX, amoxicillin, and clavulanic acid, and the internal standards (IS) lansoprazole, metronidazole, and cefadroxil, were all purchased from the United States Pharmacopeial Convention. HPLC-grade acetonitrile and methanol, tetrabutylammonium hydroxide solution ([(CH 3 CH 2 CH 2 CH 2 ) 4 N]OH), sodium di-hydrogen phosphate (NaH 2 PO 4 •2H 2 O), di-sodium hydrogen phosphate (Na 2 HPO 4 ), tri-sodium phosphate (Na 3 PO 4 •12H 2 O), potassium di-hydrogen phosphate (KH 2 PO 4 ), and other chemicals of reagent grade were purchased from Wako (Wako Pure Chemical Industries, Ltd., Osaka, Japan).</p><p>For the calibration curves, individual stock solutions were prepared by dissolving reference standards in the corresponding solvent (Supplemental File S1) at a concentration of 0.2 mg/mL. Afterwards, stock solutions were diluted to 5.0, 10.0, 20.0, 30.0, and 40.0 μg/mL aliquots to obtain five calibration samples. The concentration of all IS solutions was 20.0 μg/mL, and these were mixed with the each of the diluted reference standard solutions and sample solutions. For the assay and content uniformity sample solutions were prepared using the same solvent and diluted to the same concentration as the reference standard solutions.</p><p>The dissolution test of the samples was conducted with an NTR-VS 6P dissolution apparatus (Toyama Sangyo Co. Ltd., Osaka, Japan), and the assay was carried out by HPLC. The circular dichroism (CD) spectra of the enantiomer of omeprazole (ESM) was detected and measured using a Jasco CD-PDA detector (Chiral detector-CD 2095, and Photo diode array-MD 2018 Plus, Jasco, Tokyo, Japan) equipped with an AS-950 Jasco auto-sampler. The HPLC system for analyzing CFIX and CVA-AMPC samples consisted of a Prominence HPLC equipped with auto-sampler (SIL-10AD) and Ultraviolet-Photo Diode Array Detector (UV-PDA, SPD-20A/20AV Series; Shimadzu, Kyoto, Japan).</p><p>Mechanical calibration and performance verification tests were performed before sample testing for performance qualification to ensure the absence of technical and mechanical errors. Test methods and system suitability for each medicine were validated according to USP 41 46 . The five-point calibration curves were prepared with three replicates of injections for each vial of the calibrators, and three replicates were used during sample analysis. Calculations for quantitation were based on the peak area ratios of the analyte relative to its corresponding IS using weighted (1/x) regression. The linearity of the curves was assessed by linear calibration using the correlation coefficient (r). The r was determined using the mean of three replicates at each level of the calibrators. Details of the analytical condition, dissolution test, chromatographic condition, and the reference for compliance criteria are summarized and presented in Supplemental Table S1.</p><!><p>In the potency test (quantity), ESM samples were evaluated as meeting acceptance criteria if the amount of API lay within the range of 90.0-110.0% of the label claim. For CFIX samples, the tolerance range was 90.0-110.0% of the label claim. For CVA-AMPC samples, the range was 90.0- where, x is the mean of an individual content expressed as a percentage of label claim; M is the reference value, K is acceptability constant, and S is the sample standard deviation.</p><p>In the dissolution test, Q values for evaluation were as follows: ESM in acid, not more than (NMT) 10% of the label claim; ESM in buffer, not less than (NLT) Q = 75% of the label claim; CFIX, Q = 75% of the label claim; amoxicillin, = 85% of the label claim; and clavulanic acid, Q = 80% of the label claim.</p><p>Raman scattering analysis of the samples. Raman scattering analysis was performed to analyze the molecular structure by light scattering using a portable Raman scattering analyzer (Inspector 500, SciAps Inc., Laramie, WY, USA). The instrument was equipped with higher wavelength Raman excitation, consisting of a 300 mW power source with a 1030 nm wavelength Class III B laser and a cooled Type III-IV semiconductor detector array (spectral range 100-2500 cm −1 ). The exposure time was set at the default (maximum 8.0 s). Each of the tested samples was analyzed for five consecutive spectral data on the front, back, and side, thus generating 45 spectral data. The average of these 45 spectral data was then calculated and analyzed. Tablet samples were taken out of the blister and kept directly in front of the laser source. For capsule samples, granules were separated from the gelatin shell and kept in a thin and transparent glass tube. The glass tube was positioned in front of the laser source (three times each), and the Raman spectral data were recorded and compared with those of the other samples of the same brand or the authentic samples of the same brand. The concordance rate (match score) was calculated from the Pearson's correlation coefficient between the test sample spectrum and the reference sample spectrum on a common interpolated wavenumber scale using the NuSpec Pro software (SciAps Inc., Laramie, WY, USA), and the Raman spectral data were input into the Unscrambler (CAMO Software, Oslo, Norway) for principal component analysis (PCA). The PCA model was constructed for three sets of samples, wherein each set represents five points, averaged from the 45 spectral data obtained from the different regions of each sample. Spectral data of the reference samples were used for the PCA model as the calibration set. The optimal number of principal components was determined from the internal cross-validation where authentic samples were treated as the reference set 47 . Spectral pre-processing involved the application of Savitzky-Golay smoothing and differentiation filter (second-degree polynomial and first derivative) to remove noise and baseline signals. We then performed unit-area normalization by applying Standard Normal Variate to the smoothed and differentiated signals [48][49][50] .</p><p>Price. The prices of samples were recorded in local currency (BDT) and converted from local currency to US Dollar (USD), based on the exchange rate given by the money exchange office in Dhaka on April 09, 2018 (US$1 = Tk 83.5). Prices for the different strengths of medicine were calculated individually for each of the samples and expressed as the median unit price of an individual medicine; the observed individual prices were compared to an international reference price as recommended by the WHO/HAI manual 51 . The Median Price Ratio (MPR) was calculated by dividing the median unit price of an individual medicine by the median supplier prices from the Management Sciences for Health (MSH) 2015 52 .</p><!><p>(IBM Corp., Armonk, NY, USA). The Fisher exact test was performed online (Extended-http:// aoki2. si. gunmau. ac. jp/ exact/ fisher/ getpar. html). A significance probability of 1% was used for these analyses. Confidence intervals were calculated using descriptive statistical analysis. The criterion of significance was taken as p < 0.05. Means, standard deviations, and coefficient of variation (CV%) were calculated using Microsoft Excel 2016. Graphs for figures were generated using GraphPad Prism (version 9.0, GraphPad Software Inc., San Diego, CA).</p><!><p>Sample description. A total of 189 samples were collected in this study. Samples were purchased by selecting at least one sample of either ESM, CFIX, or CVA-AMPC from the 210 outlets visited. From the list, 28 (13.3%) outlets had to be excluded as the shops were either non-existing or found closed at the time of sampling. Additionally, two ESM samples, two CFIX samples, and three CVA-AMPC samples were provided by the manufacturers upon request. All the collected samples were found to be domestically manufactured; 100 samples (54.9%) were collected from Dhaka North City Corporation area and the rest (45.1%) were collected from Dhaka South City Corporation. Samples from retail shops accounted for 78.6% of the total samples collected, whereas 20.9% samples were collected from wholesale shops selling retail medicine. Only one CFIX sample was collected from a shop inside a hospital facility. No provider asked to see a medical prescription from the buyers during the purchase of samples. Further details of the collected samples are summarized in Table 1.</p><!><p>Samples were collected only from shops listed by the DGDA (no samples were collected from illegal shops). During sample collection, details regarding the storage conditions of the medicines were recorded. Among the visited shops, only 13 (7.2%) shops were found to be equipped with air-conditioning, although four of them were not functional; the remaining shops (92.8%) did not have air-conditioning. However, there was no significant temperature difference (p = 0.117) between shops with airconditioning (30.1 ± 1.5 °C) and without air-conditioning (31.7 ± 2.5 °C). Although a significant humidity (rela- www.nature.com/scientificreports/ tive humidity-RH) difference was observed (p < 0.05) between shops with air-conditioning (52.2% RH ± 8.7) and without air conditioning (59.6% RH ± 7.9).</p><p>Sample observation. Samples were checked for 46 points according to the FIP Tool for Visual Inspection of Medicines 38 . Overall, packaging analysis of the samples was satisfactory. One ESM sample had two different batch numbered strips from the same secondary packaging box (D-106). Two samples (A-102, B-213) did not have an insert inside the box, one of which (B-213) was confirmed to be falsified after chemical analysis. One CVA-AMPC sample (C-308) was found with a spelling error in the insert regarding the strength of the product (Supplemental Fig. S2). Although CVA-AMPC is usually film coated, no film coating was observed on the tablets from this sample when opened during analysis. The tablet surfaces were also dirty with speckles (Supplemental Fig. S2). The sample was also suspected to be falsified after chemical analysis.</p><!><p>The names and addresses of manufacturers and wholesalers were identified through the product packages and manufacturer websites. Questionnaires were sent to the respective manufacturers of ESM, CFIX, and CVA-AMPC by e-mail in July 2018, and 5 manufacturers out of 33 gave their response by 04 September 2018. Forty samples (21.2%) out of 189 were confirmed as genuine products by four manufacturers, and no falsified samples were reported (Table 2). Most manufacturers have a contact e-mail address, but in many cases, we did not obtain a response, even after sending a reminder mail to non-responders. The results of our authenticity investigation are summarized in Table 2.</p><p>Although questionnaires were sent to the DGDA by e-mail in July 2018, we initially received no response. According to our first stage of registration verification, fifty-one out of 62 ESM samples (85%) were verified for registration from the DGDA drug product registration list in their website. Eight (13.3%) samples could not be verified, as the registration numbers in the package were different to the DGDA registration number. While one sample (B-114) was registered for tablet dosage form, we collected capsules from this manufacturer with the same registration number. In case of CFIX samples, n = 47 samples (78%) could not be verified due to differences in the registration number on the package and in the DGDA list. Moreover, one manufacturer was not listed as a registered manufacturer of CFIX raising legitimacy concerns. For CVA-AMPC, 52 out of 60 samples (86.7%) from three manufacturers had different registration numbers from those in the DGDA list and therefore could not be verified initially. At the beginning of 2020, DGDA crosschecked the registration status of the collected samples and confirmed that some newly changed registration numbers had not been updated on their site. However, two manufacturers were identified who were not licensed to produce CFIX. Another manufacturer was only licensed to produce a tablet dosage form of CFIX, although a capsule dosage form from this manufacturer was available on the market. The results of our registration verification of the samples and manufacturers are summarized in Table 2.</p><p>Results of chemical analysis. Among all the tested samples, 18 samples (0.095; 95% CI 5.3-13.7) were found to be non-compliant after the final stage of at least one of the pharmacopoeial tests. Among the non-compliant samples, two samples were identified as falsified medicine (CFIX, 1; CVA-AMPC, 1). Raman spectroscopy and HPLC analysis of the falsified samples confirmed the absence of any APIs. The remaining 171 samples (90.5; 95% CI 86.3-94.7) were compliant with all pharmacopoeial tests according to the declared pharmacopoeia, i.e., USP or BP. The results of our chemical analyses are summarized in Table 3.</p><p>The average quantity of all samples compliant in the potency test was 100.1% ± 4.7 (95% CI 99.4-106.0) with a minimum mean of 90.1% API and a maximum mean of 111.8% API. The frequency distribution chart of all samples in the potency test is shown in Fig. 1a. In contrast, the average quantity of all the non-compliant samples in the potency test was 78.8% ± 7.1 (95% CI 73.7-83.8). Among the non-compliant samples, we identified samples with a minimum mean potency of 68.5% and a maximum mean potency of 86.2%.</p><p>High inter-unit variability was observed for nine (4. 8; 95% CI 1.7-7.8) of the total collected samples. Notably, among the non-compliant samples in the content uniformity test, none were from the ESM or amoxicillin samples. www.nature.com/scientificreports/</p><p>In the dissolution test, the average percent release of API in the dissolution medium within the specified time was 93.3% with ± 9.5 (95% CI 92.1-94.5) for all the samples meeting pharmacopoeia acceptance criteria. The lowest amount of API released in the medium after the pharmacopoeia specified time was 71.2% and the maximum was 112.5%. The frequency distribution chart of all samples in the dissolution test is shown in Fig. 1b. Conversely, the mean percent release of API in the samples failing to meet the pharmacopoeia specified criteria in the dissolution test was 70.3% ± 10.0 (95% CI 63.6-76.9). The mean percent release in those samples was observed to be 55.7%, with a maximum of 93.4%. Surprisingly, all of the ESM tablets and granules kept their integrity Table 2. Authenticity, legitimacy investigation, and registration verification of the samples and manufacturers. *Two esomeprazole samples were provided by the manufacturers upon request; **Two CFIX samples were provided by the manufacturers upon request; ***Three amoxicillin-clavulanic acid samples were provided by the manufacturers upon request; § One manufacturer of esomeprazole did not have manufacturing or marketing authorization to produce and sell esomeprazole capsule; ¶ One manufacturer was not listed as a registered manufacturer of CFIX and manufacturing license was cancelled for another manufacturer by DGDA but their samples were circulating in the market and one of their sample was identified as falsified; † Combining all samples total number of manufacturers were n = 33; a One esmeprazole sample was provided by AstraZeneca which was used a standard esomeprazole sample. www.nature.com/scientificreports/ during the acid stage of the dissolution test. Nevertheless, the dissolution rates of some samples were low in the buffer stage of the dissolution test, releasing less ESM than the pharmacopoeia specified range (Fig. 1c). Among the non-compliant samples, one ESM sample (B-103) demonstrated extreme failure, as two units showed as low as 1.3% release of API in the medium in the second stage of the dissolution test. Mean Quantity of the API of non-compliant samples versus their mean dissolution rate in the dissolution test are presented in Fig. 1c-f. As demonstrated in Fig. 1e and f, the potency and dissolution rate of AMPC in the CVA-AMPC samples were within the compliance range. Nevertheless, they were non-compliant due to the lower content of CVA in the co-amoxiclav samples. In spite of having lower potency than the declared amount, a comparatively higher dissolution rate was observed for CVA (Fig. 1f).</p><p>Raman scattering analysis coupled with match score and PCA. Raman spectroscopy was employed to explore the distinctive behavior between compliant and non-compliant samples, principally by match score and PCA. Raman spectra were analyzed after preprocessing and were subjected to PCA to investigate the similarity of chemical components between and/or among the samples. In most of the cases, non-compliant samples were compared against the authentic samples, if available. Otherwise, collected compliant samples from the same manufacturer were chosen for comparison. In the case of ESM samples, some manufacturers' samples showed variances in chemical analysis, where some samples of the same brand were compliant while others were not. The results were further confirmed by Raman analyses and their PCA (Fig. 2). Compliant and noncompliant representative samples from those manufacturers showed a variation in the match score and a wide distribution on the score plot, suggesting significant differences among samples even though they were of the same generic brand. The correlation and variance among three different samples from the same manufacturer were analyzed using PCA on the basis of their compliance and non-compliance in the pharmacopeial analysis (Fig. 2a-d). Although a more number of representative samples could provide more precise evidence, a negative correlation was observed in PC1 between the compliant and non-compliant samples (Fig. 2b,d). In PC2, a positive correlation was observed between two non-compliant samples. As shown in Fig. 2e and f, another ESM sample (PS-006) showed variation in the strips and batch number in the same packaging, as discussed in the observational analysis results. Granules from both the strips showed a very poor dissolution rate (59.2% and 59.0%) in the dissolution medium. Notably, another ESM sample (D-106) manufactured by the same manufacturer was found compliant. To investigate whether the observed differences between the reference and test samples were not just because of the batch-to-batch variation, the dissolution profiles of the two reference samples (authentic samples obtained from the manufacturers) were analyzed together. The reference samples almost overlapped on the different PCs, indicating that the between-batch variation was negligible.</p><p>Inconsistencies in the quality of the same generic brand samples from the same manufacturer were also observed in samples of CFIX and CVA-AMPC. In case of CVA-AMPC, though an authentic sample from a manufacturer failed to comply with the acceptance criteria, two samples out of four collected samples from this manufacturer were found to be compliant. Two authentic samples were separately provided by this manufacturer upon request. There is a possibility that any discrepancies in the physicochemical analyses between these samples may have occurred as a result of degradation of clavulanic acid resulting from the ineffective film coating (Sample B-315 in Fig. 3a and b; Fig. 3 in Supplemental File S2). Since the tablets in the samples from this manufacturer were packaged in strip, we assume that the chipping was caused by mechanical stress, and that it resulted in the permeation of moisture and consequent degradation of clavulanic acid [53][54][55] .</p><!><p>There was a reasonable level of agreement between the HPLC analyses and Raman spectroscopic analyses. In both the chromatographic and spectroscopic analyses, one CFIX and one CVA-AMPC sample appeared to be falsified (Figs. 4 and 5). The packaging labels of these samples stated that one contained CFIX and one contained CVA-AMPC, although no API was detected in either the assay or the spectroscopic analysis of these samples (see Sample B-213 in Figs. 4b and 5a; and Sample C-308 in Figs. 4d and 5b; Fig. 2 in Supplemental Fig. S2).</p><p>In case of the CVA-AMPC sample, the packaging was almost identical to the compliant samples from the same manufacturer. However, close observation of the packaging and insert showed a mistake in the insert (Fig. 2 in Supplemental File S2). For the CFIX sample, no comparator was available. To quantify the API, ten units of suspected CFIX sample and five units of suspected CVA-AMPC sample were assayed by HPLC, respectively. No chromatogram peak was observed with either of the samples (as described in Figs. 4 and 5). Additionally, 6 units of CVA-AMPC were used for the dissolution test that showed no disintegration of the drug in the medium (remained intact). Out of 20 collected CVA-AMPC samples that were manufactured by this manufacturer, one sample was found to be falsified. Presumably, the falsified sample had successfully mimicked the original. Further analysis of these samples was performed by Raman Spectroscopy, and the resulting spectra were markedly different from those of the reference standards and the other compliant samples (Figs. 3c and 5).</p><!><p>The discrepancy in price for all medicines was summarized by the minimum, 25th percentile, median, and maximum price values relative to the International Reference Price (IRP) 47 . The median unit price of a 20 mg ESM capsule was 0.084 ± 0.01 USD (minimum, 0.06 USD/unit; maximum, 0.102 USD/ unit). The tablet formulation of 20 mg ESM had a median price of 0.06 ± 0.01/unit (minimum, 0.056; maximum, 0.078/unit). The price of a 20 mg ESM capsule was slightly higher than that of the 20 mg tablet, but not significantly. The Management Sciences for Health (MSH) price comparator was not available for ESM, although it's MPR was significantly higher (p < 0.05) than that of omeprazole. Compliant ESM samples had a slightly higher price than the non-compliant ESM samples, although the difference was not significant (Fig. 6a,b). The median price of 200 mg CFIX was 0.419 ± 0.04 USD (minimum, 0.314 USD/unit; maximum, 0.603 USD/ unit) (Fig. 6a,c). The CFIX price was significantly higher than the MSH median price (t-test, p < 0.001). However, the MPR for CFIX was found to be 2.54. An MPR of 1 or less is commonly interpreted as efficient procurement in the public sector, while an MPR below 3 is considered acceptable for the private sector 56 . All of our samples were procured from private retail shops (the authentic samples were not included in the price calculations). One sample of CFIX was 400 mg and the price was 0.60 USD. The minimum price paid for 625 mg CVA-AMPC was 0.24/unit (maximum, 0.395/unit; median 0.383/unit), which was significantly higher than the MSH price (t-test, www.nature.com/scientificreports/ p < 0.001). For 625 mg CVA-AMPC, the MPR was found to be 2.35. No significant differences were observed between compliant and non-compliant samples of either CFIX or CVA-AMPC (Fig. 6a,d).</p><!><p>Recent research has revealed that an alarming percentage of analyzed medical products in LMICs were found to be unacceptable in terms of quality, with the additional concern that SF antibiotic could lead to drug resistance due under-dosing or no dosing at all 9,57,58 . It is important that tragic events like that reported in Bangladesh in 1995 are not repeated, especially in the Covid-19 pandemic era 21,59 . Despite achieving remarkable health and pharmaceutical improvements since gaining independence in 1971, adequate reporting of the quality of medicines using tested methodological practices has not been accomplished in Bangladesh 60 . This is the first comprehensive analytical study documenting the quality of selected medicines in private drug outlets in Bangladesh. A summary of the study has been presented in Fig. 7 as a flow chart. The present study reports the analytical results from medicines for a non-communicable disease and two antibiotics, including one combination drug, as representative medicines. Overall, the quality of the majority of the samples analyzed was revealed to be good; more than 90% of the analyzed samples collected from the Dhaka City Corporation region complied with the pharmacopoeial reference ranges and were found to be of acceptable quality (Table 3). The results of this study are significant, when compared with the results of summarized studies conducted in other LMICs 9,56 . The WHO previously reported a failure of 19.5% (95% CI 18.8-20.3) of samples following random sampling, whereas Ozawa et al. reported that the prevalence of SF medicines ranged from 18.7% (95% CI 12.9-24.5) in Africa to 13.7% (95% CI 8.2-19.1) in Asia 9,57 . In contrast, the prevalence www.nature.com/scientificreports/ of substandard medicines in our study was found to be 8.5% (95% CI 4.5-12.4), and this was higher than the prevalence of falsified medicines 1.1% (95% CI − 0.4-2.5). In total, these results indicate a prevalence of 9.5% for SF medicines, lower than the assumed prevalence of 15% (p < 0.05; 95% CI 5.7-14.6). While most of the samples were of good quality, 4.2% of samples (95% CI 1.4-7.1) contain less API than the stated amount in their packaging (Table 3). Another important observation was the uniformity of the content, with 4.9% of samples demonstrating high inter-sample variation; these were found to be non-compliant according to the 'acceptance value (AV)' set by the pharmacopoeia. The permissible range of AV values set by the pharmacopoeia was established to ensure that all individual units within a sample have a sufficient amount of the declared API. Indeed, it is essential that the patient receives a dosage close to that claimed in the label. However, in the cases reported above, it cannot be stated with certainty that every unit of the sample contained the exact same amount of API.</p><p>The drug dissolution test is a good predicator for the sample performance test, while allowing observation of physicochemical changes in tablets/granules in capsules [61][62][63] . The results presented in this study demonstrate that 5.3% samples (95% CI 2.1-8.6) were non-compliant (including all tested samples but excluding the falsified samples). The poor dissolution rates of the samples were a persistent problem, an observation which had previously been reported in our studies 60,61,64,65 . As an essential indicator of bioavailability and drug quality, the dissolution test has not been given as much attention as it deserves 62,66,67 . Despite the fact that all of the ESM samples contained the stated amount of API, a poor dissolution rate was observed in the buffer stage of the dissolution test in 10.9% of the ESM samples (Table 3; Figs. 1 and 4). HPLC analysis detected one inconsistent ESM sample, www.nature.com/scientificreports/ and this showed only a 1.3% dissolution rate of units in the second stage of the dissolution test. Unfortunately, a genuine comparator was not available for this sample to investigate further. The reliability of the findings was further supported by the Raman spectroscopy results, which also confirmed the detection of two falsified medicines. The status of suspect samples identified during a visual observation using the FIP checklist was confirmed by the chemical analysis result (for example, sample D-106 appeared to be substandard and B-213 appeared to be falsified). However, packaging analysis alone could not detect additional falsified CVA/AMPA medicines (Figs. 4 and 5; Supplemental Fig. S2). Identical packaging from falsified samples could not be distinguished from that of the other samples stated to be manufactured by the same manufacturer, even after close observation of the packaging. In our study, handheld Raman spectroscopy was a practical and useful tool for the screening of suspicious samples, facilitating both identification of and characterization of suspicious samples (Figs. 2, 3, and 5). Samples showing different chromatographic analysis result but similar spectra could be differentiated using principal component analysis (PCA) of the Raman spectra (Figs. 2 and 3). One falsified sample of CFIX and one falsified sample of CVA-AMPC were initially identified using the handheld Raman spectroscopy device, demonstrating its usefulness for the detection of falsified samples. Hence, this comparatively low-cost device may prove a useful tool in field settings requiring quality analysis. www.nature.com/scientificreports/ Despite the satisfactory results obtained in our chemical analyses, falsified medicines threaten the national supply and distribution chain (Figs. 3, 4, 5). No one test can serve as an ultimate tool for the detection of SF medicines, and the roles of MRAs and manufacturers remain in question, as the anticipated responses from the stakeholders regarding product authentication and legitimacy verification were below the expected level. In our authenticity investigation, a response was received from only five manufacturers, confirming the authenticity of 21.2% samples (Table 2). Better cooperation from the manufacturers in response to questionnaires is essential for any authenticity investigation. Furthermore, considering the export potential of pharmaceuticals, which is largely driven by domestic manufacturers, the governing authorities should focus on the active and effective regulatory surveillance of manufacturers.</p><p>Data from our study suggests that there is substantial scope for improving the storage situation of the distributed medicines, and for lowering the prices of the medicines in the private drug outlets. The observed sub-standard storage conditions for the medicines may be linked to degradation of the medicines, although no relationship could be established between storage and quality of medicines in our study 68 . As reported above, the prices for the collected medicines were slightly elevated relative to the international standards (Fig. 6), although the prices were not excessive. This suggests that the revised health policy intervention may be necessary to reduce the catastrophic out-of-pocket expenditure of patients, especially during long-term treatment.</p><p>The strength of this study is that it presents a comprehensive analysis of the collected samples including all stage of pharmacopoeial analysis for potency, a dissolution test using HPLC, and Raman spectroscopy combined with chemometrics. The sample size for this study was not large enough to allow prediction of the influencing factors and a comparative analysis. However, given the limited personnel resources, a larger sample size would limit the scope for such a detailed analysis. Indeed, longer processing times would make it difficult to complete the analysis before the expiry date of the collected samples. Finally, the study is not nationally representative and should be interpreted with caution, as samples were only collected from one urban area of the Capital City and illegal or unlicensed shops were excluded.</p><!><p>The threat of SF medicines exists in Dhaka City Corporation, although the proportion of SFs was revealed to be relatively lower than the estimated proportion. In addition, differences in the quality of the same branded sample from the same manufacturer may put the patient at risk. Therefore, the national MRA, in an active collaboration with the manufacturers, should identify hotspots for these life-threatening poor-quality medicines, and take prompt and effective action. In our study, a consensus was observed between the portable Raman spectroscopy results and the observational and authenticity analysis results thus, implying that scope for these screening techniques should be considered in the field detection and evaluation of medicines. A full-scale analysis, including a dissolution test, is essential to accurately estimate the prevalence of substandard and falsified medicines.</p>

Scientific Reports - Nature

A Ketogenic Diet Differentially Affects Neuron And Astrocyte Transcription

Ketogenic diets (KDs) alter brain metabolism. Multiple mechanisms may account for their effects, and different brain regions may variably respond. Here, we considered how a KD affects brain neuron and astrocyte transcription. We placed male C57Bl6/N mice on either a 3-month KD or chow diet, generated enriched neuron and astrocyte fractions, and used RNA-Seq to assess transcription. Neurons from KD-treated mice generally showed transcriptional pathway activation while their astrocytes showed a mix of transcriptional pathway suppression and activation. The KD especially affected pathways implicated in mitochondrial and endoplasmic reticulum function, insulin signaling, and inflammation. An unbiased analysis of KD-associated expression changes strongly implicated transcriptional pathways altered in AD, which prompted us to explore in more detail the potential molecular relevance of a KD to AD. Our results indicate a KD differently affects neurons and astrocytes, and provide unbiased evidence that KD-induced brain effects are potentially relevant to neurodegenerative diseases such as AD.

a_ketogenic_diet_differentially_affects_neuron_and_astrocyte_transcription

Introduction<!>KD Intervention in C57Bl6/N Mice<!>Cell Type Enrichment and RNA Isolation<!>Fluorescence Activated Cell Sorting (FACS) Cell Enrichment Analysis<!>Generation of Total Stranded RNA Library and Performance of RNA-Seq<!>RNA-Seq Data Quality Assessment and Preprocessing<!>Statistical Analysis<!>Pathways Analysis<!>Blinding<!>Figure Generation<!>Validation of Cell Population Enrichment<!>Diet Macronutrient Profiles and Physiologic Effects<!>RNA Sample Preparation and Quality Assessments<!>Changes in Individual Gene Expression<!>KEGG Molecular Pathway Analysis<!>GSEA of Molecular Pathways<!>IPA of Gene Expression<!>KEGG Pathology Analysis<!>IPA of Disease States<!>Hypothesis-Driven Evaluation of AD-Associated Transcriptional Changes<!>Discussion

<p>The ketone bodies D-β-hydroxybutyrate and acetoacetate supply and support brain energy-generating carbon fluxes (Koppel & Swerdlow 2018). In starvation states they supplement and spare brain glucose catabolism. By activating liver ketone body production ketogenic diets (KDs) increase circulating levels of D-β-hydroxybutyrate and acetoacetate, which freely cross the blood brain barrier and access neurons and astrocytes via monocarboxylate transporters (MCTs) 2 and 1 respectively. (Owen et al. 1967; Vijay & Morris 2014). In ketolytic tissues such as neurons ketone bodies are converted to acetyl-CoA, which accesses the mitochondrial TCA cycle and generates reducing equivalents that enter the electron transport chain (Fukao et al. 1997).</p><p>KDs reduce seizure frequency in persons with epilepsy (Kossoff et al. 2009), which suggests manipulating brain energy metabolism in some settings confers neurologic benefits. Different mechanisms may contribute as this intervention can support bioenergetics, enhance ROS scavenging, modulate neurotransmitters, induce post-translational protein modifications, increase neurotrophin signaling, and activate signaling receptors (Marosi et al. 2016; Rahman et al. 2014; Kimura et al. 2011; Shimazu et al. 2013; Xie et al. 2016; Rardin et al. 2013; Sullivan et al. 2004; Erecinska et al. 1996). Some advocate a potential role for KDs in treating neurodegenerative diseases that feature defective brain bioenergetics, including Alzheimer's disease (AD), in which mitochondrial dysfunction and glucose hypometabolism can precede cognitive decline (Taylor et al. 2017; Taylor et al. 2019; Cunnane et al. 2016a; Cunnane et al. 2016b; Swerdlow et al. 1989).</p><p>Some of the above-noted mechanisms may not depend on the presence of ketone bodies, although overall ketone bodies presumably play an important role. A KD may also uniquely affect distinct parts of the brain. Here, we considered whether neurons and astrocytes could respond differently to a KD. To address this, we placed C57Bl6/N male mice on either a 3-month KD or chow diet and generated enriched neuron and astrocyte fractions. We then used RNA-Seq KEGG pathway analysis, gene set enrichment analysis (GSEA), Ingenuity Pathway Analysis (IPA), and an AD-pertinent, hypothesis-based approach to analyze KD-induced transcriptome changes.</p><!><p>The University of Kansas Medical Center Institutional Animal Care and Use Committee approved all experimental animal protocols and procedures (Animal Care and Use Protocol #2020-2558). This study was not pre-registered. This study was exploratory and was designed for a 90-day dietary intervention period. Power calculations were performed using the PROPER method to estimate an appropriate number of mice (Wu et al. 2014). For power calculations we assumed a sampling of approximately 20,000 genes, primary comparison of diet effect within cell-type (comparison of 2 groups), and 5% of genes being differentially expressed with a magnitude of effect defined from a normal distribution with standard deviation equal to 1 on a log2 scale. From this we determined that groups of 8, 10, and 12 would provide marginal power of 0.72, 0.75, and 0.77 respectively with a nominal false discovery rate of 0.1.</p><p>Sixteen -week old C57Bl6/N male mice (Charles River Laboratories, RRID: MGI: 5656552) were acclimated for one week before obtaining baseline weight, blood glucose, and blood ketone levels. Mice were group-housed to a maximum of 5 littermates per cage in Touch Slim Line IVC cages (Techniplast). Mice were separated on an individual basis upon recommendation of the institutional veterinary staff for excessive fighting and fight wounds. Singly housed mice were provided additional domed plastic housing for environmental enrichment per institutional standard operating procedures. Blood was obtained by facial vein phlebotomy and metabolites measured using a Precision Xtra meter (Abbott Diabetes Care, Inc. cat no. #9881465) with blood glucose (Abbott cat no. #9972865) and ketone test strips (Abbott cat no. #7074565). Mice were distributed into dietary treatment groups via simple randomization stratified by weight to obtain equivalent mean initial body weights prior to initiating the KD.</p><p>Mice were maintained on either a standard chow (LabDiet cat no. #5053) or ketogenic diet (Bioserve cat no. #F3666) for 90 days with ad libitum access to food and water. Mouse weight and blood parameters were monitored periodically throughout the dietary intervention. Food intake was measured by weighing food prior to and following a 24-hour feeding cycle for three days with a mean intake value per mouse (g) being determined and multiplied by the associated diet kcal/g to determine the daily consumed kcal.</p><p>We began the study by utilizing ten mice for protocol development. For the dietary intervention initial cohorts of 15 mice per group were generated, to facilitate a final comparison of 12 mice per group at study termination. In total, 40 mice contributed to the present study.</p><!><p>Figure S1 illustrates the overall experimental timeline and cell enrichment procedure. At the end of the 90-day intervention mice were sacrificed for brain cell separation and RNA isolation. Euthanasia began in the morning with sample isolation completed by early afternoon. Mice that completed the diet interventions were euthanized by rapid decapitation, using an inspected and certified mouse guillotine and brains were removed from the skull and placed in digestion buffer within 30 seconds of decapitation. For these mice anesthesia and CO2 asphyxiation were not utilized to avoid disruptions to brain metabolism prior to euthanasia (Mena et al. 2010; Richerson 1995; Dean et al. 2003). Animals that were euthanized prior to study completion for humane reasons received primary CO2 asphyxiation in home cages followed by secondary decapitation.</p><p>The digestion buffer contained 2 parts Neurobasal medium (Gibco cat no. #21103-049) to 1-part Accutase (Gibco cat no. #A11105-01) supplemented with DNAse I (Sigma cat no. #DN25) to a final concentration of 0.3 mg/mL. Brains were incubated in digestion buffer at 37°C for 20 minutes. Brains were supplemented with an additional 10 mL of neurobasal medium and then triturated via 10 passages through 10 mL serological pipettes, followed by 10 passages through 5 mL serological pipettes, and a final 20 passages through 1,000 μL pipette tips in a 50 mL conical tube. We removed large debris by letting it settle to the bottom of the tube. The medium layer was collected and passed through a 70 μm cell strainer (Fisher cat no. #22-363-549) pre-hydrated with 5 mL neurobasal medium. The pass-through was collected and strained through a 40 μm cell strainer (Fisher cat no. #22-363-547) pre-hydrated with 5 mL neurobasal medium.</p><p>The final pass-through was collected and centrifuged at 300 rcf for 10 minutes at 4°C to pellet cells. The supernatant was discarded. Cells were resuspended in 1.8 mL of ice-cold cell stain buffer composed of low-fluorescence Hibernate A (BrainBits, LLC), 0.5% bovine serum albumin (Boston Bioproducts cat no. #P-753), and SuperaseIn RNAse inhibitor (ThermoFisher cat no. #AM2696) and sterile filtered. 200 μL of myelin removal beads (Miltenyi Biotec cat no. #130-096-433) were added and cells were incubated for 15 minutes at 4°C on a rocking shaker. 10 mL of fresh ice-cold cell stain buffer were added per sample and cells were again centrifuged. The supernatant was discarded, and the cell pellet resuspended in 4 mL of cell stain buffer. The cell suspension was evenly distributed by 1 mL volumes into clean, autoclaved microcentrifuge tubes. Tubes were placed into a magnetic tube rack (ThermoFisher cat no. #12321D) at 4°C for 10 minutes. The unbound supernatant was carefully removed through pipetting and transferred to a new, clean 15 mL conical tube. Collections were centrifuged under prior conditions and the supernatant discarded. Cell pellets were resuspended in 720 μL of cell stain buffer and evenly distributed into two clean microcentrifuge tubes for cell-type specific enrichment protocols.</p><p>For neuron enrichment, 80 μL of biotinylated antibody cocktail (Miltenyi Biotec cat no. #130-115-389) targeting non-neuronal cell types was added to the cell suspension per manufacturer instructions. Samples were incubated at 4°C for 10 minutes on a rotating tube rack (Fisher cat no. #05-450-127). Samples were centrifuged at 300 g for 10 minutes at 4°C and the supernatant carefully removed. Pellets were resuspended in 360 μL of cell stain buffer and 80 μL of anti-biotin magnetic beads added. Cells were again incubated at 4°C for 10 minutes on a rotating tube rack. Samples were diluted with 1 mL of ice-cold cell stain buffer and placed in the magnetic stand for 10 minutes at 4°C. Unbound supernatant containing neuronal cells was collected so as not to disturb the sedimented layer along the magnet containing glia and endothelium and transferred to a clean microcentrifuge tube. Collections were centrifuged as described above and the supernatants discarded. Cell pellets were dissolved in 1 mL of TRI reagent (ThermoFisher cat no. #15596018) for ten minutes before proceeding to RNA isolation. RNA was isolated via phenol-chloroform extraction with TRI reagent, with RNA purity and content measured by A260/280 ratio.</p><p>For astrocyte enrichment, 80 μL of biotinylated antibody targeting glutamate aspartate transporter 1 (GLAST) (Miltenyi Biotec cat no. #130-095-826, RRID: AB_2733472) was added to the cell suspension per manufacturer instructions. Samples were incubated at 4°C for 10 minutes on a rotating tube rack. Samples were centrifuged at 300 g for 10 minutes at 4°C and the supernatant was removed. Pellets were resuspended in 360 μL of cell stain buffer and 80 μL of anti-biotin magnetic beads were added. Cells were next incubated at 4°C for 15 minutes on a rotating tube rack. Samples were centrifuged as described previously and the supernatant removed. Pellets were resuspended in 1 mL of ice-cold cell stain buffer and placed in the magnetic stand for 10 minutes at 4°C. Unbound supernatant containing non-astrocytic cells was removed so as not to disturb the sedimented layer along the magnet containing astrocytes. The sedimented magnetized cells were washed in 1 mL of cell stain buffer and centrifuged to a pellet. The supernatant was removed, and pellets were dissolved in 1 mL of TRI reagent and allowed to stand for ten minutes before proceeding to RNA isolation. RNA was isolated via phenol-chloroform extraction with TRI reagent, with RNA purity and content measured by A260/280 ratio. The procedure time from sacrifice to TRI reagent was just under 3 hours.</p><!><p>To facilitate a FACS analysis of cell enrichment, we stained an initial set of cell suspensions with fluorescent antibodies directed to different brain cell populations. At different stages of the cell enrichment protocols, cell counts were ascertained and adjusted to 1 x 106 cells/mL with cell stain buffer and aliquoted into clean microcentrifuge tubes for the quantitative determination of negative, single channel positive, and combined positive labeling. The stains/antibodies were as follows: GLAST-PE antibody for astrocytes (Miltenyi Biotec cat no. #130-118-344, RRID: AB_2733722), CD11b-PE/Cy7 antibody (Miltenyi Biotec cat no. #101216) for microglia, O4-APC (Miltenyi Biotec cat no. #130-099-211, RRID: AB_2751644) for oligodendrocyte precursors, and DAPI (1 μg/mL) for live/dead discrimination. Live/dead discrimination was also performed using Ghost Dye 450 (TONBO Biosciences cat no. #13-0863-T100). Verification of nucleated events was performed using Vybrant DyeCycle Green stain (ThermoFisher cat no. #V35004). All antibodies were added at 1 μL per 1 X 106 cells.</p><p>Staining proceeded for 30 minutes on ice in the dark. Samples were then centrifuged as before and washed 3x in cell stain buffer. Following the final spin cells were resuspended in 500 μL and analyzed by flow cytometry. Samples were processed and analyzed by FACS on a BD FACSAria II cytometer (BD Biosciences, San Jose, CA RRID: SCR_018091). Cells were initially gated by forward scatter complexity (FSC) and side scatter complexity (SSC) to identify debris and multicellular events. Events that were DAPI negative but positive in more than one color channel were considered multicellular events or non-specific antibody labeling. Astrocytic events were considered as singly positive for GLAST-PE and neuronal events were considered as those negative for all glial markers.</p><!><p>Stranded Total RNA-Seq was performed using an Illumina NovaSeq 6000 Sequencing System (RRID: SCR_016387). Total RNA (input range: 315 ng – 1417 ng) was used to initiate the Stranded Total RNA-Seq library preparation protocol. The total RNA fraction underwent ribosomal reduction, size fragmentation (6, 4, 3 or 2 minutes based on %DV200 calculation), reverse transcription into cDNA, and ligation with the appropriate indexed adaptors using the TruSeq Stranded Total RNA HT Sample Preparation Kit (Illumina cat no. #RS-122-2203). Following Agilent Bioanalyzer QC of the library preparation and library quantification using the Roche Lightcycler96 with FastStart Essential DNA Green Master (Roche cat no. #06402712001), the RNA-Seq libraries were adjusted to a 2 nM concentration and pooled for multiplexed sequencing on a NovaSeq 6000. The onboard clonal clustering procedure was automated during the NovaSeq 6000 sequencing run. The 100-cycle paired end sequencing was performed using the NovaSeq 6000 S1 Reagent Kit - 200 cycle (Illumina cat no. #20012864). Following collection, sequence data were converted from .bcl file format to fastQ files and de-multiplexed into individual sequences for further downstream analyses.</p><!><p>To assess the quality of the RNA-Seq data, we used the FastQC tool (http://www.bioinformatics.babraham.ac.uk/projects/fastqc). The QC report suggested high sequence duplication level in the samples; a high duplication level may result in lower counts per sample and will subsequently reduce the power to detect differentially expressed genes. RNA-Seq by Expectation Maximization (RSEM, RRID: SCR_013027) was then used to map sequences to the mouse mm10 genome assembly (Li & Dewey 2011). The bowtie2 (Bowtie, RRID: SCR_005476) was selected as the aligner within RSEM (Langmead & Salzberg 2012). RSEM produced low alignment rates as the unique alignment rates were observed to be <10 % for most samples.</p><p>Raw gene counts were then normalized according to library size and genes with low transcription, defined as less than 1 count per million (CpM) in at least 2 out of 27 samples, were filtered out. After the filtering of low expressed genes, a total of ~23,000 genes were retained for downstream statistical analyses.</p><!><p>Comparisons of physiologic parameters and dietary intake were assessed via unpaired two-tailed Welch's T-test using GraphPad Prism 9.0.0 software (RRID: SCR_002798). The Bioconductor package "edgeR" (RRID: SCR_012802), a software package for examining differential expression of replicated count data, was used for pair-wise comparisons of gene expression between mice given a 3-month KD and mice given a chow diet (Robinson et al. 2010). Given the count nature characteristic of RNA-seq data, "edgeR" implements novel statistical methods based on the negative binomial distribution as a model for count variability, including empirical Bayes methods, exact tests, and generalized linear models. Given the assumed negative binomial distribution for the models we fit using "edgeR", no formal test for normality was conducted on our data. When appropriate, the presence of outliers was assessed using the ROUT method with q = 1% in GraphPad Prism 9.0.0 software (RRID: SCR_002798). No outliers were identified by this method. The edgeR differential gene expression analysis was performed separately for astrocytes and neuron samples. Given the large number genes, and subsequently, the large number of tests being implemented, multiple testing adjustment was carried out by computing the Benjamini-Hochberg false discovery rate (FDR).</p><!><p>KEGG (RRID: SCR_012773) pathway analysis and visualization were performed using the R Bioconductor packages "gage" (RRID: SCR_017067) and "pathview" (RRID: SCR_002732) (Luo et al. 2009; Luo & Brouwer 2013). To identify biological pathways enriched with differentially expressed genes, we performed a gene set enrichment analysis (GSEA, RRID: SCR_003199) (Subramanian et al. 2005). Datasets were analyzed in reference to the h.all.v.7.0.symbols.gmt gene set database on the chip platform Mouse_Gene_Symbol_Remapping_MSigDB of the Molecular Signature Database (MSigDB, RRID: SCR_016863) (Liberzon et al. 2011). Network activation analysis amongst the top 2,000 differentially expressed genes was also performed by Ingenuity Pathway Analysis (Qiagen Inc., RRID: SCR_008653). Differentially expressed genes were defined as genes whose transcripts differed between groups by a minimum 2-fold increase or decrease when compared to the chow diet groups. Analysis was performed in reference to the Ingenuity Knowledge Base (RRID: SCR_008117) including direct and indirect interactions with filters applied for mouse and central nervous system.</p><!><p>Investigators involved in mouse husbandry were aware of the diet intervention status through the brain cell separation and RNA quality assessment stages. They were uninvolved in further sample management or statistical analyses following the submission of RNA samples to core facilities for cDNA library generation, quality control, and subsequent pathway analysis. Genomics Core and Biostatistics Core personnel were blinded to intervention status and remained so through the completion of all quality control steps and pathway analyses.</p><!><p>Figures within this publication were generated using GraphPad Prism v 9.0.0, Microsoft PowerPoint, and BioRender (SCR_018361).</p><!><p>We initially analyzed pooled brains from 3 mice to validate our cell population enrichment procedures. Following the myelin removal step, we used FSC and SSC profiling to identify potential cells, and DAPI, which labels the nuclei of cells that lack intact plasma membranes, to assess basic cell integrity. Within the potential cell population DAPI negative events constituted 92.86 ± 1.66% of the total events (Fig. 1A&B). To further determine what percentage of DAPI negative events represented live cells and not subcellular debris we used Dye Cycle Green to label nucleated cells, and Ghost Dye eFluor 450 to label dead cells. Dye Cycle Green labeling indicated 85.45 ± 3.00% of the DAPI negative events came from nucleated cells rather than from subcellular debris. Ghost Dye labeling indicated 68.03 ± 2.69% of the events derived from viable cells.</p><p>Applying our neuronal enrichment protocol to the brains of these 3 mice increased the percentage of glial marker-free events from 37.93 ± 6.88% in an unenriched cell fraction to 95.85 ± 2.10% in the final neuron-enriched fraction (Fig. 1C&D). Applying our astrocyte enrichment protocol increased the percentage of singly positive GLAST-PE labeled events from 38.23 ± 15.23% in an unenriched cell fraction to 56.58 ± 23.83% in the enriched astrocyte fraction, with an associated reduction within the non-astrocyte fraction to 8.34 ± 8.31% (Fig. 1C&E).</p><!><p>We placed 15 C57Bl6/N mice into each of the two diet groups, with the goal of obtaining 12 samples per group for analysis. One mouse in the KD group was euthanized with cancer, and five mice (3 chow, 2 KD) were euthanized due to fight wounds. Twelve mice in each group, therefore, completed the 3-month diet intervention.</p><p>For mice on the KD calories derived mostly from fat, and for mice on the chow diet calories derived mostly from carbohydrate and protein (Fig 2A). Mice on the KD (n=12), but not the chow diet (n=12), increased their BHB level, from 0.29 ± 0.10 mM to a final value of 0.82 ± 0.19 mM (p< 0.001) (Fig 2B). Blood glucose levels progressively fell in the KD mice, from 149.13 ± 26.68 mg/dL to 75.23 ± 35.37 mg/dL (p=0.011) (Fig 2C). The weight of the KD mice fell initially from 29.10 ± 3.06 g to a low of 26.40 ± 5.36 g, before rebounding to a final value of 28.68 ± 6.67 g. Chow-fed mice, on the other hand, steadily increased their weight from 31.16 ± 4.07 g at the start of the intervention to 38.49 ± 4.80 g at the end (Fig 2D). KD mice consumed significantly less food by weight (chow 5.29 ± 0.89 g/day; KD 2.34 ± 0.43 g/day; mean ± SD, p < 0.001), but the daily amount of kcal consumed was equivalent (chow 18.03 ± 3.02 kcal/day, KD 16.96 ± 3.09 kcal/day; mean ± SD, p=0.36) (Fig. 2E).</p><!><p>Of the 48 neuron and astrocyte samples generated from these 24 mice, 32 samples (8 per group) were selected for cDNA library generation based upon %DV200 quality ≥ 30% (Hester et al. 2016; Illumina 2015). Post-library generation we found two samples did not meet QC criteria. These samples were excluded from further processing and analysis, leaving 7 chow neuron, 7 KD neuron, 8 chow astrocyte, and 8 KD astrocyte samples.</p><p>After obtaining gene counts from RSEM, the correlation between samples was calculated to assess mislabeling. The correlation analysis revealed three samples were potentially mislabeled or otherwise significantly distinct from other samples within their group. These three samples were excluded prior to any subsequent statistical analyses. This resulted in a final count of 7 chow neuron, 7 KD neuron, 7 chow astrocyte, and 6 KD astrocyte samples.</p><p>For the 14 neuron and 13 astrocyte samples subjected to RNA-Seq, DV200 assessments showed equivalent RNA quality across the separation protocols and dietary interventions. DV200 mean percentages were as follows: chow neurons 42.69 ± 16.76%, KD neurons 45.81 ± 17.95%, chow astrocytes 47.25 ± 14.31%, and KD astrocytes 49.72 ± 15.36% (p=0.9586, two-way ANOVA) (Fig. 3A). Neuron RNA yields exceeded astrocyte RNA yields, but the diets did not affect RNA amount (Fig 3B). The mean RNA yields were as follows: chow neurons 2,964 ± 2,497 ng, chow astrocytes 1,160 ± 432 ng, KD neurons 2,857 ± 1,592 ng, and KD astrocytes 1,004 ± 301 ng (p=0.005 for cell type by two-way ANOVA).</p><p>To verify cell population enrichment in these samples, we quantified select cell-type specific transcripts. RNA levels for genes primarily expressed in astrocytes were higher in the astrocyte fractions (Fig. 3C). This is not surprising given FACS data shown in Fig. 1 predict our neuron fractions should contain very little astrocyte contamination. Relative to levels in neurons, astrocyte Gfap expression was 6.1, Slc1a3 was 5.7, Aldh1l1 was 5.3, and S100b was 8.4 times higher. RNA levels for genes primarily expressed by neurons were also higher in the astrocyte fractions (Fig. 3D). This too is not surprising given FACS data shown in Fig. 1 predict our astrocyte fractions should contain some neuron contamination, and Fig. 3B data that show neurons contain more RNA than astrocytes. Relative to neuron levels, astrocyte Rbfox3 was 2.0, Syp was 2.2, Nefm was 1.7, and Nefh was 4.1 times higher (Fig. 3D).</p><p>Given the high raw neuronal gene count in astrocyte samples, we ratioed each sample's summated neuronal gene counts to its summated astrocyte gene counts. Compared to the astrocyte samples, the neuron:astrocyte gene expression ratio was higher in all but one of the neuron samples (neurons 0.9926 ± 0.302, astrocytes 0.292 ± 0.102; mean ± SD, p < 0.0001) (Fig. 3E). The single deviant sample did not meet statistical significance in an outlier analysis and was identified while considering the expression levels of defined genes. For these reasons we did not exclude this sample. For the analyzed genes, expression changes within and between KD and chow-treated mice strongly correlated most strongly by cell type (Fig. 3F).</p><p>To determine whether either protocol generated robust contamination by other cell types, we measured RNA levels of microglia, oligodendrocyte, and endothelial cell genes. The expression levels for most of these genes was quite low (Figure S2).</p><!><p>We compared gene expression levels for the KD vs. chow diet neuron and KD vs. chow diet astrocyte samples. Tables S1 and S2 shows the 30 genes with the smallest uncorrected p-values from each cell population. After correcting for multiple comparisons, no genes were identified as differentially expressed.</p><!><p>KEGG analysis found that in neurons, the KD activated 96 molecular pathways and suppressed 0 molecular pathways. In astrocytes, the KD activated 5 molecular pathways and suppressed 67 molecular pathways. The different cell types showed pathway overlap, as 58 of the 96 activated neuron pathways were among the 67 suppressed astrocyte pathways (Fig. 4A; Table S3).</p><p>Our KD-treated mice consistently showed reduced blood glucose levels, which could potentially impact insulin signaling. The unbiased KEGG molecular pathway analysis did reveal the insulin signaling pathway changed in both neurons and astrocytes, but in different directions. Neurons tended to show increased, and astrocytes decreased, insulin pathway gene expression (Fig. 4B&C). Multiple molecular signaling pathways implicated by KEGG analysis seemed to be driven by underlying shifts in JNK, PI3K, Ras/Raf, ERK1/2, and PKA expression. We also observed increased neuron expression of oxidative phosphorylation-related genes and an activation of genes relevant to protein trafficking (Table S3).</p><!><p>We used the Molecular Signatures Database (MSigDB) to perform a GSEA analysis of all isolated transcripts. As we saw in the KEGG pathway analysis, the KD favored activation of neuron transcription pathways and the top results by normalized enrichment score (NES) again identified oxidative phosphorylation, insulin-related, and protein trafficking modules. Only one module, the one for "coagulation," was significantly downregulated in neurons from KD mice (Fig 5A). Astrocyte responses again showed a combination of KD upregulated and downregulated expression. The top activated pathways in KD mouse astrocytes were primarily inflammation-associated signaling modules. Downregulated pathways in KD mouse astrocytes included Myc targets, mTORC1 signaling, and DNA repair modules (Fig. 5B).</p><!><p>We performed an IPA analysis of the 2000 genes with the greatest diet-related differential expression. This again revealed a KD-related increase in neuron pathways that could relate to insulin signaling, including those that involve mTOR, ERK/MAPK, PI3K/Akt, and SAPK/JNK, as well as oxidative phosphorylation-related gene expression (Fig. 6A). KD astrocytes increased their expression of genes that mediate glucuronidation; implicated glucuronidation-dependent pathways included ones for nicotine, serotonin, melatonin, and thyroid hormone degradation. Other activated glial responses included several inflammation-related pathways (Fig. 6B).</p><!><p>KEGG analysis found that in neurons, the KD activated 30 pathology-associated modules and suppressed 0 pathology-associated modules. In astrocytes, the KD activated 1 pathology-associated module and suppressed 14 pathology-associated modules. The different cell types showed module overlap, as all the suppressed astrocyte pathology-associated modules were activated in neurons (Fig. 7A; Table S4).</p><p>For the 30 neuron KD-activated pathology-associated transcriptional modules 11 were associated with cancer, 5 with neurodegenerative diseases, 5 with infection, 3 with metabolic disorders, 4 with addiction, and 2 with cardiovascular disease. Within astrocytes, the only upregulated pathology module was one that associates with the response to S. aureus infection. Of the remaining suppressed modules, 7 were associated with cancer, 3 with addiction, 2 with metabolic disorders, 1 with infection, and 1 with cardiovascular disease.</p><p>In this unbiased transcriptomic analysis of KD brain effects, AD was the disease/pathology module with the strongest q-value. The curated KEGG AD pathway revealed enhanced transcription of mitochondrial oxidative phosphorylation genes contributed to this finding (Fig. 7B&C). Enhanced transcription of amyloid precursor protein (APP), apolipoprotein E (APOE), the APOE receptor LRP1, tau, the tau kinase GSK3β, and multiple proteins involved in modulating intracellular calcium also contributed.</p><!><p>IPA associated 59 disease states with the observed KD-altered neuron transcription patterns. Seventeen of the 59 significant z-scores were in the negative (range −0.124 to −3.442), and 7 of the 17 predicted reduced risk (z-score range −2.919 to −3.442) (Table S5). The 7 reduced risk modules included neurodegeneration of the central nervous system, hypoplasia of the brain, degeneration of the central nervous system, neurodegeneration of the brain, congenital malformation of the brain, degeneration of the brain, and neurodegeneration. Only two modules were found to have a positive z-score, although they were not predicted to be activated by the software analysis. These two modules were damage of the hippocampus (z-score +1.406) and damage of the cerebral cortex (z-score +1.085).</p><p>IPA associated 9 disease states with the observed KD-altered astrocyte transcription patterns. Of these, only the "brain lesion" module had an associated z-score, of +1.745, although the software did not predict its activation (Table S5).</p><!><p>As the KEGG analysis specifically revealed a KD alters the expression of genes in a pattern that links it to AD, we looked for KD-induced expression changes by mean fold change of genes previously tied to AD through linkage, genome wide association, and molecular studies (Karch & Goate 2015; Kunkle et al. 2019). Neurons from KD mice showed a frequent although not exclusive pattern of enhanced expression of genes associated with APP metabolism, cholesterol metabolism, endocytosis, and cytoskeleton/axon development, and a muted expression pattern of genes associated with the inflammatory mediator Ms4a superfamily (Fig. 8A). When considered from a more direct perspective of genes that encode APP-processing proteins or proteins believed to regulate APP levels, levels of its derivatives, or related molecules such as Notch neurons from KD frequently (but not exclusively) showed increased expression. This included components of the α, β, and γ -secretases (Fig. 8B). Expression of the tau (MAPT) gene, along with tau kinase and phosphatase genes (Fig. 8C), frequently (but not exclusively) showed increased expression. Relative to changes seen in neurons, astrocyte response patterns often shifted in the opposite direction.</p><!><p>In this study, a KD increased the gene expression of multiple neuron molecular pathways while simultaneously suppressing their astrocyte expression. The observed patterns of change robustly overlapped with those identified in some neurodegenerative diseases that feature altered brain bioenergetics, including AD.</p><p>Empirical experience from the epilepsy field reveals KDs affect brain function, but the precise mechanisms and mediators remain unresolved (Koppel & Swerdlow 2018; Li et al. 2013; McNally & Hartman 2012). Molecular studies intended to address these knowledge gaps often consider the brain as a single compartment. This perhaps limits their ability to resolve such questions, especially given emerging data that indicate neurons and astrocytes are not bioenergetically equivalent (Pellerin & Magistretti 2004; Guzman & Blazquez 2001). Our data support the view that energy metabolism in neurons and astrocytes fundamentally differs and suggest activities from one cell type may complement the activities of the other. Regardless, it seems obvious that if the expression of a gene increases in one cell type and decreases in another, mixing those cell types together would limit the ability to detect intervention-induced changes.</p><p>Instead of sampling individual neurons and astrocytes, we attempted to pool as many neurons and astrocytes as possible within separate, enriched fractions. While this approach sacrifices levels of purity achievable through laser capture, it should reduce the impact of sampling variation. We appeared to generate high levels of neuron enrichment, with less robust astrocyte enrichment. Despite this, the fact that we saw inverse relationships argues the levels of enrichment achieved were adequate to reveal cell-specific pathway changes.</p><p>A KD certainly changes systemic insulin signaling endpoints and our data clearly reveal this extends to the brain. The overall view, though, is complex as the expression of genes that would favor insulin signaling increases in neurons while their expression in astrocytes decreases. Given the emerging perspective that astrocytes provide energy-rich carbon molecules to neurons (Guzman & Blazquez 2001; Pellerin & Magistretti 2004), this could indicate neurons sensitize themselves for glucose utilization, or at least increase their sensitivity to insulin (Bouzier-Sore et al. 2003; Liu et al. 2017; Stobart & Anderson 2013). On the other hand, under KD conditions astrocytes appear to spare glucose utilization, at least by neurons, and do not increase their own sensitivity to insulin.</p><p>The three pathway analysis programs we used found that in neurons, a KD favors the expression of oxidative phosphorylation-related genes. Astrocytes showed neither an increase nor decrease in oxidative phosphorylation-related genes. This could reflect fundamental differences between neuron and astrocyte mitochondria. Neuron mitochondria may emphasize ATP production through oxidative phosphorylation, and astrocyte mitochondria may emphasize the strategic production of carbon intermediates that go on to support molecule biosynthesis or neuron respiration (Pellerin & Magistretti 2004; Guzman & Blazquez 2001). An increase in neuron hypoxia-related gene expression, detected by GSEA, suggests a KD causes neurons to increase their oxygen consumption, which is potentially consistent with the possibility that an increased expression of oxidative phosphorylation-related genes does in fact enhance neuron oxidative phosphorylation.</p><p>In most cases, findings from the different analytical platforms nicely complement each other. This typically featured obvious confirmations, but also more obscure ones. For example, GSEA showed a KD activated the xenobiotic metabolism module. This is consistent with the KEGG analysis finding that a KD alters molecular pathways that are relevant to drug addiction.</p><p>The unbiased KEGG pathology analysis found a KD altered molecular pathways modified in neurodegenerative diseases that feature bioenergetic dysfunction, with the three smallest q-values implicating AD, Parkinson's disease, and Huntington's disease. Interest in using KDs to treat the bioenergetic lesions observed in some of these diseases is not new (Swerdlow et al. 1989), and recent studies suggest a KD could potentially benefit persons with AD or the mild cognitive impairment syndrome that frequently precedes its diagnosis (Taylor et al. 2017; Krikorian et al. 2012; Brandt et al. 2019; Neth et al. 2020). In AD oxidative phosphorylation gene expression declines (Brooks et al. 2007; Wang et al. 2020; Liang et al. 2008). Our data suggest that in neurons, a KD may counteract that change. Neurons from AD brains also show evidence of insulin resistance (Steen et al. 2005), and our data suggest a KD may additionally counteract that change.</p><p>Other KD-induced neuron molecular pathways with potential relevance to AD include those that involve protein processing, endosomes, and neurotrophin signaling (Joshi & Wang 2015; Nixon 2005; Xu et al. 2018; Ginsberg et al. 2019; Marcelli et al. 2018). Our astrocyte data indicate a KD can increase glial inflammation-related activity. The functional consequences of any of these observed changes are unclear; such interpretations must await perspective from clinical trials.</p><p>Our hypothesis-driven analysis of AD-relevant genes provided additional insight into KD-AD molecular pathway associations and tended to confirm or complement those of the unbiased analyses. Reminiscent of the KEGG analysis, we saw a relative increase in endocytosis-related gene expression. Converse to the increase in astrocyte inflammation-related pathway expression revealed by KEGG, GSEA, and IPA, multiple genes relating to inflammation-related Ms4a activity were downregulated in neurons. With respect to genes implicated in the biology of APP and tau, we continued to observe expression changes whose directions differed between neurons and astrocytes. These changes once again tended to feature upregulation in neurons and downregulation in astrocytes. Because it is unclear whether observed changes promote a compensation, reverse a compensation, alleviate a pathological event, or enhance a pathological event, interpretations of how these shifts impact AD clinical outcomes should await perspective from clinical trials.</p><p>One mouse in the KD group developed a large soft-tissue mass and was euthanized prior to study completion. As it is unusual for young mice to develop soft tumor masses, it is worth considering the relationship of the KD to this anomalous event. KDs are high in fatty acids, which have been shown to drive tumorigenesis in a number of tissue types, including glioma (Lin et al. 2017; Lewis et al. 2015; Mashimo et al. 2014; Griffiths et al. 2013; Peck et al. 2016). Although our study focused on brain rather than cancer biology, our KEGG analyses notably revealed effects on cancer-associated pathways including those for choline metabolism, microRNAs, central carbon metabolism, and proteoglycan biology with inverse effects observed between neurons and astrocytes. Other implicated cancer-relevant pathways include insulin/IGF-1 signaling, PI3K-Akt-mTOR, DNA Repair, Myc Targets, MAPK signaling, and Ras signaling. Overall, our data suggest studies to evaluate how a KD influences tumor incidence and progression are warranted.</p><p>Following corrections for multiple hypothesis testing we did not observe the significantly altered expression of any individual genes. Some nominally identified genes, however, share functional similarities. Three of the top ten implicated neuron genes, Rnf115 (ring finger protein 115/Rabring7), Ubac2, and Rnf14 participate in ubiquitin signaling. Rnf115 just barely missed significance with a false discovery rate of 0.16. Rnf115 and Rnf14 both function as E3 ubiquitin ligases and are thought to play a role in membrane receptor internalization, ubiquitination, and trafficking content between the endosome and lysosome (Smith et al. 2013).</p><p>In addition to our use of enriched as opposed to pure cell fractions, our study features several technical and conceptual limitations. Our RNA quality was limited. We cannot rule out a disproportionate loss of transcripts with short poly-A tails, which could lead to a possible 3′-UTR bias in our transcriptome dataset. Our data also did not inform the status of mitochondrial DNA (mtDNA) expression. In assessing the effects of a KD, we only considered RNA-Seq-detected transcriptional changes and did not pursue orthogonal validation. RNA data may not reflect levels or activation states of proteins, so our data merely implicate biological processes without determining how their changes affect functional outcomes. We enriched for cell type irrespective of brain region, and it is possible that just as neurons differ from astrocytes, neurons or astrocytes from one region may fundamentally differ from other neurons or astrocytes from a different region. From a translational perspective, our KD featured a higher proportion of fat calories and a lower proportion of carbohydrate calories than most human KDs can achieve.</p><p>Our study used only male mice. As such, we were unable to identify sex-specific responses. Sex reportedly influences various KD endpoints including degree of ketosis, extent of weight and fat reduction, GLP-1 and PYY serum levels, and γ-glutamyl transferase changes (Lyngstad et al. 2019; D'Abbondanza et al. 2020). Although our unbiased analysis specifically identified AD as a disease of interest, our experiments did not extend our neuron-astrocyte directed analysis to AD transgenic mice. Doing so could potentially inform a number of studies that find ketone-based interventions variably affect histology and behavioral phenotypes in these models (Van der Auwera et al. 2005; Yao et al. 2011; Aso et al. 2013; Beckett et al. 2013; Brownlow et al. 2013; Kashiwaya et al. 2013). Regardless, as we did not include in our study measurements of protein levels or post translational protein modifications, and as our study did not include an AD transgenic mouse model, we cannot say how a KD intervention might affect neuron versus glia gene expression, AD-relevant protein levels, or AD-relevant post-translational modifications in such models. We cannot comment on any correlations between any parameter we measured with mouse cognition or brain electrophysiology.</p><p>Our study also does not establish the extent to which different KD-induced physiologic changes contributed to the observed transcriptional changes. Even considering just direct ketone body-mediated mechanisms becomes complex. Cells evolved intricate protein and protein-based networks to monitor and respond to bioenergetic states, and bioenergetic changes impact this infrastructure through post-translational protein modification and by altering gene expression (Selfridge et al. 2015). D-β-hydroxybutyrate itself can inhibit histone deacetylases, (Rardin et al. 2013; Shimazu et al. 2013), serve as a substrate for lysine β-hydroxybutyrylation (Xie et al. 2016), and bind G-protein coupled receptors (Rahman et al. 2014; Kimura et al. 2011; Miyamoto et al. 2019). In addition to causing ketonemia, the KD mice experienced relative hypoglycemia, elevated fat intake, and avoided a longitudinal weight gain. KDs should also lower insulin levels, which could profoundly impact molecular pathways.</p><p>At a minimum, our study provides insight that can inform the design of future studies intended to address how KDs or ketone bodies affect the brain. Although our study does not address whether a KD may benefit persons with AD or any other disease, it does though support the view that KDs can affect brain molecular biology in ways that impact neurodegenerative diseases that feature bioenergetic dysfunction, including AD.</p>

PubMed Author Manuscript

Anti-Cancer Potential of a Novel SERM Ormeloxifene

Ormeloxifene is a non-steroidal Selective Estrogen Receptor Modulator (SERM) that is used as an oral contraceptive. Recent studies have shown its potent anti-cancer activities in breast, head and neck, and chronic myeloid leukemia cells. Several in vivo and clinical studies have reported that ormeloxifene possesses an excellent therapeutic index and has been well-tolerated, without any haematological, biochemical or histopathological toxicity, even with chronic administration. A reasonably long period of time and an enormous financial commitment are required to develop a lead compound into a clinically approved anti-cancer drug. For these reasons and to circumvent these obstacles, ormeloxifene is a promising candidate on a fast track for the development or repurposing established drugs as anti-cancer agents for cancer treatment. The current review summarizes recent findings on ormeloxifene as an anti-cancer agent and future prospects of this clinically safe pharmacophore.

anti-cancer_potential_of_a_novel_serm_ormeloxifene

INTRODUCTION<!>Chemistry and Synthesis<!>Pharmacokinetics and Bioavailability<!>ORMELOXIFENE AS AN ANTI-CANCER AGENT<!>Ormeloxifene and Breast Cancer<!>Ormeloxifen and Prostate Cancer<!>Ormeloxifene and Head and Neck Cancer<!>Ormeloxifene and Ovarian Cancer<!>Ormeloxifene in Chronic Myeloid Leukemia<!>CONCLUSIONS, FUTURE PROSPECTS AND CHALLENGES

<p>Targeting the estrogen receptor (ER) by a Selective Estrogen Receptor Modulator (SERM), tamoxifen, is the oldest form of molecular-targeted therapy in breast cancer treatment. However, major limitations associated with current SERMs include their inadequate specificity, loss of activity and early onset of resistance. Therefore, with an aim to increase specificity and to circumvent difficulties associated with the treatment of cancers refractory to drug(s), several classes of SERMs were synthesized in recent years. One such SERM, ormeloxifene (also known as centchroman [3,4-trans-2,2-dimethyl-3-phenyl-4-p-(b-pyrrolidinoethoxy) phenyl-7-methoxy chroman]) (Fig. 1A), was synthesized by the Central Drug Research Institute (CDRI), Lucknow, India. Because it is free of the side effects commonly associated with estrogen and progestin based oral contraceptives, this drug has been developed into an effective, safe and easy to use oral contraceptive. It is approved for use in India and multiple other countries including Thailand, Bangladesh, Russia, New Zealand, United Kingdom and some North and South American countries [1]. Like a classical SERM, ormeloxifene suppresses the functions of the ER in the reproductive organs (ovaries, uterus) and breast, whereas it stimulates the ER in other organs like the bones [1, 2].</p><p>Ormeloxifene was introduced over 20 years ago as the first non-steroidal contraceptive in India under the trade names Saheli, Novex and Novex DS. Its contraceptive action appears to mainly rely on the ability to prevent embryonic implantation in the uterine wall by suppressing endometrial proliferation and decidualization [3]. Ormeloxifene has many advantages as a contraceptive drug including 1) this is a once a week oral pill (30 mg dose, weekly), 2) it has long serum half-life, 3) it has very few to no side effects since it does not affect ovulation or the levels of female hormone, and 4) it provides an opportunity for effective and efficient reversal of fertility upon discontinuation of the drug. In addition, the drug is cost effective, highly stable and does not require any specific storage conditions. These advantages of ormeloxifene have established its use as a highly effective birth control pill (contraceptive or morning-after pill) in women [1].</p><p>Recently, ormeloxifene has also been reported as a potent anti-neoplastic agent in various cancers [1, 4–8]. In vitro and in vivo studies have demonstrated several advantages of ormeloxifene as a pharmacophore, including extended serum half-life, favorable pharmacokinetic and pharmacodynamic properties, absence of severe toxicity, and cost effectiveness. Therefore, ormeloxifene can be developed as a potent therapeutic anti-cancer agent [1, 4–9]. This review summarizes recent findings on the role of ormeloxifene as an anti-cancer agent. We also outline some future directions for developing this promising pharmacophore as an anti-cancer agent for the treatment and management of cancers.</p><!><p>Ormeloxifene (C30H35O3N.HCl) is a white crystalline substance (MW 493.5 Da) with a melting point of 165–166°C. It is soluble in organic solvents like chloroform, acetone, methanol, and ethanol, but is almost insoluble in water. Ormeloxifene is a highly stable compound under normal storage conditions with no marked change in color, appearance, purity, or biological efficacy. At room temperature, ormeloxifene has been found to be stable and retain its biological activity and characteristics for at least 3 years when stored in glass containers or aluminum strips [1, 10, 11]. The chemical synthesis of ormeloxifene can be carried out through multiple processes [1, 10]. A primary mode of synthesis involves the Grignard reaction of cis-3-phenyl- 4-p-acetoxyphenyl-7-methoxy-3,4-dihydrocoumarin with methylmagnesium iodide in tetrahydrofuran (THF) to produce erythro-2-methyl-3-phenyl-4-(p-hydroxyphenyl)-4-(2-hydroxy-4-methoxyphenyl)-butan-2-ol. This compound upon cyclization with polyphosphoric acid (PPA) at 75–80°C yields cis-2,2-dimethyl-3-phenyl-4-p-hydroxyphe-nyl-7-methoxychroman [1, 2]. The isomerization of this compound with n-butyl lithium in DMSO produces an equimolar mixture of the D and the L enantiomer of ormeloxifene [12].</p><!><p>Ormeloxifene is a lipophylic molecule that possesses highly favorable pharmacokinetic and pharmacodynamic properties. This drug demonstrates low binding affinity to plasma albumin (Kd value of 13.90× 10−6) and does not interact with steroids like testosterone, cortisol, estradiol and progesterone [13–16]. It is therefore predicted to not interfere with the binding of steroids to specific steroid binding plasma proteins [17–19]. Additionally, while ormeloxifene is a selective modulator of estrogen receptor(s), it has been shown to neither compete with nonsteroidal estrogen agonists like diethylstilbestrol nor with estrogen antagonists like tamoxifen or nafoxidene [19].</p><p>Ormeloxifene is quickly metabolized by the liver and excreted from the body primarily via feces [20]. Information on ormeloxifene metabolism, organ retention and circulating levels has been primarily obtained from animal studies [21]. These studies reveal that well-perfused organs like the liver, lungs and spleen tend to retain more ormeloxifene than the less perfused organs like the pancreas and muscle. The active metabolite of ormeloxifene is 7-desmethylated ormeloxifene, which is rapidly formed within 1h of ormeloxifene administration and usually peaks between 8–24h [22]. While the liver is one of the first sites of drug metabolism of orally administrated ormeloxifene [21], high concentrations of the drug and its active metabolite are also found in the spleen, lungs, uterus and adipose tissues. Interestingly, the accumulation of the drug or its metabolite is always higher in tissue than in plasma and the concentration of ormeloxifene in tissue or plasma is invariably higher than its major metabolite, 7-desmethylated ormeloxifene [22]. In adult rats treated with a single oral dose of 12.5 mg/kg ormeloxifene, the maximum plasma concentration of 210 ng/ml was reached after 12h. The overall half-life of ormeloxifene or desmethylated ormeloxifene in adult female rats is about 24h and 36h, respectively.</p><p>In human clinical trials, healthy female volunteers were administered with either a single oral dose of 60 mg or 30 mg ormeloxifene. The overall half-life of the drug was detected to be around 168h [23]. Therefore, rats and other lower animals appear to metabolize ormeloxifene faster than humans. The maximum serum concentration (Cmax) of ormeloxifene in humans was dose dependent (Cmax of 55.53 ± 15.43 ng/ml for 30 mg dose and Cmax of 122.57 ± 6.25 ng/ml for 60 mg dose) and was reached within 4h–6h [23]. Similar Cmax values were also detected in breast cancer patients treated with either 30 mg, twice a week for 12 weeks (Cmax 54.98 ± 14.19 ng/ml) or 60 mg of ormeloxifene on alternate days for 1 month or 1 year (Cmax 135 ± 15.5 ng/ml). The serum concentrations of the drug and its half-life in nursing and non-lactating mothers were quite similar. Interestingly, about 2.5% of the ormeloxifene dose was found to be excreted in milk [24]. However, the amount consumed by a suckling child is thought unlikely to be of any physiological concern [24, 25]. Multiple repeated dosing in adult female volunteers revealed no significant difference in either amount of drug accumulated or the time required for maximal accumulation [26]. It has also been demonstrated that ormeloxifene exhibits strong and long lasting estrogen antagonistic and weak estrogenic activities in a number of different types of assays. While ormeloxifene has also been shown to inhibit progesterone action in some bioassays, it does not possess any androgenic or anti-androgenic activities [15, 16]. The mode of its anti-progesterone action is unknown since ormeloxifene neither affects progesterone expression nor progesterone receptor synthesis or receptor-ligand binding kinetics [3, 27, 28]. At contraceptive dose, ormeloxifene has no affect on the hypothalamo-pituitary-ovarian axis. At much higher doses (ten times the contraceptive dose), ormeloxifene may alter the hypothalamus-pituitary-ovarian response as shown in rats [29]. This is thought to occur as a result of the anti-estrogenic effect of ormeloxifene on the hypothalamus and its agnostic action on the pituitary gland [1, 30]. These data suggest that ormeloxifene is a clinically safe drug, with highly favorable pharmacokinetic and pharmacodynamic properties.</p><!><p>The ability of ormeloxifene to inhibit rapid cell proliferation in the endometrium during embryonic implantation along with its favorable bioavailability, stability and safety in humans makes it an attractive repurposing molecule for controlling undesired rapid cell proliferation such as endometriosis and cancerous conditions. Ormeloxifene has been shown to prevent cancers of breast and uterus, probably due to its potent estrogen antagonistic activity [1, 2]. Additionally, it also stimulates the formation of new bone mass, perhaps due to the weak estrogenic activity of this drug in bones [1, 2]. Estrogen agonist and antagonists function through estrogen receptors, which play an important role in the normal development and function of reproductive tissues as well as non-reproductive tissues (lungs, colon, prostate, and cardiovascular system) [31, 32]. Therefore, compounds that modulate estrogen receptor functions have been used for the treatment of many cancers, like breast cancer [32]. In this section, we will review the role of ormeloxifene as an anti-neoplastic agent against different types of cancers (Fig. 1).</p><!><p>ERα and ERβ are the two main types of receptors involved in ER mediated signaling. [22]. Upon ligand binding, the receptors dimerize and translocate to the nucleus, to regulate the expression of a multitude of genes. In addition, estrogen receptors also regulate non-genomic signaling by activating specific signaling pathways such as Protein Kinase C (PKC), AKT pathways, etc., to regulate various functions within the cell. Therefore, dysregulation of ER signaling is associated with initiation and progression of several cancers, including breast cancer [33]. ERα and ERβ appear to have opposite effects on cell proliferation, apoptosis and motility [33]. The mitogenic properties of ERα play an important role in the proliferation of breast cancer cells and poor disease prognosis [31, 32]. In fact, over 70% of tissue specimens from breast cancer patients exhibit overexpression of ERα. ERβ on the other hand appears to counteract the proliferative and malignant effect of ERα by modulating the expression of many ERα regulated genes [33]. The basal level of ERβ is downregulated in invasive breast cancer. Lowered expression of ERβ is associated with increased invasion in both breast cancer and prostate cancer. Therefore, compounds that can selectively modulate the functions of the ERα signaling may aid in the treatment of breast cancer. Ormeloxifene primarily functions as an estrogen antagonist in many organs including breast tissue. It interacts with both ER subtypes, demonstrating more selectivity and higher affinity towards ERα (8.8%) as compared to ERβ (3%) [34]. This anti-estrogen role of ormeloxifene makes it a choice anti-cancer agent for the treatment of breast cancers where ER functions are up regulated. In addition, similar to many SERMs, ormeloxifene also modulates various other signaling pathways independent of ER expression to regulate cancer cell growth.</p><p>The first in vivo use of ormeloxifene as an anti-neoplastic drug on advanced stage breast cancer patients, non-responsive to conventional therapy was reported by Mishra et al. (1989) [3]. In this clinical trial, breast cancer patients were treated with ormeloxifene (60 mg, three times a week) for 4–6 weeks. About 38.5% of breast cancer patients responded to the ormeloxifene therapy. In this clinical trial, ormeloxifene showed relatively better anti-breast cancer activity in older females patients (peri and post menopausal) compared to younger (pre-menopausal) patients. The responses to ormeloxifene treatment were more promising for bone, pulmonary, soft tissue, skin, and lymph-node metastases than for visceral metastases. However, there was no correlation between the number of lesions or estrogen receptor positivity and response to ormeloxifene therapy [3].</p><p>Molecular mechanistic studies with this compound have demonstrated that ormeloxifene induces caspase and mitochondrial-dependent apoptosis in breast cancer cells [5, 6]. In ER+ (MCF-7 cells) and ER− (MDA-MB231) breast cancer cell line models, ormeloxifene efficiently inhibited cell proliferation at concentrations similar to tamoxifene. Ormeloxifene, however, was more effective against ER(+) MCF-7 cells than ER(−) MDA-MB231 [5–6] (Fig. 2). Interestingly, ormeloxifene was also found to induce apoptosis at very low concentrations by depolarizing the mitochondrial membrane potential in the cell lines. In addition, ormeloxifene arrested breast cancer cells at the G0/G1 cell cycle phase [5, 6]. This effect correlated with the enhanced expression of cell cycle regulators like p21Waf1/Cip1and p27Kip1 and down regulation of Cyclin-D1 and Cyclin-E expression [6]. Since p21Waf1/Cip1 expression is usually controlled transcriptionally and post-transcriptionally by the p53-dependent/independent pathway [35], both cell types (MCF-7; p53+ and MDA-MB231; p53−) show p21Waf1/Cip1 activation after ormeloxifene treatment leading to cell cycle arrest. Recently it has been reported that ormeloxifene mediated apoptosis involves a cross talk between the extrinsic/intrinsic pathways and the modulation of the redox system in breast cancer cells [5]. Additionally, the efficacy of ormeloxifene in combination with sensitizing agents such as cucurmin and resveratrol has also been investigated [36]. Phytochemicals like curcumin and resveratrol are known to sensitize cancer cells to cytotoxic drugs, enhancing the cellular response to drugs and thereby minimizing side effects [37]. A recent study showed that resveratrol and curcumin sensitized breast cancer cells to ormeloxifene. In combination with resveratrol or curcumin, relatively much lower concentrations of ormeloxifene induced apoptosis in breast cancer cells [36]. These phytochemicals increased pro–apoptotic JNK/p38 signaling and depolarized mitochondrial membrane potential to enhance ormeloxifene mediated apoptosis via activation of caspase-9/caspase-3 and alterations of Bax/Bcl-2 protein ratios [36]. Moreover, the efficacy of ormeloxifene in combination with glycine soya, a dietary estrogen-like compound, was examined for reducing breast cancer in a 7,12-dimethylbenz-[a]anthracene (DMBA)-induced rat mammary tumor model system [38]. In this in vivo study, DMBA-induced tumor bearing mice were fed with either ormeloxifene (5 mg/kg body weight) or glycine soya (3× 104 mg/per kg body weight) alone or in combination for 5 weeks. The combination of ormeloxifene and glycine soya produced very high (98.6%) tumor regression compared to control groups. These results suggest the remarkable potential of ormeloxifene combinatorial anti-cancer therapy [38]. These data also demonstrate that the anti-cancer activity displayed by ormeloxifene involves the modulation of both ER dependent and ER independent pathways. However, detailed mechanistic studies are warranted to precisely understand the molecular basis of its anti-cancer action.</p><!><p>Prostate cancer is the second deadliest cancer in males after lung cancer [39]. The prostate contains ER in both the stroma and epithelium. Both animal models and human epidemiologic studies have implicated estrogen as an initiator of prostate cancer. In aging males, prostate cancer occurs in an environment of rising estrogen and decreasing androgen levels [40]. Therefore, SERM can be used for prostate cancer treatment by selective inhibition of ER. A few investigations with tamoxifen alone or in combination with other anti-cancer agents (doxorubicin, 5-fluorouracil, leucovorin, mifepriston, vinblastin and quercetin) have been reported [41–43]. Rossi et al. (2011) also investigated the use of another SERM, raloxifene, for the inhibition of prostate cancer cell growth expressing varying levels of ERα and ERβ [44]. Raloxifene efficiently induced cell death and inhibited proliferation of prostate cancer cells via modulation of multiple signaling pathways [44]. However, the use of ormeloxifene as an agent to treat prostate cancer has not been reported yet. Preliminary work from our group suggests that ormeloxifene inhibits growth of androgen dependent and androgen independent prostate cancer cells via inducing apoptosis. Studies also suggest that ormeloxifene attenuates certain key oncogenic pathways to inhibit growth of prostate cancer cells (unpublished results, personal communication). Continued investigations, however, are necessary to elucidate precise molecular mechanisms of its anti-cancer action in clinically relevant prostate cancer cell lines and animal models.</p><!><p>Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer in the world and approximately 50,000 new cases will be diagnosed in 2012 in the US alone [39]. Currently, no strategies are available for the treatment of advanced stage HNSCC [45]. The role of ER in HNSCC is controversial. While some previous investigations have shown little to no expression of ER in HNSCC [46], many recent investigations attribute a significant role for ER in enhancing cell proliferation and migration in HNSCC [47, 48]. Since ormeloxifene exerts its cytotoxicity via ER dependent and ER independent pathways in cancer cells, Srivastava et al. (2011) investigated for the first time its potential utilization for the treatment of HNSCC [8]. Ormeloxifene very effectively inhibited growth of multiple HNSCC cell lines by inducing apoptotic cell death through the activation of Caspase 3. The aberrant activation of the PI3K/AKT pathway is often seen in HNSCC and therefore, strategies to target this pathway represent one of the promising cancer prevention and therapeutic modalities. Interestingly, ormeloxifene treatment inhibited the phosphorylation of AKT and thereby attenuated downstream AKT signaling. This in turn suppressed the AKT/mTOR, STAT3 signaling pathway and enhanced the expression of p21 and p27 proteins that are associated with cell cycle progression [8]. This resulted in the stalling of cell cycle progression, inhibition of cell proliferation and cell survival, and enhanced apoptosis of HNSCC cells. Although these in vitro results are very promising, further preclinical and clinical studies are warranted to determine its use as an anti-cancer drug in humans.</p><!><p>Ovarian cancer is an endocrine malignancy with estrogen receptors being expressed in 65% of ovarian cancer patients [49]. Some data suggest that hormonal therapies may have an effect on ovarian cancer in palliative settings [50]. The use of SERMs as anti-neoplastic agent for the treatment of ovarian cancer is controversial [50, 51]. Ovarian carcinoma has been associated with prolonged use of tamoxifen in patients with a previous history of breast cancer [50]. However, a recent in vivo study demonstrated that tamoxifen induced apoptosis and suppresses tumor growth in nude mice in hormonal-dependent ovarian cancer cells [52]. Additionally, tamoxifen in combination with other agents, chemo-sensitized ovarian cancer cells resistant to Pertuzumab [53, 54]. In combination with Gefitinib, it modulated COX-2 and ER expression in ovarectomized mouse uterus. Tamoxifen-loaded nanoparticles were found to circumvent drug resistance in ovarian adenocarcinoma by enhancing intracellular ceramide [55]. Unfortunately, no published work has examined the role of ormeloxifene for the management of ovarian cancer. Preliminary in vitro results, however, suggest that ormeloxifene is a more potent anti-neoplastic agent compared to tamoxifen, especially in ER expressing ovarian cancer cells (unpublished results). Interestingly, ormeloxifene seems to work both as a classical SERM as well as a genotoxic agent in patient derived primary ovarian cancer cells (unpublished results). Currently, the molecular basis of the inhibitory role of ormeloxifene in ovarian cancer is being investigated using both in vitro and in vivo systems. Results from these studies would provide valuable information to evaluate the efficacy of this drug for the treatment of ovarian cancer in clinical settings.</p><!><p>Ormeloxifene has demonstrated excellent anti-proliferative activity not only in several solid tumors, but also in hematological tumors like chronic myeloid leukemia (CML) [56]. Ormeloxifene induced a concentration dependent increase in apoptosis in multiple CML cells lines (U937, HL60, and K562), with the most prominent effect in K562 CML cells [56]. Ormeloxifene arrested the cells in G0–G1 growth phase and induced ERK mediated apoptosis in CML K562 cells. Proteomic analysis of ormeloxifene treated K562 cells using 2-D gel electrophoresis and mass spectrometer revealed 57% of the differentially expressed proteins to be involved in apoptosis, while another 30% were associated with cell-cycle pathways. The study also demonstrated the ability of ormeloxifene to inhibit cell proliferation by blocking cell cycle in the G0–G1 phase by upregulating p21 expression and inhibiting c-myc expression via ormeloxifene-induced c-myc promoter-binding protein (MBP-1) synthesis. The induction of apoptosis by ormeloxifene was associated with the ERK pathway and loss of mitochondrial membrane potential. Importantly, the results obtained from ormeloxifene treatment in the K562 cell line were translatable to mononuclear cells isolated from CML patients [56]. Ormeloxifene treatment of mononuclear cells isolated from patients showed apoptotic phenotypes, including nuclear condensation and membrane blebbing. These promising results suggest the development of ormeloxifene either alone or in combination with other chemotherapeutic agents is appropriate for the treatment of CML.</p><!><p>Cancer is one of the most widespread and deadly diseases in the world. It is a leading cause of death in both developing and developed countries, accounting for over 13% of deaths worldwide. Substantial resources and research has been dedicated towards understanding the causes of cancer and developing treatments/cures for this disease. However, very little progress has been made towards eradiation of this disease, since cancer is a complex, multi-component disease involving abnormal functioning of multiple genes or pathways within the cells, resulting in unregulated growth and proliferation [57]. The deregulation of multiple pathways that leads to cancer poses a major challenge for designing and developing new anti-cancer drugs with little or no side effects. The development of new drugs is a very expensive, involved and time consuming process, with few compounds passing the final stage for clinical use. Due to high mortality and morbidly in cancer patients worldwide, a smarter and more effective way to develop new anti-cancer agents might be the repurposing of established drugs such as ormeloxifene as anti-cancer agents. In fact, several examples exist wherein old clinically approved molecules have been successfully repurposed in new roles and functions. At present, many approved drugs, including metformin, are under investigation as anti-cancer agents. Most of the work on ormeloxifene as an anti-cancer agent has been reported in breast cancer. Recently, this drug has also been investigated for the inhibition of other cancers like chronic myeloid leukemia, as well as head and neck squamous cell carcinoma. Analysis of molecular mechanisms modulated by ormeloxifene suggests regulation of multiple signaling pathways in cancer cells (Fig. 2).</p><p>Although clinical studies have supported partial to complete remission of lesions in breast cancer patients after ormeloxifene treatment, comprehensive, multicenter clinical trials are required to approve worldwide utilization of this novel anti-cancer agent for cancer treatment. Also, research is required to reveal detailed molecular mechanisms and key cellular functions modulated by ormeloxifene in cancer cells. Some of the specific processes that need to be examined include the effect of ormeloxifene on tumor microenvironment and host immune system. In addition, investigations are required to delineate its effect on angiogenesis and epithelial mesenchymal transition (EMT) (important aspects of tumor development and metastasis). Furthermore, careful investigations on ormeloxifene in combination with other chemotherapy drug(s)/photochemical(s) are also required to determine useful synergistic combinations for treatment of refractory cancers. This information may help in developing a novel therapeutic regimen for the treatment of recalcitrant cancers.</p>

PubMed Author Manuscript

Switching Cyclic Nucleotide-Selective Activation of Cyclic Adenosine Monophosphate-Dependent Protein Kinase Holoenzyme Reveals Distinct Roles of Tandem Cyclic Nucleotide-Binding Domains

The cyclic adenosine monophosphate (cAMP)- and cyclic guanosine monophosphate (cGMP)-dependent protein kinases (PKA and PKG) are key effectors of cyclic nucleotide signaling. Both share structural features that include tandem cyclic nucleotide-binding (CNB) domains, CNB-A and CNB-B, yet their functions are separated through preferential activation by either cAMP or cGMP. Based on structural studies and modeling, key CNB contact residues have been identified for both kinases. In this study, we explored the requirements for conversion of PKA activation from cAMP-dependent to cGMP-dependent. The consequences of the residue substitutions T192R/A212T within CNB-A or G316R/A336T within CNB-B of PKA-RI\xce\xb1 on cyclic nucleotide binding and holoenzyme activation were assessed in vitro using purified recombinant proteins, and ex vivo using RI\xce\xb1-deficient mouse embryonic fibroblasts genetically reconstituted with wild-type or mutant PKA-RI\xce\xb1. In vitro, a loss of binding and activation selectivity was observed when residues in either one of the CNB domains were mutated, while mutations in both CNB domains resulted in a complete switch of selectivity from cAMP to cGMP. The switch in selectivity was also recapitulated ex vivo, confirming their functional roles in cells. Our results highlight the importance of key cyclic nucleotide contacts within each CNB domain and suggest that these domains may have evolved from an ancestral gene product to yield two distinct cyclic nucleotide-dependent protein kinases.

switching_cyclic_nucleotide-selective_activation_of_cyclic_adenosine_monophosphate-dependent_protein

<!>Cyclic Nucleotide Binding Selectivity of WT and Mutant RI\xce\xb1<!>Cyclic Nucleotide Activation of PKA with WT and Mutant RI\xce\xb1 Subunits<!>Reconstitution of WT and Mutant RI\xce\xb1 Holoenzymes in RI\xce\xb1\xe2\x88\x92/\xe2\x88\x92 Cells<!>Intracellular Activation of WT and Mutant RI\xce\xb1 Holoenzymes<!>CONCLUSIONS<!>Reagents<!>Plasmid Construction<!>Cell Culture and Transfection<!>Protein Expression and Purification<!>Cell Imaging<!>Surface Plasmon Resonance (SPR)<!>Spectrophotometric Kinase Activity Assay

<p>Cyclic AMP-dependent protein kinase (PKA; protein kinase A) and cyclic GMP-dependent protein kinase (PKG; protein kinase G) are structurally related but functionally distinct kinases within the AGC family of serine/threonine protein kinases (Figure 1A).1 In its inactive state, PKA forms a tetrameric holoenzyme complex consisting of a regulatory (R) subunit dimer and two catalytic (C) subunits.2 One R-subunit inhibits one C-subunit by binding to the active site cleft, and the two R-subunits are bound together via an N-terminal dimerization domain. In contrast to the heterotetrameric conformation of PKA, PKG exists as a homodimer of two identical subunits each containing both the R- and C-domains fused together on a single polypeptide chain.3 Furthermore, PKA and PKG are similar in that they are both activated by cyclic nucleotide binding to their respective R-domains, which contain cyclic nucleotide binding (CNB) domains related to the bacterial catabolite gene activator protein (CAP).4,5 Both kinases can be activated by cAMP and cGMP, respectively; however, while PKA is activated by considerably lower concentrations of cAMP, PKG is preferentially activated by cGMP.6–9 Isolated CNB domains have been shown to be cyclic nucleotide-selective, and this selectivity is thought to keep both signaling pathways segregated for eliciting specific responses. 10–12 In addition to pools of cyclic nucleotides, further intracellular segregation of signaling involves pathway-specific coupling between kinases and substrates through A-kinase anchoring proteins (AKAPs) for PKA and G-kinase interacting proteins (GKIPs) for PKG.13,14</p><p>PKA is part of many signal transduction pathways governing cellular processes such as migration, apoptosis, and immunity.15 Different isotypes of PKA holoenzymes (as defined by the R-subunits: RIα, RIβ, RIIα, or RIIβ) can be targeted to subcellular compartments by a large repertoire of AKAPs or show tissue-specific expression. The RIα holoenzyme can be thought of as the master regulator of intracellular PKA activity and, as such, is located ubiquitously throughout the cell.16 While RIIα knockout mice have no discernible physiological defects other than increased sensitivity to AKAP inhibitors,17 the embryonic lethality of RIα knockout mice highlights the importance of the RIα subunit in the cAMP-mediated regulation of PKA activity.18 As a master regulator, RIα expression levels can increase to compensate for loss of the other R-subunits, while loss of RIα is not adequately compensated by the remaining R-subunits resulting in increased basal PKA activity due to excess free C-subunits.17–20 Additionally, mutations leading to RIα haploinsufficiency result in Carney complex, a disorder characterized by the formation of myxomas.21 Mutations within the CNB domains of RIα can lead to the phenotypically different Carney complex or acrodysostosis, which can be traced to functional differences in PKA activity that is increased or constitutive for Carney complex and decreased for acrodysostosis.22</p><p>While PKA has a broad range of functions, PKG is a key kinase in the nitric oxide (NO)/cGMP signaling pathway and serves to promote platelet disaggregation, smooth muscle relaxation, and vasodilation.14 There are two types of PKG, I and II. Type I PKG is the main regulator of smooth muscle tone, and knockout mice exhibit hypertension and defects in smooth muscle contraction.23 On the other hand, type II PKG knockout mice show deficiencies in intestinal fluid secretion as well as dwarfism due to reduced bone growth.24</p><p>Since activation of both PKA and PKG is mediated by cyclic nucleotides, the binding selectivity of cyclic nucleotides by the R-domains is a crucial mechanism in segregating their respective signaling pathways. Although previous studies have identified T193, T317, and R297 of human PKG Iβ as key residues for selective binding of cGMP,8,11,25,26 their specific roles in cyclic nucleotide selectivity have yet to be investigated within a cellular setting. Here, we present a novel system utilizing fluorescence resonance energy transfer (FRET) microscopy and RIα-deficient mouse embryonic fibroblasts reconstituted with wild-type (WT) RIα or cyclic nucleotide binding mutants to examine the molecular mechanisms of cyclic nucleotide selectivity as well as PKA activation. The ex vivo (in-cell) activities are contrasted and compared to the activities observed for purified recombinant proteins in vitro. In vitro, we found that mutation of either CNB domain abolished the cyclic nucleotide selectivity, while the combined mutations in both CNB-A and CNB-B switched the cyclic nucleotide selectivity from cAMP to cGMP in binding and activation. When reconstituted ex vivo as functional PKA holoenzymes bearing the wild-type and mutant RIα subunits, we also found that the combined CNB-A and CNB-B mutations effectively switched cyclic nucleotide dependent activation of the kinase from a cell permeable cAMP-analog to its cGMP counterpart. In addition, the ex vivo PKA holoenzyme assays implicate the importance of the cyclic nucleotide selectivity of CNB-A over that of CNB-B. Our findings underline the relevance of the key residues introduced for cGMP specificity and imply a sequence divergence of the PKA and PKG CNB domains in evolution.</p><!><p>While PKA I and PKG I share conserved CNB domains, we previously showed that PKG I displays additional amino acid contacts that confer its cGMP specificity.11,25 These contacts are conserved threonine or serine residues (T193 in CNB-A or T317 in CNB-B of PKG Iβ) in the phosphate binding cassette (PBC) and an arginine residue (R297 in CNB-B) at β5 that specifically interact with the guanine moiety. To determine if introducing these contacts into RIα switches its cyclic nucleotide binding specificity, we mutated the analogous residues in the CNB domains of RIα to threonine and arginine (T192R/A212T in CNB-A and G316R/A336T in CNB-B) (Figure 1). In sum, we generated three mutants of RIα with these mutations engineered into only the CNB-A or CNB-B domains (RIαmutCNB-A or RIαmutCNB-B) or both (RIαmutCNB-A,B) and compared their affinities for either cAMP or cGMP (EC50) and their activation constants (Kact).</p><p>The cyclic nucleotide binding affinities of the wild-type and mutant RIα subunits were determined using surface plasmon resonance (SPR) competition experiments (Figure 2 and Table 1). As reported previously,10 the EC50 cGMP/cAMP ratio for RIαwt was about 500-fold, indicating a high binding selectivity for cAMP over cGMP. Compared to RIαwt, RIαmutCNB-A exhibited a 1000-fold decrease in the EC50 for cGMP while the affinity for cAMP was unaffected, resulting in a weak cGMP-binding specificity. For RIαmutCNB-B, a 50-fold decrease in the cGMP EC50 was observed. The cGMP/cAMP EC50 ratio of the RIαmutCNB-B subunit was about 16, implying a reduced selectivity for binding cAMP over cGMP compared to RIαwt. Strikingly, the quadruple mutant RIαmutCNB-A,B showed a 12-fold increase in the cAMP EC50 and a >3800-fold decrease in the cGMP EC50, with a cGMP/cAMP EC50 ratio of 0.01, indicating a switched cyclic nucleotide binding selectivity from cAMP to cGMP as compared to RIαwt (Table 1).</p><!><p>To investigate the effects of altered cyclic nucleotide binding on PKA activation, activation constants (Kact) for either cAMP or cGMP were measured (Figure 3). The RIαwt holoenzyme was specifically activated by cAMP (Kact 0.053 μM) with a more than 100-fold preference in comparison to cGMP (Kact 7.4 μM), which is in good agreement with previous studies.8,40 RIαmutCNB-A showed Kact values of 0.46 μM for cGMP and 0.36 μM for cAMP, indicating a loss of cAMP selectivity. As for RIαmutCNB-B, the Kact of cGMP was comparable to that of RIαmutCNB-A (0.54 μM versus 0.46 μM), whereas the cAMP Kact was 3-fold lower (0.11 μM versus 0.36 μM). Unlike RIαmutCNB-A, RIαmutCNB-B is slightly more selective for cAMP in activation. Finally, RIαmutCNB-A,B showed Kact values of 0.048 μM for cGMP and 0.80 μM for cAMP, reversing its selectivity for cAMP over cGMP (Table 2). All RIαwt and mutant holoenzymes form stable complexes in the absence of cNMPs and are fully activated by cAMP and cGMP (Supplementary Figure 1).</p><p>Our work complements and extends previous studies by Shabb et al.8,26 with additional mutations introduced in RIα that not only increase its affinity for cGMP but also decrease its affinity for cAMP. Previous studies showed that the A212T and A336T substitutions in CNB-A and CNB-B of RIα, respectively, produced PKA with increased affinity for cGMP but with little change in cAMP affinity.8 Our previous studies had revealed additional cyclic nucleotide contact residues, in this case, R297 in CNB-B of PKG that specifically interacts with the guanine moiety of cGMP.11,25 This structure-informed insight allowed us to introduce a specific arginine in conjunction with A212T and A336T, as quadruple mutations T192R, A212T, G316R, and A336T, to obtain type I PKA exhibiting both an increased affinity for cGMP binding and a decreased affinity for cAMP binding. Thus, we were able to switch the cyclic nucleotide selectivity in favor of cGMP 50-fold over cAMP. As both T192R and G316R mutations are based on the R297 residue of the PKG I CNB-B domain (Figure 1D), our results confirm the critical role of R297 in achieving high cGMP selectivity.11</p><p>The data obtained from the cyclic nucleotide binding selectivity and PKA activation studies reveal different characteristics of the CNB domains. The cAMP affinity of RIα remained unchanged when either CNB-A or CNB-B was mutated. However, mutating both CNB domains at a time reduced the affinity for cAMP. This suggests that in the single domain mutants, the respective wild-type domain compensates for the mutated domain. Mutation of the CNB-A more strongly increased the affinity for cGMP than mutation of the CNB-B, which implies that CNB-A has a stronger impact on binding selectivity than CNB-B. Mutation of the RIα CNB-A domain resulted in a loss of selectivity between cAMP and cGMP, which was corroborated with the PKA activation assay. In contrast to the binding data, mutation of CNB-A alone increased the Kact for cAMP in comparison to the wild-type, underlining the role of the CNB-A as an allosteric switch. CNB-A can switch between two conformational changes, either binding and thereby inhibiting the C-subunit or binding cAMP.41,42 In the inactive PKA holoenzyme, the CNB-A domains are thought to be masked by the C-subunits and only become accessible when the CNB-B domains are occupied. The RIαmutCNB-B showed a 16-fold preference for binding cAMP over cGMP (albeit reduced in comparison to RIαwt); however, this drastic difference was not seen in the PKA activation assay, as the RIαmutCNB-B cGMP Kact was only 5-fold higher than the cAMP Kact, and 14-fold lower compared to the RIαwt cGMP Kact. Finally, mutating both CNB-A and CNB-B (RIαmutCNB-A,B) dramatically switches its selectivity in binding and activation, turning PKA from a cAMP-dependent to a cGMP-dependent protein kinase.</p><!><p>To examine the altered cyclic nucleotide selectivity of RIα upon mutations in a cellular context, we transfected the wild-type and mutant RIα as mCherry-tagged fusion proteins into MEF cells lacking RIα expression (RIα−/−) and, correspondingly, lacking type I PKA holoenzyme. Compared to WT fibroblasts, RIα−/− cells have excess unbound and therefore active catalytic subunits, resulting in increased PKA activity even under nonstimulated conditions.18,28 This increased activity can be inhibited with the small molecule H89 or by transfecting RIα−/− cells to express the specific PKA inhibitor, PKI (as mCherry-PKI) (Supplementary Figure 2). To confirm reconstitution of the PKA holoenzyme bearing the expressed mCherry-RIα, transfected RIα−/− MEFs were fixed and stained using an antibody directed against substrates phosphorylated by PKA, as readout for PKA activity (Figure 4A, Supplementary Figure 2). When compared to nontransfected RIα−/− cells, expression of each of the four RIα constructs resulted in cells with significantly reduced staining for PKA-phosphorylated substrates, an indication of lower PKA activity under nonstimulated conditions (Figure 4B). Therefore, transfection of the mCherry-tagged wild-type and mutant RIα subunits into RIα−/− cells successfully reconstituted the PKA holoenzyme as indicated by decreased basal phosphorylation of PKA substrates.</p><!><p>To examine the cyclic nucleotide stimulation properties of the RIα CNB-A and CNB-B mutants in live cells, we cotransfected RIα−/− MEFs with the mCherry-tagged RIα constructs and a genetically encoded FRET-based biosensor for PKA activity termed AKAR (A-kinase activity reporter). Briefly, the expressed AKAR polypeptide contains a PKA substrate motif and a phosphoamino acid binding domain that is flanked by yellow (YFP) and cyan (CFP) fluorescent proteins. When the substrate site is phosphorylated by PKA, AKAR undergoes a conformational change bringing the YFP and CFP proteins in close proximity that can be visualized as increased YFP/CFP emission ratio upon CFP excitation.35 For the purposes of our investigations, we used the cytosolic targeted AKAR (NES-AKAR), as changes in cytosolic PKA activity tended to be more robust compared to that found in the nucleus or at the plasma membrane.27</p><p>First, we treated transfected cells with the cell permeable cAMP analog, 8-CPT-cAMP, and monitored PKA activity as changes in YFP/CFP fluorescence emission ratio (Figure 5). Cells expressing mCherry-RIαwt exhibited measurable increases in PKA-mediated FRET within 20 min following 8-CPT-cAMP addition (Figure 5A,C and Supplementary Table 1). We noted that approximately half of cells expressing mCherry-RIαmutCNB-B exhibited a measurable response to 8-CPT-cAMP. In contrast, the vast majority of cells (>75%) expressing mCherry-RIαmutCNB-A or mCherry-RIαmutCNB-A,B exhibited minimal changes in FRET activity when stimulated with 8-CPT-cAMP, suggesting that mutations introduced within CNB-A of RIα greatly diminished the ability of 8-CPT-cAMP to activate PKA in cells.</p><p>Next, we utilized the same analyses to monitor PKA FRET activity for transfected cells treated with the cell permeable cGMP analog 8-CPT-cGMP. The vast majority of cells (>75%) expressing mCherry-RIαwt or mCherry-RIαmutCNB-B exhibited minimal PKA-mediated FRET activity upon stimulation with 8-CPT-cGMP (Figure 5B,C and Supplementary Table 1). In contrast, the majority of cells expressing mCherry-RIαmutCNB-A,B and approximately half of the cells expressing mCherry-RIαmutCNB-A were positively stimulated with 8-CPT-cGMP. Taken together, the results obtained from ex vivo cellular PKA activation studies indicate that the mutations introduced within CNB-A favor activation by cGMP while diminishing activity mediated by cAMP. Strikingly, combining mutations in both CNB domains (RIαmutCNB-A,B) resulted in robust conversion of PKA holoenzymes from cAMP- to cGMP-dependent.</p><p>Overall, our in-cell stimulations of RIα constructs with 8-CPT-cAMP and 8-CPT-cGMP had opposing results, as each RIα construct showed PKA activation in response to one of the cyclic nucleotides but not the other. The factor determining activation to a specific cyclic nucleotide appears to be the binding selectivity of CNB-A (shown as a schematic in Figure 6). Both RIαwt and RIαmutCNB-B possessed the wild-type CNB-A, and both resulted in PKA activation upon 8-CPT-cAMP stimulation. Likewise, both RIαmutCNB-A and RIαmutCNB-A,B possessed the mutated CNB-A, and stimulation of both with 8-CPT-cGMP resulted in PKA activation.</p><p>In our in vitro studies investigating the purified recombinant RIαmutCNB-A protein, we found a loss of cAMP/cGMP selectivity for nucleotide binding as well as PKA activation. RIαmutCNB-B still favored cAMP, but its selectivity was reduced compared to RIαwt, and a similar reduction was seen in PKA activation. Similarly, we observed a loss of cAMP/cGMP selectivity in PKA activation for both RIαmutCNB-A and RIαmutCNB-B constructs in our ex vivo experiments. It is evident that binding of cyclic nucleotides by both CNB domains is required for robust activation of PKA. However, stimulations of RIα−/− cells with cyclic nucleotide analogs were at least partially successful in eliciting activation of PKA when the RIα construct contained CNB-A that was predicted to bind the selected cyclic nucleotide. This underlines the idea that full activation of PKA ex vivo is largely dependent on cyclic nucleotide binding to CNB-A, as has been shown in previous studies.43,44 Furthermore, the cyclic nucleotide selectivity of CNB-B seemed to have little to no effect on PKA activation ex vivo under the conditions used in this study.</p><p>In isolation, CNB-B of RIα is more selective for cAMP compared to CNB-A.10 The differences observed for the full-length RIα may be influenced by cooperative binding of cyclic nucleotides and by interactions involving the catalytic subunit.44 Within the PKA holoenzyme complex, CNB-A is masked by the catalytic subunit, while cAMP binding to CNB-B is essential for cAMP binding to CNB-A. These dynamics may be altered when the affinity of the cyclic nucleotide to CNB-A is sufficient to bypass the requirement of cyclic nucleotide binding to CNB-B. The results are in line with the previously proposed model of PKA activation, as CNB-A is the determining step in eliciting full activation.45 An interesting contrast is seen in PKG Iβ, where CNB-A lacks cAMP/cGMP selectivity and the CNB-B regulates activation by being highly selective for cGMP.11,25,46</p><!><p>Results presented here suggest that PKA and PKG have evolved from an ancestral gene product that developed cAMP or cGMP specificity through single amino acid changes. The resulting differences in the primary structures of both kinases enable the segregation of cAMP- and cGMP-signaling pathways. Still, both PKA and PKG seem to be regulated by a number of cyclic nucleotide species, thereby integrating several pathways through cross-talk.47 Along these lines, both cAMP and cGMP regulate the activation of both kinases.48–50 Based on the sequence homology of the CNB-A and CNB-B domains of PKA and PKG, it can be postulated that the tandem CNB domains may have evolved from an ancestral gene duplication event.51,52 Interestingly, up to now, no cyclic nucleotide effector protein has been described to have one cAMP-specific CNB-domain and one cGMP-specific CNB domain. Such a hybrid would possibly have only a marginal preference for binding as well as activation by one of the cyclic nucleotides. It is intriguing to consider that single nucleotide exchanges in two amino acid residues implicated in this study (RIα A212 codon is GCA, PKG Iβ T193 codon is ACA; RIα G316 codon is GGA, PKG Iβ R297 codon is AGA) may form the evolutionary basis for cAMP vs cGMP selectivity of the CNB domains.5,26</p><!><p>The cyclic nucleotide analogs 8-(4-chlorophenylthio)-adenosine-3′,5′-cyclic monophosphate (8-CPT-cAMP) and 8-(4-chlorophenylthio)guanosine-3′,5′-cyclic monophosphate (8-CPT-cGMP) (Sigma-Aldrich) were solubilized in H2O at a stock concentration of 20 mM. Cells were stimulated with cyclic nucleotide analogs at a final concentration of 500 μM.</p><!><p>The FRET-based PKA activity biosensor construct NES-AKAR3 was a generous gift from J. Zhang (Johns Hopkins University).27 The construct expressing wild-type human RIα is as described previously.28 Cyclic nucleotide binding domain mutations were generated by site-directed mutagenesis using the Stratagene QuikChange methodology. Mutations in the CNB-A domain were made using oligos corresponding to the sequences: 5′-GTTAACAATGAATGGGCAAGGAGTGTTGGGGAAGGAGG-3′ for T192R and 5′-TTTATGGAACACCGAGAACAGCCACTGTCAAAGCA-3′ for A212T. Mutations in the CNB-B were made with 5′-AAGAGTTTGTTGAAGTGCGAAGATTGGGGCCTTCT-3′ for G316R and 5′-GATGAATCGTCCTCGTACTGCCACAGTTGTTGC-3′ for A336T. All constructs were sequenced to verify introduction of the desired mutations. Finally, all RIα constructs were subcloned as a C-terminal fusion to mCherry in a pcDNA3 (Life Technologies) vector backbone.</p><p>For recombinant expression in E. coli, the human RIα constructs were subcloned into the pQTEV plasmid via BamHI and HindIII restriction sites using the primers 5′-TGGATCCATGGAGTCTGGCAG-3′ and 5′-AGCAAGCTTTCAGACAGACAGTGAC-3′. The expressed proteins contain an N-terminal 7×His-tag and a tobacco etch virus (TEV) protease cleavage site.29</p><!><p>Cell experiments were conducted using prkar1a−/− mouse embryonic fibroblasts (herein referred to as RIα−/− MEF)18,28 cultured in DMEM (Sigma) supplemented with 10% fetal bovine serum (FBS, Sigma), nonessential amino acids (NEAA, Life Technologies) and penicillin–streptomycin (Life Technologies). Cells were incubated at 37 °C with 5% CO2 and kept between passages 2 and 20. RIα−/− MEF cells were transiently transfected by nucleoporation using the Amaxa Nucleofector 2b Device (program A-023) and left to grow overnight prior to cell imaging studies. In additional experiments, lipid-based transient transfection was carried out using Lipofectamine 2000 (Thermo-Fisher).</p><!><p>The human RIα constructs were expressed in E. coli Δcya TP2000 cells, which lack adenylyl cyclase activity.30,31 The inoculated liter cultures were incubated at 37 °C with shaking at 180 rpm until growth reaching an optical density of ~0.6 to 0.8, following which 400 μM IPTG was added to induce protein expression overnight at RT. Cells were harvested and stored at −20 °C.</p><p>For purification, the cells were resuspended in lysis buffer containing 50 mM KH2PO4, pH 8.0, 500 mM NaCl, 20 mM imidazole, and 5 mM 2-mercaptoethanol with complete EDTA-free protease inhibitor (Sigma-Aldrich), 1 mM phenylmethylsulfonyl fluoride (PMSF), and 0.1 mg mL−1 lysozyme. The homogenate was passed three times through a French pressure cell press (Thermo Electron Corp.) at 16 000 psi. The cell debris was spun down at 45 000g for 45 min, and the supernatant was loaded onto a 1 mL Protino Ni-NTA column (Macherey-Nagel) using an ÄKTApurifier UPC 10 machine (GE Healthcare). The His-tagged protein was eluted with lysis buffer containing 250 mM imidazole. Elution fractions were checked on SDS-PAGE gels.32 The proteins were further purified by anion exchange chromatography using a 1 mL ResourceQ column or by gel filtration using a 16/60 Superdex 75 column. RIα constructs were stored at −20 °C in 20 mM MOPS, pH 7, 150 mM NaCl, 2 mM EDTA, 2 mM EGTA, and 5 mM 2-mercaptoethanol.</p><p>Human PKA-Cα was expressed in E. coli BL21 (DE3) cells and purified via a PKI5–24-resin as previously described.33,34</p><!><p>RIα−/− MEF cells transfected to express mCherry-RIα and NES-AKAR3 were replated on fibronectin-coated coverslips in serum-reduced DMEM (1% FBS) for 2 h prior to imaging. Live cell FRET-imaging was performed on an Olympus IX-81 microscope equipped with an active focus system, X-Cite exacte light source (EXFO), Proscan II stage (Prior Scientific), CoolSnap HQ2 camera (Photometrics), Metamorph software controller (Molecular Devices), and environmental incubation (37 °C, 5% CO2, and humidity). FRET fluorescence microscopy images were acquired using the 60× NA 1.35 oil objective, and were taken at defined intervals at 300 ms exposure before and after application of cNMP stimuli for up to 30 min. CFP (480/30 nm) and YFP (535/40 nm) emission stimulated by CFP excitation (438/24 nm) were captured simultaneously using an image splitter (Photometrics DV2) equipped with 505 nm dichroic mirror (Chroma). Imaging of mCherry was accomplished with 589/15 nm exciter and 632/22 nm emitter (Semrock). Only cells expressing comparable fluorescence for mCherry-RIα and NES-AKAR3 were selected for imaging. Postacquisition ratiometric image analysis was carried out on ImageJ with the Ratio Plus plugin as previously published.35 The YFP/CFP FRET ratiometric value for each cell at each time point was calculated as the average values of multiple ROIs (regions of interest) within the cytoplasm. FRET values were further normalized against the average FRET values obtained before stimulation with cyclic nucleotides. Box–whisker plots of the analyzed data (20–26 cells per condition) were generated with ggplot2.36</p><p>For imaging of immunofluorescence stained RIα−/− MEF cells expressing mCherry-RIα, cells adherent on fibronectin-coated coverslips were fixed using 3.7% formaldehyde diluted in modified TBS (50 mM Tris, 100 mM NaCl, 20 mM Na4P2O7, 2 mM NaF, pH 7.4). Cells were permeabilized with 0.1% Triton X-100, blocked with 2% BSA/2% NGS (normal goat serum), stained with α-phospho-(S/T) PKA substrate antibody (Cell Signaling Technology) and DyLight 488 conjugated goat anti-rabbit antibody (Pierce). Reagent dilution and all washes were performed in TBS. Labeled cells were mounted with ProLong Gold (Life Technologies). Widefield epi-fluorescence images were acquired using the 40× NA 0.75 objective on the Olympus IX-81 for phospho-PKA substrate staining (FITC channel) and mCherry-RIα (mCherry channel). To minimize signal intensity variance resulting from different focal distances, all images were acquired using an identical (or fixed) focus offset. For analysis, the mean fluorescence for phospho-PKA substrate staining for each cell was computed from 4 ROIs distributed within the extranuclear cytoplasm. "Relative PKA Activity" was then computed as the mean phospho-PKA substrate signal obtained for each mCherry-RIα transfected cell divided by the average obtained from 3 nontransfected reference cells within the same field of view. Scatter plots of the analyzed data were generated using GraphPad Prism.</p><!><p>All SPR experiments were performed on a Biacore T100/T200 machine (GE Healthcare) with 20 mM MOPS, 150 mM NaCl, and 0.05% P20 surfactant as running buffer. The cAMP analogs 8-(6-aminohexylamino)adenosine-3′,5′-cyclic monophosphate (8-AHA-cAMP), N6-(6-aminohexyl)adenosine-3′, 5′-cycl i c monophosphate (6-AH-cAMP) and 2-(6-aminohexylamino)adenosine-3′,5′-cyclic monophosphate (2-AHA-cAMP) (BioLog Life Science Institute) were coupled to high density to flow cells 2, 3, and 4, respectively, of an S-series CM5 sensor chip using amine coupling as previously described.37 Flow cell 1 was activated and deactivated and used as a reference surface.</p><p>For solution competition experiments,38 the RIα proteins were preincubated with various concentrations of cAMP or cGMP, respectively, and injected over all sensor chip surfaces for 150 to 900 s. The dissociation phase was monitored for 75 s. The SPR signal at the beginning of the dissociation phase was plotted against the logarithmic competitor concentration and half maximal effective concentrations (EC50) were calculated from sigmoidal dose–response curves using GraphPad Prism 6.01. At the end of each cycle, the surfaces were regenerated by injecting 0.5% SDS and 1 M NaCl for 60 s each. All sensorgrams depicted were double-referenced by subtracting buffer injection signals and the signals of the reference flow cell. To determine the active protein concentration with regard to cyclic nucleotide binding, calibration-free concentration analysis (CFCA), implemented in the Biacore T100/200 control software, was applied using high-density cAMP analog chips as described in the Biacore T100 software package.</p><!><p>PKA holoenzymes were formed by mixing human PKA-Cα with the respective RIα proteins at a 1:1.2 molar ratio and by subsequent dialysis against 20 mM MOPS, 150 mM NaCl, 10 mM MgCl2, 1 mM ATP, and 5 mM 2-mercaptoethanol overnight. Kinase activity was measured using an enzyme-coupled assay according to Cook et al.39 To measure the cyclic nucleotide-dependent kinase activation, the respective PKA holoenzyme was preincubated with variable concentrations of cAMP or cGMP, respectively, in assay mix (100 mM MOPS, pH 7, 10 mM MgCl2, 1 mM phosphoenolpyruvate, 1 mM ATP, 150 U lactate dehydrogenase, 84 U pyruvate kinase, 220 μM NADH, and 5 mM β-mercaptoethanol), and the reaction was started by adding 250 μM Kemptide substrate peptide (LRRASLG). The absorbance at 340 nm was followed for 1 min. Phosphotransferase activity was plotted against the logarithmic cyclic nucleotide concentration. Activation constants (Kact) were determined from sigmoidal dose–response curves using GraphPad Prism 6.01.</p>

PubMed Author Manuscript

Balancing Bulkiness in Gold(I) Phosphino‐triazole Catalysis

The syntheses of a series of 1‐phenyl‐5‐phosphino 1,2,3‐triazoles are disclosed, within which, the phosphorus atom (at the 5‐position of a triazole) is appended by one, two or three triazole motifs, and the valency of the phosphorus(III) atom is completed by two, one or zero ancillary (phenyl or cyclohexyl) groups respectively. This series of phosphines was compared with tricyclohexylphosphine and triphenylphosphine to study the effect of increasing the number of triazoles appended to the central phosphorus atom from zero to three triazoles. Gold(I) chloride complexes of the synthesised ligands were prepared and analysed by techniques including single‐crystal X‐ray diffraction structure determination. Gold(I) complexes were also prepared from 1‐(2,6‐dimethoxy)‐phenyl‐5‐dicyclohexyl‐phosphino 1,2,3‐triazole and 1‐(2,6‐dimethoxy)‐phenyl‐5‐diphenyl‐phosphino 1,2,3‐triazole ligands. The crystal structures thus obtained were examined using the SambVca (2.0) web tool and percentage buried volumes determined. The effectiveness of these gold(I) chloride complexes to serve as precatalysts for alkyne hydration were assessed. Furthermore, the regioselectivity of hydration of but‐1‐yne‐1,4‐diyldibenzene was probed.

balancing_bulkiness_in_gold(i)_phosphino‐triazole_catalysis

Introduction<!><!>Introduction<!><!>Results and Discussion<!>Phosphine Synthesis<!><!>Phosphine Synthesis<!><!>Gold(I) Chloride Complex Synthesis and Structural Analysis<!><!>Gold(I) Chloride Complex Synthesis and Structural Analysis<!><!>Gold(I) Chloride Complex Synthesis and Structural Analysis<!><!>Catalysis<!><!>Catalysis<!><!>Catalysis<!><!>Catalysis<!>Conclusions<!>

<p>Bulky phosphines offer significant and well‐documented advantages as ligands in metal‐catalysed reactions. The landmark contributions of Buchwald and co‐workers have led to a deeper understanding of the synthetic chemistry enabled by such bulky ligands, and ligands such as S‐Phos and X‐Phos are an often‐required component of the toolbox of today's synthetic chemists.1 There is emerging interest in the importance of categorising and evaluating steric and electronic parameters2 of these types of bulky phosphines3 in order to correlate and ultimately predict their suitability for use in metal‐mediated catalysis.4 The Tolman cone angle has been used to describe the bulkiness of ligands,5 this descriptor has been complemented by Nolan describing bulkiness in terms of a percentage buried volume (%V bur).6 The %V bur described by a ligand can be calculated by using Cavallo and co‐workers' web tool SambVca (2.0).7</p><p>Some authors of this report previously detailed the synthesis, and application to palladium‐catalysed Suzuki–Miyaura cross‐coupling reactions, of 1,2,3‐triazole‐containing phosphines including analogues of the aforementioned Buchwald‐type ligands, such as 1a (Figure 1).8 Analysis of the steric parameters (buried volume)6, 9 using the SambVca (2.0) web tool,7 confirmed that the bulkier ligands of those tested were the most effective in said catalysis.8</p><!><p>Selected examples of a previously reported 1‐aryl‐5‐phosphino triazoles (left) (1a–b) and single‐crystal X‐ray diffraction structure of 1a (right), as reported elsewhere, H‐atoms omitted for clarity.8</p><!><p>The 1‐aryl‐5‐phosphino‐triazole ligands, such as 1a, present their 1‐aryl fragment in the same orientation as the phosphorus lone pair (of the free ligand) or in the direction of the metal (of a phosphorus‐metal complex thereof), thus significantly impacting the determined buried volume.8 Noting that ortho‐aryl tris‐phenylene phosphines (first reported, and somewhat overlooked, in 1940)10 can impart favourable properties as ligands in catalysis,11 it struck us that a bulky, and thus possibly superior, version of the triazole phosphine ligand is possible if the phosphorus atom were to be flanked by up to three, rather than one, triazole. A series of ligands ranging from zero to three triazoles, for comparison of the manifested bulk about the phosphorus centre by 1,5‐disubstituted‐triazole motif(s), was proposed. As such, a series of phosphines comprising of triphenylphosphine (2a) and tricyclohexylphosphine (2b) along with mono‐triazole‐appended phosphines 3a and 3b (available from a previous study), and the previously unprepared bis‐triazoles 4a and 4b and 1‐phenyl‐5‐phosphino‐tris‐triazole, 5 (Figure 2), was conceived.</p><!><p>The series of phosphines conceived to probe the influence the number of 1‐phenyl‐5‐phosphino triazole fragments about the phosphorus centre upon their "bulkiness" as ligands.</p><!><p>The substituents attached to phosphorus and the triazole nitrogen, could in principle be varied significantly. To provide a benchmark in this initial investigation cyclohexyl (a) and phenyl (b) groups attached to phosphorus were selected and the substituent attached at the 1‐N‐position of the triazole was restricted to phenyl in the first instance (Figure 2). As such two series of aryl‐ and alkyl‐substituted phosphines were planned.</p><!><p>Dicylohexylphosphino‐ and 5‐diphenylphosphino‐ 1‐phenyl triazoles, 3a and 3b respectively, were available from an earlier study. Their synthesis required 6 [Scheme 1 (i)] to be deprotonated with n‐butyllithium12 and reacted with dicyclohexyl‐ or diphenyl‐ phosphorus chloride, Scheme 1 (ii) (see earlier reports, and citations therein8, 13).</p><!><p>(i) The synthesis of triazole 6. Deprotonation of 6 and reaction with: (ii) dicyclohexyl‐ or diphenyl‐phosphorus chloride depicting the previously reported synthesis of 3a and 3b; (iii) cyclohexyl‐ or phenyl‐phosphorus dichloride for the synthesis of 4a and 4b by respectively; (iv) phosphorus trichloride for the synthesis of tristriazole phosphine 5.</p><!><p>The same protocol for selective deprotonation of 6 at the 5‐position,12 was followed by addition of cyclohexyl‐ or phenyl‐ phosphorus dichloride leading to the formation and subsequent isolation of 4a and 4b in 71% and 54% yield respectively [Scheme 1 (iii)]. When a third of an equivalent of phosphorus trichloride was added to deprotonated 6, the expected tris‐1‐phenyl 5‐phosphino triazole 5, resulted and was subsequently isolated in 77% yield [Scheme 1 (iv)]. The proton and carbon NMR spectrums of 5 (in [D]chloroform at ambient temperature) displayed a single set of well‐defined resonances, corresponding to the three equivalent triazole arms of the 5, suggesting a high degree of symmetry in solution on the NMR timescale.</p><p>Both 4a and 4b provided single crystals suitable for structural analysis by XRD (Figure 3a and Figure 3b, respectively, and supplementary material). In both cases the 1‐phenyl substituents of the triazole components point broadly in the same general direction as the phosphorus lone pair, thus boding well for a systematic study of the effect of modulating the steric crowding or bulkiness about a coordinated metal in a catalytically relevant complex. For compound 5, a single crystal XRD structure was obtained (Figure 4), the solid‐state conformation contains two unique molecules in the unit cell, both of which show the aryl arms of the triazole are directed along the same orientation as the phosphorus lone pair, creating a bowl‐shaped ligand.</p><!><p>(a) A molecule of 4a determined by single‐crystal X‐ray diffraction structure analysis. Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity) (lower) and a chemical drawing of 4a (upper). (b) A molecule of 4b determined by single‐crystal X‐ray diffraction structure analysis. Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity) (lower) and a chemical drawing of 4b (upper).</p><p>A molecule of 5 determined by single‐crystal X‐ray diffraction structure analysis. Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity) (lower) and a chemical drawing of 5 in approximately the same orientation (upper).</p><!><p>Complemented by commercially sourced tricyclohexylphosphine 2a and triphenylphosphine 2b, two ligand sets corresponding to Figure 2 (R = Cy or Ph) were therefore available for comparison in the complexation of a chosen metal. Owing to the importance of gold(I) phosphines in catalysis,14 and that X‐ray crystal structures of gold(I) chloride complexes 2a and 2b have been previously reported by others,15 the gold(I) chloride complexes of the ligands in the series presented in Figure 2 were compared. Ligation of the triazole‐phosphines (3–5) to gold(I), through phosphorus, was achieved by performing dimethyl sulfide‐phosphine exchange reactions on chloro(dimethyl sulfide)gold(I) (Scheme 2). Isolated yields of the corresponding gold(I) chloride complexes ranged from 58 to 91%. Triphenylphosphine and tricyclohexylphosphine gold(I) chloride complexes (7a and 7b) were purchased from commercial suppliers.</p><!><p>Synthesis of gold(I) chloride complexes (i) 8a–b; (ii) 9a–b; and (iii) 10.</p><!><p>The single‐crystal X‐ray structures of gold(I) chloride complexes 7a and 7b [gold(I) chloride complexes of tricyclohexylphosphine and triphenylphosphine] have been previously reported in the literature.15 The deposited PDB files were used to render images [Ortep III for Windows and PovRay, Figure 5a and Figure 5b, (i) and (ii) respectively] and determine the percentage buried volume using SambVca (2.0) [alternative representations of the XRD structure and a steric map thus resulting are shown in part (iii) of the corresponding figures ]. Tricyclohexylphosphine gold(I) chloride 7a has a 33.9%V bur whereas the triphenylphosphine gold(I) chloride complex 7b has a 30.8%V bur. The single‐crystal X‐ray diffraction structures of 8a (Figure 6a), 8b (Figure 6b), 9a (Figure 7a), 9b (Figure 7b) and 10 (Figure 8) were obtained from the complexes synthesised herein, and similarly analysed to determine the corresponding buried volumes described by the ligands in their complexes with gold(I) chloride (presented in Figure 9). The crystal structures of the gold‐phosphorus bond lengths used in the buried volume determinations were those obtained crystallographically.</p><!><p>(a) Single‐crystal X‐ray diffraction structure of arising from literature deposited CIF file of 7a: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity); (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 7a (33.9%) steric map of ligand depicted (right).[15a] (b) Single‐crystal X‐ray diffraction structure of 7b arising from literature deposited CIF file: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 7b (30.8%) steric map of ligand depicted (right).[15b]</p><p>(a) Single‐crystal X‐ray diffraction structure of 8a: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 8a (44.3%) steric map of ligand depicted (right). (b) Single‐crystal X‐ray diffraction structure of 8b: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space filing representation. (iii) Percentage buried volume determined from the crystal structure of 8b (40.3%) steric map of ligand depicted (right).</p><p>(a) Single‐crystal X‐ray diffraction structure of 9a: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 9a (54.5%) steric map of ligand depicted (right); (b) Single‐crystal X‐ray diffraction structure of 9b: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 9b (45.1%) steric map of ligand depicted (right).</p><p>Single crystal X‐ray diffraction structure of 10: (i) Ortep representation ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 10 (60.8%) steric map of ligand depicted (right). The unit cell contains two molecules of the complex wherein the phenyl ring centroid to gold distances are on average 3.58 Å (3.456 Å, 3.581 Å and 3.634 Å in the one molecule and 3.351 Å, 3.692 Å and 3.759 Å in the other). Between the two molecules of the unit cell a head to tail arrangement is present and the closest intermolecular Au···Cl intermolecular distances are 7.365 Å (see supplementary material for depictions).</p><p>Cyclohexyl and phenyl replacement by 1‐phenyl triazole motif and corresponding; triazole number vs. XRD‐determined percentage buried volume [SambVca (2.0)].</p><!><p>Since previously prepared mono‐triazole‐containing phosphines 1a and 1b were readily available to this programme of study, and as 1a was shown previously to have been among the superior ligands of the mono‐triazole set in earlier palladium‐mediated Suzuki–Miyaura catalysis, the synthesis of gold(I) chloride complexes thereof was also attempted. The gold(I) chloride complex of 1a (11) was prepared by the aforementioned dimethyl sulfide ligand exchange reaction, in good yield (95%, 0.25 mmol scale). A single crystal XRD structure of 11 was obtained (Figure 10a), and closely matched (by visual inspection) the structure of 9a, with a 46.0%V bur. Attempts to prepare a 1:1 complex of 1b and gold(I) chloride under the same conditions were inconclusive. However, a trigonal 2:1 ligand/gold(I) chloride complex 12 was identified by single‐crystal X‐ray diffraction crystal structure determination (Figure 10b) from a mixture of otherwise unidentified products. A 76.6%V bur (Figure 10b) was determined from the obtained crystal structure. Attempts to prepare 12 by employing two equivalents of ligand 1b gave inconclusive results. It is noteworthy that this 2:1 trigonal gold(I) structure is similar to crystal structures reported for [(Ph3P)2Au(I)Cl]16 and [(Ph3P)2Au(I)(SCN)].17 Furthermore a linear cationic gold(I) chloride complex [(Ph3P)2Au+][Cl–] bearing two triphenyl phosphine ligands with a fully dissociated chloride counterion also been reported.[15b] Therefore, the crystal structure of 12 is included for completeness. Attempts to prepare two and three 2,6‐dimethoxyl phenyl triazole‐containing variants of 4(a or b) and 5 failed to deliver any desired products.18</p><!><p>(a) Single‐crystal X‐ray diffraction structure of 11: (i) Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity). (ii) Space‐filling representation. (iii) Percentage buried volume determined from the crystal structure of 11 (46.0%) steric map of ligand depicted (right). Compound 11. (b) Single‐crystal X‐ray diffraction structure of 12: Molecular drawing and Ortep representation, ellipsoid probability 50% (rendered in PovRay, H‐atoms omitted for clarity) upper; space‐filling representation and steric map [SambVca (2.0)] from which a 76.6% buried volume was determined.</p><!><p>With the set of gold chloride complexes of Figure 2 (7a–b, 8a–b, 9a–b and 10) along with complex 11 in hand, their effectiveness as precatalysts for gold‐catalysed alkyne hydration was probed (Table 1).19 Catalytically active cationic gold(I) species may be generated by silver‐mediated halide abstraction, and choice of appropriate counter‐anion has been shown to modulate catalytic effects.20 In this case, triflate was selected as a counter‐anion across all cationic gold(I) catalytic systems studied. Initially the gold‐catalysed hydration of dec‐1‐yne (13a) was performed by in situ preactivation of 0.5 mol% of the corresponding gold complex through silver triflate‐mediated halide abstraction, using an excess of silver triflate in methanol (or as explained below, dichloromethane in the case of 10), alkyne 13a was stirred in methanol to which the preactivated catalyst solution was added. Since complex 10 was not well solubilised in methanol a protocol of activation in dichloromethane and dilution in methanol was deployed. To facilitate comparison in the subsequent reactions with lower catalyst loadings (0.25 and 0.05 mol% of gold complex), the dichloromethane activation and methanol dilution protocol was used in those cases. Following the addition of water (2 equiv.) all reactions were heated at 80 °C in a sealed tube for two hours and conversions to ketone 14a were determined by gas chromatography.</p><!><p>Screening of ligands for gold‐catalysed hydration of dec‐1‐yne (13a) to 2‐decanone (14a)</p><p>Methanol used as solvent, unless otherwise started.</p><p>Catalyst precursor initially solubilised in dichloromethane and dispersed in methanol, such that the resulting solvent composition was 10% CH2Cl2 and 90% MeOH.</p><p>2,6‐DMP = 2,6‐dimethoxyphenyl. These conditions were deployed across all tests experiments at 0.25 and 0.05 mol% gold(I) complex loading.</p><!><p>At the highest catalyst loading of 0.5 mol% only the bulkiest complex (10) failed to give complete conversion to product 14a (Table 1, entry 10 vs. other entries in same table ). When lower catalyst loadings were probed 0.05 mol% of gold(I) complexes proved to be too low to achieve satisfactory conversion within two hours (the maximum conversion observed across the set was 10% under these conditions). Good gradation across the test series was seen at a 0.25 mol% catalyst loading (Table 1) and confirmed the superior ligands, for this gold‐catalysed transformation, to be those with one triazole substituent. The dicyclohexylphosphine‐containing complex 8a afforded quantitative conversion to 14a (Table 1, entry 2). The corresponding mono‐triazole‐diphenylphosphine‐containing complex 8b gave the best conversion to 14a among the phenyl‐appended phosphine series of 57% (Table 1, entry 6). The three triazole congener (10) (Table 1, entry 4) gave only 11% conversion under the same conditions. Pleasingly, however, catalysts derived by halide abstraction in the same manner from complex 11 gave quantitative conversion of 13a to 14a (Table 1, entry 8).</p><p>Synthetic modification of phosphorus‐containing ligands can result in significant differences in the outcomes of reactions catalysed by their corresponding cationic gold(I) complexes.[4c], [19a], 21 The regioselectivity of hydration of unsymmetrical internal alkynes has been previously probed by Nolan and co‐workers who identified anti‐Markovnikov‐selective gold(I)‐carbene complex‐catalysed hydration.22 As such, a model reaction, namely the hydration of unsymmetrical internal alkyne 15 was selected to probe selectivity that might arise from the structural modifications across the series of gold(I) complexes of phosphino‐triazoles prepared in this study (Table 2). The reaction of water at the benzylic position (15‐a) represents the generally expected (Markovnikov) outcome, with selective reaction at the alternate alkyne position (15‐b, anti‐Markovnikov) being more challenging.</p><!><p>Selectivity of the hydration of unsymmetrical alkyne 15 to give ketones 16a and/or 16b</p><p>(a) Ratio determined by GC analysis of reaction mixture.</p><p>Ratio determined by analysis of the proton NMR spectrums of mixtures of 16a and 16b obtained from the reactions.</p><p>Complete consumption of 16 was observed, the product mixture contained 5% of a dimethyl acetal adduct as determined by GC/GC–MS.</p><p>2,6‐DMP = 2,6‐dimethoxyphenyl.</p><!><p>Compound 15 was added to a mixture of catalyst precursor (1 mol%) (which had undergone silver triflate (2 mol%)‐mediated in situ halide abstraction) in a degassed methanol/water mixture.23 Consumption of 15 and the ratio of products 16a:16b was determined by 1H NMR spectroscopy or gas chromatography. That quantitative conversion was observed in all‐but‐one case (tris‐triazole complex 10, Table 2 entry 4),[19a] is in keeping with the results from the hydration of dec‐1‐yne 13a (Table 1). All complexes showed a preference for Markovnikov addition at the benzylic position of the alkyne (position [a]). Whilst all cases gave product 16a in preference to 16b (ranging from 4.2:1 to 2.3:1 ratios of products 16a/16b), it is noteworthy that use of complexes 8a and 11 as catalyst precursors (Table 2, entries 2 and 8 respectively) afforded a greater proportion of the challenging anti‐Markovnikov product 16b.</p><p>Having established that, for gold(I)‐mediated hydration of alkynes, cationic gold(I) complexes bearing dicyclohexyl monotriazole ligands 1a and 3a (complexes 11 and 8a respectively) were superior in terms of activity (Table 1); and the 1a derived catalyst gave regioselectivity of <3:1 for 16a:16b (Table 2); complex 11 was deployed in a brief substrate scope survey (Table 3).</p><!><p>Substrate scope survey for triazole‐appended phosphine ligate gold(I) catalysed alkyne hydration</p><p>1 mol% L‐AuCl and 2 mol% AgOTf.</p><p>Significant Boc‐deprotection observed, isolated yield of 14j given in parenthesis.</p><!><p>Linear alkyl‐alkynes 13a and 13f (Table 3, entries 1 and 6) gave slightly higher yields than the cyclic alkane ethynylcyclohexane 13b (Table 3, entry 2), 79 and 89% vs. 64% respectively. 1‐Ethynylcyclohex‐1‐ene 13c (Table 3, entry 3) gave a similar yield (69%) to 13b. Phenylacetylene 13d (Table 3, entry 4) gave 58% yield, with the yield for the methylnaphthyl derivative 13e being somewhat better (80% yield, Table 3, entry 5). The yield for use of para‐methoxy phenyl acetylene 13g was better than for phenylacetylene (88 vs. 58% yield, Table 3, entries 7 & 4 respectively). Pleasingly an aryl boronic ester 13g was accommodated under the reaction conditions employed (81% yield, Table 3, entry 8), as was thiophene derivative 13i, albeit in a slightly lower isolated yield (61% yield, Table 3, entry 9). Quaternary oxindoles containing alkynes have been of interest to co‐authors of this report and their availability to this programme allowed 13j–l to be probed as substrates for gold(I) catalysed hydration (Table 3, entries 10 to 12).24 To obtain appreciable conversion to isolable oxindole‐containing products catalyst loading had to be increased to 1 mol%. The N‐H bearing substrate 13j was converted to product 14j in relatively low conversions (27%, Table 3, entry 10). The N‐benzyl analogue 13k fared better giving 85% isolated yield (Table 3, entry 11). A major side product arose from in situ removal of the Boc‐group upon reaction of 13l (Table 3, entry 12) giving rise to a mixture of desired product 14l (in 30% isolated yield) and deprotected product 14j (61% isolated yield). The reaction of 14m gave the dimethyl acetal of overall reaction at the alkyne terminus in 94% isolated yield (14m′, Table 3, entry 13). Nitroarene 13n, aryl ester 13o, aryl bromides 13p and 13q, pyridyl substrates 13r and 13s and tertiary butyl 13t appended alkynes all failed to give any detectable hydration products under the conditions employed (Table 3, combined entry 14). Further optimisation to facilitate these transformations was not conducted.</p><!><p>1,2,3‐Triazoles bearing 1‐aryl and 5‐phosphino functionalities were prepared, wherein the phosphorus atom was appended to one, two or three such triazole motifs. In addition, the corresponding gold(I) chloride coordination complexes bearing phosphino triazoles [with ancillary cyclohexyl or phenyl groups on phosphorus to complete the valence requirement of phosphorus(III) where required] were also prepared and their relative bulkiness determined using the online tool SambVca (2.0). Two more gold(I) chloride complexes were prepared from ligands containing a phosphorus atom appended by one triazole of the same general formula, where the 1‐aryl substituent on the triazole was 2,6‐dimethoxy phenyl. These complexes, along with complexes of tricyclohexyl phosphine and triphenyl phosphine were in situ converted into their corresponding cationic gold(I) triflate phosphine ligated congeners and compared as alkyne hydration catalysts. Increasing the "triazole number" about the phosphorus atom confirmed that, among the triazole appended phosphines investigated, the monotriazole‐containing phosphines were the most effective ligands for cationic gold(I)‐catalysed hydration of alkynes.</p><p>In summary, the data derived from the above experimentation show triazole‐phosphine ligand 1a to be the most promising of the ligands reported. The ease of synthesis and modification of the triazole ligand framework should prove useful in future ligand design, screening and optimisation campaigns.</p><p>Supporting Information (see footnote on the first page of this article): General and experimental procedures are available as supporting information, which also includes a summary of the crystallographic data collection and analysis, and further analysis of structural features of the crystal structures of the complexes discussed. All crystal structures disclosed in this report have had the corresponding data deposited at the CCDC with deposition numbers 1922101–1922111. Experimental procedures, including the synthesis of non‐commercially available alkynes, characterisation data and spectroscopic information are available in supporting information as are selected, preliminary palladium‐catalysis findings. Citations therein refer to methods, data and software used.25 A pre‐peer‐review version of this manuscript has been submitted to a preprint repository.26</p><p>CCDC 1922101 (for 4a), 1922102 (for 4b), 1922103 (for 5), 1922104 (for 8a), 1922105 (for 8b), 1922106 (for 9a), 1922107 (for 9b), 1922108 (for 10), 1922109 (for 11), 1922110 (for 12), and 1922111 (for S5 (see supplementary material)) contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from The Cambridge Crystallographic Data Centre.</p><p>Author Contributions</p><p>All authors contributed in varying degrees to planning the experiments, evaluating results and writing of the manuscript, specific contributions in addition to this are listed for each co‐author in alphabetical order: BRB helped direct aspects of the research and gave input and critical assessment throughout the progress of the project; PWD suggested experiments, supervised and provided critical assessment for aspects of the work; JSF led and co‐conceived the project, providing critical assessment of data, day‐to‐day project management and oversight, directed most aspects throughout, supervised most of the experimental work and wrote the majority of the manuscript; FM and TJS synthesised some of the alkynes used in Table 3; HvN contributed to preliminary studies detailed in the supplementary material and some aspects of mass spectrometry of complexes; MGW advised on the preparation of gold complexes and conducted experiments of Table 2; YZ co‐conceived aspects of the project, conducted all ligand synthesis and most of the reactions, drafted a proportion of the ESI, offered critical suggestions and conducted the XRD data collection and analysis herein.</p><!><p>Supporting Information</p><p>Click here for additional data file.</p>

PubMed Open Access

Consequences of exchange-site heterogeneity and dynamics on the UV-visible spectrum of Cu-exchanged SSZ-13

The speciation and structure of Cu ions and complexes in chabazite (SSZ-13) zeolites, which are relevant catalysts for nitrogen oxide reduction and partial methane oxidation, depend on material composition and reaction environment. Ultraviolet-visible (UV-Vis) spectra of Cu-SSZ-13 zeolites synthesized to contain specific Cu site motifs, together with ab initio molecular dynamics and time-dependent density functional theory calculations, were used to test the ability to relate specific spectroscopic signatures to specific site motifs. Geometrically distinct arrangements of two framework Al atoms in six-membered rings are found to exchange Cu 2+ ions that become spectroscopically indistinguishable after accounting for the finite-temperature fluctuations of the Cu coordination environment. Nominally homogeneous single Al exchange sites are found to exchange a heterogeneous mixture of [CuOH] + monomers, O-and OH-bridged Cu dimers, and larger polynuclear complexes. The UV-Vis spectra of the latter are sensitive to framework Al proximity, to precise ligand environment, and to finite-temperature structural fluctuations, precluding the precise assignment of spectroscopic features to specific Cu structures. In all Cu-SSZ-13 samples, these dimers and larger complexes are reduced by CO to Cu + sites at 523 K, leaving behind isolated [CuOH] + sites with a characteristic spectroscopic identity. The various mononuclear and polynuclear Cu 2+ species are distinguishable by their different responses to reducing environments, with implications for their relevance to catalytic redox reactions.

consequences_of_exchange-site_heterogeneity_and_dynamics_on_the_uv-visible_spectrum_of_cu-exchanged_

Introduction<!>Z 2 Cu<!>ZCuOH<!>and Table S4 †).<!>Cu dimers<!>Discussion<!>Conclusions<!>Conflicts of interest<!>Acknowledgements

<p>Copper ions exchanged onto zeolites are implicated as active sites for the selective catalytic reduction of nitrogen oxides, [1][2][3][4][5][6][7][8][9][10] oxidation of NO to NO 2 , 11,12 decomposition of NO and N 2 O, [13][14][15][16] and partial methane oxidation to methanol. [17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32][33] These Cu ions are associated with charge-compensating AlO 4</p><p>À tetrahedral sites that are distributed throughout the zeolite lattice. Because these Al substitutions are not ordered under typical zeolite synthesis conditions, a given framework and macroscopic Si/Al ratio will present a heterogeneous distribution of microscopic Al site ensembles, each of which provides a distinct exchange environment for a Cu ion. The precise Cu species and their relative densities in a given material are thus functions of framework topology, [34][35][36][37] of the density and underlying distribution of framework Al (Si/Al), of the Cu loading (Cu/Al), and even of the protocols used to introduce Cu onto the zeolite supports. [38][39][40][41][42][43][44] Further, multiple Cu ion site motifs may be of similar free energy at a given Al site or ensemble, and at nite temperatures these motifs may interchange at timescales relevant to observation or to catalytic turnover. [45][46][47] Given these many sources of structural diversity, assignment of spectroscopic features to specic Cu motifs in a heterogeneous solid is a non-trivial endeavor, but a critical one to make connections between local structure and catalytic function.</p><p>One strategy to reduce this complexity is to intentionally synthesize zeolites with framework Al distributions that present one or a few distinct Al site ensembles. This strategy is particularly promising for zeolites of relatively high symmetry, such as the chabazite (SSZ-13) framework that is composed of a single symmetry-distinct tetrahedral site. SSZ-13 samples synthesized using only organic N,N,N-trimethyl-1-adamantylammonium (TMAda + ) structure-directing cations nominally contain only isolated framework Al sites, 39 and are found to contain predominantly [CuOH] + ions aer aqueous Cu ion exchange and high temperature (>473 K) oxidation, based on titrimetric analysis and vibrational and X-ray spectroscopies. 38 In contrast, SSZ-13 samples synthesized in the presence of TMAda + and Na + as costructure-directing cations 39 contain detectable fractions of paired Al sites in 6-membered rings (6MR). Aer Cu ion exchange and high temperature oxidative treatment, these paired Al sites are observed to preferentially host Cu 2+ ions (Z 2 Cu) before isolated Al sites are occupied by [CuOH] + ions (ZCuOH). 38 Z 2 Cu and ZCuOH are expected to contribute differently to ultraviolet-visible (UV-Vis) absorption spectra based on ligand-eld arguments and density functional theory (DFT) calculations, consistent with dd transition and ligand-to-metal charge transfer (LMCT) features in experimentally-measured UV-Vis spectra that are observed to change with Cu content in Cu-SSZ-13 samples. 6,[48][49][50] These single Cu species may coexist with dimeric or higher nuclearity Cu clusters. Oxygen-bridged Cu dimers are well established to be present and quantiable with CO temperature programmed reduction, 15 to be plausible active sites for NO and N 2 O decomposition 14,15,[51][52][53] and partial CH 4 oxidation in Cu-exchanged zeolites, 17,30,33,[54][55][56][57][58][59][60][61][62] including Raman 63,64 and UV-Vis 14,55 spectroscopic observations and CO temperature programmed reduction to quantify such sites. In Cu-SSZ-13 samples of certain composition (Si/Al ¼ 5, Cu/Al ¼ 0.3-1.6), oxygen-bridged Cu dimers and larger Cu oxide aggregates are detected by X-ray spectroscopy and are the dominant active sites for NO oxidation to NO 2 under dry conditions. 11 UV-Vis spectra of certain Cu-SSZ-13 samples (Si/Al ¼ 13, Cu/Al ¼ 0.45) following high temperature O 2 treatment show features consistent with those for oxygen-bridged Cu dimers, 48 and these features disappear upon reduction with CH 4 , 18,50 implicating them as active sites for partial methane oxidation. The precise relationship between sample composition and treatment history, and the numbers and structures of Cu dimers or larger aggregates formed, however, remains less well resolved than such relationships for monomeric Cu sites in Cu-SSZ-13.</p><p>Here, we report UV-Vis spectra of model Cu-SSZ-13 zeolites prepared to contain predominantly Z 2 Cu or ZuOH sites, by virtue of their different framework Al arrangements and elemental compositions. We use supercell time-dependent DFT (TD-DFT) calculations to correlate observed UV-Vis spectral features with specic Cu motifs. We nd that spectra computed for single, static, minimum energy structures are in poor agreement with experimental observation, because Cu ion dynamics have a signicant impact on computed spectral features even at ambient temperature. We construct synthetic spectra by averaging over congurations visited during nitetemperature ab initio molecular dynamics (AIMD) simulations and show that these dynamically averaged spectra correspond closely with experimental observation. Further, experimental and simulated spectra are compared to identify features associated with a confounding subset of Cu dimers or larger aggregates in samples prepared to contain predominantly [CuOH] + species. These results resolve several inconsistencies in spectral and site assignments reported in Cu-exchanged zeolites.</p><!><p>First, UV-Vis spectra were collected on a Cu-SSZ-13 sample that contains only Z 2 Cu sites. 38,65 This sample was prepared by starting from an SSZ-13 sample (Si/Al ¼ 5) crystallized using a mixture of TMAda + and Na + cations to contain a nite and detectable fraction of paired Al sites, 39 followed by aqueous ion exchange with a cupric nitrate solution to achieve a composition (Cu/Al ¼ 0.21) demonstrated previously to contain only isolated Z 2 Cu sites through X-ray absorption spectroscopy, Brønsted acid site exchange stoichiometries with Cu 2+ and Co 2+ (2H + replaced per Cu 2+ or Co 2+ ), and IR spectroscopy 38 (sample preparation details in ESI Section S2.1 and elemental analysis in ESI Table S4 †). This Cu-SSZ-13 sample was treated in owing dry air to 673 K for 2 h, cooled to 300 K, and UV-Vis spectra were collected from 7000 to 50 000 cm À1 . Spectra are reported in quasi-absorption (K-M) units in Fig. 1a, and show a broad dd transition feature from about 8000 to 16 000 cm À1 characteristic of a d 9 Cu 2+ ion and a broad ligand-to-metal charge transfer (LMCT) feature from about 30 000 to 50 000 cm À1 . The spectrum is consistent with that previously reported for a Cu-SSZ-13 sample aer high temperature oxidative treatment. 6 The Cu 2+ ions in this zeolite sample are associated with ensembles of two Al centers separated by two or one intervening Si tetrahedral (T-) sites in the same six-membered ring (6MR), which we term "para" and "meta" respectively. We used a triclinic 12 T-site SSZ-13 supercell described elsewhere 11,66 to represent each ensemble (Fig. 2). In each case, a single Cu ion was placed within the 6MR and the structure annealed at 673 K for 150 ps using thermostated AIMD (computational details in ESI Section S1 †). 400 equally spaced congurations were extracted from the trajectories and relaxed. In each Al ensemble, all congurations relaxed to one of three energy minima shown in Fig. 2a and e. These three minima differ in the combination of Al-adjacent and non-Al-adjacent framework oxygen (O f ) that comprise the rst Cu coordination sphere. The framework distorts to accommodate these different Cu-containing minima, behavior consistent with structural distortions reported previously in calculations of metal-exchanged MFI, 67,68 MOR, 67,69 and FER, 70 and inferred from the appearance of two T-O-T deformation bands (900, 950 cm À1 ) in DRIFTS spectra aer Cu exchange into SSZ-13. 71 Despite these structural differences, the three para congurations differ in energy by less than 5 and meta by less than 20 kJ mol À1 (Fig. 2a, e, and also ESI Table S1 †). The lowest-energy para and meta congurations differ in absolute energy by 21 kJ mol À1 (ESI Table S1 †).</p><p>We performed additional AIMD at 300 K to gain insight into variations in Cu coordination environment at nite temperature 38,72 that might inuence the observed spectra. Fig. 2b reports histograms of distances between Cu and each of the six 6MR O f , collected at 0.6 fs intervals over the 150 ps simulation of para Z 2 Cu. Distances segregate into two groups centered at 2.1 and 3.3 Å, corresponding to Cu-coordinated and free O f , each group having widths >0.5 Å (Table S2 †) that arise from nite temperature uctuations of the lattice and Cu. The We used TD-DFT to compute frequency-dependent dielectric tensors and corresponding optical absorption spectra of the six minima. The computational details and codes for the VASP inputs and subsequent analysis are provided in ESI Section S1.2 † and also on the external Zenodo repository (DOI: 10.5281/ zenodo.1473128). Computed spectra are shown in Fig. 2c and g, reported in arbitrary K-M units. All six spectra exhibit features in the d-d and LMCT regions not evident in the experimental spectrum in Fig. 1a. Coordinatively similar (as measured from Cu-O distances and O-Cu-O angles) para minima 1 and 3 and meta minimum 2 exhibit equivalent spectra. Absolute intensities are greatest for lowest symmetry and least for highest symmetry minima. 73 Neither any individual spectrum nor a Boltzmann weighting of all spectra at 300 K (shown in ESI Fig. S3 †) recovers the experimental spectrum in Fig. 1a.</p><p>To simulate the effects of these geometric variations on observed spectra, we computed the absorption spectra of 400 equally spaced snapshots from the 300 K AIMD trajectories (ESI Fig. S4a The LMCT features do not simply correlate with the Cu-O f distances. Fig. 2d and h show the results of averaging increasing numbers of spectra computed from congurations extracted with equal spacing along the AIMD trajectories. Sharp features in both the d-d and LMCT regions become broadened upon averaging an increasing number of spectra from 1 to 200 structures and converge at approximately 200 structures, as evidenced by the small changes resulting from further averaging up to 400 structures. In contrast to spectra computed from the minimum energy structures, the nal averaged meta and para UV-Vis spectra are indistinguishable and consistent with experimental observations in Fig. 1a.</p><!><p>For comparison to the sample containing predominantly Z 2 Cu sites, we synthesized different Cu-SSZ-13 samples that contain predominantly the ZCuOH site motif that results when Cu 2+ ions are exchanged near isolated framework Al sites. We previously reported that crystallization of SSZ-13 zeolites in the presence of only TMAda + cations produces a material with predominantly isolated framework Al, reected in the inability to exchange divalent Co 2+ cations. 39,40 A sample of SSZ-13 (Si/Al ¼ 15) was synthesized using this method and exchanged with different amounts of Cu 2+ (Cu/Al ¼ 0.15, 0.24) to generate two Cu-SSZ-13 samples of different ZCuOH density (sample preparation details and elemental analysis in ESI Section S2. 1</p><!><p>Aer Cu 2+ ion exchange, each model ZCuOH sample was treated in owing oxygen at 673 K (20% O 2 , balance He) for 120 min (details in ESI Section S2.2 †). UV-Vis spectra of these Cu-SSZ-13 samples collected at 300 K are shown in Fig. 1b and c. Spectra of both samples show absorbance features in the 8000 to 22 000 cm À1 region that are centered around $ 11 059, 13 593, 16 379, and 20 077 cm À1 , and a shoulder in the 24 000 to 30 000 cm À1 range, and these features are higher in intensity for the sample with higher Cu content. These spectra are similar to those reported following similar pretreatments of Cu-SSZ-13 materials that are expected to contain predominantly ZCuOH species, 18,48,50 by virtue of the Na + -free synthesis methods used to prepare the parent SSZ-13 sample. In contrast, these spectra are markedly different from those reported on materials expected to only contain Z 2 Cu species (Korhonen et al. 6 and Fig. 1a). The four d-d transitions have different relative intensities in the two Cu-SSZ-13 samples shown in Fig. 1b and c. Literature Cu-SSZ-13 samples prepared to contain ZCuOH sites and exposed to the same O 2 pre-treatment also share the same four d-d transitions but again with different relative intensities. 18,48,50 We conclude that a mononuclear ZCuOH species cannot be solely responsible for the quadruplet feature. The sample-dependent variation suggests the presence of additional multinuclear ZCuOH-derived species with structures and populations that depend on synthesis, treatment and composition.</p><p>Da Costa et al. 15 reported that CO reduces multinuclear Cuoxo species in Cu-ZSM-5 to isolated Cu + (d 10 ) ions that do not exhibit d-d transitions. Similarly, we hypothesize that multinuclear Cu-oxo species present in Cu-SSZ-13 samples aer treatment in O 2 at 673 K will be reduced by CO at 523 K, leaving behind only isolated ZCuOH species and any residual Z 2 Cu sites. Model Z 2 Cu and ZCuOH samples were held in a owing stream of 5% CO at 523 K until no further changes in UV-Vis spectra were observed (details in ESI Section S2.4 †), prior to sealing the UV-Vis cell and cooling to 300 K to record the spectra shown in Fig. 1, an approach similar to that applied by Ipek et al. to Cu-SSZ-13 samples containing mixtures of Z 2 Cu and ZCuOH sites. 18 As expected, no changes were observed to the dd transition region in the spectrum of the Z 2 Cu sample upon CO exposure (Fig. 1a). In sharp contrast, the d-d features at 16 379 and 20 077 cm À1 and LMCT transition at 27 000 cm À1 in the spectra of both ZCuOH samples (Fig. 1b and c) disappeared aer exposure to CO, and features at 11 059 and 13 593 cm À1 were shied to 11 350 and 13 000 cm À1 and decrease markedly in absorbance. Despite differences in the d-d quadruplet feature intensity that are detectable aer high temperature O 2 treatment, the d-d transition features of both ZCuOH samples become similar aer CO treatment. These ndings indicate that not all Cu 2+ sites in Cu-SSZ-13 are reducible to Cu + in the presence of CO, that the Cu 2+ sites remaining aer CO reduction are similar for both samples (Fig. 1b and c, red), and that these signatures are of isolated ZCuOH sites.</p><p>We used the same triclinic supercell to describe a [CuOH] + ion-exchanged near an isolated Al. Each T-site in the chabazite lattice is common to two 8MR, one 6MR, and three 4MR. We used 473 K AIMD and geometry optimizations to compare the energies of the [CuOH] + ion in each of these orientations. The two 8MR orientations are isoenergetic and the Cu-X (X ¼ Si, Al, O) radial distribution function (RDF) computed from their AIMD trajectories are identical (ESI Fig. S6 †). From a nudged elastic band calculation, the two 8MR minima are separated by a 63 kJ mol À1 barrier (ESI Fig. S7 †). Similar calculations with the [CuOH] + ion directed into a 6MR and 4MR result in congurations 15 and 45 kJ mol À1 higher in energy. We thus expect a [CuOH] + ion to adopt and remain in one of the 8MR orientations at typical conditions of observation here.</p><p>Again to explore the consequence of ion dynamics on spectroscopy, we performed additional AIMD simulations at 300 K for 150 ps on a [CuOH] + ion in one of the 8MR orientations. During the course of the simulation the Cu ion remained coordinated to the same two O f , bond distances uctuated, and the OH ligand rotated between Cu-OH rotational conformers twice. Fig. 3b reports histograms of the two Cu-O f and Cu-OH distances. The Cu-OH bond is shorter and has a narrower distribution than the Cu-O f bond. The Cu-O f mean distances are slightly shorter than those from the Z 2 Cu simulations while the standard deviations are the same as the Z 2 Cu trajectories (ESI Table S2 †). Thus, the coordination environment around ZCuOH varies less than Z 2 Cu. The inset to Fig. 3b reports the fraction of the trajectory spent in each of the two rotational conformations.</p><p>The 8MR [CuOH] + ion can exist in one of two rotational conformers that differ in energy by 6 kJ mol À1 and are distinguished by whether the OH ligand points into or out of the 8MR (Fig. 3a). Fig. 3c reports the computed UV-Vis spectra of a relaxed 8MR [CuOH] + ion; each conformer yields a spectrum with two equivalent sharp features in the d-d transition region and a single sharp LMCT band. Predicted ZCuOH absorption intensities are less than either the Z 2 Cu para and meta spectra in Fig. 2c and g, consistent with the higher symmetry of ZCuOH and prior predictions that ZCuOH may have small or unobservable d-d transitions. 49,74 However the two spectra in Fig. 3c for the two ZCuOH isomers are only in rough correspondence with the observed spectrum of the ZCuOH samples.</p><p>We computed the absorption spectra of 400 equally spaced structures chosen from the 300 K AIMD simulation; all computed spectra are overlaid in Fig. S4c. † Signicant variations are present in the d-d (7000 to 14 000 cm À1 ) and LMCT (30 000 to 50 000 cm À1 ) regions (ESI Fig. S6c †), with shorter mean Cu-O distances again correlating with shis to higher frequency d-d transitions (ESI Fig. S5 †). Fig. 3d reports spectra averaged over 1, 10, 25, 100, 200, and 400 structures. The two sharp d-d features at 8000 and 13 000 cm À1 broaden and begin to merge, while the LMCT region converges to a peak spanning 30 000 to 45 000 cm À1 . The averaged d-d and LMCT regions z1000 cm À1 are red-shied but similar in shape to those observed aer CO reduction of the ZCuOH samples (Fig. 1b and c). Further, the decrease in computed intensity of the d-d relative to LMCT bands in ZCuOH compared to Z 2 Cu models corresponds with experimental observation. These observations support both the assignment of the Fig. 1b and c spectra following CO treatment (red) to isolated ZCuOH and the conclusion that the quadruplet features aer O 2 treatment (black) cannot be solely assigned to ZCuOH but rather have contributions from multinuclear Cu complexes.</p><!><p>The additional features in Fig. 1b and c respectively (details in ESI Section S1.12 and Fig. S9 †). Within the generalized gradient and harmonic approximations, the free energies to form ZCu(OH) 2 CuZ and dehydrate to ZCuOCuZ are computed to be À108 and À116 kJ mol À1 , respectively (details in ESI Section S1.12 †). These energetics are consistent with the formation of dimers from [CuOH] + ions of suitable proximity.</p><p>We performed 150 ps AIMD at 300 K on these two dimer structures in the 12-T-site supercell. In both trajectories the dimers remain roughly in the plane of the 8MR and retain coordination to the same bridging and framework O, unlike the more dynamic Z 2 Cu behavior described above. Both dimers vibrate internally and against the framework. Histograms of the Cu-O and Cu-Cu distances are shown in Fig. 4b and c We used TD-DFT and spectral averaging methods identical to those above to simulate UV-Vis spectra of both dimers at 300 K. We observed computed spectra to be sensitive to the geometries of the dimers, similar to the monomer Z 2 Cu and ZCuOH. 8500 and 12 400 cm À1 and an LMCT band edge that begins near 30 000 cm À1 . These two are clearly distinct from one another and from the computed spectrum of ZCuOH.</p><p>The spectroscopy of Cu dimers may be sensitive to Al proximity, through its inuence on geometric and electronic structures. To test this effect, we constructed two additional ZCuOCuZ models with two Al placed third-and second-nearestneighbor in an 8MR and introduced a Cu-O-Cu dimer so as to maintain Cu-O distances and a Cu-O-Cu angle similar to previous reports. 11,18 Fig. 4a shows the optimized structures F and G used to initiate subsequent dynamics. At 3NN, the ZCuOCuZ dimer is symmetrically coordinated to O f associated with Al; at 2NN, this symmetry is broken, although both Cu remain bound to two O f . During subsequent 300 K AIMD simulations the Cu ions retain their coordination; as shown in the histograms, Cu-O distances vary across the same ranges at all Al placements while Cu-Cu distances and Cu-O-Cu angles vary considerably with Al separation. While the computed spectra of the optimized structures are different, these differences largely disappear during averaging. As shown in Fig. 4d, averaged spectra have similar d-d features and differ only in the LMCT band edge position.</p><p>The Cu dimer spectroscopy could also be sensitive to geometric isomerism. To test this effect, we considered several examples of Cu dimers bridged by two O, a well known motif that exhibits several geometric isomers that are sensitive to Al separation. 34,[74][75][76] Fig. 4a structures A, B, and D correspond to three different Al placements and three different isomers, all of which were obtained by geometry relaxations beginning from literature structures. 11,18 A and B have triplet and D has a singlet ground states, consistent with earlier results. 34,75 Computed spectra at the optimized geometries (ESI Fig. S10 †) exhibit sharp and distinct peaks in both the d-d and LMCT regions. We performed AIMD on all three isomers; distance histograms collected during the simulations are shown in Fig. 4b and c. The dioxo dimer D is the least variable across the trajectory; dimers A and B sample much larger Cu-O f and Cu-Cu distances, respectively. During the nite temperature simulation, dimer A moves from a cis to a trans m-peroxo orientation whereas the optimized geometry has a slightly twisted O-O linkage, resulting in smaller Cu-O-O angles (geometry comparison in ESI Fig. S13 †). Fig. 4d reports computed spectra averaged over 400 snapshots. Spectra differ signicantly in band location and intensity both in the d-d and LMCT regions.</p><p>UV-Vis spectra are thus sensitive to Cu dimer composition and structure and dynamics. Comparisons with the experimental spectra collected aer 20% O 2 treatment at 673 K in Fig. 1b and c are complicated by the ill-dened number and nuclearity of Cu species present in the samples. Nevertheless, we can make some useful connections. The relative intensities of the d-d transitions for some dimeric Cu species (in particular A, B, and E) are computed to be greater than that of monomeric ZCuOH, consistent with the observation of decreases in dd peak intensity following CO reduction. Structures B, E, and F have features that roughly correspond with those observed at 11 059, 13 593, 16 379, and 20 077 cm À1 (Fig. 1b and c), but red-shied by an amount similar to that found in the comparison of computed and observed ZCuOH spectra. Structures C and D contain features that are plausible candidates for the features observed at 16 000 and 19 000 cm À1 . Structures A and B contain features that could account for the broad low energy LMCT shoulder from z22 000 to 27 000 cm À1 that disappears aer the CO reduction treatment. All these results imply that the variations in the quadruplet feature in the d-d region of nominally ZCuOH samples shown in Fig. 3b and c are associated at least in part with the contributions of different numbers and/or types of dimers from sample to sample.</p><!><p>Cu-SSZ-13 samples used here and reported in the literature are now well understood to be intrinsically heterogeneous at the microscopic scale, as a result of variations in composition and location of Al 77 and charge-compensating Cu ions, among other factors. Some of these differences are readily observed spectroscopically; for instance ZCuOH sites are clearly indicated by their distinctive O-H stretch vibration at 3660 cm À1 , 38,78,79 while others are more subtle to infer. The UV-Vis results reported here provide some guidance for distinguishing three types of Cu species. Z 2 Cu species are distinguished by relatively intense dd transitions with maxima near 12 000 and 14 000 cm À1 (Fig. 1a), and these features persist upon exposure to CO. Isolated ZCuOH species are indicated by lower intensity dd features near 11 000 and 13 000 cm À1 , which also persist during exposure to CO. We nd that samples intentionally prepared to contain exclusively ZCuOH always contain a confounding set of species that have relatively intense but irregular d-d features, which we assign to a mixture of higher nuclearity Cu oxo and hydroxo species. These species contribute four relatively prominent features in the d-d region, but they do not persist following CO reduction treatments.</p><p>Discrimination within these sets is more challenging. Samples prepared to contain only Z 2 Cu species potentially contain two geometrically distinct sites distinguished by the location of the charge-compensating Al, 80 and such sites are predicted here to have distinct UV-Vis spectra if computed at one minimum energy structure. Accounting for the nite temperature uctuations in Cu location between local minima and associated uctuations in Cu coordination environment, however, attenuates these differences, such that the two sites become spectroscopically indistinguishable. Similar factors affect the interpretation of XAS, 79,81-84 EPR 49,83,85 spectroscopies, and X-ray diffraction 5,82,85,86 patterns, and caution should be applied in inferring Z 2 Cu geometric information by comparison of observations to predictions from single, minimum energy structures.</p><p>In samples in which Cu exchange is predominantly associated with isolated Al T-sites and contain a majority of ZCuOH sites, observed UV-Vis spectra contain features in the d-d and LMCT regions that cannot be accounted for by the spectrum computed of these sites. 18,48,49,87 Strategies to directly prepare and characterize Cu-SSZ-13 samples that contain exclusively ZCuOH sites, either by exchanging dilute amounts of Cu or using higher silica-content SSZ-13 supports, are unlikely to be successful because of the difficulties of observing dilute ZCuOH and the presence of a confounding set of framework Al sites that can stabilize dimeric forms of Cu even at low Al density. Therefore, a strategy that combines synthetic efforts to bias formation of predominantly one Cu site type with treatments that selectively remove minority Cu species, is more likely to allow access to individual Cu site types.</p><p>The exact shapes and positions of the features depend on the zeolite composition and the precise pretreatment conditions, including temperature, pressure and duration of O 2 exposure, all suggestive of additional Cu sites produced dynamically. Calculations here show that [CuOH] + ions can move between adjacent 8MRs with an activation barrier of 63 kJ mol À1 (ESI Fig. S7 †) and that suitably proximal ions can condense into dimers. This proposal is consistent with experimental observations of a decrease in the approximately 3660 cm À1 vibrational feature associated with the ZCuO-H stretch with increasing temperature, 41 corresponding Raman shis for multiple Cu 2 O x motifs. 18,19,55 It is also consistent with the observation of dry NO oxidation to NO 2 on nominally ZCuOHcontaining samples, ascribed to dimeric Cu sites. 11,88 Because computed UV-Vis spectra of Cu dimers (Fig. 4) are sensitive to Al proximity, extra-lattice ligands, and to nite temperature structural uctuations, assignment of specic spectroscopic features to individual types of dimeric Cu species is not possible based on results reported here.</p><p>The samples prepared to contain Z 2 Cu sites have spectral features that are invariant to reduction in CO. Literature results on similar samples nd that they are insensitive to exposure to He, O 2 , or CH 4 . 6,19 The four features observed in the d-d region of the ZCuOH samples following oxidizing treatment are similar to those reported previously in similar samples, 18,48,50 but these bands do respond differently to subsequent reducing treatments. We nd that 5% CO exposure at 523 K reproducibly preserves a portion of the d-d and LMCT features, which we assign to isolated ZCuOH sites. In contrast, exposure to CH 4 at 473 K results only in a decrease in the lower energy portion (25 000 to 38 000 cm À1 ) of the LMCT band and disappearance of the 29 000 cm À1 band. 18,50,87 These results suggest that at the same temperature CH 4 reduces a different population of Cu x -O y H z moieties to Cu + than exposure to inert (He), and exposure to CO reduces all of the Cu x O y H z observed to reduce in either CH 4 or He. 18 Spectral features we assign here to isolated ZCuOH moieties are observed in similar materials to persist aer CH 4 exposure. 18,50 Precise identication of the Cu dimer sites responsible for CH 4 activation under different conditions remains an important challenge for experiment and computation.</p><!><p>Cu-exchanged zeolites remain a topic of great scientic interest because of their intriguing performance in catalyzing difficult transformations involving the nitrogen oxides and methane. Identication of active sites is complicated by the sensitivity of Cu exchange to zeolite framework types, the number and distribution of framework Al atoms, Cu content, and pretreatment conditions. The SSZ-13 zeolite framework is constructed of a single symmetry-distinct type of T-site, in principle reducing the number of possible distinct ion exchange sites and thus simplifying spectroscopic interrogation of those sites. Here we use directed synthetic approaches to emphasize different types of exchange sites, and DFT evaluations of site structure and spectral signatures to test this principle. We nd that in SSZ-13 samples prepared to contain only isolated Al T-sites, and thus in principle a homogeneous array of [CuOH] + ion exchange sites, always contain a confounding subset of O-or OH-bridged Cu dimers and/or larger aggregates in the Cu/Al exchange regime explored here. These latter species are intrinsically heterogeneous due to heterogeneity in bridging ligands and/or in arrangements of framework Al substituents, which likely involve more than one Al T-site. The underlying [CuOH] + UV-Vis spectrum can be revealed by selective CO reduction of the polynuclear Cu species. The UV-Vis spectrum, even of isolated [CuOH] + ions, is found to be inuenced by the intrinsic, nitetemperature dynamics of the site, as revealed through AIMD and TD-DFT calculations. In SSZ-13 samples prepared to contain 2NN (meta) and 3NN (para) Al pairs in the 6MR, the same dynamical factors serve to obscure spectroscopic differences between these two distinct types of Z 2 Cu sites. While calculations performed at the optimized geometries predict that the meta and para Cu-exchange sites are spectroscopically distinct, those differences are indistinguishable aer accounting for the nite-temperature uctuations in Cu ion coordination environment.</p><p>These results highlight the potential and the practical challenges of developing correlations between observed spectroscopy and the contributions of various Cu ion exchange sites and motifs to observed chemical reactivity. They highlight that precise characterization of active sites in this and similar systems demands a careful integration of chemical and spectroscopic interrogation with computational models that account for the structural and dynamical complexities of the materials.</p><!><p>There are no conicts to declare.</p><!><p>The experimental research at Purdue on zeolite synthesis and characterization was supported by the U.</p>

Royal Society of Chemistry (RSC)

Virtual Screening of TADF Emitters for Single-Layer OLEDs

Thermally-activated delayed fluorescence (TADF) is a concept which helps to harvest triplet excitations, boosting the efficiency of an organic light-emitting diode. TADF can be observed in molecules with spatially separated donor and acceptor groups with a reduced triplet-singlet energy level splitting. TADF materials with balanced electron and hole transport are attractive for realizing efficient single-layer organic light emitting diodes, greatly simplifying their manufacturing and improving their stability. Our goal here is to computationally screen such materials and provide a comprehensive database of compounds with a range of emission wavelengths, ionization energies, and electron affinities.

virtual_screening_of_tadf_emitters_for_single-layer_oleds

Introduction<!>Singlet-Triplet Energy Splitting<!><!>Ambipolar Trap-free Transport<!>Small Energetic Disorder<!>Building Blocks<!><!>Computational Workflow<!><!>Compounds With Small Singlet-Triplet Splitting<!><!>Compounds With Small Singlet-Triplet Splitting<!>Analysis of the Excited-State Character<!><!>Analysis of the Excited-State Character<!><!>Compounds With Tn States Lying Close to S1<!><!>Charge Carrier Density of States<!><!>Conclusion

<p>For obtaining efficient organic light-emitting diodes (OLEDs), it is convenient to tune individual processes, such as charge injection, balanced hole and electron transport, and triplet and singlet exciton harvesting, by using dedicated layers. Every new material adds a degree of freedom and hence flexibility to the OLED design. For instance, doped charge transport layers ensure Ohmic injection, an appropriate host material balances transport inside the emitting layer, and the phosphorescent emitter ensures triplet harvesting. However, every new emitter requires optimization of the surrounding layers, with respect to energy levels, triplet energies, and charge-transport properties, complicating the OLED design.</p><p>Recently, it was demonstrated that a complex multilayer design can be substituted by a simple single-layer architecture (Kotadiya et al., 2019a) without compromising the balanced and trap-free electron and hole transport. The ohmic charge injection and the absence of heterojunctions resulted in extremely low operating voltages and thus power efficiency in a single-layer OLED utilizing thermally activated delayed fluorescence, which helps to convert triplet into singlet excitons (Uoyama et al., 2012; Godumala et al., 2019). An external quantum efficiency of 19% was achieved. Owing to the broad recombination zone and low operating voltages, one of the key features of the single-layer device is the improved device stability, which can be used to design a stable blue OLED, a grand challenge in OLED research (Heimel et al., 2018; Paterson et al., 2019, 2020). In view of this, it would be useful to understand if the single-layer design can be employed for blue OLEDs: the issue here is the trap-free transport for both holes and electrons, which sets limits on the transport gap. In this paper, we first formulate the chemical design rules for TADF emitters with ambipolar transport. Using these rules, we then computationally pre-screen a set of molecules comprised of acceptor, donor, and bridge blocks and grade them according to the predicted emission wavelength.</p><!><p>The important task of a TADF emitter is to convert triplet into singlet excitations. To do this, the reverse intersystem crossing rate, krISC , should be high, which is only possible if the energy difference between the first singlet and the first triplet excited state is small, ΔEST<0.1 eV . A typical example of a TADF emitter is CzDBA (Wu et al., 2018), shown in Figure 1. CzDBA has a D- π -A- π -D architecture: two carbazole (Cz) fragments, two m-xylene bridges and a central 5,10-dihydroboranthrene (DBA) core. The methyl groups on the m-xylene bridge ensure that the core unit is nearly orthogonal to the π bridge, leading to a small overlap between the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) and hence nearly zero ΔEST .</p><!><p>The molecular structure of a prototypical single-layer TADF emitter, 5,10-bis(4-(9H-carbazol-9-yl)-2,6-dimethylphenyl)-5,10-dihydroboran-threne (CzDBA). It features a D- π -A- π -D (or arm-bridge-core-bridge-arm) molecular architecture.</p><!><p>To ensure a broad recombination zone within the emission layer, the thin film of the TADF emitter should provide balanced and trap-free transport of holes and electrons. To realize this, one needs to select compounds with an ionization energy (IE) and electron affinity (EA) lying within the trap-free energy window (Kotadiya et al., 2019b), i.e., with ionization energy (IE) < 6.5 eV and electron affinity (EA) > 2.5 eV. These criteria ensure that contaminants such as oxygen or water do not serve as energetic traps for holes and electrons.</p><!><p>From a dipolar glass model, the energetic disorder present in a disordered molecular solid is proportional to the dipole moment of the composing molecule. Therefore, thin organic films with molecules with a small dipole moment (D) normally have a narrower density of states (Novikov and Vannikov, 2009; Lin et al., 2019; Mondal et al., 2021; Stankevych et al., 2021). This design criteria can be enforced by selecting centrosymmetric molecules only of the D- π -A- π -D or A- π -D- π -A type, similar to CzDBA. This molecular architecture ensures a small dipole moment and hence narrow density of states (Liu et al., 2021).</p><!><p>With these design rules in mind, and in view of the successful example of CzDBA, we build and characterize a database of emitters that fulfill the aforementioned criteria. To construct the emitters, we start with 97 potential donor and acceptor building blocks, all shown in the Supplementary Note S1. All of them are (quasi-)linear, composed of three (fused) rings and are reported in literature (synthesizable). These building blocks are further pre-screened to ensure the desired donor-acceptor architecture in an emitter. The pre-screening proceeds as follows: knowing that the IE and EA of CzDBA is already quite close to the boundary of the trap-free window (Kotadiya et al., 2019a; Liu et al., 2021) we take the IECz and EADBA as the pre-screening criteria for donors and acceptors, respectively. Only the fragments possessing IE < IECz + 0.2 eV (EA > EADBA - 0.2 eV) will be chosen as "trap-free" donors (acceptors) and enter the next round, see Supplementary Figure S2. The molecular structures of donors and acceptors that pass the prescreening step are summarized in Figure 2 . To build the emitter molecules, only the building blocks with the inversion symmetry are used as core fragments. These are shown in dark colors in Figure 2. This choice helps to fulfill the centrosymmetric requirement for the entire molecule.</p><!><p>Donors (blue), acceptors (red), and building blocks that can serve both as donors and acceptors (purple). The blocks with inversion symmetry (dark colors) can be used as either core or arm fragments. The building blocks without inversion symmetry (light colors) can only be used as arm fragments. We also included boant4, which is not centrosymmetric, as a core fragment to increase the number of compounds in the database.</p><!><p>Using the selected building blocks, we constructed the database of D- π -A- π -D and A- π -D- π -A. The simplified molecular-input-line-entry system (SMILES) strings of compounds were created through combination of the SMILES strings of the composing donor, bridge and acceptor. The initial geometry of each compound was first optimized using a semi-empirical method and then by density functional theory (DFT). Details are given in the Supplementary Note S2.</p><p>To obtain reliable predictions of solid-state IE, EA and excited-state energy, we followed the cost-effective ω -tuning protocol (Sun et al., 2016, 2017). In addition to the ΔEST , the difference in the characters of the singlet and triplet excited states are crucial to the rISC rate (El-Sayed, 1963). For this reason, the excited-state characters were evaluated using a fragment-based method (Plasser, 2020).</p><p>For compounds that pass the screening criteria, the density-of-states distributions for holes and electrons were computed via multi-scale simulations, that include morphology generation using molecular-dynamics simulations, followed by polarizable force-field evaluation of the solid-state contributions to the gas-phase energy levels (Rühle et al., 2011; Poelking and Andrienko, 2016; Andrienko, 2018; Mondal et al., 2021). The entire workflow is illustrated in Figure 3.</p><!><p>Illustration of the computational workflow for virtual screening of single-layer TADF emitters.</p><!><p>The combination of the core and the arm fragments gives in total 441 A- π -D- π -A and 504 D- π -A- π -D compounds. Due to convergence problems in geometry optimization, especially in the anionic state with implicit solvent, the final database contained 433 A- π -D- π -A and 481 D- π -A- π -D compounds.</p><p>The IE and EA of all compounds either lies within the "trap-free window" or close to the borderline of the window, showing that the effectiveness of prescreening of the building blocks. Therefore, we put our emphasis on the small ΔEST criterion. The distributions of the ES1 and ΔEST are shown in Figure 4 (A- π -D- π -A) and Supplementary Figure S2 (D- π -A- π -D). Around 50% of the compounds (206 out of 433 for A- π -D- π -A and 268 out of 481 for D- π -A- π -D) have very small singlet-triplet energy level splitting, ΔEST < 0.1 eV, which illustrates the efficiency of the design strategy, that is the use of the m-xylene bridge. Moreover, the computed S1 energy and ΔEST of CzDBA is 2.487 and 0.016 eV, which is in excellent agreement with the experimental values of 2.48 and 0.033 eV (Wu et al., 2018; Kotadiya et al., 2019a).</p><!><p>2D histogram constructed using the descriptors ( ES1 , ΔEST ) of the A- π -D- π -A database (433 molecules). The corresponding 1D histogram for each descriptor is shown on the axes.</p><!><p>Among these small- ΔEST compounds, we observed a broad distribution in the S1 energy, ranging from 0.2 to 2.9 eV. This indicates the opportunity to design single-layer emitting OLEDs of different colors, including the infrared region. The two branches in Figure 4 represent the rest (50%) of the emitters with ΔEST > 0.1 eV, where a similar branch is also observed for D- π -A- π -D (Supplementary Figure S3). This is counterintuitive as the ΔEST should be small if the HOMO and the LUMO are separated via the m-xylene bridges.</p><!><p>To better understand the origin of the large ΔEST , we calculated the charge transfer (CT) number ranging from 0 to 1, using the fragment-based analysis (see Supplementary Note S2). We define the core as one fragment (fC) and two bridge + arm pairs as the other fragment (fA). If the hole is 100% located at one fragment and the electron is 100% located at the other one, the charge transfer number is 1, representing a 100% CT character. In contrast, if the hole and the electron are both localized on the same fragment, the CT number is 0, featuring a local-excitation (LE) character. In most cases, the CT number is a fraction between 0 and 1 since most adiabatic excited states exhibit a mixture of CT and LE characters. The larger the CT number of the excited state is, the higher the CT character it has.</p><p>Figure 5 depicts the 2D histogram based on the CT numbers of T1 and S1 states for the 227 A- π -D- π -A compounds with ΔEST > 0.1 eV. Most of the scatter points are located at the upper left corner, meaning that these emitters possess a charge-transfer S1 and locally-excited T1 states. A similar result was observed in the D- π -A- π -D case (see Supplementary Figure S4).</p><!><p>2D histogram constructed using the CT numbers of T1 and S1 states of the A- π -D- π -A molecules with ΔEST > 0.1 eV (227 molecules). The corresponding 1D histogram for each descriptor is shown on the axes.</p><!><p>The emergence of the LE states can be explained utilizing the frontier molecular orbital (FMO) energies of the constituent building blocks (Blaskovits et al., 2020), which is illustrated in Figure 6A. The competition between the CT excitation and LE excitation depends on the relative ordering of the FMOs. In this context, we can define two descriptors, RA=(ELUMOA−EHOMOA)/(ELUMOA−EHOMOD) , RD=(ELUMOD−EHOMOD)/(ELUMOA−EHOMOD) , where ELUMO/HOMOA/D are the LUMO/HOMO energies of the acceptor/donor. If the RA or the RD is much larger than 1, the CT excitation is more favorable than the LE for the low-lying excited states and vice versa. Figure 6B demonstrates that this simple approximation works quite well for our A- π -D- π -A database: For RA or RD smaller than ∼1.2, the CT number of T1 becomes close to 0. The same behavior was also found in the D- π -A- π -D database, as shown in Supplementary Figure S5. This indicates that a prescreening step based on the individual building blocks saves the computational cost, similar to pre-screening of singlet fission donor-acceptor copolymers (Blaskovits et al., 2020).</p><!><p>(A) Schematic representation of the relation between the competence of LE and CT states and the relative order of FMO energies; (B) RD-RA scatter plots colored by the CT number of the T1 state of the A- π -D- π -A database (433 molecules).</p><!><p>The S1 and T1 states of most molecules that pass the first screening step ( ΔEST < 0.1 eV) exhibit CT character. According to the El-Sayed rule, the krISC is zero between two states having the same excited-state character, which implies that the rISC may not occur for these pre-screened compounds. However, the conformational disorder present in the solid state leads to a distribution of dihedral angles between the constituent donor and acceptor (Weissenseel et al., 2019). This disorder gives rise to different excited-state characters, that is different mixing of CT and LE diabatic states of the S1 and T1 states, (de Silva et al., 2019), resulting in non-zero krISC . This explains why TADF could still be observed in the thin film of CzDBA, where CTS1 and CTT1 are both close to 1 in the gas phase (Kotadiya et al., 2019a).</p><p>In addition, higher triplet states (T n with n>1 ) with different excited-state character from that of S1, can also assist in the rISC process via a two-step mechanism (Gibson et al., 2016). A large second order coupling can be achieved when the energies of S1, T1 and T n are close to each other. Compounds with close-lying S1 and T1 that already show different excited-state characters would possess large first-order coupling and hence high k rISC. Therefore, we applied additional screening criteria to the as-screened ∼500 molecules: 1) there should be at least one triplet state Tn that is close to S1 (| ES1 − ET n | < 0.1 eV); 2) for the triplet states that are energetically close to S1, the difference between the CT numbers of S1 and Tn should be larger than 0.5 (CTS1 −CTT n > 0.5) to give reasonable spin-orbit coupling.</p><p>Overall, around 100 molecules pass the criteria (49 A- π -D- π -A and 46 D- π -A- π -D), where the molecular structures are summarized in Figure 7. All of these compounds, except for A- π -D- π -A molecules with non-centrosymmetric core boant4 ( D=4−5 Debye), possess nearly zero molecular dipole moment. Therefore, they are considered promising candidates for single-layer OLED emitters. The position of the electroluminescence (EL) spectrum maximum of each compound, as shown in Figure 7, was estimated by subtracting the computed S1 energy by a value δ, which is defined as δ=ES1−λEL,max=2.480−2.214=0.266 eV , where ES1 and λ EL,max is the experimental optical gap and the wavelength of the EL spectrum maximum of CzDBA (Kotadiya et al., 2019a). These values are listed in Supplementary Tables S1,S2. We obtained a series of potential TADF emitters with various EL spectrum maximum, ranging from infrared (0.716 eV) to blue color (2.660 eV), which paves the way for future development of single-layer OLED devices.</p><!><p>The estimated EL spectrum maximum of 49 A- π -D- π -A candidates of single-layer OLED emitters. The molecular structures of the 14 selected compounds are depicted.</p><!><p>More sophisticated solid-state simulations can then be performed for the much smaller molecular dataset, which is now only ∼10% of the initial number of compounds. As a proof of concept, we computed the charge carrier density of states for the blue A- π -D- π -A emitter, 37bdt1-ant2 (as shown in Figure 8). The amorphous simulated morphology was generated using molecular dynamics, where the details can be found in Supplementary Note S4. The energetic disorder for electrons (0.11 eV) and holes (0.12 eV) is relatively small, which indicates a good hole/electron mobility. This also demonstrates the success of our design strategy regarding small molecular dipole moment. Since the simulated IE and EA of 37bdt1-ant2 lie at the border of the trap-free window, further experimental measurements are necessary to verify if it is really free from universal traps.</p><!><p>Simulated ionization energy and electron affinity distribution in an amorphous 37bdt1-ant2 film.</p><p>1. Molecular gas-phase ionization energies and electron affinities within the ∼ 6.2 eV to ∼ 2.0 eV range. These are calculated using implicit solvent with the dielectric constant of 3 and ensure trap-free transport of electrons and holes.</p><p>2. Small molecular dipole moment. This condition is imposed by the molecular symmetry and ensures a narrow density-of-states distribution in the solid state.</p><p>3. Small singlet-triplet splitting. This is provided by the orthogonal alignment of the bridge and the core units, as well as the suitable level alignment between the HOMO and LUMO of the donor and acceptor units. This is required for efficient reverse intersystem crossing.</p><p>4. Different character of singlet and triplet excitations to ensure sufficient spin-orbit coupling that enables reverse intersystem crossing.</p><!><p>Using the suggested design rules, we have proposed a set of TADF emitters with a broad range of emission wavelengths, from infrared to sky-blue. We hope that the suggested structures can serve as a clear guide towards further development of efficient and stable single-layer OLEDs.</p>

PubMed Open Access

Dual-biomarker-triggered fluorescence probes for differentiating cancer cells and revealing synergistic antioxidant effects under oxidative stress

Hydrogen sulfide (H 2 S) and human NAD(P)H:quinine oxidoreductase 1 (hNQO1) are potential cancer biomarkers and also vital participants in cellular redox homeostasis. Simultaneous detection of these two biomarkers would benefit the diagnostic precision of related cancers and could also help to investigate their crosstalk in response to oxidative stress. Despite this importance, fluorescent probes that can be activated by the dual action of H 2 S detection and hNQO1 activity have not been investigated. To this end, dual-biomarker-triggered fluorescent probes 1 and 2 were rationally constructed by installing two chemoselective triggering groups into one fluorophore. Probe 1 provides a small turn-on fluorescence response toward H 2 S but a much larger response to both H 2 S and hNQO1 in tandem. By contrast, fluorescence probe 2 is activated only in the presence of both H 2 S and hNQO1. Probe 2 exhibits a large fluorescence turn-on (>400 fold), high sensitivity, excellent selectivity as well as good biocompatibility, enabling the detection of both endogenous H 2 S and hNQO1 activity in living cells. Bioimaging results indicated that probe 2 could differentiate HT29 and HepG2 cancer cells from HCT116, FHC and HeLa cells owing to the existence of relatively high endogenous levels of both biomarkers. Expanded investigations using 2 revealed that cells could generate more endogenous H 2 S and hNQO1 upon exposure to exogenous hydrogen peroxide (H 2 O 2 ), implying the synergistic antioxidant effects under conditions of cellular oxidative stress.

dual-biomarker-triggered_fluorescence_probes_for_differentiating_cancer_cells_and_revealing_synergis

<!>Rational design of the dual-biomarker-triggered uorescence probes<!>Synthesis and optical properties of the probes<!>Differentiation of cancer cells using the probe 2<!>Conclusions<!>Conflicts of interest

<p>Dual-biomarker-triggered fluorescence probes for differentiating cancer cells and revealing synergistic antioxidant effects under oxidative stress † Introduction Cancer, one of the most life-threating diseases, is characterized as uncontrolled growth and division of normal cells beyond their natural boundaries. The mortality of cancer remains high, which is mainly due to metastasis of primary cancer tumors. 1 The early stages of cancer development carry the maximum potential for therapeutic interventions, and the survival rate of certain cancers can be signicantly improved with early diagnosis and treatment. 2 Cancer biomarkers are endogenous molecules that are differentially expressed in cancer cells relative to their normal counterparts. Altered levels of such biomarkers can be measured to establish a correlation with the disease process and are useful for cancer diagnosis and therapy. 3 Furthermore, the simultaneous detection of multiple biomarkers can signicantly increase diagnostic accuracy. 4 Recent research has demonstrated that hydrogen sulde (H 2 S) and human NAD(P)H:quinine oxidoreductase 1 (hNQO1, EC 1.6.99.2) are potential biomarkers in certain cancer biology, which suggests that uorescent probes that detect these two species simultaneously would be of signicant utility.</p><p>As the third endogenous gasotransmitter, H 2 S is enzymatically generated from cystathionine g-lyase (CSE), cystathionineb-synthase (CBS) and 3-mercaptopyruvate sulfurtransferase (3-MPST)/cysteine aminotransferase (CAT). 5 H 2 S plays important roles in various biological and pathological progress, 6 and misregulation of endogenous H 2 S is associated with numerous diseases. 7 Specially, low levels of endogenous H 2 S appear to exhibit pro-cancer effects, whereas higher concentrations of H 2 S can lead to mitochondrial inhibition and cell death. 8 We note that some cancer cells, such as ovarian and colorectal cancer cell lines, exhibit increased H 2 S production. 9 This increased H 2 S may be useful for cell growth and proliferation due to H 2 Sinduced angiogenesis. 9c hNQO1 is a FAD-dependent avoprotein that catalyzes the obligatory 2-electron reduction of quinones to hydroquinones and provides versatile cytoprotection with multiple functions. 10 Levels of this reductase are elevated in a number of cancer types, including non-small cell lung cancer, colon cancer, liver cancer and breast cancer, 11 when compared to the surrounding normal tissue, making it an important cancer biomarker as well as an activator for some anticancer drugs. 12 In addition to their roles as potential cancer biomarkers, both H 2 S and hNQO1 are also vital participants in cellular redox homeostasis. H 2 S is recognized as a potential antioxidant, 13 can reduce disulde bonds, and can react with various reactive oxygen and nitrogen species. For example, Chang et. al. reported that vascular endothelial growth factor (VEGF)triggered H 2 S production is dependent on NADPH oxidasederived H 2 O 2 . 14a More recently, we as well as other groups found that endogenous H 2 S can be generated upon simulation of H 2 O 2 through the glutathionylation and subsequent increased activity of CBS in HEK 293 cells. 14b,c In addition, hNQO1 can reduce ubiquinone and vitamin E quinone to their active antioxidant forms and can also reduce superoxide to protect cells during oxidative stress. 15 Furthermore, hNQO1 can be an intracellular source of NAD + , which can fuel the activity of sirtuins to inhibit mitochondrial reactive oxygen production. 16 Despite the importance of H 2 S and hNQO1 in these systems, the response of these two biomarkers to oxidative stress remains largely unknown. To this end, our goal was to rationally design uorescent probes for simultaneous detection of H 2 S and hNQO1 to provide new chemical tools for investigating their possible crosstalk in redox homeostasis.</p><p>Recent research has demonstrated that uorescence-based methods are highly suitable and sensitive for in situ and realtime visualization of biomolecules. 17 Numerous uorescent probes have been developed for the detection of hNQO1 or H 2 S in living systems. 18 Until now, however, none of these probes allows for the simultaneous detection of the chemical (H 2 S) and enzymatic (hNQO1) biomarkers via a single probe. To achieve this goal, we utilized a dual-reactive and dualquenching strategy, which we reasoned would improve the sensitivity and selectivity of the system. 19 Dual-activation probes have recently gained attention due to their ability to ne-tune responses by requiring the presence of two specic analytes. For example, Chang et. al. reported the dual-analyte detection of H 2 O 2 and caspase 8 activity during acute inammation in living mice. 20a Similar strategies have also been used for the successful dual-analyte detection of small molecules. 20b-d Herein, we report the rational design and preparation of H 2 S and hNQO1 dual-responsive uorescent probes 1 and 2, which were successfully utilized to differentiate cancer cells and reveal the synergistic antioxidant effects in response to the oxidative stress.</p><!><p>To enable the simultaneous detection of H 2 S and hNQO1, we installed two chemoselective trigger groups that respond to H 2 S and hNQO1, respectively, into one uorophore. Such dualactivity probes are superior to traditional single-analyte detection probes because they provide specic advantages, including: (1) avoiding inhomogeneous intracellular distribution from different probes; (2) providing an enhanced off-on response due to the dual-quenching effects; and (3) enable a simple method to investigate the cooperative relationship of the analytes.</p><p>To enable access to such dual-responsive probes, we made use of the trimethyl-lock containing quinone propionic acid (Q 3 PA) moiety reported by McCarley's group 18a as the triggering group for hNQO1. For the H 2 S detection motif, we utilized the thiolysis of NBD (7-nitro-1,2,3-benzoxadiazole) amines, 21 which has been utilized by our group as well as others for development of excellent H 2 S probes. Additionally, this H 2 S sensing motif has been used for different biological applications including tumor bioimaging in mice. 9c Therefore, we combined the Q 3 PA and NBD amine moieties onto coumarin and naphthalimide uorophores to access dual-responsive systems. The Q 3 PA moiety can switch off the uorescence of the uorophore by the photoinduced electron transfer (PET) effect, while the NBD part can quench the uorescence through the uorescence resonance energy transfer (FRET) effect. We expected that the uorescence of the coumarin and naphthalimide uorophores would be quenched efficiently from this dual-quenching strategy, and that only dual activation of both the Q 3 PA and NBD motifs would result in uorescence turn-on (Scheme 1).</p><!><p>As outlined in Fig. 1A, the synthesis of probe 1 started from a formylation reaction to generate 3, which was treated with dimethyl malonate to form the coumarin derivative 4. Then, single-reactive probe 6 was synthesized from coupling 4-nitro-7piperazinobenzofurazan (NBD-PZ) and the hydrolysis product 5. Aer N-boc deprotection and further coupling with Q 3 PA, probe 1 was obtained with relative good overall yield. Probe 2 was Scheme 1 Schematic illustration of the design for a dual-biomarkertriggered fluorescent probe, which should only be activated by the synergistic chemical reaction with H 2 S and enzymatic reaction with hNQO1.</p><p>prepared from a simple four-step synthesis from commercial available reagents (Fig. 1B). 4-Bromo-1,8-naphthalic anhydride was reuxed with N-boc-ethylenediamine to produce 8, aer which the piperazinyl group was introduced through a nucleophilic substitution to form 9. Further reaction with NBD-Cl afforded 10, which was then deprotected and coupled with the Q 3 PA motif to provide probe 2 in good yield. All compounds were characterized by 1 H and 13 C{ 1 H} NMR spectroscopy as well as high-resolution mass spectrometry (HRMS) (see ESI †).</p><p>With the probes in hand, we examined the optical response of 1 toward H 2 S and hNQO1 in phosphate buffered saline (PBS, 50 mM, pH 7.4 containing 0.007% BSA, 100 mM NADH). As shown in Fig. S1, † 1 displayed two absorption maxima around 405 nm and 500 nm due to the coumarin and NBD amine moieties, respectively. Aer reaction with both H 2 S and hNQO1, new peaks appeared at 395 and 520 nm, which corresponded to the production of coumarin uorophore and NBD-SH, respectively. 19b Notably, 1 remained water-solubile at concentrations over 25 mM (Fig. S2 †). Prior to activation, 1 was essentially non-uorescent (F 1 ¼ 0.15%) due to the PET-FRET dual-quenching effect. Aer treatment with both hNQO1 (1 mg mL À1 ) and H 2 S (200 mM) for 2 h, a large increase in emission (220-fold) appeared at 465 nm (Fig. 2A). When 1 was treated by H 2 S alone for 2 h, only a 34-fold uorescence enhancement was observed (Fig. 2B), which was far lower than the response from hNQO1 and H 2 S together. When 1 was treated with hNQO1 alone for 2 h, the emission enhancement was negligible (2-fold) (Fig. 2C), implying a more efficient quenching from the NBD moiety in 1. Stability investigations showed that 1 was stable in PBS buffer in the absence of analytes (Fig. 2D). Taken together, probe 1 can be used to detect H 2 S and hNQO1 in tandem, whereas treatment with only one of the analytes resulted in a signicantly smaller response.</p><p>To achieve a more efficient single-and dual-quenching effect, we next assessed the uorescence response of 2 toward H 2 S and/or hNQO1. Emission spectra were also recorded in PBS buffer in the presence of NADH. As shown in Fig. 3A, 2 (F 2 ¼ 0.041%) was essentially non-uorescent due to the dual- quenching effect, but a strong emission at 535 nm was observed when hNQO1 and H 2 S were added simultaneously. Aer 2 h, the uorescence increase at 535 nm was over 400-fold. Consistent with our design, treatment of 2 with hNQO1 or H 2 S alone resulted in only a negligible uorescence enhancement (3-or 7fold, Fig. 3B-D and S3 †). When compared with probe 1, we found that probe 2 not only resulted in a larger uorescence turn-on for combined H 2 S/hNQO1 activation, but also exhibited a lower single-analyte response. Because of these positive properties, we utilized probe 2 for subsequent bioimaging investigations.</p><p>Encouraged by the primary uorescence data, we further validated the chemistry associated with the sensing mechanism by using HRMS and UV-vis analysis. We rst conrmed the products of both the single-and dual-analyte reactions of 2 with H 2 S and/or hNQO1 with HRMS (Fig. 4 respectively. We did not observe the cross reaction sideproducts (e.g. hNQO1-triggered 13 or H 2 S-triggered 12) in the MS spectra. We next performed UV-vis experiments to further probe the reaction mechanism. As shown in Fig. S5A, † the absorption spectrum of 2 displayed two maximum absorbance peaks near 350 and 500 nm. Aer treatment with H 2 S and hNQO1, both of these peaks disappeared and were replaced by peaks at 400 and 520 nm, which corresponded to the uorophore and NBD-SH, respectively. When H 2 S alone was added, new peaks at 400 and 520 nm were also observed (Fig. S5B †). Furthermore, there was an obvious overlap between the absorbance prole of NBD-PZ and the emission prole of 11, indicating an intramolecular FRET effect in probe 2 (Fig. S5C †).</p><p>When 2 was treated by hNQO1 alone, the absorbance peak at 500 nm increased (Fig. S5D †), implying that the PET process was abolished because the PET effect should result in small changes in absorbance spectra. 22 In addition, probe 2 maintained water solubility of over 20 mM under the experimental conditions (Fig. S6 †).</p><p>To gain more detailed information about the sensitivity of the dual-responsive probe, we incubated 2 with different levels of hNQO1 and H 2 S for 2 h, aer which the emission proles were measured. Probe 2 was rst treated with different concentrations of H 2 S (0-200 mM) in the presence of hNQO1 (1 mg mL À1 ). As shown in Fig. 5A and B, the emission at 535 nm was linearly related to the concentrations of H 2 S from 0 to 75 mM. When added to 1 mg mL À1 hNQO1, a 10 mM H 2 S solution resulted in a 46-fold uorescence response. Similarly, we treated 2 with various levels of hNQO1 (0.2-1 mg mL À1 ) in the presence of a constant H 2 S concentration (50 mM), and observed a uorescence enhancement of 180-fold (Fig. 5C).</p><p>One major requirement for a uorescent probe is that it must exhibit a response toward the targeted analytes but not for other competing species. In order to conrm that the turn-on response of 2 was selectively caused by the dual activation of hNQO1 and H 2 S, probe 2 was incubated with different reactive sulfur species (SO 3 2À and S 2 O 3 2À ), biothiols (Cys, Hcy and GSH)</p><p>and reactive oxygen species (H 2 O 2 and HClO) in the presence of hNQO1 or H 2 S. As shown in Fig. 5D, only the co-incubation of hNQO1 and biothiols could trigger a very slight uorescence response (<10-fold, lanes 10-12), which was signicantly lower than the response triggered by hNQO1 and H 2 S (>400-fold, lane 15). No uorescence increase was observed when H 2 O 2 or HClO was added (lanes 6-7 and 13-14). Furthermore, treatment of 2 with dicoumarol, an hNQO1 inhibitor, resulted in a slower reaction rate than the inhibitor-free controls, conrming the requirement of hNQO1 for probe activation (Fig. S7 †).</p><!><p>We rst evaluated the cytotoxicity of 2 in HT29 cells (human colorectal epithelial cancer cells) by using the methyl thiazolyl tetrazolium (MTT) assay. The results showed that aer 2 h of cellular internalization of 33 mM probe, more than 90% of the cells remained viable (Fig. S8 †), implying a low cytotoxicity of 2. The cytotoxicity of 2 was further studied in HEK293A cells (human embryonic kidney cells) by monitoring of adherent cell proliferation through the xCELLigence RTCA system (Fig. S9 †). Compound 2 did not show signicant cytotoxicity from 0-15 mM aer 24 h incubation, and therefore 10 mM of 2 was used for bioimaging experiments. To investigate whether 2 could be employed to distinguish different types of cancer cells, several cell types were chosen as model biological systems. Given the elevated levels of both H 2 S and hNQO1 in some colorectal cancer cells, HT29 and HCT116 cells (human colorectal epithelial cancer cell lines) as well as FHC cells (human normal colorectal epithelial cell line) were initially selected. 9c Then HepG2 cells (human liver cancer cells) with a high level of endogenous H 2 S and HeLa cells (human cervical cancer cells) with a low level of endogenous H 2 S were also introduced. 8 We assumed that only the 2-stained cells with relatively high endogenous levels of both H 2 S and hNQO1 would display signicant uorescence. Aligned with this expectation, the confocal uorescence images showed clearly differentiable responses from the selected cells (Fig. 6A). The uorescence intensity in HT29 and HepG2 cells was much stronger than that in other cell lines. The relative uorescence increases in HT29 and HepG2 cells were about 5.3 and 3.7 fold higher than that of other cells (Fig. 6B). The signicantly different uorescence observed in cancerous versus non-cancerous cells is consistent with the probe design and suggests that the probe is differentially activated in cancerous versus non-cancerous cells.</p><p>In control experiments for single biomarker detection, two single-analyte probes NIR-H 2 S (for H 2 S detection) 9c and NIR-hNQO1 (for hNQO1 detection) 23 developed by us were separately incubated with these cells (Fig. S10 †). As shown in Fig. S11, † when cells were treated with NIR-H 2 S, the HT29, HepG2 and HCT116 cells displayed a uorescence response, implying the existence of endogenous H 2 S in the cells. When cells were incubated with NIR-hNQO1, the observed uorescence from the HT29 and HepG2 cells was stronger than that from the other three cell lines (Fig. S12 †). The results indicated the relatively high endogenous levels of both H 2 S and hNQO1 in HT29 and HepG2 cells, which is consistent with the bioimaging results of probe 2.</p><p>To further conrm the dual-activation of 2 in cancer cells, we added aminooxyacetic acid (AOAA, 200 mM), which is an inhibitor of enzymatic H 2 S synthesis, and dicoumarol (100 mM), which is an hNQO1 inhibitor. For the inhibitor-treated groups, HT29 cells were pretreated with the inhibitor for 30 min, then incubated with 2 (10 mM) for 1 h, washed and imaged (Fig. S13 †). HT29 cells showed strong uorescence aer incubation with 2 alone for 1 h. In contrast, pretreatment of one or two inhibitors led to a signicant decrease in uorescence, and the observed uorescence intensity was about a half of that in the group without inhibitors (Fig. 6C). These results clearly demonstrated the dual H 2 S and hNQO1 requirement for 2.</p><p>Investigation of the crosstalk between H 2 S and hNQO1 under oxidative stress H 2 O 2 , a common ROS, was introduced as a stimulus to investigate the potential crosstalk between H 2 S and hNQO1 in cellular redox homeostasis. HeLa cells were selected as the model biological systems due to the relative low levels of the both endogenous biomarkers. The cells were stained by 2, washed and imaged. As displayed in Fig. 7, 2-stained HeLa cells exhibited very weak uorescence. However, a signicant uorescence response was observed when cells were co-incubated with 2 and H 2 O 2 (50, 100 or 200 mM) for 1 h. To further understand the results, the inhibitors AOAA and dicoumarol were also used for control experiments (Fig. S14 †). The H 2 O 2stimulated cells displayed a signicant uorescence decrease Chem. Sci., 2019, 10, 1945-1952 | 1949 when pretreated with one or both inhibitors. The relative emission (Fig. 8A) showed that the stimulation by H 2 O 2 could trigger about 3.9-fold uorescence enhancement, which was much higher than the inhibitor-pretreated control groups (about 1.8-fold). In addition, aer co-incubation with H 2 O 2 and 2, AOAA-pretreated cells were further treated with Na 2 S (150 mM) for 30 min, and a small increase in uorescence was observed (1.5 fold) when compared with the AOAA-pretreated control group. These data suggest that endogenous H 2 S and hNQO1 could be spontaneously generated in living cells when cells were suffering from acute oxidative stress caused by exogenous H 2 O 2 .</p><p>Based on current knowledge, hNQO1 is regulated by the Keap1 (Kelch-like ECH-associated protein 1)/Nrf2 (nuclear factor-erythroid 2-related factor 2)/ARE (antioxidant response elements) pathway. 10 Nrf2 protein levels can rapidly increase in response to ROS, triggering the expression of hNQO1 to inhibit the formation of free radicals. 15a,24 Meanwhile, elevated Nrf2 can increase the expression of glutathione reductase (GSR), which can reduce GSSG to GSH. 24d,e Such GSH can be involved in the Sglutathionylation of CBS under H 2 O 2 to produce CBS -SG , which would enable more efficient biosynthesis of endogenous H 2 S. 14b,c Thus, we propose that the synergistic antioxidant effect of H 2 S and hNQO1 for handling oxidative stress in living cells is possibly regulated by Nrf2, which can trigger the expression of hNQO1 directly and improve endogenous H 2 S levels indirectly through controlling GSH (Fig. 8B). Taken together, these results support a synergistic antioxidant effect under cellular oxidative stress.</p><!><p>In summary, dual-biomarker-triggered uorescent probes were developed for the simultaneous detection of two potential cancer biomarkers. Probe 1 could detect the two biomarkers with a slight uorescence response toward one biomarker (34fold turn-on) and a signicantly enhanced uorescence by dual activation (220-fold turn-on). By contrast, the uorescence of probe 2 was signicantly enhanced and showed a greater response for the dual-activation from H 2 S and hNQO1 (>400fold turn-on). Moreover, probe 2 exhibited high sensitivity, excellent selectivity and good biocompatibility, which enabled us to differentiate activation levels in HT29 and HepG2 cells from FHC, HCT116 and HeLa cells due to the notably different endogenous levels of H 2 S and hNQO1 in the cell lines. Importantly, using the probe 2, we revealed a synergistic antioxidant effect between H 2 S and hNQO1 in living cells in response to the oxidative stress. These results clearly demonstrate the strengths of this dual reporter system, including the signicant off-on response, ability to distinguish cancer cells with both cancer biomarkers, and ability to investigate the crosstalk of analytes. We also note, however, potential limitations of this system. For example, the developed tools only provide information on the relative levels of the biomarkers in different cell lines rather than precise quantication measurements. In addition, the development of probes with longer wavelength emissions would be needed to translate these systems into more complex systems, such as animal studies. Based on these needs, we are currently working to develop related dual-responsive probes with emission in the near-infrared region for in vivo applications. Overall, our work has demonstrated the research potential of dual-responsive uorescent probes in cancer biology and intracellular redox homeostasis.</p><!><p>The authors declare no competing nancial interests.</p>

Royal Society of Chemistry (RSC)

Automated Peak Detection and Matching Algorithm for Gas Chromatography\xe2\x80\x93Differential Mobility Spectrometry

A gas chromatography\xe2\x80\x93differential mobility spectrometer (GC-DMS) involves a portable and selective mass analyzer that may be applied to chemical detection in the field. Existing approaches examine whole profiles and do not attempt to resolve peaks. A new approach for peak detection in the 2D GC-DMS chromatograms is reported. This method is demonstrated on three case studies: a simulated case study; a case study of headspace gas analysis of Mycobacterium tuberculosis (MTb) cultures consisting of three matching GC-DMS and GC-MS chromatograms; a case study consisting of 41 GC-DMS chromatograms of headspace gas analysis of MTb culture and media.

automated_peak_detection_and_matching_algorithm_for_gas_chromatography\xe2\x80\x93differential_mobil

<!>Case Study 1: Simulations<!><!>Case Study 1: Simulations<!>Case Study 2: Matching GC-DMS and GC-MS<!>Case Study 3: GC-DMS of MTb Cultures and Control<!>INSTRUMENTATION<!>Peak Detection<!>Baseline Correction<!>Alignment<!>Detecting Features in Each Voltagram<!><!>Detecting Features in Each Voltagram<!>Peak Merging<!>Peak Matching<!>Unfolding<!>Case Study 1<!>Case Study 2<!>Case Study 3<!>Shuffling<!>Principal Components Analysis<!>CONCLUSIONS<!>

<p>Today, modern analytical chemistry is dominated by the use of analytical instrumentation, e.g., coupled chromatography for data acquisition. This system is very powerful, producing multidimensional signals with rich sources of information.1,2 In the era of rapid development of analytical instruments, gas chromatography– differential mobility spectrometry (GC-DMS) has marked another milestone for advancement in analytical methods. It is a relatively new technology where chemical substances are characterized based on differences between ion mobilities under high and low electric fields at ambient pressure.3–6</p><p>GC-DMS has been employed in various applications such as arson investigation,7 environmental analysis,8 and disease diagnosis.9–13 This hyphenated system produces data in the form of a matrix whose rows correspond to GC retention times, RT, and columns to DMS compensation field strength, Vc (V/cm). Usually, the GC-DMS data handling involves using a total chromatographic profile where the resultant signal cannot directly be interpreted in terms of specific chemical compounds. Recent publications also revealed another approach for analysis of GC-DMS chromatograms i.e., wavelet transforms.14,15 Peak detection can be another alternative to the traditional approach, but it is laborious when a large volume of data is involved. In fact, peak detection procedures have received considerable attention in many other application areas such as GC-MS,16 GC-GC,17 LC-MS,18 etc. For GC-DMS, it remains unexplored and would certainly be useful for biomarker and pattern recognition study.</p><p>This paper reports an automated peak detection and matching algorithm for GC-DMS. The algorithm is demonstrated in three case studies: a simulated case study; a case study of headspace gas analysis of Mycobacterium tuberculosis (MTb) culture with three matching GC-DMS and GC-MS chromatograms; a case study of 41 GC-DMS chromatograms of headspace gas analysis of MTb culture and media.</p><!><p>Case study 1 consists of 300 matrices, each representing a GC-DMS chromatogram, with dimensions 2000 (corresponding to elution times, represented by i) × 250 (corresponding to compensation field strength, represented by j) each in turn, consisting of between 10 and 40 simulated peaks (the numbers of peaks in each chromatogram being generated using a random uniform distribution).</p><p>The shape of each peak is modeled by a 2D Gaussian function f(φi,φj)=ϑexp(−(φi−φimax)22σi2−(φj−φjmax)22σj2) where ν is the underlying intensity at the peak maximum; φimax, φjmax are the positions of the peak maxima in each dimension; and σi, σj relate to the width of each peak in each dimension.</p><!><p>The peak intensities, ν, are generated using a random normal distribution characterized by an underlying mean and standard deviation of 0.04 and 0.01. Because the mean is four times the standard deviation, no peaks of negative intensity are generated (in the contingency these would be replaced by 0).</p><p>The values of σ are obtained from an underlying random normal distribution with an underlying mean of 8 and standard deviation of 2 for RT dimension and mean of 12 and standard deviation of 3 in data points in the Vc dimension. These peaks are relatively narrow in the RT dimension compared to the Vc dimension.</p><p>The underlying peak maxima φimax and φjmax are obtained using a random uniform distribution between 100 to 1800 data points in the RT dimension and 80 to 200 data points in the Vc dimension.</p><p>Gaussian noise with a mean of 0 and a standard deviation of 0.002 in each 2D data matrix is added to each data point.</p><!><p>These simulated peaks were based on observed peak shapes in the experimentally obtained chromatograms.</p><!><p>Case study 2 consists of three matching GC-DMS and GC-MS chromatograms of headspace gas analysis of Mycobacterium tuberculosis (MTb) cultures which are designated S1, S2, and S3. MTb cultures were prepared in SPME vials by inoculating BACTEC 12B media with MTb strain-2. The cultures were well-defined with a known concentration of 1 × 108 bacilli/mL. The samples were incubated for 24 hours at 37 °C to allow volatiles to accumulate in the headspace of the vials. The extractions were performed for 30 min using a 50/30 μm divinylbenzene/Carboxen coating on a polydimethylsiloxane (PDMS/DVB/Carboxen) SPME fiber. The SPME fibers were purchased from Supelco Inc. (Bellefonte, PA) and the field portable SPME holder, TuffSyringe, from Field Forensics (St. Petersburg, FL). Prior to use, the fibers were conditioned as recommended by the manufacturer in the GC injector port at 250 °C for 1 h and reconditioned for 1 h in between each run to minimize carryover effects. Following extraction, all fibers were UV irradiated to prevent cross-contamination of MTb particles. During the entire process, blank fibers were exposed to the ambient atmosphere to check for extraneous volatile contamination. All fibers were reused following conditioning according to manufacturers' instructions (1 h at 250 °C) with internal confirmation that we did not have carryover by running fibers immediately after conditioning. All fibers were used for a maximal amount of cycles according to the manufacturers' instructions (up to 150 ×) unless they were deemed damaged by visual inspection.</p><p>In this case study, the DMS compensation field strength was scanned from −520 V/cm to +160 V/cm with an interval of 5.35 V/cm. The dimensions of GC-DMS chromatograms were 1044 (elution times) × 128 (compensation field strengths); note that we use compensation field strength19–21 which is a common alternative to compensation voltage. The scan rate in the GC dimension was 1.297 s/scan over 22.5 min. Although both positive and negative ion spectra are available, there is limited information in the negative ion spectra, so we report only results on the positive ion chromatograms in this paper. For the matching GC-MS chromatograms, the dimensions were 7066 (elution times) × 262 (mass numbers) with a scan rate of 0.19 s/scan (0–22.5 min) over the mass range of 39 to 300.</p><!><p>This case study consists of 41 GC-DMS chromatograms of which 18 were media (control) and 23 were MTb cultures. The MTb cultures were prepared using the media described of case study 2, and the media samples were control consisting of media alone. Both sets of samples were incubated for 24 hours at 37 °C to allow volatiles to accumulate in the headspace of the vials. The volatiles were extracted from the headspace of the vials by solid-phase microextraction, and further details are provided in the previous section. The aim of the MTB culture and media experiments is to be able to distinguish controls from inoculated cultures. Samples were analyzed within 5 days of extraction.</p><p>The DMS compensation voltages were scanned from −520 V/cm to +160 V/cm. The dimensions of GC-DMS chromatograms were 1044 (elution times) × 128 (compensation voltages) with a scan rate of 1.297 s/scan over 22.5 min over a compensation voltage range of −520 V/cm to +160 V/cm with an interval of 5.35 V/cm. As in case study 2, only the positive ion chromatograms are used.</p><p>The MTb and media samples were run in a randomized order. The instrumental drift and sample carryover was monitored by running a blank (22.5 min) followed by a VOC (volatile organic compound) standard (Supelco Inc., Bellefonte, PA; Supelco Inc., Bellefonte, PA) at the beginning and end of each day.</p><!><p>All experiments were performed on a duel detector Cyro-GC system consisting of a Agilent 6890N (Agilent Technologies, Palo Alto, CA) gas chromatograph interfaced to a differential mobility spectrometer (Model SVAC-V, Sionex Corporation, Bedford, MA) and Agilent 5975 quadrupole mass spectrometer (Agilent Technologies, Palo Alto, CA). The Cryogenic Trap Enrichment System (CTE) was purchased from GERSTEL (Baltimore, MD).</p><p>GC was carried out on a Rtx-200MS (trifluoropropylmethyl polysiloxane) (Restek Corporation, Bellefonte, PA) wall-coated open tubular column (30 m × 0.32 mm i.d., 1 μm film thickness). The GC injector was operated at 250 °C. The samples were thermally desorbed in splitless mode for 2 min with a purge delay of 2 min. The front of the GC column was cooled cryogenically with liquid nitrogen to −125 °C for 2 min and ramped at 20 °C/s to 240 °C. The GC oven was programmed from 50 °C (2 min hold) increased to 170 °C (3 min hold) at 4 °C/min and then to 230 °C (3 min hold) at 20 °C/min. The GC carrier gas was helium at a flow rate of 2 mL/min. Data was acquired between 0 and 22.5 min. Data after 22.5 min was excluded because the compounds coming off at these higher temperatures were from column bleed.22,23</p><p>A Y-connector (Restek Corporation, Bellefonte, PA) was used to split the column eluent to MS and DMS. The MS operated under vacuum pressure but the DMS operated at atmospheric pressures, so care was taken to ensure the analyte reach both detectors simultaneously by adjusting the length of the transfer lines from the Y-connector. To account for the differences in operating conditions, the total length of the transfer line from the Y-connector to the MS and DMS was 0.5 and 1 m, respectively. A differential mobility spectrometer with an electrode gap of 0.5 mm was used as the GC detector. The DMS was operated at a dispersion voltage of 22 kV/cm. The compensation voltage was scanned over a range of −520 to +160 V/cm with a step duration of 10 ms and a 2 ms step settle time. The scan duration was 1.297 s. The makeup drift gas for the DMS was nitrogen at a flow rate of 400 mL/min. The DMS sensor was operated at 85 °C. The part of the transfer gas line that was exposed to the atmosphere was heated to 180 °C to prevent sample condensation along the line. The mass spectrometer was equipped with an electron impact ionization source operated at 150 °C. The quadrupole was operated at 230 °C.</p><p>The GC-DMS data in Excel worksheets was converted to Matlab version 7.0 (The Mathworks, Inc., Natick, MA) into data matrices with rows correspond to GC retention times (RT) and columns to DMS compensation voltages (Vc). The matching GC-MS chromatograms in netcdf format were converted to Matlab version 7.0 (The Mathworks, Inc., Natick, MA) as matrices with dimensions 7066 × 262 with a scan range of m/z 39 to 300 and a scanning rate of 0.19 s/scan. The GC-MS chromatograms were subjected to the automated peak detection algorithm reported by Dixon et al.16</p><p>All software in this paper was developed in-house in Matlab.</p><!><p>A table of notation is presented in Table 1 listing the symbols and variables used in this paper. The schematic of the overall peak detection algorithm is illustrated in Figure 1.</p><!><p>The original matrix, X of dimensions I × J where I refers to retention time scans (RT dimension) and J to voltage scans (Vc dimension), is baseline corrected and aligned yielding a matrix U. For baseline correction, each column vector xj is divided into windows of 100 points and 10% intensity quantile of each window is determined. These points are then modeled with Piecewise Cubic Hermite Interpolating Polynomial (PCHIP) function.24 The baseline is then subtracted from the data.</p><!><p>The peak alignment procedure is based on that described by Krebs et al.;25 the details are not discussed for brevity. In summary, a reference chromatogram is chosen automatically using discrete coordinates simplex-like optimization routines;26 the remaining chromatograms are aligned according to the reference. This procedure aligned data in two dimensions using a rigid shift in the Vc dimension, allowing a flexible shift in RT dimension. The data is shifted based on the maximum cross-correlation value for Vc dimension, and the RT dimension is aligned with respect to the reference by identifying the common landmarks (found in both sample and reference) and interpolating according to piecewise linear function. The resulting matrix, U, is subjected to further analysis.</p><!><p>The key to feature detection is to find features at each voltagram obtained at each RT scanned. Matrix U is first processed with a quadratic Savitzky-Golay 5-point first derivative filter,27,28 both horizontally and vertically, producing matrices D and E of dimensions (I × (J – 4)) and ((I – 4) × J), respectively. The next step involves computing a value, t, at each RT i by taking the average absolute change of derivative over the trace as follows.29 ti=∑j=1J−3∣dij−di(j+1)∣J−4</p><!><p>The peak detection algorithm examines each row, i, of matrix D in turn. For each vector, the algorithm proceeds iteratively one point at a time from j = 2,.. .,(J – 3). A feature start is identified when (dij > 0), (di(j+1) > dij), and (dij > ti a): ti is defined as above, and a is a user-defined peak noise factor. The threshold at RT i is ti × a. The peak start is denoted by j = m. The variable a is a tunable noise level multiplier used to determine the potential signal threshold: in this paper, a is set at 5. Only peaks whose intensity are a-fold more than ti are detected; some features in row vectors of matrix D might be rejected especially those at the edges of a two-dimensional peak.</p><p>Next, the algorithm searches for a maximum for each feature; the maximum satisfies both (dij < 0) and (di(j+1) < 0), defined by j = r. For a perfect peak shape, the center corresponds to the zero-crossing point in the derivative where the signal crosses the x-axis going from positive to negative (dij = 0).</p><p>The algorithm continues to detect the end of the feature which is found when (dij > 0) and (di(j+1) > 0). At the end of feature, j = s.</p><p>The maximum of the peak between the start and the end is determined; the number of data points the maximum is away from s is given by M (equivalent to the left half width of the peak).</p><p>Each potential peak (or feature) detected in the voltagram is a candidate to be part of a true 2D peak which should consist of several features at successive RTs.</p><p>In order to determine whether this feature is a component of a true peak or an artifact, there must be a corresponding feature in the chromatographic dimension at column variable j = r. The algorithm then searches for features in the chromatographic dimension using matrix E. The aim is to take each feature in row i corresponding to a single data point to see whether there is a corresponding feature in the second dimension. If so, this feature is potentially part of a 2D peak, whose start and end in the RT dimension can be defined.</p><p>The next stage of the algorithm is to locate the feature start and end in the RT dimension. The row variable i is denoted by u. The start is characterized by (eur > 0) and (e(u–1)r < 0) where u = i – 1,i – 2,.. .2 and the end is the point where (eur < 0) and (e(u–1)r > 0) where u = i + 1,i + 2,.. .I – 3. The positions of the start and end (RT dimensions) are denoted by g and h, respectively. If a feature is not found, the candidate peak is rejected.</p><p>If the candidate feature is accepted, the algorithm continues to search for another peak start in the voltagram at RT, i, from position j = s until j > (J – 3).</p><p>The positions of feature start, maximum, and end determined from both dimensions are presented in a 'peak detail table', Y = a matrix of dimensions (P × Q) where P is the number of features found over the entire voltagram and the columns (Q = 7) describe the characteristics of the features (Table 2). Note that at this stage all that has been done is to identify which features in each RT voltagram are part of potential peaks and to identify the start and end (in time) for each of these features.</p><!><p>The peak detection algorithm is illustrated in Supporting Information Figure S-1.</p><!><p>The next step is to put together features at successive RTs to obtain a 2D peak characterized by a region in both dimensions. The peaks list in Y are one-dimensional peaks which do not yet represent peaks in two dimensions; a peak merging algorithm is used to cluster the peaks into respective regions.17 The peak merging algorithm produces a 'peak region table' (Table 3), Z (R × S) where R denotes the number of 2D peaks and S = 7.</p><p>8. The peak merging algorithm compares the first feature (yc where c = 1) with the next, the second lists in Y (yi where i = 2).</p><p>9. The first feature is a 'target' suggesting a start of a peak in the RT dimension. A true peak consists of several features at successive RTs. The algorithm searches for features eluting at successive RTs after the target. If the difference between yij and ycj (j = 1) is ≤ Kinit (initially set to 2), this implies that the features are found at successive RTs and can be merged to form part of a 2D peak (in Table 2, for example, (yij – ycj) = 600 – 599). The parameter for finding the neighboring features, Kinit, begins from 2 (data points) as some features at the edge could be missed during peak detection when noisy chromatograms are involved.</p><p>10. Some features may be found at successive RTs; however, they do not represent the same 2D profile. For example, a target is found at row 100 (RT data point) eluting between columns 120 and 130 in the Vc dimension and a successive feature at row 101, between columns 200 and 210, although adjacent in rows are well apart in the Vc dimension suggesting they originate from two different peaks. For this reason, it is necessary to confirm whether features found at successive RTs are from the same two-dimensional profile using the overlap ratio, ϕ. The overlap ratio is calculated as p/[(q1 + q2)/2] where q1 and q2 are the length (in data points) of the target and the candidate merging peak while p is the overlapping region. The calculation is modified from that reported by Peters et al.17 (overlap ratio = p/q1). The stability of the algorithm is believed to be improved with the modification. As an example, two peaks in the Vc direction eluting at different RTs are compared, the top (high RT) one between data points 1–10 and the bottom (low RT) between data points 6–11. According to Peters's equation, the top-to-bottom configuration yields an overlap ratio of 5/10 but the bottom-to-top configuration is 5/6. The overlap ratios for both configurations are reasonably different. With the modified calculation, the overlap ratios from both directions are maintained at 5/8; in addition, the overlapping region is compared to an average peak rather than relying completely on the edge peak. In this paper if ϕ ≥ 0.7, the peaks are considered to originate from the same 2D peak and are merged.</p><p>11. The algorithm next constructs a matrix, W, to record the one-dimensional features that represent the same 2D profile, in this case, row vectors yc and yi are the input of W. Row vector yi is then discarded from Y.</p><p>12. The comparison is performed with a new yi where i = 2 and Kinit = Kinit + 1. For every successive RT, the data point that satisfies the neighboring and overlap criteria, yi is added to matrix W and is subsequently discarded from Y; Kinit is increased by 1 each time. If the neighboring criterion is not obeyed, the algorithm continues to search for peaks in Y that is characterized by a higher row, maintaining the value of Kinit. If no further feature is found that can merge with a target, a 2D peak is considered fully described in both dimensions.</p><p>13. The algorithm then evaluates matrix W to determine the edges of the 2D peak. The limits of the peak are reported as the median of the one-dimensional features (starts and peak ends) forming the 2D peaks in both Vc and RT dimensions. The peak maximum is designated as the data point with the highest intensity within a peak region, and the peak area is reported as the sum of intensities for all data points within a peak region. An example is presented in Table 2 and Table 3 where two peaks are obtained from the peak detail table, Y, and the peak regions are recorded in the peak region table Z.</p><p>14. The target yc is discarded from Y. A new yc is used as a target and Kinit is reset to 2.</p><p>15. Matrix Y is reduced until all peaks are assigned to their respective two-dimensional profiles.</p><p>The flow diagram of the peak merging algorithm is depicted in Supporting Information Figure S-2.</p><p>In order to protect against an atypical edge, each peak in the peak detail table, Y, is evaluated from both ends: top-to-bottom and bottom-to-top (in the RT dimension). For each configuration, the mean differences in Vc are calculated for all 2D peaks found. Supporting Information Figure S-3 illustrates a sample consisting of seven one-dimensional peaks where the peak maximum of each voltagram is denoted by β1 to β7. The peak merging algorithm finds two features when the peak detail table is examined from top-to-bottom; the reverse configuration (bottom-to-top) however identifies three features. The mean difference for each 2D feature (only those involving more than one one-dimensional peak), δ, is calculated as δ=∑j=1ω−1∣βj−βj+1∣ω−1 where ω is the number of one-dimensional features forming a 2D peak. The overall mean differences for both configurations are compared; the configuration that gives the lower overall mean difference is preferred.</p><!><p>The third step is to determine which peaks in different DMS analyses originate from the same compound. This is done by seeing whether RTs and Vcs of successive peaks are within a given tolerance window, and if so, they are considered to have the same chemical origin.</p><p>For N samples, the algorithm above results in creation of N peak region tables, Z. To determine the number of unique peaks over all samples, a peak matching algorithm is applied.</p><p>16. Each peak in each sample is characterized by its RT and Vc in scan numbers.</p><p>17. Peaks found in all samples are extracted onto a matrix, L (T × 4 where T is the total peaks found over all N samples). The first column represents the origin of the peak according to sample 1,.. ., N, columns 2 and 3 detail the RT and Vc of the peak maxima, and the last column details the peak area.</p><p>18. Matrix L is sorted according to RT (column 2) in ascending order.</p><p>19. A sparse matrix, H with N columns is constructed and the tolerance window of the RT and Vc, V1 and V2, are determined.</p><p>20. The first peak in L is compared to the subsequent peaks found in different samples, i.e., the first peak in L is found in sample 1, the second and the third peaks are detected in samples 2 and 1, respectively. The algorithm compares the first peak with the second peak but comparison with the third is ignored.</p><p>21. Peaks differing within the allowable windows are considered originating from the same source and the peak area are placed in matrix H at columns corresponding to the sample number. The average characteristics RT and Vc are recorded in matrix F.</p><p>22. If more than one candidate matching peak from a sample is detected, the algorithm selects the one that has closer Vc and RT relationship to the target peak.</p><p>23. When relevant peaks are matched, the target peak and the matching peaks are removed from L. The algorithm repeats with a new target peak listed in L (i = 1) until all peaks are matched. Matrix H is transposed to give a peak table. This method finds matching peaks according to an ordered list of L.</p><p>The peak table H is of dimensions N × M where M is the number of detected peaks. All case studies are subjected to peak detection with the tunable parameters a = 5; Kinit = 2 and ϕ = 0.7. For case study 3, the tolerance window of RT and Vc, V1 and V2, are 20 scans (25.94 s) and 7 scans (37.45 V/cm), respectively.</p><!><p>The unfolding method30 is used to compare the 2D peak table obtained from the 41 = N samples of case study 3 with the raw GC-DMS chromatographic profiles. This method is conventionally used for analysis of GC-DMS chromatograms. In this study, each chromatogram is baseline corrected as discussed in Baseline Correction, and the negative values attributable to instrumental noise are replaced with 0s at the outset. Small misaligments by a few data points can have adverse consequences for pattern recognition, so it is usual to average the intensity over a small window of data points, called a bucket, to remove the influence of small RT shifts. The bucketed time window was varied between 5 and 30 scans, and the window that gives the best separation PCA scores plot was considered optimal and found to be 25 data points. The chromatograms are then bucketed in the RT direction at a window of 25 data points (32.43 s) and column-centered prior to formation of the 3D array (the 42nd bucket consisting of data points 1026 to 1044). After that, the three-dimensional array is unfolded into a 2D matrix, denoted B which is of dimensions N × C or 41 × 5376 where 5376 = 128 (Vc) × 42 (RT buckets) = C or the number of columns in the unfolded data matrix.</p><!><p>The algorithm was applied to 300 simulated data matrices. A graph of the number of peaks detected using the algorithm versus the underlying number of peaks in each simulation is presented in Figure 2a and has a correlation coefficient squared of 0.966. Figure 2b is of the estimated peak intensity versus the true peak intensity for all peaks that were correctly detected using the algorithm and the correlation coefficient squared obtained is 0.927. The maximum difference in number of peaks detected, ∣(R−R^)∣, is 9 and the minimum difference is 0 peaks with the mean difference being 2.21 peaks. The mean of all true peak intensities is 0.0416, and the root mean square (RMS) error in the estimation of the intensity is 0.0045, corresponding to 10.82% of the overall mean.</p><p>A DMS analyzer often has a limited resolving power in comparison with MS; the peaks are generally broad and overlapped. Nevertheless, the performance of the peak detection algorithm is comparable to the algorithm reported by Dixon et al.16 for GC-MS where the correlation coefficient squared between the estimated numbers of peaks versus the true number of peaks reported is 0.9025.</p><!><p>In this case study, we compare the peaks detected using GC-MS and GC-DMS. For GC-DMS, the identities of the compounds found at specific RT and Vc cannot be identified without the use of standards. It is anticipated that there will be peaks detected in common using both techniques,31 but due to the different detectors we do not expect to be able to detect all peaks by both methods. Peaks of GC-MS and GC-DMS with a RT difference of less than 10 s were matched for each individual sample. The topological plots of S1, S2, S3 and the total ion chromatograms (TIC) of GC-MS are shown in Figure 3 between 0 and 22.5 min with red dashed lines indicating the matching peaks using a 10 s RT shift criterion. For the GC-DMS chromatograms, the peaks identified with our algorithm are circled in yellow; the centers of the peaks matched with the MS analysis are in addition marked with filled yellow circles representing the positions of the peak maxima. For MS chromatograms, the peaks detected are indicated using black arrows.</p><p>Visual observation suggests that the algorithm described above can effectively identify peaks eluted in the GC-DMS chromatograms, with most of the peaks detected in GC-MS chromatograms also detected using GC-DMS. Nevertheless, there are some unmatched unique peaks using both DMS and MS detectors, suggesting differences in detection abilities between detectors. DMS offers resolution in the compensation voltage dimension so that compounds eluting at similar RTs with differing Vc can be distinguished; this is a consequence of formation of multiple ion peaks as a result of fragmentation of parent molecules or clustering of ionized species.9 However, it is observed that peaks detected using MS at earlier RTs (<5 min) are not found in the DMS chromatograms. Subsequent analysis of the MS data revealed that early eluting compounds are less susceptible to detection using DMS, as these compounds with low proton affinity could exhibit low spectral intensity.8,32</p><!><p>The peak detection algorithm can be tested on case study 3 by seeing whether the MTb culture samples can be distinguished from the media samples, using a peak table, and how this offers an advantage over the using of unfolded (raw) data.</p><!><p>The algorithm produces a peak table of dimensions 41 × 102 with the rows corresponding to samples and columns to variables. The repeatability of the peak table is assessed by shuffling the original order of the chromatograms and performing the peak matching on newly shuffled data. The shuffle test is used to see how close the peak table of the shuffled data correspond to that of the unshuffled data. In addition, it can be used to demonstrate that target peaks are extracted irrespective of the order of the list.16 In this study, the list of peaks is shuffled for 100 iterations and is subjected to independent peak matching. Using the original order of the GC-DMS chromatograms, 102 unique peaks are detected. When changing the order, the mean number of peaks detected is 100, suggesting that the order of the chromatograms is not significant. This is an important test of the robustness of the algorithm.</p><!><p>When using a peak table, H, the data is square rooted and the columns of the data matrix are standardized. Square rooting aims at reducing the influence of elements with high concentrations and to deal with the heterosedastic noise.33–35 Standardization however ensures that each variable (representing a chromatographic peak) has a similar influence.33 Standardization involves mean centering and diving by the population standard deviation.33,36,37 The unfolded data matrix B is column centered but not standardized or square rooted. These choices are made by visual examination of the PC scores plot.</p><p>Principal component analysis (PCA) is performed using the NIPALS algorithm,38,39 decomposing the data into scores T of dimensions N × A and loading matrices P of dimensions A × M or A × C (unfolded data matrix) and A is the number of PCs. For graphical visualization in 2D, we set A = 2.</p><p>The scores plot (Figure 4a) demonstrates that MTb and media samples can be distinguished using the peak table where four media samples are likely to be misclassified. The scores plot of the unfolded data matrix however indicates that MTb and media samples are less distinguishable because peaks of GC-DMS are often broad and overlapped (Figure 4b):40 unfolding the entire chromatographic profile could result in mismatching RTs between chromatograms, as they are not stable, whereas peak picking takes this problem into account.</p><!><p>Peak detection (and integration) is one of the key steps in the analysis process especially in metabolomic studies.41 As illustrated in Figure 3, the algorithm is able to detect most of the characteristic peaks, but there are limitations dependent on the choice of tunable parameters as follows.</p><p>The number of detected peaks is dependent on correct choice of the detection threshold, ti × a. If a is too low, there may be some false positives, whereas a high value can lead to false negatives. The peak detection method is based on the first derivative. Hence, problems may arise with closely eluting peaks, especially when there is only a shoulder between two coeluting peaks. In addition, the peaks are relatively narrow along the time axis. If a chromatographic peak is only represented by 3–5 data points in the RT dimension, it will not be recognized: this is because a peak identified in the Vc dimension is confirmed as a true peak only if a corresponding feature is found in the RT dimension.</p><p>Also, the one-dimensional peaks identified in both dimensions are merged to represent a two-dimensional feature based on the overlapping ratio, ϕ. A low value of this ratio may result in two overlapping peaks to be identified as a single 2D feature.</p><p>Third, there is a risk of mismatching if the tolerance shifts V1 and V2 are not chosen carefully, as these reflect the user's estimate of how peak positions shift in the chromatogram.</p><p>As in almost all published approaches, the performance of the peak detection method is parameter dependent. The limitations of the two-step peak detection method (detecting peaks in a one-dimensional form and merging the peaks to represent a two-dimensional feature) are further described by Vivó-Truyols and Janssen.42 Although automated peak detection methods are rarely as effective as visual identification of peaks, the increase in the size of analytical data often makes manual investigation infeasible. When complex chromatograms are analyzed, the number of peaks can easily exceed thousands. If there are 100 peaks in 50 chromatograms, this would involve 5000 visual checks which at 5 min per check would take a total of 416 h, hence the need for automated methods.</p><p>Although there have been recent reports of using GC-DMS in the literature,7–10,12 these primarily involve looking at whole profiles rather than individual peaks. In this paper, we have reported an automated peak detection and matching algorithm which we demonstrate on three case studies. The method demonstrates promising efficiency on the simulated case study. The parallel setup further allows verification of the algorithm; the results suggest that peaks can be effectively identified in the GC-DMS chromatograms and most of them are also detected using GC-MS despite some differences in detection abilities between detectors. The peak detection algorithm is automated, allowing pattern recognition study where necessary.</p><p>GC-DMS has a great potential to be used for disease diagnosis, offering a more cost-effective alternative to GC-MS. The availability of an automated approach for deconvoluting the GC-DMS chromatograms would allow automated interpretation of the data if the technique is used in routine field analysis.</p><!><p>Supporting Information. Additional information as noted in the text. This material is available free of charge via the Internet at http://pubs.acs.org.</p><p>Peak detection overview.</p><p>(a) Number of detected peaks versus the true number of peaks. (b) Estimated peak intensity versus the true peak intensity for simulations of case study 1.</p><p>The topological plots of S1, S2, and S3 with the TIC of the matching GC-MS chromatograms where the dashed lines indicate the matching peaks using a 10 s tolerance.</p><p>The scores plots of PC2 against PC1 for case study 3, with samples indicated according to groups (MTb cultures and media) for (a) peak table and (b) unfolded data matrix.</p><p>Notation</p><p>Example of Feature Detail Table, Y: Two Peaks Are Illustrated (the values are data points in the Vc and RT dimensions)a</p><p>The peaks separated by a horizontal space indicate a peak in dimensions. The italic cells are selected and transferred onto peak region table, Z, indicating the border of a peak in dimensions.</p><p>Vc dimension.</p><p>RT dimension.</p><p>An Example of Peak Region Table, Z</p>

PubMed Author Manuscript

Antibacterial and Anticancer Potentials of Presynthesized Photosensitive Plectranthus cylindraceus Oil/TiO2/Polyethylene Glycol Polymeric Bionanocomposite

The present study is concerned with the fabrication of the bifunctional Plectranthus cylindraceus oil/TiO2/polyethylene glycol polymeric film for antibacterial and anticancer activities. The suggested film is based on the utility of naturally extracted P. cylindraceus oil in the formation of the polymeric bionanocomposite film decorated with TiO2 nanoparticles. The bionanocomposite film was fabricated by incorporating 15 w% of P. cylindraceus oil with 10 w% polyethylene glycol and 5 w% TiO2 nanoparticles. The active components of P. cylindraceus oil were verified using gas chromatography coupled with mass spectrometry (GC-MS). The surface morphology of the resulted bionanocomposite film was characterized by various spectroscopic and microscopic techniques. The antibacterial potential of the fabricated bionanocomposite film was investigated against four pathogenic strains. The obtained results revealed excellent sensitivity against the bacterial strains, particularly E. coli and S. aureus, with minimum inhibitory concentration 320 µg mL−1 and minimum bactericidal concentration 640 and 1280 µg mL−1 for E. coli and S. aureus, respectively. Polymeric bionanocomposite exerted significant cytotoxicity against human lung carcinoma cell lines in a concentration-dependent manner with an IC50 value of 42.7 ± 0.25 μg mL−1. Safety assessment test against peripheral blood mononuclear cells (PBMCs) demonstrated that the bionanocomposite is nontoxic in nature. Bionanocomposite also showed potent photocatalytic effects. Overall, the results concluded that the bionanocomposite has expressed scope for multifaceted biomedical applications.

antibacterial_and_anticancer_potentials_of_presynthesized_photosensitive_plectranthus_cylindraceus_o

1. Introduction<!>2.1. Instrumentation<!>2.2. Chemicals and Reagents<!>2.3. Botanical Material<!>2.4. Bacterial Strains and Cancer Cell Line<!>2.5. Extraction of P. cylindraceus Oil<!>2.6. Gas Chromatography-Mass Spectrometry (GC-MS) Analysis of Essential Oil<!>2.7. Preparation of TiO2 Nanoparticles<!>2.8. Preparation of the Polymeric P. cylindraceus Oil/TiO2/PEG Bionanocomposite<!>2.9. Characterization of TiO2 Nanoparticles and P. cylindraceus Oil/TiO2/PEG Bionanocomposite<!>2.10. Thermal Stability and Hydrolytic Degradation of the Bionanocomposite<!>2.11. Antibacterial Effect of P. cylindraceus Oil and Bionanocomposite<!>2.12. Determination of Bacteriostatic and Bactericidal Concentrations<!>2.13. Morphological Study of S. aureus and E. coli (SEM)<!>2.14. Anticancer Effect of P. cylindraceus Oil and Bionanocomposite<!>2.15. Acridine Orange/Ethidium Bromide (AO/EtBr) Staining<!>2.16. Dichloro Dihydro Fluorescein Diacetate (DCFH-DA) Assay<!>2.17. Photocatalytic Effect of the Bionanocomposite<!>3.1. Chemical Composition of P. cylindraceus Oil by GC-MS<!>3.2. Characterization of TiO2 Nanoparticles and Bionanocomposite<!>3.3. Thermal Stability and Hydrolytic Degradation of the Bionanocomposite<!>3.4. Antibacterial Effect of the Bionanocomposite<!>3.5. MIC and MBC of the Polymeric Bionanocomposite against S. aureus and E. coli<!>3.6. Morphological Study of S. aureus and E. coli<!>3.7. Anticancer Effect of the Polymeric Bionanocomposite<!>3.8. Antiproliferative Effect of the Polymeric Bionanocomposite<!>3.9. Bionanocomposite-Induced Apoptosis via ROS-Mediated DNA Damage<!>3.10. Estimation of Loss of Mitochondrial Membrane Potential<!>3.11. Polymeric Bionanocomposite-Induced Apoptosis in A549 Cells<!>3.12. Safety Determination of the Bionanocomposite in PBMC<!>3.13. Biocompatibility of the Bionanocomposite<!>3.14. Photocatalytic Potential of the Bionanocomposite<!>4. Conclusion

<p>Nanocomposites provide attractive and cost-effective thin layers with superior features for biomedical and microelectronic applications. In recent times, nanocomposites synthesized from natural biopolymers or biological resources have attained much attention. The biodegradable polymers in combination with noble metal nanoparticles offer bionanocomposites with exceptional biomedical and environmental applications [1–3]. These bionanocomposites are potentially biodegradable and versatile, and their polymeric derivatives can be prepared from a range of renewable resources such as hydrocarbon- and oxygen-rich monomers [4, 5]. Recently, the best approach for upgradation of polymers is the formation of polymeric bionanocomposites, which enhance their scope of applications. The incorporation of nanosized inorganic metallic elements in the polymers dramatically enhanced the physicochemical, mechanical, electrical, thermal features, gas barrier property, stiffness, electrical conductivity, and dimensional stability in comparison to normal polymers or conventional composites [6]. Bionanocomposites are made up of a natural polymeric matrix, renewable resources obtained from the polymer matrix, and organic/inorganic components with one or more dimensions on the nanometer scale [7]. The bionanocomposites are derived from several natural polymeric resources such as starch, cellulose, chitosan, polylactic acid, and essential oil derivatives [8–13]. Among natural resources, plant-derived essential oils have been identified as an outstanding natural biomaterial with extraordinary bioactivity, biocompatibility, biodegradability, and nontoxicity containing multifunctional groups [14, 15].</p><p>Polyethylene glycol (PEG) is an extensive antifouling hydrophilic polymer with various biomedical and industrial applications [16, 17]. The introduction of new functionalities by surface coating of PEG on different nanoparticles, proteins, and substrates enhances the biocompatibility of the host materials. Surfaces coated with PEG are well described for their antifouling ability to restrict protein adsorption, microbial attachment, and cell adhesion [18]. The antifouling property of PEG and PEG-conjugated proteins can also lower the external immune response, increase blood circulation time, and enhance the proteolytic, mechanical, and thermal stability [19]. The performances of the sustainable resources-derived polymers are further improved by the addition of different conventional nanofillers, including calcium carbonate, clays, talc, kaolin, fumed silica glass fibers carbon nanotubes, noble metal, and metal oxides [20, 21]. The resultant biocomposite unites the benefits of low-dimensional layers with a vast surface area of nanoparticles, leading to diverse useful applications in the biomedical science and the manufacturing industry.</p><p>Among various metal and metal oxides, titanium oxide nanoparticles possess unique features and interesting biological properties. Titanium oxide (TiO2) easily combines with the polymers, relatively stable and strong with respect to both physical and chemical properties. It is an important semiconducting material with a crystalline structure and displayed remarkable applications, including photocatalysis, solar energy conversion and field emission emitters, environmental pollution control, chemical inertness, excellent mechanical properties, nontoxicity, high refractive index, low cost, high-temperature superconductors, gas sensors, batteries, strong oxidizing power, high hydrophilicity, intense UV absorption, and strong antibacterial activity against numerous pathogens, including bacteria, fungi, algae and virus [22]. It has been frequently used in food, food packaging, cosmetics, optical devices, paints, coatings, and specifically antimicrobials due to their strong effects towards photocatalytic disinfection [23]. TiO2 exists in anatase (tetragonal), rutile (tetragonal), and brookite (orthorhombic) crystalline polymorphs. Rutile is the most stable form and possesses good scattering effects and has been extensively used in pigments to protect the materials from hazardous UV light [24]. In recent times, the use of TiO2 nanoparticles in photodynamic therapies has been constantly increasing. TiO2 nanoparticles and their composites and hybrids with other molecules are frequently used in the photodynamic inactivation of bacteria resistant to antibiotics and photosensitizing agents in cancer treatments [25]. TiO2 nanoparticles were employed as inter alia in the formation of bioconjugates with specific-cell monoclonal antibodies for malignant tumor therapies or synthesis of black TiO2 nanoparticles for the antimicrobial therapy against antibiotic-resistant microorganisms [26, 27].</p><p>Nowadays, numerous studies have been addressed in the development of bio-derived TiO2-polymeric nanocomposites for diverse applications [28, 29]. However, limited scientific reports are available on the antibacterial and anticancer potential of TiO2-reinforced biopolymers [30–32]. The photocatalytic effect of biodegradable aliphatic polyester films incorporated with TiO2 particles was screened for their competence for the removal of volatile organic components [33]. A study conducted by Tang described the preparation of polycaprolactone-TiO2 nanocomposites by a solution-casting method and evaluated their mechanical and self-cleaning properties [34]. A biopolymer prepared from waste cooking oil doped with TiO2 fillers was reported as a surface coating for indoor and outdoor building applications [35]. Nevertheless, the aforementioned studies have not been screened for the antimicrobial effects of the film. Few reports are available in the literature with regard to the synthesis of TiO2 coated oil-based polymeric nanoparticles that were evaluated for antibacterial activity. Recently, TiO2 nanoparticles loaded polyethylene films have shown a potential antimicrobial effect against Staphylococcus aureus, suggesting that PE/TiO2 composite film might prevent the high relative humidity throughout the packing system, delaying spoilage of food and thus increasing their shelf life. Hence, these films can serve as the best candidate for the food packaging system [36]. Another study conducted by Feng et al. revealed that PLA/TiO2 nanofibers and films prepared by electrospinning and solution casting methods exhibited strong antibacterial effects against Escherichia coli and S. aureus. The results of this study concluded that both the nanocomposite membranes satisfied the requirements of food packaging materials [37].</p><p>The present study reports the synthesis and characterization of P. cylindraceus oil/TiO2/polyethylene glycol polymeric bionanocomposite comprised P. cylindraceus oil and polyethylene glycol decorated with anatase TiO2 nanoparticles. The synthesized bionanocomposite was evaluated for antibacterial and anticancer activities against various pathogenic strains and cancer cell lines.</p><!><p>All spectrophotometric measurements and the characterization of the synthesized nanoparticles were carried out by measuring the UV-Vis spectrum of the as-prepared TiO2 nanoparticles using an Ultrospec 2100-Biochrom spectrophotometer (Biochrom Ltd., Cambium, Cambridge, UK). Fourier-transform infrared (FTIR) spectra of the formed nanoparticles and their bionanocomposites were recorded using PerkinElmer FT-IR spectrophotometer (PerkinElmer Ltd., Yokohama, Japan). The morphologies of TiO2 NPs and the polymeric P. cylindraceus oil/TiO2/PEG bionanocomposite were measured by scanning electron microscope (JEM-2100F, JEOL Ltd., Akishima, Tokyo, Japan) and JEM-1400 transmission electron microscope (JEOL Ltd., Akishima, Tokyo, Japan). XRD patterns of TiO2 nanoparticles and the bionanocomposite were obtained by Siemens D-5000 diffractometer (Siemens, Erfurt, Germany). The pH-meter Metrohm model 744 (Metrohm Co., Herisau, Switzerland) was used to control the pH condition of the test solution. Distilled water (H2O) was used throughout the experimental study. Thermal stability analysis was conducted using a Shimadzu thermogravimetric analyzer (TGA-502 model). Nikon ECLIPSE fluorescence microscopy (Ti-E, Japan) was used to analyze cancer cell apoptosis.</p><!><p>Titanium (IV) nitrate tetrahydrate (Ti(NO3)4.H2O, ≥99.9%), pure grade of polyethylene glycol (PEG), analytical-grade acetone (anhydrous, ≥99.0%), ethanol (96%), 3% of glutaraldehyde in phosphate, 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), lactate dehydrogenase (LDH) rhodamine 123 dye, acridine orange/ethidium bromide (AO/EtBr) dye, dichloro dihydro fluorescein diacetate (DCFH-DA), and anhydrous sodium sulfate were acquired from Sigma-Aldrich (Hamburg, Germany). Ciprofloxacin (5 μg/disc, MASTDISCS™, Mast Diagnostics Ltd.), vancomycin (Glaxo SmithKline Pharmaceuticals Ltd.), and gentamycin (Alpha Laboratories Ltd.) were purchased from local drug stores.</p><!><p>Plectranthus montanus Benth. (syn. Plectranthus cylindraceus Hoechst. Ex. Benth.) was collected from the Abha region of Saudi Arabia in March 2018. Dr. Mohamed Yousef, a taxonomist at the Pharmacognosy Department of the King Saud University, identified the plant material, and a voucher specimen (P-5-2018) was deposited in the same department. Aerial parts of the collected plant were sun-dried to remove the moisture content. The dried plant material was coarse powder in a mixer grinder for oil extraction. The botanical sample was kept at 4°C in the refrigerator until further use.</p><!><p>Four bacterial strains: S. aureus (ATCC 25923), E. coli (ATCC 25922), Pseudomonas aeruginosa (ATCC 25566), and Salmonella typhi (ATCC 27736) and human lung carcinoma cells were used to test the antibacterial and anticancer effect of P. cylindraceus oil and P. cylindraceus oil/TiO2/PEG bionanocomposite. Bacterial strains and cancer cell lines were supplied by the Microbiology Department of the King Khalid Hospital, Riyadh, Saudi Arabia, and Research Center of King Faisal Hospital, Riyadh, Saudi Arabia, respectively.</p><!><p>Powdered aerial parts (300 g) of P. cylindraceus were subjected to hydrodistillation in a Clevenger apparatus for 4 h to yield colorless oil by obeying the previous method of Khan et al. with slight modification [38]. The obtained volatile oil was dried over anhydrous sodium sulfate as a dehydrating agent and stored in the refrigerator at 4°C for subsequent use. The yield of the volatile oil obtained from the P. cylindraceus was 1.2% on the basis of fresh weight.</p><!><p>The qualitative and quantitative of volatile oil was carried out on Hewlett-Packard-5890 series II plus gas chromatograph coupled with HP-5989 mass spectrometer. The separation was performed in HP5-MS capillary column (25 m × 0.25 mm) coated with 0.50 µm 5% phenyl in 95% methylpolysiloxane, programmed at 70–250°C temperature with a flow rate of 3°C/min. Helium was used as a carrier gas at 1.9 mL/min at a steady flow rate. The 250°C and 280°C temperatures were adjusted for the injector and interface, respectively. Electron ionization mass spectra of the components were obtained at temperature 250°C and ionization voltage 70 eV. Finally, the obtained unknown compounds were identified by comparing spectra available from the Wiley 8 and NISTO 5 database mass spectral library. AMDIS32 software was used to calculate the retention indices (IR) values and Linear retention indices obtained for the compounds were compared with those published in the literature [39]. The constituents of P. cylindraceus oil were identified on the basis of retention time, comparison with Wiley, 2008 database library, and fragmentation pattern of components with those already reported in the literature [40].</p><!><p>The collected aerial parts of P. cylindraceus (25 g) were washed thoroughly with water and boiled in 200 mL of distilled water at 80°C for 20 min. The obtained extract was filtered using nylon mesh (spectrum) followed by a millipore hydrophilic filter (0.22 µm). The resulted aqueous extract was used to prepare TiO2 nanoparticles. Briefly, 20 mL of prepared P. cylindraceus extract was added to 80 mL of titanium (IV) nitrate tetrahydrate (1.0 × 10−3 mol L−1) with continuous stirring for 24 h at room temperature until the formation light green precipitate of TiO2 NPs. The resulted nanoparticles were centrifuged, filtered, and dried in an air oven (Scheme 1(a)).</p><!><p>The fabrication of an ultrafine plain film of PEG polymer and P. cylindraceus oil was performed by mixing 10% of PEG dissolved in 5 mL acetone with 5–15% of P. cylindraceus oil under magnetic stirring for 12 h at ambient temperature until the formation of the homogeneous composite. The obtained plain P. cylindraceus oil/PEG composite was decorated with TiO2 NPs by mixing 5 w% of P. cylindraceus oil and 5 w% TiO2 NPs suspended in the polymeric mixture and subjected to vigorous shaking for 6 h at room temperature. The polymeric P. cylindraceus oil/TiO2/PEG bionanocomposite film was obtained and kept aside for further investigation (Scheme 1(b)).</p><!><p>Various spectroscopic analyses, including ultraviolet-visible (UV-vis.), Fourier-transform infrared (FTIR), and X-ray diffraction (XRD), were carried out to characterize the TiO2 nanoparticles and their polymeric bionanocomposites. The nanoparticles and polymeric bionanocomposites were further confirmed by microscopic analysis using a scanning electron microscope (SEM) equipped with energy dispersive X-ray (EDX) and transmission electron microscopy (TEM).</p><!><p>The thermal stability of the prepared bionanocomposite film was investigated by using thermogravimetric analysis (TGA, Seiko Exstar 6300, Tokyo, Japan). Around 5 mg of sample was heated from a temperature range of 10 to 600°C at a 10°C/min heating rate under a constant flow of argon of a Shimadzu thermogravimetric analyzer (TGA-502 model).</p><p>The hydrolytic degradation analysis of bionanocomposite film was investigated by controlling the pH and temperature of the system according to ASTM F1635-11 [41]. The analysis was performed for 8 weeks. Saline phosphate buffer (pH 7.4) was used to test the hydrolytic degradation of bionanocomposite in triplicates. The binaocomposite film was kept in 10 mL PBS solution and heated in a water bath at 37°C. The film was removed at different time intervals (2nd, 4th, 6th, and 8th weeks) from the controlled medium and oven-dried at 60°C. Subsequently, the weight loss of the bionanocomposite film was monitored through gravimetric analysis together with the degradation test. Time zero was adjusted as before initiating the experiment, and the measurements were recorded at the time of withdrawals of the film. No results of weight loss were obtained at 8th withdrawal due to the high deterioration of the bionanocomposite film. The results of weight loss were obtained through the following equation:(1)M=M0−Mf,where M, M0, and Mf are the weight loss, mass of the film before degradation, and mass of the film at different removal times, respectively.</p><!><p>The antibacterial activity of the oil and as-prepared bionanocomposite was evaluated by the disc diffusion method [42]. To culture the microorganisms, Mueller–Hinton agar plates were incubated at 37°C for 18 h. A bacterial cell density of 10 × 108 UFC/mL was adjusted with a sterile saline solution. To obtain the uniform bacterial growth, a sterile swab was dipped in each microbial suspension and used to spread microorganisms over the test and control plates. Different concentration (25–100 µgmL−1) of 10 µL of each P. cylindraceus oil and P. cylindraceus oil/TiO2/PEG bionanocomposite was loaded on the surface of the sterile disc. Afterwards, the plates were incubated for 24 h at 37°C, and the zones of inhibition were determined in millimeters. Ciprofloxacin (5 μg/disc) containing disc was used as a reference control for bacterial inhibition. All the experiments were performed in triplicates, and the mean inhibition diameter was calculated. The activity was evaluated as either sensitive or resistant with a cutoff value equal to 10 mm.</p><!><p>Minimum inhibitory concentrations (MIC) of P. cylindraceus oil/TiO2/PEG bionanocomposite against methicillin-resistant S. aureus and E. coli were determined by the microdilution method using Mueller–Hinton broth. Vancomycin and gentamycin were used as a positive reference control for S. aureus and E. coli strains, respectively. Minimum bactericidal concentration (MBC) of polymeric bionanocomposite was expressed as spreading aliquots of tubes with no visible growth and the first turbid tube in the MIC series. A sterile rod was used to spread uniformly the aliquots of treated samples on nutrient agar plates and incubated at 37°C for 12 h.</p><!><p>The morphological changes in treated and untreated S. aureus and E. coli by the prepared P. cylindraceus oil/TiO2/PEG bionanocomposite were studied under scanning electron microscopy. The treated microorganisms were cut into 5–10 mM pieces and fixed on a glass slide for 1 h in 3% of glutaraldehyde in phosphate buffer saline solution. The treated tissues were dehydrated with ethanol and dried with carbon dioxide. Silver pain vacuum coated with gold-palladium alloy was used to mount the dry tissues on the aluminum stubs and examined under SEM with 15 kV acceleration voltage.</p><!><p>The cytotoxic effect of P. cylindraceus oil and polymeric bionanocomposite against human lung carcinoma (A549) cell lines was performed according to the previously reported method [43]. The A549 cancer cells were kept in a DMEM medium comprised 10% FBS, 100 mg mL−1 streptomycin, 2 mM glutamine, and 100 IU mL−1. The cell cultures were maintained in a humidified incubator at 37°C with a continuous flow of 5% CO2. A 1 × 104 cells/well concentration of A549 cell culture was separately seeded in 96 well plates. The cells were treated with different concentrations of P. cylindraceus oil and polymeric bionanocomposite (0 to 200 µg mL−1) and incubated for one complete day. After 24 h of incubation, 100 µL of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) was added to treated cells and further incubated for 4 h at 37°C. After 4 h incubation, a purple-colored formazan dissolved in 100 µL of dimethyl sulfoxide was then added. After 30 min, optical density was recorded on ELISA multiwell plate reader at 570 nm for calculating IC50 values. A lactate dehydrogenase (LDH) leakage assay was performed to assess the damage of the cell membrane and loss of membrane integrity. Rhodamine 123 staining was employed to analyze the changes that occurred in mitochondrial membrane potential (ΔΨm) after the treatment of bionanocomposite. All the experiments were performed in triplicates to avoid errors.</p><!><p>The ability of P. cylindraceus oil and P. cylindraceus oil/TiO2/PEG bionanocomposite to induce apoptosis (A549 cells) was evaluated using double staining acridine orange/ethidium bromide (AO/EtBr). The staining procedure was performed according to the previously reported method [44]. Two six-well plates were individually seeded with A549 cancer with respect to IC50 concentrations of P. cylindraceus oil and polymeric bionanocomposite and were incubated for one day. Acridine orange (100 μg mL−1) and ethidium bromide (100 μg mL−1) fluorescent dyes were mixed with respective cells and incubated in the dark for 30 min at 37°C. Cells were immediately viewed under a fluorescence microscope with 500 nm excitation wavelength, and the emission was recorded at 570 nm. The AO/EtBr double staining test mechanism revealed that cells having normal nuclear chromatin represent green nuclear staining confirmed that AO is picked up only by viable cells. However, apoptotic cells containing condensed chromatin portray orange to red nuclei revealed that EtBr was picked up by nonviable cells.</p><!><p>The DCFH-DA staining was applied to estimate the intracellular ROS generation in A549 cancer cells by P. cylindraceus oil and P. cylindraceus oil/TiO2/PEG bionanocomposite. Two six-well plates were individually seeded with A549 cancer cells and treated separately with P. cylindraceus oil and polymeric bionanocomposite. The treated cells were incubated for 24 h, rinsed with 1x cold phosphate buffer saline (PBS, pH 7.4), stained with 10 μg/10 μL DCFH-DA and kept in dark for 39 min. The level of ROS generation was evaluated under the fluorescence microscope. The above assay was performed according to the previously reported procedure and in triplicates [45].</p><!><p>The photocatalytic effect of polymeric bionanocomposite was estimated on the basis of its ability to degrade the methylene blue (MB) dye under visible and UV radiations. The ultraviolet source used for the experiment was a Mercury vapor lamp with 120 W. A methylene blue stock solution (20 µg mL−1) was prepared. Ten milliliter of polymeric bionanocomposite was mixed with 100 mL of dye solution under continuous magnetic stirring for 1 h in the dark to enhance the equilibrium balance between MB and photocatalyst prior to exposure to sunlight and UV irradiations. The reaction mixture was irradiated by the light source, and after each 30 min time interval, around 2 mL of the suspension was taken up and centrifuged to remove the suspended nanoparticle. UV-visible spectrophotometer was employed to determine the rate of dye degradation at 660 nm, and the degradation percentage was calculated by applying the following formula:(2)% of degradation = Ci−CfCi×100,where Ci and Cf were the initial and final concentrations of dye at a time interval "t," respectively.</p><!><p>Hydrodistillation of fresh aerial parts of P. cylindraceus resulted in the extraction of a pleasant smelling colorless essential oil with an excellent yield of 0.42% (v/w). Qualitative and quantitative GC-MS analysis of P. cylindraceus oil identified 30 constituents that accounted for 96.54% of total oil composition. The identified compounds with their linear indices (LRI) and relative content are summarized in Table 1. The essential oil is composed of about 90% of monoterpenes of which 70% were oxygenated monoterpenes while 20% were nonoxygenated monoterpene. Seven of the detected constituents present in the oil were oxygenated sesquiterpenes. Camphor (38.85%) and 1,8-cineol (21.84%) were present as the major compounds of P. cylindraceus oil. Other classes of constituents such as oxygenated aliphatic alkanes (1.4%), diterpenes (1.2%), and aromatic and aliphatic hydrocarbons (0.4) were found in minute quantities (Figure 1).</p><!><p>Different spectroscopic and microscopic investigations were employed to characterize the formation of TiO2 nanoparticles and their bionanocomposites. The UV-Vis screening of TiO2 nanoparticles (from 250 to 500 nm) prepared by P. cylindraceus aqueous extract showed an absorption peak at a wavelength of 350 nm (Figure 2(a)). The obtained result was in agreement with the previously reported synthesis of TiO2 nanoparticles [46]. A particle size analyzer was used to calculate the average particle size of the formed TiO2 nanoparticles, and the mean average value was found to be around 60 nm (Figure 2(b)). The size and shape of TiO2 nanoparticles were viewed under SEM, and images were picked at different magnifications (30,000x and 50,000x) showed irregular shape nanoparticles with particle size ranging from 68 to 88 nm (Figures 3(a) and 3(b)).</p><p>Gradual addition of P. cylindraceus oil (5–15%) was applied to select the suitable oil concentration to fabricate the suggested P. cylindraceus oil/TiO2/PEG bionanocomposite. Around 5% of P. cylindraceus oil was initially added, but this quantity of oil could not give a well-characterized bionanocomposite. Therefore, gradual increments of P. cylindraceus oil were continued up to 15 w%. It was observed that 15w% of P. cylindraceus oil provides well-characterized polymeric P. cylindraceus oil/PEG solution. The surface morphology of PEG, the prepared plain polymeric mixture, and the synthesized P. cylindraceus oil/TiO2/PEG bionanocomposite was examined under SEM using 30,000x magnifications. The surface morphology of 100 nm of PEG at 30,000x showed a smooth and uniform polymeric matrix distribution (Figure 4(a)). The SEM images of plain P. cylindraceus oil at 30,000x magnifications showed the irregular coagulant mass with rounded oil droplets and tiny pores (Figures 4(b) and 4(c)). Furthermore, the SEM images of the synthesized bionanocomposite at the same magnification indicated the irregular structure of TiO2 nanoparticles with particle sizes ranging between 68 and 88 nm (Figure 4(d)).</p><p>The XRD pattern of TiO2 nanoparticles synthesized from aqueous P. cylindraceus extract was studied. The obtained results revealed that the structure was irregular. These results were in good agreement with JCPDS card number 21–1272. The XRD peaks were observed at 26°, 37°, 46°, 51°, 56°, 63°, and 74° corresponding to miller index values (1 0 1), (0 0 4), (2 0 0), (1 0 5), (2 1 1), (2 0 4), and (2 1 5), respectively (Figure 5(a)). It was noticed that as the width of the peak increases, there is a decrease in particle size, indicating that the formed particles were in nanoscale. To confirm the dispersion of P. cylindraceus oil and TiO2 nanoparticles in the polymeric matrix PEG, XRD analysis with Cu kα radiation (λ = 1 540 Ǻ) over Bragg angles in a range from 10 to 80 degrees, current-voltage 50 mA, and 35 kV was applied. The outcomes of XRD analysis demonstrated that no significant peak was recorded for the polymeric solution of PEG (Figure 5(b)-A); this was due to the amorphous nature of the plain polymer [47]. However, the XRD pattern of P. cylindraceus oil-PEG showed a broad peak near ∼20 degrees (Figure 5(b)-B), this could be due to the dispersion of P. cylindraceus oil in the PEG matrix causing a remarkable change in the peak intensity. However, the XRD spectrum of the synthesized P. cylindraceus oil/TiO2/PEG bionanocomposite showed the appearance of distinct peaks at 39.26°, 42.15°, 56.83°, and 73.52° revealed that Bragg's reflection was from TiO2 (101), TiO2 (004), TiO2 (200), and TiO2 (215) (Figure 5(b)-C). The observed peaks clearly confirmed the decoration of TiO2 nanoparticles over the surface of the plain polymeric-oil mixture, resulting in the formation of bionanocomposite film.</p><p>EDX analysis of PEG and the bionanocomposite was investigated using an EDX equipped with SEM. Figures 6(a) and 6(b) show the surface morphology of the plain polymer and the synthesized bionanocomposite. The presence of TiO2 nanoparticles indicated that the metal oxide was dispersed well in the polymeric matrix. Two signals representing C and O for PEG were displayed. However, the synthesized bionanocomposite showed the presence of distinct signals for C, O, and Ti (Figures 6(c) and 6(d)). No other elemental impurity was observed. The EDX results confirmed that the bionanocomposite was successfully fabricated and well distribution of P. cylindraceus oil and TiO2 nanoparticles in the polymeric matrix PEG.</p><!><p>The thermal stability of bionanocomposite has a great influence on inherent properties and strong intermolecular interaction between the bionanocomposite molecules. An increase in thermal energy as compared to bond dissociation energy greatly affects the dissociation of the polymeric chain of bionanocomposite [48]. The interaction between the surfaces of organic and inorganic nanomaterials was confirmed by thermogravimetric analysis. The stability of bionanocomposite was evaluated at different weight percent (5 w%, 10 w%, and 15 w%) of PEG/TiO2 NPs (Figure 7). The results showed that weight reduction started at 130°C, and measured values were 150, 175, and 230°C for the aforementioned bionanocomposite film. It was noticed that around 10% of water was lost from the surface of the bionanocomposite film between 175 and 230°C, suggesting that 230°C is the maximum thermal stability limit. Therefore, the outcomes revealed that an increase in the PEG weight percent resulted in an increase in carbonaceous material residual weight generated by polymer decomposition. However, the hydrolytic degradation showed different trends in the behavior of weight loss in the bionanocomposite. The results revealed that there was no loss of bionanocomposite weight at the beginning of the experiment. However, 5.70% and 13.54% of its weight was lost in the 2nd and 6th week, respectively (Table 2). It may be hypothesized that the degradation was affected by the addition of TiO2 NPs and PEG, which triggered the water penetration into the system that resulted in the diminution of polymeric chain and enhanced the disintegration of the system.</p><!><p>The antibacterial efficacy of the pure P. cylindraceus oil and P. cylindraceus oil/TiO2/PEG bionanocomposite were determined against microbes based on the zone of inhibition. P. cylindraceus oil and polymeric bionanocomposite were subjected to antibacterial tests against four bacterial species, namely S. aureus, E. coli, P. aeruginosa, and S. typhi. Preliminary screening results revealed that the as-synthesized polymeric bionanocomposite showed much greater antibacterial potential compared to pure P. cylindraceus oil at similar proportions. At the highest sample concentration (50 µgmL−1), polymeric bionanocomposite displayed a nearly twofold greater bactericidal effect compared to P. cylindraceus oil as measured using the disc diffusion test. After 48 h incubation, the zones of inhibition of bionanocomposite and P. cylindraceus oil against S. aureus and E. coli were (20.03 + 1.20 and 25.50 ± 1.08 mm) and (14.52 + 0.68 and 16.12 ± 0.56 mm), respectively. As shown in Table 3, P. cylindraceus oil and polymeric bionanocomposite exhibited the strongest antibacterial effect towards E. coli and S. aureus (Figure 8(a)). P. cylindraceus oil has been reported to provide the benefits of antimicrobial activity and medicinal value [49]. P. cylindraceus oil is dominated by monoterpenes; especially camphor and 1,8-cineol have extensive antibacterial potential because Gram-positive and Gram-negative bacteria were found susceptible to these oils [50, 51]. The strong bactericidal effect of polymeric bionanocomposite may be attributed to TiO2 nanoparticles that strongly bond with electron-donating groups in bioactive compounds containing oxygen, sulfur, and nitrogen. These nanoparticles disturb the cell boundary of microorganisms and as a result, the outermost rigid cell layer lost its protection. The small size of TiO2 nanoparticles facilitates the easy penetration to cell membranes, and monoterpenes present in oil disturb the lipid fraction of the plasma membrane of bacteria, resulting in alterations of membrane permeability and leakage of intercellular content. Also, the TiO2 nanoparticles and their ions have the ability to produce free radicals (reactive oxygen species, hydroxyl ion, superoxide ion, singlet oxygen, and hydrogen peroxide) that induces oxidative stress [52]. The ROS can permanently destroy the microbes and result in the death of microbes. There are two probable mechanisms involved in antibacterial action of as-synthesized polymeric bionanocomposite, that is, photogeneration of free radicals or by the intracellular interaction between microbial cell membrane and nanoparticles (negative charge of the outermost layer of cell membrane with positive charge of TiO2 ions) resulting in microbial growth inhibition and inducing the demise of microbes. A steady S-metal group is formed when the ions released from TiO2 nanoparticles bond with thiol or sulfhydryl groups (–SH) and proteins of the cell membrane with the loss of hydrogen ion, reducing the cell permeability and resulted in the death of the cell [53]. However, ROS generation contributes to the disruption of DNA, enzymes, proteins, and lipids. The enzyme inhibition caused by polymeric bionanocomposite was found to be the most efficient mechanism as it destroys the assimilatory food pathway and induces cell death [54]. The mechanism of polymeric bionanocomposite can proceed via two kinetic processes utilizing their ability to penetrate the bacterial cell membrane and killing the microorganism or by phagocytosis of TiO2 nanoparticles by the mononuclear phagocyte system (Scheme 2). Antibacterial potentials of P. cylindraceus oil and polymeric bionanocomposite against S. aureus and E. coli are demonstrated in Scheme 2. Also, antibacterial results showed that polymeric bionanocomposite as well as pure oil were more susceptible against E. coli (Gram-negative) bacteria as compared to S. aureus (Gram-positive) due to differences in cell wall composition. The cell wall of E. coli contains a thin peptidoglycan layer in contrast to S. aureus [55]. Thus, the result obtained confirmed that polymeric bionanocomposite exerted a remarkable bactericidal effect than pristine P. cylindraceus oil. The zone of inhibition against P. cylindraceus oil and polymeric bionanocomposite against S. aureus and E. coli is shown in Figure 8(a).</p><!><p>Agar well diffusion procedure was applied to determine the bacteriostatic and bactericidal concentrations of P. cylindraceus oil/TiO2/PEG bionanocomposite. The lowest concentration required for inhibition of visible growth of S. aureus and E. coli was assessed after 24 h at 37°C. The gradual increase in polymeric bionanocomposite concentration (5–1,280 μg mL−1) has shown a significant reduction in bacterial cell viability (p < 0.05). The 320 μg mL−1 represents the MIC for the S. aureus and E. coli as depicted in Figure 8(b) (A and B). MBC for S. aureus and E. coli was displayed at 640 and 1,280 μg mL−1, respectively, for the polymeric bionanocomposite (Table 4). The antibacterial properties of polymeric bionanocomposite were enhanced in contrast to pristine P. cylindraceus oil, due to the small size of TiO2 nanoparticles that provide a more dominant attack to the cell membrane of the microorganisms [56].</p><!><p>The effect of polymeric bionanocomposite on the surface morphology of S. aureus and E. coli was inspected under SEM. The treatment of polymeric bionanocomposite has changed the shape and size of selected microbes due to the coating of TiO2 nanoparticles on the surface of microbial cells as illustrated in Figures 9(b) and 9(e). The polymeric bionanocomposite can easily penetrate the peptidoglycan layer of the cell membrane of S. aureus and E. coli, leading to the destruction of the cell membrane, leakage of cell ingredients, and consequently resulting in the death of the microbial cell (Figures 9(c) and 9(f)) [52]. The untreated microbial cells were selected as a positive control for the comparison (Figures 9(a) and 9(d)).</p><!><p>Lung cancer is among the most widespread form of carcinoma experience in humans and the leading cause of cancer-related death worldwide. Epidemiological studies have shown that among all types of cancer, 17.6% of total cancer deaths occur due to lung cancer [57]. Nonsmall lung cancer (NSCLC) is the main sponsor of total lung cancer, which is classified further into large cell carcinoma (2.9%), squamous cell carcinoma (20%), and adenocarcinoma (35%) [58]. Adenocarcinomic human alveolar basal epithelial cells (A549) are frequently used as in vitro model systems for the investigation of NSCLC. The major responsible determinants of lung cancer are environmental and lifestyle factors including air pollution, smoking, alcohol, lack of physical activity, occupational exposure, and diet [59]. The sudden rise in lung cancer mortality rate is due to the lack of efficient diagnostics and therapeutic strategies. The treatment used for lung cancer includes surgery, chemotherapy, and radiotherapy alone or in combination depending on the stage of cancer. Gemcitabine in combination with cisplatin is FDA-approved chemotherapeutic drug that has been utilized as a primary drug for lung cancer therapy. Recently, combinatorial drug therapy has gained much attention due to the improvement of the therapeutic efficacy of the drug. Cisplatin in combination with paclitaxel, gemcitabine, and docetaxel drugs have expressed enhanced therapeutic efficiency [60, 61]. However, the majority of these chemotherapeutic drugs are very expensive and associated with adverse side effects such as neuronal damage, skin irritation, and acute pain. Hence, there is a serious demand for the development of nontoxic, cost-effective, ecofriendly, and targeted drugs for cancer treatment. Nanoparticles possessing unique physiochemical features offer an extraordinary interaction with proteins, nucleic acids, and lipids present on the surface of cells and within the cell body, which might establish new routes for the diagnosis and treatment of cancer [62]. Recently, several bio-based polymeric nanocomposites have shown outstanding anticancer properties against various types of cancer [63]. In the present study, the anticancer effect of P. cylindraceus oil/TiO2/PEG bionanocomposite prepared through a biogenic route using P. cylindraceus oil was evaluated against lung cancer cells (A549).</p><!><p>MTT assay and LDH leakage assay were applied to assess the in vitro cytotoxic activity of P. cylindraceus oil/TiO2/PEG bionanocomposite. Figure 10(a) clearly depicts that polymeric bionanocomposite exhibited potent cytotoxic activity against A549 cancer cell lines in a dose-dependent manner with 42.7 ± 0.25 μg mL−1 IC50 value, in contrast to cisplatin positive control (IC50 value of 27.6 ± 0.018 μg mL−1). Morphological changes occurred in the cell and membrane damage was analyzed under phase-contrast microscope. Results revealed a high-density cell population with normal epithelial morphology in the control group, whereas rounded and shrunk cells, condensed chromatin, formation of apoptotic body, and protrusions of membrane were noticed in the cells treated with polymeric bionanocomposite and positive control (Figure 10(b)). The activation caspase cascade caused by polymeric bionanocomposite and positive control might be responsible for morphological changes, wherein the component poly (ADP-ribose) polymerase (PARP) needed for the repair of DNA would be cleaved [64]. Cellular uptake of nanoparticles by endocytosis or macropinocytosis would initiate the ROS generation, trigger the apoptotic pathway, and eventually lead to cell demise [65].</p><p>The cellular or tissue damage was assessed on the basis of the level of lactate dehydrogenase (LDH) in the extracellular medium. LDH is a cytoplasmic soluble enzyme, released during the disruption of the cell membrane into the extracellular matrix. Thus, the increased level of LDH indicates cellular toxicity [66]. The cytotoxic potential of polymeric bionanocomposite was further validated by the LDH leakage assay. The results showed that the increase in the percentage of LDH leakage was found concentration-dependent in polymeric bionanocomposite-treated cells (Figure 10(c)). The above results were in accordance with the previous reports that TiO2 nanoparticles treatment would permeabilize the cell membrane, allowing the LDH leakage in lung cancer cells, causing cell demise [67].</p><!><p>Chemotherapeutic drugs and radiation therapy induce the death of cancer cells by the activation of the ROS-mediated apoptosis pathway. An increase in the levels of ROS leads to various pathological incidences such as inflammation, lipid peroxidation, protein oxidation, and DNA damage. ROS generation is considered to participate in several cellular events including inflammation, DNA damage, senescence mutation, and apoptosis [68]. In the current study, the levels of intracellular ROS were measured in A549 cancer cells to determine the mechanism behind the anticancer properties of polymeric bionanocomposite. Spectrofluorometric results showed that the fluorescence intensity in polymeric bionanocomposite treated groups was increased twofold in comparison to the control group (Figure 11(c)). The fluorescent microscopic analysis was applied to validate the results, which exhibited an enhanced fluorescent green intensity in the polymeric bionanocomposite treated cells, in contrast to the positive control and control, indicating a rise in ROS level (Figure 11(d)). These results confirmed that apoptosis induced by polymeric bionanocomposite was associated with the accumulation of ROS in A549 cancer cells.</p><!><p>One of the key players required for activation of apoptotic pathway is the loss of mitochondrial membrane potential (MMP Ψm) [69]. The disturbance of MMP in A549 cancer cells in polymeric bionanocomposite treated cells was examined by using Rhodamine 123 dye. In contrast to control cells, a potent reduction (p < 0.05) in fluorescence intensity nearly threefold was noticed in polymeric bionanocomposite treated cells, representing MMP disruption (Figure 11(b)). A reduction in fluorescent intensity was observed in the polymeric bionanocomposite treated cell under a fluorescence microscope, validating the spectrofluorometric results and further confirming the fact that polymeric bionanocomposite would cause a potential loss in MMP (Figure 11(a)).</p><!><p>The predominant mechanism of action of chemotherapeutics used in cancer treatment proceeds by triggering the apoptotic pathway to destroy the cancer cells [70, 71]. AO/EtBr dual staining procedure was applied to determine the rate of apoptosis persuade by polymeric bionanocomposite. Figure 12(a) displays the presence of uniform green-colored cells in the control vehicle, representing the presence of live cells (early apoptotic cells) in the range of 500 to 530 nm under a fluorescent microscope. However, polymeric bionanocomposite treated cells displayed the presence of red- and orange-colored cells, representing dead cells (late apoptotic cells) in the range of 510 to 595 nm under a fluorescent microscope, while in cisplatin-treated cells, dual stained green and orange-colored cells were found, indicating the necrosis. The results of the quantitative analysis showed that a significant increase in apoptotic cells was observed in polymeric bionanocomposite-treated cells (80.3 ± 0.01% dead and 19.7 ± 0.02% live cells) while compared to nontreated (88.9 ± 1.20% live and 11.02 ± 0.86% dead cells). Cisplatin-treated cells (positive control) showed the presence of 36.2 ± 0.08% live cells and 63.7 ± 0.05% apoptotic cells (Figure 12(b)). The results demonstrated that the polymeric bionanocomposite led to the apoptotic cell demise of A549 cancer cells.</p><p>To get further understanding of the mode of cell demise stimulated by polymeric bionanocomposite, 4′,6-diamidino-2-phenylindole (DAPI) staining was performed. As shown in Figure 12(c), the polymeric bionanocomposite treatment caused condensation of chromatin and morphological alteration in the nucleus of A549 cancer cells, illustrating the fact that the apoptotic potential of polymeric bionanocomposite is completely dependent on the production of ROS levels, which in turn has resulted in oxidative stress-mediated cell death in A549 cancer cells. The results concluded that TiO2 nanoparticle present on the surface of the bionanocomposite film of polymeric bionanocomposite enhanced the production of intracellular ROS level, creating oxidative stress, disrupting MMP, consequently activating apoptotic intrinsic pathway and ultimately causing the death of the cell. Above all, ROS, MTT, LDH, AO/EtBr, and DAPI results confirmed an enhanced anticancer potential against A549 cell lines using P. cylindraceus oil/TiO2/PEG bionanocomposite compared to that of the previous similar report [72].</p><!><p>In vitro toxicity studies were conducted as an alternate method to validate the toxicological profile of the newly identified drug [73]. For the determination of immunotoxicity and dosage limit of the new drug constituent under investigation, human peripheral mononuclear cells (PBMC) were applied in vitro model [74]. MTT assay was used to determine the in vitro cytotoxic potential of polymeric bionanocomposite in PBMC. Figure 12(d) demonstrates that the prepared bionanocomposite treatment induced no exceptional changes in the cell viability and membrane integrity of PBMC after one long day of incubation, and cell viability percentage was similar to untreated cells. However, H2O2-treated cells (positive control) expressed 69.9 ± 0.18% cytotoxicity, while bionanocomposite treated do not have cytotoxic effects similar to untreated cells; thus indicating they are safe for use as drugs. Further confirmation of safety aspects requires in vivo studies to ensure absolute safe usage.</p><!><p>For the drug application and engineering materials, TiO2 nanoparticles and their bionanocomposites must be nontoxic and biocompatible. The cytotoxicity of TiO2 nanoparticles and their bionanocomposites was evaluated in vitro A549 cancer cells. The viability of A549 cells exposed to TiO2 nanoparticles and its bionanocomposite with different oil concentrations is presented in Figure 13(a). The results showed that A549 cells can adhere and proliferate on both TiO2 nanoparticles and its bionanocomposite. The absorbance of all the samples increases with the increase in culture time, suggesting the well growth of the cells. No visible decrease in viability between TiO2 nanoparticles and their bionanocomposites can be noticed at 1, 5, and 10 culture days. Considering the nonsignificant variances (p < 0.05) between the investigated samples, bionanocomposite displayed comparable biocompatibility with TiO2 nanoparticles. The obtained results were in agreement with previous reports, oil-based polymeric bionanocomposite showed noncytotoxicity to many cells [75], suggesting that bionanocomposite have better cytocompatibility. However, the OD values of TiO2 nanoparticles and bionanocomposite were lesser than control (PBS) particularly after 10 days of culture at 570 nm. This can be attributed to the strong hydrophobic character of control compared to TiO2 nanoparticles and bionanocomposite; cell adhesion proteins tend to stick to the hydrophobic surface. Figure 13(b) demonstrates the SEM images of A549 cells after culture at 1, 5, and 10 days on TiO2 nanoparticles and bionanocomposite. Fibroblast cells started to express round or fusiform shape on the surface of TiO2 nanoparticles and bionanocomposite after immediately 24 h seeding, suggesting that the cells can be easily attached to these nanomaterials. After five-day culture, it was observed that fibroblasts attached to the surface of nanomaterials changed their original shape to fusiform. The proliferation of cells continued on the surfaces of the samples and became subconfluent on the 10th day of the culture. Morphological techniques were applied to further evaluate the phenotype and the interaction of the cells on each sample surface. As shown in SEM images of the 5 days cultured in control, TiO2 nanoparticles, and bionanocomposite, the original shape of fibroblasts adhered to all the surfaces changed to fusiform shape. A well-expanded spindle morphology with intercellular tight junctions with nearby cells was observed in cells grown on TiO2 nanoparticles and bionanocomposite. However, a slight difference in the cell morphology was noticed in the cells grown on TiO2 nanoparticles and bionanocomposite surfaces. More flattened cells on the surface of bionanocomposite were observed in contrast to TiO2 nanoparticles, which can be attributed to a much rough surface of the bionanocomposite (Figure 13(b)). Generally, the nanotopography characteristics and surface chemistry are the responsible factors for the cell behavior of the bioactive material. Furthermore, the roughness of bionanocomposite surface and the presence of TiO2 nanoparticles on its surface promote the attachment of the cells. Thus, the outcome of the MTT assay and morphological changes confirmed that bionanocomposite is a biocompatible material that can be applied as a future cell culture scaffold.</p><!><p>The photocatalytic effect of bionanocomposite was measured on the basis of the rate of degradation of methylene blue (MB; organic dye pollutant). The MB degradation rate was determined as the percentage of decoloration in contrast to time based at the 660 nm wavelength. The obtained results showed that bionanocomposite expressed a potent photocatalytic effect under UV as well as visible light in contrast to control (Figure 14). The as-synthesized bionanocomposite displayed approximately 7.9 times higher effects than the blank solution used under sunlight.</p><!><p>A simple, ultrasensitive, and ecofriendly method was developed to fabricate P. cylindraceus oil/TiO2/PEG bionanocomposite. The fabricated bionanocomposite film was examined using various microscopic and spectroscopic techniques to confirm the well-dispersed P. cylindraceus oil and TiO2 nanoparticles in the polymeric matrix PEG. The surface morphology of the film showed a well distribution of TiO2 nanoparticles with particle size in the range of 60–80 nm. The antibacterial potential of the fabricated film has been screened against four bacterial strains: S. aureus, E. coli, P. aeruginosa, and S. typhi. It was observed that the bionanocomposite film exhibited excellent antibacterial activity against E. coli and S. aureus. The anticancer effect of the fabricated bionanocomposite-film-enriched TiO2 nanoparticles also has been evaluated against human lung carcinoma.</p>

PubMed Open Access

Product Control and Insight into Conversion of C6 Aldose Toward C2, C4 and C6 Alditols in One‐Pot Retro‐Aldol Condensation and Hydrogenation Processes

Alcohols have a wide range of applicability, and their functions vary with the carbon numbers. C6 and C4 alditols are alternative of sweetener, as well as significant pharmaceutical and chemical intermediates, which are mainly obtained through the fermentation of microorganism currently. Similarly, as a bulk chemical, C2 alditol plays a decisive role in chemical synthesis. However, among them, few works have been focused on the chemical production of C4 alditol yet due to its difficult accumulation. In this paper, under a static and semi-flowing procedure, we have achieved the product control during the conversion of C6 aldose toward C6 alditol, C4 alditol and C2 alditol, respectively. About C4 alditol yield of 20 % and C4 plus C6 alditols yield of 60 % are acquired in the one-pot conversion via a cascade retro-aldol condensation and hydrogenation process. Furthermore, in the semi-flowing condition, the yield of ethylene glycol is up to 73 % thanks to its low instantaneous concentration.

product_control_and_insight_into_conversion_of_c6_aldose_toward_c2,_c4_and_c6_alditols_in_one‐pot_re

Introduction<!>Product Control of C6, C4 and C2 Polyols in One-Pot Reaction System<!>Product Control of C6, C4 and C2 alditols in semi-continuous reaction system<!>Conclusion<!>Experimental Section Materials<!>Catalytic Experiments<!>Product Analysis

<p>Biomass-derived alditols, including sorbitol, mannitol, xylitol, erythritol and others, could be acquired when aldehyde in sugars is reduced to hydroxy group. [1] Similar to most of sugars, they could offer a wide range of sweetness. However, excessive intake of sugar, as all know, will cause human pancreatic islet dysfunction, diabetes, obesity and other diseases. [2] Fortunately, not only can sugar alcohols be substitutes for traditional sugar to satisfy people's desire for sweet without causing obvious changes in blood sugar and insulin, but also they have less calorigenic properties. [3] In addition, alditols with low carbon numbers, e. g. 1,2-propylene glycol (1,2-PG) and ethylene glycol (EG), as essential platform molecules, are widely used in cosmetic, food and pharmaceutical industry to produce various value-added derivatives. [4] Among them, C4 alditol is not only considered as a zero-calorie sweetener but also as a potential chemical for production of C4 chemicals, such as butadiene, 1,4-butanediol, tetrahydrofuran and other butanediols which are consumed in large scale and used in many fields. [5] Dean et al. reported the successful deoxy dehydration to highly stereospecific olefin from C4-C6 sugar alcohols catalyzed by methyltrioxorhenium using another alcohol as solvent. [6] Currently, sugar alcohols mostly come from the fermentation broth of microorganisms and the hydrogenation of aldose. [7] For example, sorbitol and mannitol can be obtained by hydrogenating glucose and mannose. [8] The hydrogenation of sugar or biomass are usually catalyzed by some support catalyst based on noble metal such as Ru, Pd and Pt. [9] Perrard et al. have achieved complete conversion of glucose hydrogenation over a Ru catalyst loaded on activated carbon with a sorbitol selectivity of 99.2 %. [10] And xylitol can be hydrogenated by its corresponding sugar xylose. [11] 1,2-PG and EG can be prepared from the hydration of ethylene oxide and propylene oxide derived from petroleum cracking. [4] Erythritol, the most marketable sweet substitute, almost comes from the fermentation of glucose. [12] At present, there are two chemical processes to synthesize it, the one is first mixing acetylene and formaldehyde to obtain 2-butene-1,4-diol, and then oxygenating it into erythritol. [13] The other leverages an industrial process of starch or cellulose that contains acid and alkali treating, then oxidization by periodate and hydrogenation by nickel catalyst under high temperature and high pressure. [14] In the last decades, hydrolysis of cellulose has attracted much attention. Scientists have been able to obtain glucose from cellulose at a high yield. Fukuoka's group reported for the first time the ability of using heterogenous catalysts to depolymerize cellulose into sugar alcohols most of which was sorbitol. [15] They supplied a promising method producing polyols and essential value-added chemicals from cellulose in the presence of a large amount of hydroxy groups. Zhang et al. developed a route to acquire 61 % yield of EG from cellulose over NiÀ W/C. [16] Palkovits et al. found that cellulose could be converted into C4À C6 sugar alcohols with a total yield of 81 %, catalyzed by Ru/C combined with heteropoly acid H 4 SiW 12 O 40 under the conditions of 433 K and 5 MPa of H 2 . [17] Though the sugar alcohol yield of 81 % is of high level, the yield of erythritol is less than 6 %, and the rest of high yield is mostly contributed by sorbitol and mannitol.</p><p>Herein, we studied one-pot retro-aldol and hydrogenation process of C6 aldose toward C2À C6 alditols under the static and semi-flowing conditions, and explored the control factor influencing the product distribution of C2, C4 and C6 alditols. Under optimum conditions, the total yield of C4 and C6 sugar alcohols can exceed 60 % with the premise of C4 alditol yield of 20 % in the one-pot system, whereas the yield of EG can be even up to 73 % in the semi-flowing condition. These results provide the possibility of dynamical regulation of product distribution during the production of alditols from biomass derived sugars.</p><!><p>As Scheme 1 shows, C6 aldoses can be directly converted into C6 alditol by the hydrogenation process. They can also produce C4 alditol and EG via a cascade retro-aldol and hydrogenation process. C4 alditol is the hydrogenation product of the intermediate tetrose during retro-aldol process of C6 aldose. However, accompanied with its hydrogenation, the further retro-aldol process of tetrose toward glycolic aldehyde (GA) competitively occurs. This results in the difficulty in getting C4 alditol with high yield from C6 aldose. Obviously, for accumulating C4 alditol, the well matching between the production of tetrose by retro-aldol process of C6 aldose and its hydro-genation is required. Hence, we screened the catalyst types and their ratios as well as the reaction temperature and pressure to explore the key factors in product control of C6 aldose conversion.</p><p>Table 1 shows the ability on hydrogenation of different aldoses in the presence of different noble metal catalysts. It can be found that, among three catalysts, Ru/C exhibits the highest hydrogenation efficiency, which is in line with the previous reports. [18] Besides, it is clear that the hydrogenation of GA and C6 aldose can well perform with high conversion and high selectivity. However, the selectivity of C4 alditol is just lower than 50 % during the hydrogenation of erythrose (ERO), even though in the presence of Ru/C. Obviously, the high instability of ERO during the hydrogenation process further aggravates the difficulty in gathering of C4 sugar alcohol.</p><p>According to our previous reports, [19] tetrose could be attained with the yield of 30 % in retro-aldol condensation process of C6 aldoses firstly. Taking the above selectivity of C4 alditol in the hydrogenation of ERO into account, theoretically, only less than C4 alditol yield of 15 % can be obtained. However, our experimental results demonstrate that the actual yield of C4 alditol is less than 5 % by two-step method due to the high instability of tetrose. This means that a one-pot process of retro-aldol and hydrogenation process rather than a two-step process have to be adopted to obtain C4 alditol with high yield. If tetrose in situ produced by retro-aldol process of C6 aldose can be converted in time by hydrogenation, it is expected to improve the yield of C4 alditol. In view of this, we studied the role of product distribution of C6, C4 and C2 polyols in the one-pot retro-aldol and hydrogenation process of C6 aldose and the analyses and identifications of the various products are showed in Figure S1 and Figure S2.</p><p>It was reported that many Mo-and W-series oxides or carbides and their salts could catalyze the splitting of CÀ C bond, which was mainly due to their Lewis acidities. [20] Therefore, several kinds of W-based and Mo-based catalysts, i. e. ammonium tungstate (AT), ammonium metatungstate (AMT), ammonium paratungstate (APT) and amine phosphomolybdate (APM), were used as retro-aldol catalysts to perform the tandem conversion of glucose (GLU) toward C2À C6 alditols with Ru/C. As shown in Table 2, W-based catalysts display better catalytic performance than Mo-based catalyst in retro-aldol process. For Scheme 1. Conversion routes of C6 aldose toward various alditols. C4 sugar alcohol accumulation, the reaction system catalyzed by AT and Ru/C gets the highest carbon yield, i. e. 17.5 %. These results also indicate, as expected, that C4 alditol can be better accumulated by one-pot process. The consuming of the intermediate product tetrose and the generating of C4 alditol consecutively pull the equilibration of two tandem reactions so that one-pot synthesis is more appropriate to produce C4 sugar alcohol.</p><p>The ratio of the two used catalysts in this tandem reaction was studied to explore its influence on the product distribution. It shows that the retro-aldol process of C6 aldoses is accelerated with the increasing amount of AT, which leads to a decreasing yield of C6 sugar alcohol and an increasing C2 alditol (Table 3).</p><p>However, the yield of C4 alditol reaches the highest one and then decreases with the increasing ratio of AT:Ru/C. For example, the highest yield of C4 alditol could be obtained when the ratio of AT:Ru/C is 80 : 50 (mg:mg) (Table 3). On the one hand, the accumulation of C4 alditol requires the rapid retro-aldol process to avoid excessive direct hydrogenation of hexose. On the other hand, the rapid retro-aldol rate will lead to the formation of EG from ERO. Moreover, the low hydrogenation selectivity of ERO further limits the accumulation of C4 sugar alcohol. That's why few works were published discussing the acquisition of C4 sugar alcohol from hexose conversion. It is worth noting that, in the fermentation industry, a number of studies have focused on the deliberate coproduction of C4 alditol with another compound of interest such as C6 alditol. [21] Therefore, taking the high yields of C4 and C4 + C6 sugar alcohols into consideration, the 80 : 50 (mg:mg) ratio of AT:Ru/C is adopted.</p><p>Besides, when the ratio of AT:Ru/C is fixed at 80 : 50 (mg: mg), the conversion of GLU and the corresponding product distribution had been investigated with the change of reaction time. GLU has been consumed mostly after 1 h and the yields of the various products reach their ceiling after the reaction proceeded for 2 h (Figure 1). More specifically, the yield of sorbitol increases rapidly in the first hour, then decreases slowly owing to its hydrogenolysis to glycols in the second hour, and finally keeps constant with the extending time. [22] At the same time, C4 alditol and other sugar alcohols increase steadily and then settle after 2 h. Hence, the reaction time of 2 h is the best time point to obtain the target product. Additionally, C4 alditol and other alditols are steady and no hydrogenolysis or reduction reactions occur, implying the high stability of these alditols. Furthermore, the reaction temperature and H 2 pressure have been tuned to maximize the yields of both C4 and C4 + C6 sugar alcohols and the results were listed in Figure 2. The yield of C6 sugar alcohols dramatically decreases with the increasing reaction temperature, while the low H 2 pressure results in a low detectable carbon at high temperature. This means that high temperature contributes to the retro-aldol process of C6 aldoses rather than hydrogenation. The fast retro-aldol process and slow hydrogenation process bring about low detectable carbon owing to the instability of the retro-aldol products. However, a high H 2 pressure can accelerate the hydrogenation process of all the substrates including hexoses (GLU or MAN) and their retro-aldol products (ERO and GA). The fast hydrogenation can convert timely instable ERO and GA into stable C4 and C2 polyols and assure a high detectable carbon. The different influences of reaction temperature in the retro-aldol and hydrogenation process can be explained by their thermodynamics analyses. According to Scheme 1, the retro-aldol processes of C6 aldoses toward ERO and GA as well as ERO toward GA are process of the particle increase, which means that their entropy changes (ΔS) are positive. Thus, the increasing reaction temperature will lead to a more negative free energy change according to ΔG = ΔHÀ T • ΔS, which is responsible for the right shift of retro-aldol process with the increase of reaction temperature. However, contrary to the retro-aldol process, the hydrogenation processes of C2À C6 aldoses definitely are the process of negative entropy change (ΔS < 0). The free energy change will become more positive when the reaction temperature rises, which implies that high temperature shows an adverse effect on the hydrogenation process. For the H 2 pressure, the high H 2 pressure can promote the shift of hydrogenation reaction from left to right according to Le Châtelier's Principle. For example, the yield of sorbitol increases from 55.0 % to 88.7 % when the pressure of H 2 raises from 2 MPa to 4 MPa at 160 °C. Therefore, for the purpose of the accumulation of C4 + C6 sugar alcohols with high yield and the high detectable carbon, a low reaction temperature and H 2 pressure or a high reaction temperature and H 2 pressure are required. The total yield of C4 + C6 sugar alcohols can finally reach 60 %-70 % and the yield of C4 alditol can be higher than its theoretic one (15 %) under the above two conditions.</p><!><p>In one-pot reaction system, by carefully controlling reaction temperature and H 2 pressure, we can either obtain the highest yield for C4 alditol (20.2 %) and C6 alditol (88.7 %) or achieve over 60 % of the C4 and C6 sugar alcohols co-production. However, as the final product of retro-aldol and hydrogenation of C6 aldoses, whatever the reaction conditions are, the carbon yield of EG is lower than 50 %. Taking competition between retro-aldol condensation and hydrogenation of aldose, it is expected that reducing the instant concentration of substrate could prompt the shift of retro-aldol process towards final product. Hereinafter, we tried to use the semi-continuous method to decrease the instantaneous concentration of substrate in the system, i. e. the substrate with a certain concentration was injected into solution containing catalysts with a certain feeding rate.</p><p>As shown in Figure 3, the yield of C6 alditol is still high under low reaction temperature in spite of low instantaneous concentration of hexose. However, with the increasing reaction temperature, the yield of C6 alditol dramatically decreases and that of EG greatly increases. For example, under 2 MPa of H 2 pressure, the yield of C6 polyol decreases from 49.8 % to 2.9 % with the changing reaction temperature from 150 °C to 190 °C while EG rises from 10.9 % to 75.9 % (Figure 3a and Figure S3). This is much higher than the yield of EG reported by the present literatures on GLU conversion toward EG. [23] Moreover, the yield of EG also increases rapidly with the increase of reaction temperature even though the pressure of H 2 is increased to 4 MPa. These results further confirmed that higher temperature is conductive to the retro-aldol process rather than hydrogenation. However, when the pressure of H 2 is decreased to 1 MPa, all the detectable carbon retains low under all the temperature ranges (Figure S4). Obviously, the H 2 pressure of 1 MPa is not enough to hydrogenate the hexose and the products of its retro-aldol condensation in time under any reaction temperature. Interestingly, the yield of C4 alditol reaches a maximum value at a moderate reaction temperature (Figure 3). Here we can draw an analogy, i. e., taking the reaction system as a balance, C6 alditol and C2 alditol sit at the two ends of the balance, and the temperature is the weight mastering the balance. The higher the temperature is, the more thorough the reaction shifts to C2 alditol. Herein, C4 alditol is like the fulcrum of the balance in the continuous reaction, which reaches a maximum when the productions of C6 and C2 alditol are well-matched in rate.</p><p>In our study, the differences in product distribution of semicontinuous reaction root in the change of the instantaneous concentration of substrate in the process of fluid. To obtain the optimal condition, the conversion of hexoses under feed solution with different concentrations were studied. As illustrated in Table 4, in a particular range of substrate concentration, the yield of EG does not change dramatically, which results from the steady instantaneous concentration. However, when the concentration of substrate rises to a certain amount, the detectable carbon drops sharply, and the yields of EG are also suppressed. Moreover, the effect of the flowing rate on the EG yield displays the similar trend to that of the substrate concentration (Figure S5), and the feeding rate of 0.2 mL/min is finally adopted.</p><p>Similarly, this semi-continuous process can also be applied to some disaccharides (Table 4). For the conversion of disaccharide toward EG, there are three tandem steps, i. e. hydrolysis, retro-aldol condensation and hydrogenation. AT with Lewis acidity is expected to break β-glycosidic bond. Obviously, fast hydrolysis and retro-aldol process are achieved for sucrose, and so a low C6 alditol yield (3.9 %) and a high EG yield (44.8 %) can be observed. Moreover, a considerable amount of C3 products (~30.0 %) containing alditol and 1,2propylene glycol (1,2-PG) is detected since sucrose (SUC) is composed of glucose and fructose. Different from GLU, fructose tends to produce two molecules of C3 compounds via a retroaldol condensation. In the same way, maltose (MAL) can also be hydrolyzed into two molecules of glucose and undergo a series of reactions to acquire products. However, probably because of its slow hydrolysis process, only 34.4 % yield of EG can be obtained while the yield of C6 alditol rises to 15.1 %.</p><!><p>This work has achieved the product control of C6 aldose toward C6, C4 and C2 alditols under the static and semi-flowing conditions, respectively. It is found that a high temperature is in favour of the retro-aldol process of C6 aldoses whereas a high H 2 pressure can promote the hydrogenation process. By the well-matched between condensation and hydrogenation process, we can obtain 20 % of C4 alditol and more than 60 % of C4 and C6 sugar alcohols in the one-pot system. In the semi-flowing condition, the yield of EG can be up to 73 % thanks to its low instantaneous concentration of substrate. Such flexibility on the production of products can achieve a fast anticipation on varying market demands and prices of the produced polyols such as ethylene glycol, 1,2-propanediol, glycerol and erythritol.</p><!><p>Sucrose (SUC, 97 %), Maltose (MAL, 97 %), Mannose (MAN, 99 %), D-Threitol (97 %) and Ammonium Metatungstate (AMT, 99.5 %) were purchased from Shanghai Aladdin Biochemical Technology Co., Ltd. Erythrose (ERO, 75 %), meso-Erythtirol (99 %), Erythrulose (ERU, 85 %), 1,2-propylene glycol (1,2-PG, 98 %), Ammonium Phosphomolybdic (APM, AR) and Ammonium Paratungstate (APT, 99.5 %) were purchased from Shanghai Macklin Biochemical Technology Co., Ltd. Glycolaldehyde dimer (GA, 98 %), Glucose (GLU, 99.5 %), Fructose (99 %) and Ethylene glycol (EG, 99.8 %) were purchased from Sigma-Aldrich. Mannitol (98 %) and Sorbitol (98 %) were purchased from Sinopharm Chemical Reagent Company. Ammonium tungsten oxide (AT, 99.9 %) was purchased from Alfa Aesar. Deionized water was produced by a laboratory water purification system. All other reagents were commercially available and were used as received.</p><p>The hydrogenation catalysts used in the investigation were Ru/C (5 % Ruthenium on activated carbon, Maclin), Pd/C (5 % Palladium on activated carbon, Aladdin) and Pt/C (5 % Platinum on activated carbon, Aladdin). XRD patterns of these catalysts indicated their small metal sizes (Figure S6).</p><!><p>The following procedure was used for the experiments shown above. In a typical one-pot experiment, substrate, catalyst and deionized water (40 mL) were added in a stainless-steel autoclave (Parr Instrument Company, 100 mL). Then, the reactor was closed and flushed 5 times with H 2 . After applying the desired H 2 pressure, stirring was started (400 rpm) and the reactor was heated to the desired temperature. The starting time of the reaction was determined as the point when the reactor reached the desired temperature (approx. 30 min). To stop the reactions, the reactors were allowed to cool to room temperature with cooling water (approx. 10 min).</p><p>In a typical semi-continuous experiment, a certain concentration of C6 aldose solutions was prepared. After reaching the desired temperature and pressure, the substrate solution was injected using pump with the feeding rate of 0.2 mL/min for 100 min into the reactor containing Ru/C (50 mg), AT (80 mg) and 20 mL deionized water which were put into batch in advance, and finally keep reacted for another 20 min.</p><!><p>The reaction solution of 100 μL was taken out and diluted to 1000 μL with deionized water. The sample were analyzed on high performance liquid chromatography (HPLC, Shimadzu Corporation) equipped with refractive index detector (RID). The reaction products were separated using Bio-Rad Aminex HPX-87H column at 35 °C with 8.0 mM H 2 SO 4 aqueous solution as the mobile phase at the flow of 0.6 mL/min. Meanwhile, the reaction products were separated using COSMOSIL sugar-D column (4.6 mm lD × 300 mm) at 40 °C with an aqueous solution containing 90 % acetonitrile as the mobile phase at the flow of 0.8 mL/min. Before being injected into HPLC the samples needed to be filtered through a micro syringe filter. The retention time of detectable sugars and sugar alcohols by Bio-Rad Aminex HPX-87H column were as follow: GLU (8.7 min), MAN (9.4 min), FRU (9.6 min), ERO (11.4 min), GA (11.7 min), ERU (11.8 min), and C4 alditol can be detected by COSMOSIL sugar-D column at 11.9 min. Each product, as well as reactant, were calibrated by using its standard at different concentrations at their specific retention times. High-resolution mass spectra of reaction products were recorded in Bruker McriOTOF II mass spectrometer.</p><p>For the preparation of alditol from aldoses, conversions of substrates and carbon yields of products were calculated as follows:</p>

Chemistry Open

Neuron-targeted copolymers with sheddable shielding blocks synthesized using a reducible, RAFT-ATRP double-head agent

Successful adaptation of in vitro optimized polymeric gene delivery systems for in vivo use remains a significant challenge. Most in vivo applications require particles that are sterically stabilized but doing so significantly compromises transfection efficiency of materials shown to be effective in vitro. In this communication, we present a multi-functional well-defined block copolymer that forms particles with the following properties: cell targeting, reversible shielding, endosomal release, and DNA condensation. We show that targeted and stabilized particles retain transfection efficiencies comparable to the non-stabilized formulations. The block copolymers are synthesized using a novel, double-head agent (CPADB-SS-iBuBr) that combines a RAFT CTA and an ATRP initiator through a disulfide linkage. Using this double-head agent, a well-defined cationic block copolymer P(OEGMA)15-SS-P(GMA-TEPA)50 containing a hydrophilic oligoethyleneglycol (OEG) block and a tetraethylenepentamine (TEPA)-grafted polycation block was synthesized. This material effectively condenses plasmid DNA into salt-stable particles that deshield under intracellular reducing conditions. In vitro transfection studies showed that the reversibly shielded polyplexes afforded up 10-fold higher transfection efficiencies compared to the analogous stably-shielded polymer in four different mammalian cell lines. To compensate for reduced cell uptake caused by the hydrophilic particle shell, a neuron-targeting peptide was further conjugated to the terminus of theP(OEGMA) block. Transfection of neuron-like, differentiated PC-12 cells demonstrated that combining both targeting and deshielding in stabilized particles yields formulations that are suitable for in vivo delivery without compromising in vitro transfection efficiency. These materials are therefore promising carriers for in vivo gene delivery applications.

neuron-targeted_copolymers_with_sheddable_shielding_blocks_synthesized_using_a_reducible,_raft-atrp_

<p>Polycations are an attractive class of material for gene delivery because they self-assemble with and condense nucleic acids, can be synthesized at large scale, and offer flexible chemistries for functionalization.1 In the past few decades, many polymer compositions and architectures have been synthesized. In vitro screening of these polymers has yielded many materials that efficiently transfect cultured mammalian cells.2 However, only a small subset of these materials issuitable for in vivo use due to additional extracellular barriers. Polyplexes, complexes of polycations and nucleic acids, are colloids typically unstable in physiological conditions.3 Polyplexes are prone to protein adsorption and aggregation, which can lead to inflammation and mortality. Furthermore, when used in vivo, polyplexes must preferentially transfect the target cell type rather than the vast majority of other cells that are present.</p><p>To address the aggregation issue, a hydrophilic polymer shell, such as poly(ethylene glycol) (PEG) or N-(2-hydroxypropyl) methacrylamide (HPMA), is incorporated into most polyplex formulations designed for in vivo use.4 However, polyplexes shielded against protein adsorption and aggregation are also poorly recognized and internalized by cells, thereby compromising transfection efficiencies. Formulations with reversible deshielding properties that are usually triggered by acid conditions (found in tumor microenvironments or in endosomes after cellular uptake) or reducing conditions (found in the cell cytosol) have been investigated and shown to generally be more efficient than formulations with stable polymer shields. Notable examples that have been tested for in vivo delivery are (i) Wagner's ternary complexes containing targeting ligand conjugated to one polycation, PEG conjugated to a second polycation via an acid-labile hydrazone, and plasmids, (ii) Wang's ternary complexes of polyplexes coated with a charge-reversing polymer that deshield in mildly acidic environments, and (iii) Kataoka's block copolymers of PEG and poly[Asp(DET)] that deshield in reducing conditions.5 However, to our knowledge, the reported formulations to date either do not incorporate targeting ability and/or require multiple polymer components in the final structures, complicating scale-up and manufacturing.</p><p>The major advancement reported in this work is the development of a polymeric nucleic acid carrier that incorporates cell targeting, reversible colloidal stability and efficient intracellular delivery into a single well-defined material (Scheme 1). The key enabling technology for this material is a novel, reducible double-head agent consisting of both a reversible addition fragmentation chain transfer (RAFT) agent and an atom transfer radical polymerization (ATRP) initiator connected by a disulfide bond (Figure 1a). We used this double-head agent to synthesize an optimized reduction-responsive cationic block copolymer P(OEGMA)-SS-P(GMA-TEPA) by a combination of RAFT polymerization of oligo(ethylene glycol) monomethyl ether methacrylate (OEGMA) and ATRP of glycidyl methacrylate (GMA), followed by post-polymerization decoration of reactive epoxy groups in P(GMA) block by tetraethylenepentamine (TEPA) (Figure 1b). We demonstrate that diblock copolymers synthesized using this double-head agent possess several notable qualities. First, they have well-controlled composition with narrowly distributed molecular weight. Second, they can be easily modified with a targeting lig-and at the outer corona. Finally, the "in vivo ready" diblock formulation shows similar transfection efficiency as the in vitro optimized polycation segment due to the reversible shielding combined with targeting ability. The neuron targeting peptide, Tet1-conjugated P(OEGMA)15-SS-P(GMA-TEPA)50 is expected to transfect cells by condensing DNA efficiently to form core-shell type polyplexes with the P(GMA-TEPA)/DNA electrostatic complex building the core and the P(OEGMA) block constructing the shell. Once internalized, the polyplexes become localized within the endocytic vesicles. The protonatable amines in TEPA were included to facilitate endosomal escape through the proton sponge effect6 and glutathiones (GSH) in the intracellular environment are expected to degrade the disulfide links, leading to detachment of the hydrophilic P(OEGMA) coating and release of DNA (Scheme 1).</p><p>The double-head agent CPADB-SS-iBuBr was synthesized by N,N′-dicyclohexylcarbodiimide (DCC) coupling between the RAFT CTA 4-cyanopentanoic acid dithiobenzoate (CPADB)7 and 2-hydroxyethyl-2′-(bromoisobutyryl)ethyl disulfide initiator (OH-SS-iBuBr)8 in the presence of 4-(dimethylamino)pyridine (DMAP) as a catalyst. After purification by column chromatog-raphy, CPADB-SS-iBuBr was successfully obtained with 43.1% yield (Figure 1a).</p><p>The most commonly-used hydrophilic shell employed in polyplex stabilization is PEG due to the facile incorporation of PEG by various methods, such as using a PEG-based macro-initiator5a,9, macro-chain transfer agent (CTA)10 or via coupling reaction11. Coupling reaction through disulfide-thiol exchange usually requires synthesis of both a mercapto-functionalized and a pyridyl disulfide-functionalized polymer,11 therefore involves reaction between large macromolecules as well as extra separation steps to remove the excess homopolymers from desired block copolymers. In addition to PEG, OEGMA and HPMA are two biocompatible, hydrophilic, shell-building units of nano-carriers.12,13 However, there are no published reports thus far on the synthesis of well-defined copolymers containing sheddable OEGMA and HPMA blocks by controlled living radical polymerization (CLRP). Chain extension using different monomers by consecutive CLRP processes always leads to block copolymers with non-degradable C-C links in the block junctions. Recently Oh et al. reported the synthesis of poly(lactide)-SS-polymethacrylate amphiphilic block copolymers with SS linkages positioned at the block junction by ring-opening polymerization (ROP) and ATRP,8 but these methods are limited to the application of cyclic monomers.</p><p>The CPADB-SS-iBuBr double-head agent developed herein offers a simple and versatile means to prepare block copolymers based on diverse hydrophilic and hydrophobic monomers with cleavable links in the block junctions for various applications. The design principle of the double-head agent is to integrate RAFT and ATRP techniques, taking advantage of their different mechanisms for orthogonal synthetic strategy. The resulting double-head agent was first used as a RAFT CTA to carry out the polymerization of both OEGMA (Mn~300) and HPMA. The RAFT kinetics of OEGMA (Table S1) and HPMA (Table S2) polymerization using CPADB-SS-iBuBr was monitored by 1H NMR and SEC-MALLS. Both monomers were polymerized with first order kinetics (Figure S2a-c) and low PDI (< 1.3), demonstrating excellent synthesis control via RAFT polymerization.</p><p>The hydrophilic OEGMA and HPMA blocks were then used as a macroinitiator for ATRP of GMA. To minimize the possibility of concurrent RAFT-ATRP14, GMA, which is polymerized with fast kinetics by ATRP but with slower kinetics by RAFT, was polymerized with short reaction time. We demonstrate by NMR and GPC analysis that polymerization of GMA occurs predominantly by ATRP (see Supporting Information). Block copolymers of P(OEGMA)-SS-P(GMA) were characterized by 1H NMR to verify successful polymerization and to determine the degree of polymerization (DP) (Figure S1b). P(GMA) contains pendant reactive epoxy groups that were further functionalized by TEPA to generate the polycation block. The final diblock copolymer was characterized by 1H NMR (Figure S1c) and SEC-MALLS (Figure S2d). We, in earlier work, optimized the polycation block, screening various oligoamine and lengths of polymer backbone for optimal cell transfection efficiency (data not shown). Based on this initial screen, we identified P(GMA) of DP 50 grafted with TEPA, P(GMA-TEPA)50 (Mn = 22.5 kDa, PDI = 1.11, dn/dc = 0.216) to be the most effective carrier. Its transfection efficiencies were comparable to that of branched poly(ethylenimine) (bPEI, 25 kDa). We then further tested diblocks of P(GMA-TEPA)50 with P(OEGMA) and P(HPMA) of various lengths and identified an optimal material, P(OEGMA)15-SS-P(GMA-TEPA)50 (Mn = 31.5 kDa, dn/dc = 0.202, PDI = 1.29). As a control, the reduction-insensitive P(OEGMA)15-b-P(GMA-TEPA)50 copolymer (Mn= 30.1 kDa, PDI = 1.28, dn/dc = 0.201) was also synthesized by consecutive RAFT polymerizations using CPADB as a CTA. To incorporate cell targeting, we selected the neuron targeting pep-tide Tet1, which we previously have shown to facilitate targeted transfection both in vitro and in vivo when grafted to PEI.15 A N-terminus maleimide-functionalized Tet1 was conjugated by Michael-type addition to the terminal free thiols of P(OEGMA)15-SS-P(GMA-TEPA)50 and P(OEGMA)15-b-P(GMA-TEPA)50 that were generated by aminolysis of the dithioester end group during TEPA functionalization (Scheme 2). Tet1 functionalization efficiency was ∼33%, likely due to competing thiolactone formation16 (see Supporting Information).</p><p>The DNA binding of polymers was investigated by agarose gel retardation assay. The results (Figure S5) indicate that all the polymers exhibit similar DNA condensation above an N/P (amino to phosphate) ratio of 4∼5, and Tet1 peptide conjugation does not significantly affect DNA binding of block copolymers. The morphologies of the five polyplex formulations were visualized by TEM (Figure 2a) at an N/P of 5. All materials condensed plasmid DNA into compact particles with diameter < 50 nm. Polyplexes formed using block copolymers were more compact than those formed by the P(GMA-TEPA)50 homopolymer, and polyplexes containing Tet1 modification were more polydisperse. The uniformity of polyplex morphology and size might be improved in future work by controlling formulation, for example by slow mixing or microfluidics-facilitated mixing as opposed to the bulk mixing used here17.</p><p>Poor salt stability has been a significant obstacle for in vivo application of unshielded polyplexes.18,19 The salt stability of polymer/DNA complexes was studied in both PBS (150 mM, pH 7.4) and Opti-MEM using dynamic light scattering (Figure 2b and Figure S6 for full kinetics). The larger size measured by DLS compared to TEM might be attributed to minority populations of larger particles that skew the average diameters measured by light scattering and also to measurement in salt medium versus water. Only a slight increase in particle size was observed for the polyplexes formed with the block copolymers over a period of 20 h following the addition of physiological levels salt or Opti-MEM media, demonstrating excellent colloidal stability of formed polyplexes. In contrast, polyplexes of P(GMA-TEPA)50 homopolymer formed large aggregates with diameter above 1000 nm within 1 hr under the same conditions (Figure S6c). The overall results confirm that the 4.5 kDa P(OEGMA) block provides sufficient extracellular colloidal stability for P(GMA-TEPA)50 polyplexes. To investigate whether the polyplexes formed using reducible block copolymers would be deshielded in the cytosol to facilitate DNA release in the intracellular reducing environment, the particle sizes of polyplexes in the presence of 10 mM dithiothreitol (DTT) were monitored over time. Polyplexes of Tet1-P(OEGMA)15-SS-P(GMA-TEPA)50 and P(OEGMA)15-SS-P(GMA-TEPA)50 were found to increase in size due to reduction-triggered deshielding whereas polyplexes formed from the nonreducible polymers were stable in size. Hence, polyplexes of P(OEGMA)15-SS-P(GMA-TEPA)50 may have excellent colloidal stability in the circulation and still be readily destabilized in the cell cytosol, facilitating DNA release.</p><p>The in vitro transfection efficiency of reducible and non-reducible polyplexes was evaluated in four different cell lines, HeLa, HEK293T, HepG2, and 2-day differentiated PC-12 cells by luciferase assay, using P(GMA-TEPA)50 as a control. Figure 3a summarizes the transfection data for these four different cell lines. The reversibly shielded polyplexes formed by P(OEGMA)15-SS-P(GMA-TEPA)50 mediated significantly higher transfection efficiency in all four cell lines compared to the stably shielded analogue under identical conditions, affording up to 10-times higher transfection efficiency depending on the cell type. The results confirm that the reducible disulfide bond can improve transfection activity. However, the reducible polymers are less efficient than the in vitro optimized homopolycation P(GMA-TEPA)50. This is expected since the hydrophilic shielding layers have been shown to inhibit polyplex uptake.20</p><p>To address decreased polyplex uptake, the targeted transfection efficacy of Tet1-conjugated polyplexes was further assessed in 6-day differentiated PC-12 cells. Differentiated PC-12 cells display a neuron-like phenotype that includes increased binding of the Tet1 peptide.15b The results (Figure 3b) clearly show that conjugation of the Tet1 targeting peptide significantly enhanced transfection compared to corresponding polymer lacking Tet1. Of all the block copolymers, the Tet1-(OEGMA)15-SS-P(GMA-TEPA)50 displays the highest transfection efficacy. Its transfection efficacy is 50-fold higher than non-reducible, non-targeted complexes, 6.1-fold higher than non-targeted, reducible complexes and 2.6-fold higher than non-reducible targeted complexes. Most importantly, polyplexes that include both targeting ligand and releasable shielding coronas transfect target cells with similar efficiencies as the homopolycation.</p><p>In summary, we have successfully developed a versatile method for preparing functionalizable reduction-sensitive block copolymers by integrated RAFT and ATRP techniques using a novel, reducible double-head agent. Here, we prepared a neuron-targeted copolymer for nucleic acid delivery applications. We further showed that the resulting materials form particles that are salt stable but due to the combined properties of targeting and shielding still retain high transfection efficiencies comparable to the analogous homopolycation vectors for targeted gene delivery. The approach developed herein provides a versatile means for preparing various types of multifunctional drug and gene delivery vehicles.</p>

PubMed Author Manuscript

Ultrasensitive Online SERS Detection of Structural Isomers Separated by Capillary Zone Electrophoresis

A mixture of structural isomers was separated and identified at nanomolar concentrations (~100,000 molecules) by incorporating capillary zone electrophoresis (CZE) with a sheath flow surface-enhanced Raman scattering (SERS) detector. Baseline resolution was obtained from three structural isomers of rhodamine using a planar silver SERS substrate, demonstrating the utility of this approach for trace chemical analysis.

ultrasensitive_online_sers_detection_of_structural_isomers_separated_by_capillary_zone_electrophores

<p>The ability to identify and characterize molecules purified through separation lies at the heart of chemical analysis. For column-based separations, common methods of detection include UV-visible absorption, laser-induced fluorescence (LIF), and mass spectrometry. Despite its low cost and flexibility, on-column UV-visible absorption suffers from poor molecular specificity and a lack of sensitivity.1, 2 On the other hand, LIF offers a high degree of sensitivity but requires fluorescent labels.3–5 Since structure determination by migration times alone requires extensive knowledge of the samples beforehand, the use of these two methods is limited for explicit analyte characterization. Mass spectrometry provides exquisite analyte identification for many samples. However, many classes of molecules, such as structural isomers and other molecules with the same mass (isobars) are still challenging to characterize. The cost of high-resolution mass spectrometers necessary for characterizing similar compounds limits the utility of this technique for routine characterization.6, 7 As a result, there is a need for new detection techniques capable of providing structural information with high sensitivity and selectivity for chemical analysis.</p><p>Here we demonstrate surface-enhanced Raman scattering (SERS) for characterization of three rhodamine isomers separated by capillary zone electrophoresis (CZE). CZE is a powerful analytical technique for separation of charged analytes8, 9 and has been incorporated into microfluidic devices for high efficiency separations10–12. SERS provides extensive structural and quantitative information about a variety of molecules based on their vibrational transitions13 and can be readily performed in solution to facilitate detection in-line with chemical separations.14 Given these attributes, SERS has the potential to provide chemical identity of solutes following CZE separation.</p><p>There have been previous attempts to couple SERS to CZE. In these studies, CZE-SERS was accomplished by interfacing detection directly on-column or at-line. Direct on-column SERS detection has been achieved using running buffers containing silver colloidal solutions and by laser-induced growth of silver particles at the end of the capillary.15, 16 The use of colloidal particles has shown detection limits in the nM or pM range; however, memory effects commonly prevent the regeneration of the detection window and limit these configurations to a one-time-use only. Planar SERS substrates in CZE suffer an additional challenge; specifically, a metal in an electric field will form a bipolar electrode and cause electrochemical formation of bubbles and degradation of the sample.17 In-line CZE-SERS with planar substrates has been limited to μM limits of detection.16 An at-line approach to CZE-SERS deposits the effluent onto a moving substrate.18 Drying the sample adsorbs molecules to the surface and avoids challenges associated with mass transport. This approach also avoids challenges associated with the formation of a bipolar electrode across the SERS substrate; however, designing an interface that guarantees maintenance of the electrical current during the deposition onto the substrate is not trivial.</p><p>By incorporating our recently demonstrated sheath flow SERS detector,19 we are able to circumvent the challenges noted above and achieve online detection in CZE separations. In particular, the potential drop (bipolar electrode formation) across the SERS substrate is minimized by the increased volume of the sheath flow and confined sample near the electrical ground. Changes observed in the silver oxide background signals suggest a small electrochemical potential is still present. However, we have successfully used the same SERS substrate in CE applications for up to three days without significant signal degradation. The sheath flow SERS detector enables sequential and high throughput detection of the separated dyes at nanomolar concentrations (attomole - femtomole injections) using a 50 ms acquisition without significant "memory effect" or fouling of the SERS substrate.</p><p>The sheath flow SERS detector was coupled online to a CZE system. The CZE system is similar to the one previously reported except for the detection module.20, 21 CZE separation was performed in positive mode on a 50 cm bare fused silica capillary (Polymicro Technologies, Phoenix, AZ) with 72 μm i.d. and 143 μm o.d.. A constant potential of 300 V/cm was supplied by a Spellman, CZE 1000R power supply (Spellman High Voltage Electronics Corp., Hauppauge, NY).21 The sample, containing 10−8 M rhodamine 6G (R6G), 10−10 M rhodamine B (RB), and10−7 M 5-carboxytetramethylrhodamine (5-TAMRA), was prepared in 15 mM sodium tetraborate buffer (pH 9.4). The CZE separation was performed using a 2 s pressure injection, which injects 34 nL of sample. After injection, the capillary was placed in 15 mM sodium tetraborate buffer solution (pH 9.4) and 15kV (~40μA) was applied to the Pt electrode at the sample end of the capillary. SERS measurements were performed in kinetic series with 50 ms acquisition times and by using a sheath flow rate of 10 μL/min (a sheath flow to capillary flow rate ratio of 100:1). The Raman spectrometer used in this study has been previously described.22 Raman scattering was detected from a 633 nm laser, away from the absorption band of the rhodamine dyes and thus without the benefit of resonance enhancement. Full details on the instrument setups and experimental procedures are provided in the ESI. Figure S1 presents the schematic of the experimental setup used for the CZE-SERS experiments.</p><p>Figure 1A shows the heatmap of the SERS intensity as a function of Raman shift and migration time following the electrophoretic separation of three rhodamine isomers (R6G, RB, and 5-TAMRA). The Raman spectrum observed indicates that R6G migrates at tm=180 ± 13 s, RB at tm=220 ± 19 s, and finally 5-TAMRA at tm=290 ± 15 s. The SERS signal for each peak persists for about 1–2 s or less at these low concentrations. The short duration of the SERS signal is more clearly observed in the 2 s zooms shown in Figure 1B, which illustrate the difference in width of each migration peak.</p><p>Figure 1C shows the SERS electropherogram constructed from the SERS intensity at 1357 cm−1 as a function of migration time. This band is attributed to the combined aromatic C-C and C=N stretching modes of rhodamine compounds.23–27 The intensity profile at 1357 cm−1 provides a convenient signal to characterize the separation efficiency with SERS detection. The spectrally resolved SERS electropherogram of the three rhodamine dyes is characterized by a low and constant background.</p><p>Analysis of the SERS electropherogram (Figure 1C) shows a peak for R6G at t=180.25 s with a full width at half max (FWHM) of 1.25 s, which suggests a separation efficiency of N = 115,000 ± 35,000 theoretical plates. The SERS electropherogram peak for RB at t=219.75 s shows a more symmetric peak with a FWHM of 0.55 s. This corresponds to N = 898,000 ± 115,000 theoretical plates. The electropherogram peak for 5-TAMRA at t= 290.60 s has a FWHM of 0.40 s, which corresponds to a separation efficiency of N = 2,900,000 ± 620,000 theoretical plates. Because our analytes fluoresce when excited at shorter wavelengths, we performed laser-induced fluorescence (LIF) to compare the migration times and separation efficiency. Figure S-2 shows the electropherogram of the same three analyte mixture using LIF detection. The analyte concentrations and separation conditions were kept identical to those used in the optimized SERS experiments to provide a direct comparison. The LIF electropherogram shows three bands associated with the elution of R6G, RB, and 5-TAMRA with a separation efficiency N=1,000 – 6,000 theoretical plates (analyte dependent), which is low for a CZE separation with LIF detection. The poor separation efficiencies are the result of the large injection volume and the high concentration of analytes used for the CZE-LIF experiments. However, CZE-SERS and CZE-LIF generated identical elution order and equivalent migration times under identical separation conditions.</p><p>The difference in observed number of theoretical plates provides insight into the mechanism of SERS detection. Only molecules located within a close proximity to the SERS substrate surface can be detected. Our previous work suggests the observed signal arises from adsorbed molecules. However, it is known that Langmuir behavior inhibits analyte adsorption at low concentrations, typically below 1 nM.28 We have successfully detected RB at a concentration below this in Figure 1. This suggests that hydrodynamic confinement may provide a transiently increased concentration at the surface, such that the SERS detection is only obtained from the highest concentration portion of the migrating analyte band. This is in contrast to LIF, where the greater sensitivity enables detection of the width of the entire eluting sample. In Figures S-3 and S-4, we show the SERS results from a longer injection and a higher concentration of analytes. The apparent efficiency with SERS detection decreases to a level comparable to LIF due to molecules remaining on the surface for longer periods.</p><p>The role of adsorption is further evident in the width of the R6G peak relative to the widths observed for RB and 5-TAMRA. Increased adsorption of R6G to the surface results in a longer observed peak width in the SERS electropherogram (Figure 1C), suggesting R6G has a stronger binding affinity for silver surfaces than RB and 5-TAMRA. The CZE-SERS efficiency appears to correlate to the sample desorption rate from the substrate. In these results, the longer desorption rate observed for R6G can be directly attributed to the difference in molecular structure of the three rhodamine dyes (Figure S-5). R6G is the only dye out of the three containing a secondary amine group. The pKa of this amine group has a value of 6.13. When dissolved in borate buffer (pH 9.4), the basic form of R6G predominates (pH>pKa). As a result, the secondary amine group is deprotonated and more electron rich. Under these conditions, the nitrogen atom on the R6G molecules is more likely to adsorb to the silver SERS substrate than the other amine groups in RB and 5-TAMRA. These properties explain the higher affinity of R6G for the silver SERS substrate and the resulting slower desorption mechanism observed for this rhodamine dye. Despite these variations, it is worth noting that baseline resolution is achieved between each analyte, demonstrating no memory effects.</p><p>In our earlier publication,19 our SERS detector demonstrated a linear response from nM to μM concentrations of R6G, indicating that quantitation is possible for trace analyte detection. Our previous work further showed that chemical effects alter the desorption rate, which we are investigating to understand their impact on the observed separation.</p><p>The main advantage of using SERS over conventional detection techniques (UV and LIF) is that it can provide chemical information to identify and characterize analytes beyond migration times. Figure 2a shows a single 50 ms SERS spectrum of R6G (10−8 M) from the electrophoretic separation of the three dye mixture extracted from Figure 1A at tm=180.25 s. The main features of the R6G spectrum are the bands at 1175, 1306, 1357, 1506, and 1648 cm−1. These bands are associated with the characteristic stretching modes of the C-H band, C=N, and aromatic C-C stretching vibrations of R6G.23–27 Figure 2b shows a single 50 ms SERS spectrum of RB (10−10 M) extracted from Figure 1A at tm=219.75 s. The RB bands are assigned to the aromatic C-H bending (1197 cm−1), the C-C bridge-bands stretching (1276 cm−1), and the aromatic C-H bending vibrations (1357 cm−1, 1506 cm−1, and 1645 cm−1).29 Finally, Figure 2c shows a single 50 ms SERS spectrum of 5-TAMRA (10−7 M) extracted from Figure 1A at tm=290.60 s. The main features of the 5-TAMRA spectrum are the bands at 1197, 1276, 1354, 1506, and 1643 cm−1. These bands are assigned to the aromatic C-H bending, C-C bridge-band stretching, and aromatic C-C stretching modes of 5-TAMRA.30</p><p>Averaging the SERS signal over the duration of the electropherogram peak yields spectra with a S/N ratio ≥ 25 for all three analytes. Figure 2d shows the average SERS spectrum of R6G extracted from Figure 1A between t=179.65 and 180.05 s. The averaged SERS spectrum of RB extracted between t=220.45 and 220.90 s is shown in Figure 2e. Of note, the SERS spectrum of RB was acquired from the injection of a few attomoles (~100,000 molecules). Figure 2f shows the averaged spectrum of 5-TAMRA extracted from Figure 1A between t=290.45 and 290.80 s. While all three dyes show similar spectra, as expected based on their structures, the differences observed enable identification of the analytes.</p><p>In conclusion, we have demonstrated highly sensitive and ultrafast online SERS detection of structural isomers of rhodamine separated by CZE. SERS spectra of the analytes provided direct spectral signatures associated with the subtle structural differences of the three rhodamine dyes. The limit of detection for SERS reported here is more than 1000x better when compared to the best previously reported LOD using a planar substrate.16 The observed Raman scattering allowed differentiation of two isobaric compounds (R6B and RB, M.W=479.02 g/mol) at nanomolar concentrations, which is not achievable by mass spectrometry. The SERS flow detector should be readily incorporated into any liquid separation, such as liquid chromatography. The implementation of this robust and sensitive online SERS flow detector suggests an alternative for the characterization of pharmaceuticals, metabolites, and other analytes.</p>

PubMed Author Manuscript

Anchored plasticity opens doors for selective inhibitor design in nitric oxide synthase

Nitric oxide synthase (NOS) enzymes synthesize nitric oxide, a signal for vasodilatation and neurotransmission at low levels, and a defensive cytotoxin at higher levels. The high active-site conservation among all three NOS isozymes hinders the design of selective NOS inhibitors to treat inflammation, arthritis, stroke, septic shock, and cancer. Our structural and mutagenesis results identified an isozyme-specific induced-fit binding mode linking a cascade of conformational changes to a novel specificity pocket. Plasticity of an isozyme-specific triad of distant second- and third-shell residues modulates conformational changes of invariant first-shell residues to determine inhibitor selectivity. To design potent and selective NOS inhibitors, we developed the anchored plasticity approach: anchor an inhibitor core in a conserved binding pocket, then extend rigid bulky substituents towards remote specificity pockets, accessible upon conformational changes of flexible residues. This approach exemplifies general principles for the design of selective enzyme inhibitors that overcome strong active-site conservation.

anchored_plasticity_opens_doors_for_selective_inhibitor_design_in_nitric_oxide_synthase

<!>Inhibitor binding to iNOSox<!>A novel Gln specificity pocket in iNOSox<!>Isozyme differences in inhibitor binding<!>Determinants for inhibitor selectivity<!>Design and synthesis of selective iNOS inhibitors<!>DISCUSSION<!>Expression and purification of NOSox proteins<!>Mutagenesis<!>Synthesis of Compounds 1\xe2\x80\x9316<!>Binding affinity, inhibitory potency and in vivo activity<!>Crystallization, data collection and refinement<!>

<p>Nitric oxide (NO) is a small, diffusible, and transient molecule produced from amino acid L-arginine (L-Arg) by three nitric oxide synthase (NOS) enzymes1,2. The endothelial (eNOS) and neuronal (nNOS) NOS isozymes are constitutively expressed and Ca2+ regulated to provide NO for signaling, including vasodilatation, thermoregulation, neuroprotection, and endocrine function. The Ca2+-insensitive inducible NOS isozyme (iNOS) is expressed in response to cytokines or pathogens, and produces NO at a high rate to kill bacteria, viruses, and tumor cells. Insufficient NO bioavailability from eNOS and nNOS is associated with hypertension, impotence, atherosclerosis and cardiovascular disease, while excess NO from iNOS has been implicated in inflammation, rheumatoid arthritis, inflammatory bowel disease, immune-type diabetes, stroke, cancer, thrombosis, and infection susceptibilities3,4. Overproduction of NO by iNOS (and nNOS) has also been linked to neurodegenerative disorders including Parkinson's and Alzheimer's diseases, as well as multiple sclerosis5. Thus, the development of iNOS-specific inhibitors is highly desirable.</p><p>The three NOS isozymes share a common modular architecture and conserved active site. The N-terminal catalytic oxygenase module (NOSox) binds cofactors heme and (6R)-5,6,7,8-tetrahydro L-biopterin (H4B), substrate L-Arg, and a structural zinc ion across the dimer interface6–15. Upon calmodulin binding16, NOSox accepts electrons from the C-terminal reductase module17–19. The nearly complete amino acid conservation and structural similarity among the three NOS isozymes6–15 active sites presents a significant challenge for the design of isozyme-specific inhibitors20,21. Moreover, as NO availability is controlled at the synthesis level for signaling or cytotoxicity, NOS isozymes are a paradigmatic system to address the challenges of designing isozyme-specific inhibitors despite conserved binding pockets.</p><p>NOS inhibitors selective for iNOS are rare, and commonly exhibit only limited selectivity or significant toxicity22–25. In contrast, quinazoline26 (1–5) and aminopyridine27,28 (6–12) inhibitors possess good in vitro potency and selectivity for iNOS. In particular, the spirocyclic quinazoline (AR-C102222, 3, Fig. 1) shows excellent selectivity over eNOS (3000-fold), and exhibits significant protective, anti-inflammatory and antinociceptive activities in rodent models of adjuvant-induced arthritis, pancreatitis29, neuropathy, inflammation, and post-surgical pain30. Thus, we have chosen to focus our structural studies on quinazoline and aminopyridine inhibitors.</p><p>Here, we combined mutagenesis, biochemistry, crystallography, and drug design to elucidate the structural basis for the iNOS selectivity of some quinazoline and aminopyridine inhibitors. We demonstrate that plasticity of an isozyme-specific triad of residues distant from the active site modulates conformational changes of invariant residues nearby the active site to determine the exquisite selectivity of these inhibitors for iNOS. We design novel potent and selective iNOS inhibitors by applying an "anchored plasticity approach" (Supplementary Fig. 1 online). Selective inhibitors are designed with an inhibitor core anchored in a conserved binding pocket, and rigid bulky substituents that extend to remote specificity pockets accessible upon conformational changes of "plastic" protein residues. Fundamentally, this anchored plasticity approach is broadly applicable to the discovery of novel inhibitors against enzyme families with strong active-site conservation.</p><!><p>Quinazoline (1–2), spirocyclic quinazoline (3–5), and aminopyridine (6–12) inhibitors are potent (IC50 from 10 nM to 1.2 μM) and selective (2.7- to 3000-fold) inhibitors for iNOS over eNOS and nNOS (Fig. 1 and Supplementary Table 1 online). These inhibitors share a cis-amidine derived core, but have different substituents or tails. To determine the basis for the exquisite iNOS potency of these inhibitors, we solved x-ray structures of murine iNOSox bound to compounds 1–12 and of human iNOSox bound to aminopyridine 9 (Methods).</p><p>Inhibitors 1–5 and 6–12 belong to different chemotypes but all bind similarly in the iNOS active-site heme pocket (Fig. 2a–d, Supplementary Fig. 2 online). The NOSox active site is lined by the heme, invariant Glu (Glu371/377; murine/human iNOS numbering, respectively) and backwall residues (363–366/369–372). In all these inhibitor complexes, the cis-amidine moiety mimics the guanidinium group of substrate L-Arg, by making bidentate hydrogen bonds to Glu and stacking with the heme. Compounds 1–8 make an extra hydrogen bond to the main-chain carbonyl of invariant Trp366/372 and pack more parallel to the heme than compounds 9–12 (Supplementary Results). The bulky and rigid tails of compounds 2–5 and 9–12 all extend above heme propionate A and pack with invariant residues Gln (Gln257/263), Arg (Arg260/266), Pro344/350, Ala345/351 (not shown in Fig. 2), and Arg382/388. Hydrogen bonds tether the extended inhibitor tails to invariant Tyr (Tyr341/347), and either Arg382/388 (compound 2) or a water molecule (compounds 3–5 and 12). Our structural analysis thus suggests that both interactions of the inhibitor core with active-site residues and of the inhibitor tail with residues outside the active-site heme pocket mediate inhibitor binding.</p><p>To determine the roles of residues key to inhibitor binding, we measured the binding affinity and inhibitor potency of moderately selective compound 9 for several human iNOS mutant proteins (Table 1). Mutation of active site Glu into Ala had the most dramatic effect (KD = 0.4 μM for wild type vs. ≫ 100 μM for E377A mutant), thus revealing the crucial role of this invariant charged side chain in inhibitor, as well as substrate31, binding. The close match between binding affinity (KD) in iNOSox and inhibitory potency (IC50) in full-length iNOS (Table 1) suggests that enzyme inhibition data reflects true binding affinity. Both Gln mutations (Q263A, loss of side chain and Q263N, smaller side chain with similar functionality) result in slightly decreased iNOS affinity for compound 9 (3- and 5-fold, respectively), thus corroborating the role of Gln in inhibitor tail binding. The human iNOS Y347F/Y373F double mutant displayed a ~10-fold decrease in inhibitor potency, thus revealing a key role for the Tyr hydrogen bond to the carbonyl of the inhibitor tail. Our combined structural and mutagenesis results thus suggest that heme stacking, as well as hydrogen bonds and hydrophobic interactions within and outside the active site, all significantly contribute to the binding of these inhibitors to iNOS.</p><!><p>Bulky inhibitors promote a cascade of conformational changes up to 20 Å away from the iNOSox active site, resulting in the creation of a new pocket. The comparison of our iNOS x-ray structures with small (compounds 1, 6–7) and large-tailed (2–5 and 8–12) inhibitors reveals similar overall protein structures, with all inhibitor cores anchored in the active-site heme pocket (Fig. 2a–d and Supplementary Figs. 2 and 3 online). Yet, outside this pocket, invariant first-shell and second-shell residues adopt different side-chain conformations in these complexes (Fig. 2a–e, Supplementary Fig. 3, Supplementary Results online). To prevent collision with bulky inhibitors, the first-shell Gln side chain rotates around its χ1 and χ2 torsion angles from a "Gln-closed" position with hydrogen bond to Tyr, to a "Gln-open" position with hydrogen bonds to Arg. Similarly, the Arg side chain rotates closer to second-shell residues Asp274/280 and Asn (Thr277/Asn283). Finally, side-chain rotation of Arg382/388 closer to Asp376/382 enhances hydrophobic interactions with the inhibitor tail.</p><p>Upon binding of bulky inhibitors, the coupled rotations of first-shell Gln and Arg initiate a cascade of conformational changes, which further propagate to second-shell residues. In the human iNOS Gln-open conformation, the conformational change of first-shell Arg induces rotation of second-shell Asn towards third-shell Phe286 and Val305 (Fig. 2d, Supplementary Fig. 4 and Supplementary Movie 1 online). Thus, the conformational plasticity of human iNOS second-shell Asn allows coordinated movements of first-shell Gln and Arg.</p><p>The correlated side-chain rotations of Gln, Arg, and Arg382/388 to accommodate the rigid bulky tails of compounds 2–5 and 9–12 expose a new specificity pocket for enhanced inhibitor binding in iNOS. This "Gln specificity pocket" extends from the active-site heme pocket (Fig. 3a) and is lined by residues Gln, Arg, Trp340/346, Tyr, Pro344/350, Ala345/351, Tyr367/373, Asp376/382, and Arg382/388 (Fig. 2b–d). All residues forming the iNOS Gln specificity pocket are strictly conserved among NOS isozymes, with one exception: iNOS Asp376/382, which hydrogen bonds to Arg382/388, is replaced by Asn in eNOS (Supplementary Fig. 5 online). Interestingly, all previously reported NOSox structures present the "Gln-closed" conformation7,8,11–15 or a disordered Gln conformation10. We thus conclude that the Gln-open conformation and associated cascade of conformational changes leading to the opening of the novel Gln specificity pocket are favored or induced by the binding of quinazoline and aminopyridine inhibitors bearing a rigid and extended tail.</p><!><p>In iNOSox and eNOSox, the binding modes for moderately selective aminopyridine 9 are dramatically different despite common overall protein structures and active sites. The inhibitor aminopyridine cores bind similarly in the active-site heme pocket of both isozymes (Fig. 3a–b and Supplementary Fig. 6 online). In eNOS, the bidentate hydrogen bonds to active-site Glu anchor the aminopyridine core almost parallel to the heme plane and place the inhibitor bulky tail between the eNOSox heme propionates. Invariant Tyr and Arg hydrogen bond to Gln, thus preventing its hydrophobic interaction with the inhibitor tail (Fig. 3b and Supplementary Fig. 6a online). As a consequence, the eNOS complex with compound 9 exhibits the Gln-closed conformation and the Gln specificity pocket is not observed (Fig. 3a). In contrast, the iNOS complex with compound 9 exhibits the Gln-open conformation allowing the inhibitor tail to bind in the Gln specificity pocket (Fig. 3b and Supplementary Fig. 6b online). Not only first-shell Gln and Arg, but also second-shell Asn, present different conformations in the two complexes.</p><p>What prevents the Gln-closed to Gln-open conversion and the opening of the Gln specificity pocket in eNOS? First, we propose that Gln gates the Gln specificity pocket and must adopt the "Gln-open" side-chain conformation to allow inhibitor access. Mutation of Gln into Ala (Q246A, loss of side chain) only marginally enhanced inhibitor potency of compound 9 for eNOS, but did not achieve the potency observed for wild-type iNOS (Table 1). The Q246N mutation had no effect. Thus, removal or rotation of the Gln gate is important, but not sufficient, for potent compound 9 binding to eNOS. Second, we followed, in eNOS, the cascade of conformational changes observed in iNOS upon binding of bulky inhibitors (Fig. 3b). An iNOS-like binding mode for compound 9 in eNOS would induce conformational changes of first-shell Gln and Arg, and second-shell Asn (Fig. 4). However, in eNOS, bulky Leu290 and β-branched rigid Ile271 in the third shell block the side-chain rotation of second-shell Asn. Consequently, this Asn conformation prevents the conformational changes of Gln and Arg necessary for the opening of the Gln specificity pocket (Figs. 3 and 4). The triad of second-shell (Asn) and third-shell (Leu and Ile) residues are the only nearby residues that are not conserved among NOS isozymes (Fig. 2e and Supplementary Fig. 5 online). To test our hypothesis for the key role of third-shell isozyme-specific residues, we made the human iNOSox F286I/V305L double mutant to mimic the corresponding eNOSox residues. Binding affinity of this mutant iNOSox enzyme for compound 9 dramatically dropped to beneath detection levels, as observed for wild-type eNOS (Table 1), while binding of non-selective inhibitor 6 was unaffected (not shown). Furthermore, the x-ray structure of the human iNOSox double mutant co-crystallized with excess compound 9 reveals a Gln-closed conformation and the absence of bound inhibitor (Supplementary Fig. 7 online). Our results thus demonstrate the crucial role of third-shell isozyme-specific residues in inhibitor binding, and provide a structural basis for the exquisite iNOS-specificity of large-tailed quinazoline and aminopyridine inhibitors.</p><!><p>Based on our combined structural and mutagenesis analyses, we propose that differences in the plasticity of second- and third-shell residues between iNOS and eNOS modulate conformational changes of invariant first-shell residues to determine inhibitor selectivity. Together, our mutational and structural results suggest that the Gln specificity pocket accounts for the excellent iNOS-selectivity of the bulky aminopyridine and quinazoline inhibitors. In turn, opening of this pocket depends not only on conformational changes of invariant first-shell Gln and Arg, but also on the plasticity of isozyme-specific second-shell Asn (Fig. 4). This hypothesis is supported by several observations. First, the potent small NOS inhibitors (1, 6–7), which do not induce the Gln-open conformation, show poor selectivity for iNOS27,28 (Figs. 1 and 2 and Supplementary Fig. 2 online). Second, the bulky, but less rigid, tail of compound 8, which neither hydrogen bonds to Tyr nor induces Arg side-chain rotation (Supplementary Fig. 2 online), binds less deeply in the Gln pocket and exhibits only modest selectivity for iNOS32. Third, eNOS third-shell residues Ile271 and Leu290 block binding of compound 9 in the Gln pocket, as evidenced by our structural and mutagenesis results on the human iNOS double mutant (Table 1 and Supplementary Fig. 7 online). Fourth, bulky quinazoline and aminopyridine inhibitors present moderate selectivity against nNOS (7- to 80-fold more selective for iNOS; Fig. 1). We predict that inhibitor binding in nNOS will induce similar side-chain rotations for first-shell Gln and Arg, and partial rotation of second-shell Asn towards third-shell Phe506 and Leu525 (Fig. 4). The substitution of small Val305 in human iNOS with bulkier Leu525 in human nNOS will likely restrict side-chain rotation of second-shell Asn. We thus conclude that the plasticity of the isozyme-specific triad tunes the inhibitor selectivity by controlling the conformational changes of invariant first-shell Gln and Arg and the formation of the new Gln specificity pocket that can be effectively used for inhibitor binding.</p><p>Our results on NOS support an anchored plasticity approach for the design of selective inhibitors. Given a protein of known structure, a set of matching protein sequences (from different species or isoforms), and a binding pocket for a common class of ligand (substrate, cofactor, inhibitor, metabolite, etc…), we propose the following procedure for selective inhibitor design: 1) Identify anchor points for binding in the conserved pocket; 2) Locate variations in sequence and structure outside this pocket; 3) Delineate pathways connecting anchor points to variations (using solvent-accessible channels, for example). 4) Design selective inhibitors that incorporate both a core for anchored binding and extended rigid substituents oriented to exploit protein plasticity along pathways leading to variations (Supplementary Fig. 1 online). The core provides binding affinity via anchoring in nonspecific binding pockets, while the extended substituents determine inhibitor selectivity. Fundamentally, this anchored plasticity approach does not necessarily require serendipitous identification of isoform-specific residue movement. Furthermore, it is readily applicable to key enzyme families, such as kinases, that exhibit overlapping specificities.</p><!><p>Our combined results allowed us to propose the anchored plasticity approach for the design of specific inhibitors exploiting conserved binding sites coupled to distant isozyme-specific residues via cascades of conformational changes. This approach differs from other methods to design novel NOS inhibitors that only exploit differences in first-shell residues15,33,34. To test the applicability of our results, we rationally designed potent and selective iNOS inhibitors starting from a novel and unexploited template with a 5,7-fused heterobicyclic amidine core (compound 13; Supplementary Methods online). This inhibitor is potent but not selective (IC50 = 0.2 μM for iNOS and eNOS and 0.07 μM for nNOS). Addition of small substituents (compounds 14–15) increases the potency but does not significantly affect selectivity (Fig. 1). In contrast, addition of the bulkier and rigid isoquinolinyloxy-methyl tail (compound 16) results in a significant increase in potency and selectivity for iNOS over eNOS35 (Fig. 1). We determined the x-ray structures of murine iNOSox bound to compounds 14 and 16, and of human eNOSox bound to compound 15 (Fig. 5 and Supplementary Fig. 8 online). In all structures, the inhibitor core packs above the heme and makes bidentate hydrogen bonds to invariant active-site residue Glu and to Trp366 main-chain carbonyl. As seen for bulky quinazoline (2–5) and aminopyridine (9–12) compounds, the extended tail of compound 16 packs with first-shell residues Gln, Arg, Pro344, Ala345, and Arg382, and induces the Gln-open conformation (Fig. 5). These results thus demonstrate the applicability of our anchored plasticity approach for the design of novel potent and selective NOS inhibitors.</p><p>The combined inhibitor screening, structural, and mutagenesis results on iNOS and eNOS provide new insights for structure-based design of selective inhibitors. In iNOS, but not in eNOS, binding of inhibitors bearing an extended rigid tail is associated with a cascade of adaptive conformational changes, beginning with movements of invariant first-shell residues and leading to the opening of the novel Gln specificity pocket for enhanced potency and selectivity. The conformational changes reveal a specificity pocket that is separate from an otherwise conserved active-site heme pocket and not observed in previously determined NOSox x-ray structures. Indeed, others had predicted that NOS inhibitors larger than L-Arg would perturb the hydrogen bonding network with Tyr and Gln and extend into the substrate access channel20,21, in distinct contrast to our results. Here, we show that an isozyme-specific triad of second-shell (Asn) and third-shell residues in NOS tunes the plasticity of invariant first-shell residues (Gln and Arg), and thus determines the exquisite selectivity (125- to 3000-fold) of the long-tailed aminopyridine, quinazoline and bicyclic thienooxazepine inhibitors for iNOS over eNOS. The most selective spirocyclic quinazoline compound 3 (ref. 26) and aminopyridine compound 12 (ref. 28 and this study) also show good potency in whole cells and in vivo activity assays (Fig. 1; Supplementary Table 2 and Fig. 9 online). Further studies will be required to test the in vivo potency of the bicyclic thioenooxazepine compounds. Nevertheless, these highly selective NOS inhibitors are promising tools to investigate specific iNOS-mediated effects both in vivo and in vitro. More specifically, these results on iNOS and eNOS inhibitor structures can be applied to future inhibitor design for the treatment of inflammation, cancer, and other diseases, while reducing the risks of disrupting the crucial activity of eNOS in maintaining blood pressure.</p><!><p>The opening of an isozyme-specific pocket in iNOS results from permitted conformational changes of conserved first-shell and second-shell residues upon inhibitor binding. This selective induced-fit movement depends upon plasticity differences in conserved residues located far from the substrate-binding pocket. Exploiting such changes in flexibility to improve inhibitor potency is likely applicable to other key enzymes systems with overwhelmingly conserved active sites, including HIV reverse transcriptase36, aldose reductase37, cyclooxygenases38,39 and kinases40. In NOS, isozyme-specific second-shell and third-shell residues influence the plasticity of invariant first-shell residues, and thus determine the isozyme selectivity. In all these enzyme families, the new binding pocket is distinct from the active site and becomes accessible after adaptive conformational changes of conserved residues. Whether binding of the inhibitor induces the new conformation (induced-fit), or the inhibitor simply selects from different protein conformations in equilibrium (conformational selection) as seen in antigen recognition41, the result is an improved protein-inhibitor interaction.</p><p>Our results for prototypic NOS isozymes appear generally applicable to understanding and tuning binding affinity and specificity of enzyme inhibitors. Our structures demonstrate that differential residue plasticity can be exploited for conformational changes that create new specificity pockets suitable for the design of isozyme-specific inhibitors. Together, these results have exciting implications for drug discovery, and demonstrate that x-ray crystallography is crucial for revealing subtle, but important, differences in residue plasticity between closely related isozymes (e.g. iNOS vs. eNOS) or between homologous enzymes from different organisms. The significant roles of second- and third-shell residues in determining the plasticity of conserved first-shell residues will add to the existing challenges of accurately modeling induced-fit in proteins. Here, we show that systematic structural results combined with mutagenesis can identify selectivity-determining side-chain differences distant from the active site, thus overcoming the barriers that active-site conservation poses for isozyme-specific drug design.</p><!><p>Murine iNOSox Δ65 (residues 66–498) and human eNOSox (65–492, homologous to murine iNOSox Δ78) were expressed and purified as described42,12. Bovine eNOSox (53–492; homologous to murine iNOSox Δ65) was obtained via trypsinolysis of holo-eNOS, which was expressed and purified as published43. Human iNOSox wild-type and mutant constructs (82–508; homologous to murine iNOSox Δ78) were expressed and purified as described10 with slight modifications. Wild-type and mutant human iNOSox proteins were produced in a pT7 Escherichia coli expression vector based on pET-11a vector (Novagen). BL21(DE3) cells (Stratagene) were grown in the presence of ampicillin at 37 °C until reaching a cell density corresponding to A600 of 0.5–0.8. The culture was then induced with 0.5 mM IPTG, 6.125 mg l−1 ferric citrate and 450 μM ∂-amino levulinic acid, and grown for three days at 20 °C before harvesting. Pellets were re-suspended in buffer A (10 mM sodium phosphate pH 7.0, 0.1 M NaCl, 1 mM L-Arg, 2 mM DTT, 10 μM H4B), sonicated extensively and subsequently loaded on a heparin column (GE Healthcare). Protein was eluted in a single step by adding 0.3 M NaCl to buffer A. Fractions containing NOSox were concentrated, aliquoted and stored at −80 °C.</p><!><p>Mutations were introduced into the cDNAs of human iNOSox and full-length NOS within expression vectors by using the QuickChange™ Site-Directed Mutagenesis Kit (Stratagene). Mutagenic oligonucleotides were designed according to the manufacturer's instructions and mutations confirmed by sequencing.</p><!><p>Compounds 1–16 were prepared as described35,26,28. Details of the synthesis for compounds 13–16 are described in Supplementary Methods online.</p><!><p>The inhibitory potency (IC50) was determined in full-length wild-type and mutant NOS in the presence of cofactors (5 μM FAD, 5 μM FMN, 200 μM BH4, 1 mM CaCl2, 25 ng ml−1 CaM) and substrates (3 μM [3H]-arginine, 1 mM NADPH) as described26. Inhibitors were pre-incubated with NOS proteins in the presence of cofactors and NADPH for one hour before L-Arg addition.</p><p>Binding affinity (KD) of human iNOSox for compounds 6 and 9 was measured by imidazole displacement. Briefly, iNOSox was incubated for 2 hours with a matrix of concentrations of imidazole and compound 6 or 9. Samples were scanned by UV-Visible spectroscopy between 270 and 700 nm. Binding affinity was determined by plotting the absorbance difference A428–A396 as a function of the imidazole concentration. The decrease in apparent imidazole affinity as the concentration of compound increased was used to determine the binding affinity for compound 6 or 9.</p><p>Compound 12 was further tested for inhibition of NO production in an intact cell assay44 and in a rat model of inflammation via lipopolysaccharide (LPS)-induced nitric oxide production26, as described here. Compound 12 or vehicle was given orally to male conscious rats at time zero. Blood samples were taken after 2, 4 and 6 h. Total plasma nitrite and nitrate concentrations, indicative of NO production, were determined by using the Griess reaction after reduction of nitrate to nitrite by nitrate reductase. Oral administration of compound 12 to rats led to dose-dependent inhibition of elevated plasma NO levels (IC50= 1.8 μM) measured 4 h after LPS administration (Supplementary Fig. 9 online). All in vivo studies were approved by the AstraZeneca Ethical Review Committee and were conducted under licence from the UK Home Office.</p><!><p>Murine and human iNOSox, and bovine and human eNOSox were co-crystallized (with 2–5 mM inhibitor) by vapor diffusion as described7,8,10,12. Crystals grew overnight (crystallization pH was 7, 7.2, 6.5, and 6.0 for murine iNOSox, human iNOSox, bovine eNOSox, and human eNOSox, respectively). All crystals were cryo-cooled after transfer to cryoprotectant solution (murine iNOSox: 30 % glycerol, human iNOSox: 100 % MgSO4, bovine eNOSox: 15 % glycerol, human eNOSox: 15 % 2-methyl-2,4-pentanediol) in a cold nitrogen gas stream. Data were collected at 100 K at CHESS beamline F1 (compound 1), SSRL beamlines 7–1 (compounds 2–4, 8–10), 9–1 (compounds 5, 7), and 9–2 (compound 6), MAX-LAB beamline I711 (compounds 14, 16), ESRF beamlines FIP (compound 15) and ID2 (compound 12). All diffraction data sets were processed using DENZO/SCALEPACK programs45. The crystal structures of human iNOSox10 and murine iNOSox7 and human eNOSox10, with ligands, water and cofactors removed, were used as starting models for molecular replacement with AMoRe46 for human iNOSox, murine iNOSox, bovine eNOSox and human eNOSox, respectively. During crystallographic refinement of protein structures, the heme geometry, like the amino acid geometry, is restrained by a set of parameters (bond lengths, angles, dihedrals) derived from small molecules studies and high-resolution protein structures47. These parameters and their associated weights can strongly influence the resulting refined geometries for heme in protein structures48, especially those determined at lower resolution. Hence, direct comparisons of heme distortions from published x-ray structures refined in different ways can be problematic. Parameters from the HICUP database47 were used for refinement of the heme and pterin cofactors (HEC and H4B, respectively). Inhibitors were fit into Sigmaa-weighted49 Fo-Fc electron density maps, which confirmed the expected unmodified chemical structures. Overall structures were obtained by iterative cycles of refinement with CNS50 and manual fitting with O51, Xfit52, or Coot53 (Supplementary Tables 3–5 online). All superimpositions were performed for residues within a sphere of 10 Å around the inhibitor with the CCP4 program LSQKAB54. The r.m.s. deviations were calculated for all superimposed atoms.</p><!><p>Supplementary Information is linked to the online version of the paper at www.nature.com/nature.</p>

PubMed Author Manuscript

Detection of pesticide residues and risk assessment from the local fruits and vegetables in Incheon, Korea

This study was conducted to investigate the pesticide residue concentrations and assess potential human health risks from fruit and vegetable consumption in Incheon. A total of 1,146 samples of 20 different types of fruits and vegetables were collected from the Incheon area in 2020. The pesticide residues were analyzed by the multi-residue method of the Korean Food Code for 400 different pesticides. Among the fruit and vegetable samples, 1,055 samples (92.1%) were free from detectable residues, while 91 samples (7.9%) contained residues and 11 samples (1.0%) had residues exceeding the Korean maximum residue limit. A total of 32 different pesticide residues were found and 8 residues exceeded MRLs. The most frequently detected pesticide residues were chlorfenapyr, procymidone, etofenprox, pendimethalin, fluopyram and azoxystrobin. The highest values of short term and long term exposure were obtained in the case of consumption of lettuce(leaves) with chlorfenpyr. For chronic dietary exposure, the cumulative hazard index (cHI) were below 100%. The results of this study showed that the detected pesticides were not exposed to potential health risks through the consumption of fruits and vegetables.Pesticides are essential tools to increase agricultural productivity and cultivation convenience by protecting crops from pests pathogen and weeds. However, pesticides inevitably remain in agricultural products and soil 1,2 . Excessive use of pesticides causes these chemicals and metabolites to remain in the environment and food, causing serious problems in the ecosystem and public health [3][4][5] . Chronic human exposure to unsafe levels of pesticides can cause a wide range of diseases affecting human health. Pesticides have potential adverse effects on human health such as carcinogenesis, immunotoxicity, birth defects, genetic changes, neurological toxicity and endocrine disruption [6][7][8] . Fruits and vegetables are usually consumed directly without processing after washing, so they are the main cause of pesticide residue ingestion in humans. Human intakes of hazardous substances from pesticide residues in agricultural products can be significantly higher than intakes of these substances associated with water consumption and air intake 9,10 . Therefore, it is very important to monitor residual pesticides in fruits and vegetables and to assess if they cause a risk to human health.Fruits and vegetables are essential for human nutrition containing functional compounds such as carotenoids, phenolics, trace minerals, vitamins and fiber 11 . However, they may contain toxic residual pesticides due to the use of pesticides during the production process of agricultural products 12 . Pesticide residues in agricultural products are usually monitored with reference to maximum residue limits (MRLs), which represent the highest concentration of pesticide residues that is legally permitted or accepted in food commodities after the use of pesticides 13,14 . In Korea, the Ministry of Food and Drug Safety and the Rural Development Administration are responsible for managing pesticides. To ensure the safety of agricultural products, Korea has implemented the PLS (Positive List System) since 2019. PLS register pesticides used in domestic or imported foods and establishes pesticide MRLs in agricultural products. It is a system that pesticide MRLs in agricultural products is applied at 0.01 mg/kg uniformly except for registered pesticides 15 . Government and related organizations are trying to ensure safe use of pesticides, but pesticide residues are continuously detected in agricultural products. Some agricultural products occur in excess of MRLs and exposure to pesticide residues is likely to harm humans. The

detection_of_pesticide_residues_and_risk_assessment_from_the_local_fruits_and_vegetables_in_incheon,

Materials and methods<!>Samples.<!>Sample extraction and clean up.<!>GC-MS/MS analysis.<!>GC-ECD/NPD analysis.<!>Risk assessment estimation.<!>Approvals and permissions.<!>Results and discussion

<p>Chemical and reagents. In this study, tests were conducted on 400 pesticides that can be analyzed by the multi-residue method of the Korean Food Code. The 400 pesticide standards used to analyze pesticide residues were purchased by Accustandard (New Haven, CT, USA). For the extraction of pesticide residues, acetonitrile, acetone, dichloromethane and n-hexane used in this experiment were purchase with HPLC grade reagents (Muskegon, MI, USA) and anhydrous sodium chloride was purchased from Junsei (Tokyo, Japan). Solid phase extraction (SPE) for sample purification was purchased from Bekolut (Haupststuhl, Rhineland-Palatinate, Germany).</p><!><p>Sampling was performed by an authorized person who met the Korean Food Code Guideline.</p><p>Samples used in this experiment were collected from 1146 fruits and vegetables from markets in Incheon, Republic of Korea in 2020. The samples were collected from fresh agricultural products and quickly brought to the laboratory for analysis. All samples were analyzed within 24 h and kept in the refrigerator until extraction.</p><!><p>Sample extraction and clean up were performed according to the multi-residue method for pesticide residues according to the Korean Food Code. Part of the collected samples (1 kg) were taken and thoroughly crushed with a food grinder. After grinding, 100 mL acetonitrile was added to the ground sample and homogenized for 3 min with a high-speed homogenizer. The homogenized mixture was filtered into a bottle with 15 g of anhydrous sodium chloride and mix filtrate vigorously for 3 min to separate layers. From the upper layer, an aliquot of 20 mL was transferred into a tube and evaporated to dryness on a 40 °C bath with a gentle stream of air. For GC-MS/MS and GC-ECD/NPD analysis, the dried extracts were dissolved with 4 mL of acetone/n-hexane (20:80, v/v) and transferred to a Florisil cartridge (1 g, 6 mL), which was activated and pre-conditioned with 5 mL of acetone/n-hexane (20:80, v/v) and 5 mL of n-hexane. After sample loading, the cartridge was eluted with 5 mL of acetone/n-hexane (20:80, v/v). This solvent was then evaporated slowly to dryness at 40 °C bath under a gentle stream of air. The dried residue was re-dissolved with 2 mL of acetone/nhexane (20:80, v/v) and filtered with 0.2 µm PTFE filter (Advantec, Otowa, Tokyo, Japan) for GC analysis. For LC-MS/MS and LC-UVD analysis, sample extracts were dissolved with 4 mL of dichloromethane/methanol (99:1, v/v) and transferred to amino-propyl cartridge (1 g, 6 mL) which was activated and pre-conditioned with 5 mL of dichloromethane. After sample loading, the cartridge was eluted with 7 mL of dichloromethane/ methanol (99:1, v/v). This solvent was evaporated slowly to dryness at 40 °C bath under a gentle stream of air. The dried residue was re-dissolved with 2 mL of acetonitrile and filtered with 0.2 µm PTFE filter for LC analysis.</p><!><p>GC-MS/MS analysis was performed using 7890B gas chromatograph coupled to a triple quadrupole mass spectrometer 7000D with electron impact ionization (EI) equipped with a 7693 autosampler (Agilent Technologies, Santa Clara, CA, USA). Chromatographic separation of pesticides was conducted on DB-5 MS capillary columns (30 m × 0.25 mm × 0.25 µm film thickness, Agilent Technologies, Santa Clara, CA, USA). The oven temperature was programmed from 70 °C (hold 3 min) to 180 °C by a rate of 20 °C/min and finally increased to 300 °C (hold 2.5 min) by a rate of 5 °C/min. The temperature of injector was held at 250 °C and the injection volume was 1 µL with splitless mode. Helium carrier gas (99.999%) flowed constantly at 1 mL/ min. The mass spectrometry detector (MSD) used electron impact ionization mode (ionization energy 70 eV). The temperature of ion source and quadrupole were set at 250 °C and 150 °C, respectively. The multiple reaction monitoring (MRM) mode with minimum two ions for each pesticide was used for detection and quantification of pesticides.</p><!><p>The GC-ECD and GC-NPD system was used to analyze organochlorine and pyrethroid compounds, organophosphorus and nitrogen-containing compounds. An Agilent 6890 series GC equipped with 63 Ni electron capture detector and a nitrogen phosphorous detector were employed. Chromatographic separations were conducted on DB-5 capillary columns (30 m × 0.25 mm × 0.25 µm film thickness, Agilent Technologies, Santa Clara, CA, USA) for GC-ECD and GC-NPD. The operating conditions for GC-ECD were as follows: The oven temperature was programmed from 150 °C (hold 1 min) to 240 °C (hold 2 min) by a rate of 12 °C/min and finally increased to 280 °C (hold 13.5 min) by a rate of 10 °C/min. The injection volume was 1µL with split mode (42.2:1) and nitrogen carrier gas flowed at 1.2 mL/min. The temperature of injector and detector were at 250 °C and 280 °C, respectively. The operating conditions for GC-NPD were as follows :</p><p>The oven temperature was programmed from 120 °C (hold 1 min) to 240 °C (hold 2 min) by a rate of 12 °C/ min, increased to 280 °C (hold 10 min) by a rate of 10 °C/min and finally increased to 300 °C (hold 1 min) by a rate of 10 °C/min. The injection volume was 1µL with splitless mode. Nitrogen carrier, hydrogen and air flowed at 1.2 mL/min, at 3.0 mL/min, at 120.0 mL/min, respectively. The temperature of injector and detector were at 270 °C and 300 °C, respectively. LC-MS/MS analysis. LC-MS/MS analysis was performed using Vanquish UHPLC system coupled to a TSQ Altis triple quadrupole mass detector system (Thermo-Fisher Scientific, Waltham, Massachusetts, USA). Chromatographic separation of pesticides was performed using Accucore aQ (2.1 mm × 100 mm, 2.6 µm particle size, Thermo-Fisher Scientific, Waltham, Massachusetts, USA). Mobile phase A (0.01% formic acid and 5 mM ammonium formate in water) and mobile phase B (0.01% formic acid and 5 mM ammonium formate in methanol) were used for the gradient program. The gradient program was as follows: 0-0.5 min (80% A/20% B), Method validation. The analytical methods was validated in terms of limit of detection (LOD), limit of quantification (LOQ), recovery and precision according to the Korea Food Code pesticides guidelines 16 . Assessment of recovery was performed using a mixture of the targeted pesticides at fortification levels of 0.1, 1.0 mg/kg using pesticide free sample extracts. The LOD and LOQ were estimated from the standard deviation of the five replicated analyses of spiked sample at low concentration level (LOD = 3.3 × SD and LOQ = 10 × SD). Precision was expressed the relative standard deviation (RSD, %) and was evaluated by analyzing replicate samples. To assess linearity, the extracts from pesticide free samples were fortified with standard solutions of 0.05, 0.1, 0.25, 1.0, 2.0 mg/kg and analyzed in triplicate at each concentration.</p><!><p>The risk assessment of pesticide detected in fruits and vegetables was estimated based on the results of the survey on residual pesticides. The short term risk assessment (aHQ) was calculated based on the estimated short term intake (ESTI) and the acute reference dose (ARfD). ESTI was calculated by multiplying the highest residue level and food consumption and dividing this by the body weight. The aHQ was calculated using the formula: aHQ = ESTI/ARfD × 100%. The long term risk assessment (HQ) was performed using the estimated daily intake (EDI) and the established acceptable daily intake (ADI). EDI was calculated by multiplying the average pesticide concentration and the food consumption rate and dividing this by the body weight 17 . The HQ was calculated using the formula: HQ = EDI/ADI × 100%. The average daily intake was referred to the intake of fruits and vegetables examined by Korea Disease Control and Prevention Agency 18 . The criteria for ADI and ARfD refer to pesticides and veterinary drugs information from Ministry of Food and Drug Safety 19 . A value below 100% indicated that the exposed people were unlikely to experience obvious adverse effects. An index above 100% indicated the possibility that the exposure would induce obvious adverse effects 20 . HQ was calculated for the pesticides and agricultural products. The results were summed up to obtain a chronic hazard index (cHI). The chronic hazard index was calculated by the sum of HQs (cHI = ΣHQ). A cHI (%) > 100 indicated that the fruits and vegetables should be considered a risk to the consumers, whereas an index below 100 indicated that the consumption of the fruits and vegetables was considered acceptable 17 .</p><!><p>This study was approved by Incheon Metropolitan Government for permission to collect agricultural products/plants specimens.</p><!><p>Method validation. From 400 pesticides, 15 pesticides were selected considering the detection rate and the violation rate of the MRLs. Table 1 presents linear correlation coefficients, limits of detection (LODs), limits of quantification (LOQs) and recoveries for the validation study. A linear correlation coefficient between pesticide concentrations and peak areas was detected in the range of 0.9947-0.9999. The LODs and LOQs values for the studied pesticides ranged from 0.004 to 0.040 mg/kg and from 0.011 to 0.120 mg/kg, respectively. The recovery was 85.3-98.3% for all pesticides, which is within the acceptable recovery range of 70-120% and the RSD of less than 10% also met requirement 21 . These results indicate that the analytical method applied to this study is appropriate for the analysis of targeted pesticide residues in fruits and vegetables.</p><p>Pesticide residues in fruits and vegetables. In this survey, 1146 samples of fruits and vegetables were analyzed for 400 pesticides contamination to assess health risk. In 1055 of 1146 analyzed fruit and vegetable samples (92.1%), no detectable residues were found, while pesticide residues were detected in 91 samples (8.9%). A number of 11 samples (1.0%) contained residues above MRLs established by the MFDS in Republic of Korea. Perilla leaves (13 samples, 11.4%), welsh onions (11 samples, 12.3%), chili peppers (7 samples, 16.7%), lettuce leaves (7 samples, 4.1%), aster scabers (6 samples, 19.6%), Chinese chives (6 samples, 12.2%) and winter-grown cabbages (6 samples, 3.8%) had a number of contaminated samples. Aster scabers (3 samples, 6.5%), pimpinella brachycarpas (3 samples, 8.3%), crown daisies (2 samples 2.6%), welsh onions (2 samples, 1.9%) and schisandraberries (1 sample, 16.7%) violated MRLs (Table 2). Szpyrka et al. 10 reported that pesticide residues were detected in 36.6% of the analyzed in fruits and vegetables in Poland. In Republic of Korea, detectable residues were found in 13.9% of 34,520 samples of vegetables collected from 2010 to 2014 22 . In Algeria, pesticide residues monitoring for fruits and vegetables revealed residual pesticides in 57.5% of the analyzed samples 12 . Chen et al. 23 found residues of selected fungicide and insecticides in vegetables including lettuce and spinach from Xiamen, Vol:.( 1234567890 Incidences and MRL violation of pesticide residues. Table 3 shows the frequency and concentration of pesticide residues in the analyzed samples. Of the 400 pesticides tested, 32 pesticides were detected in the analyzed samples. The numbers of pesticides detected by function was 15 fungicides (46.9%), 14 insecticides (43.7%), 2 herbicides (6.3%) and 1 growth regulator (3.1%). The most frequently detected pesticides in fruits and vegetables were chlorfenapyr (13 samples, 1.1%), procymidone (9 samples, 0.8%), etofenprox (8 samples, 0.7%), pendimethalin (7 samples, 0.6%) and fluopyram (6 samples, 0.5%). Theses pesticides accounted for approximately 47.3% of all pesticides detected in this study. Figure 1 shows the number of pesticide residues detected in fruits and vegetables and the number of excess MRLs. It was detected mainly in aster scabers, chili peppers, lettuce leaves, perilla leaves and welsh onions. The maximum residue concentration of fludioxonil was 7.48 mg/ kg, which was the highest value among the detected pesticides. Chlorfenapyr was found in a concentration range ) samples exceeded the MRLs. Fenobucarb, flubendiamide, fluquinconazole, hexaconazole, methidathion, prochloraz, procymidone and tebuconazole had residues that violated MRLs. Aster scabers, crown daisies, pimpinella brachycarpas, welsh onions and schisandraberries were the fruits and vegetables with residues above MRLs. Schisandraberries, pimpinella brachycarpas and aster scabers had high violations rate of 16.7%, 8.3% and 6.5%, respectively. The most frequently violated pesticides were flubendiamide, fluquinconazole and procymidone. The result of pesticide residues that violated the MRLs generally showed similar results to other studies in the Republic of Korea. In a report 24 by the Gyeonggido Institute of Health and Environment, diazinon, carbofuran, fluquinconazole and procymidone were shown to have a high number of residues exceeding MRLs. And, Yi et al. 22 reported that one of the pesticides frequently detected in excess of MRLs was diazinon, paclobutrazol and procymidone.</p><p>The percentage of samples with pesticide residues exceeding MRLs in this study (1.0%) was lower than the majority of those reported in other studies. Szpyrka et al. 10 reported 1.8% exceedance of MRLs in fruits and vegetables collected in south-eastern Poland. According to a report by the United States Food and Drug Administration, pesticide residues exceeding the MRLs were detected in 2% of the domestic vegetables and 7% of the imported vegetable samples. Park et al. 25 found that 1.4% of vegetables exceeded MRLs in Republic of Korea. A study from Algeria found that 12.5% of fruit and vegetable samples contained pesticide residues that exceeded the MRLs 12 . In addition, incidences of pesticide residues above MRL were reported to be 1.4% in vegetables collected from markets in Seoul, Republic of Korea 22 .</p><p>Risk assessment. Pesticide residues in fruits and vegetables are unlikely to be completely removed by washing. Therefore, it is a very dangerous situation when consumers eat fruits and vegetables contaminated with high concentrations of pesticide residues over a long period time. The risk from pesticide residues in fruits and vegetables was performed on dietary exposure assessment for all detected pesticide in the samples. The results of human exposure to pesticides based on fruit and vegetable intake are shown in Table 4. In the short term risk assessment, the ESTIs of pesticide range from 1.2 × 10 -7 to 1.4 × 10 -2 mg/kg bw/day. The range of aHQ was 0.000-8.411%. The highest values of aHQ were obtained in case of consumption of lettuce (leaves) with chlorfenapyr. In a similar study, Mebdoua et al. 12 identified a potential acute exposure for pesticide residues in fruits and vegetables for the population of Algerian. The values of short term exposure ranged from 0.78% to 558.5% of aHQ for children and ranged from 0.23% to 237.8% of aHQ for adults. In the long term risk assessment, the EDIs of pesticide range from 3.7 × 10 -9 to 7.0 × 10 -4 mg/kg bw/day. The range of HQs was 0.000-6.384%. The HQs for fluquinconazole and prochloraz were 0.080-6.384 and 0.131-3.336%, respectively and were higher than those of other pesticides. The HQ value above 100% indicates a potential risk to consumers 20,26 . Therefore, the results indicate that the detected pesticides in this study are not harmful to human health. The cHI for all residues was 17.714%, which was less than 100%, meaning that there is no risk of side effects following a cumulative exposure to all the detected pesticides. Elgueta et al. 27 reported chronic health risk for consumers in leafy vegetables. They found HQs values for methamidophos (73.9%), cypermethrin (30.4%), mancozeb (11.5%) and cyfluthrin (4.5%), etc. The HQs were summed up and the cHI of all residues was 135%, more than 100%. Chen et al. 23 detected pesticide residues in fruits and vegetables, but there was no health risk for consumers. They found HQ values for omethoate (2.6%), methamidophos (2.2%) and chlorpyrifos (0.24%). Since pesticide residues are reduced by 8-68% with flowing water at home depending on the type of pesticide residues and characteristics of fruits and vegetables, the risk of pesticide residues can be lowered 28 . Additionally, agricultural products in excess of MRLs are seized by the government and unsuitable agricultural products are immediately discarded in Republic of Korea. However, continuous management and monitoring of agricultural products is required for the safety of consumers because intake of agricultural products changes according to consumer's preference, region and season.</p><p>In conclusion, pesticide residues were found in 8.9% of the samples and exceeded the MRLs in 1.0% of the total fruits and vegetables. The frequently detected pesticides were chlorfenapyr, procymidone, etofenprox, pendimethalin and fluopyram, while the high rate of violations were flubendiamide, fluquinconazole and procymidone. Based on the findings, the range of aHQs and HQs was 0.000-8.411 and 0.000-6.384%, respectively. Therefore, the results showed that the consumers were not exposed to health risks through the consumption of fruits and vegetables. The results provide important information about the current state of pollution in fruits and vegetables. The obtained data can be used to develop strategies and improve pesticide MRLs for the safe management of fruits and vegetables in Republic of Korea.</p>

Scientific Reports - Nature

Utilizing ion mobility spectrometry-mass spectrometry for the characterization and detection of persistent organic pollutants and their metabolites

Persistent organic pollutants (POPs) are xenobiotic chemicals of global concern due to their long-range transport capabilities, persistence, ability to bioaccumulate, and potential to have negative effects on human health and the environment. Identifying POPs in both the environment and human body is therefore essential for assessing potential health risks, but their diverse range of chemical classes challenge analytical techniques. Currently, platforms coupling chromatography approaches with mass spectrometry (MS) are the most common analytical methods employed to evaluate both parent POPs and their respective metabolites and/or degradants in samples ranging from d rinking water to biofluids. Unfortunately, different types of analyses are commonly needed to assess both the parent and metabolite/degradant POPs from the various chemical classes. The multiple time-consuming analyses necessary thus present a number of technical and logistical challenges when rapid evaluations are needed and sample volumes are limited. To address these challenges, we characterized 64 compounds including parent per- and polyfluoroalkyl substances (PFAS), pesticides, polychlorinated biphenyls (PCBs), industrial chemicals, and pharmaceuticals and personal care products (PPCPs), in addition to their metabolites and/or degradants, using ion mobility spectrometry coupled with MS (IMS-MS) as a potential rapid screening technique. Different ionization sources including electrospray ionization (ESI) and atmospheric pressure photoionization (APPI) were employed to determine optimal ionization for each chemical. Collectively, this study advances the field of exposure assessment by structurally characterizing the 64 important environmental pollutants, assessing their best ionization sources, and evaluating their rapid screening potential with IMS-MS.

utilizing_ion_mobility_spectrometry-mass_spectrometry_for_the_characterization_and_detection_of_pers

Introduction<!>Sample preparation<!>IMS-MS analyses<!>LC-IMS-MS instrumental analysis<!>Data analysis<!>Results and discussion<!>Investigation of a preferred ionization method for POP parent and metabolites/degradants<!>PCB parents and metabolites<!>PPCPs, industrial chemicals, and their metabolites<!>Analysis of pesticides and their metabolites or degradation products<!>PFAS and their degradation products in AFFFs<!>Conclusion

<p>Xenobiotics are of great concern to humans and the environment due to their wide use in agriculture, industrial processes, and ubiquitous presence in consumer products. Many xenobiotics are considered persistent organic pollutants (POPs) and have been shown to pose carcinogenic, mutagenic, or reproductive/developmental hazards (1). The Stockholm Convention on POPs is an international treaty designed to protect human health and the environment. Through this treaty, a number of chemicals were recognized as of great concern, including per- and polyfluoroalkyl substances (PFAS), pesticides, polychlorinated biphenyls (PCBs), industrial chemicals, and industrial chemical byproducts (2). In 2001, the Stockholm Convention listed twelve substances which have harmful impacts on human health or the environment (Table 1). In 2017, an additional sixteen POPs were added to the list, and currently, three more chemicals are under review. To date, the original twelve listed substances have either been phased out completely or their use restricted; however, many are persistent and are still detected in soil, water, and blood (3). There are also a large number of POPs not yet identified including emerging chemicals, and pharmaceuticals and personal care products (PPCPs)(4). These unknown POPs and other POPs formed by metabolism and degradation further complicate exposure studies. Metabolism of xenobiotics through various enzymes including cytochrome P450 oxidases, UDP-glucuronosyltransferases (SULT), and glutathione S-transferases(GULT) is extremely common (5). POPs are also commonly degraded by bacteria, or through photochemical and other processes, forming intermediates and metabolites that are often more toxic and stable than the parent compounds (6). Thus, the ability to identify not only the parent POPs but also their metabolites and degradants is essential to fully characterizing exposure and understanding potential adverse effects on human health and the environment (7, 8).</p><p>Due to the chemical diversity of parent POPs and their metabolites and degradants, to date, multiple time-consuming sample extractions and analytical methods are commonly employed (9). For example, traditional methods use a solid-phase or liquid-liquid extraction to isolate and concentrate the molecules of interest. Then, a number of analytical approaches often coupling gas, liquid, or supercritical fluid chromatography (GC, LC, or SCF) with mass spectrometry (MS) are used to accurately identify and quantify the molecules in the environmental and human samples (10). While current methods for sample preparation (e.g., extraction and derivatization) and analytical measurements of POPs and their metabolites/degradants provide high selectivity and throughput, the increasing number of POPs and the structural differences in their metabolites and degradation products make it difficult to rapidly screen for the broad range of chemicals in cases of complex or unknown exposures. Additional separation and coupled separation techniques are thus of great interest for the exposure assessment studies. For example, ion mobility spectrometry coupled with MS (IMS-MS) has shown great promise in screening diverse POPs including PFAS and PCBs and their subsequent products without extensive sample preparation (10, 11); however, extensive analyses of their metabolites and degradant products has not been performed. IMS is a rapid separation technique, occurring on a millisecond time frame and allowing gas-phase structural analyses (12). By coupling IMS with MS, both structure and m/z information can be obtained for the molecules of interest, and high sensitivity small molecule measurement have been reported for IMS-MS analyses in a variety of matrices including a limit of detection of 100 pg/mL in serum (11–14). Additionally, since IMS-MS occurs post-ionization, various ionization sources can be used and it can be implemented after GC or LC separations when multidimensional characterization is desired.</p><p>In this study, we utilized IMS-MS to characterize 27 chemicals listed by the Stockholm Convention and 37 POPs from a range of other chemical classes of concern, including PAHs, PCBs, industrial products and byproducts, PPCPs, PFAS, pesticides and their corresponding metabolites and degradation products. These chemicals were selected as many are included in the Agency for Toxic Substances and Disease Registry (ATSDR) or International Agency for Research on Cancer (IARC) monographs. Additionally, the selected PPCPs as well as their sulfate and glucuronide metabolites are frequently detected in wastewater.. In the IMS-MS evaluations, all chemicals were assessed for their preferred ionization method and isomer separations to determine if simultaneous rapid screening of both parents and metabolites was possible in the same analysis or if multiple evaluations must be performed (15). This study therefore used electrospray ionization (ESI) and atmospheric pressure photoionization (APPI) sources to assess both the parent and metabolites/degradants of the specific POPs noted in Table 1. Since these two sources provide complementary chemical assessments (e.g., ESI works well for polar molecules while APPI is commonly utilized for nonpolar molecules), evaluating the different chemicals with both is important. The results for simultaneous evaluations of the parent and metabolites/degradants in solvents and complex mixtures such as aqueous film-forming foams (AFFFs) therefore showcase capabilities and challenges for performing rapid assessments of each molecule type with IMS-MS.</p><!><p>Chemical standards for all molecules except the PCBs were obtained from US EPA (Dr. Ann Richard) or were purchased from Sigma-Aldrich (St. Louis, MO), Santa Cruz Biotechnology (Dallas, TX), and Toronto Research Chemicals (Ontario, CA) (Supplemental Table 1). All purchased chemical standards were > 97% pure according to the manufacturers. The PCB standards were synthesized as detailed in Grimm et al. (16). High-purity solvents (≥ 99.9%) including water, methanol, acetone, acetic acid, and toluene were purchased from Sigma-Aldrich. PCBs and PAHs were dissolved in toluene and diluted to a final concentration of 5 μM in 50:50 methanol:acetone. The remaining standards were diluted in 80:20:0.1 methanol:water:acetic acid to a final concentration of 1 to 10 μM (Supplemental Table 1). AFFF samples were acquired from Chemguard (Marinett, WI), FireStopper International (Varhaug, Norway), and Angus Fire (Angier, NC) and diluted 100-fold in deionized water.</p><!><p>An Agilent 6560 IMS-QTOF MS (Agilent Technologies, Santa Clara) was utilized for all nitrogen gas drift tube IMS (DTIMS) measurements in this work, and all individual standards were directly injected in triplicate into the APPI and ESI sources and evaluated in both positive and negative ionization modes. Furthermore, blanks were injected between each standard to make sure no carryover occurred during the runs. PFAS standards and AFFF solutions were run in triplicate with only ESI in negative ionization mode due to their known preference for this analysis type (17). For the IMS analyses, the ions were passed through the inlet glass capillary, focused by a high-pressure ion funnel, and accumulated in an ion funnel trap (18). Ions were then pulsed into the 78.24-cm-long IMS drift tube filled with ~ 3.95 Torr of nitrogen gas, where they traveled under the influence of a weak electric field (10–20 V/cm). Ions exiting the drift tube were refocused by a rear ion funnel prior to quadrupole time-of-flight(QTOF) MS detection, and their drift time was recorded. For each detected feature, collision cross-section(CCS) values were calculated using a single electric field voltage (19). Drift times and CCS values for each tested substance are listed in Supplemental Table 1. The detailed instrumental settings follow those previously published in an interlaboratory examination and drift tube IMS CCS analyses and can be found in Supplemental Document 1 (19). Prior to each experimental analysis, the instrument was tuned and a mass calibration was performed using Agilent Tune Mix (G2421A/G2432A, Agilent).</p><!><p>For the AFFF analyses (20), 20 μL of each sample was injected on a ZORBAX SB C-18 column (2.1 × 50 mm, 1.8 μM; Agilent) using a 1260 Infinity II system (Agilent). LC conditions were as detailed by (17) with mobile phase A consisting of 5 mM ammonium acetate in 95% water and mobile phase B made up of 5 mM ammonium acetate in 95% methanol. The initial chromatographic condition was maintained at 90% A and 10% B for 0.5 min. The gradient was ramped such that there was 30% B by 2 min, 95% B by 14 min, and 100% B by 14.5 min. The 100% B condition was held for 2 min, resulting in a total run time of 16.5 min. Following each run, the amount of B was returned to 10% for 6 min to equilibrate the column prior to the next injection. A flow rate of 0.4 mL/min was used through the entire gradient. Blank samples were performed before and after each AFFF analysis run to ensure that no carryover existed between samples. Blank subtraction was also utilized to make sure contaminates were not included in the evaluations. IMS-MS analyses were performed using ESI in negative mode.</p><!><p>The Agilent IM-MS Browser software was utilized for all single-field CCS calculations. Agilent Mass Profiler software was utilized to assess the drift times for the observed ions and calculate the CCS values. Relative standard deviations (RSDs) of < 1% were observed for all triplicate CCS measurements.</p><!><p>In this study, 64 chemicals were studied with 18 parent POPs from various chemical classes including industrial chemicals, PAH, PCB, pesticides, PFAS and pharmaceuticals, and 46 of their metabolites or degradation products (Table 1). Many of the assessed chemicals are recognized as chemicals of concern by the Stockholm Convention. For example, chemicals categorized in Table 1 as Annex A are those where production must be eliminated, Annex B must be restricted, and Annex C currently have measures being taken to reduce their unintentional release (21). While some chemicals analyzed in this study have not been listed by the Stockholm Convention, they are also considered POPs and may have adverse human health effects. For example, chemicals in the treaty such as PCBs, pesticides, and PFAS were of great interest in our study due to their known toxic effects. Pesticides that were analyzed in our study included aldrin, dieldrin, heptachlor, hexachlorobenzene, and endosulfan. These organochlorine pesticides have classic POP qualities, and exposure is associated with adverse health effects such as endocrine disruption or carcinogenicity. Studying PCBs is also important as their persistence in the environment corresponds to their degree of chlorination with half-lives varying from 10 days to one and a half years with potential endocrine disruptors and have genotoxic properties. In our study, PCB 3 and its isomeric metabolites were evaluated as the structure of these chemicals can have varying effects on their metabolism and elimination. Finally, PFAS were studied as this category of chemicals has gained worldwide attention due to their potential hazardous human health effects and ecological environments (22). In 2009, perfluorooctanoic acid (PFOA) and perfluorooctane sulfonate (PFOS) were added to the Stockholm Convention and their use was banned or severely restricted (2). PFAS as a class have a variety of applications including chemical industry, consumer products, and production of AFFFs because of their unique properties such as thermal stability, hydrophobicity, and surface activity (23). Because these characteristics allow manufactured goods to have beneficial properties such as stain and water resistance, > 5000 PFAS are thought to have been produced (24, 25). Because of their chemical properties, PFAS have long-distance transport potential, bioaccumulate, and have toxic effects such as immunotoxicity, genotoxicity, reproductive toxicity, neurotoxicity, and carcinogenicity. This makes PFAS in general of great concern and especially since some PFAS can be metabolized or degraded. Therefore, exposure to PFAS can be from direct sources such as drinking water or air inhalation, or indirect sources including the uptake of transformation products (26).</p><!><p>Analysis of the 64 chemicals with IMS-MS was performed using both an ESI and APPI source. The ESI studies were performed in both positive (ESI[+]) and negative (ESI[−]) ion modes to study the protonated and deprotonated ions, while APPI was only performed in positive mode (APPI[+]) to assess protonation and positive radical formation. While the APPI negative ion mode was initially evaluated for deprotonated ions, it was not used in this study due to its lower sensitivity when compared to the ESI[−] analyses (27). In our analysis of the 18 parent POPs and 46 metabolites and degradation products, we detected a total of 108 different ions as some molecules ionized in both the APPI and ESI sources for both polarity modes. Thus, a total of 108 CCS values are reported in Supplemental Table 1. In the comparison of the ionization modes, most chemicals (n = 42) were detected in ESI[−] mode, followed by APPI[+] (n = 29) and then ESI[+] (n = 21), with 6 chemicals found in all modes (Fig. 1A). ESI[−] also had the greatest number of unique identifications with 30, illustrating its potential as the most comprehensive ion source for these molecules. Since it is established ESI works best for polar molecules, while APPI is optimal for nonpolar compounds (28), our findings also reflect that most of our parent POPs studied were nonpolar and more commonly observed with APPI[+], while their metabolites/degradants were polar and observed in ESI. A majority of the molecules detected with APPI could also be detected using the ESI source, except for PAH, PCB, TCE/PCE, and some pesticides. The different ionization modes and compound classes were then assessed to see if linear correlations in CCS versus m/z plots occurred (Fig. 1B, C). While we noticed that the chemicals greatly overlapped in CCS for the different ionization modes (Fig. 1B), in Fig. 1C, the PFAS separated from all other POPs with a much lower slope due to their high degree of fluorination. In addition, m/z-specific areas such as those for the PAHs were also noted to be smaller than the other classes allowing their distinction. To further illustrate known metabolic and degradation pathways, specific examples for the different classes are given below.</p><!><p>PCBs were commonly used industrial chemicals in adhesives, electrical equipment, and oil-based paints (29) from the 1920s until their production in the USA was banned in 1979. Despite their discontinued production, PCBs can still be detected worldwide, including in Arctic regions (30). Furthermore, exposure to PCBs has known adverse human health effects, with recent findings suggesting some metabolites also have toxic properties (31). Specifically, some hydroxyl PCBs have higher estrogenic activity than their parents, act as disruptors of thyroid homeostasis, and have neurotoxic potential. Additionally, hydroxyl PCBs are transferrable from mothers to fetuses via the placenta (32) and some can inhibit glucuronidation and sulfation reactions, hindering their elimination (31). The specific PCBs analyzed in this study were PCB 3 and PCB 11, because these lower chlorinated PCBs have gained attention in recent years due to their detectability in air samples (33). Furthermore, these PCBs are likely to undergo cytochrome P450 enzyme catalysis where they can hydroxylate and form sulfate metabolites (31) (Fig. 2A).</p><p>In our study, the parent PCBs and their methoxy metabolites were only detected using the APPI source due to their nonpolar structures. Because the parent PCBs and methoxy metabolites do not have functional groups that are easily protonated or deprotonated, they predominantly form radicals with APPI. However, the hydroxyl and sulfate metabolites were preferentially detected using ESI[−] because of the additional polar functional groups which easily deprotonate. Relative abundances of sulfate PCB ions detected by ESI were considerably higher than those of all other PCB ions detected in all modes. The abundances of hydroxylated ions detected by ESI were comparable to those of PCB 3 and the methoxy metabolite detected by APPI (Fig. 2B). Furthermore, isomeric separation was also observed between 3′-PCB 3 sulfate and 4′-PCB 3 sulfate (Fig. 2B). However, while 2′-OH-PCB 3 could be distinguished from 3′-OH-PCB 3 and 4′-OH-PCB 3, 3′-OH-PCB 3 and 4′-OH-PCB 3 could not be separated (Fig. 2B). These findings are noteworthy because OH-PCBs have varying adverse effects. For instance, it has been reported that 4′-OH-PCB 3 is a carcinogen, while 3′-OH-PCB 3 and 2′-OH-PCB 3 are not (34). Hydroxylated and sulfated PCBs also bind to thyroid hormones with varying affinities (35). Although some isomers are indistinguishable with this technique, the ability to separate some of these molecules with IMS is still useful for screening samples and deciding when it is necessary for additional front-end separation techniques such as LC or GC. Additionally, both APPI and ESI sources have a similar response to the PCB ions detected with each mode except the sulfate ions which have much higher relative abundances in negative mode due to their higher ionizability (Fig. 2C).</p><!><p>This study also analyzed common wastewater pollutants and their metabolites. The increasing occurrence of POPs in wastewater has received attention in recent years because of potential adverse human health and ecosystem effects (36). The pollutants of concern in wastewater include PPCPs and industrial chemicals. These have been documented to have endocrine-disrupting effects, possible carcinogenicity to humans and lead to disruptions in the ecosystems, as well as being of possible concern to human health through the food chain (37). Here, we assessed common wastewater pollutants including bisphenol A (BPA) (Fig. 3), paracetamol/acetaminophen (Fig. 4), propofol, mycophenolic acid, morphine, and estrone as many have been documented as endocrine disruptors in fish (38, 39). Furthermore, these chemicals are metabolized with SULT and GULT, and sulfate and mono β-D-glucuronide conjugates are formed (Fig. 3A, Fig. 4A)(40). As expected, when analyzed individually, the parent, sulfate, and glucuronide versions of these molecules were readily separated based on m/z and IMS drift time, but the additional IMS dimension gives further confidence for identifying these compounds from other components in biological and environmental samples. BPA parent and metabolites were all detected in ESI[−] and separated based on m/z and drift time (Fig. 3B). The high polarity of the sulfate metabolite showed higher ionization than both BPA and the glucuronide metabolite. However, despite the low abundance of BPA, it could still be identified in the sample. While the APPI source could only ionize the BPA glucuronide metabolite, the relative abundance was very similar to what was observed using the ESI source (Fig. 3C). In the example of paracetamol and its sulfate and glucuronide metabolites, they were all detected using both ESI[+] and ESI[−] and readily separated based on m/z and IMS drift time (Fig. 4B). Although the sample contained each standard in equimolar concentrations, different relative abundances were observed based on ionization type. ESI[+] had higher ionization of paracetamol glucuronide as an [M+Na]+ ion but low ionization of paracetamol sulfate [M+Na]+, while the opposite was observed for the deprotonated ions using ESI[−]. Also, paracetamol sulfate was not observed as an [M+H]+ ion in ESI[+], while paracetamol glucuronide was. The parent paracetamol had similar ionization response in both positive and negative ESI modes, and while detected with APPI, the response was slightly lower (Fig. 4C). Furthermore, in ESI[+], two conformers were observed for the [M+H]+ adduct of paracetamol. These two peaks may be due to either structural flexibility of paracetamol or the differences in its protonation sites (protomers) as both situations have been observed for other small molecules. Importantly, this signature provides added identification confidence for paracetamol in complex mixtures (41, 42). When examining the metabolites for paracetamol, the sodiated glucuronide conjugate exhibited a lower drift time than its protonated form illustrating compaction due to the Na+ binding both the parent atoms and conjugate together. However, this did not occur in the sulfated conjugate and it is expected that the Na+ only bound to the sulfate oxygens.</p><p>Since it was observed that a majority of the common wastewater pollutants and their metabolite standards ionized best with the ESI source, these chemicals were further evaluated with ESI for rapid screening capabilities. In wastewater samples, the commonly found pollutants include bisphenol A, paracetamol, propofol, and mycophenolic acid and their metabolites with concentrations ranging from detection limits of 0.05–50 ng/L to levels as high as 43,000 μg/L (15, 43). To evaluate the ionization differences and IMS separations of these chemicals with IMS-MS, we combined all of the standards in an equimolar mixture to form our own "wastewater" sample. Similar to when the chemicals were analyzed individually, we were able to detect all parent chemicals except for bisphenol A in ESI[+] and propofol in ESI[−] (Fig. 5A). Ion abundances of all chemicals detected in both positive and negative ESI modes from the equimolar solutions are also illustrated, indicating detection but some ionization differences (Fig. 5B). As expected, the sulfate and glucoronide conjugates had significantly higher abundances in ESI[−] than all the parent chemicals, while several of the sodiated parent chemicals ionized best in ESI[+].</p><!><p>The original Stockholm Convention listed 12 chemicals which were primarily polyhalogenated organic compounds with high lipid solubility; this feature allows them to bioaccumulate in fatty tissue of animals and have great stability and resist hydrolysis and photolytic degradation in the environment (6). Included in these polyhalogenated organic compounds are the organochlorine pesticides aldrin, dieldrin, hexachlorobenzene, and endosulfan. These chemicals and their degradation products (Fig. 6A) were also assessed with IMS-MS since aldrin and endosulfan can degrade through oxidative pathways mediated by microbial or enzymatic processes and hexachlorobenze degrades through dechlorination processes. These chemicals were detected primarily using APPI[+] as most do not have polar functional groups. However, the molecules with sulfate were also detected using ESI and interestingly no significant difference in detected abundances for protonated and deprotonated endosulfan was observed; however, the sodiated ions had comparatively low abundances (Fig. 6B).</p><!><p>PFAS precursors are abundant in the environment and have been detected in many human samples (44). Their transformation products often have longer half-lives and can be more toxic than their precursors. Therefore, the analyses of both PFAS precursors and degradation products are necessary to assess possible human health effects. Limited studies have been conducted on the environmental occurrence and transformation of PFAS (45, 46). However, it is known that 6:2 FTAB, 6:2 FTS, PFOS and PFOA, degrade to shorter-chain PFAS including PFHpA, PFHxA, PFHxS, PFPeA, PFBS, and PFBA as shown in Fig. 7A (44). Degradation of PFAS has been observed in microorganism-mediated processes, activated sludge plants, and aerobic sediment (47). Additionally, remediation methods such as atmospheric pressure plasma jet treatment have elucidated possible degradation pathways of PFOS and PFOA (48).</p><p>Using ESI[−], PFAS were identified by either their [M−H]−and/or [M−HCO2]− ions and separated based on their m/z and IMS drift time values (Fig. 7B). Since the PFAS of interest for this study were preferentially ionizable by ESI[−], this was the only source utilized for their evaluations (17). Furthermore, abundance measurements showed that the longer-chain PFAS such as PFOS, PFHxS, PFOA, PFHpA, and PFBS had similar responses, but the short-chain PFAS including PFHxA, PFPeA, and PFBA had lower relative abundances (Fig. 7C). Next, the AFFFs Firestopper, Tridol, and Chemguard were assessed because these complex mixtures are commonly deployed in massive amounts during fire incidents and are thus released into the environment (49). Understanding the environmental presence of PFAS from these products can assist in exposure assessment and bioremediation efforts (20) (Fig. 7D). In our studies, we specifically looked for 6:2 FTAB since it is a known ingredient of AFFFs and can degrade using a pathway showed in Fig. 7A. Interestingly, in the AFFF analyses, 6:2 FTAB was only detected in Firestopper while it was observed at low levels in Tridol and absent from Chemguard. Firestopper also had high amounts of 6:2 FTS, the subsequent degradation product of 6:2 FTAB, and the only short-chain PFAS detected was PFHxA, illustrating a majority of the components stayed as 6:2 FTAB and 6:2 FTS. In Tridol, both 6:2 FTAB and 6:2 FTS were detected as well as the short-chain degradation products PFHxA and PFHxA. Since Tridol did not contain any of its expected precursors for PFHxA, either full degradation occurred or an additional pathway may exist that is currently unknown. In Chemguard, 6:2 FTAB was not detected but its degradation product 6:2 FTS was detected. PFOA and its subsequent degradation product PFHpA were also observed in this AFFF. In Chemguard, PFHpA showed isomeric separation of branched and linear PFHpA in the drift time distributions, where the branched occurs at a shorter drift time (17). Interestingly, our standard from Fig. 7B only showed the linear form and only the branched form was detected in Tridol. The short-chain PFHxA was also detected in Chemguard which may have occurred by the degradation of either 6:2 FTS or PFOA. IMS-MS analyses of the AAAFs were very impactful in this case as it helped highlight isomeric forms for the PFHpA that may be different in toxicity.</p><!><p>POPs remain a major issue for human and environmental health, and their large-scale production, diversity, complexity, and wide distribution worldwide require new analytical methods to perform rapid and confident exposure assessments. IMS-MS shows great promise for these studies due to its rapid, multidimensional characteristics enabling screening capabilities of POPs and their metabolites and/or degradation products. To demonstrate the utility of IMS-MS for environmental exposure assessment, in this study, we analyzed a wide range (n = 64) of POPs and their metabolites/degradation products, including PCB, PAH, PFAS, industrial chemicals, pesticides, and PCPP. For assessment of the PCB, PAH, and some pesticides, APPI[+] was necessary for their detection due to the nonpolar chemistry of each. However, analyses of the PCB and PAH metabolites preferred ESI[−]. For other common POPs such as PCPP and some industrial byproducts found in wastewater, ESI[+] or [−] may be sufficient for analysis of both parent and metabolite/degradant products and ESI was noted to have higher sensitivity than APPI, providing better rapid chemical screenings at lower concentrations. Additionally, in the analysis of PFAS and their degradants in AFFF solutions, ESI[−] was the optimal analysis mechanism. In all of the studies, IMS-MS illustrated separation capabilities for the isomeric species such as hydroxyl and sulfate PCB, and linear and branched PFAS, although some limitations in the separation of isomeric species, such as the separation of 3′-OH-PCB 3 and 4′-OH-PCB, did occur. While additional separation techniques such as LC may be needed in some cases for the positive identification of molecules such as the PCB metabolites, IMS-MS illustrated a potential screening capability for the POPs and the metabolites and degradants in wastewater and AFFFs without having to perform derivatization and the excessive sample cleanup needed by many current techniques.</p>

PubMed Author Manuscript

BAG3 directly associates with guanine nucleotide exchange factor of Rap1, PDZGEF2, and regulates cell adhesion

BAG3, a member of the Hsc70 binding co-chaperone BAG-family proteins, has critical roles in regulating actin organization, cell adhesion, cell motility and tumor metastasis. The PDZ domain containing Guanine Nucleotide Exchange Factor 2 (PDZGEF2) was cloned as a BAG3 interacting protein. PDZGEF2 induces activation of Rap1 and increases integrin mediated cell adhesion. The PPDY motif at the C-terminus of PDZGEF2 binds to the WW domain of BAG3 in vitro and in vivo. BAG3 deletion mutant lacking the WW domain lose its cell adhesion and motility activity. Gene knockdown of PDZGEF2 leads to the loss of cell adhesion on fibronectin-coated plates while BAG3 overexpression increases cell adhesion in Cos7 cells, but not in PDZGEF2 gene knockdown cells indicating that PDZGEF2 is a critical partner for BAG3 in regulating cell adhesion.

bag3_directly_associates_with_guanine_nucleotide_exchange_factor_of_rap1,_pdzgef2,_and_regulates_cel

Introduction<!>Two-hybrid Screens<!>Cell Culture and antibody<!>Gene silencing using shRNA expression vector<!>Generation of plasmid encoding cDNA of BAG3 and PDZGEF2 deletion mutants<!>Cell Motility and adhesion assays<!>Cloning of PDZGEF2 as an interacting partner of BAG3<!>PDZGEF2 binds BAG3 in vitro and in cells<!>WW domain of BAG3 is critical for regulating cell motility and adhesion<!>PDZGEF2 induces Rap1 activation and increases integrin mediated adhesion<!>BAG3 lose its cell adhesion activity in Cos7 cells with PDZGEF2 gene knockdown<!>Discussion<!>Conclusion<!><!>Analysis of PDZGEF2 interaction with BAG3<!>WW domain of BAG3 is critical for cell motility and adhesion<!>PDZGEF2 activates Rap1 and regulates cell adhesion<!>BAG3 lose its cell adhesion activity in PDZGEF2 gene knockdown cells

<p>The BAG3 protein contains a WW domain near its N-terminus, a proline-rich region (multiple PXXP motifs) and a BAG domain. The WW domain represents a protein interaction module that binds proteins carrying a proline-rich motif having the sequence PPXY [1,2]. Some WW domains recognize their peptide ligands in a phosphorylation-dependent manner. These domains have been identified in several signal transduction proteins that interact with plasma-membrane receptor complexes or with components of the submembranous cytoskeleton (reviewed in [2,3]. BAG3 also contains several SH3-ligand motifs of the sequence PXXP [4], indicating that BAG3 could bind specific SH3-containing proteins that would allow BAG3 to play a role in signal transduction pathways. BAG3 associates with actin at the leading edge of migrating cells and controls cell motility [5]. In many human tumor cell lines, especially adenocarcinomas, the BAG3 protein is highly expressed. Reduction of BAG3 levels by gene knockdown inhibits the invasive and metastatic activity of an epithelial cancer cell line (ALVA31) in vivo, indicating that BAG3 may contribute to the invasive or metastatic phenotype of cancers [5].</p><p>During cell movement, the actin cytoskeleton is tightly regulated in a spatial and temporal manner. Adhesion is mediated by integrin-family proteins attached to the extracellular matrix. Integrins are rapidly recycled via internalization, endocytosis, trafficking and exocytosis while adhesion molecules are internalized into the cell and transported to the leading-edge for reuse in attachment. Recent studies indicated that clathrin-mediated endocytosis of integrins has a dominant role at the leading edge of cells and polarized cells move in response to elevated recycled/internalization at the leading edge. Thus, tightly controlled signaling mechanisms of the actin cytoskeleton, endocytosis and adhesion molecules cooperate to regulate cell movement (reviewed in [6,7,8]).</p><p>To date, many small GTPase proteins are reported to be involved in cell movement. Recently, the PDZ domain containing guanine nucleotide exchange factors, PDZGEF1 (also known as RA-GEF1, nRapGEP and CNrasGEF) and PDZGEF2 were cloned and characterized [9]. Both PDZGEF1 and PDZGEF2 proteins have similar biochemical activity and activate Rap1 and Rap2 small GTPases [9]. The RA domain in PDZGEF1 interacts with Rap1, but not with Ras [10] while its PY domain binds the WW domain of Nedd4 [11]. PDZGEF2 increases integrin-mediated cell adhesion and also has roles in maturation of the adhesion junction [12,13].</p><p>Here we show that the co-chaperone BAG3 directly interacts with PDZGEF2 to regulate integrin-mediated cell adhesion.</p><!><p>Yeast two-hybrid library screening of a human Jurkat cell cDNA library was performed as described previously [14]. Mating tests were then performed using RFY206 yeast strain transformed with pGilda, pGilda BAG3 ΔBAG, or pGilda BAG3 full length. The pJG4–5 cDNAs were recovered using KC8 Escherichia coli strain that is auxotrophic for Trp [14].</p><!><p>Low passage Cos7 cells were cultured in DMEM medium with 10% Fetal Calf Serum (FCS) with penicillin/streptomycin. Transfections were performed with FuGENE (Roche, Indianapolis). Anti-PDZGEF2 polyclonal antibody generated against GST fusion protein of PDZGEF2 clone F11 and affinity purified. Anti-GFP, -Rap1, Myc (SantaCruiz Biotech, SantaCuiz). Anti-Actin (Calbiochem, La Jolla). Anti-Flag, -alpha tubulin (Sigma, Saint Louis).</p><!><p>Gene silencing of bag3 was accomplished using the shRNA expression vector pLTRH1puro as described previously [15]. For PDZGEF2 knockdown, we generated pLTRH1 puro vector with oligonucleotides (shGEF2#1 CUGCAUGAAUUGUAACUGAA, shGEF2#2 CAGGAAGAAGGGACAAACAAA) as previously reported by Dube et al [12].</p><!><p>Deletion mutants were made using two-step PCR-mediated mutagenesis as described previously [14]. Full-length BAG3 cDNA was used as a template to create a Flag tagged version of full-length bag3 and deletion WW (Deletion nucleotide 52–159 from the first ATG, amino acid 17–53) [14]. PDZGEF2 cDNA was amplified using RNA extracted from HEK293 cell, followed by RT-PCR (GenBank ID: NM_001164386). Deletion mutants were generated using two-step PCR (ΔPPGY: deletion amino acid 1516–1519, ΔPPDY: 1537–1540).</p><!><p>Cos7 cells were transfected with pEGFP-BAG3, pEGFP-BAG3ΔWW, pcDNA3-FAK, pEGFP vector with or without PDZGEF2 knockdown and used for motility and adhesion assays as previously described [5]. A paired Student's t-test was used to analyze differences between two groups, and p vales of <0.5, <0.05 or <0.01 were considered significant.</p><!><p>To identify proteins that interact with BAG3, we carried out yeast two-hybrid screens of human Jurkat T cell cDNA libraries of 2×106 clones, using the full-length BAG3 protein as bait. After performing several tests to confirm the specificity of these interactions, we found several candidate BAG3-binding proteins: i) PDZGEF2, a guanine nucleotide exchange factor that contains a PDZ domain and reportedly activates the Ras-family member Rap1 [9], ii) cDNA clones with sequences for uncharacterized SH3 or praline-rich proteins, and iii) overlapping clones of Hsc70. Three different fragments were cloned as overlapping clones of PDZGEF2 (RapGEF6), containing amino acid 1326–1609 (clone A7), 1444–1609 (5 clones including F11 clone) and 1506–1609 (2 clones) as shown in Figure 1A.</p><p>The BAG3-binding protein PDZGEF2 is a close homologue of PDZGEF1 and has similar peptide sequences at the N-terminus, but has limited similarity at C-terminus, the site of BAG3 interaction. The clone of PDZGEF2 (A7) having the longest sequence has only 39% identity and 52% similarity with the PDZGEF1. In contrast, the N-terminal domain has 72%–85% identity and 82–93% similarity, suggesting a functional similarity but different binding affinity with BAG3. In the homologous region, PDZGEF1 and PDZGEF2 contain (from N-terminus to C-terminus) a cyclic nucleotide monophosphate (cNMP)-binding homology domain, Ras-exchange domain (REM), a PDZ domain, a Ras/Rap association (RA) domain, a putative GDP/GTP Exchange Factor (GEF) domain, a PY domain (PPXY motif, which is known as WW domain interacting motif), and a C-terminal PDZ-ligand motif (S/TXV) motif [9]. The highest PDZGEF1 expression levels detected by Northern blot analysis were localized to brain, heart, placental, skeletal muscle, kidney and pancreatic tissue [9]. In contrast, PDZGEF2 is more broadly expressed among tissues.</p><p>In our screens, all BAG3-interacting clones carried the C-terminal PPXY motif. For in vitro binding assays, the A7 and F11 clones were subcloned and used for further analysis. The shortest interacting clone had only 103 nucleotides and was not used for subsequent assays.</p><!><p>PDZGEF2 binding to BAG3 was confirmed by in vivo and in vitro binding assays using endogenous protein, an overexpression system in 293T cells and GST-fusion proteins in combination with in vitro- translated 35S methionine labeled PDZGEF2.</p><p>To detect endogenous PDZGEF2, we generated an anti-PDZGEF2 polyclonal antibody (F11) against a recombinant PDZGEF2-F11 clone corresponding to amino acids 1444–1609 as described above. Rabbit anti-PDZGEF2 antibody was purified by an affinity purification method described previously [5]. Using the prostate cancer cell line ALVA31, immunoprecipitation experiments were performed using anti-BAG3 and F11 antibodies. ALVA31 cells were used for initial report describing regulation of motility and adhesion by BAG3 [5]. Immune complexes were isolated with an anti-BAG3 antibody or F11 antibody, followed by immunodection using anti-F11 or anti-BAG3 antibodies. In Figure 1B, immunoprecipitation of BAG3 co-purifies with PDZGEF2 (anti-BAG3). In the third lane, cell lysates were loaded directly to detect endogenous expression of BAG3. IgG1 was also used for immunoprecipitation as negative control (IgG). Using the anti-PDZGEF2 antibody and anti-BAG3 polyclonal antibody, we detected an endogenous protein-protein complex in ALVA31 cells.</p><p>To determine the BAG3 interaction domain of PDZGEF2, we constructed BAG3 deletion mutants lacking either the WW (ΔWW), PXXP (ΔPXXP), or BAG domains (ΔBAG) as described previously [5]. The interaction of BAG3 with PDZGEF2 C-terminal fragments (PDZGEF2A7) was demonstrated by an in vitro binding assay using GST fusion protein of the full-length and BAG3 deletion mutants. The in vitro-translated PDZGEF2 protein fragments represented the same cDNA clones (A7 clone) as those pulled out by two-hybrid screening, and encompassed the C-terminal portions of PDZGEF2 protein (A7 clone). From these results, we concluded that the WW domain of BAG3 and C-terminal portion of PDZGEF2 is sufficient for binding in vitro (Figure 1C).</p><p>Domain dependent interactions were also confirmed by co-immunoprecipitation assays using lysates from human HEK293T cells transfected with plasmids encoding Myc epitope-tagged versions PDZGEF2. Flag-tagged versions of full-length BAG3 and the three deletion mutants were recovered in Myc-PDZGEF2A7 (1326–1609) immune-complexes, as determined by immunoblotting using anti-myc antibody (Figure 1D).</p><p>The cDNA clones obtained in the 2-hybrid screens using BAG3 as bait encode the C-terminus of PDZGEF2, which includes the two PPXY motifs (PPGY, starting at amino acid 1516, PPDY, starting at amino acid 1537). We generated deletion mutants of PPGY (ΔPPGY), PPDY (ΔPPDY) and both (Δ2PY) in the C-terminal fragment of A7 clone of PDZGEF2 with a myc tag. Overexpression of Flag-tagged BAG3 and Myc-tagged wild type PDZGEF2A7 fragment or mutants were used for immunoprecipitation assays (Flag antibody), followed by immunodetection by myc antibody to detect co-precipitated C-terminal fragments of PDZGEF2. As shown in Figure 1E, ΔPPDY and Δ2PY lost binding activity with BAG3, suggesting PPDY, but not the PPGY motif specifically interacting with BAG3. These data provide supporting evidence that PDZGEF2 may be a physiologically relevant target of BAG3.</p><!><p>We observed that overexpression of BAG3 in Cos7 cells increased cell adhesion and cell motility in transient motility assays conducted under serum-depleted conditions [5]. Figure 2A shows a comparison of the motility of Cos7 cells transfected with pEGFP BAG3 ΔWW, pEGFP wild type BAG3 or pEGFP plasmids. Cell motility assays were performed using a transwell system. Deletion of the BAG3 WW domain caused a significant reduction in cell motility. Since our cDNA cloning data and interaction assay indicated that the WW domain interacts with PDZGEF2 (RapGEF6) and Rap1 is known as positive regulator of integrin-mediated cell adhesion [13,16], we performed adhesion assays using fibronectin-coated plates. Figure 2B shows that BAG3ΔWW significantly reduces adhesion activity on fibronectin coated plates, suggesting that the WW domain of BAG3 is important for integrin-mediated cell adhesion. Figure 2C indicates the stable expression of pEGFP-BAG3 and pEGFPBAG3ΔWW in this assay.</p><!><p>Full-length cDNA of PDZGEF2 was cloned by RT-PCR, using total RNA extracted from HEK293T cells. The small GTP-binding protein Rap1 is known to be a downstream target of PDZGEF2. To detect Rap1 activation by PDZGEF2, we compared the activity of Rap1 in control Cos7 vs. cells overexpressing PDZGEF2. Cell lysates were prepared, followed by recovery of the GTP-bound form of Rap1 by adsorption to immobilized GST-RalGDS fusion protein. For positive and negative controls, we used lysates incubated with non-hydrolyzable GTP (GTPγS) and GDP, respectively. Greater amounts of active Rap1 were recovered from lysates of PDZGEF2 overexpressing cells compared to control Cos7 cells (Figure 3A). In contrast, GTPγS-treated lysates demonstrated comparable total levels of Rap1 in both control and PDZGEF2 overexpressing cells (not shown). This result suggests that PDZGEF2 regulates Rap1 activity. Since Rap1 is reported to be a positive regulator of integrin-mediated adhesion, we compared the rate of attachment of PDZGEF2 overexpressing cells to fibronectin-coated (integrin-dependent ligand) plastic plates. In short-term adhesion assays (20 min), more PDZGEF2 over-expressing Cos7 cells adhered to fibronectin-coated plates than control transfected Cos7 cells (Figure 3B). Next, we used a transwell assay to address whether PDZGEF2 increases cell motility. Since PDZGEF2 interacts with the WW domain of BAG3 and BAG3ΔWW expressing cells have reduced motility, we examined whether PDZGEF2 overexpression increases cell motility in a transwell assay. As shown in Figure 3C, PDZGEF2 and control-transfected Cos7 cells displayed no difference in cell motility.</p><!><p>We examined whether inhibition of PDZGEF2 function influences BAG3 function in cell adhesion. We generated PDZGEF2 shRNA expression vector using two different oligomers, shGEF2#1 and shGEF2#2, as described in the Materials and Methods section. Cos7 cells were infected with retrovirus containing PDZGEF2 shRNA and stable lines were selected using puromycin-containing medium (1μg/ml). Expression of PDZGEF2 protein were confirmed by western blot, which indicated that shGEF2#2 knocked down PDZGEF2 expression more effectively than shGEF2#1, achieving suppression of PDZGEF2 expression by ~80% and ~50%, respectively. Using shGEF2#2 transfectants, pcDNA3mycbag3 full-length (pcDNA3BAG3) and control plasmids (pcDNA3) were transiently transfected (Figure 4A), followed by adhesion assays on fibronectin coated plates. As shown in Figure 4B, BAG3 overexpression does not increase cell adhesion in PDZGEF2 knockdown cells, indicating that PDZGEF2 interaction is sufficient for adhesion activity controlled by BAG3. Exogenously overexpressed myc-tagged bag3 was also confirmed by immunoblot with Myc antibody (not shown).</p><!><p>In this report, we demonstrate that BAG3 direct interacts with the guanine exchange factor of Rap1, PDZGEF2. The BAG3 WW domain interacts with the PPXY motif of PDZGEF2, a known motif of WW domain interaction. There are two PPXY sequences in PDZGEF2. Deletion mutant of either of the PPXY sequences indicated that only the second PPXY motif (PPDY) is critical for the BAG3 interaction. BAG3 WW domain deletion mutants were made and used for cell motility or adhesion assays. These suggested the importance of the WW domain in regulating motility and adhesion. The PPXY motif is also found in PDZGEF1, which is a close homologue of PDZGEF2. Although the C-terminal region of PDZGEF2 and PDZGEF1 has only low similarity, PDZGEF1 and PDZGEF2 have two PPXY motifs, (PPGY and PPDY), suggesting that PDZGEF1 may also interact with BAG3. The PPXY domain of PDZGEF1 binds to the WW domain of the E3 ubiquitin ligase Nedd4, which targets it for ubiquitin proteasomal degradation [17]. Both the PPGY and PPDY motifs of PDZGEF1 interact with Nedd4 ubiquitin ligase, but BAG3 may only interact with the PPDY motif. It will be of interest to investigate the Nedd4 ubiquitin ligase interaction with PDZGEF2 and whether BAG3 can compete with Nedd4 for interaction. The BAG3 WW domain also interacts with adenovirus penton base protein. The adenovirus penton base is viral capsid protein responsible for viral internalization and direct interaction of BAG3 with this protein increases nuclear migration, suggesting that BAG3 may contribute to the adenovirus lifecycle by [18].</p><p>Both of PDZGEF1 and PDZGEF2 have the C-terminal S/TXV motif, which is known as a PDZ ligand motif. Two PDZ proteins (S-CAM and MAGI-1) have been reported to associate with PDZGEF1 via this motif [19,20]. It will be interesting to find whether PDZGEF2 partners with PDZ ligand motifs in the BAG3 complex to better understand the function of BAG3 in regulating PDZGEF2 activity.</p><p>Previously we showed that BAG3 overexpression increased, and shRNA-mediated BAG3 protein reduction decreased cell adhesion on fibronectin coated plates. In this report, our results suggest that BAG3 is a regulator of PDZGEF2 and this interaction may potentially connect the activation of Rap1. Since PDZGEF2 increase cell adhesion, but not cell motility, further investigation is necessary to elucidate the molecular mechanisms of BAG3 regulation of cell motility. BAG3 interact with PLCγ and deletion of PXXP domain alter FAK phosphorylation [21,22]. It is possible that multiple SH3 proteins may link BAG3 with cell movement processes. Further analysis of physiologically relevant proteins that interact with the PXXP motifs of BAG3 may reveal the importance of SH3 proteins in BAG3-mediated regulation of cell motility.</p><p>A close homologue of PDZGEF2, PDZGEF1, is functionally similar in that both regulate GDP/GTP exchange on Rap1 and Rap2 [9], which are small GTPases known to regulate cytoskeletal assembly and cell migration (reviewed in [23]). The RA domain of PDZGEF1, which permits interaction with Rap1 but not Ras, is required for GEF activity and the PDZGEF1-Rap1 interaction stimulates Rap1 function via positive feedback [10]. The RA domain of PDZGEF2 interacts with M-Ras, which activates and translocates PDZGEF2 to the plasma membrane [24]. Gao et al. originally cloned a C-terminal truncated isoform as full-length PDZGEF2 and showed Rap1 activation with this isoform [24], which should not interact with BAG3. Although we obtained only a single cDNA fragment by RT-PCR, there are six potential alternatively spliced forms of PDZGEF2 found in the genome database that result in different C-terminal sequences. Isoforms #1 (GenBank ID: NM_001164386) and #2 (NM_016340.5) contain PPXY motifs, but other isoforms lack the C-terminal peptides, including the PPXY motif. Thus, we expect isoforms #3 (NM_001164387.1), #4 (NM_001164388.1), #5 (NM_001164389.1) and #6 (NM_001164390.1) would be functionally independent from BAG3 association or could work as dominant negative inhibitors of the BAG3/PDZGEF2 complex. According to sequence data of splicing variants, our shRNA knockdown of PDZGEF2 should be ineffective for isoform 6. Isoforms #2 and #5 have an 8 amino acid deletion in the GEF domain, although whether this affects their GEF activity is unknown. Our antibody detects one major band at a molecular weight of about 180kDa. Overexpression of full-length constructs produces the same molecular weight band. Thus, we expect isoform #1 (1609 amino acid) transcript to be the major product in our system, but it will be very important to examine expression pattern of each isoform in different cell types or tissues, especially whether the PPXY motif is present.</p><p>Recently, PDZGEF2 was found to be a partner protein of junctional adhesion molecule-A (JAM-A), which interacts with Afadin. PDZGEF2 regulates cell-cell adhesion to integrin medicated cell-ECM regulation by interacting with JAM-A [25]. Rap1 signaling is reported to be activated by E-cadherin internalization and increase intergrin mediated cell-ECM adhesion, suggesting regulation of signals from cell-cell adhesion to cell-ECM adhesion by Rap1. Maturation of the adhesion junction is also regulated by PDZGEF2 [12]. This evidence suggests that PDZGEF2 regulates both cell-cell and cell-ECM adhesion. The physiological activators of PDZGEF2 remain unclear, but our investigation may help reveal the role of BAG3 in the physiological function of PDZGEF2 in regulating cell adhesion and motility.</p><!><p>The WW domain of BAG3 interacts with the PPDY motif of PDZGEF2 (RapGEF6). Deletion of the BAG3 WW domain causes a significant reduction in cell motility and adhesion, while BAG3 over-expression does not increase cell adhesion in PDZGEF2 knockdown cells. These results indicate that PDZGEF2 is a critical partner of BAG3 in regulating cell adhesion.</p><!><p>PDZGEF2 was cloned as a BAG3 interacting protein.</p><p>The PPDY motif of PDZGEF2 binds to the WW domain of BAG3.</p><p>BAG3 deletion WW mutant loses cell adhesion and motility.</p><p>PDZGEF2 is a critical partner of BAG3 for regulating cell adhesion.</p><!><p>(A) Schematic structure of PDZGEF2 domains. The cNMP-binding homology, REM, PDZ, RA, GEF, PPXY domains are indicated by open boxes. The regions corresponding to the eight BAG3 binding cDNA clones obtained by two-hybrid screens are indicated by 3 lines located at the right upper corner. (B) Endogenous BAG3 and PDZGEF2 interact in cells. Endogenous BAG3 was immunoprecipitated by AR100 (polyclonal anti-BAG3 antibody), followed by immunoblot using anti-PDZGEF2 antibody. (C) GST-fusion proteins encoding BAG3, BAG3 deletion mutants (ΔWW, ΔPXXP, ΔBAG), CD40 cytosolic domain (negative control) were used for in vitro binding assay with 35S-L-methionine labeled PDZGEF2A7. As a control, an equivalent amount of input in vitro translated PDZGEF2A7 (IVT) was loaded directly onto the gels. (D) Immunoprecipitation assay using Myc-tagged PDZGEF2A7 and Flag-tagged BAG3 (FL, ΔWW, ΔPXXP, ΔBAG). Immune-complexes were precipitated by Flag antibody, followed by western blot detected by myc for PDZGEF2 (upper panel). Cell lysate were directly applied for immunoblotting, followed by detection with Myc antibody (middle panel) and Flag for BAG3 constructs. (E) Deletion of two PPXY motifs in PDZGEF2 indicating specific interaction with BAG3 by PPDY. Myc tagged PPGY deletion (ΔPPGY), PPDY deletion (ΔPPDY), or both (Δ2PY) or control were made in C-terminal fragments of PDZGEF2 (A7 fragment), followed by immunoprecipitation assay using BAG3 co-transfection in 293 cells. Immune-complex was prepared by Flag antibody and immunoblotted by Myc antibody (upper panel).</p><!><p>(A) Cells transfected with BAG3ΔWW deletion have impaired integrin-mediated adhesion. Cos7 cells were transfected by BAG3 wild type (BAG3), WW domain deleted (ΔWW) and pEGFP empty vector, followed by adhesion assay (*p<0.01, mean ± SD, n=6). (B) Impaired cell motility found in Cos7 cells with BAG3ΔWW (*p<0.01, mean ± SD, n=6). (C) BAG3 wild type (pEGFP-BAG3) and pEGFP-BAG3ΔWW were transfected into 293T cells. Western blotting with GFP antibody indicated stable expression.</p><!><p>(A) Rap1 activation was detected in PDZGEF2-transfected COS7 cells using GST-RalGDS pull down (1st panel). Total Rap1 expression (2nd panel). The overexpression of MycPDZGEF2 was detected by Myc antibody (3rd panel). Anti-PDZGEF2 antibody F11 detects endogenous PDZGEF2 in control transfected cells as well as overexpressed myc PDZGEF2 (4th panel). Actin as a loading control (5th panel). (B) PDZGEF2-transfected Cos 7 cell increase adhesion on fibronectin-coated coated plates. Data represent percentage change in the number of adherent cells relative to control-transfected cells, 20 minutes after plating (*p<0.01, mean ± SD, n=6). (C) Cell motility assay. Cos7 cells were transfected with plasmids encoding pcDNA3 mycPDZGEF2 or with empty control pcDNA3 plasmid, followed by serum starvation and transwell assay (8 μm pore size) for 4 hours. Data represent mean ± SD (n = 6).</p><!><p>(A) PDZGEF2 knockdown cells (shGEF2) or control retrovirus infected cells were used for overexpression with pcDNA3 myc-bag3 construct (pcDNA3bag3) or pcDNA3 vector alone, followed by immunodetection with an anti-PDZGEF2 antibody (upper panel) and anti-BAG3 antibody (middle panel). Tubulin served as a loading control (lower panel). (B) PDZGEF2 gene knockdown or control infected cells were transfected with full length expression vector of BAG3, followed by an adhesion assay using crystal violet. Data represent mean± SE (n=6, *p<0.01).</p>

PubMed Author Manuscript

Thieme Chemistry Journal Awardees - Where are They Now? Catalytic Transport with an Amine Carrier in a Fluorous Triphasic Reaction

Several aromatic aldehydes are transported by a fluorous amine from one organic phase through a fluorous phase to another organic phase. The derived imines react with phenylhydrazine to immobilize the transported product as a hydrazone and release the amine for reuse. In this way, catalytic transport is accomplished for the first time.

thieme_chemistry_journal_awardees_-_where_are_they_now?_catalytic_transport_with_an_amine_carrier_in

<p>The production of new organic molecules is typically comprised of a reaction and a separation, and these two stages are almost always conducted separately in space and time. Conducting a reaction and a separation coincidently has the potential to result in a synthesis device (or machine) where starting materials are fed in and products are harvested out continuously. Such devices could potentially be piggybacked to provide for continuous, realtime, multistep reaction and separation.</p><p>The separation component of any synthesis device requires some kind of partitioning, and the immiscibility of fluorous and organic liquids offers attractive opportunities.2 In 2001, we introduced fluorous triphasic devices for transportative deprotection of fluorous-tagged materials.3 In the example shown in Figure 2, a mixture resulting from enzymatic enantioselective hydrolysis of a racemic fluorous-tagged alcohol was simultaneously separated and detagged.4 The mixture was comprised of roughly equal amounts of the free alcohol (R)-1 and its enantiomeric counterpart (S)-2 still bearing the fluorous tag. Three steps are needed to complete the resolution: 1) the alcohol (R)-1 and the ester (S)-2 must be separated; 2) the retrieved ester (S)-2 must be hydrolyzed, and 3) the residual tag must be separated from the resulting alcohol (S)-1</p><p>The triphasic device in Figure 1, a simple U-tube, accomplishes all three tasks simultaneously. The U-tube contains two organic phases (source and receiving) separated by a fluorous phase that prevents direct contact between these phases and thereby regulates exchange. A mixture of (R)-1 and (S)-2 is added to the source side of the device of sodium methoxide is to the receiving side. Over time, the ester (S)-2 partitions through the fluorous phase over to the receiving phase, where it is detagged by the base. This strands the resulting alcohol (S)-1 in the receiving phase, while the residual tag (which is now highly fluorous) partitions back to the fluorous phase. The starting alcohol (R)-1 stays put in the source phase. Thus, in an ideal process, there is one product in each of the three phases.</p><p>This kind of 'transportative detagging' of fluorous-tagged substrates is a potentially general way to remove non-tagged contaminants along with removal of the fluorous tag. But such processes are by definition stoichiometric in the fluorous component (the tag). Herein, we report the first catalytic fluorous triphasic reaction and separation process.</p><p>In principle, it should be possible in a fluorous-organic triphasic device to convert an organic substrate to an organic product with the aid of a substoichiometric amount of a fluorous carrier. We set out to prove this principle by transporting an aldehyde from the source phase though the fluorous phase with a fluorous amine carrier via the intermediacy of a fluorous imine. Aldehydes are important raw materials, and their feedstocks can be contaminated with alkenes, alcohols, and other impurities. Thus, their separation with a simultaneous onward reaction has potential value.</p><p>The reactions involved in the process are summarized with 4-bromobenzaldehyde (4) in Scheme 1. In the source organic phase, 4 reacts with fluorous amine carrier 3 to give fluorous imine 5. In turn, the imine 5 is transported through the fluorous phase to the receiving organic phase, where it reacts with added phenyl hydrazine 6 to provide hydrazone 7. The energy released in this reaction drives the separation, and the amine carrier is also released for return through the fluorous phase to the source organic phase.</p><p>We expected the rates of both reaction steps to be relatively fast, and thus the rate of the overall reaction and separation process will be limited by the transport rate. Our previous stoichiometric triphasic reactions were also transport-limited, and often took several days to go to completion in standard U-tubes. Reducing the fluorous component to a substoichiometric amount will further slow the process, and accordingly we deemed the U-tube device design unsuitable.</p><p>In principle, the transport rate problem can be solved by engineering solutions in which the fluorous phase is circulated between the source and receiving phases. We decided to mimic this engineering solution with serial extractions, as summarized in Figure 2. Thus, the source phase with the aldehyde 5 and the fluorous phase with the amine 3 were vortexed together in a vial to allow reaction and subsequent partitioning of the resulting imine 5. The phases were then separated and the forward fluorous-phase-containing part of this imine 5 was vortexed with the receiving phase to allow reaction to form the hydrazone 7 with release of the imine 5. Then the back fluorous phase containing the amine 3 was returned to the source phase and the whole process was repeated in a series of identical cycles.</p><p>If the reactions are rapid and quantitative, and if the partitioning of the fluorous amine/imine into the fluorous phase is also quantitative, then it is possible to transport in each cycle an amount of aldehyde that is equal to the amount of amine added. Control experiments suggested that the two reactions were rapid and quantitative, as expected.5</p><p>But the partitioning of the amine 3 and imine 5 into the fluorous phase will surely not be quantitative since the fluorine content of these compounds is not nearly high enough.6,7 In that case, after a few cycles, a kind of quasi-equilibrium will be reached in which there is unproductive amine 3 in both the source (as imine) and receiving phases and productive amine 3 in the fluorous phase (the imine is present in the forward direction and the free amine in the reverse direction). In this scenario, the amount of aldehyde transported in each cycle will equal the amount of productive amine. Provided that the aldehyde is present in excess, a plot of transport amount versus time should be linear and should not depend on the aldehyde concentration.</p><p>In a typical experiment, 4-bromobenzxaldehyde (4, 23.2 mmol) and (perfluorooctyl)propylamine (3, 0.56 mmol, 25 mol%) in acetonitrile (7 mL) and FC 72 (7 mL) were vortexed in the source vial for one hour. The phases were allowed to separate, and the FC-72 phase was transferred to the receiving vial containing phenylhydrazine (6, 2.3 mmol) in acetonitrile (7 mL). This was similarly vortexed for one hour. The FC phase was then returned to source vial to start the second cycle. Cycles were repeated until phenylhydrazine was largely consumed.8 The progress of the reaction was followed by GC. We could not detect either imine 5 or aldehyde 4 in the receiving phase. Also, there was no detectable back transport of either the phenylhydrazine (6) or the product hydrazone 7 to the source phase.</p><p>Figure 3 plots the accumulation of hydrazone 7 in the receiving phase as a function of cycle. As expected, the plot is linear, and about 0.2 mmol of aldehyde 4 is transported from the source phase to the receiving phase in each cycle. This means that about 35% of amine 3 that was added was productive in each cycle.</p><p>After the 10th cycle, the yield of phenylhydrazone based on aldehyde was 89%, and the yield based on amine was 350%. This shows the catalytic nature of the transport. The GC analysis showed that the amine 3 was largely intact after the 10th cycle, distributed over the three phases, We conducted a similar experiment with about 10% of the amine catalyst, and about 40% of the phenylhydrazone 7 was produced in the receiving vial after ten cycles (about 4% transport per cycle). This is consistent with the expectations based on the amount of productive amine from the first experiment.</p><p>These experiments need to be controlled for direct (passive) transport of the aldehyde through the fluorous phase to the receiving phase. In these first experiments, we did this indirectly by measuring the partition coefficient of the aldehyde between FC-72 and acetonitrile. This was <0.01, so significant passive transport seems unlikely.</p><p>We then conducted a second series of cycles with the same procedure mentioned above with 4-fluorobenzaldehyde and 10% amine 3. This time, we did a parallel series of control cycles with all the same components, but omitting the fluorous amine. The plots of hydrazone yield versus cycle for both experiments are shown in Figure 4. The control experiment, which was stopped after the 4th cycle, shows that the background (passive) transport rate is not zero, but it is rather small (and 1% per cycle). This is presumably due to direct partitioning of the aldehyde into the FC-72.9</p><p>The results for the transport experiments with 4-fluorobenzaldeyde were similar to those with 4-bromobenzaldehyde. With 10% of the amine 3, about 40% of the aldehyde was transported to form the hydrazide after the 10th cycle. The initial rate of transport in the experiments with the amine exceeds that rate of the background transport by a factor of about 5. This again demonstrates the effect of the amine/imine on the transport. We did similar experiments with benzaldehyde and 4-trifluoromethylbenzaldeyde with generally similar results (not shown). In each of these case the initial rate of transport in the experiment with the amine exceeded that of the control by a factor of 3.</p><p>These experiments demonstrate for the first time that an organic reactant can be transported from one organic phase through a fluorous phase to another organic phase by a substoichiometric amount of a fluorous transport agent. The experiments are prototypes for synthesis devices that accomplish a simultaneous reaction and separation and that are catalytic in the fluorous component. Assuming that the reaction rates are fast, the efficiency of such devices will depend on the efficiency of the transport. This is subject to optimization through chemistry (changing solvents, catalysts, etc. to alter partition coefficients) and through engineering (for example, by using continuous extractors).</p>

PubMed Author Manuscript

Heavy-atom tunnelling in Cu(ii)N6 complexes: theoretical predictions and experimental manifestation†

The degenerate rearrangement on Jahn–Teller distorted metal complexes is a promising reaction for the observation of significant heavy atom quantum mechanical tunnelling. Herein, a family of Cu(ii)–N6 complexes are theoretically proven to exhibit rapid dynamical Jahn–Teller tunneling even close to the absolute zero. The manifestation of our predictions apparently appeared in solid state EPR experimental measurements on [Cu(en)3]SO4 more than 40 years ago, without the authors realizing that it was a quantum outcome.

heavy-atom_tunnelling_in_cu(ii)n6_complexes:_theoretical_predictions_and_experimental_manifestation†

Introduction<!>Methods<!>Results and discussion<!>Conclusions<!>Conflicts of interest

<p>The Jahn–Teller effect1,2 (JTE) predicts that a non-linear system with degenerate electronic states will distort in order to lift the degeneracy and lower its energy. In many cases, the distortion leads to a set of similar isoenergetic isomers, generating a "multi-well" degenerate potential energy surface. In Cu(ii) octahedral complexes, possibly the most studied compounds of this type,3,4 the JTE leads to tetragonal distortions due to a breakage of the degeneracy of the antibonding eg orbitals. This forms "elongated" and "compressed" geometries, generating a multi-well system known as the "warped Mexican hat" (Fig. 1).3,5–9 The nature of the JT distortion (elongated, with antibonding occupation, or compressed, ) cannot be easily predicted, and both geometries are theoretically valid for the first order.3,5,10 However, with six identical ligands (homoleptic complex) the elongated form will always be energetically favourable due to pseudo-JT correction5 (even in non-homoleptic complexes the compressed form is rarely observed).3</p><p>The interconversion between these isomers occurs with relative ease via a compressed geometry transition state (Fig. 1). Low barriers allow a rapid transition (high automerization frequency in a "dynamic" JT distortion), while high activation energy supposedly hinders the reaction ("static" JT), especially at low temperatures. However, an alternative path to the over-the barrier automerization exists even close to the absolute zero, consisting of quantum mechanical tunnelling (QMT) driven dynamic JT.</p><p>The role of heavy atom QMT11–13 in molecular systems (that is, any atom heavier than H or He) has been studied since the early 80's, starting with the degenerate π-bond shifting of cyclobutadiene.14 Since then, many other degenerate rearrangements have been seen to react by a QMT mechanism.15–22 As any distortion lowers the symmetry of a molecule, all of these reactions are driven by a double (or multi) well potential energy surface created due to the different flavors of the JTE.2 And in all of these reactions, the fast tunnelling rate is caused by the small particle mass (actually, the small reduced mass of the system in the reaction coordinate), the relatively low barrier height, and most critically, the narrow tunnelling distance.23 Common experimental indications of tunneling are a high kinetic isotope effect (KIE) and temperature independent rates (producing negligible Arrhenius activation energies and low pre-activation factors).23,24</p><p>In contrast, QMT dynamic JT in solid state systems was already proposed25,26 by Bersuker in 1963 and later confirmed experimentally mostly by cryogenic EPR detection of tunnelling splitting (3Γ) in solid solutions of JT active centres in insulators (such as Cu2+ doped MgO2,9,27–33). These systems have seen a revival due to their potential use in quantum computers, colossal magnetoresistance, and even in superconductivity.32 It is worth noting that solid and gas phase chemistry bear completely different surrounding conditions. Crystal structures are dominated by strong pressures and interactions (including counterion effects34) that might force the complexes to stay in a defined, static isomer. But in JT systems where the atomic displacement is short (measured here as the JT radius, see below) and with almost insignificant chemical changes, the crystal pressure actually enhances the QMT dynamics,32 as it constrains the atomic trajectories. This, combined with the fact that the rearrangement barriers for Cu oxides are radically low (of the order of one kJ mol−1), makes the tunnelling close to the absolute zero extremely probable. In fact, due to such low activation energies, the EPR study has to be carried out at extremely low temperatures (close to 1 K), to distinguish the QMT and classical dynamic JTE, and to avoid other dynamical effects.</p><p>The also known nitrogen based Cu(ii) complexes are tougher JT systems.7,8,34,35 Although it was speculated that QMT might play a role in the dynamic JTE of such complexes (specifically in Cu(ii)-doped hexaimidazole in a Zn(ii) matrix36 at 77 K), later on these observations were disproved.37–40 "Genuine" dynamic JT has been seen in many of these crystal cases (as seen in the temperature dependency of Cu–N bond lengths), but tunnelling from the ground state for Cu(ii)–N6 complexes seemed to be, apparently, impossible.</p><p>Herein we present computational evidence of the crucial role of QMT during the degenerate rearrangement of Cu(ii)–N6 type complexes in the gas phase under cryogenic conditions (which simulates the experimental results that can be obtained in supersonic expansion techniques,41 co-deposition with noble gas weakly interacting matrices,19,42 or even in He nanodroplets43). Even if not recognized at that time, this effect can actually be seen in a long-standing solid state EPR experiment.34</p><!><p>All the automerization rate constants were computed with semi-classical canonical variational theory (CVT)44 adding accurate multidimensional tunnelling correction with the small curvature tunnelling (SCT) method45,46 (including quantized reactant state tunnelling – QRST47 only at low temperatures, and with a small step size of 0.001 bohr, see the ESI†). In heavy atom QMT severe corner-cutting is uncommon, and the relatively small differences between ZCT and SCT values (approximately an order of magnitude) justify the use of the latter without requiring large curvature tunnelling corrections. DFT computations were carried out with Gaussian 16,48 while the rate constants were computed with Polyrate 17,49 with Gaussrate 17B50 as an interface to Gaussian.</p><p>Since QMT computations are highly demanding, a fast functional and basis set combination was carefully chosen after a benchmark procedure on the activation energy of Cu(ii) systems with en, ein, NH3, biea and timm ligands (see Fig. 2). For the reference energies we used DLPNO-CCSD(T)/aug-cc-pVQZ//MN15/Def2-TZVPD with tight PNO criteria (computed with ORCA 4.0).51–55 This method is not foolproof, but is orders of magnitude more reliable than any DFT scheme.51 No severe static correlation was found using the T1 and the % TAE(T) diagnostics,56,57 and negligible differences were found between the highest PNO levels (see the ESI†), justifying the selected reference method (especially considering the impossibility of using canonical CCSD(T) with complete basis set schemes on larger molecules). From all the functionals and basis sets tested, we found the PBE0/6-31+G(d) method to be the most accurate while still being relatively fast (see Tables S1–S3 in the ESI†).58</p><p>We acknowledge that even with the high-level SCT tunnelling method and the selection of the functional through careful benchmarking, small errors in the geometries and energies can lead to exponentially large errors in the computation of the rate constants. Therefore, the presented results are not to be taken at face value. Still, our predictions and conclusions stand as semi-quantitative values, possibly within an accuracy of one or two orders of magnitude. KIE computations are also sensitive, but they take advantage of error cancellation in the ratio of rates between isotopologues.</p><p>We must point out that even if computational results are not highly accurate, cryogenic experimental tests are also extremely sensitive to the technique (be it gas phase measurements through supersonic expansion, co-deposition with noble gases, solid-state doped complexes, a liquid state in He droplets, or any other low temperature available method). Therefore, a direct comparison between experiment and computation (or between experiment and experiment!) must be done taking all these reproducibility issues into consideration.42,59</p><!><p>Most Cu(ii) systems that experience QMT in solid matrices are oxygen based, where the metal–ligand bond strength is weak and therefore the automerization of JT structures is exceedingly easy. For this study, we sought complexes with stronger bonds, with higher rearrangement activation energies and lower probability of reaction by a classical over-the-barrier mechanism. Nitrogen-based complexes (amines and imines), being relatively strong Lewis bases, proved to be a much better choice than oxygenated systems, even if the barriers are still low (circa 6 to 9 kJ mol−1, see Table 1). Therefore, based on common ligands, we selected six 21 e− Cu(ii) complexes that show a well-defined JTE (Fig. 2). These mono-, bi-, and tri-dentate amine and imine complexes permitted us to explore the variability caused by ligand denticity and N-hybridization.</p><p>We computed the automerization rates for the Cu(ii)–N6 complexes from 4 to 400 K using the CVT semi-classical method, adding the SCT tunnelling correction computed at the benchmarked PBE0/6-31+G(d) level, as described in the Methods section. The distortion was gauged according to the JT radius (RJT, eqn (1)),7 where Δdi is the bond length difference between the average and the i'th M–L bonds,1</p><p>This measure is similar to the effective trajectory of a QMT process,17,60,61 and therefore their low values (combined with the low activation energies, see Table 1), suggest a high QMT probability. Solid-state experimental RJT seems to be shorter than our gas phase values (for example, 0.33 vs. 0.42 Å for Cu(tach)22+) due to the crystal pressure (see above),35 which points to a significantly faster QMT rate for the former. Still, as can be seen in Table 1, even if the results with nitrogen-based ligands in the gas phase are much slower than copper oxides in solid solutions, our computations show that close to the absolute zero Cu(ii)–N6 complexes can undoubtedly rearrange exclusively through QMT from the ground state.</p><p>Rate constant ratios for reactions over and through the barrier can be as high as 100 orders of magnitude at liquid He temperatures (see Cu(timm)22+). The QMT effect at low temperature is evident in the Arrhenius plots of Fig. 3 (see full tables in the ESI†). The most striking case is Cu(biea)2, with a rate constant of 2 × 104 s−1 (half-life of 34 μs). With this pincer ligand the geometry is highly constrained, lowering the RJT and enhancing the QMT (Cu(tach)22+ produces slightly faster tunnelling helped by the significantly lower barrier).</p><p>In all the studied systems SCT shows a negligible tunnelling contribution at T ≳ 75 K. This qualitatively matches the calculated crossover temperatures (see the ESI†)62–65 and explains why there is a negligible QMT in solid state CuII–N6 complexes at liquid N2 and higher temperatures37–40 (although there is a considerable difference between these studies and our gas phase computations). Below ∼75 K there is a growing influence of thermally activated tunnelling (that is, QMT from vibrationally excited states). For most systems, below ∼20 K the reaction is exclusively driven by tunnelling from the ground state.</p><p>However, Cu(NH3)62+ still shows signs of thermally activated tunnelling at an extremely low value of ∼4 K. We believe that this is caused by the almost free rotation of the amines, creating a virtually continuous band of vibrational states (in reality, hindered-rotational states) which enables the occupation of excited states even close to the absolute zero. This generates an interesting KIE profile, as we shall see below.</p><p>For comparison, Cu(H2O)62+, which as explained before must have an enormous probability of tunnelling due to its low barrier (ΔE‡ = 1.3 kJ mol−1, with RJT = 0.31 Å), has computed gas phase rates of kSCT = 2 × 1010 s−1 and kCVT = 5 × 10−8 s−1 at 4 K; this explains why the QMT was, supposedly, never observed for N-ligands, but easily seen in oxides. Still, regardless of the slower rates, all our species react well within what can be considered as "laboratory observable time". They are too slow to be observed by cryogenic EPR (a method with timescales of the order of nanoseconds), but they might be observable by peak coalescence of exchanging atoms in cryogenic solid-state NMR,68 which has a much slower timescale. For instance, using the methodology described elsewhere,16,17,61 with a 500 MHz equipment (50.6 MHz for N) and with computed Δδ values of 28, 35 and 2.9 for the central nitrogens, lateral nitrogens and lateral hydrogens of Cu(biea)22+, respectively, we obtain coalescence rate constants of 3000–4000 s−1. In this situation, we predict merged NMR peaks for this complex at any temperature, while in the absence of QMT we would see two peaks below ∼30 K.</p><p>Yet, the QMT rearrangement on one Cu–N6 complex might have been detected by EPR four decades ago, even if the authors of the study did not realize it (they were probably not aware of the QMT mechanism, as the idea was in its infancy). The rate of the solid-state dynamic69 JT automerization of Cu(en)3SO4 was measured by Bertini et al.34 from the temperature dependence of the EPR line width and Hudson's equation.70 The obtained Arrhenius plot was found to be acceptably linear, from which the activation energy was calculated (Ea = 1.22 kJ mol−1). We redraw here their original data on a different scale (inset in Fig. 3, see also Table S7 in the ESI†), which highlights the concavity of the plot, a signature of thermally activated QMT. If we consider only the first six points of the graph (the ones with apparent linearity), we can see that the steepest slope produces a higher Ea of 1.91 kJ mol−1 (our computed gas phase value is much higher – Ea = 6.1 kJ mol−1 – depicting the differences between the methods and conditions).</p><p>Due to the computational cost to obtain accurate SCT computations, we could not test larger ligands. However, it is possible to artificially set heavier atoms in simple complexes to model the ligand size effect. For that, we studied the Cu(NH3)62+ system changing the hydrogen masses from 1 to 2, 4, 8 and 16 u, equivalent to ligands' masses of 17, 20, 26, 38 and 62 u per NH3 (for comparison, imidazole, the ligand originally supposed to tunnel in solid state,36–40 has a mass of 68 u). The results are clear: at 4 K the rates are 9 × 10−2, 8 × 10−3, 4 × 10−5, 5 × 10−9 and 4 × 10−15 s−1, showing the difficulties of tunnelling if large ligands are attached (see Table S6 in the ESI†). In the case of imidazoles, it is still possible that some QMT can occur in solid state due to the crystal constraints, but in the gas phase it would be completely impossible. It is worth mentioning that with chelating ligands, like most of our cases, the effective moving mass is relatively light, as many atoms in the framework have almost negligible movement.</p><p>KIE analyses were carried out by replacing all 14N isotopes with their heavier 15N to assess the outcome of a possible experimental test. As can be seen in Fig. 4 (also Tables 1 and S8 in the ESI†), the high KIE plateau at low temperature clearly indicates ground state tunnelling.11,71</p><p>The exception is Cu(NH3)62+ which, as previously discussed, does not easily converge into the temperature independence range. The 14N/15N KIE grows first due to conventional ZPE differences and then in a much stepper way due to thermally activated tunnelling. However, below ∼10 K instead of stabilizing in a plateau, like all the other systems, it continues growing (Fig. 4 and S1 in the ESI†), due to the almost free rotation of the ammonia groups, as explained above.</p><p>Considering this, we also calculated the H/D KIE (all hydrogens substituted), as the higher mass of the deuterium should hinder the free NH3 rotation. The results were unanticipated and possibly unique, as the H/D KIE decreases below 10 K (Fig. S1 in the ESI†). This strange behaviour is caused by the ND3 system converging to QMT from the ground state at higher temperatures compared to NH3. Noteworthily, computing accurate properties from such flat potentials (especially conformational surfaces that are taken as vibrations) is problematic, and therefore we can only take this observation as a hypothesis more than a real prediction.</p><!><p>Significant heavy atom tunnelling was theoretically proven to occur in the gas-phase degenerate rearrangement of Jahn–Teller distorted Cu(ii)–N6 complexes. While similar QMT has been experimentally observed in Cu2+ oxide solids due to their much lower barriers, the evidence in common nitrogen-based complexes was found to be more elusive. Herein we show that easily synthesizable mono-, bi- and tri-dentate amine and imine ligands can react by tunnelling under cryogenic conditions, although at rates that are hard to detect by standard experimental tests. Surprisingly (and unbeknown to the authors), a solid state EPR experimental manifestation of our theoretical gas phase results apparently emerged 40 years ago.</p><p>Nitrogen KIE analysis on all the tested complexes revealed a large KIE, with an unexpected behaviour of Cu(NH3)62+, apparently due to free rotation of the ammonia groups. We plan to synthesize some of the Cu(ii)–N6 complexes to test them by EPR characterization.</p><!><p>There are no conflicts to declare.</p>

PubMed Open Access

Identification of pH-sensing Sites in the Light Harvesting Complex Stress-related 3 Protein Essential for Triggering Non-photochemical Quenching in Chlamydomonas reinhardtii*

Light harvesting complex stress-related 3 (LHCSR3) is the protein essential for photoprotective excess energy dissipation (non-photochemical quenching, NPQ) in the model green alga Chlamydomonas reinhardtii. Activation of NPQ requires low pH in the thylakoid lumen, which is induced in excess light conditions and sensed by lumen-exposed acidic residues. In this work we have used site-specific mutagenesis in vivo and in vitro for identification of the residues in LHCSR3 that are responsible for sensing lumen pH. Lumen-exposed protonatable residues, aspartate and glutamate, were mutated to asparagine and glutamine, respectively. By expression in a mutant lacking all LHCSR isoforms, residues Asp117, Glu221, and Glu224 were shown to be essential for LHCSR3-dependent NPQ induction in C. reinhardtii. Analysis of recombinant proteins carrying the same mutations refolded in vitro with pigments showed that the capacity of responding to low pH by decreasing the fluorescence lifetime, present in the wild-type protein, was lost. Consistent with a role in pH sensing, the mutations led to a substantial reduction in binding the NPQ inhibitor dicyclohexylcarbodiimide.

identification_of_ph-sensing_sites_in_the_light_harvesting_complex_stress-related_3_protein_essentia

Introduction<!>LHCSR3 Structure Modeling<!>Site-directed Mutagenesis of Acidic Residues<!>Transformation and Isolation of Site-directed Mutants<!>Recombinant Protein Overexpression, Purification, and in Vitro Refolding<!>Pigment Analysis<!>SDS-PAGE, Coomassie Staining, DCCD Binding, and Western Blot<!>Fluorescence Lifetime Measurements<!>Steady-state Absorption, Fluorescence, and Circular Dichroism Measurements<!>Dynamic Light Scattering Measurements<!>Structural Model of LHCSR3<!><!>Identification of the LHCSR3 Protonatable Sites Involved in Lumenal pH Sensing<!><!>In Vivo Mutation Analysis<!><!>In Vivo Mutation Analysis<!><!>In Vitro Reconstitution of LHCSR3 Recombinant Protein Mutated on Protonatable Sites<!><!>In Vitro Reconstitution of LHCSR3 Recombinant Protein Mutated on Protonatable Sites<!><!>[14C]DCCD Binding of Recombinant LHCSR3 Proteins<!><!>Fluorescence Lifetimes of Recombinant LHCSR3 Proteins<!><!>Fluorescence Lifetimes of Recombinant LHCSR3 Proteins<!>Fluorescence Emission Spectra at 77 K<!><!>Discussion<!>Author Contributions<!>

<p>Photosynthetic organisms convert sunlight absorbed by chlorophyll into chemical energy by reducing CO2 into sugars with electrons extracted from water, yielding O2. However, molecular oxygen can react with chlorophyll triplets (3Chl*)6 to yield singlet oxygen, one of several types of reactive oxygen species, which damage biological molecules (1). Because 3Chl* originates from 1Chl*, prevention of photooxidative stress can be obtained by quenching 3Chl*, scavenging reactive oxygen species, or by quenching 1Chl*. In addition, the photon absorption rate can be regulated by chloroplast relocation within the cell or changes in leaf orientation (1–3).</p><p>Of particular importance are the non-photochemical quenching (NPQ) mechanisms that quench 1Chl* and dissipate excess excitation energy as heat when light absorption exceeds the capacity of photochemical reactions. NPQ includes several components, the major and fastest of which is energy-dependent quenching (qE), which is sensitive to uncouplers (4, 5). qE is a feedback process triggered by thylakoid lumen acidification (4, 6–10). Saturation of downstream reactions leads to depletion of ADP and Pi, the substrates of ATPase, which prevents the efflux of protons generated by photosynthetic electron transport from the thylakoid lumen to the stroma, leading to lumen acidification.</p><p>Genetic analysis of qE activation led to the identification of PSBS and LHCSR (11–15) as gene products required for qE in the model plant Arabidopsis thaliana and the green alga Chlamydomonas reinhardtii, respectively (12, 13, 16–18). In C. reinhardtii, two LHCSR isoforms, LHCSR1 and LHCSR3, are active in qE. LHCSR3 strongly accumulates in excess light, whereas LHCSR1 is constitutively present even at low light levels (13, 19). Also, LHCSR-like proteins with qE activity have been identified in diatoms (20–23). A special case is found in mosses, where both PSBS and LHCSR proteins are present and involved in qE induction (15, 24, 25). Although the fundamental mechanisms of quenching activity by PSBS and LHCSR are the subjects of intense investigation (26), they must be different because LHCSR is a chlorophyll- and xanthophyll-binding protein where quenching of 1Chl* can be catalyzed as shown by its short fluorescence lifetime (18). In contrast, pigment-binding sites are not conserved in PSBS, suggesting that the quenching activity is elicited within interacting proteins (18, 25, 27). As for the capacity for sensing the lumenal pH, PSBS and LHCSR share the property of binding dicyclohexylcarbodiimide (DCCD), a protein-modifying agent that covalently binds to acidic residues involved in reversible protonation events (28). Indeed, we have previously shown that two glutamate residues in PSBS are responsible for both the DCCD binding in vitro and the NPQ activity in vivo (17). Sequence analysis of LHCSR proteins showed multiple conserved acidic residues exposed to the lumen as potential sites of protonation. Recombinant LHCSR3 from C. reinhardtii has been shown to be pH responsive and to undergo a switch to a dissipative state in acidic solution (18, 27). Mutation analysis has located eight putative pH-sensing residues in the C terminus of LHCSR3 (27), whereas PSII supercomplexes containing LHCSR3 with a stoichiometry LHCSR3: PSII of 0.28 were reported to undergo a decrease in fluorescence lifetime when exposed to pH 5 (29).</p><p>Here, we have performed a detailed investigation of the pH-sensing activity in LHCSR3 from C. reinhardtii, including identification of lumen-exposed protonatable residues that have been mutated to non-protonatable ones. The effect of these mutations has been analyzed both by fluorescence lifetime analysis of the proteins refolded in vitro and by measuring NPQ activity in vivo upon expression in a mutant lacking both LHCSR3 and LHCSR1. This comprehensive procedure led to the identification of three residues that are crucial for pH-dependent quenching in vivo and in vitro and are also responsible, to a large extent, for the binding of the qE inhibitor DCCD.</p><!><p>LHCSR3 protein structures were obtained using homology modeling techniques with the on-line servers I-TASSER (30, 31) version 1.1. The model with the best C-score (confidence score) was selected for further analysis.</p><!><p>The LHCSR3.1 genomic clone plasmid LHCSR3/GwypBC1 from previous complementation experiments (13) was used for site-directed mutagenesis of each acidic residue reported in the text. The QuikChange® Site-directed Mutagenesis Kit was used according to the manufacturer's instructions.</p><!><p>The plasmid LHCSR3/GwypBC1 containing each or multiple mutations was transformed into either npq4 (13) or npq4 lhcsr1 (32) and transformants were selected for paromomycin resistance. At least 300 colonies were picked for each line and patched onto HS minimal medium to grow in high light (400 μmol of photons m−2 s−1). NPQ was measured by chlorophyll fluorescence video imaging (Imaging-PAM, Walz). Selected colonies, as judged by their NPQ value relative to the parent strain, were further cultured in liquid HS to measure via a pulse-amplitude-modulated fluorometer (FMS2, Hansatech) and for immunoblots, as previously described (13).</p><!><p>LHCSR3 coding sequence was cloned in pET28 expression vector and expressed in Escherichia coli as previously described (18). Purified apoprotein was refolded in vitro in the presence of pigments as reported in Refs. 18 and 33.</p><!><p>Pigments bound by recombinant LHCSR3 were measured as described in Ref. 34.</p><!><p>SDS-PAGE was performed as reported in Ref. 18. SDS-PAGE gel was then stained with Coomassie-R as described in Ref. 35. DCCD binding properties of recombinant LHCSR3 proteins were estimated by incubating the refolded protein with [14C]DCCD and subsequent autoradiography evaluation of the binding as previously described (18, 28, 36). Western blot analysis was performed as described in Ref. 13.</p><!><p>Fluorescence decay kinetics were measured on recombinant LHCSR3 protein using a time-correlated Single Photon Counting apparatus similar to the one described in Ref. 10. 150-fs pulses centered at 820 nm are generated by a Ti:Sapphire oscillator (Coherent Mira 900F) at 76 MHz repetition rate. The pulses are frequency-doubled in a 1-mm thick BBO crystal and their repetition rate is reduced by a factor of 8 with a pulse picker (Spectra Physics model 3980). The resulting pulses are centered at 410 nm, with 12-nm bandwidth full width half-maximum, and with an energy at sample of ∼10 pJ/pulse. The fluorescence emitted by the sample passes through a polarizer set at magic angle, followed by either a monochromator (Horiba Jobin-Ivon H-20) or a long-pass filter. The detection system is composed of a MCP/PMT detector (Hamamtsu R3809U), electrically cooled to −30 °C. The detector is connected to a PC computer with a DCC-100 detector control card (Becker-Hickl). The full width half-maximum of the instrument response function is measured to be 45–55 ps. The samples are held in 1- or 2-mm thick quartz cuvettes (Starna Cells), and kept at ∼12 °C during the measurements with a home-built nitrogen cooling system.</p><!><p>Room temperature absorption spectra were recorded using an SLM-Aminco DK2000 spectrophotometer, in 10 mm HEPES, pH 7.5, 0.2 m sucrose, and 0.03% n-dodecyl-α-d-maltopyranoside. The wavelength sampling step was 0.4 nm. Fluorescence emission spectra were measured using a Jobin-Yvon Fluoromax-3 device. Circular dichroism (CD) spectra were measured at 10 °C on a Jasco 600 spectropolarimeter using a R7400U-20 photomultiplier tube: samples were in the same solution described for the absorption with an OD of 1 at the maximum in the Qy transition. The measurements were performed in a 1-cm cuvette. Denaturation temperature measurements were performed by following the decay of the CD signal at 682 nm when increasing the temperature from 20 to 80 °C with a time slope of 1 °C/min and a resolution of 0.2 °C. The thermal stability of the samples was determined by finding the T½ of the signal decay.</p><!><p>The size of aggregates induced by detergent dilution was determined by dynamic light scattering using ZETASIZER NANO S instrumentation as described in Refs. 37 and 38.</p><!><p>A model of LHCSR3 (Fig. 1) was created using as a template the three-dimensional structure of other LHC proteins, LHCII, CP29, and LHCI (39–41), with the aim of identifying protonatable residues exposed to the thylakoid lumen. The analysis of the protein model is consistent with LHCSR3 conserving the three trans-membrane α-helices (helix A, B, and C) and two amphipathic helices (helix D and E) revealed from crystallographic analysis of CP29 and LHCII; a short additional helix was predicted at the C-terminal domain, significantly more extended than in LHCII or CP29. LHCSR3 bears several acidic amino acid residues (i.e. aspartate and glutamate) predicted to face the thylakoid lumen. In particular, the C-terminal domain contains eight acidic residues: Glu231, Glu233, Glu237, Asp239, Asp240, Glu242, Asp244, and Asp254, as described in a previous report (27). Additional residues exposed to the lumen include Glu221, Glu224 at the lumenal end of Helix D, the residues Asp109 and Asp117 in the loop between Helix B and Helix E, and residue Glu218, located in the loop connecting Helix A and Helix D. To evaluate the accessibility of these residues to the solvent, the LHCSR3 protein sequence was analyzed by I-TASSER software (30, 31), which ranks the probability of a residue to be solvent-exposed within a range of 1 to 9, with a higher number indicating a higher probability. As reported in Table 1, the values obtained for most of the selected glutamates and aspartates were in the range of 3–4 with the exception of Glu233 and Glu237 (scoring 2) and Asp254 (scoring 6).</p><!><p>Three-dimensional model of LHCSR3. Panel A, LHCSR3 structure modeled on LHCII and CP29 crystallographic structures. Putative protonatable sites are indicated. Panel B, zoom view on Asp117, Glu221, and Glu224 residues; the distance between the different residues is indicated in yellow (Å).</p><p>Evaluation of the accessibility to the solvent of LHCSR3 lumen-exposed residues</p><p>Solvent accessibility was predicted using I-TASSER software which scores each protein residue with a 0–9 figure with high score indicating higher probability for solvent exposure. Lumen-exposed regions are reported together with the solvent-accessibility score for each residue shown below. Amino acid position, within the LHCSR3 sequence is indicated above.</p><!><p>The evaluation of the conservation of specific residues in sequences from different species can facilitate the identification of residues that are crucial for protein function. When the LHCSR3 protein sequence from C. reinhardtii was compared with homologous sequences from other organisms, either microalgae or mosses (Table 2), several residues appeared to be highly conserved (Fig. 2A). Most of the chlorophyll-binding sites previously identified in LHCSR3 from C. reinhardtii can also be found in LHCSR proteins from other species, namely the residues binding chlorophyll at the A1, A4, and A5 sites, according to the nomenclature previously used for LHCII chlorophyll binding sites (18, 42). However, some variability can be found for the A2, B5, and A3 sites. Finally, a residue at the position of binding site B6 (glutamate) was found only in a few sequences. As for the lumen-exposed residues from C. reinhardtii, their conservation is far from complete among the different sequences. In particular, as reported in Fig. 2B, the C-terminal domain, which was reported to be the knob of a dimmer switch to control the transition to a dissipative state (27), is present only in Chlamydomonas species, Volvox carteri, and Aureococcus anophagefferens, whereas the number of acidic residues within this domain is variable in the different organisms (Fig. 2B). LHCSR-like sequences from Ostreococcus tauri, Ostreococcus lucimarinus, and Chlorella variabilis show a single protonatable glutamate residue in the C terminus. Moreover, this domain was significantly shorter in the remaining sequences. In contrast, the residues in Helix E, Glu221 or Glu224 in LHCSR3 from C. reinhardtii, were conserved in 18 of 26 sequences analyzed. As for residue Glu218, this is only found in C. reinhardtii, Chlamydomonas moewusii, and V. carteri, whereas in Chlamydomonas sp. ICE a conservative replacement to aspartate was identified. Nevertheless, a glutamate was found very close to this position and shifted toward the N terminus in all the other accessions, suggesting it might have a conserved functional role. The analysis of the conservation of residues Asp109 and Asp117 showed that Asp109 is present in 9 accessions, whereas Asp117 is present in 17 accessions as aspartate, or replaced by glutamate in sequences from Ulva linza and Ulva prolifera. It is worth noting, however, that all the accessions that do not bear Asp117, do have one or more aspartate residues within 1–3 positions, the only exception being the sequence from Mesostigma viride lacking aspartates or glutamates in that protein domain. On the basis of these results, Asp109, Asp117, Glu218, Glu221, Glu224, Glu231, and Glu233 were selected for further investigation and renamed, respectively, D1, D2, E3, E1, E2, E4, and E5 for simplicity.</p><!><p>LHCSR-like protein sequences used for the determination of conserved residues</p><p>Protein sequences were selected by BLAST search using LHCSR3 mature protein sequence as query. Each sequence was selected for having a score >150 and e-value <6 e−41.</p><p>Alignment of LHCSR-like protein sequences.</p><!><p>To test the importance of these protonatable residues for LHCSR3 function, each of the selected aspartate or glutamate residues was mutagenized in vitro to asparagine or glutamine, respectively, and the resulting mutant LHCSR3 genomic DNA sequence was used to transform the npq4 mutant strain lacking LHCSR3 expression (13). After selection of transformants based on paromomycin resistance, these were screened to determine the effect of the amino acid replacement on the NPQ activity. Fig. 3A shows that when each individual acidic residue was mutated, the NPQ amplitude was reduced, but a significant level of quenching was still present. Indeed, the transformant lines exhibited an NPQ level proportional to the level of LHCSR3 protein accumulation as assessed by Western blotting (Fig. 3B), suggesting redundancy of the proton-sensing residues in LHCSR3. Strains transformed with LHCSR3 variants mutated at residue D1 did not show any accumulation of LHCSR3, suggesting a major role of this residue in stabilizing protein folding. Therefore, combinations of multiple mutations within the same protein were generated and tested. When D2, E1, and E2 residues were mutated together, the triple mutant had an NPQ amplitude similar to that of the npq4 strain. Indeed, of more than 300 colonies of the triple mutant D2E1E2 in the npq4 background, none had higher NPQ than npq4, despite the accumulation of the mutant LHCSR3 protein at wild-type level.</p><!><p>NPQ measurements and immunoblot analysis of LHCSR3 protein levels in npq4 lines expressing site-specific mutant versions of LHCSR3 affecting protonatable sites. Panel A, NPQ measurements on WT, npq4 mutant, and transgenic lines with LHCSR3 proteins carrying a single mutation on putative protonatable sites D2, E1–5. Panel B, immunoblot analysis of LHCSR3 accumulation on genotypes analyzed in panel A; immunoblot analysis of the D1 subunit of PSII is shown as a control for loading. Panel C, NPQ measurements on WT, npq4 lhcsr1 mutant, and transgenic lines with LHCSR3 proteins with double mutations on putative protonatable sites D2E2 and E1E2. Panel D, immunoblot analysis of LHCSR3 accumulation on genotypes analyzed in panel C. In all cases three independent biological replicates were analyzed. The experiments were reproduced two times.</p><!><p>To improve the signal to background ratio in the NPQ assays, subsequent transformations were done with the npq4 lhcsr1 double mutant (32). This system would allow for better resolution of the effect that the mutated LHCSR3 protein has on NPQ, independent of the LHCSR1 isoform that remains in the npq4 mutant. Combinations of the double mutations, D2E2 and E1E2, were made and transformed into npq4 lhcsr1 (Fig. 3, C and D). These double mutations impaired but did not completely eliminate the qE function of LHCSR3, because the expressed LHCSR3 protein still conferred some NPQ in the npq4 lhcsr1 background (Fig. 3, C and D). Lines of the E1E2 mutant accumulating LHCSR3 at 50–60% with respect to the wild-type had ∼20% of wild-type qE (Fig. 3C). Two lines from the D2E2 transformations with more than wild-type LHCSR3, had only ∼30 and ∼50% of the qE found in wild-type, respectively (Fig. 3C). The D2E1E2 triple mutant version of LHCSR3 was then transformed into the npq4 lhcsr1 genotype. As shown in Fig. 4, two independent lines expressed the D2E1E2 mutant protein at a level close to the wild-type LHCSR3 protein level, but they exhibited the lowest qE activity observed (∼25% of the wild-type level). The residual qE induction observed in the triple D2E1E2 mutant could be related to the activity of one or more of the other protonatable sites that are still present in the mutant. Unfortunately, the addition of further mutations in the LHCSR3 gene resulted in a loss of protein accumulation, suggesting a strong destabilization of the protein or some impairment in protein import into the thylakoid membranes.</p><!><p>NPQ measurements and immunoblot analysis of LHCSR3 protein levels in npq4 lhcsr1 lines expressing site-specific mutant versions of LHCSR3 affecting protonatable sites. NPQ measurements on WT, npq4 lhcsr1 mutant, and transgenic lines with LHCSR3 proteins mutated on D2, E1, and E2 protonatable sites (panel A). Panel B, immunoblot analysis of LHCSR3 accumulation; immunoblot analysis of the D1 subunit of PSII is shown as a control for loading. In all cases three independent biological replicates were analyzed. The experiments were reproduced three times.</p><!><p>The function of the three identified protonatable sites in LHCSR3 was next investigated in vitro using recombinant proteins. In particular, the LHCSR3 cDNA sequence was subjected to site-specific mutagenesis of D2, E1, and E2, as previously described (43). Following expression in E. coli and purification, the wild-type (WT) and mutant D2E1E2 apoproteins were refolded in vitro in the presence of chlorophylls and carotenoids (18, 27, 33). As shown in Table 3, the holoproteins were characterized by HPLC pigment analysis. In both cases the Chl a/b ratio was higher than 8, with a very small amount of Chl b per apoprotein compared with Chl a. The Chl/Car ratio was also similar in WT and the D2E1E2 mutant with 2 Car molecules per 7 Chls bound in LHCSR3. The carotenoids bound by reconstituted samples were mainly violaxanthin and lutein in agreement with previous reports (18, 27).</p><!><p>Pigment analysis of recombinant LHCSR3 proteins</p><p>Pigment analysis were performed by HPLC and fitting of absorption spectrum of pigment acetone extracts with chlorophylls and carotenoids spectral forms as described in Ref. 34. The experiments were performed two times, with three independent biological replicates each time.</p><!><p>The efficiency of energy transfer between pigments was investigated by recording fluorescence emission spectra at room temperature upon selective excitation of Chl a, Chl b, and xanthophylls, showing no differences between WT and D2E1E2 proteins. The absorption spectra of WT and D2E1E2 in the visible region (Fig. 5A) did not show significant differences, neither in the Soret nor in the Qy spectral regions, suggesting that the mutations introduced no changes in the pigment organization in the complex. Similarly, circular dichroism spectra of the two samples were virtually identical (Fig. 5B). To assess if the mutations introduced could induce some level of protein destabilization, the thermal stability of recombinant WT and D2E1E2 mutant proteins was measured by following the change of the amplitude of the CD signal at 681 nm when slowly increasing the temperature of the samples. The melting temperatures (Tm), calculated by fitting to a sigmoidal function, are reported in Table 4: Tm was similar for WT and the D2E1E2 mutant, 41.8 and 40.2 °C, respectively, suggesting that substitution of the three acidic residues in the D2E1E2 mutant did not alter the stability of the pigment-protein complex.</p><!><p>Absorption and circular dichroism spectra of LHCSR recombinant proteins. Absorption spectra (panel A) and circular dichroism (panel B) in the visible region of LHCSR3 WT and D2E1E2 mutant refolded in vitro in the presence of chlorophylls and carotenoids. The experiments were reproduced three times, each time with two independent biological replicates.</p><p>Thermal stability of LHCSR WT and D2E1E2 recombinant proteins refolded in vitro</p><p>Thermal stability (Tm) was evaluated following the decay of the CD signal at 682 nm when increasing the temperature from 20 to 80 °C. The thermal stability of the samples was determined by finding the T½ of the signal decay. The experiments were performed two times, with two independent biological replicates each time.</p><!><p>DCCD is an inhibitor of qE in Chlamydomonas (18). Its binding to acidic residues indicates reversible protonation events. An enhanced DCCD binding with respect to other LHC proteins has been reported for LHCSR3 (18), in agreement with its pH-sensing function. To assess the proton-binding activity of the Asp117, Glu221, and Glu224 residues, DCCD binding was measured in WT and D2E1E2 mutant proteins. In vitro refolded proteins were incubated with [14C]DCCD, and the amount of 14C bound by LHCSR3 was determined by autoradiography. The level of 14C bound by LHCSR3 WT and D2E1E2 was then normalized to the protein amount loaded into the SDS-PAGE gel quantified by Coomassie staining (Fig. 6). As reported in Fig. 6, both WT and D2E1E2 mutant bound [14C]DCCD, but binding to the D2E1E2 mutant was decreased by 40% with respect to WT. This result supports the hypothesis of multiple protonatable sites in LHCSR3, of which Asp117, Glu221, and Glu224 account for at least 40% of the DCCD-binding activity of this protein.</p><!><p>14DCCD binding in LHCSR3 recombinant WT and D2E1E2 mutant. Panel A, autoradiography of recombinant LHCSR3 WT and the D2E1E2 mutant treated with 14DCCD; microliters of sample (0.2 μg/μl of chlorophylls) loaded on SDS-PAGE are reported (15, 7.5, and 2.5 μl). Panel B, Coomassie staining of SDS-PAGE used for autoradiography. Panel C, ratio of the level of 14C observed by autoradiography signals and protein quantity obtained by densitometric analysis of Coomassie-stained gels. The experiments were performed two times; each time two independent biological replicates were analyzed with three technical replicates with different loading volume as indicated.</p><!><p>Fluorescence lifetime measurements on recombinant LHC proteins allow investigation of their excitation energy conserving versus quenching properties (44). To investigate in vitro the pH-dependent regulation of LHCSR3 quenching activity, fluorescence decay kinetics of WT LHCSR3 and the D2E1E2 mutant were measured using a single photon counting device at neutral pH (7.5) and at low pH (5) in detergent solution of 0.03% n-dodecyl-α-d-maltopyranoside (α-DM). As reported in Fig. 7A, the fluorescence decay kinetics of WT and the D2E1E2 mutant can be satisfactorily fitted with three exponentials with associated time constants of 4 ns, 1.9 ns, and ∼200 ps. The relative amplitudes were 38–44, 48–52, and 7.4–9.7%, respectively, with an average lifetime of 2.6–2.7 ns (Table 5). Decays were similar at both pH 7.5 and 5, suggesting no pH-dependent response of quenching reactions.</p><!><p>Fluorescence decay kinetics. Fluorescence decay kinetics of recombinant LHCSR3 WT (panel A) and D2E1E2 mutant (panel B) at pH 7.5 or 5.0 in the presence of high (0.03%) or low (0.003%) detergent (α-DM) concentrations. The experiment was performed two times, each time with two independent biological replicates.</p><p>Fluorescence lifetimes of LHCSR3 WT and D2E1E2 mutant</p><p>The decay traces reported at Fig. 7 were fitting using three exponentials functions. The amplitude (A) and time constants (τ) for each exponential are reported in the table. The average lifetimes for each sample are calculated as ΣAiτi.</p><!><p>This result is in agreement with a previous report on LHCSR3 fluorescence lifetime in detergent (27), suggesting that interaction of detergent micelles with the protein prevents the switch to a dissipative conformation. pH sensitivity of the LHCSR fluorescence lifetime can be better detected at a low detergent/protein ratio leading to moderate aggregation, which reproduces protein-protein interactions occurring in the protein-crowded thylakoid membrane (45). Fig. 7B shows the fluorescence lifetimes of recombinant WT LHCSR3 and the D2E1E2 mutant as measured upon incubation in a detergent concentration of 0.003% α-DM. These measuring conditions induced a faster decay of emitted fluorescence at either pH 5 or 7.5 for both LHCSR3 WT and D2E1E2 mutant. However, whereas at pH 7.5 the two proteins showed the same decay profile, at pH 5 LHCSR3 WT fluorescence decay was much faster than at pH 7.5, whereas the fluorescence decay of LHCSR3 D2E1E2 was the same as at pH 7.5. Decays of LHCSR3 WT at pH 7.5, and D2E1E2 at pH 7.5 and 5 were fitted to three exponentials with time constants of 2.5 ns, 0.9 ns, and 140 ps with amplitudes of 40, 31, and 29%, respectively, with an average lifetime of ∼1.4 ns (Table 5). LHCSR3 WT decay traces at pH 5 were similarly fitted to three exponentials, but in this case the major amplitude was associated to the fastest component (140 ps) with amplitude of 42%, whereas those with 0.8 and 2.5 ns showed amplitudes of 33 and 25%, respectively (Table 5). Because aggregation is well known to influence the lifetime of LHC proteins (46), the aggregation size of WT and D2E1E2 proteins at 0.03 and 0.003% DM was measured by dynamic light scattering as previously reported (38), yielding the average aggregate size (113.3 ± 8.6 in the case of WT and 126.9 ± 18.2). LHCSR3 WT and D2E1E2 in low detergent condition at pH 5 formed aggregates with similar size with a radius of 100 nm, implying that the difference in fluorescence quenching observed between WT and D2E1E2 is likely due to the different dissipative conformations that can be reached by the two proteins.</p><!><p>Activation of quenching mechanisms has been previously associated in vivo with induction of far-red fluorescence emission forms at 77 K (38, 47, 48). In particular, aggregation-dependent quenching in LHC proteins at low pH was shown to lead to far-red emission in both trimeric and monomeric isoforms, and this feature has been correlated with the extent of excitation energy quenching (38). The fluorescence emission spectra at 77 K of LHCSR3 WT and D2E1E2 recombinant proteins were measured to investigate the correlation of far-red emission forms with the activation of quenching mechanisms. The measurements were performed at high (0.03%) or low (0.003%) detergent concentrations and at pH 7.5 or 5.0. As reported in Fig. 8, fluorescence emission spectra were almost identical at high detergent conditions for both WT and D2E1E2 samples at pH 7.5 or 5 with a peak at 682 nm. At low detergent, instead, a clear shift of the emission peak to 685 or 687 nm was observed for D2E1E2 and WT, respectively. The most evident change in the spectra, however, was observed at low detergent concentrations and pH 5, where both WT and D2E1E2 dramatically increased their far-red emission forms, with the formation of a defined peak at 735 nm that was far more intense in WT versus D2E1E2. These results support the presence of a positive relationship between protonation of specific residues, the appearance of far-red emission forms in the spectra, and the activation of quenching mechanisms in LHCSR3.</p><!><p>77 K fluorescence emission spectra. Fluorescence emission spectra of recombinant LHCSR3 WT (panel A) and the D2E1E2 mutant (panel B) at 77 K measured at pH 5 or 7.5 in the presence of 0.03% α-DM or 0.003% α-DM. The experiments were reproduced three times, each time with two independent biological replicates.</p><!><p>All oxygenic photosynthetic organisms are endowed with mechanisms for thermal dissipation of excess absorbed light energy. The triggering of these mechanisms can either be controlled directly by light as for the Orange Carotenoid-binding protein of cyanobacteria (49, 50) or by low lumenal pH caused by excess light as in the case of PSBS in plants and LHCSR in unicellular algae (10, 13, 18, 26, 51, 52). LHCSR3 is of particular interest because the quenching and the pH-sensing activities are merged in the same protein subunit (18, 27, 53) making this protein a relatively simple system for the molecular analysis of NPQ. The case of plants is more complex because the pH is sensed by PSBS, whereas quenching occurs in an interacting pigment-binding partner (17, 26). LHCSR3 has been reported to undergo functional changes depending on pH (18, 27, 53). In this work, the function of LHCSR3 as a sensor of lumen pH has been investigated in vivo by site-specific mutagenesis of putative protonatable residues. The LHCSR3 structure was modeled on the basis of LHCII and CP29 structures (39, 40) (Fig. 1) allowing for identification of 13 potentially protonatable aspartate and glutamate residues located within lumen-exposed domains at the C terminus, at Helices D and E, and at the loops between Helices A and D and between Helices B and E. Among these residues, Asp117, Glu218, Glu221, Glu224, Glu231, and Glu233 were selected based on their high conservation among LHCSR-like sequences and were targeted for site-specific mutagenesis and functional analysis in vivo. The complementation of npq4 and npq4 lhcsr1 mutants with sequences carrying mutations affecting these protonatable residues identified Asp117, Glu221, and Glu224 as the key residues for the pH sensitivity of LHCSR3 in vivo. Any single mutation of an acidic amino acid residue failed to yield significant effects on qE. When double mutants affecting two different putative protonatable sites were obtained, a substantial decrease in qE relative to the LHCSR3 protein level could be observed, suggesting a cooperative behavior (Fig. 4). The triple D2E1E2 mutant expressed in npq4 lhcsr1 showed an even greater impairment of function, suggesting that Asp117, Glu221, and Glu224 are key residues for pH sensing in LHCSR3 from C. reinhardtii.</p><p>These results are consistent with the significant decrease observed in DCCD binding to LHCSR3 recombinant proteins mutated on Asp117, Glu221, and Glu224, i.e. a reduction by 41% (Fig. 6). The fact that DCCD can still be bound by the D2E1E2 mutant is not surprising, because structure modeling revealed the presence of 10 additional acidic residues, including 8 glutamate and aspartate residues at the C terminus that are likely to bind DCCD in the D2E1E2 mutant. It is interesting to note that the mutation of 23% of the putative protonatable residues, as in the case of D2E1E2 mutant, led to a 41% reduction in DCCD-binding activity of LHCSR3, suggesting that these residues have a special cooperative role in transducing the lumenal pH signal, as shown by a 72% of reduction in qE in vivo (Fig. 4). The presence of additional glutamate and aspartate residues at the C terminus in LHCSR3 is a peculiar feature of Chlamydomonas spp. LHCSR proteins (Fig. 2), whereas other LHCSR-like proteins have a shorter C terminus extension (Fig. 2B). Asp117, Glu221, and Glu224 are more conserved in the different LHCSR-like sequences analyzed. Nevertheless, it is worth noting that a substantial level of variability is present in the position and number of acidic residues, which might reflect the need for complementarity with interacting proteins putatively involved as partners in qE activity (54, 55). Alternatively, the density of lumen-exposed acidic residues might be related to the response sensitivity for triggering qE, which is particularly strong in Chlamydomonas, for which light saturation of photosynthesis occurs at lower irradiances (56), whereas NPQ amplitude is fully reached already at 200 μmol m−2 s−1 illumination (57).</p><p>It is interesting to compare the distribution of acidic residues in Chlamydomonas to that of Physcomitrella patens LHCSR1, which only harbors 4 of the 13 putative protonatable residues identified in LHCSR3. P. patens LHCSR1 has been shown to require zeaxanthin for a significant level of activity (25), whereas Chlamydomonas NPQ is not dependent on zeaxanthin accumulation (18). Zeaxanthin binding has been shown to confer cooperativity to NPQ in higher plants (58–60). It is tempting to propose that zeaxanthin might replace the effect of the 9 additional lumen-exposed acidic residues in promoting the switch of LHCSR from conservative to dissipative conformations. Indeed, we verified that faster fluorescence decay was triggered by acidic pH in WT LHCSR3 but not in the D2E1E2 mutant, implying that the protonation of Asp117, Glu221, and Glu224 has a special role in triggering quenching events within LHCSR3 in vitro (Fig. 7). A previous report has shown that mutation of 9 acidic residues at the C terminus, including Glu224, led to impaired pH sensitivity of LHCSR3 in vitro (27). Here we show that 72% of qE activity in vivo was dependent on the mutation of only three protonatable residues, consistent with loss of pH responsiveness in vitro. Residual NPQ activity in D2E1E2 is likely due to the presence of several other protonatable residues in the D2E1E2 mutant, partially inducing a small NPQ activation in the triple mutant. By the way it could not be excluded that the protonation of other LHCBM subunits, as LHCBM1 (18), would contribute to pH-dependent triggering of a low NPQ activity in D2E1E2.</p><p>As previously reported, LHCSR3 does not respond significantly to pH variation when the protein is dissolved in detergent such as α-DM, whereas the pH sensitivity becomes evident when detergent is substituted by nano-polymers (27) or decreased to levels below the critical micelle concentration. This latter condition induces the formation of small particle arrays, mimicking protein-protein interactions in the thylakoid membrane (47), a condition likely to also occur in PSII-LHCSR supercomplexes (53). Recent results showed that high detergent conditions favor monodispersion of LHCs and shift their conformation far from the dissipative state toward a state poorly responding to pH variations (45). This is likely due to the induction of a relaxed protein conformation that decreases pigment-pigment interactions within the complexes with respect to the state present in the native membrane environment. Structural analysis suggested that conformational changes involved in quenching are subtle (61, 62) and involve small changes in Chl-Chl and xanthophyll-Chl interactions (11) thus making the relaxed structure unfavorable to trigger quenching. The main effect induced at pH 5 on fluorescence decay kinetics of WT LHCSR3 was an increased amplitude of the 140-ps (τ3) component, which favorably compares with the 65- and 305-ps components recently identified as induced in vivo upon qE activation in C. reinhardtii (10) and the 200-ps component identified when measuring fluorescence lifetimes of LHCSR3-binding PSII supercomplexes at pH 5 (53).</p><p>The mechanism by which LHCSR3 dissipates excitation energy quenching is still debated. High yield of a carotenoid radical cation has been previously reported (18), and formation of these radical species has been previously related to NPQ in plants (63, 64). Here, we present evidence that aggregation-dependent quenching is also active in LHCSR3, as in other LHC proteins (38, 45, 46, 59, 62, 65, 66). Interestingly, a strong red-shift in fluorescence emission was associated with the low pH effect in the LHCSR3 WT, but not in the D2E1E2 mutant (Fig. 5). The formation of these far-red emitting forms is dependent on pH and protonation of Asp117, Glu221, and Glu224. The correlation between far-red emission and switch to a dissipative state has been previously reported for plant LHCII (47), possibly resulting from a strong coupling between chlorophylls (67). It is interesting to note that the only Chl-binding residues fully conserved through LHCSR-like sequences are those associated with sites A1, A4, and A5 (Fig. 2), as putative ligands for Chl 601, Chl 610, and Chl 609 (39). Site A1 has been previously reported to be crucial for protein stability in most LHC proteins (34, 43) acting as a bridge between Helices A and B. Chl-binding sites A4 and A5, instead, are located in proximity of the Car-binding site L2, with a special role in 3Chl* quenching (68). Together with the Chl in site B5, these Chls form a strongly coupled cluster (40, 69), which has been associated to Car radical cation formation (11). Interactions between Chls and between Chls and xanthophylls are likely involved in LHCSR3 quenching activity. The structural model of LHCSR3 in Fig. 1 shows that the three acidic residues Asp117, Glu221, and Glu224 are relatively close to each other, with an estimated distance of 3.69, 7.27, and 8.05 Å, respectively (Asp117-Glu221, Glu221-Glu224, and Asp117-Glu224 in Fig. 1B). The proximity of these residues, shown to be crucial and cooperative in transducing pH sensing, suggests that upon their protonation the overall structure of LHCSR3 might undergo adjustments that reduce the distance/relative orientation between the helices as a result of reduced electrostatic repulsion in the lumen-exposed domain. In particular, a different distance between Helix D and Helix E might be induced by protonation of Asp117, Glu221, and Glu224. This could be transduced into changes in the relative orientation of Helices A and B, with a consequent reorganization of Chl-Chl and Chl-Car interactions. The correspondence between protein aggregation in vitro and NPQ activation in vivo has been previously investigated, showing a similarity between the conformational change induced in vitro by aggregation of LHC proteins and conformational changes observed in vivo upon NPQ induction (14). In addition, we cannot exclude protein aggregation in vitro forces LHCSR3 subunits to establish some peculiar protein-protein interactions required for LHCSR3 activity in vivo. LHCSR proteins can be found as dimers in thylakoid membranes (18, 70), suggesting that possibly some specific protein-protein interactions are needed for LHCSR3 activity. The finding of LHCSR proteins as dimers agrees with recent crystallization of PSBS at low pH in the dimeric state (71), suggesting a possible common strategy for protein activation by formation of homo- or heterodimers and rearrangements of PSII supercomplexes. However, it should be pointed out that whereas LHCSR3 is a chlorophyll- and carotenoid-binding protein, PSBS was reported to bind a single chlorophyll at most. These different pigment-binding properties suggest that whereas LHCSR3 can be a direct quencher of excitation energy located on its pigments, PSBS function is more likely restricted to pH sensing, whereas triggering quenching is activated within interacting LHC proteins. Finally, it is interesting to compare the effects of mutation on protonatable residues in LHCSR3 as compared with PSBS. Recently, the structure at low pH of PSBS from spinach was revealed (71), showing the DCCD binding site at residue Glu173. Previously it was shown indeed that in A. thaliana mutation of two PSBS lumen-exposed glutamate residues, Glu122 (corresponding to Glu173 in PSBS from spinach) and Glu226, yielded complete loss of qE in vivo and DCCD binding in vitro. The effect of individual mutations was additive, with changes to non-protonatable residues at each residue leading to 50% loss in both functions (17). This is clearly not the case for LHCSR, because no effect was observed upon mutation at single residues and even the D2E1E2 mutant still retained 28% of qE and 59% of DCCD binding. Thus, it appears that pH-dependent triggering is far more cooperative in LHCSR3 than in PSBS, with a number of contributing protonation events depending on species that have been reported to differ in the relative contribution by ΔpH and Δψ to the transmembrane pH gradient (72, 73). Also, responsiveness of different species to light intensity and adaptation to specific environments (15, 57) might be tuned by the number and distribution of lumen-exposed protonatable residues in LHCSR. These results are complementary to those recently reported (27) showing that pH responsiveness, as determined by fluorescence lifetime in vitro, was lost by mutation of 9 acidic residues at the C terminus (including Glu224, also studied in the present work) to non-protonatable species. The observed cooperativity between acidic residues might well explain this result.</p><!><p>K. K. N., R. B., and M. B. conceived the work, designed the experiments, and wrote the paper. K. K. N. coordinated the experiments about Chlamydomonas complementation (Figs. 3 and 4), whereas R. B. and G. F. coordinated the experiments in vitro. M. B. and G. R. S. performed all the experiments reported with the exception of Chlamydomonas complementation and mutant screening and characterization. T. B. T. and E. E. performed the work described in Figs. 3 and 4. G. R. F. and E. D. R. contributed to the results reported in Fig. 7. All authors analyzed the results, contributed to writing, and approved the final version of the manuscript.</p><!><p>This work was supported in part by the U.S. Department of Energy, Office of Science, Basic Energy Sciences, Chemical Sciences, Geosciences, and Biosciences Division under field work proposal 449B. The authors declare that they have no conflicts of interest with the contents of this article.</p><p>chlorophyll triplets</p><p>non-photochemical quenching</p><p>n-dodecyl-α-d-maltopyranoside</p><p>carotenoid</p><p>chlorophyll</p><p>energy-dependent quenching</p><p>light harvesting complex stress-related</p><p>dicyclohexylcarbodiimide</p><p>photosystem II subunit S.</p>

PubMed Open Access

TRACE METALS IN GREEN SEA TURTLES (CHELONIA MYDAS) INHABITING TWO SOUTHERN CALIFORNIA COASTAL ESTUARIES

Foraging aggregations of east Pacific green sea turtles (Chelonia mydas) inhabit the Seal Beach National Wildlife Refuge (SBNWR) and San Diego Bay (SDB), two habitats in southern California, USA, located near urbanized areas. Both juvenile and adult green turtles forage in these areas and exhibit high site fidelity, which potentially exposes green turtles to anthropogenic contaminants. We assessed 21 trace metals (TM) bioaccumulated in green turtle scute and red blood cell (RBC) samples collected from SBNWR (n = 16 turtles) and SDB (n = 20 turtles) using acid digestion and inductively coupled plasma mass spectrometry. Principal component analyses of TM composition indicate that SBNWR and SDB turtles have location-specific contaminant signatures, characterized by differences in cadmium and selenium concentrations: SBNWR turtles had significantly more cadmium and selenium in RBC and more selenium in scute samples, than SDB turtles. Cadmium and selenium concentrations in RBC had a strong positive relationship, regardless of location. SBNWR turtles had higher selenium in RBCs than previously measured in other green turtle populations globally. Due to different retention times in blood vs. scute, these results suggest that SBNWR turtles have high long- and short-term selenium exposure. Turtles from SBNWR and SDB had higher trace metal concentrations than documented in green turtle populations that inhabit non-urbanized areas, supporting the hypothesis that coastal cities can increase trace metal exposure to local green turtles. Our study finds evidence that green turtle TM concentrations can differ between urbanized habitats and that long-term monitoring of these green turtles may be necessary.

trace_metals_in_green_sea_turtles_(chelonia_mydas)_inhabiting_two_southern_california_coastal_estuar

Introduction<!>Green Turtle Capture and Sampling<!>Trace Metal Analyses<!>Statistical Analyses<!>Principal Component Analyses<!>Carapace Scute<!>Red Blood Cells<!>Carapace Scute and Red Blood Cell Differences<!>Site-specific element differences<!>Cadmium<!>Selenium<!>Comparisons to other studies<!>Conclusions

<p>Understanding environmental contamination and its effects on threatened species is a critical first step for habitat restoration and conservation of species inhabiting these areas. Sea turtles are one taxon that, because of their coastal existence and proximity to human-altered landscapes, is potentially vulnerable to pollutants (Hamann et al., 2010). By pinpointing which populations are most affected by such anthropogenic stressors, wildlife managers are able to develop well-informed management strategies to protect coastal sea turtle populations and ideally reduce contaminant inputs into these areas.</p><p>East Pacific green sea turtles (Chelonia mydas) are a regional population of green sea turtles (hereby, green turtles) that inhabit the eastern Pacific Ocean from California, U.S. to Chile (Wallace et al., 2010). In recent years, green turtles from the eastern Pacific were reclassified from an endangered to a threatened species, citing improvements in at-sea conservation and recovery of nesting populations (Seminoff et al., 2015). However, egg harvesting, bycatch, pollution, and other human activities still threaten green turtle population recovery in the eastern Pacific (Seminoff et al., 2015). This species exhibits an ontogenetic shift from pelagic habitats as small juveniles to neritic habitats as larger juveniles where they reside and forage for long periods of time until they reach sexual maturity. As adults, the turtles leave the foraging ground periodically for mating and egg-laying at their mating/nesting location before returning back to their foraging ground (Lutz and Musick, 1997). These coastal foraging habitats are critical for the recovery and support of green turtle populations, yet many of these areas have been heavily impacted by urbanization thus presenting health risks with anthropogenic stressors such as pollution from trace metals (Gardner and Oberdorster, 2005; Finlayson et al., 2016).</p><p>Trace metals, often separated into essential and non-essential metals, occur in the environment at low concentrations, but can be artificially elevated via human activities (e.g. dredging, runoff, shipyard activity, etc.) (Pugh and Becker, 2001; Deheyn and Latz, 2006; Dodder et al., 2012; Dodder et al., 2016). Trace metal exposure in many organisms is mainly through diet (Rozman and Klaassen, 2007). Essential trace metals (i.e., selenium and iron), are needed for normal biological functions and can bioaccumulate in the tissues of organisms (Gardner and Oberdorster, 2005; Rozman and Klaassen, 2007). However, when accumulated to high enough levels, essential metals can overwhelm molecular signaling pathways, and alter an organism's ability to compensate for other stressors (Gardner and Oberdorster, 2005; Rozman and Klaassen, 2007). Non-essential trace metals, those not needed for biological functions, can bioaccumulate over time in tissues (e.g., bone or kidney) and cause negative health effects, such as neurological damage or sudden death (i.e., mercury and cadmium) (Rozman and Klaassen, 2007). Long-lived species, such as green turtles, are vulnerable to bioaccumulation of non-essential trace metals due to their long lifespan and extended residence in specific habitats (Finlayson et al., 2016). As a result, studying both essential and non-essential metals could help elucidate and characterize how urban areas affect green turtle contaminant loads.</p><p>Previous research has linked various bioaccumulated trace metals to negative effects on physiology and reproduction in multiple sea turtle species (van de Merwe et al., 2010a; Komoroske et al., 2011; Perrault et al., 2011; Keller et al., 2014a; Finlayson et al., 2016). For example, mercury in adult leatherback turtle (Dermochelys coriacea) blood samples and hatchling liver samples correlated with reduced hatching success (Perrault et al., 2011). However, co-exposure of mercury with selenium was shown to improve hatching success in leatherback turtles relative to mercury alone, suggesting that selenium may help detoxify mercury (Perrault et al., 2011). Another trace metal study using green turtle blood samples found copper and lead to be positively correlated with indicators of oxidative stress (i.e., lower 3-hydroxy-3methylglutaryl-CoA reductase, and increased lipid peroxidation), and green turtles with high oxidative stress were more likely to have fibropapillomatosis, a tumor-bearing disease affecting many green turtle populations worldwide (da Silva et al., 2016). While research on trace metal health effects has been expanding, their impact on sea turtle health remain poorly understood (Finlayson et al., 2016). The above studies indicate that trace metal monitoring in sea turtles is an important component for addressing sea turtle health concerns (Gardner and Oberdorster, 2005; Finlayson et al., 2016). Considering the possible negative effects of high trace metal exposure on sea turtle physiology, assessing the degree of trace metal exposure in green turtles occupying habitats influenced by heavy urbanization, where exposure to trace metals is elevated, is essential for evaluating possible health risks due to anthropogenic activities.</p><p>Along the western coast of the U. S., the northernmost resident foraging aggregations of green turtles inhabit San Diego Bay (SDB) and the Seal Beach National Wildlife Refuge (SBNWR) (Eguchi et al., 2010; MacDonald et al., 2012; Crear et al., 2016). SDB is home to a major U.S. Navy base and subject to other anthropogenic activities such as ship maintenance, dredging, fishing, and boating, which have resulted in the bay being listed (303D) as an impacted body of water (Fairey et al., 1998). SBNWR is located within a naval weapons station and green turtles have been shown to forage in the surrounding area (Crear et al., 2016; Crear et al., 2017). The SBNWR is within 15 km of the Port of Los Angeles, has a naval weapons station, and has input from many impacted waterways and water bodies influenced by industrial and residential runoff (Schiff et al., 2011; Dodder et al., 2016). The different industries and anthropogenic influences in these locations have led to distinct trace metal sediment compositions in SDB versus SBNWR (Dodder et al., 2016). Urbanization, such as ship yard use, dredging and residential and commercial development can continue to change trace metal profiles in these two urban habitats (Schiff et al., 2011; Dodder et al., 2016). Therefore, green turtles inhabiting these habitats may exhibit changes in their trace metal concentrations over time.</p><p>Previous studies on green turtles have shown that SDB green turtles bioaccumulate anthropogenic contaminants such as trace metals (Komoroske et al., 2011). However, the green turtle aggregation in the SBNWR was only recently discovered as a permanent foraging aggregation and has not been previously studied for trace metals accumulation levels (Crear et al., 2016). Considering the different sediment trace metal profiles (Dodder et al., 2016), the high site fidelity green turtles exhibit (Eguchi et al., 2010; Crear et al., 2016), and previous research demonstrating green turtle trace metal bioaccumulation from urbanized habitats (Komoroske et al., 2011), we hypothesized that green turtles from the SBNWR and SDB would have location-based differences in trace metal bioaccumulation. By assessing trace metal concentrations in both aggregations we can: (1) determine whether there are differences in trace metal concentrations between sampling locations, (2) assess if trace metal concentrations have changed since previous studies in SDB green turtles, and (3) compare SBNWR and SDB turtle trace metal concentrations with those found in previous research around the world.</p><p>Previous studies have shown that whole blood and red blood cell trace metals were positively correlated with trace metals in kidney and liver tissues (Keller et al., 2004; van de Merwe et al., 2010a), albeit at significantly lower levels. Measuring trace metals in blood and scute samples has been used to provide both short- and long-term exposure signatures that have been correlative to organ burden (Day et al., 2005; Komoroske et al., 2011). Using these non-lethal measurements, we assessed and compared trace metal bioaccumulation differences of green turtle foraging aggregations influenced by heavy urbanization to determine if trace metal composition differed in both tissue types across locations. These data will enhance our understanding of how green turtles accumulate trace metals in two different impacted environments, as well as assess green turtles as environmental sentinels of trace metals.</p><!><p>Green turtles were captured and tagged in two locations (Figure 1): SBNWR (33° 44' 06.8" N, 118° 03' 51.9" W) and SDB (32° 36' 54" N, 117° 6' 4" W), using previously established techniques (Eguchi et al., 2010; Crear et al., 2016), and in collaboration with National Marine Fisheries Service (NMFS) under Permit #16803. One to three 100-m long entanglement nets were set and checked every 30 min in areas where green turtles are known to inhabit. Once captured, green turtles were brought ashore or onboard a support vessel for processing. Sex of captured green turtles were determined using morphology for adults (as mature males have longer tails) and hormone concentration (data not shown) for immature turtles (as they do not have external morphology to distinguish sex until maturity) (see Caldwell, 1962; Allen et al., 2015), weighed (to the nearest 0.1 kg), and several morphometric measurements (e.g., curved carapace length; to the nearest 0.1 cm) were obtained.</p><p>National Institute of Standards and Technology (NIST) protocols (Keller et al., 2014b) were followed with moderate modifications to reduce sample contamination and increase sensitivity to lower detection limits. Briefly, blood was collected via the dorsal cervical sinus using 21-gauge and 3.8 cm needles (Owens and Ruiz, 1980). Blood was drawn into sodium-heparinized 10-mL glass vacutainers (Becton Dickson, San Jose, California). To account for varying keratin concentrations throughout the carapace (Schneider et al., 2015), scutes were sampled via numbering each scute and using a random number selector to randomly choose scutes to sample. Scutes were only sampled if there were no injuries on the scutes chosen or if the scute exhibited a thin keratin layer. Scute samples were collected with sterile stainless steel blades by scraping a thin layer of carapace to remove algae and microorganisms, followed by scraping scute shavings into a whirl-pak sampling bag, modified from Day et al. (2005). All scute and blood samples were placed in a cooler with cold packs and a cloth barrier to prevent hemolysis and point contact freezing until transported back to laboratory for processing. Scute samples were only collected from green turtles with no carapace injuries. To account for possible equipment and container contamination, Millipore water (Q-Pod®, Burlington, MA) was used as a field blank and collected into the same materials as would be used for collecting blood. After morphological measurements and sample collection, turtles were tagged and released near the location of capture. Typically, the entire procedure took approximately one hour. Whole blood in vacutainers was spun at 3000 rpm for 10 min to separate plasma, buffy layer, and red blood cells (RBCs). After removal of plasma and white blood cells for companion studies, RBCs were pipetted into 2 mL cryovials using sterile low-density polyethylene pipette tips. Samples were placed at –20°C overnight, then transferred to –80°C freezers until trace metal analysis. Whole blood and scute samples were collected from 20 green turtles in SDB and 16 in SBNWR.</p><!><p>To assess the degree of trace metal accumulation in green turtles, RBC and scute samples were digested using concentrated Nitric/Hydrochloric acid and sonicated for 3 hours (EPA Method 200.15). RBC samples were used instead of whole blood to avoid possible errors caused by hematocrit as seen in Komoroke et al. (2011) and Day et al. (2005). Twenty-five trace metals (B, Al, Ti, V, Cr, Mn, Fe, Co, Ni, Cu, Zn, As, Se, Sr, Ag, Cd, Sn, Sb, Ba, Pb, and Hg) were analyzed using an Agilent (7500) Inductively Coupled Plasma Mass Spectrometer (ICPMS) equipped with a reaction/collision cell (Agilent, Santa Clara, CA). Internal standard curves were based on a 6-point linear regression calibration curve with an R2 value of 0.99. Calibration standards were purchased from a NIST traceable commercial supplier (AccuStandard, New Haven, CT). The limit of detection for all elements was determined via EPA method 200.15 (see Kazi et al., 2009). Blanks with Millipore water and blank spikes were analyzed along with all samples for quality assurance. The limit of detection (LOD) for most metals was below 0.0005 μg/g, with the exception of total mercury (LOD = 0.025 μg/g; Supplementary Table 1). Each sample type was digested with a replicate sample from one individual, a corresponding matrix spike sample from the same individual, a blank, and a blank spike. Percent recovery of blank spikes was calculated as (measure/expected)*100% and matrix spike recovery was calculated as ([measured – measured non-spiked replicate]/expected)*100%.</p><!><p>All statistical analyses were done in R (version 3.3.3; R Core Team, 2018) with significance based on an alpha of 0.05. Due to low detection limits, all non-detections were treated as zeros for statistical purposes. Zeros reported throughout the study are non-detections. Turtles captured more than once had their values averaged. Principal component analyses (PCA) were used to determine location-based differences in green turtle trace metal concentrations and to identify specific metals for further comparison by location (R package vegan; Oksanen et al., 2017). Separate PCAs were conducted for scute and RBC trace metals concentrations to examine long and short-term trace metal exposure, respectively. PCA included only trace metals found at detectable levels, and the analyses focused on prevalent elements that were detected in more than six out of 36 turtles. K-means clustering analyses were done on the first two principal components to assess location-based differences (R package cluster; Maechler et al. 2016). T-tests were used to compare size and element concentrations between the two locations. Relationships between metal concentrations and turtle body size and between metal concentrations with other metal concentrations were assessed using a linear regression.</p><!><p>SBNWR green turtles had significantly (p<0.001) smaller carapace length (median ± SE = 70.7 ± 1.39 cm, range = 50.8 – 82.6 cm) than SDB green turtles (mean ± SE = 89.8 ± 3.86 cm, range = 61.6 – 114.1 cm). PCA for turtle body size and metals detected in scute tissue had two principal components (PC) that accounted for 60% of the variance (Figure 2A). Variables with the strongest loading factors for PC1 were cadmium (−0.414) and selenium (−0.388), and the strongest loading factors for PC2 were zinc (−0.470) and iron (−0.410). For scute samples, eleven SDB green turtles and eleven SBNWR green turtles fell within one 95% confidence ellipse, whereas eleven SBNWR green turtles and twelve SDB turtle fell within the second ellipse (Figure 2A). Cadmium, selenium, mercury, and silver accounted for the majority of the differences between PCA groups for scute samples (Figure 2A). PCA for turtle body size and metals detected in RBCs had two PCs that accounted for 44% of the variance (Figure 2B). Variables with the strongest loading factors for PC1 were vanadium (–0.316) and cobalt (–0.305), and the strongest loading factors for PC2 were silver (0.318) and mercury (−0.333). For RBC samples, all SDB and seven SBNWR green turtles fell within one 95% confidence ellipse, whereas one SDB individual and all SBNWR turtles were within a second 95% confidence ellipse (Figure 2B). Based on PC loadings, metals that accounted for the majority of the difference between the PCA groups for RBCs included aluminum, lead, cadmium, and selenium (Figure 2B).</p><!><p>Of 23 elements measured, 20 were detected in scute samples in SDB turtles and 19 elements were detected in SBNWR scute samples (Table 1). There were no significant differences between green turtle's essential metal concentrations, except selenium, in scute samples. Only SBNWR green turtles had detectable levels of mercury in scutes, whereas no mercury was detected above the LOD (LOD = 0.025 μg/g) in SDB green turtles. SBNWR mercury scute concentrations displayed a significant negative relationship (R2 = 0.62; p<0.001) with turtle size. SDB green turtles had significantly higher (p = 0.002) concentrations of cadmium in their scute samples than SBNWR turtles (Table 1). SBNWR green turtles (n = 16) had significantly higher (p<0.001) selenium in scute samples than scute samples of SDB green turtles (n= 20). No statistically significant relationship was found between selenium and cadmium in scute samples (p = 0.22).</p><!><p>Of 23 elements measured, 16 elements were detectable in RBCs (Table 2) in SDB and only 15 elements were detected in SBNWR. There were no significant differences between green turtle's essential metal concentrations, except selenium, in RBC. SDB green turtles had significantly higher (p = 0.01) aluminum in their RBCs than SBNWR green turtles (Table 2). SBNWR green turtles (n = 16) had significantly higher (p<0.001) selenium in RBC samples than RBC samples from SDB green turtles (Figure 3A; n= 20). SBNWR green turtles had significantly higher (p<0.001) cadmium in RBCs than SDB green turtles (Figure 3B). Green turtle selenium concentrations have a positive and significant relationship (p<0.001, R2 = 0.71) with cadmium concentrations in RBCs, regardless of location (Figure 3C).</p><!><p>With the exception of selenium, essential metal (B, Mn, Fe, Co, Cu, and Zn) concentrations measured in turtles did not vary between locations, regardless of sample type. Chromium, antimony, tin, barium and mercury were all below detection limits in green turtle RBCs but were detected in scute (LOD ~ 0.005 μg/g for all but mercury, see Supplementary Table 1). Only SDB green turtles had detectable levels (LOD < 0.0005 μg/g) of tin in scute and silver in both RBCs and scute samples (Table 1 & 2).</p><!><p>Overall, we found that green turtles residing within the SBNWR and SDB tended to display site-specific trace metal contaminant signatures. Whereas most individual elements were not significantly different between locations, there were some elements that differed (e.g., aluminum, cadmium, and selenium) between green turtles from SBNWR and SDB. Most essential elements, except selenium, were not significantly different across locations or individuals, indicating that essential element concentrations are maintained within similar ranges. Whereas size was not related with any elements measured (except nickel and mercury in scute), there were size differences in turtles sampled from each location, which may introduce indirect influences (such as behavior/age) that could not be measured by our study. Green turtle trace metal differences were expected to be related to sediment trace metal profiles of San Diego and Seal Beach. However, the concentrations of some metals in this study did not reflect location-based sediment concentrations, which could indicate that trace metal profiles in sediment are not necessarily reflected in green turtle tissue. In scute samples, only SDB green turtles had detectable amounts of silver despite both regions having similar sediment silver concentrations (~6.15 μg/g; Dodder et al., 2016). Since silver concentrations were higher in SDB prior to more stringent regulations (Schiff et al., 2011; Dodder et al., 2016), larger/older SDB turtles may have silver concentrations in scutes that reflect these patterns. Silver has been shown to bioaccumulate in green turtle food sources in higher concentrations than sediment (Komoroske et al., 2012) indicating that silver in the SDB region may be more readily available through diet than SBNWR, or that large individuals from SDB have accumulated enough silver over time to be detectable in scute samples. Aluminum concentration differences in green turtles were reflective of sediment patterns found in their respective locations (~ 34838 μg/g dw Al in SBNWR and ~54348 μg/g dw Al in SDB), indicating that aluminum input into San Diego Bay, possibly from shipyard activity and urban runoff, is bioaccumulating into green turtle food sources (Dodder et al., 2016). However, previous research has shown that aluminum does not readily bioaccumulate in eelgrass (Zostera marina), sponge, red algae, or green algae, implying that aluminum could be coming from other food sources, such as mobile invertebrates which inhabit plant material that sea turtles are known to consume (Lemons et al., 2011; Komoroske et al., 2012). Only SBNWR turtles had detectable levels of mercury in scute samples, despite both locations having similar mercury concentrations in sediment (Dodder et al., 2016). Mercury concentrations were negatively related to SBNWR turtle size, indicating that once turtles recruit to SBNWR their mercury concentrations decrease over time. Previous research (Komoroske et al. 2011) had found mercury concentrations above 0.025 μg/g in scute samples, while the current study did not find any SDB scute samples above the LOD. However, the current study and Komoroske et al. (2011) had two different methods used with two different LODs. These results possibly indicate that mercury concentrations in green turtle scute has decreased as they recruit to these coastal habitats and that SDB green turtle mercury exposure may have decreased in recent years. Additional monitoring with similar methods will be needed to better assess these patterns in scute mercury. PC analyses identified two elements, cadmium and selenium, as likely major drivers for the differences between metal signatures in southern California green turtles.</p><!><p>Consistent with the differences in cadmium levels in RBC, sediments from SBNWR (0.52 – 1.13 μg/g dry weight) have higher levels of cadmium compared to sediments from SDB (0.09 – 0.24 μg/g dry weight; Dodder et al., 2016). However, compared to SBNWR green turtles, SDB green turtles had higher scute cadmium concentrations, suggesting that SDB turtles have had more long-term cadmium exposure while SBNWR turtle have had higher short-term exposure. Considering the Seal Beach region has higher concentrations of cadmium in sediment than SDB (Dodder et al., 2016), and SBNWR turtles are generally smaller than SDB turtles, SDB green turtles could be bioaccumulating cadmium with age while residing in SDB (Dodder et al., 2016). These contaminant patterns suggest that sediment cadmium is accumulating in green turtle food sources, most likely eelgrass, which readily bioaccumulates cadmium and deposit cadmium in stem and leaf tissue (Lyngby and Brix, 1984). However, green turtles from SBNWR have been observed within the San Gabriel River, where no eelgrass habitat is present, which implies that green turtles that forage within the river may have a more algae biased diet (Crear et al., 2016; Crear et al., 2017). SBNWR green turtles move between SBNWR and the San Gabriel River, which may indicate that trace metal accumulation could be occurring from foraging items in both locations, possibly explaining differences from SDB. Previous research has shown evidence that green turtles have a high capacity to metabolize and detoxify cadmium at a high rate (Sinaei, 2016). However, SBNWR green turtles have one of the highest observed cadmium concentrations in green turtle red blood cell samples reported to date, 0.305 μg/mL in RBCs, suggesting that cadmium levels in SBNWR green turtles are abnormally high compared when compared to other green turtle studies that use whole blood (Villa et al., 2016).</p><!><p>Selenium concentrations in sediment at our two study sites were similar (approximately 0.48 μg/g); however, green turtle selenium concentrations in blood and scute samples were greater than sediment patterns (Dodder et al., 2016). Previous research has shown various green turtle food sources, such as many algae and plant species, accumulate higher selenium concentrations than surrounding sediment within SDB (Komoroske et al., 2012). Eelgrass beds are important foraging habitat for green turtles in SDB, and presumably in SBNWR, suggesting that eelgrass habitats may be one of the main sources of trace metal exposure for green turtles from southern California (Lemons et al., 2011; Seminoff et al., unpubl. data; Komoroske et al., 2012). Because the SBNWR area has extensive eelgrass habitat, it is expected that SBNWR green turtles feed on eelgrass habitat similar to SDB green turtles. However, SBNWR turtles may also feed within the San Gabriel River, which may result in a more algae-based diet providing an additional avenue for high selenium exposure (Seminoff et al., unpubl. data). These results suggest that either eelgrass habitat in the SBNWR or algae in the San Gabriel River has significantly more selenium than SDB eelgrass habitat or that other factors, such as non-essential metal exposure, explain increased selenium concentrations in SBNWR green turtles.</p><p>The presence of other, non-essential metals alters the uptake of selenium. For example, in many other organisms, selenium tissue burdens display a positive relationship with cadmium tissue concentrations, and co-exposure of cadmium with selenium has been found to reduce cadmium toxicity (e.g. Pugh and Becker, 2001; Gardner and Oberdorster, 2005; Gardner et al., 2006; Rozman and Klaassen, 2007; Perrault et al., 2011). While selenium and cadmium absorption and distribution mechanisms are not well studied, it is thought that selenium forms insoluble complexes with cadmium that aid in cadmium detoxification and excretion (Ohlendorf, 2003; Rozman and Klaassen, 2007). Therefore, the observed relationship between cadmium and selenium concentrations in green turtle RBCs in the current study suggests that green turtle selenium uptake may be influenced by co-occurring cadmium levels.</p><p>Green turtle selenium contamination represented the starkest difference in trace metals between SDB and SBNWR green turtles, suggesting that SBNWR green turtles are possibly at higher risk of selenium toxicity than SDB green turtles. Most other studies have found adult and juvenile green turtles had selenium concentrations roughly between 0.3 to 7 μg/mL in whole blood and RBC (van de Merwe et al., 2010a; Komoroske et al., 2011; Labrada-Martagon et al., 2011; Ley-Quinonez et al., 2013; Camacho et al., 2014; Villa et al., 2016), while SBNWR green turtles RBCs selenium ranged from 1.02 to 33.06 μg/mL. In many organisms, including banded water snakes (Nerodia fasciata), mallard ducks (Anus pltythynchos), and African clawed frogs (Xenopus laevis), excess selenium, without other trace element exposure, has been shown to reduce hatchling success, cause dermal damage and loss of neurological function (e.g. Heinz, 1996; Ohlendorf, 2003; Rozman and Klaassen, 2007). Avian research into egg contamination has been used as a proxy to model and predict risks in green turtles (Heinz, 1996; Lam et al., 2006). These studies have suggested that green turtle eggs with selenium loads higher than 600 μg/g could experience reduced hatchling success (Heinz, 1996; Lam et al., 2006). Considering green turtles can maternally offload contaminants (van de Merwe et al., 2010b), including cadmium (non-essential) and selenium (essential), to their eggs, future studies could examine what affect cadmium and selenium have on hatchling development and health in green turtles. All but one SBNWR green turtles sampled were female (unpublished results), which indicated that selenium offloading could be a potential issue as these immature turtles become adults.</p><!><p>A previous study has determined reference intervals of trace metal concentrations using whole blood from a "reference" green turtle population that inhabits pristine habitat in the West Barrier Reef in Australia (Villa et al., 2016). These reference intervals are ranges of trace metal concentrations found in a green turtle population not exposed to anthropogenic contaminants. Therefore, concentrations below or above these ranges indicate altered trace metals concentrations in blood (Villa et al., 2016). Most green turtles in our study from both locations had RBC trace metal concentrations above established reference intervals (Villa et al., 2016). This supports the hypothesis that the burden of trace metals found in SBNWR and SDB green turtles are above what is considered background levels. Green turtles in this study had vanadium, chromium, iron, and copper levels (all essential elements) that were within background levels (Villa et al., 2016). However, SDB and SBNWR green turtles had higher than background levels of cobalt, manganese, zinc, strontium, arsenic, cadmium, nickel, and selenium (Villa et al., 2016). Comparison of trace metal concentrations in green turtles from the current study and green turtle trace metal studies around the world (Table 3) suggest that green turtles inhabiting waterways impacted by urbanization have different trace metal accumulation patterns than non-exposed green turtle populations (van de Merwe et al., 2010a; Komoroske et al., 2011; Labrada-Martagon et al., 2011; Ley-Quinonez et al., 2013; Camacho et al., 2014; Villa et al., 2016). The current study's turtles may have high cadmium and selenium concentrations, associated with kidney failure and reduced hatchling success. This information supports increased pollution-based health risks in green turtles living in urbanized coastal zones.</p><!><p>Our study found evidence that a region's unique anthropogenic activity (e.g. ship-yard, power plant) and pollution can affect how green turtles accumulate trace metals. Particularly, SBNWR green turtles had significantly higher selenium and cadmium in their RBCs than green turtles from other recent studies. There was a difference in size observed between the two populations, which was positively related to nickel concentrations in scute and negatively related to mercury concentrations in SBNWR scute samples. Most elements were similar to previous measurements in SDB turtles, with the exception of mercury. Mercury concentrations are lower in current SDB turtle scute samples, indicating a possible reduction of exposure or intake, however previous studies used different methods. Future studies should investigate larger SBNWR individuals to assess whether adults in the SBNWR have more selenium than immature green turtles. SBNWR turtles had higher than average selenium concentrations and comparisons to other studies show evidence that urban habitats continue to expose green turtles to trace elements. With each green turtle foraging aggregation accumulating different trace metal concentrations, there is potential for differential health risks across these two foraging aggregations that future studies could elucidate. Because of their long life spans and extended residency patterns to specific foraging sites, continued monitoring of SBNWR and SDB green turtles will be necessary as trace metal concentrations can potentially change as green turtles continue to inhabit areas impacted by urbanization.</p>

PubMed Author Manuscript

Deciphering Distinct Overpotential Dependent Pathways for Electrochemical CO 2 Reduction Catalyzed by an Iron-Terpyridine Complex

Fe(tpyPY2Me)] 2+ ([Fe] 2+ ) is a homogeneous electrocatalyst for converting CO2 into CO featuring low overpotentials of <100 mV, near-unity selectivity, and high activity with turnover frequencies faster than 100,000 s -1 . To identify the origins of its exceptional performance and inform future catalyst design, we report a combined computational and experimental study that establishes two distinct mechanistic pathways for electrochemical CO2 reduction catalyzed by [Fe] 2+ as a function of applied overpotential. Electrochemical data shows the formation of two catalytic regimes at low (TOF/2 of 160 mV) and high (TOF/2 of 590 mV) overpotential plateaus. We propose that at low overpotentials [Fe] 2+ undergoes a two-electron reduction, two-proton transfer mechanism (electrochemical-electrochemical-chemical-chemical, EECC), where turnover occurs through the dicationic iron complex, [Fe] 2+ . Computational analysis supports the importance of the singlet ground state electronic structure for CO2 binding and that the rate-limiting step is the second protonation in this low-overpotential regime. When more negative potentials are applied, an additional electron transfer event occurs through either a stepwise or protoncoupled electron transfer (PCET) pathway, enabling catalytic turnover from the monocationic iron complex ([Fe] + ) via an electrochemical-chemical-electrochemical-chemical (ECEC) mechanism. Comparison of experimental kinetic data obtained from variable controlled potential electrolysis (CPE) experiments with direct product detection with calculated rates obtained from the energetic span model support the PCET pathway as the most likely mechanism. Moreover, we build upon this mechanistic understanding to propose the design of an improved ligand framework that is predicted to stabilize the key transition states identified from our study and explore their electronic structures using an energy decomposition analysis. Taken together, this work highlights the value of synergistic computational/experimental approaches to decipher mechanisms of new electrocatalysts and direct the rational design of improved platforms ASSOCIATED CONTENT Supporting Information. Supporting Information. Experimental and computational details, supplemental electrochemical and DFT geometry optimized atomic xyz coordinates is available free of charge via the Internet at http://pubs.acs.org.

deciphering_distinct_overpotential_dependent_pathways_for_electrochemical_co_2_reduction_catalyzed_b

INTRODUCTION<!>Computational Model.<!>Observation of Two Distinct Overpotential-Dependent<!>CONCLUSIONS

<p>Increasing energy demands and global climate change motivate the need for technologies that capture and utilize atmospheric carbon dioxide (CO2) and convert it into valueadded carbon products. [1][2][3] In this context, the electrochemical reduction of CO2 4,5 offers a promising way to restore balance to the carbon cycle and develop a net negative carbon footprint if these technologies can be coupled to renewable sources of electricity. A diverse array of upgraded carbon products for the electrochemical CO2 reduction reaction (CO2RR) include CO, CH3OH, and CH4. 6 Of these products, the two-electron two-proton reduction of CO2 to CO is economically most viable due to its usage in the Fischer-Tropsch process. 7 In order to realize the goal of sustainable electrochemical carbon capture and conversion, efficient catalysts are required to selectively drive CO2 reduction versus the thermodynamic and kinetically competitive hydrogen evolution reaction (HER). 5,8 The electrochemical CO2 reduction reaction has been extensively explored across materials, 9 biological, 10 and molecular systems. 11 Molecular electrocatalysts are of particular interest as they are ideal platforms that can be rationally and systematically tuned through synthetic chemistry [12][13][14] with a level of precision and in the absence of defects that is unavailable to heterogeneous congeners. 9,15 Moreover, owing to their small size relative to enzymes and their homogeneous nature, they can be investigated at a mechanistic level which aids in the understanding and directed development of improved catalytic platforms. 14,16,17 In this regard, iron-based molecular systems are especially desirable owing to the abundance of this element; the iron porphyrin platform Fe-TPP (TPP = tetraphenylporphyrin) is a robust and active catalyst making it a popular framework for ligand development with a considerable number of derivatives having been prepared in recent years. [18][19][20][21][22][23][24][25][26][27][28] The turnover frequency of the original Fe-TPP platform has been improved by several orders of magnitude from 10 2 s -1 to 10 6 s -1 for Fe-o-TMA, a tetra-substituted trimethylanilinium, derivative. The (current) top performance of 10 9 s -1 for Fe-TPP-based catalysts is achieved by introducing four bulky, methylimidazolium-containing groups. 29 The success story of the Fe-TPP system illustrates the effectiveness of systematic synthetic modification of the ligand scaffold and optimization of reaction conditions. Alternatively, simpler polypyridyl ligand platforms, including ironbipyridine (bpy) 30,31 (Figure 1a) and iron-quaterpyridine (qpy) [32][33][34][35] (Figure 1c), have received increased attention for electrochemical CO2 reduction chemistry. These systems are robust under electrochemical conditions and are synthetically accessible through facile and modular routes which facilitates precise tuning of their sterics and electronics in a rational fashion.</p><p>Against this backdrop, the development of novel metalpolypyridyl electrocatalysts for proton and CO2 reduction has been a part of a larger research program between our laboratories in energy conversion chemistry. 30,31,[36][37][38] We recently reported 39 a novel terpyridine (tpy)-based iron polypyridine complex (Figure 1b), [Fe(tpyPY2Me)] 2+ ([Fe] 2+ ), that leverages ligand non-innocence of the tpy moiety and metal-ligand cooperativity through exchange coupling. These two factors yield mild reduction potentials for the complex in comparison to other pyridine-based catalysts as illustrated in Figure 1 and enables it to electrochemically convert CO2 into CO at extremely low overpotentials (η), resulting in high product selectivity and rates under both organic solvents and aqueous conditions. The Faradaic efficiency for CO production (FECO), rates (kmax) and overpotentials of the three platforms are compared to the Fe-TPP platforms and summarized in Table 1. Comparison of these parameters clearly identifies these polypyridyl complexes as some of the best homogeneous catalysts to date, but we note that benchmarking should be done with care and caution. Indeed, determination of accurate overpotentials requires the use of the thermodynamic potential required to convert CO2 into CO, which is often unknown for the exact experimental conditions and can thus vary by over 500 mV depending on solution conditions. To normalize for this uncertainly, we give the values reported by Matsubara 40 that take into account the effect of solvent mixtures, acid additives, and homoconjugation of phenol in acetonitrile. [40][41][42][43] Similarly, estimation of kinetic performance based on the maximum turnover frequency (TOFmax) is highly dependent on the method utilized (e.g., foot-of-the-wave analysis, peak catalytic current analysis, etc). [44][45][46][47][48] As such, the methods used to determine the rates are indicated in Table 1. Specifically, in order to minimize uncertainty in rate determination for our [Fe] 2+ catalyst, we extracted the kinetic parameters directly from the averaged specific current densities for CO production obtained from variable potential CPE experiments that were conducted in triplicate as described by Savéant and coworkers. 49,50 In our initial study of the [Fe(tpyPY2Me)] 2+ system, we reported the synthesis, characterization, and electrocatalytic behavior of [Fe] 2+ for CO2RR to CO. Synthesis of the two-electron reduced product, [Fe(tpyPY2Me)] 0 ([Fe] 0 ), and spectroscopic characterization enabled us to attribute its exceptional catalytic activity to its unique open-shell singlet electronic structure that results from the antiferromagnetic coupling of the intermediate-spin Fe(II) center (SFe = 1) to a doubly reduced, triplet ligand system (Stpy = 1). We were able to establish that this electronic structure was responsible for the reduction of CO2 to CO at low overpotentials, through the preparation and comparison to a series of redox-(in)active metal and ligand controls. However, despite a detailed understanding of the electronic structure of the [Fe] 2+ system and its two-electron reduced product, the mechanism(s) through which it functions in CO2RR are insufficiently explored. In the present study, we use a combined experimental and computational approach to identify two distinct mechanistic pathways for CO2RR catalyzed by [Fe(tpyPY2Me)] 2+ depending on applied overpotential, .</p><p>We propose that at low overpotentials [Ecat/2 of -1.43 V vs Fc/Fc + (TOF/2 = 0.16 V)], [Fe] 2+ undergoes a two-electron reduction, two-proton transfer mechanism (electrochemicalelectrochemical-chemical-chemical, EECC) where turnover occurs though the dicationic iron complex, [Fe] 2+ . At higher overpotentials [Ecat/2 of -1.86 V vs Fc/Fc + (TOF/2 = 0.59 V)], an additional electron transfer event becomes possible through either a stepwise or a proton-coupled electron transfer (PCET) pathway, enabling catalytic turnover from the monocationic iron complex ([Fe] + ) via an electrochemical-chemical-electrochemical-chemical (ECEC) mechanism (see supplementary information for details surrounding the determination of Ecat/2 and TOF/2). The PCET steps considered in this manuscript corresponds to concerted electronproton transfer. Based on this detailed mechanistic analysis, we propose the design of improved ligand frameworks and explore their electronic structures with an energy decomposition analysis (EDA). This mechanistic analysis lays the foundation for further rational optimization of theoretically driven modifications of the ligand scaffold for improved catalytic activity.</p><!><p>In this section we briefly explain the model, its assumptions and expected errors (see SI for more technical details and justifications). Density functional theory calculations for free energies, activation energies, reduction potential and the LOBA 51 oxidation state analysis were performed with the Q-Chem package 52 (version 5.3.0) using the ωB97X-D 54 functional with a mixed basis for the optimization and frequency calculations (def2-SVP basis for all main group elements, def2-TZVP basis set for Fe). 53 This functional was chosen based on our previous study and extensive functional screening (see SI reference [43]). In addition, we probed several other popular density functionals but no other functional yielded better agreement in both the predicted redox potentials and rates (see Table S7). Gibbs free energies (G) were used to compute reduction potentials and adiabatic spin gaps and are based on standard thermodynamic cycles. [54][55][56] Solvation energies were approximated by performing single point calculations applying the implicit C-PCM solvent model with the larger def2-TZVP basis for all elements 53 (see reference [ 57 ] more technical details). Reduction potentials are reported with an isodesmic scheme against the ferrocene/ferrocenium couple (Fc/Fc + ) used as an internal standard. 55,56,58 This method allows accurate predictions even at a modest level of theory with reported errors within approximately ∼100 mV (∼4 kcal/mol) of experimental values. 55 Of particular relevance to this work, this approach has yielded accurate calculated reduction potentials in several pyridine based electrocatalysts. 31,35,39 The RMSD of ωB97X-D for barrier heights is ∼2 kcal/mol using gas phase high level wave function methods as the reference; 59,60 When comparing to experimental values, additional errors may arise from describing solvation by the implicit solvation model and other simplifications in the computational model versus experiment. The calculation of free energies for protonation reactions with implicit solvent models results in expected deviations versus experimental values of ±3 pKa units. 61 We tested our computational protocol for phenol (PhOH) with ωB97X-D yielding a pKa of 24.6 which is 4-5 pKa units too low. 62 While this is a large error, it also has a significant systematic component: calculated relative pKa values are more reliable because of favorable error cancellation by removing the experimental free energy of the proton. Thus, calculated pKa values should mainly be compared against each other.</p><p>We use PhOH as the main proton source for calculating reaction barriers involving protonation reactions as it was added to the reaction mixture in molar quantities for the CPE experiments. The concentration of other proton sources (H + and H2CO3) is negligible. 56 However, we report free energies of protonation reaction steps versus carbonic acid (H2CO3) because it is the strongest acid in solution and reprotonates phenolate (PhO -) via complexation of CO2 and OH -. This approach was also used in a previous study of quaterpyridine based catalysts. 35 The calculated standard potential for the reduction reaction using carbonic acid is -1.28 V vs Fc/Fc + ; this is in good agreement with the experimental estimate of -1.27 V used in our previous study. 39 Matsubara 40 estimates the standard potential in wet CH3CN depending on the mole fraction of water to be between 0.95 V and 1.63 V vs Fc/Fc + .The highest predicted reduction potential is −1.83 V vs Fc/Fc + (η = 0.56 V) in our catalytic cycles (vide infra); this translates to a total reaction energy of −25.4 (CO2 + 2H2CO3 + 2e -→ CO + H2O + 2HCO3 -).</p><p>In contrast, the calculated standard potential using PhOH of -1.94 V vs Fc/Fc + is too negative. However, the calculated standard potential is lowered to -1.33 V by incorporating the effect of homoconjugation of phenol ins acetonitrile (CO2 + 6PhOH + 2e -→ CO + H2O + 2(PhO -* 2PhOH)). 42,43 The formation constant (Kf) for homoconjugation of PhOH in acetonitrile is very large (~10 5.8 ) 41 and therefore we would expect significant formation even at very small concentrations of acid. This further complicates accurate determination of an unambiguous value of overpotential that would characterize the catalyst under these conditions (i.e., TOF/2) as the value strongly depends on the concentration of the acid. Furthermore, here we report kinetics based on effective overpotentials (eff) and the corresponding applied potentials referenced to the Fc/Fc + couple across a wide potential range. We employ the energetic span model to predict the turnover frequency (TOF) based on our calculated catalytic cycles. This model identifies the key intermediates and transition states, which control the rate of catalysis (see the corresponding section for more detail). 62 Experimental Methods. In this section we briefly outline the experimental techniques utilized to test our hypothesized mechanistic pathways. Additional details can be found in the supplementary information. Binding equilibria for CO2 coordination and CO dissociation were investigated by cyclic voltammetry (CV) with varying scan rates (0.1 to 50 V/s). Catalytic stability was probed with multi-segmented CV experiments over one-hundred cycles as well as with long-term (1 h) preparative scale controlled potential electrolysis (CPE) experiments taken across a range of applied potentials. Averaged Faradaic efficiencies for H2 (FEH2) and CO (FECO) production were determined from the one-hour CPE experiments, conducted in triplicate, with direct product detection by gas chromatography. The observed rate constants (kobs) and turnover frequencies (TOF) were extracted from the averaged specific current densities for CO production (jCO) over both short-term CPE experiments (5 min) conducted in a small CV cell and from longterm CPE experiments (1 hr) conducted in a large gastight PEEK electrolysis cell. We observed good agreement between rates obtained from both measurements. Kinetic isotope effects (KIE) were measured using 1 M solutions of phenol-H6 and phenol-D6 as the proton source. Observed rate constants for the KIE experiments were obtained using catalytic plateau current analysis despite the noncanonical catalytic behavior of [Fe] 2+ which makes this analysis inaccurate. This analysis was appropriate here because errors in rate determination are eliminated by taking the ratio of kH/kD.</p><!><p>Regimes for Electrochemical CO2 Reduction Catalyzed by [Fe(tpyPY2Me)] 2+ . Electrochemical analysis of this molecular iron CO2RR system exhibits two distinct catalytic regimes that are dependent on applied potential (Figure 2). Cyclic voltammograms collected under CO2 atmosphere with the addition of 1 M phenol (PhOH), as a proton source, shows the formation of two catalytic waves. The first regime reaches a half-maximum catalytic current at a potential of −1.43 V vs Fc/Fc + (𝐸 𝑐𝑎𝑡/2 1 ; η = 160 mV) and displays a canonical S-shaped wave indicative of pure kinetic conditions without substrate consumption. 43,44 At more negative applied potentials beyond −1.75 V vs Fc/Fc + (η = 480 mV), a second catalytic response is observed reaching a half-maximum catalytic current at a potential of −1.86 V vs Fc/Fc + (𝐸 𝑐𝑎𝑡/2 2 ; η = 590 mV). This second catalytic regime shows peak-shaped behavior indicative of substrate consumption by the rate-determining step (Figure 2). 44 The catalytic voltammogram clearly illustrates the formation of two catalytic waves with maximum current densities achieved at approximately −1.66 V vs Fc/Fc + (η = 390 mV) and −1.98 V vs Fc/Fc + (η = 710 mV). We attribute this behavior to the occurrence of two distinct catalytic regimes with disparate mechanisms. Building on this observation, we explore various reaction pathways for these two catalytic regimes in the presence of 1 M PhOH and 4-chlorophenol (Cl-PhOH) to probe the effect of the pKa on the observed rates and to investigate the origin of selectivity of [Fe] 2+ for the CO2RR relative to the HER (vide infra). The proposed mechanisms are in line with experimental kinetic data obtained from controlled potential electrolysis (CPE) experiments collected with direct product detection via gas chromatography. In order to compactly provide relevant information regarding the electronic structure of the proposed catalytic intermediates along the way, we introduce a naming scheme that incorporates the multiplicity (M, M= 2S+ 1), charge (C), and the identity of the sixth, axial ligand (X) on the iron center: M [Fe−X] C ; for example, 1 [Fe−CH3CN] 2+ corresponds to the unreduced, hexacoordinated initial iron complex in the singlet state. Low Overpotential Pathway. Figure 2 shows that the current density of the first catalytic regime reaches a plateau at approximately −1.66 V vs Fc/Fc + (η = 390 mV) with rates that are slower relative to the second catalytic regime. A mechanism for this first regime is illustrated in Figure 3 and presented in this section.</p><p>First, we propose two, sequential single electron reduction steps followed by the dissociation of CH3CN to generate the catalytically active open-shell singlet, 1 [Fe] 0 . CO2 binding and subsequent protonation steps can then occur resulting in the loss of a water molecule and the formation of 1 [Fe−CO] 2+ . Ligand exchange with exogenous CH3CN and CO release closes the catalytic cycle. Interestingly, the formal oxidation state of the central metal does not change during catalysis. The following sections will expand upon these proposed individual elementary steps before we expand our discussion to include a proposed catalytic pathway for the second catalytic region that occurs at more negative potentials. Reduction. The reduction pathway of 1 [Fe] 2+ was established in our initial report. 39 Variable temperature NMR and Mössbauer studies established that the starting 1 [Fe] 2+ complex is predominantly low-spin Fe(II) with a small population of thermally accessible spin excited states near room temperature that we attributed to the axial distortions engendered by the rigid tpyPY2Me ligand (Figure 4a). The first reduction of 1 [Fe] 2+ is ligand centered, with occupation of the tpy-π * orbital with almost no excess spin density on the metal center, yielding 2 [Fe−CH3CN] + , a ground state doublet and CO2 (purple) with the addition of 1 M PhOH as a proton source. Cyclic voltammograms were collected with a scan rate of 100 mV/s in an electrolyte of 0.10 M TBAPF6 dissolved in CH3CN. The proposed two distinct mechanistic pathways for 1 [Fe] 2+ are designated by the potentials at which they reach half of the maximum catalytic current (Ecat/2) at -1.43 and -1.86 V vs Fc/Fc + and are labeled in blue and red, respectively. The first regime turns over from the 2 + complex and undergoes a proposed EECC mechanism while the second regime at more negative potentials undergoes turnover from the 1 + complex via an ECEC mechanism.</p><p>composed of an low-spin Fe(II) center and a radical anion tpyPY2Me -ligand (Figure 4b). The addition of a second electron occupies another tpy-π * orbital and induces a spinstate transition of the iron center from low-spin (SFe = 0) to intermediate-spin (SFe = 1) and ligand dissociation of the axially coordinated CH3CN solvent molecule. This tetraradicaloid electronic configuration allows for strong exchange coupling of the two unpaired d electrons on the intermediate-spin iron center with the two electrons in the tpy-π * manifold, forming an overall open-shell singlet ground state, 1 [Fe] 0 (Figure 4c). The open-shell singlet electronic structure was validated by synthesizing and isolating 1 [Fe] 0 from the chemical reduction of 1 [Fe] 2+ with decamethylcobaltocene and fully characterizing the resulting coordination compound by single-crystal X-ray crystallography, Mössbauer spectroscopy, X-ray absorption spectroscopy, and DFT and CASSCF calculations. Through these spectroscopic studies and with comparison to control complexes, we were able to attribute the catalysis of 1 [Fe] 2+ for the CO2RR at mild overpotentials to this antiferromagnetic complex. The predicted reduction potentials for the two reductions are −1.46 V and −1.51 V vs Fc/Fc + , which are in good agreement with the experimental cyclic voltammetry data collected under argon atmosphere that show two closely spaced one-electron reduction waves centered at −1.43 V vs Fc/Fc + . The exchange-coupling shifts the second reduction wave positive by a remarkable 640 mV relative to [Zn(tpyPY2Me)] 2+ which employs the same redox-active tpyPY2Me ligand, but contains a Zn(II) metal center that is unable to participate in stabilization through antiferromagnetic coupling. 1 [Fe] 2+ (left side), we first propose two one-electron reduction steps followed by the dissociation of CH3CN to generate the catalytically active open-shell singlet, 1 [Fe] 0 . CO2 binding and subsequent protonation steps can then occur resulting in the loss of a water molecule and the formation of 1 [Fe−CO] 2+ . Ligand exchange with exogenous CH3CN and CO release closes the catalytic cycle (outer pathway). Alternatively, direct protonation of the pyridyl arm in 1 [Fe] 0 to give 1 [Fe−NPYH] + followed by CO2 coordination and proton transfer was found to be a competitive pathway (inner pathway). All reaction and activation energies are given in units of kcal/mol and all reduction potentials are referenced to the computed Fc/Fc + couple. CO2 Binding. CO2 coordination occurs after 1 [Fe] 2+ has been reduced by two electrons and the iron center has undergone a spin-state transition and subsequent ligand dissociation of the axially bound solvent molecule. The binding of CO2 under standard conditions is endergonic and remains on the singlet surface yielding 1 [Fe−CO2] 0 . We attempted to stabilize the CO2 adduct with the addition of either explicit water molecules or electrolyte (TBAPF6) but the binding remains endergonic with a free energy of +12.1 kcal/mol (∆H = ∼4 kcal/mol). The addition of an explicit phenol lead to direct protonation of the CO2 moiety. The CO2 ligand is bound via the carbon atom with an angle of 124º (Figure S1a). This structure is indicative of a closed-shell highly reduced CO2 moiety and a low-spin singlet Fe(II) center with a neutral tpyPY2Me ligand. The LOBA analysis indicate that the two excess electrons are stabilized in a delocalized molecular orbital (MO), which mainly consists of the CO2 * orbitals but also has significant contribution from the Fe-3𝑑 𝑧 2 (40%) and some contribution from the tpy * (see Figure S2).</p><p>The endergonic binding can be rationalized by the high stability of 1 [Fe] 0 , attributed to its antiferromagnetically coupled electronic structure, which yields a positive reduction potential but consequently at the cost of sluggish CO2 binding. This prediction is in line with experimental findings as high concentrations of PhOH are required to observe CO2 binding and subsequent catalysis. In order to probe the binding of CO2, we attempted to measure the equilibrium constants for CO2 binding (𝐾 𝐶𝑂 2 ), based on the potential shift of the ligand reduction wave under Ar and CO2 atmosphere (Figure S3). Under an Ar atmosphere, without the addition of PhOH as a proton source, there is no shift in the ligand reduction waves centered at −1.43 V vs Fc/Fc + (Figure S3a). When PhOH (1 M) is added, a catalytic wave is formed; however, a reversible couple could not be observed even when the experiment was performed with fast scan rates up to 50 V/s (Figure S3b,c). Taken together, these results suggest that CO2 binding and thus catalysis is hampered when the system is proton limited. Interestingly, the higher spin-state surfaces [triplet ( 3 [Fe] 0 ) and quintet ( 5 [Fe] 0 )] both exhibited much higher barriers for CO2 binding that would effectively prevent catalysis. Thus, the openshell singlet electronic structure of 1 [Fe] 0 is key not only to decrease the overpotentials required for catalysis but also to facilitate CO2 binding even though it is thermodynamically difficult. We further investigated an alternative pathway of CO2 binding to a singly reduced intermediate ( 2 [Fe] + ). This pathway results in an even more endergonic CO2 binding (14.1 kcal/mol) making it an unlikely intermediate. Protonation. In a first scenario, the first protonation of 1 [Fe−CO2] 0 is barrierless and strongly exergonic (−16.9 kcal/mol) yielding the carboxyl intermediate, 1 [Fe−CO2H] + (Figure S1b). This intermediate can be best described as a CO2H -moiety coordinated to the Fe(II) center with the neutral tpy ligand framework. The second proton transfer step is exergonic (−2.7 kcal/mol) but with a barrier of 13.9 kcal/mol as it is coupled to the cleavage of the C−O bond and results in the loss of water and the formation of the carbonyl intermediate ( 1 [Fe−CO] 2+ ) (Figure S1c). The transition state geometry is depicted in Figure S5. The influence of explicit solvent molecules added to the transition state geometry was probed, but we were unable to find lower barriers with the addition of exogenous water molecules. In a second scenario, 1 [Fe] 0 is protonated at one of the pyridine arms yielding a pyridinium intermediate, 1 [Fe−NPYH] + (Figure S1d). The calculated pKa of 8 is quite acidic (versus a calculated pKa of 16 for H2CO3); thus, the free energy of protonation coupled to carbonic acid is 11.5 kcal/mol. Previous experimental work by Matsubara shows that the pKa of the reaction mixture is significantly lower with the addition of water. The experimental pKa for CO2 saturated water-acetonitrile mixtures can range from 7.8 to 16.8 where we would expect the addition of molar quantities of PhOH to further acidify the solution thus making this intermediate competitive. 40 The binding of CO2 to 1 [Fe−NPYH] + is very exergonic (−16.3 kcal/mol) and directly yields a carboxy intermediate, 1 [Fe−CO2H] + (Figure S1b), by skipping the high energy CO2 adduct. This adduct is achieved by a simultaneous intramolecular proton transfer from the pyridinium and electron transfer from the tpy upon binding of CO2 to form CO2H -. The second protonation proceeds as described above, see Figure 3 (outer pathway). CO Release. CO release from 1 [Fe−CO] 2+ is exergonic and barrierless which can be attributed to the relatively high oxidation state of the iron center, which limits backbonding interactions. In addition, the low solubility of CO in CH3CN promotes removal of CO from solution and into the gas phase. In order to further probe CO coordination and release, CVs of 1 [Fe] 2+ were collected under CO atmosphere and compared to data collected under Ar atmosphere. The addition of CO results in the formation of very small reductive features at −0.59 and a quantitative reductive wave at −1.23 V vs Fc/Fc + . Additionally, the reversibility in the ligand reduction waves centered at −1.43 (Figure S6a) are lost at slower scan rates of 100 mV/s. We tentatively assign the new reduction wave at −1.23 V vs Fc/Fc + to the reduction of an Fe(II)-carbonyl species ( 1 [Fe−CO] 2+ ). This value agrees well with the predicted reduction potential of −1.18 V vs Fc/Fc+, supporting this assignment. We do note that attempts to spectroscopically characterize the waves at −0.59 and −1.23 V vs Fc/Fc + have been unsuccessful to date. Multisegmented CV data collected under CO2 atmosphere with the addition of 1 M PhOH, do not show any significant changes or the formation of new reductive features that could be attributed to the buildup of iron carbonyl species in the voltammograms across one-hundred cycles (Figure S6). Taken together, these data support the computational findings that CO binding is weakly endergonic by 2 kcal/mol and barrierless and thus we do not observe the Figure 5. Proposed mechanistic cycles for the high overpotential regime. The high overpotential catalytic regime turns over from 2 [Fe] + following an initial induction period. Single-electron reduction of 2 [Fe] + gives the catalytic resting state, 1 [Fe] 0 . Following formation of 1 [Fe] 0 , the pathway diverges in three directions. CO2 binding can occur first followed by protonation to give 1 [Fe−CO2H] + (outer pathway) or protonation-first can occur followed by CO2 coordination (inner pathway). The 1 [Fe−CO2H] + intermediate can then be further reduced to 2 [Fe−CO2H] 0 and following the final protonation step and loss of water generates the carbonyl complex, 2 [Fe−CO] + which regenerates 2 [Fe] + following ligand exchange. We additionally explore the possibility of overcoming the high energy barriers associated with either CO2 binding or protonation of 1 [Fe] 0 by undergoing a PCET pathway to generate 2 [Fe−NPYH] 0 (center pathway). All reaction and activation energies are given in units of kcal/mol and all reduction potentials are referenced to the computed Fc/Fc + couple. accumulation of iron carbonyl species under electrocatalytic conditions. Further reduction of an iron carbonyl intermediate 2 [Fe−CO] + to 1 [Fe−CO] 0 can be eliminated based on the predicted redox potential which is more negative than −2.0 V vs Fc/Fc + . A similar Fe(II)-carbonyl species equipped with a comparable terpyridine and pyridyl-N-heterocyclic carbene based ligand system was synthesized and structurally characterized by 1 H NMR and single crystal X-ray diffraction by Miller and coworkers. 63 Chemical reduction with two equivalents of decamethylcobaltocene resulted in the dissociation of the pyridine ligand and the formation of a pentacoordinate, low-valent iron carbonyl complex that was unfortunately shown to undergo rate limiting endergonic release of CO. The hemilability of the pyridyl arm to generate a stable 18-electron complex may be a defining feature to explain the difference in catalytic activity and speaks to the potential advantages engendered by utilizing a homoleptic tpyPY2Me ligand system. High Overpotential Pathway. The application of more reducing potentials beyond −1.7 V vs Fc/Fc + results in the formation of a second catalytic wave. A mechanism for this high overpotential regime is illustrated in Figure 5 and presented in detail in this section. Here we investigate three potential pathways and found two to be aligned with experimental findings.</p><p>In all three cases, the initial reduction steps are identical in redox potential and electronic structure to what we presented for the low overpotential pathway discussed above. However, the first electron transfer is off-pathway; where at more reducing potentials, we find that there is enough driving force to reduce the carboxy intermediate</p><p>) with a predicted redox potential of −1.83 V vs Fc/Fc + (vide infra). This observation results in catalytic turnover occurring from the singly reduced, iron complex ( 1 [Fe] + ) rather than turnover from the unreduced 1 [Fe] 2+ . The three mechanistic pathways explored below diverge following the formation of the 1 [Fe] 0 catalytic resting state.</p><p>The proposed mechanism is depicted in Figure 5. In the first two scenarios, from the 1 [Fe] 0 intermediate, CO2 coordination (Figure 5; outer cycle) or protonation (Figure 5; inner cycle) can occur first as discussed above for the low overpotential regime. Both pathways are uphill by approximately 11 kcal/mol and barrierless. The two pathways converge at the formation of the CO2H adduct, 1 [Fe−CO2H] + , that can be further reduced at a calculated applied potential of −1.83 V vs Fc/Fc + yielding 2 [Fe−CO2H] 0 where the electron populates one of the low-lying tpy-π * orbitals. The second proton transfer step to lose water and yield the carbonyl intermediate, 2 [Fe−CO] + , is strongly exergonic by −17.5 kcal/mol (Figure S7). Moreover, the activation barrier for the second protonation is 7.3 kcal/mol, which is 4 kcal/mol lower than the barrier for the C−O bond cleavage step in the low overpotential regime. The CO release becomes endergonic by 4.2 kcal/mol due to the excess electron density that is transferred from the tpy-π * orbital to the Fe−CO moiety thus strengthening the backbonding to the CO ligand. Hence, the additional reductive event significantly lowers the barrier of the second protonation at the cost of more difficult CO release. Therefore, the second protonation is not as critical as in the low overpotential regime. Depending on the pathway, either the addition of CO2 (12.1 kcal/mol) or the protonation of 1 [Fe] 0 (11.5 kcal/mol) become critical Figure 6. Free energy landscape for the low overpotential (Figure 3) and high overpotential (Figure 5) CO2RR regimes. Colors signify different possible pathways that were proposed in the catalytic cycles shown in Figures 3 and 5. From the doubly reduced 1 [Fe] 0 intermediate, the blue line signifies the CO2 binding first pathway, the red line denotes the protonation first pathway, and the green plots the PCET pathway. The purple line indicates the pathway where the 1 [Fe-CO2H] + intermediate is further reduced in the high overpotential regime (refer to Figure 5). For the reduction steps, a potential of −1.8 V vs Fc/Fc + is applied; for the protonation steps, PhOH is used to estimate barriers and carbonic acid for free energies (see main text for justification); solid lines correspond to intermediate states and dashed lines to transition states. steps as well. Unfortunately, it is not possible to distinguish those two pathways computationally due to the small relative energy difference of 0.6 kcal/mol and the error associated with the predictions of protonation and CO2 addition. It is possible that both channels are populated.</p><p>In a third scenario, following the formation of 1 [Fe] 0 , a proton coupled electron transfer (PCET) pathway is possible where one of the pyridyl arms is protonated yielding 2 [Fe−NPYH] 0 . The additional electron is localized in the pyridinium moiety; thus, the electronic structure can be described as a doubly reduced tpy-π * moiety coupled to the intermediate spin iron center and a singly reduced pyridinium; this structure is summarized in the spin density plot in Figure S8. The reduction potential is acid dependent; using H2CO3 the calculated reduction potential is -2.14 V vs. Fc/Fc + . However, taking non-standard concentrations into account using a similar approach to Song and coworkers, 55 we compute a shift by 0.3 V resulting in a potential of -1.84 V. For reference using H + the computed reduction potential is -1.17 V vs Fc/Fc + . Next, CO2 can bind to 2 [Fe−NPYH] 0 to directly yield 2 [Fe−CO2H] 0 in an intramolecular proton and electron transfer step with a free energy of −12.0 kcal/mol and thus also "skipping" the high energy CO2 adduct intermediate. The role of proton transfer during catalysis was examined by measuring the H/D kinetic isotope effects in both catalytic regimes using C6H5OH (PhOH-H6) or C6D5OD (PhOH-D6) as the proton source (Figure S9). Experimentally, normal primary H/D kinetic isotope effects were observed under both catalytic regimes (kH/kD of 1.59 and 1.22, respectively), suggesting that proton transfer is involved in the rate determining step, thus supporting the protonationfirst pathways, and excluding the CO2-first mechanism. Kinetic Analysis of Proposed Mechanistic Pathways. The free energy landscape summarizing the total catalytic cycle is depicted in Figure 6. Following elucidation of the possible reaction pathways for CO2 reduction across both potentialdependent regimes, we next sought to compare experimentally determined kinetic data against computationally derived turnover frequencies (TOFs) obtained from the energetic span computational model developed by Kozuch and coworkers. 64 The energetic span model was employed to identify key intermediates and transition states, which control the rate of catalysis. This model connects the free energy landscape of the DFT based catalytic cycles with the experimentally measured TOFs using Eyring transition state theory. The model identifies TOF-determining transition states (TDTS) and the TOF-determining intermediates (TDI) and computes rates based on the energetic span of these two states. 64 Often there are various intermediates with a significant influence on the catalytic rates. This progression can be quantified in this model by the degree of TOF control (denoted as XTOF), 65 which describes how the TOF varies by a small change in energy of that intermediate or transition state. The range of XTOF is between 0 and 1, where 0 denotes that the species has no influence on the TOF and 1 denotes that the species solely controls the TOF. 66,67 The results from the energetic span calculations are presented in Tables S1 -S6.</p><p>Experimental kinetic parameters were obtained from variable controlled potential (CPE) experiments as described by Savéant and co-workers. 49,50 Short-term (5 min; Figure 7) and long-term (1 h; Figure S10 and S11) CPE experiments were performed where the products (CO and H2) were detected and quantified by gas chromatography for the long-term CPE experiments conducted in an air-tight electrochemical cell. The observed rate constants (kobs) at each applied potential were extracted from the average specific current densities taken across the entire electrolysis experiment and were then compared to the computed rates obtained from the energetic span model (Figures S10 and S11 and Tables S1-S6). To further probe the influence of the proton transfer step on the observed rates, PhOH (see Figure 7), and Cl-PhOH (see Figures S10 -S12), which is approximately one pKa unit more acidic (in water), were utilized.</p><p>First, we explored the kinetics of the low overpotential regime (potential window of −1.40 − −1.75 V vs Fc/Fc + ). Short-term CPE data collected at applied potentials between −1.42 and −1.72 V vs Fc/Fc + overlay closely with the voltammogram collected under CO2 atmosphere with 1 M PhOH (Figure 7a). Current densities were stable across the entire short-term electrolysis (Figure 7b) and long-term electrolysis (Figures S10 and S11) and reached a plateau at ca. −1.66 V vs Fc/Fc + with an average current density of 1.53 mA/cm 2 which corresponds to a TOF of 1.74 x 10 5 s -1 (Figure 7c). When the more acidic, Cl-PhOH was added as the proton donor, the maximum TOF of the first catalytic regime increased to 5.50 x 10 5 s -1 , a three-fold increase in observed rate relative to the data obtained with PhOH as the proton source (Figure 7d) without loss of product selectivity (Figure S13). Analysis of this catalytic regime with the energetic span model shows that the catalytic rate is solely controlled by the second protonation step (C−O bond cleavage to release water) as the key rate-limiting intermediate and transition state are both associated with the second protonation. The carboxy intermediate, 1 [Fe−CO2H] + , is the TDI and the second protonation barrier is the TDTS both with XTOF values close to 1. Based on this analysis, a TOF of 400 s -1 is predicted. This value is slower than the experimentally determined TOF of 1.74 x 10 5 ; however, this translates to an energy difference of ∼3.5 kcal/mol which is still within acceptable agreement. Furthermore, when Cl-PhOH is utilized as the acid source, the barrier is lowered by 1.1 kcal/mol which translates to a rate increase by one order of magnitude. The stronger acid lowers the TDTS, leading to an increase in the importance of the CO2 binding step. This shift is illustrated by the change in the degree of TOF control: Indeed, the XTOF of both the TDI ( 1 [Fe−CO2H] + ) and TDTS Figure 8. Free energy landscape of the low overpotential regime for CO2RR and HER pathways. For the reduction steps, a potential of -1.8 V vs Fc/Fc + is applied. For the protonation steps, PhOH is used to estimate barriers and carbonic acid for free energies (see main text for justification). Solid lines correspond to intermediate states and dashed lines to transition structures. Black, blue, and red lines denote the uses of PhOH, Cl-PhOH, and NO2-PhOH, respectively. decreases from 0.97 to 0.82 and the CO2 binding step is increasingly important for the rate as the XTOF increases to 0.16 from 0.03. Distinction between the protonation first or CO2 binding first pathways by this kinetic analysis was not possible because both options occur before the rate determining protonation step.</p><p>Application of more negative onset potentials between −1.81 and −2.01 V vs Fc/Fc + allowed us to probe the kinetics of the second, high overpotential regime. Tight correlations were observed between the averaged current densities from the CPE experiments and the cyclic voltammograms collected under CO2 atmosphere with PhOH (Figure 7a) or Cl-PhOH (Figure S12a). Peak current density of 4.0 mV/cm 2 was reached at −2.01 V vs Fc/Fc + which corresponds to a maximum TOF of 1.05 x 10 6 s -1 . Addition of 1 M Cl-PhOH as the proton source results in a smaller rate enhancement with a maximum TOF of 1.5 x 10 6 s -1 , representing only a 1.4-fold enhancement relative to PhOH, about half of what was observed in the low overpotential regime (Figure 7d and S13). Catalytic Tafel plots comparing the two acids are presented in Figure 7d. In addition, there is still a proton dependence on the rate limiting step as illustrated by the normal, primary H/D kinetic isotope effect.</p><p>Next, the energetic span model was applied to predict the rates for the three possible mechanistic pathways proposed for the high overpotential regime. These rates are then compared to the experimentally determined TOFs. The three pathways are given in Figure 5 and are illustrated in a free energy diagram (Figure 6) to allow for more direct comparison. From the 1 [Fe] 0 catalytic resting state, the iron complex can either go through: (1) a CO2 coordination first pathway to give the CO2 complex ( 2 [Fe−CO2] 0 ); (2) a protonation first pathway to generate the cationic pyridinium intermediate, 1 [Fe−NPYH] + ; or a PCET pathway to generate a similar neutral pyridinium species, 2 [Fe−NPYH] 0 .</p><p>A TOF of 4000 s -1 is predicted for the CO2 binding first pathway. The TOF is completely controlled by the CO2 coordination step as the reduction of 1 [Fe−CO2H] + to 2 [Fe−CO2H] 0 lowers the (previously rate limiting) barrier for the second protonation by 6.6 kcal/mol from 13.9 to 7.3 kcal/mol. This shrinks the XTOF for both 1 [Fe−CO2H] + and the transition state for the second protonation to 0. The CO2 first pathway can be eliminated because CO2 coordination is shown by the energetic span model to be entirely rate limiting and thus fails to explain the experimentally observed normal, primary H/D kinetic isotope effect and a large catalytic enhancement with the addition of more acidic proton donors.</p><p>Next, for the protonation first pathway, we predict a TOF that is almost 3 orders of magnitude slower than what we observed experimentally (11000 s -1 ). Additionally, we find that the cycle is solely controlled by the first protonation of 1 [Fe] 0 to give the pyridinium intermediate, 1 [Fe−NPYH] + . The protonation first pathway would account for the primary kinetic isotope effect but not the rate enhancement upon addition of stronger acids as the energetic span model predicts the same rate regardless of if PhOH or Cl-PhOH are utilized as the proton source.</p><p>The TOF predicted for the PCET pathway of 1.4 x 10 7 s -1 resulted in the closest match to the experimentally measured value of 1.05 x 10 6 s -1 with the pathway being controlled by both the PCET step as well as the second protonation of 2 [Fe−CO2H] 0 as both have XTOF coefficients close to 0.5. The para-chloro substituted PhOH, Cl-PhOH, lowers the barrier for the second protonation by 2.6 kcal/mol. However, only the PCET pathway is kinetically controlled by this transition state increasing the TOF from 1.4 x 10 7 s -1 to 2.6 x 10 7 s -1 . This increase is smaller than the increase in the low overpotential regime despite a stronger effect on barrier lowering, which can be understood by the lower XTOF of that step in the low versus high overpotential regime. Based on the experimental and computational findings, these data suggest that the PCET pathway is the more likely mechanism for the high overpotential regime. Selectivity Versus HER. Achieving selectivity for CO2 reduction over the reduction of protons to H2 is critical. The HER is a competitive side reaction across a similar potential window to the CO2RR and is highly dependent on the presence and strength of the proton source. 14 The generation of an iron hydride under reducing conditions in the presence of a proton donor can thus shift the catalyst selectivity away from CO production and toward the formation of H2. It is therefore important to understand the kinetic and thermodynamic barriers associated with the protonation of the reduced iron center.</p><p>In our proposed pathway there are two critical intermediates through which the formation of a hydride is feasible (Figure 8 and S15). The first possibility is the direct protonation of the iron center of the doubly reduced intermediate, 1 [Fe] 0 (Figure 8 and S16a). The other possibility is the rearrangement of the pyridinium intermediate, 1 [Fe−NPYH] + , to a metal hydride (Figure 8 and S16b). In both cases, high activation barriers prevent these side reactions and explain the high product selectivity for CO2 reduction to CO that is observed experimentally. The direct formation of a hydride 1 [Fe−H] + from 1 [Fe] 0 albeit thermodynamically favorable with a free energy of −6.1 kcal/mol, is kinetically inhibited with a high activation barrier of 24.5 kcal/mol. The rearrangement of 1 [Fe−NPYH] + to 1 [Fe−H] + , is more exergonic (−17.5 kcal/mol); however, this pathway is also inhibited by a larger activation barrier of 21.4 kcal/mol. In both cases, the high barriers can be rationalized by the electronic structure of 1 [Fe] 0 : the metal center stabilizes the two ligand reductions through antiferromagnetic coupling and thus the metal center remains Lewis acidic. This feature is illustrated by the coordination of acid (e.g., PhOH or H2CO3) to 1 [Fe] 0 in which the acid prefers to bind via the oxygen atom to the iron center which then facilitates the protonation of the pyridyl arm, but blocks the metal center from protonation (see Figure S17). Therefore, the formation of a metalhydride is prohibited by steep kinetic barriers ultimately shutting down pathways to hydrogen production. The difference between the rate limiting transition state in CO2RR and the barrier for hydride formation is ∼10 kcal/mol which implies that 1 [Fe] 2+ should remain selective for the CO2RR even when stronger acids are used as proton donors. Indeed, selectivity is achieved by using 4-nitrophenol (NO2-PhOH), an even stronger acid (two pKa units stronger than Cl-PhOH), with a barrier of 19.3 kcal/mol for the formation of a hydride. We explored this possibility, and, in all cases, we do not detect any hydrogen formation experimentally. However, NO2-PhOH was also redox-active at the potentials applied, which also prevented the formation of CO. Rational Catalyst Design Driven by Mechanistic Insights. Based on the mechanistic studies we identified several critical intermediates and transition states that directed us to propose improved ligand designs that ideally stabilize several critical intermediates. The key takeaways from the kinetic analysis are: first, the low overpotential regime is solely controlled by the second protonation step, even when strong acids are used. Thus, modifications must either lower the energy of the transition state or decrease the stability of the carboxy intermediate. Second, the high overpotential regime is mainly controlled by either the CO2 adduct or the formation of the pyridinium intermediate, especially when stronger acids are used as the proton source. Thus, an optimal modification should have a positive effect on both the CO2 binding and second protonation. On the basis of our detailed mechanistic investigation, we propose rationally designed synthetic modifications to the tpyPY2Me ligand to further increase the catalytic performance of the iron complex. Synthetic modification to the pyridine moiety is the most promising for two reasons: first, it does not affect the initial reduction potentials and the crucial antiferromagnetic coupling of 1 [Fe] 0 ; second, substituents can affect all critical intermediates by stabilizing the bound CO2 adduct, increasing the pKa of the pyridine itself for protonation and PCET, and then stabilizing the transition state for the second protonation (Figure 9 shows the investigated candidates). In order to gain quantitative insights, we use the absolutely localized MO energy decomposition analysis with continuum solvation (ALMO-EDA(solv)) 67 (see SI for a short introduction and a more detailed review is also available 68 ) to understand how a specific modification stabilizes the reduced CO2 adduct. We employ a difference-of-differences-approach where we compare the change in interaction energy and EDA terms using the unsubstituted tpyPY2Me ligand as the reference. We investigate hydrogen bonding (−OH and−NH2) and ionic stabilization (−N(CH3)3 + ) as both moieties are known to facilitate CO2 binding in other CO2RR catalysts (see Figure 9). 12,19,22,30,31,69 The EDA decomposition of that change in interaction energy is crucial for understanding the exact stabilization pathway as substituents affect the reduced CO2 not only directly via a non-covalent interaction but also indirectly through substituent effects. Both pathways often have distinct EDA fingerprints and thus can be distinguished by an EDA scheme. We recently showed the viability of this approach for an iron tetraphenylporphyrin catalyst. 67 The chosen reference fragmentation of the complex is a doubly reduced CO2 2-and an unreduced catalyst. This fragmentation is more suitable for this type of complex as the O−C−O bond angle of 124º indicates a transfer of both electrons into the CO2 moiety. The alternative choice, a neutral but bent CO2 fragment and a doubly reduced metal complex will be solely dominated by the geometry distortion term as discussed elsewhere. 67,68 Thus, this interaction energy decomposed by the EDA scheme corresponds to the stabilization of the doubly reduced CO2 2-dianion by the unreduced catalyst. It is important to point out that this is just a part of the total free energy of CO2 addition. The EDA results for the unsubstituted complex are given in Figure 10a and shows that the attractive interactions are dominated by electrostatic interactions with significant charge transfer contributions. The short Fe−C bond distance of 2.05 Å rationalizes the high Pauli repulsion, and indicates that net binding energy results mainly from CT. The change in interactions energies as well as each EDA component for each substituent relative to the unsubstituted complex is depicted in Figure 10b. The ortho-hydroxy substituent (o-OH) strengthens the interaction by 77.1 kJ/mol (18.5 kcal/mol). The main driving force is additional favorable electrostatic interaction, which is also supplemented by favorable contributions from polarization and charge transfer, which are typical of a hydrogen bonding fingerprint. 70 The large increase in Pauli repulsion can be attributed to the repulsion of the diffuse lone pairs of the CO2 moiety and the hydroxy group. Interestingly, this intermediate was not stable without freezing the hydroxy OH bond as it otherwise directly protonates the CO2 moiety. The ortho-amino substituent has a much smaller stabilizing effect of −22.5 kJ/mol (5.4 kcal/mol). The favorable electrostatic interaction cannot overcome the additional Pauli repulsion of the amino substituent and the CO2 -2 moiety. Lastly, the charged trimethylanilinum (TMA) moiety had to be placed at the meta position due to the bulkiness of the group. The EDA results demonstrate how this group provides a purely electrostatic stabilization; however, the solvation screens most of the interaction to yield an overall stabilization of −44.6 kJ/mol (−10.7 kcal/mol). Moreover, the ortho hydrogen donating substituents have another advantage: They can stabilize the transition state for the second protonation step by forming hydrogen bonds. These interactions lower the transition state (see Figure S18) for the o-OH by 1.3 kcal/mol which translates to an order of magnitude faster TOF in the low overpotential regime (similar effect to the use of stronger acids).</p><!><p>In summary, inspired by the excellent observed catalytic performance of the molecular [Fe(tpyPY2Me)] 2+ system, we investigated the mechanistic pathways through which it electrochemically converts CO2 into CO. Cyclic voltammograms collected under CO2 atmosphere displayed the formation of two distinct catalytic regimes with 𝐸 𝑐𝑎𝑡/2 1 of −1.43 V vs Fc/Fc + (ηTOF/2 = 160 mV) and 𝐸 𝑐𝑎𝑡/2 2 −1.86 V vs Fc/Fc + (ηTOF/2 = 590 mV). For the low overpotential regime, the computed pathway shows that 1 [Fe] 2+ first undergoes two, single-electron reduction steps to generate the five-coordinate open-shell singlet, 1 [Fe] 0 . From 1 [Fe] 0 , we show that the order of CO2 addition or protonation is flexible and experimentally indistinguishable with both steps having comparable calculated barriers. We find that both CO2 binding and protonation of the pyridine arm are feasible steps with similar endergonic free energies. In any case, the rate limiting step was found to be the second protonation step resulting in cleavage of the C−O bond with subsequent CO release being barrierless and exergonic. Analysis of the complete low overpotential regime revealed that catalysis is solely controlled by the carboxyl intermediate ( 1 [Fe−CO2H] + ) and the transition state for the second protonation step. Consequently, the use of a stronger proton source, such as Cl-PhOH, resulted in a three-fold increase in the rate of catalysis without any loss in product selectivity. Computational analysis of the high overpotential regime shows a similar two initial electron transfer steps; however, at more reducing potentials, the first reduction is off-pathway, allowing for turnover from 2 [Fe] + rather than 1 [Fe] 2+ as is observed in the low overpotential regime. Following formation of the catalytic resting state, 1 [Fe] 0 , we computationally probed three mechanistic pathways and compared them to experimental kinetic data and results obtained from the energetic span model. From this analysis, we identified the PCET pathway of one of the pyridine arms as the most consistent mechanism. This intermediate readily binds CO2 and rearranged into a singly reduced carbonyl intermediate. The additional electron greatly facilitates the second protonation step explaining the increased rates observed for the higher overpotential regime. Analysis of the cycle showed that both the PCET and the second protonation steps control the rate of catalysis. As such, similar to the low overpotential regime, the use of a more acidic proton source increases the rate of catalysis, but to a lesser degree. The stronger acid only lowers the barrier for the second protonation; however, this step has less influence on the turnover frequency and thus the effect of acid pKa is lower.</p><p>Finally, the mechanistic insights gained from this study identified the pyridyl arms as promising targets for further optimization of the catalytic performance of this system. The pyridyl arms are not involved in stabilizing the excess electron density in the reduction events and thus any modifications should not alter the reduction potential of 1 [Fe] 2+ , which is already optimally matched to the standard redox potential for the conversion of CO2. The introduction of hydrogen bond donors into the second coordination sphere can further greatly stabilize the CO2 adduct and lower the transition state for the second protonation step, both of which are critical intermediates for fast catalysis. Moreover, we also expect that the substituent effects can help to stabilize the pyridinium intermediate; therefore, this synthetic modification has the potential to improve catalysis in all proposed cycles for both the low and high overpotential regimes.</p><p>SYNOPSIS TOC: [Fe(tpyPY2Me)] 2+ ([Fe] 2+ ) is a homogeneous electrocatalyst for converting CO2 into CO at low overpotentials with high selectivity and fast rates. Here we report a combined computational and experimental study that establishes two distinct mechanistic pathways for electrochemical CO2 reduction catalyzed by [Fe] 2+ as a function of applied potential. Determination of mechanistic pathways is then used to direct the computational exploration of improved ligand framework design by energy decomposition analysis.</p>

ChemRxiv

Theoretical study of the microhydration of 1-chloro and 2-chloro ethanol as a clue for their relative propensity toward dehalogenation

This work reports a computational analysis of hydrogen bonded clusters of mono-, di-, tri-and tetra hydrates of the chlorohydrins CH 3 CHClOH (1ClEtOH) and CH 2 ClCH 2 OH (2ClEtOH). The goal of the study is to assess the role of the water solvent into the facilitation of the initial step for dehalogenation of these compounds, a process of interest in several contexts. Molecular orbital methods (MP2), density functional methods (B3LYP, M06 and B97X-D) and composite model chemistries (CBS-QB3, G4) were employed to investigate the structure, electronic distribution and hydrogen-bonded structure of 7 monohydrates, 6 dihydrates, 5 trihydrates and 5 tetrahydrates of both species. Standard reaction enthalpy and standard Gibbs free reaction energy (∆ 𝑟 𝐺 298 0 ) were computed for all aggregates with respect to n independent water molecules and with respect to the dimer, trimer and tetramer of water, respectively, in order to evaluate stability and hydrogen bonding network. The influence of the water chains on the length and vibrational frequencies, especially of the C-Cl and O-H bonds, was evaluated.

theoretical_study_of_the_microhydration_of_1-chloro_and_2-chloro_ethanol_as_a_clue_for_their_relativ

| INTRODUCTION<!>| COMPUTATIONAL METHODS<!>| Monomers<!>Figure 2. IR spectra of 1ClEtOH (upper panel) and 2ClEtOH (lower panel)<!>Figure 5. Isovalue level curves of the laplacian of the electronic density for the most stable complexes of 1ClEtOH and 2ClEtOH with 1 to 4 water molecules.<!>| Energetics of the complexation<!>Figure 6. Plot of the enthalpy of reaction at 298.15 K (in kcal/mol) according to equation (1) (upper panels) and equation (2) (lower panels) for 1ClEtOH (left) and 2ClEtOH (right), with respect to the number of water molecules n<!>TABLE 2<!>| IR spectra<!>| CONCLUSIONS

<p>Halogenated hydrocarbons are organic species of considerable interest in environmental sciences, for instance in atmospheric chemistry where the role of chlorofluorocarbons (CFCs) in the depletion of the ozone layer [1-3] is extremely important. On the other side, absorbable organic halogen (AOX) species produced in the course of chlorine or chlorine dioxide bleaching of pulp in the Kraft process [4,5], and some byproducts of water disinfection [6][7][8], contribute to surface water contamination. Other products used in the industry or agriculture further contribute to the presence of halogenated hydrocarbons in the environment.</p><p>One of the simplest halogenated halocarbon families is that of halohydrins (halogenated alcohols) and one of their important reactions is dehalogenation, i.e. HX elimination, which can occur by pyrolysis in gas phase [9][10][11], and by base catalysis [12][13][14] or oxidation [15] in solution. Enzymes that produce this dehalogenation process (known as dehalogenases [16] and dehaloperoxidases [17]) are also present in numerous organisms.</p><p>Leaving aside the pyrolysis, all other mechanisms occur in aqueous solutions. For that reason, we wanted to investigate what is the role of the molecules of the solvent on the initial species to be dehalogenated. In this work, we present a study of the geometrical structure and energetics of clusters of 1ClEtOH and 2ClEtOH with n water molecules (n=1-4). While 2ClEtOH has been studied several times and is well known, much less information exists in the literature about the 1ClEtOH isomer, which nonetheless may play some important role in the dehalogenation path of 2ClEtOH.</p><p>We show here how the water chains influence the charge distribution on the solutes and the equilibrium distance of the C-Cl bond to be broken in the dehalogenation reaction. Studying the dependence of the stabilization energy with the number of water molecules, we show that a cluster of the solute with three water molecules is sufficient to single out the most important effects of the solvent.</p><!><p>Several computational methods have been used to study the clusters of 1ClEtOH and 2ClEtOH with n water molecules (n=1-4). The simple correlated molecular orbital (MO) method MP2 [18] was used to obtain optimum geometries, energies and frequencies. Composite methods, which rely on relatively simple calculations that are later stepwise corrected for extension of the basis set to the complete basis set (CBS) limit, for higher levels of correlation energy (MP4, CCSD(T)) and add empirical factors to correct for dissociation energies with respect to atoms, were also employed. In particular, in this paper we used the CBS-QB3 [19,20] and G4 [21] methods, which are all approximations, increasingly accurate, to CCSD(T)/CBS calculations, which are not feasible on molecules of this size. Finally, three density functional theory (DFT [22]) procedures were employed, namely B3LYP [23], M06 [24], and B97X-D [25] methods. B3LYP is the oldest adiabatically connected (or hybrid) generalized-gradient method, which is able to afford reasonable geometries and bond energies accurate to about 4 kcal/mol between main-group atoms and about 1.3 kcal/mol for noncovalent interactions. M06 is a member of the Minnesota group of functionals developed by Truhlar`s group. It includes the kinetic energy functional and reduces the error in main-group bond energies to 1.8 kcal/mol and the error in noncovalent interactions to 0.4 kcal/mol. Finally, B97X-D is a last generation functional developed by the group of Head Gordon, which additionally includes corrections for dispersion and long-range interactions. While mostly B97X-D and M06 behave similarly, the former performs better on the noncovalent interactions (like the hydrogen bonds explored in this paper).</p><p>It is a known fact that DFT methods do not have such a heavy dependence on the quality of the basis set as MO methods have. However, when low stabilization energy hydrogen-bonded clusters are studied, it is reasonable to have an appraisal of the situation employing different basis sets. In our case, we chose several Pople's basis sets [18], namely 6-31G(d), 6-31++G(d,p), 6-311++G(d,p), 6-311++G(2d,2p) and 6-311++G(3df,2pd) as examples of relatively low-cost (i.e. less complete) sets. Additionally, we used the correlation consistent basis sets of Dunning [18], cc-pVTZ, aug-cc-pVTZ, cc-pVQZ and aug-cc-pVQZ as examples of more complete (and costly) basis sets. In both cases, the rationale was to extend smoothly the valence, diffuse and correlation spaces of the sets to approach the CBS limit.</p><p>Several properties of the clusters were studied using the methods previously described. Optimum geometries of the clusters were obtained with all the methods. 1, 2, 3 and 4 water clusters were built incrementally, first looking for all the possible n-1 isomers and then adding the n-th water molecule in all possible positions that kept the water in contact with the solute (i.e. all waters considered were in the first solvation shell). These structures are, of course, only snapshots of the dynamical behavior of next neighbor water molecules in a real solvated chloroethanol. Second derivatives of the energy with respect to the geometrical coordinates were calculated at all levels and the IR and Raman spectra of the clusters were approximated using harmonic frequencies. Differential spectra of the solvated species with respect to the unsolvated ones were calculated to assess the influence of the water molecules in the spectra. Thermochemical properties were evaluated at standard conditions using the harmonic-rigid rotor approximation and the usual formulas of statistical thermodynamics. The charge distribution was analyzed calculating the electrostatic potential and point charges derived from a Natural Bond Orbital (NBO) [26] analysis, on one side, and derived from the electrostatic potential (HLY method [27]) on the other. This allowed us to calculate the amount of charge transfer to the water portion in each of the clusters. Finally, the net of hydrogen bonds in the clusters were investigated calculating the laplacian of the electronic density on different planes which include the hydrogen bonds to be investigated.</p><p>All calculations were performed by using the Gaussian 09 set of computer codes [28].</p><!><p>Optimized structures for each of the two monomers 1ClEtOH and 2ClEtOH, obtained with each method using the larger basis sets, are shown in Fig. 1. Full geometries for each species at each level of calculation are reported in the Supplementary Information. While to the extent of our knowledge there is no experimental determination of the structure of 1ClEtOH, two experimental studies have determined the structure of 2Cl2EtOH. Azrak and Wilson [29] analyzed the microwave (MW) spectrum, while Almemmingen et al [30] performed a study using electron diffraction (ED) to determine the anti-gauche ratio as a function of temperature. Both studies are reasonably in agreement with respect to the C-C, C-O and O-H bond distances, as well as with respect to the dihedral angle between the ClCC and CCO planes, but differ on the result for the C-Cl distance. While Azrak and Wilson found 1.7886 ± 0.0038 Å for the latter, Almemmeingen et al reported a value of 1.801 ± 0.001 Å. Although the discrepancy is not very large, it is significant. One must assume that the ED data is more accurate, since the Cl is the heaviest atom in the molecule and the determination of the C-Cl distance should be the better determined structural parameter in this study. Durig et al [31] gave an explanation of the biased value in the microwave experiment.</p><p>From the theoretical point of view, we did not find in the literature high level calculations (i.e. CCSD(T) or better) of the geometry of 2ClEtOH. Former calculations used normally HF, B3LYP or MP2 with at most 6-311++G(d,p) or aug-cc-pDZ basis sets, which are superseded by the calculations in this paper. In all cases the gauche Gg' structure [31] is the most stable one (a decomposition analysis of the interactions was performed by Baranac-Stojanović et al [32] who determined the effect to be caused mainly by electrostatic attraction). The dihedral angle between the ClCC and CCO planes is well reproduced by all the calculations. DFT results for the CCl bond are nearer to the ED data than those of MP2, these being nearer to the MW experimental results than to those obtained by ED. Two aspects in which the calculation fail are the OH bond length, that theoretically is predicted much shorter than the experimental value, and the hydrogen bond Cl-H, which the calculations predict longer than the experimental one. In this specific case, however, MP2 performs better than DFT. Both theoretical drawbacks can easily be explained by the fact that our calculations do not include anharmonicity corrections.</p><p>1ClEtOH and 2ClEtOH geometries clearly exhibit the differences caused by the different location of the chlorine atom. The C-C distance is larger by about 0.008 Å and the C-O bond is larger by about 0.027 Å in 2ClEtOH with respect to 1ClEtOH. On the contrary, the C-Cl bond is about 0.038 Å shorter. The results of these modifications are that the distance between chlorine and the H atom in the hydroxyl group is 0.081 Å shorter in 2ClEtOH than in 1ClEtOH, implying a stronger interaction in the former than in the latter. Also included in Fig. 1 are the laplacian of the density for both molecules calculated at the B97X-D/cc-pVTZ level. One can see that in neither of those diagrams is there a critical point between the H in OH and chlorine, discarding the presence of a hydrogen bond, thus confirming Baranac-Stojanović et al [32] analysis that the attraction is mainly electrostatic. The charge separation between the negative chlorine and the positive hydrogen (0.56 electrons according to the NBO analysis at the B97X-D/aug-cc-pVQZ level) also supports this view.</p><p>Finally, in Fig. 2 are included the IR spectra of both 1ClEtOH and 2ClEtOH calculated at the B97X-D/cc-pVTZ level. The most important data, for the purpose of this study, is the position of the C-Cl and O-H stretchings, because they will be the most affected by the solvating water molecules. In Table 1 is shown a compendium of the position of the bands with the different chemical models, as well as the ratio of the intensities of the CCl stretching band to the OH stretching band. Two characteristics of these bands can be noticed in Table 1. First, that there is an inverse shift of both bands. While the  str (C-Cl) frequency is smaller in 1ClEtOH than in 2ClEtOH, the  str (O-H) frequency is larger. Although the exact value varies a little with the method and the basis set, it is of about 40 cm -1 for the former and about 20 cm -2 for the latter. Also, the ratio of the intensities is larger than one for 1ClEtOH and smaller than one for 2ClEtOH, which may serve as a characteristic indicator to indicate the presence of one or the other isomer. This variation may be traced to the charge separation between the C-Cl and C-O atoms (and, therefore, the bond dipoles). The charge separation between the atoms in the O-H bond is approximately the same in 2ClEtOH (1.17 according to NBO) than in 1ClEtOH (1.16). It does not happen the same however in the atoms of the C-Cl bond. The charge separation in 1ClEtOH is actually about 70% larger than in 2ClEtOH (0.50 against 0.30 according to the NBO analysis). Consequently, the intensity of the C-Cl stretch in 1ClEtOH is larger than in 2ClEtOH and this is part of the explanation of the inversion of intensities observed in the spectra. 3 shows the structure of the clusters identified for 1ClEtOH and Fig. 4 shows the general structure of the water clusters of 2ClEtOH. Important geometrical parameters are shown in the figures at the B97X-D/cc-pVTZ level. Full geometries are given in the Supplementary Information section.</p><!><p>Four different monohydrated complexes were found for 1ClEtOH (1_1W_A, _B, _C and _D) while only two were found for 2ClEtOH (2_1W_A and _B). Letters were assigned in order of decreasing stability, with the A complex being the more stable one (see later on). The isomers 1_1W_A to 1_1W_C and 2_1W_A have interactions of the water molecule with both the chlorine atom and the OH residue, while 1_1W_D and 2_1W_B exhibit only an interaction of the water with the OH fragment. Moreover, while in 1_1W_A, 1_1W_B and 2_1W_A the OH group acts like a hydrogen donor, in 1_1W_C, 1_1W_D and 2_1W_B it acts as a hydrogen acceptor.</p><p>Isomers 1_1W_A and 1_1W_B are essentially equivalent. The only difference is whether the OH bond is syn or anti to the CH bond. The Cl-H hydrogen bond is clearly equivalent and the OH-O hydrogen bond is slightly tighter in 1_1W_A. 1_1W_C exhibits the peculiarity of an attractive interaction between the H in the CH group and the oxygen of water, besides the other two interactions with the Cl and the OH group. An important point to notice is that the more stable monohydrate is also the one that presents the larger elongation of the CCl bond with respect to the isolated 1ClEtOH. In the case of 2ClEtOH it happens something similar, 2_1W_A is more stable than 2_1W_B and exhibits a larger increase in the C-Cl bond length. Notice however that this increase is much shorter than in the 1ClEtOH corresponding complexes. Another important piece of data is the shortening of the Cl-H distance in 1_2W_B. The simple presence of a water bound to the hydroxyl group is enough to move the chlorine nearer to the hydrogen. This correlation between stability and increasing in the CCl bond length is a hint that water may be helping the abstraction of HCl, most noticeably in the case of 1ClEtOH. Four complexes of 1ClEtOH with two water molecules were found. As seen in Fig. 3, the four complexes present some form of the water dimer bound in different ways to the solute. 1_2W_A and 1_2W_B present interactions of the water dimer with both chlorine and the hydroxyl group. A true Cl-H hydrogen bond is noticed in these cases and the OH group works as a hydrogen donor. 1_2W_C instead shows the solvation of the hydroxyl group with two water molecules, and the interaction with chlorine is weaker, more an electrostatic attraction than a true hydrogen bond. In this case, the hydroxyl group acts both as a hydrogen donor and acceptor. Finally, 1_2W_D exhibits a hydrogen acceptor OH and the second water molecule interacts with the acidic hydrogen in the CH bond. This is a true hydrogen bond, although weaker than those with the O-H-O structure. In the case of 2ClEtOH only two complexes were identified. 2_2W_A presents the water dimer as a hydrogen bonded chain between the chlorine and the hydroxyl group, which in this case works as a hydrogen donor. 2_2W_B exhibits the water dimer solvating the hydroxyl group, both a hydrogen donor and a hydrogen acceptor in this case, but also an interaction with chlorine. In this case the hydrogen bond is weaker than in 2_2W_A but it is nonetheless a true hydrogen bond.</p><!><p>The complexes with three water molecules, three for 1ClEtOH and two for 2ClEtOH respectively, all involve bonds to both chlorine and the OH group. In the case of 1_3W_C also a CH-O hydrogen bond is visible. The hydrogen bonding between Cl and water is longer in all the 3W complexes than in the 2W complexes for the 1ClEtOH and, consequently, the C-Cl bond distance increase is smaller. The same effect, although smaller, is noticed for 2ClEtOH.</p><p>Finally, the larger cases studied were the complexes of 1ClEtOH and 2ClEtOH with four water molecules. Three such complexes were found for the former species and only two for the latter. Only 1_4W_A and 2_4W_A are fully interconnected clusters, both 1_4W_B and 2_4W_B exhibit an unconnected water as the third solvent molecule attached to the hydroxyl group, while 1_4W_C shows it completely unconnected to the solute.</p><p>One remaining aspect to consider is the existence or not of hydrogen bonds with respect to chlorine. As we showed previously, that is not case in the isolated molecules but it is yet to be proved whether such a hydrogen bond exists in the complexes. To avoid irrelevant information, we will comment only on the 1_nW_A and 2_nW_A complexes, with n=1, 2, 3, 4, leaving aside the least stable B, C, and D complexes. The procedure followed was to identify the three atoms (Cl,H,O) supposedly involved in the hydrogen bond and to calculate the gradient of the electronic density in the plane defined by them. The presence or not of a hydrogen bond, as opposed to a simple electrostatic interaction, was assessed by the appearance or not of a critical point of the density between the Cl and H atoms [38]. The corresponding images are collected in Fig. 5 What the diagrams in Fig. 5 show is that chlorine is not always hydrogen bonded to the hydrogens in the water molecules. This can be seen, for instance in 2_3W_A and 2_4W_A where, even if an electrostatic attraction and some shared density are noticeable, chlorine is not bound to any of the two hydrogens, but somehow to a point in between. In other cases, like in 2_2W_A, there is a clear hydrogen bond to one of the hydrogen atoms, while there is no density shared with the second one. Finally, there are clear cases in which Cl is hydrogen bonded, like all the series 1_nW_A and 2_1W_A. Again one notices the preferential effect of the water solvent on 1ClEtOH with respect to 2ClEtOH.</p><!><p>The ∆ 𝑟 𝐻 0 0 , ∆ 𝑟 𝐻 298 0 and ∆ 𝑟 𝐺 298 0 for the complexes were calculated in two ways. On the one hand, the properties were computed with respect to the solute and n isolated water molecules, like in (1)</p><p>In this case, the properties were divided by n (the number of water molecules) to obtain an estimate of the average contribution per monomer. Additionally, the properties were calculated with respect to water multimers. These monomer, dimer, trimer and tetramer of water were optimized starting from the structure of the water chains in N_nW_A. No attempt was made to find the most stable isomer of these water clusters, but just the one more similar to the structure present in the solvated complex. The equation would look as in (2).</p><p>The formal n=0 structure has also been included, which obviously corresponds to a zero energy in both equations. The optimum structures and energetics of the water molecule and water multimers are given in the Supplementary Information.</p><p>The numerical data for the complexation energies are given in Table 2, while a comparative analysis is shown in Fig. 6 (only for ∆ 𝑟 𝐻 298 0</p><p>). An important observation in Table 2 is that for most of the complexes, the free energy of reaction is positive, implying that the clusters would not form spontaneously in gas phase at 298.15K. Only in the case of 1_3W_A most of the methods (except B3LYP) predict spontaneous formation. This is not extraordinary. Even in the case of the isolated water clusters themselves, Shields et al [34] showed positive values of ∆ 𝑟 𝐺 298 0 at 25 o C while negative values were observed at smaller temperatures. We show this also for our complexes in Table 3. At very low temperatures (i.e. 10K), the entropic effect is negligible and ∆ 𝑟 𝐺 10 0 is basically identical to the reaction enthalpy. The influence of the entropic factor increases with the temperature and turns positive slightly over 200K, an analogous behavior to that observed in ref.</p><p>[33] (see their Table 8).</p><!><p>The second important piece of information in Table 2 is the larger stabilization energy of the water complexes for 1ClEtOH than for 2ClEtOH when n=1, and 3, but the reversion of stability when n=4. This behavior can be rationalized in terms of local vs global solvation. While the 1ClEtOH species has a very hydrophobic end (with the -Cl and -OH substituents) and a hydrophobic end (the methyl group), 2ClEtOH has two relatively hydrophilic ends, one with the chlorine atom, the other with the hydroxyl group. Therefore, when there is a small number of molecules, local solvation of the more hydrophilic -CHClOH group favors a larger stability of the complexes. For a larger number of water molecules, the whole molecule becomes important and then the balance between hydrophilic and hydrophobic regions favors the solvation of 2ClEtOH over that of 1ClEtOH. One presumes then that the enthalpy of solvation of 2ClEtOH will be larger than the enthalpy of solvation of 2ClEtOH.</p><p>Looking at the graphs, it is noticeable the convergence of the complexation energy per water molecule to a value around 8 kcal/mol. Again, the larger facility for saturation of the solvation sites in the case of 1ClEtOH than 2ClEtOH is apparent. A glance at the curves of stabilization energies with respect to the water clusters reveals that 1ClEtOH starts to behave like fully solvated already with 2 water molecules, while 2ClEtOH needs 3 to exhibit the same effect. In both cases, the 4-water complex has already the first solvation shell complete, so that the complexation energy with respect to the 4-water cluster is comparable to the energy of complexation of the 1 water complex for 1ClEtOH and the 2-water complex for 2ClEtOH. A fourwater complex should then be enough to study meaningfully dehalogenation of these complexes in water solution.</p><!><p>Energies of complexation (in kcal/mol) for the different clusters with 1 to 4 water molecules. The upper two group of values correspond to the energies of complexation with respect to the water clusters for 1ClEtOH and 2ClEtOH respectively. The second two groups correspond to the complexation energies with respect to n isolated water molecules divided by the number of water molecules in each complex. For MP2, B3LYP, M06 and B97X-D methods, the values reported are those obtained with the aug-cc-pVTZ basis set The upper panel in Fig. 6, which represents the average energy per added water molecule both to 1ClEtOH and 2ClEtOH, shows an interesting difference between the species. While in 1ClEtOH the additional water contributes with approximately the same energy to the complex, this is not the case for 2ClEtOH, which shows a marked decrease toward stabilization with the additional molecules. This is another manifestation of the already mentioned effect of the more distributed hydrophilic and hydrophobic regions in 2ClEtOH with respect to 1ClEtOH. This effect is not so noticeable if one uses equation (2) where the energy difference is recorded against the n-water complex. In this case it is clear that the curves exhibit similar behavior, although the ones for 1ClEtOH are more acute and reach lower energies than those for 2ClEtOH. One can notice also that the B3LYP DFT method is the one giving the worse results (i.e. largest difference with respect to the more exact G2 results) of the three DFT methods, most probably because of the lack of dispersion energy, present in the other two and important for these loose hydrogen-bonded complexes.</p><!><p>The comparison of the C-Cl and O-H stretching in 1ClEtOH and 2ClEtOH is shown in Table 4. Solvatochromism is present in both bands as a hypsochromic effect (blue shift) in 1ClEtOH and 2ClEtOH as well. While the effect on the OH stretching is monotonical for both species, with a total shift of about 600 cm -1 , the C-Cl band has a different behavior. In 2ClEtOH the band suffers practically no shift, while in 1ClEtOH varies more slowly than the OH band, does not decrease monotonically and even splits in two bands in the case of the 3W and 4W complexes. This fact is shown in Table 4 by listing the pair of frequencies at which these bands appear, involving the C-Cl stretching and several OH stretching from the water molecules. Both for the 3-and 4-water complexes the lower wavelength band is less intense, about one third of the intensity of the band at the longer wavelength. The superimposed IR spectra of free and 4-water complexed 1ClEtOH is presented in Figure 7. The most important bands involving some movement of the solute have been marked in the figure. The hypsochromic effect on the C-Cl stretching band is significant and should be possible to identify it experimentally. Moreover, there is an important bathochromic effect on the torsion band of the COH group and a smaller one on the stretching band of the CO group. Taking into account both the hypsochromic and bathochromic effect, an interesting phenomenon results. While in the free 1ClEtOH the C-Cl band should be observed at longer wavelengths than the COH torsion, the opposite should be true in solution. All these spectral characteristics should be helpful to identify the presence of the 1ClEtOH isomer and distinguish it from the 2ClEtOH isomer.</p><!><p>In this work we report a computational study of the two isomers of chloroethanol, namely 1ClEtOH and 2ClEtOH, both free and complexed with up to four water molecules. Both the density function M06 method, with several basis sets, and composite methods CBS-QB3 and G4 have been employed for the calculation of structures, energetics and spectra.</p><p>The results indicate that there are important differences in the way that water affects both isomers. On the one side, the water complexes of 1ClEtOH are more stable than those of 2ClEtOH. About three molecules of water are enough to saturate the first solvation shell of both isomers, but the chlorine atom is more involved in the interactions with water in the case of 1ClEtOH than in the case of 2ClEtOH. An increase in the length of the C-Cl bond as well as a blue shift of the C-Cl stretching band supports the view that water solvation weakens the C-Cl bond, therefore favoring dehalogenation. These results agree with those of Philips et al [13] in their study of dehalogenation of chloromethanol. This is not the case in 2ClEtOH where water addition does not affect neither the CCl bond distance nor the stretching frequency. The OH stretching frequency, on the contrary, is affected similarly in both molecules. Although experimentally this frequency will be masked by those of the solvent water molecules, it is possible to observe in the theoretical spectra the hypsochromic effect associated to solvation.</p><p>The study of the IR spectra allows to identify separately the 1ClEtOH and 2ClEtOH isomers, both in gas phase and solution. The C-Cl stretching band shows a gap of about 40 cm -1 between both isomers. Therefore, the presence of two C-Cl stretching bands in gas phase would give away the presence of the second isomer. The relative intensities could provide an estimation of the relative concentrations. What is more, it was shown that in the presence of water the combination of a hypsochromic shift of the OH stretching band and a bathochromic shift of the COH torsion band reverses the order of these transitions with respect to gas phase, therefore providing clues about the presence or not of this isomer.</p><p>It is known that many reactions that occur in cold regions of the atmosphere and the interstellar medium (ISM) are produced by heterogeneous catalysis on dust particles and ice clusters. Binding of small molecules to the water in these clusters or, conversely, binding of small water clusters to those species, is an important feature to be understood for studies in the ISM. This paper sheds light on one such phenomenon, in the same spirit than other very recent publications [35]. The binding energies of the chloroethanols to the water clusters, as well as their variation with the number of water molecules contacted, give clues as to the extent in which a small ISM water particle can participate in the reaction of dehalogenation of these species.</p><p>Work is under way on the participation of water molecules in the dehalogenation reaction properly, which involves a hydrogen transfer from oxygen in the case of 1ClEtOH, while demanding a transfer from a hydrogen on carbon in 2ClEtOH. The determination of the relation between both isomers in the reaction path for the obtention of acetaldehyde will be published elsewhere.</p>

ChemRxiv

Imidazolate ionic liquids for high-capacity capture and reliable storage of iodine

Fast, efficient capture and safe storage of radioactive iodine is of great significance in nuclear energy utilization but still remains a challenge. Here we report imidazolate ionic liquids (Im-ILs) for rapid and efficient capture, and reliable storage of iodine. These Im-ILs can chemically capture iodine to form I-substituted imidazolate ILs with an iodide counterion and the newly formed ILs can absorb iodine to form polyiodide species and low-temperature eutectic salts. For example, choline imidazolate shows iodine capture capacities of 8.7 and 17.5 g of iodine per gram of IL at 30 and 100 o C, respectively, which are, to the best of our knowledge, higher than the values (0.5-4.3 g/g) reported to date. Importantly, iodine can be stably stored in the Im-IL absorbent systems even at 100 o C. The Im-ILs have potential for application in the capture and storage of radioactive iodine.

imidazolate_ionic_liquids_for_high-capacity_capture_and_reliable_storage_of_iodine

N<!>Results<!>Discussion<!>Methods

<p>uclear energy, a clean and low carbon-emitting power resource, is an efficient and reliable alternative source to fossil energy if the challenge of nuclear waste pollution can be tackled efficaciously 1,2 . Appropriate nuclear waste management has received considerable attention especially after the explosion of the Fukushima nuclear power plant in 2011. Volatile and radioactive iodine species (e.g., 129 I and 131 I) possess exceptional issues, mainly because 129 I has a very long half-life (~10 7 years) and 131 I has direct effects on human metabolic processes despite its short half-life (8.02 days). Therefore, efficient capture and reliable storage of radioactive I 2 is of great significance. To realize the effective I 2 capture, great efforts have been devoted to developing various physico-and chemo-absorbents/adsorbents. Metal-organic frameworks [3][4][5] , microporous polymer [6][7][8] , charged porous aromatic frameworks 9 , hydrogenbonded cross-linked organic framework 10 , nonporous pillar [6] arene crystal 11 , carbon-based materials 12,13 , deep eutectic solvents 14 , and ionic liquids (ILs) [15][16][17] , have been reported for the efficient capture of I 2 via physical interaction. However, these materials usually suffer from low I 2 uptake and unstable I 2 storage. Ag 2 O@NFC aerogels 18 , silver-containing mordenites 19 , alkene/alkyne perovskites 20 , alkali-TCNQ salts 21 , and functionalized Mg-Al layered double hydroxide 22 have been reported as chemical adsorbents, whose capture efficiency are considerably dependent on their activity to react with I 2 , generally with shortcomings such as slow reaction rate and low capture capacity. Consequently, although such progress has been made in I 2 capture, the absorbents/adsorbents capable of fast capturing and reliably storing I 2 is still highly desirable.</p><p>ILs, composed entirely of organic cations and inorganic/ organic anions, can be designed with specific and tunable chemical reactivity via careful selection of component ions, and have been widely applied in gas capture and chemical wastes management, showing promising potentials [23][24][25] . For example, 1butyl-3-methylimidazolium tetrafluoroborate/water system has been demonstrated to be a tunable media for sustainable waste management 26 . Heterocyclic anions-based ILs, including azolate and pyridinolate ILs, have been reported as effient sorbents for CO 2 , SO 2 , or NO due to the formation of complexes between the ILs and the gases [27][28][29][30][31][32] . Especially, the azolate and pyridinolate anions of the ILs have been found to be capable of chemically capturing CO 2 efficiently [27][28][29] . ILs also have been applied in I 2 capture and 1-butyl-3-methylimidazolium bromide was reported to show an I 2 capture capacity of 1.9 mol I 2 per molar IL 15 .</p><p>Here we report a series of imidazolate ILs (Im-ILs) for fast capture and reliable storage of I 2 . These Im-ILs are capable of chemically capturing I 2 rapidly via the reaction of the Im anions with I 2 , forming I-substituted imidazoles together with new ILs with [I]anion, and the in situ-generated ILs can absorb I 2 to form polyiodide species, resulting in very high I 2 uptake capacity. For example, choline imidazolate ([Ch][Im]) as the initial absorbent shows increased I 2 capture capacities with temperature in the range of 30-100 °C, giving the capacities of 8.7 and 17.5 g of I 2 per gram IL at 30 and 100 °C, respectively, which is an improvement over typical state-of-the-art capacities of 0.5-4.3 g/g. More importantly, the captured I 2 can be stably stored in the absorbent system even at 100 °C. It is suggested that the Im-ILs reported here are highly efficient not only for fast capture of I 2 with high capacity, but also for the safe storage of this volatile compound.</p><!><p>Synthesis and characterization of Im-ILs. The Im-ILs shown in Fig. 1a were synthesized via the neutralization of bases (e.g, choline) with weak proton donors including imidazole (Im), methyl-substituted Im, as detailed in the Methods. The 1 H and 13 C nuclear magnetic resonance (NMR) spectra confirmed the formation of these Im-ILs (Supplementary Fig. 1), and thermogravimetric analysis (TGA) showed that they are thermally stable from 131 °C to 223 °C (Supplementary Fig. 2). These Im-ILs are Extraction of I 2 from cyclohexane solution by Im-ILs. The resultant Im-ILs were first used to extract I 2 from cyclohexane solutions at 30 °C. It was found that all Im-ILs could rapidly and efficiently remove I 2 from the cyclohexane solutions. As illustrated in Fig. 2, 51 mg of [Ch][Im] (0.3 mmol) almost removed I 2 completely from the cyclohexane solution (0.01 M, 50 mL, 126.9 mg of I 2 ) within 15 min, reaching extraction equilibrium with an I 2 uptake of 3.1 mol I 2 per molar IL corresponding to an I 2 uptake of 4.6 g I 2 per g IL (Fig. 2c). Moreover, it was found that [Ch][Im] could also rapidly and efficiently remove I 2 from other solvents, for instance, n-hexane and n-heptane, affording lower I 2 uptakes at extraction equilibrium of 4.2 and 2.7 g/g, respectively.</p><p>To reveal the interaction mechanism of extraction I 2 by [Ch] [Im], the absorbent solution obtained at the extraction equilibrium was examined by 1 H and 13 C NMR spectroscopy. Compared with those of [Ch][Im], the 1 H and 13 C NMR spectra of the absorbent solution changed significantly (Fig. 3a, b and Supplementary Fig. 3). In contrast, the signals at 141.67 and 124.22 ppm. assigned to the carbons in the [Im] − anion of [Ch] [Im] disappeared and two new peaks appeared at 134.21 and 119.18 ppm., which were attributed to the carbons in [ImH] + cation according to the 1 H and 13 C NMR spectra of [ImH][I] (Supplementary Fig. 4). As shown in Fig. 3b, a new signal appeared at 89.68 ppm, which was ascribed to the carbon of C-I bond from 2,4,5-triiodoimidazole based on the 13 C NMR spectrum of 2,4,5-triiodoimidazole (Supplementary Fig. 5). This is also confirmed by the heteronuclear singular quantum correlation (HSQC) spectrum (Supplementary Fig. 6). The heavy atom effect was responsible for the signal of I-connected carbons shifting upfield. Diffusion ordered spectroscopy (DOSY) analysis (Supplementary Fig. 7) was performed to further confirm the assignment of the carbon peaks. The diffusion sequences of the signals at 134.21 and 119.18 ppm. were similar, whereas that at 89.68 ppm. was significantly different (Supplementary Fig. 7). This indicates that the signals at 134.21 and 119.18 ppm. belonged to the same compound and the signal at 89.68 ppm. was attributed to another one, which agreed well with the above analysis. Notably, it is accepted that the carbon chemical shifts of C2 and C4/5 in 2,4,5-triiodoimidazole should be different. This is true if the NH proton is not changing position between two N atoms in 2,4,5-triiodoimidazole, whereas an exchange usually occurs in solution and on the NMR time scale, thus resulting in one peak. In addition, the signals of the carbons in [Ch] + cation only shifted slightly. The above NMR results indicate that the [Im] − anion was so active that it chemically captured I 2 to form new species including [ImH] + and 2,4,5-triiodoimidazole, whereas the [Ch] + cation kept unchanged. From the 1 H NMR spectra of [Ch][Im] before and after capturing I 2 to reach equilibrium (Supplementary Fig. 3), it was found that the amount of H atoms in the Im ring of the absorbent solution declined to half value compared to that of To further understand the chemical reactions in the I 2 extraction process, the absorbent solutions obtained at different extracting time were examined by one-and two-dimensional NMR analysis. In the spectrum of the absorbent solution obtained at 1 min, the signals at 141.67 and 124.22 ppm. assigned to C2 and C4@C5 in the [Im] − anion disappeared, accompanied with the appearance of some new signals, whereas the signals assigned to the carbons 6, 7, and 8 in [Ch] + cation shifted slightly. These results confirm that the [Im] − anion could rapidly and efficiently Encouraged by the above results, other Im-ILs shown in Fig. 1a were also used to extract I 2 from cyclohexane solutions at 30 °C and the results are included in Table 1. Obviously, these Im-ILs could efficiently extract I 2 with high efficiency and the I 2 uptake of each Im-IL was almost the same. These results imply that the anions of these Im-ILs might behave similarly in extracting I 2 . NMR analyses (Supplementary Fig. 14 DFT calculation. To understand the reactivity of the anions of these Im-ILs, we calculated the Hirshfeld charges of carbons in the Im ring of imidazole, [Im] − and [ImH] + , and the results are shown in Supplementary Fig. 16. In general, the carbon site possessing more negative Hirshfeld charge has stronger ability to attract electrophiles and is thus more possible to be the reactive site 33,34 . The calculation results indicate that the Hirshfeld charges of the carbon atoms in [Im] − become more negative compared with those in imidazole, whereas those in [ImH] + are more positive. Similarly, the Hirshfeld charges of the carbons, especially the methyl carbons, in the [methyl-substituted-Im] − anion are also negative (Supplementary Fig. 16). These results imply that both Csp 2 -H and Csp 3 -H in the anions have the capability to react with I 2 . Therefore, it can be deduced that Im anion-directed electron redistribution is favorable to promoting C-H bond activation, thus providing multiple-sites to capture I 2 chemically with different reactivity.</p><p>I 2 vapor capture by [Im]-ILs. As I 2 capture is generally performed in gas atmosphere, the resultant Im-ILs were applied in capturing I 2 vapor and the results are shown in Table 1.</p><p>Obviously, each Im-IL could capture I 2 with a very high capacity, much higher than those of various absorbents or adsorbents reported in literature (Supplementary Table 2) 6,9,14,15 . For example, using [Ch][Im] as the absorbent, its I 2 capture capacity increased with temperature in the range of 30-100 °C, reaching the highest value of 11.8 mol/mol (i.e., 17.5 g/g) at 100 °C. Further increasing temperature caused decline in I 2 capture capacity, decreasing to 9.4 mol/mol at 110 °C (Fig. 5a).</p><p>As described above, the I 2 capture capacity of [Ch][Im] includes two parts: the chemical capture capacity of 1.5 mol I 2 per molar IL and the I 2 uptake by the mixture of [Ch][I] + [ImH][I] + 2,4,5-triiodoimidazole. The latter is actually the I 2 solubility in the mixture. In general, the solubility of a solute in a solvent is related to its vapor pressure. Calculating the vapor pressures of I 2 at different temperatures in the range of 30-100 °C (Supplementary Table 3 and Supplementary Fig. 17), the dependence of the I 2 uptakes of [Ch][Im] on its vapor pressures is plotted in Fig. 5b. Obviously, the I 2 uptakes of [Ch][Im] increased with the I 2 vapor pressures, suggesting that higher vapor pressure, i.e., higher temperature, is favorable to the I 2 absorption by the mixture of [Ch][I], [ImH][I], and 2,4,5-triiodoimidazole. However, further increasing temperature to 110 °C, the I 2 uptake of [Ch][Im] reduced, which may be ascribed to the weaker interactions among I 2 and the absorbent molecules at higher temperature. I 2 storage reliability. Besides rapid and efficient I 2 absorption, it is also very important for volatile I 2 to be stably stored in the absorbent system for a long time. To examine the I 2 storage reliability in the [Ch][Im]-based absorbent system, TGA analysis with N 2 sweeping was performed on sample A and sample B as shown in Fig. 6. Clearly, the mass losses of these samples were hardly observed at 30 °C under a N 2 sweeping flow rate of 40 mL/ min for 10 h, whereas pure I 2 powder showed ~30 wt% mass loss under the same condition. At 100 °C, only < 5% of mass losses for sample A and sample B were observed after N 2 sweeping with a flow rate of 5 ml/min for 10 h, whereas the same amount of powdered I 2 almost completely evaporated within 50 min. These results indicate that even at high temperature (e.g., 100 °C), the absorbent system still has relatively reliable capability to store I 2 . Compared with the reported I 2 absorbents in literature 14,15 , this IL absorbent system exhibits much better performance for I 2 storage. ) combined with IL cations were present in the IL solution absorbing I 2 vapor, identical to that reported in literature 35 . The ESI-MS results enclose why the absorption capacity of the IL capturing I 2 vapor was much higher than that extracting I 2 from cyclohexane solution.</p><!><p>A series of Im-ILs were designed, which were found to be capable of rapidly and efficiently capturing I 2 via the reactions of the Im anions with I 2 and the formation of polyiodide species (Fig. 1b), showing high I 2 capture capacity. Moreover, the Im-ILs systems were safe materials for the storage of I 2 , and they could be tolerant to high temperature (e.g., 100 °C). This work presents green absorbent systems to capture I 2 rapidly and efficiently, which have promising potential applications in capturing radioactive I 2 from the nuclear waste with stable storage.</p><!><p>Materials. Iodine, choline chloride, imidazole, 2-methylimidazole, 4-methylimidazole, 2,4-dimethylimidazole, tetrabutylammonium bromide, and ion exchange resin (Ambersep|r 900(OH)) were obtained from Beijing Innochem Science & Technology Co., Ltd. and J&K Scientific Ltd., respectively. All chemicals were used without further purification.</p><p>General procedure for the synthesis of Im-ILs. The ILs as depicted in Fig. 1a The calculation of the vapor pressures of I 2 at different temperatures. The temperature for the indicated pressure of I 2 solid (Supplementary Table 3) was obtained from the CRC handbook of chemistry and physics (90th edition). The temperature-vapor pressures (Supplementary Fig. 17) curve fits with the Clausius-Clapeyron equation (ln P ¼ ÀΔH RT þ C), which was then used for calculating the vapor pressures of I 2 at different temperatures in the range of 30-100 °C</p>

Nature Communications Chemistry

Tuning supramolecular G-quadruplexes with mono- and divalent cations

Supramolecular G-quadruplexes (SGQs) are formed via the cation promoted self-assembly of guanine derivatives into stacks of planar hydrogen-bonded tetramers. Here, we present results on the formation of SGQs made from the 8-(m-acetylphenyl)-2\xe2\x80\xb2-deoxyguanosine (mAGi) derivative in the presence of various mono- and divalent cations. NMR and HR ESI-MS data indicate that varying the cation can efficiently tune the molecularity, the fidelity and stability (thermal and kinetic) of the resulting SGQs. The results show that, parallel to the previously reported potassium-templated hexadecamer (mAGi16\xc2\xb73K+), Na+, Rb+ and NH4+ also promote the formation of similar supramolecules with high fidelity and molecularity. In contrast, the divalent cations Pb2+, Sr2+ and Ba2+ template the formation of octamers (mAGi8), with the latter two inducing higher thermal stabilities. Molecular dynamics simulations for the hexadecamers containing monovalent cations enabled critical insights that help explain the experimental observations.

tuning_supramolecular_g-quadruplexes_with_mono-_and_divalent_cations

Introduction<!>Monovalent cations<!>Divalent metal cations<!>Thermal stability studies<!>MDS studies<!>Conclusion

<p>Guanosine is a privileged recognition motif for the development of discrete supramolecular self-assembled nanostructures (1). It is known that guanosine, deoxyguanosine and related derivatives can form hydrogen-bonded tetrameric structures (a.k.a. G-tetrads and G-quartets) that coaxially stack upon the addition of a wide variety of cations, to form supramolecular G-quadruplexes (SGQs) (Figure 1(a)) (2). G-rich oligonucleotides (e.g. DNA and RNA) can also self-assemble and/or fold into G-quadruplexes, which we term here oligo-GQs or OGQs. Both the structure and stability of SGQs can be modulated by either covalent modifications of the assembling subunits (intrinsic parameters) (3) or by changes of extrinsic parameters such as the cation/anion (4) and the solvent (5). For example, we have previously reported that modifications at the C8 of the guanine moiety with aryl or hetero aryl groups (8ArG) enable the reliable formation of precise supramolecules (i.e. discrete and of well-defined size and composition) (3). The 8-(m-acetylphenyl)-2′-deoxyguanosine moiety has been of special interest because its reliable potassium cation promoted self-assembly into hexadecameric supramolecules; this happens in both organic (3a) and aqueous media (6), and even after covalently attaching bulky groups such as dendrons (3). We have also shown that the hexadecamer formed by the lipophilic derivative 8-(m-acetylphenyl)-2′-deoxyguanosine (mAGi) (Figure 1(a)) in acetonitrile, replacing the cation from potassium to strontium, enabled the reversible high fidelity switching to the corresponding octamer (7). Such results prompted us to perform a systematic study to determine the effect of other mono- and divalent cations on the self-assembly of the same 8ArG derivative.</p><p>The wide interest in elucidating the role of the cation (8) in the structure and dynamics of SGQs is evident by the multiple studies addressing, for example, the effect on molecularity (amount of subunits), fidelity (9) (per cent of desired structures) and stability (thermodynamic and kinetic) of such structures, when in the presence of monovalent (10), divalent (11) and even trivalent (12) metal cations (2). This interest has also been fuelled by the potential use of SGQs as self-assembled ionophores due to their varied cation affinities and selectivities (13).</p><p>Here, we report the results of NMR, ESI-MS and molecular dynamics simulations (MDS) on the self-assembly of mAGi in the presence of monovalent cations such as Li+, Na+, K+, Rb+, Cs+, Tl+ and NH4+ and of divalent cations such as Sr2+, Ba2+ and Pb2+ (Figure 2(c)). Our results indicate that cation properties such as charge, size and electron density dictate the molecularity, fidelity and stability (thermal and kinetic) of the resulting SGQs. We expect this information to be useful in the further development of SGQs for the construction of functional supramolecular systems.</p><!><p>The 1H NMR spectra of mAGi (30 mM in CD3CN) with different iodide salts were used to evaluate the supramolecular assemblies promoted by each of the aforementioned cations. The resulting spectra, after addition of NaI, RbI or NH4I (0.5 equiv.), showed very similar signals to the previously reported potassium-promoted hexadecamer mAGi16 (Figure 2(c)–(f)) (3a), including a series of characteristic cross-peak patterns in the 2D NOESY spectra (Figure 3). HR-ESI MS experiments show clean spectra with prominent peaks matching the masses for [mAGi16·3Na]3+ (m/z 2825.8347), [mAGi16·3NH4]3+ (m/z 2820.7881) and [mAGi16·3Rb]3+ (m/z 2888.3225), although the latter with a lower relative abundance (Figure 3). As stated earlier, the NMR spectra show that Na+, Rb+ and NH4+ promote the formation of a very similar hexadecamer to the one promoted by K+, nevertheless, there are some subtle differences (6). Specifically, the chemical shifts for the N1H signals for to the outer (red) and inner (navy blue) tetrads of mAGi16 are affected to a greater extent. While the signal corresponding to the inner tetrad of mAGi16 remains reasonably constant at around 11.2 ppm (Figure 2 (c)–(f), navy blue signal), the signal corresponding to the outer tetrad (Figure 2(c)–(f), red signal) shows a downfield shift as the ionic radius of the cations increases. Smaller ions like Na+ strongly polarise the carbonyls of the outer tetrads, which in turn induces a deshielding of the N1H protons when compared with bigger cations like NH4+ (Figure 2(f)). In contrast, the region corresponding to the H1′/3′ reveals fairly constant chemical shifts due to their relatively further distance from the cation coordination site. Although Rb+ (1.61 Å) is not significantly larger than K+, and similar in size to NH4+, we hypothesise that the lower fidelity (e.g. small amounts of unidentified assemblies in the NMR) and stability of mAGi16·3Rb+ result from inter-cationic repulsion due to its higher relative electron density.</p><p>Moving from K+ to smaller (e.g. Li+; 0.92 Å) (14) or larger cations (e.g. Cs+; 1.74 Å) (15) is detrimental to the formation of SGQs (Figure 2). Addition of lithium iodide (0.5 equiv.) to a solution of mAGi leads to a 1H NMR spectrum with broader peaks (Figure 2(b)) relative to the one recorded prior to adding the salt (Figure 1(a)). The peaks corresponding to H1′ and H3′ are shifted upfield, while the N1H peak (~ 12 ppm) is broadened to the point of being almost undetectable, suggesting a more dynamic system (i.e. faster exchange of the mAGi subunits). HR-ESI-MS experiments reveal a prominent base peak for [mAGi8·2Li]2+ (m/z 2109.1572), which support the formation of a dynamic octamer (mAGi8) in solution. This behaviour is presumably a consequence of the greater desolvation energy of Li+ ions combined with its preferential binding coplanar to the tetrad. The latter precludes the simultaneous coordination of the O6 of two consecutive tetrads making their coaxial stacking more difficult. At the other end of the size range, experiments with Cs+ reveal it to also be unsuitable at promoting the formation of SGQs (Figure 2(g)). This phenomenon, however, might be limited to SGQs formed by mAGi and other 8ArG derivatives since Davis et al. (16) have reported the formation of SGQs promoted by Cs+ ions.</p><p>Based on its size (1.59 Å) (14), and the fact that it is monovalent, Tl+ should promote the formation of a hexadecamer. Thallium has proven useful in structural studies of OGQs by X-ray crystallography due to its high electron density (17). Our results show that, while Tl+ does in fact promote the assembly of mAGi (Figure 2(h)) with modest fidelity (58%), instead of the anticipated hexadecamer, the resulting SGQ is an octamer (mAGi8-Tl+) of low relative stability (Tm = 304 K; Figure 5). We hypothesise that this results from the relatively low solubility of TlI in acetonitrile (as indicated by the fact that some of the salt precipitated during sample preparation). This parallels our previous report on the shift in the equilibrium from a hexadecamer to an octamer by a derivative closely related to mAGi, in which the availability (i.e. activity) of the promoting K+ ions was modulated by the polarity of the solvent used (acetonitrile vs chloroform) (5a).</p><!><p>We have previously reported a metallo-responsive SGQ that switched, respectively, between hexadecameric and octameric states when the metal cation was changed from potassium to strontium. Intrigued by these results, we tested the effects of other divalent cations, specifically Ba2+ and Pb2+. NMR studies reveal that, similar to strontium, such cations promote the preferential formation of octamers, although with slightly lower fidelities (Figure 2(i)–(k)) and thermal stabilities (Tm = Sr2+ ~ Ba2+ > Pb2+). The corresponding HR-ESI MS spectra (Figure 4) show good correlation with the NMR experiments, similar to the corresponding experiments with monovalent cations.</p><p>Parallel to Tl+, the divalent metal cations Sr2+, Ba2+ and Pb2+ promote the formation of the octamer mAGi8, although with higher fidelity and thermal stability (Figure 4). Formation of a hexadecamer would require three consecutive divalent cations between the four tetrads leading to enhanced charge–charge repulsion that would shift the equilibrium towards the corresponding octamer. This behaviour is consistent to the one reported by the Wu group (11b) in which the lipophilic 2′,3′,5′-O-triacetyl-guanosine derivative formed octamers in the presence of Ca2+, Sr2+ and Ba2+. In contrast, Davis reported the formation of hexadecameric SGQs by lipophilic guanosine derivatives induced by strontium, barium (4b, 4c) and lead (11a) cations. These hexadecamers, however, were in essence dimers of octamers in which a belt of four picrate anions interacted strongly (via attractive charge–charge interactions and the formation of H-bonds) with the guanosine subunits of the two inner G-tetrads.</p><!><p>Variable temperature 1H NMR experiments (Figures 5, S20–S32) reveal that the hexadecamer templated by K+ (1.51 Å) (14) (mAGi16·3K+) is about 20 K more thermally stable than those templated by Na+ (1.18 Å) and NH4+ (1.61 Å) (15). The resulting trend in thermal stability (K+ > NH4+ ~ Na+ >> Rb+) is consistent to the one reported with other SGQs and OGQs (18). Divalent cations reveal mixed results, with the octamers (mAGi8) templated by Sr2+ and Ba2+ showing enhanced stability relative to mAGi16·3K+ (ΔTm > +5 K) while Pb2+ resulted in an octamer with lower stability (ΔTm = –216 K) closer to those of the hexadecamers templated by Na+ and NH4+. The enhanced stability induced by Sr2+ and Ba2+ is consistent with reports with both OGQs and SGQs where the enhancement has been attributed to a stronger coordination between O6 of the guanine moieties (18b).</p><p>Lead(II) ions have also been reported to be effective promoters for the formation of OGQs (19) and SGQs (11a, 19d), and this characteristic has been exploited in the development of OGQ-based sensors (20). The 8-aryl-substituted mAGi derivative, however, forms a hexadecamer of lower stability in contrast to the report by Davis (11a) indicating the formation of a lead(II)-templated hexadecamer of enhanced kinetic stability. We hypothesise this phenomenon results from a decreased activity of lead cations from PbI2 in acetonitrile (relative to the lead picrate used by Davis), which is parallel to the aforementioned results with TlI.</p><!><p>Atomistic MDS studies provided additional information that complement the NMR and MS experiments (15). MDS studies have provided critical insights into the effects of changes in the sequence, cations and binding of small molecules on the structure and dynamics of OGQs (15, 21). Specifically, detailed assessments of the dynamics of cations moving in and out of the central channel of OGQs were recently reported by Reshetnikov et al. (22) and Akhshi et al. (23) using, respectively, the thrombin-binding aptamer and a tetramolecular OGQ.</p><p>We carried out MDS with mAGi16·3X+ and mAGi8·X+ where X+ represents the aforementioned monovalent cations (with the exception of Tl+). The starting structures were constructed by replacing the potassium cations in mAGi16·3K+ and mAGi8·K+ with the corresponding cation of interest.1 The stability of the SGQs was assessed taking into consideration a combination of the following criteria: (1) changes in the RMSD values as a function of time (Figure 6(a)); (2) the probability density as a function of RMSD (Figure 6(b)) and (3) a visual inspection of structures at the end of the simulation (Figure 7). For the first criterion, large deviations in RMSD values relative to an idealised mAGi16·3K+ or mAGi8·K+ structure (based on 1D/2D NMR studies) (6) were taken as an indication of relative instability, as were systems with a positive slope in the trajectory. For the second criterion, sharper (i.e. narrower and higher) peaks reflect a lower spread of the distributions, and SGQs that have reached a relatively stable configuration. The third criterion is especially important in cases in which, despite the first two criteria suggesting a putative stable structure (e.g. flat RMSD trajectory and a narrow RMSD distribution), but where the supramolecular arrangement of subunits deviate significantly from the idealised features of SGQs (e.g. planar co-axial tetrads, presence of internal cations and appropriate syn/anti conformation of the G-subunits).</p><p>In general, and in all the simulated hexadecamers (mAGi16·3X+), the individual constituent subunits showed negligible changes in the sugar pucker (of the 2′-deoxyguanosine moieties) and relatively small changes in the glycosidic angles (Figure 7). Throughout the entire simulation, the central channel maintains an average diameter of 2.5 Å, but fluctuations of up to 5.2 Å (in the outer tetrads) enable the free movement of cations. The itra-tetrad hydrogen-bond patterns are largely conserved despite deviations from planarity, with the largest deviations found for the Rb+- and Cs+-containing hexadecamers. For all the systems, the outer tetrads (highlighted in red in Figure 7) showed the greatest deviations from planarity as expected due to the greater freedom for the constituent subunits.</p><p>Analysis of the RMSD trajectory for the average structures of mAGi16 with the smaller cations (Li+, Na+ and K+) reveals a low average RMSD value with a narrow structural distribution peak at around 2.4 Å (Figure 6). Although mAGi16·3Li+ is not observed experimentally, we can deduce from these studies that a putative hexadecamer containing Li+ would be akin to mAGi16·3Na+ due to their similar RMSD values and probability densities (Figure 6 (b)). Simulations of mAGi8·Li+ are consistent with the experimental observation of an unstable octamer with an average RMSD of 5 Å and a broader structural distribution than the one observed for mAGi8·K+ (Figures S35–S37). Bowman et al. (24) have used the terminologies condensed, glassy and coordinated to label the level of affinity (lowest to highest, respectively) of different cations towards specific nucleic acid structures (including OGQs). These classifications take into consideration multiple parameters such as their contributions to the kinetic and thermodynamic stability of the specific structure, diffusion characteristics and the extent of ion coordination. While the alkali cations Na+, K+ and Rb+ should be considered as the coordinated type, Li+ can be classified as a glassy cation, because the evidence is consistent with its interaction with the SGQs, but its small size preclude a stronger coordination of the eight oxygen atoms at the interface between two tetrads.</p><p>The results with Li+, Na+ and K+ contrast those obtained with Rb+, NH4+ and Cs+, which show RMSD values about three times larger than the former with different varying degrees of structural distributions. As shown earlier, the hexadecamer mAGi16 promoted by NH4+ shows slightly higher thermal stability than the one promoted by Na+, which correlates with reports from the literature in which ammonium cations have proven useful in studies of cation dynamics in OGQs and SGQs by NMR (25) and MS (26). Moreover, while mAGi16·3NH4+ gives a narrow structural distribution, the resulting RMSD is three times higher (~ 12 Å) than that of Na+. This seemingly conflicting result is likely a consequence of the specific coordination of the ammonium cations in which the resulting H-bonds impose relative orientations between the G-tetrads that deviate from those imposed by the spherical alkali cations.</p><p>As mentioned earlier, the two largest alkali cations evaluated (Rb+ and Cs+) induced significant structural distortions that are evident in the molecular models (Figure 7) and as reflected by the large RMSD values. The hexadecamer mAGi16·3Rb+ showed a broader structural distribution than mAGi16·3Cs+, while the latter resulted in the highest average RMSD value (Figure 6). A visual inspection of both hexadecamers (Figure 7) reveals significant structural distortions such as an offset of the tetrads and a displacement of the subunits relative to the initial central axis (Figure 7). These structural distortions result in diminished non-covalent interactions (e.g. π–π), and the loss of a cation from the inner/outer tetrad interface, all of which is consistent with the low fidelity and stability evident from the NMR studies (Figure 2). A closer look at the models show that, in the case of structures (mAGi16) containing Li+, NH4+, Rb+ and Cs+, one cation from the outer tetrad is expelled from the channel.</p><!><p>This work illustrates how the cation size, charge and electron density provide a suitable strategy to modulate the self-assembly of mAGi into SGQs of different molecularities and stabilities. Monovalent cations with a size in the range of 1.18–1.66 Å (K+, Na+, Rb+ and NH4+) promote the formation of the corresponding hexadecamer while, for smaller (Li+) or larger (Cs+) cations, the equilibrium is shifted towards an octamer or unidentified aggregates, respectively. Divalent cations within that size range also promote the formation of the corresponding octamers that are thermally more stable (at least for Sr2+ and Ba2+) than even the K+-promoted hexadecamer. We expect these findings to lead to the design and construction of multifunctional supramolecular systems with tuneable structure and dynamics for a wide variety of applications.</p>

PubMed Author Manuscript

Development and Validation of Discriminative Dissolution Method for Metformin Immediate-Release Film-Coated Tablets

The purpose of this study was to develop and validate a discriminative dissolution method for the metformin film-coated tablet with immediate release of the active substance that belongs to class III of the Biopharmaceutical Classification System (BCS). Different conditions such as type of dissolution medium, volume of dissolution medium, rotation speed, apparatus, and filter suitability were evaluated. The most discriminative release profile for the metformin film-coated tablet was accomplished by using Apparatus II (paddle) and 1000 mL of phosphate buffer pH 6.8 as the dissolution medium and maintained on 37 ± 0.5°C with a rotation speed of 75 rpm. The quantification of the released active substance was performed by UV/Vis spectrophotometry, at 232 nm. Acceptance criteria for not less than 75% (Q) of the labeled content for 45 minutes were set. The dissolution method was validated according to the current international guidelines using the following parameters: specificity, accuracy, precision, linearity, robustness, and stability of the solutions, found to be meeting the predetermined acceptance criteria. A developed dissolution method has discriminatory power to reflect the characteristics of the medicinal product and is able to distinguish any changes related to quantitative formulation and can be also applied for routine batch testing.

development_and_validation_of_discriminative_dissolution_method_for_metformin_immediate-release_film

1. Introduction<!>2.1. Material and Equipment<!>2.2. Dissolution Method<!>2.3. Preparation of Standard Stock Solutions<!>2.4. Determination of Metformin Hydrochloride in the Dissolution Samples<!>2.5. Filter Suitability Evaluation<!>2.6. Validation of the Dissolution Method<!>3. Results and Discussion<!>3.1. Selection of Dissolution Medium<!>3.2. Selection of Medium Volume, Dissolution Apparatus, and Rotation Speed<!><!>3.2. Selection of Medium Volume, Dissolution Apparatus, and Rotation Speed<!>3.3. Filter Suitability Evaluation<!>3.4. Validation of the Dissolution Method<!>4. Conclusions<!>Data Availability<!>Conflicts of Interest<!>

<p>Dissolution tests can be used to guide the development of new formulations and to assist in proper formulation selection (selection of excipients), to assess the characteristics of the active substance (AS), and to evaluate the batch-to-batch quality and stability of the medicinal product helping in the establishment of shelf life [1– 3]. It is also commonly used as a prediction of the in vivo performance of a medicinal product to provide a basis for achieving in vitro/in vivo correlation and to minimize the need for bioequivalence studies (BE). The dissolution procedure has several distinct components. These components include a dissolution medium, an apparatus, the study design (including acceptance criteria), and the method for quantification of the released AS [4].</p><p>Fundamentally, the dissolution method should be discriminatory, and it should allow evaluating the performance of the medicinal product, particularly in monitoring AS or critical formulation parameters [5]. Metformin hydrochloride (1,1-dimethylbiguanide hydrochloride) is used in the treatment of type 2 diabetes mellitus. The active substance is highly hydrophilic and is classified as class III according to the Biopharmaceutical Classification System (BCS) with high solubility and low permeability. The active substance is ionized at physiological pH (pKa value is 12.4). After oral administration, metformin HCL is absorbed by the gastrointestinal mucosa. The main site of rapid release absorption is the small intestine with negligible absorption in the stomach and colon [6]. However, to develop a dissolution method, the characteristics of the AS and its behaviour in the selected test media should be taken into consideration. Moreover, the dissolution conditions must follow the sink conditions and the quantitation method should be specific, accurate, precise, linear, and robust [7–9]. Several dissolution methods for determination of the released percentage of the AS from the medicinal product metformin immediate-release film-coated tablets are introduced in the British (BP) and American Pharmacopoeia (USP). In the BP monograph, the recommended dissolution method for metformin immediate-release film-coated tablets uses 900 mL of phosphate buffer pH 6.8 at 37°C and baskets with 100 rpm with quantitation by spectrophotometry at 232 nm. Furthermore, the USP monograph presents three procedures for determining the percentage of released active substance depending of the dosage strengths [10, 11]. Usually, either the USP or BP methods may be suitable for generic immediate-release (IR) products, but during development, we must optimize the selected method with respect to our proposed formulation.</p><p>The objective of the present study is to develop a discriminating dissolution method for metformin film-coated tablets to support development of the medicinal product and quality control efforts for all dosage strengths. The initial part of the study was focused on the selection of suitable dissolution conditions including volume of the medium using different dissolution media, type of apparatus, rotation speed, and suitability of the filter type. Additionally, validation was performed to ensure that the developed discriminatory method accomplishes its intended purpose.</p><!><p>Metformin film-coated tablets 1000 mg, 850 mg, and 500 mg, metformin hydrochloride working standard (WS), potassium dihydrogen phosphate, and sodium hydroxide were with analytical grade. Dissolution media: pH 1.2 (HCl and NaCl), acetate buffer pH 4.5, and phosphate buffer pH 6.8 were prepared according to the directions in the European Pharmacopeia (EP) monograph.</p><p>The following instruments were used: six-station dissolution apparatus (Varian-Vankel 7025 Model: 115/230) in accordance with the USP general methods, pH meter (Mettler Toledo), hotplate stirrer (IKA C-MAG HS7), analytical balance (Sartorius CPA 225D-OCE), and UV/Visible Spectrophotometer (Varian-Model: Cary 50/60) using 10 mm quartz cells).</p><!><p>The dissolution tests on metformin film-coated tablets were performed using Apparatus I and II at 37 ± 0.5°, with a rotation speed of either 50 rpm or 75 rpm for the paddle and 100 rpm for the basket. Different dissolution media (pH 1.2 HCl and NaCl, acetate buffer pH 4.5, and phosphate buffer pH 6.8) with either 900 mL or 1000 mL were tested. Sampling aliquots of 10 mL were withdrawn at 5, 10, 15, 30, and 45 min and replaced with an equal volume of the fresh medium maintained at the same temperature. After the end of each test time, the samples aliquots were filtered through 0.45 μm membrane filter (regenerated cellulose, RC), diluted with respective dissolution medium and then analyzed by the UV-Vis spectrophotometric method.</p><!><p>The stock solution was prepared by dissolving 55.55 mg of metformin hydrochloride WS in a 100.0 mL volumetric flask with the medium phosphate buffer pH 6.8. 1.0 mL of this solution was diluted to 100.0 mL with medium to obtain a concentration of 0.005 mg/mL. The solution was filtered through 0.45 μm RC membrane filter.</p><!><p>The percentage of the released active substance from the medicinal product was determined by the UV-Vis spectrophotometric method. The UV-Vis spectra of the metformin HCL solution revealed one absorption maxima at 232 nm.</p><!><p>Suitability of the filter type (0.45 µm RC membrane filter) was determined by comparison of the absorbance obtained from six individually filtered aliquots of the standard solution of metformin hydrochloride WS (concentration of 0.005 mg/mL) with the absorbance of unfiltered standard solution.</p><!><p>The proposed dissolution method was validated for specificity, accuracy, precision, linearity, robustness, and stability of sample solutions according to the current ICH and FDA guidelines [12, 13]. The specificity of the dissolution medium was tested by examining the peak interference from the dissolution medium and placebo in comparison with AS by an aliquot of the dissolution medium without AS (diluent), dissolution medium in which the placebo was added (placebo), dissolution medium containing AS at working concentration (standard solution), and dissolution medium with the metformin film-coated tablet (sample solution).</p><p>Accuracy of the method was tested by adding known amounts of metformin HCL WS to the placebo. Three concentrations (80%, 100%, and 120) of the theoretical working concentration were spiked, and the measurements were done in triplicate. The average recovery percentage was calculated.</p><p>The precision of the method was determined by testing the repeatability in the same day and intermediate precision (same analyst, using the same instrument) on a different day. The repeatability was tested by six replicate measurements of the absorbance of the standard solution in a concentration of 0.005 mg/mL and evaluated based on relative standard deviation (% RSD) of the results. Intermediate precision was determined with six sample solutions prepared individually using a single batch of film-coated tablets as per the test method by the same analyst, using the same instrument on a different day. Furthermore, comparison of the results variability was performed by statistical F-test.</p><p>Linearity was tested using six concentrations (in triplicate) of the standard solution within a concentration range from 0.00125 to 0.0075 mg/mL (25%, 50%, 75%, 100%, 125%, and 150% of the expected working concentration) and was evaluated by the linearity plot and the correlation coefficient (r2).</p><p>The robustness of the method was evaluated by variation of the three parameters: pH of the buffer (±0.2 units), rotation speed (±5 rpm), and temperature of the dissolution medium (±0.5°C). For each variation, the content of dissolved metformin HCl was calculated. The robustness of the method was examined by statistical F-test.</p><p>Stability of metformin HCl was evaluated using the standard solution (0.005 mg/mL) over the 48 hours test period. Sample solutions were prepared in the same dissolution media and at the same conditions as for the dissolution test. The drug concentrations observed in samples at 0, 5, 24, and 48 h were compared.</p><!><p>Development and validation of the discriminatory dissolution procedure was achieved following the current compendial standards in accordance with the FDA and ICH guidelines. It is worth noting that regulatory requirements at all levels of vertical follow-up (directives, regulations, procedures, recommendations, guidelines, and observations) are often strict and precise.</p><p>The dissolution procedure has several distinct components. These components include a dissolution medium, an apparatus, the study design (including acceptance criteria), and the mode of assay. All of these components must be properly chosen and developed to provide a method that is reproducible for within laboratory day-to-day operation and robust enough to enable transfer to another laboratory.</p><!><p>During the development of the dissolution procedure, one general goal is to have "sink" conditions. When "sink" conditions are present, it is more the likely that the dissolution results will reflect the properties of the dosage form. The sink conditions are defined as concentrations that yield a saturation solubility of the active substance at least three times the highest dose of the active substance dissolved in the volume of the medium used for dissolution [7]. Sink conditions are preferred because they are more likely to result in dissolution that reflects kinetics of the active release from the dosage form rather than from solubility limitations [9]. Media deaeration is usually required and can be accomplished by heating the medium or filtering the medium (more commonly) or placing it under vacuum for a short period of time. Bubbles can cause particles to cling to the apparatus and vessel walls. On the contrary, bubbles on the dosage units may increase buoyancy, leading to an increase in the dissolution rate, or may decrease the available surface area, leading to a decrease in the dissolution rate [14].</p><p>Determination of the maximum solubility or concentration of saturated metformin HCL solution in a different dissolution media (pH 1.2 HCl with NaCl, acetate buffer pH 4.5, and phosphate buffer pH 6.8) was performed using sink conditions at eight concentration levels in the range of 0.0006–0.0111 mg/mL. 250 mL from each medium were transferred into flasks and placed on a magnetic stirrer with temperature at 37 ± 0.5°C in a period of 45 minutes. The results from the solubility test of metformin HCL in different proposed dissolution media are summarized in Table 1.</p><p>The performed test for achieving the sink condition confirmed that the metformin HCL is highly soluble in all tested media with a solubility of more than 140 mg/mL. The maximum solubility of metformin hydrochloride was achieved in phosphate buffer pH 6.8. Solubility data were used as the basis for selecting a dissolution medium for further evaluation of the medicinal product.</p><p>According to the bioavailability data, after oral administration, metformin HCl is absorbed by the gastrointestinal mucosa. Taking into consideration that the main site of absorption is in the small intestine, with physiological pH value of 6.6–7.0, the phosphate buffer pH 6.8 was selected as the dissolution medium for further dissolution testing and at the same time аs a diluent where the standard stock solution would be dissolved.</p><!><p>The standard dissolution medium volumes accepted by the regulatory agencies are 500 mL, 900 mL, and 1000 mL [7]. The dissolution behaviour (variability and profile) of the dosage form itself is the best guide in choosing the volume. An effort to use one of the three standard volumes should be made to facilitate method transfer and reduce the likelihood of regulatory questions. For solid oral dosage forms, Apparatus I (basket) at 100 rpm or Apparatus II (paddle) at 50 or 75 rpm is recommended. The intent was to set a dissolution method performance that could yield data that are not highly variable and to avoid coning or mounding problems. After visual observation of the behaviour of the dosage form during the dissolution testing at 50 rpm, coning at the bottom of the Apparatus II (see Figure 1) has been noticed which leads to incomplete release of active substance and risk for obtaining variable results.</p><p>Occurrence of coning of the dosage form can be reduced by increasing the paddle speed thereby improving the results [14]. Therefore, the rotation speed of 75 rpm for the paddle was selected.</p><!><p>Case 1: Apparatus I (basket) at 100 rpm, 900 mL medium</p><p>Case 2: Apparatus I (basket) at 100 rpm, 1000 mL medium</p><p>Case 3: Apparatus II (paddle) at 75 rpm, 1000 mL medium</p><!><p>Dissolution was evaluated by measuring the amount dissolved over time and carried out on six (6) tablets. Samples of 10 mL were taken after 5, 10, 15, 20, 30, and 45 minutes, and the medium was replaced to maintain the same volume. The obtained results are presented in Table 2. The dissolution profile of metformin 1000 mg film-coated tablets is shown on Figure 2.</p><p>The obtained results show that the percentage of release of the active substance in Case 1 is lower compared to the results obtained in cases 2 and 3. Furthermore, the higher variability is evident in the dissolution profile in Case 2 compared to the dissolution profile in Case 3. Therefore, the Apparatus II (paddle), 75 rpm, and 1000 mL of phosphate buffer pH 6.8 were chosen as the conditions for the dissolution method.</p><p>The dissolution procedure requires an apparatus, a dissolution medium, and test conditions that together provide a method that is discriminating, yet sufficiently rugged and reproducible for day-to-day operation and should be able to transfer between laboratories. The ideal method will have enough power to pick up changes in critical attribute that may affect the release mechanism.</p><p>The discriminatory power of the proposed dissolution method was confirmed by comparing the dissolution profiles for the two different formulations of metformin 1000 mg film-coated tablets: original formulation and formulation with a deliberate change in the composition of excipients (around 33% excess of magnesium stearate) [15]. Results have shown that changes of the quantitative composition of magnesium stearate (lubricant) significantly affect the rate of in vitro dissolution by decreasing the percent of the released active substance (see Figure 3). This concept was used to establish the factor that has the most significant influence on the dissolution rate.</p><p>In order to confirm the applicability of the proposed dissolution method for all the strengths of metformin film-coated tablets (500 mg, 850 mg, and 1000 mg), the dissolution profiles in 1000 mL phosphate buffer at pH 6.8 with paddle rotation speed of 75 rpm were compared (see Figure 4).</p><p>The obtained profiles between the strengths with the developed dissolution method were considered as similar. Therefore, the experimental conditions described in Case 3 are proposed for the dissolution test in the formulation development of the medicinal product to determine the similarity of the dissolution profiles, as well as for selection of the biobatch that will be further used in the bioequivalent study. The quantitative determination of the released active substance from the medicinal product was performed using the UV/Vis spectrophotometric method at 232 nm. The acceptance criteria for not less than 75% (Q) of the labeled content for 45 minutes were set [15].</p><!><p>The choice of a suitable filter type is important and should be experimentally justified in the early development of the dissolution method [7, 8]. The suitability of the proposed filter (0.45 µm regenerated cellulose membrane filter) was confirmed by comparison of the recovery values of the concentration of metformin HCl between filtered and unfiltered standard solutions. The recovery values are within 99.23–100.15% (acceptable results are between 98.0% and 102.0%). The results indicate that the recovery is not affected when the standard solutions are filtered through the filter 0.45 µm regenerated cellulose membrane filter.</p><!><p>Method validation was performed on the highest strength of the medicinal product. The results are summarized in Table 3.</p><p>To evaluate the specificity of the dissolution procedure, it is necessary to demonstrate that the results are not affected by the placebo constituents in the medicinal product. A proper placebo should consist of everything in the formulation, except the AS. Comparison of the spectra (recorded between 200 nm and 400 nm) of the dissolution medium (diluent), placebo, standard, and sample solution shows that there is no interference between the spectra of the diluent and placebo with metformin HCl, indicating the specificity of the method (see Figure 5).</p><p>The accuracy expresses the agreement between the accepted value and the observed value. The method is accurate (the recovery values are within 95.0–105.0%, and RSD is not more than 5.0%) and precise (RSD values are less than 2%, and F value <5.05) [13, 15]. To assess linearity, a standard curve was constructed by plotting average absorbance versus concentration.</p><p>The correlation coefficient (r2) was greater than 0.995 for the calibration curve over the range of 25 to 150% of the working concentration, indicating a good linearity of the method.</p><p>In all the deliberately varied conditions (pH value, rotation speed, and temperature of dissolution medium), the released amount of metformin HCl from the dosage form remained unchanged, which demonstrates that the developed method is robust.</p><p>The results of the stability studies showed that the standard solution of the metformin HCL was found to be stable for 48 h at room temperature.</p><!><p>The developed dissolution method (Apparatus II with a rotation speed of 75 rpm and 1000 mL of phosphate buffer pH 6.8 as a medium) has a discriminatory power for metformin film-coated tablets and is able to distinguish any changes related to quantitative formulation of the pharmaceutical dosage form, applicable for all dosage strengths. The method can be used in the formulation development studies, as well as for the selection of biobatch for the bioequivalent study. The proposed method is specific, accurate, precise, linear, and robust and can be successfully applied for evaluation of the batch-to-batch quality and stability of the medicinal product.</p><!><p>The data used to support the findings of this study are included within the article.</p><!><p>The authors declare that they have no conflicts of interest.</p><!><p>Visual observation of metformin 1000 mg film-coated tablets using paddles with rotation speed (a) 50 rpm and (b) 75 rpm.</p><p>Dissolution profile of metformin 1000 mg film-coated tablets using different apparatus and volume of the medium.</p><p>Dissolution profiles of original formulation and formulation with a deliberate change in the composition of excipients (mismanufactured).</p><p>Dissolution profiles of different strengths of metformin film-coated tablets.</p><p>Specificity of UV spectrophotometric method for dissolution of metformin HCl. (a) Diluent. (b) Placebo. (c) Standard solution. (d) Sample solutions.</p><p>Results from the solubility test of metformin HCL in the proposed dissolution media.</p><p>Calculated from 1000 mg/250 mL × 3 = 12 mg/mL.</p><p>The obtained results of the dissolution test of metformin 1000 mg film-coated tablets using different apparatus and volume of the medium.</p><p>Validation of the proposed dissolution method for metformin film-coated tablets.</p>

PubMed Open Access

EOM-CC guide for Fock space travel: The C 2 edition

Despite their small size, C 2 species pose a big challenge to electronic structure owing to extensive electronic degeneracies and multi-configurational wave functions leading to a dense manifold of electronic states. We present detailed electronic structure calculations of C 2 , C − 2 , and C 2− 2 emphasizing spectroscopically relevant properties. We employ double ionization potential (DIP) and ionization potential (IP) variants of equation-of-motion coupled-cluster method with single and double substitutions (EOM-CCSD) and a dianionic reference state. We show that EOM-CCSD is capable of describing multiple interacting states in C 2 and C − 2 in an accurate, robust, and effective way. We also characterize the electronic structure of C 2− 2 , which is metastable with respect to electron detachment.

eom-cc_guide_for_fock_space_travel:_the_c_2_edition

I. INTRODUCTION<!>II. MOLECULAR ORBITAL FRAMEWORK AND ESSENTIAL FEATURES OF ELECTRONIC STRUCTURE OF C<!>EOM-DIP<!>III. COMPUTATIONAL DETAILS<!>IV. RESULTS AND DISCUSSION<!>State Configuration<!>V. CONCLUSION

<p>Ironically, the smallest form of neat carbon, the C 2 molecule, features the most complex electronic structure. The complexity stems from the inability of four valence electrons of carbon to form a quadrupole bond (remarkably, the bonding in C 2 is still hotly debated [1][2][3][4][5][6][7][8] ).</p><p>Because the optimal electron pairing cannot be reached, multiple electronic configurations have similar likelihood, leading to a dense manifold of low-lying electronic states. This results in rich spectroscopy: C 2 features multiple low-lying electronic transitions, which have been extensively studied experimentally [9][10][11][12][13][14] . Despite a long history of experimental work, C 2 continues to generate interest. For example, recently new band systems have been identified [15][16][17] .</p><p>Besides obvious fundamental importance, C 2 (and its anionic forms, C − 2 and C 2− 2 ), play a role in combustion 18 , plasma [19][20][21] , and astrochemistry 19,22 . For example, C 2 and C − 2 have been observed in comet tails, protoplanetary nebulae, the atmospheres of stars, and in the diffuse interstellar medium [22][23][24][25][26][27] . C 2 is responsible for the color of blue flames 18 . It is also a prominent product of electrical discharges containing hydrocarbons 20 .</p><p>From the theoretical point of view, C 2 is arguably the most difficult molecule among homonuclear diatomics from the first two rows of the periodic table. Electronic neardegeneracies lead to multiconfigurational wave-functions. Small energy separations between different electronic states also call for high accuracy. Because of its complex electronic structure, C 2 has been often described as a poster child of multi-reference methodology. The availability of high-quality spectroscopic data, complex electronic structure, and its small size make C 2 a popular benchmark system for quantum chemistry studies [28][29][30][31][32] . Among recent theoretical studies of the low-lying states of C 2 , the most comprehensive are tour-de-force MR-CISD (multi-reference configuration interaction with single and double excitations) calculations by Schmidt and coworkers 33 and by Szalay and co-workers 34 . In both studies, the effect of basis set and higher-order corrections have been carefully investigated. To correct MR-CISD energies for violation of size-extensivity, Davidson's quadruple correction was used. Szalay and co-workers have also reported results obtained with an alternative strategy, the so-called MR-average quadratic coupled-cluster (AQCC) method. In both studies, the theoretical values of the reported equilibrium distances (r e ) and term energies (T ee ) agreed well with the experimental data.</p><p>The anionic forms of C 2 , C − 2 and C 2− 2 , have received less attention. C − 2 is produced in plasma discharge from acetylene 35,36 . Electronically excited C − 2 has been observed in a carbon-rich plasma via fluorescence 21 . Recently, C − 2 has been proposed as a candidate for laser cooling of anions 37 , which makes these species interesting in the context of quantum information storage. Ervin and Lineberger 38 have measured photoelectron spectrum of C − 2 using 3.53 eV photons; they reported adiabatic electron affinity (AEA) of C 2 to be 3.269±0.006 eV. A similar value (3.273±0.008 eV) has been derived by Neumark and coworkers 39 , who reported vibrationally resolved photodetachment spectra using 4.66 eV radiation. Feller has reported an AEA of 3.267 eV calculated using a composite method based on coupled-cluster (CC) methods 40 .</p><p>Because of the highly unsaturated character of C 2 , it has relatively large electron attachment energy so that even the two lowest excited states of C − 2 are bound electronically. In contrast, C 2− 2 is metastable with respect to electron detachment. Its existence has been postulated on the basis of features observed 41,42 in electron scattering from C − 2 and confirmed by calculations 43,44 .</p><p>In this contribution, we present detailed electronic structure calculations of C 2 , C − 2 , and C 2− 2 , with an emphasis on spectroscopically relevant properties. We employ an alternative methodology based on CC and equation-of-motion CC (EOM-CC) theory [45][46][47][48][49] . We show that electronic states of C 2 and C − 2 are well described by the double ionization potential (DIP) 50 and ionization potential (IP) 51,52 variants of EOM-CCSD (EOM-CC with single and double substitutions) using a dianionic reference state. Formulated in a strictly singlereference fashion, the EOM-CC family of methods provides an accurate, robust, and effective alternative to cumbersome multi-reference calculations [45][46][47][48][49] . To describe metastable species, such as C 2− 2 , we employ the complex-variable extension of CCSD and EOM-CCSD via the complex absorbing potential (CAP) approach [53][54][55] . Due to orbital near-degeneracies, various electronic configurations of six electrons over the upper four orbitals have similar energies, leading to closely lying electronic states and multi-configurational wave-functions. In C − 2 , there are four important configurations in which the unpaired electron resides on one of the upper orbitals. In C 2− 2 , which is isoelectronic with N 2 , all four upper orbitals are doubly occupied, resulting in the</p><!><p>2 is a well-behaved closed-shell state dominated by a single Slater determinant; thus, it can be well described by single-reference methods such as, for example, CCSD. From this reference state, EOM-IP and EOM-DIP operators can generate all important electronic configurations needed for describing the electronic states of C − 2 and C 2 , respectively, as illustrated in Fig. 2. Mathematically, the EOM-CCSD target states are described by the following ansatz [46][47][48]</p><!><p>where e T1 + T2 Φ 0 is the reference CCSD wave function (the amplitudes of the excitation operator T are determined by the CCSD equations for the reference state) and operator R is a general excitation operator. In EOM-IP-CCSD, R comprises all 1h (one hole) and 2h1p (two hole one particle) operators 51,52 , whereas in EOM-DIP-CCSD it includes all 2h and 3h1p operators. In EOM-EE-CCSD (EOM-CCSD for excitation energies 56 ) and EOM-SF-CCSD (spin-flip EOM-CCSD 57,58 ), R is particle-conserving and includes 1h1p and 2h2p operators (in the SF variant, R changes the number of α and β electrons). In the EA (electron attachment) variant 59 , the operator R is of the 1p and 1h2p type. The amplitudes of R are found by diagonalization of the similarity-transformed Hamiltonian, H:</p><p>Linear parameterization ensures that different configurations can mix and interact. There are no assumptions about their relative importance-the relative weights of different configurations are defined by the EOM eigen-problem and can span the entire range of situations, from the cases dominated by a single electronic configuration to the cases of equal contributions from multiple determinants. The EOM-CC ansatz is capable of reproducing exact degeneracies (such as between the two components of Π states in linear molecules or Jahn-Teller degeneracies), which are violated by state-specific MR treatments. Since all important configurations appear at the same excitation level, they are treated in a balanced way. As a multi-state method, EOM-CC produces the entire manifold of electronic states, without requiring user input regarding state character. These features of EOM-CC make it very attractive for treating multiple electronic states and extensive degeneracies 49 . Among recent applications illustrating the power of EOM-CC, we mention calculations of electronic states of copper oxide anions 60 , Cvetanović diradicals 61 , and molecules with several unpaired electrons 62,63 .</p><p>The success of EOM-CC in treating a particular electronic structure depends on whether a proper well-behaved reference can be found from which the target-states manifold can be reached by an appropriately chosen R1 . As illustrated in Fig. 2, the electronic structure of C 2 is best described by EOM-DIP using the dianionic reference state. The DIP method is capable of describing electronic degeneracies beyond two-electrons-in-two-orbitals or threeelectrons-in-three-orbitals patterns 50,60,61,[64][65][66][67][68] , however, its applications are limited by the complications due to the use of the dianionic reference.</p><p>Isolated dianions of small molecules are usually unstable with respect to electron detachment and exist only as transient species. 69 In dianions, the competition of long-range repulsion between an anionic core and an extra electron versus stabilizing valence interactions with short-range character leads to a repulsive Coulomb barrier. The extra elec-tron is trapped behind this barrier but can leave the system by tunneling. This is similar to metastable radical monoanions where the extra electron is trapped behind an angularmomentum barrier also affording resonance character. In a computational treatment using a sufficiently large basis, the wave function of a resonance becomes more and more diffuse, approximating a continuum state corresponding to an electron-detached system and a free electron [70][71][72] .</p><p>Resonances can be described by a non-Hermitian extension of quantum mechanics 73 by using, for example, complex absorbing potential (CAP) 74,75 . If one is interested in the dianionic state itself, then the CAP-based extension of CC theory can be used 55 . However, in practical calculations using EOM-DIP-CC, the dianionic state just serves as a reference for generating target configurations. Thus, less sophisticated approaches can be used to mitigate complications due to its metastable character. The easiest and most commonly used one is to use a relatively small basis set, such that the reference state is artificially stabilized 50,60,61,[64][65][66]76 . Kuś and Krylov have investigated an alternative strategy, stabilization of the resonance using an artificial Coulomb potential with a subsequent de-perturbative correction 71,72 . Here we show that in the case of C 2 using the aug-cc-pVTZ basis provides a robust description of the dianionic reference, delivering accurate results for the target states.</p><p>To further validate these calculations, we carried out CAP-EOM-IP-CCSD calculations in which the dianionic reference is stabilized by the CAP and compare the potential energy curves of C 2− 2 and C − 2 obtained by these two calculations. In the CAP approach 74,75 , the Hamiltonian is augmented by a purely imaginary confining potential iηW (the parameter η controls the strength of the potential). This transformation converts the resonances into L 2 -integrable wave functions with complex energies</p><p>where the real and imaginary parts correspond to the resonance position (E res ) and width (Γ). In a complete basis set, the exact resonance position and width can be recovered in the limit of η → 0. In finite bases, the resonance can only be stabilized at finite values of η.</p><p>The perturbation due to the finite-strength CAP can be removed by applying first-order deperturbative correction 53,54 and identifying the special points of η-trajectories at which the dependence of the computed energy on η is minimal. When combined with the EOM-CCSD ansatz, this approach has been shown to yield accurate and internally consistent results for both bound and metastable states 55 . For example, these calculations yield smooth potential energy curves [77][78][79] and in many cases correctly identify the points where resonances become bound. We note, however, that in some polyatomic molecules spurious widths of about 0.04 eV for bound states persist 79 . In our previous calculations 53,55,[78][79][80][81][82] , we used CAP-EOM-CCSD to describe metastable EOM states from stable (bound) CCSD references. Here we present the first example of a calculation where the CCSD reference is metastable, but the target EOM-CCSD states are bound.</p><!><p>As explained above, we describe the electronic states of C − 2 and C 2 by EOM-IP-CCSD and EOM-DIP-CCSD, respectively, using the dianionic reference (see Fig. 2). In realvalued EOM-CCSD calculations, we used the aug-cc-pVTZ basis. In the CAP-augmented CCSD and EOM-IP-CCSD calculations, we used the aug-cc-pVTZ+3s3p and aug-cc-pCVTZ+6s6p6d basis sets (the exponents of the additional diffuse sets were generated using the same protocol as in our previous studies 54,81 ). Two core orbitals, σ 1s and σ * 1s , were frozen in correlated calculations except when employing the aug-cc-pCVTZ basis. In the calculations using aug-cc-pVTZ+3s3p, the CAP onset was set according to the expectation value of R 2 of the triplet UHF wave function of C 2 (at r cc =1.28 Å, this gives the following onsets:</p><p>x 0 = y 0 =1.6 Å, z 0 = 2.6 Å). In the calculations with aug-cc-pCVTZ+6s6p6d, the CAP onset was set according to the expectation value of R 2 of the dianion computed using CCSD/augcc-pCVTZ (at r cc =1.2761 Å, this gives x 0 = y 0 = 2.4 Å, z 0 = 3.6 Å). First-order correction 53 was applied to the computed total energy and then optimal values of η were determined from these corrected trajectories using our standard protocol 53,54 . All electronic structure calculations were carried out using the Q-Chem package 83,84 . The calculations of partial widths were carried out using ezDyson 85 .</p><!><p>A. C 2 Bond length (Å) The results illustrate that EOM-DIP-CCSD is capable of tackling the complexity of C 2 rather well. It describes the entire manifold of the low-lying states with an accuracy comparable to that of much more cumbersome and labor-intensive multi-reference calculations.</p><p>When compared to the experimental values, the root-mean-square (RMS) errors in the equilibrium bond lengths and term energies computed with EOM-DIP-CCSD/aug-cc-pVTZ are 0.0165 Å and 1661 cm −1 . The errors in bond length are only marginally bigger than those of MR-CISD+Q/cc-pVTZ values (0.0114 Å). Remarkably, the errors in energy are consistently smaller than a conservative estimate of EOM-CCSD error bars, which is roughly 0.3 eV (2420 cm −1 ). The relative state ordering is also correctly described. MR-CISD+Q/cc-pVTZ yields, on average, smaller errors in term energies (RMS of 469 cm −1 ), however, for three out of nine states, the EOM-DIP-CCSD/aug-cc-pVTZ values are closer to the experiment.</p><p>For a fair comparison, it is important to stress that the EOM-DIP-CCSD ansatz is very compact and includes only 2h and 3h2p configurations, whereas in MR-CISD+Q and AQCC, the size-extensivity corrections entail contributions of up to quadruply excited configurations. As with other EOM-CCSD methods, perturbative or explicit inclusion of connected c From Refs. 9-14.</p><p>triple excitations is expected to significantly reduce the errors. We note that higher excitations can also describe orbital relaxation thus mitigating the effect of an unstable dianionic reference.</p><p>To put the results presented in Table I in The size-extensivity correction is significant-the errors of MR-CISD decrease when either Davidson's correction or MR-AQCC is employed. Without size-extensivity corrections, the RMS in the equilibrium bond lengths and term energies computed with MR-CISD/cc-pVTZ are 0.0117 Å and 623 cm −1 . The effect of the basis set on the term energies is less systematic 34 . The RMS error in bond lengths within MR-AQCC/cc-pVTZ is 0.0115 Å (to be compared to 0.0114 Å of MR-CISD+Q). The errors in term energies were also comparable to MR-CISD+Q/cc-pVTZ. We note that in the MR-AQCC(TQ) calculations, the largest errors in term energies were observed for 1 ∆ u and e 3 Π g (999 cm −1 and 722 cm −1 ). Both MR-AQCC and MR-CISD+Q calculations were sensitive to the orbital choice and showed improved performance when using state-averaged CASSCF orbitals. Extrapolation to the complete basis set based on the cc-pVTZ and cc-pVQZ calculations results in a systematic decrease of equilibrium bond lengths by 0.01 Å.</p><p>Several studies have also investigated the magnitude of higher-order corrections, with an aim to achieve spectroscopic accuracy 31,86 . Schmidt and co-workers showed that the inclusion of core-valence correlation combined with scalar relativistic corrections in the framework of MR-CISD+Q brings the spectroscopic constants within 1% from the experimental values 33 .</p><p>Jiang and Wilson have reported similar trends 31 .</p><p>In addition to the states shown in Table I, we also computed two electronic states, 1 ∆ u and e 3 Π g , which have been recently identified experimentally [15][16][17] . The electronic configurations of these states are:</p><p>Thus, they cannot be generated by the 2h operator from the dianionic reference and the norm of the 3h1p EOM amplitudes becomes large (≈1). Consequently, the computed term energies are too high. In order to describe these states with the same accuracy as the states dominated by 2h configurations, the EOM-DIP ansatz needs to be extended up to 4h2p operators.</p><p>We note that several lowest state of C 2 can also be described by EOM-SF-CCSD using a high-spin triplet reference, e.g., [core] 6 (σ * 2s ) 2 (π 2px ) 2 (π 2py ) 1 (σ 2pz ) 1 . Using ROHF-EOM-SF-CCSD/aug-cc-pVTZ, vertical excitation energy from 1 Σ + g to a 3 Π u at r cc =1.2425 of C 2 is 319 cm −1 , to be compared with 1924 cm −1 computed by EOM-DIP-CCSD/aug-cc-pVTZ.</p><p>To quantify the bonding pattern in C 2 , we also computed Head-Gordon's index 87 , which characterizes the number of effectively unpaired electrons. For the EOM-SF-CCSD wave function of the ground state of C 2 at equilibrium, n u,nl =0.29. This value indicates that C 2 has substantial diradical character, comparable 63 to that of singlet methylene (0.25) or meta-benzyne (0.26). In other words, there is no support for a quadruple bond, which would be manifested by n u,nl ≈ 0. TABLE II: Equilibrium bond lengths (r e , Å) and term energies (T ee , cm −1 ) of bound electronic states of C − 2 . EOM-IP-CCSD vertical excitation energies (E ex , cm −1 ) and oscillator strengths (f l ) are also shown.</p><!><p>EOM-IP-CCSD/aug-cc-pVTZ Expt.</p><p>a Fig. 4 shows the potential energy curves of the three bound states of C − 2 computed using EOM-IP-CCSD/aug-cc-pVTZ. The respective electronic configurations, equilibrium distances, and term values are given in Table II. The Dyson orbitals 89 representing the unpaired electron in C − 2 are shown in Fig. 1.</p><p>As one can see, the computed equilibrium distances and term energies are in excellent agreement with the experimental data. The computed oscillator strengths show that transitions to both excited states are optically allowed. The computed T ee of the 2 Σ + u → 2 Σ + g transition is 2.37 eV. Vertically, at the equilibrium geometry of the 2 Σ + u state, the energy gap between two states is 2.29 eV, which is exactly equal to the fluorescence signal observed in Ref. 21. Thus, our results confirm that fluorescence observed in Ref. 21 can be attributed</p><p>We also computed adiabatic electron affinity, AEA, of C 2 . Using EOM-DIP-CCSD/augcc-pVTZ total energy of X 1 Σ + g and EOM-IP-CCSD/aug-cc-pVTZ total energy of the X 2 Σ + g state at the respective r e , the computed value of AEA is 4.57 eV (without zero-point energy), which is more than 1 eV larger than the experimental value 38,39 of 3.27 eV and high-level ab initio estimates 40 . This suggests that the current correlation level is insufficient to describe relative position of the two manifolds. The two relevant states, X 1 Σ + g and 2 Σ + g , can also be computed using an alternative EOM-CC scheme, via SF and EA using the high-spin</p><p>These calculations yield AEA of 3.44 eV when using UHF triplet reference and 3.42 eV eV when using the ROHF reference. The analysis of the total energies shows that the EOM-EA energy of the anion is very close to the corresponding EOM-IP energy whereas the EOM-SF energy of the neutral state is significantly lower than the EOM-DIP energy. We attribute this to orbital relaxation effects-while the dianionic orbitals are reasonably good for the anion, they are too diffuse for the neutral and the EOM-DIP ansatz with only 2h and 3h1p operators is not sufficiently flexible to account for that. To characterize lifetimes of the dianion and to quantify the effect of its resonance character on the computed quantities of C − 2 , we carried out CAP-CCSD and CAP-EOM-IP-CCSD calculations. The results are summarized in Tables III and IV and shown in Figs. 5 and 6.</p><p>As one can see from Fig. 5, the total energies of C 2− 2 obtained from the CAP-augmented calculations are nearly identical to the real-valued results. Moreover, the impact on the computed term energies of C − 2 is also small: at r CC =1.28 Å, the differences in excitation energies of C − 2 between the two calculations are ∼0.03 eV. The adiabatic energy gap between C 2− 2 and C − 2 is 3.16 eV computed with CAP-CCSD/aug-cc-pCVTZ+6s6p6d, only slightly smaller than the value obtained in real-valued calculations (3.41 eV).</p><p>Previous calculations using the charge-stabilization method 43 estimated the closed-shell 1 Σ + g resonance of C 2− 2 below 4 eV, roughly at around 3.4 eV, above the ground state of C − 2 . Later, CAP-augmented MR-CISD calculations 44 yielded E res = 3.52 eV and an equilibrium bond length of 1.285 Å. Thus, our results confirm the findings of these earlier studies 43,44 .</p><p>The resonance position and width are rather sensitive to the basis set employed, as Table III illustrates. For example, at the equilibrium bond length (r CC =1.28 Å), the aug-cc-pVTZ+3s3p basis yields adiabatic E res =3.7 eV and Γ=0.68 eV, whereas the aug-cc-pCVTZ+6s6p6d basis produces E res =3.16 eV and Γ=0.25 eV. A distinct stabilization point of the η-trajectory is only obtained using the larger basis set (see Fig. 6); in the small basis only first-order corrected trajectory shows stabilization point. Our best value for the resonance width (0.25 eV) is in very good agreement with the CAP-MR-CISD value (0.30 eV) 44 and also agrees qualitatively with the estimate from charge-stabilization calculations (0.26-0.55 eV) 43 Im(E)/a.u.</p><p>Re(E)/a.u.</p><p>without correction including first−order correction FIG. 6: Uncorrected and first-order corrected CAP-CCSD using aug-cc-pCVTZ+3s3p (left) and aug-cc-pCVTZ+6s6p6d (right) η-trajectories for C 2− 2 at equilibrium bondlengths.</p><p>We also estimated partial widths corresponding to the three decay channels. Within Feshbach formalism, partial widths of autodetachment can be approximated by the following matrix element 82 :</p><p>where Γ c is the partial width corresponding to detachment channel c, ω c and φ d c are the respective detachment energy and Dyson orbital, ξ ωc is the wave function of the free electron, and F is the Fock operator. Given the localized nature of F , this matrix element is bound by the value of the overlap between the Dyson orbital and the free-electron wave function. Thus, branching ratios x p corresponding to different detachment channels can be estimated as follows:</p><p>giving rise to Γ p = x p Γ. Note that the contributions from the degenerate channels (such as Π u ) should be multiplied by the respective degeneracy number (2 for Π-states). The overlap</p><p>is proportional to the norm of φ d c and is strongly dependent on the energy of the detached electron and the shape of the Dyson orbital. Fig. 7 shows the energy dependence of the computed values of the squared overlap between the normalized Dyson orbitals and the free-electron wave function approximated by the Coulomb wave. As one can see, the overlap values are zero at low detachment energies and increase at higher energies. The trends for the Σ u and Π u channels are very similar, which is not surprising given the similar shape of the respective Dyson orbitals. Fig. 7 immediately suggest that the autodetachment process will be dominated by the channels producing the two lowest states of the anion, Σ g and Π u . Table ?? lists the computed values using E res =3.41 eV (from EOM-IP-CCSD/augcc-pVTZ). As one can see, the contribution of the Σ u is negligible and the Σ g channel is dominant. When using lower energy value (3.16 eV, from CAP-EOM-IP-CCSD/augcc-pCVTZ+6s6p6d), the contribution from the Σ u channels drops even further while the ratio between the Σ g and Π u channels remains unchanged. Using Dyson orbitals from the CAP-EOM-IP-CCSD/aug-cc-pCVTZ+6s6p6d calculations leads to the increase of the relative weight of the Σ g channel. These simple estimates are in qualitative agreement with partial widths computed using CAP-MR-CISD wave function and an approach based on CAP projection 44 ; their reported values correspond to x p of 0.31, 0.66, and 0.02 for the Σ g , Π u , and Σ u channels. One important difference is that our calculations predict that the dominant decay channel is Σ g , producing the ground-state of C − 2 . We note that using plane wave to describe the state of the free electron yields an entirely different picture: the overlaps are rather large around the threshold and change much slower, resulting in comparable branching ratios for all three channels. Finally, we investigate the dependence of the resonance width on the bond length. As illustrated by Figure 8, the CAP-CCSD resonance width shrinks with an increasing bond length near the equilibrium distance while it is nearly constant beyond 1.6 Å. This is consistent with the potential energy curves of C 2− 2 and the 2 Σ + g and 2 Π u states of C − 2 becoming nearly parallel at elongated bond distances (see Figure 5). However, the behavior is different from that of valence shape resonances in diatomic molecules (for example, H − 2 or N − 2 ) that become bound if the bond is stretched somewhat. 55 It is more reminiscent of dipolestabilized resonances whose width is also insensitive towards bond length changes. 80 Figure 8 also shows that the resonance width behaves differently at the CAP-CCSD and CAP-HF levels. Within the HF approximation, Γ has a minimum around the equilibrium structure (0.08 eV) and grows when the bond is stretched. This behavior is similar to the results reported in Ref. 44 where CAP-CIS and CAP-MR-CISD also yielded Γ increasing with bond length between 1.2 and 1.4 Å. A detailed investigation of these differences is beyond the scope of the present work, but we note that the resonance width of C 2− 2 has to vanish eventually, when the bond is stretched far enough, because the 4 S ground state of C − obtained in the dissociation limit is stable towards electron detachment.</p><p>The description of the decay channels reveals a shortcoming of the CAP-CCSD approach based on a metastable reference. The CAP-EOM-IP-CCSD energies of the three bound states of C − 2 feature sizable positive imaginary parts of more than 0.3 eV (at the equilibrium bond length and optimal η values for the dianionic resonance). This is despite that real parts of absolute CAP-EOM-IP-CCSD energies agree with the CAP-free values within ∼0.1 eV. Also, it is in stark contrast to the performance of CAP-EOM-CCSD based on bound reference states 53,54 , where the imaginary energies of bound states typically stay below 0.03 eV. We note that application of the de-perturbative correction 53,54 does not rectify this problem. This is not surprising as the original analysis of E(η) in terms of perturbation theory 75 was designed for resonances but not bound states. Furthermore, the imaginary energies of the three bound states of C − 2 differ by more than a factor of two so that a single, not state-specific, correction is not realistic. However, since a positive imaginary energy is unphysical and since no stabilization of the η-trajectory is observed for the CAP-EOM-IP-CCSD states, the problem is easily discernible. Importantly, despite this shortcoming, CAP-EOM-CCSD calculations using an unstable reference clearly distinguish bound and metastable states.</p><p>Experimentally 41,42 , the C 2− 2 resonance manifest itself as a broad feature around 10 eV in electron scattering detachment spectra from C − 2 , however, the interpretation of these spectra in terms of the position of the resonance is not straightforward, as explained by Sommerfeld and co-workers 43 . We hope that our results will stimulate further experimental efforts to characterize electronic structure of C 2− 2 . a Adiabatic EOM-IP-CCSD/aug-cc-pVTZ energies (eV).</p><p>b Overlap (squared) is computed between normalized Dyson orbitals and the Coulomb wave with charge=-1 and kinetic energy corresponding to adiabatic detachment energy.</p><!><p>We reported electronic structure calculations of C 2 , C − 2 , and C 2− 2 using the CC/EOM-CC family of methods. The results illustrate that EOM-CCSD provides an attractive alternative to MR approaches. The low-lying states of C 2 and C − 2 are well described by EOM-DIP-CCSD and EOM-IP-CCSD using dianionic closed-shell reference (C 2− 2 ), despite its metastable nature.</p><p>EOM-DIP-CCSD offers a much simpler computation approach based on a single-reference formalism. In the EOM-DIP calculations, no active space selection is required and the results of the calculations do not depend on the number of states computed, in contrast to stateaveraged MR schemes. One does not need to guess what are the electronic configurations of the states to be computed-once the user specifies how many states in each irrep are desired, the algorithm will compute these states.</p><p>The electronic structure of C 2− 2 was characterized by CAP-augmented CCSD. The calculations place the closed-shell C 2− 2 resonance 3.16 eV adiabatically above the ground state of C − 2 . The computed resonance width is (0.25 eV), which corresponds to a lifetime of 2.6 fs. Importantly, the CAP-augmented calculations yield detachment energies that are very close to the real-valued EOM-CCSD calculations with the aug-cc-pVTZ basis set thus confirming the validity of the results obtained with EOM-DIP-CCSD and EOM-IP-CCSD using the dianion reference.</p>

ChemRxiv

Obtusaquinone: A Cysteine-Modifying Compound That Targets Keap1 for Degradation

We have previously identified the natural product obtusaquinone (OBT) as a potent antineoplastic agent with promising in vivo activity in glioblastoma and breast cancer through the activation of oxidative stress; however, the molecular properties of this compound remained elusive. We used a multidisciplinary approach comprising medicinal chemistry, quantitative mass spectrometry-based proteomics, functional studies in cancer cells, and pharmacokinetic analysis, as well as mouse xenograft models to develop and validate novel OBT analogs and characterize the molecular mechanism of action of OBT. We show here that OBT binds to cysteine residues with a particular affinity to cysteine-rich Keap1, a member of the CUL3 ubiquitin ligase complex. This binding promotes an overall stress response and results in ubiquitination and proteasomal degradation of Keap1 and downstream activation of the Nrf2 pathway. Using positron emission tomography (PET) imaging with the PET-tracer 2-[18F]fluoro-2-deoxy-D-glucose (FDG), we confirm that OBT is able to penetrate the brain and functionally target brain tumors. Finally, we show that an OBT analog with improved pharmacological properties, including enhanced potency, stability, and solubility, retains the antineoplastic properties in a xenograft mouse model.

obtusaquinone:_a_cysteine-modifying_compound_that_targets_keap1_for_degradation

<!>OBT Forms Reversible Covalent Adducts with Thiols.<!>OBT Activates the Nrf2 Pathway in Vitro and in Vivo.<!>OBT Is a Cysteine-Modifying Drug Targeting Keap1.<!>OBT Promotes the Degradation of Keap1.<!>OBT and Its Analog Effectively Target Tumors in Preclinical Mouse Models.<!>DISCUSSION<!>Reagents.<!>Cell Culture.<!>Cell Viability.<!>Statistical Analysis.

<p>The transcription factor Nrf2 (Nuclear factor erythroid 2 (NFE2)-related factor 2) plays a key role in maintaining cellular homeostasis in response to oxidative stress by regulating the expression of antioxidant response element (ARE) dependent genes.1 Keap1 (Kelch-like ECH-associated protein 1) is recognized as a predominant negative regulator of Nrf2 and functions as a substrate adaptor protein for the ubiquitin ligase CRL3 (cullin 3 (CUL3)-RING ubiquitin ligase). During homeostasis, Keap1 recruits Nrf2 to CUL3 thereby promoting its ubiquitination and subsequent proteasomal degradation. Nrf2 has been recognized to exert either protumorigenic or antitumorigenic properties.2 This apparent contradiction can be rationalized by differences in the cell state and the functional dependence on Nrf2 activation. Oncogenic activity is generally associated with constitutive activation of Nrf2 caused either by overexpression or somatic mutations of Nrf2 and/or other regulatory proteins.3,4 Tumor suppressor activity, in contrast, is linked to transient activation of Nrf2, e.g., by small molecules.3 In this context, the Nrf2-Keap1 module has been validated and successfully pursued as a target for small molecules that disrupt Nrf2 binding and/or induce the dissociation of Keap1 from CUL3.5,6 These efforts have identified several inhibitor classes, including cysteine-reactive natural products and synthetic compounds, that bind Keap1 and activate the Nrf2 pathway.7</p><p>We have previously identified the quinone methide obtusaquinone (OBT) as a potent antineoplastic agent with selectivity over normal cells for glioblastoma (GBM) and several other cancer types.8 Within the scope of this research, we have demonstrated that OBT treatment increases the production of reactive oxygen species (ROS) and induces DNA damage, leading to apoptotic cell death. However, the molecular mode of action of OBT, its medicinal chemistry, and in vivo pharmacology have not been understood, which has impeded preclinical development. We here developed novel OBT analogs with improved pharmacological properties and show that OBT is a thiol-reactive compound that reacts reversibly with cysteine residues and particularly binds to Keap1, leading to CRL3-mediated autoubiquitination and proteasomal degradation of Keap1 thereby activating the Nrf2 pathway.</p><!><p>OBT features a 2-hydroxy-p-quinone methide, a moiety found in other natural products such as celastrol,9 which efficiently reacts with thiol nucleophiles including cysteine side chains to form substituted catechols (Supplemental Figure 1A). It has previously been shown that the desmethoxy analog of OBT SI1 (Supplemental Figure 1B) reacts with glutathione (GSH) in aqueous buffer to form four distinct addition products, corresponding to the diastereomers derived from direct SI2a,b and vinylogous SI3a,b addition.10 To demonstrate that OBT retains the ability to form sulfhydryl adducts, we incubated OBT with β-mercaptoethanol (BME) in ethanol and found that BME was readily added to OBT, preferentially (10:1) forming the direct addition product over the vinylogous addition product (AF20, Figure 1A; AF20a and AF20b, Supplemental Figure 1C). To investigate the reversibility of the thionucleophile addition to OBT, we incubated AF20 in the presence of 5-fold excess cysteamine. Monitoring the reaction mix by LC/MS showed the formation of the corresponding amine functionalized analogs SI4a and SI4b, demonstrating that the addition of thiols is reversible or that the substitution proceeds through an SN2′ mechanism and that the corresponding adducts exist in a dynamic equilibrium (Supplemental Figure 1C,D).</p><p>Next, we explored if the reactivity with thiols could be exploited for the development of novel OBT analogs with improved pharmacological properties to overcome the limited solubility of OBT in aqueous media and to buffer the abrupt oxidative stress as a result of rapid depletion of the intracellular GSH pool caused by thiol-reactive compounds.11 We have previously shown that addition of GSH or N-acetyl cysteine (NAC) reduces the activity of OBT in cell culture, likely by extracellular scavenging of OBT.8 Based on our hypothesis that the ability to react with cysteine side chains is critical for OBT activity, we postulated that prodrugs designed to liberate OBT or analogs that retain the ability to react with cysteines would yield improved inhibitors, including compounds with enhanced solubility, while mitigating the general toxicity as a result of GSH depletion. AF20, the adduct of OBT and BME, which is more soluble and liberates OBT in PBS, potently killed patient-derived glioma stemlike cell (GSC) neurospheres, while exhibiting lower toxicity toward primary human astrocytes (HA) (Supplemental Figure 2A).8 As previously observed with OBT, the activity of AF20 is reversed in the presence of NAC (Supplemental Figure 2B). These results are consistent with our proposed mechanism that AF20 converts to OBT as an active compound and that NAC may function not only as a ROS antagonist but also as a direct scavenger of OBT.</p><p>To block the direct reversibility observed with AF20, we speculated that acetylation of the catechol would stabilize the thiol adduct in aqueous media but predicted that the phenolic esters would be cleaved intracellularly allowing for intracellular release of OBT via AF20 or following displacement by an SN2′ mechanism (see the Supporting Information). Treatment of AF20 with 2 equiv and 3 equiv of acetic anhydride yielded AEN36 Bis and AEN36 Tris (Figure 1A), respectively. Both compounds demonstrated increased stability and increased potency on different cancer cells as compared to OBT (Figure 1B and Supplemental Figure 2C).</p><p>To confirm reversible cysteine modification by OBT, we established a mass spectrometry-based approach that would allow the identification of specific cysteine side chains in native proteins that covalently react with OBT and differentiate reversible from irreversible interactions. The experimental setup is outlined in Figure 1C,D, and the potential outcome is outlined in Figure 1E. The method is based on serial exposure of a protein to OBT, the cysteine-alkylating reagents iodoacetamide (IAA) to monitor derivatization with OBT, to dithiothreitol (DTT) to probe reversibility of OBT binding, and to another alkylating reagent, N-methylmaleimide (NEM), to allow a readout of probing the reversibility. In two parallel reactions, OBT is replaced by IAA and NEM. These reactions are used as standards to enable a final reaction read-out by multiplexed quantitative mass spectrometry.12,13 We established a rule set to allow an unambiguous interpretation of the data (Supplemental Table 2) and assigned three different binding types for OBT: (1) irreversible binding, with OBT and IAA channels at an intensity at least 2-fold lower than the NEM channel and the OBT channel not being significantly (p ≤ 0.01) higher than the IAA channel intensity; (2) partially reversible binding, as irreversible binding but with an OBT channel intensity significantly higher than the IAA intensity; (3) undefined binding, the IAA channel intensity is at least as high as the NEM channel intensity. We performed this experiment using bovine catalase, a cysteine-rich antioxidant (Supplemental Figure 2D), and found that OBT binds to all detected peptides with cysteine residues (Figure 1F; Supplemental Tables 1 and 2). For Cys-377, where the cysteine is followed by a proline in the protein sequence, we observed reversible binding, while the two other monitored cysteine residues (Cys-232 and Cys-460) showed tight or irreversible binding under the tested reaction conditions (Figure 1F). Unexpectedly, we observed a high IAA signal in the NEM-conjugate, which could be attributed to deviations in the experimentally determined and predicted intensity patterns, likely due to disulfide bonds reduced and realkylated during the experiment. These results show that OBT is an effective cysteine-alkylating reagent and that the alkylation is reversible but might depend on the protein structure or amino acids adjacent to the cysteine residue, binding affinity, and/or dissociation kinetics. It is to be noted that reversibility was determined under the condition of 2-fold excess of DTT over alkylating reagents and that OBT binding may be affected differentially under other conditions.</p><!><p>To gain better insight into the molecular mechanism of OBT and its effect on the global proteome, we implemented multiplexed quantitative mass spectrometry-based proteomics using the isobaric labeling strategy with Tandem Mass Tag (TMT) reagents and the SPS-MS3 method.12,14 Global proteomics analysis in two different patient-derived GSC specimens at 20 h after treatment with a subtoxic dose of OBT identified heme oxygenase 1 (HMOX1; HO1) as the top upregulated protein in both lines (Figure 2A,B; Supplemental Table 3). Gene ontology category analysis of the most upregulated proteins after treatment (using the DAVID bioinformatics platform) showed that the most significantly enriched category was an "oxidation-reduction process" containing the upregulated proteins HMOX1, PHS2, GLRX1, VKORL, BLVRB, and CDO1. To further explore the functional network of these proteins, we interrogated the STRING database for high-confidence direct interactors of these proteins (white circles; Figure 2C) and found a network of densely connected proteins containing five of the upregulated proteins (red) as well as Nrf2 (green), a master regulator of oxidative damage response, and Keap1 (green), a substrate adaptor protein of the E3-ligase regulating the ubiquitin-mediated degradation of Nrf2 (Figure 2C). A strong increase in HO1 mRNA expression in response to OBT was detected across different cancer cells (Figure 2D and Supplemental Figure 3A). The Nrf2 target genes NQO1 and TXNRD2 were also upregulated following OBT treatment (Supplemental Figure 3B). To further support these findings independently, we designed a functional ARE luciferase reporter which was strongly activated following treatment with subtoxic doses of OBT, even higher than the positive e control tert-butylhydroquinone (tBHQ), a potent activator of Nrf215 (Supplemental Figure 3C,D). Importantly, OBT-induced ARE activation and loss of cell viability were completely reversed by the addition of various thiol nucleophilic antioxidants including NAC, dithiothreitol (DTT), and GSH but only partially reversed by Trolox, an antioxidant that is devoid of thiol groups (Figure 2E and Supplemental Figure 3E).</p><p>Finally, we confirmed OBT-mediated Nrf2 activation in vivo using a breast cancer mouse model generated by injecting MDA-MB-231 into the fat pad of nude mice.8 Treatment with OBT (7.5 mg/kg for four consecutive days) resulted in 7-fold upregulation of HO1 transcripts in the tumor, consistent with our findings in cell culture (Figure 2F). Taken together, these results confirm that OBT acts as an inducer of Nrf2 both in culture and in vivo.</p><!><p>Several thiol reactive Nrf2 activators have been shown to bind Keap1, suggesting that OBT could function in a similar fashion.16 To investigate if Keap1 is also an OBT target, we applied the mass spectrometry-based approach described in Figure 1 to Keap1. Similar to the experiment with bovine catalase, we found OBT to bind covalently to all identified cysteine sites, thus confirming that this compound can directly bind to Keap1. Importantly, under our experimental conditions, we found this modification to be partially reversible at Cys151 located in the BTB domain (CUL3 binding site) and Cys434 located in the Kelch domain (Nrf2 binding site) of Keap1 (Figure 2G; Supplemental Figure 4 and Supplemental Tables 4 and 5).</p><p>Next, we evaluated the effect of OBT treatment on cells following stable downregulation of Keap1. Silencing of Keap1 with shRNA (shKeap1) expectedly resulted in stabilization of Nrf2, leading to a major increase in ARE reporter activity (Figure 3A,B and Supplemental Figure 5A). There was no strong potentiation of OBT-mediated ARE activation in cells expressing shKeap1 as compared to a nontargeting shRNA (shCtrl), detected using the ARE reporter (Figure 3A,B and Supplemental Figure 5B) and mRNA expression of HO1 and NQO1 (Supplemental Figure 5C). Further, silencing of Keap1 decreased cell death following treatment with OBT (Figure 3C and Supplemental Figure 5D). These data suggest that either Keap1 expression is necessary for binding of OBT to Keap1 cysteine residues, thus stabilizing and activating Nrf2, or that a strong activation of Nrf2 protects against OBT-mediated cytotoxicity.</p><p>Numerous anticancer drugs have been shown to activate Nrf2.16 We tested if reactive electrophiles and oxidants known as transient activators of Nrf2 could induce cell death when added to breast and brain cancer cells, similar to OBT. We first confirmed ARE-inducing properties of the triterpenoid CDDO-Me (2-cyano-3,12-dioxoolean-1,9-dien-28-oic acid-methyl ester),17 currently being clinically tested for the treatment of leukemia and solid tumors as well as other diseases. U87 and MDA-MB-231 cells expressing ARE-Gluc reporter treated with CDDO-Me showed increased reporter activity (Supplemental Figure 6A). Higher doses of this compound led to a marked decrease in cell viability in U87 cells (Supplemental Figure 6B). Treatment of U87 cells with additional Nrf2 activators, cinnamaldehyde,18 diethyl fumarate,19 and sulforaphane20 also resulted in increased ARE activity and a moderate decrease in U87 cell viability at the doses tested (Supplemental Figure 6C,D). When combined with OBT, all three compounds showed increased cytotoxicity (Supplemental Figure 6E). These results suggest that increased electrophile or oxidant concentrations are likely to cause further cysteine modifications in the cellular proteome along with a stress response evident by an upregulation of Nrf2 signaling, thus increasing cytotoxicity.</p><!><p>We next asked whether OBT-mediated covalent modification of Keap1 affects stability of this protein. Immunoblot analysis showed a time- and dose-dependent decrease in Keap1 protein levels as early as 4 h following treatment with OBT, which were restored to physiological levels at 24 h (Figure 3D,E). Within this time frame (<24 h), we have previously shown that treatment with OBT results in early morphological changes, elevated ROS levels, and activation of apoptosis in tumor cells.8 To further confirm Keap1 protein degradation, we expressed HA-tagged Keap1 or Keap1 with a deleted BTB domain (Keap1ΔBTB)21 which is essential for Keap1 binding to CUL3 and activation of Keap1-CUL3 E3 ligase activity. We did not observe any decrease in Keap1 protein levels in cells expressing BTB-mutated Keap1 (Figure 3F), suggesting that the E3 ligase activity is required for OBT-mediated degradation of Keap1. Further, similar to Keap1 knockdown, silencing of CUL3 with shRNA prevented ARE activation following treatment with OBT and decreased cell death (Supplemental Figure 5B,D). Ectopic expression of a dominant-negative CUL3 mutant (DN-CUL3) also protected against OBT-induced cell death, further corroborating these findings (Supplemental Figure 7A). The neddylation of CUL3 is essential for its ubiquitin ligase activity.22 To determine if CUL3 activation is essential for OBT-induced cell death, we cotreated cells with OBT and the neddylation inhibitor MLN4924 and observed protection against cell death (Supplemental Figure 7B). Overall, these findings confirm that E3 ligase activity is required for targeting of tumor cells with OBT.</p><p>Since ubiquitination of Keap1 could lead to its degradation,23 we evaluated this process following OBT treatment by immunoblotting. Indeed, OBT treatment effectively resulted in ubiquitination of Keap1 (Figure 3G). Additionally, downregulation of CUL3 or cotreatment with the proteasome inhibitor MG132 prevented OBT-mediated degradation of Keap1 (Figure 3H). Among the reactive cysteines of Keap1, C151 was found to be necessary for Keap1-alkylating ARE inducers that promote the dissociation of the Keap1-CUL3 complex, thus stabilizing Nrf2. Accordingly, mutation of C151 impairs its alkylation by electrophiles such as sulforaphane, tBHQ, or AI-1 and impairs Nrf2 activation.24−26 However, serine substitution of Cys-151 (Keap1C151S) did not protect against OBT-mediated degradation of Keap1 (Supplemental Figure 7C). Overall, these results confirm that OBT treatment leads to degradation of Keap1 and that CUL3 is an essential regulator of this process.</p><!><p>Pharmacokinetic profiling of OBT in mice showed high systemic plasma clearance with terminal plasma half-life of 24 min following intraperitoneal injection (Supplemental Figure 8A). Furthermore, we found that OBT efficiently penetrates the intact blood-brain barrier (BBB) (Supplemental Figure 8B,C). To confirm that OBT is able to penetrate the brain and functionally target brain tumors, we used positron emission tomography (PET) with the PET-tracer 2-[18F]fluoro-2-deoxy-D-glucose (FDG), which measures the rate of tumor glucose uptake in a mouse orthotopic GSC model. OBT-treated mice exhibited approximately a 50% decrease in FDG tumor uptake as measured by FDG-PET imaging (Figure 3I). Finally, we selected the analog (AEN36 Tris) with most improved pharmacological properties including enhanced potency, stability, and solubility and evaluated the in vivo antineoplastic effect in a breast cancer mammary fat pad tumor xenograft model. Treatment with AEN36 Tris (10 mg/kg daily for 22 days) induced a significant decrease in tumor volume, compared to the control group, as assessed by bioluminescence imaging (Supplemental Figure 8D). In summary, OBT penetrates the BBB and can be modified to enhance its solubility while retaining its antineoplastic properties.</p><!><p>Small molecules that react with cysteine side chains within Keap126 or target the kelch domain of Keap127 have been identified. We now provide direct evidence that the natural compound OBT activates the Nrf2 pathway by binding covalently to cysteine residues within the BTB-domain of Keap1 leading to its ubiquitination and subsequent proteasomal degradation. This directly impacts the ability of the CUL3-Keap1 ubiquitin ligase complex to degrade Nrf2, resulting in Nrf2 stabilization and downstream activation of ARE-mediated transcription (Figure 3J). It is highly likely that OBT also interacts with other thiol-rich proteins; however, our data supports that Keap1 is a major functional target for OBT and that the BTB-CUL3 ubiquitin ligase complex is required for OBT-mediated degradation. The cysteine-reactive compound likely engages other secondary targets in order to promote an overall stress response. In fact, downregulation of Keap1 was not sufficient to induce the same level of cytotoxicity observed after treatment with OBT or other Nrf2 activators, confirming this hypothesis.</p><p>The transcription factor Nrf2 is often viewed as a pleiotropic gene. Whether its activation or inhibition is beneficial for tumor treatment remains a paradox and seems to depend on various factors such as the cell type, tumor stage, and genetic aberrations within the tumor.2−4 Nrf2 has been suggested to act as a tumor suppressor, thus its activation can suppress carcinogenesis.2,28−30 Nrf2 activation was shown to decrease tumor growth in established tumors,31 and several antineoplastic drugs enhance Nrf2 activity.16 In this study, we have demonstrated that, in addition to OBT, several other reactive electrophiles and oxidants known as transient activators of Nrf2 can induce cytotoxicity in tumor cells. On the other hand, activation of Nrf2 by cancer targeting drugs can also lead to unfavorable clinical outcomes because of Nrf2's ability to enhance chemoresistance.28 One plausible explanation of this paradox is that, unlike somatic mutations and oncogene-mediated signaling that promote a sustained Nrf2 activation accompanied by numerous adaptation mechanisms, pharmacological activation of Nrf2 is transient and does not necessarily phenocopy constitutive Nrf2 activation.3,4 Despite these controversies, several Nrf2 activators have been developed as antineoplastic compounds in preclinical studies,32 and at least one such compound, sulforaphane, has advanced to a phase 2 clinical trial for the treatment of metastatic breast cancer (NCT02970682).</p><p>The dose−response curve of many chemopreventive agents as well as chemotherapeutic drugs is U-shaped,2 resulting in opposing effects between low and high doses of the same agent. For example, synthetic oleanane triterpenoids exert chemopreventive functions at low doses but are also able to induce oxidative stress and apoptosis at higher doses.28,33 The same model could be applied to OBT where lower doses of the compound lead to a strong Nrf2 activation, while at higher doses, OBT acts as a potent pro-oxidant that targets cancer cells leading to DNA damage and apoptosis as we have previously shown.8</p><p>In conclusion, we have established a mechanistic understanding of the mode of action of OBT. We show that OBT is a reversible covalent modifier of cysteine residues in Keap1, targeting it for proteosomal degradation, leading to strong activation of Nrf2. In addition, we have developed novel OBT analogs with improved in vivo activity that allow the tuning of pharmacological properties, which will facilitate the preclinical development of this compound class. Finally, given that impaired Keap1 activity and Nrf2 activation lead to increased expression of antioxidant and detoxification genes, and their role in neuroprotection,19,34,35 we speculate that OBT and its analogs might have cross-disease applications for disorders such as diabetes, Alzheimer's disease, and Parkinson disease;36 however, this remains to be tested.</p><!><p>OBT was purchased from Gaia Chemicals. N-Acetyl-L-cysteine, dithiothreitol, L-glutathione, tert-butylhydroquinone, N-ethylmaleimide, iodoacetamide, cinnamaldehyde, diethyl fumarate, and sulforaphane were purchased from Sigma-Aldrich. Trolox, sulfasalazine, CDDO-Me, and MG-132 were purchased from Cayman Chemical. Nontargeting shRNA control (shCtrl) and shRNA constructs targeting Keap1 (TRCN0000156676) and CUL3 (TRCN0000012778) were purchased from Sigma and packaged into lentiviral vectors using standard protocols.</p><!><p>MDA-MB-231 and U87 (U87-MG) cells were obtained from the American Type Culture Collection (ATCC). 293T human embryonic kidney fibroblasts were provided by Dr. Xandra Breakefield (Massachusetts General Hospital). All three cell lines were grown in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum (Gemini Bioproducts), 100U penicillin, and 0.1 mg mL−1 streptomycin (Sigma). All cells were maintained at 37 °C in a humidified 5% CO2 incubator. GSCs were obtained from tumor tissues of GBM patients following surgical resection, under approval from the corresponding Institutional Review Board. These cells have been previously characterized37−39 and were maintained in culture as neurospheres in Neurobasal medium (Gibco) supplemented with heparin (2 μg/mL; Sigma) and recombinant EGF (20 ng/mL) and bFGF-2 (10 ng/mL; Peprotech). Human astrocytes were obtained from ScienCell and cultured in Astrocytes Medium (ScienCell).</p><!><p>Cells were plated in 96-well plates and treated with the corresponding compounds. Cell viability was measured by adding 25 μL/well of CellTiter-Glo (Promega) followed by 10 min incubation and transfer to a white 96-well plate. Bioluminescence was quantified using a Synergy HTX multimode reader (Biotek).</p><!><p>GraphPad Prism v6.01 software (LaJolla, CA) was used for statistical analysis of all data. A p-value of less than 0.05 was considered to be statistically significant. For analysis between multiple groups, a two-tailed Student's t-test (unpaired), ANOVA, and Tukey's posthoc test were performed as indicated. All experiments were performed at least in 3 replicates and repeated 3 independent times.</p>

PubMed Author Manuscript

Degradation Pathways and Complete Defluorination of Chlorinated Polyfluoroalkyl Substances (Clx−PFAS)

Chlorinated polyfluoroalkyl substances (Clx−PFAS) have been developed and applied for decades, but they have just been recognized as an emerging class of pollutants. This study systematically investigated the degradation of three types of Clx−PFAS structures, including omega-chloroperfluorocarboxylates (ω-ClPFCAs, n=1,2,4,8 Cl−CnF2nCOO − ), 9chlorohexadecafluoro-3-oxanonane-1-sulfonate (F-53B, Cl−(CF2)6−O−(CF2)2SO3 − ) and polychlorotrifluoroethylene oligomer acids (CTFEOAs, n=1,2,3 Cl−(CF2CFCl)nCF2COO − ) under UV/sulfite treatment. The results lead to a series of transformative insights. After initial reductive dechlorination by hydrated electron (eaq -), multiple pathways occur, including hydrogenation, sulfonation, and dimerization. In particular, this study identified the unexpected hydroxylation pathway that convert the terminal ClCF2− into − OOC−, which is critical for the rapid and deep defluorination of F-53B. The hydroxylation of the middle carbons in CTFEOAs also triggers the cleavage of C−C bonds, yielding multiple −COO − groups to promote defluorination. Hence, the Cl atoms in Clx−PFAS enhance defluorination in comparison with the perfluorinated analogs.After UV/sulfite treatment, the HO• oxidation of the residue leads to ~100% defluorination of all ω-ClPFCAs and CTFEOAs, without generating toxic ClO3 − from Cl − . This study renovates and further advances the mechanistic understanding of PFAS degradation in "advanced reduction" systems. It also suggests the synergy between "more degradable" molecular design and costeffective degradation technology to achieve the balanced sustainability of fluorochemicals.

degradation_pathways_and_complete_defluorination_of_chlorinated_polyfluoroalkyl_substances_(clx−pfas

<!>Mass Balance Analysis.<!>I<!>Å<!>E

<p>To effectively address the global pollution by per-and polyfluoroalkyl substances (PFAS), [1][2][3][4][5][6][7] it is imperative to understand the degradation mechanism and develop treatment strategies for various PFAS structures. Environmental studies on PFAS started approximately 50 years after the creation of PFAS. [8][9][10][11][12] Current efforts primarily focus on the "legacy" perfluorinated CnF2n+1−X (X = COO − , SO3 − , and (CH2)m−R, where R represents highly diverse organic moieties).</p><p>Those structures have been well known as building blocks and degradation products of surfactants 13,14 and coatings. 15,16 However, "alternative" PFAS containing −O−, [17][18][19] -H, 20,21 and -Cl 22,23 in the fluorinated moiety (RF) have also been systematically developed and extensively applied for decades. Recent studies have confirmed their negative health effects [24][25][26][27] and worldwide pollution in water [28][29][30][31][32][33][34][35][36] and soil. 37 Chlorine-containing PFAS (Clx−PFAS) are prepared by the telomerization process, where the iodinated telogen (RF−I) initiates the polymerization of fluorinated olefins (Schemes 1A and B). For enhanced stability against thermal and chemical reactions, the terminal iodine is further replaced by chlorine. The chlorination treatment costs less than fluorination (Scheme 1A). 38,39 Although the inclusion of Cl could result in higher surface tension than the perfluorinated analog, 20 the −CF3 branch could yield an even lower surface tension 22 for fluoropolymer production. 40 A long-chain ether sulfonate surfactant, F-53B, contains an omega-Cl and exhibits similar surface tension to the perfluorinated (and more costly) F-53 in a wide concentration range. 41 F-53B was used as a mist suppressant for the electroplating industry in China 39 and extensively detected in the environment. [42][43][44] Recently, omega-chloroperfluoropolyether carboxylates (ClPFPECAs, Scheme 1C) developed for fluoropolymer synthesis 45 have been detected in the soils of densely populated New Jersey of the U.S. 37 Chlorine atoms are also included in both telogens and olefins to prepare polychlorinated PFAS. 22,38 Polychlorotrifluoroethylenes (PCTFEs, Scheme 1D) are chemically inert, nonflammable, and more cost-effective than perfluoropolyethers for the use in metal lubricants and hydraulic fluids, and have shown high haptic toxicities. 46,47 In the rat liver, PCTFEs are converted into oligomer acids (CTFEOAs), which are also chemically synthesized as commercial surfactants. by UV/sulfite treatment, which produces hydrated electron (eaq − , Equation 1), [48][49][50][51] a potent species for reductive hydrodefluorination (Equations 2−3): 52,53</p><p>Based on the existing knowledge, the C−Cl bonds in Clx−PFAS are supposed to undergo hydrodechlorination. 54 Initially, we expected that omega-chloroperfluorocarboxylates (ω-ClPFCAs, Cl−CnF2nCOO − ) would rapidly yield the omega-hydro analogs (ω-HPFCAs, H−CnF2nCOO − ) and then degrade as previously elucidated (Scheme 1E). 55 ) showed little reactivity with eaq -. 52,55,56 For n=1 structures, while the defluorination from H−CF2COO − and CF3COO − were nearly 100% within 4−8 h, 55 the defluorination from Cl−CF2COO − was up to 96% (Fig. 1C). We note that the 4% disparity was not negligible. Instead, it motivated us to identify a novel reaction pathway promoted by the Cl atom. Hydrochlorination Is Not the Primary Pathway. Although previous studies on UV/sulfite treatment of Cl−CH2−COO − 57 and F-53B 54 have confirmed the hydrodechlorination pathway, our transformation products (TP) analysis found novel information. After the reductive dechlorination of ω-ClPFCA (Equation 4), the omega carbon radical can be hydrogenated to yield ω-HPFCA (Fig. 2A and Equation 3).</p><p>However, from 25 μM of Cl-C4F8COO − , the maximum concentration of H-C4F8COO − was merely 0.66 μM at 4 min (Fig. 2B). At this time point, the rapid dechlorination of Cl-C4F8COO − had just finished (Fig. 1A), while the much slower defluorination had not proceeded to a significant level. Similarly, 25 μM of Cl-C8F16COO − yielded a maximum of 0.59 μM of H-C8F16COO − at 4 min (Fig. 2C). For short-chain ω-ClPFCAs, we raised the initial concentration for 10-fold to facilitate the TP detection. From 250 μM of Cl−CF2COO − and Cl-C2F4COO − , the maximum concentration of the corresponding products, H−CF2COO − and H-C2F4COO − , were 3.7 μM at 60 min and 2.7 μM at 30 min, respectively (Figs. 2D and E). Therefore, only a small fraction of the parent ω-ClPFCAs were converted to the corresponding ω-HPFCAs.</p><p>In an early study on Cl−CH2COO − degradation by UV/sulfite, 57 , C, E). The treatment of 250 µM of Cl−CF2COO − yielded − OOC−COO − at the maximum concentration of 15.7 µM (6.3%) at 1 h (Fig. 2D). However, UV/sulfite treatment of H−CF2−(CF2)n-1COO − could not produce − OOC−(CF2)n-1COO − . Although − O3S−(CF2)nCOO − and dimeric − OOC−(CF2)2n−COO − may produce small amounts of − OOC−(CF2)n-1COO − via reductive C−S bond cleavage 52 and sequential decarboxylation, 53 respectively, they were not the primary source of − OOC−(CF2)n-1COO − . For example, the maximum − OOC−COO − concentration from 250 µM of pure -O3S−CF2COOwas only 5.4 µM, much less than that from 250 µM of Cl−CF2COO − . From 250 µM of pure -OOC−(CF2)2COO -, the formation of − OOC−COO − (after two rounds of decarboxylation) was not detected.</p><p>Because direct oxidation of all Clx−PFAS with HO• resulted in negligible defluorination (Table S1), the degradation of Clx−PFAS required the reaction with eaq -. We propose that after reductive dechlorination of Cl−(CF2)nCOO -, the •CF2-(CF2)n−1COOintermediate could react with a HO• to yield unstable perfluorinated alcohol, 58 which spontaneously evolves into − OOC−(CF2)n-1COO − (Fig. 2A). Further degradation of such TPs was also confirmed (Tables S2 and S3). Because HO• is present in UV/sulfite system, 59,60 we added methanol to scavenge HO•. 61 As expected, the yield of − OOC−(CF2)n-1COO − was lowered, and the production of H−(CF2)nCOOwas increased (Fig. S1). However, the overall deF% from Cl−(CF2)nCOOwere not impacted because (i) methanol is not a significant scavenger of eaqand (ii) ω-HPFCAs and perfluorodicarboxylates (PFdiCAs) allowed the same number of C−F bonds to be cleaved by UV/sulfite treatment. 55 The reactivity of -O3S−(CF2)nCOOis assumed to be similar to H−(CF2)nCOOor F−(CF2)nCOO -.</p><!><p>Because most TPs from Cl−CF2COO − are quantifiable with standard chemicals and the n=1 structure limited the number of reaction sites, we were able to close 94% of the total F balance at 30 min (Fig. 2D). This value is the highest F mass balance we have ever achieved from UV/sulfite treatment of PFAS. Therefore, we propose that the four pathways (hydrogenation, sulfonation, dimerization, and hydroxylation) could represent the major degradation pathways upon the dechlorination by eaq -. Both -O3S−CF2COO -(Fig. 2G) and H−CF2COOallowed 100% defluorination by UV/sulfite treatment, whereas − OOC−(CF2)2COO − allowed up to 83% defluorination. 55 Therefore, the incomplete defluorination (96%, Fig. 1C) of Cl−CF2COO − can be attributed, at least partially, to the dimerization pathway. More importantly, the experimental results have shown that (i) the previously reported hydrodechlorination is the least preferred reaction pathway among the four and (ii) oxidative species such as HO• are playing significant roles in the UV/sulfite system. However, detailed mechanistic eludication for the oxidative species warrants further studies. A complete set of degradation pathways for long-chain and perfluorinated structures remain largely unknown. Our lab is investigating that aspect and will report the findings in the near future.</p><p>New Mechanistic Insights into F-53B Degradation. Similar to ω-ClPFCAs, the ether sulfonate F-53B exhibited rapid and complete dechlorination within 6 min (Fig. 3A) and 73% defluorination at 12 h (Fig. 3B). The trend of calculated C−F and C−Cl BDEs (Fig. 3G) in F-53B versus its hydro-and perfluorinated analogs is similar to the all-carbon-chain structures (Fig. 1D).</p><p>The spontaneous C−Cl cleavage from [Cl−(CF2)6O(CF2)2SO3] 2− • (Fig. 3G) supports the reductive dechlorination mechanism. Like ω-ClPFCA, F-53B yielded H−(CF2)6O(CF2)2SO3 − (structure a in Fig. 3H), − O3S−(CF2)6O(CF2)2SO3 − (b), and − OOC−(CF2)5O(CF2)2SO3 − (c) (Fig. 3C). The dimerized structure exceeded the molecular weight limit we set for suspect TP detection. A series of shorter-chain TPs, − OOC−(CF2)nO(CF2)2SO3 − (n=4,3,2,1, structures d−g in Fig. 3H), showed decreasing abundance as the −(CF2)n− moiety became shorter (Fig. 3D). Because we did not detect any shorter-chain F-53B analog impurities (i.e., n=1−5 Cl−(CF2)nO(CF2)2SO3 − ) in the t=0 sample, the shorter-chain − OOC−(CF2)nO(CF2)2SO3 − TPs d−g most probably came from the stepwise decarboxylation 52 from c (Fig. 3H). Notably, all carboxylate TPs became non-detectable after 4 h (Fig. 3D), whereas significant defluorination continued to 8 h (Fig. 3B).</p><!><p>In comparison to c−g, the hydrogenated TP a and sulfonated TP b degraded much slower (Fig. 3E). A previous study proposed that the reaction between a and eaqtriggered C−O cleavage. 54 However, our experimental results led to a different interpretation. First, if C−O cleavage occurred in a, the short moiety •(CF2)2SO3 − (or •O(CF2)2SO3 − ) would evolve into − OOC−CF2SO3 − (h) via the unstable HO(CF2)2SO3 − . 62 Although h was detected as a significant TP (Fig. 3F), its abundance became negligible after 4 h, when a major fraction of a still remained (Fig. 3E). Thus, the formation of h from a is less likely.</p><p>Second, although spontaneous C−O cleavage has been confirmed from the reaction between perfluoroether carboxylates (PFECAs) and eaq -, 62 we wondered whether this mechanism applies to perfluoroether sulfonates (PFESAs). Thus, we tested a short-chain model, CF3CF2-O-(CF2)2-SO3 -. To our surprise, this C2+C2 PFESA did not show any decay or defluorination. This result differs entirely from short-chain PFECAs that exhibited rapid decay and significant defluorination. 62 Apparently, the C−O cleavage mechanism does not readily occur for PFESA.</p><p>Moreover, the previous study on F-53B degradation also reported complete decay and significant defluorination of F-53 (perfluorinated F-(CF2)6O(CF2)2SO3 -, not available for this study). 54 The relatively facile degradation of C6+C2 PFESA and no reactivity of C2+C2 PFESA resemble the comparison among n=6, 4, and 1 CnF2n+1-SO3 -. 52,56 Lacking a terminal −COO − , the reactivity of perfluoroalkane sulfonates strongly depends on the fluoroalkyl chain length. PFBS was substantially more recalcitrant than PFOS, and TFMS did not show any degradation (Fig. 3I).</p><p>Third, our previous study on PFECAs has confirmed that C-O cleavage occurred regardless of how many -CF2units separate the ether bond and terminal −COO − . 62 Therefore, the − OOC−(CF2)nO(CF2)2SO3 − are the most probable TPs that allow reductive C-O cleavage and thus produce h as the common TP (Fig. 3H). Notably, h allowed 100% defluorination (Fig.</p><!><p>A double hydrogenated product was observed from C4 CTFEOA (Fig. 5A), but the attempts to identify all possible H/Cl exchanged TPs from C6 and C8 CTFEOAs were not successful. This is probably because the diverse substitutions (e.g., H + +eaq -, SO3 − •, and HO•) upon dechlorination at multiple carbons significantly lowered the abundance of individual TPs. Despite of the complex reaction pathway network, we prioritized the focus on filling the remaining 6−8% gap from the goal of complete defluorination of most Clx−PFAS structures. Further Defluorination of Clx−PFAS by the Following Oxidation. The sequential treatment using UV/sulfite followed by heat/persulfate has allowed near-complete defluorination from most PFCAs and PFSAs. 56 After the UV/sulfite treatment, residues containing C−H bonds allow extensive oxidation so that the isolated CF3− or −CF2− can be hydroxylated and thus defluorinated. As expected, both SO4 -• (initial pH=2) and HO• (initial pH>12) were capable of cleaving residual C−F bonds and resulted in 99−103% overall deF% of all four ω-ClPFCAs and three CTFEOAs (Table 2, entries 1−7). The exception is F-53B (entry 8), where the following oxidation brought deF% from 76% (after UV/sulfite for 24 h) to 93%. We attribute the incomplete defluorination to recalcitrant structures containing long fluoroalkyls, such as H−(CF2)6O(CF2)2SO3 − , which remained in a significant abundance at 24 h (Figs. 3E and H).</p><!><p>We further examined the potential formation of the toxic byproduct chlorate (ClO3 − ) 63,64 from Cl − , both a ubiquitous water component and the Clx−PFAS dechlorination product. The SO4 -</p><p>• treatment of Clx−PFAS defluorination residues oxidized a small portion of Cl − into ClO3 − (Table 2). In the absence of organic residues from Clx−PFAS defluorination, the yield of ClO3 − from the same concentration of Cl − was elevated (entry 9 versus entries 1−4, and 8). In sharp contrast, the use of HO• in all cases produced negligible ClO3 − , if any (lower than the detection limit). , 3H, and 5D). Perfluorinated carboxylates have the highest degradability among all reported PFAS pollutants. 52,56 Second, a C−Cl bond integrated into sulfonate-terminal PFAS will substantially enhance the degradability by introducing −COO − . This feature is particularly important for degrading non-carboxylate short-chain structures (Fig. 2G versus 3I). Third, the reductive C−O cleavage pathway, which is critical for deep defluorination of ether structures, is exclusive for carboxylates (Fig. 3H). Fourth, the comparison of defluorination kinetics for CTFEOAs versus PFCAs (Table 1 and Figs. 4B, D, F) suggests that the inclusion of multiple C−Cl bonds can reduce the UV energy consumption by at least 50%. In addition, even if a small portion of C−Cl bonds were converted into C−H bonds, our previous study has shown that ω-HPFCAs favor the desirable decarboxylation pathway for defluorination over PFCAs. 55 Implications for PFAS Chemical Design. In the real world, PFAS chemicals cannot be immediately phased out from all fields due to their unique properties for a broad scope of applications. 65 The fluorine-free replacements could even result in a higher toxicity. 66 For the future design, manufacturing, and management of specialty PFAS products, it would be imperative to enhance the degradability without sacrificing the desirable property. The inclusion of one or more Cl atoms in the PFAS structure could be a potential solution. Earlier works have demonstrated that the negative impact of replacing a F atom with a Cl atom on surfactant properties can be offset by flexible molecular designs. 22,41 This proof-of-concept study also verifies that the use of 254 nm UV, sulfite, hydroxide, and persulfate (or other common chemicals and approaches producing HO•), all of which are essential components of practical water and wastewater treatment, 67 at 20 °C and pH 12 overnight. 68 The information on CAS numbers, purities, and vendors are listed in the Supplementary Information (SI). Sodium sulfite (Na2SO3), sodium bicarbonate (NaHCO3), sodium hydroxide (NaOH), potassium persulfate (K2S2O8), and sulfuric acid (H2SO4) were purchased from Fisher Chemical.</p><p>UV/sulfite Treatment. The reactor configuration 52,53,56 and photochemical parameters 69 have been established in our previous studies. Briefly, a 600 mL aqueous solution containing 25 μM individual PFAS, 5 mM NaHCO3, and 10 mM Na2SO3 was loaded into the photoreactor (assembled with Ace Glass parts #7864-10, #7874-38, and #7506-14, and wrapped with aluminum foil). The solution was irradiated by an 18 W low-pressure mercury lamp (GPH212T5L/HO) placed in the quartz immerse well. The temperature was maintained at 20 °C by the jacketed cooling water. Prior to the reaction, N2 sparging was not needed 52 because the initially dissolved oxygen (up to 0.25 mM at the saturated level at room temperature) was readily depleted by sulfite. 70 All three outlets of the photoreactor were sealed with rubber stoppers to prevent air intrusion.</p><p>Subsequent Oxidation by Heat/persulfate. After UV/sulfite treatment, the 30 mL aliquots of the resulting solution were amended with 5 mM K2S2O8 and added with either H2SO4 to pH 2.0 or with 12.5 mM NaOH to pH ~12.4. The solutions were loaded in glass reaction tubes and heated at 120 °C for 40 min in a pressure cooker (Farberware 6 Quart). 56 Thermal decomposition of S2O8 2− yielded two equivalents of sulfate radical (SO4 − •). The SO4 − • is preserved at pH 2.0 or fully converted into a hydroxyl radical (HO•) at pH>12. 71 Since the 1960s, thermal digestion of environmental samples using S2O8 2− has been extensively adopted [72][73][74] due to its high efficiency in mineralizing organic structures. Our proof-of-concept studies have used this approach to probe the "upper limit" of oxidative conversion of PFAS. 55,56 Sample Analyses. Fluoride ion (F -) release was measured by a Fisherbrand accumet solidstate ion-selective electrode connected to a Thermo Scientific Orion Versa Star Pro meter. This method has been validated by ion chromatography (IC) 52 and solution matrix spiking tests. 56 Chloride ion (Cl -) release was measured by IC. The percentages of defluorination (deF%) and dechlorination (deCl%) are defined as the ratios between released F -/Cl with an added eaq -(i.e., [Clx−PFAS] 2− •) followed our previous approach. 52,62,75 Results from these approaches have been consistent with those from condensed Fukui Function in terms of predicting the site of reductive PFAS transformation. 55</p>

ChemRxiv

Synthesis, Characterization, Anticancer, and Antioxidant Studies of Ru(III) Complexes of Monobasic Tridentate Schiff Bases

Mononuclear Ru(III) complexes of the type [Ru(LL)Cl2(H2O)] (LL = monobasic tridentate Schiff base anion: (1Z)-N′-(2-{(E)-[1-(2,4-dihydroxyphenyl)ethylidene]amino}ethyl)-N-phenylethanimidamide [DAE], 4-[(1E)-N-{2-[(Z)-(4-hydroxy-3-methoxybenzylidene)amino]ethyl}ethanimidoyl]benzene-1,3-diol [HME], 4-[(1E)-N-{2-[(Z)-(3,4-dimethoxybenzylidene)amino]ethyl}ethanimidoyl]benzene-1,3-diol [MBE], and N-(2-{(E)-[1-(2,4-dihydroxyphenyl)ethylidene]amino}ethyl)benzenecarboximidoyl chloride [DEE]) were synthesized and characterized using the microanalytical, conductivity measurements, electronic spectra, and FTIR spectroscopy. IR spectral studies confirmed that the ligands act as tridentate chelate coordinating the metal ion through the azomethine nitrogen and phenolic oxygen atom. An octahedral geometry has been proposed for all Ru(III)-Schiff base complexes. In vitro anticancer studies of the synthesized complexes against renal cancer cells (TK-10), melanoma cancer cells (UACC-62), and breast cancer cells (MCF-7) was investigated using the Sulforhodamine B assay. [Ru(DAE)Cl2(H2O)] showed the highest activity with IC50 valves of 3.57 ± 1.09, 6.44 ± 0.38, and 9.06 ± 1.18 μM against MCF-7, UACC-62, and TK-10, respectively, order of activity being TK-10 < UACC-62 < MCF-7. The antioxidant activity by DPPH and ABTS inhibition assay was also examined. Scavenging ability of the complexes on DPPH radical can be ranked in the following order: [Ru(DEE)Cl2(H2O)] > [Ru(HME)Cl2(H2O)] > [Ru(DAE)Cl2(H2O)] > [Ru(MBE)Cl2(H2O)].

synthesis,_characterization,_anticancer,_and_antioxidant_studies_of_ru(iii)_complexes_of_monobasic_t

1. Introduction<!>2.1. Chemicals and Instrumentations<!>2.2. Preparation of the Tridentate Schiff Bases (DAE, HME, MBE, and DEE)<!>2.3. Synthesis of Ru(III)-Tridentate Schiff Base Complexes<!>2.3.1. Synthesis of [OHC6H3OH:C(CH3):N(C2H4)N:C(CH3):NHC6H5RuCl2(H2O)]<!>2.3.2. Synthesis of [OHC6H3OH:C(CH3):N(C2H4)N:CH:C6H3OHOCH3RuCl2(H2O)]<!>2.3.3. Synthesis of [OHC6H3OH:C(CH3):N(C2H4)N:CH:C6H5(OCH3)2RuCl2(H2O)]<!>2.3.4. Synthesis of [OHC6H3OH:C(CH3):N(C2H4)N:C(Cl):C6H5RuCl2(H2O)]<!>2.4. In Vitro Antiproliferative Activity<!>2.5.1. Scavenging Activity of 1,1-Diphenyl-2-picrylhydrazyl (DPPH) Radical<!>2.5.2. ABTS: 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) Radical Scavenging Assay<!>3.1. Synthesis and Characterization<!>3.2. Molar Conductivity Measurements<!>3.3. Infrared Spectra<!>3.4. Electronic Absorption Spectra Studies<!>3.5. Antiproliferative Activity<!>3.6. Antioxidant Capacity<!>3.6.1. DPPH Radical Scavenging Assay<!>3.6.2. 2,2′-Azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) Radical Scavenging Activity<!>4. Conclusion

<p>Coordination chemistry of transition metal Schiff base complexes possessing N, O, and S-donor atoms has received consideration over the past few decades, due to the imperative roles these compounds have played in a variety of biochemical procedures like haloperoxidation [1], insulin mimicking [2, 3], fixation of nitrogen [4], inhibition of cancer growth, and prophylaxis against carcinogenesis [5, 6]. A huge variety of carbonyl compounds (>C=O) and amines (R-NH2) have been exploited in the preparation of Schiff bases [7, 8]. The reactivity of aldehyde compounds is generally faster than those of the ketones in condensation reaction, thereby resulting in the formation of Schiff bases with a centre that are less steric than the ketone's, relatively unstable and freely polymerizable [9]. This important attribute of Schiff base ligands offers prospects for prompting substrate chirality and metal centred electronic factor tuning and improving the solubility and steadiness of either homogeneous or heterogeneous catalysts [10–12].</p><p>Schiff bases have shown an interesting application as an active corrosion inhibitor that is established on their capability to spontaneously form a monolayer upon the surface to be glazed [13], as it is a type of interaction existing between an inhibitor and a metal surface known as chemisorption [14]. It is interesting to note that several commercial inhibitors contain amines and aldehydes, but seemingly because of the presence of >C=N bond, this makes Schiff bases function more resourcefully in many ways [15]. Stabilization of metal ions in various oxidation states and monitoring their reactivity for catalytic applications have been linked to Schiff bases [16]. The nitrogen-oxygen Schiff bases geometry largely relies on the diamine structural unit, nature of the ancillary ligand, and the central metal ion [17]. Schiff base-transition metal complexes have been known to be one of the most modifiable and comprehensively studied systems [18] with applications in clinical and analytical fields [19, 20]. Antioxidants derived from metal Schiff base ligand combinations have received current attention for their capability to safeguard living systems and cells from impairment caused by oxidative stress or free radicals [21].</p><p>DNA binding, cleavage potentials, scavenging potentials, and anticancer investigations of Schiff base-ruthenium(III) complexes have been accounted for [22]. Synthesis, spectral, redox, catalytic, and biological action investigation of mononuclear Ru(III)-Schiff base structures are reported [23]. 2,2′-Bipyridine and tetradentate Schiff base ancillary ligands of mixed-ligand Ru(II) complexes have been reported for their electrochemical and Na+ binding properties [24]. Catalytic and growth inhibitory activities of Ru(III) mixed ligand complexes of 2-hydroxy-1-naphthylideneimines have been reported [25].</p><p>In this study, we report the synthesis, characterization, free radical scavenging, and anticancer studies of four mononuclear ruthenium(III) complexes of Schiff bases derived from 2′,4′-dihydroxyacetophenone and ethylenediamine as the bridging ligand with RCHO moiety alongside their radicals scavenging action on 1,1-diphenyl-2-picrylhydrazyl (DPPH) and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS) and antiproliferative potentials. The Schiff base ligands containing N2O type tridentate partitions were utilized for the synthesis of the mononuclear ruthenium(III)-Schiff base complexes (Scheme 1).</p><!><p>All reagents used were of analytical grade and used as purchased commercially. Ethylenediamine, N,N′-dimethylformamide (DMF) and ascorbic acid (Vit. C) were received from Merck, 2′,4′-dihydroxyacetophenone and RuCl3·3H2O were obtained from Aldrich. 1,1-Diphenyl-2-picrylhydrazyl (DPPH), 2,2′-azinobis-3-ethylbenzothiazoline-6-sulfonic acid (ABTS), butylated hydroxytoluene (BHT), and rutin hydrate were received from Sigma Chemical Co. (St. Louis, MO, USA). Elemental analysis was carried out using Perkin-Elmer elemental analyzer. IR spectra were recorded on an FT-IR spectrometer: Perkin-Elmer System (Spectrum 2000) via KBr disk method was used for the IR spectra analysis. Freshly prepared DMF solutions of about 10−3 M containing Ru(III) complexes gave the molar conductance at room temperature with Crison EC-Meter Basic 30+ conductivity cell. Electronic absorption spectra ranging from 200 to 900 nm were recorded on a Perkin-Elmer Lambda-25 spectrophotometer. Stuart melting point (SMP 11) was used for the melting points. Four N2O type tridentate ligands, (1Z)-N′-(2-{(E)-[1-(2,4-dihydroxyphenyl)ethylidene]amino}ethyl)-N-phenylethanimidamide [DAE], 4-[(1E)-N-{2-[(Z)-(4-hydroxy-3-methoxybenzylidene)amino]ethyl}ethanimidoyl]benzene-1,3-diol [HME], 4-[(1E)-N-{2-[(Z)-(3,4-dimethoxybenzylidene)amino]ethyl}ethanimidoyl]benzene-1,3-diol [MBE], and N-(2-{(E)-[1-(2,4-dihydroxyphenyl)ethylidene]amino}ethyl)benzenecarboximidoyl chloride [DEE], were synthesized and reported previously [26].</p><!><p>Ethylenediamine (0.015 mol) dissolved in 20 mL of alcohol was slowly added to 2′,4′-dihydroxyacetophenone (0.015 mol) dissolved in same alcohol (30 mL) and allowed to stir for 60 minutes at room temperature and then followed by drop-wise addition of appropriate aldehyde (RCHO, 15 mmol) dissolved in 30 mL alcohol for 20 minutes time interval at room temperature and further stirred for 120 minutes. The mixture was left standing with continuous stirring for approximately 36 hours at room temperature, after which the desired tridentate compounds were filtered and washed with ethanol to give crystalline solid. The crude product was recrystallized from warm ethanol. The products were dried in the vacuum at 50°C overnight to give analytically pure products in good yields (64.2% to 73.8%).</p><!><p>Ru(III) complexes were prepared by adding (0.5 mmol) ethanol solution of ruthenium(III) chloride to a warm ethanolic solution (0.5 mmol) of [DAE]/[HME]/[MBE]/[DEE], respectively. The colour of the solutions changed immediately, magnetically stirred and kept under reflux for 6 hours. The precipitated solids were filtered by suction from the reaction medium, washed with ethanol and then with diethyl ether, and dried over anhydrous calcium chloride. The yields were about 55.7–61.9%. The synthesis of the complexes is explained in Scheme 1.</p><!><p>[Ru(DAE)Cl 2 (H 2 O)]·H 2 O. Dark-green solid; Yield: 156.6 mg (60.4%); F. Wt: 518.38 g; Anal. Calcd. for C18H24N3O4RuCl2 (%): C 41.71, H 4.67, N 8.11; Found (%): C 41.43, H 4.54, N 8.29; IR (KBr) ν max/cm−1: 3436 (O-H), 1621 (C=N), 1242, 1170 (C-O), 520 (Ru-N), 438 (Ru-O); UV-Vis (DMF): λ max/nm (cm−1): 281 (35 587), 310 (32 258), 391 (25 576), 452 (22 124), 525 (19 048), 613 (16 313); Decomp. Temp, °C, 238-239°C; Λμ: 31.8 μScm−1.</p><!><p>[Ru(HME)Cl 2 (H 2 O)]·H 2 O. Darkish-green Solid; Yield: 165.7 mg (61.9%); F. Wt: 535.37 g; Anal. Calcd. for C18H23N2O6RuCl2 (%): C 40.38, H 4.33, N 5.23; Found (%): C 40.58, H 4.21, N 5.44; IR (KBr) ν max/cm−1: 3422 (O-H), 1637 (C=N), 1245, 1173 (C-O), 485 (Ru-N), 437 (Ru-O); UV-Vis (DMF): λ max/nm (cm−1): 277 (36 101), 309 (32 363), 381 (26 247), 393 (25 446), 513 (19 493), 623 (16 051); Decomp. Temp, °C, 218-219°C; Λμ: 30.5 μScm−1.</p><!><p>[Ru(MBE)Cl 2 (H 2 O)]·H 2 O. Darkish-green Solid; Yield: 160.4 mg (58.4%); F. Wt: 549.39 g; Anal. Calcd. for C19H25N2O6RuCl2 (%): C 41.54, H 4.59, N 5.10; Found (%): C 41.29, H 4.32, N 4.98; IR (KBr) ν max/cm−1: 3435 (O-H), 1639 (C=N), 1244, 1171 (C-O), 548 (Ru-N), 475 (Ru-O); UV-Vis (DMF): λ max/nm (cm−1): 277 (36 101), 311 (32 155), 380 (26 316), 393 (25 446), 510 (19 608), 623 (16 051); Decomp. Temp, °C, 226-227°C; Λμ: 30.1 μScm−1.</p><!><p>[Ru(DEE)Cl 2 (H 2 O)]·H 2 O. Dark-green Solid; Yield: 145.9 mg (55.7%); F. Wt: 523.79 g; Anal. Calcd. for C17H20N2O4RuCl3 (%): C 38.98, H 3.85, N 5.35; Found (%): C 39.11, H 3.67, N 5.11; IR (KBr) ν max/cm−1: 3416 (O-H), 1617 (C=N), 1243, 1169 (C-O), 475 (Ru-N), 436 (Ru-O); UV-Vis (DMF): λ max/nm (cm−1): 275 (31 364), 306 (32 680), 385 (25 974), 521 (19 231), 632 (15 823); Decomp. Temp, °C, 228-229°C; Λμ: 38.8 μScm−1.</p><!><p>The potentials of the Ru(III)-tridentate Schiff base complexes to interfere with the growth of TK-10 renal cell line, UACC-62 melanoma cell line, and MCF-7 breast cell lines were determined by SRB assay as previously described [22]. 3–19 passages of MCF-7, TK-10, and UACC-62 cell lines with plating densities of 7–10 000 cells per well were precultured into 96-well microtitre plates for 24 h at 37°C with 95% air, 5% CO2, and 100% relative humidity in RPMI medium, supplemented with 5% fetal bovine serum (FBS), 50 μg mL−1 (gentamicin), and 2 mM L-glutamine [27]. The compounds were dissolved in DMSO and treated with the cells after 24 h and diluted in RPMI medium giving rise to 5 concentrations comprising 0.01, 0.1, 0, 10, and 100 μM.</p><p>Wells containing culture medium were used as control while the wells containing complete culture medium with no cells were used as the blanks. Parthenolide was used as the standard drug in this study. The plates were then incubated for 48 h after the addition of the compounds. Viable cells were fixed to the bottom of each well with cold 50% trichloroacetic acid, washed, dried, and dyed by SRB. The unbounded dye was separated, while the protein-bound dye was extracted with 10 mM Tris base and multiwell spectrophotometer at the wavelength 540 nm was used for its optical density determination. IC50 values were determined by plotting the percentage viability against concentration of compounds on a logarithmic graph to obtain 50% of cell growth inhibition relative to the control.</p><!><p>The antioxidant activity of the prepared Ru(III) complexes was studied using spectrophotometer by 1,1-diphenyl-2-picrylhydrazyl (DPPH) method. This compound is known as a stable readily accessible free radical, with solubility in methanol giving a purple solution, and when reacted with antioxidant species changes to an equivalent light yellow colour. The radical scavenging potentials of the complexes with DPPH radical were evaluated as described [22]. 1 mL solution of the compounds in DMF with concentrations ranging from 100 to 500 μg/mL was mixed thoroughly with equivalent amount of 0.4 mM DPPH in methanol; the mixtures were then allowed to react in the dark for half an hour. Measurement of the mixture absorbance was achieved spectrophotometrically at 517 nm. Vitamin C and rutin were used as the standard drugs. All test analysis was carried out in triplicate. The ability of the ruthenium compounds to scavenge DPPH radical was calculated via the following equation:(1)DPPH  radical  scavenging  activity  %=Absorbance  of  control−Absorbance  of  sampleAbsorbance  of  control×100.</p><!><p>ABTS scavenging ability of the Ru(III)-tridentate Schiff base complexes adopted a described method [28]. 7 mM ABTS solution and 2.4 mM potassium persulfate solution in equal amounts (1 : 1) were used for working solution preparation and allowed to react in the dark for 12 h at room temperature. An absorbance of 0.706 ± 0.001 units at 734 nm required for the analysis was obtained by diluting 1 mL ABTS+ solution. Test samples (1 mL) were mixed with 1 mL of the ABTS+ solution, and absorbance was read spectrophotometrically at 734 nm. The test samples' ABTS scavenging capacity alongside standard drugs was evaluated. Triplicate analysis was carried out. The percentage inhibition of ABTS radical scavenging activity was obtained following a previous report [28].</p><!><p>The obtained compounds were of coloured powders, stable in atmosphere with a general formula: [Ru(LL)Cl2(H2O)] (LL = monobasic tridentate Schiff base anion: DAE, HME, MBE, and DEE). They were prepared by treating [RuCl3·3H2O] with the corresponding Schiff base in an equal mole ratio in alcohol as depicted in the Scheme 1. All the complexes are dark-green and sparingly soluble in general organic solvents but soluble in polar aprotic solvent such as DMF and DMSO; the melting point analysis showed that the Ru(III) complexes were decomposing before melting. The physicoanalytical data collected for the compounds are in agreement with the structural formulae proposed, thus confirming the suggested mononuclear composition for the Ru(III) complexes (Scheme 1).</p><!><p>The molar conductance of the synthesized Ru(III) complexes was measured in DMF at 10−3 M solution. The values were found to be in the range of 30.1–38.8 μScm−1 suggesting the nonelectrolytic nature of the complexes in solution [22, 29].</p><!><p>Valuable evidence concerning the environment of the functional group attached to the ruthenium atom has been obtained from the FTIR spectra. The IR spectra of the ligands, when compared with those of the newly synthesized complexes, confirm the coordination of N2O type tridentate ligands to the ruthenium ion. The classification was achieved by comparing the spectra of the ligands with those originating from the coordination between ruthenium(III) metal ion and the active sites. The Schiff bases showed the broad bands in the 3462–3477 cm−1 range attributable to the ν(OH) cm−1 vibration. Ligand infrared spectra showed that a band at 1605–1619 cm−1 is attributed to ν(C=N) stretching of the azomethine group based on earlier reports [30]. This ν(C=N) shift to 1617–1639 cm−1 in all the complexes by about 5–23 cm−1 signifies the participation of azomethine nitrogen in the coordination sphere with the ruthenium(III) ion for all the complexes [21, 31]. A medium band that corresponds to phenolic oxygen atom ν(C-O) is observed at 1167 and 1245 cm−1 for the free ligands.</p><p>The higher shifting of ν(C-O) stretching vibrations as observed in the ruthenium(III) complexes spectra suggests that the phenolic OH group of Schiff base, DAE, HME, MBE, and DEE, is involved in coordination with ruthenium ion after deprotonation [32, 33]. Seemingly, the DAE, HME, MBE, and DEE ligands act as a tridentate chelating compound, coordinating to the metal ion via the two nitrogen atoms of the azomethine group as well as O atom of phenolic group [21, 25]. This is further supported by the displacement of ν(O-H) in the range 3462–3477 cm−1 in all the complexes. The presence of coordinated water gave a broad band that appeared in the regions 3416–3436 and 813–851 cm−1; this can be due to ν(O-H) stretching and ν(O-H) rocking vibrations, respectively, which further confirms the presence of nonligand assignable to the rocking mode of water [28, 34]. New weak nonligand bands that are not found in the DAE, HME, MBE, and DEE ligands appeared in the ranges 475–548 cm−1 and 436–475 cm−1 in the complexes spectra attributed to ν(Ru-N) and ν(Ru-O) vibrations, respectively [35, 36]. A band ranging from 311–346 cm−1 appeared in the spectra of the Ru(III)-Schiff base complexes indicating the presence of two chloride ions in trans position around ruthenium centre [37–40].</p><!><p>The UV-Vis spectra of the Ru(III)-Schiff base complexes in DMF solutions were recorded at room temperature ranging from 200 to 900 nm. The nature of DAE, HME, MBE, and DEE ligands field around the ruthenium ion was obtained from the electronic spectra. The free ligands showed absorption bands within the range of 277–393 nm attributable to π ∗ ← π and π ∗ ← n transitions relating the benzene ring (Figure 1). The shifting of these bands in the complexes spectra followed the participation of the imine group nitrogen and phenolic group oxygen in bonding [22, 25]. Ground state of ruthenium(III) is 2T2g, where initial excited doublet levels in order of increasing energy are 2A2g and 2T1g, arising from t2g 4eg 1 configuration [41].</p><p>Ru3+ ion, with a d5 electronic configuration, possesses high oxidizing properties and large crystal field parameter. Also, charge transfer bands of the type Lπy → T2g were noticeable within low energy region, obscuring weaker bands that is due to d-d transitions [22, 25]. The extinction coefficient bands around 613–632 nm regions are found to be low when compared to the charge transfer bands. These bands have been assigned to 2T2g → 2A2g transition and are in agreement with the assignment made for similar octahedral ruthenium(III) complexes [42, 43]. Absorption bands within the 452–525 nm regions were assigned to the charge transfer transitions [22, 44]. Overall, the absorption spectra of the Ru(III)-Schiff base complexes are typical of octahedral environment about the ruthenium(III) ions [22].</p><!><p>Investigation into the structure-activity relationship of the isolated Ru(III)-N2O Schiff base complexes with respect to different functional groups on the ligands used for ruthenium ion complex formation has been conducted via antiproliferative studies. Three of the Ru(III)-Schiff base compounds alongside parthenolide were subjected to cell lines tests at different sample concentrations ranging from 0.01 to 100 μM towards renal cancer cell (TK-10), melanoma cancer cell (UACC-62), and breast cancer cell (MCF-7). The cancer cell lines were incubated for 48 h, followed by the addition of the compounds of various concentrations via Sulforhodamine B (SRB) assay [22].</p><p>The ruthenium(III) compounds and standard drug (parthenolide) IC50 values are presented in Table 1 and revealed that the test samples showed significant inhibition against the tested cell lines. Figures 2 –4 represent the cell viability percentages of ruthenium(III)-Schiff base complexes and parthenolide drug against TK-10, UACC-62, and MCF-7 cell lines, at different concentrations of ruthenium(III) compounds or parthenolide. A high level of antiproliferative potentials against the studied cell lines was exhibited by parthenolide in accordance with earlier reports [45]. The obtained results revealed that treatment of cell lines with different concentrations of Ru(III)-Schiff base complexes efficiently affected cell viability towards MCF-7 cells, as displayed in Figures 2 –4 and Table 1. The Ru(III) compounds exhibited low to strong in vitro antiproliferative activities against the selected cell lines as compared to the standard drug (parthenolide). [Ru(DAE)Cl2(H2O)], [Ru(HME)Cl2(H2O)], and [Ru(DEE)Cl2(H2O)] induced more efficient cell death with IC50 values of 3.57 ± 1.09, 4.88 ± 1.28, and 3.43 ± 1.48 μM, respectively, towards human breast cancer cell (MCF-7) cells than other investigated cell lines, compared with IC50 values of 0.44 ± 2.02 μM MCF-7, for the standard cytotoxin drug parthenolide.</p><p>The order of activity of the complexes against human melanoma cancer cell (UACC-62) is as follows: [Ru(DEE)Cl2(H2O)] > [Ru(HME)Cl2(H2O)] > [Ru(DAE)Cl2(H2O)]. With respect to previous report by Shier [46], compounds exhibiting IC50 activity ranging from 10 to 25 μM are referred to as weak anticancer drugs, while those with IC50 action between 5 and 10 μM are moderate, and the compounds possessing activity less than (<) 5.00 μM are considered as strong agents. Thus, the Ru(III) complexes exhibited a weak to strong activity against the investigated cancer cell lines with the following order of activity: MCF-7 > UACC-62 > TK-10. However, [Ru(DAE)Cl2(H2O)] showed the highest antiproliferative activity with IC50 valves of 3.57 ± 1.09, 6.44 ± 0.38, and 9.06 ± 1.18 μM for MCF-7, UACC-62, and TK-10, respectively. The biochemical activity could be due to the methoxy, alkyl, chloride group substituents and bridge spacer: ethylenediamine, which could have played a vital role in antiproliferative potentials of the Ru(III)-N2O Schiff base complexes. In vitro anticancer activity of the synthesized Ru(III) complexes in this study was compared with Ru complexes reported by other authors and found that [Ru(DAE)Cl2(H2O)], [Ru(HME)Cl2(H2O)], and [Ru(DEE)Cl2(H2O)] complexes exhibited higher antitumor activities. [RuCl(CO)(PPh3)L] reported by Raja et al. [47] against human cervical carcinoma cell line, (HeLa) after exposure for 48 h, gave an IC50 value in the range of 31.6 μM and [RuCl2(AsPh3)L] with an IC50 value of 37.8 μM [48]. Raju et al. [43] reported ruthenium(III) Schiff base complexes of the type [RuX2(PPh3)2(L)] (where X = Cl or Br; L = monobasic bidentate ligand) complex to have IC50 value in the range of 45.2 μM.</p><!><p>Different antioxidant techniques and modifications have been put forward to evaluate antioxidants reactivity and functionality in foods and biological systems as a means of checkmating variety of pathological activities such as cellular injury and aging process; these damaging occurrences are caused by free radicals. Hence, two free radicals were used for in vitro antioxidants activities of the test samples in this study, namely, 1,1-diphenyl-2-picrylhydrazyl (DPPH) and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS).</p><!><p>The activity of antioxidants on DPPH radical is believed to be centred on their ability to donate hydrogen [22]. DPPH has been a stable free radical, with the ability to accept hydrogen radical or an electron and then become a stable molecule [49].</p><p>The mode of rummaging the DPPH radical has extensively been used to appraise antioxidant activities of test samples in a moderately short period of time compared to other procedures [49]. The reduction in the DPPH radical capability is calculated by the decrease in its absorbance at 517 nm prompted by antioxidants [50]. The reduction of DPPH radical intensity in this study is due to the interaction of Ru(III) complexes with radical and as such scavenging the radicals by hydrogen donation (Scheme 2). The DPPH activities by the Ru(III)-N2O Schiff base complexes exhibit strong electron donating power when compared to the standards: ascorbic acid and rutin as displayed in Figure 5. The calculated IC50 and its corresponding R 2 (correlation coefficient) values of Ru(III) compounds are listed in Table 2. Compounds [Ru(DAE)Cl2(H2O)], [Ru(HME)Cl2(H2O)], [Ru(MBE)Cl2(H2O)], and [Ru(DEE)Cl2(H2O)] with an IC50 value of 1.60 ± 0.68, 1.54 ± 0.44, 1.63 ± 1.05, and 1.51 ± 0.50 μM, respectively, exhibited higher activity against DPPH than the commercially available Vit. C and rutin (standard); however, [Ru(DEE)Cl2(H2O)] showed the highest activity of all investigated ruthenium(III) samples with an IC50 value of 1.51 ± 0.50 μM.</p><p>Scavenging ability of the test samples on the DPPH radical can be ranked in the following order: [Ru(DEE)Cl2(H2O)] > [Ru(HME)Cl2(H2O)] > [Ru(DAE)Cl2(H2O)] > [Ru(MBE)Cl2(H2O)] > [Vit. C] > [rutin]. The scavenging effect of the DAE, HME, MBE, and DEE ligands is lower as compared to their corresponding Ru(III) complexes, owing to the coordination of the organic molecules to the Ru3+ ion. It is further supported by the observed discolouration from purple DPPH radical solution to yellow solution showing scavenging of the DPPH radicals by hydrogen donation (Scheme 2). Hence, these complexes could be effective therapeutic agent's preparation for the treatment of chronic conditions such as cardiovascular, neurodegenerative, and arteriosclerosis diseases [21].</p><!><p>To further confirm the synthesized Ru(III)-N2O Schiff base complexes antiradical potential, we examined the ABTS assay in this study. A well-known protonated radical like 2,2′-azinobis-3-ethylbenzothiazoline-6-sulfonic acid (ABTS) possesses characteristic absorbance maxima at 734 nm and decreases with the scavenging of the proton radicals [51]. The assay measures radical scavenging by electron donation. The outcome of Ru(III)-N2O Schiff base complexes alongside the standard drugs on ABTS radical is presented in Table 2. At 734 nm, the absorbance of active ABTS∗ solution noticeably declined upon the addition of different concentrations of ruthenium(III) samples; the same trend was also observed for the standard drugs: butylated hydroxytoluene (BHT) and rutin hydrate with the percentage inhibition displayed in Figure 6.</p><p>The efficacy of the tested samples in quenching ATBS∗ radicals in the system was observed at 100 μg/mL, the lowest concentration, and Ru(III) complexes exhibited higher ABTS % inhibition than the standards. [Ru(DEE)Cl2(H2O)] complex exhibited the highest ABTS scavenging activity amongst the studied ruthenium(III) complexes with an IC50 value of 3.24 ± 0.93 μM and 0.855 R 2 (correlation coefficient) as listed in Table 2 while complexes of [Ru(DAE)Cl2(H2O)], [Ru(HME)Cl2(H2O)], and [Ru(MBE)Cl2(H2O)] had an IC50 value of 3.30 ± 0.89, 4.27 ± 1.17, and 3.30 ± 1.48 μM, respectively.</p><p>The ABTS scavenging activity pattern of the complexes is ranked in the following order: [Ru(HME)Cl2(H2O)] < [Ru(MBE)Cl2(H2O)] = [Ru(DAE)Cl2(H2O)] < [Ru(DEE)Cl2(H2O)]. With this result, the antiradical studies showed that the synthesised Ru(III)-N2O Schiff base complexes may be useful in developing therapeutic agent for averting cell oxidative damage and as radicals chain terminator. This is because various free radicals generated in the system often lead to cancer, cellular injury, aging process, and cardiovascular diseases [21].</p><!><p>In this study, we present the synthesis of Ru(III) Schiff base complexes formulated as [Ru(LL)Cl2(H2O)] (LL = DAE, HME, MBE, and DEE). The complexes were characterized using the microanalytical, conductance, electronic, and vibrational spectral analysis. FTIR spectral data showed that the ligand acts as tridentate chelating ligand, coordinating through azomethine nitrogen and phenol oxygen atom. The microanalyses were in conformity with the proposed structures. Conductance measurements showed the complexes to be nonelectrolytes in DMF. Octahedral structures were assigned to these complexes based on the elemental and spectral information. In vitro antiproliferative studies of the Ru(III) complexes gave a weak to strong inhibition against the studied cancer cell lines, with the following activity order: MCF-7 > UACC-62 > TK-10. Significantly, further investigation on the compounds free radical scavenging properties revealed that Ru(III)-Schiff base complexes possessed considerable antioxidant activities. The outcome from DPPH and ABTS inhibition studies revealed that the compounds are proficient in donating electron or hydrogen atom and subsequently terminate the chain reactions in a dose-dependent pattern. Scavenging ability of the test samples on the DPPH radicals can be ranked in the following order: [Ru(DEE)Cl2(H2O)] > [Ru(HME)Cl2(H2O)] > [Ru(DAE)Cl2(H2O)] > [Ru(MBE)Cl2(H2O)]. Thus, Ru(III)-N2O Schiff base complexes showed stronger inhibition of DPPH at various concentrations.</p>

PubMed Open Access

High-Performance, Single-Crystal Gold Bowtie Nano-Antennas via Epitaxial Electroless Deposition

Material quality can play a critical role in the performance of nanometer-scale plasmonic structures. Here, we highlight a novel deposition strategy for single-crystal noble metal deposition and provide a direct and quantitative comparison between the fabrication yield, durability, and efficiency of bowtie nano-antennas fabricated from monocrystalline and polycrystalline gold films using subtractive nanofabrication. Focused ion beam milling of monocrystalline Au(100) films deposited through epitaxial electroless deposition to form bowtie nano-antennas produces devices that demonstrate key performance enhancements over devices patterned identically from polycrystalline Au films deposited via physical vapor deposition. Single-crystal bowties reveal significant improvements in pattern transfer fidelity and device yield, the ability to tailor and model local plasmonic field enhancements and marked improvement in their thermal and mechanical stability over those fabricated from polycrystalline Au films. This work underscores the performance advantages of single-crystal nanoscale plasmonic materials and describes a straightforward, solution-phase deposition pathway to achieve them.We anticipate that this approach will be broadly useful in applications where local near-fields can enhance light−matter interactions, including for the fabrication of optical sensors, photocatalytic structures, hot carrier-based devices, and nanostructured noble metal architectures targeting nanoattophysics.

high-performance,_single-crystal_gold_bowtie_nano-antennas_via_epitaxial_electroless_deposition

INTRODUCTION<!>Yield and Activity as a Function of Film Quality<!>Polarization Dependence of the Nano-Antennas<!>Device Stability<!>Plasmonic Activity and Field Enhancement<!>CONCLUSION<!>Monocrystalline Silver Deposition on Silicon<!>Electroless Deposition of Monocrystalline Gold on Silver<!>Bowtie Gold Nano-Antenna Fabrication<!>Benzoic Acid (BA) Preparation for SERS<!>Finite-Difference Time-Domain Simulations<!>Surface Enhanced Raman Scattering Measurements (SERS).

<p>The coupling of extended electromagnetic waves to planar metal/dielectric interfaces through surface plasmon polaritons (SPPs) or to nanometer-scale metal structures through locally resonant surface plasmons (LRSPs) leads to confined and amplified local fields that can be exploited for application in energy harvesting, sensing, spectroscopy, catalysis and imaging. The fate of these plasmonic excitations is intimately linked with the characteristics of the materials from which they are formed. [1][2][3][4][5][6] SPP propagation lengths, SP dephasing, decay, and decoupling are influenced strongly by material crystallinity and scattering processes that are induced by material defects, grain boundaries, and other material imperfections. Single-crystal plasmonic structures are expected to yield advantages over their polycrystalline analogues through reductions in optical absorption loss, grain boundary scattering and dissipation, while providing enhanced local fields derived from well-defined faceted nanostructures.</p><p>Conventional deposition of plasmonic metals such as gold is typically carried out through physical vapor deposition (PVD) techniques and generally yield polycrystalline metal films and nanostructures.</p><p>While deposition strategies and other protocols to mitigate the polycrystalline character of these films have been developed, 7 polycrystalline metal deposition can lead to compromised fabrication yields, 8,9 as well as loss and dissipation that result in device inefficiency, and remains a significant challenge in the field. We have recently developed an alternative approach to achieve ultrasmooth monocrystalline Au(100) films via electroless deposition from highly alkaline solutions of common gold salts onto Ag(100)/Si(100) substrates. 10 The method is scalable to the wafer level, environmentally friendly, and represents a promising new approach to the integration of noble metal-based plasmonic structures into CMOS compatible device architectures. 11,12 Here, we use this approach to fabricate bowtie nanoantenna devices in a 100 nm thick single-crystal Au(100) film by subtractive patterning. Focused ion beam (FIB) milling of these single-crystal films results in high quality, low defect density, monocrystalline bowtie antenna structures. By contrast, we have also deposited 100 nm thick polycrystalline gold films by evaporation, utilizing a Si(100) wafer with a 5 nm Cr adhesion layer as a substrate (supporting information), followed by patterning identical bowtie nano-antenna structures through FIB milling. In this manuscript, we employ these bowtie antennas to provide a direct comparison between the performance of single-crystal and polycrystalline plasmonic devices.</p><p>Bowtie nano-antenna devices were fabricated by a Thermo Fisher Helios NanoLab 650 SEM/FIB system, using a focused gallium ion beam. Figure 1a,b illustrate the sequential milling of material as the focused gallium ion beam is moved over surface regions in a serial fashion to create the bowtie nano-antennas on the surface. Figure 1c shows a top-view SEM of the milled single-crystal (left) and polycrystalline (right) bowtie structures. The images reveal significant differences in the quality of pattern transfer, with the milled regions of the monocrystalline film appearing highly uniform, and those of the polycrystalline film much more irregular. The lack of milling uniformity in the polycrystalline films results from anisotropic, crystal direction-dependent ion milling rates and provides a bowtie structure defined by the remaining non-milled area, surrounded by a region of recessed roughened gold. Note that the pattern generation scheme involved milling rectangular and diamond regions sequentially. This process yields milled regions surrounding the bowtie that lie at different depths within the film which are separated by small vertical step edges. These regions can be seen readily (Fig 1c, left) in the areas of overlap of the rectangular and diamond regions. The dimensions of the milled geometrical features were chosen to create a bowtie antenna with a length, 𝐿𝐿 = 1560 𝑛𝑛𝑛𝑛, a gap, 𝑔𝑔 = 20 𝑛𝑛𝑛𝑛, and height, ℎ = 50 𝑛𝑛𝑛𝑛.</p><p>The bowtie nano-antenna dimensions were selected so that they could be resonantly excited with available 780 𝑛𝑛𝑛𝑛 laser radiation to activate the devices and produce the field confinement at the antenna feedpoint.</p><!><p>Focused ion beam milling of 3 × 10 bowtie nano-antenna arrays was performed on single-crystal and polycrystalline gold films. The performance of the bowtie arrays was assessed with a Zeiss scanning laser microscope (SLM) equipped with a 63x objective lens, and a wavelength tunable Coherent Chameleon ultrafast oscillator (80 MHz repetition rate, 140 fs pulse duration) used to activate the antennas.</p><p>Resonant excitation of the bowtie nano-antennas leads to two-photon photoluminescence (2PPL) that is well-correlated with the locally resonant surface plasmon excitation of the structures. 2PPL imaging has been used extensively to characterize the resonant behaviour of plasmonic nanostructures 1,[13][14][15][16][17][18][19][20] and is utilized here as a measure of the nano-antenna plasmonic response and local field enhancement. These structures provide a stringent test of fabrication precision and yield, with the goal of uniform, reproducible and intense field confinement at the antenna's feedpoints.</p><p>2PPL intensity maps of the bowtie arrays induced by 780 nm laser excitation are presented in Figure 2 and highlight the primary performance differences between the mono-and polycrystalline nanoantennas. The 2PPL maps demonstrate that fabrication yield is greatly impacted by the material quality and subsequently by the resulting pattern transfer characteristics. The yield of monocrystalline bowtie antennas is close to 100% as measured by the appearance of an enhanced confined local near-field demonstrated in the 2PPL intensity at the 20 nm wide antenna feedpoints, and the relative uniformity of this 2PPL intensity for the vast majority of antennas, (Figure 2a). Structures milled identically in the polycrystalline-deposited gold (Figure 2b), show a poor fabrication yield with few devices showing nearfield intensity enhancements at the antenna feedpoints, and of these, little uniformity in 2PPL intensity.</p><p>Note that the fabrication differences between the mono-and polycrystalline structures (e.g. the presence of a Cr adhesion layer in the case of the polycrystalline antennas) can potentially lead to differences in the resonant response characteristics of the antennas. However, scanning of the 10 nm bandwidth laser wavelength in the vicinity of the expected resonant excitation spectra (780 nm) did not yield improvements in the emission characteristics of the polycrystalline antennas.</p><p>2PPL emission from the polycrystalline antennas (Fig 2b ) shows poor fabrication yield with few antenna structures displaying 2PPL gap intensity. While the integrated emission intensity from the polycrystalline antennas appears brighter than that from single-crystal devices, the vast majority of the 2PPL emission from polycrystalline devices emanates from the roughened recessed regions surrounding the bowties, and not from the antenna's feedpoints, as desired. This "background" emission results from the roughened nature of the surrounding regions, as SP's scatter from polycrystalline grain boundaries and material defects that arise from non -uniform and anisotropic milling. Further, the bright, localized 2PPL emission from monocrystalline antenna feedpoints, is significantly more intense than the average le vel of background emission emanating from polycrystalline devices, reflecting larger and more uniform field enhancement factors in the single -crystal bowtie gaps.</p><!><p>The activity of the bowtie structures are known to be highly polarization sensitive. The bowtie nano-antennas fabricated on mono-and polycrystalline Au films were studied under vertically-and horizontally-polarized 780 nm laser irradiation at normal incidence. Their polarization-dependent 2PPL emission characteristics are illustrated in Figure 3, along with a numerical simulation of the anticipated response calculated using a finite difference time domain (FDTD) model of the bowtie structures (Lumerical). To compare the modelled and the experimentally measured antenna response accurately, the geometrical shapes employed in the FIB milling protocol of the fabricated devices were used to design the nano-antennas for the FDTD software model. The results presented in Figures 3a,b represent the simulated device response for plane wave excitation at 780 nm for vertically-and horizontally-polarized light with respect to the bowties. As anticipated, the electric field distribution across the device is polarization-sensitive and shows field maxima lines that lie orthogonal to the polarization direction. The milling protocol results in the formation of recessed regions of the film that define local plasmonic cavities characterized by sharply-edged walls.</p><p>Light that is orthogonally-polarized to the wall edges is edge-coupled into these cavities which are capable of supporting SP modes that appear as field intensity maxima in the FDTD simulations. These Comparison of the plasmonic response of the simulated bowties with the fabricated bowties reinforces the significant differences in pattern transfer quality of the mono-and polycrystalline devices. Finite quality factors of the milled cavities will couple a range of incident wavelengths into the structures that can lead to SP mode interferences. Complex constructive and destructive interferences resulting from the multiple SP cavities that define the milled structure may contribute to intensity differences between the simulated and 2PPL intensity maps.</p><p>Comparison of the 2PPL emission response from polycrystalline bowties (Fig 3e-f) shows very modest polarization dependence, the nature of which is significantly different from that observed from the monocrystalline antennas. Poor pattern transfer quality in the polycrystalline antennas leads to little or no well-defined mode structure as observed in the case of the single-crystal antennas. Plasmonic excitation and rapid decay through grain boundary and defect induced plasmon dissipation leads to 2PPL "background" emission with little polarization character. However, it should be noted that the overall intensity of 2PPL emission appears to be more intense for horizontally-polarized excitation, presumably due to the enhanced coupling of light that is enabled by the bowtie antenna for this polarization.</p><p>Further refinements in film quality, pattern transfer, and simulation accuracy are currently underway in our laboratory to improve the level of agreement between simulated and fabricated structures.</p><p>Nevertheless, the high quality of material deposition enabled through our electroless deposition process, provides good agreement between simulation and experiment.</p><!><p>The effect of material quality on device stability was also investigated. To do so, the 2PPL intensity emanating from bowties was evaluated upon increasing the power of the incident laser. Figure 4 ure reflect the percentage of total laser output power coupled into the scanning laser microscope. The actual power incident on the sample through the SLM 63x objective is a small fraction of this intensity, but scales linearly with the displayed percentage, as measured independently in the absence of a sample with a calibrated power meter. As the laser power is increased, both mono-and polycrystalline devices emit increased 2PPL emission intensity as expected, since 2PPL intensity is proportional to I 2 , where I is the local near-field intensity enhancement 1,13,15,20 . The antennas appear to be non-emissive at low incident intensity, however this is misleading, as the 2PPL emission intensities displayed in Fig 4 have been normalized to the maximum emission intensities observed under high intensity illumination. Figure 4 demonstrates that as the incident intensity is systematically increased, so too is the bowtie gap intensity. Further increase in incident intensity results ultimately in the catastrophic rupture of the devices as indicated by the loss of bowtie structure and saturated emission intensity in the 2PPL image maps. We attribute the catastrophic destruction of the bowtie structures to plasmonic decay via photothermal mechanisms, generating local heating effects that exceed the thermal and mechanical stability of the structures. Inspection of Figure 4 reveals that the threshold incident intensity necessary to induce catastrophic damage under these illumination conditions is approximately ten times greater for single-crystal bowtie devices (~45% incident intensity) than for polycrystalline devices (4.5% incident intensity). The power incident on the bowtie samples after the 63× objective was found to scale approximately linearly with percent power transmission as indicated on the microscope and in Figure 4. The power measured at 4% transmission corresponded to 8.8 mW of incident radiation (just below the onset of damage in the polycrystalline bowties), whereas the power at 48% transmission was measured to be 84 mW (just below the onset of damage observed in the single crystal bowties), approximately 1 order of magnitude higher. We attribute this large difference to the presence of grain boundaries and defects in the polycrystalline structures which increase the dissipation of SPs to heat over the entire milled region of the structures (bowtie and background). Further, the polycrystalline structures of these antennas are anticipated to be less thermally and mechanically stable than their corresponding single-crystal counterparts, leading to lower thresholds for bowtie destruction. The lack of grain boundaries in the monocrystalline Au films does not provide such a path for distributed SP photothermal decay. Further, the decay of SP in single-crystal structures can be mediated by additional longer-range mechanisms of thermal conduction (e.g. via phonon dissipation) that are unavailable in polycrystalline structures comprised of nanoscale grains. Thus, our stability study indicates that the single-crystal bowtie structures can support approximately 10 times more incident illumination intensity, corresponding to 10 2 greater 2PPL intensity, and therefore, a factor of 10 4 greater local field, beyond that of polycrystalline bowties, before irreversible and catastrophic loss.</p><!><p>Surface enhanced Raman spectroscopy (SERS) is a well-known and well-studied process in which the local excitation of SPs leads to a significant enhancement in the Raman scattered light collected from surface molecules [21][22][23][24] . The locally excited electric field and the Raman enhancement can be achieved using nanoparticles and nanostructures made from plasmonic noble metals, [24][25][26] or with the help of nano-scaled devices with resonating cavities that can confine the excited SPs within very small gaps [27][28][29][30][31][32][33][34][35] . Here, the SERS response from the common Raman reporter molecule benzoic acid (BA) is used to compare the SERS efficiency as a measure of the relative magnitude of the field confinement for mono-and polycrystalline bowtie nano-antennas. In a receiving antenna, the maximum power gain is directly related to the maximum effective area of the antenna, A e , which is calculated through:</p><p>where λ is the wavelength of the incident photon 36 . Since the bowtie nano-antennas on mono-and polycrystalline Au films are identical in design, we expect the A e to remain constant across the fabricated devices on both surfaces and therefore the difference in the performance can be attributed to the quality of the materials. The field confinement magnitude at the gap of the plasmonic bowtie nano-antennas is linked to the coupling efficiency of photons to SPs, which in turn, is a function of surface quality of the film from which the device is made 1,[4][5][6] . The surface roughness of the polycrystalline devices negatively impacts the intensity of excited SPs at the bowtie feedpoint by enabling photon-SP decoupling at grain boundaries and material defects, thereby reducing the magnitude of the field at the gap. This route for SP intensity decay is minimized for the monocrystalline Au nano-antennas, resulting in a larger gap field. Both mono-and polycrystalline devices were coated with 10 μL of 0.02 M BA in methanol, by drop casting, followed by solvent evaporation. SERS was carried out using a Renishaw Invia Raman microscope and a fiber coupled continuous wave 785 nm diode laser, as the excitation source. The Raman spectra were collected at 50% incident laser intensity with a 10s exposure time. The bowties were far enough apart from one another that Raman data from single devices could readily be acquired. The SERS spectra from BA coated bowties appear in Figure 5 and are representative of the mono-and pol-ycrystalline responses from many bowtie measurements. The data suggest that the larger observed SERS enhancement from single-crystal antenna can be attributed to the quality of the Au film from which the devices were fabricated and that single-crystal nanostructures support larger near-field gap intensities than their polycrystalline counterparts, suggesting significant advantages in the use of single-crystal plasmonic materials. Performance improvement and antenna efficiency are expected to result in part from device material conductivity, 37,38 reducing conduction losses as the captured field of incident photons is directed toward the antenna feedpoint. At optical frequencies, ohmic losses occur in close proximity to the surface and therefore material quality and conductivity of the metal plays a critical role in determining device impedance. 36,37 We have previously reported the improved conductivity of the solution-deposited monocrystalline gold films compared to their vapor-deposited counterparts. 10 Four-point probe measurements on both mono-and polycrystalline gold surfaces indicate that the conductivity of the single-crystal gold surfaces are greater by approximately a factor of 20 over the polycrystalline films.</p><!><p>We have demonstrated a new scalable and green solution-deposition method for the fabrication of large area single-crystal Au(100) films and the subtractive manufacture of single-crystal plasmonic bowtie nano-antennas with improved fabrication and performance yields compared to their polycrystalline counterparts. We have presented a direct and quantitative comparison of the performance of mono-versus polycrystalline plasmonic bowtie nano-antennas. Single-crystal bowties were fabricated via FIB milling of Au(100) films deposited by epitaxial electroless deposition from highly alkaline deposition baths onto Ag(100)/Si(100) substrates. Polycrystalline antennas were fabricated through an identical patterning protocol on polycrystalline films deposited by Au evaporation onto a Si(100) wafer containing a 5nm thick Cr adhesion layer. The quality and yield of pattern transfer onto single-crystal films far surpasses that of polycrystalline films and leads to significant performance advantages of the single-crystal devices. These include the uniformity, and intensity of local near-field distributions, the ability to model accurately these distributions, and the resulting stability of single-crystal devices compared to their polycrystalline analogues. Single-crystal devices demonstrate the ability to support one order of magnitude more incident intensity (and therefore 10 4 times the local field) than polycrystalline devices, before their catastrophic loss via photothermal decay. This enhanced stability is attributed to the greater thermal and mechanical characteristics of single-crystal materials. Single-crystal bow-ties have also shown a greater SERS enhancement factor than polycrystalline structures through reduced photon-surface decoupling and absorption loss, providing greater local gap fields that enhance the light-matter interaction.</p><p>While our study has focused specifically on bowtie nano-antennas, we anticipate that the performance of plasmonic device structures more generally will benefit from single-crystal materials and that the epitaxial electroless deposition approach employed here will be broadly useful for single-crystal device fabrication.</p><!><p>Silver Ag(100) deposition was carried out using a Kurt J. Lesker Company PVD-75 thermal evaporation tool with a base pressure of <2 × 10 -7 Torr. Ag (99.99% Kurt J. Lesker Company) was evaporated from an alumina coated tungsten wire basket. The substrate was heated via a backside quartz lamp and the temperature was monitored with a K type thermocouple attached to the backside of the sample chuck assembly. Deposition was carried out at a substrate temperature of 340 ⁰C and a rate of 3 Å/s. Prior to Ag deposition, substrates were immersed in either dilute HF acid solutions (10:1 with de-ionized water), or similarly diluted commercial buffered oxide etch solutions (BOE, CMOS Grade, J.T. Baker Inc.), to remove the native oxide layer from the surface of the silicon wafer. All activities, prior to characterization of the films, were carried out under class 100 clean room conditions or better.</p><!><p>A 1 x 1 cm 2 Ag(100) substrate was used as surface on which to grow a 200 nm thick monocrystalline Au(100) film electrolessly. The Ag substrate was submerged in 10 mL of 1 M NaOH which acted as the deposition bath.</p><p>Then 250 μL of 0.025 M of HAuCl 4 solution was added to the deposition bath (10 mL NaOH). The solution was placed in a water bath where its temperature was kept at 70°C for 60 minutes undisturbed to grow a 200 nm thick monocrystalline Au(100) film on the Ag(100)/Si(100) substrate. The sample was then washed with distilled water and sonicated in isopropanol alcohol for 60 s and air dried.</p><!><p>An FEI Helios Focused-Ion beam (FIB) tool (4D LABS) was used to fabricate the gold bowtie nano-antennas.</p><p>The process was carried under the pre-set conditions in the tool for Au films, in which the desired milling depth was 50 nm. The ion beam current was set at 7.7 pA for the 30 kV operating voltage. Under these conditions, for 50 nm depth etching the dose was set to be 33 pC/μm 2 and this value was doubled for the milling the monocrystalline Au film. The exposure time for fabrication of bowtie nano-antennas on the monocrystalline Au film was also increased by a factor of 2 over the parameters used for milling polycrystalline films to achieve a milling depth of 50 nm, due to the lower material removal rate for single-crystal Au. Figure 1 shows the fabricated bowtie antenna on both monocrystalline and polycrystalline Au achieved under these etching conditions. The dimensions of the nano-antennas (L=1560 nm) was designed to be twice the wavelength of the 780 nm incident photons to achieve efficient coupling.</p><!><p>The BA solution was prepared by dissolving 0.0488 g of BA solid powder (Coleman & Bell) in 20 mL of methanol to achieve 0.02 M concentration. From this solution, 10 μL of BA was drop casted on the bowtie nanoantennas described in this paper.</p><!><p>The FDTD analysis was carried out using Lumerical Solutions FDTD tool to simulate the electric field distribution across the surface of the fabricated bowtie for comparison with the experimental results. The design of the structures input to the FDTD model were as close as possible to the fabricated structures for more accurate analysis. The images shown in the figure 3a) and b) of the manuscript, are from a power monitor placed 50 nm above the structure at 0° and 90° polarization respectively. The source used in this simulation was emitting a plane wave with a bandwidth from 730 nm to 830 nm (centered at 800 nm). A uniform mesh with 1 nm x 1 nm x 1 nm size was used over the region under simulation with 1000 fs simulation time. The dimension of the FDTD simulation area was 5 x 5 x 2 μm 3 (3D simulation) and the mesh accuracy of the simulation was set at 5 ("High accuracy") with "conformal variant 1" for the mesh refinement selection and 0.25 nm minimum mesh step. The boundary conditions were set for the perfect matching layer ("PML") with 12 pml layers in all directions and 0.0001 pml reflection. The substrate on which the bowtie nano-antenna was designed, was a 10 x 10 x 2 μm 3 cuboid and the selected optical material was "Au (Gold)-CRC".</p><p>Two-Photon Photoluminescence (2PPL). Two-photon photoluminescence studies of the bowtie devices were performed using a Zeiss LSM 510 MP laser scanning microscope equipped with a 140 fs Chameleon Ultra excitation laser (Coherent) with 80 MHz repetition rate, tunable from 710 to 980 nm. The high-resolution images and bowtie power studies were collected with an LD Plan-Neofluar 63×/0.75 Korr dry objective lens, and the bowtie nanoantennas were irradiated with 780 nm wavelength. The beam waist (W) at the focuswas calculated using W = 1.22λ/NA, using a numerical aperture (NA) of 0.75 and determined to be approximately 1.3 μm.</p><!><p>Surface enhanced Raman spectroscopy was performed with a Renishaw (Invia) Raman microscope/spectrometer equipped with a fiber-coupled 785 nm diode laser source. Raman spectra were acquired using a 50x objective with 10s exposure time acquisitions. Both mono-and polycrystalline devices were coated with 10 μL of 0.02 M BA in methanol, by drop casting, followed by solvent evaporation. The BA solution was prepared by dissolving 0.0488 g of BA solid powder (Coleman & Bell) in 20 mL of methanol to achieve 0.02 M concentration. Spectra were obtained under 50% incident laser power and compared with the spectra obtained from a silicon reference sample obtained under identical illumination and collection conditions. Repeated acquisition of the SERS spectra obtained from the same illuminated region under these illumination conditions showed no appreciable signal degradation.</p>

ChemRxiv

Surfactant-free Aqueous Fabrication of Macroporous Silicone Monoliths for Flexible Thermal Insulation

sol-gel, silicone, porous materials, superhydrophobic, thermal insulation, surfactant-free Hydrophobic silicone macroporous materials prepared in an aqueous solution by the sol-gel method have been considered for various applications such as separation media, heat insulators, and liquid nitrogen adsorbents. In the conventional preparation process, surfactants are used to suppress phase separation to obtain a uniform bulk material. However, a large amount of solvent and time is required to remove them before drying, which hinders industrial-scale synthesis. By copolymerizing tetra-, tri-, and bifunctional organosilicon alkoxides in an aqueous acetic acidurea solution, flexible macroporous silicone monoliths were successfully obtained. The marshmallow-like monoliths recovered their original shape even after 80 % uniaxial compression and significant bending and water repellency. The thermal conductivity of those materials was ~0.035 W m −1 K −1 and did not increase even under 60 % uniaxial compression. This characteristic 2 property can be used for thermal insulation on surfaces with various shapes and in confined spaces under harsh conditions.

surfactant-free_aqueous_fabrication_of_macroporous_silicone_monoliths_for_flexible_thermal_insulatio

Introduction<!>Experimental<!>Sample TMOS<!>Preparation and characterization of MGs by surfactant-free process<!>Thermal conductivity change in MGs by compressive deformation<!>Conclusions<!>Corresponding Author<!>Notes

<p>Researchers are increasingly investigating the use of porous materials as heat insulators owing to growing interest in global warming and energy issues. Historically, unglazed bricks, felt, and cork have been used as porous insulators, and recently, foamed polymers such as polyurethane, styrene, and phenolic resins, as well as glass wool, are often used. 1 Those heat-insulating materials have different advantages and disadvantages in terms of heat and weather resistance and degradation over time, and they are used differently depending on the target application. As a superinsulating material that is far superior to existing ones, aerogels have long attracted the attention of researchers. [2][3][4][5] In recent years, there are some reports of flexible aerogels that overcome the brittleness that has been a problem for decades. [6][7][8][9][10] However, the industrial use of aerogels is still quite limited due to its still poorer handling and much higher production cost than existing insulation materials. Because the market penetration of aerogel is expected to take some time, there is a need to develop insulation materials that are not as good as aerogels but are more efficient to use than existing materials.</p><p>This study investigates polyorganosiloxane (silicone) monolithic macroporous materials that exhibit high thermal insulation with excellent weatherability and processability. Silicones are generally characterized by low thermal conductivity not found in ceramics and chemical stability not found in organic polymers (specifically, they are not degraded by oxygen or water vapor). [11][12][13] Further, they are flexible even when dense, and become more flexible when produced with a 3 porous structure such as foam. By taking advantage of its softness, silicone foam is already widely used in cushioning materials and soundproofing materials. The macroporous silicone materials produced by the sol-gel method have higher application potential than silicone foam. Macroporous silicone monoliths, we have reported and called marshmallow-like gels (MGs), are obtained by copolymerizing organosilicon alkoxides in an aqueous solution sol-gel reaction. 14,15 These materials are fabricated by controlling the phase separation using a surfactant in an aqueous solution system, and have a structure with a framework diameter of several micrometers and a pore diameter of several tens of micrometers. Unlike porous silica (SiO2) and silsesquioxane (RSiO1.5) materials, which also contain siloxane bonds produced by a similar reaction, silicones are highly flexible and resistant to compressing and bending, and therefore do not collapse easily.</p><p>Their chemical properties and sponge-like flexibility have been exploited for realizing a wide range of practical applications such as liquid phase (oil-water) separation, 15 liquid repellency, 16,17 liposome fabrication, 18 sound absorption, 14 thermal insulation, 19 and liquid nitrogen adsorbents 15,19 . However, there are still some problems that need to be improved for mass production and lower cost.</p><p>Although not limited to the MG case, using an aqueous sol-gel system to fabricate hydrophobic macroporous materials is disadvantageous because a surfactant must be used in the starting sol.</p><p>Even though the preparation of gels by aqueous reactions can be easily scaled up while maintaining homogeneity, the need for surfactant removal as a pretreatment for drying remains a significant problem. The larger the volume of the desired monolithic xerogel, the longer it takes to remove the surfactant owing to slower liquid diffusion inside the macroporous material. Cationic surfactants, which are also used for sterilization, are known to be cytotoxic. The generation of waste liquids containing large amounts of surfactants is undesirable because it harms the environment. Organic solvents could be used instead of an aqueous solution for the reaction; however, this would increase process costs. For realizing widespread use, therefore, completing the chemical reaction in a simple aqueous solution system remains desirable.</p><p>Previous studies have reported methods to prepare macroporous silicone monoliths without using surfactants such as adding nanomaterials and copolymerizing alkoxides containing ionic groups. 20,21 However, none of these methods could produce a material with high flexibility like MGs. In this light, I propose a method for preparing flexible porous silicones in an entirely aqueous solution system by adding hydrophilic tetrafunctional alkoxides. Alkoxides with different organic groups are known to have different hydrolysis rates and undergo different polycondensation reactions. 22 The goal of this study is to find the appropriate conditions under which multiple alkoxides can react uniformly, and to prepare samples of the order of several hundred milliliters.</p><p>The microstructure and mechanical properties of the obtained samples are investigated, and their thermal conductivity is measured to evaluate thermal insulation properties.</p><!><p>Materials. The silicon alkoxides tetramethoxysilane (TMOS, Si(OCH3)4), methyltrimethoxysilane (MTMS, CH3Si(OCH3)3), and dimethyldimethoxysilane (DMDMS, (CH3)2Si(OCH3)2), and the cationic surfactant n-hexadecyltrimethylammonium chloride (CTAC), were purchased from Tokyo Chemical Industry Co., Ltd. (Japan). Acetic acid, urea, and methanol were purchased from Kanto Chemical co., Inc. (Japan). All reagents were used as received.</p><p>Sample preparation. On a 25 mL scale, x mL of TMOS, y mL of MTMS, and z mL of DMDMS were added to 15 mL of 5 mM aqueous acetic acid solution containing 5.0 g of urea. The mixture was stirred for 15 min to hydrolyze the alkoxides. After the sol became homogeneous, it was transferred to an airtight container and allowed to stand at 80 °C for 24 h for gelation (within 1-3 h) and aging. The obtained wet gels were washed by immersion in water and underwent solvent exchange to methanol, following which they were subjected to evaporative drying. The resulting gel was named MGx-y-z. Figure 1 and Table 1 show the flowchart of the fabrication process and the combinations of alkoxides tested in this paper.</p><p>As a reference, 14 a gel with the surfactant CTAC was also prepared. In this case, 1.0 g of CTAC was added to a starting composition of MG0-3.0-2.0, and methanol was used for all washes; this sample is called MGref hereafter.</p><p>Characterization. The bulk densities were calculated based on the respective measured weights and volumes with an error margin of approximately 5 %. The microstructures were observed using a scanning electron microscope (SEM; TM3000, Hitachi High-Technologies Corp., Japan). The thermal conductivity was measured using a heat flow meter (HFM 446 Lambda Small, Netzsch GmbH, Germany) for sample panels with 110 mm × 110 mm × 10 mm. The temperatures of the top and bottom heat plates were set at 25 °C and 15 °C, respectively, and the thermal conductivity at an average temperature of 20 °C was measured. For highly flexible samples, the thermal conductivity was measured while compressing and deforming the sample from a thickness of approximately 20 mm to 5 mm within the original weight range of the device. Uniaxial compression tests were performed using a universal/tensile tester (EZ-SX, Shimadzu Corp., Japan) and a 500 N pressure gauge. For measurements, the sample was cut into a 15 mm × 15 mm × 8 mm rectangle piece. Young's modulus was calculated from stress changes under compressive strains of 5.0-10.0 %. The water droplet contact angle was calculated by capturing images using a self-made device fabricated using Raspberry Pi 4 and Camera Module V2. The images were analyzed using the Image J plug-in Contact Angle. [23][24][25] The photographs required for the calculations were taken by dropping 10 μL of water on a smooth cut surface of the samples. Fourier transform infrared (FTIR) spectra were recorded using IRSpirit-L (Shimadzu Corp., Japan) with an attenuated total reflection (ATR) attachment QATR-S. A total of 100 scans of samples were recorded at a resolution of 4 cm −1 . A moisture meter MOC63u (Shimadzu Corp., Japan) was used to measure the moisture content. Each sample was heated at 110°C for 1 h. Thermogravimetricdifferential thermal analysis (TG-DTA) was performed using Thermo Plus EVO2 TG8122 (Rigaku Corp., Japan) at a heating rate of 10 °C min −1 with air at a rate of 100 mL min −1 .</p><p>Table 1. Molar ratio of each alkoxide in the starting composition and physical properties of the obtained MG samples.</p><!><p>MG0.5-3.5-0.</p><!><p>In previous studies, MGs were prepared by adding the tri-and bifunctional silicon alkoxides MTMS and DMDMS, respectively, to an aqueous solution with a surfactant ratio of approximately 3:2 and then hydrolyzing and copolymerizing them in a two-step acid-base reaction to obtain a uniform gel. 14 When the same reaction was conducted without the surfactant, the phase separation of the organosiloxane oligomer, which became more hydrophobic with polymerization, could not be suppressed, and a bulk gel could not be obtained. If TMOS, which has no hydrophobic methyl group, was added instead of the surfactant to make the oligomer hydrophilic, phase separation would less likely occur before gelation. However, because the tetrafunctional silicon alkoxide uses all bonds to form a network, the resulting gel tends to be hard and brittle. Therefore, a range of microstructures and mechanical properties similar to those of MGs were investigated by finely varying the composition ratio of the three silicon alkoxides used for copolymerization.</p><p>The efficient hydrolysis and polycondensation of silicon alkoxides to form a three-dimensional network are realized commonly through a two-step acid-base reaction. 22,[26][27][28][29] The same reaction was used in this study; however, its conditions had to be optimized. The precursors TMOS, MTMS, and DMDMS have different numbers of methyl groups bonded covalently to silicon, resulting in different hydrolysis and polycondensation rates of the alkoxy groups. The difference of reaction rates must be minimized to form a uniform organosiloxane network using the three alkoxides. In our previous study, gels were obtained relatively easily using dilute acetic acid/ammonia water as the acid/base in the system with a surfactant. 30 However, the reproducibility of the method with TMOS was low when the fabrication scale was increased. Through various adjustments, the reproducibility was secured independently of the scale by increasing the temperature to 80 °C rapidly after the hydrolysis of the precursor using the acid for using ammonia derived from the hydrolysis of urea as a base.</p><p>Figure 1 and Table 1 show the compositions of silicon alkoxides and their physical properties.</p><p>These alkoxides were prepared as homogeneous gels with a viscoelastic phase separation structure 31 similar to MGref and flexibility to return to the original shape after 80 % uniaxial compression. All these compositions could be easily scaled up to more than 100 mm × 100 mm × 10 mm (100 mL), which is the sample size required for thermal conductivity measurements with low error. These flexible panels can be bent and will not tear when wrapped around a pipe with a diameter of 10 mm, for example. However, because the samples deformed under their own weight, it was not possible to ensure reproducibility in the three-point bending measurement. The Young's modulus of MGs was higher with increasing the percentage of tetrafunctional alkoxide TMOS in 9 the coprecursor and lower with bifunctional alkoxide DMDMS. Although flexibility is a characteristic property of MG, the samples with low Young's modulus showed poor handling before evaporative drying, because they contained liquid inside that can reach many times their weight. Highly flexible samples with a thickness of several centimeters (e.g., MG1.0-2.0-2.0) sometimes collapsed under their weight if not immersed in liquid.</p><p>Here, the physical properties of an MG1.0-2.5-1.5 sample are described in detail as a representative composition. Figure 2a shows a photograph of the resulting 500 mL sample. In previous studies, MGref samples of several hundred milliliters required immersion in alcohol for several days to remove the surfactant within. This is because the hydrophobic part (alkyl group) of the surfactant tends to stick to the hydrophobic silicone surface and diffuses slowly. If the rinsing process were incomplete, the dried sample would shrink or not exhibit its original properties. By contrast, for the MG1.0-2.5-1.5 sample, the acetic acid and urea used in the reaction were immediately washed away by immersing in warm water; this significantly reduced the time required for drying. Although the precursor compositions were different, MG1.0-2.5-1.5 had a viscoelastic phase separation structure similar to that of MGref (Figures 2b and S1) and showed high flexibility against compression and bending. Despite the addition of TMOS to make the siloxane oligomers hydrophilic during the reaction, the MG1.0-2.5-1.5 cut surfaces all showed high water repellency with a water drop contact angle of 151.7° (Figure 2c). At the same time, because of its lipophilicity, the bulk of MG1.0-2.5-1.5 was able to separate oil (organic solvent) from water in the same way as previously reported using MGref (Figure 2d and Movie S1). The maximum amount of heptane absorbed was ~6.4 times its weight, and the MG could be reused by squeezing out the absorbed liquid. However, FTIR measurement results showed unreacted silanol (Si-OH), which has hydrophilicity, at approximately 900 cm −1 in all samples prepared using 10 TMOS (Figures 3a and S2). 28 Although it is difficult to examine precisely, the silanol on the surface of the microstructure is assumed to have decreased during aging, and some remains only inside the skeleton. In the moisture meter measurement, MG1.0-2.5-1.5 released only about 1.8% of its weight of water, also suggesting that the material has few surface hydrophilic groups. Heat treatment was applied to induce reactions between the residual silanol groups to form a more stable siloxane framework. 32 Thermogravimetric-differential thermal analysis (TG-DTA) measurements were performed to determine the heat treatment temperature. A slight dehydration reaction was observed at 200 °C, followed by the oxidation of methyl groups at approximately 360 °C (Figure 3b). 28,33 From this result, the bulk sample was heat-treated at 300°C for 2 hours, and then FTIR measurement was performed. In the obtained spectrum, the Si-OH peak (~900 cm −1 ) was no longer seen. However, SEM observations of the microstructure before and after heat treatment did not reveal any noticeable change, and no loss of flexibility in compression or bending occurred. Since the decrease in silanol groups did not affect the physical properties much, it can be said that heat treatment is not essential. To investigate the heat resistance, the heat treatment at 300 °C was extended to 24 h; however, the results remained unchanged. By contrast, with heat treatment at 350 °C, the methyl groups of MG1.0-2.5-1.0 were oxidized gradually, resulting in the loss of flexibility and brittleness after 24 h.</p><!><p>Silicone is a polymer with low thermal conductivity, and macroporous silicone monolith MGs exhibit high thermal insulation properties. All MGs produced in this study have low thermal conductivities of 0.032-0.036 W m −1 K −1 ; these conductivities are comparable to those of commercially available high-performance thermal insulators. Even if the thermal conductivity is at the same level, silicones have much higher thermal and chemical stability than ordinary organic polymers and glass fibers. They do not degrade over long periods, even in environments with rapid temperature changes or high humidity. Unlike polymer foams, MG does not use an enclosed gas, so there is no reduction in thermal conductivity due to gas exchange. As a flexible silicone material, MGs can be expected to be applied as an excellent insulator that flexibly fits objects with complex shapes even in harsh environments. To investigate the change in thermal conductivity during deformation, measurements were performed while the MG was compressed uniaxially. Owing to the limited range of pressures that the thermal conductivity measurement system can apply to the sample, a panel sample with much higher flexibility was prepared by increasing the amount of acetic acid-urea solution to alkoxides by a factor of 1.67 in the starting composition of MG1.0-2.5-1.5 (Figures 4a, S3 and Table S1). When the compression of the sample started, the thermal conductivity decreased for a while (Figure 4b). This is because the pore diameter reduced due to 13 compression, thereby suppressing the heat momentum exchange of the gas inside the pores and reducing the thermal conductivity of the gas phase. [34][35][36] For a compressive strain above 40 %, the thermal conductivity increased with increased compression because the effect of the increase in bulk density on the thermal conductivity of the solid phase became more considerable than that of the decrease in the thermal conductivity of the gas phase. For a compressive strain of approximately 60 %, the thermal conductivity did not become higher than that of the uncompressed sample. The unique feature of MGs is that the monoliths do not its heat insulation properties and mechanical properties due to repeated deformation (Figure S4). This is expected to lead to special heat insulation applications, such as filling in small spaces where maintenance is complex.</p><!><p>Macroporous silicon monoliths were prepared by the reaction of tetra-, tri-, and bifunctional silicon alkoxides, TMOS, MTMS, and DMDMS, respectively, as co-precursor in dilute aqueous solutions of acetic acid and urea. The obtained macroporous monolithic materials were sufficiently flexible to recover their original shape even after 80 % uniaxial compression and bending. Their cut surface showed high water repellency with a water droplet contact angle above 150°. When the obtained samples were heat-treated at temperatures above 200 °C, the unreacted silanol groups created siloxane bonds, making the material stable even at 300 °C for 24 h. The materials kept a low thermal conductivity of approximately 0.035 W m −1 K −1 , even when they were significantly deformed. The MGs produced in this report are expected to find thermal insulation applications under particular conditions. The MGs fabricated by the new process have almost the same physical properties as the silicone materials we have reported before and may be applied in various ways such as liquid-liquid separation, liquid nitrogen adsorption, and liposome preparation tools. The environmentally friendly fabrication process, which does not require surfactants or organic solvents, has been confirmed to be reproducible on a scale of several liters. In the near future, we plan to research the production and use of MGs on an industrial scale.</p><p>ASSOCIATED CONTENT Supporting Information. The following files are available free of charge.</p><p>SEM images, FTIR spectra, stress-strain curves, and other physical properties of MGs (PDF)</p><p>Removal of heptane from water using an MG (MP4)</p><!><p>*Email: gen@aerogel.jp</p><!><p>The author declares no conflicts of interest.</p>

ChemRxiv

Efficient synthesis of enantiopure amines from alcohols using resting E. coli cells and ammonia\xe2\x80\xa0\n

\xce\xb1-Chiral amines are pivotal building blocks for chemical manufacturing. Stereoselective amination of alcohols is receiving increased interest due to its higher atom-efficiency and overall improved environmental footprint compared with other chemocatalytic and biocatalytic methods. We previously developed a hydrogen-borrowing amination by combining an alcohol dehydrogenase (ADH) with an amine dehydrogenase (AmDH) in vitro. Herein, we implemented the ADH-AmDH bioamination in resting Escherichia coli cells for the first time. Different genetic constructs were created and tested in order to obtain balanced expression levels of the dehydrogenase enzymes in E. coli. Using the optimized constructs, the influence of several parameters towards the productivity of the system were investigated such as the intracellular NAD+/NADH redox balance, the cell loading, the survival rate of recombinant E. coli cells, the possible toxicity of the components of the reaction at different concentrations and the influence of different substrates and cosolvents. In particular, the cofactor redox-balance for the bioamination was maintained by the addition of moderate and precise amounts of glucose. Higher concentrations of certain amine products resulted in toxicity and cell death, which could be alleviated by the addition of a co-solvent. Notably, amine formation was consistent using several independently grown E. coli batches. The optimized E. coli/ADH-AmDH strains produced enantiopure amines from the alcohols with up to 80% conversion and a molar productivity up to 15 mM. Practical applicability was demonstrated in a gram-scale biotransformation. In summary, the present E. coli-ADH-AmDH system represents an important advancement towards the development of \xe2\x80\x98green\xe2\x80\x99, efficient and selective biocatalytic processes for the amination of alcohols.

efficient_synthesis_of_enantiopure_amines_from_alcohols_using_resting_e._coli_cells_and_ammonia\xe2\

Introduction<!>Co-expression of ADHs and AmDH in E. coli cells<!>Influence of intracellular NAD+/NADH redox balance, cells loading and initial substrate concentration on the productivity of the bioaminations of (S)-2-hexanol using E. coli (Ch1-AA)<!>Survival rate, productivity and toxicity assays for E. coli (Ch1-AA): influence of types of substrates, intermediates, products and their concentrations<!>Influence of product toxicity and cellular redox balance on maximum productivity<!>Influence of co-solvents on amine product toxicity<!>Bioamination of racemic alcohols using E. coli (Ch1-AA-LBv)<!>Analysis of molar productivities and potential current limitations of the \xe2\x80\x98resting E. coli cells-alcohol bioamination\xe2\x80\x99s system\xe2\x80\x99<!>Scale-up of bioamination with resting E. coli cells<!>Conclusions<!>General procedure for engineering recombinant E. coli strains<!>General procedure for culturing E. coli strains<!>General optimized procedure for bioamination of alcohols on analytical scale<!>Derivatization of samples<!>Minimal inhibitory concentration assays<!>General optimized procedure for bioamination of alcohols in biphasic aqueous\xe2\x80\x93organic media on analytical scale<!>Preparative biotransformation of racemic 2-hexanol<!>Supplementary Material

<p>Many fine chemical and pharmaceutical products as well as intermediates either consist of, or are produced from α-chiral amines.1–3 Indeed, α-chiral amines comprise approximately 40% of the optically active drugs that are currently commercialized mainly as single enantiomers. These amines are typically synthesized industrially starting from ketones through multistep processes involving the hydrogenation of an activated intermediate such as an enamide, enamine or pre-formed N-substituted imine.1,2 Such chemical routes are lengthy and atom-inefficient, require the use of an expensive and unsustainable transition metal complex as catalyst (i.e., for stereoselective hydrogenation) and often result in amine products with insufficient chemical and/or optical purity. Therefore, the efficient and sustainable synthesis of enantiomerically pure amines is of critical importance. A high atom-efficient alternative is the recently developed direct conversion of alcohols into amines through a hydrogen-borrowing mechanism using either chemocatalytic or biocatalytic methods.4–15 Compared with organometallic catalysis, biocatalytic processes for amination of alcohols offer additional advantages such as elevated stereoselectivity, maximized atom-efficiency, the use of nontoxic and intrinsically biodegradable catalysts, a requirement of mild reaction conditions (neutral pH, ambient temperature, atmospheric pressure, etc.) and overall reduction of generated waste.16 </p><p>In this context, we recently developed a biocatalytic stereoselective hydrogen-borrowing cascade for the amination of alcohols via the combination of two enzymes in vitro, namely an alcohol dehydrogenase (ADH) and an amine dehydrogenase (AmDH).11 The former enzyme performs the oxidation of the alcohol to the carbonyl intermediate, whereas the latter performs the subsequent reductive amination. Another research group later presented the same concept using alternative dehydrogenases.12 In this dual-enzyme cascade (Fig. 1A), the coupling of the redox reactions enables an efficient internal recycling of the nicotinamide coenzyme (NAD+/NADH), thereby only requiring ammonia and catalytic quantities of NAD+ coenzyme. This biocatalytic process was applied for the amination of primary and secondary racemic alcohols using isolated enzymes in solution or in immobilized form, thus leading to excellent conversions, chemoselectivities and stereoselectivities.11,12,17,18 Although the hydrogen-borrowing cascade for the amination of alcohols performs efficiently by pairing purified ADH(s) and AmDH in the presence of ca. 2-5 mol% (related to the substrate) of NAD+, the costs and time associated with protein purification and the external supplementation of NAD+ might represent limitations for certain types of large-scale applications (i.e., depending on the product value). For a wider applicability of the bioamination of alcohols, another possibility is the use of whole-cell systems, which provide the advantage of a direct applicability by removing the need for enzyme purification and supplementation of coenzyme while also increasing the stability of enzymes due to the cellular environment.19,20 However, mass transfer limitations of compounds over the cell wall and membrane, possible toxicity of compounds and competition with endogenous metabolic pathways of the host are potential drawbacks to this method.21 In this study, we investigated the applicability of resting cells by co-expressing two or three dehydrogenases for enabling the conversion of alcohols into enantiopure amines (Fig. 1).</p><!><p>The following dehydrogenases were selected to enable the bioamination cascade starting from secondary (enantiopure or racemic) alcohols: a Prelog alcohol dehydrogenase (ADH) from Aromatoleum aromaticum (AA-ADH);22 an anti-Prelog ADH variant from Lactobacillus brevis (LBv-ADH)23 and a chimeric amine dehydrogenase (Ch1-AmDH).24 AA-ADH and Ch1-AmDH were applied to investigate the bioamination of (S)-configured alcohols. The optimization of expression conditions demonstrated that Ch1-AmDH is efficiently expressed in E. coli only when the gene bears either a His-tag or a GST-tag at its N-terminus, which was not the case for AA-ADH. Further tests of co-expression of AA-ADH and Ch1-AmDH using a Duet plasmid showed that equal mass production of both dehydrogenases can be achieved when the Ch1-AmDH gene precedes the AA-ADH gene in the construct (Strain 1, ESI section 2.2†). Accordingly, this strain—termed E. coli (Ch1-AA)—was used for the continuation of the work. Since the bioamination of racemic alcohols was investigated by applying AA-ADH, LBv-ADH and Ch1-AmDH, Strains 3 and 4 were created by including an additional plasmid for the expression of LBv-ADH into Strain 1 (ESI section 2.2†). Preliminary experiments (not shown) demonstrated that LBv-ADH is efficiently expressed when its gene bears either a His-tag or a GST-tag; however, approximately equal co-expression in mass of AA-ADH, LBv-ADH and Ch1-AmDH was obtained when GST-LBv-ADH was used (Strain 3; ESI section 2.2†). Notably, the selected two- or three-enzyme co-expression systems showed consistent expression levels between various E. coli batches, thus indicating that the plasmids were stably incorporated into the cells. Fig. 1C illustrates typical examples of the expression levels for Strains 1 and 3.</p><p>The bioamination of alcohols using purified ADH and AmDH in vitro required a higher molar concentration of the latter enzyme in order to achieve elevated conversions.11,18,25 Although expression levels in vivo might theoretically match the optimal molar ratio found for in vitro experiments, we illustrate in this work that many additional dynamics affect the productivity of the bioamination using engineered E. coli resting cells. Since altering the relative expression levels of ADH and AmDH would also impact the effect of these other factors, further tuning of the expression levels was not considered necessary at this stage.</p><!><p>Initially, we tested various reaction conditions for the bioamination of (S)-2-hexanol ((S)-2a, 20 mM) catalyzed by E. coli (Ch1-AA, 70 mg mL−1 cww) in NH4Cl/NH3 buffer (1 M, pH 8.7; ESI Fig. S1†). The reactions were carried out: (i) with resting or lyophilized cells; (ii) in the presence or absence of externally added nicotinamide cofactor; (iii) in the presence or absence of glucose as additive. Resting and lyophilized E. coli cells performed equally in the absence of externally supplemented NAD+ (ca. 5% conversion to (S)-2-aminohexane, (S)-2c); however, accumulation of a large fraction of ketone intermediate 2b was observed only for the bioamination with lyophilized cells (ca. 80%). Supplementing resting cells or lyophilized cells with NAD+ (1 mM) led to improved conversion into (S)-2c only using lyophilized cells (ca. 30%). This difference compared with the use of resting cells must be attributed to the impermeability of the cell membrane in resting E. coli cells to NAD+,26 whereas lyophilization is known to affect the integrity of the cell membrane. Since most of the intracellular cofactors are bound to enzymes,27,28 it is possible that cofactor shuttling from ADH to AmDH and vice versa is less efficient in the bacterial cytosol than in vitro using isolated enzymes in solution. Depending on environmental factors such as growth medium or aeration,29,30 the oxidative form of the nicotinamide cofactor is predominant in E. coli cells with a NAD+/NADH ratio ranging from 8 to 43.31–34 Therefore, we speculated that the alcohol bioamination could be limited by unbalanced redox equilibrium in the cell due to a higher proportion of oxidized cofactor under physiological conditions. The addition of glucose is a cost-effective means of balancing the NAD+/NADH in resting cells during the alcohol bioamination, as aerobic catabolism of glucose within the cells leads to net and gradual production of NADH,35 which Ch1-AmDH can utilize for the reductive amination step of the cascade reaction. Indeed, the supplementation of glucose (20 mg mL−1, 111 mM) to resting cells increased the conversion to amine up to 80% (ESI Fig. S1†), which is comparable to the level achieved in the in vitro cascade.18 Conversely, the effect of supplementing glucose to the reaction catalyzed by lyophilized cells was mediocre (ca. 10% of (S)-2c), thus demonstrating that in principle, intact metabolism is required for efficient NADH regeneration.</p><p>The above-described results were obtained using a substrate : glucose molar ratio of ca. 1 : 5.5. Fig. 2A illustrates an extensive study on the influence of conversion of (S)-2a (20 mM) to (S)-2c versus varied concentration of glucose (up to 10 eq.). Above a threshold of approximately 12 mM glucose (equal to 1: 0.6 substrate : glucose, molar ratio), additional glucose did not lead to further increase of conversion. Therefore, we henceforth used a 1 : 1 molar ratio of substrate : glucose (unless otherwise specified) in order to ensure that sufficient NADH was generated for sustaining the conversion of alcohol to amine. Interestingly, a set of experiments performed at varied but equimolar substrate and glucose concentrations demonstrated that a minimum concentration of glucose (ca. 10 mM) is required in any case for enabling efficient conversion of substrate (ESI Fig. S5†). These data indicated the existence of a certain threshold concentration of glucose that is consumed during aerobic catabolism for cell survival.</p><p>Assessing the influence of cell concentration on the conversion of (S)-2a (20 mM) demonstrated that amine production reached a maximum at an E. coli cell concentration of 50 mg mL−1 cww, remaining statistically constant above this value (Fig. 2B). Indeed, the statistical variation (i.e., standard deviation) of the conversion values for the experiments was significantly large at cell concentrations above 70 mg mL−1 cww, which we attributed to the increased viscosity of the samples resulting in less homogenous mixing. In further experiments, the cww was fixed at 60 mg mL−1 for optimal conversion.</p><p>Reproducibility of biotransformations is a particular concern when using resting cells versus isolated enzymes. A wide range of side reactions can potentially occur in a cell, which could limit substrate conversion. Additionally, work-up procedures can become more complicated, as the system contains multiple components (e.g., cell membranes, DNA, other proteins and metabolites) that could interfere with—for example—quantitative extraction of products with an organic solvent. Moreover, analytical determination of yield using an internal standard can become difficult due to the viscosity and heterogeneity of the reaction medium, which complicate extraction procedures. Therefore, in this work, we also investigated the efficiency of extraction procedures when using resting E. coli cells and validated that all the components of the reaction mixture (substrates, intermediates and products) can be extracted quantitatively with the optimized procedure (ESI† section 3.6). Another cause of reproducibility issues when using resting cells for biotransformation is batch-to-batch differences among E. coli cultures, particularly variations of protein expression levels. To demonstrate that our system is robust in this sense, we performed replicated experiments for the bioamination of (S)-2a using different batches of independently grown E. coli (Ch1-AA) cells (60 mg mL−1 cww, 20 mM substrate and 20 mM glucose). Fig. 2C depicts a plot of the average conversions per set of experiments with the related standard deviations. Notably, the average conversion into (R)-2c for several independent experiments ranges between 60–75%, thus confirming the consistency and robustness of our system.</p><p>Subsequently, we investigated whether higher substrate concentrations (up to 50 mM) could yield an increased absolute product formation. The substrate : glucose ratio was maintained at 1 : 1 to ensure that glucose would not become limiting. Whereas conversion decreased progressively in percentage with the increase of the initial concentration of (S)-2a (ESI Fig. S4†), the absolute amount of (R)-2c formed was stable at approximately 15 mM for the biotransformations at initial substrate concentrations of 20, 30 and 40 mM (Fig. 2D). Conversely, the absolute amount of (R)-2c formed decreased substantially for reactions conducted at substrate concentration below 20 mM and above 40 mM. Notably, besides a maximum productivity at initial ca. 40 mM substrate concentration, the standard deviations of conversion values (i.e., error bars) also increased substantially in the case of reactions at and above 40 mM substrate concentration, which could indicate statistical effects on the cell population during the reaction. Such effects can signify either differences in cell survival and/or cofactor availability/recycling and/or stability of the expressed proteins.</p><!><p>To test whether the reaction suffers from environmental effects, we monitored the conversion over time at varied substrate concentrations and measured the survival rates of the cells in the reaction at each time point. The conversions for the bioamination at 20, 40 and 50 mM of (S)-2a over time are plotted in Fig. 3A. The curve at 20 mM is typical for a successful biotransformation and reached a plateau after 16 h. For all three concentrations, the initial profile was similar; however, from 4 h onwards, the statistical fluctuation of the conversion values for the reactions at 40 and 50 mM increased significantly. Furthermore, after 7 h, curves for the reactions at 20 mM, 40 and 50 mM substrate concentrations started also to diverge. Although the 40 and 50 mM reaction traces showed large differences in conversion for each time point after 2 h, it is noteworthy that the maximum amount of amine formed never consistently exceeded the boundary of 12.5–15 mM established by the 20 mM reaction trace.</p><p>The survival rate of E. coli (Ch1-AA) in the reaction is plotted in Fig. 3B as measures of colony-forming units per mg of cells (CFU mgcells –1). CFU's are a measure of the number of E. coli cells that survive after being subjected to a certain condition.36 Fig. 3B clarifies that resting cells incubated in a reaction at a 20 mM substrate concentration survive longer and with higher population density than in reactions at 40 or 50 mM substrate concentrations. In fact, reactions with a 50 mM substrate concentration showed a large decrease in cell survival already after 2 h, and both reactions at 40 and 50 mM substrate concentrations exhibited almost no survival after 7 h. In contrast, cells incubated in reactions at a 20 mM substrate concentration still had significant CFU numbers after 16 h. For 20 mM reactions, cell death occurred between 16 and 24 h.</p><p>Notably, a correlation was observed between the rapid decrease in survival of E. coli cells (Fig. 3B) and the sharp increase of the standard deviation's value for conversions in the time range of 2–24 h for bioamination reactions performed at 40 and 50 mM substrate concentrations (Fig. 3A). Indeed, combined with the lower survival rates observed in Fig. 3B and the observation of maximum amine production in Fig. 2C, the large statistical variation at higher substrate concentrations observed in Fig. 3A suggests that the produced amine must be toxic to E. coli cells at a specific concentration.</p><p>To expand the substrate scope of the alcohol bioamination in vivo using resting cells (60 mg mL−1, cww), we tested other substrates (20 mM) that were previously studied in the in vitro cascade.11,25 Interestingly, each of these substrates exhibited a different conversion pattern for bioamination in vivo (Fig. 3C). In contrast to in vitro alcohol bioamination, only (S)-2a showed the expected conversion of approximately 75% among the aliphatic substrates (S)-1–4a, whereas the other substrates displayed lower conversions to the amine product (<25%) and partial accumulation of ketone intermediate. In the case of the bioamination of aromatic compound ((S)-5a), the conversion was closer to that for the bioamination of (S)-2a, although the larger standard deviation indicated that (S)-5a is not as easily and consistently converted as (S)-2a. Nevertheless, the stereoselectivities for all tested reactions were perfect (ee >99%, R), and thus identical to the ees obtained for in vitro systems using AA-ADH and Ch1AmDH.11 </p><p>As mentioned above, large statistical variations of conversions might already indicate toxicity of compounds in the reaction mixture. Moreover, the data on the decrease of E. coli survival at higher concentrations of (S)-2a (Fig. 3B) and the discrepancy in the conversion of chemically similar aliphatic substrates—which showed otherwise similar conversions using purified enzymes in vitro—suggested that either the substrates, intermediates or products were toxic to the cells. Therefore, we thoroughly investigated the probable toxicity of the substrates and/or intermediates and/or products through a minimal inhibitory concentration (MIC) assay, whereby E. coli cells were grown in the presence of varied concentrations of these compounds. The lowest concentration for which no visible growth can be established is defined as the MIC. Fig. 3D shows that most alcohols and all ketones tested did not influence visible E. coli growth up to 50 mM, whereas the amines displayed toxicity at moderate or even low concentrations (e.g., already above 1.25 mM for 4c). Among the tested alcohols, only (S)-5a was found to be toxic. Compounds 5b and 5c could not be tested, as they were unfortunately unattainable in sufficient amounts due to purchase restrictions imposed by drug laws.</p><p>To eliminate any possible leaky expression of ADH and/or AmDH and consequent conversion of compounds which would alter their actual concentration, the E. coli strain used in the initial toxicity assays contained no plasmid. However, both E. coli devoid of exogenous plasmid and E. coli (Ch1-AA) were then tested with several compounds (Fig. S6†). As expected, E. coli (Ch1-AA) exhibited higher resistance to (S)-5a, which is likely due to the partial conversion of (S)-5a to 5b; however, its resistance to 2c was lower than that of E. coli devoid of plasmid. This difference might be due to increased pressure on the E. coli cells to maintain the plasmid and concurrent leaky expression of the genes on the plasmid. Notably, as depicted in Fig. 2D and 3A, the MIC of 15 mM for 2c for E. coli (Ch1-AA) correlates nicely with the observation that the formation of (R)-2c does not exceed ca. 15 mM.</p><p>The observed toxicity of the produced amines generally explains the low product titers for some substrates and the large statistical variations between batches observed when operating at substrate concentrations around and above a certain critical value. At high substrate concentrations, even low conversions to product can build-up a toxic level of amine, thus resulting in cell death. As the different cell populations can vary in their resistance to toxic amines from batch to batch (and even per sample), this fact would explain the large variations in conversions observed in these critical situations.</p><!><p>The toxicity level of amines was investigated by repeating the alcohol bioamination at lower substrate concentrations while keeping the 20 mM glucose concentration. Fig. 4 depicts conversions at 1, 5 or 10 mM substrate concentrations, respectively. As expected, lowering the substrate concentration below the lowest MIC (e.g., 2.5 mM for 4c) resulted in highly consistent and elevated conversion values (70–80%) for the amination of (S)-2–5a. Only (S)-1a showed a different behavior, as it was converted solely to ketone. The precise reason why this substrate was not converted in the same manner is unclear; in vitro it was converted similarly to the other aliphatic alcohols.11 </p><p>Increasing the substrate concentrations to 5 and 10 mM slightly changed the conversion pattern. As expected from the MIC assays, the production of (R)-4c did not rise above approximately 1 mM (e.g., 10% conversion at 10 mM substrate concentration) because higher concentrations led to cell death. The product with the next-lowest MIC was 3c (10 mM). Accordingly, the conversion of (S)-3a exhibited conversions consistent to 80% (R)-3c for reactions at 5 and 10 mM substrate concentrations, as (R)-3c could not reach toxic levels (which was not the case for reactions at 20 mM of (S)-3a). The conversion of (S)-1a was still the outlier in this set with 10–15% amine conversion (i.e., maximum productivity ca. 1 mM). Comparing Fig. 3D and 4, the effect of build-up of toxic amine is well demonstrated, as the production of amines at concentrations above the MIC values was generally not feasible or exhibited large values of standard deviation (i.e., different cell batches can display different tolerance at amine concentration around or slightly above the MIC values). Indeed, when keeping the production of amine below the MIC, conversions were consistently good to excellent, reaching approximately 80%.</p><p>Another aspect of the cellular metabolism was revealed when performing this experiment with a 1 : 1 molar ratio of substrate : glucose (ESI Fig. S5†). The reactions at 10 mM alcohol and 10 mM glucose (Fig. S5†) exhibited the same conversions as for those at 10 mM alcohol and 20 mM glucose (Fig. 4). Below 10 mM glucose while keeping the 1 : 1 substrate : glucose molar ratio, the resulting alcohol substrates nearly completely converted into ketones plus amines (except for (S)-5a); however, amine conversion was significantly lower at the glucose and substrate concentration of 5 mM and it was not formed at all at the glucose and substrate concentration of 1 mM. In these cases, accumulation of ketone intermediate occurred. As previously stated, the intracellular cofactor is mostly present as NAD+ at physiological conditions, which is then used by the ADH for oxidation of the alcohol. The NADH generated in the first oxidative step can also be involved in other intracellular reductive processes, thus preventing reductive amination of the ketone intermediate by AmDH. Aerobic respiration of glucose increases NADH levels in cells, thus enabling the reductive amination step. The impaired conversion of ketone to amine in presence of only 1 and 5 mM glucose implies that the cellular metabolism consumes a background level of NADH that has to be regenerated in order to sustain the bioamination in vivo.</p><!><p>As product titers for the amination of (S)-3a and (S)-4a were particularly moderate, we investigated the use of co-solvents as a reservoir for the produced toxic amines, thereby enhancing cell survival.37,38 Several aspects must be considered for a proper choice of co-solvents for efficient alleviation of toxicity, such as biocompatibility, Log P value (i.e., partition coefficient of amine within liquid phases) and miscibility with the aqueous phase of the reaction.37,38 Operating with resting (living) cells makes co-solvent biocompatibility a critical factor,38,39 particularly considering the intrinsic relatively low kinetics for the conversion of ketones to amines that necessitates longer cell survival. We tested various co-solvents for cell survival at varied volumetric ratios in the aqueous phase, namely n-heptane, n-decane and n-hexadecane, and the latter yielded the highest survival rates for cells after 4 h (data not shown). Consequently, n-hexadecane was used in further experiments. We then investigated the effect of the substrate concentration and different ratios of aqueous phase: cosolvent towards conversion of (S)-3a and (S)-4a, and the results are shown as absolute amounts of amine formed in Fig. 5A and B, respectively.</p><p>Adding the co-solvent was beneficial for the reaction at a 20 mM concentration of (S)-3a but detrimental for reactions at lower concentrations (1, 5, 10 mM). In fact, if the amine does not reach toxic levels, a decrease in conversion is likely due to side-effects of the co-solvent such as disintegration of cell membrane over time,39,40 which also explains the generally lower conversions obtained in the experiments using a 50% v/v of hexadecane rather than with 5 or 10% v/v. Notably, the maximum amount of produced (R)-4c showed a two-fold improvement when using a co-solvent (from <1 mM to 2 mM). In general, the optimal amount of co-solvent was 10% v/v.</p><!><p>We have previously demonstrated that the amination of racemic alcohols is feasible by simultaneously using two stereocomplementary ADHs,11 which resulted in the creation of two E. coli (Ch1-AA-LBv) strains (Strains 3 and 4; ESI section 2.2†). Preliminary activity tests evidenced that Strain 3 performed the alcohol bioamination more efficiently than Strain 4 (ESI† section 3.1), and unlike Strain 4, Strain 3 showed equal and elevated expression levels for all three enzymes. Hence, this E. coli (Ch1-AA-LBv) cell strain was tested for further conversion of racemic as well as enantiopure 2a and 5a (20 mM; Fig. 6A and B, respectively).</p><p> rac-2a was converted to 40–65% amine by the three-enzyme E. coli system (depending on the batch of E. coli; ESI Fig. S2†). Notably, both alcohol enantiomers of the racemic mixture were converted to a similar extent because the remaining alcohol 2a at the end of the reaction gave a (R): (S) ratio of 52 : 48. Thus, both stereocomplementary ADHs possess similar apparent activity for their respective 2a enantiomer. The conversion rate was also comparable to that observed for the bioamination employing E. coli (Ch1-AA) strain with enantiopure (S)-2a as a starting material (Fig. 3A and 6A, respectively).</p><p>Conversely, conversion of 5a (20 mM) with E. coli (Ch1-AA-LBv) exhibited a different behavior (Fig. 6B). (S)-5a was converted to ca. 40%, which is somewhat lower than the 50% obtained using E. coli (Ch1-AA) and the same substrate (Fig. 3C). Standard deviation values of conversions between samples were also significantly large, as the experimental conditions are at the toxicity limit for this compound (MIC of 5a is 15 mM, Fig. 3D). Interestingly, (R)-5a was converted very poorly to amine (<10%), whereas rac-5a was converted at intermediate level (ca. 20%) between (R)-5a and (S)-5a. As the enantiomeric ratio of the remaining 5a at the end of the reaction yielded a (R): (S) ratio of 42 : 58, the poor conversion of (R)-5a did not stem from catalytic inefficiency of LBv-ADH. Generally, different enantiomers can have different effects and/or toxicity in biological systems.41 Indeed, (S)-5a was less toxic for our E. coli system than (R)-5a, thus explaining the much higher conversion to amine at 20 mM scale when starting from enantiopure (S)-5a, as well as the halved conversion obtained for the amination starting from rac-5a compared to (S)-5a.</p><!><p> Table 1 reports the maximum molar productivity for each substrate tested in this study.</p><p>The amination of alcohol 2a (20 mM) yielded a high molar productivity starting from both enantiopure S-configured and racemic alcohols (Table 1, entries 2 and 6). The reason for the high productivity stems from the relatively low toxicity of all reaction components, including the amine product (R)-2c. The behavior was different in the case of the amination of 5a. Amination of the enantiopure alcohol (S)-5a (20 mM) yielded a molar productivity of (R)-5c of 9.3 mM, whereas the amination of rac-5a (20 mM) produced less than half product concentration (Table 1, entries 5 and 7). We attributed this difference to the higher toxicity of (R)-5a compared to (S)-5a, as also supported by the amination of enantiopure (R)-5a, which afforded less than 2 mM of amine product (Fig. 6B). Finally, the amination of substrate (S)-3a also yielded a remarkable molar productivity above 8 mM (Table 1, entry 3), whereas toxicity was a more limiting factor for the amination of (S)-1a and (S)-4a (entries 1 and 4).</p><p>Broadly, we noticed a difference in the maximum attainable conversion between in vitro and in vivo bioamination, one reason for which was the general toxicity of the amine products to E. coli at certain concentrations. The addition of cosolvent could increase product formation in cases of severe toxicity (e.g., amination of (S)-3a and (S)-4a); however, the cosolvent itself seems to have an impact on survivability of the E. coli strain. The influence of the cellular environment on the availability of NAD+, NADH and NH4 + might be another factor that limits bioamination in vivo. On the one hand, in principle, the correct NAD+/NADH redox balance could be set by exploiting the aerobic catabolism of exogenously added glucose. However, as (S)-2a could be converted to 80% in the entire concentration range of 1–20 mM substrate, it seems that the NAD+/NADH cofactor availability was sufficient. On the other hand, it could be that the intracellular NH3/NH4 + concentration is lower than the 1 M value present in the reaction buffer, as intracellular cations concentrations are regulated in vivo.35 Considering that hydrogen-borrowing amination in vitro at 200 mM of NH3/NH4 + buffer afforded typically 75% conversion11 and that the K M of Ch1-AmDH for NH3 is around 350 ± 133 mM,24 it could be that quantitative conversion is partly limited by the actual intracellular NH3/NH4 + concentration. The pH is another factor that can potentially influence the thermodynamics of the system. In fact, the intracellular environment is normally buffered approximately between pH 7.2 and 7.842 regardless the pH of the reaction buffer, which was set at 8.7 in this study because it was found to be optimal for bioamination with purified enzymes. Other, more subtle factors that could prevent quantitative conversion of substrates are either the unavailability of the alcohol substrate and/or ketone intermediate due to partitioning of these compounds to cell membranes, or insufficient shuttling of the intermediate between the ADH and the AmDH.</p><!><p>To demonstrate the potential applicability of the bioamination with resting E. coli (Ch1-AA-LBv) for larger scale production of enantiopure amines, we performed a preparative biotransformation on rac-2a (511 mg, 5 mmol). In 250 mL cell suspension, (R)-2c was obtained in 40% conversion. Amine recovery from the reaction mixture was only partial due to amine volatility, giving an isolated yield of approximately 16%; however, the product was isolated with perfect chemical (>99%) and optical purity (ee >99% (R)).</p><!><p>A number of cascades for the bioamination of alcohols using isolated enzymes in vitro have been reported during this decade. One approach entailed the combination of an ADH, a ω-transaminase (ωTA) and an alanine dehydrogenase (AlaDH),14,15,43,44 whereas others utilized either an alcohol oxidase (AOx)45,46 or a laccase/TEMPO system47 for the first oxidative step, and always in combination with a ωTA. In particular, the ADH-ωTA-AlaDH system was co-expressed and tested in resting E. coli cells at a 10 mM concentration of alcohol substrates. Although the supplementation of any cofactor was not required, the addition of 2 to 25 equivalents of L-alanine as amine donor (compared to the substrate) was mandatory to attain elevated conversion.48,49 Notably, the addition of L-alanine could be omitted only when an AOx-AlaDH-ωTA module was applied for the amination of styrene diols to yield the related 1–2 amino-alcohols.50 An alternative approach consisted of an orthogonal enzyme network operating in vitro and comprising four oxidoreductases.25 A similar network was recently implemented in a hybrid system, whereby E. coli cells expressing the ADH and an extracellularly added isolated AmDH were combined.51 </p><p>In this work, we demonstrated the viability of the hydrogenborrowing amination cascade using an ADH/AmDH combination in resting E. coli cells, thus representing the simplest and most atom-efficient system for the amination of alcohols in vivo. Most of the tested substrates gave conversions to amine of approximately 80%, depending on the substrate concentration. Further studies will focus on improving the system's toxicity resistance to amines and co-solvents in order to further increase the productivity of the bioamination. Various options are available, such as the use of non-conventional co-solvents52,53 and/or solvent-tolerant bacteria,54 as well as the implementation of a biphasic system with a hollow membrane fiber55 or a constant flow set-up rather than a batch-process. Another challenging and complementary option is the engineering of efflux pumps in E. coli for the selective secretion of the toxic amine products.56–58 Finally, the bioamination reaction could be integrated into longer multistep pathways, whereby the primary amine would become an intermediate rather than the final product; thus, keeping the amine concentration below toxicity levels would prevent cell death and increase the total productivity of the system. In conclusion, the present E. coli-ADH-AmDH system represents an important advancement towards the development of sustainable, efficient and selective biocatalytic processes for the amination of alcohols.</p><!><p>All genetic constructions were implemented using standard molecular biology techniques with Phusion DNA polymerase, FastDigest restriction enzymes and T4 DNA ligase (all from Thermo Scientific). Ch1-AmDH24 and AA-ADH22 were subcloned into a pETDUET plasmid, in the first and second multiple cloning sites, respectively. Ch1-AmDH also contains a N-terminal 6xHis-tag. Transformation of the pETDuet_NHis_Ch1AmDH_AA-ADH plasmid into E. coli resulted in a strain termed E. coli (Ch1-AA). The third protein, LBv-ADH,23 was cloned separately into a pET28 vector with a N-terminal GST-tag. Co-transformation of pETDuet_NHis_Ch1AmDH_AA-ADH and pET28bv_GST_LBv-ADH into E. coli BL21 DE3 resulted in the strain named E. coli (Ch1-AA-LBv).</p><!><p> E. coli BL21 DE3 strains were inoculated in Luria–Bertani (LB) medium with either ampicillin (100 mg L−1) for E. coli (Ch1-AA), or ampicillin (100 mg L−1) and kanamycin (50 mg L−1) for E. coli (Ch1-AA-LBv). Cultures were grown overnight at 37 °C and 170 rpm. The following day, fresh LB medium with the appropriate antibiotic(s) was inoculated with the overnight culture and grown at 37 °C and 170 rpm until reaching an OD600 between 0.6 and 0.8. Protein expression was induced by adding 0.5 mM IPTG and cells were grown overnight at 25 °C and 170 rpm. Cells were harvested at 3400g for 20 minutes.</p><!><p>After harvesting, E. coli cell pellets were washed once with ammonium chloride buffer (1 M NH4Cl/NH3, pH 8.7). Cells were again pelleted by centrifugation at 3400g for 20 minutes and then re-suspended in ammonium chloride buffer. Biotransformations entailed 1 mL cell suspension in ammonium chloride buffer (60 mg mL–1 cells (cww), 20 mM glucose, 20 mM substrate) in 2 mL Eppendorf tubes. Reactions were incubated at 30 °C and 170 rpm for 24 h.</p><p>The reaction was quenched by the addition of KOH (200 μL, 10 M), followed by extraction with EtOAc (2 × 500 μL). The organic layer was dried with MgSO4 and conversion was determined by GC with an Agilent DB-1701 column. Details of the GC analysis and methods are reported in the ESI.† </p><p>Details about deviations from this general procedure for the various experiments can be found in ESI.† All experiments were performed, at least, with independent biological duplicates (two different batches of E. coli), each of which consisted of a technical duplicate (each reaction was performed twice). Therefore, each sample point is averaged from at least four samples.</p><!><p>The enantiomeric excess of the amine product was determined after derivatization to acetamido. Samples were derivatized by adding a solution of 4-dimethylaminopyridine dissolved in acetic anhydride (20 μL of 50 mg mL–1 stock solution) to 500 μL of sample. The samples were shaken in an incubator at RT for 30 minutes, after which water (600 μL) was added and the samples were shaken for an additional 30 minutes. After centrifugation, the organic layer was dried with MgSO4. Enantiomeric excess was determined by GC with a Variant Chiracel DEX-CB column. Details of the GC analysis and methods are reported in the ESI.† </p><!><p> E. coli BL21 DE3 (devoid of plasmid) cells were inoculated in LB medium without antibiotic. Cultures were grown overnight at 37 °C and 170 rpm. The following day, fresh LB medium was inoculated with the overnight culture and grown at 37 °C and 170 rpm until reaching an OD600 between 0.5 and 0.9. Cells were diluted to a titer of 105 cells per mL in LB medium. In a 96-well plate, 150 μL of diluted cells were added to 150 μL of the tested compounds in an appropriate concentration. Tested concentrations were 0, 5, 10, 15, 20, 30, 40 and 50 mM, respectively for all compounds except 4c, which was tested with 0, 0.08, 0.16, 0.31, 0.63, 1.25, 2.5 and 5 mM, respectively. Plates were covered with Easyseal (Greiner Bio-ONE) and incubated for 24 h at 37 °C. Each compound was tested in triplicate per 96-well plate, with at least two 96-well plates per compound (i.e., at least six samples per compound). Growth was determined by visual inspection and MIC was defined as the lowest concentration that prevents a visible growth in all vials. In the case of less reproducible results for a specific compound and/ or experimental condition, the concentration in which at least half of the samples showed no visible growth was taken as the MIC concentration. Assays with E. coli (Ch1-AA) were performed by growing the cells with 50 mg L−1 kanamycin in the LB medium.</p><!><p>Preparation of resting cell suspension was performed as described above. Biotransformations consisted of 1 mL cell suspension in ammonium chloride buffer (60 mg mL−1 cells (cww), 20 mM glucose, 20 mM substrate) and 0.5 mL of n-hexadecane (C16) in 4 mL glass vials. Reactions were incubated at 30 °C and 170 rpm for 24 h. For the various experiments, details about deviations from this general procedure can be found in the ESI.† </p><p>Before extraction, the total volume of co-solvent was adjusted to 500 μL. The reaction was quenched by the addition of KOH (200 μL, 10 M). The co-solvent was removed and the aqueous phase was extracted once with 500 μL EtOAc. The EtOAc and co-solvent fractions were combined and dried with MgSO4, and conversion was determined by GC with an Agilent DB-1701 column. Details of the GC analysis and methods are reported in the ESI.† </p><!><p> E. coli (Ch1-AA-LBv) was cultured (3.2 L) and harvested as described above. The preparative biotransformation consisted of 250 mL cell suspension in ammonium chloride buffer (60 mg mL−1 cells (cww), 20 mM glucose, 20 mM rac-2a (0.511 g)) in a 500 mL baffled flask. The reaction was incubated at 30 °C and 170 rpm for 24 h and monitored by GC by taking 1 mL samples (worked-up and measured as described above). Additionally, an analytical biotransformation was run in parallel with the same E. coli batch as the control experiment.</p><p>The preparative reaction mixture was acidified to pH 2–4 through the addition of a concentrated HCl solution. The reaction was extracted with methyl tert-butyl ether (3 × 60 mL) to remove the unreacted alcohol and ketone intermediate. The pH of the reaction was increased to basic pH through the addition of KOH (10 M) and extraction was performed with methyl tert-butyl ether (3 × 60 mL). The organic fractions containing the amine product were combined and dried with MgSO4. After filtration and evaporation of the solvent, the product was obtained with >99% chemical purity and >99% enantiomeric excess (R).</p><p>The authenticity of the product was confirmed by 1H-NMR (400 MHz, CDCl3, see ESI†).</p><!><p>† El I) available. See DOI: 10.1039/c9gc01059a </p>

PubMed Author Manuscript

Chaperone-like N-methyl Peptide Inhibitors of Polyglutamine Aggregation\xe2\x80\xa0

Polyglutamine expansion in the exon 1 domain of huntingtin leads to aggregation into \xce\xb2-sheet-rich insoluble aggregates associated with Huntington\xe2\x80\x99s Disease. We assessed eight polyglutamine peptides with different permutations of N-methylation of backbone and side chain amides as potential inhibitors of polyglutamine aggregation. Surprisingly, the most effective inhibitor, 5QMe2 (Anth-K-Q-Q(Me2)-Q-Q(Me2)-Q-CONH2, Anth = N-methyl anthranilic acid, Q(Me2) = side chain N-methyl Q) has only side chain methylations at alternate residues, highlighting the importance of side chain interactions in polyglutamine fibrillogenesis. Above a 1:1 stoichiometric ratio, 5QMe2 can completely prevent fibrillation of a synthetic aggregating peptide YAQ12A; it also shows significant inhibition at substoichiometric ratios. Surface plasmon resonance (SPR) measurements show a moderate Kd with very fast kon and koff. Sedimentation equilibrium analytical ultracentrifugation indicates that 5QMe2 is predominantly or entirely monomeric at concentrations up to 1 mM, and that it forms a 1:1 stoichiometric complex with a fibril-forming target, YAQ12A. 5QMe2 inhibits not only nucleation of YAQ12A, but also fibril extension, as shown by the fact that it also inhibits seeded fibril growth where the nucleation steps are bypassed. 5QMe2 acts on its targets only when they are in the PPII-like conformation, but not after they undergo a transition to \xce\xb2-sheets. Thus 5QMe2 does not disassemble pre-formed YAQ12A; this contrasts with our previously described, backbone N-methylated inhibitors of \xce\xb2-amyloid aggregation (16,17). The mode of action of 5QMe2 is reminiscent of chaperones, since it binds and releases its targets very rapidly, and maintains them in a non-aggregation-prone, monomeric state, in this case, the polyproline II (PPII)-like conformation, as shown by CD spectroscopy.

chaperone-like_n-methyl_peptide_inhibitors_of_polyglutamine_aggregation\xe2\x80\xa0

<!>Peptide Synthesis and Purification<!>Dissolving and Disaggregation of YAQ12A; Sedimentation Assays of YAQ12A Aggregation<!>Additional Size Exclusion Chromatography Experiments<!>Inhibition of Seeded YAQ12A Fibril Growth by 5QMe2<!>Electron Microscopy<!>Surface Plasmon Resonance<!>Circular Dichroism<!>Analytical Ultracentrifugation: Experimental Procedure<!>Sedimentation Screening Assays<!>Time Course of Inhibition of YAQ12A Aggregation by 5QMe2 and Electron Microscopy of Reaction Products<!>Effect of Adding Inhibitor at Various Points in the Time Course of Aggregation<!>Inhibition of Seeded YAQ12A Fibril Growth by 5QMe2<!>Inhibitor Concentration Dependency<!>Kinetics of YAQ12A and 5QMe2 Binding, as Assessed by Surface Plasmon Resonance (SPR)<!>AUC and Complex Stoichiometry<!>Secondary Structure of 5QMe2 and Related Peptides<!>Discussion

<p>Expanded polyglutamine (polyQ) tracts are responsible for at least nine neurodegenerative diseases, including Huntington's Disease (HD). HD occurs when the polyQ domain of exon 1 of huntingtin protein expands beyond a threshold of approximately 35 residues (1), leading to polyQ aggregation (2-4). PolyQ proteins can form β-sheet rich fibrils (5, 6). In contrast to many other amyloids (7), polyQ domains have polar side chains, and both these and backbone atoms can form hydrogen bonds. Recent x-ray diffraction and solid-state NMR studies of the glutamine- and asparagine-rich yeast prion proteins Sup35p and Rnq1p show parallel in-register β-sheet segments, possibly separated by non-β-sheet bend structures in longer peptides (8, 9). Prefibrillar oligomers of β-amyloid are micelle-like and show β-sheet character (10). The structure of polyQ oligomers is not known, however; in contrast to oligomers of β-amyloid, the polar nature of poly Q makes it unlikely that they would be micelle-like. Large, prefibrillar "spheroids" of huntingtin also have been observed, and these acquire β-sheet structure as they mature into fibrils (11). We have recently shown that small, soluble oligomers of short polyQ peptides adopt a polyproline-II helix-like structure (12, 18), and thus the conversion to fibrils may require a transition from this structure to β-sheet.</p><p>Inhibitors of protein and peptide aggregation are of value both as molecular probes of the aggregation pathway and as potential therapeutic agents (13-15). We have described peptidic inhibitors of β-amyloid, consisting of an aggregation domain in which alternate residues are N-methylated on backbone amides. Such inhibitors (e.g., Aβ16-20m) both inhibit fibril formation and disassemble preformed fibrils, by disrupting backbone hydrogen bonding (16, 17).</p><p>In this paper, we describe what began as a comparison of N-methylation patterns in potential inhibitors of polyQ peptide aggregation. These studies, however, soon produced several surprising results, starting with the observation that side chain N-methylations alone yield the most effective inhibitors of all the permutations tested on short polyQ peptides. Subsequent experiments highlighted the importance of side chain interactions in polyQ aggregation, and showed that the inhibitors act through a mechanism reminiscent of chaperone proteins. We examined the most effective of these inhibitors in detail, and observed that it binds to a target polyQ peptide and inhibits its aggregation by forming a 1:1 stoichiometric complex. It inhibits not only nucleation, but also fibril extension, as shown by its ability to inhibit seeded fibril growth in which nucleation steps are bypassed. Furthermore, our data indicate that the effective inhibitors adopt a polyproline II conformation, and interact with a polyQ peptide only while it is also in the PPII conformation. These results suggest a more complex scheme for the polyglutamine aggregation pathway than has been appreciated previously.</p><!><p>YAQ12A and 5Q were synthesized using standard 9-fluorenylmethoxycarbonyl (FMOC) chemistry and Rink Amide resin, on an Applied Biosystems Model 431A peptide synthesizer. In early experiments, peptides were cleaved from the resin using 9.5 mL of TFA, 0.25 mL of water and 0.25 mL of triisopropylsilane. We subsequently observed that both yields and initial purity of the mixture from the resin were significantly improved by cleaving peptide from the resin using a mixture or 9 mL TFA, 0.5 mL thioanisole, 0.3 mL ethanediol, and 0.2 mL anisole for 1 h.</p><p>Peptides with backbone N-methyl groups were synthesized manually according to Biron et al. (19). Peptides with side chain methylations were synthesized on tBoc 4-Methylbenzhydrylamine (MBHA) resin at 0.25 mmol scale, either manually for peptides containing backbone N-methyl groups, or using the synthesizer for peptides with no backbone N-methyl groups. Side chain N-methyl Gln residues were prepared by deprotecting Glu(tBu) side chains on resin (TFA:H2O, 95:5, v:v), and then amidating (22°C, overnight) under N2 with 1 mmol benzotriazole-1-yl-oxy-tris-(dimethylamino)-phosphonium hexafluorophosphate (BOP), 2 mL 2.0 M dimethylamine in methanol, 5 mL dimethylformamide (DMF); this procedure was repeated for an additional 4 h. Although BOP is a less efficient coupling agent than HATU (2-(7-Aza-1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate), the latter reagent gave guanylation side products during long incubations. In one peptide, 5MeQ-PEG-5MeQ (see Figure 1A), a PEG moiety, FMOC-8-amino-3,6-dioxaoctanoic acid ("FMOC-mini-PEG", Peptides International) was incorporated into the middle of the peptide. After incorporation of the PEG group, the remaining amino acids were incorporated into the peptide as described above.</p><p>The MBHA resin was washed with DMF, and peptide was cleaved and deprotected using anhydrous HF:p-cresol (10:1, v:v, 0°C). After the peptides were precipitated in ice-cold diethyl ether, they were purified using by C18 preparative HPLC (Zorbax) at 22 °C, either with a water:acetonitrile (both 0.1% TFA, v:v) gradient, or with an isocratic mixture. Peptide purity was ≥ 95% by analytical HPLC. Molecular masses of the peptides were verified with ESI- and MALDI-TOF mass spectrometry. A Lys residue was incorporated into all inhibitor peptides to aid cellular ingress of peptides (for studies not included herein) and possibly to increase solubility. N-terminal N-methyl anthranilic acid was added to allow detection by fluorescence and UV spectroscopy.</p><!><p>Aggregation of YAQ12A was measured as the concentration of monomeric peptide remaining in solution after sedimentation, as described elsewhere (20). The elution of 100 μM YAQ12A in Size Exclusion Chromatography was most consistent with monomeric state. Chromatography was performed using a Superdex Peptide 10/300 GL column (GE Lifesciences) equilibrated with 10 mM sodium phosphate, pH 7.40; the flow rate was 0.5 mL/min (see Figure 2F and G). Apparent molecular weight (MWapp) was estimated on the basis of a calibration curve of molecular weight standards (Supporting Figure 4). Additional size exclusion chromatography experiments are described below.</p><p>Concentration was determined by absorbance at 274.6 nm, using an extinction coefficient (ε) of 1420 M−1cm−1. The solution was diluted to 100 μM with 10 mM sodium phosphate, pH 7.40, and incubated with or without inhibitor peptides for various times at 37 °C without agitation. After incubation, the mixture was centrifuged in an Airfuge ultracentrifuge (Beckman) at 98,000 × g for 1 h at 22 °C. 100 μL of the top third of the resulting solution was injected onto an analytical C18 RP-HPLC column and followed by absorbance at 220 nm. The concentration of YAQ12A injected was calculated from the peak area data compared to a standard curve made from solutions of YAQ12A of known concentrations. In more recent experiments, we were able to shorten the time between analyses by performing isocratic analyses; using a Zorbax C18 column, and an eluent of water:acetonitrile:TFA (87.5:12.5:0.1, v:v:v), YAQ12A elutes at ~ 6 minutes and 5QMe2 elutes at ~ 22 minutes (Supporting Figure 1).</p><p>Stock solutions of inhibitor peptides at ~ 4 mM were made from pure lyophilized peptide and MilliQ purified water, adjusted to 7.00, since pH affects ε of N-methyl anthranilic acid. Concentration of the stock was determined by absorbance using ε337 nm = 2690 M−1cm−1. Aliquots at various concentrations were frozen, lyophilized, and stored at −20 °C. All fibrillation reactions were performed in siliconized microfuge tubes.</p><!><p>As stated above, initial experiments showed that YAQ12A, disaggregated as described above, eluted from a Superdex Peptide column in a position most consistent with a monomeric state. To examine the time course of this chromatographic behavior, and the effect of the aggregation inhibitor peptide 5QMe2 described in Results, we performed the following series of size exclusion chromatography experiments. A 100 μM solution of freshly disaggregated YAQ12A was chromatographed either alone or in the presence of 170 μM 5QMe2. (In other experiments, a 1:1 molar ratio of YAQ12A:5QMe2 was used; see, for example, Supporting Figure 5). To make the mixture of peptides, 5QMe2 was dissolved in water:TFA (100:0.1, v:v) and distributed into siliconized microfuge tubes. Solvent was evaporated under a stream of N2 followed by lyophilization. To the dry film, YAQ12A solutions were added. After 0, 24 and 50 h of incubation of these solutions are 37 °C, a 100 μL aliquot of each solution was injected onto a Superdex Peptide 10/300 GL column. Mobile phase was 10 mM sodium phosphate containing 10 mM NaCl, pH 7.40. Flow rate was 0.5 ml/min. The column effluent was monitored at 220, 230, 274 and 337 nm. Essentially the same chromatographic pattern was observed at each wavelength, except that only 5QMe2 could be observed at 337 nm, where the N-methyl anthranlate moiety absorbs.</p><!><p>To produce fibril seeds, YAQ12A fibrils were made by incubating a 250 μM solution of the peptide, initially dissolved as described above, at 37 °C for at least one week. Immediately prior to the seeding experiment, the seed slurry was sonicated for five minutes using a Bransonic ultrasonic bath (Model: 2510R-MT, Branson Ultrasonics Corp., CT). In preliminary experiments, we assessed the effects of adding various nominal concentrations of YAQ12A fibril seeds to a fresh, 100 μM solution of YAQ12A (see Supporting Figure 2A; Supporting Figure 2B shows that the seeds were indeed fibrillar). Nominal concentration of the seed slurry was defined from the original peptide concentration of the solution from which the seed slurry was made. From these experiments, we determined that a nominal concentration of 8 μM was optimal for this batch of seed fibrils in order to observe both accelerated fibrillation of fresh YAQ12A, and the effects of 5QMe2 on seeded fibril growth.</p><p>For the seeding experiments, we compared YAQ12A fibrillation in the presence or absence of fibril seeds, and in the presence or absence of 5QMe2. YAQ12A was freshly disaggregated, and dissolved in 10 mM phosphate (with 1 μM NaN3), to a concentration of approximately 150 μM. Actual concentration was determined by absorbance at 274.6 nm, and confirmed by injecting an aliquot onto an analytical C18 RP-HPLC column, as described above. Siliconized microfuge tubes were prepared with or without a sufficient of 5QMe2 so that the final concentration of this peptide would be 100 μM. To this mixture, sufficient buffer was added so that the to final volume would be 600 μL and the final YAQ12A concentration would be 100 μM. Lastly, an aliquot of the above fibril seed slurry was added to some samples to give a final nominal seed concentration of 8 μM. Thus, in those samples containing 5QMe2, the molar ratio of YAQ12A:5QMe2 was 1:1. The mixtures were incubated at 37 °C for 0 h, 5 h, 24 h or 50 h. At each time point, fibrillation was assessed by the above sedimentation assay, performed at least in triplicate; results are reported as mean ± S.D.</p><!><p>Electron microscopy was performed as described elsewhere (16); micrographs were recorded at magnifications of 15,000 or 39,000×, plus 1.4× magnification from the CCD camera.</p><!><p>YAQ12A that had been disaggregated and centrifuged as described (20), was immobilized on a CM5 sensor chip using standard amine coupling procedures, as described by the manufacturer's instructions (Pharmacia Biosensor AB). All assays were carried out using a Biacore 3000 instrument at 25 °C, at a flow rate of 20 μL/min, with association and dissociation times of 120 s, in HSP-B buffer (Biacore). Each concentration of analyte was tested in triplicate. Kinetics data were analyzed first using BIA evaluation software (Biacore), and then by non-linear least squares analysis using Kaleidagraph software and using a non-linear least squares fitting program described by Yamaoka et al. (21). Adsorption kinetics were fitted as a bimolecular interaction of the analyte (5QMe2) interacting with the immobilized ligand (YAQ12A), forming the 1:1 complex at the sensor surface. The best fit (see below, Figure 3), was obtained with single exponential equations for association and dissociation. Specifically, during association, the SPR signal varied with time as given by the equation: [Eq. 1]Rt=kon[I]Bmaxkon[I]+koff×(1−e−(kon[I]+koff)t) where Rt is the SPR signal ("response units" or RUs) as a function of time, kon and koff are association and dissociation constants, respectively, Bmax is maximum binding capacity, and t is time in seconds. Similarly, during dissociation the SPR signal varied according to the equation: [Eq. 2]Rt=Req(e−kofft) where Req is the signal reached at equilibrium. Both equations, for association and dissociation are as described by DeMol and Fischer (22). Figure 3, shown in Results, represents a global fit of all data using single values for kon and koff. Dissociation constant, Kd, was calculated as koff/kon.</p><!><p>Circular dichroic (CD) spectra were recorded using an AVIV 202 spectropolarimeter. Lyophilized inhibitor peptides were dissolved directly in 10 mM phosphate buffer, pH 7.40. Monomeric 50 μM YAQ12A was prepared as described above in the same buffer. Spectra were obtained at 22 °C, with a 1 mm cuvette; five scans were averaged and contributions from buffer were subtracted; data were smoothed using a macro in Kaleidagraph software.</p><p>Fibril film CD spectra were measured essentially as described by Darnell et al. (12). Briefly, a concentrated slurry of fibrils was obtained by first centrifuging the slurry in a microfuge for 5 minutes, and resuspending the pellet in the same buffer; 20 μl aliquots of the slurry in five μl droplets were placed onto cut quartz plates (12.5 mm × 12.5 mm, Starna) and allowed to coalescence into one aqueous film. All samples were dried under vacuum overnight at room temperature. Spectra were taken immediately after removing the plates from vacuum, using an Aviv model 202 spectropolarimeter (Lakewood, NJ) at one nm intervals from 260 nm to 190 nm with a one sec averaging time, a one nm bandwidth, and at 25 °C. To determine whether linear dichroism was present, spectra were again recorded after the plates were rotated by 90°. Any samples that exhibited shifts in maxima or minima were discarded from the data set. Data sets showing scatter after 90° rotation were still used. Spectra shown in Results represent the averages of five scans for each film. Spectra are presented as wavelength versus background corrected raw ellipticity signal.</p><!><p>Equilibrium sedimentation analytical ultracentrifugation was performed using a Beckman Optima XLA ultracentrifuge equipped with an An-60Ti rotor with six-sector cells. Solutions of monomeric 100 μM YAQ12A were prepared as described above in 10 mM sodium phosphate, pH 7.40. This peptide sample was added to lyophilized aliquots of 5QMe2, yielding final inhibitor concentrations of 100 μM (molar ratio 1:1, YAQ12A:5QMe2), 500 μM (molar ratio 1:5, YAQ12A:5QMe2) and 1 mM (molar ratio 1:10, YAQ12A:5QMe2). Of the four cells in the rotor, one contained buffer only (10 mM sodium phosphate, pH 7.40), two were duplicates, each containing the three above described mixtures of YAQ12A and 5QMe2, while the last contained samples of 5QMe2 alone at 100 μM, 500 μM, and 1mM. Sedimentation was monitored at 230 and 280 nm, at 20 °C, with an hour between scans. Rotor speed progressed from 3000 rpm (1 scan) to 36,000 rpm (16 scans) to 48,000 rpm (13 scans) to 60,000 rpm (19 scans). Equilibrium was demonstrated by the absence of change in absorbance profile over the course of at least 8 hours.</p><p>Sedimentation equilibrium analytical ultracentrifugation data were analyzed using an approach that derives in part from previous methods for analyzing the formation of complexes by two dissimilar species, interacting at sedimentation equilibrium in an ultracentrifugal field (23-26). Absorbances used for analysis were at 280 nm, except 5QMe2 alone at 100 μM, which was followed at 230 nm for greater sensitivity at the lower concentration. Additional details of analysis, including derivation of equations, and methods for determination of v‒ (27), are given in Supporting Information.</p><!><p>To screen peptides as potential polyQ fibrillation inhibitors, we used YAQ12A, a synthetic polyQ peptide that forms fibrils over the course of several days, making it amenable to kinetic analysis. In addition, in contrast to longer polyQ peptides, YAQ12A can be highly purified. Thus, our assay consisted of measuring fibrillation in the presence and absence of potential inhibitor peptides. The potential inhibitors tested are shown in Figure 1A. Fibrils of short polyQ peptides such as YAQ12A cause very little or no appreciable thioflavin T fluorescence, however. For this reason, we used sedimentation assays (20) to follow the aggregation of YAQ12A in the presence and absence of inhibitor peptides. For the purpose of performing screening assays, high inhibitor concentrations were used (4 mM inhibitor:0.1 mM YAQ12A). In the absence of inhibitors, after incubating 100 μM YAQ12A for 40 hr at 37 °C, 8.8 ± 0.3 μM (mean ± S.D.) monomer remained in solution (Figure 1B).</p><p>The first inhibitor tested was 5MeQMe2, which contained both backbone and side chain glutamine N-methylations to disrupt hydrogen bonds at both locations. Alternate residues were modified to leave some amides available for binding to YAQ12A (16, 17). This peptide and subsequent ones discussed in this paper also contain additional moieties, as in our earlier studies on inhibitors of β-amyloid and human prion peptide (PrP106-126) aggregation (16, 17). N-Me-anthranilic acid was added for fluorescence detection in cellular studies (not discussed herein); the single Lys residue and C-terminal carboxamidation are also used to favor ingress of the peptides into cells (17) and the Lys residue may also increase solubility.</p><p>Although incubation of 5MeQMe2 with YAQ12A did show significant inhibition (36.8 ± 3.7 μM monomeric YAQ12A remaining in solution, Figure 1B), this inhibitor was considerably less effective than expected, given that it was present a 40:1 molar concentration relative to YAQ12A.</p><p>We then attempted to optimize the inhibitor. Lengthening the core glutamine sequence from 5 residues to 7 (7MeQMe2) resulted in a less effective inhibitor, (24.7 ± 1.7 μM). Based on the hypothesis that methylating both side chain and backbone amides might hinder binding of inhibitor to YAQ12A, we synthesized 5MeQ, which has only backbone N-methylations. 5MeQ proved to be slightly more effective than the other inhibitors (42.2 ± 1.7 μM). We also attempted to model a proposed polyQ fibril structure (8, 9) by connecting two 5MeQ peptides either with a tight β-hairpin turn (DPro-Gly sequence) or a flexible linker (PEG, polyethylene glycol). These peptides were moderately effective as inhibitors, yielding 31.6 ± 1.0 and 51.5 ± 4.0 μM monomeric YAQ12A remaining in solution, respectively.</p><p>Surprisingly, 5QMe2, with only side chain methyl groups (again, on alternate residues), completely inhibited aggregation of YAQ12A, with essentially all of the YAQ12A remaining in solution at the end of the incubation (91 ± 0.6 μM). (QMe2)5, with side chain methylations on all Q residues, was less effective (58.9 ± 0.9 μM monomeric YAQ12A remaining) than 5QMe2, demonstrating the importance of the alternating pattern of methylated residues. Finally, the unmethylated peptide, 5Q, was not inhibitory; after incubating 5Q with YAQ12A, the concentration of the latter peptide remaining in solution was 11.9 ± 1.7 μM, similar to YAQ12A alone.</p><!><p>We then focused on 5QMe2, by far the most effective of these inhibitors. Size exclusion chromatography confirmed that the top ultracentrifugal fraction of disaggregated YAQ12A used for these assays was most consistent with monomeric molecular weight (Figure 2F and 2G; see also Supporting Figure 4A and B). Sedimentation assays showed that YAQ12A monomer shows a monoexponential decline over 110 h, without a lag period (k = 0.04 h−1, Figure 2A). When 100 μM YAQ12A was incubated with 100 μM 5QMe2, we observed a lag period, in which soluble YAQ12A declined slowly, followed by more rapid loss of this peptide from solution after ~ 70 h. When 100 μM YAQ12A was incubated with 1 mM 5QMe2, no aggregation was observed for the entire time period. Electron microscopy confirmed the sedimentation assay results. Incubation of YAQ12A with 5QMe2 (500 μM, for Figure 2E) abrogates fibril formation at 40 h.</p><p>As stated, in the absence of inhibitor, much of the YAQ12A initially in solution precipitated and formed fibrils. As shown in Figure 2F and 2G, however, at 0, 24 and 50 h, essentially all of the YAQ12A in solution eluted from the size exclusion chromatography column at a position most consistent with monomeric molecular weight. If any oligomeric YAQ12A was in solution at these three times, it represents a very small and probably kinetically transient fraction of the total YAQ12A in solution. The effect of adding 5QMe2 to YAQ12A was to keep the latter peptide in solution for a longer period of time, but 5QMe2 did not alter the chromatographic behavior of YAQ12A. As shown in Figure 2G, at increasing times, more YAQ12A remained in solution in the presence than in the absence of 5QMe2. The chromatographs show two well-resolved peaks, one each in the positions previously observed for YAQ12A and 5QMe2. In the presence of 5QMe2, essentially all of the YAQ12A was still in solution at 24 h, as shown by the fact that the size of the peak was essentially unchanged from that at 0 h of incubation. At 50 h, the YAQ12A peak was slightly smaller than that seen at 0 or 24 h, but no additional peaks were observed. We demonstrate below that these two peptides form a complex, with Kd ~ 1 μM. The absence of a peak for this complex is consistent with dissociation of a complex of this affinity during chromatography.</p><!><p>Many aggregating peptides are believed to form fibrils by a rate limiting nucleation step followed by elongation of the nucleus (3, 5, 28, 29). To learn whether 5QMe2 could prevent aggregation of YAQ12A when added at different stages during 70 h of fibril formation, we added 1 mM 5QMe2 to 100 μM YAQ12A 10 or 20 h after the start of incubation (Figure 2B). Monomeric YAQ12A remaining after 70 h was then assessed by sedimentation assay. 5QMe2 was able to prevent all but a modest further decline in monomer concentrations, regardless of the time of addition. These experiments suggest that 5QMe2 may act to disrupt both nucleation and elongation steps and maintain YAQ12A in its monomeric state. The inhibitor did not reverse aggregation, however, as no additional monomer appeared in solution after incubation of the mixture with the inhibitor.</p><!><p>These experiments were designed to assess the whether 5QMe2 could inhibit precipitation of YAQ12A from solution in the presence of pre-formed fibril seeds. The results shown in Figure 2A and 2E indicate that 5QMe2, when added to YAQ12A at the start of a fibrillation reaction, can act as a nucleation inhibitor. That is, 5QMe2 can prevent or delay growth of fibrils under unseeded fibrillation conditions. The data shown in Figure 2B suggest that 5QMe2 can also act as a fibril extension inhibitor, since addition of 5QMe2 10 or 20 h after the start of the reaction prevented nearly all additional precipitation of YAQ12A from solution. To test further the idea that 5QMe2 can act as a fibril extension inhibitor, we examined seeded growth of YAQ12A. In particular, pre-formed fibril seeds were added to disaggregated solutions of YAQ12A, with or without 5QMe2, as described in Methods. As a control experiment, the same incubations were performed without the addition of YAQ12A seeds. The results of these experiments are shown in Figure 3. The data in Figure 3A for unseeded fibril growth is similar to that shown in Figure 2: in unseeded, freshly disaggregated solutions of YAQ12A, peptide spontaneously precipitates and forms fibrils, and the addition of 5QMe2 inhibits fibrillation, causing the YAQ12A to remain in solution. When YAQ12A fibrils are added to fresh, disaggregated solutions of YAQ12A (Figure 3B), more YAQ12A is lost from solution at each time point than in the absence of added seeds. Thus, the added fibrils do indeed seed fibrillation. When 5QMe2 is also added to the mixture of fibril seeds and fresh YAQ12A solutions, it inhibits loss of soluble YAQ12A into the insoluble fraction. Since the addition of fibril seeds largely bypasses nucleation, these data indicate that 5QMe2 inhibits fibril extension. Thus, 5QMe2 is both a nucleation and fibril extension inhibitor.</p><!><p>We examined the inhibitor concentration dependency using sedimentation assays. 100 μM YAQ12A was incubated 40 h with various concentrations of 5QMe2, at molar ratios ranging from 40:1 to 1:10 (5QMe2:YAQ12A), and after ultracentrifugation, the top fraction was assayed for soluble YAQ12A. As shown in Table 1, inhibition was essentially complete at ratios of 1:1, but was significant even at a substoichiometric ratio of 1:10 (5QMe2:YAQ12A).</p><!><p>After observing that 5QMe2 could inhibit aggregation of YAQ12A at substoichiometric levels, we investigated the affinity and kinetics of YAQ12A and 5QMe2 binding with surface plasmon resonance (SPR). We reasoned that 5QMe2 could act at substoichiometric ratios by interacting transiently and sequentially with monomeric YAQ12A molecules. SPR experiments showed rapidly reversible binding and dissociation of 5QMe2 to immobilized YAQ12A at a range of inhibitor concentrations. Figure 4A shows a global fit of the data to monoexponential rate equations (described in Methods); Figure 4B shows residuals (differences between experimental values and theoretical fits). From these analyses, kon and koff for each set of triplicate runs were 0.026 μM s−1 and 0.024 s−1, respectively, yielding a value of 0.92 μM for Kd. Data for individual curves were also analyzed using the same equations, and the numbers obtained for kon, koff, and Kd (representing an arithmetic mean of the values) were .030 μM s−1, 0.026 s−1, and 0.87 μM, respectively, in reasonable agreement with the numbers obtained from global fitting. These data indicate very rapid association and dissociation of predominantly a 1:1 complex of YAQ12A and 5QMe2, and moderate affinity. The rapidity of binding and dissociation is reminiscent of chaperones, which act substoichiometrically by making rapid, transient complexes with sequential molecules of their targets.</p><!><p>We used sedimentation equilibrium analytical ultracentrifugation experiments to analyze the complex (IQ) formed from YAQ12A (Q) and 5QMe2 (I). We first demonstrated that the inhibitor alone is monomeric at three concentrations (100, 500, and 1000 μM, and at three rotor speeds (Figure 5). The experimental molecular weight (mean ± S.D. for all of the above conditions) for 5QMe2 at these concentrations was 975.2 ± 32.1, very close to the expected value of 976 g/mol. When the results for the three concentrations were analyzed globally, the molecular weights obtained were 988.5, 1000.0, and 980.1 for rotor speeds of 36,000, 48,000 and 60,000 rpm, respectively. This is again in agreement with the expected molecular weight for 5QMe2.</p><p>The complex formed between I and Q was analyzed using the experimental procedures described by Winzor et al. (23), which is based on earlier papers of the same group (24-26). A description of this technique, including derivations of equations as used in the experiments on these particular peptides, is given in detail in the Supporting Information. Briefly, sedimentation equilibrium data are analyzed according to the equation, [Eq. 3]At(r)=CI(rF)εIψI(r)+CQ(rF)εQ[ψI(r)]p+CIQ(rF)εIQ[ψI(r)]p+1+CI2Q(rF)εI2Q[ψI(r)]p+2+… where Ci(r) is the concentration of each species at each radial position, r, in the ultracentrifugal field. The function Ψi(r) is defined as exp[Miϕi(r2 – rF2)], where ϕi=(1−vi‒ρ)ω22RT, and Mi is molecular weight of each species, rF is an arbitrarily chosen fixed radial distance from the center of rotation, v‒ is the partial specific volume of each species, ω is angular velocity, ρ is solvent density, T is temperature (K), and R is the gas constant. For synthetic peptides, the values of Mi are known; here, MI = 976, MQ = 1860, MIQ = 2736, and so forth for other types of complexes. The parameter, p is defined as Miϕi/Mjϕj; in the present experiments, i = Q and j = I, and thus p = 1.90.</p><p>In these experiments, sedimentation equilibrium ultracentrifugation was performed using three sets of peptide concentrations, either an equimolar concentration of I and Q (100 μM of each), or excess of the inhibitor (500 or 1000 μM of I with 100 μM of Q). In addition, the sample was centrifuged to equilibrium at three angular velocities, 36000, 48000, and 60000 rpm. It was necessary to use equimolar or higher inhibitor concentration because at lower concentrations of the inhibitor, Q would not stay in solution, as is also indicated in Figure 2.</p><p>The null hypothesis to be tested in these experiments is that, in agreement with SPR data, the predominant complex between I and Q is a 1:1 complex. Under the conditions used in these experiments (equimolar or excess I compared to Q) and given the Kd (0.92 μM) obtained from SPR, the expected concentration of Q is low compared to those of either I or IQ, and also with respect to the experimental error in measurement of absorbance. For this reason, the results were analyzed in two, complementary ways, as follows.</p><p>First, the data were analyzed according to Eq. 3, and the value of Kd was evaluated from the values of the three parameters, CI(rF), CQ(rF), and CIQ(rF). Second, we used the value of Kd obtained from SPR experiments to calculate the least precise (because it has the lowest value) of the above three parameters, CQ(rF). In other words, in the second procedure data were analyzed by non-linear least squares methods, using the equation: [Eq. 4]At(r)=CI(rF)εIψI+CI(rF)εI⋅KdCIQ(rF)εIQψIp+CIQ(rF)εIQψIp+1</p><p>Figure 6 shows the results of these analyses. In the figure, the symbols are experimental data points, the thin black line is the analysis of the data using EQ. 3, and the somewhat thicker gray line represents the fit of the data using EQ. 4. As shown in the figure, the two fits are essentially superimposable. A mean value (± S.D.) of 0.74 ± 0.22 μM was obtained for Kd from the values of CI(rF), CQ(rF), and CIQ(rF), which is in reasonable agreement with the values for Kd obtained from SPR of 0.92 μM.</p><p>The analysis also indicated that nearly all of the Q (YAQ12A) remaining in solution is part of a complex. Although additional terms, corresponding to higher order complexes, could be added to the equation to fit the data, such terms did not improve the fit (Supporting Figure 3). Since the theoretical fits overlap completely and are indistinguishable from that obtained using the simpler equation, it is most parsimonious to conclude that the predominant complex is a 1:1 stoichiometric complex, also in agreement with SPR data. Finally, as discussed in Supporting Information, the fit of data to the model containing a complex was compared to that containing no complex, by using the Aikake Information Criterion (AIC, 21, 30). The AIC is a measure of the goodness of fit of an estimated statistical model that rewards goodness of fit (as determined by the sum of squares or χ2), and also penalizes the number of parameters and therefore penalizes overfitting. Thus, in general, a lower AIC indicates a preferable model, i.e., the best fit of the data with a minimum number of free parameters. As indicated in Supporting Table 1, in all cases but one (which was a virtual tie) the inclusion of a term for a 1:1 complex, rather than only free I and free Q, lowered the AIC, indicating that this was a preferable model. In conclusion, the analytical ultracentrifugation data are consistent with a model in with YAQ12A forms mainly a 1:1 complex with 5QMe2.</p><!><p>The CD spectrum of YAQ12A in solution shows a small maximum at ~ 220 nm, and a trough with a minimum at ~ 199 nm, consistent with a PPII-like helical structure (Figure 7A). When it forms fibrils, it undergoes a transition to β-sheet structure, as shown by film CD spectroscopy (Figure 7B), as do other short polyQ peptides (12, 31-34). The CD spectrum of 5QMe2 is also consistent with a left-handed, PPII-like structure (Figure 7A), though with a red-shifted maximum at ~ 232 nm, and a minimum at ~ 199 nm, possibly suggesting subtle structural differences from YAQ12A. In contrast to 5QMe2, two of the less effective inhibitors, 5MeQ and 5MeQMe2 with backbone N-methylation, had CD spectra indicative of β-sheet (Figure 7C). These data suggest that the most effective inhibitor has a structure resembling the soluble form of its PPII target peptide, YAQ12A, while the less effective inhibitors have the structure, β-sheet, of the final fibrillar product of the aggregation pathway. A 1:1 mixture of YAQ12A:5QMe2 did not match an arithmetic weighted mean of the spectra of the two individual peptides (Figure 7A), which is consistent with above observations that the two peptides form a distinct complex.</p><!><p>In this paper, we have described N-methylated inhibitors of polyQ peptide aggregation. 5QMe2, with only alternate side chain amides N-methylated, was far more effective than the others, including 5MeQ with only backbone N-methylations, and 5MeQMe2 with both backbone and side chain N-methylation. These results demonstrate the importance of side chain hydrogen bonding in polyglutamine fibrillation, which has been posited from molecular modeling (35, 36). Also, the inhibitor in which alternate side chains were methylated (5QMe2) was more effective than that in which all the side chains were methylated ((QMe2)5), indicating the need to retain some unmodified amides for binding to the aggregating target peptide.</p><p>Addition of 5QMe2 to YAQ12A at stoichiometric ratios or higher prevented disappearance of monomeric YAQ12A from solution for 40 h, although at 70 h, complete inhibition required higher stoichiometric ratios (e.g., 5QMe2:YAQ12A =10:1). These results suggest that 5QMe2 inhibits nucleation of YAQ12A, but does not eliminate it. In addition, adding 5QMe2 to solutions of YAQ12A at various times after the start of fibrillation prevented all but a minor degree of further aggregation of YAQ12A for times up to 70h, suggesting that 5QMe2 also inhibits fibril elongation. Furthermore, 5QMe2 inhibited seeded fibrillation of YAQ12A, which largely bypasses nucleation steps, suggesting that 5QMe2 is a fibril extension as well as a nucleation inhibitor. 5QMe2 causes YAQ12A to remain in a monomeric state rather than that of a soluble multimer. Thus, it is more properly called an aggregation inhibitor, rather than a fibrillation inhibitor.</p><p>SPR indicates very rapid binding and dissociation of a 1:1 complex, with moderate affinity (Kd = 0.92 μM). Both SPR and sedimentation equilibrium analytical ultracentrifugation suggest that the predominant complex formed by 5QMe2 and YAQ12A has a 1:1 stoichiometry. The rapidity with which the complex dissociated was also shown by size exclusion chromatography, where both 5QMe2 and YAQ12A eluted at a position most consistent with monomeric MWapp. Chromatography of a mixture of these peptides yielded two peaks, in the same elution positions as the individual peptide, indicating complete dissociation of complex during the chromatography. (For further discussion of this point, see Supporting Information and Supporting Figures 4A and B.) This binding pattern of target peptides is reminiscent of chaperone proteins, which also bind their targets with moderate affinity, but have very high kon and koff. Rates have been measured for Hsp70 and Hsp90 (37, 38) (though these particular chaperonins do not bind polyQ peptides). Although it is sometimes assumed that a high affinity indicates a better aggregation inhibitor, this is not necessarily the case. 5QMe2 is notable for acting at sub-stoichiometric levels, suggesting that the kinetics of complex association and dissociation are faster than the kinetics of YAQ12A aggregation. Thus, the effectiveness of 5QMe2 is likely due to fast kon and koff rates, and could be hampered by a higher affinity.</p><p>Thus, 5QMe2, like chaperones, may keep aggregation-prone peptides or proteins in a less aggregation-prone conformation, e.g., by preventing β-sheet formation. Indeed, CD spectroscopy shows that 5QMe2 does not have a β-strand structure. Rather, its spectrum is most consistent with a PPII helical conformation. In contrast, the backbone modified peptides, 5MeQ and 5MeQMe2, showed β-strand structure by CD (Figure 7C), and were less effective inhibitors than 5QMe2. It is striking that the secondary structure of 5QMe2 resembles that of its target peptide, a property that may be generalizable to all N-methylated inhibitors. Indeed, when we added 5QMe2 to YAQ12A, both reactants had structures consistent with PPII or PPII-like helices by CD spectroscopy (although not identical CD spectra), and the products in the mixture also had a PPII structure. Similarly, inhibitors of Aβ aggregation, such as the backbone N-methylated peptide, Aβ16-20m (16, 17) have similar secondary structure as their targets, but in those cases, β-sheet.</p><p>These results are consonant with our recent observations (18) that short peptides containing a QQQ triplet, aggregate into soluble oligomeric species as PPII helices, although they are too short ever to form fibrils or adopt a β-sheet structure. Longer polyQ peptides, such as YAQ12A, can form insoluble β-sheets, but as soluble monomers, have a PPII helical structure. The PPII conformation is also stabilized in polyQ segments by a polyproline segments adjacent to the C-terminal end of the polyQ segment (12, 39).</p><p>The foregoing also helps to explain the fact that, in contrast to Aβ16-20m and other backbone-modified aggregation inhibitors of β-amyloid and prion peptide aggregation (16, 17, 40, 41), 5QMe2 did not disassemble pre-formed fibrils. 5QMe2 interacts with soluble (PPII-like) YAQ12A, but apparently not with fibrillar (β-sheet) YAQ12A. 5QMe2 also prevents fibril elongation, possibly by interacting with YAQ12A still in solution. A model of 5QMe2 action is depicted in Figure 8. According to this model, YAQ12A exists in solution as PPII-like monomer. It is possible that this peptide can also forming oligomers, but we did not observe this, suggesting that oligomers, if present, were at very low concentrations or were short-lived. Progression to fibril formation requires a transition from PPII to β-sheet conformation, of YAQ12A monomers, oligomers, or both. 5QMe2 forms transient 1:1 complexes with YAQ12A monomers, thereby maintaining the latter peptide in a PPII-like state. Binding of 5QMe2 appears to inhibit formation of hydrogen bonds between side chains of YAQ12A molecules. The fact that 5QMe2 both maintains YAQ12A in a monomeric PPII-like state, and also prevents this peptide from forming fibrils suggests that some species with PPII structure may be on-pathway for fibril formation.</p><p>Finally, it is appropriate to address the relevance of YAQ12A to aggregation of longer polyQ proteins in biological contexts. PolyQ peptides and proteins have a sharp threshold for aggregation into fibrils, but this threshold is context dependent (12, 42, 43). In the context of the protein huntingtin, this threshold is ~ 35 glutamine residues, as is shown by the occurrence of disease in people with CAG triplet expansions of the gene (1), in model systems such as C. elegans and S. cerevisiae (44, 45), and in studies of recombinant htt exon 1-encoded protein in vitro (2, 3). In spinocerebellar ataxia (SCA) type 6 however, the threshold for disease is considerably shorter. The gene for this disease encodes the α1A-voltage-dependent calcium channel subunit, and disease and intracellular protein aggregates occur when the polyQ expansion exceeds 20-30 residues (46). For peptides without neighboring domains, even as few as six Q residues is sufficient to form β-sheet-rich fibrils (12), and YAQ12A clearly is above this threshold (Figure 7B). Thus, while the threshold might change with context, inhibitors such as 5QMe2 can serve as structural probes of the aggregation pathway of biologically relevant polyQ peptides and proteins. Finally, although we have used this inhibitor peptide primarily as a structural probe, it could possibly also serve as the basis of an avenue of therapeutic approach.</p>

PubMed Author Manuscript

SAR and molecular mechanics reveal the importance of ring entropy in the biosynthesis and activity of a natural product

Macrocycles are appealing drug candidates due to their high-affinity, specificity, and favorable pharmacological properties. In this study, we explored the effects of chemical modifications to a natural product macrocycle upon its activity, 3D geometry, and conformational entropy. We chose thiocillin as a model system, a thiopeptide in the ribosomally-encoded family of natural products that exhibits potent antimicrobial effects against gram-positive bacteria. Since thiocillin is derived from a genetically-encoded peptide scaffold, site-directed mutagenesis allows for rapid generation of analogs. To understand thiocillin\xe2\x80\x99s structure-activity relationship, we generated a site-saturation mutagenesis library covering each position along thiocillin\xe2\x80\x99s macrocyclic ring. We report the identification of eight unique compounds more potent than WT thiocillin, the best having an 8-fold improvement in potency. Computational modeling of thiocillin\xe2\x80\x99s macrocyclic structure revealed a striking requirement for a low entropy macrocycle for activity. The populated ensembles of the active mutants showed a rigid structure with few adoptable conformations while inactive mutants showed a more flexible macrocycle which is unfavorable for binding. This finding highlights the importance of macrocyclization in combination with rigidifying post-translational modifications to achieve high potency binding.

sar_and_molecular_mechanics_reveal_the_importance_of_ring_entropy_in_the_biosynthesis_and_activity_o

<p>Natural products are critical modulators of microbial and multicellular biology. Roughly one-third of the pharmacopeia are derived from small molecule natural products. Many of these exceed Lipinski's rules with sizes in the range of 500–1500 Da but retain favorable pharmacokinetic properties, enabling oral dosing in many cases. Macrocycles are conformationally constrained by cyclization, which has been suggested to reduce their apparent size and pre-organize the compound into a low entropy state, facilitating permeation and target binding. However, macrocyclic rings of the same size can vary dramatically in their conformational flexibility, due to the presence or absence of rigidifying elements such as double bonds or backbone rings. Here we assess the effect of ring entropy on binding and activity in a model natural product system, by combining systematic mutational analysis with computational modeling of ring entropy.</p><p>We chose to study the natural product thiocillin, a thiopeptide in a class called RiPPs, ribosomally synthesized and post-translationally modified peptides. Thiocillin undergoes a cascade of post-translational modifications (PTMs) to form a mature macrocyclic natural product.1 Thiopeptide antibiotics inhibit the growth of gram-positive bacteria including MRSA and VRE at nanomolar concentrations.2 Like many other thiopeptides, thiocillin targets the interface between ribosomal protein L11 and the 23S rRNA.3 Thiocillin's prepeptide contains a C-terminal core peptide and an N-terminal leader sequence, which is removed once modifications are completed.4 Common thiopeptide PTMs include thiazole (from Cys), oxazole (from Ser), methyloxazole (from Thr), dehydroalanine (Dha) (from Ser), and dehydrobutyrine (Dhb) (from Thr); all of these modifications rigidify the peptide.5 The modified core peptide undergoes an enzyme-catalyzed [4+2] cycloaddition reaction to close the macrocycle and forms a pyridine core.6,7 One of our goals was to understand the importance of rigidifying modifications in the peptide to macrocyclization and potency.</p><p>We present a systematic SAR analysis of thiocillin by saturation mutagenesis and computational modeling. Thiocillin has previously been subjected to alanine scanning, cysteine-to-serine scanning, ring-size variants, incorporation of noncanonical amino acids, and various point mutants.8–11 In this work, we conducted saturation mutagenesis on macrocycle residues 2–9, the 8 residues not involved in macrocycle linkage, thus producing a total of 152 single point mutants. This provided a comprehensive understanding of tolerated amino acid replacements, including variants with enhanced antibacterial activity. Importantly, we noted sharp SAR that separated active and inactive analogs. Molecular mechanics modeling combined with NMR studies showed the steep loss of activity seen with certain mutants was the result of dramatic increases in ring entropy. Thus, conformational constraints beyond macrocyclization are critical for the activity of this natural product.</p><p>To rapidly generate mutants of thiocillin, we used a plasmid complementation strategy (Figure 1B). By expressing the prepeptide gene, tclE, in Bacillus cereus ΔtclE-H, a strain lacking the endogenous prepeptide gene, we demonstrate the rescue of thiocillin production (Figure S1).</p><p>The mature form of thiocillin contains a large macrocycle closed via a pyridine ring which absorbs light at 350 nm. To screen for macrocycle formation, methanolic extracts from small-scale 1.5 mL cultures were analyzed by LC/MS. Presence of a 350 nm peak and a mass consistent with the mutation indicated the successful macrocyclization of 25 mutants (Figure 2). Our results suggest mutations to non-thiazole forming residues 3, 4, 6, and 8 are tolerant to mutations, with 6 and 8 being the most tolerant without loss of activity. Only residue 8 was able to accept large aromatic side chains such as phenylalanine and tyrosine. Mutating thiazole-forming Cys residues was poorly tolerated, consistent with a previous publication.8 These mutants were not detectable in our small scale high-throughput assay, which deliberately focused on identifying highly-expressing and well- tolerated mutants.</p><p>To screen for active mutants, an overlay assay was used against Bacillus subtilis 168, a representative grampositive bacterium. Engineered B. cereus strains producing an active thiopeptide create a zone of inhibition when overlaid with B. subtilis (Figure S3). This screen identified 18 thiocillin mutants with antibiotic activity (Figure 2). Production of these 18 active mutants and an inactive negative control mutant were scaled up and purified for quantitative minimal inhibitory concentration (MIC) determination. Since each mutant produces multiple unique compounds due to auxiliary PTMs on residues 6 and 8, we isolated and screened all unique compounds with yields greater than 0.1 mg/L. In total, 33 unique analogs were purified from the active mutants and 2 from the negative control. The MIC assay resulted in 7 compounds more active than WT (Table 1). The most potent compound was V6A2 with an MIC of 0.06 μg/mL, an 8-fold increase in potency. This particular mutant was previously discovered to have a 4-fold increase in potency when assayed as a mixture of compounds.8 Interestingly, we identified mutants such as T8F, which showed activity on solid media, but not in liquid culture. This may be due to the hydrophobic Phe side-chain limiting its solubility.</p><p>To examine the influence of each single point mutation on conformational entropy, we sampled the potential energy landscape for every analog using BRIKARD.12 BRIKARD applies inverse kinematics to enable efficient generation of low-energy conformations of macrocycles. We initially validated our computational approach by reproducing near-native states extracted from crystallography data for similar thiopeptides, including one bound to ribosomal protein L11.3 Detailed methods are available in the SI. The conformations generated for each thiopeptide by BRIKARD were then clustered using a stringent Cartesian RMSD metric (0.25 Å) to eliminate redundant (nearly identical) conformations. While this collection of low-energy states cannot be interpreted as a true thermodynamic ensemble, the number of such states varied dramatically among the thiocillin variants (Table 2), reflecting the rigidity of the macrocycle and hence its entropy.</p><p>Our results activity data, with a clear division between the highly rigidified active analogs having a maximum of 47 conformational states compared to the >250 conformational states accessible to the inactive mutants (Table 2). This finding suggests that, within the thiocillin family, mutations that dramatically increase the flexibility of the macrocycle ablate binding, due at least in part to the increased entropic loss required for binding. We cannot, of course, rule out other effects of the mutations also impacting binding affinity and activity. Conversely, conformational rigidity does not guarantee activity, i.e., the data suggests that rigidity is necessary but not sufficient for activity.</p><p>One dramatic example with mutant T4A shows that breaking the planar character of the Dhb residue at position 4, which resulted in a mature macrocycle with no detectable activity. In this case, simply changing a planar sp2 alpha carbon to a tetrahedral sp3 geometry leads to a dramatic increase in backbone entropy (Figure 3). These striking changes in macrocycle entropy exemplify the need for computational sampling of ring entropy.</p><p>Following our computational study of the potential energy landscape of thiocillin and our mutant analogs, NMR was used to determine the structure of WT thiocillin in DMSO. The 3D HNHA NMR experiment was used for structural determination because it is an accurate method for measuring homonuclear three-bond 3JHNHa coupling constants and was used to help elucidate the structure of lassomycin.13 The 3JHNHa values obtained from 15N-labelled thiocillin were used to estimate dihedral phi angles for Thr-3, Val-6, and Thr-8 according to the Karplus equation (Table S5).14 These values were used as restraints in BRIKARD to generate an ensemble of conformations consistent with the NMR data (Figure S4A). Five similar conformations were obtained with and without experimental restraints, further supporting the use of BRIKARD for conformational sampling. Our predominant NMR structure (Figure S4B) (31% occupancy) showed a similar folded conformation to a previous NMR structure obtained using ROESY data.8</p><p>Based on our experimental and computational SAR, we developed a second generation combinatorial thiocillin library randomizing residues 6 and 8, the most tolerated sites. 1200 colonies were screened by overlay assay, observing many previously known single mutants as well as 16 new double mutants. Six double mutants that were combinations of potent single mutants were purified to produce 8 unique compounds and subjected to MIC testing. Only the V6A-T8V double mutant showed a 2-fold increase in potency over WT (Table S2). These double mutants did not show the expected synergy, suggesting that other global molecular properties such as solubility or permeability may be limiting their activity.</p><p>The three PTMs characteristic of thiopeptides are formation of (1) five-membered heterocyclic thiazoles and oxazoles, (2) sp2 side chains of Dha/Dhb, and (3) a macrocycle enclosing pyridine/dehydropiperidine ring. All are rigidifying modifications. The conformational analyses in this study suggest the entropy restricting property of these planar Dha/Dhb side chains may be as important as their previously appreciated chemical reactivity, as electrophiles in lanthionine residue formation and as dual electrophiles and nucleophiles in pyridine/dehydropiperidine ring formation. If generalizable, the Dha/Dhb-forming PTM may reveal yet another layer of chemical logic used in nature to create high affinity molecular scaffolds.</p><p>Our studies suggest that a rigid macrocycle is a requirement for binding in this system and that very small chemical changes can lead to substantial increases in entropy as measured by computational modelling. The formation of a macrocycle greatly reduces backbone entropy, however, macrocyclization alone appears insufficient for preorganization of the compound. Particularly in thiopeptides, the presence of rigidifying PTMs such as heterocyclization and dehydrations appears to lower the entropic barrier for binding. In other macrocycle systems that are not heavily modified, entropy reduction can be achieved by intramolecular hydrogen bonding. Entropy is an often overlooked parameter in macrocycle design and the principles learned here can be applied to other natural and synthetic macrocycle systems.</p>

PubMed Author Manuscript

Concentrated dual-cation electrolyte strategy for aqueous zinc-ion batteries

Rechargeable Zn-ion batteries are highly promising for stationary energy storage because of their low cost and intrinsic safety. However, due to the poor reversibility of Zn anodes and dissolution of oxide cathodes, aqueous Zn-ion batteries encounter rapid performance degradation when operating in conventional low-concentration electrolytes. Herein, we demonstrate that an aqueous Zn 2+ electrolyte using a supporting Na salt at a high concentration is efficient to address these issues without sacrificing the power densities, cycling stability, and safety of zinc-ion batteries. We show that the highconcentration solute minimizes the number of free water molecules and the changes in the electronic state of the electrolyte. A combination of experimental and theoretical investigations reveals that a unique interphase, formed on the Zn anode, enables reversible and uniform Zn plating. Utilizing a cathode of sodium vanadate synthesized through a scalable strategy, the Zn-sodium vanadate battery with the concentrated bi-cation electrolyte shows improved cycling stability, decent rate performance, and low self-discharge. This work provides new insights on electrolyte engineering to achieve highperformance aqueous batteries. Broader contextElectrolytes play a critical role in determining the performance of electrochemical storage devices, such as cycling stability, rate performance, and selfdischarge. To date, the promising aqueous Zn-ion batteries still suffer from low cycling performance due to the low reversibility of the Zn anode and serious dissolution of active cathode materials during cycling in conventional aqueous electrolytes. Recently, highly concentrated ''water-in-salt'' electrolytes have been proposed to decrease the amount of free water molecules for improved cyclability, though the selection is still limited to expensive organic or toxic solutes. Herein, to address the challenges of developing formulations of aqueous electrolytes, we design a dual-cation electrolyte that features high ionic conductivity, low cost, and intrinsic safety. Importantly, the high concentration can not only suppress water activity and decrease the amount of free water molecules, but also induce the change of electronic state of the electrolyte. The former prohibits cathode dissolution, while the latter ensures a high reversibility for the anode. These factors together enable stable aqueous Zn-ion batteries with decent rate performance and improved self-discharge performance.

concentrated_dual-cation_electrolyte_strategy_for_aqueous_zinc-ion_batteries

Introduction<!>Results and discussion<!>Paper<!>View Article Online<!>À<!>2+

<p>Compared to the prominent energy storage technologies (e.g., Li-ion batteries and Pb-acid batteries), aqueous Zn-ion batteries (ZIBs) show the advantages of safety, affordability, and environmental friendliness. 1 Aqueous ZIBs benefit from the direct use of a Zn metal anode with high capacity (820 mA h g À1 and 5854 mA cm À3 ) and decent electrochemical potential (À0.76 V vs. standard hydrogen electrode (SHE)). These features make aqueous ZIBs particularly promising for grid-scale storage and other applications (e.g., home batteries) in which energy density is not a major concern. 2 Importantly, aqueous ZIBs can be compatible with the existing Li-ion battery manufacturing infrastructure, in which billions of dollars have been invested, 3 a significant indicator for potentially commercializing aqueous ZIBs. Their full a Materials Science and Engineering, King Abdullah University of Science and levelized cost of electricity should be further reduced to promote their application at scale. This will benefit from the cost reduction expected for scale-up fabrication (e.g., preparing cathode materials at scale) and increased battery cycle life. 4 The working mechanism of an aqueous ZIB depends on the shuttling of Zn 2+ between a Zn metal anode and a cathode capable of reversible Zn 2+ (de)insertion. Different formats of Zn metals (e.g., foil, powder, and mesh) are commercially available at a low price. However, the conventional aqueous electrolytes lead to fast performance degradation during battery operation because of: 2,5-10 (1) the poor reversibility of the Zn metal anode caused by the side reactions (e.g., hydrogen evolution reaction and corrosion) and non-uniform Zn plating, and (2) the dissolution of cathodes into the bulk electrolytes upon cycling, which is more conspicuous when operating at low current densities. These issues are strongly related to the existence of a large number of free water molecules in the conventional aqueous electrolytes (e.g., 2 M ZnSO 4 ). The recent development of ''water-in-salt'' (WIS) electrolytes and deep-eutectic electrolytes, both of which feature limited or even no free water molecules, renders an unprecedented opportunity to overcome the aforementioned shortcomings of aqueous ZIBs. 5,[11][12][13][14][15][16][17][18] Based on the eutectic effect, WIS electrolytes with an extremely high concentration of 75 m (m: mol kg H 2 O À1 ) could be even achieved. 19 On the one hand, the dissolution of cathodes can be impeded, enabling improved cycling stability. 13,14 On the other hand, the reorganized solvation-sheath structure of cations (e.g., Zn 2+ ) (typically dominated by binding anions) suppresses the hydrolysis effect efficiently, providing a high possibility for reversible Zn plating/stripping with high coulombic efficiency (CE). 11,20,21 Accordingly, engineering electrolytes through using solutes with intrinsically high solubility or additives can enable high-performance aqueous rechargeable batteries. 22,23 In addition, the physicochemical properties (i.e., ionic conductivity and electrochemical stability window) and economic and environmental aspects of the aqueous electrolytes should be considered when developing aqueous ZIBs for potential practical applications. 22 However, most WIS electrolytes with excess reaction-irrelevant anions (e.g., costly fluorinated metal salts) lead to cost and safety concerns, which compromises the economic and environmental benefits anticipated for aqueous ZIBs. 2 For instance, the frequently used lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) (42.8 euros per g) in the classical WIS electrolytes shows acute dermal and oral toxicity as well as chronic aquatic toxicity. 13 Additionally, the classical WIS electrolytes show high viscosity and low ionic conductivity, thereby resulting in poor rate performance and low energy efficiency. 13,24,25 Recently, 30 m ZnCl 2 WIS electrolyte with high viscosity of B1000 mPa s and low ionic conductivity of B2 mS cm À1 has been shown as a low-cost electrolyte for ZIBs with decent cyclability and rate capability, though few cathodes are likely workable in this electrolyte. 21,26 All these factors make it extremely challenging to apply the current WIS electrolyte chemistry into potential commercial energy storage devices for intermittent renewable energy storage.</p><p>Significant progress has been achieved in identifying cathode materials for aqueous ZIBs. 2,10,27 The synthesis of these cathode materials typically utilizes hydrothermal methods with prolonged heat treatment, which not only increases the instrumental cost but also leads to low product yield (e.g., B100 mg per batch). To achieve the potential application of aqueous ZIBs for future energy storage, a scalable approach to fabricate high-performance cathodes should be developed. Herein, we develop a room-temperature, scalable dissolutionrecrystallization method to prepare Na 2 V 6 O 16 ÁnH 2 O (hereafter referred to as NVO) nanofibers as an efficient cathode for aqueous ZIBs. To improve the cycling stability of our NVO cathode, we design a low-cost and environmentally benign WIS electrolyte of Zn(ClO 4 ) 2 with a highly concentrated supporting salt of NaClO 4 , specifically, 0.5 m Zn(ClO 4 ) 2 with 18 m NaClO 4 (see the price comparison of different solutes in Fig. S1, ESI †). The supporting salt of NaClO 4 can efficiently adjust the solvation structure and electronic states of the electrolyte upon increasing the concentrations, as discussed below. Therefore, the highly concentrated electrolyte shows the scarcity of free water molecules, low viscosity, and high ionic conductivity, which can efficiently prohibit the dissolution of the NVO cathode to enable remarkable cycling stability and high rate performance.</p><!><p>In conventional aqueous electrolytes, vanadium-oxide-based cathodes experience rapid capacity degradation because of the dissolution of cathodes during cycling (Fig. S2, ESI †). 1,2,8 One strategy to address this issue is to add salt additives into the electrolytes, wherein the cation of the additive is the same as the metal cation in the cathode (e.g., Na + for NaV 3 O 8 ÁnH 2 O and Mn 2+ for MnO 2 ). 28,29 This may shift the dissolution equilibrium of active cathodes, thereby suppressing their continuous dissolution during battery operation. However, this strategy is still inefficient in retaining the capacity when operating at low current densities (e.g., 0.1 A g À1 ). This problem correlates with: (1) the structural disintegration of vanadium oxides at a high degree of Zn 2+ intercalation (high capacities), and (2) the slow cathode dissolution and low Zn plating/ stripping efficiency (increased side reaction of hydrogen evolution at low current densities) because of the existence of a large number of free water molecules. The recent development of superconcentrated WIS electrolytes enables the stable operation of aqueous ZIBs even at low current densities through minimizing or even removing free water molecules, which, however, comes at the cost of the rate capability because of the significantly increased viscosity and decreased ionic conductivity. 11,26 Additionally, the rate performance of aqueous ZIBs is intuitively associated with the resistance of Zn 2+ transport in the electrode. Our NVO electrode has ample vacancies in the layered structure, showing structural stability and flexibility to accommodate fast Zn 2+ intercalation at high capacities. Inspired by the WIS concept using an additional supporting salt, [11][12][13]25 we here present a high-concentration Zn-ion electrolyte</p><!><p>Energy & Environmental Science</p><p>Open Access Article. Published on 03 July 2021. Downloaded on 7/6/2022 12:29:13 AM. This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.</p><!><p>(hereafter denoted as HCZE) using highly soluble NaClO 4 as the supporting salt (0.5 m Zn(ClO 4 ) 2 + 18 m NaClO 4 ), which enables stable operation of the ZIB using an NVO cathode with a remarkable rate capability. HCZE possesses a low viscosity of 8.16 mPa s and a high ionic conductivity of 98.5 mS cm À1 , which are superior to those of the classical WIS electrolytes 14,22,[30][31][32] and state-of-the-art eutectic electrolytes featuring no free water 5,12,13 (Fig. 1a and Fig. S3, ESI †). The superior properties of our HCZE electrolyte play an essential role in enabling high-rate, stable Zn 2+ storage using the NVO cathode, as discussed later. Even at a high concentration (n water : n cation = 3), the low viscosity and high ionic conductivity of HCZE is related to the absence of bulky organic ligands (e.g., acetate and trifluoromethanesulfonate) and additives (e.g., urea, acetamide, and succinonitrile) that are typically utilized to prepare WIS and eutectic electrolytes, respectively. Additionally, the presence of a minor amount of ''free'' water molecules in HCZE, as confirmed in the following section, induces fast exchange between ClO 4 À /water and highly concentrated Na + to ensure high ionic conductivity. Therefore, our HCZE electrolyte shows a pronounced improvement in the trade-off between ionic conductivity and the amount of free water molecules in aqueous electrolytes.</p><p>As an important parameter to evaluate the rechargeability of Zn metal anodes, CE was determined using an asymmetric Ti/Zn half cell at a capacity of 0.4 mA h cm À2 , in which a Ti foil</p><p>was used as the working electrode and Zn metal foil as the counter and reference electrodes. Smooth and stable voltage profiles can be realized in HCZE (Fig. S4, ESI †), along with a high average CE of 98.2% calculated from cycle 1 to 100 at 0.4 mA cm À2 (Fig. 1b). The electrochemical performance of Zn plating/stripping achieved in HCZE outperforms that of the conventional low-concentration electrolytes, including 0.5 m Zn(ClO 4 ) 2 and 2 m ZnSO 4 (Fig. S5, ESI †). The improved performance indicates that the plated Zn on the Ti substrate is efficiently recovered in the following stripping process, and parasitic reactions (e.g., hydrogen evolution) can be significantly prohibited. 5,11,33 In dilute electrolytes (e.g., 0.5 m Zn(ClO 4 ) 2 ), Zn 2+ ions are fully hydrated by dipolar water molecules to form [Zn(OH 2 ) 6 ] 2+ . 34 Such Zn 2+ -H 2 O interaction would greatly weaken the O-H bond of water molecules, 35 leading to the deprotonation of H 2 O to generate an acidic electrolyte.</p><p>The deprotonation process is confirmed by the mildly acidic nature of 0.5 m Zn(ClO 4 ) 2 (pH B 5). In contrast, HCZE shows a high concentration of solute, which can efficiently mitigate the parasitic reactions like hydrogen evolution, corrosion, and passivation that can compete with Zn plating/stripping. Accordingly, HCZE enables a remarkable stability of Zn/Zn 2+ redox processes with high efficiency, superior to those of the dilute control electrolytes (Fig. S5 and S6, ESI †). Additionally, cyclic voltammetry (CV) tests of Zn plating/stripping (Fig. 1c) reveal that compared to the low-concentration electrolyte, HCZE enables decreased voltage offset and onset potential for Zn deposition/dissolution, suggesting improved reversibility and kinetics in the concentrated electrolyte.</p><p>The stability and reversibility of the Zn metal anode in our HCZE electrolyte were further evaluated in a Zn/Zn symmetric cell under galvanostatic conditions. As shown in Fig. 1d, the symmetric cell exhibits a long-term stability at a low current density of 0.2 mA cm À2 without signs of potential fluctuation or cell short-circuiting. This is expected given the enhanced CE of Zn plating/stripping and the scarcity of free water molecules in HCZE. The overpotential is even smaller than that in dilute aqueous electrolytes of ZIBs (Fig. S7, ESI †), suggesting the improved Zn plating/stripping kinetics in our HCZE electrolyte. These results reveal the capability of HCZE in enabling stable and reversible Zn chemistry.</p><p>To investigate the Zn plating/stripping mechanism in more detail, Zn was plated with a capacity of 1 mA h cm À2 onto the Zn electrode (0.5 mA cm À2 ) and then stripped away, followed by multiple plating/stripping cycles with a capacity of 0.5 mA h cm À2 at every half cycle. The scanning electron microscopy (SEM) image shows the corrosion and dendrite formation on the Zn electrode in 0.5 m Zn(ClO 4 ) 2 (Fig. 1e), which is the reason for cell shortcircuiting and premature failure. In sharp contrast, Zn foil shows a relatively flat surface without obvious dendrites during cycling tests in HCZE (Fig. 1e and Fig. S8, ESI †). X-ray diffraction (XRD) patterns of the Zn electrodes at different cycles show the typical diffractions from the hexagonal phase of zinc (Fig. 1f) when using HCZE, which is different from the formation of undesired byproducts (e.g., Zn 4 ClO 4 (OH) 7 , ZnO) upon Zn plating/stripping in 0.5 m Zn(ClO 4 ) 2 (Fig. S9, ESI †).</p><p>X-ray photoelectron spectroscopy (XPS) reveals the generation of a characteristic 2p 1/2 and 2p 3/2 doublet that is attributed to Cl À of ZnCl 2 (Fig. 1g). This indicates the efficient reduction of ClO 4 À to form a Cl À -containing layer on the surface of Zn, 36 which could ensure a potential gradient across this surface layer to allow Zn plating underneath. This is in a much similar manner to the LiCl-containing solid-electrolyte interface (SEI) layer on Li metal towards stable Li plating/stripping. 37 Although ClO 4 À reduction could happen in 0.5 m Zn(ClO 4 ) 2 , other competitive reactions like corrosion and hydrogen evolution 1,2 led to the formation of a low-quality protection layer (Fig. S10 and S11, ESI †). This results in low coulombic efficiency and fast cell degradation upon using the low-concentration electrolyte of 0.5 m Zn(ClO 4 ) 2 . Such findings demonstrate the synergistic effect from high-concentration ClO 4</p><p>À and limited ''free'' water in HCZE to achieve stable and reversible Zn deposition. Given the high-entropy state of ClO 4 À and a minimal amount of free water molecules in HCZE, the insoluble Cl À -containing layers most likely exist in the form of stable and coordination-saturated ZnCl 2 -based species. 5,37 As a consequence, the solvation structure of HZCE on uniform Zn deposition is dependent on Zn 2+ desolvation from the active ZnCl 2 -associated Zn solvates and prohibited side reactions at the interface (see the theoretical calculations later).</p><p>The unusual interface formed on the Zn metal anode is closely related to the state changes of electrolytes in HCZE. We analyzed the interaction among Zn 2+ , Na + , ClO 4 À , and H 2 O molecules using Raman, Fourier-transform infrared spectroscopy (FT-IR), differential scanning calorimetry (DSC), and nuclear magnetic resonance (NMR) spectroscopy. The Raman spectra show strong O-H stretching vibration modes of water molecules in dilute solutions (Fig. 2a), which is attributed to various hydrogen-bonding environments in water clusters. 38 Increasing the concentration of solutions causes a pronounced blue shift and peak intensity weakening of the typical O-H vibrations at 3250 and 3400 cm À1 (Fig. 2a and Fig. S12, ESI †).</p><p>A similar phenomenon has been found in the FT-IR spectra of the electrolytes upon increasing the concentrations (Fig. 2b and Fig. S11, ESI †). A new sharp peak at 3550 cm À1 in the Raman spectra can be assigned to the signature of crystalline hydrates. 5,12 These changes suggest that the minimized quantity of free water molecules has been stabilized in the concentrated electrolytes, namely, the extensive disruption of the water network connected through hydrogen bonding.</p><p>Thermodynamically, HCZE is a stable liquid at room temperature (Fig. S13, ESI †). The dilute electrolyte of 0.5 m Zn(ClO 4 ) 2 shows clear melting transitions at À3.9 1C (Fig. 2c), characteristics of low-concentration solutions with a completely dissociated salt. 39 Differently, HCZE exhibits an exothermic peak at À68.5 1C, which is attributed to the crystal formation of a salt-water eutectic mixture. Further heating ceases the continuous rearrangement of the eutectic mixture, thereby leading to the melting process after crystal formation. The two endothermic peaks are similar to the features of eutectic electrolytes with a minimum amount of free water molecules. 24,25,30 The aforementioned observations reveal that the solvation shells of different ions interpenetrate when reaching the limit of high concentration. We further conducted liquid-state NMR to study the aggregation behaviors of the two salts in HCZE (Fig. 2d, e, f and Fig. S13, ESI †). Compared to the water reference and diluted solutions, the 17 O (water) response of HCZE shows a pronounced upfield shift (Fig. 2d). An insignificant peak broadening can be assigned to a slightly decreased T 2 relaxation time, 13 which in turn confirms the low viscosity of HCZE. This is consistent with our previous analysis and in sharp contrast to those of the classical WIS electrolytes using expensive fluorine-containing anions (Fig. 1a). Notably, the 17 O (ClO 4 À ) signal of HCZE shows a slight shift compared to the diluted reference electrolyte (Fig. S14, ESI †). These results reflect that the solvation structures of different ions strongly interpenetrate in the concentrated electrolyte.</p><p>As evidenced by Raman and FT-IR, HCZE shows a very limited amount of free water molecules, wherein most of the water molecules would be a part of cation solvation shells. Due to the higher charge density of Zn 2+ , 11,13 Zn 2+ possesses a stronger binding ability to water and ClO 4</p><p>À compared to Na + .</p><p>In the most concentrated electrolyte, the Zn 2+ solvation shells would be primarily occupied by ClO 4 À , 11 while the interaction between the lone-pair electrons on water oxygen and Zn 2+ could be efficiently suppressed, leading to the shielding effect of 17 O (water) and 67 Zn resonance signals (Fig. 2d and e). For the concentrated electrolytes, the line width of the 67 Zn signal increases marginally, further suggesting the enhanced interaction and complexation between Zn 2+ and the abundant anions in the electrolytes. On the other hand, the presence of water and ClO 4 À in the coordination shell of Na + results in increased electron density around Na + , thus leading to the upfield peak shift of 23 Na for HCZE (Fig. S14, ESI †). The relatively narrow linewidth of the 23 Na NMR signal is associated with the fast exchange between Na + and ClO 4 À /water in the solvation shell, 40 which ensures the high ionic conductivity of HCZE.</p><p>The Hofmeister series indicates that ClO 4 À salts feature a strong capability to destabilize the bulk water structure, implying the probability of adjusting solvation-sheath structures of ion aggregations. At a high concentration for HCZE (0.5 m Zn(ClO 4 ) 2 + 18 m NaClO 4 ), the water-breaking ClO 4 À ions can form complex ion networks with water molecules and cations, giving rise to unique solvation structures and disrupting the formation of hydrogen bonding networks. Therefore, we conducted 35 Cl NMR to study the role of ClO 4 À in changing the solvation-sheath structure of cations (Fig. 2f). The upfield peak shift for highly concentrated electrolytes is indicative of ion shielding. Compared to Na + , the stronger binding capability of Zn 2+ to ClO 4 À could deliver an enhanced ion shielding effect.</p><p>Our findings are in good agreement with previous molecular dynamics (MD) simulations, [11][12][13]24,25 which demonstrated that a significant increase in the number of ion aggregates could prevent the formation of an extended hydrogen-bonding network of water molecules in concentrated electrolytes.</p><p>To further obtain insights into the solvation structures, we performed ab initio MD simulations on our electrolytes with three representative concentrations at different temperatures. Compared to Na + , the stronger binding ability of Zn 2+ towards H 2 O and ClO 4</p><p>À results in a longer residence time of ClO 4 À in the vicinity of Zn 2+ . Therefore, a series of MD simulations were conducted from 300 K (room temperature) to 350 K and 400 K Energy & Environmental Science Paper (overheated system), aiming at accelerating the dynamics and ensuring the equilibrium of the Zn 2+ solvation-sheath structure (Fig. 3 and Fig. S15, S16, ESI †). In a typical dilute solution of 0.5 m Zn(ClO 4 ) 2 (Fig. 3a), Zn 2+ is expected to be solvated by six water molecules with insignificant contribution from ClO 4 À . The remaining water molecules generate a network through hydrogen bonding. Upon increasing NaClO 4 to the intermediate concentration (i.e., 0.5 m Zn(ClO 4 ) 2 + 9 m NaClO 4 ) (Fig. 3b), the ClO 4</p><p>À ions penetrate into the first solvation shell of Zn 2+ .</p><p>The formation of cation-anion aggregates has been confirmed to be beneficial for anion reduction on the anode rather than water, 24,41 which is vital for prohibiting the side reaction on the Zn anode. For the electrolyte with the highest concentration (0.5 m Zn(ClO 4 ) 2 + 18 m NaClO 4 ) (Fig. 3c), the cationic hydration sheaths demonstrate a high degree of interpretation with a significantly decreased amount of free water molecules.</p><p>The first hydration shell of Zn 2+ is mainly occupied by the ClO 4 À ions with four oxygen atoms from ClO 4 À . Radial distribution functions (RDFs) were further performed to illustrate the distributions of neighboring molecules of solvated Zn 2+ ions (Fig. 3d). RDFs of 0.5 m Zn(ClO 4 ) 2 indicate that water molecules dominate both the first and second solvation shells of Zn 2+ , as revealed by two sharp peaks at 2.05 Å and 4.25 Å. In contrast, two main peaks at 2.05 Å and 3.04 Å appear in the most concentrated electrolyte. This suggests that the water molecules replaced by ClO 4 À enter the second hydration shell, wherein the ClO 4 À ions still contribute significantly to the second shell. In combination with Raman and FT-IR spectra (Fig. 2a and b), we confirm that the high concentration of cations in HCZE can efficiently decrease the number of water molecules at a free state. This is an important indication to prevent cathode dissolution when using HCZE in aqueous Zn-ion batteries.</p><p>We further conducted density functional theory (DFT) calculations to investigate the influence of solute concentration on the electronic structures of our electrolytes. For the dilute electrolyte of 0.5 m Zn(ClO 4 ) 2 (Fig. 3e), the lowest unoccupied molecular orbital (LUMO) is primarily contributed by water. Accordingly, water molecules would be predominantly reduced on the Zn anode, leading to fast capacity degradation of the battery. At the highest concentration, the ratio of contact ion pairs and aggregates increases, along with a drastically decreased amount of free water molecules. The extensive coordination of water to cations leads to a significant downshift of the orbital levels of ClO 4 À , delivering a dominant contribution of ClO 4</p><p>À in both the LUMO and the highest occupied molecular orbital (HOMO) levels (Fig. 3f). As a result, ClO 4 À is predominantly reduced to yield an anion-derived protection layer. Similar features have also been found for different WIS and eutectic electrolytes such as ZnF 2 16 and NaF. 25 The anionderived passivation layer has been proven to show a low interfacial resistance, 12,16,24 thus ensuring fast kinetics of Zn plating/stripping. Noticeably, the passivation layer of electrically insulating ZnCl 2 could potentially establish a potential gradient across the layer to drive Zn 2+ diffusion through the passivation layer, which allows for uniform Zn deposition. The function of the ZnCl 2 layer is similar to that of the LiCl-containing SEI layer on the Li metals. 37,42 The passivation layer of ZnCl 2 could efficiently prohibit parasitic reaction, as evidenced by our DFT calculations (Fig. 4). The adsorption energy of H 2 O on ZnCl 2 (À0.87 eV) is lower than that on Zn (À0.04 eV) (Fig. 4a). The H 2 O molecule could interact with both Zn and ZnCl 2 surfaces via the Zn-O bond, which can be characterized by the electron accumulation (green area) located above the ZnCl 2 surface and the electron depletion (purple area) around the O atom. The intensity of the charge redistribution further reveals that the Zn-O bond in the case of ZnCl 2 is considerably stronger than that on the Zn surface. Accordingly, the passivation layer induces a higher energy barrier for water dissociation towards hydrogen evolution. 43,44 Moreover, the insulating feature of ZnCl 2 makes it difficult to obtain electrons from the Zn anode, thus making it kinetically sluggish to proceed hydrogen evolution and the corresponding formation of OH À on the Zn anode. This would significantly mitigate the generation of OH À -involved detrimental byproducts (e.g., Zn 4 ClO 4 (OH) 7 and ZnO) on the Zn anode. Compared to the Zn 2+ adsorption on the Zn metal surface (Fig. 4b), the lower adsorption energy of Zn 2+ on ZnCl 2 could promote the in-plane diffusion of Zn ad-atoms with a low diffusion barrier, similar to the roles of metal chlorides or fluorides in the SEI of different metal anodes (e.g., Li, Zn, and Al). 16,37,45,46 This is also confirmed by the larger charge density difference at the Zn 2+ /ZnCl 2 interfacial region, where the electron accumulation appears above the ZnCl 2 surface and the electron depletion is localized in the Zn 2+ ion. These metal chlorides or fluorides have been acknowledged as the primary rigid-frame materials to protect metal anodes, wherein they can induce the homogeneous metal nucleation and thus inhibit dendrite growth efficiently. The schematics of interfacial reactions between the Zn anode and electrolytes are shown in Fig. 4c and d. The Zn 2+ ions in the most concentrated electrolyte show much higher desolvation energy compared to that of [Zn(H 2 O) 6 ] 2+ in the dilute electrolyte (Fig. S17, ESI †). Therefore, the presence of fragile [Zn(H 2 O) 6 ] 2+ in the dilute electrolyte causes different irreversible parasitic reactions like hydrogen evolution and corrosion. These reactions could compete with Zn plating/ stripping at low potentials, thereby promoting Zn dendrite formation and decreasing the coulombic efficiency and cycling stability. However, the primary solvation shell of Zn 2+ in the most concentrated electrolyte is mainly occupied by the ClO 4 À ions. The reduced affinities between Zn 2+ and H 2 O, as a result of the strong coordination of ClO 4</p><p>À to Zn 2+ , could efficiently mitigate the parasitic reactions on the Zn anode. As such, the synergistic effect of the ZnCl 2 passivation layer from perchlorate anion reduction and the intimate Zn 2+ -ClO 4 À interaction through increasing Zn 2+ desolvation energy could enable stable Zn plating/stripping with high efficiency. Even if the substitution of water by perchlorate ions occurs for Zn 2+ solvation, water molecules would be released from the first solvation shell but most like stay in the second solvation shell through a persistent but relatively weak attraction with the Zn 2+ center. Accordingly, the improvement of anodic stability can be rationalized. Of note is that the high-concentration feature of HCZE provides abundant internally induced dipole interaction between H 2 O and ClO 4</p><!><p>. This effect could saturate the bipolar sites of water molecules, 47,48 further depressing their activity. Given the aforementioned features of our HCZE electrolyte, it is promising to suppress the dissolution of oxide cathodes (e.g., vanadate nanostructures) in conventional low-concentration electrolytes of aqueous ZIBs.</p><p>To demonstrate the feasibility and efficiency of HCZE in aqueous ZIBs, we developed NVO nanofibers as the cathodes to pair with Zn metal anodes in HCZE. NVO nanofibers were fabricated via an aqueous processing technique, which is inexpensive, green, and easy for mass production (see synthesis details in ESI †). Here we demonstrate a batch capability of 100 g NVO using a 1 L reactor (Fig. S18, ESI †). Our technique can be potentially scaled up using larger reactors, which shows the first example to produce cathodes of aqueous ZIBs at scale and low cost. NVO shows nanofiber morphology with high purity, aspect ratio, and homogeneity (Fig. S19, ESI †). The monoclinic structure of NVO is constructed from the V 3 O 8 layers featuring two different zigzag chains (chains of edgesharing V 3 O 8 square pyramids and chains of distorted VO 6 octahedra) (Fig. S19, ESI †), while hydrated Na ions are located among the layers. The layered structure and structural flexibility of NVO can enable a good electrochemical performance towards reversible cation (e.g., Zn 2+ ) intercalation. As the crystalline structures of monoclinic NVO and orthorhombic V 2 O 5 have no structural relationship, it is unlikely that a solid-state transformation led to the formation of NVO in aqueous solutions. A dissolution-recrystallization process for the structural transformation has been confirmed (Fig. S20-S25, see Supplementary Note 1 for a detailed analysis, ESI †).</p><p>HCZE shows high ionic conductivity, scarcity of free water molecules, and capability to enable stable Zn plating/stripping with high efficiency. The cyclic voltammetry (CV) curve of the NVO electrode shows two pairs of peaks (Fig. 5a), which can be assigned to the redox reactions between V 5+ and lower oxidation states during charge/discharge. The electrochemical redox 49,50 During the charge/discharge process, no features related to Na + insertion/removal could be observed even though a supporting salt of NaClO 4 at a high concentration was used, as evidenced by the ex situ Na 1s core level (Fig. S27, ESI †). Additionally, the Na 1s core levels remain unchanged at fully discharged and charged states, suggesting the stable and immobile Na + pillars in the layered structures of NVO upon charge/discharge. 51 Similar results have been found in the literature using ''water-in-bisalt'' electrolytes. 11,13 The absence of Na + intercalation in our case can be explained by the larger cation size of Na + (1.02 Å) than that of Zn 2+ (0.74 Å) (Fig. S28, ESI †). 13,20 Galvanostatic tests were further carried out to evaluate the rate capability of the NVO electrode (Fig. 5b). The predominantly sloping charge-discharge plots signify a solid-solution type process of cation (de)intercalation. At 0.1 A g À1 , a high capacity of 253 mA h g À1 can be obtained. Increasing the rate to 4 A g À1 delivers a decent capacity of 94 mA h g À1 (Fig. 5b). The rate capability of our NVO cathode in HCZE outperforms the intercalation cathodes in conventional WIS and eutectic electrolytes. [11][12][13]16,26 This points to the synergistic effect between the high ionic conductivity of HCZE and the flexible cationconducting structural framework of the NVO electrode.</p><p>As another critical parameter to evaluate the self-discharge of a battery, 52,53 the voltage decay of the NVO electrode on the shelf was recorded upon fully charging to 1.5 V at 0.1 A g À1 . After resting for 15 days, 80.9% of the original capacity can be maintained, apparently superior to the performance using conventional dilute electrolytes (Fig. S29, ESI †). This low selfdischarge rate suggests that the parasitic reactions (e.g., Zn corrosion and water dissociation) and dissolution of the active cathode are efficiently suppressed in our battery (Fig. S1, ESI †). The evaluation of cycling performance is another critical approach to study the effectiveness of HCZE in prohibiting the side reactions and cathode dissolution. As expected, nearly no capacity decay can be observed over 2000 cycles at 4 A g À1 (Fig. 5c), featuring high coulombic efficiency of 499%.</p><p>Coincident with the electrochemical performance, the nanofiber morphology of our NVO electrode can be maintained after cycling tests (Fig. S30, ESI †). As such, the robust stability of the NVO cathode is strongly associated with the minimized amount of free water molecules to inhibit irreversible parasitic reactions and the flexible, stable layered framework of NVO. We further conducted different ex situ techniques, including XRD, X-ray absorption spectroscopy (XAS), XPS, solid-state NMR, FT-IR, SEM, and transmission electron microscopy (TEM), to illustrate the electrochemical energy storage mechanism of NVO in HCZE. Ex situ XRD patterns of the NVO cathode demonstrate a slight contraction of the gallery spacing of the (001) plane (2y = 11.11) upon discharge to 0.65 V (Fig. 5d and Fig. S31, ESI †), indicating the decrease of interlayer spacing. This is related to an improvement of structural coordination because of the strong electrostatic interaction between the intercalated cations and the V 3 O 8 bilayers of NVO. In the subsequent discharge region of 0.65-0.3 V, we observe the continuous shrinkage of interlayer spacing, which is accompanied by the evolution of a set of new peaks. These diffractions can be assigned to Zn 4 ClO 4 (OH) 7 (Fig. S31, ESI †), 5,29,54 implying a proton intercalation process. The diffractions of Zn 4 ClO 4 (OH) 7 disappear upon the potential reversal to 1.5 V, along with the recovery of the NVO phase. To reveal the electronic structure of vanadium, we collected V L-edge and V K-edge X-ray absorption spectra for the NVO cathode at certain charging/discharging voltages (Fig. 5e and Fig. S32, ESI †). The line shape and intensity of the V L 3 -edge at 510-520 eV are used to determine the oxidation state of vanadium. The increasing intensity ratio of 516.8 eV/514.5 eV peaks for NVO electrodes upon charging to 1.5 V indicates higher oxidation of vanadium at charged states (Fig. S31a, ESI †). V K-edge X-ray absorption near edge structure (XANES) spectra show negative shifts of the adsorption edge upon discharge, which reflects the reduction of vanadium (inset of Fig. 5e). During the recharge process, the absorption edge shifts back to higher energy indicating the oxidation of vanadium. Both V L-edge and K-edge spectra reveal the reversible changes of V oxidation states during charge and discharge, which is consistent with ex situ XPS analysis (Fig. S33, ESI †). Wavelet transform analysis shows that the V 3 O 8 bilayers of NVO are well maintained during charge and discharge (Fig. S32, ESI †). The framework stability of NVO is associated with the stable Na + pillars featuring highly negative formation energy in the presence of structural water. 51,55 Accordingly, the reversible contraction/expansion of the gallery spacing of NVO during discharge/charge plays a vital role in ensuring its long-term cycling stability.</p><p>Typically, H + co-intercalation with Zn 2+ occurs in conventional electrolytes of aqueous ZIBs. The consumption of H + from water dissociation leads to the formation of electrolyte-dependent products on the cathodes, e.g., Zn 4 SO 4 (OH) 6 ÁxH 2 O in ZnSO 4 , 29 Zn 5 (OH) 8 Cl 2 ÁxH 2 O in ZnCl 2 , 26 and triflate-containing Zn-based layered double hydroxides (LDHs) in Zn(CF 3 SO 3 ) 2 . 56 During discharge, the H + -intercalation byproduct of Zn 4 ClO 4 (OH) 7 sheets could be formed on the surface of NVO in HCZE (Fig. 5g and Fig. S34-S37, ESI †). Ex situ 1 H NMR spectra show noticeable yet reversible changes in intensity, peak position, and line width of the 1 H signal that is attributed to structural water (6.5 ppm for the pristine NVO) (Fig. 5f). This indicates the participation of structural water in assisting H + and Zn 2+ intercalations through providing an electrostatic shielding effect. 55,57 The peaks at 1.1 and 5.4 ppm correspond to the intercalated protons. Another reversible peak at 1.8 ppm is assigned to the proton in Zn 4 ClO 4 (OH) 7 . 28,29 Combining the reversible changes of Zn and Cl of energy-dispersive X-ray spectroscopy (EDS) with SEM and TEM at different states (Fig. S36, S37 and Table S1, ESI †), the simultaneous insertion/extraction of H + and Zn 2+ can be confirmed. Such observation is different from previous aqueous ZIBs, such as the sole storage of Zn 2+ or H + in the cathodes 26,58 and stepwise intercalation of H + and Zn 2+ . 59 During discharge, the H + -intercalation byproduct (i.e., Zn 4 ClO 4 (OH) 7 ) can also be generated on the NVO cathode in the dilute electrolyte of 0.5 m Zn(ClO 4 ) 2 . This phenomenon, together with the overlapping of galvanostatic charge/discharge curves to that in HCZE, indicates nearly the same energy storage mechanism in these two electrolytes (Fig. S38, ESI †). The difference between low-concentration aqueous electrolyte and HCZE lies in the changes in solvation structures induced by the high concentration of solutes. In the dilute aqueous electrolyte featuring enough available water molecules, all the Zn 2+ ions can be solvated by dipolar water molecules to generate octahedral Zn(H 2 O) 6</p><!><p>. Such cation-water interaction can induce electron departure from the 3a1 orbital of the coordinated water to the empty Zn 2+ orbitals, thus significantly weakening the O-H bond of the coordinated water. 60 For the dilute electrolyte, such effect could induce deprotonation of water to generate a mildly acidic solution (pH B 5 for 0.5 m Zn(ClO 4 ) 2 ). This can be confirmed by a negative dehydrogenation energy of À3.83 eV for the H 2 O in octahedral [Zn(H 2 O) 6 ] 2+ , as calculated by DFT (Fig. S39, ESI †). The water deprotonation can be efficiently suppressed in concentrated electrolytes because of the unique inter-molecular interaction (e.g., Zn 2+ -ClO 4 À and ClO 4 À -H 2 O).</p><p>However, a minor amount of free water molecules still exists in the most concentrated electrolytes (Fig. 2 and 3), which shows three advantages of: 30,61 (1) ensuring the high ionic conductivity for high rate performance through decreasing the electrolyte viscosity, (2) minimizing cathode dissolution to ensure cycling stability, and (3) featuring hydrogen bonding networks and providing sources for proton co-intercalation into NVO, which is confirmed to be thermodynamically feasible (Fig. S40, ESI †) (see the detailed energy storage mechanism in Supplementary Note 2, ESI †). Compared to the previously reported highly concentrated dual-cation aqueous electrolytes (e.g., Zn-Li hybrid battery 11 and Li-ion battery 13 ), our battery chemistry shows two advantages:</p>

Royal Society of Chemistry (RSC)

Extraction, purification, kinetic and thermodynamic properties of urease from germinating Pisum Sativum L. seeds

BackgroundUrease, one of the highly efficient known enzymes, catalyzes the hydrolysis of urea into ammonia and carbon dioxide. The present study aimed to extract urease from pea seeds (Pisum Sativum L). The enzyme was then purified in three consequence steps: acetone precipitation, DEAE-cellulose ion-exchange chromatography, and gel filtration chromatography (Sephacryl S-200 column).ResultsThe purification fold was 12.85 with a yield of 40%. The molecular weight of the isolated urease was estimated by chromatography to be 269,000 Daltons. Maximum urease activity (190 U/g) was achieved at the optimum conditions of 40°C and pH of 7.5 after 5 min of incubation. The kinetic parameters, K m and V max , were estimated by Lineweaver-Burk fits and found to be 500 mM and 333.3 U/g, respectively. The thermodynamic constants of activation, ΔH, E a , and ΔS, were determined using Arrhenius plot and found to be 21.20 kJ/mol, 23.7 kJ/mol, and 1.18 kJ/mol/K, respectively.ConclusionsUrease was purified from germinating Pisum Sativum L. seeds. The purification fold, yield, and molecular weight were determined. The effects of pH, concentration of enzyme, temperature, concentration of substrate, and storage period on urease activity were examined. This may provide an insight on the various aspects of the property of the enzyme. The significance of extracting urease from different sources could play a good role in understanding the metabolism of urea in plants.

extraction,_purification,_kinetic_and_thermodynamic_properties_of_urease_from_germinating_pisum_sati

Background<!>Methods<!>Urease extraction and purification<!>Acetone precipitation<!>DEAE-cellulose chromatography<!>Gel filtration chromatography<!>Enzyme characterization<!>Determination of molecular weight of urease<!>Effect of pH on the activity of pisum sativum urease<!>Effect of storage at −4°C on enzyme activity<!>Effect of different concentration of enzymes<!>Effect of temperature<!>Thermodynamic studies<!>Effect of different concentration of substrates<!>Purification of urease<!><!>Purification of urease<!><!>Molecular weight determination<!>Effect of pH on the activity of pisum sativum urease<!><!>Effect of storage period<!>Effect of enzyme concentration on the activity of pisum sativum urease<!>Effect of temperature on enzyme activity<!>Thermodynamic parameters (Ea, ∆H, and ∆S)<!><!>Effect of substrate concentration on the activity of pisum sativum urease<!><!>The kinetics constants (Km and Vmax)<!>Discussion<!>Conclusions<!>Abbreviations<!>Competing interests<!>Authors’ contributions<!>Acknowledgements

<p>Catalyzing the hydrolysis of urea into ammonia and carbon dioxide, ureases (urea amidohydrolases, EC 3.5.1.5) are a one of known highly efficient enzymes that belong to amidohydrolase and phosphotriesterase superfamily [1]. Several reports have been published on the extraction of urease from various bacteria [2-4], and plants [5-11]. The high molecular mass nickel-containing metalloenzyme [12] is believed to play an important role in the nitrogen transport cycle in plants [13]. In addition, the enzyme decomposes urea formed from arginase that is found in seed germination [14]. Urease is also important in human bodies due to the fact that many urinary tract and gastroduodenal diseases [15,16], including cancer [17,18], are related in some ways to this enzyme. The increased need in finding proper ways to remove urea from different environments brought great attractions in the biotechnology field [19]. Some of urease's applications include treatment of industrial waste [20], the industry of alcoholic beverages [21], use in haemodialysis [22,23], and its potential use in space missions as life supporter [24].</p><p>The plant and fungal ureases are homo-oligomeric with identical proteins repetition. On the other hand, the bacterial ureases are composed of complex repetitions of two or three subunits of different sizes [25]. The crystal structure of protein is often the key to its enzyme function. This configuration is governed by its primary structure and environment. Any environmental factor, that alters the shape of the enzyme or blocks the access to the active site in substrate, will affect enzyme activity. Such environmental factors include matrix salt concentration, pH, temperature, substrate concentration, activators, and inhibitors [26,27].</p><p>The purpose of this study is to extract, purify, and characterize urease from plant source (pisum sativum L). Also, the activity of the enzyme was evaluated based on the change of the environmental factors that carried out during the purification procedure. This may provide an insight on the various aspects of the property of the enzyme.</p><!><p>Pisum Sativum L. seeds were obtained from faculty of agriculture, Kafr Elshaikh University, Kafrelshaikh city, Egypt. The seeds were soaked in distilled water for 6 hours, germinated in the dark at 22°C for 2, 4, 6, 8, 10, 12, 14 and 16 days. The germinated seeds were stored separately in deep freezer (−20°C) for further experimental purposes. Dextran polymer particles (Sephadex G-200), bovine serum albumin (BSA), standard proteins, and DEAE-cellulose were purchased from Sigma Chemicals Ltd., USA. All other chemicals used for this research were of analytical grade. All absorbance measurements were performed using Lambda 35 PerkinElmer UV/V is spectrometer.</p><!><p>Unless mentioned otherwise, all of the following procedures were done at 4°C. Ten grams of germinated seeds of pisum sativum were pasted in a mortar and pestle and then suspended in 40 mL of 20% chilled acetone (−20°C). Occasional stirring for 3 h was required. Double layer cheese cloth was used for filtrating of the suspension. After 15 minutes of centrifuging of the filtrate, the supernatant was isolated and used as "crude extract".</p><p>The urease assay was performed as described by Sharma et al. [28]. Enzyme extract (0.25 μL) was added to 10 mL of urea solution (0.4 g urea in 25 ml of phosphate buffer). One millilitre of the previous solution was added to each test tube containing 5 mL of Nessler's reagent, and incubated at 40°C for 5 min. They were followed by the addition of 1.0 M HCl to terminate the reaction after specific time. Absorbance measurements were taken for the resulting solutions (at 405 nm). The estimation of urease was carried out using the standard curve of ammonium sulphate. One unit of urease activity is defined as "the amount of enzyme required to liberate 1.0 μM of NH3 from urea per minute at pH 7.5 and temperature 40°C" [29].</p><p>Proteins were determined according to Lowry et al.[30] using BSA as standard material. Different concentrations of BSA were prepared ranging from (0 to 25) μg/mL. The linear calibration curve was used to determine the concentration of protein in the assay and estimated for the original sample.</p><p>The enzyme was purified to homogeneity by the following successive steps which carried out at 4°C:</p><!><p>The "crude extract" was adjusted to 50% saturation by addition of acetone (chilled to −20°C) under constant and gentle stirring. The resulting precipitate was centrifuged, collected, dissolved in minimum volume of pre-cold 50 mM phosphate buffer (pH = 7.4), and finally dialyzed against the same buffer for 24 h. The resulting solution was then centrifuged for 10 min and the clear supernatant was designated as "crude enzyme solution".</p><!><p>The "crude enzyme solution" was dialyzed against 50 mM phosphate buffer, pH 7.4. It was then loaded on pre-equilibrated DEAE-cellulose column (15 × 3.0 cm) (with 50 mM phosphate buffer, pH 7.8). The bound proteins were eluted with a linear gradient of NaCl (100 – 500 mM), prepared with phosphate buffer, pH 7.8, at a flow rate of 0.5 mL min−1. After collecting the active fractions (the fractions that shows urease activity), the proteins binding to the column were eluted using gradient of (0–0.5 M) KCl and (20 mM) phosphate buffer, pH 7.5. The absorbance of these fractions was measured at wavelength of 280 nm. The active fractions were combined and the volume was measured for the determination of the urease activity and protein contents in the assay.</p><!><p>The enzyme obtained from the ion exchange step was concentrated with acetone and loaded on the Sephacryl S-200 column (1.5 × 65) at a flow rate 30 mL/h using 50 mM phosphate buffer, pH = 7.8. Five millilitre eleunts were collected. The enzymatically active fractions were pooled and dialyzed against 50 mM phosphate buffer, pH 7.4 for 24 h. The absorbance measurements at 280 nm were used to determine the protein concentration and urease activity.</p><!><p>The isolated enzyme activity was characterized and studied as a function of pH, temperature, storage period, enzyme concentration, and substrate concentration using the following procedures:</p><!><p>Two millilitres of blue dextran-2000 solution (6 mg into 3 mL of PEM buffer pH 7.5, 0.1 M PIPES, 1 mM EGTA, 1 mM MgSO4, at pH = 6.6; where PIPES is piperazine-N, N′-bis(2-ethanesulfonic acid, and EGTA is ethylene glycol-bis-(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid) were passed through Sephacryl S-200 column. 20 mM of PEM buffer pH 7.5 was added. Fractions of 5 ml were eluted and the absorbance at 600 nm for each fraction was measured. The column void volume (V 0 ) was determined by estimating the total volume of the fractions characterized with the starting point movement of the dextran to climax of absorbance of the blue dextran. Same procedure was done for the standard proteins; BSA, aldolase, catalase, ferrtin, and thyroglobulin. The eluted fractions, which give a maximum absorbance at 280 nm, were determined, and the eluted volume (V e ) was calculated for each standard protein. The linear calibration curve of VeV0 against logarithm value of molecular weight of standard protein was plotted. The curve was used for determining the molecular weight of native urease.</p><!><p>The pH profile for the purified urease was estimated using urea as a substrate. The pH range used was from 3 to 10 using 50 mM phosphate buffer.</p><!><p>To determine the effect of storage of the enzyme on the urease activity, the enzyme was stored at different time internals of 0–60 days. The enzyme activity was measured after each separate time period.</p><!><p>The optimum enzyme concentration was determined by varying the amount of the pure enzyme.</p><!><p>The optimum temperature for urease activity was determined over the temperatures from 10 to 40°C using the standard conditions of the assay.</p><!><p>The relationship between the rate of an enzymatic reaction and activation energy is given by the empirical formula of the Arrhenius equation:</p><p>(1)Ea=R⋅lnV2V1⋅1T1‒1T2</p><p>where V 1 and V 2 are the enzyme activities at the temperatures T 1 and T 2 ; E a is the energy of activation (kJ mol−1) which can be determined from the slope of the Arrhenius plot of ln(V) against 1T.</p><p>The activation enthalpy (ΔH) can be calculated by eqn. 2.</p><p>(2)ΔH=Ea−RT</p><p>Finally the entropy (ΔS) was calculated by eqn 3 (Eyring-Polanyi), which correlates ∆H, E a , and Arrhenius equation (eqn. 1);</p><p>(3)lnVmaxT=lnKBh+ΔSR−ΔHR⋅1T</p><p>where T, K B , h, and R are absolute temperature, Boltzmann constant, Planck constant and gas constant respectively.</p><!><p>The effect of urea concentration on the activity of enzyme was examined. Urea solution of different concentration was taken in different test tubes and the enzyme activity was measured. K m and V max for urease were calculated using Lineweaver-Burk double reciprocal plot [31].</p><!><p>After 6 days of germination the activity increased gradually and showed maximum activity on the ninth day after germination and then declined rapidly (Data not shown). Therefore, we used the 9th day of germination for further experimental purpose. The results of the purification activity of the germinated pisum sativum seed urease were summarized in Table 1. The total activity which represents the summation of the activity of all proteins in the extract samples showed a decrease over the period of the purification procedure. However, the most important parameter is the specific activity of the extracted enzyme which represents the actual activity of the active proteins only. The results showed an increase in the specific activity throughout the purification steps. The final purification fold achieved was nearly 12.85. The specific activity after finishing the purification was 5833.3 Umg−1.</p><!><p>Yield and purification fold at different steps of purification of urease from germinated pisum sativum seeds</p><!><p>Ion-exchange chromatography was carried out using the anion exchanger DEAE-cellulose. Approximately 20 ml of crude extract was passed through the column. As shown in Figure 1a, four peaks of proteins appeared at fractions of 20, 41, 68, and 85. Peaks of fraction number 20 and 68 showed urease enzyme activity. The values of the urease activity were of 190 unit/ fraction for the fraction number 20 and 73 unit/ fraction for the fraction number 68. The fractions of the first peak were collected for the gel chromatography separation. The fractions that showed enzyme activity from the gel chromatography filtration were concentrated by dialysis and applied to Sephacryl S-200 column. At the end of the purification procedure, one peak for protein and enzyme activity was observed at fraction 23 (Figure 2b).</p><!><p>A typical elution profile for the ion exchange chromatography of Pisum Sativum L. urease from, (a) only DEAE-cellulose column (15 × 3.0 cm), and (b) DEAE-Cellulose followed by Sephacryl S-200 column (90 × 1.6 cm i.d.). is for absorbance at 280 nm, and is for urease activity.</p><p>Calibration curve for standard proteins; bovineserum albumin, aldolase, catalase, ferritin, and thyroglobulin using Sephacryl S-200 column. The empty filled squares are for the standard proteins, and the solid sphere is for the urease extracted from the germinating pisum sativum seeds used in this study.</p><!><p>The molecular weight of urease was determined using gel filtration chromatography to be equal to 269,000 ± 200 Da (linearity regression parameter of 5 standard proteins, R2 = 0.997) as shown in Figure 2.</p><!><p>The pH profiles for the purified urease were estimated using urea as a substrate. The pH range used was from 3 to 10 using 50 mM phosphate buffer. The activity of urease was the highest at pH 7.5 (Figure 3a).</p><!><p>Effect of (a) pH (at incubation temperature = 40°C), (b) time of storage (at −4°C), (c) enzyme concentration (pH = 7.5, temperature = 40°C), and (d) temperature (pH = 7.5), on urease activity extracted from Pisum Sativum L. seeds.</p><!><p>The enzyme activity decreased with time even when stored at −4°C. It represented 100% on the first day but decreased to 80% on the tenth day. However, it retained about 14.1% even after 2 months. The exponential decay fits showed a half life time of 22.4 days (see Figure 3b).</p><!><p>As indicated in Figure 3c, the urease activity increased significantly rapidly by increasing the enzyme concentration until reaches a value of 100 ug/mL. After that the activity kept rising but in slower rate. The maximum activity reached was 102 units/ assay at 200 ug/mL of enzyme.</p><!><p>Figure 3d showed the temperature optimum curve for urease. The complete assays of enzyme were incubated at different temperatures from 10 to 80°C for 10 minute. Results showed that urease had an optimum temperature at 40°C.</p><!><p>The activation energy (Ea) can be determined from the slope of the empirical formula of the Arrhenious plot of natural logarithm of the urease activity versus the reciprocal value of temperature (see Figure 4a) The activation energy was found to be 23.7 kJ/mol. Both enthalpy of activation (∆H) and entropy of activation (∆S) were calculated using Arrhenius plot as shown in Figure 4b and were found to be 21.20 kJ/mol, 1.18 kJ/mol respectively.</p><!><p>Arrhenius plot for the activity of urease that is extracted from germinating pisum sativum L. seeds, for the calculation of (a) activation energy, and (b) entropy of activation and enthalpy of activation.</p><!><p>As indicated in Figure 5a by increasing urea concentration, the activity increased until nearly constant maximum activity 102 units/assay at 200 mM of substrate. Further increase in urea concentration resulted in a gradual decrease in enzyme activity.</p><!><p>Effect of substrate concentration on the activity of urease extracted from pisum sativum L. seeds. (a) Dependence of initial reaction rate on substrate concentration of urease; pH 7.5, temperature 40°C. (b) Lineweaver-Burk double reciprocal plot.</p><!><p>The kinetics constants (K m and V max ) for the purified urease were determined by incubating a fixed amount of enzyme with varying the concentration of urea solution (urea used here as substrate). Km and Vmax for urea were calculated using Lineweaver-Burk double reciprocal plot, Figure 5b, and were found to be 500 mM and 333 U/g, respectively</p><!><p>Urease plays an essential role in the nitrogen metabolism in both germination and seedlings of plants. Two peaks were found to have urease activity. However, the highest activity peak was chosen for further experiments. The low activity peak (high salt peak) has low concentration and low urease ratio that limit our ability to consider it for further experimentation. The fact for having two peaks is unusual. However, as this study is the first to use the germinating pisum sativum seeds, there is a possibility of having more than one isozyme with different elution times and conditions. The results of germinated pisum sativum seed urease, Table 1, shows that the final purification fold achieved was 12.85 and the specific activity was 5833.3 Umg−1. In comparison with other studies, the purification results of germinated chickpea specific activity was 489.57 and the final purification fold was 45 [26]. Also, for Proteus mirabilis urease, the Specific activity of the extracted enzyme was 22932.86 and the final purification fold was 13.86 [32].</p><p>The molecular weight of the pisum sativum seeds urease reported in this investigation was 269,000 Da, compared to 480,000 Da for jack bean [33], and 540,000 Da for dehusked pigeonpea (Cajanus cajan L.) [7]. The molecular weight of enzymes is known to change relative to the source and even stage of plant growth.</p><p>The pH plays an important role in the activity of enzyme. The urease isolated from Pisum Sativum seed was found to yield maximum activity at pH 7.5 (Figure 3a) which means that the seeds may belong to the category of basic urease. Despite the fact that Mulberry leaves have shown neutral optimum pH [3], many other studies reported basic pH as an optimum value for the ectracted urease. For example the optimum pH was found to be 8 in jack beans [34], pigeonpea[7], pathogenic fungus (Coccidiodes immitis)[35], aspergillus niger[36], and bacillus pasteurii[37]. These results may be explained by the fact that acidic pH has an inhibitory effect on the enzyme resulting in reducing its activity. Also, the existence of the active sites in amino acids will be influenced by the change in pH which may alters the ionization of these amino acids [38].</p><p>The optimum temperature, where the greatest urease activity carries out, is equal to 40°C. This result is comparable to several studies, reported by Das et al. [7], and Srivastava et al. [39]. The kinetic energy of molecules increases with an increase in temperature which results in seeding up the rate of reaction. When the temperature was further increased, the molecules of enzyme exceed the barrier of energy. This causes the breakage of hydrogen and hydrophobic bonds that are responsible for maintaining the 3D structure of enzyme [4,40].</p><p>The optimum value of substrate concentration, where the urease activity has the largest value, was found to be 120 mM. After that the activity starts to gradually decrease. The decrease in the activity could be explained by substrate inhibition at higher urea concentrations. The enzyme showed the highest activity when incubated for 5 min under standard conditions; temperature = 40°C and pH = 7.4. The rate of hydrolysis of urea increases with increasing urea concentration until reaching a maximum, beyond that hydrolysis activity starts to decrease [41,42]. Loest [43] and Shepard and Lunceford [44] obtained maximum urease activity at 0.25 M and 0.008 M concentration of urea, respectively.</p><p>The Kinetics constants (K m and V max ) for urease extracted from germinated pisum Sativum was calculated using Lineweaver-Burk double reciprocal plot and were found to be 500 mM and 333.3 U/g respectively. These values indicated a low affinity of substrate to urease.</p><!><p>Urease was purified from germinating Pisum Sativum L. seeds. The purification fold was 12.85 with a yield of 40%. The molecular weight was estimated as: 269,000 Daltons factors adjusted for its action was pH 7.5, temperature 40°C, but further studies are required to elucidate its significance in the metabolism of urea in plants.</p><!><p>DEAE: Diethylaminoethanol; BSA: Bovine serum albumin; PEM buffer: 0.1 M PIPES, 1 mM EGTA, and 1 mM MgSO4; PIPES: Piperazine-N,N′-bis(2-ethanesulfonic acid; EGTA: Ethylene glycol-bis-(β-aminoethyl ether)-N,N,N',N'-tetraacetic acid.</p><!><p>The authors declare that they have no competing interests.</p><!><p>ME and MS conceived the study, carried the purification procedure, activity measurements, and drafted the manuscript. AI participated in the activity measurements, coordination, and helped drafting the manuscript. EA participated in the activity measurements, and coordination. All authors read and approved the final manuscript.</p><!><p>The authors would like to thank King Abdulaziz University, Rabigh College of Arts and Science, Tabuk University and Tanta University for offering the required facilities and funding.</p>

PubMed Open Access

The polyglutamine-expanded androgen receptor has increased DNA binding and reduced transcriptional activity

Expansion of a polyglutamine-encoding trinucleotide CAG repeat in the androgen receptor (AR) to more than 37 repeats is responsible for the X-linked neuromuscular disease spinal and bulbar muscular atrophy (SBMA). Here we evaluated the effect of polyglutamine length on AR function in Xenopus oocytes. This allowed us to correlate the nuclear AR concentration to its capacity for specific DNA binding and transcription activation in vivo. AR variants with polyglutamine tracts containing either 25 or 64 residues were expressed in Xenopus oocytes by cytoplasmic injection of the corresponding mRNAs. The intranuclear AR concentration was monitored in isolated nuclei and related to specific DNA binding as well as transcriptional induction from the hormone response element in the mouse mammary tumor virus (MMTV) promoter. The expanded AR with 64 glutamines had increased capacity for specific DNA binding and a reduced capacity for transcriptional induction as related to its DNA binding activity. The possible mechanism behind these polyglutamine-induced alterations in AR function is discussed.

the_polyglutamine-expanded_androgen_receptor_has_increased_dna_binding_and_reduced_transcriptional_a

<!>Introduction<!>Reagents, plasmids and constructs<!><!>Quantification of intranuclear AR by specific [3H]-R1881 binding<!>Quantification of the MMTV transcription by S1-nuclease protection and specific DNA binding by DMS in vivo footprinting<!>ARQ25 and the ARQ64 expression and quantification in Xenopus oocytes<!><!>Functional comparison of ARQ25 and ARQ64<!><!>No difference between ARQ25 and ARQ64 in androgen agonist- or antagonist-dependent nuclear translocation<!>AR with shorter polyQ repeats, ARQ0 and ARQ13, showed no difference in transcriptional activity compared to ARQ25<!>Final remarks<!>Conflict of interest<!>

<p>Spinal bulbular muscular atrophy is caused by a polyQ expanded androgen receptor.</p><p>Function of AR with expanded polyQ tract was analyzed in Xenopus oocytes.</p><p>AR with expanded polyQ tract has increased DNA binding but reduced gene activation.</p><!><p>Androgenic hormones play a vital role in many biological processes in various parts of the body including reproductive organs, kidney, liver, bone, muscle and brain. They exert their role via binding to the androgen receptor (AR), a ligand-activated steroid hormone receptor that acts as a transcription factor to control the expression of androgen-dependent genes [1]. The N-terminal transactivation domain (NTD) of the AR protein contains a polymorphic polyglutamine (polyQ) tract which has been linked to spinal and bulbar muscular atrophy (SBMA, Kennedy's disease) [2], a disorder characterized by progressive neuromuscular weakness which develops when its length exceeds 37 residues [3]. The expanded polyQ tract in AR has been demonstrated to alter transcriptional activity of AR in different ways in different cell types. Several studies have shown that AR transcriptional activity inversely correlates with the length of this tract [4], [5], [6], [7], however not all reports are in agreement. Thus, it was shown that AR transcriptional activity is positively affected by increasing polyQ repeat length in skeletal muscle cells [8] thus arguing that the effect of an extended polyQ repeat on AR function is context dependent, for example due to interactions with tissue-specific co-activators. Interestingly, the polyQ repeat length also affects AR stability, possibly because of altered protein folding [7], [9] and recent studies demonstrate beneficial effects on the AR polyQ disease in a mouse model by disrupting the SUMOylation of AR [10].</p><p>Although the cause of SBMA is expansion of the CAG repeats in the AR gene the exact disease mechanism remain unclear. We decided to use Xenopus oocytes to look more closely at the function of the AR with an expanded polyQ tract. The large size of these cells allows quantification of intranuclear receptor concentration, sequence specific DNA binding and AR target gene activation [11]. As a gene target we used the enhancer and promoter of the mouse mammary tumor virus (MMTV) since this is a useful model system for studies of hormone regulation by glucocorticoids [12] progestins and androgens [11].</p><p>An advantage of the Xenopus oocyte system is that proteins may be expressed in variable amounts by injection of corresponding in vitro transcribed mRNAs [12]. The DNA reporter is introduced by intranuclear injection of circular single-stranded (ss) DNA, which in our case yielded approximately 600 million gene copies of the MMTV long terminal repeat and all copies are active in terms of specific protein-DNA binding and chromatin remodeling [12]. Importantly, intranuclear injection of ssDNA in Xenopus oocytes leads to second-strand DNA synthesis coupled to assembly of a tightly organized chromatin structure [12]. Because of the high copy number of the injected DNA, specific transcription factor-DNA interactions can be quantified with high precision by dimethylsulfate (DMS) in vivo footprinting [11], [13]. It is straightforward to isolate the cell nucleus of the oocyte by manual dissection and hence to analyze its protein content.</p><p>Here we show that AR with a pathological polyQ tract of 64 residues (ARQ64) has increased capacity for specific DNA binding. Interestingly, this increase did not correlate with an increase in transcription induction at the MMTV promoter. Hence the transcriptional activity of ARQ64 was significantly reduced in comparison to the wild type ARQ25 as related to its DNA binding activity. The possible mechanism for this effect is discussed.</p><!><p>AR ligands used were R1881 (PerkinElmer Inc., Waltham, MA), as 1×10−3 M in EtOH and MDV3100 (enzalutamide) as 1×10−2 M in DMSO (from Selleck Chemicals Co. Ltd., Houston, TX). The reporter pMMTV:M13 contains the 1.2 kb MMTV LTR fused to the HSV TK gene and its transfer to M13 was described [12], as has the production of mRNAs. The cDNA coding for the different AR variants were based on pβhAR described before [11], that contains the full length human AR, a kind gift from Dr. Jeming Wong [14]. AR variants with polyQ tracts of different length were generated by restriction cloning. A fragment within AR containing the CAG repeat flanked by XmaI and EcoRI restriction sites was excised from pβhAR and replaced with fragments containing 0, 13, 25, or 64 CAG repeats. The AR-Q64 cDNA was shown to be contaminated with AR variants containing shorter repeat(s) than Q64. The contaminants were removed by re-transformation of the pβAR-Q64 plasmid in E-coli cells with reduced recombination activity (SURE cells™, Stratagene).</p><!><p>Experimental design and quantification of AR in oocyte nuclei. (A) Xenopus oocytes were injected with mRNA into the cytoplasm and with ssDNA into the nucleus and harvested for analysis at indicated time (h). (B) Increasing concentrations of [3H]-R1881 were incubated with oocytes either injected or not injected with mRNA coding for ARQ25 and next day taken for analysis of [3H]-R1881 in manually isolated nuclei. (C) Oocytes were injected with the indicated amounts of mRNA coding for ARQ25 or ARQ64 and were then analyzed either by quantification of nuclear [3H]-R1881 or (D) by Western blot (WB), A.U. indicates arbitrary units.</p><!><p>Eight oocytes from each pool of mRNA injected oocytes were placed in a separate 96-well plate containing OR2 buffer [15] and 100 nM concentration of [3H]-R1881 (PerkinElmer, 81.2 Ci/mmol). Since Xenopus oocytes contain endogenous androgens [16] the concentration of radioactive hormone was titrated to obtain near complete exchange of any endogenous ligand with the radioactively labeled probe this titration showed that [3H]-R1881 concentration above 50 nM was enough to fully saturate the expressed AR with radioactive ligand (see results, Fig. 1B). 28 h after mRNA injections, the intranuclear AR protein was analyzed as the radioactivity detectable in isolated oocyte nuclei. A [3H]-toluene standard (PerkinElmer) was used to estimate the efficiency for tritium (see results, Fig. 1C). The AR concentration was calculated assuming that one oocyte nucleus has a volume of 40 nl [17]. Routinely, three nuclei were dissected in duplicate for each data point. Nuclear [AR] was defined as the average of the two samples after subtraction of nonspecific radioactivity (see results).</p><!><p>Quantification of the MMTV transcription by S1-nuclease protection and specific DNA binding by DMS in vivo footprinting was done as described previously [18]. Nonlinear model fit of data points were generated using the software CurveExpert Pro V.2.2.0. The amount of intranuclear plasmid DNA was recovered as described for DMS in vivo footprinting and quantified by primer extension in parallel with a DNA standard [13].</p><!><p>We previously used Western blot (WB) to monitor the amount of AR in Xenopus oocytes [11]; an alternative strategy to quantify steroid receptors is to monitor the hormone-receptor complex using a tritiated hormone-ligand as a probe [19]. The latter strategy provides information about the absolute amount of hormone-receptor complex since there is one hormone-binding site in each receptor molecule. Xenopus oocytes contain endogenous androgens so it was important to test whether a tritiated hormone ligand could compete efficiently for binding to the receptor in presence of the unlabeled endogenous hormone. Oocytes previously injected with ARQ25 mRNA and incubated with increasing concentration of the tritium labeled androgen agonist R1881 [20] were taken for nuclear dissection. Oocytes not injected with AR mRNA served as negative controls, indicated as blank oocytes in Fig. 1B. Radioactivity analysis of nuclear extracts showed that a [3H]-R1881concentration above ∼50 nM efficiently competed for any endogenous ligand since the AR dependent nuclear radioactivity reached a plateau at this level and no further increase occurred even at 400 nM [3H]-R1881 (Fig. 1B). The nonspecific radioactivity recovered in blank oocyte nuclei showed a linear relationship with increasing [3H]-R1881concentration in the medium. The addition of non-radioactive R1881 together with the tritiated ligand competed efficiently as expected and reduced the nuclear radioactivity accordingly (data not shown).</p><p>Xenopus oocytes are filled with yolk protein and lipophilic substances. Steroids such as [3H]-R1881 are also rather lipophilic. We addressed this potential source of error in quantification of AR based on radioactive ligand by measuring the amount of [3H]-R1881 ligand in extracts from whole oocytes and from manually isolated nuclei that were either not injected or injected with ARQ25 mRNA (Supplement 1A). This showed a 0.52 µM concentration of nuclear AR in the oocytes injected with AR mRNA and a signal to noise ratio of 5.1 corresponding to 0.02 pmol of specifically bound hormone ligand per nucleus based on a nuclear volume of 40 nl [17]. The total amount of hormone in the intact oocyte was 1.2–1.5 pmol/cell and thus a 45-fold higher amount than in AR containing nuclei or 260-fold higher than nuclei in oocyte not injected with AR mRNA (Supplement 1A). The level of radioactivity in the cytosol was about the same whether AR was expressed or not and higher than in the media. This illustrates the differential distribution of the lipophilic steroid R1881 in water solution compared to the lipid-containing oocyte cytosol. This was in sharp contrast to the nucleus, where the background level of radioactive ligand was low enough to allow a reproducible quantification of the AR-dependent amounts.</p><p>mRNAs coding for the polyQ variants ARQ25 and ARQ64 were injected into the cytoplasm of two different pools of Xenopus oocytes and resulting protein expression was analyzed in manually isolated oocyte nuclei either by monitoring [3H]-R1881 radioactivity (Fig. 1C) or by WB analysis of nuclear extracts (Fig. 1D). Both methods show that the stepwise increase of injected of ARQ25 mRNA rendered a linear increase of expressed ARQ25 protein (Fig. 1C and D). However, the ARQ64 expression generated by mRNA amounts above 6 ng per oocyte reached a plateau indicating that saturation in protein expression was achieved (see Fig. 1C and D). This comparison between the quantification of nuclear AR based on retained [3H]-R1881 and WB showed the two methods to correlate well. However, we found that WB tends to produce more variable results and furthermore the WB method did not provide information about absolute amounts of receptor. Both methods have been used in this work, but we favor the [3H]-R1881-based approach for intranuclear AR quantification.</p><!><p>Comparison of ARQ25 and ARQ64. (A) Quantification of MMTV transcription by S1 nuclease protection analysis of oocytes injected with either 2.1; 3.1; 4.6 or 6.9 ng ARQ25 mRNA 2.7; 4.1; 6.1 or 9.2 ng of ARQ64 mRNA and then 3 ng ssDNA pMMTV:M13 and exposed to 100 nM R1881. A.U. indicates arbitrary units. The diagram below shows MMTV transcription as a function of nuclear [AR](µM) based on [3H]-R1881 analysis. Gray shadow indicates the curve fitting of all data by software Curve Expert Pro v.2.2.0. (B) Autoradiogram of primer extension from DMS in vivo footprinting of aliquots of the same oocytes as in Fig. 2A. Specific DNA sites for AR (ARE) are indicated on the left side together with binding sites for other proteins (not expressed here), radioactive bands protected in presence of AR, i.e. DMS methylation protected, are marked to the right with empty circles and reference bands for loading control as filled circles. Quantification of the average value of the protected bands, two lanes per oocyte pool, is shown as columns below with the average deviation of double samples as error bars. The last lane was lost in ARQ25. (C) AR-DNA binding, based on DMS methylation protection, plotted as a function of intranuclear [AR], based on quantification of nuclear [3H]-R1881. The curves are calculated based on the Curve Expert Pro v 2.2.0 software. (D) Transcription of MMTV RNA analyzed by S1 nuclease was plotted as a function of AR-DNA binding activity from DMS in vivo footprinting.</p><!><p>Groups of oocytes were injected with increasing amounts of mRNA coding for ARQ25 or ARQ64 as indicated (Fig. 1A) followed by 3 ng of ssDNA of pMMTV:M13 harboring the MMTV enhancer and promoter. Aliquots of oocytes from each group were collected after mRNA injection and incubated with ∼100 nM [3H]-R1881 and after 28 h the oocytes were harvested and analyzed for nuclear AR concentration, AR-driven MMTV transcription (Fig. 2A), and sequence specific AR-DNA binding (Fig. 2B) by DMS in vivo footprinting.</p><p>A comparison of the two AR variants did not show any difference in the capacity to induce MMTV transcription when related to the nuclear AR concentration (see Fig. 2A, lower diagram). However, the sequence-specific DNA binding, defined as DMS methylation protection (Fig. 2B, see protected bands marked with open circles) plotted as a function of nuclear AR concentration (Fig. 2C) demonstrated a stronger DNA binding by ARQ64. Importantly, a diagram showing AR-driven transcription as a function of AR-DNA binding activity revealed a 2.5-fold increased transcriptional response for ARQ25 as compared to ARQ64 (Fig. 2D). This result was reproduced twice (c.f. Supplement 2). In all three experiments a more robust transcriptional response was seen by ARQ25 as compared to ARQ64 in relation to its specific DNA binding activity. These results argue for a functional dissociation between the DNA binding event and the transcriptional induction seen with the ARQ64 variant.</p><!><p>Comparison of nuclear and cytosolic distribution of ARQ25 and ARQ64 (A) in presence of androgen agonist R1881 or (B) androgen antagonist enzalutamide. Oocytes were injected with 3.5 ng ARQ25 mRNA or 6.9 ng of ARQ64 mRNA followed by 3 ng ssDNA pMMTV:M13 as in Fig. 1A. 28 h later oocytes were harvested and processed for SDS PAGE and WB (see Section 2). 0.75 of oocyte equivalent of cytosol or nuclear extract was applied on each lane. The smaller ARQ64 sub-band of MW ∼120 kDa was also present in this experiment since the re-amplification of the pβARQ64 described above was done later. The ratio of the main band and the smaller band of ARQ64 remained constant when comparing cytosolic and nuclear ARQ64 (data not shown).</p><!><p>We conclude that there is no difference in the nuclear and cytoplasmic distribution of the ARQ25 and the ARQ64. The experiment also indicates that the ligand binding specificity for the ligands tested here is the same for both AR variants.</p><!><p>Our finding of a reduced capacity by ARQ64 to drive MMTV transcription as related to its specific DNA binding activity (c.f. Fig. 2D) encouraged us also to address the biological activity of AR with shorter polyQ repeats; hence we developed ARQ0 and ARQ13 constructs. The in vitro transcribed mRNA was injected into oocytes and 28 h later nuclei and cytosol was separated by manual dissection and the relative amounts of AR in nuclear and cytosol extracts estimated by WB. This indicated an average nuclear AR localization of 98, 97 and 96% for ARQ0; Q13 and Q25, respectively. Thus there was no difference in nuclear uptake between AR25Q and AR variants with shorter polyglutamine tracts (data not shown).</p><p>Oocytes injected with the three AR variants Q0, Q13 and Q25 followed by the ssDNA injection, 3 ng MMTV: M13, three independent experiments did not generate any consistent difference in expression (Supplement 3B), DNA binding or transcription (Appendix A, Appendix A) when relating these activities to the relative intranuclear AR levels, here monitored by WB. We conclude that there was no major difference in AR function for the AR variants containing non-pathological polyQ repeats, suggesting that the effects observed for ARQ64 are specific for the polyQ-expanded AR. However, this finding does not exclude functional differences of AR with shorter repeats in another promoter- or cellular context. As mentioned above, both increased [8] and decreased [5] transcriptional activity have been reported with increasing length of the polyQ repeat.</p><!><p>We observed a difference in sequence-specific DNA binding by ARQ64. Unexpectedly, this AR variant showed increased capacity to bind specific DNA sequence in a chromatinized template in vivo. This difference was distinct when comparing DNA binding as a function of nuclear AR concentration (Fig. 2C and Supplement 2C), but it was even more robust when relating the DNA binding to its capacity to induce transcription (Fig. 2D and Supplement 2D). In the latter case the ARQ25 had a much stronger capacity to induce transcription than the ARQ64 variant. Importantly, this difference was not apparent when relating MMTV transcription to the nuclear AR concentration (Fig. 2A, lower diagram and Supplement 2, lower diagram). This is unexpected since all steroid receptors show hormone-dependent specific DNA binding activity as an important step in the chain of events involved in the hormone-driven gene induction. It suggests that ARQ64 has a reduced capacity to convert the DNA binding event into a transcriptional response.</p><p>Interestingly, inhibiting SUMOylation of AR with an expanded polyQ repeat counteracted the significant loss of transactivation caused by polyQ expansion and ameliorated the mutant AR-mediated disease in a mouse model [10]. A deubiquitinating enzyme Usp12 was recently shown to act as a coactivator for AR in prostate cancer cells [22]. If such modifications occur in oocytes and play a role in the process involved in converting a DNA binding event into a transcriptional response then the elongated polyQ of the ARQ64 variant might reduce or alter this process. We speculate that posttranslational modifications are changed in the context of an extended polyQ and that this contributes to the reduced transcriptional response with ARQ64.</p><p>We are the first to report increased specific DNA binding capacity of the ARQ64 in vivo. This important step in the mechanism of action of androgens is confined to the chromatin target and the transactivating factor, i.e. AR. It remains to be determined whether this result also applies to other AR driven enhancers and in other cellular context.</p><!><p>The authors have declared no conflicts of interest [23].</p><!><p>Supplementary material</p><p>Supplementary material</p><p>Supplementary material</p><p>Supplementary material</p><p>Supplementary material</p>

PubMed Open Access

DrugTargetSeqR: a genomics- and CRISPR/Cas9-based method to analyze drug targets

To identify the physiological targets of drugs and bioactive small molecules we have developed an approach, named DrugTargetSeqR, which combines high-throughput sequencing, computational mutation discovery and CRISPR/Cas9-based genome editing. We apply this approach to ispinesib and YM155, drugs that have undergone clinical trials as anti-cancer agents, and demonstrate target identification and uncover genetic and epigenetic mechanisms likely to cause drug resistance in human cancer cells.

drugtargetseqr:_a_genomics-_and_crispr/cas9-based_method_to_analyze_drug_targets

<!>Chemical compounds<!>Cell Biology<!>Selection of resistant clones<!>Cell proliferation assays<!>Dose Response Analysis<!>Immunofluorescence<!>Nucleic Acid Purification, PCR, and Sanger Sequencing<!>Vector construction<!>Transfection and Selection<!>SURVEYOR Nuclease Assay for Genome Modification<!>RNA-Seq library construction and sequencing<!>Overall bioinformatics strategy

<p>Deconvolving the mechanisms of action of chemical inhibitors is a major challenge in drug discovery and chemical biology research1, 2. When the target of a drug is not known, it is difficult to improve its efficacy or reduce any unanticipated toxicity, and its use as a probe for studying cellular mechanisms is restricted. Therefore, several approaches have been developed to identify the targets of bioactive chemicals1, 2. Recently, pooled shRNA-based knockdown and CRISPR/Cas9-mediated gene deletion-methods have been developed to unravel the mechanisms of action of chemical inhibitors and toxic agents3–5. A major limitation of these approaches is that determining if a candidate protein is the drug's physiological target depends on correlations between protein knockdown and pharmacological inhibition phenotypes. These correlations often fail due to differences between cellular responses to fast-acting (typically, minutes) chemical inhibitors and the cumulative direct and indirect effects of protein knockdown, which can require significant time (typically, hours)6.</p><p>High confidence in establishing a protein as a drug's direct target is achieved when a mutation in the protein confers resistance to the chemical inhibitor in cells and also suppresses drug activity in a biochemical assay, e.g., drug-binding or kinase assay7. To achieve this 'gold standard' (or 'genetic') proof of a drug's target we have developed an approach that uses next-generation sequencing-based discovery of high frequency drug-resistance conferring mutations in human cancer cells7. Our findings suggest that resistance via mutations in the drug's direct target arises at frequencies sufficient for our approach to be effective in human cells that have large complex genomes8. However, testing whether any single mutation can confer drug resistance in human cells typically involves transgene overexpression and may fail for several reasons, such as toxicity. We reasoned that direct genome editing would circumvent this major obstacle and developed an integrated approach for drug target identification. This method, which we name DrugTargetSeqR, (with 'Seq' for transcriptome sequencing and 'R' for CRISPR), combines high-throughput sequencing, computational mutation discovery and CRISPR/Cas9-based genome editing9, 10.</p><p>To develop this method, we analyzed ispinesib, an inhibitor of kinesin-5 that has entered clinical trials as an anticancer agent (Fig. 1a)11–13. We isolated 12 clones (hereafter referred to as "drug-resistant clones"), that were 70–300-fold less sensitive to ispinesib than the parental cells, (Supplementary Results, Supplementary Fig. 1, Supplementary Table 1). We next analyzed all clones for resistance to five known MDR (multi-drug resistance) substrates. Eight of twelve clones showed minimal to no cross-resistance (Fig. 1b, Supplementary Fig. 2 and 3). Four clones revealed moderate to substantial resistance to the MDR substrates and were not prioritized for further analyses. As expected, ispinesib treatment resulted in monopolar mitotic spindles in parental cells (Fig. 1c)14. In contrast, bipolar spindles similar to those observed in vehicle-treated controls (Fig. 1d) were observed in ispinesib-treated drug-resistant clones (Fig. 1e). The mitotic indices of ispinesib and nocodazole treated drug-resistant and parental cells were similar (Supplementary Table 4). Together, these data suggest that ispinesib-resistance in these 8 clones is not conferred by indirect mechanisms, such as suppression of the spindle assembly checkpoint or MDR.</p><p>Transcriptome sequencing was performed on the ispinesib-resistant clones and parental cells. Known mechanisms of resistance to kinesin-5 inhibition include overexpression of kinesin-12 or centrosome separation secondary to EGFR activation15, 16. No significant differences were observed in the expression of kinesin-12 and EGFR transcripts between ispinesib-resistant clones and the parental cell line (Supplementary Fig. 4). A more extensive analysis indicated that the expression levels of a small number of genes (19 up-regulated and 4 down-regulated) were significantly altered in drug-resistant clones (Supplementary Fig. 5). We cannot exclude the possibility that some of these genes are involved in drug resistance. However, the magnitude of changes and the number of differentially expressed genes was lower than what we have observed in clones resistant to other drugs7 and therefore, we did not prioritize analyzing these genes further.</p><p>We next focused on identifying genetic mutations that may confer drug resistance. We applied computational analysis to the transcriptome sequencing reads to identify expressed genes in each clone that have mutations that are absent or undetectable in the parental cells7. Central to our approach is finding genes that are most frequently mutated in 'independent', i.e. least genetically related drug-resistant clones, as these genes are likely to express the drug's direct target7. Using methods previously reported the 8 clones were clustered into 5 'independent' groups7. Only nine genes were mutated in more than one group (Fig. 2a, Supplementary Tables 5–13). Kinesin-5 was the only gene mutated in more than two groups. In fact, this gene was mutated in each of the 8 drug-resistant clones, and three different mutations were identified (Supplementary Table 13).</p><p>In order to analyze whether any one of the identified kinesin-5 mutations is sufficient to confer ispinesib-resistance, we used the CRISPR/Cas9 'nickase' system and homology directed repair (HDR)9, 10. As HDR can be inefficient, selectable markers have been employed to obtain cells with desired mutations9, 17. We postulated that the drug itself could be used to select for genome-edited clones. HeLa cells were transfected with the Cas9 'nickase' and homologous template DNA with or without the kinesin-5 A133P mutation. Wildtype transfectants produced no surviving colonies after drug selection (not shown). In mutant transfectants, mutagenesis of the A133 residue was confirmed using the SURVEYOR mutation detection assay18 (Supplementary Fig. 6), and Sanger sequencing of the genomic locus (Fig. 2b). We found that the A133P mutation conferred 150-fold resistance to ispinesib (Fig. 2c). This mutation, along with the other two kinesin-5 mutations we identified, map to the protein's drug-binding pocket (Supplementary Fig. 7). Point mutations (e.g. D130, A133) and deletions (e.g. residues in 'loop-5') within this pocket have been shown to confer resistance to ispinesib-analogs in vitro19. Therefore, with these data, we can confirm kinesin-5 to be ispinesib's direct physiologic target in a human cancer cell line.</p><p>We also successfully applied the CRISPR/Cas9 'nickase' system combined with drug selection to introduce point mutations that confer resistance to the proteasome inhibitor Velcade20 (Supplementary Fig. 8) in two cancer cell lines. Together, these data indicate that this genome-editing protocol can overcome a major bottleneck in establishing the 'genetic' proof of a drug's physiological target.</p><p>Sequencing data from our ispinesib-resistant clones revealed that the mutant kinesin-5 alleles represent 85 – 100% of the sequencing reads (Supplementary Tables 5–12) in each clone. A similar pattern of mutant kinesin-5 allele expression was observed in the genome-edited, drug-selected HeLa kinesin-5 A133P cells (Supplementary Fig. 9). This near-homozygosity suggests that the wildtype kinesin-5 allele is lost in a large fraction of the resistant cells or that the mutated allele is amplified. These two scenarios are unlikely as kinesin-5 mutations are heterozygous at the DNA level (Fig. 2d) and kinesin-5 transcript levels are slightly decreased in resistant clones compared to parental cells (Supplementary Fig. 10). We propose that wildtype kinesin-5 is silenced by an epigenetic mechanism, e.g. promoter DNA hypermethylation or repressive histone modifications. We reason that since kinesin-5 functions as a homotetramer21, cells undergo selective pressure to preferentially express the mutated, drug-resistant allele in order to generate bipolar spindles and complete mitosis. As the chemical inhibitor acting on any motor domain may inhibit activity, only the tetramer comprised of four mutants is likely to be able to confer drug resistance (Supplementary Fig. 11). We propose that such regulation of gene expression to confer drug resistance may also be important for other targets that exist in multiple-copies within multi-protein complexes.</p><p>We next used DrugTargetSeqR to examine the mechanisms of action of YM155 (sepantronium bromide), a cytotoxic drug that has entered clinical trials as an anticancer agent22. While the discovery of YM155 was based on its ability to suppress survivin expression, questions have been raised about its mechanism of action22. We selected YM155-resistant clones, analyzed their multi-drug resistance, and sequenced their transcriptomes (Supplementary Fig. 12). Remarkably, the number of mutations identified in the drug-resistant clones was ~10-fold higher than in the other cases we have analyzed7 (see also Supplementary Tables 5–12). These data indicate that YM155 is likely to be a mutagenic agent, consistent with YM155's chemical structure which suggests it may intercalate DNA and other reports that YM155 treatment activates the DNA damage response23. While analyses of genetic mutations do not reveal a specific resistance mechanism, the gene expression data indicates that resistance to YM155 is likely due to reduced cell proliferation (Supplementary Fig. 9d). We have also applied our methodology to multi-targeted agents (e.g. sorafenib) and have had difficulty selecting clones resistant to the drug, but not other MDR substrates, likely due to a low probability that a single cell could acquire 2 (or more) resistance-conferring mutations simultaneously without a substantial loss in fitness.</p><p>Overall, our findings indicate that DrugTargetSeqR is an effective method for identifying the targets of cytotoxic drugs (Fig. 2e). The CRISPR/Cas9-based genome editing step has several advantages, as mutations can be engineered in to the endogenous locus, and interaction between multiple genetic alterations can be analyzed. Further, when two drugs are effective in combination and resistant mutations to one are known, we can readily introduce these mutations and analyze the second agent's target and resistance mechanisms. At this stage we favor the use of transcriptome sequencing, as most drugs target expressed proteins, and the sequencing data include gene expression levels and mutations. However, other strategies such as exome-capture followed by sequencing may be used to detect copy number alterations and mutations in genes expressed at low levels24. We also believe that compounds that target regulatory sequences or noncoding transcripts could be analyzed using DrugTargetSeqR. Specifically, observed changes in gene expression could be followed by targeted sequencing of specific promoters and regulatory regions.</p><p>We are optimistic that we can further improve the general applicability of DrugTargetSeqR. To analyze mechanisms of action of non-cytotoxic agents, drug-resistant clones could be selected using reporter gene expression or a phenotype that can be readily measured for a few or single cells, e.g. by high-throughput microscopy. Furthermore, the use of suspension cell lines in place of adherent lines may allow for high-throughput analysis of drug targets, since the selection and expansion of drug resistant clones could be carried out by dilution, a step that is more readily compatible with automated liquid transfers.</p><!><p>Ispinesib (S1452) and mitoxantrone (S2485) were purchased from Selleck Chemicals. Nocodazole (74151), paclitaxel (T7402), and vinblastine (V-1377) were purchased from Sigma. Irinotecan was purchased from LC Laboratories (I-4122). All compounds were dissolved in dimethylsulphoxide (DMSO).</p><!><p>For isolation of drug-resistant clones, we chose the HCT116 cell line as it is known to be mismatch-repair deficient and hence genetically unstable, and is known to express low levels of drug-efflux pumps responsible for multidrug resistance (MDR) 7, a common mechanism of drug resistance in human cancer cells. However, prior work has revealed that under drug selection, even HCT116 cells may be selected for high MDR expression7. In order to test whether any single mutation could confer resistance, the comparatively genetically-stable, mismatch-repair intact cell line HeLa was used for CRISPR-based genome editing. The HCT116 cell line was purchased from ATCC. HeLa and HEK293E were kind gifts from Dr. Charles Sawyers, MSKCC. HCT-116 cells and clonal lines were cultured in McCoy's 5A medium (Invitrogen), while HeLa and HEK293E were cultured in Dulbecco's Modified Eagle's medium (Invitrogen). All cultures were supplemented with 10% FBS (Atlanta Biologicals) and penicillin-streptomycin (100 U/ml and 100 ug/ml, respectively, Invitrogen) and grown at 37°C in a humidified chamber with 5% CO2. HeLa and HEK293E were also supplemented with 2 mM L-glutamine (Invitrogen).</p><!><p>Resistant clones were generated as previously described7. Briefly, 0.5 – 1.0 × 106 HCT116 cells were plates in 10 cm culture dishes with media containing ispinesib at a final concentration of 50 – 125 nM and 0.1% DMSO. Media with compound was exchanged every three days for two to four weeks. Most cells did not survive, but a few per plate grew into colonies. Typically fewer than 10 colonies were found on each plate. Colonies were picked by ring cloning and transferred to a new plate where they were maintained in media containing drug at the same concentration as the selections.</p><!><p>In order to quantify cell growth in the presence of drug, cells (2000 cells in 100 μl of media per well) were plated in flat-bottomed 96-well plates and treated with 4 or 8 doses of a serial dilution of the compound of interest. Each condition was plated in triplicate. After three days, cell proliferation was determined using a WST1 assay (Millipore) according to the manufacturer's instructions. Absorbance was read at 440 nm and 690 nm using a BioTek Synergy Neo HTS Plate Reader. Data were used to generate dose response curves as described below.</p><!><p>For each experiment, cellular proliferation (mean absorbance of 3 reads) and error bars (standard deviation) versus concentration of drug were plotted. For each experiment, data were fit using Eq. 1 to find the LD50. Three independent experiments were performed for each condition. The mean LD50 and standard deviation of 3 independent experiments are shown in Supplementary Tables 1, 2, and 3.</p><!><p>All experiments were performed on a Zeiss Axioplan2 microscope (Carl Zeiss MicroImaging), with a 40x objective (Plan Neo, NA 0.75) and DeltaVision Image Restoration Microscope (Applied Precision). HCT116 cells were plated on glass coverslips (Fisher Scientific) in 6-well dishes 24 hours before fixation. Cells were exposed to DMSO only (vehicle control), ispinesib (50 and 100 nM) for 4 hours, or 50 ng/mL (166 nM) nocodazole for 14 hours at 37 °C in complete media. Cells were fixed for 10 min at 37°C in fix solution (4% formaldehyde, 0.2% Triton X-100, 10 mM EGTA, 1 mM MgCl2, 100 mM PIPES pH 6.8). Coverslips were washed 3 times with TBS-tx (TBS + 0.1% Triton X-100), blocked with AbDil (2% BSA in TBS-tx buffer) and incubated for 1 hr at room temperature with FITC-conjugated mouse anti-tubulin monoclonal antibody (Sigma # F2168; 1:2000 dilution in AbDil). Coverslips were washed three times in TBS-tx, and DNA was stained with Hoechst 33342 (Sigma; 1:10,000). Coverslips were mounted in 0.5% p-phenylenediamine (Sigma) in 20 mM Tris, at pH 8.8, with 90% glycerol and sealed with nail polish.</p><!><p>Total RNA was isolated from cells using the RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Genomic DNA was isolated from cells using the DNeasy Mini Kit (Qiagen) according to manufacturers instructions. PCR amplification of the genomic locus for vector construction, SURVEYOR assay, and sequencing reactions was performed using AccuPrime™ Pfx DNA Polymerase according to manufacturer's recommendation.</p><!><p>The SpCas9n "nickase" targeting vector pX335 [Plasmid 42335: pX335-U6-Chimeric_BB-CBh-hSpCas9n(D10A)] was kindly provided by Dr. Luciano Marraffini (Rockefeller University). Vector was digested using BbsI (NEB), and a pair of annealed oligos were ligated into the guide RNA construct.</p><p>The synthetic guide RNA targeting kinesin-5 exon 5 was generated using oligos 5′-caccttcagtcaaagtgtctctgt-3′ and 5′-aaacacagagacactttgactgaa-3′. The synthetic guide RNA targeting psmb5 was generated using oligos 5′-cacccaccatggctgggggcgcag-3′ and 5′-aaacctgcgcccccagccatggtg-3′.</p><p>Template DNA containing the single base pair change uncovered in our screen (kinesin-5 A133P) was generated using PCR-amplification of a pBlueScript vector containing the desired point mutation flanked by 1kb homology arms to generate a 2kb template for Homology Directed Repair. Point mutants were generated using Quikchange site-directed mutagenesis (Stratagene) of pBlueScript engineered to contain the 2 kb fragment of the genomic sequence flanking the desired mutation. Wildtype genomic DNA inserts were amplified from parental cells, and blunt ligated in to EcoRV-digested pBluescript. All constructs were sequence verified using Sanger sequencing in both directions and compared to reference sequence GRCh37.p13. PCR amplified template was gel purified before use in transfection. For amplification of the genomic region of kinesin-5 exon 5, genomic DNA from parental cells was amplified using primers 5′-taaagtgatggggtcccactg-3′ and 5′-tgaccatctgtctcccacact -3′. For amplification of the genomic region of psmb5, the primers 5′-cagtagccacaagccacaca-3′ and 5′-aggcctcttgggttgattcc-3′ were used. Site-directed mutagenesis using Quikchange employed the following primers: kinesin-5 A133P was mutagenized using primers: 5′-taatttcaggatcccttgcctggtataattccac -3′ and 5′-gtacgtggaattataccaggcaagggatcctg-3′. Psmb5 was mutagenized using 5′-cgcccccagccacggtgccaagcagg-3′ and 5′-cctgcttggcaccgtggctgggggcg-3′ (M104V) and 5′-accatggctgggggcacagcggattgcagct-3′ and 5′-aagctgcaatccgctgtgcccccagccatggt-3′ (A108T).</p><!><p>Cells were seeded onto 12-well plates (BD Falcon) at a density of 400,000 cells/well, 24 hours prior to transfection. Cells were transfected using FuGene6 transfection agent at 80%–90% confluency following the manufacturer's protocol. A total of 1000ng pX335 Cas9 plasmid and 1000 ng of HDR template PCR product was transfected to each well. Cells were incubated at 37°C for 48 hours post-transfection prior to drug exposure, at which point they were expanded to 6cm dishes in ispinesib 10 nM. Transfected cells were maintained in escalating doses of drug to a final concentration of 100nM for 10 days, during which time most cells died. Media with drug was exchanged every three days. Cells were split, expanded, and harvested for genomic DNA as described above in Nucleic Acid Purification.</p><!><p>A 390 bp genomic region flanking the CRISPR target site for each gene was PCR amplified using primers listed above and gel purified using QiaQuick Spin Column (QIAGEN) following the manufacturer's protocol. 400 ng total of the purified PCR products were eluted using 10mM Tris-HCl pH 8.8, 15 mM MgCl2, 50 mM KCl to a final volume of 20 μl and were subjected to a reannealing process to enable heteroduplex formation: 95°C for 10 min; 95°C to 85°C ramping at −2°C/s; 85°C to 25°C at −0.25°C/s; and 25°C hold for 1 min. After reannealing, products were treated with SURVEYOR nuclease and SURVEYOR enhancer S (Transgenomics) following the manufacturer's protocol and were analyzed on 4%–20% Novex TBE polyacrylamide gels (Life Technologies). Gels were stained with EtBr and imaged using ultraviolet light.</p><!><p>Library construction for RNA-Seq was performed as previously described7, following standard Illumina protocols using Illumina reagents by the Weill Cornell Genomics Core Facility. Sequencing was performed using the Illumina HiSeq2500 using SR 51bp. Three samples were run per lane.</p><!><p>Identification of reads mapping to exons and across known exon junctions was performed as previously described7. All mapped reads were remapped to the hg18 reference human genome using custom programs based on the June 2010 RefSeq gene annotation. Mutation detection, clone clustering, and merging was performed as previously described7, with one modification. In previous work, we had initially reasoned that a two-hit mutation involving the same exact nucleotide (mutation of two alleles) would be unlikely, and were therefore using a filter eliminating homozygous or near-homozygous mutations. We have removed this filter from the current analysis and recommend not using that filter in light of the results seen here. (Removing the filter had no influence on the previously published results.) All bioinformatics protocols can be found at http://icb.med.cornell.edu/wiki/index.php/Elementolab/TargetID.</p>

PubMed Author Manuscript

Computational engineering of low bandgap copolymers

We present a conceptual approach to low bandgap copolymers, in which we clarify the physical parameters which control the optical bandgap, develop a fundamental understanding of bandgap tuning, unify the terminology, and outline the minimum requirements for accurate prediction of polymer bandgaps from those of finite length oligomers via extrapolation. We then test the predictive power of several popular hybrid and long-range corrected (LC) DFT functionals when applied to this task by careful comparison to experimental studies of homo- and co-oligomer series. These tests identify offset-corrected M06HF, with 100% HF exchange, as a useful alternative to the poor performance of tested hybrid and LC functionals with lower fractions of HF exchange (B3LYP, CAM-B3LYP, optimally-tuned LC-BLYP, BHLYP), which all significantly overestimate changes in bandgap as a function of system size.

computational_engineering_of_low_bandgap_copolymers

Introduction<!>Methodology<!>Geometries<!>Key parameters of bandgap engineering<!><!>Key parameters of bandgap engineering<!><!>Benzoid-quinoid and donor-acceptor copolymers<!><!>Bandgap tuning via the copolymer concept<!>Frontier orbital localization<!>Prediction of (co)polymer bandgaps<!><!>Prediction of (co)polymer bandgaps<!><!>Prediction of (co)polymer bandgaps<!><!>Prediction of (co)polymer bandgaps<!>Conclusion<!>Conflict of interest statement

<p>Low bandgap (co)polymers have attracted much attention for use in polymer-based bulk heterojunction (BHJ) organic solar cells (OSCs) to efficiently harvest the near infrared portion of the solar spectrum. The possibility to combine donor (D) and acceptor (A) units in a DA copolymer structure (Roncali, 1997), and to introduce functional side chains with electron donating/withdrawing moieties and/or sterical demands, which mainly control the polymer morphology, has opened sheer endless possibilities to effectively tune the optical bandgap (Roncali, 1997; Ajayaghosh, 2003; Bundgaard and Krebs, 2007; Kroon et al., 2008; Chen and Cao, 2009; Cheng et al., 2009; Son et al., 2011; Zhou et al., 2012), but also to achieve a broad absorption spectrum (Beaujuge et al., 2010), balanced electronic levels (Li, 2012; Takimiya et al., 2013), improved processability, and controlled packing (Chen et al., 2009).</p><p>However, the optoelectronic properties and processes in OSCs are a complex matter; the polymers' excited state features (energies and oscillator strengths of singlet and triplet states), the absolute position of the highest (lowest) (un)occupied molecular orbitals, HOMO (LUMO), interchain packing, and interfacing with the BHJ acceptor material (usually fullerene variants), sensitively influence not only the photogeneration of excitons, but also exciton transport, charge generation, recombination, transport and extraction at the electrodes; for a recent review, see (Coughlin et al., 2013). Facing these challenges, quantum chemical (QC) calculations have emerged as an indispensable tool to understand properties and processes in OSCs (Brédas et al., 2009; Risko et al., 2011), but also to prescreen polymers prior to synthesis. However, the relevant length- and time-scales of the various polymer properties and optoelectronic processes span several orders of magnitude, and despite steadily increasing computer speeds, only the shortest of these length- and time-scales can be treated with the most accurate of today's methods. Thus, the rather modest task of calculating optical transitions of single (co)polymer chains taxes even cost-effective methods, previously based on semi-empirical configuration interaction singles theory (CIS), but now almost exclusively based on (time-dependent) density functional theory, (TD)DFT.</p><p>Nevertheless, despite providing significantly higher accuracy than semi-empirical methods in many cases, DFT and its TD extension suffer from several pitfalls (Dreuw and Head-Gordon, 2004; Cohen et al., 2012). One such pitfall is that currently popular and widely used DFT functionals such as B3LYP typically provide higher accuracy for small molecules than large ones and as such they are unable to correctly predict changes in properties as a function of system size. In particular, they provide an incorrect evolution in bond-length alternation (BLA) (Jacquemin et al., 2006; Sancho-García and Pérez-Jiménez, 2007; Körzdörfer et al., 2012), ionization potentials (Körzdörfer et al., 2011), and HOMO-LUMO band gaps and electronic transition energies as a function of system size (Gierschner et al., 2007; Körzdörfer et al., 2011). The latter failure limits the utility of TD-DFT methods in the accurate prediction of the optical properties of conjugated polymers by extrapolation of oligomeric transition energies. Gierschner et al. (2007) showed clearly for several classes of conjugated polymers that B3LYP overestimates the decrease in transition energy as the chain length increases, mainly due to the overestimation of MO delocalization (Milián-Medina and Gierschner, 2012a), resulting in significantly underestimated extrapolated polymer values.</p><p>While the experts' discussions are occupied by a forest of new functionals adapted to solve the pitfalls of (TD)DFT for specific applications, B3LYP has become the workhorse for non-experts often without consideration of its limitations. At the same time, combinatorial approaches are screening millions of small oligomer building blocks around the clock (Hachmann et al., 2011; Kanal et al., 2013). Both parties seem in some way to hamper the legitimate quest for both reliable prediction of the optoelectronic properties at the polymer limit and a proper understanding of the underlying physics. It also does not help that mutual understanding between experiment and theory is often lacking.</p><p>This article aims to develop such understanding in the specific case of low bandgap (co)polymers. For both experiment and theory, we adopt the oligomer approach, where the (optical) bandgaps of oligomers are plotted against the reciprocal chain length n (Meier et al., 1997; Müllen and Wegner, 1998; Meier, 2005). Only by studying evolution with n can the pitfalls of the QC methods be properly detected. In fact, in literature, the extreme care required when comparing calculated and measured optical bandgaps is rarely taken. Gas phase calculations on an oligomer are often directly compared to experimental polymer values without consideration of environmental effects and hence little can be learned about the reason for the success or failure of different QC methods.</p><p>There are essentially only two relevant physical parameters which control the chainlength evolution and the polymer bandgap, which can be extracted by a Kuhn-fitting procedure (Gierschner et al., 2007). In light of this parameter set we will discuss concepts for low bandgap materials, unify the terminology by comparing the formally divided donor-acceptor (DA) and benzoid-quinoid (BQ) strategies, demonstrate the need for a proper QC description of the 2-parameter set and shortly discuss MO localization phenomena. In the last section, selected DFT-based QC methodologies will be systematically tested by a careful comparison with experimental values for oligothiophenes and for low bandgap co-oligomers, allowing us to suggest a viable approach to polymer bandgap prediction.</p><!><p>Ground state optimizations as well as frontier orbital calculations (energy and topology) were performed using DFT with various functionals, including B3LYP with 25% HF exchange (Becke, 1993a,b; Stephens et al., 1994), BHLYP [as implemented in Gaussian09, similar but not identical to Becke (1993b)] with 50% HF exchange, M06HF (Zhao and Truhlar, 2006) with 100% HF exchange, the long-range corrected (LC) CAM-B3LYP (Yanai et al., 2004), and finally a LC optimally tuned (OT)-LC-BLYP for which the range parameter was chosen to tune both the HOMO and LUMO and hence the bandgap for each oligomer. This was achieved by minimization of the function (1)J2(μ)=[εHOMOμ(N)+IPμ(N)]2+[εHOMOμ(N+1)              + IPμ(N+1)]2 where μ is the range parameter, εHOMO is the HOMO energy, IP is the vertical ΔSCF ionization potential, and N is the number of electrons, (Foster and Wong, 2012) using a home-built code incorporating an automatic minimization routine. The code was tested and reproduced the optimized range-separation parameters obtained in (Foster and Wong, 2012) for five systems. Optimized OT-LC-BLYP geometries were obtained by alternating between successive geometry and J2(μ) optimizations to self-consistency. The 6-31G** basis set was used throughout. Vertical absorption transition energies Evert were then calculated using all the above functionals coupled to a TD-DFT scheme. Semiempirical Evert were calculated by the ZINDO/S method, i.e., intermediate neglect of differential overlap method as parameterized by Zerner et. al. for spectroscopic applications (Zerner, 1994) coupled to an single configuration interaction (SCI) scheme taking into account all π/π* type orbitals. All (TD)DFT calculations were done in vacuum within the Gaussian09 program package (Frisch et al., 2009). Orbital pictures were drawn with MOLEKEL 4.3 (Flukiger et al., 2000–2002).</p><!><p>All calculations were performed on hydrogen-terminated oligomers. Ground state equilibrium structures of dimeric (co-)oligomers were optimized to determine the starting structures for the longer oligomers. For the TD-DFT studies, geometries were optimized with various functionals described above in order to establish to what extent structural differences, in particular BLA and inter-ring torsion angles, affect excitation energies and their chain-length evolution. In general, higher BLA and larger torsion angles will shorten the conjugation length and lead to larger bandgaps. Full optimization of geometries within both planar and non-planar symmetry groups allowed the impact of differences in predicted BLA and torsions to be distinguished. Accordingly, thiophene oligomers (nT, n = 2–6) and thienopyrazine-alkyl-thiophene co-oligomers (nTTP, n = 1–5) were optimized within the planar groups C2v and C2h (odd and even n, respectively)and the non-planar groups Cs/C2. Diketopyrrolopyrrole thiophenes (nTDPP, n = 1–3) were optimized in the planar and non-planar groups C2h and C2, respectively. Oligophenylenevinylenes (nPV, n = 2–5) were optimized in the planar and non-planar groups C2h and Ci, respectively.</p><p>For the ZINDO/S bandgaps geometries were optimized at the B3LYP/6-311G* level. Here, (alkoxy-)thiophene oligomers (nT, nROT) were kept coplanar as shown earlier (Milián Medina et al., 2008); also thiophene methines (nTM) were found to be planar systems and were calculated at the maximum possible symmetry (C2h, C2v, Cs). Benzene-based systems, benzothiadiazole (nBT) and its co-oligomers (F8BT) were found to be far from planarity (inter-ring torsional angles of 33 and 35°, respectively) (Huang et al., 2011); no symmetry restrictions were imposed here. Isothionaphthenes (ITN) and it's co-oligomers with thiophene were also found to be non-planar.</p><!><p>Figure 1 shows the evolution of the experimental optical bandgap E (adiabatic transition energies) of unsubstituted oligothiophenes (nT) where n is the number of the repetition units (in this case thiophene rings), where the optically allowed singlet transition from the ground to first excited state (S0→S1) of nTs [and the majority of the conjugated (co)polymers] is sufficiently described by the HOMO→LUMO transition. Bandgaps in vacuo were obtained from the intersections of absorption and emission spectra utilizing different solvents and extrapolation to vacuo as described elsewhere (Gierschner et al., 2002, 2003, 2007). This procedure allows for a precise comparison with QC calculations of adiabatic (as well as vertical) transition energies, which are conventionally performed in vacuo. Plotting EN against 1/N, N being the number of double bonds along the shortest conjugated pathway between the terminal carbon atoms (for nT, N = 2 · n), a fairly linear relation for small oligomers (N ≤ 12) is obtained (see Figure 1). This however saturates for long oligomers above the effective conjugation chain length (which can be graphically determined) (Gierschner et al., 2007), reaching the polymer limit at E8 = 2.5 eV for nT (in vacuo). It has been shown (Gierschner et al., 2007), that the chain length evolution can be reasonably described by a simple 2-parameter model according to Kuhn (1948). This model is based on a linear coupling of harmonic oscillators (double bonds) responsible for the electronic transition (Figure 2), where E1 is the transition energy of a formal double bond (N = 1) and Dk is a relative force constant (Dk = 2 · ks/kd), measuring how strongly the double bonds (isolated oscillators of force constant kd) are coupled by the single bonds through the force constant ks.</p><!><p>Optical bandgaps as a function of 1/N with N as the number of double bonds on the shortest conjugation pathway between the terminal carbons. Solid circles: Experimental values (in vacuo) for oligothiophenes nT (where n = 2N) (Gierschner et al., 2007). Solid line: Fit according to the Kuhn equation (Figure 2) with E1 = 6.88, Dk = 0.867. Variation of the Kuhn parameters Dk (left) and E1 (right).</p><p>Parameters of the Kuhn equation and their primary and secondary control in conjugated (co)polymers.</p><!><p>The saturation for E(N→∞) in Kuhn's coupled oscillator model has its quantum-chemical equivalent in the localization of the frontier MOs in the center of the molecule with increasing N. This confines the change in BLA upon S0→S1 excitation (i.e., from a benzoid pattern with a large BLA to a more delocalized pattern with a small BLA) to a finite number of units at the center of the molecule (Milián-Medina and Gierschner, 2012a). In fact, DFT-calculated HOMO-LUMO gaps also show a Kuhn-like evolution (Milián Medina et al., 2007), also for very large n (Zade et al., 2011), as recently reviewed (Torras et al., 2012). Within the Kuhn model, Dk is implicitly related to the change in BLA upon S0→S1 excitation; smaller changes in BLA are associated with values of Dk closer to unity, and hence smaller bandgaps. Thus, in order to promote smaller bandgaps, a smaller ground state BLA is indeed required as stated frequently in literature (Kertesz et al., 2005); starting from a ground state which already has a low BLA decreases the change in BLA upon excitation, and results in a Dk closer to unity.</p><p>It should be noted that for certain classes of materials, empirical fitting functions, e.g., the 3-parameter exponential fit introduced in Meier et al. (1997) give somewhat better fits to the experimental values (Karsten et al., 2008). However, we want to stress that in contrast to the exponential fit, the Kuhn model allows for a physical interpretation of the 1/N evolution and thus molecular design rules, vide infra. The sometimes unsatisfactory Kuhn fit reflects to the simplicity of the model, which assumes a constant Dk with increasing n. Although the latter is true for a large variety of materials (Gierschner et al., 2007), there are important exceptions all arising from "end effects." Firstly, monomeric species (n = 1), can usually not be fitted with the same parameters as longer oligomers due to pronounced resonance stabilization in the monomer, which is weakened for longer chains. Other exceptions are oligomers with strong terminal donor and/or acceptor substituents, whose impact on the π-electron system significantly decreases with chainlength (Meier, 2005; Gierschner et al., 2007), and thus do not allow for a simple homogenous description. A third exception arises from systems, in which the BLA changes strongly between the rings at the end of the chain and those at the center, e.g., in terminal tetracyano-methylene-substituted oligothiophenes (TCnT), which introduce a quinoid structure in the end groups (Ponce Ortiz et al., 2007). All these cases require a modification of the Kuhn equation, with e.g., an additional exponential term (Gierschner et al., 2007).</p><p>For the vast majority of conjugated materials with n ≥ 2 however the Kuhn model performs reasonably well (Gierschner et al., 2007). This immediately provides two simple strategies for achieving lower bandgaps, which are summarized in Figure 2:</p><!><p>an increase of Dk (while keeping E1 constant) leads to an increase of the slope and diminishes the curvature at the polymer limit (Figure 1), thus increasing the effective conjugation length for Dk < 1 and yielding a zero bandgap for the limiting case of Dk = 1. This corresponds to a decreased BLA by increasing the quinoid character of the conjugated chain (vide infra), as predicted early on for poly-isothionapthene (PITN; Figure 3) (Kürti and Surján, 1990), or for the previously mentioned TCnT series, which both show a substantial reduction of the lowest optical transition relative to the nT series as predicted by relatively simple semi-empirical QC calculations, see Figure 4.</p><p>Besides this primary effect to manipulate Dk, the planarity of the molecular backbone constitutes a secondary effect, which can have a considerable impact on the slope. This can be seen e.g., in the strongly sterically hindered oligophenylenes (Milián-Medina et al., 2011); here the introduction of a torsional twist leads to a significant decrease of Dk. This effect should be distinguished however from only thermally induced non-planarity, which blue-shifts the absorption maximum (which, to a first approximation, corresponds to Evert), but not the optical bandgap (which corresponds to the adiabatic transition energy) as previously demonstrated (Gierschner et al., 2002, 2003). For side-chain substituted systems, regio-regularity is crucial to promote planarity of the chain as has been shown for polyhexylthiophene (P3HT) (Gierschner et al., 2007).</p><p>a decrease of E1 (and thus of the excitation energy of the repetition unit ERU) leads to smaller bandgaps at the polymer limit; however the slopes of the chainlength evolution get significantly smaller as long as Dk is kept constant, see Figure 1. To reduce ERU, electron withdrawing moieties are used, e.g., by replacing benzene-based RUs by thiophene, or better, by annulated rings which reduce ERU mainly due to the stabilization of the LUMO (Figure 3), and/or using electron-donating substitution patterns. For instance, electron-donating side-chain substituents in all rings lead to somewhat lower bandgaps also at the polymer limit (compare e.g., P3HT vs. PT with a gain of ca. −0.1 eV due to the inductive +I effect of the alkyl groups) (Gierschner et al., 2007). On the other hand, terminal substitution with strong D/A groups can have a very large effect on small oligomers in generating small bandgaps (Meier, 2005), however this effect will vanish at the polymer limit (Meier, 2005; Gierschner et al., 2007). This is also true for the recently discussed dicyano- (Pina et al., 2006) or dicyanovinyl-thiophene (DCVnT) series (Pappenfus et al., 2003; Schulze et al., 2006), see Figure 4. A secondary effect in this category is the promotion of dense packing in the solid state by minimizing the disorder, which significantly lowers the effective optical bandgap due to enhancement of the anisotropic polarizability of the material via closely packed parallel chains (Egelhaaf et al., 2002), whereas H/J-aggregation play only a very minor role in polymers (Gierschner et al., 2009).</p><p>In summary, low bandgaps with E∞ ≤ 1 eV are only expected if narrow bandgap repetition units are combined with a small BLA. An example for a homopolymer fulfilling these conditions is poly-thiophenopyrazine (PTP; Figure 3), with a small optical bandgap of ca. 1.2 eV in solution, (1.0 eV in films) (Wen et al., 2008). The low bandgap of PTP is further promoted by planarization due to attractive sulfur-nitrogen interactions (Özen et al., 2007), evidenced by well-structured vibronic side bands of the absorption spectrum (Milián Medina et al., 2008; Wen et al., 2008). Appropriate side chains will further promote dense packing and thus lowered bandgaps; however this has to be balanced with the solubility requirements for polymer processing.</p><p>Examples of small bandgap monomers, with calculated HOMO and LUMO levels as well as Evert obtained at the (TD-)DFT B3LYP/6-311G* level of theory.</p><p>Evert of different conjugated oligomers as a function of 1/N as calculated at the ZINDO/S level on DFT (B3LYP/6-311G*) optimized geometries. Solid lines for nT and nTM are fits according to the Kuhn equation (Figure 2); for the DCVnT mono-exponential fits, and for nITN and TCnT biexponential fits were used.</p><!><p>Although homopolymers like PITN might reach very low bandgaps, the application of such materials is limited, both due to the rather limited possibilities for electronic tuning (e.g., of the absolute MO energies) and processability through side chain functionalization. Thus, copolymers emerged as an alternative, combining different conjugated repetition units and substitution patterns to provide better processability and a large variation in optoelectronic properties and packing motifs. Copolymers can be formally divided into BQ and DA copolymers (Kertesz et al., 2005; Zhou et al., 2012). Herein, BQ copolymers are formed by insertion of quinoid repetition units into a benzoid backbone; (Kertesz et al., 2005), where (partial) quinoid character in the repetition units is introduced through resonance-stabilized annulated rings (e.g., the benzene ring in the ITN unit in Figure 2), whose aromaticity can be additionally enhanced by electron-donating substituents. It should be noted however that e.g., poly(thiophene methine), PTM (Chen and Jenekhe, 1995), see Figure 4, although it connects B and Q units, behaves very similar to the nT series (Figure 4). This is due to the fact that the BLA path of both subunits is not broken, i.e., the alternation is pursued throughout the chain, since B and Q units do not compete for the same bonds. Thus, PTM might be rather called a pseudo-BQ copolymer.</p><p>In DA copolymers, decreased energy of the repetition unit is accomplished by appropriate DA pairs with low lying HOMO and LUMO levels of A compared with those of D, as shown for D = fluorene (F8) and A = benzothiadiazole (BT) in Figure 5. Consequently, the LUMO energy of DA is similar to that of A, and the HOMO to D, thus yielding a small bandgap for DA (Havinga et al., 1993; Cornil et al., 2003; Karsten et al., 2009). To realize narrow bandgaps, annulated electron-withdrawing moieties as acceptor units are used (Figure 3), which usually belong to the same family as the quinoid moieties, so that in practice, a distinction between BQ and DA copolymers is not required; in fact, as we will see later all copolymer pairings have an impact in both E1 and Dk.</p><!><p>Left: Orbital correlation diagram (B3LYP/6-311G*, no symmetry restrictions) for the (F8BT) and BT3 from their oligomer fragments. Right: frontier orbitals of (F8BT)3.</p><!><p>Although bandgaps of DA pairs are indeed smaller compared to those of single D and A units, this is not true at the polymer limit as one could suspect in a first approximation (Havinga et al., 1993). Instead, acceptor-only A∞ homopolymers exhibit lower bandgaps than (DA)∞ copolymers (Karsten et al., 2009). In fact, it is not the relative energy of the (isolated) D and A frontier MOs which is relevant in determining the optical bandgap of the copolymer (Salzner and Köse, 2002), but E1 and Dk, as we can conveniently see for F8BT (Figure 5). Following the idea of the last section, the bandgap of the DA (F8BT) repetition unit has to be compared to an (A)n oligomer with the same number of double bonds N, since this governs the chainlength evolution. In the present example this is (BT)3, which has in fact a smaller bandgap than F8BT (Figure 5) due to the larger coupling Dk of the BT oligomers compared to the coupling between F8 and BT. Thus, the F8BT copolymer bandgap is expected to be found between that of polyfluorene and the (hypothetical) BT-polymer. The same principle holds for other copolymers, e.g., composed of (alkyl-substituted) thiophene (T) and thienopyrazine (TP) units, where Dk of poly-TP (PTP) is considerably smaller than that of polythiophene (PT). Indeed, for the copolymer P[T2TP1], both Dk and the absorption maximum (and thus the bandgap) lie between the corresponding homo-polymers (Karsten et al., 2009). However, the bandgap for P[T2TP1] is smaller than might be expected from its stoichiometry [i.e., E(P[T2TP1]) < 2/3·E(PT) + 1/3·E(PTP)], which is due to a substantial quinoid character of TP introduced into the copolymer chain (Karsten et al., 2009), vide supra. Further detailed discussions of the variation in bandgaps as a function of D/A stoichiometry can be found in the quantum-chemical studies of Gil-bernal et al. (2010) and Hung et al. (2013).</p><p>Hence, to reach small bandgaps via the DA concept, small bandgaps of the A moiety are decisive, assisted by quinoid contributions as discussed above. It should be noted that the quinoid character introduced in the copolymer chain might significantly vary with the chemical composition of the D/A moieties, and thus change Dk. Hence, prescreening copolymer MO properties through correlation with simple DA pair calculations as is now frequently done (Blouin et al., 2008; Hachmann et al., 2011; Kanal et al., 2013), might not result in appropriate selections for fully-optimized polymers (vide infra). On the other hand, screening of copolymers via periodic calculations (Longo et al., 2012; Bérubé et al., 2013) encounters the same problem from the other side, relying on a correct evolution of the electronic and optical properties with chain length. Thus, a reliable prediction of both E1 and Dk by QC methods will be crucial, vide infra.</p><!><p>Electronic level tuning by the DA concept might imply orbital localizations of the frontier MOs within one of the moieties (Dutta et al., 2009; Milián-Medina and Gierschner, 2012b). The main factor for MO localization is the energy offset between the frontier orbitals of the HOMOs (LUMOs) of D and A segments; other factors include orbital symmetries and conjugation breaks by sterical demands (Schmidtke et al., 2007). F8BT might once more serve as an example, see Figure 5. Here the HOMO is formed by both the F8 and the BT unit, whereas the LUMO is localized on the BT unit (Cornil et al., 2003); thus the HOMO→LUMO transition shows partial intermolecular charge transfer (ICT) character. However, MO topologies might significantly change when going from a simple DA pair to the polymer limit, and so the chainlength evolution has to be carefully studied (Karsten et al., 2008; Dutta et al., 2009). For longer F8BT oligomers, the LUMO is indeed localized on the acceptor moieties, but homogeneously distributed on all BT units within the conjugation length, rather than on a single BT unit (Figure 5). This topology differs significantly from a report on a phenylene-ethynylene based copolymer with a highly localized HOMO (and delocalized LUMO) (Dutta et al., 2009). There, the localization was spatially limited to one single unit of the polymer chain and thus produced a low-lying ICT state of small oscillator strength, which gave rise to a large Stokes shift between emission and main absorption band.</p><p>The question in which way MO localization influences the electronic properties of the copolymer in an OSC BHJ blend is an intriguing one. As the charge-carriers in the copolymer phase are holes, LUMO localization will probably not be a problem, while (partial) HOMO localization might be detrimental. In any case, the polar structure induced by the DA topology could be even helpful for the charge separation at the BHJ interface (Carsten et al., 2011). Moreover, localization resulting in spatial separation of HOMO and LUMO promotes smaller singlet-triplet energy gaps (Milián-Medina and Gierschner, 2012b), which might be of some importance since the low lying triplets seems to play a crucial role in the mechanism of charge recombination in OSCs (Veldman et al., 2009).</p><!><p>The conceptual tools developed above allow us to assess the performance of different QC methods to properly describe the chainlength evolution of (co)oligomers, and to identify their pitfalls. These failures of standard DFT are known, being due to the approximate form of the exchange potential resulting in spurious self-interaction, and in particular, its incorrect distance-dependence. The admixture of exact Hartree-Fock (HF) exchange into the functional (in the so called hybrid functionals), typically improves the accuracy (Milián-Medina and Gierschner, 2012a). LC functionals aim to provide a more balanced description of exchange by varying the fraction of exact HF exchange as a function of distance, and have been applied with some success to long-range charge transfer states for which both popular pure and hybrid functionals typically fail (Dreuw and Head-Gordon, 2004). We will thus especially appraise whether recently developed LC functionals coupled to a TD scheme provide an accurate description of the evolution of optical bandgaps as a function of oligomer length, and hence whether they can offer the predictive power necessary for use as a screening tool in the design of novel conjugated (co)polymers for OSC applications.</p><p>In order to assess the reliability of QC methods, we will perform a careful comparison with experimental oligomer evolution. As a test case for a homopolymer we will use Evert of the nT series (Gierschner et al., 2003), followed by a comparison for low bandgap co-oligomers. Unfortunately, there are only a few examples of systematic experimental studies on co-oligomers in literature; notable exceptions are systems synthesized in the group of R. Janssen, i.e., the nTTP series (Karsten et al., 2008), see Figure 6, and diketopyrrolopyrrole thiophenes nTDPP (Karsten and Janssen, 2011).</p><!><p>Evert of nTs (A) and nTTPs (B) computed using a range of DFT functionals for both geometry optimization (planar C2v/C2h) and transition energies (solid symbols) and Kuhn fits (lines). Experimental results (open symbols; nTs: in DCM, nTTP: in toluene) are shown for comparison.</p><!><p>Improper comparison can produce non-negligible errors which can sum up to 1 eV at the polymer limit, see Gierschner et al. (2007). This concerns the proper definition (and experimental extraction) of the optical transitions (adiabatic vs. vertical transition vs. HOMO-LUMO gap), environmental effects (solvent and solid-state shifts, temperature), and substituent effects. Here we strive for a compromise between accuracy and computational efficiency in order to develop a procedure suitable for rapid screening of (co)polymers with extended repeat units. As such, we focus on vertical absorption energies in solution, and assume that bathochromic solvent shifts and temperature effects do not vary significantly with chain-length, nor between (co-)oligomer series, and can thus be modeled as a rigid shift in transition energies. The available data on nTs (n = 2–6) suggests that this is a reasonable approximation; vacuum to solvent shifts in Evert are 0.19 ± 0.01 eV at 293 K in dichloromethane (DCM) and the 15K–293 K shift of Evert in DCM is 0.16 ± 0.01 eV (Gierschner et al., 2007). Equivalent data is not available for the nTTP, nTDPP series, however this approximation seems to work quite well (vide infra).</p><p>Experimental Evert bandgaps of the nT and nTTP series are shown as open symbols in Figure 6 with the corresponding Kuhn fits through data points with n ≥ 2. The difference between the two series is striking, giving not only much lower values of E0 for the nTTP series (see Table 1) but also stronger couplings Dk, indicative of significant quinoid contributions. Concomitantly, the slope m of the nT series (obtained from a straight-line fit in the quasi-linear regime of the data points; n = 2 to ca. 6), is 50% higher than that of nTTP. This directly demonstrates that reliable predictions of polymer bandgaps via the oligomer approach cannot be obtained from calculations on monomers (n = 1), but requires at least the calculation of two data points (n = 2, 3).</p><!><p>Kuhn fit parameters E1 (in eV) and DK extracted from the experimental and computed Evert as given in Figure 6.</p><p>E∞ is the extracted polymer value (in eV) at the polymer limit; m is the slope from a straight-line fit. ΔEoff is the difference of Evert(measured)–Evert(calculated) averaged over each oligomer series and σ is its standard deviation, representing an average error within a series arising from incorrect chain length evolution.</p><p>TD, time-dependent calculation.</p><p>GO, geometry optimization.</p><p>S, symmetry; p, planar; np, non-planar; see geometry subsection for details.</p><!><p>Figure 6 also plots calculated Evert using a range of DFT functionals. For clarity, results are shown only for planar geometries optimized using the same functional as used to compute the transition energies (a subset of the calculations shown in Table 1). The B3LYP functional strongly overestimates the slope for both series (Figure 6), as previously reported for various homo-oligomers (Gierschner et al., 2007; Milián-Medina and Gierschner, 2012a). OT-LC-BLYP (for which the range-separation parameter had been optimized for each oligomer) performs similarly poorly to B3LYP, significantly overestimating Dk and the slope. As such, B3LYP and OT-LC-BLYP are the two worst-performing functionals we tested when it comes to chain-length evolution of optical bandgaps. The CAM-B3LYP functional gives significantly smaller Dk values (Table 1), yet still strongly overestimates the slopes relative to experimental values. Thus, in contrast to what one might expect, the LC functionals tested here do not correct for the errors in popular hybrid functions which lead to significant overestimation of changes in bandgaps as a function of system size. It is important to note that while the CAM-B3LYP extrapolated polymer value E8 for the nTTP series is close to experiment [ΔE(nTTP) = 0.04 eV], in light of the poor performance on the slope, this must be ascribed to a fortuitous cancellation of errors. However, these errors in E8 do not always cancel, but are strongly system dependent; for the nT series, they result in the rather unsatisfying error ΔE(nT) of 0.22 eV. Thus, we cannot recommend these functionals for unbiased pre-synthesis polymer prediction.</p><p>As the fraction of exact HF exchange is increased from 20% in B3LYP to 50% in BHLYP, the slope decreases, but still significantly overestimates m compared to experiment, see Table 1. Increasing the HF exchange to 100% by using M06HF (for both the geometry and TD parts) yields slopes in remarkable agreement with experiment for both nT and nTPP, consistent with the previous finding that RCIS/HF also provides accurate slopes (Gierschner et al., 2007). Although M06HF bandgaps feature a strong hypsochromic offset relative to experiment, the shift is surprisingly similar for the nT (ΔEoff = 0.70 eV) and the nTTP series (0. 81 eV) despite the strong differences between the two series with respect to both E1 and Dk. We have further tested the M06HF performance on other homo- and co-oligomer series (oligophenylenevinylene; nPV, and nTDPP), showing similarly good results with Δ Eoff (nPV) = 0.76 and Δ Eoff (nTDPP) = 0.69, as depicted in Figure 7 where the experimental values are compared with the offset-corrected M06HF results with ΔEoff = 0.75 eV. The average and maximum absolute deviation of predicted E∞ from experimental value are 0.05 and 0.09 eV, suggesting that this method fulfills our criteria for accurate polymer screening (accuracy to within 0.1 eV). It should be stressed that ΔEoff is basis set dependent. Thus, while the value of Δ Eoff = 0.75 ± 0.10 eV is valid for the 6-31G** basis set used in Figure 7, our preliminary tests on nTs using a 6-311G**+ basis require a value Δ Eoff = of 0.46 ± 0.09 eV. It has been shown that the RCIS/HF method also provides correct slopes in transition energies as a function of chain length (Gierschner et al., 2002, 2003). As for TD-M06HF, RCIS/HF transition energies also overestimate the measured values. This observation was not apparent in the original reports nor in subsequent papers (Gierschner et al., 2007; Milián-Medina and Gierschner, 2012a) as the reported energies are in fact vertical emission energies Evert,em, and not adiabatic energies (E00) as stated therein (by chance, for nTs, Evert,em from RCIS/HF/6-311+g* are almost identical to measured E00 gas phase values).</p><!><p>Experimental Evert (black open symbols) and offset-corrected M06HF results for the nPV,nT, nTTP, and nTDPP series (R,R′ = alkyl chains) using; (A) planar M06HF geometries (solid symbols) and non-planar M06HF geometries (colored open symbols), both with ΔEoff = 0.75, and (B) non-planar BHLYP geometries using ΔEoff = 0.70. Lines are Kuhn fits to the data points.</p><!><p>While the slopes for nT, nTTP, and nPV are in excellent agreement with experiment, the slope of nTDPP is significantly overestimated. This is likely due to our performing calculations on fully planar geometries, while nTDPP oligomers with n > 1 are significantly twisted due to steric hindrance where two alkyl-thiophenes connect. Indeed, using non-planar geometries results in a chain-length evolution in excellent agreement with experiment for nTDPs (see Figure 7, dashed magenta line). It should be noted however that the same procedure applied for the nT series results in an underestimation of the chain length evolution and significant increase in the hypsochromic shift relative to experiment (see Figure 7, dashed red line). This is due to the pronounced deviation from planarity; e.g., in 2T the inter-ring torsion is 156°, and similar values are found in 6T. This runs contrary to experimental evidence that while 2T is twisted, longer oligomers are essentially planar: the lack of mirror symmetry of the absorption and fluorescence spectra of 2T shows that its ground state is far from planar in agreement with ab initio benchmark calculations and experiments (Raos et al., 2003). But as the chain length is increased, the absorption and fluorescence spectra become increasingly symmetric, suggesting that the ground state is, like the excited state, planar in longer oligomers (Macchi et al., 2009). The results suggests that M06HF does not provide sufficiently accurate torsional potentials to correctly describe bandgap chain-length evolution in quasi-planar systems without symmetry constraints, but for such systems M06HF can provide the correct chain-length evolution if planarity is enforced.</p><p>The role of geometry was further investigated by combining different functionals for transition energy calculations and geometry optimizations and comparing results of planar and non-planar symmetries (see Table 1). TD-BHLYP using non-planar BHLYP geometries provides slightly smaller slopes than planar geometries, but still overestimates them relative to experiment, suggesting torsions alone are not responsible for the slope overestimation. For planar geometries, slopes decrease in the order TD-B3LYP/B3LYP > TD-B3LYP/BHLYP > TD-BHLYP/B3LYP > TD-BHLYP/BHLYP, suggesting that the choice of functional used in the TD part has a larger impact than the functional used for geometry optimization. Nevertheless, the geometry clearly has a strong impact; TD-M06HF calculations on planar and non-planar B3LYP, BHLYP, and M06HF nT geometries show a significant spread in E∞, slopes and offsets relative to experiment. Interestingly, TD-M06HF on non-planar BHLYP geometries provided quite good results for nTs (almost as good as TD-M06HF on planar M06HF geometries). This combination was therefore also tested on nTTP, nPV, and nTDPP in the hope that it might provide a general approach to polymer extrapolation, free from assumptions about planarity. Figure 7B plots the results (with an offset of −0.70 eV). Consistent with the data in Table 1, the chain-length evolution of nTs is indeed very well described, and as hoped for, the slope for nTDPP is significantly improved relative to the planar M06HF/M06HF calculations (Figure 7A). However, this does come at the cost of a slight loss of accuracy for nTTPs and nPVs. Nevertheless, the average and maximum absolution deviation of predicted E∞ from experimental values are 0.08 and 0.11 eV, respectively, suggesting this approach could provide sufficient accuracy to be useful in polymer screening.</p><!><p>The conceptual approach to low bandgap copolymers adopted in the present work allowed us to analyze and interpret both the benzo-quinoid and DA strategies to low-bandgap polymer design within the unified framework of the simple, but physically meaningful Kuhn model. The two parameters which determine the optical bandgap at the polymer limit (but also for oligomers) are essentially the bandgap of the repetition unit and the coupling between them. The latter is intimately related to the change of the BLA upon electronic excitation, which should be minimized to achieve low bandgaps and is favored by quinoidal structures in the ground state. In fact, the annulated electron-poor acceptor units generally used to generate low bandgap copolymers additionally enhance the quinoidal character in the polymer chain through resonance stabilization of the annulated rings. Because this can significantly vary depending on the D/A pairing, simple monomer-polymer correlation can be misleading, and the reliable performance of quantum-chemical methods on the chainlength evolution has to be carefully checked through proper comparison with experiments on (co-)oligomers. Such tests are also recommended before using the QC methods in periodic schemes.</p><p>Our tests on nT homo-oligomers and low bandgap nTPP co-oligomers included different TD-DFT methods ranging from the popular B3LYP and other hybrids incorporating higher fractions of HF exchange (BHLYP, M06HF) to LC functionals (CAM-B3LYP, and optimally gap-tuned OT-LC-BLYP). All methods with low to medium HF exchange including the LC variants significantly overestimated the slopes of the chainlength evolution and only occasionally predict polymer optical gaps in agreement with experiment in cases of fortuitous cancellation of errors, and hence cannot be recommended for the prediction of polymer bandgaps. TD-M06HF, when combined with planar M06HF geometries, results in slopes in very good agreement with experiment, except for systems which are clearly non-planar (e.g., nTDPP), for which it overestimates slopes. Removing symmetry constraints and re-optimizing non-planar geometries fixes the slopes of such non-planar systems, but overestimates inter-ring torsion angles in nTs, resulting in underestimated slopes and a large blue-shift. The combination of non-planar BHLYP geometries and M06HF bandgaps seems to provide a more balanced description of both planar and non-planar systems, but with a slight loss of accuracy for both classes. For both approaches a strong hypsochromic offset is found, which however is fairly constant for the four series of in total 17 oligomers we have tested them on. For TD-M06HF/M06HF on planar geometries the offset is 0.75 eV with σ = 0.05, for TD-M06HF/BHLYP it is −0.70 eV with σ = 0.07. The largest error in E∞ relative to extrapolated experiment values are 0.11 and 0.09 eV, respectively, suggesting that both approaches could be useful for predicting polymer bandgaps prior to synthesis. Nevertheless, it is expected that the ongoing development of new functionals will lead to further improvements in the description of the chain-length evolution of optical gaps of conjugated oligomers with non-negligible inter-ring torsion angles.</p><!><p>The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.</p>

PubMed Open Access

MgNb ₂ O ₆ Modified K _0.5 Na _0.5 NbO ₃ Eco‐Piezoceramics: Scalable Processing, Structural Distortion and Complex Impedance at Resonance

In this work, piezoceramics of the lead-free composition K 0.5 Na 0.5 NbO 3 with an increasing amount of MgNb 2 O 6 (0, 0.5, 1, 2 wt.%) were prepared through conventional solid-state synthesis and sintered in air atmosphere at 1100 °C. The effect of magnesium niobate addition on structure, microstructure and piezoelectric properties was evaluated. The ceramics maintain the orthorhombic Amm2 phase for all compositions, while an orthorhombic Pbcm secondary phase was found for increasing the concentration of MgNb 2 O 6 . Our results show that densification of these ceramics can be significantly improved up to 94.9 % of theoretical density by adding a small amount of magnesium-based oxide (1 wt.%). Scanning electron microscopy morphology of the 1 wt.% system reveals a well-packed structure with homogeneous grain size of ~2.72 μm. Dielectric and piezoelectric properties become optimal for 0.5-1.0 wt.% of MgNb 2 O 6 that shows, with respect to the unmodified composition, either higher piezoelectric coefficients, lower anisotropy and relatively low piezoelectric losses (d 33 = 97 pC N À 1 ; d 31 = À 36.99 pC N À 1 and g 31 = À 14.04 × 10 À 3 mV N À 1 ; Q p (d 31 ) = 76 and Q p (g 31 ) = 69) or enhanced electromechanical coupling factors (k p = 29.06 % and k 31 = 17.25 %).

mgnb_₂_o_₆_modified_k__0.5_na__0.5_nbo_₃_eco‐pi

Introduction<!>Results and Discussion<!>Conclusions<!>Synthesis and Sintering of Ceramics<!>Structural, Microstructural and Morphological Characterization<!>Piezoelectric, Dielectric and Elastic Characterization at Resonance

<p>Pb(Zr 1-x Tix)O 3 (PZT) and PZT-based ferro-piezoelectric ceramics are, today, the most commercially exploited systems due to their dielectric, piezoelectric and electromechanical coupling coefficients suitable for a wide number of classical technological applications such as buzzers, gas ignitors, sensors, ultrasonic motors and so on. [1] However, the high toxicity of lead oxide, commonly used as reagent in PZT preparation (up to 60 wt.%), has moved the attention for alternative systems which could replace lead-based materials while miming their performance in emerging applications. [2][3][4][5] Among them, sodium potassium niobate, K 1-x Na x NbO 3 (KNN) is one of the most promising and widely studied lead-free system with high Curie Temperature (420 °C), good ferroelectric properties (Pr = 33 μC cm À 2 ) and appealing electromechanical planar coupling factor for x = 0.5. [6] On the other hand, KNN still present some disadvantages related with their preparation and, in particular, with the extremely narrow interval of temperatures for sintering, high volatility of light elements, and poor densification which affect the piezo properties. [7] To overcome these drawbacks, soft chemical routes, like hydrothermal, microwave assisted and solgel methods, have been explored during the last years. On the other hand, solid-state reaction route followed by air sintering still remains the most common, easiest and cheapest method. [8,9] Recently, some efforts have been made to improve the scalability of the manufacturing process, which is still an open issue for KNN ceramics. In this context, the mechanochemical activation has aroused particular interest due to its many benefits including shorter synthesis time, enhanced chemical homogeneity and density of sintered pellets. [10] In addition, over the years, various "additives" have been suggested to increase the densification and lower the sintering temperature of KNN-based ceramics. In particular, Cu 2 + and Zn 2 + based oxides such as CuO, K 4 CuNb 8 O 23 , CuNb 2 O 6 , K 5.4 Cu 1.3 Ta 10 O 29 , ZnO and K 1.94 Zn 1.05 Ta 5.19 O 15 , have proved to effectively promote densification through the formation of a secondary phase during sintering. [11][12][13][14][15][16] Furthermore, it is frequently reported that these modifying agents, in particular Cu-based compounds, act as acceptors dopants that create anion (oxygen) vacancies and associated defect dipoles. Due to the high mobility of the oxygen vacancies, the defect dipoles are easily oriented by the external field and make difficult the overall switching of the polarization. This domain wall pinning effect is responsible to the increase of the "hard" behavior of the ceramics. [17][18][19] Likewise, Mg-based compounds have shown beneficial effects on the sintering performance of KNN through the formation of a liquid phase, which favor the mass transportation then promoting the densification process. Furthermore, the addition of MgO seems to enhance the electrical properties of the KNN-modified ceramics (d 33 > 100 pC N À 1 ; k p > 30 %; ɛ T 33 > 600). [20] Nevertheless, unlike to the Cu/Zn-based additives, Mg-based systems have been poorly explored, up to know, on KNN ceramics.</p><p>In this work, MgNb 2 O 6 (hereinafter abbreviated as MN) was tested on KNN ceramics, for the first time by our knowledge. The synthesis was conducted with the mechanochemicalassisted activation method, which is an easy and scalable process. The effect of MN on sintering behavior, structure, microstructure, and piezoelectric, elastic and dielectric properties at resonance have been evaluated.</p><!><p>In this study, the composition K 0.5 Na 0.5 NbO 3 was selected due its highest piezoelectric properties. The density of the as-prepared KNN ceramics, sintered at 1100 °C, as a function of MN concentration (wt.%) is summarized in Figure 1. With the increasing of MN content, the density firstly increased from 4.02 g cm À 3 (pure KNN) to 4.10 g cm À 3 (MN 0.5 wt.%), until reaching an optimal value of 4.28 g cm À 3 for the sample modified with 1 wt.%, corresponding to 94.9 % of the theoretical density (TD) of pure KNN (4.51 g cm À 3 ). [21] For higher amount of MN (2 wt.%), the density decreased to 4.14 g cm À 3 , in accordance with trends reported in the current literature for other KNN-modified systems. [22,23] The system with 4 wt.% of MN, exhibits a similar density with respect to 2 wt.%, but it resulted too conductive for determining its electromechanical properties, most probably due to the presence of secondary phases segregated at the grain boundaries. [24] For this reason, the structural and microstructural characterization has been carried out only for the 0.5, 1 and 2 wt.% modified systems, together with the pure KNN for comparison purposes. Starting from the crucial point that the structures of piezoceramics significantly influence their properties, the XRD patterns of calcined KNN and sintered KNN-xMN (x = 0, 0.5, 1, 2 wt.%), at room temperature, were recorded to get light on their structure. The as-calcined pure KNN pattern appears as a complex multiphase mixture (Figure 2).</p><p>The Rietveld refinement were performed using the orthorhombic Amm2 KNbO 3 , the tetragonal P4mm K 0.3 Na 0.7 NbO 3 and the orthorhombic Pbcm NaNbO 3 as a starting model, according to what reported by other authors (Figure 2a). [10,25] The good agreement between the model and the experimental data was established by the low R-factor (Rwp = 6.91 %) achieved by this analysis. Shoulders on the right side of the peaks are attributable to the orthorhombic Pbcm phase of the NaNbO 3 (Figure 2b). The presence of mixed phases was due to the monoclinic polymorph of Nb 2 O 5 used as starting reagent (Figure S1 in the Supporting Information) in the solid-state reaction, in agreement with what was observed by Hreŝĉak et al. [26] The milled monoclinic Nb 2 O 5 formed an inhomogeneous mixture of solid solutions with various Na : K ratios.</p><p>The estimated cell volumes here evaluated for the orthorhombic Amm2, the tetragonal P4mm and the orthorhombic Pbcm phases (Table S1 in the Supporting Information) are slightly different compared to the corresponding values reported in the literature for purely stoichiometric compounds. In particular, we observed a shrinkage of KNbO 3 and an expansion of NaNbO 3 and K 0.3 Na 0.7 NbO 3 unit cell volumes. This evidence suggests that a partial diffusion of Na + and K + into KNbO 3 and NaNbO 3 respectively take place during calcination. This may be justified by the smaller Na + radius (rNa = 1.02 Å) compared to K + (rK = 1.38 Å). Since the reaction rate of Na 2 CO 3 / Nb 2 O 5 is one order of magnitude higher than K 2 CO 3 /Nb 2 O 5 , it is reasonable to expect that the average crystallite size dimensions will be larger for NaNbO 3 (2000 Å vs 1115 Å). At the same time, the micro-strain generated by the presence of K + in the NaNbO 3 unit cell is about twice compared to KNbO 3 (Table S1 in the Supporting Information). The tetragonal P4mm, which is the majority phase in the powder mixture, shows a slightly larger cell volume (63.773 Å 3 ) compared to the stoichiometry composition (62.572 Å 3 ), suggesting that this phase probably contain a larger amount of potassium than the nominal composition.</p><p>The XRD analyses of sintered pellets are reported in Figure 3. All ceramics possessed the typical perovskite-type structure. For x < 1 a tetragonal tungsten-bronze type phase has been detected in trace (see details of secondary phases in Figure S2 in the Supporting Information), which was formed during sintering due to the volatilization of alkaline ions. [27] A secondary phase, with orthorhombic symmetry and s.g Pnma, was also detected for x > 1. Since this impurity phase appears for the high concentration of MN, it is reasonable to relate it to the addition of dopant, which reaches the solubility limit around x ~1. The diffraction peaks of MN could not be observed for any pattern (Figure 3a). This seems in apparent contrast with the post calcined KNN-xMN patterns, which presented the Bragg reflections ascribable to MN phase (see Figure S3 in the Supporting Information). The disappearance of the MN peaks is then related to the full reaction between KNN and MN, which took place during the sintering process.</p><p>The patterns of sintered samples have been analyzed by the Rietveld method and the results are shown in Table 1. As it can be surmised, the single-phase homogeneity was obtained during the subsequent sintering step at 1100 °C, with the formation of a with formula K 0.5 Na 0.5 NbO 3 . Cell parameters and cell volume turn out to be in good agreement with the literature data (Table 1). The system with 0.5 wt.% of MN (KNN-0.5MN) showed a similar crystalline structure with a dominant orthorhombic Amm2 phase, while the Rietveld analysis performed on the KNN-1MN pattern, revealed the presence in trace (~3 wt.%) of the orthorhombic Pnma secondary phase. This still unknown phase possesses a KTiNbO 5 -type structure, also observed in other ceramics such as Rb(Mg 0.34 Nb 1.66 )O 5 and K(Fe 0.43 Nb 1.57 )O 5 and, as already mentioned above, its presence is probably related to the solubility limit of MN in KNN ceramics and/or volatilization and segregation of alkali elements during the thermal treatment. [28] Increasing the content of MN (KNN-2MN), a further increase of this second phase has been detected (~5 wt.%). The results show that the crystallographic parameters of the Amm2 phase are roughly similar for all samples (Table 1). However, as it can be better visualized in the magnification in Figure 3b for the (022) and ( 200) reflection peaks, a slight shift toward higher two theta angles, for increasing amount of MN, has been observed. Since the ionic radius of Mg 2 + (0.72 Å, 6 CN) is larger than Nb 5 + (0.64 Å, 6 CN), the addition of a low amount of Mg 2 + inside the KNN matrix is expected to produce an expansion of the unit cell, as observed by Li and co-authors. [29] For higher amount of Mg 2 + this effect is limited by the high oxygen vacancy concentration which produces a reverse tendency and thus a shrinkage of the unit cell. In this case, a decrease in the cell volume is observed with an increasing concentration of MN. This phenomenon could be explained by the formation of A-site vacancy in the ABO 3 perovskite formula, due to the addition of an excess of Nb 5 + ions (B-site) which causes an overall decrease in the cell volume. Similar behavior was found for KNN modified with another columbite-type structure dopant as CuNb 2 O 6 . [30] The equivalent cell volume for KNN-1MN and KNN-2MN, further confirms that the solubility limit of MN in the KNN matrix is roughly x ~1. Furthermore, the root-mean-square strain (r.m.s strain), which is a measure of the lattice distortion, for the Amm2 phase increases as a function of the amount of MN. Pure KNN a strain of about 6.0*10 À 4 (Table 1) while KNN-2MN has a strain of about 1.8*10 À 3 , which is one order of magnitude higher than pure KNN.</p><p>The SEM micrographs of surfaces of the KNN-xMN ceramics sintered at 1100 °C, are illustrated in Figure 4. The images show that the pure KNN and KNN-0.5MN (Figure 4a and Figure 4b) reveal a similar microstructure characterized by cubic-shape grains. [31] The KNN-1MN (Figure 4c) presents a well-packed microstructure characterized by a more homogeneous grain size. An evident limit of the grain growth was reached for the KNN-2MN sample, which showed a less compact microstructure, an increase of porosity and roundish grain morphology (Figure 4d). This evidence suggests that an excess of the orthorhombic Pnma secondary phase, most probably located at grain boundaries, inhibits the growth of the grains, also preventing the formation of cubic grains. The statistical calculation of the grain size reveals that KNN and KNN-0.5MN have a lognormal grain distribution with an average grain size of 3.38 μm and 3.25 μm, respectively. For a high amount of MN (x > 1) large particles of > 6 μm are not detected anymore and the grain size distributions, both for KNN-1MN and KNN-2MN, are characterized by a narrower, quasi-Gaussian curve, with an average grain size of 2.72 μm and 2.48 μm, respectively.</p><p>A similar behavior for higher amount of the impurity phase, has been observed in other ceramics systems, such as Mndoped BiFeO 3 and KCuTa 3 O 9 -modified KNN. [32] Figure 5 shows the dielectric permittivity and losses of KNN-1MN and KNN-2MN as a function of the temperature for increasing frequency from 1 kHz. The low temperature (T1) dielectric anomaly corresponds to the orthorhombic to tetragonal (O-T) polymorphic phase transition, whereas the high temperature (T2) corresponds to the phase transition to the paraelectric cubic phase (T-C). KNN-0.5MN ceramic has similar temperature dependence as KNN-1MN one (Figure 5) with welldefined anomalies. T1 takes place at 175.7, 175.8 and 94.4 °C, for KNN-0.5MN, KNN1-MN and KNN-2MN at 1 kHz, respectively, and is always lower than the one for pure KNN (200 °C). [6] Thus, the decrease of T1, as the amount of MN increases, is not linear. For KNN-2MN this low-temperature anomaly is much wider. This diffuse phase transition could reflect the crystal disorder confirmed by XRD analysis, in turn reflected in the distinct morphology of the grains in this ceramic (Figure 4d). T2 for the maximum permittivity is 396.1, 400.0 and 413.8 °C at 1 kHz, for 0.5, 1 and 2 % of MN, respectively. All temperatures are also lower than that for pure KNN (420 °C). [6] This reduction of temperatures is also found for other dopants. [19] The higher porosity of the KNN-2MN ceramic gives place to a lower value of the maximum permittivity at T2. Figure 6 shows d 33 as a function of the electric poling field. We can see that the KNN-0.5MN sample shows the highest d 33 value for 20 kV cm À 1 . The threshold value, beyond which the d 33 starts to decrease due to the mechanical deterioration of the sample, seems to be 20 kV cm À 1 for the pure KNN and KNN-0.5MN. The KNN-1MN shows endurance against degradation by electric field until 30 kV cm À 1 , though the saturation of d 33 is reached at 20 kV cm À 1 .</p><p>This behavior is most probably related to the increase in the amount of additive, which results in better densification and optimized microstructure for KNN-1MN (Figure 4c). The pure KNN and KNN-0.5MN continue to worsen while KNN-2MN undergoes the degradation only after the maximum d 33 is achieved at 30 kV cm À 1 . In Figure 6 is apparent that the increase of MN content above 1 % makes it necessary to use higher poling fields to achieve similar, slightly lower, d 33 . This slightly lower value is most likely a microstructural effect of the higher porosity of the ceramic (Figure 4d). In KNN-2MN, the need of a higher field to get almost similar d 33 is mainly due to the lower grain size of this sample. It is well-known that the ferroelectric domain width is proportional to the grain size. The more complex domain configuration and the higher surface pinning of domain walls, causing their reduced mobility, results in higher energy needed for their reorientation, thus higher electric fields. [33] The inhibition of grain growth, as already explained, is a compositional effect.</p><p>Table 2 shows a comparison between some properties of the ceramics in our work with the higher d 33 values (Figure 6) and ceramics of KNN doped with common sintering aids, such as Cu and Zn based compounds. Cu-doped ceramics show a much higher mechanical quality factor (Q m ) than the ceramics here studied due to what appears to be a major electromechanical "hardening" effect of Cu-based compounds, however Q m values are comparable with those of Zn-based compounds. [19,[34][35][36] However, the piezoelectric coefficient (d 33 ), planar coupling factor (k p ), relative dielectric permittivity (ɛ T 33 ) and dielectric losses (tanδ e ) are comparable.</p><p>All piezoelectric, elastic and dielectric complex parameters obtained from the radial resonance of the ceramics studied in this work are shown in Table 3. As a representative example, Figure 7a shows the measured complex impedance curves (modulus and phase) for the fundamental mode of the extensional radial resonance of a thin disk of KNN-0.5MN ceramic. Figure 7b also shows the equivalent plot of R and G, both the experimental and reconstructed peaks, used for the calculation of parameters in Table 3. It is apparent that the model used is accurate and reproduces well the experimental curves (Fig- ure 7b) as quantified by a R 2 factor close to 1. This happens to all ceramics (Table 3). Instead, the losses commonly reported are restricted to one dielectric (tanδ e ) and one mechanical (Q m ) factor. Piezoelectric losses arise from the friction of the vibration or strain that result as a response to an electric field or from the dielectric losses associated with the voltage or charge generated as a response of a mechanical stress.</p><p>They are an important parameter in the design of devices as they emerge as undesired hysteresis of the piezoelectric response and heat generation. What clearly emerges from Table 3 is that the addition of the doping agent increases the piezoelectric response with respect to the undoped composition. In particular, the enhancement of the charge coefficients, d 33 and d 31 , and coupling factors, k p and k 31 has been proved in the compositional range between 0.5-1 % of MN. The addition of MN also reduces the piezoelectric anisotropy as quantified by the ratio d 33 /d 31 . The wide low temperature dielectric anomaly of KNN-2MN (Figure 5b) influences the room temperature values of ɛ T 33 , also at the frequency of planar resonance. Permittivity at room temperature otherwise should be lower for this ceramic, since it contains higher porosity (Figure 4d), but, nevertheless, is the highest of the modified ceramics. All ceramics have higher compliances (s ij ), as well as lower stiffness (c ij ) and frequency number (N p ) than pure KNN, but also than the KNN-K 4 CuNb 8 O 23 doped ceramic, favoring the sensor performance. [19] All modified ceramics are less stiff at resonance than pure KNN. Losses are also calculated by the ratio between the frequency of the maximum of the recalculated R and G peaks and their width at half height, Q s and Q p , respectively, as losses affect the shape of these peaks. In this way the R, G peak plots are the fingerprint of the losses of the ceramic.</p><!><p>In this work lead-free K 0.5 Na 0.5 NbO 3 -xMgNb 2 O 6 ceramics (KNN-xMN) were successfully synthesized through an easy scalable conventional solid-state method. The effect of MN on KNN matrix has been evaluated. The sintered pellets show a perovskite structure with orthorhombic symmetry and Amm2 space group. An orthorhombic Pbcm secondary phase appears for concentration of MN higher than 1.0 wt.%. Furthermore, the addition of MN causes an increase of the lattice distortion of the Amm2 phase. Experimental results show that the densification of KNN ceramics can be significantly improved by adding an appropriate amount of MN (1 wt.%). KNN-1MN also shows a better microstructure, characterized by well packed grains with narrower, quasi-Gaussian, size distribution. The Pnma secondary phase seems to act as an inhibitor of the grain growth: in fact, the KNN-2MN sample shows smaller particle size and different particle shape characterized by roundish grains. A non-linear dependence of the phase transition temperatures T1 (O-T) and T2 (T-C) with the amount of MN was found. The addition of MgNb 2 O 6 reduces both temperatures with respect to those of pure KNN. The best properties were obtained in the compositional range between 0.5-1 wt.% of MN. KNN-0.5MN shows the highest charge and voltage coefficients with relatively lower anisotropy and piezoelectric losses (d 33 = 97 pC N À 1 ; d 31 = À 36.99 pC N À 1 and g 31 = À 14.04 × 10 À 3 mV N À 1 ; Q p (d 31 ) = 76 and Q p (g 31 ) = 69), while KNN-1MN possess the best electromechanical coupling factors (k p = 29.06 % and k 31 = 17.25 %), which can be correlated with the best microstructure achieved. This work provides a positive evaluation of the addition of MN on KNN ceramics. The simple manufacturing method is easily scalable and suitable for medium-large scale productions. Further optimization of the milling and sintering conditions could lead to even better electromechanical properties and higher densities.</p><!><p>KNN powders were prepared through solid-state reaction, starting from a mixture of K 2 CO 3 (Sigma Aldrich, � 99.995 %), Na 2 CO 3 (Sigma Aldrich, � 99.5 %) and Nb 2 O 5 (Alfa Aesar, 99.9985 %), in a molar ratio 1 : 1 : 2, respectively. Manipulations of the starting reagents have been conducted in an Ar Glove box machine (MBraun) with level of oxygen and moisture below 2 ppm, in order to prevent hydration, contamination and side reactions. 8 grams of powders were transferred into a stainless-steel vial together with 1 ball (stainless steel) of 10 g. The powders were mechanically treated for 12 hours at 875 rpm by using a Spex 8000 M Mixer/Mill and then transferred into an alumina crucible. The calcination step was conducted from room temperature to 825 °C for 4 hours using a heating rate of 3 °C/min and then cooled up to 25 °C with a cooling rate of 10 °C/ min. MgNb 2 O 6 powders were synthesized through solid-state reaction exploiting a similar procedure reported in the current literature. [37] MgO (Sigma Aldrich, � 99 %) and Nb 2 O 5 (Alfa Aesar, 99.9985 %) in a stoichiometric molar ratio 1 : 2, were mixed by high energy ball-milling (Spex 8000 M Mixer/Mill), for 24 hours at 875 rpm, and then thermally treated up to 1000 °C (dwell time: 1 hour) by using a heating rate of 5 °C min À 1 (the cooling step was realized with a rate of 10 °C min À 1 ). The corresponding X-ray diffraction pattern and Rietveld analysis are reported in Figure S4 in the Supporting Information. The modification process of KNN powders, was made by mixing appropriate amount of KNN and MN: four samples were prepared with increasing weight percentage of MN (0, 0.5, 1, 2 and 4 wt.%, respectively). KNN and MN were milled, with 5 ml of ethanol (Sigma Aldrich, purity > 95 %), for 24 hours at 875 rpm. The as-obtained slurry was then transferred in a beaker and heated in an oven at 150 °C for 4 hours to eliminate the solvent. The powders were finely ground in a mortar to obtain a fine particulate and mixed with few drops of polyvinyl alcohol (PVA) solution (3 wt.%) before compacting into a disk by means of a hydraulic press (220 kg min À 2 for 30 minutes). The pellets were thermally treated for 10 hours at 550 °C to eliminate all traces of PVA. Sintering was conducted at 1100 °C for 3 hours using a heating rate of 5 °C min À 1 and cooling rate of 10 °C min À 1 . Finally, bulk densities were measured by geometric method.</p><!><p>Structural investigations were conducted using a SMARTLAB diffractometer with a rotating anode source of copper (λ = 1.54178 Å) working at 40 kV and 100 mA. The spectrometer is equipped with a graphite monochromator and a scintillation tube in the diffracted beam. The patterns were collected in the angular range from 18°to 110°with a step size of 0.05°and a fixed counting time of 4 seconds per point. Quantitative analysis of the crystalline phases and structure determinations were performed with the MAUD software (Materials Analysis Using Diffraction), a Rietveld extended program. [38] Lattice parameters of the constituent phases were refined from the line peak positions after allowing a correction for the zero-offset, while crystallite size and lattice disorder contributions to the peak broadening were separated because of the wide angular range explored. Microstructure and morphology of the samples have been characterized by Quanta FEI 200 scanning electron microscope (SEM). Grain size distributions of the sintered pellets, evaluated on 100 grains, were estimated using the ImageJ software.</p><!><p>To measure electric properties, pellets were reduced in thickness by polishing to the proper thickness/diameter aspect ratio. Silver paste was attached on both surfaces of the thin disks and sintered at 400 °C for 30 minutes. After that, samples were increasingly poled in thickness under 10-35 kV cm À 1 at 130 °C for 15 minutes in a silicone oil bath, followed by field cooling (FC). Permittivity vs. temperature curves at frequencies above 1 kHz were measured using an automatic temperature control and capacitance-loss tangent data acquisition from an impedance analyzer (HP 4194 A). The quasi-static d 33 piezoelectric charge coefficient, which characterizes the sensor performance of the ceramic in the poling field direction, was measured with a Berlincourt d 33 -meter at 100 Hz. Complex impedance as a function of the frequency was measured with an impedance analyzer (HP 4192 A-LF) at the radial extensional resonance of the thickness poled thin disks. The related piezoelectric, dielectric and elastic material coefficients, including all the losses, were determined using the software for automatic iterative analysis of the complex impedance vs. frequency curves. The numerical method used was developed by Alemany et al. [39,40] A set of non-linear equations, which result when experimental data of complex impedance (Z*) or admittance (Y* = 1/Z*) at specific frequencies are introduced into the appropriate analytical solution of the wave equation, is solved iteratively until a convergence criterion is fulfilled. This set of equations is established for as many frequencies, which are automatically selected by the program, as unknown material coefficients. For the automatic determination of these frequencies, the program uses an alternative representation to the classical plots for Z* or Y* of modulus and phase as a function of the frequency. Instead, the peaks of resistance (R) and conductance (G) values are plotted vs. frequency, being Z* = R + iX and Y* = G + iB. The two main frequencies, f s and f p , used to establish the set of four equations to solve, are determined as those corresponding to the maximum values in the R and G peaks, respectively. For each iteration, the other two values of frequency are also determined automatically. The reconstruction of the R and G peaks is carried out once the material coefficients are calculated. The calculated coefficients are inserted in the analytical expression of the resonance and calculation of the complex admittance is performed by this model as a function of the frequency and the reconstructed peaks plotted. The residuals for these reconstructed R and G peaks to the experimental ones, quantified by the regression factor (R 2 ) accounts with the validity of the model for the resonance mode. The closer is the model to the experimental curves; the closer is R 2 to 1. For the planar mode, the complex material coefficients (P* = P'À iP'') directly determined in this analysis are the piezoelectric charge coefficient, d 31 , the dielectric permittivity, ɛ T 33 , and the elastic compliances, s E 11 and s E 12 . The value of d 31 allows analyzing the sensor performance in the perpendicular plane to the applied field, and the ratio d 33 /d 31 gives an insight of the anisotropy of the material. Besides, a number of other material coefficients are determined by the software from those, using well known relationships. [40] These allow us to analyze the performances of the ceramics as generator (piezoelectric voltage coefficient g 31 ) and energy transducer (electromechanical coupling factors and frequency number k p , k 31 , N p = f s (kHz)D(mm), where D is the diameter of the disk). For anisotropic materials, as the piezoelectric ceramics, the Poisson's ratio (σ) depends on the direction of extension and transverse deformation. Ceramics have two σ values, for planes parallel and perpendicular to the poling direction. From the radial mode of resonance, we can calculate the latter as σP = À s E 12 /s E 11 . A perfectly incompressible isotropic material deformed elastically at small strains would have a Poisson's ratio of exactly 0.5. Most piezoceramics exhibit values of about 0.3. In piezoceramics, σ allows also to quantify the in-plane anisotropy, the higher the σ the lower the anisotropy and the higher the mechanical coupling between various vibrational modes. Losses can be expressed for each material complex coefficient as loss tangent factor (tanδ = P''/P'), commonly used for the dielectric coefficients, or as a quality factor (Q=P'/P''), commonly used for the elastic coefficients. The mechanical Q factors (Q m ) calculated by the iterative method here used was compared with that calculated according to international standard methods. The latter was considered very accurate for both low and high loss materials, as well as for low and high electromechanical coupling factor materials. [41] The calculation of piezoelectric coefficients in complex form allows obtaining piezoelectric losses, not so commonly reported, but equally important for the design of devices.</p>

Chemistry Open

Multianalyte Physiological Microanalytical Devices

Advances in scientific instrumentation have allowed experimentalists to evaluate well known systems in new ways as well as gain insight into previously unexplored or poorly understood phenomena. Within the growing field of multianalyte physiometry (MAP), the microphysiometer development is building, instruments capable of electrochemically measuring changes in the concentration of various metabolites in real time. By simultaneously quantifying multiple analytes, these devices have begun to unravel the complex pathways that govern biological responses to ischemia and oxidative stress while contributing to basic science discoveries in bioenergetics and neurology. Patients and clinicians have also benefited from the highly translational nature of MAP, and the continued expansion of the repertoire of analytes that can be measured with multianalyte microphysiometers will undoubtedly play a role in the automation and personalization of medicine. This is perhaps most evident with the recent advent of fully integrated sensor arrays that can continuously monitor changes in analytes linked to specific disease states using non-invasive sensors and can deliver a therapeutic agent as needed without the need for patient action.

multianalyte_physiological_microanalytical_devices

Introduction<!>In Vitro Microfluidic Studies<!>Multianalyte Microphysiometer<!>Microclinical Analyzer (MCA).<!>Electrophysiological Event Monitoring<!>Multiplexed Reactive Oxygen and Nitrogen Species Sensing.<!>Microchip Electrophoresis with Amperometric Detection<!>Clinical Devices and Testing<!>Multiplexed Electrolyte Detection<!>Additional Biofluids<!>In Vivo Multianalyte Physiometry using Microanalytical Devices.<!>Strategies for Avoiding Electrochemical Interference<!>MAP Utilizing Microdialysate.<!>MAP Using Implantable Electrodes<!>Wearable Devices for MAP<!>Conclusions<!>

<p>The development of analytical devices incorporating sensors for multiple analytes has enabled new studies of physiological systems. For example, although electrochemical glucose sensors(1–3) have a well-established clinical presence in monitoring diabetics (1) and pre- and post- op surgical patients (4), fundamental studies on aerobic and anaerobic respiration are better accomplished through the detection of lactate, oxygen, and pH in addition to glucose (Figure 1) (5). When observing physiological effects with complex methodologies varying from monoculture microfluidic devices to clinical samples, sensing strategies often need to accommodate small devices, moderate-throughput experiments, and limited volumes of patient samples.</p><p>The subfield of multianalyte physiometry (MAP) has emerged to broaden the understanding of physiological pathways with multianalyte analysis using either a sensor array or single multiplexed detector to probe complex biochemical processes in higher dimensions (spatial and temporal) of phase space. Researched applications of electrochemical MAP sensors include observations of biological and chemical responses to stimuli, determination of the health status of organs on chips (OoCs), and development of novel clinical and point of care (POC) devices. When approaching the more complex task of detecting multiple analytes in living organisms, implantable ultramicroelectrodes are desirable due to decreased disruption in animal models and fast scan cyclic voltammetry (FSCV) minimizes adverse effects to the subject by minimizing the diffusion layer (6, 7). MAP devices' breadth of applications includes fundamental cellular studies, clinical and point of care testing platforms, and in vivo/in situ research.</p><!><p>Multianalyte physiological microanalytical devices currently encompass many different types of experimental methodologies and techniques. Microfluidic devices that focus on electrochemical measurements within the device are the focus of this review, as the electrode sensors are built within these self-contained systems. Examples of the different types of these devices include microphysiometers, microclinical analyzers, microfluidic electrophysiology, multielectrode arrays, and microchip electrophoresis coupled to electrochemical detection. Each of these sub-types is reviewed in the sections below.</p><!><p>Conventional microphysiometry was developed by Harden McConnell and co-workers leading to the Cytosensor Microphysiometer from Molecular Devices, an instrument that measured the extracellular acidification (i.e. pH changes) produced by the metabolism of approximately 100,000 to 1,000,000 live cells in a 3 μL microfluidic chamber using a light-addressable potentiometric sensors (LAPS) (Figure 2a). The Cytosensor has been modified (Figure 2b) into the multianalyte microphysiometer (MAMP) for detecting glucose, oxygen, and lactate in addition to pH inside the cell chamber. These additional analytes are detected amperometrically with immobilized glucose oxidase (GOx), immobilized lactate oxidase (LOx), and O2 is directly detected (Figure 2b) (5, 8–16). Investigations with the MAMP instrument include the metabolic effects on cellular bioenergetics of nutrient concentrations (12), chemical toxins (5, 8–11), harmful proteins (10, 11) non-invasive fluorescent chemical probes (13), and oxidative burst (14–18). Major shifts in cellular bioenergetics observed in the MAMP demonstrate shifts in extenet of utilization of aerobic or anaerobic respiration pathways (5, 8).</p><p>Multiplexing cellular bioenergetics and insulin detection can offer valuable information on the complicated physiology and metabolism of primary islets. For example, excised pancreatic islet physiology and metabolism were studied by combining an insulin selective electrode with electrodes for the previously mentioned analytes in the MAMP. This allowed the time-dependent study of insulin release resulting from a hyperglycemic environment (19).</p><p>Cellular bioenergetics studies of neurons using the MAMP has also revealed a strong correlation between neuronal adaptation and survival by developing a new model of transient ischemic attack (TIA) (12). Metabolic adaptation to stress is critical for cell survival but poorly understood. The metabolic compensation and survival of nutrient-deprived neurons relies on neuronal-glial chemical and physical communication. In this model of transient ischemic attack (TIA), primary neurons primed with short deprivations of oxygen and glucose in the days leading up to an otherwise lethal deprivation achieved increased survival rates over unprimed neurons (20). It was found that after priming the neurons, cellular ATP and neuronal oxygen consumption increased above baseline. After treating these primed neurons with an oxygen deprivation triggering an anaerobic pathway shift that would normally be lethal to the neurons, these primed neurons recovered normal lactate production in 30 min signalling a return to aerobic sufficiency. The rapid increase in aerobic respiration (monitored via oxygen consumption) observed after priming, coupled with the increase in cellular ATP indicates that the protective pathways in this ischemic model include the immediate increase in production of energy stores through aerobic respiration. This increase in energy production may help to prevent cellular starvation upon subsequent stress. This provides the first dynamic measurement used to identify essential events mediating neuronal injury in vitro and can aid in the identification of predictive biomarkers in injury. These microphysiometry data reveal that the greatest single predictor of neuronal survival is extracellular acidification; however, lactate levels, which are currently a key clinical indicator of injury, were not in fact, correlated with neuronal cell fate (12, 21).</p><!><p>Unlike in multianalyte microphysiometry where the sensors are embedded into the microfluidic cellular chamber, in a microclinical analyzer, the electrochemical measurements are in performed in separate chamber downstream. Thus, MCAs measurements can be multiplexed between multiple organs on a chip using a valve to select which organ effluent is to be measured. Thus, the MCAs trade off temporal and spatial resolution for increased resistance to biofouling of the electrochemical sensors. Our MCA is a self-calibrating device consisting of a microfluidic pump and valve that directs calibrant and sample solutions to a 26μL sample chamber where an array of platinum screen-printed electrodes (SPEs) amperometrically detects hydrogen peroxide produced by spatially separated enzyme films and potentiometrically detects pH via open circuit potential (OCP) shifts using electrodeposited iridium oxide. SPEs are robust, highly reproducible, and low-cost, resulting in a device that is easily customizable with interchangeable sensors capable of measuring glutamate, acetylcholine, calcium, potassium and/or sodium (18, 54). Custom multichannel potentiostats were created to accommodate simultaneous MCA measurements (56). Our investigations with the MCA instrument have focused on monitoring tissues and organs on chips (OoCs), (52, 53).</p><p>Other researchers have developed similar MCAs that are useful in toxicological testing of healthy or cancer cells (57, 58). Combined with other metabolites of bioenergetics, oxygen sensing inside their 10μL device increases the likelihood that changes are a result of cellular aerobic respiration. Running multiple simultaneous experiments with multi-sensor microsystems allows for a moderate-throughput screening of potential drug candidates (58). An electrolyte-focused MCA has been developed with four ion selective and three partially selective electrodes, specifically selective for potassium, sodium, hydrogen, calcium, amines, cations, and anions. These electrolyte MCAs are primarily concerned with signaling flux from ionic gradients. While direct pathway elucidations are not always exact, using chemometric analysis, training sets can be used to define a particular physiological state of interest. The combination of electrolyte sensing and multiplexed analysis was used to determine the ratio of dead to live cells in a cancer therapeutics screening (57).</p><!><p>Exocytosis of redox active species can be measured potentiometrically or amperometrically (22) and has led to numerous discoveries in vesicle transport. For example, the measurement of the relative concentrations of specific analytes, including octopamine, following release have been reported (23). While differentiating the molecules released is not always possible, the timing and frequency of events marked by the release of neurotransmitters can be recorded (24) and microelectrode arrays allow for both single cell and population analysis (25). Additionally, using experimentally determined parameters, the initial pore size before full release has been mathematically modeled (24). Nanopipette potentiometric electrodes and carbon fiber nanoelectrodes can be used separately or in combination to measure postsynaptic potential shifts and neurotransmitter release events (22, 24, 26). Measurements of synaptic events were performed in a microfluidic device which co-cultured superior cervical ganglion neurons and smooth muscle cells (Figure 2c) to confirm that synaptic events mimic in vivo conditions, in addition to improved axon orientation (26).</p><!><p>Oxidative stress is marked by the production of reactive oxygen species (ROS) and reactive nitrogen species (RNS), specifically hydrogen peroxide (H2O2), nitric oxide (NO-), nitrite (NO2-), and peroxynitrite (ONOO-). By amperometrically measuring at four different potentials, the concentrations of ROS and RNS are determined with a physiological model consisting of four linear equations delimiting current contributions of each compound. The detection of ROS and RNS concentrations has been used to quantify anti-oxidant capacity, to help determine basic cellular function of macrophages, and to dcompare the effectiveness of cancer therapies. The antioxidant capacity of a superoxide dismutase mimicking manganese complex (27) was determined with amperometric detection of ROS/RNS produced by macrophage cells. Utilizing the redox nature of ROS/RNS reactions, a new cell-free method for determining antioxidant capacity showed that trolox, ascorbic acid, gallic acid, and caffeic acid, compounds with known antioxidant properties, electrochemically simulated the redox effect of ROS in a microfluidic device (28). In a study of amacrophage's ability to control ROS/RNS release and implement protective measures, two platinum electrodes were used to observe dynamic intracellular and extracellular ROS/RNS concentrations. By inserting and sealing one electrode into a macrophage cell while simultaneously measuring outside the same cell, ROS/RNS leakage through phagolysosmes and neutralization to avoid premature oxidative damage were observed (29). Experimental drug molecules of the ferrocifen class were tested on two breast cancer cell lines to determine how treatment with each effected mechanical depolarization of the cells. As the multiplexed detection of ROS/RNS species allows for the concentration of each species to be quantified, the analysis of each drug's effectiveness showed that H2O2, NO-, NO2-, and ONOO- production was not uniformly affected and that the differences in the mechanism of action could be determined with more study (30). After being established as an adaptable method, the multiplexed amperometric detection of ROS/RNS at four potentials was integrated into a microfluidic device that separates each analytical measurement into individual 38μL cell chambers (31).</p><!><p>As an alternative approach to multielectrode formats, microchip electrophoresis methods have been applied to multianalyte physiometry. By adding a separation step prior to electrochemical detection, these devices can increase the number of analytes detected on a single electrode. The challenge in pairing microchip electrophoresis (ME) and amperometric detection is noise from high ME operating potentials, but noise reduction techniques, such as in-channel detection and an electrically isolated potentiostat, can be implemented (32). ME with amperometric detection (Figure 2d) has been used to detect redox active biomolecules, interfering species that are electroactive at the same potentials as target molecules (32–40), and glucose using a glucose oxidase modified working electrode (34). Methods have been optimized to observe nitric oxide production in macrophages stimulated with lipopolysaccharide (LPS) by analyzing nitrite ions (35) and dopamine metabolites from rat brain slices (39). ME with amperometric detection is a versatile technique that can detect multiple analytes, and does not required specific modification of each electrode sensor. This enables ME to be more generic as the same basic device does not need specific analyte modifications to detect another set of species.</p><!><p>Clinical testing and point of care (POC) devices incorporates multianalyte physiometry to monitor, diagnose, or treat a patient. With hundreds of possible tests available, clinical MAP devices focus on patient monitoring, organ function status, or disease diagnosis These focused assays significantly narrow the number and type of tests performed. Typically, a physician starts with a few dozen analytes in clinical blood tests that are used for monitoring a patient during recovery, treatments and procedures; and transitions to organ function and disease biomarker specific tests if irregularities in the basic tests are present (41–43).</p><p>The i-STAT ® handheld blood analyzer (Abbott Point of Care) has been a staple in hospitals and doctor's offices around the world for almost 25 years and frequently a featured point of care device in clinical chemistry literature (41, 43–47). The self-contained testing cartridges (Figure 3a) have nineteen commercially available variations that contain either single-analyte or multianalyte electrode arrays (41, 48). Low operating volumes (17–95μL) accommodate comparative validation of a newly developed MAP device or improved detection schemes, especially if the amount of sample is limited. The cartridges contain a calibrant solution and one or more electrode arrays with exposed contact pads to easily connect to the handheld device. Quality control during the analysis includes a pseudo-sample statistical treatment to determine validity of the cartridge, followed by the sample, which is evaluated for concentrations, bubbles, and volume collected. If the calibrant and sample measurements meet the statistical demands, a digital output reports concentrations determined via single concentration point calibration (41, 47). In most circumstances, pseudosample analysis is not ideal for treating data sets. To limit variability in the experimental conditions and determine data skews within the population, high precision commercial microfabrication of cartridges and regular clinical chemistry evaluations are relied upon (45–47, 49, 50). As a result of the reliability and immunoassay availability, an i-STAT device and cartridge has been modified to detect point mutations in genetic diseases and strain differentiation in viral and bacterial infections after an eight minute on-device PCR (51). The small, rugged nature of the i-STAT device lends to use in the emergency room and by care providers who travel between resource-constrained hospitals (46). Even though the i-STAT has low operating volumes conducive to some of the larger OoCs and bioreactors, the i-STAT lacks automated fluid handling and is targeted to physiometry measurements of macroscale samples.</p><!><p>A prominent category of physiological tests is monitoring of electrolytes, specifically sodium, potassium, calcium, magnesium, and chloride (41). Electrochemical ion detection is most commonly achieved via ion selective electrodes and potentiometric techniques (56). While many ion selective membranes designed for ISEs are commercially available, innovative multiplexed, low-interferent, ion selective sensors for ecological and physiological purposes could reduce the number of electrodes needed in an array potentially leading to reduced detection volumes. Recently described layered devices utilized cyclic voltammetry to measure concentrations of multiple ions based on a thin membrane embedded with multiple ionophores covering an ion-to-electron transducing material. Ion-transfer at the membrane surface occurs at specific potentials for different ions depending on the ionophore used (59–62). While not all materials utilizing the thin membrane sensors are robust enough for some biological samples, a recent polyurethane ionophore was able to detect lithium and potassium ions in a single scan in undiluted plasma and serum (61). ISEs enhanced with ion-exchange exclusion membranes have improved the separation efficiencies and sensor lifetimes enabling them to be used in complex clinical samples (63–66). In addition to multiple electrolyte detection on a single sensor, fully automated ISE arrays have detected monoatomic and polyatomic ions for analysis of the carbon cycle in freshwater ecosystems (67).</p><!><p>POC sensors need to be able to handle measurements in various biological matrices, such as blood, cerebrospinal fluid, urine, intraocular fluid and saliva, yet handle a wide range of concentrations and potential interferences that vary between matrix fluid types (42–44, 68–70). Many new devices are being designed to avoid finger-stick blood draws by optimizing testing in fluids requiring less invasive sampling (70, 71). One such MAP device was shown capable of measuring glucose, lactate, and cholesterol in blood and saliva with organic electrochemical transistors. In this device, analyte specific oxidase enzymes were immobilized using chitosan modified with a ferrrocene redox mediator, onto poly(3,4-ethylene-dioxythiophene):poly(styrene sulfonic acid) (PEDOT:PSS) coated gold electrodes, which effectively lowered the operating potential to 100–200mV and reduced the number of redox interferences. An additional electrode was included for the purposes of background subtraction and was coated with all the same materials except bovine serum albumin (BSA) replaced the enzymes. The microfluidic flow of the sample in PBS through the device was controlled by manual compression of the poly(dimethylsiloxane) housing. Sample injection was accomplished by manual release of the pressure on the device resulting in sample flow to electrodes upon channel re-expansion (Figure 3d) (71).</p><!><p>Due to the failure of in vitro models to accurately predict the efficacy of certain drugs in vivo, the need for animal and human trials remains. Several factors may contribute to the discrepancy between predicted vs observed therapeutic effects, including poor bioavailability of the compound and off-target toxicity of the drug or its metabolites. Although reductionist approaches have provided insight into the complex biochemical pathways that exist within the human body, it can be particularly difficult to predict off-target effects prior to in vivo testing. Additionally, the shift of the medical field toward personalized medicine makes it advantageous to develop methodologies that allow for continuous patient monitoring for an extended period of time.</p><!><p>Measurements made in the undefined environmental conditions within an organism may encounter electrochemical interference. This interference can come from both direct sources, such as redox active species, and indirect sources, such as biofouling, and commonly manifests as artificial changes in current or potential, plaguing electrochemical investigations. The need to measure signal coming only from analytes of interest has led investigators to develop several methods capable reducing the impacts of electrochemical interference, including 1) using electrostatic polymer barriers (6, 76, 81, 82, 84, 91, 95), 2) shifting the overpotential through electrode modifications (4, 77, 78, 89, 90, 96, 97), and 3) limiting electrode exposure time to the interfering species (98).</p><p>Polymer barriers such as polypyrrole and Nafion films can reduce the interference of redox active species, such as dopamine (DA) and ascorbic acid (AA), commonly encountered when making electrochemical measurements in vivo. These polymers are charged, and as such they form an electrostatic barrier around the electrode that can prevent the approach of charged species. For example, 3,4-dihydroxyphenyl-L-alanine (L-DOPA) is used to treat Parkinson's disease but polymerizes on the surface of carbon microelectrodes forming melanin and acting as an interferent (84). This biofouling is largely mitigated at high concentrations of L-DOPA by electrochemically preconditioning a carbon microfiber electrode and electrodepositing a thin film of Nafion onto the surface of the electrode, maintaining the sub-second response times associated with fast-scan cyclic voltammetry (FSCV) (84). Without the Nafion layer, L-DOPA treatment appeared to inhibit the release of DA, but with the Nafion layer, L-DOPA treatment was observed to enhance DA release in neural pathways associated with Parkinson's disease, which highlights the role of electrode design in understanding biochemical pathways in vivo (84).</p><p>Direct interference can also be avoided by shifting the potential at which amperometric measurements are made (4, 77, 78, 89, 90, 96, 97). Shifting of the needed overpotential can be accomplished by either coating or impregnating the working electrode with dyes such as Meldola's blue (MB) and Prussian blue (PB). For example, carbon electrodes modified with MB can be used to monitor enzymatic reactions involving the production of NADH, but AA interferes with these measurements (97). NADH is frequently detected at ~0.6 V, the same potential as AA. Incorporation of MB into the electrode shifts the needed potential down to +0.1 V 24 where AA does not become oxidized and thus does not contribute to the measured current. Measuring NADH has been used to diagnose and monitor the intrahepatic cholestasis of pregnancy (97). When diagnosing and monitoring this it is imperative that biosensors utilizing 3-HSD respond only to analytes of interest where artificially large currents resulting in a false positive could lead to unnecessarily inducing labor.</p><!><p>Monitoring metabolic changes within the brain can help elucidate the biochemical basis of a variety of neural pathologies. Because AA is a known interferent in experiments such as these, shifting the potential used can help signal overlap with analytes of interest. Secondary enzymes such as horseradish peroxidase (HRP) can be included to shift the potential needed for amperometric detection of glucose and lactate down to −0.1 V by adding ferrocene to the buffer used to carry the dialysate sample through FIA systems (4, 76–80). The use of rapid sampling microdialysis (rsMD) allows for on-line measurements of patient samples in FIA systems (4, 76–80). When coupled with other methods of monitoring brain activity such as electrocorticography, correlations can be made between extracellular concentrations of glucose and lactate and spreading depolarizations (SDs) in perilesional tissue following traumatic brain injury (76, 78). After the SD wave passes through the perilesional tissue, significant decreases in glucose and increases in lactate concentrations are noted, likely the result of cellular recovery (78). The observation that the changes in extracellular glucose and lactate concentrations occur after the SD wave has passed successfully demonstrates that the SD wave is likely the cause and not the effect of changes in cellular metabolism. For patients experiencing frequent SDs, the extracellular glucose concentration fails to reach pre-SD levels before the next wave again depolarizes the cells, suggesting that a prolonged hypoglycemic milieu may contribute to the poorer outcome for these patients (78).</p><p>As with traumatic brain injuries, localized physiological monitoring after surgical procedures can provide insight into recovery mechanisms as well as provide a personalized approach to medicine to predict patient outcome and alter the course of treatment as necessary. One such procedure is surgical anastomosis, the tethering of tubes or channels such as blood vessels or intestines. Following an anastomosis, the lactate to glucose (L:G) ratio determined through FIA using rsMD may continue to increase near the surgical site indicating local hypoxia (4, 77). Ischemic conditions, which often require surgical intervention if prolonged, have been evaluated in porcine bowels (77). Using an intramurally inserted microdialysis probe into porcine bowels, glucose and lactate were measured both before and after repeated simulated ischemic conditions (77). As ischemia set in, the concentration of glucose decreased with a simultaneous rise in lactate concentration. Upon reperfusion, the levels returned to baseline, but the effects of a secondary ischemic event were more severe than those seen in the first, suggesting that the tissue did not fully recover from the first ischemic event or was more susceptible to future events (77). In experiments such as these, the ratio of the analytes provides a better picture of the metabolic processes as slight changes in the position of the microdialysis probe can alter the measurements dramatically. This is in part due to the flushing effect of the microdialysis perfusion fluid over time causing artificially decreased absolute concentrations of analytes being measured under conditions where nutrients are not being replaced as seen in ischemia.</p><p>Monitoring of free flap surgical procedures through FIA with rsMD predicts patient outcome using the L:G ratio (4). After anastomosis of the arteries and veins of the free flap, the L:G ratio is expected to return to near pre-operative baselines as the nutrient supply is restored and wastes are carried away. Cases in which the ratio did not return to baseline but instead continued to steadily increase were those where the anastomosis had failed or a thrombosis developed (4). Interestingly, the effect of the topical vasodilator papaverine was also observed in real time as absolute concentrations of glucose and lactate both increased due to the increase in blood flow to the area (4).</p><p>The microdialysis platforms used in these studies have limited temporal resolution in part due to the downstream injection system. However, the use of microdialysis probes is not limited to FIA platforms. Continuous measurements of the microdialysate using traditional enzymatic biosensors can improve temporal resolution (76). In the case of SD wave passages, where the L:G ratio increases as discussed above (78), on-line sensors are capable of continuously measuring glucose and potassium ion concentrations. Using a potassium ion selective electrode and the glucose biosensor to monitor induced SD in mice, temporal resolution was improved by a factor of 60, down to 1 second (76). Using this method, the decrease in extracellular glucose was temporally resolved from the increase in potassium ion concentration, demonstrating that the drop is glucose is in response to the increased metabolic demands of the neurons undergoing repolarization (76).</p><!><p>Electrodes may also be directly implanted in vivo to probe complex biochemical pathways that are difficult or impossible to reproduce in vitro, such as those seen in the brain. These in vivo electrodes can therefore preserve information about sub-second cellular responses, a daunting task for microdialysis and flow injection methods. However, implanted electrodes present their own set of challenges: the sensitivity of the electrodes must remain constant and they must be very small or risk a trade-off of spatial resolution for temporal resolution.</p><p>Two broad categories of electrodes have proven efficient in maintaining spatial resolution for in vivo physiological measurements: metallic microelectrode arrays (6, 81, 82) and carbon fiber microelectrodes (7, 83–88). Metal microelectrode arrays are slightly larger than carbon fiber electrodes because they spatially separate the surface at which each analyte is detected by electrode-specific functionalization such as enzymatic or ion-selective films. However, this separation allows for more specific detection of each analyte of interest by decreasing the degree of interference between different analytes while maintaining high sensitivity for individual analytes (6, 81, 82, 89–92). Some metal electrodes are not suitable for implantation despite their ubiquity in in vitro devices. For example, Ag/AgCl film electrodes have been shown to elicit inflammatory responses in vivo (81). Alternatively, IrO2 is a biocompatible reference electrode that exhibits long-term stability and low noise in vivo. This electrode is, however, sensitive to pH changes and may only be suitable in regions of the body where pH changes are minimal, such as the brain (81). Metal electrodes have been used to measure physiological changes in the brain in response to mechanical and chemical stimuli. For example, the release of neurotransmitters in the striatum of rats in response to electrical stimuli and physical stresses can be monitored with sub-second temporal resolution using platinum electrodes (6). The second category of implantable electrodes, carbon fiber microelectrodes, has superior spatial resolution relative to metal electrodes. Carbon fiber microelectrodes are frequently used with FSCV to quantitate catecholamines such as DA (7, 84–87). Because DA adsorbs well to the surface of carbon fiber microelectrodes, faster scan rates are used to maximize signal. Traditionally, this is accomplished by using a triangular waveform, but novel waveforms such as the "sawhorse" scan which briefly holds the potential above +1.0 V further improve sensitivity by opening up sites on the electrode for DA to bind (85). This can be used to deconvolute signals arising from mixtures of redox active molecules such as adenosine, ATP, and H2O2 (93).</p><p>Long-term studies with implantable electrodes are limited by their decreased sensitivity over time. Loss of sensitivity is primarily due to biofouling and is compounded by the inability to recalibrate electrodes once implanted (82). Biofouling of GOx, LOx, and pyruvate oxidase (POx) coated electrodes results in a diffusion barrier shown to decrease sensitivity in vivo (82). However, the purposeful use of diffusion barriers, such as microdialysis membranes has been shown to extend the linear range of electrodes as well as reduce biofouling. For example, LOx films protected by a microdialysis membrane better retained sensitivity compared to GOx and POx films without the protective membrane (82). Alternatively, FSCV can be used to extend the lifetime of implanted electrodes by eliminating the assumption that the electrode sensitivity is constant during the course of an experiment. It does so by fitting the total background current and switching potential to a model with four regression coefficients (83). This method has accurately predicted the sensitivity of a carbon fiber microelectrode for a variety of analytes in vivo, including DA, AA, H2O2, and H+ (83).</p><p>Spatial mapping can be accomplished by electrochemically evaluating how cells respond to different stimuli. By determining how cells respond to DA, medium spiny neurons can be subtyped, mapping the nucleus accumbens in behavioral studies (87). When iontophoresed stimulants are not redox active in the potential range being swept with FSCV, redox active compounds such as DOPAC (88) and acetaminophen (87) can be added to the iontophoresed solution to act as internal standards. Glutamate, an iontophoresed stimulant that is not redox active, can be accurately quantified by such internal standards (88).</p><p>By simultaneously using electrodes at different locations, responses in different regions of the brain may be seen as a result of a single stimulus (7). For example, DA can be tracked in the nucleus accumbens while 5-HT is tracked in the substantia nigra pars reticulata to investigate the mechanisms governing their release. Upon stimulation, DA release in the nucleus accumbens is 300 times greater than 5-HT release in the substantia nigra pars reticulata despite them being similar in overall concentration within their respective tissue (7). By selectively inhibiting enzymes responsible for the synthesis, packaging, release, uptake, and metabolism of these two neurotransmitters, the dependence of DA transmission on synthesis and repackaging and the tight regulation of 5-HT transmission by reuptake and degradation pathways were revealed. Additionally, the severe neurological consequences resulting from the co-administration of 5-HT transporter and monoamine oxidase inhibitors were observed (7).</p><!><p>As manageable chronic conditions such as diabetes mellitus, hypertension, and hyperlipidemia become more prevalent, the demand for minimally invasive continuous monitoring devices increases. Wearable devices are now capable of monitoring many different analytes simultaneously, have readouts that interface with mobile phones, and can deliver a therapeutic agent (89). Devices that utilize bodily fluids other than serum and cerebral spinal fluid such as sweat and urine are becoming increasingly common in part due to the ease by which samples can be obtained (89–92). The 'diabetes patch' sits on the surface of the skin, monitors temperature, humidity, pH, and glucose levels, and delivers Metformin as needed. In the device, glucose is measured by the amperometric detection of peroxide on gold-doped graphene modified with PB and GOx at −0.05 V. By using additional sensors to assess pH and temperature, corrections can be made for the activity of GOx, yielding more accurate glucose readings. When hyperglycemic conditions are detected, a thermal actuator will melt polymeric microneedles incorporating Metformin, releasing it into the bloodstream. This patch is capable of continuously monitoring glucose levels over an entire day, and such non-invasive platforms may someday replace current methods of monitoring and managing blood glucose levels.</p><p>Wearable MAP microanalytical devices have also been utilized in exercise training regimens to gain insights into individual performance. Two of the most promising are a 3D-printed microfluidic device and 'smart bands.' The 3D-printed microfluidic devices have a subcutaneous microdialysis probe that allows external needle-based electrodes to continuously monitor glucose and lactate levels (94). The needle electrodes are removable from the device housing and can be easily modified to measure different analytes as desired. Although technically wearable, this device is still reliant upon an external potentiostat and requires a clinician to subcutaneously insert the microdialysis probe (94). On the other hand, a recently developed 'smart headband/smart wristband' can simultaneously and non-invasively measure glucose, lactate, Na+, and K+ levels in sweat (90). In these smart bands, GOx- and LOx-chitosan-SWCNT films on PB are used for chronoamerpometric detection of glucose and lactate, respectively, whereas Na+ and K+ levels are determined via open circuit potential using ionophores (Na ionophore X and valinomycin) in ion-selective membranes atop PEDOT:PSS. When access to water is restricted during exercise while wearing the flexible integrated sensor array (FISA), dehydration can be clearly seen when concentrations of sodium and potassium begin to significantly increase. In combination with the real-time profiles of glucose and lactate, athletes undergoing intense training can avoid over-exertion and gain physiological insights into individual performance.</p><p>Wearable FISAs have also incorporated ion-selective membranes to continuously measure Ca2+ and pH levels to determine hydration and electrolyte levels. Ca2+ concentration and pH are measured through films incorporating calcium ionophore II on PEDOT:PSS and electropolymerized polyaniline.(92) As with sodium and potassium, calcium ion concentration can be monitored with open circuit potential measurements in sweat during the course of a workout, and the in vivo stability of the reference electrode can be enhanced by the incorporation of a poly-vinyl butyral layer to maintain chloride ion saturation at the Ag/AgCl film with little interference from other common cations, including NH4+, Mg2+, K+, and Na+. The results obtained with the FISA for both pH and Ca2+ aligned well with the results from commercial pH meters and ICP-MS, the gold-standard for Ca2+ measurement, indicating that these wearable platforms can be used in clinically relevant conditions such as hyperparathyroidism and kidney stones.</p><p>The quantitation of other metal species bears clinical relevance due conditions such as Wilson's disease and acute heavy metal poisoning. Heavy metals can be detected by square wave anodic stripping voltammetry relying on the inherent redox potentials of the metal species being evaluated. The redox potentials of zinc, copper, cadmium, lead, and mercury are sufficiently separated to allow their quantitation in a complex mixture (91). The strong correlation between values obtained from a sweat sample using ICP-MS and the wearable sensor (91) indicates that these wearable sensors can also be used reliably for heavy metal detection in addition to the analytes previously discussed (90, 92). These advances significantly expand the number of cationic species that can be continuously evaluated in biofluids such as sweat and urine with wearable sensors, thus enlarging the diagnostic toolbox in the field.</p><!><p>The in-house modification of commercially available instruments to convert single analyte detection systems into those capable of measuring multiple analytes has largely become obsolete in part due to the advent of screen printed electrodes and 3-D fabrication techniques. The inclusion of customizable sensors into microfluidic devices has enabled multianalyte detection under physiological conditions, allowing researchers to make advances in the fields of bioenergetics, toxicology, and neurology. These devices also show great promise in clinical settings due to their capabilities of continuously monitoring vitals and predicting patient outcome. As the medical field shifts toward more personalized approaches, the widespread adoption of devices that can perform multianalyte detection is inevitable as evidenced by the continued success of the i-STAT. Outside of research and clinical settings, non-invasive smart bands are beginning to allow athletes to optimize training regimens, and the self-contained nature of these robust fully integrated sensor arrays in combination with their low power usage makes them ideal for use in low resource monitoring settings as well. Overall, multianalyte investigations will continue to grow in number as the monitoring of a panel of analytes is more broadly recognized for its utility in gaining a better understanding of fundamental cellular processes and disease states.</p><!><p>Schematic of a cell showing cellular bioenergetics pathways. Metabolites commonly detected with multianalyte physiometry (lactate, glucose, oxygen, and acid) are highlighted. Abbreviations: ADP, adenosine diphosphate; ATP, adenosine triphosphate; G6P, glucose 6-phosphate; NAD, nicotinamide adenine dinucleotide; NADH, nicotinamide adenine dinucleotide hydrate; TCA, tricarboxylic acid. Adapted with permission from Reference 10. Copyright 2006, Elsevier.</p><p>Multianalyte physiometers. (a) Side cross section and (b) bottom view of a CytosensorTM modified into a multianalyte microphysiometer by electrode addition. The four added platinum electrodes include three working electrodes and one counter electrode. The working electrodes detect glucose, lactate, and oxygen. Panels a and b adapted with permission from Reference 5. Copyright 2004, American Chemical Society. (c) Samples are stop-flowed into the device, and pH is measured through light-emitting diode (LED)-illuminated light-addressable potentiometric sensors (LAPS). Panel adapted with permission from Reference10. Copyright 2006, Elsevier. (d) Schematic of a multianalyte physiometer based on a glass chip that combines a cell cultivation chamber, microfluidics, and metabolic monitoring. Oxygen and pH are measured in the cell culture area, and biosensors for lactate and glucose are connected downstream by microfluidics. The wafer-level fabrication features thin-film platinum and iridium oxide microelectrodes on a glass chip, microfluidics in an epoxy resist, a hybrid assembly, and an on-chip reference electrode. Panel adapted with permission from Reference 21. Copyright 2014, Royal Society of Chemistry.</p><p>Photograph of a microclinical analyzer inset with a schematic of a screen-printed electrode. The pump and valve work together to flow 26μL of buffer, calibrants, and/or sample into the sample chamber containing the electrodes. From left to right, electrodes are modified to detect pH (blue), glucose (yellow), oxygen(middle), and lactate (pink). The far right electrode is an Ag/AgCl quasi-reference. Photograph courtesy of Dmitry Markov.</p><p>Clinical and wearable devices. (a) i-STAT handheld device. Adapted with permission from Reference 48. Copyright 1998, American Chemical Society. (b) Photograph and schematic of the selective multianalyte detection in complex media using the finger-powered OECT array. Photograph shows a red-colored solution that was pressure driven from the inlet through the sensing areas, as indicated by the vertical arrow. Adapted with permission from Reference 70. Copyright 2016, John Wiley & Sons. (c) A fully integrated wearable multiplexed sensing system on a subject's arm. Adapted with permission from Reference 91. Copyright 2013, Royal Society of Chemistry. Abbreviations: BSA, bovine serum albumin; OECT, organic electrochemical transistor</p><p>Summary of physiological systems and their multianalyte readouts in microanalytical devices.</p>

PubMed Author Manuscript

Fine-tuning the spike: role of the nature and topology of the glycan shield in the structure and dynamics of the SARS-CoV-2 S

The dense glycan shield is an essential feature of the SARS-CoV-2 spike (S) architecture, key to immune evasion and to the activation of the prefusion conformation. Recent studies indicate that the occupancy and structures of the SARS-CoV-2 S glycans depend not only on the nature of the host cell, but also on the structural stability of the trimer; a point that raises important questions about the relative competence of different glycoforms. Moreover, the functional role of the glycan shield in the SARS-CoV-2 pathogenesis suggests that the evolution of the sites of glycosylation is potentially intertwined with the evolution of the protein sequence to affect optimal activity. Our results from multimicrosecond molecular dynamics simulations indicate that the type of glycosylation at N234, N165 and N343 greatly affects the stability of the receptor binding domain (RBD) open conformation, and thus its exposure and accessibility. Furthermore, our results suggest that the loss of glycosylation at N370, a newly acquired modification in the SARS-CoV-2 S glycan shield's topology, may have contributed to increase the SARS-CoV-2 infectivity as we find that N-glycosylation at N370 stabilizes the closed RBD conformation by binding a specific cleft on the RBD surface. We discuss how the absence of the N370 glycan in the SARS-CoV-2 S frees the RBD glycan binding cleft, which becomes available to bind cellsurface glycans, and potentially increases host cell surface localization.

fine-tuning_the_spike:_role_of_the_nature_and_topology_of_the_glycan_shield_in_the_structure_and_dyn

Introduction<!>Results<!>N234-Man5<!>N234-Man3<!>N234-Man9 with FA2G2 at N370<!>Discussion<!>Conclusions

<p>Spike (S) glycoproteins mediate the adhesion and fusion of enveloped viruses to the host cell, initiating viral infection. [1][2][3] The interaction with the cell-bound receptor leads to a complex conformational change, dependent on the S architecture, 4,5 that terminates with the fusion of the viral envelope with the host cell's membrane, giving the virus access to the cellular machinery for replication. 3 Viral envelope S are heavily coated with a dense layer of complex carbohydrates, also known as a glycan shield, that performs different intrinsic and extrinsic functions, from modulating protein folding, stability and traf-cking, to masking the virus from the immune system and mediating contacts with lectins and antibodies. 1 The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has a trimeric S glycoprotein 6,7 protruding from the viral envelope surface. 8,9 The SARS-CoV-2 S has 22 N-glycosylation sequons per protomer, of which at least 18 appear to be consistently occupied in different constructs, 8,[10][11][12][13][14][15] and O-glycosylation sites with signicantly lower occupancy. 11,12,16 A unique feature of the SARS-CoV-2 S glycoprotein's architecture is the key role of the glycan shield in its activation mechanism. 17,18 Binding of SARS-CoV-2 S to its primary receptor, namely the angiotensin-converting enzyme 2 (ACE2), 19,20 requires the opening of one (or more) receptor binding domains (RBDs), which need to emerge from the glycan shield to become accessible. 9,17,18,[21][22][23] The RBD opening creates a cavity within the SARS-CoV-2 S prefusion trimer's structure, 6,7 see Fig. 1, which extends deeply into the trimer's core. In the absence of strategically positioned N-glycans as a support, 17 upon RBD opening this large pocket would be lled by water molecules, likely weakening the S prefusion structural integrity, especially considering the SARS-CoV-2 S pre-cleaved polybasic furin site at the S1/S2 boundary. 24,25 Multi-microsecond molecular dynamics (MD) simulations supported by biolayer interferometry experiments 17 have shown that this structural weakness is effectively recovered by the N-glycan at position N234, where a site-specic large oligomannose 8,10,14,15 is able to ll the cavity, supporting the RBD open conformation. 17 Furthermore, MD simulations have also shown that the Nglycans at positions N165 and N343, see Fig. 1, are directly involved in important interactions with residues of the open RBD, supporting its open conformation 17 and mediating (or gating) its transition from open to closed, 18 respectively.</p><p>The crucial role of the N-glycans at N234, N165 and N343 is exerted through the contacts these structures can make with protein residues, both in the open RBD (of chain B, following the PDBid 6VYB nomenclature) and in the adjacent closed RBD (chain C) at either side of the empty cle, 17 see Fig. 1. Their ability to engage in effective interactions is intrinsically linked to the nature, size, sequence and branching of the N-glycans at these sites, opening the oor to a broader discussion on the relative structural stability and dynamics of different S glycoforms. This is a very important point to explore, especially in view of the design of specic antiviral therapeutic strategies targeting glycosylation, 29 yet a very difficult (if not impossible) one to systematically address experimentally.</p><p>In this work, we present the results of a set of multimicrosecond MD simulations of different SARS-CoV-2 S glycoforms aimed at characterizing the effect of changes in the type of glycosylation at positions N234, N165 and N343, see Fig. 1, while the rest of the glycan shield is represented consistently with a stable recombinant S prefusion trimer (S trimer ). 10,17 More specically, we investigated if shorter oligomannoses structures at N234, such as Man5 and paucimannose (Man3), see Fig. 1, could occupy the empty cle in the open SARS-CoV-2 S as effectively as larger oligomannoses such as Man9, characteristic of a highly stable prefusion S trimer . 10 As an important note, Man5, rather than Man7/8/9 is found to be present, or even to be the dominant glycosylation type at N234 in the virus, vaccine epitopes and in some recombinant S constructs. 8,11,30 We also explored how the role of oligomannose structures at N234 is supplemented by the complex N-glycans at N165 and N343, which can form a tight network of glycan-glycan and glycanprotein interactions that stabilize the orientation and dynamics of the open RBD across different possible orientations.</p><p>Finally, we also investigated the effect of a mutation, unique to the SARS-CoV-2 Wuhan-Hu-1 (NCBI reference sequence: NC_045512.2) and derived strains, 31 that changes the RBD glycan shield's topology. More specically, in SARS-CoV and MERS, 32 as well as in the bat RaTG13 and pangolin CoV SARS-CoV-2 variants, 25,33,34 position N370 on the RBD is part of an occupied NST sequon, 32 which is lost in SARS-CoV-2 due to a T372A mutation. We performed ancestral sequence reconstruction of selected SARS S sequences to investigate the gain and loss of glycosylation sequons during S evolution. To address the effect of this change in the glycan shield's topology, we re-introduced the N-glycan at N370 in the SARS-CoV-2 S native sequence and ran MD simulations to study its effect on the stability of both, the open and closed RBDs. Our simulations indicate that within the SARS-CoV-2 S architecture, glycosylation at N370 does not interfere with the glycan network at N234, N165 and N343, but actively contributes to it by stabilizing very effectively the RBD open conformation.</p><p>Interestingly, analysis of the closed protomers shows that the N370 glycan from RBD (A) is rmly bound to the RBD (C) surface, where it occupies a specic cle. This interaction results in tying the closed RBD (C) to the adjacent closed RBD (A), very much like the laces in a shoe, thus potentially hindering the opening of the RBDs. Based on these ndings, we propose that the recent loss of N-glycosylation at N370 allows for a higher availability of open S conformations by lowering the energetic cost of the opening reaction, which is likely to be benecial by providing higher infectivity of SARS-CoV-2 relative to closely related variants carrying this sequon. We also discuss how the cle on the closed RBD surface in SARS-CoV-2 S, which is occupied by the N370 glycan in S glycoforms with the sequon, may be used by other glycans found on the cell surface, such as glycosaminoglycans, [35][36][37][38][39][40] sialogangliosides, and blood group antigens, 39 where these interactions may contribute to increasing the S cell-surface localization. 26 and drawn with DrawGlycan 27 (http://www.virtualglycome.org/ DrawGlycan/). Molecular rendering done with VMD. 28</p><!><p>In this section we will present the results of extensive conformational sampling based on multi-microsecond MD simulations of SARS-CoV-2 S models with different glycosylation at N234, N165 and N343, see Table 1. Specically, we determined the effects of site-specic glycan structure on the SARS-CoV-2 S ectodomain's structure and dynamics by systematically reducing the size of the oligomannose at N234; from Man9 (N234-Man9), found in highly stable prefusion SARS-CoV-2 S constructs, 10,12,30 and also studied in previous work, 17,41 we progressed to a shorter Man5 (N234-Man5), found on the virus and on other SARS-CoV-2 S constructs, 11,30 and to paucimannose (N234-Man3) as a hypothetical size limit. In these models the glycosylation at N165 and N343 is biantennary complex (FA2G2), and in one case we have also considered bisecting GlcNAc F/A2B forms at N343 and N165, respectively, see Fig. 1 and Table S1. † The results obtained for the N234-Man5 models were compared to a SARS-CoV-2 S model with a uniform immature glycosylation, 42 i.e. in which all N-glycans are Man5 (all-Man5). Finally, to assess the effect of loss of glycosylation at N370 on SARS-CoV-2 S, we added a complex N-glycan (FA2G2) at N370 in the N234-Man9 model with bisecting GlcNAc (F/A2B) Nglycans at N165 and N343, respectively. In all of these models the glycosylation at all sites other than the ones listed above is consistent and the same as the prole experimentally determined for the stable prefusion S trimer , 10 see Table S1. † Results are based on the analysis of multiple uncorrelated MD trajectories (replicas) run in parallel for each model (see details in the Computational methods section in ESI †).</p><!><p>The structure and dynamics of the RBD (residues 330 to 530) from a model with a Man5 at N234 and FA2G2 at N165 and N343 was analysed based on three replicas with sampling times indicated in Table 1. Note that the rst 300 ns of each replica were omitted from the analysis to allow conformational equilibration, see Fig. S1. † The dynamics of the RBD is rather complex and for consistency with previous work 17 Results shown in Fig. 2 and Table 1 indicate that in the N234-Man5 S model the open RBD can access a larger conformational space relative to the N243-Man9 model along the lateral displacement coordinate, while it is still able to adopt a stable open conformation, indicated by the axial tilt values. To note, the results from the MD simulations of the N243-Man9 ectodomain model presented here agree with the results obtained for the whole N243-Man9 S glycoprotein model. 17 glycan is relatively free on the opposite side of the pocket, or alternatively the RBD can be in an intermediate conformation between the wide-open and the closed. In the latter the Nglycans at all three positions N165, N234 and N343 are involved in complex interactions with protein residues in the RBD, namely with the loop from L460 and F489, and in and around the pocket, and with each other, see Fig. S3. † Within this framework, the higher lateral degree of exibility of the open RBD in the N234-Man5 S model is due to the smaller size of the Man5 glycan and its specic conformational propensity relative to Man9, which adopts a more 'tree-like' conformation, supported by extensive inter-arm interactions, 43 that lls the pocket much more effectively, see Fig. 2 (panels e and f). More specifically, the Man5 at N234 appears to be less competent than Man9 at forming interactions that bridge both sides of the pocket as it accesses the cavity le open by the RBD, while it mainly interacts with residues at the base of the open RBD, leaving it 'unhinged' when in the "wide-open" conformation, relative to the N234-Man9 model. As an interesting point to note, in all the trajectories the core fucose of the N343 FA2G2 glycan is exposed to the solvent, potentially allowing for its recognition. This is in agreement with cryo-EM studies reporting interactions involving the core fucose of the N-glycan at N343 and human neutralizing antibodies. 44 To further assess the role of the type of glycosylation around the N234-Man5 in the RBD dynamics, we ran three MD simulations of a model with uniform Man5 glycosylation (all Man5 model), see Table 1 and Fig. S4. † The results indicate that the replacement of the complex glycans with immature Man5 structures is problematic for the stability of the open RBD. More specically, the shorter size and 3D architecture of the Man5 at N343 does not allow it to engage effectively with residues in the RBD, while the Man5 at N234, as seen for the N234-Man5 model, is not able to effectively engage with the closed RBD (chain C) that ank the opposite side of the empty pocket. These results agree with recent work indicating that immature glycosylation, achieved in GnTIÀ/À mutant cells, leads to a less competent S glycoprotein relative to the fully glycosylated variant. 42 For comparison we tested a N234-Man9 model where only the N-glycans at N165 and N343 were Man5. The results shown in Table 1 and Fig. S5 † indicate that the presence of a large oligomannose structure at N243 helps recover the instability by more effectively occupying the empty pocket. In the equilibrium conformation sampled through 2.1 ms, the RBD conformation is more closed, shown also by the positive value of the axial angle, see Table 1, where both Man5 at N165 and N343 can interact with each other and with the RBD residues, see Fig. S5. †</p><!><p>To gauge the implications of the size of the oligomannose at N243 for the orientation and dynamics of the open RBD, we studied a model were N243 is modied with a paucimannose (Man3). Note, to our knowledge, the presence of Man3 at this or at any sequon in the SARS-CoV-2 S has not been detected to date; 8,10-13,30,45,46 nevertheless, as Man3 is one of the smallest oligomannose structures, this S glycoform represents a case study to account for the effect of the reduction in the size of the N-glycan to an extreme at this strategic position. The results shown in Table 1 and Fig. 3 in addition to the visual analysis of all trajectories, show that Man3 at N234 is less competent than Man 5 and Man9 in supporting a fully wide-open RBD. In only one of the three replicas, namely R3 shown in Fig. 3 (panels b and d), the Man3 accesses the interior of the open pocket, although its size does not allow for it to form any interactions that contribute to stability. In both R1 and R2, Man3 interacts with residues outside or at the edge of the pocket, as shown in Fig. S6. †</p><p>In the N234-Man3 model, the open RBD does not show any signicant lateral excursions, see Fig. 3 (panel a), as its position is rmly held in place by interactions with the complex Nglycans at N165 and N343, which contribute to pulling it towards a conformation intermediate near to a closed RBD, shown in Fig. S2 and S6. † In this conformation the N165 and N343 N-glycans interact extensively with residues in the receptor binding motif (RBM), possibly preluding to a nal closing, as reported in earlier work, showing the gating activity of the glycans at N343. 18 When the RBD is near closing there is a very limited degree of freedom in the lateral angle coordinate. As an interesting point, when the FA2G2 N-glycan at N343 interacts with the RBD, its core fucose becomes less accessible to a potential recognition, see Fig. 3 (panel b).</p><!><p>Analysis of the glycan shield topology through the ancestral sequence reconstruction of select SARS S sequences, shown in Fig. 4 (panel d), indicates the loss of a sequon at N370 in the SARS-CoV-2 Wuhan-Hu-1 strain, due to a mutation to NSA of the conserved NST present in closely related strains of current interest, namely SARS-CoV, the bat RaTG13G S and the pangolin CoV S. To study the effect on SARS-CoV-2 S structure and dynamics of this additional ancestral glycan at N370, we restored the sequon at N370 and glycosylated this position with a complex FA2G2 N-glycan, consistently with data reported in the literature for SARS-CoV. 47 In this model, position N234 is modied with Man9, while positions N343 and N165 are glycosylated with complex N-glycans with a bisecting GlcNAc and core fucosylation at N343 (FA2B), and without core-fucosylation at N165 (A2B). Our results, shown in Fig. 4 and Table 1, indicate that the wide-open RBD conformation in the N370-glycosylated SARS-CoV-2 S is as stable as the one observed for the corresponding N234-Man9, where the N-glycan at N370 lls the interior of the empty pocket together with the Man9. In this model, the Man9 at N234 is only able to access the entrance of the pocket due to steric hindrance with the N370 glycan that occupies the core. This suggest that once the RBD opens, an S glycoform with glycosylated N370 would be highly competent in exposing the RBD to the ACE2 receptor, as the SARS-CoV-2 S with a large oligomannose at N234.</p><p>The analysis of the dynamics of the closed RBDs of protomers A and C (PDB 6VYB numbering) shows that the N370 Nglycan of RBD (A) is tightly bound to the surface of the adjacent closed RBD (C), threading the two closed RBDs together, see Fig. 4 (panel b, e and f) and Fig. S7. † The 3D architecture of the complex FA2G2 N-glycan at N370, characterized by independent dynamics of the arms, 48 allows for stable interactions of the (1-6) arm within a cle the RBD, anked by residues between N448 and Y453 on one side and F490 and Y495 on the other, see Fig. 4 (panel e and f) and Fig. S7, † that support binding through hydrogen bonding and hydrophobic interactions with different monosaccharide units along the (1-6) arm. Because of the stability of the N370 glycan-protein interaction in the closed RBD, N370 glycosylation may hinder the opening mechanism. Thus, the loss of glycosylation at N370 likely contributes to enhancing the binding activity of S and infectivity of SARS-CoV-2 relative to other variants with an N-glycosylation sequon at this position.</p><!><p>The SARS-CoV-2 S glycosylation prole with expression in (or infection of) mammalian cells has been reported by several studies, 8,[10][11][12][13][14][15]45 almost all of which nd a large oligomannose Nglycan, such as Man9-7, as the most common structure at N234. This is especially true in highly stable prefusion SARS-CoV-2 S trimer glycoforms, 10,30 which bear the 2P mutation. 6,49 Meanwhile, a shorter Man5 appears to be present or even the dominant structure at N234 in the virus 8,30 and in the secreted ChAdOx1 nCoV-19 (AZD1222) vaccine epitope, 14,15 suggesting a higher degree of accessibility to that site by alphamannosidases in the ER. The results of multi-microsecond simulations presented in this work indicate that a reduced degree of lling of the pocket le empty by the opening of the RBD by a smaller N-glycan at N243 leads progressively higher degree of instability of the wide-open conformation of the RBD. More specically, while the N234-Man5 model appears competent in exposing the open RBD despite its higher dynamics relative to the N234-Man9 model, the N234-Man3 model leads to a dominant "more closed" conformation, see Fig. S2, † where the paucimannose cannot form stable interactions within the pocket, interacting only with residues at the pocket's gate or outside it, see Fig. 3 and S6. † The progressive destabilization upon reduction of the N-glycan size at N234 is in agreement with the results obtained for the N234A/N165A mutant, 17 designed to account for the complete removal of glycosylation at N234, which leads to 60% less binding to the ACE2 receptor as determined through biolayer interferometry assays. 17 It should be noted that the compact architecture of the S trimer does not allow for very large excursions in terms of lateral and especially of axial angles, nevertheless changes in these parameters in the same numerical range observed for the N234A/N165A mutant 17 have been shown to correspond to dramatic difference in ACE2 binding.</p><p>In agreement with recent work that discusses the roles of the glycan at N343 in supporting the RBD intermediate dynamics between open and closed conformations, 41 and of the glycan at N165 in supporting the open RBD, 17 we observed that the stability of the N234-Man5 is signicantly reduced when all the N-glycans in the shield are reduced to Man5 (All-Man5 model). In agreement with recent work, 41 this destabilization is due primarily to the lack of interactions that the shorter Man5 at N343 can make with the open RBD, and in particular with residues in the disordered loop within the RBM (400 to 508), in addition to the Man5 at N234 allowing the RBD to be relatively unhinged. We have shown that this destabilization is partially recovered when the N234 is occupied by a larger oligomannose, such as Man9. However, within this framework, the RBD adopts a more closed conformation, where the Man5 at both N165 and N343 can interact with the open RBD and with each other, see Fig. S5. †</p><p>The introduction of an additional glycosylation site at N370, which SARS-CoV-2 S has lost due to the T372A mutation in the Wuhan-Hu-1 and derived strains, illustrates the importance of an effective lling of the cavity le empty by the opening of the RBD. Indeed, the complex FA2G2 glycan at N370 can easily access the pocket in addition to Man9 at N234, which in this specic case can only partially ll it due to steric hindrance. We have shown that the presence of an N370 glycan contributes effectively to the stability of the wide-open RBD state, which nevertheless needs to be achieved starting from a closed S conformation. 41 To this end, we nd that the N370 N-glycan on the SARS-CoV-2 S throughout all our simulations occupies a specic cle on the surface of the closed RBDs. This binding mode is stabilized by a network of hydrogen bonds and hydrophobic interactions between the protein and the monosaccharides of the FA2G2 N370 glycan (1-6) arms, see Fig. 4 and S7. † Because the N370 glycan involved in this interaction originates from the adjacent closed RBD, its binding results in tying the closed RBDs together, likely hindering opening in the rst place. This is in agreement with recent work 50 suggesting that the introduction of a N370 sequon in SARS-CoV-2 S negatively affects binding to human ACE2, contributing to increased replication of SARS-CoV-2 S in human cells relative to its putative ancestral variant. Furthermore, it is also interesting to note that cryo-EM S structures from the bat RatG13 and pangolin CoV variants, both carrying the N370 sequon, have been only solved in their closed states, 25,33,34 possibly also suggesting opening is less favoured in these S glycoproteins relative to the SARS-CoV-2 S.</p><p>Ultimately, the tight binding we observed between the N370 N-glycan and the surface of the SARS-CoV-2 S closed RBD surface is not only interesting in terms of the implications for higher ACE2-binding activity through loss of the corresponding sequon suggests, but also indicates the presence of a glycan binding site in that cle, which is occupied in CoV variants that retain that sequon. In this context, recent work 35 provided evidence that heparan sulfate (HS) binds the SARS-CoV-2 S RBD in a ternary complex with ACE2, as an essential interaction for cell infection, meanwhile unfractionated heparin, non-anticoagulant heparin, heparin lyases, and lung HS are found to potently block SARS-CoV-2 S binding to ACE2 and infection. Notably, the same study provides evidence that SARS-CoV S binding to heparin-BSA is signicantly reduced, yet not completely negated. 35 Furthermore, recent work has also shown evidence that the RBD of SARS-CoV-2 S specically binds sialogangliosides, such as GM1 and GM2, with the same affinity observed for glycosaminoglycans (GAGs), as well as blood group antigens with a lower affinity. 39 Our results suggest that in SARS-CoV2 S a glycan-binding cle on the RBD surface is available and broadly accessible to be occupied by GAGs, sialogangliosides as well as blood group antigens, provided that they t the structural and electronic constraints that the site imposes. Meanwhile, in SARS-CoV S binding of these species may be disfavoured, because the cle is occupied by the N370 glycan from the adjacent protomer, further stabilizing the closed conformation. Furthermore, because of the high density of GAGs and sialogangliosides displayed on the surface of mammalian cells, we can speculate that this recently acquired topological change of the glycan shield may be advantageous for the virus towards cell surface localization and increased affinity to ACE2, where these glycans act as co-receptors. 35,39 Further investigation on these topics is underway.</p><p>Analysis of reconstructed SARS S ancestral sequences indicates that while the N370 sequon was recently lost in SARS-CoV-2 S, this sequon was only quite recently acquired within the phylogeny. However, proximal N-glycosylation sequons for example at position 364 (D364-YS in CoV2) have been gained and lost alternatively, see Fig. 4 (panel d). Based on our results, it is reasonable to think that glycosylation within this S topological region may have been evolutionarily conserved, because of its role in effectively stabilizing the open RBD, despite the higher energetic cost involved in the transition from the RBD down-to-up state. In SARS-CoV-2, glycosylation at N234, which according to our analysis, shown in Fig. 4 (panel d), also appeared recently, would then functionally take the role of a glycan at N370. This evolutionary exibility in the precise positions of glycosylation sites would make the presence of any specic glycan dispensable, with the consequent advantage of ensuring easier RBD opening reaction, and thus a more active S.</p><!><p>In this work we have used multi-microsecond MD simulations to determine the effect of changes in the nature and topology of the SARS-CoV-2 S N-glycosylation at sites known to be involved in its function. Our results indicate that reducing the size of the N-glycans at N234 led to the instability of the "wide-open" RBD conformation, with a consequent increase in RBD dynamics and a progressive stabilization of conformations favouring the closed protomer. Additionally, the structure of the N-glycans at N165 and N343 also affects the stability of the open RBD with shorter structures unable to effectively interact with the RBD disordered loop within the RBM. This effect is especially dramatic when a shorter N-glycan, such as Man5, is also present at N234. To account for changes in the glycan shield topology, we explored the effect of re-introducing N-glycosylation to a recently lost sequon at N370. Our results indicate that while the N-glycan at N370 is highly effective in stabilizing the open RBD in conjunction with the N-glycan at N234, it tightly binds a specic cle on the surface of the closed RBD, tying the closed protomers together and likely increasing the energetic cost of the RBD opening. Because the architecture of this RBD cle is particularly able to bind multiple monosaccharides through a network of hydrogen bonds and dispersion interactions, we suggest that in SARS-CoV-2 it can be occupied by other diverse glycan structures, such as glycosaminoglycans, proteoglycans or sialylated species, which have also been shown to bind S. This exibility in glycan-binding preference would provide an additional advantage in terms of increasing localization at the host cell surface. Finally, comparative analysis of reconstructed SARS ancestral sequences suggests that specic changes in the glycan shield topology at and around N370, in conjunction with the gain of N-glycosylation at N234, may have contributed to an increase in S activity, and thus of the infectivity of the SARS-CoV-2 relative to closely related coronaviruses.</p>

Royal Society of Chemistry (RSC)

Sub-3V, MHz-Class Electrolyte-Gated Transistors and Inverters

Electrolyte-gated transistors (EGTs) have emerging applications in physiological recording, neuromorphic computing, sensing, and flexible printed electronics. A challenge for these devices is their slow switching speed, which has several causes. Here we report the fabrication and characterization of n-type ZnO-based EGTs with signal propagation delays as short as 70 ns. Propagation delays are assessed in dynamically operating inverters and five stage ring oscillators as a function of channel dimensions and supply voltages up to 3V. Substantial decreases in switching time are realized by minimizing parasitic resistances and capacitances that are associated with the electrolyte in these devices. Stable switching at 1-10 MHz is achieved in individual inverter stages with 10-40 m channel lengths, and analysis suggests that further improvements are possible.

sub-3v,_mhz-class_electrolyte-gated_transistors_and_inverters

Introduction<!>Results and Discussion<!>Conclusions<!>Experimental Section

<p>The use of polarized electrolytes to modulate or "gate" the conductance of a semiconducting channel is an old idea that has its origins in the development of the first solid-state transistors at Bell Labs in the 1940s 1 . There are now many examples of semiconductor devices that can generally be called electrolyte-gated transistors (EGTs) 2,3 , though they go by different names. Over the last 70 years, EGTs have had several rebirths associated with the development of new EGT applications, or new materials that enhance performance, or discovery of new transport phenomena within EGTs. The ion-sensitive field effect transistor (ISFET) developed in the 1970s is an early example of an EGT 4,5 , where ion adsorption on gate oxides of Si FETs is exploited to make high sensitivity ion sensors. Microelectrochemical transistors, extensively developed by Wrighton and coworkers in the 1980s 6,7 , were another important milestone. These devices employ gate potentials to modulate the conductivity of conducting polymers in direct contact with electrolyte. Currently, EGTs are again enjoying a resurgence of interest associated with their applications in physiological recording [8][9][10] , neuromorphic computing [11][12][13][14][15][16] , and biosensing [17][18][19][20][21][22] , as well as in fundamental physics [23][24][25][26] , where the high charge densities achieved in electrolyte-gated semiconductors are leading to exciting discoveries.</p><p>A principal problem with EGTs for many applications is that their switching speed is limited 2 . This is easily understandable because the mechanisms of electrolyte gating rely on ion fluxes to form electrical double layers at the gate/electrolyte and semiconductor/electrolyte interfaces, orin the electrochemical modeto penetrate into the semiconductor. Ion mobilities are orders of magnitude smaller than electron or hole mobilities in semiconductors, which means the ionic resistance of the intervening electrolyte between the gate electrode and the semiconductor channel is relatively large. This ionic resistance can be viewed as a parasitic resistance Rp in series with the gate and it is dependent on intrinsic electrolyte conductivity, electrolyte thickness, and cross-sectional area [27][28][29] + Rp, where Rp dominates [30][31][32] ). We, and others, have measured ionic conductivity for gel electrolyte films based on ionic liquids and polymers and found it to be as large as 4 mS/cm depending on the precise gel composition [33][34][35] . Thus, for typical EGTs with channel areas of ~1000 m 2 , and gel film thicknesses of ~1 m, Rp is on the order of 1 k, which is a large value. Less resistive, ultrathin gate electrolyte films are a clear goal for EGTs 36 .</p><p>Khodagholy and coworkers 8 have recently minimized electrolyte polarization times by mixing gel electrolytes with a semiconducting polymer to make a composite mixed ionic and electronic channel material. In their EGTs, single current step rise and fall times of 1 s were achieved, which they argued were due to short migration path lengths for ions between electronically conducting polymer domains and internal, phase-separated ion reservoirs. However, Rp was not reported in that work and continuous MHz-level switching was not demonstrated.</p><p>Another challenge for EGT switching speed is parasitic capacitance, Cp [30][31][32] where it is also a problem 37 . Cp is minimized in conventional TFTs by reducing the projected areal overlap between the gate and the source and drain electrodes. For EGTs, the parasitic penalty on  is exacerbated by the enormous specific capacitance associated with metal/electrolyte interfaces, of order 10 F/cm 2 , which means that essentially any source/electrolyte or drain/electrolyte contact results in relatively large Cp.</p><p>In this work, we apply these ideas to fabricate fast EGTs and EGT circuits, and we establish that MHz-class devices can be realized with switching delays below 100 ns for the fastest devices.</p><p>In our view there are clear reasons why EGTs will not rival the speed of Si MOSFETs 2 (though perhaps not everyone agrees 38 ), but for envisioned applications of EGTs, gigahertz speeds do not appear necessary. As EGTs are being developed as physiological recording amplifiers 9,10 , and integrated into circuits 36,[39][40][41] and neural nets 4,11,13,14,16 , it is nevertheless important to establish the limits of performance that can be obtained by rational design. We note that there are other limitations to EGT performance like high dielectric loss tangents and quasi-static leak currents associated with electrolytes that are important for power consumption 2 , but these are not our focus here.</p><!><p>Our EGT device design is shown in Figure 1a. We employ a thin film of ZnO as the semiconductor channel because of its high electron mobility (~5 cm 2 /Vs), its stability in contact with electrolytes, and its ease of growth and patterning by atomic layer deposition (ALD) and photolithography 32,42 . The source and drain electrodes are Au/Ti and channel lengths L are systematically varied from 10 -40 m, scaling the channel width W such that the aspect ratio is fixed at W/L = 20, unless otherwise noted. The electrolyte is a soft solid called an ion gel that we and others have developed previously for EGTs [43][44][45] . Its attributes are high ionic conductivity (~4 mS/cm), a wide electrochemical stability window, chemical inertness, hydrophobicity, which allows it to serve as a kind of channel encapsulation, and its facile patterning by aerosol jetting.</p><p>The gate electrode is also aerosol jet printed and is make of the conducting polymer PEDOT:PSS, which has a large volumetric capacitance (~1 F/cm 3 ) by virtue of its permeability to ions. The large volumetric capacitance reduces the impedance of the PEDOT/ion gel interface and ensures that when a gate voltage VG is applied, that the majority of the voltage drops at the electrical double layer formed at the smaller, impermeable ion gel/ZnO interface. Importantly, the broad top surfaces of the Au source and drain electrodes are insulated from the ion gel by patterned SiOx overlayers in order to reduce the parasitic capacitance. Figure 1b shows an optical micrograph of a completed ZnO EGT; images of the devices at various stages of completion are shown in Supporting Information (see Figure S1).</p><p>Typical quasi-static drain current (ID) -gate voltage (VG) curves are displayed in Figure 1c for an EGT with W/L = 400 m/20 m. One sees minor hysteresis in the forward and reverse sweeps at a gate voltage sweep rate of 3 V/s. The threshold voltage VT = +1 V, the sub-threshold current onset occurs at VG = 0 V, and the ON/OFF ratio is 10 5 . The relatively large OFF current of 10 nA (see Figure S2 in Supporting Information) can be attributed to the large size of the device and the inherent leakage through electrolytes, which is often due to impurities such as H2O (the PEDOT gate is printed from an aqueous ink). The corresponding output curves (ID vs drain voltage VD) are shown in Figure 1d. The device exhibits reasonable saturation and square law current scaling up to VG = +0.8 V. Beyond VG = +0.8 V, ID increases, but not quadratically (see Supporting Information, Figure S3). This behavior reflects the mobility saturation that occurs at larger gate voltages.</p><p>To characterize parasitic resistances and capacitances that affect dynamic performance, we have undertaken impedance analysis of the ZnO EGTs as a function of voltage and frequency. To assess the dynamic response, we have fabricated EGT-based inverters in the internal feedback configuration in which the gate of the load EGT is connected to Vout, Figure 3a. A typical quasi-static inverter transfer characteristic is shown in Figure 3b and the corresponding gain (Vout/Vin) for VDD = +2 V is displayed in Figure 3c. One can see in Figure 3b that a rail-to-rail voltage swing is achieved over a very small voltage interval Vin = 50 mV. The gain is consequently 60 at VDD = +2 V, a high gain for an EGT inverter, and the gain increases with increasing VDD from +0.5 -+2 V as expected, Figure S4 (Supporting Information). Figure S5 demonstrates that the inverter noise margins are 0.6 V and ~80% of the theoretical maximum.</p><p>Further, dynamic inverter performance at 130 kHz drive frequency is displayed in Figure 3d, where it is evident that Vout follows Vin nicely with a full rail-to-rail voltage swing.</p><p>To further quantify dynamic performance, we have fabricated five-stage ring oscillators based on six EGT inverters in which the static load transistor is tuned with a separate supply voltage VBias, Figures 4a,b. Figure 4c shows the dynamic output at VDD = +3 V for a ring oscillator , and therefore 𝐶 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 , are proportional to channel area and thus will increase as L 2 . For conventional TFTs, 𝑅 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 does not vary with L when W/L is fixed, so the prediction would be that  = 𝑅 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 𝐶 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 ∝ L 2 .</p><p>However, for our EGTs the dominant role of Rp potentially introduces a different L dependence.</p><p>Considering the ionic resistance between the gate and channel, one expects Rp to scale inversely with device area, i.e., the larger the device footprint, the smaller the ionic resistance. With this line of thought, one predicts Rp ∝ L -2 and thus the product 𝑅 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 𝐶 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 ∝ L -2 . L 2 would be independent of L. That is, scaling up the device area increases 𝐶 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 , but decreases 𝑅 𝑡𝑜𝑡𝑎𝑙 𝑂𝑁 commensurately.</p><p>The data in Figure 4d are clearly at odds with this simple prediction.</p><p>Here</p><!><p>In summary, we have demonstrated here that by minimizing parasitic resistance and capacitance, and by selecting a high mobility thin film semiconductorin this case ZnOpropagation delays for EGT inverters can be decreased into the 100 ns regime for applied biases less than 3V. For our devices, parasitic resistance remains a significant factor and further decreases in switching time will require decreases in the electrolyte, gate electrode, and source/drain contact resistances. Standardization of EGT fabrication procedures will also enhance our ability to develop quantitative models for how the EGT static I-V characteristics and dynamic performance scale with device dimensions. Overall, continuous improvements in EGT performance should boost ongoing efforts to employ these devices in sensors, amplifiers, neural networks, and various types of flexible, wearable devices.</p><!><p>Materials: Ion gels were made from the ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide ([EMI][TFSI]) and the triblock copolymer poly(styrene-b-ethyl acrylate-bstyrene) (SEAS). The SEAS polymer was synthesized in-house using previously reported procedures. 43 [EMI][TFSI] ionic liquid was purchased from EMD Chemicals (Gibbstown, NJ, USA) and stored in an inert atmosphere. Aerosol jettable ion gel inks were made by preparing a solution with a mass ratio of 1:9:90 SEAS polymer:[EMI][TFSI] ionic liquid:ethyl acetate solvent.</p><p>The aqueous PEDOT:PSS ink, PH1000, was purchased from Heraeus (Germany), and 6 vol% ethylene glycol was added to the ink to enhance the PEDOT conductivity on drying.</p><p>Device Fabrication: A similar fabrication process is described in prior work 30,32 . Briefly, a 50 nm thick ZnO film was deposited using atomic layer deposition (ALD) (Savannah Series, Cambridge Nano Tech) with diethylzinc and water vapor as precursors. The ZnO film was annealed in a rapid thermal annealer (RTP-600S, Modular Process Technology). The first anneal at 300 °C in N2 lasted 15 min. The second anneal followed at 400 °C in O2 for another 15 min. The ZnO was then patterned with standard photolithographic procedures with aqueous HCl used as an etchant.</p><p>Standard lithography processes using Shipley S1813 resist were used to make the contact pads, interconnects, and source and drain electrodes. Interconnects and contact pads were deposited in an electron beam evaporator (Temescal) with 5 nm Cr as an adhesion layer followed by 25 nm thick Au. The source and drain electrodes, which were insulated with SiOx, were made in a separate series of lithographic steps. The metal layers (60 nm Au on 10 nm Ti) were deposited via e-beam evaporation (Varian 3118). Then another photolithography process was used to deposit 10 nm Ti on top of the Au, followed by 200 nm SiOx, again by e-beam evaporation. The gate dielectric (ion gel) and gate contact (PEDOT:PSS) were sequentially printed with an aerosol jet</p>

ChemRxiv

Structural basis of molecular recognition of helical histone H3 tail by PHD finger domains

The plant homeodomain (PHD) fingers are among the largest family of epigenetic domains, first characterized as readers of methylated H3K4. Readout of histone post-translational modifications by PHDs has been the subject of intense investigation; however, less is known about the recognition of secondary structure features within the histone tail itself. We solved the crystal structure of the PHD finger of the bromodomain adjacent to zinc finger 2A [BAZ2A, also known as TIP5 (TTF-I/interacting protein 5)] in complex with unmodified N-terminal histone H3 tail. The peptide is bound in a helical folded-back conformation after K4, induced by an acidic patch on the protein surface that prevents peptide binding in an extended conformation. Structural bioinformatics analyses identify a conserved Asp/Glu residue that we name ‘acidic wall’, found to be mutually exclusive with the conserved Trp for K4Me recognition. Neutralization or inversion of the charges at the acidic wall patch in BAZ2A, and homologous BAZ2B, weakened H3 binding. We identify simple mutations on H3 that strikingly enhance or reduce binding, as a result of their stabilization or destabilization of H3 helicity. Our work unravels the structural basis for binding of the helical H3 tail by PHD fingers and suggests that molecular recognition of secondary structure motifs within histone tails could represent an additional layer of regulation in epigenetic processes.

structural_basis_of_molecular_recognition_of_helical_histone_h3_tail_by_phd_finger_domains

Introduction<!>Structural basis of H3 recognition by BAZ2A PHD.<!>Crystallographic data processing and refinement statistics<!>Summary of thermodynamic-binding parameters for complex formation between different H3 peptides and WT and mutant BAZ2A/B PHD fingers<!><!>‘Acidic wall’ residue is conserved among PHD fingers that recognize helical H3 tail.<!>Prevalence of helical H3 N-terminal tail-recognizing human PHD fingers<!>CSPs induced by the H3 10-mer peptide on BAZ2A PHD.<!>Characterization of the interaction between H3 N-terminal tail and BAZ2A/B PHD in solution by NMR.<!>Role of the acidic wall residue of BAZ2A/B PHDs in H3 N-terminal tail recognition.<!>Summary of the thermodynamic-binding parameters for complex formation between H3 WT 10-mer peptide and mutant BAZ2A/B PHD<!><!>Changes in H3 N-terminal tail helicity correlate with different binding affinities for BAZ2A and BAZ2B PHD fingers<!>MD simulations of helicity of H3 N-terminal tail, WT and mutant.<!>Changes in H3 N-terminal tail helicity correlate with different binding affinities for BAZ2A and BAZ2B PHD fingers<!>Helical propensity of H3-derived peptides studied by CD.<!>Discussion<!>Protein expression and purification<!>Site-directed mutagenesis<!>Peptide synthesis<!>NMR spectroscopy<!>Isothermal titration calorimetry<!>X-ray crystallography<!>System set-up<!>Simulation protocol<!>Sequence alignment<!>Circular dichroism<!>Accession numbers

<p>The plant homeodomain (PHD) finger is one of the largest families of epigenetic reader domains present in chromatin-related proteins, with over 170 PHD fingers identified in the human genome [1]. Early pioneering studies led to PHD fingers being classified as domains that specifically recognize histone H3 trimethylated at K4 [2–5]. However, the diversity of PHD fingers in terms of their ability to recognize a wide array of post-translational modifications (PTMs) and unmodified tails has now become apparent [6–9]. Several PHDs have been characterized that recognize different PTMs on the H3 tail, including di- and tri-methylation of K4 [2], trimethylation of K9 [10], acetylation of K14 [11] and trimethylation of K36 as well as PTMs on H4 such as acetylation [6]. An additional layer of complexity in the molecular recognition by PHD fingers is imparted by the recurrent presence of adjacent domains that aid combinatorial, multivalent readout of histone tails, intra- or inter-nucleosomal [8,12]. Indeed, PHD fingers are often found in close proximity with a bromodomain (BRD) [13,14], as well as other PHD fingers [11,15], bromo-adjacent homology domains [16], tudor domains [17] and chromodomains [18]. Structural studies have elucidated diverse modes of combinatorial readout by PHD fingers and their tandem domains for individual and multiple histone tails, which typically involve recognition of the peptide in a fully extended conformation [9,19]. While PTM-specific and combinatorial readout modalities of histone tails are well understood [20], much less is known about the recognition of secondary structure features within the histone tail itself.</p><p>Members of the BAZ family of proteins, which includes BAZ1A, also known as Acf1 [21,22], BAZ1B, also known as Wstf [23,24], BAZ2A, also known as TIP5 (TTF-I/interacting protein 5) [25,26], and BAZ2B [26], are all characterized by the presence of a PHD–BRD tandem module at their C-terminus. BAZ2A is the best characterized member of the BAZ family from a functional standpoint. BAZ2A binds to the ATPase SNF2h (sucrose nonfermenting protein 2 homolog) to form the chromatin remodeling complex NoRC (nucleolar remodeling complex), which plays an essential role in silencing ribosomal DNA (rDNA) genes [27]. Experiments performed with truncated versions of BAZ2A showed that the PHD–BRD module plays an important role in NoRC formation, with the PHD domain being required for interaction with the nucleosome to trigger transcriptional silencing of rDNA [28]. In a recent study, BAZ2A was found to be overexpressed in prostate cancer and a role was proposed for the protein in establishing epigenetic alterations that favor an aggressive phenotype of the cancer [29]. The related protein BAZ2B [30] is yet poorly characterized and its biological role remains unclear. We recently biochemically and structurally characterized the PHD fingers and BRDs of both BAZ2A and BAZ2B, and identified the N-terminal tail of histone H3 as the preferred binding partner of the PHD domains [26]. Structural studies with PHD–BRD tandem constructs have pointed to rather elongated and rigid structures with the two domains probably recognizing distinct regions of H3 histone tails independently [26]. NMR spectroscopy has been combined with computational studies to throw light on the molecular recognition features of histone H3K14ac recognition by the BAZ2B BRD [31]. However, the complete molecular picture of H3 tail recognition by the PHD fingers of BAZ2A and BAZ2B had remained elusive.</p><!><p>(A–C) Crystal structure of BAZ2A PHD (shown in gray) in complex with H3 10-mer (shown in green). (A) Surface and ribbon representation of BAZ2A PHD (regions of positive and negative electrostatic potential are shown in blue and red, respectively) in complex with H3 10-mer shown in a ribbon and stick representation. Residues of the H3 10-mer peptide are labeled. (B) The 2Fo–Fc map contoured at 1σ (shown in gray) for H3 10-mer. (C) Close-up view of the interaction between the BAZ2A PHD and the H3 10-mer peptide. Residues of BAZ2A PHD interacting with the H3 10-mer peptide are shown in a stick representation and labeled in black. Residues of the H3 10-mer peptide are labeled in red. (D) ITC-binding curves of different H3-derived peptides titrated into BAZ2A PHD.</p><!><p>Values in parentheses are for the highest resolution shell.</p><!><p>Error values reported on dissociation constant (KD), stoichiometry of binding (N) and binding enthalpy (ΔH) are generated by the Origin program and reflect the quality of the fit between the nonlinear least-squares curve and the experimental data. Errors reported on TΔS and ΔG were propagated from the errors of KD and ΔH. Raw ITC data are shown for each titration in Supplementary Data (Supplementary Figure S1).</p><!><p>Abbreviations: N.D.: not determined.</p><!><p>(A) Sequence alignment of PHD fingers whose structure was solved in complex with an H3 N-terminal tail peptide. The column corresponding to E1689 of BAZ2A is highlighted through the alignment with a red box, and Asp or Glu residues in this column are colored in red. The column corresponding to the absolutely conserved tryptophan of PHD fingers that recognize methylated-K4 is highlighted through the alignment with a black box, and tryptophan residues in this column are colored in magenta. PHD fingers that induce the H3 tail to adopt a helical (cyan box), bent (green box) or extended (magenta box) fold are grouped (see Supplementary Figure S4 for full alignment). (B) Structures of PHD fingers (gray cartoon) that have an acidic residue (shown in red) in the position corresponding to E1689 of BAZ2A and that induce the H3 N-terminal peptide (green cartoon) to adopt a helical folded-back conformation. (C) Sequence alignment of PHD fingers of KDM5 proteins. The columns corresponding to E1689 of BAZ2A and to the absolutely conserved tryptophan of PHD fingers that recognize methylated-K4 are highlighted through the alignment as described in A (see Supplementary Figure S5 for full alignment).</p><!><p>It is remarkable that PHD fingers recognizing H3 in a bent or extended conformation do not have an acidic residue in the position corresponding to E1689 of BAZ2A PHD (Figure 2A and Supplementary Figure S4). To investigate the prevalence of the acidic wall residues in all human PHD fingers, we extended our bioinformatics analysis to the entire human genome [1]. We found that 36 of the 172 sequences annotated as PHD fingers have an acidic residue in the position that corresponds to E1689 of BAZ2A (Supplementary Figure S5). Among these are all the PHD fingers of the BAZ family, CREBBP [33] and the homologous EP300 [34], all members of the DPF family of proteins: DPF1, DPF2 and DPF3, as well as members of the KDM5/JARID1 histone lysine demethylase family: KDM5A, KDM5B, KDM5C and KDM5D [35,36]. Interestingly, we noted that, in all four KDM5 members, only the first PHD domain, which like BAZ2A/B recognizes unmodified K4, but not the second or third, has an acidic residue at this position, and this is mutually exclusive with the presence of the conserved tryptophan residue characteristic of the aromatic cage for methyl-K4 recognition (Figure 2C) [37]. Indeed, only 5 of the 36 sequences containing the acidic wall residue also contain this tryptophan (Supplementary Figure S5). Based on this observation, we postulate that there could be a level of incompatibility between methyl-lysine readout and helical H3 recognition by PHD finger domains. Five PHD fingers bear both an acidic wall residue and the tryptophan needed for methyl-K4 recognition: ASH2L [38], the MLL2 and MLL3 members of the KMT2 family of lysine methyltransferases [39], PHF20 [40] and UBR7 (Supplementary Figure S5). Structural information is available only for ASH2L PHD, which unveils an atypical PHD fold with only one zinc ion coordinated, suggesting that the ASH2L PHD structure is incompatible with histone binding [38]. The remaining four PHD fingers are poorly characterized, their substrate specificity is not known and it is difficult to conclude if they represent genuine exceptions to the observed mutual exclusivity between acidic wall residue and conserved Trp residue.</p><!><p>Overlay of [15N-1H]-HSQC spectra recorded on 15N-BAZ2A PHD with increasing concentrations of the H3 10-mer peptide. Spectra were recorded at the following protein:peptide molar ratios: 1:0 (blue), 1:2 (cyan), 1:4 (yellow) and 1:8 (red). For a set of peaks, the direction of the shifts is indicated with black arrows. The horizontal dotted lines indicate peak pairs corresponding to the side-chain of Asn and Gln.</p><!><p>Chemical shift differences induced by H3-derived peptides on BAZ2A/B PHDs were weighted as described in the Experimental section and plotted against BAZ2A/B PHD sequences. The resulting histograms were used to group residues based on the extent of their CSPs: weak (weighted chemical shift difference value equal or above the average chemical shift), medium (equal or above the average chemical shift plus the standard deviation) and strong (equal or above the average chemical shift plus two times the standard deviation). The CSPs observed were mapped on BAZ2A/B PHDs structures (PDB: 5T8R and 4QF3, respectively) by coloring residues with weak shifts in yellow, medium in orange and strong in red. Residues with a weighted chemical shift difference value lower than the average chemical shift are in white. The H3 10-mer peptide is shown as sticks and colored in green and its residues are labeled in red. In the middle panel, the peptide is omitted for clarity.</p><!><p>(A and B) ITC-binding curves of the H3 10-mer peptide titrated into WT and mutant BAZ2A PHD (A) and BAZ2B PHD (B).</p><!><p>Titrations were performed at 25°C in triplicate, except where indicated, and values reported are the means ± s.e.m. Raw ITC data are shown for representative titrations in Supplementary Data (Supplementary Figure S2).</p><!><p>Titrations performed in duplicate.</p><p>N was fixed to 1 during the data fitting.</p><!><p>To gain a better understanding of the histone molecular recognition, we investigated the energetic contribution of different H3 residues in binding to the PHD fingers of BAZ2A and BAZ2B. We performed an alanine scan where residues 2–6 of the H3 10-mer were mutated individually to alanine and the resulting mutant peptides tested for binding with BAZ2A PHD and BAZ2B PHD by ITC (Supplementary Figure S1 and Table 2). The R2A and T3A mutations abolished binding. The K4A mutation did not affect the binding affinity with BAZ2A PHD and even increased the affinity toward BAZ2B PHD. The Q5A mutation improved binding, and the simultaneous introduction of K4A and Q5A mutations remarkably increased binding affinities by 4-fold (BAZ2A) and 15-fold (BAZ2B) (Table 2). Finally, T6A did not affect the binding affinity of H3 10-mer toward either protein. Our data show that K4-T6 residues are not critical for binding to the PHD fingers of BAZ2A and BAZ2B, while R2-T3 are crucial. These results are consistent with those recently reported by Chakravarty et al. [37] for BAZ2A PHD and the first PHD domain of KDM5B but are distinct from the results of the first PHD of AIRE, which is known to bind H3 in an extended conformation. In that case, the T3A mutation was tolerated, whereas the K4A mutation abolished binding [37].</p><p>The increase in binding affinity observed for the mutant H3 peptides was unexpected, especially the ones harboring the K4A mutation as both BAZ2A and BAZ2B PHD fingers recognize unmodified K4 [26]. The strong contacts formed by the K4 side chain in the deep surface groove of the PHD surface (Figure 1A–C) would not be recapitulated upon K4A mutation, and hence, a loss of binding affinity was anticipated. To investigate the structural basis for the unusual increase in binding affinity of the H3 10-mer AA mutant peptide (ARTAATARKS), we mapped its binding site by NMR using the so-called minimal shift approach (Supplementary Figure S8). Overall, we observed equivalent CSP maps for the H3 10-mer AA mutant compared with WT peptide (Supplementary Figure S8), the major difference being present at the N-terminus of BAZ2A PHD where the side chain of H3K4 is accommodated. Consistently with the H3 K4A mutation, the shifts induced by the H3 10-mer WT peptide at the BAZ2A PHD N-terminus are reduced for the AA mutant peptide (Supplementary Figure S8). Importantly, we did not observe any extra cluster of shifts for H3 10-mer AA mutant peptide that would suggest different binding site(s) exploited by this mutant peptide (Supplementary Figure S8). In light of our crystal structure and of the helical fold of bound H3 peptide, we reasoned that the K4A and Q5A mutations could stabilize the peptide helicity accounting for the increased affinity. Indeed, alanine has the highest helix propensity among natural amino acids [41].</p><!><p>(A) Helical character of each peptide residue in the WT, double-Ala and double-Gly mutants during the last 60 ns of MD simulations, represented as a percentage of time with the secondary structure of α-, 310- or π-helix, and shown as median ± interquartile range, in complex with BAZ2A PHD (left) and in aqueous solution (right). (B and C) A superposed cartoon representation of the last frame of 4 MD replicas of H3 10-mer (B) and H3 10-mer K4A/Q5A (C) in complex with BAZ2A PHD (shown as a surface in gray). (D and E) Superposed cartoon representation of the last frame of four MD replicas of H3 10-mer (D) and H3 10-mer K4A/Q5A (E) in aqueous solution.</p><!><p>We hypothesized that the observed increase in helical stability upon alanine mutation could also be reflected in their unbound state. To analyze this effect, we modeled both peptides in aqueous solution (Figure 6A, right panel, and D,E). In the absence of the PHD protein, there is still some helicity, albeit weak, persisting in the WT peptide (∼5% of the time). The helical character of the peptide in the unbound state was consistently increased by the introduction of K4A and Q5A mutations to 12% of the time (Figure 6A, right panel). To further investigate the relationship between the helical propensity of H3 10-mer and binding affinity toward BAZ2A/B PHDs, we aimed to reduce the peptide helicity by replacing K4 and Q5 with a Gly residue. Indeed, excluding proline, glycine has the lowest helix propensity among natural amino acids [41]. The resulting H3 10-mer GG mutant peptide (ARTGGTARKS) showed markedly reduced binding affinity toward both BAZ2A/B PHD fingers (Supplementary Figure S1 and Table 2). ITC data revealed that the decreases in binding affinity of the H3 10-mer GG mutant are contributed entirely by large entropic penalties (Table 2), consistent with a significant reduction in conformational freedom of the peptide upon binding relative to H3 10-mer WT or H3 10-mer AA mutant peptides. MD simulations further showed a significant weakening of the intramolecular hydrogen-bond network along with a modest decrease in the helical character in the H3 10-mer GG peptide compared with H3 10-mer WT and AA mutant (Figure 6A and Supplementary Figures S9 and S10). Taken together, our data reveals that H3 tail helicity plays an important role in recognition by the PHD domains of BAZ2A/B.</p><!><p>The CD spectra of the H3 10-mer WT (ARTKQTARKS), H3 10-mer AA (ARTAATARKS) and H3 10-mer GG (ARTGGTARKS) peptides at different TFE concentrations (Supplementary Figure S11) were deconvoluted (Supplementary Tables S1–S3) and the content of the regular α-helix found in the best matching solution was plotted against the TFE concentration (v/v).</p><!><p>Epigenetic regulatory processes modulate human physiology and disease; thus, reaching a comprehensive understanding of their molecular basis is important. Molecular recognition of secondary structural features within histone tails by epigenetic reader domains has received little attention to date. Herein, we have examined the structural and biophysical basis for the recognition of the helical histone H3 tail by PHD fingers, using the PHDs of BAZ2A and BAZ2B as the model system.</p><p>Our structural insights into the molecular recognition of histone H3 by the BAZ2A and BAZ2B PHD fingers add to the emerging evidence for specific recognition of the helical H3 tail by this reader family, provided by peptide-bound structures recently solved for the PHD fingers of UHRF1, MOZ and DPF3 (Figure 2B). Identification of a strict conservation for an Asp/Glu residue at the acidic wall position on these family members suggests a simple consensus signature for this subclass of PHD domains. Sequence alignment of the whole human PHD finger-ome identified a single putative exception to this rule, namely the first PHD finger of KDM5B (KDM5B-PHD1) that bears an Asp as the acidic wall residue, in spite of being found to bind H3 in an extended conformation based on an NMR structure of the complex (PDB: 2MNZ [47], Supplementary Figures S4A and S12). However, in that structure, the H3 peptide fits into a groove close to the N-terminus of KDM5B-PHD1 rather than running parallel to the domain as observed for the PHD fingers that bind H3 peptide in an extended conformation (Supplementary Figure S12). Such arrangement resembles the conformation observed for H3 bound to UHRF1 PHD in the co-crystal structure reported by Wang et al. [48] (Supplementary Figure S12). However, there are six other independent structures of UHRF1 PHD in complex with H3 N-terminal peptide bound in a helical fold (Supplementary Figure S12) [17,49–53]. We therefore propose that, as observed for UHRF1, the KDM5B-PHD1 can also recognize the H3 N-terminal tail in a helical fold.</p><p>We identify a subclass of 36 human PHD fingers containing an acidic wall residue Asp/Glu as potential consensus to molecular recognition of the helical H3 tail. The minimal overlap observed between the acidic wall subclass and the subclass comprising the key conserved Trp residue corresponding to specific readout of methylated-K4 suggests a level of incompatibility between these two molecular recognition features. Indeed, methylation of H3K4 was found to weaken or completely abrogate histone binding in several PHD fingers that recognize helical H3, such as BAZ2A/B [26], DPF3b [11] and MOZ [54]. This trend of incompatibility is particularly evident in the KDM5 subfamily, where, in all its members, only the first PHD finger (PHD1) has an acidic wall and this is mutually exclusive with the conserved Trp for methylated-K4 recognition that is instead present in PHD2 and PHD3 (Figure 2C). The interaction between unmodified H3 N-terminal peptide and both KDM5A-PHD1 and KDM5B-PHD1 has been recently characterized using NMR [35,36,47]. Interestingly, the patterns of CSPs observed were in both cases consistent with the one observed for BAZ2A/B PHDs (Figure 4).</p><p>The region corresponding to the acidic wall residue is often found as highly acidic, with additional Asp/Glu residues found either immediately before or after the acidic wall residue (Figure 2A–C and Supplementary Figure S5). Sequence analysis showed that 22 of these 36 PHD sequences contain at least two adjacent acidic residues (Supplementary Figure S5), suggesting the prevalence of a double acidic patch. We provide evidence that full neutralization of this double acidic patch abrogates H3 binding in BAZ2A PHD, highlighting its important role (Figure 5 and Supplementary Figure S2). We propose that the negatively charged patch at the acidic wall helps to stabilize the helical fold of H3 by forming electrostatic interactions with the positive dipole of the histone helix. The carboxylate side chain(s) at the acidic wall can help to induce the helical bound conformation in H3 by interacting with the basic side chains of K4 and Q5 at the start of the helix. In addition, it can form a salt bridge with the guanidinium group of R8, as shown by recently solved crystal structures of the DPF of DPF3b (PDB: 5I3L [55]), MOZ (PDB: 5B75 [56]) and MORF (PDB: 5U2J [57]), and thus probably occurring in BAZ2A PHD. Interestingly, mutation A275D in the DPF of MOZ just upstream of acidic wall residue D276, thus installing a double-negative charge as in BAZ2A, enhanced the binding of H3K14ac peptide by 3- to 4-fold [54].</p><p>It has been shown that PTMs can affect the secondary structure of histones [58]. It is tempting to speculate that induction or stabilization of the helicity of histone tails could represent an additional layer of regulation in epigenetic processes beyond or in cross-talk with PTMs. Within the context of tandem epigenetic reader domains, the H3 helical fold has been shown to be important for simultaneous recognition of distinct regions of the H3 tail by two epigenetic reader domains on the same protein. For example, the H3 helical conformation induced by UHRF1 PHD binding was found to be essential for productive recognition of K9 modification states by the neighboring tudor domain [49]. Similarly, the H3 helical fold was found to be critical for simultaneous recognition of K4 and K14 modifications by the double PHD finger domain of MOZ [15,54,56]. Our work suggests that, in BAZ2A/B and related proteins, the helical fold of the bound H3 N-terminal tail could facilitate productive simultaneous recognition of both unmodified K4 and downstream marks, e.g. K14ac, by the neighboring PHD finger and BRD, respectively, which warrant future investigation.</p><p>In conclusion, we propose that among the large PHD family exists a class of PHD fingers with a distinct recognition mode of the histone H3 tail that induces H3 to adopt a helical fold after K4. PHD fingers that belong to this class are characterized by the presence of a conserved Asp/Glu residue within a short acidic patch made of a helical turn or loop just before the first β-strand. We show that H3 helicity is critical for molecular recognition by this subclass of PHD fingers and identify mutations at K4 and Q5 in H3 that either enhance or weaken the binding affinity by stabilizing or disrupting the peptide helicity, respectively. This mutagenesis approach may provide a rapid and direct strategy to identify other reader domains that also recognize helical H3 tails. Our work has also implications for drug design. The growing interest in targeting epigenetic reader domains with small molecules has led to many examples of successful campaigns delivering potent chemical probes in particular for BRDs [59,60], but also for methyl-lysine reader domains such as malignant brain tumor domains [61–63] and chromodomains [63–66]. In contrast, relatively little progress has been made on targeting PHD fingers, with only two examples reporting weak-binding fragments [67] and screening-active compounds [68], suggesting low ligandability for this class of reader domains. Drug design approaches to stabilize the helical conformation, e.g. by using stapled peptides, or to mimic the helix recognition pharmacophore could provide attractive new strategies to aid the development of epigenetic chemical probes that disrupt this class of reader–histone interactions.</p><!><p>Expression and purification of BAZ2A PHD and BAZ2B PHD were performed as described recently [26]. 15N and 15N/13C uniformly labeled BAZ2A PHD and BAZ2B PHD were expressed in modified M9 minimal medium [69] where the sole sources of nitrogen and carbon were 1 g/l 15NH4Cl (Goss Scientific) and 2 g/l 13C-d-glucose (Goss Scientific) as appropriate. The expression conditions and the purification procedures used for labeled proteins were the same as for unlabeled samples [26].</p><!><p>Mutations were introduced into BAZ2A/B PHD fingers by polymerase chain reaction (PCR) amplification of the original construct using Phusion DNA Polymerase (Thermo Fisher Scientific) and a specific pair of primers for each mutation (Supplementary Table S4). The PCR amplification product was incubated with Dpn I (New England BioLabs) for 1 h at 37°C to digest the parental DNA strands and then used to transform Escherichia coli DH5α cells. Transformed cells were grown on lysogeny broth (LB) agar plates supplemented with 100 µg/ml ampicillin for 16 h at 37°C. Single colonies were picked to inoculate 5 ml of LB plus 100 µg/ml ampicillin and grown for 16 h at 37°C. The DNA was extracted from the bacterial cultures using the QIAprep Spin Miniprep Kit (Qiagen) and the presence of the desired mutation was checked by DNA sequencing. Electrospray ionization mass spectrometry analyses confirmed that the mutant constructs were successfully translated into the correctly mutated proteins.</p><!><p>Peptides were synthetized by standard automated solid-phase synthesis on a ResPep SL peptide synthesizer (Intavis) using Fmoc-protected amino acids and Rink Amide resin (Novabiochem). Amino acids were coupled twice adding 1.05 equivalents of Fmoc-protected amino acid, 1 equivalent of N,N,N′,N′-tetramethyl-O-(1H-benzotriazol-1-yl)uranium hexafluorophosphate and 1 equivalent of N-methylmorpholine with 5-fold excess over the resin. Peptides were cleaved from the resin and deprotected by incubation of the resin for 3 h with 1 ml of cleavage mixture containing 97.5% trifluoroacetic acid (TFA) and 2.5% water, leaving all peptides amidated at the C-terminus. Peptides were precipitated by the addition of 5 ml of ice-cold diethyl ether and pelleted by centrifugation. The resulting pellets were washed twice with diethyl ether. High-performance liquid chromatography (HPLC) purification was performed on a Gilson Preparative HPLC System using a Zorbax 300SB-C18 column (5 µm particle size, 250 × 9.4 mm) run at 4 ml/min. The solvents used were A (99.9% water and 0.1% formic acid or TFA) and B (94.9% acetonitrile, 5% water and 0.1% formic acid or TFA). A linear gradient from 0% to 10% B was used. All the peptides were retained during the run but were eluted before the gradient, i.e. in 100% A, except for ARTAATARKS and ARTKQTARKS, which were eluted at the beginning of the gradient. Removal of formic acid and TFA was performed using the VAriPure IPE column, and the absence of TFA was confirmed by 19F NMR. Purified peptides were submitted to liquid chromatography–mass spectrometry (LC–MS) analysis (Supplementary Figure S13). LC–MS analyses were performed with an Agilent Technologies 1200 series HPLC connected to an Agilent Technologies 6130 quadrupole spectrometer and a diode array detector. Chromatography runs were conducted with a Waters XBridge C18 column, 50 mm × 2.1 mm, with a 3.5 µm particle size with a mobile phase of water/acetonitrile +0.1% formic acid using a gradient from 95:5 to 10:90 over 7.5 min.</p><!><p>NMR spectra were acquired from 15N or 15N/13C-labeled samples of BAZ2A PHD and BAZ2B PHD at a concentration of 350 µM in a buffer containing 50 mM KCl, 1 mM dithiothreitol (DTT), 0.02% (w/v) NaN3, 10% D2O and 25 mM K2HPO4 at a pH of 6.9 for BAZ2A PHD and a pH of 6.5 for BAZ2B PHD. All NMR experiments were performed at 25°C using an AV-500 MHz Bruker spectrometer equipped with a 5 mm CTPXI 1H-13C/15N/D Z-GRD cryoprobe. Sequence-specific backbone assignments were obtained for BAZ2A PHD and BAZ2B PHD from the identification of intra- and inter-residue resonances in the following spectra: [15N-1H]-HSQC, 15N/13C/1H HNCO, HNCA, HNCACB and HN(CO)CACB. Acquisition times used in the [15N-1H]-HSQC experiments were 120 ms for 1H and 60 ms for 15N. Typical acquisition times in the three-dimensional experiments were: 100 ms for 1H, 14–19 ms for 15N and 7–12 ms for 13C. All the NMR spectra were processed using the program TopSpin (Bruker) and analyzed using the package CcpNmr Analysis [70].</p><p>In CSP experiments, the chemical shift differences in proton (ΔδH) and nitrogen (ΔδN) were combined to obtain a weighted chemical shift difference (Δδweighted) using the following equation: Δδweighted = |ΔδH| + |ΔδN| * 0.14, where 0.14 is a scaling factor required to account for the difference in the range of amide proton and amide nitrogen chemical shifts [71]. Shifted residues were clustered based on the extent to which they showed a CSP into strong (Δδweighted value above the average chemical shift plus two times the standard deviation), medium (Δδweighted value above the average chemical shift plus the standard deviation) and weak (Δδweighted value above the average chemical shift). CSPs in the slow exchange regime on the NMR timescale were analyzed using the 'minimal shift approach' [72]. The chemical shift change for each backbone amide group was measured from the peak detected in the HSQC spectrum recorded on the free form to the nearest peak detected in the HSQC spectrum recorded on the bound form. ΔδH and ΔδN were combined as described before to obtain a minimal Δδweighted.</p><!><p>All calorimetric experiments were performed on a MicroCal iTC200 microcalorimeter (GE Healthcare) at 25°C in a buffer containing 20 mM HEPES at a pH of 8, 150 mM NaCl and 0.5 mM Tris(2-carboxyethyl) phosphine. All ITC experiments were carried out titrating peptide solutions (1.5–3 mM) into protein solutions (80–120 µM) loaded in the calorimeter cell, performing one first injection of 0.4 µl followed by 19 injections of 2 µl. The data were analyzed using the MicroCal™ software package subtracting the data from an independent titration of peptide into buffer to account for heat of dilution, and then fitted using a single-binding site model. Protein concentration was determined by measuring absorbance at 280 nm using the following extinction coefficients: BAZ2A PHD ε280 = 6990 M−1 cm−1 and BAZ2B PHD ε280 = 8480 M−1 cm−1. Lyophilized peptides were weighted and dissolved in an appropriate volume of buffer to obtain the desired concentration.</p><!><p>Crystals of BAZ2A PHD in the apo form were grown at 18°C using the sitting drop vapor diffusion method by mixing equal volumes of protein [6 mg/ml in 20 mM Tris–HCl (pH 8), 150 mM NaCl and 2 mM DTT] and crystallization buffer (2.2 M Na/K phosphate buffer at a pH of 8.5). To obtain crystals of the BAZ2A PHD-H3 10-mer complex, preformed apo BAZ2A PHD crystals were transferred and soaked overnight into a solution containing 2 mM H3 10-mer (ARTKQTARKS) in crystallization buffer and cryoprotected in 1.6 mM H3 10-mer, 1.7 M Na/K phosphate and 20% glycerol. The data sets were collected at the beamline ID29 at European Synchrotron Radiation Facility and processed with XDS [73,74] and AIMLESS [75], to 2.4 Å of resolution. The structure of the complex was determined by isomorphous replacement with the apo form of BAZ2A PHD (PDB entry: 4QF2 [26]). Manual model building and refinement were carried out using Coot [76] and Refmac5 [77]. The quality of the models was checked by MolProbity [78], and all structure figures were generated using PyMOL (The PyMOL Molecular Graphics System, Version 1.7.05, Schrödinger, LLC).</p><!><p>The X-ray crystal structure of BAZ2A PHD in complex with H3 10-mer (ARTKQTARKS) was used as the starting structure of the corresponding simulation. The missing residues K9 and S10 were added choosing a suitable low-energy rotamer from PyMOL and minimized for 4000 steps with the rest of the protein fixed. The initial structures of the H3 10-mer with K4A/Q5A and K4G/Q4G mutations to generate ARTAATARKS and ARTGGTARKS, respectively, were built from the WT structure, with the point mutations performed in PyMOL. The structures of the H3 peptides obtained this way were used to simulate the peptides in the unbound state, i.e. in aqueous solution, as well. All models were solvated in a TIP3P water box with a padding of 15 Å from the edge of the box to any protein atom. The system charges were neutralized with sodium or chloride ions as appropriate.</p><!><p>MD simulations were carried out using the NAMD program [79] and the CHARMM 36 force field [80]. Initially, the solvated systems were minimized for 3000 steps with the protein restrained to eliminate residue unfavorable interactions between the protein and the solvent, followed by another 5000 steps with all atoms free to move. Heating of the systems from 0 to 300 K was achieved in 100 ps (time step of 1 fs), with fixed protein backbone atoms to allow relaxation of the solvent. The systems were subsequently equilibrated for 600 ps (time step of 2 fs) with all atoms free to move. The NPT ensemble was used during the production simulations, which involved four replicates of 80 ns each (time step of 2 fs). The temperature was controlled with a Langevin thermostat at 300 K, and the pressure with a Nose–Hoover Langevin piston barostat at 1 bar. A SHAKE constraint was applied to all bonds containing hydrogen atoms. Short-range non-bonded interactions were switched at 10 Å and cut off at 12 Å, and particle mesh Ewald summation was employed for long-range non-bonded interactions. Consistency and stability throughout the MD replicas were assessed (Supplementary Table S5). The per-residue secondary structure calculation was performed using the Timeline plugin v.2.3 and the hydrogen-bond contacts with the HBonds plugin v.1.2, both contained in VMD v. 1.9.2 [81]. Pair-wise distribution differences among simulated systems were assessed statistically using the two-tailed Mann–Whitney U-test, as implemented in the statistical package R [82].</p><!><p>Sequences of domains that belong to the PHD family and whose structures were solved in complex with an H3 N-terminal tail peptide were identified with the software Dali [83] using as input the structure of BAZ2A PHD (4QF2). The sequences of human PHD fingers were obtained from the Structural Genomic Consortium database [1]. The multiple sequence alignment was performed using MAFFT (Multiple Alignment using Fast Fourier Transform) [84] and analyzed using Jalview [85].</p><!><p>CD spectra were acquired from H3-derived peptides dissolved in water (30 µM) at increasing concentrations of TFE using a Bio-Logic CD spectrometer with a cuvette with a path length of 1 mm, at a temperature of 20°C, with a bandwidth of 0.5 nm and a sampling time of 0.5 s. Each spectrum represents the average of three accumulations minus the signal from the blank. Additionally, a constant was added or subtracted to CD spectra so that ellipticity at high wavelengths was 0. Spectra deconvolution was performed using the CONTIN algorithm [86] implemented into DichroWeb [87].</p><!><p>The atomic coordinates and structure factors have been deposited in the PDB with the accession number: PDB ID: 5T8R (BAZ2A PHD in complex with H3 10-mer). NMR assignments for BAZ2A PHD and BAZ2B PHD have been deposited in the BMRB with deposition numbers 26 754 and 25 988, respectively.</p>

PubMed Open Access

Nitrite Reductase Activity in Engineered Azurin Variants

Nitrite reductase (NiR) activity was examined in a series of dicopper P.a. azurin variants in which a surface binding copper site was added through site-directed mutagenesis. Four variants were synthesized with copper binding motifs inspired by the catalytic type 2 copper binding sites found in the native noncoupled dinuclear copper enzymes nitrite reductase and peptidylglycine \xce\xb1-hydroxylating monooxygenase. The four azurin variants, denoted Az-NiR, Az-NiR3His, Az-PHM, and Az-PHM3His, maintained the azurin electron transfer copper center, with the second designed copper site located over 13 \xc3\x85 away and consisting of mutations Asn10His,Gln14Asp,Asn16His-azurin, Asn10His,Gln14His,Asn16-His-azurin, Gln8Met,Gln14His,Asn16His-azurin, and Gln8His,Gln14His,Asn16His-azurin, respectively. UV\xe2\x80\x93visible absorption spectroscopy, EPR spectroscopy, and electrochemistry of the sites demonstrate copper binding as well as interaction with small exogenous ligands. The nitrite reduction activity of the variants was determined, including the catalytic Michaelis\xe2\x80\x93Menten parameters. The variants showed activity (0.34\xe2\x80\x930.59 min\xe2\x88\x921) that was slower than that of native NiRs but comparable to that of other model systems. There were small variations in activity of the four variants that correlated with the number of histidines in the added copper site. Catalysis was found to be reversible, with nitrite produced from NO. Reactions starting with reduced azurin variants demonstrated that electrons from both copper centers were used to reduce nitrite, although steady-state catalysis required the T2 copper center and did not require the T1 center. Finally, experiments separating rates of enzyme reduction from rates of reoxidation by nitrite demonstrated that the reaction with nitrite was rate limiting during catalysis.

nitrite_reductase_activity_in_engineered_azurin_variants

INTRODUCTION<!>Reagents<!>Mutagenesis of the Azurin Scaffold<!>Protein Expression, Purification, and Metal Ion Reconstitution<!>Spectroscopic Methods<!>Electrochemistry<!>Activity Assays<!>Design and Synthesis<!>UV\xe2\x80\x93Visible Absorption Spectroscopy<!>EPR Spectroscopy<!>Electrochemistry of T1 and T2 Centers<!>Nitrite Reduction Activity: Control Reactions<!>Nitrite Reduction Activity: Michaelis\xe2\x80\x93Menten Kinetics<!>Nitrite Reduction Activity: Single Turnover Reduction vs Reoxidation Rates of Azurin Variants<!>Rates of Reoxidation of Azurin Variants, in Comparison to Michaelis\xe2\x80\x93Menten Rates<!>CONCLUSIONS<!>

<p>Copper nitrite reductase (NiR) catalyzes the reduction of nitrite ions to nitric oxide within bacterial dissimilatory denitrification pathways that convert nitrate ions to dinitrogen gas.2,3 The enzyme contains two different copper centers.4 One center is classified as a type 1 (T1) copper center, with a characteristic intense blue color (absorption ~600 nm), narrow EPR hyperfine splitting value (~60 G), and an electron transfer function.5,6 The other center is a characteristic type 2 (T2) copper center with a weakly absorbing Cu(II) ion, broad EPR hyperfine coupling constants (~150 G), and catalytic function.5,6 NiR and related proteins have been referred to as noncoupled dinuclear copper proteins, as the two ions have undetectable magnetic coupling and are separated by over 12 Å.7,8</p><p>Native NiR activity, or the acceleration of conversion of nitrite to nitric oxide by the enzyme, utilizes two copper ions to fulfill the one-electron reduction of nitrite. The native enzyme is quite efficient with reported turnover numbers of 8 s−1 9 to 4000 s−1 10.11–20 Since KM values for nitrite by native NiRs are often in the micromolar range (14 µM18 to 500 µM,21 with 50 µM being common9,12,13,20,22–25), the catalytic efficiencies (kcat/KM) fall in the 105–107 s−1 M−1 range.9,13 NiRs are therefore very efficient, as diffusion-controlled limits for enzymes in general are 108–109 M−1 s−1.</p><p>Synthetic models of copper NiR have been prepared and characterized (see for example refs 26–28), including models that bind nitrite or NO. Because only a single copper ion is theoretically needed for the one-electron reduction of nitrite, some of these complexes demonstrated activity by reaction with nitrite to produce NO either in organic solvent29 or in aqueous systems.30–35 The observed activity has ranged from single turnover30,36–43 between the reduced form of the model and nitrite to full catalytic turnover in a few models.29–35 For example, recent studies of de novo designed copper binding helical peptides report nitrite reduction to NO at significant levels above background.35 Small-molecule models using a bipyridylamine ligand functioned electrocatalytically on an electrode surface.31–33 The reported turnover numbers for these and other model systems are less than those for the native systems and range from 1 × 10−4 to 5 s−1,29,31–35 while reported KM values are larger at 1–15 mM,31–33 and the resulting kcat/KM values are 10–60 s−1 M−1.31–33 These complexes all contain a single copper ion per molecule that mimics the catalytically active T2 center in native Cu NiRs.</p><p>Here we report the nitrite reduction activity of four models of NiR designed into the small soluble T1 copper protein azurin. The four variants were designed after comparisons of azurin to native enzymes containing T2 catalytic Cu sites. NiR and peptidylglycine α-hydroxylating monooxygenase (PHM) contain solvent-exposed T2 catalytic copper sites, in addition to a distant electron transfer copper center. Two models in azurin, denoted Az-NiR and Az-PHM,44 were constructed by adding a second surface copper binding site in addition to the existing blue copper center (Figure 1). The amino acids mutated for the designed T2 copper site were Asn10His,Gln14Asp,Asn16His in Az-NiR and Gln8Met,Gln14His,Asn16His in Az-PHM.44 These particular residues were identified because their arrangement on an antiparallel β-sheet region of azurin resembled that of the T2 copper sites in NiR and PHM, respectively. The three residues in azurin emulated the triangular facially coordinating arrangement of residues in the native copper centers.44 Furthermore, as in the native systems, these residues were solvent exposed and estimated to be 14 and 15 Å away from the T1 Cu center in Az-NiR and Az-PHM, respectively. These two models in azurin therefore contained an added T2 copper center with two His and a potential third ligating amino acid: Asp in Az-NiR or Met in Az-PHM. In addition, the third and fourth azurin variants presented here incorporate three histidines in the designed T2 copper binding site by changing the Asp or Met in the above variants, respectively, to give rise to Az-NiR3His (Asn10-His,Gln14His,Asn16His-azurin) or Az-PHM3His (Gln8His,Gln14His,Asn16His-azurin). Here we report the NiR activity of these four NiR models along with their spectroscopic characterization.</p><!><p>All chemicals were of reagent grade or higher and used without further purification. The gene for P. aeruginosa azurin in a pET9a plasmid was kindly provided by Prof. Yi Lu (University of Illinois Urbana-Champaign, Urbana, IL) with permission from Prof. John H. Richards (California Institute of Technology, Pasadena, CA).45</p><!><p>The models of NiR were designed into the P.a. azurin scaffold as described previously.44 To design the 3-histidine-containing active sites in azurin, the crystal structure of azurin (PDB ID: 4AZU1) was compared with that of NiR (PDB ID: 1NIR46) using RasWin Version 2.7.5.2. Upon identification of the desired mutations, pDRAW32 1.0 (ACACLONE) was used to assist in the design of DNA primers for the site-directed mutagenesis experiments. Primers (57 bps in length) for Quikchange mutagenesis (Stratagene, San Diego, CA) on the WT azurin-pET9a plasmid were synthesized and HPLC purified by Eurofins MWG Operon (Huntsville, AL). In addition to the amino acid changes, the primers incorporated (+) or deleted (−) new restriction sites for each mutation ((−) EcoRV for Az-NiR and Az-NiR3His, (+) NsiI for Az-PHM and Az-PHM3His) for screening of the desired plasmids. Quikchange mutagenesis and DNA isolation methods were described previously.44 The resulting plasmids were sequenced at the University of Minnesota Genomics Center, confirming the mutations and the integrity of the entire azurin gene. The names given here are Az-NiR (Asn10His,Gln14Asp,Asn16His-azurin; same as "Az-NiR" in ref 44), Az-NiR3His (Asn10His,Gln14His,Asn16His-azurin), Az-PHM (Gln8Met,Gln14His,Asn16His-azurin; same as "Az-PHM/DBM" in ref 44), and Az-PHM3His (Gln8His,Gln14His,Asn16His-azurin).</p><!><p>The plasmid DNA was transformed and the protein was expressed in BL21*(DE3) E. coli (Novagen, Madison, WI). Azurin proteins were purified as described earlier.44 Proteins were overexpressed from E. coli grown at lower temperatures (26 °C) to minimize background expression during cell growth and thus unhealthy cell cultures. Typical yields for the purified azurin variants ranged from 15 to 60 mg/L of cell culture. The protein identities were further verified by ESI-MS (Table S1 in the Supporting Information). Proteins were isolated in the apo form and reconstituted with Cu2+ in an experiment-dependent manner: First, type 1 Cu only samples were prepared by the addition of 0.75 equiv of aqueous CuSO4 per apo azurin. Second, fully reconstituted T1 and T2 copper samples were made by the addition of 5 equiv of CuSO4 per apo azurin, followed by removal of excess ions with a size-exclusion PD-10 column with Sephadex G-25. EPR showed that this protocol resulted in 100% type 1 copper loading and >90% type 2 copper loading in 20 mM ionic strength47 sodium phosphate buffer pH 6.35. Third, type 2 copper only samples were prepared by blocking the T1 site with addition of 1.5 equiv of aqueous HgCl2 per protein, followed by size exclusion chromatography to remove excess ions. Then, 0.75 equiv of CuSO4 was added to populate T2 sites.</p><!><p>UV–visible absorption spectra were obtained on a Shimadzu UV-2401 PC spectrophotometer at room temperature. Proteins were ~0.1 mM buffered in 50 mM ammonium acetate pH 5.1. The absorptivity of azurin (ε625 = 5000 M−1 cm−148,49) was used to determine those for the variants in this study by integrating X-band EPR signal intensities.50 Reported absorptivities were the average of three measurements and have an error of ±10%. EPR spectra were obtained on an X-band Varian EC-1365 spectrometer fit with a Wilmad (Buena, NJ) liquid-nitrogen sample Dewar at 77 K. Data were measured on ~0.5 mM protein samples in 50% glycerol and 25 mM ammonium acetate pH 5.1, with instrument parameters of frequency ~9.27 GHz, microwave power 0.5 mW, and modulation amplitude and frequency of 5 G and 100 kHz, respectively. All EPR spectra were simulated with the SIMPOW6 program.51 ESI-mass spectra were obtained at the University of Minnesota, Department of Chemistry Mass Spectrometry Facility, on a Bruker BioTOFII. Samples were prepared in 50 mM ammonium acetate buffer pH 5.1 with 1 equiv of added copper. The MS data were used to confirm the identity of each variant and the binding of the higher affinity T1 Cu(II) ions (Table S1 in the Supporting Information).</p><!><p>Reduction potentials of the variants were measured with cyclic voltammetry techniques using a CH Instruments 620B electrochemical analyzer and Faraday cage. A pyrolytic graphite edge (PGE) working electrode, aqueous standard calomel reference electrode (E = 0.241 V vs NHE at 25 °C), and platinum-wire counter electrode were used. The working electrode was constructed as described previously.52 Prior to each experiment, the working electrode was polished for 1–5 min with a 1 µm alumina powder paste, rinsed with deionized water, and briefly sonicated. The 200 µL ~0.1 mM deaerated protein sample was incubated at 25 °C in a 1 mL cell blanketed with argon gas inside a Faraday cage. Direct electrochemistry of the variants with a 0.025 V/s scan rate gave strong pseudo-reversible signals in 50 mM ammonium acetate pH 5.1 buffer. Broader than ideal anodic and cathodic peak separations of ~100 mV were due to slower macromolecular diffusion rates, as confirmed by scan rate dependent analyses of peak separations.53–55 All reduction potentials are reported versus NHE unless otherwise specified.</p><!><p>All nitrite reductase activity assays were conducted in a Coy Laboratories anaerobic chamber (Grass Lake, MI, USA) under a nitrogen + 2% hydrogen gas atmosphere. Anaerobic conditions resulted in more precise data for reactions of nitrite with Cu(I)-azurin, as competition with O2 was significant over the longer (minutes to hours) reaction time scales and varied with mixing. In addition, stoichiometric amounts of the reductant ascorbate could be used to reduce copper ions anaerobically without noticeable consumption by the side reaction with O2. For the Michaelis–Menten kinetic experiments, samples were 300 µL in volume in 20 mM sodium phosphate buffer pH 6.35,47 with 86 µM azurin variant (1.2 mg/mL), 33 mM ascorbate as reductant, and sodium nitrite concentrations varying from 1–150 mM. Nitrite aliquots were added to stirred solutions of protein followed by addition of ascorbate solution to initiate the reaction. The ascorbic acid and sodium nitrite stock solutions were prepared in buffer, and pHs were adjusted to 6.35 if necessary to maintain a constant pH for all assay conditions. Ascorbic acid solutions are thus referred to as ascorbate, since addition of some NaOH was required to reach pH 6.35. The rate of nitrite reduction was determined by using the Griess method56–59 with 20 µL of the reaction removed every 1–15 min and added to Griess reagents. The 20 µL aliquots were diluted in buffer (to a concentration of less than 400 µM NO2− for the most concentrated nitrite samples) to 500 µL and then mixed with 500 µL of 58 mM sulfanilic acid (SAN) in 3 M HCl followed by 500 µL of 3.85 mM N-(1-naphthyl)ethylenediamine dihydrochloride (NED). Care was taken to add NED rapidly and consistently after the addition of SAN to minimize the side reaction of the diazonium intermediate species with ascorbate.56 If necessary, reacted samples were then diluted again with buffer to give an absorbance of less than 1 au. Absorbances were measured on a Thermo Scientific Genesys 10S scanning UV–vis spectrophotometer. Initial rates of reaction were determined from the slope of the first 5–10% loss in nitrite concentration with time. Assays at each nitrite concentration were repeated at least three times.</p><p>To determine the rates of reduction, varied concentrations of ascorbate were added to 2 mL of 100 µM azurin in a stirred cuvette in 20 mM sodium phosphate buffer pH 6.35.47 The rate of reduction of the blue type 1 Cu center (A625 nm) was monitored with time, and a rate constant was derived by fitting the curve with a first-order exponential function. The resulting rate constants were plotted versus ascorbate concentration to derive the rate dependence.</p><p>Rates of reoxidation of the type 1 Cu center with the addition of nitrite were determined by first reducing 2 mL of 100 µM azurin variant in a stirred cuvette in 20 mM sodium phosphate buffer pH 6.3547 with 1 equiv of ascorbate. Ascorbate provided two electrons to reduce both copper centers. Once the protein was fully reduced (A625 nm of the T1 copper center disappeared), varied amounts of NO2− were added. The single turnover of the enzyme was observed by monitoring the recovery of the A625 nm value with time. The rates of reoxidation were determined by finding the initial slope and plotting it against the concentration of NO2−. Experiments were performed in triplicates.</p><p>The formation of NO(g) as the reaction product was verified by bubbling 1 mL of catalyzed reaction head space gas into an Fe[EDTA]2− solution (Figure S3 in the Supporting Information).60,61 The UV–vis absorption spectrum of the readily formed characteristic yellow [Fe(NO)(EDTA)]2− complex was observed.35,37,62 In the anaerobic chamber, 0.4 mL of 0.1 M H2SO4 and 0.3 M FeSO4 and 0.6 mL of 0.5 M NaOH and 0.2 M EDTA were mixed in a cuvette to generate [Fe(EDTA)]2−. The spectrometer was base-lined with this solution before injection of 1 mL of head space gas from a reaction with 260 µM Az-NiR3His, 100 mM nitrite, and 33 mM ascorbate.</p><p>Reaction reversibility was confirmed by measuring an increasing nitrite concentration with the Griess test after adding NO(g) to a solution of 86 µM Az-NiR3His and 1 mM NaNO2. NO(g) was generated either in a sealed anaerobic mixture of copper metal and aqueous HNO3 or in a sealed reaction vessel with 86 µM Az-NiR3His, 33 mM ascorbate, and 100 mM NaNO2. Headspace gas from these reactions (2 mL) was added to the headspace of the stirred Az-NiR3His solution.</p><!><p>Four models of NiR were made in P.a. azurin by adding a T2 copper binding site on a surface region of antiparallel β-sheet resembling that of native NiR (Figure 1). Two azurin variants, termed Az-NiR3His and Az-PHM3His, were intended to mimic the native NiR catalytic T2 Cu sites that contain three histidines as primary coordinating ligands. These azurin variants share a similar arrangement of mutations to the other two models, named Az-NiR and Az-PHM, which themselves were designed after native catalytic T2 copper sites.44 However, the 3His variants contain a third His residue, which is known to favorably coordinate metal ions.</p><!><p>The addition of CuSO4 to all azurin variants resulted in blue solutions with UV–vis absorption spectra characteristic of azurin and T1 copper proteins. Spectra of the variants were very similar to each other and to that of WT azurin (Figure 2 and Table 1). The absorptivity values, as determined by integration of EPR signals, of the dominant S(Cys) → Cu charge transfer transition at 625 nm were all within experimental error and were around 5000 M−1 cm−1.48,49 The added mutations for the second copper binding site onto the azurin surface did not significantly perturb the T1 copper center, as might be expected from their large separation (>13 Å). Absorption peaks from the T2 Cu(II) center were not apparent in the T1 Cu(II) spectra, as absorptivity values of the former are typically around 2 orders of magnitude smaller than the latter.</p><p>We were able to observe the absorption spectra of the T2 Cu(II) centers alone by blocking the higher affinity T1 site of our azurin models with colorless Hg(II). Azurin with a T1 site saturated by Hg(II) yielded weak-intensity blue samples following reconstitution of the T2 site with CuSO4 (Figure 3). Samples required protein concentrations of over 0.25 mM in order to discern the weakly absorbing Cu(II) centers. Overall, the spectra of the mutants were similar to each other in that they contained broad and low-intensity Cu(II) d → d transitions located around 700 nm (Figure 3 and Table 1) with absorbance maxima of 722, 690, 677, and 715 nm for Az-NiR, Az-NiR3His, Az-PHM, and Az-PHM3His, respectively. Their absorption spectra were distinct from that of free CuSO4 in buffer. The absorbance maxima and absorptivity values were consistent with those of other native and designed T2 copper centers, which vary depending on geometry, ligand number and type, and protonation state of the Cu(II) complex. They were similar to other histidine-rich Cu(II) centers in distorted-square-pyramidal geometries.65–71 T2 Cu(II) centers in superoxide dismutase (SOD), for example, demonstrated broad Cu(II) d–d transitions at 670–700 nm,66,72 and those of the trihelical peptides were around 640 nm,34,35 while those of other N,O-rich square-pyramidal synthetic model complexes were ~650–720 nm.36,43,68–70 The spectrum of a native NiR T2 copper center was reported to be 790 nm with an absorptivity of 85 M−1 cm−1.64 The absorptivity values of the T2 centers on azurin were low, as expected for Cu(II) d → d transitions. They were ε722 = 51, ε690 = 65, ε677 = 31, and ε715 = 43 M−1 cm−1 for Az-NiR, Az-NiR3His, Az-PHM, and Az-PHM3His, respectively. They were on the low end of values of ε ≈ 70–150 M−1cm−1 reported for similar aqueous systems.34,35,43,64,66,67,72</p><p>Furthermore, the effects of the small exogenous ligands nitrite, azide, chloride, and imidazole on the absorption spectra of T2 Cu(II) only samples were explored, in order to probe their substrate binding abilities (Figure 4 for Az-NiR3His and Figure S1 in the Supporting Information for all variants). Overall, the results were similar to other related T2 Cu(II) systems where the absorption spectra were not very sensitive to small ligands, as the observed changes occurred only with addition of multiple equivalents.35,73,74 Titration with azide or imidazole perturbed the spectra for all of the azurin variants in the same way. Azide (at 10–20 equiv and higher) resulted in peaks that were blue-shifted for each variant with an ~50% increase in absorptivity. The addition of azide to free CuSO4 resulted in a similar increase in absorptivity, but the peak was about 100 nm lower in energy than the azurin variants (Figure 4 and Figure S1). The exogenous ligand imidazole did not initially affect the spectra of any variants up to 10–20 equiv. Upon addition of greater amounts (20–100 equiv), the spectrum began to shift to an intensity and position approaching that of free CuSO4 and imidazole in solution (Figure S1), thus indicating that Cu(II) was pulled from the protein binding site. Finally, the addition of nitrite or chloride ions to the variants and free CuSO4 did not appreciably influence their spectra up to 100 equiv (25 mM). Furthermore, there was no significant difference between titrations performed in 50 mM ammonium acetate pH 5.1 buffer and 20 mM sodium phosphate pH 6.35 buffer for the Az-NiR3His variant. The absorption titrations, particularly with azide, suggest that ligand binding is possible to the designed T2 copper centers on azurin. Data on other systems are sparse due to T2 Cu(II)'s low absorptivity and frequent interference from other copper centers. However, the weak response of these azurin variants, to nitrite in particular, were consistent with the UV–vis absorption data for T2 centers in the CuM site of PHM, which required 1500 equiv of nitrite to observe full perturbation of the MCD spectrum.73 EPR, however, was more sensitive in this regard.</p><!><p>X-band EPR spectra of the copper sites in the azurin variants were obtained on all proteins reconstituted with a single T1 Cu(II), both T1 Cu(II) and T2 Cu(II), and a single T2 Cu(II) (T1 Hg(II)–T2 Cu(II) sample) (Figure 5). All spectra were simulated for accurate determination of g and A values51 (Figure 5 and Table 1).</p><p>The T1 Cu(II) EPR spectral parameters for all the variants were similar to those of WT azurin (Figure 5A) and characteristic of the T1 blue copper protein family with axial line shapes and narrow Cu hyperfine splittings.75 The g∥ values were ~2.26, while gx and gy were similar to each other, in the range of 2.04–2.05 across the variants. The T1 Cu(II) A∥ hyperfine splitting values were similar for all azurin variants, ranging from 47 to 53 G, in comparison to the WT A∥ value of 55 G. The common EPR parameters across the variants indicated a conserved T1 Cu(II) site structure that was largely unperturbed by the addition of the second surface T2 Cu binding site.</p><p>EPR spectra of the T2 Cu(II) centers in each variant were similar to each other (Table 1 and Figure 5B), likely resulting from the common His-supported, solvent-exposed surface binding site (Figure 1). The EPR line shapes and parameters were characteristic of T2Cu(II) proteins, including native NiR. The axial spectra (g∥ ≈ 2.30, gx,y ≈ 2.05–2.07) were slightly more rhombic than the T1 centers (g∥ ≈ 2.26, gx,y ≈ 2.04–2.05) and were similar to those of native NiR T2 Cu(II) sites, which have comparable values of g∥ ≈ 2.30–2.35 and gx,y ≈ 2.04–2.10.10,14,17,18,63,76–79 The g∥ values and the A∥ hyperfine splittings (160–165 G) were clearly discerned from that of free Cu(II) in solution (Figure 5B) and from that of 1 equiv of Cu(II) added to WT azurin (g∥ ≈ 2.37 and A∥ ≈ 142 G).44 The A∥ values of the variants were similar to those of other T2 Cu(II) proteins and slightly larger than those of native NiR T2 Cu(II) sites, which have A∥ ≈ 110–150 G.10,14,17,18,63,76–79 The EPR axial line shape and spectroscopic parameters of the azurin variant T2 Cu(II) centers were consistent with those of other five-coordinate copper sites of mixed nitrogen- and oxygen-donating ligands in a distorted-square-pyramidal geometry.36,43,66–68,80 The readily discerned copper hyperfine splittings in the g∥ region demonstrated a dx2–y2 ground state common in T2 Cu(II) centers.81 Small differences in the spectroscopic values between the variants reported here are likely results of the different residues in the four azurin variants in varied locations on the protein surface.</p><p>Finally, the T1 Cu(II)–T2 Cu(II) doubly bound samples were analyzed by EPR spectroscopy. This was an effective tool for discerning the T1 and T2 coppers in these models, as well as the presence of any free Cu(II) ions. Simulations of these spectra were accomplished by simply summing the simulated parameters for the T1 Cu(II) site with those of the T2 Cu(II) site. The sum of these two species accurately recreated the doubly bound spectra (Figure 5C) and illustrated their noncoupled nature.7</p><p>These EPR simulations also allowed us to measure the relative loading of the T1 versus T2 sites. The relative amount of the T2 Cu(II) signal needed to accurately simulate the experimental EPR curves never exceeded the amount of T1 Cu(II) signal and was in the 70–100% range. Our method for reconstituting these model systems with Cu(II) prior to assays was effective. The simulations demonstrated that, in 50 mM ammonium acetate at pH 5.1, the samples bound T1 Cu(II) with 70–100% T2 Cu(II) loadings and no apparent free Cu(II). Similar experiments using assay conditions (20 mM sodium phosphate buffer pH 6.35) showed 90–100% Cu(II) loadings of the T2 copper site with no free copper.</p><p>EPR was more sensitive to the influence of small exogenous ions on the T2 Cu(II) site, with similar but more clearly discerned changes in comparison to those observed above by UV–vis absorption spectroscopy. EPR spectra demonstrated small changes with added nitrite, while azide and imidazole gave the largest perturbations (Figure S2 in the Supporting Information). Nitrite up to 200 mM resulted in small decreases (3–5 G) in the hyperfine coupling values for the four azurin variants, with very small decreases in g values of ~0.005. Azide effects on the EPR spectra of all four azurin variants were observed at smaller concentrations (50 mM) with decreased A∥ values (2–6 G) and large decreases in g∥ of ~0.05 (~2.29 to ~2.24). Finally, addition of imidazole to the variants resulted in the most dramatic perturbations, where ~30 G increases in A values and about 0.05 decreases in g values were observed with addition of up to 20 mM imidazole. Addition of further amounts of imidazole resulted in the appearance of a free copper signal, consistent with the UV–visible absorption titration data discussed above.</p><p>The overall trends in small-molecule interactions with the T2 Cu(II) sites as observed by EPR were consistent with those found by UV–vis absorption data and consistent with those of native systems. Azide and imidazole resulted in the largest changes, while the spectra were only slightly perturbed by the addition of nitrite. In native NiR and related systems, the effects were similar; however, they required as little as a few equivalents of added nitrite76,78 to as much as 1500 equiv.73 Where reported, the changes in EPR parameters upon nitrite binding included either small increases14,73,76,78 or decreases18,35,79 in A∥ values of around 10 G, with a correspondingly decreased18,76,78,79 or sometimes unperturbed14,35,73 g∥ value. In summary, the data on the azurin variants suggest that exogenous ligands interact with the designed surface copper centers in a manner consistent with T2 Cu(II) centers, such as those of native NiR.</p><!><p>Cyclic voltammetry was used to measure the reduction potential of the azurin variants. The cyclic voltammograms for WT and the four azurin variants were obtained by direct electrochemistry on ~100 µM samples with a PGE working electrode (Figure 6 and Table 1). Quasi-reversible signals with peak separations of about 100 mV were observed. The larger peak separations were due to the slower diffusion rate of macromolecules, consistent with experimental work on similar samples.53 We first measured samples with only a T1 Cu(II) ion. The measured reduction potentials of all of the T1 Cu centers were very similar within experimental error and were around 355 mV vs NHE.</p><p>Cyclic voltammetry was performed on T2 Cu(II) only samples by first blocking the T1 center with Hg(II) before reconstituting the designed T2 centers with Cu(II) (Figure 6 and Table 1). The reduction potentials of the resulting T2 Cu(II) centers were about 100 mV lower than those of the T1 Cu(II) centers and showed some variation within the variants. The potential for Az-NiR3His, Az-NiR, Az-PHM3His, and Az-PHM were 270, 257, 258, and 251 mV, respectively, vs NHE. The reduction potentials were similar to those reported for the T2 Cu(II) sites of native NiRs that range from 230 to 280 mV76,82–84 at pH 7.0 and 310–340 mV82,83 at pH 6.0. Furthermore, the native T2 Cu(II) reduction potentials are reported to be similar to or higher than those of the native T1 sites.5 The finding that the azurin variants have lower T2 Cu(II) reduction potentials than the T1 Cu(II) reduction potentials indicates that the equilibrium could be shifted toward a reduced T1 Cu center.</p><!><p>All four azurin variants were found to catalytically reduce nitrite using ascorbate as a reductant. Nitrite degradation was monitored using the colorimetric Griess assay that detects nitrite. Assays were conducted with 86 µM protein loaded with 1 or 2 equiv of Cu(II) in the presence of varied sodium nitrite (1–150 mM) and excess ascorbate (33 mM). The milder ascorbate was used as the reductant because sodium dithionite was harmful to azurin's stability and because dithionite was found to have a significant background rate of reaction with nitrite. It was determined in control reactions with varied ascorbate concentrations from 1 to 60 mM that concentrations as low as 10 mM could be used with no change in observed rates of nitrite reduction. An ascorbate concentration of 33 mM, about 400-fold higher than that of the enzyme, was used in the assays. In addition (vide infra), it was found that the rate of azurin reduction by ascorbate was at least ~100 times faster than the rate of azurin oxidation by nitrite. Therefore, the reduction of the azurin variants by ascorbate was not rate limiting. Finally, the resting Cu(II) form of the azurin variants did not result in any NO2− loss; consequently, reactions were initiated by the addition of ascorbate.</p><p>As with native NiR enzymes, the reaction pH was important. A reaction pH of 6.35 was used in our assays. The azurin variants were similar to native enzymes that have catalytic rate optima around pH 5.5 with slower product formation at higher pHs (>7.0).2,5 However, at lower pHs (<6.0) our assays showed significant background or uncatalyzed reaction rates between nitrite and ascorbate. At pH 6.35 only ~0.05% of the nitrite disappeared in the background uncatalyzed reaction between 33 mM ascorbate and 100 mM nitrite over the 2 h reaction times, whereas in the presence of azurin variants under these conditions, 5–20% of the nitrite was reduced.</p><p>NO(g) was found to be the product of nitrite reduction with the azurin variants by generating the characteristic yellow NO adduct of [Fe(EDTA)]2− (Figure S3 in the Supporting Information).35,37,60–62 Addition of 1 mL of reaction headspace gas to a sealed [Fe(EDTA)]2− solution generated the signature absorption spectrum of the Fe(II)–NO adduct with a peak at 432 nm, thus demonstrating the ready production of NO by the azurin variants.</p><p>WT azurin was tested as a catalyst, prepared in the same way as the protein models by addition of 5 equiv of Cu(II) followed by treatment with a desalting column to remove excess ions. WT azurin, lacking a T2 Cu(II), gave little to no reduction of nitrite with 33 mM ascorbate and varied nitrite concentrations from 1 to 150 mM (Figure S4 in the Supporting Information). Second, the rates with WT azurin were comparable to the background uncatalyzed rate of nitrite reduction (Figure S4). Third, the rates of nitrite reduction of the four azurin variants with only a T1 Cu(II), lacking the second T2 Cu(II) ion, yielded kinetics that were the same as that of WT azurin, with rates at least 100 times slower than the dicopper(II) populated proteins. These experiments highlighted the critical role of a catalytic T2 Cu center in the reaction with nitrite, which is well-known with native NiR species.5,19,77</p><p>In further control reactions the reversibility of the chemistry catalyzed by the azurin variants was demonstrated by the production of nitrite from NO gas. NO was generated under an inert atmosphere either by mixing aqueous nitric acid with copper metal or by mixing 86 µM Az-NiR3His with 33 mM ascorbate and 100 mM NaNO2. Headspace gas (2 mL) from this reaction was added to the headspace of a buffered solution of azurin variant in the oxidized Cu(II) state. The resulting solution faded to blue, indicating a reduction of the azurin variant T1 copper center, presumably as a result of the T2 Cu center being reduced first, since reactions with WT azurin or T1 Cu(II) only variants did not result in a reduced protein. Wever et al. observed T1 Cu(II) sites to reduce slightly under 1 atm of NO(g) (~15% yield of Cu+–NO+), while here no reaction was observed under such mild addition of NO.85 Removal of reduced solutions from an inert atmosphere to an oxygen atmosphere resulted in a full recovery of the blue color, demonstrating the turnover capability of the azurin variants. Furthermore, we used the Griess assay (as well as the appearance of a 350 nm peak characteristic of nitrite) to confirm that nitrite was formed and increased in concentration over time with the addition of NO(g). No nitrite was observed with WT azurin. This confirmed the ability of the azurin variants to catalyze the nitrite redox chemistry in both directions, as was demonstrated with native NiR systems.13,86</p><p>In single enzyme turnover experiments, the azurin variants were reduced with ascorbate and then reoxidized with excess nitrite. For reduction, the addition of 0.5 equiv of ascorbate was required to fully reduce the blue T1 copper center of WT azurin, which contains only one Cu(II) center. However, the addition of 1.0 equiv of ascorbate was required to fully reduce both Cu(II) centers in the model systems. Addition of 0.5 equiv of ascorbic acid to these dicopper variants resulted in a decrease in blue color of ~50%. This demonstrated that two copper ions were present in the model proteins, as verified by EPR. The reduction reaction was also probed by EPR. After 60 s of mixing Az-NiR3His with 1 equiv of ascorbate, EPR showed that more T2 copper was reduced than T1 copper, indicating that T2 copper reacted with reductant ahead of the T1 Cu (Figure S5 in the Supporting Information). All reduced azurin variants remained stably reduced under the inert atmosphere but fully returned to the blue color if exposed to oxygen gas or nitrite.</p><p>The addition of excess sodium nitrite to reduced azurin variants resulted in a return of blue color or reoxidation of the T1 and T2 copper centers. This reaction was also monitored by EPR, which showed the T2 copper center reoxidized before the T1 center. The amount of nitrite reduced upon this single turnover reoxidation of the azurin variant was quantified, and results indicated that two nitrites were consumed per azurin protein. This was demonstrated by the loss of 185 µM nitrite following the addition of 1 mM nitrite to 100 µM Az-NiR3His, which was prereduced with 100 µM ascorbate. On the other hand, with the singly T2 Cu(II) loaded T1 Hg(II)–T2 Cu(II) Az-NiR3His protein, about 60 µM nitrite was reduced using 100 µM protein (100 µM T2 Cu(II)), which was prereduced with 1/2 equiv of ascorbate (50 µM). This showed that electrons from both the T1 and the T2 copper centers were delivered to reduce nitrite by consecutive reaction with two different nitrite molecules. Finally, although WT azurin was readily reduced by ascorbate, the T1 Cu(I) center did not revert to blue or reoxidize with the addition of nitrite over days of monitoring in the glovebox. This demonstrated that the designed T2 copper surface site was important for catalysis in the azurin variants and that any reduction of nitrite directly by the T1 copper center was negligible.</p><p>Finally, to evaluate the robustness of the azurin catalysts, their stability was tested by repeating a single turnover experiment multiple times. In the experiment, 100 µM Az-NiR3His was equilibrated with 10 mM NaNO2 in an anaerobic chamber. The absorbance at 625 nm, of the T1 copper center in the dicopper azurin variant, was monitored. The addition of 2 equiv of ascorbic acid reduced the T1 copper center, as evidenced by the decrease in blue color. Nitrite was consumed and the blue color returned, indicating reoxidation by NO2−. Upon recovery of A625, 2 equiv more of ascorbic acid was added. This process was repeated multiple times for 10 h (Figure 7). The multiple turnover cycles and the full recovery of T1 Cu(II) confirmed the catalytic capability of the azurin variants and their robust nature. This indicated that the protein was not readily denatured under these conditions and that it was capable of catalyzing multiple turnovers.</p><p>Results from all of the above nitrite reduction experiments demonstrated the robust nature and catalytic capabilities of the azurin variants. The formation of NO product from nitrite reduction, as well as the reversible production of nitrite from NO, was similar to that of native copper NiRs. The above reactions also demonstrated the requirement of the T2 surface copper site for activity, as WT azurin and all variants with a T1 Cu(II) only had no activity above background. Finally, it was shown that electrons from both the T1 copper site and T2 copper were donated to nitrite by following the gain in characteristic blue color with stoichiometric reduction of nitrite. We therefore further determined the catalytic Michaelis–Menten parameters of the four azurin variants with regard to nitrite.</p><!><p>Michaelis–Menten parameters of our azurin variants were determined by monitoring the concentration of nitrite with time using the colorimetric Griess assay. The initial rate of reduction (v0) was found by following the concentration of nitrite until a 5% decrease was observed. Assays for each variant were run in triplicate and at nitrite concentrations from 1 to 150 mM. Tests of solution ionic strength influences were conducted where the ionic strength of the solution was held to a constant level by supplementing varied amounts of NaNO2 with NaCl to maintain a total salt concentration of 100 mM. The rates derived from these reactions were the same as those where only NaNO2 was varied without the addition of NaCl. Each reaction assay used 86 µM protein, with 2 equiv of Cu(II) as confirmed by EPR, and excess ascorbic acid (33 mM) with pH adjusted to 6.35 in 20 mM phosphate buffer. The rates were plotted versus the corresponding nitrite concentrations to generate the Michaelis–Menten plots (Figure 8).</p><p>The azurin variants show significant nitrite reduction activity (Figure 8) in comparison to WT azurin, which showed no detectable activity above background (Figure S4 in the Supporting Information). The KM and Vmax values are reported in Table 2, and the turnover numbers were determined using 86 µM azurin variant. Among the azurin variants, Az-PHM3His and Az-NiR3His had the highest turnover numbers of 0.59(13) and 0.56(5) min−1, respectively. Az-NiR and Az-PHM had lower turnover numbers with 0.34(5) and 0.38(9) min−1, respectively. The higher turnover numbers of Az-NiR3His and Az-PHM3His may reflect the importance of the third histidine. The Az-PHM and Az-NiR variants have a Met or Asp in addition to two added histidines on the designed T2 site, whereas Az-NiR3His and Az-PHM3His have three histidines, similar to the native NiR. We have observed by EPR that three histidines in the T2 site retain the copper(II) ion better following size exclusion chromatography and thus could indicate a more stable active site. The histidines could also assist in the important role of providing the necessary proton in the formation of the product OH−/H2O from NO2− reduction.</p><p>The KM and Vmax values for the variant Az-NiR3His with only a T2 copper ion were also determined. The T1 site was blocked with Hg(II) ions prior to reconstitution of the T2 site with Cu(II). This variant displayed kinetic parameters that were very similar to those of Az-NiR3His with two copper ions (Figure 8 and Table 2). This result indicated that the T1 copper center had no influence on the rates of catalytic nitrite reduction by these catalysts. Although single enzyme turnover experiments above showed that electrons from both copper sites can be delivered to nitrite, as examined further below it was observed that the rate-limiting step of catalysis is the reaction with nitrite, presumably at the T2 copper center, since this was required for activity. If reduction by ascorbate is much faster than nitrite reduction, then it is possible that turnover could occur with a T2 copper alone. This same phenomenon was observed in native NiRs containing a T2 copper site only, where significant activity was observed using small-molecule reductants, which presumably directly reduced the T2 copper center.77</p><p>The turnover numbers of the azurin variants were smaller than those of native NiRs and comparable with those of other reported synthetic small-molecule models. The rates are 3–4 orders of magnitude slower than those of many native NiRs (1630 min−1,18 8870 min−1,19 90000 min−1,13 and 24960 min−1)11 but similar to those of other model copper NiR's (200 and 270 min−1) in methanol solvent29 or aqueous solution (0.021 min−1 at pH 5.834). There are likely multiple factors leading to higher rates in the native NiRs. A primary factor that is missing in our models is a dedicated proton donor for the leaving oxygen atom as nitrite transitions to NO and OH− products upon reduction. The native enzyme has conserved His and Asp residues that are thought to provide the proton through a hydrogen-bonding network with water.2,8,87,88 Another possible reason for the slower kinetics in our models is the lack of a direct covalent link between the T1 and T2 copper centers, like the one observed in native NiRs.8,46 The link allows for efficient electron transfer from the reductant through the T1 to T2 copper centers upon nitrite binding in native systems.87,88</p><p>The KM values of the azurin variants were similar within error (Table 2) and found to be around 40 mM. This value is 100–1000 times larger than the KM values of native NiRs, which range from 0.5 to 0.034 mM.13,21–23,25 The lower nitrite affinity in the azurin variants can be explained by a number of factors. The designed copper binding site on the azurin surface has two nearby anionic aspartic acid residues that could decrease the affinity for an anionic substrate such as nitrite.89,90 In addition, in this first phase of copper site design, we did not add additional hydrogen bonding capable amino acids. The extra aspartate and histidine residues found in native NiR proteins, in addition to those residues binding the copper ion, are important for nitrite binding.4 Studies with native NiRs demonstrated the strong influence of such residues on KM. For example, mutating the Asp residue in native AxNiR to Ala, Asn, or Glu increased KM values by factors of 14–170.91 Addressing these points in future rounds of design may improve the activity of our designed variants, particularly when the reaction with nitrite is rate limiting.</p><!><p>To further probe the reaction kinetics of nitrite and ascorbate with the azurin model-NiRs, we independently examined the reduction behavior of the azurin variants with ascorbate and then their reoxidation by nitrite. First, rates of reduction were determined by varying ascorbic acid concentrations and monitoring the loss of blue color, or A625, of azurin variants with time. As mentioned above, EPR studies demonstrated the T2 copper was also reduced under these conditions and was slightly ahead of the T1 center (Figure S5 in the Supporting Information). The reduction of Az-NiR3His with 0.5 mM ascorbic acid is shown in Figure 9, with a single-exponential fit to the data. All of the variants demonstrated fits comparable to that shown in Figure 9. Deviations from a perfect fit with a single-exponential function could be attributed to possible trace oxygen in the sample, to any non fully T1 + T2 Cu reconstituted azurin proteins, and to the fact that different protonation states of ascorbate exist in solution,92,93 all of which would result in another component in the reduction curve.</p><p>The reductions of azurin variants by excess ascorbic acid were consistent with a first-order reaction with respect to azurin. The resulting pseudo-first-order rate constants, k′, were plotted and linear fits were found versus the square root of ascorbic acid concentration (Figure 10). The linear fits thus supported a half-order reaction with respect to ascorbic acid, which has been observed in other systems94–96 and could arise from a rate-determining outer-sphere electron transfer to the T1 Cu center via the T2 Cu center or some other surface site. Indeed, the reduction of the T2 Cu center before the T1 Cu center is consistent with the EPR experiment discussed above (Figure S5 in the Supporting Information). The resulting rate constant, k, for the overall 1.5-order reaction between the azurin variant and ascorbic acid was determined from the slope (Table 3).</p><p>The overall rate constant (k) for the reaction of azurin with ascorbic acid (Table 3) was found to be highest for Az-NiR3His (1.05 min/(mM Asc1/2)), followed by Az-NiR (0.57 min/(mM Asc1/2)). The values for Az-PHM (0.26 min/(mM Asc1/2)) and Az-PHM3His (0.32 min/(mM Asc1/2)) were even lower. The rate constants were slightly higher or comparable to those of WT azurin (0.26 min/(mM Asc1/2)).</p><p>For ready comparison, the absolute rates of reduction of 100 µM azurin with 0.5 mM ascorbate were calculated using the experimentally derived rate constants (Table 3). These reduction rates for the variants with a surface T2 copper site were slightly faster than for WT azurin (0.017 (mM Az)/min), with some variation among the variants themselves. The rate of reduction of Az-NiR3His with 0.5 mM ascorbic acid was the fastest at 0.083 (mM Az)/min. The rate was slower for Az-PHM3His (0.021 (mM Az)/min) and Az-PHM (0.024 (mM Az)/min), while the reduction rate for Az-NiR (0.043 (mM Az)/min) was intermediate in the group. Given that the T1 Cu(II) sites of azurin are conserved in all of these variants, the varied rates of reduction could be due to influences of the different T2 copper centers. Overall, reduction rates for the azurin variants by ascorbate was at least 2 orders of magnitude faster than their reoxidation by nitrite (vide infra), particularly at catalytic ascorbate concentrations of 33 mM.</p><p>Following the determination of reduction rates of the T1 copper sites, their rates of reoxidation by nitrite were determined. Reduction of azurin variants with 1 equiv of ascorbic acid and the subsequent recovery of blue color (reoxidation of T1 site) with the addition of NO2− was monitored. Since each ascorbic acid molecule provides two electrons, addition of 1 equiv was required to reduce both the T1 and T2 Cu(II) sites in 1 equiv of azurin variant. Experiments with WT azurin required only 0.5 equiv of ascorbic acid for complete reduction, indicating that only one electron was needed per WT azurin. With the protein stably reduced (Figure 11) in the absence of oxygen, NO2− was added in concentrations ranging from 1 to 150 mM. Wild type azurin was not found to reoxidize with addition of up to 150 mM nitrite and after hours of monitoring. However, all azurin variants fully returned blue with the addition of nitrite. As mentioned above, EPR studies demonstrated that the T2 Cu center also reoxidized, but ahead of the T1 Cu center (Figure S5 in the Supporting Information).</p><p>The resulting rates of reoxidation were determined from plots of A625 with time (Figure 11). Griess tests of the reaction or monitoring the A350 nitrite absorption confirmed that nitrite was consumed as the azurin variants reoxidized. The reoxidation curves were fit to an exponential function (data not shown) supporting a first-order reaction with respect to reduced azurin. The initial reoxidation rates were plotted versus nitrite concentration, and the slope of the linear plots yielded reoxidation rate constants for the reaction of nitrite with reduced azurin (Figure 12 and Table 4).</p><p>The rate constants for reoxidation of the reduced azurin variants by NO2− were found to be much lower than those found above for azurin reduction by ascorbate. Under typical catalytic Michaelis–Menten conditions discussed above, for example 10 mM nitrite and 33 mM ascorbate, the rate of reduction of azurin is over 100 times faster than reoxidation by NO2− (0.61 versus 0.0023 mM Az/min, respectively, for Az-NiR3His). This therefore makes the reaction with NO2− the rate-determining factor in catalysis.</p><p>If reaction with nitrite is rate-limiting, then the catalytic rates might be comparable to the rates of reoxidation. Indeed, Az-NiR3His had the highest rate constant for single turnover reoxidation by NO2− ([2.3(7)] × 10−4 (mM Az)/(min (mM NO2−)) (Table 4), consistent with its highest activity from catalytic Michaelis–Menten parameters (Table 2). Similarly, rates of reoxidation by NO2− for Az-NiR ([1.5(4)] × 10−4 (mM Az)/(min (mM NO2−))) and Az-PHM ([1.3(2)] × 10−4 (mM Az)/(min (mM NO2−))) were slower, consistent with their slower Michaelis–Menten rates. The rate of reoxidation for Az-PHM3His (0.7(1) × 10−4 (mM Az)/(min (mM NO2−))) on the other hand, was the slowest among the variants and inconsistent with its faster Michaelis–Menten parameters. The reason for this anomaly with the Az-PHM3His variant is unclear but could be explained by a slower transfer of electrons from the T1 Cu site to the catalytically necessary T2 Cu center in the single enzyme turnover experiments. Michaelis–Menten parameters were found by observing nitrite loss with time under an excess of reductant, where it was shown that the turnover could be sustained by a T2 copper alone, without the T1 copper necessarily participating. It is possible, therefore, when observing the T1 copper center reoxidize without excess reductant in a single enzyme turnover experiment, that a slower transfer of electrons from the T1 center to the T2 center or nitrite could result in a slower rate.</p><p>To summarize, the recovery of the A625 peak upon the addition of NO2− demonstrated that the reoxidation of T1 Cu(I) is correlated with the reduction of NO2− at the T2 center. First, WT azurin did not reoxidize or react with nitrite under these conditions, demonstrating the necessity of a T2 copper center. This suggests that NO2− reacts with the T2 copper site. Furthermore, these reoxidation experiments demonstrated that electrons are transferred from the T1 site to reduce nitrite, thus lending support to the possibility of electron transfer between the T1 and T2 copper centers during catalysis. As further evidence, we used the Griess test to determine that, with 100 µM Az-NiR3His, there was 185 µM nitrite consumed from an addition of 1 mM nitrite. This corresponds to each of the T1 and T2 copper centers donating an electron to reduce nitrite. Finally, for three of the azurin variants (Az-NiR3His, Az-NiR, and Az-PHM), the rates of reoxidation of the T1 copper center are comparable to the rates of nitrite reduction by the Griess assays (see below), indicating a correlation between the rates of nitrite consumption to the rates of loss of electrons from the T1 copper center.</p><!><p>The rate of reduction of the azurin variants by ascorbate was at least 2 orders of magnitude larger than the rates of reoxidation with nitrite. This suggested that the rate-limiting step in catalysis is not the reduction of the enzyme but rather the reaction with nitrite. If this were true, then the rates of reduction of nitrite and the rates of reoxidation of the azurin variants would be comparable, depending on the rates of electron transfer between the T1 and T2 copper centers. We indeed found the rates of nitrite reduction, obtained from the catalytic Michaelis–Menten assays using the Griess test, to be comparable within error to the rates of reoxidation, obtained with the azurin variants' single turnover reoxidation by nitrite. For example, the experimentally determined rate of nitrite reduction using the Griess assay (0.0038(5) mM NO2−/min at 5 mM NO2− (Figure 8)) was similar to the initial rate of reoxidation for Az-NiR3His (0.0032(9) mM Az/min at 5 mM NO2− (Figure 12a)). The same was true for Az-NiR and Az-PHM variants. For example, the Griess assay rates at 5 mM NO2− for Az-NiR was 0.0016(7) (mM NO2−)/min and the reoxidation rate for same nitrite concentration was 0.0011(2) mM Az/min, while the Griess assay rates at 5 mM NO2− for Az-PHM was 0.0023(7) mM NO2−/min and its reoxidation rate was 0.0018(8) mM Az/min. On the other hand, as discussed above, Az-PHM3His showed slower reoxidation rates than Griess assay rates. For example, the Griess assay rate at 5 mM NO2− for Az-PHM3His was 0.0033(9) (mM NO2−)/min and the reoxidation rate at the same nitrite concentration was about 2 times smaller at 0.0014(4) (mM Az/min. As mentioned above, this could be due to a slower rate of electron transfer from the T1 copper site in Az-PHM3His than in the other variants under the single enzyme turnover conditions. Overall, the similarity between the nitrite reduction rates obtained by the Griess assay and the reoxidation rates suggest that the reaction with nitrite was rate limiting.</p><!><p>The catalytic nitrite reduction behavior of four T2 copper site variants of azurin was characterized. Experiments showed that a T2 copper ion was required for nitrite reduction to NO product and that electrons from the reduced T1 copper center could also be transferred to reduce nitrite. Furthermore, the catalysis was reversible, with the addition of NO(g) resulting in reduction of the T1 copper center and concomitant formation of nitrite. Single turnover experiments showed that the fast step was reduction of the copper centers by ascorbate, thus making the reaction with NO2− the rate-limiting step. This study does not more specifically clarify which part of the NO2− reaction is rate limiting, such as its binding to the active site, electron transfer to bound NO2−, the protonation of the leaving oxide, or product dissociation.</p><p>Michaelis–Menten parameters Vmax and KM were reported for the four T2 copper site variants in azurin. The rates of nitrite reduction of azurin variants lacking a surface T2 copper center were the same as WT azurin and background rates. However, T2 Cu only proteins demonstrated rates that were comparable to the T1 Cu + T2 Cu doubly loaded proteins, suggesting that the T1 copper was not involved in the rate-limiting step. This was consistent with the finding that reduction of Cu(II) is very rapid in comparison to the reaction with nitrite, making it likely that the T2 Cu(II) center can be directly reduced by ascorbate, as was suggested previously for native NiRs lacking a T1 Cu center where significant nitrite reduction rates were achieved with small -molecule reductants.77</p><p>The differences in Michaelis–Menten parameters for the four T2 azurin variants can be attributed to differences in the T2 copper site motifs. In addition, the discovery that the rate-limiting step is the reaction of nitrite with the T2 copper center puts focus on its redesign as a future direction to improve activity. This could therefore include future rounds of design with variants that improve the KM value for nitrite, such as by modulating nearby charged surface residues that interfere with anion binding.89,90 It could also include variants with proton-donating residues, such as those in native NiR, which were shown to be important for triggering electron transfer in addition to proton transfer to nitrite during the formation of the OH−/H2O product.5,87,88,97</p><!><p> ASSOCIATED CONTENT </p><p> Supporting Information </p><p>The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.5b03006.</p><p>Additional data as described in the text (PDF)</p><p>The authors declare no competing financial interest.</p>

PubMed Author Manuscript

Using Triplex-Forming Oligonucleotide Probes for the Reagentless, Electrochemical Detection of Double-Stranded DNA

We report a reagentless, electrochemical sensor for the detection of double-stranded DNA targets that employs triplex-forming oligonucleotides (TFOs) as its recognition element. These sensors are based on redox-tagged TFO probes strongly chemisorbed onto an interrogating gold electrode. Upon the addition of the relevant double-stranded DNA target, the probe forms a rigid triplex structure via reverse Hoogsteen base pairing in the major groove. The formation of the triplex impedes contact between the probe\xe2\x80\x99s redox moiety and the interrogating electrode, thus signaling the presence of the target. We first demonstrated the proof of principle of this approach by using a well-characterized 22-base polypurine TFO sequence that readily detects a synthetic, double-stranded DNA target. We then confirmed the generalizability of our platform with a second probe, a 19-base polypyrimidine TFO sequence that targets a polypurine tract (PPT) sequence conserved in all HIV-1 strains. Both sensors rapidly and specifically detect their double-stranded DNA targets at concentrations as low as ~10 nM and are selective enough to be employed directly in complex sample matrices such as blood serum. Moreover, to demonstrate real-world applicability of this new sensor platform, we have successfully detected unpurified, double-stranded PCR amplicons containing the relevant conserved HIV-1 sequence.

using_triplex-forming_oligonucleotide_probes_for_the_reagentless,_electrochemical_detection_of_doubl

<!>Polypurine TFO Probes and Targets<!>Polypyrimidine TFO Probes and Targets<!>Reagents<!>Sensor Fabrication<!>PCR of HIV-1 RNA<!>Electrochemical Measurements<!>RESULTS<!>DISCUSSION<!>

<p>Many DNA detection methods are based on hybridization and thus require the generation of single-stranded DNA prior to analysis. For example, to properly generate a signal, methods such as DNA microarrays, Southern blotting, and in situ hybridization require denaturation of the double-stranded DNA of interest into single strands, followed by their subsequent renaturation with specific probes. In response, a number of approaches have been developed in which double-stranded targets are detected directly, such as the use of intercalating dyes.1–3 As these approaches lack sequence specificity, however, they are prone to false positives arising due to, for example, spurious amplification.4 The development of assays that are both sequence specific and avoid the cumbersome need to generate single-stranded DNA targets would thus significantly simplify DNA detection.</p><p>Several approaches have been recently proposed for the direct, sequence-specific detection of double-stranded DNA. Most of these employ non-DNA recognition probes and optical read-outs. Such probes include low molecular weight polyamides, which bind to specific minor groove sequences,5,6 or DNA binding proteins, which target specific duplex sequences.7–9 However, while several of these methods have been reduced to practice in the laboratory, their use in "real-world" settings is beset by drawbacks. Protein-based methods, for example, showed promising features of selectivity and sensitivity toward double-stranded targets,10 but they require cumbersome selection processes for the production of the relevant affinity reagents, and the use of proteins as recognition elements reduces the stability of the platform. Polyamide probes are likewise limited by the length of the targeted DNA sequence, which is generally between four to six base pairs, although, with recognition footprints of 10–12 base pairs, head-to-head bis-hairpin minor groove binders (MGBs) are a recent exception.11,12 The utility of these approaches to the direct, sequence-specific detection of double-stranded DNA has thus proven limited to date.</p><p>A potential route toward the direct, sequence-specific detection of double-stranded DNA is the observation that many such sequences support the formation of triplex structures in which a third strand of DNA runs along the major groove of the double helix, where it forms Hoogsteen base pairs.13–15 Specifically, triplex-forming oligonucleotides (TFOs), which are homopurine or homopyrimidine oligonucleotides that bind in the major groove of homopurine–homopyrimidine duplexes,16 exhibit high affinity and specificity, suggesting that they might serve as viable probes for the detection of double-stranded DNA targets without invoking the prior generation of single-stranded DNA. Motivated by these arguments, several groups have reported the use of TFOs in the optical detection of double-stranded DNA.17–20 Here, we expand this approach into a reagentless, electrochemical platform for the direct detection of double-stranded DNA targets.</p><!><p>The modified polypurine 22-base TFO was obtained from Biosearch Technologies (Novato, CA) and employed as the probe DNA without further purification.</p><p>The polypurine probe sequence is as follows: 5′-HS-(CH2)6-CGTTC-GAAGG-AGGAA-GGAGG-GA-(CH2)τNH2-MB-3′.</p><p>The probe is modified at the 5′-end with a mercaptohexanol moiety and at the 3′-end with a methylene blue (MB) redox label. The MB redox moiety conjugation has been performed at the 3′-end of the oligonucleotide via succinimide ester coupling to a 3′-amino modification. The 15 internal bases of this sequence (underlined above) target the duplex obtained by hybridizing (before the injection in the working solution) oligos T1-R and T2-R. The complementary 15-base single-stranded target was also tested. Sequences of these oligonucleotides are reported below:</p><p>T1-R (15 bases, 5′-GGAGG-AAGGA-GGAAG-3′).</p><p>T2-R (15 bases, 5′-CTTCC-TCCTT-CCTCC-3′).</p><p>ss-DNA target complementary to the polypurine TFO probe (15 bases, 5′-CCTCC-TTCCT-CCTTC-3′).</p><!><p>The modified polypyrimidine 19-base TFO was obtained from Biosearch Technologies (Novato, CA) and employed as the probe DNA without further purification. The polypyrimidine probe sequence is as follows:</p><p>5′-HS-(CH2)6-TATTT-TTCTT-TTCCC-CCCT-(CH2)τ-NH2-MB-3′.</p><p>The probe is modified at the 5′-end with a mercaptohexanol moiety and at the 3′-end with a methylene blue (MB) redox label. The 15 internal bases of this sequence (underlined above) target the duplex obtained by hybridizing (before the injection in the working solution) oligos T1-Y and T2-Y or by directly injecting the hairpin self-complementary target (T3-Y) whose sequences are reported below:</p><p>T1-Y, oligopurine target strand (15 bases, 5′-AAAAG-AAAAG-GGGGG-3′).</p><p>T2-Y, oligopyrimidine target strand (15 bases, 5′-CCCCC-CTTTT-CTTTT-3′).</p><p>T3-Y, oligopurine–oligopyrimidine hairpin target (34 bases, 5′-AAAAG-AAAAG-GGGGG-TTTT-CCCCC-CTTTT-CTTTT-3′).</p><p>The polypyrimidine TFO probe was also tested for specificity by challenging with a hairpin target sequence mutated at four positions. The hairpin is still able to form fully complementary double-stranded DNA, but four base pairs (underlined in the sequence below) are not complementary to the TFO probe:</p><p>Mutated oligopurine–oligopyrimidine hairpin for specificity tests (34 bases, 5′-AAACG-CAAAG-GTGGT-TTTT-ACCAC-CTTTG-CGTTT-3′).</p><p>The complementary 15-base single-stranded target forming a double-strand DNA with the TFO was also tested. Sequences of this oligonucleotide are reported below:</p><p>ss-DNA target complementary to the polypyrimidine TFO probe (15 bases, 5′-GGGGG-GAAAA-GAAAA-3′).</p><p>The polypyrimidine TFO was also tested with PCR amplification products. In order to demonstrate the feasibility of detecting the PCR products with this sensor, a synthetic 63 base-pair PCR duplex sequence, which mimics the actual HIV1 double-stranded PCR product, was preliminarily tested (target region is underlined):</p><p>plus strand, 5′-GTAGATCTTAGCCACTTTTTAAAAGAAAA-GGGGGGACTGGAAGGGCTAATTCACTCCCAAAGAA-3′.</p><p>minus strand, 5′-TCTTTGGGAGTGAATTAGCCCTTCCAGTC-CCCCCTTTTCTTTTAAAAAGTGGCTAAGATCTAC-3′.</p><!><p>Reagent grade chemicals, including 6-mercapto-1-hexanol (C6-OH), sulfuric acid, potassium phosphate monobasic and dibasic, and sodium chloride (all from Sigma-Aldrich, St. Louis, MO), were used without further purification.</p><!><p>The sensors were fabricated by depositing the relevant TFO probe on gold rod electrodes (3.0 mm diameter) as previously described.21 Prior to use, the electrodes were cleaned using a series of oxidation and reduction cycles in 0.5 M H2SO4, 0.01 M KCl/0.1 M H2SO4, and 0.05 M H2SO4.21 The thiol-containing oligonucleotide we have employed is supplied as a mixed disulfide of 6-mercaptohexanol in order to minimize the risk of oxidation. The first step in sensor fabrication is the reduction of the DNA probes (100 μM) for 1 h in a solution of 0.4 mM tris(2-carboxyethyl)phosphine hydrochloride (TCEP) in 1 M NaCl/10 mM potassium phosphate, pH 7. This solution was then diluted to 50 nM with 1 M NaCl/10 mM potassium phosphate buffer, pH 7.</p><p>Electrodes (thoroughly rinsed with DI water) were incubated in 250 μL of this DNA probe solution for 60 min. Electrodes were rinsed with DI water and incubated in 2 mM mercaptohexanol in 1 M NaCl/10 mM potassium phosphate buffer (pH 7) for 2 h to displace nonspecifically adsorbed DNA and passivate the remaining electrode area. After thoroughly rinsing with DI water, electrodes were stored in the working buffer solution for 30 min before use. Polypurine TFO probes were tested in 0.1 M pH 7 Tris buffer containing 10 mM MgCl2, while polypyrimidine TFO probes were tested in 0.1 M pH 6.5 Tris buffer solution also containing 10 mM MgCl2. For the PCR experiments, the PCR product solution was diluted 1:5 prior to measurement using a highly acidic, pH 2.3 Tris buffer (0.2 M) containing 11 mM of MgCl2 to achieve a final, postdilution pH of 6.5.</p><!><p>HIV-1 RNA was obtained by extraction from inactivated, intact viral particles (Optiqual HIV-1 RNA Positive Controls, Acrometrix, Benecia, CA) using a QIAamp Viral RNA kit (Qiagen, Valencia, CA). To improve the yield of target DNA for electrochemical detection, we amplified this starting material using a nested, reverse transcriptase-PCR (RT-PCR) protocol. This protocol requires two sets of primers, here termed outer and inner primers. All primers were obtained from IDT (Coralville, IA).</p><p>The outer primers, which flank the targeted duplex region, contain the following sequences:</p><p>Outer_F, 5′-CACAAGAGGAGGAGGAGGTG-3′.</p><p>Outer_R, 5′-TGGCCCTGGTGTGTAGTTCT-3′.</p><p>The sequences for the inner primers are as follows:</p><p>Inner_F, 5′-GTAGATCTTAGCCACTTTTTAAAAG-3′.</p><p>Inner_R, 5′-TCTTTGGGAGTGAATTAGCCCTTCCA-3′.</p><p>The following procedure was adopted for the nested RT-PCR. One microliter of a 1 μg/μL solution of extracted viral RNA was added to a RT-PCR tube containing the following reaction mixture (all from Qiagen, Valencia, CA): 1× OneStep RT-PCR buffer mix [Tris-Cl/KCl/(NH4)2SO4/1.25 mM MgCl2/DTT/pH 8.7]/2.5 mM MgCl2/400 μM of each dNTP/0.6 μM of each outer primer/2 units of OneStep RT-PCR enzyme mix. The reaction tubes were then placed in a standard thermal cycler programmed with the following settings: reverse transcription at 50 °C for 30 min; polymerase activation at 95 °C for 15 min; then 30 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, and elongation at 72 °C for 1 min. Finally, a 10 min elongation step at 72 °C was used to complete the polymerization of any less than full length products. Immediately following RT-PCR, 1 μL of the amplified products was added to a new tube containing a standard PCR mixture (all products from Qiagen, Valencia, CA) composed of the following: 1× Taq polymerase buffer [Tris-Cl/KCl/(NH4)2SO4/1.5 mM MgCl2/pH 8.7]/2.5 mM of MgCl2/400 μM of each dNTP/0.5 μM of each inner primer/2.5 units of HotStar Taq polymerase. The reaction tubes were placed in a standard thermal cycler programmed with the following settings: initial polymerase activation step at 95 °C for 15 min; 35 cycles of denaturation at 94 °C for 30 s, annealing at 55 °C for 30 s, and elongation at 72 °C for 1 min. A final elongation step of 10 min at 72 °C was applied to complete polymerization of any less than full-length products. Visualization of the products was performed using gel analysis with 4–20% TBE polyacrylimide gels and stained with SYBR gold (Invitrogen, Carlsbad, CA).</p><!><p>The sensor response was measured by incubating the electrodes in a solution containing the appropriate target DNA. The sensors were interrogated at different intervals in the same target solution until a stable current peak was obtained (typically after 20 min). The ratio between the stabilized current peak in the presence of target DNA and the current peak in the absence of target DNA gives the measure of the signal suppression caused by the target.</p><p>Between target detection experiments, the electrodes were rinsed with an 8 M guanidine hydrochloride and subsequently interrogated in target-free buffer. This method provides a measure of the extent to which each sensor can be regenerated. All measurements were performed at room temperature using an Autolab potentiostat (EcoChemie, Utrecht, The Netherlands). Square wave voltammetry (SWV) was recorded at 60 Hz, 50 mV amplitude, and with an increment potential of 1 mV over a potential range from −0.1 to −0.45 V in a standard cell with a platinum counter electrode and a Ag/AgCl (3 M NaCl) reference electrode.21</p><!><p>We have employed TFOs as recognition elements for the direct detection of specific double-stranded DNA targets in an electrochemical sensor format analogous to the previously described E-DNA platform.21 Our sensors are comprised of a redox-tagged TFO probe that is strongly chemisorbed to an interrogating gold electrode (Figure 1). In the absence of its double-stranded DNA target, the probe is flexible, allowing an attached methylene blue to approach the electrode and exchange electrons. Upon the addition of its specific double-stranded target, the probe forms a rigid triplex structure that impedes contact between the methylene blue and the interrogating electrode. Thus, the TFO probe serves as an E-DNA sensor23 and maintains the various positive features of this class of devices.21,23–29 For example, we show below that this new class of sensor is reagentless, reusable, and suitable for deployment directly in complex matrices.</p><p>To test the principle of a triplex-forming E-DNA platform, we first fabricated a sensor using a well-characterized 22-base polypurine TFO probe designed to bind a specific double-stranded target DNA via reverse Hoogsteen base pairing in the major groove18 (Figure 1). The sensor was fabricated by the formation of a mixed self-assembled monolayer (SAM) comprised of the TFO probe and 6-mercaptohexanol on the gold electrode, which gives rise to a sharp, well-defined peak at ~260 mV (vs Ag/AgCl), consistent with the formal potential of the methylene blue redox moiety (Figure 1, right). This TFO probe readily binds to its specific double-stranded DNA target (formed by the previous hybridization of oligonucleotides T1-R and T2-R) via triplex formation,18 producing a readily measurable decrease in Faradaic current (Figure 1, right, Figure 2, left). Support for the proposed formation of triplex is provided by the observation that the 65% signal suppression observed in the presence of the double-stranded target is somewhat greater than the 55% suppression obtained when the TFO probe is instead hybridized to its fully complementary, single-stranded DNA target to form a simple duplex (Figure 2, left). This presumably occurs because the greater bulk and charge of triplex DNA is more effective at reducing the efficiency with which the reporting redox tag approaches the electrode. As control experiments we also investigated the effects of adding the two single-stranded oligonucleotides (T1-R) employed in the double-stranded target in isolation. We find that the first of these, T1-R, does not lead to any detectable change in current, even at high concentrations (data not shown). In contrast the addition of the other, T2-R, leads to a nontrivial signal change, presumably because an 11-base portion of this strand is fully complementary to the TFO probe. Despite this, the signal suppression observed in the presence of this strand is much less than that obtained in presence of the duplex target (data not shown), further suggesting that the signal change produced by the duplex target arises due to triplex formation and not due to the dissociation of the duplex target and the consequent formation of a duplex with the probe DNA.</p><p>The polypurine TFO-based sensor rapidly and specifically detects its double-stranded DNA target at concentrations as low as ~10 nM. The ~60 nM dissociation constant observed when the TFO probe is challenged with double-stranded target is, as expected,30,31 about an order of magnitude poorer than the 6 nM dissociation constant produced with fully complementary single-stranded target (Figure 2, left). Triplex formation is relatively rapid, exhibiting an equilibration half-life of ~5 min (Figure 3) that, perhaps because of the greater negative charge and increased hydrodynamic radius of the duplex target, is slightly slower than the rate of simple duplex formation (half-life ~3 min). The sensor is also specific: random sequences of single- and double-stranded DNA oligonucleotides ranging from 15 to 25 bases do not produce any detectable signal change at concentrations as high as 1 mM (data not shown).</p><p>In order to determine the generality of our approach to the sequence-specific detection of double-stranded DNA, we fabricated a second sensor employing a TFO probe targeting the 15 base pair polypurine tract (PPT) sequence conserved in all HIV-1 strains and present twice in HIV-1 proviral DNA.32 In contrast to our first sensor, which is a polypurine tract, the TFO probe in this sensor is a polypyrimidine sequence. We find that this second sensor achieves a dissociation constant of 33 nM (again, only an order of magnitude higher that that obtained for the single-stranded target) and supports the detection of its double-stranded target at concentrations as low as 10 nM (Figure 2, right). In order to rule out a strand-exchange mechanism in which the sensor is instead detecting dissociated, single-stranded target, we also tested this system using a self-complementary hairpin target containing a poly-T linker between the two strands. Because the hairpin target is self-complementary across 15 bases, duplex dissociation is thermodynamically unfavorable and simple duplex formation with the TFO probe is unlikely. The observation that the 37 nM affinity of this hairpin target is effectively indistinguishable from that of a double-stranded target comprised of separate strands (Figure 2, right) further supports the proposed triplex-based detection mechanism.</p><p>Observations regarding the effects of divalent cations on the response of our sensors provide still more support of the triplex-formation mechanism. Specifically, while triple-helix formation requires the presence of magnesium or other divalent cations,17,33 duplex formation does not. Consistent with this, we do not observe any signal change in the presence of even very high concentrations of duplex target if magnesium and other divalent cations are absent (Figure 4, left). In contrast, when the sensor is challenged with single-stranded target (and thus signaling is linked to the formation of a double helix), we observe similarly large signal changes in both the presence (10 mM) and the absence of magnesium (Figure 4, right).</p><p>This new electrochemical sensor architecture is convenient and selective. For example, because the TFO probe is strongly adsorbed to the sensor electrode, the sensor is reagentless and reusable: the original current is completely recovered via a 30 s wash with room temperature 8 M guanidine hydrochloride, allowing us to perform repetitive measurements with a single device (Figure 1, right and Figure 5, left). Likewise, because the signal change upon triplex formation is solely due to a binding-specific change in DNA flexibility, and not simply to the adsorption of charge or mass on the sensor surface, it performs well even when deployed directly in complex, multicomponent samples. For example, the sensor detects its double-stranded DNA target directly in blood serum that is diluted 1:10 with buffer with a signal change similar to that observed in buffer alone (Figure 5, left). The specificity of the TFO sensor was tested by using a synthetic hairpin target mutated at four positions.46 The mutated target is similar to the hairpin target (T3-Y) in that it contains the poly-T linker that holds the complementary strands together, creating a stable duplex DNA molecule. However, while both strands are complementary to one another, four base pairs are mutated such that reverse Hoogsteen interactions with the TFO probe are compromised. When challenging the sensor with high concentrations (200 nM) of this mutated duplex, the sensor does not produce any measurable signal change, thus demonstrating the high specificity of the TFO E-DNA platform (Figure 5, right).</p><p>The above proof-of-principle studies employed fully synthetic targets that would be of little if any interest in a clinical setting. Thus motivated, we have also explored the utility of employing this new sensor architecture in the detection of authentic, unpurified PCR amplicons generated from HIV-1 samples. We amplified a region of the HIV-1 genomic RNA containing the PPT sequence recognized by the polypyrimidine TFO probe employed above. As a first step toward this goal, we determined the ease with which we can use our sensor to quantify synthetic, 63-base pair oligonucleotides equivalent to the PCR amplicon (Figure 6, left). We find that these produce signal changes similar to the large, easily measurable signal changes associated with the short hairpin targets described above. The dissociation constants for these larger targets, however, are slightly poorer than those obtained for the short hairpin target. This is presumably due to steric and/or electrostatic effects associated with the larger targets.</p><p>Moving forward, we find that our new sensor platform readily detects double-stranded DNA produced via reverse transcriptase polymerase chain reaction amplification of HIV-1 genomic RNA (Figure 6, right). To demonstrate this, we employed a conventional nested, amplification protocol to produce a high concentration of amplified products. To inhibit any potential hybridization of the primers–or primer-dimer or other primer-sequence-containing amplification artifacts–to the TFO probes, we placed the sensor's recognition footprint 20 bases internal to the end of the amplicon. Starting from 1 μg of HIV-1 genomic RNA, we obtain about 100 nM of the appropriate, double-stranded DNA amplification product (nucleic acid concentrations obtained using Nanodrop instrument). After incubation of the E-DNA sensors with the PCR amplicon solution, we observe about a 40% drop in the sensor current signal, which is indicative of the presence of the expected target sequence (Figure 6, right). Negative controls performed exclusively with the primers sets and the TFO probe produced no change in signal (data not shown), and a control experiment using nontemplate negative control RT-PCR did not show any significant signal change over the same time period (Figure 6, right).</p><!><p>Among the numerous methods recently proposed for the sequence-specific detection of DNA, E-DNA sensors, the electrochemical analog of optical molecular beacons,22,29,34–38 present several advantages over other optical or electrochemical hybridization detection methods.21 E-DNA sensors are based on the hybridization-induced folding of an electrode-bound, redox-tagged DNA probe. The E-DNA platform is reagentless, electronic (electrochemical), and highly selective (they perform well even when challenged directly in complex, multicomponent samples such as blood serum or soil) and can discriminate between a perfect match and a single mismatch target.21,23 For all these reasons, they appear to be a promising and appealing approach for the sequence-specific detection of DNA and RNA.39,40 But like other hybridization-based methods, traditional E-DNA sensors require the generation of single-stranded DNA targets prior to detection. Here, in contrast, we describe an E-DNA sensor that overcomes this limitation by hybridizing directly to double-stranded DNA targets. And while this approach is limited to the detection of homopurine or homopyrmidine tracks, these triplex-forming sequences are nevertheless common enough that it is straightforward to generate 16–20 base probes with sufficient specificity to target unique sites in human or pathogen genomes.41,42</p><p>Our approach is not the first to employ triplex formation in the detection of double-stranded DNA. Several groups, for example, have previously reported the use of TFOs in the optical detection of double-stranded DNA.16–19 Similarly, two electrochemical methods have been proposed to date where triplex formation is monitored via a direct electrochemical signal from guanine43 or via HPLC separation of the triplex coupled with electrochemical detection.44 While these latter approaches are interesting and achieve remarkable detection limits, their detection principles are limited by several drawbacks: the direct oxidation of guanine appears impractical for realistic applications due to the high overpotential required, the possibility of electrochemical interferences in complex samples (from, for example, exogenous, nontarget DNA), and a high background currents.45 Likewise, the use with HPLC renders the second platform complicated and less suitable for PCR coupling or rapid measurements. For these reasons, the E-DNA sensor described here would appear to offer potentially important advantages for the rapid, electrochemical detection of specific double-stranded DNA sequences.</p><p>Like earlier E-DNA counterparts,21 sensors based on triplex-forming oligonucleotides are label-free, reusable, and selective enough to employ directly in complex sample matrices such as blood serum. Directly measuring double-stranded DNA targets makes these TFO sensors an optimal candidate for use with the PCR amplification process. We have demonstrated this coupling here by performing a nested, reverse-transcriptase PCR of HIV-1 genomic RNA followed by the measurement of the amplified target with the TFO-based E-DNA sensors. As E-DNA sensors have proved suitable for low-cost sensors and portable instrumentations,26 our TFO sensors provide many optimal features that demonstrate the possibility of adopting these sensors in real-world applications. All these attributes suggest that E-DNA sensors may be better suited for clinical applications than the previous, mostly optical methods supporting the direct detection of double-stranded DNA, including approaches based on non-DNA probes.</p><!><p>An E-DNA sensor employing a triplex-forming oligonucleotide (TFO) probe readily detects double-stranded DNA targets. (Left) The sensor consists of a polypurine or polypyrimidine TFO probe modified at its 3′-terminus with a methylene blue redox tag and at its 5′-terminus with a mercaptohexanol moiety for attachment on a gold electrode. (Right) The Faradaic current arising from the flexible TFO probe is significantly reduced in the presence of the double-stranded DNA target, presumably because triplex formation reduces the efficiency with which the terminal redox tag collides with the electrode surface and transfers electrons.</p><p>The polypurine (left) and polypyrimidine (right) TFO based E-DNA sensors respond well to their specific double-stranded DNA targets (triplex formation). As expected, they also respond to fully complementary, single-stranded targets (via the formation of duplex DNA). As previously reported, the affinity of TFO probes for their double-stranded DNA targets is approximately an order of magnitude poorer than that for their single-stranded targets.30,31</p><p>The TFO-based E-DNA sensor is rapid. We observe an equilibration half-life (time required for half of the total signal change to occur) of ~5 min for the double-stranded (triplex-forming) DNA target and ~3 min for a fully complementary single-stranded target. The results shown here were obtained with the polypurine TFO probe.</p><p>The proposed triplex formation mechanism, shown here with our polypyrimidine TFO probe, is supported by experiments in which the concentration of magnesium ions is altered. As triplex formation is dependent on magnesium ion concentration, the absence of magnesium produces no signal, even at high concentrations of double-stranded target (left). A duplex-forming target, in contrast, is readily detected in both the presence and absence of magnesium ions (right). The magnesium ion concentration was 10 mM in both experiments.</p><p>The TFO-based E-DNA sensors are selective, reusable, and specific. Because signal change upon triplex formation is solely due to a DNA binding-specific event, the TFO-based E-DNA sensor performs well, even when deployed directly in complex, multicomponent samples, including 1:10 diluted blood serum (left). The original current is completely recovered via a 30 s rinse in room temperature 8 M guanidine hydrochloride solution, thus allowing us to perform repetitive measurements with a single sensor (left). Moreover, the TFO E-DNA probe displays specificity when challenged with mutated double-stranded DNA (right). Here are shown voltammograms obtained when challenging the polypyrimidine TFO probe with 200 nM of double-stranded target and 200 nM of mutated target (right).</p><p>TFO sensor responds to unpurified PCR amplicons containing a triplex-forming element in the HIV-1 genome. (left) Specifically, test results using a synthetic, duplex oligonucleotide equivalent to the relevant HIV-1 sequence binds to the probe, albeit with a poorer dissociation constant and somewhat poorer signaling than those seen with a short hairpin target containing the same recognition footprint. These effects are presumably due to the larger size of the amplicon. (right) The sensor also responds rapidly and robustly to authentic PCR amplified target molecules (at 100 nM), while producing no signal change in response to negative control PCR samples. Polyacrylimide gel (right, inset) containing amplicons from HIV-specific PCR, a negative control (no template added to the PCR), and synthetic HIV duplex show the specificity of the PCR reaction. The size of our HIV target DNA is 63 bp.</p>

PubMed Author Manuscript

Epigenetically Enhanced Photodynamic Therapy (ePDT) is Superior to Conventional Photodynamic Therapy for Inducing Apoptosis in Cutaneous T-Cell Lymphoma

Conventional photodynamic therapy with aminolevulinate (ALA-PDT) selectively induces apoptosis in diseased cells and is highly effective for treating actinic keratoses. However, similar results are achieved only in a subset of patients with cutaneous T-cell lymphoma (CTCL). Our previous work shows that the apoptotic resistance of CTCL correlates with low expression of death receptors like FAS, and that methotrexate upregulates FAS by inhibiting the methylation of its promoter, acting as an epigenetic derepressor that restores the susceptibility of FAS-low CTCL to caspase 8-mediated apoptosis. Here, we demonstrate that methotrexate increases the response of CTCL to ALA-PDT, a concept we refer to as epigenetically enhanced PDT (ePDT). Multiple CTCL cell lines were subjected to conventional PDT versus ePDT. Apoptotic biomarkers were analyzed in situ with multispectral imaging analysis of immunostained cells, a method that is quantitative and 5\xc3\x97 more sensitive than standard immunohistology for antigen detection. Compared to conventional PDT or methotrexate alone, ePDT led to significantly greater cell death in all CTCL cell lines tested by inducing greater activation of caspase 8-mediated extrinsic apoptosis. Upregulation of FAS and/or TRAIL pathway components was observed in different CTCL cell lines. These findings provide a rationale for clinical trials of ePDT for CTCL.

epigenetically_enhanced_photodynamic_therapy_(epdt)_is_superior_to_conventional_photodynamic_therapy

INTRODUCTION<!>Cell Lines<!>Conventional PDT and ePDT<!>Western blotting<!>Immunohistochemistry (IHC)<!>Multispectral imaging analysis (MIA)<!>Quantitation of PpIX<!>Statistical analysis<!>The response of CTCL cell lines to conventional PDT is positively correlated with baseline expression of FAS<!>ePDT upregulates both FAS/CD95 and FASL in CTCL<!>ePDT increases TRAIL-R1(DR4)/TRAIL in CTCL<!>ePDT induces greater apoptosis than conventional PDT in CTCL<!>ePDT enhances predominantly the extrinsic apoptotic signaling in CTCL<!>DISCUSSION

<p>Photodynamic therapy (PDT) utilizes the accumulation of a photosensitizer in diseased tissue and subsequent illumination with visible light. The resultant photochemical reaction leads to formation of reactive oxygen species (ROS), which destroy cells by inducing apoptosis and/or necrosis (1–3). In PDT with topical aminolevulinic acid (ALA), the targeted cells metabolize the porphyrin precursor ALA into the endogenous photosensitizer protoporphyrin IX (PpIX), rendering this treatment highly selective and virtually non-toxic (4). Conventional ALA-PDT represents a first-line therapy for actinic keratoses (AKs) and is increasingly used as an alternative therapeutic option for a wide range of skin disorders (5–7). Nevertheless, the excellent outcomes achieved in patients with AKs are not easily reproduced in other indications. This is exemplified by mycosis fungoides (MF), the most common form of cutaneous T-cell lymphoma (CTCL). Long-term complete responses to ALA-PDT were observed in up to 50% of patients with patch/plaque MF in certain studies as well as in some cases of recalcitrant and/or tumor-stage disease (8–14). Such results are promising, however, the lack of success in many other cases of MF remains unexplained.</p><p>Epigenetic gene alterations, i.e. defective gene expression due to modifications other than those of DNA sequence, are recognized as crucial to the pathogenesis of disease (15). In CTCL, defects involving death receptor/ligand pairs such FAS/FASL, TNF-R1/TNFα and TRAIL-R2(DR5)/TNF-related apoptosis-inducing ligand (TRAIL) are correlated with resistance to apoptosis (16–18). Our previous work shows the interindividual heterogeneity of epigenetic alterations in CTCL, reflected in variable expression of apoptotic proteins such as FAS in lesional skin of MF patients (16). We have also demonstrated that methotrexate (MTX) acts as an epigenetic regulator by inhibiting gene promoter methylation, thus upregulating proteins that are essential to T-cell apoptosis and restoring the sensitivity of CTCL to apoptotic cell death (19,20). Moreover, we established that multispectral imaging analysis (MIA) is 5× more sensitive in the detection of biomarkers in skin lesions as compared to standard light microscopy, making it ideally suited to preselect patients for targeted MTX therapy and to monitor protein changes in response to treatment (21). Therefore, we hypothesize that the limited response of CTCL to ALA-PDT might be due to epigenetically suppressed expression of death receptors and/or their ligands and that MTX increases the efficacy of PDT by enhancing extrinsic apoptosis (Figure 1). The data presented here support our theory, showing that the epigenetic action of MTX in concert with PDT allowed significantly greater induction of extrinsic apoptotic cell death as compared to conventional PDT. We propose to introduce epigenetically enhanced PDT (ePDT) as a novel treatment concept for disorders which require the eradication of malignant and/or activated T-lymphocytes.</p><!><p>The human CTCL cell lines MyLa, Hut78, HH, SZ4 and SeAx were acquired as described previously (19,20). The cells were cultured in RPMI 1640 medium with 2mM L-glutamine, 10% fetal bovine serum, and 1mM sodium pyruvate (all cell lines) as well as 10mM 4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acid was added to the media (HH and SZ4 only).</p><!><p>For conventional PDT, cells were incubated for 6 hours with 1mM 5-aminolevulinic acid hydrochloride (Sigma-Aldrich, Saint Louis, MO), followed by exposure to 630 nm light at 3.22 J/cm2 using the Luzchem Expo Panel light source equipped with a red filter (Luzchem Research, Ottawa, Canada) and collection at various time points post irradiation. For ePDT, cells were treated with methotrexate hydrate (Sigma-Aldrich) for 48 hours and subjected to ALA-PDT as described above. ALA only, light only, MTX only, and untreated controls were included in all studies. Experiments involving ALA were performed in the dark to avoid photobleaching.</p><!><p>Cells were washed with ice-cold phosphate-buffered saline (PBS) and lysed with radioimmunoprecipitation assay buffer. DC Protein Assay (Bio-Rad, Hercules, CA) was used to establish the amount of protein per sample. Subsequently, 30µg of protein was subjected to sodium dodecyl sulphate-polyacrylamide gel electrophoresis and transferred onto a nitrocellulose membrane. Blots were exposed to primary and secondary antibodies (anti-PARP polyclonal Ab, anti-FAS mAb clone C18C12, anti-rabbit IgG horseradish peroxidase-linked Ab, all from Cell Signaling, Danvers, MA) followed by chemiluminescent detection.</p><!><p>Cells were washed with PBS, centrifuged and re-suspended in PBS to create cytopreparations on glass slides, which were exposed to the following primary antibodies: anti-FAS/CD95 antibody clone DX2 (Enzo Life Sciences, Farmingdale, NY), anti-FASL antibody clone G247-4 (BD Pharmingen San Jose, CA), anti-cleaved caspase-3 antibody clone 5A1E (Cell Signaling), anti-DR4 polyclonal antibody sc-6823 (Santa Cruz Biotechnology, Dallas, TX), anti-TRAIL antibody clone III6F (Enzo Life Sciences), anti-cleaved caspase-8 antibody clone 12F5 (Enzo Life Sciences), and anti-cleaved caspase-9 antibody clone Asp330 (Cell Signaling). Proteins were visualized using the MACH4 Universal HRP-Polymer detection system (Biocare Medical, Concord, CA) and 3,3-diaminobenzidine (DAB). All slides were counterstained with methylene blue (MB).</p><!><p>The quantitative analysis of protein expression was carried out using the Nuance system (Perkin-Elmer, Waltham, MA). Multispectral 8-bit image cubes of relevant areas on each slide were acquired in brightfield mode at 20nm intervals from 420 to 720nm, using a 20× objective lens and 1×1 binning. A spectral library containing the characteristic wavelength emission curves of MB and DAB was created by sampling pure spectra from slides stained with single dyes, i.e. MB only and DAB only. Blue and red pseudo-colors were assigned to visualize MB and DAB, respectively. The image cubes, initially shown as pseudo-color composites of both blue and red, were unmixed into individual components by the Nuance 3.0.0 software version. For quantitation of protein expression, 100–300 cells per slide were manually designated as regions of interest (ROIs). Using the spectral library as reference, the system measured the amount of target signal (intensity of red pseudo-color) within each ROI, automatically converting spectral data into optical density (OD) units. Final data are displayed as average OD/cell. All histograms shown in Figures 3–7 are based on the quantitative assessment of protein expression by MIA. Exemplary MIA-generated high-power images of protein expression and quantitative analysis are shown in Figure 2.</p><!><p>cells were cultured and incubated with 5-ALA as described above. After 6 hours of incubation time, the cells were centrifuged, re-suspended in PBS and placed on glass slides for immediate capture of PpIX fluorescence with the Nuance microscope. Multispectral images cubes were acquired in fluorescence mode using a 405 nm light source, 20× objective lens and 1×1 binning. The characteristic brick-red fluorescence of PpIX was captured to build the PpIX spectral library, which was subsequently used to quantitate intracellular PpIX as outlined above. Final data were displayed as average signal/cell.</p><!><p>Statistical analysis was performed by using Student's t-test. P-values < 0.01 were considered statistically significant.</p><!><p>The CTCL cell lines used throughout our studies are characterized by known FAS gene abnormalities, resulting in a wide range of FAS protein expression with complete absence of FAS gene in SeAx cells (16, 22). Conventional ALA-PDT caused caspase 3 cleavage in all cell lines, peaking at 24–48 hours (Figure 3A). However, the FAS-high prototypes MyLa and Hut78 demonstrated significantly greater cleaved caspase 3 levels than their FAS-low (HH, SZ4) and FAS-null (SeAx) counterparts. Figure 3B,C shows that there was an approximately 10-fold increase in intracellular PpIX levels following incubation with ALA. Notably, PpIX accumulation following ALA incubation did not differ significantly among cell lines, thus excluding inadequate ALA-uptake and/or insufficient PpIX formation as a possible cause of differences in response to ALA-PDT. These data mirror the variable response of CTCL to PDT encountered in the clinical setting, and indicate a positive correlation between baseline FAS levels and ALA-PDT-induced apoptosis.</p><!><p>Conventional PDT did not significantly affect FAS expression in any of the cell lines tested (Figure 4A). In contrast, both MTX alone as well as ePDT resulted in a 3- and 4-fold increase of FAS in FAS-low prototypes HH and SZ4, respectively. This effect was achieved at very low doses of MTX, as demonstrated in Figure 4B. The baseline levels of FASL were relatively low in all cell lines, and remained unchanged after MTX treatment (Figure 5). Conventional PDT led to a 2–3-fold FASL upregulation of FASL across all cell lines as compared to untreated controls. Moreover, ePDT resulted in significantly greater FASL increase as compared to ALA-PDT in MyLa, Hut78 and SZ4 cells. Taken together, these data demonstrate that unlike conventional PDT or MTX alone, ePDT restores FAS expression in FAS-low cell lines while simultaneously upregulating FASL.</p><!><p>The baseline levels of DR4 were low in all CTCL cell lines, and were significantly increased by both components of ePDT (predominantly by MTX and to a lesser degree by PDT) in 4 of the cell lines tested (Figure 6A). In Hut78 cells, the increase in DR4 was observed following MTX but not conventional PDT. Consequently, ePDT resulted in greater DR4 increase as compared to conventional PDT in all cell lines (Figure 6A). Figure 6B shows baseline TRAIL levels as measured by MIA, revealing high-medium levels in SZ4 and MyLa cells, and very low levels in the remaining three cell lines. MTX alone resulted in upregulation of TRAIL only in MyLa cells, while conventional PDT led to TRAIL increase only in HH cells. Whereas ePDT led to TRAIL upregulation in both MyLa and HH cells, TRAIL expression was significantly higher post ePDT versus conventional PDT only in MyLa cells. These findings underscore the involvement of multiple death receptor/ligand pairs in the response of CTCL to ePDT. Table 1 summarizes the changes in FAS/FASL and DR4/TRAIL expression in response to the individual components of ePDT, namely MTX and PDT.</p><!><p>Compared to conventional PDT, ePDT resulted in greater apoptosis in all cell lines as evidenced by caspase 3 cleavage post treatment (Figure 7A). The most pronounced increase in apoptotic cell death following ePDT versus conventional PDT was observed in the FAS low cell lines HH and SZ4. Figures 7B and 7C depict the effects of ePDT versus conventional PDT in HH cells in greater detail, showing that ePDT with low-dose MTX elicited significantly greater cleavage of caspase 3 as well as PARP.</p><!><p>As demonstrated in Figure 8A, ePDT produced significantly greater caspase 8 activation as compared to MTX alone as well as to conventional PDT in all cell lines. These data suggest that the ePDT-induced increases in FAS/FASL and DR4/TRAIL summarized in Table 1 are positively correlated with the capability of malignant T-cells to undergo extrinsic apoptosis. Figure 8B shows that ePDT also resulted in greater caspase 9 activation in three cell lines, further enhancing the overall greater efficacy of this treatment as compared to PDT or MTX alone.</p><!><p>There are no curative treatment options for MF and its leukemic variant, Sézary syndrome (SS), both of which represent > 75% of all CTCLs. Recently developed therapies include antibody-drug conjugates which target specific markers expressed primarily on the surface of lymphoma cells (23, 24). However, the indolent nature of early-stage MF precludes aggressive systemic approaches. Consequently, there is an urgent need for effective yet non-toxic treatment strategies for refractory early-stage CTCL.</p><p>Unpredictable clinical outcomes despite optimized drug and light delivery continue to represent a major drawback of ALA-PDT. The current trend is therefore its combined use with other therapies. Pretreatment of basal cell carcinomas with retinoids was found to enhance the efficacy of ALA-PDT (25), but equivalent trials in patients with CTCL are lacking. In the CTCL cell line Hut78, a combination of 8-methoxy-psoralen and ALA followed by UVA irradiation was superior to either conventional PUVA or PDT alone (26). However, UVA is carcinogenic and penetrates skin far less deeply (1mm) than red light (up to 5mm) (4). This may be the reason why MF patients who receive PUVA, UVA1 or narrow band UVB therapy require at least 15–20 treatments, while PDT usually achieves comparable results with far fewer sessions (13).</p><p>Very few studies explore the mechanisms of ALA-PDT-induced apoptosis in activated and/or malignant T lymphocytes. Involvement of death receptor/ligand pairs in PDT-induced apoptosis has been demonstrated in the setting of PDT with Pc4, Calphostin-C, hypocrellin and hypericin (27–30), but not in that of ALA-PDT and CTCL. ALA-PDT induces apoptosis predominantly via the intrinsic pathway due to preferential localization of PpIX in the mitochondria, however, exogenous ALA administration also leads to PpIX accumulation in cellular membranes and in the cytosol (31–33). In the oral cancer cell line Ca9-22, ALA-PDT activated both caspase 8 and caspase 9 via the NF-κB/JNK pathway, whereas p38 MAPK and JNKs were involved in the initiation of ALA-PDT-mediated apoptosis in non-small cell lung carcinoma cells (34, 35). Interestingly, the first report of PDT-induced apoptosis in a malignant T-cell line featured ALA-PDT (36). Subsequent studies on leukemia/lymphoma cells with photosensitizers other than ALA confirmed the susceptibility of T cells to PDT-induced apoptosis (2, 37–41). Notably, the majority of research involving the response of malignant or activated T-cells to PDT has been conducted on Jurkat cells, a cell line derived from T-cell acute lymphoblastic leukemia, which is clinically very different from CTCL and therefore not at all representative of MF/SS with regard to its tumor biology (36, 42). In contrast, the spectrum of human CTCL cell lines studied here encompasses a wide range of phenotypes in terms of death receptor/ligand expression, allowing us to study the associations between varying levels of these proteins in CTCL and the efficacy of MTX, PDT and ePDT.</p><p>The proposed mechanisms underlying the efficacy of ePDT are depicted schematically in Figure 1. Figure 1A shows a lymphoma cell with downregulated or absent expression of death receptor(s) such as FAS. The formation of ROS following conventional ALA-PDT induces apoptosis via the intrinsic caspase 9-mediated pathway. ROS activate numerous signaling pathways, such as the JNK/c-Jun pathway, which is known to upregulate some death receptor ligands (43). As summarized in Figure 1B, cells with fully functioning death receptor/ligand pairs or cells with MTX-mediated derepression of death receptors/ligands are more likely to undergo apoptosis after PDT for the following reasons: 1) death receptor/ligand binding boosts an additional cell death mechanism, the extrinsic apoptotic pathway (19,20), 2) increased caspase 8 cleavage leads to Bid truncation, which in turn boosts the intrinsic apoptotic pathway (44), 3) MTX leads to increased production of ROS (45), further contributing to caspase 9 activation.</p><p>The data presented in Figures 3–8 support the above concept, demonstrating significantly greater activation of apoptosis primarily via the extrinsic pathway as well as by increased cleaved caspase 3 and PARP cleavage products in ePDT as compared to PDT or MTX alone. The expression of FAS in CTCL was unaffected by PDT. These findings are in line with a previous report that showed no change in FAS levels in Jurkat cells after PDT (41). Our observation that ALA-PDT increases FASL in CTCL constitutes another novel finding in this setting, albeit one that was anticipated, since ROS such as H2O2 and singlet oxygen are major known upregulators of FASL in activated T-lymphocytes (46, 47). Since MTX increased FAS levels in FAS-low cell lines, ePDT created the ideal prerequisite for FAS/FASL engagement irrespective of baseline protein levels, correlating with increased activation of the extrinsic pathway. ePDT surpassed the efficacy of conventional PDT even in FAS-high cell lines, and resulted in increased cleaved caspase 8 levels in FAS-null SeAx cells, indicating the involvement of other death receptor/ligand pairs as well as additional mechanisms in the response of CTCL to ePDT. All five cell lines studied here showed a low baseline level of DR4. MyLa and SZ4 were TRAIL-high, whereas the remaining cell lines exhibited low TRAIL baseline levels. MTX increased DR4 in all cell lines and upregulated TRAIL in MyLa cells, while PDT led to increased DR4 levels in MyLa, HH, SZ4 and SeAx and increased TRAIL in HH cells. Taken together, the above data support the hypothesis that multiple death receptor/ligand pairs contribute to the enhanced extrinsic apoptosis observed in ePDT. Future studies are planned to investigate whether these effects are due to decreased methylation of DR4 and/or TRAIL promoter by MTX. MTX also exerts effects that are independent of its function as an epigenetic regulator, e.g. S-phase inhibition of mitosis. In addition, it has been shown that MTX promotes cellular differentiation and increases intracellular PpIX in epithelial cancers by upregulation of coproporphyrinogen III (CPIII) oxidase, the enzyme that catalyzes the oxidative decarboxylation of CPIII to PpIX (48, 49). Both ALA-PDT and MTX are well-known and approved treatment modalities. The concept of ePDT could therefore be tested in clinical trials without further need for in vitro- or animal studies. In addition to oral dosing of MTX, this work might also prompt the formulation of MTX as a topical drug. MIA-based quantitative IHC lends itself as an excellent tool to a) pre-select the patients who would most likely benefit from ePDT and b) monitor the effects of ePDT by objective measurements of pertinent proteins directly in skin biopsies (19, 23, 50, 51). By employing ePDT in this way, the success of traditional ALA-PDT seen in AKs could be potentially replicated in CTCL. ePDT would also represent a non-carcinogenic alternative to radiation therapy in cases of facial and/or limited extent CTCL. Lastly, the applications of this novel modality extend well beyond CTCL: 1) we expect ePDT to exert therapeutic effects in cutaneous disorders in which activated T-cells play a central role, such as psoriasis, and 2) epigenetic dysregulation is a hallmark of other skin cancers as well, all of which represent targets for future studies and potential candidates for ePDT.</p>

PubMed Author Manuscript

The Quasi-Static Self Quenching of Trp-X and X-Trp Dipeptides in Water: Ultrafast Fluorescence Decay

Time-resolved fluorescence decay profiles of N-acetyl-L-tryptophan-amide (NATA) and tryptophan (Trp) dipeptides of the form Trp-X and X-Trp, where X is another aminoacyl residue, have been investigated using an ultraviolet upconversion spectrophoto fluorometer with time resolution better than 350 fs, together with a time correlated single photon counting apparatus on the 100ps to 20ns time scale. We analyzed the set of fluorescence decay profiles at multiple wavelengths using the global analysis technique. Nanosecond fluorescence transients for Trp dipeptides all show multiexponential decay, while NATA exhibits a monoexponential decay near 3 ns independent of pH. In the first 100 ps, a time constant for the water \xe2\x80\x9cbulk relaxation\xe2\x80\x9d around Trp, NATA and Trp dipeptides is seen near 1-2 ps, with an associated preexponential amplitude that is positive or negative depending on emission wavelength, as expected for a population-conserving spectral shift. The initial brightness (sub-ps) we measure for all these dipeptides is less than that of NATA, implying even faster (<200fs) intra-molecular (quasi) static quenching occurs within them. A new, third, ultrafast decay, bearing an exponential time constant of 20-30 ps with positive amplitude, has been found in many of these dipeptides. We believe it verifies our previous predictions of dipeptide QSSQ (\xe2\x80\x9cquasi static self quenching\xe2\x80\x9d) \xe2\x80\x93the loss of quantum yield to sub-100ps decay process (Chen et al., Biochemistry, 1991, 30, 5184). Most important, this term is found in proteins as well (J.A.C.S., 2006, 128, 1214; Biophysical Journal 2008, 94, 546; 2009, 96, 46a), suggesting an ultrafast quenching mechanism must be common to both.

the_quasi-static_self_quenching_of_trp-x_and_x-trp_dipeptides_in_water:_ultrafast_fluorescence_decay

I. Introduction<!>Upconversion spectrophotofluorometer<!>TCSPC apparatus<!>Samples<!>III. Results<!>IV. DISCUSSION<!>V. Conclusions

<p>Understanding peptide dynamics in aqueous solution is critical. Their functions depend strongly on their structures (and structural change) during biological processes such as protein folding, protein hydration, and protein-peptide recognition.1-4 In order to gain insight into the relationships among the solution environment, conformation changes and peptide function, Tryptophan (Trp) and other indole-containing compounds have been used as probes. Trp reveals structure and dynamics in its emission maximum and width, quantum yield, lifetime and anisotropy. Among amino acid residues, Trp possesses the highest UV extinction coefficient and highest quantum yield of emission.5-8 In a polar environment, Trp fluorescence is emitted from the singlet 1La state because ultrafast (<100 fs) internal conversion converts a mixture with 1Lb to 1La.9. Trp is very sensitive to the polarity and dynamics of the immediate environment; spectral energy and width depend strongly on exposure to water (and/or the polarity of the electrostatic environment inside protein).10,11 The ability to probe time resolved fluorescence of tryptophan, which, if not already present, can often be incorporated in a peptide without significant structural changes, opens a window into peptide/protein dynamics inaccessible to other biophysical techniques.12,13 Recently, interest in Trp has grown, as tunable ultrafast Ti: sapphire lasers have made it possible to monitor the molecular vibrational modes and femtosecond solvent relaxation. 14-16</p><p>It is well known that many peptides and proteins containing even a single Trp yield multiexponential fluorescence decays linked to different Decay Associated Spectra (DAS).17-19 The origin of these DAS is still controversial. Basically, three explanations have been proposed for complex fluorescence decays. First, these phenomena are most easily explained on the basis of ground state heterogeneity; e. g. the presence of several conformers, each with different fluorescence lifetimes.20,21 NMR-derived rotamer population data has thus been used to predict amplitudes.21,22 Further, single-Trp proteins in crystals provided angle-dependent preexponential factors that support a rotamer concept.23 Second, in addition to frank rotamers of the Trp side chain, microconformational states of peptides or proteins can create different local environments of the indole ring that constitute a source of ground-state heterogeneity,24 and third, relaxation of the peptide/protein matrix (or solvent water) surrounding the indole ring during the lifetime of the excited state will also produce a complex decay.25-27 The latter interpretation, driven by the preponderance in literature of shortlived bluer and longlived redder DAS, requires generalized solvent relaxation on the nanosecond timescale (in response to the increased dipole moment of the excited state).28 In this model, decay curves change character across the emission surface; rapid decay at the blue end of the fluorescence spectrum is replaced by a central region of simple exponential decay, and fast rising behavior is seen at the red end of the spectrum.14,15 A negative preexponential amplitude is thus generated at the red tail of spectrum that is always associated with a short lifetime (i.e. "short negative DAS"). The presence or absence of this negative term has become an important signature for relaxation, although mixtures of relaxation and heterogeneity processes might obscure it. The DAS recovered in the relaxation model should generally show a trend of longer decay times for the longer wavelength components.29</p><p>Most Trp fluorescence studies in peptides/proteins have been made with photomultipliers providing sub-nanosecond time resolution.30-32 In such previous work, analysis of lifetime and quantum yield data for Trp-X dipeptides provided evidence for quasi-static self quenching (QSSQ).33 Very short lifetimes (e.g., under 100ps) can be mistaken for scattered light or obscured by noise on the nanosecond scale, so they would be underrepresented or lost. Their presence was inferred from quantum yield defects.</p><p>In this work, extended time resolved fluorescence transients, including fluorescence decay kinetics and DAS, are collected from both an upconversion spectrophotofluorometer and a time correlated single photon counting (TCSPC) apparatus coupled to fs and ps laser sources. The full dynamics are dissected to reveal contributions from both the heterogeneous environs of the peptides and the evolving solvent shell. Recently, Larsen et al.34 also found a 16 ps fluorescence component for Trp in a small 22-mer peptide, and they proposed that this lifetime originated from a rotamer of Trp, which reinforces the conclusions of this paper.</p><!><p>The experimental setup has been partly described elsewhere.14 Briefly, a mode-locked Ti: sapphire laser was used to generate a 400 mW pulse train with a typical pulse duration of 120 fs at a repetition rate of 80MHz. After seeding a Ti:sapphire regenerative amplifier, the amplified infrared pulses at 885nm each had an energy of ~160 J and an autocorrelation pulse width of 350 fs at a repetition rate of 5 kHz. They were frequency-doubled and tripled to generate ultraviolet excitation (295 nm) with an average power of 30 mW. The UV beam (pump pulse) was then separated from the infrared beam (885nm, probe pulse), and blue beam (doubled) by two dichroic mirrors, and the average power for the excitation of the sample was carefully attenuated to less than 1mW to avoid photodegradation, saturation, hole burning, and other undesirable effects. The sample was placed into "sandwich" cells with a path length of 1 mm that were, in turn, mounted in a delrin disk for attachment to a chopper motor and continuously spun so the tangential velocity of sample through the beam was greater than 1 m/s. The residual infrared pulse was retroreflected from a hollow cube corner on a computer-controlled stage, and this variably delayed pulse was used as a sampling probe pulse for the upconversion process. The fluorescence emission was collected by a pair of parabolic mirrors and focused into a BBO crystal, and the upconversion signal was produced from 232nm to 280 nm via type I sum frequency generation with the sampling pulse in the crystal. To reject the strong background signals (infrared laser, remnant UV and fluorescence) accompanying the upconverted signal, a non-collinear configuration was set up between infrared laser and fluorescence. The polarization of the excitation beam (295 nm) was chosen by a motor-controlled zero-order half-wave plate. Since polarizations of laser pulses and fluorescence were strictly determined by the orientation of nonlinear crystals,35 no extra linear polarizer was necessary in our experiments (in the terms of normal fluorometers, G=1.). The upconverted fluorescence was detected by a monochromator with a bandwidth of 0.5nm and a solar blind photomultiplier tube. Amplified signals were discriminated and then recorded by a Ortec 994 gated single photon counter The "lamp" (or instrument response) function under the current nonlinear geometry was around 400 fs as determined from the cross-correlation between UV-generated Raman scattering at 328 nm in water and the infrared laser. Instrument calibration with the linear fluorophore p-terphenyl yielded an initial anisotropy of 0.40+/−.01 and a single rotational correlation time of 41ps in cyclohexane (both from Sigma/Aldrich).</p><!><p>A tunable Spectra-Physics cavity dumped dye laser (3520) was synchronously pumped by a mode-locked DPSS (Vanguard) green laser, and dye laser output was doubled into the ultraviolet. For the present work, rhodamine 6G was used as the laser dye, and vertical excitation was employed at 295nm with pulses having a FWHM of 2ps. The fluorescence was recorded from 310nm to 405 nm through a magic angle polarizer by use of a cooled microchannel plate photomultiplier. Fluorescence decays for the samples and reference were measured to 1 × 104 counts in the peak channel. The instrumental response time was about 90ps, so the measurement of lifetimes of ~60ps and greater can be made in this instrument. A JYH10 monochromator with 1.5mm slit width was used to select the emission. Melatonin in water was used as a standard, for which separate experiments showed monoexponential decay and a lifetime of 5.4ns. Lifetimes were obtained by fitting the decay data to a multiple-exponential model, according to the weighted, least-squares method. Goodness of fit was assessed with the χ2R function.36 The fits for which data are given in this work yielded values of 1.01-1.2. For decay-associated spectra (DAS), time-resolved data were obtained at every 5 nm over the emission band from 310 nm to 405 nm. Alternating the sample with the scatterer, stepping of the emission monochromator, data collection, and transfer of data from the multichannel buffer to the computer was done automatically. In the analysis of the multiple curves obtained for a DAS, they were all fit to the same multiexponential model:37 I(λ, t) = ∑αi(λ) · e−t/τi.</p><p>Steady-state absorption and fluorescence spectra were characterized with a diode array spectrophotometer (HP 8452A) and Fluorolog-3 spectrophotofluorometer (SPEX), respectively.</p><!><p>N-acetyl-L-tryptophan-amide (NATA), L-Trp, and the Trp dipeptides (Trp-Leu, Trp-Ala, Trp-Phe, Trp-Gly, Leu-Trp, Val-Trp, Gly-Trp) were purchased from Sigma-Aldrich chemical Co. They were used without further purification. The solutions were prepared in a 10 mM sodium acetate buffer at pH 5.2 or sodium borate buffer at pH 9.3 using distilled deionized water. A typical concentration of the dipeptides in water for upconversion was 1 mM. All samples were made at the room temperature. A fresh sample solution was prepared for each time-resolved measurement.</p><!><p>All Trp compounds studied here had a typical indole absorption band at a peak around 280nm (data not shown). Figures 1A, 1B show the steady state fluorescence spectra (295nm excitation) of NATA and Trp dipeptides (Trp-Leu, Leu-Trp) in water at pH 5.2 and 9.3, respectively. Several features should be noted: First, the spectrum of NATA (~355 nm) does not depend on pH. Second, the dipeptides show red (Leu-Trp) or blue (Trp-Leu) shifts relative to NATA (355nm) when pH is 5.2. For example, the peak for Trp-Leu at pH 5.2 is shifted by 11 nm to the blue compared with that of Leu-Trp. However, when pH is 9.3, the spectrum maximum matches that of NATA, indicating the identity of X does not govern spectral position when pH is high (anionic form). This is in accord with prior observations 33 and recent simulations.45 A series of upconversion transients for Leu-Trp (taken with magic angle excitation geometry) at various emission wavelengths is shown in Fig. 2. Clearly, there is a rapid process (corresponding to a 1-2ps exponential) that is depleting the blue side of the spectrum and creating a corresponding rising term on the red edge. This 1-2ps term is seen in all picosecond Trp studies and is ascribable to the dipolar relaxation of bulk water in response to the larger dipole of Trp in the excited (1LA) singlet state.14-16 This term provides smaller amplitude contributions near the center of the spectrum (where its sign changes).</p><p>The upconverted transient for the central (~355nm) portion of the spectrum for Trp-Leu is shown in Fig. 3. It is clear that a biexponential fit is inferior to one including not only the ns mean lifetime (gleaned from TCSPC) and the 1-2 ps water relaxation term, but also a positive exponential of 20-30ps. To ascertain the origin of this term, one must collect a full decay surface. As we have previously shown for the proteins monellin and IIAGlc, the shape of the spectrum composed of preexponential amplitudes (DAS) can help distinguish heterogeneous vs. solvent relaxation mechanisms.16 The DAS from a full decay surface for Trp-Leu at pH 9.3 is shown in Fig. 4. Importantly, the DAS for the 1-2ps exponential has the "positive blue, negative red" signature of solvent relaxation while the DAS for 23ps is positive at all measured wavelengths (up to 400nm, where our detection system becomes inefficient). Further, the DAS shape does not suggest an imminent axis crossing; in fact, the 23ps DAS has about the same shape and width as those found by TCSPC (likely for heterogeneous conformers with distinct nanosecond decay times). The positive portion of a shortlived, solvent relaxation DAS that becomes negative in the red tail will necessarily be more narrow than the entirety of a positive-definite DAS arising from heterogeneity. One can never rule out the coexistence of some solvent relaxation in these spectra on similar timescales, but the data here do not require it.</p><p>We have compiled transients and DAS for several commercially available nonpolar Trp dipeptides, both in Trp-X and X-Trp configurations. In all cases, the 1-2ps DAS changes sign with wavelength while the 20-30ps term remains positive even at the red tail. Thus, the 1-2ps term represents bulk water relaxation and the 20-30ps term represents a rapidly decaying subpopulation (quenching mechanism yet to be determined). The bulk relaxation process reduces the energy of, but does not depopulate, the singlet (shifting bluer emitters to the red side). Hence it has no first order influence on the quantum yield. The 20-30ps decay term, however, contributes to the previously measured (and predicted) QSSQ of the peptides where it is found.</p><p>We have compiled distillates of these dipeptide decay characteristics in Table I and II. We define the parameter δ as the ratio of the sum of amplitudes of all terms with lifetimes greater than 2ps to all those above 100ps. In other words, δ is the ratio of molecular population that can be observed decaying in this new ultrafast measurement to that observable by TCSPC alone. The same ratios could be expected for any method limited by the intrinsic jitter of a photomultiplier detector (even though a few virtuoso TCSPC and phase/modulation measurements have been made down to ~40ps). We have studied the confidence limits on δ by examining the relationship between the fitting error (reduced chi-squared, χ2R) and excursions in the fixed amplitude of the 20-30ps term (all other parameters free). The F-test for such a system with ~90 degrees of freedom sets bounds at an approximately 10% χ2R increase; thus we expect (averaging multiple runs) that tabulated δ are accurate to approximately +/− 6%.</p><p>Examining Table I and II, it appears the presence or absence of the 20-30ps QSSQ term does not follow a predictable pattern vis-à-vis pH or Trp-X vs. X-Trp. On the other hand, it is present in a subset of these hydrophobic dipeptides but not in NATA or Trp itself, at either zwitterionic or anionic pH. This comports with prior evidence for specific configurations of the peptide bond as an important source of ultrafast quenching.37</p><p>We also sought to quantify the remaining fraction of the molecules that were quenched in subpicosecond events and/or, in truly static fashion, the ground state. This can be accomplished by recalculating <t> 33 with carefully normalized ps and ns measurements, or more simply, by meaning "Initial Brightness" (IB). If a time-resolved emission curve contains all exponential terms (i.e. none were below the instrument resolution) and no ground state molecules were selectively sequestered, a measurement at the peak of the curve would essentially be the sum of amplitudes times the radiative rate (and some other instrumental factors). Thus, equimolar (OD 295) solutions of compounds sharing the same radiative rate will show the same IB unless they have unresolved QSSQ. For our instrument, only events faster than ~350fs will be unrecorded in an initial brightness measurement.</p><p>In Table I and II, we use the same natural (radiative) lifetime surrogate, <τ>/QREL, used by Chen et al.33 to survey QSSQ across the range of peptides. A minor exception is that we choose NATA as the Q=1 standard, since it is essentially pH independent while Trp and the peptides have differing pKa values. As a reminder, these are apparent radiative lifetime ratios only; the fact that they exceed the value for a standard like NATA is what provides evidence of unreconciled QSSQ. The underlying principle we use is that spectra in the same approximate location for the same molecule should share a common true radiative rate.</p><p>In Table I and II, we show how each peptide which has a quantum yield defect relative to <τ> (e.g., QSSQ) is affected by both 20-30ps and subpicosecond events. The accompanying bar graphs (Fig.5 and 6) show how the initially elevated <τ>τ> τ>/QREL levels, when multiplied by first δ and then IB factors, lead to essentially constant <τ>τ> τ>/QREL values that agree with the standards (within error).</p><p>The 1-2ps solvent relaxation DAS seen for all of these compounds provides a small but important contribution to our IB measurements taken at 350nm; the final row of bars in Figs. 5 and 6 represent a correction to the IB factor that reduces it by the ratio (all amplitudes except 2ps/ all amplitudes). This extracts and excludes the presumed solvent relaxation term from IB. This presumes the 2ps term is a population conserving relaxation (if the 2ps decay term were entirely quenching rather than relaxation, the uncorrected IB term would be appropriate). The fact that the <τ>τ> τ>/QREL levels are both more consistent and closer to standards for the corrected IB approach suggests, again, that most of this 1-2ps term is "zero-sum" wrt yield and therefore arises from relaxation. The symmetry of the 2ps DAS wrt the abscissa also support that view.</p><p>The final reconciliation of <τ>τ> τ>/QREL values in Fig.5 and 6 leads to values consistent with NATA. Thus we can generally ascribe the QSSQ previously identified by Chen et al.32 to two ultrafast processes: either 20-30ps or sub-ps processes that deplete the population prior to conventional measurements.</p><!><p>The role of water in time resolved fluorescence of Trp has long been controversial. Clearly, an increased Trp dipole moment in the excited state, inside a macromolecule dissolved in a polar solvent like water, cannot be immune to generalized relaxation. Every Trp analog, peptide or protein we have studied possesses the strong 1-2ps transient with "blue positive/red negative" DAS that are signatory of excited state reaction).39 Bulk water relaxation near 1-2ps is known from a variety of measurements.14-16</p><p>What is more difficult to ascertain is the role, if any, played by slow relaxation in the macromolecular matrix that strongly couples to water 40 or simpler processes like water desorption 41 in the period of a few ps to ~100ps. If one focuses on the TDFSS (Time Dependent Fluorescent Stokes Shift) alone, it is impossible to discern relaxation from heterogeneity. Only a study of DAS, perhaps reinforced with TDFSS spectral width evolution profiles, can identify which mechanism is dominant. 16 Further peeling the 10-200ps time span apart into mixed relaxation and heterogeneity terms 42 can only be done with outside information (e.g., ad hoc, by postulating that terms faster than ca. 200ps have pure relaxation origins).</p><p>In contrast, we have shown that exponential terms near 16ps can have origins in heterogeneity in protein;16 here we show that even simple dipeptides display the same ultrafast quenching decay terms. One might expect that dipeptides would display a much simpler set of slow conformational fluctuations than complex polypeptides in folded configurations. Thus it is likely that the agreement between peptides and proteins about a 20--30ps quenching term is due to a common quenching mechanism. A plethora of other evidence points to the peptide bond as a potent quencher of Trp in agreement with our observations.43,44</p><p>It is not clear why the QSSQ proceeds at ~20ps in only a subset of these peptides; all have subpicosecond quenching as well. Further, NATA possesses the equivalent of two bonds but does not suffer the quench. It will require a rather large series of QM-MM simulations upon a variety of seeded peptide structures to learn whether only certain rotamers of both indole and the nonpolar sidechain lead to charge-transfer proximity with the peptide bond (or perhaps other quenching moieties). The fact that Trp also exhibits IB below the IB for NATA suggests that, at least for sub-ps decay, other (non peptide bond) quenching processes are still possible. As a final aside, we note for completeness that an atypical sharp reduction in kr upon relaxation can, in limited cases, also mimic a quenching. 39</p><!><p>The ultrafast emission of NATA and Trp dipeptides in water has been studied with femtosecond time resolution. Trp in many of these dipeptides exhibits a multiexponential fluorescence decay including a 20-30ps term in addition to the water relaxation of 1-2ps. This multiexponentiality is apparently associated with ground-state heterogeneity and may arise from different rotamers of Trp or the 'X' sidechains. Experiments on constrained system will be needed to confirm this view.44 The corresponding DAS (decay associated spectra) are positive throughout the measurement range. Since these ultrafast components are not detected in conventional time-resolved or phase/modulation experiments, their presence might frustrate the matching of ns decay populations to either nmr-derived or computational predictions for conformers. Accounting for QSSQ may thus improve our understanding of polypeptides. On a more challenging level, this range of ultrafast quenching rates provides signposts for QM-MM simulation of Trp and its environment. 45 Also, we must leave open the possibility that some lifetimes could be surrogates for rates of internal motions that quench certain subpopulations. In any case, heterogeneity provides the dominant source of ultrafast QSSQ.</p>

PubMed Author Manuscript

Inhibition of ischemia-induced angiogenesis by benzo[a]pyrene in a manner dependent on the aryl hydrocarbon receptor

We have investigated the effect of benzo[a]pyrene (B[a]P), a carcinogen of tobacco smoke and an agonist for the aryl hydrocarbon receptor (AHR), on hypoxia-induced angiogenesis. Ischemia was induced by femoral artery ligation in wild-type and AHR-null mice, and the animals were subjected to oral administration of B[a]P (125 mg/kg) once a week. Exposure to B[a]P up-regulated the expression of metallothionein in the ischemic hindlimb and markedly inhibited ischemia-induced angiogenesis in wild-type mice. The amounts of interleukin-6 and of vascular endothelial growth factor (VEGF) mRNA in the ischemic hindlimb of wild-type mice were reduced by exposure to B[a]P. These various effects of B[a]P were markedly attenuated in AHR-null mice. Our observations suggest that the loss of the inhibitory effect of B[a]P on ischemia-induced angiogenesis apparent in AHR-null mice may be attributable to maintenance of interleukin-6 expression and consequent promotion of angiogenesis through up-regulation of VEGF expression.

inhibition_of_ischemia-induced_angiogenesis_by_benzo[a]pyrene_in_a_manner_dependent_on_the_aryl_hydr

<!>Animals<!>Systolic blood pressure and laser Doppler perfusion imaging<!>Oral exposure to B[a]P<!>RT-PCR and immunoblot analyses<!>Assay of IL-6<!>Statistical analysis<!>Body weight and systolic blood pressure<!>Blood flow ratio<!>Expression of VEGF, Ang-1, and their receptors<!>Expression of HIF-1\xce\xb1, ARNT, AHR, and CYP1A1<!>Expression of MT-1 and MT-2<!>Expression of IL-6<!>Discussion

<p>Cigarette smoke is a major risk factor for ischemic heart disease, peripheral vascular disease, and chronic obstructive pulmonary disease [1,2]. It has adverse effects on vascular biology, inducing endothelial dysfunction and arterial stiffness [3], and it inhibits angiogenesis by pulmonary artery endothelial cells in the setting of severe vascular obstruction and lung tissue ischemia [4]. Angiogenesis, the development of new blood vessels from existing vessels, is a tightly controlled physiological process [5]. Vascular endothelial growth factor (VEGF) is a key angiogenic factor produced by ischemic tissue and growing tumors [6], and up-regulation of VEGF expression at the transcriptional level is thought to be responsible for the progressive development of the collateral circulation [7]. Although the association between smoking and vascular disease is well established, the mechanism by which cigarette smoke inhibits angiogenesis has remained unclear.</p><p>The effects of benzo[a]pyrene (B[a]P) in cigarette smoke and of other polycyclic and halogenated aromatic hydrocarbons in the environment are mediated by the aryl hydrocarbon receptor (AHR) [8]. The ligand-bound form of the AHR and AHR nuclear translocator (ARNT), which belongs to the Per-Arnt-Sim (PAS) family of basic helix-loop-helix transcription factors, form a heterodimeric transcription factor [9] that binds to xenobiotic response elements (XREs) in the promoter regions of target genes such as that encoding cytochrome P4501A1 (CYP1A1) [10]. Activation of the VEGF gene is mediated by the binding of another basic helix-loop-helix transcription factor, hypoxia-inducible factor-1α (HIF-1α) [11], which also dimerizes with ARNT [12]. We previously showed that AHR-ARNT signaling plays an important role in the regulation of ischemia-induced angiogenesis [13]. Given that smoking is a risk factor for ischemic heart disease and peripheral vascular disease, we hypothesized that B[a]P might impair angiogenesis and that AHR signaling might contribute to such an effect. We have now investigated this hypothesis with the use of AHR-null mice and an animal model in which ischemia is induced surgically in a hindlimb.</p><!><p>Twelve-week-old male wild-type (WT) or AHR-null mice backcrossed with C57BL/6N mice for more than eight generations were studied. AHR-null mice were obtained from the National Cancer Institute colony [14] and maintained by the Animal Resource Facility at Nagoya University Graduate School of Medicine. Animals were subjected to left femoral artery ligation as described previously [13]. All experimental procedures were performed in accordance with Institutional Guidelines for Animal Research, and the study was approved by the Animal Ethics Committee of Nagoya University Graduate School of Medicine.</p><!><p>Systolic blood pressure of conscious mice was measured by the tail-cuff method (BP-98A, MCP-1; Softron, Tokyo, Japan) immediately before and each week after arterial ligation. Evidence for ischemia-induced functional changes in vascularization was obtained by laser Doppler perfusion imaging with a laser Doppler blood flow analyzer (Moor LDI; Moor Instruments, Axminster, UK) immediately as well as each week after surgery, as previously described [13]. After scanning, the stored images were analyzed to quantify blood flow, and the average flows of the ischemic and nonischemic hindlimbs were calculated. To avoid interference from ambient light and body temperature, we expressed blood flow in the ischemic leg relative to that in the nonischemic leg.</p><!><p>The effect of B[a]P treatment on angiogenesis induced by femoral artery ligation was examined in WT and AHR-null mice. B[a]P (Sigma–Aldrich Japan, Tokyo, Japan) was first administered to WT mice by oral gavage in corn oil (10 ml/kg) at 125, 25, or 5 mg/kg once per week as described [15], with the first dose being administered immediately before surgery. Ischemia-induced functional changes in vascularization were evaluated by laser Doppler perfusion imaging immediately as well as 7, 14, and 21 days after surgery. Given that B[a]P at only the dose of 125 mg/kg per week induced significant inhibition of the recovery of blood flow in the ischemic hindlimb of WT mice, the effects of B[a]P were compared between WT and AHR-null mice at this dose.</p><!><p>Total RNA was extracted from ischemic or nonischemic thigh muscles isolated 3 days after surgery and was subjected to quantitative reverse transcription (RT) and polymerase chain reaction (PCR) analysis with primers and probes specific for HIF-1α, ARNT, AHR, CYP1A1, VEGF, VEGF receptor (Flt-1, Flk-1), Ang-1, Ang-1 receptor (Tie2), metallothionein (MT)-1, MT-2, and interleukin (IL)-6 mRNAs. The mRNA for β-actin was used as an internal standard. Protein extracts of ischemic or nonischemic thigh muscles isolated 1 week after surgery were subjected to immunoblot analysis with rabbit polyclonal antibodies to mouse VEGF (Santa Cruz Biotechnology, Santa Monica, CA), to mouse Ang-1 (Alpha Diagnostic, San Antonio, TX), and to human MT (Santa Cruz Biotechnology) at dilutions of 1:500, 1:1000, and 1:200, respectively, as well as with a mouse monoclonal antibody to mouse β-actin (Sigma, St. Louis, MO) at a dilution of 1:2000. Immune complexes were detected with enhanced chemiluminescence (ECL) reagents (GE Healthcare Bio-Science, Piscataway, NJ).</p><!><p>Protein extracts prepared from ischemic or nonischemic thigh muscles isolated 1 week after surgery were assayed for IL-6 with an enzyme-linked immunosorbent assay (ELISA) kit (Pierce/Endogen, Rockford, IL) as described [16]. The assay was performed in duplicate, and absorbance at 450 nm was measured with a microtiter plate reader. The tissue content of IL-6 was expressed as picograms per milligram of protein.</p><!><p>Data are presented as means ± SEM. Statistical significance of differences was evaluated by one-way analysis of variance followed by Dunnett's post hoc test for comparisons among four or eight means. A P value of <0.05 was considered statistically significant.</p><!><p>Body weight did not differ between AHR-null and WT mice either before or for up to 4 weeks after arterial ligation to induce hypoxia in the left hindlimb (Fig. 1A). Weekly oral administration of B[a]P (125 mg/kg) beginning immediately before surgery did not affect body weight in either mouse strain (Fig. 1A). Systolic blood pressure remained at basal levels for up to 4 weeks after surgery and also did not differ between AHR-null and WT mice (Fig. 1B). Exposure to B[a]P did not affect systolic blood pressure in AHR-null or WT mice (Fig. 1B).</p><!><p>The ischemic/nonischemic blood flow ratio of AHR-null mice was significantly greater than that of WT mice at 1 week after surgery, and this difference remained apparent for up to 4 weeks (Fig. 1C and D). Blood flow recovered in the ischemic hindlimb of WT mice exposed to B[a]P. However, 1 week after surgery, the ischemic/nonischemic blood flow ratio of WT mice exposed to B[a]P was significantly smaller than that of vehicle-treated WT mice, and this difference also remained apparent for up to 4 weeks (Fig. 1C and D). Furthermore, the blood flow ratio was significantly greater in AHR-null mice exposed to B[a]P than in WT animals exposed to this agent.</p><!><p>The amounts of mRNAs for the angiogenic factors, VEGF and Ang-1, in the ischemic hindlimb were significantly greater for AHR-null mice than for WT mice at 3 days after surgery (Fig. 2A). The abundance of VEGF mRNA in the ischemic hindlimb was also significantly reduced in WT mice exposed to B[a]P compared with that in vehicle-treated WT mice. The amounts of VEGF and Ang-1 mRNAs in the ischemic hindlimb were significantly greater for AHR-null mice exposed to B[a]P than for WT animals exposed to this agent. Immunoblot analysis revealed similar changes in the abundance of VEGF and Ang-1 proteins at 1 week after surgery (Fig. 2B). With regard to the expression of genes encoding receptors for VEGF (Flt-1, Flk-1) or for Ang-1 (Tie2), the amount of Flk-1 mRNA in the ischemic hindlimb was significantly greater for AHR-null mice than for WT mice (Fig. 2C). However, B[a]P had no effect on the abundance of Flt-1, Flk-1, or Tie2 mRNAs in the ischemic hindlimb of either mouse strain, and there were no differences in the amounts of these mRNAs in the ischemic hindlimb between AHR-null or WT mice exposed to B[a]P (Fig. 2C).</p><!><p>Quantitative RT-PCR analysis revealed that, whereas the amounts of both HIF-1α and ARNT mRNAs were significantly greater for the ischemic hindlimb than for the nonischemic hind-limb of AHR-null or WT mice at 3 days after surgery, they were also significantly greater for the ischemic hindlimb of AHR-null mice than for that of WT mice (Fig. 3A and B). Exposure to B[a]P did not affect the amounts of HIF-1α or ARNT mRNAs in AHR-null or WT mice. The amount of AHR mRNA in skeletal muscle of WT mice did not differ between the ischemic and nonischemic hindlimbs and was also not affected by exposure to B[a]P (data not shown). There was no difference in the abundance of CYP1A1 mRNA between the ischemic and nonischemic hindlimbs of either mouse strain (Fig. 3C). Exposure to B[a]P induced significant increases in the amount of CYP1A1 mRNA only in WT mice (Fig. 3C).</p><!><p>The amounts of MT-1 and MT-2 mRNAs were increased in the ischemic hindlimb of both AHR-null and WT mice compared with those in the corresponding nonischemic hindlimb at 3 days after surgery (Fig. 4A). Exposure to B[a]P resulted in significant increases in the abundance of MT-1 and MT-2 mRNAs in the ischemic hindlimb of WT mice but not in that of AHR-null mice (Fig. 4A). Immunoblot analysis also revealed similar changes in the abundance of MT at 1 week after surgery (Fig. 4B).</p><!><p>The amount of IL-6 mRNA was increased in the ischemic hindlimb of both AHR-null and WT mice compared with that in the corresponding nonischemic hindlimb at 3 days after surgery (Fig. 4C). This effect of ischemia was more pronounced in AHR-null mice than in WT mice. The abundance of IL-6 mRNA in the ischemic hindlimb of AHR-null mice exposed to B[a]P was also significantly greater than that for WT mice exposed to this agent. The amount of IL-6 immunoreactivity was also increased in the ischemic hindlimb of both AHR-null and WT mice compared with that in the corresponding nonischemic hindlimb at 1 week after surgery (Fig. 4D). IL-6 immunoreactivity in the ischemic hindlimb of WT mice exposed to B[a]P was significantly reduced compared with that for vehicle-treated WT mice. It was also significantly greater in the ischemic hindlimb of AHR-null mice exposed to B[a]P than in that of WT animals exposed to this agent.</p><!><p>Exposure to cigarette smoke has a negative effect on endothelial function, serum lipid profile, and hemostatic factors [17]. Smoking plays a central role in the initiation and progression of Buerger's disease, a nonatherosclerotic condition that most commonly affects small and medium-sized arteries and veins as well as nerves of the extremities [18]. To examine the effects of cigarette smoke on angiogenesis, we treated mice with B[a]P, one of the polycyclic and halogenated aromatic hydrocarbons found in tobacco smoke. Oral exposure to B[a]P resulted in significant inhibition of the increase in blood flow induced by hypoxia in WT mice. Furthermore, we found that this inhibition of angiogenesis by B[a]P was associated with inhibition of hypoxia-induced up-regulation of VEGF expression. Cigarette smoke was previously shown to inhibit secretion of the soluble form of the VEGF receptor Flt-1 in a dose-dependent manner [19]. Inhibition of the VEGF receptor Flk-1 was also shown to augment the vascular and endothelial dysfunction induced by cigarette smoke [20]. However, we detected no significant effect of B[a]P on the abundance of Flt-1 or Flk-1 mRNAs in ischemic hindlimb of WT mice in the present study.</p><p>MTs are cysteine-rich metal-binding proteins with diverse physiological functions, including protection against metal toxicity and oxidants [21]. Furthermore, MT expression has been shown to be increased at the transcriptional level by hypoxia acting through metal response elements in the proximal promoter region of MT genes [22]. In the present study, we detected a marked increase in the expression of MT-1 and MT-2 in the ischemic hindlimb of both AHR-null and WT mice. Moreover, B[a]P induced a further marked increase in MT expression in the ischemic hindlimb of WT mice. This observation is consistent with previous studies showing that MT expression was induced in response to B[a]P exposure [23]. Such an increase in MT expression may reflect a protective response to B[a]P-induced oxidative stress. We also found that B[a]P induced a decrease in the amount of IL-6 in the ischemic hindlimb of WT mice. IL-6 has been shown to induce MT expression by binding to IL-6-responsive elements in the gene promoter [24]. On the other hand, the lipopolysaccharide-induced increases in the circulating concentration of IL-6 as well as in the expression of IL-6 in lung, kidney, and liver were markedly greater in mice deficient in MT-1 and MT-2 than in WT mice [25]. Furthermore, overexpression of MT-1 inhibited up-regulation of ectopic IL-6 expression in astrocytes of transgenic mice [26]. IL-6 has also been shown to support angiogenesis by up-regulating the expression of VEGF [27]. The increase in MT expression in the ischemic hindlimb of WT mice induced by B[a]P in the present study may thus result in down-regulation of IL-6 expression, which in turn may lead to impairment of hypoxia-induced angiogenesis.</p><p>AHR-null mice are relatively unaffected by the potent AHR ligand 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) at doses that induce severe toxic and pathological effects in WT mice [28]. Moreover, genotoxic and carcinogenic responses to B[a]P or dibenzo[a,l]pyrene are greatly diminished in AHR-null mice [29]. In the present study, the increase in the amount of CYP1A1 mRNA induced by B[a]P in WT mice was not apparent in AHR-null mice, consistent with our finding that AHR deficiency attenuated the inhibition of ischemia-induced angiogenesis by B[a]P. The expression of MT induced by B[a]P in the ischemic hindlimb of WT mice was not detected in that of AHR-null mice. The amount of IL-6 in the ischemic hindlimb of AHR-null mice was not affected by exposure to B[a]P, consistent with the lack of an effect of B[a]P on MT expression and possibly contributing to the diminution of the impairment of ischemia-induced angiogenesis by B[a]P in these mice.</p><p>In conclusion, oral exposure to B[a]P resulted in a marked impairment of angiogenesis in response to surgically induced hindlimb ischemia. Furthermore, the impairment of ischemia-induced angiogenesis by B[a]P was greatly reduced in AHR-null mice. Our observations suggest that the loss of the inhibitory effect of B[a]P on ischemia-induced angiogenesis apparent in AHR-null mice may be attributable to maintenance of IL-6 expression and consequent promotion of angiogenesis through up-regulation of VEGF expression.</p>

PubMed Author Manuscript

Convergent and Stereospecific Synthesis of Complex Skipped Polyenes and Polyunsaturated Fatty Acids

Skipped polyenes (i.e. 1,4-dienes and higher homologues) are stereodefined components of a vast array of biologically important natural products, including polyunsaturated fatty acids. While widespread in nature, these architectures are generally considered to represent significant barriers to efficient chemical synthesis. While partial reduction of skipped poly-ynes provides a pathway to a subset of such structures, general chemical methods for the preparation of skipped polyenes that contain varied stereochemistries and substitution patterns are lacking. Here, we describe a metal-promoted reductive cross-coupling reaction between vinylcyclopropanes and alkynes (or vinylsilanes) that provides stereoselective access to a diverse array of skipped polyenes through a process that establishes one C\xe2\x80\x93C bond, generates up to three stereodefined alkenes, and can be used to introduce stereogenic centers at the central positions of the skipped polyene motif. We also demonstrate the significance of the present bond construction by preparing substituted and stereodefined polyunsaturated synthetic fatty acids.

convergent_and_stereospecific_synthesis_of_complex_skipped_polyenes_and_polyunsaturated_fatty_acids

<!>Reaction design<!>Exploring the reactivity of the vinylcyclopropane component<!>Probing stereospecificity<!>Skipped trienes from the reductive coupling of vinylcyclopropanes with alkynes<!>Synthesis of novel polyunsaturated fatty acids (PUFAs)<!>Conclusion<!>General procedure for the reductive cross-coupling of vinylcyclopropanes with alkynes

<p>Fatty acids are a subset of small molecules that are essential for life, playing not only central roles in compartmentalization and membrane function but also impacting cellular pathways that regulate blood pressure, clotting and lipid levels as well as the immune response and inflammation.(1–3) Polyunsaturated fatty acids (i.e. arachidonic acid, α-linoleic acid, γ-linoleic acid and eicosapentaenoic acid) are a subset of this large class of biomolecules that are present in all higher organisms and play critical roles in human health (Figure 1A).(4,5) These ubiquitous biomolecules share a common structural motif that imparts their unique properties – methylene interrupted polyenes (highlighted in blue). This central architectural feature is a stereodefined motif that is encountered throughout nature, with examples including structurally complex bioactive natural products from polyketide, terpene and alkaloid biosynthetic pathways. These more architecturally intricate molecules, identified as potent antibiotic, antifungal and cytotoxic agents, house skipped polyenes that contain stereochemically diverse di- and tri-substituted alkenes (i.e. ripostatin A, madangamine A).(6–10) Notably, more complex examples include secondary metabolites that house 1,4-dienes with stereogenic sp3 carbons at the central position of the isolated non-conjugated diene (i.e. jerangolid D and phorbasin C).(11–14)</p><p>The stereoselective preparation of these structural motifs remains a significant challenge in organic chemistry.(15–18) While methods based on carbonyl olefination, alkylation and partial reduction of acetylenes have been employed in numerous campaigns in stereoselective synthesis, these often multi-step processes are each plagued by significant limitations in scope, selectivity and efficiency. Here, we describe a titanium-mediated stereoselective fragment coupling reaction that delivers a variety of complex skipped polyenes by a process that establishes five unique stereochemical relationships across the skipped polyene backbone in concert with C–C bond formation (Figure 2). While defining a unique stereoselective transformation in organic chemistry, this convergent coupling reaction illuminates a powerful solution to the synthesis of a complex stereodefined structural motif observed throughout nature. In addition to presenting the basic reaction scope and selectivity of the process, we demonstrate the application of this reaction to the preparation of complex synthetic polyunsaturated fatty acids (PUFAs).</p><!><p>In designing a mode of reactivity suitable for the stereoselective construction of acyclic skipped polyenes, we were inspired by the well-established sigmatropic rearrangement chemistry of vinylcyclopropanes (Figure 3). In the case of cis-divinylcyclopropanes 1, Cope rearrangement (a) is driven by the release of ring strain associated with the cyclopropane, and a cyclic 1,4-diene 2 is produced (Figure 3a).(19,20) In a mechanistically related rearrangement, cis-disubstituted vinylcyclopropanes (3) can undergo 1,5-hydrogen migration (b) to deliver acyclic 1,4-dienes (4) (Figure 3b).(21–24)</p><p>We speculated that a related six-electron process could ensue from an organometallic intermediate of general structure 5 (Figure 3c). Here, we reasoned that fragmentation would have the potential to proceed in a stereospecific manner, where the stereochemistry of the organometallic intermediate (5) could be translated directly to the stereochemistry of the skipped polyene product (6).</p><p>To prepare the required stereodefined metalated cyclopropanes (5), we speculated that a titanium-mediated, alkoxide-directed fragment union reaction between a substituted vinylcyclopropane 7 and a suitably functionalized coupling partner (i.e. 8), could deliver tricyclic titanacyclopentane intermediates of general structure 9 (Figure 3d).(25) In addition to overcoming the sluggish reactivity of substituted alkenes in carbometalation chemistry, the hydroxy group of 7 would orchestrate encapsulation of the titanium center while establishing a rigid organometallic intermediate 9 suitably functionalized for the planned fragmentation. If rupture of the tricycle 9 proceeds in a stereospecific fashion, then the stereoselective annulation process (7 + 8 → 10) would lead, after hydrolysis, to the establishment of a stereodefined skipped polyene 11. Overall, this reaction design would define a convergent coupling reaction that has the potential to establish up to three stereodefined alkenes and one new stereogenic center (), while generating a complex yet common stereochemically defined structural motif in nature. As synthetic approaches to stereodefined cyclopropanes of general structure 7 are readily available, we reasoned that the proposed coupling process would define a potentially powerful entry to skipped polyenes.</p><!><p>The feasibility of this proposal was first examined in coupling reactions of chlorodimethylvinylsilane with substituted vinylcyclopropanes, themselves derived from well-established cyclopropanation chemistry of allylic diazoacetates (i.e. 12–16) (Table 1).(26,27) As depicted in entry 1, initial exploration provided support for the proposed stereoselective coupling process. Here, vinylcyclopropane 17 was converted to 1,4-diene 23 in 58% isolated yield (82% yield based on recovered starting material). While illustrating a useful reaction for the synthesis of a 1,4-diene bearing a central quaternary carbon, this reaction proceeded in a stereoselective manner and established a central (E)-disubstituted alkene (E:Z ≥ 20:1).</p><p>Similarly, conversion of vinylcyclopropane 18 to 1,4-diene 24 proceeded in 54% isolated yield (entry 2). Here, high E- selectivity (≥ 20:1) was accompanied by the generation of a 1 , 4-diene possessing a central chiral carbon. Entries 3 and 4 demonstrate that this coupling process is compatible with more highly substituted substrates. While conversion of 19 to the 1,4-diene 25 establishes a trisubstituted alkene, the trisubstituted vinylcyclopropane 20 is converted to a 1,4-diene that possesses a 1,1-disubstituted alkene and an allylic stereocenter (26). In both cases, high (E)-selectivity is observed in the formation of the central disubstituted alkene (≥ 10:1).</p><p>This coupling reaction can be used for the preparation of substituted skipped polyenes that house stereogenic centers at the 3-and 6-positions of the central 1,4-diene. As depicted in entries 5 and 6, vinylcyclopropanes 21 and 22 can be converted to stereodefined 1,4-dienes 27 and 28 in 55 and 52% yield, respectively. In each case, products were isolated as single isomers (dr ≥ 20:1, E:Z ≥ 20:1).</p><!><p>This complex coupling reaction appears to proceed in a stereospecific fashion. As depicted in entries 7 and 8 of Table 1, coupling of the isomeric vinylcyclopropanes 30 and 32 proceeds to deliver the skipped Z,E-diene 31 and E,E-diene 33. While the establishment of the central E-trisubstituted alkene occurs with apparently high levels of stereoselectivity (≥20:1), the disubstituted alkene of these products is generated in a stereospecific fashion. On close spectroscopic analysis of the crude products from these coupling reactions, no evidence could be found for the production of minor products derived from conversion of 30 to 33 or from 32 to 31. This intimate relationship between relative stereochemistry of the starting material and olefin geometry of the 1,4-diene product is consistent with a mechanistic proposal that follows from stereoselective formation of the tricyclic titanacyclopentanes I and J, followed by stereospecific fragmentation.</p><p>While defining a unique entry to stereodefined 1,4-dienes, to the best of our knowledge, this coupling reaction represents a unique transformation in organic chemistry. Alkoxide-directed metal centered [2+2+1] annulation enables stereoselective access to a fleeting organometallic intermediate (a configurationally defined cyclopropylcarbinyl anion) whose subsequent fragmentation occurs in a stereospecific manner.(28–34)</p><!><p>The titanium-mediated cross-coupling reaction of vinylcyclopropanes with alkynes enables direct stereoselective access to skipped trienes. As depicted in Figure 4a, reaction of vinylcyclopropane 34 with TMS-alkyne 35, followed by desilylation (TBAF, THF) delivers complex triene 36 in 34% isolated yield over the two-step process (82% based on recovered starting material in the T i-mediated coupling reaction). While regioselective functionalization of the alkyne proceeds in a manner where C–C bond formation occurs distal to the TMS-substituent,(35) the sense of regio- and stereoselective functionalization of vinylcyclopropane 34 is consistent with the observations made during study of the vinylcyclopropane–vinylsilane coupling. Overall, this convergent coupling reaction establishes one C–C bond, sets three stereodefined alkenes, one sp3 stereogenic center, and delivers a complex skipped triene in a single step. Notably, each sp3 stereogenic center in this product is positioned between the alkenes of the skipped polyene, defining a product that contains five contiguous stereogenic relationships that span eight contiguous carbons.</p><!><p>As a test of our methodology, we targeted application to the synthesis of novel stereochemically complex polyunsaturated fatty acids. While polyunsaturated fatty acids define a region of natural product chemical space that is rich in biological function, the dearth of methods available for the efficient synthesis of stereodefined skipped polyenes has generally limited medicinal investigation. The targets selected to explore the utility of our vinylcyclopropane-based skipped polyene synthesis were fatty acids that display complex stereochemical features for which a general synthetic method has not been available. Specifically, we aimed to access polyunsaturated fatty acids that contain: 1) di- and tri-substituted alkenes, 2) E- and Z- olefins, and 3) contain stereochemistry at the central positions of the 1,4-diene motifs. Such features represent the most complex examples of skipped polyenes observed in bioactive natural products.</p><p>As illustrated in Figure 4b, a sequence involving 1) reductive cross-coupling of suitably substituted vinylcyclopropanes with TMS alkynes, 2) desilylation, and 3) oxidation provides polyunsaturated fatty acids in a highly stereoselective fashion. While equation 2 demonstrates this process is effective for the preparation of a simple C-14 skipped triene-containing fatty acid, equations 3 and 4 confirm that this synthetic sequence is equally effective for the generation of higher homologues that contain stereodefined tetra-enes (39) and branched alkyl substitution (41).</p><p>While serving as a platform to challenge our chemical technology, this exercise has delivered complex stereodefined synthetic fatty acids. Likely due to the difficulties associated with preparing such architectures with existing chemical methods, such structures have never been reported. The mode of chemical reactivity described here provides a unique and powerful means to initiate investigations aimed at elucidating the physical and biological properties of complex and diverse polyunsaturated fatty acids.</p><!><p>Skipped polyenes are structural motifs present in organic molecules observed throughout nature and biology. Natural products that possess this motif are known to play key roles in compartmentalization and signal transduction, while others possess potent anticancer and antibiotic properties. Although recognized as a central stereodefined architectural feature of a variety of biologically significant natural products, skipped polyenes have remained as significant challenges in stereoselective chemical synthesis. We have described a chemical process that enables direct and stereoselective access to complex and diverse skipped polyenes by the direct union of vinylcyclopropanes with alkynes and vinylsilanes. While appearing complex, the vinylcyclopropane substrates for this process are available in a few steps from allylic diazoacetates by well-known cyclopropanation chemistry. Due to the previously established barriers associated with the preparation of this type of stereodefined architecture, and the relative ease with which the current method delivers these complex systems, we are looking forward to scientific advances that follow from this initial report.</p><!><p>To a round bottom flask (RBF) containing alkyne (2 mmol) and ClTi(Oi-Pr)3 (2 – 2.5 mmol, 1.0 M in hexanes) in toluene (0.1 M) at −78 °C is added c-C5H9MgCl (4 – 5 mmol) by syringe and the mixture is warmed to −30 °C and stirred for 1 h (becoming deep brown in color). In a separate RBF containing a vinylcyclopropane carbinol (1 mmol) in diethyl ether (Et2O) (<0.3 M) at −78 °C is added n-BuLi (1.2 mmol) by syringe and the mixture is warmed to 0 °C over 20 min and then added by cannula to the titanium complex that has been re-cooled to −70 °C. The mixture is then slowly warmed to RT over 2–3 h and treated with 1N HCl (~5 mL / mmol of vinylcyclopropane used) with rapid stirring. After 10 min the now colorless mixture is further diluted with ethyl acetate (EtOAc) and then filtered through a pad of silica rinsing with additional EtOAc. After concentration in vacuo, the non-polar product fractions are separated from the unreacted vinylcyclopropane by flash column chromatography (0 to 5% Et2O in hexanes followed by flushing with EtOAc).</p><p>In the examples presented, subsequent functionalization by desilylation and/or oxidation was accomplished by the following protocols: To this crude product in a plastic vial dissolved in acetonitrile (MeCN) / dichloromethane (CH2Cl2) (5 mL, 1:1) at 0 °C is added HF • pyridine by plastic syringe in 1 mL / 5 min increments. When the reaction is complete (indicated by TLC), the reaction mixture is poured slowly into a plastic beaker containing stirred saturated aqueous NaHCO3 (100 mL) at 0 °C, then diluted with Et2O. The layers are separated, the aqueous phase extracted with Et2O, organic extracts combined and filtered through MgSO4, concentrated in vacuo, and purified by flash column chromatography (10 – 15% EtOAc in hexanes). The resultant alcohol, which could contain small amounts of reduction side products, was used directly in the following oxidation. The purified skipped polyene alcohol (ie 0.5 mmol) in dimethylformamide (DMF) (3 mL) is added pyridinium dichromate (PDC) (2.0 mmol) and two drops of water at RT and the reaction is stirred for 12 h. The reaction mixture is then poured into brine (15 mL) and extracted three times with Et2O, extracts combined, washed with brine, filtered through MgSO4, concentrated in vacuo, and purified by flash column chromatography (20% EtOAc in hexanes) to yield the pure product.</p>

PubMed Author Manuscript

Luminescent Properties of Eu(III) Chelates on Metal Nanorods

In this article, we report the change of optical properties for europium chelates on silver nanorods by near-field interactions. The silver rods were fabricated in a seed-growth method followed by depositing thin layers of silica on the surfaces. The europium chelates were physically absorbed in the silica layers on the silver rods. The silver rods were observed to exhibit two plasmon absorption bands from longitudinal and transverse directions, respectively, centered at 394 and 675 nm, close to absorption and emission bands from the Eu(III) chelates. As a result, the immobilized Eu(III) chelates on the silver rods should have strong interactions with the silver nanorods and lead to greatly improved optical properties. The Eu\xe2\x80\x93Ag rod complexes were observed to have enhanced emission intensity up to 240-fold in comparison with the Eu(III) chelates in the metal-free silica templates. This enhancement is much larger than the value for the Eu(III) chelates on the gold rods or silver spheres indicating the presence of stronger interactions for the Eu(III) chelates with the silver rods. The interactions of Eu(III) chelates with the silver rods were also proven by extremely reduced lifetime. Moreover, the Eu\xe2\x80\x93Ag rod complexes exhibited a polarized emission, which was also due to strong interactions of the Eu(III) chelates with the silver rods. All of these features may promise that the Eu(III)\xe2\x80\x93Ag rod complexes have great potential for use as fluorescence imaging agents in biological assays.

luminescent_properties_of_eu(iii)_chelates_on_metal_nanorods

INTRODUCTION<!>EXPERIMENTAL SECTION<!>Preparation of Metal Nanorods<!>Depositing Silica Layers on Metal Nanoparticles<!>Immobilizing Eu(III) Chelates on Metal Nanoparticles<!>Ensemble Spectral and Anisotropy Measurements<!>TEM Image<!>RESULTS AND DISCUSSION<!>SUMMARY<!>

<p>It is known that a lanthanide emits luminescence with a large Stokes shift, narrow emission band, and long lifetime, so the use of lanthanide dyes may offer an opportunity to create a sensitive time-resolved bioassay or cell imaging with a low luminescence background.1,2 However, the emission from the lanthanide may involve an electron transition in its 4f orbits, which is basically forbidden, so in comparison with a conventional organic fluorophore the lanthanide dye often has an extremely low absorbance coefficient and a very slow emissive rate.3 Consequently, the lanthanide dye often displays a low quantum yield and brightness, and a lanthanide-based fluorescence bioassay often results in a low detection sensitivity. In addition, the lanthanide dye has an extremely long decay time up to milliseconds, which may often bring up a low photon count rate in the emission and result in low emission intensity. Therefore, the single molecule detection (SMD) of the lanthanide dye is not reported to date.4</p><p>With advances in coordination chemistry, new lanthanide chelates have been developed with significantly improved optical properties.1,2 Basically, these lanthanide chelates are created by coordinating chromophores with lanthanide ions, in which the chromophores can act as antennas or sensitizers to absorb photons and subsequently transfer the photons to the lanthanide centers for increasing their emission rates and efficiencies. In comparison with the lanthanide ions, these lanthanide chelates have significantly increased optical properties including brightness and quantum yields. Because of absorption from the ligands, the lanthanide chelates often exhibit a broader excitation range leading to a significant shift of maximal excitation wavelength to the red region.5 Hence, it is believed that the formation of lanthanide chelates may overcome some inherent weakness of lanthanide dyes especially on the absorption cross-section and excitation wavelength. However, it is also noticed that the decay rate of the lanthanide center in the chelate is not really increased with the coordination.6 Therefore, there is a basic research need to develop a strategy that can increase the radiative rate of the lanthanide chelate and furthermore increase the cyclic number for the excitation/emission of Eu(III) chelate in a period of time and the photon count rate in the measurement.</p><p>The near-field interaction approach is considered to enable use for such a purpose. Fundamentally, as induced by an incident light, the metal nanoparticle of subwavelength size can create a local electric field around it.7,8 When a fluorophore is put in the local field within a near-field range from the metal nanoparticle, the coupling interaction between the fluorophore and metal nanoparticle can increase the radiative rate of the fluorophore and furthermore enhance the emission intensity of the fluorophore up to 10- to 103-fold.9–12 Increased radiative rate of the fluorophore may also bring a decrease in the lifetime.9 Thus, the metal-enhanced fluorescence (MEF) by the near-field interaction effect is often accompanied with a decrease of lifetime for a fluorophore. In this study, the near-field interaction approach was employed to improve the optical properties of lanthanide chelates. Typically, the lanthanide chelates were immobilized on the metal nanoparticles within a near-field range to initiate the coupling interactions. The feasibility of this approach has been demonstrated by one of our earlier reports in which the lanthanide chelates were localized in the cores of silver shells and the emission could be greatly enhanced due the near-field interactions.13 More work is needed to optimize the conditions for achieving maximally enhanced luminescence.14–16 Importantly, the silver shells in the earlier report were found to exhibit only one plasmon band that is localized between the absorption and emission bands of Eu(III) chelate.13 Even though the interactions in the near-field range cannot be completely reflected by the observations on the far-field spectra,9 we still believe that it is important to achieve the maximally spectral overlapping from the fluorophores and metal nanoparticles to achieve the efficient near-field interaction between them. Of course, the couplings between the lanthanide chelates and metal shells were considered to be insufficient.9 Actually, it is almost impossible to fabricate a metal shell, which displays only a single plasmon band enabling a sufficient coupling with a lanthanide chelate, which has a very large Stokes shift at both its excitation and emission bands. Therefore, new metal substrates are needed.</p><p>Metal nanorods have been widely reported as new nanoparticle substrates.17,18 We were interested in the metal rods and intended to use them as the substrates to couple with the lanthanide chelates because the metal rods can display two separated plasmon bands from the longitudinal and transverse directions, respectively, and the band maxima can be tuned with the aspect ratio of metal rods.17,18 Particularly, we have research interest in silver nanorods because the silver rods can display two plasmon bands close to the absorption and emission maxima from the Eu(III) chelates. In this study, the silver rods were fabricated in a chemical method followed by depositing the silica layers on the surfaces. The Eu(III) chelates were physically absorbed in the silica layers on the silver rods within the near-field range from the metal surfaces. Optical properties from the Eu(III) chelates on the silver rods were determined on an ensemble fluorescence spectrophotometer with the time-gated conditions. Compared with the emission properties of Eu(III) chelates in the metal-free silica templates, the influences from the silver rods to the emission of Eu(III) chelates could be explored. The gold rods and silver spheres were also fabricated and immobilized with the Eu(III) chelates in the same strategy. The optical properties from the Eu–Au rod and Eu–Ag sphere complexes were determined as controls of the properties on the silver rods.</p><p>Polarized emission is regarded as an important feature for a fluorophore in biological and medical applications.4,19,20 Because the emission from a lanthanide chelate arises from a high-spin-high-spin transition,21,22 which involves multiple transition dipole moments that are energetically degenerated, it is generally unpolarized. Moreover, an extremely long lifetime may offer the lanthanide chelate enough time to make a rotational motion during its excited state so the emission is mostly unpolarized. Hence, there is a particular research interest to find a strategy that can be used to initiate a polarized emission from the lanthanide chelate. So far, some reports appear,23–25 in which the lanthanide chelates were mostly immobilized in aligned matrixes such as liquid crystals and stretched polymer films, or the single crystals of lanthanide complexes were used to restrict the motions and realize the polarized emission. However, neither of these approaches is inappropriate for the uses of lanthanide-based assays. A recent publication shows us another alternation,26a in which the near-field interaction on the metal nanoparticle can alter the transition moment of a fluorophore and finally initiate a polarized emission from the fluorophore. Therefore, aside from enhanced luminescence, we expected that the interactions for the Eu(III) chelates on the silver rods could initiate the polarized emission from the lanthanide chelates.</p><!><p>All reagents and spectroscopic grade solvents were used as received from Fisher or Sigma/Aldrich. Tris-(dibenzoylmethanate)mono(5-amino-1,10-phenanthroline) europium chelates were commercially available from Sigma/ Aldrich. Nanopure water (>18.0 MΩ∙cm−1) purified on a Millipore Milli-Q gradient system was used in all experiments.</p><!><p>In a modified seed growth method, the silver nanorods were fabricated in aqueous solutions with a defined aspect ratio.26–28 In brief, small silver nanoparticles were first generated as a seed solution to initiate the growth of silver rods. Typically, 20 mL aqueous solution with 0.25 mM AgNO3 and 0.50 mM sodium citrate was added by 1 mL of ice-cold 0.1 M NaBH4 solution with rigorous stirring. The color of solution was turned to yellow representing the formation of small silver nanoparticles, which had an average diameter of 4 nm. For the silver rod fabrication, 10 mL of aqueous solution with 0.25 mM AgNO3, 80 mM cetryltrimethylammoium bromide (CTAB), and 0.5 mM ascorbic acid were added to a 30 µL seed solution. The mixing solution was continuously stirred for 30 min. The color of the solution changed from yellow to green representing the formation of silver rods. The formed silver rods were collected by centrifugations followed by dispersing in 1 mL of mixing solvent of water/ethanol (v/v = 1) for depositing the silica layers on the external surfaces.</p><p>The gold nanorods and silver spheres were also fabricated and used as controls of silver rods. The gold rods were fabricated in the same strategy as the silver rods. In the strategy of our earlier report, the silver spheres were fabricated to have an average diameter of 40 nm.15 Like the silver rods, the gold rods and silver nanospheres were respectively redispersed in 1 mL of mixing solvent of water/ethanol (v/v = 1) for depositing the silica layers on the external surfaces.</p><!><p>In a modified Stöber method, thin silica layers were deposited on the metal nanoparticles including silver rods, gold rods, and silver spheres for immobilizing the europium chelates.29,30 Typically, 1 mL of metal nanoparticle solution was added by 1 mL of water/ethanol (v/v = 1) solution containing 1 × 10−3 M tetraethyl orthoorthosilicate. Subsequently, 10 µL of 30% ammonia solution was added dropwise under vigorous stirring. The solution was stirred overnight at room temperature to form the silica layers on the metal nanoparticles. The suspension solution was removed by centrifugation, and the residual solid was dispersed in 1 mL of a mixing solvent of N,N-dimethylformamide (DMF)/alcohol (v/v = 1/1) for immobilizing the europium chelates.</p><!><p>Tris(dibenzoylmethanate)mono(5-amino-1,10-phenanthroline) europium chelates were immobilized into the silica layers on the metal nanoparticles via physical absorptions.13 The Eu(III) chelates (1 × 10−5 M) were codissolved in a mixing solution of DMF/alcohol (v/v = 1/1) containing the silica-treated metal nanoparticles (1 × 10−8 M). The solution was stirred for 12 h at room temperature, and the suspension was removed by centrifugation. The residual solid was subsequently washed with DMF, ethanol, and water. The collected metal nanoparticles were dispersed in water for the ensemble spectral measurements.</p><!><p>Absorption spectra were conducted on a Hewlett-Packard 8453 spectrophotometer. Ensemble fluorescence spectra were recorded with a Cary Eclipse Fluorescence Spectrophotometer. Because of a slow decay time of Eu(III) chelate, a time-gated technique was used to collect the emission from the lanthanide, in which the gate pulse width was 0.5 ms and the delay time was 0.1 ms. All Eu(III) chelate-associated samples were excited at 380 nm for their emission spectral collections. In the current measurements, the polarized emission of Eu(III) chelate-associated samples were determined with two orthogonal polarizations, one parallel and one perpendicular to the excitation polarization vector.4 The emission anisotropies of Eu(III) chelate-associated samples were determined in the single-channel method under the same time-gate conditions.4 In general, the emission anisotropies can be expressed as,4 (1)r=I‖/I⊥−1I‖/I⊥+2 in which I‖ is the intensity of the emitted light when the emission polarizer is oriented parallel to the direction of the excitation polarization, whereas I┴ is the intensity of emitted light when the emission polarizer is perpendicular to the direction of the polarized excitation. The monochromator will have different transmission efficiency for vertically and horizontally polarized light, and as a result the excitation intensity can change when the excitation polarizer is rotated. Thus, in the single-channel method measurements, the emission intensities (I‖ and I┴) have been normalized to correct for the polarization sensitivity of the detection pathway in eq 2. (2)I‖/I⊥=IVVIHHIVHIHV IVV, IHH, IVH, and IHV are the emission intensities measured on the different directions for two polarizers installed on the excitation and emission sides. Angle-dependent luminescence from the Eu–meal complexes were measured by adjusting the angle between the two polarizers of excitation and emission sides and corrected using the emission from the free Eu(III) chelates in solution that was considered to be completely unpolarized.</p><!><p>For the TEM measurements, the nanoparticle samples were diluted to nanomolar concentrations followed by casting onto the copper grids (200 mesh) with standard carbon-coated Formvar films (200–300 Å). The samples were dried in air. TEM images were taken with a side-entry Philips electron microscope at 120 keV. The distributions of nanoparticle sizes were analyzed with Scion Image Beta Release 2 on the base on at least 200 images.</p><!><p>In this study, the silver nanorods were fabricated in a modified seed growth method.26–28 The achieved silver rods could be distinctly outlined by TEM images (part a of Figure 1) showing an average length of 80 nm and width of 18 nm. As a control, the gold rods were also fabricated in the similar strategy and outlined by TEM image showing an average length of 60 nm and width of 20 nm (part b of Figure 1). The aspect ratios of metal rods hence were estimated to be ca. 5 for the silver and ca. 3 for the gold. These metal rods were stabilized by cetryltrimethylammoium bromide (CTAB) on the surfaces. As expected, the metal rods exhibited distinct two plasmon absorption bands from the longitudinal and transverse directions, respectively, in aqueous solution (Figure 2), on which the bands from the silver rods were centered at 394 and 675 nm and the bands from the gold rods were centered at 536 and 982 nm. Tris(dibenzoylmethanate)mono(5-amino-1,10-phenanthroline) europium chelate was employed to immobilize on the metal rods for exploring the near-field interactions. It was observed that this Eu(III) chelate exhibited an absorption band and an emission band centered at 353 and 614 nm (Figure 3), respectively, close to the plasmon bands from the silver rods. Thus, there were sufficient spectral couplings between the Eu(III) chelates and silver rods at both the excitation and emission processes, and as a result the Eu(III) chelates in the near-field range from the metal rods were expected to have strong interactions with the silver rods leading to greatly improved optical properties. As the control of silver rods, the gold rods were observed to display the plasmon bands significantly shifting to the red region and, thus, could not sufficiently couple with the absorption and emission bands from the Eu(III) chelates, and the emission from the Eu(III) chelates on the gold rods should be less enhanced.</p><p>The Eu(III) chelates were immobilized in the silica layers on the metal rods. Because the Eu(III) chelates were required to localize in the near-field range, the silica layers on the metal rods were thin. Typically, the metal rods were deposited by thin silica layers through a hydrolysis reaction of tetraethyl orthosilicate.29,30 By the ratio of orthosilicate monomer over the metal rod in the reaction solution, the silica layer on the metal rod was controlled to be ca. 5 nm thick. The silane-treated metal rods were dispersed in a mixed solvent of DMF/ alcohol containing Eu(III) chelates to absorb the Eu(III) chelates from solution by physical interactions . The Eu–metal rod complexes were recovered by configuration and dispersed in water for the spectral measurements. It was difficult to outline the silica layers from the metal rods on the TEM images because of their thin thickness. But they could be identified by the emission spectral measurements. For the silane-treated metal rods, there was significant luminescence detectable after absorbing the Eu(III) chelates, whereas for the untreated metal rods there was no luminescence after absorbing the Eu(III) chelates supporting that there were the silica layers formed on the metal rods after the silane-treatments.</p><p>In the absorption process of Eu(III) chelates into the silane-treated metal rods, the Eu(III) chelates and silane-treated metal rods were codissolved at a molar ratio of 1000/1 in solution. According to the concentration change of Eu(III) chelate in solution prior to and after the absorption treatment, the consuming amount of Eu(III) chelate in solution was estimated, and thus, the amount of Eu(III) chelate absorbed on the metal rods was inferred. As a result, the number of Eu(III) chelate on one metal nanoparticle could be approximately estimated. This number was estimated to be ca. 650 on the silver rod and ca. 560 on the gold rod. The loading number of Eu(III) chelate on the silver rod corresponds to approximately 5 M, which is extremely high for most organic dyes in the solid matrixes. But for the Eu(III) chelates, there was no serious quenching to occur among them because of a large Stokes shift.1,2 Actually, most lanthanide-based nanoparticles are fabricated with high loading number of chelate for maximally increasing their brightness. The loading number of Eu(III) on the metal rod hence was considered to be acceptable in this study.</p><p>The Eu(III)–Ag rod complexes were observed to display two plasmon bands and the band maxima remained almost unchanged from those from the CTAB-coated silver rods, indicating the silver rods were chemically stable in the surfaces reactions. The Eu(III) chelates were immobilized on the metal rods but the immobilized Eu(III) chelate so the metal rods could not be determined by the absorption spectrum because of strong interference from the plasmon bands from the metal rods. Thus, the emission spectrum was used to determine the Eu(III) chelates on the silver rods. It was shown that upon excitation at 380 nm the Eu(III)–Ag rod complexes displayed an emission band centered at 610 nm (Figure 4), a 4 nm shift to blue in comparison with the free Eu(III) chelates. The emission band from the Eu(III)–Ag rod complexes was also significantly broadened, which was primarily due to the restriction of the movements of Eu(III) chelates after they were immobilized in the solid templates as well as the structural heterogeneity of Eu(III) chelates in the nearby environments.31 In fact, it has been widely reported that the emission band becomes broadened with the immobilization of fluorophores into the solid substrates in comparison with that in solution. These fluorophores include not only lanthanide chelates but also many other organic fluorophores. The breakdown of beta-diketonate by the acidic conditions in the silica is also considered as a possible reason causing such broadness. But because there is no significant change in the emission spectrum except the broadness with the immobilization of Eu(III) chelates in the silica, we suggest that the restriction of movement for the immobilized fluorophore in the solid substrate and influence from the heterogeneous surrounding is still the primary reason causing such a change in the band broadness.</p><p>The near-field interactions of Eu(III) chelates with the silver rods were supposed to explore by the optical property change for the Eu(III) chelates in the silica templates in the presence or absence of silver rods. In brief, the silver rods could be removed by adding several drops of 0.1 N NaCN solution into the Eu–Ag rod complex solution. With the dissolution of metal rods, the metal-free Eu(III) chelate-contained silica templates were consequently released.32 Apparently, it was observed that the color of plasmon from the metal rods disappeared progressively, and, simultaneously, the emission intensity from the Eu(III) chelates was decreased due to the loss of near-field interaction. Finally, the solution became completely colorless and the emission from the Eu(III) chelates was reduced to a minimum (Figure 4). However, the wavelength and broadness of the emission band from the Eu(III) chelates were not significantly altered with the dissolution of metal rods indicating that the Eu(III) chelates were still immobilized in the metal-free silica templates. To test the possible influence from adding NaCN to luminescence from the Eu(III) chelates in the silica templates, the emission from the Eu–silica templates were subsequently monitored by addition of NaCN aqueous solution. It was shown that the emission remained almost unchanged indicating that the emission from the Eu(III) chelates in the silica templates was not significantly influenced by NaCN in solution.</p><p>The emission intensity from the Eu–Ag rod complexes was reduced with the dissolution of metal rods representing that the emission from the Eu(III) chelates was indeed enhanced on the metal rods due to the near-field interactions. The enhancement efficiency was estimated by the ratio of the emission intensity from the Eu–Ag rod complexes prior to the NaCN treatment over the intensity after the treatment. The value was estimated to be 240-fold (part a of Figure 5). As the control, the Eu–Au rod complexes were also treated in the same strategy and the enhancement efficiency was estimated to be 14-fold. The enhancement for the Eu(III) chelates on the gold rods was much smaller than that on the Ag rods representing that the Eu(III) chelates had stronger interactions with the silver rods as expected early. Basically, the silver rods have their transverse and longitudinal plasmon bands close to the absorption and emission bands of Eu(III) chelates, which might lead to stronger interactions on the silver rods.</p><p>For a fluorophore, the near-field interaction at the emission process can increase its radiative rate and furthermore enhance its luminescence.9 However, the increased radiative rate of a fluorophore can be reflected by its decreased lifetime. The lifetimes from the Eu(III) chelates on the metal rods were collected by the time-gated method in this study. The decay curves from the Eu–Ag rod complexes and metal-free silica templates were presented in Figure 6 showing different decay rates. The curves were analyzed in terms of a two-exponential model, and the lifetime data were presented in average values (part b of Figure 5). It is shown that the free Eu(III) chelates in solution have a lifetime of 75 µs (not shown in figure). After immobilizing the Eu(III) chelates in the silica templates, the lifetime was increased to 0.5 ms, which was primarily due to the restriction of movements of Eu(III) chelates in the solid templates as discussed earlier.33 On the silver rods, the lifetime of Eu(III) chelates was dramatically reduced to be less than 10 µs. We suppose that it was primarily because of strong near-field interactions of Eu(III) chelates with the silver rods. As the control, the lifetime of Eu–Au rod complexes was also measured and estimated to be 70 µs. This value is shorter than the lifetime on the metal-free silica templates but much longer than the lifetime on the silver rods implying that the Eu(III) chelates had interactions with the gold rods but the interactions were much weaker than those with the silver rods. This result is approximately consistent with the observation achieved from the emission intensity for the Eu(III) chelates on the metal rods.</p><p>For a near-field interaction, a fluorophore is known to interact with a metal nanoparticle at either the excitation or emission process of the fluorophore. But the changed lifetime of the fluorophore by the metal nanoparticle can only reflect the interaction at the emission process.9 In this case, the lifetime of Eu(III) chelates on the silver rods was reduced by 50-fold (Figure 6). But the emission intensity was enhanced by 240-fold. The enhancement fold on the emission intensity is significantly larger than the decrease fold on the lifetime. Thus, it is suggested that the enhanced luminescence for the Eu(III) chelates on the silver rods should arise at least partially from the interactions at the excitation process. A sufficient spectral coupling between the absorption band from the Eu(III) chelates and the transverse plasmon band from the silver rods can support this speculation. In contrast, the enhanced luminescence from the Eu(III) chelates on the gold rods should arise almost completely from the emission process rather the excitation process due to the lack of sufficient spectral coupling at the excitation region.</p><p>It is important to understand how the near-field interaction on the metal rod can influence how the emission lines form the europium center for the chelate. It is known that the Eu(III) center in the chelates can accept the energy from all 5D orbits.1,2 However, although the Eu(III) centers in the chelates can emit a luminescence from the 5D1 level, the main transitions may originate from the 5D0 level that correspond to the main emission line at 614 nm (Figure 3). With the immobilization of Eu(III) chelates in the silica layers and interactions with the metal nanoparticles, the emission bands from the Eu(III) chelates were significantly broadened. As a result, the detail lines from the emission bands could not be distinguishable. Thus, it is difficult to understand the influence from the near-field interactions on the transitions of Eu(III) centers in the chelates. But because of sufficient spectral overlapping of absorption and emission bands from the Eu(III) chelates with the plasmon bands from the silver rods, we expect that in this case the Eu(III) chelates can interact efficiently with the silver rods at all of their energy levels.</p><p>To demonstrate the role of silver rods in the near-field interactions, the 40 nm silver spheres were also fabricated and the 5 nm thick silica layers were deposited on the silver spheres for immobilizing the Eu(III) chelates. Different from the silver rods, the silver spheres exhibited only one plasmon band centered at 460 nm15,16 localized between the bands of absorption and emission from the Eu(III) chelates. As a result, the spectral coupling on the silver spheres should be not sufficient as it is on the silver rods. Thus, it is no surprise that the Eu(III) chelates expressed weaker interactions with the silver spheres relative to the Eu(III) chelates on the silver rods. The enhancement efficiency on the silver spheres was determined to be 80-fold lower than that on the silver rods, and the lifetime was reduced to 30 µs, longer than the lifetime on the silver rods. Both data showed that the interactions of Eu(III) chelates on the silver spheres were weaker than those on the silver rods. According to reports, the silver ions that were dissolved from the metal rods were supposed to interact with the ligands from the chelates, which could probably enhance the emission from the Eu(III) chelates.13 But because most silver on the nanoparticles were presented as a noble metal rather than as ions, it is believed that the influence from the silver ions should be much less in comparison with the near-field interaction effect.</p><p>It is also known that the near-field interaction for a fluorophore on a metal nanoparticle is distance-dependent.9 When the fluorophore is localized too close to the metal nanoparticle, the absorption effect is a domain factor and the emission from the fluorophore is quenched, whereas with an increase of distance from the fluorophore to the metal nanoparticle but still within the near-field range, the scattering effect becomes a domain effect and the emission from the fluorophore becomes enhanced.9,14 Plenty of studies have reported on this topic. In this work, we deposited 5 nm thick silica layers on the metal rods and the Eu(III) chelates were physically absorbed in the silica to explore the near-field interactions. We believe that all Eu(III) chelates should be localized within the near-field range from the metal rod. The emission from a portion of Eu(III) chelates proximate to the metal surfaces would be quenched but the emission from most Eu(III) chelates would be enhanced. There should be a gradient through the silica layer from the metal-silica surface to the external surface in which the Eu(III) chelates could emit luminescence with the intensity from quenching to enhancement dependent on the distance. But the total intensity was observed to enhance due the near-field interactions. Both the quenching and enhancement for the Eu(III) chelates on the metal rods could be accompanied with a decrease of lifetime. In this study, we expected that the near-field interaction for the Eu(III) chelate on the silver rod could increase the radiative rate of Eu(III) chelate and furthermore increase the cyclic number for the excitation/emission of Eu(III) chelate in a period of time and increase the photon count rate in the measurement.</p><p>Polarized emission from a fluorophore is an important feature for its applications. But the lanthanides generally involve multiple transition dipole moments that are energetically degenerated during their emission. As a result, the emissions from the lanthanide dyes are a lack of polarization.23 In addition, extremely long lifetimes of the lanthanides may bring up their free rotational motions in the excited states leading to loss of polarized emission. Some efforts have been done to initiate the polarized emission from the lanthanides but few progresses were achieved.24,25 It is reported that the near-field interaction of a fluorophore with a metal nanoparticle can initiate a directional emission.26a In this study, we pursued to initiate a polarized emission from the Eu(III) chelates on the metal rods by the near-field interactions. The emission measurements were conducted in a single-channel method that was equipped with two orthogonal polarizations – one parallel and one perpendicular to the excitation polarization vector.4 The emission intensities (I‖ and I┴) have been normalized to correct for the polarization sensitivity of the detection pathway. The angle-dependent data were collected and presented in Figure 7. As expected, the metal-free Eu-contained silica templates displayed almost unpolarized emission (part a of Figure 7) because of their inherent characteristics, whereas the Eu–Ag rod complexes displayed a significant polarized emission (part b of Figure 7). As controls, the Eu–Au rod complexes and Eu–Ag sphere complexes were also determined under the same conditions. It was shown that the emission from the Eu–Au rod complexes was completely unpolarized but the emission from the Eu–Ag sphere complexes displayed a weak polarization. The emission anisotropies were calculated in eq 2. The value was 0.09 for the Eu–Ag rod complexes and 0.04 for the Eu–Ag sphere complexes. These values were lower in comparison with the anisotropies of most organic fluorophores in the solid substrates but acceptable in comparison with the anisotropies of the Eu(III) chelates in the aligned matrixes.23–25 The polarized emission from the Eu–Ag rod complexes or Eu–Ag sphere complexes should arise from the near-field interactions. A stronger interaction on the silver rod might result in a larger decrease in the lifetime of the Eu(III) chelate, and thus the rotational motions of Eu(III) chelate on the silver rod should be severely restricted in a shorter excitation period. In addition, a stronger interaction of Eu(III) chelate with the silver rod might probably break up more efficiently the multiple transition dipole moments of Eu(III) chelate and become energetically undegenerated. Consequently, the emission from the Eu(III) chelate became polarized. We also notice that the emission from the Eu(III) chelate on the gold rod was completely unpolarized, which was primarily due to weak near-field interaction.</p><p>We consider that the plasmon resonance from the silver rods can extend to far field and contribute the polarized emission from the Eu–Ag rod complexes.34,35 It is believed that when a fluorophores is coupled with a metal rod, the enhanced emission from the fluorophore–metal complex should have a component of plasmon energy by the metal nanoparticle extended to the far field at the same wavelength of fluorophore emission.9,26a The plasmon resonance from the metal nanoparticle is extended with the specific directions polarized,35 so the fluorophore–metal complexes can emit the luminescence with the polarization. Relative to the silver spheres, the silver rods are considered to emit the plasmon resonances with more specifically directional property that may bring up more polarized emission.</p><!><p>We studied the improved optical properties for the Eu(III) chelates on the metal nanorods due to the near-field interactions. The silver rods were fabricated in a wet chemical method followed by depositing thin silica layers on the surfaces. The Eu(III) chelates were followed by physically absorbing the silica layers on the silver rods. The silver rods displayed two plasmon bands, which had maxima close to the absorption and emission bands of the Eu(III) chelates. Thus, the Eu(III) chelates were expected to have strong interactions with the silver rods leading to greatly improved optical properties. The emission intensity could be enhanced up to 240-fold. Considering there were ca. 650 Eu(III) chelates on one silver rod, it was estimated that one Eu–Ag rod complex should be more than 1.5 × 105-fold brighter than one single free Eu(III) chelate. Because of the near-field interactions, the radiative rate of Eu(III) chelates on the silver rods is greatly increased leading to a dramatic decrease of lifetime from hundreds to several microseconds. However, this value is much longer than the lifetimes of most organic dyes, so the Eu–Ag rod complexes can be considered to use for time-resolved cellular assays. The gold rods or silver spheres were fabricated and conjugated with the Eu(III) chelates. The optical properties were measured as the controls of Eu–Ag rod complexes. It was found that both the gold rods and silver spheres had their plasmon bands only in part overlapping with the absorption and emission bands from the Eu(III) chelates leading to weaker near-field interactions. Consequently, the Eu–Au rod complexes and Eu–Ag sphere complexes had their optical properties less improved in comparison with the Eu–Ag rod complexes. It was also interesting to notice that the emission from the Eu(III) chelates on the silver rods was polarized, whereas the emission from others was almost not. This observation p supports that the sufficient near-field interactions for the fluorophores with the metal nanoparticles are essential for polarized initiation.</p><!><p>The authors declare no competing financial interest.</p>

PubMed Author Manuscript

Use of a Novel 5\xe2\x80\xb2-Regioselective Phosphitylating Reagent for One-Pot Synthesis of Nucleoside 5\xe2\x80\xb2-Triphosphates from Unprotected Nucleosides

The 5\xe2\x80\xb2-triphosphates are the building blocks for the enzymatic synthesis of DNAs and RNAs. This unit presents a protocol for the convenient synthesis of 2\xe2\x80\xb2-deoxyribo- and ribo-nucleoside 5\xe2\x80\xb2-triphosphates (dNTPs and NTPs) containing all the natural bases and the modified bases. This one-pot synthesis can also be applied to prepare the triphosphate analogs that contain sulfur or selenium atoms replacing the non-bridging oxygen atoms of the alpha phosphates of the triphosphates. These S- or Se-modified dNTPs and NTPs can be used to prepare diastereomerically-pure phosphorothioate nucleic acids (PS-NAs) or phosphoroselenoate nucleic acids (PSe-NAs, i.e., one type of selenium-derivatized nucleic acids: SeNA). Even without extensive purification, the synthesized dNTPs or NTPs are of high quality and can be directly used in DNA polymerization or RNA transcription. Synthesis and purification of the 5\xe2\x80\xb2-triphosphates, analysis and confirmation of natural and sulfur-or selenium-modified nucleic acids are described in this protocol unit.

use_of_a_novel_5\xe2\x80\xb2-regioselective_phosphitylating_reagent_for_one-pot_synthesis_of_nucleos

INTRODUCTION<!>BASIC PROTOCOL 1: ONE-POT SYNTHESIS OF NATIVE NUCLEOSIDE 5\xe2\x80\xb2-TRIPHOSPHATES (dNTP AND NTP)<!><!>Materials<!><!>BASIC PROTOCOL 2: ONE-POT SYNTHESIS OF NUCLEOSIDE 5\xe2\x80\xb2-(ALPHA-P-THIO) TRIPHOSPHATES<!><!>BASIC PROTOCOL 3: SYNTHESIS OF NUCLEOSIDE 5\xe2\x80\xb2-(\xce\xb1LPHA-P-SELENO) TRIPHOSPHATE<!><!>Background information<!>Application of the natural and modified nucleoside 5\xe2\x80\xb2-triphosphates<!>Critical parameters and troubleshooting<!>Anticipated results<!>Time Considerations

<p>The protocol described here is based on work published previously (Caton-Williams et al., 2011; Lin, et al., 2011; Caton-Williams et al., 2012) for the convenient synthesis of 2′-deoxyribo- and ribo-nucleoside 5′-triphosphates containing all the natural bases, the modified bases, and the S- or Se-modifications at the alpha phosphate positions. Nucleoside 5′-triphosphates (dNTPs and NTPs) play crucial roles in the synthesis of nucleic acids (DNA and RNA), and are involved in many biological regulations and pathways (Storz, 2002; Golubeva et al., 2008; Soutourina et al., 2011). In order to synthesize the natural and modified DNAs and RNAs enzymatically, the appropriate dNTPs and NTPs need to be prepared first.</p><p>Because of the presence of multiple functionalities (hydroxyl and amino groups) in the structure of the starting nucleosides, the nucleosides usually need to be protected. Protection and deprotection of these functionalities are achieved by several synthetic steps. To avoid the lengthy synthetic processes, Huang's research laboratory has developed a straightforward approach to conveniently synthesize the series of native and modified dNTPs and NTPs without the need for protecting the starting nucleosides. Basic Protocol 1 (Caton-Williams et al., 2011) describes the one-pot synthesis of the native 5′-triphosphates 8O and 9O (Figures 1 and 2). This protocol can also be applied to synthesize the 5′-triphosphates containing the modifications on the sugar and base moieties. The purification, analysis and confirmation of the native 5′-triphosphates are illustrated in Figures 3 and 4.</p><p>Basic Protocol 2 (Caton-Williams et al., 2012) describes the synthesis of the nucleoside 5′-(α-P-thio)triphosphates 8S and 9S (Figure 5) where a non-bridging oxygen atom of the alpha phosphate is replaced by a sulfur atom. Because of this modification at the alpha phosphate, diastereomers are formed. Typical HPLC profiles of diastereomers are illustrated in Figures 6 and 7. Basic Protocol 3 is a slight modification of the published article (Lin et al., 2011) and describes the synthesis of the nucleoside 5′-(α-P-seleno) triphosphates, 8Se and 9Se (Figure 8), where a non-bridging oxygen atom at the alpha phosphate is replaced by a selenium atom, resulting in diastereomeric mixtures, similar to the products of Basic Protocol 2. An HPLC profile demonstrating diastereomeric mixtures is shown in Figure 9. In all three scenarios, the facile synthesis is achieved by first reacting 2-chloro-1,3,2-benzodioxaphosphorin-4-one (salicyl phosphorochloridite, 1) with pyrophosphate (2) to generate a phosphitylating reagent in situ, which offers high regioselectivity at the 5′-hydroxyl group of the unprotected nucleoside. Following oxidation and hydrolysis, each 5′-triphosphate is synthesized by a one-pot synthesis.</p><p>The products from Basic Protocol 1 are the building blocks for the synthesis of natural nucleic acids, whereas Basic Protocol 2 and Basic Protocol 3 products can be used to synthesize diastereomerically-pure phosphorothioate and phosphoroselenoate nucleic acids, respectively. Even without the HPLC or ion-exchange purification, the precipitated crude 5′-triphosphates are of sufficient quality for direct DNA polymerization and RNA transcription. The one-pot synthesis developed by our laboratory is convenient. Since the reaction conditions are mild and the synthesis is cost-effective, this novel strategy can be used to synthesize the triphosphates with various modifications.</p><p>CAUTION: Carry out all reactions in a well-ventilated fume hood, and wear lab coat, gloves and protective glasses. All reactions should first be performed on a small scale.</p><!><p>5′-dNTP or 5′-NTP is synthesized as depicted in Figure 1. This protocol describes the preparation of deoxy- and ribo-nucleoside 5′-triphosphates without any protecting groups on the starting materials. The first step of the synthesis involves the reaction of 1 with 2 in the presence of DMF and tributylamine to form an intermediate 3 in situ (Figure 1), which selectively phosphitylates the 5′-hydroxyl group of the unprotected nucleoside (A, C, G T, or U). After subsequent displacement of the phenolic group, the nucleoside cyclic phosphite (6 or 7; the key intermediate) is formed. This intermediate can then be converted to the native or alpha-modified 5′-triphosphate using the appropriate oxidizing agent to afford the desires nucleoside 5′-triphosphates (8 or 9; see Figure 2). Iodine/water solution is the commonly used oxidizing reagent for preparing native 5′-dNTPs and 5′-NTPs.</p><!><p>2′-Deoxyadenosine monohydrate (Sigma-Aldrich)</p><p>2′-Deoxycytidine monohydrochloride (ChemGenes Corporation)</p><p>2′-Deoxyguanosine monohydrate (ChemGenes Corporation)</p><p>2′-Thymidine 99% (Alfa Aesar)</p><p>Adenosine (Sigma-Aldrich)</p><p>Cytidine (ChemGenes Corporation)</p><p>Guanosine (ChemGenes Corporation)</p><p>Uridine (ChemGenes corporation)</p><p>Tributylammonium pyrophosphate (Sigma-Aldrich)</p><p>2-Chloro-1,3,2-benzodioxaphosphorin-4-one (salicyl phosphorochloridite, Sigma-Aldrich)</p><p>Anhydrous N,N-dimethylformamide (DMF, Sigma-Aldrich)</p><p>Tributylamine (TBA, Sigma-Aldrich)</p><p>Argon gas (Dried)</p><p>Methanol (MeOH)</p><p>Dichloromethane (methylene chloride, CH2Cl2)</p><p>Iodine solution (Glen Research)</p><p>Water (deionized)</p><p>Isopropanol (ip-OH)</p><p>Ammonium hydroxide (NH4OH)</p><p>Sodium chloride (3 M, NaCl)</p><p>200 PROOF pure ethanol (KOPTEC)</p><p>2′-Deoxyadenosine 5′-triphosphate (dATP, Epicentre)</p><p>2′-Deoxycytidine 5′-triphosphate (dCTP, Epicentre)</p><p>2′-Deoxyguanosine 5′-triphosphate (dGTP, Epicentre)</p><p>2′-Thymidine 5′-triphosphate (TTP, Epicentre)</p><p>Adenosine 5′-triphosphate (ATP, Epicentre)</p><p>Cytidine 5′-triphosphate (CTP, Epicentre)</p><p>Guanosine 5′-triphosphate (GTP, Epicentre)</p><p>Uridine 5′-triphosphate (UTP, Epicentre)</p><p>Oven</p><p>15-, 10- and 5-mL round-bottom flasks</p><p>8 × 1.5-mm magnetic stir bar</p><p>Septa</p><p>Parafilm</p><p>High vacuum / vacuum pump</p><p>Balloons</p><p>1 ml syringes (Norm Ject)</p><p>Needles IM 1½ 23GTW (Becton Dickinson)</p><p>Magnetic stir plate</p><p>Silica-coated thin-layer chromatography (TLC) plate with fluorescent indicator</p><p>Kieselgel 60F254 (Dynamic Adsorbents Inc. and Sorbent Tech.)</p><p>UV lamp</p><p>3 ml syringes</p><p>Disposable glass pipette (9″)</p><p>15- or 50-mL tube (Falcon)</p><p>Marker</p><p>−80 or −20° C Freezer</p><p>UV spectrophotometer</p><p>HPLC System</p><p>Centrifuge</p><p>Additional reagents and equipment for thin-layer chromatography (TLC: APPENDIX 3D)</p><!><p>NOTE: Dry all glassware in an oven and perform all reactions in an argon atmosphere. An argon atmosphere can be carried out through the use of a rubber septum that seals the flask, and an argon-filled balloon is inserted into the septum. Liquid chemicals are added using a syringe inserted through a rubber septum. A slightly positive pressure is maintained in the system to provide an anhydrous and oxygen-free atmosphere.</p><!><p>To a 10- or 15-mL oven-dried, round-bottom flask, add a dried 8 × 1.5-mm magnetic stir bar.</p><p>Weigh 20 mg of starting nucleoside (0.07 to 0.08 mmol) directly into the flask.</p><p>For the deoxynucleoside as starting nucleoside: 2′-deoxyadenosine 4a, 2′-deoxycytidine 4c, 2 ″-deoxyguanosine 4g, thymidine 4t. For the ribonucleoside as starting nucleoside: adenosine 5a, cytidine 5c, guanosine 5g, uridine 5u.</p><p>NOTE: dissolve adenosine in a mixed solvent of 0.16 mL DMF and 0.06 mL DMSO, and guanosine in a mixed solvent of 0.11 mL DMF and 0.12 mL DMSO).</p><p>To another 10-mL oven-dried round-bottom flask, add a dried 8 × 1.5-mm magnetic stir bar and directly weigh 73 mg of tributylammonium pyrophosphate 2 (0.16 mol, 2 eq.) into the flask.</p><p>Seal each flask with a cream rubber septum and wrap with parafilm.</p><p>Place each flask on high vacuum to dry for over 2 h at room temperature.</p><p>Prepare argon filled balloons by using parafilm to wrapping a deflated balloon to the top end of a 1mL syringe.</p><p>Inflate two balloons with argon and insert into each septum of the flasks.</p><p>Quickly weigh 20 mg of 2-chloro-4-H-1,3,2-benzodioxaphosphorin-4-one 1 (0.1 mmol, 1.2 eq.) into a 5- or 10-mL oven-dried round-bottom flask containing a 8 × 1.5-mm magnetic stir bar.</p><p>Insert an argon filled balloon to the septum and purge for 30 seconds.</p><p>Insert another inflated balloon into the cap of the anhydrous DMF bottle and purge. Be careful not to contaminate the DMF solvent during transfer.</p><p>Using a 1 mL syringe equipped with a 1½ ″ needle, transfer 0.2 mL purged DMF to the flask containing 1.</p><p>Place the flask on a magnetic stirring plate and stir to dissolve.</p><p>Using a 1 mL syringe equipped with a 1 ½ ″ needle, transfer 0.2 mL purged DMF to the flask containing dried 2 (step 3). Stir to dissolve, keeping the solution under an argon atmosphere.</p><p>Using another clean syringe, add 0.3 mL (approximately 18 eq) of dried TBA to the flask containing the dissolved pyrophosphate 2.</p><p>Using the same syringe, transfer the solution in step 14 to the septum of the flask containing 1 and let the reaction stir for 30 min.</p><p>Remove the nucleoside from the high vacuum and insert an argon filled balloon on the septum of the flask.</p><p>Use a syringe to inject 0.2 mL dry DMF to dissolve the nucleoside.</p><p>Slowly (for 5 min) inject the reaction solution in step15 (mixture of 1 and 2) into the dissolved nucleoside and let the reaction stir vigorously for 2 h.</p><p>Monitor the formation of the intermediate 6 (or 7) by TLC (using standard silica gel plates, Dynamic Abs., APENDIX 3D) and eluent: 10% MeOH in CH2Cl2 for the intermediates of compound 4a and 4t; 15% MeOH in CH2Cl2 for the intermediates compound 5a and 5u; and 20% MeOH in CH2Cl2 for the intermediates of compound 4c, 4g, 5c and 5g (to monitor the consumption of the nucleoside)</p><p>The nucleoside cyclic phosphite intermediates 6(7) are visualized by UV lamp (254 nm) as fluorescent spots [Rf = 0.20, 0.13, 0.20, and 0.3, respectively for the 2′-deoxynucleosides (A, T, C, G) and 0.19, 0.12, 0.16, 0.12, respectively for the ribonucleosides(A, U, C, G)].</p><p>NOTE: under these reaction conditions the starting nucleosides are approximately 70 % completed according to TLC observations.</p><p>Using a syringe, inject 1 mL iodine solution(0.02 M) to the reaction containing the nucleoside reaction until a permanent brown color is maintained, similar to that of the iodine solution.</p><p>Let the reaction continue with stirring for another 30 min. If the reaction solution becomes colorless, add drop wisely more iodine solution to main the brown color during this time. The oxidized products are not normally monitored by TLC in this step, since it is not stable. After this oxidation step, using a 3 mL syringe with a 1 ½ ″ needle, inject 2 volumes of water (twice the reaction volume) and let the cyclic phosphate hydrolyze for 1.5 h under stirring</p><p>Monitor the formation of the product by TLC using standard silica gel plates (Sorbent Tech., APENDIX 3D) and the eluent: 5:3:2 (v/v/v) iso-Propanol/NH4OH/water.</p><p>The product is visualized by UV lamp (254 nm) (Rf = 0.47, 0.37, 0.35, 0.42, respectively for the deoxynucleoside 5′-triphosphates (8O; A, T, C, G) and 0.35, 0.28, 0.23 and 0.30 for the ribonucleosides 5′-triphosphates (9O; A, U, C, G).</p><p>After the reaction is completed, use a disposable glass pipette (9″) to transfer the resulting solutions into two 15-mL tubes, or one 50-mL tube. Use a marker to label each tube.</p><p>For a 1 mL of reaction solution, add 0.1 mL of 3 M NaCl to each tube and shake, followed by the addition of 3 mL ethanol. Note the volumes and use this information to calculate the amount of NaCl and ethanol to use for your volume measured.</p><p>Place the two tubes at either −80 °C or −20 °C for 1 h and centrifuged at approximately 3000 rpm for 30 min.</p><p>[NOTE: A refrigerated centrifuge (4°C) gives better results than a non-refrigerated centrifuge. A refrigerated centrifuge (−10 °C) is preferred.]</p><p>Using a glass pipette, transfer the supernatant to another tube and air dry (for 30 min) the white residues by slanting on a 5- or 10-degree incline to remove excess ethanol.</p><p>Re-dissolve the dried pellet in 200 μL deionized water and repeat step 25–27 to reprecipitate the triphosphate.</p><p>Determine the concentration of the triphosphate sample using a UV-vis Spectrophotometer.</p><p>The crude 5′-triphosphate (dNTP or NTP) is now ready for use in DNA polymerization or RNA transcription.</p><p>For product with high purity, centrifuge the solution of the crude product to remove any solid particles and use the supernatant to conduct HPLC purification.</p><p>Purify the crude nucleoside 5′-triphosphate on reversed-phase HPLC (RP-HPLC) using the following recommended conditions:</p><p>Column: 21.2 × 250–mm Welchrom C18 (or Ultisil C18)</p><p>Buffer A: 20 mM TEAA buffer, pH 7.1</p><p>Buffer B: 50% CH3CN in buffer A</p><p>Gradient: 0% to 40% buffer B over 20 min</p><p>Flow rate: 6 mL/min</p><p>Detection wavelength: 260 nm.</p><p>Collect the nucleoside 5′-triphosphate fractions and lyophilize them.</p><p>Dissolve the residue in 200 μL deionized water and precipitate it with NaCl (20 μL, 3 M)/ethanol (660 μL) to afford the 5′-triphosphate as the sodium salt.</p><p>A typical profile of RP-HPLC analysis of crude thymidine 5′-triphosphate is shown in Figure 3.</p><p>After the RP-HPLC purification, the 5′-triphosphates are characterized by 1H-NMR, 31P-NMR and high-resolution mass spectrometry (HR-MS). (Expected data is shown for the 2′-deoxynucleoside 5′-triphosphates.)</p><p>8aO (7.3 mg, 19% yield) as the triethylammonium salt, followed by re-precipitation with NaCl/ethanol to afford the sodium salt. 1H NMR (400 MHz, D2O): δ 8.50 (s, 1H, H-2), 8.27 (s, 1H, H-8), 6.54 (t, J = 6.5 Hz, 1H, H-1′), 4.41 (m, 1H, H-3′), 4.31 (m, 1H, H-4′), 3.88 (m 1H, H-5′), 3.24 (m 1H, H-5′), 2.84 (m, 1H, H-2′), 2.61 (m, 1H, H-2′); 31P NMR (162 MHz, D2O): δ −22.79 (t, Jβ = 19.60 Hz, 1P, β-P), −11.21 (d, Jα = 19.60 Hz, 1P, α-P), −9.49 (d, Jγ = 19.60 Hz, 1P, γ-P); UV (H2O): λmax = 259 nm; HR-MS (ESI): molecular formula C10H16N5O12P3; [M-H+]−: 489.9938 (calculated: 489.9936).</p><p>8cO (18.8 mg, 46% yield) as the triethylammonium salt, followed by re-precipitation with NaCl/ethanol to afford the sodium salt. 1H-NMR (400 MHz, D2O): δ 7.98 (d, J = 7.6 Hz, 1H, H-6), 6.34 (t, J = 6.0 Hz, lH, H-1′), 6.14 (d, J = 7.6 Hz, 1H, H-5), 4.63 (br, 1H, H-3′), 4.21 (m, 2H, H-4′ & 5′), 3.65 (m, 1H, H-5′), 2.39 (m, 1H, H-2′), 2.33 (m, 1H, H-2′); 31P-NMR (162 MHz, D2O): δ −22.65 (t, Jβ = 19.60 Hz, 1P, β-P), −10.85 (d, Jα = 19.60 Hz, 1P, α-P), −10.16 (d, Jγ = 19.60 Hz, 1P, γ-P); UV (H2O): λmax = 271 nm; HR-MS (ESI): molecular formula C9H16N3O13P3; [M-H+]−: 465.9814 (calculated: 465.9817).</p><p>8gO (11.5 mg, 30% yield) as the triethylammonium salt, followed by re-precipitation with NaCl/ethanol to afford the sodium salt. 1H-NMR (400 MHz, D2O): δ 8.12 (s, 1H, H-8), 6.33 (t, J = 6.8 Hz, lH, H-1′), 4.71 (m, 1H, H-3′), 4.21 (m, 3H, H-4′& 5′), 2.84 (m, 1H, H-2′), 2.52 (m, 1H, H-2′); 31P-NMR (162 MHz, D2O): δ -22.73 (t, Jβ = 19.44 Hz, 1P, β-P), −11.10 (d, Jα = 19.44 Hz, 1P, α-P), −9.50 (d, Jγ = 19.44 Hz, 1P, γ-P); UV (H2O): λmax = 252 nm; HR-MS (ESI): molecular formula C10H16N5O13P3; [M-H+]−: 505.9893 and calculated: 505.9885.</p><p>8tO (15 mg, 39% yield) as the triethylammonium salt, followed by re-precipitation with NaCl/ethanol to afford the sodium salt. 1H-NMR (400 MHz, D2O): δ 7.75 (s, 1H, H-6), 6.35 (t, J = 6.8 Hz, 1H, H-1′), 4.69 (m, 1H, H-3′), 4.22 (m, 3H, H-4′& -5′), 2.38(m, 2H, H-2′), 1.94 (s, 3H, Me-5); 31P-NMR (162 MHz, D2O): δ −21.22 (t, Jβ = 18.6 Hz, 1P, β-P), −10.50 (d, Jα = 18.6 Hz, 1P, α-P), −6.65 (d, Jγ = 18.6 Hz, 1P, γ-P); UV (H2O): λmax = 267 nm; HR-MS (ESI): molecular formula C10H17N2O14P3; [M-H+]−: 480.9831 (calculated: 480.9820).</p><!><p>Basic protocol 1 describes the synthesis of native 5′ triphosphates (Figures 1 and 2). In Basic protocol 2, this synthetic strategy is applied to the preparation of nucleoside 5′-(alpha-P-thio)triphosphates (5′-dNTPαS and 5′-NTPαS). In this protocol, the oxidation step utilizes a sulfurizing agent as the oxidant instead of iodine in step 3 (Figure 1). Sulfurization of the nucleoside cyclic phosphite 6 (7) is performed using 3-[(dimethylaminomethylidene)amino]-3H-1,2,4,dithiazol-e-3-thione (sulfurizing Reagent II), followed by hydrolysis to afford the crude dNTPαS and NTPαS analogs as mixtures of Sp and Rp diastereomers (Figure 6), with yields up to 60%. When using this synthetic strategy, only one diastereomer of dGTPαS and GTPαS is observed, and can be confirmed by comparing with commercially available compounds as the standards (Figure 7).</p><!><p>Acetonitrile, anhydrous 99.8% (Sigma-Aldrich)</p><p>Pyridine (Sigma-Aldrich)</p><p>3-[(Dimethylaminomethylidene)amino]-3H-1,2,4,dithiazol-e-3-thione (Sulfurizing Reagent II) from Glen Research</p><p>2′-Deoxyguanosine 5′-(α-P-thio)triphosphates (dGTPαS, TriLinks Inc.)</p><p>Guanosine 5′-(α-P-thio)triphosphates (GTPαS, TriLinks Inc.)</p><p>For the synthesis of the deoxynucleoside 5′-(alpha-P-thio) triphosphates and the ribonucleoside 5′-(alpha-P-thio) triphosphates, follow the synthesis in Basic Protocol 1 from step 1 to step 19.</p><p>Weigh 33 mg of Sulfurizing Reagent II (0.16 mmol, 2 eq) directly into a 5-mL round-bottom flask.</p><p>Seal the flask with a septum, wrap it with parafilm and dry it on high vacuum for over 2 h.</p><p>Insert an argon filled balloon on the septum.</p><p>Using a 1 mL syringe with a 1 ½″ needle, transfer 0.3 mL of dried pyridine to Sulfurizing Reagent II.</p><p>Use another 1 mL syringe and add 0.2 mL of anhydrous acetonitrile to the flask. Stir to dissolve (a yellow solution results).</p><p>Using the same syringe, slowly transfer this reaction mixture to the flask containing the cyclic phosphite (step 19 of Basic Protocol 1).</p><p>Let the reaction continue with stirring on a magnetic stir plate for 2 h. NOTE: The cyclic phosphate is not normally monitored by TLC in this step, since it is unstable.</p><p>Add two volumes of water (twice the reaction volume) to the reaction flask and let stirring continue for another 2 h to afford the desired 5′-triphosphate 8S or 9S.</p><p>Monitor the formation of the product by TLC (APENDIX 3D) using the eluent: 5:3:2 (v/v/v) ip-OH/NH4OH/water.</p><p>The products are visualized by UV lamp (254 nm) [Rf = 0.58, 0.45, 0.43, and 0.52, respectively for the deoxynucleoside 5′-triphosphates (8S; A, T, C, G) and 0.47, 0.33, 0.32, and 0.40 for the ribonucleosides 5′-triphosphates (9S; A, T, C, G)].</p><p>After the reaction is completed, use a disposable glass pipette to transfer the resulting solutions into two argon-purged 15-mL tubes, or one 50-mL tube. Use a marker to label each tube.</p><p>Measure the volume and calculate the amounts of 3 M NaCl (10% of the volume) and ethanol (3 fold of the volume) for the precipitation. NOTE: purge the NaCl and ethanol with argon for over 5 min prior to use.</p><p>Place the two tubes at either −80 °C or −20 °C for 1 h and centrifuged at approximately 3000 rpm for 30 min.</p><p>A refrigerated centrifuge (−10 °C) is preferred.</p><p>Using a glass pipette, transfer the supernatants to another tube and air dry the yellow residues by slanting on a 5- or 10-degree incline to remove excess of ethanol for over 30 min.</p><p>Re-dissolve the dried pellet in 1 mL deionized water and determine the concentration with a UV-vis Spectrophotometer.</p><p>The crude product is now ready for use in DNA polymerization or RNA transcription.</p><p>Keep the product under argon in −80 °C for extended storage.</p><p>Ensure to centrifuge the crude sample prior to the HPLC purification.</p><p>For a high purity compound, purify the crude 5′-triphosphate on reversed-phase HPLC using the following recommended conditions:</p><p>Column: 4.6 × 250 mm Welchrom C18 (or Ultisil C18)</p><p>Buffer A: 20 mM TEAA buffer, pH 6.5</p><p>Buffer B: 50% CH3CN in buffer A</p><p>Gradient: 0% to 20% buffer B over 25 min</p><p>Flow rate: 1 mL/min</p><p>Detection wavelength: 260 nm.</p><p>Collect the individual HPLC peaks of both diastereomers of the nucleoside 5′-triphosphate and evaporate the solvent by lyophilization.</p><p>A typical HPLC profile of uridine 5′-(alpha-P-thio)triphosphate is shown in Figure 6.</p><p>Only one diastereomer of dGTPαS or GTPαS is obtained with this protocol (Figure 7).</p><p>Further analyses of the 5′-triphosphates can be performed and compared with the commercially available native dNTPs.</p><p>Confirm the integrity of all synthesized nucleoside 5′-(alpha-P-thio) triphosphates by HR-MS analysis (Table 1).</p><!><p>Basic Protocol 1 describes the general procedure for the synthesis of nucleoside 5′-(α-P-seleno)triphosphates (Figure 1). Since oxygen, sulfur and selenium are in the same family of the periodic table, a suitable selenizing reagent can be used to introduce selenium (Basic Protocol 3; Figure 8) replacing a non-bridging oxygen atom at the alpha phosphate of the triphosphate. BTSe (commercially available at SeNA Research, Inc.), a selenium-introducing oxidant, can be used to introduce the selenium functionality during the oxidative step. The synthesis yields of the nucleoside 5′-(α-P-seleno)triphosphates are generally greater than 30%. The 5′-triphosphate products of Basic Protocol 3 are diastereomeric mixtures, similar to the products in Basic Protocol 2.</p><!><p>Dioxane (Sigma-Aldrich)</p><p>Triethylamine (Sigma-Aldrich)</p><p>3H-1,2-Benzothiaselenol-3-one (BTSe, SeNA Research, Inc.)</p><p>For the synthesis of the 2′-deoxynucleoside 5′-(alpha-P-seleno) triphosphates and the ribonucleoside 5′-(alpha-P-seleno) triphosphates, follow the synthesis in Basic Protocol 1 from step 1 to step 19.</p><p>Weight 35 mg of 3H-1,2-benzothiaselenol-3-one (BTSe, 0.16 mmol, 2 eq.) directly into a 5- or 10-mL round-bottom flask.</p><p>Seal the flask with a septum, wrap it with parafilm and dry it on high vacuum for over 2 h.</p><p>Insert an argon filled balloon on the septum.</p><p>Using a 1 mL syringe equipped with a 1 ½″ needle, transfer 0.3 mL of dioxane to dissolve BTSe.</p><p>Use another 1 mL syringe and add 0.2 mL of dried triethylamine (TEA) to the flask and stir to dissolve (a yellow solution or suspension results).</p><p>Using the same syringe, slowly transfer this reaction mixture to the flask containing the cyclic phosphite (step 19 of Basic Protocol 1).</p><p>Let the reaction continue with stirring on a magnetic stir plate for 2 h. NOTE: The cyclic phosphate is not normally monitored by TLC in this step, since it is unstable.</p><p>Add two volumes of water (twice the reaction volume) to the reaction flask and let stirring continue for another 2 h to afford the desired 5′-triphosphates 8Se or 9Se.</p><p>Monitor the formation of the product by TLC (APENDIX 3D) using the eluent: 5:3:2 (v/v/v) ip-OH/NH4OH/water.</p><p>The product is visualized by UV lamp (254 nm) as a fluorescent spot [Rf = 0.67, 0.53, 0.52, and 0.58, respectively, for the deoxynucleoside 5′-triphosphates (8Se; A, T, C, G) and 0.54, 0.40, 0.38 and 0.48 for the ribonucleosides 5′-triphosphates (9Se; A, T, C, G)].</p><p>After the reaction is completed, use a disposable glass pipette (9″) to transfer the resulting solutions into two argon purged 15-mL tubes,or one 50-mL tube. Use a marker to label each tube.</p><p>Measure the volume and calculate the amounts of 3 M NaCl (10% of the volume) and ethanol (3 fold of the volume) for the precipitation.</p><p>NOTE: purge the NaCl and ethanol with argon for over 10 min prior to use.</p><p>Place the two tubes at either −80 °C or −20 °C for 1 h and centrifuged at approximately 3000 rpm for 30 min.</p><p>A refrigerated centrifuge (−10 °C) is preferred.</p><p>Using a glass pipette, transfer the supernatant to another tube and air dry the yellow residue by slanting at a 5- or 10 degree incline to remove excess of ethanol for over 30 min.</p><p>Re-dissolve the dried pellet in 0.6 mL deionized water and determine the concentration with a UV-vis Spectrophotometer.</p><p>The crude product is now ready for use in DNA polymerization or RNA transcription.</p><p>Keep the product under argon in −80 °C for extended storage.</p><p>Ensure to centrifuge the crude sample prior to the HPLC purification or analysis.</p><p>For a high purity compound, purify the crude 5′-triphosphate on reversed-phase HPLC using the following recommended conditions:</p><p>Column: 4.6 × 250 mm Welchrom C18 (or Ultisil C18)</p><p>Buffer A: 20 mM TEAA buffer, pH 6.5</p><p>Buffer B: 50% CH3CN in buffer A</p><p>Gradient: 0% to 30% buffer B over 30 min</p><p>Flow rate: 6 mL/min</p><p>Detection wavelength: 260 nm.</p><p>Collect the individual HPLC peaks of both diastereomers of the nucleoside 5′-triphosphate and evaporate the solvent by lyophilization.</p><p>A typical HPLC profile of guanosine 5′-(alpha-P-seleno) triphosphates is shown in Figure 9.</p><p>Further analyses of the 5′-triphosphates can be performed and compared with the commercially available native dNTPs and NTPs.</p><p>After the RP-HPLC purification, the integrity of all synthesized nucleoside 5′-(alpha-P-seleno)triphosphates is confirmed by 1H-NMR, 13C-NMR, 31P-NMR and HR-MS. (The expected data is shown for the ribo-nucleoside 5′-(alpha-P-seleno)triphosphates; Tables 2–5.)</p><!><p>Nucleoside 5′-triphosphates (NTPs and dNTPs) play key roles in biochemistry, molecular biology and medicine (Eckstein, 1985; Bogdanov et al., 2010). They are the essential building blocks to the synthesis of nucleic acids (DNA and RNA) both in vivo and in vitro and depend on template, primer and polymerases to perform their functions. To further understand their roles played in nucleic acids and protein regulations in vitro and in vivo and their biological and medicinal significances, there are urgent needs to synthesize their analogs. The native and modified nucleoside triphosphates are often prepared via chemical synthesis of phosphorylated derivatives and analogs. They are very challenging to synthesize chemically and isolate in high purity (>99%). Burgess and Cook (review article) have discussed the general practical strategies and problems associated with preparing, isolating, characterizing, and storing of 5′-triphosphates (Burgess and Cook, 2000).</p><p>The first chemical synthesis of nucleoside 5′-triphosphate was achieved over six decades ago (Baddiley, 1949). Currently, there are numerous strategies (Zou et al., 2005; Horhota, et al., 2006; Sun, et al., 2008; Warnecke and Meier, 2009; Schultheisz et al., 2010; Jansen, 2011) developed to synthesize nucleoside 5′-triphosphates, but no one is universal. This scenario is still challenging largely due to the multiple functionalities (i.e., 5′-, 2′-, and 3′-sugar hydroxyl groups and nucleobase amino groups) of the natural nucleosides as well as the modified nucleosides containing other functionalities. Therefore, from the starting nucleosides to the final 5′-triphosphate products, the functional groups need to be introduced and removed, causing longer synthesis steps (Burgess and Cook, 2000; Wu et al., 2004; Sun et al., 2008). Protection of these groups is necessary in order to minimize the formation of by-products and region-isomers, and to ensure good yields and minimum by-products. The major region-isomers and by-products in the 5′-triphosphate synthesis are the 3′- and 2′-triphosphates, in addition to the mono-, di and oligo-phosphates, which are difficult to remove during the purification process. Despite many synthetic strategies that have been developed, such as the "one-pot, three-step" method developed by chemists (Yoshikawa et al., 1967; Ludwig and Eckstein 1989; He et al., 1998; Zou et al., 2005; Cheek, 2008), there is still a pressing need to develop convenient strategies for synthesizing nucleoside 5′-triphosphates from unprotected nucleosides with high 5′-regioselectivity.</p><p>To address the concern of regioselectivity, thereby avoiding the protection and deprotection steps, and simplifying the triphosphate synthesis, Huang's research group has generated a mild and selective phosphitylating reagent to distinguish these functionalities, showing greater selectivity toward the 5′-hydroxyl group of the starting nucleoside. This mild reagent is attained through the reaction of the highly reactive salicyl phosphorochloridite and pyrophosphate, affording a weak, bulky phosphitylating agent, when compared to salicyl phosphorochloridite (Caton-Williams et al., 2011). This bulky phosphitylating reagent can offer a better selectivity to distinguish the primary 5′-OH group from the secondary 2′- and 3′-OH groups. Furthermore, at reduced temperature (0 or −10 °C), the 5′-selectivity is further increased, but the reaction requires a longer time to consume the nucleosides (Gillerman and Fischer, 2010). After oxidation and hydrolysis, the nucleoside 5′-triphosphates are synthesized.</p><p>The strategy developed to synthesize native nucleoside 5′-triphosphates can be extended to the one-pot synthesis of the modified triphosphates starting from nucleoside derivatives, such as ethanodeoxyadenosine (a base modified deoxyadenosine) (Caton-Williams et al., 2011) and the S or Se-alpha phosphate. With the appropriate oxidizing reagent, either a sulfur or a selenium atom can be substituted for a non-bridging oxygen atom at the alpha-phosphate, resulting in Sp and Rp diastereomers of modified 5′-triphosphates (Ludwig and Eckstein 1989; He, et al., 1998; Lin and Shaw, 2000; Carrasco and Huang, 2004; Carrasco, et al., 2005; Brandt, et al., 2006).</p><p>It was reported that polymerases recognize the Sp diastereomer and not the Rp (Romaniuk and Eckstein, 1982; Eckstein and Gindl, 1983), therefore HPLC separation of the diastereomeric mixture is not necessary for the enzymatic synthesis of modified nucleic acids (PS-NA and PSe-NA). This is an important step toward therapeutic applications of the modified nucleic acids (Mori et al., 1989; Levin, 1999; Juliano et al., 2008; Lin et al., 2009, Gan, et al., 2011) and X-ray structural determination by using the Se-derivatized nucleic acids (SeNA; Buzin et al., 2004; Carrasco et al., 2004; Salon et al., 2007; Caton-Williams and Huang, 2008; Salon et al., 2008; Sheng et al., 2008; Hassan et al., 2009; Lin et al., 2011).</p><!><p>The strategy developed by Huang's research group to synthesize natural and modified nucleoside 5′-triphosphates is straightforward and convenient. Such synthesis will contribute tremendously to the fields of biochemistry, molecular and cellular biology, as well as medicine. It is quite useful for researchers interested in small scale synthesis of 5′-triphosphates to perform various molecular biology applications, such as polymerase chain reaction (PCR), real-time PCR, cDNA synthesis, primer extension, nick translation, DNA sequencing, DNA labeling, and RNA transcription.</p><p>The sulfur- and selenium-derivatized nucleoside 5′-triphosphates possess unique properties. Nucleoside 5′-(α-P-thio)triphosphates are the building blocks for the enzymatic synthesis of phosphorothioate nucleic acids (PS-NAs). Because of their bioavailability and nuclease resistance properties, PS-NAs have shown to be promising as potential therapeutics in antisense DNA, siRNA and microRNA to selectively inhibit gene expression by mRNA inactivation (Kunkel et al, 1981; Levin, 1999; Juliano et al., 2008).</p><p>Phosphorothioate modification has been used in combination with other modifications, such as boranophosphates (Summerton et al., 1997; Lorenz et al., 1998; Krishna and Caruthers, 2011) to further increase their usefulness as active components of drugs and mechanistic probes. Phosphoroselenoate modification can be introduced into DNA and RNA through enzymatic synthesis, utilizing nucleoside 5′-(α-P-seleno)triphosphates to prepare SeNA. Although phosphoroselenoate nucleic acids have not been well studied in therapeutic applications as its sulfur counterpart, there is great potential in therapeutics. Phosphoroselenoate modification, however, has great potential in X-ray crystallography, contributing enormously to the phase problem determination in structure study of non-coding RNAs and protein-RNA complexes, as well as DNAs (Ferre-D'Amare et al. 1998; Ke and Doudna, 2004; Keel et al., 2007; Ferre-D'Amare, 2010; Koldobskaya et al., 2011).</p><p>Such a convenient synthetic strategy for 5′-triphosphates will contribute greatly to fundamental and applied research in cellular and molecular biology.</p><!><p>Small-scale reaction should be carried out first. It is essential that all glassware, starting materials and reagents are thoroughly dried. During the synthesis process, it is important to keep the reaction environment dry and free of oxygen. During the ethanol precipitation step, it is important to purge the ethanol thoroughly with argon prior to use. To ensure consumption of the starting nucleoside and formation of the nucleoside cyclic phosphite intermediate, the reaction could be extended to more than 5 hr prior to oxidation and hydrolysis. For the extended storage, the sulfur- and selenium-modified 5′-triphosphates should be kept under argon at −80 °C.</p><!><p>The 5′-triphosphates synthesized according to Basic Protocols 1, 2 and 3 are of high quality. Even without HPLC and ion-exchange purification, they can be directly used as substrates for DNA polymerization and RNA transcription. Because of the replacement of the non-bridging oxygen atom at the alpha-phosphate with either sulfur or selenium, a chiral center is resulted. The modifications creating this chirality have increased the difficulty in chemically synthesizing diastereomerically-pure DNAs and RNAs. The chemically synthesized PS-NAs or PSe-NAs are diastereomeric mixtures as the current chemical synthesis is unable to fully control the diastereomer formation at each phosphorus center. On the other hand, DNA polymerization and RNA transcription enable diastereomer-pure synthesis of phosphate-modified nucleic acids. Herein, DNA polymerase (Klenow) (Eckstein, 1979; Brody et al, 1982) and RNA polymerase (T7 RNA) (Burgers and Eckstein, 1978; Ueda et al., 1991)accept only the Sp diastereomers of dNTPαS (dNTPαSe) and NTPαS (NTPαSe) analogs, respectively. Since the Rp diastereomers are neither substrates nor inhibitors, fortunately, the S- or Se-modified triphosphates can be directly used to conveniently synthesize diastereomerically-pure sulfur- and selenium-derivatized nucleic acids with DNA (or RNA) polymerases, without HPLC or ion-exchange purification of the triphosphates. The expected results of the PS-DNAs and PS-RNAs utilizing crude dNTPαS and NTPαS are shown in Figure 10 and 11. This type of substitutions has been used extensively to study the conformational properties of nucleic acids and enzymatic cleavage of phosphodiester bonds (Burgers et al., 1979; Burgers and Eckstein, 1979; Frey, 1982; Eckstein, 1985).</p><!><p>The synthesis of the nucleoside 5′-triphosphate (the native or the modified ones) is easy and convenient and can be achieved in one day. Fortunately, both the sulfurizing and selenizing reagents are commercially available.</p>

PubMed Author Manuscript

Divergent Synthesis of Indolenine and Indoline Ring Systems by Palladium-Catalyzed Asymmetric Dearomatization of Indoles

Dearomatized indole derivatives bearing a C3-or C2stereocenter exist ubiquitously in natural products and biologically active molecules. Despite remarkable advances in their synthesis, stereoselective and regio-divergent methods are still in a high demand. Herein, a Pd-catalyzed intermolecular asymmetric spiroannulation of 2,3-disubstituted indoles with internal alkynes has been developed for the efficient construction of indoline structures with a C2-quaternary stereocenter. Stereospecific aza-semipinacol rearrangement of these indoline derivatives under acidic conditions afforded indolenine products bearing a C3-quaternary stereocenter, where the migrating group could be controlled by the reaction sequence. The asymmetric spiroannulation together with the subsequent aza-semipinacol rearrangement enabled a divergent access to dearomatized indole derivatives with either a C3-or a C2-quaternary stereocenter.

divergent_synthesis_of_indolenine_and_indoline_ring_systems_by_palladium-catalyzed_asymmetric_dearom

<!>Entry Sub

<p>Indolenines bearing a C3-quaternary stereocenter and indolines bearing a C2-quaternary stereocenter are widely occurring core structures in many natural products and biologically active molecules, [1,2b,2d] and the synthetic methods that enable efficient construction of these core structures is highly demanded. [2] Catalytic asymmetric dearomatization (CADA) reactions serve as one of the most straightforward approach to chiral building blocks from aromatic substrates. [3] Recently, a series of CADA reactions have been developed for the construction of chiral indolenine/indoline derivatives based on the reactivity of C(sp 2 )-Pd species (Scheme 1). [4][5][6] The majority of these reactions were performed in an intramoleculcar manner (Scheme 1a), in which a haloarene or an acetylene is tethered to the N1-, C2-or C3position of indole as the C(sp 2 )-Pd precursor. The You group achieved enantioselective C3-arylation of C3-tethered indoles; [4p] Jia [4b,4k] and Fukuyama [4n] developed the dearomative Heck reaction of N1-and C2tethered indoles to build a C2-stereocenter; Jia, [4h-j,4m,4o] Liang, [4c,4d,4g] Lautens, [4h] and Dai [4e] reported the asymmetric dearomative difunctionalization of C2-tethered indoles by employing an external nucleophile.</p><p>Compared with intramolecular cyclization, the intermolecular version allows for a more divergent approach to dearomatized products. In this regard, Zhang and co-workers reported recently an enantioselective annulative dearomatization of C3-bromoarylindoles with alkynes using the Pd/Sadphos catalytic system to access various spiro-indolenines with a C3-quaternary stereocenter (Scheme 1b). [5] However, to date the intermolecular asymmetric dearomatization of C2-arylindoles has not been reported yet, and stereoselective construction of structurally diversified indoline/indolenine derivatives remains a formidable challenge. [7] Herein, we report a Pd-catalyzed asymmetric intermolecular spiroannulation reaction of C2-arylindoles with internal alkynes, leading to the formation of spiro-indolines bearing a C2-quaternary stereocenter (Scheme 1c). These products could undergo a stereospecific azasemipinacol rearrangement under acidic conditions, affording fusedindolenines with a C3-quaternary stereocenter. Furthermore, the migration selectivity of this process could be tuned by the selection of reaction sequence. The asymmetric spiroannulation together with the stereospecific rearrangement enabled a divergent access to dearomatized indole derivatives with either a C3-or a C2-quaternary stereocenter. At the outset, we envisioned to attempt the reaction between Bocprotected 2-(2-bromophenyl)-3-methyl-1H-indole (1a) and alkynes as a model reaction. We first focused on the aryl-aryl rotational barrier of 1a, since similar substrates exhibit axial chirality [5,6d] and facile interconversion between the two enantiomers is crucial for efficient dynamic kinetic resolution (DKR). [8] We found that, the enantiomers of 1a could be well resolved by chiral HPLC, indicating that they interconvert slowly under room temperature (Scheme 2). The separation of both enantiomers allowed us to determine the kinetic parameters of the racemization process by performing the reaction under elevated temperatures. [9] On the basis of the measured enantiomerization rate constants, the enantiomerization barrier of 1a was determined to be 25.4 kcal/mol at 298 K by applying the Erying equation, which ensures a rapid racemization process at elevated temperature. Bearing this information in mind, we commenced the study on the reaction between 1a and diphenylacetylene (2a) by employing Pd(dba)2 as the precatalyst and toluene as the solvent (Table 1). It was found that, chiral NHC ligand L1, BINAP (L2), spiro-phosphoramidite ligand L3, and Feringa-type ligand L4 resulted in unsatisfactory results (entries 1-4). To our delight, a promising result was observed by employing the Carreira ligand L5 with t-BuOLi as the base at 120 °C, affording product 3aa in 55% NMR yield with 92:8 er (entry 5). A better result was achieved by lowering the reaction temperature to 90 °C and switching the base to MeOLi, providing 3aa in 88% isolated yield with 95:5 er (entry 9). Further optimization on ligand structure indicated that more sterically hindered ligands L6 and L7 exhibited a negative effect on enantioselectivity (entries 10-11). The fluoro-substituted ligand L8 was found to give comparable yield and slightly inferior enantioselectivity (entry 12). A comparison showed that phosphoramidite ligand L9 was inferior compared with L5 (entry 13). A brief screen of the N-protecting group on indole indicated that the Boc group was the best amongst tosyl, acetyl, and methyl group (entries 14-16).</p><p>To illustrate the generality of the reaction, the scope of the indole coupling partners was initially investigated (Table 2). Gratifyingly, 2aryl-3-methylindole substrates with fluoro, chloro, methoxyl, and methyl substituents (1b-e) at the 5-position of the phenyl ring worked well under the optimal reaction conditions, giving the corresponding spiro-indoline products 3ba-3ea in 70-86% yield with 90:10 to 96:4 er. Substrate bearing an electron-withdrawing trifluoromethyl group (1f) was also tolerated to afford 3fa in 92:8 er, whereas the reaction yield decreased to 23%. 6-Methyl and 4,5-methylenedioxy substituent on the phenyl ring (1g-h) and a more sterically congested naphthyl group in place of the phenyl group (1i) were tolerated to provide the desired products 3ga-3ia in 39-99% yield with 93:7-95:5 er. However, introduction of a methyl group at the 6-position (1j) of the phenyl moiety resulted in decreased enantioselectivity. Substituents on the indole moiety (5'-methyl, 5'-trifluoromethyl, 5'-chloro, and 4'-chloro) of the substrates (1k-n) were compatible with the reaction, delivering 3ka-3na in 33-82% yield with 93:7-95:5 er, while the 7'-fluoro substituent significantly decreased the yield of the desired product 3oa. The use of 2-aryl-3-ethylindole substrate 1p led to a good yield of spiroannulation product 3pa as a C=C bond geometrical isomer mixture with diminished enantioselectivity.</p><p>Table 1. Optimization of the reaction conditions.</p><!><p>Base L T/ °C Yield(%) [a] Er [b] 1 [c] 1a t-BuONa L1 120 9 53.5:46.5 2 [d] 1a Cs2CO3 L1 120 23 51.5:48.5 2 [e] 1a Cs2CO3 L2 120 0 -3 [f] 1a [e] Using [Pd(C3H5)Cl]2 (2.5 mol%) and L2 (7.5 mol%).</p><p>[f] Using Pd(dba)2 (5.0 mol%) and L3 (7.5 mol%).</p><p>[g] Using Pd(dba)2 (10.0 mol%) and L4 (15.0 mol%).</p><p>[h] MeOH (1.0 equiv.) as additive.</p><p>[i] Yield of isolated product.</p><p>[j] Not detected.</p><p>[a] Reaction conditions: 1 (1.0 equiv), 2 (1.5 equiv), and toluene (0.1 M) under Ar. Yields of isolated products are reported. The Z/E ratios were determined by crude Then the scope of the internal alkyne was explored (Table 2). Symmetrical diaryl acetylenes with various substituents at the para-(2be) or meta-(2f-g) position of both phenyl rings participated in the reaction smoothly to give the corresponding products 3ab-3ag in 57-97% yield with 92:8-95:5 er. Di(thiophen-2-yl)acetylene (2h) could undergo the reaction with 1a, but afforded a low yield of cyclization product 3ah in 91:9 er. The ortho-substituent on the aryl groups of diarylacetylene generally exhibited a negative effect on the reaction: ortho-fluoro substituted 1,2-bis(2-fluorophenyl)ethyne 2i afforded product 3ai in a moderate yield with a low enantioselectivity (77:23 er), and no desired product was observed for the more sterically hindered ortho-methyl substituted diphenylacetylene 2j. To our delight, the symmetrical dipropylacetylene 2k was found to be suitable to undergo the annulation with substrates 1a, 1p, and 1q. Notably, a gram-scale synthesis of 3ak (1.31 g) was carried out to afford the product in 98% yield with 90:10 er, indicating the scalability of the present method. Moreover, the reactions of 2-aryl-3-ethylindole substrate 1p and 2-aryl-3-tert-butoxycarbonylmethylindole substrate 1q with alkyne 2k afforded the corresponding products 3pk and 3qk in good yields and enantioselectivities. Finally, the unsymmetrical alkyl/aryl mixed acetylenes 2l and 2m produced cyclization products in good yields with decent enantioselectivities and satisfactory regioselectivities.</p><p>Interestingly, we found that the present protocol could be extended to a 2-aryl-3-methylbenzofuran substrate by employing L4 as the ligand, affording spiro[benzofuran-cyclopentane] 3ra and 3rk in comparable enantioselectivities.</p><p>The C2-spiroindoline structure in the cyclized products enabled a potential aza-semipinacol rearrangement [4f, 10] leading to the formation of a C3-stereocenter, and if the rearrangement proceeded stereospeficially, enantioenriched C3-substituted indolenine derivative could be obtained. Therefore, we attempted to perform an acid-promoted aza-semipinacol rearrangement on the spiroindoline products (Table 3). We found that, by treating the C3-methylene (3aa, 3aj, 3am) or C3ethylidene (3pj) substituted spiro-indolines with TFA (Condition A), the rearranged products bearing a C3-quaternary stereocenter (4aa, 4aj, 4am, and 4pj) were obtained in moderate to excellent overall yield without loss of enantiopurities. In this process both the cleavage of the N-Boc group and the stereospecific rearrangement of the C2-substituent proceeded smoothly, and the aryl migration was favored over the alkenyl migration.</p><p>To figure out whether the rearrangement step occurs prior to or after the elimination of the N-Boc group, we investigated another reaction sequence involving stepwise deprotection and rearrangement for the same set of C2-spiroindoline derivatives (Condition B). The deprotection of N-Boc group proceeded smoothly by treatment with TMSOTf/2,6-lutidine, and the crude deprotected product was treated with TFA as before. To our surprise, this reaction sequence delivered alkenyl migration product (4aa', 4aj', 4am', 4pj') as the major product with complete retention of enantiopurities. This result indicates that the nature of the N-substituent has a remarkable effect on the azasemipinacol rearrangement, and under Condition A the rearrangement occurred prior to deprotection. Therefore, by choosing an appropriate reaction sequence, rearrange of C2-spiroindoline to C3-substituted indolenine with selective aryl/alkenyl migration could be achieved.</p><p>In order to gain more mechanistic insights into the Pd-catalyzed cascade spiroannulation and the regio-divergent aza-semipinacol rearrangement, we performed a DFT computational study (Scheme 3). The reaction between indole 1a and alkyne 1b catalyzed by Pd/L5 was selected as the model, and the calculated reaction pathway is shown in Scheme 3a. It was found that, the reaction starts with the Pd-ligand complex, Pd(L5)2, [11] and proceeds through oxidative addition, alkyne coordination and insertion, indole insertion, and β-H elimination steps. The alkyne insertion step via TS2 turned out to be both turnover-limiting and selectivity-determining, and the overall activation barrier for the formation of 3aa was 20.4 kcal/mol in terms of Gibbs free energy. The other pathway leading to ent-3aa starts from the coordination of ent-1a with the Pd catalyst, and the key alkyne insertion step via TS2' was less favored by 1.1 kcal/mol compared with TS2, in agreement with the preference for (R)-product observed experimentally. It is notable that the barrier for the interconversion of atropisomers of 1a (∆G ≠ = 25.4 kcal/mol) seemed to be higher than that for the catalytic dearomatization via TS2 (∆G ≠ = 20.4 kcal/mol). Given that TS2 involves three components (catalyst, substrate 1a, and alkyne 2a at 0.01 M, 0.1 M, and 0.15 M, respectively), the corrected barrier for the dearomatization reaction (∆G ≠ = 25.6 kcal/mol) fits the requirement for DKR. The fact that unconsumed 1a was determined to be racemic at different conversions confirmed the DKR scenario experimentally (see the Supporting Information for details). The aza-semipinacol rearrangement was also investigated by DFT calculation to understand the nature of the migration selectivity, employing the transformation of 3aa to 4aa and 4aa' as the model (Scheme 3b). The result indicated the aza-semipinacol rearrangement of both the N-Boc substrate (starting from INT7) and the N-H substrate (starting from INT7') favors alkenyl migration (TS5a vs TS5b and TS5a' vs TS5b'). We attribute this trend to the fact that alkenyl migration transition states TS5a and TS5a' could better stabilize the positive charge due to the existence of the aryl rings, whereas this stabilizing effect lacks in the aryl migration transition states TS5b and TS5b'. Interestingly, for the N-Boc substrate the rearrangement is reversible due to a more energy-demanding Boc-deprotection step, while for the N-H substrate the rearrangement is irreversible and thus determines the migration selectivity. As a result, the migration selectivity of the N-Boc substrate is dictated by the Boc-deprotection step via TS6a and TS6b, and the preference for the aryl-migrated product in this step inherits from the thermodynamic stability of the rearranged intermediates INT8a and INT8b.</p><p>In summary, we have developed a palladium-catalyzed enantioselective intermolecular dearomatization of C2-arylindoles with internal alkynes, leading to C2-spiroindolines bearing a C2-quaternary stereocenter with good yields and enantioselectivities. The stereospecific aza-semipinacol rearrangement afforded enantioenriched indolenine derivatives bearing a C3-quaternary stereocenter via an tunable aryl/alkenyl migration. The combined steps enabled a divergent access to dearomatized indole derivatives with either a C3-or a C2quaternary stereocenter.</p>

ChemRxiv

DOUBLE PULSED FIELD GRADIENT (DOUBLE-PFG) MR IMAGING (MRI) AS A MEANS TO MEASURE THE SIZE OF PLANT CELLS

Measurement of diffusion in porous materials and biological tissues with the pulsed field gradient (PFG) MR techniques has proven useful in characterizing the microstructure of such specimens noninvasively. A natural extension of the traditional PFG technique comprises multiple pairs of diffusion gradients. This approach has been shown to provide the ability to characterize anisotropy at different length scales without the need to employ very strong gradients. In this work, the double-PFG imaging technique was used on a specimen involving a series of glass capillary arrays with different diameters. The experiments on the phantom demonstrated the ability to create a quantitative and accurate map of pore sizes. The same technique was subsequently employed to image a celery stalk. A diffusion tensor image (DTI) of the same specimen was instrumental in accurately delineating the regions of vascular tissue and determining the local orientation of cells. This orientation information was incorporated into a theoretical double-PFG framework and the technique was employed to estimate the cell size in the vascular bundles of the celery stalk. The findings suggest that the double-PFG MRI framework could provide important new information regarding the microstructure of many plants and other food products.

double_pulsed_field_gradient_(double-pfg)_mr_imaging_(mri)_as_a_means_to_measure_the_size_of_plant_c

1 INTRODUCTION<!>2.1 Overview of Theory<!>2.2 Double-PFG MR pulse sequence<!>3.1 Glass Capillary Array (GCA)<!>3.2 Celery<!>4 DISCUSSION and CONCLUSION

<p>The nuclear magnetic resonance (NMR) signal can be sensitized to diffusive motion of spin bearing molecules, e.g., via the application of pulsed field gradients (PFGs).[1] When the specimen contains pores that restrict the random motion of these molecules, the resulting signal possesses information about the underlying microstructure of the porous medium. Consequently, diffusion NMR provides a means to probe microscopic length scales, which are impossible to resolve using conventional magnetic resonance imaging (MRI).</p><p>The most widely-used approach to provide diffusion sensitization involves the application of a single pair of PFGs. These "single-PFG" acquisitions have been successful in characterizing many useful features of the porous specimens. However, as pore size gets smaller, the experiment demands stronger gradient strengths to characterize the diffusion process. It has been realized since '90s that when multiple diffusion gradient pulse pairs are employed,[2] the dependence of the signal on the angle between the different gradient vectors could yield information that is descriptive of the particular diffusional motion and the underlying porous structure.[3,4]</p><p>Such experiments can be regarded as multidimensional acquisitions yielding information about correlations of motion during different encoding intervals.[5] If the diffusion process can be assumed to be locally Gaussian, a two-dimensional Laplace transform can be utilized to produce maps of diffusion coefficients illustrating the motional correlation during the separate encoding intervals. Such an analysis has been performed on chive samples.[6]</p><p>In recent years, the multiple-PFG extensions of the single-PFG experiments have attracted a great deal of interest in the context of restricted (non-Gaussian) diffusion. It was shown that the non-monotonicity of the signal characteristic of the so-called "diffusion-diffraction" phenomenon in single-PFG experiments[7] is replaced by zero-crossings when an even number of gradient pulse pairs are employed, making the phenomenon robust to microscopic heterogeneity of the examined specimen.[8]</p><p>When the angle between the gradients of a multiple-PFG sequence is varied, the signal can be sensitized to anisotropy at different length scales.[9-13] Of particular interest for the purpose of this study is a special case of the multiple-PFG acquisitions, where two pairs of gradients are employed and the delay or mixing time between the successive gradient pairs is short. In such double-PFG acquisitions, the signal is sensitive to microscopic anisotropy (μA), which is induced by the walls restricting the motion of molecules. This anisotropy is present even in isotropic pores[14]. The resulting angular dependence of the double-PFG signal is characterized by a bell-shaped curve from which the presence of restrictions to diffusion is inferred. The size of the compartments can be obtained by fitting an appropriate model to data. Because this angular dependence emerges in the quadratic term of a Taylor expansion of the NMR signal decay, restricted diffusion can be characterized and microscopic dimensions can be measured using small gradient strengths. Such angular dependence has been observed in a radish specimen.[15]</p><p>In this work, we demonstrate the application of the double-PFG MRI technique to measure the diameter of cylindrical pores of microscopic dimensions and adopt the technique to measure the cell size in plants. First, a description of the employed theory, which involves incorporating the DTI-derived fiber orientation in a previously developed double-PFG framework is presented. Next, the double-PFG imaging sequence is introduced. The technique is subsequently validated through experiments on a stack of controlled glass capillary arrays (GCAs) with different pore sizes to illustrate the contrast in the computed diameter maps. Finally, the method is applied to measure the size of phloem and xylem cells making up the vascular tissue of the celery stalk.</p><!><p>There are two main approaches taken to relate the microstructural features of the specimen to the double-PFG MR observations of restricted diffusion. Continuing the work of Mitra,[3] the first approach treats the data as if they were acquired by setting the timing parameters to their limiting values.[15,16] More specifically, the duration of the gradient pulses (δ) is assumed to be infinitesimally short, while the separation of the pulses (Δ) are assumed to be sufficiently long for the molecules to traverse distances large compared to the dimension of the compartment they reside in. These assumptions greatly simplify the analysis. For example, when the separation between the two encoding blocks (mixing time) is zero, the small-q behavior (q=|q|, where q=(2π)−1γδG, γ is the gyromagnetic ratio, and G is the gradient vector) of the MR signal attenuation is given by[3] E(ψ)=1-4π2q2ρ2/3 (2+cos(ψ)), where ρ2 is related to the mean-squared radius of gyration of the pores, and ψ is the angle between the two gradients. Frequently, one encounters pore shapes that can be characterized by a single size. For example, many biological cells are nearly spherical or cylindrical. In this case, the ρ2 value can be related to the actual diameter of the pores. For cylinders, the signal modulation is given simply by E(ψ)=1-π2q2r2 (2+cos(ψ)), where r is the radius of the cylinders. However, this expression is valid only when the special conditions for the timing parameters can be met, which is difficult to satisfy in practice.</p><p>In contrast, the second approach employs the assumption for the pore shape from the outset. This approach has enabled the incorporation of all timing parameters into the formalism,[9] thus significantly improving the size estimates.[17,18] The corresponding expressions are relatively complicated and will not be reproduced here.</p><p>Another important improvement is the extension of the formalism to arbitrary q-values. To this end, a theoretical framework that is valid for arbitrary gradient waveforms would be exceptionally useful. Such a theory was introduced by Özarslan et al.[19] It generalizes the multiple correlation function method[20] to pulse sequences that involve gradients applied along different directions. This generalization is critically important for the multiple-PFG sequences where the gradients along different orientations are applied in different PFG blocks. Moreover, it opens the door to incorporate the effects of other gradients in the sequence such as spoilers and imaging gradients.</p><p>In this framework, a general gradient waveform is expressed as a piecewise constant function involving N separate intervals. A diagonal infinite-dimensional matrix operator, Λ, and a vector operator, A, are constructed whose components are given by infinite-dimensional matrices, though finite-sized matrices are employed in practice. The elements of these operators are related to the eigenvalues of the Laplacian operator and the size of the pore. Then, the MR signal attenuation due to restricted diffusion is given simply by the very first element of the matrix (1)∏n=1Nexp(−δnΛ+i2πqn⋅A) where δn is the duration of the nth interval, and qn is the corresponding q-vector. Equation (1) enables one to compute the signal attenuation for simple geometries. The elements of the operators Λ and A for spherical and cylindrical pore shapes are provided elsewhere[19] and will not be reproduced here for brevity. This procedure enables one to compute the MR signal attenuation due to restricted diffusion for any pulse sequence as long as the gradients are applied in a direction perpendicular to the walls of the enclosing surface.</p><p>Of particular interest for the purposes of this study is the cylindrical geometry wherein the orientation of the cylinder should be taken into account. As shown before[12], decomposing the q-vector into components parallel with (q∥) and perpendicular to (q⊥) the cylinder's symmetry axis makes it possible to write the attenuation due to diffusion within the cylinder as the product of two attenuations, i.e., (2)Er=E⊥(q⊥)E‖(q‖). Here, E⊥(q⊥) can be computed using Eq. (1) as described above. As for E∥(q∥), using the corresponding expression for a slab geometry would yield the signal attenuation for diffusion taking place inside a capped cylinder[12]. Alternatively, if the cylinders are very long so that diffusion is not influenced significantly by the presence of the caps, an infinite cylinder model could be used. In this paper, we follow the latter option, in which case the E∥(q∥) function can be derived from the solution for free diffusion, denoted by Ef. The procedure to obtain Ef is well-known for arbitrary pulse sequences.[21] For example, when applied to the double-PFG sequences with vanishing mixing times, the relevant expression is given by[9] (3)Ef(q1,q2)=exp(−4π2D0[(Δ−δ∕3)(q12+q22)−(δ∕3)q1q2cos(ψ)]), where D0 is the bulk diffusivity of the molecules, q1 and q2 are, respectively, the q-vectors associated with the first and second blocks of the bouble-PFG acquisition, and q1=|q1| and q2=|q2|. Therefore, as a special case of the above expression, (4)E‖(q‖)=exp(−4π2D0[(Δ−δ∕3)(q1‖2+q2‖2)]), where q1∥ and q∥ are, respectively, the components of the vectors q1 and q2 along the cylinder's axix.</p><p>It is possible for a voxel to contain freely diffusing water molecules in addition to those confined to restricted domains. In this case, the above formulation for restricted diffusion can be combined with the expected signal behavior for free diffusion[9] in a bicompartmental model.[22] The aggregate signal is then given by E=frEr+ffEf, where Er and Ef are the signal attenuations from restricted and free domains, respectively, and fr and ff are the corresponding volume fractions that sum up to unity. It should be noted that this model assumes two populations of molecules to be isolated from each other. Although this is an idealization of the diffusion process for many complex systems, the values obtained by employing such a model is expected to be meaningful when the exchange rate between the different pools is limited. When a substantial level of exchange is expected during the course of the diffusion encoding, a more sophisticated method that accounts for the permeability of membranes could be useful.[23]</p><!><p>We illustrate the double-PFG filtered MRI pulse sequence[18] employed in this study in Figure 1. In this pulse sequence, double-PFG encoding is performed as a filter prior to spatial encoding.[24] Since we are interested in microscopic anisotropy, we would like to minimize the delay between the two diffusion blocks of the double-PFG sequence. This is achieved by applying the second gradient of the first pair simultaneously with the first gradient of the second pair. Thus, the resulting sequence has three pulses (of duration δ where the separation of the successive gradient pulses is denoted by Δ. G1 and G2 are the gradient vectors of the two encodings. The second gradient of the three-pulse sequence is given by the vector sum G1+G2. Once the diffusion encoding is performed, the spatial encoding starts, which is achieved via slice selection as well as phase and frequency encoding.</p><!><p>We performed double-PFG experiments on a diffusion MRI phantom[18] built from four stacked glass capillary array wafers (Photonis, Sturbridge, MA) filled with pure water. The phantom consisted of GCAs with a disk diameter of 13 mm and nominal pore diameter values of 25 and 10 μm (stacked in an alternating manner) with a maximum variation of 5% between GCAs of the same nominal diameter. MR acquisition was carried out on a 7T vertical-bore Bruker DRX system (Bruker BioSpin, Germany). The sample temperature was set to 19°C. The double-PFG parameters were: δ/Δ=3.15/75 ms, q-values of 9.9, 13.9, 19.8, 23.8 and 29.7 mm−1 were achieved by applying gradients of strength 74, 103, 148, 177, and 221 mT/m. The sample was placed parallel to the main magnetic field, which defines the z-axis. The first gradient was fixed along the x-axis while the direction of the second gradient was varied on the transverse (xy) plane such that the angles between the two gradients (ψ) were 0, 45, 90, 135, 180, 225, 270, 315 and 360 degrees. In addition, one data set with no diffusion gradients was collected. The delay between the diffusion and spatial encoding sections of the pulse sequence was set to 35 μs. The parameters for the latter section were: TE/TR=12/7000 ms, slice thickness=2 mm, field of view=15.5 mm, matrix size=128×128, resolution=121×121×2000 μm3. The number of averages was 2, yielding a total acquisition time of 30 minutes for each image.</p><p>In Figure 2, we illustrate our experimental results on the GCA phantom. The top row of this figure contains four double-PFG images, two at a low q-value and two at the highest q-value. In both cases, images with parallel (ψ=00) and antiparallel (ψ=1800) gradients are included. Clearly, at the higher q-value, there is a significant attenuation of the signal in the parallel case relative to the case in which gradients are oriented antiparallel. Note that this anisotropy of the signal is not due to the pores being elongated, as the gradients are perpendicular to the walls of the pore in all cases. Rather, it is induced by the mere presence of the walls, and as such, it is indicative of the reflective character of the diffusion process. The arrows indicate the location from which the signal values are plotted on the right. In this plot, the symbols indicate the data points while the fit to these points are depicted via continuous lines.</p><p>The fits were obtained by computing the signal values using the theoretical framework outlined above. The unknown parameters to be estimated via the fitting were: the radius of the cylinders, and the signal values with no diffusion weighting corresponding to the restricted and free compartments, which we shall denote by S0r and S0f, respectively. Note that the aggregate signal with no diffusion weighting is just sum of these two signal intensities, i.e., S0=S0r+S0f. The volume fractions of the two compartments are also obtained from these signal intensities via the relationships f0r= S0r/S0 and f0f= S0f/S0. An IDL implementation of the Levenberg-Marquardt algorithm available at http://www.physics.wisc.edu/~craigm/idl/fitting.html was employed in the fitting, which also provided error estimates in the results.</p><p>In the bottom row of Figure 2, we illustrate the results obtained via voxel-by-voxel fitting of the model to the data. The first image is a map of the estimated inner diameter (ID) values. A region of interest analysis over the respective GCAs revealed inner diameter values of 27.1±0.4, 10.0±0.7, 27.3±0.4, 10.2±0.6 μm for the voxel-by-voxel pore size estimates. The fitting of the theory to the values obtained by averaging the signal values over the same regions within each of the GCAs led to estimates of 27.36±0.12, 10.04±0.08, 27.43±0.14, 10.17±0.08 μm for the pore diameter. Note that all these values are reasonably close to their nominal values. Moreover, the consistency of the results obtained from the voxel-by-voxel analysis with that achieved by averaging the signal values prior to fitting indicates the adequacy of the signal level for a local estimate of the pore diameter. The image with no-diffusion weighting (S0) obtained from the fit is shown in the second map of the bottom row in Figure 2. Unlike in the case of the ID-valued map, the S0 image does not yield significant contrast between the GCAs with vastly different pore diameters. The last two figures depict the restricted and free volume fractions (fr and ff), which are useful in distinguishing regions containing freely diffusing water from those with restricted diffusion.</p><!><p>We used a similar framework to image a specimen of celery stalk inside water. The MRI protocol included a series of 18 double-PFG scans followed by a diffusion tensor imaging (DTI)[25] protocol with 44 single-PFG spin echo acquisitions. The parameters for the DTI acquisition were: TE/TR=59/3000 ms, δ/Δ=3/50 ms, field of view=22 mm, matrix size=128×128, resolution=172×172×2000 mm3. Two images with no diffusion gradients were acquired followed by 42 diffusion weighted images (21 directions, 2 b-values up to 340 s/mm2, where b=4π2q2(Δ-δ/3)). The number of averages was 1, and the total acquisition time was 4 hours and 42 minutes.</p><p>Double-PFG filtered imaging was performed on the same geometry. The double-PFG parameters were: TE/TR=12/3000 ms, δ/Δ=3.15/50 ms. A total of eighteen double-PFG images were acquired, one at q=0 mm−1, three at q=9.9 mm−1 and seven images were collected at each of q=13.9 and 19.8 mm−1. The number of averages was 8, yielding a total acquisition time of 51 minutes for each image.</p><p>In Figure 3, we illustrate the images obtained from the DTI acquisition. Specifically, we show the image with no diffusion weighting and the quantitative maps of mean diffusivity (MD) and fractional anisotropy (FA)[26]. The figure on the bottom left corner depicts the direction encoded color (DEC) map[27] obtained from the principal eigenvector of the diffusion tensor. This map shows the primary orientation of the fibers within each voxel. Note the sharp contrast the vascular bundles yield in DEC and FA maps suggesting the coherence and elongation of the cells in these regions. Moreover, in these regions, the cells appear to be oriented in-and-out of the image plane as expected.</p><p>In this study, these DTI findings were exploited for two purposes: (i) Due to the realization of the sharp contrast in the direction encoded color (DEC) maps between the vascular and the surrounding tissue, the voxels that are contained in the manually drawn ROIs were further pruned by excluding those with fiber orientations making an angle larger than 150 with the z-axis. (ii) Any small deviation of the orientation of the cells from the z-axis was accounted for by feeding the DTI-derived orientation into the double-PFG fitting procedure as a priori information. Consequently, the diffusion gradients, which are applied on the xy-plane, are decomposed into two components, one along the fiber and one perpendicular to it as described in the Theory section. The former component thus leads to signal attenuation consistent with free diffusion, while for the latter the solutions for restricted diffusion was employed. Apart from this important difference, the fitting procedure followed similar lines as for the GCA data set. The unknown parameters that were determined via the fitting procedure were the fiber radius, S0r, S0f, and the bulk diffusivity, D0. One double-PFG scan with low diffusion-weighting is illustrated in Figure 4. A region-of-interest (ROI) was drawn manually that includes each of the regions containing the vascular bundles; these regions were later pruned via the DTI-derived DEC map as mentioned above. The signal values were averaged over the respective ROIs. The values obtained from the fitting can be seen in Figure 4 as well. The estimates are consistent with their expected values.</p><!><p>It is instructive to consider traditional (k-space) MR imaging as another potential method to directly image the pore space and measure the size of the compartments from such images. However, it is impossible to resolve the individual cylinders using our instrument since, due to the size of the specimens (GCAs or celery stalk), such very high resolution scans would demand a very large matrix size, hence prohibitively long acquisition times. Further, the loss of SNR is expected to be another major limiting factor in such acquisitions. As described in this work, by exploiting the diffusion of molecules and its influence on the double-PFG method, we were able to measure the diameter of the cylinders with reasonable accuracy and precision.</p><p>Double-PFG MR is a promising new alternative to more traditional single-PFG acquisitions that provides novel information while making more modest demands on the gradient hardware. For example, to probe different length scales via single-PFG scans one could vary the diffusion pulse separation[28]. On the other hand, because two pairs of gradients are employed in double-PFG acquisitions, two new experimental parameters (the delay between the two pairs, and the angle between the two gradients) can be systematically varied as well. Of particular interest in this study was the short separation time experiments, which are sensitive to restrictions to diffusion at very low q-values. Consequently, such data can be acquired even when strong gradient coils are not available and the microstructural features of the specimen can be obtained.</p><p>This paper employed a theoretical framework that we presented in a series of articles, and extended it by incorporating the fiber orientation. Although the fiber orientation can be determined from the double-PFG acquisitions as well, it would require a gradient sampling scheme more sophisticated than the circular sampling employed in this work. Here, the deviation of the vascular fiber orientation from the z-axis was assumed to be small, which made it possible to estimate the size of the cells by using the circular sampling scheme even when the plane of the sampled circle is not perfectly perpendicular to the fiber direction. This small deviation can be obtained via an independent method—in the present application, that method was a DTI acquisition performed in tandem. Other contributions of the paper included the illustration of a sharp size-dependent contrast in a well-controlled phantom of GCAs, and the application of our method to a plant tissue.</p><p>The validation of the technique involved experiments on a phantom comprising glass capillary arrays (GCAs). Subsequently, the method was applied to quantify the size of phloem and xylem cells—the two main constituents of the vascular tissue in celery stalk. Considering that incorporation of fast imaging techniques like echo planar imaging would considerably shorten the acquisition times, the satisfactory results of this study combined with the less demanding nature of the double-PFG experiments suggest it as a feasible non-invasive method to characterize the microstructure of vascular plant cells, which are implicated in the transport of water and nutrients along the plant stem. Moreover, the technique is expected to be useful in tackling other problems in food science such as characterizing the size of droplets in emulsions[29] (e.g., cheese and margarine[30]).</p>

PubMed Author Manuscript

Aggregation induced emission behavior in oleylamine acetone system and its application to get improved photocurrent from In2S3 quantum dots

Blue emission giving nanoscale molecular clusters of Oleylamine-Acetone system was formed by an aging assisted hydrogen bond formation between the interacting molecular systems, at room temperature. The as-formed nanoscale molecular clusters were found to be self-assembled into flowerlike aggregates and shifted the emission wavelength to red colour depicting an exciton delocalization in the aggregate system. Interestingly aging process has also produced imine type binding between Oleylamine and Acetone due to the condensation reaction. The experimental conditions and formation mechanism of hydrogen bond assisted Oleylamine-Acetone molecular aggregates and imine bond assisted Oleylamine-Acetone is elaborated in this paper in a systematic experimental approach with suitable theory. Finally we have introduced this Acetone assisted aging process in In 2 S 3 QD system prepared with Oleylamine as functional molecules. It was found that the aging process has detached Oleylamine from QD surface and as a consequence In 2 S 3 QD embedded Oleylamine-Acetone aggregates was obtained. When this In 2 S 3 QD embedded molecular cluster system was used as an active layer in a photo conductor device then a maximum photo current value of the order of milli Ampere was obtained. The surfactant molecules normally inhibit the charge transport between QD systems and as a result it is always problematic to have the functional molecules in the QD based transport devices. Our approach has a solution to this problem and the present paper discusses the outcome of the results in detail.The study on "Molecular aggregates" 1-4 is an exciting topic due to its interesting coulombic as well as vibrational coupling between nearby molecules leading to the better exciton transport within the aggregate 5,6 . Michael Kasha, American photochemist has clearly explained the concept of photophysics in molecular aggregates. According to kasha's rule, molecular aggregates can be classified into two types such as H-aggregates and J-aggregates 5,7,8 . H-aggregate is the aggregate in which transition dipoles of neighbouring molecules are aligned in a parallel fashion, exhibiting hypsochromic shift in the absorption spectrum compared to monomer absorption 9,10 . On the other hand, when transition dipoles of neighbouring molecules are aligned in a head to tail manner, it exhibits bathochromic shift in absorption spectrum compared to monomer absorption which is coined as J-aggregates 9,11,12 . Packing of molecules to form aggregate is highly related to intermolecular interactions between the molecules. The intermolecular interactions in the form of magnetic dipole moment, electric dipole moment,

aggregation_induced_emission_behavior_in_oleylamine_acetone_system_and_its_application_to_get_improv

<!>Experimental methods<!>Immisible Oleylamine-Acetone hybrid system with low Acetone concentration. A 5 ml of<!>Results and discussion<!>Conclusions

<p>pi-pi stacking, halogen 13 or hydrogen bonding allows ordered structure of molecules 14 . Among which hydrogen bonding triggered self-assembly is very interesting one [15][16][17][18][19][20] . High boiling point of water arises from OH… HO type hydrogen bond in the system 21 . NH…O=C type hydrogen bond exist in DNA system make them semiconducting 22,23 . Hydrogen bonded pigment indigo is a very good organic semiconductor. NH…O=C type hydrogen bond with neighbouring molecules along one particular crystal axis and π-stacking along perpendicular direction leads to higher mobility in indigo. Optical absorption spectra of most of the hydrogen bonded pigments undergo extensive bathochromic shift from dilute solution to the solid state. Hydrogen bond enhanced delocalization effect [24][25][26] was briefly explained in ϒ-quinacridone 27 . This pigment is used as an active layer in photovoltaic device and shows considerable external quantum efficiency of nearly 10%. Herewith we have illustrated the NH…O=C type hydrogen bond interaction found between Oleylamine and Acetone and studied their self-assembling nature upon aging for the first time.</p><p>Oleylamine [28][29][30] is a well-known capping ligand 31 used for the synthesis of various metal and semiconducting nanomaterials 30,32,33 . This molecule has received much research attention in the area of phosphine-free growth of colloidal quantum dots 34,35 . Oleylamine can be used as a passivating layer in silicon solar cell to enhance the device efficiency 36 . Doping of Self-assembled Oleylamine networks 37 in MoS 2 transistor can increases the carrier concentration from 0.7 × 10 12 cm −2 to 1.9 × 10 13 cm −2 . Oleylamine belongs to the family of primary aliphatic amines and the chemical structure of Oleylamine is similar to that of Oeic acid except the end group COOH (the end group is NH 2 in Oleylamine). It is well known that amine-carbonyl type hydrogen bonding is much studied in pigments like Indigo, tyrian purple, quinacridone, epindolidione and biomolecules such as DNA and proteins 16 . Because Oleylamine contains two hydrogen atoms covalently bonded to the nitrogen atom, it can easily form hydrogen bond with neighbouring molecules. With this basic knowledge, here we have chosen carbonyl group containing molecule Acetone and amine group containing molecule Oleylamine and studied their possibility to form amine-carbonyl type intermolecular hydrogen bonding interaction. Interestingly, aging assisted formation of hydrogen bond in the Oleylamine-Acetone was found at room temperature and that leads to the formation of nanoscale molecular clusters and their flower like self-assemblies. The present paper gives an insight into the formation of such a structures and their application in a systematic approach. Interestingly we have also observed the room temperature formation of imine type interaction between Oleylamine and Acetone molecules during the aging process and that was due to the condensation reaction between Oleylamine and Acetone across the interface between solute and solvent. This process has converted Oleylamine into Oleylimine. By optimizing the ratio between Oleylamine and Acetone an interface free condition was set to avoid the formation of Oleylimine and as a result hydrogen bond assisted molecular cluster/aggregates were only formed.</p><p>Nanoscale cluster to macro scale aggregates and their reversible process was controlled by changing the dilution level of the cluster system and as a result rapid reversible photoluminescence behavior was also noticed. This paper discusses about the reversible emission behavior of the molecular system and the formation of mechanism of the molecular clusters in detail. Finally we have introduced this strategy, that is, Acetone assisted aging process in In 2 S 3 quantum dot system prepared with Oleylamine as surfactant molecule, and this is in order to understand the consequences of aging process in photo current generation in In 2 S 3 QD system. Interestingly the aging process has removed the Oleylamine surfactant molecules from In 2 S 3 QDs. As evidence slightly size enlarged In 2 S 3 QDs were found to be coexisting with Oleylamine-Acetone molecular cluster system. It was found that the molecular cluster system worked as conducting matrix, and thus helped to improve the overall photo current generation to few milli amps. In the present paper the formation conditions of hydrogen bond assisted Oleylamine-Acetone molecular clusters/aggregates and covalent interactions between the molecules will be discussed first and followed by In 2 S 3 QD preparation/device results will be discussed.</p><!><p>Chemicals. Oleylamine (≥ 98%, Primary amine) purchased from Aldrich and Acetone (99%) obtained from Loba chemicals were used in this study.</p><!><p>Oleylamine mixed with 2 ml of Acetone stored in centrifuge tubes initially looked like transparent in nature and after ten minutes it was converted into whitish cloud solution. Then the solution was separated into two layers when the aging process proceeds for four weeks. The upper part seems to be yellowish in colour and the lower part seems to be colourless. When we excite the sample with UV laser, then we have attained dual colour emission from the sample in such a manner that the upper part looks yellow in colour and the lower part looks blue in colour.</p><p>Miscible Oleylamine-Acetone hybrid system with high Acetone concentration. A 5 ml of Oleylamine mixed with 10 ml of Acetone exhibited transparent colour during the day of mixing, then every week its colour changed from pale yellow to dark red after one month. Almost similar experiments were conducted with different ratio as mentioned in the main text.</p><p>In 2 S 3 quantum dot preparation. Indium Chloride (99.9%, Sigma Aldrich), Elemental Sulfur (99.9%, Sigma Aldrich) were used as a source material for In 3+ and S 2− ions respectively. Dodecanthiol (99%, Sigma Aldrich) and Oleylamine (98%, Aldrich) were used for the capping purpose. Acetone (99%, Loba Chemicals) was used for aging study. A 0.001 mol of indium chloride was dissolved in 1-Dodecanthiol and stirred under Argon atmosphere at 110 °C for half an hour. Elemental Sulfur was dissolved in Oleylamine and stirred under vacuum at 90 °C for half an hour. After that elemental sulfur solution was quickly injected into indium chloride solution and mixed under Argon atmosphere for 20 min to grow indium sulfide quantum dots. As obtained quantum dots were purified by means of using equimolar ratio of Hexane (Solvent) and Methanol (Antisolvent) under centrifugation at 5000 rpm for 30 min. As prepared Oleylamine and 1-Dodecanthiol capped blue emitting In 2 S 3 quantum dots were dissolved in Acetone and kept under aging for 1 month.</p><!><p>A 5 ml of technical grade Oleylamine (C 18 H 35 NH 2 ) obtained from Sigma Aldrich was mixed with 10 ml Acetone (CH 3 COCH 3 ) and the resulting sample's optical absorption and emission spectra were recorded before aging. The absorption and emission 38 maximum were at 350 nm and 410 nm respectively for this sample as shown in Fig. 1A,B. The as prepared mixture solution was then kept under aging for about 30 days and found to be converted into a thick red colour solution. The corresponding samples absorption and emission maximum were found to be strongly red shifted to 585 nm and 640 nm respectively. Further the emission intensity of this aged samples was found to be quenched. The dark red solution was diluted by adding Acetone to look for an enhanced red emission. But surprisingly due to dilution effect we have obtained blue shifted emission spectra. An effort was taken to study this reversible blue shift by taking different micro liters of aged solution and diluting them in a fixed volume of Acetone (5 ml) and by doing so it was possible to shift the emission maximum towards blue side in a controlled manner as in (Fig. 1C,D). For example, when 500 μl red solution was diluted in 5 ml Acetone then an emission maximum at 555 nm corresponding to the excitation at 470 nm was observed. When 200 μl red solution was diluted then emission spectra has shown slight enhancement in emission intensity and blue shifted emission at 537 nm. For 100 μl red solution, the dilution has resulted blue shifted emission at 494 nm. Compared to 100 μl diluted sample, when 20 μl aged sample is diluted then it was possible shift the emission to 444 nm corresponding to the excitation at 367 nm. Thus, by controlling the dilution level preciously it was possible to fix the emission wavelength to a desired value and therefore a continuous tunability of emission www.nature.com/scientificreports/ wavelength was possible in this molecular solution. Such an observation is similar to the one from quantum dots, where emission tunability is done by varying the size 39 . By varying the size of the quantum dots, quantum confinement level was controlled to tune different emission wavelengths. In our case the samples are not QDs, but molecular solution and therefore the possibility of forming molecular clusters/aggregates could be the main reason behind the observed emission shift. In the case of molecular system, aggregation induced emission [40][41][42] (AIE) and aggregation caused quenching (ACQ) are the two important phenomenon that occur in many of the organic molecules. Aggregation of fluorophore can yield any of the following emission characteristics in comparison to that of its diluted state (i) Quenching of fluorescence intensity. (ii) Enhancement of fluorescence intensity and unchanged fluorescence intensity. If the fluorescence intensity of aggregates is enhanced in comparison to that of dilution state then the process is called aggregation induced emission [43][44][45][46] . In contrast, if the fluorescence intensity of aggregates is reduced in comparison to that of dilution state then the process is called aggregation caused quenching. Some of the fluorophores exhibited aggregation induced redshifted emission (AIRSE) 43,47 . Herewith we have obtained the aggregation induced redshifted emission at first through aging process and a rapid dilution assisted reversible blue shifted emission. The observed photoluminescence behaviour clearly indicates a fact that there is a possibility of forming some types of clusters/aggregates and dis-aggregates.</p><p>In order to find out the correlation between the change in the emission wavelength and micro structure of the samples, we have subjected the samples to HRTEM analysis (Fig. 2). The 20 μl diluted (in 5 ml Acetone) sample has exhibited spherical shaped particles having an average size of 3 nm. In the case of 100 μl diluted sample, we have obtained micron sized aggregates having a flower like morphology as shown in Fig. 2E. A close look into this micro structure has revealed a beautiful stacking arrangement of nanoparticles as shown in Fig. 2D. The stacking pattern has extended over an average lateral distance of 180 nm and width of 18 nm. The structural observation clearly tells the fact that the fundamental building block of the flower like pattern is nanosized particles and those nanoparticles were nothing but could be the molecular clusters made of Oleylamine-Acetone complex system. A similar flower like morphology with denser stacking of nanoparticles with an average lateral distance of 390 nm and width of 36 nm was obtained from 200 μl diluted sample (Fig. 2F).Thus from the morphological analysis of the samples we can understood the following facts. During aging process Oleylamine-Acetone interaction is established leading to the formation of nanoscale heterotype molecular clusters at first and then their selfassembly by Ostwald ripening, leading to the formation of flower-like aggregation. The formation of stable nanoscale sized molecular clusters took several days thus requiring an aging process, but the as formed clusters self-assembled into flower like aggregates immediately. When the well grown aggregates becomes responsible for the red emission, the nanoclusters were becoming responsible for blue emission. When the molecular structure was switching between these two extreme ends through a change in dilution process then the emission maximum was shuttling between these two wavelength regions. More importantly when nanocluster concentration of the diluted sample was increased then we have noticed immediate red shift, does not requiring any aging process. This observation clearly indicated a fact that the aging process is required to form nanoscale clusters, and once they are formed then the formation of flower shaped aggregates and dis-aggregates are faster process. It is therefore a rapid red to blue (or) blue to red emission wavelength tunability was possible by varying the dilution level of the solution.</p><p>In order to know about the vibrational level changes upon the formation of nanoscale molecular clusters and their self-assemblies we have subjected the samples to Fourier transform infrared spectroscopy analysis. Figure 3A shows the FTIR spectra of deep red colour aggregated sample (Oleylamine-Acetone aged sample) and for the comparison purpose the FTIR spectrum from Oleylamine and Acetone are also given. FTIR spectra for pure Oleylamine solution shows dominant sharp band at 2922 cm −1 and 2852 cm −1 corresponding to methyl asymmetric stretching from the terminal CH 3 groups and methyl asymmetric C-H stretching from the CH 2 groups in Oleylamine chain. The broad band at 1629 cm −1 corresponds to N-H bending and NH 2 scissoring combined motions. FTIR modes in between 650-900 cm −1 corresponds to N-H wagging. The bands at 909, 964 and 993 cm −1 corresponds to NH 2 bending mode. FTIR spectra for Acetone solution shows dominant mode at 1712 cm −1 corresponds C=O stretching mode. The other vibrational modes at 1426 and 1361 cm −1 corresponds to CH 3 asymmetric and symmetric deformation. CCC asymmetric stretching mode of Acetone is at 1222 cm −1 . The modes at 904 and 1091 cm −1 corresponds to CH 2 rock vibrations. Oleylamine-Acetone aged samples show both vibrational modes from Oleylamine and Acetone molecules. C=O stretching mode of Oleylamine-Acetone aged samples red shifted to 1706 cm −1 compared to that of Acetone (1712 cm −1 ) as shown in Fig. 3B and that could be because of the formation of NH…O=C type hydrogen bond 48 between hydrogen in the Oleylamine 49,50 and oxygen in the Acetone. Figure 3C Schematic diagram illustrating the formation of hydrogen bond interaction between the Oleylamine and Acetone.</p><p>Raman spectroscopy analysis was used to confirm the formation of molecular clusters/aggregates. Figure S1 shows the Raman spectra recorded from Oleylamine-Acetone sample at the time of mixing and aged Oleylamine-Acetone molecular aggregate samples. For comparison purpose we have also recorded the Raman spectra from Oleylamine and Acetone samples as well 51,52 . The Oleylamine-Acetone sample at the time of mixing has shown the respective samples Raman modes as in Figure S1. The tables (Table S1 and Table S2) given in the supporting information explains about the details of the Raman modes. We have not seen any noticeable changes in the Raman modes of Oleylamine and Acetone at the time of mixing. However in the aged samples we have obtained a broad Raman spectrum as shown in Fig. S1. This interesting observation clearly tells a fact that there is correlation between the formation of molecular clusters/aggregates and broadening of Raman modes. We believe that the discrete vibrational levels of monomer type molecules becomes a relatively dense packed energy levels due to the clustering of molecules and when Raman excitation/de-excitation occurs through these dense energy levels a broader Raman band is resulted. The observation of broad Raman modes can be therefore taken as a confirmation for the molecular cluster formation. Further in order to know about the crystalline nature of the sample X-ray diffraction analysis was carried out from the Oleylamine as well as Oleylamine-Acetone aged sample by coating it as thin film on borosilicate glass substrate. We have obtained a broad diffraction peak at 22° www.nature.com/scientificreports/ corresponding to the 2θ of amorphous carbon 53 (Fig. S2) and it indicated the non-crystallinity of the molecular cluster/aggregates. Time correlated single photon counting technique was used to find the excited state lifetime of as-prepared samples, which is shown in Fig. 4. Oleylamine-Acetone sample at the time of mixing exhibited an average lifetime value of 9.58 ns. After the aging process for 1 month the sample exhibited lower average lifetime value of 4.56 ns compared to initial day sample. The observed results indicated a fact that the photoexcited charge carriers are undergoing a rapid relaxation through the coupled aggregates resulting of decrease in the average lifetime value in the case of aged sample. We have also repeated the experiment with different ratio of solute (Oleylamine)/solvent (Acetone). The solute:solvent ratio taken are 5 ml: 0.5 ml, 5 ml: 1 ml, 5 ml: 1.5 ml, 5 ml: 2 ml, 5 ml: 2.5 ml, 5 ml: 3 ml, 5 ml: 4 ml, 5 ml: 5 ml, 5 ml: 6 ml, 5 ml: 7 ml, 5 ml: 8 ml, 5 ml: 9 ml and 5 ml: 10 ml. The mixed solutions www.nature.com/scientificreports/ were kept under aging for 30 days. When 0.5 ml of Acetone is mixed with 5 ml Oleylamine, then we have noticed blue emission even after aging for several days. However, when 1 ml Acetone is mixed with Oleylamine of 5 ml, an interesting observation is noticed. Now the aged solution did not turn completely into red. Instead the top portion of the container has turned into orange, leaving the bottom portion as transparent solution (Immiscible solution). When we have slowly increased the Acetone volume to 2 ml, 3 ml, 4 ml, 5 ml and 6 ml then there we have noticed a change in the ratio between the orange solution to the transparent solution, upon aging. In particular the orange portion of the solution (volume) was increasing with the increase of Acetone addition. When 7 ml, 8 ml, 9 ml and 10 ml Acetone were added then the entire solution was turned into orange to deep red solution upon aging (Miscible solution). The critical quantity of Acetone required to turn the entire solution to red colour (upon aging) was found to be 7 ml with 5 ml Oleylamine.</p><p>When the samples having two regions were seen through UV light excitation, interestingly we have noticed two different emission colours as shown in Fig. S4. The top portion has slowly turned from blue to yellow colour after one month and yellow to orange colour after two months, whereas the bottom portion remained blue in colour always. To understand the nature of the solution in two different regions, FTIR spectrum was taken for both portions of the samples (Fig. S4). The aged bottom portion has shown the signature for the presence of Water and Acetone. The presence of Water indicated a fact that there is a possibility for the condensation reaction between Oleylamine and Acetone leading to the formation of Oleylimine. Further from FTIR analysis of the bottom portion of sample we have noticed hydrogen bond interaction between Water molecule and Acetone and that reflected with a redshift in vibrational mode of C=O stretching to 1694 cm −154 . The top portion has shown the signature for the presence of both imine and hydrogen bonded Oleylamine-Acetone molecules. The FTIR modes at 1662 cm −1 corresponds to C=N stretching and 1710 cm −1 corresponds to hydrogen bonded C=O stretching were the evidences for the presence of both Oleylimine and hydrogen bonded Oleylamine-Acetone clusters 55 , respectively. The Oleylimine molecule (imine bonded Oleylamine-Acetone molecular structure) formed at the bottom of the beaker during the condensation reaction reaches the top portion of the solution and become responsible for the FTIR mode at 1662 cm −1 . Thus from overall experimental observations we understand the Theoretical work also has been conducted in order to understand the formation of hydrogen bond and its stability in Oleylamine-Acetone molecular system, through quantum chemical calculations. Theoretical emission spectrum of Oleylamine were obtained in Air and Acetone solvent medium and this is to understand the environmental effect on Oleylamine emission spectra. Figure S5 is the simulated emission spectra of Oleylamine in gas medium and Oleylamine in Acetone solvent medium, intended for the study of solvatochromic effects on Oleylamine in Acetone solvent. The spectral curve is clear indicative of solvatochromic effect leading to red shifting of Oleylamine molecule emission spectrum in Acetone solvent medium. In gas medium spectral distribution of Oleylamine seems to be in shorter UV range of wavelength, suggestive of its colourless property. Though Oleylamine shows well distinguishable red shifted spectral behaviour in Acetone solvent medium in the experimental case, its spectrum still in longer UV range in theoretical data. These observations are suggestive that though solvatochromism of Oleylamine in Acetone solvent red-shifts its gas phase spectra, solvatochromism alone could not explain the experimentally observed vast spectral changes occuring in Acetone solvent upon aging. Hence we speculate that the experimentally observed red shifted emission probably involves reasons beyond solvatochromism. In this case there is a possibility for hydrogen bond interaction between solute-solvent molecules. Since Oleylamine has amine (-NH 2 ) ending and Acetone has open oxygen atom (-C=O), hydrogen bonding could possibly happen between those molecules. Possibilities of other non-covalent interactions in other sites of molecules are seems to be low. Hence we carried out quantum chemical analysis on probable hydrogen bonding sites where electronegative atoms and hydrogen atoms are co-existent. Figure 5. represents optimized geometries of interacted Oleylamine-Acetone complexes of various combinations. It is important to note that stronger interactions are characterised by shorter bond length and larger interaction energy. Table 1. enumerates interaction energies of each of these complexes. From bond length and interaction energy it could be observed that Ace_Ace (Acetone interacting with another Acetone) interaction is the weakest among all interactions, OLA_OLA (Oleylamine interacting with another Oleylamine) interaction, OLA1_Ace (Hydrogen H1 in the Oleylamine interact with one Acetone) and OLA2_Ace (Hydrogen H2 in the Oleylamine interact with one Acetone) possess comparable interaction strength, while OL12_Ace ( Hydrogens H1 and H2 in the Oleylamine interacts with two Acetones) shows strongest interaction among the studied complexes. When solute-solute interaction and solvent-solvent interaction are stronger than solute-solvent interaction, one would observe solvophobic effect that results in immiscible mixture of solute and solvent. However from the obtained results we could see that solution-solution interaction (Ace-Ace) is clearly weaker than solute-solvent interactions (OLA1_Ace, OLA2_Ace and OL12_Ace). Further, though solute-solute (OLA_OLA) shows stronger interaction than solute-solvent interactions (OLA1_Ace, OLA2_Ace and OL12_Ace), in terms of bond distance, and viceversa in terms of interaction energy based values we see solute-solvent and solute-solute interactions are close in interaction strength. Therefore, we conclude that the discussed mixture might show partial solvophobic effect, where Acetone solvent will try to penetrate Oleylamine solute to form hydrogen bonding with Oleylamine molecules. On the other hand, hydrogen bonding interactions of Oleylamine molecules with another Oleylamine will try to mitigate the hydrogen bonding interaction between Acetone molecules and Oleyamine, as a result we assume a persistent competition of interactions between solute-solvent and solute-solute species will be there leading to the formation miscelle type configuration. Figure 6 shows the HOMO-LUMO distributions observed in the studied complexes, which clearly shows that HOMO of Oleylamine and LUMO of Acetone are crucial orbitals involved at hydrogen bonding sites. HOMO and LUMO distributions in solute-solvent interacted complexes shows localization of HOMO on Oleylamine and localization of LUMO on Acetone, which is indicative of charge transfer type excitations in these systems (donor-acceptor kind of complex). Such complexes are usually stabilized by electrostatic attractions. Among all the studied complexes, OL12 shows much reduced HOMO-LUMO gap (Table S3). The relative energy level position for all the studied complexes are also given in Table 1. Interaction energies of solute (Oleylamine) and solvent (Acetone) calculated at sB3LYP/6-311G (d,p) level of theory with BSSE corrections. (i) Acetone-Acetone complex, (ii) Oleylamine-Oleylamine complex, (iii) Oleylamine (OLA1)-Acetone complex, (iv) Oleylamine (OLA2)-Acetone complex, (v) Oleylamine (OLA12)-Acetone complex. www.nature.com/scientificreports/ the sensitiveness of the ratio between the solute and solvent in deciding the type of reactions between the molecules. As we observe the formation of molecular aggregates at a critical solute-solvent ratio, wherein the solution (Acetone) concentration is higher, we believe that Oleylamine (solute) will be surrounded by Acetone (solution) resulting into a micelle structure and in this pattern hydrogen bonding is favoured between the Oleylamine and Acetone molecules, which finally leads to the formation of nanoscale molecular clusters and flower like aggregates. When there is no enough coverage to surround solute molecules, such micelle type structures are not seems to be formed and therefore hydrogen bond type interaction between the molecules is not favoured. The excess solute therefore settles at the bottom as a separate phase, thus producing two different regions in the sample. The formation of an interface between two liquid phases could be the driving force of condensation reaction between Oleylamine and Acetone. Interface assisted (Immiscible liquids) polymerization 56 , liquid-air interface assisted peptide bond formation 57 are the best examples for the conversion of monomers into a big molecule. At the interface when monomer molecules try to diffuse from one region to other there is a high possibility to have interaction between them and therefore monomer becomes large sized molecules. Here we believe that when monomer molecules (Oleylamine and Acetone) diffuse from one part of the liquid to other there is a close interaction between them leading to the formation of Oleylimine, leaving Water molecules as byproduct. Therefore from overall experimental/theoretical studies we come to a conclusion that the hydrogen bond type chemical interaction between Acetone and Oleylamine is due to the formation of micelle structure and relative stronger solute-solvent and solute-solute interaction between the molecules. On the other hand Oleylimine is formed due to the interfacial effect and that happens at a different ratio between solute-solvent mixing. The photographic image of of different dilutions of Oleylamine-Acetone aged samples under UV laser illumination as shown in Fig. S7. Since Oleylamine is an important surfactant molecule used to prepare size controlled quantum dots, Acetone assisted aging process was done in In 2 S 3 quantum dot samples, as an example system. First we have prepared Oleylamine and 1-Dodecanthiol capped blue emitting In 2 S 3 QDs by Hot Injection Technique (Details are given in the experimental part). Then the QDs were dissolved in an optimized Acetone concentration and kept under room temperature aging process for one month. Figure 7 shows the HRTEM images of as prepared In 2 S 3 QD system, which shows the clustered type QDs with an average size of 3 nm. In contrast to this as prepared sample's HRTEM image, the aged sample's HRTEM image has shown a different morphology, consisting of In 2 S 3 QDs with relatively larger size (average size: 19 nm) and as well as Oleylamine-Acetone molecular clusters, as we www.nature.com/scientificreports/ have seen previously. The size distribution corresponding to larger sized In 2 S 3 QD is given in the supplementary information. In 2 S 3 QDs were found to embedded in the Oleylamine-Acetone molecular aggregate matrix as shown in Fig. 8. The carbonyl group in the Acetone molecules formed hydrogen bond interaction with amine group in the Oleylamine attached to the In 2 S 3 QDs and that leads to the detachment of Oleylamine molecules from QD surface and as a consequence there is an increase in QD size. In the absence of sufficient surface coverage smaller sized In 2 S 3 QDs started to grow into larger sized QD system is the reason for this observation. The optical absorption and emission spectra were recorded from the aged sample and compared it with the data of as prepared samples. Interestingly we have seen a shift in the emission peak position from blue to red wavelength, as noticed previously in the molecular aggregate samples. This shift could be due to the possible energy transfer interaction between blue emission giving In 2 S 3 QD and red emitting Oleylamine-Acetone cluster system. As the broad absorption band position of the molecular cluster has overlapped with the emission maximum of blue emitting In 2 S 3 QD, there is a high possibility to have energy transfer interaction between In 2 S 3 QD and molecular cluster system. The presence of In 2 S 3 QD in the matrix of molecular cluster thus enable the transfer of photoexcited charge carriers from In 2 S 3 QD to molecular system. To use this advantage in the photo current generation, a photo current device was constructed using the In 2 S 3 QD/Oleylamine-Acetone molecular aggregate hybrid system as an active layer. For comparison purpose devices were also constructed with In 2 S 3 QD free molecular cluster sample as an active layer as well.</p><p>The structure of the photocurrent device fabricated using the as formed cluster solutions as photoactive material and the relative energy level diagram are shown in (Fig. 9). The device constructed using Oleylamine-Acetone sample at the time of mixing did not produce any photo current upon illumination with solar simulator. On the other hand device constructed using Oleylamine-Acetone aged sample shows good photocurrent value of 1.3 µA at 50 s ON/OFF ratio time with the biasing voltage of 1 V (Fig. 10). The I-V curves of the samples were also recorded at dark as well as light illumination condition. There is a considerable increment in the photocurrent value after light illumination compared to the dark condition reveals that the molecular clusters are sensitive to incident light and able to produce photocurrent. It has to be noted that non-zero dark current in the samples could be due to leakage of charges through the defect levels present in the samples. Since the active and other supportive layers are sandwiched between two electrodes with different work functions, a potential drop will be www.nature.com/scientificreports/ there across the device structure to drive the trapped charge carriers through the defect levels under dark condition resulting into a non-zero dark current at zero applied bias. The magnitude of this dark current shall follow a linear trend when external bias is applied. Photo current device fabricated with aged sample can produce high magnitude photocurrent because exciton delocalization through self assembled aggregates is much easier. On the other hand the photo current device fabricated with diluted samples shows relatively lesser magnitude photo current, and that can be related to dissociation of self-assemblies which can decrease the exciton delocalization.</p><p>The photo current values of 0.16 μA and 0.06 μA were obtained from the photo current devices made up of 500 μl and 200 μl diluted samples. In the device, if the In 2 S 3 QD aged sample was used as an active material then we have noticed a high magnitude photocurrent in mA range as shown in Fig. 11B. For comparison purpose In 2 S 3 QD un-aged samples photocurrent data are also given in Fig. 11A, and which shows lesser magnitude photocurrent. The presence of In 2 S 3 QD in the matrix of Oleylamine-Acetone aggregates pumps in more photo generated charge carriers into the molecular cluster system and therefore a high magnitude photocurrent is possible in the hybrid sample. Since the as prepared aggregates are J-type with head to tail dipole arrangement and as this arrangement is helping to delocalize the charge carriers, the photo excited charge carriers transferred from In 2 S 3 QD can be easily conducted through the molecular aggregates, therefore, high magnitude photocurrent is possible. From the comparison of all of the photocurrent data, it is concluded that there is an overall increase in the photocurrent magnitude when aged In 2 S 3 QD sample is used and which supported our claim that the photoexcited charges are transferred to molecular cluster system through energy transfer interaction. Since, most of preparation experiments of colloidal QDs are done by organic surfactants/ligands such as Oleylamine, Oleic acid and tri-octyl phosphine oxide, which are introduced during colloidal QDs synthetic procedures to control nucleation and growth, their presence in the device structure has some disadvantageous effect on the device performance. For example, these ligands on the colloidal QDs surface are usually electrically insulating and prohibits the hopping of photo induced excited charge carriers from one QD to other and as a result, a less number of photo excited charge carrier reaches to electrode. Hence therefore, the replacement of bulky surfactants are essential to increase the charge carrier transport within the colloidal QDs or from QDs to another nanostructured system (for eg: to TiO 2 electrode) for an improved photocurrent generation. It is therefore various efforts have been taken to post synthesis replacement of capping ligands with short chain molecules or mono atomic ligands (Cd, S, Br or I) to form interconnected QD network structure. Here we have used a different approach to remove the surfactant molecule (Oleylamine) from In 2 S 3 QD surface by aging assisted cluster formation method. Interestingly the as formed molecular aggregates are found to be acting as a conducting platform to transport the photo excited charge carriers from In 2 S 3 QD system and enhanced photocurrent is generated. It is therefore believed that our approach can solve the problem associated with the removal of surfactant from QD system, in the future.</p><!><p>In conclusion, through experimental and quantum chemical calculation results we have realized hydrogen bond mediated heterostructure type molecular cluster formation and their aggregation, resulting of red shift in the emission wavelength for the first time. Aging assisted formation of NH…O=C type hydrogen bond between Oleylamine and Acetone leads to the formation of nanoscale molecular clusters. Once the nanoscale molecular clusters were formed then the formation of flower like aggregates and their dissociation as well as reformation were rapidly occurred by means of controlling the dilution level of flower like aggregates in Acetone. Dissociation www.nature.com/scientificreports/ and reformation of flower like aggregates leads to the red to blue and blue to red shift in emission characteristics in a rapid manner. The sensitiveness of the ratio between solute/solvent in deciding covalent or non-covalent hydrogen bond type interactions were elaborated in Oleylamine-Acetone molecular system. Covalent interaction between Oleylamine and Acetone was possible due to the formation of liquid-liquid interface between molecular cluster and non-cluster solution, followed by diffusion of monomers through interface. Further high current giving photoconductive device was also realized from the molecular cluster/aggregate system. When this implication (molecular cluster formation) was introduced in a real quantum dot system, there we noticed an further increase in the photocurrent generation.</p>

Scientific Reports - Nature

Molecular Mechanism of Cytokinesis

Division of amoebas, fungi, and animal cells into two daughter cells at the end of the cell cycle depends on a common set of ancient proteins, principally actin filaments and myosin-II motors. Anillin, formins, IQGAPs, and many other proteins regulate the assembly of the actin filaments into a contractile ring positioned between the daughter nuclei by different mechanisms in fungi and animal cells. Interactions of myosin-II with actin filaments produce force to assemble and then constrict the contractile ring to form a cleavage furrow. Contractile rings disassemble as they constrict. In some cases, knowledge about the numbers of participating proteins and their biochemical mechanisms has made it possible to formulate molecularly explicit mathematical models that reproduce the observed physical events during cytokinesis by computer simulations.

molecular_mechanism_of_cytokinesis

INTRODUCTION<!>MECHANISMS THAT POSITION THE CLEAVAGE FURROW<!>Contractile Ring Placement in Animal Cells<!>Seven-level signaling pathways control cytokinesis.<!>Regulation of timing by cell cycle kinases.<!>Two master regulatory complexes.<!>Ect-2 activates Rho-GTPase.<!>Rac GTPases.<!>Mechanisms localizing the master regulators.<!>Contractile Ring Placement in Fission Yeast Cells<!>Contractile Ring Placement in Budding Yeast Cells<!>MECHANISMS THAT ASSEMBLE CONTRACTILE RINGS<!>Contractile Ring Assembly in Fission Yeast Cells<!>Contractile Ring Assembly in Budding Yeast Cells<!>Contractile Ring Assembly in Animal Cells<!>MECHANISMS THAT CONSTRICT CONTRACTILE RINGS<!>Contractile Ring Tension and Constriction in Fission Yeast Cells<!>Measurements of ring tension in fission yeast.<!>Mechanism of ring tension production in fission yeast.<!>Mechanism of ring constriction in fission yeast.<!>Contractile Ring Tension and Constriction in Budding Yeast Cells<!>Mechanisms of ring tension production and constriction in budding yeast.<!>Contractile Ring Constriction and Disassembly in Dictyostelium<!>Structure of the contractile ring in Dictyostelium.<!>Mechanisms of ring tension production and constriction in Dictyostelium.<!>Contractile Ring Tension and Constriction in Animal Cells<!>Measurements of ring tension in echinoderm embryos.<!>Mechanisms of ring tension production and constriction.<!>Other Proposed Mechanisms of Ring Tension Production

<p>Cytokinesis is the most dramatic event in the life of a cell. After the mitotic apparatus partitions the duplicated chromosomes into separate nuclei, a cleavage furrow separates these nuclei into two daughter cells. Accurate partitioning of daughter nuclei into separate cells is required for viability and avoiding aneuploidy.</p><p>Pioneering experiments four decades ago established that animal cells assemble a contractile ring composed of actin filaments (1, 2) and myosin-II (3) around the equator of the dividing cell and that both proteins (1, 4) are required to form a cleavage furrow. Mechanical measurements established that the cleavage furrow of echinoderm eggs produces a force of ~25 pN (5), enough to divide the cell but much less than contracting muscle cells. Genetics experiments showed that cytokinesis by cellular slime mold cells (6) and fission yeast cells (7, 8) also depends on actin and myosin-II. Much has been learned about the proteins required to assemble and constrict cytokinetic contractile rings (9–11), but many fundamental questions still remain (12).</p><p>This review focuses on the assembly and mechanics of cytokinetic contractile rings (Figure 1). Contractile rings appeared in the common ancestor of animals, fungi, and amoebas approximately 1 billion years ago concurrent with the evolution of myosin-II. Plants, algae, and vast numbers of single-celled organisms lack myosin-II (13) and depend on fusion of membrane vesicles (14) or other means to divide. For example, forces produced by intracellular flagellar axonemes divide the early branching eukaryote Giardia (15). Metazoan organisms redeployed the cytokinesis proteins in contractile cells that evolved into muscle cells. Our deep knowledge of muscle contraction strongly influences thinking about cytokinesis, although some caution is required.</p><p>Fortunately, most genes required for cytokinesis are ancient, so insights from work on model organisms such as budding yeast, fission yeast, and nematode zygotes often apply to vertebrates. The mechanisms are not identical owing to hundreds of millions of years of divergent evolution. We evaluate what is known about the molecular mechanisms of four aspects of cytokinesis: location, assembly, constriction, and disassembly of the contractile ring. Others have reviewed abscission (16, 17), the final separation of the membranes of the daughter cells. Prokaryotes use different mechanisms for cytokinesis (18).</p><!><p>Placing the plane of cell division between the daughter nuclei is essential to assure viability, so cells evolved robust, apparently redundant mechanisms to assure the fidelity of the process. These mechanisms appear to have diverged more than any other aspect of cytokinesis, so budding yeast, fission yeast, and animal cells use different strategies to separate daughter nuclei into two cells. The mitotic spindle is the prime source of positional information in animal cells (19). Fission yeast cells use cues from the poles of the cell and the nucleus to place the contractile ring. Budding yeast cells use polarity signals to position the site where the daughter cell buds from the mother cell and then later guide one set of chromosomes into the bud before cytokinesis.</p><!><p>Micromanipulation of echinoderm zygotes by Rappaport and Hiramoto showed that chemically undefined signals move from the mitotic spindle to the cortex of the cell, where after a couple of minutes, they stimulate the formation of a cleavage furrow (20). Other experiments showed that one mitotic spindle could stimulate multiple furrows. As furrows can form between spindle poles that do not straddle chromosomes, Rappaport argued that the signal must come from the poles and travel along the astral microtubules. However, equally persuasive evidence showed that the middle of the mitotic spindle also provides positioning information in small animal cells (21). Reexamination of the process in large echinoderm and frog zygotes (22) led to the conclusion that the "spindle midzone must be a source of a diffusible signal, which, under normal conditions, is spatially focused in the equator by astral microtubules" (23, p. 4049). This positioning mechanism stimulates the assembly of actin filaments and myosin-II associated with the plasma membrane in between the poles of the mitotic spindle, whether they are located centrally or positioned asymmetrically Given this conceptual framework, investigators set out to find the molecules that position the cleavage furrow.</p><p>Experiments on zygotes of the nematode Caenorhabditis elegans have been particularly informative about the signaling pathways that specify the site of cytokinesis in animal cells (19) (Figure 2). Parallel studies on flies (24), echinoderm zygotes (25), frog zygotes (26), and cultured fly cells (27) have yielded complementary information. Given the widespread distribution of these species on the phylogenetic tree, many animals must use the pathways shown in the figure but with some variation in the mechanistic details (19). No cell type has been shown to use all of the mechanisms, so our understanding of the general principles is still forming. Some aspects of the positioning mechanism have been reconstituted in frog egg extracts (28), but much work is still needed on both the biochemical and the cellular aspects of the positioning mechanisms.</p><!><p>Two decades of research identified multiple biochemical pathways targeting Rho GTPases to position and assemble contractile rings (Figure 2). Cell cycle kinases including both cyclin-dependent kinases (Cdks) and Polo-like kinases at the top coordinate the whole process by controlling two master regulatory complexes, the centralspindlin complex and chromosomal passenger complex (CPC). These multi-subunit complexes specify the location of the cleavage site and activate downstream signaling proteins including the key guanine nucleotide exchange factor Rho-GEF named Ect-2, Rho family GTPases, and proteins that act locally to assemble the contractile ring.</p><!><p>Strong Cdk1/cyclin B activity drives the cell into mitosis but suppresses cytokinesis until anaphase by phosphorylating sites on proteins that inhibit various steps (29–31). For example, phosphorylation of kinesin-6 in the centralspindlin complex inhibits binding to microtubules (32), and phosphorylation of Ect-2 Rho-GEF inhibits its activity and interactions with centralspindlin and membranes (33–35). Degradation of cyclin B allows the transition from metaphase to anaphase and releases the pathways controlled by the two master regulatory complexes. Simply inhibiting Cdk1 (but not other Cdks) during mitosis can drive premature cytokinesis (29, 30).</p><p>The Polo-like mitotic kinase can both inhibit and promote interactions relevant to cytokinesis. For example, phosphorylation of the microtubule cross-linking protein PRC1 inhibits formation of the central spindle until anaphase (36), whereas phosphorylation of the centralspindlin protein MgcRacGAP allows it to bind and activate Rho-GEF Ect-2 during anaphase (37).</p><!><p>The centralspindlin complex (38) is essential for positioning the contractile ring. It consists of a homodimer of the kinesin-6 isoform Kif23 and a homodimer of the Rac GTPase-activating protein (GAP) named Cyk-4 or MgcRacGAP (39). The kinesin-6 motor protein steps toward the plus ends of microtubules (40). MgcRacGAP has at least four functions. First, it binds kinesin-6 and promotes microtubule cross-linking in the central spindle (39). Second, the GAP domain can inactivate Rac and Cdc42 by promoting GTP hydrolysis in biochemical assays (41, 42). Third, a C1 domain interacts with acidic lipids to anchor the whole complex to the plasma membrane (43). Fourth, it binds the Rho-GEF Ect-2, the pivotal activator of the Rho-GTPase that controls the assembly of the contractile ring. This interaction depends on the Polo-like kinase phosphorylating S157 and A164 of MgcRacGAP (37), creating binding sites for the BRCT (BRCA1 C-terminal domain) domains of Rho-GEF Ect-2 (44). MgcRacGAP not only concentrates Ect-2 at the cleavage site but also activates its GEF activity (45).</p><p>The CPC consists of four proteins: borealin, survivin, INCENP, and Aurora B kinase (46). The CPC concentrates at centromeres early in mitosis, but when Cdk activity declines during anaphase, INCENP is dephosphorylated and kinesin-6/Kif20A targets CPC to the overlapping microtubules of the central spindle. Some CPC accumulates in the cortex surrounding the central spindle. During telophase, CPC concentrates in the midbody. The Aurora B kinase in the complex regulates several aspects of mitosis including attachment of microtubules to kinetochores, the mitotic spindle assembly checkpoint, and concentration of the centralspindlin complex to the spindle midzone (46). Aurora B kinase also promotes contractile ring formation by inactivating a 14—3–3 protein that inhibits the oligomerization of centralspindlin (47) (Figure 2).</p><!><p>The primary positive signal to assemble the contractile ring comes from Ect-2 (GEF) that activates Rho GTPase (48, 49). In fact, local activation of RhoA induces a furrow anywhere on the cell surface (50). Autoinhibition (51) and phosphorylation of threonine 341 by Cdks (33–35) keep Ect-2 activity low during G2 and early mitotic phases. When this phosphate is removed in anaphase, Ect-2 associates with phosphorylated MgcRacGAP of the centralspindlin complex and concentrates along with active Rho-GTP on the plasma membrane adjacent to the middle of the mitotic spindle (35, 52). A PH (pleckstrin homology) domain and a region of basic residues allow Ect-2 to bind to the plasma membrane, an interaction required to activate Rho (35). Rho-GAPs named RGA-3/4 and MP-GAP/ARHGAP11A turn off activated Rho by stimulating GTP hydrolysis (53–55).</p><p>C. elegans has a second, nonessential pathway dependent on the nop-1 gene that also activates Ect-2 in early embryos (45, 56). This uncharacterized pathway complicates the analysis of the cellular functions of centralspindlin unless NOP-1 is depleted.</p><!><p>Rac GTPases influence cytokinesis indirectly by activating a pathway stimulating the Arp2/3 (actin-related protein 2 and 3) complex to form branched actin filaments that may interfere with furrow formation (57). Because depleting Rac mRNAs or activators of Arp2/3 complex can compensate for certain mutations in the GAP domain of MgcRacGAP in C. elegans, the GAP activity may be responsible for inactivating Rac during furrowing (58, 59). However, other experiments suggest that the GAP domain (but not always the GAP activity) is required to activate Ect-2 (45).</p><!><p>The mitotic spindle determines where the master regulators—centralspindlin complex and CPC—accumulate and thus determines the position of the cleavage furrow (19). Various types of cells emphasize different aspects of the known mechanisms, which are still not fully understood.</p><p>Kinesins move both complexes toward the plus ends of microtubules. The kinesin-6 subunits move the centralspindlin complex and associated Ect-2 Rho-GEF (23, 60), whereas another kinesin-6 (Kif20AE) and a kinesin-4 (Kif4A) move the CPC (28). Consequently, both master regulators accumulate along with Ect-2 at two locations (Figure 3) where the protein PRC1 and kinesin-Kif4 cross-link the plus ends of overlapping antiparallel microtubules (61). The universal concentration of Ect-2 and centralspindlin in the middle of the anaphase mitotic spindle is striking but not required to direct equatorial furrow formation in some cells (62).</p><p>More important is the localization of Ect-2 and centralspindlin on the plasma membrane, where astral microtubules overlap midway between the poles of the mitotic spindle (23). In small cells, active Ect-2 may diffuse locally from the central spindle to the membrane (23, 60), but in large egg cells, the overlapped astral microtubules are the likely source of Ect-2 on the plasma membrane (23, 60). Less is known about the accumulation of CPC on the equatorial plasma membrane, where it contributes to stabilizing the zone of active Ect-2 by promoting oligomerization of centralspindlin due to inhibitory phosphorylation of 14—3–3 by Aurora B (47) (Figure 2). Ect-2 also binds to lipids and membrane-anchored Rho-GTP. The Ect-2 activates Rho and drives the assembly of the contractile ring (63). Simultaneously, anaphase movements of chromosomes carry a PP1 protein phosphatase to the poles of the mitotic spindle, where it may contribute to relaxing the overlying cortex by dephosphorylating the ERM protein moesin and reducing actin polymerization (64).</p><p>Remarkably, some of these positioning events can be reconstituted with purified proteins (65) and in frog egg extracts with artificial asters of microtubules (28). Kinesin-4 and kinesin-6 motors concentrate the CPC, the centralspindlin complex, and active Rho-GTPase, where the plus ends of microtubules from adjacent asters overlap. Active Rho stimulates actin polymerization next to these overlap zones. In the extracts, the microtubules stop growing when they encounter oppositely polarized microtubules from an adjacent aster. Thus, cross-links formed by PRC1 between antiparallel microtubules may regulate the elongation of the microtubules and help to localize the proteins that control the assembly of the contractile ring.</p><p>In addition to positive signals guided by the mitotic apparatus, 50 years of biological observations suggested that astral microtubules participate in generating a negative signal that suppresses furrow formation around the poles of the cell but allows the activation of Rho around the equator (19). A candidate mechanism is recruitment of the protein TPXL-1 to the tips of astral microtubules, where it activates Aurora A kinase to inhibit the local accumulation of contractile ring proteins by uncharacterized pathways (66).</p><!><p>Fission yeast cells establish the cleavage plane early in the G2 phase of the cell cycle by positioning the precursors of the contractile ring in the middle of the rod-shaped cells (9) (Figure 3). These precursors are protein assemblies termed nodes bound to the inside of the plasma membrane. Two types of nodes comprised of different proteins form separately but join together during the G2 phase of the cell cycle and mature into the cytokinesis nodes that form the contractile ring (see the next section) (67).</p><p>Type I nodes form around the equator early in the G2 phase. The Cdr2 kinase is the main constituent and scaffold of type I nodes; smaller numbers of the protein kinases Cdr1p and Wee1p bind Cdr2p and contribute to regulating cell size (68–70). C-terminal lipid-binding sites anchor Cdr2p to the plasma membrane. During interphase, anillin Mid1p leaves the nucleus and joins the stationary type I nodes around the equator (71). Adjacent C2 and PH domains at the C terminus of Mid1p interact with membrane lipids (72).</p><p>Type II nodes consist of the membrane-binding scaffold protein Blt1p, Rho-GEF Gef2p, and kinesin Klp8 (variously classified as a kinesin-3 or -10). These nodes are present in the constricting contractile ring and are released on the sides of the cleavage furrow into the daughter cells as the ring disperses at the end of cytokinesis (70). They diffuse from the division site along the plasma membrane until they encounter and bind a stationary type I node near the center of the cell (67). The combined type I and type II nodes mature into cytokinesis nodes by accumulating the proteins that assemble the contractile ring.</p><p>Signals from both ends of the cell exclude type I nodes from the poles, restricting them to the middle of the cell. The Pom1 kinase generates part of this repulsive signal (68, 69, 73). Gradients of Pom1p extending from both poles are established by three mechanisms: microtubules transport Pom1p along with Tea1p and Tea4p to the poles; a PP1 phosphatase associated with Tea4p removes a phosphate from Pom1, which allows it to bind the membrane; and Pom1p phosphorylates other Pom1p molecules at the poles to promote dissociation from the membrane (74). Simulations of a mathematical model of these reactions reproduced the observed gradient of Pom1p at the tips of the cell (74).</p><p>Pom1p repels type I nodes in proportion to its local concentration by phosphorylating Cdr2p, which reduces its interactions with the membrane and Mid1p and inhibits its kinase activity (75). Therefore, nodes only form in the middle of the cell where the Pom1p concentration is low (68, 69). However, other factors must participate, because cells lacking Pom1p exclude type I nodes from only one pole of the cell (68, 69, 76, 77). Another kinase, Kin1p, concentrates at the opposite pole from that with more Pom1p and phosphorylates similar substrates but on different residues (78).</p><p>Manipulation of the position of the nucleus inside living cells shows that it can secondarily adjust the division site. Normally this is not required, as the type I nodes are located around the nucleus in the middle of the cell during interphase. Dynamic instability of microtubules attached to the nuclear envelope positions the nucleus in the middle of the cell by pushing against the ends of the cell (79). Centrifugation or optical traps can displace the nucleus if microtubules are depolymerized with a drug or absent from the cytoplasm as in the late G2 phase (80, 81). A contractile ring forms at the appropriate time wherever the displaced nucleus is located.</p><p>As the absence of anillin Mid1p results in poorly positioned contractile rings (82, 83), and as Mid1p moves from the nucleus to nodes during interphase (71, 84), many authors have proposed that the local transfer of Mid1 from the nucleus to nodes in the adjacent cortex dominates over the repulsive signal from the poles to position the contractile ring as the cell approaches mitosis. At that point, polo kinase Plo1p phosphorylates one of the nuclear export signals of Mid1p and promotes its transport from the nucleus (85). Consistent with this hypothesis, cortical nodes marked with Mid1p-green fluorescent protein (GFP) move adjacent to a displaced nucleus (80). The simplest explanation would be that interphase nodes with Cdr2p and Mid1p move intact or turn over and reappear next to the displaced nucleus. However, the type I interphase node scaffold protein Cdr2p is reported to stay behind, in the middle of the cell, when the nucleus is displaced (86). Another experiment supporting the local transfer hypothesis used a mutated Mid1p lacking both nuclear localization signals, so it never entered the nucleus. These cells place nodes with mutant Mid1p and the contractile ring correctly in the middle of the cell, even if the nucleus is displaced to one end (presumably because no Mid1p is available for local transfer from the nucleus) (86). We do not yet understand how this nuclear positioning mechanism works (87).</p><p>During mitosis the septation initiation network (SIN) of signaling proteins, the yeast version of the animal Hippo pathway (88), disperses the type I node proteins from the cytokinesis nodes into the cytoplasm by phosphorylating the Cdr2p kinase (89). Type I nodes reassemble on the plasma membrane around the daughter nuclei when the SIN activity declines at the end of mitosis (90).</p><!><p>Budding yeast cells proliferate by growing a cellular protrusion termed a bud on the side of a mother cell. During the G1 phase of the cell cycle, a bud emerges near the site of the previous cytokinesis (Figures 1 and 3). This well-studied example of cellular polarization depends on Cdc42, the most conserved of the small GTPases, and a positive feedback mechanism that concentrates active Cdc42-GTP on the inside surface of the plasma membrane (91) (Figure 3). The positive feedback works like this: Cdc42-GTP, associated with the plasma membrane by a C-terminal prenyl modification, binds a cytoplasmic kinase named PAK that is associated with a scaffold protein Bem1 and Cdc24, the GEF for Cdc42. The GEF activates nearby Cdc42 proteins, which in turn bind more of the activating GEF, to amplify the process. Simulations of mathematical models showed that rapid diffusion of the activating complex in the cytoplasm, binding to active Cdc42 on the membrane, and local activation of more slowly diffusing Cdc42 on the membrane build up a high concentration of active Cdc42 (92). In keeping with this mechanism, optogenetic targeting of Bem1 or Cdc24 to a spot on the plasma membrane can trigger the self-amplifying local accumulation of active Cdc42-GTP (93).</p><p>Both positive and negative cues assure that the new bud emerges next to the bud scar where the mother and daughter cells separated (91). A Cdc42 GAP named Rga1 is left behind at the cleavage site as a negative cue to block Cdc42 accumulation. Positive cues are positioned next to the cleavage site. They include transmembrane proteins from the secretory pathway that accumulate around or next to the cleavage site. One of these proteins binds and activates a GTPase named Rsr1, which activates Cdc24 GEF to trigger a local accumulation of Cdc42-GTP. Active Cdc42 recruits effector proteins including formins, which contribute to the assembly of the precursor of the contractile ring (see the next section).</p><!><p>Contractile rings from amoebas to humans are comprised of the same major proteins. In all well-studied systems, formins assemble the actin filaments, and myosin-II produces force to arrange the filaments in a continuous ring at the cleavage site. Biochemical reconstitution established that the main components of contractile rings can self-assemble; polymerizing actin with a crowding agent (methylcellulose) produces rings of actin filaments inside aqueous droplets lined with membrane lipids and surrounded by oil (94). Including myosin-II with actin promotes the formation and constriction of rings of actin filaments.</p><p>Contractile ring proteins are present at constant concentrations throughout the cell cycle, at least in fission yeast (95), so conditions change during mitosis to direct the assembly of the contractile ring. The following sections explain the different strategies used by fission yeast, budding yeast, and animal cells.</p><!><p>Fission yeast assemble a contractile ring during a period of 10–12 min early in mitosis. Fungi lack a cortical actin filament network during interphase, so the contractile ring actin filaments are assembled de novo. During the 10 min prior to separation of the spindle pole bodies in prophase, cytokinesis nodes formed during interphase from type I and type II nodes accumulate the proteins required to form a contractile ring: IQGAP Rng2p links anillin Mid1p to the C terminus of myosin-II Myo2 (96), and F-BAR protein Cdc15p interacts with Blt1p, Mid1p, Rng2p, and formin Cdc12p (97). Super-resolution fluorescence microscopy of live cells expressing fluorescent fusion proteins showed that a cluster of approximately 10 Mid1p proteins anchors approximately 10 dimers each of IQGAP Rng2, F-BAR Cdc15p, and myosin-II Myo2 to the plasma membrane with a smaller number of formin Cdc12p dimers. A bouquet of myosin heads project into the cytoplasm (Figure 4). An actin filament grows from each formin (82, 98), which presumably anchors its barbed end to the node. Formin For3p contributes additional actin filaments (99). If an actin filament growing from a formin comes close to another node, Myo2 can bind and pull the two nodes together (100). A myosin-V Myo51, located between nodes augments the force produced by Myo2 (101, 102).</p><p>Simulations of a molecularly explicit mathematical model based on the number of proteins present and their biochemical activities showed that a search, capture, and pull mechanism assembles rings in 10 min as observed in cells, providing the connections between the nodes break approximately every 20 s to avoid clumping of nodes (100). Experiments with cofilin mutants with low severing activity showed that cofilin is required to break these connections (103). This release step is an error-correcting mechanism. A 3D simulation of ring assembly on a cylindrical representation of the cell produced contractile rings from nodes polymerizing semiflexible actin filaments bundled by cross-linking proteins (104). After assembly, the ring adds unconventional myosin-II Myp2, more F-BAR Cdc15p, and more polymerized actin during a 20-min maturation period before the onset of constriction (95, 99). Anillin Mid1p is replaced by a related protein Mid2p.</p><!><p>The GTPase Cdc42 serves as the polarity marker to specify the assembly of a stable collar of septins on the plasma membrane at the border between the mother cell and bud (105). The septins organize the assembly of the contractile ring. First, septins bind an adapter protein named Bni5 that links to the tail of myosin-II (named Myo1). Later, polo kinase stimulates a GEF to recruit the GTPase Rho1, which activates formin Bni1p to polymerize actin filaments. As the cell cycle progresses, IQGAP takes over from Bni5 to anchor Myo1, and the collar of septins splits into two rings with the contractile ring in between.</p><!><p>Active Rho-GTP drives contractile ring assembly in animal cells by activating Rho-kinase (ROCK) to phosphorylate the regulatory light chains of myosin-II (106) and by activating formins to assemble actin filaments (107). Myosin-II accumulates in bipolar filaments (108, 109) at the site of furrowing independent of actin filaments (110, 111) and even in the presence of the myosin inhibitor blebbistatin (112). The myosin tethering mechanism is not known. Genetics identified the crucial formins in C. elegans (Cyk-1) (24) and Drosophila (diaphanous) (113), but multiple formins appear to contribute in vertebrate cells (114). In addition, preexisting cortical actin filaments formed by the Arp2/3 complex (115) may also contribute to contractile rings.</p><p>Interactions of myosin-II with actin filaments produce forces that generate cortical flow of actin filaments and clusters of myosin toward the future cleavage site and align the filaments around the equator (116). Electron micrographs confirmed that myosin filaments interact with actin filaments during the formation of contractile rings in sea urchin zygotes (109). This arrangement is consistent with a search, capture, and pull mechanism, but the molecular details must be verified and incorporated into models and simulations of contractile ring assembly for animal cells (116).</p><!><p>Contractile rings containing actin, myosin, and other components (as described above) constrict and generate inward contractile force (5, 117–120) that mechanically drives or guides furrow ingression and cell division. Rappaport (5) measured an inward radial force of ~25 nN in the cleavage furrow of echinoderm eggs, and Schroeder identified actin filaments in the contractile ring and suggested a sliding filament mechanism similar to muscle (2, 121). Subsequently myosin-II was identified in the contractile ring (3) and shown to be required for constriction (4).</p><p>Although the evidence pointed to a muscle-like mechanism of tension generation and constriction, many questions remained. First, how does tension emerge from the organization of contractile rings, which appears remote from the spatially periodic sarcomeric organization of striated muscle? Second, given tension production in the ring, what mechanism sets the constriction rate and what is the role of other processes that accompany constriction such as change in cell shape, remodeling of the cell cortex in animal cells and amoebas, and septum growth in organisms with cell walls? Third, as contractile rings disassemble as they constrict (95, 121), how do they maintain functionality as a contractile machine while shedding their parts?</p><p>Here, we review the current status of understanding of these questions for four model systems. Investigators have proposed models of ring tension and constriction that can be tested by computer simulations and perturbing live cells. We emphasize the importance of experimental measurements of the organization of components in the ring and of ring tension, if realistic quantitative models of the fundamental mechanisms are to emerge.</p><!><p>For fission yeast significant information on the composition and structure of constricting rings is available, so that realistic models are now feasible. At the onset of constriction, the fission yeast contractile ring has a cross section of ~ 125 nm and a circumference of 11 μm. It consists of approximately 200,000 molecules of actin (500 μm of polymer) (99, 122, 123), 5,000 myosin-II polypeptides [3,000 of Myo2 (1,500 dimers) and 2,000 of Myp2], 650 IQGAP dimers, 250 α-actinin dimers, and 200 formin Cdc12p dimers (95, 99, 124). Assuming every formin dimer initiates an actin filament, the ~350 filaments have a mean length of ~ 1.4 μm in a bundle of ~50 filaments in a cross section (96).</p><p>Punctate structures similar in size and composition to cytokinesis nodes persist in constricting rings (96) and anchor actin filaments and Myo2 to the plasma membrane (see the sidebar titled Radial Anchoring of Constricting Rings). The number of nodes is likely 120–140, as the number of Myo2 molecules changes little during maturation (95). Myp2 lies farther from the membrane than Myo2 (102, 124), and its presence in the ring depends on actin filaments (125). In cells with the myo2-E1 temperature-sensitive mutation, straight bridges containing actin filaments and Myp2 detach from the plasma membrane (102) (Figure 4d). Nevertheless, despite Myo2p-E1 having minimal ATPase activity (126), these rings appear tensile as they constrict in myo2-E1 protoplasts (120). Thus, even when manifestly unanchored, Myp2 still appears to exert tension, so that being unanchored is likely its normal state (see also Figure 4c).</p><p>Rings shed material as they constrict while increasing the densities (number per ring length) of node components (Myo2, IQGAP Rng2p, F-BAR Cdc15p, and formin Cdc12p), of Myp2, and of α-actinin (95). The number of polymerized actin molecules falls in proportion to the circumference (99), suggesting the mean filament length decreases.</p><!><p>Contractile ring tension was measured in protoplasts (see the sidebar titled Fission Yeast Protoplasts), whose cell walls were removed by enzyme digestion (120). A mean ring tension of ~400 pN was inferred from a force balance at the furrow produced by the contractile ring, using the membrane tension measured by micropipette aspiration during interphase. A more recent study found that the membrane tension is higher during mitosis, showing that the average ring tension was ~640 pN and increased ~2-fold during constriction (S. Wang, H.F. Chin, S. Thiyagarajan, Z. McDargh, E. Karatekin, et al., unpublished manuscript). Average ring tensions were 400 pN in protoplasts of Δmyp2 cells and 220 pN in protoplasts of myo2-E1 cells that lack Myo2 ATPase activity, suggesting that both myosin-II isoforms contribute to the total tension in wild-type cells.</p><!><p>Stochastic simulations of molecularly explicit 2D (120) and 3D (S. Wang, H.F Chin, S. Thiyagarajan, Z. McDargh, E. Karatekin, et al., unpublished manuscript) models produced tensions similar to the experimental measurements in protoplasts. In the initial 2D model, the amounts and the biochemical properties of proteins in the ring were taken from experiment, and formins and node-like clusters of Myo2 were assumed to be independently anchored to the plasma membrane. The model did not include Myp2. Myo2 exerted 1 pN per head on an actin filament it binds, similar to muscle myosin (127).</p><p>The simulations generated ring tension of ~400 pN, because formins associated with nodes anchored the barbed ends of the actin filaments to the plasma membrane. Thus, every actin-myosin interaction pulled the filament (productive tensile contribution) rather than pushing it (compressive contribution) (Figure 4a; see the sidebar titled A Mechanism for Ring Tension). The mechanism does not require regular muscle-like architecture, as actin filaments are made tense almost independently. Consistent with the mechanism, nodes marked by Myo2 move bidirectionally at ~22 nm/s in constricting rings (96), much slower than the ~350 nm/s maximum load-free actin gliding velocity (128). Thus, forces opposing node motion (likely drag on membrane anchors) present sufficient resistance for myosin-II to operate close to the maximum stall force. Simulations showed that turnover of actin and myosin is required to generate tension, because without turnover, ring components aggregated into unproductive clusters.</p><p>The simulations also explained how rapid turnover and self-organization enable rings to disassemble as they constrict without loss of functionality (Figure 5b; see the sidebar titled Disassembly and Maintenance of the Constricting Ring). Dissociating components were replaced by incoming components that continuously self-assembled into the ring organization, a tight actomyosin bundle maximizing actin-myosin interactions (120). A key self-organizing process was Myo2-mediated zippering of formin-nucleated actin filaments into the bundle. Thus, every ~30 s the ring disassembles and reassembles itself, a mechanism that would enable it to quasi-statically shorten during the ~20 min of constriction without traumatizing the organization and compromising ring tension. This mechanism requires turnover and self-assembly rates to far exceed the relative constriction rate.</p><p>A new 3D molecularly explicit, stochastic simulation of constricting fission yeast rings (S. Wang, H.F. Chin, S. Thiyagarajan, Z. McDargh, E. Karatekin, et al., unpublished manuscript) included Myo2 and Cdc12p colocalized in membrane-bound nodes, as shown by super-resolution microscopy (95, 96), and with Myp2 binding only actin filaments (Figure 4c). The molecule numbers and turnover rates of components were set by experiment (95, 99, 130). During simulations, the components self-organized into an actomyosin bundle ~130 nm in diameter with a dense core of 30–35 filaments bundled by the anchored Myo2 and a corona of ~10 additional filaments cross-linked to the core by unanchored Myp2. With stall forces ~1.7 pN and 1.1 pN per Myo2 or Myp2 head, respectively, simulations reproduced the mean tension of ~640 pN, the increasing tension as rings constrict, and the loss of tension and organization in mutants of Myo2 and Myp2. The tension mechanism required actin barbed-end anchoring (Figure 4a; see the sidebar titled A Mechanism for Ring Tension), but unanchored Myp2 contributed one-third of the total tension by migrating to locations of zero net polarity in the actin bundle (131) (Figure 4c).</p><p>An analytical 1D model of constricting rings based on nodes anchoring Myo2 and formin Cdc12p reproduced the observed node motions and showed that contractile rings require lateral anchoring for organizational stability and tension (129) (see the sidebar titled A Mechanism for Ring Tension). Solutions of partial differential equations evolved bidirectional movements of nodes at ~22 nm/s, as observed (96, 132). A realistic ~2 pN of stall force per Myo2 head reproduced the experimental tension of 390 pN (120). Anchor drag resisted Myo2 pulling, enabling filaments to develop tension (Figure 4a). Half of the tension derived from a stochastic version of the sliding filament concept (2, 121), with Myo2 clusters pulling sliding actin filaments past them. The other fixed filament contribution was produced by chains of interconnected like-oriented nodes encircling the ring and also required firm lateral node anchoring, without which this contribution was rapidly destroyed by clumping instability.</p><p>Another molecularly explicit 3D simulation of the fission yeast contractile ring represented myosin-II as both two-headed membrane-anchored and four-headed unanchored molecules (133). Beads representing myosin heads pull actin filaments represented by flexible chains of beads, cross-linked by springs. The numbers of myosins and actin filaments in the cross section match experimentally measured values (95, 123). After testing different anchoring and clustering schemes, the authors favored a model with unclustered, anchored myosins, unanchored myosin, and unanchored actin. This model generated ~250 pN of tension. A node-like organization (clustered anchored myosin and actin) was rejected, because it produced severe membrane puckering. The node-induced puckering appears to result from a model rule that the surface representing the membrane and cell wall grows inward only where force is applied, but we are unaware of evidence supporting such a rule in cells.</p><!><p>Fungi are enclosed by a cell wall, so cytokinesis must couple ring constriction to growth of a new cell wall, termed the septum, in the wake of the constricting ring (134, 135) (Figure 6b). The membrane is pinned against the cell wall by an internal turgor pressure of ~ 1.5 MPa (136), and the leading edge of the septum, the membrane lining furrow, and the contractile ring constrict together at the rate of inward radial septum growth, ~1.1 nm/s (130, 137).</p><p>Septum growth sets the constriction rate, because the tension produced by a constricting contractile ring [T ~400 pN in protoplasts (120)] is no match for either the turgor pressure or the cell wall stiffness E ~50 MPa (136) (see the sidebar titled The Relation Between Ring Tension and Construction Rate). A ring 125 nm wide (96) with radius R ~1 μm produces an inward pressure ~T/wR ~2 kPa, almost three orders of magnitude below the reported turgor pressure, and negligible strains in the septum material ~T/wRE ~0.01% (130).</p><p>However, ring tension might influence the β-glucan and α-glucan synthases Bgs1–4 and Ags1 that synthesize the septum (134). Assuming that septal growth is mechanosensitive and mechanically coupled to the ring, a mathematical model reproduced characteristic edge roughness of septum edges in live cells and showed that ring tension would then modify septal growth rates in a curvature-dependent fashion, suppressing roughness and maintaining a nearly circular septum hole (Figure 6c) (130). The model predicted a Bgs1 mechanosensitivity of ~0.15 pN−1, the relative increase in synthesis rate per applied force (130). Consistent with curvature-dependent growth, septa grew faster in the more curved portions of elliptically shaped septum edges in mechanically deformed fission yeast cells (138).</p><p>In permeabilized protoplasts that lack cytoplasm, entire sections of rings detached from the membrane and shortened ~30-fold faster than rings in intact cells until becoming straight (139) (see the sidebar titled Fission Yeast Protoplasts). Both Myo2 and Myp2 contributed to this shortening, but neither actin assembly nor actin disassembly was required. A molecularly explicit simulation accurately reproduced and explained the experimental constriction rate and showed that segments of the bundle with actin filaments anchored by their barbed ends reel in unanchored actin filaments (S. Wang, B. O'Shaughnessy, unpublished manuscript) (Figure 4b). Simulations with other hypothetical anchoring schemes failed to constrict partially detached rings. Thus, the experimental findings and modeling explicitly support the barbed-end anchoring mechanism of tension generation (Figure 4a).</p><!><p>Budding yeast cells assemble contractile rings from the same proteins as fission yeast but by a different pathway (explained above). Isolation of partially purified contractile rings confirmed the presence of actin, myosin-II (named Myo1), IQGAP, and an F-BAR protein (140). Electron micrographs show filaments thought to be comprised of myosin-II around the bud neck flanked by networks of septins (141).</p><p>Fluorescence recovery after photobleaching experiments showed that contractile ring proteins exchange with cytoplasmic pools at widely different rates that vary across the cell cycle (142). Tropomyosin exchanges in seconds, but myosin-II (Myo1) does not exchange until the ring constricts. Normal disassembly of the ring during constriction depends on the anaphase-promoting complex APC/C, a ubiquitin ligase that also controls the transition from metaphase to anaphase (143).</p><!><p>Like fission yeast, ring constriction in budding yeast is coupled to inward growth of the septum (105), but surprisingly some strains manage to divide (with difficulty) after deletion of the Myo1 gene (144). Other strains require myosin-II, but the myosin heads, which interact with actin filaments, are dispensable (145), so septum growth and closure can occur even without forces from Myo1. The essential tails of Myo1 may serve as scaffolds.</p><p>A mathematical model of disassembly driven constriction assumed that the actin filaments depolymerize from their pointed ends and that cross-links of myosin and other proteins grab retreating pointed ends of overlapping filaments, so they slide as they depolymerize, tending to restore overlap and constrict the ring (146). The model accounted for experimental data showing correlations between rates of ring disassembly and constriction, including the independence of relative constriction rate on initial ring length. Molecular aspects of this model require testing.</p><!><p>Although Dictyostelium is more distantly related to animals than fungi, it lacks a cell wall and handles cytokinesis more like an animal cell (147). The protein toolbox for cytokinesis is the same as fungi and animals and includes actin filaments, formins, IQGAP, and myosin-II (148, 149). When the cells are grown in liquid suspension, myosin-II is essential for cytokinesis (6), but these highly motile cells can divide by pulling themselves in two on a solid surface without myosin-II.</p><!><p>Dictyostelium has a very broad cleavage furrow associated with a network of short (100-nm average length) cross-linked actin filaments (150). Approximately 10% of total myosin-II concentrates in cortex across this broad cleavage furrow (151) as puncta similar to thick filaments (149) and exchanges with a half time of 7 s (152). The minimum force required to cleave the cell was estimated to reach a peak value of ~6 nN (151).</p><!><p>A mathematical model treated the cortex as a viscoelastic solid and the cytoplasm as a viscoelastic liquid and incorporated passive and myosin-mediated cortical stresses and forces from adhesion, protrusion, and pressure (153). In simulations, both nonadherent and adherent cells elongated and developed ingressing furrows. Without myosin-II, furrows ingressed only in adherent cells. The authors concluded that myosin or traction-mediated protrusive forces can initiate furrow ingression, whereas passive cortical tension and surface curvature forces complete ingression. An analytical model representing the cell as two lobes connected by a cylindrical furrow assumed axial compressive stresses and inward radial myosin stresses (154). The model predicted that myosin exerts a furrow stress of 0.1 nN μm−2 and that constriction rates are set by myosin forces, Laplace pressure, and resistive forces controlled by RacE and dynacortin. Mechanosensitivity of myosin-II and the cross-linker cortexillin-I helps to recruit these proteins and stabilize the furrow (147).</p><!><p>In animal cells, a cortical zone of active RhoA drives the assembly of actin filaments and myosin-II (25). Actin filaments of both polarities are largely aligned with the ring (2, 155) and interdigitated with rods the size of myosin-II minifilaments (156). Super-resolution imaging confirmed that these rods and fluorescent spots of myosin-II in cleavage furrows (157, 158) are bipolar myosin-II filaments in contractile rings of vertebrate cultured cells (108) and sea urchin zygotes (109). The molecules anchoring the contractile ring to the plasma membrane are still under investigation (see the sidebar titled Anchoring Mechanisms).</p><p>The concentration of myosin-II–GFP in constricting rings of four-cell C. elegans embryos was approximately constant throughout constriction (57) rather than increasing as in fission yeast (95). Photobleached zones of myosin-II–GFP in such rings regained fluorescence over 60–90 s from the sides of the bleached segment, suggesting that myosin in the ring is laterally mobile but does not exchange rapidly. The actin-sequestering drug Latrunculin A blocked ring assembly and slowed the constriction when administered after rings formed. The authors suggested that actin filaments in constricting rings shorten without replenishment from the cortex or cytoplasm (Figure 5a; see the sidebar titled Disassembly and Maintenance of the Constricting Ring). However, actin and myosin-II flow from the cortex into the cleavage furrows of single-cell C. elegans embryos (116), and the concentration of myosin-II-GFP increases exponentially over time in their rings (159), a behavior attributed to compression of the cortex as it flows into the furrow. The constriction rate per unit length of ring (minus the strain rate) increases in parallel, suggesting that myosin-II is responsible for the increasing magnitude of the strain rate.</p><!><p>Rappaport (5) used microneedles to measure a net inward radial force at the furrow of ~ 10–5 0 nN. Hiramoto (118) measured similar forces from the deformation of a ferrofluid droplet. The net force sums the ring tension and opposing forces from the cortex, viscous drag, and possibly other sources (Figure 6a). Yoneda & Dan (119) used axial force to compress dividing echinoderm eggs to measure cortical tension and applied an approximate force balance at the furrow to estimate the minimum ring tension for ingression. The lower bound of tension peaked at 45–60 nN and decayed as rings constricted. This approach assumed equal spherical cap lobes and uniform cortical tension, but the cortical tension is likely to vary spatially (160).</p><!><p>Antibody injection (4) and treatment with the myosin inhibitor blebbistatin (112) established that myosin-II is normally required for contractile ring constriction in animal cells. However, Cos7 cells can complete cytokinesis with one myosin-II isoform lacking ATPase activity (161), so other myosin isoforms may participate. Insufficient molecular details are known to establish the molecular mechanism of ring tension, but it likely entails myosin-II pulling actin filaments anchored to the membrane at their barbed ends (Figure 4).</p><p>Active gel models do not explicitly describe molecules but address large-scale behavior of the actomyosin cortex in a continuum framework, with force produced by myosin represented by active stress terms (162). In a model of constriction of dividing sand dollar zygotes, the cortex was treated as an incompressible viscoelastic layer with contractility from an isotropic active stress term proportional to cortex thickness, which tends to be restored locally by turnover (163). An equatorial zone with elevated active stress, assumed to represent myosin-II activated by RhoA-GTP, drove constriction, which was resisted by cortical viscosity and osmotic pressure enforcing almost constant cell volume. The model reproduced the observed evolution of cell shape and showed that isotropic equatorial activation is theoretically sufficient to drive furrow ingression.</p><p>A model of asymmetric furrowing of C. elegans embryos represented actin properties with continuous fields and described membrane- and ring-bending forces, membrane tension, osmotic pressure, and viscous forces. The model assumed a positive feedback loop where contractile forces bend the membrane, which tends to align the actin filaments and increase force locally (164). Solutions reproduced asymmetric furrowing characteristic of these cells owing to the mutually reinforcing effects of membrane curvature and filament alignment.</p><p>Constriction rates follow a scaling law in some organisms, with faster constriction in larger cells. In C. elegans embryos, shortening rates in successive divisions are proportional to initial ring length so that constriction times are invariant, suggestive of possible organization of the ring in repeat contractile units reminiscent of striated muscle sarcomeres, such that longer rings have proportionally more units (57) (Figure 5a). However, the authors concluded (159) that their new observation of parallel increases in the concentration of myosin in the ring and the constriction rate in single cell C. elegans embryos does not support the contractile unit model. Constriction rates in the filamentous fungus Neurospora crassa also increase with cell size but less than proportionally, such that constriction takes longer in larger cells (165).</p><!><p>Fission yeast and possibly animals appear to generate contractile ring tension by myosin-II exerting force on actin filaments whose barbed ends are membrane anchored (Figure 4a). Several alternative possible mechanisms have been explored theoretically (discussed below).</p><p>Some theoretical mechanisms rely on internal cross-linking schemes between myosin and actin. A general 1D model assumed sliding actin filament pairs, represented either as rods or by a continuous density field, and using either bipolar myosin motors that track actin filament barbed ends or cross-linkers that track pointed ends of treadmilling filaments (166). Simulations generated tension, but to our knowledge, end-tracking myosins or cross-linkers have not been identified in cells. Another study assumed myosin is attached to unanchored actin filament barbed ends in interconverting monopolar and bipolar actin-myosin structures, described by continuous fields (132). Solutions for various parameters showed rings with stationary and/or moving clusters or homogeneous rings, reminiscent of experiments showing circumferentially moving clusters in fission yeast but not in HeLa cells (132). This model produces ring tension, but we are unaware of the proposed structures or kinetics in cells.</p><p>Another proposal is based on actin filament buckling. A 1D mean field model described actin filaments subject to forces from myosins assumed to follow a distribution of force-velocity relations (167). In solutions of the model, some compressive segments buckled, leaving tensile segments in the majority and rendering the bundle tensile. Some filaments buckled during contraction of reconstituted actomyosin bundles, but it remains to be seen if such buckling is relevant to cells.</p><p>Actin filament treadmilling has been proposed as a factor in ring tension production. A 1D model represented actin filaments as arcs on a circle and myosins as points with two binding sites mediating filament sliding (168). Actin filaments treadmill, and newly polymerized actin must wait to bind randomly to a free myosin site, so that myosin becomes depleted near filament barbed ends and hence tension is generated. However, actin would presumably rapidly bind myosin, as myosin-II is not known to be processive and the mechanism of actin filament turnover remains to be established.</p>

PubMed Author Manuscript

Revisiting the Language of Glycoscience: Readers, Writers and Erasers in Carbohydrate Biochemistry

AbstractThe roles of carbohydrates in nature are many and varied. However, the lack of template encoding in glycoscience distances carbohydrate structure, and hence function, from gene sequence. This challenging situation is compounded by descriptors of carbohydrate structure and function that have tended to emphasise their complexity. Herein, we suggest that revising the language of glycoscience could make interdisciplinary discourse more accessible to all interested parties.

revisiting_the_language_of_glycoscience:_readers,_writers_and_erasers_in_carbohydrate_biochemistry

<!>Conclusion<!>Conflict of interest<!>

<p>S. Dedola, M. D. Rugen, R. J. Young, R. A. Field, ChemBioChem 2020, 21, 423.</p><p>The perception that glycoscience—the chemistry and biology of carbohydrates—is both complex and ubiquitous in nature1, 2 has led to the notion that "carbohydrates in molecular biology are like dark matter in the universe… poorly studied yet crucial to a full understanding of how things actually work".3 In contrast to nucleic acids and proteins (DNA codes for RNA codes for protein), the lack of template‐encoding disconnects the "glyco code"4 from direct gene sequence control. This results in carbohydrate biosynthesis and the biological function of glycans being dependent upon a series of protein–carbohydrate interaction events. Overall, the concerted actions of lectins, glycosyltransferases and/or glycoside hydrolases achieve the integrity of mature bioactive glycan structures. The intricacies of this landscape are made worse by the tendency of the glycoscience community to emphasise the complexities of the field, perhaps making it less accessible to the casual reader—the informed non‐expert—than it needs to be. The glycoscience community are not alone in this shortcoming, as highlighted by a Comment in Nature that suggests that "Antibiotic resistance has a language problem. A failure to use words clearly undermines the global response to antimicrobials′ waning usefulness".[5] Technological6 and informatics7 advances in glycoscience, alongside combinations of the two,8 are providing new ways to cut through the complexity, whilst comprehensive books of glycobiology topics provide entries in to the field.9 The introduction of stylized symbol nomenclature for glycans (SNFG; Figure 1) also represents an important step towards simplifying communication within and between interested disciplines10 along with guidelines for experimental design and data curation,11 and a repository for glycan structures.12</p><p>Representing glycan structures: simplification and standardization with stylized SNFG. Taken from ref. 10.</p><p>As discussed recently by Gabius,13 there are notable parallels between aspects of glycobiology and the epigenetic regulation of chromatin structure and function. The latter processes, which occur with exquisite precision, are typically referred to in stripped‐down terms as a series of read, write and erase events, making the field immediately accessible to outsiders. Indeed, this approach emulates computer programming's create, read, update and delete (CRUD)14—the four basic functions employed for persistent data storage.15 Herein, we consider the potential to recapitulate glycoscience language in the terms of epigenetic vocabulary.</p><p>In simple terms, epigenetics concerns small chemical changes (marks) in the chemical structure of chromatin—typically the histone proteins that organize and package DNA in chromosomes.16 Dynamic changes in these epigenetic protein marks impact on the physical accessibility of gene sequences for expression, rather than on the alteration of the genetic code per se. The profound biological consequences of these processes have attracted enormous attention over the past decade, given their central role in life and their disruption in disease.17 The molecular hallmarks of epigenetic regulation comprise a dynamic series of enzymatic modification steps that introduce or remove marks to the histone protein structure. Epigenetic writers, which introduce epigenetic marks on amino acid residues of the histones, include histone acetyltransferases (HATs, which N‐acetylate lysine), histone methyltransferases (HMTs, which N‐methylate lysine), protein arginine methyltransferases (PRMTs) and protein kinases (which O‐phosphorylate serine/threonine), amongst others. Epigenetic readers, which bind to epigenetic marks and amplify their impact on DNA packaging and hence gene accessibility for expression, include proteins containing bromodomains, chromodomains and Tudor domains. Epigenetic erasers, such as histone deacetylases (HDACs), lysine demethylases (KDMs) and phosphatases, catalyse the removal of epigenetic marks (Figure 2).</p><p>Epigenetic writers, readers and erasers. DNA packaged around histones gives a condensed genomic information package (top) that can be selectively unwound by epigenetic modification (e.g., acetylation, methylation of phosphorylation) to expose genes for transcription (turn on). Abbreviations used are given in the text. Adapted from ref. 17a.</p><p>The impact of the lysine N‐acetylation epigenetic mark is perhaps simplest to appreciate. Writing this mark results in the loss of a positive charge on the lysine side chain of a histone, thus removing the potential for interaction with the negatively charged DNA backbone and causing loosening the DNA–histone complex. The resulting opening up of the chromosome structure enables the localized activation (turning on) of gene expression. In the opposite sense, erasing a lysine acetylation mark drives a tighter assembly of the histone–DNA complex and silencing (turning off) gene expression.</p><p>The general principle of readers, writers and erasers prompts consideration of potential parallels between epigenetics and the control of glycan biosynthesis, structure and function. That is, does the notion of lectin readers, glycosyltransferase writers and glycosyl hydrolase erasers ring true in glycobiology? A convenient segue from epigenetics into glycoscience is provided by the reversible O‐GlcNAc modification of Ser/Thr residues in proteins.18 This central metabolic "rheostat"19 comprises a nutrient status‐responsive, post‐translational modification that impacts on protein–protein and protein–nucleic acid interactions. In turn regulating of cellular events including transcription and signal transduction, with implications in diabetes, Alzheimer's disease and cancer.</p><p>So how does the O‐GlcNAc cycle work? O‐GlcNAc transferase (OGT) writes and O‐GlcNAcase erases, providing a simple and reversible modification cycle that is orthogonal to protein phosphorylation and which has far‐reaching physiological impact (Figure 3).20</p><p>The O‐GlcNAc cycle and its impact on the modulation of cellular processes. Adapted from ref. 19.</p><p>In addition to glycosyltransferase writers and glycosyl hydrolase erasers, there are also potential readers in glycoscience—a function performed by lectins21 and the carbohydrate‐binding modules (CBMs)22 in multidomain CAZymes. The full read, write, erase combination in glycoscience is most easily exemplified by the proofreading and editing cycle associated with N‐linked glycoprotein biosynthesis. These processes are essential to ensuring the correct integrity and dynamics of cell‐surface glycoproteins, which contribute to the glycocalyx that dominates cell–cell interactions in the maintenance of healthy tissue and which underpin sperm–egg interactions during fertilisation, but which also serve as cellular receptors for a wide range of microbial pathogens.9</p><p>Asparagine‐linked protein N‐glycosylation starts in the endoplasmic reticulum, whereas the peptide chain is unfolded, and proceeds through protein folding to the Golgi apparatus, where the glycan components are processed to a mature state. This requires a highly organised distribution of processing machinery to achieve the fidelity and quality control needed to ensure biological function.23 Approximately 80 % of the proteins entering the secretory pathway are glycosylated in the ER and most of the proteins assembled in the ER feature N‐linked oligosaccharides. Most of the glycoproteins featuring mature N‐glycans are, as described by Aebi, "precisely heterogeneous" in their carbohydrate composition—a result of kinetically controlled processing.24 Nonetheless, to reach their final mature and bioactive form, in the early stage of biosynthesis all N‐linked glycoprotein are homogeneously glycosylated. This is a result of a precise lectin chaperone (reader) based proofreading mechanism in the ER, which discriminates between correctly folded and misfolded glycoproteins (Figure 4).25 Here the oligosaccharide plays a key role in presenting each glycoprotein for scrutiny by the sophisticated biological checkpoint process, which is referred to as glycoprotein quality control.26 This process ensures that only correctly folded glycoproteins are transported to the Golgi for further glycan processing in to mature glycoproteins. Unfolded and misfolded glycoproteins are retained in the ER for further folding attempts and are eventually degraded if the correctly folded status is not achieved.</p><p>Carbohydrate writers, readers and erasers oversee the quality control of glycoprotein folding in the ER by modification of then‐linked glycan high mannose oligosaccharide core structure.</p><p>The glycoprotein quality control system presents clear parallels to the read, write and erase processes of epigenetic regulation. In the first step of glycosylation, Glc3Man9GlcNAc2 is transferred en bloc from an oligosaccharyl dolichol diphosphate to the nitrogen of an asparagine side chain in the nascent polypeptide chain by the writer oligosaccharyltransferase (OST).27 Immediately after the Glc3Man9GlcNAc2 is transferred, the eraser glucosidase‐I28 cleaves off the terminal glucose (Glc) residue, which is necessary to prevent the glycoprotein product rebinding to the OST. Subsequently, the eraser glucosidase‐II29 catalyses cleavage of a second glucose residue and the resulting monoglucosylated polypeptide is promptly sequestered by the calnexin (CNX)30 and calreticulin (CRT)31 lectin chaperone32 readers. These chaperones prevent aggregation of the unfolded glycopolypeptide chains, and assist in their correct folding by presentation to the oxidoreductase ERp57, which is responsible for effecting correct disulfide bond formation.33 Once the folded glycoprotein is released from the lectin chaperones readers, the eraser glucosidase‐II removes the final glucose residue and the glycoprotein undergoes inspection by the UDP‐glucose:glycoprotein glycosyltransferase (UGGT)29 (Figure 4).</p><p>If the correct glycoprotein folding is not accomplished, UGGT serves as a writer and re‐glucosylates the misfolded glycoprotein in preparation for recycling to the chaperones/ERp57 machinery34—the so‐called calnexin/calreticulin cycle (Figure 4).35 Following repeated failed folding attempts, the glycoprotein is degraded by the endoplasmic reticulum associated degradation system (ERAD).36 If correct folding is achieved, the glycoprotein is transported into the Golgi apparatus for further processing of the glycan to provide the mature glycoprotein.</p><!><p>It is widely recognised that carbohydrates play important roles in biological molecular recognition, and have a profound impact on human health and medicine. Nonetheless, there is merit in simplifying the language of glycoscience to make it more accessible to the uninitiated. In turn, this might facilitate a focus on the principles and implications of glycosylation in biology, rather than risking drowning in the detail of structural complexity. The notion of accessible vocabulary in glycoscience is not new: it was already evident in Hood, Huang and Dreyer's 197737 description of differentiation antigens as cell‐surface "area codes"; and the potential of cell‐surface carbohydrates, lectins, enzymes and carbohydrate‐binding antibodies in Feizi's 198138 "cellular addresses", "postmen, policeman and traffic signs" "involved in the obedient interpretation of area codes". Similar thoughts were explored in Brandley and Schnaar's 198639 "potential carbohydrate "language" involved in intercellular interactions", while Hakomori's 200240 "glycosynapse"—microdomains of glycolipids—seeks to draw parallels to the "immune synapse" assembly that contributes to cell adhesion and signalling. As highlighted in the cross‐disciplinary article by Bertozzi and Kiessling in 2001,41 "chemical tools have proven indispensable for studies in glycobiology". Perhaps it is time to revisit the terminology of glycoscience, to make interdisciplinary communication more straightforward and to support marketing and engagement beyond the immediate field. Reference to lectin readers, glycosyltransferase writers, and glycosyl hydrolase erasers could therefore be worth wider (re)consideration.</p><!><p>The authors declare no conflict of interest.</p><!><p>As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.</p><p>Supplementary</p><p>Click here for additional data file.</p>

PubMed Open Access

Be ₃ Ru: Polar Multiatomic Bonding in the Closest Packing of Atoms

The new phase Be 3 Ru crystallizes with TiCu 3 -type structure (space group Pmmn (59), a = 3.7062(1) Å, b = 4.5353(1) Å, c = 4.4170(1) Å), a coloring variant of the hexagonal closest packing (hcp) of spheres. The electronic structure revealed that Be 3 Ru has a pseudo-gap close to the Fermi level. A strong charge transfer from Be to Ru was observed from the analysis of electron density within the Quantum Theory of Atoms in Molecules (QTAIM) framework and polar three-and four-atomic BeÀ Ru bonds were observed from the ELIÀ D (electron localizability indicator) analysis. This situation is very similar to the recently investigated Be 5 Pt and Be 21 Pt 5 compounds. The unusual crystal chemical feature of Be 3 Ru is that different charged species belong to the same closest packing, contrary to typical inorganic compounds, where the cationic components are located in the voids of the closest packing formed by anions. Be 3 Ru is a diamagnet displaying metallic electrical resistivity.

be_₃_ru:_polar_multiatomic_bonding_in_the_closest_packing_of_atoms

Introduction<!>Results and Discussion<!>ChemistryOpen<!>Conclusion<!>Experimental Section

<p>Beryllium finds different technological applications as elemental metal (e. g., X-ray windows, satellite mirrors or nuclear shields) or as alloying component in diverse light-weight materials. Despite this, the knowledge of beryllium-containing intermetallic compounds is scanty in comparison with other series of intermetallic compounds. Moreover, the crystal chemistry of beryllium intermetallic compounds is barely investigated, its binary and ternary phase diagrams are scarcely studied and only the industrially relevant ones have been explored. This is mainly due to the challenges associated with preparation, characterization and the notorious toxicity of beryllium. [1] In general, the structural chemistry of beryllium intermetallic compounds is governed by the relatively small atomic radius, low valence electron count, and covalent character of beryllium. Moreover, the relatively small size of beryllium atoms leads to higher coordination numbers, especially for the beryllium-rich phases. [2] Beryllium is a good electrical conductor with the bulk superconductivity observed below T c = 0.026 K. Superconductivity was also reported for several beryllium intermetallic compounds, an example is the recently discovered Be 21 Pt 5one of the few complex metallic alloys displaying super-conductivity below T c = 2.06 K. [3] In contrast to typical intermetallic compounds, a study on Be 5 Pt revealed a semiconducting behavior at very low valence-electron count. [4] From a chemical point of view, in order to obtain an intermetallic compound with semiconducting characteristics, that is, showing a narrow gap at the Fermi level, some criteria regarding the combination of different bonding types between the components should be fulfilled (Figure 1). Noteworthy examples of semiconductors are the elements belonging to the p block (e. g. Ge and Si). A combination of elements from the group 13 with those from the group 15 or between elements from group 12 with those from group 16, can give rise to semiconducting materials -for example GaAs [5] or ZnSe, [6] which are characterized by twocenter-two-electron moderately polar bonds. Moving to the left side of the Periodic Table , where the number of electrons in the last shell (ELSA [7] ) is reduced, different bonding scenarios can be observed, depending on the difference in electronegativity between the components. When the difference in electro- ChemistryOpen www.chemistryopen.org negativity is very large, Zintl phases are the protagonists: the cationic part donates its valence electrons to the anionic one to form two center-and/or multiatomic bonds with moderate polarity in the anion sublattices. The anionic part of the Zintl phases is preferably formed by elements belonging to the p block, combined with elements of low electronegativity belonging to the 1-3 groups (e. g., SrSi 2, [8] BaSi 2 [9] ). Such materials show mostly gaps or pseudo-gaps in the electronic density of states (DOS) near the Fermi level. Compounds of late transition metals and alkaline or earth alkaline metals (e. g., Cs 2 Pt, [10] RbAu [11] ) or with elements belonging to the p block (e. g., FeGa 3, [12] RuGa 3 [13] ) can also show a gap at the Fermi level. While in the first family -aurides and platinates -the gap may be large here, it is essentially smaller in the second family. In the former, the difference in electronegativity between the components is high, suggesting the formation of ionic, polar two-and multi-atomic bonds; in the second, the difference in electronegativity is not so high, encouraging the formation of two-and multi-atomic bonds with moderate polarity. The last two groups represent relatively rare cases among materials with semiconducting properties. During the recent studies on the BeÀ Pt system, it was found that Be 5 Pt is a semiconductor, [4] belonging to the first of the last two groups mentioned above. Motivated by this interesting result, we decided to investigate the BeÀ Ru system, with the aim of finding new semiconducting materials. Differently from what we expected, the new phase reported here, Be 3 Ru, does not show semiconducting behavior, being a metal. It shows an unusual crystal chemical feature, representing a closest packing structure formed by differently charged species.</p><!><p>The existence of a phase with composition Be 10 Ru 3 was first suggested by Obrowski [14] who claimed that it is of γ-brass type, crystallizing with the cubic Cu 5 Zn 8 -type structure. Analysis of Xray diffraction data measured on the single crystals selected from the sample with the nominal composition Be 3 Ru (cf. Experimental Section), revealed instead an orthorhombic lattice, and the systematic extinctions were compatible with the space group Pmmn (Table 1).</p><p>The crystal structure contains three crystallographically unique atoms in the unit cell (Table 2), which includes two different Be and one Ru site. The structure was refined with anisotropic displacement parameters only for the heavy Ru atom. The refinement shows that all sites are fully occupied. Be 3 Ru crystallizes with TiCu 3 -type structure. [15] The final difference Fourier map did not reveal significant residual density. The refined atomic coordinates, together with the atomic displacement parameters, are listed in Table 2. Information about the interatomic distances and coordination numbers of the atoms can be found in Table 1S (Supporting Information). The crystallographic data are deposited in the ICSD database with deposition number 2169466.</p><p>The crystal structure of Be 3 Ru can be derived from a distorted hexagonal structure (2c/b = 1.947 instead of p 3 = 1.732) in which the atoms are organized in closest-packed layers perpendicular to [100], in an hcp sequence ABAB… [16] Typically for closest-packed structures, the coordination number for all the atoms is 12, and the coordination polyhedra around Be and Ru atoms are distorted hexagonal analogues of cuboctahedra (anticuboctahedra, Figures 2 and 1S). Ru atoms are found at a rather long distance of 3.12 Å from each other (longer than 2.65 Å, the RuÀ Ru distance in elemental Ru), [17] therefore there are no homoatomic RuÀ Ru contacts in the structure. The coordination environment of Ru is made only of Be atoms. The anticuboctahedron around Be1 atom contains 4 Ru, 4 Be1, and 4 Be2 atoms, while the one around Be2 atom contains 4 Ru and 8 Be1 atoms. The anticuboctahedra are condensed by sharing the triangular faces and form infinite columns along [100]. The columns are shifted with respect to each other along [100] in a way that the anticuboctahedra belonging to one column share the rectangular faces with the anticuboctahedra belonging to the second column (Figure 2). [18] The BeÀ Ru bond lengths range from 2.29 Å to 2.57 Å, which can</p><p>0.0251, 0.0270 Largest diff. peak and hole (e À Å À 3 ) À 0.5/0.6</p><p>[a] X-ray powder diffraction data. [b] [a] U eq = 4/3 [U 11 (a*) ChemistryOpen be interpreted using the covalent or atomic (metallic) radii, or combining both of them. Moreover, the Be1-Be2 distance of 2.28 Å is longer than the Be1À Be1 distance (2.24 Å) and is also longer than the distance of 2.08 Å in elemental Be. [17] The coordination environment of Be2 atoms contains only one type of Be atom (8 Be1) and 4 Ru atoms, while the coordination environment of Be1 atoms contains 4 Be1, 4 Be2, and 4 Ru atoms (Figure 1S). The crystal structure of Be 3 Ru belongs to the TiCu 3 -type structure. [15,19] With the 118 representatives, [20] it is one of the basic structural prototypes of intermetallic compounds. The members of the TiCu 3 family are mainly formed by elements with similar electronegativities, like MNi 3 (M = Mo, Nb, Ta) [21] or the electronegativity of the majority component is higher, like in REAu 3 (RE = GdÀ Lu). [22] The atomic arrangement in TiCu 3 (Be 3 Ru, space group Pmmn), together with that of Mg 3 Cd (P6 3 / mmc), represents a superstructure ('coloring' variant) of the hexagonal closest packing (hcp) of spheres. Both basic structure types differ in the ordering pattern within the closest packed layer -orthorhombic in TiCu 3 , and hexagonal in Mg 3 Cd. Such way of 'coloring' opens a possibility to form a variety of potential intermediate structures with the component ratio 1 : 3. [16b] All these circumstances would likely yield a characteristic metallic band structure with non-vanishing electronic density of states (DOS) at the Fermi level.</p><p>The band structure calculations in semi-and full-relativistic approximation indeed reveal a non-zero DOS at the Fermi level E F (Figure 3). Similar to other compounds of the transition metals with the s and p elements, the total electronic DOS is composed of three regions. The low-energy region (E < À 6 eV) is formed mainly by the s states of Be with some admixtures of the BeÀ p and RuÀ d and -s states. The d states of Ru make the majority contribution to the intermediate range (À 6 eV < E < À 1.3 eV). While the separation of the first and second regions is not very pronounced, the third part can be clearly recognized just below the Fermi level (À 1.3 eV < E < E F ), being mainly formed by a mixture of the RuÀ d and Be2-p states. The strong structuring of the electronic DOS, in particular in the two last regions, indicates special features of the chemical bonding. Furthermore, in contrast to typical metallic materials, the electronic DOS of Be 3 Ru shows a clear dip in the vicinity of the Fermi level. This resembles Zintl-and Wade-type phases (cf. SrGe 6, [23] Sr 3 Li 5 Ga 5, [24] Ca 2 Ga 4 Ge 6, [25] Dy(Cu 0.18 Ga 0.82 ) 3.7 [26] ) more than 'classical' intermetallic compounds. [27] On the other hand, the DOS of Be 3 Ru in the vicinity of the Fermi level is similar to that of the recently investigated Be 21 Pt 5 [3] and Be 5 Pt compounds, [4] which show a strong electron transfer from Be to Pt. The nonzero value of the DOS at the Fermi level E F is consistent with metallic-type behavior of the electrical resistivity (Figure 4S, Supporting Information).</p><p>These DOS features, in combination with a low electron number in the last shell per atom (ELSA [7,28] ) raise questions about their origin from the point of view of chemical bonding interaction.</p><p>The analysis of chemical bonding was performed by quantum chemical techniques in position space, which has recently been shown to be a powerful bonding investigation tool, in particular for intermetallic compounds with low ELSA and multi-atomic bonding. [3][4]7] The effective charges of all atoms were determined from the calculated electron density.</p><!><p>First, the zero-flux surfaces in the gradient vector field of the electron density were determined. They form the boundaries of electron density basins, which represent atomic regions according to the Quantum Theory of Atoms in Molecules (QTAIM). [29] The shape of the QTAIM Ru species in Be 3 Ru reveals some characteristic features (Figure 4, top). It is far from a spherical one and has concave surfaces toward neighboring beryllium species, similar to the platinum and beryllium atoms in Be 21 Pt 5 [3] or iridium and magnesium atoms in Mg 3 Ga 1-x Ir 3 + x , [7] where this feature is correlated with the strong charge transfer. The beryllium species look like 'soft tetrahedrons' resembling atomic shapes of Be and Mg in the two above-mentioned compounds. Taking into account the twelve-coordination of Ru by Be, one may expect that the Ru QTAIM shapes do not contact each other. Nevertheless, plane common surfaces between neighboring Ru atoms are observed in Be 3 Ru (Figure 4, bottom), indicative also of the RuÀ Ru interactions in some form (cf. orange bonding basin in Figure 5).</p><p>In the next step, the electron density was integrated in spatial regions, defined above as atomic shapes, yielding their electronic population. Subtraction of the electron numbers in the neutral atoms (atomic number) from the latter results in the QTAIM effective atomic charges. Ruthenium atoms reveal a large negative charge of À 3.58. The beryllium species are clearly playing the role of cations in Be 3 Ru. In particular, Be2found at the shorter distance to Ru (2.29 Å) -shows a larger charge of + 1.32, while Be1, found at a greater distance from Ru</p><p>(2.38 Å), shows a smaller charge of + 1.13. This variation indicates the different roles of Be1 and Be2 in the structural organization of Be 3 Ru. In addition, the diamagnetic properties (Figure 5S, Supporting Information) of Be 3 Ru might be also be related to such a charge transfer.</p><p>Further information about the bonding between atoms was obtained by applying the electron localizability approach, [30] based on the combined analysis of electronlocalizability indicator and electron density. The electron localizability indicator (ELIÀ D) shows spherical distribution around the nuclei of the non-interacting (isolated) atoms. Due to the bonding interaction, the spherical distribution is violated and attractors may appear in the regions of valence or penultimate shells, signaling bonding and indicating its geometrical organization. [30] Each of the so-formed attractors has its own ELIÀ D basin, which is determined, like the QTAIM atomic basins, by the zero-flux surfaces in the gradient vector field of ELIÀ D. The number of common surfaces of a bonding basin with the attached core (penultimate shell) basins defines the synapticity of the bonding basin and characterizes the number of atomic species participating in this bond (bond atomicity). In Be 3 Ru, only four types of bonding basins are observed, visualizing different components of bonding. The basins of the first group involve one Ru and three Be atoms and characterize the respective four-atomic bonds (red and green in Figure 5, top). The basins of the second group characterize the three-atomic bonding of one Ru and two Be species (blue, Figure 5). The basins of the third type (Figure 5, orange) involve two Ru and two beryllium atoms and highlight the RuÀ Ru interaction, which is already indicated by an analysis of the electron density (see above), but within a four-atomic bond. The population of all bonding basins is formed mainly by the Ru atoms with minor contributions from the beryllium species (Figure 5, middle), indicating the pronounced polar character of bonding in Be 3 Ru. This can be confirmed on the level of atomic volumes (Figure 5, bottom panel), where most parts of the volumes of bonding basins is located within the QTAIM atomic volume of ruthenium. For the quantitative characterization of the bond polarity for multi-atomic interactions, the concept of bonding polarity in position space [31] may be expanded from the two-atomic to multi-atomic bonds if only two sorts of atoms are participating in this bonding. Less dependent on the number of participating atoms, the polarity [32] of the BeÀ Ru interactions in Be 3 Ru varies from 0.44 (red and blue basins in Figure 5) to 0.56 (orange basin) and 0.58 (green basin) on a scale between 0 (non-polar, covalent bond) and 1 (close-shell configuration, ionic bond). This is consistent with the pronounced charge transfer (cf. QTAIM charges discussed above, Figure 4).</p><!><p>Be 3 Ru is one of the rare representatives of the TiCu 3 -type structure, formed by the valence electron-poor element as a majority component. From the geometric point of view, the crystal structure of Be 3 Ru can be derived by 'coloring' the hexagonal closest packing of spheres characteristic for large groups of intermetallic compounds. This is consistent with the low number of electrons in the last shell per atom (ELSA), typical for 'metallic' structures. Be 3 Ru is a diamagnet and its metallic electrical resistivity is confirmed by electronic structure calculations as well as experimental measurements. The pseudo-gap in the DOS is however, unusual. The calculated QTAIM charges of À 3.58 for Ru and + 1.13 or + 1.32 for Be reveal strong charge transfer. Both observations can be explained by the formation of polar three-and fouratomic BeÀ Ru bonds. The closest packing in Be 3 Ru is jointly formed by cationic and anionic components. This is in contrast to traditional inorganic compounds where one of the components forms the closest-packing motif, while the other one is located in the voids of the previous one.</p><!><p>Complete sample preparation was performed in a laboratory specialized for work with Be, inside an argon-atmosphere glovebox system (MBraun, p(H 2 O/O 2 ) < 0.1 ppm). [33] Polycrystalline samples with nominal composition Be 3 Ru were synthetized by arc melting from elemental Be (sheets, Heraeus, 99.2 wt %) and pre-melted Ru powder (Alfa Aesar, 99.95 wt %). To ensure the homogeneity, all samples were re-melted three times, with final mass losses of at most 5 wt %. To compensate for the beryllium loss during arc melting, a small excess ( � 5 wt %) of the latter was used. The samples were placed into BeO crucibles and sealed in Ta tubes. In order to obtain single-phase material, several experiments with varied annealing time have been carried out. After annealing at 1300 °C for 1 day, impurities of the Be-rich neighbor phase were observed. Only after annealing for 7 days at 1300 °C, a single-phase material was obtained (Figure 2S and Figure 3S). Be 3 Ru does not show an evident homogeneity range and does not exhibit air or moisture sensitivity. The thermal behavior of the prepared materials was studied in a differential scanning calorimeter (DSC) Netzsch DSC 404 C Pegasus, using a ZrO 2 crucible with lid, sealed in a Ta ampoule (purity of Ta 99.995 %). The phase was found to decompose at 1510 °C.</p><p>Powder X-ray diffraction measurements were performed on a Huber G670 Image plate Guinier camera using LaB 6 as internal standard (Cu Kα1 radiation, λ = 1.54056 Å; see Supporting Information). The lattice parameters were determined by a least-squares refinement using the peak positions, extracted by profile fitting.</p><p>Single crystals of Be 3 Ru were selected from the crushed annealed samples. They were glued to thin glass fibers and were analyzed at room temperature using a Rigaku AFC7 diffraction system equipped with a Saturn 724 + CCD detector (MoKα radiation, λ = 0.71073 Å). Absorption correction was performed by a multi-scan procedure. All crystallographic calculations were made with the program package WinCSD. [34] Details and results of the data collection are listed in Table 1. The electrical and magnetic properties of Be 3 Ru were determined experimentally from measurements on a few mm-sized irregular shaped solid chunk. Electrical resistivity measurements were performed by a Quantum Design Physical Property Measurement System (PPMS) 9 T, using a standard 4-terminal technique. Two Pt wires were used for making voltage contacts and two for current contacts. The wires were glued to the surface of the sample by using DuPont 4922 N silver conducting paste. Ac electrical resistivity 1 was measured at fixed temperatures between 2 and 300 K in 0 T and ChemistryOpen 9 T magnetic fields by applying 1 mA current pulse with frequency 23 Hz for 1 s. The measured data is shown in Figure 4S in Supporting Information. Magnetic measurements were conducted in a Quantum Design Magnetic Property Measurement System (MPMS) XL-5 superconducting quantum interference (SQUID) magnetometer equipped with a 7 T magnet. The sample was mounted on a capillary glass tube with varnish glue and its magnetic moment was measured vs. temperature in a stable magnetic field. The direct current (dc) magnetic susceptibility χ = M/H in the temperature range between 300 and 2 K for magnetic field μ 0 H = 1 T, is shown in Figure 5S of Supporting Information.</p><p>Electronic structure calculations on Be 3 Ru were performed by using the all-electron, full-potential local orbital (FPLO) method. [35] The experimental values of lattice parameters and the optimized values of atomic coordinates were used for calculations with a full relativistic model for the DOS and a semirelativistic model for the electron density and ELIÀ D calculation. All results were obtained within the local density approximation (LDA) to the density functional theory through the Perdew-Wang parametrization for the exchange-correlation effects. [36] A mesh of 12 × 12 × 12 k points was used for calculations.</p><p>The analysis of the chemical bonding was performed using the electron localizability approach in position space. [30c,37] For this purpose, the electron localizability indicator (ELI) in its ELIÀ D representation [30a,b] and the electron density (ED) were calculated with a specialized module within the FPLO code. [38] The topologies of the calculated three-dimensional distributions of ELIÀ D and ED were evaluated by means of the DGrid program. [39] The atomic charges from ED and bond populations for bonding basins from ELIÀ D were obtained via the integration of ED within the basins (space regions), bounded by zero-flux surfaces in the according gradient field. This procedure follows the Quantum Theory of Atoms in Molecules (QTAIM). [29] Deposition Number 2169466 (for Be 3 Ru) contains the supplementary crystallographic data for this paper. These data are provided free of charge by the joint Cambridge Crystallographic Data Centre and Fachinformationszentrum Karlsruhe Access Structures service.</p>

Chemistry Open

Signaling Functions of Reactive Oxygen Species\xe2\x80\xa0

We review signaling by reactive oxygen species, which is emerging as a major physiological process. However, among the reactive oxygen species, H2O2 best fulfills the requirements of being a second messenger. Its enzymatic production and degradation, along with the requirements for the oxidation of thiols by H2O2, provide the specificity for time and place that are required in signaling. Both thermodynamic and kinetic considerations suggest that among possible oxidation states of cysteine, formation of sulfenic acid derivatives or disulfides can be relevant as thiol redox switches in signaling. In this work, the general constraints that are required for protein thiol oxidation by H2O2 to be fast enough to be relevant for signaling are discussed in light of the mechanism of oxidation of the catalytic cysteine or selenocysteine in thiol peroxidases. While the nonenzymatic reaction between thiol and H2O2 is, in most cases, too slow to be relevant in signaling, the enzymatic catalysis of thiol oxidation by these peroxidases provides a potential mechanism for redox signaling.

signaling_functions_of_reactive_oxygen_species\xe2\x80\xa0

<!>SUPEROXIDE, HYDROXYL RADICAL, AND SINGLET OXYGEN ARE NOT SECOND MESSENGERS<!>HYDROGEN PEROXIDE FITS THE CRITERIA TO BE A SECOND MESSENGER<!>Cysteine Oxidation Products<!>Enzymatic Oxidation of Cysteine: The Lesson from Thiol Peroxidases<!><!>Enzymatic Oxidation of Cysteine: The Lesson from Thiol Peroxidases<!>SIGNALING BY CYSTEINE OXIDATION<!>The Case of PTP1B<!>FLOODGATE HYPOTHESIS<!>CONCLUSIONS<!>

<p>There are multiple components of signaling by ROS1 that are the subject of current investigations: the pathways in which ROS participate, the endogenous sources of ROS, the generation of secondary species that participate in signaling, and how ROS affect signaling and eventually gene expression. While these are all important, this article will focus on the chemical reactions through which ROS could act as signaling molecules. More emphasis will be given in this work to signaling by H2O2, which alone, among ROS, clearly fulfills all the chemical and biological requirements to be a "second messenger".</p><p>Reactions of ROS similar to those occurring in signaling can be relevant as mechanisms of pathology and toxicology and in the hormetic response in which nonlethal toxicity results in adaptive responses. Cyclooxygenase requires hydroperoxides derived from ROS for its activation that leads to the production of prostaglandins (1). Also, the initiation of metal-catalyzed lipid peroxidation requires ROS and then hydroperoxides and results in production of isoprostanes (2) and α,β-unsaturated aldehydes (3). The prostaglandins and isoprostanes cause signaling through G-protein-coupled receptors, while low concentrations of α,β-unsaturated aldehydes can activate several signaling pathways (4–7). However, as the production of these compounds is an indirect effect of ROS, they will not be considered further.</p><p>In the 1970s, a few studies noted that exogenously added H2O2 could mimic growth factor activity and that the growth factors could stimulate in cells the endogenous production of H2O2 (8–10). Nonetheless, it was not until studies in the late 1980s and early 1990s showed that low concentrations of exogenous H2O2 added to cells caused proliferation (11) and that exogenous H2O2 added to T-cells activated NF-κB (12) that a role of H2O2 in signaling became noticed by researchers outside of a small community of free radical experts. Although NF-κB activation in stimulated macrophages by endogenously generated H2O2 was then demonstrated (13), it was not until the existence NADPH oxidases of nonphagocytic cells was discovered (14–16) that physiological signaling by reactive oxygen species became a major topic of interest.</p><p>Concepts of how H2O2 participates in signaling have involved studies in which it has been proposed that the ability of enzymes to remove H2O2 must be overcome (the "floodgate" hypothesis that will be discussed later) while the signaling reactions by H2O2 occur nonenzymatically. Unfortunately for the field, these studies did not pay attention to the very important fact that H2O2 reacts slowly with protein thiols, the principal targets in H2O2-dependent signaling. Thus, a rigorous discussion about the biochemical logic of redox transitions caused by H2O2 is needed. The kinetics of enzymatic and nonenzymatic H2O2 reactions will, therefore, be the major focus of this work.</p><!><p>A crucial feature of a second messenger is having specificity in its interaction with effectors in signaling pathways. The specificity of signaling by ROS is defined by both kinetics and spatial relationships. Unfortunately, the kinetics of the reactions of ROS with potential targets, in competition with the enzymes that remove them, has often not been given due consideration.</p><p>Superoxide oxidizes thiols to thiyl radical, which can initiate a chain reaction, but the rate constants for this reaction are quite slow, probably no more than 103 M−1 s−1 at pH 7.4 (17). Indeed, the oxidation of thiols by superoxide is more likely a reaction with the protonated HO2⋅ (the pKa for O2⋅− is 4.7) than with O2⋅− itself. So at pH 7, the rate constant for oxidation of thiols by O2⋅− is insignificant in comparison to the rate constants (>109 M−1 s−1) for the cytosolic and mitochondrial superoxide dismutases (18) that are abundant in their compartments. When O2⋅− is generated on the outside of cells, such as by the NADPH oxidase isoform, NOX2, in endothelial cells, it appears to be able to enter cells, resulting in signaling (19); however, once inside the cell, it would be rapidly dismuted to H2O2 and O2. Therefore, although a role for O2⋅− in signaling has been proven, targets that react with O2⋅− have not been demonstrated to do so in vivo. The signaling role of O2⋅− is more likely as a precursor of H2O2.</p><p>Hydroxyl radical has also been thought to be involved in signaling (20); however, HO• has no specificity as it reacts with almost any organic molecule with rate constants near the limit of diffusion. Some of the products of lipid peroxidation, initiated by different free radicals, including HO•, such as the isoprostanes can signal through receptors (4), while others, such as 4-hydroxy-2-nonenal, can participate in signaling by modifying signaling proteins, such as Keap1 (21) or SHP-1, a protein tyrosine phosphatase (22). However, this is not evidence of a direct role of HO• as a second messenger.</p><p>Singlet oxygen is produced in cells by nonenzymatic reactions, such as photoactivation of some drugs and the reaction of hypohalous acids with H2O2. Studies with heme peroxidases suggest either a very low yield by prostaglandin G2 reductase (23) or a high yield in the chloroperoxidase-catalyzed oxidation of halides. The site and time specific generation of singlet oxygen in cells, which would be required for it to fit the definition of a second messenger, has not been demonstrated and seems very unlikely. Regardless, although singlet oxygen is a potent dienophile, no normal physiological process for biological oxidation by singlet oxygen has been discovered, which also precludes it as a second messenger.</p><!><p>What qualifies H2O2 as a suitable second messenger are its enzymatic production and degradation that provide specificity for time and place and its interesting chemistry that provides specificity for oxidation of thiols. The way H2O2 is produced in nature is remarkably similar to the modern industrial process for the production of tons of it. In the industrial process, molecular dioxygen is reduced during autoxidation of anthrahydroquinone, which is reduced back by hydrogen in the presence of a metal catalyst. In nature, the redox moiety of the flavin in aerobic dehydrogenases transfers two electrons to oxygen and is thereafter reduced back by hydrogens extracted from the reduced substrate of the enzymatic reaction. Superoxide anion is the one-electron transfer intermediate in the reaction, and in some cases, such as in xanthine oxidase or during redox cycling of quinones reduced by microsomal electron transport systems, it can be released as the final product. Another relevant source of H2O2 results from a leak in the mitochondrial electron transport chain. Again, H2O2 comes from dismutation of O2⋅−, which is primarily generated by the autoxidation of ubisemiquinone: (1)2O2•−+2H+→SODO2+H2O2</p><p>Although dismutation can occur nonenzymatically with a rate constant of 105 M−1 s−1, the rate constants for the superoxide dismutases are much faster, in the range of 109 M−1 s−1 (18). Mitochondrial O2⋅− production is regulated by the energy state (24–27), O2 concentration (26), uncoupling proteins (28), nitric oxide, and peroxynitrite (29, 30). The uncoupling proteins, which cause oxidation of the electron transport components, decrease the level of O2⋅− production (28), while nitric oxide appears to increase the level of O2⋅− production by inhibiting cytochrome oxidase, thereby resulting in a greater reduction of the electron transport components (29). Nonetheless, the rate of mitochondrial O2⋅− production is also determined by its dismutation to H2O2 (31, 32). The effect of superoxide dismutase in accelerating mitochondrial H2O2 production occurs because dismutation pulls forward the autoxidation of semiquinones which is thermodynamically unfavorable (33).</p><p>Although all the mechanisms for producing H2O2 for signaling described above cannot be definitively ruled out, the major source for which there is positive evidence of a signaling function is O2⋅− generated by NADPH oxidases: (2)NADPH+2O2→NADPHoxidaseNADP++H++2O2•− followed by dismutation (reaction 1).</p><p>The NADPH oxidases are regulated by either assembly of the components into a functioning enzyme complex (as for NOX1 and NOX2) or transcription of the oxidase (as for NOX4). The related DuOX (originally given this acronym because of proposed dual oxidase and peroxidase functions) enzymes appear to primarily catalyze two-electron transfer to O2 to produce H2O2 (34, 35), which, as with other flavoproteins that appear to produce H2O2 (36), adds electrons to O2 in two steps with the second step happening more quickly than the O2⋅− can escape from the active site.</p><p>H2O2 is a polar compound (dipole moment of 2.2 D) because of the dihedral angle of 120° and the H–O–O angle of 95°, in the most stable structure (37). A bond strength typical of covalent bonds, 90 kcal/mol, has been measured for the H–O bond, while the O–O bond is much weaker (51 kcal/mol) (38). Accordingly, this bond is easier to break in biochemical reactions. H2O2 is a strong oxidant (E° = 1.776 V for the reduction to water). The oxidation to O2 is also possible (E° = 0.682 V) and takes place in the catalytic site of catalase using the energy from the reduction of the other H2O2: (3)2H2O2→catalaseH2O+O2 In biological systems, the oxidizing potential of H2O2 is exploited by two mechanisms. By one-electron reduction (ΔG0 = −8.8 kcal/mol), H2O2 is reduced usually by a transition metal ion. The hydroxyl radical (HO•) produced in the reaction is an extremely strong oxidant that reacts with practically all the biological molecules with a nearly diffusion-limited rate constant. As pointed out above, this radical cannot play any specific role in signaling.</p><p>The other mechanism is the two-electron nucleophilic substitution reaction that has been suggested for the oxidation of thiols. This reaction requires the dissociation of the thiol to a thiolate anion (RS−) to achieve the required nucleophilicity (39): (4)H2O2+RS−→OH−+RSOH While this reaction occurs spontaneously with a rate constant of 18–26 M−1 s−1 (40), the activation energy is too high for the unassisted reaction to be fast enough to be biologically relevant. If a proton-donating group is, however, in sufficient proximity to allow the hydroxyl ion to leave as water, then the reaction (5)H++H2O2+RS−→H2O+RSOH would proceed at a much faster rate. In solution, reaction 5 would not occur, as the thiolate would be protonated. Special features of active sites surrounding the peroxidatic Cys (Cp) of proteins such as Prxs and invertebrates or plant glutathione peroxidases that rely on a cysteine for catalytic activity, Cys-GPxs, allow reaction 5 apparently by favoring both the dissociation of thiol and the protonation of the leaving group. Understanding how Prxs and Cys-GPxs catalyze reaction 5 provides a general mechanism for the H2O2-dependent oxidation of protein thiols in a biologically relevant context. It should be noted that in most cases, any physiological hydroperoxide or peroxynitrite could substitute for H2O2. The exceptions would be when enzyme specificity discriminates among these compounds. Indeed, there are cases in which the alternative reactive compounds are more likely involved in signaling than H2O2 (41, 42).</p><!><p>Cysteine can be oxidized to form several different products. These include the thiyl radical (–S•) by a one-electron transition, sulfenic acid (–SOH) and disulfide (–S–S–) by a two-electron transition, sulfinic acid (–SO2H) by a four-electron transition, and eventually sulfonic acid (–SO3H) by a six-electron transition.</p><p>In water, at neutral pH the electrochemical oxidation of a cysteine thiolate anion against a saturated calomel electrode takes place at an Eox of approximately 0.1 V (43). To the best of our knowledge, there are not data available for the oxidation of a cysteinesulfenic acid derivative, but it seems reasonable assuming that the tautomerism facilitates a further oxidation to sulfinic acid, which, instead, must be more resistant to oxidation because of the presence of two electronegative atoms. A disulfide is also relatively resistant to oxidation (Eox ~ 0.8 V in water at pH 7 or 1.2 V in an aprotic solvent) (44, 45). By inference, we can reasonably conclude that, when permitted, the formation of a disulfide that competes with the formation of a sulfinic acid residue prevents further oxidation of a cysteine thiol.</p><p>The sulfinic acid derivative has indeed been shown to be formed, at least in part, at the active site of Prx-2 (46, 47), but requiring a second reaction with H2O2, the reaction seems more likely to occur under severe oxidative stress than during physiological signaling. The same holds for the sulfonic derivatives that practically do not exist under biological conditions. It follows that the oxidation states of a Cys residue relevant for signaling are most likely to be the sulfenic acid derivative and the disulfide only.</p><!><p>Enzymatic oxidation of substrates containing thiol groups provides a 2-fold example of the mechanism for sulfur oxidation. A Cys or a Sec residue is oxidized at the active site of a peroxidase (first mechanism), which in turn shifts the oxidation potential to a thiol-containing substrate (second mechanism), thus regenerating the ground-state enzyme. GSH and Trx or other Trx-like proteins are the low-molecular weight substrates oxidized by the specific thiol peroxidases that use H2O2 as the oxidizing substrate. Trx is oxidized by Prxs 1–5 in vertebrates and by Cys-GPxs in nonparasitic invertebrates or plants. (6)H2O2+Trx(SH)2→Cys−GPxsPrxs1−52H2O+TrxS2</p><p>GSH is oxidized by mammalian glutathione peroxidases, Sec-GPx, which typically contain a peroxidatic Sec residue or 1-Cys Prx-6 (48). Prx-6 appears to interact with GSTP; however, while GSTP appears to be required for Prx-6 to be active, the transferase more likely preserves activity of the peroxidase rather than participates in the peroxidase catalytic activity (A. Fisher, personal communication). (7)H2O2+2GSH→Sec−GPxsPrx−62H2O+GSSG In these peroxidases, the Cys or Sec at the active center is oxidized at a very fast rate by the hydroperoxide to a primary product, a SOH or Se-OH, due to an assist from amino acids in the distinct active sites, which, remarkably, are strongly conserved among the members of the two families of enzymes. Both Prx and GPx active sites promote dissociation of the thiol (or selenol), polarize the O–O bond of the hydroperoxide, and protonate the leaving OH− group (49, 50).</p><p>Sulfenic acid as the reactive primary product of the oxidation by the peroxide can also be formed at the active site of some other non-peroxidase proteins. Using a specific antibody against the protein adduct formed by reaction with dimedone, a reagent that reacts rather specifically with sulfenic acid, this epitope was found in normal cells, with even greater amounts in cancer cells (51). The actual physiological or pathological relevance of this oxidation needs to be clarified; however, although nonspecific oxidations are possible, the proteins undergoing a physiological redox switch must fulfill the mandatory criteria for lowering the activation energy of the nucleophilic displacement reaction described above. In Prxs or Cys-GPxs, this primary oxidation product evolves into other species, depending on the specific environment. When there is a suitable Cys in the same subunit, i.e., the resolving Cys (Cr) as in atypical 2-cysteine Prxs (Prx-5 of mammals) and in the majority of Cys-GPxs, an intrachain disulfide is formed as the oxidized intermediate of the peroxidase. This is usually associated with some unfolding of the protein (52). In the case of typical 2-cysteine Prxs, such as Prxs 1–4 of mammals, a Cys residue of a second subunit is the thiol reacting with the –SOH group thereby producing an intermolecular disulfide as the oxidized intermediate. The fast reduction of both these types of disulfides by Trx or Trx-like proteins accounts for the reductive step of the peroxidatic cycle in atypical/typical 2-cysteine Prxs and plant/invertebrate Cys-GPxs. Notably, fast kinetic analysis of the peroxidatic reaction of thioredoxin peroxidase of Mycobacterium tuberculosis showed that the Cp is directly reduced by Trx, although a Cr also exists (53). The formation of the disulfide between Cp and Cr, observed in the absence of Trx, was interpreted not as a necessary intermediate of the peroxidatic cycle but as a kind of "parking of the oxidized form". The reaction is seen as "protective" for the enzyme, since the formation of an irreversibly overoxidized form in the absence of the reducing substrate was prevented. Lacking fast kinetic data for other peroxidases, we do not know how often this interesting mechanism occurs in nature.</p><p>When there is not a suitable Cr, such as in Sec-GPx or mammalian Prx-6, the reaction of the primary oxidation product occurs with GSH, yielding a mixed disulfide with the reducing substrate GSH as the oxidized intermediate of the enzymatic reaction. The overall specificity for the enzymatic oxidation of the target emerges from the specificity of the interaction of the enzyme first with the hydroperoxide and then with the thiol-containing substrate. Accordingly, the kinetics of the enzymatic reaction follows a ping-pong mechanism in which ternary complexes are not formed (54, 55).</p><p>The steady-state kinetic analysis of different Prx subtypes indicates that the rate constants for the oxidation of the Cys residue at the active site range from 2 × 103 to 4 × 107 M−1 s−1, depending on the hydroperoxide and the individual Prx (55–58). These values are comparable with those obtained for the Cys-GPxs (from 4 × 103 to 1.3 × 106 M−1 s−1), while values lower than 3.4 × 105 M−1 s−1 have never been obtained for the Sec-GPxs (54). From these measurements, we can conclude that the structure of the active site fits the requirements for a physiological oxidation of a thiol to sulfenic acid, where the nucleophilic displacement reaction is accelerated by the donation of a proton to the leaving HO− group.</p><!><p>(a) direct oxidation of a thiol to sulfenic acid, when the suitable Cys is assisted in the active site as in reaction 5 (8)H++H2O2+protein−S−→H2O+protein−SOH</p><p>(b) formation of a disulfide (9)protein−SOH+RSH→protein−SSR+H2O</p><p>(c) thiol–disulfide exchange (10)protein−SSR+R′SH⇄protein−SSR′+RSH or (11)protein−SSR+R′SH⇄protein−SH+RSSR′ where R(R')SH is any accessible thiol.</p><!><p>Only reaction 8 requires an oxidant, H2O2, while in reactions 9–11, scrambling takes place among cysteine residues but the overall redox status of the thiols is maintained. It should be noted, however, that in reaction 9, either the protein–SOH or RSH species would need to be in the anionic form for the reaction to proceed, while in reactions 10 and 11 the R'SH species would need to lose its proton as well.</p><!><p>The cysteinesulfenic acid derivative could be stable when in an apolar environment not accessible to solvent, and the approach by reactive molecules is sterically inhibited (59). Oxidation of protein cysteine to sulfenic acid can be detected using dimedone, a compound that reacts quite specifically with sulfenic acid. Several proteins containing the reactive group have been detected using a specific antibody (51) or by using dimedone linked to a fluorescent probe (60). More specifically, besides the active site of some oxidoreductases (61), the formation of a cysteinesulfenic acid has been reported or suggested for proteins involved in cell signaling. OxyR is a nuclear factor that in bacteria stimulates, through interaction with RNA polymerase, transcription of genes enhancing resistance to oxidants (62). The activation of the nuclear factor takes place through the formation of a cysteinesulfenic acid in the presence of submicromolar concentrations of H2O2 with a rate constant for reaction in the range of 105 M−1 s−1 (63). This cysteinesulfenic acid then reacts relatively slowly with a cysteine that is not in proximity (17 Å), resulting in the formation of a disulfide (64). This implies that, although the overall mechanism is still a matter of debate (65), a major conformational shift has to take place following oxidation, driving the modification of the interaction with DNA. The reversal of the reaction is accomplished by glutaredoxin by a thiol–disulfide exchange reaction (59).</p><p>Less conclusive evidence has been obtained for the actual role of a cysteinesulfenic acid in the redox regulation of Fos and Jun, an AP-1 transcription factor complex (66), while definitive evidence supports the role of Trx and HAP/Ref1 in reduction of cysteine in AP-1 that is required for its transcriptional activity and binding to DNA (67–69).</p><p>In Saccharomyces cerevisiae, exposure to an oxidizing environment activates adaptive gene expression through the transcription factor, Yap-1, that is activated by oxidation. This does not take place by direct interaction with H2O2 but requires a Cys-GPx subtype, yeast GPx-3. The oxidation of the Cp, Cys-36, of GPx-3 results in the formation of a mixed disulfide with Cys-598 of Yap-1. Cys-303 of Yap-1 than displaces it, and an intramolecular disulfide between Cys-598 and Cys-303 of the transcription factor is formed, eventually accounting for the shift to its active form. Trx turns off the signaling pathway, by reducing both the sensor and the regulator (70). The emerging function of yeast GPx-3 as a sensor of oxidants and redox translator underscores the peculiar role of the active site of GPxs in reacting specifically and at a high rate with H2O2. The outcome of the reaction also suggests the role of a cysteinesulfenic acid intermediate, although its involvement has been only indirectly assumed.</p><p>There is an increasing number of proteins that appear to form a protein–glutathione mixed disulfide under physiological conditions, including PTP1B (71), cytosolic NADP+-dependent isocitrate dehydrogenase (72), the yeast 20S proteasome (73), ryanodine receptor type 1 (74), signal transducer and activator of transcription 3 (also called STAT 3), (75), caspase-3 (76), and the phosphatase and tensin homologue deleted from chromosome 10 (also called PTEN) (77). When considering glutathionylation in a biological environment, the most reasonable mechanism must fulfill the constraints described above for the formation of a sulfenic acid, in turn reacting with GSH. (12)protein−SO−+GSH→protein−SSG+OH− A disulfide exchange reaction between a cysteine residue in a protein and GSSG is also possible and is catalyzed by the Trx-like protein glutaredoxin (Grx): (13)RSH+GSSG⇄GrxRSSG+GSH The oxidizing substrate GSSG is formed by oxidation of GSH in reaction 7. It is well established that formation of protein–glutathione mixed disulfides occurs under severe oxidative stress and in the endoplasmic reticulum (78, 79). In the cytosol, under physiological signaling conditions, however, the low concentration of GSSG is unlikely to push this reaction toward formation of the mixed disulfide. Moreover, it should be pointed out that for thiol–disulfide exchange the actual barrier that limits the oxidation of proteins by GSSG is thermodynamic. Indeed, the mixed disulfide formed must have a redox potential higher than that of the [GSH]2/[GSSG] couple, which although theoretically possible in some extreme cases, is extremely unlikely. In addition, we also must consider that the kinetics of this reaction is expected to be slow unless it is catalyzed.</p><p>Although Trx and GSH can react with H2O2 nonenzymatically, the Prx-catalyzed rate of reaction to form TrxS2 is several orders of magnitude faster (reaction 6) (56). Oxidized Trx is known to be involved in at least one signaling pathway, the activation of ASK1, an upstream protein kinase kinase kinase in the p38MAPK and Jun N-terminal kinase (JNK) pathways (80). ASK1 is normally inhibited by reduced Trx; however, when the Trx is oxidized, it dissociates from ASK1, allowing it to dimerize and self-activate and then activate downstream protein kinase kinases (80). This has been shown to occur with endogenous generation of H2O2 produced in macrophages through stimulation of NOX2 (81).</p><p>Another example is in TRP14, a Trx-related protein, in which its active site cysteine appears to form a specific disulfide link with the LC8 cytoplasmic dynein light chain (82).</p><!><p>Possibly the most studied signaling protein proposed to form a sulfenic acid intermediate is PTP1B. The rate constant for the oxidation of the PTP1B active site cysteine by H2O2 is, however, only in the range of 9–43 M−1 s−1 (71, 83, 84). This is approximately the same rate constant as for the thiolate moieties of low-molecular weight compounds, which are in the range of 18–26 M−1 s−1 (40). Thus, while it has been demonstrated that PTP1B can react in vitro with H2O2 to form a sulfenic acid that goes on to form a sulfenyl–amide intermediate (85, 86), this is rather unlikely to occur in a cell where it would be in competition with Prxs and Sec-GPxs, which are abundant in the cytosol and react with H2O2 1 million times more rapidly. The intracellular concentration of GSH in its thiolate form would be 12.4 μM, calculated from an intracellular pH of 7, the GSH thiol pKa of 8.9 (87), and 1 mM GSH, which is near the low end of the intracellular GSH concentration range (88). With a similar rate constant for reaction with H2O2 as the PTP1B active site cysteine, 12 μM GS− would outcompete PTP1B [estimated to be 8.3 nM in HepG2 cells (89)] and prevent its oxidation even in the absence of Prx and Sec-GPx activity. In theory, a very high local concentration of H2O2 could allow a reaction with a slow rate constant to occur in competition with the hydroperoxide reducing enzymes; however, in the one study where the endogenous source of the H2O2 was identified, it was produced on the other side of the plasma membrane from PTP1B (90). Nonetheless, reversible oxidation of the PTP1B active site cysteine does occur (83), forming a mixed disulfide with GSH during cell signaling (90). Assuming that the thrice measured rate constant for the oxidation is correct, mechanisms other than direct oxidation of the enzyme followed by reaction with GSH must be involved. The use of the active site thiolate in place of the resolving cysteine or GSH by a peroxidase would be consistent with the principles for thiol oxidation described above.</p><!><p>In this hypothesis, inactivation of Prx occurs by overoxidation of the active site cysteine to sulfinic acid and this permits H2O2 to then react with its targets. Prx, which reacts very rapidly with H2O2, can react a second time to produce the sulfinic acid: (14)Prx−SOH+H2O2→Prx−SO2H+H2O This reaction, which can occur during oxidative stress, has also been proposed to occur during physiological signaling and to act as a floodgate that regulates redox signaling (46). An energy-dependent restoration of the thiolate in the Prx active site is catalyzed by the enzyme, sulfiredoxin (91, 92), which allows Prx to once again inhibit H2O2-dependent signaling. It has been demonstrated that exogenous H2O2 can cause the overoxidation of Prx-2 and cause cell cycle arrest (93); however, this has yet to be shown under physiological conditions. Mutation of Prx-2 has been shown to enhance signaling in response to platelet-derived growth factor, which is consistent with the floodgate hypothesis (94); however, Prxs interact directly with several signaling proteins, consistent with a role different from the removal of H2O2. A yeast form of Prx-1, Tpx1, forms a peroxide-induced disulfide complex between itself and Sty1, a yeast JNK (95). Mammalian Prx-1 (also called PAG) binds to the GSTP–JNK complex, thereby preventing the release of JNK from the complex and inhibiting its activation (96). Prx-1 also specifically binds and inhibits c-Abl, a nonreceptor tyrosine kinase (97) and binds to and alters the activity of c-Myc (98).</p><p>There are three major problems with the floodgate hypothesis. First, the role of Sec-GPxs has been overlooked. Cells, if not supplemented, are severely depleted of selenium and express the cytosolic GPx at a very low level (99). Thus, in many cell studies that showed Prxs to be the major enzyme eliminating H2O2, GPx deficiency actually existed. Nonetheless, in vivo, where selenium depletion is rare, inactivation of Prxs by overoxidation would still leave a significant GPx activity that would prevent the flood. The second major problem is the requirement of the hypothesis for a second hit on the same Prx molecule to eliminate a significant amount of activity before H2O2-dependent signaling occurred. The rate of oxidation of Prx-1 was thought to result in the inactivation of 0.072% of Prx-1 during each round of catalysis (94). Whether this relatively slow rate of inactivation is sufficient to allow the floodgate hypothesis to account for H2O2-dependent signaling is, therefore, questionable. Data suggesting a more rapid rate of reversal of the oxidation of Prx isoforms to sulfinic acid (100) and questioning of the extent of Prx overoxidation (101) cast further doubt on the hypothesis. The third major problem is that, even in the absence of both Prx and GPx activities, the cytosolic 12.4 μM GS− (see the section on PTP1B above) would outcompete the nanomolar concentration of any target protein for nonenzymatic reaction with H2O2. Finally, even if the floodgate hypothesis were to be shown to be correct, it does not address the mechanism by which H2O2 actually signals.</p><!><p>From a review of the chemistry and biochemistry of ROS, it has become clear that H2O2 is the only one that clearly has the characteristics of a second messenger. Superoxide is a major precursor of H2O2 rather than a direct participant in signaling, while hydroxyl radical lacks specificity.</p><p>The specificity of H2O2 as a second messenger comes from its reactions with specific, "oxidation prone" protein Cys residues in local environments that lower the pKa. Nonetheless, having a Cys in the thiolate form is not sufficient for the definition of an oxidation prone Cys. The reaction must be fast enough to compete with peroxide-removing enzymes and can occur only with assistance in breaking the O–O bond of H2O2, resulting in the formation of a cysteinesulfenic acid and water.</p><p>The protein cysteinesulfenic acid formed either is stable or reacts with (a) another Cys in the same or another protein, giving rise to a conformational shift; (b) GSH, producing a glutathionylated protein; (c) an amide in the backbone of the protein; or (d) another hydroperoxide. Options a–c are readily reversed, while option d requires more complex reactions to restore the original cysteine thiolate. Glutathionylated proteins in cytosol are produced by either reaction b or exchange of GSH with a protein disulfide catalyzed by glutaredoxin.</p><p>The formation of glutathionylated signaling proteins, such as PTP1B, or intramolecular disulfides, such as in the thioredoxin that modulates ASK1, is the posttranslational modification that is essential to redox signaling. The kinetics of thiol oxidation suggests that only enzymatically catalyzed oxidations are likely to be part of signaling and that it is the enzymatic use of H2O2 rather than its overcoming of antioxidant defense as in the floodgate hypothesis that defines signaling by ROS.</p><!><p>Supported by National Institutes of Health Grant ES05511 to H.J.F. and Progetti di ricerca di Ateneo, Università di Padova, Grant CPDA087343/08 to M.M.</p><p>Abbreviations: ASK1, apoptosis signaling kinase 1; Cys-GPxs, cysteine-dependent glutathione peroxidases; HO•, hydroxyl radical; Cp, peroxidatic cysteine; Cr, resolving cysteine; GPxs, glutathione peroxidases; GSTP, glutathione S-transferase π; NOX, NADPH oxidase; Prxs, peroxiredoxins; PTP1B, protein tyrosine phosphatase 1B; ROS, reactive oxygen species; Sec-GPxs, selenocysteine-dependent glutathione peroxidases; Trx, thioredoxin.</p>

PubMed Author Manuscript

Genetic Variants in the Fibroblast Growth Factor Pathway as Potential Markers of Ovarian Cancer Risk, Therapeutic Response, and Clinical Outcome

Background The fibroblast growth factor (FGF) and FGF receptor (FGFR) axis plays a critical role in tumor-igenesis, but little is known of its influence in ovarian cancer. We sought to determine the association of genetic variants in the FGF pathway with risk, therapeutic response, and survival of patients with ovarian cancer. Methods We matched 339 non-Hispanic white ovarian cancer cases with 349 healthy controls and geno-typed them for 183 single-nucleotide polymorphisms (SNPs) from 24 FGF (fibroblast growth factor) and FGFR (fibroblast growth factor receptor) genes. Genetic associations for the main effect, gene\xe2\x80\x93 gene interactions, and the cumulative effect were determined. Results Multiple SNPs in the FGF\xe2\x80\x93FGFR axis were associated with an increased risk of ovarian cancer. In particular, FGF1 [fibroblast growth factor 1 (acidic)] SNP rs7727832 showed the most significant association with ovarian cancer (odds ratio, 2.27; 95% CI, 1.31\xe2\x80\x933.95). Ten SNPs were associated with a reduced risk of ovarian cancer. FGF18 (fibroblast growth factor 18) SNP rs3806929, FGF7 (fibroblast growth factor 7) SNP rs9920722, FGF23 (fibroblast growth factor 23) SNP rs12812339, and FGF5 (fibroblast growth factor 5) SNP rs3733336 were significantly associated with a favorable treatment response, with a reduction of risk of nonresponse of 40% to 60%. Eleven SNPs were significantly associated with overall survival. Of these SNPs, FGF23 rs7961824 was the most significantly associated with improved prognosis (hazard ratio, 0.55; 95% CI, 0.39 \xe2\x80\x93 0.78) and was associated with significantly longer survival durations, compared with individuals with the common genotype at this locus (58.1 months vs. 38.0 months, P = 0.005). Survival tree analysis revealed FGF2 rs167428 as the primary factor contributing to overall survival. Conclusions Significant associations of genetic variants in the FGF pathway were associated with ovarian cancer risk, therapeutic response, and survival. The discovery of multiple SNPs in the FGF\xe2\x80\x93FGFR pathway provides a molecular approach for risk assessment, monitoring therapeutic response, and prognosis.

genetic_variants_in_the_fibroblast_growth_factor_pathway_as_potential_markers_of_ovarian_cancer_risk

<!>Study Design<!>Data Collection<!>SNP Selection and Genotyping<!>Statistical Analysis<!>eQTL Analysis<!>Population Characteristics<!>Association of Individual SNPs with Ovarian Cancer Risk<!>Association of FGF and FGFR SNPs with Chemotherapeutic Response<!>Overall Survival for FGF and FGFR Variants<!>SNP\xe2\x80\x93Gene Association in cis eQTL<!>Discussion

<p>Ovarian cancer is the leading cause of death from gynecologic cancers and the fifth most lethal malignancy in women in the US. An estimated 22 240 new cases and 14 030 deaths from ovarian cancer will occur in the US in 2013 (1). The overall dismal 46% 5-year survival rate for ovarian cancer has remained unchanged for several decades (1). The main reason for this poor outcome is the lack of success in diagnosing ovarian cancer at an early stage, owing to an absence of obvious symptoms, clinical indications, and effective screening tests. A majority of women are diagnosed with a high-grade invasive cancer that is difficult to treat. In contrast, women have a 90% to 95% probability of survival when their ovarian cancer is detected at an early stage (2). The results obtained with current screening strategies to reduce mortality in women with ovarian cancer— which use the serum biomarkers cancer antigen 125 (CA125)5 and human epididymis protein 4 (HE4) along with the Risk of Ovarian Malignancy Algorithm and transvaginal ultrasonography— have not been encouraging (3). Currently, these serum biomarkers (CA125 and HE4) are being used mainly to monitor chemotherapeutic response and to detect recurrence after therapy, but none of the current biochemical markers are sufficient to guide the prediction, screening, and prognosis of ovarian cancer (4, 5). Therefore, the search for ovarian cancer biomarkers—in particular genetic markers—for risk assessment, monitoring of therapeutic response, and outcome prediction of ovarian carcinoma is of profound importance.</p><p>A number of common germline genetic alterations, including those identified via genome-wide association studies, have been associated with ovarian cancers (6–9). Candidate-gene and pathway-based approaches have also successfully identified ovarian cancer–susceptibility loci and loci associated with clinical outcomes (10). Our group previously demonstrated that nucleotide-excision repair polymorphisms are associated with recurrence and survival in ovarian cancer patients (11). Transforming growth factor β (TGF-β) enhances ovarian tumor metastasis, and we recently demonstrated that genetic variants in the TGF-β signaling pathway are associated with variation in the risk of developing ovarian cancer (12). In addition, we have identified several microRNA-related genetic polymorphisms that are associated with ovarian cancer risk and clinical outcomes (13); however, the full spectrum of the genetic loci contributing to ovarian cancer susceptibility and outcome remains to be revealed.</p><p>Ovarian cancer is a multifactorial and polygenic malignancy; therefore, any variation in a single gene will not be sufficient to provide comprehensive disease information. Fibroblast growth factors (FGFs) are a large family (24 members) of growth and differentiation factors. FGFs mediate their effects by binding to FGF receptors (FGFRs) on the cell surface. The signaling axis of FGFs and their receptors plays important roles in regulating cellular proliferation, migration, angiogenesis, wound repair, and differentiation (14). The FGF–FGFR axis has been demonstrated to modulate tumor stroma and cancer progression (15). On the other hand, FGF signaling may have tumor-suppressive functions in certain contexts (16). Compelling evidence has shown that FGF signaling pathways are implicated in cancer progression by inducing mitogenesis, cell migration, and tumor angiogenesis (16, 17). Therefore, aberrant FGF signaling can promote cancer development.</p><p>Several studies have demonstrated altered expression of genes encoding FGF receptors to be associated with ovarian cancer (18, 19). Furthermore, serum FGF2 (fibroblast growth factor 2) concentrations are increased in patients with ovarian cancer (20, 21), and amplification of FGF1 is correlated with poor survival in patients with advanced-stage serous ovarian cancer (22, 23). Similarly, altered expression of genes encoding FGFs have been reported for other human cancers (24, 25). This growing evidence for the role of FGF signaling in tumorigenesis has led to proposals for therapeutic strategies that target ovarian cancer via the FGF signaling pathway axis (14, 26, 27) The problem with developing such strategies, however, is that reports on the relationships between germline alterations in genes encoding FGFs or FGFRs and ovarian cancer are limited. Johnatty et al. investigated single-nucleotide polymorphisms (SNPs) in the FGF26 [fibroblast growth factor 2 (basic)] gene for ovarian cancer risk and observed no statistically significant associations (28); however, no studies have been conducted for other members of the FGF–FGFR axis or for associations with response to therapy or clinical outcome. Therefore, we investigated genetic variants within 24 FGF and FGFR genes for any associations with risk of ovarian cancer, therapeutic response to chemotherapy, and overall survival of patients with ovarian cancer.</p><!><p>We recruited 417 ovarian cancer cases newly diagnosed and histologically confirmed at The University of Texas MD Anderson Cancer Center between August 1991 to January 2009. There were no age, ethnicity, and clinical-stage restrictions on recruitment. In parallel, we recruited a group of healthy women (n = 417) without prior history of cancer (except nonmelanoma skin cancer) from a large pool of individuals seeing a physician for routine health checkups or addressing health concerns at the Kelsey–Seybold Clinics, a large private multispecialty physician group in the Houston metropolitan area. Cases and controls were matched by age (±5 years) and ethnicity. To minimize population admixture, we included only non-Hispanic white individuals in the current analysis (339 cases and 349 controls).</p><!><p>Epidemiology, demographic, clinical, and follow-up data were obtained from medical records. Overall survival was calculated from the date of diagnosis to the date of death or the end of patient follow-up. Response to platinum-based chemotherapy was defined as evidence of residual disease, as indicated by various clinical measures, such as positron emission tomography and computed tomography scans, second-look surgery, and postchemotherapy CA125 concentration. Each patient and control individual signed a written informed-consent form. The study was approved by the Institutional Review Boards of The University of Texas MD Anderson Cancer Center.</p><!><p>A peripheral blood sample was obtained from each study participant. Genomic DNA was extracted from peripheral blood with the QIAamp DNA Blood Maxi Kit (Qiagen) according to the manufacturer's protocol and stored for future use. For each gene encoding an FGF or FGFR, we extracted tag SNPs within 10 kb upstream of the transcriptional start site and 10 kb downstream of the transcriptional stop site. Selected tag SNPs had r2 values ≥0.80 and minor-allele frequencies (MAFs) ≥0.05. In addition, we identified potentially functional SNPs with MAF values ≥0.01, including coding SNPs and SNPs located in potential regulatory regions (promoter, splicing site, 5′ untranslated region, and 3′ untranslated region). We identified 183 SNPs in 24 genes encoding FGFs or FGFRs and sent a set of SNPs to Illumina technical support for custom iSelect Infinium II BeadChip design with Illumina's proprietary program. Genotyping followed the standard protocol for Illumina's Infinium iSelect HD Custom Genotyping BeadChip. For quality control, we randomly selected at least 2% to 3% of the samples for replicates. The concordance for all replicates was 100%. The call rate for all SNPs was 99.86%, and genotypes were autocalled with BeadStudio software (Illumina). All laboratory personnel were blinded to the case/control and outcome status of the study participants.</p><!><p>The distributions of categorical variables and continuous variables between cases and controls were evaluated by Pearson χ2 tests and Student t-tests, respectively. The χ2 test was used to evaluate each SNP for the Hardy–Weinberg equilibrium in the population of control individuals; SNPs with P values <0.01 were removed from further analysis. Multivariate logistic regression analysis was performed to estimate odds ratios (ORs) and 95% CIs for each SNP's main effect while adjusting for age. The test with the highest level of statistical significance among the 3 genetic models of inheritance (dominant, recessive, and additive) was used to determine the statistical significance of each SNP. If the frequencies of the homozygous variant genotypes were <5% in the cases or controls, however, only the dominant model with the highest statistical power was considered. The effect of each SNP on survival was assessed with a multivariable Cox proportional hazards regression analysis adjusted for age, clinical stage, histology, and treatment regimen. We used Kaplan–Meier plots and log-rank tests to assess differences in overall survival for each SNP by genotype. For response to therapy, we carried out unconditional multivariate logistic regression analysis for each SNP while adjusting for age, clinical stage, histology, and treatment regimen. The cumulative effects of multiple unfavorable genotypes were evaluated for the SNPs that showed statistical significance in the main analysis (i.e., P < 0.05). As an alternative to external validation, we performed bootstrap resampling to internally validate the results. For single-SNP analysis, we conducted bootstrap re-sampling of 1000 runs. In each run, we performed bootstrap resampling 50 times to calculate the bootstrap P value. We then counted the number of times bootstrap P values were <0.05. For classification and regression tree (CART) analysis, survival tree analysis, and unfavorable-genotype analysis, we reported the bias-corrected bootstrap confidence intervals on the basis of performing bootstrap resampling 10 000 times. We used STATA software (version 10; StataCorp) for the statistical analyses described above, and we used HelixTree software (Golden Helix) for CART analysis to explore higher-order gene– gene interactions and to classify the study participants into distinct risk groups. Survival tree analysis was conducted with the STREE program (http://c2s2.yale.edu/software/stree) to build a decision tree via the recursive-partitioning method. In brief, the root nodes contained all the patients, and we defined the measure for goodness of split with the log-rank P value to select the optimal initial split and subsequent splits of the data set until no statistically significant split was identified (29). Owing to the small number of events for terminal node 1, we used terminal node 2 as the reference group to provide more-reliable estimates of the effect and the SE, thereby providing smaller 95% CIs. The terminal nodes were classified into low-risk, medium-risk, and high-risk groups according to their relative risk compared with terminal node 2. All P values reported were the results of 2-sided tests. Multiple hypothesis testing was conducted with the "q-value" package in R by controlling the false-discovery rate to <10% (30).</p><!><p>To identify functional relevance in our findings, we checked for the potential functional effect of identified SNPs on gene expression by analyzing gene–SNP association in expression quantitative trait loci (eQTL) studies with the Genevar (GENe Expression VARiation) database (http://www.sanger.ac.uk/resources/software/genevar/) (31) in the HapMap3 data set. All analyses were performed for the CEU population with MAF values ≥0.05.</p><!><p>The characteristics of the study population have been described (13). In brief, 417 case individuals and 417 control individuals were included in this study; 339 of the cases (81.3%) and 349 of the controls (83.7%) were for non-Hispanic white individuals. The mean age of the case and control individuals was 60.73 and 60.30 years, respectively. The difference in age between the cases and the controls was not significant (P = 0.554). To minimize the effects of treatment type on survival in clinical-outcome analyses, we focused the analysis on the non-Hispanic white patients who had received surgery and platinum-based chemotherapy (n = 319). For this group, 88% of the patients had a diagnosis of advanced-stage (stages III, IV) ovarian cancer. The majority (62%) of the tumors were of the serous sub-type. The median survival time was 48.26 months (median age, 60.73 years; range, 26 – 88 years). Slightly less than half of the patients (46%) had died by the end of the follow-up period, with 48% showing cancer recurrence and 33% not responding to treatment.</p><!><p>Of the 183 SNPs we analyzed from 24 genes encoding FGFs or FGFRs, 22 SNPs from 7 genes showed a significant association with ovarian cancer risk (i.e., P < 0.05, and q < 0.10; Table 1). The SNP with highest statistical significance was FGF1 [fibroblast growth factor 1 (acidic)] rs7727832. Individuals carrying at least 1 variant allele exhibited a 2.27-fold (95% CI, 1.31-fold to 3.95-fold; P = 0.0035) increased risk of ovarian cancer, compared with individuals with the wild-type genotype. The association with this SNP remained significant in >1000 bootstrap resamplings, providing strong support for the validity of this result. Carriers of at least 1 variant allele for rs3809495 in the FGF7 (fibroblast growth factor 7) gene had the highest risk of ovarian cancer (OR, 2.99; 95% CI, 1.07– 8.42; P = 0.038). Approximately half of the significant associations were associated with increased risk, whereas the other half conferred a protective effect. The protective SNP with the highest statistical significance was rs2288696 in the FGFR1 (fibroblast growth factor receptor 1) gene. Under the dominant model, this locus was associated with a 30% reduction in risk (95% CI, 0.51– 0.93; P = 0.017); this association remained significant in >90% of the bootstrap resamplings.</p><p>Because of the large number of variants in the FGF–FGFR axis that were significantly associated with ovarian cancer risk, we conducted a CART analysis to explore higher-order gene– gene interactions among these 22 significant SNPs. The final tree structure contained 5 terminal nodes with dramatically different risks for ovarian cancer (Fig. 1). The first split on the decision tree was FGF1 rs7727832, indicating that this SNP was the primary risk factor for ovarian cancer in the study population. With individuals of terminal node 1 used as the reference, the HRs for the other 4 terminal nodes ranged from 1.80 to 6.33.</p><!><p>Four SNPs were significantly associated with treatment response after platinum-based chemotherapy and surgery (Table 2). Interestingly, each of these loci was associated with a favorable response (a 35% to 56% reduction in risk of a poor response). These SNPs were FGF18 (fibroblast growth factor 18) rs3806929 (OR, 0.64; 95% CI, 0.44 – 0.94), FGF7 rs9920722 (OR, 0.65; 95% CI, 0.44 – 0.98), and FGF23 (fibroblast growth factor 23) rs12812339 (OR, 0.65; 95% CI, 0.43– 0.98) under the additive model, and FGF5 (fibroblast growth factor 5) rs3733336 (OR, 0.044; 95% CI, 0.19 – 0.98) under the recessive model. None of the SNPs remained significant, however, after adjustment for multiple comparisons (q > 0.1), and only 1 variant, rs3806929, reached significance in >50% of the bootstrap resamplings.</p><!><p>Eleven SNPs from 7 genes were significantly associated with overall survival of ovarian cancer patients, with the results remaining highly significant in the bootstrap resampling (Table 3). FGF23 SNP rs7961824 had the highest significance. This association remained significant after multiple comparisons (q = 0.058) and was significant in 100% of the bootstraps. Individuals with at least one rs7961824 allele had a 45% reduction in risk of dying during the follow-up period (95% CI, 0.39 – 0.78), compared with women with the wild-type genotype. This favorable prognosis dramatically improved the median survival time (by 20 months), from 38.0 months for patients with the wild-type genotype to 58.1 months for those with the variant allele.</p><p>The remaining variants were borderline significant after adjustment for multiple comparisons (q values, 0.19 – 0.22), suggesting that they could have an effect on overall survival in the appropriate context. To assess this possibility, we performed a survival tree analysis for these 11 variants. The resulting tree structure comprised SNPs from 5 genes: FGF2, FGFR2 (fibroblast growth factor receptor 2), FGFR4 (fibroblast growth factor receptor 4), FGF14 (fibroblast growth factor 14), and FGF23 (Fig. 2A). The first split on the survival tree was FGF2 rs167428, indicating that this SNP is the primary factor contributing to overall survival. When we used individuals of terminal node 2 as a reference, the HRs for the other 6 terminal nodes ranged from 0.21 to 2.21. Grouping these terminal nodes into 3 risk groups—low, medium, and high— identified dramatic differences in survival durations. The median survival times for the patients in the low-risk, medium-risk, and high-risk groups were 128.95, 53.06, and 25.16 months, respectively (P = 8.21 × 10−7, log-rank test; Fig. 2B).</p><p>Next, we determined the cumulative effects of multiple unfavorable genotypes on ovarian cancer survival. Compared with patients carrying ≤2 unfavorable genotypes (low risk), patients carrying 3– 6 unfavorable genotypes (medium risk) and 7–10 unfavorable genotypes (high risk) exhibited a progressively increased risk of death, with HRs of 4.41 (95% CI, 1.97–9.87) and 7.86 (95% CI, 3.34 –18.52), respectively (P for trend = 2.49 × 10−7). This increase in risk produced highly significant differences in the median survival time. Patients in the medium-risk and high-risk groups had survival times of only 49.44 and 26.64 months, respectively, compared with 99.24 months for those in the low-risk group (P = 9.09 × 10−7, log-rank test; Fig. 3).</p><!><p>To check for possible functional effects of our identified SNPs on gene expression, we investigated whether any of the significant variants showed eQTL with gene expression with the Genevar database tool. One SNP (FGF1 rs17099029) associated with survival was in cis eQTL with FGF1 expression (P = 0.03; see Fig. 1 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol60/issue1). Individuals carrying a genotype with the variant had decreased FGF1 expression compared with those with the wild-type homozygous genotype. Interestingly, this SNP was also in linkage with the FGF1 SNP rs7727832 (r2 = 0.5) that was highly associated with ovarian cancer risk.</p><!><p>Ovarian cancer is one of the most lethal gynecologic malignancies (1). The high mortality is often attributable to late diagnosis. Thus, considerable efforts have been made to identify high-risk populations to look not only for genetic markers associated with ovarian cancer risk that could improve early detection and screening approaches, but also for predictors of clinical outcomes. Several studies have been conducted to identify variants associated with ovarian cancer risk and clinical outcomes, but many of the findings of these studies are inconsistent (6 –10). Because of the increasing knowledge regarding the role of the FGF– FGFR axis in tumorigenesis in general and ovarian cancer specifically, we performed a comprehensive analysis of genetic variation in this pathway with respect to its association with ovarian cancer risk and outcomes.</p><p>One of the most important findings is that multiple SNPs in the FGF–FGFR pathway are associated with increased risk of ovarian cancer. In particular, FGF1 rs7727832 showed the most significant association with ovarian cancer in the main-effect analysis, and this result was supported by the CART gene– gene interactions analysis. Also showing significant associations with ovarian cancer were variants in 6 other genes, including FGFR1, FGFR2, FGF7, FGF9 (fibro-blast growth factor 9), FGF10 (fibroblast growth factor 10), and FGF14. Studies have shown that FGF10 over-expression leads to epithelial hyperproliferation (32) and that the FGF10 –FGFR1 or FGF10 –FGFR2 signaling axis plays a potential role in oncogenic transformation (33). Furthermore, evidence suggests that FGF1 overexpression may lead to increased angiogenesis and autocrine stimulation of cancer cells (23, 34). Specifically for ovarian cancer, there is strong evidence that these specific ligands and receptors are important in ovarian cancer tumorigenesis. For example, FGF1 overexpression has been observed in ovarian cancer and been associated with clinical prognosis (23, 24). Likewise, FGF7 is expressed in the majority of ovarian tumors and is detectable in malignant ovarian cancer– associated ascites (35). Cole et al. showed that FGFR2 and FGF7 stimulate ovarian cancer proliferation (36). An early study observed that the FGF2 and FGFR1 genes are highly expressed in human ovarian tumor endothelium (37). FGFR2 has been suggested to be involved in ovarian cancer pathogenesis (18). Indeed, increased FGFR2 levels have been found in ovarian cancer (36). The interaction of FGF7 with FGFR2 ligands and with FGFR2-IIIb induces proliferation, motility, and protection from cell death in epithelial ovarian cancer cell lines (38). Interestingly, approximately half of the significant associations observed were with a reduced risk of ovarian cancer, suggesting that FGF– FGFR signaling can exhibit tumor-suppressive functions in certain contexts (16).</p><p>We also identified that 4 SNPs (FGF18 rs3806929, FGF7 rs9920722, FGF23 rs12812339, and FGF5 rs3733336) were significantly associated with improved response to treatment. Experimental data have indicated that inhibition of FGFR along with the use of standard chemotherapeutic agents, including those used in ovarian cancer treatment, interacted synergistically to improve the therapeutic effect on endometrial cancer cells (39). Even more intriguing is that Cole et al. showed that the inhibition of FGFR1 and FGFR2 increased cisplatin sensitivity in ovarian cancer (36). Studies have demonstrated that FGFR4 is a potential therapeutic target and associated with a better therapeutic response, although we observed no association of variants within FGFR4 with any treatment response in our study (19, 20). Nevertheless, our results support targeting of the FGF–FGFR signaling axis as a potential therapeutic opportunity in ovarian cancer. Such targeting may provide a future approach to stratifying the patient population to improve the response to these agents.</p><p>Another major finding is that 11 SNPs from 7 genes were significantly associated with overall survival in ovarian cancer patients. Of these SNPs, FGF23 rs7961824 showed the most significant association with survival. Survival tree analysis revealed FGF2, FGFR2, FGFR4, FGF14, and FGF23 gene– gene interactions, indicating FGF2 rs167428 to be the primary factor contributing to overall survival. Patients in the low-risk group carrying these alleles have much longer survival times (>40 months) than those in the medium- and high-risk groups. In addition, the highly significant dose–response effect evident for an increasing number of unfavorable genotypes provides support for the global effect of variation within these important signaling genes. Previous studies have indicated that FGF1 overexpression may lead to increased angiogenesis, leading to poorer overall patient survival (23), and a recent study demonstrated that FGF1 expression adversely influenced survival in patients with ovarian tumors (24).</p><p>Interestingly, we found that one of the FGF1 SNPs associated with ovarian cancer survival (rs17099029) demonstrated a significant eQTL with FGF1 expression (Genevar analysis) (40). This SNP shows a degree of linkage with FGF1 rs7727832 (r2 = 0.5), the SNP with the highest risk, suggesting that changes in FGF1 expression may modify risk for both cancer susceptibility and outcome. rs17099029 is located in the 3′ untranslated region of FGF1, thus hinting at a possible role in affecting gene expression. Taken together, the novel findings of these genetic variants associated with ovarian cancer survival suggest, in the context of established tumor biology, that they have potential for use as prognostic markers for ovarian cancer.</p><p>While recognizing the original and significant findings of this current study, we understand that our study has some limitations. It is a hospital-based case control study, and thus selection bias may be present. The effects of these variants observed for non-Hispanic whites may not be generalizable to populations of other ethnicities. In addition, the SNPs we genotyped in this study are primarily tagging SNPs and most likely are not the true causal or functional variants. Therefore, it is difficult to ascertain the underlying biological mechanisms for these significant associations. Additional functional characterizations are needed. Nevertheless, our genetics-driven study used a well-characterized patient population with detailed clinical, treatment, and follow-up information. Although internal validation was conducted with bootstrap methods to add a level of confidence to the findings, further independent or external validation is necessary to verify the findings in our study.</p><p>In conclusion, we have conducted the first comprehensive study to identify significant associations of genetic variants in the FGF–FGFR pathway with ovarian cancer risk, therapeutic response, and survival. The association of multiple SNPs in the FGF pathway could provide a molecular approach for the development of new ovarian cancer biomarkers for these key end points, and the use of such markers could have a major impact on the survival of patients with this devastating disease. Furthermore, these findings lend support to the development of the FGF–FGFR axis as potential therapeutic targets in ovarian cancer.</p>

PubMed Author Manuscript

Structural Assembly of Molecular Complexes Based on Residual Dipolar Couplings

We present and evaluate a rigid-body molecular docking method, called PATIDOCK, that relies solely on the three-dimensional structure of the individual components and the experimentally derived residual dipolar couplings (RDC) for the complex. We show that, given an accurate ab initio predictor of the alignment tensor from a protein structure, it is possible to accurately assemble a protein-protein complex by utilizing the RDC\xe2\x80\x99s sensitivity to molecular shape to guide the docking. The proposed docking method is robust against experimental errors in the RDCs and computationally efficient. We analyze the accuracy and efficiency of this method using experimental or synthetic RDC data for several proteins, as well as synthetic data for a large variety of protein-protein complexes. We also test our method on two protein systems for which the structure of the complex and steric-alignment data are available (Lys48-linked diubiquitin and a complex of ubiquitin and a ubiquitin-associated domain) and analyze the effect of flexible unstructured tails on the outcome of docking. The results demonstrate that it is fundamentally possible to assemble a protein-protein complex based solely on experimental RDC data and the prediction of the alignment tensor from three-dimensional structures. Thus, despite the purely angular nature of residual dipolar couplings, they can be converted into intermolecular distance/translational constraints. Additionally we show a method for combining RDCs with other experimental data, such as ambiguous constraints from interface mapping, to further improve structure characterization of the protein complexes.

structural_assembly_of_molecular_complexes_based_on_residual_dipolar_couplings

Introduction<!>Methods<!>Formulation<!>Efficient Computation of the Alignment Tensor<!>Algorithm<!>Additional Constraints<!>Results and Discussion<!>Docking Using Ideal Synthetic Data<!>Robustness of RDC-Guided Docking to Experimental Noise<!>Docking using Experimental RDC Data<!>Docking Using Experimental RDC Data: Combining Alignment and Translation<!>Application to a Real System: Ubiquitin/UBA Complex<!>Application to a Real Dual-Domain System: Lys48-linked di-Ubiquitin<!>Docking Using Experimental RDC Data Combined with Ambiguous Interface-Related Restraints<!>Docking Using Unbound Structures<!>Conclusions<!>

<p>Detailed understanding of molecular mechanisms underlying biological function requires knowledge of the three-dimensional structure of biomacromolecules and their complexes. Nuclear magnetic resonance (NMR) spectroscopy is one of the main methods for obtaining information on molecular structure and interactions at atomic-level resolution. A major challenge in using NMR for accurate structure determination of multidomain systems and macromolecular complexes is the scarcity of long-distance structural information. Intermolecular Nuclear Overhauser Effect (NOE) contacts are often scarce, difficult to detect, and could be affected by intermolecular motions. Chemical shift perturbation (CSP) mapping is another powerful method for general identification of the interface. However, its informational content is highly ambiguous because CSPs do not identify pair-wise contacts and should be used with caution, since a perturbation of the local electronic environment of a nucleus does not necessarily indicate direct involvement of the corresponding atom in the interactions. Moreover, both NOEs and CSPs are limited to the contact area and could be insufficient for accurate spatial arrangement of the interacting partners. Residual dipolar couplings (RDCs), resulting from partial molecular alignment in a magnetic field,1,2 could supplement the scarce interdomain data, because they contain valuable structural information in terms of global, long-range orientational constraints (reviewed in3). In addition, RDCs also inevitably reflect (hence are sensitive to) the physical properties of the solute molecule responsible for its alignment. Thus, a commonly used method for aligning proteins in solution takes advantage of the anisotropy of molecular shape by imposing steric restrictions on the allowed orientations of the molecule. Such steric alignment can often be modeled as caused by planar obstacles (see e.g.,2,4); we will refer to this simplified model of molecular alignment as the barrier model.</p><p>The alignment of a rigid molecule can be characterized by the so-called alignment tensor. Several methods have been developed4–8 to use the barrier model for predicting the alignment tensor (and with it the RDCs) either directly from the 3D shape of the molecule or indirectly, using an ellipsoid representation. The RDCs' sensitivity to molecular shape has the potential for improving structure characterization, especially in multi-domain systems and macromolecular complexes, by fully integrating RDC prediction into structure refinement protocols to directly drive structure optimization. In fact, RDCs have been used to orient domains and bonds relative to each other either directly, using rigid-body rotation,9–13 or by incorporating RDCs as orientational restraints into protein docking14–16 (see e.g., the reviews17,18). However, none of these methods has used the information on the shape of the molecule (including not only the intervector/interdomain orientation but also the actual positioning of the individual domains) embedded in the measured RDCs.</p><p>Another physical property sensitive to molecular shape is the overall rotational diffusion tensor, characterizing the rates and anisotropy of the overall tumbling of a molecule in solution. Interestingly, although they reflect distinct physical phenomena (rotation versus orientation) the diffusion and the alignment tensors are oriented similarly, provided the alignment is caused by neutral planar obstacles.19 As demonstrated recently by Ryabov and Fushman,20 the sensitivity of the overall rotational diffusion tensor to molecular shape can be utilized to guide molecular docking. One would expect that the alignment tensor could be used similarly. Given that accurate RDC measurements for a wide variety of bond vectors are readily available, the use of the alignment tensor to guide molecular assembly could be of significant value for a broad range of macromolecular systems. However, to our knowledge, the ability to dock molecules using the alignment tensor has not been demonstrated, and RDCs have never been used to completely drive molecular docking, i.e. not only orient but also properly position molecules/domains relative to each other in a complex.</p><p>In this paper we demonstrate that it is possible to determine the structure of a complex by utilizing the sensitivity of RDCs to molecular shape, provided that the structures of the individual components of the complex are available. We describe a method for rigid-body molecular docking based solely on the orientation- and shape-related information embedded in the experimental RDCs/alignment tensor of the complex. This method, called PATIDOCK, uses a recently developed computationally efficient algorithm (PATI8) for ab initio prediction of the alignment tensor from the three-dimensional shape of a molecule. We demonstrate that PATIDOCK can deterministically and efficiently perform rigid-body docking based on the alignment tensor. In addition, we analyze the robustness of PATIDOCK under certain types of experimental errors, examine its performance in applications to real experimental data, and discuss challenges and various ways of refining the results by including other available experimental restraints and integrating our method into more sophisticated docking approaches.</p><!><p>Here we present a method, called PATIDOCK, for rigid-body assembly of a molecule made up of two distinct sets of atoms (hereafter called domains) whose structures are known, by using experimental RDC values exclusively. The method is based on first rotating/aligning the two domains using the corresponding subsets of the RDC values (see e.g.,9,11,13) and then translating/positioning them relative to each other in order to minimize the difference between the predicted A and the experimental à alignment tensors. A is computed for the complex using the barrier-model-based algorithm PATI, while à is derived directly from the RDC values, measured for the whole molecule, using a linear least squares approach (see e.g.,8,21) and the (already aligned) 3D structures of the individual domains. As discussed in Berlin et al.,8 PATI predicts RDCs with the same accuracy as program PALES,4 while its computational efficiency is achieved by using numerical integration and a convex hull representation of the molecular surface. Note that while some parts of the docking algorithm are specific to the use of PATI, the general algorithm and key concepts can be applied to any current or future method for alignment tensor prediction.</p><!><p>We formulate the docking algorithm as a minimization problem. The algorithm is based on minimizing the difference between the predicted alignment tensor A, computed based on the structure/shape of the molecule, and the experimental alignment tensor Ã, derived directly from the experimental RDC values.</p><p>Let the set S of atoms of a molecule be subdivided into two distinct sets (domains), S1 and S2, such that S1 ∩ S2 = ∅, S1 ∪ S2 = S, no RDC-active bond is shared between the two sets, and each set contains enough bond vectors/RDCs associated with it to provide a proper sampling of the orientational space required for accurate determination of the alignment tensors.22 We define A(Rc, x) as the predicted alignment tensor of S, where the coordinates of atoms in S1 remain static and the coordinates of atoms in S2 are rotated by some rotation matrix Rc and then translated by x = [x1, x2, x3]. Our goal is to first properly orient the two sets by finding the optimal rotation matrix, R*, and then to find the optimal translation vector x* that minimizes the difference between A(R*,x) and Ã. The separation of orientation from translation is possible because inter-domain orientation can be obtained directly from the experimental RDCs and bond vectors for each set,9,11,13 regardless of their relative position.</p><p>To solve for R* we simply align S1 and S2 relative to each other using experimental RDC data, as described in.9,11,13 We first compute the experimental alignment tensors, Ã1 and Ã2, of S1 and S2, respectively. The alignment tensors have eigendecompositions A∼1=R1D1R1T and A∼2=R2D2R2T, where R1, R2 are rotation matrices (orthogonal matrices with determinant of 1) and D1, D2 are the diagonal matrices of principal components of the corresponding alignment tensors. Therefore, R* can be derived by solving the equation R*R2 = R1: (1)R∗=R1R2T.</p><p>Note that due to orientational degeneracy of the alignment tensor there is a four-fold ambiguity in the relative alignment of domains, hence four possible solutions for R*.13 One can find these possible solutions by computing an eigendecomposition of Ã2, determining the four assignments of signs to the columns of R2 that make det(R2) = 1, and using Equation (1) for each one. Note that in the case when two or more eigenvalues of the alignment tensor are close to each other (e.g. very low rhombicity) it might not be possible to accurately orient the two domains. In this case additional experimental information, e.g. in the form of interdomain contacts (see below), could come to rescue.</p><p>Knowing the optimal rotation matrix R*, we find the optimal translation vector x* by solving a nonlinear least squares problem. Since R* is derived directly from the experimental RDC data independent of x*, in the rest of the paper (except for the last sections) we assume that the two subsets are already properly aligned and simplify the notation from A(Rc, x) to A(x). Our nonlinear least squares problem is then formulated as: (2)x∗=argminxχ2(x), where the target function is defined as</p><p> (3)χ2(x)=∑i,j=13(Aij(x)−A∼ij)2. and the computation of A(x) is described in the next section.</p><!><p>In this section we reformulate PATI, from the formulae presented in Berlin et al.,8 to one that can be efficiently recomputed multiple times on S under different translations of S2.</p><p>From equations for PATI in Berlin et al.,8 given a set of atoms S and a unit vector b = [b1, b2, b3] in the direction of a static magnetic field, the predicted alignment tensor A of S can be expressed as: (4)Aij=1N∫02π∫−11Fij(α,u)η(α,u)dudα,i,j=1,2,3,N=h−14π∫02π∫−11η(α,u)dudα, where 2h is the distance between the planar barriers oriented orthogonal to the z axis. η(α,u) is the difference between the z-coordinate of the center of the molecule and the minimum z-coordinate value of all points in S at a given orientation of the molecule, specified by the zyz Euler rotation angles [α, β, γ], and u = cosβ. See the Supporting Material for definition of the Euler rotation and matrix F, and see Berlin et al.8 for how h is defined and how to compute η(α,u) from a set of atoms of a molecule by building a convex hull. In practice, the interbarrier distance can be estimated directly from the bicelles concentration (see e.g.8,23). In the case of PEG/hexanol medium, our analysis based on the available experimental RDC data (see8 and the Results section) suggests that h = 400–500Å provides a reasonable estimate. Given the computational efficiency of our method (see below), this value could be further adjusted iteratively.</p><p>Since the molecule consists of two domains with an unknown translation x* between them, η will depend on translation x, α, and u. (This implies that A and N also depend on x.) Therefore, we modify our notation from η(α,u) to η(x,α,u), where x is the vector of translation of the coordinates of all atoms of S2.</p><p>Without loss of generality, let the center of S1 be at 0, and the center of S2 be at m̃, both of which are inside their associate convex hulls. We compute η for S1 and S2 separately, and call them η1(α,u) and η2(α,u). Note that η1(α,u) and η2(α,u) do not depend on x. The combined η(x,α,u) of the two sets (domains) is the largest of the two η, where η2 is adjusted to reflect that S2 is centered at m̃ + x, and is computed as</p><p> (5)η(x,α,u)={η1(α,u)ifη1(α,u)≥η2(α,u)−ϒ(x),η2(α,u)−ϒ(x)otherwise, where</p><p>Precomputing F(α,u), η1(α,u), η2(α,u), and R(α, arccos u,0) for a fine enough set of [α,u] allows us to quickly compute A(x) for multiple values of x.</p><!><p>In this section we describe how to solve the minimization problem posed in Equation (2). We use a nonlinear least squares solver, specifically the Levenberg-Marquardt algorithm,24 due to the limited number of local minima, local convexity, and smoothness of our target function.</p><p>An efficient nonlinear least squares solver requires a Jacobian to be computed or approximated using finite differences. Fortunately in this case, the Jacobian elements can be computed: (7)∂Aij(x)∂xk=1N(x)∫02π∫−11Fij(α,u)∂η(x,α,u)∂xkdudα+Aij(x)4πN(x)∫02π∫−11∂η(x,α,u)∂xkdudα, where</p><p> (8)∂η(x,α,u)∂xk={0ifη1(α,u)≥η2(α,u)−ϒ(x),−R3k(α,arccosu,0)otherwise, and i, j, k = 1,2,3.</p><p>Due to translational symmetry of the problem, there can be two significant local minimizers of our target function: the actual minimizer, and the incorrect minimizer where domain S2 is located on the opposite side of domain S1 (see an example in the Results section). In addition, if the convex hull of S2 is fully inside S1 then our target function has derivatives of 0, and the minimization algorithm might become trapped on a plateau. Therefore, picking the right set of initial guesses is important.</p><p>To assure that the convex hull of S2 is not inside S1 we place any initial starting point x0i at a distance d = maxα,uη1(α,u) from the center of S1. We pick a set of six initial positions, [d, 0, 0], [−d,0,0], [0,d,0], [0,−d,0], [0,0,d], and [0,0,−d], to make sure that during the minimization we approach S1 from different directions and therefore are likely to find all the minimizers. We refer to this method for finding the optimal translation between two domains as PATIDOCK-t. Additionally, we refer to the method that first aligns the two domains using Equation (1) and then finds the optimal translation using PATIDOCK-t as PATIDOCK.</p><!><p>As demonstrated earlier,8 there is inaccuracy in barrier model-based prediction of the alignment tensor of a molecule. This inaccuracy would contribute to errors in the docking solution if we just minimized the target function χ2(x) (Equation (3)). In order to mimic a real situation, when additional experimental data are available, we examine whether the RDC-based docking could be improved by introducing additional restraints to enforce intermolecular distance constraints and avoid steric clashes.</p><p>Obviously, introduction of specific intermolecular distance constraints (e.g. from NOEs) would significantly improve docking by positioning the corresponding atoms (hence the domains carrying them) at the proper distance from each other. However, intermolecular NOEs are often unavailable or averaged out by molecular motions such as domain dynamics, association/dissociation events, etc. Therefore, we analyze the effect of adding "milder", ambiguous restraints, often used for molecular docking based on interface mapping16,25,26 using chemical shift perturbations (CSPs). CSPs quantify NMR signal shifts in the presence of a binding partner, and their observation represents the basic and perhaps the simplest way to monitor intermolecular interactions by NMR. The CSPs provide a general qualitative map of atoms/residues involved in the interface, without any specific information about pair-wise contacts. Thus, we construct a "CSP-like" energy function based on ambiguous information on intermolecular contacts. To prove the concept of including additional constraints into RDC-guided docking, we forgo the complicated modeling and data refinement of the actual CSPs. Instead we simply label an atom as being "CSP-active" if the CSP for it is significantly high. For the molecules for which we do not have CSP data, for simple testing purposes we generate a synthetic CSP-active list by selecting all the atoms in one domain that are within a certain distance, dΩ, of any atom in the other domain, and would therefore potentially experience a CSP in an experimental setting. We define the subsets of atoms from S1 and S2 that are CSP-active as I1 and I2 respectively.</p><p>Let Dij(x) be the distance between two atoms, si ∈ S1 and sj ∈ S2, when the atoms in S2 are translated by x. To generate the energy function for the CSP-like constraints we weigh an atom in the CSP-active set as 0 if it is currently interacting with atoms in the other domain; otherwise we assign some penalizing value as the atom's weight. To handle outliers we stop the growth of the penalty at a cutoff distance dΩcut. Specifically, the CSP-active weights for the two domains are</p><p> (9)Ω1i(x)={0ifminjDij(x)≤dΩorsi∉I1,minjDij(x)−dΩifdΩ<minjDij(x)≤dΩcutandsi∈I1,dΩcut−dΩotherwise, and</p><p>Note that in this proof-of-principle study we use a single dΩ value for all atoms/residues. A future refinement of the method might require adjusting this parameter depending on the length and the nature of the contacting side chains. We sum the average weights to form the target function for the CSP-like interactions: (11)χΩ2(x)=∑i[Ω1i(x)]2∣I1∣+∑j[Ω2j(x)]2∣I2∣, where |·| is the cardinality of the set.</p><p>To prevent physically impossible overlap (steric clash) of the domains we assign a penalizing value to atoms that are closer than a given distance dΨ to atoms in the opposing domain. The weights</p><p> (12)Ψ1i(x)={dΨ−minjDij(x)ifminjDij(x)<dΨ,0otherwise, (13)Ψ2j(x)={dΨ−miniDij(x)ifminiDij(x)<dΨ,0otherwise, form the target function for the domain-overlapping constraints: (14)χΨ2(x)=∑i[Ψ1i(x)]2+∑j[Ψ2j(x)]2.</p><p>We now combine the alignment tensor, CSP-like, and domain-overlapping constraints into one energy function</p><p>In this study dΩ= 4Å, dΨ= 0.9Å, and dΩcut=10Å. The weight of 100 for χΨ2 was chosen as just a very large value that would penalize even minimal overlap significantly more than any violation of a CSP-like interaction. We set κ = 1.23 × 105 (see Supporting Material for derivation of the constant κ).</p><p>We reformulate Equation (2) to use χF2 instead of χ2, and solve this problem to improve the minimizer from PATIDOCK. We refer to this method as PATIDOCK+. The new target function cannot be solved using local minimization. Therefore, we use a branch and bound method27 to deterministically solve Equation (15) for the global minimizer.</p><!><p>In order to examine the feasibility of molecular docking guided by RDCs, we applied PATIDOCK-t, PATIDOCK, and PATIDOCK+ to several protein systems. Potential sources of inaccuracy in our docking approach are errors in the experimental data (RDCs) and the inaccuracy in the barrier model prediction of molecular alignment. To separate and quantify these errors we tested our method on two distinct datasets as well as two protein-protein systems. The first dataset, which we refer to as COMPLEX, is a set of 84 protein-protein complexes described in Mintseris et al.28 This dataset provides a wide variety of interprotein contacts and molecular shapes, but it contains no experimental RDC data. We used this dataset to generate synthetic RDC data and examine the validity of our docking method and its sensitivity to common measurement errors due to experimental imprecision. This allowed us to test our method under "ideal experimental conditions", i.e. when the simple barrier model is an adequate physical model for molecular alignment, and the only errors in the data originate from (random) experimental noise in the measurements.</p><p>The second dataset, which we refer to as SINGLE, consists of 7 monomeric proteins for which experimental RDC data (in bicelles- or PEG/hexanol-based media) are available in the BMRB database.29 We utilized this dataset previously to test PATI predictions.8 These experimental RDC data are used here to gauge the accuracy of our docking method under real experimental conditions and the inaccuracies inherent to the barrier model's prediction of the alignment tensor. Similar to the COMPLEX dataset we also generated synthetic RDC data for this set of proteins, as a control. Since these are single-domain proteins, to use this dataset for testing docking, we artificially created a molecular "complex" using a plane to arbitrarily bisect each protein molecule into two distinct sets of atoms. See Figure 1A and Figure 1B for an illustration of how Cyanovirin-N is cut into two domains by a plane.</p><p>Finally we applied our method to two protein-protein systems for which we have experimental RDC and CSP data: ubiquitin/UBA complex30 (PDB code 2JY6) and lysine-48-linked di-ubiquitin14 (PDB code 2BGF). These complexes allow us to present a "real world" practical application for PATIDOCK. We show that it is possible to quickly get a good solution for a complex using only the alignment tensor. In addition we show that combining our method with a more complicated energy function that accounts for additional factors such as van der Waals interactions and CSPs can yield an accurate solution in practice.</p><p>We implemented PATIDOCK in MATLAB 7.8.0 and performed all calculations and timing on a single core of 3.16 GHz Pentium Core 2 Duo E8500 processor with 3.25 GB of RAM, running Windows XP Service Pack 3. The set of [α,u] values for which we precompute F, η1, η2, and R was determined by the adaptive numerical integration of Equation (4) with an absolute error of 0.05 (using MATLAB's quad function, see e.g.,31). The terminating condition for the Levenberg-Marquardt algorithm (MATLAB's lsqnonlin function) was set to the step size less than 0.1Å. The 0.05 value was determined empirically based on the highest tolerance value which still gave docking solutions accurate to within 0.3Å for synthetic RDCs for all complexes in the COMPLEX and SINGLE datasets. Accuracy can be increased, at the expense of time, by changing the tolerance to the numerical integration routine. Note, however, that the improvement in accuracy is limited by the inherent inability of the barrier model to fully model the physical conditions.</p><p>Due to the four-fold ambiguity of the relative orientation of domain S2 with respect to S1 and the existence of multiple local minimizers (with regard to translation) for each orientation, we expect to have at least eight potential solutions. The solutions are ranked by the backbone RMSD between the experimental structure of the complex and the predicted one, where the atom positions in S2 are adjusted by R* and x* (recall that S1 is fixed in space). Only the results for the lowest-RMSD solution are shown in this paper. Since R* can be directly computed from the experimental RDC data independent of our model, we first focus our analysis on the minimizers that come from the correct orientation of the two domains. We then present the results for the complete docking method that also includes automatic alignment of the two domains, in addition to their positioning relative to each other.</p><!><p>In order to demonstrate the feasibility of structural assembly of molecular complexes based solely on RDC data, we first applied PATIDOCK-t to synthetic data generated for proteins from the COMPLEX and SINGLE datasets.</p><p>To test our ability to find the correct minimizer under ideal conditions, for each complex we generated a synthetic alignment tensor Ãsyn using PATI prediction. From this and the three-dimensional structure of the complex we calculated RDCs for all amide NH bonds, which we call synthetic RDCs, assuming that there is no noise in experimental measurements. The synthetic RDCs along with the three-dimensional structures of the two domains comprise the input to our minimization algorithm. We will rate our results based on the "Δc", the smallest distance between the original and all the predicted center of the second domain. The results for PATIDOCK-t, using Ãsyn as the "experimental" alignment tensor, are presented in Table 1 (columns "0 Hz", "Time(s)", and "#Sol.") for the SINGLE dataset. The results for the COMPLEX dataset under ideal conditions (labeled "0 Hz" in Figure 2) are very similar (also see Supporting Information). These results clearly demonstrate that it is possible, under ideal conditions, to accurately and efficiently assemble molecular complexes based solely on RDC data.</p><!><p>In reality, RDC values always contain measurement errors, which are usually below 1 Hz. To assess the effect of such errors on the RDC-guided docking we added to the synthetic RDCs normally distributed noise with standard deviation of 1 Hz or 3 Hz. This allowed us to test whether it is possible to accurately dock a complex based solely on the alignment tensor in the presence of considerable (1 Hz) or extreme (3 Hz) noise in the data. Figure 2 shows errors in the docking solutions for the COMPLEX dataset in the presence or absence of random noise in the generated RDC values. Very similar results were obtained using synthetic RDC data (with noise) generated for the SINGLE dataset; see Table 1, columns "1 Hz" and "3 Hz".</p><p>From these results (Figure 2 and Table 1) we conclude that PATIDOCK-t is able to find correct docking solutions for a wide variety of proteins even under heavy (3 Hz) experimental noise. These results validate the concept of molecular docking based exclusively on the alignment tensor.</p><p>PATIDOCK-t is also extremely fast, as it takes only seconds to dock two domains on a single PC. This speed makes it feasible to perform RDC-based docking at each iteration step of a more complicated flexible docking algorithm, for example by analyzing docking of multiple conformers at each minimization iteration. Another potential consequence of the speed is that it opens up the possibility of extending the docking algorithm to three or more molecules. Since we are able to accurately dock molecules given perfect prediction of the alignment tensor, the accuracy of the results in practice will depend on how well we can predict the alignment tensor in an experimental setting.</p><!><p>Having established the ability to accurately assemble molecular complexes using synthetic data, we next test our method on the alignment tensors derived from actual experimental data, in order to understand how errors in prediction of the alignment tensor affect the overall accuracy of docking. We use for this purpose the 7 proteins of the SINGLE dataset. The alignment tensor prediction and the limitations of the barrier model for these proteins were addressed in detail in our previous publication.8 Since the errors in the experimental RDC data for these proteins are about or smaller than 1 Hz, based on our results with synthetic data (Table 1) we expected to get a good solution provided that the barrier model is a good predictor of the alignment tensor. The results for PATIDOCK-t are shown in Table 2.</p><p>Surprisingly, these solutions are worse than one would expect based just on the errors in the experimental data. Given that with synthetic RDC data these proteins were docked properly (see Table 1) this suggests that the alignment tensor predicted using a simple barrier model differs from the actual tensor, and this discrepancy could translate into an error (about 4.3Å) in the docking solution. In fact, as shown previously in Berlin et al.,8 the inaccuracy in alignment tensor prediction can be approximately separated into an error in the magnitude (scaling) of the tensor and an error in its orientation. On the positive side, however, the results in Table 2 show that by using only RDC data we are able to place the second domain on average within a radius of 4.3Å of its proper position.</p><!><p>The docking efforts presented above focused on domain translation, while keeping interdomain orientation the same as in the original structure. We now combine our method for determining the correct translation with the method for aligning the two domains based on the orientations of the alignment tensor of the complex "reported" by each individual domain.9,11,13 This is the complete method, PATIDOCK, that takes two domains with arbitrary positions and orientations, and the associated experimental RDC values, and assembles their complex automatically with no human intervention at any step.</p><p>We first align the two domains by extracting (from the experimental RDC data for the complex) the alignment tensors "seen" by each domain and using Equation (1) to properly orient the second domain relative to the first one. Once the domains are oriented, we compute the experimental alignment tensor of the whole complex, Ã, from the RDC data and the combined bond vectors of the first domain and the newly oriented bond vectors of the second domain. This step helps average out experimental error and improve the accuracy of the resulting experimental alignment tensor by increasing the number of bond vectors used (generally resulting in improved orientational sampling22 and statistical averaging). We then use PATIDOCK to compute the proper translation between the now aligned domains. Due to the four-fold ambiguity in alignment we expect the number of solutions and the computation time to increase by a factor of four. The results for PATIDOCK with all potential solutions are shown in Table 3. Note that no domain alignment was performed in the PATIDOCK-t docking shown in Table 2, so the values in "Δc" column of that table are also "RMSD2" values as defined in Table 3.</p><p>The error in the relative position of the second domain (see Δc in Table 3) changed only slightly (an increase by 0.08Å on average) compared to the PATIDOCK-t method. Combined with the small increase (0.15Å on average) in RMSD2 values from the fixed-orientation assembly in Table 2 (values in the "Δc" column) to the align-and-translate assembly in Table 3, these results indicate that alignment of domains by using experimental RDC values is a robust and accurate technique and is not a significant contributor of error to structure assembly. As expected, there is a four-fold increase in the number of possible solutions and the running time, but the combined algorithm still completes in less than four seconds.</p><!><p>We now test our method on a protein complex for which experimental RDC and CSP data are available: the complex of human ubiquitin (Ub) with the UBA domain of ubiquilin-130 (PDB code 2JY6). Using the experimental CSP data we defined as CSP-active residues L8, T9, G10, K48, E51, R54, Q62, H68, L71, and L73 in Ub, and M557, G558, L560, I570, A571, N577, E581, R582, L584 in UBA. See Figure 3 for the mapping of the CSP-active residues onto the Ub/UBA complex. In this section we will only use the RDC data, while the CSP data will be included in a later section.</p><p>A potential complication for the rigid-body docking approach arises in the case of the Ub/UBA complex from the fact that both proteins have extended unstructured and highly flexible tails. In fact, residues 73–76 in Ub and 536–544 in the UBA construct used in the experimental study experience large-amplitude motions30 on a ps-ns time scale, which is many orders of magnitude faster than the time scale (~100 ms) of a NMR experiment. These motions are also present in the Ub/UBA complex, reflecting the fact that the tails do not participate in the binding.30 Naturally, such tails present a significant challenge for shape-sensitive computations like those in the current study, because no single structure can represent the ensemble/motion-averaged molecular shape relevant for a particular experiment. This raises important questions that have not been addressed in the literature so far: could flexible tails simply be neglected (clipped off) in such calculations or should they be represented by a structural ensemble, and how large does the latter need to be? In order to address these questions, we performed docking for both the structural ensembles and the clipped (tailless) structures. Because the RDC data were measured in the PEG/hexanol medium,12 the actual inter-barrier distance was unknown and had to be estimated. We set h = 400Å, a value that gives the correct scaling between the predicted and experimentally determined alignment tensor at the known solution.</p><p>To sample various orientations of the tails (not present in the original PDB structure of the complex), we extracted 10 representative orientations of Ub's C-terminus from the NMR ensemble of Ub monomer (PDB code 1D3Z36) and 10 possible orientations of the N-terminus of the UBA domain from its NMR ensemble in the monomeric state (PDB code 2JY530). These conformations of the tails were superimposed onto the corresponding domains in the complex structure (2JY6), thus creating an ensemble of 100 possible models for the Ub/UBA complex (shown in Figure 3). We refer to this Ub/UBA complex as Structure 2jy6-I. From the 100 models of Structure 2jy6-I we were able to estimate the variance in the docking solutions that the two tails introduce. The results are presented in Table 4.</p><p>Because averaging by fast reorientations of the tails is expected to diminish the tails' effect on the alignment tensor, we clipped off the two tails from the structures of the corresponding proteins and then docked the two tailless molecules using PATIDOCK-t and PATIDOCK. We refer to the tailless Ub/UBA complex as Structure 2jy6-II; the results are presented in Table 4. Figure 4 shows the isosurface plot of the energy function χ2 for the tailless Ub/UBA complex and the visualization of the two solutions from PATIDOCK-t. The isosurface plot clearly demonstrates that there are two distinct minima in the energy function, both of which were found by our program. As can be seen from Figure 4C and Figure 4D, the reason for the two minima is that both solutions have very similar convex hulls due to the geometric symmetry inherent in the problem.</p><p>As evident from Table 4, the conformation(s) of the tail can have a profound effect on the results of docking. The solution varies on average by 2Å over all the possible combinations of tail orientations, whereas removing the tails improves the results significantly. This suggests that a potential solution for dealing with flexible tails in RDC-guided docking is to clip them off rather than use a specific conformation or try to deduce the "averaged" conformation of the tail. Without the tails, using PATIDOCK, we get the Δc and RMSD2 of about 3.7Å, which are smaller than the expected average position error of about 4.4Å (see Table 2 and Table 3).</p><!><p>Finally, we tested our method on a dual-domain system for which both experimental RDC and CSP data are available: the Lys48-linked di-Ubiquitin12–14 (PDB code 2BGF). Using the experimental CSP data we define hydrophobic-patch residues L8, I44, and L70 on both domains to be CSP-active. See Figure 5 for the mapping of the CSP-active residues onto the di-Ubiquitin (Ub2) structure. The CSP data will be used in a later section. Because the RDC data were measured in the PEG/hexanol medium,12 the actual inter-barrier distance was unknown and had to be estimated. We set h = 550Å, a value that gives the correct scaling between the predicted and experimentally determined alignment tensor at the known solution.</p><p>As in the case of the Ub/UBA complex, a potential complication for the rigid-body docking approach arises from the unstructured and highly flexible C-terminal tails comprising residues 73–76 of each domain,13 though the tail in Ub is much shorter than that of UBA. We therefore performed a similar analysis to that in the previous section. However, instead of superimposing the tails onto the Ub2 complex, we simply took the ensemble of the 10 models from the Ub2 structure 2BGF (shown in Figure 5). We refer to this ensemble as Structure 2bgf-I. Similarly, we created Structure 2bgf-II by taking the first model in 2BGF and clipping off residues 73–76 of both domains. The results for the ensemble and the clipped (tailless) structures are presented in Table 5.</p><p>As above, the conformation of the tail has noticeable effect on the results of docking, although significantly less than in the Ub/UBA complex. The solution varies on average by 1Å among all the possible tails' conformations, and removing the tails improves the results slightly. These results further support the conclusion that the potential solution for dealing with flexible tails in RDC-guided docking is to clip off the tails. Without the tails, using PATIDOCK, we get the errors in positioning of the second domain of Δc = 3.6Å and RMSD2 = 3.7Å, which are smaller than the expected average value of about 4.4Å (see above).</p><!><p>The results in previous sections using real experimental data give a good hint at the errors that one can expect when using the barrier model as the alignment tensor predictor. Thus, we expect that in practice the error in domain positioning using PATIDOCK would be less than 5Å. The fact that the results are a relatively short distance from the actual solution demonstrates that the alignment-tensor-based χ2 is a useful constraint.</p><p>We now seek to improve upon the previous results by combining CSP-like constraints along with the alignment tensor constraints by minimizing χF2 (see Equation (15)). The combination of constraints should lead to a better and more reliable overall solution. The results of applying PATIDOCK+ to the SINGLE dataset, Ub/UBA, and Ub2 are presented in Table 6. Note that we are now able to select the correct structure out of all possible solutions by picking the one with the lowest χF2 value. The cartoon representations of the solutions for the two protein-protein systems are presented in Figure 6.</p><p>As evident from Table 6, the addition of ambiguous, CSP-like restraints significantly improved the solution for all proteins, compared to the results in Table 3, Table 4, and Table 5. The docked solutions for the two "real" complexes (Ub/UBA and Ub2) based entirely on experimental RDC and CSP data have both Δc and RMSD2 below 2Å. This indicates that combining RDCs with other experimental intermolecular constraints in a real situation could be a powerful method for quickly yielding good docking solutions. The additional benefit of using CSP-like restraints is that we now are able to correctly identify the best solution from the eight or more possible symmetry-related solutions based just on the χF2 values.</p><!><p>In some docking applications structures of the individual components in the bound state might not be known in advance, but are to be determined in the process of docking, for example, using the "unbound" structures of the domains as the starting point. We therefore examine how accurately our method positions two domains relative to each other given only the RDC data for the bound complex and the unbound structures of the two domains, i.e. how robust our method is with regard to structural rearrangements in the individual components resulting from binding interactions. Generally, we anticipate several sources of inaccuracy in the resulting RDC-guided complexes when using unbound structures of the individual components. These include (i) inaccuracy in the derived experimental alignment tensor(s), due to a different orientation of the RDC-active bond vectors, and (ii) a different 3D shape of each component (and their complex), which would affect the predicted alignment tensor. Perturbations in intermolecular contacts at the interface, reflecting different orientations of the side chains, could also affect the accuracy of docking, when contact-based restraints are included (see above).</p><p>Here we take advantage of the availability of both bound and unbound structures for the 84 proteins of the COMPLEX dataset.28 The synthetic RDCs generated for each bound complex as described above (zero noise) were used as input "experimental" RDC data for the same complex, but applied to unbound structures of each domain. Using the NH bond vectors of the unbound structures and the synthetic RDCs, we computed the alignment tensors "reported" by each of the domains, and used the same docking procedure as above (PATIDOCK-t or PATIDOCK) to assemble the corresponding complex of the unbound individual components.</p><p>We compare the resulting structures (docked "unbound" complexes) with the corresponding complexes of the bound structures in Figure 7. The results are presented in terms of RMSDs for all backbone atoms. These numbers should be compared to the "Base" RMSD level (red bars in Figure 7) that reflects the structural differences between the unbound and bound structures of the individual domains, calculated by superimposing the unbound structure of each domain onto the bound structure in the complex and computing the overall (backbone) RMSD. The results show that structural/dynamic rearrangements in the individual components upon complex formation do not dramatically affect the relative domain positioning in the resulting RDC-guided structures. The average error in the position of the second domain (Δc) for PATIDOCK-t and PATIDOCK was about 5Å.</p><p>Finally, we examine the performance of RDC-guided docking of unbound structures when using real experimental RDC data. We use the unbound tailless structures of ubiquitin (PDB code 1D3Z) and UBA (PDB code 2JY5) to assemble the Ub/UBA and Ub2 complexes using experimental RDCs for their bound complex. The results, shown in Table 7, are similar to the COMPLEX dataset using synthetic data, shown in Figure 7.</p><p>These results indicate that the RDC-guided docking is relatively robust with respect to structural rearrangements induced by complex formation. This is likely due to statistical averaging during the RDC to alignment tensor conversion. Moreover, this finding also suggests that the unbound structures of the individual components could be used as a crude, initial approximation for the complex assembly, to be followed by more rigorous docking steps that allow structural flexibility and adaptation necessary for final adjustment of the individual components in the complex.</p><!><p>In this paper we demonstrated that it is fundamentally possible to assemble a protein-protein complex based solely on experimental RDC data and the prediction of the alignment tensor from three-dimensional structures, provided that the structures of the individual components are available. To achieve this, we reduced the multitude of experimental RDCs to a single alignment tensor consisting of five independent parameters, and then used the latter to guide positioning and orientation of one domain relative to the other. During the docking process, the alignment tensor acts as a "mechanical" constraint applied to the interdomain vector and forcing the individual components to adopt a particular position within the molecule such that the molecular shape of the resulting complex matches that of the real one (as far as the alignment tensor is concerned). The ability to assemble a molecular complex using RDCs is remarkable, because it shows that despite the purely angular nature of residual dipolar couplings, they can be translated into distance/translational constraints. This is due to RDC's sensitivity to molecular shape and reflects the fact that it is the shape of the molecule that causes its steric alignment.</p><p>The PATIDOCK method is robust with respect to large experimental errors in RDC data, provided there are a sufficient number of experimental RDCs. This is not surprising since the alignment tensor "averages" the information contained in the RDCs. By extension, the inherent averaging of RDCs in the alignment tensor makes PATIDOCK also somewhat robust against local structural rearrangements/dynamics associated with complex formation. When applied to real experimental data, PATIDOCK gives on average a less than 5Å error in the relative positioning of the molecules. We demonstrated that the resulting structure could be further refined by including other available experimental data (PATIDOCK+). Moreover, the presence of extended unstructured/flexible parts (e.g. tails) in a molecule can potentially affect the solution by more than 2Å, depending on which structure/conformation of such parts is chosen. We propose removal of the flexible tails as a potential solution to this problem.</p><p>The PATIDOCK methods are extremely fast, and therefore we do not foresee a need for a faster, but less accurate, method for prediction of the alignment tensor than PATI. Potential improvements in the prediction of the alignment tensor will most likely involve (i) representing individual molecular components as structural ensembles rather than single structures and (ii) using a weight function inside the integrals in Equation (4), to account for possible non-steric interactions with the aligning medium. For example, such a function could weigh η differently, or introduce charge potentials in case of non-neutral alignment media (see e.g.,23). We foresee such an addition as being easily adapted into our docking method.</p><p>It is worth mentioning that accutate characterization of protein-protein complexes should account for contributions to the experimental RDC data from free components in fast exchange with the complex (see e.g.40). This is particularly true for weak macromolecular interactions. Application to such systems would require modification of the target function in equation (3) to include the contributions to experimental data from the free form of the interacting partners.</p><p>The PATIDOCK approach presented in this paper can potentially be used in several ways. First, it provides a quick rigid-body docking method whose solutions can be utilized to significantly limit the search space of a more complicated flexible-docking algorithm. The robustness of the approach with respect to structural rearrangements suggests that the RDC-guided docking could be used early on in the process of molecular complex assembly, e.g., starting with the unbound structures of the individual components, and subsequently refining them as the computation progresses. Second, our energy functions can be included as an additional term into a more general energy function that accounts for all other structure-related constraints such as distance and torsional angle restraints, hydrogen bonding, electrostatic and van der Waals potentials, etc. Moreover, the computational efficiency of the PATIDOCK method makes it feasible to perform RDC-based docking at each iteration step of a more complicated flexible-docking algorithm, for example by analyzing docking of multiple conformers at each minimization iteration. The molecular-shape-based RDC-guided docking can be incorporated into existing structure determination/refinement protocols (e.g. HADDOCK,25 XPLOR-NIH41). This would allow us to account for side chain and backbone flexiblity at the inerface and integrate with all other available experimental data. A recent XPLOR-NIH implementation42 of the diffusion-tensor-guided docking method20 serves as an example. Third, PATIDOCK can be used as the main method for driving molecular docking in a situation where there is a lack of unambiguous intermolecular structural information, like NOEs. This last application will become more practical as methods for prediction of the alignment tensor improve. Fourth, the energy function designed here could potentially also be used to evaluate and refine protein structures, including those for single-domain proteins, based on how well the 3D shape of the molecule agrees with experimental RDC data.</p><p>The fact that our docking method is extremely fast for two-domain complexes opens up the possibility of extending the PATIDOCK approach to three or more domains. Even though each additional domain gives rise to an exponential increase in complexity and time, it is still possible to quickly evaluate our energy function for a multitude of domains.</p><!><p>Illustration of the bisection of Cyanovirin-N (PDB code 2EZM). (A) Van der Waals surface of Cyanovirin-N. (B) Illustration of how the protein is split into two domains with approximately equal number of atoms by a plane. The first domain is colored green, the second domain is red.</p><p>PATIDOCK-t docking results for the 84 complexes in the COMPLEX dataset using synthetic RDC values with no noise (0 Hz, red circles) or in the presence of a Gaussian noise with the standard deviation of 1 Hz (green squares) or 3 Hz (blue diamonds) (see Supporting Table S1). (A) PATIDOCK-t docking results when all of the NH bond vectors are used in the computation of the alignment tensor. (B) PATIDOCK-t docking results when only 100 randomly selected NH bond vectors from the complex are used. Similar results were obtained when using only 50 randomly selected NH bond vectors (Supporting Table S1). The height constant h was adjusted for each complex to give a Da value of 20 Hz for Ãsyn, which corresponds to the average Da value of the SINGLE dataset, ubiquitin/UBA complex, and di-ubiquitin complex. In the case of noisy data, docking of each complex was performed six times, with individual RDC errors randomly selected from a normal distribution. All six results for each complex with RDC errors are plotted. For the purposes of visualization a few outliers for complexes 43 and 46 are not displayed. Bigger errors for some complexes reflect a much lesser sensitivity of the molecular shape (hence of the alignment tensor) of these specific complexes to translations of one domain relative to the other. (C-F) Van der Waals surface representation of the major outliers: (C-D) complex #43, PDB code 1I4D (mass 47 kDa, S1=chain D, S2=chains A and B); (E-F) complex #46, PDB code 1IBR (mass 77 kDa, S1=chain B, S2=chain A). The structures in (D) and (F) are rotated counterclockwise around the z-axis by 90°. The individual domains are colored green (S1) and red (S2), the convex hull of the complex is colored light blue.</p><p>A cartoon representation of the ensemble of 100 possible models for the Ub/UBA complex (Structure 2jy6-I). Ub is colored green, UBA is in red, the flexible tails are colored blue, and the CSP-active residues are represented by spheres around their Cα atoms.</p><p>The results of RDC-guided docking for the tailless Ub/UBA complex (2jy6-II) using PATIDOCK-t. Shown are (A–B) isosurface plots of the χ2(x) function and (C-D) the associated van der Waals surfaces (wrapped by their convex hulls) of the two solutions corresponding to the two local minima of χ2(x). The isosurfaces correspond to (A) minxχ2(x) + 0.1σ and (B) minxχ2(x) + 0.6σ, for all x inside the grid, where σ is the standard deviation of the values of χ2 in the grid. The isosurface data were collected on a 100 × 100 × 100 Å grid around 0. (C) The best (closest) solution with the UBA domain positioned to the right of Ub, with χ2 = 2.01 × 10−7 at the solution. (D) The incorrect solution where the UBA domain is to the left of Ub, with χ2 = 1.24 ×10−7 at the solution. In these van der Waals surface plots Ub is colored green and UBA is red. Both solutions have a very similar convex hull, hence similar predicted alignment tensor. The camera angle relative to Ub's orientation is the same in both figures. Note that the best solution has a higher χ2 value.</p><p>A cartoon representation of the ensemble of 10 models for the di-Ubiquitin complex (Structure 2bgf-I). Proximal domain is colored green, distal domain is in red, the flexible tails are colored blue, and the CSP-active residues are represented by spheres around their Cα atoms.</p><p>A cartoon representation of the actual structure (green) vs. the docked structure (red) for the (A) Ub/UBA complex and (B) Ub2 molecule based on minimization of χF2. Only the adjusted domain (S2, right) is shown for the docked structures, the other domain (S1, left) superimposes exactly with the corresponding domain in the actual structure.</p><p>The results of PATIDOCK-t (green bars) and PATIDOCK (blue bars) assembly of complexes of "unbound" structures of the proteins from the COMPLEX dataset, using synthetically generated alignment tensors from the corresponding "bound" complexes as the target experimental alignment tensor to guide the docking. Shown are backbone RMSDs between the resulting (unbound) complex and the original (bound) complex. "Base" RMSDs (red bars) reflect the structural differences between the unbound and bound structures of the individual domains, calculated by superimposing the unbound structure of each domain onto the bound structure in the complex and computing the overall RMSD. Missing bars correspond to those few complexes where we were unable to properly match the atoms between the bound and the unbound coordinate sets.</p><p>The results of RDC-guided docking using PATIDOCK-t for the SINGLE dataset based on synthetic RDC data with added experimental noise.</p><p>The RCSB Protein Data Bank code for protein coordinates. First model from the ensemble of NMR structures was used for all calculations.</p><p>Δc (in Å), the best distance between the original and the predicted centers of the second domain. The values in brackets represent the RMSD (in Hz) between the synthetic RDCs and the predicted RDCs at the solution. The column labels represent the size of the standard deviation of the normally distributed noise added to synthetic RDCs. "0 Hz" corresponds to no noise added to synthetic RDCs.</p><p>The values represent an average of twelve independent runs.</p><p>The average elapsed time (in seconds) for PATIDOCK-t based on all the runs for "0 Hz", "1 Hz", and "3 Hz".</p><p>The number of possible solutions, all of which have a very similar predicted alignment tensor.</p><p>The results of RDC-guided docking using PATIDOCK-t for the SINGLE dataset based on experimental RDC data.</p><p>The RCSB Protein Data Bank code for protein coordinates. First model from the ensemble of NMR structures was used for all calculations.</p><p>The distance (in Å) between the original and the predicted center of the second domain.</p><p>The RMSD (in Hz) between the experimental and the predicted RDC values at the best predicted minimizer.</p><p>The elapsed time (in seconds) required for docking.</p><p>The number of possible solutions, all of which have a very similar predicted alignment tensor.</p><p>The results of RDC-guided docking using PATIDOCK for the SINGLE dataset based on experimental RDC data.</p><p>The RCSB Protein Data Bank code for protein coordinates. First model from the ensemble of NMR structures was used for all calculations.</p><p>The backbone RMSD (in Å) between the original complex structure and the predicted complex. The structures are optimally rotated and centered using the center of mass.39</p><p>The backbone RMSD (in Å) between the coordinates of atoms of the second domain for the original and the predicted complex.</p><p>The distance (in Å) between the original and the predicted center of the second domain. The center is computed as the average of the positions of all the atoms in the domain.</p><p>The RMSD (in Hz) between the experimental and the predicted RDC values at the best predicted minimizer.</p><p>The elapsed time (in seconds) required for docking of all four orientations.</p><p>The number of possible solutions, all of which have a very similar predicted alignment tensor.</p><p>The results of docking the Ubiquitin/UBA complex using PATIDOCK-t and PATIDOCK.</p><p>2jy6-I is the ensemble of 100 structures representing various conformations of Ub and UBA tails (see text), whereas in 2jy6-II the tails were clipped off.</p><p>The method that was used to dock the complex.</p><p>The backbone RMSD (in Å) between the original complex structure and the predicted complex. The structures are optimally rotated and centered using the center of mass.39</p><p>The backbone RMSD (in Å) between the coordinates of atoms of the second domain for the original and predicted complex.</p><p>The distance (in Å) between the original and the predicted center of the second domain. The center is computed as the average of the positions of all the atoms in the domain.</p><p>The RMSD (in Hz) between the experimental and the predicted RDC values at the best predicted minimizer.</p><p>The number of possible solutions, all of which have a very similar overall alignment tensor.</p><p>Values are the means of the individual values for the best solution of each of the 100 models.</p><p>Values in the parentheses are the standard deviations of the individual values for the best solution of each of the 100 models.</p><p>The results of docking Lys48-linked di-Ubiquitin using PATIDOCK-t and PATIDOCK.</p><p>2bgf-I is the ensemble of 10 structures representing various conformations of the C-terminal tails of both Ub molecules (see text), whereas in 2bgf-II the tails were clipped off.</p><p>The method that was used to dock the complex.</p><p>The backbone RMSD (in Å) between the original complex structure and the predicted complex. The structures are optimally rotated and centered using the center of mass.39</p><p>The backbone RMSD (in Å) between the coordinates of atoms of the second domain for the original and the predicted complex.</p><p>The distance (in Å) between the original and the predicted center of the second domain. The center is computed as the average of the positions of all the atoms in the domain.</p><p>The RMSD (in Hz) between the experimental and the predicted RDC values at the best predicted minimizer.</p><p>The number of possible solutions, all of which have a very similar overall alignment tensor.</p><p>Values are the means of the individual values for the best solution of each of the 100 models.</p><p>Values in the parentheses are the standard deviations of the individual values for the best solution of each of the 100 models.</p><p>The results for PATIDOCK+ using a combination of CSP-like and alignment tensor constraints.</p><p>See previous tables and Results section for structure references.</p><p>The backbone RMSD (in Å) between the original complex structure and the predicted complex. The structures are optimally rotated and centered using the center of mass.39</p><p>The backbone RMSD (in Å) between the coordinates of atoms of the second domain for the original and predicted complex.</p><p>The distance (in Å) between the original and the predicted center of the second domain. The center is computed as the average of the positions of all the atoms in the domain.</p><p>The RMSD (in Hz) between the experimental and the predicted RDC values at the best predicted minimizer.</p><p>The number of possible solutions, all of which have a very similar χF2.</p><p>The results of docking the unbound Ubiquitin/UBA and Lys48-linked di-Ubiquitin complexes using PATIDOCK-t and PATIDOCK.</p><p>For this docking we used unbound tailless structures of ubiquitin (PDB code 1D3Z) and UBA (PDB code 2JY5). The resulting structures of the ubiquitin/UBA and di-ubiquitin complexes were compared with the corresponding tailless (bound) complexes, 2JY6-II and 2BGF-II, respectively.</p><p>The method that was used to dock the complex.</p><p>The structural differences (in Å) between the unbound and bound structures of the individual domains, calculated by superimposing the unbound structure of each domain onto the bound structure in the complex and computing the overall RMSD.</p><p>The backbone RMSD (in Å) between the original complex structure and the predicted complex. The structures are optimally rotated and centered using the center of mass.39</p><p>The backbone RMSD (in Å) between the coordinates of atoms of the second domain for the original and the predicted complex.</p><p>The distance (in Å) between the original and the predicted center of the second domain. The center is computed as the average of the positions of all the atoms in the domain.</p><p>The RMSD (in Hz) between the experimental and the predicted RDC values at the best predicted minimizer.</p><p>The number of possible solutions, all of which have a very similar overall alignment tensor.</p>

PubMed Author Manuscript

Optimised GMP-compliant production of [18F]DPA-714 on the Trasis AllinOne module

BackgroundThe translocator protein 18 kDa is recognised as an important biomarker for neuroinflammation due to its soaring expression in microglia. This process is common for various neurological disorders. DPA-714 is a potent TSPO-specific ligand which found its use in Positron Emission Tomography following substitution of fluorine-19 with fluorine-18, a positron-emitting radionuclide. [18F]DPA-714 enables visualisation of inflammatory processes in vivo non-invasively. Radiolabelling of this tracer is well described in literature, including validation for clinical use. Here, we report significant enhancements to the process which resulted in the design of a fully GMP-compliant robust synthesis of [18F]DPA-714 on a popular cassette-based system, Trasis AllinOne, boosting reliability, throughput, and introducing a significant degree of simplicity.Results[18F]DPA-714 was synthesised using the classic nucleophilic aliphatic substitution on a good leaving group, tosylate, with [18F]fluoride using tetraethylammonium bicarbonate in acetonitrile at 100∘C. The process was fully automated on a Trasis AllinOne synthesiser using an in-house designed cassette and sequence. With a relatively small precursor load of 4 mg, [18F]DPA-714 was obtained with consistently high radiochemical yields of 55-71% (n=6) and molar activities of 117-350 GBq/µmol at end of synthesis. With a single production batch, starting with 31-42 GBq of [18F]fluoride, between 13-20 GBq of the tracer can be produced, enabling multi-centre studies.ConclusionTo the best of our knowledge, the process presented herein is the most efficient [18F]DPA-714 synthesis, with advantageous GMP compliance. The use of a Trasis AllinOne synthesiser increases reliability and allows rapid training of production staff.

optimised_gmp-compliant_production_of_[18f]dpa-714_on_the_trasis_allinone_module

Introduction<!>Materials and reagents<!><!>Trasis allinOne synthesiser<!>Cyclotron<!><!>Synthesis preparation<!><!>Preparation of reagent kit and cartridges<!>GMP production of [18F]DPA-714<!><!>HPLC for identification, chemical and radiochemical purity determination<!>GC for residual solvent and ethanol content determination<!>Thin layer chromatography spot test for determination of tetraethylammonium content<!>Bacterial endotoxins test<!><!>Quality control<!>pH<!>Filter integrity test<!>Radionuclidic identity and purity<!>Sterilty<!>Environmental monitoring<!>Stability studies<!>Optimisation of radiochemistry<!>Automated production<!><!>GC<!>Bacterial endotoxins test<!>Tetraethylammonium<!><!>Routine clinical productions<!>Discussion<!>Conclusion<!>

<p>The translocator protein (18 kDa), TSPO in short, is found on the outer mitochondrial membrane and has the highest expression in steroidogenic tissues (Lee et al. 2020; Choi et al. 2011). The protein was formerly known as the peripheral benzodiazepine receptor (PBR), the name, however, was judged to misrepresent its role, following extensive characterisation. Neither were benzodiazepines its only binding targets (omitting cholesterol, among others), nor was "peripheral" an accurate description of its localisation, as one its expression sites is the brain itself (Papadopoulos et al. 2006). Typically, TSPO expression in the central nervous system (CNS) is modest and limited primarily to glial cells, particularly astrocytes and microglia (Selvaraj and Stocco 2015). A marked increase in TSPO density is observed as a result of neuroinflammation mediated by the activation of microglia. This mechanism is implicated in various neuropathologies and is the common denominator in their early stages. TSPO expression has been suggested as a biomarker for brain disorders characterised by activated microglia, such as Alzheimer's disease (AD), multiple sclerosis (MS), stroke and cancer (James et al. 2008).</p><p>Positron Emission Tomography (PET) is a powerful molecular imaging technique for detection of TSPO density changes, with the potential to become a useful diagnostic tool for neuroinflammatory responses, particularly relevant in early disease stages. The first TSPO targeting tracer, (R)-[11C]PK11195 (1-(2-chlorophenyl)-N-[11C]methyl-N-(1-methylpropyl)-3-isoquinoline carboxamide), was employed in several studies to investigate microglial activation, however its suboptimal non-specific binding component deteriorates image quantification due to the low signal-to-noise ratio (Guo et al. 2012). Extensive research led to the elucidation of numerous ligands, including [11C]PBR28 (N-acetyl-N-(2-[11C]methoxybenzyl)-2-phenoxy-5-pyridinamine) and [18F]PBR111 ((2-(6-chloro-2-(4-(3-[18F]fluoropropoxy)phenyl)imidazo[1,2-a]pyridin-3-yl)-N,N-diethylacetamide), known collectively as second-generation TSPO tracers. The former exhibited particularly favourable kinetics and specificity for the protein of interest (Collste et al. 2016). A clear advantage of [18F]PBR111 over [11C]PBR28 is the use of fluorine-18, exhibiting superior decay properties and consequently, higher resolution images. In addition, similar improvements with respect to the signal-to-noise ratio of the PET images were reported (Eberl et al. 2017). Recently, a structural derivative of [18F]PBR111, [18F]DPA-714 or (N,N-diethyl-2-(2-(4-(2-[18F]fluoroethoxy)phenyl)-5,7-dimethylpyrazolo[1,5-a]pyrimidin-3-yl)acetamide received significant attention as a superior TSPO ligand, owing to high specific binding and brain-blood barrier penetration (Golla et al. 2015). Biodistribution and blocking studies with unlabelled DPA-714 and PK11195 were performed by Vicidomini and co-workers using µPET in the brain and peripheral tissues of mice (Vicidomini et al. 2015). Golla et al. employed the tracer in a study with 10 AD patients and 6 healthy volunteers to determine the appropriate kinetic model for data quantification (Golla et al. 2015). Recently, Hagens et al. performed a proof-of-concept study with 8 MS patients and 7 healthy controls to evaluate the imaging power of the tracer in progressive MS (Hagens et al. 2018). Promising results of various investigations have fuelled efforts into optimising radiosynthesis of [18F]DPA-714 for clinical use, requiring a robust and high-yielding process, with an appropriate degree of automation for multi-patient and multi-centre clinical trials.</p><p>Initial synthetic efforts were described by James et al., who achieved 18F-labelling of the tosylate DPA-714 precursor via SN2 substitution with [18F]fluoride in the presence of Kryptofix-222 and potassium carbonate in refluxing acetonitrile for 10 min. The production was performed on a GE TRACERlab MXFDG system. Decay-corrected (dc.) radiochemical yields (RCYs) were modest, reaching 21% only with a high precursor loading of 6-12 mg (James et al. 2008). Damont and co-workers performed the radiolabelling analogously, however, exchanging acetonitrile for DMSO at 165∘C and shortening the reaction time to 5 min (Damont et al. 2008). Pleasingly, this led to an increase of RCY (ndc.) up to 20%, corresponding to 35% dc., with molar activities (Am) of 37-111 GBq/µmol at the end of synthesis (EOS). Kuhnast et al. published a detailed protocol of their high-yielding radiolabelling of the tracer on a TRACERlab FXFN synthesiser (Kuhnast et al. 2012). The authors performed the reaction with tosylate DPA-714 in the presence of Kryptofix-222 and potassium carbonate to effectuate substitution of the tosylate moiety with [18F]fluoride. The transformation was achieved in either DMSO at 165∘C or MeCN at 120∘C, for 5 min in both circumstances, however the authors obtained significantly higher RCYs of 43-50% (dc.) with the high-boiling solvent. High Am values were also obtained, reaching 222 GBq/µmol at EOS. Kuhnast et al. successfully led their [18F]DPA-714 synthesis through the GMP validation process, a critical step for clinical studies and potential marketing authorisation.</p><p>Herein, we describe efforts towards development of a robust [18F]DPA-714 radiosynthesis for clinical use at Radboud University Medical Center (Radboudumc), following a short yet fruitful optimisation, that resulted in, to the best of our knowledge, the most efficient and highest-yielding radiolabelling of the tracer reported in literature. Our fully GMP-compliant process was designed in-house on a Trasis AllinOne (AIO) synthesiser, making use of its simplicity, robustness and versatility, and in addition, enabling rapid training of radiopharmacy personnel.</p><!><p>All reagents and solvents, unless specified otherwise, were purchased from Sigma Aldrich (Zwijndrecht, Netherlands) and VWR International (Amsterdam, Netherlands), and used without further purification. All materials were purchased from VWR International (Amsterdam, Netherlands), unless specified otherwise. Saline solution for injection (0.9% NaCl w/v%) and water for injection were acquired from B. Braun (Melsunger, Germany). The cassette was prepared in-house, starting from a commercially available [18F]FDOPA cassette from Trasis (Ans, Belgium). The GMP-grade precursor to [18F]DPA-714, DPA-714 tosylate, was purchased from Pharmasynth (Tallinn, Estonia). Sep-Pak Accell Plus QMA carbonate and Sep-Pak C18 Plus cartridges were purchased from Waters Corporation (Etten-Leur, Netherlands). Fluorine-18 for labelling was obtained on-site using a Siemens Eclipse HP cyclotron, in the form of [18O]H2O-bound [18F]fluoride via a 18O(p,n)18F nuclear reaction in a tantalum target. 18O-enriched water (≥97%) was purchased from Rotem Industries (Dimona, Israel). Cyclotron-specific parts are described in "Cyclotron" section.</p><!><p>Results of the optimisation study with tosylate DPA-714 precursor in anhydrous acetonitrile at 90∘C after 10 min with different PTC/base systems. Each run was performed with the indicated volume of the PTC/base combination dissolved in acetonitrile/water 70/30 v/v% for Entry 1 and 85/15 v/v% for Entries 2 and 3</p><p>anon-isolated, estimated by (radio)HPLC</p><!><p>Radiosynthesis of [18F]DPA-714 was performed on a Trasis AllinOne (AIO) module (Ans, Belgium). The synthesiser is equipped with 36 rotary actuators (arranged in two rows) to which a dedicated disposable cassette can be attached. The cassette for [18F]DPA-714 productions was assembled in-house, starting from a commercially available Trasis [18F]FDOPA cassette, as described in "Cassette preparation" section.</p><!><p>[18F]fluoride is produced using a Siemens Eclipse HP (11 MeV, 2 x 80 µA) cyclotron using 2-3 mL 18O-enriched water (>98%) contained in a tantalum target with a 50 µm thick Havar window. Typically, irradiation is performed with 60 µA 11 MeV protons for 20 min, yielding 30-40 GBq of [18F]fluoride at the end of bombardment (EOB). Transport of [18F]fluoride in [18O]H2O from the cyclotron target to the designated hot cell, through 35-40 m long 1/16" OD PTFE tubing from Bohlender GmbH (Grüsfeld, Germany), is effectuated using argon and helium gas (Argon Scientific 6.0, Helium Scientific 6.0, both from Linde Gas, Schiedam, Netherlands) in a two-stage delivery system. To ensure high Am and RCYs, transfer lines towards the designated hot cell are rinsed with 18O-enriched water using an in-house rinsing system on the day of the production (preferably shortly before irradiation).</p><!><p>Layout of the cassette designed for the radiosynthesis of [18F]DPA-714 on a Trasis AIO module. The process can be controlled via the user software interface</p><p>Layout of the commercially available cassette designed for the radiosynthesis of [18F]FDOPA on a Trasis AIO module by Trasis</p><p>Outline of the [18F]DPA-714 cassette on the Trasis AllinOne radiosynthesiser, with description of reagent and materials position. Positions outlined in boldface are original parts of the commercially available [18F]FDOPA cassette</p><!><p>Preparation of Trasis AIO synthesiser Prior to the production, the synthesiser underwent internal tests. After completion, the cassette was installed and evaluated using an automatic sequence to ensure that all connections had been secured in place and all components functioned properly.</p><p>Preparation of dispensing Dispensing was performed in a GMP Grade A hot cell equipped with fully automated open vial dose divider system from Von Gahlen. The incoming product line from the Trasis AIO was first rinsed with ethanol, followed by water for injection. A new dispensing set and appropriate sterile filters were then installed.</p><!><p>QMA eluent solution. Tetraethylammonium bicarbonate (TEAB; 11.5 mg, 60.1 mmol) in acetonitrile/water (85/15 v/v%; 1.5 mL) for elution of [18F]fluoride from the QMA cartridge - attached to position 2 of the cassette.</p><p>Precursor solution. Tosylate-DPA 714 precursor (4 mg; 7.3 mmol) dissolved in anhydrous acetonitrile (1 mL) - attached to position 8 of the cassette.</p><p>Quench solution. 20% EtOH in water for injection - attached to position 10 of the cassette.</p><p>Ethanol solution. 70% EtOH in water for injection for elution of the tracer from the Sep-Pak Plus C18 cartridge - attached to position 12 of the cassette.</p><p>0.9% NaCl solution. Saline solution for injection - attached to position 16 of the cassette.</p><p>HPLC product collection vial. Vial prefilled with 14 mL water for injection for receipt of the HPLC product fraction - attached to the cassette via position 13 and the HPLC unit.</p><!><p>Prior to attachment to the cassette, the Sep-Pak C18 Plus cartridge was washed with absolute ethanol (5 mL), followed by water for injection (10 mL) and dried with air (5 mL). The preconditioned QMA cartridge supplied with the [18F]FDOPA cassette was used as such.</p><!><p>Cyclotron-produced [18O]H2O-bound [18F]fluoride (30-40 GBq) was first introduced into a Von Gahlen Activity Distribution System (ADS), where it was collected in a vial fixed inside an ISOMED 2010 dose calibrator from MED Nuclear Medizintechnik (Dresden, Germany), allowing for accurate and automatic activity measurements. Thereafter, it was distributed into the appropriate hot cell under a flow of helium ("Cyclotron" section). The difference between the accumulated activity (following target unloading) and the leftover activity (following receipt of the bulk [18F]fluoride volume in the designated Trasis AIO) was recorded as the starting activity. In the module, [18F]fluoride was trapped on a QMA cartridge, from which it was released with a solution of TEAB (40 mM; 0.75 mL) into a pre-heated reactor (60∘C). [18F]fluoride was dried by azeotropic distillation under stepwise heating. The reactor was cooled to 65∘C before addition of the tosylate DPA-714 precursor solution (4 mg) in anhydrous acetonitrile (1 mL). The mixture was heated to 90∘C for 10 min. After cooling to 60∘C, the reaction was quenched with 20% EtOH (1.5 mL) and injected on the HPLC unit for purification. The reactor and HPLC load lines were rinsed with the quench solution twice more (total volume of 8 mL) to ensure that the maximum amount of activity was recovered from the reactor. HPLC purification was performed on a Phenomenex Luna C18(2) column (100 Å 5 μm; 250 ×10 mm) at room temperature, with 0.1 M ammonium acetate/ethanol (55/45 v/v%) as the mobile phase, running at 4 mL/min and a recording wavelength of 254 nm. The fraction containing [18F]DPA-714 (elution starting approximately 13-14 min post-injection) was collected into a sterile vial prefilled with water for injection (14 mL). The solution was withdrawn and passed through a Sep-Pak C18 Plus cartridge to remove the mobile phase. The cartridge was subsequently washed with 5 mL of saline, followed by elution of the trapped tracer with 70% ethanol (2 mL) and dilution with 0.9% NaCl solution for injection (8 mL).</p><!><p>QC part 1 vial. Initial QC was performed immediately after the production. It aimed to determine the product identity, chemical and radiochemical purity, pH, the presence of bacterial endotoxins and of residual tetraethylammonium (TEA) ions. Typically, 0.5-1 mL of the active product was filled.</p><p>QC part 2 vial. Final QC was performed after decay (between 1-8 days post-production). It aimed to determine radionuclidic impurities, sterility and the presence of ethanol and residual solvents. Typically, 0.5-1 mL of the active product was filled.</p><p>Sterility vial. The vial was sent for sterility testing to a third party. Results were made available a few weeks post-production. The vial typically contained 2 mL of the active product.</p><p>Reference and retention vial. The vial was kept at Radboud Translational Medicine for reference purposes and unforeseen circumstances.</p><p>Customer vials. These vials were filled according to customer requests, 400 MBq/mL at EOS – activity concentration at a required time.</p><p>Spare vial. This vial was kept for emergency circumstances.</p><p>Stability vial. This vial was filled only when stability studies were being undertaken.</p><!><p>Method validation was performed for HPLC identification and determination of chemical and radiochemical purity of the [18F]DPA-714 injection in compliance with ICH Guideline Q2(R1) Validation of Analytical Procedures and European Pharmacopoeia Chromatographic Separation Techniques 07/2016:20246.</p><p>The HPLC module consisted of the following Shimadzu (Kyoto, Japan) components: SPD-20A(V) UV-Vis detector coupled to a SIL-20AHT autosampler and injection unit, LC-20AT tandem plunger, CMB-20A light system controller, CTO-20AC column thermostat with a FCV-14AH6 6-way column switching valve and a DGU020A5R degassing unit. The system was coupled in series to an Berthold POMO PET radio flow detector (Bad Wildbach, Germany) for radioactivity measurements. The entire setup was controlled by DataApex Clarity software (Prague, Czech Republic).</p><p>Analysis was performed on a Waters XTerra Shield RP-18 column (125 Å 5 µm; 250 × 4.6 mm) at a flow rate of 1 mL/min using isocratic elution with 0.1 M ammonium acetate/acetonitrile (50/50 v/v%). The spectrum was recorded at 254 nm. Prior to [18F]DPA-714 injection, system suitability was validated by injection of a blank sample (water, 20 µL), followed by the DPA-714 reference solution (10 µg/mL, 20 µL) and an additional blank injection. Analysis of the radioactive sample was performed with the designated QC part 1 vial ("GMP production of [18F]DPA-714" section).</p><p>Validation of chemical impurities (UV detector) The following criteria were assessed during the validation: specificity, accuracy, linearity and precision. Range, detection and quantification limits were also determined. Linearity was essential for performing Am calculations, using the slope and intercept of the line of best fit.</p><p>Validation of radiochemical impurities (radiodetector) The following criteria were assessed during the validation: specificity, accuracy, linearity and precision. Range, detection and quantification limits were also determined.</p><!><p>Method validation was performed for gas chromatography (GC) identification and control of residual solvents in compliance with European Pharmacopoeia guidelines of Gas Chromatography 01/2008:20228, Residual Solvents 07/2016:50400, Identification and Control of Residual Solvents 07/2017:20424 and Chromatographic Separation Techniques 07/2016:20246, as well as the ICH guideline Q2(R1) Validation of Analytical Procedures. Ethanol was used as an excipient to enhance stability of the product and was not considered a residual solvent. The content of ethanol was determined to verify its compliance with a maximum of volumetric concentration of 10% v/v%.</p><p>The GC module consisted of the following Shimadzu components: GC-2010 PLUS unit equipped with an AOC-20 autoinjectior/autosampler and a hydrogen flame ionisation detector. The system was controlled by DataApex Clarity software. Analysis was performed on a Restek Rxi-624Sil MS column (30 m × 0.25 mm × 1.4 µm) with a gradient heating rate (0-2 min 40∘C, 2-7 min 80∘C, 7-11 min 160∘C) at a linear velocity of 15 cm/s with helium as carrier gas. Prior to [18F]DPA-714 injection, system suitability was validated by injection of a blank sample (water, 0.1 µL), followed by the reference solution containing 1-propanol, acetonitrile and ethanol (0.1 µL) and additional blank injection. Analysis of the batch sample was performed with a designated QC part 2 vial ("GMP production of [18F]DPA-714" section).</p><p>Validation of GC analysis of residual solvents The following criteria were assessed during the validation: specificity (acetonitrile, ethanol), linearity of ethanol and precision.</p><!><p>Despite the fact that TEAB is a commonly used reagent in [18F]fluorinations, to the best of our knowledge, no limit tests have been described in literature or the European Pharmacopoeia. The limit of TEA content was therefore chosen based on: 1) toxicity data for TEA obtained through animal studies and 2) the limit described for tetrabutylammonium (TBA), a longer alkyl chain derivative of TEA, used for [18F]PSMA-1007 radiosynthesis at Radboud Translational Medicine B.V. The lowest LD50 TEA value for intravenous injection (i.v.) was described for mice at 36 mg/kg by Pindell and co-workers (Pindell et al. 1961). Similar values were reported in the Merck Index (The Merck Index Online). For TBA, mouse median lethal dose (LD50) i.v. was estimated at 10 mg/kg, as reported by Meyer and co-workers (Meyer et al. 1995). Setting the limit of TEA content in radiopharmaceutical injections at the TBA value (0.26 mg/mL) was considered a generous and secure estimation.</p><p>Initial investigation of suitability of thin layer chromatography (TLC) spot test for detection of TEA in the [18F]DPA-714 injection was positive. Distinguishing of the TEA spot from other matrix components (ethanol, 0.9% sodium chloride) was done unambiguously using a TLC plate and iodine staining. The results are qualitative only, with no attempt to quantify the TEA content.</p><p>The following solutions were spotted adjacently on a TLC plate: matrix solution (10% ethanol in 0.9% NaCl), TEA reference solution (0.26 mg/mL) and undiluted [18F]DPA-714 solution. The plate was then placed in an iodine-saturated chamber at 50∘C for 2 min. Immediately, a picture of the plate was taken and assessed visually. TEA formed a distinct brown spot, while no spot were visible for the matrix solution. For compliance, intensity of the [18F]DPA-714 solution spot should be less than of the TEA reference solution. The amount of TEA for all validation productions was below the limit of detection.</p><p>Validation of TLC spot test for TEA The following criteria were assessed during the validation: specificity and detection limit.</p><!><p>Method validation was performed for determination of bacterial endotoxins in [18F]DPA-714 injection according to the guidelines of the European Pharmacopoeia on Bacterial Endotoxins 2.6.14.</p><p>Analysis was performed on a portable Endosafe PTS reader from Charles River Laboratories (Wilmington, USA). The following tests were performed during the validation: inhibition/enhancement screening and selection and verification of the final dilution factor.</p><!><p>Tests, methods and acceptance criteria for the quality control of [18F]DPA-714</p><!><p>Quality control parameters that did no require method validation for this particular tracer were: pH, filter integrity test (FIT), radionuclidic identity, radionuclidic purity, sterility and environmental monitoring. The aforementioned were already validated for other radiopharmaceuticals produced on-site.</p><!><p>Measurements were performed on a QUANTOFIX Relax reflexion photometer from Macherey-Nagel (Düren, Germany) using their pH-Fix 2-9 indicator strips.</p><!><p>Integrity of the final sterilising filter was assessed by a bubble point test. This was performed automatically via an in-line FIT system of the Von Gahlen open vial dose divider used for the dispensing process.</p><!><p>Identity Radionuclidic identity was determined from half-life measurements, on the ISOMED 2010 dose calibrator with ISOMED software control, after dispensing.</p><p>Purity Radionuclidic purity was determined using an Osprey Multi-channel Analyser from Canberra Industries (Meriden, United States) using the resulting gamma ray spectrum.</p><!><p>Sterility of the final product was determined by direct inoculation in soybean casein digest and fluid thioglycollate media. This was performed by a third party, Eurofins Bactimm (Nijmegen, Netherlands).</p><!><p>Environmental monitoring was performed using contact and settle Trypticase Soy Agar plates with lecithin and polysorbate 80 isolator pack, acquired from Eurofins Bactimm, placed within 30 cm of critical operations in all GMP Grade A manipulation areas concerned - dispensing hot cells and airlocks. The plates were sent to Eurofins Bactimm for analysis.</p><!><p>For stability studies, three [18F]DPA-714 productions were performed. QC analysis involving visual inspection, HPLC and pH measurements were performed at release and 6 hours post-EOS. The sample was stored in a designated vial, in an inverted position, to mimic vial flipping during transport.</p><!><p>In optimisation studies, radiofluorination of tosylate DPA-714 was investigated with different PTC/base systems in acetonitrile. The subset included the most commonly used PTC, Kryptofix-222, in conjunction with potassium bases - carbonate and bicarbonate. TEAB was also tested. The procedure was very similar for each experiment. The reaction mixture was heated to 90∘C for a fixed time interval of 10 min. Table 1 summarises the results. Differences in elution volumes are attributed to the efficiency of [18F]fluoride release from the Sep-Pak QMA cartridge, measured as trapped activity with a dose calibrator.</p><p>A similar reaction profile was obtained for all three runs, with [18F]DPA-714 as the main reaction product, with unreacted [18F]fluoride as the only other radioimpurity. [18F]Fluorination proceeded most efficiently in the presence of TEAB. Almost quantitative RCY, 94%, was obtained, as estimated by (radio)HPLC. With optimised conditions in hand, preparation of the first automated production on a Trasis AIO followed thereafter.</p><!><p>Suitable semi-preparative HPLC purification conditions were also identified (described in full detail in "GMP production of [18F]DPA-714" section).</p><!><p>Radiosynthesis of [18F]DPA-714 using optimised conditions. OTs = tosylate</p><p>HPLC purification chromatogram of a [18F]DPA-714 production on the Trasis. Upper trace: Radioactivity. Note: the scale is not calibrated, so radioactivity values are not accurate. Lower trace: UV measurement, λ = 254 nm</p><p>Batch information: radioactivity parameters of the automated process</p><p>aDilution with sterile 0.9% NaCl</p><p>bFrom arrival of activity in AIO to measurement of activity in dispensing unit (pre-dilution)</p><p>Typical HPLC chromatogram (λ = 254 nm) obtained with a DPA-714 reference standard (10 µg/mL)</p><p>Typical HPLC radiochromatogram obtained with [18F]DPA-714 solution. RCP = 100%</p><!><p>The GC method described in "GC for residual solvent and ethanol content determination" section allowed for accurate and specific measurement of acetonitrile (residual solvent) and ethanol (excipient) in the presence of other components of the sample matrix of a sterile [18F]DPA-714 solution for injection. The symmetry factor of the solvents was, according to the acceptance criterion, smaller than 2: 0.995 for ethanol, 1.103 for acetonitrile and 1.181 for 1-propanol. Resolution between acetonitrile/1-propanol and ethanol was 10.479 and 10.765, respectively. Linearity and range for ethanol were evaluated by constructing a calibration curve. R was found to be 0.9994 and the concentration of ethanol linear within the range of 1-100 mg/mL. System repeatability for acetonitrile and ethanol was assessed through the relative standard deviation (RSD), which did not exceed the specified 15% - it was 0.55 for the ethanol/1-propanol and 0.90 for acetonitrile/1-propanol. The presented results prove suitability of the chosen GC method for the determination of acetonitrile and ethanol content in a [18F]DPA-714 solution for injection.</p><!><p>Dilutions of 10 and 100 complied with the specified spike recovery between 50 and 200%, however the former exhibited better accuracy. The results proved suitability of the test for the determination of bacterial endotoxins in a [18F]DPA-714 solution for injection.</p><!><p>LOD was qualitatively (visually) established as 0.01 mg/mL of TEA ions. No spots corresponding to any component of the matrix or the [18F]DPA-714 solution were visualised using the method described in "Thin layer chromatography spot test for determination of tetraethylammonium content" section, apart from TEA itself. The presented results prove suitability of the method for TEA determination in a [18F]DPA-714 solution for injection.</p><!><p>Quality control results at product release for three validation batches of [18F]DPA-714. Detailed explanation of acceptance specifications can be found in Table 3</p><p>aOnly tested during the validation phase. Not part of regular QC</p><p>Results of the stability study performed 6 hours post-EOS for three validation batches of [18F]DPA-714. Detailed explanation of acceptance specifications can be found in Table 3</p><!><p>Until the time of submission of this publication, three GMP-complaiant productions were performed for patient studies. RCYs were 55, 68 and 55%, with molar activities of 350, 232 and 227 GBq/µmol, respectively. The productions all conformed to the specification criteria at and post-release.</p><!><p>Following a short series of manual labelling experiments, highly promising labelling conditions were identified, eradicating the need for further investigation. Radiolabelling follows nucleophilic aliphatic substitution (SN2) on a tosylate leaving group (Figure 3). The reaction proceeded cleanly with no distinct radioimpurities. The unreacted [18F]fluoride was easily separated during semi-preparative HPLC purification. Interestingly, when a test run was performed without HPLC purification and relying solely on solid-phase extraction, co-eluting impurities were observed which significantly lowered RCP to <95%. Unsurprisingly, RCY obtained during manual labelling could not be reproduced on an automated system due to limited control over some variables (such as mixing and rinsing) and inherent losses of solutions/reaction mixture during numerous transfer stages.</p><p>Automation on the Trasis AIO module greatly reduces the level of complexity that characterises, still commonly used, non-cassette based synthesisers. The use of single-use GMP-grade cassettes, tubing and reagent vial not only reduces the risk of cross-contamination and reduces cleaning but very importantly, lowers bioburden of the process. The programmed sequence is robust and based on the presented results, the level of batch-to-batch variability is very low. This greatly simplifies training of production personnel. The sequence could naturally be automated to greater extent, however it was deemed less beneficial from the radiosynthetic aspect. For example, automatic HPLC collection, especially when a fixed duration is set, may jeopardise the entire production, if yields are lower than expected. Nevertheless, our proposed strategy could easily be translated to any other automated synthesiser.</p><p>Following successful validation procedure, including three fully automated productions and aseptic dispensing, the described GMP-compliant [18F]DPA-714 radiosynthesis delivered the tracer in, to the best of our knowledge, the highest RCY reported in literature and the first automated protocol designed on a Trasis AIO module. Several patients can be scanned with a single batch of the tracer, following short cyclotron irradiation. Our method offers a clear advantage over other published methods - significantly higher RCYs, 55-71% (n=6) vs 43-50%, both dc., obtained with the same precursor loading and with similar starting activities. The main difference between the approach of Kuhnast et al. is in the PTC. TEAB could therefore be a better enhancer of [18F]fluoride nucleophilicity than Kryptofix-222, perhaps due to better stabilisation of the precursor in the reaction. Fookes et al. reached very high [18F]DPA-714 of 55-85% (dc.) for a smaller precursor loading of 2 mg (Fookes et al. 2008). The difference could be ascribed to the purification and formulation protocol - following HPLC purification, the solvents were removed in vacuo, rather than by SPE, hence reducing the number of transfers and losses due to tracer sticking to tubings, etc. Given the scale of clinial production, this would be neither practical nor safe, with a low degree of batch-to-batch reproducibility.</p><p>Quality control tests performed on all the batches were fully compliant with in-house and European Pharmacopoeia specifications for the synthesis of 18F-based tracers for routine clinical use. The product remains intact (RCP of 100%) at high activity concentrations for a minimum of 6 hours, without the need of extra stabilisers, such as sodium ascorbate.</p><p>Radboudumc is currently involved with a clinical study employing [18F]DPA-714 as a neuroinflammation biomarker, produced using our GMP-compliant method, on a regular basis.</p><!><p>[18F]DPA-714 was synthesised in 55-71% RCY (n=6) in a nucleophilic aliphatic substitution on a tosylate leaving group with [18F]fluoride mediated by TEAB in acetonitrile. The method represents a significant improvement to the existing protocols, which use a similar or increased precursor loading. One of its assets, apart from robustness and reliability, is, undeniably, the use of an automated module with a disposable cassette, easily implementable in a modern radiopharmacy laboratory. The tracer can be produced in sufficient amounts, following short bombardment, to allow for shipment to more remote centres with no impact on tracer stability. The process and quality control comply with GMP standards, rendering it suitable for routine human use.</p><!><p>AllinOne</p><p>Molar activity</p><p>Alzheimer's disease</p><p>Colony-forming unit</p><p>Coefficient of variation</p><p>Decay-corrected</p><p>End of synthesis</p><p>Ethanol</p><p>2-Deoxy-2-[18F]fluoroglucose</p><p>Filter integrity test</p><p>Gigabecquerel</p><p>Gas chromatography</p><p>Good manufacturing practice</p><p>Hour</p><p>High performance liquid chromatography</p><p>International unit</p><p>Intravenous injection</p><p>Kryptofix-222</p><p>Median lethal dose</p><p>Limit of detection</p><p>Limit of quantification</p><p>Megabequerel</p><p>Multiple sclerosis</p><p>Non-decay-corrected</p><p>Outside diameter</p><p>Peripheral benzodiazepine receptor</p><p>Positron emission tomography</p><p>Polytetrafluoroethylene</p><p>Polyvinylidene fluoride</p><p>Radiochemical purity</p><p>Radiochemical yield</p><p>Phase-transfer catalyst</p><p>Quality control</p><p>Quaternary methylammonium</p><p>Correlation coefficient</p><p>Relative standard deviation</p><p>Solid-phase extraction</p><p>Tetrabutylammonium</p><p>Tetraethylammonium</p><p>Tetraethylammonium bicarbonate</p><p>Thin layer chromatography</p><p>Translocator protein</p><p>Publisher's Note</p><p>Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.</p>

PubMed Open Access

Comparison of endpoints relevant to toxicity assessments in 3 generations of CD-1 mice fed irradiated natural and purified ingredient diets with varying soy protein and isoflavone contents

Diet is an important variable in toxicology. There are mixed reports on the impact of soy components on energy utilization, fat deposition, and reproductive parameters. Three generations of CD-1 mice were fed irradiated natural ingredient diets with varying levels of soy (NIH-41, 5K96, or 5008/5001), purified irradiated AIN-93 diet, or the AIN-93 formulation modified with ethanol-washed soy protein concentrate (SPC) or SPC with isoflavones (SPC-IF). NIH-41 was the control for pairwise comparisons. Minimal differences were observed among natural ingredient diet groups. F0 males fed AIN-93, SPC, and SPC-IF diets had elevated glucose levels and lower insulin levels compared with the NIH-41 group. In both sexes of the F1 and F2 generations, the SPC and SPC-IF groups had lower body weight gains than the NIH-41 controls and the AIN-93 group had an increased percent body fat at postnatal day 21. AIN-93 F1 pups had higher baseline glucose than NIH-41 controls, but diet did not significantly affect breeding performance or responses to glucose or uterotrophic challenges. Reduced testes weight and sperm in the AIN-93 group may be related to low thiamine levels. Our observations underline the importance of careful selection, manufacturing procedures, and nutritional characterization of diets used in toxicological studies.

comparison_of_endpoints_relevant_to_toxicity_assessments_in_3_generations_of_cd-1_mice_fed_irradiate

Introduction<!>Animals<!>Diets<!>Chemical analyses of diets<!>Body weights, food consumption, and F1 pup allocation<!>Uterotrophic assay<!>Markers of sexual development and vaginal cytology<!>Glucose challenge<!>DXA<!>Necropsy, sperm analyses, and testicular histopathology<!>Statistics<!>Diet analyses and serum isoflavones<!>Body weights, food consumption, and metabolic efficiency<!>F0 and F1 mating and litter summaries<!>Markers of sexual development and estrous cycle<!>Uterotrophic assay<!>Glucose tolerance<!>Clinical chemistry<!>Organ weights<!>DXA scans<!>Sperm evaluation and testicular histopathology<!>Discussion

<p>Diet has long been recognized as an important variable that can influence the results of rodent toxicology studies (Brown and Setchell, 2001; Conner and Newberne, 1984; Cooper et al., 2006; Fullerton et al., 1992; Greenman and Fullerton, 1986; Greenman et al., 1987; Heindel and vom Saal, 2008; Kozul et al., 2008; Newberne and Conner, 1986; Rao and Knapka, 1998; Thigpen et al., 2013; Thigpen et al., 2004). In the case of studies of potential endocrine disruptors, and, in particular, agents with potential estrogenic activity, there has been much concern about the possible confounding effects of soy components, particularly estrogenic soy isoflavones (Boettger-Tong et al., 1998; Muhlhauser et al., 2009; Patisaul et al., 2012; Thigpen et al., 2004; Wang et al., 2005). This focus is due to the widespread use of varying levels of soy meal in natural ingredient diets typically used in rodent studies and fluctuations in isoflavone levels in diets of equivalent amounts of soy meal based on weather and other growing conditions (Brown and Setchell, 2001; Heindel and vom Saal, 2008; Thigpen et al., 2007; Thigpen et al., 1999). The interpretation of studies on the impact of diets containing soy is complicated by the presence of many non-isoflavone bioactive components in soy meal that can vary independently of the isoflavones (Kang et al., 2010; Velasquez and Bhathena, 2007). In addition, other dietary factors, including energy content, impact study results (Odum et al., 2001; Odum et al., 2004; Thigpen et al., 2002).</p><p>The present study was prompted by consideration of conducting long-term studies with hormonally active agents in a mouse model. We had previously used a soy-free diet in several studies with strong and weak estrogens in Sprague-Dawley rats, including multigenerational and chronic studies (Delclos et al., 2014; National Toxicology Program, 2008a, b, 2010a, b). Similarly, others have used soy-free diets in multigenerational reproductive studies in rats and mice (e.g., Kendig et al., 2012; Sprando et al., 2000); however, some literature suggested that the specific soy- and alfalfa-free diet that we had used might be problematic in studies in the CD-1 mouse (Ruhlen et al., 2008). This diet, designated 5K96, was originally designed to be nutritionally equivalent to the open-formula NIH-31 diet. The soy and alfalfa meal present in NIH-31 were replaced by casein in 5K96 to reduce considerably the estrogenic isoflavones and coumestrol found respectively in soy and alfalfa meals. Ruhlen et al. (2008) reported that CD-1 mice fed the 5K96 diet from one week prior to mating produced offspring that were obese as adults, with males, but not females, also showing increased serum leptin and impaired glucose regulation at PND 90 relative to mice fed soy meal-containing 5008 diet. Weights of the testes, epididymides, and seminal vesicles were decreased in the soy-free diet group, while prostate weight was increased. These effects were hypothesized to be due to the increase in fetal estrogen levels measured in mice fed 5K96 diet relative to those in mice fed the soy-containing diet. Cederroth et al. (2007) reported a similar finding with respect to body weight and fat deposition in CD-1 mice maintained from conception on a soy-free modified natural ingredient diet (Ziegler Phytoestrogen Reduced Rodent Diet I) similar to that used by Ruhlen et al. (2008). Cederroth and Nef (2009a) further reported that the beneficial effects of a high phytoestrogen diet on obesity were confined to postnatal exposure and beneficial effects on glucose tolerance were confined to intrauterine exposure, with the latter effects dependent on the intrauterine position of the pups. In addition, Cederroth et al. (2010) reported opposite effects on the male reproductive tract from Ruhlen et al. (2008), with the high soy diet resulting in reduced sperm counts, reduced seminal vesicle weight, and reduced fertility.</p><p>While the studies listed above were major motivators for the study described here, other literature also presents an unclear picture of the effects of dietary soy components on body weight, fat deposition, and reproductive function, as well as on the ability to detect the effects of weak estrogens in rodents. Multiple studies have suggested beneficial effects of phytoestrogens and soy protein on obesity (reviewed in Orgaard and Jensen, 2008; Velasquez and Bhathena, 2007). Soy has been implicated as a protective factor against the development of metabolic syndrome in rodents (Cederroth and Nef, 2009b), and both isoflavones and soy protein, due to its amino acid composition or peptides derived from the protein (Moriyama et al., 2004; Tachibana et al., 2010), have been implicated in this beneficial effect. Similar to the results mentioned above in CD-1 mice (Cederroth et al., 2007; Ruhlen et al., 2008), Badger et al. (2001) reported that Sprague-Dawley rats fed an AIN-93 based diet with a soy protein isolate replacing casein from conception had lower body weights than the animals fed the casein diet. Andreoli et al. (2015) also reported that feeding adult male Wistar rats a low phytoestrogen diet induced obesity and impaired glucose metabolism. Conversely, exposure of Wistar rats to a soy meal-containing diet during development has been reported to result in increased adult body weight in males, but not females (Cao et al., 2015), while developmental exposure of Sprague-Dawley rats to genistein was reported to result in obese adult females, but not males (Strakovsky et al., 2014). Lifetime exposure of Sprague-Dawley rats to 500 ppm, but not 5 or 100 ppm, genistein reduced body weight gain in the F1 and F2 generation females and in the F1 generation males (National Toxicology Program, 2008a). Developmental exposure of Wistar rats to soy protein-containing diet was reported to increase body weight of both male and female rats as adults compared to those fed a casein diet, but glucose intolerance was more pronounced in adult males (Jahan-Mihan et al., 2011). The effects of the dietary protein difference were largely attributed to the dams' diet during pregnancy (Jahan-Mihan et al., 2012). The data regarding soy diet effects on body weight and obesity are thus somewhat difficult to interpret and may depend on species and strain, the specific soy preparations or components examined, or other diet constituents or properties (e.g. pellet hardness or moisture content) that differ across studies.</p><p>The goals of the present study were to assess the suitability of a soy- and alfalfa-free modified natural ingredient diet for use in toxicology studies with CD-1 mice, to explore the utility of modified purified ingredient diets to alleviate the variability inherent in natural ingredient diets, and to better separate the effects that are due to soy isoflavones from those due to soy protein or other soy components. The results affirm marked diet-related differences in basic animal parameters, underline the difficulties in attributing diet effects to specific dietary components, and point to the importance of careful consideration of micronutrient variations when using irradiated purified diets.</p><!><p>Weanling CD-1 mice [Crl:CD1(ICR)], approximately three to four weeks old, were obtained from Charles River Laboratories (Wilmington, MA). The experiment was loaded in two equal sized animal groups spaced two weeks apart. Males born a week earlier than females were obtained to ensure that littermates were not mated. Upon receipt, animals were fed NIH-41 Irr diet and housed in polysulfone cages with hardwood chip bedding (P.J. Murphy, Montville, NJ), wire lids, and microisolator tops. Millipore-filtered tap water was provided in glass bottles with silicone stoppers and stainless steel sipper tubes. Feed and water were provided ad libitum. Animal rooms were maintained at 23° ± 3°C with a relative humidity of 50% ± 20%. Lights were on for 12 hours per day starting at 6 AM.</p><p>Within a week of arrival, animals were randomly allocated to one of six experimental diet groups by a stratified randomization procedure by sex based on body weight to give approximately equivalent mean body weights in each diet group. Males and females were randomly paired one-to-one within diet groups when the F0 females were approximately 11–12 weeks of age. Pairs were separated after 10 days. Except for the mating period and the pre-weaning period when dams and pups were co-housed, animals were individually housed to allow for more accurate evaluation of food consumption. Sentinel animals were maintained with NIH-41 Irr diet in each animal room and one animal per room was removed and evaluated for viral and bacterial pathogens at the midpoint and end of the study in accordance with the sentinel animal program at the National Center for Toxicological Research (NCTR). No pathogens were detected in any sentinel animal. All animal procedures were approved by the NCTR Animal Care and Use Committee.</p><!><p>There were six diet groups and 10 diets in the experiment. Four of the groups had separate growth (G) and maintenance (M) diet formulations, while the remaining two groups had one diet formulation fed through all stages of the experiment. The six diet groups were: 1) NIH-41, an irradiated formulation of NIH-31, 2) 5K96, 3) 5008(G)/5001(M), 4) AIN-93 G and M, 5) SPC G and M, which were modifications of AIN-93 in which casein was replaced with ethanol-washed soy protein concentrate (SPC, Arcon® SJ, product 066408, Archer Daniels Midland, Decatur, IL), and 6) SPC-IF G and M, the SPC diet formulations supplemented with isoflavones (IF, NovaSoy® 650, containing 73.08% isoflavones, product 152650, Archer Daniels Midland). The open formula purified diets AIN-93, SPC, and SPC-IF were obtained from Envigo (Harlan) Teklad Diets (Madison, WI). SPC was used in this study rather than soy protein isolate because a commercially available ethanol-washed, and thus isoflavone-free, soy protein isolate product could not be identified at the time that the study started. NIH-41 open formula, and 5K96 and 5008/5001, which are closed formula diets, were obtained from PMI Nutrition International (LabDiet/TestDiet), St. Louis, MO. Single lots of each diet were used in this study. All of the diets were provided in pelleted form and were irradiated by the manufacturers prior to shipment. Natural ingredient diets from Lab Diet were irradiated with Cobalt 60 (gamma irradiation) at an average dose of 20–25 kGy for approximately one hour (personal communication, Dr. Carrie Schultz, PMI Nutrition International). The other diets were irradiated with a Cobalt 60 source at 20–50 kGy for approximately six hours (personal communication, Dr. Barbara Mickelson, Envigo (Harlan) Teklad). Diets were confirmed to be free of microbiological contaminants. All were stored at 4°C until given to the animals and fresh diet was supplied weekly. Compositions and calculated energy contents of the natural ingredient diets are summarized in Table 1. Compositions and calculated energy contents of the three AIN-93 based diets are shown in Table 2. The AIN-93 based diets had higher energy densities than did the natural ingredient diets, including the natural ingredient growth diet, 5008. F0 animals were assigned as described above to one of the six maintenance diets within one week of arrival at NCTR. One week before mating, the F0 and F1 breeding animals were fed the growth diet in the four diet groups that had growth and maintenance diets. F1 animals were switched to maintenance formulations on PND 21.</p><!><p>Results of analyses of selected dietary components and contaminants assayed after arrival at NCTR are reported in Table 3. Fat was extracted with a Soxhlet apparatus and analyzed by gravimetry; total protein was quantified using a FP28 nitrogen/protein determinator (LECO Corporation, St. Joseph, MI); vitamins A and E were determined by high pressure liquid chromatography (HPLC) with fluorescence detection, and thiamine by the method of Gehring et al. (1995). Organochlorine pesticides, organophosphate pesticides, and polychlorinated biphenyls were determined following a modified FDA Pesticides Analytical Manual (PAM) method by gas chromatography with electron capture detector. Aflatoxins were measured by thermospray mass spectroscopy after derivitization with iodine (Holcomb et al., 1991). Fumonisins were measured with a RIDASCREEN®FAST fumonisin immunoassay kit (R-Biopharm Inc, Marshall, MI) according to the manufacturer's instructions. Heavy metals (arsenic, cadmium, selenium, and mercury) were analyzed by Applied Research and Development Laboratory, Inc. (Mount Vernon, IL), using standard EPA methods. Daidzein, genistein, and zearalenone levels in feed were determined by LC tandem mass spectrometry following an acid hydrolysis standard addition method. Background levels of bisphenol A (BPA) in feed and cage bedding were determined by LC-MS/MS following liquid-liquid extraction and cleanup using solid phase extraction.</p><!><p>Animals were weighed on arrival and weekly thereafter, except for the period when they were paired for mating. Litter weights were recorded by sex on PND 1 (day of birth = PND 0) and body weights of individual pups were recorded on PND 4, 7, and weekly thereafter. Body weights were also collected on the day on which vaginal opening and preputial separation were observed, on the day of glucose challenge, and at removal for necropsy. Food consumption was measured weekly except for the mating period. Food consumption data are approximate, since waste was not accounted for.</p><p>For each diet group, the first 20 litters born that met the minimal litter criterion of at least three male and three female pups were continued on the study beyond weaning. Litters with more than 10 pups were randomly culled to 10 on PND 4 with a distribution of five males and five females when possible. At PND 21, one pup per sex per litter was euthanized and the carcass scanned by dual x-ray absorptiometry (DXA). Two other pups per sex per litter were randomly selected to continue on the study after weaning at PND 21. One retained pup of each sex was mated with a non-littermate at 11 – 12 weeks of age to produce F2 litters (breeding or BR arm), and the other pup of each sex underwent glucose challenge and necropsy (GN arm) as described below. F2 litters were treated similarly to the F1 litters, except that there was no minimum litter size requirement and all pups retained after the culling of the litter on PND 4 were euthanized at PND 21.</p><p>In some cases, pups from any litters produced above the 20 above or that did not match the litter size criterion were used in the uterotrophic assay (see below) or as sentinels.</p><!><p>On PND 17, 180 F1 female pups (30 per diet group; maximum two per litter) were weaned and housed up to six per cage within diet groups so that no animals in any given cage were littermates. Starting on PND 17, daily subcutaneous injections of 0, 0.1, 1, 10, or 100 μg/kg body weight (bw)/day ethinyl estradiol (EE2; Sigma product # E4876-10G, lot # 028K1411) in corn oil (Sigma, product # C8267, lot # MKBG0329V) were administered for three consecutive days, with animals in a given cage all receiving the same EE2 dose. On the morning of PND 20, the mice were euthanized and body weights recorded. The uterus was removed as prescribed in OECD Test Guideline 440 (OECD, 2007) and the wet (i.e., uterus and luminal fluid) and blotted (i.e., uterus after expression and blotting of luminal fluid) uterine weights were recorded.</p><!><p>Anogenital distance was measured using an ocular micrometer (Delclos et al., 2014) at PND 21 on all F1 pups kept in the study after the PND 21 sacrifice. The anogenital distance measured at weaning has been reported to reflect the intrauterine androgen environment in mice (Vandenbergh and Huggett, 1995; Hotchkiss and Vandenbergh, 2005). Preputial separation and vaginal opening were monitored in F1 pups from PND 22 until occurrence. Vaginal smears were taken on the day of vaginal opening and daily thereafter for 20 consecutive days to determine the time of first estrus.</p><!><p>On PND 90 ± 3, one F1 animal per sex per litter (GN arm) had food removed from the cage at 6 AM. Starting at 1 PM, the animals were weighed and a zero-time blood sample collected by a tail vein prick made with a 3 mm lancet. Each animal received an intraperitoneal injection of 2.0 g/kg bw of glucose in 0.9% saline, and blood drops were collected from the tail vein at 15, 30, 45, 60, and 120 minutes after dosing for measurement of glucose levels using a glucometer (Accu-chek Aviva, Roche Diagnostics, Indianapolis, IN). After the final glucose measurement, animals were returned to their cages and continued on the study until necropsy on PND 105 ± 5.</p><!><p>On PND 21 (F1 and F2 males and females) and PND 105 (F0 male and F1 male and female breeders), animals were sacrificed and analyzed on the same day using a DXA PIXImus densitometer (Lunar GE Medical Systems, Madison, WI). The instrument was calibrated daily using the manufacturer's phantom mouse. The whole carcasses were scanned in the ventral position, with the limbs held splayed and the tail wrapped around the left side of the body. The head of each mouse was excluded from analysis by placing it outside the X-ray field and/or using the software's exclusion area.</p><!><p>F0 breeder males and females were sacrificed after the 10 day mating period or after weaning of litters, respectively. Animals were anesthetized with gaseous carbon dioxide and blood was collected by cardiac puncture into a serum separator tube. The blood was allowed to clot and centrifuged at 1000 × g for 10 minutes. Serum was utilized for the following clinical chemistry measurements: cholesterol, triglycerides, glucose, insulin, leptin, and adiponectin. Clinical chemistry analyses were conducted on an Alfa Wassermann ALERA (West Caldwell, New Jersey), as previously described (Delclos et al., 2014). Leptin concentration was measured using a mouse leptin ELISA kit (Millipore, St. Charles, MO). Plates were read on an ELx800 Universal Microplate Reader (Bio-Tek, Winooski, VT). Insulin was assayed using a "Coat-a-Count" radioimmunoassay (RIA) (Siemens, Los Angeles, CA) and adiponectin was assayed with a mouse adiponectin RIA kit (Millipore, St. Charles, MO). The insulin and adiponectin tubes were then counted on a Wizard2 (PerkinElmer, Shelton, CT) gamma counter. An aliquot of the male serum was also used for isoflavone analyses (Twaddle et al., 2002). For males, the terminal body weight was recorded and the carcass was scanned by DXA for body composition analysis. For females, the uterus was removed and placed in 10% ammonium sulfide for enumeration of implantation sites. The uteri of F0 dams that did not deliver litters or delivered litters with less than 3 pups per sex were also evaluated for implantation sites and non-viable fetuses.</p><p>For F1 animals in the GN arm, food, but not water, was removed at 6 AM on the morning of scheduled necropsy on PND 105 ± 5. After a mean fast of 7.6 ± 0.8 (SD) h, animals were anesthetized with gaseous CO2, blood was collected by cardiac puncture, and serum prepared as described above. The serum was used for the following assays: total protein, albumin, urea nitrogen, creatinine, alanine aminotransferase, gamma glutamyl transpeptidase, sorbitol dehydrogenase, aspartate aminotransferase, alkaline phosphatase, total bile acids, glucose, cholesterol, triglycerides, insulin, leptin, and adiponectin. The following organs were removed and weighed: adrenals, brain, prostate (ventral and dorsolateral after fixation), epididymides, kidneys, liver, ovaries, seminal vesicles with coagulating gland, testes, epididymal, ovarian and parametrial (combined), and retroperitoneal fat pads, thyroid (after fixation), and uterus. Paired organs were weighed together, except for testes and epididymides, which were weighed separately because of the sperm evaluations described below. A vaginal smear was taken at the time of necropsy to determine the stage of the estrous cycle; in all diet groups, except the NIH-41 group, there was one animal with a poor smear from which the estrous cycle stage could not be determined. The cycle stage at necropsy was not taken into account in the analyses of the clinical chemistry or organ weights. With the exception of the testes, which were processed for microscopic evaluation, all weighed tissues were fixed in 10% neutral buffered formalin and held as wet tissue. The right testes was removed, fixed in modified Davidson's fixative, embedded in infiltrating media (Formula R, Leica, Buffalo Grove, IL). Five μm sections were stained with Periodic Acid Schiff's stain for histological evaluation. The left testis was frozen on dry ice and stored at −80°C for later use in determining testicular spermatid head counts. The left epididymis was dissected away from the testis, weighed, and processed for determination of sperm motility and morphology and total sperm counts as previously described (Delclos et al., 2014).</p><!><p>NIH-41, the standard institutional diet at NCTR, was considered the reference control diet for pairwise comparisons among the six diet groups. In addition, exploratory analyses using AIN-93 as the reference control diet for comparisons were conducted within the three diet groups based on the AIN-93 diet (AIN-93, SPC, and SPC-IF). Results of the latter analyses are discussed in the text and summarized in Supplemental Table 1. Analyses were performed separately for females and males and for the BR and GN study arms. Adjustment for multiple comparisons of diet groups to the reference control diet was performed using Dunnett's method for analysis of variance (ANOVA), analysis of covariance (ANOCOVA), the log-rank test, logistic regression, and Poisson regression; pairwise comparisons of diets to the reference control diet were performed within contrasts. Holm's method of adjustment was used for Fisher's exact test. For analyses of repeated time point measures, within-group correlations were modeled using a heterogeneous first-order autoregressive (ARH(1)) correlation structure; for correlation between litter mates, correlation was modeled using a compound symmetric correlation structure. All statistical tests were conducted as two-sided at the 0.05 significance level.</p><p>Body weights, food consumption, and metabolic efficiency (g of body weight gained per g of food consumed in a given time period) were analyzed using a repeated measures mixed model ANOVA with terms for diet, week, and interaction. Food consumption outliers were defined as daily average greater than 80% of an animal's body weight at the end of the week; outliers for average weekly percent metabolic efficiency were identified using Grubb's test.</p><p>Mating success, defined as the proportion of females that littered to number mated, was analyzed using Fisher's exact test. Pup counts at birth (number alive, number of males, number of females, and number of unsexed) were analyzed using Poisson regression. Litter size was defined as the total number of pups (male, female, and unsexed) born alive. Sex proportions within litters were analyzed using logistic regression. Unsexed pups were assigned as male for the analysis reported, but assignment of the unsexed pups as female did not change the results. For litter weight data, analysis was performed using ANOVA. For mean animal weights, ANOCOVA was performed adjusted for litter size. Numbers of implants and resorptions were analyzed using ANOVA; the number of resorptions was calculated as the number of implants minus the sum of the number of pups born alive and dead. For anogenital distance (AGD), repeated measures mixed model ANOVA was performed for the mean of three measurements of AGD and for anogenital distance index (AGI), defined as the mean of AGD divided by the cube root of body weight.</p><p>Age and body weight at vaginal opening for females and at preputial separation for males were analyzed using ANOVA. Daily vaginal swabs were reported as estrus (E), proestrus (P), or diestrus (D), in addition to some reported as indeterminate (E/D, P/E, and D/P). For the analysis, E/D and P/E were categorized as E and D/P as D. Time to first estrus was defined as the first occurrence of an estrus smear during data collection up to 21 days following vaginal opening. Log rank analysis was performed for time to first estrus, with animals considered censored if dead or moribund, or if estrus was unobserved. The distribution of D, E, and P animals at necropsy for each diet compared to the reference control was analyzed using Fisher's Exact test with Monte Carlo estimation.</p><p>Glucose levels at baseline and at 15, 30, 45, 60, and 120 min after the injection of glucose were analyzed using a repeated measures mixed model ANOCOVA with terms for diet group, time, interaction, and covariate baseline glucose level. Area under the curve (AUC) was also calculated using the trapezoidal rule. One male animal was excluded due to an accidental early death.</p><p>For clinical chemistries, ANOVA was performed using a nonparametric method with midranks and an unstructured covariance. Measurements beyond detection limits were defined as one-half the lower limit or equal to the upper limit of quantification; samples with insufficient volume for analysis were considered missing.</p><p>ANOVA was performed for DXA scan of percent fat of tissue using arcsine square root transformation. In addition, ANOVA was performed for bone mineral density, bone mineral content, bone area, tissue area, and total tissue mass. Organ weights were analyzed using ANOVA for absolute weights. ANOCOVA was performed with covariate brain or terminal body weight in separate analyses. Ovaries with grossly observable cysts were excluded from the ovary weight analysis; in addition, the ovarian weights of one animal in the NIH-41 diet group and one animal in the AIN-93 diet group were excluded because the ovaries had been weighed together with the oviduct. Fisher's Exact test was performed to compare diet groups to NIH-41 control for the number of these exclusions. For seminiferous tubule degeneration of the testes in the study diet groups compared to the NIH-41 control, Fisher's Exact test was used for incidence, and Shirley's method, modified by Williams, was performed for severity scores; tests were conducted as one-sided and no adjustment was made for multiple comparisons.</p><!><p>The results of analyses of diets for the estrogenic substances, macronutrients, and vitamins in the single lots of each irradiated diet used in the study are shown in Table 3. The estrogenic compounds assayed (genistein, daidzein, coumestrol, zearalenone, and BPA) were specific for this study and the other analyses are standard analyses conducted at NCTR for all diets. As expected, the natural ingredient diets contained varying levels of soy-derived isoflavones in the order 5008/5001 > NIH-41> 5K96. Minimal soy isoflavones were reported in the AIN-93 G/M and SPC G/M diets, while SPC-IF had the highest measured isoflavone content of all diets used. Assessment of the serum of F0 breeder males confirmed that only animals fed 5001, SPC-IF, or NIH-41 had measurable levels of circulating isoflavones (Table 4). Coumestrol levels were highest in 5001 and NIH-41. Levels of other measured contaminants varied across diets, but were within tolerances established by the National Toxicology Program (National Toxicology Program, 2011) with the exception of malathion levels in 5K96, which exceeded the 0.5 ppm tolerance by approximately 20%. Low ppb levels of BPA were also found in all diets tested for the present study.</p><p>Vitamin A and B1 (thiamine) were higher in the natural ingredient diets than in the AIN-93-based diets, while vitamin E levels were higher in the AIN-93-based diets, particularly in the SPC and SPC-IF diets. All vitamins, except thiamine, exceeded NRC recommended levels in all diets (National Research Council, 1995). As shown in Table 3, several of the AIN-93-based diets (AIN-93 G and M, SPC M, SPC-IF M) had thiamine levels that were not detectable by the standard method used for diet thiamine analysis at this institution. After all of the study data had been collected and analyzed, samples of the purified diets were returned to the manufacturer for reanalysis, which confirmed low levels of thiamine in the diets (Table 3, footnote), although the length of time that had elapsed between diet preparation and analysis precludes the exact determination of the thiamine level at the time the diet was fed to the animals.</p><!><p>For F0 females, pairwise comparisons to NIH-41 indicated that the only significant difference for the females was at 6 weeks with the mean weight of the 5K96 group higher (6%) than that of the NIH-41 group (Supplemental Figure 1). There were no significant differences between any diet and NIH-41 at any week in pairwise comparisons for F0 males (Supplemental Figure 1).</p><p>Mean male pup PND 1 weights, adjusted for litter size, were 5% lower in the SPC-IF group relative to the reference NIH-41 group in both the F1 (p = 0.041) and F2 (p = 0.030) generations (Supplemental Table 2). In both generations, body weights of both sexes in the SPC and SPC-IF groups were significantly less than those in the NIH-41 group at PND 4, 7, 14, and 21 (Figure 1). Although no formal statistical comparison was conducted, these weight differences were similar for SPC and SPC-IF groups relative to the other natural ingredient diets. The mean body weights in animals fed SPC or SPC-IF ranged from 11 – 15% lower than those fed NIH-41 at PND 4 and 25 – 31% lower at PND 21 (Figure 1). For the AIN-93-related diet subset, there was no difference in PND 1 weights among the AIN-93, SPC, and SPC-IF groups. At PND 4 and after, both sexes in both the SPC and SPC-IF groups generally had significantly lower mean body weights than the AIN-93 group (Supplemental Table 1).</p><p>After weaning, the animals that were assigned to the BR and GN arms were analyzed separately, with similar results in both study arms. As with the pre-weaning animals, the body weights of the SPC and SPC-IF groups were lower than those of the NIH-41 reference animals during the post-weaning period of both sexes (Figure 2). The animals in the SPC-IF group had the lightest body weights, ranging from approximately 30% lighter than the NIH-41 group in the week after weaning to approximately 10 – 18% at week 12 (BR arm) or 13 (GN arm). For both pre- and post-wean animals, the analysis of the modified AIN-93 diet subsets showed similar significant differences for SPC and SPC-IF groups versus the AIN-93 control group, with the SPC and SPC-IF animals having lower body weights than the AIN-93 group (Supplemental Table 1). Males in the 5001 group had mean body weights that were 10 – 12% lighter than the reference NIH-41 group animals at weeks 10, 11, and 12 for the BR arm (Figure 2, top left) and at week 12 in the GN arm (Figure 2, bottom left). For the F1 BR arm, the AIN-93 group females had a higher body weight (11 – 15%) than NIH-41 animals during the post-weaning period (Figure 2, top right).</p><p>Food consumption was measured only for the F1 post-weaning animals and similar results were seen in both the BR and GN study arms. Food consumption was lower than that in the NIH-41 group for the AIN-93, SPC, and SPC-IF groups in both sexes (Figure 3). Consumption in the 5K96 group was higher (24 – 59%) in females in all weeks after week 4 in the BR arm (Figure 3, top right) and in weeks 5, 6, and 8 in the GN arm (Figure 3, bottom right). Males did not show a consistent difference in food consumption for 5K96 versus the NIH-41 control, with higher consumption in week 6 and lower consumption in week 9 in the GN arm (Figure 3, bottom left) and higher consumption in weeks 5 and 8 in the BR arm (Figure 3, top left). For the AIN-93-related diet subset, females and males in both the SPC and SPC-IF groups in both the GN and BR study arm had significantly lower food consumption than the AIN-93 group over the course of the study, although the numerical differences were less than those between these diets and NIH-41 (Figure 3 and Supplemental Table 1). Food consumption data were also plotted as calories consumed, since the diets differed in energy density, with similar results to those of grams consumed (Supplemental Figure 2).</p><p>Metabolic efficiency, while variable over the course of the study, was generally numerically higher for AIN-93, SPC, and SPC-IF relative to NIH-41, and was higher in all significant comparisons (Supplemental Figure 3). For the AIN-93-related diet subset, significant differences in metabolic efficiencies were observed in a few specific weeks that varied by sex and study arm (Supplemental Table 1).</p><!><p>Mating success of the F0 and F1 generations and basic litter parameters are shown in Supplemental Table 2. There were no statistically significant differences in the proportion of mated pairs producing litters, implant sites, resorptions, litter size, or pup sex ratio in either generation. The percentage of mated pairs producing litters ranged from 83 – 97% in all diet groups in both generations, except in the F1 AIN-93 group, where it was 70%; however, this decrease was not significantly different from the NIH-41 diet control.</p><!><p>A summary of the markers of pubertal development measured in both study arms is shown in Table 5. Statistically significant effects on the timing of vaginal opening were confined to the SPC diet group in the BR arm (1.9 day delay). The difference of 0.9 days in the GN study arm was in the same direction, but it was not statistically significant. The body weight at vaginal opening was significantly lower in the SPC and SPC-IF diet groups, consistent with the generally lighter body weights of the animals in these groups reported above. In both study arms, the time of first estrus was delayed relative to that in the NIH-41 diet group in the 5001, SPC, and SPC-IF groups. The time of first estrus was significantly less than that in the NIH-41 group in the AIN-93 group in the BR arm, but not in the GN arm, where the median time of first estrus was numerically later than that in the NIH-41 group. All smears in the breeding group for 21 days were read regardless of the time that the first estrus smear was detected and evaluated for cycle parameters; however, the majority of mice did not show regular cycles within this time period, consistent with previous reports that regular cycles in mice can take weeks to become established after vaginal opening and first vaginal estrus are observed (Nelson et al., 1990). The results of these analyses, which indicated limited statistical differences across diets, are not shown. The distributions of D/E/P animals at necropsy were as follows: NIH-41, 8/12/0; 5K96, 9/9/1; 5001, 13/4/2; AIN-93, 9/8/2; SPC, 12/5/2; SPC-IF, 10/7/2. These proportions did not differ significantly in comparisons of diets to the reference control.</p><p>AGD and AGDI were measured in all retained pups at PND 21. There were no AGD differences among diet groups that were not explained by the body weight reductions in the SPC and SPC-IF groups, as indicated by the lack of statistically significant differences for the AGDI (Supplemental Table 3).</p><!><p>Data from the uterotrophic assay were analyzed based on both absolute and body-weight adjusted uterine wet and blotted weights. Body weight-adjusted results are discussed because of the significant effect of the SPC and SPC-IF diets on body weight. Both wet and blotted uterine weight relative to body weight (Figure 4, top and bottom, respectively) showed an increase over the corn oil control at 1 μg EE2/kg bw/day and above. For the SPC-IF group, the wet uterine weight was significantly higher than the NIH-41 control at 0.1, 1, and 100 μg EE2/kg bw/day; similarly, the blotted uterine weights were significantly higher than the NIH-41 control at 0, 0.1, and 100 μg EE2/kg bw/day. Both wet and blotted uterine weights were lower in 5001 than that in the control diet group at 10 μg EE2/kg bw/day.</p><!><p>The baseline glucose level for the AIN-93 diet was 30% higher compared to the NIH-41 diet for males; however, there were no significant differences in post-baseline glucose levels, adjusted for baseline, between any of the diet groups compared to the NIH-41 diet for either females or males (Supplemental Figure 4). In addition, there were no differences across diets in the area-under-the-blood-glucose*time curve (Supplemental Table 4).</p><!><p>Selected clinical chemistry measurements are shown in Figure 5; terminal blood was collected from unfasted F0 animals, while F1 animals were fasted for approximately 7 hours.</p><p>In F0 males, the AIN-93, SPC, and SPC-IF diet groups had significantly elevated glucose levels (18, 19, and 14%, respectively) and depressed insulin levels (40, 28, and 29%, respectively) relative to those in the NIH-41 diet. Cholesterol was elevated by 16% in the SPC group and triglycerides depressed by 26% in the 5001 group relative to levels in the NIH-41 diet group. For females, the only statistically significant difference in the clinical chemistry parameters measured was a 12% increase in glucose in the SPC group relative to levels in the NIH-41 animals.</p><p>For F1 animals, significant differences between the NIH-41 control diet and the other diets for the analytes shown in Figure 5 were confined to males. In AIN-93 males, glucose was elevated 21%, while insulin, leptin, and triglycerides were decreased by 66%, 51%, and 36%, respectively. The insulin levels in SPC-IF were decreased by 60%, but glucose, leptin, and triglycerides were not significantly different from the control diet level. None of these parameters differed in the SPC diet versus the NIH-41 diet. In the 5001 group, cholesterol, triglycerides, leptin, and insulin were all significantly lower than NIH-41 levels by 18%, 28%, 68%, and 61%, respectively. Triglycerides were increased by 23% in the 5K96 diet compared to NIH-41 diet. There were few statistically significant changes in the other serum components measured (Supplemental Tables 5 and 6 for males and females, respectively). All three of the AIN-93-related diet groups had reduced alanine aminotransferase levels relative to the NIH-41 control in both sexes.</p><!><p>Selected organ weights recorded at necropsy are shown for males and females in Tables 6 and 7, respectively. Other recorded organ weights are listed in Supplemental Tables 7 (males) and 8 (females). The percentage differences mentioned in this paragraph are based on the least squares means generated by the statistical model. In males, brain, kidney, testes, and thyroid gland showed significant differences across diets. In females, significant diet effects were confined to the kidney and thyroid gland. Male brain weight, corrected for body weight, in the SPC-IF group was 12% higher than that in the NIH-41 group. Kidney weights were lower than those in the NIH-41 group in the AIN-93, SPC, and SPC-IF groups in both males (9%, 19%, and 20%, respectively) and females (11%, 21%, and 26%, respectively). In males, thyroid weights were significantly lower than those in the NIH-41 group in the AIN-93, SPC, and SPC-IF groups (29%, 33%, and 30%, respectively). In females, only the SPC-IF diet group had a significantly different thyroid to body weight ratio (35% lower). Mean thyroid weights were numerically lower than the NIH-41 control mean in the 5001, AIN-93, and SPC diet groups, but these differences were not statistically significant. Testes weight in the AIN-93 group was significantly reduced compared to the mean weight in the NIH-41group (21% less). Summaries of the significant results from the analysis of organ weight adjusted for body weight in the AIN-93-related diet subset are shown in Supplemental Table 1. In both sexes, kidney weights were lower in the SPC and SPC-IF diet groups compared to the AIN-93 group, while liver weight was lower in both sexes in the SPC group and in females in the SPC-IF group. Testes weights were significantly higher in the SPC and SPC-IF groups than in the AIN-93 group.</p><p>While there were no diet-related differences in ovary weights, grossly observable cysts led to the exclusion of many ovaries from weighing at the time of necropsy, and the percentage of animals affected was significantly different across diets (Table 7). The 5K96, AIN-93, and SPC diet groups had 9 of 20 (45%), 11 of 20 (55%), and 12 of 20 (60%), respectively, excluded for ovarian cysts compared to 1 of 20 (5%) in the NIH-41 group. All of these low isoflavone diet groups differed significantly from the NIH-41 control (p = 0.025, 0.005, and 0.002, for 5K96, AIN-93, and SPC, respectively). Histological examination of the ovaries confirmed an association between the gross cysts and bursal cysts (data not shown).</p><!><p>DXA scans were conducted on the carcasses of F0 males, F1 and F2 PND 21 weans, and F1 males and females from the BR arm. The results for percent body fat are shown in Table 8. PND 21 females and males in the AIN-93 diet group had significantly increased percent body fat in both the F1 (13–14% higher) and F2 (28–30% higher) generations compared to the NIH-41 diet group. There were no statistically significant differences between any of the diets and the NIH-41 control in percent fat in F0 adult males or F1 adult males or females, except for F1 adult females fed SPC-IF. The DXA software also calculated bone area, bone mineral content, bone mineral density, total area, and total tissue mass (Supplemental Tables 9 and 10 for males and females, respectively).</p><!><p>No significant differences in F1 caudal sperm motility or morphology were found between any diet group and the NIH-41 group (data not shown). Testicular spermatid head counts were significantly lower in the AIN-93 group than in the NIH-41 group, but caudal epididymal sperm counts did not differ among diet groups (Table 9). Analysis of the AIN-93-related diet subset indicated that both SPC and SPC-IF groups had significantly higher mean testicular spermatid head counts than the AIN-93 group (Supplemental Table 1). Data from the microscopic evaluation of the testes are summarized in Table 10 and Figure 6. In the AIN-93 diet group, there was a 65% incidence of seminiferous tubule degeneration characterized by degenerate and necrotic spermatogenic cells with reduced numbers of germinal epithelial cells and with multinucleated giant cells. In some areas, the germ cells were markedly depleted and only flattened Sertoli cells remained. Sixty-two percent (8 or 13) of the affected animals had a marked (grade 4, the highest grade) lesion severity. Low incidences (5 – 10%) of minimal severity (grade 1) lesions were seen in the NIH-41, 5K96, 5001, and SPC diet groups and were considered background.</p><!><p>The primary purpose of this study was to assess the suitability of a soy- and alfalfa-free modified natural ingredient diet for use in toxicology studies with CD-1 mice. The study was initiated in light of reports of both low and high soy diets interfering with evaluations of estrogenic agents or directly inducing adverse effects (Cederroth et al., 2007; Cederroth et al., 2010; Ruhlen et al., 2008). Uncertainty with regard to the soy components that are responsible for various soy diet-related effects prompted us to investigate the purified ingredient diet AIN-93 and modifications of that diet with casein, the sole protein source in that diet, replaced with an ethanol washed soy protein concentrate with or without added isoflavones. As indicated previously, the soy- and alfalfa-free modified natural ingredient diet 5K96 was tested because it has been the diet used for several rat studies of chemicals with estrogenic activity at this institution (Delclos et al., 2014; Delclos et al., 2009; National Toxicology Program, 2008a, b, 2010a, b) and because it was reported to induce a neonatal estrogenization-like metabolic syndrome in CD-1 mice (Ruhlen et al., 2008). NIH-41, an irradiated version of NIH-31, was tested because it is the formulation on which 5K96 is based, with soy and alfalfa meals replaced by casein, so it is nutritionally equivalent to NIH-41, but has a lower soy content than diets to which soy-free diets have been compared in previous mouse studies (e.g., Cederroth et al., 2007; Cederroth et al., 2008; Cederroth et al., 2010; Ruhlen et al., 2008). 5008 and 5001, a G/M combination of diets, were used because they are the comparison diets used by Ruhlen et al. (2008) and they have a variable, but relatively high, phytoestrogen content (Brown and Setchell, 2001; Thigpen et al., 2004). Under the conditions of this study, where feeding of a single lot of 5K96 was started shortly after weaning of the F0 generation and continued throughout the study, we did not observe significant differences between the 5K96 group and the NIH-41 group with regard to body weight, fat pad weights, serum leptin, adiponectin, insulin, triglycerides, total cholesterol or glucose, response to glucose or estrogen challenges, sperm parameters in F1 animals, or reproduction over two generations. Although no formal statistical comparisons were conducted comparing the 5K96 diet group to the 5008/5001 diet group, no marked differences were apparent in the means of these endpoints between these groups. It should be noted that the specific lots of 5008 and 5001 used in this study had lower chemically measured isoflavone levels (Table 3) than have been reported previously (Brown and Setchell, 2001; Thigpen et al., 2004). The level of phytoestrogens in the diet could influence the observed effects. For example, Lephart et al. (2004) reported lower post-weaning body weights and serum leptin and insulin levels in Long Evans rat pups exposed to a soy-containing diet with 600 ppm isoflavones from conception compared to rats exposed similarly to a low phytoestrogen diet. Our results show no major differences among the natural ingredient diets tested and suggest that all should be suitable for multigenerational toxicology studies in mice.</p><p>Advantages and disadvantages of using natural ingredient diets versus purified diets have been recognized and discussed (Barnard et al., 2009; Rao and Knapka, 1998). Among the potential advantages of purified diets are the minimization of contaminants and reduction of batch-to-batch variability. Conversely, natural ingredient diets have higher micronutrient levels than purified diets (Oller et al., 1989; Rao and Knapka, 1998) and marginal micronutrient levels may contribute to occasional reports of breeding difficulties on the purified diets (US Food and Drug Administration, 2007). There is a lack of data supporting the suitability of the purified ingredient diets for chronic studies (Duffy et al., 2002; Lewis et al., 2003). The AIN-93 G and M purified diets used in the present study, which contain casein as their sole protein source, are widely used phytoestrogen-free diets that can be modified to accommodate dietary component additions. The ethanol-washed soy protein concentrate used in the SPC diet, Arcon® SJ, consisted of soy protein with a relatively low content of isoflavones or other ethanol-extractable low-molecular-weight phytochemicals. A soy isoflavones concentrate, NovaSoy® 650, was added to the SPC-IF diet to supply isoflavones. The most marked and consistent effect observed in the study was the significantly lower body weights of animals of both sexes after birth in the F1 and F2 generations in both the SPC and SPC-IF diet groups, suggesting that the soy protein was the prime driver of this effect. F0 animals fed these diets only during the post-weaning period did not show significant differences in body weights from the natural ingredient diets or from the AIN-93 group, indicating that the diet consumed by the dam during gestation and lactation influenced the pup body weights. The importance of the diet consumed by the dam on pup body weight throughout life has been reported in other studies of soy- versus casein-containing diets, although inconsistent effects have been found, with the soy-consuming F1 CD-1 mice having lower body weights (Cederroth et al., 2007; Ruhlen et al., 2008) and Wistar rats having higher body weights (Jahan-Mihan et al., 2011; Jahan-Mihan et al., 2012). A study in Sprague-Dawley rats reported no differences in adult body weight between rats fed purified diets with either casein or soy protein isolate as the protein sources (Badger et al., 2001).</p><p>There were no notable organ weight differences seen in the F1 animals among the natural ingredient diets, but in males, the kidney and thyroid weights, adjusted for body weight, were lower than the NIH-41 organ weights in all three AIN-93-related diets. The same was true for kidneys in females, but in females, only the thyroid weight of the SPC-IF diet group was lower than that in the NIH-41 group. Within the AIN-93-related diet subset, the SPC and SPC-IF females had lower kidney and liver weights than the AIN-93 females and the SPC-IF females also had a significantly lower retroperitoneal fat pad weight. For males, the SPC and SPC-IF groups had lower kidney weights than the AIN-93 group, liver weight was decreased in the SPC group and adrenal and brain weights were lower in the SPC-IF group. Testis weight differences are discussed below. Soy protein has been reported to have beneficial effects on kidney function (Ogborn et al., 1998; Ogborn et al., 2010), and soy protein has been reported to lower liver weight, as well as reduce serum and liver cholesterol and triglycerides (Ascencio et al., 2004; Nagata et al., 1981; Torre-Villalvazo et al., 2008; Wanezaki et al., 2015), although the latter effects were not observed in the present study. While there were no significant diet effects on ovary weights, there was a higher incidence of gross ovarian cysts in the three low isoflavone diets compared to the NIH-41 diet. Patisaul et al. (2014) recently reported an increased incidence of cystic follicles in Wistar rats fed a soy meal containing diet.</p><p>An unexpected aspect of this study that limits the ability to compare it with other published studies using AIN-93 diets and modifications thereof was the low thiamine level reported in the AIN-93, SPC, and SPC-IF diets (Table 3). Despite these low thiamine levels, there was no apparent effect on growth or general health of the animals in any of the diet groups, and gestation and lactation were supported by all diets. In fact, throughout the course of the study, none of the classical signs of thiamine deficiency in rodents, including poor growth or neurological symptoms, were noted in any of the diet groups reported to have low thiamine. The U.S. National Research Council (NRC) has estimated a thiamine requirement for mice of 5 mg thiamine hydrochloride/kg diet (National Research Council, 1995), although the similar estimate for rats (4 mg thiamine hydrochloride/kg diet) has been questioned in that approximately 5 to 7-fold lower levels were able to maintain adequate growth in rats (Rains et al., 1997; Shibata and Fukuwatari, 2013). In both rodents and humans, the thiamine requirement has long been known to vary according to the carbohydrate content of the diet, with higher thiamine levels required in high carbohydrate diets (Elmadfa et al., 2001; Reinhold et al., 1944; Wainio, 1942). Generally, natural ingredient diets have higher levels of thiamine than the NRC estimated requirement, while purified diets have been reported to have lower and variable levels (Oller et al., 1989; Rao and Knapka, 1998). Vitamin A levels were also lower in the AIN-93-related diets, although all were higher than the NRC estimated mouse requirement of 0.72 mg/kg (2.4 IU/g) diet. Vitamin E levels were higher in the AIN-93-related diets, particularly in the SPC and SPC-IF diets. As discussed below, the low levels of thiamine in the diet may have impacted significantly the results of the current study.</p><p>The reason for the low thiamine levels in the AIN-93-related diets in the present study is not clear, although irradiation seems to have been a factor (Table 3). It is known that heat and irradiation can reduce thiamine and vitamin A levels, and this is taken into account when formulating diets that will be autoclaved or irradiated. Irradiation has been reported to have relatively modest effects in a natural ingredient diet matrix (Caulfield et al., 2008), while thiamine was reported to be unstable on storage at room temperature in the purified AIN-76 diet relative to a natural ingredient diet (Fullerton et al., 1982). In the current study, the vitamin mix added to the AIN-93-based diets should have been sufficient to account for loss due to irradiation, and the diets were stored refrigerated to maintain nutrient levels. However, specifics of pellet preparation were likely a contributing factor as, in an attempt to normalize hardness across all purified diets, some had higher moisture content than others. Factors that contribute to thiamine loss in irradiated purified diets include a combination of moisture content, irradiation dose, and form of thiamine (personal communication, Dr. B. Mickelson, Envigo (Harlan) Teklad). The longer irradiation time and higher dose range used by the manufacturer of the purified diets (see Materials and Methods) may have contributed to the lower thiamine levels in these diets. In any case, although the thiamine levels fed to the animals in the study were below 5 ppm, they were clearly adequate in all cases to support growth, gestation, and lactation.</p><p>Endpoints to evaluate spermatogenesis (testes weight, testicular spermatid head count, caudal epididymal sperm count, motility, and morphology, and histopathology) were included based on the reports of soy/phytoestrogen detrimental effects on spermatogenesis in CD-1 mice (Cederroth et al., 2010). Although our data did not support such a link, abnormal spermatogenesis was noted in animals fed AIN-93, and its low thiamine may have contributed to this observed effect. Testicular degeneration had been reported decades ago in rats subjected to severe thiamine deficiency (Morris and Dubnik, 1947), and spermatogenesis defects have been reported in thiamine transporter knockout mice in the absence of the more severe thiamine deficiency manifestations (Fleming et al., 2003; Oishi et al., 2004). An interesting aspect of the present study is that defects in spermatogenesis were observed only in the AIN-93 diet group, which had non-detectable levels of thiamine reported in both the G and M formulations, while the SPC and SPC-IF groups, which had detectable thiamine levels in the M, but not G, formulation, did not show effects on testicular weights, testicular spermatid head counts, or histology. This could indicate that an early developmental subclinical thiamine deficiency may be critical to the noted effect on spermatogenesis and/or that there is an interaction between the subclinical thiamine deficiency and the lack of soy protein. Furthermore, although consistent effects on testes were noted, there was no observed deficit in epididymal sperm in the AIN-93 diet group. This could indicate a delayed effect on testicular spermatogenesis, similar to the delayed effect on spermatogenesis reported in aromatase knockout mice (Robertson, 2002). Further targeted experiments in mice fed low thiamine diets would be needed to address these questions more definitively. The issue of thiamine deficiency affecting reproductive toxicity assessments in rodents is unlikely to be an issue when natural ingredient diets are used, due to their high thiamine content; however, this may be a potential issue for rodent studies when purified diets, and particularly sterilized purified diets, are used. It must also be noted that, while the low measured thiamine levels are possibly involved in the observed effect, other micronutrients that were lower in the AIN-93-related diets may have also played a role. For example, as noted previously, vitamin A levels were lower in these diets than in the natural ingredient diets and vitamin A also plays a critical role in the development of the testis and the maintenance of spermatogenesis (Chihara et al., 2013; Hogarth and Griswold, 2010).</p><p>A significant question is whether or not subclinical nutritional deficiencies might contribute to defects in spermatogenesis or to an individual's susceptibility to reproductive toxicants. This issue seems to have received relatively little attention, although it has been reported that the effect of bisphenol A on the number and motility of sperm in mice is enhanced under conditions of vitamin A deficiency (Aikawa et al., 2004; Nakahashi et al., 2001). Thiamine deficiencies severe enough to produce beriberi are rare in the modern developed world and severe thiamine deficiency associated with severe neurological effects (Wernicke's encephalopathy) is largely confined to alcoholic adults or other specific conditions or medical procedures (Francini-Pesenti et al., 2009; Frank, 2015; Lallas and Desai, 2014; Stroh et al., 2014; World Health Organization, 1999). Low thiamine levels in diabetics have been noted, and it has been hypothesized that consumption of high energy high carbohydrate diets may lead to low, although subclinical, thiamine levels (Kerns et al., 2015; Lonsdale, 2015; Page et al., 2011; Rosner et al., 2014). In any case, the data reported here and by Oishi et al. (2004) in mouse models suggest that defects in spermatogenesis produced by low thiamine appear before other overt signs of deficiency. The mechanism and relevance of this observation to humans may be worthy of further study.</p><p>As noted in the Introduction, multiple studies and reviews over the years have pointed out the potential influence of diet on the outcomes of animal toxicology studies. Despite this, many investigators continue to provide little information on the diet utilized in their studies and likely select diets based largely on familiarity or convenience. The results of this study underline the importance of diet selection and formulation, and thorough nutrient characterization and reporting of the diets tested to enhance interpretability of study results and provide sufficient information to investigators attempting to replicate or build on others' results.</p>

PubMed Author Manuscript

IsoQuant: A Software Tool for SILAC-Based Mass Spectrometry Quantitation

Accurate protein identification and quantitation are critical when interpreting the biological relevance of large-scale shotgun proteomics datasets. Although significant technical advances in peptide and protein identification have been made, accurate quantitation of high throughput datasets remains a key challenge in mass spectrometry data analysis and is a labor intensive process for many proteomics laboratories. Here, we report a new SILAC-based proteomics quantitation software tool, named IsoQuant, which is used to process high mass accuracy mass spectrometry data. IsoQuant offers a convenient quantitation framework to calculate peptide/protein relative abundance ratios. At the same time, it also includes a visualization platform that permits users to validate the quality of SILAC peptide and protein ratios. The program is written in the C# programming language under the Microsoft .NET framework version 4.0 and has been tested to be compatible with both 32-bit and 64-bit Windows 7. It is freely available to non-commercial users at http://www.proteomeumb.org/MZw.html.

isoquant:_a_software_tool_for_silac-based_mass_spectrometry_quantitation

INTRODUCTION<!>SILAC labeling and hippocampal slice culture<!>Preparation of cell lysates and protein digestion<!>Liquid chromatography-tandem mass spectrometry<!>Database searching and data processing<!>1) Overview of IsoQuant SILAC-based quantitation<!>2) IsoQuant minimizes under-sampling problem when a data-dependent method is used to acquire MS2 spectra<!>3) Summary of IsoQuant Peak Selection Algorithm<!>4) Quantitative accuracy of IsoQuant using sPRG2009 SILAC peptide standards and complex biological sample<!>5) Validation of IsoQuant quantitation accuracy using SILAC-labeled rat hippocampal slices<!>CONCLUSIONS

<p>Mass spectrometry-based proteomics approaches can provide valuable qualitative and quantitative information regarding not only protein identification, but also protein-protein interaction, and the dynamics of protein expression levels 1. Accurate protein quantitation is required in order to make meaningful biological inferences from mass spectrometry data. In proteomics experiments, qualitative data is obtained by identifying candidate proteins with an acceptable false positive identification rate. Several commercially available database search programs, such as SEQUEST 2 and Mascot 3, and open-source database search programs including X!Tandem 4, Protein Prospector 5, ProbID 6, OMSSA 7, ProSight 8 and InsPecT 9 are available for this purpose. Although many software programs have been developed for quantitative proteomic analysis, certain challenges pertaining to reliable and accurate quantitation still remain 10.</p><p>Several methodologies are currently available for mass spectrometry-based protein quantitation including stable isotope labeling (chemical, enzymatic, or metabolic) and label-free strategies. SILAC (stable isotope labeling by amino acids in cell culture) 1112 is a metabolic labeling method whereby stable isotope-labeled amino acids (typically Lys and Arg) are incorporated into proteins as they are synthesized, thereby resulting in robust peptide and protein quantitation ratios. SILAC has significant advantages as compared to chemical labeling strategies, such as iTRAQ and TMT, since the labeling event occurs at the beginning of the sample preparation procedure. Consequently, less confounding factors influence the calculation of peptide and protein ratios. Additionally, SILAC yields more robust quantitation ratios than label-free strategies since the lack of reproducibility in certain label-free strategies can negatively impact the generation of consistent ratio reports. In general, SILAC based proteomic analyses are limited to only those cells or animals that can metabolically incorporate stable isotope labeled amino acids. Although it has not been widely adopted by most clinical proteomic analyses, recently the SILAC method has been modified and extended to the analysis of clinical biopsy samples that have not traditionally been amenable to this type of approach 13 14.</p><p>Several methods have been developed for the mass spectrometry-based determination of SILAC ratios. One method is based on peak area integration of Extracted Ion Chromatograms (XIC) using software programs such as Xpress 15, ASAPRatio 16 and MSQuant 17. In general, XIC peak area is calculated by integrating peak intensity over retention time for a specified m/z value. However, it is often difficult to objectively determine the boundaries of a peak corresponding to a specific peptide, especially when the XIC peak or elution profile of a peptide is not well defined. Another method for determining SILAC-labeled peptide ratios is based on ratios generated from MS1 scans surrounding the MS/MS spectrum of an identified peptide. The programs Census 18 and RelEx 19 use a linear regression approach to calculate quantitation ratios. MaxQuant 2021 determines SILAC pairs based on peptide features and subsequently uses the database search results from Andromeda 22, a novel peptide search engine using a probabilistic scoring model, to calculate ratios. A newly developed software program, UNiquant 23, provides a quantitation method based on overall peak intensity ratio and has been shown to significantly increase the number of quantified peptides in the analysis of quantitative proteomics data using stable isotope labeling.</p><p>A critical issue commonly associated with the processing of quantitative SILAC-based proteomic data is the validation of the accuracy of the large number of SILAC peptide ratios generated from a shotgun proteomics experiment. Since every software program uses a different algorithm and computational method to calculate SILAC peptide ratios, we and others have observed a rather large discrepancy among the SILAC peptide ratios reported by different programs for the same peptides, especially for low abundant peptides. Therefore, the manual validation of SILAC peptide ratios at both the protein and peptide levels is often required for each analysis. Currently, the Qurate 24 graphical tool can be used to manually validate isotopically labeled peptide quantitation events. The Rover 25 tool can be used to validate the quantitation reports from several different software packages such as MaxQuant and Census.</p><p>In this paper we describe the framework and quantitation accuracy of IsoQuant, a software tool we developed for SILAC-based mass spectrometry quantitation. IsoQuant uses unique peak selection and absolute peak intensity integration algorithms combined with high mass accuracy and dynamic chromatographic retention time filters in order to maximize the accuracy of peptide and protein quantitation. To improve the efficiency of manual validation, we have included a graphic user interface in IsoQuant, which allows end users to visualize original MS1 spectra and examine how each individual SILAC peptide ratio was calculated in order to improve the confidence of protein quantitation. IsoQuant is freely available to non-commercial users at http://www.proteomeumb.org/MZw.html.</p><!><p>Hippocampal slice cultures were prepared and maintained as described previously 26. Briefly, hippocampi were dissected from 5- to 7-day-old rat pups. Slices of 400 µm thickness were cut with a tissue chopper and attached to glass coverslips in a chicken plasma clot. The coverslips and slices were placed in individual sealed test tubes containing semi-synthetic medium and maintained on a roller drum in an incubator for 14 days. On day 15, the cultures were transferred to DMEM-based SILAC medium (Pierce/Thermo Fisher Scientific) supplemented with D4 l-Lysine-HCl (100 µg/ml) (purity 96–98%, Cambridge Isotope Laboratories), 13C6 l-Arginine-HCl (100 µg/ml) (purity 99%, Cambridge Isotope Laboratories), 10% dialyzed fetal bovine serum (Invitrogen) and 1% penicillin/streptomycin. Slices were harvested on the 14th day of being cultured in SILAC DMEM.</p><!><p>SILAC-labeled cultured hippocampal slices were homogenized in a buffer containing 8 M urea, 75 mM NaCl, 50 mM Tris-HCl, pH 8.2, protease inhibitor cocktail (Roche), 1 mM NaF, 1 mM ß-glycerophosphate, 1 mM sodium orthovanadate, 10 mM sodium pyrophosphate, and 1 mM PMSF using a microtube pellet pestle rod and motor (Kontes). Insoluble material was pelleted by centrifugation at 10,000 × g for 10 min at 4 °C. The concentration of the soluble proteins was determined by the BCA protein assay (Pierce/Thermo Fisher Scientific). Reducing Laemmli sample buffer was added to three aliquots of 25 µg protein prior to a 5 min incubation at 95 °C and protein separation via SDS-PAGE. Gel bands were visualized by Coomassie blue staining. Each gel lane was divided into 15 regions and the corresponding gel regions from each lane were combined and subjected to in-gel tryptic digestion as described previously 27.</p><p>The complex cell lysate sample used to prepare standard mixtures and evaluate the reproducibility of peptide quantitation by IsoQuant was derived from SKBR3 cells cultured in SILAC DMEM supplemented with 10% dialyzed fetal bovine serum (Invitrogen) and 1% penicillin/streptomycin. Five 15-cm dishes of SKBR3 cells were cultured in supplemented SILAC DMEM to which 12C6 l-Lysine-HCl (100 µg/ml) (Cambridge Isotope Laboratories) and 12C6 l-Arginine-HCl (100 µg/ml) (Cambridge Isotope Laboratories) were added; this medium is referred to as SILAC "light": Lys0, Arg0. Five 15-cm dishes of SKBR3 cells were cultured in supplemented SILAC DMEM to which 13C6 l-Lysine-HCl (100 µg/ml) (purity 99%, Cambridge Isotope Laboratories) and 13C6 l-Arginine-HCl (100 µg/ml) (purity 99%, Cambridge Isotope Laboratories) were added; this medium is referred to as SILAC "heavy": Lys6, Arg6. Cells were lysed by sonication in the same buffer as described for the hippocampal slices. The concentration of the soluble proteins was determined by the BCA protein assay and 1 mg of total proteins from the SILAC "light" and "heavy" labeled proteins were combined at the following ratios: 1:1, 1:2, 1:5, 1:10, 1:25, 1:50 and 1:100. The standard mixtures in triplicates were then fractionated by SDS-PAGE. Following SDS-PAGE, protein bands were excised and subjected to in-gel trypsin digestion and LC-MS/MS analysis as described below.</p><!><p>Peptides were separated by nanoscale reverse-phase liquid chromatography using an Xtreme Simple nanoLC system (CVC/Micro-Tech). The analytical column was prepared by packing 1.7µm 200Å C18 resin (Prospereon Life Sciences) into a laser-pulled fused silica capillary (75 µm inner diameter, 10.5 cm length, 10µm tip; Sutter Instruments) using a pressure injection cell (Next Advance). Peptides were injected into the sample loop using an Endurance autosampler (Spark Holland, Brick, NJ) and were loaded onto the column with 95% solvent A (0.5% acetic acid in water). A 120 min LC gradient method from 5 – 35% B (80% acetonitrile, 0.5% acetic acid) with a post-split flow rate of 0.3 µl/min was used to elute the peptides into the mass spectrometer. The LTQ-Orbitrap mass spectrometer (Thermo Electron) was equipped with a nanospray ionization source. The spray voltage was 1.5 kV and the heated capillary temperature was 180 °C. MS1 data were acquired in the profile mode in the Orbitrap with a resolution of 60,000 at 400 m/z and the top ten most intense ions in each MS1 scan were selected for collision induced dissociation in the linear ion trap. Dynamic exclusion was enabled with repeat count 1, repeat duration 30 sec, and exclusion duration 180 sec. Other mass spectrometry data generation parameters were as follows: collision energy 35%, ion selection threshold for MS/MS 500 counts, isolation width 3 m/z, default charge state 3, and charge state screening enabled.</p><!><p>MS/MS spectra were searched against a UniProtKB human protein database (version Oct 5, 2010; 20,259 reviewed sequences; 75,498 nonreviewed sequences) using Bioworks 3.3.1 SP1 with the SEQUEST algorithm. Search parameters included 20 ppm peptide mass tolerance, 1.0 Da fragment tolerance, static Cys + 57.02510 (carbamidomethylation) modification and the following differential modifications: Met + 15.99492 (oxidation), Lys + 6.02013 (13C6) and Arg + 6.02013 (13C6). Fully tryptic peptides with up to two missed cleavages and charge-state-dependent cross-correlation scores ≥ 2.5, 3.0, and 3.5 for 2+, 3+, and 4+ peptides, respectively, were considered as positive identifications for further quantitative analyses. The same raw data were processed by Proteome Discoverer (version 1.2, Thermo Scientific) using SEQUEST. The protein database, database search parameters and peptide filters were the same as indicated for the workflow using Bioworks.</p><!><p>In this report, we have developed a new SILAC-based proteomics quantitation method, named IsoQuant. Figure 1 illustrates the framework of IsoQuant for calculating SILAC peptide quantitation ratios. Briefly, IsoQuant consists of three major components: .raw file extraction, Peptide/Protein Quantitation, and Data Validation and Visualization. Figure 2A is a screen shot of the three major IsoQuant user modules. To automate and streamline the data processing, IsoQuant takes original .raw files generated from an Orbitrap mass spectrometer as input and extracts MS1 spectra using the Xcalibur Developer's Kit (XDK, Thermo Electron, San Jose CA). This .raw file conversion program selects all of the MS1 scans in the .raw binary file and exports them to a single ASCII format text file. The built-in IsoQuant .raw file conversion program is convenient for users since it does not require any third party software tools for mass spectrometry data conversion. Next, a summary report of peptide quantitation ratios is generated. Since IsoQuant performs its peptide quantitation after the database search, the filtering and selection of peptides based on their false discovery rates can be readily defined by the user. The third step is protein identification whereby IsoQuant checks the database search results and groups proteins with homologous peptides into protein clusters. Lastly, IsoQuant calculates protein quantitation ratios and compiles protein quantitation reports based on the peptide quantitation ratios in each protein cluster. The graphical user interface of IsoQuant's peptide and protein quantitation module is presented in Figure 2B.</p><!><p>Mass spectrometer data acquisition methods with dynamic exclusion enabled are commonly used in most proteomic experiments. Mass spectrometers operated in a data-dependent data acquisition mode have a bias toward selecting the most abundant peaks for MSn fragmentation. As a result, the use of a dynamic exclusion method often reduces the number of times a given peptide is selected for MSn fragmentation, thereby increasing the likelihood that lesser abundant peaks will be selected for fragmentation.</p><p>In this study, we found that data-dependent data acquisition methods often result in the lack of fragmentation of both the heavy- and light-labeled versions of each SILAC peptide pair in cases where the intensity of one of the peptides is below the threshold used to trigger MS/MS fragmentation. To compensate for the loss of potentially valuable quantitation data in cases such as these, many current software packages, such as RelEx 19 and ProRata 28, only require either the heavy or light peptide to be selected for MS/MS fragmentation and identified by database search in order for a SILAC peptide ratio to be calculated. Once the peptide is correctly identified, corresponding SILAC-labeled heavy and light peptide MS1 isotopic envelopes will be extracted. IsoQuant has adopted this important feature and can automatically extract the XICs of heavy and light peptides when one of these variants is identified. We found this feature has significantly increased the number of peptide pairs that can be quantified by IsoQuant.</p><p>Figure 3 illustrates one of the limitations of data-dependent acquisition instrument methods and indicates that MS1 peaks selected for MS/MS fragmentation do not always have clear isotope distributions for accurate SILAC-pair quantitation. As shown in Figure 3A, an MS1 peak corresponding to 572.77 m/z at retention time (RT) 32.70 min and RT 32.75 min (indicated by *) was selected for MS/MS fragmentation. The MS1 scan of RT 32.70 min has a distinct isotopic peak envelope at 572.77 m/z (heavy SILAC peptide), whereas 569.76 m/z (light SILAC peptide) has an incomplete isotopic peak envelope (Figure 3B). The use of dynamic exclusion in this instrument method precluded the selection of the 572.77 m/z (heavy SILAC peptide) and 569.76 m/z (light SILAC peptide) peaks for MS/MS fragmentation after RT 32.75 min, as indicated in Figure 3A. On the other hand, at RT 32.84 min, which coincides with the apex of the peak in the extracted ion chromatogram of 572.77 m/z, complete isotopic peak envelopes were obtained for both the light and heavy SILAC peptides (Figure 3C). However, these two peaks at RT 32.84 min were never selected for MS/MS fragmentation since they had already been added to the dynamic exclusion list. To overcome this issue, IsoQuant uses both high mass accuracy and retention time filters to determine candidate SILAC peptide pairs for quantitation if either the heavy or light peptide was identified by the database search engine. Therefore, IsoQuant can overcome the limitations of dynamic exclusion-based data-dependent acquisition methods that affect the accuracy of other quantitation programs.</p><!><p>Two common challenges associated with the accuracy of peptide quantification in a shotgun proteomic analysis are the co-elution of contaminant isobaric peptides and the lack of a clear distinction between low abundant peptides and background noise. As a result, these two issues often affect the clear identification of the XIC derived from the target peptide because the boundaries, or the starting and ending points, of the XIC can be difficult to discern. To overcome these issues, both Census 18 and RelEx 19 use a least-squares regression method to quantify peptides with poor S/N ratios. Recently another program, Vista 29, also demonstrated that the quantitation accuracy of low abundant SILAC peptide pairs can be significantly improved when both high mass accuracy and signal/noise ratio were included as part of the algorithm.</p><p>IsoQuant uses the following three filters to identify XIC boundaries: retention time (default setting: 50 MS1 scans preceding and following the MS2 scan of the identified peptide), high mass accuracy (default setting: 5ppm), and detection of complete isotopic envelopes corresponding to both the identified and theoretical SILAC peptides. Figure 4 is a representative example of a SILAC peptide pair quantified by IsoQuant. In this case, the MS2 spectrum of the heavy-labeled SILAC peptide was identified by the SEQUEST search engine at scan #7605. IsoQuant then searches for the MS1 isotopic envelopes corresponding to the parent mass of identified peptides using its high mass accuracy filter (5ppm) and retention time filter. Because of the detection limits of mass spectrometers and the wide dynamic range of most proteomes, it is often difficult to detect the complete isotopic envelopes of low abundant peptides. Therefore, for low abundant peptides or peptides with poor S/N ratios, only monoisotopic peaks are used for XIC construction. In order to discriminate between the true signals derived from low abundant peptides and the signals generated from the background noise, the current version of IsoQuant requires at least two monoisotopic peaks to be identified for each peptide during XIC reconstruction.</p><p>As indicated in Figure 4B, IsoQuant's MS1 extraction module will only extract MS1 scans containing at least two monoisotopic peaks from both the heavy and light peptides. XICs corresponding to heavy and light labeled peptides are reconstructed and areas under the curves are then calculated. Since the light labeled peptide was not identified by SEQUEST database search in this case, IsoQuant used the final SEQUEST search result and automatically identified the corresponding MS1 peaks of the light labeled peptide. Using the same high mass accuracy and retention time filters described earlier, the boundaries of XIC corresponding to isotopic envelopes of the light labeled peptide were then defined. We found that an advantage of using isotopic envelopes (or at least two monoisotopic peaks) to define the boundaries of XIC for low abundant peptides is the simultaneous exclusion of contaminant background peaks since most of these background ions do not have a complete isotopic envelope. In addition, we have also found that requiring a complete set of isotopic envelopes to reconstruct the XIC of target peptides automatically allows IsoQuant to eliminate co-eluting low abundant isobaric peptides, which significantly improves the accuracy of SILAC pair quantification.</p><!><p>In order to evaluate the performance and accuracy of IsoQuant, we used the SILAC-labeled peptide standards provided by the Association of Biomolecular Resource Facility's (ABRF) Proteomics Standards Research Group (sPRG) 30. The natural and synthetic ([13C6, 15N2]-lysine and [13C6, 15N4]-arginine) peptides from five recombinant human proteins (serum albumin, histidyl-tRNA synthetase, ribosyldihydronicotinamide dehydrogenase, peroxiredoxin-1, and ubiquitin) were combined in three defined ratios and quantities. Two technical LC-MS/MS replicates were performed and IsoQuant was used to determine the peptide and protein ratios. The accuracy of the calculated quantitation ratios was assessed using the following criteria. First, the ratios for all occurrences of the same peptide identified in the same run should be consistent. Second, the ratios for all occurrences of the same peptide identified in both runs should be consistent. Lastly, the ratios for all unique peptides identified from the same protein should be consistent.</p><p>Table 1 lists the SILAC ratios calculated by IsoQuant as compared to the published consensus ratios from other laboratories who participated anonymously in the ABRF sPRG09 study. Using IsoQuant, we successfully quantified all 15 peptides and five proteins present in the sample. The average absolute deviation for all protein ratios was 4.99%. Therefore, this study suggests that IsoQuant is an accurate method that can be used to reliably determine peptide and protein ratios.</p><p>To further demonstrate the accuracy of IsoQuant in quantifying the relative abundance of proteins in a complex sample, heavy 13C-Lys/Arg-labeled cell lysates were mixed with light 12C- Lys/Arg-labeled cell lysates at the following H/L ratios = 1:1, 1:2, 1:5, 1:10, 1:25, 1:50 and 1:100. These standard mixtures were then separated by SDS-PAGE in triplicate and analyzed by trypsin in-gel digestion and LC-MS/MS. After the SEQUEST database search, the accuracy of SILAC peptide ratios was compared between IsoQuant and Proteome Discoverer (Figure 5, see Supplemental Tables 1–4 for the details of peptide and protein quantification). IsoQuant was able to quantify more peptides and proteins in the samples with expected SILAC H/L ratios of 1:25, 1:50 and 1:100 (Figures 5A and B, Supplemental Tables 5 and 6). The accuracy of IsoQuant at the H/L ratios = 1:1, 1:2, 1:5, 1:10 and 1:25 is very comparable to Proteome Discoverer (Figure 5C). However, the ratios of the H/L = 1:50 standard mixture as calculated by IsoQuant and Proteome Discoverer were 53.19 ±3.20 and 40.00 ±0.39, respectively. Because of the less than 100% labeling efficiency of the heavy-labeled cell lysate (99%) and the failure to identify low abundant isotopic envelopes of light labeled peptides, the ratios of the H/L = 1:100 standard mixture as calculated by IsoQuant and Proteome Discoverer were 82.53 ± 4.53 and 58.09 ± 3.57, respectively.</p><p>IsoQuant uses a weighted average method to calculate the final protein ratio. The following formula is used to calculate protein ratios: ratio=∑wixi∑wi, wi is the XIC area of heavy and light peptide, is the ratio of heavy and light pair.</p><p>An example of the final assembly of peptide ratios into protein ratios is illustrated in Table 2. It is clear that the weighted average method improves the quantitation of IsoQuant. If many peptides belong to a group of proteins, IsoQuant will group these proteins into a common protein group and use a weighted average method to obtain an average protein group ratio. However, if there is a unique peptide from a specific isoform that can be identified within a protein group, IsoQuant will calculate and record the protein ratio of that specific isoform based on the ratio of the quantified unique peptide pair. Therefore, no outlier removal is used during the process of protein ratio calculation in order to maintain the ratios of the unique peptides for each specific protein isoform. Figure 5D summarizes the accuracy of IsoQuant protein quantitation. The accuracy of IsoQuant protein quantitation at the H/L ratios = 1:1, 1:2, 1:5, 1:10, 1:25 and 1:50 is very close to the ratios of the known standard mixtures. Due to the limitations of the mass spectrometer on the extraction of peptides with low signal/noise ratios, the calculated protein ratio of IsoQuant for the H/L = 100 standard mixture is 68.97 ± 3.77 (Figure 5D, Supplemental Table 6).</p><!><p>We used IsoQuant to conduct a quantitative analysis of a sample derived from SILAC-labeled rat brain hippocampal slices. A representative table of peptide quantitation results and the overall peptide ratio distribution from this sample are illustrated in Figure 6A. If the quantitation of a specific SILAC peptide pair requires further visual validation, the IsoQuant peptide quantitation viewer/visualization module will display all corresponding SILAC MS1 m/z pairs from the original mass spectrometer .raw file that were used to calculate that specific peptide ratio. For instance, the highlighted SILAC peptide MATDPENIIK (Figure 6A, indicated by MS2 scan# 2473) from synaptophysin was quantified with a SILAC ratio of 5.8 ± 0.85. Figure 6B is a representative spectrum of a MATDPENIIK SILAC peptide pair generated from IsoQuant's visual validation module. The "light" MATDPENIIK peptide has an isotope envelope at 574.28 m/z (relative intensity 5%) and the "heavy" MATDPENIIK peptide has an isotope envelope at 576.30 m/z (relative intensity 30%); heavy/light = 5.88 ± 0.85. Therefore, this visual validation module allows users to quickly confirm the accuracy of SILAC ratios generated by IsoQuant.</p><!><p>In summary, we have developed a new SILAC-based mass spectrometry quantitation software tool, named IsoQuant, which provides accurate quantitation ratios and avoids some of the common quantitation limitations related to data-dependent mass spectrometry data acquisition methods. The results using SILAC-labeled hippocampal slice cultures and cell lysate datasets prove that IsoQuant can be used for accurate peptide and protein SILAC-based quantitation of complex samples. Most importantly, IsoQuant offers an easy to use graphic interface peptide/protein quantitation report, and it also includes a visualization platform, which permits users to validate the quality of SILAC peptide and protein ratios.</p>

PubMed Author Manuscript

The zebrafish Cytochrome b5/Cytochrome b5 reductase/NADH system efficiently reduces Cytoglobins 1 and 2 \xe2\x80\x93 Conserved activity of Cytochrome b5/Cytochrome b5 reductases during vertebrate evolution.

Cytoglobin is a heme protein evolutionarily related to hemoglobin and myoglobin. Cytoglobin is expressed ubiquitously in mammalian tissues; however its physiological functions are yet unclear. Phylogenetic analyses indicate that the cytoglobin gene is highly conserved in vertebrate clades, from fish to reptiles, amphibians, birds, and mammals. Most proposed roles for cytoglobin require the maintenance of a pool of reduced cytoglobin (FeII). We have shown previously that the human cytochrome b5 / cytochrome b5 reductase system, considered a quintessential hemoglobin/myoglobin reductant, can reduce human and zebrafish cytoglobins up to 250-fold faster than human hemoglobin or myoglobin. It was unclear whether this reduction of zebrafish cytoglobins by mammalian proteins indicates a conserved pathway through vertebrate evolution. Here we report the reduction of zebrafish cytoglobins 1 and 2 by the zebrafish cytochrome b5 reductase and the two zebrafish cytochrome b5 isoforms. In addition, the reducing system also supports reduction of Globin X, a conserved globin in fish and amphibians. Indeed, the zebrafish reducing system can maintain a fully reduced pool for both cytoglobins, and both cytochrome b5 isoforms can support this process. We determined the P50 for oxygen being 0.5 torr for cytoglobin-1 and 4.4 torr for cytoglobin-2 at 25 \xc2\xb0C. Thus, even at low oxygen tension, the reduced cytoglobins may exist in a predominant oxygen-bound form. In these conditions, the cytochrome b5/cytochrome b5 reductase system can support a conserved role for cytoglobins through evolution, providing electrons for redox signaling reactions such as nitric oxide dioxygenation, nitrite reduction or phospholipid oxidation.

the_zebrafish_cytochrome_b5/cytochrome_b5_reductase/nadh_system_efficiently_reduces_cytoglobins_1_an

Introduction<!>Reagents \xe2\x80\x93<!>Expression and Purification of Recombinant Globins \xe2\x80\x93<!>Sequence analysis \xe2\x80\x93<!>Steady-state kinetics of CYB5R with CYB5a and CYB5b \xe2\x80\x93<!>Reduction of Cytoglobins and Globin X by CYB5/CYB5R \xe2\x80\x93<!>Redox potentials \xe2\x80\x93<!>Determination of melting temperatures \xe2\x80\x93<!>Determination of oxygen binding affinities \xe2\x80\x93<!>Protein expression and spectral properties \xe2\x80\x93<!>Steady-state kinetics \xe2\x80\x93<!>Redox potentials \xe2\x80\x93<!>Thermal denaturation of zebrafish cytochromes b5a and b5b \xe2\x80\x93<!>Oxygen affinity of zebrafish cytoglobins \xe2\x80\x93<!>Reduction of zebrafish cytoglobins by the native zebrafish Cytochrome b5 reductase/Cytochrome b5 systems\xe2\x80\x93<!>Discussion

<p>Vertebrate organisms have retained at least eight globin genes in their genome,1 with five globin genes identified in mammals to date.2 These proteins are usually associated with oxygen transport functions, following the roles of the most conspicuous hemoglobin (Hb) and myoglobin (Mb). However, newly discovered globins have been associated with functions that defy this paradigm. Neuroglobin, androglobin and cytoglobin (Cygb) show cellular levels nowhere near the high concentrations of hemoglobin and myoglobin and appear to be involved in other processes not necessarily related to oxygen transport and storage. Proposed functions include the detoxification of reactive oxygen species, regulation of nitric oxide (NO) levels, signaling reactions and lipid peroxidation.3–8</p><p>Cygb is a heme protein ubiquitously expressed in mammalian tissues.9–11 Unlike hemoglobin and myoglobin, but similar to neuroglobin and androglobin, the heme in Cygb is coordinated by two histidine side chains, resulting in a six-coordinate conformation. The presence of a distal ligand to the heme iron causes important changes in heme properties.4–6, 11–14 Mammals carry one copy of the Cygb gene, however two copies of the gene have been identified in teleost fish (Cygb1 and Cygb2), probably arising from an ancient whole genome duplication event.15, 16 Studies in zebrafish indicate that both genes are expressed.17 The in vitro characterization of the two proteins suggests that Cygb1 may be involved in oxygen transport, whereas Cygb2 has biophysical properties more alike the mammalian Cygb protein.17, 18</p><p>As for other globins, the Cygb reactivity is largely dependent on the oxidation state of its heme group. As most biological reactions will oxidize the heme iron to the ferric (FeIII) form, any catalytic cycle requires a source of electrons to reduce Cygb (equations 1–3). (Equation 1)FeII−O2 → FeIII + O2·− (Equation 2)FeII−O2 + NO → FeIII + NO3− (Equation 3)FeII + NO2− → FeIII+NO+OH−</p><p>To that purpose, it appears that ascorbic acid can fulfill this role in certain conditions.19 New data suggests that the cytochrome b5 (CYB5) and cytochrome b5 reductase 3 (CYB5R) system is a feasible candidate for the physiological reduction of Cygb.20, 21 To our surprise, we observed that the human CYB5/CYB5R system catalyzes human and zebrafish Cygb reduction at rates up to 250-fold higher than those for their bona fide substrates, Hb and Mb.20</p><p>Globin X (GbX) is an ancestral globin protein recently discovered in fish and amphibians.22 The function of this globin is yet unknown. It has been noted that it can be lapidated and associated with membranes23 and it is also present in fish thrombocites (the equivalent to mammalian erythrocites), where it may be involved in oxygen or nitric oxide related functions.24 In our previous work we have observed that human CYB5b/CYB5R has some ability to reduce GbX heme, whereas 5mM Ascorbate shows almost no effect.20</p><p>One caveat in our previous study is that we used human CYB5/CYB5R proteins to monitor the reduction of the zebrafish Cygbs and GbX.20 In order to establish unequivocally whether this reaction is functional in zebrafish, a complete system of zebrafish proteins should be used. Moreover, we only tested the ability of human CYB5b to support Cygb and GbX reduction. As the properties of the CYB5a and CYB5b isoforms may differ notably,25, 26 and two CYB5 isoforms are present in zebrafish, as in mammalian genomes, the ability of both proteins to reduce Cygb and GbX may differ and hint towards tissue- or cell-specific capabilities.</p><p>Homologous genes for the 4 CYB5R isoforms found in mammals have been identified in the zebrafish genome. It has been reported that a membrane-bound CYB5R protein can reduce fish Hb and maintain proper Hb reduction levels.27, 28 The NADH diaphorase activity of fish red blood cell (RBC) extracts, a surrogate for CYB5R activity, correlates with the resistance of fish to methemoglobinemia-inducing conditions.27, 28 Although the presence of a CYB5R3 homolog – the isoform found in mammalian RBCs and mitochondria – in fish has been described,29, 30 and the protein shows ≈ 65% identity with the characterized mammalian enzymes, the properties of the fish CYB5Rs are unknown. Thus we also characterized the properties of the zebrafish CYB5R3 protein.</p><p>Here we study the ability of the reconstituted zebrafish CYB5/CYB5R system to reduce zebrafish Cygbs and GbX. Our results indicate that these zebrafish proteins provide a viable system for the reduction of the Cygbs, with both CYB5 proteins supporting the fast reduction of Cygb1 and Cygb2 at physiological temperatures. The complete zebrafish electron transfer system also catalyzes the reduction of GbX faster than reported for the human CYB5/CYB5R system or ascorbate, and may also provide a physiological source of electrons for GbX in vivo. These observations suggest a conserved role of CYB5/CYB5R supporting Cygb function that possibly predates its role as hemoglobin and myoglobin reductases.</p><!><p>All reagents were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise specified.</p><!><p>Zebrafish GbX, Cygb1 and Cygb2 were cloned into the pET-28 plasmid (Novagen) and expressed in E. coli cells and purified as previously described.20, 24 The cDNA clones for zebrafish CYB5a, CYB5b and CYB5R (cytochrome b5 reductase 3) were obtained from the IMAGE consortium through Thermo Fisher/Open Biosystems. The accession numbers and IMAGE identifiers are as follows: CYB5R, BC_066624 (IMAGE 6960174); CYB5a, BC_154824 (IMAGE 9002123); CYB5b, BC_066748 (IMAGE 6525311). CYB5R was cloned into the pET28a plasmid (Novagen) using the NdeI/HindIII restriction sites. The residues 1–20 necessary for membrane insertion were removed, and the 21–298 sequence, coding the soluble protein, was cloned (Figure 1). The protein sequence includes an N-terminal 6xHis tag. CYB5a and CYB5b were cloned into pET11a plasmid using the NdeI and NcoI restriction sites. The C-terminal residues 101–137 of CYB5a that form a membrane insertion helix were removed. For CYB5b, the N-terminal residues 1–17 and the C-terminal residues 115–153 were removed to produce the soluble protein as reported for mammalian proteins (Figure 1).25, 26 Purification of zebrafish CYB5 proteins and CYB5R was carried out as described for the human proteins.20, 31 Spectral data were recorded using either a Cary 50 spectrophotometer (Agilent) or an Agilent HP8453 diode array spectrophotometer. Proteins were studied in 100 mM sodium phosphate buffer, pH 7.4.</p><!><p>CYB5 sequences were analyzed for possible posttranslational modifications using publicly available software. Acetylation motifs were scanned using Terminus32 and NetAcet33 servers. Palmitoylation sequences were screened using the CSS-Palm 4.0 software.34</p><!><p>Steady state parameters for the reaction of CYB5R with NADH and CYB5a/CYB5b were determined as follows: NADH diaphorase activity was assayed using dichlorophenol-indophenol (DCPIP) or potassium ferricyanide as electron acceptors. In the experiments with DCPIP, zebrafish CYB5R (0.028 μM) was incubated with 95 μM DCPIP and the reaction was initiated by adding variable amounts of a 1 mM NADH stock solution for final NADH concentrations between 0 and 22 μM. The reduction of DCPIP was monitored at 600 nm. In the experiments with potassium ferricyanide, zebrafish CYB5R (0.84 nM) was incubated with 1 mM potassium ferricyanide and the reaction was initiated by adding variable amounts of a 1 mM NADH stock solution for final NADH concentrations between 0 and 50 μM. The reduction of potassium ferricyanide was monitored via the decay of NADH absorbance at 340 nm. The initial reduction rates were plotted versus the concentration of NADH and the data was fitted to the Michaelis-Menten equation to determine the kcat and KM parameters. To determine the apparent KM for CYB5a and CYB5b, the reaction samples included CYB5R (7.3 nM), NADH (60 μM) and variable amounts of CYB5a or CYB5b (0 to 140 μM). Reduction of CYB5a or CYB5b was monitored at 555 nm. The initial rates were fitted as a function of the initial CYB5 concentrations and the data was fitted to the Michaelis-Menten equation to determine the kinetic parameters as above. Experiments were carried out in 50 mM Bis-Tris propane buffer, pH 7.4, at 25 °C.</p><!><p>Steady-state reduction of zebrafish Cygbs and GbX by the CYB5/CYB5R was studied as previously reported.20 The reduction was performed in anaerobic conditions and monitored in an Agilent HP8453 spectrophotometer housed in an anaerobic glovebox (Coy Laboratories, Grass Lake, MI). Oxidized globins were prepared by treatment with excess potassium ferricyanide and then the excess ferricyanide was removed by passing the sample through a Sephadex G25 column (PD10, GE healthcare). The globin solution was diluted to a 20 μM final concentration and zebrafish CYB5R (0.2 μM) and either CYB5a or CYB5b (2 μM) was added. The reaction was initiated by addition of 100 μM NADH. The fraction reduced was determined by monitoring the absorbance changes in the characteristic peak for the ferrous form of the globins (≈ 560 nm) as reported.20</p><!><p>Oxidation-reduction potentials for the zebrafish CYB5R and CYB5 proteins were determined in anaerobic conditions at 25 °C in 100 mM sodium phosphate buffer, pH 7.0. Spectral data were recorded using an Agilent HP8453 spectrophotometer and redox potential was collected via an Accumet 15 connected to a MI-480 electrode (Microelectrodes, Inc). Proteins (≈10 μM) were titrated with either sodium dithionite (reductive titrations) or potassium ferricyanide (oxidative titrations) in the presence of redox mediators. The redox mediators (1–5 μM) used were phenosafranine (Em = −252 mV) for CYB5R. In the case of CYB5a and CYB5b a modified version of the method described by Efimov et al35, using sodium dithionite as reductant instead of the xanthine/xanthine oxidase system and an anaerobic setup that circumvents the use of the glucose/glucose oxidase system to remove oxygen. Indigo tetrasulfonate (Em = −46 mV) was used as mediator. Spectral and redox potential readings were analyzed using the Nernst equation to calculate the midpoint potentials as described.13, 17, 35</p><!><p>Thermal denaturation of CYB5a and CYB5b proteins in the ferric heme state was monitored by UV-Vis spectroscopy using a Cary50 spectrophotometer. Experiments were carried out in 10 mM phosphate buffered saline, pH 7.4. The temperature was increased from 20 °C to 100 °C; for each step the temperature was maintained for ~5 minutes and the spectral changes were then recorded. Changes in the Soret peak were fitted to the Santoro-Bolen equation36 to determine the melting temperature (Tm) of each protein.</p><!><p>In order to determine the oxygen partial pressures at half-saturation of the ferrous-oxygen complex (P50) of Cygb1 and Cygb2, we measured oxygen equilibrium curves using a thin-layer modified diffusion chamber described elsewhere37–39 with some modifications.14 Samples (5 μl) contained 200 μM heme Cygb1 or Cygb2. As a reducing system, 40 μM human CYB5b, 4 μM human CYB5R and 600 μM NADH were added. 20 mM sodium formate and 10 mU of formate dehydrogenase were added to regenerate the consumed NADH. Samples were equilibrated for 5 minutes at room temperature under N2 atmosphere to allow heme reduction before transfer to the modified diffusion chamber. Experiments were conducted at either 25°C or 37°C in 100 mM potassium phosphate buffer at pH 6.8 or 7.4. Fit of the oxygen saturation data to the Hill equation also provided the Hill coefficient measuring the degree of cooperativity in oxygen binding.</p><!><p>Zebrafish CYB5R was expressed in soluble form by removing the N-terminus sequence (amino acids 1–20, Figure 1A) to prevent membrane binding. In mammalian CYB5R, Gly2 has been found to be myristoylated to direct the protein to mitochondria and other membrane fractions; the initial fragment (amino acids 1–23 in mammalian proteins) is not present in the soluble isoform expressed in mammalian erythrocytes.40 The sequence of the zebrafish CYB5R lacks the glycine in position 2. Nevertheless, the Protein in fish erythrocytes appears to be membrane-bound28. Although the post translational modification responsible has not been identified, sequence analysis software predicts a possible acetylation site in Ser2 (Terminus32, NetAcet33) and palmitoylation of Cys12 (CSS-Palm34). Zebrafish CYB5a (microsomal isoform) was expressed in its soluble form without its C-terminal membrane binding-domain (amino acids 101–137, Figure 1B). Zebrafish CYB5b (mitochondrial outer membrane isoform) was also expressed in a soluble form, omitting its N-terminal amino acids (amino acids 1–17, Figure 1B) and C-terminal transmembrane domain (amino acids 115–153, Figure 1B).</p><p>The zebrafish CYB5R and cytochromes CYB5a and CYB5b were expressed in E. coli and purified following similar protocols to those used previously for the mammalian proteins.20, 31 We did not observe notable differences in the expression and purification as compared to the human proteins. The three proteins show spectral properties similar to their mammalian counterparts.25, 41 Spectra for the fully oxidized and fully reduced species are shown in Figure 2. Oxidized CYB5R shows the characteristic peaks of the FAD cofactor, with maxima at 388 nm and 460 nm (Figure 2A). CYB5a and CYB5b showed similar wavelengths for their Soret peaks in the oxidized state (413 nm) and an additional peak at 531 nm for CYB5a or 528mn for CYB5b, consistent with the spectra of other six-coordinated heme proteins in the ferric form. CYB5a and CYB5b in the reduced state show spectra characteristic of a bis-His, six-coordinate heme, with two peaks around 520 nm and 560 nm. The observed maxima were 423 nm, 527 nm and 555 nm (CYB5a) and 423 nm, 527 nm and 556 nm (CYB5b) (Figures 2B, 2C).</p><!><p>In order to characterize the ability of CYB5R to oxidize NADH and catalyze electron transfer reactions to its electron transfer partner proteins CYB5a and CYB5b, we studied the kinetics of CYB5R with different electron acceptors in steady-state conditions.</p><p>We determined the KM of zebrafish CYB5R for NADH in the presence of DCPIP and ferricyanide as electron acceptors. The observed rates for the reaction of CYB5R with DCPIP are shown in Figure 3A. The values were fitted to a Michaelis-Menten equation yielding a kcat of 26 s−1 and a KM of 0.6 μM (Figure 3A, Table 1). This activity is comparable to that reported for the other CYB5R proteins (Table 1). When potassium ferricyanide was used as electron acceptor, we observed rates also consistent with a Michaelis-Menten fit. Calculated values were kcat of 235 s−1 and a KM of 3.7 μM (Figure 3B, Table 1).</p><p>The kcat values are comparable to those of the fly CYB5R (Table 1); although 3–4-fold lower than those of mammalian enzymes (~600–800 s−1, Table 1). The KM value is comparable to the reported KM values for other CYB5R enzymes (~0.6–6.0 μM, Table 1) indicating a high affinity towards NADH. We then determined the activity of CYB5R with its partner proteins CYB5a and CYB5b. These experiments were conducted in saturating concentrations of NADH (60 μM).</p><p>Experiments on the absence of CYB5R showed that NADH can also reduce directly CYB5a and CYB5b at a significant rate; background activity in the absence of CYB5R was subtracted from the rates in the presence of CYB5R to determine reduction rate due to CYB5R (Figure 3B,C). The fit of the initial rates yielded values of kcat = 165 s−1 and KM = 36 μM for CYB5a and kcat = 121 s−1 KM = 40 μM for CYB5b (Table 1). Thus, CYB5R reduces both substrates with very similar kinetic parameters, though at slightly faster rates with CYB5a. The KM values are higher than the expected physiological CYB5 concentrations, suggesting a linear dependence of the reduction rates with CYB5 concentrations in vivo.</p><!><p>In order to further characterize the properties of the zebrafish CYB5R/CYB5 system, we determined the redox potential of the three zebrafish proteins (Figure 4). The titration of CYB5R shows a single transition, fitted to a slope of ~29.5 mV, consistent with a two-electron step (Figure 1A). Similar observations have been reported for other CYB5R proteins, where the one-electron reduced, semiquinone state is not stable and does not accumulate during the reduction step.43, 47–49 Thus, the two one-electron transitions show a very similar midpoint potential and the two steps are merged in a single apparent Em value. We observe a midpoint potential of 264 mV, in the range of the values reported for mammalian proteins (Table 2).</p><p>Previous work has indicated notable differences in the redox potential of CYB5a and CYB5b in mammalian systems. Reported CYB5a midpoint potential values are around 0 mV, whereas CYB5b proteins show more negative values. In particular, human and rat proteins show values of −40 and −102 mV, respectively (Table 2).25, 26 We observe values in a similar range for the zebrafish proteins (Table 1). The potential for CYB5a is −28 mV, slightly more negative than the value observed for mammalian proteins. In the case of CYB5b, the observed value of −62 mV is between the observed values for rat and human CYB5b proteins. Notably, the observed values for the zebrafish Cygb1 and Cygb2, −58 mV and −26 mV respectively17, are in a similar range to those of the zebrafish CYB5 proteins (Table 2).</p><!><p>As noted above we observe a 35 mV difference in redox midpotential values between zebrafish CYB5a and CYB5b. This is not unlike the situation for mammalian CYB5 proteins, where differences of up to 95 mV have been reported52–54. Given the high identity and similarity of the CYB5 sequences, the source of these differences is not evident (Figure 1). The changes within CYB5 sequences are limited to structural elements away from the heme binding regions, with nearly 100% identity in the regions close to the heme within CYB5a or CYB5b sequences, and limited changes between the two CYB5s. Given the high variability between mammalian CYB5a and CYB5b redox potentials (Table 3), it has been speculated that the redox potential is in part modulated by the strength of the bond between the FeIII-heme and the proximal histidine.25 Although the heme exposure to the solvent is one of the main contributors to the regulation on the heme redox potential in proteins,55 the differences in heme binding in the ferrous and ferric states can also cause large variations in the redox potential, even in conditions where the heme solvent accessibility is unchanged.56 As a proxy for the FeIII-His bond strength we studied the thermal denaturation of the zebrafish CYB5a and CYB5b proteins. Our experimental determinations of the melting temperatures (Tm) are shown in Figure 5 and Table 3. Our results indicate that the two zebrafish proteins have similar Tm values (Table 3). The value for CYB5a is in line with mammalian CYB5a proteins, however the Tm for zebrafish CYB5b appears much lower than the reported values for mammalian CYB5b (Table 3). The spectral changes also indicate clear differences between the thermal denaturation processes for CYB5a and CYB5b. The denaturation of CYB5a shows a shift from the spectra of the bis-His hexacoordinated globin towards a spectrum similar to that of free heme, with a Soret peak around 385 nm (Figure 5A). A similar increase in the 350–400nm area, indicating heme dissociation, is observed for CYB5b, but in this case the heme release correlates with visible light scattering as noted by the increase of the signal at 700nm and throughout the spectral range (Figure 5B). Without further structural data, the nature of the changes cannot be unequivocally assigned, but we can speculate that the denaturation of the zebrafish CYB5a leads to the formation of a stable apoprotein, whereas the zebrafish CYB5b aggregates when the heme is dissociated form the protein, suggesting a less stable apoprotein in our experimental conditions.</p><!><p>We have shown that zebrafish Cygb1 and Cygb2 can be reduced efficiently by the mammalian CYB5R/CYB5/NADH system.20 This allows for a catalytic cycle of NO dioxygenation with continuous regeneration of the deoxy (FeII) species, and in the presence of oxygen, the subsequent fast formation of the oxygen-bound species (FeII-O2), as shown in the following equations: (Equation 2)FeII−O2+NO→FeIII+NO3− (Equation 4)FeIII+e−→FeII (Equation 5)FeII+O2→FeII−O2 (Equation 6)NO+O2+e−→NO3−</p><p>Where equation 4 indicates the regeneration of the Cygb deoxy species by reaction with the pool of reduced CYB5 and the net sum of the three reactions yields the catalytic deoxygenation of NO (Equation 6).</p><p>In order for zebrafish Cygbs to support NO dioxygenation reactions, the reduced form of the protein should be able to bind oxygen in physiological conditions, and therefore it is important to determine the oxygen affinity of these Cygbs. We used a met-reducing enzymatic system57 to determine P50 values for Cygb1 and Cygb2. We have used a similar system in the determination of the P50 for other heme proteins, including several mutant Cygbs and GbX.14,23 The system uses ferredoxin NADP+-reductase, ferredoxin and NADPH to regenerate the pool of globin oxidized due to endogenous autoxidation. However this system was unable to maintain zebrafish Cygb2 in the ferrous reduced state due to its very fast autoxidation rate.17 As we previously observed the efficient reduction by the human CYB5R/CYB5 system,20 we used these proteins to keep the zebrafish Cygbs in the reduced state. Formate and formate dehydrogenase were included to regenerate the NADH consumed.</p><p>Our experimental results for the determination of the P50 values at pH 6.8 and 7.4 at either 25 °C or 37 °C are shown in Figure 6 and Table 4. We observe that Cygb1 has a higher oxygen affinity than Cygb2, with values between 0.4–2.5 torr for Cygb1 compared to 3.6–6.1 torr for Cygb 2. The P50 values for Cygb2 appear higher than those reported for mammalian Cygbs at pH 7.0 (0.2–2.8 torr)11, 14, 58, 59 which is unexpected given that most properties of Cygb2 are otherwise more similar to the mammalian protein than Cygb1.17 On the other hand, Cygb1 shows very high affinity towards oxygen with similar or lower P50 values than those of the mammalian Cygbs. Comparing the data at different pH values, neither Cygb shows a significant Bohr effect, in agreement with previous data on mammalian Cygbs.14</p><p>Analysis of the Hill coefficients (n values in Table 4) indicates that both zebrafish proteins bind oxygen cooperatively. In particular, Cygb2 has very high n values (3.0–4.3), as also evident from the highly sigmoidal oxygen equilibrium curves (Fig. 6), suggesting formation of complexes larger than tetramers. It is difficult to speculate further from these data on the nature of these complexes. It has been shown that oligomerization of mammalian Cygbs, with formation of tetramers and octamers, may occur at high concentrations, probably at supraphysiological concentrations.59 It is thus unclear if the oligomers observed in our study are physiologically relevant or are due to the high protein concentrations used and do not necessarily reflect the cellular conditions. Zebrafish GbX binds O2 cooperatively with a overall lower affinity (P50 1.3–12.5 torr at 20 °C)23 compared to zebrafish Cygbs, and thus may not be significantly oxygenated in vivo.</p><!><p>In previous work we have shown that human cytochrome b5 reductase, in combination with human CYB5b, can efficiently reduce zebrafish Cygb1 and Cygb2. We were thus interested in evaluating if the native zebrafish systems, using CYB5R in combination with either zebrafish CYB5 isoform, could also reduce the Cygbs, and if so how they perform as compared to the human reducing system.</p><p>In order to test the reduction of the Cygbs by CYB5/CYB5R/NADH we monitored the reduction of zebrafish Cygbs 1 or 2 (20 μM) in the presence of zebrafish CYB5a or CYB5b (2 μM, consistent with reported physiological levels,60, 61 either zebrafish or human CYB5R (0.2 μM), and 100 μM NADH to ensure CYB5R saturating conditions.</p><p>Our results are shown in Figure 7. As in the case of the human Cygb in the presence of human CYB5/CYB5R,20 the zebrafish CYB5/CYB5R system is able to completely reduce Cygb1 or Cygb2 in around 100 s. Both CYB5 isoforms can support fast Cygb1 or Cygb2 reduction, with the CYB5b performing slightly faster than CYB5a (Figure 7). Increasing the temperature from 25 °C to 37 °C shows a modest effect on the reduction rates. In order to compare the reduction of zebrafish Cygbs by the zebrafish CYB5/CYB5R system, traces for previous experiments using the human CYB5b/CYB5R are included (Figure 7 D–E).20 In the case of Cygb1, the human proteins can reduce the zebrafish Cygb1 at a comparable rate (Figure 7D). However in the case of Cygb2, the zebrafish proteins clearly outperform the human CYB5/CYB5R system, and can support full Cygb2 reduction in ≈50 s as compared to ≈200 s for complete reduction with the human proteins (Figure 7E).</p><p>We have previously hypothesized that GbX could be reduced by the CYB5/CYB5R system in zebrafish.20 The traces for the reaction of the zebrafish CYB5/CYB5R system with GbX are shown in Figure 7 (Panels 7C and 7F). Our results indicate that this reduction system, although not as efficient as for Cygb, can reduce GbX at an appreciable rate. In particular CYB5b is efficient in this reduction even at 25 °C-consistent with zebrafish physiological conditions-catalyzing GbX reduction at a faster rate than the human system at 37 °C (Figure 7F). Thus, we conclude that CYB5/CYB5R can provide a pathway for GbX reduction in zebrafish.</p><!><p>Our previous work strongly supported the role of the CYB5/CYB5R/NADH system as physiological reductant of Cygb in human cells.20 It also noted the ability of the human proteins to reduced fish Cygb1 and Cygb2 as well; this observation suggested that this could be a pathway not limited to mammals but conserved in vertebrates.</p><p>To test this hypothesis we expressed and purified the components of the CYB5 reducing system from zebrafish in E. coli. This is to our knowledge the first characterization of recombinant CYB5R and CYB5 proteins from fish. Of note, the enzymatic activity of CYB5R in fish is particularly relevant to aquaculture, where the accumulation of nitrites in water can cause methemoglobinemia–accumulation of oxidized hemoglobin in the erythrocytes–63, 64. Fish can recycle methemoglobin to its reduced form through several pathways, including reaction with glutathione, ascorbic acid, or NADPH-dependent methemoglobin reductase; however the main contributor is CYB5R. Indeed, the differences in CYB5R activity can partly explain fish species sensitivity to methemoglobinemia.27</p><p>The CYB5R/CYB5R proteins in fish showed comparable properties to the mammalian proteins. The main differences were the more negative midpoint potential for zebrafish CYB5a and the lower thermal stability of CYB5b as compared with the more stable mammalian CYB5b paralogues. However there are differences may be partly related to the lower body temperature of the fish and do not appear to indicate large differences in their physiological properties. As the FeIII-His interaction seems to be less strong in CYB5b, contrary to what is expected given its more negative redox potential,56 it suggests that other factors such as heme exposure53 may be more relevant to zebrafish CYB5b redox potential. The cause for the reduced stability is yet unknown but we notice that important residues from the hydrophobic core (Ala18, Ile32, Leu36, Leu47; rat CYB5b numbering) identified in rat CYB5s25, 26 are not conserved in zebrafish CYB5b; notably, Ala18 and Leu36 correspond to Gly41 and Met59, respectively, in zebrafish CYB5b.</p><p>Our experiments of Cygb reduction show that CYB5 and CYB5R are able to reduce Cygb1 and Cygb2 in the reduced state. This reduction is particularly important for Cygb2, as this protein shows autoxidation rates 5-fold faster than human Cygb.17</p><p>A limitation of our previous study20 is the use of the CYB5b isoform, which is mostly present in the mitochondrial outer membrane, thus probably isolated from cytoplasmic cytoglobins. However, the CYB5a isoform is attached to the endoplasmic reticulum, with the heme domain remaining in the cytoplasmic side and thus available for interactions with cytoplasmic molecules.65–67 CYB5R3 is located in the mitochondrial outer membrane and also attached to the endoplasmic reticulum67 and can thus provide electrons to cytosolic cytoglobins through CYB5a. GbX has been found to associate with the plasma membrane when expressed in mammalian cells23. In these conditions a cytosolic, soluble CYB5 could provide electrons to GbX. The presence of soluble CYB5 proteins in zebrafish is not established. However, large scale protein expression studies in human tissues indicate that human CYB5a and CYB5b can be found in the cytosol-dependent or independent of membrane binding-68, 69, Thus, we hypothesize that membrane-bound GbX could be reduced in vivo by the NADH/CYB5/CYB5R system either via CYB5 soluble isoforms present in zebrafish cells – as will be the case in thrombocytes, where GBX has been detected24, or by localizing to the endoplasmic reticulum membrane, where CYB5a and CYB5R are available.</p><p>Mammalian Cygb may play an important role in the regulation of NO levels in the vascular wall.19, 21, 70 It is unclear if Cygbs are present in fish blood vessels, but given the important physiological differences between NO metabolism and vascular biology of the fish and mammals, it is conceivable that Cygb in the fish accomplish other unrelated functions. However, we have shown that fish, like mammals, have a very effective Cygb reduction system, suggesting that NO detoxification or other roles requiring the reduced Cygb can be maintained. We observe very low P50 for both Cygbs, indicating that these proteins would be largely saturated with oxygen in vivo and that their biological roles may specifically involve the ferrous-oxy species, at least for Cygb1. Alternatively, Cygb2 can be involved in O2 transport but its P50 may be optimal for oxygen sensing functions. Experiments in different fish species have shown increases of Cygb1 mRNA transcripts under hypoxia – but usually not Cygb2 – which is consistent with this hypothesis71–73. The deoxy Cygb2 species can be formed in hypoxic tissues, where it could fulfill other functions not related to oxygen binding.</p>

PubMed Author Manuscript

Synthesis of <i>N</i>-aryl amines enabled by photocatalytic dehydrogenation

Catalytic dehydrogenation (CD) via visible-light photoredox catalysis provides an efficient route for the synthesis of aromatic compounds. However, access to N-aryl amines, which are widely utilized synthetic moieties, via visible-light-induced CD remains a significant challenge, because of the difficulty in controlling the reactivity of amines under photocatalytic conditions. Here, the visible-light-induced photocatalytic synthesis of N-aryl amines was achieved by the CD of allylic amines. The unusual strategy using C 6 F 5 I as an hydrogen-atom acceptor enables the mild and controlled CD of amines bearing various functional groups and activated C-H bonds, suppressing side-reaction of the reactive N-aryl amine products. Thorough mechanistic studies suggest the involvement of single-electron and hydrogen-atom transfers in a well-defined order to provide a synergistic effect in the control of the reactivity. Notably, the back-electron transfer process prevents the desired product from further reacting under oxidative conditions.Scheme 1 Visible-light-induced catalytic dehydrogenation.

synthesis_of_<i>n</i>-aryl_amines_enabled_by_photocatalytic_dehydrogenation

Introduction<!>Results and discussion<!>Conclusions<!>Conflicts of interest

<p>Catalytic dehydrogenation (CD) offers a useful strategy for constructing unsaturated systems from readily available sp 3 carbon-rich scaffolds. 1 For example, the CD of hydrocarbons provides valuable olen feedstocks, 1c and alcohol dehydrogenation can promote the formation of carbonyl functional groups. 1a By exploiting the thermodynamic preference towards aromatization, a number of transformations to obtain various aromatic compounds have been developed. 2 Nevertheless, despite the signicant advances in the CD process, the requirement for harsh reaction conditions, such as high reaction temperatures and the use of strong oxidants still limits the practical application of the strategy.</p><p>Utilization of visible-light in organic transformations via photoredox catalysis has led to novel and practical synthesis of target molecules under relatively mild reaction conditions. 3 Visible-light photoredox catalysis has also been applied to CD reactions (Scheme 1). 4 Since the initial development of a photocatalyst and the use of an external oxidant such as O 2 , 4a tert-butyl peroxybenzoate, 4b and sulfonyl chloride, 4h the combination of a visible-light photocatalyst (Ir, Ru, organocatalyst) and a hydrogen-evolution metal catalyst (Co, Pd) has afforded the aromatization of saturated N-heterocycles such as tetrahydroquinoline and indoline via an acceptorless CD process to efficiently produce the corresponding aromatic heterocycles (Scheme 1A). 4c-f,i-k,m Further elaboration of CD strategies using an additional hydrogen-atom transfer (HAT) catalyst enabled direct access to more challenging substrates. For example, Kanai designed a triple hybrid catalysis (photoredox, HAT, and metal catalysts) to achieve the acceptorless photoredox CD of tetrahydronaphthalene derivatives (Scheme 1A). 4e,g Very recently, the developed strategies were applied to CD of acyclic alcohols or secondary amines, providing carbonyl or imine compounds. 4l,n,o Despite these advances, the synthesis of functionalized benzenes, such as N-aryl amines, has been challenging, despite the widespread use of these moieties in organic, materials, and biological chemistries. Direct access to N-aryl amines via visiblelight photocatalytic dehydrogenation is difficult (Scheme 1B) because the strategy previously applied in photocatalytic dehydrogenation relies on the single-electron transfer (SET) of a nitrogen atom to produce reactive amine radical cation species. 5 Saturated N-heterocyclic systems, which contain nitrogen atoms inside the ring system, allow control of the reaction to promote the desired dehydrogenation; however, systems with an exo-nitrogen functionality are more prone to unwanted side-reactions (Scheme 1B). 6 These possibilities raise the question of how the reactivity of the exo-nitrogen atom is transferred to the target cyclic system to selectively achieve the desired dehydrogenation reaction. Moreover, the produced Naryl amine would become a competing and more reactive substrate in the course of the reaction. This has the potential to transform N-aryl amine products into unwanted a-amino radical, iminium, and enamine species under visible-light photoredox catalysis (Scheme 1B). 5,6 Very recently, Leonori and co-workers successfully overcame the above-mentioned hurdles via photocatalytic dehydrogenation of in situ generated enamines with photoredox/cobalt dual catalytic system. 7 Herein, we demonstrate the synthesis of N-aryl amines from a 2-cyclohexenyl amine precursor via visible-light photoredox CD (Scheme 1C). The use of peruoroarene as a hydrogen-atom acceptor is key to achieving selective aromatization on the cyclohexenyl moiety, and wide ranges of functional groups and bioactive motifs are tolerated under the reaction conditions. The operating mechanism was revealed to involve the synergistic relay of SET and HAT processes mediated by the organic photoredox catalyst and a pentauorophenyl radical, in combination with the protective back-electron transfer (BET) process, based on the nature of the amine substrates.</p><!><p>The 2-cyclohexenyl amine A1 was initially selected as a model substrate to explore the target reaction. Aer extensive screening (Tables S1-S5 †), 3DPAFIPN was identied as the photocatalyst, iodopentauorobenzene (C 6 F 5 I) as the hydrogenatom acceptor, and K 2 CO 3 as the base to induce the desired oxidative aromatization in high yield (95%). It is noteworthy that no imine compounds derived from either A1 or B1 were observed, although the CD of secondary amines under visiblelight photoredox catalytic conditions was reported in the synthesis of imines. 4o Further variation of the standard reaction conditions was implemented to check the effect of each component (Table 1). Changing the photocatalyst to a wellknown iridium photocatalyst ([Ir(dF(CF 3 )ppy) 2 (dtbbpy)](PF 6 )) signicantly decreased the reactivity (52%, entry 2), and photocatalyst-free conditions yielded only a small amount of the desired product (4%, entry 10). Replacing C 6 F 5 I with C 6 F 5 Br (entry 3) completely shut down the reaction, although it was previously reported that C 6 F 5 Br can be converted to a penta-uorophenyl radical under visible-light irradiation in the presence of Eosin Y. 8 In addition, no pentauorophenylation of the aniline aromatic ring as a side reaction was observed in all entries. The developed reaction is highly sensitive to the solvent (entries 4-6, Table S3 †), and dichloromethane can only be replaced with 1,2-dichloroethane (1,2-DCE) without loss of the reactivity (91%, entry 6). Potassium carbonate was essential for obtaining a high yield of the product (entries 7, 8, 12), and a control experiment under dark conditions (entry 13) proved that the reaction proceeded under visible-light irradiation.</p><p>With the optimal conditions in hand, the substrate scope of the reaction was evaluated (Table 2). Several benzylamine derivatives were initially tested for this reaction. Notably, various functional groups, including halides (B3, B4, B8), ester (77%, B5), triuoromethyl (61%, B6), and boronate ester (44%, B7), were tolerated under the reaction conditions and gave the desired N-aryl amines in good yields. A piperonylamine derivative (77%, B9) and other heteroaromatic methyl amine derivatives (85%, B10 and 32%, B11) were reactive under the reaction conditions. Other secondary amine precursors, such as amethyl benzylamine (84%, B12), linear (70%, B14), or cyclic (70%, B15) aliphatic amines, and even a sterically bulky 1-adamantylamine derivative (74%, B16) could be applied in the reaction. Notably, the cyclohexyl ring of B15 remains intact under the dehydrogenative conditions, implying the necessity of an allylic a-amino C-H bond for the desired reaction. The Table 2 Evaluation of the substrate scope a a Reaction conditions: A (0.2 mmol), 3DPAFIPN (0.012 mmol), C 6 F 5 I (0.8 mmol), K 2 CO 3 (0.6 mmol), CH 2 Cl 2 (1 mL) in a 4 mL reaction vial, irradiated with a 34 W Kessil blue LED for 24 h under fan cooling. Yields of the isolated products were described. b 36 h. c 48 h. efficiency of the reaction was maintained even with different types of tertiary allyl amines (B17-B21). To our delight, no oxidation at the benzyl, 4o,9 tetrahydronaphthyl, 4e,g,7 or a-amino 10 positions was observed (e.g., reactions with A22 or A23, highlighted in orange), demonstrating the advantage of the mild reaction conditions compared to previously reported oxidative aromatization conditions.</p><p>The cyclohexenyl scaffold was then varied and the synthesis of ortho-/meta-/para-substituted N-aryl amines was achieved using the developed method, albeit with a reduced yield of the product in some cases (B24-B29). It is noteworthy that sulde (A29) is compatible with the reaction, producing a sulfurcontaining N-aryl amine in good yield (68%, B29), considering that suldes are known to be readily oxidizable under oxidation conditions and tend to inhibit transition metal-mediated CD. 11 The developed method was further applied to bioactive amines (Table 3). Amines containing polycyclic structures, amino acid moieties, and terpenoid structures all remained intact to form corresponding N-aryl amines in moderate to good yields. Various types of reactive positions, such as benzylic (62%, B30), tertiary (67%, B32), and a-oxygen carbons (79%, B34), were tolerated under the developed reaction conditions. When the Sertraline derivative (A33) containing both cyclohexenyl and tetrahydronaphthyl groups was used, selective aromatization of the cyclohexenyl group was achieved to form the corresponding N-aryl amine (45%, B33) with the tetrahydronaphthyl group remaining intact. This unique chemoselectivity may be benecial for the synthesis of bioactive Naryl amines while maintaining a specic structure. Further elaborations with different cyclohexenyl moieties were conducted using Mexiletine as a starting amine. As depicted in Table 3, the introduction of a simple phenyl group (B34) and substituted phenyl groups (48% for B35, 72% for B36) was realized under the standard reaction conditions with good efficiency.</p><p>The developed protocol for photocatalytic dehydrogenation provides a high degree of reactivity and selectivity for a wide range of allylic amines. This raises fundamental questions regarding the mechanism of the reaction (Fig. 1). First, the origin of the selective dehydrogenation leading to aromatization needs to be addressed to understand the underlying principle of the reaction. Several reactions with A containing a reactive benzylic position did not show any sign of activation of the benzylic position, which is in clear contrast to the reported functionalizations of benzylic amines. 5,12 In addition, the role of C 6 F 5 I as a hydrogen-atom acceptor in the reaction must be further investigated. Lastly, the preservation of the synthesized N-aryl amines under photocatalytic conditions (E ox p (B1) ¼ 0.88 V vs. SCE in CH 3 CN) 13 is the most noticeable aspect of the reaction, which enables controlled dehydrogenation with high efficiency.</p><p>To answer the questions about the mechanism of the developed reaction, a number of control experiments were conducted (Table 4). The reactions with C 6 F 5 Br as the sole reagent did not provide the desired amine B1 (entry 2; Table 1, entry 3), although C 6 F 5 Br can provide a C 6 F 5 radical and promote peruoroarylation under visible-light photocatalytic conditions. 8 Noticeably, the existence of both C 6 F 5 I and C 6 F 5 Br exhibited a comparable reactivity, and the conversion of only C 6 F 5 I to C 6 F 5 H was observed by 19 F NMR spectroscopy (72%, entry 3, see Fig. S2 †). The effect of the counteranion was not observed in either case (Table 4, entries 4 and 5), suggesting that halide anions did not affect the reactivity. CV measurements revealed that the single-electron reduction of C 6 F 5 I (E red p ¼ À1.36 V vs. SCE in CH 3 CN) is more facile than that of Table 3 Utilization of bioactive amines a a Reaction conditions: A (0.2 mmol), 3DPAFIPN (0.012 mmol), C 6 F 5 I (0.8 mmol), K 2 CO 3 (0.6 mmol), CH 2 Cl 2 (1 mL) in a 4 mL reaction vial, irradiated with a 34 W Kessil blue LED for the indicated time under fan cooling. Yields of the isolated products were described.</p><p>Fig. 1 Major mechanistic questions regarding the reaction.</p><p>C 6 F 5 Br (E red p ¼ À1.64 V vs. SCE in CH 3 CN), 14 which may account for the difference in the reactivity of the two C 6 F 5 radical precursors. The addition of a radical scavenger ((2,2,6,6tetramethylpiperidin-1-yl)oxyl (TEMPO) or galvinoxyl) unambiguously hampered the reaction, indicating the involvement of radical intermediates in the reaction (Table 4, entries 6 and 7).</p><p>Evaluation of the effect of 3DPAFIPN (Table 1, entries 1 and 10) indicated the formation of a radical intermediate by photoexcited 3DPAFIPN. Stern-Volmer quenching experiments with A1 and C 6 F 5 I (Fig. S6 †) provided insight into the reductive quenching process (Scheme 2A, le). In addition, the involvement of HAT of the C 6 F 5 radical was substantiated by the observation of C 6 F 5 H aer the reaction (Fig. S1 †). However, the detailed aspects of HAT should be discussed by considering a number of possible scenarios. In particular, the target of HAT by the C 6 F 5 radical should be addressed. First, it is possible for the reactive C 6 F 5 radical to abstract activated hydrogen-atom from the neutral amine (A) (Scheme 2A, right). When amine A1 deuterated at the allylic position (A1-D) was applied in the reaction, only a trace amount of C 6 F 5 D was observed in the reaction mixture (Scheme 2B, Fig. S7 and S8 †), which is in clear contrast with the results of signicant deuterium incorporation in the hydrodeuorination of peruoroarene with an a-deuterated aliphatic amine. 15 This result strongly suggests that the direct HAT of A1 by the C 6 F 5 radical is not operative, considering that HAT with A1 would occur at an allylic a-amino C-H bond due to the cumulative stabilization effect of the nitrogen atom and the olen (Scheme S1 †). Next, the standard reaction was conducted with A1 using CD 2 Cl 2 as the solvent (Scheme 2C). Notably, no C 6 F 5 D was detected by 2 H NMR spectroscopy, suggesting that the formation of a dichloromethyl radical by HAT between the C 6 F 5 radical and CH 2 Cl 2 does not occur in the reaction mixture. 16 From the experimental observations, together with the reductive quenching between 3DPAFIPN and A1, it is expected that the generated amine radical cation (A1-radcat) would expel the allylic proton (deprotonation) to generate the a-amino radical, which undergoes HAT at the C4 position with the C 6 F 5 radical to generate the cyclic dienamine intermediate (A1-diene) (Scheme 2D). The involvement of A1-radcat as an intermediate would result in selective functionalization of the allylic structures, which can provide a more stable allylic a-amino radical (A1-rad) than the radical intermediate (A1-rad 0 ) derived from deprotonation of the benzylic position. The thermodynamic preference of A1-rad to A1-rad 0 in 3 kcal mol À1 would result in a >150 times higher concentration of A1-rad in equilibrium. The same trend is expected in the corresponding transition states of the deprotonation steps. 17 The observed solvent effect in Table 1 could be derived from the thermodynamically preferred deprotonation, which is favoured in dichloromethane, as recently reported by Liu and Ready. 17 The formation of double bonds via HAT in radical species has been demonstrated with a Co complex, 18 a tert-butoxy radical, 4b and TEMPO. 19 Because the reactive C 6 F 5 radical can be generated only aer the initial a Reaction conditions: A1 (0.2 mmol), 3DPAFIPN (0.012 mmol), additive, K 2 CO 3 (0.6 mmol), CH 2 Cl 2 (1 mL) in a 4 mL reaction vial, irradiated with a 34 W Kessil blue LED for 24 h under fan cooling.</p><p>b Measured by GC using dodecane as an internal standard.</p><p>Scheme 2 Mechanistic studies of HAT process.</p><p>reductive quenching of the photoexcited 3DPAFIPN by A1, such a HAT event occurs in a highly selective manner between the C 6 F 5 radical and A1-rad. This order of the elementary steps would be key to driving an array of oxidation steps toward the desired aromatization process. HAT at the C6 position, which provides a different type of diene (A1-diene 0 ), is a thermodynamically less favoured pathway (Scheme S2 †), and SET between the C 6 F 5 radical and A1-radcat would not be viable, based on the calculated SET barrier of the reaction (31.8 kcal mol À1 , Scheme S2 †). Unfortunately, a number of attempts to trap the proposed diene via Diels-Alder trapping with dienophiles were not successful (Scheme S6 †), presumably due to the facile oxidation of A1-diene (E ox (calc.) ¼ 0.43 V vs. SCE in CH 3 CN) by 3DPAFIPN for further transformations and deconstruction of the catalytic cycle by the external dienophiles.</p><p>The reaction with in situ generated imine species (A1-imine) gave only trace amount of B1, 20 implying that the formation of the imine intermediate from A1-rad via HAT or A1-radcat via halogen-atom transfer and deprotonation is not operating in this reaction. 21 The synergistic SET and HAT processes could proceed again on the generated A1-diene, performing deprotonation on C5 position and HAT on C6 position, to provide the aromatized product (B1). The involvement of the proposed HAT pathway was further supported by the reaction with A1-D 3 (Scheme 2E), which provided about 1 : 1 ratio of C 6 F 5 H and C 6 F 5 D as a result of sequential HAT processes on C4 position of A1-D 3 and C6 position of the corresponding diene intermediate.</p><p>The developed reaction provides an N-aryl amine as a product, but over-reactions of the generated N-aryl amine, such as the formation of corresponding imines, were not observed. Because the proposed mechanisms involving SET and HAT are also possible with N-aryl amines as reactants (Fig. S6 †), it was envisaged that an additional mechanism that discriminates the starting alkyl allyl amines (A) and the products (B) may be operative in the reaction. Insight into this aspect of the mechanism was gained from the reactivity of an aniline-derived allylic amine substrate (A37), which has a lower oxidation barrier (E ox p ¼ 0.88 V vs. SCE in CH 3 CN) than that of the model amine substrate (A1, E ox p ¼ 1.29 V vs. SCE in CH 3 CN) (Scheme 3A). The conversion of A37 and the formation of C 6 F 5 H were very low under the optimized reaction conditions (Fig. S3 †), implying that the productive catalytic cycle (Scheme 3A, orange arrow) did not operate efficiently. From the proposed mechanism that includes deprotonation of the amine radical cation, it was hypothesized that the stability of the amine radical cation would affect the reaction kinetics. Specically, we focused on the possibility of back-electron transfer (BET) between the reduced form of 3DPAFIPN (PCc À ) and the amine radical cation (Scheme 3A, green arrow).</p><p>BET affects the reactivity of a number of visible-light photoredox catalysis, and the control of this elementary step has been elaborated to improve the reactivity and selectivity. 22 When the triplet-state photoexcited 3DPAFIPN is quenched by the amine substrate, the radical ion-pair is immediately generated from two neutral species. Immediate and facile deprotonation of the amine radical cation then occurs via the desired catalytic cycle (Scheme 3A, orange arrow) under basic conditions (K 2 CO 3 ). 23 However, in the case of the N-aryl amine radical cation, which is known to have higher stability than the aliphatic amine radical cation and hence undergoes slower deprotonation, 24 BET can regenerate the starting amine and 3DPAFIPN (Scheme 3A, green arrow). This unproductive catalytic cycle clearly excludes the N-aryl substrate from the productive cycle, while no consumption of any reagents or the catalyst happens. The concept of BET in product protection is well established in the photocatalytic oxygenation of benzene to phenol, where the electron-rich phenol does not undergo oxidation due to the facile BET between the aryl radical cation and the photocatalyst. 22a,22b To verify the effect of the rate of deprotonation of various amine radical cations on the reaction, an array of N-aryl amines bearing ortho-substituents and N-substituents were prepared. Dinnocenzo and co-workers demonstrated the stereoelectronic effect of N-Ar substituents in the deprotonation of N-aryl amine radical cations. 24c Incorporation of a substituent at the orthoposition of the aryl group led to structural distortion through the N-C 1 bond, and the structurally-disfavoured conjugation induced destabilization of the amine radical cation and thus faster deprotonation (Scheme 3B). 24c Inspired by this stereoelectronic effect on the rate of deprotonation of amine radical cations, the correlation between the amine reactivity and the dihedral angle (:C 2 C 1 NR N ) of the amine radical cation was investigated with independently prepared N-aryl amine substrates (Scheme 3C). Notably, aryl allyl amines with large dihedral angles in the corresponding amine radical cations (A38-Me and A39-Me) were reactive toward CD, producing N,Ndiaryl amines (B38-Me and B39-Me) in noticeable yields (28% and 34%, respectively). These results support the involvement of BET in the reaction by demonstrating that preventing BET could initiate the productive catalytic cycle (Scheme 3A, orange arrow). In the synthesis of N-aryl amines from aliphatic allylic amines, BET is highly advantageous for achieving the targeted synthesis in an efficient and selective manner by preventing further undesired transformation of the product B (Scheme 3D).</p><p>Based on mechanistic studies and previous literature, we propose an overall mechanism for the photoredox CD of 2-cyclohexenyl amines (Fig. 2A). Transformation of the allylic amine substrate (A) is initiated by single-electron oxidation by the tripletstate photocatalyst ( 3 PC) to generate the amine radical cation (Aradcat), which undergoes facile deprotonation to produce the aamino radical species (A-rad). The pentauorophenyl radical ðC 6 F 5 Þ, which is generated by the subsequent reduction of C 6 F 5 I by the reduced form of the photocatalyst (PCc À ), abstracts a hydrogenatom from another allylic position of A-rad, forming the dieneamine intermediate (A-diene). With this intermediate, a similar mechanistic scenario is possible, considering the redox potential of the dienamine derived from amine A1 (A1-diene, E ox (calc.) ¼ 0.43 V vs. SCE in CH 3 CN). The generated dienamine radical cation (Adieneradcat) undergoes the facile deprotonation at the C5 position (DDG (C5 and C6) ¼ À6.5 kcal mol À1 , Scheme S3 †) to provide the 7p-enaminyl radical, as previously reported in the production of the 5p-enaminyl radical. 25 Further HAT by the C 6 F 5 radical of the highly conjugated enaminyl radical intermediate (A-dienerad) provides an aromatized product (B). The involvement of a chain-propagating mechanism for the generation of radical intermediates was excluded, based on the low quantum yield of 0.08 (Fig. 2B).</p><p>Throughout the photoredox catalytic cycle, both the productive pathway (Fig. 2B, black arrow) and the non-productive pathway (Fig. 2B, green arrow) are operative, leading to discriminatory activity where the substrates react and the product is preserved in the reaction media. The possibility of halogen bonding between the substrate amine (A) and C 6 F 5 I was excluded by the absence of noticeable interaction in the 19 F NMR spectra of the mixture (Fig. S15 †), 26 along with the UV-Vis observation of that the direct SET via the donor-acceptor complexation and photoinduced electron transfer (PET) do not likely operate in the reaction (Fig. S16 †). 27</p><!><p>In conclusion, the visible-light photoredox catalytic dehydrogenation of allylic amines was developed to synthesize invaluable N-aryl amines. Synergistic single-electron transfer and hydrogen-atom transfer enabled controlled dehydrogenation of the 2-cyclohexenyl amines bearing several reactive sites, thus affording a wide range of N-aryl amines under mild reaction conditions. Notably, the protective back-electron transfer pathway plays a crucial role in achieving the photoredox catalytic dehydrogenation protocol for the synthesis of N-aryl amines by preventing unwanted side-reactions.</p><!><p>There are no conicts to declare.</p>

Royal Society of Chemistry (RSC)

Self-Assembly and Regrowth of Metal Halide Perovskite Nanocrystals for Optoelectronic Applications

ConspectusOver the past decade, the impressive development of metal halide perovskites (MHPs) has made them leading candidates for applications in photovoltaics (PVs), X-ray scintillators, and light-emitting diodes (LEDs). Constructing MHP nanocrystals (NCs) with promising optoelectronic properties using a low-cost approach is critical to realizing their commercial potential. Self-assembly and regrowth techniques provide a simple and powerful “bottom-up” platform for controlling the structure, shape, and dimensionality of MHP NCs. The soft ionic nature of MHP NCs, in conjunction with their low formation energy, rapid anion exchange, and ease of ion migration, enables the rearrangement of their overall appearance via self-assembly or regrowth. Because of their low formation energy and highly dynamic surface ligands, MHP NCs have a higher propensity to regrow than conventional hard-lattice NCs. Moreover, their self-assembly and regrowth can be achieved simultaneously. The self-assembly of NCs into close-packed, long-range-ordered mesostructures provides a platform for modulating their electronic properties (e.g., conductivity and carrier mobility). Moreover, assembled MHP NCs exhibit collective properties (e.g., superfluorescence, renormalized emission, longer phase coherence times, and long exciton diffusion lengths) that can translate into dramatic improvements in device performance. Further regrowth into fused MHP nanostructures with the removal of ligand barriers between NCs could facilitate charge carrier transport, eliminate surface point defects, and enhance stability against moisture, light, and electron-beam irradiation. However, the synthesis strategies, diversity and complexity of structures, and optoelectronic applications that emanate from the self-assembly and regrowth of MHPs have not yet received much attention. Consequently, a comprehensive understanding of the design principles of self-assembled and fused MHP nanostructures will fuel further advances in their optoelectronic applications.In this Account, we review the latest developments in the self-assembly and regrowth of MHP NCs. We begin with a survey of the mechanisms, driving forces, and techniques for controlling MHP NC self-assembly. We then explore the phase transition of fused MHP nanostructures at the atomic level, delving into the mechanisms of facet-directed connections and the kinetics of their shape-modulation behavior, which have been elucidated with the aid of high-resolution transmission electron microscopy (HRTEM) and first-principles density functional theory calculations of surface energies. We further outline the applications of assembled and fused nanostructures. Finally, we conclude with a perspective on current challenges and future directions in the field of MHP NCs.

self-assembly_and_regrowth_of_metal_halide_perovskite_nanocrystals_for_optoelectronic_applications

<!>Introduction<!>Self-Assembly of MHP NCs<!><!>Self-Assembly of MHP NCs<!>Phase Transition, Morphological Evolution, and Mechanism of MHP NC Regrowth<!><!>Phase Transition, Morphological Evolution, and Mechanism of MHP NC Regrowth<!>Applications of Assembled and Regrown MHP Nanostructures<!><!>Applications of Assembled and Regrown MHP Nanostructures<!><!>Applications of Assembled and Regrown MHP Nanostructures<!><!>Applications of Assembled and Regrown MHP Nanostructures<!>Conclusion and Outlook<!>Special Issue<!>Author Contributions<!>

<p>LiuJ.; SongK.; ShinY.; LiuX.; ChenJ.; YaoK. X.; PanJ.; YangC.; YinJ.; XuL.-J.; YangH.; El-ZohryA. M.; XinB.; MitraS.; HedhiliM. N.; RoqanI. S.; MohammedO. F.; HanY.; BakrO. M.Light-Induced Self-Assembly of Cubic CsPbBr3 Perovskite Nanocrystals into Nanowires. Chem. Mater.2019, 31, 6642–6649.1In this work, the light-induced synthesis of CsPbBr3 nanowires through the regrowth of nanocrystals (NCs) was studied, with a systematic investigation of their phase transition, shape evolution, anisotropic growth mechanism, and growth preference.</p><p>PanJ.; LiX.; GongX.; YinJ.; ZhouD.; SinatraL.; HuangR.; LiuJ.; ChenJ.; DursunI.; El-ZohryA. M.; SaidaminovM. I.; SunH.-T.; MohammedO. F.; YeC.; SargentE. H.; BakrO. M.Halogen Vacancies Enable Ligand-Assisted Self-Assembly of Perovskite Quantum Dots into Nanowires. Angew. Chem., Int. Ed.2019, 131, 16223–16227.2This work explored a halide-vacancy-driven regrowth mechanism of metal halide perovskite (MHP) NCs and provided insights into the corresponding defect-correlated dynamics and defect-assisted fabrication of devices.</p><p>ZhangY.; SunR.; OuX.; FuK.; ChenQ.; DingY.; XuL. J.; LiuL.; HanY.; MalkoA. V.; LiuX.; YangH.; BakrO. M.; LiuH.; MohammedO. F.Metal Halide Perovskite Nanosheet for X-ray High-Resolution Scintillation Imaging Screens. ACS Nano2019, 13, 2520–252530721023.3This work reported the self-assembly of MHP nanosheets for application in X-ray high-resolution scintillation detectors.</p><p>LiuJ.; SongK.; ZhengX.; YinJ.; YaoK. X.; ChenC.; YangH.; HedhiliM. N.; ZhangW.; HanP.; MohammedO. F.; HanY.; BakrO. M.Cyanamide Passivation Enables Robust Elemental Imaging of Metal Halide Perovskites at Atomic Resolution. J. Phys. Chem. Lett.2021, 12, 10402–1040934672588.4This work reported the ligand-induced regrowth of CsPbBr3 NCs into nanoplates via an interface-assisted technique. The obtained nanoplates exhibited ultrahigh stability against electron irradiation and were elementally mapped via atomic-resolution X-ray energy dispersive spectroscopy.</p><!><p>Constructing nanomaterials with a desired structure and function is an aim of nanotechnology. The spontaneous arrangement of individual components into organized structures, that is, self-assembly and regrowth, is one of the most facile approaches for achieving this goal.5−7 The self-assembly of nanocrystals (NCs) into larger, long-range-ordered macroscopic arrays,8,9 superlattices,10,11 and larger crystals12 can result in a set of unique properties, including enhanced mechanical strength,13 electronic couplings,14 improved charge-carrier transport,15−17 and superior stability,18 compared with those of the individual constituents of such organized structures. Therefore, these macroscopic assembled nanostructures have stimulated the development of a wide range of applications in optoelectronic and thermoelectric devices and catalysis.15,19 Moreover, the bottom-up self-assembly and regrowth strategy provides a simple but effective platform for producing diverse NC ensembles, in contrast to top-down techniques that require elaborate facilities and produce limited structures.5,9</p><p>Unlike traditional chalcogenide NCs, metal halide perovskite (MHP) NCs are soft ionic materials that possess unique features,20−23 such as highly dynamic surface ligands,24,25 rapid anion exchange,26−28 and ease of ion migration,29 which facilitate their regrowth and the rearrangement of their overall appearance (Scheme 1). Consequently, the regrowth of MHP NCs can occur after the self-assembly process.1,30 The organization and fusion of MHP NCs, which are ideally suited for self-assembly and regrowth, into targeted nanostructures has been pursued as one method to modulate their optoelectronic properties. One example is self-organized three-dimensional (3D) superlattices, which exhibit key signatures of superfluorescence with red-shifted photoluminescence (PL)31,32 and a long exciton diffusion length.3,33 These closely packed superlattices with long-range order accommodate a high density of exciton states of low energetic disorder and a long dephasing time, enabling the construction of macroscopic quantum states.34 Importantly, the electronic properties of NCs, such as their conductivity and carrier mobility, can be substantially modulated when they are assembled into close-packed structures with strong NC coupling, which facilitates charge transport and is beneficial for fabricating high-performance devices. Moreover, the fused MHP nanostructures can lower the defect density and thus exhibit considerably enhanced stability against moisture,35,36 light,37 and electron-beam irradiation4,38 compared with their individual NC counterparts, offering a route for improving the inherent vulnerability that plagues MHPs. Consequently, these nanostructures have been employed in fabricating light-emitting diodes (LEDs),32 X-ray scintillators,3 and lasers34 and have potential applications in nanoantennas39 and photoelectric-compatible quantum processors.34 However, MHP self-assembly and regrowth have not been as intensively studied as MHP NC syntheses. Thus, the mechanisms of facet-directed connections and shape modulation have not yet been clearly elucidated. Therefore, many of the finer details regarding the self-assembly and fusion mechanisms of NCs and their unique roles in optoelectronic devices remain unclear. Summarizing the current progress in MHP NC self-assembly and regrowth and their property implications would be beneficial in stimulating further research efforts toward realizing the full potential of these highly ordered materials in optoelectronic applications.</p><p>In this Account, we review the following topics concerning advances in the self-assembly and regrowth of MHP NCs: the mechanisms, driving forces, and techniques for controlling the assembly process; morphological and phase evolution at the atomic level; investigations of oriented attachment mechanisms; the potential applications of assembled and fused nanostructures. We conclude by discussing the current challenges facing this field and forecasting possible opportunities.</p><!><p>Self-assembly is a spontaneous process that organizes individual components into orderly nanostructures, e.g., one-dimensional (1D) superlattice chains,1 two-dimensional (2D) layered superlattices,3 and 3D superlattices,27 as shown in Figure 1b–d. The self-assembly is driven by NC–NC interactions, including van der Waals forces between inorganic cores and between surface ligands, as well as osmotic, electrostatic, and elastic contributions.5 The balance of these forces can be illustrated by the effective interparticle pair interaction potential, U (Figure 2a). In colloidal nanoparticle solutions, the repulsive potential dominates and favors the monodispersion of the nanoparticles (Figure 2a, dark-green trace). During the self-assembly process, the effective interparticle interaction changes from repulsive to attractive (Figure 2a, light-green trace). The total removal of solvent results in the curdling of the NCs into a superlattice, with a balance between ligand elastic repulsion and van der Waals attractive forces.</p><!><p>(a) Schematic of the self-organization of MHP colloidal NCs into highly ordered superlattices (b–d) and further regrowth into large, bulky crystals (e–g). (b) One-dimensional (1D) superlattice chains. Reproduced with permission from ref (1). Copyright 2019 American Chemical Society. (c) Two-dimensional (2D) layered superlattices. Reproduced with permission from ref (3). Copyright 2019 American Chemical Society. (d) Three-dimensional (3D) superlattice. Reproduced with permission from ref (31). Copyright 2018 Nature Publishing Group. (e) Nanowires. Reproduced with permission from ref (1). Copyright 2019 American Chemical Society. (f) Nanoplates. Reproduced with permission from ref (4). Copyright 2021 American Chemical Society. (g) Nanocuboids. Reproduced with permission from ref (40). Copyright 2019, Wiley.</p><p>(a) Evolution of the effective pair interaction potential, U, at different self-assembly stages. Reproduced with permission from ref (5). Copyright 2016 American Chemical Society. (b) Scheme depicting the capillary forces associated with solvent evaporation. (c) Schematic illustration of the formation of self-assembled 3D CsPbBr3 superlattices by ultrasonication. Reproduced with permission from ref (32). Copyright 2018 Wiley. (d–f) TEM images of binary ABO3-type (d), binary NaCl-type (e), and ternary ABO3-type (f) superlattices from MHP nanocubes. Reproduced with permission from ref (47). Copyright 2021 Nature Publishing Group.</p><!><p>van der Waals forces are speculated to be the most dominant interaction at the nanoscale, usually acting in a manner that brings particles together. As one of the three types of van der Waals forces, dipole–dipole interactions dominate the initial stage of self-assembly because of the long interaction distance of dipolar attractions.41 Inorganic MHPs (i.e., CsPbX3), in which a dipole moment should not intrinsically exist, have exhibited a perfect distortion-free cubic structure.1 Electric polarization can emerge from the disruption of crystal symmetry, which is a process that involves the migration of A cations, the movement of B cations away from the center of the BX6 octahedra, and distortion of the BX6 octahedra, all of which can induce a dipole moment.42 When a polar solvent (i.e., ethanol) induces CsPbI3 lattice distortion, the adsorption of polar molecules causes the migration of Cs+ as well as distortion of the PbI6 octahedra, resulting in the breaking of symmetry and the polarization of the CsPbI3 NC.43 When a third polarized CsPbI3 NC approaches two polarized CsPbI3 NCs, the arrangement along the rectilinear direction would create the smallest gradient of the dipole potential field, which accounts for linear alignment into a 1D superlattice chain. This characteristic is conspicuous in organic–inorganic hybrid perovskites (i.e., MAPbX3), in which the asymmetry of organic cations results in the absence of an inversion center in the structure.42,44</p><p>Self-assembly of NCs is commonly triggered by solvent evaporation or by varying the polarity of the reaction system and destabilizing the NC capping ligands. Figure 2b shows the preparation of assembled nanostructures by evaporation of the colloidal solvent. During solvent evaporation, the interparticle distance decreases, and NCs can potentially assemble in an orderly manner to maximize the total entropy of the system. A strong capillary interaction is further exerted to drive parallel alignment between neighboring NCs (Figure 2b).19 Once the NCs are closely spaced, they start to align and stack at the interface to form superlattices. The formation of a superlattice via drying-mediated self-assembly can be modulated by (i) the initial NC concentration, (ii) the temperature of solvent evaporation, and (iii) the concentration of the capping ligands.45 For example, the addition of oleic acid and oleylamine (OAm) can passivate bare NC surfaces, prevent solvent dewetting, and stimulate depletion attraction, thereby assisting in the formation of CsPbBr3 NC superlattices,31 where the building blocks (CsPbBr3 nanocubes) are atomically aligned exclusively via the four vertical {100}c facets46 (subscript "c" represents cubic crystal structure). The solvent evaporation technique has also been extended to the construction of binary and ternary superlattices via coassembly of cubic CsPbBr3 NCs and other types of NCs (e.g., Fe3O4, NaGdF4, and PbS),47 enriching the membership of MHP ensemble families (Figure 2d–f).</p><p>The assembled nanostructures are strongly affected by the interactions between the solvent and the capping ligands.48,49 By exploiting the dynamic ligand–surface interaction of MHPs and their sensitivity to solvent polarity, researchers have used appropriate solvents to trigger self-assembly. In these assemblies, solvent selection is critical, and a uniform stable dispersion of NCs is undesirable. For example, CsPbBr3 NCs can disperse stably in toluene but tend to arrange into chain-like assemblies in hexane.50 This arrangement occurs because when MHP NCs are dispersed in a nonpolar solvent such as hexane, the excess aliphatic ligand complex with ionic species forms via oleophilic interactions, and alkyl ligand chains connect to one another through strong van der Waals interactions, resulting in the self-assembly of CsPbBr3 NCs into 1D superlattice chains.</p><p>In addition to the aforementioned investigations of the contributions of solvent evaporation and polarity, efforts have been dedicated to investigating the anisotropic and isotropic assembly of MHP NCs, including surfactant interactions,51,52 external forces32,53 (e.g., sonication in Figure 2c), and template-assisted assembly.54 For example, prepatterned polydimethylsiloxane templates were used for the template-induced self-assembly of CsPbBr3 NCs into large-area 2D supercrystals.54</p><p>Because of the inherently soft ionic nature of the MHP crystal structure and low formation energy, the obtained MHP superlattices tend toward continuous regrowth into large crystals (Figure 1e–g). Although the ligands render MHP NCs resistant to aggregation, these ligands loosely attach to the surface of MHP NCs because of the highly dynamic binding between the surface-capping ligands and the oppositely charged NC surface ions. Therefore, in contrast to most conventional NCs (e.g., chalcogenide), whose self-assembly preferentially ends with superlattices,45 MHP NCs self-assemble and regrow simultaneously into different nanostructures (e.g., Figure 1b,e)1,30 but are particularly prone to regrowth.</p><!><p>Regrowth is commonly driven by thermodynamics and occurs when NCs are physically attached. Surface atoms usually exhibit higher chemical reactivity than interior atoms because of the considerable number of dangling bonds between surface atoms. Removing surface ligands exposes NC surfaces, and this process is associated with a substantial increase in the attraction potential among NCs; consequently, the NC units in a superlattice can continue to grow (Scheme 1).</p><p>The regrowth of MHP derivatives is usually accompanied by a phase transformation. However, the atomic structure of MHP NCs remains poorly understood. For example, whether the crystal structure of CsPbBr3, which is the most prevalent type of MHP NC, is cubic or orthorhombic remains a topic of debate.55,56 The orthorhombic phase evolves from a slight tilting of the PbBr6 octahedra in the cubic structure, which preserves the 3D network of corner-sharing octahedra while introducing structural differences between axially and equatorially coordinated halides. This small tilting of the PbBr6 octahedra cannot be distinguished by powder X-ray diffraction analysis. Aberration-corrected scanning transmission electron microscopy (STEM) has thus been used to elucidate the atomic details of the crystal structure of CsPbBr3 NCs. The high-angle annular dark-field (HAADF)-STEM image shown in Figure 3b indicates that CsPbBr3 NCs are of a cubic phase (ICSD 29073; Pm3̅m (221); a = 0.5874 nm) and exhibit truncated cubic shapes. During the regrowth process, the initial cubic CsPbBr3 NCs undergo a phase transformation from the cubic phase to the thermodynamically more stable orthorhombic phase (ICSD 97851, Pbnm (62), a = 0.8207 nm, b = 0.8255 nm, c = 1.1759 nm), as displayed in Figure 3c. A similar phenomenon has been observed among iodized derivatives.43</p><!><p>(a) Geometrical relationship between cubic and orthorhombic unit-cell axes and faces. (b) Atomically resolved high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM) image of cubic-phase CsPbBr3 NCs viewed along the [001]c zone axis and (c) regrowth of orthorhombic-phase CsPbBr3 bulk crystals viewed along the [11̅0]o zone axis. (d–f) High-resolution transmission electron microscopy (HRTEM) images showing the coalescence of CsPbBr3 NCs via oriented attachment along the [110]o (d), [001]o (e), and [100]o (f) crystallographic directions. (g–i) Investigation of the vacancy distribution in initial CsPbBr3 NCs. Inset is a simulated HRTEM image of CsPbBr3. Red circles represent Br vacancies. Panels b–e reproduced with permission from ref (1). Copyright 2019 American Chemical Society. Panels f–i reproduced with permission from ref (4). Copyright 2021 American Chemical Society.</p><!><p>With respect to the isotropic cubic phase, understanding the alignment of the orthorhombic structural axes is paramount for elucidating their anisotropic shape evolution (e.g., into nanowires and nanoplates). Orthorhombic CsPbBr3 NCs are modeled as possessing four side {110}o facets (subscript "o" represents orthorhombic crystal structure), two bottom {001}o facets, and 12 edge facets (four {100}o and eight {112}o facets).57,58 Density functional theory (DFT) calculations indicate that both the {001}o and {110}o surfaces of CsPbBr3 NCs are terminated with a CsBr surface.59 However, the surface Cs+ ions are inclined to be replaced with surfactants;60 this replacement process, combined with abundant surface Br– vacancies,61 leads to the possible exposure of PbBr2 termination. Undoubtedly, the coexistence of CsBr- and PbBr2-terminated surfaces ensures continuous regrowth, which has been confirmed by atomic-resolution STEM.1 The relatively higher surface energy of the {100}o surfaces terminated with CsPbBr2+ and Br22– facets makes these surfaces less stable than the {001}o and {110}o surfaces (surface energies {100}o > {110}o > {001}o).4 Different surface-atom configurations of the NC surfaces are particularly important because they can determine the preferred growth direction.</p><p>Oriented attachment (OA), one of the most important mechanisms for controlling NC growth, is becoming a prevalent approach for controlling the design of nanostructures. NCs in a dispersed colloidal solution collide frequently because of Brownian motion; however, not all of these collisions result in NC attachment. Only NCs that share a common crystallographic orientation can undergo an effective collision.62 Otherwise, the NCs undergo continuous rotations until they match with a perfect lattice.63 Afterward, the NCs undergo oriented attachment at the contacted facets with a common crystallographic orientation.62 As illustrated in Figure 3d–f, MHP NCs can coalesce in various ways, including face-to-face (e.g., [110]o or [001]o direction), edge-to-edge (e.g., [100]o direction), and even corner-to-corner.1,4,64 Even if the {100}o surfaces are thermodynamically less supported, the electrostatic interaction that originates from their charged character (facets terminated with CsPbBr2+ and Br22–) can promote the likelihood of edge-to-edge coalescence. Long-distance interactions (e.g., van der Waals forces and Coulombic interactions) are critical to bringing nanoparticles sufficiently close together for OA.65 For instance, in a colloidal solution, if two MHP NCs are far away from each other, the primary driving force for OA is van der Waals interactions, while interatomic Coulombic interactions are negligible in comparison because the interatomic Coulombic interactions are screened. By contrast, when two MHP NCs are in close proximity, electrostatic interactions become dominant, driving the NCs to approach each other and eventually fuse.65,66 From a thermodynamic viewpoint, the fusion in a coherent crystallographic orientation eliminates the interfaces of the NCs; thus, a reduction in surface energy is assumed to be the thermodynamic driving force.67 The total energy change of MHPs, as soft ionic materials, is largely derived from the interatomic Coulombic interactions arising from both surface and interior atoms, where the former also contribute to a reduction in surface energy.66 Simulation68 and transmission electron microscopy (TEM) results revealed that the presence of abundant Br vacancies in MHP NCs (Figure 3g–i) may further promote the intrinsic Coulombic interactions.4 Because of dangling Pb bonds and a substantial concentration of vacancies on the surface of MHP NCs, their surface atoms exhibit high chemical reactivity toward the absorption of ions and contribute to their regrowth into various microstructures.2</p><p>After NCs attach, the interface region starts to regrow until the interface boundary completely disappears. In addition to the perfect perovskite structure, planar defects (e.g., Ruddlesden–Popper (RP) planar faults, symmetric grain boundaries (GBs), or asymmetric GBs)69 can also be introduced at the surface where attachment occurs. Defects form mainly because of the relatively fast dynamics of attachment, in conjunction with the low formation energy of MHPs, where coalescence can occasionally occur even if the NCs have not reached identical crystallographic orientations. For example, the merging of two NCs with the same terminated surface leads to a RP planar fault. If two NCs with the same surface termination do not coalesce in parallel, a symmetric GB will form. If the NCs are fused with different surface terminations, an asymmetric GB will form.</p><p>Shape evolution is known to be a kinetic process in which high-energy surfaces grow faster than low-energy surfaces.70,71 According to DFT calculations, the {001}o facets maintain the densest atomic stacking mode and possess lower surface energy than the {110}o facets. Consequently, the orientation with {001}o facets and growth along the four [110]o directions is the thermodynamically most favorable shape-evolution mode and ensures maximally exposed, low-energy {001}o facets.4 Ideally, CsPbBr3 NCs evolve into nanoplates along the [110]o directions.4 However, the effects of capping ligands, external forces, or assembly techniques, for instance, could alter the growth trajectory and lead to the development of other morphologies. For example, the contribution of the surfactant OAm accounts for the anisotropy associated with the production of nanowires.1 The basal (largest) facets of the nanowires are two {001}o and two {110}o facets, which are side nanowire facets, whereas the two remaining bottom planes are {110}o planes. From a thermodynamic perspective, nanowires grow along the [110]o direction to maximally expose low-energy {001}o surfaces (Figure 3d). In addition, research on vacancy-assisted regrowth has revealed that the {100}o surfaces possess a much larger Br-vacancy (VBr) formation energy than the {110}o surfaces and that the higher vacancy density on the {110}o facets leads to the growth of nanowires along the [110]o direction (Figure 3d).2 Pradhan et al. reported that the direction of facet connection can be further tailored by controlling the reactant composition ratios.72 The inclination of Pb-rich (Br-deficient) NCs enables the connection of {100}o edge facets (Figure 3f), resulting in zigzag 1D nanowires. However, Pb-deficient (Br-rich) compositions promote merging along the {110}o facets, which indicates that Br– ions assist in the elimination of the {100}o active edge facets and allow for connection mostly via {110}o facets.72 These observations could seem in contrast to the aforementioned Br-vacancy-induced growth along the [110]o direction.2 Therefore, additional experimental and theoretical investigations (related to, for example, the selective adherence of capping ligands to facets) are needed to reconcile the various evident observations of MHP nanostructure formation. In general, ligands contribute to the growth kinetics, preferential orientation, and shapes associated with NC self-assembly. For PbS NCs, oriented attachment occurs exclusively via the (100) facets (those with lower surface energy) because oleate ligands are strongly bound to the (110) facets, preventing them from oriented attachment through that surface orientation.45 Unlike the strongly bound oleate ligands in PbS, the surface ligands in soft ionic perovskites are loosely attached and highly dynamic. This contrasting behavior offers a plausible explanation for the differing roles of ligands in the oriented attachment of MHP NCs (versus chalcogenide NCs), in which the higher-energy (100)o facets are still involved in oriented attachment.</p><!><p>The self-assembly and regrowth of MHPs has afforded their use in numerous promising applications, including the fabrication of high-performance LEDs,32,73 lasers,34,49 and X-ray scintillators,3 and they are appealing candidates for use in nanoantennas39 and photoelectric-compatible quantum processors.34 The self-assembly of NCs into close-packed assemblies results in more efficient electronic behaviors (e.g., improved conductivity and carrier mobility). In addition, the assembled MHP NCs exhibit novel optoelectronic properties (e.g., superfluorescence,31 renormalized emission,32 longer phase coherence times, and longer exciton diffusion lengths3,33) because of the electronic and physical coupling of NCs. Moreover, further regrowth into fused MHP nanostructures with the removal of ligand barriers between NCs could facilitate charge carrier transportation and enhance stability against moisture, light, and electron-beam irradiation by eliminating surface point defects, making them useful in practical applications. These attractive properties and concerns are addressed in this section, with a discussion of representative cases.</p><p>Assembled superstructures with novel properties arising from electronic coupling between NCs are expected to find applications in light-emitting devices. Compared with the individual NC building units, self-organized superlattices display obvious signatures of superfluorescence (short, intense bursts of light)31 (Figure 4a) because of the many-body quantum phenomenon induced by collective coupling; they also exhibit red-shifted narrowing emission31,32,74 (Figure 4b) with accelerated radiative decay (Figure 4d),31,34,75 longer phase coherence times, and fluorescence resonance energy transfer (FRET)-mediated long exciton diffusion lengths (Figure 4c).3,33</p><!><p>(a) Schematic of the build-up of superfluorescence in CsPbBr3 NC superlattice. Reproduced with permission from ref (31). Copyright 2018 Nature Publishing Group. (b) Optical properties and energy diagram of CsPbBr3 NCs and superlattices. Reproduced with permission from ref (32). Copyright 2018 Wiley. (c) Steady-state exciton diffusion measurement. Normalized profile of PL intensity emitted by sparse (top) and close-packed (bottom) MHP NC monolayer when excited with a diffraction-limited laser spot. Reproduced with permission from ref (33). Copyright 2019 American Chemical Society. (d) Time-resolved PL decay of uncoupled (blue) and coupled (dark red) NCs. Reproduced with permission from ref (31). Copyright 2018 Nature Publishing Group.</p><!><p>To further exploit such superfluorescence characteristics (Figure 4a), Zhou et al. developed a perovskite-based quantum-dot superlattice microcavity (QDSM) that exhibits cavity-enhanced superfluorescence behavior and an optically stimulated amplification effect, as displayed in Figure 5a.34 The typical Q-factor of a QDSM can reach ∼2000 (Figure 5b). During the QDSM lasing process, the cavity field in the QDSM further accelerates the superfluorescence process, leading to a picosecond-scale radiative time (Figure 5c). Moreover, the coherent nature of these exciton states, which arises from the strong mesoscopic coupling between NCs, can enable the development of entangled multiphoton quantum light sources and may allow for the application of QDSMs in ultrafast, photoelectric-compatible quantum processors. However, decoupled multiquantum-well 2D superlattices, which exhibit a high photoluminescence quantum yield (PLQY), narrowband emission, and enhanced light outcoupling, have also emerged as desirable candidates for many practical applications, such as LEDs and nanoantennas.39</p><!><p>(a) PL spectra showing cavity-enhanced superfluorescence from a quantum-dot superlattice microcavity (QDSM). (b) Power dependence of the PL intensity in cavity mode. (c) Radiation dynamics. Panels a–c reproduced with permission from ref (34). Copyright 2020 Nature Publishing Group. (d) Schematic showing the self-assembly of CsPbBr3 nanosheets. (e) Schematic showing the energy transfer process from thin to thick nanosheets, with a fluorescence resonance energy transfer (FRET) efficiency of 74%. (f) Photographic and X-ray images of a standard central processing unit panel with a Si chip integrated underneath. Panels d–f reproduced with permission from ref (3). Copyright 2019, American Chemical Society. (g) STEM image revealing the self-assembly of CsPbBr3 NCs with an atomic-scale interparticle distance. (h) Measurement of the exciton diffusion length. (i) Current–voltage traces and carrier mobility measurements. Panels g–i reproduced with permission from ref (76). Copyright 2020 Nature Publishing Group.</p><!><p>The self-assembly of MHP NCs into electronically coupled superlattices, resulting in the formation of a delocalized, extended electronic state, can renormalize their emission energy.74 These ensembles retain the high PL efficiency of their NC subunits, which can be exploited to develop a new set of electronic applications. For example, the assembled CsPbBr3 3D superlattices exhibit red-shifted emission at 535 nm (Figure 4b), whereas the individual NCs emit cyan-green light (515 nm), allowing them to overcome the "green gap" and thus enabling the fabrication of efficient pure-green LEDs that satisfy the Rec. 2020 standard.32</p><p>Because of FRET (Figure 4c), CsPbBr3 NC assemblies exhibit extremely efficient exciton diffusion, demonstrating a record diffusion length of 200 nm with a diffusivity of 0.5 cm2 s–1; this performance is substantially better than that of chalcogen-based NCs.3,33 Their self-assembly in close-packed systems promotes communication between neighboring NCs by enabling the FRET of excitons, which results in the transport of excitonic energy in multiple steps before the excitons recombine. The occurrence of FRET in assemblies is critical for enhancing optoelectronic device performance. For example, an X-ray high-resolution scintillator fabricated from self-assembled CsPbBr3 nanosheets (Figure 5d) demonstrated markedly enhanced scintillation performance because of the energy transfer process inside the stacked nanosheets (Figure 5e), with a FRET efficiency of 74%. Such a simple prototype enabled spatial resolution within 0.2 mm (Figure 5f).3</p><p>Self-assembly is an effective method of surface engineering to achieve high-performance photovoltaics (PVs) and LEDs. The self-assembly of NCs into densely packed assemblies affords another level of modification for tailoring electronic properties (e.g., conductivity and carrier mobility). For example, a solvent-assisted assembly strategy can be used to modulate the prototypical, long, insulated ligands via surface engineering, leading to a close-packed and smooth surface.77 Connecting the NCs after removal of the surfactant barriers could considerably improve charge injection and carrier transport.76,77 Consequently, a substantially boosted external quantum efficiency (EQE) has been achieved, indicating considerable enhancement in the properties associated with the CsPbBr3 emitting layer, such as carrier transport and radiative decay.77 Furthermore, assisted by surface-functionalized self-assembly, an ultrasmooth monolayer nanocube thin film with a root-mean-square roughness of ∼4 Å has been fabricated.78 The short-chain ligands introduced into the system enable more efficient charge transport and substantially reduce the NC–NC interactions, thereby greatly promoting self-assembly. Similarly, a bipolar-shell-resurfacing strategy has been proposed to stabilize CsPbBr3 NCs, where ligand exchange leads to close-packed films with a long diffusion length, high carrier mobility, and reduced trap density.76 The TEM image in Figure 5g shows the features of assembled CsPbBr3 NCs. The contribution from FRET in the assembled films and the reduced trap density originating from a resurfaced bipolar shell led to elongated exciton diffusion lengths of ∼70 ± 30 nm (Figure 5h) and improved carrier mobility (Figure 5i). Self-assembled MHP NCs with atomic-scale interparticle distances and a resurfaced bipolar shell yielded EQEs of 12.3% and 22% for blue and green devices, respectively.76</p><p>The inherent susceptibility of MHP derivatives to degradation remains a major obstacle to their practical application. Self-assembly and regrowth may offer a new avenue for overcoming their instability. For example, CsPbBr3 nanoplates obtained from the regrowth of their NCs exhibited ultrahigh stability under a 300 kV electron beam, and the first atomic-resolution X-ray energy dispersive spectroscopy elemental mapping data of MHPs were successfully acquired, as shown in Figure 6a.4,38 Similarly, fused CsPbBr3 nanoplates with RP defects (Figure 6b) showed substantially improved stability against UV light compared with as-synthesized NCs (see Figure 6c).37 Another example is the pressure-driven regrowth of MHP NCs into nanosheets (Figure 6d); compared with their NC units, these nanoplates demonstrate enhanced properties, including a higher PL intensity and remarkable resistance to water.35 The attenuation of surface defects and the disappearance of grain boundaries in perovskites by self-assembly and regrowth contribute to enhanced stability against moisture. In addition, a strategy was developed to fabricate 2D/3D perovskite films by the self-assembly of low-n-value 2D perovskite crystals with enhanced device stability against moisture.36 In summary, these results demonstrate that the stability of assembled or fused MHP derivatives is superior to that of unassembled NCs.</p><!><p>(a) HAADF-STEM image and atomic-resolution X-ray energy dispersive spectroscopy elemental mapping images of a CsPbBr3 nanoplate. Reproduced with permission from ref (4). Copyright 2021 American Chemical Society. (b) Atomic-resolution HAADF image of a Ruddlesden–Popper planar defect with an overlaid atomic model. (c) Normalized PL intensity dynamics of as-synthesized and fused CsPbBr3 NCs upon continuous ultraviolet (UV) light exposure. Panels b and c Reproduced with permission from ref (37). Copyright 2018 American Chemical Society. (d) Relative photoluminescence intensity of NCs and fused nanosheets immersed in water. Reproduced with permission from ref (35). Copyright 2020 American Chemical Society. (e) Performance of assembled nanowires for LEDs: a constant driving current of 6 mA (150 mA cm–2) led to an increase in luminance (L0) from 0 to 11500 cd m–2. The estimated operational half-lifetime (T50) at 100 cd m–2 was 694 h. Reproduced with permission from ref (73). Copyright 2020 Wiley. (f) PL lifetime measurements of CsPbBr3 NCs and the nanowires obtained through light-induced regrowth. Reproduced with permission from ref (1). Copyright 2019 American Chemical Society.</p><!><p>The formation of defect-free MHPs is key to the successful implementation of MHP nanostructures in optoelectronic devices.21 Surface point defects can be effectively self-healed during the self-assembly and regrowth process, resulting in low density of trap states in MHP nanostructures. For instance, self-assembled nanowire arrays are known to exhibit very high PLQY of 91% at a wavelength of 600 nm because of their low trap density and strong quantum confinement.73 This ultralow trap density could contribute to remarkable structural and environmental stability. As shown in Figure 6e, the fabricated LEDs based on these assembled nanowires exhibited a record luminance of 13644 cd m–2 with an EQE of 6.2%, along with substantially improved operational lifetimes (T = 13.5 min at 11500 cd m–2, T = 694 h at 100 cd m–2). In a similar study, fused highly crystalline nanowires demonstrated a carrier lifetime 2 orders of magnitude longer than that of the initial NCs (Figure 6f), which provides clear evidence that regrowth can reduce the defect concentration and thus hinder the nonradiative recombination in these fused CsPbBr3 nanowires.1</p><!><p>In this Account, we focused on advances in the self-assembly and regrowth of MHP NCs, along with their assembly and fusion mechanisms, driving forces, preparation strategies, and potential applications. Undoubtedly, self-assembly and regrowth is a facile and powerful approach for controlling the structure, shape, and dimensions of MHPs. The novel optoelectronic properties (e.g., superfluorescence) of assembled MHP NCs, their promoted electronic behaviors, and their enhanced stability against electron-beam irradiation and light facilitate the fabrication of high-performance devices. Therefore, exploring the self-assembly and regrowth behavior of MHP NCs is important, and numerous opportunities lie ahead.</p><p>Self-assembled MHP NCs exhibit stronger coupling than conventional NC films, which enhances optoelectronic device performance. The electronic coupling of MHP NCs enables the high PLQY of the individual NCs to be retained, which can facilitate charge-carrier injection and reduce the probability of trap-assisted recombination.</p><p>The successful study of MHP NC regrowth provides a path toward comprehensively understanding the nucleation and growth kinetics of perovskite materials. The soft ionic nature of MHPs enables them to exhibit fast nucleation and growth kinetics such that the intermediate states during the dynamic growth process are undetectable by normal reaction-tracking approaches; consequently, the mechanism by which MHP NCs are formed remains unclear. The self-assembly and regrowth of MHP NCs can dramatically slow NC growth kinetics, enabling the tracking of NC trajectories. In addition, recently developed ultralow-dose electron microscopy techniques79 enable atomic-resolution imaging of MHPs to provide routes toward addressing the aforementioned uncertainties and fully elucidate the orientation growth process.</p><p>The formation of electron irradiation-stable MHP NCs via fusion provides a platform for studying defect species in perovskite semiconductors. Sufficient stability achieved against electron beams provides additional opportunities for much deeper investigations using TEM, such as research into the types of defects and mechanism of defect formation at the atomic scale. Understanding the origin of defects in MHP NCs is paramount in attaining long-term structural stability and improved optical efficiency.</p><p>Obstacles to fully understanding the complexities of self-assembly persist. First, the driving forces that control the formation of MHP assemblies remain largely unexplored; a fundamental analysis of these forces would assist the development of diverse assembled nanostructures. Accordingly, more attention should be devoted to investigating the strongly coupled MHP assemblies, which may provide avenues for new developments in optoelectronic devices with properties that exploit efficient charge transport and enhanced conductivity in those assemblies. However, predicting appealing assembled structures and phases will likely involve well-understood simulations (e.g., Monte Carlo and molecular dynamics simulations) coupled with experimental verification.</p><!><p>This Account was originally intended to be part of the Accounts of Chemical Research special issue "Transformative Inorganic Nanocrystals", which was completed April 6, 2021.</p><!><p>∥ J.L. and X.Z. contributed equally to this work.</p><!><p>This work was supported by King Abdullah University of Science and Technology (KAUST).</p><p>The authors declare the following competing financial interest(s): O.M.B. is a founder of Quantum Solutions, a nanotechnology company that develops and manufactures quantum dot materials for optoelectronics.</p><p>Jiakai Liu received his Ph.D. in Materials Science and Engineering from King Abdullah University of Science and Technology (KAUST) in 2020. He is currently a postdoctoral researcher in the research group of Prof. Peigang Han and Prof. Wang Zhang at Shenzhen Technology University (SZTU). His research focuses on perovskite nanocrystal synthesis, self-assembly, and regrowth.</p><p>Xiaopeng Zheng received his Ph.D. in Materials Science and Engineering from KAUST in 2020. He is currently a postdoctoral researcher in Prof. Joseph M. Luther's research group at the National Renewable Energy Laboratory (NREL). His research focuses on perovskite photovoltaics and light-emitting diodes for clean-energy harvesting and displays.</p><p>Omar F. Mohammed is an Associate Professor of Materials Science and Engineering at KAUST. He received his Ph.D. degree in Physical Chemistry from Humboldt University of Berlin, Germany. His research group focuses on the study of ultrafast charge-carrier dynamics in photoactive materials with the aid of cutting-edge laser spectroscopy and ultrafast electron imaging.</p><p>Osman M. Bakr holds a B.Sc. in Materials Science and Engineering from MIT (2003) and an M.S. and Ph.D. in Applied Physics from Harvard University (2009). He is currently a Professor of Materials Science and Engineering at KAUST. His research group focuses on the study of hybrid organic–inorganic materials, particularly on advancing their synthesis and self-assembly for applications in optoelectronics and catalysis.</p>

PubMed Open Access

2^nd coordination sphere controlled electron transfer of iron hangman complexes on electrodes probed by surface enhanced vibrational spectroscopy

Iron hangman complexes exhibit improved catalytic properties regarding O 2 and H 2 O 2 reduction, which are attributed to the presence of a proton donating group in defined vicinity of the catalytic metal centre.Surface enhanced resonance Raman (SERR) and IR (SEIRA) spectro-electrochemistry has been applied concomitantly for the first time to analyse such iron hangman porphyrin complexes attached to electrodes in aqueous solution. While the SERR spectra yield information about the redox state of the central iron, the SEIRA spectra show protonation and deprotonation events of the 2 nd coordination sphere. To investigate the influence of a proton active hanging group on the heterogeneous electron transfer between the iron porphyrin and the electrode, two hangman complexes with either an acid or ester functional group were compared. Using time resolved SERR spectroscopy the electron transfer rates of both complexes were determined. Complexes with an acid group showed a slow electron transfer rate at neutral pH that increased significantly at pH 4, while complexes with an ester group exhibited a much faster, but pH independent rate. SEIRA measurements were able to determine directly for the first time a pK a value of 3.4 of a carboxylic hanging group in the immobilized state that shifted to 5.2 in D 2 O buffer solution. The kinetic data showed an increase of the heterogeneous electron transfer rate with the protonation degree of the acid groups. From these results, we propose a PCET which is strongly modulated by the protonation state of the acid hanging group via hydrogen bond interactions.

2^nd_coordination_sphere_controlled_electron_transfer_of_iron_hangman_complexes_on_electro

Introduction<!>Materials and methods<!>Results and discussion<!>SERR spectroscopy of immobilised hangman complexes<!>SEIRA spectroscopic probing of the hanging group<!>TR-SERR spectroscopic determination of the HET rate<!>Mechanistic implications of the effect of the hanging group<!>Conclusions

<p>Second coordination sphere assisted reactions are crucial for the efficiency of numerous catalytic transformations. For example, in nature a variety of different reactions are catalysed by heme cofactors where selectivity of the reaction is induced by acidic or basic amino acids in the coordination sphere around the heme environment. [1][2][3] This highly ordered arrangement of proton donating or accepting groups denes the catalysed reaction by the heme group, which can range from reduction of hydrogen peroxide to water (catalase), substrate oxidation (peroxidase), binding of molecular oxygen (myoglobin, hemoglobin), oxygen reduction to water (cytochrome c oxidase) or hydroxylation of different compounds (cytochrome P450).</p><p>Synthetic biomimetic molecular catalysts copy the essence of the reaction centres of their enzymatic analogues, and exploit the optimally evolved active structures for maximal performance. 4,5 In this respect, they exhibit numerous advantages compared to their biological idols. On the one hand, due to their smaller size, better substrate accessibility and higher stability, these compounds bear a high potential to be used in technological applications such as biomimetic fuel cells. On the other hand, the study of molecular catalysts is highly valuable in general. Well-dened synthetic catalysts allow detailed investigations of the catalytic mechanism at a molecular level and precise ne-tuning of desired catalytic activity using synthetic chemistry. Understanding the structure-function relationships of catalytically active sites can in turn enhance the knowledge of biological catalysis.</p><p>One of the challenges in catalyst design is to mimic the electron/proton transfer interplay that has been naturally optimized in enzymatic catalysis. In this regard, hangman porphyrin complexes that carry a heme group and an arbitrary functional "hanging" group positioned in a dened distance to the reaction centre, constitute an interesting model system to study the inuence of the 2 nd coordination sphere. [6][7][8] Hangman complexes that exhibit a hanging carboxylic acid group have been shown to signicantly enhance the catalase 9 and oxidase reaction in solution in comparison to complexes with nonacidic hanging groups. 10,11 For similar iron hangman corroles, a catalase like reaction mechanism has been proposed that involves the carboxylic acid group as a proton donor site. 12 Cobalt hangman porphyrins and corroles have also been successfully tested in electrocatalytic dioxygen reduction and hydrogen evolution. [13][14][15] For a technological application in fuel cells, the hangman complexes have to be immobilized on an electrode surface. In contrast to homogeneous reactions, the adsorption provides numerous advantages such as site isolation of catalytically active centres, facilitated catalyst recycling and the general use of aqueous solvents. [16][17][18][19] Importantly, the created direct electronic contact can lead to enhanced electron transfer (ET) between catalyst and electrode. 20 ET processes play a crucial role in the electrocatalytic mechanism as reaction intermediates are generated through electron acceptance/ donation. Therefore, the rate of this process may not only determine overall catalytic activity but has also been shown to directly inuence the reaction products in case of oxygen reduction. 21 The study of adsorbed compounds is challenging and requires adaptation of suitable spectroscopic methods that provide structural insights into the catalytic processes at the surface. The elucidation of these heterogeneous reactions is a major prerequisite for promoting technological application of hangman compounds. In this regard, surface enhanced vibrational spectroscopy has the surface sensitivity to investigate sub-monolayer concentrations of immobilized molecules. In particular, the two vibrational spectroscopic techniques, surface enhanced Raman (SER) spectroscopy and surface enhanced infrared absorption (SEIRA), are able to provide different and oen complementary information at a molecular level that can be used to monitor both, redox changes and protonation events of adsorbed compounds. Particularly, for heme containing molecules, laser excitation with violet light allows exploitation of the molecular resonance effect yielding surface enhanced resonance Raman (SERR) spectroscopy to selectively monitor the vibrational modes of the absorbing porphyrin ring. Hence, SERR spectro-electrochemistry has been used extensively in the past, in particular, to analyse the redox and catalytic properties of surface bound heme enzymes. [22][23][24][25][26] Recently, also SERR measurements of surface bound heme containing molecular catalysts were presented providing interesting insights into their catalytic mechanism by inter alia monitoring direct product transformation at the heme using a RDE-SERR setup. [27][28][29][30][31][32] SEIRA spectroscopy on the other hand monitors all vibrations of the surface bound molecules. It is, however, especially sensitive to polar vibrations, such as carboxylic acid groups, and has been used in the past e.g. to analyse the protonation of a single glutamic acid residue in a complex protein matrix. 33 The combination of both types of surface enhanced vibrational spectroscopies has been applied to understand the effect of protein reorientation in enzymatic electrocatalysis. 34 In the present work it is used for the rst time to study small electrocatalytic active complexes on surfaces and to correlate electron transfer with proton delivery events in the coordination sphere. Thus, this technique is able to provide unique insight into the 2 nd coordination sphere controlled heterogeneous electron transfer (HET) of molecular catalysts on surfaces in operando. In this paper, we present the rst results regarding electron and proton transfer processes of surface bound heme based hangman complexes in the absence of substrate using SERR and SEIRA spectroscopy.</p><!><p>Iron hangman porphyrin compounds were synthesised according to published procedures. 6,11 Briey, the free base porphyrins POH and POMe were synthesised as described in ref. 6 and reacted with iron(II) chloride in dimethylformamide as described in ref. 11. Aerobic acid workup yields in the formation of the corresponding chloroiron(III) porphyrin complexes. 11 For SERR measurements, an electrochemically roughened Ag ring electrode was used as solid support prepared by a previously described procedure. 35 For SEIRA measurements, a Si prism was coated chemically with a thin Au layer that was used as electrode interface. A detailed description of the process and measurement geometry can be found here. 36 The respective electrodes were incubated overnight (>16 h) in an ethanolic solution containing 0.6 mM and 0.3 mM of 1-heptanethiol (98%, Sigma Aldrich) and 1-(11-mercaptoundecyl)imidazole (96%, Sigma Aldrich), respectively. This procedure leads to the formation of a mixed self-assembled monolayer (SAM) on top of the electrode's surface. The electrodes were cleaned with abundant ethanol prior to use. Hangman adsorption was achieved by incubation of the SAM coated electrodes with a ca. 10 mM solution of the hangman compound in DCM. 28 Immobilisation was nished aer 2 h, and unspecic bound, i.e. physisorbed, compounds removed by rinsing with abundant DCM (>99.8%, Sigma Aldrich). 28 The electrodes were subsequently mounted into a homemade spectro-electrochemical cell prepared for potential controlled SERR experiments and rotated (10 Hz) during measurements to avoid laser induced degradation. Rotation of the electrode is further necessary to minimize diffusion limitation of substrate or protons. 37 SEIRA measurements were carried out using a home-built spectro-electrochemical cell in the ATR mode in Kretschmann geometry using the Si prism as waveguide. 36 For measurements in aqueous phosphate buffer (PBS) solution, an Ag/AgCl 3 M KCl reference electrode was used (DriRef, WPI). Unless otherwise mentioned, PBS buffer always refers to pH 7 and 100 mM concentration. Catalysis tests were performed using diluted H 2 O 2 (30% in water, Sigma Aldrich) in buffer using a commercial rotating Au disc electrode setup (Pine Instruments). All employed solvents and chemicals were purchased and used without further purication. All experiments were performed under Ar atmosphere.</p><p>SERR spectra were acquired using the 413 nm line of a krypton ion laser (Coherent Innova 300c) coupled to confocal Raman setup with a single-stage spectrograph (Jobin Yvon LabRam 800 HR) equipped with a liquid-nitrogen-cooled CCD detector in 180 back scattering geometry. The laser light was focused using a Nikon 20Â objective (N. A. 0.35) with a working distance of 20 mm. Laser power on the sample was about 1 mW. Spectra acquisition times varied from 5 to 60 s for stationary and from 120 to 180 s for time resolved measurements, respectively. All experiments were repeated several times to ensure reproducibility. For time resolved (TR) SERR experiments, potential jumps of variable height and duration were applied to trigger the redox reaction as previously described. 38 The SERR spectra were measured at different delay times following the potential jump using synchronized laser light modulators. Aer background subtraction the spectra were treated by component analysis, in which the spectra of individual species, i.e. components, were tted to the measured spectra using a home-made analysis soware. 39 SEIRA measurements were carried out using a Bruker IFS 66v/s spectrometer equipped with a photoconductive MCT detector. 400 scans were co-added for a spectrum with a nal resolution of about 4 cm À1 .</p><!><p>Two iron hangman porphyrin complexes with a different hanging functional group were synthesized according to published procedures. 6,11 In the rst complex, the hanging group consists of a proton active carboxylic acid terminus while the second exhibits a carboxylic ester group (see Fig. 1). The former complex is abbreviated as FePOH, the latter as FePOMe. Immobilization of the hangman complexes on SERR active Ag supports was achieved by coating the supports with a mixed monolayer following a recently published procedure. 28 The monolayer consists of two types of molecules: a shorter methyl terminated (HS-(CH 2 ) 6 -CH 3 ) and a longer imidazole terminated (HS-(CH 2 ) 10 -Im) alkanethiol. Specic binding of the hangman compound is expected to occur by coordination of the imidazole nitrogen to the heme iron as present in heme-histidine systems of biological heme enzymes. Unspecic bound compounds were removed by rinsing with abundant dichloromethane (DCM). Subsequently, the solution was either changed to acetonitrile (ACN) or PBS buffer. As a rst step, it was checked whether the hangman complexes could preserve their catalase function upon immobilization. For this, catalytic catalase activity of the immobilized hangman complexes towards H 2 O 2 oxidation was detected by chronoamperometry. A rotating Au electrode coated with SAM and hangman complexes was immersed into a 100 mM PBS pH 7 solution and the potential set to +0.1 V. Upon stepwise H 2 O 2 substrate addition, increasing catalytic currents were observed conrming catalytic activity in the immobilized state in aqueous PBS buffer solution (for details see ESI Section 1 †). For the concentration of H 2 O 2 at half maximum current, a value of 8 AE 4 mM and 3 AE 1 mM H 2 O 2 was determined for FePOH and FePOMe, respectively. Moreover, a ca. threefold higher catalytic current was observed for FePOH than for FePOMe under identical experimental conditions (ESI Fig. S1 †).</p><!><p>In a second step, SERR spectroscopy was performed on the hangman/electrode system using a rotating Ag ring electrode. supporting the proposed direct binding of the heme iron to the imidazole (see ESI section 2 † for more information). Fig. 2 shows the SERR spectra of the imidazole immobilised FePOH hangman complex on Ag electrodes at À0.4 V and 0.15 V applied electrode potential in Ar purged PBS buffer. The spectra resemble typical spectra of heme compounds, exhibiting strong marker bands around 1370 cm À1 (n 4 ) and 1575 cm À1 (n 2 ) as well as a broader band with lower intensity at around 1495-1500 cm À1 (n 3 ). These bands are indicative for the redox, spin and ligation state of respective heme compounds. 40 To extract the contribution of the different redox and congurational species, a component t analysis was performed (for details of the tting procedure see section 3 in ESI †). 39 Briey, known component spectra of different heme types were used and modied accordingly. As a result, the presence of two heme spin states (high spin (HS) and low spin (LS)) were found, each appearing in two different redox states (Fe III and Fe II ). The molecular nature of the different (spin) species is not known a priori. However, regarding the surface functionality of the SAM, we propose that the HS species is represented by an iron complex with the imidazole group of the SAM as h axial ligand. A water or hydroxide molecule loosely attached to the heme iron as a sixth ligand is very likely and usually does not lead to a change in spin state (vide infra). 40 To induce the observed LS state, a stronger binding sixth ligand is required. The nature of this 6 th ligand is yet unknown and it might be a residue of the synthesis procedure or an unwanted side-product that is formed on the electrode surface. This species furthermore shows only a limited redox activity (vide infra). For the following data evaluation, we therefore concentrate on the Fe-HS species.</p><p>To transform SERR intensities into relative surface concentrations, spectral intensities of the different heme species, determined from the component analysis, were multiplied with respective SERR cross sections accounting for the different RR scattering efficiency and summed up to a total intensity. [40][41][42] Relative concentration of a particular heme species was derived by calculating the relative spectral intensity of this species in the overall intensity. Calculation and determination of the cross sections followed established procedures for heme proteins (for details see section 4 in the ESI †). [40][41][42] The relative surface concentrations of each species are shown in Fig. 3A as a function of applied potential. At the starting potential of 0.15 V, a mixture of oxidized HS and LS species is observed with the HS species as the major fraction. A reduced species that contains both, a LS and a HS conformation, arises at more negative potentials at the expense of the oxidized HS species. In contrast, the concentration of the oxidized LS species seems to be largely independent from the applied electrode potential.</p><p>A t of the Nernst equation to the values of the oxidized HS species as a function of potential yields the redox potential E 0 for the redox couple Fe III /Fe II -HS. The so derived values for E 0 of FePOH and FePOMe are plotted in Fig. 3B as a function of pH. Here, signicant differences are observed for the two types of hangman complexes. While FePOH shows a distinct dependence of E 0 on pH, E 0 of FePOMe remains almost pH independent. A linear t of the data for FePOH yields a slope of À57 AE 5 mV pH À1 .</p><!><p>In contrast to SERR measurements, SEIRA experiments can visualize non-heme-related changes. For SEIRA measurements, the hangman complexes were immobilized in the same way as in the SERR experiment albeit in this case a nanostructured gold lm deposited on a Si prism was used as electrode. 36,43 First, adsorption of the SAM onto the electrode was followed. Here, a band pattern in the region from 3000 cm À1 to 2800 cm À1 and a band at 1113 cm À1 is observed (Fig. S3 †). These bands can be attributed to modes with high contributions from C-H stretching vibration due to the methylene groups of the SAM molecules. 44 Furthermore, a band at 1510 cm À1 is observed, which is assigned by comparison with literature and by DFT calculations to the n(C]C) stretching mode of the deprotonated imidazole ring. 45 A more detailed discussion on the band assignment is provided in section 2 of the ESI. † Fig. 4 shows the SEIRA spectrum of the immobilized FePOH and FePOMe compound in ACN solution using the SAM coated electrode as a reference. Upon addition of the compounds, the imidazole band at 1510 cm À1 disappears or shis conrming that the hangman complexes indeed bind via the proposed coordinative Fe/N(imidazole) bond. Furthermore, for FePOH a prominent band arose at around 1737 cm À1 that can be assigned to the n(C]O) stretching vibration of the protonated carboxylic acid of the hangman motif. In the case of FePOMe, this band is located at around 1727 cm À1 in accordance with an expected downshi for carbonyl stretching frequencies of esters with respect to acids. 44 For FePOH and FePOMe, a shoulder at 1706 cm À1 is observed, which is more pronounced for FePOMe. As this band is observed for both complexes, we exclude that it originates from the carbonyl stretching vibration of the hanging group itself. More likely, the band may arise from a high shied n(C] N) vibration probably of the heme pyrrole or the imidazole C]N group. Interestingly, the n(C]O) vibration of FePOH disappeared when the solution was changed to aqueous PBS buffer at pH 7. This observation can be explained with a deprotonation of the carboxylic acid group. Upon changing the pH of the buffer solution to low pH values, the band at 1737 cm À1 reappeared clearly associated with a decrease in intensity of the band at 1565 cm À1 as shown in Fig. 5A. This band most likely represents the asymmetric n(COO À ) stretching of the associated carboxylate base. The intensity of the 1737 cm À1 band was used to create a pH titration curve presented in Fig. 5B. From these measurements, we determine the pK a value of the hanging acid group in aqueous solution to be 3.4 AE 0.2. Upon D 2 O exchange, the band of the FePOH shis to 1715 cm À1 (ESI Fig. S7 †). This constitutes a rather drastic downshi and might be caused by an additional overall change in the hydrogen/deuteron bonding network around the acid group. The pK a of the acid group in D 2 O buffer was determined to be 5.2 AE 0.4 (Fig. 5B). Qualitatively such a shi in pK a is in line with a predicted increase of basicity upon deuteration of carboxylic acids. 46 Finally, SEIRA difference spectra were measured as a function of potential in ACN (with 10% MeOH) and PBS buffer. In both cases, no potential induced changes of the 1737 cm À1 band were observed (ESI Fig. S8 †) indicating that the protonated/deprotonated form of the carboxylic hanging group is stable over the scanned potential range.</p><!><p>Using time resolved SERR spectroscopy, the heterogeneous electron transfer rates k HET of FePOH and FePOMe in aqueous and deuterated phosphate buffer solution were measured by following the oxidation state of the heme as a function of time. 38 Measurements were performed at pH 7 and pH 4 to investigate the inuence of the protonation state of the carboxylic acid group on the ET kinetics. The relative contribution of the oxidized Fe III -HS species was monitored as a function of delay time subsequent to a potential jump (see ESI section 6 †). The initial potential was set to 0.10 V. The nal potential was set in a way to yield an overpotential of h ¼ E À E 0 ¼ À0.30 V taking into account the difference in redox potential at different pH values. For such a high overpotential, the rate observed for reduction can be set equal to the heterogeneous electron transfer rate k HET . 35 Fig. 6 shows a typical relaxation curve of the Fe III -HS species to the surface redox equilibrium at the nal potential induced by the potential jump. In all measurements, we observed a fast initial drop of the Fe III -HS concentration that was sometimes followed by a much slower relaxation phase until equilibrium conditions were reached (ESI Fig. S11 †). In order to achieve a consistent evaluation of the kinetic data, all kinetic traces were tted with monoexponential functions. The rate constant is obtained by k ET ¼ 1/s where s is the time constant of the t function (details on the TR SERR measurements are given in section 6 of the ESI †). 35,38,41,42 The heterogeneous ET (HET) rate constants were obtained from several different experiments and their respective mean values are listed in Table 1. Due to scattering of the data, an average relative error of 15% for k ET is stated. The determined HET rates obtained at signicantly high overpotential show values in the order of several thousand per second. Among other factors, these fast ET rates might be a result of the direct wiring of the heme iron to the electrode affording a good electronic coupling and/or electron tunneling path. 20 Again, a distinctly different behavior of FePOH compared to FePOMe was observed. While for FePOMe, the ET rates do not depend on the buffer pH within the given accuracy, the ET rate of FePOH is more than 25 times higher at pH 4 than at pH 7. A similar behavior is observed upon switching to deuterated phosphate buffer solution. For FePOMe, a slight decrease of the rate constant is observed upon changing from pH 7 to pH 4 in D 2 O. In contrast, the rates increased for FePOH by 6 times for the same set of measurements. Comparing the rates at the same pH in H 2 O vs. D 2 O, another remarkable observation is made. FePOMe shows almost no kinetic isotope effect (KIE) at both measured pH values. In contrast, the ET rates of FePOH at pH 7 increase more than tenfold affording, in fact, an inverse KIE ¼ 0.08. A smaller inverse KIE ¼ 0.3 is observed at pH 4.</p><!><p>The distinctly different behaviour of both compounds points to a direct perturbation of the redox thermodynamic and kinetic behaviour by the hanging group. While FePOMe shows almost no variation of E 0 as a function of pH, FePOH exhibits a shi of E 0 by À57 mV per pH unit. This nding strongly indicates a PCET step involved in the redox transition of FePOH from Fe III / Fe II . In this vein, the pH dependent shi further implies a transfer of one proton to the compound upon one electron reduction. 31,[47][48][49] A simple redox transition induced de-/protonation of the carboxylic group was not observed in potential dependent SEIRA experiments over a broad potential range and can therefore not account as accompanied PT process. More likely is the scenario of a water or hydroxyl ligand bound at the heme iron as 6 th ligand in the ferrous and ferric state, respectively. This additional protonable ligand may be able to induce a PCET reaction as found already for other transition metal (e.g. Ru) complexes. 49,50 The existence of such a ligand is difficult to probe spectroscopically as the addition of a hydroxyl or aqua ligand does not change considerably the electronic conguration of the heme. Therefore, RR spectra of ve coordinated and six coordinated HS-heme complexes with an aqua/hydroxyl ligand exhibit high resemblance with only minor alterations in the low frequency region from 210-450 cm À1 . 51,52 Nevertheless, hydroxyl as 6 th ligand has been identied in the crystal structure of the ferric FePOH and similar complexes have already been reported for a range of other heme compounds. 7,51 Although OH and H 2 O as 6 th ligand exhibit, in general, only weak to moderate binding affinities towards the heme iron, the presence of the carboxylic acid hanging group might be able to stabilise the ligation due to hydrogen bonding interactions. 51 In this sense, the rigid carboxylic group xes the water/OH À molecule at the heme iron cavity. A similar situation was already reported for a picket-fence Fe-heme complex in which a Fe II -OH 2 was stabilized by hydrogen bonding interaction with an amide group in the 2 nd coordination sphere. 51,52 Based on this, we propose a reaction pathway of a possible PCET reaction of the FePOH presented in Scheme 1. In this scheme, state 1, i.e. ferric FePOH, carries a hydroxyl ligand that is protonated upon reduction to the ferrous state. This means, the Fe II -OH 2 complex formation is achieved through 1e À /1H + transfer. In this assumption, low pH values would afford the thermodynamic destabilisation of state 1 in favour of state 2. This will lead to a facilitated reduction, which is consistently perceived as a positive shi of the redox potential upon lowering the pH. In the case of FePOMe, the lack of hydrogen bond interactions might result in a vacant axial position, thus, suspending a PCET reaction. Alternatively, in line with the observations, is also a scenario of ferric OH-bound FePOMe in which the hydroxyl ligand detaches upon reduction. This may also explain the subtle increase of the redox potential at lower pH, which affords lower hydroxide concentrations in solution facilitating the detachment.</p><p>The kinetic data obtained for the HET between electrode and the different hangman compounds supports the hypothesis. Here, FePOMe shows almost no deviation of k HET upon changing pH and isotopic exchange. This observation is in line with both of the proposed scenarios for FePOMe above and points to a fast and unimpeded direct ET process. Moreover, the absolute rate constants lie in the range expected for direct electrode-wired heme domains and most likely involves "pure" electron tunnelling. 20 In contrast, a distinctly different behaviour is noted for FePOH as is expected for a PCET reaction. 53,54 Here, a drastic impact of the hanging group on the HET kinetics is observed. Specically, a dependence of the HET rate on the protonation degree of the hanging carboxylic acid group was found. Fig. 7 shows the derived kinetic rate constants for FePOH from Table 1 plotted against the protonation degree calculated via the Henderson-Hasselbalch equation (see ESI section 7 †):</p><p>x COOH denotes the molar fraction of the protonated carboxylic acid group at a given pH value. A clear correlation between HET rate and protonation degree can be seen in Fig. 7. The k HET rates could be tted reasonably with an exponential function. The general dependence of the kinetic constants on the protonation degree of the hanging group can be rationalised by considering a perturbation of the ET/PT equilibrium. In this regard, two major effects may have to be distinguished. In the simplest view, FePOH carries either a protonated acid or deprotonated carboxylate hanging function. These two different pH dependent states exhibit a different net charge resulting in an altered electrostatic environment close to the heme, and altered hydrogen bonding interactions with the bound OH/OH 2 at the iron. Both factors are expected to exhibit a major impact on the stability/energy of the different states 1-4. Therefore, these factors might also signicantly modulate the pathway of the PCET shown in Scheme 1. Following this argumentation, the potential jump induced redox transition may also proceed differently for the two species leading to the different observed kinetic behaviour. Scheme 2 summarises the possible interactions of the protonated and deprotonated acid with the 6 th OH/ OH 2 ligand. Note that Scheme 2 is shown in a very minimalistic way to highlight the different reaction pathways. In principle, it cannot be excluded that an additional water molecule is placed between the 6 th ligand and the hanging group. This, however, does not lead to a principle change in the proposed reaction schemes. In the case of the protonated acid (Scheme 2A), formed at lower pH values, a hydrogen bond interaction between the acid function and bound OH is present that may allow efficient formation of Fe III -OH 2 , i.e. state 2 via PT1. Hence, equilibrium between 1 and 2 is shied to the latter, and ET may predominantly proceed via state 2 / 4. In contrast, in the case of the deprotonated acid function that lacks this H-bond, direct ET1 from state 1 / 3 is rather expected (Scheme 2B). PT2 would then occur subsequent to ET by a proton from the bulk that might be pre-coordinated at the carboxylate function (not shown in the Scheme). As the TR-SERR spectroscopic experiment, however, only follows changes in the heme redox state, the water formation at the axial ligand binding site is not monitored. Comparing the two ET routes, i.e. ET1 and ET2, one would intuitively assume that latter is more efficient independent from the protonation state of the acid group, affording faster ET rates. In fact, ET1 involves a formation of the high energetic intermediate 3 that accommodates closely situated negative charges. This is also in line with energetic considerations that generally hold for PCET reactions. 54 In our system, PT1 and ET1 are energetically uphill, while the corresponding transfer reactions ET2 and PT2 are downhill. 49,53,54 Therefore, one would expect an increase of ET rate constants upon lowering the pH as ET2 becomes the dominating process. Alternatively, the stepwise ET/PT reaction might also be replaced by a concerted PCET reaction at neutral to basic pH values to proceed directly from 1 to 4 circumventing the formation of 3. The coupling of a fast ET to a most likely slower PT process will afford signicantly decreased apparent HET rate constants measured by TR-SERR spectroscopy, also in line with our experimental observations. 54 Kinetic measurements in D 2 O can reveal the existence of a concerted PCET as the isotopic exchange would lead to a more pronounced deceleration of the HET rate constants. 54 However, in our experiments the isotopic exchange also afforded a distinctly shied pK a value of the hanging group. The observed inverse KIE might therefore be rather related to an acceleration of ET rates through shi of the protonation/deuteration equilibrium in the same vein as mentioned above. Furthermore, the possible existence of a concerted PCET process, as proposed for these complexes 11,55 is supported by the measurements of catalytic activity regarding H 2 O 2 dismutation. As this reaction requires both electrons and protons, its reaction rate will be controlled by the slowest of the two charge transfer processes. At pH 7 catalytic activity of FePOH is equal or even better than for FePOMe albeit the apparent HET rate is 2 orders of magnitude lower. If a stepwise ET/PT process would be present with a constant rate for PT, the result should afford lower catalytic activity for FePOH at pH 7. In a concerted PCET process, however, the slowest reaction could be equally fast or even faster than in the case of FePOMe. Although it is not possible to pin down unambiguously the exact reaction route, we have conclusively shown that the protonation of the hanging group is strongly inuencing the HET of immobilised hangman complexes. This observation points to a strong coupling of the HET rate with the availability of protons in the 2 nd coordination sphere. Interestingly, such modulated ET has not been observed before in solution under non-turnover conditions. However, homogeneous reactions using an electrode as electron supplier afford slow ET rates (10 À2 cm s À1 ). 15 It might very well be that the inuence of the 2 nd coordination sphere becomes only observable when high HET rates are present, which holds true for direct electrode wired complexes. 20 This effect might be highly important for electrocatalytic efficiency of surface bound hangman complexes, and has to be investigated in the future in more detail.</p><!><p>For the rst time, the electron transfer properties of immobilised iron hangman complexes were analysed in aqueous solution via surface enhanced vibrational spectroscopy. The inuence of a proton active hanging group in the 2 nd coordination sphere on the non-turnover redox thermodynamics and kinetics of the hangman complexes was studied by investigating two different hangman complexes that exhibit either an acid or an ester functionality as hanging group. Signicant differences were found for these two compounds as only the acid containing complex showed a strong dependence of redox potential and heterogeneous ET kinetics on pH and H/D exchange.</p><p>Concomitantly performed SERR and SEIRA measurements were able to correlate the HET rates to the pK a of the carboxylic acid hanging group, which was determined experimentally for the rst time in aqueous buffer solution. The obtained data provides evidence for an increased HET rate with increased protonation degree of the carboxylic acid function. As possible explanation, a PCET reaction is proposed for the proton active complex that is strongly modulated by the pH dependent redox equilibrium of the hanging acid group. The overall ndings shed light on the reaction mechanism of heterogenised hangman complexes in aqueous environment and demonstrate the impact of the 2 nd coordination sphere on the redox and kinetic properties of these catalysts immobilized on electrodes. This effect might be of high relevance for the heterogeneous catalytic activity of Fe hangman complexes or similar molecular electrocatalysts. Finally, our research proves the capability of the combination of (TR) SERR and SEIRA spectroscopy to probe 2 nd coordination sphere mediated reactions.</p>

Royal Society of Chemistry (RSC)

Synthesis and Quantitative Structure\xe2\x80\x93Activity Relationship of Imidazotetrazine Prodrugs with Activity Independent of O6-Methylguanine-DNA-methyltransferase, DNA Mismatch Repair and p53

The antitumor prodrug Temozolomide is compromised by its dependence for activity on DNA mismatch repair (MMR) and the repair of the chemosensitive DNA lesion, O6-methylguanine (O6-MeG), by O6-methylguanine-DNA-methyltransferase (EC 2.1.1.63, MGMT). Tumor response is also dependent on wild-type p53. Novel 3-(2-anilinoethyl)-substituted imidazotetrazines are reported that have activity independent of MGMT, MMR and p53. This is achieved through a switch of mechanism so that bioactivity derives from imidazotetrazine-generated arylaziridinium ions that principally modify guanine-N7 sites on DNA. Mono- and bi-functional analogs are reported and a quantitative structure-activity relationship (QSAR) study identified the p-tolyl-substituted bi-functional congener as optimized for potency, MGMT-independence and MMR-independence. NCI60 data show the tumor cell response is distinct from other imidazotetrazines and DNA-guanine-N7 active agents such as nitrogen mustards and cisplatin. The new imidazotetrazine compounds are promising agents for further development and their improved in vitro activity validates the principles on which they were designed.

synthesis_and_quantitative_structure\xe2\x80\x93activity_relationship_of_imidazotetrazine_prodrugs_w

Introduction<!>Compound Design<!>Synthesis<!>In vitro QSAR: influence of MGMT and MMR<!>NCI60 Panel Data<!>Further in vitro evaluation<!>Discussion and Conclusions<!>Synthesis<!>General Method A<!>Methyl 3-(methyl(phenyl)amino)propanoate 12a15, 33<!>Methyl 3-((4-fluorophenyl)(methyl)amino)propanoate 12c<!>Methyl 3-((4-chlorophenyl)(methyl)amino)propanoate 12d<!>Methyl 3-((4-methoxyphenyl)(methyl)amino)propanoate 12e33<!>Methyl 3-(tert-butoxycarbonyl-(4-chlorophenyl)amino)propanoate 12f<!>General Method B<!>3-(Methyl(phenyl)amino)propanehydrazide 13a<!>3-((4-Fluorophenyl)(methyl)amino)propanehydrazide 13c<!>3-((4-Chlorophenyl)(methyl)amino)propanehydrazide 13d<!>3-((4-Methoxyphenyl)(methyl)amino)propanehydrazide 13e<!>tert-Butyl-4-chlorophenyl(3-hydrazinyl-3-oxopropyl)carbamate 13f<!>General Method C<!>3-(2-(Methyl(phenyl)amino)ethyl)-4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide 2a<!>3-(2-((4-Chlorophenyl)(methyl)amino)ethyl)-4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide 2d<!>tert-Butyl 2-(8-carbamoyl-4-oxoimidazo[5,1-d][1,2,3,5]tetrazin-3(4H)-yl)ethyl(4-chlorophenyl) carbamate 2f<!>3-(2-(Methyl(4-methyl-3-nitrophenyl)amino)ethyl)-4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide 2h<!>3-(2-((4-Methoxy-2-nitrophenyl)(methyl)amino)ethyl)-4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide 2i<!>3-(N-(4-Fluorophenyl)-N-methylamino)propanoyl azide 14c<!>2-(N-(4-Fluorophenyl)-N-methylamino)ethyl isocyanate 15c<!>3-(2-(N-(4-Fluorophenyl)-N-methylamino)ethyl)-4-oxo-3H,4H-imidazo[1,5-d][1,2,3,5]tetrazine-8-carboxamide 2c<!>Modified Procedure for tetrazine 2e 3-(N-Methyl-N-(4-nitrophenyl)amino)propanoyl azide 14e and 3(N-methyl-N-(2-nitrophenyl)amino)propanoyl azide<!>2-[(4-Nitrophenyl)methylamino]ethylisocyanate 15e<!>3-(2-N-Methyl-N-(4-nitrophenyl)-N-methylamino)ethyl)-4-oxo-3H,4H-imidazo[1,5-d][1,2,3,5]tetrazine-8-carboxamide 2e and 3-(2-(N-methyl-N-(2-nitrophenyl)-N-methylamino)ethyl)-4-oxo-3H,4H-imidazo[5,1d][1,2,3,5]tetrazine-8-carboxamide<!>meta-NO2 isomer<!>General Method D<!>Dimethyl 3,3\xe2\x80\xb2-(phenylazanediyl)dipropanoate 8a37<!>Dimethyl 3,3\xe2\x80\xb2-(4-fluorophenylazanediyl)dipropanoate 8c37<!>Dimethyl 3,3\xe2\x80\xb2-(4-methoxyphenylazanediyl)dipropanoate 8e37<!>General Method E<!>3,3\xe2\x80\xb2-(Phenylazanediyl)dipropanehydrazide 10a38<!>3,3\xe2\x80\xb2-(4-Fluorophenylazanediyl)dipropanehydrazide 10c<!>3,3\xe2\x80\xb2-(4-Methoxyphenylazanediyl)dipropanehydrazide 10e<!>General Method F<!>3,3\xe2\x80\xb2-(2,2\xe2\x80\xb2-(Phenylazanediyl)bis(ethane-2,1-diyl))bis-(4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide) 3a<!>3,3\xe2\x80\xb2-(2,2\xe2\x80\xb2-(4-Fuorophenylazanediyl)bis(ethane-2,1-diyl))bis-(4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide) 3c<!>3,3\xe2\x80\xb2-(2,2\xe2\x80\xb2-(4-Methoxy-2-nitrophenylazanediyl)bis(ethane-2,1-diyl))bis-(4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide) 3f<!>3,3\xe2\x80\xb2-(2,2\xe2\x80\xb2-(4-Methoxyphenylazanediyl)bis(ethane-2,1-diyl))bis-(4-oxo-3,4-dihydroimidazo[5,1-d][1,2,3,5]tetrazine-8-carboxamide) 3e<!>In vitro Chemosensitivty<!>NCI Data Handling<!>

<p>The imidazotetrazine prodrug temozolomide 1 (TMZ), a DNA methylating agent, in combination with radiotherapy, is now first line treatment for glioblastoma multiforme (GBM) in North America and Europe. However, intrinsic and acquired resistance significantly limits the ultimate efficacy of therapy. In particular, tumor re-growth up to two years following TMZ treatment is aggressive and resistant to further TMZ therapy.1-3 This paper reports new compounds of the TMZ class that have activity independent of the two principal constraints on the ability of a tumor to respond to TMZ therapy: MGMT activity which directly repairs TMZ GO6-methylation of DNA, and MMR which enables the tumor response to TMZ therapy.2</p><p>At neutral pH, TMZ is relatively unstable (t1/2 = 1.24 h, pH 7.4)4 and undergoes hydrolytic ring-cleavage to the open chain triazene 5-(3-methyltriazen-1-yl)-imidazole-4-carboxamide (MTIC, t1/2 = 8 min, pH 7.4),5 which then fragments to the highly reactive electrophile, methyldiazonium (t1/2 = 0.39 sec, pH 7.4).6-8 Methyldiazonium reacts with nucleophilic groups on DNA, resulting in DNA methylation, Scheme 1. Approximately 70% of the methyl groups transferred to DNA appear at N7-guanine, 10 % at N3-adenine, and 5 % at O6-guanine sites. The N-methylation products are readily repaired by the base-excision repair pathways so are not major contributors to chemosensitivity.9 In contrast, O6-MeG lesions are reversed by direct removal by the MGMT protein. The MGMT gene is silenced by promoter methylation in approximately 35% of GBM.1 In these tumors, persistent O6-MeG lesions form wobble base-pairs with thymidine during replication; O6-MeG=T pairings result in futile cycles of mismatch repair, leading to stalled replication forks, DNA double-strand breaks and ultimately cell death.1 Low levels of MMR expression (or mutations leading to non-functional MMR) lead to a tolerant phenotype that cannot respond to TMZ. Consistent with these mechanisms of O6-MeG processing and repair, high-level expression of the MGMT protein is a major mechanism of inherent TMZ resistance in primary tumors. Down regulation or mutations of the MMR pathway or up-regulation of MGMT expression are important causes of acquired TMZ resistance in GBM.2 Herein we report the synthesis, pre-clinical activity in vitro and QSAR of new imidazotetrazines that have activity independent of MMR and MGMT.</p><!><p>The imidazotetrazine bicycle is a prodrug of alkyldiazonium ions which are liberated by pH-dependent hydrolysis.6 Careful control of these reactive intermediates, in particular the suppression of competing side reactions, such as hydrolysis, elimination or re-arrangement,10-12 is essential in the design of effective new agents.13, 14 Furthermore, of the methyl groups transferred from TMZ to DNA, only a small fraction becomes the therapeutically-beneficial O6-MeG lesion: we sought to achieve therapeutic benefit though generation of N7-G-adducts, the major products of reaction of TMZ with DNA, thereby making more efficient use of the imidazotetrazine prodrug. We have designed novel series of mono-(2a-i) and bi-functional (3a–f) imidazotetrazines in which a 3-(2-anilinoethyl) group was substituted for the 3-methyl group of TMZ. The new compounds are efficient precursors of aziridinium ions, 4, Scheme 2. This strategy provides for effective control of the reactivity of incipient diazonium ions using a neighboring group participation mechanism not available to TMZ. Moreover, aziridinium ions are reactive intermediates of proven clinical utility, being closely related to those generated by nitrogen mustard drugs, and are established as working through N7-guanine adducts.15 An additional feature of the drug design is the aniline para-substituent "X" of derivatives 2 and 3 that can be optimized to fine-tune pharmacological activity. This group affects the electron density at the aniline nitrogen by resonance or inductive effects and thereby controls the basicity and nucleophilicity of this site: i.e. the propensity either to protonate or to form an aziridinium ion. In addition, the bi-functional molecules would be expected to generate DNA cross-links, that would not be processed by MMR or MGMT and so avert those constraints on activity; previous studies have demonstrated that polar16 and bulky17 GO6 adducts cannot be processed by MGMT.</p><!><p>The new imidazotetrazines were prepared by variants of the established TMZ synthesis where a diazoimidazole 5 is reacted with an isocyanate 6 to yield the imidazotetrazine, Scheme 3.18 The target compounds all included an aminoethyl group in the 3-position. The requisite β-aminoisocyanate precursors 7 are inherently unstable compounds as they bear incompatible functional groups. In the two aniline series, the anilines were sufficiently deactivated to allow isolation of the isocyanates without recourse to protecting group strategies. The formation of the relatively fragile tetrazine ring was usually planned to be the last step in the synthesis.</p><p>To prepare the dimers 3a–f, Cu(I)-catalyzed Aza-Michael conjugate addition of methyl acrylate to the requisite aniline furnished the diesters 8 in good (60–98%) yields, Scheme 4. Conversion to isocyanates 9 was achieved through the intermediate hydrazides 10 and azides 11 followed by Curtius rearrangement. For the electron rich anilines (X=CH3, OMe), hydrazide formation had to be performed under very mildly acidic (0.17 M AcOH) conditions to avoid nitration of the aniline ring (10e → 11f → 3f). This is an attractive route as the bis-hydrazides are usually solids and easily purified while the Curtius rearrangement produces pure isocyanates directly, free from contaminating by-products.</p><p>The mono-imidazotetrazine analogs 2a–i were similarly prepared from N-methyl anilines, Scheme 5. In one example, the secondary aniline version 2g (R=H, X=Cl) was achieved using a BOC protection strategy that exploited the relative acid stability of the imidazotetrazine ring. Carefully controlled diazotization (12f → 13f → 14f) yielded BOC-protected isocyanatoethylaniline 15f; tetrazine ring closure followed by final treatment with TFA gave secondary aniline 2g. As a similar yield was obtained omitting the BOC group, the protection was shown to be redundant. The p-nitro derivative 2e was accessed from hydrazide 13a by simultaneous diazotization and nitration under more vigorous conditions (10 M HCl); the mixture of regioisomers was carried through to the final step where the p-NO2 isomer, tetrazine 2e, was isolated by flash column chromatography.</p><!><p>Screening of all new compounds for chemosensitizing efficacy was undertaken following a protocol by Margisson et al that used the MGMT and MMR-proficient ovarian carcinoma cell line A2780, and its MMR-deficient derivative A2780-cp70, Figure 1.19 To assess dependence of cytotoxicity on MGMT function, each cell line was further treated with the MGMT inactivator PaTrin2.20 The data are presented graphically in Figure 1 and fully tabulated in Table S1. As expected, sensitivity to TMZ and mitozolomide (MTZ) was highly dependent on both lack of MGMT activity and presence of wild-type MMR capacity. For TMZ, the IC50 was >250 μM in both A2780 and A2780-Cp70 cell lines, compared with IC50 = 8.5 and 231 μM respectively for the same cell lines co-treated with PaTrin2. In contrast, the novel imidazotetrazines were significantly more potent, with IC50 values in the A2780 wild-type line ranging from 22–73 μM for the mono-functional (2a–i) and 1–15 μM for the bi-functional (3a–f) compounds. Dependence on MMR can be examined by comparing the shaded bars with the grey bars (i.e. MMR+/MMR− with MGMT inactivated in both cases). For temozolomide the IC50 was >27–fold lower in the MMR-proficient cell line. For the new bi-functional agents 3, this ratio was reduced to 5–5.8-fold and for the mono-functional agents 2, 2.8–10-fold. The extent of MGMT-mediated resistance can be assessed by comparing the shaded and the black bars (i.e. MGMT−/MGMT+ with MMR competent in both cases). Here, temozolomide was >30-fold more potent when MGMT was inactivated whilst for the new agents, this ratio was 0.5–5-fold for the bi-functional and 1.1–2.9-fold for the mono-functional agents. Importantly, in the absence of MMR, all compounds showed activity greater than temozolomide irrespective of the MGMT status of the cells (white bars and grey bars) showing that MMR-dependent toxicity and MGMT-mediated resistance are now only minor determinants of the chemosensitizing effect.</p><p>The similarity of response of examples 2d and 2g indicate that the N-methyl aniline is not essential to activity as the secondary and tertiary aniline compounds are equipotent (see Figure 1 and Table S1). Retention of the BOC group in analog 2f significantly reduced potency, presumably by impeding the aziridine/aziridinium formation step. Nitration in the aniline ring, whether as the sole or as an additional substituent, resulted in reduced potency. This is consistent with electron withdrawal reducing nucleophilicity of the aniline and hence the propensity for aziridinium ion formation, (for examples, compare the compound pairs 3e/3f, 2a/2e, and 2b/2h).</p><p>To probe the structure-activity relationship and examine MGMT and MMR dependence in more detail, the results for dimeric compounds 3a-e were presented in Hammett plots, Figure 2. The Hammett constant σ-p is a structural parameter that measures the electron donating or withdrawing effect of the substituent "X" on the aniline ring. The plot of IC50 against σ-p is informative in several respects, Figure 2A. For each cell type, the graph shows a pair of lines, the dotted line being MGMT-inactivated data (i.e. PaTrin 2 present); the difference between the solid and dotted lines shows the effect of MGMT on the cellular response. For most analogs this separation is of the same order as the error bars on the data, so in effect is zero. This is a clear demonstration that these new compounds have activity independent of MGMT function and is consistent with a mechanism of action independent of guanine-O6 alkylation (or involving a guanine-O6 adduct that cannot be repaired by MGMT). The separation between the two pairs of lines is the MMR effect which ranges from about 2–5-fold on the IC50, which is greatly reduced from the >27-fold effect measured for TMZ.</p><p>p-Methyl-substituted analog 3b lies on a minimum in the graph so is optimally potent, with a modestly electron-donating substituent on the aniline ring, Figure 2A. The data for all analogues except 3e (X = OMe) lie on exponential curves (see the semi-log plot, Figure 2B). The anomalous data for 3e are likely due to enhancement of the basicity of the aniline nitrogen lone pair leading to appreciable protonation at the pH of the experiment, and thereby a reduced propensity to act as a nucleophile in the aziridinium ion-forming reaction. These data clearly demonstrate the direct effect of electron density at the aniline nitrogen in determining the in vitro activity of the compounds, so provide further evidence that the aziridinium ion mechanism occurs in cells.</p><!><p>Compounds 2d, and 3c-e were selected for full screening in the NCI 60 cell line panel. Mean graph data are presented in Figures S1-6 and are summarized in Table 1. These data show that the NCEs have pharmacological activity distinct from TMZ. TMZ is essentially inactive against the panel (GI50 > 10−4 M for 57/60 cell lines, Figure S1). In contrast, the new agents exhibit strong patterns of discrimination over a 2–3-log range of GI50 (Table 1 and Figures S2-5); this is important as it shows that the new compounds possess a degree of selectivity and are not uniformly cytotoxic against cells grown in culture. These data compare favorably with the logGI50 mean and range data for established agents CHB, MEL, CP, BCNU and CCNU; this shows there is good reason to suppose that the new compounds may have an acceptable therapeutic index and be no more systemically toxic in vivo than these other clinically useful agents.</p><p>Particular sensitivity is evident in the leukemia, CNS and ovarian sub-panels. In the CNS sub-panel, the pattern of activity is similar to MTZ and CP, with particular sensitivity in the glioblastoma lines SF295 and SNB75. In the ovarian sub panel, particular sensitivity is seen in the OVACR-3 and SK-OV-3 lines. Here, in contrast to the CNS activity, the new compounds show improvement over CP (see Figure S6) this is a particularly interesting result given the wide clinical use of CP against ovarian adenocarcinoma and the fact that TMZ lacks clinical activity against ovarian cancer.21</p><p>COMPARE analysis of mean graph data can be valuable in discerning molecular mechanisms of action. The similarity in patterns of response is assessed by the Pearson rank correlation coefficient, P: values > 0.7 are considered highly significant, values 0.6–0.7 less so.22, 23 Matrix COMPARE was valuable in confirming the novelty of mechanism of action of the new agents. In this method of data handling, the new compounds were compared with standard agents selected for similarities of chemical structure or predicted mechanism of action. Agents selected were other imidazotetrazines (TMZ, MTZ), nitrogen mustard cross linkers (chlorambucil CHB, melphalan MEL), nitrosoureas (1,3-Bis(2-chloroethyl)-1-nitrosourea BCNU, N-(2-chloroethyl)-N′-cyclohexyl-N-nitrosourea CCNU) and CP, Table 2. Correlations amongst the four new imidazotetrazine compounds examined were all strong (0.92 ≤ P ≥ 0.77) and, interestingly, there was no clear discrimination between the mono- and bi-functional agents: for monofunctional 2d, 0.83 ≤ P ≥ 0.77 and in particular P = 0.79 with the equivalent p-Cl dimer 3d. Gratifyingly, the new compounds showed only moderate correlations (0.52 ≤ P ≥ 0.38) with the established imidazotetrazines TMZ and MTZ, so are distinct as new members of this compound class. The absence of correlation with MTZ is significant as this compound, exhibits a characteristic pattern of differential activity across the panel, unlike TMZ.</p><p>The putative DNA lesion of the bifunctional agents 3 is a five-atom crosslink, related in structure to those formed by the nitrogen mustard prodrugs; however, no drugs of this class showed strong correlations 0.59 ≤ P ≥ 0.29. Notably, there was no similarity to the nitrosoureas which are also diazonium ion precursors, 0.45 ≤ P ≥ 0.05. The new compounds all showed strong correlations with dacarbazine (DTIC, highest dose = 1 μM). This observation is not easy to interpret, as this prodrug requires hepatic oxidative demethylation in order to release its active electrophile, so it is somewhat surprising that DTIC scored so highly against the new imidazotetrazines. Part of the reason for this may lie in the two relatively featureless datasets of DTIC [mean log GI50(SD) = −4.30(0.28), range = 0.99, high conc. = 104 M; mean log GI50(SD) = −3.66(0.548), range = 2.55, high conc. = 10−3 M] and a shared sensitivity in the leukemia sub-panel. Simple COMPARE using DTIC GI50 (high conc. = 10−3 M) data as seed in the Standard Agents database yields no correlations P>0.537.</p><p>Simple COMPARE analysis was performed with the Standard Agents and Molecular Targets databases using the GI50 data as seed, Table 3. The Standard Agents data again identified similarity with DTIC, as noted above. In addition, modest correlations with 6MP and 6TG appeared for all four compounds, which is presumably linked to residual MMR-dependence,24, 25 a property that is shared with MTZ and DTIC. With the exception of compound 3e, correlations in the Molecular Targets database were not strong. Notable amongst the correlations of analog 3e are the lymphocyte and leukocyte related proteins LCP1, LRMP, RASSF5 and LAIR1 which may account for the high sensitivity of the leukemia sub-panel to these new compounds. Also intriguing are the correlations with the RAS-related proteins ARHGAP4 and RASSF5 and the moderate inverse correlation with RAB1A, a RAS superfamily GTPase.26</p><p>All of the NCI cell lines are characterized for expression of MGMT and components of MMR.27 Plots of protein expression vs. log GI50 were prepared to probe correlation or inverse correlation between GI50 and expression of MGMT, hMLH1 (the core MMR protein) or MSH6 (the MMR domain that recognizes O6-alkylG-DNA damage), see Figures S6-8. For each of the compounds 2d, 3c-e, the graphs showed scatters with no connection between the protein expression levels and GI50: a result consistent with the independence of activity from MGMT and MMR found in the A2780 cell line screen.</p><!><p>The new compounds were further assessed in selected drug-resistant cell lines (Tables 4 and 5). HCT116 is a human colon carcinoma cell line that is deficient in both MGMT and MMR, in consequence it is insensitive to both TMZ and CP.28-30 Both mono- and bi-functional agent families showed considerably more potent activity than TMZ and MTZ in this cell line and, in further contrast, activity was retained in the p53-deficient mutant, Table 5. TMZ activity is known to be p53 dependent.31 Compounds were also evaluated against an isogenic pair of SNB19 GBM cell lines competent in MMR that either expressed MGMT, SNB19(MGMT), or were negative for MGMT, SNB19(vector). This pair of cell lines highlights the improvements the new bi-functional agents offer over TMZ. The MGMT-proficient line being remarkably insensitive to TMZ (IC50 is unattainable) and the new bi-functional agents showing equivalent activity against the two lines (IC50 6-12 μM) with increased potency against both SNB19(vector) and SNB19 (MGMT). A2058 is an aggressive melanoma line that is TMZ-sensitive and even so, the new bi-functional agents show a 10-fold increase in potency.</p><!><p>The aims of generating new imidazotetrazine prodrugs with activity independent of MGMT and MMR have been achieved through a rationally-designed switch in molecular mechanism. This complements other work, where alternative GO6-lesion dependent strategies have been pursued.29 Two new families of mono- and bi-functional 3-anilinoethyl imidazotetrazines have been prepared. These were designed to shift the mechanism of action from dependence on GO6-DNA modification to GN7-modification. This was achieved through trapping the diazonium electrophile released on hydrolysis of the imidazotetrazine prodrug as an aziridinium ion. The mono-functional agent 2b is efficiently converted to an aziridinium ion (Scheme 2) and 2b and 3b are both DNA GN7-alkylators.15 It is proposed, on the basis of analogy with reactive intermediates generated by nitrogen mustard and similar prodrugs, that biological activity is likely associated with this lesion. However it is not yet possible to distinguish definitively between a mechanism of action deriving from GN7-adduct formation and one proceeding from GO6-adducts that are resistant to MGMT-mediated repair.</p><p>Screening in A2780 and A2780-cp70 cell lines showed that the new compounds are more potent than TMZ. The mono-functional agents are approximately equipotent with MTZ and the bi-functional agents more active. This activity is independent of MGMT activity as determined by use of the MGMT inactivator PaTrin2. Dependence of the IC50 on MMR is greatly reduced from approximately 30-fold for TMZ to 2–5-fold for the new bi-functional agents. The Hammett plot, Figure 2, for the bi-functional agents demonstrates the critical rôle of the para-substituent in directing the reactivity and bioactivity of the compounds, consistent with the aziridinium ion mechanism occurring in vitro and corroborating the design principles. The data identified the p-toluidine derivative 3b as the optimal analog for MGMT and MMR independence and potency.</p><p>In the NCI60 screen, the new agents were not uniformly cytotoxic and showed distinct and individual patterns of response over a near 3-log range of GI50. The patterns of biological response to the new agents were distinct from standard agents drawn from the imidazotetrazine, nitrogen mustard and nitrosourea families which, by chemical consideration, may be predicted to share similar mechanisms of action. For all compounds evaluated in the full 60 cell line screen, there was neither correlation with MLH1 or MSH6, nor inverse correlation with MGMT protein expression levels. Correlations in the Molecular Targets database suggest there may be a specific molecular mechanism that accounts for the sensitivity of the leukemia sub-panel to the new agents and, moreover, that activity could be linked to RAS family oncogenes.</p><p>Collectively, the in vitro data show that the initial aims of improved potency and MGMT and MMR independence have been achieved. The switch of chemical mechanism has resulted in new agents that are distinct from existing imidazotetrazines, nitrogen mustards and nitrosoureas. This implies that an increased diversity of tumor types may now be able to respond to drugs of the imidazotetrazine class. NCI60 panel data identified CP-resistant ovarian carcinoma cell lines as responsive to the new agents. Chemosensitivity in selected drug-resistant colon carcinoma and glioma cell lines (HCT116 and SNB19) showed that there is activity against TMZ- and CP-resistant tumor lines of these types, and moreover, that the HCT116 response is independent of p53 status. The ability of cells to repair damage by the new agents is linked to the ATR and FANC pathways for the bi-functional and ATM and FANC pathways for mono-functional agents.32 Mutation in one of these genes would hypersensitize a tumour to these new agents.</p><p>Overall these findings validate our understanding of the underlying chemistry of imidazotetrazine prodrugs. The members of our compound library are distinctive new agents, worthy of further development and application against a more diverse range of tumor types than TMZ.</p><!><p>Reagents were purchased from Sigma-Aldrich, Alfa Aesar and Fluka, solvents from Fisher Scientific. TLC was performed on highly-purified silica gel plates with UV indicator (silica gel F245), manufactured by Merck and visualized under UV light (366 or 254 nm) or stained with iodine. Melting points were determined with an Electrothermal IA9200 digital melting point apparatus. Infrared data were obtained using a Perkin Elmer (Paragon 1000) FT-IR Spectrophotometer. NMR spectra were acquired on a JEOL GX270 Delta, or where indicated ECA600, spectrometers observing 1H at 270.05 and 600.17 MHz and 13C at 67.80 and 150.91 MHz respectively. 13C assignments were made with the aid of the DEPT135 experiment. Mass spectra were obtained from the EPSRC National Mass Spectrometry Service Centre at the University of Swansea, UK, and the Analytical Centre at the University of Bradford using a Micromass Quattro Ultima mass spectrometer. Elemental analyses were obtained from the Advanced Chemical and Material Analysis Unit at the University of Newcastle upon Tyne, UK. All compounds entering biological evaluation were dried in vacuo for 3–4 days ≤35 °C; ≥95% purity was adjudged by combustion analysis and hplc, solvation is indicated where appropriate; corresponding analytical hplc chromatograms and highfield 1H NMR spectra are shown in ESI.</p><p>Compounds 2b and 3b were prepared as described in reference 15.</p><!><p>Single Conjugate Addition to Anilines.</p><p>4-Chloroaniline (20.0 g, 157 mmol) was mixed with methyl acrylate (54.0 g, 627 mmol, 4 eq), cuprous chloride (1.56 g, 15.68 mmol, 0.1 eq) and AcOH (30 mL) and heated under reflux at 160 °C for 2 h. The solvents were removed by evaporation and the residue partitioned between chloroform (500 mL) and water (500 mL). The organic layer was dried (MgSO4) in the presence of charcoal and filtered through celite. The solvents were removed leaving an oily residue. Petroleum ether (200 mL) was added and heated to reflux. The solvent was decanted and the product precipitated upon standing. The solid was collected by filtration, washed with petroleum ether and dried to give the ester (23.7 g, 71 %): 1H NMR (CDCl3) 7.11 (d, J = 8.8 Hz, 2H, 2-H & 6-H), 6.55 (d, J = 8.8 Hz, 2H, 3-H & 5-H), 3.68 (OCH3), 3.41 (t, J = 7.2 Hz, 2H, CH2N), 2.61 (t, J = 7.2 Hz 2H, CH2CO); 13C NMR (CDCl3) 172.8 (C=O), 146.1 (C-1), 129.2 (C-2 & C-6), 122.5 (C-4), 114.3 (C-3 & C-5), 51.9 (OCH3), 39.7 (CH2N), 33.6 (CH2CO); MS (ES): m/z 214.1 (M+H)*+; IR (film) 3400, 1725, 1600, 1500, 1435 cm−1.</p><!><p>Prepared by General Method A. Ester 12a was obtained as a colorless oil (13.3 g, 74 %). 1H NMR (CDCl3) 7.16 (t, 2H, J = 8.6 Hz, 3-H & 5-H), 6.64 (m, 3H, 2-H, 4-H & 6-H), 3.58 (m, 5H, CH2N & CH3N), 2.85 (s, 3H, OCH3), 2.49 (t, J = 7.3 Hz, 2H, CH2CO) 13C NMR (CDCl3) 172.8 (C=O), 148.6 (C-1), 129.4 (C-3 & C-5), 116.9 (C-4), 112.6 (C-2 & C-6), 51.8 (OCH3), 48.7 (NCH2), 38.3 (NCH3), 31.6 (CH2CO); MS (ES): m/z 194.1(M+H)*+; νmax (film) 1725s (C=O), 1175m (C-O) cm−1</p><!><p>Prepared according to General Method A. Ester 12c was obtained as an orange oil (2.22 g, 75%). 1H NMR (CDCl3) 6.92 (m, 2H, 3-H & 5-H), 6.66 (m, 2H, m, 2-H & 6-H), 3.65 (s, 3H, OCH3), 3.60 (t, J = 7.3 Hz, 2H, NCH2), 2.86 (NCH3), 2.52 (t, J = 7.3 Hz, 2H, CH2CO); 13C NMR (CDCl3) 172.8 (C=O), 155.7 (d, 1JCF = 235 Hz, C-4), 145.5 (C-1), 115.7 (d, 2JCF = 22 Hz, C-3 & C-5), 114.1 (d, 3JCF = 7 Hz, C-2 & C-6), 51.8 (CH3O), 49.5 (CH2N), 38.7 (CH3N), 31.5 (CH2CO); MS (ES): m/z 212.1 (M+H)*+; IR (KBr) 3025m (Ar C-H), 2950m (C-H), 1725s (C=O), 1525m (Ar C=C), 1175m (C-O) cm−1.</p><!><p>Prepared according to General Method A. Ester 12d was obtained as a yellow oil (9.1 g, 54%). 1H NMR (CDCl3) 7.15 (d, J = 8.9 Hz, 2H, 3-H & 5-H), 6.63 (d, J = 8.9 Hz, 2H, 2-H & 6-H), 3.63 (s, 3H, OCH3), 3.62 (t, J = 7.2 Hz, 2H, NCH2), 2.89 (s, 3H, NCH3), 2.54 (t, J = 7.2 Hz, 2H, CH2CO); 13C NMR (CDCl3) 172.6 (C=O), 147.2 (C-1), 129.1 (C-3 & C-5), 121.7 (C-4), 113.9 (C-2 & C-6), 51.9 (OCH3), 48.8 (NCH2), 38.4 (NCH3), 31.5 (CH2CO); MS (ES): m/z 228.1 (M+H)*+; IR (KBr) 3025m (Ar C-H), 2950m (C-H), 1725s (C=O), 1525m (Ar C=C), 1175m (C-O) cm−1.</p><!><p>According to General Method A. Ester 12e was obtained as a colorless oil (4.72 g, 58%): 1H NMR (CDCl3) 6.76 (d, 2H, J = 9.0 Hz, 3-H & 5-H), 6.67 (d, 2H, J = 9.0 Hz, 2-H & 6-H), 3.68 (s, 3H, CH3O-Ar), 3.59 (s, 3H, CH3OCO), 3.51 (t, J = 7.3 Hz, 2H, CH2N), 2.78 (s, 3H, CH3N), 2.46 (t, J = 7.3 Hz, 2H, CH2CO); 13 C NMR (CDCl3) 172.4 (C=O), 151.6 (C-4), 143.1 (C-1), 114.6 & 114.3 (C-3, C-5 & C2, C6), 55.3 (CH3O-Ar), 51.2 (CH3OCO), 49.4 (NCH2), 38.4 (CH3N), 30.9 (CH2CO); MS (ES): m/z 224.1 (M+H)*+; IR (film) 1725s (C=O), 1175m (C-O) cm−1.</p><!><p>Ester 12g (2.27 g, 10.6 mmol) was mixed with di-tert-butyl carbonate (9.3 g, 42.6 mmol, 4 eq) in the absence of solvent. The mixture was heated at 100 °C for 18 h. The mixture was partitioned between H2O and petroleum ether. The organic phase was washed with H2O (5 × 50 mL), dried and evaporated to give ester 4h as an oil (3.2 g, 97%). 1H NMR (CDCl3) 7.29 (d, J = 8.8 Hz, 2H, 2-H & 6-H), 7.11 (d, J = 8.8 Hz, 2H, 3-H & 5-H), 3.91 (t, J = 7.2 Hz, 2H, NCH2), 3.53 (OCH3), 2.57 (t, J = 7.2 Hz, 2H, CH2CO), 1.41 (s, 9H, C(CH3)3); 13C NMR (CDCl3) 172.3 (OC=O), 154.4 (NC=O), 141.0 (C-1), 130.9 (C-4), 129.1 (C-2 & C-6), 113.7 (C-3 & C-5), 79.8 (C(CH3)3), 51.8 (OCH3), 48.2 (NCH2), 32.8 (CH CO), 28.4 (CH3); MS (ES): m/z 314.1 (M+H)*+; IR (film) 1725s (C=O), 1175m (C-O) cm−1.</p><!><p>Preparation of monohydrazides.</p><p>Ester 12g (3.7 g, 11.8 mmol, 1 eq) was mixed with hydrazine hydrate (5.9 g, 118 mmol, 10 eq) in propan-2-ol (10 mL) for 48h at RT. The volatiles were removed with evaporation to leave hydrazide 13g as colorless oil (3.5 g, 95%). 1H NMR (DMSO-d6) 9.03 (br s, 1H, NHNH2), 7.05 (d, 2H, J = 8.9 Hz, 2-H & 6-H), 6.53 (d, 2H, J = 8.8 Hz, 3-H & 5-H), 5.76 (br s, 1H, NH-aniline), 4.28 (s, br, 2H, NH2), 3.19 (t, J = 7.2 Hz, 2H, NCH2), 2.41 (t, J = 7.2 Hz, 2H, CH2CO); 13C NMR (DMSO-d6) 170.5 (C=O), 148.0 (C-1), 129.1 (C-3 & C-5), 119.4 (C-4), 113.9 (C-2 & C-6), 39.9 (CH2N), 33.7 (CH2CO); MS (ES): m/z 214.1 (M+H)*+; IR (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1.</p><!><p>Prepared according to General Method B. Hydrazide 13a was obtained as a colorless oil (12.7 g, 98%). 1H NMR (CDCl3) 7.94 (br s, 1H, NH), 7.20 (t, 2H, J = 7.8 Hz, 3-H & 5-H), 6.69 (m, 3H, 2-H, 4-H, 6-H), 3.62 (br, 4H, CH2N & NH2), 2.86 (s, 3H, CH3N), 2.33 (t, J = 6.8 Hz, 2H, CH2CO); 13C NMR (CDCl3) 172.5 (C=O), 148.7 (C-1), 129.4 (C-3 & C-5), 117.1 (C-4), 112.8 (C-2 & C-6), 49.2 (NCH2), 38.6 (NCH3), 31.9 (CH2CO); MS (ES): m/z 194.1 (M+H)*+; IR (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1.</p><!><p>Prepared according to General Method B. Hydrazide 13c was obtained as a yellow oil (2.16 g, 97%): 1H NMR (CDCl3) 7.47 (br s, 1H, NH), 6.91 (m, 2H, 3-H & 5-H), 6.66 (m, 2H, 2-H & 6-H), 3.56 (t, J = 6.8 Hz, 2H, CH2N), 3.5 (br s, 2H, NH2), 2.83 (s, 3H, NCH3), 2.34 (t, J = 6.8 Hz, 2H, CH2CO); 13C NMR (CDCl3) 172.4 (C=O), 156.0 (d, 1JCF = 237 Hz, C-4), 145.6 (C-1), 115.8 (d, 2JCF = 22 Hz, C-3 & C-5), 114.8 (d, 3JCF = 8 Hz, C-2 & C-6), 50.1 (CH2N), 39.3 (CH3N), 31.8 (CH2CO); MS (EI): m/z 212.1 (M+H)*+; IR (KBr) 3300m (NH), 3050m (Ar C-H), 1650s (CONH), 1525m (Ar C=C) cm−1.</p><!><p>Prepared according to General Method B. Hydrazide 13d was obtained as white solid (4.0 g, 44%): 1H NMR (CDCl3) 7.15 (d, 2H, J = 8.9 Hz, 3-H & 5-H), 7.03 (s, 1H, NH), 6.62 (d, 2H, J = 8.9 Hz, 2-H & 6-H), 3.86 (br s, 2H, NH2), 3.63 (t, J = 6.8 Hz, 2H, CH2N), 2.88 (s, 3H, NCH3), 2.35 (t, J = 6.8 Hz, 2H, CH2CO); 13 C NMR (CDCl3) 172.2 (C=O), 147.3 (C-1), 129.2 (C-3 & C-5), 122.1 (C-4), 114.0 (C-2 & C-6), 49.3 (NCH2), 38.8 (NCH3), 31.8 (CH2CO); MS (ES): m/z 226.9 (M+H)*+; IR (KBr) 3300m (NH), 3050m (Ar C-H), 1650s (CONH), 1525m (Ar C=C) cm−1.</p><!><p>Prepared according to General Method B. Hydrazide 13e was obtained as an oil which crystallized upon standing (4.5 g, 95%): 1H NMR (CDCl3) 7.63 (br s, 1H, NH), 6.80 (d, 2H, J = 9.2 Hz, 3-H & 5-H), 6.75 (d, 2H, J = 9.2 Hz, 2-H & 6-H), 3.72 (s, 3H, CH3O-Ar), 3.47 (t, J = 6.6 Hz, 2H, CH2N), 3.3 (br s, 2H, NH2), 2.79 (s, 3H, NCH3), 2.33 (t, J = 6.6 Hz, 2H, CH2CO); C NMR (CDCl3) 172.7 (C=O), 152.9 (C-4), 143.7 (C-1), 116.5 (C-2 & C-6), 114.9 (C-3 & C-5), 55.8 (CH3O-Ar), 50.7 (CH2N), 40.0 (NCH3), 31.9 (CH2CO); MS (ES): m/z 224.1 (M+H)*+; IR (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1.</p><!><p>Prepared according to General Method B. Hydrazide 13f was obtained as a colorless oil (3.5 g, 95%). 1H NMR (CDCl3) 7.76 (br s, 1H, NH), 7.22 (d, 2H, J = 7.8 Hz, 2-H & 6-H), 7.04 (d, 2H, J = 7.8 Hz, 3-H & 5-H), 3.86 (t, J = 6.8 Hz, 2H, NCH2), 3.65 (s, br, 2H, NH2), 2.41 (t, J = 6.8 Hz, 2H, CH2CO), 1.41 (s, 9H, C(CH3)3); C NMR (CDCl3) 173.6 (C=O), 154.3 (NC=O), 141.8 (C-1), 131.1 (C-4), 129.2 (C-2 & C-6), 114.1 (C-3 & C-5), 79.8 (C(CH3)3), 48.7 (NCH2), 34.2 (CH2CO), 20.3 (ArCH3); MS (ES): m/z 314.1 (M+H)*+; IR (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1</p><!><p>Preparation of mono-imidazotetrazines</p><p>Hydrazide 13g (1.0 g, 4.68 mmol) was dissolved in dichloromethane (20 mL). Water (20 mL) was added followed by HCl (2.5 mL, 37%). The mixture was cooled in an ice/CaCl2 bath and a solution of NaNO2 (0.39 g, 56.2 mmol, 1.2 eq) in water (10 mL) was added with strong agitation at below −5 °C. The ice bath was removed and dichloromethane (20 mL) was added. The reaction mixture was stirred for 40 min and the organic layer was separated, dried over MgSO4 and evaporated to give the crude azide 14g as an oil, identified by IR. The oil was diluted with toluene (100 mL) and heated under reflux for 1 h under N2. The volatile components were removed to give the crude isocyanate 15g as an oil. Diethylether (150 mL) was added and the mixture heated with strong agitation. The hot ether solution was decanted leaving an oily residue of impurities behind. The ether layer was evaporated to leave pure isocyanate as pale yellow oil (IR νmax 2260s). The isocyanate (0.1 g, 0.48 mmol) was mixed with diazo-IC 518 (0.07 g, 0.48 mmol, 1 eq) in dry DMSO (0.3 mL) under N2 at RT in the absence of light for 24 h. The reaction was quenched with water (10 mL) and the resultant solid collected by filtration and washed with copious amounts of water to leave imidazotetrazine 2g as a brown solid (0.16 g, 29%): m.p. 160–161 °C; 1H NMR (DMSO-d6, 600 MHz) 8.84 (s, 1H, imidazole CH), 7.82 & 7.70 (2 × br s, 2H, CONH2), 7.62 (d, J = 8.8 Hz, 2H, 3-H & 5-H), 7.54 (d, J = 8.8 Hz, 2H, 2-H & 6-H), 4.52 (t, J = 6.0 Hz, 2H, NCH2), 4.42 (t, J = 6.0 Hz, 2H, CH2CO), 3.3 (br s, NH-aniline & H2O); 13 C NMR (DMSO-d6, 151 MHz) 161.8 (C=O amide), 140.0 (C-1), 139.4 (C=O tetrazine), 134.4 (Cq tetrazine), 132.7 (C-4), 131.8 (Cq imidazole), 130.0 (C-3 & C-5), 129.5 (C-H imidazole), 122.3 (C-2 & C-6), 45.9 (CH2N-aniline), 42.0 (CH2N-tetrazine); MS (ES): m/z 334.1 (M+H)*+; IR (KBr) 3450m & 3300 (CONH2), 1750s (C=O), 1650s (CONH2), 1600m & 1500s (Ar-H) cm−1. Anal. C20H19N13O4·0.25H2O, CHN.</p><!><p>Prepared according to General Method C. The product was purified by flash column chromatography (2% MeOH/CHCl3) and imidazotetrazine 2a was obtained as a white solid (0.34 g, 21%): m.p. 186–187 °C. 1H NMR (DMSO-d6, 600 MHz) 8.77 (s, 1H, imidazole CH), 7.79 & 7,68 (2 × br s, 2H, CONH2), 7.06 (t, J = 8.6 Hz, 2H, 3-H & 5-H), 6.68 (d, J = 8.6 Hz, 2H, 2-H & 6-H), 6.50 (t, J = 8.6 Hz, 1H, 4-H), 4.46 (t, J = 6.4 Hz, 2H, NCH2), 3.80 (t, J = 6.4 Hz, 2H, CH2CO), 2.91 (s, 3H, NCH3); 13C NMR (DMSO-d6, 151 MHz) 161.9 (CONH2), 148.9 (C-1), 139.6 (C=O tetrazine), 134.7 (Cq tetrazine), 131.3 (Cq imidazole), 129.4 (C-3 & C-5), 129.2 (C-H imidazole), 116.6 (C-4), 112.4 (C-2 & C-6), 50.3 (CH2N-aniline), 46.4 (CH2N-tetrazine), 38.3 (CH3); MS (ES): m/z 314.2 (M+H)*+; 627.4 (2M+H)*+; IR (KBr) 3450m (CONH2), 1750s (C=O), 1675s (CONH2), 1600m (Ar-H) cm−1. Anal. C14H15N7O2, CHN.</p><!><p>Prepared by General Method C. The product was precipitated with water, collected by filtration and the crude solid dried and suspended in chloroform (20 mL). The impurities were removed by filtration and the filtrate evaporated to yield imidazotetrazine 2d as a yellow solid (0.084 g, 25%): m.p. 154–155 °C 1H NMR (DMSO-d6, 600 MHz) 8.76 (s, 1H, imidazole CH), 7.75 & 7.65 (2 × br s, 2H, CONH2), 7.07 (d, J = 8.8 Hz, 2H, 3-H & 5-H), 6.66 (d, J = 8.8 Hz, 2H, 2-H & 6-H), 4.43 (t, J = 6.6 Hz, 2H, CH2N), 3.77 (t, J = 6.6 Hz, 2H, CH2CO), 2.87 (s, 3H, NCH3); 13C NMR (151 MHz, DMSO-d6) 161.9 (CONH2), 147.8 (C-1), 139.6 (C=O tetrazine), 134.7 (Cq tetrazine), 131.4 (Cq imidazole), 129.2 (C-H imidazole), 129.1 (C-3 & C-5), 120.3 (C-4), 113.9 (C-2 & C-6), 50.4 (CH2N-aniline), 46.3 (CH2N-tetrazine), 38.5 (CH3); MS (ES): m/z 348.2 (M+H)*+; 695.3 (2M+H)*+; 717.3 (2M+Na)*+; IR (KBr) 3450m (CONH2), 1750s (C=O), 1675s (CONH2), 1600m & 1500s (Ar-H) cm−1. Anal. C14H14ClN7O2·0.5 H2O, CHN.</p><!><p>Prepared according to General Method C. The product was purified by flash column chromatography (10% AcOH/CHCl3) and imidazotetrazine 2f was obtained as a white solid (0.15 g, 51%): m.p. 177–178 °C; 1H NMR (DMSO-d6, 600 MHz) 8.91 (s, 1H, imidazole CH), 7.83 & 7.70 (2 × br s, 2H, CONH2), 7.40 (d, J = 8.8 Hz, 2H, 3-H & 5-H), 7.32 (d, J = 8.8 Hz, 2H, 2-H & 6-H), 4.40 (t, J = 6.7 Hz, 2H, NCH2), 4.04 (t, J = 6.5 Hz, 2H, CH2CO), 1.11 (s, 9H, C(CH3)3); 13C NMR (DMSO-d6, 151 MHz) 161.8 (CONH2), 154.3 (N(CO)O), 140.8 (C-1), 139.6 (CO tetrazine), 134.6 (Cq tetrazine), 131.4 (Cq imidazole), 131.0 (C-4), 129.4 (C-H imidazole), 129.2 & 129.1 (C-3 & C-5,C-2 & C-6), 80.74 (C(CH3)3), 48.3 & 48.0 (CH2N-tetrazine & CH2N-aniline), 28.0 (CH3); MS (ES): m/z 434.3 (M+H)*+; 456.2 (M + Na)*+; IR (KBr) 3450m (CONH2), 3150m (CONH2), 1750s (C=O), 1700s (CONH2), 1600m & 1500s (Ar-H) cm−1. Anal. C18H20ClN7O4·1/3 H2O, CHN.</p><!><p>Prepared according to General Method C from hydrazide 13b. On completion of the reaction, the mixture was quenched with propan-2-ol (10 mL) and the solid collected by filtration and washed with propan-2-ol then methanol to leave imidazotetrazine 2h as an orange solid (0.43 g, 24%): m.p. 133–134 °C. 1H NMR (DMSO-d6, 600 MHz) 8.76 (s, 1H, imidazole CH), 7.77 & 7,66 (2 × br s, 2H, CONH2), 7.38 (d, J = 2.0 Hz, 1H, 2-H), 7.20 (d, J = 8.4 Hz, 1H, 5-H), 7.14 (br d, J = 8.4 Hz, 1H, 6-H), 4.44 (t, J = 6.2 Hz, 2H, CH2N-aniline), 3.45 (t, J = 6.2 Hz, 2H, CH2N-tetrazine), 2.75 (s, 3H, NCH3), 2.12 (s, 3H, ArCH3); 13C NMR (DMSO-d6, 151 MHz) 161.9 (CONH2), 143.0 (C-3), 142.5 (C-1), 139.4 (C=O tetrazine), 134.6 (Cq tetrazine), 134.3 (C-2), 131.4 (Cq imidazole), 131.3 (C-4), 129.2 (C-H imidazole), 125.3 (C-6), 121.8 (C-5), 53.4 (CH2N-aniline), 46.4 (CH2N-tetrazine), 40.7 (CH3N), 20.0 (ArCH3); MS (ES): m/z 373.2 (M+H)*+; 395.2 (M + Na)*+. Anal. C15H16N8O4·0.15 H2O, CHN.</p><!><p>Prepared according to General Method C from 13e. The product was purified by flash column chromatography (10% AcOH/CHCl3). The solid was washed with 50% MeOH/H2O and imidazotetrazine 2i was obtained as a purplish red solid (0.21 g, 26%): m.p. 166–168 °C. 1H NMR (DMSO-d6, 600 MHz) 8.80 (s, 1H, imidazole CH), 7.79 & 7.67 (2 × br s, 2H, CONH2), 7.36 (d, J = 8.9 Hz, 1H, 6-H), 7.18 (d, J = 3.0 Hz, 1H, 3-H), 7.07 (dd, J = 3.0, 8.9 Hz, 1H, 5-H), 4.42 (t, J = 6.0 Hz, 2H, CH2N-aniline), 3.68 (s, 3H, OCH3) 3.35 (t, J = 6.0 Hz, 2H, CH2N-tetrazine), 2.70 (s, 3H, NCH3); 13C NMR (DMSO-d6, 151 MHz) 162.0 (CONH2, 155.2 (C-4), 145.8 (C-2), 139.4 (C=O tetrazine), 138.6 (C-1), 134.7 (Cq tetrazine), 131.3 (Cq imidazole), 129.2 (C-H imidazole), 125.1 (C-6), 119.9 (C-3), 108.9 (C-5), 56.4 (CH3O), 54.5 (CH2N-aniline), 46.6 (CH2N-tetrazine), 42.0 (NCH3); MS (ES): m/z 389.2 (M+H)*+; 411.2 (M + Na)*+; IR (KBr) 3425m (CONH2), 3150m (CONH2), 1750s (C=O), 1700s (CONH2), 1600m & 1525s (Ar-H) cm−1. Anal. C15H16N8O5·0.1 H2O, CHN.</p><!><p>Hydrazide 13c (0.25 g, 1.2 mmol) was dissolved in a mixture of AcOH (0.66M 15 mL) and DCM (15 mL), the mixture was then stirred on a CaCl2-ice bath at 0 °C, a solution of NaNO2 (0.49 g, 7.1 mmol) in H2O (25 mL) was added dropwise keeping the exothermic reaction at 0–5 °C. A further amount of DCM (20 mL) was added, the DCM layer separated, washed with H2O (20 mL) dried over MgSO4, then filtered. Formation of azide 14c was confirmed by IR. 1H NMR (CDCl3, 600 MHz): 6.93 (t, 2H, JHH = 3JHF = 9.1 Hz, 3′,5′-H), 6.66 (dd, 2H, JHH = 9.1 Hz, 4JHF = 4.3 Hz, 2′,6′-H), 3.60 (t, 2H, J = 7.1 Hz, 3-H), 2.86 (s, 3H, NCH3), 2.55 (t, 2H, J = 7.1 Hz, 2-H); 13C NMR (CDCl3, 151 MHz) 179.2 (C-1), 155.3 (d, 1JCF = 237 Hz, C-4′), 145.2 (C-1′), 115.8 (d, 2JCF = 21.7 Hz, C-3′,5′), 114.3 (d, 3JCF = 7.2 Hz, C-2′,6′), 49.4 (C-3), 38.9 (NCH3), 34.2 (C-2); IR (liq. film) 2914w (CH st), 2137s (CON3), 1711s (CO), 1511m (C=C), 815s (p-di-substituted aromatic ring) cm−1.</p><!><p>The anhydrous DCM solution of azide 14c was stirred under nitrogen at RT overnight. Isocyanate 15c formation was confirmed by IR and 1H NMR. The DCM was evaporated under reduced pressure at low temperature (an ice bath was used to lower the temperature). The isocyanate was collected as a yellow oil (0.2 g, 77%). 1H NMR (CDCl3, 600 MHz) 6.95 (t, 2H, JHH = 3JHF = 9.1 Hz, 3′,5′-H), 6.69 (dd, 2H, JHH = 9.1 Hz, 4JHF = 4.3 Hz, 2′,6′-H), 3.43 (m, 4H, 1,2-H), 2.94 (s, 3H, NCH3); 13C NMR (CDCl3, 151 MHz) 157.2 (NCO), 155.3 (d, 1JCF = 237 Hz, C-4′), 145.2 (C-1′), 115.8 (d, 2JCF = 21.7 Hz, C-3′,5′), 114.4 (d, 3JCF = 7.2 Hz, C-2′,6′), 54.5 (C-3), 40.9 (C-2), 39.6 (NCH3); IR (liq film) 3057w (Ar-CH st), 2917w (CH st), 2270s (NCO), 1512m (C=C), 1228m, 816s (p-di-substituted aromatic ring) cm−1; MS (ES) m/z 195 (100%) (M+H)+.</p><!><p>Isocyanate 15c (0.2 g, 1.03 mmol) was diluted with DMSO (1.5 mL) under nitrogen then added to a suspension of diazo-IC 5 (0.14 g, 1.03 mmol) in DMSO (1.5 mL), the mixture was stirred at rt protected from light for 48 h. The 1H NMR spectrum showed product 2c formation. The reaction mixture was suspended in H2O (30 mL) and filtered, the residue was washed with copious amounts of H2O until the washings came through colorless and then washed with Et2O. The solid was purified by flash column chromatography, 10% AcOH in CHCl3 was used for the elution; the fractions containing the product were evaporated to give an orange solid. The solid was then re-dissolved in CHCl3 and filtered, the CHCl3 evaporated and the residue re-suspended in Et2O and then collected by filtration to give 2c (0.057 g, 17%), m.p. 172-173 °C. 1H NMR (DMSO-d6, 600 MHz): 8.77 (s, 1H, 6-H), 7.75 & 7.65 (2 × br s, 2H, CONH2), 6.91 (t, 2H, JHH = JHF = 9.1 Hz, 3′,5′-H), 6.69 (dd, 2H, JHH = 9.1 Hz, JHF = 4.3 Hz, 2′,6′-H), 3.45 (t, 2H, J = 6.4 Hz, 1-H), 3.77 (t, 2H, J = 6.4 Hz, 2-H), 2.90 (s, 3H, NCH3); 13C NMR (CDCl3, 151 MHz) 161.89 (CONH2), 155.0 (d, 1JCF = 235.5 Hz, C-4′), 145.9 (C-1′), 139.5 (C-4), 134.7 (C-8a), 131.3 (C-8), 129.2 (C-6), 115.7 (d, 2JCF = 21.7 Hz, C-3′,5′), 113.6 (d, 3JCF = 7.2 Hz, C-2′,6′), 50.9 (C-2), 46.3 (C-1) 38.7 (NCH3); IR (KBr) 3439s (NH), 3117m (Ar-CH st), 2914w (CH st), 1749s (C(4)O), 1682s (CONH2), 1515 s, 1458m (C=C), 820m (p-di-substituted aromatic ring) cm−1; MS (ES): m/z 331.9 (20%) (M+H)+, 353.9 (10%) (M+Na)+, 195.0 (100%) (C10H11FN2O+H)+. Anal. C14H14FN7O2·0.75H2O, CHN.</p><!><p>Hydrazide 13a (0.3g, 1.6 mmol) was dissolved in a mixture of DCM (10 mL) and HCl (10 mL, 14.8%), the solution was stirred at 0 °C on a CaCl2-ice bath, NaNO2 (0.66 g, 9.6 mmol) solution in H2O (10 mL) was added gradually, keeping the exothermic reaction between 0–5 °C. A further portion of DCM (15 mL) was added, the DCM layer was separated, washed with two portions of H2O (20 mL) dried over MgSO4, then filtered. Formation of the azide 14e was confirmed by IR and 1H NMR which also showed the formation of small amount of the ortho-nitrated azide in ratio 1:5. Azide 14e 1H NMR (CDCl3, 600 MHz) 8.12 (½AB, 2H, J = 9.5 Hz, 3′,5′-H), 6.61 (½AB, 2H, J = 9.5 Hz, 2′,6′-H), 3.77 (t, 2H, J = 7.0 Hz, 3-H), 3.10 (s, 3H, NCH3) 2.62 (t, 2H, J = 7.0 Hz, 2-H); 13C NMR (CDCl3) δ: 178.5 (CON3), 152.8 (C-4′), 133.2 (C-1′), 126.4 (C-3′,5′), 110.6 (C-2′,6′), 48.0 (C-3), 39.1 (NCH3), 34.4 (C-2); IR (liq film) 2918w (CH st), 2140s (N3), 1711s (CO), 1597s (NO2 st as), 1518 s (C=C), 1311s (NO2 st) cm−1. o-NO2-Azide: 1H NMR (CDCl3, 600 MHz): 7.72 (d, 1H, J = 8.9 Hz, 5′-H), 7.42 (t, 1H, J = 8.9 Hz, 4′-H), 7.21 (d, 1H,, J = 8.9 Hz, 2′-H), 6.96 (t, 1H, J = 8.9 Hz, 3′-H), 3.46 (t, 2H, J = 7.1 Hz, 3-H), 2.80 (s, 3H, NCH3) 2.62 (t, 2H, J = 7.1 Hz, 2-H).</p><!><p>The anhydrous DCM solution of the crude azide 14e was stirred under nitrogen at RT overnight. Isocyanate formation was confirmed by IR. The DCM was evaporated under reduced pressure at low temperature (an ice bath was used to lower the temperature) and the crude isocyanate 15e collected as a yellow oil (0.27 g, 90 %). 1H NMR (CDCl3, 600 MHz) 8.15 (½AB, 2H, J = 9.4 Hz, 3′,5′-H), 6.67 (½AB, 2H, J = 9.4 Hz, 2′,6′-H), 3.66 (t, 2H, J = 6.2 Hz, 2-H), 3.57 (t, 2H, J = 6.2 Hz, 1-H), 3.15 (s, 3H, NCH3); 13C NMR (CDCl3, 151 MHz)153.1 (NCO), 145.2 (C-4′), 138.1 (C-1′), 126.4 (C-3′,5′), 110.7 (C-2′,6′), 52.6 (C-2), 40.7 (C-1), 39.5 (NCH3); IR (liq film) 2919w (CH st), 2270s (NCO), 1597s (NO2 st as), 1517 s (C=C), 1311s (NO2 st) cm−1. Isocyanate ortho-nitro isomer: 1H NMR (CDCl3, 600 MHz) 7.74 (d, 1H, J = 8.4 Hz, 5′-H), 7.45 (t, 1H,, J = 8.4 Hz, 4′-H), 7.18 (d, 1H,,J = 8.4 Hz, 2′-H), 6.99 (t, 1H, J = 8.9 Hz, 3′-H), 3.48 (t, 2H, J = 6.0 Hz, 3-H), 2.90 (s, 3H, NCH3), 3.31 (t, 2H, J = 7.1 Hz, 2-H).</p><!><p>The crude mixture of the isocyanate 15e (0.27 g, 1.22 mmol) was diluted with DMSO (1.5 mL) under N2 then added to a suspension of diazo-IC 5 (0.17 g, 1.22 mmol) in DMSO (1.5 mL), the mixture was stirred at RT protected from light for 48 h. The reaction mixture was then suspended in water (30 mL) and filtered. The solid on the filter was washed with copious amounts of H2O until the washings came through colorless, then with Et2O. The dry solid was purified by flash column chromatography eluted with a gradient 5–20% AcOH in CHCl3. 1H NMR showed impurities from silica, so the imidazotetrazinone solid was dissolved in DMF, filtered and the imidazotetrazinone precipitated as a yellow solid by H2O addition. The solid was collected by filtration, washed with copious amounts of water and then dried to give 2e (0.025g, 6 %), m.p. 179–180 °C. 1H NMR (DMSO-d6, 600 MHz)8.81 (s, 1H, 6-H), 8.00 (½AB, 2H, J = 9.2 Hz, 3′,5′-H), 7.77 & 7.67 (2 × br s, 2H, CONH2), 6.82 (½AB, 2H, J = 9.2 Hz, 2′,6′-H), 4.52 (t, 2H, J = 6.3 Hz, 1-H), 3.96 (t, 2H, J = 6.3 Hz, 2-H), 3.07 (s, 3H, NCH3); 13C NMR (DMSO-d6, 151 MHz) 161.9 (CONH2), 154.0 (C-4′), 139.7 (C-4), 136.6 (C-1′), 134.7 (C-8a), 131.6 (C-8), 129.5 (C-6), 126.3 (C-3′,5′), 111.4 (C-2′,6′), 50.3 (C-2), 46.4 (C-1) 39.2 (NCH3); IR (KBr) 3443m (NH), 2919w (CH st), 1749s (C(4)O), 1684s (CONH), 1597s (NO2 st as), 1457m (C=C), 1316s (NO2 st) cm−1; MS(ES): m/z 359.1(80%) (M+H)+, 381.1(100%) (M+Na)+; Anal. C14H14N8O4·0.6 AcOH·0.2 CHCl3, CHN.</p><!><p>1H NMR (DMSO-d6, 600 MHz) 8.83 (s, 1H, 6-H), 7.83 & 7.71 (2 × br s, 2H, CONH2), 7.65 (d, 1H, J = 9.1 Hz, 5′-H), 7.46 (t, 1H, J = 9.1 Hz, 4′-H), 7.31 (t, 1H, J = 9.1 Hz, 2′-H), 6.92 (d, 1H, J = 9.1 Hz, 3′-H), 4.52 (t, 2H, J = 6.1 Hz, 1-H), 3.59 (t, 2H, J = 6.1 Hz, 2-H), 2.83 (s, 3H, NCH3).</p><!><p>Double Conjugate addition to anilines.</p><p>4-Chloroaniline (10 g, 78.4 mmol) was mixed with methyl acrylate (67.5 g, 784 mmol, 10 eq), cuprous chloride (1.24 g, 12.5 mmol, 0.16 eq) and AcOH (100 mL) and heated under reflux at 140 °C for 48 h. The reaction mixture was allowed to reach RT and water (300 mL) was added with strong agitation. The batch was allowed to stand in the fridge overnight and the water layer was decanted leaving the oil behind. The oil was washed with more water (2 × 300 mL), diluted with diethylether (200 mL), washed with water (300 mL), dried over MgSO4 and evaporated to give diester 8d as a light brown oil (19.0 g, 88%). 1H NMR (CDCl3) 7.17 (d, J = 8.8 Hz, 2H, 3-H & 5-H), 6.68 (d, J = 8.8 Hz, 2H, 2-H & 6-H), 3.65 (s, 6H, 2 × CH3), 3.61 (t, J = 7.2 Hz, 4H, 2 × CH2N), 2.57 (t, J = 7.2 Hz, 4H, 2 × CH2CO); 13C NMR (CDCl3) 172.3 (C=O), 145.0 (C-1), 129.4 (C-3 & C-5), 129.3 (C-4), 114.1 (C-2 & C-6), 51.9 (CH3), 47.2 (NCH2), 32.1 (CH2CO); MS (ES): m/z 300.1 (M+H)*+; IR (film) 1725s (C=O), 1175m (C-O) cm−1.</p><!><p>Prepared according to General method D. Diester 8a was obtained as an orange oil (18.8 g, 67%). 1H NMR (CDCl3) 7.16 (m, 2H, 3-H & 5-H), 6.62 (m, 3H, 2-H, 4-H, 6-H), 3.60 (m, 10H, 2 × CH2N-aniline & 2 × OCH3), 2.52 (t, J = 7.1 Hz, 4H, 2 × CH2CO); 13C NMR (CDCl3) 172.6 (C=O), 146.7 (C-1), 129.6 (C-3 & C-5), 117.2 (C-4), 112.6 (C-2 & C-6), 51.8 (CH3), 47.0 (NCH2), 32.3 (CH2CO); MS (ES): m/z 266.1 (M+H)*+; IR (film) 1725s (C=O), 1175m (C-O) cm−1.</p><!><p>Prepared according to General Method D. Diester 8c was obtained as a yellow oil (0.99 g, 75%). 1H NMR (CDCl3) 6.92 (m, 2H, 3-H & 5-H), 6.74 (m, 2H, m, 2-H & 6-H), 3.64 (s, 6H, 2 × CH3), 3.56 (t, J = 7.2 Hz, 4H, 2 × CH2N), 2.52 (t, J = 7.2 Hz, 4H, 2 × CH2CO); 13C NMR (CDCl3) 172.6 (C=O), 156.0 (d, 1JCF = 237 Hz, C-4), 143.5 (C-1), 115.9 (d, 2JCF = 22 Hz, C-3 & C-5), 114.9 (d, 3JCF = 7 Hz, C-2 & C-6), 51.8 (CH3), 47.7 (CH2N), 32.3 (CH2CO); 19F NMR (CDCl3) −127.62; MS (ES): m/z 284.0 (M+H)*+; IR (KBr) 3025m (Ar C-H), 2950m (C-H), 1725s (C=O), 1525m (Ar C=C), 1175m (C-O), cm−1.</p><!><p>Prepared according to General Method D. Diester 8e was obtained as an orange oil (21.7 g, 90%). 1H NMR (CDCl3) 6.84 (d, 2H, J = 8.1 Hz, 3-H & 5-H), 6.74 (d, 2H, J = 8.1 Hz, 2-H & 6-H), 3.75 (s, 3H, CH3), 3.64 (s, 6H, 2 × CH3), 3.53 (t, J = 7.1 Hz, 4H, 2 × NCH2), 2.51 (t, J = 7.1 Hz, 4H, 2 × CH2CO); 13C NMR (CDCl3) 175.4 (C=O), 152.9 (C-4), 141.0 (C-1), 121.4 (C-3 & C-5), 114.9 (C-2 & C-6), 55.8 (CH3), 51.9 (CH3), 50.5 (NCH2), 33.5 (CH2CO); MS (ES): m/z 296.1 (M+H)*+; IR (film) 1725s (C=O), 1175m (C-O) cm−1.</p><!><p>Preparation of bis-hydrazides</p><p>Diester 8d (19 g, 63.4 mmol) was mixed with hydrazine hydrate (12.6 g, 0.25 mol, 4 eq) in propan-2-ol (60 mL) for 48 h at RT. The resulting solid was collected by filtration, washed with propan-2-ol and dried in vacuo to give hydrazide 10d as white solid (18.8 g, 99%): 1H NMR (DMSO-d6) 9.03 (s, 2H, 2 × NH), 7.15 (d, 2H, J = 9.1 Hz, 3-H & 5-H), 6.66 (d, 2H, J = 9.1 Hz, 2-H & 6-H), 4.18 (br s, 4H, 2 × NH2), 3.46 (t, J = 7.2 Hz, 4H, 2 × NCH2), 2.23 (t, J = 7.2 Hz, 4H, 2 × CH2CO); 13C NMR (DMSO-d6) 170.4 (C=O), 146.4 (C-1), 129.3 (C-3 & C-5), 119.8 (C-4), 113.8 (C-2 & C-6), 47.4 (NCH2), 31.9 (CH2CO); MS (ES): m/z 300.1 (M+H)*+; IR (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1.</p><!><p>Prepared according to General Method E. Hydrazide 10a was obtained as a white solid (6.3 g, 72%). 1H NMR (DMSO) 9.04 (s, 2H, 2 × NH), 7.15 (t, 2H, J = 7.4 Hz, 3-H & 5-H), 6.66 (d, 2H, J = 7.4 Hz, 2-H & 6-H), 6.58(t, 2H, J = 7.4 Hz, 4-H), 4.18 (br s, 4H, 2 × NH2), 3.48 (t, J = 7.2 Hz, 4H, 2 × CH2N), 2.25 (t, J = 7.2 Hz, 4H, 2 × CH2CO); 13C NMR (DMSO) 170.5 (C=O), 147.6 (C-1), 129.7 (C-3 & C-5), 116.2 (C-4), 112.4 (C-2 & C-6), 47.4 (NCH), 32.1 (CH2CO); MS (ES): m/z 266.1 (M+H)*+; IR (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1.</p><!><p>Prepared according to General Method E. Hydrazide 10c was obtained as a white solid (0.63 g, 64%): m.p. 131.4 °C. 1H NMR (DMSO-d6) 8.91 (br s, 2H, 2 × NH), 6.88 (m, 2H, 3-H & 5-H), 6.56 (m, 2H, 2-H & 6-H), 4.07 (br s, 4H, 2 × NH2), 3.33 (t, J = 7.2 Hz, 4H, 2 × CH2N), 2.11 (t, J = 7.2 Hz, 4H, 2 × CH2CO); 13C NMR (DMSO-d) 170.5 (C=O), 154.8 (d, 16 JCF = 228 Hz, C-4), 144.5 (C-1), 116.0 (d, 2JCF = 22 Hz, C-3 & C-5), 113.7 (d, 3JCF = 7 Hz, C-2 & C-6), 47.8 (CH2N), 31.9 (CH2CO); 19F NMR (CDCl3) -129.53; MS (EI): m/z 283.0 (M+H)*+; νmax (KBr) 3300m (NH), 3050m (Ar C-H), 1650s (CONH), 1525m (Ar C=C) cm−1</p><!><p>Prepared according to General Method E. Hydrazide 10e was obtained as a white solid (20.5 g, 95%): 1H NMR (CDCl3) 7.39 (br s, 2H, 2 × NH), 6.85 (d, 2H, J = 8.1 Hz, 3-H & 5-H), 6.80 (d, 2H, J = 8.1 Hz, 2-H & 6-H), 4.10 (br s, 4H, 2 × NH2), 3.75 (s, 3H, CH3), 3.32 (t, J = 7.1 Hz, 4H, 2 × CH2N), 2.28 (t, J = 7.1 Hz, 4H, 2 × CH2CO); 13C NMR (CDCl3) 173.4 (C=O), 155.0 (C-4), 142.4 (C-1), 121.4 (C-3 & C-5), 114.7 (C-2 & C-6), 55.6 (CH3), 50.7 (NCH2), 33.3 (CH2CO); MS (ES): m/z 296.1 (M+H)*+; νmax (KBr) 3300s (NH), 3050m (Ar C-H), 1650s (CONH) cm−1.</p><!><p>Preparation of bis-imidazotetrazines</p><p>Hydrazide 10d (1.0 g, 3.18 mmol, 1 eq) was dissolved in dichloromethane (10 mL). Water (10 mL) was added followed by HCl (2.5 mL, 37 %). The mixture was cooled in an ice/CaCl2 bath and a solution of NaNO2 (0.57 g, 8.26 mmol, 2.6 eq) in water (10 mL) was added with strong agitation below 5 °C. After the addition, the reaction was allowed to reach RT and stirred overnight. The organic layer was separated, dried over MgSO4 and evaporated to give the azide 11d as a crude oil, identified by IR. The oil was diluted with toluene (100 mL) and heated under reflux for 2 h under N2. The volatile components were removed to give the crude isocyanate 9d as an oil. Diethylether (150 mL) was added and the mixture heated with strong agitation. The hot solution was decanted leaving a residue of oily impurities behind. The ether was evaporated to leave pure isocyanate as pale yellow oil (IR νmax 2260s). The isocyanate (0.29g, 1.03 mmol) was mixed with diazo-IC 5 (0.3 g, 2.17 mmol, 2.1 eq) in dry DMSO (0.1 mL) under N2 at RT in the absence of light for 24 h. The bis-imidazotetrazine was purified by flash column chromatography eluting with CHCl3:AcOH (1:1) to give imidazotetrazine 3d as a light brown solid (0.1 g, 6%): m.p. 143-144 °C. 1H NMR (DMSO-d6, 600 MHz) 8.79 (s, 2H, imidazole CH), 7.77 & 7.66 (2 × br s, 4H, 2 × CONH2), 7.09 (d, J = 8.9 Hz, 2H, 3-H & 5-H), 6.80 (d, J = 8.9 Hz, 2H, 2-H & 6-H), 4.42 (t, J = 6.6 Hz, 4H, 2 × NCH2 aniline), 3.78 (t, J = 6.6 Hz, 4H, 2 × CH2N-tetrazine); 13C NMR (DMSO-d6, 151 MHz) 161.9 (C=O amide), 146.2 (C-1), 140.0 (C=O tetrazine), 134.7 (Cq tetrazine), 131.4 (Cq imidazole), 2 × 129.3 (C-3 & C-5, C-H imidazole), 120.9 (C-4), 114.2 (C-2 & C-6), 49.1 (CH2N-aniline), 46.4 (CH2N-tetrazine); MS (ES): m/z 540.2 (M+H)*+; 562.3 (M + Na)*+; IR (KBr) 3450m (CONH2), 1750s (C=O), 1675s (CONH2), 1600m & 1500s (Ar-H) cm−1. Anal. C20H18ClN13O4·0.85 AcOH, CHN.</p><!><p>Prepared according to General Method F. The product was purified by flash column chromatography (5% AcOH/CH3CN) and imidazotetrazine 3a was obtained as a yellow solid (0.38 g, 60%): m.p. 194-195 °C; 1H NMR (DMSO-d6, 600 MHz) 8.79 (s, 2H, imidazole CH), 7.77 & 7.66 (2 × br s, 4H, 2 × CONH2), 7.06 (t, J = 7.5 Hz, 2H, 3-H & 5-H), 6.80 (d, J = 7.5 Hz, 2H, 2-H & 6-H), 6.51 (t, J = 7.5 Hz, 1H, 4-H), 4.44 (t, J = 6.9 Hz, 4H, 2 × CH2N-aniline), 3.79 (t, J = 6.9 Hz, 4H, 2 × CH2N-tetrazine); 13C NMR (DMSO-d6, 151 MHz) 161.9 (C=O amide), 147.2 (C-1), 139.7 (C=O tetrazine), 134.7 (Cq tetrazine), 131.3 (Cq imidazole), 129.6 & 129.3 (C-3 & C-5, C-H imidazole), 117.1 (C-4), 112.5 (C-2 & C-6), 48.8 (CH2N-aniline), 46.6 (CH2N-tetrazine); MS (ES): m/z 506.3 (M+H)*+; νmax (KBr) 3450m & 3150m (CONH2), 1740s (C=O), 1675s (CONH2), 1600m & 1510s (Ar-H) cm−1. Anal. C20H19N13O4·H2O·0.8AcOH, CHN.</p><!><p>Prepared according to General Method F. Imidazotetrazine 3c was obtained as dark yellow solid (0.04 g, 42%): m.p. 290-291 °C; 1H NMR (DMSO-d6, 600 MHz) 8.78 (s, 2H, imidazole CH), 7.77 & 7.66 (2 × br s, 4H, 2 × CONH2), 6.91 (m, 2H, 3-H & 5-H), 6.80 (m, 2H, 2-H & 6-H), 4.41 (t, J = 6.7 Hz, 4H, 2 × CH2N-aniline), 3.76 (t, J = 6.7 Hz, 4H, 2 × CH2N-tetrazine); 13C NMR (DMSO-d6, 151 MHz) 161.9 (C=O amide), 155.0 (d, 1JCF = 243 Hz, C-4), 144.1 (C-1), 139.7 (C=O tetrazine), 134.7 (Cq tetrazine), 131.3 (Cq imidazole), 129.3 (C-H imidazole), 116.0 (d, 2JCF = 23.1 Hz, C-3 & C-5), 114.1 (d, 3JCF = 7.2 Hz, C-2 & C-6), 49.4 (CH2N-aniline), 46.5 (CH2N-tetrazine); MS (ES): m/z 524.3 (M+H)*+; 546.2 (M + Na)*+; IR (KBr) 3450m (CONH2), 1725s (C=O), 1675s (CONH2), 1600m & 1510s (Ar-H) cm−1. Anal. C20H18FN13O4·0.35 H2O·0.3 AcOH requires, CHN.</p><!><p>Prepared according to General Method F. The product was purified by flash column chromatography (5% AcOH/CH3CN) and imidazotetrazine 3f was obtained as an orange/yellow solid (0.21 g, 26%): m.p. 139–140 °C. 1H NMR (DMSO-d6, 600 MHz) 8.72 (s, 2H, imidazole CH), 7.77 & 7.67 (2 × br s, 4H, 2 × CONH2), 7.49 (d, J = 9.2 Hz, 1H, 6-H), 7.16 (d, J = 3.0 Hz, 1H, 3-H), 7.04 (dd, J = 9.2, 3.0 Hz, 1H, 5-H), 4.32 (t, J = 6.2 Hz, 4H, 2 × CH2N-aniline), 3.67 (s, 3H, OCH3), 3.47 (t, J = 6.2 Hz, 4H, 2 × CH2N-tetrazine); 13C NMR (DMSO-d6, 151 MHz) 161.9 (C=O amide), 156.3 (C-4), 147.8 (C-2), 139.5 (C=O tetrazine), 135.6 (C-1), 134.8 (Cq tetrazine), 131.3 (Cq imidazole), 129.1 (C-H imidazole), 126.9 (C-5), 119.4 (C-6), 109.1 (C-3), 56.4 (CH3O), 52.1 (CH2N-aniline), 46.9 (CH2N-tetrazine); MS (ES): m/z 581.3 (M+H)*+; 603.1 (M + Na)*+; IR (KBr) 3450m (CONH2), 1750s (C=O), 1675s (CONH2), 1600m & 1525s (Ar-H) cm−1. Anal. C21H20N14O7·1.4 H2O·0.4 AcOH, CHN.</p><!><p>Prepared by a variation on General Method F using AcOH (0.17 M) in place of HCl. The product was purified by flash column chromatography (10% AcOH/CHCl3) and 3e was obtained as red solid (0.2 g, 11%): m.p. 178-179 °C. 1H NMR (DMSO-d6, 600 MHz) 8.75 (s, 2H, imidazole CH), 7.76 & 7,65 (2 × br s, 4H, 2 × CONH2), 6.73 (d, J = 8.1 Hz, 2H, 3-H & 5-H), 6.62 (d, J = 8.1 Hz, 2H, 2-H & 6-H), 4.39 (t, J = 6.6 Hz, 4H, 2 × CH2N-aniline), 3.72 (t, J = 6.6 Hz, 4H, 2 × CH2N-tetrazine), 3.53(s, 3H, CH3); 13C NMR (DMSO-d6, 151 MHz) 162.0 (C=O amide), 152.0 (C-4), 141.5 (C-1), 139.6 (C=O tetrazine), 134.7 (Cq tetrazine), 131.2 (Cq imidazole), 129.2 (C-H imidazole), 115.2 & 114.9 (C-2 & C-6, C-3 & C-5), 55.6 (CH3), 49.6 (CH2N-aniline), 46.9 (CH2N-tetrazine); MS (ES): m/z 536.3 (M+H)*+; 558.3 (M + Na)*+; IR (KBr) 3450m (CONH2), 1750s (C=O), 1675s (CONH2), 1600m & 1500s (Ar-H) cm−1. Anal. C21H21N13O5, CHN.</p><!><p>Cells used were A2780 (human ovarian carcinoma, from European Collection of Cell Cultures), the MMR-deficient derivative A2780-Cp70 (gift of Professor G Margisson, University of Manchester, UK). Isogenic HCT116 p53+/+ and HCT116 p53−/- came from Bert Vogelstein.39 Cells were plated into 96-well culture plates at 1 × 103 cells per well and incubated over night at 37 °C in a CO2 enriched (5%) atmosphere to enable cells to adhere to the plate. Culture medium was removed and replaced with fresh medium containing test compound at concentrations ranging from 0 (controls) to 250 μM. Following 5 days incubation at 37 °C, cell survival was determined using the MTT assay. All TMZ-related compounds were dissolved in DMSO, and the final concentration of DMSO in the culture plates was <0.1% (v/v). PaTrin2 was used as an inhibitor of MGMT and cells were incubated with test compounds in the presence or absence of 10 μM PaTrin2</p><!><p>Standard COMPARE and matrix COMPARE were run using the NCI database Data Build Date: 2012-04-28. GI50 data were selected and where multiple datasets were available, those averaged from the larger number of individual experiments were used.</p><!><p>Includes tabulated IC50 data and statistical analysis for Figures 1 and 2; NCI mean graphs and COMPARE results for compounds 2d, 3c-e; HPLC analysis of new tetrazines 2, 3.</p>

PubMed Author Manuscript

Free Energetics of Carbon Nanotube Association in Aqueous Inorganic NaI Salt Solutions: Temperature Effects using All-Atom Molecular Dynamics Simulations

In this study we examine the temperature dependence of free energetics of nanotube association by using GPU-enabled all-atom molecular dynamics simulations (FEN ZI) with two (10,10) single-walled carbon nanotubes in 3 m NaI aqueous salt solution. Results suggest that the free energy, enthalpy and entropy changes for the association process are all reduced at the high temperature, in agreement with previous investigations using other hydrophobes. Via the decomposition of free energy into individual components, we found that solvent contribution (including water, anion and cation contributions) is correlated with the spatial distribution of the corresponding species and is influenced distinctly by the temperature. We studied the spatial distribution and the structure of the solvent in different regions: intertube, intra-tube and the bulk solvent. By calculating the fluctuation of coarse-grained tube-solvent surfaces, we found that tube-water interfacial fluctuation exhibits the strongest temperature dependence. By taking ions to be a solvent-like medium in the absence of water, tube-anion interfacial fluctuation also shows similar but weaker dependence on temperature, while tube-cation interfacial fluctuation shows no dependence in general. These characteristics are discussed via the malleability of their corresponding solvation shells relative to the nanotube surface. Hydrogen bonding profiles and tetrahedrality of water arrangement are also computed to compare the structure of solvent in the solvent bulk and intertube region. The hydrophobic confinement induces a relatively lower concentration environment in the intertube region, therefore causing different intertube solvent structures which depend on the tube separation. This study is relevant in the continuing discourse on hydrophobic interactions (as they impact generally a broad class of phenomena in biology, biochemistry, and materials science and soft condensed matter research), and interpretations of hydrophobicity in terms of alternative but parallel signatures such as interfacial fluctuations, dewetting transitions, and enhanced fluctuation probabilities at interfaces.

free_energetics_of_carbon_nanotube_association_in_aqueous_inorganic_nai_salt_solutions:_temperature_

INTRODUCTION<!>Force Fields<!>Potential of Mean Force Calculation<!>Hydrophobe-Solvent Fluctuation Analysis<!>Potential of Mean Force of CNT Association in NaI Solution<!>Solvent Structure: Tube Center to Solution Bulk<!>Solvent Structure: Intertube Region<!>CONCLUSION

<p>With the plethora of future applications of carbon nanotube materials rapidly being realized and exploited, the assembly and aggregation of hydrophobes are inherent in a wide variety of fundamental and industrial processes1–9. Modulating the self-assembly process (either preventing it or exploiting it to create extended structures) often requires the addition of co-solutes. How these additives influence the hydrophobic assembly, directly or indirectly, is the central issue to understanding/explaining considerable fundamental questions. In the case of single-walled carbon nanotubes (SWNTs), recent research has focused on preventing the assembly/aggregation of SWNTs in solution, by involving addition of surfactants or surfactant-like dispersants1,2,10–12. These additional molecules effectively present a barrier for close contact between SWNTs, thus affording some control of aggregation behavior.</p><p>On the other hand, inorganic ions have recently been found to present significant effects on regulating dispersing properties of carbon nanotubes or other hydrophobic solutes. Niyogi et al. reported that without adding any co-surfactant, the use of metal chloride salts leads to a reduction in internanotube interactions13,14. While considering hydrophobic interactions, it is proposed that the different spatial preference of ions around the hydrophobes is critical to the types of (in)stability conferred on the solutes in the self-assembly process15. Larger ionic charge density species may amplify the hydrophobic interaction between nonpolar solutes and show different effects with smaller ionic charge density species16–18. In a biological context, these specific ion effects such as those observed by Hofmeister19–27 relating to the "salting-in" (increasing the solubility of proteins) and "salting-out" (antisolvent crystallization, precipitation crystallization) of proteins in aqueous electrolyte solutions of varying ionic species and concentrations have been pursued from theoretical, modeling, and experimental approaches for decades; studies continue to probe the fundamental phenomenology of this effect15,20–22,25,26,28–52. Horinek et al. investigated the free energetics for Na+, Cl−, Br− and I− to transfer from bulk aqueous solution to a hydrophobic self-assembled monolayer (SAM)-water interface in an infinite dilution48 and reported that soft polarizable monovalent anions (I− and Br−) prefer to accumulate around the hydrophobic interface. Nelson and Schwartz examined the effects of ions taken from opposite ends of the Hofmeister series (F− and SCN−) on the dynamics of a hydrophobic probe molecule at the SAM-aqueous interface via single-molecule total internal reflection fluorescence microscopy53 and revealed that the adsorption rate of the probe molecule to the SAM increased systematically in the presence of F− anions, which are expected to increase hydrophobic interactions. Using the association of two model hydrophobic plates, Zangi et al. reported "salting-out" as being entropically-driven by ions with high charge density (the effect increasing with charge density)35,41, and less polarizability, that form strong hydration complexes away from the hydrophobic surfaces, whereas "salting-in", is caused by ions with lower charge density, and greater polarizability, that exhibit preferential binding at the hydrophobic surfaces and stabilized by entropic or enthalpic effects.</p><p>The propensity of larger halide anions to the water-hydrophobe interface is consistent with the scenario of anions at liquid-vapor interfaces (which represent the limit of the most hydrophobic interface)15,51,52. Recent studies54–59 have begun to consider differential perturbations of liquid-vapor interfacial fluctuations induced by different ions along with ionic surface affinities. It is observed that the surface stable (an)ions induce larger interfacial fluctuations compared to the non-surface active species, thus demonstrating a strong correlation between induced interfacial fluctuations and ion surface stability as observed from molecular simulations. Ou et al. traced these differences in induced interfacial fluctuations to the nature of the hydration environment around the ions; water molecules in the hydration shells of surface stable species are shown to be more dynamic and less persistent compared to those in proximity to the non-surface stable species. When approaching the liquid-vapor interfaces, coupling of local solvent around ions with solvent further away and near an interface leads to different perturbations of the interface by ions, and thus vastly different contributions to interfacial fluctuations, and ultimately surface stability. They also showed that the surface stability of I− decreases as the temperature increases, as well as the corresponding induced fluctuations. In contrast, Cl− shows no such temperature dependence either for surface stability or the induced fluctuations56; for this reason, we consider only I− in the present study. Because of this implied connection of the behaviors of ions at ideally hydrophobic aqueous liquid-vapor interfaces, we seek to address effects of ions at other hydrophobe-water interfaces. We anticipate that similar qualitative trends and behaviors should arise in the hydrophobe context as observed at liquid-vapor interfaces.</p><p>The present study continues to probe the effects of simple inorganic salt solutions, sodium iodide (NaI), on the energetics of assembly in such systems, and provides a reference point from which to consider systems where salts are used in combination with larger dispersants (i.e., surfactants, polymers, etc.) in order to effect desired physical behaviors leading to desired physical properties and structures. Particularly, we are interested in the temperature dependence of the free energetics of association, and how the contributions of water and ions change with temperature, as well as the hydrophobe-solvent interfacial fluctuation. We hope to be able to explain the temperature dependencies within the context of our model invoking the spatial distribution of interacting species as a mode for conferring (de)stabilizing effects for association.</p><p>Furthermore, as an extension, we adopt the use of Graphical Processing Units (GPUs) as a new platform for molecular simulations of fully atomically resolved systems60–69. Specifically, we use GPU resources to extend the sampling times of our all-atom systems in order to compute well-converged potentials of mean force (vis-a-vis, free energies or reversible work) that require substantial sampling for statistical precision. We use current state-of-the-art algorithms for treatment of electrostatics and short-range dispersion interactions, in order to faithfully retain the quality of well-validated force fields in our studies.</p><p>This article is organized as followed. First we describe the force fields and computational details. In the next section we present our results in three topics. We consider potential of mean force for nanotube association and the contribution from system components along with the temperature dependence of this property. We continue with examining the density of water molecules and ions at the water-tube interface facing the water bulk, as represented by the local hydrogen bonding patterns and tetrahedrality. The profiles for the region between the nanotubes are discussed to explain the difference observed in potential of mean force. We further present our results of tube-water and tube-ion surface fluctuations. The last section concludes our findings and general discussion.</p><!><p>Molecular dynamics simulations were performed using a custom-developed code, FEN ZI70,71 on two GPU clusters, a local GPU cluster at the University of Delaware and a XSEDE GPU cluster called Keeneland (KIDS) at Oak Ridge National Lab. The local cluster includes 48 dual six-core compute nodes (576 cores), 96 Fermi S2070 GPU systems, four GPUs per node. The cluster utilizes both an Infiniband fabric and a Gigabit Ethernet interconnect. KIDS is a HP SL-390 (Ariston) cluster with Intel Westmere hex-core CPUs, NVIDIA 6GB Fermi NVIDIA M2070 GPUs, and a Qlogic QDR InfiniBand interconnect. The system has 120 nodes with 240 CPUs and 360 GPUs, three GPUs per node. GPUs provide substantial speed-up (up to 10X) and allow the scalable study of larger molecular systems compared to traditional many-core CPU-based clusters. We provided detailed performance comparisons of GPU-based MD simulations using Fen Zi versus CPU-based MD simulations using traditional codes such as CHARMM in our previous work71,72. Details of performance tests can be found in Reference68,69,71,72.</p><p>We chose to run canonical ensemble (NVT) simulations using the geometry shown in Figure 1; the central simulation cell consists of two liquid-vapor interfaces. This geometry allows the solvent density to relax in response to the presence of the carbon nanotubes. With respect to calculation of potentials of mean force, to be discussed further below, the methodology adopted in this work has been shown to reproduce the results for large hydrophobe association free energetics using constant pressure molecular dynamics simulations73. Temperature was maintained at T = 300, 310, 320, 330, 340, 350 and 360 K using a Nosé-Hoover thermostat74. The simulation cell was rectangular with dimensions of Lx = 40 Å, Ly = 40 Å and Lz = 200 Å, in which z is the direction normal to the liquid-vapor interface. Two (10,10) single wall carbon nanotubes are parallel and oriented along the x-direction. The SWNTs are constructed with the Nanotube Modeler package75. Each (10,10) SWNT is comprised of 400 carbon atoms with the bond length of 1.421 Å 76. The diameter (D) and the length (l) of each tube are 13.56 Å and 24.0 Å, respectively. Throughout the simulation the positions of all tube atoms were fixed. The Lennard-Jones (LJ) parameters of the tube atoms were ε = 0.0663 kcal/mol, Rmin = 4.0195 Å; however, the interactions among carbon atoms on the same tube were excluded. The water-carbon interactions were fixed to be εij = 0.1015 kcal/mol, Rmin = 3.7793 Å. These parameters are adopted from previous studies15,77, which have been validated to model hydrophobic effects in understanding the nature of surface friction in nanoscale fluid transport. For two SWNTs in pure water, a bulk slab consists of 7000 water molecules, which are represented by the extended simple point charge (SPC/E) model78. To generate 3 m salt solutions, 736 water molecules were replaced randomly with 368 pairs of anions and cations (yielding a final system of 6264 SPC/E water molecules, 368 anions, 368 cations). The molar concentrations used throughout this work are only approximate concentrations since they depend on the actual physical volume of the solution, which fluctuates slightly in this system with liquid-vapor interfaces. The actual molality (moles of ion over the mass of water) is 3.255 mol kg−1; we retain the approximate molar concentrations for convenience. In order to obtain the potential of mean force (PMF) as a function of the separation between two SWNTs, we define a collective variable (or order parameter), d, as the distance between the center axes of the tubes In this study we sample more than 50 separations/windows between 16.4 Å and 26.0 Å.</p><p>For SPC/E water, the LJ parameters are εOO = 0.1554 kcal/mol, Rmin = 3.5537 Å and are assigned only to the oxygen atoms. Point charges of oxygen and hydrogen atoms are qO = −0.8476e and qH = +0.4238e, respectively. Ions were treated as non-polarizable particles with interaction parameters based on those by Fyta et al.79 which have been shown to reproduce experimental solvation free energies and the osmotic coefficients. All the LJ parameters are summarized in Table 1.</p><p>The non-bond interactions were treated via the standard Lennard-Jones "12-6" potential</p><p>The parameters of ion-tube and ion-water interactions were determined by the Lorentz-Berthelot combining rules: (2)εij=εiεj,Rmin,ij=Rmin,i+Rmin,j2</p><p>Lennard-Jones interactions were gradually switched off at interparticle distance of 12 Å, with a gradual switching between 10 Å and 12 Å using the switching function: (3)S(rij)={1rij≤ron(roff2-rij2)2(roff2+2rij2-3ron2)(roff2-ron2)3ron<rij≤roff0rij>roff</p><p>In considering the force field to use for this study, our philosophy is to attempt to apply models developed and validated to some degree based on considerations of the hydration free energy of ions in pure water. We believe that consideration of ionic solutions requires attention to this chemical aspect of the physics of these systems. We acknowledge that the precise absolute values of hydration free energies are fraught with assumptions necessarily arising from the inseparability, in reality, of electroneutral anion-cation pairs. Furthermore, the nature of the water-hydrophobe interface is also critical when considering simulations of the nature presented here, and within this context, we have chosen force fields that, again, have been validated with the nature of this interface in mind.</p><p>Conditionally convergent long-range electrostatic interactions were treated using Particle Mesh Ewald (PME)80 approach with a 40 × 40 × 200 point grid and κ = 0.330. SHAKE81 was used to maintain the rigid geometry of each water molecule. Dynamics were propagated using a Velocity-Verlet integrator with a 1.0 fs timestep under 3D periodic boundary conditions. The sampling time for each window was at least 150 ns. Properties were calculated from all but the first 5 ns, which was treated as equilibration, resulting in more than 145 ns per window for data analyses. Details of performance tests of FENZI can be found in Reference68,69. Empirically, the performance can reach 6 ns per day using a single C2090 GPU for each simulation window.</p><!><p>The PMF for nanotube association following an order parameter that brings the parallel-oriented tubes from an large-separation, dissociated state to the associated/contact state is calculated directly from the average forces acting on a tube: (4)W(ξ)=-∫〈F(ξ)〉dξ where ξ is the collective variable taken as the separation distance between the center of mass of the two carbon nanotubes. We consider the average of forces 〈F(ξ)〉 acting on each nanotube over the sampled configurations at each separation distance; the average force used in the integration is the average of the values for the two SWNTs after negating the forces acting on the CNT located in the negative z-region of the simulation cell. All profiles of the PMF are shifted such that W(ξdissociated) = 0 kcal/mol, therefore, the difference in PMF from dissociated state to the global minimum (the contact state) is defined as: (5)ΔWtotal=W(ξcontact)-W(ξdissociated)</p><p>We can consider the total force acting on the nanotubes as well as the decomposition of the total force into constituent contributions: (6)ΔWtotal=ΔWnon-solv+ΔWsolv=-∫〈Fnon-solv(ξ)〉dξ-∫〈Fsolv(ξ)〉dξ where the subscript stands for either solvent induced or non-solvent induced contribution. Since tubes are the only non-solvent component and being fixed throughout the simulation for each separation distance, the tube contribution to the PMF is identical in all systems, giving ΔWnon–solv = ΔWtubes as a constant. Therefore, any differences in the PMF for our systems are solvent (water and ions) driven. Considering the equivalence of the Gibbs and Helmholtz potentials by neglecting small ambient-condition P-V contributions to the former, ΔWsolv is then our Gibbs free energy change.</p><!><p>From individual snapshots/configurations we can construct the coarse-grained instantaneous surface defined by Willard and Chandler82. Gaussian mass distributions are assigned to each water oxygen atom (or ion): (7)Φ(r;ξ)=(2πξ2)-d/2exp(-r2/2ξ2) where r is the magnitude of r, ξ is correlation length, and d stands for dimensionality. We use correlation length = 3 Å, which is roughly the size of a single water molecule in the bulk, to assign the Gaussian mass to each water oxygen. However, unlike the water, the total number of ions is less, as a result we need to consider a larger correlation length (6 Å) to render the coarse-grained surface (further validation of the choice of 6 Å is included in supporting information). At space-time point r, t, we have the coarse-grained density as</p><p>The interface is then determined as the (d − 1)-dimensional manifold with a constant value c. In the case of liquid-vapor interface, the surfaces are often orthogonal to an normal vector (z-vector), coordinate (x, y, z) for each grid points in space is set up and the surface is obtained as the manifold by setting ϱ(x, y, z) = ϱbulk/2 (based on the usual definition of Gibbs dividing surface). The mean surface of all the instantaneous surfaces is then described as the function 〈h(x, y)〉. Subtracting the mean values from the instantaneous ht(x, y), we obtain δht(x, y) as surface height and the height fluctuations 〈δh2(x, y)〉. Owing to the cylindrical geometry of nanotubes, we set up a series of (ρ, θ, l) (cylindrical) based grid points and treat ρ as the tube-water surface height, as shown in Figure 1d. ρ is radial distance of end point of the radius vector from the tube axis with the range 6.8 Å ≤ ρ ≤ 20 Å since we are interested in the tube-water surface outside the tube; θ is polar angle, which is defined as intersection angle between the radius vector and the positive y-vector with the range to ensure that only the hemisphere of interest is selected (facing the bulk: 0 ≤ θ ≤ π or intertube region: −π ≤ θ ≤ 0); l is the position along x-dimension with −11 Å ≤ l ≤ 11 Å. In practice, we set 66 × 18 × 22 grid points along (ρ, θ, l), which gives a resolution of (0.2 Å,10°,1 Å) in each dimension. The tube-water and tube-ion surfaces are defined as 3/4 bulk density of the corresponding species. We use a different constant value c here compared with liquid-vapor interface case to avoid the ambiguity during constructing the tube-solvent interface, as discussed in the Supporting Information. We note that in the current convention, instantaneous tube-solvent interface is then expressed as ht(θ, l), mean surface as 〈h(θ, l)〉, and the surface height fluctuation as 〈δh2(θ, l)〉.</p><!><p>We start by considering the free energetics of nanotube association and its decomposition into enthalpic and entropic components to assess the relative contributions of these thermodynamic quantities. We show the PMFs of nanotube association in 3 m NaI solutions at different temperatures in Figure 2a. The inset shows the minimum region of the PMF profiles. Numerical results (ΔWtotal, as defined in Equation 5) are summarized in Table 2. We also include ΔWb = W(ξpeak) − W(ξdissociated) as the value from the dissociated state to the first barrier (which is usually treated as the desolvation barrier for hydrophobic association process). The increase in stability, as represented with ΔWtotal, is linear in temperature with a slope of −0.071 kcal mol−1 K−1. This slope is consistent with the slope of stability for nanotubes in pure SPC/E (in the absence of ions, data not shown), which is not surprising since the water dominates the contribution to the PMF (will be discussed later). The barrier is seen to shift slightly toward larger ξ and exhibits a similar temperature dependence with a weaker slope −0.058 kcal mol−1 K−1. Using two hydrophobic plates (diameter ~ 21 Å) with the same water model, Zangi and Berne83 obtained a slope of −0.032 kcal mol−1 K−1 which is smaller than our result. Earlier, Mancera and Buckingham reported the aggregation of ethane in aqueous solution has a different dependence with temperature, in which they found at the intermediate 317 K ethane particles have the maximum tendency to aggregate 84. Although not the main scope in this study, this might be related to the difference in solute size between ethane with the model hydrophobes in Zangi's and our work, since the thermodynamic driving force (enthalpic or entropic) is length-scale dependent8,85.</p><p>We next consider the enthalpic and entropic components. In our system the volume change between the contact state and dissociated state is negligible, therefore the change in enthalpy (ΔH) is obtained as a difference in potential energy of the system at a particular tube-tube separation and that at the largest separation (dissociated state). To be consistent with ΔWsolv, we compute the total interaction energy between tubes and all other components of the system (which we consider the solvent—water, anions, and cations as relevant); this also includes interaction energy between elements of the solvent (water-ion, ion-ion, water-water). Since the tube-tube interaction is independent of the systems, we only look at the solvent induced enthalpic changes (ΔHsolv), which can be obtained by subtracting the direct tube-tube potential part from the overall enthalpic change. The entropic change is therefore determined from −TΔSsolv = ΔGsolv − ΔHsolv 15,83. Particularly, we plot ΔWsolv, ΔHsolv and ΔSsolv between dissociated state and contact state as functions of temperature in Figure 2b and c. The decorrelated uncertainties of ΔHsolv were obtained from block averages.86.</p><p>We notice that values of ΔWsolv and ΔHsolv are positive for the temperature range in this study. Both Gibbs free energy change and enthalpic change are less repulsive at higher temperature, with a stronger slope/dependence for ΔHsolv than ΔWsolv to the temperature. This negative slope of enthalpic change with temperature is similar to that observed in the folding of proteins87–89, although our system is under the condition of constant volume. ΔHsolv has a larger slope than ΔWsolv (as shown in Figure 2b), which indicates a compensation in entropic change. Since the tube atoms do not have any degree of freedom, ΔSsolv is essentially equal to solvent entropy. At high temperature, the ΔSsolv term is almost zero, consequently the association process is dominated by enthalpy; at low temperatures, ΔSsolv becomes 5 times larger and there is more entropic contribution to association. Similarly, Lüdemann et al.90 studied the hydrophobic interactions of two methane-like particles in water and found that at 300 K the association is controlled by entropy, while the internal energy takes over and dominates at high temperature. They reported that both enthalpy and entropy changes show crossover within the temperature range 300 K and 500 K. Here we will not argue whether the association process at the low temperature is dominated by enthalpic or entropic contribution, since the origin of that depends on the strength of solvent-tube interactions83. However, our main goal here is to reflect that the temperature dependencies of the three thermodynamic quantities in this work are similar to other studies at the nanometer scale qualitatively83,91. It implies that the parameters in this study are reasonable and also that there are underlying global similarities in the descriptions of atomic-level interactions.</p><p>Our next question is then, what is the temperature dependence of each component in the solvent? By using the same method in Reference15,92 individual solvent-induced contributions to the PMF can be decomposed into: (9)ΔWsolv=ΔWwater+ΔWNa++ΔWI-=-∫〈Fwater(ξ)〉dξ-∫〈FNa+(ξ)〉dξ-∫〈FI-(ξ)〉dξ</p><p>The contribution from each component to the PMF is displayed in Figure 3; numerical values for individual component of ΔWsolv are listed in Table 2. As the major component to the solvent, the contribution to the PMF from water dominates the solvent contribution and disfavors nanotube association, which agrees with previous work using different nonpolar solutes83,91,93–96. It is also evident that the barriers and minima are largely affected by the water contribution. These barriers/minima are the consequences of solvent reorganization around the nanotubes and will be discussed later in this manuscript. Here we focus on the temperature effect to the free energetics. As the temperature increases, ΔWwater becomes less positive/repulsive/destabilizing. On the other hand, ΔWI− is negative/stabilizing at low temperature (which is consistent with our previous study15), but becomes destabilizing at high temperature. ΔWNa+ shows no significant difference among the temperatures (≈ 1.3 kcal/mol).</p><p>We explain this difference of temperature dependence among the three species from the mechanistic point of view. The z-component of the total force from the individual species on all the tube atoms (〈Fz,i〉) can be written as a function of the Cartesian positions (x, y, z) of the corresponding species. Then, by integrating over the length of nanotube, we obtain an average force that depends only on the y and z positions of the individual species as the integrated force map: (10)〈Fz,i(y,z)〉=〈∫-l′l′Fz,i′(x,y,z)dx〉 where Fz,i′(x, y, z) corresponds to the sum of the forces applied on all tube atoms from given species i at position (x, y, z) (refer to Figure 6 in Reference15). However, the total force on the tubes is determined from the sum of the distribution/population of the force at the corresponding position, in other words, the density of the species, which is defined as: (11)ρi(y,z)=∫-l′l′ρi′(x,y,z)dx where ρi′(x,y,z) represents the number density of a given species i at position (x, y, z). We therefore have to multiply the density map of the species with the respective force map to explain how temperature results in different behavior among the three species.</p><p>We take water contribution as an example. In our system, tube atoms are neutral, therefore only van der Waals forces are considered for tube-water interactions. Shown in Figure 4a and b are 〈Fz,water(y, z)〉 at d = 16.4 Å and 19.4 Å. Again we emphasize that the force map is only related to how different tube separation affects the water spatial distribution, and is independent from the temperature. From the perspective of waters residing in the hemisphere facing the bulk solution, a negative value in the figure denotes that the net force in z-direction for the corresponding water at position (y, z) is repulsive and "pushes" the tubes to associate (stabilizing), while a positive value indicates an attractive net force that "pulls" the tubes apart (destabilizing). The opposite holds for waters within the inter-tube region. The figures are symmetrized by averaging forces from both tubes to increase the statistics since the tubes are equivalent about the center of the simulation cell. We can then categorize the "stabilizing water" and "destabilizing water" based on the spatial position of the water molecule. Since the force map is independent from the temperature, the difference in density-weighted forces between two temperatures is then determined by the difference of density maps. The difference of densities for water oxygen between 300 K (low temperature) and 360 K (high temperature), which can be written as ρi,300K(y, z) – ρi,360K(y, z), are shown in Figure 4c and d, for the tube separation d = 16.4 Å and 19.4 Å, respectively. The difference of spatial density-weighted force map is then the product of panel a and c (or panel b and d), as presented in Figure 4e and f. In Figure 4a and b, the major difference is the strongly repulsive/destabilizing water band at the intertube region for d = 19.4 Å; on the other hand there is an excess of water in this intertube region at low temperature (as presented in Figure 4d). Meanwhile, due to more "destabilizing" water in the intertube region at 300 K, water contributes more repulsive/destabilizing forces in total. We include similar spatial maps for forces induced by the anions at 300 K and 360 K in the Supporting Information. We admit that without actually summing up the force, the current force map may not be quantitatively sufficient to explain the difference of solvent-induced PMFs; what we are suggesting is that the contributions from the waters/anions to the PMF for nanotube association are different and arise from the spatial distributions of each species at the different temperatures. This differences in spatial distribution for the bulk environment, in the vicinity of the nanotubes, and the intertube regions, are studied in the following sections to probe the origin of the temperature dependence of nanotube association.</p><!><p>In this section we will study the waters/ions residing in the hemisphere facing the bulk solution, as a reference/baseline for the behavior and structure of the solvent at different temperatures. We start with radial number density profiles of water molecules and ions about the tube. We restricted sampling for these profiles to a half cylinder on the side of the tube extending into the bulk solution. This prevents sampling in the intertube region where the distributions will contain interference effects from the other tube. After each half cylinder is considered (for each tube) the symmetry of the two-tube system allows us to combine the sampling volumes into a full cylinder. Also, we limited the length of the sampling cylinder to −l′ ≤ x ≤ l′ (l′ = 11.0 Å), which is shorter than the tube length. Although being slightly reduced from the true length of the tube, it can avoid artifacts arising from the edge of the tubes. So the radial density profiles (ρ(r)) can be written as: (12)ρi(r)=Ni4πl′rdr where i is the atomic/molecular species, Ni is the corresponding number of species i within the region r − dr/2 to r + dr/2. Figure 5 shows the radial density profiles of water oxygen atoms and ions as a function of r which is relative to the center of the tube axes (at r = 0.0 Å).</p><p>In Figure 5a, we observe enhanced water density in the first solvation shell of the nanotubes which is consistent with earlier studies97–99. At higher temperatures, the structuring is reduced, though even up to 360 K the first shell retains a significantly enhanced density relative to the bulk. This effect may be attenuated with modified interaction potentials between water and hydrophobic moiety, but the essence of the physical effect will remain as long as physically relevant interaction energies/forces are incorporated in the force fields. We believe that the models for water-hydrophobe interactions applied here are well-tuned for this interface.</p><p>Figure 5b shows the density of I− in the same regions corresponding to Figure 5a for water. We observe that the anion also exhibits enhanced interfacial density relative to the bulk; the temperature dependence is similar to that of water. This enhancement recapitulates previous studies showing the propensity of large anions to hydrophobic interfaces48,100–102. As the anion replaces water in the immediate vicinity of the tubes, the cation is forced to the next outer solvation layer, thus creating an electrical double layer in the immediate vicinity of the tube, as shown in Figure 5c. The enhancement of the cation is substantially lower than exhibited by the anion near the nanotube surface. The differences in anion and cation affinity at the hydrophobe interface are consequences of the differential hydration propensities; the smaller cation is well hydrated and prefers the bulk environment. Via experiments and simulations, previous work54,56,103,104 showed that the total entropic contribution disfavors the presence of the ion at the hydrophobic interface consistent with the negative adsorption entropies of alkanes (from pentane up to octane) at an aqueous surface105 suggesting that hydrophobic solvation might be significant in determining the affinities of ions near hydrophobic surfaces59,106,107.</p><p>Furthermore, we observe that the cation is more probable to reside in the nanotube interior (r < 4.0 Å, intra-tube region) as demonstrated by the density profiles of Figure 5c. This is not surprising since the cation, being smaller, is also well solvated and does not lose its local hydration shell upon entering the intra-tube region. For this region, the average number of intra-tube water molecules (〈Nintra–tube,water〉), anions (〈Nintra–tube,anion〉) and cations (〈Nintra–tube,cation〉) are summarized in Table 3. These values are calculated as direct averages of the number of species in the tube normalized by the number of observations. Alternatively, though not done here, one can integrate the radial density profiles to obtain the equivalent value. Uncertainties included in the parentheses are determined from the standard deviations of 〈Nintra–tube〉. The average number of intra-tube water molecules is greatest at 300 K and follows the same temperature dependence as the first solvation shell; however, the average numbers of ions are low (about or less than one ion pair within the tube per snapshot) and they are not very sensitive to the system temperature. Similar to the interfacial propensities on the bulk solvent side of the tubes, we see that the anion, even when in the nanotube interior, displays enhanced density at the edge of the intra-tube volume, while the cation resides in the interior where it is well-solvated.</p><p>Recently, some studies found that the extent of air/water interfacial fluctuations induced by different ions correlates with the surface stability of the corresponding ion54–58,108. We next present the interface fluctuations of coarse-grained tube-water interfaces and tube-ion surfaces for the bulk sides of the tubes, to discuss the interfacial propensity of each species at different temperatures, as an application of the fluctuation analyses. The tube-ion surface was defined as the interface formed between the tube and the ions taken to be a solvent-like medium in the absence of water. This treatment is not unique as Feig and Pettitt also used ions as part of a solute's solvation shell109. Furthermore, since we use a high concentration of bulk anion, and considering the fact that the anion in this study prefers the tube-water interfacial region, analyses of the region in proximity to the interface by treating the ions as a continuous solvent-like medium places more validity on the approach. Ideally, because of the cylindrical symmetry of the tube, the surface fluctuation should be identical and independent of θ and l, except for the regions near the termini/edges of the tube. To represent the fluctuation values and prevent edge effects, we show the 〈δh2(θ, l)〉 as an average in the region −5 Å ≤ l ≤ 5 Å and 85° ≤ θ ≤ 95°. The uncertainties were obtained via standard deviations. The results for tube-water and tube-ion fluctuations at different temperatures are shown in Figure 6a and b. A few salient features are noteworthy in this figure: (i) generally, the magnitude of tube-ion fluctuations is larger than tube-water fluctuations. (ii) All the solvent components show enhanced fluctuations as the temperature increases; however, in terms of the absolute magnitude, the difference (between 300 K and 360 K) in tube-I− (≈ 1.6 Å2) is the greatest, followed by the difference in tube-water (≈ 0.4 Å2), and finally tube-Na+ is rather modest (≈ 0.3 Å2). If we consider the relative magnitude of each component (normalized with the inherent fluctuation at 300 K), we have tube-water (53%) > tube-I− (34%) > tube-Na+ (4.6%). Therefore, the results suggest that tube-Na+ shows basically no temperature dependence, while tube-I− and tube-water are dependent upon temperature. Interestingly, this is consistent with what has been observed in the PMF and radial density profile.</p><p>We use the deformability/malleability of solvent shell to explain these features. From Figure 5a, the density ratio between the first peak (r = 10.1 Å) and the first minimum (r = 11.7 Å) of water is 0.095 Å−3/0.015 Å−3 = 6.33; from Figure 5b this ratio for I− is 0.0031 Å−3/0.00095 Å−3 = 3.26; and for Na+ this ratio is 0.0022 Å−3/0.0015 Å−3 = 1.47. The ratios demonstrate different manners in which the solvation shells of the nanotubes couple with the solvent component at the interface. Water retains the strongest, most unambiguous, well-defined solvation shells, and does not support the intermixing with water from outer shells and therefore the tube-water interface is the least malleable. I− holds the second strongest ordered solvation shells while Na+ is the weakest, subsequently the tube-ion surfaces are more deformable. As the temperature increases, the ratios for water and I− decrease (as highlighted in the insets), their malleabilities of solvation shells are increased, and consequently, the fluctuations arise. In the case of Na+, the ratio between the first peak and the first minimum is insensitive to the temperature, hence no significant temperature dependence is observed.</p><p>Previously, we investigated the connection between the stability of a solute at an aqueous liquid-vapor interface and corresponding interface fluctuation. The fluctuation can be separated into inherent and solute-induced contributions. When temperature increases, the enhancement of inherent thermal water surface fluctuation rises, thus reducing the impact of any solvent structural perturbation induced by the solute56. This behavior accompanies the disappearance of an adsorption free energy minimum (i.e., interfacial free energy stability) of surface stable species near the interface. In this study, this behavior is shown as the reduction of the first peak of anion density (Figure 5b) as the temperature is increased.</p><p>As we suggested, the structure of solvation is important in explaining the temperature dependence of fluctuations and PMF; we consider the hydrogen bond network and geometric configurations at hydrophobic surfaces next. Specifically, we investigate hydrogen bonding numbers (average hydrogen bonds per molecule, 〈NHB〉) and the angular tetrahedral parameters (q, to be defined further below) at various locations. Two water molecules were considered hydrogen bonded if the oxygen-oxygen distance is less than 3.5 Å and the HO ···O angle is less than 30°15,73,110. The average number of hydrogen bonds, normalized by the number of hydrogen bonds in the bulk for T = 300 and 360 K for the configuration d = 16.4 Å, are shown in Figure 7a as functions of r using the same cylindrical sampling volume that we used to compute the radial density profile. In aqueous salt solutions, due to the presence of ions, the bulk value of 〈NHB〉 is less than the pure water case (about 3.5 hydrogen bonds per water molecule, refer to Reference15). The inset shows the numbers of hydrogen bonds per water molecule in the bulk at various temperatures in this study. The uncertainties were obtained via standard deviations. As the temperature increases, less hydrogen bonds are formed for bulk water. In the intra-tube region, water tends to hold more hydrogen bonds than water in the bulk region. This should not be surprising since on average only 1–2 ions are found in this region, the equivalent ion concentration is about half compared to the bulk. In the vicinity of the nanotubes, we found a slight enhancement followed by a depletion of hydrogen bonds; as the temperature elevates, the number of hydrogen bonds at the interface is reduced. Consider the deformation of hydrogen bond network as a result of lacking strong electrostatic interaction between hydrophobe and water: water-water interactions are therefore more important in shaping the orientation of water molecules111. Since the loss of hydrogen bonds is more likely to happen when there is another partner (i.e., another water molecule in our case) in the neighborhood to exchange the bond with, the location near the tube surface (which is liquid-vapor like) can promote the maximization of hydrogen bonding. However, the incompatibility of the tube surface with the tetrahedral network induces a distortion in the donor-acceptor angle for the water at the interface than in the bulk83. With increased temperature, the rotational mobility of water increases. At high temperature, the water molecules in the bulk are correlated and therefore have to maintain a certain degree of tetrahedral network connectivity; however, the orientational preference of water near the tubes limits the ability to simultaneously rotate and maintain hydrogen bond connectivity. Shown in our supporting information, we observe a positive excess of hydrogen atoms in the vicinity of the tubes, indicating that the hydrogen atoms point closer to the tube and is consistent with this scenario.</p><p>We next apply the conditional tetrahedral parameter (q) introduced by Godec et al.112 to describe the configurations of water near the nanotubes. For a tetrahedral configuration there are five points: the four vertices of the tetrahedron and the center. In a perfectly tetrahedral arrangement, all the angles between bonds hold the same value, and the cosine of this angle is −1/3. However, in the vicinity of the tubes, water molecules may not have four nearest hydrogen-bonded neighbors. Physically, at least two neighboring water molecules are necessary for a meaningful evaluation of tetrahedral order. The generalized form can be written as: (13)qi=1-1N(N-1)98∑j=1N-1∑k=j+1N(cosψj,k+13)2 where ψj,k is the angle subtended at the central water between the jth and kth bonds. N is the number of neighboring water molecules with 2 ≤ N ≤ 4. The squaring ensures the contribution from each inter-bond angle is always greater than or equal to zero. The normalization factor is chosen so that qi = 1 when the configuration is perfectly tetrahedral (all the cosine values are −1/3); qi = 0 indicates the extreme non tetrahedral arrangement. The average tetrahedral order parameter, q, normalized by the corresponding bulk values (qbulk) for T = 300 and 360 K with the configuration d = 16.4 Å, are shown in Figure 7b as functions of r using the same cylindrical sampling volume that we used to compute the radial density and hydrogen bond profiles. The inset shows qbulk values as a function of temperature, the uncertainties were obtained via standard deviations. Consider the radius of the nanotube, which is roughly 6.8 Å, the tetrahedrality of water molecules shows slight enhancement, which extends up to 5.5 Å from the tube atoms. At 300 K, q reaches a maximum of 0.97, which is about 1.1qbulk. qbulk is not sensitive to temperature, gives only 0.95% difference between 300 K and 360 K, while 〈NHB,bulk〉 gives 7.26% difference between these two temperatures. Qualitatively, our results with temperature dependence are consistent with the experiment performed by Davis et al.113. The authors reported temperature-dependent Raman scattering measurements to investigate the hydrophobic hydration shells of linear alcohols. At low temperatures, the hydration shells were found to hold a hydrophobically enhanced water structure with greater tetrahedral order along with fewer weak hydrogen bonds than the surrounding bulk water. By increasing the temperature the structure is replaced by a less ordered structure with weaker hydrogen bonds than bulk water.</p><!><p>Finally, we study the solvent structure of the intertube region. To investigate the behavior of water molecules and ions within the intertube region, we define a rectangular sampling volume as follows and illustrated in Figure 1b and c. The length (L) of the box is along the x-axis and identical with the cylinder used to sample for the radial density profiles, which makes L = 22.0 Å; the width (W) is along the z-axis and determined by subtracting from the tube separation (d) with the tube diameter (D = 13.56 Å). Finally, the height (H) equals D – 2 Rmin,OC and projects along the y-axis, with Rmin,OC = 3.7866 Å based on Lorentz-Berthelot combining rules. The choice of this sampling region is not unique; we provide further analysis and an alternative scheme in the supporting information. Results are self-consistent using both sampling volumes. The average numbers of water oxygen atoms (〈Ninter,water〉), anions (〈Ninter,anion〉) and cations (〈Ninter,cation〉) in the rectangular box as functions of d are shown in Figure 8. Insets show focused view of the separation that there is a dramatic increase of 〈Ninter,water〉 compared with other separations. As temperature increases, the number of intertube water is reduced. The plateaus in Figure 8a suggest the existence of stable solvation shells at particular tube separations, and are also found in other instances of hydrophobic association91,114. We plot the water oxygen density profile along z-dimension at 300 K and 360 K in Figure 9 to illustrate this behavior. Each z ~ z + dz region has the same x and y-dimension length resulting in a sampling volume of L × H × dz. The second solvation shell of water only appears at the largest separation, therefore the total number of 〈Ninter,water〉 is determined by the first solvation shell. At the same tube separation, the density of water at low temperature is substantially larger than the density at high temperature, which correlates with the water densities in the outer regions. Giovambattista et al. also found the same behavior for the water confined between two hydrophobic plates115. The effective hydrophobicity of the tube decreases as temperature decreases, consistent with the suppression of the vapor phase upon cooling. The numbers of anion and cations in the intertube region are not sensitive to the temperature, as shown in Figure 8b and c. We should point out that the difference of 〈Ninter,water〉 in our system is very small (less than 3 water molecules at d = 22.0 Å); meanwhile, even at the largest separation, only 1.3 pairs of ions are observed, resulting in an equivalent concentration also about half compared to the bulk environment.</p><p>We next address the surface fluctuation for intertube region. Due to the scarcity of ions in this region, it is impractical to construct tube-ion interfaces. We therefore focus on the tube-water interface fluctuation. The results for tube-water fluctuations for various tube separations at 300 K and 360 K are shown in Figure 10. Refer to our supporting information for the results at other temperatures. At the dissociated state, we found 〈δh2(θ, l)〉 ranges from 0.6 to 0.65 Å2 at all temperatures, which results from the analogous concentration/environment at this confined region. Also because of the lower local concentration in this region, we found a smaller 〈δh2(θ, l)〉 value relative to the fluctuation at the side facing the bulk (as shown with red dashed lines, along with the uncertainties, which are plotted as dotted lines). As the concentration of the solvent increases, surface tension increases and strengthens the hydrophobic effect. In our case, the reduction of ion concentration leads to less hydrophobicity and less induced interface fluctuations. In the supporting information, we include our results of 〈δh2(θ, l)〉 of the sides facing the solvent bulk, at different concentration (0 m, 1 m, 3 m) of NaI aqueous solution.</p><p>It is worth noting that a crossover of 〈δh2(θ, l)〉 relative to the bulk side fluctuation happens between d = 22.0 Å and 23.0 Å (we do not consider the crossover between d = 21.0 Å and 22.0 Å due to the large uncertainties observed at all temperatures). First, at these separations, less than 0.5 pairs of ions are observed (refer to Figure 9b and c). Considering that there are ~ 20 water molecules in the intertube region, the "crossover" should be dictated by the water behavior. Again we use the "malleability" of solvation shell to explain this "crossover" of fluctuations. From Figure 9a, when d = 22.0 Å and 23.0 Å, the intertube densities are similar; but the ratio between the density peak and density minimum (i.e. z = 0 Å) are considerably different, as listed in Table 4. The shell "malleability" for d = 23.0 Å is therefore weaker than the case at d = 22.0 Å. As a consequence, the fluctuation for d = 23.0 Å is less. Recall that for the sides facing the solvent bulk, we have a ratio ~ 6.33 for the water solvation shell in the bulk side of the tube. This value 6.33 is between the ratios at d = 22.0 Å and d = 23.0 Å (both in low and high temperature scenarios), therefore a "crossover" behavior should not be surprising. This argument also holds when we compare the ratio at the same separation but different temperatures. At d = 22.0 Å, the ratio for 300 K (~ 2.76) is larger than the ratio (~ 1.94), therefore the former solvation shell is more rigid and holds a smaller fluctuation; meanwhile, at d = 23.0 Å, the ratio for 300 K (~ 22.76) is substantially greater. Consequently its corresponding fluctuation value is significantly lower.</p><p>We next address the hydrogen bonds and tetrahedrality of the water molecules as functions of z in the intertube region. The same sampling volume (L × H × dz) is used as for the analysis of intertube densities. The results of 〈NHB〉 /NHB,bulk and q/qbulk at various tube separations at 300 K and 360 K are shown in Figure 11 and Figure 12. Usually, under the confinement of hydrophobes, the (CNT) walls do not form any hydrogen bonds with water molecules, and so the average number of hydrogen bonds per molecule in confined system is expected to be smaller than in bulk environment116,117. The addition of salts has little effect on the O-H and H-H structures but still distorts the arrangement of the water network118. Therefore the relatively lower concentration in the intertube region induces less hydrogen bond breaking (when d > 21.0Å), as shown in Figure 11. This concentration dependence of hydrogen bonding profile is consistent with the neutron diffraction data from monovalent ionic solutions reported by Mancinelli et al.119</p><p>For intertube water tetrahedrality, at small separation (d < 22.0 Å) all the q(z) are smaller than the bulk value. When d = 22.0 Å (for both low and high temperatures) we observe that the largest q(z) (at z = 0 Å) barely coincides with the bulk value. At d = 23.0 Å, strong enhanced q (even larger than the qbulk) is discovered in the intertube region, indicating a highly-ordered structure of water molecules. Again, we use the idea of equivalent concentration in the intertube region to explain this behavior. From Figure 8, below the separation of 21 Å, the presence of ion is not found. There is also a limited amount of water in the intertube region to introduce tetrahedral structure. As the tube separation increases, this region allows the entry of ions, with a lower equivalent concentration relative to the bulk. The corresponding tetrahedral structure among water molecules is therefore less perturbed by the ions, in agreement with Galamba's results of tetrahedrality of water in different concentration of aqueous salt solutions120. Another feature at d = 23.0 Å is that, considering qbulk is insensitive to the temperature, we have q(z) reduced for all values of z as temperature increases, suggesting that the local water structure becomes less ordered in this respect consistent with the temperature dependence reported by Giovambattista et al. using two hydrophobic plates115. However, due to the geometry of nanotubes and the resulting intertube environment, we observe no plateau region in our computed profile nor the wider profiles in q(z) upon cooling115,121.</p><!><p>In the preceding sections, we studied the effect of temperature on the interfacial properties of carbon nanotube assembly in 3 m NaI aqueous solution using GPU-enabled molecular dynamics simulations (FEN ZI). We examined the temperature behavior for the free energetics of two parallel single-walled carbon nanotubes by computing the potentials of mean force (PMF) between the contact state and dissociated state. Decomposition of solvent-induced PMF into changes in enthalpy and entropy suggests that all three thermodynamical quantities decrease with increasing temperature. At the highest temperature (360 K) in this manuscript, we found barely any entropic contribution to the free energetics. Further decomposition of PMF into each solvent component (water, I− and Na+) implies that as the dominant component, water contributes to the tube association in a less destabilizing manner as the temperature increases. I− shows reversed trend, and Na+ presents no dependence on temperature. From the mechanistic point of view, this difference in force contribution originates from differences in spatial distribution of the various system species. We investigated the spatial distribution and the structure of the solvent in different regions: intertube, intra-tube and the solvent bulk. For the bulk side of the nanotubes, by calculating the fluctuation of coarse-grained tube-solvent surfaces, we found that proportionally, tube-water interfacial fluctuation exhibits the strongest temperature dependence. Tube-I− interfacial fluctuation also shows similar but weaker dependence to the temperature, while tube-Na+ interfacial fluctuation shows no dependence in general. These characteristics are discussed via the malleability of their corresponding solvation shells. Hydrogen bonding profile and tetrahedrality of water arrangement are also computed to describe the structure of solvent at different temperatures. The analysis of hydrogen bonding in the vicinity of the nanotubes suggests an orientational preference of water molecules, implying a higher orientational order relative to the bulk water. In the intertube region, the hydrophobic confinement induces a relatively lower concentration environment (compared with bulk environment), therefore causing different behaviors depending on the tube separation.</p><p>In general, our work explained how association/dissociation between carbon nanotube materials can be modulated by controlling the temperature and inorganic additives. Essentially, the self-assembly strength is determined by the corresponding solvation structures. From the biological perspective, it is natural to connect the temperature and salt (concentration and type) dependencies to the studies of protein folding, micelle formations, protein-protein binding affinity and even protein-based materials122. Both experimentally and theoretically, there is already progress addressing temperature and solvent dependencies123,124; however, the heterogeneity of the biomolecular surfaces are much more complicated than the simple hydrophobic models125–130. How these solvation structures behave, or how the malleability of the solvation shells can be evaluated as a response to the change of temperature and ionic strength are therefore ongoing avenues of inquiry.</p>

PubMed Author Manuscript

Analytical and Biological Methods for Probing the Blood-Brain Barrier

The blood-brain barrier (BBB) is an important interface between the peripheral and central nervous systems. It protects the brain against the infiltration of harmful substances and regulates the permeation of beneficial endogenous substances from the blood into the extracellular fluid of the brain. It can also present a major obstacle in the development of drugs that are targeted for the central nervous system. Several methods have been developed to investigate the transport and metabolism of drugs, peptides, and endogenous compounds at the BBB. In vivo methods include intravenous injection, brain perfusion, positron emission tomography, and microdialysis sampling. Researchers have also developed in vitro cell-culture models that can be employed to investigate transport and metabolism at the BBB without the complication of systemic involvement. All these methods require sensitive and selective analytical methods to monitor the transport and metabolism of the compounds of interest at the BBB.

analytical_and_biological_methods_for_probing_the_blood-brain_barrier

1. INTRODUCTION<!>1.1. The Blood-Brain Barrier<!><!>1.2. Blood-Brain Barrier Transport Mechanisms<!>1.3. Importance of the Blood-Brain Barrier<!>2. IN VIVO METHODS FOR STUDYING TRANSPORT OF SUBSTANCES ACROSS THE BLOOD-BRAIN BARRIER<!>2.1. Intravenous Injection Methods<!>2.2. Brain Perfusion Techniques<!>2.3. Tomographic Methods<!>2.4. Microdialysis Sampling<!>2.4.1. Benefits of microdialysis sampling for blood-brain barrier studies<!>2.4.2. Drawbacks of microdialysis sampling for blood-brain barrier studies<!>2.4.3. Quantitation and calibration issues<!>2.4.4. Analytical methods for microdialysis samples<!>3. IN VITRO MODELS OF THE BLOOD-BRAIN BARRIER<!>3.1. Primary Brain Microvessel Endothelial Cells<!>3.2. Immortalized Cell Lines<!>3.3. Methods for Measuring Transport and Metabolism with In Vitro Models<!>3.3.1. Permeability markers<!>3.3.2. Scintillation counting<!>3.3.3. Capillary electrophoresis<!>3.3.4. Liquid chromatography/mass spectrometry<!>3.3.5. Microfluidic applications<!>4. SUMMARY AND FUTURE DIRECTIONS

<p>The blood-brain barrier (BBB) is a unique biological interface that maintains brain homeostasis by preventing and regulating the permeation of endogenous substances, ions, and xenobiotics (toxins, pollutants, and drugs, for example) into the extracellular space of the brain (1–3). Although beneficial for neurobiological purposes, this interface is also the major obstacle in the development of drugs for treatment of central nervous system (CNS) disorders and brain cancers (4–6). To better understand and evaluate the role of the BBB in drug delivery as well as the chemical communication between the CNS and the peripheral nervous system (PNS), a number of biological and analytical methods have been developed. This article reviews the different in vivo and in vitro approaches for studying the BBB and the analytical methods that are used to measure the transport, metabolism, and release of compounds at the blood-brain interface.</p><!><p>The major component of the BBB is the brain microvessel endothelial cell (BMVEC). Specialized proteins, such as claudin, occludin, and cadherins, hold the endothelial cells together to produce tight junctions (TJs), areas where adjacent endothelial cells are physically held together, making passive transcellular transport of small hydrophilic molecules extremely difficult. The "tightness" of these junctions can be evaluated by measuring the transendothelial electrical resistance (TEER). The cells that make up the BBB exhibit TEER values that are almost three orders of magnitude higher than those in peripheral capillaries. In fact, the resistance across these endothelial cells is so great that even the movement of small hydrated ions, such as Na+ and Cl−, is significantly restricted. The surface of BMVECs is also strongly anionic and creates an electrostatic barrier for the transport of negatively charged compounds. In addition to the physical and electrostatic barriers to transport, these cells also create a metabolic barrier. There are a number of intracellular and extracellular enzymes, including peptidases, nucleotidases, monoamine oxidase, and cytochrome P450, that convert substrates into less permeable or less toxic compounds. Additionally, the BBB is an immunological barrier that prevents bacteria and viruses from entering the brain.</p><p>The endothelial cells that make up the BBB are part of the larger neurovascular unit (NVU) that also contains pericytes, astrocytes, microglia, and neurons (Figure 1) (1, 3). All the cells in the NVU play a role in the maintenance of the BBB as well as in blood-brain signaling in both directions (7). The endothelial cells are surrounded by a basement membrane that is shared with pericytes, generally undifferentiated cells that differentiate into support cells in the brain vasculature (e.g., vascular smooth muscle cells). Known functions of the pericytes include helping to build and maintain the basement membrane at the BBB. Astrocytes are connected to the endothelial cells via their end feet and provide growth factors and nutrients to both endothelial cells and local neurons (2, 3). These cells, therefore, play an important role in blood-to-brain communication. Microglia are also a part of the NVU and are involved in the immune response (8). Normally, these cells are in a resting state, but they quickly become activated if there is a disturbance in the homeostasis of the brain such as during ischemia, infection, or an influx of albumin from the blood.</p><p>As mentioned above, the primary role of the BBB is to maintain the homeostasis of the brain by inhibiting the uncontrolled influx of molecules from the blood into the brain. There are many natural substances that, if allowed into the brain, would severely disrupt neuronal activity or cause brain damage. For example, if glutamate, a neuroexcitatory amino acid that is also present at high concentrations in food, were allowed to pass freely into the brain, severe neurotoxicity would result. A complementary second role of the BBB is to supply nutrients to the brain in a regulated manner. This task is accomplished using specific transport systems for molecules needed to maintain the cells in the brain (9). Examples of such transport systems include transporters for glucose, insulin, amino acids, and neurotransmitter precursors. A third role of the BBB is to protect the brain against toxins present in the blood. Examples include poisons, pollutants, and drugs, as well as endogenous metabolites or proteins. Specialized transport proteins, expressed on the apical (blood) side of BBB endothelial cells, are responsible for the efflux of undesirable compounds, such as lipophilic toxins, that may otherwise cross the cell membrane.</p><p>The BBB is also an important participant in the brain's immune system, and it can direct inflammatory cells to act quickly in response to changes in the neurovascular space. For example, if albumin crosses the BBB (owing to head trauma or stroke), an inflammatory reaction in the brain, microglia cellular activation, and programmed cell death may result (10). Lastly, the BBB serves as an important chemical messaging system between the CNS and the PNS. Substances released in the periphery can be transported to the brain and generate a neuronal response. Likewise, substances that are released from the brain into the periphery can generate a physiological response in a remote tissue. In particular, cytokines and neuropeptides are important mediators of such signaling. This transport/messaging system can be involved in a variety of disorders, including depression, drug addiction, Alzheimer's, and Parkinson's (6).</p><!><p>Transcellular passive diffusion: Small lipophilic compounds can passively diffuse across the endothelial membrane. In general, the more lipophilic a molecule is, the greater its ability to permeate. Examples of compounds that are transported by this mechanism include acetaminophen and fluoxetine.</p><p>Carrier-mediated transport: Various transporters are used to bring essential polar molecules into the brain. For example, there are specific amino acid, nucleoside, peptide, vitamin, and glucose transporters. Neurotransmitters such as dopamine and serotonin do not cross the BBB; however, their precursors, levodopa and tryptophan, are transported through this mechanism.</p><p>Receptor-mediated endocytosis: This is a common method of transport for large peptides and proteins and is facilitated by binding to a receptor on the membrane surfaces followed by endocytosis. Examples of molecules transported by this approach include insulin, transferrin, cytokines, and other large peptides.</p><p>Adsorptive-mediated endocytosis: Due to the highly anionic character of the BBB, cationic molecules adsorb nonspecifically to the membrane and undergo endocytosis. This mode of transport has a lower affinity and a higher capacity than receptor-mediated endocytosis. Highly positively charged molecules such as histones, cationized albumin, and arginine-containing peptides are examples of molecules transported by this approach.</p><p>TJ modulation: Typically, the TJs between the BMVECs restrict the passage of even very hydrophilic compounds from crossing the BBB via paracellular diffusion; however, if the TJs are disrupted, nonspecific passage into the brain of molecules that would normally be excluded can occur. Changes in the resistance of the TJs are usually due to a disease or administration of a drug that disrupts the proteins that make up the TJs. Experimentally, this can be accomplished through the administration of hyperosmolar solutions such as 25% mannitol. The high ionic strength causes the endothelial cells to shrink, opening the TJs. Clinically, this is employed for the treatment of brain tumors. Additionally, leukocytes and other immune cells can modify TJs or cross the BBB via transcellular mechanisms. Changes in the TJs can also occur during ischemia or brain trauma.</p><!><p>Along with these transport mechanisms of substances from the blood to the brain, efflux mechanisms also shuttle substances out of the endothelial cells and back into the blood before they have a chance to enter the brain. These carrier-mediated efflux mechanisms pose a significant challenge to pharmaceutical scientists attempting to deliver drugs into the brain. Examples of carrier-mediated efflux systems that are present at the BBB include the multidrug-resistance proteins such as P-glycoprotein (P-gp).</p><!><p>Understanding the BBB and the NVU is important for several reasons. If the integrity of the BBB is compromised due to diseases such as AIDS, undesired substances could leak into the brain, generating an immune or inflammatory response (12). Conversely, if drugs are unable to pass through the BBB, they will be ineffective for the treatment of neurological and psychiatric disorders, including depression, schizophrenia, Alzheimer's, and Parkinson's. Likewise, anticancer drugs must be able to enter the brain to treat brain tumors. In addition, the BBB plays an important role as a chemical messaging system between the CNS and the PNS. Peptides and other substances can be released in the brain or periphery, traverse the BBB, be transformed by a separate set of enzymes, and trigger a neurological or physical response at a remote location.</p><!><p>In response to the great amount of interest in the BBB as a means to understand and treat neurological diseases, several methods have been developed to investigate BBB transport. These range from in silico and cell-culture models to the use of live animals and positron emission tomography (PET) (13–15). Here, we discuss the different approaches for studying BBB transport and metabolism as well as the analytical methods that are employed to determine the kinetics of the transport processes.</p><!><p>The most common approach for monitoring the permeation of a substance across the BBB is the intravenous injection method (16). In this case, the compound is administered to the animal intravenously, and the analyte concentration is determined at different times in the brain, plasma, and cerebrospinal fluid. With this approach, all the physiological and metabolic systems remain intact, thereby providing the most realistic assessment of what actually gets into the brain. However, a major disadvantage of the intravenous injection method is that a separate animal must be used for each data point. Because animal-to-animal variability is also an issue, large numbers of animals must be employed for brain pharmacokinetic studies to obtain statistically relevant data. It is also difficult to determine the specific action of hormones, ions, nutrients, and proteins on the delivery of a drug or other substances to the brain with this approach, and there is no straightforward way to study efflux mechanisms.</p><!><p>For a more direct indication of the permeation of a substance across the BBB, the brain perfusion technique was developed by Smith & Allen (17) and Takasato et al. (18). In this technique, the compound of interest is dissolved in artificial blood, plasma, or saline and directly infused into the heart or a major vessel that leads directly to the brain by use of a perfusion pump. At a predetermined time, the perfusion is stopped, the animal is decapitated, and the amount of substance in the brain is determined. For ease of detection, radiolabeled compounds are frequently employed in these studies. This approach has an advantage over the intravenous injection method described above in that the compound does not undergo first-pass metabolism prior to entering the brain. In addition, the composition of the perfusate, the concentration of the substance under investigation, and the duration of the perfusion experiment can be carefully controlled. Inhibitors for metabolic enzymes and/or efflux transporters can be introduced into the perfusate to clarify their role in the transport and metabolism of the substance under investigation. By controlling the concentration of albumin in the perfusate, the effect of protein binding on BBB transport can also be elucidated. These experiments are most easily performed using radiolabeled substances, although liquid or gas chromatography/mass spectrometry (LC-MS or GC-MS, respectively) can also be employed.</p><p>The brain perfusion method also has several disadvantages. It is very animal and labor intensive because a new animal must be used for each experiment. As with the previous method, if kinetic studies are being performed, a large number of animals may be needed to obtain statistically relevant data. Nonspecific adsorption of a drug to the brain tissue can also lead to erroneous conclusions concerning the activity of the drug because only the free concentration of the drug is physiologically active. It is also impossible to make any correlation between drug concentrations in the brain and animal behavior with this technique.</p><!><p>Recently, PET and single-photon emission-computed tomography (SPECT) have been employed to study brain uptake kinetics, cerebral blood flow, BBB integrity, and efflux mechanisms (19–22). In these experiments, a small amount of compound labeled with a positron-emitting radionuclide is injected intravenously and allowed to distribute to the different tissues in the body. The emitted γ radiation is then measured as a function of tissue depth by the instrument. Computer software is employed to create a three-dimensional image of the distribution of the substance in the brain and other tissues. PET provides higher-resolution images than SPECT does. Both techniques are noninvasive and can be used on both human and animal subjects. A major advantage of these methods is that it is possible to obtain excellent spatial resolution regarding drug distribution in a noninvasive manner. Figure 3 shows the use of PET to investigate the BBB permeation and tissue distribution of a series of 18F-labeled fluoropyridinyl compounds in rats (22). The authors of this study determined the kinetics of BBB permeation by measuring the radioactivity in the cerebral areas of the PET images and plotting the percentage of the original injected dose in that region over time.</p><p>Some disadvantages of tomography are that it requires expensive instrumentation and that radiolabeled analogs of the substance under study must be synthesized (13). PET uses 11C- or 18F-labeled compounds, whereas SPECT employs 123I. The short half-life of 11C-labeled compounds (20 min) means that they must be synthesized on site prior to administration. Compounds labeled with radioactive fluorine or iodine have longer half-lives, but they can exhibit different transport properties that can confound the results. Another major drawback of this method is that, because only radioactivity is measured, it does not distinguish between transport of the drug and its metabolites. It is also impossible to determine the free versus bound fraction in vivo (14).</p><!><p>Microdialysis sampling was developed in the 1970s (23–27) as a minimally invasive method for monitoring neurotransmitters in the brain. It has been extensively used to investigate the transport and metabolism of drugs and other substances at the BBB. Compared with the aforementioned techniques, microdialysis sampling offers several key advantages. In microdialysis sampling, tissue concentrations are determined by direct sampling of the chemical makeup of the interstitial fluid. Therefore, along with drug concentrations, other analytes of importance, such as neurotransmitters and metabolic markers, can be measured at the same time. In addition, because long-term sampling can be accomplished on a single animal, it can serve as its own control, leading to a significant reduction in the number of animals needed to obtain statistically significant results in tissue-distribution studies.</p><p>To perform microdialysis sampling, a probe consisting of a short length of hollow-fiber dialysis membrane connected to inlet and outlet tubing is implanted in the brain or other tissue. A solution that is similar in composition and ionic strength to the extracellular fluid (ECF) of the tissue of interest is then pumped slowly through the probe. Small molecules in the extracellular space that are not present in the perfusate diffuse across the membrane based on their concentration gradients and are transported by the syringe pump to a fraction collector or online analysis system. Figure 4 shows a typical microdialysis sampling setup.</p><p>For studies involving the BBB, a stainless-steel concentric cannula probe is most commonly used for sampling the brain ECF, whereas a flexible probe can be used to monitor the blood levels (28). The concentric cannula probe is very rigid, whereas the flexible probe is composed of side-by-side fused silica capillaries or flexible tubing and can bend when the animal moves, minimizing any damage to blood vessels. An alternative method of blood sampling is the Culex automated blood sampler that is used to sample blood using a push-pull method; it then replaces the lost fluid with isotonic saline. The Culex blood-sampling method has been employed in conjunction with brain microdialysis sampling to investigate the transport of nicotine across the BBB (29). Due to the small size and relatively noninvasive nature of the microdialysis sampling probes, multiple probes may be used in a single animal. Therefore, it is also possible to measure blood, brain, and tissue concentrations of drugs or endogenous substances simultaneously.</p><!><p>A major advantage of microdialysis sampling over the in vivo methods discussed above is that there is no net fluid loss from the tissue as a result of the sampling process. Thus, long-term studies are possible with minimal disruption of the physiological system. A unique advantage of this method for BBB studies is the possibility of continuously monitoring the concentrations of a substance in both the brain and blood of an awake and freely moving animal, which makes it feasible to correlate brain concentrations of a substance with behavior. Also, because of the low-molecular-weight cutoff of the dialysis membranes, samples obtained via microdialysis sampling are protein free, meaning that only the physiologically active free-drug concentration is measured by this technique. Enzymes are also excluded, preventing metabolic degradation of the sample contents.</p><p>Another important advantage of microdialysis sampling for CNS-active drugs is the ability to simultaneously monitor drug concentrations and neurotransmitter release in the brain and correlate them to both blood levels and behavior. Figure 5 shows a nice example of the use of microdialysis sampling for a BBB study. A rat was given a single intravenous injection of methylphenidate (Ritalin) (30). Brain and blood sampling were accomplished using a concentric cannula probe and a flexible probe, respectively. Dialysates were collected off-line, and the concentrations of methylphenidate and dopamine in both the brain and blood were determined using LC with electrochemical detection. In this manner, it was possible to measure the transport of Ritalin across the BBB as well as monitor its effect on catecholamine release. Lastly, through the use of a RatTurn©, the extracellular concentration of these substances could be directly correlated with the overall activity level of the rat.</p><p>In addition to looking at transport across the BBB, microdialysis sampling can also be employed for investigations involving site-specific transport and metabolism of drugs. An excellent example of such sampling is a report by de Lange et al. (31, 32), who examined differences in the distributions of an intravenous injection of methotrexate in tumor versus healthy brain tissue. Microdialysis sampling has also been used to investigate the metabolism of drugs and neuropeptides in specific areas of the brain (33–35). For substance P, the results obtained for metabolism in the neurovascular space were then compared with those obtained using cell-culture models of the BBB.</p><!><p>A major concern with microdialysis sampling for BBB transport studies is the effect of probe implantation on the brain and the integrity of the BBB (26, 36). In particular, if the studies are to be performed over a long period of time, tissue damage associated with probe implantation and the potential for an immune response must be taken into consideration because fibrosis or gliosis occurs several days after probe implantation (37–39). A study by Jaquins-Gerstl & Michael (40) used immune labeling of tissue slices for glial fibrillary acidic protein to investigate glial activation by implantation of the microdialysis sampling probe. They found a 200% increase in glia immunoreactive cells 24 h after probe implantation (40). Jaquins-Gerstl et al. (41) also recently reported that inclusion of dexamethasone in the perfusate can reduce this glial response.</p><p>The integrity of the BBB following probe implantation has also been a controversial issue in brain microdialysis sampling (26, 42). Studies conducted using autoradiography with 14C-α-aminoisobutyric acid (which does not cross the BBB under normal conditions) as well as transport characteristics of hydrophilic and moderately lipophilic drugs (following intravenous injection) post surgery indicate that the BBB integrity is maintained overall (42, 43). However, other studies have shown a significant effect of probe implantation on BBB permeability by using 51Cr-EDTA transport (44). More recently, Gerhardt's group (45) showed that, although there was tissue damage and glial activation following probe implantation for both ceramic microelectrode arrays and microdialysis sampling, the BBB remained intact.</p><!><p>To quantitate the amount of an exogenous substance entering the brain from the blood, it is necessary to calibrate the microdialysis sampling probe (46). Calibration is usually accomplished by measuring the relative recovery, which is defined as the ratio of the concentration of the analyte inside the probe to that in the sampling solution. This number depends on a number of parameters, including the charge, size, and lipophilicity of the analyte, the microdialysis sampling flow rate, and the composition of the dialysis membrane. Although in vitro calibration can be accomplished by placing the probe in a stirred beaker containing the analyte of interest, it is rarely equivalent to the recovery in vivo (27, 46, 47).</p><p>The no-net-flux method is the "gold standard" for quantitation by microdialysis sampling (48). In this method, the animal is administered a constant intravenous dose of a drug until steady-state conditions are confirmed in the tissue of interest. After steady-state conditions have been reached in the brain, specified concentrations above and below the expected values for the compound of interest are perfused through the brain microdialysis probe. When the concentration in the brain is equal to that of the perfusate, the concentration in the dialysate does not change. This concentration of the substance of interest in the ECF of the brain can be determined by plotting the concentration of the dialysate minus the concentration of the perfusate (Cd–Cp) versus the concentration of the substance in the perfusate (Cp). The zero intercept is the concentration of the substance in the brain ECF (46, 49). This method provides the most accurate measurement of exogenous substances in the ECF of the brain but is a very time-consuming experiment.</p><p>The technique of retrodialysis has been evaluated to quantitate the free concentration of a substance in the brain ECF (27, 47, 50). This technique can be performed in one of two ways. The first uses the actual exogenous substance of interest as its own control and assumes that analyte delivery through the probe into the tissue is equivalent to the recovery. In this experiment, the substance under study is dissolved in the perfusate and delivered via the microdialysis probe to the brain region of interest. The delivery is then calculated as the ratio of the concentration of the substance coming out of the probe to the original perfusate concentration. Following the delivery experiment, the substance is washed out of the tissue by perfusion with artificial cerebrospinal fluid, and the drug is administered peripherally for the BBB experiment. The delivery is evaluated again at the end of the experiment. Experiments performed with caffeine and acetaminophen obtained ECF concentrations with the delivery method that were higher for caffeine but lower for acetaminophen in comparison to those obtained using the no-net-flux method (50). The higher values obtained for caffeine were believed to be due to saturable active transport across the BBB.</p><p>An alternative to the delivery experiment described above is retrodialysis using a calibrator (51, 52). This approach uses a compound with structural and physical properties similar to those of the analyte of interest as an internal standard. This calibrator is added to the perfusion medium, and the delivery of the substance through the probe is monitored continuously throughout the BBB experiment. Changes in the delivery of the calibrator during the course of the experiment correlate with changes in analyte recovery. One of the difficulties is finding a calibrator compound that has structural characteristics similar to those of the analyte of interest but is not physiologically active. Recently, deuterated analogs were shown to be useful calibrators when used with mass-spectrometric analysis (53).</p><p>Another calibration approach is to sample at a very low flow rate (100 nl min−1) in which the perfusate approaches complete equilibrium with the ECF, yielding close to quantitative recovery (within error) (54). The disadvantage of this approach is that only very small sample volumes (100 nl or less) are obtained per minute, so highly sensitive analytical methods that can analyze such small samples are necessary for good temporal resolution. Capillary and microchip electrophoresis are extremely useful methods for the analysis of such small sample volumes (55, 56). Another approach is to use the MetaQuant probe, which samples at a very low flow rate to obtain quantitative recoveries but then supplies makeup flow after sampling to provide an adequate sample for conventional analysis methods (57).</p><p>Despite some of the concerns regarding quantitation and calibration, microdialysis sampling is still extensively employed to evaluate the transport of drugs across the BBB (14, 26). Significant advantages of microdialysis sampling for these studies are that the animal can be used as its own control and several parameters (different concentrations of drug, administration of efflux inhibitor, etc.) can be investigated using the same animal. Doing so provides a clear advantage over the previously described methods, in which a different animal is needed for each time point and experimental variable. It is also impossible to determine the free (active) concentration of a drug in the brain of a freely moving animal by any other technique (58).</p><!><p>LC is the most common method for analysis of microdialysis samples. To maintain good temporal resolution and high sensitivity, microbore and capillary columns are commonly employed. Detection methods include ultraviolet (UV), laser-induced fluorescence (LIF), and electrochemical (EC) detection, as well as MS (59). In an early BBB study, microbore LC-UV was used in conjunction with microdialysis sampling to determine the effects of anesthesia on the transport of tacrine across the BBB (60). More recently, microdialysis sampling, in conjunction with LC-EC, has been used to correlate brain concentrations of mercaptoproprionic acid, a chemical model for epilepsy, with electrical activity in the brain (61).</p><p>LC-MS has proven to be a popular method for the analysis of microdialysis samples due to its high specificity for the analytes of interest and potential for low limits of detection. Special considerations must be made to remove the high salt concentrations of the dialysate sample prior to analysis to minimize ion suppression (62). Microdialysis sampling with LC-MS analysis was used to investigate the metabolism of substance P in the neurovascular space and compare it with the metabolism of the peptide by bovine BMVECs (63). Using LC-MS, researchers can employ deuterated calibrators for more accurate quantitation of brain extracellular concentrations. Bengtsson et al. (53) used deuterated morphine as a calibrator to monitor the recovery of morphine and evaluate the performance of the probes.</p><p>The transport of L-3,4-dihydroxyphenylalanine (L-DOPA) into an "on-rat" collection system was investigated using LC-EC (64). Figure 6 shows a diagram of the experimental setup. An Alzet™ osmotic pump implanted in the back of the rat was used to provide the flow through the probe implanted in the brain. Because osmotic pumps do not require a power source, on-rat collection was made possible by attaching a small vial with a septum to the outlet of the microdialysis probe. The rat was then injected with L-DOPA, and the collection vial was removed from the rat's head at specific time intervals. The contents were subsequently analyzed for DOPAC (3,4-dihydroxyphenylacetic acid), HVA (homovanillic acid), and 5-HIAA (5-hydroxyindoleacetic acid) by LC-EC.</p><p>Although most analyses of substances for BBB transport studies are developed using LC as the separation method, capillary electrophoresis (CE) is useful for the analysis of microdialysis samples owing to its low-sample-volume requirements. Thus, it is possible to employ very low flow rates to improve analyte recovery or, alternatively, to employ a single microdialysis sample of 1 to 10 μl to determine several different analytes. The ability to measure substances in very small volumes also improves temporal resolution, which can be particularly important if neurotransmitter release is being measured in conjunction with brain drug concentrations.</p><p>One of the first experiments using CE to analyze microdialysis samples involved the detection of kynurenine following peripheral administration of tryptophan (65). Samples were collected off-line and measured using CE-EC. A more recent example is the use of microdialysis sampling in conjunction with CE-LIF for the simultaneous determination of the concentrations of carbamathione in the brain following administration of disulfuram and the release of GABA (γ-aminobutyric acid) and glutamate (66, 67).</p><p>Microdialysis sampling coupled to microchip electrophoresis has recently been employed to simultaneously monitor neurotransmitter concentrations and BBB permeability (68). Rats were injected with a bolus dose of fluorescein that was used as an indicator of BBB integrity. The amino acids in the microdialysate were derivatized online using a microchip reactor. The detection of the excitatory amino acid neurotransmitters aspartate and glutamate, along with fluorescein that had entered the brain, is shown in Figure 7.</p><p>Other approaches for the analysis of microdialysis samples include AMS (accelerator mass spectrometry) and electron spin resonance. AMS has been employed to investigate the transport of 14C-labeled polyphenols across the BBB as well as to monitor brain pharmacokinetics of morphine at very low concentrations (69, 70). Electron spin resonance has also been employed to detect the permeation of free-radical compounds. Representative examples of the different analytical methods in conjunction with microdialysis sampling are provided in Table 1 and in Reference 56.</p><!><p>Cell-culture systems have evolved as a first-pass screening tool for BBB permeability to drugs and other compounds. These systems can also be used to investigate the effects of the metabolic component of the BBB on transport. Using primary cultures isolated from the gray matter of bovine brains, Audus & Borchardt (71) developed one of the first cell-culture models of the BBB. These primary cultures express metabolic enzymes crucial for investigating peptide metabolism (72–74), and they exhibit many of the physical characteristics typical of the BBB, including decreased pinocytosis and minimal fenestration. Efflux transporters as well as specific uptake transport systems are expressed by these cells and have been extensively characterized, making them a useful method for investigating BBB permeability (75).</p><!><p>Bovine and porcine brain microvessel endothelial cells, known as BBMECs and PBMECs, respectively, are the most common primary cell models for BBB transport studies (75). For transport studies using the BBMEC culture system, cells are isolated from whole bovine brains, grown to confluency on polycarbonate membranes, and mounted in Side-by-Side™ diffusion chambers (Figure 8). This setup enables temperature maintenance via external circulating water baths as well as constant stirring, driven by an external console, to eliminate the effect of a stagnant layer at the monolayer surface. The Side-by-Side diffusion chambers provide access to both sides of the cell monolayer, allowing bidirectional permeation to be investigated by spiking either the apical or basolateral chambers with the compound of interest and monitoring transport in either direction. Once the compound is added to the donor side of the cell monolayer, aliquots are removed at discrete time points from the receiver chamber and stored for further analysis. To avoid changes in volume that could affect flux in the system, fluid is always replaced following the removal of each aliquot. Another advantage of this in vitro method is the ability to manipulate variables, including temperature and compound concentration, to evaluate their effects on permeability. For example, by performing experiments at 4°C, the effect(s) of reduced ATP activity on the BBB transport of a specific compound can be characterized (76), which enables the researcher to distinguish between passive and active transport processes. Passive transport does not require energy; however, active transport mechanisms, such as carrier- and receptor-mediated transport, require ATP. If a compound is actively transported across the BBB, the temperature dependence of transport can be used to determine the activation energy of the transport process. Receptor-mediated processes can be further elucidated by determining the effect of competing ligands for the same receptor on the transport process.</p><p>In addition to the advantages that the primary cell line affords in these assays, the Side-by-Side diffusion chambers also provide advantages over Transwell® culture plates that have been used in Caco-2 studies. In the Transwell system, cells are grown on a polycarbonate insert tray that fits within a larger multiwelled culture plate. An issue that can arise when using this system with BBMECs, however, is the propensity for the cells not to reach confluency at the edges of the inserts, negatively impacting permeability data. Temperature control is also more difficult in these systems, and constant stirring presents a challenge. However, the Transwell system has some merit in studies involving the BBB, as has been demonstrated with coculture systems and immortalized cell lines (described below).</p><p>The BBMEC culture model also expresses the efflux transporter P-gp, which is a common obstacle to the delivery of anticancer agents for brain tumors. Therefore, in addition to transport studies aimed at determining permeation across the BBB, Rhodamine 123 uptake assays with BBMECs serve as a convenient in vitro assessment of a compound's affinity for P-gp. The fluores-cent dye Rhodamine 123 is known to be a substrate for P-gp, and accumulation of this dye within cultured BBMECs can be characterized via intracellular fluorescence determinations. Competitive assays in which BBMECs are incubated with both Rhodamine 123 and the compound of interest exhibit an increased accumulation of Rhodamine 123 intracellularly if the unlabeled drug of interest is also a P-gp substrate. In other words, a substrate for P-gp prevents the efflux of Rhodamine via P-gp; therefore, an increase in the fluorescence signal is observed due to a higher-than-normal accumulation of the dye within the cells, as can be seen with the anticancer agent Paclitaxol in Figure 9 (76, 77). This assay is useful for investigating structural modifications that may alter a drug's affinity for P-gp, thus providing insight into the modifications necessary to formulate CNS-active therapeutics that circumvent efflux mechanisms (76, 77).</p><!><p>In recent years, work with in vitro models of the BBB has focused on the development of immortalized brain endothelial cell lines (78) to decrease the time necessary to reach confluency (7 versus 14 days) as well as lessen the workload on the laboratory scientist because the isolation procedures are time and labor intensive for primary cells. The immortalized cell lines form monolayers but not complete TJs, leading to a so-called leaky barrier. Therefore, they can be used to explore endothelial uptake of compounds, but they are not useful for transport studies. Given the leakiness of the immortalized cell lines, cells derived from other tissues have been evaluated as models for BBB transport. The Mardin-Darby canine kidney cell line transfected with the multidrug resistance gene (MDR1) is currently the best noncerebral epithelial model for BBB permeability studies (15, 79).</p><p>Additionally, the role of other cell types in the behavior of the brain endothelium is of interest and has driven the development of coculture systems. Specifically, the role of astrocytes in TJ formation has been investigated using cocultures with Transwell systems and endothelial cells grown in astrocyte-stimulated media (75, 80–82). Typically, the endothelial cells are grown on the Transwell insert tray, whereas the astrocytes are cultured on the bottom of the Transwell multiwell plate (Figure 8).</p><!><p>Once an in vitro method is chosen as a BBB model, researchers must have in place analytical techniques with sufficient sensitivity to monitor the transport of the compounds of interest across the cell monolayers. Low limits of detection are often required, especially for potent compounds. Therefore, several factors must be kept in mind when choosing an analytical technique to be used for the investigation of the transport or metabolism of drugs in cell-culture studies. These factors include the anticipated concentrations of the drug on both the donor and receiver sides of the transport assay, the ability to label the compounds of interest (fluorescently or radioactively) for subsequent detection, sample-volume requirements, and the compatibility of the cellular matrix with the desired technique.</p><!><p>Once in vitro transport studies are completed using the BBMEC culture model, further analyses must be performed to quantitate the percent transport or apparent permeability coefficient of the compounds of interest. Essential to these determinations is the use of proper control compounds to assess cell monolayer integrity and, therefore, validate the experimentally determined transport properties of the drug or other compound. Typically, low-molecular-weight compounds that are membrane impermeant are added to the apical side of the diffusion chambers at the end of a study to assess the integrity of the endothelial cell TJs. Radiolabeled sucrose or fluorescein is most commonly employed. Samples are removed from the basolateral side of the membrane over the course of 1 h and are easily analyzed via scintillation counting or fluorescence spectrophotometry (often performed on basic benchtop plate readers). The ability of these compounds to permeate must fall within a defined range to accurately assess the transport of the compound of interest (83).</p><p>Another important consideration for in vitro transport studies is the maintenance of healthy cells during transport and metabolism studies. To achieve this, experiments must be carried out under appropriate conditions. Transport media must maintain isotonicity, TJ integrity, and physiological pH (7.4) as well as provide an energy source (typically glucose) for studies lasting longer than 1 h (71). Once the conditions are met to maintain healthy cells throughout the study, the sample requirements for the downstream analytical technique must also be taken into consideration.</p><!><p>One of the simplest methods to determine the transport of compounds across the BBB is the use of radiolabeled compounds and scintillation counting (84, 85). Scintillation counting is frequently employed in competitive studies that are performed to determine whether saturable, carrier-mediated processes are responsible for a compound's permeation. In these studies, a radiolabeled substance is added to the diffusion chamber, and its permeation is determined. Then, the effect of the unlabeled compound on the permeation of the radiolabeled compound is evaluated. Alterations in the permeation of the labeled compound indicate that the two species are competing for the same transport system. Once samples are diluted in scintillation fluid, additional sample preparation is unnecessary, adding to the simplicity of this technique. This method, however, can be costly due to the expense of producing labeled compounds and the required special training for handling radioactive materials.</p><!><p>Because the cell studies employ relatively small (milliliter) volumes of sample on either side of the membrane in the Transwells, sample-volume requirements of the analytical method can become an issue for studies where multiple samples will be removed from the diffusion cell. CE provides considerable advantages in addressing this aspect of cell-based studies. In particular, fluorescence detection has been utilized with CE to analyze peptide transport samples. However, one drawback is the need for sample derivatization prior to analysis (86). Most labeling techniques require reaction with amine groups such as the peptide N terminus or lysine residues. Therefore, peptides with modified N termini or those lacking a lysine residue are not detected. Nonetheless, the small sample volumes necessary for CE analysis and the sensitivity of fluorescence detection still make it an attractive technique.</p><p>Utilizing CE-LIF Freed et al. (33, 86) investigated the metabolism and transport of substance P at the BBB following sample derivatization by NDA/CN−. Transport of this neuropeptide was characterized in both the apical-to-basolateral (blood-to-brain) and basolateral-to-apical (brain-to-blood) direction. Substance P was found to cross the BBB at similar percentages over the concentration range studied. Additionally, the temperature dependence of transport was determined by examining transport at 37, 25, and 4°C; transport of the peptide was completely abolished at 4°C.</p><!><p>The matrices with high salt content that are essential for cellular function can present several downstream analytical challenges, especially when LC-MS is the desired technique. First, high-salt solutions are incompatible with electrospray ionization because significant damage to the ESI probe can occur. The first line of defense against such damage to expensive MS equipment is the use of a diverter valve to flush the salts present in the void volume to waste prior to introduction of the effluent into the MS. A second line of defense is to dilute samples in the mobile phase to decrease the salt concentration. Alternatively, liquid-liquid or solid-phase extraction can be employed prior to analysis (83).</p><p>Chappa et al. (63, 87) investigated substance P transport across a BBMEC monolayer, employing LC-MS/MS for peptide quantitation. One advantage of the LC-MS method over CE-LIF is the ability to investigate the effects of lower substance P concentrations on BBB transport owing to improved limits of detection. In this work, the transport of substance P was found to be a saturable process (87). Competitive studies with radiolabeled substance P provided further evidence for a carrier-mediated transport mechanism. The saturation kinetics in conjunction with the competition studies with radiolabeled peptides, as well as the temperature dependence of substance P transport, are all indicative of a carrier-mediated transport process and also suggest the involvement of a specific receptor, neurokinin-1. The presence of this receptor in the BBMEC culture model was confirmed via Western blot (87). The BBB can also act as a significant metabolic barrier for the delivery of peptides into the brain. The metabolism of substance P by enzymes present at the BBB was also investigated using the BBMEC model and LC-MS analysis (63). Figure 10 shows the products of substance P degradation by the BBB as a function of time.</p><p>Another neuropeptide of interest, dynorphin (Dyn), is currently being investigated because of its involvement in pain and addiction through its interaction with κ-opioid receptors (89). Work with analogs of dynorphin, specifically Dyn A(1–11)NH2, utilized BBMECs and LC-MS/MS to evaluate the effects of structural changes to the peptide backbone on BBB permeability. All the dynorphin analogs exhibited a similar BBB permeability lag time, indicating a carrier-mediated transport mechanism was responsible for their transport (Figure 11) (90). However, the structural changes altered their ability to permeate the BBB. Peptide analogs that exhibit improved blood-to-brain transport have the potential to serve as improved therapeutics for cocaine addiction, whereas those that do not significantly enter the brain could be used to treat peripheral pain, selectively targeting κ receptors in other tissues.</p><p>BBMEC assays are also employed to determine the BBB permeation of small-molecule drugs. Desino et al. (91) investigated the BBB permeation of TCP-FA4, a derivative of the monoamine oxidase inhibitor, tranylcypromine, using the BBMEC culture model and subsequent analysis via LC-MS/MS. Compared with the parent drug, the derivative demonstrated improved BBB permeability as well as decreased monoamine oxidase inhibition. Temperature and inhibition studies suggest that passive diffusion is responsible for the compound's permeation. TCP-FA4 improved neuronal cell survival in the presence of the toxic protein β-amyloid, and it increased the level of BDNF, a protein involved in neuronal survival and differentiation, in HUVECs (human umbilical vein endothelial cells), suggesting its potential use as a neuroprotective agent.</p><!><p>More recently, microfluidic devices have been used to investigate the transport and metabolism of substances by endothelial cells as well as the release of signaling compounds (92). These devices can be fabricated from biocompatible materials and can incorporate multiple functions, including cell culture, stimulation, and analysis (93–97), into a single device. An advantage of the microfluidic platform for studies of endothelial cell function is that the chip can be designed so that the size of the channels and flow of fluid over the cells mimic physiological conditions, and cells can be cultured directly into a microchip device (98). Because the overall channel volumes are extremely small (nanoliters), molecules secreted into the perfusing fluid are only slightly diluted and can be detected with better sensitivity than with bulk analysis methods.</p><p>Nitric oxide (NO) is a vasodilator that increases the permeability of the BBB (99). Cytokines and other inflammatory agents can stimulate inducible NO synthase and increase NO production of brain endothelial cells. Therefore, methods that are capable of monitoring NO production by endothelial cells would be very useful for understanding the role that inducible NO synthase plays in BBB permeation. Under physiological conditions, NO has a very short half-life. Therefore, the NO oxidation products nitrite (NO2−) and/or nitrate (NO3−) are normally used as an indirect measurement of NO production. However, the rapid analyses afforded by microfluidic devices allow NO to be measured directly.</p><p>The detection of NO release by bovine pulmonary artery endothelial cells (bPAECs) has been investigated using cells grown inside the channels of a polydimethylsiloxane (PDMS) microchip. A Nafion®-coated carbon electrode was employed to detect NO (100, 101). Release of NO from the bPAECs was accomplished by changing the perfusate from pure buffer to ATP-containing buffer. ATP binds the P2y purinergic receptors in the bPAECs and stimulates the release of NO. Due to the small dimensions of the channel and the close proximity of the electrode, it was possible to measure the NO release from the cells with good sensitivity and temporal resolution.</p><p>A benefit of microfluidics is the relative ease of fabricating devices that can combine multiple processes into a single substrate. Researchers have constructed devices that provide a representative model of the vasculature and that monitor the release of NO from erythrocytes (102, 103). Figure 12 shows a diagram of the current version of such a device. Flow-through channels were fabricated in a PDMS slab and then sealed to a second piece of PDMS containing sample wells with a polycarbonate membrane sandwiched in the middle. Cells were then seeded onto the membrane and allowed to grow to confluence. The device was built so that the TEER of the bPAEC layer could be measured to monitor the growth of the cells (103). Once the endothelial cells were grown to confluency, red blood cells were pumped through the fluidic channels in close proximity to the endothelial cells on the membrane. Release of ATP from the red blood cells stimulated the release of NO from the endothelial cells. DAF-FM was incubated with the endothelial cells to quantify the concentration of NO produced in the system. Unlike previous static models, this system provides a realistic in vitro representation of blood vessel circulation and the secretion of NO within the cells. Furthermore, these microchips were designed to allow many cells to be run in parallel and analyzed via high-throughput means. Multiple detection wells have been fabricated onto a single chip and analyzed using a standard 96-well-plate reader (102). This device could also have applications for BBB transport and metabolism studies.</p><p>Devices that can be used for both cell permeation and transport studies have been reported. In one study, investigators fabricated a microchip in which a top PDMS layer containing cell-culture chambers was sealed to a bottom glass fluidic network with a membrane sandwiched in the middle. Unlike the open-air well described above, the cells were immobilized in a closed channel. Pneumatic valves were incorporated into the design and used to trap HUVECs in predefined regions of the chip. Once the cells reached confluency, FITC (fluorescein isothiocyanate)-labeled albumin was perfused through the microfluidic channels and allowed to permeate the cell layer (104). After a designated period of time, a fluorescence measurement was made in the portion of the cell channel not in direct contact with the microfluidic channel. This was done to avoid detecting nonpermeated FITC flowing through the channel below. Introducing a compound known to increase permeation led to an increased amount of fluorescence in the reservoir, as expected (104).</p><p>Other cell types that are representative of those present in the NVU have also been investigated in the microfluidic format. One recent example is the development of a PDMS device to monitor the release of dopamine and norepinephrine from PC-12 cells. This device employed a pneumatic valve to isolate the cell perfusion channel from the electrophoresis channel. The valve could be rapidly actuated to allow a small plug of the cell perfusate sample to be introduced into the separation channel. Analytes were separated electrophoretically and then detected by fluorescence following postchannel derivatization (105, 106).</p><!><p>Methods are continuing to be developed to monitor the transport and metabolism of molecules at the BBB as well as to evaluate the integrity of the BBB. A number of new in vitro cell lines have been developed and evaluated for properties that mimic the BBB. The integration of these cell lines into a microfluidic format will facilitate the high-throughput screening of drugs and other substances for their ability to cross the BBB.</p><p>The ability to correlate behavior with concentrations of drugs or other substances in the brain will become increasingly important in drug development and in studies aiming to better understand neurological diseases. Microdialysis sampling enables the simultaneous monitoring of drug concentrations and neurotransmitter release in awake, freely moving animals. However, these animals are normally tethered to the syringe pump by tubing and a liquid swivel. Studies on freely roaming animals have been performed employing an osmotic pump and on-animal collection (64). However, sample analysis has thus far been performed only off-animal. In the future, the development of sampling and analysis methods that could be placed on-animal, allowing the animals to be freely roaming, would enable researchers to better investigate the effects of CNS-active drugs on behavior. Therefore, the next necessary advancement for neurochemical investigations is the development of portable on-animal separation-based sensors for drugs and neurotransmitters (107). Microfluidics, microelectronics, and telemetry are all tools that can be employed to make these sensors a reality. The advent of such devices will serve to elucidate neurochemical associations between drug administration, neurotransmitter release, and resulting behaviors.</p>

PubMed Author Manuscript

Cysteine Conjugate \xce\xb2-Lyase Activity of Rat Erythrocytes and Formation of \xce\xb2-Lyase-Derived Globin Monoadducts and Cross-Links after in Vitro Exposure of Erythrocytes to S-(1,2-Dichlorovinyl)-L-cysteine

S-(1,2-Dichlorovinyl)-L-cysteine (DCVC), a mutagenic and nephrotoxic metabolite of trichloroethylene can be bioactivated to reactive metabolites, S-(1,2-dichlorovinyl)-L-cysteine sulfoxide (DCVCS) or chlorothioketene and/or 2-chlorothionoacetyl chloride, by cysteine conjugate S-oxidase (S-oxidase) and cysteine conjugate \xce\xb2-lyase (\xce\xb2-lyase), respectively. Previously, we characterized reactivity of DCVCS with Hb upon incubation of erythrocytes with DCVCS and provided evidence for formation of distinct DCVCS-Hb monoadducts and cross-links both in isolated erythrocytes and in rats given DCVCS. In the present study, we investigated DCVC bioactivation and Hb adduct formation in isolated rat erythrocytes incubated with DCVC (9 and 450 \xce\xbcM) at 37 \xc2\xb0C, pH 7.4. The results suggested no DCVCS monoadducts or cross-links were formed; however, LC/ESI/MS and MALDI/MS of trypsin digested globin peptides revealed presence of \xce\xb2-lyase\xe2\x80\x93derived globin monoadducts and cross-links. Adducts and cross-links in which the sulfur atom of the reactive sulfur intermediates were replaced by oxygen have also been detected. Use of SDS-PAGE provided additional evidence for globin cross-link formation in the presence of DCVC. Interestingly, the MS results suggest the observed peptide selectivity of the \xce\xb2-lyase\xe2\x80\x93derived reactive sulfur/oxygen-containing species was different than that previously observed with DCVCS. While these results suggested erythrocytes have \xce\xb2-lyase but not S-oxidase activity, further support for this hypothesis was obtained using S-(2-benzothiazolyl)-L-cysteine, an alternative substrate for \xce\xb2-lyases. Collectively, the results demonstrate the utility of Hb adducts and cross-links to characterize the metabolic pathway responsible for DCVC bioactivation in erythrocytes and to provide distinct biomarkers for each reactive metabolite.

cysteine_conjugate_\xce\xb2-lyase_activity_of_rat_erythrocytes_and_formation_of_\xce\xb2-lyase-deriv

Introduction<!>Caution<!>Materials<!>Animals<!>Incubations of Erythrocytes with DCVC<!>Hemolysis<!>Globin Chain Cross-link Analysis by SDS-PAGE<!>Mass Spectra of Trypsin Digest<!>Mass Spectral Analyses of Tryptic Peptides<!>\xce\xb2-Lyase Activity in RBCs<!>S-Oxidase Activity in RBCs<!>HPLC Analyses<!>Statistics<!>Hemolysis<!>Mass Spectral Analyses. Monoadducted Peptides Modified by \xce\xb2-Lyase-derived Reactive Intermediates<!>Cross-linked Peptides Modified by \xce\xb2-Lyase-derived Reactive Intermediates<!>Peptides Modified by S-Oxidase-derived Monoadducts and Cross-links<!>Globin Chain Cross-link Analysis by SDS-PAGE<!>\xce\xb2-Lyase Activity in RBCs<!>S-Oxidation in RBCs<!>Discussion<!>

<p>Trichloroethylene (TCE) is a common environmental pollutant listed in the Eleventh Report on Carcinogens as "reasonably anticipated to be a human carcinogen" (1). Renal-cell carcinomas from occupationally exposed workers exhibited mutation in the von Hippel-Landau tumor suppressor gene. S-(1,2-Dichlorovinyl)-L-cysteine (DCVC), a mutagenic and nephrotoxic metabolite of TCE, has been implicated in the renal toxicity and carcinogenicity of TCE (2–4). Formation of DCVC occurs via a GSH-dependent pathway that primarily occurs in the liver and initially converts TCE to S-(1,2-dichlorovinyl)-L-glutathione (DCVG). Following systemic distribution and/or secretion into the bile duct, DCVG is converted to DCVC by γ-glutamyl transpeptidase and cysteinylglycine dipeptidases present in the kidneys, bile duct, and intestines (3–5). After human exposure to TCE, DCVG was detected in blood within 30 min and its presence persisted for up to 12 h (3). DCVC was also detectable in the blood of male rats for up to 48 h after exposure to TCE (5). DCVC can be converted to the mercapturic acid derivative, N-acetyl DCVC, in the liver or the kidneys (4). Excretion of N-acetyl DCVC has been detected in the urine over 48 h after a 6 h human exposure to TCE (6).</p><p>Bioactivation of DCVC in the kidneys via the cysteine conjugate β-lyase (β-lyase) pathway was suggested to play a major role in DCVC cytotoxicity in proximal tubule cells of Sprague-Dawley rats (reviewed in 4, 7, 8). β-Lyases are pyridoxal 5′-phosphate-dependent enzymes that catalyze β-elimination reactions (7). At least eleven β-lyase enzymes that catalyze cysteine S-conjugate β-elimination reactions have been identified (7). β-Elimination proceeds by cleavage of the thioether linkage and results in generation of pyruvate, ammonia, and depending upon the chemical structure of the cysteine-S-conjugate, a stable thiol or an electrophilic reactive sulfur-containing fragment (8, 9). In the case of DCVC, β-elimination reaction yields a highly reactive thiol species, 1,2-dichloroethenethiolate that could tautomerize to yield 2-chlorothionoacetyl chloride or lose HCl to form chlorothioketene (Figure 1). Both reactive metabolites could covalently bind macromolecules, such as proteins and DNA (9–11). Renal β-lyase activity was shown to be most abundant in the cytosol of rats and humans, however, overall renal cytosolic β-lyase activity in human samples was approximately 10% that of rats (12, 13). Because purified recombinant human kidney glutamine transaminase K (GTK) exhibited higher β-lyase activity with cysteine S-conjugates than rat kidney GTK (14), the low activity detected in human kidney samples warrants further investigation to better characterize the variability of renal cytosolic β-lyase activity in human kidneys.</p><p>DCVC is also a substrate for S-oxidation (S-oxidase) by flavin-containing monooxygenase 3 (FMO3) (15, 16). The resulting metabolite, S-(1,2-dichlorovinyl)-L-cysteine sulfoxide (DCVCS), is a reactive electrophile that was a much more potent nephrotoxicant to rats in vivo than DCVC (2). DCVCS induced necrosis, apoptosis, GSH depletion, and mitochondrial dysfunction in primary cultures of human proximal tubule cells (17). DCVCS treatment depleted hepatic and renal GSH and a DCVCS-GSH adduct was detected in the bile of rats given DCVCS (18). Although the FMO3 bioactivation pathway may be contributing to DCVC nephrotoxicity and/or mutagenicity, the expression levels of FMO3 in the human kidney are low in comparison to human liver (16, 19, 20). Incubations of DCVC with human liver microsomes in the presence of NADPH resulted in formation of DCVCS, whereas DCVCS was not detectable with human kidney microsomes (16). FMO3 mRNA levels in human kidney were also detected at much lower levels than in liver (19). Low levels of S-oxidase and possibly cytosolic β-lyase (see above) in the human kidney suggest that humans are poorly susceptible to DCVC-induced toxicity. Alternatively, bioactivation of DCVC may occur primarily extrarenally, and reactive metabolites may get translocated from tissues such as the liver to the kidney via the circulation.</p><p>Because currently there is no evidence for translocation of DCVC-derived reactive metabolites among tissues, we recently developed methods that would allow us to investigate the presence of DCVCS in the circulation after exposure of rats to DCVC or TCE. These methods were based upon the expected reactivity of DCVCS as a Michael acceptor with Hb. Several globin cysteine-containing peptides monoadducted and cross-linked by DCVCS were identified both in vitro in red blood cells (RBCs) and in vivo in rats after exposure to DCVCS (21). Because DCVCS readily formed three monoadducts and one cross-linked diadduct upon incubation with N-acetyl-L-cysteine (NAC) at physiological conditions but not with N-acetyl-L-lysine or L-valinamide (22), cysteine residues of Hb have been implicated as primary reaction sites. While the results suggested Hb adducts can serve as biomarkers for DCVCS in the circulation, there was no prior information on whether RBCs have the ability to bioactivate DCVC. Therefore, in the present study we characterized Hb adducts formed after incubation of rat erythrocytes with DCVC. We also investigated β-lyase activity in RBCs using the model cysteine S-conjugate β-lyase substrate, S-(2-benzothiazolyl)-L-cysteine (BTC).</p><!><p>DCVC and DCVCS are hazardous and should be handled with care. DCVC was shown to be a strong, direct-acting mutagen by the Ames test (23).</p><!><p>Trifluoroacetic acid (TFA), alpha-ketobutyric acid sodium salt monohydrate (KBA), aminooxyacetic acid (AOAA), 2-mercaptobenzothiazole (2-MBT), horseradish peroxidase (Type II, 224 purpurogallin units/mg solid), cytochrome P450 reductase, and human Hb were purchased from Sigma-Aldrich Research (St. Louis, MO). Sodium borate decahydrate was purchased from ICN Biomedicals (Aurora, OH). Acetone and 30 % H2O2 were purchased from Fisher Scientific (Pittsburgh, PA). Trypsin (reductively alkylated) was obtained from Promega (Madison, WI). SDS, 15 % Tris-HCl Criterion precast gels, DTT, glycine, Kaleidoscope Prestained Standards were purchased from Bio-Rad Laboratories (Hercules, CA). SilverSnap Stain Kit II was obtained from Pierce (Rockford, IL). Heparin was supplied by American Pharmaceutical Partners (Schaumburg, IL). DCVC, DCVCS, and BTC were synthesized and characterized as previously described (15, 18, 24, 25). Purity of all synthetic compounds was determined to be >95% by HPLC analysis.</p><!><p>Male Sprague-Dawley rats (180–220 g), purchased from Harlan (Madison, WI), were maintained on a 12 h light/dark cycle and given water and feed ad libitum. Heparinized whole blood was obtained through cardiac puncture and processed immediately.</p><!><p>After removal of the plasma fraction from the whole blood, the red blood cells were washed three times with an equal volume of saline and centrifuged at 1,855 × g for 5 min in between washes. Erythrocytes (approximately 3.52 × 106/220 μL) were resuspended in an equal volume of PBS (10 mM, pH 7.4) before incubation with and without DCVC (9, 90, 450 μM) in a shaking water bath for 2 h at 37°C. At the end of the incubation, RBCs were lysed with equal volume of cold doubly deionized H2O and globin was isolated using acidified acetone as described previously (26).</p><!><p>To ascertain that the DCVC concentrations used to investigate globin adduct formation were not associated with RBC toxicity, the extent of hemolysis was measured for erythrocytes incubated alone and with DCVC (90 or 450 μM) at 0–2 h using a spectrophotometric assay as previously described (21).</p><!><p>Globin samples from in vitro incubations (9–450 μM DCVC and control) were investigated for globin chain cross-link formation using SDS-PAGE as described previously (21). Briefly, 2.5 μg of globin dissolved in 2.5 μL doubly deionized H2O was added to treatment buffer (0.5 M Tris-HCl, 10% glycerol, 10% SDS, and 0.01 g/mL bromophenol blue, 10% doubly deionized H2O). Samples were incubated at room temperature for 30 min after addition of 200 mM DTT to reduce disulfide bonds before being loaded onto a 15 % Tris-HCl Criterion Precast Gel. Kaleidoscope Prestained Standards (1 μL diluted 50 fold with the treatment buffer) were used for molecular weight determination. Silver stained gels were analyzed for dimer formation using density measurements with Quantity One software (Bio-Rad Laboratories; Hercules, CA).</p><!><p>Trypsin digest of globin was obtained as described previously (21). Samples (control and 450 μM incubation) were desalted using C-18 solid-phase zip tips before being loaded onto a Zorbax (Agilent) C18 stable bond column (0.075 mm × 150 mm, 5μ, 300 Å) equipped with a Micromass Electrospray Hybrid Quadrupole Orthogonal Time-Of-Flight mass spectrometer (ESI-QTOF/MS) (21). Trypsin digests from 450 μM DCVC and control incubations were also subjected to a MALDI-TOF2 4800 mass spectrometer (Applied Biosystems; Foster City, CA) (21). The digest samples with and without 9 μM DCVC were subjected to an HPLC system carrying a Zorbax (Agilent) C18 stable bond column (0.075 mm × 100 mm, 3 μ, 300 Å) and analyzed by a Linear Trap Quadrupole (LTQ) Orbitrap XL mass spectrometer (Thermo Scientific; Waltham, MA) (21). The latter instrument was used due to suspension of ESI-QTOF/MS availability.</p><!><p>Peak lists of experimental monoisotopic peptide masses were imported into Microsoft Access for comparison with the theoretical masses of modified peptides generated from the virtual digest of rat globin chains (α1, α2, β1, and β2) using Protein Prospector (www.prospector.ucsf.edu). Same methodology and search criteria were applied as described previously (21). Calculations were based on the 35Cl monoisotopic masses for adducts containing one Cl atom (monoadducts). The isotopic pattern was then evaluated for the presence of 37Cl. We limited our analyses to cysteine-containing peptides (including one and two missed cleavages) since DCVCS preferentially reacts with sulfhydryl groups (21, 22). Theoretical lists of monoisotopic masses for modified peptides were calculated for the following S-oxidase-derived monoadducts as described previously (20): DCVCS (+194.9757 Da, addition of DCVCS moiety and loss of HCl), DCVCS-GSH (+466.0828 Da, addition of DCVCS-GSH conjugate and loss of 2 HCl), and NA-DCVCS (+236.9862 Da, addition of NA-DCVCS and loss of HCl). Data were also analyzed for masses that matched formation of dimers between cysteine-containing peptides with DCVCS as a cross-linker (+158.9990 Da; addition of DCVCS and loss of 2 HCl) or NA-DCVCS as a cross-linker (+201.0096 Da; addition of NA-DCVCS and loss of 2 HCl).</p><p>Trypsin digests were also analyzed for peptides and peptide dimers modified by β-lyase-derived metabolites as shown in Figure 1. Theoretical monoisotopic masses were calculated for peptides modified by β-lyase-derived sulfur-containing monoadducts (type 1; +91.9488 Da) and by monoadducts formed between GSH and the reactive thiol species (type 2; +363.0559 Da, addition of GSH conjugate and loss of HCl). 2-Chlorothionoacetyl chloride and chlorothioketene could also give rise to 2-chlorothiolacetic acid (ClCH2C=OSH), an oxygen-containing reactive intermediate, in the presence of water (27). Hydrolysis could occur before and/or after reaction with Hb resulting in substitution of sulfur for oxygen. Therefore, formation of peptides modified by the oxygen-containing fragments (type 3; +75.9716 Da) and by monoadducts formed between GSH and the oxygen-containing fragment (type 4; +347.0787 Da, addition of GSH conjugate and loss of HCl) were also investigated. Due to the expected reactivity of β-lyase reactive intermediates with nucleophilic residues besides cysteines, we extended our search to include up to 4 monoadducts (reactive intermediates and their GSH conjugates) on the same cysteine-containing peptides.</p><p>We also analyzed for the presence of cross-links formed by the sulfur-containing reactive intermediate (cross-link type 1; +55.9721 Da, addition of the sulfur-containing fragment and loss of 2 HCl), and cross-links formed due to the oxygen-containing reactive intermediate (cross-link type 2; +39.9949 Da, addition of the oxygen-containing fragment and loss of 2 HCl).</p><!><p>A time course enzymatic assay was performed to determine if a prototype cysteine S-conjugate β-lyase substrate, BTC, can be metabolized by the β-lyase activity in RBCs (24). Formation of its stable metabolite, 2-MBT, was monitored over 60 min by HPLC. Freshly isolated washed RBCs from rats (n=3) were lysed with equal volume of cold doubly deionized H2O and hemolysate was centrifuged twice (10,621 × g, 20 min at 4 °C) (28). The supernatant was removed and dialyzed with 10 mM PBS (8.4 mM Na2HPO4, 1.6 mM KH2PO4, 154 mM NaCl, pH 7.4) against Spectrapor 7 dialysis membrane with the molecular mass cut-off of 15,000 Da (Spectrum Laboratories; Rancho Dominguez, CA). RBC hemolysate (RBCh) was used in incubations after being stored overnight at 4 °C. Protein concentrations were determined according to the method of Lowry et al (29). RBCh (7–9 mg protein/ml) was incubated with 1 mM BTC in the presence and absence of 1 mM KBA, an α-ketoacid used to enhance β-lyase activity (30). Incubations were also performed in the presence and absence of AOAA (0.2 mM), a potent inhibitor of renal β-lyases (31). All chemicals were initially dissolved in 10 mM PBS pH 7.4 except for BTC which was dissolved as described by Dohn and Anders (25). Briefly, initial dissolution of BTC (10 mM) in doubly deionized H2O (40% of final volume) with addition of 10 μL of 1 M NaOH was followed by addition of sodium borate decahydrate buffer (0.2 M pH 8.6) (50% of final volume) and final volume was adjusted with doubly deionized H2O. Incubations were initiated by addition of BTC (30 μL) after a 5 min preincubation time in the shaking bath at 37°C. The final pH of the reaction mixture was 8.3. At each time point protein from reaction aliquots (60 μL) was precipitated with ice cold ethanol (90 μL) and the samples were subjected to centrifugation (20,817 × g, 10 min at 4 °C). The resulting supernatant was filtered through a 0.2 μM Acrodisc LC13 membrane filters (Pall; East Hills, NY) before HPLC analyses as described below.</p><!><p>In order to determine if S-oxidase activity could arise due to monooxygenase-like activity of Hb in the presence of H2O2 which could form in RBCs during oxidative stress (32), we set up several reactions to model these catalytic processes that may result in DCVC oxidation. Experimental designs similar to those used to demonstrate dibenzothiophene and aniline oxidations by Hb were followed (33, 34). Briefly, DCVC (1 mM) was incubated with H2O2 (1–5 mM) in 10 mM PBS at pH 7.4, 37°C in the presence and absence of human Hb (0.1–2 mM) and formation of DCVCS over 1 h was monitored by HPLC as described below. To further investigate DCVC oxidation over 1 h by Hb catalysis we incubated human MetHb (1 μM), P450 reductase (0.02 units, 1 unit equals 1 μmole of cytochrome c reduced per minute), DCVC (10 mM) with and without NADPH (0.2 mM) in KH2PO4 buffer (20 mM, pH 6.8) (34). The concentration of MetHb was determined by the method of Winterbourn et al (35). Reactions (200 μL) were stopped by the addition of ice cold acetone (600 μL), centrifuged as described above, and dried using a stream of N2 before being reconstituted in PBS (10 mM, pH 7.4) and analyzed by HPLC.</p><p>Alternatively, horseradish peroxidase (0.002 mM) was used as a model for catalytic activity of Hb in the presence of H2O2 (0.5 mM) and DCVC (1 mM) and DCVCS formation was monitored at 37 °C over 1 h. The buffer used with horseradish peroxidase experiments contained 0.1 M KH2PO4, 0.15 M KCl, and 2 mM EDTA adjusted to pH 7.4 (33).</p><!><p>HPLC analyses were performed using a Gilson gradient controlled HPLC system (Model 306 pumps) equipped with a Beckman Ultrasphere 5-μm analytical (4.6 mm × 25 cm) column. HPLC methods listed below were developed from the original methods described by Elfarra and Hwang for 2-MBT analyses and by Ripp et al. for DCVCS analyses (15, 24). UV detection was set at 321 nm (for 2-MBT analyses) and 220 nm (for DCVCS analyses).</p><p>Mobile phases for 2-MBT analyses were: pump A, doubly deionized H2O, pH adjusted to 2.5 with TFA and pump B, 50% ACN, pH adjusted to 2.5 with TFA. Initial 70% B was maintained for 5 min, then increased to 100% over 2 min and held for 4 min. The percent B was then decreased to 70 over 1 min and held for 3 min. Quantitation of 2-MBT was based on peak areas of the standard curve (r>0.99) generated using synthetic 2-MBT (limit of detection was 0.39 μM).</p><p>Mobile phases for DCVCS analyses were: pump A, 1% acetonitrile (ACN), 0.1% TFA and pump B, 20% ACN, 0.1% TFA. Initial 12.5% B was maintained for 5 min, then increased to 90% over 3 min and held for 4 min. The percent B was then decreased to 12.5 over 3 min and held for 3 min. Quantitation of DCVCS diastereomers I and II was based on peak areas of the standard curve (r>0.99) generated using synthesized DCVCS (limit of detection was 2.16 μM for each diastereomer).</p><!><p>Statistical analyses were executed using One Way ANOVA test followed by Student-Newman-Keuls method using SigmaStat (San Jose, CA). Significance was assigned if the values were <0.05.</p><!><p>In vitro incubations of rat RBCs with DCVC (90 and 450 μM) at physiological conditions (pH 7.4, 37 °C) did not cause hemolysis as determined by the hemolysis assay (data not shown). Erythrocytes remained intact after 2 h similar to control incubations without DCVC.</p><!><p>Formation of β-lyase-derived monoadducts with globin was investigated using trypsin digested peptides from 9 and 450 μM DCVC incubations that were subjected to LC/ESI/MS. In the event that reactive thiol species undergo hydrolysis and the sulfur is replaced by the oxygen before or after reaction with Hb (27), we have also analyzed globin peptides for the corresponding monoadduct type 3 (Figure 1). In addition, we analyzed for peptides modified by GSH conjugates of reactive sulfur and oxygen-containing intermediates (monoadducts types 2 and 4) because GSH is abundant (0.8–2 mM) in RBCs (32, 36). At 450 μM DCVC we detected three modified peptides on both α and β chains (Table 1). The majority of those peptides also displayed the characteristic 35Cl, 37Cl isotopic pattern. Peptide β(121-132) was modified by both monoadducts types 2 and 3. The overlapping sequence between two modified peptides (8-31 and 12-40 on α chains) suggests that four nucleophilic sites (Cys13, Lys16, His20, Arg31) may be involved in reactivity with monoadducts 1 or 2. When RBCs were incubated with 9 μM DCVC, no monoadducts were detected.</p><!><p>We also analyzed data for cross-links between two globin peptides with sulfur or oxygen-containing intermediates as cross-linkers (cross-links types 1 or 2, respectively) (Figure 1). Two unique peptide cross-links were formed between α chains, β chains, and α and β chains at 450 μM DCVC (Table 2). Peptide (1-16) (or the short peptide 8-16) is predominantly involved in cross-link type 1 or 2 formation between α chains suggesting that Lys11, Cys13, and Lys16 may be potential targets for these electrophiles. Short β chain peptides [(83-95) and (121-132)] and peptides containing missed cleavages, such as (77-104) and (121-144)/(121-146) were also detected in cross-link types 1 and 2. When globin from 9 μM DCVC incubation was analyzed, one cross-link was detected each between the two β chains and between α and β chains. The cross-link containing peptides (12-31) and (121-146) has an overlapping amino acid sequence with the cross-link detected at 450 μM DCVC suggesting that reactivity of the β-lyase- derived fragment is selective toward these peptides. Target nucleophilic amino acids involved in this cross-link formation may be Cys13 and Lys16 of the α chains and Cys125, Lys132, His143, Lys144, or His146 of the β chains.</p><!><p>Formation of DCVCS in RBCs incubated with DCVC was assessed by examining trypsin digested cysteine-containing peptides of globin from 450 μM DCVC incubation with RBCs. Digest subjected to LC/ESI/MS and MALDI/MS was analyzed for peptides modified by DCVCS, NA-DCVCS, and DCVCS-GSH monoadducts as described before (21). Since a diadduct of DCVCS with two moieties of NAC was characterized previously and globin cross-links were detected with DCVCS in vitro and in vivo (21, 22), peptide cross-link formation with DCVCS or NA-DCVCS as cross-linkers was also analyzed with DCVC incubation. However, no monoadducted or cross-linked peptides containing DCVCS-derived intermediates were detected with either instrument suggesting that S-oxidation of DCVC does not occur in RBCs.</p><!><p>To determine if there is any cross-link formation between Hb chains, we evaluated cross-link formation in globin from incubations of RBCs with DCVC (9–450 μM) using SDS-PAGE. Upon addition of 200 mM DTT to reduce disulfide bond formation, dimer bands persisted at all DCVC concentrations and not in the control (Figure 2) which suggested that treatment of RBCs with DCVC gave rise to cross-linking of globin chains that is not due to disulfide bonds.</p><!><p>In order to assess β-lyase activity with cysteine S-conjugates in RBCs, we used a prototype β-lyase substrate, BTC, that generates a stable thiol, 2-MBT, (24) eluting at 7.6 min on HPLC. After cell lysis and dialysis to remove extraneous protein, RBCh was incubated with and without 1 mM BTC and in the presence or absence of 1 mM KBA and 0.2 mM AOAA and formation of 2-MBT was monitored over 60 min. Control reaction with RBCh incubated alone did not give rise to any coeluting peaks. A time-dependent product formation was observed when RBCh was incubated with BTC with and without KBA (Figure 3) corresponding to the elution time of 2-MBT standard suggesting β-lyase presence in RBCs. Formation of 2-MBT in the presence of AOAA was significantly inhibited in reactions containing KBA at 30 and 60 min (p<0.05) (75 and 83%, respectively) as compared to the corresponding 2-MBT levels in reactions without AOAA providing additional evidence for the activity being mediated by a β-lyase present in RBCs.</p><!><p>We assessed the extent of DCVC oxidation over 1 h when human Hb (0.1–2 mM) or horseradish peroxidase (0.002 mM) was incubated with DCVC (1 mM) in the presence of H2O2 (1–5 mM) and (0.5 mM), respectively. We also incubated metHb (1 μM) and DCVC (10 mM) in the presence of P450 reductase (0.02 units) and NADPH (0.2 mM) to monitor formation of DCVCS over 1 h. Increase in area of DCVCS peaks eluting at 4.0 min (diastereomer I) and 4.3 min (diastereomer II) was not detected in any of these systems suggesting that DCVC is not a substrate for Hb or horseradish peroxidase catalysis. Collectively, these results suggest that oxidation of DCVC to DCVCS in RBCs is unlikely to occur under physiological conditions.</p><!><p>In the present study we have used Hb adducts to characterize DCVC bioactivation in intact erythrocytes. Our detection of Hb adducts with β-lyase-derived reactive intermediates after incubation of RBCs with DCVC suggested that DCVC is bioactivated in RBC via β-lyase. β-Lyase activity has been detected in a variety of mammalian tissues (testes, pancreas, spleen, heart, muscle, and brain) besides liver and kidney (37). The presence of aspartate aminotransferase has been previously characterized in erythrocytes (38). Bioactivation of DCVC by aspartate aminotransferases has been demonstrated in the liver, kidneys, and brain (7, 39, 40). In this study we report data showing that cysteine conjugates (BTC and DCVC) are substrates for β-lyase activity in RBCs. The finding that BTC β-lyase activity was inhibited by AOAA provided further support for BTC being a substrate for β-lyase in male Sprague-Dawley rat erythrocytes. Because β-lyases have different selectivities for cysteine conjugates (41) and BTC was shown to display a much lower activity than DCVC with mitochondrial aspartate aminotransferase (39), BTC may not fully portray total β-lyase activity in RBCs. Bioactivation of DCVC to DCVCS via S-oxidase pathway in RBCs was also studied by investigating the presence of Hb adducts modified by DCVCS and its related products (NA-DCVCS, DCVCS-GSH). Due to high reactivity of DCVCS with N-acetyl-L-cysteine, GSH, and Hb (18, 21, 22) we expected formation of monoadducts or cross-links between DCVCS and Hb if DCVCS was formed after RBCs were incubated with DCVC. Lack of detectable DCVCS-derived Hb adducts or cross-links by the various MS techniques used suggested lack of formation of DCVCS and no S-oxidase activity in erythrocytes. These results suggest that RBCs are capable of contributing to further DCVC metabolism via the β-lyase pathway, but not via the S-oxidase pathway. Therefore, DCVC biactivation to DCVCS may only occur outside of the circulation.</p><p>Formation of monoadducts (types 1–4) occurs due to nucleophilic addition reaction with nucleophilic residues of Hb resulting in a corresponding mass shift for the unmodified peptide. Although we limited our analyses to cysteine-containing peptides, presence of multiple monoadducts on the same peptides suggests that other nucleophilic residues, such as lysines, histidines, or arginines are also reactive toward sulfur/oxygen-containing reactive species generated by the β-lyase pathway. A β-lyase-dependent reactive metabolite of S-(1,2,2-trichlorovinyl)-L-cysteine (TCVC), dichlorothioketene, was shown to covalently modify renal and hepatic proteins in vivo and an Nε-(dichloroacetyl)-L-lysine adduct was detected after rat treatment with tetrachloroethene or TCVC (42, 43). In addition, upon treatment of rats with fluorinated ethylenes, thioanoacyl halide derivative of the resulting cysteine-S-conjugate was shown to react with lysine groups of mitochondrial proteins (44). These results suggest that both chlorothioketene and 2-chlorothionoacetyl chloride of DCVC may also react with ε-amino groups of proteins. Our study revealed that Hb adducts containing GSH were formed much less than globin monoadducts and cross-links at 450 μM DCVC. This could be due to the higher concentration of Hb (8–10 mM) in RBC in comparison with that of GSH (0.8–2 mM) (32, 36) making Hb more accessible to reactive electrophiles than GSH. Preference of β-lyase-derived intermediates for amino groups over sulfhydryl groups may have also contributed to their selectivity. Thus, β-lyase-derived metabolites of DCVC appear to be less selective in their reactivity with proteins than DCVCS which reacts only with cysteines (21, 22) suggesting that these species have a higher potential than DCVCS for reactivity with a wider variety of macromolecules.</p><p>A second nucleophilic addition reaction between another nucleophilic residue of Hb and a monoadducted β-lyase-derived intermediate followed by elimination of HCl results in formation of a cross-link as evidenced by SDS-PAGE and determined by a corresponding mass shift on LC/ESI/MS and MALDI/MS. Cross-linking between globin chains (due to β-lyase-derived intermediates) was prevalent in contrast to the lack of monoadducted peptides detected at low DCVC concentration (9 μM) consistent with our previous study where cross-links (due to DCVCS) but not monoadducts were detected at low DCVCS concentration (9 μM) (21). It is possible that some cross-linking may occur after the DTT treatment of globin during the trypsin digestion step. However, that would only concern sulfhydryl residues of Hb that remained unreacted in intact RBCs and formed disulfide bonds upon globin isolation procedure. Although reactivity of chlorothioketene with DNA base (cytosine) and formation of a monoadduct, N4-(chlorothioacetyl)cytosine, and an intramolecular cross-link, (3-N4-thioacetyl)cytosine, was previously characterized in vitro in organic solvents (45), to our knowledge cross-linking due to β-lyase-derived intermediates has not been previously demonstrated under physiological conditions.</p><p>In evaluating similarities between predominant α chain monoadducted and cross-linked peptides modified by β-lyase-derived reactive intermediates at 9 and 450 μM DCVC, potential target nucleophilic sites could be narrowed down to Cys13 and Lys16 because these residues overlap in detected peptides (1-16, 8-16, 8-31, 12-31, and 12-40) (Tables 1 and 2). Similarly, Cys125 and Lys132 of β chains may be involved because majority of modified peptides (121-132 and 121-146) at high DCVC concentration contain these residues. Peptides from the β chains involved in cross-link formation at 9 and 450 μM DCVC contained amino acid sequence 83–95 suggesting that His87, His92, Cys93, or Lys95 could be potential targets for electrophilic attack by β-lyase-derived sulfur/oxygen-containing intermediates at high and low DCVC concentrations. Overall, the relative amount of peptides modified by the sulfur-containing fragment (monoadduct types 1 and 3) was similar to peptides modified by the oxygen-containing fragment (monoadduct types 2 and 4) as indicated by β-lyase-derived monoadduct and cross-link data (Tables 1 and 2). Attempts to obtain adequate MS/MS fragmentation of modified peptides were not successful possibly due to complexity of the mixture, low abundance of these ions, and the presence of multiple signals representing different types of adducts/cross-links. However, the use of mass spectrometry with different ionization methods provided corroboration for the presence of multiple monoadducted and cross-linked peptides.</p><p>To confirm lack of detectable DCVCS-Hb adducts with lack of S-oxidase activity in RBCs, we investigated the presence of S-oxidase activity in RBCs by investigating monooxygenative-like properties of Hb. Environment inside erythrocytes is rich in oxygen and heme-containing Hb (8–10 mM) predisposing these cells to constant oxidative stress (36). Autooxidation of Hb (2–3% a day) to metHb followed by dismutation of superoxide is a major source of H2O2 generation in RBCs (32). In the presence of H2O2 certain substrates, such as dibenzothiophene and chlorpromazine, are oxidized by monooxygenative and peroxidative catalysis of Hb and peroxidases, respectively, in vitro (33, 46). In addition, because P450 can mimic monooxygenase activity of Hb in RBCs, a reconstituted system containing rat liver cytochrome P450 reductase, NADPH, and human Hb with aniline as a substrate gave rise to an oxidation product of aniline (47, 48). Failure to detect DCVCS formation with any of these experimental designs suggested that RBCs do not contribute to bioactivation of DCVC via S-oxidation.</p><p>In the present study, we detected multiple globin monoadducts and cross-links with reactive intermediates generated by the β-lyase pathway and characterized β-lyase activity in RBCs which suggests that RBCs represent an additional compartment for DCVC metabolism via the β-lyase pathway. Because we did not identify any DCVCS-Hb adducts and due to the lack of S-oxidase activity in RBCs, bioactivation of DCVC to DCVCS is not likely to occur in the circulation. Thus, any detection of DCVCS in the circulation would likely be due to its formation and translocation from other tissues, i.e. liver. Unlike DCVCS which reacts only with cysteine residues (21, 22), identification of β-lyase-derived monoadducts and cross-links with different nucleophilic residues of Hb in this study suggests that DCVC bioactivated via β-lyases could yield reactive sulfur/oxygen-containing species that react with a wide spectrum of nucleophilic residues on macromolecules forming monoadducts and cross-links.</p><!><p>Cysteine conjugate β-lyase dependent metabolism of S-(1,2-dichlorovinyl)-L-cysteine (DCVC) and proposed mechanism for formation of Hb monoadducts and cross-links with sulfur (oxygen)-containing reactive intermediates. Hydrolysis that could occur before and/or after reaction with Hb results in substitution of sulfur for oxygen (27) and formation of the corresponding Hb monoadducts (types 3 and 4) and cross-link (type 2) could then result where the sulfur atom is replaced by the oxygen. Hb is used here to describe general types of adducted globins without specification of which globin chain is involved. aMonoadducts (types 2 and 4) can also arise by initial conjugation with GSH followed by reaction of GSH conjugate with Hb.</p><p>SDS-PAGE of globin (2.5 μg loaded) after incubation of RBCs with and without DCVC (9–450 μM) for 2 h at pH 7.4 and 37 °C. The table below contains typical values for dimer band intensities using optical density measurements. Fold values refer to densities with DCVC over density without DCVC. Monomers display two bands (α chains- 15153 (α1), 15197 (α2) Da and β chains- 15848 (β1), 15861 (β2) Da). Standards and globin were run on the same gel.</p><p> DCVC [μM]Dimer Band Density (ODu/mm2)Relative Density00.521.0090.841.62901.112.134500.871.67</p><p>Formation of 2-MBT over time when RBC hemolysates (RBCh) were incubated with 1 mM BTC in the presence and absence of 1 mM KBA and 0.2 mM AOAA. Values are means ± SD from 3 experiments. aSignificantly higher (p<0.05) than the corresponding values in the presence of AOAA at 30 (with KBA) and at 60 min (with and without KBA).</p><p>LC/ESI/MS and MALDI/MS Results of β-lyase-derived Globin Peptide Monoadducts (types 1, 2, 3, and 4; see Figure 1 for structures) after Incubation of RBCs with 450 μM DCVC for 2 h at 37 °C.</p><p># refers to the number of monoadducts on a peptide</p><p>Peptides α1 (8-31), (12-40), and (93-127) can also be from α2.</p><p>Peptides β1 (83-95) and (121-132) can also be from β2.</p><p>Residues in bold represent possible nucleophilic sites where reaction may occur</p><p>Monoisotopic masses displayed characteristic 35Cl, 37Cl isotopic pattern</p><p>LC/ESI/MS and MALDI/MS Results of β-lyase-derived Globin Peptide Cross-links (types 1 and 2; see Figure 1 for structures) after Incubation of RBCs with DCVC for 2 h at 37 °C.</p><p>Peptides α1 (8-16), (12-31), and (100-127) can also be from α2</p><p>Peptides β1 (77-104), (83-95), (83-104), (121-132), (121-144), and (121-146) can also be from β2</p>

PubMed Author Manuscript

Determining Excited-State Structures and Photophysical Properties in Phenylphosphine Rhenium(I) Diimine Biscarbonyl Complexes Using Time-Resolved Infrared and X-ray Absorption Spectroscopies

We have explored the structural factors on the photophysical properties in two rhenium(I) diimine complexes in acetonitrile solution, cis,trans-[Re(dmb)(CO) 2 (PPh 2 Et) 2 ] + (Et(2,2)) and cis,trans-[Re(dmb)(CO) 2 (PPh 3 ) 2 ] + ((3,3)) (dmb = 4,4'-dimethyl-2,2'-bipyridine, Ph = phenyl, Et = ethyl) using the combination method of time-resolved infrared spectroscopy, time-resolved extended Xray absorption fine structure, and quantum chemical calculations. The difference between these complexes is the number of phenyl groups in the phosphine ligand, and this only indirectly affects the central Re(I). Despite this minor difference, the complexes exhibit large differences in emission wavelength and excited-state lifetime. Upon photoexcitation, the bond length of Re-P and angle of P-Re-P are significantly changed in both complexes, while the phenyl groups are largely rotated by ~20º only in (3,3). In contrast, there is little change in charge distribution on the phenyl groups when Re to dmb charge transfer occurs upon photoexcitation. We concluded that the instability from steric effects of phenyl groups and diimine leads to the smaller Stokes shift of the lowest excited triplet state (T 1 ) in (3,3). The large structural change between the ground and excited states causes the longer lifetime of T 1 in (3,3).

determining_excited-state_structures_and_photophysical_properties_in_phenylphosphine_rhenium(i)_diim

Introduction<!>Experimental<!>Experimental and calculated spectra and vibrational mode assignments<!>Temporal evolutions of TR-IR spectra<!>TR-EXAFS determination of atomic displacements adjacent to Re atom<!>NTOs and geometries in S 0 and T 1 10<!>Correlation between photophysical properties and excited state structures

<p>Rhenium(I) diimine biscarbonyl complexes that bear two phosphine ligands with various numbers of allyl groups or alkyl groups are widely known as efficient redox photosensitizers in photocatalytic systems for CO 2 reduction [1][2][3][4][5][6]. These complexes are good building blocks for linear-shaped [7,8] or ring-shaped [9][10][11][12] Re(I) multinuclear complexes. Such complexes have been used in various applications as photofunctional materials such as light harvesting materials [8,13], efficient photosensitizers [9,12], and multielectron storage molecules [10]. A useful characteristic of these complexes for these applications is that their photophysical and photochemical properties are controllable by varying the number and type of functional groups of the two phosphine ligands [4,6].</p><p>In general, the redox potentials, absorption, and emission wavelengths in these complexes are linearly dependent on the sum of the Tolman's  values, which are an index of the electron accepting ability of phosphine ligands [4,6]. There are deviations from these trends when the complexes bear different numbers of aryl groups [6]. In particular, the Stokes shifts, which are the difference between absorption and emission energies, significantly decrease and the excited state lifetimes significantly increase when the complexes bare three aryl groups. These characteristics in the complexes bare three aryl groups are favorable for redox photosensitizers.</p><p>In a previous report [6], these deviations were attributed to - interaction between the aryl groups of the two phosphine ligands and a diimine ligand. The - interaction is a generic concept that includes various interactions from attractive to repulsive forces between aromatic rings [14][15][16][17].</p><p>In the current study we examined the electronic and structural properties of two prototypical complexes in the excited-state using time-resolved infrared vibrational spectroscopy (TR-IR), time-resolved extended X-ray absorption fine structure (TR-EXAFS), and quantum chemical calculations.</p><p>TR-IR is a powerful tool for studying metal complexes in the excited state [18]. In particular, the IR spectra of double bonds such as C=C and C=N stretching vibrational modes in the fingerprint region are good probes for the electronic and structural character of aromatic groups in the excited state, for aromatic ligands in metal complexes [19][20][21][22][23], and organic compounds [24][25][26][27].</p><p>There are two major advantages to studying intramolecular interactions in molecules containing aromatic groups using TR-IR. One is that a geometry for the lowest excited triplet state (T 1 ) can be determined in combination with quantum chemical calculations [20][21][22][23]. This is because T 1 is the lowest triplet state so optimized geometries can be calculated without considering configuration interactions or having to use time-dependent density functional theory (TD-DFT). This was confirmed using prototypical [Ru(bpy) 3 ] 2+ and its ligand-and isotope-substituted complexes [20,23]. The other major advantage is that correlations between photophysical properties and structural dynamics can be derived from temporal evolutions of TR-IR spectra [19,21,[24][25][26][27]. The wavenumber and intensity of vibrational modes including double-bond vibrations are sensitive to the bond order and structure. Variations in bond order and structure can therefore be traced after photoexcitation in real-time. Using this method, we revealed that inter-and intramolecular interactions between aromatic groups play an important role for photophysical properties such as photoinduced phase transition materials [19,24,25], photoactive liquid crystals [26], and thermally-activated delayed fluorescence materials [27].</p><p>The lowest photoexcited state in Re(I) diimine complexes is the metal-to-ligand charge transfer (MLCT) state. Thus, the positions of atoms adjacent to the central Re(I) atom are key for understanding the structural dynamics upon photoexcitation. EXAFS is one of the best methods to determine the positions of atoms adjacent to metal atoms [28,29]. It has recently become possible to measure TR-EXAFS of a metal complex in solution in sub-nanosecond time resolution with high sensitivity comparable to that of TR-IR [30][31][32][33][34][35][36]. TR-EXAFS provides complementary information to TR-IR on the structure of metal complexes. This is because TR-IR spectra above 1000 cm -1 are sensitive to vibrations within ligands but not to vibrations between a central metal atom and ligands.</p><p>In this study, we focus on two Re(I) diimine biscarbonyl complexes that have very different photophysical properties despite their similar molecular structures. One is cis, trans- 1a) and the other is cis, trans-[Re(dmb)(CO) 2 (PPh 3 ) 2 ] + [PF 6 ] -((3,3), Figure 1b).</p><p>According to the previous report [6], the absorption wavelengths of Et(2,2) and (3,3) to 1 MLCT (metal-to-ligand charge-transfer state) are almost the same at 403 and 402 nm, respectively.</p><p>However, their emission wavelengths from 3 MLCT are very different at 622 and 600 nm, respectively. This indicates that the Stokes shift of (3,3), 8200 cm -1 , is significantly smaller than that of Et(2,2), 8700 cm -1 . The lifetime of 3 MLCT of (3,3) is 1280 ns, which is more than twice as long as that of Et(2,2) at 646 and 203 ns (dual emission). The higher energy (because of the small Stokes shift) and longer lifetime of 3 MLCT of (3,3) should make (3,3) a better redox photosensitizer than Et(2,2).</p><p>To explore the difference in photophysical properties between Et(2,2) and (3,3) in terms of excited-state character, we measured TR-IR spectra in the wavenumber range from 1000 to 2200 cm -1 and carried out spectral simulations using DFT calculations. To confirm the assignments of vibrational modes, we carried out measurements on deuterated (3,3) in which all hydrogen atoms of phenyl groups were replaced by deuterium ((3D, 3D), Figure 1c). We measured TR-EXAFS under almost the same conditions as the TR-IR measurements, and analyzed the positions of atoms adjacent to Re(I). We discuss in detail the electronic and structural character based on geometries confirmed by TR-IR and TR-EXAFS. We found little difference in the electronic properties but a large difference in the excited-state structures between Et(2,2) and (3,3), which 5 should cause the difference in photophysical properties.</p><!><p>2.1. Materials. All complexes were synthesized according to a previous report [6].</p><p>2.2. Instrumentation. Detailed procedures of steady-state IR and TR-IR measurements are reported elsewhere [18,23]. The steady-state IR spectra were measured by a Fourier-transform infrared (FT-IR) spectrometer (Shimadzu Prestage-21). KBr pellets for FT-IR measurements were prepared from 1.0 mg of sample and 150 mg of KBr powder. TR-IR spectra were acquired by a purpose-built system based on the pump-probe method, using a femtosecond Ti:sapphire chirped pulse amplifier (Spectra Physics Spitfire Ace, pulse duration = 120 fs, wavelength = 800 nm, repetition rate = 1 kHz). The pump pulse (wavelength = 400 nm, fluence at sample position = 7 mJ/cm 2 ) was generated using the second harmonic generation of part of the amplifier output. The tunable probe pulse (wavenumber range = 1000-4000 cm -1 , bandwidth = 150 cm -1 ) was obtained by an optical parametric amplifier equipped with a crystal for difference frequency generation (Lighconversion, TOPAS-Prime). The pump and probe pulses were irradiated on an infrared flow cell with BaF 2 windows (optical path length = 0.1 or 0.5 mm). The polarizations of the pump and probe pulses on the flow cell were set to the magic angle using half wave plates for visible and mid-infrared wavelengths to avoid the effect of rotational relaxation. The spectrum of each probe pulse after passing through the flow cell was recorded by a mercury-cadmium-tellurium (MCT) infrared linear array system (Infrared Systems Development FpAS-64166-D). TR-IR spectra were obtained by averaging the difference between those with and without the pump pulse using an optical chopper synchronized with the pump pulse. The wavenumber resolution was 1-3 cm -1 , which depends on wavenumber because of the linearity of polychromator. The accuracy of the wavenumber was calibrated by comparing to the spectrum of a polystyrene film to within 2 cm -1 . Typical samples for TR-IR measurements were 3 mM acetonitrile solutions, which were bubbled with nitrogen gas before and during each measurement. All measurements were carried out at room temperature. TR-EXAFS experiments with the pump-probe method were conducted at the AR-NW14A beamline in the Photon Factory Advanced Ring (PF-AR) [37]. The pump pulse (wavelength = 343 nm, fluence at sample position = 6 mJ/cm 2 ) was generated by the third harmonic generation of a femtosecond Yb fiber laser (Amplitude Tangerine). TR-EXAFS spectra were collected in fluorescence mode using a plastic scintillation detector. To measure the spectral change by the laser excitation precisely, the fluorescence X-ray signals before and after laser excitation were measured by gated integrators synchronized with the laser pulse (397 kHz). The sample (5 mM acetonitrile solution) was circulated using a diaphragm pump to reduce radiation damage by the laser and X-ray, and was shaped to a stable 450-m-thick jet using a stainless steel nozzle.</p><p>2.3. Quantum chemical calculations. All quantum chemical calculations were performed using the Gaussian16 software package [38]. The geometries of the ground state (S 0 ) and T 1 were optimized by DFT calculation using the mPW1PW91 functional and the LanL2DZ basis set.</p><p>Solvent effects of acetonitrile were considered using the conductor-like polarizable continuum model. The IR vibrational spectra of S 0 and T 1 were simulated on the basis of the optimized geometries, and difference spectra corresponding to experimental TR-IR spectra were obtained by subtracting the S 0 spectra from the T 1 spectra. To fit the calculated spectra to the experimental spectra, we considered the bandwidth = 10 cm -1 and a scaling factor = 0.97 for vibrational modes at wavenumbers less than 1650 cm -1 [20] and a scaling factor = 1.02 for CO stretching vibrational modes ranging from 1750 to 2050 cm -1 [39]. To discuss changes in charge distribution upon electronic transition by photoexcitation, we performed TD-DFT calculations and obtained natural transition orbitals (NTOs) [40].</p><!><p>Figures 2a and 2c show the FT-IR spectra and Figures 2b and 2d show the calculated spectra of Et(2,2) and (3,3), respectively. The intensities less than 1650 cm -1 are shown enlarged for clarity. The FT-IR and calculated spectra are in good agreement with each other for both complexes, so we assigned the observed vibrational bands to the normal vibrational modes obtained by the calculations. The colors indicate the ligand at which the vibrations of each vibrational mode are mainly localized, judging from the view of the vibrational mode obtained by the calculations (Figure S1-S4). Peaks in blue, green, red, and gray are mainly localized at CO, dmb, Ph, and Et, respectively. Roughly speaking, the bands at 1850-1950 cm -1 are assigned to CO stretching vibrational modes while those at <1650 cm -1 are assigned to dmb and Ph.</p><p>Figures 3a and 3c show the TR-IR spectra at 100 ps of Et(2,2) and (3,3), respectively, and Figures 3b and 3d show their corresponding calculated spectra. These spectra show the difference in absorbance (abs) spectra between the ground and excited states. Upward and downward bands are assigned to vibrational transitions in the excited (transient absorption band) and ground states (bleach band), respectively. The TR-IR and calculated spectra are also in good agreement with each other for both complexes, so it is reasonable to discuss the character of T 1 based on these calculations. In the same manner as above, we assigned the bands to normal vibrational modes and the colors indicate the same groups as those for the FT-IR results. There are two distinctive features in these spectra. One is that the wavenumbers of bands assigned to CO (indicated by blue) are increased by 40-70 cm -1 compared with the corresponding bleach bands. These blue shifts are well known and indicate that the charge at Re is reduced by the charge transfer from Re to dmb upon photoexcitation, and the C-O bands are strengthened by a decrease in  backdonation from Re [41][42][43][44]. The other distinctive feature is that almost all bands at wavenumber <1650 cm -1 (indicated by green) except that at 1100 cm -1 (indicated by red) are assigned to vibrational modes of dmb, which is largely different from the ground state spectra. This feature strongly indicates that charge variations by the charge transfer occur mainly on dmb. This speculation is supported by the NTO analysis from DFT calculations, which is described later.</p><p>To confirm our vibrational mode assignments, we measured the TR-IR spectra and carried out spectral simulations of (3D,3D). The measured and calculated results are shown in Figure 4a and 4b, respectively. The normal mode vibrations assigned to Ph are strongly localized in Ph (Figure S4), so the deuteration of Ph only affects the vibrations in Ph. The overall features of both the measured and calculated spectra are almost the same as those of (3,3) (Figure S5), indicating that there are few bands assigned to vibrational modes localized in Ph. However, carefully comparing the spectra of (3,3) in Figure 3c and (3D,3D) in Figure 4a reveals that the shapes of the bands assigned to Ph (indicated in red at approximately 1100 cm -1 in Figure 3c) are different.</p><p>To show this difference more clearly, the TR-IR spectra of (3,3) and (3D,3D) are compared in the range from 1085 to 1130 cm -1 in Figure 4c. Although there is little difference in spectral shape at 1100-1120 cm -1 , the transient absorption and bleach bands at 1085-1100 cm -1 are only observed in (3,3). This indicates that these bands are assigned to vibrations localized in Ph. This result strongly suggests that the bands at 1085-1100 cm -1 are a good probe for changes in the character of Ph upon photoexcitation.</p><!><p>Temporal evolutions of TR-IR are useful for understanding the electronic and structural character of ligands in photoexcited states. We measured temporal evolutions of the TR-IR spectra of Et(2,2) and (3,3) in three wavenumber regions where typical bands assigned to each ligand are located: 1825-2050 cm -1 for CO (Figure 5), 1180-1260 cm -1 for dmb (Figure 6), and 1075-1120 cm -1 for Ph (Figure 7). In each figure, (a) and (b) show data of Et(2, 2) and (3, 3), respectively.</p><p>The effects of rotational relaxation on the spectra were canceled by the magic angle configuration.</p><p>In Figures 5a and 5b, the two bleach bands show little spectral change up to 100 ps, whereas the transient absorption bands show slight blueshift and bandwidth narrowing within 10 ps. These trends in spectral evolution are almost the same as those in Figures 6a and 6b. On the basis of the spectral simulations obtained by the quantum chemical calculations, the bands in Figures 5 and 6 are assigned to the vibrations purely localized on CO and dmb, respectively. Except for the slight blueshift and narrowing, which result from vibrational relaxation in the excited states [41], the large wavenumber shifts of these transient absorption bands from those in the ground state occur 8 less than 1 ps after photoexcitation. This indicates the charge transfer from Re to dmb occurs immediately after photoexcitation. No change in bleach bands up to 100 ps indicates that there is no relaxation process to the ground state up to100 ps.</p><p>In contrast, the temporal evolutions of bands in Figure 7 are very different from those in Figures 5 and 6. On the basis of the calculations, vibrational bands in this region largely contain vibrations of Ph. Ph therefore has different dynamics from the other ligands in the early process after photoexcitation. The band at 1100 cm -1 in Figure 7a, which consists of normal vibrational modes strongly localized on Ph, increases gradually from 1 to 100 ps though it starts at a negative value because of overlapping with a bleach band. The band at 1100 cm -1 , which consists of vibrational modes of both dmb and Ph, also increases gradually from 1 to 30 ps after the sudden increase at 1 ps. In Figure 7b, the temporal behaviors of the bands are almost the same as those in Figure 7a, except that the overlapping between the transient absorption band and bleach band is much larger in the range from 1087 to 1097 cm -1 .</p><p>To show the different dynamics of each ligand clearly, we plot the intensities of typical bands at 1971-1976 cm -1 for CO (blue squares), at 1212-1214 and 1214-1217 cm -1 for dmb (green squares), and at 1098 and 1093 cm -1 for Ph (red squares), as a function of delay time in Figures 8a and 8b for Et(2,2) and (3,3), respectively. The intensities of the bands for CO and dmb were estimated by fitting to Gaussian functions to cancel the effects of the blueshift and narrowing. For the intensities of the bands for Ph, the Abs at 1198 cm -1 for Et (2,2) and the Abs at 1193 cm -1 for (3,3) were simply plotted because it is difficult to extract one band from the experimental spectra consisting of several transient absorption and bleach bands. For all three plots, the intensities change greatly less than 1 ps after photoexcitation followed by a gradual increase over 10 ps for CO and dmb and over 50 ps for Ph. The time constants of these gradual increases were estimated using a single exponential function. The estimated time constants for Et(2,2) in Figure 8a are 1.2 ± 0.1 ps for CO, 2.0 ± 0.1 ps for dmb, and 30.8 ± 1.4 ps for Ph, and those for (3,3) in Figure 8b are 1.2 ± 0.1 ps for CO, 3.7 ± 0.3 ps for dmb, and 25.2 ± 0.7 ps for Ph. These results strongly indicate that phenyl groups in the phosphine ligands in both complexes do not undergo structural changes immediately after charge transfer from Re to dmb by photoexcitation. Rather, it takes more than 20 ps for their structural changes to occur after charge transfer.</p><!><p>To determine directly the positions of atoms adjacent to the central Re(I) atom, we measured TR-EXAFS of (3,3) at the Re-LIII absorption edge. Details of the analysis can be found in the supporting information. For S 0 and T 1 , EXAFS spectra weighted by k 3 (k: wavenumber) and their Fourier-transformed (FT) spectra are shown in Figure 9a-d. The EXAFS spectrum of T 1 was measured at 100 ps after photoexcitation in which the movement of the ligands subsided. In the 9 FT spectra of both S 0 (Figure 9b) and T 1 (Figure 9d), a dominant contribution at ∼1.0-2.3 Å is attributed to the first nearest neighbor (NN) of the Re-C in CO, the Re-N in dmb, and the Re-P in PPh 3 . The feature appearing at ∼2.3-3.2 Å results from contributions from the second NN of the Re-C in dmb and the Re-O in CO, and the multiple scatterings in dmb and CO. To investigate the local molecular structures around the Re atom in more detail, curve-fitting analysis was performed.</p><p>The R range employed in the curve-fitting analysis was Δ R ≈ 1-3.2 Å. Fitting results are shown in Figure 9 as dotted lines. The bond lengths of the first NN obtained from the EXAFS analysis are listed in Table 1 together with those obtained by the DFT calculations from the TR-IR analysis.</p><p>These values are in good agreement with each other. Also, the values for S 0 are very close to those in a single crystal obtained by X-ray diffraction analysis [6]. These results strongly indicate that the optimized geometries by DFT calculations are highly reliable.</p><p>The differences in bond length between S 0 and T 1 are qualitatively explained by the fact that T 1 is the MLCT state. Charge transfer from Re to dmb decreases the charge on Re and increases that on dmb. The bonds between Re and CO and between Re and dmb have d- interaction as well as coordination bond whereas that between Re and a phosphine ligand has only coordination bond. A decrease in charge on Re decreases the d- interaction in Re-C bond between Re and CO (a decrease in -back donation) and increases the bond length between Re and C. The bond between Re and dmb undergoes an increase in charge on dmb in addition to a decrease in charge on Re; thus, the d- interaction somewhat increases and the bond lengths between Re and N increase. In contrast, the bond length between Re and P increase by a decrease in charge on Re because the bond between Re and a phosphine ligand has no d- interaction.</p><!><p>The calculated difference spectra between T 1 and S 0 are in good agreement with the TR-IR spectra and also the optimized geometries are confirmed by the TR-EXAFS spectra. The optimized geometries in S 0 and T 1 are therefore sufficiently reliable for further discussion on molecular orbitals and molecular structures. Figure 10 shows NTOs in S 0 and T 1 for Et(2, 2).</p><p>NTOs were calculated by TD-DFT based on the optimized geometries, and they express the variation in charge distribution upon transition [40]. For both S 0 and T 1 geometries, the highest occupied NTO (HONTO) is distributed mainly at Re and slightly at CO and dmb, whereas the lowest unoccupied NTO (LUNTO) is distributed mainly at dmb and slightly at CO but little at Re. This variation in orbital distribution indicates that charge at Re transfers to dmb, which agrees well with T 1 being assigned to the 3 MLCT state. In contrast to CO and dmb, NTOs are little distributed at Ph and there is almost no change in orbital distribution between the HONTO and LUNTO, indicating that phenyl groups do not undergo changes in charge distribution. These results explain why the structural change of phenyl groups occurs more slowly than charge transfer from Re to dmb after photoexcitation (described above). The phenyl groups are electronically isolated through the C-P-Re bond, so they are relocated by weak interactions with dmb and/or CO, which can be regarded as - interactions. This process should take several tens picoseconds using the analogy for the reorientation of surrounding solvent molecules. This mechanism also applies in (3,3) because the HONTO and LUNTO are almost the same as those of Et(2, 2), as shown in Figure 11.</p><p>We now compare the optimized geometries between S 0 and T 1 . For Et(2, 2), there are two geometries that have slightly different energies: one geometry has both Et groups on the same side (cis-Et(2, 2) in Figure 12) and the other geometry has one Et on either side (trans-Et(2, 2) in Figure 13). In these figures, the lower and upper geometries are S 0 and T 1 , respectively. These geometries correspond to two of the conformers in the previous report [6], and they can be verified by comparing their calculated spectra with TR-IR spectra. The calculated difference spectra of these two geometries are almost the same, as shown in Figure S6. Thus, the predominant geometry in solution cannot readily be determined from the TR-IR spectra. The energy difference between these geometries is very small: 0.0069 eV in S 0 and 0.0207 eV in T 1 , but the potential barrier between the two geometries should be much larger considering the reaction path. According to the previous report [6], the barrier heights are around 10 kJ/mol, which is much higher than the energy at room temperature (~2.5 kJ/mol). Thus, these two geometries are not likely to exchange at room temperature.</p><p>Figure 14 shows the optimized geometries in S 0 (lower) and T 1 (upper) for (3,3). Tables S1 and S2 summarize structural parameters obtained from the calculations for Et(2, 2) and (3, 3), respectively. When the geometrical changes from S 0 to T 1 are considered, the common features in structural change between Et(2, 2) and (3,3) are the elongation of Re-P bonds (0.05 Å in Et(2, 2) and 0.05 Å in (3, 3)) and the reduction of P-Re-P angle (-4.6º in Et(2, 2) and -8.1º in (3,3)). These structural changes should occur immediately after photoexcitation because P is coordinated directly to Re and Re is oxidized by the charge transfer from Re to dmb upon photoexcitation. Similar structural changes are reportedly observed in the dimer consisting of the same Re complex units [21]. The structures of CO and dmb are little changed by the charge transfer because the double or triple bonds in these ligands are stronger than P-Re. The electronic states of these ligands are largely changed by charge transfer because large spectral changes are observed immediately after photoexcitation.</p><p>In addition to these common features in structural change between Et(2, 2) and (3, 3), the rotations of phenyl groups on the C-P axis are different between Et(2, 2) and (3,3). As shown in Tables S1 and S2, the phenyl groups are rotated by 1º or 2º in Et(2,2) and by 16-27º in (3,3), with respect to the ground state angles. The intensities of vibrational bands assigned to Ph increase more slowly than the other bands, so these rotations occur much later after the charge transfer upon photoexcitation. There is little change in charge distribution in the phenyl groups upon photoexcitation, indicating that there is little change in electronic interaction of the phenyl groups with other ligands. Thus, these structural changes occur not by direct change in electronic states by the charge transfer but by indirect change induced by the elongation of Re-P bonds.</p><!><p>According to the report [6], there are two characteristics of photophysical properties in (3,3) compared with (2,2) and other complexes:</p><p>(1) The Stokes shift of (3,3) is much smaller than that of Et(2,2),</p><p>(2) The lifetime of T 1 in (3,3) is much longer than that in Et(2,2).</p><p>Our experiments and calculations revealed the following characteristics in terms of electronic state and molecular geometry:</p><p>(i) Charge transfer occurs only in the plane consisting of Re, CO, and dmb, and has no direct effect on Ph, (ii) Only Ph undergoes structural change slowly (25-30 ps) after photoexcitation.</p><p>Therefore, phenyl groups in the phosphine ligands show different behavior upon photoexcitation, and this behavior is expected to originate from indirect interactions.</p><p>The next question is how these experimental observations explain the difference in photophysical properties. Regarding characteristic (1), the quantum chemical calculations of T 1 , which are well supported by TR-IR and TR-EXAFS, effectively reproduce this characteristic as shown in the energy diagram in Figure 15. This means that the calculations include the origin of this characteristic. Close inspection of the distances among the functional groups in the filling models obtained from the calculations (Figure 16) shows that the phenyl group above dmb in Et(2,2) is more parallel to the dmb plane than that in (3,3). This is probably because the phenyl groups in (3,3) are so crowded that they are rotated to avoid steric repulsions. This rotation reduces the attractive - interaction between the phenyl group and dmb in (3,3) and increases the repulsive force among the phenyl groups and dmb. This is why T 1 in (3,3) is higher in energy than that in Et(2,2). Figure 17 shows schematic potentials depicting this situation. The crowded phenyl groups in (3,3) create the shallow potential because of steric repulsion, which makes the phenyl groups readily rotate. The deeper potentials because of the non-crowded phenyl groups in Et (2,2) restrict the rotation of phenyl groups. This difference in potential also explains why the Stokes shift of (3,3) is smaller than that of Et(2,2).</p><p>Regarding characteristic (2) requires considering how the lifetime of T 1 is determined in these complexes. In general, a lowest triplet excited state has a very long lifetime of more than a millisecond because of spin forbidden transition. The main relaxation paths to the ground state in these types of metal complexes are direct relaxation from T 1 to S 0 potentials unless another excited state exists that is close in energy to T 1 . This type of non-radiative relaxation was studied semiquantitatively by Engleman and Jortner [45]. They derived a simple law named the "energy gap law": The relaxation rate is determined by the overlap between wavefunctions of the excited and ground state potentials at the same energy level (Franck-Condon factor). Roughly speaking, a larger energy gap between two potentials gives a smaller Franck-Condon factor and a smaller relaxation rate if the equilibrium positions of the two potentials are close (Figure 18a). A larger difference in equilibrium position gives a larger Franck-Condon factor and a larger relaxation rate if the energy gap is the same (Figure 18b). However, these simple predictions are not reasonable for a molecule in which there is a large difference in molecular structure between the ground and excited states such as in our complexes. Particularly the assumption described in Ref. [45]: "We assume that the normal modes and their frequencies are the same in the two electronic states except for displacements in the origins of the normal coordinates.", is not valid for a complex that undergoes a large structural change upon photoexcitation such as (3,3). This deviation from the simple model is known as the Duschinsky effect [46,47]. Nevertheless, quantitative estimation of this effect is difficult for a complicated system such as in our complexes, so here we consider it qualitatively. If the normal coordinate of T 1 is largely different from that of S 0 , the Franck-Condon factor becomes small, as shown schematically in Figure 18c. The energy gap law considering this effect can explain characteristic (2). While the structural change between T 1 and S 0 in (3,3) is large, that in (2,2) is small. Thus, the total Franck-Condon factor in (3,3) is smaller than that in (2,2), so the lifetime of T 1 in (3,3) becomes longer than that in (2,2). This situation is the same as that for the dimer consisting of the same Re complex units we reported previously [21]. The natural transition orbital analysis indicates that photoexcitation induces charge transfer from Re to the diimine ligands but no change in charge at the phenyl groups. From the optimized geometries of S 0 and T 1 , the bond length of Re-P and the angle of P-Re-P are changed to a similar degree in both complexes, while the phenyl groups are largely rotated by 16-21º only in (3,3).</p><p>These results indicate that that displacements of phenyl groups occur indirectly via steric interaction among the aromatic groups, phenyl groups, and diimine. We therefore concluded that the triplet excited state in (3,3) being located higher in energy than that in Et(2,2) originates from the steric instability; this is because of the congestion of the phenyl groups in (3,3). The longer lifetime of (3,3) is attributed to the reduced non-radiative relaxation from the smaller overlap of wavefunctions between S 0 and T 1 because of the large displacements of phenyl groups, which is a special case of the energy gap law. This correlation between photophysical properties and structural dynamics indicates that photophysical properties in metal complexes can be controlled by carefully placing ligands while considering their steric interactions. In addition, our study shows that the combination method of TR-IR and TR-EXAFS makes it possible for us to explore photophysical and photochemical properties in metal complexes in terms of excited-state structure.</p>

ChemRxiv

Pd(II)-Catalyzed C(alkenyl)-H Activation Facilitated by a Transient Directing Group

Palladium(II)-catalyzed C(alkenyl)-H alkenylation enabled by a transient directing group (TDG) strategy is described. The dual catalytic process takes advantage of reversible condensation between an alkenyl aldehyde substrate and an amino acid TDG to facilitate coordination of the metal catalyst and subsequent C(alkenyl)-H activation by a tailored carboxylate base. The resulting palladacycle then engages an acceptor alkene, furnishing a 1,3-diene with high regio-and E/Z-selectivity. The reaction enables the synthesis of enantioenriched atropoisomeric 2aryl-substituted 1,3-dienes, which have seldom been examined in previous literature. Catalytically relevant alkenyl palladacycles were synthesized and characterized by X-ray crystallography, and the energy profiles of the C(alkenyl)-H activation step and the stereoinduction model were elucidated by density functional theory (DFT) calculations.

pd(ii)-catalyzed_c(alkenyl)-h_activation_facilitated_by_a_transient_directing_group

<!>Scheme 1. Synopsis of prior work and current study<!>Figure 3. Synthesis of alkenyl palladacycle complexes.<!>Scheme 2. Proposed catalytic cycle

<p>Alkenes react in a myriad of organometallic processes, 1 including nucleometallation, 2 migratory insertion, 3 C-H activation, 4 and isomerization. 5 Controlling selectivity among these pathways is critical for developing synthetically useful alkene functionalization methods. During the past few years, substrate-directed alkene functionalization has emerged as an enabling approach, in which selectivity control arises from coordination of the metal catalyst to Lewis basic sites on the substrate and subsequent formation of metalacyclic intermediates. 6 We recently described methods for enantioselective hydroarylation and 1,2-arylfluorination of alkenyl aldehydes using a transient directing group (TDG) [7][8] strategy. In these systems an amino acid or amino amide co-catalyst reversibly condenses 9 with the alkenyl aldehyde substate to generate an imine intermediate that is capable of coordinating to the palladium catalyst and directing arylpalladium(II) migratory insertion and downstream elementary steps. This TDG approach overcomes limitations associated with auxiliarybased methods, 10 which are widely used but add steps for auxiliary installation and cleavage. Based on these precedents, we questioned whether it would be possible to perturb this TDG-mediated alkene addition process such that C(alkenyl)-H activation 1b,4a,4b,11 would occur in preference to migratory insertion (or nucleometallation) from a common π-alkenepalladium(II) complex. This would generate an exo alkenyl palladacycle capable of engaging in catalytic coupling with a potentially wide arsenal of reaction partners, thereby complementing other advances in TDG-based C-H functionalization, 12 which have largely focused on C(alkyl)-H, 13 C(benzyl)-H 14 , and C(aryl)-H, 15 substrates (Scheme 1A). 16,17 Herein, we describe the development of a TDG approach to C(alkenyl)-H alkenylation in which two C(alkenyl)-H bonds are oxidatively cross-coupled to generate 1,3-diene products.</p><!><p>To reduce this idea to practice, we carried out optimization with alkenyl aldehyde substrate S1 and tert-butyl acrylate as the acceptor alkene. Using a previously published method as the starting point, 11d we quickly recognized that the most important aspect of reaction optimization was identifying the optimal combination of TDG and carboxylate base to promoted C(alkenyl)-H activation via a concerted metalation/deprotonation (CMD) process. 18,19 Screening of carboxylic acid additives as CMD bases (see Supporting Information) revealed that fluorine-containing benzoic acids are particularly effective in promoting the transformation. 14b In particular, A10 emerged as the optimal additive, presumably due to having three electron-withdrawing groups without steric bulk at the 2 or 6 positions (vide infra). Having identified a suitable carboxylic acid additive, screening of αamino acid TDGs 7,14a showed the importance of the steric properties of the α-substituent. TDGs containing a branched α-substituent were more effective (Table 1), among which tert-leucine (TDG6) was the highest-yielding. In comparison α,α-disubstituted amino acid TDG7 was much less effective, With an efficient method in hand, we evaluated the substrate scope (Table 2). First, different substituents on the aromatic ring of the benzaldehyde moiety were examined. Electrondonating and electron-withdrawing groups were tolerated, providing good to high yields (3-8). In general electron-poor alkenyl benzaldehyde substrates were lower-yielding, as exemplified by the para-CF3 in example 2. Then, we tested substrates containing different substituents attached to the alkene (9-14). Beyond stilbene derivatives, Z-configured alkyl-substituted alkenes were also compatible. Whereas methyl-substituted substrate 9 gave only 32% yield, branched alkyl groups were generally high-yielding (10-13). Sterically hindered tri-substituted alkenes, which are a challenging class of substrates in C(alkenyl)-H activation, 1b,4a,4b were competent substrates in the case of benzylidenecyclobutane (14) and benzylidenepiperidine (15), furnishing highly substituted 1,3-diene products that would be otherwise difficult to prepare.</p><p>In terms of the coupling partner scope, in addition to tertbutyl acrylate, other conjugated alkenes including N,Ndimethylacrylamide ( 16), acrylonitrile (17), and phenyl vinyl sulfonate (18) were effective. Moreover, non-conjugated alkenes were also viable coupling partners (19-21), though yields in these cases were lower. This reaction system is special compared to many previous Pd-catalyzed C-H alkenylation reactions because it successfully incorporates unreactive 1-hexene (21). 20,21 The method was demonstrated in a gram-scale reaction to prepare 1, which proceeded in 74% yield.</p><p>In preliminary experiments, we have found that this TDGmediated C(alkenyl)-H activation method can be extended to a non-aromatic aldehyde substrate, namely (Z)-5-phenylpent-4-enal, albeit in low yield (Eq. 1). Compared to the standard conditions in Table 2, screening a small panel of conditions revealed that 1,3-diene was produced by using PivOH as CMD promoter and conformationally constrained cyclopropanebased TDG8 as the TDG. Improved performance was achieved with a higher loading of the Pd catalyst and the TDG. To demonstrate the utility of this C(alkenyl)-H alkenylation method, a representative product (1) was converted into a variety of useful derivatives (Figure 1). We were able to selectively reduce or oxidize the aldehyde moiety to prepare (1) benzyl alcohol (23) and benzoic acid (24), respectively. Reductive decarboxylation could then be carried out from the benzoic acid to yield 25, 22 allowing the aldehyde to function as a traceless directing group. Straightforward deprotection of the t-Bu ester with TFA yielded free dienyl acid 26. Alternatively, both the aldehyde and the diene moieties can be simultaneously engaged in various annulation reactions. 3,4-Dihydroisoquinoline nitrone analogue 27 was prepared by treatment of 1 with hydroxylamine, which triggered condensation followed by aza-Michael addition. 23 Tetralone analogue 28 was obtained by Rh-catalyzed C(formyl)-H activation. 2, rotation about the C(aryl)-C(dienyl) bond was found to be restricted at ambient temperature. We thus questioned whether using an enantioenriched TDG could be used to develop an atroposelective version of this transformation (Figure 2). [25][26][27][28][29][30][31][32][33][34] In comparison to axially chiral styrenes, synthesis of atropoisomeric 1,3-dienes are less explored owing to synthetic difficulties and facile product racemization (Figure 2A). [35][36][37][38] In one study, a C2-symmetric cyclic 1,3-diene with large alkenyl substituents possessing a chiral axis along the C(alkenyl)-C(alkenyl) bond was synthesized, and the two atropoisomers were separated through chiral resolution. 38 Recently Shi reported the synthesis of enantioenriched 1,3dienes containing a chiral C(aryl)-C(dienyl) axis via thioetherdirected Pd(II)-catalyzed C(alkenyl)-H activation to form an endo-palladacycle intermediate with a spirocyclic phosphoric acid as the chiral ligand. 39 Given that our approach proceeds via an exo-palladacycle, we imagined that it could offer access to a complementary collection of axially chiral 1,3-diene products. Thus, we optimized reactions conditions with respect to yield and ee for 3-substituted-2-alkenyl benzaldehyde substrates. We found that using L-tert-leucine (L-TDG6) as TDG and carrying out the reaction at room temperature for four days afforded 1,3-dienes (3, 4, 6, and 7) with good yield and excellent atroposelectivity (Figure 2B). Single X-ray crystallography was used to establish the absolute stereochemistry of the major isomer of 4 as Ra. The high selectivity and critical role of the TDG and carboxylic acid promoter prompted us to example the reaction mechanism through experimentation and theory. First, control experiments showed that these reaction conditions developed for C(alkenyl)-H activation were ineffective for analogous substrates bearing similarly positioned C(alkyl)-H and C(aryl)-H bonds (see SI), demonstrating the unique aspects of C(alkenyl)-H activation in terms of electronic properties and transition state geometry. Next, by combining substrate S1, Pd(OAc)2, and a TDG in pyridine, 40,41 we were able to prepare two alkenyl palladacycles, 29 and 30 (Figure 3). These complexes were obtained as a mixture of E/Zstereoisomers, 5 with the product ratio influenced by the nature of the TDG. Complex 29 was further characterized by single-crystal X-ray diffraction confirming the Zstereochemistry of the major isomer (aryl and Pd cis to each other). Notably, the complex 29 is monomeric in contrast to the dimeric exo-alkenyl palladacycle complex obtained previously in our investigations of 8-aminoquinoline-amidedirected C(alkenyl)-H activation. 11d</p><!><p>A plausible catalytic cycle (Scheme 2) is proposed based on the experimental mechanistic studies and prior work. 7,8,11d Following coordination with the condensed imine, a p-alkene complex is formed, and the carboxylate-assisted C−H metalation occurs via a concerted metalation-deprotonation (CMD) mechanism to generate a six-membered palladacycle. Ligand exchange with the alkene is followed by migratory insertion to form an eight-membered palladacycle. The diene product is then formed via b-hydride elimination and directing group dissociation. Oxidation of the Pd(0) species and coordination of another condensed imine substrate regenerate the catalyst/reactant complex.</p><!><p>We performed density functional theory (DFT) calculations to investigate the proposed mechanism and the origin of the atroposelectivity (Figure 4). 42 Reaction free energy profile of the C−H alkenylation of alkenyl aldehyde S1 with tert-butyl acrylate using L-tert-leucine as the TDG and 3,4,5trifluorobenzoic acid (A10) additive was computed at the M06/6-311+G(d,p)-SDD(Pd)/ SMD(MeCN)//M06/6-31G(d)-SDD(Pd) level of theory (Figure 4A). 43 From the most stable isomer of the N-,O-coordinated π-alkene complex IM1, the carboxylate-assisted alkenyl C−H metalation occurs via the CMD mechanism via transition state TS1. 44 The resulting sixmembered palladacycle IM3 undergoes ligand exchange to replace the coordinated benzoic acid with tert-butyl acrylate to form more stable intermediates IM4a and IM4b, where two opposite π-faces of the tert-butyl acrylate bind to the Pd. Alkene migratory insertion from IM4a and IM4b (via TS3a and TS3b) leads to eight-membered palladacycles IM5a and IM5b that are both stabilized by coordination of the π bond on the g carbon to the Pd center. A relatively small (1.5 kcal/mol) energy difference between TS3a and TS3b is observed-here, TS3a is slightly more stable due to less steric repulsion between the tert-butyl acrylate and the carboxylate oxygen on the TDG (see SI for 3D structures of TS3a and TS3b). Upon β-hydride elimination and directing group dissociation, IM5a and IM5b form the same 1,3-diene product.</p><p>Next, we investigated the origin of atroposelectivity in the C−H alkenylation. Because the C−H metalation step is irreversible and the rotation about the C(aryl)-C(alkenyl) bond is hindered after palladacycle formation, the atroposelectivity of the 1,3-diene product is determined in the C−H metalation step. We computed the alkenyl C−H metalation pathways from two π-alkene complexes (IM1 and IM2, Figure 4B), where two different π-faces of the alkene bind to the Pd center, leading to two atropoisomers. The chiral center on the TDG significantly impacts the relative stabilities of these two π-alkene complexes and subsequent C−H metalation transition states (TS1 and TS2). In the less stable π-alkene complex IM2, the Pd center is significantly distorted from square planar geometry, making it 9.2 kcal/mol higher in energy than IM1 (Figure 4b). Similar distortion is observed in TS2, which is 11.0 kcal/mol higher in energy than TS1. The distortion in IM2 and TS2 is caused by steric repulsion between the t-Bu group on the TDG and the imine carbon, which are syn-periplanar in IM2 and TS2. Such steric repulsion is diminished in IM1 and TS1, where the Ph group on the imine is not co-planar with the t-Bu. Although the atroposelectivity of the 1,3-diene product 1 is ablated due to the lack of ortho-substituent on the Ar group allowing free rotation about the C(aryl)-C(dienyl) bond, the predicted atroposelectivity is consistent with the X-ray crystal structure of the o-Oi-Pr substituted product 4.</p>

ChemRxiv

Influence of Nanoparticle Processing on the Thermoelectric Properties of (Bi _x Sb _1−X ) ₂ Te ₃ Ternary Alloys

The synthesis of phase-pure ternary solutions of tetradymitetype materials (Bi x Sb 1À x ) 2 Te 3 (x = 0.25; 0.50; 0.75) in an ionic liquid approach has been carried out. The nanoparticles are characterized by means of energy-dispersive X-ray spectroscopy (EDX), powder X-ray diffraction (PXRD), scanning electron micro-scopy (SEM), and transmission electron microscopy. In addition, the role of different processing approaches on the thermoelectric properties -Seebeck coefficient as well as electrical and thermal conductivity -is demonstrated.

influence_of_nanoparticle_processing_on_the_thermoelectric_properties_of_(bi__x_sb_<sub>1−

Introduction<!>Material Synthesis<!>Characterization<!>Thermoelectric Properties<!>Parameters and Reproducibility Study of Hot-Pressing Procedure<!>Investigating the Effect of Annealing<!>Comparison of Nanoparticle Processing by Hot Pressing and Spark Plasma Sintering<!>Thermal Conductivity of the Alloy Series<!>Conclusions<!>Experimental Section Experimental Details<!>Synthesis of Ternary (Bi x Sb 1 À x ) 2 Te 3 Nanoparticles<!>Thermal Analyses<!>NMR Spectroscopy<!>Powder X-Ray Analysis<!>Electron Microscopy<!>Thermoelectric Properties

<p>Today, the alarming risks of global warming and energy shortage continue to rise with the depletion of unsustainable energy sources such as coal, petroleum, and natural gas.</p><p>Despite an increasing demand for energy, considerable amounts of untapped energy in the form of waste heat derive from almost every electrical and mechanical process, including machine operation, oil refining, steel making and food production. In recent years, waste-heat harvesting and recovery using thermoelectric (TE) materials have attracted significant attention as a promising technology to generate clean energy and reduce carbon emission. [1] By definition, TE materials are capable of converting heat to electricity via the Seebeck effect, and electrical energy into cooling via the Peltier effect. TE materials can therefore be used in a wide range of technical applications which include power generators, cooling devices, IR detectors, and gas sensors. [2][3][4][5] The energy conversion efficiency of a thermoelectric material is typically represented by the dimensionless figure-ofmerit zT = (α 2 σT/k), where α is the Seebeck coefficient, σ the specific electrical conductivity, T the absolute temperature in Kelvin, and k the thermal conductivity as the sum of the electronic k el and lattice k L contributions. [1,[6][7][8][9] In order to achieve a high thermoelectric figure-of-merit, the material of interest must exhibit a sufficiently large Seebeck coefficient, a high electrical conductivity, and a low thermal conductivity. Notably, antimony (Sb 2 Te 3 ) and bismuth telluride (Bi 2 Te 3 ) as well as their solid ternary solutions (Bi x Sb 1À x ) 2 Te 3 have already found their way in many commercialized thermoelectric devices due to their remarkable performance near room temperature. [10][11][12][13][14][15][16] These compounds belong to a class of so-called tetradymitetype semiconductors which exhibit heavy doping and narrow band gap characteristics, as well as a unique quintuple-layered structure. [12,13,[16][17][18][19] Despite their high potential as thermoelectric candidates, group V chalcogenides have a strong tendency to generate anti-site defects due to the spontaneous occupation of the Te lattice sites by Sb and Bi atoms. The occurrence of non-stoichiometric compositions reduces the Seebeck coefficient, and hence has limited existing zT values to less than 1. [20][21][22] Though many efforts were made to overcome such intrinsic limitations and optimize the zT value, [23][24][25] the interdependence of the thermoelectric transport properties renders it rather difficult to enhance one physical parameter without diminishing the others. [26][27][28] Interestingly, nanostructuring (i. e. nanograins and metallic nano-inclusions) as well as point defect engineering are long-proven strategies that have helped achieve significant optimization of thermoelectric materials by decoupling the transport parameters. Such phenomena selectively reduce the lattice thermal conductivity in nanostructured materials such as nanoparticles and thin films due to the additional scattering of heat carrying phonons at grain boundaries and interfaces. [9,11,25,27,[29][30][31][32][33] As a result, the synthesis of chalcogenide-based nanostructures of various shape and size has been widely studied using a variety of methods such as electrochemical deposition, solvothermal or hydrothermal approaches [18,[34][35][36] as well as microwave-assisted approaches, [37] mechanical alloying, [38] and gas phase processes (i. e. atomic layer deposition ALD, [39,40] metal organic vapor deposition MOCVD, [41][42][43][44] physical vapor deposition PVD). [45,46] Nevertheless, (Bi x Sb 1À x ) 2 Te 3 ternary solid solutions have demonstrated higher zT values than their binary counterparts, mainly due to their larger unit cell size, lower crystal symmetry, and higher site-occupancy disorder. [47,48] For instance, Xie et al. reported a maximum zT value of 1.56 at room temperature for a (Bi 0.26 Sb 0.74 ) 2 Te 3 bulk material of low dimensional structure, only to achieve an outstanding enhancement of 50 % in comparison to commercial Bi 2 Te 3 . [49] Similarly, using a combination of ball milling and hot-pressing techniques Poudel et al. prepared a ptype nanocrystalline BiSbTe alloy with record-high zT = 1.4 at 380 K. [11] More recently, magnetron co-sputtering methods were employed to fabricate p-type (Bi x Sb 1À x ) 2 Te 3 thermoelectric thin films of various chemical compositions. Song et al. investigated the influence of the Bi content (x = 0-0.57) on the microstructure and electrical transport properties and found that the grain size of the nanocrystalline (Bi x Sb 1À x ) 2 Te 3 films decreased with higher Bi content. Correspondingly, while the carrier concentration of the films decreased with increasing x value, the Seebeck coefficient increased from 114 to 240 μV K À 1 , leading to a maximum power factor of 31.3 μW K À 2 cm À 1 for x = 0.45. [50] On the other hand, Shang and co-workers reported the fabrication of a flexible TE generator using p-type Bi 0.5 Sb 1.5 Te 3based heterostructure films deposited on polyimide substrates. In this work, the random distribution of nanoscale heterojunctions among Bi 0.5 Sb 1.5 Te 3 and newly introduced phases (Te and Sb 2 Te 3 nanoinclusions) helped suppress the thermal conductivity (0.8 Wm À 1 K À 1 ) in order to achieve a good power factor of 23.2 μ WK À 2 cm À 1 . [51] Meanwhile, several groups have also performed low temperature syntheses of (Bi x Sb 1À x ) 2 Te 3 nanoparticles using wetchemical based approaches, however very few have reported on the consolidation of the resulting materials for transport measurements. [48,[52][53][54] Notably, Liu and co-workers hydrothermally synthesized Bi 0.5 Sb 1.5 Te 3 nanocrystals by reaction of ECl 3 (E = Sb, Bi) with elemental Te in the presence of NaBH 4 as reducing agent; the nanocrystal powders were then consolidated by cold-pressing and sintering, whereas Dharmaiah and co-workers applied spark plasma sintering instead. [55,56] Unfortunately, the reductive conditions and low thermal stability of traditional precursors [i. e. BiCl 3 , Bi(AOC) 3 , Bi(NMe 2 ) 3 ] that are typically used in solution-based methods often lead to Te-or Birich materials due to the incorporation of unwanted metal impurities into the crystal lattice. [57][58][59] As a result, off stoichiometric compositions and high number of defects could ultimately poison the Seebeck coefficient and lower the electrical conductivity due to suboptimal changes in the charge carrier concentration. [9,60] Even though wet chemical processes offer more control over the size-and shape-selective syntheses of (Bi , Sb) 2 Te 3 nanoparticles, the presence of organic residues stemming from the stabilizing capping agents (i. e. EDTA, OA, OAC, ODT, DDT, PVP) was shown to drastically compromise the electrical conductivity. [48,56,61,62] Nonetheless, we successfully demonstrated in our previous works the general applicability of ionic liquids (ILs) not only as an effective solvent medium, but also as a stabilizer and shape-directing template. Using weakly coordinating ILs, we synthesized Sb 2 Te 3 , Bi 2 Se 3 , Bi 2 Te 3 , and (Bi x Sb 1À x ) 2 Te 3 nanoparticles with extremely low traces of carbon impurities and surface contamination. [62][63][64][65] As a result, the materials showed exceptional zT values of up to 1.5 due to significant optimization of the electrical conductivity in comparison to commercial powder samples. [65] We herein report on the synthesis of phase-pure ternary tetradymite-type materials (Bi x Sb 1 À x ) 2 Te 3 (x = 0.25, 0.5. 0.75) using a surfactant-free ionic liquid-based approach. The alloys series were synthesized by reacting the Bi-containing ionic liquid [C 4 C 1 Im] 3 [Bi 3 I 12 ] [64] with the single source precursor (Et 2 Sb) 2 Te and were subsequently subjected to a post-thermal treatment. This modified route ensures the formation of pure nanoparticles by removing any ionic liquid residues. Aside from the effect of thermal annealing on the chemical composition and nanostructure of the resulting nanoparticles, we herein also report on the effect of the nanoparticle processing on the resulting transport properties. The chemical composition, phase purity, morphology, and particle surface of the obtained nanomaterials were examined by energy-dispersive X-ray spectroscopy (EDX), powder X-ray diffraction (PXRD), scanning electron microscopy (SEM), and transmission electron microscopy (TEM).</p><!><p>Low temperature syntheses of tetradymite type materials often require reactive metal organic precursors with well-defined decomposition pathways to guarantee the formation of highly stoichiometric products. The thermal decomposition of the single source precursor (Et 2 Sb) 2 Te [61,65] and others [66,67] was previously demonstrated as an effective route for obtaining binary Sb 2 Te 3 nanoparticles with precise chemical composition and defined defect concentration. The low defect and carrier densities of these materials were proven crucial for the optimization of the electronic transport properties, especially the Seebeck coefficient. Similarly, the reactive IL [C 4 mim] 3 [Bi 3 I 12 ] was applied in the wet-chemical synthesis of binary Bi 2 Se 3 and Bi 2 Te 3 nanoparticles. [62] The high reactivity yet increased thermal stability of [C 4 mim] 3 [Bi 3 I 12 ] compared to many other bismuth precursors eliminated the occurrence of unwanted side reactions such as homolytic bond breakage reactions and formation of Bi double layers. Together, the combined decomposition of these precursors will ensure the production of ternary alloys with the correct stoichiometry and defined concentrations of the metal (Bi/Sb) dopant.</p><p>We herein synthesized (Bi x Sb 1-x ) 2 Te 3 (x = 0.25, 0.5, 0.75) nanoparticles from a modified route by reaction of (Et 2 Sb) 2 Te with different amounts of [C 4 mim] 3 [Bi 3 I 12 ] in [C 4 C 1 Im]I. [64] Upon thermolysis of (Et 2 Sb) 2 Te at 150 °C, the reaction was stirred for 12 h, after which the resulting metal chalcogenide powders were repeatedly washed with acetonitrile and isolated by centrifugation. All experimental steps including nanoparticle synthesis and purification were strictly performed under inert gas conditions to avoid surface oxidation of the resulting materials (Scheme 1). For high quality materials, the (Bi x Sb 1À x ) 2 Te 3 nanoparticle series were annealed at 250 °C under a vacuum medium of 10 À 6 À 10 À 7 mbar for 24 h and were then naturally cooled down to room temperature.</p><!><p>According to PXRD, the as-prepared (Bi x Sb 1-x ) 2 Te 3 nanoparticles (x = 0.25, 0.5, 0.75) were confirmed as phase-pure ternary materials (Figure 1). All observed Bragg reflections can be indexed to the rhombohedral (Bi 0.5 Sb 0.5 ) 2 Te 3 (PDF 072-1853, ICSD database). The increasing substitution of the antimony sites by bismuth atoms can be observed from the reflection shifts towards lower reflection angles and from the increase of the lattice parameters. In this case, the gradual exchange of the smaller antimony atoms by larger bismuth atoms would eventually expand the unit cell volume. Most importantly, the steady shift of the highest intensity reflection proves the alloying of a single-phase material rather than the formation of a separate Bi x Te y component.</p><p>In addition, the gradual peak broadening of the full width at half-maximum (fwhm) is consistent with the decreasing size of the crystalline domains in response to the increasing bismuth concentration. Rietveld refinements on the PXRD patterns of (Bi x Sb 1À x ) 2 Te 3 revealed crystallite sizes of 47, 22, and 20 nm for x = 0.25, 0.5, and 0.75 using the Scherrer equation (Table 1). [68] However, the exclusive sharpening of the (110) reflection at 41.61°(2θ), that is observed in the sample containing the highest bismuth content, indicates a preferential growth along the ab-plane perpendicular to the c-axis. Shouldering of the ( 015) and (1010) reflections, which would indicate the presence of elemental Te, are absent in any sample of the alloy series. Other impurity phases including crystalline oxidation products (i. e. Sb 2 O 3 , Bi 2 O 3 , TeO 2 ), metal Sb or metal Bi, as well as additional Bi x Te y phases were also not observed.</p><p>Previous studies proved that the processing temperature of post-thermal treatments may heavily influence the microstructures and transport properties of (Bi x Sb 1À x ) 2 Te 3 based materials by eliminating crystal defects and modifying the chemical composition, grain size, and carrier concentration. [50,55,60,70] We herein annealed the (Bi x Sb À x ) 2 Te 3 nanoparticles in a glass ampoule at a base pressure of 10 À 6 À 10 À 7 mbar for 24 h. The heat treatment temperature was kept below 300 °C to avoid the excessive evaporation of tellurium, which would otherwise lead to Te-deficient materials. Upon completion of the heat treatment, a yellow to orange film was deposited on the cold end of the glass wall, which closely resembles the thermochromic behavior of S2). As expected, the color intensity and glass coverage of the ionic liquid residue that was extracted from the (Bi x Sb 1-x ) 2 Te 3 samples increased with higher concentrations of bismuth. The heat treatment serves as a final purification step to remove ionic liquid residues that remained on the particle surface after [69] Table 1. Refined lattice parameters and crystallite sizes of (Bi x Sb 1-x ) 2 Te 3 nanoparticles. The calculated crystallite sizes should be handled with care due to the anisotropic nature of the samples.</p><p>As prepared the washing procedure. The presence of such impurities could especially compromise the electrical mobility of these nanomaterials. The XRD peak assignment of the alloy series after annealing verified the phase purity of these materials as ternary solid solutions of (Bi x Sb 1À x ) 2 Te 3 (Figure 2). The XRD patterns for all annealed samples show no traces of elemental Bi, Sb, or Te and no sign of additional reflexes such as metal oxide phases, which would indicate a possible oxidation of the materials during the annealing process.</p><p>The calculated crystallite size of the annealed (Bi x Sb 1À x ) 2 Te 3 nanoparticles is almost 3.5 times larger (Scherrer's equation) compared to the original samples (Table 1). As previously observed, the decreasing crystallite size order of the annealed samples is consistent with the increasing concentration of bismuth (x). When directly comparing the XRD patterns of the alloy series before and after annealing, the narrowing of the fwhm of the annealed samples indicates a higher degree of crystallinity and an increase in the crystallite size.</p><p>Quantification of the EDX spectra within standard deviations confirmed the stoichiometric compositions of the ternary alloy series which closely correspond to the theoretical Bi,Sb : Te ratio of 40 : 60 (Table 2). The bismuth to antimony elemental proportions in the sum formula (Bi x Sb 1À x ) 2 Te 3 could be approximated to the desired compositions where x = 0.28, 0.51, and 0.65. The EDX results agree with the PXRD measurements, hence confirming the high stoichiometry of the (Bi x Sb 1À x ) 2 Te 3 nanoparticle alloys.</p><p>EDX elemental mapping analyses revealed a homogeneous distribution of Bi, Sb, and Te within the materials (Figure 3).</p><p>No significant changes in the material compositions were observed after annealing, which is consistent with the results obtained from PXRD, and no signals originating from ionic liquid contaminations were detected in the EDX measurements. SEM images of the (Bi x Sb 1À x ) 2 Te 3 ternary materials showed the formation of anisotropic nanoparticles with the hexagonal plate-like morphology (Figure 4). The average edge lengths of the as-prepared materials are 70, 55, and 53 nm for x = 0.25, 0.5, and 0.75, respectively. After the annealing process, the average edge length increased to 97, 87, and 80 nm, respectively, while the average particle thickness increased from 16 to 35 nm.</p><p>TEM bright field images revealed the hexagonal plate-like shape of the (Bi x Sb 1À x ) 2 Te 3 nanoparticles ranging from 30 to 200 nm in size and 10 to 20 nm thickness (Figure 5).</p><p>Due to the elimination of shape and size-directing capping agents, large size distributions of small spherical particles as well as large intergrown plates can be observed in the resulting (Bi x Sb 1À x ) 2 Te 3 samples. High-resolution TEM measurements (HR-TEM) confirmed the crystallinity and clean surfaces of the (Bi x Sb 1À x ) 2 Te 3 nanoparticles. In particular, the HR-TEM dark field image of the (Bi 0.75 Sb 0.25 ) 2 Te 3 ternary phase in [110] orientation displays uniform quintuple layers of 1.02 nm thick, which confirms that there are no additional Bi-bilayers in the lattice structure before (as-prepared) and after annealing (Figure 6). Hence, the results prove that the annealing step does not negatively affect the crystallinity of the final materials. Analogous findings were observed for the (Bi 0.25 Sb 0.75 ) 2 Te 3 ternary phase (Figures S4,S5). Additionally, the SAED pattern of (Bi 0.75 Sb 0.25 ) 2 Te 3 (Figure 5c) once again verifies the crystalline nature of the material. The crystal structure of the resulting (Bi x Sb 1À x ) 2 Te 3 materials is thus proven to adopt a layered and infinite arrangement of quintuple layers (QLs) that are stacked along the c-axis (Figure S3). In each QL, one could observe five atomic layers with a Te(1)-Bi-Te(2)-Bi-Te(1) stacking sequence that is terminated by a Te(1) atomic layer on both sides. More specifically, the bonding between the Bi or Sb and Te atoms inside the quintuple QL are best described as strong covalent and ionic interactions, whereas the interactions among the adjacent QLs consist of weak van der Waals forces. [16]</p><!><p>Thermoelectric properties of alloyed chalcogenides are often optimized by the approach of high energy ball milling followed by rapid compaction like SPS. [73] The alloy composition hereby may even be adjusted purely mechanically during the milling procedure. This approach provides a pragmatic guideline towards high zT materials, but makes an assessment of the influences on the thermal conductivity reduction difficult by the multitude of potential sources like alloy composition itself, nanoparticle structure, or other defects that are typically produced by the ball milling. In part, even the amorphization of the nanoparticles by the high energy ball milling may be expected. Motivated by the excellent purity and crystalline integrity of the here-synthesized nanoparticles, we assess these influences in our compacted nanoparticle pellets. The synthesis and all subsequent processing steps were performed under inert gas conditions, hence avoiding the effect of oxygen impurities. We further optimize the nanoparticle processing steps and discuss the in-plane and through-plane thermal conductivity of the obtained nanocrystalline bulk pellets.</p><!><p>The limited availability of powder per batch obtained by the chemical synthesis made it necessary to adopt methods and tools appropriately to be able to produce highly dense pellets of small quantities of powder in the order of 70 mg per pellet. Note that it is not common to process this small quantity of nanoparticles, since usually hot pressing or spark plasma sintering is carried out with batches of several grams. 70 mg of the nanoparticles were hot-pressed at comparatively mild temperature of 300 °C and 100 MPa, being this the optimized parameters of the process with respect to the density of the final product. Therefore, we first performed a reproducibility study of this hot-pressing procedure to identify the variations from sample to sample and verify stable processing conditions. For reasons of simplicity, we used the binary compound Sb 2 Te 3 for this study, characterized in-depth in an earlier work, [72] synthesized by an IL approach that is similar to the one developed for the alloy nanoparticles. [72] A set of hot-pressed samples of Sb 2 Te 3 were produced by keeping the processing parameters identical and the complete thermoelectric properties were characterized (Figure 7).</p><p>The hot-pressing procedure resulted in reproducible over-all properties and almost identical zT values (~1 at 550 K) of all samples within the given error bars. These results are comparable with published data, [70] despite the severely reduced amount of powder and a different kind of processing. Often, variation in the processing of the powder might result in a strong variation of the transport properties such as the thermal conductivity, electrical conductivity and so on. Therefore, we performed this part of the study to emphasize that both synthesis and processing are suitable for revealing the intrinsic properties of the materials rather than existing impurities and defects induced by the processing.</p><!><p>The here-synthesized alloy nanoparticles were in part objected to an annealing procedure. This additional annealing step was intended to verify the cleanliness of the nanoparticles since residual IL could be removed by exposing the nanoparticles to a base pressure of 10 À 6 À 10 À 7 mbar for 24 h at temperatures below 300 °C. The Seebeck coefficient is the transport coefficient that is most sensitive to any changes in the Fermi level of the semiconductor and therewith the chemical composition of the nanoparticles. Therefore, in Figure 8, we compare the temperature dependent Seebeck coefficients of the nanoparticles of the alloy series, all processed by the optimized hotpressing, with and without annealing. It is evident that the annealing procedure does not alter the Seebeck coefficients and therewith the Fermi level of the materials. The trends of the measured Seebeck coefficients represent the expected behavior of this alloy series with respect to the Bi content. This therefore proved as a beneficial additional step in the processing to further purify the surfaces of the nanoparticles, without affecting the chemical composition and electronic band structure that were intentionally designed by the complex synthesis procedure.</p><!><p>The consolidation of the powder may affect the transport properties of the pellet. The sintering affects the microstructure of the macroscopic pellets, but also the reconstruction of the individual interfaces at the atomic level. [74] We here use a bottom-up approach for the nanoparticle synthesis that emphasizes on structural and crystalline integrity of the nanoparticles. Therefore, the main objective of the processing is compaction rather than reconstruction of the individual grains.</p><p>This experiment was designed in a way to directly compare the influence of hot-pressing and spark plasma sintering. For this, two pellets of the same nanoparticle batch of the composition (Bi 0.75 Sb 0.25 ) 2 Te 3 were produced by the optimized hot-pressing procedure and a spark plasma sintering procedure, respectively. Because of the different technologies used to densify the nanoparticles, parameters like temperature and pressure had to be adapted. Figure 9a) shows the obtained electrical in-plane conductivities of the two pellets, together with an analysis of carrier mobility and carrier density (Figure 9b). Hereby, the electrical conductivity provides a good measure for this comparison since it is sensitive towards the quality of crystalline interfaces and grain boundaries. Both samples differ in their density: The hot-pressing procedure resulted in a density of 6 g cm À 3 (81 %), while the SPS processed sample showed only a density of 5.1 g cm À 3 (70 %) after the typical signatures of sample shrinking detected within the SPS machine by tracking the movement of the tools. Note that higher temperatures or longer hold times realized after the observation of sample shrinking typically only result in partial melting. [71] As a result, it was found that the hot-pressing procedure resulted in an electrical conductivity of around 250 S cm À 1 at RT that was higher than the electrical conductivity of 50 S cm À 1 of the SPS sample. This was caused by both, an improved charge carrier density as well as an improved charge carrier mobility of this n-type alloy sample. We conclude that the hot-pressing procedure generates the conditions for an efficient thermal activation of carriers in the pellet. Further, defects, especially pores, that reduce the carrier mobility are less favored by the hot-pressing than by the SPS processing. Hence, hot-pressing is better suitable to allow for a dense arrangement of the here-used nanoparticles in the compacted pellet, also resulting in improved electrical properties.</p><!><p>The parts of the study presented above were all intended to develop a reliable processing for small quantities of nanoparticles. By the processing in a glove box cluster, we avoid the incorporation of unwished oxygen impurities. The hot-pressing provides a high density and high reproducibility of the obtained pellets (Figure 7). The additional annealing of the nanoparticles improves the purity with respect to potential incorporation of residues of the IL, without affecting the chemical composition (Figure 8). Therefore, the combination of the sophisticated nanoparticle synthesis combined with this processing provides a means to understand the influence of the nanoparticle alloying on the thermal conductivity in-plane and throughplane of the pellets. Influences that are typically found in alloyed nanoparticles like the incorporation of defects from mechanical alloying by ball milling or the incorporation of impurities by the chemical synthesis are not expected to play a major role here.</p><p>Figure 10 shows the in-plane and through-plane temperature dependent thermal conductivity of annealed nanoparticles of the composition (Bi 0.75 Sb 0.25 ) 2 Te 3 and (Bi 0.25 Sb 0.75 ) 2 Te 3 . The change in the thermal conductivity with Bi content is consistent with what is expected from previous literature. [75] With less Bi content, the thermal conductivity increases. Clearly, the thermal conductivity of the pellets exhibits an anisotropy indicating that the nanoparticles align within the pellets with a preferred orientation of the crystallographic ab-plane in-plane of the pellets. This was found before for similarly synthesized and compacted Bi 2 Te 3 and Bi 2 Se 3 nanoparticles. [62] Further we find that the thermal conductivity is higher than expected for nanocrystalline bulk of this alloy compositions. Single crystalline reference data of similar compositions is added to Figure 10. Hence, the thermal conductivity of our samples is in-between the single crystalline reference and nanocrystalline bulk of other processes. For instance, for the alloy composition (Bi 0.25 Sb 0.75 ) 2 Te 3 a ball milling and hot-pressing approach resulted in isotropic samples, which showed thermal conductivities of around 1 W m À 1 K À 1 at room temperature. While this value corresponds to the here-presented throughplane direction, it is significantly reduced compared to the thermal conductivity in in-plane direction, [73] and would not represent the isotropic average over all crystalline directions. For (Bi 0.7 Sb 0.3 ) 2 Te 3 , the in-plane thermal conductivity was found to be around 1.3 W m À 1 K À 1 at room temperature for a sample that was also exposed to a combination of ball milling and SPS. [70] A recent report on a micro-wave assisted solutionsynthesis of (Bi x Sb 1À x ) 2 Te 3 nanoparticles followed by SPS processing [37] yielded pellets, which featured a low thermal conductivity of below 1 W m À 1 K À 1 at room temperature for a composition of (Bi 0.25 Sb 0.75 ) 2 Te 3 . This comparison shows that the thermal transport properties of the obtained product are not mainly defined by the chemical composition, but rather by the sum of the synthetic details such as the used additives, as well as the processing details including the compaction method. In an earlier study, we already showed that high purity Bi 2 Te 3 and Bi 2 Se 3 compacted nanoparticles exhibited an exceptionally high thermal conductivity close to that of single crystals or even abinitio calculations for bulk. [62] We now showed for the alloyed nanoparticles a consistent behavior. Obviously, the oftenreported low thermal conductivity of samples prepared by ball milling followed by rapid compaction originates from the morphological peculiarities like point defects or dislocations rather than the chemical composition or the nanostructure itself.</p><p>While the here-synthesized nanoparticles provide a close-toperfect model system to study the influence of chemical composition and nanostructure on the transport properties, they obviously cannot compete with the best thermoelectric materials of the nominally identical composition due to their high thermal conductivity which results from the extraordinary high purity and crystalline integrity of the nanoparticles. Consequently, zT values of these alloyed nanoparticles range from zT = 0.06 to zT = 0.15 at 400 K for the composition of (Bi 0.25 Sb 0.75 ) 2 Te 3 in in-plane direction, depending on the details of the applied processing.</p><!><p>Highly stoichiometric, phase-pure ternary solid solutions of the type (Bi x Sb 1À x ) 2 Te 3 were synthesized by thermolysis of the single source precursor (Et 2 Sb) 2 Te with various amounts of the Bi source [C 4 mim] 3 [Bi 3 I 12 ] in a surfactant-free ionic liquid-based approach in [C 4 mim] 3 I. The phase purity as well as the perfect stoichiometric composition and homogenous element distribution within the resulting nanoparticles was proven by XRD, SEM, EDX and TEM. After compaction of the nanoparticles, the thermoelectric transport properties, Seebeck coefficient, electrical conductivity and thermal conductivity were determined. Our studies clearly proved the crucial role of nanoparticle processing on the resulting transport properties.</p><!><p>Nanoparticle synthesis, thermolysis experiments, fabrication of pellets, and electrical contact preparations were performed under inert conditions (Ar atmosphere) in a glovebox or using standard Schlenk techniques to avoid any oxidation reactions. The powders and pellets were carried using sealed and Ar filled vessels until the materials were finally transferred to the measurement devices. Solvents were dried over CaH 2, stored over molecular sieves, and degassed prior to use. 1-N-Methylimidazole (99 %, Sigma Aldrich), 1-halobutanes (99 %, Acros) and CH 3 CN (99.9 + %, Acros) were commercially available, while [C 4 C 1 Im]I, [C 4 C 1 Im] 3 [Bi 3 I 12 ], and (Et 2 Sb) 2 Te were prepared by literature methods. [64,76]</p><!><p>The corresponding amount (see Table 3) of [C 4 C 1 Im] 3 [Bi 3 I 12 ] was dissolved in 5 mL of [C 4 C 1 Im]I, the red ionic liquid solution was stirred under an Ar atmosphere at 150 °C. 1028 μl (4.1 mmol) of (Et 2 Sb) 2 Te were added dropwise yielding a black suspension which was stirred at 150 °C for 12 h. The resulting black precipitate was centrifuged and repeatedly washed with 20 mL of dry acetonitrile (6x). The solid product was dried in vacuum at ambient temperature. An aliquot was transferred to a closed ampoule to be annealed at 250 °C for 24 h under dynamic vacuum (10 À 7 mbar).</p><!><p>TGA/DTA studies were performed using a Mettler-Toledo DSC Star1 system that is operated under inert gas conditions.</p><!><p>1 H (400 MHz) and 13 C{1H} (75.5 MHz) NMR spectra (δ in ppm) were recorded using a Bruker AVNEO 400 MHz spectrometer and were referenced to internal DMSO-d 6 ( 1 H: δ = 2.50; 13 C: δ = 39.52).</p><!><p>PXRD patterns were collected with a Bruker D8 Advance diffractometer with Cu Kα radiation (λ: 1.5418 Å, 40 kV, 40 mA) using a Si single crystal as sample holder to minimize scattering. The powder samples were re-dispersed in EtOH on the Si surface and investigated in a 2θ range from 10 to 90°with a step size of 0.01°2 θ (counting time 0.6 s). Rietveld refinement was done with the program package TOPAS 4.2 (Bruker) to determine lattice parameters and average crystallite sizes by using the Scherrer equation. [68] The background was modelled using Chebyshev polynomials. The structure model of Bi 2 Te 3 (PDF 15-863) from the ICSD database was used. For each Rietveld refinement, the instrumental correction determined using a standard powder sample LaB 6 from NIST (National Institute of Standards and Technology) as reference material (SRM 660b; a(LaB 6 ) = 4.15689 Å) was considered.</p><!><p>The particle morphology and elemental composition of the powdered samples were analyzed by scanning electron microscopy (SEM) using Apreo S Lovac microscopes equipped with a Bruker Quantax 400 units (EDX). TEM studies were carried out on a Jeol JEM 2200 fs microscope equipped with probe-side Cs-corrector operated at 200 kV acceleration voltage.</p><!><p>Pellets were prepared using a manual hydraulic press (Specac Atlas) and a custom-made heating mantle equipped with a PID temperature control. A pressure of 100 MPa was applied for 90 minutes while the temperature of the stainless-steel die was kept at 300 °C. One pellet was sintered by spark plasma sintering procedure using SPS 210-Gx (AGUS). For sintering, a pressure of 35 MPa was applied for 1 minute at temperature of 240 °C. The density of the pellets was determined geometrically. Thermoelectric properties were measured from room temperature up to 240 °C. The thermal diffusivity was measured along the pressing direction and perpendicular to the pressing direction (Linseis LFA). Thermal conductivity was calculated from the diffusivity data with the relation k = D 1 c p where D is the thermal diffusivity, 1 the density, and c p the heat capacity. Bulk literature values were used for c p . [70] The Seebeck coefficient and the electrical conductivity were measured perpendicular to the pressing direction (Linseis LSR-3). The majority carrier concentration and carrier mobility were determined perpendicular to the pressing direction at room temperature with Hall measurements in the van-der-Pauw geometry (Quantum Design Versalab).</p><p>Supporting Information: PXRDs including the Rietveld refinement of (Bi x Sb 1-x ) 2 Te 3 (x = 0.25, 0.5, 0.75) nanoparticles, crystal structure of (Bi x Sb 1-x ) 2 Te 3 materials, and 1 H-NMR spectrum of ionic liquid contamination from nanoparticle annealing.</p>

Chemistry Open

Designed Single-Step Synthesis, Structure, and Derivative Textural Properties of Well-Ordered Layered Penta-Coordinate Silicon Alcoholate Complexes

The controllable synthesis of well-ordered layered materials with specific nanoarchitecture poses a grand challenge in materials chemistry. We report the solvothermal synthesis of two structurally analogous 5-coordinate organosilicate complexes via a novel transesterification mechanism. Since the polycrystalline nature of the intrinsic hypervalent Si complex thwarts the endeavor in determining its structure, a novel strategy concerning the elegant addition of a small fraction of B species as an effective crystal growth mediator and a sacrificial agent is proposed to directly prepare diffraction-quality single crystals without disrupting the intrinsic elemental type. In the determined crystal structure, two monomeric primary building units (PBUs) self-assemble into a dimeric asymmetric secondary BU via strong Na+-O2\xe2\x88\x92 ionic bonds. The designed one-pot synthesis is straightforward, robust, and efficient, leading to a well-ordered (10\xc4\xab)-parallel layered Si complex with its principal interlayers intercalated with extensive van der Waals gaps in spite of the presence of substantial Na+ counterions as a result of unique atomic arrangement in its structure. On the other hand, upon fast pyrolysis, followed by acid leaching, both complexes are converted into two SiO2 composites bearing BET surface areas of 163.3 and 254.7 m2 g\xe2\x88\x921 for the pyrolyzed intrinsic and B-assisted Si complexes, respectively. The transesterification methodology merely involving alcoholysis but without any hydrolysis side reaction is designed to have generalized applicability for use in synthesizing new layered metal-organic compounds with tailored PBUs and corresponding metal oxide particles with hierarchical porosity.

designed_single-step_synthesis,_structure,_and_derivative_textural_properties_of_well-ordered_layere

Introduction<!>Solvothermal synthesis<!>PXRD analysis<!>SCXRD, 29Si and 13C MAS NMR analyses<!>11B MAS NMR analysis<!>ATR-FTIR spectroscopy<!>EA<!>TG and DTG analyses<!>SEM observations<!>Gas sorption analyses on the pyrolyzed products<!>Conclusion<!>Experimental Section<!>Synthetic procedures<!>Solvothermal synthesis of the B-undoped Si complex<!>Solvothermal synthesis of the B-assisted Si complex<!>Pyrolytic synthesis of porous silica composites

<p>In contrast to the ubiquitous tetrahedrally coordinated silicate species, the reports of 5-coordinate Si ([5]Si) species have so far been infrequent in alkoxide-based crystalline solids.[1-6] Consequently, there are still some unknown aspects about their physicochemical properties. The laborious preparations of several [5]Si complexes were disclosed in multiple synthetic steps[1,3] by reacting SiO2 grains with alkali hydroxide in excessive ethylene glycol (EG) solvent with continuous distillation of EG and coincident distillative removal of liberated water to drive the dissolution process to proceed smoothly. Additionally, this technique encountered severe foaming in the presence of Na+ cations over the course of the synthetic process. More importantly, water-sensitive or hygroscopic silicoalcoholate complexes cannot be produced by this method due to the unavoidable water interference therein. Although the conventional transesterification reaction involving carboxylic acid esters has already been well-exploited in organic synthesis, e.g., biodiesel production from triglyceride and light alcohols,[7] it is not applicable for the preparation of metal-organic compounds (e.g., layered 5-coordinate organosilicates) without any modification of the reaction.</p><p>Crystallography and NMR are the most powerful techniques available for unraveling unknown molecular structures. However, neither polycrystalline powders nor single crystals of 5-coordinate alkoxysilicates with mixed methanolate/glycolate ligands have been thus far disclosed, not to mention their crystal structure determination[1,3a]. In general, growing large single crystals indispensable for crystallography can be classified into two major categories: direct and indirect methods. In the latter case, it manages to transform the polycrystalline powders into the respective single crystals by time-intensive recrystallization techniques provoked primarily by controlled solvent evaporation, slow cooling, vapor diffusion, liquid-liquid inter-diffusion, sublimation, and convection, etc. Another disadvantage associated with the indirect method is that the recrystallized single crystals are occasionally phase transformed.[8] Particularly, the effective preparations of crystallographically resolvable single crystals are a significant challenge for Si-containing compounds owing to the strong propensity of silicates towards fast gelation even at the reactant mixing stage. As a general countermeasure against the undesirable gelling, excessive strong chelators and/or sparingly soluble silica sources, e.g., quartz, are admixed together with dilute alkaline solutions to regulate the relative contribution between nucleation and crystal growth,[9-11] thereby causing dramatically extended synthetic duration and aggravated waste disposal loads.</p><p>The versatility of well-ordered layered materials makes them quite attractive for several important applications including lithiumion batteries,[12] supercapacitors,[13] drug delivery,[14] adsorbents,[15,16] catalysts,[17] and catalyst supports.[18] Among the known layered materials, those with their interlayers intimately linked together only by long-range attractive force of van der Waals type (i.e., so-called van der Waals gap as typically found in the FeOCl[12a] and TiS [12c]2) are quite rare in the presence of charge-balancing alkali metal counter-ions. This feature could be quite useful for advanced lubricants and precursors towards preparing nanoflake/polymer composite membranes for gas-/liquid-mixtures separations.</p><p>In this contribution, we report a novel single-step solvothermal preparative methodology, i.e., a modified transesterification mechanism, for preparing two layered [5]Si complexes with a similar asymmetric unit. In addition, a new B-mediated approach that does not involve any B incorporation into the intrinsic host silicate matrix is formulated herein for the facile synthesis of large-sized single crystals of the hitherto poorly-characterized binary-ligand [5]Si complex, with which its crystal structure is for the first time determined in this study. Also, we demonstrate a proof-of-concept study regarding the rapid pyrolytic conversion of these Si complex precursors into high-surface-area amorphous SiO2 composites with minor residual carbon loadings.</p><!><p>A modified transesterification strategy using pseudo-covalent metal alkoxide ester instead of the traditional carboxylic acid ester is proposed for the one-pot synthesis of two well-ordered layered [5]Si complexes under anhydrous circumstances. Over the course of solvothermal reaction involving B mediator, the effective deprotonation of EG reactant (only 1.5-fold excess from stoichiometry) is initially accomplished by the operative strong base catalyst of excessive sodium methoxide to initiate the SN2 nucleophilic substitution reaction, leading to the formation of ethylene glycolate anions. These incoming anions thus formed then compete with stoichiometric amounts of methanolates for nucleophilic attack of the Si electrophiles, thereby causing the concurrent departure of CH3O− anions from Si(OCH3)4 reactant. Due to the chelating nature of the bidentate glycolate, Si electrophile obviously possesses a coordinative preference for glycolate over monodentate methanolate, thus affording a unique [5]Si complex at an ultimate methanolate/glycolate ratio of 1:2 in the empirical formula of the B-assisted Si complex (Table S1). The designed transesterification mechanism is advantageous because it only involves alcoholysis but precludes any hydrolysis side-reaction, thereby effectively eliminating the coincident generation of metal hydroxide precipitate as a contaminant. It also offers the opportunity and flexibility for designing novel moisture-sensitive or hygroscopic compounds.</p><p>It is worth noting that initial studies were dedicated to detailed characterization of the intrinsic [5]Si complex without any B predosing. However, this attempt was thwarted by the failure in directly achieving X-ray quality single crystals. Inspired by the intriguing borosilicate glass and zeolite chemistries,[19-21] where B heteroatom shows the difficulty in isomorphously inserting in considerable proportion into the silicate network matrix due to the substantial mismatching between d(Si-O) (1.58-1.64 Å)[22] and d(B-O) (1.44-1.52 Å) both in 4-coordinate geometries,[23] an efficient strategy, by which crystallographically resolvable single crystals are successfully achieved by taking advantage of B-mediated transesterification synthesis, is developed herein in order to address the hitherto unknown crystal structure of the [5]Si complex bearing mixed alcoholates. It is found that B(OCH3)3 additive allows for a well-mobile thin gel precursor formulation rather than the instantaneous creation of an undesirable sluggish thick gel entity in the absence of B mediator. In this case, the vast majority of B species are supposed to still partition in the dynamically equilibrated bulk solution in close contact with the gel phase. The dynamic bi-phasic B distribution effectively inhibits diverse Si species from full condensation into a thick gel mass that is detrimental for large single crystal growth. It is well-known that B forms diester complexes with polyol chelators more readily than does Si. It therefore seems likely that organosilicate crystal growth is seeded by rapidly created bis(glycolato)borate crystal nuclei. When the crystallization process progresses, B would be gradually expelled from the growing crystal lattice because it cannot support 5-coordination unlike Si. The hypothetical [5]B species isomorphously substituting for [5]Si sites is highly energetically unfavorable due to the large negative charge density (i.e., [5]B2−) localized on the small B center, preliminarily validating the lack of detectable B insertion into the 5-coordinate organosilicate matrix. Hence, this B-mediated synthetic methodology proposed here provides a unique opportunity to study in depth Si chemistry by rationally growing the relevant single crystals.</p><!><p>Experimental powder X-ray diffraction (PXRD) patterns of the intrinsic Si complex (A), B-assisted counterpart (B), and simulated PXRD pattern of a single crystal of the latter (C) are graphically stacked in Figure 1. Both samples are highly crystalline, as evidenced by negligible background intensities and very sharp major peaks. The zoomed-in profiles in the high-angle regime (Figure S1 A and B) reveal that following B dosing (5.3% relative to Si), the XRD pattern differs to some extent from that of the intrinsic Si complex in terms of peak 2θ position and relative intensity. Interestingly, regardless of the preparative methodology, the intrinsic Si complex exhibits the XRD pattern analogous to those reported in Refs.[2,3] for Na2Si2(OCH2CH2O)5 complex but with the most intense peak shifting from 11.02° 2θ for the former down to 10.81° 2θ for the latter, presumably originating from the swelling effect of EG solvent adopted in the latter syntheses. Namely, they are isostructural with each other. Combining B and C in both Figures 1 and S1, the experimental PXRD pattern of the B-assisted Si complex closely matches the simulated pattern derived from the single-crystal XRD (SCXRD) structure, manifesting the high phase purity of the well-developed single crystalline grains. The strongest peak corresponding to (10ī) reflection at 11.27° 2θ (7.84 Å in interplanar d-spacing) is indicative of a well-organized layered B-assisted Si complex with its dominant interlayers parallel to the (10ī) plane (Figure 1 B).</p><p>After rapid pyrolysis of both Si complexes at 800 °C for 8-10 min, the resultant samples of Si_py and Si+B_py are proven X-ray amorphous, regardless of B dosing (data not shown).</p><!><p>The molecular geometry and atom-labeling scheme for the B-assisted Si complex are illustrated in the thermal ellipsoid plot of Figure 2. The complex crystallizes in the monoclinic crystal system and the space group P21/n (Table S1) with the asymmetric unit containing one dimeric sodium tetrakis(glycolato)bis(methanolato)silicate complex Na2Si2(OCH3)2(OCH2CH2O)4. Both Si centers are uniquely 5-coordinate; each is bound by two bidentate glycolate chelators and one monodentate methanolate ligand. 29Si magic angle spinning nuclear magnetic resonance (MAS NMR) (Figure 3 A) reveals two distinct [5]Si resonances located at -105 and -108 ppm further supporting the assignment of two crystallographic Si sites.[2,3a,4] The 13C{1H} MAS NMR spectrum (Figure 3 B) shows two well-resolved regions which can be assigned to the ethylene glycolate (63 ppm) and methanolate (53 ppm) ligands.[24] The carbon ratio between the glycolate and methanolate sites using both direct excitation and cross polarization (CP) gives ca. 4:1, corroborating the monomeric BU structure determined by SCXRD studies. Since the carbon atoms (C5 and C6) of one of the four glycolates are position-disordered, we speculate the disorder of the glycolates is the cause of the extra shoulder resonances observed in the 13C MAS NMR spectrum (i.e., 66 and 54 ppm in Figure 3 B). Similarly, two discernible regions are also observed in the 13C{1H} MAS NMR spectrum of the B-undoped Si complex Na2Si2(OCH2CH2O)5 (Figure 3 C) with the least intense resonance at 52.5 ppm probably attributable to a very small number of the unbound MeOH solvent molecules tightly occluded in the interplanar voids, as found in Ref.[24]. There remain three resolvable resonances in the bound glycolate domain with an intensity ratio of ca. 1:2:2. Among them, the weakest resonance is assigned to the sole bridging glycolate ligand with the other two resonances corresponding to the remaining four glycolate chelators. The position-disordered carbon atoms of glycolates like C5 and C6 in Figure 2 may account for the resonance splitting of these four glycolates situated in a locally similar chemical environment.</p><p>As shown in Figure 2, concerning one monomeric BU NaSi(OCH3)(OCH2CH2O)2, the central [5]Si1 atom exclusively adopts an quasi-trigonal-bipyramidal geometry. Oxygens O1 and O3 occupy the axial positions and form the elongated Si1-O bonds of 1.7656(10) and 1.7422(10) Å, respectively.[32] Oxygens O2, O4, and O9 take up the equatorial positions at a little shorter Si1-O bond lengths of 1.695(10), 1.6858(10), and 1.6762 (10) Å, respectively. On the basis of the respective atomic coordinates,[32] it is calculated that the Si1 center resides 0.051 Å below the plane defined by O2, O4, and O9 in reference to the O1 apex. These bond lengths are in close proximity to the Si-O bond distances found in Refs.,[3,25] where [5]Si complexes with their atomic structures different from this study were reported, but slightly longer than d(Si-O) of 1.59-1.64 Å for [SiO4] tetrahedra contained in zeolites.[26,27] Partially arising from inherently asymmetric stereo-configuration of mixed alkoxy ligands, distortions from the idealized trigonal-bipyramidal geometry towards a square pyramid with O9 at the apex are conceivable, as has been found for other [5]Si compounds. On the other hand, there exist six Na-O contacts in the 2.3032(11)-2.4106(11) Å scope definitely contributing more or less to such distortions.</p><p>Packing plots of the B-assisted Si complex viewed along the crystallographic b and [101] axes are provided in Figure 4 A and B, respectively. By coupling both figures, it can be envisaged that the 3D crystal structure is constructed of periodically aligned layers parallel to (10ī) plane instead of parallel chain-like BUs, which agrees well with the above PXRD results. Interestingly, both Si and Na atoms are in-plane with the layers with their peripherals decorated with the arrays of methanolates and glycolates, resulting in the interlayers uniquely held together by extensive van der Waals attractive force. The weak interaction of this type may readily bring about the interplanar slip. In-plane Na+ ions behave both as charge-counterbalancing cations and cross-linkers, strongly tethering the monomers in each layer together by strong Na+-O2−ionic bonding interactions. Due to full condensation of the EGs catalyzed by NaOCH3, the absence of free hydroxyl groups in its structure interprets the lack of identifiable hydrogen bonding interactions on inter-atoms in the entire crystal lattice.</p><!><p>SCXRD analysis shows no evidence of B heteroatom incorporation within the crystal lattice of the B-assisted Si complex. However, one single crystal may not be representative of the entire sample. Its bulk powders is thus qualitatively analyzed via 11B MAS NMR (data not shown) in light of the inherent sensitivity for detecting 11B chemical environments due to its high natural abundance (80.1%). No detectable B species are identified, suggesting the B content is << 1 wt%,[28,29] if any.</p><!><p>Attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectra of the B-undoped sample (A), B-assisted one without (B) and with (C) exposure to air for 10 min are presented in Figure 5. For the B-undoped polycrystalline powders, a broad band at 3000-3600 cm−1 attributable to hydroxyl absorption signal (Figure 5 A) is a combined result of physisorbed moisture and encapsulated MeOH solvent residues, as corroborated by 13C MAS NMR (Figure 3 C). In contrast, this weak broad band is significantly enhanced in intensity even subjecting to ambient exposure for 10 min (Figure 5 B and C), highlighting strong hygroscopicity of the B-assisted Si single crystals. As shown in Figure 5 A and B, the absorption band at 2750-2980 cm−1 is assigned to the sp3 C-H bond stretching vibrations, whereas the band at 1340-1500 cm−1 corresponds to the C-H bond bending vibrations. More specifically, two discernible peaks at 2870/2950 cm−1 (methyl C-H symmetrical/asymmetrical stretching vibrations) confirm the existence of methyl groups within both Si complexes but with the methyl origin differing from each other (residual MeOH solvents for the intrinsic sample vs. methanolate ligands for the B-assisted one). Strong Si-O-C bond stretching vibrations at 933-1110 cm−1 are observed in both samples, in which the 1030-1070 cm−1 absorption band originating from the C-O bond stretching vibrations dominates the above spectral range. The absorption band at 600-900 cm−1 is attributed to varying Si-O bond vibrations including O-Si-O bond vibrational frequency.[30] Any distinguishable peak at ca. 670 cm−1 corresponding to the O-B-O bond bending vibrations[31] does not appear in the B-assisted Si complex (Figure 5 B and C), revealing the B content below the detection limit of IR spectroscopy.</p><!><p>Bulk elemental analysis (EA) is conducted to quantitatively determine the final whereabouts of the B species. Table 1 tabulates the elemental composition of both Si complexes along with the refined SCXRD empirical formula for comparison sake. Based upon earlier PXRD and 13C MAS NMR analyses, Na2Si2(OCH2CH2O)5 formula is assigned to the B-undoped Si complex, which is qualitatively in accord with the apparent composition of Na1.94Si2C9.80H21.44O10.82. In view of the deliquescence-prone nature of the B-assisted sample (Figures 5 and 6), the drying step preceding EA is supposed to drive off a small number of hydrolytically produced MeOH and physisorbed moisture, thus resulting in a slight underestimation of both carbon and hydrogen contents in relation to SCXRD result. On the other hand, the finding that the B content is too low to be accurately quantified provides the most direct evidence, lending further support to the conclusion that the B-assisted Si complex is devoid of identifiable B incorporation.</p><!><p>The thermal stabilities of both Si complexes are evaluated by thermogravimetric (TG) technique both operating in flowing N2 streams, as illustrated in Figure 6. For the B-undoped polycrystalline powders, between 30 and 370 °C, there exists a minor weight loss (1.2%) likely corresponding to the evaporation of physisorbed water and residual MeOH solvents. With increasing temperature, one pronounced weight loss is observed before the final plateau. A dramatic weight loss associated with the framework collapse is observable over 370-550 °C with a temperature inflection point at 430 °C. In comparison with the theoretical overall weight loss of 56.4%, the corresponding value of 53.8% is experimentally obtained while ramping the temperature from RT to 800 °C in an anaerobic environment, ultimately leaving behind the blackened end powders of decomposition. Namely, after pyrolysis a residual carbon loading of ca. 2.6% for the Si_py composite is theoretically predicted.</p><p>In contrast, the onset degradation temperature of the B-assisted single crystals occurs at 320 °C with a total weight loss of 2.8% from RT to 320 °C. A larger weight loss is responsible for stronger hygroscopicity of the B-assisted Si complex relative to the undoped counterpart. In the entire temperature range, a total weight loss of 52.0% is measured. The theoretical weight loss for the conversion from Na2Si2C10H22O10 to an oxide of stoichiometry Na2O·2SiO2 is 54.9%, which means that the Si+B_py composite theoretically contains a carbon loading of approximately 2.9%. Both materials are hence quite thermally stable under inert gas atmospheres.</p><!><p>Figures 7 and 8, respectively, show the SEM images of Si_py, Si_py_al, Si+B_py, and Si+B_py_al, as well as the corresponding magnified images of one single grain in the insets. SEM images of the intrinsic silica composite (Figure 7 A) indicates that the unwashed sample is composed of some large irregular grains with a wide particle size distribution. The magnified image (inset) reveals that the large secondary grain consists of the aggregation of numerous small primary spherical particles (ca. 3.4 μm in mean diameter) and interspersed micro-lamellas, which are tightly held together by the thermally induced sodium silicate binder. After acid leaching, the secondary grains are broken up into smaller irregular pieces as a result of the dissolution of the binder (Figure 7 B). In contrast, the unwashed secondary particles of Si+B_py are more uniform and smaller (Figure 8 A) together with the primary spheres of roughly 1.2 μm in average grain diameter (inset in Figure 8 A). Notably, despite the wide dispersity of particle sizes, a portion of silica nanoparticles are successfully achieved upon experiencing simple acid-leaching treatment (Figure 8 B).</p><!><p>Figures 9 and 10, respectively, show the N2 sorption isotherms of Si_py, Si_py_al, Si+B_py, and Si+B_py_al, as well as their corresponding PSDs (insets). Textural parameters for both sets of amorphous silica composite materials are summarized in Table 2. As seen in both figures, all samples except for Si+B_py exhibit Type IV isotherms with widespread yet distinct hysteresis loops of Type H3 in the IUPAC classification, characteristic of mesoporous materials with relatively broad PSDs that are also reflected by the respective PSDs (insets in Figures 9 and 10). It is found that the sorption isotherms give dramatic increases in SBET from 35.5 for Si_py to 163.3 m2 g−1 for Si_py_al, and from 30.0 for Si+B_py to 254.7 m2 g−1 for Si+B_py_al. The corresponding total pore volumes of these four composites are respectively 0.051, 0.211, 0.046 and 0.302 ml g−1. In both cases, the treatment with 2.2 wt% nitric acid solutions leads to remarkably improved sorption properties in terms of Vt and SBET, which can definitely extract Na+ cations from the pyrolyzed products with an empirical stoichiometry C(0.4-0.46)·Na2O·2SiO2, thereby creating an appreciable volume fraction of voids occupied by constituent Na2O porogen. Upon acid leaching, the microporosity decreases from 13.7% for Si_py to 10% for Si_py_al, whereas that increases from 13% for Si+B_py to 20.5% for Si+B_py_al. The opposite trend can be explained by the inter-particulate void change that occurs over the course of acid treatment. It is found that acid leaching cannot completely alter the agglomerated properties of the particles derived from the raw polycrystalline Si complex without any B mediator (Figure 7 B) in sharp contrast to the resulting better-divided fine powders from the B-assisted single crystals (Figure 8 B). It is worth noting that acid leaching turns out not to notably modify the mean mesopore widths (Table 2). The large surface area of these silica composites is principally an outcome of very small constituent primary spheres (Figures 7 and 8 insets). The coincident carbon deposits even in significantly low yield are expected to serve as an effective inhibitor minimizing the sintering events to occur on the growing adjacent silica particles during pyrolysis so as to partially offset the adverse effect of sodium silicate binder produced simultaneously, thus contributing to the generation of unwashed silica composites of modest surface area. On the other hand, the resultant silica composites with hierarchical micro-/meso-porosity could inherit to some degree a legacy of porosity from the corresponding layered Si complex precursors. Finally, the atomic arrangement featuring the Si centers free of any direct Si-O-Si linkages (Figure 2) is favorable for the production of high-surface-area silica powders.</p><!><p>In this work, we have utilized a modified transesterification mechanism to solvothermally prepare two well-ordered layered [5]Si complexes with similar crystal phase and empirical formula. Additionally, a novel B-assisted crystallization approach without affecting the elemental kind of the host organosilicate matrix is presented here allowing for the facile preparation of X-ray quality single crystals, whereas no any single crystals but polycrystalline powders are directly attained without the aid of B species as an effective crystal growth modifier. The crystal structure of the B-assisted Si complex with mixed methanolate and glycolate ligands is for the first time determined by SCXRD technique, highlighting the major interlayers intercalated with extensive van der Waals gaps even in the presence of substantial Na+ counter-ions due to unique atomic arrangement in its crystal structure. The synthetic strategy proposed here offers a generalized route for controllable synthesis of diverse layered metal-organic compounds, which can optionally serve as valuable precursors towards the flexible and facile production of high-surface-area metal oxide composites. Herein, the fabrication of amorphous carbon-silica composites with an appreciable specific surface area is typically exemplified through rapid pyrolysis under inert atmosphere, followed by acid leaching. As such, distinct from the classic sol-gel technique, the present findings afford a new example on how to rapidly synthesize high-surface-area metal oxides. The synthetic methodology of fast pyrolysis is expected to open up a pathway to effectively prepare such functional metal oxides with a broad compositional diversity.</p><!><p>For general methods of characterization and analysis see the SI.</p><!><p>A distinct transesterification synthetic strategy was created to prepare two layered 5-coordinate organosilicate complexes with different empirical formula. No special care was taken to exclude the exposure to extraneous moisture, and all manipulations were carried out in a well-ventilated fume hood. All chemicals were purchased from Sigma-Aldrich and used as received without further purification.</p><!><p>A recipe of Si(OCH3)4:3EG:NaOCH3:15CH3OH on a molar basis was originally formulated. In the synthesis, 3.20 g of sodium methoxide powders (97%) were completely dissolved in 18.14 g of anhydrous methanol solvent (99.8%) under intense stirring before the addition of 10.81 g of anhydrous ethylene glycol (EG, 99.8%). After thorough homogenization of the aforementioned solution, 8.91 g of tetramethyl orthosilicate (TMOS, 98%) was rapidly poured into it at room temperature (RT) under agitation. Subsequently, the highly viscous gel thus formed was transferred into an autoclave, and the solvothermal synthesis was statically carried out at 140 °C under autogenous pressure for 7.5 d. The solid fraction was recovered via vacuum filtration and repeated rinse with copious amounts of anhydrous methanol. The collected solid was vacuum dried at 80 °C for 9 h yielding 8.5 g of dry white polycrystalline powders and then stored in a desiccator for future characterizations.</p><!><p>To obtain the large single crystals required for crystal structure elucidation, a small fraction of B species was introduced acting as an effective mediator between nucleation and crystal growth. A gel precursor with a molar composition of 0.95Si(OCH3)4:0.05B(OCH3)3:3EG:NaOCH3:15CH3OH was prepared by sequentially mixing calculated amounts of anhydrous methanol, NaOCH3, trimethyl borate (≥ 99%), TMOS and EG under intense stirring. The resulting thin gel was aged under stirring overnight at RT. The solvothermal synthesis was statically carried out at 181 °C for 4 d. The solid fraction was collected via vacuum filtration and multiple rinses with plenty of anhydrous methanol. The recovered solid was vacuum dried at 85 °C for 9 h affording 3.3 g of transparent diffraction-quality single crystals, e.g., 0.35 × 0.20 × 0.15 mm3 in dimension.</p><!><p>Fast pyrolysis operations were conducted on both of as-synthesized Si complexes in a flowing N2 atmosphere (25 ml min−1) in a TGA furnace using a ramping rate of 10 °C min−1 up to the targeted 800 °C and a soaking time of 8-10 min. In addition, part of these pyrolyzed samples were subsequently rinsed overnight at 50 °C under intense stirring with 2.2 wt% nitric acid aqueous solutions in adequate excess to substantially acid leach the thermally formed Na2O component. Afterwards, they were collected via vacuum filtration and complete rinse with copious amounts of deionized (DI) water until the pH of the filtrate was neutral, and finally dried at 120 °C overnight for future PXRD, SEM, and N2 sorption analyses. The resultant samples were denoted as Si_py, Si_py_al, Si+B_py, and Si+B_py_al, where py and al represented pyrolysis and acid leaching treatments, respectively.</p>

PubMed Author Manuscript

The impact of 5-formyltetrahydrofolate on the anti-tumor activity of pralatrexate, as compared to methotrexate, in HeLa Cells in vitro

Purpose To investigate the impact of 5-formytetrahydrofolate on the activities of pralatrexate, as compared to methotrexate (MTX), in vitro. Methods Cells were exposed to (6S)5-formyltetrahydrofolate (5-formylTHF) for 24h, before or after a 6h exposure to antifolates following which the cellular accumulation and activities of the drugs were evaluated in HeLa cells. Results A 24h delay between a 6h exposure to antifolates and a subsequent 24h exposure to 4 \xce\xbcM 5-formylTHF sustained the full activities of both antifolates. A 72h interval was required between a single exposure of up to 4 \xce\xbcM 5-formylTHF and subsequent exposure to drugs to sustain activities of the antifolates. When cells were incubated with 4 \xce\xbcM 5-formylTHF for 24h weekly, for 4 weeks, there was no significant increase in the IC50 for pralatrexate, but the MTX IC50 increased 2.5-fold as compared to cells growing continuously in 25nM 5-formylTHF. This cyclical exposure to 5-formylTHF increased the cell folate pool by 16%, had no significant effect on the intracellular pralatrexate level, but decreased intracellular MTX by 15%. An extracellular concentration of methotrexate 50-fold higher than that of pralatrexate was required to achieve an intracellular level, and growth inhibition, comparable to that of pralatrexate. Conclusions Cyclical exposures to 5-formylTHF at levels in excess of what is achieved in most clinical \xe2\x80\x9crescue\xe2\x80\x9d regimens do not affect pralatrexate accumulation nor antitumor activity in HeLa cells, in contrast to MTX. An important element in preserving pralatrexate activity is achieving a sufficient interval between exposure to 5-formylTHF and the next dose of antifolate.

the_impact_of_5-formyltetrahydrofolate_on_the_anti-tumor_activity_of_pralatrexate,_as_compared_to_me

Introduction<!>Reagents<!>Cell lines<!>Experimental Design and Growth inhibition assays<!>Measurement of total cellular folate and antifolate levels<!>Statistical analysis<!>Analysis of 5-formylTHF \xe2\x80\x9crescue\xe2\x80\x9d of HeLa cells as a function of time between exposure to antifolates and exposure to 5-formylTHF<!>The impact of the 5-formylTHF concentration on \xe2\x80\x9crescue\xe2\x80\x9d<!>An analysis of the \xe2\x80\x9cprotective\xe2\x80\x9d potential of 5-formylTHF on the subsequent exposure to pralatrexate and MTX in HeLa cells<!>Antifolate \xe2\x80\x9cprotection\xe2\x80\x9d as function of the 5-formylTHF concentration<!>Cumulative effect of 5-formylTHF on total cell folate<!>The impact of four weekly exposures to 5-formylTHF on pralatrexate and MTX accumulation<!>Discussion<!><!>The impact of exposure to 5-formylTHF after exposure to antifolates on growth inhibition<!>Impact of the interval between exposure to 5-formylTHF followed by exposure to the antifolates on growth inhibition<!>Impact of a single or four-weekly exposures to 5-formylTHF, 72h prior to exposure to pralatrexate or MTX, on antifolate activities<!>Effect of 4 weekly exposures to 5-formylTHF on total intracellular folate and antifolate accumulation

<p>Pralatrexate, the 10-propargyl-10-deaza analog of aminopterin, is a potent inhibitor of dihydrofolate reductase (DHFR) [1], the enzyme required for the maintenance of tetrahydrofolate (THF) cofactor pools within cells when 5,10-methylene THF is oxidized to dihydrofolate (DHF) during the synthesis of thymidylate mediated by thymidylate synthase [2,3]. Pralatrexate's pharmacological advantage is based upon its superior properties relative to methotrexate (MTX) as a substrate for the reduced folate carrier (RFC) and folylpolyglutamate synthase (FPGS) resulting the rapid formation, and accumulation to high levels, of its polyglutamate derivatives that sustain suppression of its target enzyme for long intervals within cells [1,4]. Unlike its very high affinity for RFC, pralatrexate has a lower affinity for the proton-coupled folate transporter (PCFT) than MTX so that this carrier is unlikely to contribute to the transport of this agent into tumor cells. However, the low affinity for PCFT, the mechanism of intestinal absorption of folates and antifolates, would tent to increase the fecal excretion and clearance of this agent [4]. Pralatrexate is approved for the treatment of relapsed and refractory peripheral T-cell lymphoma (PTCL) and transformed mycosis fungoides (T-MF) [5,6,7] and is currently being evaluated for its efficacy in other malignancies [8,9].</p><p>Pralatrexate is ~10-fold more potent than MTX with continuous exposure, and the difference becomes much greater with a 6h exposure to the antifolates [4]. However, pralatrexate can be administered parenterally at a weekly dose of 30 mg/m2 with folic acid supplementation, comparable to the weekly dosing for MTX [10,5]. Based upon the relative potencies of pralatrexate and MTX, current pralatrexate regimens might be considered comparable to "high-dose" MTX without leucovorin (6(R,S)5-formylTHF) "rescue". These observations indicate that pralatrexate has a greater degree of selectivity than MTX; nonetheless, mucosites is often dose-limiting for the drug [5]. Hence, studies have been initiated and planned to evaluate the impact of "rescue" regimens on the incidence and intensity of the mucosites associated with this agent [11,12]. The objective of the current study is to determine the extent to which 5-formyltetrahydrofolate alters the activity of pralatrexate, as compared to MTX, using the active "S" isomer (5-formylTHF) and an in vitro protocol that simulates elements of how this folate would be employed in the clinical setting.</p><!><p>[3′,5′,7,9-3H](6S)5-formylTHF, [3′,5′,7-3H]MTX and generally labeled [3H]pralatrexate were obtained from Moravek Biochemicals (Brea, CA), The agents were purified as necessary, and purity monitored by liquid chromatography as described previously [13,4]. Pralatrexate was purified using a 5-mm OSD2 4.6 x 250 mm reversed-phase column (Waters Spherisorb) by isocratic elution with 100 mM sodium acetate pH 5.5 (solvent A) and 15% acetonitrile (solvent B). The mobile phase was delivered at 1 ml/min, reaching 100% solvent B in 30 min. Nonlabeled pralatrexate (Folotyn) and (6S)5-formylTHF (Fusilev) were provided by Spectrum Pharmaceuticals (Irvine, CA). MTX was obtained from Sigma-Aldrich (St. Louis, MO).</p><!><p>HeLa cells were maintained in folate-free RPMI 1640 medium supplemented with 10% dialyzed fetal bovine serum (Gemini Bio-Products, CA), 100 units/ml penicillin and 100 μg/ml streptomycin (Gibco Life Technologies, CA) at 37°C in a humidified atmosphere of 5% CO2. (6S)5-formylTHF (25 nM) was the folate source in the medium.</p><!><p>Experiments were designed to simulate exposures to 5-formylTHF as it is administered in clinical regimens. The experiments evaluated the impact of 5-formylTHF administered after pralatrexate as in "rescue" along with its impact on the activity of the "next" dose of pralatrexate as would occur in a weekly regimen with this agent. The studies were performed within the context of a comparison with MTX at concentrations that produced comparable growth inhibition. The details of each type of experiment are described in the Results section and the legends to the figures. HeLa cells were seeded in 96-well plates at a density of 2 × 103 cells/well. At some point, pralatrexate or MTX was added to achieve a spectrum of concentrations. After 6h, the cells were washed then grown in drug-free medium for 3–5 days. The cells were then assayed by sulforhodamine B staining. Absorbance was measured at 540 nm with the VERSAmax plate reader (GE Intelligent Platforms, Charlottesville, VA).</p><!><p>For determination of total cellular folates, HeLa cells were grown for a week in folate-free medium supplemented with GAT (0.2 mM glycine, 0.1 mM adenosine, and 0.01 mM thymidine) to deplete endogenous folates. The cells were then replated in medium supplemented with 25 nM [3H](6S)5-formylTHF. One portion was maintained in this medium, the other portion was exposed to 4 μM [3H](6S)5-formylTHF for 24h weekly. After 4 weeks (four, 7-day, cycles) the cells were washed in ice-cold HBS buffer (20 mM Hepes, 140 mM NaCl, 5 mM KCl, 2 mM MgCl2 and 5 mM dextrose; adjusted with 1 N NaOH to achieve a pH of 7.4) and then digested with 500 μl of 0.2 N NaOH at 65°C for 45 minutes. Twenty μl was used for protein determination by the bicinchoninic acid protein (BCA) assay (Pierce Chemical, IL), 400 μl for assessment of tritium on a liquid scintillation spectrometer.</p><p>For determination of cellular antifolate levels, HeLa cells were grown in medium supplemented with 25 nM (6S)5-formylTHF containing either 0.1 μM [3H]pralatrexate or 5μM [3H]MTX for 6h then washed twice in ice-cold HBS. These concentrations of antifolates approximated the level required to produce comparable growth inhibition (IC50's). One portion of cells was assessed for total tritium and the other portion for total protein as described above. Intracellular radioactivity is expressed as picomoles of tritiated substrate per mg of protein.</p><!><p>The IC50 values were calculated from an analysis of growth inhibition (as percent of control growth) as a function of the log of the extracellular antifolate concentration. In some experiments, the data is plotted as the ratio of the IC50 of the cells exposed to 1 ≥ μM 5-formylTHF to cells grown in 25 nM 5-formylTHF as a function of the extracellular (6S)5-formylTHF concentration. Statistical comparisons were performed by the two-tailed Student's paired t test. All statistical analyses utilized GraphPad Prism (version 6.0 for Windows, GraphPad Software).</p><!><p>On day 1, HeLa cells were incubated for 6h with a spectrum of pralatrexate or MTX concentrations. Twenty-four or 48h later, the cells were incubated with 1 μM 5-formylTHF for 24h following which the cells were grown without either agent but in presence of 25nM 5-formylTHF, representative of the "normal" blood folate level. The IC50 (the antifolate concentration at which growth was 50% that of cells not exposed to drugs) for pralatrexate (Fig. 1A) or MTX (Fig. 1B) was not affected by these exposures to 5-formylTHF. Comparing the two panels, it can be seen that pralatrexate was ~130-fold more potent than MTX under these conditions.</p><!><p>Because a 24h interval was sufficient to sustain the cytotoxicity of MTX and pralatrexate at an extracellular concentration of 1 μM 5-formylTHF, studies evaluated the impact of increasing the 5-formylTHF concentration. On day 1 HeLa cells were incubated for 6h with pralatrexate or MTX, 24h later the cells were incubated with 1 to 16 μM 5-formylTHF for 24h following which the cells were grown for an additional 4 days in 25nM 5-formylTHF. As indicated in Figure 1C, the pralatrexate and MTX IC50 ratios (the ratio of the IC50 of the cells exposed to 1≥ μM 5-formylTHF to cells grown in 25 nM 5-formylTHF) were unchanged at 5-formylTHF concentrations up to 4 μM; the small increase in the pralatrexate IC50 at 4μM was not significant. Comparing the average IC50 of the lowest three with the average of the highest three 5-formylTHF concentrations, there was only a small increase in the pralatrexate IC50 (47±2 vs 71±2 nM, P=0.006) and the MTX IC50 (2.9±0.15 vs 3.8±0.2 μM, P=0.08).</p><!><p>The previous experiments indicated that a single 6h incubation with pralatrexate or MTX followed 24h later by growth in up to 4 μM 5-formylTHF did not alter the activities of these antifolates. Other studies assessed the extent to which exposure to the folate prior to the exposure to the antifolates affects the antitumor activity. HeLa cells were exposed for 24h to 1 μM 5-formylTHF, then replated in medium with 25 nM 5-formylTHF 24h, 48h or 72h before a 6h incubation with pralatrexate or MTX. As indicated in Figure 2A, pralatrexate activity was decreased ~2-fold (IC50 increased from ~28 nM to ~60 nM) when cells were exposed to the drugs 24h or 48h after treatment with 5-formylTHF. MTX activity was comparably decreased (IC50 increased from ~3 μM to ~7 μM) under similar conditions (Fig. 2B). The activities of both drugs were fully restored when exposure to the antifolates was delayed until 72h after exposure to the 5-formylTHF.</p><!><p>To determine the extent to which protection was dependent upon the 5-formylTHF concentration, HeLa cells were exposed to from 1 to 16 μM 5-formylTHF for 24h, then washed and incubated in medium with 25nM 5-formylTHF; 72h later the cells were incubated for 6h with pralatrexate or MTX then grown in absence of antifolate until the end of the assay. Figure 3A indicates that the IC50 ratio for pralatrexate was unchanged up to 8 μM 5-formylTHF, the small increase at 16 μM 5-formylTHF was not significant. Similarly, while there was a trend towards an increase in the IC50 ratio for MTX at the highest 5-formylTHF concentrations, this did not reach statistical significance.</p><p>Next, the cumulative effect of repeated exposures to 5-formylTHF on pralatrexate and MTX activities was assessed. HeLa cells were subjected to 24h incubations with 4μM 5-formylTHF weekly for four weeks. Seventy-two hrs after the last (4th) exposure, the cells were incubated with pralatrexate or MTX for 6h then replated with 25 nM 5-formylTHF for an additional 5 days. Figure 3B indicates that pralatrexate activity was minimally affected by exposures to 5-formylTHF. Figure 3C illustrates the slope of the IC50 ratio as a function of the extracellular concentration of 5-formylTHF. While the slope for pralatrexate is significant (p=0.03, Panel C), there was only a negligible increase in the IC50 ratio at 4 μM 5-formylTHF. The pattern for MTX was different as indicated in Figure 3B and 3C. The IC50 ratio increased as the 5-formylTHF concentration increased reaching a level >2.5-fold greater for cells growing in 4 μM as compared to 25 nM 5-formylTHF (P=0.005, Fig. 3C). The inset to Figure 3C amplifies the slope of the IC50 ratios within the 4 μM range. It can be seen that the slope of the MTX IC50 ratio as a function of the extracellular 5-formylTHF concentration exceeded that of pralatrexate by a factor of 10 (0.4 vs 0.04, P=0.01). Hence, loading cells with 5-formylTHF clearly impacts on the activity of MTX but has a negligible effect on pralatrexate activity particularly at concentrations of the folate to which cells are exposed in clinical rescue regimens (see Discussion).</p><!><p>The impact of the multiple exposures to 5-formylTHF on total cell folate was assessed by first depleting cells of endogenous folates by growth in GAT for one week. The cells were then grown continuously with 25 nM [3H](6S)5-formylTHF except for a 24h exposure to 4 μM [3H](6S)5-formylTHF at the same specific activity, each week for four weeks; this was the concentration at which the difference between MTX and pralatrexate growth inhibition was maximal. The control cells were maintained in 25 nM [3H](6S)5-formylTHF. Figure 4,A indicates that there was a small (16%) but significant increase in the total folate pool (58.8 ± 1.0 vs 68.3±2.7 pmol/mg of protein respectively, P=0.02) after the fourth cycle.</p><!><p>To evaluate the impact of four cycles of exposure to 5-formylTHF, as described above, on the accumulation of the antifolates, 72h after the 4th exposure to 5-formylTHF, cells were exposed to 0.1[3H]pralatrexate or 5 μM [3H]MTX for 6h following which the cells were analyzed for their intracellular antifolate content. As indicated in Figure 4B, the small decrease in the accumulation of [3H]pralatrexate (16.84±0.1 vs 15.16±1.3 pmol/mg of proteins) was not significant (p=0.4). The 15% decrease in accumulation of [3H]MTX (16.95±0.79 vs 14.44±0.64) was highly significant (p=0.004) in cells exposed to multiple cycle of 5-formylTHF as compared with the cells maintained in 25 nM 5-formylTHF. Under these conditions, >80% of intracellular antifolates were the polyglutamate forms. It is noteworthy that a 50-fold higher concentration of [3H]MTX (5 μM) was required to achieve a total intracellular antifolate level comparable to what was achieved with 0.1 μM [3H]pralatrexate over the 6h incubation, consistent with the more efficient transport and polyglutamation of the latter antifolate. These concentrations also reflected the differences in IC50's for these agents.</p><!><p>Pralatrexate has a much higher therapeutic index than MTX. However, the degree of toxicity is far less than might be expected from an agent that is so potent relative to MTX but can be administered weekly intravenously at a dose only slightly less than that of MTX [6]. While sharing the same target, DHFR, pralatrexate's 8–10-fold higher affinity for RFC, and 10-fold greater catalytic activity mediated by FPGS, relative to MTX, results in the formation and accumulation of high levels of its active polyglutamate derivatives that are retained in tumor cells and produce sustained inhibition of its target enzyme [1,4]. This is illustrated by the observation that a 6h incubation with MTX required a ~50-fold higher extracellular concentration than pralatrexate to achieve a comparable intracellular level and growth inhibition. However, the mucosites associated with pralatrexate can be dose-limiting and an impediment to the use of this agent. This has raised the possibility that a leucovorin "rescue" regimen might improve the clinical utility of this agent [12]. Some patients have been treated with a "rescue" regimen (a single 50 mg dose 24h after pralatrexate) with amelioration of toxicity and, while anecdotal, antitumor activity was preserved at least in part [12].</p><p>Previous studies from this laboratory described the adverse impact of 5-formylTHF on the activities of a variety of antifolates [14,15]. The impact on pemetrexed activity was shown to be due to inhibition of the formation of polyglutamate derivatives of this agent required for inhibition of its target enzymes [14]. Recently, a similar phenomenon was observed for new generation GARFTase inhibitors that also have a high affinity for FPGS [16]. When the exposure to MTX is brief, its activity is very sensitive to the cellular folate pool. This reflects the much lower affinity of MTX for FPGS relative to pralatrexate and the much longer time required to accumulate its polyglutamate derivatives in cells [17,18,14]. It was important to understand the effect that 5-formyTHF might have on the activity of pralatrexate in order to minimize, as much as possible, a negative impact on its efficacy.</p><p>The natural (S) isomer of 5-formylTHF was utilized in the current study. The unnatural (D) isomer has essentially no biological activity; it is a very poor substrate for RFC [19] and what does enter cells cannot be metabolized. Hence, the concentrations of (6S)5-formylTHF employed in this study are chemically equivalent to half the concentration of the racemic mixture but equal to the biologically active fraction. The exposures to 5-formylTHF encompassed and far exceeded the concentration of this folate, and 5-methyltetrahydrofolate to which it is rapidly converted, achieved in "rescue" regimens in vivo. For instance, following a 50 mg intravenous dose of leucovorin, the same dose as in the case reports with pralatrexate described above [12], the peak (6S)5-formylTHF blood level was ~4.4 μM with a t½ of 32 min, the peak 5-methylTHF level was ~1.6 μM with a t½ of 224 min so that at 6h the former was ~0.02 μM and the latter was ~0.8 μM; by 12h the latter was ~0.2 μM [20]. When leucovorin was administered p.o. at 25 mg every 8h, the mean 5-methylTHF blood level was only ~0.7 μM at 48h [20]. Likewise, after an intravenous dose of 10 or 25 mg/m2 leucovorin, the peak 5-methylTHF blood levels were 0.23 and 0.6 μM, respectively, with an average t½ of 6.6h. With p.o. dosing, the peak 5-methylTHF levels were 0.32, 0.47 μM, respectively with an average t½ of 3.1h [21]. At these 5-methylTHF blood levels, there would be essentially no impact on pralatrexate activity.</p><p>A 24h interval between exposure to the antifolates and 5-formylTHF in the current study was sufficient to sustain full growth inhibition by MTX and pralatrexate to a 5-formylTHF concentration of 4μM with only a small (50%) increase in IC50 at 16 μM 5-formylTHF. On the other hand, when the impact of 5-formylTHF on the "next" dose of antifolate was considered, a 72h delay after exposure to 5-formylTHF was necessary to preserve antifolate activities. Even when the exposure to 5-formylTHF was repeated weekly for four weeks, there was no increase in the pralatrexate IC50 ratio at blood levels that exceeded those achieved in "rescue" regimens; however, the MTX IC50 increased to 2.5-fold as the 5-formylTHF concentration was increased to 4 μM. Hence, it would appear that a critical element in preserving antifolate activity is the delay between 5-formylTHF and the next dose of antifolate. Indeed, in a weekly regimen the delay would be 4 days if the "rescue" interval was restricted to 24h.</p><p>The rationale for high-dose MTX with leucovorin "rescue", and its putative selectivity, is based upon elements at the tumor, cellular, and biochemical levels [22,3]. High concentrations of drug facilitate passive diffusion into the poorly perfused interstitium of solid tumors. The subsequent provision of much lower doses of leucovorin poorly penetrate tumors but are readily accessible to precursor cells of the intestine and bone marrow with their intact vascular system. Likewise, high concentrations of MTX would passively diffuse into tumor cells with impaired transport due to low expression or loss-of-function mutations of RFC. The low concentration of leucovorin during "rescue" which enters cells via the same transporter would have limited access to the tumor cells but would readily enter normal tissues with intact RFC. At the biochemical level, MTX polyglutamate derivatives are potent direct inhibitors of thymidylate synthase and AICAR transformylase [23,24,25]. This interferes with the utilization of 5,10-methyleneTHF and 10-formylTHF, respectively, formed by the interconversion of leucovorin [26]. However, utilization of these one-carbon donors in bone marrow and intestine is less impeded since only low levels of MTX polyglutamate derivatives accumulate in these cells [27,28,29,30]. Finally, the interconversion of leucovorin to dihydrofolate results in the displacement of the monoglutamate of MTX from DHFR reactivating tetrahydrofolate synthesis. However, this does not occur when high levels of MTX polyglutamates accumulate in tumor cells [27,31,32,26]. There is no information, as yet, as to the extent to which pralatrexate polyglutamates are direct inhibitors of tetrahydrofolate cofactor-requiring enzymes in cells.</p><p>There are preclinical in vivo data that support the concept that leucovorin "rescue" diminishes the toxicity, while preserving the activity, of MTX and aminopterin [33,34]. Leucovorin "rescue" clearly decreases the toxicity of high-dose MTX regimens; however, the extent to which the antitumor activity of MTX is diminished is uncertain. One clinical study compared standard-dose MTX (40 mg/m2 weekly for 8 weeks) with or without low-dose leucovorin "rescue" (10 mg/m2 p.o. every six hours, starting 24h after MTX, for four doses) in patients with squamous cell head and neck cancer. Leucovorin decreased drug toxicity but also diminished antitumor activity (response rate, 17.1 and 36.7%, respectively, P=0.047) [35]. Data in the current report indicate that there is a greater adverse impact of loading cells with folates on MTX than pralatrexate anti-tumor activities. Because of the high therapeutic index of pralatrexate relative to MTX, it may be that very modest doses of 5-formylTHF administered over short intervals will obviate the mucosites without an adverse impact on pralatrexate's therapeutic efficacy, as suggested by the initial report described above [12].</p><p>Finally, the observations in this report are focused on one cell line. It is possible that the critical intervals between exposures to antifolates and 5-formylTHF, along with the magnitude of the suppressive effects of 5-formylTHF, may vary among cells of different origin. However, irrespective of these considerations, these studies indicate that the impact of "rescue" with intermittent exposures to 5-formylTHF will be less for pralatrexate than for MTX. Since, as indicated above, there is evidence that leucovorin "rescue" is associated with decreased efficacy of MTX, this represents an important advantage for this next-generation DHFR inhibitor.</p><!><p> Conflict of Interest </p><p>This study was supported by a grant from Spectrum Pharmaceuticals, Inc (Irvine, CA).</p><p>(6S)5-formyltetrahydrofolate</p><p>d,l,5-formyltetrahydrofolate</p><p>phosphoribosylaminoimidazolecarboxamide formyltransferase</p><p>dihydrofolate reductase</p><p>folylpolyglutamate synthetase</p><p>glycinamide ribonucleotide formyltransferase</p><p>thymidylate synthase</p><p>reduced folate carrier</p><p>methotrexate</p><!><p>HeLa cells were exposed to pralatrexate (Panel A) or MTX (Panel B) for 6h at the indicated concentrations. After 24h (open square) or 48h (open triangle) the cells were incubated for 24h with 1 μM 5-formylTHF then grown in 25 nM 5-formylTHF for additional 4 and 3 days, respectively. Control cells (filled circle) were maintained continuously in 25 nM 5-formylTHF. Growth in the absence of drug is indicated as 100%. The vertical line intercepts the x axis at the concentration at which growth inhibition is 50% of the level of growth in the absence of drug (IC50). Data are the mean ± S.E.M from three independent experiments. Panel C: HeLa cells were incubated with a spectrum pralatrexate or MTX concentrations for 6h. After 24h, the cells were exposed to 5-formylTHF at the indicated concentrations for 24h and then grown in 25 nM 5-formylTHF for an additional 4 days. Control cells were maintained in 25 nM 5-formylTHF. The Y axis is the ratio of the IC50 of the cells exposed to 1 ≥ μM 5-formylTHF to cells grown in 25 nM 5-formylTHF. Cells were always washed after each manipulation. Data are the mean ± S.E.M from three independent experiments.</p><!><p>HeLa cells were exposed to 1 μM 5-formylTHF for 24 (open square), 48 (open triangle) or 72h (open circle) before exposure to pralatrexate or MTX for 6h at the indicated concentrations. The cells were then grown for additional 5 days in 25 nM 5-formylTHF. Control cells (filled circle) were maintained continuously in 25 nM 5-formylTHF. Growth in the absence of drug is indicated as 100%. The vertical line intercepts the x axis at the concentration at which growth inhibition is 50% of the level of growth in the absence of drug (IC50). Cells were always washed after each manipulation. Data are the mean ± S.E.M from three independent experiments.</p><!><p>HeLa cells were exposed once (Panel A), or weekly X 4 (Panel B), for 24h to 5-formylTHF at the indicated concentrations. Seventy-two hours after the last exposure, the cells were exposed to a range of pralatrexate or MTX concentrations for 6h and then grown for additional 5 days in 25 nM 5-formylTHF. Control cells were maintained continuously in 25 nM 5-formylTHF. Panel C is a representation that illustrates the increase in IC50 ratio as a function of the extracellular concentration of 5-formylTHF. The inset amplifies this relationship over the 1–4 μM 5-formylTHF range. The IC50 ratio is described in the legend to Figure 2. Cells were always washed after each manipulation. Data are the mean ± S.E.M from three independent experiments.</p><!><p>(Panel A) HeLa cells were grown for one week in GAT to deplete endogenous folates and then replated for 4 weeks in 25 nM [3H]5-formylTHF. Once a week for four weeks a portion of cells was treated for 24h with 4 μM [3H]5-formylTHF, another portion was maintained in 25 nM [3H]5-formylTHF. Seventy-two hours after the last exposure to [3H]5-formylTHF, total intracellular radioactivity was measured. (Panel B) HeLa cells were grown in 25 nM 5-formylTHF; once each week the cells were exposed for 24h to 4 μM 5-formylTHF. Control cell were maintained in 25 nM 5-formylTHF. Seventy-two hours after the last exposure to 5-formylTHF, the cells were exposed to 0.1 μM [3H]pralatrexate or 5 μM [3H]MTX for 6h and intracellular radioactivity measured. Cells were always washed after each manipulation. Data are the mean ± S.E.M from three independent experiments.</p>

PubMed Author Manuscript

Fabrication of potato-like silver molybdate microstructures for photocatalytic degradation of chronic toxicity ciprofloxacin and highly selective electrochemical detection of H2O2

In the present work, potato-like silver molybdate (Ag 2 MoO 4 ) microstructures were synthesized through a simple hydrothermal method. The microstructures of Ag 2 MoO 4 were characterized by various analytical and spectroscopic techniques such as XRD, FTIR, Raman, SEM, EDX and XPS. Interestingly, the as-prepared Ag 2 MoO 4 showed excellent photocatalytic and electrocatalytic activity for the degradation of ciprofloxacin (CIP) and electrochemical detection of hydrogen peroxide (H 2 O 2 ), respectively. The ultraviolet-visible (UV-Vis) spectroscopy results revealed that the potato-like Ag 2 MoO 4 microstructures could offer a high photocatalytic activity towards the degradation of CIP under UVlight illumination, leads to rapid degradation within 40 min with a degradation rate of above 98%. In addition, the cyclic voltammetry (CV) and amperometry studies were realized that the electrochemical performance of Ag 2 MoO 4 modified electrode toward H 2 O 2 detection. Our H 2 O 2 sensor shows a wide linear range and lower detection limit of 0.04-240 μM and 0.03 μM, respectively. The Ag 2 MoO 4 modified electrode exhibits a high selectivity towards the detection of H 2 O 2 in the presence of different biological interferences. These results suggested that the development of potato-like Ag 2 MoO 4 microstructure could be an efficient photocatalyst as well as electrocatalyst in the potential application of environmental, biomedical and pharmaceutical samples.Nowadays, antibiotics are main class of antimicrobial drugs in today's medicine and widely used to prevent the bacterial infection for both human and animals. In particular, ciprofloxacin (CIP), as an antibiotic drug, plays an important role for the treatment of urinary, digestive infections and pulmonary diseases 1 . The CIP can be entered into the aquatic environment through the intentional disposal of surplus drugs, the release of excreta from human and animals, malapropos treatment in the hospitals or in pharmaceutical industries, improper removal of waste water treatment plants and the use of animal's feces as agricultural fertilizers 2,3 . Probably, the CIP drug is not completely metabolized and the continuous release of CIP into the environments displays the chronic toxicity to bacteria, which causes toxicity to the microorganism and retarding to aquatic vertebrates 4,5 . It also interacts with photosynthesis process and inhibits the growth of spinach plants 6 and cause antibiotic-resistance bacteria growth in the environment 7 . Therefore, the removals of CIP from the water sources are major concern to protect the aquatic system and soil environment. Several methods have been reported for the removal of CIP from water including adsorption 8 , oxidation 9 , sonolysis 10 , sorption 11 , (photo)-Fenton process 12 , photocatalytic degradation [13][14][15] and ozonation 16 . Among them, photocatalysis is an efficient, cost-effective and eco-friendly method

fabrication_of_potato-like_silver_molybdate_microstructures_for_photocatalytic_degradation_of_chroni

<!>Results and Discussion<!>54<!>Amperometric determination of H 2 O 2 at Ag 2 MoO 4 modified GCE. The amperometric i-t technique<!>Experimental Section

<p>for the environmental remediation, which degrades the hazardous organic pollutants into easily bio-degradable or nontoxic molecules 17,18 .</p><p>On the other hand, hydrogen peroxide (H 2 O 2 ) is an essential intermediate in the various food manufacturing and also involved in our life process. The detection of H 2 O 2 is a paramount issue because it is a major reactive oxygen species, generated by most oxidases in mitochondria and related to the several bodily disorders such as myocardial infarction, atherosclerosis, and Alzheimer's disease, cancer, etc. 19 . H 2 O 2 is also acts contrarily in cell growth, differentiation, physiological signaling pathways, migration and immune function system 20 . Moreover, it is most significant chemical in clinical, pharmaceutical industries and atomic power stations, which dramatically have an effect on the cloud and rainwater 21 as well as a precursor to the formation of more reactive and potentially harmful hydroxyl radicals 22 . Therefore, the trace level detection of H 2 O 2 is an important concern for the environment and medicinal fields. Up to now, various methods are available to detect the H 2 O 2 and some of them have been explored a satisfactory results to the concern of H 2 O 2 detection 23 . However, the electrochemical methods hold more advantages such as simplicity, portability, rapid analysis, high sensitivity and low-cost instrumentation. Due to these attractive features, the electrochemical methods are represent a promising alternative technique for the determination of H 2 O 2 .</p><p>Transition metal-based molybdates (M = Fe, Ni, Co, Ag, Mn etc.,) are considered as an important inorganic material which are widely explored in different applications such as Li-ion storage batteries 24 , birefringent filters 25 , supercapacitors 26,27 , optical fibers 28 , photoluminescence 29 , scintillation crystal 30 , photocatalyst 31,32 , humidity sensors 33 , magnetic properties 34 and catalysts 35 . However, low dimensional metal molybdates have attracted more curiosity in current years. In particular, silver molybdate (Ag 2 MoO 4 ) has attracted considerable attention because of its unique properties such as environmental friendly, photoluminescence, high electrical conductivity, excellent antimicrobial activity, good photocatalytic activity and remarkable electrochemical energy storage performance 36 . Due to these properties, the Ag 2 MoO 4 is potentially used in several applications including ion-conducting glasses 37 , high-temperature lubrication 38 , gas sensor 39 , antibacterial material 40 , photoswitches 41 and ceramics 42 . In photocatalysis, Ag 2 MoO 4 has paid significant attention owing to it's photosensitivity which make this material with high photocatalytic activity under UV or visible-light irradiation. Recently, few kinds of literature are reported based on Ag 2 MoO 4 and its composite that act as a photocatalyst for the degradation of organic dyes into the wastewater [43][44][45] . The photocatalytic activity mainly depends on the crystal and electronic structures of materials that affect the energy band structure and the efficiency of charge carrier transfer. Moreover, to improve their physicochemical properties of the photocatalyst, researchers have growled a number of approaches to obtain the different morphologies of Ag 2 MoO 4 including nanoparticles, nanorods, nanowires, wire-like nanostructures, nanoclusters, broom-like, flower-like microstructures and microcrystals 36,41,[46][47][48][49][50] . However, to the best of our knowledge, we reported the synthesis of potato-like Ag 2 MoO 4 microstructure, its applications for the photocatalytic degradation of CIP and electrochemical detection of H 2 O 2 for the first time.</p><p>In this present work, we developed a simple one-pot hydrothermal synthesis of potato-like Ag 2 MoO 4 microparticles with the assistance of urea and characterized using various analytical and spectroscopic techniques in detailed and further evaluated for electrochemical sensing and photocatalytic applications, as illustrated Fig. 1. Fascinatingly, we find that the as-prepared potato-like Ag 2 MoO 4 microparticles exhibited a high-performance electrochemical sensor for the detection of H 2 O 2 . Moreover, their photocatalytic activity towards the removal of CIP antibiotic into the environment was also investigated with efficient degradation rate.</p><!><p>The crystalline structure and phase purity of the as-synthesized samples were determined by using XRD pattern as shown in Fig. 2A. The distinctive diffraction peaks obtained at 16.48°, 27.09°, 31.87°, 33.31°, 38.65°, 42.27°, 47.84°, 50.93°, 55.81°, 63.10°, 65.70°, 66.57°, 76.49° and 78.90° in the 2θ range which well agreed to the (111), (220), (311), ( 222), (400), (331), (422), (511), (440), (620), (533), (622), (642) and (731) reflection planes, respectively. Aforementioned planes are well related to the standard XRD report of cubic phase structured Ag 2 MoO 4 with the space group of Fd3m 51 . From the XRD pattern, it was clearly revealed that the as-synthesized product is β -Ag 2 MoO 4 36 . The appearance of sharp and high intense peaks demonstrated the higher crystalline nature of cubic Ag 2 MoO 4 phase. There is no any other peaks were appeared which related to the Ag 2 O or MoO 3 phase, indicates the as-synthesized Ag 2 MoO 4 was homogeneous solid.</p><p>FTIR and Raman spectroscopy is an important tool for analyzing the involvement of functional groups present in the as-synthesized Ag 2 MoO 4 . In the FTIR spectra (Fig. 2B), the absorption peaks at 3280 and 1650 cm −1 correspond to the O-H stretching and bending vibrations of the water molecules, respectively 52 . The peak at 645 cm −1 is confirmed the Ag-O stretching vibration of Ag 2 MoO 4. The sharp peak at 891 cm −1 attributed to the anti-symmetric Mo-O stretching in tetrahedral MoO 4 2− ion 53 . Raman spectra showed (Fig. 2C) the peaks at 896, 782, 382 and 305 cm −1 which are due to the ν 1 (A g ), ν 3 (E g ), ν 4 (B g ) and ν 2 (A g ) symmetric and asymmetric stretching vibration modes of Ag 2 MoO 4 , respectively. The vibration modes at 896 and 782 cm −1 could be ascribed to the symmetric stretching vibration of Mo-O bond of the MoO 4 unit and the asymmetric stretching vibrations of the molybdate ion, respectively. The peaks at 382 and 305 cm −1 were clearly indicated the ν 4 and ν 2 bending vibration modes of tetrahedral MoO 4</p><!><p>.</p><p>The surface morphology of the as-prepared microstructures was investigated by SEM. Figure 3(A-C) illustrated the general views of the different magnifications of the obtained Ag 2 MoO 4 microstructures. The images are clearly displayed the potato-like structure of Ag 2 MoO 4 which seems like bunches of potatoes with clean and fairly smooth surfaces, the average diameters of microstructures about 1-2 μ m. Energy dispersive x-ray spectra (EDX) were used to identify the elements present in the as-prepared Ag 2 MoO 4 microparticles, as depicted in Fig. 3D. The EDX spectra showed the peaks at approximately 0.5, 2.4 and 3 keV reveal the presence of O, Mo, and Ag elements in the material without any other significant impurities. The bandgap energy of the Ag 2 MoO 4 is an important parameter for the selection of suitable kind of light source needed for the degradation purposes. The UV-Vis (Diffuse reflectance) absorption spectrum of Ag 2 MoO 4 microparticles is shown in Fig. 4A. The results shows that the relation between the normalized absorption of the photocatalyst and wavelength with a range of 200-800 nm. The most part of absorption spectra of Ag 2 MoO 4 falls in the UV region, a broad steep from 300 to 420 nm which corresponds to the band gap energy value from 3-3.34 eV. The band gap value was calculated using Tauc's equation and the graph plotted (α hν ) 2 against (hν) as can be seen in Fig. 4B. The calculated band gap energy value is 3.14 eV. On the other hand, the bandgap of Ag 2 MoO 4 is significantly altered compared to that of previous reports 36,43,44,50 . The oxygen vacancy created in the crystal lattice of the Ag 2 MoO 4 is leads to the distortion in the energy levels and influenced the bandgap which may be attributed to the effect of hydrothermal environment on the surface microstructures.</p><p>X-ray photoelectron spectroscopy (XPS) was used to evaluate the information about the chemical composition and chemical status of the as-synthesized Ag 2 MoO 4 , as shown in Fig. 5. The overall XPS spectrum in Fig. 5A shows the coexistence of elements Mo, C, Ag and O within the as-prepared Ag 2 MoO 4 microparticles and no other impurities were detected, which are in good agreement with EDX report. In addition, the presence of C peak at 284.9 eV is ascribed to the adventitious hydrocarbon from the XPS instrument and it is inherent. High resolution scanning XPS spectra clearly confirms the Ag 3d, Mo 3d, and O 1s level, which is fitted by using the Gaussian fitting method, as shown in Fig. 5(B-D). In Fig. 5B, the Ag 3d spectra displays the two peaks at 368.7 and 374.4 eV attributed to the Ag 3d 5/2 and Ag 3d 3/2 electron binding energy in Ag 2 MoO 4 , respectively 55 . The peaks at 233.2 and 236.3 eV ascribed to the Mo 3d 5/2 and Mo 3d 3/2 binding energy of Mo 3d, respectively. The major binding energy peaks Mo 3d 5/2 and Mo 3d 3/2 are separated by 3.1 eV, which belongs to the Mo 6+ oxidation state as depicted in Fig. 5C 56 . The high intense peaks at around in the range of 530.5-533.6 eV revealed the presence of O 1s core level 57 in Ag 2 MoO 4 (Fig. 5D). Hence, the obtained XPS results clearly confirmed that the valence of Ag, Mo and O are + 1, + 6 and − 2, respectively, which is very good agreement with the phase structure of Ag 2 MoO 4 .</p><p>Electrochemical impedance spectroscopy (EIS) was used to investigate the changes of the electrode surface during the fabrication process. The nyquist curves of the EIS spectra was observed using bare GCE (a) and Ag 2 MoO 4 modified GCE (b) in 0.1 M KCl containing 5.0 mM K 3 Fe(CN) 6 /K 4 Fe(CN) 6 (Fig. 6). The diameter of Photocatalytic activity. The photocatalytic behavior of as-prepared Ag 2 MoO 4 microparticles was performed against the degradation of CIP under UV-light illumination, as illustrated in Fig. 7A. The absorbance spectrum shows the progressive degradation of CIP and the main absorption peak of CIP was observed at 276 nm and other small peaks were also completely diminished within 40 min. The degradation percentage of CIP solution was estimated from the relative intensity of absorbance in UV-visible spectra. The relative intensity of absorbance was decreased and reached almost zero within 40 min, reveals that the Ag 2 MoO 4 microparticles degraded the 98% of the CIP solution. Initially, the utmost decrement of CIP was observed which could be attributed to the competence of CIP with hydroxyl radicals generated by UV-light photoexcitation of Ag 2 MoO 4 microparticles. However, as the reaction proceeds, the formation of by-products from degradation might compete with the hydroxyl radicals and adsorption sites on the catalyst surface. Hence, the gradual degradation of CIP was observed.</p><p>Under similar degradation conditions, commercially available Ag 2 MoO 4 and commercial TiO 2 powder were also employed as a photocatalyst and their accurate comparison was carried out which depicted in Fig. 7B. The photocatalytic activity of commercial Ag 2 MoO 4 and TiO 2 on CIP degradation exhibited a poor efficiency and the corresponding degradation percentages are observed around 38% and 32%, respectively. Moreover, there is no significant degradation was observed either in the absence of light or in the absence of a catalyst. The results clearly confirmed the as-synthesized Ag 2 MoO 4 shown a superior photocatalytic efficacy over the commercial Ag 2 MoO 4 and TiO 2 powder for the degradation of CIP solution.</p><p>Catalyst dosage is an important parameter that can significantly influence the rate of photodegradation. Figure 7C shows the efficiency of CIP degradation (%) against the effect of catalyst loading by varying the catalyst amount from 30 to 80 mg/mL, wherein the concentration of CIP and intensity of the light were maintained as constant. It can be seen that the rate of photocatalytic degradation was maximum at 50 mg/mL, due to the generation of more number of electron-hole pairs. However, the addition of excess amount of photocatalyst can blocks the existing active sites and interfere with the diffusion of photons. As a result, the photocatalytic degradation of CIP was decreased when increase the concentration of photocatalyst over 50 mg/mL 58 .</p><p>Effect of initial concentration of CIP solution on the photodegradation was also investigated and the concentration varied from 10 to 30 mg/L under identical conditions, the results are shown in Fig. 7D. In the present case ~98% degradation was achieved at 20 mg/L whereas 83% and 77% degradation were observed at 25 and 30 mg/L concentration of CIP, respectively. This is due to the screening of light by the CIP solution and the less number of photons to reach the Ag 2 MoO 4 surface. Hence, the electron-hole pair generation is reduced greatly and consequently, the degradation efficiency decreased.</p><p>Generally, the reactive oxidative species (ROS) viz hydroxyl radical (•OH), superoxide radical anion (O 2 •− ), hole (h + ) and electron (e − ) involved in the photocatalytic reaction 59 . The photocatalytic mechanism of the degradation of CIP is represented in the following equations</p><p>The semiconductor photocatalyst generally undergoes excitation under light illumination with energy greater than the bandgap while the e − excited from the valence band (VB) to the conduction band (CB) leaves h + in the VB 60 . In the present study, Ag 2 MoO 4 was irradiated by UV light which produced the e − in the CB and h + in the VB, as illustrated in the Eq. 1. The e − in the CB were reacted with the oxygen molecule to form O 2 •− and the O 2</p><p>•− was reacted with the water molecule to form •HO 2 , as shown in Eqs 2 and 4. Furthermore, h + in the VB adsorbed water molecules and reacted to form •OH, as given in Eq. 2. The ROS formed in the photocatalytic reaction facilitated the degradation of the CIP by the stepwise photocatalytic reduction process (Eqs 5 and 6). The high photocatalytic activity as well as recycling ability of the catalyst is a vital issue for long-term use in practical applications. To evaluate the sustainability of Ag 2 MoO 4 , the recycling experiments were carried out for the degradation of CIP, as illustrated in Fig. 7E. The photocatalytic activity of the Ag 2 MoO 4 on CIP degradation did not show a significant loss and attains more than 90% of degradation rate even after the five successive cycles with every cycle lasting for 40 min, which indicating that the potato-like Ag 2 MoO 4 microstructure is a very effective and highly stable photocatalyst. Thus, the obtained results also indicated a good reusability of the catalyst. Moreover, the poor loss of catalytic activity during the recycle performance is due to the accumulation of impurities on the surfaces of the photocatalyst. S1. When the pH value increased from 3 to 5, the reduction peak current increased and then decreased gradually while increasing the pH more than 7. However, higher peak currents are observed at pH 7. Therefore, Ag 2 MoO 4 modified GCE has good electrocatalytic activity at pH 7 and the reduction of H 2 O 2 is pH dependence. Therefore, we chosen pH 7 is optimized pH and this pH was used further electrochemical measurements.</p><!><p>is one of the most important method to determine the electrocatalytic activity of modified electrodes in electrochemical sensor and biosensor applications. In the present work, we have utilized an amperometric method to estimate the performance of Ag 2 MoO 4 modified GCE toward H 2 O 2 detection. In this regards, Ag 2 MoO 4 modified rotating disc glassy carbon electrode (RDGCE) was performed in continuously stirred pH 7 solution by applying constant potential at − 0.5 V with rotation speed at 1200 rpm. Figure 10 reveals the amperometric i-t performance obtained at Ag 2 MoO 4 modified RDGCE upon the different addition of H 2 O 2 (0.049 to 247 μ M) in the PBS solution. These results undoubtedly shows that Ag 2 MoO 4 modified film demonstrates a fast and well-defined response obtained in each different addition of H 2 O 2 concentration. The response time of H 2 O 2 detection on Ag 2 MoO 4 modified RDGCE was observed for 5 s, it's clearly suggesting that the fast electron movement process was occurred in the electrolyte and electrode interface when introducing the H 2 O 2 . The linear response current increases with increasing the concentration of H 2 O 2 (low concentration to high concentration) from 0.049 to 240 μ M (linear range inset; Fig. 10), the obtained sensitivity and limit of detection (LOD) of the sensor is around 9.8 μ Aμ M −1 cm −2 and 0.03 μ M, respectively. The above results suggesting that the Ag 2 MoO 4 modified RDGC electrode has good electrocatalytic activity towards H 2 O 2 . The analytical parameters, such as LOD, linear range, and sensitivity, of H 2 O 2 sensor was compared with various modified electrodes are summarized in Table 1. Clearly, the Ag 2 MoO 4 modified RDGCE, reported here, exhibits good sensitivity and LOD over a wide linear range of H 2 O 2 concentration compared to other reports [62][63][64][65][66][67][68][69][70][71][72][73][74][75] .</p><p>Selectivity is very important study in the electrochemical sensor. In order to investigate selectivity, the proposed sensor was used to detect the H 2 O 2 in the presence of variety of biological interferences such as catechol (b), fructose (c), lactose (d), sucrose (e), glucose (f), hydroquinone (g), ascorbic acid (h), uric acid (i), dopamine (j) and epinephrine (k) with 50-fold excess concentration of each analytes, as shown in Fig. 11. The Ag 2 MoO 4 modified RDGCE exhibited well-defined response towards each 100 μ M H 2 O 2 (a). No remarkable responses were monitored for the 50-fold excess concentration of biological interferences as mentioned above. Hence, the Ag 2 MoO 4 modified RDGCE has excellent selectivity towards the determination of H 2 O 2. Furthermore, we have studied the stability of our proposed sensor by amperometric (i-t) techniques and the results obtained as it can be seen in Figure S2. The prepared sensor retains its 98.2% of its initial response of H 2 O 2 after prolongs runs up to 2600 s, which suggesting the good stability of the sensor.</p><p>In summary, a potato-like Ag 2 MoO 4 microparticles were successfully prepared through a simple hydrothermal method. Different analytical and spectroscopic methods were used to confirm the structural nature of Ag 2 MoO 4 microparticles. All the obtained results are strongly evidenced that the prepared compound shows like pristine Ag 2 MoO 4 without any other impurities. The as-prepared Ag 2 MoO 4 microparticles explored excellent photocatalytic activity towards the degradation of chronic toxicity CIP under UV-light illumination and 98% of the CIP was degraded within 40 min. On the other hand, the Ag 2 MoO 4 microparticles were used to fabricate the sensor electrode for the detection of H 2 O 2 in the potentially interfering biological substances. The Ag 2 MoO 4 modified GCE revealed a high electrocatalytic activity with wide linear range, good stability and low detection limit than the previous reports. Finally, the Ag 2 MoO 4 represents an interesting and promising candidate for photocatalytic</p><!><p>Materials. Silver nitrate (AgNO 3 ), sodium molybdate dihydrate (Na 2 MoO 4 .2H 2 O), urea (CH 4 N 2 O), hydrogen peroxide (H 2 O 2 ) and all chemicals were purchased from Sigma-Aldrich and used further purification. Ciprofloxacin drug was purchased from Thiruvengadam Medicals, Virudhunagar, India. All other chemicals were of analytical grade and used without further purification. The phosphate buffer solution (PBS) was prepared using 0.05 M Na 2 HPO 4 and NaH 2 PO 4 and all the required solutions were prepared using deionized water (DI).</p><p>Synthesis of silver molybdate. In a typical synthesis, 0.5 M of Na 2 MoO 4 .2H 2 O and 0.1 M of AgNO 3 were dissolved in 60 mL DI water. Then, 0.3 g of urea (10 mL) was gradually added into the above mixture under vigorous stirring at room temperature for 30 min. The mixture was transferred into 100 mL Teflon-lined autoclave and maintained at 120 °C for 8 h in an oven. The autoclave was then cooled down to room temperature naturally, and the obtained yellow product was collected by centrifugation and washed thoroughly with DI water and ethanol for three times and dried overnight at 80 °C. Characterization. The powder X-ray diffraction (XRD) analysis was carried out using PANalytical X'Pert PRO X-ray diffractometer measured with Cu-Kα radiation ( λ = 1.54178 Å) in the 2θ range of 10-100°. X-ray photoelectron spectroscopy (XPS) measurements were carried out using ULVAC-PHI 5000 VersaProb instrument. Scanning electron microscope (SEM) and Energy dispersive X-ray spectra (EDX) were probed using Hitachi S-3000 H and HORIBA EMAX X-ACT, respectively. Raman spectra were collected in an NT-MDT confocal Raman microscopic system with an exciting laser wavelength of 488 nm and the laser spot-size is around 0.5 μ m. Fourier transform infrared spectroscopy (FTIR) was recorded by Shimadzu FT-IR 3000 spectrometer in the diffuse reflectance mode at a resolution of 4 cm −1 , the sample was pressed into KBr disc with a weight ratio of sample to KBr of 1:100 in the range of 4000-400 cm −1 . UV-Visible diffused reflectance (DRS) spectrum of the sample was taken from Shimadzu UV-2600 spectrophotometer and BaSO 4 was used as a reflectance reference material. The absorption spectra in the photocatalytic degradation process were monitored by Shimadzu 2100 UV-Visible spectrometer. The electrochemical impedance spectroscopy (EIS) was performed by XPOT (ZAHNER elektrik instrument). Cyclic voltammetry (CV) and Amperometric (i-t) experiments were carried out using CHI 405a work station and PINE instrument, respectively. All the electrochemical studies were carried out in three-electrode cell system, glassy carbon electrode (GCE surface area = 0.071 cm 2 ) was used as a working electrode, platinum wire and standard Ag/AgCl electrodes were used as a counter and reference electrode, respectively.</p><p>Photocatalytic activity. The photocatalytic activity of the as-prepared Ag 2 MoO 4 was evaluated via degradation of CIP solution under UV light illumination (λ = 200~400 nm). In a typical activity, 50 mg/mL of catalyst was dispersed in 100 mL aqueous solutions of CIP (20 mg/L). Prior to illumination, the solution mixture was stirred magnetically for 30 min in the dark to establish the adsorption-desorption equilibrium between CIP and as-prepared Ag 2 MoO 4 photocatalyst. Then, the solutions were illuminated by UV light (λ max = 365 nm) to induce photocatalytic reaction. During the irradiation, 5 mL of the solution was withdrawn at 5 min time intervals and centrifuged to remove the catalyst. The obtained clear liquor was analyzed by UV-Vis spectrometer to determine the concentration changes of CIP.</p><p>Fabrication of silver molybdate modified GCE. Before surface modification, the GCE was mirror like polished with 0.05 μ m alumina slurry and rinsed with DI to remove the alumina particles on the GCE surface.</p><p>After that the GCE was sonicated for 1 min containing ethanol and water (1:1 ratio). About 5 mg/mL of the as-prepared Ag 2 MoO 4 was re-dispersed in DI water and about 8 μ L (optimized concentration) of the dispersed Ag 2 MoO 4 was drop casted on the GCE surface. Then it was allowed to dry at room temperature, followed by the dried Ag 2 MoO 4 modified GCE was rinsed with DI water to remove the loosely attached catalyst on the GCE surface. The obtained Ag 2 MoO 4 modified GCE was used to further electrochemical experiments. Then, it was stored in room temperature when not in use.</p>

Scientific Reports - Nature

Reversible and Quantitative Photoregulation of Target Proteins

Noninvasive, precise, and reversible regulation of specific proteins is essential to elucidating fundamental biological processes. To achieve this goal, Tan and colleagues have developed a universal, aptamer-based, photoresponsive nanoplatform. By modulating light irradiation, they were able to reversibly and quantitatively manipulate the activity of target proteins in both complex serum and living bacteria samples.

reversible_and_quantitative_photoregulation_of_target_proteins

INTRODUCTION<!>Construction and Optimization of the Nanoplatform<!>Reversibly Regulating Protein Activity<!>Quantitative Activation of the Target Protein<!>Photoregulation of Blood Clotting in Human Plasma<!>Photoregulation of Protein Activity in Complex Living Systems<!>DISCUSSION<!>EXPERIMENTAL PROCEDURES<!>DNA Synthesis<!>Synthesis of AuNPs<!>Aptamer Modification of AuNPs<!>Fluorescence Measurements<!>Real-Time Monitoring of Thrombin Catalytic Reaction<!>Real-Time Monitoring of Lysozyme Activity<!>AFM Assay<!>Quantitative Activation of Thrombin<!>Human Plasma Tests

<p>Proteins, as key components of organisms, are involved in virtually all biological processes. Protein dysregulation can lead to biological dysfunction and disease. Thus, precise and reversible regulation of protein activity at the single-molecule level is essential to understanding complex biological signaling networks or identifying new therapeutic targets. 1,2 Photons, as efficient and controllable external triggers, have been used to perform precise temporal manipulation over many protein systems. 3,4 For example, by covalently attaching a synthetic photoisomerizable small molecule to endogenous ion channels, Kramer and colleagues have developed a light-activated chemical gate for regulating neuronal activity. 5 Lin and co-workers proposed a fluorescent-inducible protein through genetically fusing photochromic Dronpa domains to both termini of the protein, enabling precise optical manipulation of both protease domains and guanine nucleotide exchange factor. 6 Heo and co-workers developed a light-activated reversible inhibition strategy to inhibit protein function by sequestering targets into large assembled protein structures. 7 Although all of these methods can rapidly and reversibly control protein behavior at a defined time and location, they all involve chemical or genetic protein modification. This process is complicated and time consuming, and it potentially affects protein activity, as well as related biological events, thus limiting practical applications. Moreover, none of these methods has been able to achieve quantitative activation of target proteins. Alternative strategies with simple design, convenient operation, little influence on natural protein structure, and potent capability of regulation at the single-molecule level are being sought.</p><p>The advancement of nanotechnology has allowed the construction of many versatile nanoplatforms for biological applications, 8,9 especially protein regulation. 10,11 To</p><p>The Bigger Picture Temporally precise, noninvasive control of target protein activity at the single-molecule level is essential to understanding many protein-dominant biological processes. Here, we report a DNA-based, noninvasive, photoresponsive nanoplatform capable of reversible and quantitative regulation of target protein activity. Adjusting light irradiation allows the target proteins to be captured around the nanoparticle or released into the medium, thus leading to inhibition or activation of the protein, respectively. Moreover, modulating irradiation time allows quantitative control of protein release. Consequently, this nanoplatform can be a potent tool for intensive and precise investigation and manipulation of specific proteins without altering their natural structures, which is significant for studying complex biological signaling networks, identifying new therapeutic targets, and developing intelligent strategies for precision medicine.</p><p>perform targetable control on proteins of interest, specific recognition units are always required. Particularly suited to this task, aptamers are artificial oligonucleotides selected via an in vitro method 12 on the basis of their specific binding affinity 13,14 to the target molecules. Accordingly, they have been applied as promising alternatives to protein antibodies, owing to their intrinsic advantages 15,16 of easy synthesis, convenient modification with various functional groups, and flexible design. The combination of aptamers with nanomaterials is expected to offer unlimited opportunities for protein research. 17,18 Based on many reports, consensus would seem to hold that the formation of high-order protein clusters has the potential to trap and, hence, inactivate protein function. 7,19 The hypothesis is plausible because the active sites of proteins are expected to be trapped and blocked within the assembled protein architectures. Therefore, in the present work, we developed an aptamer-functionalized nanoplatform to reversibly control the formation and dissociation of protein clusters with the expectation of first inhibiting and then restoring protein activity without affecting natural protein structure. As shown in Scheme 1, aptamers able to specifically recognize a target protein were functionalized onto the surface of nanoparticles, and photoresponsive azobenzene (AZO) molecules were incorporated into the aptamer probe, together with a short cDNA sequence. Subsequently, light-driven conformational change of the aptamer controls the capture and release of proteins. More specifically, irradiation with UV light drives a trans-to-cis isomerization of AZO, leading to dehybridization of the aptamer/cDNA duplex, allowing the aptamer sequence to capture the protein of interest, thus leading to inactivation. On the other hand, under illumination with visible light, AZO undergoes a cis-totrans conversion, and the strong affinity between aptamer and cDNA, compared with that between aptamer and protein, again forms, blocking the aptamer and liberating protein, leading to restoration of its activity. Importantly, by modulating the irradiation time, protein release can be quantitatively controlled. Consequently, this nanoplatform can be a potent tool for intensive and precise investigation and manipulation of specific proteins. Protein-specific aptamers modified with several light-responsive molecules are functionalized onto the surface of nanoparticles. By adjusting light irradiation, the light-responsive molecules can drive the conformational switch of aptamer between the blocked status and liberated status. In this way, the target proteins are captured around the nanoparticle or released into the medium, thus leading to inhibition or activation of the protein, respectively. Moreover, by modulating light irradiation, protein activity can be quantitatively controlled.</p><!><p>With the aim of developing a simple, universal, and potent protein regulation nanoplatform with high potential for practical applications, we chose gold nanoparticles (AuNPs; diameter = 13 nm)-which are among the most well-studied nanoparticles with advantages of easy synthesis in aqueous solution, uniform size and shape, low biotoxicity, 20 and efficient functionalization with oligonucleotides 21,22 -as the ''core'' for protein assembly. Thrombin, a key protein implicated in blood coagulation 23 and one whose activity can be easily evaluated by common measurement techniques, was used as the model protein. A 29-mer thrombin-binding aptamer 24 (TBA) able to specifically recognize thrombin but with little inhibition on thrombin activity was used to functionalize AuNPs via a gold-thiol bond, and the resultant product was termed TBA-AuNPs. In principle, active thrombin can catalyze the conversion of fibrinogen to insoluble fibrin, resulting in a rapid increase of scattered light intensity. 25 Figure 1A shows dose-dependent inhibition of thrombin activity by TBA-AuNPs, achieving complete inhibition at the TBA-AuNP concentration of 0.125 nM (green line in Figures 1A and 1B). However, as shown in Figure 1B, negligible thrombin inhibition was observed in the samples containing excess TBA molecules (400 nM, which was about 50 times higher than the TBA concentration of the TBA-AuNPs). In addition, neither equivalent AuNPs nor rDNA 29 -AuNPs (AuNPs modified with 29-mer random DNA sequences) induced observable changes on thrombin activity, indicating that the thrombin inhibition originates from the TBA-specific thrombin assembly around the AuNP core, which is consistent with a previous report. 19 To examine the interaction between TBA-AuNPs and thrombin, atomic force microscopy (AFM) was performed. Larger nanostructures were observed in the mixtures of thrombin and TBA-AuNPs (Figure 1E); neither thrombin only (Figure 1C) nor the mixture of rDNA 29 -AuNPs and thrombin (Figure 1D) showed any assembled morphology.</p><p>To demonstrate the universality of this aptamer-based nanoassembly strategy for inhibiting target protein activity, another protein (lysozyme)-aptamer 26 system was tested. Theoretically, active lysozyme can destroy the cell walls of Micrococcus lysodeikticus, which can be monitored via a decrease in absorbance at 450 nm. 27 As shown in Figure S3, the activity of lysozyme was efficiently inhibited by LA (lysozyme aptamer)-AuNPs in a concentration-dependent manner. However, little inhibitory impact on lysozyme was caused by excess LA molecules (400 nM), AuNPs (0.3 nM), or rDNA 40 -AuNPs (0.3 nM), indicating specific lysozyme inhibition induced by LA-AuNPs. Moreover, AFM was performed to confirm the protein cluster morphology obtained in the LA-AuNPs + lysozyme sample, but not in the lysozyme only or rDNA 40 -AuNPs + lysozyme samples (Figure S4). These results demonstrated the high potential of the current aptamer-based nanoplatform for specific and efficient inhibition of protein activity.</p><p>To exert temporal and reversible control over target proteins, the 5 0 end of TBA was extended with a short cDNA sequence, and AZO molecules, which have been widely introduced into DNA and subsequently shown to change the conformation of DNA via photons, 28,29 were incorporated into this cDNA sequence. Initially, part of the TBA sequence was blocked through hybridization with cDNA, preventing the interaction between TBA and thrombin. However, irradiation by UV light (365 nm) could drive a trans-to-cis isomerization of the azobenzene moiety. This led to dehybridization of the TBA/cDNA duplex, thus releasing the intact functional TBA sequence to capture thrombin around the AuNP core to form a protein cluster, leading to inhibition of thrombin activity. Next, by illumination with visible light at around 450 nm, azobenzene underwent a cis-to-trans conversion. Because the binding affinity between TBA and cDNA was stronger than that between TBA and thrombin, as proved by a competition experiment (Figure S5), TBA/cDNA duplex was again formed and blocked the functional structure of TBA. Thus, thrombin was liberated and its activity restored. Consequently, by adjusting the light illumination, thrombin activity can be reversibly controlled.</p><p>To verify the light-driven conformational change of TBA, a carboxyfluorescein (FAM) fluorophore was modified on one of its termini. To optimize the photoregulation efficiency, three aptamer probes (see sequences in Table S1) containing different numbers of AZO were separately synthesized and modified on AuNPs. The resultant aptamer-modified AuNPs were termed TBA-AZO n -AuNPs, where n represents the number of AZO moieties. When TBA partially hybridizes with cDNA, a hairpin structure forms, and the FAM fluorophore approaches the surface of AuNPs, resulting in fluorescence quenching. However, upon illumination with UV light, AZO undergoes a trans-to-cis conversion, leading to dehybridization of the TBA/cDNA duplex. The fluorophore then withdraws from the surface of AuNPs, and FAM fluorescence is, to some extent, restored. In this light-driven design, enhancement of fluorescence intensity was associated with the time of UV illumination (Figures 2A, 2B, and S6A-S6C). On the other hand, irradiation with visible light triggered cis-to-trans isomerization of AZO, accompanied by the formation of DNA hairpin structure and a decrease in fluorescence intensity (Figure 2A, 2B, and S6D-S6F), indicating light-dependent switching of DNA conformation.</p><p>By setting the fluorescence intensity of FAM-TBA-AuNPs at 100%, the fluorescence ratiometric changes of TBA-AZO 3,4,5 -AuNPs between UV and visible irradiation were 4.12%, 56.06%, and 6.36%, respectively (Figure S7). TBA-AZO 4 -AuNPs achieved the best photoregulation of the aptameric conformation and were therefore used as the optimal design in subsequent experiments. With fluorescence spectrometry following a reported protocol, 30 the average number of DNA molecules (TBA-AZO 4 ) immobilized on each AuNP surface was quantified as 58.34 G 4.82. The reversible conformation switches of TBA-AZO 4 -AuNPs in response to alternate illumination with UV (10 min) and visible (5 min) light were also investigated for multiple cycles. As shown in Figure 2C, in each cycle, FAM fluorescence intensity increased upon UV-light irradiation and decreased upon visible-light irradiation. Of note, because intact TBA forms a G-quadruplex structure, which can partially quench FAM fluorescence, the restored fluorescence intensity of TBA-AZO 4 -AuNPs after illumination with UV light was lower than that induced by addition of cTBA. Also, FAM moves farther away from the AuNP surface by complete cTBA/TBA hybridization. Meanwhile, negligible attenuation was observed, even after alternating seven cycles of light irradiation, and little negative impact of UV irradiation on DNA integrity was proven with gel electrophoresis (Figure S8), indicating excellent reversibility and stability of this nanoplatform.</p><!><p>Having demonstrated the reversible conformational changes of this aptamer/ AZO-based nanoplatform, we proceeded to test its ability to photoregulate protein activity. As shown in Figure S9A, the inhibition of thrombin activity was gradually enhanced with incremental addition of TBA-AZO 4 -AuNPs and approached a plateau at 0.125 nM, corresponding well with the results for the TBA-AuNP samples (Figure 1A). To test whether the visible-light-driven conformational change of the TBA probe could restore the activity of thrombin, UV-pretreated TBA-AZO 4 -AuNPs were irradiated with visible light for different lengths of time, followed by immediate mixing with the substrate solution to test their influence on thrombin activity. As shown in Figure 3, thrombin activity gradually recovered with the extension of time of visible-light irradiation. The ability of TBA-AZO 3 -AuNPs and TBA-AZO 5 -AuNPs to reversibly photoregulate thrombin activity was also evaluated (Figure S10), and the results were well correlated with those of the fluorescence measurements just described (Figure S6). Meanwhile, different light treatments induced little change in thrombin activity in the samples containing TBA-AZO 0 -AuNPs with the same design as TBA-AZO 4 -AuNPs, but with no AZO incorporated (Figure S9B), indicating that regulation of thrombin activity originates from light-directed AZO isomerization.</p><p>The ability of this aptamer/AZO-based nanoplatform to photoregulate thrombin activity was further confirmed by monitoring the thrombin-catalyzed conversion process from fibrinogen to fibrin with confocal laser scanning microscopy. To perform this experiment, an Alexa-Fluor-546-labeled fibrinogen, whose fluorescence can be significantly intensified after conversion into fibrin, was used as the substrate. For fibrinogen only, no obvious fluorescence signal was observed, even after 66 min of incubation (Figure S11). However, upon thrombin catalysis (Figures 4A and S12), fibrinogen rapidly turned into small fluorescent fibrin units (t = 3 min) and gradually formed larger fluorescent aggregates (t = 11 min), indicating a feasible system for evaluating thrombin activity. When treated with UV-pretreated TBA-AZO 4 -AuNPs, in which the functional TBA sequence was unblocked (Figures 4B and S13), thrombin activity was efficiently inhibited with no observable morphological change. However, upon illumination with visible light, dynamic fluorescence enhancement was observed to be compatible with that of the thrombin-only sample, indicating effective recovery of thrombin activity. The reversible ON/OFF regulation of protein activity was further verified by alternating UV-and visible-light illumination (Figure 4C, 4D, S14, and S15). The fluorescent morphological difference among these samples could be clearly observed through 3D projection of the z stack image at 66 min. The average fluorescence intensity of each image (Figure S16), calculated with ImageJ software, was consistent with the above results. Together, these results verify that the aptamer/AZO-based nanoplatform can be used to reversibly manipulate the activity of target proteins.</p><!><p>Although challenging, quantitative activation of specific proteins, especially at the single-molecule level, is essential for understanding their functional behavior. To demonstrate whether this nanoplatform has the potential to precisely activate target proteins, UV-pretreated TBA-AZO 4 -AuNPs were first used to capture and inactivate thrombin, and then different visible-light irradiations were performed during the catalysis reaction. As shown in Figures 5A and S17, the scattered light intensity rose abruptly after each exposure to visible light. The initial catalysis reaction rate (V 0 ) was calculated according to the slope of the kinetic curve (Figure 5B). A sharp increase in V 0 was induced by each pulse of light, indicating sudden release and activation of thrombin. Moreover, V 0 was proportional to visible-light exposure time.</p><p>To further investigate the quantitative relationship between thrombin activation and visible-light irradiation, TBA-AZO 4 -AuNP-pretreated thrombin was irradiated with visible light for different times, and the thrombin released was collected and mixed with fibrinogen. The catalytic reaction was monitored by reading the kinetic absorption spectrum (at 450 nm) with a 96-well plate reader, which can measure multiple samples at one time (Figure S18). The kinetic absorption spectrum of thrombin at different concentrations was also collected, and a standard curve was prepared by plotting V 0 versus the given thrombin concentration (Figure S19). As shown in Figures 5C and S20, a good linear relationship between amount of activated thrombin and visible irradiation time was obtained, ranging from 0 to 240 s. Meanwhile, 86.42% G 0.40% of the captured thrombin could be released with 240-s visible irradiation, indicating high efficiency of protein activation. Moreover, on the basis of these calculations, 20 thrombin molecules were initially captured per TBA-AZO 4 -AuNP, and activation of one thrombin molecule could be achieved by visible irradiation of 14.4 s (Figure 5D), indicating the strong potency of this aptamer/AZO-based nanoplatform for quantitative protein regulation at the single-molecule level. Moreover, protein detection methods with higher sensitivity and proper light devices could potentially provide the current system with the ability to activate one protein at a time, thus providing a powerful tool for studying elusive biological processes.</p><!><p>After demonstrating the excellent capability of this aptamer/AZO-based nanoplatform for precisely regulating thrombin activity in buffer solution, we next tested its performance in human plasma by measuring thrombin clotting time (TCT). As shown in Figure 6A, TCT induced by thrombin only is 17.1 s, which is within the normal test range (14.0-21.0 s) provided by the manufacturer. After treatment with the UV-pretreated TBA-AZO 4 -AuNPs, however, the TCT value showed a dose-dependent increment and even reached 83.2 s (at 10 nM), which is 3.9 times longer than that induced by thrombin only, indicating that UV-pretreated TBA-AZO 4 -AuNPs can efficiently inhibit the coagulation process of human plasma samples. Next, we examined whether the current nanoplatform could be used to reversibly regulate the coagulation of human plasma samples. As shown in Figure 6B, by alternating irradiation with UV (10 min) and visible light (5 min), the TCT value of the plasma samples was switched accordingly. In addition, the attenuation of photo-driven conversion efficiency was negligible, even after nine cycles of alternating treatment, demonstrating the excellent performance of this nanoplatform for reversibly regulating the activity of target proteins in complex physiological fluids.</p><!><p>To further evaluate the potential of this aptamer/AZO-based nanoplatform for regulating protein activity in more complex living systems, live Bacillus subtilis, a wellstudied Gram-positive bacterium that exists in the gastrointestinal tract of ruminants and humans, was used as a model. The cell wall of B. subtilis can be digested by S21D and S21E) showed red and green fluorescence, respectively, further demonstrating that the proposed mechanism to inactivate protein worked in this system. Large protein-AuNP aggregation architectures (as indicated by yellow arrows in the bright channel) were observed only in the LA-AuNPs-treated lysozyme sample (Figures S21D and S21E), consistent with the AFM results (Figures S4B and S4C) and indicating aptamer-specific lysozyme aggregation around the AuNP core.</p><p>To exert photoregulation on lysozyme activity, the 5 0 end of LA was extended with a cDNA sequence incorporated with five AZO molecules (Table S1). The thiolated probe was modified onto the AuNP surface and the resultant product was termed LA-AZO 5 -AuNP. The live SYTO9/PI-stained B. subtilis were mixed with LA-AZO 5 -AuNP-pretreated lysozymes and then irradiated with visible light for different times (Figure 7A). As shown in Figure 7B, the red fluorescence signal (as indicated with yellow arrows) gradually increased, and the green fluorescence decreased when extending the visible irradiation time, indicating visible-light-induced activation of lysozyme. The death rate of B. subtilis was increased by extending the visible irradiation time (Figure S22), corresponding well with the red-to-green fluorescence ratio calculated with ImageJ software. Specifically, with visible irradiation for 5 min, the percentage of dead B. subtilis reached 38.09% G 9.27% (Figure 7C). However, with no visible-light treatment, only 3.69% G 1.12% dead B. subtilis was observed. Of note, the visible-light irradiation itself had no negative influence on the viability of B. subtilis (Figure S21B). All these results demonstrate that our aptamer/AZO-based nanoplatform can be used to precisely photoregulate target protein activity in complex biological systems.</p><!><p>In summary, we have developed an aptamer/AZO-based nanoplatform for quantitative and reversible regulation of target protein activity. Unlike conventional photoregulation strategies, no modification of proteins is required in our design. By using thrombin and its TBA aptamer as demonstration models, we proved that TBA/AZO-functionalized AuNPs can reversibly regulate thrombin activity in both buffer solution and human plasma, indicating a promising strategy for intelligent anticoagulation therapy. Of note, this protein inactivation strategy was based on aptamer-specific protein capture around the AuNP core without a need for involving special aptamers that themselves are capable of protein inhibition. Because it is much more common and easier to select aptamers that can bind only with their target proteins but with no interference on the protein activity, this design is expected to provide a more universal strategy for protein regulation. Meanwhile, modulating visible-light irradiation achieved quantitative activation of thrombin, providing a potent tool for the precise and dynamic regulation of target protein, which is essential for studying the myriad of protein-dominant biological processes. With a proper light source and highly sensitive detection technique, this photoresponsive nanoplatform is expected to enable activation of one protein molecule at a time. Moreover, another protein (lysozyme)/ aptamer-AuNP system was successfully constructed and applied to a complex living B. subtilis system, revealing that this aptamer-based nanoplatform can be potentially expanded for studying and manipulating different proteins in more complex biological systems.</p><!><p>Materials and Reagents DNA synthesis reagents were purchased from Glen Research (Sterling, VA). KCl, CaCl 2 , KH 2 PO 4 , K 2 HPO 4 $3H 2 O, sucrose, and mercaptoethanol were purchased from Sinopharm Chemical Reagent Co. (Shanghai, China). NaCl, MgCl 2 $6H 2 O, human a-thrombin, fibrinogen from human plasma, and lysozyme were obtained from Sigma-Aldrich (Shanghai, China). Sodium citrate was purchased from Tianjin Fengchuan Chemical Reagent Science and Technology (Tianjin, China). Tween 20 was obtained from Beyotime Institute of Biotechnology (China). PBS 103 (pH 7.4) and Alexa-Fluor-546-labeled fibrinogen were purchased from Life Technologies. Chloroauric acid, tris-(2-carboxyethyl)-phosphine hydrochloride (TCEP), and 1 M Tris-buffer sterile solution (pH 7.4) were purchased from Sangon Biotech (Shanghai, China). B. subtilis was obtained from the China General Microbiological Culture Collection Center (Beijing, China). Milli-Q water (resistance >18 MU cm) was used to prepare all solutions.</p><!><p>All DNA sequences (Table S1) were synthesized on a PolyGen DNA synthesizer. Synthesis and deprotection were performed according to the instructions provided by the manufacturers of the reagents. Deprotected DNA was then precipitated by adding 40 mL (1/10 volume) of 3 M NaCl and 1 mL (2.53 volume) of cold ethanol. The mixtures were placed in a freezer at À20 C for 30 min and then centrifuged at 4,000 rpm at 4 C for 30 min. After removing the supernatants, the DNA products were dissolved in 400 mL of 0.1 M triethylamine/acetate (TEAA) and purified by high-performance liquid chromatography (Agilent 1260) with a C18 column (5 mm, 4.6 3 250 mm, 100 A ˚; GL Science, Inertsil ODS-3). The purified DNA was dried, detritylated by dissolving in 80% acetic acid (200 mL) for 20 min, and then precipitated with 20 mL of 3 M NaCl and 500 mL of cold ethanol. The mixtures were chilled at À20 C for another 30 min, collected through centrifugation at 14,000 rpm at 4 C for 5 min, and then resolved with ultrapure water. The final DNA products were desalted with illustra NAP-5 columns (GE Healthcare), and their concentration was determined by detecting the UV absorption at 260 nm on a BioSpec-nano (Shimadzu).</p><!><p>AuNPs with an average diameter of 13 nm were synthesized by reduction of HAuCl 4 with sodium citrate. 31 Chloroauric acid solution (0.01 wt %, 100 mL) was added to a clean round-bottom flask and heated with stirring and refluxing until boiling. Then, sodium citrate solution (3 wt %, 1 mL) was added rapidly to the solution, which was heated with continuous stirring for another 30 min. The solution was then cooled to room temperature and stored in the dark for subsequent experiments. The size of AuNPs (Figure S1) was characterized with a JEM 2100 transmission electron microscope (Hitachi). A UV-2450 spectrophotometer (Shimadzu) was used to measure the absorption of the AuNPs (Figure S2A). The hydrodynamic diameter of AuNPs (Figure S2B) was measured with a Zetasizer Nano ZS90 DLS system (Malvern Instruments, Worcestershire, UK). The concentration of AuNPs was determined by UV-visible absorbance via Beer's law (A = εbc).</p><!><p>Thiolated DNAs were attached to AuNPs as described in previous literature. 32,33 First, the disulfide bond of thiolated DNAs was cleaved by treatment with TCEP (503) at pH 5.0 for 1 hr, and then the DNAs were desalted with illustra NAP-5 columns. TBA-AZO 0 , TBA-AZO 3 , TBA-AZO 4 , TBA-AZO 5 , and LA-AZO 5 were dissolved in 13 PBS, heated at 95 C for 5 min, and then immediately cooled to 4 C for 2 min to form a hairpin structure. Other DNA sequences (Table S1) were dissolved directly with ultrapure water with no heating/cooling treatment. Then, 500 mL of 13-nm AuNP solution (2.5 nM), 5 mL of Tween 20 (1% v/v), 10 mL of thiolated DNA (100 mM) and 10 mL of citrate-HCl buffer (pH 7.0, 0.5 M) were mixed in a centrifuge tube. After that, NaCl solution (3 M) was slowly added to the mixture to reach a final concentration of 0.1 M. After salt aging at room temperature for 6 hr, the mixtures were centrifuged at 16,200 3 g at 4 C for 15 min. After removal of the supernatants, the precipitate was washed with 5 mM Tris-HCl buffer (containing 0.01% v/v Tween 20, pH 7.4) three times via centrifugation. The resultant DNA-AuNPs were resuspended in PBT solution (25 mM Tris-HCl [pH 7.4], 0.01% v/v Tween 20, 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , and 1 mM CaCl 2 ) and stored at room temperature in the dark for subsequent experiments. A UV-2450 spectrophotometer (Shimadzu) was used to measure the absorption of the DNA-AuNPs (Figure The hydrodynamic diameter of DNA-AuNPs (Figure S2B) was measured with a Zetasizer Nano ZS90 DLS system (Malvern Instruments, Worcestershire, UK).</p><!><p>UV light at 365 nm was provided by a Spectroline Model SB-100P/FA UV lamp (100 W), and visible light at 450 nm was provided by a 250 W high-pressure mercury lamp (Shanghai Jiguang) with a 450 nm glass filter (Shenzhen Zhonglai Technology). The distance between the samples and the UV-visible light source was fixed at 5 cm. To evaluate the light-driven conformational switch of the TBA/AZO probes, FAM fluorophores were modified on one of their termini. Generally, the TBA-AZOn-AuNP samples (200 mL, 2.5 nM) were treated with UV or visible light for a certain length of time, and the fluorescence spectra were recorded with a Fluoromax-4 spectrofluorometer (HORIBA JobinYvon, Edison, NJ). The fluorescence peak intensity at 520 nm was used to analyze the performance of this system. The fluorescence intensity of equivalent AuNPs and FAM-TBA-AuNPs was used as the baseline and the internal standard, respectively. All fluorescence data were calculated by subtracting the baseline and normalized according to the internal standard.</p><!><p>Active thrombin can convert fibrinogen into insoluble fibrin, leading to an increase in scattering light intensity (650 nm) and the light absorbance value at about 450 nm. In brief, the samples were pretreated with or without UV-visible light, as described above, and mixed with thrombin for 10 min. The mixtures were added to the fibrinogen solution, which was dissolved in physiological buffer (25 mM Tris-HCl [pH 7.4] containing 150 mM NaCl, 5 mM KCl, 1 mM MgCl 2 , and 1 mM CaCl 2 ). The thrombin catalytic reaction was monitored by measuring the intensity change of the scattered light with a Fluoromax-4 spectrofluorometer (HORIBA JobinYvon, Edison, NJ). The final concentrations of thrombin and fibrinogen in the mixtures (V = 200 mL) were 2.5 and 1.14 nM, respectively. The excitation and emission wavelengths were both set at 650 nm with a bandwidth of 5 nm for excitation and emission. For calibration of systematic error, the catalytic reaction induced by thrombin only was always performed as an internal standard in each independent experiment.</p><p>In some experiments, to guarantee that most TBAs would be liberated to capture thrombin, the functional AuNPs were pre-irradiated with UV light for 10 min. Meanwhile, the pretreated AuNPs were mixed and incubated with thrombin (2.5 nM) for 10 min before the fibrinogen substrate was added to initiate the catalytic reaction.</p><!><p>Lysozyme is an enzyme that can catalyze the hydrolysis of Micrococcus lysodeikticus, resulting in decreased absorbance at 450 nm (A450). In the detection assay, Micrococcus lysodeikticus (Huich Bio-tech) was suspended in the reaction buffer (66 mM potassium phosphate [pH 6.2]), and its working concentration was adjusted according to the A450 value of 0.6-0.8. Initially, a certain concentration of free lysozyme-or lysozyme pretreated with LA, AuNPs only, LA-AuNPs, or rDNA 40 -AuNPs-was added into the Micrococcus lysodeikticus solution, and the kinetic absorbance spectra were immediately recorded with a UV-2450 spectrophotometer (Shimadzu). The concentration of LA and AuNPs was 400 and 0.3 nM, respectively.</p><!><p>For AFM assay, 10 mL of the prepared sample was dropped onto a mica surface. After 5 min at room temperature, the mica was washed with 200 mL of Milli-Q water four times and dried under nitrogen. All samples were imaged on a Multimode 8 (Bruker, USA) with ScanAsyst mode at 0.977 Hz scan rate.</p><p>In Situ Imaging of the Thrombin Catalytic Reaction For the fluorescence imaging assay, thrombin (2.5 nM) was treated with different TBA-AZO 4 -AuNP samples. Then, Alexa-Fluor-546-labeled fibrinogen (1.14 mM) was added. Subsequently, 30 mL of the mixtures was rapidly transferred to a 35-mm Petri dish and immediately imaged with an inverted microscope (Nikon Ti-E, Nikon, Japan) at different reaction time points. The excitation wavelength was 561 nm. The z stack images of the mixtures were scanned at 66 min. The images were recorded with a 203 dry objective.</p><!><p>In this experiment, we changed to a 96-well reader, which can read multiple samples at a time, to test the activity of thrombin by measuring the kinetic absorbance at 450 nm. To set up a standard curve, pure thrombin at different concentrations was mixed with fibrinogen (1.14 mM), and the kinetic absorption spectra were measured immediately with a Synergy 2 multimode reader (BioTek, USA). The initial catalytic reaction rate (V 0 ) was calculated and plotted versus the thrombin concentration. To guarantee that most TBAs would be liberated to capture thrombin in experiments of quantitative thrombin activation, TBA-AZO 4 -AuNPs were pre-irradiated with UV light for 10 min. Then, the pretreated AuNPs were mixed and incubated with thrombin (2.5 nM) for 10 min. The mixtures were irradiated by visible light for different lengths of time and then centrifuged at 16,200 3 g at 4 C for 15 min. The supernatants were collected and mixed with the fibrinogen substrate, and the kinetic absorbance spectra (at 450 nm) were recorded immediately. The active amount of thrombin was calculated according to the standard curve.</p><!><p>We studied the ability of the aptamer/AZO-based nanoplatform to reversibly regulate thrombin activity in the process of blood clotting of human plasma. To accomplish this, healthy human blood samples were collected with the permission of Xiangya Hospital and processed according to a standard procedure to extract plasma for subsequent analysis. Thrombin with different treatments was added to the human plasma, and the TCTs of samples were measured with an Automated Coagulation Blood Analyzer CS-2000i (Sysmex Shanghai, Shanghai, China).</p>

Chem Cell

Biomaterials for Bioprinting Microvasculature

Microvasculature functions at the tissue and cell level, regulating local mass exchange of oxygen and nutrient-rich blood. While there has been considerable success in the biofabrication of large and small-vessel replacements, functional microvasculature has been particularly challenging to engineer due to its size and complexity. Recently, three-dimensional bioprinting has expanded the possibilities of fabricating sophisticated microvascular systems by enabling precise spatiotemporal placement of cells and biomaterials based on computer-aided design. However, there are still significant challenges facing the development of printable biomaterials that promote robust formation and controlled 3D organization of microvascular networks. This review provides a thorough examination and critical evaluation of contemporary biomaterials and their specific roles in bioprinting microvasculature. We first provide an overview of bioprinting methods and techniques that enable the fabrication of microvessels. We then offer an in-depth critical analysis on the use of hydrogel bioinks for printing microvascularized constructs within the framework of current bioprinting modalities. We end with a review of recent applications of bioprinted microvasculature for disease modeling, drug testing, and tissue engineering, and conclude with an outlook on the challenges facing the evolution of biomaterials design for bioprinting microvasculature with physiological complexity.

biomaterials_for_bioprinting_microvasculature

Introduction<!>General Introduction of Vessel Formation<!>Angiogenesis<!>Vasculogenesis<!>The Roles of Growth Factors<!>PDGF Family<!>VEGF Family<!>FGF Family<!>TGF-\xce\xb2 Family<!>Angiopoietin family<!>Other Molecules<!>Endothelial Cells<!>Human umbilical cord vein endothelial cells (HUVECs)<!>Human microvascular endothelial cells (HMVECs)<!>Induced pluripotent stem cell derived endothelial cells (iPSC-ECs)<!>Endothelial Progenitor Cells (EPCs)<!>Circulating EPCs (cEPCs)<!>Endothelial colony forming cells (ECFCs) and endothelial outgrowth cells (EOCs)<!>Pericytes<!>Mesenchymal stem cells (MSCs)<!>Fibroblasts<!>Vascular smooth muscle cells (vSMCs)<!>The Role of Extracellular Matrix<!>Bioprinting Techniques and Biomaterials Considerations for Bioprinting Microvasculature<!>Droplet-based Bioprinting<!>Extrusion-based Bioprinting<!>Embedded 3D Bioprinting<!>Light-assisted Bioprinting<!>Laser-assisted Direct Writing<!>Laser-based Stereolithography<!>Projection-based Stereolithography<!>Scaffold-free Bioprinting<!>Biomaterials for Bioprinting Microvasculature<!>Naturally Derived Hydrogel Bioinks<!>Collagen.<!>Fibrin<!>Gelatin<!>Gelatin Methacryloyl (GelMA)<!>Decellularized ECM<!>Agarose<!>Alginate<!>Carbohydrate Glass<!>Hyaluronic Acid<!>Synthetic Hydrogel Bioinks.<!>Poly(ethylene glycol) (PEG)<!>Poloxamers<!>Applications of Bioprinted Microvasculature<!>Bioprinting Microvasculature for In Vitro Disease Modeling and Drug Testing<!>Cardiac Tissue Model<!>Lung Tissue Model<!>Liver Tissue Model<!>Kidney Tissue Model<!>Intestinal Tissue Model<!>Placental Tissue Model<!>Vascular Model<!>Cancer Model<!>Bioprinting Microvasculature for Tissue Engineering and Regeneration<!>Bone Tissue<!>Dental Tissue<!>Cardiac Tissue<!>Skeletal Muscle Tissue<!>Skin Tissue and Wound Healing<!>Conclusions and Outlook

<p>The human cardiovascular system consists of a sophisticated hierarchical network of blood and lymphatic vessels that conduct fluids to and from tissues and organs.1,2 Each level of this hierarchy plays a distinct role in maintaining homeostasis throughout the body. Larger vessels like arteries and veins are responsible for transporting large volumes of blood between organ systems. Following Murray's Law, large blood vessels branch into progressively smaller vessels to control local blood pressure and volumetric flow to the tissues and cells within each organ system.3,4 Capillaries are the smallest and most densely distributed vessels in the cardiovascular system and have a specialized role in directly exchanging fluid with cells deep within tissues. The exact distribution and orientation of microvasculature is influenced by the metabolic activity of the given tissue.5 While large and small vessels have specialized roles, they operate in unison to efficiently maintain homeostasis throughout the body.</p><p>The anatomy of large vessels differs from that of small vessels.6 Large vessels (i.e. arteries and veins) have three layers: an inner layer composed of endothelium, a middle layer composed of smooth muscle, elastic tissue and collagen fibers, and an outer layer composed of elastic tissue and collagen fibers.7 The percentage of elastic tissue in arteries is much higher than those in veins since arteries conduct blood at higher pressures.8 Small vessels (i.e. arterioles, venules, capillaries) are much narrower and thinner than arteries and veins. Arterioles and venules have thin layers of smooth muscle and fibrous tissue, respectively.7 Capillaries are the smallest vessels in the body and are only one cell layer thick to allow for fluid permeability and mass exchange. The cell types within large and small vessels also differ slightly. Arteries, arterioles and veins are composed of endothelial cells (ECs), smooth muscle cells (SMCs), and pericytes.9 Venules are usually made up of ECs and pericytes, along with SMCs, which have distinct characteristics compared to SMCs derived from arteries.10–12 Capillaries are composed of a single layer of ECs and some pericytes for stabilization.13 The large vessels are mainly responsible for mass transport while the small vessels, especially capillaries, are involved in multiple biological processes, including mass exchange, immune response, lymphocyte migration and homing, etc.1,2,13 The varied structures and functions of blood vessels exemplify their remarkable complexity.</p><p>Engineering the complexity of microvasculature has been a key obstacle in the field of tissue engineering since its inception.14,15 Diffusion of oxygen and nutrients within tissues is effectively limited to 100–200 μm. Therefore, engineered tissues larger than these dimensions require endogenous microvasculature for proper nutrient delivery and survival in vivo.15–18 Numerous biofabrication methods have been developed to create vascular networks in vitro, which generally involve microfluidics-based molding techniques.19–23 In addition, controlled delivery of proangiogenic factors like VEGF within biomaterials (e.g. hydrogel scaffolds) has also been a popular strategy to promote vascularization.24,25 Despite the significant progress made with these techniques, they generally lack the spatiotemporal precision and control required to replicate the physiological complexity and function of 3D vascular networks.</p><p>To address this challenge, 3D bioprinting has emerged as a powerful means of fabricating vascularized tissues with structural complexity unattainable by traditional fabrication methods.26–32 The ultimate ambition for the bioprinting field is to resolve the organ donor shortage by creating patient-specific, transplantable replacement tissues and organs in the lab.33–36 However, while there has been success in creating large and small-diameter vessels using bioprinting approaches, fabricating functional microvasculature in constructs of human scale is still an unmet need and a key hurdle in the clinical translation of bioprinted tissues and organs. Biomaterials play a central role in the bioprinting process and serve as writing materials, or "bioinks", for printing the desired tissue construct. Therefore, the development of biomaterials for bioprinting microvasculature is a key driving force in the evolution of the field.</p><p>There have been essentially two approaches to using biomaterials for bioprinting microvasculature – indirect and direct. Indirect approaches employ sacrificial bioinks to print hollow tubes that can conduct fluid within a tissue construct (Figure 1A). Sacrificial, or "fugitive", bioinks can be printed as solid filaments during printing then removed after printing to leave behind hollow channels that can be perfused and endothelialized. While indirect approaches could theoretically be used to print capillary networks, the resolution of most indirect bioprinting platforms (>100 μm) does not approach that of capillaries (5–10 μm). Alternatively, direct approaches exclude the use of sacrificial materials and employ vascular-inductive bioinks containing endothelial cells to guide their self-assembly into capillary networks after printing via cell-cell and cell-matrix interactions (Figure 1B). Since this strategy leverages cells, scaffolds, and signaling molecules to assemble vasculature endogenously, it is more suitable for promoting the formation of smaller vessels (e.g. capillaries) than indirect bioprinting.37 However, there is a limited availability of proangiogenic biomaterials with high printability for direct bioprinting. Indirect and direct approaches for bioprinting microvasculature will be further discussed in Section 3.</p><p>In this article, we will review and discuss the use of biomaterials for bioprinting microvasculature. Before we begin, it is first necessary to define microvasculature in the context of this review. A universally accepted definition of microvasculature is unclear, as it may vary between disciplines. For example, a surgeon may define microvasculature differently than an engineer. The medical definition of microvasculature is "the part of the circulatory system made up of minute vessels (such as venules or capillaries) that average less than 300 μm in diameter". However, reports from the engineering community have described vessels larger than 300 μm as microvasculature, with or without a lining of ECs.38–40 In the bioprinting field, there has been limited consideration given specifically to microvasculature. A search for "microvascular bioprinting" in the PubMed database yields 17 results. In contrast, "vascular bioprinting" yields close to 300 results. While microvascularization has been achieved in numerous vascular bioprinting platforms, there have been limited efforts to intentionally design biomaterials for bioprinting microvessels/capillaries. Therefore, in the interest of breadth, we define microvasculature through an engineering lens. Bioprinted microvasculature satisfying the following criteria was considered for this review: 1. The diameter of the microvessel(s) are around or smaller than 500 μm with preference given to the latter; 2. The microvessels conduct fluid with or without a lining of ECs; 3. Endothelial "cords" or primitive networks without lumens are also included, since they may precede the formation of more patent microvessels.</p><p>Since the selection and utilization of these biomaterials rely on an understanding of biological mechanisms underlying blood vessel development, Section 2 will offer a brief introduction of the fundamental biology of microvessel formation (i.e. angiogenesis and vasculogenesis), including the roles for growth factors, supporting cell types, and ECM. Since the selection of biomaterials also depends on the requirements of the specific technique it will be applied in, Section 3 will review techniques for bioprinting microvessels and their associated printability considerations for biomaterials. Section 4 will critically review the current landscape of biomaterials and bioinks used for bioprinting microvasculature. We categorize bioinks based on the source of the scaffold materials, which include naturally derived and synthetic hydrogels. Section 5 will review recent applications of bioprinted microvessels for in vitro disease modeling, drug testing, tissue engineering, and regenerative medicine therapies. We end with an outlook on future challenges facing the development of biomaterials for bioprinting microvasculature.</p><!><p>Microvessel formation is mediated through highly sophisticated biological mechanisms. Several different models of vessel formation and remodeling are shown in Figure 2.41 Among these, angiogenesis and vasculogenesis are the most extensively studied.41,42 There are significant distinctions between these two models during organogenesis. Vasculogenesis gives rise to the primitive vascular plexus during embryonic development through the differentiation and growth of mesodermal-derived hemangioblasts.43,44 Vasculogenesis also occurs in adults via differentiation of endothelial progenitor cells into ECs. Angiogenesis is characterized by endothelial sprouting and tube formation from pre-existing vessels.45 Angiogenesis and vasculogenesis have been extensively studied and utilized in tissue engineering and regenerative medicine strategies for therapeutic vascularization.41,42 Therefore, the following sections will provide some background on the biological mechanisms driving these processes.</p><!><p>Angiogenesis is the process of new blood vessel formation from pre-existing vessels.42 In addition to physiological conditions, angiogenesis is associated with multiple pathological conditions (e.g., atherosclerosis, chronic inflammation and cancer). Significant progress has been made in revealing the underlying mechanisms of angiogenesis.46 Numerous comprehensive reviews about angiogenesis can be found in references 41,42,47. Therefore, the following sections provide a brief introduction of the current consensus of angiogenesis mechanisms. In addition, the effects of growth factors, cell sources and ECM will also be reviewed.</p><p>There are two distinct mechanisms of angiogenesis: sprouting angiogenesis and intussusception. During sprouting angiogenesis, growth factors such as vascular endothelial growth factor (VEGF), angiopoietin-2 (Ang2), and fibroblast growth factor (FGF) trigger proangiogenic gene activation in quiescent vessels. Pericytes detach from the vessels, proteases break down basement membrane, and cell-cell junctions loosen to facilitate sprouting from the vessel wall. A subtype of ECs called "tip cells" migrate along the chemokine gradient and establish the path of the new sprouting vessel (Figure 2A). The neighboring cells of the tip cells, "stalk cells", support the tip cells invading into remodeled ECM and pericytes help stabilize the integrity of the nascent vasculature. During intussusception (Figure 2C), interstitial cellular columns insert into the lumen of pre-existing vessels. Further expansion and growth of these inserted columns lead to vessel branching, eventually causing the remodeling of the vascular networks.452.1.2. Vasculogenesis</p><!><p>Vasculogenesis is initiated by angioblasts during embryonic development to form the primitive capillary plexus.42 In adults, vasculogenesis occurs via migration and differentiation of endothelial progenitor cells (EPCs) from bone marrow into mature ECs (Figure 2B). While vasculogenesis is mostly referred to in a developmental context, vasculogenesis has also been reported in cultures of mature ECs and supporting cells (e.g. pericytes and fibroblasts).48–51 In the tissue engineering field especially, vasculogenesis is used loosely to describe de novo formation of vascular networks from dissociated suspensions of endothelial cells. Some of these models can be found in Table 1.</p><!><p>As described above, sprouting angiogenesis is initiated by proangiogenic signaling molecules (e.g. growth factors). These signals control and direct vessel development during angiogenesis.41,42 Physiologically, these signals are released by cells under hypoxia52, and include but are not limited to platelet derived growth factor (PDGF), vascular endothelial growth factor (VEGF), fibroblast growth factor (FGF), transforming growth factor β (TGF-β), angiopoietins, epithelial growth factor (EGF), and insulin-like growth factor (IGF). We refer readers to references 41,53 for comprehensive reviews about these factors. Here, we will give a glance at each category and their role in vascular morphogenesis.</p><!><p>The most well-studied members of the PDGF family are PDGF-A and PDGF-B, which are encoded by the PDGF gene. They can form three different forms of dimers, PDGF-AA, PDGF-AB, and PDGF-BB. Recent publications have discovered additional PDGF genes and proteins, PDGF-C and PDGF-D.54</p><p>PDGFs play a critical role during development.60 Although the current understanding of the functions of PDGFs in physiological and pathological conditions remains incomplete, emerging literature shows a correlation between the altered expression levels of PDGFs and the pathological/regeneration progression of blood vessels.61 Several PDGF-targeted therapies have been developed. Especially, recombinant human PDGF-BB based therapy has been utilized clinically as a wound-healing therapy for diabetic ulcers.62,63</p><p>There are three known types of PDGF receptors, PDGFR-αα, PDGFR-αβ and PDGFR-ββ. PDGFR activation can affect a variety of signaling pathways (e.g. Ras-MAPK, PI3K and PLC-γ). As a result, the activation of PDGFRs is highly involved in many types of organogenesis, including vascular development. In addition to supporting the fundamental functions of ECs (e.g. survival and proliferation), PDGFs also play critical roles in the function of multiple supporting cell types, such as pericytes64 and SMCs.65 Specifically, PDGF-B targets PDGFR-β as a paracrine signaling mechanism between ECs and perivascular cells.66 Studies have also shown that PDGF-B/PDGFRβ signaling is responsible for the recruitment of pericytes61 and vascular smooth muscle cells.67</p><p>PDGF is also an important molecular mediator of vasculogenesis. The PDGF family functions as a major mitogen for many mesenchymal/neuroectodermal originating cells. PDGFs also have chemo-attractive properties during multiple tissue remodeling processes, such as wound healing, bone formation, and the development of various organs.68</p><p>In summary, the PDGF family has significant roles in angiogenesis and vasculogenesis, especially for mural cell recruitment and vessel stabilization. The following section will introduce VEGF, which is closely related to the PDGF family, as a detailed example.</p><!><p>VEGF is the most well-studied and one of the most critical signaling molecules for angiogenesis. VEGF has several isoforms, including VEGF121, VEGF145, VEGF165 and VEGF189. Some isoforms are matrix-bound while others are soluble. Each isoform plays a distinct role in promoting angiogenesis.69 There are three primary receptors for these VEGF isoforms: VEGFR1 (Flt-1), VEGFR2 (Flk-1 or KDR) and VEGFR3 (Flt-4). Targeting to these different receptors can lead to different effects during angiogenesis. For instance, VEGFR2 (Flk-1 or KDR) is thought to be the primary receptor for EC proliferation and migration, while VEGFR1 (Flt-1) is believed to be an important modulator during vessel development through the VEGF signaling pathway.70,71 VEGF is usually required for angiogenesis in vitro, which can be either exogenously introduced or locally secreted by cells.69 VEGF gradients control filopodia extension and tip cell migration for endothelial sprouting during angiogenesis as well as vessel permeability.72–74</p><p>Interestingly, positive VEGF gradients trigger endothelial cell sprouting while negative gradients inhibit it.75 Furthermore, it has been reported that different forms of VEGF, either enzyme-releasable or permanently-immobilized, contribute to the formation of enlarged or branching vessels, respectively.76 VEGF also plays a substantial role in vasculogenesis, promoting angioblast differentiation from hemangioblasts77–80 and EPC differentiation into ECs via binding of VEGFR2.81</p><p>Overall, VEGF is a central mediator of neovascularization. We refer readers to ref. 82 for a more detailed discussion on the biology of VEGF and its receptors.</p><!><p>FGFs belong to another important protein family for angiogenesis. Among them, basic fibroblast growth factor (bFGF or FGF-2) was the first identified molecule claimed to have "angiogenic effects".83 To date, there are around 20 different FGF isoforms discovered in the FGF family. FGF-1 and FGF-2 are the most studied molecules.84 The correspondent receptors for FGFs are FGFR-1, −2, −3 and −4.</p><p>Molecular biology studies have demonstrated the activation of different FGF receptors leads to distinct functions. For instance, FGFR-1 has been shown as a critical receptor for vascular development during embryonic stages.85 During vasculogenesis, FGF can synergize with VEGF to influence angioblast differentiation or EPC differentiation to ECs.77,78,81 Inactivation of the gene encoding FGFR-3 causes abnormalities in mouse skeletal development.86 Furthermore, several angiogenesis-related pathways are activated by FGFR-1-mediated signaling pathways, such as Ras, PI3K, and PLC pathways, which leads to survival, proliferation, and migration of ECs and supporting cells.87–90 FGFs have high binding affinity to heparan sulfate proteoglycans (HSPGs), making HSPGs function as a reservoir of FGFs, mediating the local concentration and gradient of FGFs. Inspired by this, several biomaterials systems have incorporated HSPGs to sequester and prolong the delivery of FGFs for angiogenesis.91–93 We refer readers to ref. 94 for a detailed review of the biology and therapeutic potential of the FGF family.</p><!><p>The transforming growth factor-beta (TGF-β) family is composed of more than 30 different isoforms.95 TGF-β1 is the most studied among them. TGF-β is secreted in an inactivated form, which forms a large latent complex (LLC). The LLC can be activated by integrin αvβ6 and αvβ8 subunits by multiple cell types through different mechanisms.96–98 In ECs, active TGF-β binds to its receptor and promotes the phosphorylation and activation of type I TGF-β receptor (ALK-1). The signal is transduced through Smad1/5/8 and enhances the secretion of angiogenic factors, such as ID1 or IL1.99</p><p>TGF-β regulates angiogenesis in a context-dependent manner. For instance, angiogenesis is enhanced at a low expression level of TGF-β but inhibited at a high level of expression.100,101 They hypoxic condition of tissues can augment the concentration and effects of TGF-β.102,103 TGF-β can control angiogenesis through different mechanisms. For instance, TGF-β manipulates its targeted receptors (ALK1 and ALK5) to switch between two different signaling cascades, which lead to varied levels of vessel remodeling and maturation.104 Furthermore, TGF-β is capable of changing the expression level and altering the function of other angiogenic factors like VEGF.105,106 TGF-β also has significant roles during pathological angiogenesis.107</p><p>TGF-β also has multiple distinct roles during vasculogenesis. It can induce EC differentiation while also inhibiting endothelial tube formation.104 It enhances VEGF synthesis by MSCs, but inhibits the proliferation of the vascular supporting cell types.108,109 TGF-β signaling is reviewed in detail in ref. 95.</p><!><p>Angiopoietins and Tie signaling play important roles in vascular morphogenesis, homeostasis and remodeling. 110,111 The angiopoietin (Ang) family includes four different isomers, Ang1, Ang2, Ang3 and Ang4.112 Ang1 and Ang2 were initially recognized as agonistic and antagonistic ligands for the Tie2 receptor, respectively.113,114 Ang3 and Ang4 were subsequently discovered as human and mouse orthologues of Ang1 and Ang2.112 In general, the Ang/Tie system controls sprouting angiogenesis, vascular remodeling, EC activation, and mural cell recruitment.111 Recent studies revealed that Ang/Tie system is also involved in regulation of lymphatic system development115, lymphangiogenesis116, inflammation117 and even tumor development.118</p><p>During Ang/Tie signaling cascades, Ang1 and Ang2 have distinct roles. In quiescent endothelium, pericytes release Ang1 to promote EC survival and vessel stabilization.119 During angiogenesis, matrix-bound Ang1 mediates EC migration and adhesion while Ang2 behaves as a competitive antagonist against Ang1 for the Tie2 receptor and promotes pericyte dissociation and vascular permeability to allow tip cells to sprout and respond to angiogenic cues.120 However, the relationship of Ang1 and Ang2 to vessel development is more complicated than "binary opposition". Ang2 regulates vessel formation and regression through Tie2 signaling in a context-dependent manner. As mentioned above, Ang2 is released during EC activation and potentially performs as a stimulator of Tie2 signaling in activated endothelium.121 However, it also maintains quiescent endothelium by balancing the activities of Ang/Tie signaling.122</p><p>The Ang/Tie system plays a critical role in maintaining vessel integrity through pericyte recruitment, as severe defects in recruitment of pericytes in Ang-1 and Tie-2 deficient mice have been observed, leading to edema and localized hemorrhage.123 While the exact mechanisms involved in Ang/Tie-mediated SMC recruitment are not fully understood, Ang1 has been shown to enhance EC-stimulated SMC migration by a mechanism involving up-regulation of endothelial-derived heparin binding EGF-like growth factor (HB-EGF), which is a known effector of SMC migration and recruitment via ErbB1 and ErbB2 receptors.124</p><p>Angiopoietins also perform as co-factors to regulate vessel development and remodeling. There are multiple reports that demonstrated the synergistic effects between Angiopoietins and VEGF118,125,126. In the absence of VEGF, Ang2 induces EC apoptosis and vessel regression. In the presence of VEGF, Ang2 promotes EC migration, proliferation and vessel sprouting in tandem with VEGF.125 The ratio between VEGF and Ang2 also governs vessel development. For instance, Oshima and coworkers showed that a higher ratio of Ang2 to VEGF causes vessel regression while low ratio of Ang2 to VEGF leads to angiogenesis.126,127 In addition to VEGF, Angiopoietins also cooperate with cytokines such as TNF-α 117and IL-6128. For more detailed information about angiopoietins and Tie signaling for vessel development, we refer readers to a detailed review in ref. 111.</p><!><p>Other growth factors, such as BMP, EGF, and IGF, have shown vasculogenic and angiogenic potential. However, effects of most of these factors are mediated through supporting cell types instead of ECs. For instance, it was confirmed that EGF would stimulate A431 cells (a cell type in human epidermoid carcinoma) to secrete VEGFs and promote HUVEC migration.129 Further information about these molecules can be found in references 130–132.</p><p>In addition to growth factors, cytokines are also reported to promote angiogenesis. The activation of IL-8 not only promotes EC proliferation, but it also enhances MMP-2 and MMP-9 secretion in ECs.133,134 Stromal cell derived factor-1 (SDF-1) has been shown to synergize with VEGF to promote angiogenesis.135 SDF-1 is also a potent homing factor that promotes mobilization of endothelial progenitor cells to sites of vascular injury via binding of its receptor CXCR4.136 Growth-regulated peptide-α/growth-regulated oncogene-1 has also been shown to induce EC proliferation.137</p><p>We have provided an overview of growth factors that are known to be important molecules in vessel development through vasculogenesis and angiogenesis. The general interactions of these factors during neovascularization are represented in Figure 3.</p><!><p>Endothelial cells (ECs) are the primary cell type that make up the inner lining of blood vessels.138 Currently, there are a variety of EC subtypes used for vascular biology research.139 For organ-on-chip models and translational applications, human-derived ECs are especially favorable. The selection of ECs for specific fabrication purposes is critical because differences have been shown in the expressions of both surface marker and RNA profiles of ECs derived from different tissues.140 In addition, it has been shown that there are differences in EC cell type between large vessel endothelium and small vessel endothelium.141 We will provide a background on three common EC types used for tissue engineering: human umbilical vein endothelial cells (HUVECs), human microvascular endothelial cells (HMVECs), and induced pluripotent stem cell-derived endothelial cells (iPSC-ECs).</p><!><p>HUVECs are derived from the endothelium of veins from the umbilical cord and are the most popular model endothelial cell type used to study vascular function and pathology. Since the umbilical cord is usually discarded as medical waste, HUVECs are an economical and abundant source of human ECs. Early passage HUVECs present EC markers like CD31 (PECAM-1), von Willebrand factor (vWF) as well as most receptors for growth factors, cytokines, and vascular signaling molecules.142–144 HUVECs can easily be distinguished from vascular progenitor cells since they are negative for the expression of vascular progenitor markers, such as CD133.145 HUVECs can form both large and small blood vessels in vitro.146–148 HUVECs are also widely used in a variety of engineering applications, including tissue fabrication (bioprinting) and organ-on-a-chip.149,150 However, since HUVECs are derived from large veins, they may not fully recapitulate native microvessels like arterioles and capillaries. Further discussion of the potential of HUVECs for microvascular tissue engineering can be found in references 151,152.</p><!><p>HMVECs can be derived from human microvessels in multiple different types of tissues. Based on their original organs, HMVEC could be further categorized into several subtypes, including human adipose-derived microvascular endothelial cells (from adipose tissue), human liver sinusoidal microvascular endothelial cells (from liver) and human cardiac microvascular endothelial cells (from cardiac tissues), etc. Though derived from different tissues, these ECs share common markers, such as vWF and CD31 as well as being LDL uptake positive.151 HMVECs can be incorporated with parenchymal cell types from their tissue of origin to mimic the tissue-specific vascular microenvironment.153,154 Since they originate from microvessels, HMVECs inherently have excellent potential for forming microvasculature in engineered tissues.</p><!><p>Induced pluripotent stem cell-derived ECs are an autologous source of ECs and have been obtained through multiple differentiation methods from several different cell lines.156–159 The markers of iPSC-ECs are CD31, CD34, and VEGFR. They respond to shear stress and can form tubular networks on Matrigel.156–159 In addition, they respond to inflammatory stimulus (e.g. IL-1β, TNF-α and lipopolysaccharide).158,160 These characteristics of iPSC-ECs in immune, transport, hematological, and mechanical response qualifies them as a valuable autologous alternative to primary ECs. iPSC-ECs have been used in various applications, including the fabrication of patient-specific vasculature in vitro for disease modeling and precision medicine.161 However, ECs isolated from a particular organ could lose their organ-specific features once they depart from their native environment.162 In addition, the relatively complicated differentiation protocols from iPSC to EC have greatly hindered the further use of iPSC-ECs.</p><!><p>Endothelial progenitor cells (EPCs) were first identified in human peripheral blood based on shared antigens with hematopoietic stem cells (HSCs).163 EPCs were found to differentiate into endothelial cells and contribute to neovascularization in adults, similar to the paradigm of vasculogenesis previously thought to be restricted to embryonic vascular development.42,164 The capacity of EPCs to augment collateral vessel growth to sites of ischemia has made them a popular cell source for therapeutic vascularization and vascular tissue engineering.165,166 Some studies have even shown EPCs to outperform vascular-derived ECs in forming vascular networks in vitro and in vivo.167–169 One of the most studied EPC types, circulating EPCs, will be further discussed in the following section.</p><!><p>Circulating EPCs (cEPCs) were the first of several types of EPCs initially discovered from blood.170 They are generally two subtypes based on their origin, hematopoietic EPCs and nonhematopoietic EPCs.165,166 Hematopoietic EPCs are derived from hematopoietic stem cells, which originate from bone marrow.171 Nonhematopoietic EPCs, based on their nomenclature, are not derived from HSCs but are instead believed to derive from organ or tissue-derived EPCs, including blood cells. 172,173 Due to the heterogeneity of these two subtypes, there is no consensus on the general phenotype, surface markers, and stable origins for cEPCs.</p><p>Hematopoietic EPCs were originally identified through CD34+ cells from peripheral blood and they were probably the earliest portion of cEPCs which were proven to contribute to the treatment of ischemic diseases in vivo by neovascularization.163 In subsequent studies, markers for these EPCs were suggested as CD34+/CD133+/VEGFR2+, which was supported by the correlations between this EPCs phenotype and cardiovascular conditions through clinical observations.163,174 Though these markers are still widely utilized for EPC sorting, a series of additional markers were also recommended, including CD45, CD105, CD106, CD117 and CD144.175 Meanwhile, some characteristics, such as the uptake of acetylated low-density lipoproteins and activated aldehyde dehydrogenase, were also suggested as co-evidences of the phenotype of these EPCs.176 Additionally, some populations of hematopoietic EPCs were also observed to present many similar characteristics with monocytic cells, such as uptake of lectin and acetylated low-density lipoproteins, as well as expressing monocytic marker, CD14.177 Because of these similarities, these EPCs were called "early EPCs" (eEPCs) in some studies.177–180</p><p>Nonhematopoietic EPCs, were believed to derive from nonhematopoietic tissue and vessel walls were one of the most possible sources.172,173 The heterogeneity among isolated nonhematopoietic EPCs has led to little consensus on the typical general marker(s) for isolating nonhematopoietic EPCs. Importantly, nonhematopoietic EPCs have demonstrated less proliferative capacity and progressive senescence during culture, making them less suitable for clinical applications.181</p><p>The unique functions of hematopoietic and nonhematopoietic EPCs are complex and still not fully understood. It is generally agreed upon that these two cell types can differentiate into endothelial lineages and secrete and respond to angiogenic/vasculogenic factors for paracrine effects.169,180,182–184 Recently, subpopulations of EPCs known as endothelial colony forming cells (ECFCs) and endothelial out-growth cells (EOCs) have received additional attention because of their unique functions and good potential for clinical therapies.169,178,185</p><!><p>ECFCs were initially discovered through an endothelial colony formation assay, which was developed with the purpose of clearly distinguishing EPCs and HSCs for precisely sorting EPC phenotype.186–188 Similarly, EOCs were reported through another type of endothelial colony formation assay system.178,185</p><p>The subcategorical definition of ECFCs in the EPC family is still debated. It is suggested that ECFCs are hematopoietic EPCs because they initially were isolated from blood-derived mononuclear cells and they exhibit some classical hematopoietic markers, such as CD34 or CD133.185 However, due to the heterogeneity of ECFCs, the defined markers for these cells are still under development and the most widely used protocols have employed the markers CD146+/CD45-/CD133-, which suggested that these cells originated from vessel walls rather than bone marrow.189 More recently studies even recommended a unique profile of "CD45-/CD34+/CD31low" because it could generate pure endothelial populations.173</p><p>Compared to ECFCs, EOCs are most likely categorized as one type of nonhematopoietic EPCs since they cannot be placed into any classic hematopoietic related cell types due to the undetermined origin.165 In addition, the major approach to obtain EOCs still relies on the endothelial colony formation assay system.</p><p>Both ECFCs and EOCs exhibit the capacity to differentiate into endothelial lineages and directly contribute to the de novo vessel formation.181,190 Cell populations which can secrete angiogenic/vasculogenic factors have also been discovered in ECFCs and EOCs, which would offer the paracrine effects for neovascularization.183,190 ECFCs display robust proliferative potential, form capillary networks in vivo, and functionally anastomose with host vasculature in vivo, making them a strong cell source for vascular tissue engineering and regeneration.191</p><!><p>Mural cells (e.g. pericytes) play an important role in the regulation of vascular dynamics both in embryonic and adult stages.192 Pericytes support ECs through not only physically wrapping around them but also by modulating ECs through paracrine effects.193 In addition to stabilizing established vessels, pericytes also provide mechanical support, manage the diameter of vessels, and remodel the vascular ECM microenvironment.194–196 Furthermore, recent publications have demonstrated that pericytes also regulate the permeability of vessels and the barrier function in blood-brain barrier system.194,197</p><!><p>Mesenchymal stem cells are multipotent stem cells with potential for osteo-, chondro-, adipo-, and myogenic differentiation.198,199 MSCs are defined by their multilineage potential and ability to self-renew, along with expression of several cell-surface markers, including CD44, CD73, CD105, and CD90, and lack of endothelial or hematopoietic cell-surface markers such as CD45 and CD34.200,201 MSCs are typically harvested from bone marrow or adipose tissue but can also be obtained through isolation from umbilical cord or placenta.200</p><p>MSCs play an important role in angiogenesis and the development of vascular networks. Paracrine effects are one mechanism through which MSCs promote blood vessel formation. MSCs produce and secrete several growth factors and vesicles that enable cell communication and the regulation of vascular development.202Additionally, MSCs have been recognized as perivascular progenitor cells and have the ability to differentiate into vascular phenotypes such as smooth muscle and endothelial cells.203,204 Furthermore, there are many links between MSCs and pericytes, including cell-surface markers and functions such as stabilizing endothelial cells and secretion of pro-angiogenic growth factors.198,205,206 Importantly, MSCs have been shown to have antithrombogenic effects when incorporated into vascular grafts.207 Accordingly, MSCs have been widely used in tissue engineering strategies to facilitate the generation of functional vasculature. Taken together, the autologous availability, low immunogenicity, multilineage potential, and proangiogenic characteristics of MSCs make them excellent supporting cell types in therapeutic angiogenesis and vascular tissue engineering, as reviewed in ref. 208.</p><!><p>Fibroblasts are an important and widely used supporting cell type for vascular studies. Their main function is to secrete ECM scaffold proteins like collagen to reinforce ECM mechanical properties and promote vascular network and lumen formation.209,210 More information about the topic of ECM secreted by fibroblasts for angiogenesis can be found in ref. 209. Fibroblasts also release numerous proangiogenic paracrine factors for the modulation of angiogenesis.211</p><!><p>In mature vessels, vascular smooth muscle cells are responsible for contraction and regulating blood pressure. During embryonic vascular development, vSMCs have a high proliferation rate and produce a large number of ECM components for blood vessel wall assembly.212 In addition, vSMCs still hold a remarkable plasticity in mature animals.213 A detailed review of the features of vSMCs can be found in ref. 214.</p><!><p>The ECM plays a central role in vascular morphogenesis. In the quiescent state, there is a dense basement membrane surrounding blood vessels, which is mainly composed by type IV collagen and laminin proteins. In addition to serving as physical scaffolds, they also maintain blood vessel homeostasis through cell-ECM signaling. During angiogenesis, the basement membrane is degraded by proteases (e.g., MMPs) secreted by the cells activated by angiogenic stimuli (e.g., hypoxia, growth factors). This disrupts the basement membrane and exposes the sprouting ECs to the interstitial ECM to facilitate their proliferation and migration. Glycoproteins in the interstitial ECM, such as fibronectin, collagen, and laminin, directly engage cell surface integrins to support vessel formation. It has been demonstrated that different types of integrin activation can promote distinct orientation and density of nascent blood vessels.215 The interstitial ECM is also rich in proteoglycans and glycosaminoglycans (GAGs), which can bind to angiogenic growth factors (e.g., VEGF and FGF) and sequester their release in a precise spatiotemporal manner for vessel patterning. Overall, the ECM functions as a dynamic biomolecular scaffold to guide and support neovascularization.216 A general relationship between the ECM and ECs during angiogenesis is illustrated in Figure 4.</p><!><p>There have been numerous techniques developed for bioprinting microvasculature. Though not the focus of this review, we feel it is important to have a basic understanding of these techniques as each have unique requirements for printability and therefore require distinct biomaterials properties. A basic understanding of the different bioprinting techniques will provide the reader with necessary context before analyzing biomaterials for bioprinting in Section 4.</p><p>Each modality has unique advantages and disadvantages for bioprinting microvasculature in terms of speed and resolution. Bioprinting techniques are commonly categorized as droplet-based bioprinting (DBB), extrusion-based bioprinting (EBB), and light-assisted bioprinting (LAB). Numerous in-depth review articles have been published detailing these modalities and their applications in bioprinting.32,36,218,219 Here we focus on bioprinting platforms in the context of bioprinting microvasculature. In each section, we will provide a brief introduction to the bioprinting techniques, their associated biomaterials requirements for printability, and applications in vascular tissue engineering before critically analyzing and comparing their suitability for bioprinting microvasculature.</p><!><p>Droplet-based bioprinting, or DBB, is an approach that involves the serial deposition of droplets of biomaterials and/or cells in precisely defined 2D or 3D arrangements. Like commercial inkjet printers that propel droplets of ink onto paper to reproduce a digital image, inkjet bioprinters propel droplets of "bioinks" onto a bioprinting substrate, sometimes referred to as "biopaper". Bioinks are formulations of biomaterials and/or cells that serve as the writing material for bioprinting and are discussed further in Section 4. The resolution of droplet-based bioprinting is generally around 50–300 μm, making it suitable for printing microvasculature. Capillary network formation in DBB approaches relies on self-assembly of ECs in the printed bioink. Therefore, proangiogenic bioinks or biopaper substrates are ideal to promote microvascularization after printing. DBB can be further categorized into inkjet bioprinting, acoustic-droplet-ejection bioprinting, and micro-valve bioprinting, depending on the means of droplet formation. For more details on DBB techniques, we refer readers to a comprehensive review of droplet-based bioprinting in ref. 220.</p><p>In general, bioinks for droplet-based bioprinting must have a low viscosity (<10 mPa s) as it becomes increasingly difficult to generate droplets in high viscosity bioinks, which may cause clogging at the nozzle orifice. Cell density also affects droplet formation, with higher densities leading to increased droplet size, decreased droplet velocity, and increased breakup time.221 To preserve the integrity of the printed structure, it is ideal to use biomaterials that can be rapidly crosslinked to form a solid hydrogel after deposition. This can be accomplished by printing bioinks into a liquid solution containing crosslinker, applying crosslinker solution to the printed bioink through another nozzle or by mist, or by using photopolymerization for photosensitive biomaterials. There are numerous biomaterials that can be crosslinked instantaneously via physical or chemical methods, which will be discussed further in Section 4. To bioprint microvasculature using DBB methods, it is imperative to consider the proangiogenic features of both the bioink and the printing substrate. The properties of the bioink and substrate should be complementary in promoting both high printability and rapid self-assembly of endothelial cells into functional vascular networks during culture. Benning and others recently conducted a side-by-side comparison of conventional hydrogel bioinks and found that collagen and fibrin were most suitable for inkjet bioprinting of endothelial cells as they best supported HUVEC proliferation in 2D and sprouting from HUVEC spheroids after 3D printing.222</p><p>DBB techniques are valuable tools for microvascularized tissue engineering due to their high resolution, precision, and cytocompatibility. Biomaterials, cells, and other biologics may be deposited with great spatiotemporal control in droplets that are nano- or picoliters in volume. Boland's group pioneered the modification of commercial inkjet printers for direct droplet-based bioprinting of microvasculature. The first demonstration used a modified Hewlett-Packard (HP) inkjet printer to deposit bovine aortic endothelial cells and smooth muscle cells onto Matrigel and collagen, respectively.223 The cells remained highly viable after 3 days of culture. Nakamura and others also demonstrated an electrostatically driven inkjet system that was highly biocompatible with endothelial cells.224 To generate 3D tube-like constructs, Boland's group suspended rat smooth muscle cells in an alginate hydrogel bioink for layer-by-layer printing in a CaCl2-containing bath.225 The cells remained viable after two weeks of culture and, interestingly, exhibited vasoreactivity to a vasoconstricting agonist Endothelin-1. A later study demonstrated that endothelial cells could adhere to the pores of the alginate-based printed vascular structures.226 Boland's group has applied their inkjet bioprinting platform to fabricate microvascularized bilayer skin grafts to treat full-thickness wounds in mice.227 Compared to a commercial skin graft, the bioprinted graft promoted wound contraction and formation of healthy, vascularized skin with both dermal and epidermal layers of normal thicknesses. In another study, Atala's group used an inkjet bioprinter to create complex 3D heterogenous constructs228 and showed that the bioprinted structures significantly improved functional vascularization and bone tissue formation in vivo compared to manually seeded scaffolds. Three-dimensional vascular tube-like structures with bifurcations have also been fabricated by valve-based printing of alginate bioinks layer-by-layer into a CaCl2-containing bath.226,229 These studies demonstrate the capabilities of DBB techniques to position multiple cell types in user-defined arrangements with excellent precision and viability, leading to enhanced vascularization and overall function of the tissue construct. The accessibility, affordability, and mobility of droplet-based bioprinters is also very advantageous for translational applications of DBB in microvascularized tissue engineering. Accordingly, recent studies have modified droplet-based bioprinters for in situ bioprinting of cell-laden hydrogels for skin tissue regeneration in small and large animal models.230,231 Lastly, inkjet bioprinting is uniquely advantageous for printing microvasculature as it was recently revealed that thermal inkjet bioprinting triggers activation of the VEGF pathway in human microvascular ECs, as illustrated in Figure 5.232</p><p>Despite the advantages of DBB, there are still important concerns associated with these approaches. One major concern for droplet-based bioprinting is the hydration of printed cells. Since printed droplets are quite small, they may evaporate quickly during the printing process, leaving cells dehydrated. Therefore, the printing substrate should have a high-water content to keep cells within the droplets hydrated. Furthermore, the small droplet sizes generated by DBB methods makes scaling the production of larger tissues or organs a serious challenge. Conventional DBB strategies are mostly limited to 2D structures since the discontinuous droplets may be mechanically unstable when printed in multiple layers.233 Therefore, DBB may be most suitable for bioprinting microvasculature within 2D patches (i.e. skin or cardiac) for tissue engineering or for patterning chemokine gradients onto a 2D surface to study endothelial cell behavior. DBB methods are also relatively slow since the bioinks are printed drop-by-drop, though the throughput of DBB methods can be massively improved with multi-nozzle and multi-material printheads.234,235 Finally, the low viscosity required of bioinks for printing with DBB reduces the versality of these techniques and the breadth of compatible biomaterial formulations. Therefore, novel bioinks containing biomaterials that are both printable and proangiogenic should be emphasized for DBB applications.</p><!><p>Extrusion-based bioprinting, or EBB, uses pneumatic-, piston-, or solenoid-driven actuators to extrude bioinks through a nozzle onto a printing substrate. EBB is a widely used approach due to its accessibility, compatibility with high viscosity bioinks, and fast multi-layer printing times. In EBB approaches, cylindrical bioink filaments can be printed layer-by-layer to form a lattice-like macroporous construct. The continuous extrusion of cylindrical filaments allows larger 3D constructs to be printed with superior mechanical integrity compared to DBB. However, shear stress-induced cell death is more of a concern with EBB due to higher pressures generated at the nozzle during extrusion.236 In addition, EBB techniques generally have the lowest resolution of the bioprinting platforms with a minimum feature size above 100 μm, making them less suitable for fabricating capillary-like structures.237</p><p>EBB approaches have been widely utilized for bioprinting vascular constructs. Due to resolution limitations, generation of capillary networks in the printed structures using EBB approaches mostly relies on vasculogenesis and angiogenesis within the filaments after printing, while large vessel-like channels can be printed directly or indirectly.238–242 Most studies have used EBB techniques to rapidly fabricate large channels first, followed by endothelization to form functional vasculature. There have also been demonstrations that achieved formation of capillary networks through angiogenesis from the larger parent vessels during culture.243,244 Vascular networks printed using EBB have been shown to improve mass transport and diffusion within the printed construct.243 In addition to capillary formation, several supporting cell types have also been incorporated into EBB platforms to improve vessel stabilization and maturation, including but not limited to pericytes, smooth muscle cells, and fibroblasts245–247. For instance, Ma's group has incorporated mouse fibroblasts into bioprinted hollow constructs and demonstrated good viability of the fibroblasts after 7 days' culture.247 Zhang and others employed human coronary artery smooth muscle cells (HCASMCs) and human bone marrow-derived mesenchymal stem cells (hMSCs) to facilitate 3D small-diameter vasculature formation.248 Furthermore, there are also demonstrations using tissue spheroids rather than single cells as building blocks for EBB.249 Overall, EBB techniques are among the most popular for vascular bioprinting due to their capacity to rapidly print tubular structures and multi-layer constructs, as well as their accommodation of a wide range of bioinks. For a comprehensive analysis of current advances in EBB, we refer readers to a detailed review by Ozbolat and Hospodiuk in ref. 250.</p><p>Coaxial extrusion is a popular type of EBB approach for printing microvasculature. Coaxial nozzles are composed of an inner and outer compartment, allowing simultaneous extrusion of a bioink and a crosslinker solution in a core-shell fashion for rapid gelling at the dispensing head. The immediate crosslinking at the nozzle orifice in coaxial systems enables printing accuracy to be decoupled from bioink rheological behavior251 and allows for the fabrication of multi-layer constructs with low viscosity bioinks.149 Hollow tubular fibers or bulk fibers can be printed by extruding crosslinking solution in either the core or shell compartment, respectively (Figure 6). The core/shell element of coaxial extrusion is a powerful feature, as it allows for rapid fabrication of perfusable tubular constructs with one nozzle. Further, different biomaterials and cells can be incorporated into the core and shell compartments to generate heterogenous tubular structures. For instance, Liu and others used a custom Dual Ink Coaxial Bioprinter to fabricate vascularized pancreatic constructs.252 Islets were housed in the core compartment and were surrounded by EPCs or regulatory T cells in the shell compartment. The coaxial positioning of these cell types improved vascularization of the construct while providing immunoisolation to the islets. Coaxial bioprinting systems also enable user-defined control over the sizes of printed channels by adjusting nozzle size and geometry as well as extrusion settings (i.e. pressure/flow rate) of inner and outer compartments, based on the requirements of the applications.247 For example, Millik and others used customized coaxial nozzles of varying diameters and extrusion conditions to generate perfusable hydrogel tubes with different cross-sectional geometries.253 The shape and orientation of the printed vasculature can also be managed through computer-aided design and the 3D printing process. Along with the aforementioned studies, there have been multiple coaxial bioprinting systems developed for printing smooth and continuous lumens in any predesigned length, confirming the power of this technology.247,254</p><p>To further improve the heterogeneity of extrusion-based techniques, microfluidics-assisted multi-material EBB systems have been developed. These are necessary for printing heterogeneous constructs with tunable features that mimic the spatial complexity of human tissues at the microscale.256,257 Most multi-material EBB systems to date have used multiple syringes to sequentially print bioinks one at a time. This is relatively low-throughput and requires the nozzles to be carefully calibrated. The frequent start-and-stop of flow between extrusion can also introduce defects and discontinuity in the extrudate. Extruding multiple materials from one nozzle can increase the throughput and allow for the fabrication of structures with encoded composition and variable properties along the print path.258 To this end, Hardin and others developed microfluidic printheads that could seamlessly switch between two viscoelastic PDMS bioinks "on-the-fly" during printing.259 To switch between inks during printing, syringe B is compressed while syringe A is simultaneously decompressed. This results in a rapid pressure change that permits flow from syringe B while prohibiting flow from syringe A. The timing of this switch may be precisely controlled for programmable microscale properties in the printed construct. Active mixing printheads have also been developed for controlled blending of two bioinks in one nozzle immediately before extrusion.260 Khademhosseini's group recently developed a method to rapidly dispense up to 7 bioinks in one nozzle by bundling several capillary extrusion tips into one dispenser housing and independently programming the flow of each bioink.261 These approaches offer exciting potential to rapidly multiplex different biomaterials and cells within an engineered tissue to enhance its biomimicry.</p><!><p>In most EBB systems, bioinks are directly written onto substrates in open air without supports. This limits the complexity of printed structures and can lead to gravity-induced sagging during the printing process, especially when using soft hydrogel bioinks (<100 kPa). Embedded 3D printing addresses this problem by printing directly into a physical support matrix to prop up the extrudate during printing. This allows for omnidirectional extrusion within the support matrix and minimizes gravity-induced sagging. The hydrated support matrix also helps maintain cell viability during printing by providing an aqueous environment with tight control over pH, temperature, and sterility. During embedded printing, the nozzle generates void space in its wake as it moves through the support matrix. Ideally, the support matrix should exhibit shear-thinning and self-healing viscoplastic properties to accommodate nozzle translation and fill the void space to maintain support of the extrudate.262 Thixotropic hydrogels, which yield to higher loads and fully recover afterwards, are ideal support matrices for embedded 3D printing. Jennifer Lewis's group first developed embedded 3D bioprinting for fabricating acellular microvascular networks within a Pluronic F127 support matrix.241 This platform has also been used for soft robotics applications263 and to embed strain sensors into elastomeric hydrogels.264 Since then, several other embedded 3D bioprinting platforms have emerged with more biocompatible support matrices.</p><p>Granular hydrogels are an excellent support medium for embedded 3D bioprinting.265,266 Granular microgels "jam" to form solid-like matrices at low shear strains, but can yield locally to high shear strains. After the strain is removed, granular hydrogels recover back to their solid-like "jammed" state. More details on the properties of granular hydrogels can be found in a review by Riley and others.267 During embedded printing, granular gels exhibit thixotropic properties, fluidizing around the nozzle then quickly recovering around the extrudate.266 The tip speed and flow rate can be adjusted to tailor the diameter of the extruded filaments. Intricate hierarchical networks containing hollow tubes with diameters of 100 μm have been printed in granular support mediums.266 One approach for fabricating vascular structures using embedded 3D bioprinting involves directly embedding a bioink within a sacrificial support bath (Figure 7A), as demonstrated by Hinton and others269 with an alginate bioink and calcium-containing granular gelatin hydrogel support bath (Figure 7B). Once the bioink is crosslinked inside the matrix, the bath may be removed, and the printed construct retrieved. Another approach relies on embedding a sacrificial bioink within a polymerizable support matrix to indirectly pattern a perfusable microvascular network (Figure 7C). After crosslinking the support matrix around the sacrificial bioink, the bioink can be removed to leave behind a biomimetic vascular network, as demonstrated by Wu and others241 using a sacrificial Pluronic F127 bioink and photopolymerizable Pluronic F127-diacrylate support matrix (Figure 7D). These different approaches allow for the use of a variety of different biomaterials as the bioink and support matrix to fabricate tissue- and organ-specific constructs with perfusable microvasculature.</p><p>Several conventional hydrogel bioinks, including poly(ethylene glycol) (PEG), hyaluronic acid (HA), and alginate can be used as biomaterials for bioprinting in granular matrices.266 Likewise, numerous different biomaterials can serve as the granular medium. In FRESH bioprinting (freeform reversible embedding of suspended hydrogels), developed by Feinberg and others, granular gelatin microparticles are used as the supporting medium (discussed further in Section 4.1.1.3.).269,270 In "GHost writing", developed by Burdick and others, hyaluronic acid hydrogels modified for supramolecular host-guest interactions are used as the supporting medium (discussed further in Section 4.1.2.4.).271 SWIFT bioprinting (sacrificial writing into functional tissue), developed by Lewis and others, uses sacrificial bioinks written into support matrices composed of dense compactions of cellular aggregates, or organ building blocks (discussed further in Section 3.5) to fabricate tissues with physiological cell density.272</p><p>As a relatively emergent approach, 3D embedded bioprinting has shown great promise for its application in microvascularized tissue engineering. Free-standing biological structures have been fabricated with impressive complexity in vitro using embedded bioprinting techniques, including models of the heart269,272,273, brain270, cardiac patches273, and perfusable vascular structures with biomimetic features244,265,271,274,275. Several of these studies have demonstrated physiological cell- and tissue-level function within the printed structures, making them applicable for in vitro drug testing and vascular modeling. Further studies are necessary to demonstrate biocompatibility and organ-level functions of these structures to fully realize their potential as replacements for human tissues and organs.</p><!><p>Light-assisted bioprinting, also known as laser-assisted bioprinting or LAB, uses light energy to manipulate cells and photoreactive biomaterials in 2D, 3D, and recently in 4D, based on a digital design. Laser-assisted techniques are arguably the most suitable for bioprinting microvasculature as they have exceptionally high resolution, with feature sizes less than 10 μm.276 Accordingly, LAB techniques have been used for many tissue engineering applications, including bone277,278, skin279,280, and cardiac281 regeneration, as well as in vitro models of microvasculature for lab-on-a-chip studies282. Tissue engineering applications using LAB are reviewed in detail in refs. 283–286. Light-assisted methods can be categorized into laser-assisted direct writing, laser-based stereolithography, and projection-based stereolithography. While the principles of these methods are discussed in detail elsewhere284,287, we will review and analyze LAB techniques for printing microvasculature and their associated biomaterials considerations.</p><!><p>Laser-assisted direct-write approaches can be additive or subtractive. Laser-induced forward transfer (LIFT) is a common laser-assisted additive technique that uses laser energy to deposit cells and biomaterials directly onto substrates with high resolution and reproducibility.284,287 LIFT setups are typically composed of a pulsed laser source (e.g. Nd:YAG crystal laser), a print ribbon coated in cell-laden bioink, and a collector substrate or biopaper on a motorized stage. When the ribbon is irradiated with laser energy, heat and pressure are generated and a droplet of bioink is ejected onto the collector substrate. To protect cells and biological materials from damaging laser exposure during LIFT, an energy-absorbing layer (e.g. metal or biopolymer) can be placed between the print ribbon and the bioink.288,289 Droplet volume during LIFT is dependent upon laser pulse energy and repetition, and the energy needed for droplet formation depends on the rheological properties of the bioink and the cell density used.290 LIFT principles and physical parameters are discussed in detail in ref. 291.</p><p>Various cell types can be printed with high viability (>95%) using LIFT since it is a non-contact approach.292 ECs have been printed with nearly 100% viability using biological laser printing (BioLP).288 In a more recent study, Wu and Ringeisen used BioLP to print HUVECs into capillary-scale branch/stem structures resembling the complex vein networks of a leaf. Bioinks used in LIFT typically have low material concentration and low viscosity (1–300 mPa·s) to facilitate droplet formation, but they can accommodate relatively high cell densities (up to 60 million cells/mL). For example, 1% wt. alginate has been used as a bioink for printing ECs via LIFT.293 The bioink had a viscosity of 100 mPa·s and, depending on the laser energy, could be printed in droplets around 50 μm in diameter. The viscosity increased 20% when ECs were incorporated at 40 million cells/mL. To promote capillary network formation after printing, proangiogenic biomaterials like Matrigel and collagen can be used as collector substrates.278,281,293,294 Kérourédan and others have optimized LIFT parameters for bioprinting ECs onto collagen biopaper (Figure 8).278</p><p>Subtractive laser-assisted techniques are also powerful platforms for direct writing of capillary networks (Figure 9).295 These approaches mostly rely on photoablation, where focalized high-intensity pulsed lasers cause local ablation of material to etch patterned networks. Nano- or femto-second pulsed lasers have energies of around 80–150 mW and 500–900 mW, respectively, which are enough to break covalent bonds.296 Early work used laser-assisted direct writing to etch microfluidic mixers and artificial capillary networks onto 2D silicon and Pyrex surfaces.297,298 Photodegradation techniques commonly employ synthetic hydrogels since they can easily be modified with photolabile functional groups for tuning of their chemical and physical properties.299 These hydrogel systems are discussed further in Section 4.2.1. The main advantage of laser-assisted direct writing is the simplicity of generating perfusable capillary-scale networks without involving the complex steps necessary for removing sacrificial materials in indirect bioprinting approaches. The main drawback of laser-assisted direct writing techniques is that they are relatively slow and become slower with increasing vessel size. Higher intensity lasers can be used to ablate larger channels, but this comes at the expense of increased cell death near the laser and compromising the structural integrity of the bulk construct. Therefore, photoablation techniques are only practical for fabricating submillimeter-scale vasculature in cellularized hydrogels. For true multiscale vascular bioprinting, laser-assisted direct writing would need to be combined in tandem with a complementary approach capable of printing larger vessels.</p><p>Due to their relative simplicity, laser-assisted direct writing techniques have been widely used to bioprint microvasculature for tissue engineering applications. In a recent study, an LAB bioprinter was developed for in situ patterning of endothelial cells into a mouse calvaria bone defect.277 When printed onto a collagen substrate containing human MSCs and VEGF, the printed cells self-assembled into organized vascular networks that contributed to improved vascularization and bone regeneration compared to randomly seeded endothelial cells, providing evidence of the clinical applicability of LAB. In another study, LIFT was used to pattern human stem cells and endothelial cells in a defined pattern on a Polyester urethane urea (PEUU) cardiac patch.281 The patches were cultured in vitro before being transplanted to infarcted rat hearts, where the LIFT-printed patches improved cardiac functional recovery, capillary density, and functional anastomosis with host vasculature compared to patches with randomly seeded cells. Laser-assisted direct writing has also been used to fabricate skin substitutes by patterning keratinocytes and fibroblasts onto Matriderm, a commercial dermal substitute composed of collagen and elastin.279 The substitutes formed skin-like structures in vitro and promoted blood vessel migration towards the printed cells in vivo when transplanted to a dorsal skin fold chamber in mice. These studies demonstrate the versatility and translational potential of laser-assisted direct writing approaches. However, they are mainly limited to engineering planar tissue constructs.</p><!><p>The stereolithography apparatus (SLA) is the most popular 3D laser-assisted fabrication modality. There are two main types of SLA – laser-based and projection-based. Laser-based SLA utilizes raster scanning of a focused UV or near-UV laser to crosslink a photopolymerizable resin based on a digital CAD model. Laser-based SLA is a bottom-up approach as each layer is polymerized point-by-point. Each cured layer is lowered on a stage in the Z-direction for printing of the next layer and the process is repeated, eventually yielding a 3D object. The CAD models for stereolithography can be derived from 3D drawings (e.g. in PowerPoint slides) or from magnetic resonance imaging (MRI) and computed tomography (CT) scans. Micro-CT scans of corrosion casts can be used for generating CAD models of microvasculature.301 We refer readers to a review by Melchels, Feijen, and Grijpma in ref. 302 for details about the principles of the SLA method.</p><p>Biomaterials used in laser-based additive manufacturing methods like SLA must be photocrosslinkable. They should behave as a liquid in the printing reservoir and rapidly solidify when illuminated with light. There are numerous photocrosslinkable hydrogels and photoinitiators that are suitable for SLA and they are discussed in detail in ref. 286. Synthetic polymers like PEG and PVA and natural polymers like gelatin and hyaluronic acid can be modified with photoreactive acrylate/methacrylate groups for printing with SLA.285 The mechanical properties of constructs printed with SLA can be tailored by varying material concentration, composition, laser exposure time, and laser intensity. To print live cells via SLA, biomaterials should be hydrophilic and crosslinked under mild conditions. Synthetic photopolymerizable polymers can be modified with cell-adhesive RGD peptides and growth factor-sequestering heparan sulfate proteins to enhance their bioactivity.303 Several water-soluble photoinitiators have been identified as cytocompatible in UV and visible light-based systems.304,305 A more comprehensive discussion on materials and additives for stereolithography can be found in ref. 306.</p><p>Stereolithography has proven useful in many biomedical applications, including vascular bioprinting and tissue engineering. Early studies leveraged SLA for rapidly prototyping patient-specific anatomical models using data from imaging modalities like MRI and CT. For example, life-size patient-specific models of aortic aneurysms211 and other arterial pathologies307 have been fabricated from CT data using SLA to help surgeons plan individual procedures, design novel stent grafts, and study physiologically accurate flow dynamics in the altered anatomy. These studies paved the way for using SLA to fabricate cellularized constructs out of photoreactive biomaterials for vascular tissue engineering.</p><p>While SLA approaches have traditionally relied on single-photon UV absorption, two-photon photopolymerization (TPP), or multiphoton polymerization, has been used as a more precise alternative to single-photon polymerization as TPP excitation is highly localized to a small focal volume, enabling nanoscale resolution.283,308–310 Far-red laser light is often used for TPP, which is relatively safe for cell culture. Accordingly, cell-laden constructs have been printed with high viability using TPP.308 Remarkably, vascular structures with lumen diameters <20 μm have been printed using TPP.311 However, the lumens collapsed once they reached 4 μm, indicating a lower threshold for vascular dimensions in TPP methods. Emerging applications of TPP include bioactive site-selective protein modification of biomaterials to guide cell morphogenesis.312–314 In an early study, TPP was used to micropattern RGDS, an adhesive ligand, in PEG hydrogels to guide 3D fibroblast migration, demonstrating the capacity of TPP to guide tissue regeneration at the microscale.315 For vascular tissue engineering, DeForest's group has pioneered the use of multiphoton polymerization and photoablation for laser-based direct writing of capillary networks in cell-laden hydrogels.316 In a recent study, they used multiphoton photoablation to engineer 5–10 μm channels in collagen hydrogels to model biophysical and biomolecular interactions of malaria-infected erythrocytes in human capillaries.317 As evidenced by these studies, the unprecedented resolution of TPP methods holds great promise for engineering capillaries with physiological scale and function within biocompatible hydrogels. The main disadvantages of TPP techniques are their relatively slow speed and short penetration depth, which may be restricted to small constructs (~1 mm thick), limiting the scalability of TPP approaches. Complementary techniques with more robust fabrication capacities (e.g. extrusion-based methods) would likely need to be applied in parallel with multiphoton polymerization to produce multiscale features within human-scale scaffolds.</p><p>The emerging development of 3D holography bioprinting shows promise to significantly accelerate print times for multi-photon approaches.318 This technology has been pioneered by Prellis Biologics. Holographic bioprinting essentially combines the precision of multi-photon polymerization with the speed of projection-based stereolithography, effectively decoupling speed from resolution. A holographic projection is cast onto a photocurable substrate to simultaneously crosslink multiple voxels at once with submicron resolution. The holograms are projected as a series of images at high speeds (up to 300 Hz) to print structures within a 3D field of view without the need for a moving stage. Objects can also be printed inside structures that have already been printed using this method. While the technology is proprietary, Prellis Biologics has developed a holographic bioprinter in collaboration with CellINK to offer commercial products that enable 3D cell culture in microvascular networks.</p><!><p>Projection-based stereolithography is often preferred over traditional laser-based lithography as it provides much faster build times. While laser-based SLA approaches polymerize hydrogels point-by-point, projection-based SLA techniques are top-down approaches that crosslink planes of photocurable material at once according to a digital image. Shaochen Chen's research group pioneered the use of projection-based stereolithography for rapidly fabricating complex 3D microenvironments.319 In their approach, known as Dynamic Optical Projection Stereolithography (DOPsL), a digital micro-mirror device (DMD) and an objective lens are used to project UV light in a 2D image across a plane of a photocurable solution. The DMD contains an array of mirrors that can be flipped to either reflect light or not and can be continuously switched within microseconds, allowing for a dynamic "maskless" projection of custom digital images. A 3D image obtained by CT can be divided into 2D slices and used as projections for rapidly prototyping the scanned object layer-by-layer.</p><p>Perfusable multiscale vascular networks can be patterned within photoreactive hydrogels using projection-based stereolithography (Figure 10).320 These hierarchical networks mimic physiological vasculature, with millimeter- and micron-scale vessels forming continuous vascular trees and capillary networks. Projection-based SLA allows for these different vessels sizes to be fabricated concurrent with each other layer-by-layer, offering unprecedented speed and complexity for advanced vascular tissue engineering. Accordingly, Chen's group has applied projection-based SLA to fabricate prevascularized tissues with complex microarchitecture.321 Their study showed that hierarchical vascular networks could be patterned directly into cell-laden hydrogels using projection SLA, and that these prevascularized constructs significantly improved vascularization and anastomosis when implanted in vivo. In the future, projection-based SLA could be utilized to engineer complex microvascular networks within tissue-specific constructs, such as cardiac or liver, to enhance their biological relevance in vitro and improve engraftment in vivo.322 Furthermore, proangiogenic growth factors such as VEGF and PDGF could be embedded within the tissue-specific architectures for controlled release to promote vascular morphogenesis, as demonstrated by Wang and others.323</p><p>Most SLA techniques rely on ultraviolet (UV) light as an energy source. This is problematic since exposure to UV light can cause DNA damage-induced cell death by apoptosis.324 Next-generation lithography methods are turning to visible light crosslinking for better biocompatibility and clinical translation potential. To this end, visible light-sensitive photoinitiators like eosin Y have been incorporated into hydrogel mixtures of PEG and GelMA for printing highly viable cells.325,326 Grigoryan and others recently identified tartrazine (yellow food coloring), curcumin (from turmeric), and anthocyanin (from blueberries) as nontoxic additives that could absorb visible light for projection-based stereolithography bioprinting.327 Tartrazine was identified as the best candidate since it is FDA-approved and could easily be washed out of the printed construct. This allowed for the fabrication of intravascular topologies of unprecedented complexity within biocompatible hydrogels, offering promise for clinical applications of projection SLA in vascular tissue engineering.</p><!><p>Scaffold-free bioprinting, as the name suggests, excludes the use of biomaterials as scaffolding material in the bioink formulation. Instead, dense populations of cells are forcibly aggregated together and produce their own ECM to support their shape. As opposed to traditional scaffold-based "cells-in-gels" approaches, scaffold-free approaches allow the use of much higher cell densities that approach physiological levels. Cell-cell interactions are much more prevalent in these systems since cells are not separated by scaffold materials and the cell aggregates, or spheroids, can quickly fuse to form tissue/organ building blocks. There are numerous methods for generating spheroids for scaffold-free bioprinting, with hanging-drop328 and micro-molding329 being the most common. Though the generation of spheroids itself does not involve exogenous biomaterials, spheroids are often printed within hydrogels to support their arrangement into 3D constructs. This is because spheroids by themselves have poor printability and cannot be bioprinted into a freeform 3D structure without collapse. Direct embedding of spheroids into hydrogels has been demonstrated in droplet-based, extrusion-based, and recently laser-assisted LIFT bioprinting approaches.</p><p>Some of the first bioprinting platforms for vascularized tissue fabrication utilized 3D cell aggregates. The self-assembling capacity of spheroid microtissues has made them attractive "building blocks" for vascular tissue engineering. Gentile et al developed hollow vascular spheroids by treating E8.5 mouse allantois-derived spheroids with VEGF.330 This method yielded uniluminal spheroids with distinctive inner and outer layers of ECs and SMCs, respectively. These spheroids can fuse to form even larger vascular microtissues while still retaining their hollow core.331 Vascularized macrotissues can also be fabricated by "coating" spheroids with ECs. During culture, endothelial cell-coated microtissues fuse to form macrotissues with endogenous vasculature.332 Importantly, human endothelial cell-based vascular spheroids can form robust blood and lymphatic vasculature after in vivo implantation.333 In a recent study, Pattanaik and others found that prevascularized endothelial-fibroblast aggregates can anastomose with host vasculature in as little as 6–12 hours which could enable high viability postimplantation.334</p><p>While early studies demonstrated impressive engineering of vascularized micro- and macrotissues, their architectural complexity was mostly limited to spherical shapes or patches. To address this, scaffold-free bioinks were adapted for 3D bioprinting. Forgacs and others have made pioneering advances in the field of scaffold-free biofabrication. In their foundational work, Jakab and others established mathematical and experimental models of spheroid fusion to demonstrate their potential as building units for bioprinting.335 They manually printed aggregates of Chinese Hamster Ovary (CHO) cells (~500 μm diameter) into collagen hydrogels in ring and tube-like structures and modeled their fusion into the printed shape. In a later study, aggregates of embryonic cardiac cells and ECs were used to print vascularized cardiac constructs.336 Interestingly, the ECs migrated to the boundaries between spheroids and lined the void space between them. To print the aggregates, they were placed in a glass capillary so that spheroids were densely packed at the tip and could be deposited sequentially into a collagen I substrate. In another study, Tan and others printed ring-shaped molds with alginate and then deposited spheroids composed of ECs and SMCs (1:1 ratio).337 The spheroids fused to form toroid-shaped tissue units and secreted endogenous collagen during in vitro culture. Alternatively, cell aggregates can be suspended in alginate-based bioink and printed onto a moving stage in a CaCl2 bath to fabricate zigzag cellular tubes with the need for supports..338</p><p>Precisely controlling the placement of individual spheroids can be challenging. Spheroid fusion is closely dependent on their packing density339 and inconsistent placement of spheroids may lead to inhomogeneous fusion.30 To address this, spheroids have been pre-fused into solid cylinders overnight before the printing process. The fusion of cylindrical tissues units is faster and more continuous at a large scale compared to spherical units.340 These "tissue strands" can be printed into vessel-like structures and subsequently fuse into a large vascular tube.249</p><p>Ozbolat's research group first proposed the concept of bioprinting tissue strands along with vascular channels to promote multiscale vascularization of tissue constructs (Figure 11).341 A hybrid approach with a multi-arm bioprinter could be used for coaxial dispensing of hollow alginate microfibers in tandem with endothelialized tissue strands. Perfusion of the vasculature during culture would theoretically drive angiogenesis in the tissue strands, eventually establishing perfusable microvasculature throughout the tissue. While this has not been demonstrated experimentally, chondrocyte-laden tissue strands have been used to print cartilage patches.342</p><p>Cell aggregate-based bioinks are promising tools for bioprinting vascularized tissues with clinically relevant cell densities. However, the high costs and laborious cell culture needed to generate the required number of spheroids for scaffold-free bioprinting is a major obstacle to translation. To address this, De Moor and others have developed a platform for high-throughput fabrication of prevascularized spheroids with controlled size and high yield using agarose micromolds (Figure 12A).343 Combining HUVECs with fibroblasts and ADSCs improved capillary formation within the spheroids (Figure 12B). The spheroids were also able to fuse into large constructs and produce robust endogenous capillary networks (Figure 12C). This platform could be a promising approach to regulate the production of a high number of vascularized microtissues for bioprinting capillarized macrotissues.</p><p>Besides being used as writing materials, cellular aggregates can also be used as granular support materials. Sacrificial writing into living tissue, or SWIFT, is a recently developed technique based on the principle of embedded 3D bioprinting. Instead of a self-healing hydrogel support medium, dense compactions of organ building blocks (OBBs) cured in a thermoresponsive Matrigel/collagen matrix are used to support the embedding and subsequent perfusion of a sacrificial gelatin bioink. Jennifer Lewis's group developed the SWIFT method to print elaborate perfusable constructs of physiological cell density (~108 cells/mL).272 The OBBs may be embryoid bodies, organoids, or other cellular aggregates, depending on the desired application. OBBs could be produced using iPSC-derived cells for fabricating patient-specific constructs. The OBB-based matrices effectively behave as a self-healing, viscoplastic matrix, yielding to nozzle translation and recovering in its wake to support 3D embedding. Hundreds of thousands of OBBs could be incorporated into a SWIFT construct, improving upon typical "cells-in-gels" approaches where cell densities are 1 to 2 orders of magnitude lower than that in native tissues. Depending on print speed, filaments between 400–1000 μm in diameter could be embedded. This enabled fabrication of multiscale hierarchical networks within embryoid body (EB) support matrices. Attempts at endothelializing SWIFT constructs with HUVECs resulted in incomplete lining of the sacrificial channels, though some VE cadherin-positive monolayers were observed. Such dense constructs may contract significantly during culture as the dense populations of OBBs remodel their environment, which may impact the stability of embedded vasculature. Spatiotemporal patterning of OBBs in SWIFT constructs would also be a beneficial improvement, as the current version only allows for homogenous casting of the OBB matrix into a mold.</p><p>There have been few studies focusing on microvascularization within scaffold-free systems. Going forward, it is crucial that bioprinted scaffold-free tissues and organs have hierarchical branched vascular networks that span the scale of macro- and microvasculature.344,345 Cell aggregates can be densely compacted and their endogenous ECM can inhibit diffusion of nutrients to cells in the middle of the spheroid. This can lead to a necrotic core if vasculature is not established. Future studies should focus on the formation of functional microvascular networks within macro-assemblies of scaffold-free vascular building units to realize the scalability of these approaches.</p><p>We have presented a necessary background on the large variety of techniques available for bioprinting microvasculature, along with their unique advantages and disadvantages. We feel that this provides the reader with necessary context and knowledge to understand the state-of-the-art in bioprinting approaches for vascular tissue engineering and biofabrication. In the next section, we will turn our focus towards a comprehensive analysis of the role of biomaterials within the various printing techniques for bioprinting microvasculature.</p><!><p>The "raw materials" of bioprinting are formulations of printable biomaterials known as "bioinks". Hydrogels are the most common biomaterials used for bioprinting as they mimic the physical properties of native ECM. Hydrogels can be processed for additive manufacturing by tailoring the nature of their gelation, crosslinking, and polymer composition.346,347 Hydrogels can be further categorized as natural or synthetic, depending on their source. Hydrogels can also be blended into hybrid formulations to customize their properties.</p><p>In this section, natural and synthetic hydrogel bioinks for bioprinting microvasculature will be thoroughly reviewed. In each section, we will critically analyze and discuss the advantages and disadvantages of each platform. The structure and composition of each material will be introduced before focusing on their applications in vascular tissue engineering and bioprinting microvasculature. Hydrogel blends will be reviewed in sections corresponding to their base material. Each section will include critical discussion and outlook on the application of the respective hydrogel in bioprinting microvasculature. For this review, we focused on hydrogel bioinks that have had at least some application in vascularized bioprinting with preference given to those that considered angiogenesis and capillary-scale microvascularization. Both in vitro and in vivo formation of microvasculature in printed bioinks were considered. There are many more hydrogels available for tissue engineering applications that are not discussed here because they have not been established in vascularized bioprinting platforms. We refer readers to references 286,346,348,349 for more comprehensive reviews of hydrogels for tissue engineering. In general, bioinks that are highly printable, biocompatible, and form functional microvascular networks are considered ideal for bioprinting microvasculature Figure 13.</p><!><p>Naturally derived hydrogels originate from a biological source, which may be mammalian or non-mammalian. Naturally derived biomaterials are often isolated by extraction via solvents or enzymatic digestion. The preparations of natural hydrogels are reviewed in ref. 350. Naturally derived hydrogels are favored for their biocompatibility and some are inherently proangiogenic. Naturally derived hydrogels are widely used in tissue engineering applications351 and can be processed for bioprinting in numerous techniques as reviewed in ref. 352. Naturally derived hydrogels can be further categorized as protein-based or polysaccharide-based. Here we review protein-based and polysaccharide-based hydrogel bioinks that have been used for bioprinting microvasculature. Only those hydrogels that have been applied in bioprinting microvasculature will be reviewed but we acknowledge that other naturally derived hydrogels not included in this review are also suitable for bioprinting (e.g., chitosan, silk fibroin).</p><!><p>Collagen is the primary structural component of mammalian ECM and is essential for tissue formation and homeostasis. There have been 29 different types of collagen proteins identified, with type I being the most abundant. Types I, II, III, V, and XI can form fibers. Collagen IV forms a sheet and is the main scaffold component of the basement membrane that surrounds and stabilizes blood vessels.353 Collagen has a triple helical structure with three alpha polypeptide chains composed of thousands of amino acids based on the repeating Gly-X-Y motif.354,355 The chains form stable fibers via hydrogen and covalent bonds. Collagen-based biomaterials are widely utilized in biomedical research and tissue engineering applications and have been reviewed extensively in ref. 356.</p><p>Collagen I hydrogels are excellent biomaterials for therapeutic vascularization as they provide an ideal microenvironment for angiogenesis. Collagen hydrogels can be derived from multiple animal sources, with the most common being bovine. Previous studies have shown that Collagen I may stimulate angiogenesis by binding endothelial cell-surface integrins α1β1 and α2β2 via the GFPGER sequence of the collagen fibril.357,358 Endothelial cells are able to degrade and invade collagen matrices via MMPs to establish vascular networks.359 Collagen I activates Src and Rho to initiate capillary morphogenesis in ECs.360 This leads to disruption of VE-cadherin and basement membrane proteins to drive sprouting in the surrounding matrix followed by maturation of the newly formed network Figure 14. The proangiogenic capacity of collagen biomaterials depends on numerous factors, including polymer concentration and crosslinking.361,362 Rigid collagen gels promote the formation of thick and sparsely distributed microvessel networks while softer collagen hydrogels promote thinner, more dense networks.363 Collagen gels with low matrix density (0.7 mg/mL) cannot support endothelial cell adhesion and migration, while collagen gels that are too dense (>3 mg/mL) inhibit migration and sprout formation. Collagen hydrogels with intermediate matrix density (1.2 to 1.9 mg/mL) promote long, stable sprout formation by balancing endothelial cell proliferation and migration.363</p><p>Collagen has been used in EBB, DBB, and LAB approaches. The printability of collagen bioinks mostly depends on their storage and loss moduli before and after printing.365 Collagen solutions at concentrations of 0.5–1.5% exhibit shear-thinning properties.366 The moduli and gelation kinetics of collagen hydrogels are dependent on temperature and pH, with the storage modulus peaking around a pH of 8 at 37°C (Figure 15A,B).365 Gelation of collagen is optimal at 37°C and a pH of 8 (Figure 15C). In general, the physical crosslinking of Collagen I is slow, which is unfavorable for extrusion-based bioprinting. The slow gelation results in spreading across the substrate upon deposition, which lowers the printing resolution. Gelation of collagen I at 37°C can take several minutes. This can be quantified by the crossover time where the storage modulus (G') becomes greater than the loss modulus (G") (Figure 15D). High cell densities can increase the gelation time of collagen367 and the cells may sediment before gelation is complete, leading to an inhomogeneous suspension.218 This makes collagen-only bioinks generally unsuitable for extrusion-based 3D bioprinting. A stronger material like polycaprolactone can be used as a "framework" to support multi-layer deposition of collagen-only bioinks.368</p><p>The printability of collagen bioinks can be improved by blending with fast-gelling materials or by using alternative crosslinking mechanisms. Crosslinking reagents like EDC/NHS369 and glutaraldehyde (GA)370 have been used to improve the viscosity and mechanical properties of collagen, but these reagents are cytotoxic and not suitable for bioprinting live cells. Instead, cytocompatible crosslinking reagents like genipin371 and tannic acid (TA)372 can be used to improve the printability of collagen for EBB. Importantly, cells remain highly viable when using these reagents. Crosslinking collagen with tannic acid improves its stability by lowering its sensitivity to collagenase.373 Recently, tannic acid crosslinked collagen bioinks were used in core/shell EBB to print a freestanding intestinal villi structure with an endogenous capillary network.374</p><p>Collagen bioinks have been used in a newly developed EBB technique termed "pre-set extrusion". With this method, multiple bioinks can be printed simultaneously in a pre-defined shape through a single nozzle.375 Bioinks are placed into a precursor cartridge with a specific configuration and then attached to the printing nozzle. The bioinks can then be extruded into filaments containing the design of the cartridge. Printing ECs (20 million cells/mL) and hepatic cells (30 million cells/mL) in separate collagen bioinks pre-set in a hepatic lobule design resulted in cell viability and proliferation that were similar to a "mixed" or homogenous design. Notably, HepG2 cells showed the indication of the improved functionality in the pre-set design, as evidenced by higher expression of the CYP3A4 enzyme following rifampicin exposure. Therefore, interactions between vascular and parenchymal cells in heterogeneous printed structures are different than those in homogenous structures and may promote more organotypic function.</p><p>For DBB, collagen I has been blended with alginate to enable instantaneous crosslinking and prevent spreading after printing.376 After the collagen had time to fully crosslink, alginate could be chelated to leave behind a pure collagen hydrogel. This improved the shape fidelity of printed collagen, but some collapse of the droplet was still evident. To address this problem, Gettler and others used a superhydrophobic surface to preserve droplet morphology while printing adipose-derived stromal vascular fraction (SVF) cell-laden spheroids with a collagen I bioink.377 Droplets were immobilized on a hydrophobic coating of polydimethylsiloxane (PDMS) modified with hexamethyldisilazane. Spheroid morphology was preserved using this method and SVF cells remained highly viable after 14 days in static and dynamic culture. Angiogenic sprouting phenotypes were also observed. Colocalization of endothelial-specific lectin Griffonia simplicifolia (GS-1) and alpha smooth muscle actin (α-SMA) indicated the formation of stabilized capillary networks.</p><p>Using collagen I as a "biopaper", or printing substrate, a capillary-like network can be patterned by laser-assisted bioprinting.278 Kerouredan et al patterned endothelial progenitor cells onto an MSC-containing collagen hydrogel substrate (2 mg/mL) using LIFT and then overlaid it with a collagen hydrogel after printing to study subsequent microvascular network formation. Network formation depended on cell density, with higher densities (70 million cells/mL) yielding the most extensive network formation. In a follow-up study, Kérourédan et al used LIFT for in situ micropatterning of HUVECs and stem cells from the apical papilla (SCAPs) directly onto mouse calvaria bone defects.277 A layer of collagen substrate was first spread across the defect before cells were printed either randomly or in ring-shaped, disc-shaped, or crossed-circle-shaped designs, generating a vascularized network to promote bone regeneration. Interestingly, the extent of vascularization differed among designs. The crossed circle shape significantly enhanced vascular network formation and bone regeneration after two months. This in vivo printing platform could possibly bypass the need for an in vitro construction and maturation phase, which would shorten the time to patient bedside.</p><p>Collagen was recently used to bioprint compartments of the human heart using an embedded 3D bioprinting technique. Utilizing an updated version of FRESH (freeform reversible embedding of suspended hydrogels) bioprinting, organ-level anatomical structures were printed, including a tri-leaflet heart valve, multiscale vasculature, and a neonatal-scale human heart.269 Unmodified collagen was printed as an aqueous, acidified solution into a pH 7.4 buffered granular support bath of monodisperse gelatin microparticles. Upon deposition, the collagen rapidly neutralized and gelled to form a filament, exhibiting excellent gelation kinetics. Compared to casted collagen gels, FRESH-bioprinted collagen gels loaded with VEGF promoted more extensive in vivo microvascularization. The collagen bioink used in this study was loaded into the printing reservoir as an acidic solution (pH=3.5), which is cytotoxic. Therefore, this collagen bioink could not be seeded with cells before printing. Nevertheless, this study is a major step forward for vascularized organ bioprinting and greatly improves the printability of collagen bioinks for embedded 3D bioprinting.</p><p>Laser-based direct writing has been used for in situ patterning of capillary networks in collagen hydrogels.300 Early work by Liu and others established an optimal collagen concentration for substrate patterning and cell viability as well as a laser fluence threshold for ablation in collagen hydrogels.378 Hribar and others have used a near-infrared femtosecond laser to pattern microvascular networks in cell-laden collagen hydrogels.379 Gold nanorods were mixed into the hydrogel to help convert laser energy to heat and stimulate thermal denaturation of the surrounding collagen matrix. Interestingly, ECs suspended in the collagen matrix migrated towards the etched microchannels and elongated adjacent to the hollow tube structures. However, perfusion of the patterned channels was not assessed. The authors noted a tradeoff between cell viability and laser power, with higher laser energy (>150 mW) causing significant cell death and complete denaturation of the bottom of the collagen hydrogel where the NIR makes first contact. In a recent study by Brandenberg and Lutolf, a focalized pulsed laser was used to etch hollow microchannels in a collagen matrix. The authors first established feasible dimensions for the gel as well as the geometry of microfluidic networks that could be fabricated using photoablation (Figure 16A). Importantly, microchannels well below 100 μm in diameter could patterned with their approach. The patterned microchannels (~50 μm) could then be seeded with HUVECs by microfluidic perfusion. Within five days, a confluent layer of HUVECs expressing CD31 and VE-cadherin was established within the channels (Figure 16B). This approach was also compatible with other naturally derived hydrogels such as agarose, gelatin, and Matrigel. While this direct writing approach can fabricate true capillary-scale microchannels, it is only feasible in millimeter-scale constructs, limiting the scalability of the technique. Nevertheless, laser-based direct writing is a promising strategy for fabricating intricate capillary-scale networks in collagen hydrogels that could not otherwise be achieved with other printing techniques. However, the intense heat generated during photoablation raises concerns about cell viability and structural integrity with these techniques. Therefore, the laser energy and writing speed must be carefully optimized to prevent significant cell death.</p><p>Collagen shows great promise as a biomaterial for bioprinting microvasculature but still faces some challenges. The limiting factor in collagen bioinks is their relatively low printability, which makes it difficult to print at capillary-scale resolution. Blending with other materials or chemical modification of collagen is currently needed to improve its printability. Laser-assisted approaches may be necessary to achieve high resolution bioprinting with collagen.300 Collagen can be modified with methacrylate groups to be rendered photocrosslinkable, making it feasible to print collagen via SLA approaches.380 However, this has not yet been demonstrated. Developing novel collagen bioinks for SLA could enable high-resolution free-form fabrication and may be a promising avenue for printing microvasculature with collagen bioinks going forward. Another major obstacle is that collagen I hydrogels contract significantly (up to 50% of their initial surface area) during culture, which may compromise the intended geometry of printed constructs.381 Mitigating this contraction is an important consideration when printing collagen-based constructs. Crosslinking collagen with succinimidyl glutarate polyethylene glycol (PEG-SG), for example, can help preserve the initial surface area of collagen hydrogels.381 In addition, the exact composition of collagen is not well understood and can vary depending on its source and processing conditions. This lack of characterization affects the regulation of collagen biomaterial properties and raises concerns about the reproducibility of its bioactivity. Further characterization of collagen is necessary to understand its function and to tailor its printability and vasculogenic properties. As mentioned earlier, collagen sourced from bovine skin is commonly used in tissue engineering and bioprinting. Since this is a xenogenic source that carries the risk of pathogen transfer, the use of bovine collagen is not suitable for clinical translation. Bacterial engineered recombinant collagen may be a better alternative as its composition can be regulated and it poses less risk of immunotoxicity.382 Overall, there is currently a limited availability of collagen blend bioinks tailored specifically for printing microvasculature. Leveraging the proangiogenic properties of collagen in printable bioink formulations could enable robust fabrication of 3D microvascular networks for in vitro and in vivo applications.</p><!><p>Fibrin is the main matrix component of blood clots. Thrombin-mediated proteolysis of cryptic binding sties in soluble fibrinogen proteins results in polymerization of fibrin monomers. The fibrin matrix is covalently crosslinked and stabilized by transglutaminase Factor XIIIa in the coagulation cascade.383,384 Fibrin has been widely used as a biomaterial for wound healing applications and is FDA-approved as a surgical adhesive.385 Fibrin networks form relatively soft viscoelastic hydrogels but exhibit shear stiffening properties under high strains due to the stretching of fibrin monomers.386 The exact properties of fibrin gels depend on the nature of their polymerization, which is influenced by thrombin concentration, salt concentration, Factor XIII concentration, and pH.387</p><p>Fibrin bead assays have been used extensively as a model to study the fundamentals of angiogenic sprouting.146 HUVECs sprout from microcarrier beads when cocultured with fibroblasts in fibrin gels (Figure 17A). The microporous and nanofibrous topography of fibrin networks is conducive to cell adhesion and migration (Figure 17B). Fibrinogen binds integrin αVβ3388, which is required for angiogenesis.389 Cytoskeletal changes via integrin binding and Rho signaling regulate capillary sprouting and lumen formation in fibrin gels.390 Growth factors like FGF-2 and VEGF can bind fibrinogen to regulate endothelial cell proliferation and heparins stabilize and retain these factors within the fibrin matrix.391,392 Prevascularization of fibrin-based constructs is known to accelerate anastomosis with host vasculature after transplantation.393 Fibrin-based biomaterials for tissue engineering have been reviewed in ref. 394.</p><p>Fibrin-only bioinks generally have poor printability. Fibrin gels have poor mechanical integrity and degrade rapidly during culture. Furthermore, fibrin can take several minutes to fully polymerize.396 Solutions of fibrinogen and thrombin must be kept separate until the moment of printing. These solutions generally have very low viscosity and are therefore suitable only for droplet-based bioprinting. Droplets of thrombin crosslinker solution can be printed onto a fibrinogen substrate to print a 2D fibrin construct.396,397 An early study by Cui and Boland determined optimal conditions for printing human microvasculature with fibrin via thermal inkjet bioprinting.396 Solutions of thrombin, CaCl2, and HMVECs were printed drop-by-drop onto a fibrinogen substrate to form cell-laden fibrin lattice structures (Figure 18A). Printed fibers were less than 100 μm in diameter but had some minor deformations (Figure 18B, C). The HMVECs proliferated and formed multicellular networks over 21 days (Figure 18D).396 These networks were quite immature, however, and this platform was limited to 2D printing.</p><p>For 3D extrusion-based bioprinting, fibrin can be blended with more printable biomaterials and/or printed within a support scaffold material. Piard and others developed a bioink blend of fibrin (5 wt.%) and gelatin (5%) to print an osteon-like scaffold containing an inner region of HUVECs and an outer region of hMSCs. Printing these discrete cellular regions led to significantly enhanced neovascularization in vivo compared to casted controls.398 However, blending with gelatin can lead to significant mass loss during culture due to the melting and dissolution of gelatin.</p><p>Though fibrin is a favorable proangiogenic material, it is far from an ideal bioink. Therefore, fibrin is usually blended with more printable biomaterials. Fibrin bioinks can be used when angiogenesis is desired after bioprinting, but they are not suitable as the major bulk material of the printed construct since they degrade rapidly. Fibrin-only bioinks have poor shape fidelity and have mostly been used for droplet-based bioprinting of 2D tissues like skin231 or cardiac patches399. Solid freeform extrusion of fibrin hydrogels is difficult due to the need to keep fibrinogen and thrombin solutions separated until after deposition. While it has not yet been explored, fibrinogen and thrombin could possibly be printed in coaxial nozzles for instantaneous crosslinking upon extrusion. Granular mediums are also a promising platform for fabricating 3D structures with poorly printable materials like fibrin.270 To overcome the rapid degradation of fibrin, fibrinolytic inhibition via addition of aprotinin and tranexamic acid can be used to slow proteolysis.400 The use of recombinant fibrinogen and thrombin is ideal to prevent immunogenicity of fibrin hydrogels and enable clinical translation of fibrin bioinks.401</p><!><p>Gelatin is a mixture of polypeptides formed from denatured collagen. Gelatin is thermo-responsive and forms a hydrogel below 37°C by aggregation of gelatin monomers through hydrogen bonding. At temperatures higher than 37°C, the monomers dissociate and gelatin melts, returning to a liquid state. Gelatin hydrogels contain adhesive peptide sequences like Arg-Gly-Asp (RGD) as well as protease-sensitive sites,402,403 making it a useful biomaterial for tissue engineering and regenerative medicine.404,405</p><p>Human microvascular endothelial cells can form capillary networks on 2D surfaces of micropatterned gelatin.405 Unmodified gelatin dissolves completely after 24 hours of incubation406, but can be crosslinked with glutaraldehyde to slow its degradation. Glutaraldehyde is cytotoxic, though, and should be avoided when using live cells. Phenolic hydroxyl groups can be added to gelatin for enzymatic crosslinking and tailoring of its proteolytic degradability, but this requires the presence of potentially cytotoxic hydrogen peroxide and horseradish peroxidase for crosslinking.406 Alternatively, gelatin can be enzymatically crosslinked with microbial transglutaminase (mTG), though the gelation time is quite slow.407 Sacrificial poly(N‐isopropylacrylamide) (PNIPAM) microfibers have recently been incorporated into mTG-crosslinked gelatin hydrogels for developing 3D microvascular networks. At room temperature, the PNIPAM fibers melt and can be washed out to leave behind perfusable microchannels <100 μm. Refinement of this platform for 3D printing could be further explored.</p><p>Due to its fast melting at normal incubation temperatures (37°C), gelatin is used mostly as a sacrificial bioink. Lee and others have extruded HUVEC-laden sacrificial gelatin tubes within layers of collagen to yield perfusable channels.408 HUVECs were able to line the channel and sprout into the surrounding collagen matrix via angiogenesis. Perfusion of 10 μm fluorescent microbeads confirmed the presence of luminal structures within these capillary sprouts. In another study, the same group printed an endothelial cell-laden fibrin hydrogel bioink between sacrificial gelatin channels to generate robust multi-scale vasculature in a thick collagen matrix (Figure 19).243 Large sacrificial channels (lumen size of ~1 mm) were first printed with gelatin on top of a bottom layer of collagen. A fibrin hydrogel bioink seeded with endothelial cells and fibroblasts was then printed between the gelatin channels and followed by a top layer of collagen. The gelatin was melted and perfused with endothelial cells to form large vascular channels. During culture, capillary networks formed within the fibrin hydrogel and then connected to the large channels by sprouting angiogenesis (Figure 19A). By day 14, the microvascular bed in the fibrin hydrogel functionally connected to parent vascular channels and could be perfused. The presence of the capillary network increased the overall diffusional permeability of the construct compared to without the capillaries (Figure 19B). This platform demonstrated the power of combining direct and indirect approaches to fabricate thick tissues with functional multi-scale vascular structures down to the capillary scale.</p><p>In the FRESH bioprinting method developed by the Feinberg Lab, granular gelatin microgels have been used as sacrificial support mediums for embedded 3D bioprinting of vascular structures.269,270,274 After printing, the gelatin matrix can easily be liquified at 37°C and the embedded structure can be removed without any loss of structural integrity. In the first version of FRESH published in 2015, granular gelatin slurries were created by mechanical blending of a gelatin block gel in a commercial blender. The microparticles formed with this approach were relatively large (65 um), polydisperse, and amorphous, which lead to irregular shapes and sizes of extruded filaments and limited the printing resolution of collagen to 200 μm.270 An updated version of FRESH, published in a 2019 study, used a coacervation technique to produce smaller (25 um), monodisperse, and spherical gelatin microparticles.269 This greatly improved the printing resolution to as low as 20 μm.</p><p>In general, gelatin is a versatile biomaterial for bioprinting microvasculature. The biocompatibility and thermo-reversibility of gelatin hydrogels makes them ideal sacrificial bioinks to pattern vascular networks within 3D constructs. However, since gelatin is commonly printed using droplet-based or extrusion-based techniques, its resolution is limited >100 μm, which is larger than the dimension of capillaries. Therefore, gelatin must be printed with complementary biomaterials to induce capillary sprouting. Also, the thermosensitive gelation property of gelatin requires a cautious regulation on temperature during bioprinting. This may limit the application of gelatin within other biomaterial systems that have conflicting thermogelation requirements with gelatin.</p><!><p>Modification with methacrylamide and methacrylate groups is the most popular strategy to improve the stability of bioprinted gelatin. GelMA is a semisynthetic hydrogel as it is based on a naturally derived material but contains synthetic functional groups. Methacrylation enables covalent crosslinking of gelatin macromers by photopolymerization in the presence of a photoinitiator (Figure 20A,B).409,410 Depending on their degree of methacrylation and gel concentration, GelMA hydrogels can be tuned for user-defined mechanical properties (Figure 20C).409 GelMA hydrogels have been used in a wide range of tissue engineering applications, including bone, cartilage, and cardiac tissues.411 Since GelMA hydrogels retain RGD sequences, they promote endothelial cell adhesion and microvascular network formation.410,412–414 The bioactivity and tunability of GelMA makes it an excellent bioink candidate for direct bioprinting of microvasculature.</p><p>Khademhosseini's research group has made pioneering efforts in developing GelMA bioinks for bioprinting. Their early efforts optimized GelMA bioinks for direct-write bioprinting of hepatocytes.238 Since GelMA prepolymer solutions have very low viscosity, the bioinks were pre-polymerized in the printing nozzle and extruded as solid filaments. This method was limited by the requirement of relatively high concentration bioinks that would be too stiff to accommodate vascular morphogenesis. In more recent studies, GelMA has been blended with alginate and printed using the coaxial extrusion technique to produce complex microfibers. These include solid, hollow, morphology-controllable, and multi-layer microfibers (Figure 21).</p><p>A printable low-viscosity bioink was developed by blending GelMA with alginate and using a coaxial microfluidic printing head to extrude calcium chloride crosslinker in the outer shell.255 The alginate immediately crosslinked upon extrusion and preserved the cylindrical fiber structure, preventing the collapse of the otherwise slow-gelling GelMA. This microfluidic approach allows for either bioink to be printed one at a time or both to be printed simultaneously in a Janus structure. After printing, the GelMA can be crosslinked by UV exposure and the alginate washed out. HUVECs can be incorporated into GelMA/alginate bioinks and printed with good viability, depending on the UV exposure time. Remarkably, encapsulated HUVECs can migrate to the periphery of the bioprinted fibers after 10 days of culture, self-assembling into tubular structures (Figure 21A, B).415</p><p>In another study using a multi-channel coaxial nozzle, hollow GelMA/alginate microfibers were digitally tuned "on-the-fly" by varying flow rates of the bioinks in the channels (Figure 21C).147 In this study, Pi and others developed a "GAP" bioink blend of alginate, GelMA, and eight-arm poly(ethylene glycol) (PEG) acrylate with tripentaerythritol core (PEGOA) to print single and double-layer hollow microfibers.147 PEGOA allowed for UV-induced covalent crosslinking of the fibers after printing and improved the mechanical properties of the bioink compared to GelMA/alginate bioinks without PEGOA. The layered microfibers have been continuously tuned during printing by controlling the flow rates of CaCl2 crosslinker in the core nozzle with the bioinks in the inner and outer shell nozzles. These microfibers have been directly embedded with ECs and SMCs, which exhibited high viability and formed CD31+ and α-SMA+ networks over 14 days.</p><p>Most traditional coaxial systems are only capable of producing cylindrical filaments with no control over their morphology, which limits the complexity of these methods. A novel coaxial bioprinting method was developed by Shao and others to generate morphology-controlled GelMA microfibers that resemble small-diameter blood vessels (Figure 21D).416 Based on the "liquid-rope-coil effect," straight, wavy, and helical GelMA microfibers were printed within a progressively crosslinked alginate matrix. The GelMA bioink in the core was crosslinked via UV light exposure at the nozzle while sodium alginate in the shell was ionically crosslinked upon printing into a Ca2+-containing bath. The flow rate of sodium alginate has been varied to tailor the diameters of straight and helical fibers, and some straight fibers had diameters below 100 μm. HUVECs directly encapsulated within the microfibers were viable, proliferative, and migrated to the periphery of the fiber during in vitro culture, forming a continuous lumen in both the straight and helical fibers. Microfibers generated in this study were not immediately perfusable, however the GelMA could theoretically be degraded from the core to leave a hollow endothelialized channel. While this platform enables more versatility in directly bioprinting microvasculature, it is mostly limited to fiber-shaped tissues like muscle fibers, nerve fibers, and blood vessels.</p><p>While coaxial nozzles are limited to the number of bioinks that can be extruded at one time, custom bioprinting nozzles have been developed to print multiple bioinks in one step. Liu and others developed a multi-material EBB platform that could extrude 7 different GelMA bioinks in a continuous, programmable fashion through a single nozzle.261 GelMA bioinks containing different human cell types (HDFs, HUVECs, HepG2, and hMSCs) were used in this platform to create a heterogeneous heart-like structure and vascularized tissue construct. All four cell types were viable after 7 days, but vascular morphogenesis was not reported. Nevertheless, this platform could greatly increase the throughput of fabricating heterogeneous vascularized constructs.</p><p>Though GelMA contains cell-adhesive RGD sites, its vascular activity can be further enhanced through bioconjugation. For example, VEGF can be chemically immobilized to GelMA via EDC/NHS coupling chemistry (Figure 22A).418 The GelMA+VEGF bioink significantly improved vascular morphogenesis compared to native GelMA after 5 days (Figure 22B). The bioink was applied in fabricating a pyramidal structure containing different amounts of VEGF in radial layers to create a graded vasculogenic niche. A soft GelMA bioink with low methacrylation was printed in the middle of the construct and degraded during culture to form a hollow channel within the structure. HUVECs and MSCs were able to line the channel after perfusion, forming hollow vascular lumens around 500 μm in diameter. However, angiogenic sprouting from the parent vessel into the surrounding GelMA+VEGF matrix was not demonstrated. GelMA with low methacrylation is also a poor sacrificial material, as it took 3 days to fully degrade before the channel could be perfused.</p><p>For more defined control over vascular morphogenesis in GelMA hydrogels, VEGF-mimetic peptides can be conjugated to the polymer backbone. The VEGF-mimetic "QK" peptide developed by D'Andrea and others can bind VEGFR-2 and stimulate angiogenesis.419 Covalently immobilizing an acrylated QK peptide onto GelMA (Figure 23A) can enhance microvascular network formation (Figure 23B).420 Cui and others developed a catechol-functionalized GelMA (GelMA/C) bioink with an immobilized VEGF-mimetic peptide to coaxially print small-diameter vasculature along with a sacrificial HUVEC-laden Pluronic F127 slurry.421 Catechol groups can be crosslinked rapidly in the presence of a trace amount of sodium periodate (NaIO4) and adhere strongly to tissue surfaces. The fugitive slurry contained NaIO4, which rapidly crosslinked the GelMA/C bioink after extrusion.248 The VEGF peptide-functionalized GelMA/C bioink enhanced vasculature development in vitro and, after in vivo implantation, anastomosed with host vasculature and promoted capillary invasion from host tissue after 6 weeks. The use of synthetic biomimetic peptides is encouraged as they can be used to engineer chemically defined hydrogels tailored for specific cell engagement.422 The stiffness of GelMA can also be varied to mimic the unique properties of a tissue-specific niche. Soft GelMA hydrogels promote vasculogenesis and capillary-like network formation of human dermal microvascular endothelial cells (HDMECs) while stiffer GelMA matrices can support osteogenesis and bone matrix formation by hASCs.423 GelMA-collagen blend bioinks have been developed for droplet-based bioprinting of hMSCs and HUVECs to support capillary network and lumen formation after 14 days.424 Blending 2.8% or 4% GelMA with 0.208% or 0.16% collagen I, respectively, leads to a bioink blend with shear-thinning properties and a higher elastic modulus. The main disadvantage to using GelMA as a bioink material is the need to use UV light exposure for crosslinking. UV light causes base damage in DNA, and while most studies have reported decent cell viability (>75%), cytotoxicity and mutagenesis are still a concern when using UV-crosslinked materials.324 The scaling up of bioprinted tissues using GelMA will require longer UV exposures to fully crosslink the entire construct, which could compromise the viability of the encapsulated cells. Visible light crosslinkable gelatin using photoinitiators like eosin Y or Rose Bengal may be a more biocompatible alternative.425,426 However, current visible light photoinitiators take much longer to polymerize than UV-sensitive photoinitiators.</p><!><p>The extracellular matrix is nature's scaffolding material. Cell-ECM interactions are critical for vascular morphogenesis.364 Native ECM is highly complex and its composition varies across different tissues and even regionally within the same tissue. Conventional hydrogels cannot entirely replicate the structure and function of ECM, limiting their capacity to direct cell behavior. To address these limitations, extracellular matrix can be harvested through a tissue biopsy and decellularized to leave behind just the ECM scaffolding. This yields an ideal biological template that can either be reseeded with autologous cells or processed for other tissue engineering applications as a biomaterial.</p><p>There have been many decellularization protocols developed for tissues and whole organs, with the exact methods being dependent on tissue density, geometry, and intended application. In general, decellularization protocols involve physical and chemical agents for lysing cells before rinsing them out of the tissue to leave behind pure ECM. An overview of tissue and whole organ decellularization and their applications in regenerative medicine can be found in references 427–429. While most of the ECM proteins and gross architecture can be preserved, decellularization protocols always result in some loss of ECM surface structure and composition. Microvasculature and other microscale features of native ECM are particularly difficult to preserve during decellularization.428,430,431 Therefore, seeding decellularized ECM scaffolds with endothelial cells often results in incomplete microvascularization. Inadequate microvasculature is a major source of failure for decellularized organ transplants, as leaky vessels and exposed ECM cause edema and blood coagulation.430,431</p><p>Decellularized ECM can be processed for bioprinting to build vascularized tissues from the bottom-up. These bioinks possess proangiogenic ECM proteins and growth factors that many conventional hydrogels lack unless supplemented exogenously. Even then, conventional hydrogel bioinks provide a matrix that mimics broad aspects of soft tissue, while dECM bioinks sourced from specific tissues can more accurately recapitulate the ECM content of that tissue. For example, lyophilized dECM from adipose, cartilage, and heart tissues have been used to print tissue-specific analogs. After decellularization, the dECM was converted into a powder and solubilized into pre-gel solutions that were liquid below 10°C and gels above 37°C (Figure 24A). Interestingly, the rheological properties of the dECM bioinks varied across tissue sources (Figure 24B-D). The dECM bioinks supported long-term cell viability 14 days after printing.432 While structural elements of the dECM were lost during conversion to a powder, the tissue-specific bioinks still promoted lineage differentiation and structural maturation of human tissue-derived mesenchymal stem cells. However, the bioinks had to be kept below 15°C while printing to prevent gelation, which may compromise cell viability during long printing sessions. Also, the slow physical gelation of the dECM bioink prevented it from being printed in multiple layers without a PCL support scaffold.</p><p>The mechanical properties of dECM bioinks can be tailored by incorporating photosensitive crosslinking agents. For example, vitamin B2-induced UVA crosslinking can be used to increase the mechanical strength and stability of dECM bioinks.433 Incorporating various PEG-based crosslinkers (linear, 4-arm, and 8-arm) can also be used for fine-tuning the mechanical properties of dECM-based bioinks.434 Methacrylate groups have been added to kidney-derived dECM (KdECM) bioinks for covalent photocrosslinking.435 This allowed for tunable stiffness before and after printing and KdECMMA hydrogels were significantly more stable in culture compared to non-methacrylated KdECM.</p><p>The capacity of dECM to promote vascular morphogenesis depends on its source. Interestingly, human dermal microvascular endothelial cells (hDMVECs) secrete significantly more proangiogenic factors and express more angiogenesis-related genes when cultured on dECM derived from vascularized tissues (e.g. tracheal mucosa) compared to avascular tissues (e.g. cornea).436 Therefore, it may be best to source dECM bioinks from tissues that are naturally vascularized since their composition would provide microenvironmental cues of vascular niches. Accordingly, dECM bioinks derived from vascular tissue (i.e. aorta) have been developed to treat ischemic disease437 and volumetric muscle loss.438 Gao et al used a hybrid bioink of vascular-derived dECM (VdECM) and alginate to print tubular "bio-blood-vessels". The hybrid bioink provided a suitable environment for proliferation, differentiation, and neovascularization of EPCs and the dual ionic and thermal crosslinking gave the bioink good printability. Loading the bioink with PLGA microparticles for controlled release of atorvastatin, a proangiogenic drug, significantly enhanced functional recovery, capillary density, and arteriole density in a murine hindlimb ischemia model.437 Cho et al used a VdECM bioink and granular gelatin support bath to 3D print a prevascularized muscle construct that improved vascularization, innervation, and functional recovery in a rat model of volumetric muscle loss.438 Coaxial printing of the bioinks led to more robust CD31+ networks along the fibers in vitro compared to if the bioinks were homogenously mixed. Therefore, using the coaxial technique to compartmentalize different bioinks improved their overall performance, highlighting the synergy between method and material to enhance the spatial organization and biomimicry of printed vascularized tissues.</p><p>Vascular-derived dECM (VdECM) bioink was further utilized in another study to coaxially bioprint freestanding, perfusable, and functional microvessels. VdECM was combined with alginate in the shell while Ca2+-containing Pluronic F127 (CPF1-27) was printed in the core (Figure 25A).439 CPF-127 preserved the patency of the microchannels while simultaneously crosslinking alginate in the VdECM-containing bioink. Pluronic F127 and poloxamer bioinks are discussed further in Section 4.2.2. The diameter of the printed vessels depended on the needle size, with 25G nozzles capable of printing channels with a diameter around 250 μm. When HUVECs were printed within the channel, they formed a stable monolayer after a week in culture. A collagen hydrogel with or without growth factors (VEGF and bFGF) was cast around the endothelialized vessels to model angiogenic sprouting from the parent vessel towards a chemokine gradient. Endothelial cells from the main channel only sprouted into collagen containing growth factors (Figure 25B,C), eventually forming lumenized capillaries by day 3.</p><p>While decellularized ECM offers unique advantages over other natural hydrogels in terms of biomimicry, it is often sourced non-autologously. This poses a threat of immune response or pathogen transfer in humans if the ECM is not processed thoroughly. For example, alpha gal antigen, which is a carbohydrate found in mammals and not primates, was found remaining in porcine-derived dECM even after aggressive decellularization.440 In contrast, autologous dECM from human omentum tissue did not stain positive for alpha gal antigen and elicited a lower immune response compared to xenogenic and allogenic dECM, highlighting source-dependent differences in biocompatibility of dECM. Moreover, human omentum dECM hydrogels could efficiently reprogram patient-derived iPSCs into multiple lineages, including endothelial cells, demonstrating that ECM from omentum, which is a highly vascularized tissue, possesses requisite physical and biochemical cues to generate personalized hydrogels for engineering autologous vascularized tissue replacements.</p><p>Noor and others converted dECM from patient-derived human omentum tissue into a personalized bioink to bioprint personalized cardiac patches and perfusable heart-like structures.273 Omentum biopsies could be decellularized and converted into concentrated aqueous solutions (1–2.5% w/v) that were thermoresponsive, forming a weak gel at room temperature and increasing in storage modulus upon incubation at 37°C (Figure 26A-D). Microvascularized cardiac patches were fabricated using cardiomyocyte-laden omentum bioink and a sacrificial gelatin bioink containing ECs (Figure 26E, F). The ECs adhered to the surrounding omentum matrix during incubation (Figure 26G) and the gelatin could be evacuated, leaving behind viable endothelial cell-lined channels around 300 μm in diameter (Figure 26H). Spreading and vascular morphogenesis of the ECs was not demonstrated, however. Using a granular support bath made of gelatin microparticles and xanthan-gum supplemented culture medium, the omentum dECM bioink was then used to print a perfusable small-scale bifurcated blood vessel and vascularized heart-like structure with discrete ventricle compartments. The printed vasculature in this application was relatively large, and incorporation of smaller microvasculature would be needed to enable the viability and function of such thick tissues.</p><p>Overall, dECM bioinks offer unique advantages over conventional hydrogel bioinks. They better recapitulate the biochemical, and microenvironmental properties of native ECM compared to conventional hydrogels, making dECM bioinks an attractive biomaterial for supporting vascular morphogenesis. The extrudability of dECM bioinks can be improved by blending with fast-gelling materials like alginate or using a support scaffold like PCL or granular support medium for embedded 3D bioprinting. For printing vascularized tissues and organs, ECM derived from tissues that are naturally vascularized (i.e. omentum) may suitable for providing a proangiogenic microenvironment to ECs. The greatest promise of dECM bioink platforms is the possibility of printing fully autologous, personalized tissue and organ replacements. Recent work from Tal Dvir's research group has provided proof for this concept, but microvasculature will need to be more of a focus in future studies. Current dECM bioinks have mostly been developed for extrusion-based bioprinting, which has limitations in resolution. Future efforts could explore photocrosslinkable dECM bioinks for high-resolution printing via SLA. Improvements in decellularization methods are also necessary to preserve more of the ECM components that are lost during the decellularization process. Preservation of these ECM components through the refinement of decellularization protocols can further improve the bioactivity of dECM bioinks.</p><p>This section has reviewed protein-based naturally derived hydrogels for bioprinting microvasculature. These hydrogels are summarized in Table 2. In the following section, we will review naturally derived polysaccharide-based hydrogels and their applications in bioprinting microvasculature.</p><!><p>Agarose is a linear polysaccharide composed of repeating units of agarobiose – a disaccharide of D-galactose and 3,6-anhydro-l-galactopyranose. Agarose gelation is thermoreversible, with gelation typically occurring between 30–40°C and melting between 80–90°C. These properties depend on agarose concentration, molecular weight, and number of side groups.441 Aqueous agarose solutions undergo a three-stage gelation process: induction, gelation, and pseudoequilbirium.442 At its gelation point, agarose molecules form helical structures by electrostatic interactions and hydrogen bonding between oxygen and hydrogen in the side groups.443 The self-gelling feature of agarose makes it easy to use and highly biocompatible since potentially toxic crosslinking agents are not needed. Therefore, agarose-based biomaterials have been used for tissue engineering and regenerative medicine.444</p><p>As a bioink, agarose is most suitable for extrusion-based bioprinting due to its viscoelasticity and shear-thinning properties at high concentrations, but concentrations below 4% may be suitable for DBB if printing temperatures are maintained above 37°C to prevent clogging.222 Low concentration agarose bioinks spread across the substrate and have poor integrity after extrusion, but can be blended with alginate for better printability.445 Agarose has also been used as a bioink for LIFT bioprinting.446 Agarose hydrogels have suitable swelling and degradation properties for bioprinting and can provide structural support to 3D printed constructs.222</p><p>Agarose rods have been used as a bioprinting template to indirectly pattern microchannels ranging from 250–500 μm in diameter within casted methacrylated gelatin (GelMA) hydrogels (Figure 27). The microchannels were perfusable and could be seeded with ECs that adhered and spread along the GelMA surface. After the GelMA hydrogel was crosslinked, the agarose rods had to be physically removed, making this approach only feasible for microchannels with simple geometries. Otherwise, physical removal of biomaterials would disrupt the shape and integrity of the printed structure, especially for delicate capillary-scale microvasculature.</p><p>While agarose is biocompatible, agarose hydrogels do not readily support endothelial cell adhesion.447 Therefore, agarose has been blended with other biomaterials that are cell-adhesive (e.g., Matrigel, collagen).448 Agarose-collagen blend bioinks can support adhesion and spreading of human umbilical artery smooth muscle cells.449 In Agarose-collagen blend bioinks, agarose provides long-term mechanical stability in culture while collagen provides microenvironmental cues for vasculogenesis. Kreimendahl and others found that blending 0.5% agarose with 0.5% collagen supported HUVEC assembly into capillary-like networks after 14 days without significantly affecting the printability of agarose.450 Blending agarose and collagen led to a significantly altered shear modulus compared to pure agarose or collagen and the shear modulus of AGR0.5COLL0.5 bioinks increased with increasing temperature. The microvascular networks contained lumens from around 5 to 30 μm in diameter and HUVECs were found to spread alongside collagen fibrils. This provides a promising platform to directly print microvasculature with agarose-based bioinks.</p><p>Besides blending, agarose may also be chemically modified for innate vasculogenic properties. Carboxylation of the agarose backbone can switch the secondary structure of agarose hydrogels from α-helix to β-sheet.451 Carboxylation diminishes hydrogen bonding between polymer chains and leads to less helical-helical interactions, resulting in softer hydrogels. The degree of carboxylation can be varied to tailor the mechanical properties of agarose hydrogels. Modifying soft, carboxylated agarose with RGD peptides led to apical polarization in HUVECs and lumen formation around 50–100 μm. RGD-functionalized carboxylated agarose (60%) matched the stiffness of fibrin clots (~0.5 kPa) and, upon injection, stimulated angiogenesis, capillary stabilization, and recruitment of CD11b+ myeloid and CD11b+/CD115+ monocytes in vivo.452 Carboxylated agarose was recently shown to be extrudable and possess shear-thinning properties depending on the degree of carboxylation.452 MSCs had higher viability when printed with carboxylated agarose bioinks compared to native agarose.453 However, the ability of this emergent material to promote microvascularization in a bioprinted construct has yet to be shown.</p><p>Hollow channels with diameters less than 200 μm have been fabricated within photolabile agarose hydrogels to guide neural cell migration454, but this technique has not been explored for vascular engineering. Further optimization of laser-assisted bioprinting approaches with agarose could enable high-resolution patterning of microvascular channels for endothelial cell growth within mechanically stable agarose-based hydrogels.</p><p>Overall, agarose-based bioinks are mostly used for improving the mechanical integrity of printed structures due to their stability at physiological temperatures. For direct bioprinting of microvasculature, agarose must be blended with bioactive materials or chemically modified with adhesive motifs. Indirect bioprinting of small-diameter vascular channels has been accomplished with agarose-based bioinks, but capillary-scale microvasculature remains a challenge. More systematic studies are needed to optimize agarose-based bioinks for high-resolution printing of microvasculature.</p><!><p>Alginate is a naturally derived anionic polysaccharide commercially obtained from brown algae (Phaeophyceae) and is structurally characterized as an unbranched linear copolymer of (1,4)-linked β-D-mannuronate (M) and α-L-guluronate (G) residues. Carboxylate groups in the G-blocks of alginate can be ionically crosslinked via divalent cations like Ca2+, Ba2+, Mg2+, and Sr2+ in an "egg-box" model as proposed by Morris et al.455 The properties of alginate hydrogels depend on the molecular weight and ratio of M and G blocks in the polymer chain, which vary depending on the source of the alginate.456 We refer readers to a review by Lee and Mooney in ref. 457 for more details on the properties of alginate.</p><p>Alginate has been extensively used as a bioink and can be adapted for several printing modalities, including DBB, EBB, and LAB, as reviewed in ref. 458. The rheological behavior of alginate solutions for bioprinting have been studied in detail.459 Alginate possess several features that are favorable for bioprinting. Alginate has excellent gelation kinetics and can be crosslinked instantly in the presence of divalent cations (e.g., calcium chloride). Sodium alginate is shear-thinning459,460, which reduces the shear stresses experienced by cells in alginate bioinks. The degradation profiles of alginate hydrogels can be tailored by periodate oxidation of uronic acid residues,461 or by gamma irradiation to modify alginate's molecular weight distribution.462 The viscosity of alginate bioinks mainly depends on alginate concentration, molecular weight, cell density, temperature, and pH. Our lab has systematically evaluated oxidized alginate bioinks for printability and cytocompatibility in droplet-based bioprinting (Figure 28).463</p><p>Xu and others have printed multiple cell types into complex heterogeneous tissue constructs containing microvascular networks using inkjet bioprinting.228 The cell types included human amniotic fluid-derived stem cells (hAFSCs), canine smooth muscle cells (dSMCs), and bovine aortic endothelial cells (bECs). Each cell type was suspended in a solution of calcium chloride (CaCl2) before being printed in a "pie" shape onto an alginate/collagen printing substrate. The cells have been printed in discrete regions and promoted tissue vascularization after 2 weeks subcutaneous implantation in mice. Importantly, printing cells in a specific pattern improved vascularization compared to manually seeded cells.</p><p>For extrusion-based bioprinting, Khalil and Sun were the first to explore a "fabrication window" for bioprinting viable ECs using alginate bioinks.464 Low wt.% alginate bioinks best support endothelial cell viability, but these hydrogels suffer from poor printability and mechanical stability during culture. Alginate can be crosslinked before printing to enable better shape fidelity after extrusion, but this may lead to deformities in printed constructs. To address this, alginate has been mixed with gelatin to improve its printability since gelatin is slightly viscous at room temperature.465 Gao and others have recently taken a systematic approach to optimize alginate-gelatin composite bioinks for solid freeform extrusion.466 Their study identified an important role for the loss tangent (G″/G') in balancing printability and cytocompatibility.</p><p>Coaxial bioprinting systems have become the most popular modality for printing vascularized constructs with alginate bioinks. Alginate is the most widely used biomaterial for coaxial bioprinting since it can be rapidly crosslinked with calcium chloride upon extrusion. Calcium chloride crosslinker solutions in the core or shell can allow for the extrusion of hollow or solid alginate microfibers. Alginate is often blended with proangiogenic biomaterials for coaxial bioprinting of vascularized constructs.</p><p>Besides coaxial extrusion, alginate bioinks have recently been utilized in embedded 3D bioprinting approaches. By including calcium chloride in the support matrix, alginate bioinks can be crosslinked rapidly during embedding. Cell-laden alginate bioinks have been printed with high resolution in Ca2+-containing gelatin supports via FRESH bioprinting.269,270 Wang and others recently used alginate as a sacrificial bioink to embed 3D vascular networks within prepolymers of agarose/gelatin and GelMA hydrogels.467 The alginate crosslinked within the Ca2+-loaded prepolymer, leaving behind a channel that could be washed out after solidifying the pre-polymer (Figure 29A). While hierarchical networks were fabricated with this approach (Figure 29B,C), the smallest diameter channel was around 400 um, which is much larger than capillaries (Figure 29D). Therefore, the pre-polymer support matrix would need to encourage angiogenesis from the parent channels for adequate microvascularization of the construct.</p><p>While alginate hydrogels are highly printable, bare alginate is bioinert and does not support vascular morphogenesis. Therefore, there have been numerous approaches to improving the bioactivity of alginate. Alginate bioinks can be loaded with proangiogenic growth factors to stimulate vascularization and controlled release from alginate improves the potency of VEGF when compared to adding it directly to bulk media.468 Laponite, a synthetic clay, has been blended with alginate to improve its printability and enable sustained release of growth factors.469 Faramarzi and others have developed patient-specific alginate-based bioinks by incorporating platelet-rich plasma (PRP) as an autologous source of angiogenic growth factors (Figure 30).470 The alginate/PRP bioink enhanced vascular network formation with HUVECs in vitro compared to native alginate, but this was only demonstrated in 2D. Branched vascular structures were also printed with the alginate/PRP bioink, but multilayer 3D organization of microvascular networks was not demonstrated. Nevertheless, using autologously sourced growth factors could broadly improve the translational potential of proangiogenic bioinks.</p><p>Besides growth factor supplementation, alginate can be modified with synthetic bioactive peptides like RGD to support cell adhesion. Injectable RGD-alginate hydrogels have been shown to promote endothelial cell adhesion and proliferation in vitro and angiogenesis in vivo.471,472 Torres and others found that injectable RGD-alginate microgels with higher M-to-G ratios can guide vascular morphogenesis in vitro and promote anastomosis with host vasculature in vivo due to lower crosslink density in the alginate matrix.473 The applications of proangiogenic alginate hydrogels in bioprinting have not been fully realized. RGD-alginate hydrogels could be useful for bioprinting microvasculature and their printability should be evaluated along with other peptide-modified alginates in future studies. This should be relatively straightforward, as certain material properties required for injection are also in agreement with extrusion (i.e. shear-thinning).</p><p>Overall, the printability and biocompatibility of alginate hydrogels make them an excellent biomaterial for bioprinting. The rapid gelation of alginate allows one to essentially blend their hydrogel of choice with alginate for improved gelation and shape fidelity during printing. However, alginate bioinks are mostly suitable for extrusion-based systems where resolutions below 100 μm are rarely achieved, necessitating angiogenesis in the construct after printing to generate microvasculature. New developments in photoreactive alginate may enable printing structures less than 100 μm using techniques like stereolithography.474 While there have been many studies using growth factors or bioactive peptides to promote angiogenesis in alginate-based hydrogels, their potential for bioprinting microvasculature has not been explored. Future investigations should evaluate the printability of peptide-functionalized alginate hydrogels to expand the design space of vasculogenic and angiogenic alginate bioinks.</p><!><p>Carbohydrate glass is formed by the solidification of melted sugar and sugar-alcohol solutions upon cooling. Isomalt, a sugar-alcohol commonly used in culinary applications, is the most common form of carbohydrate glass used for bioprinting. Sugar glass is an ideal sacrificial material because it is highly printable, water-soluble, and biocompatible. Furthermore, carbohydrate glass is cheap and readily available. Nanoscale glass fibers can be produced via melt-spinning sugar (cotton candy). These fibers have been embedded in PDMS to create a sacrificial microvascular network similar in size and density to capillaries.475 Sugar glasses can also be extruded as filaments if kept above their glass-transition temperature during printing. The filaments cool rapidly upon extrusion at room temperature and form stiff, brittle filaments. Due to their rapid curing, sugar glass filaments can be precisely printed as freestanding structures using model-guided design.476,477</p><p>In their seminal work, Chen and others used a sacrificial network of carbohydrate glass to pattern vascular networks within cell-laden hydrogels.242 Depending on nozzle travel speed, filaments between 200–1000 μm could be extruded and a variety of ECM biomaterials could be cast around the networks. The channels can be dissolved within minutes, leaving behind perfusable vascular networks within the construct. Importantly, perfusion of the networks with HUVECs led to endothelialized channels that exhibited angiogenic sprouting from the main channel into the bulk fibrin hydrogel. In a subsequent in vivo study, 3D printed carbohydrate glass networks within a fibrin bulk matrix (10 mg/mL) promoted angiogenesis and integration with host vasculature (Figure 31).478 This improved perfusion in animal models of hind limb ischemia and myocardial infarction. A disadvantage of these approaches was that the bulk ECM could only be cast homogenously instead of 3D printed, inhibiting precise placement of cells and ECM around the sacrificial network.</p><p>A related study investigating surgical anastomosis of vascular networks made by sacrificial sugar glass determined that the networks could withstand pulsatile flow as evidenced by Doppler perfusion in a hindlimb ischemic model.479 However, this was only confirmed up to 3 hours post-implantation. Vascularization of the surrounding PDMS bulk matrix was not demonstrated, likely due to the inability of PDMS to support cell adhesion and migration. More biocompatible hydrogels like alginate can be cast around sugar glass filaments loaded with calcium chloride.480</p><p>Overall, carbohydrate glass is an excellent fugitive material for indirectly patterning vascular networks within biocompatible hydrogels. To date, applications of carbohydrate glass in printing microvasculature have mostly used simple cylindrical structures >100 μm. Below these diameters, sugar glass is quite fragile and may break during casting. Future work should develop sugar glass formulations for printing smaller, more intricate microvascular networks that remain intact during fabrication. Furthermore, temporospatial printing of cell-laden hydrogels around sacrificial glass networks would be a more biomimetic approach to printing microvascularized matrices than homogenous hydrogel casting. Concerns about hyperglycemic response of suspended cells to the concentrated sugar solutions should also be addressed.</p><!><p>Hyaluronic acid (HA), or hyaluronan, is a non-sulphated glycosaminoglycan found ubiquitously in human ECM. It helps maintain the structure and fluid homeostasis in loose connective tissues.481 HA can influence cell morphogenesis by directly binding the cell surface hyaluronan receptor CD44, as well as other receptors.482 Furthermore, HA is highly hydrophilic and forms hydrated networks that permit intercellular signaling.483 The polymer structure of HA is a linear polysaccharide composed of repeating units of a disaccharide, β−1,4-D-glucuronic acid - β−1,3-N-acetyl-D-glucosamine. The molecular weight of HA ranges from 103 to 104 kDa depending on the source. Low molecular weight HA oligosaccharides stimulate endothelial cell proliferation, migration, and angiogenesis by binding CD44 receptors expressed by ECs.484 In physiological conditions, HA proteins take on an expanded random coil conformation and entangle to form continuous hydrated networks. Their conformation can be further influenced by HA-binding proteins known as hyaladherins.485</p><p>Physical gelation of HA yields fragile hydrogel networks that rapidly degrade. Therefore, many alternative crosslinking strategies have been developed to improve the stability of HA hydrogels, as reviewed in references 486 and 487. The carboxylic acid and N-acetyl groups of HA can be targeted for chemical modification by various chemistries. Methacrylated HA (MeHA), developed by Burdick and others, has been widely used to generate photopolymerizable HA hydrogels with tunable mechanical and degradation properties.488 MeHA hydrogels have been used to support self-renewal and differentiation of hESCs.489 Bioactive peptides like RGD can be conjugated to acrylated hyaluronic acid (AHA) for further control over cellular adhesion and migration in HA-based hydrogels.490,491 Gerecht's research group has pioneered the development of HA hydrogels to engineer microvascular networks from pluripotent stem cells in vitro (Figure 32). We refer readers to refs. 492–496 for more about hyaluronic acid hydrogels and their applications in engineering microvasculature.</p><p>Crosslinking HA with acrylated synthetic polymers can improve its rheological properties and suitability for bioprinting.497,498 Methacrylated collagen has also been combined with thiolated HA to develop a hybrid bioink for 3D bioprinting liver microenvironments.499 Vessel-like constructs have been bioprinted with extrudable MeHA and GelMA blend bioinks in a two-step UV crosslinking approach.500 Tetracylated PEG can be added to these blends to further improve their mechanical properties after printing.501 However, these bioinks have been limited to large-diameter vessel constructs.</p><p>The Burdick laboratory has developed HA bioinks based on guest-host supramolecular chemistry for embedded 3D bioprinting.271 Adamantane (Ad, guest) and β-cyclodextrin (CD, host) can be conjugated to HA and physically crosslink to form shear-thinning and self-healing hydrogels. This allows for shear-thinning guest-host HA bioinks to be printed directly into self-healing guest-host HA hydrogel matrices in a technique known as "GHost writing". During nozzle translation, Ad-HA and CD-HA interactions are disrupted but quickly recover to solidify around the extruded material in the wake of the nozzle. Microfibers with diameters as low as 35 μm can be printed with this method, depending on the nozzle size. Guest-host HA hydrogel bioinks can either be used as sacrificial or stabilized bioinks depending on their chemical modifications (Figure 33). Primary modification with guest and host functionalities yields a sacrificial bioinks while secondary modification with methacrylate groups allows for stabilization bioink stabilization after printing. Accordingly, perfusable microchannels can be fabricated by printing sacrificial guest-host HA bioinks within stabilized HA bioinks. These channels can last up to 30 days in culture, depending on the degree of methacrylation.502 Incorporating RGD peptides and protease-degradable crosslinkers into the support hydrogel for GHost writing can support robust capillary sprouting from parent channels.244 GHost writing with HA bioinks represents an outstanding emergent method to print high-resolution microvasculature with HA hydrogels.</p><p>Since methacrylated HA is photoreactive, it can also be printed using laser-assisted methods, which offer better resolution. Glycidal methacrylate-hyaluronic acid (GMHA) and GelMA bioinks have been used to print an iPSC-derived vascularized hepatic lobule model with DLP-SLA.503 This method used an interesting approach whereby digital masks were applied in a two-step sequential manner to print a first layer of hiPSC-derived hepatic cells in 5% (w/v) GelMA followed by a layer of endothelial and mesenchymal supporting cells (HUVECs and ADSCs) in 2.5% GelMA and 1% GMHA. The HUVECs formed microvascular networks after 7 days of culture and the triculture model significantly enhanced liver-specific gene expression and functions in 3D. In another study, Zhu and others used DLP-based microscale continuous optical bioprinting (μCOB) to create prevascularized constructs containing a "base" layer of HepG2 cells in 5% (w/v) GelMA and a "vascular" layer of HUVECs and 10T1/2 cells in 2.5% GelMA and 1% HA.504 LAP (lithium phenyl-2,4,6-trimethylbenzoylphosphinate) was used as a cytocompatible photoinitiator. By using different digital masks, heterogeneous vascularized constructs could be printed with gradient channel widths regionally controlled biomaterials properties. Encapsulated HUVECs formed lumen-like structures and microvascular networks after 1 week of in vitro culture. Furthermore, the prevascularized networks could functionally anastomose with murine host vasculature after two-week subcutaneous implantation. Non-prevascularized constructs were shown to integrate poorly with host vasculature.</p><p>The versatility and biocompatibility of hyaluronic acid makes it a useful biomaterial for bioprinting microvasculature. The mechanical and bioactive properties of HA hydrogels can be finely tuned through various chemical methods, making HA an excellent canvas biomaterial for engineering "semi-synthetic" bioinks with highly controllable microenvironments. There have been impressive platforms developed recently to promote human microvascular network formation in engineered HA hydrogels.494 These platforms can potentially be adapted for the development of bioinks with highly defined proangiogenic microenvironments.</p><p>This section has reviewed polysaccharide-based naturally derived hydrogels for bioprinting microvasculature. These hydrogels are summarized in Table 3. Synthetic hydrogels will be reviewed in the following section along with their applications in bioprinting microvasculature.</p><!><p>Synthetic hydrogels are based on hydrated networks of polymers synthesized using chemical methods. The chemical and physical properties of synthetic polymers can be tightly controlled depending on the monomers used and the nature of their crosslinking. Synthetic hydrogels are known for their inability to support cell adhesion but can be functionalized with bioactive peptides for cell-mediated adhesion and degradation.505,506 This makes synthetic hydrogels excellent canvas materials for user-defined functionalities. Synthetic hydrogels are highly printable since their physicochemical properties can be tailored to meet the printability requirements of a given technique. In this section we will review common synthetic hydrogel bioinks applied in bioprinting microvasculature.</p><!><p>Poly(ethylene) glycol, or PEG, is a popular synthetic biomaterial for a variety of biomedical applications. PEG is a linear polyether compound and is favored for its hydrophilicity and resistance to protein and cell adsorption. PEG is generally biocompatible and elicits minimal immune response, though there is some emerging evidence demonstrating anti-PEG antibodies produced in rodents.507 PEG is often modified with diacrylate (DA) or methacrylate (MA) groups for free-radical polymerization. This affords PEG hydrogels with highly tunable and defined mechanical properties. Importantly, PEG is non-toxic and other biomaterials that have limited tunability can be "PEGylated" or crosslinked with PEG monomers to tailor their physical properties.</p><p>PEG hydrogels do not possess cell attachment sites and are non-biodegradable. Therefore, various strategies have been developed to engineer PEG hydrogels with proangiogenic properties. Growth factors like VEGF, bFGF, and PDGF as well as proangiogenic signaling ligands like EphrinA1 can be covalently bound to PEGDA hydrogels to promote EC migration and tubulogenesis.508–510 Protease-sensitive degradation sites can also be engineered into multi-arm PEG matrices for cell-mediated matrix remodeling and self-assembly into vascular networks Figure 34.511–513 Interestingly, conjugating the VEGF-mimetic peptide QK to PEG hydrogels significantly enhances vascular morphogenesis compared to PEG-VEGF or VEGF alone.514 Multi-arm PEGs offer numerous sites for conjugating different bioactive substrates, enabling the engineering of complex proangiogenic microenvironments within PEG hydrogels.</p><p>PEG hydrogels have been used in various bioprinting approaches. For EBB, PEGs are mostly used to tune the mechanical integrity of extrudable bioinks. Multi-arm PEGs can increase the shear modulus of hydrogel bioinks due to higher crosslinking density. Co-crosslinking multi-arm PEGs with thiolated hyaluronic acid and gelatin can enable 3D printing of vessel-like constructs that support cell viability for up to four weeks.501 Adding four-armed PEG acrylate (PEG-4A) to methacrylated gelatin and methacrylated hyaluronic acid cryogels improved their mechanical properties and supported capillary network formation in a coculture of ADSCs and HUVECs.515</p><p>In a microfluidics-assisted bioprinting approach, bioinks containing PEG monoacrylate-fibrinogen (PF) and alginate were used to print multicellular lattice constructs of varying geometries.516 The role of alginate was to immediately crosslink the bioink to preserve its shape during printing, while PF could be covalently crosslinked after printing for long-term stability as well as promoting cell attachment. Alginate could be removed by EDTA after printing, leaving behind pure PF fibers around 100 μm in diameter. After 7 days, especially in Janus constructs, HUVECs migrated to the periphery of the fibers and formed vascular-like tubes with lumens around 150 μm. These vascular networks were able to anastomose with host vasculature in mice after subcutaneous implantation for 15 days.</p><p>Most of the studies using PEG biomaterials employ crosslinking methods that require UV exposure. More biocompatible crosslinking strategies are necessary to maximize cell viability in PEGylated hydrogels. Rutz et al have developed a PEGX toolkit to manipulate the properties of PEG-based bioinks before and after printing (Figure 35).256 In the PEGX method, PEG is functionalized with reactive groups on both ends which represent the "X". PEGX can then be used to crosslink a variety of polymers with a diverse selection of chemical methods. The mechanical properties of the resulting gel can be tuned depending on the concentration of PEGX as well as its molecular weight and display of functional groups. For cell-encapsulating PEGX bioinks, cytocompatible crosslinking chemistries like click chemistry and Michael-type additions can be used. In a recent study, Rutz and others used Thiol Michael type addition and tetrazine-norbornene click chemistry to tailor post printing mechanical properties and cell viability in gelatin-based bioinks.518 HUVECs can be printed using the PEGX method but vascular morphogenesis has not been demonstrated. Nevertheless, PEGX bioinks could offer great flexibility in expanding the printability of otherwise poorly printable hydrogels like gelatin. It would be interesting to investigate vascular morphogenesis in PEGX systems that place more consideration on angiogenesis in the future. For example, protease-sensitive proangiogenic PEG hydrogels like those described in the previous paragraph could theoretically be developed into a bioink for direct bioprinting of microvasculature with the PEGX method.</p><p>Most PEG hydrogels are modified for UV-based crosslinking, which makes them amenable to laser-assisted bioprinting approaches. Zhang and others have patterned biomimetic 3D capillary structures in PEGDA hydrogels using projection-based stereolithography.319 In a more recent study, Grigoryan et al fabricated vascular structures with unprecedented complexity in PEGDA hydrogels using a custom projection stereolithography apparatus for tissue engineering (SLATE). These structures included entangled multivascular networks with intravascular features that mimic the sophistication of in vivo vasculature. A voxel resolution of 50 μm was achieved using cytocompatible tartrazine as a photoabsorber. Furthermore, the vascular networks could be endothelialized and incorporated into a hydrogel carrier to support engraftment of hepatic aggregates in nude mice. Though the networks generated with this approach were highly sophisticated in form, their lumen diameter was above 100 μm and therefore did not recapitulate capillaries. The 20% (w/v) PEGDA hydrogel formulation would also be much too stiff to promote angiogenesis from the parent vascular network. It is unclear whether the light blocking additives would allow for the fabrication of patent capillary-like networks. Incorporation of the identified light blockers into a more proangiogenic hydrogel may facilitate the formation of capillary networks in vitro and in vivo using this approach.</p><p>PEG hydrogels are useful biomaterials for direct writing via LAB. Culver and others used two-photon laser scanning lithography to pattern immobilized bioactive peptides in a PEG scaffold.519 In this study, acrylated PEG was functionalized with a matrix metalloproteinase peptide (GGPQGIWGQGK, abbreviated PQ) for cell-mediated degradation and an RGDS peptide for cell adhesion. Both peptides were acrylate-terminated and a photoinitiator was incorporated into the PEG matrix so that, upon excitation with a tightly focused two-photon laser, site-specific immobilization could take place via addition polymerization. Unbound peptides could then be washed away, leaving behind a precisely organized 3D pattern of immobilized biomolecules with features as small as 5 μm in the axial direction. This strategy was used to pattern biomimetic features derived from various tissues (e.g. cerebral cortex capillaries) that could be used to pattern the organization and tubule structures of HUVECs.</p><p>In a more recent study, Brandenberg and Lutolf used focalized short-pulsed lasers for in situ bioprinting of microfluidic networks in 3D cellularized PEG and collagen hydrogels.300 Protease-sensitive and integrin-binding peptides were functionalized onto the PEG hydrogel to promote cell-mediated degradability in the bulk hydrogel. Importantly, cells as close as 20 μm away from the pulsed laser were still viable, highlighting the cytocompatibility of this technique. However, the use of photoablation with high-intensity lasers still raises concerns over structural integrity as it induces nonspecific chemical bond photolysis and microcavitation. To address this, Arakawa and others used a multiphoton laser with cytocompatible wavelength and intensity for 4D patterning of microvasculature and cell-instructive ligand presentation in PEG hydrogels (Figure 36).316 In this study, a diazide-modified synthetic peptide was used to crosslink a PEG-tetrabicyclononyne hydrogel via strain-promoted azide-alkyne cycloaddition (SPACC). The synthetic peptide contained a photodegradable ortho-nitrobenzyl linker (oNB) that could be degraded by pulsed near-infrared light to generate multiscale vascular networks as small as 10 μm in diameter. This circumvents the use of high-intensity laser for photoablation, which is potentially harmful to cells. The synthetic peptide crosslinker also contained RGD and protease-cleavable sequences to promote cell adhesion and remodeling of the hydrogel matrix (Figure 36A). Perfusable, hierarchical 3D vascular networks ranging from 300 μm to 25 μm in diameter were patterned within the hydrogel (Figure B,C) as well as biomimetic capillary structures (Figure 36D). Endothelialization of patterned microchannels with cross-sections of 60 μm × 60 μm (Figure E,F) and 45 μm × 45 μm was obtained after perfusion with HUVECs, which is currently the smallest endothelialized synthetic vessel generated within a cytocompatible biomaterial (Figure 36G,H).316 These channels can be fabricated in the presence of stromal cells to fabricate multicellular tissues with multiscale vasculature, highlighting the potential of this approach to produce functional, heterogeneous, capillarized tissues.</p><!><p>Poloxamers, commonly known by their trade name Pluronics®, are tri-block copolymers composed of hydrophilic poly(ethylene oxide) and hydrophobic poly(propylene oxide) in linear alternating PEO-PPO-PEO blocks. Gelation of poloxamers is thermoreversible and they form micellar liquids below their sol-gel point (~20°C) and gels at physiological temperature via physical aggregation of micelles.520 The exact sol-gel point of poloxamers depends on polymer concentration, with higher concentrations having lower sol-gel temperatures.521 Poloxamers have commonly been used as a controlled release vehicle for hydrophobic drugs.522 They have also been used as a wound dressing.523 Poloxamer 407, or Pluronic® F-127, is a poloxamer that is FDA-approved for use in humans and is the most commonly used poloxamer biomaterial.</p><p>While Pluronics are generally cytocompatible, they are not ideal for cell encapsulation. Concentrations starting at 10% (w/w) have significant negative effects on cell viability.524 This is problematic, as the minimum concentration of Pluronic F127 needed to form a gel in mammalian cell culture medium is 14%.525 Encapsulation of HMVECs in 15.6% (w/w) F127 gels led to complete cell death in 5 days.524 Membrane-stabilizing agents (hydrocortisone, glucose, and glycerol) can rescue cell viability in F127 gels, but may have inadvertent effects on cell function.</p><p>Poloxamers are frequently used as sacrificial bioinks due to their thermoreversible gelation. Jennifer Lewis's group has made major contributions to bioprinting perfusable vascular networks using Pluronic F127 bioinks.241 In their foundational work, they used Pluronics to pattern omnidirectional microvascular networks within a physical hydrogel.241 Photopolymerizable diacrylate-functionalized Pluronic F127 was used as a fluid reservoir to fill in voids created in the physical F127 gel substrate by nozzle translation. Nozzle size and printing pressure could be varied to create channels as small as 150 μm in diameter that could be perfused after UV curing of Pluronic F127-diacrylate from the fluid reservoir. In a subsequent study, Kolesky and others used Pluronic F127 to pattern fugitive channels within vascularized heterogeneous cell-laden GelMA hydrogels.526 In a later report, Kolesky used Pluronic F127 to pattern perfusable vascular networks within constructs several millimeters thick.148 In this study, a multi-layer vascular lattice was first printed with Pluronic F127 "vascular ink" and a gelatin "cell ink". Then, a GelMA/fibrin hydrogel was cast around the lattice. The fugitive Pluronic bioink was washed out and the open channels were seeded with HUVECs to form an endothelialized channel around 200 μm in diameter after two days (Figure 37). The networks could be perfused and support the viability of cells throughout the large construct. However, angiogenic sprouting from parent channels was not demonstrated, even after 45 days of perfusion. Furthermore, casting of the bulk hydrogel around the sacrificial network limits heterogeneity of the final construct. In another study, Millik et al used coaxial extrusion of unmodified Pluronic F127 in the core and Pluronic F127-bisurethane methacrylate (F127-BUM) in the shell to fabricate hollow tubes with diameters as low as 150 μm.253 Adding collagen to the F127-BUM bioink enabled cell adhesion to the luminal surfaces of the tubes and promoted monolayer formation during in vitro culture.</p><p>The previously mentioned studies did not demonstrate sprouting of ECs from the main channels, and therefore did not recreate proper hierarchical vascular networks with microchannels <100 μm. This is mostly due to the resolution limits imposed by extrusion. To integrate capillary-scale networks, Jacoby and others fused a dense "fluff" of melt-spun shellac microfibers (5–500 um) with networks of manually extruded Pluronic F127.527 After sacrificing the materials, ECs and SMCs could be perfused throughout, self-assembling into hierarchical structures resembling arterioles, venules, and a capillary bed. This study demonstrated how sacrificial networks printed with poloxamers can be complemented by alternative biofabrication approaches to integrate capillary-scale features.</p><p>Poloxamer bioinks are highly printable and can be used to fabricate sacrificial vascular networks with excellent shape fidelity. However, poloxamers are currently only suitable for extrusion-based bioprinting, which is limited in resolution as has been previously discussed. Therefore, bioprinting capillary networks with sacrificial poloxamer bioinks is not currently feasible. However, patterned vasculature printed with poloxamer bioinks can be surrounded by proangiogenic hydrogels to promote capillary sprouting from the fugitive network. Another drawback of using poloxamer bioinks is the need to use subphysiological temperatures to liquefy and evacuate the sacrificial poloxamer network.</p><p>In this section, we have reviewed synthetic hydrogels for bioprinting microvasculature. Synthetic hydrogel bioinks for bioprinting microvasculature are summarized in Table 4.</p><p>We have now critically analyzed the biomaterials currently available for bioprinting microvasculature. We have focused on individual biomaterials (and blends) in each section and how they are formulated into printable bioinks for fabricating microchannels and microvascular networks in vitro and in vivo. Sections 3 and 4 have provided a thorough analysis of techniques and biomaterials used to bioprint microvasculature. An important aspect of these technologies that has not yet been discussed yet is how they are applied. In the next section we will review how bioprinted microvasculature can be applied for disease modeling and drug testing as well as tissue engineering and regenerative medicine.</p><!><p>Applications of bioprinting microvasculature have improved the quality and scope of existing disease modeling and tissue regeneration methods. In this section, we begin by providing examples of pathophysiological models of several tissue types that incorporate bioprinted microvasculature and move to examples of bioprinted microvasculature for tissue engineered constructs intended for regenerative therapies. In our discussions, we include the advantages of incorporating microvasculature and how microvascular structures can improve the biomimicry and efficacy of constructs intended for disease modeling, drug testing, and regeneration strategies.</p><!><p>Conventional in vitro tissue models have until recently remained in the realm of 2D structures and microfluidic devices. These models have proven to be useful in the collection of inexpensive data but fall short in recapitulating complex and physiologically relevant tissues. Three-dimensional structures, however, can better represent in vivo environments in which there are complex interactions, such as those between cells, growth factors, and the extracellular matrix.528 Additionally, traditional microengineering methods have limitations in the use of multiple cell types and ECM environments in unique spatial arrangements that mimic in vivo conditions.529 Bioprinting can be used to position biomaterials and cells in precise positions, while maintaining control over various spatiotemporal elements in the 3D structures.36 Therefore, bioprinting is an enticing technique for applications in in vitro disease modeling and drug testing studies.</p><p>The current unmet need to develop effective in vitro models for various pathologies lies in the fact that animal models are inadequate in representing human diseases.530 Additionally, existing disease models such as 2D cell culture and macroscale hydrogels do not sufficiently recapitulate pathophysiological conditions.531 The bioprinted tissue-specific 3D models can facilitate drug testing. Approximately 25% of all drugs that are withdrawn from clinical trials are attributed to toxicity and pharmacokinetics.532 The development of in vitro models that better mimic in vivo conditions is essential to advance drug development and testing.533 One significant feature that can be introduced to the advanced 3D tissue models for drug testing is microvasculature; a vascularized component in models will permit the biomimetic transportation of nutrients, oxygen, and drugs throughout the construct.534 To date, various bioprinted constructs for pathophysiological modeling and drug testing have been developed to introduce microvasculature and simulate the conditions of cardiac, lung, liver, kidney, intestinal, placental, vascular, and cancer tissues. These are summarized in Figure 38 and will be reviewed in the follow sections.</p><!><p>The myocardial tissue of the heart is made of cardiomyocytes that are uniquely aligned to exhibit electrical and mechanical functions necessary for contraction of the heart.535 The creation of cardiac tissue models is met with several challenges. In terms of constructing physiologically relevant cardiac models, spatial control over cardiomyocytes and the 3D architecture is necessary to completely recapitulate aspects of native myocardium such as signal propagation and cardiomyocyte contraction.536 Additionally, to create thick constructs for cardiac tissue models, it is necessary to introduce microvasculature to promote nutrient diffusion and waste removal throughout the models.537 To meet these challenges, bioprinting has proven to be a useful technique to both implement microvasculature and mimic the organization of native cardiac tissue.36,538</p><p>In one study, Zhang et. al created multilayer microfibrous structures utilizing bioprinted microvasculature to model cardiac tissues (Figure 38A).539 The group used neonatal rat cardiomyocytes seeded onto a bioprinted endothelialized microfibrous scaffold. The subsequent endothelialized myocardial structure demonstrated uniform beating that lasted up to at least 2 weeks while undergoing perfusion culture. This bioprinting approach to cardiac modeling serves several advantages, including control of scaffold parameters, the ability to create scaffolds with multiple cell types, and further applications of the endothelialized structures outside the study of cardiac tissues. Furthermore, the model can be translated to human studies in a step toward personalized medicine using a hiPSC-derived cardiomyocyte (hiPSC-CM) model, which was proven by using hiPSC-CMs as the source of cardiomyocytes rather than neonatal rat cardiomyocytes. The model in this case demonstrated uniform and synchronized beating, much like the prototype with the neonatal rat cardiomyocytes.</p><p>Studies on the effects of drugs on cardiovascular functions are of high priority due to the fact that cardiac safety is the leading cause for the discontinuation of drugs.540 Zhang et al. additionally used their endothelial myocardial model described above to test the toxicity of several drugs by combining their microfibrous model with a bioreactor.539 The group found that their model was able to predict cardiovascular drug toxicity, as proven by dose-dependent responses when treated with the anti-cancer drug doxorubicin. Following exposure to doxorubicin, both the beating rate of cardiomyocytes the secretion of vWF by ECs decreased. The response to doxorubicin, which has known cardiovascular toxicity effects,541 demonstrates that their model shows promise for testing potential cardiotoxicity of other drugs. Furthermore, their model demonstrates potential in the field of personalized medicine since hiPSC-CMs seeded onto the scaffolds showed dose-dependent responses to doxorubicin.</p><!><p>Blood is oxygenated via respiration and diffusion of inhaled air from the lung alveoli to the pulmonary capillaries.542 Capillary networks make up a large portion of the lungs, with ECs covering a surface area of approximately 130 m2.543 In lung models, it is important that both the lung tissue and the endothelial networks are recapitulated due to the close association between the alveoli and capillary networks in the diffusion of oxygen and removal of carbon dioxide.544 In vitro lung modeling has consisted of simulations of the alveolar and capillary interface on microfluidic devices544 that simulate breathing conditions through cyclic positive and negative pressure loops.545 Bioprinting has emerged as a useful tool for creating improved and more tunable 3D airway models, specifically through their advanced simulation of the interface between alveolar and capillary tissues.436,545</p><p>A vascularized airway-on-a-chip model was made through a bioprinted dECM bioink laden with ECs and lung fibroblasts (Figure 38B).436 The construct included a vascular platform consisting of bioprinted ECs that organized into an interconnected vascular network after 7 days. In the model, endothelial cell orientation was responsive to shear stress and inflammatory mediators. Additionally, the model was able to recapitulate the epithelium-blood interaction of the physiological airway. Park et al. also applied their lung model for disease modeling.436 In this system, they utilized their bioprinted airway-on-a-chip to create an asthmatic airway epithelium model that responded with increased mucus secretion. Furthermore, they used their bioprinted design to create an asthma exacerbation model to mimic environmental exposures that contribute to chronic inflammation.436</p><!><p>The major purpose of the liver is detoxification and metabolism of foreign substances.549 Recapitulation of native liver functions by engineered models relies heavily on facilitating proper interactions between hepatocytes and supporting cell structures. Since these interactions rely on 3D assembly of the cell types, bioprinting is a useful tool in engineering liver models. 503 In the native liver, microvasculature is necessary for the execution of detoxification; therefore, liver models can become more physiologically relevant when microvasculature is included. Several studies have demonstrated that bioprinting can aid in the incorporation of microvasculature to fabricate advanced liver models.534,550</p><p>Through controlling the placement of hiPSC-derived hepatic progenitor cells (hiPSC-HPCs) in culture with HUVECs and adipose-derived stem cells, hepatic lobule constructs have been fabricated (Figure 38C).503 The inclusion of controlled geometries and materials and multiple cell types through bioprinting led to enhanced function of the hiPSC-HPCs. Furthermore, this bioprinted model was specifically advanced since it was created with hiPSC-derived hepatic cells and, therefore, makes strides towards applications as a patient-specific model.</p><p>Drug testing in liver models with microvasculature is of specific importance due to the role the liver plays in the breakdown of drugs and the fact that liver damage caused by drug toxicity is a major concern in drug development.549 Bioprinting has been implemented to create 3D liver tissue chips capable of high-throughput drug testing by sandwiching a layer of HepG2 cells between two layers of HUVECs.550 In another study, Massa et al. used bioprinting of microvasculature to study drug toxicity in a liver model.534 The group printed microchannels by printing a sacrificial agarose fiber in a cell-laden GelMA hydrogel and then seeding the channel with ECs, creating endothelialized microchannels within a 3D liver model. They applied this technique to drug toxicity screening and showed that the implementation of the microvascular channel within the liver model delayed the permeability of molecules into the construct and demonstrated a protective role, both of which are physiologically relevant characteristics of the microvasculature in the liver. Additionally, the incorporation of the endothelialized channel led to higher viability of the HepG2/C3A cells.</p><!><p>The kidney serves the important role in the human body of filtering solutes in the blood through the interactions between renal compartments and a vascular network.551 Due to the precise structural interaction requirements of kidneys, 3D models are helpful in recapitulating kidney functions. For instance, 3D printing of microfluidic chambers has been used in conjugation with kidney organoids to facilitate the formation of vascular networks.551 Furthermore, 3D bioprinting of microvasculature with endothelial cell-laden bioinks within renal constructs has emerged as an attractive technology for disease modeling and regenerating kidney functions.552</p><p>To improve upon current kidney-on-chip models that do not have substantial 3D architecture, Lin et al. made use of bioprinting microvasculature to study renal reabsorption in 3D kidney proximal tubule model (Figure 38D).546 Using a fugitive ink, they created ~200 μm channels seeded with glomerular microvascular endothelial cells to create a microvascular structure adjacent to a 3D bioprinted kidney epithelium. Proximal tubule epithelial cells printed along the microvascular structures demonstrated selective uptake of albumin and reabsorption rates of glucose that could be compared to in vivo values. Additionally, they demonstrated that their bioprinting model could be used to study diabetes through the induction of hyperglycemic conditions.</p><!><p>The functions of the human intestine include digestion of ingested foods, nutrient absorption, and pathogen defense553,554. To facilitate these functions, a complex structure is found in intestinal tissue.554 Intestines are made of multiple cell types along with symbiotic microbes that come together to create a 3D structure consisting of villi and crypts.555 Therefore, to model intestinal tissue, 3D constructs are needed to provide more accurate recapitulations of physiological conditions. Specifically, there is a need for models that have control of geometry and architecture, which can be achieved through bioprinting.</p><p>For proper function of the intestine, capillary systems are vital. The vessels serve the purpose of absorption and transportation of nutrients and drugs that come into contact with the intestinal epithelium.556 Therefore, physiologically relevant intestinal models and gut-on-a-chip models have benefited from the introduction of microvasculature, specifically through bioprinting.374 In an application in which a shell of colon epithelial cells and a vascular core of HUVECs were printed to create an intestinal villi model, Kim et al. conducted simulation of the epithelial barrier function of the intestine (Figure 38E).374 This was achieved by measuring permeability and glucose absorption. Permeability and glucose uptake were highest in the structures printed with the vascular network, indicating that the vascularization created by bioprinting play a large role in the reproduction of the physiology of the human intestine.</p><!><p>The female reproductive system is unique in the fact that the development of new blood vessels does not follow the typical pattern of being in a quiescent state as is the case throughout almost all tissues in the adult human body. In the female reproductive system, angiogenesis occurs regularly and cyclically during the menstrual cycle and to maintain pregnancy.557 Additionally, vascular remodeling is an important aspect in preparation for embryo implantation and placentation.558</p><p>The placenta is vital in providing nutrients to a developing fetus. Understanding the physiology of the human placenta is necessary to realize the source of pregnancy complications.559 The placenta is developed in part from trophoblasts that provide nutrients to the early fetus, and understanding the interactions between trophoblasts and ECs can provide insight into irregular trophoblast invasion that can lead to pregnancy complications.547 Current in vitro models have not been able to examine trophoblasts and ECs. However, bioprinting using GelMA hydrogels laden with HUVECs and HTR8 extravillous trophoblasts allowed Kuo et al. to create a placental model capable of examining trophoblast-endothelium interactions (Figure 38F).547 HUVECs printed in the model demonstrated outgrowth and network formation that was later impaired by the incorporation of trophoblasts. The direct co-culture of HUVECs and trophoblasts in a dynamic environment showed that this placenta model could be used to help researchers in their understanding of the interactions between trophoblasts and the endothelium, which shows potential for gaining insights concerning preeclampsia and other human reproductive pathologies.</p><!><p>The endothelium of blood vessels in the body serve as the primary interface between blood and tissues and serves an important function in controlling the movement of nutrients and soluble factors in the blood.560 Modeling of the vascular endothelium is important in understanding various pathologies, including tissue overgrowth and cancer and other vascular abnormalities that occur in the cardiovascular, neurovascular, and musculoskeletal system.561 Several fabrication techniques have been used for modeling various vascular parameters and creating vascular systems. For instance, standard photolithography methods have been used to model vascular properties that become abnormal during disease states while recapitulating physiologically relevant properties of the endothelial barrier such as intimal stiffness, self-deposition of the basement membrane, and self-healing of the endothelial barrier.562 Additionally, 3D printing has been used to create a template within a GelMA hydrogel construct containing 10T1/2 cells and a lining of ECs.563 Bioprinting, however, has demonstrated the ability to improve techniques in vascular modeling due to the capacity to create structures with different cell types and materials and overcome the challenge of fabricating complex vascular models that exists in current free-standing microfluidic models.561</p><p>Using bioprinted microvasculature on the scale of ~500 μm, Gao et al. modeled various vascular parameters (Figure 38G).439 In this model, they were able to observe permeability, antiplatelet adhesive effects, response to shear stress, and microvessel sprouting. Following formation of their bioprinted vascular model, Gao et al. modeled pathological changes in response to airway inflammatory stimulation.</p><p>In silico modeling has also shown potential in the applications of bioprinting microvasculature.339,564 In two studies, Wang and colleagues modeled the fusion of cell aggregates in the bioprinting of several tissue architectures. The group simulated layer-by-layer deposition of scaffold-free cell spheroids to fabricate various tissue geometries, including a bifurcated vascular junction. Through their model, they showed that computer-aided design tools can be used to gain insight into the bioprinting process and ideally be implemented for use in bioprinting vascular networks.339 In a second study, the group looked at the fusion of cell spheroids to fabricate vascular networks through kinetic Monte Carlo simulations.564 They relied on parameters pertaining to cell-cell and cell-medium interactions to create a model to simulate the fusion and cell sorting that would occur during bioprinting of vascular networks.564</p><!><p>Effective recapitulation of the tumor microenvironment, including chemical and biophysical cues, is necessary for the fabrication of cancer models.531 3D bioprinting have been used to fabricate spatially precise and complex constructs with the structures necessary for mimicking tumor environments.565 One particular structure that should be introduced in cancer models is microvasculature due to the critical roles it plays in tumor survival and metastasis.566</p><p>To study glioblastoma (GBM) and the roles angiogenesis plays in the development of GBM, Wang et al. developed a tumor model consisted of bioprinted glioma stem cells.567 Glioma stem cells were printed into a grid structure with porous channels using an alginate-gelatin-fibrinogen hydrogel. The stem cells formed tubular networks across the pores and expressed higher amounts of angiogenesis-related genes versus those from the suspension culture. This observation demonstrates that the 3D bioprinted structure recapitulates the tumor microenvironment more closely and has the potential to be used for the study of the roles glioma stem cells play in tumor angiogenesis. Furthermore, these bioprinted models could be used in research for anti-tumor angiogenesis therapies.</p><p>Bioprinting was also used by Meng et al. to build vascularized tumor constructs to study characteristics of the tumor microenvironment in metastasis (Figure 38H).548 In this study, bioprinting was especially useful in manipulating chemical gradients as well as the placement of cell types, including tumor cells, stromal cells, and infused vascular cells. The metastatic tumor models consisted of a primary tumor cell droplet, endothelialized microchannel, a fibroblast-laden hydrogel to serve as the stroma, and release capsules serving as gradients of chemotactic agents. The models were capable of recapitulating mechanisms of tumor spreading and were additionally used for anticancer drug screening.</p><!><p>With the ultimate goal of engineering whole replacement tissue and organs, microvasculature is an essential step in ensuring nutrient diffusion in complex and large-scale tissue.568 Additionally, microvasculature play a critical role for proper vascular integration of therapeutic tissue engineered constructs in vivo,.569 Various challenges persist in the fields of tissue engineering and regenerative medicine in terms of the implementation and engineering of microvasculature. These include facilitating proangiogenic interactions between cells and 3D matrices,570 creating hierarchical vascular architectures that meet diffusion requirements, fabricating capillary-size vasculature,569 achieving functional anastomosis in vivo.569571 Importantly, bioprinting have demonstrated the ability to conquer many of these challenges and advance tissue engineered structures through the introduction of microvasculature. In this section, we will review and evaluate different microvascularized tissue models created by bioprinting. These models are summarized in Figure 39.</p><!><p>Bone tissue is comprised of bone matrix and vascularized tissue that serve several functions in the body, including providing structural support, protecting internal organs, and holding cells responsible for hematopoiesis with marrow.572 The vascular network within bone is especially important in supplying nutrients to cells within the bone matrix and removing waste.573 Additionally, microvasculature is necessary for providing bone with hormones, growth factors, and neurotransmitters, as well as controlling hematopoiesis.574 Proper bone function relies on interactions between both vascularized and bone matrix regions of bone tissue. Therefore, it is vital that both regions are recapitulated in tissue engineered constructs. Bioprinting has demonstrated potential in printing multi-element bone structures.575 For example, bioprinting has been used to position HUVEC spheroids and calcium phosphate into GelMA hydrogels create constructs that have an osteogenic outer layer and angiogenic inner layer.576 Furthermore, bioprinting of microvasculature can be applied to fabricate vascularized grafts, which is one of the largest challenges facing fabrication of constructs for the treatment of large bone defects.398</p><p>Bioprinting 3D structures for bone tissue engineering has been implemented to induce prevascularization of bone constructs. For instance, sequential seeding of hSMCs followed by HUVECs on porous 3D-printed Hyperelastic Bone (HB) scaffolds promoted the formation of vascular structures across 3D-printed fibers of the HB scaffolds.577 HUVECs cultured on these scaffolds expressed large amounts of endothelial cell-related genes and microscale tube-like structures were seen in HB scaffolds with pores up to 1000 μm.577 Additionally, seeding of HUVECs and MSCs onto 3D printed scaffolds of calcium phosphate cement and alginate-gellan gum preloaded with VEGF promoted the formation of tubular networks.578 When implanted into bone defects in rat femurs, the printed scaffolds promoted infiltration of microvasculature into the wound area.578</p><p>Utilization of bioinks laden with ECs has also proven to be effective in bone tissue engineering and regeneration. Bone-like 3D structures have been created through bioprinting of cylindrical rods with two different GelMA bioinks to create osteogenic and vasculogenic niches.418 This approach to fabricating bone-like structures shows promise in applications such as the treatment of large bone defects. Printing of HUVEC-laden hydrogels has also proven useful in bioprinting hybrid scaffolds that consist of bioprinted microvasculature.150 Through printing a HUVEC-laden gel into the pores of a PDACS/PCL scaffold laden with Wharton's jelly mesenchymal stem cells (WJMSCs), both osteogenesis and angiogenesis can be enhanced. Although they did not show any in vivo application of the bioprinted scaffold, the proof of promotion of both angiogenesis and osteogenesis demonstrates that their methods hold potential in future regeneration of bone defects.</p><p>Biphasic bioprinted structures have also been used to mimic the native osteons of cortical bone.398 Cortical bone consists osteons in which osteogenic cell types are organized into lamellae and vasculogenic cell types create branched Haversian systems that run through the centers of the lamellae throughout the cortical bone.572 Piard et al. used bioprinting of microvasculature to recapitulate the osteogenic and angiogenic potential of cortical bone osteons to enhance neovascularization.398 The group recreated the Haversian canals of cortical bone using a HUVEC-laden bioink printed in the center of a fibrin bioink containing hMSCs and demonstrated the angiogenic potential of the scaffolds through in vivo studies, proving the potential use of biphasic osteon-mimicking scaffolds for vascularized bone tissue engineering. An additional application of bioprinted microvasculature has been conducted to recapitulate the complex, hierarchical architecture of bone through printing a soft organic bioink into a hard, mineral structure (Figure 39A).579 The structure itself promoted the ECs within the GelMA hydrogels to form capillary-like networks with lumen-like channels. When functionalized with VEGF- and BMP2-mimetic peptides, enhanced angiogenic and osteogenic potential was observed. Furthermore, dynamic culture of the constructs in which the vascular channels were under shear stress improved the formation of vascular lumen. The structure shows potential of clinical translation and therapeutic bone regeneration.</p><p>Methods of in situ bioprinting to pattern ECs for prevascularization have been used to promote bone regeneration. As shown by Kerouredan et al., bone regeneration was enhanced depending on the pattern in which the ECs were printed into calvaria bone defects. Bone regeneration was evaluated based on the percentage of bone formation, and the authors concluded that the prevascularization due to bioprinting corresponded with printing pattern, in that they observed enhanced bone regeneration when the ECs were patterned in disc and crossed circle patterns.277</p><!><p>The tooth and supporting tissue include the periodontium, dentin-dental pulp complex, tooth root, blood vessels, and nerves. The interactions between the complex architecture of teeth and their supporting tissues is necessary for functions performed by teeth such as speech, chewing, and aid in digestive functions.581 Without complete recapitulation of these tissue elements, tooth regeneration will not be possible. Microvasculature is especially important in the engineering of dental pulp due to its high levels of vascularization and the supporting vasculature of teeth.582 Bioprinting can be implemented to control the architecture of tooth tissues and promote synergistic activity of the various elements within dental tissues.</p><p>Duarte Campos et al. created a novel bioprinting strategy that utilized the bioprinting of microvasculature to aid in the engineering of dental pulp tissue (Figure 39B).580 Their design included a hand-held bioprinter that would print an agarose and collagen type I hydrogel bioink directly into the tooth root canal. The bioink induced the formation of hollow vascular tubes by HUVECs inside the ink. Their design holds promise in future clinical application for dental pulp regeneration.</p><!><p>Cardiovascular disease is the leading cause of death worldwide and accounts for approximately one third of deaths in the United States.583 Therefore, there is a large need for therapeutic strategies to maintain heart function. Due to the low regenerative capacity of the heart, there is an pressing need for research in cardiac tissue engineering and regenerative medicine.584 The purpose of cardiac tissue engineering is to stabilize areas of the heart through revascularization and restoration of function.585 Although there are several elements important in cardiac tissue engineering, microvasculature is vital in regeneration of cardiac tissue. The average inter-capillary distance in myocardial tissue is approximately 20 μm, which is often not recapitulated with tissue engineered constructs that span several hundred μm to several mm.586 Additionally, microvasculature is necessary for transportation of nutrients to keep both cells in engineered constructs and native cells viable. Therefore, vessel architecture must be imposed on tissue engineered structures.586</p><p>Sacrificial methods for the bioprinting of microvasculature have been implemented to create microvasculature in cardiac constructs. One method involves application of the SWIFT method for the fabrication of cardiac tissue.272 Using a support bath of iPSC-derived cardiac organ building blocks, a perfused branching architecture could be printed into the tissue, yielding high viability and sarcomeric structure development in the cardiac tissue. These printing methods were then used to recapitulate patient-specific cardiac structures, including a segment of the left anterior descending artery.272 Additionally, using patient left ventricle data from CT images, several mm thick structures consisting of two bioinks—one with cardiomyocytes and the other of a sacrificial ink containing ECs and fibroblasts—were used to bioprint microvasculature.273 Incorporation of microvasculature, as demonstrated in these applications, allowed the bioprinted structures to be several millimeters thick, which shows the potential for this technology to be applied to clinically relevant cardiac patches.273 In another technique, ECs were pre-organized to create a vascularized cardiac patch through seeding of ECs in patterned channels.478 The patches were useful in promoting anastomosis of native vessels in the host through paths created in vitro. Given the limitation of resolution of sacrificial printing, other methods must be implemented to make smaller scale capillary-like structures.</p><p>An improvement on the use of sacrificial bioprinting for the introduction of microvasculature in tissue engineered cardiac structures included a sacrificially printed branched channel networks used in conjunction with an oriented microporous scaffold.587 The pores within the structures allowed for elongation of cardiomyocytes and confluent seeding of ECs that were perfused through the channels. Additionally, the pores allowed for migration of ECs, which could facilitate the formation of microvascular networks.587</p><p>Using endothelial cell-laden bioinks, significant strides in bioprinting microvasculature have been made. 149,478,585 For example, microfluidic techniques can be coupled with bioprinting to create multiple-cell type tissue constructs for cardiac tissue engineering. Bioinks laden with ECs have been printed to create a prevascular structure that supports synchronous beating of cardiomyocytes seeded on top of the construct.149 Cardiac patches have been constructed by bioprinting a methacrylated collagen hydrogel laden with human coronary artery endothelial cells, as shown by Izadifar et al.585 To create the implant framework, alginate was printed into a calcium chloride bath, followed by the removal of calcium chloride and the printing of the cell-laden hydrogel. In the three-layer patch, which consisted of the alginate framework and cell-laden MeCol gel, lumen-like structures were formed by the ECs.</p><p>Several developing techniques in bioprinting have emerged to have large potential in advancing bioprinting microvasculature for cardiac applications. FRESH 3D bioprinting has demonstrated the potential to print small scale vessels (Figure 39C).269 Feinberg and colleagues printed multiscale vasculature based on MRI-derived CAD models that contained vascular structures patent at ~100 μm. Additionally, their technology demonstrated the ability to print organ-scale structures. Their proof-of-concept models exhibit the potential for future applications in which FRESH 3D bioprinting can be used for introducing microvasculature in tissue engineered constructs for cardiac applications.</p><!><p>Muscle tissue consists of organized bundles of muscle cells capable of contraction that are supported by nerves, blood vessels, and ECM.588 Blood vessels are important in supplying muscle cells with oxygen and nutrients and removing waste created following muscle contractions. Therefore, for proper engineering of muscle tissues with thicknesses greater than 100–200 μm, a functional vascular system must be engineered into constructs.568</p><p>Tissue-derived bioinks have been shown to be viable options for the treatment of volumetric muscle loss (VML), as shown by Choi et al (Figure 39D).438 The group's approach induced extensive endothelial vessel network. Following implantation in VML space, the group found red blood cells in the lumens of implanted areas and high co-localization of human CD31 and mouse CD31 structures, indicating anastomoses between the implanted bioprinted microvasculature and host vasculature. The coaxial nozzle printing was successful in recapitulating the hierarchical architecture of the muscle and led to 85% of functional recovery.</p><!><p>In clinical therapies for wounds, rapid wound treatment and the promotion of tissue regeneration is vital to prevent excess scaring and the worsening of wounds with time.230 Strategies for treatment of wounds in the clinical setting include split thickness autografts and skin substitutes with or without cells.230,589,590 However, several limitations in these methods persist, and improved therapies are needed. Tissue engineering has become an increasingly attractive technique for the fabrication of skin tissue and for wound healing therapies. However, one major challenge that faces the development of tissue engineered constructs is that to create full-thickness constructs, vasculature must be present. Without sufficient vasculature, problems may occur such as lack of integration of the skin substitute, presence of infection, and necrosis of the tissue.591 Fortunately, bioprinting has emerged as a useful technique in the fabrication of full thickness tissue constructs due to the ability to generate organized microvasculature591 and recapitulate the architecture of native skin.590</p><p>One application of bioprinting consisted of fabrication of a multi-layered skin equivalents consisting of fibroblasts and keratinocytes on top of Matriderm.279 These constructs were then implanted into full thickness wounds to fill the defect site and promote the infiltration of native vasculature. Additionally, in situ bioprinting methods have been used to induce microvascular infiltration into wound sites (Figure 39E).230 A fibrin and collagen bioink laden with dermal fibroblasts and keratinocytes was used to deliver cells directly to wound sites based on the topography of a patient's wound. This study showed promise in healing full thickness wounds and demonstrated a robust technique for delivering cells in a manner unique to a patient's wounds. However, these methods did not specifically print microvasculature but only harnessed the ability of the bioinks to promote vascular infiltration into the wound site.498,510</p><!><p>Bioprinting has tremendous potential for on-demand fabrication of human replacement tissues and organs. However, its clinical translation remains impeded by the inability to print vasculature that recapitulates the multiscale nature and function of human vascular systems. Most studies to date have focused on bioprinting tubular structures and millimeter-scale vascular constructs with simple cylindrical or branched geometry. While these structures are useful as replacements for large and medium diameter vessels, more delicate capillary networks and intra-organ vasculature are necessary for fabricating functional tissues and organs. With the development of advanced bioinks and innovative bioprinting techniques, there has been a surge in progress towards bioprinting vascularized tissues and organs with physiologically accurate architectures. In this review, we have evaluated the performance and potential of these techniques and bioinks for bioprinting microvessels. In addition, we have also examined the current applications of bioprinted microvascular constructs for in vitro assays and in vivo tissue regeneration. Despite the exciting progress made so far, there are still numerous challenges remaining in the field. They include: 1) developing novel bioinks that are both proangiogenic and highly printable; 2) printing microvasculature with physiologically relevant heterogeneity and function; 3) enabling controlled organization of microvascular networks in 3D printed tissues; 4) developing novel printing techniques that fulfill appropriate speed, resolution, and biocompatibility requirements to fabricate clinically applicable vascularized tissues and organs.</p><p>The current shortage of bioinks with both good printability and high proangiogenic activity has created a major bottleneck in bioprinting systems for fabricating microvasculature. It is well-understood that a "biofabrication window" exists for most conventional bioinks wherein biocompatibility and printability are negatively correlated.347 Most highly printable existing bioinks are often not cytocompatible. There is an urgent need for the development of novel, versatile bioinks that extend this biofabrication window. Tremendous innovation has been demonstrated in the design of injectable proangiogenic hydrogels for therapeutic vascularization. It is highly feasible to adapt these hydrogels for bioprinting since the rheological requirements for injectability (e.g. shear-thinning, in situ gelation) are compatible with extrudability. Recent work has demonstrated the use of nanomaterials (e.g. nanoparticles and nanofibers) to reinforce the rheological properties of bioinks for printing vascular structures.592,593 Bioprinted nanoparticles can also be loaded with growth factors to induce angiogenesis after printing.594 In addition, physical and chemical crosslinking as well as blending strategies with tunable materials like PEGs could allow for further fine-tuning of the printability of proangiogenic biomaterials. Photocrosslinkable functionalities could also render proangiogenic biomaterials suitable for light-assisted bioprinting of capillary-scale networks.</p><p>Ultimately, bioinks need to be formulated with biomaterials that closely mimic provascular microenvironments within native ECM. As discussed in Section 2.4, the ECM plays a critical role in regulating physiological angiogenesis and tissue formation, and the exact mechanisms by which this occurs is still an active area of research. Therefore, novel bioink platforms should adapt to a contemporary understanding of the interplay between matrix biology and vascular morphogenesis. Extracellular matrix and growth factor engineering in designer biomaterials systems should take advantage of advances in matrix biology and materials science towards developing novel proangiogenic bioinks for bioprinting microvasculature.216</p><p>Another major challenge for bioprinting microvasculature is recapitulating physiologically relevant heterogeneity and maturity of microvascular networks in engineered tissues. As discussed in Section 2, the biological mechanisms of vascular morphogenesis are exceptionally complex, and there are still gaps in knowledge of the basic science of developmental and adult vascular biology. While sophisticated engineering approaches have been developed to mimic the processes of vessel generation, the resultant microvascular networks still fail to recapitulate the physiological complexity of capillary beds. For example, the controlled release of single or dual growth factors (e.g. VEGF, PDGF) is a common strategy to promote angiogenesis, but it is not enough to drive the robust formation of functional microvascular networks. In addition, conventional proangiogenic hydrogels do not provide dynamic cell-matrix interactions and anisotropic microenvironments that orchestrate native vascular morphogenesis. Going forward, controlled release of multiple pro-angiogenic and pro-maturation factors in vascular-inductive matrices can enhance microvascular network formation and maturation within printed tissues.595 This could be addressed by adjusting the physical properties of biomaterials to achieve desired release profiles matching natural cascades during wound healing. Another potential approach would be incorporating growth factor-loaded nanoparticles within functionalized microporous scaffolds to mimic native ECM components, as reviewed in ref. 596. Furthermore, post-printing maturation through perfusion bioreactors and mechanical loading may be essential to develop mature microvasculature. Combining advanced release strategies for delivery of multiple growth factors in multi-material bioink platforms along with mechanical stimulation may vastly improve the function of vascular networks and tissue morphogenesis in bioprinted constructs.</p><p>The heterogeneity of cell types involved during microvascular fabrication is another important consideration to build physiologically relevant microvessel systems. HUVECs have been used in most of the existing bioprinting studies due to their high availability and robust expansion in culture. However, they may not be an optimal cell type for engineering microvasculature. Microvascular endothelial cells may provide a more suitable alternative since their native function is to form capillaries in vivo. Notably, there are significant genotypic and phenotypic variations among microvascular endothelial cells from different organs and microvascular-beds.597 Therefore, bioprinting strategies will need to take into account the heterogeneity of microvasculature and their microenvironmental niches across different systems in the body. In addition to ECs, supporting cell types like stem cells and fibroblasts should also be included to further promote tissue formation, maturation, and establishment of an endogenous basement membrane around vessel networks, which lessens the likelihood of thrombotic events after in vivo implantation.598 Therefore, biomaterials used in bioprinting should promote normal function of these supporting cell types, including appropriate guidance of stem cell differentiation for the desired tissue. Finally, mural cells (i.e. pericytes) will need to be included in bioink formulations to stabilize mature microvascular networks. This is especially important for tissues like the brain, where pericytes control the permeability of the blood-brain barrier.194 More studies focused on the formation of mature, pericyte-supported capillaries surrounded by basement membrane (i.e. collagen type IV and laminin) in bioprinted tissues are necessary. For clinical translation and patient-specific applications, iPSC-derived ECs and supporting cells are ideal to promote immunotolerance of implanted constructs. Furthermore, recent developments in organoid and microtissue fabrication offer powerful new building blocks for engineering vascularized organs and microsystems with physiologically relevant cell density and 3D organization.</p><p>Controlling the 3D patterning of microvascular network formation in bioprinted constructs represents another critical challenge. In most bioprinting platforms, there is little control over the organization of developing microvessels. After printing, vascular cells are essentially left to self-assemble into networks in an unmanageable fashion. This causes poor reproducibility of microvascular network formation in bioinks, especially in biomaterials that have batch variations in their composition. While it is relatively straightforward to control the patterning of microvascular networks on 2D substrates through site-specific functionalization, manipulating their organization in 3D poses significant difficulties. Indirect bioprinting approaches can pattern 3D sacrificial networks, but subsequent endothelialization and angiogenesis from the parent vessels is difficult to control. Emerging biomaterials systems that offer 4D control over matrix properties like growth factor delivery and bioactive ligand presentation have tremendous potential for controlling microvascular network patterning in hydrogel bioinks.519 For example, photosensitive biomaterials and growth factors can be spatiotemporally manipulated using laser-based direct writing to control site-specific ligand presentation and guide 3D endothelial cell migration in hydrogels.312,316 These strategies could provide elegant control over the organization of microvasculature in bioprinted materials and should be further explored for building functional microvasculature.</p><p>Lastly, although numerous advanced modalities have been developed for vascular bioprinting, as discussed in Section 3, each approach has inherent limitations and no single technique on its own can fabricate microvascularized tissues with sufficient speed and biomimicry. Since these methods are being developed in pursuit of a common goal to fabricate human tissues and organs, it is foreseeable that next-generation bioprinting techniques would combine DBB, EBB, and LAB approaches into modular hybrid methods to complement their strengths, offset their individual limitations, and accommodate more diverse bioink formulations. For example, Ozbolat's group proposed a hybrid platform using coaxial extrusion and scaffold-free bioprinting to fabricate scalable tissues and organs with perfusable microvasculature and physiological cell density.341 Alternatively, EBB and LAB methods could be combined to complement the speed and resolution of these approaches, respectively, towards high-throughput fabrication of scalable tissues with intricate microvasculature. Dual 3D bioprinting systems combining extrusion-based methods and stereolithography have recently been developed to fabricate perfusable small-diameter vasculature248 and hierarchically vascularized bone biphasic constructs579, demonstrating proof-of-concept and the strong potential of hybrid bioprinting platforms for vascular tissue engineering. The evolution of similar hybrid platforms should unlock new possibilities for generating sophisticated microvascularized tissues and organs for clinical applications in the future. Furthermore, emergent applications of volumetric additive manufacturing in tissue engineering may massively increase build times for vascularized tissues and organs.599–601</p><p>Bioprinting technology is evolving rapidly and is undoubtedly poised to make major contributions to healthcare. Bioprinting microvasculature represents one of the most critical challenges in the evolution of the field. Complementary innovation in high-resolution bioprinting methods, highly printable sacrificial and proangiogenic biomaterials, and autologous sources of cells and biomaterials are necessary to bioprint microvasculature capable of mimicking native vascular physiology. Overcoming these challenges will bring the field closer to printing functional human organ-scale vasculature, which has been referred to as the "Mars mission of bioengineering".602</p>

PubMed Author Manuscript

Electrostatic Origins of CO2-Increased Hydrophilicity in Carbonate Reservoirs

Injecting CO 2 into oil reservoirs appears to be cost-effective and environmentally friendly due to decreasing the use of chemicals and cutting back on the greenhouse gas emission released. However, there is a pressing need for new algorithms to characterize oil/brine/rock system wettability, thus better predict and manage CO 2 geological storage and enhanced oil recovery in oil reservoirs. We coupled surface complexation/CO 2 and calcite dissolution model, and accurately predicted measured oil-on-calcite contact angles in NaCl and CaCl 2 solutions with and without CO 2 . Contact angles decreased in carbonated water indicating increased hydrophilicity under carbonation. Lowered salinity increased hydrophilicity as did Ca 2+ . Hydrophilicity correlates with independently calculated oil-calcite electrostatic bridging. The link between the two may be used to better implement CO 2 EOR in fields.Oil will be an important energy source for the rest of the 21 st century 1 and carbonate reservoirs host most of the world's oil (>60%) 2 . However, the recovery factor is low (<40%) 3 , so there is enormous motivation to improve recovery cost-effectively, and with environmentally friendly manners. CO 2 EOR is attractive because it produces more oil without the expense of chemicals although CO 2 injection costs energy to compress before injection, and the CO 2 source availability needs to be also considered. Moreover, CO 2 -EOR can combat global warming by injecting CO 2 into geological formations. CO 2 EOR techniques include miscible 4 and immiscible continuous injection 5,6 , carbonated water flooding 7 , huff and puff injection (injecting CO 2 in a single well and producing from the well after CO 2 equilibration with the crude oil) 8,9 , and water-alternating-CO 2 injection 10-12 . CO 2 techniques work through some combination of immiscible drive, first contact miscible drive, vaporizing-gas drive, condensing-gas drive, and vaporizing-condensing gas drive, and multiple-contact miscible drive. At the microscopic level, these processes can: promote oil-swelling, reduce oil viscosity, mitigate gravity segregation by reducing the density difference between oil and water, and, lower oil interfacial tension, all of which can increase oil recovery. The net impact of CO 2 addition can be quite large, amounting to recovery of an extra 4-15% of the original oil in place in conventional reservoirs 13 . Moreover, CO 2 huff-n-puff can achieve 14% additional oil recovery from unconventional reservoirs 14 .While much is known about the effect of CO 2 on oil fluid properties, oil-CO 2 -brine-carbonate system wettability is not well understood, which triggers intrinsic uncertainties to predict and manage the CO 2 injection and reservoir performance although CO 2 -brine-rock system wettability has been well documented 15,16 . This is largely because system wettability governs subsurface multiphase flow and residual saturations 17 . To examine the wettability, contact angle test has been perceived as an effective means together with interpretation using Derjaguin-Landau-Verwey-Overbeek (DLVO) [18][19][20] and surface complexation modelling [21][22][23] showed that dissolving CO 2 into seawater decreases oil contact angles on calcite, thus increasing hydrophilicity. Decreased salinity also decreases contact angles. Teklu et al. 10 noted several potential explanations for their contact angle trends and called for a closer examination of the surface controls over wettability alteration. Venkatraman et al. 24 used Gibbs free-energy function to integrate phase-behaviour computations and geochemical reactions to find equilibrium composition, but quantitative work remains to be made to understand how dissolved CO 2 governs oil-brine-calcite interaction, thus wettability. Here we constrain surface chemical controls over wettability in carbonate reservoirs undergoing CO 2 EOR by interpreting new oil-on-calcite contact angles

electrostatic_origins_of_co2-increased_hydrophilicity_in_carbonate_reservoirs

<!>Results and Discussion<!>Implications and Conclusions<!>Methods

<p>in the presence of model reservoir brines containing NaCl and CaCl 2 using a coupled surface complexation/CO 2 and mineral dissolution model.</p><!><p>To examine the wettability of oil-brine (CO 2 )-carbonate system, we measured contact angle of oil on calcite substrates in the presence of carbonated brine or non-carbonated brine. Figure 1 shows oil-on-calcite contact angles measured at 25 °C and 3000 psi pressure in model brines under carbonated and non-carbonated conditions. Carbonated water lowers contact angles and produces a strongly water-wet system regardless of salinity and ion type compared to non-carbonated water. For example, non-carbonated 1 mol/L NaCl yielded a contact angle of 120°, meaning an oil-wet system. However, carbonated 1 mol/L NaCl gave a contact angle of 39°, meaning a strongly water-wet system. Similarly, Teklu et al. 10 observed a contact angle shift from 116.6-133.6° (non-carbonated seawater, pH = 6.6) to 36.1-40.8° (carbonated seawater, pH = 5.5 at atmospheric condition). A secondary effect of lowered salinity decreasing contact angles and moving the system towards water wetness is also seen in Fig. 1, and was observed before by Teklu et al. 10 . Divalent cations (Ca 2+ ) gave a lower contact angle compared to monovalent cations (Na + ) regardless of concentration.</p><p>To understand how carbonation increases hydrophilicity, we develop a geochemical model that couples CO 2 dissolution, mineral dissolution, and oil and calcite surface chemistry (Table 1). CO 2 and calcite dissolution into brines is calculated by a standard equilibrium approach. Oil surface species are assumed to be -NH + , -COO − and -COOCa + 23,25,26 , polar surface groups expressed at, and attached to, the oil-water interface. Calcite surface species are assumed to be >CO 3</p><p>, and >CO 3 Ca + 21-23,27 (Fig. 2); where ">" denotes a calcite surface species. The primary electrostatic bridges between oppositely charged oil and calcite surface species are then the pairs, -NH + and >CO 3 − , -NH + and >CaCO 3 − , -COOCa + and >CO 3 − , -COOCa + and >CaCO 3 − , -COO − and >CaOH 2 + , and -COO − and> CO 3 Ca + . A quantitative measure of electrostatic attraction is termed the bond product sum 21,23 , BPS, which is equal to</p><p>where bracketed terms are calculated surface concentrations (μmol/m 2 ). Bond product sum (electrostatic bridges) is an explicit way to reflect the electrostatic force change between the oil/brine and rock/brine interfaces. Our previous studies 25,28 show that DLVO and surface complexation modelling predict similar wettability trends. This is because the physics of DLVO and surface complexation is the same as a result of diffuse double layer. We therefore decided to use BPS to reveal the interaction of oil-brine-carbonate because BPS can be practically modelled using the geochemical reactions with reservoir simulators for waterflooding and EOR studies. Speciation of Oil/Brine Interfaces. Figures 3 and 4 show calculated oil surface speciation in non-carbonated and carbonated NaCl and CaCl 2 brines. Calculation results are listed in Tables 2 and 3 in Supplementary Information. The calculated amount of -NH + decreases with increasing pH regardless of ion type and salinity for both non-carbonated and carbonated brines as pH controls the amount of-NH + through Reaction 1 (Table 1) shifting to the left 25,29 . The calculated amount of -COO − increases with increasing pH but decreases due to the formation of -COOCa + for non-carbonated brines (Fig. 3). The same trend is observed in carbonated brines (Fig. 4), but with an increase of -COO − with increasing pH due to the formation of CO 3 2−</p><p>, which decreases Ca 2+ . Note: the amount of -COOCa + depends on dissolved Ca levels and to a lesser extent ionic strength because of their effect on surface species concentrations and the Ca 2+ activity coefficient 25 . Keeping in mind that the surface speciation responds to dissolved phase concentrations that, through calcite equilibria, are set by pH and amount of carbonation (in situ P CO2 ). For example, Ca 2+ levels and ionic strength are higher at low pH and in carbonated brine. The PHREEQC surface complexation calculation tracks each of the competing factors while maintaining equilibrium with calcite.</p><p>Speciation of Calcite/Brine Interfaces. Figures 5 and 6 show calculated calcite surface speciation in non-carbonated and carbonated NaCl and CaCl 2 brines. Calculation results are listed in Tables 4 and 5 in Supplementary Information. Note that the legends in Figs 5-8 refer to initial solution compositions. Final solution compositions are influenced by calcite dissolution and P CO2 . Because in the CaCO 3 -H 2 O-CO 2 system CO 2 , pH, and Ca 2+ are coupled, ionic strength is particularly sensitive to pH and P CO2 -dependent calcite dissolution reactions. So calcite dissolution in the pH < 4 for example causes calculated ionic strengths to be well above 1 M.</p><p>In both non-carbonated and carbonated solutions, low pH calcite surface charge is dominated by >CaOH 2 + . >CaOH 2 + is the most abundant surface species at high pH as well in non-carbonated solutions. Increasing pH favors a decrease in >CaOH 2</p><p>+ and an increase in >CaCO 3 − , >CO 3 − , and >CO 3 Ca + for a given available Ca +2 . In carbonated solutions, high pHs and bicarbonate prompt appreciable formation of >CaCO 3 − , >CO 3 − , and >CO 3 Ca + .</p><p>Two shifts that stand out between the non-carbonated and carbonated cases are the conversion of >CaOH 2 + to >CaCO 3 − and >CO 3 − to >CO 3 Ca + with increasing CO 2 . These reactions are driven by respectively the higher bicarbonate and calcium levels in CO 2 -charged brine:</p><p>Calculation of Oil-on-Calcite Wetting. We combined the calculated oil and calcite speciation above into a bond product sum, BPS, the number of electrostatic bridges between the oil and calcite. Again, the BPS is a measure of electrostatic attraction between oil and calcite, is proportional to measured contact angles 21,30 , and is therefore a useful predictor of wetting. For our system, the bond product sum is the total of six concentration products that quantify the strength of six electrostatic bridges between oppositely charged oil and calcite species, as noted above. For natural systems containing sulphate the BPS would also include, for example, a [>CaSO 4 − ] [-NH + ] term.</p><p>Figures 7 and 8 show the bond product sum for non-carbonated and carbonated conditions. Calculation outputs are listed in Table 6 and The pH in non-carbonated brines increased from 7 to 10 after equilibration with calcite which decreases the BPS (Fig. 7) and the contact angle (Fig. 1). In contrast, in carbonated brines, the pH decreases to below 4 which decreases the BPS to almost three times than the non-carbonated brine (Fig. 8), accounting for the contact angle decrease in a range of 20 to 80° with various brines (Fig. 1) thus more hydrophilicity system. The pH difference</p><p>Table 1. Surface complexation model input parameters. ">" represents the negatively charged site on the carbonate surface while the "−" represents the negatively charged site on the oil surface. Given that directly sorbed oil probably doesn't respond to low salinity waterflooding, and that only oil-rock with an intervening water layer will respond 23 , our analysis focusses solely on the water-present situation which can be modelled using surface complexation theory 23,29 . In our geochemical modelling, we did not consider the interaction between non-polar oil and calcite surfaces, e.g., hydrogen bonding, Van der Waals interaction, and ligand bridging, etc. 45 . However, we can reasonably assume that acidic and amine functional groups governs the electrostatic surface species at oil surfaces, which dominates the adhesion force between oil and rock surfaces [46][47][48] . Moreover, water assisted EOR (e.g., carbonated water and low salinity water) plays a main role in the interaction of polar part and rock surfaces 46,49,50 , but the interaction between non-polar oil and calcite surfaces plays a little effect in water assisted EOR 51,52 . Our assumption therefore can be reasonably justified.</p><p>between calcite-equilibrated carbonated and non-carbonated brine largely accounts for why Teklu et al. 10 and we observed a dramatic contact angle decrease in carbonated brine (Fig. 1). Specifically, electrostatic adhesion decreases with carbonation because of a decrease in pH. In a carbonate reservoir the reduction in electrostatic adhesion with carbonation ultimately means greater oil recovery because it causes an increase in oil relative permeability 31 .</p><p>Although BPS prediction appears to be in line with contact angle measurements on calcite surfaces, to complement the BPS estimates and provide deeper thermodynamic insights to the nature the physics which controls wettability of brine-oil-carbonate, we computed surface potential of brine-oil and brine-calcite in light of diffuse double layer to calculate total disjoining pressure of oil-brine-carbonate system in non-carbonated and carbonate brine within DLVO framework 32 as the sum</p><p>Total e lectrical 3 where II Total is the disjoining pressure of the specific intermolecular interactions which reflects the interactive forces between the interfaces of brine-oil and brine-rock. II electrical is the electrostatic forces due to the development of the charges between interacting surfaces. A brief introduction of the forces and calculation procedures were presented elsewhere 33 . The Hamaker constant for oil-brine-rock in water is approximately 1 × 10 −20 J 34 .</p><p>Melrose 35 used Hamaker constants ranging from 0.3 to 0.9 × 10 −20 J. In this study, 0.81 × 10 −20 J was used as the Hamaker constant 34 . We did not consider the structural forces to model the total disjoining pressure due to the fact that the structural forces are short-range interactions over a distance of less than 5 nm compared with long-range interactions 36 , e.g., London-van der Waals and electrical double layer forces. We computed the brine chemistry and surface potential of fluid-fluid and fluid-rock with considering calcite dissolution and water uptake of CO 2 for carbonate brine using PHREEQC, as shown in Table 9 and Table 10 listed in Supplementary Information. Constant charge (CC) and constant potential (CP) conditions based on the linear P-B expression 37 were used to compute total disjoining pressure versus interfacial separation of the oil and calcite surfaces. The two conditions represent upper and lower bounding curves on the total disjoining pressure 25,37 .</p><p>Figure 9 shows the isotherms of total disjoining pressure versus separation distance between oil and calcite surfaces across various brines. Positive pressure indicates repulsion, and negative pressure implies attraction. In the presence of 0.01 mol/L CaCl 2 and NaCl solution, carbonated brine gave a positive disjoining pressure which exhibits a progressively more repulsive barrier on approach, implying a strongly water-wet system in line with contact angle measurements. Non-carbonated brine gave a negative disjoining pressure indicates an oil-wet system except 0.01 mol/L NaCl at constant charge (CC). This however contradicts contact angle results, which show a water-wet system although contact angle is 20 to 30° more than the contact angle in carbonated brines. This is because non-carbonated brine gives a strongly negative surface potential, whereas the surface potential of brine-calcite remains positive, triggering attractive forces. We believe that both electrostatic and nonelectrostatic physisorption together with competitive ion chemisorption (ion exchange and surface complexation modelling) 38 would be combined to better account for the total disjoining pressure.</p><p>In the presence of 1 mol/L CaCl 2 and NaCl solution, carbonated brine yields a relatively lower adhesion compared to non-carbonated brine, indicating a lower contact angle in carbonated brine in line with experiments. However, the same contradiction remains, showing that both carbonated and non-carbonated brine gave negative disjoining pressure, signifying an oil-wet system. We believe that double layer collapse may be one of the main reasons to account for this negative disjoining pressure due to the high ionic strength 39 . Together, although the disjoining pressure does not completely predict the contact angle results, DLVO and surface complexation modelling predict the same trend over the contact angle results, showing that carbonated brine leads to less adhesion compared to non-carbonated brine thus hydrophilicity in line with contact angle results. In addition, our surface potential results also show that carbonation plays a minor role in surface potential of brine-calcite in the presence of CaCl 2 (Table 10 in Supplementary Information), confirming that Ca level likely dominates brine-calcite surface potential thus zeta potential rather than pH in line with Mahrouqi et al. 39 .</p><!><p>To better predict and manage CO 2 geological storage and enhanced oil recovery in carbonate oil reservoirs, we aimed to understand oil-CO 2 -brine-carbonate system wettability by measuring oil contact angle in carbonated and non-carbonated brines. We also coupled surface complexation/CO 2 and calcite dissolution model, and accurately predicted measured oil-on-calcite contact angles in NaCl and CaCl 2 solutions with and without CO 2 . To further complements surface complexation modelling, DLVO theory was used to calculate disjoining pressure at constant charge and constant potential conditions, confirming that DLVO and surface complexation modelling predict the same trend. Contact angle results show that carbonated water increases hydrophilicity. Reduced salinity increased hydrophilicity as did Ca 2+ . Our coupled surface complexation/CO 2 and mineral dissolution through zeta potential measurements. The impact of salinity on oil and calcite surface complexation in high TDS solutions must be verified. Alternative calcite surface complexation stoichiometries than those in Table 1 exist 27 . Our preliminary calculations using the calcite surface stoichiometries of Song et al. 27 predict the same trends seen above although the absolute values of the calculated BPS are different (Fig. 9 and Table 8 in Supplementary Information).</p><!><p>Substrates. Calcite minerals supplied by Ward's Science were used in the contact angle tests. X-Ray Diffraction (XRD) tests confirmed that the composition of substrates were 100% calcite. To avoid any hysteresis and contamination the natural mineral surfaces (cleavage) were used as pendent spots.</p><p>Prior to experiments, substrates were cleaned with solvents (e.g., toluene and methanol) to remove any traces of organic and inorganic contaminants. Substrates were then rinsed with equilibrated deionised water to prevent undesired dissolution and dried in an oven at moderate temperature of 60 °C. Then, clean and dry substrates were exposed to air plasma for 10 min to remove organic surface contamination 40 . We also imaged the cleaved calcite substrate to obtain the surface roughnesusing atomic force microscopy (AFM) (WITec, ALPHA 300 RA for combined Raman-AFM imaging). Results show that the surface roughness was in a range of 0 to 4.8 nm 30 , implying that the surface roughness effect on contact angle should be negligible 16,41 .</p><p>Liquids Preparation. Texas crude oil from the United States was used in contact angle tests. Chemical analysis of crude oil indicated the acid and base number were 1.7 and 1.2 mg KOH/g, respectively. To prepare carbonated brines, 1.0 mole and 0.01 mole of NaCl and CaCl 2 brines were prepared and individually loaded in a reactor. CO 2 gas was injected in the reactor through a syringe pump with the aid of a compressor and mixed with the brine at 3000 psi and 25 °C until the brine was saturated with CO 2 gas. Saturated brine was transferred into an accumulator and maintained under pressure until the experiment was carried out.</p><p>Experimental Procedure. Contact angle experiments were measured using a Vinci IFT apparatus (see Fig. 1 in Xie et al. 28 ). All contact angles were measured at 3000 psi and 25 °C conditions. Calcite substrates were mounted on the apparatus turn table and placed inside the high pressure high temperature (HPHT) cell and sealed and vacuumed until state of vacuum was attained. The pressure cell was then filled with the desired brine and pressurised to 3000 psi. Subsequently, the experimental oil was slowly and steadily introduced into the cell through a capillary needle (0.64 mm diameter) until a droplet was formed. The droplet was then released on the substrate, and integrated software was utilised to measure left and right contact angles between substrate and the oil droplet. Contact angles were continuously recorded until equilibrium was achieved where contact angle became stable. This process was repeated for CO 2 -saturated brines. Throughout the experiment test pressure was closely monitored and maintained to prevent depressurisation of cell, and desaturation of the brine.</p><p>Simulation Methods. Surface complexation modelling (and DLVO theory) presumes an electrical double layer at each interface and the existence of charged surface species whose concentrations depend upon the chemical makeup of the water and the oil and mineral surface 28 . Surface equilibria and constants 23,[42][43][44] are listed in Table 1. The surface species concentrations were calculated using PHREEQC version 3.3.9 (Parkhurst and Appelo 2013) and a diffuse layer surface model. The calcite surface site density was assumed to be 5 sites/nm 2,22 . The oil/ calcite surface area was set to 0.11 m 2 /g 22 .</p>

Scientific Reports - Nature

MOLECULAR PATHOGENESIS OF HEPATIC FIBROSIS AND CURRENT THERAPEUTIC APPROACHES

The pathogenesis of hepatic fibrosis involves significant deposition of fibrilar collagen and other extracellular matrix proteins. It is a rather dynamic process of wound healing in response to a variety of persistent liver injury caused by factors such as ethanol intake, viral infection, drugs, toxins, cholestasis and metabolic disorders. Liver fibrosis distorts the hepatic architecture, decreases the number of endothelial cell fenestrations and causes portal hypertension. Key events are the activation and transformation of quiescent hepatic stellate cells into myofibroblast-like cells with the subsequent up-regulation of proteins such as \xce\xb1-smooth muscle actin, interstitial collagens, matrix metalloproteinases, tissue inhibitor of metalloproteinases and proteoglycans. Oxidative stress is a major contributing factor to the onset of liver fibrosis and it is typically associated with a decrease in the antioxidant defense. Currently, there is no effective therapy for advanced liver fibrosis. In its early stages, liver fibrosis is reversible upon cessation of the causative agent. In this review, we discuss some aspects on the etiology of liver fibrosis, the cells involved, the molecular pathogenesis and the current therapeutic approaches.

molecular_pathogenesis_of_hepatic_fibrosis_and_current_therapeutic_approaches

1. Introduction<!>2. Etiology of hepatic fibrosis<!>2.1. Alcohol<!>2.2. Chronic viral hepatitis<!>2.3. Other causes of hepatic fibrosis<!>3.1. Hepatic stellate cells<!>3.2. Portal fibroblasts<!>3.3. Bone marrow-derived mesenchymal stem cells<!>3.4. Hepatocytes and biliary epithelial cells<!>3.5. Fibrocytes<!>4. Molecular pathogenesis of hepatic fibrosis<!>4.1. Cell-cell and cell-matrix interactions<!>4.2. Oxidative stress<!>4.3. Role of MMPs and TIMPs<!>5. Current therapies for hepatic fibrosis<!>6. Molecular therapy for hepatic fibrosis<!>7. Conclusion<!>

<p>Fibrosis is the wound healing response to a variety of acute and/or chronic stimuli, including to name a few, ethanol, viral infection, drugs and toxins, cholestasis and metabolic disease [1–2]. Hepatic fibrosis develops due to an increase in fibrillar collagen synthesis and deposition along with insufficient remodeling [3–4]. Fibrosis is associated with a number of pathological and biochemical changes leading to structural and metabolic abnormalities, as well as with increased hepatic scarring [5–6]. The progression of liver fibrosis leads to cirrhosis, a condition characterized by distortion of the normal architecture, septae and nodule formation, altered blood flow, portal hypertension, hepatocellular carcinoma and ultimately liver failure [7].</p><!><p>Most chronic liver diseases are associated with fibrosis and are characterized by parenchymal damage and inflammation. Alcohol abuse, chronic viral hepatitis (HBV and HCV), obesity, autoimmune hepatitis, parasitic diseases (i.e. schistosomiasis), metabolic disorders (hemochromatosis and Wilson's disease), biliary disease, persistent exposure to toxins and chemicals and drug-induced chronic liver diseases are the most common causes of hepatic fibrosis.</p><!><p>Alcohol consumption is a predominant etiological factor in the pathogenesis of chronic liver diseases worldwide, resulting in fatty liver, alcoholic hepatitis, fibrosis/cirrhosis, and hepatocellular carcinoma [8]. Acetaldehyde, the product of alcohol metabolism via alcohol dehydrogenase, increases the secretion of transforming growth factor β1 (TGFβ1) and induces TGFβ type II receptor expression in hepatic stellate cells (HSC), the key collagen-producing cell within the liver [9]. Both, ethanol and acetaldehyde induce the COL1A2 promoter and up-regulate collagen I protein expression [10]. In cultured human HSC, acetaldehyde up-regulates COL1A1 mRNA expression via distinct mechanisms in the early and late responses [11]. Acetaldehyde-induced fibrogenesis involves a complex signaling pathway, which differs from that mediated by TGFβ1 in the early time points to up-regulate COL1A2 gene expression [11].</p><p>TGFβ 1 is a critical factor in the progression of alcoholic liver disease (ALD) in patients with steatosis and steatohepatitis [12]. Acetaldehyde does not alter the Smad3 and Smad4 protein concentration; however, it selectively induces phosphorylation of Smad3 but not of Smad2 [13]. Weng et al. [12] identified a significant correlation of Smad2 phosphorylation with the fibrosis stage and the inflammation score. In addition, an association between serum pro-collagen III N-pro-peptide and TGFβ 1 has been reported in patients with ALD [14]. These results demonstrate a significant role for TGFβ 1 as mediator of alcohol-induced liver fibrosis.</p><p>Hepatic alcohol metabolism generates reactive oxygen species (ROS) causing significant cell death [15]. Indeed, oxidative stress, likely by increasing mitochondrial permeability transition, promotes hepatocyte necrosis and/or apoptosis. Generation of ROS within hepatocytes may be a consequence of an altered metabolic state, as it occurs in nonalcoholic fatty liver disease and non-alcoholic steatohepatitis. Alternatively, it could result from ethanol metabolism as in alcoholic steatohepatitis, with ROS being generated mainly by the mitochondrial electron transport chain, cytochrome P450 isoforms such as cytochrome P450 2E1 (CYP2E1), damaged mitochondria, xanthine oxidase, NADPH oxidase and generation of lipid peroxidation-end products [16]. In addition, it is known that chronic alcohol consumption lowers glutathione levels; thus, contributing to liver injury [17]. ROS-derived mediators released by damaged neighboring cells can directly affect the HSC behavior. ROS up-regulate the expression of critical genes related to fibrogenesis including pro-collagen type I, monocyte chemoattractant protein 1 (MCP-1) and tissue inhibitor of metalloproteinase-1 (TIMP1), possibly via activation of a number of critical signal transduction pathways and transcription factors, including c-jun N-terminal kinases (JNKs), activator protein 1 (AP-1) and nuclear factor kappa B (NFκB) [18].</p><!><p>Chronic hepatitis B and C virus are the most common causes of liver disease worldwide, with an estimated 350 and 170 million of individuals with chronic infection, respectively [19]. In addition, these infections are the primary cause of hepatocellular carcinoma (HCC). In both cases, there is significant chronic liver injury with subsequent progression to advanced liver fibrosis and in many cases cirrhosis. While HBV can be integrated into the host genome leading to changes in genomic function or chromosomal instability, HCV cannot integrate into the host genome. Various HCV proteins, including the HCV core protein, the envelope and non-structural proteins present oncogenic properties. In HBV infection, antiviral therapy and vaccination decrease the risk of HCC. Current antiviral therapies for HCV such as ribavirin significantly reduce the risk of HCC.</p><!><p>In addition to alcoholism and chronic viral hepatitis, other factors contributing to hepatic fibrosis are obesity and steatosis, which can lead to nonalcoholic fatty liver disease and to chronic steatohepatitis. Nonalcoholic fatty liver disease has also been reported in non-obese individuals in developing countries [20].</p><p>Autoimmune hepatitis, the anomalous presentation of human leukocyte antigen class II in hepatocytes, causes cell-mediated immune responses against the host liver, and may lead to liver fibrosis as well [21]. Parasitic infections like schistosomiasis, have been shown to trigger advanced liver fibrosis and portal hypertension [22].</p><p>Metabolic disorders such as hemochromatosis and Wilson's disease are typically accompanied by chronic hepatitis and fibrosis [23]. In hereditary hemochromatosis, the excessive absorption and accumulation of iron in tissues and organs including liver is related to mutations in the HFE (High-iron) gene [24]. Wilson's disease or hepatolenticular degeneration is a genetic disorder leading to copper accumulation in the liver and it is due to a mutation in the APTase (ATP7B) that transports copper [25].</p><p>Lastly, cholestasis due to bile duct obstruction, leads to chronic portal fibrosis and eventually cirrhosis. Moreover, chronic exposure to toxins or chemicals such as N-nitrosodimethylamine, carbon tetrachloride (CCl4) or thioacetamide leads to severe hepatic fibrosis in experimental animal models [26–28]. Exposure to these chemicals in humans is rare and generally occurs in the industry during manufacture and in places where these chemicals are routinely used.</p><!><p>Several cell types are involved in the pathogenesis of hepatic fibrosis. HSC reside in the space of Disse between hepatocytes and sinusoidal endothelial cells [29]. Quiescent HSC are characterized by significant expression of desmin and vitamin A storage. Following liver injury, HSC lose their vitamin A content, increase the expression of α-smooth muscle actin (α-SMA), acquire a myofibroblast-like phenotype losing their typical star-shape, become proliferative, motile, pro-fibrogenic, contractile and show abundant rough endoplasmic reticulum [30].</p><p>Many factors have been identified to contribute to HSC activation. Damage to hepatocytes and Kupffer cell activation are still considered the primary effectors driving HSC activation [31–32]. Mediators released from damaged hepatocytes, such as lipid peroxidation products, intermediate metabolites of drugs or hepatotoxins, acetaldehyde and 1-hydroxyethyl radical from alcohol metabolism as well as ROS (hydrogen peroxide, superoxide radical and others) are strong inducers of HSC activation.</p><p>Activated Kupffer cells release ROS and cytokines that are crucial for HSC activation as well [32]. They are a major source of TGFβ and platelet-derived growth factor (PDGF), two potent profibrogenic cytokines that traditionally have been considered key fibrogenic and proliferative stimuli to HSC, respectively [33]. In addition, the Kupffer cell phagocytic activity generates large amounts of ROS that could further activate HSC and induce their fibrogenic potential.</p><p>We have previously demonstrated that cytochrome P450 2E1-dependent generation of ROS is critical for increased collagen I protein synthesis in co-cultures of hepatocytes and HSC [31]. Furthermore, addition of ethanol and arachidonic acid synergized to activate Kupffer cells and modulated the fibrogenic response by a mechanism involving TNFα, reduced glutathione and TGFβ [34].</p><p>It has been also demonstrated that in vivo ablation of TNFα, TLR4, CD14 and lipopolysaccharide-binding protein protects from the fibrogenic response [35]. Despite the close association of inflammation and fibrosis, little is known on the crosstalk between these two key events and the intracellular signal transduction pathways activated. For example, TLR4 is activated by lipopolysaccharide in Kupffer cells leading to NFκB and IRF3 activation, and the subsequent transcriptional activation of pro-inflammatory mediators such as TNFα and IFNγ. Moreover, TNFα activates the NFκB signaling pathway in hepatocytes, which is key for their survival [8, 36]. However, there is no crosstalk with the TGFβ pathway that results in the activation of Smad3 and Smad4 and the associated induction of TGFβ-responsive genes.</p><!><p>The portal connective tissue in healthy liver is surrounded by quiescent portal fibroblasts, which constitute a second population of liver cells implicated in portal fibrosis [37]. Derived from small portal vessels, they express markers distinct from HSC (e.g. elastin) [38]. Proliferation of biliary cells is often accompanied by proliferation of portal fibroblasts, which form onion-like configurations around biliary structures and acquire a myofibroblast phenotype, and are thus implied in the early deposition of extracellular matrix (ECM) in portal zones [39]. It is generally believed that substantial signaling from biliary epithelial cells leads to portal fibroblast activation, although the key factors remain to be identified.</p><!><p>Several studies have indicated that bone marrow derived mesenchymal stem cells could be a source of multi-lineage cells for various organs. They have the capacity to differentiate into hepatocytes, biliary epithelial cells, sinusoidal endothelial cells and even Kupffer cells in the presence of a suitable hepatic microenvironment [40–41]. There is growing evidence to suggest that bone marrow-derived stem cells are recruited during both progression and regression of liver fibrosis. During regression from CCl4 induced hepatic fibrosis, bone marrow-derived mesenchymal stem cells migrate into the fibrotic liver, where they can express matrix metalloprotease-13 (MMP13) and MMP9 [42]. In addition, granulocyte colony-stimulating factor and hepatocyte growth factor treatment significantly enhance migration of bone marrow-derived cells into the fibrotic liver and accelerate the regression of liver fibrosis [43]. Over-expression of hepatocyte growth factor together with granulocyte colony-stimulating factor, synergistically stimulate MMP9 expression, which is followed by accelerated resolution of fibrotic scars [44]. A significant contribution of bone marrow-derived cells has been shown in human liver fibrosis, but it is unclear the specific type of mesenchymal stem cells [45].</p><!><p>Epithelial-to-mesenchymal transition (EMT) is now emerging as a possible source of injury-associated mesenchymal cells, derived either from resident hepatocytes or from biliary epithelial cells [46–47], although its role in liver fibrosis is still controversial [48–49]. The main molecules inducing EMT are TGFβ, epidermal growth factor, insulin-like growth factor-II and fibroblast growth factor-2 [50]. Hepatocytes that express albumin also express fibroblast-specific protein-1 in response to CCl4 in vivo or to TGFβ1 in vitro [50]. Kaimori et al. [51] reported that hepatocytes express COL1A1 in response to TGFβ1 in vitro, and that Smad signaling mediates EMT.</p><p>On the contrary, a recent study shows that hepatocytes isolated from mice chronically treated with CCl4, neither express mesenchymal markers nor exhibit a phenotype clearly distinguishable from control hepatocytes [52]. In order to confirm this, Scholten et al. [53] studied EMT using the Cre-LoxP system to map the cell fate of CK19+ cholangiocytes in CK19 (YFP) or fibroblast-specific protein-1 (FSP-1) (YFP) mice. Mice were bile duct ligated or subjected to CCl4-induced liver injury and the livers were analyzed for expression of mesoderm and epithelium markers. The results demonstrated that EMT of cholangiocytes does not contribute to hepatic fibrosis in mice. Likewise, GFAP (Cre)-labeled HSC showed no co-expression of epithelial markers, providing no evidence for EMT in HSC in response to fibrogenic liver injury [53].</p><p>Biliary epithelial cells have been described to be involved in EMT in liver fibrogenesis. In primary biliary cirrhosis, it has been shown that cells of the bile duct express fibroblast-specific protein-1 and vimentin, early markers of fibroblasts [54]. A consequence of EMT in biliary epithelial cells is the amplification of the pool of portal fibroblasts, contributing significantly to portal fibrosis. In vitro studies with human biliary epithelial cells have confirmed these clinical observations [55]. Thus, EMT could be considered a mechanism participating in the pathogenesis of chronic cholestatic liver disease. However, additional research is necessary to validate this possibility.</p><!><p>Fibrocytes constitute a circulating bone marrow-derived CD34+ cell subpopulation with fibroblast-like properties initially associated with tissue repair in subcutaneous wounds [56]. They comprise a fraction of about 1% of the circulating pool of leukocytes expressing markers of mesenchymal cells [57]. Subsequently, two studies have demonstrated the bone marrow origin of fibrogenic cell populations in the CCl4 mouse model of fibrosis and in the bile duct ligation model of biliary hyperplasia [58–59]. Roderfeld et al. [60] hypothesized that bone marrow-derived circulating CD34+ fibrocytes represent key mediators of liver fibrogenesis in the Abcb4−/− mice, which represent a highly reproducible, well-characterized non-surgical mouse model for cholangiopathy in humans.</p><!><p>Figure 1 summarizes key concepts involved in the molecular pathogenesis of liver fibrosis. A recent review by Hernandez-Gea et al [61] provides more insight into the pathogenesis of hepatic fibrosis.</p><!><p>Alterations in normal cell-cell and cell-matrix interactions play a significant role in pathogenesis of hepatic fibrosis. When normal cell-cell and cell-matrix interactions are altered due to hepatocyte necrosis or invasion of inflammatory cells, new interactions are established that trigger a fibrogenic response. In the fibrotic liver, significant quantitative and qualitative changes occur in the composition of ECM in the periportal and perisinusoidal areas. During cirrhosis, the amount of fibrillar collagens and proteoglycans can be up to six times higher than in healthy livers [62–63]. The scar is typically composed of fibrillar collagen type I and III, proteoglycans, fibronectin and hyaluronic acid [64]. As a result, alteration in the physiological architecture of the liver occurs, particularly in the space of Disse, where the low electron-dense ECM is replaced by one rich in fibrillar collagens and fibronectin [61]. This leads to loss of endothelial cell fenestrations, impaired exchange of solutes among neighboring cells, altered hepatocyte function and subsequent non-parenchymal cell damage [61]. Previous work from our group has shown that co-culture of Kupffer cells or hepatocytes with HSC induces a more activated phenotype, greater proliferation rates and increased collagen synthesis in HSC compared to HSC cultured alone [31, 32, 65].</p><!><p>Chronic HBV infection and long-term consumption of alcohol induce cell damage through increased generation of ROS [66]. Indeed, oxidative stress, which favors mitochondrial permeability transition, is able to promote hepatocyte necrosis and/or apoptosis. In some clinically relevant conditions, generation of ROS within hepatocytes may represent an altered metabolic state as in non-alcoholic fatty liver disease and non-alcoholic steatohepatitis, or significant ethanol metabolism as it occurs in alcoholic steatohepatitis. ROS are generated mainly via the mitochondrial electron transport chain or via activation of cytochrome P450 -mostly cytochrome P450 2E1-, NADPH oxidase, xanthine oxidase or via mitochondrial damage. The ROS generated can directly affect the HSC and myofibroblasts behavior [32, 67]. ROS up-regulate the expression of critical fibrosis-associated genes such as COL1A1, COL1A2, MCP1 and TIMP1 via activation of signal transduction pathways and transcription factors, including JNK, activator protein-1 and NFκB [18]. ROS generation in HSC and myofibroblasts occurs in response to several known pro-fibrogenic mediators, including angiotensin II, PDGF, TGFβ and leptin [68]. Overall, a decrease in the antioxidant defense such as GSH, catalase or SOD, in conjunction with enhanced lipid peroxidation leads to a pro-fibrogenic response by enhancing collagen I protein expression [69].</p><!><p>The ECM is a highly dynamic milieu subject to constant remodeling whereby synthesis of new components occurs with simultaneous degradation. Life-threatening pathological conditions arise when ECM remodeling becomes excessive or uncontrolled. Among the cells involved in hepatic ECM degradation are HSC, neutrophils and macrophages. MMPs are the main enzymes responsible for ECM degradation and TIMPs have the ability to inhibit MMPs [70]. Therefore, regulation of the MMP-TIMP balance is crucial for efficient ECM remodeling. The MMP-TIMP ratio is tipped because of multiple pro-fibrogenic insults. Activated HSC not only synthesize and secrete ECM proteins such as collagens type I and type III, but also produce MMP1 [71] and MMP13 [72]. However, MMP1 and MMP13 expression decreases as HSC activation progresses, while the activity of other MMPs remains relatively constant, except for MMP2 and MMP9 [73]. The increase in MMP2 activity is associated with distortion of the normal lobular architecture, which further activates HSC [73]. Moreover, activated HSC up-regulate the expression and synthesis of TIMP1 and TIMP2 [74]. TIMP1 not only prevents the degradation of the rapidly increasing ECM by blocking MMPs, but also inhibits the apoptosis of activated HSC [75]. The net result is the deposition of mature collagen fibers within the space of Disse and thus scaring.</p><p>As indicated above, scar formation is regulated by the balance between MMPs and TIMPs, which are induced by ROS and RNS in the CCl4 model of liver fibrosis [76]. This was elucidated in a study using a mutant form of MMP9 that scavenged TIMP1 and inhibited CCl4-induced fibrosis [77]. Thus, TIMP1 activity may be a crucial factor in the regression of fibrosis. Several factors can activate TIMP1, including leptin [78], angiotensin II [79], and sphingosine 1 phosphate [80]. The delineation of the signaling pathways elicited by these factors is important for successful inhibition of TIMP1 activity. Increased liver MMP activity either by recruitment of bone marrow-derived cells or by decreasing TIMP1 level may lead to regression of fibrosis [81].</p><!><p>Despite significant advances in understanding hepatic fibrosis and defining targets for therapy, there are limited anti-fibrotic drugs approved for clinical use in patients with advanced liver disease. Regression of established fibrosis can be accomplished in selected individuals with chronic liver diseases that have effective interferon therapies [82]. However, a large cohort of patients does not respond to conventional treatment and thus remain at risk for progression of fibrosis to cirrhosis.</p><p>Ideally, the anti-fibrotic therapy should be liver-specific, selective for targeting of the fibrogenic cascade, including inhibition of matrix deposition, collagen synthesis, modulation of HSC activation, enhancing matrix degradation, stimulating HSC death or apoptosis, and it should be well tolerated if administered for a prolonged period. Several compounds including colchicine and malotilate, with potential antifibrotic activity, have been studied in human trials but found to be not very effective [83–84]. The ideal anti-fibrotic agent, which should be safe when used over a long period, liver-specific, non-hepatotoxic, for oral administration and inexpensive is not yet available.</p><p>Many agents such as malotilate, genistein, curcumin and silymarin have been shown to be effective in vitro and in experimental animal models [84–86]. Although there are no perfect and effective anti-fibrogenic agents, the potential candidates include agents that can reduce inflammation and the immune response such as corticosteroids, colchicines and IL-10; agents that reduce the activation of ECM-producing cells as well as their pro-fibrogenic potential such as inhibitors of TGFβ1, INFα and peroxisome proliferator activated receptor-γ agonists [87]; antioxidants such as vitamin E [88], phosphatidylcholine [89] and S-adenosyl-L-methionine [90]; N-acetylcysteine and 5-nitroso-N-acetylcysteine [91], which appears to be more efficient, agents capable of increase the degradation of ECM fibrillar such as MMPs and uroplasminogen activator [92] and interferons [93]. Combination therapy that works at different mechanistic levels would be more appropriate to block HSC activation and the pathogenesis of liver fibrosis.</p><!><p>Compared with other antisense strategies such as antisense oligonucleotides, ribozymes or DNAzymes, siRNA has been proved to be a potent knockdown of any given target gene with high sequence specificity [94]. Hence, siRNA would be a powerful strategy to treat liver fibrosis in the future. There are three ways to deliver siRNA: synthetic duplex, plasmid and viral vectors. While viral vectors give high transduction efficiency, their immune reactions limit their application in therapeutics. On the other hand, plasmid DNA complexes with cationic liposomes may not pass through sinusoidal gaps, because endothelial cells fenestrations are closed during liver fibrosis. In contrast, low molecular weight synthetic duplex siRNA is likely to pass through sinusoidal gaps in fibrotic livers, and thus they may be ideal candidates for treating liver fibrosis [95].</p><p>Most of the targeted genes are those critical for HSC activation, proliferation, and/or collagen synthesis and deposition, which are usually markedly up-regulated during hepatic fibrogenesis, including CTGF, TGFβ1, PDGF, TIMPs and plasminogen activator inhibitor-1 [96]. However, these molecules may also play a role in many physiological processes and their inhibition may lead to adverse side effects. Cheng et al. [95] have successfully designed and validated TGFβ1-specific siRNAs and then converted two potent siRNA sequences into shRNAs, which effectively silenced TGFβ1 gene expression in HSC cells. TGFβ1 gene silencing significantly reduced the production of collagen I, TIMP1 and inflammatory cytokines [95]. Their results suggested that silencing TGFβ1 by siRNA and shRNA may be an efficient and more specific approach for treating liver fibrosis [95]. Another study [97] showed that inhibiting the receptor for advanced glycation end products (RAGE) gene, which is involved in migration of activated HSC and myofibroblasts [98], by a specific RAGE siRNA expression vector, inhibited the expression of collagen I in CCl4-induced rat liver fibrosis and dramatically reduced the levels of serum pro-collagen type III, hyaluronic acid and laminin. These results indicated that inhibition of RAGE had important antifibrogenic effects [98]. However, the mechanisms of the effects of RAGE on the accumulation of ECM in hepatic fibrosis need further investigation.</p><p>Using the RAGE specific siRNA strategy, they effectively inhibited RAGE gene expression in rat liver fibrosis and successfully prevented experimental liver fibrosis in rats [98]. The suppression of the up-regulated expression and activity of NFκB and HSC activation via inhibition of IκBα degradation, inhibition of ECM production and attenuation of liver injury may be possible strategies to prevent liver fibrosis by inhibiting RAGE. These findings strongly suggests that RAGE may be a new target for combating liver fibrosis and RAGE specific siRNA might be an effective candidate to prevent liver fibrosis.</p><p>Based on recent studies, micro RNAs (miRNAs) have gained significant interest as diagnostic biomarkers and as therapeutic targets. miRNAs belong to a class of small, non-coding RNA molecules that control protein expression at the post-transcriptional level [99]. Their mechanism of action involves imperfect binding to complementary sequences in the 3'-untranslated region of target mRNAs, leading to either cleavage of the mRNA [100] or suppression of protein translation [101]. In liver fibrosis, activation of HSC, which is regulated by multiple signal transduction pathways, is considered a key event; thus, proteins from the pathways involved could be important targets for miRNAs.</p><p>In order to understand the critical pathways of HSC activation, Guo et al. [102] performed comprehensive comparative bioinformatics analysis of microarrays of quiescent and activated HSC. Changes in miRNAs associated with HSC activation status revealed that 13 pathways were up-regulated and 22 pathways were down-regulated by miRNA. Lee and colleagues dissected a novel mechanism for cystogenesis involving miRNA [103]. They demonstrated that levels of the miRNA, miR15a are decreased in livers of patients with autosomal recessive and autosomal dominant polycystic kidney disease and congenital hepatic fibrosis as well as in the PKC rat model of autosomal recessive polycystic kidney disease. This results in increased expression of the cell-cycle regulator Cdc25A, which is a direct target of miR15a, and increased cellular proliferation and cystogenesis in vitro.</p><p>Roderburg et al. [104] studied the regulation of miRNAs in experimentally induced hepatic fibrosis by microarray. They found that three members of the miR-29-family are significantly down-regulated in livers of CCl4-treated mice as well as in mice that underwent bile duct ligation [104]. Specifically, their data indicate that miR-29 mediates the regulation of liver fibrosis and it is part of a signaling nexus involving TGFβ- and NFκB-dependent down-regulation of miR-29 family members in HSC with the subsequent up-regulation of ECM genes [104]. Therefore, identifying potential miRNAs for the arrest of HSC activation and proliferation could lead to therapeutic intervention of hepatic fibrosis.</p><!><p>Understanding the mechanisms and the pathways involved in the pathogenesis of the fibrogenic response could provide novel therapeutic approaches for liver diseases in which fibrosis is a detrimental component. Identifying potential new therapeutic agents that target specific pathways involved in fibrosis is critical to overcome the barriers to progress in the field of liver fibrosis. Thus, prevention of hepatic fibrosis could help ameliorate the complications of cirrhosis, and improve the quality of life of many patients worldwide.</p><!><p>We evaluated different aspects of the etiology of liver fibrosis</p><p>Delineated the molecular mechanism of the pathogenesis of hepatic fibrosis</p><p>Discussed all current therapeutic approaches including molecular gene therapy</p><p>This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.</p><p> Conflict of interest </p><p>The authors have no conflicts of interest to disclose</p>

PubMed Author Manuscript

Cleavage of Peptide Bonds bearing Ionizable Amino Acids at P1 by Serine Proteases with Hydrophobic S1 Pocket

Enzymatic hydrolysis of the synthetic substrate succinyl-Ala-Ala-Pro-Xxx-pNA (where Xxx = Leu, Asp or Lys) catalyzed by bovine chymotrypsin (CHYM) or Streptomyces griseus protease B (SGPB) has been studied at different pH values in the pH range 3 to 11. The pH optima for substrates having Leu, Asp, and Lys have been found to be 7.5\xe2\x80\x938.0, 5.5\xe2\x80\x936.0 and ~ 10, respectively. At the normally reported pH optimum (pH 7 to 8) of CHYM and SGPB, the substrate with Leu at the reactive site is more than 25,000-fold more reactive than that with Asp. However, when fully protonated, Asp is nearly as good a substrate as Leu. The pK values of the side chains of Asp and Lys in the hydrophobic S1 pocket of CHYM and SGPB have been calculated from pH-dependent hydrolysis data and have been found to be about 9 for Asp and 7.4 and 9.7 for Lys for CHYM and SGPB respectively. The results presented in this communication suggest a possible application of CHYM like enzymes in cleaving peptide bonds contributed by acidic amino acids between pH 5 and 6.

cleavage_of_peptide_bonds_bearing_ionizable_amino_acids_at_p1_by_serine_proteases_with_hydrophobic_s

INTRODUCTION<!>MATERIALS AND METHODS<!>RESULTS AND DISCUSSION

<p>Serine proteases are ubiquitous and exist in a wide spectrum of specificities [1, 2] with the capability of hydrolyzing virtually any peptide bond. Serine proteases that preferentially hydrolyze peptide bonds contributed by hydrophobic (Ala, Val, Leu, Ile, Met, Phe, Tyr, Trp), cationic (Lys, Arg), anionic (Asp, Glu), and polar (Ser, Gln) side chains at P1 (Schechter-Berger nomenclature [3]) as well as proline at P1 are known to exist [1]. Of these speicific classes of serine proteases, the ones that hydrolyze at hydrophobic and cationic amino acids at P1 are the best characterized.</p><p>A great deal has been learned about the size, shape, and chemistry of the primary specificity pocket (S1 pocket) of serine proteases by studying their interaction with Standard Mechanism protein inhibitors [4-6]. The inhibitor-protease complex mimics the substrate protease transition state complex, and, therefore, the good correlation often found [7,8] between association equilibrium constant (Ka) values for a set of inhibitor variants and the corresponding kcat/Km values for substrates is not surprising. In Standard Mechanism inhibitors the reactive site peptide bond never shifts, even if the P1 residue is deleterious for the target enzyme [9, 10]. This is not the case with substrates, which can readily shift their scissile peptide bond (their reactive site) so as to undergo cleavage according to the specificity of the serine protease. Thus serine proteases with a hydrophobic S1 pocket, such as bovine chymotrypsin (CHYM) and Streptomyces griseus protease B (SGPB), have not been reported to hydrolyze peptide bonds in which the P1 residue is contributed by Asp or Glu. In contrast to substrates, Standard Mechanism inhibitors with a P1 Asp or Glu are able to bind, albeit weakly, with CHYM or SGPB [4]. Our investigations on the binding of serine proteases to variants of turkey ovomucoid third domain (OMTKY3) containing all possible ionizable residues at P1 led to the finding of large shifts of the pKa values of these P1 residues when in the inhibitor-protease complex [11]. We ask a similar question here for the hydrolysis of synthetic substrates bearing Asp and Lys at the reactive site by two serine proteases that have a neutral hydrophobic S1 pocket, CHYM and SGPB. The pH-dependence of the hydrolysis of these substrates reveals interesting results that are discussed in this paper.</p><!><p>TPCK treated bovine chymotrypsin was obtained from Sigma. SGPB was purified from a commercial (Sigma) preparation of Pronase as described earlier [12]. The purity of SGPB was confirmed by ion exchange and size exclusion column chromatographies, while its identity was confirmed by amino acid composition determination. Synthetic protease substrates, such as Suc-Ala-Ala-Pro-Leu-pNA (Suc-AAPL-pNA), Suc-Ala-Ala-Pro-Asp-pNA (Suc-AAPD-pNA), and Suc-Ala-Ala-Pro-Lys-pNA (Suc-AAPK-pNA), were purchased from BACHEM. Various buffers and solutions used with their pH ranges in parentheses were: 0.1 M NaCl-NaOH (pH 9.5–11.0), 0.1 M Tris-HCl (pH 7.0–9.0), 0.1 M Bis-Tris (pH 6.0–7.0), 0.1 M sodium acetate-acetic acid (pH 4.0–5.5), 0.1 M glycine buffer (pH 3.0–3.5). Stock solutions of 100 mM substrates were prepared in dimethyl sulfoxide (DMSO).</p><p>Kinetics of enzymatic hydrolysis of substrates were performed at different pH values and at 22 °C by following the change in absorbance over a period of 120–300 s in the wavelength range 380–410 nm on a Hewlett-Packard HP8453 diode array spectrophotometer. The kinetic data were automatically corrected for any light scattering by continuously subtracting the absorbance values in the wavelength range 650–700 nm during the kinetic run. The rates of hydrolysis of substrates obtained as change in absorbance per second were converted to moles of product (p-nitroaniline) produced per second (Vo) by using a molar extinction coefficient of 8,800 M-1cm-1 [13] for p-nitroaniline. Km and kcat values were calculated from the rates of hydrolysis of substrates at different substrate concentrations using the Lineweaver-Burk equation. The expression of the P1Q variant of OMTKY3 and the procedure for the measurement of the association equilibrium constant are described in an earlier publication [4].</p><!><p>Hydrolysis of Suc-AAPL-pNA by CHYM and SGPB was studied at different pH values between pH 4.0 and pH 10.0 (Fig. 1A). CHYM and many other serine proteases show maximum enzymatic activity between pH 7 and 8. The decline in enzyme activity below pH 7 is well known and has been attributed to protonation of catalytic histidine 57 (chymotrypsinogen numbering). The decrease in activity above pH 8 can be attributed to the deprotonation of the α-amino group of Ile16 and the disruption of charge-charge interaction between Ile16 and the side chain carboxyl group of Asp194 [14]. The salt bridge between Ile16 and Asp194 is characteristic of CHYM like enzymes but is either absent in other serine proteases or is substituted by a different type of salt bridge. For example, SGPB, a bacterial serine protease, has a salt bridge involving Asp194 and the side chain of Arg138 [14]. Because the arginine side chain has a very high pKa value (>12), the interaction between Asp194 and Arg138 in SGPB is unlikely to be disrupted significantly below pH 11. This is consistent with the observation that the pH-dependence of kcat/Km for Suc-AAPL-pNA with SGPB shows a flat region between pH 8 and pH 10 (Fig. 1A). Substrates with other neutral hydrophobic side chains (such as Ala, Phe, Val) in place of Leu gave results similar to those shown in Fig. 1A.</p><p>The pH-dependence of the association equilibrium constant, Ka, of inhibitors with non-ionizable residues at P1 has the same shape as that for the pH-dependence of the hydrolysis of Suc-AAPL-pNA (Fig. 1A). The pH-dependence of Ka of P1 variants of OMTKY3 with different serine proteases has been extensively studied in our laboratory at Purdue. As an example, the pH-dependence of Ka of P1Q variant of OMTKY3 with CHYM and SGPB is shown here (Fig. 1B). The strong correlation between free energy of association of inhibitors (log Ka) and transition state free energies of substrates (log kcat/Km) has been observed by us as well as by others [4,8,15,16].</p><p>The two serine proteases used here have a hydrophobic S1 pocket and therefore have a preference for amino acids with hydrophobic side chains at P1. Thus OMTKY3 variants with Asp, Glu, Lys or Arg at P1 bind weakly to these serine proteases [4]. We performed pH dependent measurements of association equilibrium constants of such inhibitors with different serine proteases and found that when a charged side chain is present at P1, it undergoes large pK shifts in the direction that makes it neutral in the S1 pocket. Thus Asp and Glu undergo large increases in their pKa values whereas Lys undergoes a decrease in its pKa value [11, 15].</p><p>The pH-dependence of kcat/Km determined for Suc-AAPD-pNA and Suc-AAPK-pNA with CHYM and SGPB (Fig. 2) shows a large shift in the optimum pH with both enzymes. When Asp is present at the scissile peptide bond, the optimum shifts to pH 5.0–5.5, whereas when Lys is present, the pH optimum is 10 or higher. These shifts are consistent with the binding of Asp and Lys in their neutral state. Suc-AAPD-pNA is an extremely poor substrate for CHYM and SGPB at pH 8 but becomes a moderately good substrate (~100 fold better) at pH 5.5. The pH-dependence of Asp and Lys substrates can be used to determine the pKa values of these two residues when they are inserted in the S1 pocket of the enzyme (designated as pKc) in the transition state complex. We have used a procedure we initially described for inhibitors [11] to determine the pKc values of Asp and Lys in the substrates. In this procedure an assumption is made that the difference in pH-dependence of log kcat/Km of two similar substrates, one with an ionizable side chain at P1 and the other with a nonionizable side chain at P1, can solely be explained by the pKa value of ionizable group in free-state (pKf) and in bound state (pKc). Free and bound states refer respectively to the substrate that is not bound to a protease and to the substrate that is bound to a protease. We used Suc-AAPL-pNA with Leu at P1 as our representative nonionizble P1 residue. As an example, plots obtained with CHYM for Suc-AAPD-pNA and Suc-AAPK-pNA are shown in Fig. 3. In these plots, the difference in log kcat/Km values between Suc-AAPL-pNA and Suc-AAPD-pNA (or Suc-AAPK-pNA) designated as log R is plotted against pH. The pKf values for Suc-AAPD-pNA and Suc-AAPK-pNA were determined by NMR spectroscopy [17]. The plot is fitted to the equation we developed earlier [11] for the pH-dependence of association equilibrium constants for the interaction of P1 variants of OMTKY3 with serine proteases; (1)LogR=LogRo+Log(1+10(pH-pKc))−Log(1+10(pH-pKf)) In this equation, Ro is the ratio of kcat/Km values for an ionizable residue to a nonionizable residue at a pH where the ionizable residue is completely protonated and pKf and pKc are the pKa values of the ionizable group in the substrate and enzyme-substrate transition state respectively.</p><p>Fitting the data according to the relation given above allowed us to calculate the pKc values for the ionizable residues, Asp and Lys, in the transition state complex of the substrates with the enzymes (Table 1). It is interesting to note that in the transition state complex with CHYM, Lys is a stronger acid (pKc = 7.44) than Asp (pKc = 8.92). The pK shifts of the Asp side chain were nearly identical for CHYM and SGPB (pKc – pKf = ΔpK = 5.0 ± 0.1) and represent one of the largest shifts reported for acidic amino acid side chains in proteins [11, 18]. The pK shifts for the Lys side chain were quite different for CHYM and SGPB (see Table 1). The probable cause of this difference is the flexibility of the lysine side chain and its ability to adopt different conformations in the S1 pocket of proteases. In an earlier study we found that the pK shift for the P1 lysine of BPTI is much smaller than that for the P1 Lys variant of OMTKY3 in complex with CHYM [15]. The molecular explanation emerged from comparison of the X-ray crystallographic structures of the complexes with CHYM of BPTI [19, 20] and P1Lys-OMTKY3 [1hja, unpublished]. In the P1Lys-OMTKY3-CHYM complex, the P1Lys side chain is inserted deep into S1 pocket of CHYM, whereas in the BPTI-CHYM complex, the P1Lys side chain goes into the S1 pocket but bends back in such a manner that the ε-amino group is able to make a couple of hydrogen bonds with the backbone oxygens of Ser217 of the enzyme and Pro13 (P3) of the inhibitor. The difference in the microenvironment of Lys in the BPTI-CHYM complex and in the P1Lys-OMTKY3-CHYM complex accounts for the difference in the pKc value of these two lysine side chains. Thus it is likely that the a similar mechanism accounts for the difference in the pKc values of Lys in Suc-AAPK-pNA in its transition state complexes with CHYM and SGPB.</p><p>Why chymotrypsin-like enzymes are not reported to hydrolyze peptide bonds contributed by Asp and Glu is obviously due to extremely poor activity of these enzymes towards such peptides in the pH range 7 to 8. For example, at pH 8, kcat/Km for the hydrolysis of Suc-AAPD-pNA by CHYM was found to be 0.74 M-1s-1 compared to 25,000 M-1s-1 for Suc-AAPL-pNA. This is because in order bind to the S1 pocket of CHYM, the Asp side chain must become protonated – energetically an expensive process at pH above 7. However, at low pH values (pH = 4), kcat/Km for Suc-AAPD-pNA approaches that of Suc-AAPL-pNA (10.7 M-1s-1 vs 28.2 M-1s-1). This is also indicated by a near zero log R value at low pH values (log R value of zero means that kcat/Km for Suc-AAPD-pNA and Suc-AAPL-pNA are equal).</p><p>The pKc values reported here are very similar to the one found for P1 Asp and Lys variants of OMTKY3 in complex with SGPB, CHYM, porcine pancreatic elastase and subtilisn Carlsberg [11, 15, Qasim and Laskowski – unpublished). Thus, the strong correlation reported between association equilibrium constants in inhibitors and kcat/Km values in substrates at pH 8.3 [4, 8, 15, 16] for P1 variants extends over the whole ionizable range of pH 3 to 11.</p><p>The results presented here clearly show that the pH optimum of a neutral serine protease is dependent on the nature of the side chain at the scissile peptide bond. The large preferential increase in the catalytic activity of neutral serine proteases for acidic amino acid side chains at pH 5.0 to 5.5 may have a physiological role. For example, granulocytes are rich in chymotrypsin-like serine proteases and have a pH between 5.0 and 6.0 [21] – a pH range well suited for cleavage at peptide bonds contributed by Asp and Glu. Regardless of any physiological role of the action of serine proteases on acidic amino acids at low pH, the results presented here suggest that such proteases can be used for cleavage at acidic amino acids by performing the cleavage at an appropriate low pH.</p>

PubMed Author Manuscript

Monofunctional Platinum(II) Compounds and Nucleolar Stress: Is Phenanthriplatin Unique?

Platinum anticancer therapeutics are widely used in a variety of chemotherapy regimens. Recent work has revealed that the cytotoxicity of oxaliplatin and phenanthriplatin is through induction of ribosome biogenesis stress pathways, differentiating them from cisplatin and other compounds that mainly work through DNA damage response mechanisms. In order to probe the structure-activity relationships in phenanthriplatin\xe2\x80\x99s ability to cause nucleolar stress, a series of monofunctional platinum(II) compounds differing in ring number, size and orientation were tested by nucleophosmin (NPM1) relocalization assays using A549 cells. Phenanthriplatin was found to be unique among these compounds in inducing NPM1 relocalization. To decipher underlying reasons, computational predictions of steric bulk, platinum(II)-compound surface length and hydrophobicity were performed for all compounds. Of the monofunctional platinum(II) compounds tested, phenanthriplatin has the highest calculated hydrophobicity and volume but does not exhibit the largest distance from platinum(II) to the surface. Thus, spatial orientation and/or hydrophobicity caused by the presence of a third aromatic ring may be significant factors in the ability of phenanthriplatin to cause nucleolar stress.

monofunctional_platinum(ii)_compounds_and_nucleolar_stress:_is_phenanthriplatin_unique?

Introduction<!>Reagents and Synthesis<!>Cell culture and treatment<!>Immunofluorescence<!>Imaging and quantification<!>Computations<!>Oxaliplatin and Phenanthriplatin cause NPM1 relocalization<!>Picoplatin does not cause NPM1 relocalization<!>NPM1 relocalization is not a general property of monofunctional platinum compounds<!>Steric bulk is not sufficient to predict NPM1 relocalization<!>Hydrophobicity is not sufficient for predicting NPM1 relocalization<!>Conclusions

<p>Platinum-based drugs are an important class of chemotherapeutics. After the initial discovery of the anti-proliferative capabilities of cisplatin, the drug was FDA-approved in 1978 and continues to be in significant use over 40 years later [1]. Two additional Pt(II) compounds were subsequently approved by the FDA, carboplatin in 1989 and oxaliplatin in 1996. Improvements upon these three drugs have been attempted and some new compounds even entered into clinical trials, but none have been approved by the FDA [2].</p><p>The three FDA-approved drugs are all considered classical platinum compounds. The characteristics of classical compounds are a result of early structure-activity relationship (SAR) studies that determined the necessary properties for platinum compounds to exhibit anti-proliferation activity [3]. These required components are that the platinum compound be square planar, have a neutral overall charge, and contain two non-labile cis-am(m)ines and two labile cis anionic ligands. Although these rules led to the drugs that are used today, research into compounds that would not be within a traditional SAR study have produced non-classical platinum drugs with anti-proliferative activity. These non-classical compounds include Pt(IV) prodrugs, monofunctional, trans-platinum, polyplatinum, and tethered platinum complexes [3, 4]. One of the most effective and well-studied non-classical compound is the monofunctional Pt(II) phenanthriplatin [4, 5] (Figure 1). In addition to having only a single exchangeable anionic ligand, the N-heterocyclic ligand of phenanthriplatin and others of this class, such as pyriplatin (Figure 1), are rotated perpendicular to the square-planar Pt ligand plane.</p><p>Phenanthriplatin has exhibited unique activity in the NCI-60 cell line screen when compared to other platinum chemotherapeutics [5]. Phenanthriplatin is significantly more potent with a 7–40x higher toxicity than cisplatin [3–5]. It has higher cellular uptake than cisplatin or pyriplatin [5]. In addition, the phenanthridine ligand of phenanthriplatin may facilitate rapid DNA binding through reversible intercalation between nucleobases before platinum binding occurs [6]. Studies have also revealed some of the biological targets of phenanthriplatin. It has been shown to act as a topoisomerase II poison [7]. Phenanthriplatin also was demonstrated to inhibit RNA polymerase II [8], but allows efficient DNA polymerase η bypass [9]. Overall, these studies have shown that phenanthriplatin can affect biological processes in a variety of ways, and this has led researchers to suggest that the effectiveness of the compound is through multiple cellular pathways [10].</p><p>In a recent study, the classical platinum compound oxaliplatin and non-classical phenanthriplatin were both shown to induce ribosome biogenesis stress as the primary pathway to cell death [11]. This surprising observation is in contrast with cisplatin and carboplatin, which were shown to cause cell death through DNA damage as is expected for classical compounds. The ability to induce nucleolar stress shared between oxaliplatin and phenanthriplatin is perplexing considering the major structural differences between the two compounds. We endeavored to determine whether there were structural similarities between these two molecules which would explain this similar activity, and determine whether the ability to induce nucleolar stress was inherent to the family of non-classical monofunctional platinum(II) compounds. In order to do this, we synthesized a suite of monofunctional and related platinum compounds (Figure 1) and analyzed their ability to cause nucleolar stress using an assay for nucleolar phosphoprotein B23 (NPM1) relocalization. We further compared structural and electronic properties of these compounds based on DFT calculations. We find that phenanthriplatin, but not related quinoplatin or isoquinoplatin, induces nucleolar stress as measured by NPM1 relocalization in human lung carcinoma A549 cells. Although phenanthriplatin has the largest total volume and hydrophobicity of the compounds tested, quinoplatin and isoquinoplatin may have similar potential to disrupt intermolecular interactions based on Pt-ligand distances. We conclude that the unique ability of phenanthriplatin to induce nucleolar stress is conferred by the third aromatic ring. The ligand disposition of these monofunctional N-heterocyclic Pt(II) compounds is sufficiently different from oxaliplatin to suggest that separate properties of oxaliplatin and phenanthriplatin lead to their abilities to both cause nucleolar stress.</p><!><p>Cisplatin [12], picoplatin [13], and pyriplatin, quinoplatin, isoquinoplatin, and phenanthriplatin [5] were synthesized as previously reported. Oxaliplatin was purchased from TCI America. Actinomycin D was purchased from Thermo Fisher Scientific. A549 cell line was acquired from the American Type Culture Collection.</p><!><p>A549 human lung carcinoma cells (#CCL-185, American Type Culture Collection) were cultured at 37 C°, 5% CO2 in DMEM supplemented with 10% FBS and 1% antibiotic-antimycotic. A549 cells have been used previously to study nucleolar stress pathways [14, 15]. Cells between passage 11–25 and at confluency of 70% were used in the treatments. Cells were treated for 24 hours with 10 µM compound, with the exception of phenanthriplatin and phenanthridine which were administered at 0.5 µM and actinomycin D at 5 nM. Stock solutions of 5 mM compound in DMF were made and used with the exception of oxaliplatin, which was made in water and actinomycin D which was made in DMSO. Immediately prior to treatment, platinum compounds were diluted into media. Final DMF and DMSO concentrations were 0.2% in media. We chose to use 0.5 µM phenanthriplatin to account for the higher cellular accumulation of phenanthriplatin and to be more in line with reported 72 hour IC50 values which are not exhibited by the other studied compounds [5].</p><!><p>Cells to be imaged were grown on coverslips (Ted Pella product no 260368, Round glass coverslips, 10mm diam, 0.16–0.19mm thick) as described above. Following treatment, cells were washed twice with PBS. They were then fixed for 20 minutes at room temperature in 4% paraformaldehyde diluted in PBS. Cells were permeabilized with 0.5% Triton-X in PBS for 20 minutes at room temperature followed by two 10-minute blocking steps with 1% BSA in PBST. The cells were incubated for 1 hr using primary antibody (NPM1 Monoclonal Antibody, FC-61991, from ThermoFisher, 1:200 dilution in PBST with 1% BSA) and 1 hour in secondary antibody (Goat Anti-Mouse IgG H&L Alexa Fluor® 488, ab150113, Abcam, 1:1000 dilution in PBST with 1%BSA). Between each incubation and before mounting, slides were washed 3 times for 5 min each using PBST. Coverslips were mounted on slides with ProLong™ Diamond Antifade Mountant with DAPI (Thermo Fisher) according to manufacturer's instructions.</p><!><p>Images were taken using a HC PL Fluotar 63x/1.3 oil objective mounted on a Leica DMi8 fluorescence microscope with Leica Application Suite X software. Quantification of NPM1 relocalization was performed in an automated fashion using a Python 3 script. Images were preprocessed in ImageJ [16, 17] to convert the DAPI and NPM1 channels into separate 16-bit greyscale images. Between 50–250 cells were analyzed for each treatment group. Nuclei segmentation was determined with the DAPI images using Li thresholding functions in the Scikit-Image Python package [18]. The coefficient of variation (CV) for individual nuclei, defined as the standard deviation in pixel intensity divided by the mean pixel intensity, was calculated from the NPM1 images using the SciPy Python package. All data was normalized to the no treatment control in each experiment. NPM1 imaging results for each compound were observed on two separate testing days. Duplicates of treatments were performed and analyzed and are available upon request from the corresponding author.</p><!><p>Based on the experimental results, we hypothesized that the size, shape or hydrophobicity of the platinum(II) compounds may be instructive in correlating the biological activity with the chemical structure because of biological implications of these structural components in an interaction between two biomolecules that may be disrupted. Thus, we optimized all platinum(II) compounds using density functional theory (DFT) as implemented in Gaussian09 [19] so that we might quantitatively assess the structural differences and hydrophobicity of the compounds.</p><p>Geometry optimizations were performed with an RMS force convergence criterion of 10−5 hartree. The electronic wavefunction was minimized using the GGA functional PBE [20, 21], with the DEF2TZP basis set. Relativistic effects were not explicitly included, however, these were not expected to significantly impact the geometries of the platinum(II) complexes.[22] Solvent was implicitly included using the Solvent Model Density method [23].</p><p>The solvent-dependent difference in Gibbs free energies (ΔGH2O−Octanol) was calculated using ΔGH2O−Octanol=ΔGH2O−ΔGOctanol where (ΔGH2O) and (ΔGOctanol) are the change in free energies of the system in water and n-octanol, respectively. (ΔGH2O) was computed using the structure optimized in the pseudo solvent, water. This optimized structure was kept constant for all subsequent computations, including calculation of the compound in pseudo solvent, n-octanol, which yielded (ΔGOctanol). This approach minimizes the reorganizational energetic differences. Thus, (ΔGH2O−Octanol) is a measure of the hydrophobicity for each compound.</p><p>Further calculations were required to assess the size and shape of the platinum(II) compounds. Two measures of size were considered, i) volume, and ii) the longest vector between the platinum atom and the surface of the molecule. The latter characteristic represents the main steric component of the ligand in each compound.</p><p>To quantitatively assess the volume of each compound, a definition of size is necessary. Thus, we will use the presence of electron density to signify the location of the chemical system. Since DFT yields both the electron density and electrostatic potential of the optimized, non-hydrolyzed platinum(II) compound structures and we have previously developed a tool to analyze the electrostatic potential of chemical systems[24], we will use the same file format to analyze the electrostatic potential. As a result, the electrostatic potentials of the optimized structures were computed by minimizing the electronic wavefunction using a 500 eV planewave cutoff, a gamma-only k-grid, and the PBE [20, 21] functional utilizing a plane-augmented wave (PAW) [25, 26] basis as implemented in the Vienna Ab initio Software Package (VASP) [27–30]. All compounds were calculated within a sufficiently large computational box to minimize self-interaction.</p><p>The electric field is the gradient of the electrostatic potential; thus, the electric field embodies the direction of greatest increase in electrostatic potential. This is significant because the increased slope of the electric field enables us to more clearly define the edge of a chemical system in space. Therefore, deriving the electric field from the electrostatic potential returned by DFT allows us to assess the size of each compound by sampling the electric field. However, in order to achieve this the definition of a surface needs to be addressed.</p><p>We will define the edge of a chemical system as the point where the electric field magnitude no longer changes, which is intuitive considering the definition of the electrostatic potential. Since DFT calculations return electrostatic potential values on the order of 10−6 eV, a change in less than 10−5 eV is considered negligible. This approach is based on previous atomic radii calculations, which employ negligible change in electron density to assess the size of atoms [25,26].</p><p>Using the area of each compound defined by sampling the electric field, the longest vector between the platinum atom and the surface was calculated for each compound, capturing the main steric component of each ligand.</p><!><p>A previous study examining cell death mechanisms of phenanthriplatin (1) and oxaliplatin (2) has shown that both compounds cause cell death through ribosome biogenesis stress [11]. For the current studies we monitored NPM1 relocalization from the nucleolus to the nucleoplasm, which is a hallmark of nucleolar stress resulting from the disruption of ribosome biogenesis [31]. Under non-stressed conditions NPM1 is localized to the nucleolus; however, NPM1 is distributed throughout the nucleoplasm following nucleolar stress. We set out to measure the extent of NPM1 relocalization when cells were treated with a series of platinum compounds with cyclic ligands and either monofunctional or bifunctional substitution properties.</p><p>We first examined NPM1 relocalization following treatment with oxaliplatin and phenanthriplatin. As expected, known ribosome biogenesis stress inducer actinomycin D caused NPM1 relocalization to the nucleoplasm while the negative no-treatment control showed NPM1 localized in the nucleoli (Figure 2A). Both oxaliplatin and phenanthriplatin caused relocalization of NPM1 throughout the nucleus, confirming their ability to cause nucleolar stress as previously reported [11].</p><p>In order to determine the extent of nucleolar stress we quantified the heterogeneity of nuclear NPM1 intensity distribution by its coefficient of variation (CV). The CV is the standard deviation of the pixel intensity populations corresponding to NPM1-based immunofluorescence normalized by the mean intensity of each nucleus. In cells that were undergoing stress, NPM1 is relatively evenly diffused throughout the nucleus, leading to homogeneous intensities and a small CV. Histograms of representative cells show a large population of medium intensity pixels across the cell for compounds that cause NPM1 relocalization (figure 2B). For cells that are not undergoing stress, NPM1 is concentrated in the periphery of the nucleolus while being absent in the nucleoplasm, resulting in a heterogeneous population of pixel intensities and a high CV. Histograms of cell images from compounds that do not cause NPM1 relocalization show large populations at the two extremes of the pixel intensity which would result in a large CV (figure 2B). CVs were calculated for each cell in a population and the distribution of these CVs was evaluated for each treatment condition. Corresponding to our representative NPM1 images (Figure 1A), compounds that caused no NPM1 redistribution had median CVs around 1 (when normalized to the no treatment control) while compounds that caused NPM1 relocalization had medians lower than 0.6 (normalized to the no treatment control). NPM1 relocalization was observed upon treatment with oxaliplatin, phenanthriplatin and actinomycin D (figure 2C). Additionally, treatment of the ligand phenanthridine alone is not sufficient to induce nucleolar stress (Figure 2C).</p><!><p>There are large structural differences between oxaliplatin and phenanthriplatin; however, these disparate compounds are both able to activate nucleolar stress pathways whereas cisplatin does not. Both the DACH ligand of oxaliplatin and the phenanthridine ligand of phenanthriplatin add significant steric bulk in comparison with cisplatin. However, phenanthriplatin is a monofunctional compound. In addition, unlike the case of oxaliplatin, in phenanthriplatin the phenanthridine rings are oriented perpendicular to the square-planar Pt ligand plane [5]. Picoplatin (3) is one compound that bridges these differences in that the picoline ring is oriented perpendicular to the platinum plane [32]. Picoplatin is also a classical bifunctional platinum compound and enabled us to determine whether the added ligand bulk regardless of orientation was sufficient to induce NPM1 relocalization. In A549 cells treated with picoplatin, NPM1 did not relocalize to the nucleoplasm (figure 2A) as quantified by a median CV of around 1 (figure 2C), indicating that picoplatin does not cause nucleolar stress.</p><!><p>After determining that the classical compound picoplatin did not cause NPM1 relocalization despite having some similarities to oxaliplatin in terms of added ring and steric bulk, we next examined the properties of non-classical monofunctional platinum compounds. We synthesized three additional monofunctional compounds that have one or two aromatic rings in order to test whether nucleolar stress was inherent to ring-containing monofunctional platinum(II) compounds as a whole or whether it was a phenomenon only exhibited by phenanthriplatin.</p><p>We had tested picoplatin and determined that the perpendicular orientation of the picoline ligand is not sufficient to cause NPM1 relocalization. To further explore the influence of ligand orientation and the binding mode of platinum, we next tested pyriplatin (4). Similar to picoplatin, pyriplatin contains a single aromatic ring. However, unlike picoplatin, pyriplatin has more possible orientations of the aromatic ring due to lack of steric clashes involving the methyl of the picoline [33]. In addition, pyriplatin is more similar to phenanthriplatin in being a monofunctional compound with an overall positive charge. Following a 24 hour treatment at 10 µM, pyriplatin did not cause NPM1 relocalization and samples had a median CV of around 1 (figure 2). From this, we concluded that the ability to cause NPM1 relocalization was not inherent to the class of monofunctional platinum(II) compounds containing N-heterocyclic ligands.</p><p>We next considered whether steric bulk was a factor in NPM1 relocalization by examining the influence of the addition of a second ring. We synthesized the structural isomers quinoplatin (5) and isoquinoplatin (6) (figure 3), in order to test whether a second aromatic ring would be sufficient to cause NPM1 relocalization. We tested these compounds using the NPM1 assay and determined that neither quinoplatin nor isoquinoplatin caused increased NPM1 relocalization, with NPM1 intensities from cells treated with both compounds having a median CV of around 1 (figure 2). From this we concluded that for monofunctional Pt(II) compounds, the steric bulk from a second ring alone does not induce NPM1 relocalization regardless of ring orientation. This added further evidence that NPM1 relocalization was not an inherent property of this non-classical class of platinum compounds and was unique to phenanthriplatin.</p><!><p>From our data, we have determined that phenanthriplatin and oxaliplatin are unique to our suite of compounds. We next examined whether there are any trends present in steric bulk that could explain whether compounds caused NPM1 relocalization. All platinum(II) compounds were optimized using DFT (figure 4) and two variables were calculated to assess steric bulk. First, the volume of the optimized, non-hydrolyzed structure is obtained by sampling the respective electrostatic potential (table 1). Oxaliplatin and phenanthriplatin were the compounds with the largest volume; however, this included the aquation-labile ligands which accounts for a large portion of oxaliplatin's volume.</p><p>Second, the magnitude of the maximum vector between platinum and the surface of the compound, where the surface of the compound is defined as the extent to which the electrostatic potential permeates in space (table 1), was calculated. No trend was found with these distance measurements. Oxaliplatin, which caused NPM1 relocalization, had a similar maximum distance as that of quinoplatin, which did not cause NPM1 relocalization. Additionally, phenanthriplatin, which caused NPM1 relocalization had a similar distance to that of isoquinoplatin which did not cause NPM1 relocalization (table 1). Thus, while phenanthriplatin exhibits the largest steric bulk, it does not have the maximum steric reach from platinum to the surface of the compound.</p><!><p>Hydrophobicity of the non-labile ligand may be an important factor in interrupting biomolecular interactions, or in partitioning into cellular compartments or regions of the nucleolus. We examined if there was a trend in hydrophobicity that would explain why oxaliplatin and phenanthriplatin caused NPM1 relocalization while all other compounds in our library did not. We used our optimized structures to calculate ΔGH2O−Octanol (table 2). As expected, compounds with more aromatic rings were more hydrophobic and had more positive differences in ΔGH2O−Octanol, while compounds with less rings showed the opposite trend. Phenanthriplatin is more hydrophobic than all other compounds except picoplatin, which does not cause NPM1 relocalization and is the most hydrophobic compound tested with a Gibbs solvation energy of 2.54 kcal/mol. Overall, this measure of hydrophobicity was not able to produce a trend that provides a satisfactory explanation for why oxaliplatin and phenanthriplatin cause NPM1 relocalization while others did not. Therefore, we conclude that hydrophobicity alone is not sufficient for causing NPM1 relocalization.</p><!><p>This work aimed to find a structural relationship between oxaliplatin and phenanthriplatin which would provide information on necessary and sufficient structural components required for these platinum compounds to induce cell death via nucleolar stress. In comparison with cisplatin, which does not cause nucleolar stress, oxaliplatin and phenanthriplatin both have significantly larger ring-containing ligands. Phenanthriplatin is also a monofunctional Pt(II) compound. In order to explore this question, we synthesized a library of ring-containing platinum compounds, most being monofunctional Pt(II) compounds. This library was tested for the ability to induce nucleolar stress by monitoring NPM1 relocalization, and quantifying the resulting images. First, we tested oxaliplatin and phenanthriplatin to confirm that they caused NPM1 relocalization in agreement with previous literature proposing that they cause nucleolar stress [11]. We then tested whether a heterocyclic ligand oriented perpendicular to the square-planar platinum(II) ligand plane would be sufficient by testing picoplatin, and found that picoplatin did not cause nucleolar stress as measured by NPM1 relocalization. Thus, for bifunctional platinum compounds, a ring is insufficient to cause nucleolar stress.</p><p>We investigated the importance of ring number and distribution in other compounds of the monofunctional platinum(II) class by testing pyriplatin, quinoplatin and isoquinoplatin. None of these compounds caused NPM1 relocalization, indicating that phenanthriplatin was unique in this class of monofunctional compounds. We note that this limited study has been performed at a single concentration and treatment time for all compounds. It is possible that longer treatment time or higher concentrations might lead to different effects, and this is being explored in further studies. None of the non-phenanthriplatin compounds cause significant levels nucleolar stress at relatively high (10 µM) treatment concentrations, indicating that they are in a different class than phenanthriplatin in terms of activities.</p><p>We performed DFT calculations to optimize structures and calculate the solvent-dependent difference in Gibbs free energy between water and n-octanol, a measure of hydrophobicity. To further investigate structural characteristics, we calculated the maximum distance from the platinum atom to the surface of each structure and volume from the DFT optimized structures. We found no correlation between this distance and the ability to cause NPM1 relocalization. Further, there was no strong correlation between the solvent-dependent difference in Gibbs free energy between water and octanol for compounds that were able to induce NPM1 relocalization.</p><p>In view of these results, we suggest that phenanthriplatin is a unique compound in the monofunctional platinum(II) compound class in its ability to cause NPM1 relocalization. We suggest that the addition of a third aromatic ring in phenanthriplatin may play a large role in differentiating phenanthriplatin from other monofunctional platinum(II) compounds we tested for inducing nucleolar stress. The presence of a third aromatic ring increases steric bulk both above and below the square-planar platinum ligand plane. Additionally, a third ring increases hydrophobicity and provides intercalation potential to phenanthriplatin [6] in comparison to quinoplatin and isoquinoplatin. Phenanthriplatin exhibited the largest volume and was the most hydrophobic compound of the monofunctional platinum(II) compounds but did not exhibit the longest distance from platinum atom to the edge of the non-labile ligand. Consequently, spatial orientation and/or hydrophobicity caused by the presence of a third aromatic ring may be significant factors in differentiating phenanthriplatin from the rest of its family. Derivatization of phenanthriplatin could further elucidate the structural components of this third aromatic ring that is responsible for causing NPM1 relocalization. We also note that the fast kinetics of DNA binding exhibited by phenanthriplatin may play a role in why phenanthriplatin is unique in the class of monofunctional platinum(II) compounds [6].</p><p>While oxaliplatin and phenanthriplatin both contain extended ligand structures around platinum(II), we find that steric properties alone are insufficient to explain the shared ability of these compounds to cause nucleolar stress. It is possible that monofunctional and bifunctional platinum(II) compounds may induce NPM1 relocalization through differential binding effects or mechanisms.</p>

PubMed Author Manuscript