Datasets:

griffin

/

ChemSum

like 12

Efficient Acid-catalyzed 18F/19F Fluoride Exchange of BODIPY Dyes

Fluorine containing fluorochromes represent important validation agents for PET imaging agents as they can be easily rapidly validated in cells by fluorescence imaging. In particular, the 18F-labeled BODIPY-FL fluorophore has emerged as an important platform but little is known about alternative 18F-labeling strategies or labeling on red shifted fluorophores. Here we explore the acid-catalyzed 18F/19F exchange on a range of commercially available N-hydroxysuccinimidyl ester and maleimide BODIPY fluorophores. We show this method to be a simple and efficient 18F-labeling strategy for a diverse span of fluorescent compounds, including a BODIPY modified PARP-1 inhibitor, and amine- and thiol-reactive BODIPY fluorophores.

efficient_acid-catalyzed_18f/19f_fluoride_exchange_of_bodipy_dyes

<!>General<!>Isotopic Exchange Reaction<!>18F-Benzyl-BFL (18F-Bn-BFL, 18F-7)<!>18F-Cysteine-BFL (18F-Cys-BFL, 18F-8)

<p>Fluorescence technology has revolutionized biological imaging through the development of biocompatible fluorescent small molecule, nanoparticles and proteins[1]. Despite these advances, translation into the clinic has proven more difficult and clinical trials remain scarce. Conversely, several thousand positron emission tomography (PET) imaging agents have been imaged in vivo in animals, but most of them have not been validated at the cellular level given the inherent spatial resolution limitations of PET.[2] Chemically and biologically equivalent bimodal, PET/fluorescence imaging agents circumvent this problem offering a number of advantages such as rapid screening, direct cellular validation (via microscopy or flow cytometry), and cost-effective testing of the stable isotope compound prior to rapid precursor scale-up and labeling of the optimized radiotracer.</p><p>Our and other groups have recently developed methods for labeling boron dipyrromethene fluorochromes (BODIPY) with the radionuclide fluorine-18.[3–5] Chemical removal of 19F yields a stable intermediate, which when treated with [18F]fluoride ion gives the chemically equivalent but is otopically distinct BODIPY, Scheme 1.</p><p>We reported the 18F-labeling of BODIPY dyes by a modification of the procedure of Hudnall[6] and during the course of our studies found acidic conditions were required for the labeling of the reactive intermediate[3]. Here, we describe and explore the relative reactivity of the acid-catalyzed exchange of 19F with 18F of a number of commercially available N-hydroxysuccinimidyl ester and maleimide BODIPY fluorophores, Scheme 2. Additionally, we show that the integrity of the N-hydroxysuccinimidyl (NHS) ester and maleimide (Mal) functional groups is conserved by demonstrating subsequent reaction with benzylamine and L-cysteine, respectively. Finally, using this method, we also describe the direct 18F-labeling small molecule, which is based on the PARP-1 inhibitor AZD2281 (Olaparib) modified with BODIPY-FL.</p><p>Following our previously published procedure[3], 4,4-difluoro-1,3,5,7-tetramethyl-4-bora-3a,4a-diaza-s-indacene-8-propionic succinimidyl ester (B493-NHS, 1) was electrophilically activated with trimethylsilyl triflate then treated with 2,6-lutidine to give a stable but reactive inter mediate. Addition of azeotropically dried 18F to the stable intermediate in the presence of acid gave the desired 18F-labeled BODIPY 18F-1 in high radiochemical yield in less than 2 minutes. We generat ed trifluorosulfonic acid (TfOH) in situ by the addition of Tf2O with tBuOH to the reaction mixture. These substances in the presence of residual water from the 18F were sufficient to generate the acidic conditions required for 18F fluorination of the stable intermediate. When these same conditions were applied to the labeling of BFL-NHS (2), we found the rate of the reaction significantly reduced. Upon further investigation, we found that structurally different BODIPY dyes (BODIPYs 1–6, Figure 1) display significant differences in their relative 18F/19F exchange rates.</p><p>Preliminary exchange experiments were conducted by treating NHS ester 2 with a mixture of Tf2O, tBuOH (40 mM final reaction concentration) and azeotropically dried 18F at 50 °C. Within minutes, radio-HPLC analysis showed 18F incorporation into 2. When 2 was treated with Tf2O or tBuOH separately only 1% exchange was observed after 2 h. With these observations we proceeded to study the exchange of 18F for 19F in the general BODIPY structure in more detail by varying the reagent concentrations, time, and temperature as well as the starting BODIPY fluorophore. The reaction rate was found to be dependent on both Tf2O/tBuOH and BODIPY.</p><p>Figure 2A and Table 1 show the dependence of the rate on the concentration of Tf2O/tBuOH concentrations (10, 14, and 40 mM) while all ot her concentrations were held constant. Linear regression of the data from these kinetic experiments provided a second order rate constant with respect to the acid concentration of (6.86 ± 1.01) × 10−5 M−1 s−1. Interestingly, we found that there was no incorporation of 18F at Tf2O/tBuOH concentrations that were below the concentration of the alkaline tetrabutylammonium bicarbonate (TBAB) phase transfer catalyst. We hypothesize that below this threshold, the reaction mixture is too basic for exchange to proceed. In addition to the second order dependence on Tf2O/tBuOH, we determined that there is a second order dependence on the concentration of the BODIPY substrate, as shown in Figure 2B. The BFL-NHS (2) concentrations were varied (0.5, 1.0 and 2.0 mM final reaction concentrations) while all other reagent concentrations were held constant. Linear regression of these observed rate constants provided a second order rate constant, (4.19 ± 0.15) × 10−4 M−1 s−1.</p><p>We also tested TfOH and methanesulfonic acid (MsOH) at the same final reaction concentration as the Tf2O/tBuOH mixture (40 mM). TfOH showed rapid 18F/19F exchange but also rapid decomposition of the fluorophore and thus the release of free 19F− into the reaction mixture and decreasing the incorporation of 18F. The MsOH experiment shows a significantly slower rate of 18F/19F exchange (Supplementary Information Figure S2A and S2B).</p><p>While the above kinetic experiments were conducted at 50 °C, two other temperatures, 0 and 23 °C, were also explored (Supplementary Information Figure S2C). The observed reaction rate at 50 °C (7.95 ± 1.01) × 10−4 s−1 was 11.7 times faster than that observed at 23 °C, (5.91 ± 0.89) × 10−5 s−1, and 33.1 faster than observed at 0 °C, (2.08 ± 0.38) × 10−5 s−1. Increasing the amount starting activity (2.0, 16.5 and 34.5 mCi) added to the reaction while maintaining a constant starting BFL-NHS (XYZ μmol) concentration resulted in a linear increase of specific activity of 6.5, 45.4 and 73.6 mCi/μmol, respectively.</p><p>In a series of separate experiments, individual exchange reactions on 2 were set up for each time point (15, 30, 60, 90, 120 min) in an effort to compare loss of activity due to evaporation, presumably in the form of [18F]-HF, while opening and closing the reaction tube for serial analysis of a single reaction. At each time point, the reaction activity was measured before and after removal of the aliquot for analysis. These data were decay corrected to the time of initial addition of activity and compared. Less than 5% loss of radioactivity to evaporation was observed during the course of running a single reaction with multiple analyses and less then 3% loss per reaction when running multiple reactions in parallel with single analysis per reaction.</p><p>Recognizing the potential utility of this method for the generation of dual PET-optical molecular imaging probes, we sought to broaden the scope of this labeling method therefore we tested these conditions (125 nmol BODIPY and 2.5 μmol Tf2O/tBuOH, 2 and 40 mM final reaction concentrations, respectively, at 50 °C) against a number of other commercially available BODIPY fluorochromes (B493-NHS (1), B530-NHS (3), BTMR-X-NHS (4), B630-X-NHS (5) and BFL-Mal (6)) as well as a biologically relevant BODIPY labeled small molecule PARPi (9)[7]. The structures of these compounds are shown in Figure 1 and the results of the exchange experiments are summarized in Figure 3 and Table 2. The previously studied B493-NHS (1) was the most reactive, showing over 87% 18F exchange within 15 min. Maleimide 6 and NHS esters 2 and 4 were found to be similar in reactivity, 60, 64, and 65% 18F exchange at 30 min. NHS esters 3 and 5 were considerably slower with only 16 and 24% exchange after 30 min, respectively. For the specific activity, each reaction has therefore is theoretical maximum, governed by the 18F activity incorporated and the amount intact radiolabeled plus cold BODIPY dye after the reaction. The specific activities range from 1.9 to 12.8 mCi/μmol (Table 2) correlating with the radiochemical yields, which range from 16.2 to 90.6% after 30 min reaction time.</p><p>Following the labeling of 18F-2 and 18F-6 after 45 min at 50 °C, these two compounds were purified from unreacted [18F]fluoride ion by passage through a silica gel cartridge. Once loaded onto the cartridge, elution with DCM and EtOAc secured the radiochemically pure 18F-2 and 18F-6 in 75 and 66 % radiochemical yield (non decay-corrected), repectively. Addition of benzylamine to 18F-2 in the presence of triethylamine (Et3N) and L-cysteine to 18F-6, provided in 18F-7 and 18F-8, Figure 4A. Presence of the desired stable isotope conjugates was confirmed by detection of the correct mass during LC/MS analysis and the presence of radio-labeled conjugates was verified by HPLC co-elution with the stable isotope 7 and 8.</p><p>A BODIPY conjugate previously prepared in our laboratory, PARPi (9) was subjected to acid-catalyzed 18F/19F exchange (conditions: 125 nmol PARPi, 2.5 μmol Tf2O/tBuOH, 40 °C) with 49% 18F incorporation observed after 30 min, Figure 4B.</p><p>The acid-catalyzed exchange of 19F for 18F on BODIPY fluorophores was found to have second order kinetics with respect to the acid as well as the dye added. At 50 °C, the incorporation of 18F into BFL-NHS (2) achieved over 50% incorporation at 15 min and 64% at 30 min. The incorporation proceeded to equilibrium over the next 1.5 h, reaching a maximum of approximately 85% at 2 h total reaction time. An increase in the BODIPY concentration would not only increase the final concentration of the 18F-labeled species but also the rate at which this equilibrium would be achieved.</p><p>Exchange reactions on other commercial BODIPY fluorochromes (B493-NHS (1), B530-NHS (3), BTMR-X NHS (4), B630-X-NHS (5) and BFL-Mal (6)) all demonstrated 18F− incorporation, although differences in reactivity were observed. Our results indicate that for compound 2, increasing the concentration of Tf2O/tBuOH increases the 18F/19F exchange rate. We anticipate that the other compounds tested will follow this trend to approach the exchange rate of 1. Both amine-reactive N-hydroxysuccinimidyl esters and thiol-reactive maleimides are well tolerated under these reaction conditions and demonstrated reactivity in subsequent conjugation reactions. To our knowledge this is the first example of a direct 18F-labeling of a maleimide. Other reported 18F-labeling protocols require multiple steps to obtain the desired thiol-reactive maleimide prosthetic group.[8–11]</p><p>As a general labeling strategy this method does have a limitation in specific activity due to the degeneracy of the starting material and product and therefore an inability to chromatographically separate the desired labeled product from starting material. To address this, we have shown that increasing the amount of starting activity will increase the specific activity and that the relationship was found to be linear over the activity range studied, 2 – 35 mCi (74–1295 MBq), shown in Figure 5. Although the resulting activities of 70 mCi/μmol (2590 MBq/μmol) are lower than other commonly used radiotracers FES (2.5–5 Ci/μmol[12, 13]), the specific activity generated is likely sufficient even for clinical applications. For example, for a 18F-PARPi scan, only 2 μmol would have to be injected (35 mCi, 70 mCi/μmol), which represents less than 0.1% of a daily clinical dose of the parent compound, Olaparib (based on a 300 mg (390 μmol) bi-daily oral administration, ClinicalTrial.gov identifier: NCT01844986). We envision by varying the starting quantity of 18F activity and/or starting material, the radiochemical yield and specific activity may be optimized for a particular application.</p><!><p>Unless otherwise noted, solvents and reagents were purchased from Sigma-Aldrich (St. Louis, MO) and used without further purification. All NHS and maleimide BODIPY compounds were purchased from Invitrogen. Synthesis of 4-[[4-fluoro-3-(4-(5-oxopentanamide)piperazine-1-carbonyl)phenyl]methyl]-2H-phthalazin-1-one 9 was described earlier[14]. [18F]Fluoride ion (n.c.a.) in 18O-enriched water was purchased from PETNET (Woburn, MA) and dried by azeotropic distillation of water with acetonitrile (MeCN) in the presence of tetrabutylammonium bicarbonate (TBAB; ABX) using a Synthra RN Plus automated synthesizer (Synthra GmbH, Hamburg, Germany) operated by SynthraView software. The dried 18F/TBAB was reconstituted in MeCN, collected in a 2-mL vial, and diluted to achieve a 12 mM TBAB solution. For non-radioactive compounds, LC-ESI-MS analysis and HPLC-purifications were performed on a Waters (Milford, MA) LC-MS system. For LC-ESI-MS analyses, a Waters XTerra® C18 (4.6 × 50 mm, 5 μm) column was used (Method A: eluents 0.1% formic acid (v/v) in water (A) and MeCN (B); gradient: 0–1.5 min, 5–100% B; 1.5–2.0 min 100% B; 5 mL/min). Preparative high performance liquid chromatography (HPLC) runs for synthetic intermediates utilized an Atlantis® Prep T3 OBD™ (19 × 50 mm, 5 μm) column (Method B: eluents 0.1% TFA (v/v) in water (A) and MeCN (B); gradient: 0–1.5 min, 5–100% B; 1.5–2.0 min 100% B; 30 mL/min). Analytical HPLC of radiolabeled compounds was performed employing an Agilent 1200 Series HPLC and a Poroshell 120 EC-C18 (4.6 × 50 mm, 2.7 μm) reversed-phase column (Method C: eluents 0.1% TFA (v/v) in water (A) and MeCN (B); gradient: 0–0.3 min, 5% B; 0.3–7.5 min, 5–100% B; 7.5–10 min, 100% B; 2.5 mL/min) with a multichannel-wavelength UV-vis detector, fluorescence detector and a flow-through gamma detector connected in series. Solid-phase extraction cartridges used were Oasis C18 3-cc cartridge (60mg, 30 μm particle size, Waters, MA) and Sep-Pak Silica 3-cc Vac cartridge (500 mg, 55–105 μm particle size, (Waters, MA)). The two-step, one-pot 18F− labeling procedure employing TMS-OTf was previous described.[3]</p><!><p>To azeotropically dried 18F/TBAB (30 μL, 12 mM TBAB in MeCN) in a 1.5-mL centrifuge tube, triflic anhydride (Tf2O; 10 μL, 250 mM in MeCN), t-butanol (tBuOH; 10 μL, 250 mM in MeCN) and the BODIPY dye (12.5 μL, 10 mM in 1.5:1 DCM:MeCN, B493-NHS (1), BFL-NHS (2), B530-NHS (3), BTMR-X-NHS (4), B630-X-NHS (5), BFL-Mal (6), or PARPi (9)) were added sequentially in 2 minute inter vals. The radioactivity of the reaction tube was measured in a well counter then placed in 50 °C shaker. At 15, 30, 60, 90, 120, and 150 min (with exception of 1, only 15 and 30 min data points were obtained), the tubes were removed from heat, cooled in an ice bat h for 20 s, and the activity of the reaction tube measured. An aliquot (1–3 μL) was removed from the reaction tube, radioactivity measured in a well counter and analyzed by HPLC (Method C). Radioactivity of the reaction tube was measured and then returned to the heated shaker. Kinetic data points were obtained from the area of the HPLC radio-chromatograms and plotted as a percent fraction. Observed rate constants were generated from the data by the program KINETIC of Dr. R. Fink, a gift from the late Prof. William von Eggars Doering, which handles kinetic schemes containing up to seven components, and incorporates a calculation of Marquardt that generates error limits in the rate constants at the 95% confidence level.[15, 16] Second-order rate constants were calculated using Graph-Pad Prism 4.0c (GraphPad Software, Inc, San Diego, CA). Exchange experiments were repeated for 2 in the same manner as described above varying starting concentrations of 2 (final reaction concentrations of 1.0 and 0.5 mM) or Tf2O/tBuOH (final reaction concentrations of 14, 10, and 7 mM) while maintaining a constant final reaction volume. Additional experiments were conducted for 2 at 0 and 23 °C.</p><!><p>The crude 18F-BFL-NHS mixture, prepared as described above in the acid-catalyzed exchange reaction after 45 min at 50 °C, was loaded on to a SPE silica gel cartridge (1.0 mCi) conditioned with pentane (500 μL). The cartridge was washed with pentane (2×150 μL) then the 18F-BFL-NHS eluted with dichloromethane (DCM, 3×150 μL) followed by DMSO (3×150 μL). Activity collected in the elution fractions was as follows: 1.7 and 2.7 μCi for pentane fractions 1 and 2; 179.5, 273.0 and 14.3 μCi for DCM fractions 1, 2 and 3; 48.6, 192.1, and 14.3 μCi for DMSO fractions 1, 2 and 3; and 229 μCi remaining on the silica cartridge. To the combined DCM fractions of 18F-BFL-NHS was added Et3N (10 μL, 250 mM in DCM) and benzylamine (37.5 μL, 25 mM in DCM) and stirred at rt for 40 min. HPLC analysis (10 μL aliquot) demonstrated full conversion of 18F-BFL-NHS to 18F-Bn-BFL. Bn-BFL, 7, LC-ESI-MS analysis found: 404.28 [M+Na+]+, 362.18 [M−F−]+; calculated: [M+Na+]+ = 404.17, [M−F−]+ = 362.18.</p><!><p>The crude 18F-BFL-Mal mixture, prepared as described above in the acid-catalyzed exchange reaction after 45 min at 50 °C, was loaded on to a SPE silica gel cartridge (1.0 mCi) conditioned with pentane (300 μL). The cartridge was washed with pentane (2×150 μL) then the labeled compounds eluted with dichloromethane (EtOAc, 3×150 μL) and DMSO (3×150 μL). Activity collected in the elution fractions was as follows: 2.6 and 2.5 μCi for pentane fractions 1 and 2; 198.9, 291.9 and 95.9 μCi for EtOAc fractions 1, 2 and 3; 18.8, 30.5, and 5.6 μCi for DMSO fractions 1, 2 and 3; and 317 μCi remaining on the silica cartridge. To the combined EtOAc fractions of 18F-BFL-Mal was added Et3N (10 μL, 250 mM in DCM) and L-cysteine (37.5 μL, 25 mM in DCM) and stirred at rt for 40 min. HPLC analysis (10 μL aliquot) demonstrated full conversion of 18F-BFL-Mal to 18F-Cys-BFL. Cys-BFL, 8, LC-ESI-MS analysis found: 516.31 [M−F−]+, 534.25 [M−H+]− calculated: [M−F−]+ = 516.19, [M−H+]+ = 534.18.</p>

PubMed Author Manuscript

Applications of Quantum Mechanical/Molecular Mechanical methods to the chemical insertion step of DNA and RNA polymerization

We review theoretical attempts to model the chemical insertion reactions of nucleoside triphosphates catalyzed by the nucleic acid polymerases using combined quantum mechanical/molecular mechanical methodology. Due to an existing excellent database of high resolution x-ray crystal structures, the DNA polymerase \xce\xb2 system serves as a useful template for discussion and comparison. The convergence of structures of high quality complexes and continued developments of theoretical techniques suggest a bright future for understanding the global features of nucleic acid polymerization.

applications_of_quantum_mechanical/molecular_mechanical_methods_to_the_chemical_insertion_step_of_dn

I. Introduction<!>II. Methods for describing reactive pathways<!>III. DNA Polymerase \xce\xb2<!>Abashkin et al., (2001): A cluster quantum mechanical calculation at steps along a proposed reaction path<!>Rittenhouse et al. (2003): A cluster quantum mechanical calculation on the pre-chemistry complex of Pol \xce\xb2<!>Radhakrishnan and Schlick (2006): QM/MM study of Pol \xce\xb2 (based on pdb=1BPY Sawaya et al. (1997))<!>Lin et al. (2006): QM/MM study of Pol \xce\xb2 (based on pdb=2FMS Batra et al., (2006))<!>Lin et al., (2008): Pol \xce\xb2 incorrect insertion QM/MM study (based on pdb=2C2K: Batra et al., 2008)<!>Batra et al., (2013): A QM/MM study based on the x-ray crystal structure of the Pol \xce\xb2 variant D256E<!>The broader picture: Pol \xce\xb2<!>IV. Application of QM/MM to systems similar to Pol \xce\xb2<!>a) Pol \xce\xbb insertion<!>b) Dpo4 NTP insertion<!>c) Pol \xce\xba insertion<!>V. RNA Polymerase<!>VI. Thoughts for future QM/MM simulations on ternary substrate complexes of nucleic acid polymerases<!>VII. Conclusions

<p>The evolution of modern x-ray crystallography has led to a rapidly increasing wealth of information about the three dimensional structures of both DNA and RNA polymerases (Wu et al. 2014). The resulting high resolution structures, capture intermediates that span the reactants to products path, have created a fertile ground for computational theoreticians to develop, test and apply methods that can expose the finer details of the bonds that form and break. In this article we focus on several nucleic-acid polymerization systems for which sufficient structures exist to reasonably explore the energy requirements of possible pathways.</p><p>An overview of the current state of the rapidly changing knowledge of DNA repair enzymes can be found in the more recent review by Sobol (2014) and the more structural review by Wu et al. - (2014). The nucleic acid (DNA and RNA) polymerase chemistry and structures have been systematically discussed in the recent volume edited by Murakami and Trakselis (2014). The basic premise is that good structures of pertinent complexes may serve as starting points for theoretical studies along a reaction path between reactant (pre-chemistry) and product.</p><p>The 2013 Nobel prize for Chemistry recognized the pioneering theoretical developments by M. Karplus, M. Levitt and A. Warshel, the central contribution being the quantum mechanical/molecular mechanical (QM/MM) approach to break a large biological molecular system up into a quantum mechanical core where bonds can break and form and a molecular mechanical system where covalent bonds are not formed or broken. In the latter case, classical force fields can be used to describe the motion and long range electrostatics (Warshel and Levitt, 1976 and Bash et al., 1991). Recent reviews that focus on the development of the method are available (Lin and Truhlar (2007), Senn and Thiel (2009), Lodola and De Vivo (2012), Groenhof (2013) and Konig et al., (2014)).</p><!><p>There are many variations of the QM/MM methodology. Most have to do with the manner in which the boundary between the quantum mechanical and classical mechanical regions is treated. (There is the story of the party of mathematicians that you happen to be attending. One of the group is expounding on a particular idea. The best way to become included in the group, the story goes, is to ask, at a particular point of emphasis in the exposition: "But, what happens at the boundary?"). How to treat the boundary and what happens at the boundary are the essential questions that we face. Let us consider an example of a popular version of QM/MM—the ONIOM method of the Morokuma group (Vreven et al., 2006)). The ONIOM method considers 4 energies—E(QM, all), E(MM, all), E(QM, Q) and E(MM, Q): the QM and MM energies of the complete system and the QM and MM energies of the specified quantum region, Q, respectively. Then the boundary is chosen (in principle) in such a way that the differences between E(QM, all) and E(MM, all) and between E(QM, Q) and E(MM, Q) are equal or nearly so. Then, we would have</p><p> (1)E(QM,all)=E(MM,all)+[E(QM,Q)-E(MM,Q)] that can be iterated to consistency.</p><p>Clearly, if we had sufficient computer resources, we would compute E(QM, all) directly by solving the time dependent Schrodinger equation directly. Lacking that overall ability at the present time for large enzymatic systems, we can, however, if the quantum region size (Q) is modest, determine estimates for the three quantities on the right hand side of equation (1). In the case of the ONIOM method, the boundary is chosen in practice by identifying single bonds in the side chains of residues that project into the proposed catalytic region but that are not expected to take part directly in the bond forming/breaking of the reaction. These bonds are then terminated with hydrogen atoms which are included in the calculation of E(QM, Q) and E(MM, Q). Since E(QM, Q) and E(MM, Q) include the electrostatic fields from atoms outside the Q region (defined as electronic embedding - in practice it becomes necessary to reduce the charges on atoms within several bonds of the Q region). The ONIOM method is flexible in that the critical catalytic region can be treated at a high quantum level, while the immediate surrounding region can also be treated at a lower level quantum region thereby allowing for polarization. A recent application of the ONIOM (QM:MM) method by Ding et al., (2013) to a photochemically-induced decarboxylation reaction of a green fluorescent protein illustrates the power and flexibility of the method. Developments leading up to this application are considered by Chung et al. (2012).</p><p>The major computer modeling packages- (AMBER (Wang, et al., 2004), CHARMM(Brooks, et al., 2009), GROMOS(Scott, et al., 1999), and NAMD(Phillips, et al., 2005)) - all have QM/MM software with varying degrees of flexibility. Individual laboratories have also contributed novel approaches, for example, the pseudobond approach (Zhang et al., 1999), the QTCP approach (Rod and Ryde, 2005) and a high-dimensional string QM/MM free energy method combined with an enhanced-sampling technique (Rosta, Yang & Hummer, 2014).</p><p>A particularly insightful and recent review of the application of QM/MM methods to enzymes has been provided by van der Kamp and Mulholland (2013) and a measure of how good certain variants of the method perform vs a full QM calculation can be found in a review by Hu, Soderhjelm and Ryde (2011).</p><!><p>It is probably true that more crystallographic data has been collected on DNA polymerase (Pol) β than any other polymerase (Beard and Wilson, 2014). This 39 kD, DNA gap-filling enzyme has occupied much of the chemical, structural and theoretical effort of our laboratories in recent years. The definitive structure for defining the active site for NTP insertion was published in 2006 (Batra et al., 2006) and was made possible by the use of a nonreactive NTP to impede the reaction. Previous approaches relied on using a dideoxy-terminated DNA substrate that lacked an atom that participated in catalysis; i.e., O3′ (Pelletier et al., 1994). The comparative analysis of structures and kinetics of Asp256 (wild-type) along with D256E and D256A variants (Batra, et al., 2013) has established that Asp256 is the catalytic base for triggering the insertion reaction – Specifically, the transfer of a proton from the sugar O3′ to the catalytic base initiates the reaction. As described below, this system universally utilized nucleic acid polymerases.</p><!><p>A brave and early attempt at an atomic level understanding the chemistry of the insertion step of Pol β was by Abashkin, Erickson and Burt (2001). This study, while not QM/MM, motivated our later work in that it focused closely on what was known from structural experiments at the time to define a realistic initial state for theory. In this case, the reference crystallographic structure (2.9 Å resolution) was determined by Sawaya et al., (1997). The theoretical study that followed was a set of DFT(double zeta)/QM calculations on an electrostatically neutral cluster of 67 atoms that defined various steps along two postulated reaction paths - (Fig. 1).</p><p>A path with an intermediate PO3 was found to have an unreasonably high reaction energy barrier and so an alternate scheme involving a penta-coordinated transition state was explored. In this alternate scheme, the reaction initiates by the proton on the O3′ (modeled, since the x-ray structure used the absence of this group to avert reaction) jumping to the free oxygen on the α phosphate from which it is further transferred, after some adjustments, to an oxygen atom of the β phosphate of the departing pyrophosphate product. The original modeled position on the O3′ proton may be an essential element in determining much of what happens downstream. The QM energies of seven structures, including the initial and final minimized complexes were determined. The product state was 19 kcal/mol more stable than the initial state. This is a reasonable result, but mostly based on the 1997 crystal structure. Several modifications to the crystallographic structure were made to arrive at a stable initial QM form which was electrostatically neutral: the necessary O3′H was added, a hydroxide ion was added to the catalytic magnesium ion, a proton was added to Asp190, which helps bridge the two active site magnesium ions, and this proton was placed so as to hydrogen bond with the Pα-oxygen that also bridges the magnesium ions. Finally, a hydrogen bond was defined between the side chain of Arg149 and a γ-phosphate oxygen. The catalytic magnesium may be therefore somewhat under-coordinated and the hydrogen added to Asp190 has no structural basis (in fact, it was added for ensuring calculated stability of the initial structure).</p><!><p>Although this study (Rittenhouse et al. 2003) did not map the reaction, it does highlight the details of the pre-chemistry active site. By building possible models of the active site based on the 1997 x-ray crystal structure of Pol β (Sawaya et al. 1997), an active site is proposed in which Pα-O, Asp-190 and Asp-192 bridge the two magnesium ions in a largely symmetrical fashion. In addition a water molecule was proposed to be tightly bound to the catalytic magnesium ion. Asp-256 is also proposed to form a tight bond with the catalytic magnesium ion. The three active sites aspartates are chosen to be −1 formal charge while the NTP was set up with a −4 charge. The primer DNA and substrate (NTP) bases were included along with water molecules bound to each magnesium ion. However, the charge balancing arginines and the proton on the metal bridging oxygen present in the Abashkin et al. (2006) study, were not present. Thus the overall charge of the cluster was −3. In their model, derived from quantum ab initio and DFT optimizations, the O3′ proton has an orientation that is poised for transfer to the unbound negatively-charged α-phosphate oxygen.</p><!><p>This study as well as work by Lin et al. (2006) that follows, appear to be the first works to apply the QM/MM methodology to the Pol β system. The initial structures for the reaction are taken from earlier studies by the authors (Radhakrishnan and Schlick, 2005, 2004) that had focused on subdomain motions associated with NTP binding. Both the correct NTP (Pol β/DNA/dCTP for G:dCTP) and incorrect insertion (Pol β/DNA/dATP for G:dATP) reactions were considered. While based on the 1BPY structure (Sawaya et al. (1997)), modifications were made to the active site so that three aspartates (190, 192, 256) are somewhat involved (precise geometries are not given, particularly for the units that bridge the two magnesium ions). The initial incorrect insertion structure was modeled from the initial correct insertion structure. The QM region including the link hydrogen atoms at the boundaries consists of 86 atoms: 7 atoms from each aspartate side chain (all assumed to bear −1 charge), 9 atoms from the O3′H region of the primer sugar, 17 atoms from the NTP (assumed to bear −4 charge and terminate with a (link H)-CH2-O-Pα unit), 7 atoms from the Ser 180 side chain, 19 atoms from the Arg 183 side chain (+1 charge), two magnesium ions (+2 charge) and 7 H2O molecules (each Mg ion apparently has a water molecule bound in the initial structures). The charges add up to −2 although the charge was indicated to be −1. When using the 1BPY structure as the starting structural model, O3′H must be added; apparently the location modeled for the proton on O3′H was such that the proton was not hydrogen-bonded to Asp 256. A later high resolution structure suggests this feature (pdb=2FMS; Batra et al. (2006)), where O3′H binds the catalytic magnesium (Mgcat). The QM/MM calculations employed an existing interface between GAMESS-UK (Schmidt et al. 1993) for the QM calculations (6-311G basis) and CHARMM (Brooks et al. 1983) for the MM calculations. The reactions were studied by generating new structures by constrained MD on modeled structures with modified O3′-Pα and O3′-Mgcat distances. These new structures were subjected to energy minimization and of these (50 total), 4 intermediates and a final product were found. These forms were additionally subjected to QM/MM dynamics to reach a total of six structure (reactant, four intermediates, and product) for both the correct and incorrect insertion reactions. For both reactions, the first intermediate is found to involve the -O3′H proton jump to a water molecule and, although the paths of the proton after that are somewhat different for the two reactions, it ends up in the product structure on the γ-phosphate of the NTP. The transition barriers for a presumed bipyramidal transition state are estimated to be greater than 18 kcal/mol for the correct insertion and greater than 21 kcal/mol for the incorrect insertion; values that are consistent with those estimated from experimental studies (Beard, et al., 2002; Ahn, et al., 1998). Due to the manner for generating intermediates, a reaction path construction was not possible. Shortly after this study, structural studies were published implying that the correct insertion initial system (pdb=2FMS, Batra et al. (2006)) has the –O3′-H proton strongly hydrogen-bonded to Asp 256 and the incorrect insertion initial system (Batra et al., (2007)) for G:dATP has O3′ displaced from Mgcat with a water molecule completing the catalytic metal hydration sphere. Both of these structural findings were accommodated in the next three studies discussed.</p><!><p>New structures of Pol β indicated that the catalytic metal site can be occupied by sodium or magnesium ions (Batra et al., 2006). Refinement of the latter magnesium structure was at 2.0 Å. The major change from the 1997 lower resolution structure was that the catalytic metal was now octahedrally coordinated, with clear density for all water ligands, Asp 256 and the O3′ oxygen of the primer terminus sugar.</p><p>After performing cluster calculations (DFT/B3LYP) on several models suggested by the 2FMS structure, and motivated by the earlier DFT/QM study of Abashkin et al., (2001) and by the QM study of Rittenhouse et al., (2003), Lin et al., (2006) were able to map a stable QM/MM reaction path (not shown) that did not involve a proton on the bridging Pα oxygen. Arginine residues 183 and 149 also were not included. Asp 256, which is bound to the catalytic metal, serves as the catalytic base for transfer of the O3′ proton. Separate QM/MM calculations to determine the most stable position of the O3′ proton for the cluster found that it was located between O3′ and a Asp 256 oxygen atom as the central part of a hydrogen bond. The initial geometry of the O3′H proton is consistent with the O3′-OD2-Asp 256 distance (2.81 Å) of the 2FMS structure. Both metals remain six coordinate throughout the reaction. To gain the effect of the entire protein, QM/MM calculations were performed using the ONIOM method (with electrostatic embedding) discussed in Methods (Lin, et al. 2006). The QM region consisted of three water molecules, the NTP with a proton on the γ-phosphate and a link atom after the O3′ sugar oxygen, three aspartates residues with link atoms between the β- and α-carbons, and the primer sugar with a link atom at the exit from the sugar ring (Fig. 3) for a total of 64 quantum atoms. In the crystallographic structure (Batra et al., 2006), there is an apparent hydrogen bond linking the O3′-H to an oxygen of Asp-256, and this oxygen, along with O3′, is also coordinated to the catalytic metal. This hydrogen bond telegraphs the transfer of the O3′ proton to Asp 256 which activates the nucleophile (O3′ anion) for attack on the α-phosphate. The cluster study of Abashkin, et al. (2001), based on a less detailed x-ray crystal structure, especially near the catalytic metal, had O3′ proton initially transferred to the α-phosphate.</p><p>The initial equilibrium structure (before insertion) was obtained by adding protons to the 2FMS structure (Batra et al., 2006). The non-hydrolyzable analog, −2′-deoxy-uridine-5′(α,β)-imido triphosphate, -located in the active site of the x-ray crystal structure was changed to dTTP to facilitate the reaction. Protonation states of amino acids were set at pH=7.0 via http://propka.chem.uiowa.edu). All crystal waters were preserved. Water and counter ions were added to provide a box that included 21,367 water molecules, 25 sodium ions and was electrically neutral. The SANDER module of the Amber 8 package (Case et al., 2005) with the Amber ff99 force field (Wang et al., 2004)) was employed for dynamics simulations and minimizations and the particle-mesh Ewald code (Darden et al., 1993; Essmann et al., 1995) was used for long range electrostatic interactions. The ONIOM module implemented in Gaussian 03 (Vreven et al., 2005) was the base QM/MM method with electronic embedding adopted for the study.</p><p>The strategy was to map the energy as a function of two variables: the forming O3′-Pα bond and the breaking Pα-O(Pβ) bond. Thus, using the scan keyword in Gaussian 03, a map of the energy versus these two variables was generated using the quantum method/basis set of B3LYP/6-31G*. In the early stages of the reaction, the proton on O3′ transfers to the Asp 256 carboxylate group with a low barrier of about 3.5 kcal/mole. Once this step occurs, the O3′-Pα distances closes until a transition state (with no stable intermediate) occurs. The geometry at the transition state is defined by R-O3′-Pα=2.2 Å and R-Pα-O(Pβ) =1.9Å. The transition state barrier above the initial reactant state is 21.5 and 18.0 kcal/mol above the deprotonated intermediate state. These barriers are consistent with experimental estimates of a free energy of activation of 16 kcal/mol (Vande Berg et al., 2001). At the last point along the reaction path, the product state is −5.2 kcal/mol below the initial reactant state and the reactant bond has now closed to R-O3′-Pα=1.65 Å and the broken bond has expanded to R-Pα-O(Pβ) to 3.27 Å.</p><p>An electrostatic energy decomposition study was undertaken at the transition state to ascertain which amino acids were stabilizing and which were destabilizing. The energy of interaction was taken to be</p><p>Then, the difference between the residue's position at the transition state and at the initial deprotonated state (i.e., O3′ deprotonated early in the path) is</p><p>The two residues that contributed the most to stabilizing the transition state were Arg-149 and Arg-183 (−4.6 kcal/mol and −7.1 kcal/mol, respectively). These stabilization energies are in accord with their vicinity to the pyrophosphate leaving group in the initial state (modeled from the x-ray crystal structure) and so, being charged themselves, they can modify the response of the breaking bond as the reaction proceeds.</p><p>A central observation of this QM/MM study was the remarkable stability of the geometry of the two magnesium ions, both of which interact relatively symmetrically with a O-Pα atom, which together form a scaffold upon which the reaction appears to evolve.</p><!><p>The appearance of a Pol β mismatch structure (G:dATP) (Batra et al., 2008) provided a new opportunity for understanding of misincorporation at the atomic level. Experimentally, the misincorportation step is experimentally estimated to be > 600 fold slower for G:dATP than G:dCTP (Ahn et al., 1998; Bakhtina et al., 2005), but it does occur and is measurable. Examination of the x-ray crystal structure pdb=2C2K of this mismatch (Batra et al., 2007) suggests why. The O3′H of the primer terminus is no longer coordinating the catalytic metal; it has been displaced by a water molecule. A direct path was tested to determine if O3′H would bond with Pα after the O3′ proton transferred to the Pα free oxygen. The energy barrier was found to be very high for this path (i.e., 48 kcal/mol) suggesting that another path must instead be viable. Another alternative was to instead propose a two step mechanism, in which the intrusive water molecule was synchronously moved from the magnesium coordination at the same time the O3′H group gained coordination. Constrained molecular dynamics was used to force this transfer and provide a prechemistry state that had O3′ strongly interacting with the catalytic metal. The resulting structure, found by equilibration molecular dynamics, was very similar to the Lin et al (2006) prechemistry structure. The barrier for this process was found to be about 14 kcal/mol from a B3LYP/6-311G** calculation on a cluster that included key coordinations for the two metals. The O3′-Mgcat distance served as the driving coordinate for the transformation. The idea was that the ground state structure, derived from careful equilibration of the x-ray crystal structure by molecular dynamics, and which had a high reaction barrier, underwent at conformational change (O3′H moves to displace the Mgcat coordinating water) that costs 14 kcal/mol, but which created a prechemistry structure from which the reaction path energetics could be determined from QM/MM procedures as in Lin et al., (2006). The energy barrier, using QM/MM procedures with the ONIOM method similar to Lin et al., (2006) for the misinsertion step is about the same as found for the correct insertion. The ratio of the insertion rates (correct/incorrect) is 12.5/.019, which can be converted to an energy difference of 3.8 kcal/mol, and then interpreted as the difference between the ground states of G:dCTP vs G:dATP. In this view, for correct insertion, the ground state is the prechemistry state, while for incorrect insertion, the ground state is separated thermodynamically from the prechemistry state by 3.8 kcal/mol and by a non-rate limiting barrier of 14 kcal/mol (water<->O3′ switch). The switch must occur before reaction can occur. Electrostatic effects of residues lying outside the quantum region on the transition state energies were found to be similar to those found for the correct insertion study.</p><!><p>The two previous studies (Lin et al., 2006, 2008) concluded that the catalytic base for the de-protonation of the –O3′H proton of the primer terminus was the oxygen atom of Asp 256 that is also bonded to the catalytic magnesium ion. In order to test this idea further, two variant structures (Fig. 4) were determined: D256E and D256A (PDB codes: 4JWM and 4JWN, respectively). The D256E structure (pdb=4KWM) is charge-conservative but significantly different than the D256 structure (pdb=2FMS) in two important ways: the carboxylate side chain is not bonded to the catalytic metal as in wild type (the coordination position on the catalytic metal is now occupied by a water molecule) and, Arg 254, which helps anchor the side chain of Arg 256 in place in the wild type, no longer interacts with the 256 position carboxylate. The –O3′H oxygen is, however, still in essentially the same place as for the wild type (pdb=2FMS) structure, approximately 3.5 Å from the Pα of the NTP. Also, for both the D256E and wild type structure, there exists an apparent hydrogen bond between –O3′H and a carboxylate (256 position) oxygen. The change is more drastic for the D256A system—the structure (pdb= 4JWN) does not have the bimetallic active site, but instead has only one magnesium ion which is located in the nucleotide binding position.</p><p>The catalytic metal site is empty (Fig. 4) and the -O3′H oxygen is displaced to 4.9 Å from the Pα of the NTP (as compared to about 3.5 Å in the 2FMS wild type structure. Kinetic measurements of kpol of the insertion rate indicate that the rate of insertion for the D256E system was reduced by three orders of magnitude as compared to the wild type while the D256A system had no measurable activity at pH 7.4. Increasing the reaction pH recovered significant activity (~10-fold/pH unit) indicating that the required deprotonation event had a pK above 10. QM/MM reaction paths, using a QM region were similar to those of the Lin et al. (2006) study. In addition, the basis set was now 6-31G*, an extra −CH2 group was included in the QM region on the aspartates and only the O3′–Pα distance was varied were mapped using the electronic embedding methodology of the GAUSSIAN 09/ONIOM interface (Case et al., 2010; Vereven et al. 2006) for both the wild type (a QM/MM study was included for internal reference) and the D256E variant systems. Because the alignment of O3′-Pα-Oβ is similar to that of the wild-type system preprocessing to obtain a pre-chemistry state was not necessary. To test the location along the reaction path of the -O3′H de-protonation, two paths were investigated: one where O3′ de-protonates early and one where it de-protonates late; i.e., near the transition state. The former case occurs for the wild type (as in the Linet al. (2006) study) and the latter case is found for the variant form. The transition state barriers are 14 and 21 kcal/mol for the wild type and variant cases, respectively, while the O3′-Pα distance at the transition state is the same for both cases. The higher energy barrier for the variant is consistent with the greatly lowered experimental insertion rate observed for the variant. A charge analysis using the Merz-Kollmann (Besler et al., 1990) option of GAUSSIAN 09 showed that a more effective nucleophile (O3′ anion) is developed at the transition state for the wild type as compared to the variant system. An electrostatic energy decomposition analysis comparison between the wild type and D256E variant did not reveal a clear reason as to why the variant transition barrier is higher; however, when the electronic embedding of the ONION procedure was turned off, thereby minimizing the transition state lowering effects of residues outside the quantum region, the barrier increase was much greater for the variant (increased from 21 to 58 kcal/mol) than for the wild-type enzyme (increased from 14 to 42 kcal/mol). Thus the location of the Glu256 side chain in the variant D256E is more destabilizing overall relative the location of the Asp256 side chain in the wild-type system, even though both systems are aligned well for the reaction and both have an apparent hydrogen bond for the -O3′H hydrogen with the carboxylate of the catalytic base. This particular paper is especially interesting because it combines structure determination, kinetic studies and theoretical estimation of a reaction path to investigate the nature of the chemistry step of the Pol β system.</p><!><p>Some DNA polymerases such as Pol β undergo subdomain conformational changes when the NTP substrate binds whereas others (e.g., Pol λ) do not (Wu et al., 2014). When the conformational change is present, the question arises as to whether these pre-chemistry events are kinetically and/or thermodynamically influence the overall mechanism. In addition, there are related questions as to what property (or properties) of the system control substrate discrimination (i.e., fidelity). Theoretical viewpoints differ somewhat, as seen by the attempt by Mulholland et al. (2012) to adjudicate a lively discussion between Schlick et al. (2012) and Prasad et al. (2012) about the relative contributions to the mechanism of pre-chemistry conformational changes to the observed barrier for catalysis. Further insight into the mechanism of DNA polymerases, if not resolution of conflicts, was provided by a commentary by Tsai (2014) and the experimental work (Olson et al., 2013) that showed that for the free energy difference between correct and incorrect insertion (averaged over two families and many substrates) was of the order of 5 kcal/mol. This result would imply that the moderate fidelity of the Pol β may simply be due to this large thermodynamic difference, although it does not tell us about the underlying molecular details. Finally, the relationship of Pol β's mechanism of action to accumulating structural data, kinetics and computation has been summarized recently (Beard and Wilson, 2014).</p><!><p>Five of the major DNA polymerase families (A, B, C, X and Y) were recently compared structurally for similarities (Wu et al. 2014) with a focus especially on the geometry of the active site of the ternary complexes (polymerase, double stranded DNA and nucleoside triphosphate) that lead to the chemistry of the formation of the O3′-Pα bond and the breaking of the Pα-O-Pβ bond to form pyro-phosphate. Essential elements that appear to span these five families in the active site are the existence of two negatively charged aspartate side-chains that bridge two divalent metal ions that are separated by 3.5–4.0 Å with nothing directly between these highly charged ions, the full NTP unit, the primer terminal -O3′H. Another essential element appears that for optimal function, including Watson-Crick insertion fidelity, the two divalent metals ions should be magnesium ions. Given this background, let us consider three recent QM/MM studies, one of which consider chemical insertion in another member of the X family (Pol λ) and two of which consider chemical insertion in members of the Y family.</p><!><p>For Pol λ (X-family DNA polymerase) insertion (Cisneros et al, 2008) is a QM/MM study based on a published x-ray crystal ternary substrate complex structure (pdb=2PFO; Garcia-Diaz et al., 2007) which had many features similar to the structure of Pol β (pdb=2FMS). The x-ray crystal structure (Fig. 5) consisted of the pre-catalytic complex of double strand DNA, Pol λ and a non-hydrolysable NTP (dUMPNPP) served as the starting template for a QM/MM study of the chemical insertion reaction. To prepare the system, the Mn(II) ion in the catalytic metal site was replaced by a magnesium ion and the dUMPNPP was replaced by dUTP, hydrogen atoms were added, system was solvated in a large box of water and equilibrated with a 2 ns PMEMD simulation (Case et al., 2005). All atoms within 30 Å of the catalytic metal were retained for the starting system. The QM/MM calculations were performed with the pseudo-bond method (Zhang, Lee, and Yang (1999); Zhang (2005)) which employed a modified version of Gaussian 03 (Frisch, 2004) with TINKER (Ponder, 1998) to compute energies along the reactions paths studies. Reaction coordinates for the proposed paths were chosen as described (see below) in Cisneros et al. (2008). A quantum mechanical subsystem was chosen that consisted of the NTP through the C5′ sugar atom, side chains of Asp490 (the equivalent of Asp256 in Pol β) and Asp427 and Asp429 (Asp190 and 192 in Pol β), the two magnesium ions, part of the primer sugar terminus (excluding the phosphate and C5′) and two metal bound water molecules for a total of 72 atoms (Fig. 5). The boundary atoms defining the pseudo-bond locations are the aspartate Cα's, C5′ of the primer dC and C4′ of the incoming NTP. After protonating the γ-phosphate, the NTP had a formal charge of −3 and the QM region a net charge of −2. The QM method was B3LYP (Becke, 1993; Lee, Yang, Parr, 1988) with a combined basis of 6-31G*for atoms involved in the paths proposed, the 3-21G for the non-reactive atoms and the LANL2DZ pseudo-potential (Wadt and Hay, 1985) was employed for Mn (paths in which the catalytic metal was either Mn2+ or Mg2+ were investigated; Mg was always the NTP binding metal). An extra diffuse function was included on Pα to accommodate Pα hybridization changes. The techniques used to produce the reaction path coordinates (Cisneros, 2008) required a product state structure; this was produced from the reactant state using modeling and QM/MM optimization. This non-experimental structure which anchors the product end of the reaction path thereby interjects a degree of uncertainty into the process. Once a given test reaction path is specified, a reaction coordinate can be defined. Unfortunately the equations defining the reaction coordinates for the reaction paths are not stated explicitly. In this study, the several paths (with magnesium in both sites) that were investigated were initiated by i) transfer of the O3′H proton to Asp 490, the analog of Asp 256 in the Pol β system, ii) the transfer of the O3′H proton to Asp 429, one of the metal bridging aspartates, iii) transfer of the O3′H proton to one of the water molecules bound to the metal ions and iv) transfer of the O3′H proton to the free oxygen on Pα. All of the paths, except the first, which transfers the O3′ proton to the non-bridging asp on the catalytic metal, have high energy transition states. The systematic generation of the coordinates along the reaction path (equilibrated experimental reactant structure) to (equilibrated generated in silico product structure) permitted the determination of the energy versus reaction coordinate profile for the systems with either magnesium or manganese ions at the catalytic metal site. These profiles for both metal ions gave approximately the same value for the activation energy (~17 kcal/mol) with the suggestion of a weakly bound intermediate between two transition states. It may be that the intermediate seen is an artifact of the constraint imposed by the method of defining the reaction coordinates. The distances between the metals change very little over the path (3.5 Å for Mg-Mg and 3.7 Å for Mn-Mg). Overall, despite the fact that different QM/MM methods were employed, the conclusions about the path of the reaction and magnitude of the activation energy determined, is similar to that found in the QM/MM study for the wild type Pol β system (Batra et al., 2013).</p><!><p>The Dpo4 (Y-family DNA polymerase) insertion reaction (Wang and Schlick, 2008) was a QM/MM study based on the x-ray crystal structure of the Dpo4/DNA complex of 8-oxoG:dCTP (pdb=2ASD: Rechkoblit et al., (2006)). The x-ray crystal structure has calcium ions in the active site and -O3′H is missing by design to stop insertion during structure determination. Thus this structure is somewhat distorted; for instance, the modeled -O3′-Mgcat distance is 5.3 Å (approx. 3 Å too long), the modeled-in O3′-Pα distance is 4.7 Å (approximately 1 Å too long) and the Me-Me distance is approximately 0.8 Å too long. The distorted x-ray crystal structure was then modeled to be similar to the high resolution Pol β structure (Batra et al., 2006; pdb=2FMS), although it is not clear from the paper's descriptions if the coordination state of the NTP-coordinating Mg ion is complete or the distance between the active site metals. The quantum part of the QM/MM was chosen to include the two bridging aspartates, a Mgcat coordinating glutamate, a small part of the primer terminus, two modeled magnesium ions, the incoming dCTP and four water molecules that coordinate Mgcat and dCTP. The total charge on the quantum system is apparently −3 as the dCTP is fully charged (−4). Once the initial pre-chemistry state was established, several possible paths were investigated. The QM/MM procedure was apparently chosen to be similar to the earlier study on Pol β by Radhakrishnan and Schlick (2006). The boundary between QM and MM is handled by a link hydrogen atom. The assumption was that the O3′ must initially be deprotonated before the O3′-Pα bond can form and the Pα-O-Pβ bond can break. Given this, the paths tested were to transfer the proton from O3′ to: i) one of the water molecules, ii) to the free oxygen on the Pα, iii) to an oxygen on Glu108, which is in the same position as Asp256 in the Pol β structure, or iv) to one of the bridging aspartates. The best energy path was concluded to be the transfer of the proton to water molecules with a subsequent transfer to the γ-phosphate. Three intermediates are found along the path to products (Pα-O-Pβ bond broken, O3′-Pα bond formed) with the highest energy intermediate one in which the proton is transferred through water molecules to the free oxygen on Pα. Although a transition state for this transfer is shown (at 20 kcal above the initial state and 5 kcal above the intermediate), a description of how this curve was determined, and subsequent transition states is not given. The overall conclusion is that the rate-limiting step (O3′ proton transferred to water to Pα oxygen) occurs about 20 kcal above the reactants.</p><!><p>Pol κ is also a Y-family polymerase that bypasses certain lesions such as benzopyrene. This study by Lior-Hoffmann et al. (2012) employs the pseudo-atom method (Zhang et al., 1999). A ternary substrate complex X ray crystal structure (pdb=2OH2; Lone et al., 2007) is employed. This structure, which is missing the -O3′H group and does not have a catalytic ion in place, requires significant modeling to obtain a structure suitable for defining an adequate pre-chemistry system. The modeled system involved establishing octahedral coordination at each magnesium ion. A backbone oxygen of Met108 occupies the sixth position of the NTP Mgcat and Glu 199 occupies the same position as is occupied by Asp256 in Pol β. A reaction coordinate driving procedure which proceeds by "stepping along a proposed reaction coordinate and performing energy minimizations with respect to the remaining coordinates" is employed (Zhang et al., 2000). In is not clear what the actual reaction coordinates used are in the paper, but generally they appear to partially describe the transfer of protons. Also, not all of the geometric parameters are defined. The choice as to which atoms to include in the QM part of the QM/MM scheme is somewhat different than for the Dpo4 system. For the Pol κ system, 81 atoms are included in the QM part: the NTP, two water molecules (one of which resides on Mgcat), the two Mg ions, the primer terminal sugar and base and the side chain of Glu199. The two metal ion bridging aspartates and the backbone atoms of Met 108 (which coordinates the Mgnuc) are relegated to the MM sub system. The NTP is taken to be fully charged (−4) so that the total system charge is −1. Several reaction paths for the de-protonation of the modeled O3′H were tested: i) employing Glu199 as a catalytic base, or ii) employing the free oxygen on Pα as the catalytic base. These potential paths were rejected when the protonated species appeared to be unstable for short interspersed QM/MM-MD simulations. The path (reaction paths tested are never precisely defined) that was found to be desirable was iii) a transfer of the O3′H proton to the γ-phosphate through the two waters included in the QM sub system leading to a stable intermediate, followed by the transfer of this proton to the β-phosphate as the O3′-Pα bond formed and the Pα-O-Pβ bond broke. Similar to the Dpo4 system, the initial proton transfer step was found to be rate limiting. The free energy of activation was found to be approximately 11 kcal/mol.</p><p>These last two QM/MM simulations discussed (Dpo4 and Pol κ), although using different ways of handling the boundary between QM/MM, share similarities. Both of the reference x-ray crystal structures are significantly distorted, are missing the O3′ unit, and do not have a Mg-Mg ion pair at the core of the active site until modeled. Addiitonally, neither paper gives geometry details about the modeled active site. For instance, it is impossible to deduce from these papers the distances from O3′ to the oxygen atoms of the glutamate that occupies the Asp256 position in Pol β in their pre-chemistry structures. If the Lin et al. (2006, 2008) and Batra et al. (2013) QM/MM papers are correct for the mechanism of Pol β and other similar systems, the reaction of the chemistry insertion is initiated by the transfer of the O3′H proton to the side chain of Asp256 that is nearest O3′ and together Asp256 and O3′ are bound to the catalytic magnesium. Both the Dpo4 and Pol κ papers consider the possibility of this mechanism: for Dpo4, the transfer of the O3′ proton to Glu108 and for Pol κ, the transfer of the O3′ proton to Glu199. In both cases, the cause for rejecting this path is stated to be that the transfer state is found unstable. In both cases, however, not enough detail (which carboxylate oxygen for transfer was tested, what was the distance from the O3′ to this oxygen in the modeled pre-chemistry structure, what was the precise geometric system tested for stability) is given to be assured that this path (O3′H to glutamate) was thoroughly vetted. This issue becomes important later.</p><!><p>The understanding of the mechanism of the synthesis of mRNA was greatly advanced by the publication of a group of structures of RNA Polymerase II by Wang et al. (2006) derived from S. cerevisiae. One of these structures, pdb=2E2H, was chosen by Carvalho, Fermandes and Ramos (2011) as the basis of a QM/QM study on the mechanism of extension of mRNA by Pol II. The structure (resolution 3.95 Å) does not have an O3′ on the ribose primer and the α subunit of the GTP substrates is not perfectly seated between the two Mg ions (as in the case of higher resolution structures of DNA polymerases). The Mg ions are 3.43 Å apart which is consistent with known DNA polymerase structures. While Asp 485 is firmly attached to the catalytic Mg ion and two aspartates serve as bridges between the two magnesium ions, the resolution is insufficient to observe a complete coordination shell about the magnesium ions. The ONIOM/Gaussian 03 (Vereven et al., 2006) method is used for optimizing trial structures generated to satisfy trial paths chosen for investigation. A QM/QM procedure in which there is a higher level QM (DFT B3LYP/6-31G(d)) applied to inner core atoms and a lower level QM (PM3MM, Stewart, 1989) applied to an outer shell of atoms. The ONION/Gaussian 03 programs allow both QM/MM and QM/QM procedures. Final energies of the optimized structures were then computed at the B3LPY/6-311++G(2p,2d) level. A neutrally charged model of the active site was extracted (Fig. 6) that consisted of parts of 4 aspartates, GTP, the primer terminal ribose, the two Mg ions, parts of a histidine, a lysine and three arginines for a total of 226 atoms. For the inner core, the two Mg ions, GTP, the ribose, histidine and last three side chain atoms of the four aspartates were chosen. There apparently was no water molecules included in the region of quantum calculation. After preparing the models by adding the missing –O3′H unit to the ribose and missing H atoms, and relaxing in neutralized water for 20 ns, four hypothetical reaction pathways were generated. These were: i) HYP1 (the O3′ proton jumps to an α phosphate oxygen on the GTP and ends up protonating the pyrophosphate after the bond forming (O3′-Pα) and bond breaking (Pα-Oβ)), ii) HYP2 (a hydroxide ion materializes near the -O3′H, deprotonates O3′ so that it can attack Pα, and the departing pyrophosphate product is stabilized by a proton transfer from a nearby histidine), iii) HYP3 (an OH- group is added on the catalytic magnesium ion initially, which deprotonates the O3-H, while the product pyrophosphate is stabilized by a proton from the histidine and iv) HYP4 (the non-bridging aspartate bound to Mgcat accepts the proton from –O3′H so that the –O3′ anion can attack Pα and the histidine protonates the leaving pyrophosphate). The energy cost of generating a hydroxide from bulk water for the HYP2 path was rationalized to be 7.5 kcal/mol by the use of free energy perturbation theory and concentration in bulk considerations. Only HYP2 (path 2) had a low activation energy (~10 kcal/mol), the other three paths had barriers of > 29 kcal/mol. In this path, a hydroxide is created near O3′H, which facilitates deprotonation of the O3′. Then, the positively charged His1085 loses its proton to one of the free oxygens on Pβ. The consequence of this is the weakening of the Pα-Oαβ bond as the Pα-O3′ bond forms. It is this last step that defines the limiting reaction step in the overall path. If it is the case, as we speculated earlier, higher resolution structures ultimately lead to an initial active site that has what we think are the essential features (O3′ in place, an Asp/Glu group at the Asp 256 Pol β position and a water molecule coordinating Mgcat, the Pα-O symmetrically (approximate isosceles triangle) bridging the two magnesium ions and there are two metal-bridging aspartates), this mechanism will require reinvestigation. Indeed, in the same paper that pdb=2E2H originates, one can also find pdb=2E2J. Also, in the latter structure, O3′ is present (a non-hydrolyzable NTP is used) and it is tightly bound to the Mgcat. In this structure, the aspartate coordinating to the Mgcat is within H-bonding distance (2.52 Å) of the naturally-present O3′. And a later structure (pdb=3S1Q) has appeared (Liu et al. 2011) from the same group, where the aspartate at Mgcat exists and superposition of Pα-O and the two magnesium ions with pdb=2FMS is almost perfect even though O3′ is missing in the former structure.</p><p>Very recently, the insertion pathway for RNA Pol II has been further investigated (Zhang, 2013) in the Salahub lab at the University of Calgary. Several starting systems were considered: i) model A, based on pdb=2E2H (Wang et al, (2006)), ii) model B, based of pdb=2E2J (Wang et al., (2006)), (Fig. 7) iii) models C1 and C2 where, for both of these models, the NTP(GMPCPP) of pdb= 2E2J was changed to the GTP of pdb=2E2H and optimized and the system equilibrated for either 1ns (Model C1) or 12 ns (Model C2). The sizes of the various structures were reduced by fixing all atom coordinates beyond 20 Å of the NTP Pα. The quantum subsystem included parts of sidechains of 4 aspartates (481, 483, 485, 837), 3 arginines (446, 766, 1020), three water molecules, the entire GTP or NTP substrate, the ribose of the primer terminus and the two magnesium ions for a total of 144 atoms and charge of −1 for the quantum region. An in-house QM/MM program was employed that utilized hydrogen link atoms at the boundary. The MM part of the calculation was performed with the CHARMM27 force field (McKerell, et al., 1998, Foloppe & Mackerell, 2000) and the QM part of the calculation was performed with the semi-empirical AM1/d-PhoT method (Cui, Gao & York, 2007). The four models were subjected to the defined QM/MM procedure with the apparent conclusion that model C is most appropriate and that direct transfer of the O3′H proton to the α-phosphate will be the lowest energy path to products. We note that a path with initial transfer of the O3′H proton to Asp 485 (the structural equivalent of Asp 256 for Pol β), the side chain carboxylate oxygen of which is 2.52 Å from O3′H in pdb=2E2J) was not investigated.</p><!><p>Has the experimental structural evidence that has accumulated to date reached the critical amount needed to be able to develop a more unified, consensus view of how the two metal site at the core of polymerase functions? Twenty years ago, this issue was visited with some controversy (Pelletier, 1994; Steitz, 2004) and left unresolved. The diverse mechanisms presented in the various QM/MM applications to polymerases that are presented in this manuscript reflect the lack of a unified view. However, we now believe that a solid case can be made for the "O3′H → initial proton transfer to an active site acidic residue" as the initial step in polymerase activity. For Pol β, this is Asp 256. The data for this suggestion is collected in Table 1.</p><p>This data appear consistent with the proposal that if the crystal structure is performed with an non-hydrolyzable NTP substrate, so that O3′ is present, and if the active site contains two magnesium ions, then the O3′H group appears to be within a distance characteristic of a strong hydrogen bond in its interaction with the Lewis base occupying the Asp 256 (Pol-β) position. We note that the data in Table 1 spans several nucleic acid families. These observations support the idea that the essential active site for DNA/RNA polymerase activity may involve two closely spaced magnesium ions, an NTP, two metal bridging aspartates, an O3′ that coordinates with the catalytic metal and with a Lewis base bound to the catalytic metal. In the higher resolution structures in Table 1, there is also a water molecule bound to the catalytic magnesium. If any of these features is missing and must be modeled, the modeling must be carefully done to restore all essential geometry for function. One reasonable way to interpret the data in Table 1 is with the assertion that the insertion chemistry is triggered by the proton transfer of the O3′ proton through its hydrogen bond to the catalytic base (an aspartate or glutamate). We propose that this jump may take place by quantum mechanical tunneling. Proton tunneling is now a well-established concept for enzyme reactions (Kuznetsov and Ulstrup, 1999; Truhlar et al., 2004; Truhlar, 2009, Hay and Scutton, 2012, Klinman and Kohen, 2013 and Kippenstein, Pande and Truhlar, 2014). Once this jump occurs," the cat is out of the proverbial bag". Instantaneously there is a reorganization of electronic charge involving at a minimum the proton, the catalytic metal, the Lewis base and O3′. And, it is likely that the conserved water at Mgcat is involved, perhaps by its induced ionization to lose a proton to a water network with almost simultaneous transfer of the nearby just-added Asp256 proton. It is important to appreciate that there are at least three separated time scales in action here: the widely separated time scales of electrons and heavy nuclei (C, O, N), and the intermediate time scale of protons. The time scale of charge reorganization should be on the order of femtoseconds since it is electronic in nature. It is possible that the proton on the Lewis base at the Asp256 position (Pol β), once transferred from O3′, is short lived (it may be transferred through a water network that includes the conserved water molecule on the catalytic metal ion to the region of the NTP near Oαβ). All of these reasonable events will happen with the heavy nuclei essentially static. Instantaneously, we have the Pα-Oαβ bond weakened and the O3′-Pα interaction strengthened. As a result, inversion at Pα begins. Bond forming/breaking happens, perhaps nearly spontaneously. In this view, there is little activation energy of these steps, presumably the rate controlling step that defines the turnover number is, in fact, the collective energy needed to snap all of the molecular parts in place (or perhaps to clear the active site once the reaction in over). Once in place - with the NTP symmetrically bridging the two metals at Pα, the O3′-Asp256 position hydrogen bond with both O3′ and the residue at the 256 position bound to the catalytic metal—the reaction is initiated by the tunneling event between O3′ and Asp256. (In the QM/MM simulations on Pol β (Lin et al 2006 and 2008; Batra et al, 2013), the proton was transferred to Asp256 in the slower process of classical barrier crossing and then retained on the Asp256 position Lewis base while the insertion reaction was forced to occur. Tunneling of the proton was not evaluated and a sufficient water network to bridge Mgcat and Oαβ was not in place.) The structural data on the DNA/RNA polymerases (Table 1) thus lead us to a view of polymerase reactivity that may only be fully revealed through the consideration of proton tunneling and multiple time scale events.</p><!><p>All QM/MM studies to date on polymerase reactions, either begin with models that do not correctly account for the necessary strong hydrogen bond at the O3′-Asp/Glu interaction, do not have enough water molecules present to allow for the proton to end up at Oαβ with little energy cost or have quantum sub-systems which are charged, rather than neutral. The ideal pre-calculation crystal structure would have this key hydrogen bond in place, which requires the use of a non-hydrolysable NTP. In this inferred ideal structure, the two magnesium ions would be about 3.5 Å apart, the NTP would be in place, binding the Pα subunit symmetrically to the two metals and two aspartates would bridge the two magnesium ions. The model extracted from this would require three neutralizing positively charged amino acids around the periphery. These would ideally be arginines, which are less resistant to donating protons to the NTP than lysine. The structure of DNA Pol β has three arginine groups ideally located for model building. Tunneling has generally not been considered in the DNA/RNA insertion chemistry reactions to date. Hopefully future work in this area will strive to be described in such a manner as to be reasonably reproducible.</p>

PubMed Author Manuscript

Quest for breathing: proliferation of alveolar type 1 cells

There is much evidence that the vertebrate lung originated from a progenitor structure which was present in bony fish. However, critical basic elements for the evolution of breathing in tetrapods, such as the central rhythm generator sensitive to CO2/pH and the pulmonary surfactant, were present in the lungless primitive vertebrate. This suggests that the evolution of air breathing in all vertebrates may have evolved through exaptations. It appears that the capability for proliferation of alveolar type 1 (AT1) cells is the “critical factor” which rendered possible the most radical subsequent innovation—the possibility of air breathing. “Epithelial remodeling,” which consists in proliferation of alveolar cells—the structural basis for gas diffusion—observed in the alimentary tract of the gut-breathing fishes (GBF) has great potential for application in biomedical research. Such a process probably led to the gradual evolutionary development of lungs in terrestrial vertebrates. Research on the cellular and molecular mechanisms controlling proliferation of squamous epithelial cells in the GBF should contribute to explaining the regeneration-associated phenomena that occur in mammal lungs, and especially to the understanding of signal pathways which govern the process.

quest_for_breathing:_proliferation_of_alveolar_type_1_cells

Introduction<!>Background<!>Research models<!>From water to land<!>Exaptation<!>Promising natural model<!><!>Promising natural model<!>Natural model organism<!>At the end<!>In conclusion<!>

<p>The organs used for air breathing in vertebrates are very diverse (Duncker 2014). In spite of the differences in their structure and ventilation mechanisms, however, the respiratory organs always have some typical modifications: simple squamous epithelium and the distribution of numerous capillary vessels among the epithelial cells. The two characteristics must co-occur. Such an adaptation causes a significant reduction in the thickness of the air–blood barrier, allowing for gas diffusion. The adaptation is also observed in air-breathing fishes which use their stomach as an accessory respiratory organ (Satora 1998; Podkowa and Goniakowska-Witalińska 2003; Cruz et al. 2009; Cruz and Fernandes 2016). The lungs are the main organ of the respiratory system in mammals. They possess a unique architecture: millions of alveoli. Each of them is lined mainly by thin squamous epithelial cells. The apical parts of epithelial cells are strongly attenuated and form an exceedingly thin, continuous layer covering capillary blood vessels located between the bodies of epithelial cells; such structures are primarily sites for gaseous exchange (Liem 1988; Ciechanowicz 2019). Although mammalian lungs are slow-turnover organs that are highly quiescent at steady state, they have the ability to repair epithelial damage (Liem 1988; Ciechanowicz 2019).</p><p>Lung diseases are among the most common medical conditions all over the world. Moreover, chronic obstructive pulmonary disease (COPD) and lower respiratory infections are associated with high morbidity and mortality. Currently, they are ranked by the World Health Organization (WHO) as the third and fourth leading cause of death worldwide, respectively (WHO 2020). Both chronic and acute respiratory diseases affect the interstitium, for example sarcoidosis, idiopathic pulmonary fibrosis, autoimmune diseases, pneumonia, pulmonary edema (Kaku et al. 2020; Meyer et al. 2021). Parenchymal diseases are characterized by progressive remodeling of lung parenchyma combined with destruction and fibrosis of alveoli and, consequently, progressive respiratory distress. With the developing inflammatory reaction, the pneumocytes desquamate, which is accompanied by production of hyaline membranes. The altered alveoli cease to fulfill their primary role in gas exchange (Sims et al. 2005).</p><!><p>In mammals, gas exchange takes place in lung alveoli, which ensures a large surface area for diffusion of oxygen and carbon dioxide. The respiratory epithelium, which is the main component of the alveolar wall, contains two main types of cells (Liem 1988; Desai et al. 2014; Ciechanowicz 2019; Parekh et al. 2020). Alveolar type 1 (AT1) cells, which maximize surface area while minimizing the gas–blood barrier, occupy almost 95% of the lung surface and form a thin, continuous lining in the alveolar wall. AT1 cells supply a short diffusion pathway for gas exchange—creating a gas–blood barrier of about 0.2–2.5 µm (Liem 1988; Miettinen et al. 1997; Desai et al. 2014). Alveolar type 2 (AT2) cells (cuboid cells) are almost twice as abundant as AT1 cells (Ciechanowicz 2019), but they occupy only 7–10% of the lung surface and express high levels of surfactant protein C (Ciechanowicz 2019; Parekh et al. 2020). The surfactant ensures low surface tension and contributes to the elastic properties of the lungs (Bensch et al. 1964; Clements et al. 1970; Pattle 1976; Sullivan et al. 1998; Hawgood et al. 1998). In mammals, AT1 cells have lost their capacity for proliferation by cell division and, when damaged, they are replaced by AT2 cells (Liem 1988; Desai et al. 2014; Zacharias et al. 2018; Ciechanowicz 2019; Parekh et al. 2020). In fully developed lungs, the microenvironment regulates proliferation and differentiation potential of populations of multipotent endogenous stem cells located in niches (Ciechanowicz 2019). Furthermore, it is proposed that AT1 cells are completely differentiated, since there is little evidence to indicate that they can divide, whereas AT2 cells are regarded as bifunctional alveolar progenitor lung stem cells, which can differentiate into AT1 cells (Desai et al. 2014; Zacharias et al. 2018; Parekh et al. 2020). These functions are regulated by the bone morphogenetic protein (BMP) signaling pathway. During this process, BMP4 prevents proliferation of AT2 cells and promotes differentiation; its antagonists, such as Noggin, promote proliferation (Parekh et al. 2020). Other factors and signaling pathways are implicated in the self-renewal of AT2 cells after distal lung injury. In this case, stromal cell-derived factor 1 (SDF1) activates yes-associated protein (YAP), which leads to the production of growth factors, such as epithelial growth factor (EGF), and paracrine signals released by macrophages. However, following injury, AT2 cells possess limited proliferative ability. Further subclassification of AT2 cells, and ascertaining their role in lung regeneration processes, is still necessary (Desai et al. 2014; Zacharias et al. 2018; Parekh et al. 2020). However, without recognizing the mechanisms which control human lung development, the precise identity and function of human lung stem and progenitor cell types, and the genetic and epigenetic control of human lung fate, progress toward the development of strategies for lung regeneration following injury is impossible (Desai et al. 2014; Ciechanowicz 2019; Parekh et al. 2020).</p><!><p>Studies on rodents—a group characterized by large disparities in the size, structure, cellular composition and physiology of their airways compared to humans—impose limitations on the use of this model as a preclinical animal model system (Parekh et al. 2020). On the other hand, differentiation of AT2–AT1 cells in 3D organoid culture in research on cell lineages is still the main challenge (Parekh et al. 2020). For example, freshly isolated cells of human alveoli quickly lose their differentiation status during culture, and this leads to failure to detect types of cells in vivo (Sims et al. 2005). Understanding the cellular and molecular mechanisms which control the development of the gas exchange surface and differentiation of the lungs is crucial for understanding the pathogenesis of acute and chronic lung diseases. This pertains especially to regeneration after exposure to damaging factors. Unfortunately, for obvious reasons, there is no direct physiological evidence, and the lung evolution can only be studied in extant species, followed by extrapolations (Randal et al. 1981). In this situation it seems crucial to find an adequate model to observe the initial stages of lung formation in terrestrial vertebrates (Satora et al. 2020b).</p><!><p>In the almost four billion years since life on earth emerged, evolution has generated a number of marvelous metamorphoses. One of the most spectacular changes is that which produced terrestrial creatures bearing limbs, fingers and toes from water-bound fish with fins. The replacement of fins with limbs was a crucial step in this transformation, but was by no means the only crucial step (Clack 2005). Land is a radically different medium from water, and to conquer this medium, tetrapods had to evolve novel ways to breathe and become equipped with a respiratory organ for air breathing. Most accounts of vertebrate evolution describe early air-breathing fishes, and stress the importance of aerial respiration in the origin of the tetrapods (Randal et al. 1981; Graham 1997). Nevertheless, the focus of these treatments usually shifts to the tetrapods themselves and the changes occurring in the phyletic progression from amphibians to mammals. Such accounts rarely consider the evolution of fishes beyond the Paleozoic and as a result succeed, more often than not, in conveying the impression that both fish evolution and the importance of air breathing to fishes ended with the appearance of amphibians. Similarly, comparative surveys of air-breathing fish respiratory adaptation have not considered the phyletic histories of fish, and thus most often treat both the primitive and modern air-breathing fishes similarly, as inferior grades of the mammalian specialization, evolutionary curiosities, or both (Randal et al. 1981; Graham 1997). Air breathing has persisted throughout the evolutionary history of the fishes and has played a fundamental role in the evolution of this group (Graham 1997; Icardo 2018). In general, air-breathing organs sequester a bubble of air out of contact with the water but in contact with a thin epithelium through which O2 diffuses into the blood. In contrast to water breathing, ventilation of the air-breathing organ is periodic (Kramer and Braun 1983).</p><!><p>Exaptation is defined as existing structures that now enhance fitness but were not produced by natural (or sexual) selection for their current role (Gould and Vrba 1982; Tattersall 2009). There is much evidence to indicate that the vertebrate lung originated from a progenitor structure present in bony fish (Randall et al. 1981; Graham 1997; Nelson and Dehn 2011; Hoffman et al. 2016), but crucial structures for the evolution of air breathing were present in the vertebrate ancestors (lungless) prior to the evolution of the lung (Sullivan et al. 1998; Hoffman et al. 2016). In 2016, Hoffman and co-authors proposed a completely novel hypothesis, namely that the evolution of air breathing in all vertebrates occurred through exaptations (Gould and Vrba 1982; Tattersall 2009; 2014) derived from critical basic elements (Hoffman et al. 2016). One of them is the central rhythm generator sensitive to CO2/pH present in lamprey—a lungless vertebrate (Hoffman et al. 2016).</p><p>Additionally, Sullivan and co-authors point out that the evolution of air breathing must have been preceded by evolution of the surfactant system—evolved initially in the gut and subsequently utilized and modified in the lung (Sullivan et al. 1998). Pulmonary surfactant (mixture of lipids and proteins) is present in all air-breathing vertebrates, synthesized in the endoplasmic reticulum of cuboid alveolar cells (Bensch et al. 1964; Clements et al. 1970; Pattle 1976; Haagsman and van Golde 1991) and stored in dense multilayered structures called lamellar bodies (Bensch et al. 1964; Chevalier and Collet 1972). The surfactant forms a thin, amorphous alveolar lining, spreading over all the cells in contact with air (Bensch et al. 1964; Clements et al. 1970; Pattle 1976; Haagsman and van Golde 1991; Satora 1998). The surfactant reduces surface tension at the air–liquid interface and protects the cells against drying and the toxic effects of oxygen (Clements et al. 1970; Pattle 1976; Smits et al. 1994; Sullivan et al. 1998; Satora 1998). The study of surfactant protein A (SP-A) in members of all the major vertebrate groups implies that the surfactant had a single evolutionary origin in the vertebrates (Sullivan et al. 1998).</p><p>However, the presence of AT1 epithelial cells is indispensable for gas diffusion. The appearance of such epithelium, combined with the capability for proliferation, in the alimentary tract is the next critical element for the evolution of breathing in Tetrapoda. In addition, the ability of these cells to proliferate seems to be a "critical factor" of the practical breakthrough in the evolution of lung. The environmental factor—hypoxia—has turned out to be the main driving force of such changes leading to the origin of the lung (Randal et al. 1981; Graham 1997; Nelson 2014). According to Tattersall, exaptations, combined with "critical factors," constitute a powerful evolutionary mechanism, and they are the driving force of development. On the other hand, all new genomic variants must arise as exaptations, mutations occur at random, and new functions cannot be adopted without prior new structures (Tattersall 2006; 2009; 2014).</p><p>Sometimes, a combination of pre-existing elements (exaptations) results in something totally unexpected (Tattersall 2006). In the case of limb development, Clack discovered in 2005 that many of the critical innovations arose while vertebrates were still largely aquatic (Clack 2005). Furthermore, Clack suggested that the first changes appeared to have been related not strictly to locomotion but to an increased dependence on breathing air (Clack 2005).</p><!><p>It is believed that regular occurrence of aquatic hypoxia (low oxygen conditions), being a primary factor, led to the evolution of air breathing in the Late Silurian fishes and as a result enabled vertebrates to invade land in the Devonian period (Randall et al. 1981; Graham 1997). Also, in modern air-breathing fishes, hypoxia is the greatest inducing force for air breathing (Graham 1997; Seymour et al. 2008; Nelson 2014). Among the air-breathing fishes, the gut-breathing fishes (GBF) seem especially interesting; they must have special adaptations to use their alimentary tract as an accessory respiratory organ during low oxygen levels in the water (Nelson and Dehn 2011; Nelson 2014).</p><!><p>a Transmission electron micrograph of the gas–blood barrier in the corpus of the stomach of Ancistrus multispinnis (Loricariidae). The gas–blood barrier is composed of three layers: external, thin cytoplasmic sheets of respiratory epithelial cells (remodeled gastric epithelial cells), narrow interstitial space, with basement membrane (BM), and thin parts of endothelial cells (EN). E erythrocyte with nucleus; GE gastric epithelium; GL gastric lumen; N nucleus of epithelial cell.</p><p>Taken from Satora (1999). Scale bar = 1 µm. b Transmission electron micrograph of a section of the stomach corpus epithelium of Ancistrus multispinnis. Flattened epithelial cells with nucleus (N) and lamellar bodies (arrows) are visible. BM basement membrane; E erythrocyte; GE gastric epithelium; GL gastric lumen. Taken from Satora (1999). Scale bar = 1 µm. The cell body with large nucleus (N) is situated between capillaries covered by thin epithelial sheets (insert in 1b). Taken from Satora (1999). Scale bar = 1 µm</p><p>a Transmission electron micrograph of the stomach of Ancistrus multispinnis (Loricariidae). Epithelial cell with nucleus (N) and lamellar bodies (arrows). GL gastric lumen.</p><p>Taken from Satora (1999). Scale bar = 1 µm. b Ultrastructure of neuroendocrine-like cell of the stomach corpus Ancistrus multispinnis. The cytoplasm contains characteristic secretory vesicles (dense core vesicles). Taken from Satora and Winnicki (2000). (TEM) Scale bar = 1 µm</p><p>Summary schematic of "epithelial remodeling" in the gastrointestinal tract in gut-breathing fishes under hypoxic conditions. Aquatic hypoxia causes degranulation of graininess within putative chemoreceptors (NEC), which triggers a "cascade of events." As a result, proliferation of oval epithelial cells situated between columnar enterocytes follows, combined with a change in the shape of the cells—gradual flattening and stretching. The capillaries (c) get closer in relation to the future respiratory surface. At the same time, the number of lamellar bodies contained within decreases. In the final stage of this process, the epithelial cells differentiate into the enlarged basal part (with nucleus) located between the capillary (c) and strongly flattened peripheral extensions. Numerous capillaries are covered only with flattened projections of epithelial cells, thereby creating a gas–blood barrier which enables gas diffusion</p><!><p>Lungs, as well as respiratory and non-respiratory bladders of chondrosteans, appear to have originated from a respiratory, posterior pharynx through proliferation of the squamous cells and gradual enlargement. The fish groups which have lungs, or a pulmonoid/respiratory swim bladder, tend to develop only the skin as an accessory aerial gas exchange organ, whereas those with non-secretory or secretory swim bladder also modify their gills, opercular or branchial cavities, pharynx, pneumatic duct, stomach or intestine (Perry et al. 2019). It is suggested that this mechanism has developed independently in several species of the GBF (Satora et al. 2020b). The epidermal growth factor receptors (EGFR) seem to be among the factors responsible for the adaptation of the gastrointestinal tract to the role of additional respiratory organ in the GBF (Satora et al. 2017; Mytych et al. 2018). In mammals, the EGFR plays an important role in lung maturation; EGFR deficiency results in a mild respiratory distress syndrome and delayed lung maturation (Miettinen et al. 1997). Other essential elements include secretory neuroepithelial-like cells (NECs), putative chemoreceptors (Zaccone et al. 2017, 2018, 2019, 2020, Capillo et al. 2021; Lauriano et al. 2021), which are probably responsible for the control of proliferation of AT1 cells (Fig. 3) in the digestive tract in the GBF during hypoxia (Satora et al. 2020b).</p><p>In vertebrates, specialized sensory cell types called neuroepithelial sensors, or neuroendocrine cells (NECs), display characteristics of both neurons and hormone-secreting endocrine cells (Lauriano et al. 2021). In the mammalian lung, pulmonary neuroendocrine cells (PNEC) are widely distributed throughout the airway mucosa as solitary cells and as distinctive innervated clusters—called neuroepithelial bodies (NEB). They can detect airborne allergens and relay signals to stimulate immune cells and induce tissue/organ-wide responses. Their increase is associated with a wide range of congenital and infantile lung disorders (Cutz 2015; Jonz et al. 2016; Whitsett et al. 2019). The PNEC and NEB also play an important part in mammalian lung development (Cutz 2015; Whitsett et al. 2019). It is suggested that the groups of neuroendocrine cells represent an ancient mechanism for environmental sensing that integrates epithelial receptors with innate immunity (Lauriano et al. 2021). Understanding their role in lung regeneration and aging is of utmost importance (Cutz 2015; Branchfield et al. 2016; Sui et al. 2018; Whitsett et al. 2019). The chemoreceptors have both receptor and secretory function, and initiate reflex responses to hypoxia (Jonz et al. 2016); they were observed to be active (releasing granules) in hypoxic conditions (Tzaneva et al. 2011)—the strongest air-breathing-inducing factor (Randall et al. 1981; Graham 1997; Nelson 2014).</p><p>Additionally, in the respiratory intestine of the bronze corydoras (Corydoras aeneus), a hypoxia-inducible factor-1α (HIF-1α) has been found, which is considered the main transcriptional regulator of the cellular and the developmental response to hypoxia (Satora et al. 2018).</p><!><p>Proliferation of the squamous epithelial cells observed in the alimentary tract of the GBF in conditions of water hypoxia (Fig. 3) has probably led to the gradual evolutionary development of lungs in terrestrial vertebrates (Satora et al. 2020a). HIF-1α, depending on the normoxic/hypoxic conditions, is one of the most important downstream effector molecules of the EGFR pathway (Lu et al. 2012). Also, NECs—putative chemoreceptors (Figs. 2b,  3)—were found to play an important role in stimulating the development of organs for air breathing in the early terrestrial vertebrates (Jonz 2018; Smatresk 1990). Thus, the most important factors associated with proliferation of AT1 cells, such as HIF-1α, NECs and EGFR, are present in the GBF (Satora and Winnicki 2000; Satora et al. 2017; Mytych et al. 2018; Satora et al. 2018). GBF antibodies directed against human EGFR and HIF-1α were successfully used in immunohistochemical and western blot studies (Mytych et al. 2018; Satora et al. 2018), which additionally facilitates the observations. The research on signals and interactions between those elements in conditions of hypoxia makes it possible to observe a "switching pulse" initiating the proliferation of squamous epithelial cells (Fig. 3). In addition, the state of normoxia causes inhibition of the proliferation process. The GBF can be considered a natural research model of great potential, enabling a breakthrough in research on AT1 cell proliferation.</p><p>The presence of NECs was detected in developing sites of gas exchange in the GBF (Satora and Winnicki 2000; Podkowa and Goniakowska-Witalińska 2002, 2003). Thus, understanding the function of NECs in the formation of the squamous epithelium which enables gas diffusion in the GBF seems crucial (Satora et al. 2020b). On the other hand, experimental studies on NECs of a simple model—GBF—may lead to a breakthrough and contribute to an understanding of the processes which govern proliferation of squamous epithelium, and thus regeneration of respiratory epithelium in the lungs. Therefore, the GBF would seem to be an ideal, low-cost model organism for developmental and molecular biology, but also for physiology.</p><!><p>Understanding the mechanism of proliferation of AT1 cells which enable gas diffusion is critically important. However, models using cell cultures are too simplistic and may lead to misinterpretations (Sims et al. 2005). In turn, studies using mammalian models constitute highly interactive models (Sonnenschein and Soto 2018). Thus, a relatively simple natural model which allows for easy stimulation of squamous epithelial cell proliferation is extremely valuable. Numerous experiments have shown that fishes are promising models for molecular studies, with great potential. For example, the zebrafish (Danio rerio) is a vertebrate model widely used in biomedical research (Bradford et al. 2017).</p><p>There is an increasing body of evidence that air breathing in tetrapods arose as an exaptation. Furthermore, the proliferative ability of squamous epithelial cells, observed in the GBF, seems to be a practical breakthrough which came into existence under the effect of environmental stimulus—hypoxia (Capillo et al. 2021). Using the GBF as a natural model organism opens a completely new avenue which is not available with other models such as mammals and cell lineages. In the studies on such models, tools dedicated to mammals (such as antibodies) can be used successfully (Mytych et al. 2018). Moreover, the model is not expensive and the experiments are relatively easy to conduct.</p><!><p>Despite efforts, it has been impossible to identify the mechanisms which control human lung development, the precise identity and function of human lung stem and progenitor cell types, and the genetic and epigenetic control of human lung fate, without which progress toward the development of strategies for lung regeneration following injury is impossible.</p><p>The suggestion that the evolution of air breathing in all vertebrates occurred through exaptations opens a completely new research perspective. Studies on the mechanisms that control the proliferation of squamous epithelium in the alimentary canal in GBF may contribute to a precise understanding of the signal pathways which govern this process in mammals. This in turn may lead to a breakthrough in the study of mammalian lung regeneration.</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

Electronic Properties and Carrier Trapping in Bi and Mn Co-doped CsPbCl3 Perovskite

Metal halide perovskites exhibit impressive optoelectronic properties with applications in solar cells and light-emitting diodes. Co-doping the high-band gap CsPbCl3 perovskite with Bi and Mn enhances both material stability and luminescence, providing emission on a wide spectral range. To discuss the role of Bi3+ and Mn2+ dopants in tuning the CsPbCl3 perovskite energy levels and their involvement in carrier trapping, we report state-of-the-art hybrid density functional theory calculations, including spin–orbit coupling. We show that co-doping the perovskite with Bi and Mn delivers essentially the sum of the electronic properties of the single dopants, with no significant interaction or the preferential mutual location of them. Furthermore, we identify the structural features and energetics of transitions of electrons trapped at Bi and holes trapped at Mn dopant ions, respectively, and discuss their possible role in determining the optical properties of the co-doped perovskite.

electronic_properties_and_carrier_trapping_in_bi_and_mn_co-doped_cspbcl3_perovskite

<!>Computational Details<!>

<p>Although metal halide perovskite semiconductors1,2 have recently emerged as inexpensive absorber layers in solar cells,3−5 these materials have also shown high mobility,6−8 narrow band emission, a tunable band gap,9−12 photon recycling,13 and bright emission,14 features that are appealing for solid state lighting applications. As an example, the CsPbCl3 perovskite has an appropriate band gap for exciton energy transfer15 and exhibits excellent optical properties like a narrow emission band, a wide color range, and overall promising optoelectronic applications.16 However, this material may suffer from the low photoluminescence quantum yield of the blue–violet radiation that it emits (still <10%).17 Doping lead halide perovskites with different metal ions is an effective approach to tuning their optical, electronic, and magnetic properties,18 through energy or charge transfer interaction between the host and dopant.19 Very recently, by partial substitution of Pb sites with Bi3+ in all-inorganic cesium lead bromide perovskites, Miao et al.20 found the resulting material to show an enhanced absorption over the entire visible spectrum together with a low trap density and a high carrier mobility. Hu et al.21 succeeded in stabilizing the α phase of CsPbI3 by incorporating Bi3+ ions into the perovskite host. The doped compound exhibited enhanced photoelectric performance and moisture stability compared to those of the pure perovskite. Snaith and co-workers pointed out by means of 207Pb NMR and ellipsometry spectroscopy the band gap of MAPbI3 perovskite to be not significantly affected by the introduction of the Bi3+ dopant, which solely contributes to the increase in the number of defects in the material.22 Some of us confirmed this result through density functional theory (DFT) computational analyses highlighting the presence of deep traps associated with Bi in doped MAPbI3 perovskites, which are responsible for the modified optical properties.23 Kang et al.24 found that the free electrons that originated from Bi3+ doping in CsPbCl3 are significantly compensated by the formation of native acceptor defects, because the BiPb substitutional defect (bismuth replacing lead in the crystal lattice) predominantly exists in the 1+ charge state. In a recent investigation of Bi-doped MAPbBr3 thin-film perovskites, Ulatowski et al. observed an enhancement of electron trapping by defects with a resulting ultrafast charge carrier decay in the IR region (∼1.2 eV), possibly associated with the presence of bromine interstitials that originated from compensation of the Bi3+ charged dopant.25</p><p>Mn2+ doping has been widely investigated over the past three decades in the case of semiconductor quantum dots, such as ZnS, ZnSe, CdSe, and ZnO, with the purpose of obtaining luminescence in the orange region of the spectrum.26−28 The incorporation of Mn2+ into semiconducting nanocrystals provides the emergence of a broad photoluminescence peak at ∼600 nm that is widely attributed to a spin-forbidden transition arising from the decay of the 4T1 (t2g4eg1) excited state of Mn2+ to the 6A1 (t2g3eg2) ground state.29−31 Significant efforts were made to understand more about the excitation and de-excitation processes related to the Mn2+ dopant. It is generally accepted that the excited state of the Mn2+ dopant is activated by impact excitation from the optically excited carriers of the semiconducting host to the dopant ion, thus generating the 4T1 state by energy transfer.32−35 Auger recombination has also been demonstrated as a de-excitation pathway under electrochemically controlled charging conditions.36 Very recently Gahlot et al. performed time-resolved spectroscopy studies that support the involvement of a transient Mn3+ species that was proposed to mediate the excitation of Mn2+ in CdxZn1–xSe quantum dots.37</p><p>Mn2+ doping has been also successfully implemented in metal halide perovskites.38 Despite the difference between the ionic radius of Pb2+ (∼133 pm) and that of Mn2+ (∼57 pm), which is responsible for lattice contraction, the stability of doped perovskite nanocrystals was much the same as that of undoped ones.15 The typical Mn2+ emission is still observed in CsPbX3 perovskites with features similar to those of conventional Mn-doped quantum dots. Interestingly, it was found that the nature of the X halide impacts the quantum yield of the dopant photoluminescence, with CsPbCl3 delivering the maximum quantum efficiency.15 Pandey et al. investigated the band structure of Mn:CsPbCl3, finding Mn 3d orbitals within the perovskite band gap, which could contribute to the observed luminescence.39 A similar band structure was found by Pradeep et al. for Mn:CsPbBr3 perovskite, additionally revealing a significant phonon coupling associated with Mn and Pb modes related to dopant/host charge transfer.40 The typical photoluminescence due to Mn was detected also in CH3NH3PbxMn1–xCl3 nanocrystals that show Mn2+ dopant emission at ∼610 nm with a high quantum yield.18 Mn-doped CsPbCl3 shows an optimal photoluminescence quantum yield at low Mn doping and exhibits an emission centered at ∼590 nm.41 In addition to tuning the optoelectronic properties, introducing Mn2+ ions into the perovskite was shown to significantly stabilize the crystal lattice, as illustrated by Akkerman et al. for CsPbxMn1–xI3 nanocrystals.42</p><p>Overall, these results indicate43 that replacing Pb2+ with different metal ions such as Mn2+ (orange-red emission44) and Bi3+ (blue emission45,46) is successful for covering a wide luminescence range.43 Practically white emission can be obtained by a co-doping with such dopant metal ions.47 Shao et al.9 reported that dual ion Bi3+/Mn2+ co-doping of the CsPbCl3 perovskite facilitates stable multicolor and white light emission, exhibiting tunable emission spanning the wide range of correlated color temperature.9 The same authors showed that Bi doping induces a broad photoluminescence band ranging between 440 nm (2.82 eV) and 550 nm (2.25 eV) associated with a progressive decrease in the host photoluminescence at 410 nm. As Mn is added to the material, a sharp emission at ∼600 nm appears in the red region. The simultaneous presence of Bi3+ and Mn2+ was shown to exhibit properties similar to the sum of those detected for the singly doped CsPbCl3 perovskite.9,17</p><p>Given the relevance of Mn/Bi co-doped CsPbCl3 perovskite for optoelectronics applications, we report here DFT calculations of individual Mn2+- and Bi3+-doped perovskites and those of co-doped systems to provide a quantitative understanding of the electronic and structural properties of these materials with inference to the possible carrier trapping at the dopant sites. We employ a state-of-the-art computational strategy combining hybrid DFT and spin–orbit coupling (SOC) that turns out to be crucial both for obtaining a proper energy level alignment and for obtaining reliable structural geometries.48,49 Our results show that Mn2+ can trap a hole and be consequently oxidized into its +3 state, whose state appears deep in the band gap. Notably, the transition energy for recombination of a conduction band electron and a Mn-trapped hole in Mn-doped CsPbCl3 almost coincides with that of the typical 4T1 → 6A1 transition related to the observed orange luminescence, possibly constituting an additional recombination pathway in the doped CsPbCl3 perovskite. The incorporation of Bi, similarly to what happens in MAPbI3,22,23 provides a trap state for electrons that may act as a recombination center with a valence band hole. We additionally investigated the role of the interaction between the two different dopants, simulating adjacent and non-adjacent doping lattice sites. We find that the electronic properties of the co-doped perovskite are not drastically affected by the separation between the Bi and Mn heterometals, suggesting negligible interaction between the two co-dopants.</p><p>Our computational setup delivers a 3.05 eV band gap for the pristine CsPbCl3 perovskite, in excellent agreement with the experimental value of 3.10 eV.9,17 The projected density of states (PDOS) reported in Figure 1 indicates that the description of the CsPbCl3 perovskite electronic structure is, as expected, similar to that of MAPbCl3.50 The valence band is characterized by a major contribution (∼66%) of Cl 3p orbitals (Figure 1a–b), with a significant involvement (∼32%) of Pb 6s orbitals (Figure 1a–c), while the conduction band is almost entirely contributed by Pb 6p orbitals.</p><p>(a) Projected density of states of the pristine CsPbCl3 computed at the HSE06-SOC level of theory with partial Pb contributions (red), Cl contributions (purple), and Cs contributions (blue). (b) Contributions of Cl (s) and Cl (p) orbitals to the density of states of CsPbCl3. (c) Contributions of Pb (s) and Pb (p) orbitals to the density of states of CsPbCl3.</p><p>Compared to the prototypical MAPbI3 perovskite and related lead iodide perovskites, we notice an increased contribution of Pb 6s orbitals to the valence band in CsPbCl3. This is likely due to the stronger electron donation to Pb from chlorine compared to iodine, which increases the energy of the antibonding Pb 6s/Cl 3p combinations leading, together with band gap opening, to the emergence of occupied Pb 6s states.</p><p>Both Mn- and Bi-doped CsPbCl3 perovskites, investigated in the more stable 2+ and 3+ states, respectively, present a small band gap increase (∼0.1 eV) due to the structural distortion introduced by the heteroatoms (Table S1). While this effect could vanish in the limit of infinite dilution, i.e., at significantly lower defect densities, doping clearly introduces hole/electron trap states in the perovskite band gap that play a key role in the perovskite electronic properties. The Bi3+-doped CsPbCl3 perovskite presents an unoccupied state located 2.73 eV above the valence band (VB), and thus quite close to the conduction band (CB) of the pristine perovskite (3.04 eV), which is characterized by an antibonding combination of Bi 6p orbitals and Cl 3p states, as it is visible in the PDOS and in the isodensity plot of the corresponding single-particle orbital reported in panels b and c of Figure 2. When an electron is added to this system, it becomes trapped at the Bi3+ site, formally leading to a reduced Bi2+ center, with the corresponding occupied state now lying 2.17 eV above the VB (Figure 2e−f). Structural data suggest that the lattice does not undergo a significant change upon trapping, with Bi–Cl distances slightly increasing from ∼2.7 to ∼2.8 Å (Figure 2a,d). Notably, by neglecting the effect of SOC in the structural relaxation of the defect-trapped electron, we predicted a significant structural distortion involving the axial Bi–Cl distances, which increase from 2.72 to 3.13 Å (see Figure S1). This structural difference is due to different energetics of the singly occupied orbital where the added electron is located against the CB energy, which would be otherwise unoccupied at the PBE-SOC level of theory. Considering the anticipated impact of both SOC and hybrid functional in precisely tuning the perovskite CB and dopant energy levels, we further investigated the stability of the Bi-doped perovskite with an added electron by carrying out SOC-HSE06 energy evaluations along a linear path connecting the distorted (SR-) and undistorted (SOC-PBE) optimized geometries. This analysis shows a shallow potential energy profile with the SOC-HSE06 minimum located in an intermediate geometry between the SR-PBE and SOC-PBE, lying 0.06 eV below the former. Despite the flat energy surface, the different geometries have a significant impact on the singly occupied orbital representative of the Bi-trapped electron, whose energy decreases by as much as ∼0.7 eV when going from the SOC-PBE to the estimated SOC-HSE06 minimum. This result should be considered as a warning for PBE structural optimizations not predicting the correct geometries for carrier trapping/detrapping in metal halide perovskites. In fact, structural optimizations of neutral iodine vacancies in MAPbI3 incorrectly predict the formation of a Pb–Pb dimer, which is instead not favored by SOC-HSE06.48 This is not the case for Bi3+ where the corresponding orbital is unoccupied.</p><p>Main geometrical parameters calculated for (a) Bi3+ and (d) Bi2+ individually doped structures where equatorial distances are reported in blue and axial ones are reported in red. Projected densities of states (PDOS) computed for the singly doped (b) Bi3+ and (e) Bi2+ perovskites with dashed lines highlighting the states associated with the Bi. The diagrams are aligned with the CsPbCl3 pristine system using the 5d j = 1.5 orbital peak, and the energy reference (zero) is the VB of the pristine perovskite. Isodensity plots of the Kohn–Sham states located under the conduction band for the (c) oxidized and (f) reduced forms. Bi is colored yellow.</p><p>The electronic structure of the Mn2+-doped CsPbCl3 perovskite is not significantly different from that of the pristine material, with the exception of the aforementioned slight increase in the band gap. Occupied states related to Mn2+ are found deep in the VB. By order of decreasing energy, the first Mn contribution to the DOS starts ∼0.65 eV below the VB edge, being represented by partly hybridized Mn–Cl states (Figure 3b−c). These hybridized states extend for ∼2.5 eV, while a main peak related to the Mn 3d shell is found at a still lower energy. This is consistent with previous results for Mn-doped MAPbI3 employing the same level of theory42 but is at variance with the results of ref (39) showing in-gap Mn states, which were obtained by non-hybrid/non-SOC DFT. Similar results are indeed obtained by us by PBE (see Figure S2), clearly pointing to a significant interplay of exact exchange and SOC in determining the electronic properties of Mn-doped perovskites, consistent with ref (40), which found Mn unoccupied states buried in the CB when SOC was included. Also interesting is the fact that the extent of Mn hybridization with perovskite states seems to depend on the nature of the halide (see Figure S2). In the oxidized Mn3+ form, we notice the emergence of an unoccupied orbital, placed 0.87 eV above the VB (Table S1 and Figure 3e), which originates from an antibonding combination between a Mn state and chlorine p orbitals, as is shown by the isodensity plot reported in Figure 3c–f. A Jahn–Teller distortion in the equatorial plane occurs for Mn3+, which is signaled by the considerable decrease in the Mn–Cl bond length compared to the Mn2+ case, with a bond shortening from ∼2.6 to ∼2.4 Å (Figure 3a–d).</p><p>Main geometrical parameters calculated for (a) Mn2+ and (d) Mn3+ individually doped structures where equatorial distances are reported in blue and axial ones are reported in red. Projected density of states (PDOS) computed for the singly doped (b) Mn2+ and (e) Mn3+ perovskites with dashed lines highlighting the states associated with the Mn. The diagrams are aligned to the CsPbCl3 pristine system using the 5d j = 1.5 orbital peak, and the energy reference (zero) is the VB of the pristine. Isodensity plots of the Kohn–Sham states located in the band gap for the (c) reduced and (f) oxidized forms. Mn is colored orange.</p><p>We can now combine the electronic structure data described above to identify the possible trapping/detrapping pathways for non-interacting Bi3+- and Mn2+-doped CsPbCl3 (Figure 4). The presence of a Bi state below the CB (2.73 eV) traps the photoexcited electrons forming a Bi2+ species (Figure 4a). Subsequently, the electron may recombine with a hole in the VB and regenerate the Bi3+ species. This process is associated with an emission falling at 1.49–2.17 eV (SOC-HSE06 or SOC-PBE minimum), thus underestimating the experimental value of 2.6–2.8 eV.9 Closer agreement with the experiment is obtained if one considers a transition from the unoccupied state in the gap (∼2.7 eV), although emission should be accompanied by occupation of the state and its associated structural relaxation, so unless nonthermalized emission takes place, the former values should be representative of the process.</p><p>(a) Electron trapping pathway in Bi3+-doped CsPbCl3 perovskite and (b) hole trapping pathway in Mn2+-doped CsPbCl3 perovskite.</p><p>For the Mn2+-doped perovskite, our analysis suggests the process schematized in Figure 4b. After electronic excitation, hole trapping at Mn2+ leads to geometrical distortion due to the Jahn–Teller effect on the Mn3+ site that causes the emergence of a hole state located 0.87 eV above the VB. This state may provide a possible de-excitation pathway by recombining with a CB electron with an associated transition energy of 2.21 eV, which is close to the experimental value of ∼2.1 eV.17,48 We notice, however, that most of the literature on Mn2+-doped semiconductors agrees on such luminescence as originating by the crystal field 4T1 → 6A1 transition of the Mn2+ ion, populated by energy transfer from the host to the Mn2+ ion.15 In this case, the energy of the 4T1 → 6A1 transition related to the luminescence, experimentally observed at 2.1 eV, can be overlapping with the transition corresponding to hole trapping at Mn2+, thus constituting a possible additional recombination channel. We notice that the contribution of the transient Mn3+ intermediate state precursor to the 4T1 → 6A1 luminescence has been very recently documented in Mn2+-doped CdZnSe quantum dots,37 suggesting a similar intermediate process taking place in CsPbCl3, as well.</p><p>To understand if the presence of the two dopants in the same CsPbCl3 perovskite host would influence the electronic properties of the material, we investigated Mn/Bi co-doped crystals where the two metals are placed in a non-interacting position lying on two different layers of the lattice at a distance of ∼12 Å and in two adjacent octahedra (see Figure 5a−b). We then computed the Mn3+/Bi3+ and Mn2+/Bi3+ co-doped perovskites for both the interacting and non-interacting cases. We found only a slight variation of the energy levels [∼0.1 eV at most (see Figure 6)], suggesting that singly doped systems are reliable models of the Mn/Bi co-doped systems. Furthermore, the concomitant presence of the two heterometals in an interacting position at a relative distance of ∼6 Å (Figure 5b) does not significantly alter the electronic properties (compare panels c and d of Figure 6 to panels a and b of Figure 6) or the relative stability of the two systems, with the interacting Mn2+/Bi3+ configuration lying within 0.02 eV with respect to the non-interacting Mn2+/Bi3+ one. These findings are in agreement with the experimental data where the photoluminescence observed in co-doped nanocrystals is essentially unmodified with respect to that observed in individually doped systems.9,17</p><p>Supercells employed for the modeling of (a) non-interacting and (b) interacting Mn/Bi CsPbCl3 co-doped perovskites. Mn is colored orange, and Bi yellow.</p><p>Projected densities of states (PDOS) computed for the non-interacting co-doped (a) Mn2+/Bi3+ and (b) Mn3+/Bi3+ perovskites and for interacting (c) Mn2+/Bi3+ and (d) Mn3+/Bi3+ ones. The PDOS diagrams are aligned with the CsPbCl3 pristine system using the 5d j = 1.5 orbital peak, and the energy of the VB of the pristine is set to zero.</p><p>The simultaneous presence of Bi3+ and Mn2+ thus shows properties similar to the sum of those detected for the singly doped CsPbCl3 perovskite without significant interaction effects of the two dopant sites. This can be visualized in the global dopant energy levels of Figure 7.</p><p>Graphical representation of band edge and dopant trap states for CsPbCl3 individually doped by (a) Bi3+ and Mn3+ and Mn3+/Bi3+ co-doped in (b) non-interacting and (c) interacting configurations.</p><p>In summary, we calculated the structural and electronic properties of singly doped and co-doped Bi/MnCsPbCl3 perovskites by state-of-the-art first-principles calculations. We find a significant combined effect of spin–orbit coupling and hybrid functional on structural features of electrons trapped at the Bi dopant site, with scalar relativistic geometry optimization leading to large structural deformation. Electron trapping at the Bi dopant site introduces a defect state within the material band gap, as well as hole trapping at Mn. A significant Jahn–Teller distortion occurs upon hole trapping at the Mn2+ defect, while the extent of structural relaxation upon electron trapping at Bi depends on the considered level of theory, SOC-HSE06 results delivering a partial distortion. The energy of the 4T1 → 6A1 transition related to the typical luminescence related to Mn doping in conventional semiconductors is overlapping with the transition corresponding to hole trapping at Mn2+, thus constituting an additional recombination channel in the CsPbCl3 perovskite. A possible role of Mn3+ in mediating excitation transfer from the perovskite host to the excited Mn2+ dopant is proposed, which may constitute an additional recombination channel for photogenerated charge carriers.</p><!><p>The equilibrium structures of individually doped, co-doped, and pristine CsPbCl3 perovskites were modeled in a 2 × 2 × 2 supercell with a total of 160 atoms, employing the tetragonal structure of the material, using the Perdew–Burke–Ernzenhof (PBE) exchange-correlation functional51 including scalar relativistic (SR-PBE) and spin–orbit corrections (SOC-PBE) and relaxing ion positions until forces on atoms were less than 0.001 Ry Å–1. All PBE calculations were performed with ultrasoft pseudopotentials and the plane wave basis set implemented in the Quantum Espresso Program Package.52 Cutoffs on the plane waves and the charge density of 25 and 200 Ry, respectively, were used, sampling the Brillouin zone at the k-point Γ. The dopants, Mn and Bi, were disposed in a substitutional position, in place of the metal lead, assuming a doping concentration of 3.12%, and the supercell was relaxed with the same procedure explained previously with fixed cell parameters. In the case of co-doped perovskite, structures with different relative positions of the two substitutional cations were considered, simulating interacting and non-interacting dopants. To investigate the optical properties of the systems, all PBE-SOC geometries involving pristine, doped, and co-doped perovskites were refined employing the Heyd-Scuseria-Ernzerhof 2006 (HSE06) hybrid functional,53 including spin–orbit coupling corrections. Norm-conserving (NC) pseudopotentials were used with a cutoff on the wave function of 40 Ry and a cutoff on the Fock grid of 80 Ry, sampling at the Γ point of the Brillouin zone. An increased fraction of exact exchange has been included in the HSE06 functional α = 0.43. It has been shown that this computational setup provides accurate band gaps for MAPbI3 perovskites compared to the experiment and to accurate GW calculations,49 and it is normally employed for the quantitative prediction of thermodynamic ionization levels of defects in these materials.48,54</p><!><p>Aligned energy levels and SR-PBE relaxed structures and DOS (PDF)</p><p>jz0c01567_si_001.pdf</p><p>The authors declare no competing financial interest.</p>

PubMed Open Access

RNase I Regulates Escherichia coli 2\xe2\x80\x99,3\xe2\x80\x99-Cyclic Nucleotide Monophosphate Levels and Biofilm Formation

Regulation of nucleotide and nucleoside concentrations is critical for faithful DNA replication, transcription, and translation in all organisms, and has been linked to bacterial biofilm formation. Unusual 2\xe2\x80\x99,3\xe2\x80\x99-cyclic nucleotide monophosphates (2\xe2\x80\x99,3\xe2\x80\x99-cNMPs) recently were quantified in mammalian systems, and previous reports have linked these nucleotides to cellular stress and damage in eukaryotes, suggesting an intriguing connection with nucleotide/nucleoside pools and/or cyclic nucleotide signaling. This work reports the first quantification of 2\xe2\x80\x99,3\xe2\x80\x99-cNMPs in Escherichia coli and demonstrates that 2\xe2\x80\x99,3\xe2\x80\x99-cNMP levels in E. coli are generated specifically from RNase I-catalyzed RNA degradation, presumably as part of a previously unidentified nucleotide salvage pathway. Furthermore, RNase I and 2\xe2\x80\x99,3\xe2\x80\x99-cNMP levels are demonstrated to play an important role in controlling biofilm formation. This work identifies a physiological role for cytoplasmic RNase I and constitutes the first progress toward elucidating the biological functions of bacterial 2\xe2\x80\x99,3\xe2\x80\x99-cNMPs.

rnase_i_regulates_escherichia_coli_2\xe2\x80\x99,3\xe2\x80\x99-cyclic_nucleotide_monophosphate_level

Introduction<!>Bacterial strains, plasmids, general culture conditions, chemicals, and statistical analyses<!>Extraction of 2\xe2\x80\x99,3\xe2\x80\x99-cNMPs<!>Quantification of 2\xe2\x80\x99,3\xe2\x80\x99-cNMPs<!>Extraction of c-di-GMP and pGpG<!>Quantification of c-di-GMP and pGpG<!>LC-MS/MS parameters<!>Quantification of 2\xe2\x80\x99,3\xe2\x80\x99-cNMPs in \xe2\x88\x86rna expressing pCA24N-rna<!>Cytoplasm/periplasm fractionation<!>Addition of exogenous 2\xe2\x80\x99,3\xe2\x80\x99-cAMP<!>Quantification of 2\xe2\x80\x99,3\xe2\x80\x99-cNMP levels following growth +/\xe2\x88\x92 casamino acids<!>Chloramphenicol-mediated induction of RNA degradation<!>Total RNA quantification +/\xe2\x88\x92 chloramphenicol treatment<!>\xce\xbbDE3 lysogenization of BW25113 WT<!>Overexpression of non-translatable mRNA pACYC-noRBS-mRNA<!>2\xe2\x80\x99,3\xe2\x80\x99-cNMP quantification in shaking and static cultures<!>Assessment of metabolic state in shaking and static cultures<!>Biofilm quantification by Congo Red staining<!>Biofilm quantification by crystal violet staining<!>Quantification of biofilm-related gene transcript levels<!>Quantitative reverse transcription PCR (RT-qPCR) analysis<!>2\xe2\x80\x99,3\xe2\x80\x99-cNMP levels fluctuate during E. coli growth<!>RNase I generates 2\xe2\x80\x99,3\xe2\x80\x99-cNMPs in vivo from RNA degradation<!>RNase I modulates biofilm formation<!>Curli production is upregulated in the rna mutant<!>Discussion

<p>Regulation of nucleotide and nucleoside pools is a crucial physiological process in all organisms. Concentrations of (d)NTPs and (d)NDPs fluctuate in Escherichia coli culture throughout the various phases of growth [1], and imbalanced nucleotide concentrations impair essential cellular processes, including the fidelity of DNA replication, transcription, and translation [2–4]. Numerous other biological functions also are mediated directly by nucleotide second messengers such as adenosine 3',5'-cyclic monophosphate (3',5'-cAMP), cyclic dimeric-3':5'-guanosine monophosphate (c-di-GMP), and guanosine 3'-diphosphate, 5'-triphosphate (pppGpp) in prokaryotes [5–8]. In addition, recent work has implicated primary nucleotide metabolism in the regulation of bacterial biofilm morphology, suggesting involvement in the regulation of other, unidentified processes [9–13]. Along with paradigmatic nucleotides, atypical 2',3'-cyclic nucleotide monophosphates (2',3'-cNMPs) have been quantified in mammalian organs and cells [14–17]. Further investigation demonstrated that concentrations of 2',3'-cAMP increase in response to acute organ stress in eukaryotes, suggesting an interesting link with cellular damage [17, 18]. However, less is known about these unusual nucleotides in bacteria, despite their initial detection in E. coli over 50 years ago [19]. A recent study quantified intracellular and extracellular 2',3'-cCMP and -cUMP levels in Pseudomonas fluorescens culture [20], and 2',3'-cAMP has been quantified in Staphylococcus aureus [21]. However, the biological relevance of these nucleotides in bacteria has not been identified. To date, physiological quantification of the four 2',3'-cNMPs of the canonical RNA nucleobases has not been reported in any prokaryote, enzymes involved in modulating 2',3'-cNMP concentrations have yet to be identified, and the physiological roles are unknown.</p><p>In vitro studies have suggested that the bacterial enzyme ribonuclease I (RNase I) may play a role in the formation of 2',3'-cNMPs [22, 23]. RNase I is an endoribonuclease in the widely distributed RNase T2 superfamily that hydrolyzes short oligoribonucleotides (oligoRNAs) in vitro to generate 2',3'-cNMP monomers, regardless of sequence context. The resulting 2',3'-cNMPs are then slowly hydrolyzed to 3'-NMPs in a second non-specific catalytic step [22, 23]. The gene for RNase I encodes a periplasmic localization sequence and the protein originally was isolated from the periplasm of E. coli [24, 25]. Thus, RNase I was suggested to function in catabolism of extracellular RNA [26, 27], but the physiological substrate(s) remains unclear due to the paucity of in vivo investigations.</p><p>More recently, a cytoplasmic variant of RNase I encoded by the same rna gene was purified from E. coli and characterized in vitro, with the cytoplasmic enzyme exhibiting differences in pH stability and thermal denaturation in vitro, as compared to periplasmic RNase I [28, 29]. A recent study demonstrated that E. coli RNase I activity is inhibited by the 16S rRNA of the 30S ribosomal subunit [30]. However, the physiological relevance of this interaction is unclear due to the use of chimeric rRNA. The study also failed to delineate which form of RNase I (i.e. periplasmic and/or cytoplasmic) binds 16S rRNA. Thus, the biological function of periplasmic RNase I remains incompletely understood, and even less is known about the role of the cytoplasmic variant. A role for cytoplasmic RNase I in the final steps of mRNA degradation was suggested, but never investigated experimentally [28]. Consequently, the physiological function of RNase I remains ambiguous, particularly the putative role of RNase I in mRNA degradation [31, 32]. Initiation of mRNA decay in bacteria typically involves endonucleolytic cleavage of the transcript by RNase E of the RNA degradosome. The inactivated transcript is subsequently degraded by 3' to 5' exoribonucleases, such as RNase II and polynucleotide phosphorylase (PNPase), with the aid of auxiliary degradosome proteins [33]. The short oligoRNAs remaining after RNase II and PNPase processing [34–36] are ultimately degraded by oligoribonuclease (oligoRNase) [37, 38], and a role for RNase I in this final catabolic step has been postulated, but never confirmed [28].</p><p>The present work quantifies 2',3'-cAMP, -cGMP, -cCMP, and -cUMP concentrations in E. coli and demonstrates for the first time that RNase I generates 2',3'-cNMPs in vivo. Experimental perturbation of RNA degradation has validated that RNase I degrades cytoplasmic RNA to generate 2',3'-cNMPs. Furthermore, studies employing a recombinant 2',3'-cyclic nucleotide phosphodiesterase have revealed a role for 2',3'-cNMPs in regulating biofilm formation. This report constitutes the first progress toward understanding the biological functions of 2',3'-cNMP pools in bacteria and offers insight into the physiological processes regulated by RNase I, providing a foundation to further elucidate the roles of these cyclic nucleotides in processes linked to bacterial nucleotide metabolism.</p><!><p>The E. coli strain BW25113 (wild-type, WT) (lacIq rrnBT14 ∆lacZWJ16 hsdR514 ∆araBADAH33 ∆rhaBADLD78) [39] and Keio deletion mutant rna::kanR (deficient in RNase I, ∆rna) in the BW25113 strain background [40] were used for all studies (Supplementary Table S1 and Figure S14), unless specified otherwise. The pKT-CNP plasmid was generated by subcloning the catalytic domain (corresponding to the final 221 amino acid residues) of the Rattus norvegicus CNP gene [41] (UniProtKB-P13233; codon-optimized for E. coli; synthesized by GenScript) into the pKT vector [42] via double digest with restriction enzymes NdeI and SpeI, placing the gene under control of the TetA promoter (inducible with anhydrotetracycline). A catalytically-inactive variant of CNPase (H73L/H152L, numbering based on catalytic domain) [41] was generated via QuikChange mutagenesis (for primer sequences, see Supplementary Table S2 ). To construct plasmid pACYC-noRBS-mRNA, polymerase incomplete primer extension (PIPE) cloning was utilized [43]. To this end, the pACYCDuet-1 vector (EMD Millipore) was amplified by polymerase chain reaction (PCR) to delete both multiple cloning sites (including the ribosome binding site, T7 promoters, and T7 terminator). The 162-bp noRBS-mRNA insert containing its own T7 promoter and T7 terminator (54% GC, purchased as a gBlock fragment from Integrated DNA Technologies) was PCR amplified to install 5'- and 3'-terminal regions complementary to the pACYC vector PCR product for PIPE cloning into the vector (for detailed PIPE cloning procedure and insert sequence, see Supplementary Protocol S2). T7-mediated expression was required because genes lacking a ribosome binding site are poorly transcribed by E. coli RNA polymerase [44]. Plasmid pCA24N-rna was obtained from the ASKA collection [45]. For bacterial growth, isolated colonies from Luria Broth (LB)-agar plates were cultured overnight at 37°C with 225–240 rpm shaking in 3 mL of M9 minimal medium (supplemented with 0.4% glucose and 0.2% casamino acids), unless otherwise noted. The resulting starter culture then was inoculated 1:100 into 10 mL of the same medium in 50-mL Celltreat® conical tubes (sterile, polypropylene; lids left loose for gas exchange) and incubated under the aforementioned conditions, unless specified otherwise. Kanamycin, chloramphenicol, and carbenicillin were used at working concentrations of 25, 30, and 100 μg mL−1, respectively. Prior to 2',3'-cNMP extraction, cells were harvested by centrifugation at 3000 g at 20°C for 10 min, frozen in liquid N2, and stored at −80°C, unless otherwise noted. Analytical standards of adenosine 2',3'-cyclic monophosphate (2',3'-cAMP) and cytidine 2',3'-cyclic monophosphate (2',3'-cCMP) (monosodium salts) were purchased from Carbosynth (Berkshire, UK); standards of guanosine 2',3'-cyclic monophosphate monosodium salt (2',3'-cGMP), uridine 2',3'-cyclic monophosphate monosodium salt (2',3'-cUMP), cyclic dimeric-3':5'-guanosine monophosphate sodium salt (c-di-GMP), and 5'-phosphoguanylyl-3':5'-guanosine sodium salt (pGpG) were purchased from BioLog (Bremen, Germany). The sodium salt of 8-bromo adenosine 3',5'-cyclic monophosphate (8-Br 3',5'-cAMP) was obtained from Sigma-Aldrich. Adenosine 3'-monophosphate (3'-AMP) was purchased from Sigma-Aldrich as the free acid. All data represent at least three biological replicates. Statistical significance was evaluated using a two-sample t-test, where equal or unequal variance was assessed via an F-test. Data were considered statistically-significant for P < 0.05.</p><!><p>Aliquots (10-mL) were harvested from exponential-phase WT cultures (OD600 ~0.4–0.6) and stationary-phase cultures (16 or 24 h post-inoculation) by centrifugation. For 2',3'-cNMP extraction, frozen cell pellets were suspended in 500 μL of ice-cold acetonitrile/methanol/water (2/2/1, v/v/v), as previously described [14]. The cells were lysed by sonication on ice and subsequently centrifuged at 4°C at 3000 g for 10 min. The lysate was concentrated to dryness using a vacuum centrifuge and resuspended in 250 μL of sodium phosphate buffer (50 mM, pH 7.4) containing 0.5 μM 8-Br 3',5'-cAMP as internal standard. The extracts were centrifuged at 12000 g for 30 min at 4°C and transferred to an LC-MS autosampler vial.</p><!><p>Quantification of 2',3'-cNMPs was performed via an internal standard (IS) method, using 8-Br 3',5'-cAMP as the IS. Calibration curves for 2',3'-cAMP, -cCMP, -cGMP, and -cUMP analytes were constructed by plotting the peak area ratio of 2',3'-cNMP/IS against the concentration ratio of cNMP/IS, as described previously [14]. 2',3'-cNMP concentrations were adjusted based on the recovery efficiency of each analyte (Supplementary Figure S1) and normalized to cell density. The concentration of IS was 0.5 μM in all samples for calibration. The concentrations of authentic 2',3'-cNMP analytes ranged from 0.02 – 20 μM (a range over which the analytical response remained linear). A linear regression model was used to generate the calibration curves. All nucleotide concentrations in stock solutions were determined via UV-Vis spectrophotometry (Cary Series, Agilent Technology, Santa Clara, CA, USA).</p><!><p>WT E. coli were cultured overnight (18 h) at room temperature without shaking to late-exponential/early stationary phase (OD600 ~0.7–1 A). The nucleotides were extracted essentially as described previously [46]. The protocol was performed analogously to the 2',3'-cNMP extraction described above, except that cell pellets were suspended in 0.5 mL of ice-cold sodium phosphate buffer (50 mM, pH 7.4) with 1 mM EDTA (0.05 mL of extraction buffer added per 1 mL of bacterial culture harvested).</p><!><p>Quantification of c-di-GMP and pGpG was performed using an IS method, in analogy to that detailed above for 2',3'-cNMP quantification. The concentration of IS was 0.1 μM in all samples for calibration. The concentrations of authentic c-di-GMP and pGpG analytes ranged from 0.0125 – 0.2 μM (a range over which the analytical response remained linear).</p><!><p>The LC-MS/MS methodology was performed as previously described [14], with minor modifications. A Thermo Electron LTQ-FTMS was employed for sample analysis. Chromatographic analysis was performed using a Shimadzu autosampler and a Dionex Ultimate 3000 dual gradient pump. LC-MS instrumentation was controlled by Xcalibur and DCMSlink software (Thermo Scientific). Samples were separated using a reversed-phase Leapsil C18 column (2.7 μm, 150 x 2.1 mm) (Dikma Technologies, Inc; Lake Forest, CA, USA). The mobile phase consisted of water with 0.1% formic acid (A) and methanol with 0.1% formic acid (B). The flow rate was 0.3 mL/min and the following chromatography program was employed: 0% B from 0 to 4 min, then a gradient from 0 to 1.5% B from 4 to 15 min, followed by a gradient from 1.5 to 8% B over 15 to 20 min, followed by holding at 8% B from 20 to 25 min, then a gradient from 8 to 15% B from 25 to 28 min, followed by holding at 15% B from 28 to 35 min, and finally a gradient back to 0% B from 35 to 35.1 min. The column was re-equilibrated by holding at 0% B from 35.1 to 45 min. This chromatography method separates 2',3'-cNMPs from the 3',5'-cNMP regioisomers (Supplementary Figures S8-S11) [14]. The column was washed after analysis of every 2–4 extracts using the following chromatographic method: a gradient from 0 to 100% B from 0 to 2 min, followed by holding at 100% from 2 to 10 min, then a gradient from 0% to 100% C (acetonitrile) from 10 to 12 min, followed by holding at 100% C from 12 to 20 min, followed by a final gradient from 0% to 100% A over 20 to 25 min. The column was re-equilibrated to 100% A from 25 to 40 min. 2',3'-cNMPs were quantified via 10 to 30 uL injections; pGpG and c-di-GMP were quantified via a 45 uL injection. Electrospray ionization was performed in positive-ion mode in the LTQ-FTMS using a capillary voltage of 35 V, a 5 kV needle voltage, a capillary temperature of 275°C, and a 110 V tube lens voltage. Samples were detected in the ion trap using a 1 amu isolation window, and a normalized collision energy of 35 eV. An activation Q of 0.250 was used, with an activation time of 30 ms. Nucleotides were detected based on the protonated parent ions and quantified using the protonated nucleobase fragment ions (Supplementary Figure S12). Peaks were integrated using Xcalibur software (Thermo Fisher).</p><!><p>Cultures of ∆rna harboring plasmid pCA24N-rna were cultured to OD600 ~0.1–0.2 and subsequently induced by addition of 10 μM IPTG. Incubation was continued to OD600 ~0.5–0.6; the cells were harvested and the 2',3'-cNMPs were extracted and quantified, as described above.</p><!><p>Separation of cytoplasmic and periplasmic fractions was performed according to a published procedure, and efficiency of the fractionation procedure was evaluated via SDS-PAGE analysis as described previously [47] (Supplementary Figure S2). Samples collected during exponential growth were resuspended in 100 μL TSE buffer (200 mM Tris-HCl pH 7.8, 500 mM sucrose, 1 mM EDTA). After incubation on ice for 30 min, the suspension was centrifuged at 14000 g at 4°C for 40 min. The supernatant (final periplasmic extract) was stored at −80°C until LC-MS/MS and the pellet (spheroplast) was stored at −80°C until 2',3'-cNMP extraction. Spheroplasts were extracted in the same way as outlined above for cell pellets.</p><!><p>WT cultures (60-mL) were grown to OD600 ~0.4–0.5 A in 250-mL glass Erlenmeyer flasks. Each culture then was split into two equal portions (one for 0.1 mM 2',3'-cAMP addition and one for 0.1 mM 3'-AMP addition). 10-mL samples were collected 20 min after addition of the nucleotides for 2',3'-cNMP extraction.</p><!><p>WT E. coli were cultured in 10 mL of either M9 minimal (0.4% glucose, 0.2% casamino acids) or M9 minimal (0.4% glucose, 1.2% casamino acids) in 50-mL Celltreat® conical tubes (sterile, polypropylene). Upon reaching OD600 ~0.4–0.6, 10-mL samples were harvested for 2',3'-cNMP extraction.</p><!><p>WT cultures (50-mL) were grown to early exponential phase in 250-mL glass Erlenmeyer flasks and split into two equal portions. One portion was treated with 200 μg mL−1 chloramphenicol [48], and the other portion was treated with an equal volume of ethanol as a control. After incubation for 30 min, 10 mL were harvested from all cultures by centrifugation.</p><!><p>BW25113 were cultured as described above for Chloramphenicol-mediated inhibition of mRNA degradation. From these cultures, 1-mL samples were collected pre-chloramphenicol treatment and 30 min post-treatment by centrifugation at 12000 g at 24°C for 5 min to quantify total intracellular RNA via the RNAsnap™ procedure [49]. The cell pellets were suspended in 300 μL of RNAsnap™ extraction solution (95% formamide, 18 mM EDTA, 0.025% SDS, 1% β-mercaptoethanol) and incubated for 7 min in a 95°C sand bath. The samples were centrifuged at 14000 g at 24°C for 10 min and the 260 nm absorbance (A260) of the supernatant was quantified using a NanoDrop™ 1000. The total RNA concentration was calculated from the A260 using an extinction coefficient of 0.025 μg mL−1 cm−1 and normalized to the OD600- and volume-dependent cell density of each sample [50].</p><!><p>The WT BW25113 strain was lysogenized using the λDE3 Lysogenization Kit (Novagen) according to the manufacturer's instructions.</p><!><p>BW25113 (DE3) and BW25113 (DE3) harboring plasmid pACYC-noRBS-mRNA were cultured to early exponential phase (OD600 ~0.2–0.3). All cultures then were treated with 0.4 mM isopropyl-β-D-thiogalactopyranoside (IPTG) to induce expression of the non-translatable mRNA. Upon reaching an OD600 of 0.5–0.6, the cultures were harvested by centrifugation.</p><!><p>WT cultures (10-mL) were grown in 50-mL Celltreat® conical tubes (sterile, polypropylene) at either 37°C with 225 rpm shaking to mid-logarithmic phase (OD600 ~0.4–0.6) or at room temperature without shaking overnight to allow biofilm formation. Cells were harvested from 9 mL of culture and lysed for 2',3'-cNMP quantification as detailed above. Biofilm formation was qualitatively confirmed in the static cultures by crystal violet staining, in analogy to a published procedure [51].</p><!><p>WT E. coli were grown in 100 μL cultures in a 96-well microplate (Corning Costar, sterile, non-treated, polystyrene). One set of cultures was incubated at 37°C with shaking to exponential phase, while the other set was incubated at room temperature without shaking for 24 h. The metabolic state then was assessed using the XTT Cell Proliferation Kit II (Roche) according to the manufacturer's protocol with minor modification. Upon reaching the desired cell density, the XTT labeling mixture (50 μL) was added to each culture and the 490 nm absorbance was immediately recorded using a microplate reader. The A490 was normalized to cell density using the OD600 of each culture.</p><!><p>Congo Red assays were conducted as previously described [52]. Individual colonies of BW25113 WT and ∆rna from LB-agar plates were inoculated into 5 mL LB and cultured overnight in 15-mL plastic culture tubes. In addition, WT E. coli harboring plasmid pKT-CNP or inactive control pKT-CNP-inact were cultured in the same way. The overnights were inoculated 1:50 into 7 mL of YESCA (1% casamino acids, 0.12% yeast extract) containing 0.0025% Congo red in 50-mL Celltreat® conical tubes (sterile, polypropylene) (lids left loose for gas exchange). After reaching an OD600 ~0.3–0.4, 1 mL of each culture was transferred to a 1.6-mL Eppendorf tube and either treated with vehicle or with 25 ng mL−1 anhydrotetracycline (AHT) to induce expression. The cultures were incubated for 48 h at room temperature without shaking (lids left open and tubes were loosely covered in plastic wrap and foil). For biofilm quantification, samples were centrifuged at 12000 g for 15 min and 200 μL of supernatant were transferred to a 96-well microplate (Corning Costar; sterile, non-treated, polystyrene). The absorbance at 500 nm was recorded using a microplate reader. For normalization, each culture was disturbed by pipetting and 200 μL were transferred to a 96-well microplate prior to recording the OD600 using a microplate reader.</p><!><p>Cultures of WT and ∆rna (2-mL) were incubated in 24-well Corning Costar microplates (sterile, non-treated, polystyrene) for 24 h at room temperature without shaking. Biofilm formation was quantified by crystal violet staining according to a published procedure with minor modification [51]. Non-adherent cells were poured out and the microplate was gently submerged twice in a beaker of water. A 0.1% aqueous solution of crystal violet (2.5 mL) was added to each well and the microplate was incubated at room temperature for 15 min. The crystal violet solution was poured out and the microplate was gently submerged three times in a beaker of water to remove residual crystal violet (blotting the plate on a stack of paper towels after each wash). The plate was allowed to dry overnight at room temperature. The crystal violet in each well was dissolved by addition of 3 mL of 30% aqueous acetic acid, and the 570 nm absorbance was measured using a microplate reader and normalized to CFU (quantified by drop plating, according to published procedure [53]).</p><!><p>Analysis of mRNA transcript levels for genes related to E. coli biofilm formation were quantified by the Emory Integrated Genomics Core and analyzed by the Emory Integrated Computational Core. Six E. coli pellets (three biological replicates of WT and three of ∆rna) were submitted for extraction and expression profiling on the Affymetrix E. coli Genome 2.0 Array. RNA was extracted using Qiagen miRNEasy kit w/ on column DNAse. Cells were lysed using 700 μL Qiazol + 100 mg acid-washed beads (150–600 μm) on the Qiagen tissue lyser at 30 Hz for 5 min. RNA was eluted in 30 μL nuclease free water. 1 μL was used to determine OD values on a Nanodrop 1000. 1 μL was used to assess sample profiles on the Agilent 2100 using the RNA 6000 Nano assay.</p><p>Whole-Transcript Expression Analysis (Gene ST Arrays) was performed as follows. 10 ng of RNA was processed according to the GeneChip® WT Pico Reagent Kit protocol. Labeled cDNA was hybridized to the E. coli Genome 2.0 microarray for 16–18 hours at 45°C. Hybridized microarrays were washed and stained on an Affymetrix GeneChip 450 fluidics station using the appropriate chip dependent fluidics script. Intensity data was extracted using an Affymetrix 7G scanner and the Command Console software suite.</p><p>The obtained expression data from the microarray experiment were analyzed using 'limma' package in R/Bioconductor (http://www.r-project.org). The raw data were log2 transformed, and Robust Multi-array Average (RMA) normalized to normalize the intensity data between the samples. The differentially expressed genes were identified on the basis of Benjamini-Hochberg (BH) multiple test adjusted P values (i.e. FDR) and fold changes (the increase in number of gene copies). Genes with an FDR value <0.05 and log2 fold change ≥1.0 were considered significantly differentially expressed. Heat maps were created on the z-score-normalized probe signal using the R/Bioconductor function 'hclust' from the 'heatmap3' package. PCA was done using the R/Bioconductor function 'princomp' also applied to z-score normalized expression data.</p><p>Gene expression data obtained from the microarray experiment have been submitted to ArrayExpress at EMBL-EBI (http://www.ebi.ac.uk/arrayexpress/) under accession number E-MTAB-6095.</p><!><p>Total RNA was extracted using the Direct-zol™ RNA MiniPrep Kit (Zymo Research, Irvine, CA; Cat. no. R2050) according to the manufacturer's instructions with optional on-column DNase treatment. Subsequently, 1 mg of total RNA was used as template to synthesize cDNA with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA; Cat. no. 4368814). Primers for all assays were designed using Primer 3 [54] (also see http://www.ncbi.nlm.nih.gov/tools/primer-blast/index.cgi). For primer sequences, see Supplementary Table S6. Melting curve analysis was performed to insure single-product amplification for all primer pairs. Real time PCR was performed on the BioRad CFX384 Real Time System (BioRad, Hercules, CA) using assays specific to the genes of interest. Each reaction well contained 5 mL of PowerUp™ SYBR Green Master Mix (Applied Biosystems; Cat. no. A25742), cDNA equivalent to 20 ng of total RNA and 250 nM each of forward and reverse amplification primers in a final reaction volume of 10 mL. Cycling conditions were as follows: 95°C for 10 minutes for polymerase activation, followed by 40 cycles of 95°C for 15 seconds and 60°C for 1 minute. Data analysis was performed using CFX Manager software from BioRad, version 3.1. The experimental Cq (cycle quantification) was calibrated against the endogenous control products DNA-directed RNA polymerase subunit beta (rpoC). Samples were analyzed for relative gene expression by the DDCt method [55].</p><!><p>Physiological concentrations of 2',3'-cAMP, -cCMP, -cGMP, and -cUMP were quantified in wild-type (WT) E. coli BW25113 [39] during growth in M9 minimal medium using a sensitive LC-MS/MS-based protocol [14]. In exponentially growing E. coli BW25113 cultures, the various 2',3'-cNMPs exist at intracellular concentrations of approximately 10–30 μM (Figure 1). After 16 h of growth, the four 2',3'-cNMP concentrations fall to undetectable levels (limit of detection [LOD] of the LC-MS/MS assay is approximately 150–500 fmol for the different 2',3'-cNMPs [14]). Intriguingly, the 2',3'-cNMP concentrations then increase in 24 h-old cultures to nanomolar levels, approximately 40–240 fold lower than the exponential-phase concentrations, depending on the particular nucleotide (Figure 1). E. coli also exhibit distinct relative concentrations of the different 2',3'-cNMPs, maintaining 2-fold and 5-fold higher concentrations of the 2',3'-cyclic purines relative to the pyridimines in exponential phase and stationary phase (24 h-old) cultures, respectively (Figure 1). These data demonstrate that relative and absolute 2',3'-cNMP concentrations are regulated during E. coli growth (Figure 1).</p><!><p>To identify the enzyme(s) involved in 2',3'-cNMP production, cyclic nucleotide levels were quantified in an E. coli strain deficient in RNase I (BW25113 ∆rna) [40], as this enzyme generates 2',3'-cNMP monomers in vitro from short (~2–12-nt), unstructured oligoribonucleotides [23]. The results demonstrate that essentially all 2',3'-cNMPs produced during exponential and stationary phase growth arise from RNase I activity (Figure 2A and Supplementary Figure S11), as 2',3'-cNMP levels in the ∆rna strain were below the limit of detection. To solidify the role of RNase I in 2',3'-cNMP production, the ∆rna strain was transformed with a plasmid encoding the rna gene, which restored 2',3'-cNMP production (Supplementary Figure S7).</p><p>Although the rna gene encoding RNase I contains a periplasmic localization sequence [27], previous reports of a cytoplasmic RNase I variant encoded by the same gene [28, 56] necessitated separate quantification of periplasmic and cytoplasmic 2',3'-cNMP concentrations to determine the cellular localization of 2',3'-cNMPs. Importantly, 2',3'-cNMPs exist in both the periplasm and the cytoplasm, demonstrating that cytoplasmic RNase I degrades cytosolic RNA (Figure 2B). To probe the possibility that elevated 2',3'-cNMPs arise due to import following periplasmic degradation of extracellular RNA by RNase I, 0.1 mM exogenous 2',3'-cAMP was added to exponentially-growing cultures of the wild-type BW25113 strain in a separate experiment. This addition amounts to approximately 20,000 pmol exogenous 2',3'-cAMP added per 1×108 cells - over 4000-fold greater than the typical physiological 2',3'-cAMP concentration of ca. 5 pmol per 1×108 cells in exponential phase (Figure 1). 2',3'-cAMP levels were unaffected by exogenous nucleotide addition, showing no change compared to cultures treated with exogenous 0.1 mM 3'-AMP as a control (Figure 3A), further supporting that 2',3'-cNMPs are generated from RNA degradation in the cytoplasm.</p><p>Currently, the biological role of RNase I is unknown, but the nuclease is not essential for growth in E. coli [56]. Based on in vitro studies, cytoplasmic RNase I has been proposed to complete the catabolism of short oligoribonucleotides generated from mRNA degradation in vivo, but this function remains speculative [28]. In accord with the inabitility to degrade structured RNA substrates in vitro, RNase I is not involved in the initial inactivation of mRNA, as previously determined by quantifying the half-life of functional β-galactosidase transcript [57]. However, the function of the enzyme in the downstream degradation of short mRNA fragments resulting from transcript inactivation has not been investigated experimentally. To this end, mRNA degradation was perturbed by amino acid starvation and by overexpression of a non-translatable mRNA to probe the effect on 2',3'-cNMP levels. The role of RNase I in the degradation of rRNA also was investigated by chloramphenicol-induced ribosome turnover. Amino acid starvation has been demonstrated to induce expression of a number of Escherichia coli toxin-antitoxin systems, including RelE and MazF, that cleave mRNA [58, 59]. Moreover, E. coli lacking endoribonuclease toxin RelA display dysregulated activation of amino acid biosynthetic genes in the wake of nutrient deprivation, demonstrating the importance of toxin-antitoxin systems in responding to amino acid limitation [60]. Therefore, if 2',3'-cNMPs are formed during mRNA degradation, 2',3'-cNMP levels should be dependent on the presence of amino acids in the growth media. Indeed, E. coli BW25113 grown in minimal media with 1.2% casamino acids exhibit markedly lower concentrations of 2',3'-cNMPs than the same strain grown in minimal media with 0.2% casamino acids (Figure 3B).</p><p>To further solidify a function for RNase I in mRNA degradation, a plasmid-borne gene lacking a ribosome-binding site (pACYC-noRBS-mRNA) was overexpressed to increase the intracellular mRNA concentration, with the expectation that 2',3'-cNMP levels would increase upon expression of the mRNA substrate. In accord with the hypothesis, E. coli cultures expressing the non-translatable mRNA displayed ~2–2.7-fold higher 2',3'-cNMP levels compared to control cultures lacking the plasmid, providing additional validation that 2',3'-cNMPs arise from RNase I-mediated degradation of mRNA (Figure 3C). Collectively, these experiments identify a novel role for RNase I in mRNA catabolism.</p><p>The role of RNase I in ribosome decay was interrogated by treating WT E. coli cultures with chloramphenicol to stall translation and concomitantly increase rRNA turnover [48, 61]. Cultures treated with chloramphenicol displayed higher 2',3'-cNMP levels relative to concentrations in the ethanol-treated control cultures (Figure 3D), demonstrating that the increased 2',3'-cNMP levels arise from rRNA degradation. Moreover, a control experiment confirmed that chloramphenicol treatment altered the total RNA concentration, as expected (Supplementary Figure S6). These studies provide evidence that cytoplasmic RNase I is involved in degradation of mRNA and rRNA to yield 2',3'-cNMPs. Additional work is underway to determine whether RNase I also degrades tRNA.</p><!><p>Several reports have established intriguing links between nucleoside/nucleotide pools and bacterial biofilms [9–13], which are microbial communities of aggregated cells growing in an extracellular matrix of polysaccharides, nucleic acids, and other biopolymers [62]. Therefore, the roles of RNase I and 2',3'-cNMPs in biofilm formation were interrogated. Levels of 2',3'-cNMPs for E. coli BW25113 WT cells grown in shaking versus static culture first were investigated. Significant differences in 2',3'-cNMP levels were observed; quantification yielded approximately 15-fold lower levels of all 2',3'-cNMPs for cells grown in static, biofilm-forming cultures, as compared to shaking cultures (Figure 4A). To test the possibility that the decreased 2',3'-cNMP levels observed in sessile culture were simply a result of reduced metabolism compared to planktonic cells, the metabolic state of the cultures was assessed via a colorimetric tetrazolium-based assay. Although the static/20°C cultures exhibited approximately 1.3-fold decreased metabolism relative to the shaking/37°C cultures (Supplementary Figure S3), the metabolic difference is not sufficient to explain the 15-fold lower 2',3'-cNMP concentrations in the static cultures (Figure 4A). Furthermore, compared to WT BW25113, which is known to form a relatively poor biofilm [63], biofilm formation increased over 10-fold in the ∆rna strain (Figure 4B and Supplementary Figure S5), which does not have observable levels of 2',3'-cNMPs. These results demonstrate that 2',3'-cNMP concentrations are correlated with biofilm formation, with sessile cells having low levels of 2',3'-cNMPs.</p><p>One possible explanation for the dysregulated biofilm formation in the rna mutant is aberrant c-di-GMP signaling. Previous work with Pseudomonas aeruginosa demonstrated that deletion of oligoRNase, which degrades 2–5-nucleotide RNAs, increases the concentration of 5'-phosphoguanylyl-3':5'-guanosine (pGpG), the immediate degradation product of the important biofilm regulator c-di-GMP [64, 65]. Accumulation of pGpG inhibits c-di-GMP-specific phosphodiesterases, thereby increasing the c-di-GMP concentration in P. aeruginosa lacking oligoRNase, resulting in upregulated biofilm production [64, 65]. However, E. coli contains both oligoRNase and RNase I [66], while P. aeruginosa lacks a known homolog of RNase I. Due to the similar capability of oligoRNase and RNase I to hydrolyze short oligoRNAs in vitro [28, 67], the effect of RNase I deletion on intracellular pGpG and c-di-GMP levels in E. coli was investigated. Intriguingly, neither c-di-GMP nor pGpG levels were altered in ∆rna relative to WT E. coli (Figure 4C), suggesting alternative c-di-GMP-independent mechanisms for RNase I and/or 2',3'-cNMPs in modulating biofilm formation in this bacterium.</p><p>To independently investigate the role of RNase I versus the role of 2',3'-cNMPs in biofilm formation, the catalytic domain of a mammalian 2',3'-cyclic nucleotide phosphodiesterase (CNPase, UniProtKB-P1323) [68] was developed as an inducible tool to hydrolyze 2',3'-cNMPs in WT E. coli expressing RNase I. WT cells harboring plasmid pKT-CNP or inactive variant pKT-CNP-inact as a control were assayed for biofilm formation via Congo red assay. Biofilm formation increased in WT cultures expressing CNPase relative to control cultures expressing the inactive CNPase variant, both in the presence of the inducer (anhydrotetracycline, AHT) and under basal expression conditions in the absence of AHT (Figure 4B), thus demonstrating a functional link between 2',3'-cNMPs and biofilm formation. Importantly, expression of active CNPase in sessile cultures decreased levels of the 2',3'-cyclic purine nucleotides below the quantification limit of the LC-MS/MS assay, while reducing concentrations of 2',3'-cCMP and -cUMP 25-fold and 14-fold, respectively, compared to levels in cells expressing the inactive CNPase control (Supplementary Figure S4). These results further indicate that decreasing 2',3'-cNMP levels upregulates biofilm formation.</p><!><p>To provide mechanistic insight into the hyper-biofilm phenotype of the RNase I-deficient strain, which lacks 2',3'-cNMPs, the effect of rna deletion on transcript levels of biofilm-associated genes was investigated. Analysis of the transcriptome indicated 1.5-fold higher expression of curli structural gene csgB and 1.8-fold increased expression of curli accessory gene csgC in ∆rna compared to WT E. coli (Figure 5). Thus, the upregulated biofilm production in the mutant strain is due, at least in part, to increased curli synthesis. Curiously, ∆rna displays decreased expression of the divergently transcribed csgDEFG operon, which activates csgBAC transcription (CsgD) and regulates curli transport and assembly (CsgEFG) (Figure 5) [69]. Upregulated expression of curli genes is consistent with the increased Congo red staining observed in the rna mutant and in WT cells expressing active CNPase (Figure 4B), as Congo red primarily binds to amyloid curli fibers and cellulose [70]. Notably, decreased expression of the pgaABCD locus in ∆rna indicates that elevated poly-N-acetyl-β−1,6-D-glucosamine (PNAG) production is not contributing to the hyper-biofilm phenotype [71, 72]. To validate the surprising downregulated expression of the PNAG biosynthetic operon, quantitative reverse transcription PCR (RT-qPCR) was performed to quantify abundance of the pgaA transcript. The rna mutant displayed reduced pgaA expression, further confirming that increased PNAG synthesis is not responsible for hyper-biofilm production in RNase I-deficient E. coli (Supplementary Table S5).</p><!><p>Absolute and relative nucleotide concentrations are maintained by elaborate regulation of de novo synthesis and salvage pathways. These processes are vital not only in primary metabolism, but also in the coordination of specialized signal transduction cascades which rely on nucleotide second messengers. The present work demonstrates that 2',3'-cNMP concentrations are regulated over the Escherichia coli growth curve and are generated by RNase I-catalyzed degradation of mRNA and rRNA (Figures 1, 2, and 3), presumably, based on the inability of RNase I to digest structured RNA substrates [28], as one of the final steps in RNA catabolism. RNase I homologs exist in several classes within Proteobacteria, indicating that 2',3'-cNMPs likely govern certain biological processes in other bacterial taxa. In addition, genes encoding other RNase T2 superfamily enzymes are conserved across bacteria, eukaryotes, and viruses [73], alluding to possible 2',3'-cNMP-dependent pathways in diverse kingdoms of life. The present results suggest that 2',3'-cNMP pools constitute a previously unknown facet of primary nucleotide metabolism and/or a novel nucleotide second messenger signaling system. 2',3'-cNMPs and the corresponding 3'-NMPs resulting from enzymatic hydrolysis possibly function as intermediates in a novel salvage pathway, as the nonspecific nucleotidase SurE in the cytoplasm accepts 3'-NMPs as substrates [74]. Analysis of previously published NTP, NDP, NMP, and nucleoside concentrations in E. coli fails to suggest many obvious parallels between the 2',3'-cNMP ratio and other nucleotide/nucleoside pools [1, 75]. However, the finding that 2',3'-cNMP levels decrease in stationary phase E. coli cultures relative to exponential phase cultures mirrors the previously observed growth-dependent fluctuation in dNTP concentrations [1] (Figure 1). The present study also reveals an increased concentration of 2',3'-cAMP and 2',3'-cGMP compared to 2',3'-cCMP, and -cUMP in exponential and stationary phase cultures (Figure 1). The elevated 2',3'-cAMP level could be due to poly-A polymerase (PAPase) activity, as 3'-polyadenylation of mRNA facilitates exonucleolytic degradation in bacteria [33]. The different 2',3'-cNMP concentration ratios observed in exponential and stationary phase E. coli cultures (Figure 1) cannot be explained by preferential activity of RNase I because the enzyme does not display strong sequence or nucleobase specificity in vitro [28]. These results allude to a complex regulation of 2',3'-cNMP metabolism, which likely intersects with processes governing other nucleotide levels. Understanding the regulation of 2',3'-cNMP concentrations will require further investigation of growth-dependent relationships between 2',3'-cNMP, 3'-NMP, and other nucleotide/nucleoside concentrations.</p><p>Recent studies indicate that changes in nucleoside and nucleotide metabolism can alter biofilm formation [9–13], demonstrating the importance of exploring 2',3'-cNMP pools in the context of overall nucleotide metabolism. This work demonstrates that decreasing 2',3'-cNMP levels increases biofilm formation in E. coli due to upregulated production of curli, the major protein constituent of biofilms (Figures 4B and 5, Supplementary Figure S5). Intriguingly, pyrimidine auxotrophy impairs synthesis of curli fibers, and conditions favoring UMP synthesis via pyrimidine salvage, as opposed to de novo UMP biosynthesis, also modify the biofilm matrix by increasing cellulose production [9]. In addition, the pyrimidine antimetabolite cancer drug 5-fluorouracil inhibits E. coli biofilm formation by up-regulating expression of AriR, a transcriptional repressor of biofilm-related genes [11]. Though some of the effectors involved in these processes have been identified in certain bacterial species, such as the CytR transcription factor in Vibrio cholera that de-represses pyrimidine metabolic genes in response to cytidine [12], additional mechanistic details of the pathways connecting nucleoside/nucleotide levels to biofilm formation remain elusive. Notably, published data have shown that CytR is not involved in modulating pyrimidine-dependent biofilm phenotypes in E. coli [9], suggesting that additional unknown factors mediate this process in certain species. These findings suggest that disrupting normal 2',3'-cNMP regulation may alter biofilm formation by perturbing primary nucleotide/nucleoside metabolism, perhaps ultimately impacting c-di-GMP signaling. Although the total c-di-GMP concentration does not differ between WT and ∆rna E. coli (Figure 4C), it remains possible that cells lacking RNase I and 2',3'-cNMPs exhibit dysregulated levels of spatially isolated c-di-GMP pools, as local concentrations of this second messenger mediate biofilm formation [76]. Alternatively, the aberrant biofilm phenotype observed in ∆rna and in WT E. coli expressing CNPase (Figure 4B) potentially could be elicited by a novel second messenger signaling pathway mediated directly by 2',3'-cNMPs, as the micromolar 2',3'-cNMP concentrations in exponential phase E. coli cultures are similar to the basal level of 3',5'-cAMP [75, 77], a canonical second messenger.</p><p>Ongoing work seeks to investigate potential 2',3'-cNMP-mediated signal transduction and elucidate the roles of the different 2',3'-cNMPs in modulating bacterial phenotypes. Gene expression data reported herein demonstrate that E. coli lacking RNase I exhibit aberrant expression of several transcripts relevant to biofilm production (Figure 5 and Supplementary Table S4). Published phenotypic and deep sequencing investigations using E. coli deficient in RNase II, PNPase, or RNase R have linked these processive exoribonucleases to biofilm formation via perturbation of biofilm-associated transcripts [78, 79]. Although the mechanistic intricacies of the uniquely altered transcriptome in these different RNase mutants remain ambiguous, RNase II, PNPase, and RNase R directly impact transcript half-life [80, 81], suggesting that altered mRNA decay is directly influencing the biofilm phenotype. Conversely, previous work has shown that rna deletion does not directly affect transcript stability [57], and the present study demonstrates that RNase I deletion does not perturb global c-di-GMP or pGpG levels (Figure 4C). These data allude to a more complex regulatory mechanism involving 2',3'-cNMPs, which is further demonstrated by the finding that inducing hydrolysis of 2',3'-cNMPs upregulates biofilm formation in WT cells expressing RNase I (Figure 4B).</p><p>The present study quantifies 2',3'-cNMPs in E. coli, demonstrates that RNase I generates 2',3'-cNMPs via mRNA and rRNA degradation, and identifies a role for 2',3'-cNMPs in regulating biofilm formation. The identification of RNase I as the enzyme responsible for generating 2',3'-cNMP pools provides the first insights into the phenotypic consequences of aberrant 2',3'-cNMP concentrations and RNase I levels on biofilm formation in bacteria. Additional experiments are in progress to elucidate the mechanisms that control the relative concentration ratios of 2',3'-cNMP pools and the link to biofilm formation in E. coli and in other bacterial taxa. Given the importance of biofilm formation in virulence and pathogenesis of numerous bacterial species [5, 7], elucidation of the mechanism(s) by which RNase I and 2',3'-cNMPs alter biofilm formation will provide insight into new methods to alter bacterial phenotypes.</p>

PubMed Author Manuscript

Nitro-Group Functionalization of Dopamine and its Contribution to the Viscoelastic Properties of Catechol-Containing Nanocomposite Hydrogels

Linear polyacrylamide (PAAm) is modified with dopamine or nitrodopamine (PAAm-D and PAAm-ND, respectively) to evaluate the effect of nitro-group modification on the interfacial binding properties of polymer-bound catechol. Nanocomposite hydrogels are prepared by mixing PAAm-based polymers with Laponite and the viscoelastic properties of these materials are determined using oscillatory rheometry. The incorporation of a small amount of catechol (\xe2\x89\x880.1 wt% in swollen hydrogel) drastically increases the shear moduli by 1\xe2\x80\x932 orders of magnitude over those of the catechol-free control. Additionally, PAAm-ND exhibits higher shear moduli values than PAAm-D across the whole pH range tested (pH 3.0\xe2\x80\x939.0). Based on the calculated effective crosslinking density, effective functionality, and molecular weight between crosslinks, nitro-group functionalization of dopamine results in a polymer network with increased crosslinking density and crosslinking points with higher functionality. Nitro-functionalization enhances the interfacial binding property of dopamine and increases its resistant to oxidation, which results in nanocomposite hydrogels with enhanced stiffness and a viscous dissipation property.

nitro-group_functionalization_of_dopamine_and_its_contribution_to_the_viscoelastic_properties_of_cat

1. Introduction<!>2.1. Materials<!>2.2. Synthesis of Catechol-Functionalized PAAm<!>2.3. Characterization of Catechol-Functionalized PAAm Polymers<!>2.4. Preparation of Nanocomposite Hydrogels<!>2.5. Characterization of Nanocomposite Networks<!>2.6. Oscillatory Rheometry<!>2.7. Statistical Analysis<!>3.1. Synthesis of Catechol-Functionalized PAAm Polymers<!>3.2. Preparation of Nanocomposite Networks<!>3.3. Oscillatory Rheometry<!>3.4. Effective Crosslinking Density and Molecular Weight between Crosslinks<!>3.5. Effect of Laponite Concentration on Viscoelastic Properties<!>4. Conclusion

<p>The mixing of polymer with nanoparticles is an efficient strategy to fabricate nanocomposite systems with high performance and multiple functionalities.[1–6] Nanocomposite materials rely on weak physical interactions (e.g., hydrogen bonding, ionic interactions, etc.) between the polymer matrix and the inorganic nanoparticles to achieve elevated mechanical properties and extensibility, as well as the ability to recover after large, repeated deformation. Strong interfacial binding is necessary to acheive effective stress transfer between the nanoparticles and the polymer matrix.[7–9] A strong interfacial bond is especially critical for materials that experience cyclic loading to prevent rapid breakdown at the interface.[10] Although the use of covalent bonds to bridge the polymer matrix with incorporated nanoparticles has been described,[11–13] covalently crosslinked composite materials exhibit a significant increase in mechanical properties but a reduction in flexibility and recovery efficiency when the irreversible covalent bonds are broken during deformation.[14] A strong, reversible bond found in marine adhesive chemistries is exploited herein to investigate the effect of interfacial binding strength on the viscoelasticity of nanocomposite hydrogels.</p><p>Marine mussel secretes protenaceous adhesives to enable its attachment to various substrate surfaces under wet, saline conditions.[15,16] The excellent bonding ability of these adhesive proteins is attributed to the presence of a unique amino acid, 3,4-dihydroxyphenylalanine (DOPA).[17] The catechol side chain of DOPA is capable of forming strong reversible bonds with metal oxide with a measured adhesion strength averaging around 800 pN, roughly 40% that of covalent bonds.[18] The use of network-bound catechol (e.g., DOPA, dopamine) to increase the interfacial binding strength between polymer networks and encapsulated nanoparticles in creating nanocomposite materials (e.g., hydrogels,[19] films,[20] nanofibers,[21] rubber,[22,23] and tissue adhesive[24] with drastically improved materials properties has been described. However, the effect of pH on the material properties of these nanocomposites has not yet been reported. pH plays an important role in the oxidation state of catechol and its ability to bind to metal oxide and ions.[25–27] When the pH approaches and exceeds the first dissociation constant of the catechol–OH group (pKa1 = 9.0), DOPA auto-oxidizes to its quinone form with reduced adhesive trength.[18,25,28]</p><p>DOPA is also found in adhesive proteins utilized by sandcastle worms to cement sand fragments into tube-shaped dwellings.[29] The catechol side chain in these adhesive proteins is further modified with an electron withdrawing chloro-functional group at the para position (2-chloro-4,5-dihydroxyphenylalanine), which was proposed as a natural adaptation to increase interfacial binding strength.[30] Substituting −H with an electron withdrawing group (EWG) lowers the dissociation constant (pKa) of the catechol hydroxyl groups and their redox potential, thus making the oxidation of the catechol group more difficult.[31] The oxidized quinone forms weak bonds with the substrate surface when compared to its reduced catechol counterpart.[25,28,32] Similarly, nitro-substituted catechols have been found to form complexes with metal oxides that are more stable than unsubstituted catechol.[33–35] Nanoparticles modified with nitro-DOPA- or nitrodopamine-functionalized polymers were found to have significantly higher colloidal stability when compared to other catechol groups (e.g., DOPA, dopamine, etc.) and could be repeatedly heated to 90 °C without noticeable agglomeration.[34] An electron paramagnetic resonance study indicated that enhanced electron delocalization for nitrocatechols may have contributed to its remarkable binding strength to metal oxide surfaces.[33] Most recently, the nitro-functionalization of dopamine has been found to drastically alter the rate of crosslinking and the degradation rate of nitrodopamine-modified bioadhesives.[36,37]</p><p>The present work seeks to enhance the interfacial binding strength of the catechol moiety with EWG modification and evaluate its effect on the viscoelastic properties of nanocomposite hydrogels. To this end, a linear and water soluble polyacrylamide (PAAm) polymer was functionalized with catechol with and without EWG modification (e.g., nitrodopamine and dopamine, respectively). These polymers were mixed with nanoparticles to yield a homogenous nanocomposite hydrogel using a biocompatible nanosilicate, Laponite (Na0.7+(Mg5.5Li0.3Si8) O20(OH)4)0.7−). Laponite was chosen for this study as its reversible interaction between dopamine has been previously established.[19,24] Additionally, pH was used to further modulate the interfacial binding strength of the polymer-bound catechol moieties and to determine its effect on the viscoelastic properties of the nanocomposite hydrogels.</p><!><p>Dopamine hydrochloride (99%) and ammonium persulfate (APS, 98.0%) were purchased from Acros (New Jersey, USA). Sulfuric acid (H SO , ACS PLUS), hydrochloric acid (HCl, ACS grade), acrylamide (AAm, ≥ 99%), acrylic acid N-hydroxysuccinimide ester (AA-NHS, 90%), and N,N,N′,N′-tetramethylethylenediamine (TEMED, ≈99%) were purchased from Sigma-Aldrich (USA). Laponite XLG was donated by Southern Clay Products Inc. (Texas, USA). All reagents were used as received. Deionized water (D.I. H2O) was used to prepare all aqueous solutions. 0.1 M NaCl solutions with a pH range from 3.0 to 9.0 were prepared as buffer solutions for hydrogel soaking. pH values of 3.0 and 5.0 were adjusted with 0.1 m HCl, and pH values of 7.4 and 9.0 were adjusted using 0.1 m Na2HPO4 and 0.1 m HCl solutions at appropriate ratios. All reagents were used as received. Nitrodopamine hemisulfate was synthesized according to a published protocol.[38]</p><!><p>Dopamine- and nitrodopamine-functionalized polyacrylamide (PAAm-D and PAAm-ND, respectively) were prepared in two steps (Scheme 1). In the first step, a precursor polyacrylamide (PAAm) copolymer containing reactive −NHS ester groups (PAAm-co-AA-NHS) was prepared by free radical polymerization. 50 mL of D.I. H2O containing 0.6 m of AAm and 2 mol% of AA-NHS (relative to AAm) was purged with nitrogen for 30 min before APS and TEMED (APS: 0.2 mol% relative to AAm, TEMED:APS = 2.3:1 mol:mol) were added to trigger the polymerization at room temperature under a nitrogen atmosphere for 3 h. The as-made precursor polymer was precipitated in chilled acetone to remove the redox initiator and unreacted monomers to yield the PAAm-co-AA-NHS precursor polymer.</p><p>In the second step, PAAm-co-AA-NHS was redissolved in 50 mL of D.I. H2O and mixed with 40 mL of borax buffer (0.025 m Na2B4O7·10H2O adjusted to pH 8.0) containing either dopamine hydrochloride or nitrodopamine hemisulfate. The molar ratio between the catechol amine and AA–NHS was fixed at 1.5:1. The reaction mixture was stirred under a nitrogen atmosphere for 24 h. The polymer was dialyzed in dialysis tubing with a molecular weight cut off of 3500 Da for 72 h in D.I. H2O acidified to pH 2.5 by adding 0.1 m HCl and collected by precipitating in chilled acetone to obtain the catechol-functionalized polymers, and dried. The PAAm control was also prepared using same free radical polymerization process without the addition of AA-NHS. 1H NMR (D2O:DMSO-d6, 4:1 v:v, δ): PAAm-D 7.7–7.4 (m, 2H; −C(=O)−NHH), 6.9 (m, H; −C(=O) −NHH), 6.8–6.5 (m, 3H; C6H3), 3.2 (s, 2H; −NH−CH2−CH2−), 2.8 (s, 2H; −NH−CH2−CH2−), 2.1 (m, 1H; −CH−CH2−), 1.5 (m, 2H; −CH−CH2−) (Figure S1, Supporting Information); PAAm-ND 8.0 (s, 1H; C6HH), 7.7–7.4 (m, 2H; −C(=O)−NHH), 6.9 (m, H; −C(=O)=NHH), 6.7 (s, 3H; C6HH), 3.2 (s, 4H; −NH−CH2−CH2−), 2.1 (m, 1H; −CH−CH2−), 1.5 (m, 2H; −CH−CH2−) (Figure S2, Supporting Information).</p><!><p>The molecular weight of the synthesized polymers was analyzed using a gel-permeation chromatograph (GPC, Agilent 1260) equipped with an Agilent G1362A 1260 refractive index detector. Three Waters Ultrahydrogel columns were used, aligned in series, and an aqueous mobile phase (0.1 m sodium nitrate, 0.02 wt% sodium azide; flow rate = 1 mL min−1). Shodex Standard P-82 pullulan was used as the standard. Polymers were dissolved in D.I. H2O at a concentration of 2 mg mL−1, stirred for 5 h, and injected with a 0.45 μm filter. The catechol content in PAAm-D and PAAm-ND was quantitatively determined by UV-vis spectrometry (Lambda 35, Perkin Elmer Inc., USA).[39,40] Briefly, a known amount of the PAAm-D or PAAn-ND was dissolved in D.I. H2O acidified to pH 2.5 and its absorbance was determined at 279 nm and 310 nm, respectively (Figure S3, Supporting Information). Standard curves for dopamine (279 nm) and nitrodopamine (310 nm) were used to determine the amount of catechol coupled to each polymer.</p><!><p>PAAm polymers with or without dopamine or nitrodopamine were homogenously mixed with Laponite nanoparticles to form nanocomposite hydrogels (Scheme 2). Unless otherwise specified, hydrogels were prepared with a ratio of 60:40 (wt%:wt%) polymer:Laponite (i.e., starting polymer and Laponite contents were 60 and 40 wt%, respectively). The polymer (0.100 g) was dissolved in pH 4.4 D.I. H2O (20.0 mL) with vigorous magnetic stirring overnight to give a clear solution (5.0 mg mL−1). Separately, Laponite nanoparticles (0.067 g) were dispersed in D.I. H2O (26.8 mL) with vigorous magnetic stirring for 20 min, and then 0.15 mL of 0.1 m NaCl solution was added with sonication for 5 min to give 2.5 mg mL−1 of clear Laponite suspension containing 0.56 × 10−3 M NaCl. The polymer solution was added dropwise to the Laponite suspension with vigorous magnetic stirring and the stirring was continued for an additional 4 h to obtain a homogenous mixture. The mixture was then transferred into a pre-weighed petri dish and evaporated in the fume hood to give a condensed nanocomposite mixture (≈2.0 wt%) and further added to a circular mold with a diameter of 20 mm and a depth of 2 mm (Figure S4, Supporting Information) and freeze dried. To reconstitute the hydrogels, the freeze dried nanocomposite discs were soaked in excess 0.1 m NaCl buffer solution at a desired pH (3.0, 5.0, 7.4, and 9.0) for 48 h.</p><!><p>The chemical composition of the freeze dried nanocomposite network was determined using Fourier transform infrared (FTIR) spectroscopy (Perkin Elmer Spectrum One). The morphologies of the dried nanocomposite networks were characterized using field emission scanning electronic microscopy (FE-SEM, Hitachi S-4700). A freshly formed cross-section was created by cutting the network in half and coating with Pt/Pd alloy. The equilibrium water content (EWC) of the reconstituted hydrogels was calculated by measuring the mass of the hydrogel before (Ms) and after (Md) drying the samples in vacuum for at least 48 h. EWC was determined using Equation (1).[41]</p><!><p>Oscillatory rheometry was carried out using an HR-2 rheometer (TA Instruments Inc., USA) to perform frequency sweep experiments (0.01–600 rad s−1 with a strain of 10%) to collect storage (G′) and loss (G″) moduli. The reconstituted nanocomposite hydrogel discs (diameter = 20 mm, replicate n = 3) were tested using parallel plate (20 mm) with a gap that was 85% that of the individual hydrogel thickness, as measured by a digital calliper. Mineral oil was applied around the edge of the hydrogel disc to avoid dehydration during testing.</p><!><p>Statistical analysis was performed using JPM Pro 9 software (SAS, Cary, NC). The student t-test and one-way analysis of variance (ANOVA) with Tukey-Kramer HSD analysis were performed for comparing means of two and more than two groups, respectively. A p-value less than 0.05 was considered significant.</p><!><p>PAAm modified with either dopamine or nitrodopamine (PAAm-D and PAAm-ND, respectively) was prepared using a two-step sythesis process (Scheme 1). In the first step, free-radical polymerization was carried out to yield a PAAm copolymerized with AA-NHS, which contained pendant activated ester groups. The reactive −NHS groups were further reacted with the primary amine on either dopamine or nitrodopamine to covalently link these adhesive moieties to the PAAm chain. Both 1H NMR (Figure S1 and S2, Supporting Information) and UV–vis (Figure S3, Supporting Information) confirmed the presence of dopamine and nitrodopamine in PAAm-D and PAAm-ND, respectively. On average there were nearly 3 catechol pendant groups coupled to each polymer chain based on UV-vis (Table 1). For both PAAm-D and PAAm-ND, the resulting yield of the coupled catechol accounted for around 0.5 mol% relative to that of AAm repeating units, corresponding to around a 25% yield based on the theoretical starting concentration of AA-NHS (2 mol%) used during the polymerization step. This yield is comparable to published results (13–35%) for polyacrylate-based polymer containing NHS activated esters.[42] PAAm homopolymer had a higher molecular weight compared to the copolymers, indicating that the presence of AA-NHS affected the growth of the polymer chain. The polydispersity index (PDI) was found to be around 2–2.4. This relatively high PDI was attributed to the fast reaction rate of the free radical polymerization, which resulted in a weight-average molecular weight (M̄w) of around 106 Da within 3 h. Additionally, the high viscosity achieved during the polymerization process of this high molecular weight polymer also contributed to the measured PDI. Nevertheless, both PAAm-D and PAAm-ND were prepared with comparable catechol content and molecular weight to investigate the effect of nitro-group functionalization on the polymer-nanoparticle interfacial binding within nanocomposite hydrogel networks.</p><!><p>To ensure a homogenous distribution of Laponite within the PAAm-based hydrogel network, both the nanoparticle and polymer were mixed in a dilute concentration (combined concentration of 3.5 mg mL−1) and sequentially condensed and freeze dried (Figure S4, Supporting Information). The freeze dried PAAm-D nanocomposite network was white, indicating that no oxidation of the dopamine side chain had occured during the process of network formation. On the other hand, the mixture of PAAm-ND and Laponite both in solution and in the freeze dried network was yellowish in color. This color indicated the presence of nitrodopamine in the nanocomposite network, as nitrodopamine powder is dark yellowish-brown (not shown) in color. FTIR spectra (Figure 1) of various PAAm nanocomposites revealed the characteristic primary amide (3305, 3178, and 1606 cm−1 for primary −NH2 and 1647 cm−1 for C=O) and aliphatic acrylate backbone (2964 and 1448 cm−1 for −CH2−) of PAAm. Additionally, these composites also exhibited a strong contribution from the Si−O−Si stretching of the nanoparticle (984 cm−1). For both PAAm-D and PAAm-ND, catechol peaks were not readily visible due to the low content of the adhesive moieties in the polymer. SEM images of the freeze-dried nanocomposites confirmed that these materials formed porous networks (Figure 2).</p><p>Dry nanocomposite networks were reconstituted in excess aqueous buffer solutions and equiliated for 48 h to form hydrogels. The equilibrium water content (EWC) of these nanocomoposite hydrogels was between 92 and 97 wt% (Figure 3). EWC is an important property of a hydrogel and is inversely proportional to the crosslinking density and mechanical properties of the network.[27,43,44] At all the pH levels tested, the EWC values of PAAm-D and PAAm-ND hydrogels were significantly lower than those of PAAm. This indicates that nanocomposite hydrogels containing the adhesive catechol moieties were more densely crosslinked, presumably due to strong interfacial binding between polymer-bound catechol and Laponite.[19,24] Although there was no significant difference for EWC values measured for PAAm-D and PAAm-ND at pHs between 3.0 and 7.4, EWC for PAAm-ND measured at pH 9.0 was lower than that of PAAm-D. These results suggested that the nitrodopamine still formed strong bonds with Laponite even at a pH above the proton dissociation constant of the first −OH group (pKa1 = 6.5 and 9 for nitrodopamine and dopamine, respectively.[33,45]</p><!><p>The viscoelasticity properties of PAAm and Laponite nanocomposite hydrogels with or without the incorporation of dopamine and nitrodopamine moieties were determined using oscillatory rheometry (Figure 4,5 and 6). Regardless of the polymer or pH, nanocomposite hydrogels exhibited very similar responses. For all samples, the measured storage modulus (G′) values were greater than those of the loss modulus (G′′) values across all the frequency ranges tested, indicating that these materials behaved as crosslinked hydrogel networks. For PAAm-D and PAAm-ND, crosslinks were formed between polymer-bound catechol and Laponite.[19] The catechol-free PAAm control also formed a hydrogel network but with lower G′ and G″, presumably due to the weak PAAm-Laponite interaction.[46,47] Additionally, both G′ and G″ increased with increasing frequency, similar to other reported physically crosslinked hydrogels.[46–48] At a frequency less than 100 rad s−1, both G′ and G″ exhibited power law behavior in response to the shear frequency (ω), where G′ ~ ωn′, G″ ~ ωn″, and n′ and n″ are the relaxation exponent for G′ and G″, respectively.[49] The relaxation exponent values reported here (Table S1, Supporting Information) are in agreement with physically crosslinked nanocomposite hydrogels composed of PAAm and Laponite (0.1–0.18).[47] n′ for both PAAm-D and PAAm-ND were lower than that of PAAm, indicating that the catechol-containing nanocomposite behaved more elastically.[49] G′ rose sharply while G″ decreased sharply when the frequency increased beyond 100 rad s−1. When the hydrogels were repeatedly deformed at a high frequency, the polymer chains failed to rearrange themselves within the short time scale of the imposed mechanical deformation, which led to stiffening of the networks.[50,51] This result further confirmed the formation of hydrogel networks.</p><p>Nanocomposite hydrogels containing catechol adheisve moieties exhibited G′ and G″ values that were 1–2 orders of magnitude higher than those of the catechol-free hydrogel (Figure 7). This drastic difference in the measured moduli values was attributed to the strong wet adhesive properties of the catechol group. Increased measured G′ values for catechol-containing materials indicated an elevated stiffness as a result of increased effective crosslinking density due to the formation of strong catechol-Laponite bonds. Increased crosslinking density in the catechol-containing network was corroborated with reduced EWC (Figure 3). Elevated G″ values indicated strong viscous dissipation properties resulting from the breaking of reversible physical bonds between catechol and Laponite.[19,24] The measured increase in materials properties as a result of catechol-functionalization was remakable, considering the extremely low content of catechol moiety in the nanocomposite hydrogel (≈0.1 wt% in the swollen network).</p><p>Both PAAm-D and PAAm-ND nanocomposite hydrogels exhibited higher G′ and G″ values under acidic conditions (Figure 7). For PAAm-ND, maximum G′ and G″ values (4200 and 1000 Pa, respectively) were observed at pH 5.0 and both moduli decreased in value with a further increase in pH. On the other hand, G′ and G″ was maintained at around 1700 and 500 Pa, respectively, for PAAm-D tested at pH values between 3.0 and 7.4, before these values decreased to 1000 and 300 Pa, respectively, when the pH was raised to 9.0. When comparing the rheological response of PAAm-D and PAAm-ND nanocomposite hydrogels, PAAm-ND demonstrated equivalent or higher G′ and G″ values when compared to those of PAAm-D. Measured shear moduli values for PAAm-ND were statistically higher at pH 5.0 and 9.0 when compared to those of PAAm-D.</p><p>Rheological data further confirmed that the interfacial binding properties of the catechol side chain were responsible for the elevated materials properties. For example, 4-nitrocatechol demonstrated maximum absorption to inorganic substrates at a pH of around 6,[52] which corresponded well with the pH dependent behavior of the shear moduli measured for PAAm-ND. Additionally, the point of zero charge (PZC) for silica oxide (SiO2) occurs at around a pH of 5,[53] and under a more acidic or basic condition, proton or hydroxide ions can compete with catechol for binding onto the charged oxide surface.[54,55] Furthermore, the reduced form of catechol is responsible for the strong moisture-resistant adhesion to inorganic substrates and its binding strength decreases with oxidation.[25,28,32] When both nitrodopamine and dopamine exist predominantly in their reduced forms at pH 5.0, PAAm-ND significantly out performed PAAm-D, indicating that EWG modification greatly enhanced the affinity of catechol towards oxide surfaces. When the pH is raised above the pKa1 of nitrodopamine (pKa1 = 6.5) but below that of dopamine (pKa1 = 9) at pH = 7.4, the shear moduli of PAAm-ND decreased and there was no statistical difference between PAAm-ND and PAAm-D samples. Further increasing the pH to around the pKa1 of dopamine (pH = 9.0) resulted in a further decrease in the measured shear moduli values. However, PAAm-ND exhibited higher shear moduli than PAAm-D, suggesting that the presence of an electronegative −NO2 group rendered the nitrodopamine less prone to oxidation.[56–58]</p><!><p>The rheological and EWC data were used to calculate the effective crosslink density (veff) of the nanocomposite hydrogels using the Equation (2) while assuming an affine phantom network:[59,60]</p><p>where G is the shear modulus, feff is the effective functionality of the crosslinks, R is the gas constant, T is the temperature in Kelvin, and φ2 is the volume fraction of the polymer in the hydrogel. G′ recorded at a frequency of 1 rad s−1 was used as G.[61] In the nanocomposite hydrogels used here, Laponite acts as the site for forming multifunctional crosslinking points and feff was calculated based on Equation (3):[62]</p><p>where Mw,L is the molecular weight of Laponite (2.5 × 106 g mol−1)[61] and CL is the mass concentration of Laponite in the hydrogel. φ2 was calculated based on the EWC data using the density of water (1 g cm−3), PAAm (1.42 g cm−3),[63] and Laponite (2.53 g cm−3),[64] while assuming the mass ratio of PAAm and Laponite remain the same as the starting mixture at a 60:40 wt%:wt% ratio.</p><p>For a given pH, veff values for PAAm were significantly lower than those of PAAm-D and PAAm-ND (Table 2), indicating that the presence of catechol drastically increased crosslinking density. Similarly, PAAm-ND were more densely crosslinked than PAAm-D. This increase in the crosslinking density was associated with the formation of crosslinking points with higher functionality (i.e., a number of elastically effective network chains extending from Laponite) as measured by feff. These results are in agreement with published observations, where increased interfacial binding strength between polymer matrix and encapsultated nanoparticles increased both the crosslinking density and functionality of the network.[65] Specifically, PAAm-ND exhibited equal or higher veff and feff values when compared to those of PAAm-D. Additionally, PAAm-ND equillibriated at pH 5.0 exhibited the highest veff and feff values reported for all the formulations tested. At pH 9.0, both veff and feff values for PAAm-ND and PAAm-D were equivalent. Given that PAAm-ND exhibited significantly higher moduli values than PAAm-D (Figure 7), our results further confirmed that nitrodopamine formed stronger bonds with Laponite even at a basic pH. Both veff and feff values reported here are of the same order of magnitude as other physically crosslinked nanocomposite hydrogels.[61,62,65]</p><p>The calculated effective crosslinking density was further used to determine the molecular weight between crosslinks (M̄c ) using Equation (4):[66]</p><p>where ρpolymer is the density of the PAAm polymer chain. Calculated M̄c values revealed that the amine side chain of PAAm was involved in interfacial binding. Both the catechol-functionalized PAAm contained approximately 3 catechols per polymer chain (Table 1), which would yield M̄c values of around 300 kDa (i.e., M̄w of polymer divided by the number of catechols) if only these adhesive moieties were involved forming crosslinking points with Laponite. Given that the calculated M̄c values were 5–10 times lower for PAAm-D and PAAm-ND, the strong interfacial binding properties of catechol likely faciliated PAAm-Laponite interactions, potentially by enhancing the proximity between the polymer network and the nanoparticles. Finally, data reported here further confirmed that nitro-functionalization greatly enhanced the interfacial binding strength of the catechol. Equation (4) assumes an ideal network consisting only of effective network chains. As such, the reported M̄c values are likely to be overestimated.[62]</p><!><p>Up to this point, testing was performed using nanocomposite formulated with a starting polymer:Laponite wt%:wt% ratio of 60:40 (i.e., 40 wt% Laponite). The effect of increasing the starting Laponite content on the viscoelastic properties of these nanocomposite hydrogels was also investigated (Figure S5, S6 and S7, Supporting Information). Both G′ and G″ values for PAAm-ND and PAAm-D showed little or no increase when the starting Laponite content increased from 40 wt% to 60 wt% (Figure 8). Shear moduli values did not decrease, even though there was a reduction in the catechol content with increasing nanoparticle wt%. On the other hand, there was a drastic increase in the measured moduli values for PAAm with increasing Laponite content (43 and 55 times increase for G′ and G″, respectively). Both shear moduli values for PAAm nanocomposite approached those of PAAm-ND and PAAm-D at the highest Laponite content tested. These results indicated that, while catechol was responsible for strong wet adhesion with the nanoparticles, the primary amine side chain of PAAm also contributed to the interaction with Laponite.[46,47]</p><p>Taken together, this report demonstrates that the viscoelasticity property of a nanocomposite hydrogel is highly dependent on the strength of the interaction between the polymer matrix and the nanoparticles. When no catechol moieties were immobilized onto PAAm chains, the interaction between the polymer matrix and nanoparticles was dominated by weak physical bonds (i.e., hydrogen bonding, electrostatic interactions etc.). Although these weak interactions supported network formation, the weakly associated networks demonstrated reduced material properties, especially when the nanoparticle content was low. When a small amount of catechol was introduced, nanocomposite hydrogels exibited a significant increase in measured shear moduli.</p><p>Laponite was chosen in the current study as its strong, reversible interaction with network-bound dopamine has previously been reported.[19,24] Additionally, Laponite is known to be biocompatible and support cellular attachment and proliferation,[67,68] which will aid in the future design of mechanically strong biomaterials. The exact molecular interaction between catechol and Laponite is still unknown. Catechol can potentially form hydrogen bonds with silica oxide.[69] However, hydrogen bonding alone cannot account for the drastic increase in the measured storage and loss moduli (an increase of 1–2 orders of magnitude) when an extremely small amount of the catechol group (≈0.1 wt%) was introduced into the polymer network, especially given the fact that acrylamide side chains are capable of hydrogen bond formation. From density functional theory analysis, the binding energy of catechol with a silica oxide surface was estimated to be 33 kcal mol−1, which is significantly higher than that of hydrogen bonds (binding energy = 14 kcal mol−1).[70,71] The theoretical catechol–SiO2 binding energy value approaches the experimental and theoretical values of catechol bound to a titanium oxide surface.[18,72] As such, the combination of hydrogen bonding and the dispersion interaction of the phenylene ring are likely to be involved in the interfacial binding of catechol and Laponite.[71] A more comprehensive study will be required to verify the specific nature of this interaction.</p><p>Both EWG modification and pH changes were performed to further modulate the interfacial binding strength between the network-bound catechol and the nanoparticles. The results confirmed that −NO2 group modification greatly enhanced the binding strength of catechol towards Laponite, which resulted in the formation of a more densely crosslinked network with enhanced mechanical properties. The presence of nitrodopamine contributed to forming multifunctional crosslinking points on Laponite, likely through facilitating nanoparticle interactions with PAAm chains. Finally, the material properties of both dopamine- and nitrodopamine-functionalized nanocomposites were highly dependent on the pH and the oxidation state of these catechol moieties. Although measured shear moduli decreased with increasing pH as a result of catechol oxidation, these materials still exhibited moduli values that were 1–2 orders of magnitude higher than catechol-free control. Nitrodopamine was also less prone to oxidation and exhibited stronger binding to Laponite when compared to dopamine when tested above the dissociation constants of these adhesive moieties. Recently, polymer-bound nitrodopamine was demonstrated to be susceptible to light-mediated degradation,[73] which may represent an additional opportunity to tune the material properties of these nanocomposite materials.</p><!><p>PAAm nanocomposite hydrogels were formulated with Laponite nanosilicate and network bound mussel-mimetic catechol adhesive. The catechol exhibited a strong affinity toward Laponite, as a small amount of catechol incorporation (ca. 0.1 wt% in the swollen hydrogel) resulted in a drastic enhancement in material properties and the crosslinking density of the nanocomposite hydrogels. The catechol group was further modified with an electron withdrawing nitro-functional group in the form of nitrodopamine, which further enhanced the measured shear moduli across the whole pH range tested. Although, the moduli decreased under basic conditions as a result of catechol oxidation, PAAm-ND still exhibited significantly higher moduli values when compared to those of PAAm-D, potentially due to the increased resistant to oxidation of the nitrodopamine. This paper confirms that strong interfacial binding between the polymer matrix and encapsulated nanoparticles is critical to the fabrication of nanocomposite hydrogels with improved material properties.</p>

PubMed Author Manuscript

CRISPR/Cas9 ablating viral microRNA promotes lytic reactivation of Kaposi\xe2\x80\x99s sarcoma-associated herpesvirus

The CRISPR (clustered regularly interspaced short palindromic repeats)/Cas9 (CRISPR-associated gene 9) system is an RNA-guided, DNA editing method that has been widely used for gene editing, including human viruses. Kaposi\xe2\x80\x99s sarcoma-associated herpesvirus (KSHV/HHV8), following latent infection in human cells, can cause a variety of malignancies, such as Kaposi\xe2\x80\x99s sarcoma (KS), primary effusion lymphoma (PEL), and multicentric Castleman disease (MCD), with a high prevalence in immunocompromised patients. Of significant concern, the latent infection with KSHV has been shown to lead to increased resistance to antiviral therapies. MicroRNAs (miRNAs) are a set of non-coding, small RNA molecules that regulate protein-coding genes at the post-transcriptional and translational levels. KSHV has its miRNAs, most of which are expressed in latently infected cells and play a crucial role in maintaining KSHV latency. Notably, by regulating the expression of the downstream target genes in host cells, KSHV miRNAs can interact with the host environment to promote the development of KSHV-related diseases. Although CRISPR/Cas9 has been reported to edit KSHV protein-coding genes, there is no published literature on whether the CRISPR/Cas9 system can regulate the expression of KSHV miRNAs. In this study, we used CRISPR/Cas9 to inhibit the expression of KSHV miRNAs by directly editing the DNA sequences of individual KSHV miRNAs, or the promoter of clustered KHSV miRNAs, in latent KSHV-infected PEL cells. Our results show that CRISPR/Cas9 can ablate KSHV miRNAs expression, which in turn leads to the upregulation of viral lytic genes and alteration of host cellular gene expression. To the best of our knowledge, our study is the first reported demonstration of the CRISPR/Cas9 system editing KSHV miRNAs, further expanding the application of CRISPR/Cas9 as a novel antiviral strategy targeting KSHV latency.

crispr/cas9_ablating_viral_microrna_promotes_lytic_reactivation_of_kaposi\xe2\x80\x99s_sarcoma-assoc

Introduction<!>Cell culture<!>qRT-PCR analysis<!>Plasmids construction and transfection<!>DNA isolation and PCR assay<!>Western blot<!>Immunofluorescence assay<!>Oxygen consumption rate (OCR) and spare respiratory capacity (SRC)<!>CRISPR/Cas9 can target KSHV miRNAs by editing the KSHV genome in host PEL cells<!>CRISPR/Cas9 editing individual KSHV miRNAs in host PEL cells<!>CRISPR/Cas9 editing KSHV miRNAs can affect host cell genes expression and physiological processes<!>Discussion

<p>Kaposi's sarcoma-associated herpesvirus (KSHV), also known as human herpesvirus-8 (HHV8), has been documented as a causative agent for Kaposi's sarcoma (KS), primary effusion lymphoma (PEL), and multicentric Castleman disease (MCD), which are the malignancies prevalent in immunocompromised populations [1]. Like other herpesviruses, KSHV has two phases in its life cycle: the latent and lytic phases. When the latent infection is established in host cells, KSHV expresses only a few essential latent genes to maintain the viral extrachromosomal episomes without producing infectious virions, which allow KSHV to escape from the host's immune surveillance and remain inactive throughout its life status [2]. Notably, many antiviral drugs only effectively target lytic KSHV, but cannot specifically eliminate latent KSHV [3].</p><p>Recent studies have revealed that microRNAs (miRNAs) are a group of non-coding small RNA molecules that play a significant role in the viral life cycle [4]. KSHV expresses 25 mature miRNAs, encoded by 12 pre-miRNA genes, which are pivotal in developing KSHV-related pathogenesis and tumorigenesis [4-6]. Most of KSHV miRNAs are expressed during the latent phase; however, they can be encapsulated into the virions when KSHV is reactivated and switches to the lytic phase [4]. Previous studies already extensively reported the critical role of viral miRNAs in maintaining KSHV latency; however, the regulatory function of these miRNAs in KSHV reactivation and switching to the lytic phase has not been well-studied.</p><p>The clustered regularly interspaced short palindromic repeats/CRISPR-associated gene 9 (CRISPR/Cas9) system is an RNA-guided DNA editing method that has been widely used in gene editing [7]. Over the past five years, the application of CRISPR/Cas9 as a robust and convenient gene-editing tool has rapidly expanded to almost all biomedical fields, including the study of human viruses [8-11]. A recent study reported that CRISPR/Cas9 could disrupt KSHV latency by targeting latency-associated nuclear antigen (LANA) and ORF57 in KSHV-infected cells [12,13]. However, to the best of our knowledge, there are no reports of studying KSHV miRNAs with CRISPR/Cas9.</p><p>Our previous study has demonstrated the capacity and efficacy of CRISPR/Cas9 to edit human miRNAs in both in vitro and in vivo models [14]. In this study, we utilized CRISPR/Cas9 to target KSHV miRNAs in BCBL-1 and BCP-1, two primary effusion lymphoma (PEL) cell lines with latent KSHV infection. After editing the promoter of the KSHV-miRNA cluster with CRISPR/Cas9, we found ten individual KSHV miRNAs in the cluster were ultimately downregulated. We further targeted two individual KSHV miRNAs, miR-K12-1 and miR-K12-9, for CRISPR/Cas9 editing. Our data show that not only their expression was downregulated, but also their downstream target genes were altered accordingly, demonstrating that CRISPR/Cas9 is a robust and precise tool for modulating KSHV miRNAs. Of significance, the use of CRISPR/Cas9 to target viral miRNAs led to the lytic reactivation of KSHV, resulting from the upregulation of lytic genes, along with metabolic alterations and increased oxygen consumption in host PEL cells. Thus, our results provide novel insights into the development of precision medicine with CRISPR/Cas9 editing viral miRNAs to eliminate KSHV latent infection.</p><!><p>Body cavity-based lymphoma cells (BCBL-1) were maintained in RPMI 1640 medium (Gibco) with supplements as previously described [15]. The cell lines BCP-1 were purchased from American Type Culture Collection (ATCC) and maintained in complete RPMI 1640 culture medium (Gibco) supplemented with 20% FBS.</p><!><p>The premier sequences for stem-loop reverse transcription (RT) and quantitative real-time PCR (qRT-PCR) to examining mature KSHV miRNAs were referred to a previous report [16]. The relative quantitation of target genes was calculated using the 2-ΔΔCT method, as described previously [14]. The primer sequences are included in Supplementary Table S1.</p><!><p>Plasmid lenti-CRISPR was purchased from Addgene (#49535). Protospacer sequences of CRISPR/Cas9 targeting the promoter of KSHV-miRNA cluster or individual KSHV miRNAs were designed using CRISPR DESIGN (http://crispr.mit.edu/). CRISPR/Cas9 was transiently transfected into cells using electroporation, followed by clonal selection with puromycin (5 μg/mL) and then cell expansion in the conventional medium. Primer sequences are included in Supplementary Table S1.</p><!><p>DNA was extracted from the cells and KSHV using a Wizard Genomic DNA Purification Kit (Promega, A1120). DNA fragments were amplified using GoTaq DNA Polymerase (Promega). The purified PCR products were ligated to pGEM-T Easy vectors (Promega, A1360) for DNA sequencing.</p><!><p>Western blot was performed using our protocol published previously [14]. The primary antibodies, anti-CDKN1A (Cell Signaling Technology, #2946), β-Actin (Cell Signaling Technology, # 3700S), were diluted at 1:1000.</p><!><p>The immunofluorescence staining of NF-κB/P65 was performed using our protocol published previously [17]. The primary antibody NF-κB/P65 (BD Biosciences, #610868) was diluted at 1:200. The secondary fluorescent antibody (Thermo Fisher, # A-11002) was diluted at 1:400. DAPI (Thermo Fisher, #62248) was used at a concentration of 1 μg/ml. The images were taken using a confocal microscope (Nikon Eclipse Ti2, Tokyo, Japan).</p><!><p>OCR and SRC were analyzed using an Agilent Seahorse XFe24 Analyzer. The seahorse plate was pre-coated with a Cell Tak (Corning, #354240) for 20 min at room temperature, and a total of 200,000 cells were seeded and tested following the manufacturer instructions of Seahorse XF Cell Mito Stress Test Kit (Agilent, #103015–100).</p><!><p>The KSHV genome contains 12 viral miRNAs genes, as shown in Supplementary Fig. S1A. MiR-K12-1 to −9 and miR-K12-11 are located in the intron of KSHV Kaposin (Open Reading Frame K12) gene and are transcribed as a cluster [4,5,18]. Apparently, the clustered KSHV miRNAs share the same promoter [19]. To edit the KSHV episomal DNA with CRISPR/Cas9, we designed two single guide RNAs (sgRNAs), sgRNA-1, and sgRNA-2, to guide this process in the promoter of the KSHV-miRNA cluster. Supplementary Fig. S1B illustrates the design of the sgRNAs. The PCR result shows a deletion of 200 bps in the promoter after CRISPR/Cas9 editing (Supplementary Fig. S1C), which was further confirmed with DNA sequencing (Supplementary Fig. S1D). When examining the expression of the clustered KSHV miRNAs using qRT-PCR, we found that the promoter editing with CRISPR/Cas9 could robustly downregulate individual mature KSHV miRNAs in BCBL-1 and BCP-1 cells (Fig. 1A & B). We further analyzed the proliferation of host PEL cells and determined that the KSHV miRNA promoter-deficient cells (BCBL-Pro-del and BCP-1-Pro-del) grew more slowly than the vector control cells (sg-Ctrl, Fig. 1C). Of significance, when examining the selected KSHV genes associated with the latent and lytic phases, we found that KSHV lytic genes, RTA, vGPCR, K8.1, and ORF57 were significantly elevated, while the latent gene, LANA remained steady (Fig. 1D). Given the previous study reporting that KSHV miRNAs play a vital role in maintaining KSHV latency [20], our results support the notion that genomic editing of viral miRNA with CRISPR/Cas9 can lead to the lytic reactivation of KSHV.</p><!><p>Since the clustered KSHV miRNAs can be edited by CRISPR/Cas9 targeting their shared promoter, we further investigated whether CRISPR/Cas9 could also precisely target selected individual KSHV miRNAs. We designed the sgRNAs to specifically target miR-K12-1 and miR-K12-9, which are located at the ends of the KSHV miR-K12-1 to 9-11 cluster, as shown in Supplementary Fig. S2A. PAM sequences and the design of sgRNAs are illustrated in Supplementary Fig. S2B. Fig. 2A shows that the miR-K12-1-sgRNA solely targets miR-K12-1 and dramatically downregulates the expression of mature miR-K12-1 in BCBL-1 and BCP-1 cells. Notably, the expression of mature miR-K12-9 was not significantly disrupted. We also examined the miR-K12-9-sgRNA and found that it only targeted mature miR-K12-9 but not mature miR-K12-1 in BCBL-1 and BCP-1 cells. To further demonstrate the precise editing of CRISPR/Cas9, we examined all other mature KSHV miRNAs in this cluster in BCBL-1 and BCP-1 cells, but no off-targeting effect was observed (Supplementary Fig. S2C-S2F). These data provide lines of evidence in support of the accuracy of CRISPR/Cas9 editing. In addition, we tested the expression of KSHV latent and lytic genes and found that the lytic genes, RTA, vGPCR, K8.1, and ORF57, were all significantly upregulated in both KSHV miR-K12-1 and miR-K12-9 knockdown BCBL-1 and BCP-1 cells, but the latent gene, LANA remained steady (Fig. 2C and D). These results support the effectiveness and accuracy of CRISPR/Cas9 in targeting and editing viral miRNAs in KSHV-infected cells.</p><!><p>Since the previous study reported that KSHV miRNAs played a central role in maintaining latent KSHV infection by regulating host cell gene expression [20], we further examined the expression of their downstream target genes. In particular, we examined the target genes of KSHV miR-K12-1, CDKN1A (p21) and NFKBIA (IKBα) in host PEL cells. As shown in Fig. 3A, the mRNA expression levels of CDKN1A and NFKBIA were significantly increased in BCBL-1-sg-miR-K12-1 and BCP-1-sg-miR-K12-1 cells, which were consistent with the previous reports [16,21]. We further determined the protein expression of CDKN1A in BCBL-1 and BCP-1 cells by Western blot. As shown in Fig. 3B, CDKN1A was significantly induced in BCBL-1-sg-miR-K12-1 and BCP-1-sg-miR-K12-1 cells compared to the vector control cells (sg-Ctrl). Given NFKBIA that is an inhibitor of the NF-κB signaling pathway, KSHV miR-K12-1 was reported to directly target and downregulate the expression of NFKBIA to enhance the NF-κB activity in host cells [16]. Utilizing the immunofluorescence assay, we examined the distribution of NF-κB/P65 in BCBL-1 cells and showed that BCBL-1-sg-miR-K12-1 cells had more NF-κB/P65 in the cytoplasm than the vector control cells (sg-Ctrl; Fig. 3C). Thus, our results demonstrate that the NF-κB activity in host PEL cells can be modulated through CRISPR/Cas9 editing KSHV miR-K12-1 to upregulate NFKBIA. We also examined selected downstream target genes of individual KSHV miRNAs that have been reported previously (Supplementary Table S2), and our results show the ultimate upregulation of these target genes in the promoter-deficient BCBL-1 cells after CRISPR/Cas9 editing (Pro-del; Supplementary Fig. S3).</p><p>KSHV miRNAs were reported to be involved in the metabolism of host cells [12]. Of particular interest, we measured the oxygen consumption rate (OCR) using the Seahorse XFe24, given that KSHV miRNAs have been previously reported to reduce the OCR of host cells [18]. As shown in Fig. 4A, OCR was significantly increased in promoter-deficient BCBL-1 cells after CRISPR/Cas9 editing (Pro-del), compared to the vector control cells (sg-Ctrl). We further calculated the spare respiratory capacity (SRC), which can indicate the capacity of cells to respond to an energetic demand and how closely a cell respires to its theoretical maximum. As shown in Fig. 4B, SRC is positively correlated with OCR in both CRISPR/Cas9 edited (Pro-del) and the vector control cells (sg-Ctrl). Altogether, these data support the notion that CRISPR/Cas9 targeting KSHV miRNAs can efficiently induce the expression of their target genes and alter the physiological process in host cells.</p><!><p>The CRISPR/Cas9 system is an innovative and robust tool for precise genome editing, with added utility with human DNA or RNA viruses and their DNA intermediate in their life cycle [8-11]. It has been reported that CRISPR/Cas9 can edit two KSHV protein-coding genes (LANA and ORF57) in KSHV-infected cells [12,13]. These studies demonstrated that KSHV-targeted CRISPR/Cas9 could be utilized for KSHV gene editing, significantly reducing the KSHV episomal burden over time [12,13]. However, there is no report to date on whether the CRISPR/Cas9 system can regulate the expression of KSHV miRNAs. KSHV can produce 25 mature miRNAs, which are involved in the process of the KSHV entry, latency, host cells survival, immune response, and KSHV-associated tumor development [4,5]. Understanding the exact role and function of these viral miRNAs remains a challenging area of research. Currently, most studies are still utilizing mimics/sponges to upregulate or downregulate individual or clustered KSHV miRNAs to determine their functions. However, these traditional approaches are limited to the short-term cellular dynamics and cannot simulate the entire life cycle of KSHV in host cells [4]. Editing KSHV episome DNA appears to be a more effective strategy for altering the individual or clustered viral miRNAs.</p><p>Our group previously reported that the CRISPR/Cas9 system could robustly edit human miRNAs expression in colon cancer cells, in both in vitro and in vivo models [14]. In this study, we further demonstrate that the CRISPR/Cas9 system has the capacity to edit KSHV episomal DNA in two KSHV-infected PEL cells, BCBL-1 and BCP-1 cells. First, we edited the promoter of the KSHV-miRNA cluster with CRISPR/Cas9, and our results showed the significant downregulation of these viral miRNAs. Of significance, the host PEL cells with CRISPR/Cas9 editing grew more slowly than the control cells, supporting the potential utility of CRISPR/Cas9 in treating KSHV-associated malignancies. We further examined the specificity of CRISPR/Cas9 by targeting two individual viral miRNA, miR-K12-1 and miR-K12-9, in the KSHV-infected PEL cells. Notably, we did not see any significant off-target effects with the two specific sgRNAs. Our results support the robustness and accuracy of CRISPR/Cas9 in the editing of KSHV miRNAs.</p><p>Since KHSV miRNAs are known to play a critical role in maintaining viral latency, we found that KSHV lytic genes (RTA, vGPCR, K8.1, and ORF57), but not the latent gene LANA, were significantly upregulated in KSHV-infected PEL cells with ablated KSHV miRNAs after CRISPR/Cas9 editing. The upregulation of lytic genes indicates the lytic reactivation of KSHV. We further evaluated the expression of selected downstream target genes of KSHV miRNAs and determined the physiological alterations in host PEL cells after CRISPR/Cas9 editing. CDKN1A and NFKBIA are the targets of KSHV miR-K12-1, and they are known to play tumor-suppressive roles in host cells. CDKN1A is a cyclin-dependent kinase inhibitor (CKI) and is the downstream target of P53 that controls cell cycle progression [22]. NFKBIA is a cytosolic protein that functions to mask the nuclear localization signals (NLS) of NF-κB protein and sequester it in the cytoplasm [23]. Our results demonstrate that CDKN1A and NFKBIA are ultimately upregulated in KSHV-infected PEL cells after CRISPR/Cas9 editing. We furthered our study and examined the translocation of NF-κB/P65 in the host cells. As expected, most of the NF-κB/P65 signals appeared in the cytoplasm in KSHV-infected PEL cells edited by CRISPR/Cas9 to target KSHV miR-K12-1, compared to the vector control cells. We also investigated the metabolic changes of host PEL cells after CRISPR/Cas9 editing KSHV miRNAs, specifically OCR and SRC, measured their oxygen consumption and respiratory capacity. Our results show that OCR and SRC were dramatically increased in CRISPR/Cas9-edited PEL cells. These results support the notion that altered expression of clustered KSHV miRNAs expression by CRISPR/Cas9 can increase host cell metabolism, suggesting that more energy is produced for lytic reactivation.</p><p>Therapeutic manipulation of viral miRNAs has been utilized in antiviral therapy, with some using the specific antagomir of miR-122 as a treatment method for patients infected with the hepatitis C virus (HCV) [24]. Given the distinct sequences of KSHV miRNAs from the human genome, targeting viral miRNAs should be a feasible approach to develop safe and robust antiviral strategies. Recently, the systematic safety and feasibility of CRISPR/Cas9 have been further demonstrated in patients with refractory cancer [25], making it a potentially promising antiviral treatment option in the near future. In this study, we utilized CRISPR/Cas9 to edit individual or clustered KSHV miRNAs in host PEL cells. Our results demonstrate that CRISPR/Cas9 editing can alter the expression of mature KSHV miRNAs, leading to the upregulation of their downstream target genes and the viral lytic genes, which are associated with the psychological changes to facilitate the lytic reactivation of KSHV in host cells. Therefore, our study supports the premise that CRISPR/Cas9 is a robust and precise tool to edit KSHV miRNAs and can be developed as a potent antiviral therapeutic strategy to impair viral replication.</p>

PubMed Author Manuscript

L-edge sum rule analysis on 3d transition metal sites: from d10 to d0 and towards application to extremely dilute metallo-enzymes

According to L-edge sum rules, the number of 3d vacancies at a transition metal site is directly proportional to the integrated intensity of the L-edge X-ray absorption spectrum (XAS) for the corresponding metal complex. In this study, the numbers of 3d holes are characterized quantitatively or semi-quantitatively for a series of manganese (Mn) and nickel (Ni) complexes, including the electron configurations 3d10\xe2\x86\x923d0. In addition, extremely dilute (<0.1% wt./wt.) Ni enzymes were examined with two different approaches: 1) by using a high resolution superconducting tunnel junction (STJ) X-ray detector to obtain XAS spectra with very high signal-to-noise ratio, especially in the non-variant edge jump region; and 2) by adding an inert tracer to the sample that provides a prominent spectral feature to replace the weak edge jump for intensity normalization. In this publication, we present for the first time: 1) L-edge sum rule analysis for a series of Mn and Ni complexes that include electron configurations from an open shell 3d0 to a closed shell 3d10; 2) a systematic analysis on the uncertainties, especially on that from the edge jump, which was missing in all previous reports; 3) a clearly-resolved edge jump between the pre-L3 and the post-L2 regions from an extremely dilute sample; 4) an evaluation of an alternative normalization standard for L-edge sum rule analysis. XAS from two copper (Cu) proteins measured with a conventional semiconductor X-ray detector are also repeated as bridges between the Ni complexes and the dilute Ni enzymes. The differences between measuring 1% Cu enzymes and measuring < 0.1% Ni enzymes are compared and discussed. This study extends L-edge sum rule analysis to virtually any 3d metal complex and any dilute biological samples that contain 3d metals.

l-edge_sum_rule_analysis_on_3d_transition_metal_sites:_from_d10_to_d0_and_towards_application_to_ext

1. Introduction<!>Samples<!>XAS Measurements<!>3. Results and discussions<!>XAS Spectra of Ni and Mn Complexes<!>L-edge sum rule<!>Error Analysis<!>Dilute Samples<!>Measuring <0.1% Ni Enzymes Using A STJ detector<!>Measuring < 0.1% Ni Enzymes with Alternative Standard<!>Comparison with K-edge XAS<!>4. Summary

<p>The oxidation state is an indicator of the degree of oxidation for a particular atom and is one of the most-pursued quantities in chemistry, because the distribution of electron/charge density in inorganic complexes or enzymatic metal sites determines their chemical, physical and biological functions/properties. Unlike alkali or alkaline earth metals, transition metals can have different oxidation states, and thus different chemical properties. For example, manganese (Mn) complexes1, 2 can have possible MnII to MnVII including electron configurations from 3d5 to 3d0, nickel (Ni) complexes3, 4 can have Ni0 to NiIV (from 3d10 to 3d6), while copper complexes can have CuI and CuII sites. In inorganic and bioinorganic chemistry, resolved oxidation states, i.e. the measured number of electrons/holes localized in the bonding orbital, have helped understand the chemical and biochemical roles of many metal sites,4–6 while unresolved oxidation states have contributed to longstanding controversies in many systems.3, 7–11 The oxidation state is typically characterized by an integer, which is the hypothetical charge in an atom assuming the bonding is 100% ionic. Since a real 100% ionicity does not exist, and there is ambiguity of the assignment of the oxidation state of the electronegative ligands, metal oxidation states may be represented more quantitatively by the number of vacancies (holes) localized at the metal sites (e.g. Mn or Ni).4</p><p>X-ray absorption spectroscopy (XAS), especially L-edge XAS (or L XAS), is one of the best methods to investigate the oxidation states of 3d transition metals via absorption edge positions as well as their spectral features.12,4, 13 In an L XAS, the L3 edge's centroid energies, the branching ratios of IL3/(IL2+IL3), and the rich multiplet structures are all sensitive to electronic structures.4, 14–16 For example, L XAS exhibit about 2 eV per oxidation state change (eV/oxi) for the Mn complexes and 0.9 eV/oxi for Ni complexes. L-edge centroids will often be sufficient to assign oxidation states.4,34–36 For Mn complexes, the branching ratio has also been used to determine the Mn oxidation states.1, 17 Nevertheless, the shifts in L3 centroids are also affected by the changes in the final state in addition to the differences in the ground states. The spectral multiplets and branching ratios are also sensitive to metal's electronic spin states and its coordination geometries, in addition to their oxidation states.</p><p>Sum rules link the integrated XAS absorption intensity to the number of holes localized in the X-ray absorbing metal.4, 18 For 3d metals, L-edge XAS probes electronic transitions at 2p→3d, 2p→4s, and 2p→continuum, as shown in Fig. 1 (middle panel). A typical L-edge XAS spectrum therefore has a pair of strong absorption peaks corresponding to 2p3/2→3d and 2p1/2→3d transitions (2p→4s is 20-fold weaker) and an invariant edge jump step between the pre-L3 and post-L2 regions corresponding to 2p→continuum transitions. Therefore the total number of 3d holes localized in the X-ray absorber is proportional to the integrated L2,3-edge peaks (2p1/2, 3/2→3d) intensity when normalized to this invariant edge jump, as reported previously.4–6, 19–21 Besides being able to obtain a number of 3d holes (H3d), the integrated L intensity IL is a pure ground state property and is well-suited for investigating d-shell vacancies.4 In the past, L-edge sum rule analysis has been applied to one set of Ni4 and one set of Cu5, 6 model complexes, and a few Cu proteins with Cu concentrations of order ~1% wt./wt.5, 6 However, the error bars for their XAS (especially for the weak edge jump) were not well-discussed or controlled. In addition, L-edge sum rule analysis on Mn or other 3d metal complexes and on samples with < 0.1% wt./wt. metal concentration are not available.</p><p>In this publication, we first characterize a series of Mn and Ni complexes with L-edge sum rule analysis. We then extend this analysis to NiFe hydrogenase (H2ase) and CO-dehydrogenase (CODH), which have an extremely dilute Ni concentration of < 0.1% wt./wt. Two approaches to make this extension available are discussed. The differences between measuring 1% Cu enzymes and measuring < 0.1% Ni enzymes are also addressed. For the first time, we present: 1) L-edge sum rule analysis on a series of 3d metal complexes, which include electron configurations from an open shell 3d0 to a closed shell 3d10; 2) a systematic analysis of the error bars for L-edge XAS, especially for their edge jump regions; 3) a clearly-resolved edge jump between the pre-L3 and the post-L2 regions for an extremely dilute sample (< 0.1% Ni); and 4) a detailed evaluation on using an alternative normalization standard for L-edge sum rule analysis.</p><!><p>NiIIF2, NiIICl2, NiIIBr2, NiIIO, NiII(OH)2, MnIIO, LiMnIIIO2, Mn2IVO3, MnIVO2, and KMnVIIO4 were purchased from Sigma-Aldrich, stored in desiccator and used without further treatment. Ionic K3NiIIIF6 complex22 and LiNiO2, Ni2O3, KNiIVIO623, 24 were provided by Dr. Neil Bartlett from UC Berkeley and Dr. Melendres from Argonne National Lab respectively. Covalent Ni-S complexes Na2[Ni0(SR)4], Na[NiI(SR)4], [NiII(SR)4], where (SR)4 = bis(diphenylbis((methylthio)methyl)borate and [PhTttBu]NiICO, [PhTttBu]NiII(Cl), where PhTttBu = phenyltris((tert-butylthio)methyl)borate,25, 26 were prepared in Dr. Charles Riordan's group from the University of Delaware. All solid model complexes were finely ground and pressed onto a piece of UHV compatible carbon tape inside a nitrogen (N2) atmosphere glovebox (although not all the samples are air sensitive). Such prepared samples were loaded into the UHV measurement chamber with a vacuum loadlock.</p><p>Blue Cu protein from the construct engineered azurin and from plastocyanin were prepared5, 27–29 by Professor E.I. Solomon's laboratory at Stanford University. They were used as examples of metalloenzymes with moderate metal concentration (e.g ~1% Cu, wt./wt.). Clostridium thermoaceticum-CODH (or Ct-CODH) was purified and prepared30, 31 in Dr. P. W. Ludden's laboratory. The 310 kDa tetramer has four Ni sites, corresponding to about 770 ppm (or 0.077%) Ni concentration. The as-isolated and H2 reduced NiFe H2ase solution samples (< 0.067% wt./wt.) were prepared from D. gigas H2ase12, 32, 33 at Lawrence Berkeley National Laboratory. The two Ni enzymes were examples of extremely dilute samples with < 0.1% metal concentration. For most enzyme samples, partially dehydrated films were made by drying the solution samples on sapphire substrates under a H2 atmosphere (for the H2 reduced NiFe H2ase sample) or a N2 atmosphere (for other samples). For Ct-CODH enzyme, a frozen solution sample was prepared instead of partially dehydrated films.</p><p>The Ce M5 edge at 881.7 eV is close to Ni L3 edge at 852.6 eV and can be used as an intensity normalization standard in the Ni L-edge sum rule analysis instead of the more commonly used edge jump. To establish the method, a testing complex mixture was first prepared and evaluated. Inside the glovebox, 0.1M NiBr2-TRIS-HCl solution was prepared, providing a sample with ~ 0.6% Ni concentration (wt./wt). The Ce(NO3)3•6H2O (or Ce(NO3)3 for short) was then added to the 0.1M solution. Due to the extremely short penetration depth by soft X-rays (0.5 µm in H2O and ~200Å in typical solids), the Ni, Ce and buffer mixture must be homogenized on a microscopic scale in solution. Experimentally, the best molar ratio for NiBr2:Ce(NO3)3 was found to be ~ 1:5.5, when compatible signal intensities for Ni L3 and for Ce M5 edges were obtained in the XAS spectra. For NiFe H2ase, Ce(NO3)3 was added to the as-isolated enzyme solution. The H2-reduced solution was then prepared by incubating the as-isolated H2ase with added Ce(NO3)3 under pure H2 for >8 hours. This maintains a constant Ni:Ce ratio in the as-isolated and the reduced samples, although their absolute concentration may be different.</p><!><p>L-edge XAS was measured at Stanford Synchrotron Radiation Laboratory (SSRL) beamline 8–234 and at Advanced Light Source (ALS) beamline 9.3.235 and 4.0.236 inside a UHV chamber under windowless operation.13, 37 These beamlines have an energy resolution of 1.0 eV, 0.4 eV and 0.2 eV, respectively, at the Ni L-edge at 850 eV. For concentrated Mn and Ni complexes, the XAS spectra were measured by total electron yield (TEY)15 using a Galileo 4716 channeltron electron multiplier as photoelectron detector. The incident beam intensity (I0) was monitored through a gold-plated grid. For biological samples or dilute NiBr2 matrix samples, XAS were measured by partial X-ray fluorescence yield (PFY)12, 31 using a 30-element germanium (Ge) array detector with 180 eV energy resolution, or a 9- or 36-element superconducting tunnel junction (STJ) detector with 15 eV energy resolution.39,40 For measuring < 0.1% Ni Ct-CODH, the use of an STJ detector was necessary to extract the weak Ni signal from the high spectral background. During these measurements, one region was defined around Ni L, Cu L or Ce M partial fluorescence signal (PF), which is proportional to the metal absorbance and incident X-ray beam intensity, while the second region was set at around the oxygen (O) K fluorescence signal used as a measure of I0.</p><p>Superconducting tunnel junctions (STJs)38–41 are high-resolution X-ray detectors that consist of two superconducting electrodes separated by an extremely thin layer of insulating material. Electric current can pass through the STJ junction via the quantum-mechanical tunneling effect. X-rays absorbed in one of the electrodes will generate free excess charges in proportion to the X-ray energy, and the subsequent increase in tunneling current can be directly read out with a room temperature amplifier. STJ X-ray detectors exploit the extremely small (~1 meV) energy gap in superconductors to achieve an order of magnitude higher energy resolution than conventional semiconductor detectors, which have an energy gap of ~1 eV. In general, our Nb-based STJ detector has a 10–20 eV energy resolution39, 42, 43 while a semiconductor detector has 100–300 eV38, 44, 45. The higher energy resolution allows a more effective rejection of unwanted background counts and increases the signal-to-noise ratio (S/N) for the L- and M-edge PFY XAS spectra for dilute samples.39, 42, 43 This is especially true for resolving the weak features, such as the edge jump steps.</p><p>For the measurements at SSRL, each concentrated Ni complex spectrum was the sum of 5–6 raw scans, each blue Cu protein spectrum was the sum of 20 scans, while each NiFe H2ase spectrum represented the sum of 40 raw scans. The Ni complexes were measured at room temperature, while the blue Cu enzymes and NiFe H2ases were measured at 10K maintained with a liquid helium flow cryostat. At ALS, each Mn complex spectrum was the sum of 3–5 raw scans, each dilute NiBr2 spectrum was the sum of 6 raw scans, while each Ct-CODH spectrum represented the sum of 10 raw scans. The energies were calibrated with MnO at 638.7 eV, NiF2 at 852.7 eV, and CuO at 932.7 eV for the respective samples.46 As the beam intensity is much stronger at ALS in comparison with that at SSRL, all the spectra were recorded at 10K (using a LHe cryostat). To further minimize the possible radiation damage and photoreduction, the position of the X-ray beam on the sample was moved for every scan. We also tested multiple short scans at one spot for all the samples at the ALS and at SSRL, but observed no spectral change under our experimental conditions.</p><p>The L-edge data processing for complex samples involves subtraction of the spectral background, using their sample holders XAS as reference (as will be addressed in detail in the Error Analysis section). These L XAS spectra were then normalized to the invariant edge jump between the pre-L3 and the post-L2 regions to produce the integrated L-edge spectra. The non-resonant edge jumps were then removed by subtracting a simulated two-step function. The L3 and L2 intensities (IL3 and IL2) were obtained from these spectra by integration over 929–936 and 950–955 eV for Cu proteins, over 851–858 and 868–875 eV for Ni samples, and over 634.2–657.6 eV for Mn complexes. For STJ measured Ct-CODH XAS, a process similar to the one for processing Ni complex data was used to calculate IL3 and IL2. For the NiFe H2ase and the dilute NiBr2 XAS, the integrated Ce M5-edge intensity was used as an alternative intensity normalization standard, because these spectra do not exhibit an observable edge jump.</p><p>K-edge XAS for MnO, Mn2O3 and KMnO4 were measured at BL08B2 of the SPring-8 synchrotron radiation facility in Hyogo Prefecture, Japan. The main X-ray optical elements consist of a first Rh-coated vertical collimation mirror, a double-crystal monochromator and a second Rh-coated vertical refocusing mirror. A water-cooled Si (111) double-crystal monochromator was used to produce a ~1 eV bandwidth X-ray beam with a beam size of 2.0 mm in the horizontal direction and 0.5 mm in the vertical direction. The K-edge XAS data were measured in transmission mode over an energy range from −330 to 1500 eV with respect to the Mn absorption edge by using ionization chambers to record the incoming and transmitted intensities. The transmission measurement is possible because Mn K-edge has a transmission depth of ~1 mm. The Mn K absorption edge was defined as the first peak in the first derivative spectrum of XAS data. A Mn foil with an XAS absorption edge at 6539 eV was used as the energy calibration standard for the monochromators.</p><!><p>Five published Ni XAS spectra4 were either re-measured (on NiF2, K3NiF6 and KNiIO6) or cited (for [NiI/II(SR)4]−/0) and compared with the XAS for five new Ni and five additional Mn complexes. The errors for all fifteen L-edge XAS spectra were evaluated or re-evaluated with the new method.</p><!><p>Examples of the observed Ni and Mn L-edge spectra are shown in Fig. 1, with Ni complexes from d10 to d6 in the left panel and Mn complexes from d5 to d0 in the right panel. Since [Ni0(SR)4]2− has a closed 3d10 shell, its 2p→3d resonance intensity should be absent, and this is indeed the case (Fig 1a, dashed red line). The L-edge XAS spectrum is then dominated by the 2p→continuum transition, also called the edge jump. This edge jump does not change with the number of 3d holes and can therefore be used to normalize XAS spectra. All spectra in Fig. 1 were normalized to their corresponding edge jumps between the pre-L3 and post-L2 regions. Although [Ni0(SR)4]2− may not be an interesting compound in synthetic chemistry, it did provide us with the first L-edge XAS for a real d10 sample, which is rare. We emphasize that although Ni metal is also nominally Ni0, it has an electronic configuration of 3d84s2 rather than 3d10, and a typical NiII L-XAS feature as shown in Fig. S1 in the supplemental information (SI).</p><p>In contrast, [NiI(SR)4]− has a 3d9 configuration, opening up one d hole for obvious 2p3/2→3d and 2p1/2→3d resonant transitions (Fig. 1a, solid black line). As the oxidation state further increases to NiII, NiIII, and NiIV, the number of d-holes increases gradually, and so does the L-edge resonance intensity (Fig. 1a→1d).4</p><p>In the right panel, although the d5 MnO2 has the smallest intensity among the Mn complexes, it still has higher L3 and L2 intensities than the highest Ni intensity (for KNiIVIO6). As the Mn oxidation state continues to increase (d5→d4→d3→d0 in Fig. 1e→1h), the L-edge resonance intensity also increases significantly as expected. In addition to the integrated intensities, these Ni and Mn spectra contain additional information, such as L3 centroid energies, branching ratios of IL3/(IL2+IL3), and multiplet structures.4, 14–16 Nevertheless, as mentioned in the introduction, only integrated L intensity is a pure ground state property.</p><!><p>With the L-edge XAS spectra normalized to the invariant edge jump, the integrated L-edge absorption intensities IL for our Ni and Mn complexes are listed as in Table-1 and shown as in Fig. 2a. For example, we observe 0 for [Ni0(SR)4]2− (d10), 14.7 for NiIIO (d8), 29.4 for KNiIVIO6 (d6), 36.8 for MnIIO (d5), 40.9 for MnIII2O3 (d4), and 59.6 for KMnVIIO4 (d0). Here we did not attempt to evaluate the trace amount "resonant" intensity for [Ni0(SR)4]2− but simply set it to zero. There is a roughly linear dependence between the integrated L-edge intensity (IL) and the nominal number of 3d holes (N3d) as expected. The slope, which we call R, shows an averaged normalized intensity per nominal 3d hole of about R = 6.9 in Fig. 2a. Each individual data point has an errorbar, which will be discussed in the next section.</p><p>In Fig. 2a and 2b, some additional trends are worth noticing: 1) S and N based covalent Ni complexes have obviously lower R values in comparison with O and F based ionic complexes, because the S, N have lower ionicity than the more electronegative O and F; 2) For the complexes with the same ligand donor (e.g. O), the L intensity per nominal 3d hole (R value) decreases as the number of nominal 3d holes increases. This again is consistent with the lower expected ionicity for higher-valence complexes; 3) Comparison of F vs. Cl and Br complexes show RF > RCl > RBr as well (Table-1). This is caused by different negativities of the different ligands47.</p><p>To convert the measured IL to the numbers of actual 3d holes (H3d, not nominal N3d) on a calibrated absolute scale, we use standard samples for which the real numbers of 3d holes (H3d) and their L XAS intensities (IL) are accurately known4. In this study, we have: 1) Ni metal, which has an IL = 13.1±0.7 with a band structure calculated H3d = 1.5±0.1 per Ni atom16,48,49; 2) NiO, for which we measure IL = 14.7, with an estimated Ni H3d = 1.72 from references.50, 51 With these two standards, the integrated L-edge intensity per 3d hole for the calibration samples is obtained as Rcal = 8.6. The average ionicity for our Ni and Mn complexes is therefore R/Rcal = 6.9/8.6 ≈ 80%.</p><!><p>If IL represents the L-edge peak integral and J represents the edge jump height, we can define the normalized integrated L intensity as α=IL/J. Then the error for this normalized intensity is given by: (1)δα/α=[(δIL/IL)2+(δJ/J)2]1/2or (2)δα=(IL/J)[(δIL/IL)2+(δJ/J)2]1/2Note that both δIL/IL and δJ/J contribute to δα. The δIL is primarily from the integral's statistical error and the uncertainty in the choice of the integration range, while δJ is due to the uncertainty in judging the height of the edge jump. As shown in the Table-1, δIL/IL varies between ±2.5% and ±5.0%. The relatively small uncertainty δIL is consistent with the fact that IL is an integrated value, which has an averaging effect. On the other hand, δJ/J is larger and varies between ±9.0 and ±11.8%. As illustrated in Fig. 3, the higher δJ/J is due to the shape of the spectral baseline (especially its tilt angle) in the edge jump region, which can be difficult to define and therefore increases the uncertainty in δJ. As it dominates the uncertainty of δα, δJ or δJ/J must be evaluated carefully. However, the errors for δα in most previous publications included only the statistical errors of δIL plus 10% "instrumental error",45, 6 without any consideration of δJ.</p><p>To estimate the error δJ, we demonstrate the data processing for the L-edge sum rule analysis as in Fig. 3, which includes the removal of a spectral background to produce an intermediate XAS spectrum with a linear background (b, black) and the subsequent removal of this linear background to produce the final spectrum (c). Although the blank sample holder spectrum was measured every time, the real XAS background used in the data process is often its manipulation (tilt or bend) or simply a theoretical polynomial (a, dashed black line). In short, the initial background is just the fit of the XAS spectra in the non-resonance regions. The choice of the tilted background slope introduces a major error to δJ, with two extreme cases for NiO shown in Fig. 3 (b, red and blue spectra), although the judgement on edge jump step's height may also bring in a minor error. The difference between the slopes for the two background lines is about 25%, while the difference in the peak heights for the two "final" NiO spectra (c) is about 20%. The final value for δJ/J is ± 9.9%, while its δIL/IL is ±3.4%. The δIL/IL and δJ/J for other complexes are listed in Table-1. According to equation (1), δα/α is about ±10.5% for NiO and ±9.7% – ±11.7% for other complexes. These errors are presented in Table-1 and Fig.2. Most of the "instrumental error" is actually calibratable while the rest random error should already be included in our statistical error of the data, therefore we do not need to add an arbitrary 10% instrumental error to the analysis.</p><p>When the number of 3d holes is small, the above approach (with δα/α ~ ±0.11) is still accurate enough to at least semi-quantitatively identify the metal's oxidation state and its number of 3d holes (H3d). For example, nominal-d8 [NiII(SR)4] has an observed H3d ~ 1.30–1.62 (1.46±0.16), which is clearly higher than d9 [NiI(SR)4]− (0.59–0.73); similarly, d7 K3NiIIIF6 has a H3d ~ 2.49–3.11, which is also significantly higher than d8 NiIIF2 (1.53–1.89). This is true even for some ultra-covalent complexes which have very small difference between different oxidation states. For example, derived from IL=8.69, 11.33, 13.25,52 the 38–50% covalent (Ph4As)2Ni[S2C2(CF3)2]2, (Bu4N)Ni[S2C2(CF3)2]2, and Ni[S2C2(CF3)2]2 complexes have distinguishable (meaning difference > δα defined by δα/α ~ ±0.11) H3d=1.01, 1.32 and 1.54. Their reported L XAS52 are cited as in Fig. S2 in SI for a reference.</p><p>However, as α increases, the observed H3d values for the two consecutive oxidation states can overlap if error-bars are included. Thus a clear assignment of the oxidation state is less obvious. For example, as shown in Table 1, the d4 MnIII2O3 has a α value of 40.89±4.78, ranging from 36.11 to 45.67, while d3 MnIVO2 has 44.78±4.77 (from 40.01 to 49.55) which is overlap with the α region for d4 MnIII2O3. Under these circumstances, one has to either improve the XAS to reduce uncertainties in the edge jump δJ, or to search for alternative features in the XAS as a normalization standard for IL. These approaches are the central topic of this publication, and we will discuss them with the L-edge sum rule analysis for dilute samples in the following sections.</p><!><p>For concentrated samples, the absorption features from the element of interest dominate the XAS, like those in Fig. 1. X-ray absorption of other elements such as C, N and O still contributes to the spectra as background, but their spectral features tend to be much smaller than those from the metal of interest. For dilute samples whose metal signal is weak, most spectral counts and features are due to the large background rather than due to the metal of interest. These background intensities (B) also contributes significant statistical noise δB ~ B1/2 to both IL and J, which increases the uncertainty of α to the point that accurate analysis eventually becomes impossible. This background problem for dilute samples can be resolved by using only the X-ray fluorescence from the element of interest as a measure of absorption in XAS, instead of using the total electron yield. If the fluorescence from the element of interest can be separated efficiently from the huge background fluorescence after X-ray excitation, the XAS detection limit can be improved significantly.</p><p>A metal concentration of ~ 1% is the highest concentration for L XAS with fluorescence detection before saturation effects start to distort the spectra.53 Therefore blue copper proteins, which have a Cu concentration ~ 1% wt./wt.,5, 54 represent the most favorable candidates for L XAS and L-edge sum rule studies.5, 556 In this publication, we repeated Cu L-edge sum rule analysis on the blue Cu enzymes from plastocyanin (Fig. 4 insert i1)5 and from the construct engineered azurin (b or i2)6 with a Ge X-ray fluorescence detector and use them as bridges between the concentrated Ni complexes and the <0.1% Ni enzymes. Our spectra reveal a H3d = 0.4 (i1) and 0.2 (i2) per Cu atom, respectively.</p><!><p>CODH catalyzes CO oxidation and acetyl-CoA synthesis. It is found in acetogenic, methanogenic, and sulfate-reducing bacteria, and fixes carbon on a global scale.56 The Ct-CODH has a Ni concentration of 770 ppm, only 1/13 of the Cu concentration in blue Cu proteins, which complicates L XAS and L-edge sum rule analysis. In frozen solution samples, the Ni concentration is further reduced. Fig. 4(b) illustrates an idealized fluorescence signal for a hypothetical sample with 0.077% Ni, 1.0% Cu and O in balance, assuming negligible background from pile-up or second order excitation (the detector's electronic background). While it is fine to resolve the 1% Cu signal with a 180 eV resolution Ge detector (Fig. 4a, orange), it is harder to resolve a signal from <0.1% Ni with the same detector and the same procedure (Fig. 4a, blue), even without any additional background. A lower metal concentration and a Ni X-ray energy of 850 eV (closer to the O background fluorescence) make the resolution of a <0.1% Ni signal difficult, let alone the resolution of its weak edge jump step J for L-edge sum rule analysis. With a 180 eV resolution Ge detector, the edge jump is often washed out or buried under the huge background and pile-up from light elements like C, N, O and Na.12, 3157, 58</p><p>STJ detectors have an energy resolution of ~15 eV instead of 180 eV. Measuring XAS with an STJ therefore produces a high resolution X-ray fluorescence spectrum, in which the hypothetical 0.077% Ni signal becomes clearly resolvable from the huge O background (Fig. 4b). In Fig. 5, the STJ measured L2,3 XAS of Ct-CODH (<0.077% Ni, green) is compared to the channeltron measured XAS for the concentrated covalent [NiII(SR)4] complex (red), both spectra show a clear low spin NiII feature and both have a clear edge jump J. The L-edge sum rule analysis leads to a α= IL/J = 12.2 and a H3d = 12.2/8.6 = 1.42, corresponding to a typical NiII site. With STJ detectors, L-edge sum rule analysis can readily be extended to samples with a Ni concentration <0.1%, such as Ct-CODH frozen solution (<0.077% Ni), because its edge jump can be clearly observed. The same procedure will also be useful in studying other Ni enzymes with a Ni concentration below 0.1%, such as Ni in NiFe H2ases. The advantage of this measurement is that it keeps the same ratio of normalized L intensity α and hole density H3d for different measurements or for different samples. On the other hand, the disadvantage is: this approach will become difficult when H3d is large (e.g. for Mn) because a perfect edge jump is almost impossible to obtain even with STJ detectors and a less perfect edge jump will introduce too much noise to the Δα and ΔH3d because Δα (ΔH3d) is large.</p><!><p>Alternatively, we can introduce an external feature S to the XAS with its integrated intensity IS comparable to IL and use it as the intensity normalization standard instead of using J. For example, as Ce has a M5 edge at 881.7 eV (close to Ni L2,3 at 852.6 eV), we add chemically inert Ce(NO3)3 to the samples to be measured and use the Ce M5 edge (IS) as the normalization standard for Ni L2,3. In this case we define α'= IL/IS. This was first tested on a NiBr2/Ce(NO3)3/buffer matrix with a 0.6% Ni concentration, as detailed in the experimental section. Compatible XAS intensities for Ni L3 and for Ce M5 edges were obtained for a molar ratio of NiBr2:Ce(NO3)3 = 5.5:1 (Fig. 6a). The advantages of this procedure includes: one can still use a Ge detector with 180 eV resolution because it can resolve the Ni L and Ce M5 peaks from the background; and one can have a minimum errorbar for δα'/α'. For the 0.6% NiBr2 sample, the uncertainty δIL/IL is about 4%, while the uncertainty of the Ce M5 δIS/IS is about 4.1%, leading to a total error of ~5.7% for α', much smaller than when using J for normalization. This minimized δα' increases the accuracy of the intensity ratio α'=IL/IS. Note that IL/IS depends on the accurate knowledge of the Ni:Ce concentration ratio in a sample, and can therefore be different in different samples because the Ce concentration can vary. That complicates extracting an absolute value for the hole density H3d from α'=IL/IS directly and makes this approach less attractive in comparison with the STJ option.</p><p>Nevertheless, it successfully opens another pathway to extend L-edge sum rule analysis to <0.1% Ni enzymes, such as H2ases (0.067% Ni)12. These enzymes catalyze the reversible reaction of hydrogen (H2) production and consumption, and monitoring their Ni oxidation states in their catalytic circle8, 12, 59–61 is critical to understanding their catalytic mechanisms. As discussed in the experimental section, we added Ce(NO3)3 as intensity normalization standard S in the as-isolated sample before the H2-reduction, both H2ase samples therefore have the same Ce:Ni ratio, although their absolute metal concentration may be different. We obtained IL/IS values of 0.92 and 0.66 (Fig. 6b), leading to a ratio of 0.92/0.66=1.39 between them. This number of 1.39 is independent of the Ce concentration and is consistent with a NiIII → NiII reduction for these two H2ase samples studied12, 59, 60. This approach should also be a good way to study dilute systems with higher H3d, such as Mn sites.</p><p>Besides Ce, a series of potential intensity standards in the spectral region from 600 to 1000 eV are also shown in the top inset of Fig. 6. For example, the Cs M5-edge at 726 eV is a good candidate as an intensity standard for Fe L2,3-edges at 706 and 720 eV. In addition, to avoid the uncertain concentration ratio of Ce:Ni, some intrinsic elements in the enzymes can also be used as a standard. For example, the as-isolated D. gigas NiFe H2ase contains 12 Fe (66 3d holes) and 1 Ni atoms (3 3d holes). In principle, both Fe and Ni could have a charge change when the H2ase oxidation state varies by ±1. However, even if we assume all the ±1 change occur in Fe, its total number of 3d holes should only change 1/66=1.5%. Fe can therefore be treated as a constant and be used as a normalization standard for L-edge sum rule analysis for Ni. In this special case, the Fe and Ni concentrations will always be the same for any given protein samples (e.g. D. gigas NiFe H2ase).</p><!><p>L-edge XAS and K-edge XAS offer complementary advantages. For example, K-edge XAS is bulky sensitive, and the edge energies of different elements are widely spaced, which allows XAS be extended to the useful EXAFS region (Fig. S3).62 In addition, since K X-rays have higher energies than L X-rays, K-edge XAS does not require a vacuum chamber, and a higher fluorescence yield which enables the analysis of more dilute samples.</p><p>On the other hand, L XAS has richer spectral features that provide a finger print of the sample's electronic structure. L-edges also have a much narrower line widths than K-edges (0.2 eV vs. 1 eV) that allow measuring these spectral features. For sum rule analysis, L XAS is favored over K XAS because L XAS measures 2p→3d transitions and therefore provides direct information about 3d bonding orbitals. These transitions are dipole allowed and therefore much stronger and more stable than the pre-K 1s→3d. As illustrated in Fig. 7, the maximum 1s→3d feature in K XAS (Fig. 7a, black) is still smaller than its edge jump, while most of the 1s→3d features are barely visible (Fig. 7a, green and red). Meanwhile, L XAS has a signal size several times stronger than its corresponding edge jump and 10–500 times stronger than the corresponding 1s→3d features in K XAS, because 2p→3d transitions are dipole allowed but 1s→3d is not. Moreover, as 1s→3d is not dipole allowed, the weak feature's intensity heavily depends on the extent of its orbital mixture with other dipole allowed transitions and thus on the geometric structures of the complex. This also makes 1s→3d transitions dependent on changes in the electronic structure and difficult to interpret. For example, according to the L-edge sum rule analysis, H3d for MnO should be about 60% of that for KMnO4. However, this ratio is only a few percent in K-edge XAS (Figure 6a, green vs. black), and MnIIIO3 actually has slightly lower intensity than MnIIO in K-edge XAS.</p><!><p>According to sum rules, the integrated L XAS intensity IL (when normalized by the invariant edge jump J) is proportional to the number of 3d holes localized in the X-ray absorber (H3d). We have used L-edge sum rules to obtain the H3d from their corresponding IL/J for a series of Ni and Mn complexes (3d10→3d0). In complementary to the previous reports, errors in estimating the edge jump (δJ) are identified as dominant and must be carefully controlled. When the number of 3d holes is small, the IL/J approach is accurate enough to at least semi-quantitatively define the metal's oxidation state. However, as IL/J increases, the H3d observed values for the two consecutive oxidation states can overlap and a clear assignment of the oxidation state is less obvious. In this case, one has to either improve the accuracy of the setup and measure XAS with a clearer edge jump J (with a smaller error δJ), or to use alternative spectral features IS to normalize IL.</p><p>We have examined these two approaches to measure and extended the method to studies of extremely dilute biological samples (<0.1% Ni), such as Ct-CODH and NiFe H2ase. For higher-accuracy XAS, we use a ~15 eV high resolution STJ detector to separate the extremely weak Ni fluorescence from the huge background to obtain the edge jump J with high signal-to-noise ratio, and continue to use the L-edge sum rule procedure developed for Ni complexes. The STJ measured L XAS for Ct-CODH (< 0.077% Ni) illustrates this option and conclude it has a NiII site. As an alternative spectral feature, we have added inert Ce to our samples and use the Ce M5 as the new intensity normalization standard for IL. This reduces the total error in α= IL/IS, as demonstrated with a dilute NiBr2 (0.6% Ni) samples and with NiFe H2ase samples (0.067% Ni). Although this analysis depends on the Ce concentration, changes in oxidation state for a given sample do not change the Ni:Ce ratio. This is shown in the comparison of as-isolated and H2-reduced NiFe H2ase samples, whose ratio of 1.39 is consistent with a NiIII→NiII reduction. These experimental approaches will enable us to extend L-edge sum rule analysis to virtually any dilute biological metals in the future.</p>

PubMed Author Manuscript

High-throughput quantification of carboxymethyl lysine in serum and plasma using high-resolution accurate mass Orbitrap mass spectrometry

BackgroundCarboxymethyl lysine is an advanced glycation end product of interest as a potential biomarker of cardiovascular and other diseases. Available methods involve ELISA, with potential interference, or isotope dilution mass spectrometry (IDMS), with low-throughput sample preparation.MethodsA high-throughput sample preparation method based on 96-well plates was developed. Protein-bound carboxymethyl lysine and lysine were quantified by IDMS using reversed phase chromatography coupled to a high-resolution accurate mass Orbitrap Exactive mass spectrometer. The carboxymethyl lysine concentration (normalized to lysine concentration) was measured in 1714 plasma samples from the British Regional Heart Study (BRHS).ResultsFor carboxymethyl lysine, the lower limit of quantification (LLOQ) was estimated at 0.16 μM and the assay was linear between 0.25 and 10 μM. For lysine, the LLOQ was estimated at 3.79 mM, and the assay was linear between 2.5 and 100 mM. The intra-assay coefficient of variation was 17.2% for carboxymethyl lysine, 9.3% for lysine and 10.5% for normalized carboxymethyl lysine. The inter-assay coefficient of variation was 18.1% for carboxymethyl lysine, 14.8 for lysine and 16.2% for normalized carboxymethyl lysine. The median and inter-quartile range of all study samples in each batch were monitored. A mean carboxymethyl lysine concentration of 2.7 μM (IQR 2.0–3.2 μM, range 0.2–17.4 μM) and a mean normalized carboxymethyl lysine concentration of 69 μM/M lysine (IQR 54–76 μM/M, range 19–453 μM/M) were measured in the BRHS.ConclusionThis high-throughput sample preparation method makes it possible to analyse large cohorts required to determine the potential of carboxymethyl lysine as a biomarker.

high-throughput_quantification_of_carboxymethyl_lysine_in_serum_and_plasma_using_high-resolution_acc

Introduction<!>Sample preparation<!>Chromatography and HRAM mass spectrometry<!>Isolation and detection<!><!>Isolation and detection<!><!>Limits of blank, detection and quantification<!>Serum-based QC<!><!>Batch-to-batch variation<!><!>Batch-to-batch variation<!>EDTA plasma versus serum samples<!><!>Throughput<!><!>Comparison with other IDMS methods<!><!>Comparison with other IDMS methods<!><!>Comparison with other IDMS methods<!>Conclusion

<p>Carboxymethyl lysine (CML) is an advanced glycation end product (AGE), produced in vivo, particularly under hyperglycaemic conditions, and available from the diet.1,2 Increased CML concentrations are associated with cardiovascular disease, diabetic nephropathy and retinopathy, chronic kidney disease and others.2–4 CML has been proposed as a potential biomarker of cardiovascular disease; however, conflicting results have been found.4–6 Enzyme linked immunosorbent assays (ELISAs) are available but suffer from steric hindrance of the antigen and interference from endogenous anti-AGE antibodies.1,2 Isotope dilution mass spectrometry (IDMS) methods are available for quantification of protein-bound CML, and samples require chemical reduction, protein denaturation, hydrolysis and drying prior to IDMS analysis. Published sample preparation methods are individual tube-based, which have limited throughput.1,5 Therefore, we developed a high-throughput sample preparation method using 96-well plates.</p><p>Published IDMS methods for CML quantification rely on multiple reaction monitoring (MRM) detection using triple quadrupole mass spectrometers (MS).7 In MRM, the first quadrupole is optimized to select for the parent ion of interest (CML) based on the mass to charge ratio (m/z). A collision cell fragments the parent ions into product ions,7,8 while the third quadrupole is optimized to select for specific product ions: the quantifier (for quantification) and the qualifier (for verification the identity).7–9 As triple quadrupole detectors provide excellent sensitivity and specificity, even in complex biofluids, they are widely used in clinical chemistry laboratories (e.g. toxicology, endocrinology and new born screening).7–9</p><p>There is increasing interest in the use of high-resolution accurate mass (HRAM) MS for absolute quantification of ions, including for routine clinical analysis and clinical research.10,11 HRAM MS relies on superior mass accuracy (typically sub 3 ppm), which allows excellent ion selectivity, provided appropriate mass-extraction windows (based on the theoretical m/z of the ion of interest) are chosen.10–13 HRAM MS analysis is commonly run in full scan mode, enabling the detection of all ionized compounds, without the need to optimize quadrupoles and collision energies for individual ions.12 The quantitative performance of HRAM MS now equals that of triple quadrupole mass spectrometry: in terms of sensitivity, mass accuracy, selectivity, although this does depend on the conditions, parameters and the metabolite of interest used.10–12,14 Some of the major advantages of HRAM MS are that data can be reanalysed retrospectively to investigate further biomarkers and that those biomarkers can be more easily identified (based on molecular formula).10–12,14 HRAM MS has been successfully used to quantify a number of small molecule groups: over 50 metabolites (including amino acids);14 amino acids (within 3 min);15 drugs and drug metabolites;16 circulating steroids11 and plasma metanephrines.11</p><p>We therefore optimized an HRAM IDMS method to quantify protein-bound CML and lysine and their deuterated internal standards using an Orbitrap Exactive mass spectrometer. To account for variation in plasma total protein concentration and variation introduced during sample preparation, the CML concentration was normalized to the lysine concentration.5,6 The three measures reported are CML (μM), lysine (μM) and normalized CML (μM per M lysine).</p><!><p>Serum collected from one healthy volunteer, stored at –80°C in multiple aliquots, was used as a quality control (QC) sample. The Glasgow University Ethics Committee provided ethical approval for collection of anonymized samples for QC (Project number 200140133). Unthawed fasting EDTA plasma samples (n = 1714) from the 30th year re-examination of the British Regional Heart Study (BRHS),17,18 stored at –80°C, were randomized to 21 batches. Approval for collection was obtained from the local research ethics committees of the 24 towns where participants were recruited. All participants provided written informed consent to participate in the study.18 The study is consistent with the World Medical Association Declaration of Helsinki. The BRHS is a prospective study which recruited 7735 men between 1978 and 1980 from 24 British towns. At the 30-year re-examination, samples were collected from 1722 men between 2010 and 2012, with the men then being aged 71 to 92 years.17</p><p>A high-throughput 96-well deep-well plate method of sample preparation based on previously published methods was developed.1,4–6 Sodium tetraborate (Sigma, Dorset, UK), sodium borohydride (Alfa Aesar, Lancashire, UK), trichloroacetic acid (Sigma, Dorset, UK), hydrochloric acid (Sigma, Dorset, UK) were used for sample preparation. Plasma samples were defrosted for 90 min and centrifuged at 20,000 × g for 5 min. Ten microlitres of plasma were added to 300 μL sodium tetraborate (0.2 M)/borohydride (0.1 M) buffer in a 96-well deep-well polypropylene plate (Thermo Fisher Scientific, Hemel Hempstead, UK). The samples were chemically reduced overnight at 4°C to prevent further production of CML (or other advanced glycation end products) during subsequent hydrolysis.1 The protein was denatured in 20% trichloroacetic acid, and the pellet was washed in 20% trichloroacetic acid. The protein pellet was then hydrolysed at 110°C in 600 μL 6 M hydrochloric acid for 24 h, using a ceramic bead-bath. After hydrolysis, the samples were dried to completion at 95°C (approximately 24 h). Immediately prior to analysis, the samples were spiked with 10 μL of 20 μM CML-d4 (Toronto Research Chemicals, Ontario, Canada, 98% pure) and 10 μL of 150 mM universally 13C labelled L-lysine:2HCl) (Cambridge Isotope Laboratories Inc., MA, USA, 98% pure) as internal standards (ISs) and reconstituted in 270 μL of 5 mM nonafluoropentanoic acid (NFPA) (Sigma, Dorset, UK) as an ion-pairing agent.</p><p>CML (Toronto Research Chemicals, Ontario, Canada, 96% pure) and lysine (Sigma, Dorset, UK) were used to prepare calibrator samples: made up in water and then mixed with IS and NFPA. A seven-point calibration curve (CML: 0, 0.25, 0.5, 1, 2, 5, 10 μM and lysine: 0, 2.5, 5, 10, 20, 50, 100 mM) was used to quantify both CML and lysine relative to their ISs. The concentration ranges were chosen based on the concentrations previously reported using IDMS quantification.1,6 Previous studies demonstrated that acid hydrolysis of calibration solutions did not alter peak area; therefore, calibrator samples were not hydrolysed.1</p><!><p>Chromatography was carried out on an UltiMate 3000 RSLC system (Thermo Fisher Scientific, Hemel Hempstead, UK) using an ACQUITY UPLC BEH C18 column (100 mm × 2.1 mm, 1.7 μm column, Waters, Wilmslow) with VanGuard pre-column (Waters, Wilmslow, UK). Mobile phase A was 5 mM NFPA (Sigma, Dorset, UK) in HPLC grade water (Fisher, Loughborough, UK). Mobile phase B was HPLC grade acetonitrile (Fisher, Loughborough, UK). The column was maintained at 50°C, and samples were eluted with a linear gradient over 9.0 min at a flow rate of 0.3 ml/min. Starting conditions were 90% mobile phase A, decreasing to 20% between 0.1 to 4.6 min; this was held between 4.6 and 6.1 min, then increased to 90% at 6.2 min and held until 9.0 min to re-equilibrate the column. The injection volume was 5 μL, and samples were maintained at 5°C prior to injection. For HRAM MS, an Orbitrap Exactive (Thermo Fisher Scientific, Hemel Hempstead, UK) was operated in high-resolution full scan positive mode, at a scan range of 120–250 m/z, a probe temperature of 150°C and capillary temperature 275°C. The mass resolution was 50,000, providing a mass accuracy of less than 1 ppm. A mass calibration was performed prior to each batch using Pierce LTQ Velos positive ion calibration solution (Thermo Fisher Scientific, Hemel Hempstead, UK). TraceFinder 3.3 (Thermo Fisher Scientific, Hemel Hempstead, UK) was used to automatically detect peaks within expected retention time and mass extraction windows.</p><p>Protein-bound CML and lysine concentrations were calculated based on integrated areas relative to those of the ISs. The normalized CML concentration (μM/M lysine) was calculated from the measured CML and lysine concentrations observed (CML/lysine × 1,000,000). This allows for variation in total protein concentration to be accounted for, analogous to reporting HbA1c in relation to haemoglobin concentration.19 It also allows variation in hydrolysis to be accounted for. Once the method was optimized, 1714 samples from the BRHS were analysed.</p><!><p>Mean Pearson correlation coefficients, mean response factors (gradients) and mean y-intercepts for protein-bound CML and lysine calibration were calculated (Table 1).</p><!><p>Figures demonstrating assay performance and median concentration observed in BRHS plasma samples.</p><p>CML: carboxymethyl lysine; LOQ: limit of quantification; LOB: limit of blank; LOD: limit of detection.</p><!><p>Separation of CML from the closely eluting peak, suspected to be valylserine or serylvaline, was good in most (1614 [97%]) samples (Figure 1). Valylserine and serylvaline are dipeptides composed of valine and serine with the same molecular formula (C8H16N2O4) and m/z as CML and are likely to be produced during acid hydrolysis. Since this is an isobaric interference, reduction of the mass extraction window cannot eliminate this interferent. The CML and CML-d4 appear to have isomerized, resulting in two peaks with the same m/z at two different retention times (1.86 and 2.38 min). The peak with the latter retention time was chosen for integration due to better peak shape.</p><!><p>Extracted-ion-chromatograms (0–5 min): (a) Close-up of CML (and other metabolites) in serum with m/z of 205.1183 and retention time of 2.39 min (note splitting of the CML peak [at 1.87 and 2.39 min] and closely eluting peak thought to be valylserine or serylvaline; (b) CML (and other metabolites) in serum with m/z of 205.1183 and retention time of 2.39 min; (c) deuterated CML in serum with m/z of 209.1343 and retention time of 2.38 min (note splitting at 1.86 and 2.38 min); (d) lysine in serum with m/z of 147.1128 and retention time of 3.11 min; (e) universally labelled 13C-lysine in serum with m/z of 153.1329 and retention time of 3.11 min.</p><p>CML: carboxymethyl lysine.</p><!><p>In IDMS, the lower limit of detection (LLOD) and quantification (LLOQ) are usually determined by calculating the signal to noise ratio (SNR) in a spectrum, with an SNR of 3 being used as an LLOD and an SNR of 10 (with accuracy of 80–120% and <20% imprecision) being used as an LLOQ.8 Due to the signal processing of the Orbitrap Exactive, with baseline removal inherent in Orbitrap data acquisition, there is generally no noise in the extracted ion chromatogram.11 Therefore, the SNR for all peaks was infinity, and SNR cannot be used to estimate LLOD and LLOQ. Instead, the LLOD and LLOQ were estimated based on the slope and SD of the y-intercept of the calibration curve.20</p><p>In the lowest water-based calibrator sample (0.25 μM), the mean CML concentration (over 21 batches) was 0.29 μM (<20% bias), with an inter-assay coefficient of variation (CV) of 10.1% (<20% variation). In the water-based calibrators, the results were linear between 0.25 μM and 10 μM. The limit of blank (LOB) for the zero-calibrator sample was estimated as 0.13 μM based on an average concentration of 0.04 μM and a standard deviation of 0.06 (LOB = mean + 1.645 SD).21 The LLOD was estimated as 0.12 μM, by multiplying the SD of the y-intercept by 3.3 and dividing by the slope of the calibration curve.20 The LLOQ was estimated as 0.16 μM, by multiplying the SD of the y-intercept by 10 and dividing by the slope of the calibration curve.20 The concentration of the lowest water-based calibrator (0.25 μM) was chosen as the LLOQ, in order to avoid extrapolation,22 since CML is endogenous to serum samples. Only one plasma sample was observed with a CML concentration below the LLOQ of 0.25 μM. Thirteen samples were observed with a CML concentration of over 10 μM (ranging between 11.9 and 17.4 μM). Since there was no evidence of detector saturation with the closely eluting isobaric contaminant peak present at approximately 1000 times higher concentrations (based on peak area), the linearity can be extrapolated.</p><p>In the lowest water-based calibrator sample, the mean lysine concentration (over 21 batches) was 2377 μM (<5% bias), with an inter-assay CV of 18.3% (<20%).22 In the water-based calibrators, the results were linear between 2500 μM and 100,000 μM. The LOB for the zero-calibrator sample was estimated as 136 μM based on an average concentration of –539 μM and an SD of 410 (LOB = mean + 1.645 SD).21 The LLOD was estimated as 1250 μM, by multiplying the SD of the y-intercept by 3.3 and dividing by the slope of the calibration curve.20 The LLOQ was also estimated as 3789 μM, by multiplying the SD of the y-intercept by 10 and dividing by the slope of the calibration curve.20 The concentration of the lowest water-based calibrator, 2500 μM, was chosen as the LLOQ, in order to avoid extrapolation,22 since lysine is endogenous to serum samples. No serum samples were observed with a lysine concentration below 2500 μM (all ≥ 8,343 μM). No serum samples were observed with a lysine concentration over 100,000 μM (all ≤ 58,231 μM).</p><!><p>Six serum QC samples were prepared and re-injected 16 times each. The mean intra-assay, intra-sample CVs were 2.7% for CML, 2.1 for lysine and 3.9% for normalized CML. This demonstrates that repeated analysis of the same sample preparation is robust. The CVs for normalized CML are increased, since the variability of both the CML measurement and the lysine measurement is contributing to the overall variability.</p><p>The intra-assay CV (based on 30 freshly prepared and reconstituted samples, each injected only once) was 17.2% for CML, 9.3% for lysine and 10.5% for CML normalized to lysine. The CV for normalized CML is lower than that of directly measured CML in serum samples, as normalization accounts for variation incorporated during individual sample preparation.</p><p>The overall inter-assay CV was 18.1% for CML, 14.8% for lysine and 16.2% for CML: lysine. The inter-assay CVs for CML (directly measured and normalized) are outside the target CV of 15% recommended by Food and Drug Administration and other guidelines for validation of bioanalytical methods within regulated environments.22 In non-regulatory environments, a CV of 20–25% is a commonly used target.23 The variation may have been introduced during high-throughput sample preparation, particularly during hydrolysis or during HRAM MS analysis.</p><p>It is recommended that the normalized CML concentration is used for clinical research studies, as it accounts for variation in sample preparation and in blood total protein concentrations. This is analogous to reporting HbA1C in relation to haemoglobin concentration.19</p><p>To assess the variability of the HRAM MS analysis and the sample preparation, a serum-based QC was prepared and run with every plate. The normalized CML concentration obtained was within QC limits, according to the Westguard Multi-rules, for all but one batch, and all but three within 2 SD of the mean of 12 run-in samples (Figure 2). Unfortunately, in some runs, this serum QC sample was contaminated – either from an unknown contaminant eluting over a wide retention range or from co-elution of the isobaric peak (suspected to be valylserine or serylvaline) (Figure 1).</p><!><p>Levy-Jennings plot displaying the variability of the measured concentration of CML (normalized to the measured lysine concentration) in the quality control serum samples run with every batch. The mean normalized CML concentration derived from previous analysis of 12 quality control serum samples is referenced as the grey line. The green, yellow and red lines reference the mean ± 1 2 and 3 SD, respectively. Note for some runs co-elution with an isobaric interferant or contamination of the QC sample meant that CML concentration could not be measured.</p><p>CML: carboxymethyl lysine.</p><!><p>Since no external or commercial QC material exists, the overall mean, median and interquartile range (IQR) was monitored for each batch of 95 samples to check for batch effects. The batch-to-batch values are expected to vary, since each batch includes different samples; however, no obvious trends were observed in normalized lysine concentration (Figure 3).</p><!><p>Box plots showing median (line), box (interquartile range) and whiskers (<1.5 × IQR) for normalized CML concentration in BRHS plasma samples run in each batch, arranged by plate number (n = 95 per batch). The circles represent outliers (<3 × IQR) and the stars represent extreme outliers (>3 × IQR). No obvious trends are observed from batch to batch; samples were randomized before sample preparation. Normalized CML concentrations of ≥ 200 mM/M lysine were excluded from the figure for clarity.</p><p>CML: carboxymethyl lysine.</p><!><p>To attempt to investigate the stability of the dried hydrolysed samples, 12 samples which were stored for three months at –80°C were reconstituted with NFPA and analysed by HRAM MS. The mean CML, lysine and normalized CML concentrations were not significantly different. To check for unwanted trends in the data over time, the box plots were arranged by date run on the LC-MS, by date of sample preparation and by delay between sample preparation and LCMS run. No obvious trends were observed (data not shown).</p><!><p>Paired samples of serum (serum separator vacutainers) or plasma (K+EDTA vacutainers) from seven healthy volunteers were run in triplicate (Table 2). The concentrations observed were not statistically significantly different for all three analytes. CML concentrations were also found to be similar in plasma versus serum samples in a previous study.4</p><!><p>Comparison of results from paired serum and EDTA plasma samples (seven paired samples run in triplicate). Results of paired t-test demonstrated no significant (ns) difference between the two sample types.</p><p>EDTA: ethylenediaminetetraacetic acid; CML: carboxymethyl lysine.</p><!><p>We developed a 96-well deep-well method of sample preparation based on previously published methods individual tube methods.1 The hands-on preparation time (not including incubations) is reduced more than five-fold compared with the estimated sample preparation time reported for individual tube methods1 (Table 3). It also uses lower volumes of reagents. MS analysis time is slightly increased: 9-min run compared with the 7.5-min run published.4 However, analysis of a 96-well plate can still be completed within an overnight run.</p><!><p>Comparison of hands-on sample preparation time using the 96-well versus individual tube method and comparison of HRAM MS analysis time for 1000 samples.</p><!><p>The CML concentration was measurable in 1664 samples from the BRHS. For 50 samples (3%), the CML peak co-eluted with the isobaric peak (suspected to be valylserine or serylvaline) or there was broad contamination. The CML and normalized CML concentrations were positively skewed and results for lysine were negatively skewed (Figure 4).</p><!><p>Histogram of (a) CML concentration (μM); (b) lysine concentration (μM); (c) normalized CML concentration (μM/M lysine) in 1664 BRHS EDTA plasma samples.</p><p>CML: carboxymethyl lysine.</p><!><p>A mean of 2.7 μM (full range: 0.2–17.4 μM) with an SD of 1.4 μM and a median of 2.5 μM (IQR: 2.0–3.2 μM) was found for CML concentration in the BHRS samples (n = 1664). A reference range of 1.1 to 5.6 μM was calculated (mean ± 1.96 SD, after log transformation). It should be noted that the samples obtained at the 30th year re-examination of the BRHS were from white European males aged 71 to 92 years and that up to 444 of the participants had a diagnosis of CVD.17</p><p>The mean CML concentration of 2.7 μM measured is similar to the mean of 2.8 μM (SD 0.4 μM), previously reported for 10 healthy controls1 (Table 4). The BRHS median CML concentration of 2.5 μM is also similar to the median concentration of 2.9 μM (range 1.7 to 4.4 μM) observed in 31 individuals with type 1 diabetes mellitus with normal renal function.5</p><!><p>Comparison of CML and normalized CML concentration observed by HRAM MS with high-throughput sample preparation vs. LC-MS/MS with individual tube-based sample preparation.</p><p>AGE: Advanced glycation end-product; COPD: chronic obstructive pulmonary disease; CVD: cardiovascular disease; EDTA: ethylenediaminetetraacetic acid; eGFR: estimated glomerular filtration rate; IQR: interquartile range; NA: not applicable; SD: standard deviation; T1DM: type 1 diabetes mellitus; T2DM: type 2 diabetes mellitus; CML: carboxymethyl lysine.</p><!><p>The mean lysine concentration in the BHRS samples was 39,490.5 μM, with an SD of 6268.6 μM, after hydrolysis. The median was 39,773.5 μM (IQR: 36,109.0–43,210.6). A reference range of 26,182 to 57,677 μM was calculated (mean ± 1.96 SD, after log transformation). To our knowledge, no reference ranges have been reported for lysine concentration in serum or plasma hydrolysate, as this is not a physiologically relevant measure. The lysine concentration correlated with the albumin concentration (r2 = 0.24), demonstrating that variation in protein concentration and variation in sample preparation contribute to the variation in lysine concentration.</p><p>A mean normalized CML concentration of 69 μM/M lysine (range: 19–453; SD 34) was reported in the BRHS. The median was 65 μM/M lysine (IQR: 54–76). A reference range of 34 to 123 μM/M was estimated from the log-transformed data (mean ± 1.96 SD).</p><p>The median normalized CML concentration of 65 μM/M lysine measured is similar to the median of 68 μM/M lysine, previously reported for 18 sedentary individuals26 (Table 4). It is also broadly similar to the medians of 51 μM/M lysine and 80 μM/M lysine reported for 70 individuals in the top and bottom tertile of AGE score, respectively.29</p><p>The mean of 69 μM/M lysine observed in the BRHS samples was far lower than the medians reported by Hull et al.: 132 to 140 μM/M lysine4 (Table 4). They measured lysine concentration using IDMS; however, they did not report the raw CML or lysine concentrations observed. Their range is based on repeated analysis of a pooled sample from 10 healthy volunteers analysed under different preanalytical conditions, none of which were found to significantly affect the CML concentration.4 It is also far lower than the mean of 158 μM/M lysine reported in 21 healthy children25 and lower than the mean of 83 μM/M lysine reported in 44 ex-smokers30 and of 93 μM/M lysine reported in 169 individuals without diabetes.31</p><p>The median normalized CML concentration is about double that reported by Hanssen et al.6: 34 μM/M lysine (IQR: 29 to 39) based on 558 individuals without prior CVD. They derivatized CML with 1-butanol: HCl, as an alternative to using NFPA as an ion-pairing agent, before IDMS analysis. The sample was split after hydrolysis and the lysine concentration was measured separately, again using IDMS. Hanssen et al. did not report the raw CML or lysine concentrations observed. Perhaps incomplete derivatization or differences in sample population account for the difference in ranges observed between our study and theirs. It is also approximately double that reported for over 200 individuals in the top tertile for diastolic function28 and more than double that reported in five healthy individuals.34</p><!><p>CML is a challenging AGE to measure, as it is being detected in the presence of other amino acids (including lysine) present at 1000 times higher concentrations in the sample. However, the results suggest that this is a robust method for the quantification of CML (normalized to lysine), despite reducing the hands-on time (and reagent volumes used) for sample preparation substantially. There are no gold standard methods available for comparison at present, only ELISAs and other IDMS methods. Our method appears to compare relatively well to other IDMS-based methods. At present, we recommend this method for research use only. Further work is required in clinical research to determine whether CML is indeed a useful biomarker. The increased throughput provided by this sample preparation method will aid this endeavour. If CML is a useful biomarker, then it will be appropriate to evaluate whether the improved mass accuracy of an ultra high-resolution instrument is better for sensitivity and selectivity than a triple quadrupole instrument for this analyte.</p>

PubMed Open Access

Strategies for quantifying C60 fullerenes in environmental and biological samples and implications for studies in environmental health and ecotoxicology

Fullerenes are sphere-like molecules with unique physico-chemical properties, which render them of particular interest in biomedical research, consumer products and industrial applications. Human and environmental exposure to fullerenes is not a new phenomenon, due to a long history of hydrocarbon-combustion sources, and will only increase in the future, as incorporation of fullerenes into consumer products becomes more widespread for use as anti-aging, anti-bacterial or anti-apoptotic agents. An essential step in the determination of biological effects of fullerenes (and their surface-functionalized derivatives) is establishment of exposure-assessment techniques. However, in ecotoxicological studies, quantification of fullerenes is performed infrequently because robust, uniformly applicable analytical approaches have yet to be identified, due to the wide variety of sample types. Moreover, the unique physico-chemistry of fullerenes in aqueous matrices requires reassessment of conventional analytical approaches, especially in more complex biological matrices (e.g., urine, blood, plasma, milk, and tissue). Here, we present a review of current analytical approaches for the quantification of fullerenes and propose a consensus approach for determination of these nanomaterials in a variety of environmental and biological matrices.

strategies_for_quantifying_c60_fullerenes_in_environmental_and_biological_samples_and_implications_f

1. Introduction<!><!>1. Introduction<!><!>2. Water-soluble fullerene dispersions and their detection<!><!>2.1.1. Dispersion of C60 fullerenes in water<!>2.1.2. Impact of nC60 preparation method on nC60 chemistry<!>2.1.3. Alternative nC60-preparation methods<!><!>2.1.3. Alternative nC60-preparation methods<!>2.1.4. Impact of nC60 solvent-exchange method on downstream ecotoxicological assays<!>2.2. Phase 2 \xe2\x80\x93 nC60 extraction from aqueous media<!>2.2.1. Liquid-liquid extraction<!>2.2.2. Solid-phase extraction<!><!>2.2.2. Solid-phase extraction<!>2.3.1. Analytical separation<!>2.3.2. Analytical detection<!>2.3.3. Establishing the reference/calibration standard<!>2.3.4. Calculating C60 recovery<!>2.3.5. Factors affecting the recovery<!>3. Conclusions<!>

<p>Fullerenes are molecules of pure carbon (e.g., C60 and C70) that are increasingly investigated for use in biomedical, cosmetic and industrial applications, ranging from drug delivery systems to anti-aging formulations to electrical components. Commercial and scientific interest is spawned by their unique physico-chemical properties that make them both robust and versatile. Fullerenes possess a thermodynamically-stable carbon shell of about 10 Å in diameter (for C60) that can withstand heat, pressure and radiation [1–3] but, due to their unique electron-hybridization pattern of sp2 bonds, they are also highly configurable. These electron double bonds allow pristine fullerenes, which are intrinsically hydrophobic [4], to become readily derivatized and water soluble through the addition of various functional groups {e.g., oxygen-, hydroxyl-, polyvinylpyrrolidone (PVP)- [5], and phenyl Cn-butyric acid methyl-ester moieties [6]}. This functionalization enables fullerene derivatives to permeate through cellular membranes [7,8], where they may interact with and inhibit active sites of enzymes [9]. Hence, potential biomedical applications are numerous and have been comprehensively reviewed [10]. Briefly, fullerenes and their derivatives were found to show potential as antiviral and antibacterial agents, slow-release drug- or gene-delivery systems, electron-transfer shuttles, and antioxidant and radical scavengers, although they also can generate reactive oxygen species {e.g., singlet oxygen, superoxide and hydroxyl radicals (references in [10]).</p><p>The marketable value of fullerenes is anticipated to lead to increased human and environmental exposure on a global scale. This process is already under way, most notably in the widespread dermal application of cosmetic products [11], inhalation of dust particles [12] or soot from combustion processes [13], and discharge of fullerene-containing products into waterways [14]. To unravel the biological effects of fullerenes, numerous groups have initiated ecotoxicological [15–18], pharmacokinetic [19–21], environmental-fate [14] and environmental-impact studies [1,22–26].</p><p>To interpret the toxicological data accurately, establish the pharmacokinetics, or determine the body burden, the concentrations that organisms or cells are (intentionally) exposed to during experiments should be unambiguously determined using a robust methodology. Moreover, the availability of robust, overarching methods would further improve exposure assessments, which, along with biological effects, are required for meaningful risk assessments, and can be a prerequisite for considering potential regulations for limiting environmental discharge and human exposure. The scientific community has therefore initiated an effort to deliver such methods, which are reviewed below and are extrapolated to human and ecological risk-assessment studies.</p><p>A tremendous amount of data has been generated regarding the environmental implications of fullerenes. For readers specifically interested in the field of environmental exposure to fullerenes, we would like to refer to recently published comprehensive reviews on:</p><!><p>the quantitative analysis of fullerenes in environmental samples [27];</p><p>the colloidal properties of fullerenes [28];</p><p>the behavior and ecotoxicity of carbon nanoparticles in the aquatic environment [29];</p><p>the physico-chemical interactions between nanoparticles and biological systems [30]; and,</p><p>the mechanisms driving toxicity of fullerenes [31].</p><!><p>In the past few years, significant advances have been made in the analytical quantification of fullerenes [18,20,27,32]. Traditional detection strategies include visualization techniques (e.g., transmission electron microscopy) to ascertain the presence of fullerenes [23,33], or use of radioactively or fluorescently labeled fullerenes for fate studies of carbonaceous nanoparticles in vivo and in vitro [34–36]. Since each method comes with its own benefits and limitations, it is frequently desirable to use two or more strategies for fullerene tracking. For example, labeling techniques offer convenient ease of use but may also affect the behavior of the nanomaterials during the course of the experiment. While these methods are valid and yield valuable information, future investigations on human subjects (e.g., to determine the body burden) will require a less invasive, more quantitative approach. More recent approaches therefore employ liquid chromatography coupled to ultraviolet spectroscopy (LC/UV) or mass spectrometry (LC/MS) to quantify fullerenes in plasma or skin [37].</p><p>In this article, we review analytical quantification approaches for pristine C60 fullerenes, and place them in the framework of studies in environmental health, ecotoxicology and pharmacokinetics. Most of the manuscripts discussed are readily available through scientific databases (e.g., NCBI PubMed and ISI Web of Knowledge). Our goal was to elucidate and discuss three phases in the workflow process that are pivotal for human health and ecotoxicological studies involving fullerenes and that have been shown to affect significantly the outcome of nanometrology, which we identified as:</p><!><p>nC60-stock suspension preparation;</p><p>spiking and extraction; and,</p><p>detection, quantification and recovery determination.</p><!><p>Experimental protocols for the quantification of fullerenes in biological systems encompass three distinct phases that can significantly affect the outcome of the ecotoxicological and analytical findings (Fig. 1):</p><!><p>During the initial phase, the C60-calibration standards and the aqueous nC60-stock suspension are prepared. While the former is used exclusively for quantification (with some exceptions [37]), the latter is used for both quantification [22] and spiking of fullerene aggregates into aqueous and biological matrices. Preparing an aqueous-stock suspension of the hydrophobic C60 molecules has proven to be a pivotal step in assessing the toxicity of fullerenes and is required for, e.g., intravenous injections, topical administrations, bronchoalveolar lavages, and in vitro toxicity assays [20,36,38,39]. The dispersion of C60 in the air has also been reported, mainly to assess the impact of inhalation of nanomaterials during in vivo exposure studies [39].</p><p>The second phase concentrates on the exposure experiment and encompasses the spiking/dosing and extraction of the nanomaterials from the environmental and biological samples. In this section, the matrix-dependent flexibility of the experimental design is reviewed. As discussed in more detail below (sub-section 2.2.), the selection of the spiking strategies and extraction protocols from blood, milk, urine and tissue is driven mostly by the availability of source materials and by existing analytical limitations.</p><p>The final phase comprises the actual quantification of the fullerenes. Various protocols have been described to determine the recovery efficiency of fullerenes in biological matrices. Those approaches are reviewed in the framework of the type of exposure experiment conducted.</p><!><p>Pristine C60 fullerenes are highly hydrophobic (log Kow = 6.67), due to their composition of pure carbon [4]. As a result, they are only poorly dissolvable in aqueous matrices (solubility in water 7.96 ng/L – 11 µg/L) [4, 40]. When C60 powder is introduced into water (Fig. 1), it can take between two weeks and several months of stirring to obtain a stable aqueous suspension [38,41–44]. As they slowly transition from the bulk phase to the water phase, pristine fullerenes form stable aggregates, which exhibit colloidal properties and are generally referred to as nano-C60 (nC60). These water-stable nC60 particles exhibit physico-chemical duality, as they form aggregates with a hydrophobic core surrounded by a hydrophilic shell or cluster of water molecules [45]. This polar cloak is stabilized by weak donor-acceptor type bonds between the hydrogen of the water molecules [45], and is potentially the source of the typical electronegative surface of the nC60 aggregates [46,47].</p><p>Direct transfer of fullerene powder into water by stirring or sonication is considered the most representative method to create nC60 [24], as it approaches the real-world transition of C60 from the particulate or commercialized form to the aqueous form. nC60 prepared by direct transfer into water is denoted as aqu/nC60 and it was shown to be the most stable form of nC60 nanoparticles compared to the aggregates generated with other methods (see sub-section 2.1.3) [48]. The direct dissolution method avoids the use of solvents (e.g., THF) that could impart artifacts to (eco)toxicological studies, but the preparation can be extremely time consuming [49].</p><p>Electrophoretic mobility, electrokinetic potential, surface charge or zeta potential is commonly measured using phase-analysis light scattering (PALS) instrumentation [48,50]. The size range of the colloids can be determined directly in suspension using dynamic light scattering (DLS) [44] or field flow fractionation (FFF) [51]. These data can then be validated after evaporation of the sample, using transmission electron microscopy (TEM) [44,52] and atomic force microscopy (AFM) [53]. The size ranges of spherical and hexagonal-shaped nC60 particles are often mono-dispersed and change only slightly over time [38,44,54,55], but some have reported polydispersity of the nC60 particles [47,51]. The reported sizes vary in the ranges from ~10–500 nm in ultrapure water [52,56] to ~500–1200 nm in the presence of ionic compounds (i.e. salts) (Table 1).</p><p>"Salting out" of fullerenes occurs when the electron double layers of the fullerenes are compressed by increasing ionic strength, and intermolecular repulsion is decreased, so the fullerenes form larger colloidal structures due to the attractive Van der Waals forces [55]. Nonetheless, the size distribution can be skewed due to the presence of elongated nC60 particles 300–860 nm long [44]. Also, in some cases, TEM can be used to screen for even smaller particles in the range <10–20 nm, often not captured by DLS [44]. However, neither approach is powerful enough to confirm or deny the presence of single hydrated C60 molecules or aggregates of 3.4 nm [45,57]. Accurately characterizing the colloidal properties of the fullerene-nanoparticle dispersions will prove important for interpretation and interlaboratory comparison of ecotoxicological findings.</p><!><p>During the nC60 preparation, extended contact times or high-energy sonication procedures may influence the interactions between the C60 molecules and the matrix components [48,58] (e.g., water molecules, ions, dissolved gasses, (non-)ionic surfactants, and biological macromolecules). Prolonged exposure of fullerenes to these matrix components, in combination with the high energy provided through sonication, was shown to yield novel covalently-bound polar moieties [59] or increase the measured oxygen content in a suspension of pure-carbon nC60 particles [48]. Hence, the nC60 particles could comprise a hydrophobic core of pristine fullerenes surrounded and stabilized by a small population of polar and amphiphilic C60-derivative molecules [22,48]. Because these hydrophilic C60 derivatives remain unidentified to date, C60 remains, within our ability to detect with 13C NMR, underivatized in these aggregates [22]. Nevertheless, several groups have observed that the elemental composition of aqu/nC60 contained 0.3% oxygen [1] or increased from 1.6% in C60 powder to 2.9–5.0% after being stirred for 40 days in water [48]. These findings could indicate the presence of oxidized C60 derivatives. In the same framework, other authors have confirmed the presence of ions with an m/z ratio corresponding to C60O [60,61], with a peak area of approximately 2.6% of the C60 peak (m/z 720) [62], while others reported its absence [44]. Even though polar C60 derivatives will not interfere with the m/z-based quantification of pristine fullerenes [62], they might remain undetected in some cases because the extraction methods are based on the use of non-polar solvents only. In addition, while the oxidized forms are currently detected to a lesser extent (due to the above-mentioned analytical challenges), cells or organisms used for in vitro assays and during in vivo experiments will be co-exposed to these derivatives. Because their relative contribution to the toxicity of fullerenes has not been established to date, this co-exposure should be considered in future ecotoxicological assays. The oxidized derivatives may possess inherent toxicity or may alter that of the fullerenes under investigation.</p><!><p>Alternative approaches have been devised to facilitate or accelerate the transition of fullerenes into the water phase for research purposes only:</p><!><p>sonication [42,51,63];</p><p>solvent exchange [22,37,43,55,56,60,62–66];</p><p>co-solubilization/encapsulation [42,43,63];</p><p>derivatization/functionalization [34]; or.</p><p>combinations of the above (Fig. 1).</p><!><p>The first two [(1) and (2)] are by far the most commonly-reported approaches in laboratory-based investigations (e.g., oral gavage in rats, exposure of in vitro cell cultures, detection in porcine plasma, skin samples, and wastewater and surface-water analogs), as they are less time-consuming than direct dissolution in water [41–43]. By contrast, the next two approaches [(3) and (4)] are used in the search for applications of fullerenes [6, 36]. For this reason, they will not be discussed further and are not shown in Fig. 1.</p><p>There are many variations of the solvent-exchange protocol to generate nC60 particles. Table 1 reveals that 19 papers reported four different approaches, most often with a small modification of the previous protocol (yielding about 26 variations). In general, C60 powder is dissolved in an organic solvent {e.g., benzene [62,63], toluene [24,47], tetrahydrofuran (THF) [55,56], dimethylsulfoxide (DMSO) [60,67], or dimethylformamide (DMF) [67]}, then, in some cases, filtered through membranes of varying pore sizes (0.2–0.45 μm). Thereafter, the C60 solution (tol/C60, ben/C60, or THF/C60) is transferred to ultrapure water at a given rate, in some cases in the presence of a co-solvent (e.g., acetone, ethanol, chloroform) or a co-solubilization substrate (e.g., PVP) [55,62–64]. Research showed that the rate at which C60 is transferred from the organic to the aqueous phase can significantly affect the size of the particles [22]. The organic solvents are ultimately removed either by boiling or rotary evaporation (Table 1).</p><p>The nC60 suspension obtained can be sonicated, allowed to settle, and/or be filtered using membranes with pore sizes of 0.22–10 μm [22,43,64,65] (Table 1). Varying the transfer solvent, temperature, C60 concentration, and mixing regime appears to affect the colloidal properties {e.g., size [52,68], structure [45,52,56,57], and charge [46,56] of the nC60 particles}. However, the reported size and zeta potential of the different nC60 suspensions in ultrapure water are similar (Table 1). Nonetheless, one must consider how the different preparation methods can affect other less frequently reported colloidal properties (e.g., surface area and particle shape) or the outcome of the toxicological assays. The colloidal properties are of particular importance, as they will determine the target, the mechanism and the strength of interaction between nanoparticles and biological interfaces [30].</p><!><p>Even though it is conventional to use organic solvents to prepare stock solutions of hydrophobic small molecules (e.g., endocrine-disrupting compounds) to spike cell cultures accurately, this approach is not so straightforward for fullerenes. Despite the characteristically high vapor pressures of the solvents (most often THF, Table 1), they are intercalated in the crystalline-like lattice of the nC60 or are adsorbed to the nC60 particles [19,22,24,46,47,69], instead of completely evaporating from the aqueous sample. This causes nC60 to appear more toxic and reactive during toxicological studies when prepared with THF as opposed to water [18,22,24,38,44,70–72]. Moreover, these studies illustrate synergistic toxic effects between nC60 and residual THF [22,42,69] or THF-derivatives (i.e. THF-hydroxyperoxide, 2-tetrahydrofuranol, or γ-butyrolactone) [38,73]. As a convention, aqueous nC60 suspensions prepared using a specific transfer solvent are denoted as "solvent"/nC60, e.g., THF/nC60.</p><p>Ultrafiltration of the samples followed by multiple rinsing steps (Fig. 1) is efficient in removing the residual THF from the nC60 lattice and mitigates the observed nC60 toxicity for lung-cell cultures [38]. Moreover, solvent-free nC60 (prepared directly in water) was shown to be non-toxic to E. coli at concentrations of up to 500 ppb [56], while THF/nC60 completely inhibited microbial growth [22]. Nevertheless, the toxic effects might not be limited to residual solvent alone, as differences in colloidal properties between the aqu/nC60 and THF/nC60 have been shown [30,38,66,74,75], so these differences should not be overlooked when assessing the origin of the different biological effects.</p><p>The classic strategy of assessing the impact of the vehicle solvent (i.e. solvent control) is not ideal when working with fullerenes because it is difficult to quantify the residual transfer solvent in these nC60 dispersions. Moreover, the adverse effects of THF differ between directly applied THF and THF delivered as THF/nC60 [24,38]. To resolve the issues with THF in toxicological studies, ethanol has been chosen by some investigators as the sole transfer solvent because it was found not to influence a genotoxicity assay with human lymphocytes [44]. However, in another study, the cytotoxicity of EtOH/nC60 was intermediate to that of THF/nC60 and aqu/nC60 [70]. Hence, we recommend evaluating the transfer solvents case by case when it is not possible to use the more relevant aqu/nC60.</p><!><p>The following section describes liquid-liquid extraction (LLE) and solid-phase extraction (SPE) methods that are used for nC60 in environmental and biological samples.</p><!><p>LLE is the most commonly used technique for extracting C60 from aqueous samples. LLE is a traditional method for extracting non-polar molecules from aqueous phases into non-polar solvents where the choice of solvent is paramount to the extraction efficiency. In one case, benzene, heptane-isoamyl alcohol, and chloroform-methanol were all inefficient for extracting C60 solubilized with PVP from rat plasma [76], perhaps due to solubility in that solvent. Summaries of solvents with high C60 solubility are available [77], and toluene has become the most widely-used solvent for LLE.</p><p>After discovering that nC60 is not extracted in non-polar solvents alone [22,45,62], addition of salt to the aqueous nC60 solution was shown to be essential to destabilize the electric double layer of the nC60 aggregates and facilitate extraction into the toluene phase with high recoveries [22]. Many salts have been investigated for this purpose, including Mg(ClO4)2 [22,37,41], KCl [37], and NaCl [64,65] in ranges up to 3.4 M [64]. At least 10 mM Mg(ClO4)2 was found to be required to extract trace amounts (60 ng) of nC60 from water [37] (Fig. 2). The use of these various salts would suggest that destabilization of nC60 occurs as a result of increased ionic strength [37], although oxidative effects of salts [e.g., Mg(ClO4)2] have not been thoroughly investigated. After all, oxidation of nC60 would produce more polar fullerene molecules, which are less soluble in toluene [22,78] (see sub-section 2.1.2.). The fact that salts improve extraction of aqueous fullerenes is an interesting observation for the future investigation of human and murine urine.</p><p>LLE requires thorough mixing of the aqueous and solvent phases, which has been done in various ways. Horizontal and rotary shakers have been used to promote mixing over times ranging from 10 min [37] to 4 h [65]. A simplified, non-exhaustive LLE was shown to be effective for just 10 min of mixing on a rotary shaker at 500 rpm [37]. This vigorous mixing can cause formation of emulsions between the two phases in the presence of organic material (e.g., proteins and lipids) and/or surfactants. In extractions of nC60 from bovine serum albumin (BSA) solutions and porcine plasma, it was found that a ratio of glacial acetic acid (GAA) to sample volume of greater than 2.25 [37] was necessary to eliminate formation of emulsions. This addition of GAA was also found to eliminate the need for addition of Mg(ClO4)2, possibly due to the low pH and subsequent lower zeta potential of the nC60 aggregates [48].</p><!><p>SPE is a separation technique in which molecules extracted from aqueous solution through sorption onto a solid matrix, after which the target compound is eluted into the desired solvent. SPE has been used previously to extract nC60 (Table 2). Reversed-phase SPE is most often used in which a hydrocarbon sorbent removes nC60 from the aqueous phase to be eluted into a non-polar solvent. Various types of SPE columns, including Sep-Pak C18 cartridges [41,76], 500 mg/6mL Strata C18-E [64], and 500 mg/6mL Strata SDB-L [65], have been used with varying results. Column-preparation protocols have also varied with a range of conditioning solvents (e.g., methanol, acetonitrile, and toluene) and the use of salts (KCl and Mg(ClO4)2) in the nC60 aqueous sample [41,76].</p><p>SPE is especially advantageous for extractions of large sample volumes, as it conserves solvent use (compared to LLE), and the samples can be concentrated from the L range down to hundreds of µL for analysis [64,65]. C60 is often eluted from SPE cartridges with 10–15 mL of toluene, although 2.5 mL were used for C18 Sep-Paks with groundwater and surface water [41]. The extraction of fullerenes from large-sample volumes into small-solvent volumes allows SPE to have high concentration factors more economically compared to LLE (Table 2). Although a common method in analytical chemistry is to evaporate to dryness and reconstitute into a known volume of solvent, this technique can lower the recovery of C60 [37]. This decrease in efficiency could be due to aggregate formation and/or adsorption to the glassware [37]. These various methods of applying SPE to aqueous samples (with varying success, see sub-section 2.3.) demonstrate an opportunity to standardize SPE protocols for nC60 extraction based on experience with environmental samples.</p><p>Despite the popularity of (automated) SPE to extract other organic compounds from biological and environmental matrices in "real-world" concentrations, the use of SPE for the recovery of nC60 dispersions from environmental and biological samples has been limited, probably due to the early successes with LLE [22] and the matrix-dependent recoveries with a simplified SPE protocol [64]. Hence, the potential of an optimized SPE protocol seems not to have been investigated to date. Nevertheless, most of the SPE-related issues could be resolved with some basic optimization steps, by:</p><!><p>diluting the sample to improve sample flow-through rates;</p><p>adding solvents to the sample to destabilize emulsions (e.g., in milk or blood);</p><p>adding a sample-homogenization step and avoiding filtration to overcome loss of C60 retained in the biomass or the filter itself [65];</p><p>investigating the best cartridge types, volumes, pore sizes, cleaning solvents or solvent mixtures;</p><p>adding a protein- and lipid-removal step; and,</p><p>investigating the ideal elution conditions for C60 and the different fullerene derivatives.</p><!><p>Once all these factors have been taken into account, SPE could prove a valuable approach due to its high reproducibility (when automated), low solvent consumption, high sample-volume-loading capacity, and its known capacity to separate analytes from interferences at high efficiency.</p><!><p>Chromatographic separation of fullerenes has been thoroughly investigated since the 1990s [77,79]. Detection of C60 in environmental samples has occurred rather recently, and chromatographic conditions are similar among researchers (Table 3). LC columns for fullerene analysis are commonly 3.9–4.6 mm i.d. by 150–250 mm packed with 5 μm particles of C18, pyrenylpropyl, pyrenylethyl, or related hydrophobic stationary phases [4,41,51,60,65]. Toluene is the most common eluent for Buckyprep columns because of its high solubility for C60, and mixtures with acetonitrile (20–45% ACN) have improved separation on C18 columns. Flow is often isocratic at rates around 1 mL/min, and elution of C60 can be achieved in as little as 3 min [64].</p><!><p>Following extraction into solvent, detection of C60 can be carried out by UV-Vis spectrometry, LC/UV-Vis or LC/MS (Fig. 1). For environmental and biological samples, UV-Vis spectrometry can be impaired by interfering organic compounds. For these samples, chromatography offers separation of C60 from background interferences enabling detection with UV-Vis spectrometry or MS.</p><p>The ionization of fullerenes (mainly yielding m/z 720 in negative mode) has been achieved using different approaches [e.g., electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), atmospheric pressure photoionization (APPI), matrix-assisted laser desorption/ionization (MALDI), laser desorption/ionization (LDI)] reviewed previously [27], with the first two being the most frequently used (Table 2). MS has been used to detect C60 down to 0.2 pg of injected mass [60] while UV-Vis has been reported to achieve a limit of detection (LOD) of 50 pg [76].</p><p>In one case, the LOD was extremely low due to the high injection volume of 150 µL (Table 2) [14]. In that study, a remarkably high ESI source temperature (725°C) was also reported [14] compared to the commonly used temperature ranges of 180–220°C and 200–450°C for APCI [32,64,65,76] and 125°C and 250°C for ESI [60] as source and desolvation temperatures, respectively. Whole-method LODs are a function of numerous parameters including LC injection volumes, initial sample volume, and sample pre-concentration (Table 2).</p><!><p>The most common way to quantify fullerenes in standard ultrapure-water dispersions (i.e. spike suspensions) is to conduct exhaustive LLE into toluene, and then to quantify the concentrated extract using UV-Vis, LC/UV, or LC-MS with nominal calibration standards of C60 dissolved in toluene [22,37,41,60,65,76] (Fig. 1). To validate this approach as a reference method, gravimetric analyses have been performed by measuring the residual dry mass from nC60 stock dispersions after complete evaporation of the aqueous phase [22]. Thermal optical transmittance (TOT) analysis has been used to quantify the total-carbon content of multi-walled carbon-nanotube-stock solutions [80] and could be applied as a method of verification of the C60 concentrations in nC60-stock solutions obtained by LLE with LC/UV or LC/MS detection. Similarly, total organic carbon (TOC) analyses can be used to provide verification of nC60-stock concentrations [66]. Finally, filtration or ultrafiltration followed by resuspension of the filter cake in toluene can also be applied to isolate fullerenes {even in matrices with low concentrations of hydrophobic/amphiphilic biosolids (e.g., urine) [42,44,62]}.</p><p>Spike and recovery experiments are challenging because of the uncertainty in concentrations of the nC60-stock solutions. Nominal concentrations of the nC60-stock dispersions that are based on the ratio of a weighed mass of dry C60 powder added to a volume of ultrapure water can be up to 33% higher than quantification based on the reference method (i.e. exhaustive LLE followed by LC/UV or LC/MS analysis) [60]. This discrepancy between the nominal and measured values could be a result of loss of C60 during nC60-preparative steps {e.g., filtration, adsorption to glassware [27,37,64], or sonication in an unsealed vessel [31]}, or due to aggregation and settling of nC60 in aqueous matrices [56,65]. Identifying the source and the extent of this disparity could be of significance for ecotoxicological studies, as cells are presumably exposed to higher fullerene concentrations than we are able to recover and to measure.</p><p>Table 3 summarizes many ways in which researchers have quantified C60 concentrations in nC60 dispersions, and determined recovery efficiencies after spiking of sample matrices.</p><!><p>To calculate C60 recovery, the amount of nC60 that is recovered from the sample matrix is compared to the amount of C60 recovered from ultrapure water or in a nominal calibration standard in toluene (see sub-section 2.3.1.). As shown in Table 3, recoveries can be 2–110%. To understand why recoveries can vary to such an extent, we need to investigate the method of recovery determination in every cited study. First, it is necessary to determine with respect to which type of calibration curve the recoveries are being calculated. When nC60 is spiked into distilled or ultrapure water and recovered using effective LLE protocols with LC/UV detection, recoveries are 92–101% compared to a nominal C60 standard in toluene (Table 3). LLE was also efficient at recovering 88–97% of nC60 spiked into raw and treated wastewater samples when compared to an extracted nC60 ultrapure-water standard [65].</p><!><p>Beside the recovery-calculation approach, the efficiency of the chosen extraction procedure itself can also influence the recovery, as not all extraction procedures are suitable for all sample matrices. In one case, SPE was not applicable for wastewater samples [65] because C60 adsorbed to the biomass and recoveries of 9–18% were obtained, while, in complex water matrices with less biosolids (e.g., surface water and groundwater), recoveries for LLE and SPE were comparable (75–81%) [41].</p><p>Similar to recovery from various environmental samples, recoveries of nC60 from biological samples (e.g., tissues, blood, and plasma) can be very efficient. Fullerenes spiked at 10 ng/mL into BSA solutions and porcine plasma were recovered at 94–100% using LLE [37]. Recoveries of nC60 from rat liver, spleen, and blood were 84–92% using a liquid-liquid tissue-extraction method into toluene with LC/UV detection (compared to an nC60 in ultrapure-water standard) [32]. Similarly, nC60 was recovered from an embryonic zebrafish homogenate at 90 ± 3% [60]. These extraction and detection protocols have improved in recent years, as an early attempt at extracting nC60 (solubilized with polyvinylpyrrolidone) from rat plasma achieved 12.5%, 7.7% and 0% recoveries using LLE into benzene, heptane-isoamylalcohol, and chloroform-methanol, respectively, with LC/UV detection [76]. SPE recovery was increased with a benzene elution, although it remained relatively low at 62.1% [76].</p><p>Two important caveats to successful LLE are the use of a destabilizing agent [e.g., Mg(ClO4)2] and avoiding sample evaporation to dryness when salts are present [37,60] (Fig. 2). Acetic acid is added in LLE approaches to prevent emulsion formation and can also be used as a destabilizing agent [41]. However, our personal experience has indicated that residual GAA in the toluene extract can suppress MS detection of C60 using APCI and a toluene/acetonitrile (55:45) eluent flowing at 1 mL/min. Bringing the samples to dryness mitigates GAA-induced interference by volatilizing GAA and yields higher recoveries (unpublished results), despite the previously described issues with evaporation to dryness [37] that is due to the presence of salts [64].</p><p>The use of SPE for extraction of spiked C60 has yielded more varied results compared to LLE. C60 recoveries were increased from plasma by using SPE instead of LLE, although toluene was not used as the extraction solvent [76]. Similar recoveries were obtained between SPE and LLE for recovering C60 from surface water and groundwater [41]. However, dramatically lower recoveries from wastewater samples were obtained using SPE, although this was partly due to the removal of nC60 during sample pre-filtering [65]. Low recoveries were acquired using SPE in tap-water samples, although recoveries were also low for LLE (32–42%) [64].</p><p>Even though SPE has not been successfully applied across a variety of aqueous and biological matrices by many researchers, it can be advantageous for treating larger sample volumes [64,65] with lower solvent use, a high degree of automation, and ease of field use compared to LLE [41], so it can lead to lower LODs and better reproducibility. We believe that SPE protocols may simply have not been sufficiently optimized for universal application, but it is possible that SPE is more susceptible to matrix variations (e.g., the presence of biosolids).</p><p>Currently, many questions remain regarding the appropriate SPE cartridge (type and size), sample matrix pretreatment (e.g., digestion of matrix components), conditioning-rinse-elution solvents, and sample loading rate, to name a few. Therefore, we propose the protocol described in Fig. 2 for extraction of nC60 from biological samples and detection of C60, based on the review of LLE techniques found in the literature (references in Table 2).</p><!><p>Fullerenes are unique chemicals with potential for many beneficial applications in biomedical and technological fields. If fullerenes are to become widely used, analytical techniques that are able to detect their presence in various environmental and biological matrices are indispensable, so we may conduct the necessary assays to determine their environmental occurrence, pharmacokinetics, and body burden, and ultimately evaluate the potential hazard for human health.</p><p>Past studies have illustrated the implications of different nC60-preparation methods on the colloidal properties and have evaluated the impact of transfer solvents in ecotoxicological assays. In-depth material characterization (based on, e.g., FFF, TEM, or zeta-potential analysis) in the wide spectrum of environmental and biological matrices will no doubt be the subject of many future investigations.</p><p>Our ability to quantify fullerenes in biological samples largely depends on our ability to extract these compounds from these complex matrices. LLE has been optimized to isolate fullerenes from a wide range of matrices using a simple consensus protocol. LLE is therefore currently considered the most robust method for fullerene extraction. However, as LLE is now widely used for the extraction of fullerenes from small-volume samples only, a drawback will ultimately be the LOD (and the large volumes of solvents that LLE requires). For this reason, we see an opportunity for (automated) SPE to become a crucial route to allow detection of fullerenes in "real-world" concentrations, especially for population studies, where hundreds of small volumes can be pooled, concentrated, and analyzed using LC/UV or LC/MS.</p><!><p>The three phases of a fullerene-exposure experiment. In phase 1, the C60 solution and nC60 aggregate suspension are prepared, and both can be used to administer fullerenes to a biological matrix. In phase 2, the fullerenes are spiked in a biological matrix and then extracted using liquid–liquid extraction (LLE) or solid-phase extraction (SPE). The nC60 dispersion in ultrapure water (UPW) is used as a reference for the extraction efficiency. In phase 3, the C60 and nC60 stocks are characterized and the recovery from the biological matrix is determined by comparing to the calibration standard of C60 in toluene or the reference solution of C60 after exhaustive LLE. TGA: thermogravimetric analysis; TOT: thermal/optical transmission analysis; TOC: total organic carbon analysis; TEM: transmission electron microscopy; AFM: atomic force microscopy; FFF: field flow fractionation; DLS: dynamic light scattering; PALS: phase analysis light scattering. The asterisk indicates the same C60 solution.</p><p>Consensus liquid-liquid extraction protocol based on a comprehensive review of the literature on quantification of nC60 fullerenes in various biological and environmental matrices.</p><p>Overview of preparation approaches for nC60-stock suspensions and characteristics of the colloidal fullerenes</p><p>ben, benzene; CLF, chloroform; Conc., concentration; dist., distillation; DMSO, dimethylsulfoxide; EtOH, ethanol; Evap., evaporation; LC-UV, online ultraviolet spectroscopy following liquid chromatography; LLE, exhaustive liquid-liquid extraction; 13C NMR, nuclear magnetic resonance targeting 13C; NOM, natural organic matter; Nominal, weighed mass per added volume of solvent; NR, information was not reported; MS, mass spectrometry; PTFE, polytetrafluoroethylene; PVP, polyvinylpyrrolidone; rotary evap., rotary evaporation; SDS, sodium dodecyl sulfate; son., sonication; TGA, thermogravimetric analysis; THF, tetrahydrofuran; tol, toluene; TOC, total organic carbon analysis; UF, ultrafiltration; UPW, ultrapure water (18.2 MΩ/cm); UV, ultraviolet spectroscopy measurement in 1 cm cuvette.</p><p>under anoxic conditions;</p><p>after exhaustive LLE the concentration was 60% of the nominal concentration;</p><p>filtration prior to addition of water;</p><p>more than 75% of C60 remained in water, a yellow film was observed sticking to the glass wall of the vial;</p><p>pH 10.25–3.75;</p><p>at 10 μM NaCl pH 7.0;</p><p>in the presence of 50 and 5 mg/L-C NOM, respectively.</p><p>Published nC60 extraction and quantification protocols and respective limits of detection in various aqueous matrices</p><p>APCI, atmospheric pressure chemical ionization; BSA, bovine serum albumin; Conc., concentration; DMSO, dimethyl sulfoxide; ESI, electrospray ionization; Evap., evaporation; Filtr., filtration; GW, groundwater; hex, hexane; LOD, limit of detection; MS, mass spectrometry; NA, not applicable; NOM, natural organic matter; NR, information is not reported; PVP, polyvinylpyrrolidone; Rec., reclaimed; SAA, same as above; SDS, sodium dodecyl sulfate; Sec. effluent, secondary effluent; SW, surface water; THF, tetrahydrofuran; TIC, total ion count; tol, toluene; UPW, ultrapure water; WW, wastewater; WWTP, wastewater treatment plant.</p><p>five different wastewater matrices;</p><p>instrument detection limit, 0.05 ng at S, N=2;</p><p>Detection limit, 0.0002 ng and quantification limit, 0.001 ng;</p><p>Methodology for whole-method LOD determination is not reported;</p><p>After inhalation and distribution among the organs via blood circulation;</p><p>2.22 mg/m3 C60 at an inhalation rate of 140 mL/min;</p><p>instrument detection limit = 0.1 ng of C60 injected</p><p>Approaches for recovery determination of nC60 extraction from aqueous and biological matrices</p><p>aqu/nC60: nC60 suspension prepared without transfer solvent; BSA: bovine serum albumin; Evap: evaporation; GW: groundwater; LLE: liquid-liquid extraction; NA: not applicable; ND: not detectable; NOM: natural organic matter; NR: information not reported; PVP: polyvinylpyrrolidone; rel. PA: relative peak area; SAA, same as above; SPE: solid-phase extraction; SW: surface water; TGA: thermogravimetric analysis; THF: tetrahydrofuran; tol: toluene; Tap W: tap water; UPW: ultrapure water; WW: wastewater; WWTP: wastewater-treatment plant.</p><p>Recovery is calculated compared to the extraction from UPW using the simplified LLE method (same as samples);</p><p>since no matrix interferences were observed upon standard addition, 0.1–150 ppb, quantification based on ratio of analyte peak area to the 13C60 peak area;</p><p>the ratio of the amount extracted to the amount that was spiked;</p><p>C60 sorbed to particulates were filtered out;</p><p>the C60 peak area/total peak area = ≥98% (wavelength not reported);</p><p>extraction recoveries can be best determined through exhaustive suspension extraction to determine the total analyte mass in suspension;</p><p>recovery was verified through comparison with a thermogravimetric procedure.</p>

PubMed Author Manuscript

Determining factors in the growth of MOF single crystals unveiled by in situ interface imaging

Crystal interface evolution is monitored by super-resolution technique during the growth of small MOF crystals into larger ones, in solution. The reaction orders for metal ion and organic linker are determined by isolation method and found to be independent from the corresponding MOF chemical formulas. This leads to the proposal of a new mechanism for MOF growth, involving the assembling and fragmentation of secondary building units, followed by fragment accumulation in a reversible transition layer at the interface of the MOF crystal.

determining_factors_in_the_growth_of_mof_single_crystals_unveiled_by_in_situ_interface_imaging

INTRODUCTION<!>ionic<!>RESULTS AND DISCUSSION<!>MOF crystal growth rate obtained by super-resolution technique<!>Article<!>Organic linker as the determining factor for MOF crystal growth<!>Impact from reaction temperature on MOF growth<!>Article<!>Linkage and dissociation of molecular building blocks at MOF crystal interface<!>Article<!>Conclusions<!>EXPERIMENTAL PROCEDURES<!>Materials availability<!>Data and code availability<!>DFM platform<!>In situ interface imaging for MOF crystal growth<!>Kinetic study for the quantitative analysis of MOF crystal growth determining factors<!>Dynamics of crystal interface upon crystal growth interruption<!>SUPPLEMENTAL INFORMATION

<p>The behavior of crystal interface is always intriguing, not only because it is the frontier for various substance exchange and chemical reactions but also due to its dynamic nature, where the composition and structure evolve over time. [1][2][3][4] Such variety and dynamics make it extremely challenging to characterize crystal interface, especially in a liquid environment. 5 In this study, we use a dark-field microscope (DFM) to trace the evolution of metal-organic framework (MOF) crystal interface during their solvothermal synthesis (Scheme 1); attempting to understand the behavior of MOF crystal interface in solution, we investigated (1) the linkage and dissociation of the building blocks, (2) the activity at specific crystal facets, and (3) the evolution of crystal facets during crystal growth. A super-line localization method was developed for the quantitative measurement of crystal facet progression in correlation with the metal ion, organic linker, and the reaction temperature. Based on this quantitative analysis, kinetic parameters of crystal growth, including reaction orders and activation energy, were derived for MOFs with different structures and crystal shapes. The determining factors for the interface evolution of MOF crystals are also discussed, providing new insight into their growth mechanism and guidance for their industrial production.</p><p>MOFs, a fast-developing class of network crystals, are chosen here as an ideal subject for the in situ crystal interface study. These frameworks, formed by the coordination of organic linkers to metal ions, offer molecularly defined pore spaces within which matter can be manipulated and controlled, exhibiting excellent storage, separation, and conversion properties. [6][7][8][9][10][11] Different from molecular crystals, 12</p><!><p>The bigger picture</p><p>Interface is an attractive topic in modern chemistry. So far, little is known on the evolution of crystal interface, especially at the stage where small nuclei grow into large single crystals. Here, a nondestructive optical method is used to outline the crystal interface of MOFs, a typical class of porous crystals formed by coordination bonds, during its growth in solution. The progression of the crystal interface is quantified and correlated with the concentration of metal ions and organic linkers, as well as the reaction temperature. The crystal interface itself is also identified as an additional factor to crystal growth, indicating the existence of a reversible transition layer. These findings provide humble steps toward a full understanding of the crystal interface behavior. Differences are found in the determining factors for the growth of these MOFs. For all MOFs, the reaction temperature is important for their crystal size expansion, as reflected in the activation energy. Separate influence is determined from the compositional metal ions and organic linkers, revealed by the apparent reaction orders. The values are found to be inconsistent with the ratio of metal ion to organic linker in the formula of the corresponding MOF. This is in stark contrast to the prevailing view that these molecular building blocks are correlated during MOF growth. 25,26 A previous attempt successfully investigated the separate roles of metal ions and organic linkers by their alternative introduction, using layer-by-layer synthesis. 27 In this study, these building blocks were mixed in the same solution, under conditions commonly applied to the solvothermal synthesis of MOFs. A control variable method was used to extract the impact from each building block, with the other conditions fixed. Different reaction orders for metal ions and linkers were observed in the growth of different types of MOFs, providing an experimental basis for the discussion of a mechanism for MOF crystal growth. A new mechanism is proposed for the growth of Cu-MOF-2-BDC and -NDC crystals, which involves the assembling and fragmentation of secondary building units (SBUs), followed by the accumulation of the fragments at the crystal interface. It is also interesting that unprecedented negative reaction orders for organic linkers were identified in the case of HKUST-1 and Cu-MOF-74.</p><p>Furthermore, the use of flow cell in this study allows for real-time pausing and resuming of the growth reaction. A retreat and recovery of MOF crystal interface are observed upon interruption to interface evolution of Cu-MOF-2-NDC. This can be associated with the linkage and dissociation of the molecular building blocks, where a reversible transition layer is likely to exist between the MOF crystal bulk and the solution environment. The thickness of the transition layer was measured, and it decreased as the crystal grew larger. Based on this observation, the critical size is estimated, at the point of which crystalline domain is formed as the core, marking the emergence of MOF crystal.</p><!><p>Tracing the growth interface of MOF single crystal by in situ DFM The evolution of MOF growth interface was traced in situ by DFM. Unlike atomically precise methods, such as electron microscopes and atomic force microscopes (AFMs) that were usually required to obtain insight at the molecular level, 17,25,[28][29][30][31] optical methods are rarely applied to the study of crystal growth owing to their limited spatial resolution. Here, we demonstrate that accessible, facile, and nondestructive optical methods, such as DFM, 32,33 are also suitable for the in situ study of crystal interface, where an appropriate setup was designed (Scheme 1; Figure S7). In this regard, three aspects are critical for the accuracy of the measurement: (1) the use of flow cell, (2) specific optic geometry of the DFM, and (3) precise temperature control. These three aspects were demonstrated in the visualization of Cu-MOF-2-NDC crystal growth interfaces as an example (Figure 1). First, a flow cell was designed to provide a stable concentration of metal ions and organic linkers during the period of MOF crystal growth (Scheme 1A; Figures S8, S9, and S75). Based on multiple trials, an appropriate flow rate was chosen (Figure S72) to minimize the difference in molecular diffusion coefficient for metal ions and linkers (Figure S73), and to avoid concentration change of these molecular building blocks as they were consumed along the crystal growth process in traditional closed systems. 34,35 Slip boundary conditions and shear stress might exist in the flow cell, but these have little influence on MOF crystal growth in this study, based on fluid flow simulations and experimental results (Figure S74-S77). With the precise control of flow rate, the progression profile of MOF crystal interface was kept linear against time, making it much easier to extract the contributions from metal ions and organic linkers. This flow cell design is also different from microfluidic systems for the synthesis of MOFs by microdroplet flow (MF) reaction, where the resulting MOF crystals keep on moving as carried by the liquid flow. 36,37 Here, the majority of MOF single crystals remained stationary during the entire growth process (Figure S74), with their positions being accurately calibrated by inert Au Article nanoparticles as position markers under the same camera view (Figure S28). This allows for real-time tracing of specific crystal interface for every crystal in the same batch.</p><p>Second, the growth of multiple individual MOF crystals in the same view (200 3 200 mm) was tracked simultaneously by in situ DFM camera (Figures 1A and 1B), with a frame rate of 20 Hz through a period that varied from 1,680 to 3,060 s to give a typical total frame of 61,200 for the analysis (Figures 1A and 1B). Conical incident light was used to illuminate the MOF crystals in DFM camera (Schemes 1B and 1C; Figures 1A, 1B, and S7). The optic geometry of the microscope kept the reflected and refracted light away from the camera, leaving the crystal center dark in the DFM image, while only the scattering light at the edges of each crystal was captured by the camera in the microscope, outlining the crystal interface (Schemes 1B and 1C; Figure S7). This unique feature of DFM allows for focusing on the crystal interface at specific crystal facet throughout the crystal growth process, which is different from classic bright-field optical microscope where the areas of crystal body and interface are usually obscured. The selective highlighting of the crystal interface combined with the super-resolution technique [38][39][40] provide accurate positional information of the crystal edges with a typical trace resolution of 4.1 nm (Figures 1C-1E and S35). The reliability of our data is based on an accurate statistical analysis taking into account the subtle difference in the individual crystal's local chemical environment (Figures 2 and 3). It is worth noting that both the positional information of the crystal and the movement of the crystal facet are accurately measured for each individual crystal in the same batch, where multiple crystals are growing simultaneously. This is distinct from the diffraction methods such as X-ray, [41][42][43] where the crystal size is averaged out from multiple crystals without knowing the specific position of each crystal in the solution.</p><p>Last but not least, the temperature of the flow cell was precisely controlled by the power of the electrical heating mantle placed beneath the flow cell and was calibrated by an infrared camera (Schemes 1A and 1B; Figure S8). This avoided the potential influence from heating of the solution by incident light (Figure S36). The use of a temperature control system here guarantees a stable reaction temperature, ranging from room temperature to 90 C, during the crystal growth process, which is critical for the linearity of the progression profiles of crystal facets. This entire optic setup was applied to monitor the growth of all five MOFs in this study, demonstrating the universality of the DFM platform to study crystal growth (Figures S10-S14).</p><p>Prior to the image processing, other structure characterizations such as scanning electron microscope (SEM) and X-ray diffraction were performed on these MOFs (Fig- ures S15-S24). The detailed morphologies of the crystal facets are shown in SEM images, in good accordance with the crystal interface outlined in DFM images (Figures 1B, 3D-3F, and S15-S19). Sharp peaks were observed in the powder X-ray diffraction (PXRD) patterns of these MOFs, confirming their crystallinity and phase purity (Figures S20-S24). The crystal facets of Cu-MOF-2-NDC were indexed by single-crystal X-ray diffraction analysis of the square-shaped crystal. Most of the crystals observed in the camera view were laying on the (001) facet, and the crystal interface moved along [010] and [100] directions extending the edges of the square (Figure S25).</p><!><p>The growth rate for a MOF crystal was derived from the progression profile of its crystal edges. The distance between each crystal edge and the body center was monitored as a function of time (Figures 1C, 1D, and S27), using the square-shaped ll Chem 8, 1637-1657, June 9, 2022 1641</p><!><p>Cu-MOF-2-NDC crystal as an example. Since only the crystal interface was highlighted in each DFM image (Figures 1B and 1C), the precise location of the crystal edge could be determined, applying Gaussian fitting based on the corresponding contour map (Figures 1C, 1D, and S27), and marked as the position of the resulting peak. This is termed super-line localization, a super-resolution technique similar to the typically super-point localization used for the tracking of a single molecule and nanoparticle in a solution. [38][39][40] The intensity of the peak, reflecting the brightness of the crystal edge in the DFM image, might be related to the thicknesses of the crystals, due to the good accordance with the increase in the height profile of MOFs along the [001] direction in their ex situ AFM images taken at areas identical to the DFM images (Figure S26). However, the increase in thickness did not affect the accurate tracing of the crystal edge, since it was perpendicular to the evolution plane of the crystal edges along [010] and [100] directions highlighted in the DFM camera.</p><p>The evolution of crystal edges was accurately traced by applying super-resolution technique, where adjacent 10 frames were merged (accumulated over 0.5 s) to give high-quality images, albeit of a relatively low pixel resolution, 400 nm. The progression profiles of four edges in the same crystal were converged to one crystal growth profile (Figures 1E, S29, and S30), where the subtle drift of the entire crystal along the growth process was corrected. These allowed for achieving a typical trace resolution of 4.1 nm, with precision at unit cell level (Figures 1E and S35). When the temperature and the concentrations of metals and organic linkers were fixed in the flow cell, a linear relationship was observed between edge extension (Dl) and reaction time (t) (Figure 1E). Since the increase in the crystal thickness ([001] direction) was negligible in comparison with the crystal size extension along [010] and [100] directions (Figure S38), the crystal growth rate (r) can be represented by the edge extension (Dl) in the square plan, consistent with the analysis applied to a molecular single crystal. 44 Therefore, the slope of the crystal growth profile gave the crystal growth rate (r), dividing edge extension (Dl) by the reaction time (Dt) (Figure 1E):</p><p>The initial point of the crystal growth profile, t 0 , was chosen when there was a sufficient amount of crystals reaching suitable sizes (1-2 mm) with their crystal interfaces visible (Figures 1A and 1B). This allowed us to only focus on the evolution of crystal interface at the stage of crystal size extension, rather than at the nucleation stage, at the very beginning, where the crystal interface was not distinguishable from the body in DFM images.</p><p>Growth profiles of four Cu-MOF-2-NDC single crystals in the same camera view exhibited similar growth rates, 1.26 G 0.06 nm$s À1 (Figure 1E). The difference in their crystal sizes, possibly originating from their different nucleation times, did not influence their growth rates; thus, the kinetics of MOF crystal growth was independent from the crystal size during the test period of 1,800 s. The uniformity in their growth rates also demonstrated the precision in temperature control and the advantage of using flow cell. These crystals were always in contact with fresh reactant solution of constant concentration (Figures S8 and S9), minimizing the difference in local chemical environment at crystal interface of various crystals.</p><p>During the growth process, the feed solution was flushed twice, breaking the growth profile into three sections, reflected in the small steep steps at 422.45 and 851.55 s (Figure 1E). In this test, a solution with an identical composition was used, where all three sections exhibited an identical growth rate for each crystal, showing that the mechanic change in solution did not affect the crystal growth rate. In the test of other MOF crystals with different morphologies, linear growth was also observed (Figures S31-S34). These tests paved the way for the study on crystal growth kinetics.</p><p>It is worth noting that the subtle difference in growth rates among individual crystals was clearly observed here. This valuable information on crystals is largely missing in space-averaged characterizations, such as X-ray diffraction and light scattering. We found that the growth rate of MOF crystal was slightly affected when there were other crystals nearby (Figure S41). The unambiguous identification of these crystals in the DFM image allowed us to exclude them from the statistical analysis.</p><p>Metal ion as the determining factor for MOF crystal growth Kinetic parameters, such as apparent reaction order and activation energy, are critical to unveil the determining factors for MOF crystal growth. Systematic variation of the reaction conditions was introduced in the flow cell by changing the reaction temperature and feed solution with different concentrations of metal ion or organic linker during the continuous growth of MOF crystals. The reaction orders of metals and organic linkers were obtained by isolation method, where the concentration of one component was varied systematically with all other conditions fixed. 45 The values of reaction orders were not restricted to integer numbers here and were therefore closely reflecting the actual reaction kinetics. We noted that the reaction orders for molecular building blocks were successfully obtained for the growth of extended covalent networks of a different kind, covalent-organic frameworks, 46,47 but a different method, known as the integration method, was applied. In those attempts, the experimental growth curves were fitted with existing models of integer reaction order to give an overall estimation, whereas the separate role of each building block remained unknown. The isolation method used here does not rely on any existing model, where the reaction order can be non-integer, and the impact from each molecular building block is also given.</p><p>In the case of Cu-MOF-2-NDC, feedstock solutions of Cu 2+ concentrations in the range of 0.015, 0.025, 0.035, 0.045, and 0.055 mol$L À1 were added sequentially into the flow cell (Figure 2A), each for a duration of 408.6 s corresponding to 8,172 frames. During the switching, the flow cell was flushed with the new feedstock shortly (30 s) with a much higher flow rate (2,000 mL$min À1 , 100 times that of the experimental rate used for crystal growth, 20 mL$min À1 ) to achieve steady concentration prior to recording (Figure S8). Each section of MOF crystal growth at a specific Cu 2+ concentration exhibited a nearly linear curve with a unique slope, based on which the growth rate during this section was obtained. As the Cu 2+ concentration increased gradually, faster growth rate was observed. The log values of these growth rates were plotted against the log values of their Cu 2+ concentrations, where a linear fitting was performed with the satisfactory regression (Figures 2D and S46). The slope of the resulting line was the apparent reaction order of Cu 2+ (m) for this MOF crystal.</p><p>Each MOF single crystal was traced separately to provide a unique value of reaction order. Multiple parallel experiments were performed under identical conditions, generating data of multiple (typically 36-75) individual crystals. Similar results were obtained to provide sufficient samples for statistical analysis, thus giving a profile of reaction orders (Figure S47). The peak position of the Gaussian fitting represented the reaction order of 1.09 for Cu 2+ (Figure 2G). A narrow full width at half maximum (FWHM) was observed in the Gaussian distribution histogram, demonstrating the accuracy of the statistical analysis, where the subtle difference ll Chem 8, 1637-1657, June 9, 2022 1645</p><p>Article in the apparent reaction orders between different crystals might originate from their structure features and local chemical environments during the growth process. This allows for the unveiling of apparent reaction order of metal ion independently and accurately for the crystal growth of a MOF. This previously unknown aspect also reveals that the apparent reaction orders for different MOF single crystals are not the same, and not necessarily integers.</p><p>Furthermore, this in situ interface imaging method was also applied to study the apparent reaction orders for other MOFs, including 1D MOF with rod shape, Cu-MOF-74 (Figures 3A, 3G, and S42); 2D MOFs with square shape, Co-ZIF (Figure S49) and Cu-MOF-2-BDC (Figures 3B, 3H, and S44); and 3D MOF with octahedron shape, HKUST-1 (Figures 3C, 3I, and S51). In the case of Cu-MOF-2-BDC, isoreticular to Cu-MOF-2-NDC, the apparent reaction orders for Cu 2+ was 0.81, in the same ballpark of those for Cu-MOF-2-NDC, indicating similar reaction kinetics. In the case of HKUST-1, the apparent reaction order for Cu 2+ was 1.10, close to that of Cu-MOF-2-NDC, 1.09, and Cu-MOF-2-BDC, 0.81 (Figure 3I). All three MOFs were composed of Cu 2+ paddle-wheel-shaped SBUs (Figures 3B, 3C, and S5), unveiling a possible correlation between the reaction order and the type of SBU.</p><!><p>The apparent reaction order for the organic linker (n) was also obtained, by systematic variation of the linker concentration, while fixing the temperature and concentration of metal ions (Figures 2H, 3J-3L, S43, S45, S48, S50, and S52). In the case of Cu-MOF-2-NDC and Cu-MOF-2-BDC with isoreticular crystal structure, the reaction orders for NDC and BDC linkers were determined as 0.49 and 0.43, respectively (Figures 2H, 3K, S45, and S48). Such proximity in apparent reaction orders indicated that the growth mechanism is similar for these two 2D MOFs with almost the same crystal structure. The reaction orders of metal ion and organic linker of another MOF with a 2D-layered structure, Co-ZIF, were also measured, as 2.92 and 2.12, respectively, which were distinctively different from those of Cu-MOF-2-NDC and Cu-MOF-2-BDC (Figures S49 and S50). Such difference might originate from the types of SBUs as well as the organic linkers (Figures S2-S4). Nevertheless, both the reaction orders of metal ion and organic linker were positive for all three MOFs with a 2D structure, indicating that both constituents are the determining factors for their crystal growth.</p><p>The reaction order of organic linker, however, was not always positive. In the study of HKUST-1 with a 3D morphology, an apparent negative reaction order of the tri-topic organic linker BTC (trimesic acid), À1.14, was observed. This does not mean that the crystal stopped growing or even dissolved upon the addition of BTC linker (Figures 3L and S52) but rather that when a higher concentration of BTC was introduced into the flow cell, the MOF crystals grew slower. This is counterintuitive to the Le Chatelier's principle where an increase in the amount of starting materials will push the reaction toward the products. This is also different from the prevailing view that both metal ions and organic linkers are responsible for driving crystal formation. In order to double check whether this was an experimental error, more than three parallel experiments were performed, yielding the same negative reaction order (Figure S53). In addition, experiments were performed with gradually decreasing concentration of linkers (Figure S54), instead of the increasing sequence (Figures 3L and S52). Still, a negative reaction order was observed. Both experiments confirm that an increase in the concentration of organic linkers will lead to a slowdown of crystal growth in the case of HKUST-1. It is worth noting that the SBU for HKUST-1 is identical to that in The growth of another Cu MOF, Cu-MOF-74 with a 1D pore structure, was also studied. A positive reaction order, 0.34, was observed for Cu 2+ , whereas the apparent reaction order for organic linker DHTP (2,5-dihydroxyterephthalic acid) was found slightly negative (À0.14) (Figures 3G, 3J, S42, and S43). The absolute value of the reaction order of organic linker for this 1D MOF was relatively small, indicating that the growth of this MOF is less sensitive to the concentration of DHTP linker. This might be related to the shape of 1D SBU in Cu-MOF-74 (Figures 3A and S1), where the formation of Cu containing SBUs is driving the growth of the MOF. In fact, the crystal is also needle shaped, propagating along the c axis, which is the same direction of SBU extension.</p><p>All these results show that the metal ion and organic linker play separate roles in the crystal growth of MOFs, different from the prevailing view that these components are generally correlated. [48][49][50] With the knowledge of reaction orders, the rate constant can be derived for each MOF single crystal, which is in good accordance with the experimental result (Figure S46B; Table S8; supplemental information).</p><p>For MOFs with different structures, compositions, and morphologies, the determining factors for their crystal growth vary (Table S7). Here, for three MOFs with a 2D-layered structure, i.e., Cu-MOF-2-NDC, Cu-MOF-2-BDC, and Co-ZIF, both metal ions and organic linkers are responsible for their crystal growth; for MOFs with a 1D pore structure, Cu-MOF-74, the metal ion favors the extension of SBUs and hence the development of a needle-shaped crystal. As for the MOF with a 3D structure, HKUST-1, the organic linker slowed down the growth rate, whereas the metal ion accelerated the crystal growth. For all five MOFs, the values of apparent reaction orders for metal ions and organic linkers are different from the ratio in the corresponding MOFs, further demonstrating the independent role of these constituents in MOF crystal growth.</p><!><p>In addition to the study of reaction orders, the other critical aspects of MOF crystal growth kinetics, the activation energies (E a ), were also assessed (Figure 2C, 2F, 2I, 3M-3O, and S55-S59). For all five MOFs, as the reaction temperature increased gradually, the growth rates became faster in the flow cell, when the concentrations of metal ion and organic linker were fixed (Figure 2C). Based on the Arrhenius equation, systematical and sequential variation of the reaction temperatures revealed the accurate activation energies for Cu-MOF-74, Cu-MOF-2-BDC, Cu-MOF-2-NDC, Co-ZIF, and HKUST-1 as 115.6, 59.9, 82.1, 73.6, and 71.3 kJ$mol À1 , respectively (Table S9). These values were a result of the collective and statistical analysis of 32-163 crystals, where the activation energy of each individual crystal was obtained separately (Figures 2I, 3M-3O, and S55-S59). This is different from the activation energies obtained based on an average size of multiple crystals in the solution. 31,42 The study of determining factors provides valuable guidance for the production of MOFs. The growth rates of 1D, 2D, and 3D MOFs can be accelerated by 1.7-5.6 times through adjusting the reactant concentration and reaction temperature. In the case of Cu-MOF-74, a rapid growth of single crystal was achieved at the rate of 10.2 nm$s À1 at 68.7 C, with Cu 2+ and linker concentrations of 0.10 and 0.02 mol$L À1 , respectively, which is among the fastest crystal growth rate for MOF production (Table S10). Moreover, the acceleration of crystal growth was also verified in a closed system for Cu-MOF-2-NDC, where the growth rate of crystal was ll Chem 8, 1637-1657, June 9, 2022 1647</p><p>Article effectively increased by rational application of the corresponding determining factors (Figure S60). This acceleration matched well with the theoretical growth rate calculated from the experimental results of apparent reaction orders for metal ions and organic linkers, demonstrating the potential application of these factors in the precise control of crystal size and growth rate for the industrial production of MOFs.</p><p>Reversible transition layer at the crystal interface Beyond the impact from the external solution environment, including the concentration of building blocks and the reaction temperature, is there any other factor influencing the crystal growth of MOFs? With this question in mind, we tested the dynamic behavior of the crystal interface by pausing and resuming the crystal growth during the solvothermal synthesis of Cu-MOF-2-NDC with fixed temperature (Figure 4A). Specifically, the crystal growth was paused, after linear growth for 300 s, by introducing bare solvent without metal ions or organic linkers at the same flow rate of 20 mL$min À1 , for the following 300 s, until the position of the crystal interface became steady. It was then resumed by switching back to the original reaction solution to give a new linear growth curve starting from 700 to 1,050 s (Figures 4A and S61). Such interruption was induced 4 times during a continuous test of 3,200 s in total.</p><p>Two important observations were made. First, the crystal interface quickly retreated right after the environment was switched to bare solvent, lasting about 50 s for every interruption (Figures 4B and 4C). The retreat of the interface is unexpected, with an obvious movement of position ranging from 10 to 40 nm, which is way beyond the trace resolution of 4.1 nm. Such retreat of the interface was observed at all four edges of the square-shaped single crystal simultaneously, ruling out the possible drift of the entire crystal (Figure S61). Even when the crystal growth was stopped, the interface would remain roughly at the same position, rather than moving backward to the crystal center. The retreat of the interface indicates that the interface is dynamic, in stark contrast to the stable crystal bulk. Different from direct attachment in classic models, 45 the molecular building blocks, metal ions and linkers, do not directly become part of the highly ordered MOF unit cell at the interface. Instead, they are likely to form a dynamic transition layer with less ordering, which mediates the crystal growth.</p><p>The other observation was also unexpected. Here, the growth rate slowed down to 1.80, 1.53, 1.39, and 1.21 nm$s À1 after the 1 st , 2 nd , 3 rd , and 4 th interruption, respectively (Figures 4D and S63), compared with the observations for unaltered crystal growth rate without interruption (Figure 1E). This infers that the transition layer is another factor to influence the growth of MOF crystals. When it is interrupted, the local composition and/or the structure of the transition layer might change; thus, the growth continuum is broken at the crystal interface. Notably, the influence from the transition layer is relatively small in comparison with that from the environmental factors, such as the concentration of metal ions and organic linkers, as shown in the experiments with systematic variation of their concentrations after crystal growth interruption (Figures S65 and S66). The overall trend in the change of crystal growth rate is still dominated by these environmental factors (Figures S67-S69).</p><p>Nonetheless, with the high trace resolution offered by this interface imaging method, the influence from the possible transition layer was still visible in all interruption experiments (Figures 4A, 4B, and S63). These observations were not accidental; they occurred to each crystal in the batch and in every repeated experiment ll 1648 Chem 8, 1637-1657, June 9, 2022</p><!><p>(Figures S63 and S64). Results from additional experiments also ruled out the possibility of systematic artifact (Figures S65 and S66). The accuracy in trace resolution allowed us to analyze the retreat distance (Figure 4B) of the crystal interface quantitatively, which might provide valuable information to estimate the thickness of the possible existing transition layer. Regardless of the sequence of change in concentration of the reaction solution, a similar trend was observed (Figures S65 and S66).</p><p>The retreat distance of the interface was the longest after the first interruption, when the crystal size was small, and this distance became shorter after each following interruption, as the crystal became larger. This trend was also in good accordance with the tests in which identical solution concentration was applied after each 4F, S63, and S64). Therefore, a correlation emerges between the retreat distance of the interface (DH) and the size of the crystal, represented by the distance from crystal center to the edge (l), which is one half of the edge length (a) in a square-shaped crystal (Figures 4F and S70). This infers that the influence from the transition layer becomes smaller as the crystal size becomes larger, where the ordering of the building blocks in the crystal bulk is strengthened and reflected in the increased crystallinity. Furthermore, a quantitative linear fitting was performed between the interface retreat thickness (DH) and the size of the crystal (l) (Figure 4F), giving the critical size for the emergence of crystalline domain for the MOF crystal. In the case of Cu-MOF-2-NDC, the crystal size is estimated to be 150 nm. This observation is in good accordance with the existence of the amorphous cluster before the emergence of crystalline domain in the growth of other MOFs, and the value of critical size here is also in the same ballpark to the size of the amorphous cluster in a previous study. 18</p><!><p>The above new knowledge is directly obtained from experiments, which provides a solid base for the discussion on the behavior of molecular building blocks at the crystal interface during MOF growth. Both the retreat and recovery of the interface upon interruption (Figure 4A), as well as the negative apparent reaction order for organic linkers BTC (Figure 3L), indicate that the conversion from molecular building blocks to crystal bulk is not an elementary reaction. Although there are only limited examples of elementary reactions, the multi-step nature of MOF crystal growth is confirmed here, which is distinctively different from that of the molecular crystal or metal salt with relatively simple components. 12,13 A reversible surface coordination process is likely to take place during MOF growth, in the transition layer described here, at the crystal interface. Such transition layer is active for substance and energy exchange with the outer environment and also responsible for the regulation of building blocks to form highly ordered crystal bulk (Figure S71). This reversible surface coordination process involves the accumulation of growth unit and the dissociation of the crystal bulk (Figure S71). The molecular building blocks, not yet mature to become a solid part of the crystal, can still be dissociated from the transition layer, once the feedstock solution is removed and switched to bare solvent (Figures 4A and S63). Upon switching back to the original growth solution, where new building blocks are supplied from the environment, a new balance is established at the crystal interface to sustain the crystal growth (Figure S71).This phenomenon also reflected the difference between the transition layer and the crystalline core in regard to their solubility in pure solvent; whereas the transition layer can be dissolved, the crystalline core remains almost unaltered in the flow of pure solvent. However, this feature is not exactly the same as the dissolution behavior of classic ionic solids as described by Madelung energy. The underlying reason is similar though, where the binding energy between the linker and metal containing building blocks in the crystalline core is much higher than that in the transition layer.</p><p>The existence of a reversible transition layer provides a reasonable explanation for both the crystal growth interruption and the negative reaction order. The rate of accumulation for metal ion and organic linker to form growth units and the rate of their dissociation from the crystal bulk can be separately tuned by reaction conditions, since these might involve two different steps in the crystal growth process (Figure S71). When the accumulation is faster than dissociation, the crystal interface advances forward as reflected in the increase of the crystal size (Figure S71).</p><!><p>Otherwise, the crystal interface moves backward as the building blocks are dissociated from the transition layer, reflected in the interface retreats (Figures 4A and S62B). Further experiments were also performed where the reactant solutions were flowed at higher rates to wash the crystal interface (200 mL$min À1 as shown in Figure S62A). The crystal interfaces extended continuously as expected, rather than the growth interruption. It is worth noting that the retreat and recovery of the interface are only observed when the crystal growth is interrupted by washing with pure solvent. In the continuous growth of crystal, without interruption, such a reversible transition layer might also exist. It provides a steady equilibrium for the size extension of the MOF single crystal (Figure 1E), instead of reversing the movement of the interface. In the case of HKUST-1, as the concentration of BTC linker increases gradually, the rate of building block dissociation might be accelerated more than the rate of accumulation; thus, the crystal growth is slowed down, manifested in an apparent negative reaction.</p><p>Possible multi-step growth mechanism for Cu-MOF-2-BDC and -NDC The apparent reaction orders accurately measured also allow us to consider the possible reaction mechanism. Here, we use Cu-MOF-2 as an example for discussion.</p><p>If the growth mechanism were a classic monomer-by-monomer addition (Figure 5A), where the metal (Cu 2+ ) and linker (L, L = BDC or NDC) were added to the MOF interface in an alternating manner, the ratio between the corresponding apparent reaction order would have been 1:1, according to the MOF formula, depending on how many pairs of metal ions and linkers are connected at the same time. However, the experimental values were 1 and 0.5 for Cu 2+ and organic linker, respectively, ruling out such possibility. Another possible mechanism involves the formation of a cluster of Cu 2 L 4 , corresponding to the underlying SBU for Cu-MOF-2. 22,51 It was well established that the SBU, coordination complex of linkers with metal ions in the center, was critical for the design of MOFs, 6,7 as well as the nucleation process of MOF growth. 17,18 Is it possible that the metal ion and linker coordinate to form Cu 2 L 4 , followed by its accumulation at the MOF crystal interface (Figure 5B)? If this were the case, the apparent reaction orders would have been 2 for Cu 2+ and 1 for organic linkers for Cu-MOF-2, still different from the experimental data. Thus, it was unlikely that the paddle-wheel-shaped Cu 2 L 4 SBU aligned directly at crystal growth interface during MOF growth process.</p><p>The analysis of the above two possibilities encouraged us to further propose a third mechanism (Figure 5C). This involves both the formation and fragmentation of Cu If this mechanism proves to be a major possibility, Cu 2 L 4 SBU plays a structural directing role, consistent with previous studies of MOF growth at the nucleation stage, 17,18 but it is the fragmentation product, CuL 2 , that functions as the important species bridging the formation of SBU and the growth of MOF. Given all of the above, we note that there might be other possible explanations or mechanisms for MOF crystal growth. The proposed mechanism merely provides an example of how the results obtained in this study could be used as new experimental guidelines for continuing exploration in this field.</p><!><p>Here, we show that optical methods can be used to study crystal growth in a solution with a trace resolution of several nanometers. The DFMs combined with super-resolution techniques are well suited for the study of interface evolution for a complex crystal, MOF, and this might also be used to investigate other important crystals in solution. Identifying and quantifying the determining factors for MOF growth, including the apparent reaction orders of metal ions and organic linkers and the activation energy, provides new guidelines for the industrial production of MOFs and a new understanding of MOF growth at molecular level.</p><p>It is also shown that the crystal growth of MOFs is not an elementary reaction. A multi-step process is likely to exist involving the accumulation of the building blocks and their dissociation from the crystal in separate steps. In addition to the classic environmental factors, the crystal interface itself will also influence crystal growth, manifested in its retreat and recovery at the possible transition layer. The existence of a reversible transition layer is likely to mediate the crystal formation process as well as the dissociation of the crystal, offering the dynamics in the crystal growth of MOFs. This study depicts an alternative way of describing the evolution of crystal interface, compared with existing models.</p><!><p>Resource availability Lead contact Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Hexiang Deng (hdeng@whu.edu.cn).</p><!><p>This study did not generate new materials.</p><!><p>Data relating to the materials, methods, experimental procedures, mathematical models and other characterization are available in the supplemental information. All other data are available from the authors upon reasonable request.</p><!><p>The in situ interface imaging of MOF single crystals in solution was performed on Olympus IX71 DFM platforms. The DFM configuration is composed of white light source, condenser lens, home-designed flow cell, heating ring, objective lens, and electron-multiplying charge coupled device (EMCCD) camera placed beneath the sample (Scheme 1A; Figure S7). The individual MOF single crystals were illuminated The scattering light signal from the crystal interface was collected by a 40X NA0.60 air objective and followed by the detection using an ANDOR Ixon unknown date-897D-CS0-#BV EMCCD camera operated with a frame rate of 50 ms to generate a movie. The images in the movies were processed by super-resolution technique to trace the interface location for each MOF single crystal along its growth form small crystal to large one. The growth kinetic parameters of each individual MOF crystal were obtained by quantitative analysis. Constant bright spots from gold nanoparticles in DFM images were used as markers for the correction of the inevitable image drift with nanometer accuracy.</p><!><p>The flow cell was designed by parallel arrangement of two quartz slides with a distance of 0.32 mm sealed by AB glue, and it was placed between condenser lens and objective on DFM platform (Figure S8). The inlet and outlet of the liquid is using perfluoroalkoxy (PFA) tubes with an inter diameter of 0.5 mm. A suitable observation field in the eyepieces of the microscope was identified by adjusting the microscope focal distance. The inlet tube was connected to a vial containing stock solution of metal ions and organic linkers, while the outlet of flow cell was connected to the micro syringe pump to extract the solution and accurately control the growth rate. A steady rate of 20 mL$min À1 was used during MOF growth, unless specified otherwise. It was worth noting that the extraction design here avoided potential cracking of flow cell caused by excessive pressure induced from injection. The temperature of the flow cell was precisely controlled by the power of electrical heating mental placed beneath the flow cell and further calibrated by infrared camera. The image view covered an area of 200 3 200 mm in the quartz flow cell, corresponding to 512 3 512 pixels in the DFM image. The EMCCD camera was started recording after the flow rate and reaction temperature were steady. The in situ interface imaging method on DFM platforms was suitable for MOFs with various dimensionalities, Cu-MOF-74, Cu-MOF-2-BDC, Cu-MOF-2-NDC, Co-ZIF and HKUST-1. When recording was finished, the flow cell was taken off from the DFM setup and opened carefully for ex situ characterizations.</p><!><p>In the kinetic study of MOF growth by in situ interface imaging, one of the determining factors, concentration of metal ions and organic linkers, as well as temperature, was isolated and varied sequentially in a continuing test. In the case of Cu-MOF-2-NDC as an example, the detailed conditions were shown in Table S3, and the experimental procedures were described as follows. In the investigation of Cu 2+ as the determining factor, the concentration of organic linkers (0.035 mol$L À1 ) and temperature (50 C) were fixed, while the concentrations of Cu 2+ were varied sequentially from low to high (0.015, 0.025, 0.035, 0.045, and 0.055 mol$L À1 ). The stock solution with Cu 2+ concentration of 0.015 mol$L À1 was made by mixing the Cu 2+ solution, prepared by dissolving 36. Article (Cu 2+ : 0.025 mol$L À1 , linker: 0.035 mol$L À1 ) for 30 s with a rate of 2,000 mL$min À1 prior to a resumed recording. This was achieved by switching the inlet of the flow cell into new growth solution. In the new growth section, the flow rate was adjusted back to 20 mL$min À1 for another 410 s. This continued for the rest of solutions in the sequence, and after the recording of the last section, the camera was stopped, and the heating device was turned off. The remaining solution in the flow cell was extracted out at a flow rate of 1 mL$min À1 until empty. The investigation of the organic linker as determining factor was performed in similar way, where the concentration of Cu 2+ (0.025 mol$L À1 ) and temperature (50 C) were fixed, while the concentrations of organic linker NDC were varied sequentially from low to high (0.010, 0.015, 0.025, 0.035, and 0.045 mol$L À1 ). The impact from growth temperature was also assessed in the same way, where the concentrations of metal ions and organic linkers were kept constant (Cu 2+ : 0.075 mol$L À1 , organic linkers: 0.075 mol$L À1 ), while the growth temperature were varied sequentially from low to high (62.</p><!><p>A stock solution with the concentrations of Cu 2+ ions, 0.075 mol$L À1 , and organic linker NDC, 0.050 mol$L À1 , was applied as the initial condition at a flow rate of 20 mL$min À1 , while the temperature was fixed at 50 C. Then, the feed solution was changed to bare solvent of DMF, EtOH, and MeOH with volumetric ratio of 1:1:1, without Cu ions or NDC linker, at 300 s from the starting time of crystal growth (t 0 ). Different from the flashing with the rate of 2,000 mL$min À1 applied in the kinetic study, where the stock solution of the next concentration of metal ion and linkers were used by following the sequence, here, the bare solvent was used, and at the same flow rate to the prior section, 20 mL$min À1 , without flushing. The location change of the interface was traced for 300 s with the bare solvent in flow cell. Then, it was switched back to the original stock solution with the same concentration of Cu 2+ ions, 0.075 mol$L À1 , and organic linker NDC, 0.050 mol$L À1 , at the same flow rate for another 300 s. This alternating switching process was applied as the interruption induced to the crystal growth of Cu-MOF-2-NDC, 4 times in a continuous test. The detailed growth condition was listed in Table S4. Other interruption experiments were also performed using the same bare solution, but it was switch to stock solution of different concentration of metal ion and organic linker, either in a sequence of increasing or decreasing concentration.</p><p>Other experimental procedures and characterizations Chemicals and detailed synthetic conditions, SEM, PXRD, crystal face indexing, AFM, and other experimental procedures are provided in the supplemental information.</p><!><p>Supplemental information can be found online at https://doi.org/10.1016/j.chempr. 2022.03.006.</p>

Chem Cell

MEASUREMENT OF HbA1c IN PATIENTS WITH CHRONIC RENAL FAILURE

Background Carbamylated hemoglobin (carbHb) is reported to interfere with measurement and interpretation of HbA1c in diabetic patients with chronic renal failure (CRF). There is also concern that HbA1c may give low results in these patients due to shortened erythrocyte survival. Methods We evaluated the effect of carbHb on HbA1c measurements and compared HbA1c with glycated albumin (GA) in patients with and without renal disease to test if CRF causes clinically significant bias in HbA1c results using 11 assay methods. Subjects included those with and without renal failure and diabetes. Each subject\xe2\x80\x99s estimated glomerular filtration rate (eGFR) was used to determine the presence and degree of renal disease. A multiple regression model was used to determine if the relationship between HbA1c results obtained from each test method and the comparative method were significantly (p<0.05) affected by eGFR. These methods were further evaluated for clinical significance using difference between the eGRF quartiles of >7% at 6 or 9% HbA1c. The relationship between HbA1c and glycated albumin (GA) in patients with and without renal failure was also compared. Results Some methods showed small but statistically significant effects of eGFR; none of these differences were clinically significant. If GA is assumed to better reflect glycemic control, then HbA1c was approximately 1.5% HbA1c lower in patients with renal failure. Conclusions Although most methods can measure HbA1c accurately in patients with renal failure, healthcare providers must interpret these test results cautiously in these patients due the propensity for shortened erythrocyte survival in renal failure.

measurement_of_hba1c_in_patients_with_chronic_renal_failure

1. Introduction<!>2. Methods<!>3. Results and Discussion<!>4. Conclusions

<p>Renal failure is common in patients with diabetes, and HbA1c is widely used as an index of mean blood glucose in these patients. Many factors can affect interpretation of HbA1c measurements in patients with chronic renal failure (CRF). Several reports have suggested that erythrocyte survival is substantially lowered in most patients with CRF; this would be expected to lower HbA1c results [1–3]. Although shortened erythrocyte lifespan would presumably not interfere with the measurement of HbA1c, it could adversely affect the interpretation of HbA1c results.</p><p>Carbamylated Hb (carbHb) is formed by non-enzymatic condensation of cyanate with the N-terminal valine of hemoglobin. In chronic renal failure carbHb is increased due to elevated urea, which is dissociated in vivo to yield cyanate ions [4]. A number of old reports have suggested that HbA1c methods, especially those based on charge separation (e.g. ion-exchange HPLC) may have interference from carbHb that would be expected to falsely increase HbA1c results [5–7], but many of these methods are no longer in use. Subsequent reports evaluated newer ion-exchange HPLC assay methods which showed improved separation of the HbA1c fraction from other hemoglobin adducts [8,9] and therefore did not show interference from carbHb.</p><p>The present study is twofold; we first evaluated several current HbA1c methods for interference from carbHb in patients with and without renal failure. Although carbHb was not measured directly in the present study, there is a large amount of data showing that this hemoglobin modification is significantly increased in patients with renal failure and the carbamylated fraction (HbA1d3) as well as other measures of carbHb (measurement of valine hydantoin by HPLC) are correlated with plasma creatinine, serum urea and time-averaged urea concentrations [10–12]. Studies have also shown that the amount of carbHb depends upon both the duration and severity of real failure [13–15]. We therefore used eGFR as an indicator of overall renal function in place of direct measurement of carb Hb. We used a boronate affinity chromatography HPLC method as our reference method since this method has been shown to have no interference from carbHb [6,7,16,17].</p><p>In addition to the possible method-specific interference of carbHb, CRF, especially end stage renal disease, may also cause changes in erythrocyte lifespan which might alter the interpretation of HbA1c results. Several studies propose the use of glycated albumin (GA) measurement in place of HbA1c as a more accurate assessment of glycemic control in patients with renal disease. One study showed that GA was a better predictor of risk of death and hospitalization in these patients, compared to HbA1c [18]. Serum GA levels were also shown to be better correlated with average glucose (based on 4-point profiles, 3 days per week for 4 weeks) than HbA1c [19]. In this present study, we investigated the relationship between HbA1c and glycated albumin (GA) in patients with and without renal failure using the same patient samples as for the method-specific carbHb interference study.</p><!><p>We evaluated eight ion-exchange HPLC methods: G7 and G8 (Tosoh Bioscience), Variant II NU, Variant II Turbo, Variant II Turbo 2.0, D-10 and D-10 Dual (Bio-Rad Laboratories), HA-8160 (A. Menarini Diagnostics), two immunoassay methods: Tina-quant HbA1c gen.2 on Integra 800 (Roche Diagnostics) and DCA 2000 (Siemens Healthcare Diagnostics), and one enzymatic method: Direct Enzymatic HbA1c (Diazyme Laboratories) on the Hitachi 917 (Roche Diagnostics). Presumably, hemoglobin species modified by reactants other than glucose and not displaying a cis-1,1-diol group should not interfere with measurement of HbA1c by boronate affinity methods. Published data support this lack of interference of carbHb with boronate affinity methods [6,16,17]. Therefore, we used the boronate affinity ultra2 HPLC (Trinity Biotech) as our comparative method.</p><p>This study was approved by the ethics review committee at DynaLIFEDX in Edmonton, Canada where the samples originated. Whole blood samples (n=120) from subjects normal renal function and subjects in various stages of renal failure were residual samples from routine testing that had been collected in EDTA-containing tubes. The samples were shipped on cold packs to the Diabetes Diagnostic Laboratory at the University of Missouri (Columbia, MO). Several small whole blood aliquots were made from each sample and stored at −70°C until they were shipped on dry ice to various sites for analysis. One aliquot was centrifuged and the plasma was separated and stored at −70°C until analysis of GA. Each patient's eGFR, calculated using the MDRD equation, was used to estimate the degree of renal disease and the level of carbHb. A multiple regression model was used to determine if the relationship between HbA1c results obtained from each test method and the ultra2 method were significantly (p<0.05) affected by eGFR. For those methods' results that were significantly affected by eGFR, results were evaluated for clinical significance by dividing the samples into quartiles based on eGFR results (eGFR ≤11, 11< eGFR ≤ 45, 45< eGFR ≤ 84, eGFR >84). Deming regression was then used to compare the relationships between each method and the ultra2 for the highest and lowest quartiles; a difference between the quartiles of >7% at 6 or 9% HbA1c was defined as being clinically significant [20]. The relationship between HbA1c (ultra2 HPLC) and GA was evaluated comparing patients with normal eGFR (eGFR>90 ml/min, no renal disease, n=18) and those with renal failure (eGFR<60 ml/min, n=73). Data analyses were performed using SAS and Excel.</p><!><p>The D-10, D-10 Dual, DCA 2000, G7 and Direct Enzymatic methods showed very small but statistically significant effects of eGFR; clinical significance was therefore evaluated. The differences in HbA1c from the reference method between the lowest and highest eGFR quartiles is shown in figure 1 as a box plot of the HbA1c simple linear regression residuals for each method. In this way, any inherent calibration bias is removed and only the difference between the highest and lowest quartiles can be seen. Importantly, none of the methods evaluated showed any clinically significant effects of eGFR.</p><p>The relationship between HbA1c and GA is shown in figure 2. The difference in the relationship between the normal and renal failure groups was both statistically (Linear regression, p<0.0001) and clinically significant. If we assume that GA is providing an accurate measure of glycemic control in patients with renal failure, HbA1c results are lowered by approximately 1.5% HbA1c in patients with renal failure at critical treatment levels. These results are consistent with the findings of others that have found lower HbA1c results in renal failure when compared to measures of glycated plasma protein or plasma albumin [2,3]. The studies showing that GA is superior to HbA1c use in CRF are somewhat convincing but far from definitive [21]. There are studies showing a linear increase in all-cause mortality with increasing HbA1c levels [22,23] and there is no evidence as yet that physicians can achieve better glycemic control using GA instead of HbA1c in these patients. In addition, as with HbA1c, there may be factors (e.g. proteinuria, altered albumin homeostasis) that interfere with measurement or interpretation of GA. There is ongoing debate about which assay is most useful for monitoring glycemic control in this vulnerable population [24,25].</p><!><p>We conclude that most current HbA1c methods can provide valid analytical results for patients with CRF. However, healthcare providers need to be aware of potential interferences when interpreting HbA1c results in clinical settings due to alteration in erythrocyte lifespan in many patients with chronic renal failure which can cause falsely lowered HbA1c results.</p>

PubMed Author Manuscript

Rapid Identification of Virulence Determinants in Influenza Viruses

To date there is no rapid method to screen for highly pathogenic avian influenza strains that may be indicators of future pandemics. We report here the first development of an oligonucleotide-based spectroscopic assay to rapidly and sensitively detect a N66S mutation in the gene coding for the PB1-F2 protein associated with increased virulence in highly pathogenic pandemic influenza viruses. 5\xe2\x80\x99-thiolated ssDNA oligonucleotides were employed as probes to capture RNA isolated from six influenza viruses, three having N66S mutations, two without the N66S mutation, and one deletion mutant not encoding the PB1-F2 protein. Hybridization was detected without amplification or labeling using the intrinsic surfaced-enhanced Raman spectrum of the DNA-RNA complex. Multivariate analysis identified target RNA binding from non-complementary sequences with 100% sensitivity, 100% selectivity, and 100% correct classification in the test data set. These results establish that optical-based diagnostic methods are able to directly identify diagnostic indicators of virulence linked to highly pathogenic pandemic influenza viruses without amplification or labeling.

rapid_identification_of_virulence_determinants_in_influenza_viruses

Introduction<!>Reagents<!>Preparation of Ag Nanorod SERS Substrates<!>DNA Probes<!>Influenza Viruses<!>Viral Influenza RNA Samples<!>Immobilization of DNA Probes onto Ag Nanorod Arrays<!>Raman Spectroscopy<!>Data Analysis<!>Results<!>Conclusions

<p>Influenza A virus is a ubiquitous negative strand RNA virus having pandemic potential.1,2 Numerous studies have suggested that specific mutations in the HA, PB1, and NA genes are related to influenza virulence and pandemic potential.3-6 The PB1-F2 protein has especially been linked to virulence since it is considered proapoptotic and pathogenic.7-10 A N66S mutation in the PB1-F2 sequence is consistent among pathogenic influenza viruses, including the pandemic 1918 H1N1 and 1997 H5N1 highly pathogenic avian influenza strains, and is considered a virulence determinant.11 Research shows that the N66S mutation correlates with significantly increased pathogenicity and mortality in mice, and that PB1-F2 promotes secondary bacterial infections; the mechanism of increased virulence may be related to inhibition of interferon induction.12 A recent global database analysis of the PB1-F2 protein revealed that the N66S mutation was present in only 3.8% of the H5N1 strains, however the mutation was specifically found associated with the highly pathogenic strains.13 In particular, all six H5N1 human isolates having the N66S mutation in the PB1-F2 protein isolated from Hong Kong influenza outbreaks were found to be highly pathogenic.13 Given these data, it is apparent that the N66S mutation is relevant and critical for determining the pathogenic potential of influenza.</p><p>Development of a rapid and sensitive method for identifying emerging influenza viruses and determinants of virulence or pandemic potential is critical for control of transmission and disease intervention strategies. Currently, only genomic techniques such as PCR are available for laboratory diagnosis of virulence markers.14,15 While these techniques provide identification of prognostic indicators, they rely entirely on genomic sequencing and alignment, and can be limited by issues of reliability, standardization, and cost. Some studies of a commercial PCR test for influenza showed a relatively low sensitivity (~75%);16 the authors suggest the use of a more sensitive reference test to confirm negative results. The inability to provide definitive screening highlights the need for a diagnostic platform with high sensitivity, specificity, and expediency.</p><p>Our research groups have previously shown that surface-enhanced Raman spectroscopy (SERS) is a highly sensitive and specific method for direct, label-free detection of DNA-RNA binding.17-22 The intrinsic Raman spectra of oligonucleotide probe-target complexes have been shown to be spectrally unique and sensitive to the hybridization of both matched and mismatched target sequences.23-29 We recently reported on a SERS-based assay for identification of virulence factors associated with pathogenesis in influenza in model systems.30 The current work shows that oligonucleotide-modified Ag nanorod arrays can be used for rapid and sensitive detection of pathogenicity determinants isolated from highly pathogenic and pandemic influenza viruses through direct identification of RNA and genetic mutations in PB1-F2 without amplification or labeling of the virus. The findings reported here provide the basis for oligonucleotide-based SERS screening of influenza with pandemic potential in a point-of-care application.</p><!><p>6-mercapto-1-hexanol (MCH) was purchased from Sigma-Aldrich (St.Louis, MO). All other chemicals were of analytical grade and used without any further purification. The hybridization buffer was prepared by dissolving 20 mM Tris HCl, 15 mM NaCl, 4 mM KCl, 1 mM MgCl2, and 1 mM CaCl2 in molecular biology grade water at pH 7.3; it was stored at 4°C when it is not in use. The buffer and working tools were DNase free.</p><!><p>Oblique-angle vapor deposition (OAD) was used to produce aligned Ag nanorod substrates for SERS applications, according to previously published methods.31,32 In brief, standard glass microscope slides were cleaned using piranha solution, rinsed several times with deionized water, and dried using N2 before being placed into a custom-designed, high vacuum electron beam vapor deposition chamber. Uniform thin film layers of Ti (20 nm) layer and Ag (500 nm) were first deposited onto the glass substrate at rates of 2.0 Å/s and 3.0 Å/s, respectively. The substrates were then rotated to 86° relative to the incident vapor source and Ag nanorods were deposited at a constant rate of 3.0 Å/s until a nominal thickness of 2000 nm, as determined by a quartz crystal microbalance in the deposition chamber. These vapor deposition conditions result in optimal high aspect ratio Ag nanorod SERS substrates with overall nanorod lengths of ~900 nm, diameters of ~80-90 nm, densities of ~13 nanorods/μm2, and a tilt angle of 71° with respect to the substrate normal.32 Following nanofabrication, a patterned multi-well array was produced on the Ag nanorod substrate according to previously published procedures.33</p><!><p>DNA probes were purchased from Integrated DNA Technologies (IDT, Coralville, IA). The 5'C6 thiolated ssDNA probes were received lyophilized, and dissolved in molecular biology grade water to a concentration of 1000 nM. DNA probes were designed for viruses having determinants of low and high virulence in the PB1-F2 RNA, as previously described.30</p><!><p>Three wild type influenza viruses were used in in these studies: A/Mute Swan/MS451072/06 (H5N1), A/CK/TX/167280-04/02 (H5N3), and A/CK/PA/13609/93 (H5N2).34 The first two of these wild type viruses are examples of strains containing the N66S mutation, while the third did not have the mutation. Three additional reverse genetics viruses were used in these studies. These were the WH, WH N66S, and WH ΔPB1-F2 strains. These three viruses are 7:1 reassortants of A/WSN/33 (H1N1) with the PB1 segment (segment 2) of A/Hong Kong/156/97 (H5N1) highly pathogenic avian influenza virus. Two of these reverse genetics viruses contained either the wild type, intact PB1-F2 protein (WH), or the PB1-F2 protein with the N66S mutation (WH N66S). The third of the reverse genetics viruses was a negative control in which the PB1-F2 protein was deleted by removal of the start codon and introduction of two stop codons within the PB1-F2 open reading frame (WH ΔPB1-F2). 11,35</p><p>MDCK cells were used to propagate the WH influenza viruses and were maintained in Dulbecco's Modified Eagles Medium (DMEM; GIBSO BRL Laboratories, Grand Island, NY) with 5% heat-inactivated (56 °C) FBS (Hyclone Laboratories, Salt Lake City, UT). For virus production, MDCK cells were rinsed 3 times with PBS, overlayed with 5 ml MEM + TPCK trypsin (1μg/ml; Worthington Biochemical, Lakewood, NJ) + virus and grown for 3-5 days at 35°C until ~70% cells were released from the flask surface. Supernatants containing virus were collected, centrifuged to remove cellular debris, aliquoted and stored at −80°C until use. Virus titers were quantified by hemagglutination (HA), 50% tissue culture infectious dose (TCID50) , and plaque assays as previously described.36 The virus stock titer and PFU in 0.2 ml final volume for each of the influenza viruses used in this study are summarized in Table S.1 in Supporting Information.</p><!><p>Viral RNAs isolated from six strains of influenza were used. This include three examples of N66S mutations (WH N66S, A/Mute Swan/MS451072/06, A/CK/TX/167280-04/02), and two without the N66S mutation (WH, A/CK/PA/13609/93). An influenza deletion mutant not containing the PB1-F2 sequence was used as a negative control (WH ΔPB1-F2). A PureLink Viral RNA/DNA mini Kit (Invitrogen, Carlsbad, CA) was used to isolate influenza virus RNA. Viral RNA was extracted by mixing 200 μL of each strain with 25 μL of Proteinase K in 1.5 mL followed by addition of 200 μL of 1X PBS/0.5 % BSA in a microcentrifuge tube. The resulting solution was mixed for 15 seconds and then the lysate was incubated at 56 °C for 15 minutes. Subsequently, 250 μL of 96-100 % ethanol was added and then the lysate was mixed for 15 seconds followed by the incubation for 5 minutes at room temperature. The lysate was transferred onto the Viral Spin Column® and centrifuged at 5,000 rpm for 1 minute. The flow-through was discarded and the spin column was placed in a new collection tube. The washing step was repeated one more time with 500 μL of the wash buffer. The collection tube was discarded and the spin column was transferred into a new collection tube and spun at 13,000 rpm for 1 minute to dry the column. The column was placed into a new recovery tube and 50 μL of sterile, RNase-free water was added to top the column. The resulting solution was incubated for one minute at room temperature. The column was then centrifuged at 13,000 rpm for 1 minute to elute the viral nucleic acids. Virus RNA purity and concentration was quantified by UV-Vis spectrometry (Thermo Fisher NanoDrop 1000, Wilmington, DE).</p><!><p>5'-thiol single stranded DNA (ssDNA) oligonucleotide probes were immobilized on the Ag nanorod array surface to capture and detect RNA strains corresponding to the PB1-F2 gene mutation. Preparation of self-assembled monolayers (SAMs) of ssDNA probes on the Ag nanorod substrates followed previously published procedures.20,22,30 Briefly, 20 μL of 1000 nM of the oligonucleotide solution was added to a patterned microwell and incubated overnight at room temperature. After the incubation, any unbound oligonucleotide solution was removed from the microwell by rinsing it three times with molecular biology grade water and blow dried with N2. Then, 20 μL of 100 nM solution of the spacer molecule 6-mercapto-1-hexanol (MCH) was added to the microwell in order to minimize non-specific binding of DNA/RNA molecules to the surface of Ag nanorod substrates and for the correct oligonucleotide conformation. The spacer molecule was incubated for 6 hours at room temperature followed by the rinsing and drying steps. 20 μL of 20 ng/μL (~5 nM) RNA solution diluted in the binding buffer was added to the oligonucleotide-functionalized Ag nanorod to accomplish the hybridization and incubated at 37°C for 2 hours under a humid environment to avoid dehydration. After the incubation, any non-specifically adsorbed RNA molecules were removed by rinsing with the binding buffer with the final wash using molecular biology grade water. The rinsed substrate was then dried with a gentle stream of N2.</p><!><p>Raman spectra were collected using a confocal Raman microscope (InVia, Renishaw, Inc., Hoffman Estates, IL) equipped with a 785 nm diode laser as the excitation source. The sample was illuminated through a 20× (Leica, Germany) N.A. = 0.40 objective with a spot size of approximately 4.8 × 27.8 μm; laser power was ~0.42 mW as measured at the sample. Spectra were collected between 2000 – 500 cm−1 using a 30 second acquisition time. Spectra were acquired from five different spots in each individual micro well on the Ag nanorod substrate. Four micro wells were used for each sample; therefore, twenty spectra were collected for each sample and used for further data analysis.</p><!><p>Prior to multivariate analysis, the raw spectra were preprocessed using a 1st order Savitzky-Golay derivative filter (15-point, 2nd order polynomial), normalized to unit vector length, and mean centered. These preprocessing methods removed any spectral variations caused by instrumental drift, non-uniformity between different micro wells on the substrate, and environmental changes. Initial spectral quality was assessed using principal component analysis (PCA). Determination of spectral outliers was based on calculation of their PCA scores with their corresponding Hotelling T2 and Q residuals values.37 Out of more than 140 spectra used in this analysis, only one spectral outlier was found and eliminated prior to analysis.</p><p>Multivariate analysis for classification was performed using partial least squares discriminate analysis (PLS-DA)38 and support vector machine discriminate analysis (SVMDA)39,40. All data processing was performed with PLS Toolbox version 6.2 (Eigenvector Research Inc., Wenatchee, WA) in MATLAB R2012a (The Mathworks Inc., Natick, MA).</p><!><p>Six influenza strains were used in this study. Three of these influenza strains contained the N66S mutation, representative of the putative PB1-F2 mutation consistent with increased virulence; these were the WH N66S, A/Mute Swan/MS45107206, and A/CK/TX167280-04/02 strains. Two other viruses were used that did not contain the N66S mutation and were representative of low virulence; these were the WH and A/CK/PA/13609/93 strains. A negative control virus was included, WH ΔPB1-F2, that had the open reading frame for the PB1-F2 protein deleted. For simplicity viruses containing the N66S determinant are referred to as "high virulence" while the viruses not containing the N66S determinant are referred to as "low virulence." In addition to these samples, spectra of the DNA probe alone were collected and used in the analyses.</p><p>SERS spectra are shown in Figs. 1A and 1B for the high and low virulence strains, respectively; each spectrum is an average of 20 individual spectra and is presented without processing. Figure 1A presents SERS spectra of the high virulence DNA probe-spacer complex before hybridization (Fig. 1A I), DNA-probe hybridized with complementary high virulence viral RNA strains (Fig. 1A II), and the DNA-probe incubated with non-complementary low virulence viral RNA strains (Fig. 1A III). Figure 1B shows SERS spectra of low virulence DNA probe-spacer complex alone (Fig. 1B I), the spectra of DNA-probe incubated with non-complementary high virulence viral RNA strains (Fig. 1B II), and DNA-probe incubated with complementary low virulence viral RNA target sequence (Fig. 1B III). The dominant features found in the SERS spectra in Fig. 1 correspond to nucleic acid vibrations, e.g. 1332, 1089, 1023, 793, and 623 cm−1.22,30</p><p>The high virulence target RNA was distinguished from low virulence and control RNA using a whole-spectrum, multivariate statistical analysis of the Raman spectra. This method has been previously employed for detection, identification, and classification of pathogens.41-43</p><p>Partial least squares discriminant analysis (PLS-DA) was utilized to build multivariate classification models to discern high virulence RNA binding to the substrate. The classification model was designed such that 2/3 of spectra of the high virulence and low virulence RNA complexes were designated as a calibration/training sets, while the remaining 1/3 of the spectra in each class were designated as the validation/prediction sets. This separation allowed the calibration model to contain all possible variances needed to explain the validation set. The spectra were randomly assigned to each set in order to minimize any correlation between spectral variances and order sequence. Cross-validation (Venetian blinds, 10 splits) was used for internal validation of the calibration model. The optimal number of latent variables (LV's) was selected based on the cross-validated class error.</p><p>Figures 2A and 2B represent the PLS-DA prediction results for the high virulence and low virulence assays, respectively, as a function of sample number. The prediction results for the samples in Fig. 2 include both the calculated values for the calibration sets as well as the predicted values for the validation sets. Each icon in Fig. 2 represents a SERS spectrum; the color code and shape of the symbol represents the particular class of samples the spectrum belongs to, as defined in the caption to Fig. 2. The optimum threshold value for sample classification is represented by the red dashed line in Fig. 2; the threshold is calculated using Bayes' Theorem based on the minimization of total classification errors.44 Spectra with predicted values greater than the Bayesian threshold are designated as belonging to a particular category, as defined by the classification model.</p><p>Figure 2A represents the results of a PLS-DA classification model designed to identify high virulence strains. Samples 1 – 40 in Fig. 2A represent the predicted PLS-DA classification values for the training set of SERS spectra containing the high virulence strains, including WH N66S, (), A/CK/TX167280-04/02 (), and A/Mute Swan/MS451072/06 (). Samples 41 – 105 represent the predicted classification values for the training set of LPAIV strains and controls, including LPAIV RNA isolated from strains WH (), A/CK/PA/13609/93 (), and WH ΔPB1-F2 (); controls included the buffer () and the DNA HPAIV capture probe alone (). It is clear from Fig. 2A that this method unambiguously separates the spectra of the high and low virulence strains in the calibration sets with complete accuracy.</p><p>This high virulence classification model was tested using samples 106 – 124 and 125 – 160, which were the validation sets for the high virulence and low virulence / control samples, respectively. Figure 2A qualitatively indicates that this high virulence model accurately classified both low and high virulence validation sets; Table 1 provides the quantitative values. The results show 100% calculated sensitivities and specificities, with root-mean square error of prediction (RMSEP) values of 0.21 for both classes. The overall percentage of test samples correctly classified by the high virulence PLS-DA model was 100%.</p><p>Figure 2B represents the complementary situation for a classification model designed to identify low virulence strains. In similar fashion to Fig. 2A, samples 27 – 38 and 65 – 76 in Fig. 2B represent the predicted PLS-DA classification values for the calibration sets of SERS spectra containing the low virulence strains WH () and A/CK/PA/13609/93 (), while samples 112 – 120 and 140 – 146 represent the validation sets used to test this model. Similar to the high virulence model in Fig. 2A, the low virulence model in Fig. 2B indicates high classification accuracy. Table 1 provides the quantitative results for the low virulence model: calculated sensitivities and specificities of 100%, and RMSEP of 0.22, with an overall percentage of correctly classified test samples of 100%. Results from the PLS-DA models show extremely high sensitivities, specificities, and percent correct classification, albeit with relatively high values for RMSEP.</p><p>While PLS-DA is a powerful tool for classification and regression, it is not optimized for use with complex, non-linear data sets.45,46 Support vector machine-discriminant analysis (SVM-DA) is a relatively new classification and regression method that can produce a unique, global solution when presented with high-dimensional inputs.47,48 We applied SVM for classification of the high and low virulence strains described above.</p><p>For SVM-DA analysis, a radial basis function (RBF) kernel was used and the SVM model was calculated by grid searching within a range of paired values of cost (C = penalty error) and radial width (γ). In this formulation, SVM required fitting two parameters for optimization. The first is γ defined as γ=12σ2, where σ is the radial width of the RBF that determines the shape of the hyperplane that best separates the different classes. The second parameter C takes into account the regression errors of the training set and controls the complexities of the class boundaries. Once the optimal parameters were determined for the calibration set, the test set was loaded and class membership probabilities calculated using the established SVM-DA calibration model.</p><p>The SVM-DA model structure in terms of calibration and validation sets was identical to that described for PLS-DA (see above). The calibration set was first compressed by choosing an optimized rank of latent variables as determined from a cross-validated principal least squares (PLS) calculation. The optimal pair of SVM parameters (C, γ) was chosen by cross validation (Venetian blinds, 5 splits) of the calibration set. The values used were γ = 100 and C = 0.316 for the high virulence assay, and γ = 100 and C = 0.001 for the low virulence assay. Nineteen support vectors were used in the calculations for both assays.</p><p>Figure 3 illustrates the results from the SVM-DA calculations for a high virulence classification model. As in the case of the PLS-DA model (Fig. 2A), samples 1 – 40 represent the training set of SERS spectra containing the HPAIV strains, samples 41 – 105 represent the training set of low virulence strains and controls, while samples 106 – 124 and 125 – 160 are the test sets for the high virulence and low virulence / control samples, respectively. The ordinate axis in Fig. 3 is the predicted class membership probability, as calculated by the SVM-DA model. In a binary classification model, the closer the class predicted probability is to 0.0 or 1.0, the more likely the sample is to belong to that particular class. Figure 3A shows that SVM fully separates the high virulence test samples from the low virulence and control samples. Table 2 quantifies the results: the SVM model provides 100% specificity and sensitivity for prediction with 100% of test samples correctly classified. In addition, the SVM model has a root-mean-square error of class predicted probability (RMSECPP) of 0.07, showing a much lower prediction error compare to PLS-DA model, which had an RMSEP value of 0.21.</p><p>A similar situation occurs for the low virulence assay illustrated in Fig. 3B, in which samples 27 – 38 and 65 – 76 represent the predicted SVM class membership probabilities for the calibration sets of SERS spectra containing the low virulence strains WH and A/CK/PA/13609/93, while samples 112 – 120 and 140 – 146 represent the validation sets used to test this model. Table 2 provides the quantitative results: the SVM low virulence model showed 100% sensitivity and specificity for prediction with 100% of the test samples correctly classified. The SVM model also had a RMSECPP of 0.06, compared with the prediction errors associated with PLS-DA, i.e. an RMSEP value of 0.22.</p><!><p>We report here the first use of oligonucleotide-modified substrates as diagnostic tools for the direct identification of a PB1-F2 mutation in the influenza virus genome related to virulence, specifically the N66S gene mutation within the PB1-F2 protein. The method employed 5'-thiol-modified ssDNA sequences as probes to capture RNA isolated from avian and reverse genetics influenza viruses containing low virulence or high virulence determinants. We used a label-free and amplification-free optical read-out method, i.e. Raman spectroscopy, to determine the efficacy of binding. The Raman spectra of both high virulence and low virulence DNA-RNA target complexes showed high similarity; therefore, multivariate analysis was used to identify target binding. Binary classification models were developed to distinguish complementary from non-complementary DNA-RNA target hybrids. The SVM-DA model that was developed using a radial basis function kernel resulted in calculated values of 100% sensitivity, 100% specificity, and 100% correct classification of the test samples with a small root-mean-square error of prediction (RMSECPP ~0.07).</p><p>The current study was designed to demonstrate the ability of the SERS methodology to identify different virulence genotypes from real RNA virus-containing specimens, not to determine a lower limit of detection of the assay. However, a previous study using the same ssDNA PB1-F2 probes employed in this article demonstrated that these SERS-based methods were an order of magnitude more sensitive than ELISA for the capture of synthetic influenza RNA target sequences (10 nM vs. 100 nM).30 Also, in terms of the use of these methods for complex biological samples, we have previously shown that the SERS methods described in this paper were simultaneously able to identify eight human rotavirus strains and classify each according to its G or P genotype with >96% accuracy.49 Other studies showed that our SERS-based methods had equivalent-or-better detection limits than qPCR for analysis of pathogens in complex clinical samples.50 Therefore, based on our previous experience, we feel confident that the methods described here can be extended to analyze biologically complex mixtures.</p><p>These studies establish that optical-based Raman diagnostic methods are able to sensitively and accurately detect influenza virus RNA mutations linked to pathogenicity in emerging highly pathogenic avian and pandemic influenza viruses without amplification or labeling. The results are also the first demonstration of the use of real influenza viral RNA for direct identification of diagnostic indicators of influenza virulence. Future work will address the applicability and robustness of this platform for more relevant containing the target viral RNA in complex influenza isolates.</p>

PubMed Author Manuscript

Enantioselective dioxytosylation of styrenes using lactate-based chiral hypervalent iodine(III)

A series of optically active hypervalent iodine(III) reagents prepared from the corresponding (R)-2-(2-iodophenoxy)propanoate derivative was employed for the asymmetric dioxytosylation of styrene and its derivatives. The electrophilic addition of the hypervalent iodine(III) compound toward styrene proceeded with high enantioface selectivity to give 1-aryl-1,2-di(tosyloxy)ethane with an enantiomeric excess of 70–96% of the (S)-isomer.

enantioselective_dioxytosylation_of_styrenes_using_lactate-based_chiral_hypervalent_iodine(iii)

<!>Findings<!><!>Findings<!><!>Findings<!><!>Findings<!><!>Findings<!>

<p>This article is part of the Thematic Series "Hypervalent iodine chemistry in organic synthesis".</p><!><p>Hypervalent aryl-λ3-iodanes have been widely used for metal-free oxidation with high selectivity in organic synthesis [1–3]. The reactivity of an aryl-λ3-iodane is controlled by the electronic and steric properties of the aryl group and the heteroatomic ligand coordinated to the iodine atom. Optically active hypervalent iodine compounds contain chiral ligands or chiral aryl groups. Several types of optically active hypervalent iodine reagents and catalysts have been developed for highly stereocontrolled oxidative transformations [4–14]. The enantioselective vicinal difunctionalization of alkenes constitutes one type of attractive transformation achieved by chiral hypervalent iodine compounds. As a seminal example in this field, Wirth et al. [15–17] reported the dioxytosylation of styrene (1a, Scheme 1). Chiral hypervalent iodine reagents 2 bearing a 1-methoxyethyl side chain were used for enantiocontrol of the dioxytosylation, and the maximum enantiomeric excess (ee) of the product 3a reached 65%. Despite recent rapid progress in the field of asymmetric oxidation achieved by chiral hypervalent iodine compounds, there has been no subsequent examination of dioxytosylation, which can be used as a standard reaction for comparing the enantiocontrolling ability of chiral hypervalent iodine reagents.</p><!><p>Enantioselective dioxytosylation of styrene as a seminal example.</p><!><p>The design of chiral hypervalent iodine reagents using a lactate motif has been employed for several types of oxidation reaction since we first reported this procedure [18]. Enantioselective oxidative transformations include the dearomatization of phenols [19–24], α-functionalization of carbonyl compounds [25–29], and vicinal difunctionalization of alkenes [18,30–50]. Here, the efficiency of the lactate-based chiral hypervalent iodine reagents 4a–e (Figure 1) was assessed using the dioxytosylation of styrenes as a reference reaction.</p><!><p>Series of lactate-based hypervalent iodine reagents.</p><!><p>A series of lactate-derived aryl-λ3-iodanes 4a–e was used for the oxidation of styrenes 1 in the presence of p-toluenesulfonic acid (TsOH) in dichloromethane. The reaction proceeded at −50 °C to give the 1,2-dioxytosylated product 3 and the rearranged product 5. The yields of 3 and 5 were determined by 1H NMR using an internal standard. The ee of 3 was determined by chiral HPLC analysis. The results for the yields and ee are summarized in Table 1.</p><!><p>Enantioselective dioxytosylation of styrenes 1 using aryl-λ3-iodanes 4.a</p><p>aThe reaction was carried out at −50 °C in dichloromethane containing 4 (47 mM), TsOH (86 mM), and 1 (43 mM) for 4 h. bThe yield was determined by 1H NMR using an internal standard. cThe ee was determined by chiral HPLC using a Daicel CHIRALPAK AD column (ø 4.6 mm × 250 mm). dPreferential configuration of product 3. The absolute stereochemistry of 3b and 3c was not determined. eThe reaction was carried out for 20 h.</p><!><p>The reaction of styrene (1a) with 4a gave the 1,2-dioxytosylated product 3a with 70% ee of the (S)-isomer (Table 1, entry 1). An ee of equal to or greater than 70% was also achieved in the reactions with the other lactate-based reagents 4b–e (Table 1, entries 2–5). The reaction with the 2,6-bis(lactate)aryl reagent 4e provided a high ee of 92%. The reactions of p-chlorostyrene (1b) gave 3b with a similar ee, and the ratios of 3 to 5 (3b to 5b) were higher than those in the reaction of 1a (Table 1, entries 6–8). In the reactions of o-methylstyrene (1c), the ee of the 1,2-dioxytosylated product 3c was slightly higher than those of 3a and 3b, but the regioselectivity for 3c over 5c was poor (Table 1, entries 9 and 10).</p><p>Scheme 2 illustrates possible reaction pathways that lead to 3 and the achiral byproduct 5. The treatment of (diacetoxyiodo)benzene with TsOH readily gives Koser's reagent [PhI(OH)OTs] [51], which has a higher electrophilicity toward the carbon–carbon double bond in 1. The dioxytosylation of alkenes with Koser's reagent was found to proceed via an SN2 reaction of a cyclic intermediate such as I1, judging from the syn selectivity of the dioxytosylation [52–53]. The attack of the tosylate ion on I1 possibly takes place at the benzylic position or at the methylene carbon atom. The positive charge of I1 may be stabilized by the aryl group and localized at the benzylic position. This may allow the preferential formation of I3 from I1. If I2 was the major intermediate in the pathway leading to 3, the stereochemical purity of 3 would have decreased owing to the facile elimination of the iodonium group [54] at the benzylic position of I2 (SN1). The high enantiomeric ratio of 3 can be rationalized via a preference for the I1→I3→3 pathway over the I1→I2→3 pathway. The product ratio of 3 to 5 was affected by the ring substituent in styrenes 1: the electron-withdrawing chloro substituent in 1b increased the amount of 3, whereas the electron-donating methyl substituent in 1c decreased the amount of 3. An electron-donating aryl group increases the rate of participation of the aryl group (I3→I4). In other words, a reaction pathway that bifurcates from I3 to 3 and 5 agrees well with the regioselectivity for 3 over 5 observed for the substituted styrenes. The phenonium cation intermediate I4 contains two reaction sites on the ethylene bridge. Electron donation due to the lone pair on the oxygen atom of the internal tosyloxy group may weaken the bond between the tosyloxy-bonded carbon and the quaternary carbon in I4.</p><!><p>Plausible pathways in dioxytosylation of styrenes.</p><!><p>The reaction of styrene with 4a–e preferentially gave (S)-3, which forms via an electrophilic addition of the iodane toward the Si face of styrene, followed by an SN2 reaction with the tosylate ion. If an SN1 mechanism were involved in the oxytosylation of I1, the enantiomeric ratio of 3 would decrease owing to the planar structure of the benzylic cation. Thus, the tosylate ion may act as an effective nucleophile for the SN2 reaction of I1. The stereoface-differentiation in the dioxytosylation reaction using the lactate-derived aryl-λ3-iodanes is similar to that in preceding reactions [14], which include the diacetoxylation [38–3950] and diamination [30,49] of styrene.</p><p>In summary, the reaction of styrenes with lactate-derived aryl-λ3-iodanes gave the dioxytosylated product with an ee of 70–96%.</p><!><p>Experimental procedures, characterization data, and copies of 1H and 13C NMR spectra are available.</p>

PubMed Open Access

Accurate calibration and control of relative humidity close to 100%\nby X-raying a DOPC multilayer

In this study, we have designed a compact sample chamber that can achieve accurate and continuous control of the relative humidity (RH) in the vicinity of 100%. A 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) multilayer can be used as a humidity sensor by measuring its inter-layer repeat distance (d-spacing) via X-ray diffraction. We convert from DOPC d-spacing to RH according to a theory given in the literature and previously measured data of DOPC multilamellar vesicles in polyvinylpyrrolidone (PVP) solutions. This curve can be used for calibration of RH close to 100%, a regime where conventional sensors do not have sufficient accuracy. We demonstrate that this control method can provide RH accuracies of 0.1 to 0.01%, which is a factor of 10\xe2\x80\x93100 improvement compared to existing methods of humidity control. Our method provides fine tuning capability of RH continuously for a single sample, whereas the PVP solution method requires new samples to be made for each PVP concentration. The use of this cell also potentially removes the need for an X-ray or neutron beam to pass through bulk water if one wishes to work close to biologically relevant conditions of nearly 100% RH.

accurate_calibration_and_control_of_relative_humidity_close_to_100%\nby_x-raying_a_dopc_multilayer

Introduction<!>Chamber design<!>Calibration of the DOPC d-spacing vs. relative\nhumidity curve<!>I. Derivation of osmotic pressure<!>II. Fitting data from Hristova and White16 and Tristram-Nagle et al.11<!>III. Direct calculation of RH vs. DT using thermodynamic theory<!>I. Comparison with other literature<!>II. Discussion of errors<!>III. Advantages of the present method<!>IV. Further discussion

<p>There are currently two commonly used methods for relative humidity (RH) control. One utilizes air/water vapor flow, for which the accuracy is usually ±1 to 2% for the RH range from 0% to 95%. The second method involves placing a reservoir with saturated salt solution in the chamber, which gives a discrete number of values of the RH, depending on the kind of salt used, e.g., NaCl for 75% RH and K2SO4 for 97% RH. Both methods require a uniform temperature environment. A small temperature fluctuation or a temperature gradient would easily result in ±1% to ±2% error in RH. To our knowledge, accurate and continuous humidity control with an error of less than ±0.1% for high humidity values (95–100% RH) has not been shown with these methods. To achieve high accuracy humidity control close to 100% RH, one must control the temperature gradient and have an accurate measure of the RH. No existing RH sensor in the market can measure with accuracy close to or better than 0.1%. To design such accurate RH control, one needs to address both issues carefully.</p><p>Temperature uniformity and stability throughout the whole sample chamber is very difficult to control within such a small tolerance. This is exactly the cause of the widely debated "vapor pressure paradox" for lipid membranes,1 where better than 99% RH was not achieved. It has been experimentally proved by Katsaras that once the temperature gradient is eliminated, 100% RH can be achieved and no paradox exists.2,3 The Nagle group has also designed a chamber to achieve 100% RH for lipid bilayer X-ray measurements,4,5 and neutron measurements (see www.humidity.frank-heinrich.net).</p><p>In order to achieve not only 100% RH, but also accurate and continuous control for a range of high relative humidities close to 100%, we have developed a chamber which controls a temperature differential. This method has been used previously for surface wetting studies.6–8</p><p>In order to obtain an accurate measurement of RH, we need to use a calibration sample that responds very sensitively to RH changes close to 100%. We have chosen to use the lamellar repeat spacing of a 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) multilayer as a calibrant. It is well known that the water uptake of lipids responds very sensitively when the RH gets close to 100%. Although the possibility of using a supported lipid multilayer to measure RH has been previously mentioned in the literature,9 no rigorous calibration of the d-spacing vs. RH curve has been carried out directly with vapor chambers. This is mainly due to the lack of RH sensors with sufficient accuracy. In this study, we will try to establish this calibration standard by consolidating the theory given in the literature with the published data of DOPC multi-lamellar vesicles in polyvinylpyrrolidone (PVP) solution, and use this curve as the calibration curve for our data with supported DOPC multilayers in a vapor chamber.</p><!><p>There are two main parts in our chamber design: the reservoir and the sample. A reservoir consisting of a 1% (mole fraction) K2SO4 solution serves as the humidity source which generates a constant water vapor pressure (we call this relative humidity the reference RH). The sample is located where the desired RH is created. There are two independent temperature control loops: temperature control for the reservoir and temperature control for the sample. The two parts are connected via a weak thermal link. A schematic of the temperature control setup is given in Fig. 1.</p><p>By controlling the temperature of the reservoir, Tres, and the temperature of the sample, Tsam, we can control the temperature difference ΔT = Tres − Tsam. The distribution of water molecules in the water vapor will re-arrange according to the temperature gradient, which results in a re-distribution of relative humidity. As demonstrated in Fig. 2: when ΔT = 0 the sample is at the reference RH for 1% K2SO4 solution. Note that the use of an unsaturated salt solution produces an approximately temperature independent RH. When ΔT < 0, RH of the sample is lower than the reference RH (Fig. 2(c)); similarly when ΔT > 0, it is higher (Fig. 2(a)).</p><p>Fig. 3 shows pictures of our humidity controlled sample chamber used as a cell for X-ray diffraction and optical microscopy measurements. The chamber consists of two parts: the top (Fig. 3(b)) with reservoir and the bottom (Fig. 3(c)) with sample. The two parts have independent temperature control loops, and are thermally separated by a Teflon© ring. The reservoir solution is contained in a sponge. The sponge and the sample are kept in close contact with the respective part of the sample chamber to favor temperature equilibration. The sample chamber is constructed with copper to ensure good thermal uniformity. A Lakeshore temperature controller with two control loops is used for the temperature control.</p><!><p>Measurements of a DOPC multilayer as a standard sample were carried out to measure the RH of the sample environment. X-ray diffraction measurements of the lamellar repeat distance or d-spacing of the DOPC multilayer sample are a sensitive measure of the RH of the sample environment, since the uptake of water between the bilayers depends sensitively on RH particularly as the RH tends to 100%. In our experiments, the sample temperature is kept constant at 31 °C while the reservoir temperature is raised to increase the RH at the sample. The DOPC multilayers are deposited using spreading method developed by Li et al.10 and were annealed at 50 °C for 1–2 days in a humidity chamber after taking out from the vacuum.</p><p>The X-ray measurements were taken on the diffractometer at sector 33 BM at the Advanced Photon Source, Argonne National Laboratory with a 20 keV X-ray beam. Fig. 4 shows one set of diffraction measurements of a typical DOPC sample over a range of temperature gradient. The RH ranges from 97.1% to 100.000% if converted from measured d-spacing with the standard curve discussed below. As RH increases the diffraction peaks shift to lower qz, which means the d-spacing is increasing. The gradual distortion and disappearance of higher orders of Bragg peaks is due to the increased undulations due to increased hydration, as explained by Nagle and Tristram-Nagle,11–13 as well as by Salditt.14</p><p>There is no obvious standard established in the literature for converting from d-spacing to RH. The data which do exist contradict each other.15,16 We have resolved this conflict by using a theoretical model combined with existing experimental measurements.</p><!><p>In order to calibrate our data for d-spacing as a function of RH, we will compare our results with measurements of the d-spacing of DOPC multilammelar vesicles in solution, where the osmotic pressure of the solution has been modified by the addition of PVP, a high molecular weight polymer. These measurements should be comparable, since the osmotic pressure in solution, and the RH in vapor should have identical effects on the chemical potential of water in the multilayers. According to Petrache et al.,17 the osmotic pressure of the multilayer, Posm, can be decomposed into three contributions, a Helfrich fluctuation pressure, Pfl, a hydration pressure, Ph, and a van der Waals pressure PvdW; (1)Posm=Pfl+Ph+Pvdw. </p><p>The fluctuation pressure, Pfl, depends on the thickness of the water layer, a and can be approximated with an exponential function with a decay length λfl of the form, (2)Pfl=Afle−a/λfl. </p><p>The hydration pressure, Ph can be written in a similar fashion in terms of a decay length λh, (3)Ph==Ahe−a/λh. </p><p>The van der Waals pressure,18 PvdW has the form, (4)Pvdw=H6π(−2(DB′+a)3+1(2DB′+a)3+1a3). </p><p>Here, H is a Hamaker constant, DB′ is the bilayer thickness and a the water thickness. DB′ + a = d, is the d-spacing of the multilayer.</p><p>In order to determine the values for the parameters Afl, λfl, Ah, λh, H and DB′ for DOPC, we need to fit this expression for Posm to the experimental data.</p><!><p>In the paper by Hristova and White published in 1998,16 a list of d-spacings vs. RH from 34% to 93% and PVP weight fractions from nominal 60% to 5% is given, as well as the number of water molecules per lipid nw. In order to get a calibration curve for RH > 95%, we took the 60% to 5% weight fraction PVP data and translated it into osmotic pressure.</p><p>To convert PVP concentration to osmotic pressure, we use the method described by Vink19 in 1971. The concentration c can be calculated from the PVP weight fraction w, (5)c=wwυ2+(1−w)υ1. </p><p>Osmotic pressure can then be calculated using the relation; (6)P=A1c+A2c2+A3c3. </p><p>The values of υ1, υ2, A1, A2, A3 are taken from the same paper.19 The PVP weight fractions are taken with values that are determined via refractive index measurements by Hristova and White.16 The calculation results are listed in Table 1.</p><p>To calculate DB′, we used the method described by Nagle and Tristram-Nagle,12 (7)DB′=2DC+2DH (8)DC=VCA (9)VC=VL−VH (10)nw=Ad2−VLvw where DC is the lipid tail group thickness and DH is the lipid head group thickness. VC is the lipid tailgroup volume, VH is the lipid headgroup volume and VL is the total volume of one lipid molecule. A is the lipid cross sectional area, vw is the volume of one water molecule and nw is the number of water molecules per lipid.</p><p>One can solve for A and feed into the expression for DB′ and get (11)DB′=VCdnwvw+VL+2DH </p><p>Put in values for DH = 9 Å obtained by Büldt et al. with neutron diffraction,20 VH = 319 Å3 by Sun et al. with X-ray diffraction,21 VL = 1303.3 Å3 by Tristram-Nagle et al.11 with X-ray neutral flotation measurements, vw = 30 Å3, nw and d data given by Hristova and White,16 we get values for DB′ and a. These are given in Table 1.</p><p>Tristram-Nagle et al.11 also have done detailed studies of DOPC swelling with osmotic pressure and published data of osmotic pressure vs. DOPC multilayer water spacing a, and osmotic pressure vs. d-spacing.22 When we compare the results of Tristram-Nagle and Hristova, we found that there is a discrepancy in the number of water molecules per lipid, which leads to a discrepancy in the calculated DB′. As listed in Table 1, DB′ is between 47.7–48.3 Å, while DB′ from ref. 11 is between 45.3–46.5 Å in the same hydration range. This ~2 Å discrepancy in bilayer spacing would lead to a shift of plots of osmotic pressure vs. water spacing a for the two published data sources.</p><p>However, when plotting the two data sources of osmotic pressure vs. d-spacing of DOPC, they agree very well, as shown in Fig. 5. So we decided to combine the two published data of osmotic pressure vs. d-spacing, and fit with the function (4) by making DB′ a fitting parameter together with Afl, λfl, Ah, λh, H. The combined data set should give better accuracy than fitting either data alone.</p><p>There are multiple sets of parameters which can fit the data equally well if all the parameters are allowed to vary. We have chosen to fix the values of λfl and Ah at the values obtained by Tristram-Nagle et al.11 Since our purpose is to obtain a calibration curve for the d-spacing vs. osmotic pressure, we do not concern ourselves with the significance of the actual values of the fitted parameters. The result of the fitting is shown in Fig. 5 and Table 2. The fitted DB′ value is close to ref. 11.</p><p>With this calculated standard curve of Posm vs. d-spacing, we can convert to RH vs. d-spacing using the relation between osmotic pressure and relative humidity1 (12)Posm=−(kTvw)ln(RH). </p><p>The fitted curve of d-spacing vs. RH is plotted in Fig. 6 together with the data of Hristova and White16 and Tristram-Nagle et al.11 Also shown are our measured values of the DOPC d-spacing at various values of ΔT. The measured DOPC d-spacings vs. ΔT are plotted in Fig. 7(a). By putting these measured d-spacings onto the standard curve, we can establish the RH vs. ΔT plot for our chamber environment (see Fig. 7(b)) which we shall now discuss in the next section.</p><!><p>Besides the experimental approach, we can also directly calculate the RH vs. ΔT. from thermodynamic theory. This can serve as a reliability check for our experimental calibration.</p><p>Assuming ideal behavior, the RH at the reservoir is (13)rreservoir=PPH2O*=xH2Ovapour=xH2Oliquid=1−φ </p><p>Here, P is the partial pressure of water vapor at the reservoir, PH2O* is the saturated water vapor pressure, xH2Ovapor is the mole fraction of water in vapor and xH2Oliquid is the mole fraction of water in liquid and φ is the mole fraction of solute in the reservoir.</p><p>According to the Clausius–Clapeyron equation for a liquid– gas equilibrium, (14)dln(P)dT≈ΔHmRT2 where ΔHm is the enthalpy of vaporization of water and T is the sample temperature which is not near the critical temperature Tc. The equation can be re-written as (15)ln(1+ΔPP)≈ΔPP≈ΔHmΔTRT2 where ΔP, ΔT and ΔP are respectively the temperature difference and the difference in partial pressure of water vapor between the sample and the reservoir.</p><p>Thus we have (16)rsample≈rreservoir−ΔHmΔTRT2 where rsample and rreservoir are RH at the sample and the reservoir respectively. This shows that to first order, the change in RH is proportional to the change in temperature ΔT. By putting in numbers of ΔHm = 40.68 kJ mol−1, RT = 25.249 J mol−1 and T = 304 K, we obtain ΔHmRT2=0.0530. Comparing this result to our experimental result in Fig. 7(b), our experimental result also shows a linear relation except for the last 4 points at RH > 0.9995. The linear fit of this data gives a slope of 0.0209, which is less than half of the theory predicted value.</p><p>After careful examination of the d-spacing equilibration time, we hypothesize that the reason for the last four points falling off the straight in Fig. 7(b) is that we did not allow enough time for the d-spacing to equilibrate. The waiting time at each temperature before measurement was around 20–25 min, which is not enough when RH gets very close to 100%.</p><p>We also carefully examined the temperature gradient in our chamber, and concluded that the discrepancy in the slope of linear fit is caused by a small temperature gradient between the sponge and the copper top. The temperature sensor for the reservoir is embedded in the copper top for good thermal contact. When heating up the reservoir relative to the sample, the temperature gradient is always negative from the copper to the sponge, which means the sponge is a little cooler than the sensor reading. This leads to a smaller experimental slope than the theoretically predicted value. In conclusion, the temperature differentials plotted are nominal temperature differentials, not the actual temperature between the sample and reservoir. The fact that our data fall onto a linear relation predicted by thermodynamic theory when translated to RH vs. dT plot provides further support for the RH vs. d-spacing standard.</p><p>Detailed analysis of more lipid multilayer data using this humidity control setup will be presented separately in other papers by Y. Ma et al., (in preparation).</p><!><p>There are other papers in the literature reporting the evolution of the d-spacing of DOPC with RH, such as the paper by Caracciolo et al.15 We make a comparison of our simulations according to Petrache's method, Hristova and White's data, Tristram-Nagle's data with Caracciolo et al.'s data for the range of RH very close to 100% (Fig. 6). As can be seen, there is a significant discrepancy in Caracciolo's data compared to the rest. The theory used by Petrache et al. from literature predicts a non-linear and diverging behavior of lipid d-spacing with change in RH at high RH range, and Hristova and White and Tristram Nagle's data also suggest that, while Caracciolo et al.'s data shows an almost linear relation at the same range.</p><p>We think that the discrepancy can come from 2 sources. Firstly, Caracciolo et al.'s study was time dependent. The measurement of d-spacing was done while the humid air was continuously flowing into the chamber and increased RH in real time. At lower humidities, the d-spacing changes slowly with RH, so it can still catch up and be close to equilibrium; however, at high RH values close to 1, the d-spacing changes are much larger for the same amount of change in RH due to the divergent behavior, and thus in real-time the d-spacing no longer catches up with the change of RH therefore the measurements "on the fly" are not under equilibrium conditions. For example, measurements by Servantes23 show that it can take up to several hours for a multilayer to equilibrate for RH near 100%.</p><p>A second possibility is the non-accurate reading from the RH sensor. The humidity is measured with a humidity sensor in Caracciolo et al.'s study while Hristova and White's data are PVP weight fraction calculated from refraction index measurements on the sample. It is well known that for the current humidity sensors on the market, the accuracy is around ±1%, and would not be able to determine changes on the order of 0.1% or less. So in this case, it is quite possible that the RH sensor is already saturated when RH is close to 1 and yields readings larger than the actual humidity in the chamber. On the other hand, the refraction index measurements can be more accurate for determining the PVP weight concentration and thus give a more accurate measure when converted to RH.</p><!><p>To estimate the error of RH, we need the error of d-spacing measurements, and also the error in the standard conversion curve. The error bars for the d-spacing measurements in Fig. 7(a) are between 0.014 Å to 0.031 Å, which are 0.02–0.05% errors, much smaller than the symbol size to plot. We can estimate the errors of the standard conversion curve from the reduced Chi-square of the fitting. The reduced Chi-square is 0.10 in the ln P fit, so the dP/P is approximately 0.10=0.32. From the differential of eqn (12), we can get the error of RH between 0.32% to 0.00047% in the RH range 99% to 99.999%.</p><!><p>We believe that our compact and economic chamber design together with using a calibration standard would be helpful for future studies of soft materials and bio materials which rely on a high humidity environment. In our own experiments we put a standard DOPC sample with an actual sample side by side. By switching the two in and out of the X-ray beam, one can get the RH value from measuring the d-spacing of the DOPC multilayer and also get real measurements from the actual sample under the same conditions. If one is confident about the thermal contact between the sponge and the cell, as well as the accuracy of salt solution, one can also use the calibration curve to control ΔT without using a DOPC sample once the cell is calibrated.</p><p>There are three main advantages of this method. Firstly, it is clear that in the multilayer case, compared to measurements in solution with PVP, our results using a vapor chamber have better accuracy (smooth curve) and stronger signal (we can still see the third order diffraction peak at 100% RH). Secondly, our method makes it possible to change the RH of the environment by simply tuning the temperature differential, which enables measurements under different conditions on the same piece of sample. In the PVP method, one has to make a different sample for each PVP concentration. For experiments with large sample-to-sample variance but looking for subtle changes in a given sample under different conditions (which might be true for a lot of soft matter experiments), this can be a big advantage. Finally, samples under saturated vapor pressure are more amenable to studies using X-rays and neutrons since problems associated with absorption and scattering in the water overlayer are not present.</p><!><p>There are some points to be noted for designing and using such a chamber. First of all, using a non-saturated salt reservoir instead of pure water can help because it lowers the reference RH, at the same time increasing the required temperature differential. Secondly, the extremely compact design of the sample chamber makes a difference. As demonstrated, our chamber is 2.5 inch in outer diameter, which can fit in one's palm. The small volume makes temperature control much easier – less non-uniformity and faster equilibration. The parts are easy to make, assemble, maintain and transport.</p><p>It is also worth noting that this chamber design works best at temperatures a few degrees above ambient temperature. With only heating elements, the chamber will not operate below ambient temperature; on the other hand, if the temperature is set too far above from ambient temperature, water condensation on the inner window can create problems and frequent wiping is required. For lower and higher temperatures (such as 10 C and 50 C), an additional temperature regulated layer of enclosure outside our described cell is recommended to raise or lower the ambient temperature.</p><p>Last but not least, the chamber can be equipped and used for a wide range of non-contact measurements. For example, our chamber can do X-ray experiments at the same time as optical microscopy. For contact experiments, similar principles apply, one simply has to pay special attention to the sealing of the chamber and avoidance of cold spots.</p>

PubMed Author Manuscript

Characterization of sphingosine-1-phosphate lyase activity by ESI-LC/MS/MS quantitation of (2E)-hexadecenal

Sphingosine-1-phosphate (S1P) is a sphingolipid signaling molecule crucial for cell survival and proliferation. S1P-mediated signaling is largely controlled through its biosynthesis and degradation, and S1P lyase (S1PL) is the only known enzyme, which irreversibly degrades sphingoid base-1-phosphates to phosphoethanolamine and the corresponding fatty aldehydes. S1PL-mediated degradation of S1P results in the formation of (2E)-hexadecenal, while hexadecanal is the product of dihydrosphingosine-1-phosphate (DHS1P) degradation. Fatty aldehydes can undergo biotransformation to fatty acids and/or alcohols, which makes them elusive and renders the task of fatty aldehyde quantitation challenging. We have developed a simple, highly sensitive, and high-throughput protocol for (2E)-hexadecenal quantitation as a semicarbazone derivative by liquid chromatography-electrospray ionization-tandem mass spectrometry. The approach was applied to determining S1PL activity in vitro, with the ability to use as low as 0.25 \xc2\xb5g microsomal protein per assay. The method is also applicable to the use of total tissue homogenate as the source of S1PL. A correction for (2E)-hexadecenal disappearance due to its biotransformation during enzymatic reaction is required, especially at higher protein concentrations. The method was applied to confirm FTY720 as the inhibitor of S1PL with the IC50 of 52.4 \xc2\xb5M.

characterization_of_sphingosine-1-phosphate_lyase_activity_by_esi-lc/ms/ms_quantitation_of_(2e)-hexa

Introduction<!>Reagents<!>Preparation of fatty aldehyde semicarbazone derivatives<!>Hydrogenation of unsaturated aldehydes<!>ESI-LC/MS/MS analysis of fatty aldehyde semicarbazone derivatives<!>Tissue homogenization and S1P lyase reaction<!>ESI-MS/MS analysis of fatty aldehydes as semicarbazone derivatives<!>ESI-LC/MS/MS analysis of fatty aldehydes as semicarbazone derivatives<!>Optimization of derivatization conditions<!>Optimization of S1PL reaction with liver homogenate<!>Characterization of S1PL activity using liver microsomes<!>Inhibition of S1PL by FTY720<!>

<p>Sphingosine-1-phosphate (S1P) is a sphingolipid signaling molecule crucial for cell survival and proliferation [1,2]. S1P-mediated signaling is largely controlled through its biosynthesis and degradation, and S1P lyase (S1PL) is the key enzyme in the regulation of S1P and sphingolipid homeostasis. The importance of S1PL for mammalian physiology has been revealed only recently. It was shown that the abnormal expression of S1PL is associated with cancer development, with the resistance to anticancer therapies, and with the developmental pathologies [3,4]. In the immune system, by regulating circulatory S1P level, S1PL contributes to the control of lymphocyte egress from lymph nodes [5], and is now considered as a promising target to treat inflammatory disorders [3].</p><p>S1PL is the only enzyme that irreversibly degrades sphingoid base-1-phosphates to phosphoethanolamine and the corresponding fatty aldehydes, thus eliminating sphingoid bases from the total pool of sphingolipids [6]. By degrading S1P, S1PL controls S1P-mediated signaling and also generates metabolites that may elicit potent physiological responses [7]. S1PL degrades both S1P and its endogenous analog dihydro-S1P (DHS1P). S1PL-mediated cleavage of S1P results in the formation of (2E)-hexadecenal while hexadecanal is the product of DHS1P degradation. Fatty aldehydes are reactive and can undergo further biotransformation to fatty acids and/or alcohols [8], making them elusive and rendering the task of fatty aldehyde quantitation and characterization of S1PL activity challenging. Two approaches have been used for characterization of S1PL activity in biological specimens. One is based on radiometric quantitation of the radioactive products of [4,5-3H]DHS1P degradation after separation by TLC [8]. The other approach is based on the use of fluorescently labeled (BODIPY- or NBD-) S1P, HPLC separation, and quantitation of the corresponding BODIPY- or NBD-labeled aldehydes [9,10]. There are notable disadvantages of each method, such as the use of a radioactive substrate or the complicated HPLC profile of fluorescently labeled products of the S1PL reaction; furthermore, neither method allows monitoring of the endogenous levels of (2E)-hexadecenal. Most importantly, neither method allows exhaustive correction for the ongoing degradation/bioconversion of the (2E)-hexadecenal formed during the reaction. In the current study, we take advantage of the high sensitivity of liquid chromatography-electrospray ionization-tandem mass spectrometery (ESI-LC/MS/MS) and use of commercially available internal standard for (2E)-hexadecenal quantitation, d5-labeled (2E)-hexadecenal, to develop a simple, sensitive, and high-throughput protocol for characterization of S1PL activity in vitro based on isotope dilution quantitation of (2E)-hexadecenal as a semicarbazone derivative.</p><!><p>Methanol, water, and acetonitrile (LC/MS grade) were purchased from Fisher (Pittsburgh, PA, USA). (2S,3R)-1,3-Dihydroxy-2-amino-4E-octadecene-1-phosphate (S1P) and (2E)-hexadecanal-(15,15,16,16,16-d5) were obtained from Avanti Polar Lipids (Alabaster, AL, USA). (2E)-Hexadecenal was made by Horner-Wadsworth-Emmons olefination of myristaldehyde with triethyl phosphonoacetate, followed by alane reduction of the α,β-unsaturated ester [11] and oxidation of the resulting primary alcohol with pyridinium chlorochromate in dichloromethane. FTY720 and (S)-FTY720-phosphate (FTY720-P) were synthesized as described in [12]. cis-11-Hexadecenal was purchased from Sigma (St. Louis, MO, USA).</p><!><p>Hexadecanal, cis-11-hexadecenal, and (2E)-hexadecenal, pure or as part of the total lipid extract, were converted to the corresponding semicarbazone derivatives by heating with 0.2 ml of 5 mM semicarbazide hydrochloride in methanol containing 5% formic acid at 40°C for 2 h. The derivatives were directly subjected to MS/MS or LC/MS/MS analyses. To study the effect of acid concentration on the effectiveness of derivatization, the reactions were performed in the presence of varying (0–5%) concentration of formic acid in methanol.</p><!><p>(2E)-Hexadecenal and cis-11-hexadecenal were converted to hexadecanal by hydrogenation over ~1 mg platinum oxide in 1 ml of methanol for 1 h. After the catalyst was removed by centrifugation, hexadecanal was analyzed as its semicarbazone derivative by MS. The completeness of hydrogenation was confirmed by ESI-MS/MS by the absence of the 2- or 11-hexadecenal semicarbazone [M+H]+ ion.</p><!><p>An API-4000 Q-trap triple quadrupole ion trap mass spectrometer (Applied Biosystems, Foster City, CA, USA) interfaced with an Agilent 1100 liquid chromatograph with autosampler (Agilent Technologies, Wilmington, DE, USA) was employed for ESI-LC/MS/MS analysis of the aldehyde semicarbazone derivatives. Positive ion ESI-LC/MS/MS was employed for detection of the fatty aldehyde semicarbazone derivatives. The ion source conditions and gas settings for positive ESI-LC/MS/MS analysis were as follows: ion spray voltage = 5500V, ion source heater temperature = 550°C, collision gas setting = 4, ion source gases 1 and 2 settings = 50, curtain gas setting = 30. Three MRM transitions were monitored for detection of (2E)-hexadecenal semicarbazone and cis-11-hexadecenal semicarbazone - m/z 296/279, m/z 296/253, and m/z 296/97. The monitored MRM transitions for hexadecanal semicarbazone derivative were m/z 298/281, m/z 298/255, and 298/238. The MRM transitions for the internal standard d5-(2E)-hexadecenal semicarbazone were m/z 301/284, 301/258, and 301/79. The optimized parameters for (2E)-hexadecenal semicarbazone and d5-(2E)-hexadecenal semicarbazone positive ion ESI LC/MS/MS analysis were as follows: declustering potential = 90V, collision energy (CE) for the m/z 296/279 and 301/284 transitions = 23V with collision cell exit potential (CXP) = 8V; CE for m/z 296/253 and 301/258 transitions = 21V with CXP = 8V; CE for m/z 296/79 and m/z 301/79 transitions = 33V with CXP = 6V. Chromatographic separation of semicarbazone derivatives of (2E)-hexadecenal, cis-11-hexadecenal, and hexadecanal was achieved by using a Discovery C18 column (50 × 2.1 mm, 5 µm particle size) from Supelco (Bellefonte, PA, USA) and a gradient elution program in which methanol, water, and formic acid (60:40:0.5, v/v) containing 5 mM ammonium formate was used as solvent A, and acetonitrile:chloroform:water:formic acid (90:10:0.5:0.5, v/v) containing 5 mM ammonium formate was used as solvent B. The elution protocol was composed of a 2-min column equilibration with 100% solvent A at a flow rate of 0.5 ml/min, followed by sample injection in methanol, a 0.5 min period with 100% solvent A at a flow rate of 0.5 ml/min, a 2.5-min linear gradient to 100% solvent B and 0.425 ml/min flow rate, a 3-min period with 100% solvent B at a flow rate of 0.425 ml/min, and a 0.75-min linear gradient to 100% solvent A and a flow rate of 0.5 ml/min. The program included three cyclic needle washes consisting of duplicate needle washes per cycle prior to sample injection.</p><!><p>Frozen liver tissue was homogenized on ice in the tissue lysis buffer consisting of 5 mM MOPS, 1 mM EDTA, 0.25 M sucrose, 1 mM PMSF, 10% (v/v) glycerol, pH 7.4. Cell debris was removed by low speed (1,000 × g for 5 min) centrifugation, and the supernatant was used as a total tissue homogenate. To isolate the microsomal fraction, the total tissue homogenate was centrifuged at 10,000 × g for 15 min, and the supernatant was subjected to 100,000 × g centrifugation for 1 h. The resulting supernatant was used as a cytosolic fraction. The membrane fraction was gently dispersed in the lysis buffer using tip sonication and stored at −80°C until use.</p><p>Protein concentrations were determined using the Pierce bicinchoninic acid (BCA) reagent (ThermoFisher Scientific, Rockford, IL) and adjusted to 0.025 mg/ml with tissue lysis buffer (up to 0.25 mg/ml in some experiments). The S1PL reaction was initiated by mixing 0.025 ml of 0.4 mM S1P in 1% Triton X-100 in water, 0.175 ml of reaction buffer (35 mM potassium phosphate buffer, pH 7.4, 0.6 mM EDTA, 70 mM sucrose, 36 mM sodium fluoride, 0.57 mM pyridoxal-5'-phosphate (P5P), and 0.05 ml of protein preparation in lysis buffer (5 µg protein in the standard reaction, and up to 50 µg protein in experiments where the rate of (2E)-hexadecenal biotransformation was determined) in 8-ml glass screw-capped tubes. The reaction was performed for 20 min at 37°C, and was stopped by the addition of 2 ml of methanol. The internal standard ((2E)-d5-hexadecenal, 20 pmol) and chloroform (2 ml) were added, the tubes were vortexed, and phase separation was initiated by the addition of 1.55 ml of 0.9% KCl in water. After the tubes were extensively vortexed, the chloroform phase was separated by centrifugation, the solvent was evaporated by a stream of nitrogen, and aldehyde semicarbazone derivatives were obtained as described above.</p><!><p>Fatty aldehydes are poorly ionized by themselves but can be easily modified into derivatives with strong ionization properties. Semicarbazide is one of several possible reagents that will provide a derivative with a strong molecular ion in positive ion mode during ESI (Fig. 1). The semicarbazones of (2E)-hexadecenal, cis-11-hexadecenal, and hexadecanal produced strong molecular ions at m/z 296, m/z 296, and m/z 298, respectively. The study of the product ion formation during collision-induced decomposition (CID) revealed a substantial difference in the breakdown pattern of these aldehyde derivatives (Figure 2A–C). Thus, both (2E)-hexadecenal and cis-11-hexadecenal derivatives produced major products at m/z 279, m/z 253, m/z 251, and m/z 97. However, (2E)-hexadecenal semicarbazone gave ions at m/z 279, m/z 253, and m/z 97 with almost equal output when CID conditions were optimized for the maximal generation of each product ion, whereas the cis-11-hexadecenal semicarbazone produced a minimal amount of the ion at m/z 97. Hexadecanal semicarbazone had the only major product ion at m/z 281 and did not form a product ion at m/z 97 (Figure 2C). The study of the ion precursor-product relationships for (2E)-hexadecenal semicarbazone revealed a direct relationship between ions at m/z 97 and m/z 279; the latter also produced the ion at m/z 251 while the ion at m/z 253 was formed directly from the molecular ion. Suggested ion structures and their relationships are shown in the Figure 2D. The fact that cis-11-hexadecenal semicarbazone also produced a minimal amount of the ion with m/z 97 indicates that this ion formation is a function of the presence and position of the double bond in the aliphatic chain, with a Δ2-position providing maximal ion output.</p><!><p>Next we developed a LC/MS/MS protocol to separate aldehyde semicarbazone products and to provide an additional degree of selectivity for the entire analysis. We employed a Supelco Discovery C8 column (50 × 3.5 mm, 5 µm particle size) and a gradient from a methanol:water:formic acid system to an acetonitrile:chloroform:water:formic acid system. All three transitions ([M-17]+, [M-43]+, and [M-199]+) were monitored. Figure 3 demonstrates complete peak base separation of all three aldehydes tested (only one transition based on the loss of the amino group is shown). It is especially important to achieve complete peak base separation between the semicarbazone products of (2E)-hexadecenal and hexadecanal as palmitaldehyde is a common constituent of plasmalogens [13] and may interfere with the analysis by overlapping (from its [M+3] natural isotopic analog) with the signal from the stable isotope-labeled internal standard.</p><p>(2E)-Hexadecenal semicarbazone eluted as a major peak with Rt of 4.98 min having a small shoulder at Rt of 4.89 min (Fig. 3). (2E)-d5-Hexadecenal semicarbazone demonstrated the same LC behavior (Fig. 3). However, cis-11-hexadecenal and hexadecanal semicarbazones eluted as single peaks (Fig. 3). This suggests that the presence of the double bond at C2 allows the potential formation of two stereoisomers during the reaction with semicarbazide, which are partially resolved using our chromatographic conditions. When both (2E)-hexadecenal and (2E)-d5-hexadecenal were hydrogenated over platinum oxide, they both were fully converted to hexadecanal, and their semicarbazones had the same Rt without the shoulder peak characteristic for (2E)-hexadecenal semicarbazone (Fig. 3). Therefore, when quantifying the signal from (2E)-hexadecenal semicarbazone during ESI-LC/MS/MS analysis, we took the sum of the areas of the major peak and of its shoulder peak.</p><!><p>To prepare the semicarbazones, we initially used 5 mM semicarbazide in methanol with 1% formic acid in the reaction medium. Both methanol and formic acid are components of the solvent system used for column equilibration and initial chromatographic separation. Hence, injection of the methanol-based reaction mixture is the most compatible with the employed solvents and allows sample injection without additional steps of sample preparation before injection. Further, we explored the effect of the increasing acidity of the reaction mixture on the effectiveness of (2E)-hexadecenal derivatization with semicarbazide. Figure 4 shows that the increase in acidity has a direct effect on the rate of the semicarbazone formation performed at 40°C, with 5% formic acid in methanol giving almost complete conversion within 2 h. Based on these findings, all further derivatizations were performed with 5% formic acid in methanol and a 2-h derivatization time. These derivatization conditions ensured reliable detection of as low as 5 fmol (2E)-hexadecenal semicarbazone on-column (data not shown).</p><!><p>Rat [8] and mouse [9] liver homogenates were used previously to characterize S1PL activity with radioactive or fluorescently labeled substrates. We decided to use mouse liver as the source of the enzyme activity and employed the same lysis and reaction buffers as previously described [9] to optimize the assay conditions. However, our original attempts to detect S1P hydrolysis and the formation of (2E)-hexadecenal failed because of interference of buffer components with all targeted transitions chosen to detect both (2E)-hexadecenal and (2E)-d5-hexadecenal. Detailed investigation of the components of the reaction/lysis buffer(s) revealed that dithiothreitol (DTT) interferes in the reaction with semicarbazide. When DTT was omitted from the lysis buffer we detected the aldehydes as their semicarbazone derivatives.</p><p>Another concern we had during the initial steps of method development was the reduction or oxidation of the aldehyde to 2-hexadecen-1-ol or 2-hexadecenoic acid after being formed from S1P upon degradation by S1PL [8,9]. One of the approaches to improve the assay conditions is to purify, even partially, the enzyme. S1PL is a microsomal enzyme with its active site exposed to the cytosol [8]. To diminish potential contributions of cytosolic proteins and cofactors to (2E)-hexadecenal biotransformations we compared S1PL activity in the cytosolic and membrane fractions with that of the total homogenate. Equal amounts of protein (5 µg) were taken into the reaction mixture. Figure 5 demonstrates that S1PL activity is mostly associated with the microsomal fraction. Hence, we continued characterization of S1PL activity using the microsomal fraction as the source of the enzyme. It is important to note that the 100,000 × g fraction prepared from mouse liver homogenized in the lysis buffer without DTT did not diminish S1PL activity for at least one month when stored at −80°C.</p><p>To investigate if (2E)-hexadecenal is degraded/converted during the S1PL enzymatic reaction performed in our conditions, we checked the disappearance of a standard (2E)-hexadecenal from the reaction mixture when all components of the reaction were present except for S1P. Figure 6 shows that the degree of (2E)-hexadecenal disappearance is directly related to the amount of the microsomal protein added to the reaction as well as to the duration of the incubation. Hence, to diminish or eliminate the interference with S1PL assay from the aldehyde biotransformation, it is best to minimize both the protein concentration and the reaction time. In the case of a mouse liver microsomal preparation, a reaction period of 20 min with 5 µg protein per reaction eliminated hexadecenal biotransformations. For better precision of S1PL-mediated S1P degradation and (2E)-hexadecenal quantitation we continued to determine the rate of (2E)-hexadecenal (20 pmol) biotransformation in each further experiment and incorporated the determined "disappearance" correction factor in our calculations. However, in the case of 5 µg mouse liver microsomal protein preparation per reaction and 20-min reaction time, the rate of (2E)-hexadecenal bioconversion was stably below 1%.</p><!><p>Having determined a method to correct for (2E)-hexadecenal biotransformations during the enzymatic reaction, we proceeded to further characterize S1PL activity. First, we determined the linearity of the S1PL reaction with 40 µM S1P as a function of microsomal protein concentration with a 20-min reaction time. Figure 7 shows that the S1PL reaction is linear within a broad range of protein concentration, with (2E)-hexadecenal biotransformation data correction required at protein concentration above 5 µg per reaction. It should be noted that the sensitivity of the LC/MS/MS approach allowed reliable quantitation of (2E)-hexadecenal when 2.5, 1, 0.5, and even 0.25 µg protein per reaction was used. Clearly, at these low protein concentrations there was no loss of (2E)-hexadecenal during the reaction because of its negligible rate of biotransformation. Badhuvula et al. [9] employed 5–50 µg total tissue or cell protein per assay from control or S1PL overexpressing cells, with 25 µg protein per assay in most experiments. They demonstrated (2E)-hexadecenal biotransformation under their experimental conditions; however, no approach to correct for the aldehyde loss was suggested. In the LC/MS/MS method we describe here, with a control for the (2E)-hexadecenal disappearance, a precise quantitation for S1PL-mediated S1P degradation is achieved even at conditions favorable for (2E)-hexadecenal biotransformation.</p><p>Next, we performed a kinetic analysis of the S1PL reaction by employing 5 µg microsomal protein per reaction, a 20-min reaction time, and various S1P concentrations. Figure 8 demonstrates that the S1PL activity was nearly maximal at 40 µM S1P, with a Km of 5.7 µM. This is by far the lowest published Km value for the S1PL reaction. Previous Km values of 9.0–20.1 µM were reported with DHS1P as the substrate [8–10]. Similar Km values were obtained with NBD-S1P (7–14.6 µM) [9] and BODIPY-S1P (35 µM) [10] as substrates. Our approach uses the natural and probably the best substrate for S1PL that results in a higher rate of enzymatic activity in comparison with all previously tested compounds. Also, by using S1P we avoid the limitations imposed by the use of radioactive substrate ([3H]-DHS1P), thus simplifying the entire procedure.</p><!><p>To test the feasibility of the developed LC/MS/MS approach to detect changes in S1PL activity, we performed an in vitro assay in the presence of a known inhibitor of S1PL, FTY720 [14]. We also tested the effect of (S)-FTY720-phosphate (FTY720P), which was previously shown not to inhibit S1PL [14]. FTY720 and FTY720P were solubilized in 1% Triton X-100 together with S1P and the reaction was performed with 5 µg mouse liver microsomal protein and 40 µM S1P, in the presence or absence of 30 µM FTY720 or FTY720P for 20 min. Consistent with the published data, we found that FTY720 but not FTY720P inhibited S1PL, with about 40% inhibition of S1PL activity by 30 µM FTY720 (Fig. 9). Further, to confirm that the described assay is applicable not only to a microsomal protein preparation but can be converted to a more practical setup with the use of total tissue homogenate, we evaluated the dose-dependent inhibition of S1PL activity with FTY720 using 25 µg total liver homogenate per reaction (Fig. 10). An IC50 value of 52.4 ± 0.04 µM was determined for the FTY720-dependent inhibition of S1PL activity. We also found almost identical inhibition when the reaction was performed using either the total tissue homogenate or the microsomal protein preparation (Figs. 9, 10). However, to achieve a comparable rate of (2E)-hexadecenal formation, we needed to increase the amount of protein per reaction five-fold when using total liver homogenate as the source of the enzyme.</p><p>We also tested if S1P is degraded by S1PL when the co-factor for the reaction, P5P, is not added to the microsomal protein preparation. As expected, we found only minimal enzyme activity when P5P was omitted (about 20% of that at optimal conditions), which we attribute to a residual presence of P5P in the endogenous protein preparation. It is known that the endogenous P5P is bound to a multitude of apoenzymes but the P5P endogenous level is insufficient to saturate them [15]. This requires P5P supplementation when performing P5P-dependent in vitro assays [16].</p><p>In summary, we developed a sensitive assay to quantitate (2E)-hexadecenal as its semicarbazone derivative by ESI-LC/MS/MS using an isotope-dilution approach, and applied this method to characterize S1PL activity. Derivatization of (2E)-hexadecenal is performed as a one-step reaction, and the semicarbazone product is directly injected for LC/MS/MS analysis without any further purification. The method requires a correction for (2E)-hexadecenal biotransformation during the S1PL reaction at microsomal protein concentrations above 5 µg per reaction. The correction is performed by quantifying the disappearance of the standard (2E)-hexadecenal at the same experimental condition when the substrate (S1P) is absent. The described approach allows the use of total tissue homogenates as the source of the enzyme and demonstrates for the first time the IC50 of 52.4 µM for FTY720 as the inhibitor of S1PL. In overall, the method is ideal to search for novel S1PL inhibitors, to look for S1PL association with different pathologies, and to characterize the effect of xenobiotics or genetic manipulations on the enzyme activity.</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>

PubMed Author Manuscript

Structure-Based Design, Synthesis, and Biological Evaluation of Dihydroquinazoline-Derived Potent \xce\xb2-Secretase Inhibitors

Structure-based design, synthesis, and biological evaluation of a series of dihydroquinazoline-derived \xce\xb2-secretase inhibitors incorporating thiazole and pyrazole-derived P2-ligands are described. We have identified inhibitor 4f which has shown potent enzyme inhibitory (Ki = 13 nM) and cellular (IC50 = 21 nM in neuroblastoma cells) assays. A model of 4f was created based upon the X-ray structure of 3a-bound \xce\xb2-Secretase. The model revealed critical interactions in the active site.

structure-based_design,_synthesis,_and_biological_evaluation_of_dihydroquinazoline-derived_potent_\x

<p>β-Secretase (BACE1, memapsin 2) is an important molecular target for the treatment of Alzheimer's disease (AD).1 This membrane-bound aspartic protease catalyzes the cleavage of β-amyloid precursor protein (APP) leading to the formation of amyloid-β-peptide (Aβ) in the brain. The neurotoxicity of Aβ leads to brain inflammation, neuronal death, dementia, and AD.2,3 Early on, we designed a potent substrate-based transition-state inhibitor 1.4 An X-ray structure of 1-bound β-secretase led to the structure-based design of a series of BACE1 inhibitors.5 Over the years, we and others have designed a variety of BACE1 inhibitors incorporating novel peptidomimetic and nonpeptide scaffolds.6–8 A number of BACE1 inhibitors have been shown to reduce Aβ-levels in transgenic AD mice. Furthermore, we recently showed that administration of β-secretase inhibitor GRL8234 (2) rescued the decline of cognitive function in transgenic AD mice.9 While the development of a clinically effective inhibitor has not yet emerged, important progress has been made with small-molecule inhibitors belonging to different structural classes.1</p><p>In 2007, Baxter and co-workers reported an interesting class of 2-amino-dihydroquinazoline-derived BACE1 inhibitors.10 A representative example is inhibitor 3a (Figure 1) with reported a Ki of 11 nM. An X-ray structure of 3a-bound β-secretase provided the important molecular interactions responsible for its high affinity.10 Based upon this reported X-ray structure, we have designed a number of functionalities and heterocyclic scaffolds, to make specific interactions in the β-secretase active site to improve enzyme affinity and potency. Herein, we report the design, synthesis and evaluation of a series of potent 2-amino-dihydroquinazoline-derived inhibitors incorporating urethanes and heterocycles such as pyrazoles and thiazoles. A number of compounds exhibited potent β-secretase inhibitory activity and displayed good potency in cellular assays. To obtain molecular insight into the possible ligand-binding site of β-secretase, we have created an energy-minimized model of 4f based upon the 3a-bound β-secretase X-ray structure.</p><p>The synthesis of various dihydroquinazoline derivatives is shown in Scheme 1. The Boc-protected (R)-cyclohexylglycine methyl ester 4 was reduced by DIBAL-H at −78 °C and the resulting aldehyde was reacted with (ethoxycarbonylmethylene)- triphenylphosphorane in a mixture of CH2Cl2 and methanol to provide the corresponding α,β-unsaturated ester. The ester was subjected to hydrogenation over Pd-C for 12 h to provide the corresponding saturated derivative. Saponification of the ester with aqueous LiOH afforded the acid 5 in 70% overall yield. For the synthesis of quinazoline derivative 3a, acid 5 was coupled with N-cyclohexylmethyl amine to provide amide 6 in 77% yield. Removal of the Boc -group was carried out by treatment with trifluoroacetic acid and the resulting amine was subjected to reductive amination with aldehyde 710,11 to afford nitro compound 8 in 62% yield. Hydrogenation of 8 over 10% Pd-C in ethyl acetate at 23 °C provided the corresponding amine. Reaction of this resulting amine with BrCN in ethanol at reflux furnished dihydroquinazoline derivative 3a in 84% yield. For the synthesis of inhibitor 3b, (R)-phenylglycinol derivative 9 was oxidized under Swern conditions and the resulting aldehyde was reacted with (ethoxycarbonylmethylene)triphenylphosphorane in THF to provide α,β-unsaturated ester 10. Hydrogenation of the ester over 10% Pd-C in ethyl acetate and saponification of the ester provided the corresponding acid which was coupled with N-cyclohexylmethyl amine to provide amide 11 in 53% overall yield. Amide 11 was converted to dihydroquinazoline derivative 3b by following the same sequence of reactions as 3a.</p><p>We have investigated the potential of various urethane derivatives as BACE1 inhibitors. A representative synthesis of urethane derivative 3c is shown in Scheme 2. Phenylglycinol derivative 9 was reacted with 4-nitrophenylchloroformate to provide the corresponding mixed carbonate.12 Reaction of N-cyclohexylmethyl amine with this mixed carbonate afforded urethane derivative 12 in 90% yield in two steps. Removal of the Boc group with trifluoroacetic acid and reductive amination of the resulting amine with aldehyde 7 furnished nitro derivative 13 in 53% yield, in two steps. Hydrogenation of nitro compound 13 over 10% Pd-C in ethyl acetate followed by exposure of the resulting amine to BrCN afforded inhibitor 3c. Urethane derivatives 3d and 3e were prepared by an analogous procedure using Boc protected (R)-cyclohexyl glycinol as the starting material. For synthesis of urethane 3f, racemic amino ester 14 was prepared using a multicomponent reaction developed in our laboratory.13 Reaction of 14 with Boc2O in the presence of DMAP in CH3CN at 23 °C afforded the corresponding Boc derivative. Treatment of the resulting ester with magnesium powder in methanol under sonication afforded the corresponding Boc amino ester.14 Reduction of the resulting ester with LAH afforded the racemic alcohol 15 in 63% overall yield. This alcohol was converted to urethane derivative 3f and amide derivative 3g as described above.</p><p>Based upon the X-ray structure of 3a-bound β-secretase, we planned to incorporate various heterocyclic derivatives in place of the methyl group in 3a to interact with the β-secretase active site. We were particularly interested in designing inhibitors with thiazole and pyrazole derivatives since these heterocycles contain hydrogen bond donor and acceptor groups for interactions in the enzyme active site. Also, these heterocycles are inherent to numerous bioactive natural products and FDA approved therapeutic agents.15, 16 Accordingly we designed a number of inhibitors containing thiazole and pyrazole derivatives. The synthesis of various dihydroquinazoline derivatives are shown in Scheme 3. N-cyclohexyl derivative 16 was prepared by reductive amination of the corresponding 4-thiazolecarboxaldehyde17 and cyclohexyl amine. Similarly, amines 17–20 were prepared from the corresponding aldehydes.18–21 Various known pyrazole aldehydes22 were converted to amines 21–23 by reductive amination with cyclohexylamine. For synthesis of dihydroquinazoline 4b, acid 5 was coupled with amine 16 in the presence of EDC, and HOBt to provide amide 24. Removal of Boc and subsequent reductive amination with aldehyde 7 furnished nitro derivative 25 in 65% overall yield. This was converted to quinazoline 4b as described above. Other inhibitors 4c–4i were prepared following a similar protocol as 4b.</p><p>The β-secretase inhibitory activity of various inhibitors was determined against recombinant β-secretase using our previously reported assay protocols.23 The results are shown in Table 1. As can be seen, our synthetic dihydroquinzoline derivative 3a has shown Ki value of 25 nM. The corresponding phenyl derivative 3b exhibited nearly a 6-fold lower enzyme inhibitory potency. We have then investigated the corresponding urethane derivative 3c. However, this inhibitor displayed nearly a 5-fold loss of potency compared to 3b. However, the corresponding cyclohexyl urethane derivative 3d, improved potency by 20-fold over 3c (entry 4). The presence of methyl group is important as the cyclohexyl urethane derivative 3e is less potent. We have also incorporated tetrahydropyran ring in place of the cyclohexyl group in 3d. As shown, racemic mixture (1:1) 3f has shown reduced potency over cyclohexyl derivative 3d. We have also examined the effect of a ring oxygen in carboxamide derivative 3g. This derivative too lost nearly 5-fold potency compared to cyclohexyl derivative 3a. We have also evaluated the cellular inhibition of β-secretase in neuroblastoma cells.24 Inhibitor 3a has shown an average cellular IC50 value of 71 nM. The corresponding phenyl derivative displayed an IC50 of 482 nM. The urethane derivative 3c was significantly less potent compared to inhibitor 3b. Similarly, urethane derivative 3f showed an IC50 value of 1·1 μM (entry 6).</p><p>Based upon the reported X-ray structure of 3a-bound β-secretase, we have incorporated various thiazole and pyrazole-derived ligands in an effort to interact with residues in the β-secretase active site. As can be seen in Table 2, incorporation of (2-methylthiazol-4-yl)methyl substituent in place of the methyl group in 3d, resulted in nearly a 40-fold loss of enzyme inhibitory activity (entry 1). The corresponding amide derivative 4b also exhibited a loss of potency over 3a (entry 2). Interestingly, this compound displayed poor cellular β-secretase activity in neuroblastoma cells over 3a. Introduction of a methyl group on the methylene side chain in 4b resulted in 4c which showed a loss of potency (entry 3). Furthermore, chain elongation in 4d and incorporation of (2-isopropylthiazol-4-yl)methyl substituent in 4e did not improve potency (entries 4 and 5). Interestingly, incorporation of a methoxymethyl substituent on the thiazole side chain in 4f resulted in nearly a 6-fold improvement of enzyme activity over 4b. Furthermore, inhibitor 4f has shown very potent cellular inhibitory properties in neuroblastoma cells (entry 6). We have also investigated various substituted pyrazolylmethyl groups (entries 7–9). Methyl substituted pyrazole derivative 4g is more potent than the unsubstituted derivative 4h in both enzyme inhibitory and cellular assays. Incorporation of methyl group on the pyrazole ring in 4i did not improve potency over the alkylated or unalkylated derivatives.</p><p>To gain insight into specific ligand-binding site interactions, an energy minimized model structure of 4f was created in the active site of BACE1, based upon the crystal structure of 3a-bound β-secretase.10 The conformation of 4f was optimized using the CHARMM force field.25 As shown in Figure 2, 2-amino dihydroquinazoline functionality forms a unique hydrogen bonding network with the catalytic aspartic acids Asp32 and Asp228 and the (S)-cyclohexyl group nicely filled in the S1′-subsite as reported in the X-ray structure.10 The 2-methoxymethylthiazole moiety appears to fill in the hydrophobic pocket in the S2-subsite. Furthermore, the methoxy oxygen is within proximity to form a hydrogen bond with Thr232. This may explain a 6-fold enhancement of enzyme inhibitory potency over the 2-methylthiazole derivative 4b. The P1-cyclohexamide fits well into the S1-site of β-secretase.</p><p>In summary, we have carried out structure-based modifications of 2-amino-3,4-dihydroquinazoline -derived β-secretase inhibitors. In particular, we have incorporated thiazole and pyrazole-based P2-ligands to make specific interactions in the S2-subsite. These efforts resulted in inhibitors with improved potency and cellular inhibitory properties compared to methyl-substituted inhibitor 3a. Inhibitor 4f has shown enhanced enzyme inhibitory activity as well as very good cellular inhibitory potency in neuroblastoma cells. A protein-ligand X-ray structure-based model of 4f-bound β-secretase has provided important molecular insight into the ligand-binding site interactions. Further design and improvement of inhibitor properties are currently in progress.</p>

PubMed Author Manuscript

Electrochemistry in Media of Exceptionally Low Polarity: Voltammetry with a Fluorous Solvent

This work demonstrates the first cyclic voltammetry in a perfluorocarbon solvent without use of a cosolvent. The novel electrolyte tetrabutylammonium tetrakis[3,5-bis(perfluorohexyl)phenyl]borate (NBu4BArF104; 80 mM) allows for voltammetry of ferrocene in perfluoro(methylcyclohexane) by lowering the specific resistance to \xce\xa9268 k cm at 20.8 \xc2\xb0C. Despite significant solution resistance, the resulting voltammograms can be fitted quantitatively without difficulty. The thus determined standard electron transfer rate constant, k\xc2\xb0, for the oxidation of ferrocene in perfluoro(methylcyclohexane) is somewhat smaller than for many solvents commonly used in electrochemistry, but can be explained readily as the result of the viscosity and size of the solvent using Marcus theory. Dielectric dispersion spectroscopy verifies that addition of NBu4BArF104 does not significantly raise the overall polarity of the solution over that of neat perfluoro(methylcyclohexane).

electrochemistry_in_media_of_exceptionally_low_polarity:_voltammetry_with_a_fluorous_solvent

1. Introduction<!>2.1 Chemicals<!>2.2 Resistance Measurements<!>2.3 Voltammetry<!>2.4 DOSY 19F NMR<!>2.5 Dielectric Dispersion Spectroscopy<!>3.1 Electrolyte Characteristics<!>3.2 Electrochemical Window<!>3.3 Voltammetric Measurements<!>3.4 Heterogeneous Rate Constant<!>3.5 Dielectric Spectroscopy<!>4. Conclusions<!>5. Supplementary Material<!>

<p>Organic solvents have found extensive application as solvents for electrochemistry [1]. Their wide-ranging solvent environments provide electroactive species with distinctive solubility, stability, and reactivity characteristics [1,2]. In particular, through the development of electrochemically stable lipophilic electrolytes [3], electrochemistry was demonstrated with low polarity solvents such as cyclohexane [4], heptane [5], and super-critical CO2 [6]. While many of these solvents have no permanent dipole moment, they still exhibit a significant degree of polarizability. Fluorous solvents, on the other hand, have a very low polarizability, resulting in very weak van der Waals forces between their molecules and making them the most nonpolar solvents known [7]. For example, on the π* scale of solvent polarity/polarizability [8], where dimethylsulfoxide is defined as 1 and cyclohexane 0, perfluorooctane can be found at −0.41. In fact, octane and perfluorooctane do not mix at room temperature precisely because octane is too polarizable [9].</p><p>This extremely low polarity makes fluorous phases very useful for probing the electrochemical behavior of species at near gas-phase conditions [10]. Advantages also arise from the exceptional level of chemical inertness of perfluorocarbons and—in view of spectroelectrochemistry [11,12]—their optical transparency down to 160 nm and the low absorbance in a wide range of the IR spectrum [11]. Since fluorous phases have become important tools in catalytic synthesis [7,13] and separation techniques [14], electrochemistry in these media could allow online monitoring of reaction progress and separation efficiency. Other possible applications lie in battery technology and the study of fluorous monolayers on macroscopic surfaces [7,15] and nanoparticles [7,16]</p><p>Despite the many possible uses of electrochemistry with fluorous solvents, voltammetry in a perfluorocarbon solvent has not been reported yet. Even the most lipophilic salts known are only sparingly soluble in a fluorous solvent, and any salt that does dissolve is heavily ion-paired [17,18]. Geiger and LeSuer demonstrated voltammetry with perfluoro(methylcyclohexane) (1) and an electrolyte consisting of the tetrakis(3,5-trifluoromethyl)phenyl borate anion and an imidazolium cation with a fluorous ponytail, but they had to use the comparatively polar benzotrifluoride with a dielectric constant of 9.2 as cosolvent in a 1:1 ratio to perfluoro(methylcyclohexane) [19].</p><p>We demonstrated previously the first potentiometric sensors based on fluorous membranes using perfluoro(perhydrophenanthrene) as the membrane matrix and the sodium salt of tetrakis[3,5-bis(perfluorohexyl)phenyl]borate to provide for ionic sites [17,18]. The receptor-free and receptor-doped [20] sensors displayed excellent selectivities and show promise to reduce biofouling. However, the low solubility of this salt (1 mM) and strong ion pairing resulted in bulk resistances too large for any electrochemical technique requiring significant current. Using the novel fluorophilic electrolyte prepared from this tetrakis[3,5-bis(perfluorohexyl)phenyl]borate and the tetrabutylammonium cation (NBu4BArF104) (2), we report in this work the first demonstration of voltammetry in an undiluted perfluorocarbon solvent, specifically perfluoro(methylcyclohexane). This compound is an excellent representative of fluorous solvents as its π* value (−0.48) and dielectric constant (1.86) are very low even among fluorous phases [21]. With its boiling point of 76 °C at 1.00 atm (partial pressure 0.139 atm [22] at 25 °C), perfluoro(methylcyclohexane) is much more convenient to work with than the more volatile perfluorohexane or perfluorocyclohexane. Moreover, unlike the commercially available perfluoro(dimethylcyclohexane) and perfluoro(perhydrophenanthrene), perfluoro(methylcyclohexane) is not a mixture of isomers.</p><!><p>All reagents were of the highest commercially available purity. Perfluoro(methylcyclohexane) and tetrabutylammonium chloride were used as received from Alfa Aesar (Ward Hill, MA, USA) and Fluka (Milwaukee, WI, USA), respectively. Sodium tetrakis[3,5-bis(perfluorohexyl)phenyl] borate tetrahydrate was prepared according to a previously described procedure [17].</p><p>NBu4BArF104 was synthesized by metathesis from tetrabutylammonium chloride and sodium tetrakis[3,5-bis(perfluorohexyl)phenyl]borate. 10.0 g of sodium tetrakis[3,5-bis(perfluorohexyl)phenyl]borate tetrahydrate and 1.03 g of tetrabutylammonium chloride were added to a separatory funnel containing 300 mL water and 300 mL FC-72 (perfluorohexanes). The mixture was shaken until all of the salt dissolved. The FC-72 layer was collected, washed three times with 300 mL water, dried with MgSO4 and filtered. The solvent was removed by rotary evaporation and further drying under vacuum at 75 °C for 48 hours, yielding NBu4BArF104 as a colorless, wax-like material in quantitative yield. 1H NMR (500 MHz, FC-72, δ): 0.507 (t, JHH = 7.2 Hz, –CH3, 12H), 0.75–0.90 (m, –CH2CH3, 8H), 1.04–1.19 (m, NCH2CH2–, 8H), 2.42–2.58 (m, NCH2–, 8H), 7.31 (s, p-ArH, 4H), 7.51 (s, o-ArH, 8H). The corresponding tetraethyl- and tetrahexylammonium salts were prepared analogously. Differential scanning calorimetry (DSC, see Supplementary Material for details) revealed a glass-transition temperature of −21 °C, showing that 2 is a highly viscous ionic liquid at room temperature.</p><!><p>The specific resistance of the 80 mM NBu4BArF104/perfluoro(methylcyclohexane) solution was measured by impedance spectroscopy using a homemade cell. A Solartron SI 1287 electrochemical interface was used in combination with a Solartron 1255B frequency response analyzer (Solartron Analytical, Farnborough, Hampshire, England) at an AC amplitude of 100 mV, swept over a frequency range from 100,000 Hz to 10 Hz, using a DC offset equal to that of the open circuit potential. Solution resistance was determined by a fit of the resulting complex impedance plot to a model circuit consisting of a resistor in series with a parallel resistor/capacitor. The cell was made of two polished stainless steel disks separated by a Teflon ring (0.2 cm thick, 0.7 cm inner diameter). In each measurement, the fluorous solution was injected into the cell, the cell was placed in a Teflon pocket, and the pocket was suspended in a water bath thermostated at 20.8 °C. All samples were placed in the bath for at least 20 minutes prior to measurement. The cell constant was determined by a KCl conductivity standard purchased from Sigma-Aldrich (St. Louis, MO, USA).</p><!><p>Voltammetry was performed with a CHI600C Potentiostat in combination with a CHI200B Picoamp Booster and Faraday Cage (CH Instruments, Austin, TX, USA). A 10 μm diameter Pt microelectrode from Bioanalytical Systems (BAS; West Lafayette, IN, USA) was used as the working electrode, a Au disk electrode (BAS) as the auxiliary electrode, and a silver wire (>99%, Alfa Aesar) as a quasi-reference electrode. Cyclic voltammograms (CVs) of ferrocene were performed at concentrations of 5.43, 2.72, 1.36, and 0.68 mM. Each concentration was scanned at 10, 50, 100, and 1000 mV/s. The resultant CV curves were then fitted using the DigiElch software package (ElchSoft, Kleinromstedt, Germany). All voltammograms are resistance corrected using the method outlined in the Supplementary Material before fitting.</p><p>Polishing equipment was purchased from Buehler (Lake Bluff, IL, USA). Prior to each experiment, the working electrode was polished on Microcloth polishing pads, first with 5.0 μm Micropolish II deagglomerated alumina, then with 1.0 μm and 0.25 μm MetaDi Supreme polycrystalline diamond suspension, and lastly with 0.05 μm Micropolish II deagglomerated alumina. The polished electrode was then ultrasonicated in a Triton X-100 detergent solution for 3 minutes, rinsed, and dried. Water used for the detergent solution and for rinsing the electrode was deionized and charcoal-treated (≥18.2 MΩ cm specific resistance) with a Milli-Q PLUS reagent-grade water system (Millipore, Bedford, MA, USA). Typically, samples were not purged of oxygen in order to minimize loss of solvent (purging did not have a significant effect on the shape of the CV in the potential range necessary to observe ferrocene), except in the case of the measurement of the electrochemical window (Figures 2 and 3), in which case the samples were purged with solvent-saturated argon for 30 min prior to measurement.</p><!><p>DOSY 19F NMR was performed on a Varian INOVA 300 spectrometer equipped with a 4-nucleus probe capable of pulsed field gradients operating at 282.12 MHz. A bipolar pulse pair stimulated echo (BPP-STE) sequence, as described in ref. 23, was used to determine the diffusion coefficients of tetrakis[3,5-bis(perfluorohexyl)phenyl]borate and the solvent in a solution of 80 mM NBu4BArF104 in perfluoro(methylcyclohexane). The field gradient was calibrated using the self-diffusion coefficient of 6.1 × 10−6 cm2 s−1 for perfluoro(methylcyclohexane) at 20 °C. This value was calculated using equation 1 from the self diffusion coefficient for 25 °C, as reported in ref. 24.</p><!><p>Dielectric dispersion spectroscopy was performed with an HP8720 network analyzer and an open-ended coaxial [25] HP85070B probe (both Hewlett-Packard). In addition, a TE01-mode cylindrical dielectric resonator similar to the ones described in ref. 26 was used to measure the complex permittivity at 2.45 GHz. The value of 1.95 measured with the latter probe for the reference heptane agrees well with the literature value of 1.97 [27]. For perfluoro(methylcyclohexane), the fluorous electrolyte solution, and heptane as reference, a consistent discrepancy of 0.30 was noted for the measurements with the two probes. Reasons for the discrepancy may be power reflection from the extremities of the specimen back to the probe face, or an air gap between the open-ended coaxial probe and the liquid sample [28,29], caused by surface roughness and the poor wettability of the Ni-coated probe with liquids of low polarity. The resulting shift in the dielectric constant (capacitive in nature in case of the latter explanation) was estimated from the TE01-mode cylindrical dielectric resonator measurements of heptane, and applied as a frequency independent correction for all measurements performed with the open-ended coaxial-line probe.</p><!><p>NBu4BArF104 was found to have a high solubility (greater than 80 mM) in perfluoro(methylcyclohexane), perfluorohexanes, perfluoroheptanes, perfluorodecalin, and perfluoro(perhydrophenanthrene), and a solubility of 8 mM in n-perfluorohexane. An 80 mM solution of NBu4BArF104 in perfluoro(methylcyclohexane) was measured to have a specific resistance of 268 kΩ cm at 20.8 °C. To explore how the alkyl substituents on the tetraalkylammonium cation affect the properties of BArF104− salts, tetraethylammonium (NEt4+) and tetrahexylammonium (NHx4+) salts of BArF104− were also prepared. NEt4BArF104 was found to have a high solubility in 1, but an 80 mM solution of NEt4BArF104 in 1 exhibited a solution resistance approximately five times greater than solutions with 2 as the supporting electrolyte. This is indicative of stronger ion pairing for NEt4BArF104 than for the tetrabutylammonium salt 2. On the other hand, NHx4BArF104 was shown to have a high solubility in 1 (greater than 80 mM) but a somewhat lower solubility than 2 in n-perfluorohexane and perfluorodecalin. It follows that among the three tested electrolytes, NBu4BArF104 is the most advantageous one since its cation is large enough to weaken ion pairing but not too large to compromise the solubility in fluorous solvents.</p><!><p>The electrochemical window provided by an electrolyte solution of 80 mM NBu4BArF104 in perfluoro(methylcyclohexane) spans 4.2 V, from +1.9 to −2.3 V vs. the ferrocenium/ferrocene couple (Fc+/Fc, Fig. 2). While the reduction limit is similar to that reported by Geiger and LeSuer for the 1:1 perfluoro(methylcyclohexane)/benzotrifluoride mixed solvent system [19], the oxidation limit of the perfluoro(methylcyclohexane) solution is 0.6 V greater than that of the solvent mixture.</p><p>To determine the cause of the reduction limit (i.e., to determine whether it is caused by the solvent or by the electrolyte salt), the reduction potentials of perfluoro(methylcyclohexane) and the electrolyte were determined individually in a Na-distilled THF solution containing 0.1 M NBu4ClO4. Each solution was purged for 30 min with Ar before measurements were taken. Addition of 20 mM NBu4BArF104 showed no significant change in the cyclic voltammogram. However, addition of perfluoro(methylcyclohexane) yielded an ill-defined reduction peak at −2.9 V vs. Fc+/Fc (Figure 3), which is identical to a previously reported value for reduction of this compound in THF solution [30]. This shows that the reduction limit for our fluorous electrolyte solutions is given by the fluorous solvent. Since it has been shown that branched perfluorocarbons are more readily reduced than unbranched ones [30, 31], it appears likely that use of a monocyclic or straight-chain perfluorocarbon as fluorous solvent would further extend the electrochemical window.</p><p>To explore the source of the oxidation limit, a solution of 0.1 M NBu4PF6 in anhydrous acetonitrile was used as the electrochemical solvent medium, which exhibited an electrochemical window from +2.9 V to −2.6 V vs. Fc+/Fc. Samples were again purged with argon prior to voltammetric measurements. Upon addition of 25 mM NBu4BArF104 or 25 mM perfluoro(methylcyclohexane), no new oxidation peak could be observed, showing that the oxidation of both NBu4BArF104 and perfluoro(methylcyclohexane) in acetonitrile must occur at positive potentials that are outside the electrochemical window of acetonitrile. This indicates that the oxidation limit in the electrochemical window of the 80 mM NBu4BArF104/perfluoro(methylcyclohexane) solution is not caused by the electrolyte or the solvent. Interestingly, upon taking a voltammogram of the NBu4BArF104/perfluoro(methylcyclohexane) solution from −2.30 to +7.45 V vs Fc+/Fc, a peak of relatively low intensity was observed with a half-wave potential of +1.9 V vs Fc+/Fc (see Supplementary Material), suggesting that the oxidative limit of the electrochemical window is caused by a species of comparatively low concentration. Since this peak was observed even when perfluoro(methylcyclohexane) was purified by slow filtering through silica gel and storage over KOH, the impurities of perfluoro(methylcyclohexane) that are responsible for the oxidation limit of the electrochemical window do not appear to be polar in character. Moreover, since 1H NMR spectra of commercial perfluoro(methylcyclohexane) show the presence of hydrogen-containing impurities in the mM range (in comparison, the solvent has a self concentration of 4.96 M), it may be that a low concentration of imperfectly perfluorinated solvent molecules with one (or more) hydrogens is responsible for the oxidative limit.</p><!><p>Figure 4 shows CVs of 5.43 mM ferrocene in 80 mM NBu4BArF104 at various scan rates. The CV measured with a scan rate of 10 mV/s exhibits, in both the forward and backward scan, the typical shape expected for hemispherical diffusion at a microelectrode, i.e., the current reaches a plateau. At scan rates of 100 and 1000 mV/s, the forward scan still has the characteristics of hemispherical diffusion, while the reverse scan exhibits a peak maximum indicative of planar diffusion to the electrode. This suggests that the diffusion coefficient of Fc, which is oxidized in the forward scan, is significantly larger than the diffusion coefficient of Fc+, which is reduced in the backward scan. Indeed, upon fitting of the voltammograms, the diffusion coefficient for Fc was determined to be (2.05±0.01) × 10−6 cm2 s−1 while that of Fc+ was determined to be (2.85±0.01) × 10−7 cm2 s−1. This disparity in diffusion coefficients is not surprising since in a fluorous solvent the cation Fc+ is expected to form very stable ion pairs (and possibly even higher aggregates) with the electrolyte ions. Stability constants for ion pair formation in these solvents (log Kip values from 14 to 16) [17,18,20] are among the strongest ones reported for any condensed phase, which is explained by the extremely low polarity/polarizability of fluorous media.</p><p>The accuracy of the diffusion coefficient of Fc+ as determined by fitting of the CVs is supported by the observation that it is within 5.0% of the diffusion coefficient of the fluorophilic borate in 80 mM NBu4BArF104 [(3.0 ± 1.5) × 10−7 cm2 s−1], as determined by 19F Diffusion-Ordered NMR SpectroscopY (DOSY). This result is anticipated when considering that the ion pairs of the fluorophilic borate with tetrabutylammonium likely have very similar geometrical dimensions as the ion pairs of the fluorophilic borate with Fc+. The DOSY 19F NMR data along with the Stokes-Einstein equation</p><p> (1)D=kBT6πrη also permit a comparison of the effective radii of the BArF104− anion, ranion, and perfluoro(methylcyclohexane), rsolvent. Using equation 5, it follows that ranion/rsolvent = Danion/Dsolvent for a solution of the anion. Since the solution viscosity, η, is the same for both species, it does not affect ranion/rsolvent. With the experimental values of Danion and Dsolvent (3.0 × 10−7 and 4.2 × 10−6 cm2 s−1, respectively), ranion/rsolvent is calculated to be 14. This value is larger than what would be expected for an ion pair of BArF104− and NBu4+, confirming that not only ion pairs but also larger ionic aggregates are formed.</p><p>Figure 5 shows CVs of ferrocene in the concentration range from 0.68 to 5.43 mM. As anticipated, the CVs overlap very well upon normalization (see Supplementary Material), and the limiting oxidation current is directly proportional to the ferrocene concentration (see inset, Fig. 4). The CVs are indicative of a quasi-reversible electron transfer, as suggested by the asymmetry of the oxidation wave and supported by the fits that give a heterogeneous rate constant (k°) of (7.13±0.04) × 10−4 cm s−1. The transfer coefficient determined through fitting was found to be 0.611±0.002. This value falls well within the range of 0.45 to 0.74 that has been published for Fc+/Fc in a variety of solvents from dichloromethane to methanol [32]. As shown in Figure 6, the predicted voltammetric curves overlap well with the experimental ones.</p><!><p>The k° of (7.13±0.04) × 10−4 cm s−1 as determined by the above described fitting is somewhat smaller than the 0.03 to 8.4 cm/s for ferrocene oxidation at a Pt electrode in solvents ranging from chloroform to acetonitrile, as previously reported [32]. Two possible reasons for this observation appeared to be specific adsorption of the electrolyte onto the electrode surface [33,34], thereby inhibiting the electron transfer, or a slow rearrangement of the solvent and electrolyte, as described by the Marcus theory of electron transfer [35,36,37,40].</p><p>To test for the formation of an adsorbed electrolyte layer, a voltammogram of 80 mM NBu4BArF104/perfluoro(methylcyclohexane) was measured at 10 V/s in the range of −1.1 to 0.9 V vs. Fc+/Fc (see Supplementary Material). The results do not confirm the formation of an adsorbed layer since peaks indicative of adsorption or desorption events were not observed. However, the formation of such a layer cannot be excluded entirely since a particularly strongly adsorbed layer might not be desorbed in the potential range accessible.</p><p>An interpretation of k° based on the Marcus theory is more straightforward. The effect of solvent dynamics on k° has been studied with great detail and has been experimentally confirmed for solvents with a permanent dipole [35,36]. Modeling of the relationship between k° and the solvent dynamics involves the equation [35,37]: (2)ko=KpκELτL−1(ΔGOS∗4πRT)1/2exp(−ΔG∗RT) where Kp is the precursor formation constant, κEL is the adiabicity parameter, τL is the solvent longitudinal relaxation time, ΔGOS* is the free energy of activation for outer sphere reorganization, and ΔG* is the sum of the inner and outer sphere reorganization free energies of activation. The longitudinal relaxation time is related to the Debye relaxation time (τD) by [38]</p><p> (3)τL=(ε∞εs)τD where ε∞ is the dielectric constant in the infinite frequency limit and εs is the static dielectric constant. For perfluoro(methylcyclohexane) at 20 °C, ε∞ and εs are 1.859 and 1.85, respectively [39]. In cases where it is not known, τD can be estimated using the relation [38]</p><p> (4)τD=4πα3ηkBT where α is the radius of the solvent molecule and η is the solution viscosity. The combination of equations 1 to 3 gives: (5)ko=KpκELεskBTε∞4πα3η(ΔGOS∗4πRT)1/2exp(−ΔG∗RT)</p><p>It has been reported previously that for ferrocene as the redox-active analyte and a variety of solvents ranging from chloroform to acetonitrile [32,41] (see also Figure 7) a plot of log k° versus log τL shows the expected linear relationship. This shows that for ferrocene the term KpκEL (ΔGOS*)1/2 (4πRT)−1/2 exp(−ΔG*/RT) has only a small dependence on the solvent.</p><p>In order to apply equation 5 to the fluorous electrolyte solutions, it had to be determined to what extent the addition of electrolyte affects the viscosity of perfluoro(methylcyclohexane). Upon addition of 80 mM NBu4BArF104 to perfluoro(methylcyclohexane), the diffusion coefficient of the solvent, as determined by DOSY 19F NMR, decreased only 33% from 6.1 × 10−6 cm2 s−1 [24] to (4.19 ± 0.50) × 10−6 cm2 s−1, which suggests that the addition of electrolyte only moderately increases the viscosity of the solution. Because the k° determined in this work is 42 and 1.2 × 104 times smaller than for chloroform and acetonitrile, respectively, any effect of the electrolyte on the viscosity of perfluoro(methylcyclohexane) is comparatively small. Therefore, the published value for η of perfluoro(methylcyclohexane) of 1.56 cP [24] was used for all further calculations.</p><p>Using literature values for the viscosity and self-diffusion coefficient (6.2 ×10−6 cm2 s−1) of perfluoro(methylcyclohexane) along with equation 1, a radius of 2.26 Å is obtained for perfluoro(methylcyclohexane). This radius along with equations 3 and 4 gives τL as 56 ps at 20 °C and permits the comparison of the relationship of k° and τL from this work with corresponding data for non-fluorous solvents. Figure 7 illustrates that the experimentally determined k° for the oxidation of ferrocene in the fluorous electrolyte solution falls on the same line as k° values for non-fluorous solvents from the literature. This shows that while k° for the oxidation of ferrocene in the fluorous electrolyte solution is smaller than for solvents commonly used in electrochemistry, its value can be readily explained as the result of the large viscosity and molecular size of the solvent perfluoro(methylcyclohexane).</p><!><p>While this is the first report on voltammetry with a cosolvent-free perfluorocarbon, the question arises whether these experiments were in fact performed with the least polar organic phase to date. Or does the addition of the NBu4BArF104 electrolyte increase the polarity of the perfluoro(methylcyclohexane) phase to an extent that the electrolyte solution has a much more polar character than the fluorous solvent alone? This question may be addressed through dielectrometry, which provides the dielectric constant as a measure of the molecular dipole moments and polarizability of a sample. Because of the inherent ionic conductivity of electrolyte solutions, the dielectric properties of the fluorous media discussed here cannot be determined in the frequency range around 100 kHz, which is typically used for conventional dielectrometry. Instead, dielectric spectroscopy in the GHz range [42] was used in this study to assess the polarity of the electrolyte solution. Since relevant dynamic processes occur in the ps and ns range, the static dielectric constant can be determined by extrapolation of the real component, ε′ (ν), of the frequency-dependent complex dielectric permittivity, ε̃ (ν), from the GHz range to zero frequency. ε̃ (ν) is given by [42]</p><p> (6)ε∼(ν)=ε′(ν)−−1ε″(ν) where ε′ (ν) stands for dielectric dispersion, and the imaginary component ε″ (ν) is the dielectric absorption. Because ε″ (ν) is related to ε′ (ν), it does not carry independent information and will not be discussed here.</p><p>As Figure 8 shows, ε′ (ν) of solutions of the electrolyte salt NBu4BArF104 in perfluoro(methylcyclohexane) in the frequency range 0.2 to 20 GHz exceeded ε′(ν) of the pure perfluorocarbon by no more than 6.5%. The linear extrapolation of ε′ (ν) in the range from 0.2 to 5.0 GHz suggests a static dielectric constant of the electrolyte solution of 1.96 and of pure perfluoro(methylcyclohexane) of 1.89. The latter value is in good agreement with the literature value of 1.86 [21].</p><!><p>Using the novel fluorous electrolyte salt NBu4BArF104, we demonstrated that voltammetry can be performed with a perfluorocarbon solvent without the use of a cosolvent. Even though fluorous phases are the least polar of all condensed phases, the observed CVs can be quantitatively fitted. The thus obtained k° is 1.6 orders of magnitude smaller than the smallest k° for oxidation of ferrocene in a set of common non-fluorous solvents, for which k° values cover a range of 2.4 orders of magnitude. However, using Marcus theory, the small k° can be readily explained as the result of the large radius and the high viscosity of perfluoro(methylcyclohexane). While ion pair formation in these fluorous phases is extremely strong and the formation of local pockets of higher polarity at the submolecular level is possible, dielectric spectroscopy confirms that the addition of electrolyte has only a minimal effect on the overall polarity of the fluorous electrolyte solutions.</p><p>The unique solvent environment of fluorous phases should provide an interesting medium for further experimentation. We are currently investigating the use of fluorous media as new matrixes [43,44] for voltammetric and amperometric sensors as they are expected to exhibit selectivity patterns differing significantly from those of conventional hydrophobic phases [14,17,45] and have the potential to reduce chemical and biological fouling. Also, in view of a further extension of the already large solvent window, we are exploring the electrochemistry of different perfluorinated and partially fluorinated solvents.</p><!><p>Details of the differential scanning calorimetry, resistance correction, determination of the oxidation potential of the electrochemical window, and a voltammogram showing the absence of a desorptive peak are described in the Supplementary Material.</p><!><p>Structures of perfluoro(methylcyclohexane) (1) and NBu4BArF104 (2).</p><p>Cyclic voltammogram (CV) of perfluoro(methylcyclohexane) containing 80 mM NBu4BArF104, scan rate = 100 mV/s, showing the electrochemical background. Data is corrected for solution resistance.</p><p>CVs of a 0.1 M NBu4ClO4/THF solution containing 0, 25, 50, or 75 mM perfluoro(methylcyclohexane): scan rate = 10 mV/s, T = 21°C.</p><p>CVs of 5.43 mM ferrocene and 80 mM NBu4BArF104 in perfluoro(methylcyclohexane) with various scan rates. Data is corrected for solution resistance.</p><p>CVs of varied concentrations of ferrocene in 80 mM NBu4BArF104/perfluoro(methylcyclohexane): scan rate = 10 mV/s, T = 21°C. The inset shows the linear relationship between the limiting current and the ferrocene concentration.</p><p>CV of 1.36 mM ferrocene normalized to the diffusion-limited current (solid) along with a fit based on α = 0.61, k° = 7.13 × 10−4 cm/s, D(Fc) = 2.05 × 10−6 cm2 s−1 and D(Fc+) = 2.85 × 10−7 cm2 s−1(dots).</p><p>Log-log plot of k° versus τL along with a linear fit for the literature data (open circles, [32,40,41]) only. The fit is extrapolated to the τL for perfluoro(methylcyclohexane) (filled circle).</p><p>Dielectric dispersion spectrum of perfluoro(methylcyclohexane) with and without 80 mM NBu4BArF104.</p>

PubMed Author Manuscript

Running wheel activity protects against increased seizure susceptibility in ethanol withdrawn male rats

Ethanol withdrawal is a dysphoric condition that arises from termination of ethanol intake by dependent individuals. Common withdrawal symptoms include anxiety, increased reactivity to stimuli and increased seizure susceptibility as well as the risk of increased seizure severity. We use an animal model of dependence and withdrawal to study withdrawal behaviors and potential underlying neurobiological mechanisms. For a number of years, we have quantified pentylenetetrazol seizure thresholds as an assessment of ethanol withdrawal at both one day and three days of withdrawal. Typically, we see a significant decrease in seizure threshold (increased sensitivity to seizure induction) that persists through three days of withdrawal for male rats. Increasing evidence indicates that voluntary exercise affords protection against various challenges to physical and psychological health, including ethanol-related challenges. Therefore, the current study investigated the effect of voluntary wheel running on seizure susceptibility following chronic ethanol administration and withdrawal. We found that voluntary wheel running attenuated the increased sensitivity to pentylenetetrazol-induced seizures observed with ethanol withdrawal, at both the one-day and three-day time points. This result was especially interesting as animals with access to the running wheels consumed more of the ethanol-containing diet. These findings showed that chronic voluntary wheel running reduces the severity of ethanol withdrawal in our animal model and suggest that exercise-based interventions may have some utility in the clinical management of heavy drinking and alcohol withdrawal.

running_wheel_activity_protects_against_increased_seizure_susceptibility_in_ethanol_withdrawn_male_r

1. Introduction<!>2.1. Animals<!>2.2. Materials<!>2.3. Activity wheel procedure<!>2.4. Liquid diet procedure<!>2.5. PTZ Seizure threshold procedure<!>2.6. Data analysis<!>3.1. Body weights and ethanol consumption<!>3.2. Running wheel activity<!>3.3. Seizure threshold<!>4. Discussion

<p>Alcohol (ethanol) withdrawal arises when an ethanol-dependent individual stops consumption. Ethanol withdrawal (EW) symptoms reflect a rebound central nervous system hyperexcitability resulting from removal of this central nervous system depressant, and include anxiety, agitation, insomnia, general dysphoria and tremors that may progress to seizures (Ballenger and Post, 1978; Goldstein and Pal, 1971). Preclinical studies with laboratory rats and mice have shown them to be useful models for studying ethanol dependence and withdrawal, as EW animals display a significant increase in seizure susceptibility and severity during early EW (Becker et al., 1997; Devaud et al., 1995a; Devaud and Morrow, 1999; Finn et al., 1995; Finn and Crabbe, 1999; Veatch et al., 2007). Seizure susceptibility measurements are used in these animal models as a quantifiable reflection of the rebound hyperexcitability that is unmasked during EW.</p><p>The CNS hyperexcitability of EW is believed to arise from neuroadaptations engendered by persistent ethanol intake. A number of brain adaptations in key neurotransmitter systems and cellular modulators occur (see Moonat et al., 2010; Olsen et al., 2007; Spanagal, 2009; Vengeliene et al., 2008 for review). Ethanol acts as a CNS depressant, largely by enhancing GABAergic transmission and inhibiting glutamatergic activity. Therefore, the homeostatic drive to limit the effects of persistent ethanol exposure results in a reduced responsiveness of GABAA receptors and increased responsiveness of glutamatergic systems, particularly NMDA receptors. These chronic ethanol-induced adaptations are believed to involve alterations in subunit composition of both of these receptor types (Alele and Devaud, 2005; Cagetti et al., 2003; Devaud et al., 1995b; 1998; Devaud and Morrow, 1999; Devaud and Alele, 2004; Mhatre and Ticku, 1994; Mehta and Ticku, 2005). While these adaptations are believed to contribute to several manifestations of withdrawal, such as the increased anxiety and seizure susceptibility, it is likely that the stress of withdrawal itself also exacerbates seizure sensitivity (Friedman et al., 2011).</p><p>Increasing evidence indicates that exercise exerts protective effects against a variety of challenges to physical and psychological health. In rats and mice, these effects can be effectively modeled by housing animals with free access to running wheels. Studies have shown that adaptation to running wheels promotes neuronal health by enhancing synaptic plasticity and neurogenesis, even in adult animals (Cotman and Berchold, 2002; Stranahan et al., 2007; van Praag et al., 1999a, 1999b). Voluntary wheel running improves the ability to manage stress exposure by reducing the HPA response and increasing production of brain growth factors, such as BDNF (Nyhius et al., 2010). These effects appear to account for the anxiolytic and antidepressant effects of running-wheel activity (Duman et al., 2008; Salam et al., 2009). Chronic voluntary wheel running has also been shown to reduce the intoxicating effects of acute ethanol administration in a mouse model (Mollenauer et al., 1991, 1992) while antagonizing both the antiproliferative (Crews et al., 2004) and neurotoxic effects of repeated binge-like ethanol administration in rats (Leasure and Nixon, 2010). Further, chronic intermittent ethanol exposure reduced running, especially during the active (night) phase (Logan et al., 2010), a finding that suggests ethanol dependence and withdrawal may reduce the beneficial effects of voluntary wheel running. Pentylenetetrazol (PTZ) is a chemoconvulsant and has been used in numerous investigations to assess seizure susceptibility in animal models. We have published a number of reports studying drug effects on PTZ seizure thresholds during EW and have now extended this approach to determine whether free access to running wheels modulates the increased sensitivity to pentylenetetrazole-induced seizures seen during EW.</p><!><p>Male CR rats (Charles Rivers Lab) were approximately 42 days old at the start of experimental procedures.</p><!><p>Pentylenetetrazol (PTZ) from Sigma-Aldrich (St. Louis, MO) was dissolved in normal saline at a concentration of 5 mg/ml.</p><!><p>Animals were individually housed and randomly assigned to one of three running wheel conditions: (1) standard rat cages without wheels (No Wheel), (2) standard rat cages with wheels that are locked (Locked Wheel) or (3) standard rat cages with functioning running wheels (Free Wheel). The locked wheel condition was included to separate the possible effects of environmental complexity from exercise. The wheel condition was constant for 24 h each day throughout the course of the experiment. Running wheel activity was recorded by use of an external electronic LCD counter that was attached to the side of each free running wheel cage. Activity was recorded twice daily at 7:00 a.m. (start of rest phase) and at 5:00 p.m. (start of active phase). We chose these two times as lights on (7:00 a.m.) and approaching lights off (7:00 p.m.). Counters were manually reset after obtaining counts. All animals were housed under their respective conditions for 10 days prior to introduction of the liquid diets to allow for acclimation and adaptation to running wheels in Free Wheel animals.</p><!><p>Animals were made ethanol-dependent by administration of 6% ethanol, v/v, in a nutritionally complete liquid diet, which was slightly modified from the Frye liquid diet (Frye et al., 1983). Diet components were purchased individually with diet made at least twice per week and fresh diet was provided daily (MP Biomedical, Costa Mesa, CA) and administered for 14 days as previously described (Devaud and Morrow, 1994; Devaud et al., 1995a). Control animals were pair-fed the same liquid diet but with dextrose substituted isocalorically for the ethanol to ensure equivalent caloric intake and comparable nutritional status. The amount of liquid diet consumed was recorded daily.</p><p>After 14 days of liquid diet administration, the liquid diet was removed and regular lab chow provided ad libitum to all animals to maintain equivalent diet conditions. Seizure threshold testing was scheduled at 1 day or 3 days EW. All procedures were conducted in accordance with approved University of Maine Animal Welfare Protocols and NIH guidelines for the humane care and use of animals in an AAALAC-accredited facility.</p><!><p>Constant tail vein infusion of the chemoconvulsant was used for the induction of seizures. A 25 g butterfly needle was inserted into a lateral tail vein while the animals were gently restrained and needle taped into place. The animal was then allowed to move freely while the observer gently held the tip of its tail. PTZ was infused at 1.6 ml/min and the time to the first myoclonic twitch of the face and/or neck indicated the endpoint of infusion (Alele and Devaud, 2007). Seizure thresholds were calculated from the time of infusion (minutes) times the dose (5 mg/ml×1.6 ml/min) of PTZ infused per body weight of the animal and are presented as mg PTZ per kilogram body weight.</p><!><p>Data were analyzed using one-way and two-way ANOVA, and post-hoc pair wise comparisons were performed using the least-significant difference (LSD) procedure to control type-1 error rate (SPSS, Chicago IL, USA).</p><!><p>Animals in all groups showed substantial weight gain over the course of the experiment from initial weights (day 1) until final weight determinations (day 25 or 27) (Table 1). Two-factor ANOVA conducted on bodyweights revealed a significant main effect of housing conditions (No Wheel, Locked Wheel, Free Wheel) [F(2,72)=16.64, P<0.001] but there was no effect of ethanol treatment group (1 day EW, 3 day EW, control) nor a housing by treatment interaction. Post-hoc pairwise comparisons indicated that Free Wheel animals weighed less than both No Wheel and Locked Wheel animals, which did not differ from each other.</p><p>Despite the fact that Free Wheel animals gained less weight relative to the other two housing conditions, Table 2 shows that Free Wheel animals actually consumed nearly 10% more of the ethanol-containing liquid diet than either No Wheel or Locked Wheel animals. One-way ANOVA showed a significant effect of housing conditions on ethanol diet consumption [F(2,48)=12.133, P<0.001], while post-hoc comparisons indicated that Free Wheel animals consumed significantly more than either No Wheel or Locked Wheel animals, but that No Wheel and Locked Wheel animals did not differ. Consumption for all wheel conditions was at levels of intake required to engender dependence (Devaud et al. 1995a; 1996). Blood ethanol concentrations were not assessed in these experiments due to the disruptive nature of collection and that emphasis was on assessing behavioral outcomes. Previous investigations by our group and others found that blood ethanol concentrations were highly variable throughout the course of ethanol administration by liquid diet, as the animals tend to drink in unpredictable bouts.</p><!><p>Activity levels increased gradually over the 10-day habituation period (Fig. 1). Following introduction of liquid diet, activity levels increased somewhat in animals on the control (non-ethanol) liquid diet, but significantly declined in ethanol-fed animals. Mean levels of 24-hour wheel-running activity were determined for days 6–10 of the habituation period and for all days of liquid diet exposure. Two-factor mixed ANOVA was used to compare activity levels in ethanol-fed and control diet animals under baseline and liquid-diet conditions (Fig. 2). This analysis revealed significant main effects of liquid diet [F(1,27)= 5.332, P=0.029] and ethanol feeding [F(1,27)=11.667, P<0.001], as well as a liquid-diet by ethanol-feeding interaction [F(1,27)=101.117, P<0.001]. This interaction was explored using post-hoc pairwise comparisons, which showed that ethanol-fed and control animals differed during liquid diet administration but not under baseline conditions. Further, ethanol-fed animals had reduced overall running wheel activity under liquid diet conditions. In contrast, control diet-fed animals showed increased activity with the continued access to free running wheels (Fig. 2).</p><p>Similar analyses were also conducted on dark-phase and light-phase activity levels. Not surprisingly, analysis of dark-phase activity produced results that were essentially identical to those seen for total 24-hour activity. Values were 6170±136 wheel turns for animals assigned to control diet and 5240±233 wheel turns for animals assigned to ethanol diet. Specifically, ethanol-fed animals showed significantly reduced activity (reduced by 34% to an average of 3434±182 wheel turns per phase) during liquid-diet conditions relative to their own baseline. In contrast, however, ethanol-fed animals showed significant increases in light-phase activity relative to both baseline and control-diet conditions (from 35.4±6.3 to 117.8±19.8 wheel turns per phase; a 230% increase). Thus, as seen in Fig. 2, ethanol feeding reduced overall activity levels but also attenuated normal day–night differences in activity.</p><!><p>As shown in Fig. 3, basal seizure thresholds (mg/kg PTZ) did not differ across the three running wheel conditions [F(2,22)=0.267]. No Wheel thresholds were 27.3±1.2 mg/kg PTZ, Locked Wheel thresholds were 26.4±0.8 mg/kg PTZ and Free Wheel thresholds were 27.3±0.7 mg/kg PTZ. These data showed that the running wheel condition in and of itself did not alter basal seizure susceptibility. Two-factor ANOVA revealed significant main effects of both housing conditions [F(2,65)=4.245, P=0.019] and ethanol treatment group [F(2,65)=24.145, P<0.001] on seizure thresholds, but failed to detect a significant housing by treatment interaction. Nevertheless, since post-hoc pairwise comparisons showed that seizure thresholds in Free Wheel animals differed significantly from both No Wheel and Locked Wheel animals, while No Wheel and Locked Wheel groups did not differ, we conducted an additional ANOVA in which the No Wheel and Locked Wheel groups were combined and analyzed as a single condition. This analysis showed significant main effects of housing conditions [F(1,65)=18.015, P<0.001] and ethanol treatment group [F(2,65)=23.119, P<0.001], as well as a significant housing by ethanol treatment interaction [F(2,65) =3.648, P=0.032]. Further, post-hoc pairwise comparisons indicated that both 1 day EW and 3 day EW groups differed from their respective controls in both No Wheel and Locked Wheel animals, while only 1 day EW, not 3 day EW, animals differed from controls in Free Wheel animals.</p><!><p>The intent of the present study was to determine if voluntary wheel running afforded benefit against ethanol withdrawal severity. We, and others, have shown that one indicator of the CNS hyperexcitability of EW is increased seizure susceptibility. We consistently find an average 25% decrease in the amount of chemoconvulsant required to initiate first signs of a seizure in EW animals compared to ethanol naïve animals (Devaud et al., 1995a, 1995b, 1996; Devaud and Chadda, 2001). While most human alcoholics do not present with seizures during withdrawal, the ability to quantify changes in seizure susceptibility is a validated measure, which is indicative of the increased neuronal hyperexcitability that occurs with EW. Increased sensitivity to seizure induction during EW was found in this study, consistent with our previous reports, and supports that animals were made dependent by the liquid diet paradigm. No differences in basal seizure thresholds were observed across wheel conditions. This suggests that it was the voluntary exercise, not the cage configuration, which influenced EW seizure thresholds.</p><p>The major finding of this paper is that voluntary wheel running was protective against the increased seizure susceptibility of EW. This occurred in spite of the significant reduction in voluntary running caused by intake of the ethanol-containing liquid diet. Ethanol-fed animals ran 61% less during the active/night phase compared to control-fed animals. During the day (rest phase), animals significantly reduced running compared to the active/night phase. However, ethanol liquid diet administration disrupted the normal temporal pattern of activity by eliciting a 3-fold increase in running during the rest phase compared to activity for control diet-fed animals. The observed alteration in wheel running following ethanol exposure was consistent with a previous report showing that chronic intermittent ethanol exposure reduced wheel running in C57/BL6 mice, and that this reduction persisted for several days after termination of the ethanol exposure (Logan et al., 2010).</p><p>Ethanol withdrawal is known to be stressful and common symptoms are general dysphoria, increased anxiety, seizure risk and stress reactivity. The expression of EW is believed to result from a series of neuroadaptations that occur in response to the development of ethanol dependence. Major adaptations include reductions in the responsiveness of GABAergic systems, increased responsiveness of glutamatergic systems as well as alterations in cell signaling, all which likely contribute to the hyperexcitable state of EW. There are several possible mechanisms for the observed attenuation of the increased seizure risk of EW by voluntary wheel running. The most parsimonious explanation could involve pharmacokinetic factors. The ethanol diet animals may be less intoxicated, and, therefore, less dependent if there is enhanced clearance of ethanol in the voluntary running groups. Ardies et al. (1989) reported that exercising female rats showed increased clearance of ethanol following a single, low-dose intraperitoneal injection. However, a recent report by Leasure and Nixon (2010) reported that 14 days of voluntary wheel running did not change ethanol pharmacokinetics in their binge exposure model. Blood ethanol concentrations were not determined in the present set of studies due to the disruptive nature of this procedure and the desire to focus on identifying associations between exercise and EW seizure thresholds. A separate study that will monitor the time course for ethanol clearance after administration of a bolus ethanol injection at the completion of the liquid diet paradigm is needed to directly address this question.</p><p>It is also possible that neuronal adaptations engendered by persistent exercise also contribute to its beneficial effects during ethanol diet consumption and withdrawal. EW is stressful and it has been shown that stress exacerbates risk for seizures in human epilepsy as well as in animal models (Friedman et al., 2011). Therefore, activation of the HPA axis and release of stress mediators may contribute to the increased seizure susceptibility of EW. Regular exercise has been shown to reduce the risk of stress-associated disorders by attenuating the HPA response to stressors (Nyhius et al., 2010). For example, voluntary wheel running was found to reduce the HPA axis response to audiogenic stress (Sasse et al., 2008). Cellular mechanisms contributing to these effects may include exercise-induced increases in brain-derived neurotrophic factor levels, synaptic spine density and hippocampal neurogenesis (Cotman and Berchold, 2002; Sartori et al., 2011; Stranahan et al., 2007; van Praag et al., 1999a, 1999b). Exercise was found to protect against hippocampal damage caused by stress (Snyder et al., 2009) as well as the hippocampal neurotoxicity associated with binge ethanol exposure (Leisure and Nixon, 2010). To provide further support of a direct protective effect of exercise on seizure susceptibility, a recent report showed that three weeks of voluntary wheel running reduced the severity of kainic acid-induced seizures (Reiss et al., 2009). Taken together, these studies suggest that exercise enhances the ability of the CNS to respond to stressful challenges and may contribute to protection against EW severity.</p><p>In summary, the present findings support the suggestion that regular physical activity affords protection against negative CNS responses to excessive ethanol consumption. The present observations of attenuation in seizure susceptibility during EW by chronic voluntary exercise could result from improved clearance of consumed ethanol and/or the enhanced neuronal plasticity and ability to adapt to challenges. Continued exploration of this question may help inform clinicians in their management of alcoholic patients.</p>

PubMed Author Manuscript

Radiation persistently promoted oxidative stress, activated mTOR via PI3K/Akt, and downregulated autophagy pathway in mouse intestine

While acute effects of toxic radiation doses on intestine are well established, we are yet to acquire a complete spectrum of sub-lethal radiation-induced chronic intestinal perturbations at the molecular level. We investigated persistent effects of a radiation dose (2 Gy) commonly used as a daily fraction in radiotherapy on oxidants and anti-oxidants, and autophagy pathways, which are interlinked processes affecting intestinal homeostasis. Six to eight weeks old C57BL/6J mice (n=10) were exposed to 2 Gy \xce\xb3-ray. Mice were euthanized two or twelve months after radiation, intestine surgically removed, and flushed using sterile PBS. Parts of the intestine from jejunal-ilial region were fixed, frozen, or used for intestinal epithelial cell (IEC) isolation. While oxidant levels and mitochondrial status were assessed in isolated IEC, autophagy and oxidative stress related signaling pathways were probed in frozen and fixed samples using PCR-based expression arrays and immunoprobing. Radiation exposure caused significant alterations in the expression level of 26 autophagy and 17 oxidative stress related genes. Immunoblot results showed decreased Beclin1 and LC3-II and increased p62, PI3K/Akt, and mTOR. Flow cytometry data showed increased oxidant production and compromised mitochondrial integrity in irradiated samples. Immunoprobing of intestinal sections showed increased 8-oxo-dG and nuclear PCNA, and decreased autophagosome marker LC3-II in IEC after irradiation. We show that sub-lethal radiation could persistently downregulate anti-oxidants and autophagy signaling, and upregulate oxidant production and proliferative signaling. Radiation-induced promotion of oxidative stress and downregulation of autophagy could work in tandem to alter intestinal functions and have implications for post-radiation chronic gastrointestinal diseases.

radiation_persistently_promoted_oxidative_stress,_activated_mtor_via_pi3k/akt,_and_downregulated_aut

1. Introduction<!>2.1. Mice<!>2.2 Irradiation<!>2.3 Tissue harvesting and intestinal epithelial cell (IEC) isolation<!>2.4. RNA isolation and quantitative real time PCR (qRT-PCR) using PCR array<!>2.5. Intracellular ROS measurements<!>2.6. Mitochondrial superoxide assay<!>2.7. Mitochondrial membrane potential (MMP) analysis<!>2.8. Immunohistochemistry in intestinal sections<!>2.9. Immunoblot analysis<!>2.10. Data analysis and statistics<!>3.1. Increased oxidant production and compromised mitochondrial status accompanied decreased anti-oxidant gene expression after radiation exposure<!>3.2. Ionizing radiation downregulated autophagy-associated gene expression in intestine<!>3.3. Radiation exposure decreased autophagy pathway specific proteins, and activated pro-growth pathways<!>3.4. Radiation caused increased oxidative DNA damage and cell proliferation in intestine<!>4. Discussion

<p>Autophagy is a complex catabolic process involved in removing and recycling damaged or unwanted cellular constituents in autophagolysosomal vesicles to sustain energy supply during nutritional stress. The mammalian target of rapamycin (mTOR), a member of the phosphatidylinositol kinase-related kinase (PIKK) family and a nutrient sensor, is a major regulator of autophagy [1]. Formation of the autophagosome, and its subsequent fusion with lysosomal vesicle to form the autophagolysosome, is essential for autophagy and at least thirty autophagy-related (Atg) genes have been identified and demonstrated to be involved in different sub-types of autophagy [2,3]. However, the core autophagy pathway requires seventeen genes and in mammalian cells the process is initiated when the mammalian homolog of yeast Atg1 complex comprising ULK1, ULK2, Atg13, Atg101 and FIP200 is activated due to mTOR inhibition [1-3]. Major proteins involved in the elongation of the autophagosome membrane are Vps34/PI3PIII complex, Vps15, and Beclin1. Subsequently, the Atg12-Atg5-Atg16L protein complex joins the process and usher in maturation of the autophagosome [1-4]. Elongation and maturation of autophagosomes also involve mammalian orthologs of yeast Atg8 such as LC3, Gabarapl1 and Gabarapl2. Conversely, Atg7, Atg4, and Atg3 are involved in activation of these proteins to generate their mature forms. For LC3, it is initially converted to LC3-I or LC3A and finally to LC3-II or LC3B and similar to LC3, the GABARAP members are finally converted to their active forms of GABARAPL1-II and GABARAPL2-II. The activated LC3 and GABARAP family of proteins on the autophagosome allow binding of adaptor proteins such as p62 and Nbr-1, which in turn recruits ubiquitinated proteins for autophagolysosomal degradation [1-4]. Importantly, the autophagy pathway through its influence on cell death and proliferation, oxidative stress and inflammation, and on immune responses is also involved in the maintenance of cellular homeostasis in intestine, and dysregulation of autophagy has been linked to altered intestinal functionality and development of diseases such as inflammatory bowel diseases and cancer [1-5].</p><p>Maintenance of gastrointestinal (GI) homeostasis is essential for health and radiation exposures have been shown to affect cellular functionality in the intestine, a rapidly proliferating tissue. Radiation effects on intestine are dependent on radiation dose, and duration of exposure. Radiation exposure to intestine has been reported to cause a myriad of short- and long-term ailments due to perturbation of intestinal cell functions [6-8]. While higher doses of radiation exposure invariably leads to cell death-associated normal tissue complications, lower sub-lethal doses of radiation exposure cause most of the alterations at the molecular level and cells survive with damage leading to long-term heath risks [9]. Furthermore, epidemiological studies in atomic bomb survivors and nuclear workers, most of whom were exposed to sub-lethal doses of <1 Gy, have demonstrated that radiation is a long-term risk factor for non-cancer diseases as well as solid cancers [10-13]. Furthermore, radiation exposure has been intimately linked to increased reactive oxygen species (ROS) production and persistent oxidative stress in cells, and oxidative stress has been reported to activate the pro-growth PI3K/Akt pathway, which in turn is known to activate mTOR [14,15]. On the other hand oxidative stress and autophagy are in a co-dependent but inverse relationship where decreased autophagic activity increases oxidative stress and increased oxidative stress in turn is known to downregulate the autophagy pathway [16-18]. Notably, most of the radiation exposure studies on autophagy are in relation to cancer therapy, and effects of a sub-lethal radiation dose on autophagy has not been explored [19-24]. Here we demonstrated that radiation exposure led to persistently increased oxidant production and decreased anti-oxidant gene expression leading to oxidative stress and activation of proliferative PI3K/Akt and mTOR signaling. When considered along with our results showing radiation-induced downregulation of autophagy pathway associated factors at the mRNA as well as at the protein level, our observations have implications for radiation-induced persistent alterations of intestinal functions.</p><!><p>Six to eight weeks old female C57BL/6J mice were purchased from Jackson Laboratories (Bar Harbor, ME) and housed at Georgetown University's research animal facility, which is an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC) accredited facility. Mice were housed in autoclaved cages and bedding materials in a separate room with 12 h dark and light cycle maintained at 22 °C in 50% humidity. All animals were provided certified rodent diet with filtered water ad libitum and CO2 asphyxiation was used for euthanasia as per animal care facility guideline. Post-irradiation, any mouse showing signs of declining health assessed using approved criteria such as hunched posture, ruffled fur, diarrhea, and reduced activity was euthanized by CO2 asphyxiation and was excluded from the specific study group. All animal procedures used in the study were performed in accordance with a protocol approved by the Institutional Animal Care and Use Committee and we followed Guide for the Care and Use of Laboratory Animals by the Institute of Laboratory Animal Resources, National Research Council, and U.S. National Academy of Sciences for our research.</p><!><p>We used a 137Cs source to expose the mice (n = 10 per group) to a sub-lethal whole-body γ radiation dose of 2 Gy (dose rate 0.7 Gy/min). For this study, the radiation dose of 2 Gy was chosen due to the fact that it is a common daily fraction used in radiotherapy to deliver planned total radiation dose to cancer patients. Additionally, we have used the 2 Gy dose to correlate the current and earlier [25] radiation related persistent effect studies in intestinal cells to radiation-induced colorectal carcinogenesis studies in mouse models [26]. Post-irradiation mice were returned to their home cages and monitored regularly. Irradiation experiments were repeated three times and mice were euthanized two months after each exposure. In all experiments the control mice were sham-irradiated.</p><!><p>Mice were euthanized as per approved protocol and small intestine was surgically removed two or twelve months after radiation exposure. For our previous radiation-induced persistent effect studies in different tissues [25,27-29] we choose two and twelve months time points and the same time points were used in the current study. Intestinal lumen was flushed with phosphate buffered saline (PBS) at room temperature, and the jejunal-ileal region was used for intestinal epithelial cell (IEC) isolation, tissue fixation, or freezing. IEC was isolated, characterized, and viability assessed according to a protocol described previously [25]. Isolated cells were used for flow cytometry experiments and samples from six mice in each group were processed in triplicate for each experiment. For immunohistochemistry and immunoflurescence, 3 cm of intestinal tissue was fixed in 10% buffered formalin, paraffin embedded, and 4 µm sections were made. Tissues were also snap-frozen in liquid nitrogen and stored at -80 °C either for RNA isolation or for immunoblot analysis.</p><!><p>For RNA extraction, we used pooled samples of five mice from each experimental group. Total RNA was isolated with Trizol reagent (Invitrogen, Carlsbad, CA) and RNeasy column according to the manufacturer's instructions (Qiagen, Germantown, MD) from flash frozen intestine. RNA concentration and quality was determined using a Bioanalyzer (Agilent Technologies, Palo Alto, CA) and stored in aliquots at -80 °C for further analysis. Mouse autophagy (PAMM-084Z) and oxidative stress (PAMM-065Z) RT2 Profiler PCR Array was obtained from SA Biosciences (Frederick, MD) and intestinal RNA samples were assayed by qRT-PCR according to manufacturer's instructions. Briefly, RT2 First Strand Kit (SA Biosciences) was used to reverse-transcribe RNA into cDNA and RT2 real time SYBR green PCR master mix was used for qRT-PCR in an iCycler (Bio-Rad, Hercules, CA) following a protocol provided by the manufacturer (SA Biosciences). The PCR array probes expression of 84 genes in each of the autophagy and oxidative stress arrays and relative changes in gene expression were calculated using β-actin as an endogenous control following the comparative Ct (AACt) method using web-based tools provided by the manufacturer (SA Biosciences). Results were expressed relative to sham-irradiated control samples.</p><!><p>Intracellular ROS level was measured in IEC as per protocol described previously using the fluorescent probe 2'-7'-dichlorodihydrofluorescein diacetate (H2DCFDA, Invitrogen, Carlsbad, CA) [25,30].</p><!><p>Mitochondrial superoxide level was measured in IEC using a mitochondrial superoxide specific fluorescent probe, MitoSOX red (Invitrogen), as per manufacturer's instructions and described previously [25].</p><!><p>Changes in MMP were measured in IEC using Rhodamine 123 as per protocol described previously [25,31].</p><!><p>Staining for 8-oxo-dG was performed using a primary antibody from Trevigen (Gaithersburg, MD; dilution-1:200) according to a protocol described previously [25]. For proliferating cell nuclear antigen (PCNA) staining, sections were depraffinized, antigen retrieved in pH 6.0 citrate buffer (Dako, Carpinteria, CA), and endogenous peroxidase quenched. Sections were incubated in blocking buffer (0.1% bovine serum albumin in PBS) before exposing to PCNA antibody (sc7907; Santa Cruz Biotechnology, Dallas, TX; dilution-1:100). SuperPictureTM 3rd Gen IHC detection kit (87-9673; Invitrogen) was used for signal detection and color development. All the IHC slides were mounted and visualized under a bright field microscope and images were captured at microscopic magnification (8-oxo-dG at 40X and PCNA at 20X microscopic magnification). For immunofluorescence of LC3-II, following the blocking step described for PCNA, tissue sections were incubated overnight with anti-LC3-II antibody (PA1-33197, Thermo Scientific, Pittsburgh, PA; dilution 1:100) at 4 °C. After necessary washing steps sections were incubated with AlexaFluor488-conjugated goat-anti-rabbit antibody (A-1 1034, Life Technologies, Grand Island, NY) for 1 h in dark at room temperature. Samples were washed and mounted using DAPI containing VECTASHIELD mounting medium (H-1200, Vector Laboratories, Burlingame, CA). Sections were visualized and images captured at 20X magnification using an Olympus BX61 DSU fluorescent microscope and images were analyzed using SlideBook v5.0 software (Intelligent Imaging Innovations, Inc, Denver, CO). To determine specificity of the staining, appropriate controls were run in parallel with the experimental sections.</p><!><p>Intestinal tissues from five mice were pooled and subjected to immunoblot analysis according to a protocol described previously [28] and repeated in three sets of experimental samples. Briefly, tissues were homogenized in ice-cold extraction buffer (0.5% sodium deoxycholate, 0.5% NP-40, 10 mM EDTA in phosphate-buffered saline containing protease inhibitor cocktail obtained from Sigma-Aldrich, St. Louis, MO). Supernatants from the homogenates were collected by centrifugation, proteins were resolved by SDS-PAGE, transferred onto polyvinylidene fluoride membrane, and incubated with appropriate primary antibody for LC3-II (PA1-33197, Thermo Scientific, Pittsburgh, PA; dilution-1:1000), Beclin 1 (sc11427, Santa Cruz Biotechnology; dilution-1:400), p62 (5114s, Cell Signaling Technology, Danvers, MA; dilution-1:500), p85 (4292s, Cell Signaling Technology; dilution-1:500), mTOR (PA5-17780, Thermo Scientific; dilution-1:1000), phospho-mTOR (2971s, Cell Signaling Technology; dilution-1:500), Akt (sc8312; Santa Cruz Biotechnology; dilution-1:400,), and phospho-Akt (9277s; Cell Signaling Technology; dilution-1:500,), β-actin (sc4778, Santa Cruz Biotechnology; dilution-1:2500). Immunoblot membranes were developed with horseraddish peroxidase conjugated secondary antibody and enhanced chemiluminescence detection system. Images were captured on photographic films and scanned, and representative results are displayed. Densitometric quantification of the immunoblots was performed by normalizing to β-actin band intensity using ImageJ v1.46r (National Institutes of Health, Bethesda, MD).</p><!><p>Immunohistochemistry images were analyzed using color deconvolution and/or Image-based Tool for Counting Nuclei (ITCN) plug-ins of ImageJ v1.46r software by two observers blinded to treatment groups as per protocol described earlier [25,32,33]. We used five random image frames from each section for analysis and mean data from six mice in each group are presented graphically and a representative image from one animal of each group is shown in the results. We used WinMDI v2.9 to analyze flow cytometry data and average percent change of mean fluorescence from triplicate samples of six mice are presented in bar graphs. A representative histogram comparing sham-irradiated (black) to irradiated (red) samples is shown in the results. Two independent viewer blinded to the experimental groups counted LC3-II staining by observation in 24 random image frames (4 frames per section; n= 6 per group) captured at 20X magnification and results are expressed as average number of positive cells per 20X field and a representative image (20X magnification) from one animal of each group of one experiment is shown in the results. Statistical significance between the two groups was determined using a two-tailed paired student's t test and p<0.05 was taken as statistically significant. Error bars represent ± standard error of the mean (SEM).</p><!><p>Earlier we reported that 2 Gy irradiation of mice (same strain, gender, and age) used in this study was associated with oxidative stress in intestinal cells even twelve months after exposure [25]. In the current study, significantly increased level of ROS was observed in intestinal epithelial cells two months after exposure to 2 Gy γ radiation relative to controls (Figure 1A and B; p<0.007). Concurrently, we also observed increased mitochondrial superoxide (Figure 1C and D; p<0.001) and decreased mitochondrial membrane potential (Figure 1E and F; p<0.005) two months after radiation exposure relative to controls. Quantitative PCR analysis of oxidative stress related gene expression showed significant (determined using a fold change cutoff of 1.25 and a p-value of <0.05) alterations of 17 genes (Table 1). While expression of only one anti-oxidant gene (Gpx2) was upregulated, expression of a total of 14 anti-oxidant genes were downregulated (Gpx1, Gpx3, Gpx8, Prdx1, Prdx2, Prdx3, Prdx4, Prdx6, Sod1, Sod2, Sod3, Txnrd3, Cat, and Gstk1) in intestine two months after radiation exposure. In contrast, expression of two oxidant production related genes was perturbed. While expression of Nox4 was downregulated, expression of Nos2 was significantly increased in irradiated samples (Table 1).</p><!><p>PCR array results showed significant (determined using a fold change cutoff of 1.25 and a p-value of <0.05) perturbation of 10 autophagy pathway specific (Figure 2A) as well as 16 autophagy regulatory genes (Figure 2B) in two-month post-irradiation samples. Among the autophagy specific genes, Gabarapl1, Gabarapl2, Map1lc3a, and Map1lc3b involved in elongation and closer of autophagosomes were significantly downregulated. On the contrary, expression of six autophagy pathway specific genes (Atg16l1, Atg16l2, Atg7, Dapk1, Atg10, and Ctsb) were significantly upregulated. Additionally, we also observed downregulation of five (Eif2ak3, Pik3r4, Irgm, Snca, and Fas) and upregulation of eleven (Cdkn2a, Fam176a, Hgs, Hsp90aal, Hspa8, Ifna2, Ifna4, Ifng, Ins2, Esrl, and Tgm2) genes, which are known to influence the autophagy pathway.</p><!><p>In immunoblot analysis of two-month post-irradiation intestinal samples, levels of autophagosome forming LC3-II (p<0.001) and Beclinl (p<0.003) were decreased (Figure 3A and B). However, p62, an autophagy scaffolding protein was significantly (p<0.004) increased in irradiated samples (Figure 3A and B). Additionally, we also observed significantly increased p85a (p<0.0005), a regulatory subunit of PI3K, in the irradiated samples (Figure 3A and B). Levels of Akt (p<0.01), phospho-Akt (p<0.004) and its downstream target mTOR (p<0.002) and its active form phospho-mTOR (p<0.0001) were significantly increased in intestine two months after radiation exposure (Figure 3A and B). Significantly decreased LC3-II (p<0.0004), and increased p62 (p<0.0003), mTOR (p<0.01), and phospho-mTOR (p<0.00003) was also observed in twelve-month post-irradiation samples (Figure 4A and B). We immunostained intestinal sections for LC3-II, a protein known to be associated with autophagosome membrane, and observed decreased staining in both the two- (Figure 4C and D) and twelve-month (Figure 4E and F) post-irradiation samples relative to sham-irradiated controls.</p><!><p>Cellular consequence of increased oxidative stress was assessed using 8-oxo-dG staining. We show that two months after irradiation there was significantly greater 8-oxo-dG staining in the intestinal crypts relative to sham-irradiated control (Figure 5A and B; p<0.01). Furthermore, we also observed significantly increased number of intestinal epithelial cells stained positive for PCNA in irradiated samples relative to controls (Figure 5C and D; p<0.04).</p><!><p>Exposure to ionizing radiation, still a major modality for cancer treatment, has been demonstrated to have long-term side effects in exposed normal tissues including intestine. Effective use of radiation is playing its role in increasing cancer therapy success rate and consequently, there is increased number of cancer survivors who could potentially develop chronic enteropathies or even a second cancer in the colorectal region especially after pelvic irradiation [34]. Although the dose (2 Gy) used in the current study is one fraction of a total radiation dose delivered in planned fractionated radiotherapies, it has allowed us to discern underlying molecular events that we believe could play a role in the long-term sequela of radiation exposures in normal intestine and is not reported previously. We show that radiation exposure caused decreased expression of autophagy and anti-oxidant genes. Additionally, increased intracellular ROS production was associated with decreased levels of key autophagy proteins and activation of proliferative pathways.</p><p>Autophagy, which is a self-cannibalistic process and provides energy during stress, is widely studied for its roles in normal development and homeostasis as well as in pathological states such as inflammatory bowel disease, neurodegeneration, and cancer [3]. However, autophagy in relation to radiation has mostly been studied for radiosensitization of cancer cells [19-24]. Notably, these studies, which commonly involved short-term cell culture system dissecting molecular pathways involved in radiation-induced autophagy activation and cancer cell survival, aimed to develop strategies to inhibit autophagy and thus sensitize cancer cell to therapy [19-24,35]. While autophagy activation provides an alternate energy source to cell for survival and may contribute to promotion and progression of established cancers, its downregulation in normal cells has been implicated in functional alterations and cancer initiation [24]. The autophagosome is formed de novo and is a complex process involving multiple steps of vesicle induction, nucleation, elongation, and finally closure [2]. The Map1lc3a and Map1lc3b along with Map1lc3c are the three isoforms of Map1lc3 (microtubule-associated protein light chain 3), which is the yeast Atg8 ortholog. The Map1lc3 or commonly termed LC3 undergoes extensive post-translational modification including protease cleavage by Atg4 enzyme leading to formation of LC3-I and then phospholipidated by Atg7 and Atg3 to LC3-II, which is required for elongation of autophagosome membrane. While LC3-I is cytosolic, LC3-II is associated with autophagosome membrane and its level denotes number of autophagosome in cell [36]. Furthermore, the Gabarapl1 and Gabarapl2, essential for autophagosome vesicles closure, are also activated by the Atg4 to Gabarapl1-I and Gabarapl2-I and phospholipidated by Atg7 and Atg3 to form autophagosome closure elements Gabarapl1-II and Gabarapl2-II. Radiation exposure associated decreased expression of four important autophagy specific genes (Gabarapl1, Gabarapl2, Map1lc3a, and Map1lc3b) we believe, will reduce overall autophagy activity in intestinal cells. Decreased expressions of Map1lc3 were further supported by decreased LC3-II in immunoblots demonstrating that decreased levels of active form, in spite of increased/unaltered expression of activating proteins, was due to transcriptional downregulation and reduced availability of LC3. Importantly, Beclin1 as well as Pik3r4, which are components of the class III PI3K complex and are involved in the nucleation phase of the autophagy [37], are significantly decreased after radiation exposure. Although expression of a number of genes (Dapk1, Cdkn2a, Fam176a, Hgs, Hsp90aa1, Hspa8, Ifna2, Ifna4, and Ifng) implicated in triggering autophagy in response to different stimuli [38-42] were upregulated, our results show that exposure to a sub-lethal dose of radiation led to long-term downregulation of the core components involved in the three critical autophagy steps of nucleation, elongation, and closure of autophagosomes. Our results further show that radiation-induced perturbation of the autophagy specific factors could act in tandem with downregulated autophagy regulating factors such as Eif2ak3, Irgm1, Snca, and Fas to produce an inhibitory effect on autophagic processes. Indeed, accumulation of p62, which is an adaptor protein facilitating recruitment of ubiquitinated proteins to autophagosome via its interaction with LC3 and Gabarap family of proteins and is degraded during autophagic process [43], further argues in favor of autophagy downregulation after radiation exposure. Considering that Eif2ak3 and Irgm1 respectively are required for endoplasmic reticulum (ER) [44], and bacterial pathogen [45] mediated stress-induced autophagy activation, decreased expression of these factors will cause a faulty autophagic response to ER stress and immune challenges and increased vulnerability of intestinal cells to functional deregulation and pathologic infection. On the contrary, increased level of Snca [46] as well as ligand-mediated activation of Fas [47] is known to activate autophagy and downregulation of both the factors observed in our study further supports long-term downregulatory effects of sub-lethal radiation exposure on autophagy. Although we noted upregulation of a number of genes (Ctsb, Atg10, Ins2, Esr1, and Tgm2) linked to autophagy, they play mostly regulatory roles in autophagosome formation. On the contrary, upregulation of these genes have been reported to promote inflammation, oxidative stress, insulin resistance, and proliferation with adverse cellular consequences [48-53]. Taken together we have shown that although the activators/regulators such as Atg7, Atg10, Atg16l1, and Atg16l2 of key autophagy molecules are upregulated, the key molecules themselves involved in initiation (Beclin 1), elongation (LC3-II, and closure (Gabarapl1 and Gabarapl2) of the autophagosome membrane were downregulated suggesting potential implications for autophagy inhibition. When combined with immunofluorescence evidence of decreased LC3-II widely used as a marker for autophagosome number, our results strongly suggest that exposure to a sub-lethal radiation dose persistently downregulated autophagy with implications for chronic cellular functional deregulation and transformation.</p><p>Oxidative stress, which could be due to increased oxidant production, decreased antioxidant activity, or a combination of both, has been linked to radiation as well as autophagy. Radiation is known to impart its damaging effects directly by depositing energy into the traversed biomolecules or indirectly by oxidant production especially ROS. While radiation at higher doses causes cell death in part by producing lethal levels of ROS, at lower doses it causes sub-lethal oxidative stress, DNA damage, and activation of proliferative signaling pathway thus allowing damaged cells to proliferate with increasing transformation potential [25]. Persistent oxidative stress after radiation exposure observed in the current study is consistent with others as well as our earlier studies [25,54] and was due to increased ROS production as a results of compromised mitochondrial function, as well as due to transcriptional downregulation of important antioxidant genes including a number of Gpx and Prdx isoforms. Considering that histological examination of intestinal sections did not show differences in inflammatory cell number between control and irradiated samples and isolated epithelial cells were typified (data not shown) [55,56], our results are suggestive of epithelial cell dysregulation and oxidative stress and is in agreement with our earlier twelve month post-radiation results [25]. Importantly, there was increased Nos2 (inducible nitric oxide synthase or iNos) expression, which could in combination with ROS promote reactive nitrogen species (RNS) generation and more oxidative stress. Apart from its effects on furthering oxidative stress, increased expression of Nos2 via increased NO production has also been reported to adversely affect autophagosome formation [57]. Additionally, oxidative stress is known to downregulate autophagy and reduced autophagy is in turn known to promote cellular stress not only due to increased accumulation of damaged and aggregated proteins but also due to decreased removal of damaged mitochondria (mitophagy) [36]. Taken together, our results of decreased autophagy and increased oxidative stress lead us to propose that radiation exposure is propelling cells into a cycle of persistent dysregulated homeostasis (Figure 6). Our belief of radiation-induced cyclical deregulation is further strengthened by the fact that there is increased PI3K/Akt signaling invariably linked to increased ROS production as well as Akt-mediated activation of mTOR, which is known to downregulate autophagy (Figure 6).</p><p>The prevailing linear no-threshold (LNT) model and the seventh Biological Effects of Ionizing Radiation (BEIR VII) report proposed that no radiation dose is safe and that irrespective of dose, radiation exposure has short as well as long-term cellular consequences [58], and fundamental understanding of molecular events will support developing preventive and therapeutic strategies against adverse long-term sequel of radiation exposure. While the current study focused on long-term effects of radiation on small intestine and investigating persistent radiation effects on colonic and other tissues remains a future goal, we believe that our results have relevance to radiation associated chronic GI tract ailments. To our knowledge this is the first report on long-term effects of a sub-lethal dose of radiation on oxidative stress in relation to autophagy, which is considered a major mechanism involved in maintaining intestinal homeostasis. The current study showed that radiation exposure is initiating a chronic chain of events starting with oxidative stress leading to PI3K/Akt and mTOR activation, downregulation of autophagy pathway, and further promotion of oxidative stress and has implications for chronic intestinal pathologies such as inflammatory bowel diseases, chronic enteritis, and cancer (Figure 6).</p>

PubMed Author Manuscript

Biosynthesis of the Thiopeptins and Identification of an F420H2-Dependent Dehydropiperidine Reductase

Thiopeptins are highly decorated thiopeptide antibiotics similar in structure to thiostrepton A and harbor two unusual features. All thiopeptins contain a thioamide, a rare moiety among natural products, and a subset of thiopeptins present with a piperidine in the core macrocycle rather than the more oxidated dehydropiperidine or pyridine rings typically observed in the thiopeptides. Here, we report the identification of the thiopeptin biosynthetic gene (tpn) cluster in Streptomyces tateyamensis and the gene product, TpnL, which shows sequence similarity to (deaza)flavin-dependent oxidoreductases. Heterologous expression of TpnL in the thiostrepton A producer Streptomyces laurentii led to the production of a piperidine-containing analogue. Binding studies revealed that TpnL preferentially binds the deazaflavin cofactor coenzyme F420, and in vitro reconstitution of TpnL activity confirmed that this enzyme is an F420H2-dependent dehydropiperidine reductase. The identification of TpnL and its activity establishes the basis for the piperidine-containing series a thiopeptides, one of the five main structural groups of this diverse family of antibiotics.

biosynthesis_of_the_thiopeptins_and_identification_of_an_f420h2-dependent_dehydropiperidine_reductas

INTRODUCTION<!>Identification of the Thiopeptin Biosynthetic Gene (tpn) Cluster.<!>Heterologous Expression of TpnL in S. laurentii.<!>TpnL Is an F420H2-Dependent Dehydropiperidine Reductase.<!>DISCUSSION<!>CONCLUSION

<p>Thiopeptides are ribosomally synthesized and post-translationally modified peptides with extensive chemical modifications that impart potent antiinfective properties, particularly against drug-resistant strains of Gram-positive pathogens.1 Despite the impressive antibacterial activity of thiopeptides, their clinical use has been stymied in large part by their low aqueous solubility. There is therefore interest in the discovery, design, and production of thiopeptide analogues with improved pharmacokinetic parameters.2,3 For example, semisynthetic derivatization of the thiopeptide GE2270A led to the development of an analogue that retained antibacterial efficacy and was improved in aqueous solubility.2 The elucidation of biosynthetic pathways of thiopeptide metabolites, especially those that include enzymes with additional biochemical activities and associated modifications, is essential to expand the biosynthetic toolkit available to produce derivatives that can combat the ever-growing threat of antibiotic-resistant bacterial pathogens.</p><p>To date, over 100 distinct thiopeptides have been identified and are classified as structural series a–e on the basis of the oxidation state of the central, six-membered nitrogenous ring in the core macrocycle (Figure 1A).1 Bearing a strikingly similar structure to the prototypical thiopeptide thiostrepton A (Figure 1B, 1), the thiopeptins (Figure 1B, 2–5) are a complex of thiopeptide antibiotics produced by Streptomyces tateyamensis ATCC 21389 (S. tateyamensis).4 Both 1 and 2–5 are 17 amino acids long, differing in sequence only in the first residue, and both include a second quinaldic acid-containing macro-cycle. The thiopeptins, however, are distinguished by a thioamide moiety and either a piperidine (series a) or dehydropiperidine (series b) ring in the core macrocycle.5 Thioamides are a rare feature in natural products, with only a handful of known examples, and only a few members of series a thiopeptides have been identified to date, which include 2, 3, and Sch 18640 (Figure 1B, 6).4–8 In addition to piperidine and dehydropiperidine rings, thiopeptides can also possess either a dihydroimidazopiperidine (Figure 1A, series c), pyridine (Figure 1A, series d), or hydroxypyridine (Figure 1A, series e).1 Recent studies established that the series d pyridine and core macrocycle are installed by a dedicated synthase that condenses two serine-derived dehydroalanine (Dha) residues of a linear precursor in a [4 + 2] cycloaddition, followed by the dehydration and elimination of the leader peptide to generate the aromatic nitrogenous heterocycle.9,10 The core macrocycle of series a and b thiopeptides also originates from the condensation of two Dha residues, although it is not yet understood how this macrocyclization process leads to retention of the N-terminal portion of the peptide and a more reduced piperidine or dehydropiperine ring.11 In this work, we identify the S. tateyamensis thiopeptin biosynthetic gene (tpn) cluster and establish the biosynthetic basis of a piperidine-containing series a thiopeptide.</p><!><p>To pinpoint the genes responsible for 2–5 biosyntheses, the S. tateyamensis genome was partially sequenced to 6.92 Mb on 2,838 contiguous fragments and scanned for a nucleotide sequence that could encode the core peptide (VASASCTTCICTCSCSS). The gene encoding the thiopeptin precursor peptide, tpnA, was identified, and the sequence of the tpn locus was fully assembled by sequencing the fosmid clones from an S. tateyamensis genomic library that were identified in a PCR screen for tpn genes. The tpn cluster encompasses 23 open reading frames very similar in composition and organization to the thiostrepton biosynthetic gene (tsr) cluster (Figure 1C and Table S1). On the basis of sequence similarity to proteins encoded in the tsr and other thiopeptide biosynthetic gene clusters (Table S1), TpnA, together with TpnBCDEFG, comprise the minimum set of proteins expected to be required to assemble the core thiopeptide scaffold. The thiazol(in)es are likely installed by TpnEFG. TpnF and TpnG are similar to leader peptide binding proteins and YcaO domain-containing cyclodehydratases, respectively, which operate in concert to form thiazoline rings from the amide backbone and Cys residues of the thiopeptide.10 The dehydrogenase TpnE is expected to oxidize the intermediate thiazoline to the thiazole.10,12 A split LanB-type dehydratase appears to be encoded by tpnBC and likely catalyzes the dehydrations to generate the Dha and dehydrobutyrine residues.10 Two Dha residues are likely coupled together by a putative [4 + 2] cyclase, TpnD, to form a dehydropiperidine and the thiopeptin core macrocycle.</p><p>The quinaldic acid moiety of 1, and likely of 2–5, is derived from L-tryptophan. TpnOPUVW are expected to convert the amino acid into 4-(1-hydroxyethyl)quinoline-2-carboxylic acid (HEQ) by a process that includes an oxidative ring expansion (Scheme S1).13–15 TpnO is similar to TsrM, a radical SAM and cobalamin-dependent methyltransferase that generates 2-methyl-L-tryptophan as the first step in this process (Scheme S1).16 The second step is catalyzed by an aminotransferase (Schemes S1 and S2).14 TpnW shares homology with histidinol-phosphate transaminases, demonstrating 40% identity (50% similarity) with the thiostrepton pathway transaminase TsrV.14 To determine if TpnW could be involved in 2–5 biosynthesis, we reconstituted TpnW activity in vitro. TpnW does convert 2-methyl-L-tryptophan to 3-(2-methylindolyl)pyruvate in the presence of α-keto acid acceptors 3-indolylpyruvate, p-hydroxyphenylpyruvate, and phenylpyruvate, but not in the presence of pyruvate, oxaloacetate, or α-ketoglutarate (Figures S1–S3). Similar preferences for α-keto acid acceptors were observed with TsrV.14,15 These results indicate that TpnW is a 2-methyl-L-tryptophan aminotransferase involved in HEQ biosynthesis for 2–5.</p><p>The quinaldic acid-containing macrocycle is likely installed in the thiopeptin scaffold by the actions of TpnHIQ.17–19 TpnI is similar to adenylyltransferases and is expected to attach HEQ to Thr12 of the peptide, likely followed by TpnH-catalyzed epoxygenation.14,17,19 TpnQ, similar to the macro-cyclase TsrB, is a candidate to mediate macrocyclization of the quinaldic acid containing loop through attack of the core peptide's N-terminus upon the epoxide.18 The thiopeptins from S. tateyamensis possess either a methyl ester (2 and 4) or a carboxylic acid (3 and 5) at their C-termini; none containing the C-terminal amide observed in 1 have been reported. Accordingly, a homologue encoding the thiostrepton amidotransferase TsrT is absent.14 TpnT is similar to TsrP and may function as a carboxymethyltransferase to produce the C-terminal methyl ester.14,20 A homologue of the methylesterase TsrU is not present in the tpn cluster.20 Instead, TpnR, with similarity to other α/β-hydrolases, likely serves this role.</p><p>Three genes of the tpn cluster, tpnLMN, do not share homology with known thiopeptide biosynthetic genes and are therefore candidates for the thioamide and piperidine features unique to 2–5. TpnMN show sequence similarity to YcaO proteins and TfuA-like proteins, respectively, and are similar to the TvnHI pair encoded in the thioviridamide gene cluster, another thioamidated metabolite.7 The involvement of a YcaO–TfuA pair in the post-translational thioamidation of methyl-coenzyme A reductase in Methanosarcina acetivorans was recently established.21,22 The YcaO protein activates the peptide backbone in an ATP-dependent manner.22 Although the function of the TfuA protein is unclear, the YcaO and TfuA proteins are both required for the thioamidation of the peptide substrate.22 It is therefore likely that TpnM (YcaO protein) and TpnN (TfuA protein) work together to install the thiopeptin thioamide.</p><p>TpnL, on the other hand, shares homology to proteins annotated as FMN-dependent pyridoxine/pyridoxamine 5′-phosphate oxidases and other enzymes of the flavin/deazaflavin oxidoreductase subgroup B (FDOR-B).23 The only remaining thiopeptin modification unaccounted for is a reduction to produce the piperidine ring of the series a thiopeptins 2 and 3, and an oxidoreductase is a candidate enzyme for that step.</p><!><p>The structures of 1 and 2–5 are closely related, and we reasoned that TpnL may be able to intercept the appropriate thiostrepton biosynthetic intermediate as a substrate, generating a new metabolite and shedding light on the function of this protein. To this end, TpnL was heterologously expressed in Streptomyces laurentii ATCC 31255 (S. laurentii), a 1 producer.14 S. laurentii HI1 contains the integrative plasmid pSET1520-TpnL, expressing TpnL under control of the constitutive ermE* promoter, and S. laurentii HI0 houses an empty pSET1520 vector. S. laurentii HI1 still produced 1 (tR =23.8 min and [M + 2H]2+ m/z 833) but also produced the new thiopeptide metabolite 7 (tR = 23.7 min and [M + 2H]2+ m/z 834) with a mass increased by 2 Da relative to 1 (Figure 2A and Figure S4). 7 was isolated, and a molecular formula of C72H88N19O18S5 was determined by HR-ESI-MS (observed m/z 1666.5157 [M + H]+, calculated m/z 1666.5158). Additional MS/MS and NMR analyses indicated that 7, thiostrepton Aa, is a piperidine-containing analogue of 1 (Scheme 1 and Figures S5–S13 and Tables S2–S3). 1H−1H ROESY correlations of 7 between Pip-6-H and Pip-4-HB and between Pip-4-HB and Pip-2-H suggest that these atoms are all in axial positions of the sixmembered piperidine ring (Figures S7, S14, and S15 and Tables S3 and S4). On the basis of the previously determined structure of 1, we propose that the newly introduced stereocenter in 7 presents the same absolute configuration as is observed in the other series a thiopeptide metabolites 2 and 3.5,24 The piperidine ring in 7, characteristic of a series a thiopeptide, suggests that TpnL reduces a precursor dehydropiperidine.</p><p>The aqueous solubility of 7 (26.01 ± 0.05 μM) is slightly improved relative to the 21.88 ± 0.05 μM value for 1.25 The antibacterial activities of 7 are comparable to those of 1, demonstrating minimum inhibitory concentrations of 0.018–0.098 μg/mL (0.010–0.058 μM; Table S5). Since 1, and presumably 2–5, bind to the ribosome and impair its activity, inhibition of prokaryotic protein synthesis by 7 was determined using an in vitro transcription-translation assay.25,26 Under these conditions, the half-maximum inhibitory concentration of 7 was 0.26 ± 0.04 μM, similar to the 0.50 ± 0.04 μM value for 1 and consistent with the activities of the two metabolites against the indicator bacterial strains. The piperidine modification does not significantly alter the antibacterial properties of 7 under the conditions tested, consistent with prior examination of the semisynthetic piperidine-containing 1 analogue.27</p><!><p>FDOR-Bs are a diverse protein family and can utilize cofactor F420, FAD, FMN, or heme, and the F420-dependent proteins catalyze a variety of reductions.28–31 TpnL contains a lysine residue (K55) conserved in most F420- dependent FDOR-Bs, part of a region including several positively charged residues, that facilitates binding the polyglutamate tail of the cofactor (Figures S16 and S17).23,32 To determine whether TpnL preferentially binds F420 or a flavin, fluorescence binding assays were performed. TpnL was heterologously expressed in E. coli with an N-terminal His6-tag. The purified protein (Figure S18) was colorless, suggesting that it does not co-purify with the flavins present in E. coli, a microorganism that does not produce F420. The intrinsic tryptophan fluorescence quenching in the presence of F420 or the flavins was used to indicate cofactor binding to TpnL (Figure S19).33 The KD value of F420 for TpnL (8.5 ± 1.4 μM) is 10 times lower than those for FAD and FMN (88.3 ± 24.4 and 105.9 ± 13.2 μM, respectively). Although the reduced form, F420H2, is most likely the native cofactor for TpnL, this protein does preferentially bind F420 over FAD or FMN.</p><p>To confirm that TpnL reduces the thiopeptide dehydropiperidine, we reconstituted its activity in vitro. Piperidine formation in series a thiopeptides is expected to be among the latter maturation steps, as the [4 + 2] condensation that assembles the core macrocycle occurs after many of the post-translational modifications on the precursor peptide.10,34 We therefore reasoned that 1 may be a substrate of TpnL (Scheme 1). To generate the reduced coenzyme F420H2 in situ, F420- dependent glucose-6-phosphate dehydrogenase (FGD) and glucose-6-phosphate were included in the assay mixture with TpnL and 1.33 TpnL catalyzed the reduction of 1 to 7, as detected by HPLC and LC-MS (Figure 2B and Figure S20). TpnL activity appears to be subject to substrate inhibition at concentrations greater than 2 μM 1 (Figure S21). To examine if the in situ generation of F420H2 could be limiting TpnL activity, F420 reduction by FGD was monitored during the time course of the TpnL assay and was found to be unaffected by TpnL and 1 (Figure S22). An additional assessment of TpnL activity at 2 and 10 μM 1 using increased concentrations of F420 (by 4.4-fold) and FGD (by 2-fold) did not demonstrate a significant alteration of TpnL activity (Figure S21). The assay conditions examined here prevented a determination of KM, but an apparent kcat/KM (2.80 × 104 M−1 s−1) was derived from the linear region of substrate dependence on rate. Our results clearly demonstrate that TpnL is an F420H2-dependent dehydropiperidine reductase.</p><!><p>A defining characteristic of thiopeptides is the core macrocycle linked through a nitrogenous heterocycle derived from two Dha residues via a formal [4 + 2] cycloaddition.10 Series d thiopeptides utilize a dedicated pyridine synthase to catalyze this cyclization, followed by dehydration and elimination of the leader peptide.9 A similar process might also lead to the core macrocycle of series a and b thiopeptides, although it is not yet understood how this permits retention of the N-terminal portion of the peptide and a more reduced ring system. There are, however, no obvious reductase candidates in either the 1 or siomycin (another series b thiopeptide) biosynthetic gene clusters, leaving the origin of the dehydropiperidine ring enigmatic.14,35 Identification of TpnL as a dehydropiperidine reductase demonstrates that the series a thiopeptide piperidine originates from the reduction of a series b thiopeptide intermediate. TpnL does accept 1, a mature metabolite bearing both the core macrocycle and the quinaldic acid loop, as a substrate. It is not yet known, however, if TpnL would act upon a dehydropiperidine-containing intermediate in which the second, quinaldic acid containing macrocycle has not yet been installed.18,36 The TpnL-dependent reduction shown here occurs independently of additional ancillary recognition elements such as the leader peptide, in contrast to many modifications in thiopeptide biosynthesis that do require the presence of the leader peptide.10</p><p>An in silico search for TpnL homologues in the NCBI database revealed potential thiopeptide biosynthetic gene clusters co-localized with a putative F420H2-dependent reductase (Figure 3). Two of these clusters appear to be capable of producing close structural homologues to 2–5, identified in Amycolatopsis saalfeldensis DSM 44993 (A. saalfeldensis) and Micromonospora carbonacea DSM 43168 (DSM 43168). The TpnL homologue identified in A. saalfeldensis (SEO90151.1) shares 70% sequence identity (78% similarity) to TpnL. The associated gene cluster is identical in composition and organization to the tpn cluster, including homologues of tpnMN that are proposed to be involved in thioamidation of the peptide backbone (Table S6). The core peptide sequence encoded in the A. saalfeldensis cluster (ASSSSCTTCICTCSCSS; SEO89922.1) is very similar to the sequence in TpnA and the thiostrepton precursor peptide TsrA.14 The putative thiopeptide biosynthetic gene cluster in DSM 43168 differs in organization from the tpn and A. saalfeldensis clusters but does contain a full set of tpn-/tsr-like genes (Figure 3 and Table S7) and also encodes homologues of TpnLMN, with the TpnL-like protein (SCF235261.1) bearing 52% identity (64% similarity) to TpnL. The core peptide sequence in DSM 43168 (SCF23644.1) differs from the thiopeptin sequence only at the first residue, is identical with the thiostrepton amino acid sequence, and is identical with that predicted for Sch 18640 6 (IASASCTTCICTCSCSS). Thus far, 6 has only been isolated from Micromonospora arborensis, and it is likely that DSM 43168 is another producer of this metabolite.6 The gene clusters identified in A. saalfeldensis and DSM 43168 may therefore be involved in the generation of thioamidated series a and b thiopeptides, similar to the complex of 2–5 produced by S. tateyamensis.</p><p>To date, only one dihydroimidazopiperidine-containing series c thiopeptide has been identified, Sch 40832 (Figure S23) isolated from Micromonospora carbonacea var. Africana ATCC 39149 (ATCC 39149).37 The genome sequence of ATCC 39149 contains a candidate biosynthetic gene cluster for this unusual thiopeptide (Figure 3), and the precursor peptide found encoded in this locus (EEP73621.1) includes a match for a predicted Sch 40832 core peptide (TSSSSCTTCICTCSCSS). Like 1 and 2–5, Sch 40832 houses a second, quinaldic acid containing macrocycle, and the genes encoding HEQ biosynthetic enzyme homologues are present in the proposed cluster (Figure 3 and Table S8). The structure reported for Sch 40832 included a disaccharide tethered to the β-hydroxyl of Thr1, and a potential glycosyltransferase (EEP73625.1) is encoded in this cluster (Table S8). A homologue of TpnL (EEP73610.1) is present in the ATCC 39149 cluster, but there do not appear to be any homologues to TpnMN, the proposed thioamidation YcaO-TfuA-like proteins. Accordingly, the reported structure of Sch 40832 lacks a thioamide.37 The role of an F420H2-dependent reductase in the biosynthesis of a series c thiopeptide is not entirely clear. It could reduce a dehydropiperidine precursor, forming a nascent piperidine subjected to additional oxidative modifications elaborating the central nitrogenous ring into the dihydroimidazopiperidine scaffold. Alternatively, the TpnL homologue could mediate an entirely different reductive step in the maturation of a series c thiopeptide.</p><p>Actinoalloteichus hoggarensis DSM 45943 (A. hoggarensis) also identifies with a putative thiopeptide biosynthetic gene cluster harboring a TpnL-like protein and contains many similarities to the components of the ATCC 39149 biosynthetic gene cluster (Figure 3 and Tables S8 and S9). The A. hoggarensis core peptide (WP_093941958.1) is identical with the proposed Sch 40832 core peptide of ATCC 39149. The A. hoggarensis cluster, however, also includes homologues to the proposed TpnMN thioamidating system. It is therefore possible that A. hoggarensis may generate another series c thiopeptide bearing a thioamide.</p><p>This is the first report of involvement of the unusual coenzyme F420 in thiopeptide biosynthesis. F420 is widely produced by methanogenic archaea and Actinobacteria, including Streptomyces and Mycobacteria, and F420-related pathways have attracted attention as targets for the selective inhibition of pathogenic Mycobacteria.29,38 F0, a biosynthetic precursor of F420, is derived from the condensation of L-tyrosine and 5-amino-6-ribitylamino-2,4(1H,3H)-pyrimidinedione, both primary metabolites (Scheme S3).39 F0 synthase (FbiC) marks the first committed step in F420 biosynthesis, and at least one copy of an FbiC candidate is present in S. tateyamensis and the S. laurentii genomes (Figure S24).40 FbiA and FbiB are involved in the attachment of the phospho-L-lactyl and polyglutamyl tail on F420, respectively.41 The incomplete sequence of the S. tateyamensis genome prevented identification of encoded FbiA and FbiB homologues, but both are encoded in the complete genome sequence of S. laurentii (Figures S25 and S26).40 S. laurentii and, presumably, S. tateyamensis both appear to be capable of producing the mature cofactor F420, and the biosynthetic pathway for this cofactor is often present in actinomycetes.38 Indeed, FbiABC homologues are encoded in several thiopeptide-producing strains, including Planobispora rosea (series d; GE2270A), Streptomyces actuosus (series e; nosiheptide), and the putative series a–c producers identified above (Table S10).42,43</p><p>Several homologues of TpnL were identified in the NCBI database, only a few of which appear to be involved in thiopeptide biosynthesis. A phylogenetic analysis of TpnL was conducted alongside its homologues and established F420H2- dependent FDOR-Bs from Mycobacterium species and Streptomyces rimosus.23,28,32,44–46 The TpnL-like proteins affiliated with a known or predicted thiopeptide biosynthetic gene cluster are in a distinct clade from the other FDOR-B proteins (Figure 4). The two proteins that may be involved in biosynthesis of dihydroimidazopiperidine-containing (series c) thiopeptides form a separate subclade from the homologues identified from bacteria that are or may be involved in the biosynthesis of series a thiopeptides.</p><p>A growing number of F420H2-dependent enzymes have been proposed to reduce a wide variety of complex substrates in actinomycete secondary metabolism. In addition to the reduction of the dehydropiperidine ring in a heavily decorated thiopeptide macrocycle as shown here (Scheme 1), an F420H2- dependent FDOR-B family protein from Streptomyces rimosus reduces an alkene in the biosynthesis of the polyketide oxytetracycline.28 A separate subset of enzymes binding F420 belonging to the luciferase-like monooxygenase (LLM) family appear to be more broadly distributed and have been implicated in the biosynthesis of the 4-alkyl-L-proline substituent found in the nonribosomal peptide metabolites hormaomycin and the pyrrolo[1,4]benzodiazepines: e.g., lincomycin.38,47,48 LLM-family F420H2-dependent reductases are also suggested to be involved in the biosyntheses of the aminoglycoside kasugamycin and a coronafacoyl phytotoxin produced in a streptomycete.49,50 It is likely that the number and types of transformations attributable to F420H2-dependent enzymes will continue to expand as more metabolic pathways are characterized.</p><!><p>In summary, this work presents the initial biochemical characterization of an enzyme responsible for the formation of a series a thiopeptide and establishes the basis for biosynthesis of the thiopeptins. TpnL is an unusual F420H2- dependent enzyme that reduces the dehydropiperidine housed within the highly modified macrocycle of 1 to a piperidine and is expected to carry out a similar transformation for thiopeptin production. Further characterization of TpnL will be needed to determine whether this reduction reflects the terminal step in series a thiopeptide biosynthesis or if the true biological substrate is an earlier intermediate. Regardless, TpnL represents a biosynthetic tool that may be useful in the development of additional thiopeptide metabolites.</p>

PubMed Author Manuscript

Phosphorylation by Nek1 Regulates Opening and Closing of Voltage Dependent Anion Channel 1

VDAC1 is a key component of the mitochondrial permeability transition pore. To initiate apoptosis and certain other forms of cell death, mitochondria become permeable such that cytochrome c and other pre-apoptotic molecules resident inside the mitochondria enter the cytosol and activate apoptotic cascades. We have shown recently that VDAC1 interacts directly with never-in-mitosis A related kinase 1 (Nek1), and that Nek1 phosphorylates VDAC1 on Ser193 to prevent excessive cell death after injury. How this phosphorylation regulates the activity of VDAC1, however, has not yet been reported. Here we use atomic force microscopy (AFM) and cytochrome c conductance studies to examine the configuration of VDAC1 before and after phosphorylation by Nek1. Wild type VDAC1 assumes an open configuration, but closes and prevents cytochrome c efflux when phosphorylated by Nek1. A VDAC1-Ser193Ala mutant, which cannot be phosphorylated by Nek1 under identical conditions, remains open and constitutively allows cytochrome c efflux. Conversely, a VDAC1-Ser193Glu mutant, which mimics constitutive phosphorylation by Nek1, remains closed by AFM and prevents cytochrome c leakage in the same liposome assays. Our data provide a mechanism to explain how Nek1 regulates cell death by affecting the opening and closing of VDAC1.

phosphorylation_by_nek1_regulates_opening_and_closing_of_voltage_dependent_anion_channel_1

INTRODUCTION<!>Cell culture<!>Antibody production and purification<!>Site-directed Nekl mutagenesis<!>Production and purification of VDAC1 fusion proteins<!>Kinase assays<!>Imaging with atomic force microscopy (AFM)<!>Liposome transport assays<!>Purification of recombinant VDAC1<!>Examination of VDAC1 by AFM<!>Conductance of cytochrome c through VDAC1 in liposomes and erythrocyte membranes<!>DISCUSSION<!>CONCLUSION<!>Supplemental Figure S1 Purification of GST-VDAC1 fusion proteins<!>

<p>Intrinsic apoptotic pathways that respond to cytotoxic stress, including DNA damage, activate a hierarchical series of caspases that disassemble cells without inciting inflammation in bystanding cells. Permeabilization of mitochondria is seminal to the regulation of cell death induced by cytotoxic or genotoxic stress [1]. Among the most important proteins released from mitochondria is cytochrome c. One established paradigm, supported by abundant evidence, holds that cytochrome c exits through a permeability transition pore composed of the outer mitochondrial membrane protein VDAC1 [2], the inner mitochondrial membrane protein ANT (adenine nucleotide translocator), and the inner membrane associated protein cyclophylin D [2–4]. A more recently proposed paradigm suggests that the primary function of VDAC in pro-apoptotic conditions is to regulate ATP flux by closure rather than by opening, such that cytochrome c release from the mitochondrial intermembrane space occurs as part of a general rupture of the outer membrane rather than passage specifically through open VDAC [5]. In either case, cytochrome c initiates the cascade of events that results in mitochondrial-mediated apoptosis. After its release into cytoplasm, cytochrome C complexes with Apaf-1 (apoptosis protease activating factor) and cleaves a series of caspases that ultimately dismantle the cell by breaking down cell-cell contacts, cytoskeletal elements, and nuclear structures [6]. The mitochondrial transition pore is affected by low intracellular pH, a relative deficiency of ATP, and Ca2+ overload [4], as well as by the balance between pro-apoptotic and anti-apoptotic members of the Bcl-2 family of proteins, which bind to VDAC1 [7, 8]. To date, however, relatively little is known about how individual components of the transition pore, including VDAC, change their conformation when post-translationally modified by specific cytoplasmic proteins, like kinases.</p><p>We have recently shown that the VDAC1 is regulated by specific phosphorylation, on serine residue 193, by never-in-mitosis A (NimA) related kinase 1 (Nek1). In the basal state and in response to injury that includes DNA damage, Nek1 phosphorylates VDAC1 to limit mitochondrial cell death [9]. Nek1 is a mammalian homologue of the NimA, a stress protein kinase in Aspergillus and in other fungi that responds to DNA damage, regulates the G2-M phase transition, and keeps chromosome transmission to daughter cells faithfully [10, 11]. We have also shown that Nek1 protein, like its fungal counterpart, but in unique ways, is likewise important in mammalian cells for efficient DNA damage responses and for proper check-point activation [12, 13].</p><p>In this report, we use atomic force imaging and cytochrome C efflux to demonstrate the consequences of Nek1's phosphorylation of VDAC1: it regulates the channel's opening and closing, and its conductance of cytochrome c. Our data support a direct role for VDAC1 in conducting cytochrome c to initiate the mitochondrial-mediated cell death cascade.</p><!><p>Human HK2 human proximal renal tubular epithelial cells were obtained from American Type Tissue Collection (Rockville, MD) and cultured in recommended media containing 10% fetal bovine serum and antibiotics.</p><!><p>Details of the production and determination of the specificity of the anti-Nek1 antibodies have already been reported [9, 12]. Normal rabbit IgG was purchased from Sigma-Aldrich (St. Louis, MO, USA).</p><!><p>A Quick-Change kit (Stratagene, LaJolla, CA) was used to generate the VDAC1-S193A and VDAC1-S193E mutant cDNAs, as previously reported [9].</p><!><p>GST-VDAC1 proteins were expressed using a bacterial system. To obtain soluble fusion proteins, cells were incubated at 37C and then the protein was induced by IPTG at either 37°, 25° or 16°C. VDAC1 (wild type), VDAC1-S193A, and VDAC1-S193E cDNAs were individually fused in-frame to glutathione S-transferase (GST) or streptadivin binding protein (SBP) expression constructs. The fusion proteins were expressed in BL21 competent E. coli and purified using glutathione sepharose or strepadivin sepharose beads. After washing with phosphate buffered saline (PBS), the GST-VDAC1 proteins were eluted with glutathione and SBP-VDAC1 proteins were eluted with biotin. After removing the free glutathione or biotin, the purified proteins were then stored at 4°C.</p><!><p>Immunopurified Nekl was used as kinase source as previously described [9]. Purified GST-VDAC1 proteins were incubated in separate reactions with immunopurified Nekl in a kinase buffer, or in a mock reaction containing pre-immune IgG immunoprecipitates instead of immunopurified Nekl. The kinase reactions were carried out in a total volume of 30 µl, with 20 µl of immune complexes and 3 µg of GSTVDAC1 (or mutants) in the presence of 3 µCi of γ-32P-ATP. After incubation for 30 minutes at 37°C, equal volumes of SDS sample buffer with EDTA were added to final concentration of 2 mM for the kinase reactions prior to the AFM and liposome transport studies.</p><!><p>The GST-VDAC1 fractions after immune kinase reactions were kept on ice. Two microliters of purified GST-VDAC1, diluted in 20 mM Tris buffer (pH 7.5) to a final concentration of 0.1 mg/ml, were deposited on a freshly cut mica surface and mounted on a wet chamber of a Nanoscope IIIa atomic force microscope (Digital Imaging, Veeco Instruments, Inc., Woodbury, NY). Imaging was performed as previously described [14]. Standard plain fit and flattening options provided with the NanoScope IIIa software were applied in the height mode to obtain multiple views to show channel pores. All experiments were repeated at least twice. Two different observers assessed the open or closed state of >200 individual VDAC1 channels. Similar expression and purification strategies were used for the GSTVDAC1-S193A and S193E mutants, which were subjected to the same analysis. Detailed measurements of individual channels for base and outside pore sizes was accomplished with Digital Imaging software.</p><!><p>Purified cytochrome c was purchased from Sigma. A very small (MW 509) Oregon Green®-488 fluorochrome label was conjugated to cytochrome c according to instructions described in the labeling kit (Molecular Probes, Eugene, OR). To make relatively large and uniform liposomes, a sonication-freeze-thaw method was used. 500 mg of phospholipid mixture (soybean Type II-S) was dispersed and sonicated in 10 ml of buffer containing 10 mM K3PO4, 50 mM KC1, 20 mM Hepes, 1 mM EDTA, pH 7.5. Following sonication, 100 µl of lipid mixture was mixed with purified GST-VDAC1, GST-VDAC1-S193A, or GST-VDAC1-S193E fusion protein and Oregon Green® 488-conjugated cytochrome c, then subjected to three freeze-thaw cycles (2 minutes freeze in liquid N2 and 30 minutes thaw at room temperature). One hour after the freeze-thaw cycles, the export of cytochrome c from individual liposomes was observed and photographed using a fluorescence microscope. More than 400 individual liposomes were assessed in duplicate or triplicate experiments to determine whether labeled cytochrome c was inside or outside. Mouse erythrocyte membranes were prepared from 200 µl of blood [15] and used instead of liposomes for some of the transport assays.</p><!><p>In order to characterize VDAC1's biochemical activity, we first produced pure wild-type and key mutant forms of it. We cloned the human VDAC1 gene and fused it to a GST tag for purification [9]. In order to avoid any denature-renature processes [16–18] and to assure that the resultant tagged VDAC1 proteins retained their native configurations, we used low culture temperature to obtain soluble forms of VDAC1 directly. GST-VDAC1 was almost completely insoluble when induced and cultured at 25–37°C, but approximately 15–20% of the protein became soluble when the culture was induced and incubated at 16°C (Supplemental Figure S1a and b) [19, 20]. With this increased yield of soluble VDAC1, we were able to purify the soluble fractions in good amounts. Identical conditions were used to produce wild type GST-VDAC1 and with the key serine 193 residue mutated to a negatively charged amino acid (glutamic acid), to mimic constitutive phosphorylation (VDAC1-S193E), and with residue 193 changed to alanine and therefore unable to be phosphorylated by Nek1 (VDAC1-S193A) [9]. We then used the purified GST-VDAC1 proteins as substrates in immune kinase studies. After phosphorylation of complexes immunoprecipitated from HK2 cells with either anti-Nek1 antibody or normal rabbit (or mouse) IgG, the GST-VDAC1 proteins were subjected to additional purification to remove any contaminants. All three GST-VDAC1 proteins migrated as single, silver stained bands in SDS-PAGE (Figure 1a). They were therefore deemed pure.</p><!><p>To explore in more detail the molecular consequences of the Nek1-VDAC1 kinase-substrate interaction, we used atomic force microscopy to examine GST-VDAC1, GST-VDAC1-S193A, and GST-VDAC1-S193E, with or without phosphorylation by Nek1 immune complexes. Individual channels could be assessed by AFM for their open or closed status (Figure 1). When only nonspecific immune complexes were included in the kinase reaction, most of the wild type GST-VDAC1 channels were open; but when Nek1 immune complexes were used, most of these channels were closed. In contrast, almost all of the GST-VDAC1-S193A mutant channels were open (more wide open even than wild type GST-VDAC1 unphosphorylated by Nek1), and all of the GST-VDAC1-S193E channels were tightly closed, whether Nek1 was included in the kinase reaction or not. Detailed analysis of many individual channels by AFM (Figure 2) showed outside dimensions of the pores in wild type, unphosphorylated VDAC1 to be slightly larger than dimensions reported recently using NMR and x-ray crystallographic data [17, 18], but not much different than previous reports based on electron microscopic and AFM measurements [21–23]. The pore dimension of the VDAC1-S193A mutant, which cannot be phosphorylated by Nek1 and which is always open, was even larger and is migrated slower than wild-type and S193E mutant channels on SDS-PAGE (Figure 1c). The open channels should be large enough to conduct cytochrome c, the largest dimension of which is approximately 3.5 nm, in either oxidized or reduced state and with its associated water molecules [24–26].</p><!><p>To confirm that open GST-VDAC1 and GST-VDAC1-S193A channels can allow cytochrome c to pass through, we performed transport assays for efflux of labeled cytochrome c through VDAC1 channels inserted into liposome membranes [8]. Results were entirely consistent with those observed for individual channels examined by AFM: phosphorylation of wild type VDAC1 by Nek1 kept the channels closed and cytochrome c inside the liposomes, but had no effect on the VDAC1-S193A channels, which remained open and allowed cytochrome c efflux with or without phosphorylation by Nek1. VDAC1-S193E channels, which mimic constitutive phosphorylation on serine residue 193, always kept cytochrome c inside, with or without phosphorylation by Nek1 (Figure 3a and b).</p><p>The N-terminal GST tags on the recombinant VDAC1 proteins did not create any significant artifacts with regard to VDAC1 channel structure or function, since nearly identical results were obtained in AFM images and liposome transport assays when VDAC1 proteins were purified with a different and even smaller streptavidin binding protein (SBP) tag (Figure 3c). The VDAC1 fusion proteins, even with different tags and when examined by AFM after insertion into liposomes (Figure 4a), assumed their native barrel-type channel structures [17, 18], and were able to conduct cytochrome c when open.</p><p>We also confirmed that the cytochrome c transport assays in artificial liposomes were not specious. Similar assays with GST- or SBP-VDAC1 constructs inserted into freshly isolated mammalian red blood cell membranes, which have intact, biological, lipid bilayers containing cholesterol, were nearly identical to those in the liposomes (Figure 4b). Taken together, the AFM and liposome/erythrocyte transport data show that phosphorylation of VDAC1 on serine 193 keeps VDAC1 closed to prevent efflux of cytochrome c, and reinforce with channel imaging and functional assays the concept that we reported previously, that regulation of VDAC1's open-closed status by Nek1 may be crucial for prevention of aberrant, mitochondria-mediated cell death after injury [9].</p><!><p>We have already determined that the biochemical and mechanistic observations reported here are relevant biologically to the phenotypes that result from a naturally occurring Nek1 mutation, by examining mitochondria and apoptosis in cells from Nek1-deficient mice [9]. The so-called kat2J (kidneys-anemia-testis) spontaneous mutant mice develop pleiotropic defects, most notably growth retardation and polycystic kidney disease (PKD), and they die prematurely [27]. The kat2J mutation results in early truncation of Nek1, eliminating all of the coiled-coil domains as well as part of the N-terminal kinase domain [28]. Cells from Nek1kat2J −/− mice express no detectable Nek1 protein and are hypersensitive to the lethal effects of DNA-damaging radiation [13]. We have also shown that the Nek1-dependent phosphorylation of VDAC1 on serine 193 is biologically important, since ectopic overexpression of the constitutively closed VDAC1-S193E mutant transiently protects Nek1 −/− cells from aberrant apoptosis after DNA damage [9].</p><p>Nek1 is involved early in the DNA damage response. Its kinase activity is increased and a portion of cellular Nek1 relocalizes from cytoplasm and mitochondria to nuclear sites of DNA damage, within minutes after gamma irradiation. Nek1-deficient cells are markedly more sensitive to the lethal effects of DNA damage compared to cells expressing functional Nek1 [13]. Nek1 expression is also upregulated in renal tubular epithelial cells after ischemic injury, before the cells either undergo frank apoptosis or necrosis, or before they repair the injury ([29] and manuscript in preparation).</p><p>Our data showing that unphosphorylated, wild-type VDAC1 and the VDAC1-S193A mutant remain widely open to allow cytochrome c efflux, and that the phosphorylated, wild-type VDAC1 and the VDAC1-S193E mutant remain closed, strongly support a direct role for VDAC1 in conducting cytochrome c in initiating the mitochondrial-mediated cell death cascade. They demonstrate furthermore and for the first time how a specific kinase, Nek1, regulates VDAC1 channel activity. The serine 193 residue that Nek1 phosphorylates on VDAC1 is predicted to be at a crucial site at the junction between a C-terminal transmembrane domain and a putative cytoplasmic protein binding domain, such that its phosphorylation would have a significant impact on the configuration of the barrel-like channel formed by VDAC1 [30, 31]. Two recent reports that used NMR and x-ray crystallography to characterize recombinant human VDAC1 structure in detail have identified a helical protrusion within the VDAC1 pore [17, 18]. This protrusion is comprised by N-terminal amino acids and is thought to be less stable than other regions of the VDAC1 barrel structure, such that it may switch between different conformations to control voltage gating [17]. It is possible that the serine 193 phosphorylation of VDAC1 by Nek1 affects movement of the helical protrusion to control closing of the channel pore. It is also possible that Nek1 phosphorylation affects dimerization or oligomerization of VDAC1 [23]; such a property could also account for the large size of the pores we examined by AFM, which had characteristics of dimers or trimers. Finally, we were careful to use relatively physiologic conditions (pH 7.2 to 7.6) in our AFM and cytochrome c conductance assays. In preliminary studies, we found that changes in pH outside of the physiologic range influenced the open or closed conformation of GST-VDAC1 by AFM, irrespective of Nek1 phosphorylation (data not shown), and therefore we standardized conditions at pH 7.5. Many reported studies have used much more alkaline pH, which could influence or prevent the opening and closing of native VDAC1, as well as its oligomerization or association with other proteins.</p><!><p>We have shown previously that mitochondrial transition pores are easily opened in the absence of functional Nek1 [9, 13], which seems in the basal state to phosphorylate VDAC1 and keep it closed. AFM and cytochrome c transport assays showing here that the VDAC1-S193E mutant remains tightly closed, and thereby keeps cytochrome c from leaking out, add support to the notion that Nek1 regulates mitochondrial apoptosis through specific phosphorylation to affect the configuration of VDAC1. Our previous studies and the results presented here suggest that Nek1 may function in defense against oxidative cellular and/or DNA damage, and that relatively trivial environmental injury may lead to aberrant apoptosis in cells deficient in Nek1. Targeting Nek1 expression or the Nek1-VDAC1 interaction may be a novel way to prevent apoptosis after injury in normal cells or to enhance apoptosis in cells that overexpress Nek1.</p><!><p>Lowering culture temperature increases the fraction of soluble GST-VDAC1 fusion proteins. Coomassie blue-stained gel (a) and Western blot with anti-GST (b) from lysates separated by SDS-PAGE after at indicated temperature. Arrow marks GST-VDAC1, which migrates with an apparent molecular mass of approximately 50 kDa. For lane markers, T: total lysate, P: insoluble pellet, and S: soluble supernatant.</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>atomic force microscopy</p><p>dithiothreitol</p><p>ethylenediaminetetraacetic acid</p><p>gultathione S transferase</p><p>never-in-mitosis A related kinase 1</p><p>phosphate buffered saline</p><p>streptavidin binding protein</p><p>sodium dodecyl sulfate polyacrylamide gel electrophoresis</p><p>voltage dependent anion channel</p>

PubMed Author Manuscript

Hydroxyl-Directed Cross-Coupling: A Scalable Synthesis of Debromohamigeran E and Other Targets of Interest

A hydroxyl functional group positioned \xce\xb2 to a pinacol boronate can serve to direct palladium-catalyzed cross-coupling reactions, apparently through the agency of a transiently formed palladium alkoxide. This feature can be used to control the reaction site in multiply borylated substrates and can activate boronates for reaction that would otherwise be unreactive.

hydroxyl-directed_cross-coupling:_a_scalable_synthesis_of_debromohamigeran_e_and_other_targets_of_in

<p>Complexity-generating reactions that apply to alkenes are critical tools for asymmetric synthesis. In this regard, the stereoselective diboration reaction is a versatile approach for the efficient transformation of olefin substrates: the vicinal 1,2-bis(boronate) products possess reactivity that facilitates functionalization of both carbons of the original alkene framework.1 While oxidation of vicinal bis(boronates) to diols is broadly applicable to an array of substrates, differential transformation of the two boronates requires a substrate bias that allows site-selective transformation of one boronate in preference to the other. In this context, the sensitivity of the Suzuki-Miyaura cross-coupling to steric effects allows selective transformation of a primary, terminal boronate without detectable reaction of an adjacent secondary boronate group. Our group has employed this feature in a selective tandem diboration/cross-coupling (DCC) strategy that allows the transformation of terminal alkenes into an array of functionalized non-racemic products via A (Scheme 1).2 It was considered that a different array of target structures would be accessible if the Suzuki-Miyaura cross-coupling could be coaxed to occur preferentially at the secondary boronate in the substrate to give B.3 Importantly, features that overturn the steric bias present in terminal bis(boronates) might also enable the site-selective cross-coupling of bis(boronates) derived from internal alkene substrates. In this Communication, we present a strategy that employs a neighboring hydroxyl group to control the site-selectivity of a Suzuki-Miyaura cross-coupling reaction. A noteworthy feature of this process is that it appears to rely upon a substrate-directed transmetallation that may involve a covalent linkage between the directing hydroxyl group and the Pd center. This reactivity mode is not commonly employed to facilitate catalytic cross-coupling reactions4 and may be a useful element in reaction design.</p><p>One strategy for the stereospecific cross-coupling of secondary alkyl boronates employs Lewis basic functional groups to enhance the reactivity of an adjacent boronic acid derivative. In this context, Suginome,5 Molander,6 Hall,7 and Takacs8 have employed adjacent amide or ester functional groups as electron donors to facilitate cross-coupling. Inspired by these studies, we considered that related functional groups positioned proximal to a reacting vicinal bis(boronate) might facilitate site-selective cross-coupling. Considering that protected alcohols are readily accommodated in asymmetric diboration, and since alcohols are an inherent feature in directed metal-free diboration reactions,9 initial experiments probed the ability of a hydroxyl group to control regioselectivity of cross-coupling (Table 1). In these experiments, purified borylated alkyl silyl ethers were first subjected to TBS deprotection with catalytic para-toluenesulfonic acid, then reagents for a Pd/RuPhos10 mediated cross-coupling were introduced. After reaction for 12 hours at 70 °C, the reaction mixture was subjected to oxidative work-up. Preliminary experiments (see Supporting Information) indicated that the reaction solvent plays a critical role in controlling the ratio of elimination products to cross-coupling products with a THF/toluene/H2O combination providing optimal selectivity.</p><p>When the above described cross-coupling conditions were applied to non-functionalized bis(boronate) 1, benzylic alcohol 2 was furnished in 71% yield. In contrast, when bis(boronate) 3 was subjected to the deprotection/cross-coupling/oxidation sequence, the presence of the hydroxyl group results in a complete turnover in regioselectivity such that 4 was provided in 72% yield. The remainder of the substrates in Table 1, most notably structures 9, 11 and 12 provide clear evidence that while directing effects from β-position are efficient (entry 5), when the directing group is positioned at the α or γ site (entries 6 and 7) the cross-coupling does not benefit from the presence of a neighboring hydroxyl group.</p><p>While the ability of a β hydroxyl group to activate a secondary pinacol boronate for cross-coupling is reminiscent of the activation provided by β-acyl groups, the reactions in Table 1 appear to operate by a distinct mechanistic principle relative to the acyl-promoted cross-couplings developed by Suginome, Molander, and Hall.5–7 This distinction is most clearly indicated by the observation that acyl-promoted cross-couplings proceed with inversion of configuration, ostensibly by generalized chelated "ate" complex C (Scheme 2a), whereas the hydroxyl-directed cross-coupling occurs with complete retention of configuration at carbon (Table 1, entry 8). The stereochemical outcome with hydroxyl direction is reminiscent of Takacs' secondary-amide-directed stereoretentive cross-coupling, although the mechanistic underpinnings for the this process are unresolved.8 Additional mechanistic information about the hydroxyl-directed reaction was obtained by examining the cross-coupling of bis(boronate) 15 (Scheme 2b). In line with the observations in Table 1, when 15 was subjected to Pd/RuPhos catalyzed coupling with 1.05 equivalents of bromobenzene, the reaction exhibited complete selectivity for the B(pin) group positioned β to the hydroxyl substituent. Of note, in addition to cross-coupling product 16, ketones 17 and 18, and unreacted 15 were isolated. When the reaction was conducted in the presence of excess bromobenzene, complete conversion of 15 occurs and only 16 and 17 were isolated as the reaction products. Considering the known capacity for Pd(II) complexes to oxidize alcohols with aryl halides as terminal oxidant,11 a plausible rationale for formation of 17 and 18 involves β-hydrogen elimination from a substrate-derived L(Ar)Pd(alkoxide). Thus the observation of products 17 and 18 establishes the capacity for substrate-derived Pd(alkoxide) formation during the course of catalytic cross-coupling reactions. A last piece of data pertaining to the origin of hydroxyl-direction arises from the X-ray structure of 15 (Scheme 2c). In the solid state structure, both the β B(pin) and the hydroxyl group occupy axial positions; however, the planar geometry of boron suggests that the two groups are too far removed for internal HO→B coordination. Thus, even in a case where "ate" complex formation appears to be precluded, the cross-coupling is site selective. Taking the above observations into account, a plausible mechanism for the directing effect observed in Table 1 involves binding of the substrate hydroxyl to an LPdAr complex, perhaps by displacement of a halide; subsequent internal delivery of Pd through a complex such as D (Scheme 2d), would generate an organopalladium complex through an inner-sphere stereoretentive transmetallation, and ultimately deliver the corresponding coupling product.</p><p>The hydroxyl-directed transmetallation can merge seamlessly with recently developed hydroxyl-directed metal-free diboration and allows for rapid stereoselective and site-selective functionalization of homoallylic alcohols. As depicted in Scheme 3, sequential directed diboration/cross-coupling, when followed by silylation or acylation, furnishes γ-oxygenated boronates from a range of substrates. Importantly, the reaction can be conducted with aryl, heteroaryl, and alkenyl electrophiles and applies equally well to terminal alkenes, internal olefins, and trisubstituted alkenes. While the terminal and internal olefins furnish 1,3-syn relative stereochemistry, the trisubstituted alkene furnishes modest 1,3-anti induction, an outcome that is inline with the stereoinduction observed in the directed diboration reactions. Lastly, the use of Pd(OAc)2 in reactions of alkenyl electrophiles appeared to minimize alkene-containing side-products derived from β-deborylation of intermediate organopalladium complexes.</p><p>Tandem diboration/directed cross-coupling reactions can allow alkene functionalization in ways that are not straightforward with current methods. These strategies are illustrated by the examples in Scheme 4. To target CCR1 antagonist 4012, we considered the reaction of homoallylic alcohol 37, a starting material that is readily assembled by allylation of Boc-protected phenylalanine.13 Tandem directed diboration/directed cross-coupling allows stereoselective C-C bond formation at the internal alkene carbon of 37. Upon TEMPO catalyzed oxidation, this reaction sequence delivers lactone 38, which is a precursor to key building block 39.12 In a related fashion, enantioselective double arylation of alkene 41 (Scheme 4b) can be accomplished with Pt-catalyzed enantioselective diboration to initiate the sequence.1c,d With the alcohol protected as a silyl ether, steric effects dominate the cross-coupling of the intermediate bis(boronate) such that bond formation occurs at the terminal carbon. Subsequent TBS deprotection and directed cross-coupling furnishes non-racemic alcohol 43, a plausible intermediate in the synthesis of vitronectin receptor antagonist 44.14</p><p>As a last synthesis example, we considered construction of debromohamigeran E (Scheme 5).15 This structure belongs to a large group of related cyclopentane-containing compounds16, but has not yet been addressed by laboratory synthesis.17 We considered that the target might be accessed by directed cis double alkylation of alkene 46. To effect this transformation, 4618 was subjected to diboration followed by directed cross-coupling with 48. Remarkably, even though the reacting boronate in 47 is at a secondary carbon and resides between quaternary and tertiary carbons, the cross-coupling proceeds stereoselectively and furnishes a single isomer of product. Of note, this reaction was readily accomplished on a preparative scale and was used to furnish 2.7 grams of silyl protected product 49, an intermediate that was characterized by X-ray crystallography. A subsequent Evans-Zweifel olefination19 with 2-lithiopropene furnished 2.1 grams (93% yield) of double alkylation product 50, a versatile intermediate for the synthesis of a number of hamigerans. After C-C bond installation, a portion of 50 was reduced and subjected to silyl deprotection. Subsequent Ru-catalyzed oxidation20 occurred at the primary alcohol site and the benzylic methylene thereby furnishing 52. Saponification afforded the target structure.</p><p>In summary, the hydroxyl-directed cross-coupling enables regioselective, stereoretentive C-C bond formation in the context of multiply borylated reagents. Moreover, the induced proximity conferred by the putative Pd-O linkage appears to facilitate reaction of boronate groups that are likely to be recalcitrant in the absence of such activating effects.</p>

PubMed Author Manuscript

In Situ Study of Nanoporosity Evolution during Dealloying AgAu and CoPd by Grazing-Incidence Small-Angle X-ray Scattering

Electrochemical dealloying has become a standard technique to produce nanoporous network structures of various noble metals, exploiting the selective dissolution of one component from an alloy. While achieving nanoporosity during dealloying has been intensively studied for the prime example of nanoporous Au from a AgAu alloy, dealloying from other noble-metal alloys has been rarely investigated in the scientific literature. Here, we study the evolution of nanoporosity in the electrochemical dealloying process for both CoPd and AgAu alloys using a combination of in situ grazing-incidence small-angle X-ray scattering (GISAXS), kinetic Monte Carlo (KMC) simulations, and scanning transmission electron microscopy (STEM). When comparing dealloying kinetics, we find a more rapid progression of the dealloying front for CoPd and also a considerably slower coarsening of the nanoporous structure for Pd in relation to Au. We argue that our findings are natural consequences of the effectively higher dealloying potential and the higher interatomic binding energy for the CoPd alloy. Our results corroborate the understanding of electrochemical dealloying on the basis of two rate equations for dissolution and surface diffusion and suggest the general applicability of this dealloying mechanism to binary alloys. The present study contributes to the future tailoring of structural size in nanoporous metals for improved chemical surface activity.

in_situ_study_of_nanoporosity_evolution_during_dealloying_agau_and_copd_by_grazing-incidence_small-a

Introduction<!>Alloy Preparation<!>In Situ Grazing-Incidence X-ray Diffraction (GISAXS)<!><!>In Situ Grazing-Incidence X-ray Diffraction (GISAXS)<!>Scanning Transmission Electron Microscopy (STEM)<!>Kinetic Monte Carlo Simulation<!>Results<!><!>Results<!><!>Characteristic Length Scale during Dealloying<!><!>Characteristic Length Scale during Dealloying<!>Dealloying Kinetics<!>Dealloying versus Electrolytic Coarsening<!><!>Dealloying versus Electrolytic Coarsening<!>Summary and Conclusions<!>

<p>Owing to their high surface-to-volume ratios, nanoporous metals are employed for applications in catalysis1−3 and energy.4−7 Among various routes for the synthesis of such nanoporous metals,8 electrochemical dealloying is a particularly versatile yet simple approach. By selectively dissolving the least noble component from an alloy in an electrolyte under potential control, various bicontinuous nanoporous metals and alloys can be produced.1,9 Pore size and composition of dealloyed metals can be controlled via process parameters, such as the type of electrolyte,10,11 temperature,12 or applied potential.13−15</p><p>Since Erlebacher devised a microscopic description of the dealloying mechanism based on two simple rate equations for dissolution and surface diffusion,16 this model has become the scientific consensus that forms the basic understanding of the dealloying process. Yet several interesting aspects of dealloyed materials are still subject of contemporary research, such as the structural stability and coarsening behavior of the ensuing nanoporous structures17 or their mechanical properties.18 A broad variety of in situ methods have been utilized to study the dealloying process itself, ranging from dilatometry,19 resistometry,20 magnetometry,21 UV–vis spectroscopy,22 transmission X-ray microscopy (XTM),23 X-ray diffraction (XRD),24−26 and Raman spectroscopy27 to real-space imaging techniques such as scanning tunneling microscopy28 and transmission electron microscopy (TEM).29 While microscopic techniques, in general, allow the extraction of size information for the nanoporous structures, they suffer from the disadvantage of averaging over a limited sample area only. Furthermore, definitions of pore size, ligament size, and ligament–ligament distance from images can be inconsistent. A software-assisted evaluation of sizes from electron micrographs improves the reliability of the extracted sizes,30,31 although such an approach requires high-quality images. An ideal complement to obtain reliable size information averaged over a larger volume are small-angle scattering methods. Small-angle neutron scattering (SANS)32 and small-angle X-ray scattering (SAXS)33−38 have been applied to study the dealloying process in situ. Although experimental scattering data contain all relevant information about the nanoporous structure, the extraction of sizes characteristic for dealloyed structures has only been shown recently.37</p><p>Here, we present in situ grazing-incidence small-angle X-ray scattering (GISAXS) experiments with enhanced local sensitivity on the surface of the nanoporous materials. In combination with Kinetic Monte Carlo simulations (KMC) and scanning transmission electron microscopy (STEM), we track the evolution of characteristic sizes during electrochemical dealloying of both AgAu and CoPd alloys and apply our findings in the context of current dealloying literature. Ligament–ligament distance and pore size are obtained by fitting GISAXS scattering curves following a model function proposed by Choi and Chen39 and first applied to nanoporous metals by Welborn et al.37 For comparison of those sizes with KMC and STEM, reciprocal space scattering curves are also calculated from atomic coordinates from the KMC simulation and from STEM using a two-dimensional fast Fourier transformation (2D-FFT) of the images. Ligament–ligament spacing, as the length corresponding to the position of the main scattering peak in reciprocal space, is identified as the proper parameter to describe the size evolution during electrochemical dealloying. Two distinct stages of nanoporosity formation (i.e., dealloying) and a subsequent structural coarsening in the electrolyte are identified in surface area, which is deduced from a fit to GISAXS data following an alternative model after Teubner and Strey.40 Characteristic length d during coarsening of nanoporous (np)Au from AgAu is found to follow a d ∼ t1/4 law, typical for surface-diffusion-controlled coarsening. A retarded coarsening behavior was found for npPd from CoPd, which is interpreted as a consequence of the higher activation energy for surface diffusion. Our results emphasize the validity of the dealloying mechanism beyond the paramount alloy system of AgAu and provide a general understanding of the electrolytic coarsening of nanoporous metals during dealloying.</p><!><p>Ag75Au25 and Co75Pd25 alloys were prepared via arc melting (Edmund Bühler MAM1) from pure metal precursors (Au, Chempur, 99.95%; Pd, Alfa Aesar, 99.95%; Co, Alfa Aesar, 99.95%; and Ag, Mateck, 99.995%). The AgAu alloy was homogenized at 800 °C for 15 h under Ar atmosphere. While AgAu forms a single-phase over the full compositional range, the Co75Pd25 alloy was annealed for 1 h at 900 °C in Ar atmosphere and subsequently quenched in water to obtain a single-phase alloy. Both alloys exhibited an fcc crystal structure, as evidenced by X-ray diffraction (XRD) measurements, which are shown in Figures S1 and S2. Both alloys were mechanically rolled to a thickness of 200–250 μm, ground, and polished to a mirror-like finish. After rolling, the AgAu samples were annealed at 650 °C for 1 h in a vacuum furnace. Specimens were cut from the alloy foils to a size of 2 × 12 mm2 for in situ GISAXS dealloying.</p><!><p>Grazing-incidence X-ray diffraction experiments were conducted at the Austrian SAXS beamline at the ELETTRA synchrotron in Trieste, Italy.41 The incoming beam was set at 16 keV (0.77 Å) and directed onto the sample at an incidence angle of α = 0.2°. 2D scattering patterns were recorded by means of a pixel detector (Pilatus3 1 M, Dectris Ltd.) at a distance of 194.63 cm from the sample. The alloy platelets used for dealloying were positioned in a properly designed, 3 mm thin, electrochemical cell optimized for in situ GISAXS experiments,42 which is schematically shown in Figure 1. The samples were electrically contacted from the top and two Kapton windows allowed X-ray transmission across the cell. Two syringe pumps (Teledyne 500D) were connected to the side plugs of the cell and were used to continuously flow fresh electrolyte above the sample (0.1 M H2SO4 diluted from a 75% H2SO4 solution with a Milli-Q water flow rate of 3 mL/min) to remove atoms dissolved from the sample, thus reducing X-ray absorption.</p><!><p>Experimental setup for in situ GISAXS during dealloying. Electrochemical dealloying was conducted in a three-electrode electrochemical cell with a continuous electrolyte flow (center left). The corresponding behavior of current over time is schematically depicted in the top left corner. During etching, a detector continuously recorded GISAXS patterns (top right), of which horizontal and radial cuts were extracted for quantitative fitting of size parameters. Alloy specimens were in contact with the electrolyte from the top, which ensured dealloying progress into the sample from top to bottom (bottom right). For STEM imaging, lamellas were cut as indicated in the sketch.</p><!><p>The three-electrode setup in the present case consisted of an alloy platelet as the working electrode, which was fixed and aligned in the cell compartment using setscrews. Connection to a potentiostat (Autolab, PGSTAT204) was established using a Au wire. A homemade Ag/AgCl wire and a Au wire were used as the reference electrode for dealloying CoPd and AgAu, respectively. A coiled Pt wire served as the counter electrode. The electrolyte was 0.1 M H2SO4 diluted from a 75% H2SO4 solution with Milli-Q water. For coating the AgCl layer on the Ag/AgCl reference electrodes, Ag wires (0.25 mm, Mateck, 99.995%) were immersed in 3 M KCl solution, when a current density of +1 mA/cm2 was applied for 3 min. After every 30 s, the current was reversed for 5 s to obtain a denser coating.</p><p>A potentiostatic dealloying potentiostatic dealloying procedure was used for in situ measurements, which recorded current as a function of time at a constant potential of UD = 0.73 V (vs Au) for AgAu and UD = 0.55 V (vs Ag/AgCl) for CoPd, as schematically shown in Figure 1.</p><p>During dealloying, exposure time to the X-ray beam was adjusted as a function of the elapsed time of dealloying as follows: for the first minute of dealloying, the exposure time was set to 0.095 s; during the second minute of dealloying, the exposure time was increased to 0.495 s; and during the third minute of dealloying, the exposure time was then set to 0.995 s. For the following seven minutes of dealloying, the exposure time was set to 8 s and, for the remaining part of the measurement, the exposure time was set to 20 s. The initial, short exposure time was used to record the fast evolution taking place during the very early stages of the process, while its progressive increase allowed us to record the evolution of the sample with a higher signal-to-noise ratio.</p><p>The IGOR Pro software (IGOR Pro 7.0.8.1, Wavemetrics) was used for data reduction and fitting. For each sample, the background signal constituted by the experimental cell filled with the electrolyte was subtracted at first. Then, for both samples, analysis was performed from a horizontal cut (in-plane direction) calculated at the height of the Yoneda wings; only for the CoPd sample, an additional radial cut was calculated to reveal the isotropic scattering in transmission, as shown in Figure S3. A first qualitative analysis was carried out by observing the evolution of the calculated scattering correlation length ξ of the horizontal cut.43,44 The correlation length ξ was calculated within the range of 0.15–2 nm–1 for AgAu and 0.3–2 nm–1 for CoPd. For a quantitative analysis, the focus is set on the relative changes of the structural parameters. As a consequence, we have used a simplified analytical model to interpret the in-plane cuts rather than the full application of the distorted wave Born approximation (DWBA).45 To monitor the evolution of the nanoporous structure, one-dimensional (1D) scattering patterns of the horizontal cuts were fitted by means of two models: the so-called Choi–Chen model39 was used at first, while the Teubner–Strey model40 was employed to complete the characterization by including the analysis of the evolution of the surface area, and it was adapted to also model the behavior in the low-q region. The Choi–Chen model was developed for studying two-phase systems developing in a three-dimensional environment by means of three main parameters: an interdomain distance (L̃), here used to describe average ligament–ligament distance, a coherence length of the local domain order (R), here used to describe the average pore radius, and an interfacial length (δ), which takes into account the surface roughness in between the two phases.39 From these parameters, the average pore radius was calculated as DP = 2(R + δ). A power law term was added to take diffuse scattering into account of large structures as well as for the contribution of the specular reflected beam in the in-plane direction. The complete model can be written aswhere a = 2π/L̃, b = 1/R, and c = 1/δ. Although the Choi–Chen model provides a detailed description of a two-phase system, it does not include any information about average surface area. Thus, for monitoring the evolution of the surface area during dealloying, also the Teubner–Strey model was used to fit the experimental data. The Teubner–Strey model is based on an order parameter expansion of the Landau mean free energy density up to second order including two gradient terms. The expansion coefficients ai and cj are fixed to zero, except a2 > 0, c1 < 0, and c2 > 0.40 Indeed, the coefficient c1 is related to the creation of domain walls between two heterogeneous phases, while the coefficient c2 concurs in system stabilization. In this way, the term (a2 + c1q2 + c2q4)−1, for negative values of c1, gives rise to a broad structure peak, which is followed by a decaying intensity, proportional to q–4 in the Porod regime. From the Fourier transformation of the corresponding correlation function (γ(r)=L̃(2πr)−1e–r/R sin(2πr/L̃)), two length scales can be found40 describing the average size of the poreand describing the average ligament–ligament distance.The coefficient a2 was set to 1. The average surface area was estimated to be proportional to the ratio of forwarded scattering intensity and the coefficient c2, ITS/c2. An additional power law was added to account for large buried aggregates, crack formation, and/or changes of the surface roughness, while the diffuse scattering of the off-specular contribution in the low-q region of the horizontal cut was approximated by means of the so-called Guinier–Porod model by Hammouda, H(ID,Rg,d,q).46 Here, the important contributions are ID (in the Guinier–Porod model corresponding to the Guinier scaling factor) and the power law d representing changes of the surface roughness. Thus, the Teubner–Strey model used for data fitting can be expressed asHere, the results of the power laws and the Guinier–Porod model were considered as qualitative measures and not further analyzed. Additionally, the contribution of the Guinier–Porod model was not added when analyzing the radial cut, due to the reduced range of integration in the low-q region.</p><!><p>For STEM investigation of the nanoporous structures, samples were prepared in the same flow-cell setup as used for in situ GISAXS, while all experimental conditions stayed the same. Dealloying was stopped after 1100 s for the AgAu alloy and 4500 s for the CoPd alloy to obtain similar structures as in the late stages of the respective in situ dealloying experiments. An alloy backbone was conserved in either case to ensure mechanical stability for the preparation of TEM lamellas, which is schematically shown in Figure 1. Only the topmost layer (5–10 μm) of the nanoporous structures was prepared for STEM investigation utilizing a focused ion beam (FIB) to monitor a similar morphology as in the scattering patterns via GISAXS. The analysis has been performed using the ASTEM probe-corrected Titan3 G2 60-300 microscope (Thermo Fisher) operated at 300 kV (beam diameter of 1 Å, convergence angle of 19.6 mrad). Images were acquired by annular and high-angle annular dark-field detectors (ADF and HAADF). The data have been processed by the Gatan Microscopy Suite 3 (GMS).</p><!><p>Kinetic Monte Carlo simulations of the dealloying process were based on rate equations originally proposed by Erlebacher.16 The rate equations for surface diffusion, allowed for both atomic species in the alloy, and dissolution of the less noble alloy components (Ag or Co only) are1and2where kdiff is the diffusion rate, νD is the attempt frequency for surface diffusion (set as 1013 s–1), n is the coordination number of the respective lattice site, Eb is the single-bond binding energy, kdiss is the dissolution rate, νE is the attempt frequency for dissolution (set as 104 s–1), ϕ is the parametrized electrode potential, and T is the temperature, with e and kb being the usual physical constants.</p><p>Binding energy Eb for AgAu was adopted from ref (16), while for the binding energy of CoPd, we used the AgAu value scaled by a factor of 1.2 accounting for differences in melting temperature of the two alloys (Eb = 0.15 eV for Ag75Au25, Eb = 0.18 eV for Co75Pd25). The electrode potential ϕ for AgAu was set as 1.14 V, which corresponds to the critical dealloying potential of this materials system in the simulation.16 Accounting for the higher applied potential for AgAu in the dealloying experiment, as well as the intrinsically different dissolution potentials in the electrochemical series of Ag (+0.8V)47 and Co (−0.28 V),47 ϕ was fixed to 1.5 V for CoPd. The value of ϕ determines the balance between dissolution and diffusion only on short timescales (τ < 100 s) in the simulation. Simulated real-time τ was calculated incrementally each step followingwhere K is the sum of the rates for all processes and u ∈ (0, 1] is a uniform random number.</p><p>In the post-dealloying coarsening regime, differences in binding energy Eb are responsible for the time evolution of the structural sizes. The KMC simulation of the dealloying process in this work was implemented in MATLAB. Atomic coordinates of both alloy constituents were exported every 104 steps.</p><p>For the calculation of scattering patterns, simulated boxes were randomly rotated using 90°-rotation matrices and replicated on a triclinic lattice with 3 × 4 × 5 sites. This larger structure was created to suppress box-size periodicity in reciprocal space. Lattice constants of the fcc structures were fixed to 3.65 Å for CoPd and 4.07 Å for AgAu. The Crysol48 software allowed us to calculate the small-angle X-ray scattering patterns corresponding to Au and Pd nanoporous structures retrieved by means of KMC simulations (maximum order of harmonics: 21, order of Fibonacci grid: 18). Simulated structures were displayed by means of the open-source software OVITO.49</p><!><p>For our comparative study of the dealloying processes for CoPd and AgAu alloys, in a first step, STEM images of dealloyed nanoporous Pd and Au structures are shown in Figure 2a,b in late stages of the dealloying process, i.e., after complete conversion of the alloy into the nanoporous structure in the topmost region (∼100 nm), which is the region investigated experimentally. A significantly coarser structure is observed for nanoporous Au in Figure 2a, compared to nanoporous Pd produced from the analogous experiment in Figure 2b. The disordered and irregular nature of the nanoporous structures apparent in the STEM images illustrates the problem of identifying a representative length on the basis of such images, as one could define ligament width and length, pore size, or interligament spacing. To bypass this problem, we calculated radially averaged 2D fast Fourier transformations (STEM-FFT) of both images, which includes size information in the reciprocal space50 (see Figure S4 in the Supporting Information for a schematic representation of the calculation steps). These curves are depicted in Figure 2c for nanoporous Au and 2d for nanoporous Pd. On the basis of peak position q in reciprocal space, one could estimate a characteristic length as , which yields ∼10 nm for npAu and ∼3 nm for npPd, which agrees with structural sizes from real-space images. The second sharp peak in the STEM-FFT curve in Figure 2d at lower q-values is attributed to the changing contrast in the STEM image in Figure 2b and is considered as an artifact.</p><!><p>STEM HAADF and small-angle scattering. STEM images after dealloying (a) AgAu and (b) CoPd for 1100 and 4500 s, respectively (scale bar is 20 nm). A comparison of radially averaged 2D fast Fourier transformations calculated from STEM images displayed in black and the corresponding measured GISAXS scattering patterns in red for (c) AgAu and (d) CoPd. Time-resolved in situ GISAXS horizontal cuts recorded for (e) AgAu and (f) CoPd undergoing the respective dealloying procedure. The corresponding time-resolved in situ SAXS radial cuts for CoPd are shown in the inset in (f). The correlation length ξ from experimental data is plotted in (g, h) for AgAu and CoPd, respectively. Note: The transition from scattered data points to a smooth curve is caused by the increasing exposure times to the X-ray beam.</p><!><p>As this information in reciprocal space is equivalent to information from small-angle scattering techniques, we directly introduce results from grazing-incidence small-angle X-ray scattering measurements in Figure 2c,d, which were acquired using an equivalent dealloying procedure as for STEM samples. Horizontal cuts from 2D GISAXS patterns for npAu and npPd are depicted in red, alongside the corresponding STEM-FFT curves in black. Intensities for the GISAXS cuts have been rescaled to enable comparability with the STEM-FFT curves. For npPd from CoPd, a radial cut (SAXS geometry) is additionally shown in orange. The shape of GISAXS scattering curves for npAu in Figure 2c closely resembles the STEM-FFT image in the same plot. The knee in the curve, which is indicated with a vertical line, appears at a larger q-value for STEM-FFT in black compared to the GISAXS scattering curve in red, which points toward a smaller structural size observed via STEM. Scattering curves for npPd in Figure 2d generally appear at larger q-values compared to npAu, which reflects the larger structural sizes for npAu. For npPd, the shift of the scattering curves from STEM-FFT and GISAXS is more apparent, as again seen in the position of the knee, marked by the vertical lines. This larger shift in q corresponds to a larger divergence of the structural sizes observed in STEM and GISAXS experiments for npPd. Finer nanoporous structures on the surface compared to the underlying bulk are a known phenomenon for npAu11,51−53 and might be responsible for the observed divergence of sizes from scattering patterns and electron micrographs here. STEM images were recorded on the topmost surface layer, where such a finer structure can be expected. GISAXS, on the other hand, has a larger penetration depth of ∼60 nm assuming a classical absorption law due to the rough interface layer, which is enough to observe the larger structural sizes in the bulk nanoporous material.</p><p>Finally, Figure 2e,f shows results from our time-resolved in situ GISAXS experiments for dealloying of AgAu and CoPd, respectively. Colored traces represent scattering patterns at different points in time, from blue at the start to orange at the end of the measurement. Scattering patterns evolve from featureless power law behavior attributed to surface roughness and partly to the diffuse scattering in the electrolyte, to showing a pronounced scattering peak already after the first seconds of dealloying. With ongoing dealloying, the scattering peaks move toward smaller q-values, indicating an increase in the average ligament–ligament distance. The evolution of the size of the system is then reflected in the rising correlation length ξ calculated from the GISAXS patterns, which is shown in Figure 2g for AgAu and in Figure 2h for CoPd.</p><p>Moreover, concerning the dealloying of CoPd, a more detailed analysis is required. At the beginning, the GISAXS mode is dominant as the off-specular reflected contribution is very strong in the out-of-plane direction. This can be also observed in Figure S3c in the sharp interface due to the sample horizon. At the end (Figure S3e), there is a more or less isotropic scattering due to the transmission geometry. Over the time course, a significant drop of the specular reflected intensity can be observed. Such a sharp variation is attributed to the mechanical bending of the sample during the experiment induced by the mechanical stress, which is visible also in the variation of the out-of-plane scattering distribution (data not shown). This change of sample geometry caused the incident X-ray beam to pass through the sample and changed the experimental conditions to a transmission experiment. Thus, small-angle scattering analysis needed to be shifted from the initial GISAXS mode to the transmission SAXS mode by performing a radial cut on the 2D scattering pattern, which is shown in the inset in Figure 2f. Although the transition from GISAXS to SAXS does not occur at a precisely defined time, from the evolution of the structural parameter of both techniques, we determined the middle of their overlap regime at a time of 655 s (marked by the black arrow) as the point of transition. This overlap regime was defined within the region in which the dimension of the average ligament–ligament distance retrieved by data fitting was giving roughly the same value.</p><p>The effects of this geometrical variation of the CoPd sample are mainly observed in the low-q area of the scattering pattern (Figure S5) and were quantified by fitting the two data sets with the Teubner–Strey model; whereas in AgAu, the reduction of the scaling factor ID for the Guinier–Porod contribution (off-specular intensity) in the low-q region is only slightly decreasing (reduction factor of 10) (Figure S6), in CoPd, there is a quasi-disappearance of this scaling factor ID (reduction factor of 400).</p><p>For a systematic analysis of our in situ GISAXS experiments of AgAu and CoPd dealloying, we performed complementary KMC simulations of the dealloying process for both alloys. Coordinates of atoms were tracked in the course of the simulation and exported every 104 simulation steps. In Figure 3, unit cells from the KMC simulations of 30 × 30 × 30 atoms are shown after a dealloying time of τ = 20 s and 10 000 s for AgAu (a) and CoPd (c). Apparently disconnected regions in the box are artifacts arising from the periodic boundary conditions in two dimensions. Dealloyed nanoporous structures appear coarser for npAu compared to npPd after both 20 and 10 000 s of dealloying, which is in line with results from STEM and GISAXS in Figure 1. For both npAu and npPd, a coarsening with progressing time is also apparent. For a direct comparison with GISAXS data, scattering curves from STEM micrographs for different dealloying times were calculated from the atomic coordinates using the Crysol software (see the Materials and Methods Section), which are displayed in Figure 3b,d. A pronounced peak is visible for both npAu and npPd, which shifts toward lower q-values with increasing time. Simulated scattering patterns accurately reflect the behavior from the GISAXS experiments in Figure 2e,f. Both scattering curves from experiment and simulation were modeled to extract quantitative size parameters, characteristic for the nanoporous structures, with the fits according to the Choi–Chen model39 drawn as black lines in Figure 3b,d. The temporal evolution of characteristic size parameters during dealloying extracted from this model and their interpretation are the subject of the next section.</p><!><p>Scattering curves from KMC simulations. Representative real-space structures of a single unit cell with a size of 30 × 30 × 30 atoms from the KMC simulation after τ = 20 s and 10 000 s are shown in (a) for AgAu and (c) for CoPd. Note the finer ligament size (corresponding to a higher q-value for the scattering peak) for CoPd. Calculated scattering patterns for (b) AgAu and (d) CoPd from atomic coordinates exported during the KMC simulation. The calculated scattering curves from atomic coordinates for the coarsening of npAu from ref (54) are added as dotted lines. Black lines represent fits to the scattering curves following the Choi–Chen model.39 Scattering curves are averaged over several unit cells in a triclinic structure to reduce artifacts related to the finite box size.</p><!><p>Nanoporous metals prepared via electrochemical dealloying bear morphological resemblance to two-phase systems obtained via spinodal composition. Both represent examples for nonperiodic, bicontinuous structures of two intertwined phases, generated via a phase separation process. Nowadays, such structures are successfully modeled using Gaussian random fields, where the two different phases are separated using a level cut.55 Random fields are generated as a superposition of standing waves of random phase and direction but with only a single underlying wavelength λ. For nanoporous structures, this wavelength λ has been used to define a characteristic spacing L̃(55,56) between centers of neighboring ligaments via the first maximum in the correlation function. In the KMC simulation study of Li et al.56 on the coarsening behavior of nanoporous gold, this characteristic spacing L̃ was shown to be the size parameter most suitable for studying the size evolution during coarsening, compared to apparent ligament sizes deduced from electron micrographs or inverse surface area. For the present study on the size evolution during dealloying, we use small-angle X-ray scattering to uncover the underlying wavelength of the nanoporous structure. Through model fits to reciprocal space data from the experiment and simulated scattering curves from KMC simulations, we probe the size parameter L̃ consistently for both methods.</p><p>Size information at a given point of time t in the experiment or τ in the KMC simulation is extracted from fits to the experimental scattering curves in Figure 2e,f and the simulated scattering curves in Figure 3b,d, as described in the Materials and Methods Section. The Choi–Chen model39 enables a robust fitting of scattering curves from nanoporous materials with a broad scattering peak using three underlying length scales, first applied to dealloyed nanoporous materials by Welborn et al.37</p><p>Two of these length scales are of special relevance for the dealloying process: (1) The characteristic ligament–ligament spacing L̃, as introduced above. As L̃ is connected to an underlying wavelength, it can be considered as a measure for periodicity in the nanoporous structure. This quantity L̃ scales inversely with the peak position in the scattering curve. (2) A ligament or pore diameter Dp. In terms of scattering, this is interpreted as a measure for the spatial decay of electron density fluctuations.37</p><p>In our comparative study for dealloying of Au and Pd, the characteristic parameters L̃ and the additional parameter Dp from the Choi–Chen model are extracted in Figure 4 for both elements from experimental GISAXS data (a,b) and from the calculated scattering curves from the KMC simulation (c,d) as a function of experimental time t and simulated time τ, respectively.</p><!><p>Evolution of structural sizes during dealloying. Pore size DP and interpore spacing L̃ as a function of etching time t from experiment (a, b) and as a function of simulated time τ from KMC simulations (c, d) in double-logarithmic representation. Note the different x-axis scaling for the lower and upper graphs. Size parameters are obtained via fitting horizontal cuts of the GISAXS patterns in Figure 2 and calculated scattering curves from KMC simulation in Figure 3 using the Choi–Chen model.39 For CoPd, both radial and horizontal cuts of the GISAXS patterns were fitted. The same fitting procedure was used on calculated scattering curves from literature KMC data54 for the temperature-driven coarsening of nanoporous Au at 900 K, which are included in (c, d) as a reference with a shifted timescale (green crosses). Data points for AgAu (CoPd) are shown as black crosses (red circles) in all subplots. Continuous lines correspond to a slope of 0.25, which is characteristic for surface diffusion-driven coarsening (d ∼ t1/4).</p><!><p>Our approach of extracting the ligament–ligament diameter L̃ and ligament/pore size Dp from both experimental and simulated scattering curves from KMC simulations enables direct comparability of all relevant length parameters between experiment and simulation. In contrast to the previously used calculated L̃ from specific surface area and porosity,56 the direct fitting of simulated scattering curves offers additional advantages if deviations from a self-similar structure evolution occur, i.e, if not only the scale but also the morphology of the structure changes. The evolution of the characteristic ligament–ligament distance L̃ and pore/ligament size DP in the dealloying process is discussed in the following.</p><!><p>First, we consider the experimental length scales for AgAu and CoPd in Figure 4a. In the double-logarithmic representation, DP increases monotonously up to a plateau at DP = 7 nm for AgAu and a plateau at DP = 5 nm for CoPd. The ligament or pore size DP is systematically larger for AgAu compared to CoPd by about 2 nm for all data points. The behavior of the ligament–ligament spacing L̃ as a function of time t in Figure 4b shows a monotonous increase for all data points for both AgAu and CoPd, with a noticeably larger slope for AgAu. The gray line in Figure 4b represents an increase of L̃ following a t1/4 law, which points toward a surface-diffusion-driven coarsening.57 Data points for the dealloying of AgAu roughly follow this t1/4-behavior, which is in line with previous studies on the coarsening of npAu.12,38,58 The ligament–ligament spacing L̃ for CoPd, on the other hand, shows a considerably smaller slope, following a coarsening behavior with a much smaller coarsening time exponent of about t1/20 (dashed gray). A small jump is apparent in the data points for CoPd at the transition from GISAXS fitting of horizontal cuts to the SAXS fitting of radial cuts, which arises due to systematic scaling differences for the two fitting methods as GISAXS probes the in-plane component and SAXS probes the radially averaged components.</p><p>In Figure 4c,d, the same parameters DP and L̃ were obtained for scattering patterns calculated from KMC simulation data. It should be noted that experimental timescales t in Figure 4a,b and timescales for the simulation τ in Figure 4c,d cannot be directly compared, due to an influence of the box size and rate parameters on the simulated time τ from the KMC simulation. Nonetheless, all data points for CoPd and AgAu in the respective subplots are fully comparable, regardless of the x-axis being experimental time t or simulated time τ. For further comparison with literature data, KMC simulations for the thermal coarsening of nanoporous Au with a porosity φ = 0.35 at 900 K have been evaluated. Data for these simulations have been previously published in the study of Li et al.,54 with atomic coordinates at various coarsening stages available online. In analogy to our in-house KMC simulation of the dealloying process, atomic coordinates from the reference KMC coarsening simulation54 were used to calculate scattering patterns, which in turn were fitted using the Choi–Chen model. Times from the reference were rescaled to match size parameters from the later stages of the KMC dealloying simulation and enable a comparison of electrolytic and thermal coarsening.</p><p>The ligament or pore size DP for both AgAu and CoPd in Figure 4c shows similar features compared to the experimental values of DP in Figure 4a. Systematically larger values are again observed for the AgAu dealloying simulation compared to CoPd, while both are in decent agreement with the values for DP obtained from the experiment in Figure 4a. The plateau at higher dealloying times in Figure 4a, however, does not exist for pore sizes extracted from the KMC simulations in Figure 4c. Data points from the high-temperature coarsening KMC reference54 in green perfectly match the slope from the dealloying KMC simulation at larger times in black, indicating that the same mechanism is responsible for both high-temperature and electrolytic coarsening. The ligament–ligament distance L̃, as the characteristic parameter for the size evolution, is depicted in Figure 4d for the KMC simulation of AgAu and CoPd. The curves show the same trend as values for L̃ from the experiment in Figure 4b. After an initial period of constant L̃, a linear increase is observed for larger simulated times τ. This is in excellent agreement with the theoretical prediction of a surface-diffusion-controlled growth with a coarsening exponent of n = 4 (L̃ ∼ t1/4), represented by the gray lines. Values for L̃ from the simulation are generally lower compared to the experimental values in Figure 4b, possibly connected to an overestimation of the binding energy parameter in the KMC simulation.</p><p>An unexpected new finding is the retarded coarsening behavior of npPd compared to npAu, which is evident in the size parameter L̃ from both the GISAXS experiment and the KMC simulation in Figure 4b,d. In the following, we argue that this is a natural consequence of the higher binding energy for Pd compared to Au and interpret it as a period of dominant structural faceting.</p><p>Studies on the coarsening of nanoporous Au12,38,58 generally report a decent agreement with a kinetic t1/4 -scaling law behavior, typical for a surface-diffusion-driven process. In contrast, Son et al.59 observed a considerably slower thermal coarsening behavior of npAu using a combined experimental and KMC simulation approach. They found that the typical coarsening exponent for surface diffusion n = 4 describes the temporal evolution during annealing only at sufficiently large temperatures. For lower annealing temperatures of T = 450 °C, a much larger coarsening exponent in the order of n ∼ 12.5 (L̃ ∼ t0.08) was extracted, which indicates a retarded coarsening behavior similar to our observation for the electrolytic coarsening of npPd.</p><p>In the recent KMC simulation study by Li et al.56 on the coarsening behavior of npAu, a slower coarsening behavior has been found at a temperature of T = 900 K compared to a higher annealing temperature of T = 1800 K, which has been ascribed to a more pronounced stage of faceting at lower T. They argue that a single activation energy can account for both low- and high-temperature coarsening kinetics, as both coarsening curves coincide when rescaling the time axis with a constant factor (corresponding to a shift on the logarithmic x-axis). An analogous argument is advanced in the present study for electrolytic coarsening, as the evolution of the characteristic size parameter L̃ for AgAu and CoPd in Figure 4b,d also follows the same behavior on different timescales. In Figure 4b,d, a period of retarded coarsening is observed for npPd (L̃ ∼ t1/20), with a stronger increase in characteristic size commencing only at larger times. Simulated scattering data indicate a transition to a typical L̃ ∼ t1/4 behavior for large enough dealloying times. In analogy to the argument of lower annealing temperatures being responsible for faceting, we argue that higher binding energy for CoPd compared to AgAu induces a ligament faceting during electrolytic coarsening in a similar way. Indeed, when considering the rate equations for surface diffusion and dissolution in eqs 1 and 2, one finds that a higher/lower annealing temperature formally causes the same change in the rate constants as would be expected for a lower/higher binding energy. In a simple microscopic picture, a higher temperature promotes atomic movement processes in the same manner as a weaker binding to neighboring atoms does. Different annealing temperatures are equivalent to different binding energies, and thus different starting alloys for dealloying. Our results confirm the applicability of Erlebacher's dealloying model16 to other material systems beyond AgAu. Furthermore, our results suggest that it is possible to obtain the same microstructure and even the same dealloying kinetics for AgAu and for CoPd at an elevated temperature.</p><!><p>Dealloying and coarsening per se are distinctively different processes. Dealloying, as the formation of a nanoporous structure from solid alloy, increases the surface area over time. Coarsening, on the other hand, with the growth of nanoporous ligaments and pores, decreases the surface area over time. Despite the converse effect of dealloying and coarsening on the overall surfaces area, both processes crucially depend on the same microscopic process of surface diffusion. While the process of electrolytic coarsening has been addressed in the previous section, here, we focus on the dealloying step solely.</p><p>Small-angle X-ray scattering in transmission geometry yields a signal that averages over all sample parts, from the newly formed porous layer right at the dealloying front to late-stage coarsening. Using the small-angle X-ray scattering technique in GISAXS geometry, a constant volume defined by the interaction volume of the incident X-rays with the sample is examined. Depending on the material system, the penetration depth of X-rays in the sample in GISAXS geometry is in the order of 60 nm (AgAu) and 130 nm (CoPd), which allow the quasi-local investigation of the dealloying process in that limited volume and thus a time-resolved study of both dealloying and electrolyte coarsening.</p><p>Here, we introduce ITS/c2 as a measure of specific surface area (in the dimension of m–1), which is computed from the scattering forward probability ITS and the fit parameter c2 from the Teubner–Strey model fit to GISAXS scattering data (see the Materials and Methods Section). The value ITS/c2 is presented as a function of time t for Au and Pd in Figure 5, while the equivalence of the results retrieved from fitting the dataset with the Choi–Chen and Teubner–Strey models is shown in Figure S7.</p><!><p>Evolution of surface area during dealloying. Temporal evolution of the ratio between scattering forwarded probability ITS and the parameter c2 retrieved from fitting the measured data sets with the Teubner–Strey model.40 Data points for AgAu (CoPd) are shown as black crosses (red circles). For CoPd, both radial and horizontal cuts of the GISAXS patterns were fitted. The black line represents a decreasing surface area following a t–1/4 law.</p><!><p>Two different dealloying regimes for AgAu are clearly visible in Figure 5 (black crosses). After an initial increase of surface area for the first 20–30 s, the surface area decreases over time for the remaining data points, roughly following a t–1/4 law as indicated by the black line. We assign the initial surface area increase to the true dealloying stage in the sampled volume, i.e., the conversion of solid alloy to nanoporous metal. This conversion occurs at a sharp phase boundary between alloy and nanoporous structure, which is referred to as the dealloying front (see Figure S8). Based on the interaction depth of X-rays in the experiment of about 60 nm and the period of increasing surface area, which marks the dealloying time, a velocity of the dealloying front can be estimated to vD = 2–3 nm/s. In view of a decreasing dealloying front velocity with decreasing etching agent concentration for AgAu in free corrosion experiments,23 such a value appears reasonable considering the dilute acid concentration of 0.1 M here. A faster dealloying process with higher corrosion potentials can be expected for electrochemical dealloying compared to free corrosion, which is a direct result of the stronger driving force for dissolution imposed via the applied potential. This directly relates to the results for CoPd in Figure 5, where no period of increasing surface area can be observed for the data points in red. As the corrosion potential was effectively higher for CoPd, as shown in the Materials and Methods Section, a faster dealloying can be expected, which occurs prior to the first data point for CoPd in Figure 5. A linear relation between current density and etch front velocity has been suggested for electrochemical dealloying.23 Using this relation, a higher dealloying front velocity for CoPd is confirmed by the initial etching currents, which were about 5 times larger for CoPd compared to AgAu (not shown). Over the whole time span, the surface area for CoPd decreases, with a small jump apparent at the transition of GISAXS to SAXS due to intrinsic scaling differences in the two underlying models. No unambiguous power law behavior can be identified for the coarsening-related surface area decrease for the dealloying of CoPd.</p><p>Despite equal outer dimensions, the total etching process for CoPd takes a longer time of 4500 s compared to 1100 s for AgAu. While the dealloying step itself is faster for CoPd, as shown in Figure 5, the longer total etching time is related to a longer phase of dissolution from the already porous structure behind the dealloying front. This dissolution phase concurs with the observed coarsening stages for AgAu and CoPd in Figure 5. The preceding formation of nanoporosity in the dealloying process is governed by dissolution kinetics and thus faster for CoPd, which is observed as a faster dealloying front in Figure 5.</p><p>Trends in surface area extracted from the Teubner–Strey model in Figure 5 can also be compared to the ligament size parameters from the Choi–Chen model in Figure 4 exploiting the inverse proportionality between specific surface area and ligament size. Related to our measurements, an inverse behavior of Dp in Figure 4 and surface area Ip/c2 in Figure 5 upon electrolytic coarsening (i.e., after 20–30 s for AgAu and right from the start for CoPd) can therefore be expected, which is confirmed by the extracted values from the experiment.</p><!><p>In this study, we investigated the dealloying and electrolytic coarsening behavior of AgAu and CoPd alloys using in situ grazing-incidence small-angle diffraction supported with a combination of scanning transmission electron microscopy and kinetic Monte Carlo simulations. Results from different techniques were compared in reciprocal space, utilizing fast Fourier transforms of the STEM images and calculated scattering patterns from the KMC simulation. Scattering curves were fitted using the Choi–Chen model, where length scales of the nanoporous structures were extracted as a fit parameter. A discussion of dealloying kinetics was conducted on the basis of ligament–ligament distance as the characteristic length for nanoporous metals upon coarsening. While the kinetic behavior for AgAu confirms previous literature findings, a slower coarsening kinetics for CoPd was detected. We argue that dealloying of both alloys follows the same surface-diffusion-driven coarsening mechanism, with slower kinetics being a natural result of surface faceting due to the higher binding energy of CoPd.</p><!><p>Additional information on the experimental results and details about the conducted analysis (PDF)</p><p>jp1c09592_si_001.pdf</p><p>Open Access is funded by the Austrian Science Fund (FWF). Thank you.</p><p>The authors declare no competing financial interest.</p>

PubMed Open Access

In Vivo Ambient Serotonin Measurements at Carbon-Fiber Microelectrodes

The mechanisms that control extracellular serotonin levels in vivo are not well-defined. This shortcoming makes it very challenging to diagnose and treat the many psychiatric disorders in which serotonin is implicated. Fast-scan cyclic voltammetry (FSCV) can measure rapid serotonin release and reuptake events but cannot report critically important ambient serotonin levels. In this Article, we use fast-scan controlled adsorption voltammetry (FSCAV), to measure serotonin\xe2\x80\x99s steady-state, extracellular chemistry. We characterize the \xe2\x80\x9cJackson\xe2\x80\x9d voltammetric waveform for FSCAV and show highly stable, selective, and sensitive ambient serotonin measurements in vitro. In vivo, we report basal serotonin levels in the CA2 region of the hippocampus as 64.9 \xc2\xb1 2.3 nM (n = 15 mice, weighted average \xc2\xb1 standard error). We electrochemically and pharmacologically verify the selectivity of the serotonin signal. Finally, we develop a statistical model that incorporates the uncertainty in in vivo measurements, in addition to electrode variability, to more critically analyze the time course of pharmacological data. Our novel method is a uniquely powerful analysis tool that can provide deeper insights into the mechanisms that control serotonin\xe2\x80\x99s extracellular levels.

in_vivo_ambient_serotonin_measurements_at_carbon-fiber_microelectrodes

<!>Solutions<!>Carbon-Fiber Microelectrodes<!>Data Collection<!>Flow Injection Analysis<!>Data Analysis<!>Statistical Analysis<!>Exclusion Criteria<!>Animal Surgeries<!>Serotonin FSCAV<!>Serotonin FSCAV is Stable, Selective, and Sensitive<!>In Vivo Serotonin FSCAV<!>CONCLUSIONS

<p>Dysfunctions of the serotonin system are thought to underlie numerous neuropsychiatric disorders, such as depression, anxiety, and schizophrenia.1–3 A better interpretation of serotonin neurochemistry is crucial for understanding the roles of this neurotransmitter but insight into serotonin's chemistry has been limited by the difficulty of in vivo chemical measurements. Serotonin is particularly challenging to detect electroanalytically in vivo because of an inauspicious combination of low extracellular concentrations and the propensity of serotonin and serotonin metabolites to foul electrodes.4</p><p>In 2009, we optimized fast-scan cyclic voltammetry (FSCV) for measurement of endogenous serotonin release and reuptake in vivo using carbon fiber microelectrodes (CFMs).5 Since then, we have uncovered various important aspects of serotonin neurochemistry. For example, evoked serotonin release is under much tighter regulation than dopamine (DA),6 being subject to prolonged autoreceptor control and multiple reuptake mechanisms.7 Furthermore, a single dose of a selective serotonin reuptake inhibitor (SSRI) rapidly alters serotonin neuro-chemistry;8 an important finding because a chronic SSRI regimen must often be followed for clinical therapy.2</p><p>While FSCV continues to provide important insights into the mechanisms that regulate extracellular in vivo serotonin, the method has limitations. In particular, because FSCV is background-subtracted, it only reports information about concentration changes. These changes allow us to probe serotonin release and reuptake, however it would be invaluable to also determine serotonin's ambient (steady-state, extracellular or basal) concentrations.</p><p>There are very few methods that can quantitatively measure extracellular serotonin concentrations. By far, the most commonly used method is microdialysis which reports extracellular serotonin levels in the lower nanomolar range.9–13 We thus sought to design a reliable and accurate method for measuring serotonin's ambient concentrations with our carbon fibers that we could couple to our FSCV measurements.</p><p>We recently made basal measurements of dopamine with fast-scan controlled-adsorption voltammetry (FSCAV), which exploits the adsorption capabilities of activated CFMs. The method is robust, selective, fast, and sensitive with the additional allure that CFMs measure from discrete brain localities14 where tissue damage is minimized.15 Our studies revealed a fundamental coaction between evoked and ambient dopamine.16 Here, we orient this method toward serotonin.</p><p>We find that unique FSCAV serotonin signals (that resemble FSCV responses) are stable during repeated recordings over 120 min in vitro. We confirm that the majority of interfering analytes are unlikely to contribute to the serotonin FSCAV signal, bar serotonin's major metabolite, 5-hydroxyindoleacetic acid (5-HIAA) which gives a small signal at high concentrations. Subsequently we show that 5-HIAA is unlikely to contribute to the signal when at physiological concentrations. Next, our method is utilized to report ambient, in vivo serotonin levels in the mouse hippocampus as 64.9 ± 2.3 nM (n = 15 mice, weighted average ± standard error) using a functional linear statistical model that we developed based on electrode and in vivo variability. Using a pharmacological approach, the in vivo signal is validated and interferences by 5-HIAA, dopamine, and norepinephrine17–21 are disqualified. Finally, we take a closer look at the time course of the pharmacological data using a chemometric approach incorporating the uncertainty inherent to in vivo recordings, as well as the variability between CFMs to show that the drug response is significant earlier than shown by conventional statistical tests.</p><p>Accurate measurements of ambient serotonin are essential for establishing serotonin's physiological impact, and here, we present an ideal tool and analyses for this measurement. Furthermore, FSCV and FSCAV can be combined at a single sensor, providing the distinctively powerful analytical capability of measuring both phasic and ambient serotonin.</p><!><p>Dopamine hydrochloride, serotonin hydrochloride, 5-hydroxyindole acetic acid, ascorbic acid, 3,4-dihydroxyphenylacetic acid, uric acid, norepinephrine hydrochloride, hydrogen peroxide, adenosine, histamine hydrochloride, pargyline hydrochloride, desipramine hydrochloride, and GBR 12909 were purchased from Sigma-Aldrich (St. Louis, MO). Liquion (LQ-1105, 5% by weight Nafion) was purchased from Ion Power Solutions (New Castle, DE). Buffer solution was composed of 15 mM Tris, 126 mM NaCl, 2.5 mM KCl, 25 mM NaHCO3, 2.0 mM NaH2PO4, 1.2 mM Na2SO4, 1.2 mM CaCl2, and 2.0 mM MgCl2 (all purchased from Sigma-Aldrich (St. Louis, MO)).</p><!><p>The carbon-fiber micro-electrodes were assembled by aspirating a single T-650 carbon fiber (7 μm, Goodfellow, Coraopolis, PA) into cylindrical glass capillaries (internal diameter: 0.4 mm, external diameter: 0.6 mm, A-M Systems, Carlsborg, WA). The carbon filled capillaries were positioned vertically in a pipet puller (Narishige Group, Setagaya-Ku, Tokyo, Japan) to form a carbon–glass seal under gravity. The carbon fibers were then cut to approximately 150 μm in length. Nafion solution (Liquion-1105-MeOH, Ion Power, DE) was electrodeposited on the exposed carbon fibers as previously described.5 The microelectrodes were dried at 70 °C for 10 min.</p><!><p>FSCV and FSCAV were performed using software (WCCV 3.05) and instrumentation developed by Knowmad Technologies LLC (Tucson, AZ). FSCAV was performed using a CMOS precision analog switch, ADG419 (Analog Devices) to control the application of the computer-generated waveform to the electrode. The logic was controlled programmatically and either a series of ramps (0.2–1.0 V to −0.1–0.2 V, scan rate = 1000 V/s) was applied every 10 ms (100 Hz), or a constant potential (0.2 V) was applied to the electrode for a specified period (10 s) (controlled adsorption period).</p><!><p>FSCV in vitro analyses were performed using flow injection analysis (FIA). CFMs were placed into a flangeless short 1/8 nut (PEEK P-335, IDEX, Middleboro, MA) with 2 mm of the tip exposed from the nut. The microelectrode-containing nut was fastened into a modified HPLC union (Elbow PEEK 3432, IDEX, Middleboro, MA). The other end of the elbow union was secured into the out-flowing stream of the FIA buffer. Two holes were drilled into the union to incorporate a reference electrode and a "waste" flow stream. The flow was maintained using a syringe infusion pump (kd Scientific, model KDS-410, Holliston, MA) at a rate of 2 mL min−1. Serotonin was introduced using a rectangular pulse into the flow stream for 10 s through a six-port HPLC loop injector (Rheodyne model 7010 valve, VICI, Houston, TX).</p><!><p>FSCV and FSCAV signals were processed using software written in-house using LabVIEW 2009. The processing includes filtering, smoothing, and signal deconvolution. For FSCAV, the cyclic voltammogram (CV) of the third scan (after controlled adsorption period) was extracted and the peak that occurred between approximately between 0.4 V to 0.85 V was integrated. The resulting charge value in pC was plotted vs serotonin concentration to create calibration curves which were then utilized to report in vivo values.</p><!><p>On the basis of the calibration data, linear models relating charge with both concentration and different electrodes were developed. These linear models incorporated interactions between the concentration and the electrode to accommodate for the inherently varying responses between electrodes.22,23 These models allow different intercept and slope in the linear relationship between charge and concentration for each electrode. The fitting was performed using linear model via the lm command in the R programming package. The results show significant differences in the intercept and slope for each electrode (Figure S-1). Using these fitted linear models, and given charge measurements collected in vivo at successive time points, estimates of the concentration levels at each time point were obtained. This was done by inverting' the fitted linear relationship between concentration and charge, and a weighted pooling of the concentration estimates from each of the electrodes was performed to obtain an overall concentration level estimate at each time point. The weights are based on the inverse of the estimated variance of the concentration estimates. Point-wise confidence intervals (CI) were constructed by fitting the functional model to the time and concentration values. These 95% point-wise confidence intervals were constructed when the functional model was fitted to the pairs of time and concentration values using the predict.lm command in the R package.</p><!><p>FSCV was performed before FSCAV collection in vivo to verify the presence of serotonin. CVs collected during an evoked response in mice were compared with previously well-established signals.5 Mice in which the CVs did not match the characteristics of a serotonin CV were excluded from this study. Furthermore, mice that died before the end of the collection time were excluded. All other mice were included in this study.</p><!><p>Six-week-old male C57BL/6J mice, 20–25 g, were purchased from Jackson Laboratories (Bar Harbor, ME). The mice were housed in 12-h light/dark cycles and were offered food and water ad libitum. Animal care and procedures were in agreement with the Guide for the Care and Use of Laboratory Animals, accepted by the Institutional Animal Care and Use Committees (IACUC) of the University of South Carolina. After an intraperitoneal (i.p.) injection of the anesthetic urethane, (25% dissolved in 0.9% NaCl solution, Hospira, Lake Forest, IL) at a volume of 7 μL per 1 g mouse weight, stereotaxic surgeries (David Kopf Instruments, Tujunga, CA) were performed. A heating pad (Braintree Scientific) was used to maintain ideal mouse body temperature of 37 °C. Bregma was used as a reference for stereotaxic coordinates of Medial Forebrain Bundle (MFB) [AP, −1.58; ML, +1.0; DV, −4.8 to −5.0] and CA2 [AP, −2.9; ML, +3.35; DV, −2.5 to −3.0] from Franklin and Paxinos (2008). To access the CA2 and MFB, holes were drilled in accordance to the stereotaxic coordinates. A stainless steel electrode (diameter = 0.2 mm; Plastics One, Roanoke, VA) was implanted into the MFB for stimulation. The nafion coated CFM was then lowered into the CA2. A silver wire (diameter: 0.010 in A-M Systems, Sequim, WA), electroplated with chloride by immersion of the wire in hydrochloric acid (0.1 M, 4 V vs tungsten), was implanted into the opposite hemisphere of the CA2 electrode placement. A 60 Hz biphasic 350 μA, 120 pulse stimulation, 2 ms per phase was employed through linear constant current stimulus isolator (NL800A Neurolg; Digitimer Ltd.). All drugs were administered i.p.; Pargyline at a dose of 75 mg kg−1 and GBR 12909 at a dose of 15 mg kg−1, both dissolved in 90% saline and injected at a volume of 0.1 mL 20 g−1.</p><!><p>A robust analytical measurement of ambient serotonin would lend nuance to our understanding of this complex neurotransmitter. Researchers have conventionally relied on microdialysis for basal measurements; however, it is greatly desirable to us to measure this ambient concentration at CFMs. Because FSCV relies on background-subtraction to remove a large charging current, FSCV could not, until recently, report basal neurotransmitter concentrations. We recently described a novel modification to the FSCV technique that allowed us to measure ambient in vivo DA levels, which we coined FSCAV.16 FSCAV is similar in concept to adsorptive stripping voltammetry whereby analytes adsorb onto the electrode surface for a controlled period of time before electrochemical characterization. We utilize the terminology "ambient" to denote a measurement made at a temporal resolution (20 s) that is neither on the same scale as FSCV (milliseconds) or microdialysis (minutes). The temporal scale of FSCV allows it to measure phasic changes whereas microdialysis can measure tonic or basal changes in the brain. Conversely, FSCAV is an average of both. Here, FSCAV was applied to serotonin analysis.</p><p>Electrochemical measurements of serotonin are fundamentally challenging because of serotonins' and serotonin metabolites' detrimental effects on the electrode surface. In 1995, Jackson et al. developed a solution for electrode fouling for serotonin measurements by developing a waveform, at a very high scan rate, to "outrun" fouling reactions.4 The "Jackson waveform" was later combined with an electrodeposited Nafion coating on the CFM for in vivo serotonin FSCV.5 We therefore applied the Jackson waveform (0.2 to 1.0 to −0.1 to 0.2 V, 1000 V s−1) to a Nafion coated CFM for in vivo serotonin FSCAV (100 Hz) with a 10 s controlled adsorption period.</p><p>Figure 1A(i) shows a color plot of 100 nM serotonin collected in vitro with FSCV using the Jackson Waveform at 10 Hz. Figure 1A(ii) shows a color plot of 100 nM serotonin in vitro with FSCAV using the same waveform at 100 Hz. The black area in Figure 1A(ii) is the controlled adsorption period, and the CV taken at the third scan (denoted by star) after waveform reapplication is shown in Figure 1B. Importantly, this CV contains peaks characteristic of serotonin's redox potentials.4,5 This CV is superimposed onto a CV taken from an FSCV color plot (denoted by star). The orange lines illustrate the integration limits used for FSCAV analysis (see Experimental Section). The CVs show good agreement, with the exception of a slight potential shift in the FSCAV signal which we attribute to the higher waveform application frequency (100 Hz vs 10 Hz for FSCV).24</p><!><p>We characterized this method's stability, selectivity, and sensitivity in vitro. First, in Figure 2 FSCAV files were collected in a 100 nM serotonin solution every 30 s over 120 min. An ANOVA test performed under linear models on this response shows that the measurement is highly reproducible (p = 0.95), affirming the method's stability in vitro.</p><p>Second, we identified 9 electroactive species in the hippocampus that could potentially interfere with the FSCAV signal.18,25–32 Figure 3 shows CVs collected in vitro from these different species at concentrations that mimic a range of reported or predicted physiological values.4,5,16,24,29,33–39</p><p>We applied the integration limits for serotonin analysis (+0.4 to +0.85 V) and analyzed these CVs. Histamine (HA) (1 μM), adenosine (1 μM), DOPAC (2 μM), norepinephrine (NE) (1 μM), uric acid (UA) (1 μM), DA (100 nM), ascorbic acid (AA) (200 μM), and hydrogen peroxide (H2O2) (1 mM) showed no significant features within the serotonin integration limits (n = 4). 5-HIAA (10 μM)5 could prove problematic due to the presence of a peak within the integral limits.</p><p>We show that 5-HIAA interference is unlikely in Figure 4. In this figure, charge is plotted against concentration for serotonin (orange). Here, the linear portion of the serotonin calibration is shown with orange markers with the following linear regression: (1)y=0.0207(±0.0005)x+1.51(±0.14),R2=0.997</p><p>The serotonin plot shows linearity up to 600 nM, with a sensitivity of 0.021 ± 0.0005 pC nM−1 (n = 4 ± SEM), and a limit of quantification of 5 nM. When both serotonin and 5-HIAA are present in solution, we postulate that there is a competition for adsorption sites on the carbon fiber surface. The rationale here is that the analyte with higher adsorption equilibrium constant (Kads) on the CFM will exhibit a more favorable thermodynamic adsorption profile and will thus outcompete the other.40 A Langmuir monolayer adsorption isotherm model was used with FSCAV data to calculate Kads for serotonin and 5-HIAA. Kads for serotonin and 5-HIAA were 9.57 × 1010 and 7.02 × 108, respectively. The much higher Kads for serotonin adsorption onto CFMs means that 5-HIAA added to serotonin does not affect the signal as shown in the inset in Figure 4. The green stars signify a separate data set where approximately 100× more concentrated 5-HIAA was added to the serotonin solution and the blue markers show the signal. The close agreement of the blue and orange markers make it clear that 5-HIAA, at physiological concentrations (typically 10 μM),41 would not impact the signal. In addition to favorable adsorption, the much improved sensitivity for serotonin vs 5-HIAA can be credited to Nafion on the CFM.5</p><!><p>To apply FSCAV to in vivo serotonin measurements, we first employed a retrograde stimulation of the medial forebrain bundle (MFB) and confirmed electrically stimulated serotonin release in the CA2 region of the mouse hippocampus (example of evoked release can be seen in Figures 6 and 7). Subsequently FSCAV was performed at the same electrode. Figure 5A shows the in vivo FSCAV color plot (i) adjacent to a color plot of 100 nM serotonin in vitro (ii).</p><p>The close agreement of the CVs shown in Figure 5B is strong evidence that this signal is serotonin. We took a chemometric approach to report the concentration in vivo with uncertainty that incorporated not only in vivo variability but also the variability of individual electrodes. These variations occur mainly due to the nonuniformity between carbon fiber surfaces that arise as a result of the fabrication process. A linear functional model was developed using the calibration data of the different electrodes used in the in vivo experiments. In 15 mice, the weighted average extracellular serotonin level was 64.9 ± 2.3 nM (n = 15 mice, weighted average ± standard error) (see inset in Figure 5B).</p><p>Previous reports of ambient serotonin with microdialysis in different brain regions have estimated extracellular serotonin in low nanomolar to 10s of nanomolar.9–13 Our value is slightly above this range. Our method is performed on a fundamentally different spatial scale. For example, commercial microdialysis probes typically have a diameter of 200 μm and are 2 mm in length, whereas CFMs are 7 μm in diameter and 150 μm in length. The tissue volume impacted by a CFM is orders of magnitude smaller than that of a typical microdialysis probe15 and because we optimize the electrode's placement based on stimulated serotonin release, the electrode is in a "hot spot" (area of high serotonin activity), accounting for slightly higher levels.</p><p>Above, we assessed FSCAV's selectivity in vitro, however, the in vivo matrix is far more complicated than can be reproduced on the bench. It is therefore critical to verify the signal pharmacologically in vivo. For this task, we employed pargyline (75 mg kg−1), an irreversible monoamine oxidase B (MAO-B) inhibitor. By inhibiting MAO-B, and hence the metabolism of serotonin in the brain, an increase in serotonin and a decrease in 5-HIAA is expected.42,43 Figure 6 shows experiments that test the effects of pargyline on the FSCAV signal. First, FSCV was used to optimize the position of the CFM by evoking serotonin release (a representative example color plot is shown inset on the top left). Individual animal FSCAV responses (faint blue makers) and the averaged response (dark blue dots) 60 min before and 60 min after an i.p. injection of pargyline are shown on the central trace. Using conventional statistical analysis, pargyline administration caused a significant increase in the FSCAV signal at 29 min and thereafter (two way repeated measures ANOVA: p < 0.0001, n = 5 mice ± SEM with Dunnett's multiple comparison post hoc, p < 0.01, n = 5 mice ± SEM). The effects of pargyline were verified with FSCV following FSCAV data collection (inset top right color plot is a representative color plot and [serotonin] vs time traces (α = predrug and β = postdrug)), where pargyline increased evoked serotonin amplitude and reduced the rate of reuptake as previously seen in rats.6 This experiment eliminates the possibility of 5-HIAA interference and would verify our signal as serotonin, save for one final concern; that DA and NE are also substrates for MAO-B44,45 and are present at appreciable levels in the CA2 region of the hippocampus.17,20,21</p><p>Very little sensitivity was established in vitro for DA or NE. To further verify no interference from DA we administered GBR 12909, a potent DA transporter inhibitor, to a separate set of mice. We have previously shown that GBR 12909 causes an increase in ambient DA,16 but not in evoked serotonin levels.5 The faint, red markers in Figure 6 show individual FSCAV animal responses to 15 mg kg−1 GBR 12909, while the dark red dots show the averaged responses 60 min before and after i.p. administration (n = 5 mice ± SEM). As above, FSCV was used to assess the effects of this manipulation on the evoked serotonin response. The lack of an increase in the FSCAV signal (two way repeated measures ANOVA, p > 0.05, n = 5 mice ± SEM with Dunnett's multiple comparison post hoc, p > 0.05, n = 5 mice ± SEM) and FSCV signal allows us to exclude DA as interference.</p><p>To eliminate the possibility of interference from NE, we administered desipramine (15 mg kg−1) to a separate set of mice. Desipramine is a norepinephrine transporter (NET) inhibitor that selectively blocks NETs but has negligible effect on DA or 5HT transporters.46,47 The faint green, markers in Figure 7 represent the individual FSCAV responses, while the dark green dots represent the averaged response 60 min before and after i.p. drug administration (n = 5 mice ± SEM). Injection is immediately before first data point at 0 min (yellow bar). FSCV color plots and CVs taken before and after FSCAV data collection from a representative experiment are inset. There was no change in the FSCAV signal (two way repeated measures ANOVA, p > 0.05, n = 5 mice ± SEM with Dunnett's multiple comparison post hoc, p > 0.05, n = 5 mice ± SEM). There was no increase in the release amplitude as measured with FSCV, however there was a dramatic dip below baseline after stimulation. We previously showed that dips such as this were mediated by prolonged autoreceptor activation7 and because desipramine has agonist activity at the 5H1B receptor,48,49 it is likely we are observing a potentiation of the autoreceptor effect. This experiment allows us to exclude norepinephrine as a possible interference.</p><p>Conventional statistical tests (e.g, two way repeated measures ANOVA with Dunnett post hoc test) show that the serotonin levels increase significantly 29 min after pargyline administration. However, visually it is seen that the serotonin levels begin to rise much earlier than that. To address this, we expanded the fitted linear model to encompass the concentration values prior to and post pargyline, GBR 12909, and desipramine administration. The equations of the model were as follows</p><p>serotonin/pargyline</p><p> (2)C(t)=65.217+0.0041×t+0.5268×[max(0,62.6-t)]-0.0042×[max(0,62.6-t)]2 serotonin/GBR12909</p><p> (3)C(t)=58.82+0.0055×t-0.000088×t2 serotonin/desipramine</p><p> (4)C(t)=C(t)=71.54+0.0013×t-0.000004×t2 where C(t) is change in concentration with time, t is time and max(a, b) is the larger value between a and b. The fitted model (blue line) over the averaged serotonin data (black dots) with pargyline, GBR 12909, and desipramine administration is seen in Figure 8A–C, respectively.</p><p>For pargyline, the functional continuous model consisted of a linear part over the time portion where no drugs were administered plus a time lag Delta (0 to 60 + Delta), and is parabolic over the time interval from 60 + Delta to 120 (Figure 8A). Using this model, the estimate for this time lag Delta or simply the time point where the drug causes a change in the slope is 2.60 min. This estimate is obtained by maximizing the coefficient of determination (R2) with respect to the possible values of Delta. Note that the final fitted model has a high R2 equal to about 95%, indicating an excellent fit of the linear-parabolic model for relating concentration to time for this serotonin study.</p><p>For the GBR12909 and desipramine, the model showed no effect of time on serotonin concentration and there was basically no change over the whole period of study. The plot of this fitted model is presented in Figure 8B and 8C, respectively, which is almost flat, together with the estimated concentration levels (the solid circles) at each of the time points, and the 95% point-wise confidence intervals.</p><p>It is important to note that using repeated measures ANOVA with Dunnett test, the only information that was available is the time point at which pargyline caused a change that was statistically significant (p < 0.05, 29 min). On the other hand, through employing a statistical model that was built to take into account electrode variability, we were able to determine the point at which pargyline changed the ambient serotonin concentrations as soon as 2.60 min. This may be a more accurate reflection of the pharmacological profile of this agent.</p><p>The combination of electrochemical and pharmacological characterizations performed in vivo and in vitro allows us to confidently assert that FSCAV is able to selectively measure ambient serotonin in vivo. The synergy with a chemometric approach introduces a new wealth of information that allows for more accurate electrode calibrations and a more comprehensive understanding of the time course of in vivo data.</p><!><p>Imbalances in serotonin neurochemistry are important to study in the context of neuropsychiatric disorders. While FSCV can provide real-time chemical information, the method reports only phasic changes. Ambient serotonin levels are critical to establishing the fundamental extracellular mechanisms that control serotonin. Here, we reported FSCAV for ambient serotonin measurements. We performed a characterization of the FSCAV waveform for sensitive and selective serotonin measurements. In vivo, we utilized the waveform to report a basal serotonin level in mouse CA2 as 64.9 ± 2.3 nM (n = 15 mice, weighted average ± standard error). We pharmacologically verified the in vivo signal against perceived interferences. Finally, we developed a statistical model to further analyze the FSCAV readings and report the uncertainty caused by measuring in vivo using different CFMs. Serotonin FSCAV yields information about serotonin's basal behavior in vivo and, when coupled with FSCV at a single CFM, will provide a deeper chemical insight into serotonin's mechanisms in the brain.</p>

PubMed Author Manuscript

Enantioselective Addition of Pyrazoles to Dienes

We report the first enantioselective addition of pyrazoles to 1,3-dienes. Secondary and tertiary allylic pyrazoles can be generated with excellent regioselectivity. Mechanistic studies support a pathway distinct from previous hydroaminations: a Pd(0)-catalyzed ligand-toligand hydrogen transfer (LLHT). This hydroamination tolerates a range of functional groups and provides a breakthrough in hydrofunctionalization of dienes.

enantioselective_addition_of_pyrazoles_to_dienes

<!>Table 3. Hydroamination of Various 1,3-Dienes with Pyrazole (1a)

<p>Nitrogen-containing heterocycles, such as pyrazoles, represent valuable scaffolds for drug discovery and thus remain an inspiration for synthetic methods (Figure 1A). 1 The direct addition of a pyrazole to a double bond represents an attractive and atom-economical approach for forging C−N bonds. With regards to the coupling partner, conjugated dienes are ideal building blocks, 2 with many being raw materials for various industrial applications, including polymerizations. 3,4 Within the asymmetric hydroamination of dienes, there have been breakthroughs using anilines (Hartwig), 5 secondary amines (our lab and Malcolmson), 6,7 and primary amines (Mazet). 8 In comparison to previously studied amines (with nucleophilicities N = 13-18 on Mayr scale 9 ), pyrazoles present a challenge and opportunity because of their lower nucleophilicity (N = 9.6). Given the two reactive nitrogen atoms: a pyrrolic and a pyridinic nitrogen, the coupling of pyrazoles with unsymmetrical dienes can provide 32 isomers (Figure 1B). With both Rh and Pd-catalysts, Breit achieved enantioselective hydroamination of allenes using pyrazoles (Figure 1C). 10 Concurrent with our studies, Chen and coworkers were independently pursuing the Pd-catalyzed hydroamination of isoprene. With indazoles and select pyrazoles, they were able to generate either achiral or racemic products. 11 In this communication, we showcase the first asymmetric addition of pyrazoles to dienes. This mild hydroamination tolerates a variety of functional groups and occurs via a mechanism distinct to those previously proposed for diene hydroamination. On the basis of our previous hydroaminations of dienes, we initiated investigations with Rh-catalysts. 6,12 We chose pyrazole (1a) and 1-phenyl-butadiene (2a) as the model substrates and observed no desired reactivity (see SI). In contrast, under Pdcatalysis, the desired allylated pyrazole 3aa was obtained when using a range of achiral bisphosphine ligands (see SI). In search of an asymmetric variant (L1-L8), we found that atropoisomeric bisphosphine ligands gave the most promising results (Table 1). Thus, we focused on this ligand family to achieve enantioselective catalysis. The DTBM analogs L5-L7 afford desired pyrazole 3aa in 70-82% yield and good to excellent selectivity (>20:1 rr, 90:10-95:5 er). Substitution on the aryl groups most likely enhances reactivity by promoting ligand-substrate dispersion interactions in the transition state, a concept in accordance with literature observations by Buchwald and others. 13 We observed similar ligand trends in Rh-catalyzed hydrofunctionalizations of alkynes. 14 With further optimization, we found commercially available MeO-BIPHEP ligand L8 afforded 3aa in 91% yield with >20:1 rr and 96:4 er. With L8 in hand, we examined the addition of various pyrazoles 1 to diene 2a (Table 2). Generally, allylated pyrazoles 3ab-3aj form with high enantioselectivity (88:12-97:3 er) and >20:1 rr in the coupling of diene 2a with pyrazoles 1b-1j. The electronic properties of the pyrazoles show negligible impact on enantioselectivity and regioselectivity. However, electronwithdrawing substituents show more sluggish reactivity and require extended reaction times (3ad-3af, 3aj) or higher temperature (3ag, 3ah) to obtain good yields (64-86%). Halogenated products (3ad, 3ae) are tolerated despite the potential for competing oxidative addition into the aryl halide bond; no side products from oxidative addition are observed. Electrondonating substituents allow for facile reactivity and shorter reaction times (3ai, 69%, 89:11 er). Other unsymmetrical pyrazoles provide moderate to excellent N 1 :N 2 regioselectivity (3ak-3am, 11:1-20:1 rr). The N 1 :N 2 regioselectivity favors allylation at the less sterically-congested nitrogen atom. Pyrazole tautomerization is known to occur in solution; for example, 5-Me pyrazole (1k) exists in a nearly 1:1 ratio of the 3-and 5-substituted pyrazoles on the basis of NMR spectroscopy. 16 Despite the presence of tautomers, we observe high selectivity for formation of 3ak, which indicates tautomerization occurs faster than C-N bond formation. Together, these results represent the first enantioselective hydroamination of 1,3-dienes with azoles. 17 In addition to pyrazole substrates, several other azoles show promising reactivity under the standard conditions. The addition of 1H-1,2,3-triazole (3an) gives 70% yield and 91:9 er with high regioselectivity (>20:1 rr, >20:1 N 2 :N 1 ). 18 The coupling of 1Hbenzotriazole and 1H-indazole with diene 2a provides the corresponding allylated azoles (3ao, 3ap) with promising reactivity and chemoselectivity, however, further optimization is needed. 19 In stark contrast, pyrrole showed no reactivity, a result that supports our hypothesis on the mechanism (vida infra).</p><p>Next, we studied the hydroamination of fourteen different 1,3dienes 2 with pyrazole 1a (Table 3). Varying substitution on the aryl dienes results in a range of chiral allylated pyrazoles 3ba-3ka (31-95% yield, >20:1 rr, 89:11-96:4 er). Hydroamination of both electron-rich methoxy-substituted (3ba, 3fa, and 3ha) and electron-poor fluoro-substituted (3ea) 1,3-dienes react well. In contrast, chloro-substituted phenyl diene 2d affords 3da in only 31% yield due to competing oxidative addition into the C-Cl bond. Sterically encumbered ortho-substituted dienes undergo addition to 3ha and 3ia in 40% and 62% yield, respectively. The protocol transforms heterocyclic substituted dienes 2j (R 1 = 2-furyl) and 2k (R 1 = 2-thiophenyl) into 3ja in 34% yield, 90:10 er, and >20:1 rr, and 3ka in 95% yield, 93:7 er, and >20:1 rr. Hydroamination of alkyl substituted 1,3-diene 2l yields the allylic pyrazole 3la in 45% yield, 95:5 er, and 2:1 rr. Cyclic dienes such as 1,3cyclohexadiene (2m) couple with pyrazole (1a) to generate 3ma in 71% yield and >20:1 rr, albeit with a lower enantioselectivity of 81:19 er. Hydroamination of feedstocks, isoprene and myrcene, provide the tertiary allylic amines (3na, 3oa) as single structural isomers.</p><!><p>Reaction conditions: 1a (0.1 mmol), 2 (0.5 mmol), [Pd( 3 -C3H5)Cl]2 (5 mol%), MeO-BIPHEP L8 (15 mol%), CPME (0.4 mL), 23 °C, 18 h. Isolated yields. Regioselectivity determined by 1 H NMR analysis of the unpurified reaction mixture. Enantioselectivity determined by chiral SFC.</p><p>Electronic circular dichroism (ECD) is a powerful technique to determine absolute configuration. 20 By using this method, we elucidated the absolute stereochemistry of the chiral allylated azoles. Comparing theoretical calculations and experimental data, a qualitative match (i.e., similar shapes) enabled assignment of the absolute configuration. 20 TDDFT calculations produced the ECD spectra of (S)-3aa and (S)-3ab. Qualitative comparison to the experimental results suggests that the major enantiomer bears the (S)-configuration as drawn (see SI).</p><p>Previous reports on the hydrofunctionalization of dienes, including Chen's hydroamination of isoprene, feature mechanisms that occur by Pd(II)-H catalysis. 4,11 In these scenarios, alkene insertion into a Pd-H forges the new carbonhydrogen bond and these transformations occur at elevated temperatures. On the basis of recent reports and our own observations, we propose that our ambient hydroamination occurs via the mechanism depicted in Figure 3. The palladium pre-catalyst interacts with a bisphosphine ligand to form active Pd(0) catalyst I. Both pyrazole 1 and 1,3-diene 2 bind to complex I to generate palladium intermediate II. Given Pd(II)'s preference to adopt a square planar geometry, we reason that diene coordinates to the Pd in an  2 fashion. 11,21,22 From here, we imagine that the hydrogen atom is transferred directly from pyrazole 1 to 1,3-diene 2 through ligand-to-ligand hydrogen transfer (LLHT). Ionization of intermediate III followed by outersphere nucleophilic attack with pyrazole anion on the C3 carbon affords the desired allylic pyrazole 3 and regenerates Pd(0) complex I. In support of a Pd(0) pathway, Huang 23 and Rutjes' 24 computations have shown that the formation of the Pd(II)-H complex from [Pd( 3 -C3H5)Cl]2 is kinetically infeasible at temperatures up to 80 C. In line with this, we do not observe Pd-H when studying a mixture of [Pd( 3 -C3H5)Cl]2 with ligand in d8toluene. Additionally, alternative Pd(0) precursors, including Pd(PPh3)4 and Pd(P t Bu3)2, afford the allylic pyrazole albeit in lower yields and selectivity (see SI). In these cases, there is no acid additive, which makes Pd-H unlikely. By using Burés' variable time normalization analysis (VTNA) method, 25 we studied the kinetic profile and observed first order in catalyst and zero order in both the pyrazole (1) and diene (2). This rate law supports coordination of diene and pyrazole to Pd to generate intermediate II as the catalyst resting state.</p><p>In Zi's study on hydrosulfonylation of dienes, theoretical calculations show that diene migratory insertion into Pd(II)-H is energetically unfavorable compared to LLHT. 21 In analogy, we propose an LLHT that is the turnover-limiting step (Figure 3). Comparing the initial rates of deuterated pyrazole d-1a against 1H-pyrazole 1a in parallel, we observe a KIE of 1.4 (Figure 4A). When we subjected deuterium-labeled pyrazole d-1a to the standard conditions (Figure 4B), we see quantitative deuterium incorporation at the C4 position of d-3aa; the recovered diene shows no deuterium labelling. Together, these results suggest that hydrogen transfer is highly selective and irreversible. Of note, in Malcolmson's hydroamination, the analogous experiment demonstrated deuterium scrambling. 7 Additionally, hydroamination of diene 1a with pyrrole shows no reactivity. In comparison to pyrazole, the pyrrole lacks a second nitrogen atom. We reason the second nitrogen coordinates to Pd to provide the geometry needed for LLHT. Next, we performed a crossover study by adding pyrazole 1b to (S)-3aa in the presence of the Pd-catalyst (Figure 4C, entry 1). The crossover product (S)-3ab was generated where the major isomer possessed the same absolute configuration as the (S)-3aa starting material. A similar crossover experiment using a racemic mixture of 3aa was performed (Figure 4C, entry 2). After 18 h, the crossover product (S)-3ab (60:40 er) is afforded along with an enantioenriched mixture of (R)-3aa (34:66 er). Enantioenrichment of the R-enantiomer suggests that (S)-3aa reacts faster, transforming into (S)-3ab in these crossover experiments. These experiments suggest that Pd insertion into the C-N bond can be reversible, especially under conditions where pyrazole is present in large excess (see SI). Based on these experiments, we favor a mechanism that involves outer sphere nucleophilic attack of pyrazole to complex III. In line with Tsuji-Trost transformations, the ionization of allylic pyrazole (S)-3 with Pd-catalyst would invert the configuration at the reactive center (Figure 4C). To afford the S-enantiomer of the crossover product, the nucleophile must attack through an SN2-like mechanism on the face of the olefin opposite to Pd. Our proposed mechanism fits with the convention of classifying nucleophilic attack on  3 -Pd-π-allyl intermediates. Pyrazole, which is considered a "soft" nucleophile (pKa~19.8), would be expected to proceed through this outer-sphere pathway. 26 In our and Malcolmson's independently reported Pdcatalyzed diene hydrofunctionalizations, a competition experiment was performed using a mixture of E and Z dienes. In their studies, both (Z)-and (E)-1-phenylbutadiene (2a) converged to the same major enantiomer. 7,27 Moreover, deuterium scrambling into the diene was observed. These results supported a Pd-H mechanism where hydropalladation is reversible. In this pyrazole study, however, we find when using a mixture of (Z)-and (E)-1-phenylbutadiene (2a) isomers, only the (E)-2a transforms to allylic pyrazole (3aa) (50% yield brsm), while the (Z)-2a is recovered (Figure 4D). These contrasting results point to a mechanism which differs from those previously invoked. Here, we reason that the (Z)-diene does not transform due to increased steric strain in the LLHT step.</p><p>Hydroamination represents an attractive way to transform dienes into nitrogen-containing building blocks. By using Pdcatalysis, we achieved the first enantioselective hydroamination of dienes with aromatic heterocycles. The allylation tolerates a broad range of substituted pyrazoles and dienes, and both secondary and tertiary allylated pyrazoles are obtained in good to excellent yields with high regio-and enantioselectivities. Insights from this study will guide the invention of future heterocyclic hydrofunctionalizations and pave a route to biologically-relevant azole-containing molecules.</p>

ChemRxiv

Synthesis, structural characterization, and optical properties of benzo[f]naphtho[2,3-b]phosphoindoles

Phosphole-fused π-conjugated acenes have been attracting interest because of the attractive features of the phosphole moiety, such as fluorescence and chemically modifiable properties. Herein, 6-phenyl-6H-benzo[f]naphtho[2,3-b]phosphoindole was prepared by reacting dichlorophenylphosphine with a dilithium intermediate derived from 3,3′-dibromo-2,2′-binaphthyl. Various derivatives, such as a phospholium salt and a borane–phosphole complex with functional groups on the phosphorus atom were synthesized using the obtained phosphole as a common starting material. Single-crystal X-ray analysis of the parent benzo[f]naphtho[2,3-b]phosphoindole revealed that the pentacyclic ring is almost planar. Fluorescence spectroscopy data showed that the phosphole derivatives, such as phosphine oxide and the phospholium salt and borane complex exhibited photoluminescence in chloroform.

synthesis,_structural_characterization,_and_optical_properties_of_benzo[f]naphtho[2,3-b]phosphoindol

Introduction<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Conclusion<!>

<p>Phosphole-based heteroacenes are attracting increasing interest in various fields, such as organic synthesis, structural chemistry, and materials science [1–4]. The phosphorus atom of trivalent phosphorus compounds has a high chemical reactivity. Therefore, this phosphorus center can be easily chemically modified and converted to phosphole derivatives with different electronic properties by reactions such as oxidation, alkylation, and coordination to a Lewis acid [1–8].</p><p>Theoretically, pentacyclic benzonaphthophosphindole contains six structural isomers, in which the position of the fused benzene rings is different; of these, three are shown in Figure 1. The synthesis, crystal structure, and dynamic behavior of benzo[e]naphtho[2,1-b]phosphindole (A) with the C2 symmetry axis on the binaphthyl skeleton have been reported [9–11]. Synthetic approaches for phosphine oxide B [12–14], alkylated products C [15], and transition metal complexes D [16–17] have also been developed. However, for isomers E [18–19] and G [20], only the synthetic method for pentavalent phosphine oxides has been reported. To the best of our knowledge, the synthesis and derivatization of trivalent phosphole F and the optical properties have not been clarified. The systematic knowledge of the properties of a new family of fused phospholes is valuable for the design of new types of functional π-electron-containing materials. This paper presents the synthesis, molecular structure, and optical properties of 6-phenyl-6H-benzo[f]naphtho[2,3-b]phosphoindole (F) and the derivatives in which the phosphorus atom is chemically modified, such as a phospholium salt and the borane–phosphine complex.</p><!><p>Benzonaphthophosphindoles.</p><!><p>Treatment of 3,3′-dibromo-2,2′-binaphthyl (1) [21] with n-butyllithium in dry THF at −78 °C and subsequently with dichlorophenylphosphine resulted in ring closure, affording the desired product containing 6-phenyl-6H-benzo[f]naphtho[2,3-b]phosphoindole (2) in 52% yield via the 3,3′-dilithio-2,2′-binaphthyl intermediate. The chemical modification of the phosphorus atom of 2 was carried out; the results are shown in Scheme 1. The reaction of 2 with hydrogen peroxide or elemental sulfur afforded the corresponding phosphine oxide 3 and sulfide 4 in 92% and 88% yield, respectively. Treatment of 2 with methyl triflate afforded phospholium triflate 5 in 81% yield. The reaction of 2 with borane in THF formed a borane complex 6 in 91% yield, and the reaction with chloro(dimethyl sulfide)gold afforded the gold complex 7 in 39% yield.</p><!><p>Synthesis of benzo[f]naphtho[2,3-b]phosphoindoles.</p><!><p>The molecular structures of the synthesized phospholes 2−7 were confirmed by spectral analyses (1H, 13C, and 31P NMR as well as MS and IR). All the corresponding aromatic proton and carbon atoms on the two naphthalene rings were equivalent in the 1H and 13C NMR spectra of phospholes. These results show that all phosphole derivatives had a symmetric structure in solution. The 31P NMR spectra of these show the typical low-field shift for P-modified phospholes 3–7 (δ = 22.5−39.3 ppm) relative to that of the parent compound 2 (δ = −13.27 ppm). These results suggest that the electron densities of the latter were reduced in comparison to that of 2. Single crystals of 2 suitable for X-ray analysis were obtained by repeated recrystallization. The molecular structure of 2, determined through single-crystal X-ray diffraction analysis, is illustrated in Figure 2, and selected geometrical parameters are shown in Table 1. The results revealed that the naphthalene and fused phosphole rings are almost coplanar (mean deviation = 0.030 Å). The angle between each naphthalene ring containing ten carbon atoms is 1.64°, which is smaller than that for group 15 analogs (i.e., N-phenyldibenzo[b,h]carbazole: 4.47° or 2.57° [22], for crystal data, see Figure S2, Supporting Information File 1 and Sb-phenyldinaphtho[2,3-b:2′,3′-d]stibole: 8.05° [23]). The sum of the bond angles around the phosphorus atom is 295.99°, and hence the phosphorus atom is sp3-hybridized and has a trigonal pyramidal geometry. X-ray analysis revealed that the packing structure of 2 had π–π-stacking, with a distance of approximately 3.427 Å between two benzonaphthophosphoindole planes (Figure 2b).</p><!><p>Crystal structure of 2: different views.</p><p>Selected bond lengths and angles.</p><!><p>The photophysical properties of the synthesized phospholes were evaluated. The corresponding data are shown in Figure 3 and Table 2. Parent compound 2 shows the absorption maximum (λmax) at 362 nm, which is longer than that of 2,2′-binaphthyl (λmax = 300 nm) [24]. The introduction of a phosphorus atom to 2,2′-binaphthyl results in the formation of a bridge that fixes two naphthalene rings in a coplanar axis and forms a heterole-fused system. The λmax values of derivatives 3–7 are similar to each other and have a shorter absorption band than that of parent 2. Furthermore, it is known that λmax of [n]helicenes generally has a shorter wavelength than for the corresponding linear poly(acene)s [25]. The λmax of benzo[e]naphtho[2,1-b]phosphindole (A), 358 nm [16], is the same as that of parent 2. In contrast, the corresponding oxide B has a λmax wavelength that is approximately 50 nm longer than that of our oxide 3 [16]. The fluorescence wavelength, including the maximum emission (λem), and the quantum yield depend on the nature of the P-modification. Phosphine oxide 3, cation 5, and boron complex 6 emitted blue fluorescence in the visible-light region, with λem at 395–426 nm (Table 2). P-methylated cation 5 exhibited the longest wavelength and the highest quantum yield.</p><!><p>a) Absorption spectra and b) normalized fluorescence spectra for selected compounds in CHCl3.</p><p>Absorption and emissions spectroscopy data.a</p><p>aMeasured in CHCl3. bExcited at 335 nm. cMeasured using anthracene as a standard.</p><!><p>The electrochemical properties of benzonaphthophosphoindoles were investigated by using cyclic voltammetry, and the electrochemical data are summarized in Table 3 and Figure S4, Supporting Information File 1. Parent compound 2 and borane complex 6 showed reversible reduction peaks (Ered = −1.25 and −1.26 V, respectively). Due to the increased electron–acceptor character of the phosphorus center, P-modification compounds 4 and 5 show a more positive oxidation potential than parent 2 [26]. Unfortunately, the electrochemistry of some compounds could not be determined under the conditions available to us.</p><!><p>Cyclic voltammetric dataa (E vs Ag/AgCl) and calculated HOMO and LUMO levelsb of phospholes.</p><p>aMeasured in DCB with TBAP. bDFT calculation at the level of B3LYP/LanL2DZ. cIrreversible. dReversible. eCation part only.</p><!><p>Therefore, computational investigations are particularly useful for understanding the trends of the electrochemical and photophysical properties of molecular materials. Density functional theory (DFT) calculations [27] were performed at the B3LYP/LanL2DZ level of theory to gain additional understanding of the electronic structures. The HOMO and LUMO energies of the selected compounds are listed in Table 3. The DFT calculations showed that these compounds have HOMO–LUMO gaps of 3.60–3.96 eV. For parent compound 2 and derivatives 3, 5, and 6, the HOMO and LUMO correspond to the π and π* orbitals of the benzonaphthophosphoindole skeleton, respectively (Figure S5, Supporting Information File 1). In contrast, the corresponding π orbitals of phosphole sulfide 4 comprise HOMO−2 energy levels (Figure 4). The HOMO and HOMO−1 levels of oxide 3 are delocalized in the π orbitals of the benzonaphthophosphoindole skeleton, while HOMO−2 and HOMO−3 have a large contribution from the lone-pair and π orbitals. In sulfide 4, the sulfur analog of the oxide 3, the HOMO and HOMO−1 energy levels are delocalized in the lone-pair orbitals on the sulfur atom, resulting in considerable destabilization and significantly small HOMO–LUMO gap compared to those of other phospholes 2, 3, 5, and 6. This result is similar to the reduction potentials of compounds 2 and 6. According to time-dependent DFT calculations for 4, the S0 to S1 transitions are mainly dominated by the dipole-forbidden lone pair–π* HOMO–LUMO transitions. This phenomenon may be responsible for the nonfluorescence of 4. Both the π and π* energy levels in the all-functionalized phosphole derivatives 3–6 are lower than those of parent phosphole 2, owing to the increased electron deficiency of the phosphorus center in the former. In particular, the energy levels in cationic phospholium 5 are significantly stabilized because of the cationic nature of the phosphorus center. These results are consistent with the 31P NMR observations discussed above.</p><!><p>The spatial plots of the HOMO−3 to LUMO of compounds 3 and 4. The calculations were performed at the level of B3LYP/LanL2DZ.</p><!><p>In summary, a novel pentacyclic phosphole, 6-phenyl-6H-benzo[f]naphtho[2,3-b]phosphoindole, was prepared by performing the ring-closure reaction of 3,3′-dilithio-2,2′-binaphthyl with dichlorophenylphosphine. The obtained product was used as a key compound for the chemical modification of the phosphorus atom. X-ray crystal analysis showed that the parent trivalent phosphole has a considerably planar benzonaphthophosphoindole skeleton in its crystal structure. 1H and 13C NMR observations revealed that all the phospholes obtained in this study had a highly symmetric structure in solution. Fluorescence spectroscopy data showed that the phosphole derivatives, such as a phosphine oxide, phospholium salt, and borane complex, exhibited photoluminescence in chloroform. The π and π* levels in all-functionalized phosphole derivatives are lower than those of the parent phosphole, owing to the increased electron deficiency of the phosphorus center in the former. This electronic nature is supported by the low-field shift of the P-modified derivatives relative to parent phosphole 2 in 31P NMR. Further investigation of the design, synthesis, and theoretical and spectroscopic studies of new functional π-electron materials for organic electronics applications is under progress, and the results will be reported in due time.</p><!><p>Further analytical and experimental data.</p><p>X-ray crystal structure of 2.</p><p>X-ray crystal structure of N-phenyldibenzocarbazole.</p>

PubMed Open Access

Environmental effects on soil microbial nitrogen use efficiency are controlled by allocation of organic nitrogen to microbial growth and regulate gross N mineralization

Microbial nitrogen use efficiency (NUE) is the efficiency by which microbes allocate organic N acquired to biomass formation relative to the N in excess of microbial demand released through N mineralization. Microbial NUE thus is critical to estimate the capacity of soil microbes to retain N in soils and thereby affects inorganic N availability to plants and ecosystem N losses. However, how soil temperature and soil moisture/O2 affect microbial NUE to date is not clear. Therefore, two independent incubation experiments were conducted with soils from three land uses (cropland, grassland and forest) on two bedrocks (silicate and limestone). Soils were exposed to 5, 15 and 25 \xc2\xb0C overnight at 60% water holding capacity (WHC) or acclimated to 30 and 60% WHC at 21% O2 and to 90% WHC at 1% O2 over one week at 20 \xc2\xb0C. Microbial NUE was measured as microbial growth over microbial organic N uptake (the sum of growth N demand and gross N mineralization). Microbial NUE responded positively to temperature increases with Q10 values ranging from 1.30 \xc2\xb1 0.11 to 2.48 \xc2\xb1 0.67. This was due to exponentially increasing microbial growth rates with incubation temperature while gross N mineralization rates were relatively insensitive to temperature increases (Q10 values 0.66 \xc2\xb1 0.30 to 1.63 \xc2\xb1 0.15). Under oxic conditions (21% O2), microbial NUE as well as gross N mineralization were not stimulated by the increase in soil moisture from 30 to 60% WHC. Under suboxic conditions (90% WHC and 1% O2), microbial NUE markedly declined as microbial growth rates were strongly negatively affected due to increasing microbial energy limitation. In contrast, gross N mineralization rates increased strongly as organic N uptake became in excess of microbial growth N demand. Therefore, in the moisture/O2 experiment microbial NUE was mainly regulated by the shift in O2 status (to suboxic conditions) and less affected by increasing water availability per se. These temperature and moisture/O2 effects on microbial organic N metabolism were consistent across the soils differing in bedrock and land use. Overall it has been demonstrated that microbial NUE was controlled by microbial growth, and that NUE controlled gross N mineralization as an overflow metabolism when energy (C) became limiting or N in excess in soils. This study thereby greatly contributes to the understanding of short-term environmental responses of microbial community N metabolism and the regulation of microbial organic-inorganic N transformations in soils.

environmental_effects_on_soil_microbial_nitrogen_use_efficiency_are_controlled_by_allocation_of_orga

Introduction<!><!>Sampling sites<!>Soil physical and chemical properties<!>Incubation experiments<!>Isotope pool dilution assays<!>Determination of microbial growth rate and microbial NUE<!>Potential enzyme activities<!>Calculations and statistics<!>Bedrock and land use effects on basic soil properties<!>Bedrock and land use effects on soil N processes and microbial NUE<!>Temperature effects on soil N processes and microbial NUE<!>Moisture/O2 content effects on soil N processes and microbial NUE<!>Discussion<!>Land use and bedrock effects on microbial NUE and soil N processes<!>Temperature effects on microbial NUE and soil N processes<!>Soil moisture and oxygen effects on N processes and microbial NUE<!>Conclusions<!>Supplementary Material

<p>Microbes break down soil organic matter by enzymes which enables them to assimilate organic N compounds directly such as oligopeptides, free amino acids and amino sugars (Barraclough, 1997; Jones et al., 2005; Hu et al., 2018). However, due to widespread resource imbalances and C limitation in soils (Mooshammer et al., 2014b), microbes tend to take up organic N in excess of what they require, resulting in a partitioning of organic N acquired between biomass formation and excretion of excess N through N mineralization. The proportion of organic N allocated to biosynthesis (mainly growth) relative to acquired N (organic N uptake) is termed microbial nitrogen use efficiency (NUE) (Mooshammer et al., 2014a). On the one hand, microbial NUE reflects the capacity of organic N retention in microbial biomass. It has been demonstrated that under N limitation, most acquired organic N is allocated to growth and microbial biomass, resulting in a high microbial NUE. In contrast, under conditions of N excess or C limitation a smaller fraction of organic N is used for biomass formation and more is used to meet the C demand while excess N is mineralized, resulting in low microbial NUE (Mooshammer et al., 2014a). When microbes allocate organic N to produce new biomass this improves the stabilization of N in soils. On the other hand, microbial NUE regulates the gross production of ammonium from organic N. High gross N mineralization rates are observed in soil microbes with low microbial NUE, providing inorganic N to plant growth as well as fostering N losses via nitrification coupled to nitrate leaching and denitrification. Therefore, microbial NUE determines the balance between anabolic and catabolic N processes and controls the fluxes at the intersection of organic and inorganic soil N cycles.</p><p>Despite the importance of microbial NUE, fundamental mechanisms and controls of microbial NUE are still unclear. Microbial NUE is likely affected by soil physicochemical and biological properties, such as substrate quality and quantity, microbial community composition, soil pH and clay content, which are strongly shaped by types of land use and bedrock. Forest soils are characterized by abundant C and high C/N ratios compared to grasslands, while cropland soils are managed with applications of inorganic fertilizers and receive less organic C inputs from litter and roots (Booth et al., 2005). Soil pH is strongly affected by bedrock and liming and was suggested as a prime factor influencing microbial community composition and microbial growth rates (Bååth and Arnebrant, 1994; Rousk et al., 2009, 2010). Since microbial NUE is regulated by substrate availability and microbial nutrient imbalances (Mooshammer et al., 2014a), microbial NUE might therefore be affected by land use and bedrock.</p><p>Soil temperature and moisture are among the most important environmental factors to potentially alter the balance between microbial biomass synthesis and N mineralization. It has been widely accepted that enhanced temperatures tend to raise microbial activity and accelerate turnover rates of soil N (Dalias et al., 2002; Guntiñas et al., 2012; Wang et al., 2017). Growth rates of both fungi and bacteria increased as temperature increased from 0 to 45 °C (Pietikäinen et al., 2005) and from 5 to 25 °C (Zheng et al., 2019), causing increasing allocation of N to anabolic processes fueling microbial growth. Differently, soil moisture regulates microbial metabolism and N cycling in two opposite ways (Sierra, 1997). At low soil water content microbial NUE might increase with moisture due to slow substrate diffusion and water limitation (Borken and Matzner, 2009). Microbial activity and N processes are thought to be promoted in moister soils and gross N mineralization was observed to increase in the range of 30%–90% WHC (Zaman et al., 1999; Zaman and Chang, 2004; Cheng et al., 2014), indicating a decreasing efficiency of microbial biomass formation and of microbial NUE. However, O2 becomes limiting in soils as they become water-logged (Grable and Siemer, 1968; Hollesen and Matthiesen, 2015) and obligate aerobic microbes as well as their oxidative metabolic processes are then inhibited in soils and sediments - when turning anoxic/suboxic – as indicated as microbial growth decreased markedly while respiration showed only little response (Bastviken et al., 2003; Zheng et al., 2019). This indicates that microbial NUE may decline under oxygen deficiency due to depression on microbial growth rates caused by C (energy) limitation.</p><!><p>Microbial NUE is impacted by land use and bedrock. Microbial NUE is expected to be high in forest soils being high in organic C but relatively low in available N. Microbial NUE is expected to be low in cropland soils with less plant C input but strong N fertilization. Moreover, microbial NUE is expected to be higher in limestone than in the silicate soils, resulting from more favorable conditions for microbial growth in soils with higher pH.</p><p>Temperature and moisture/O2 changes play a prominent role for microbial NUE in the short-term, therefore causing consistent responses of microbes (microbial NUE and growth) across all soils.</p><p>Microbial growth and gross N mineralization rates are accelerated by temperature in the range of 5–25 °C. Microbial NUE is promoted as a result of greater increases in anabolic processes (growth rates) than in catabolism at higher temperatures.</p><p>Microbial NUE increases from 30 to 60% WHC as a consequence of elevated substrate diffusion while microbial NUE declines dramatically at 90% WHC and 1% O2 due to energy limitation of microbial growth caused by O2 depletion.</p><!><p>Cropland, grassland and forest soils were sampled at two adjacent sites differing in bedrock (silicate and limestone) in June 2016. The two sampling sites were located in the central Enns valley, Styria, Austria and soils developed on silicate (LFZ Raumberg-Gumpenstein 47° 29′ N, 14° 6′ E, 690 m a.s.l.) and limestone (Moarhof in Trautenfels-Pürgg 47° 30′ N, 14° 4′ E, 708 m a.s.l). The soils were classified as Cambisols on silicate, and Luvisols on limestone. Mean annual temperature (MAT) is 7.2 °C and mean annual precipitation (MAP) is 980 mm (period 1980–2016). The forest was dominated by spruce (Picea abies L.) on silicate and by spruce and ash (Fraxinus excelsior L.) on limestone. Both grasslands were fertilized and regularly grazed by livestock, being sheep on silicate and cattle on limestone. Cabbage, onion, potato and beans were cultivated on cropland soils on silicate; barley, wheat and oat were grown on limestone soils. Mineral soils were collected to a depth of 15 cm in four replicates at each site using a root corer (diameter 8 cm) the replicates being sampled across major differences in topography and processed independently. Plants, litter and organic layers on the surface were removed by hand prior to soil sampling. All soil cores were collected within two days, and passed through 4 mm sieves on the day of sampling. Soils were transported back to the laboratory in Vienna and sieved to 2 mm within a day.</p><!><p>Soil water content (SWC) was measured gravimetrically after oven drying fresh soils in aluminum dishes for three days at 85 °C. Water holding capacity (WHC) was determined by repeated saturation of fresh soil (10 g) in a funnel with a filter paper and drainage for 2.5 h to compute the water retained in soils at field capacity on a dry matter basis. Soil pH was measured in water with a soil to solution ratio of 1:2.5 using an ISFET electrode (Sentron, Austria). Prior to measurements of soil organic C and total N, carbonate was removed from limestone soils by 2 M HCl. The acid-treated soils were oven dried and finely ground using a ball mill (MM2000, Retsch, Germany). Soil organic C (SOC) and N (TN) contents were subsequently quantified by an elemental analyzer (Carlo Erba 1110, CE Instruments) coupled to a Delta Plus isotope ratio mass spectrometer (Finnigan MAT) via a Conflo III (Thermo Fisher).</p><p>Dissolved C and N contents were measured in 1M KCl extracts (1:7.5 (w: v)). Soil total dissolved nitrogen (TDN) and dissolved organic carbon (DOC) contents were measured with a DOC/TN analyzer (TOC-VCPH/CPN/TNM-1, Shimadzu, Austria). Free amino acid (FAA) concentrations were determined fluorimetrically following the modified method from Jones et al. (Jones et al., 2002; Prommer et al., 2014). NH4+ and NO3− concentrations were quantified by colorimetric methods as described by Hood-Nowotny et al. (2010). Dissolved organic nitrogen (DON) was calculated as the difference between TDN and inorganic N in the KCl extracts. Microbial biomass C and N (Cmic, Nmic) were estimated by the chloroform fumigation extraction method (Vance et al., 1987) and corrected by an extraction factor of 0.45. Soil texture, exchangeable cations (K+, Na+, Mg2+, Ca2+, Fe3+, Al3+), base saturation (BS) and effective cation-exchange capacity (CEC) were assayed by the Institute of Sustainable Plant Production, Federal Agency for Food Security (AGES, Vienna, Austria), according to standard procedures (ÖNORM). Acid ammonium oxalate- and dithionite/citrate-extractable Fe and Al were measured as proxies of humusbound, amorphous and crystalline metal (hydr)oxides (Courchesne and Turmel, 2007) by ICP-MS at the Institute of Soil Sciences, University of Natural Resources and Life Sciences, Vienna, Austria.</p><!><p>To investigate the responses of microbial NUE and gross N transformation processes (N mineralization, NH4+ immobilization and nitrification) to soil temperature and soil moisture/O2 in the short term, we set up two separate laboratory incubation experiments with microcosms.</p><p>Before the temperature experiment, 250 g sieved soil of each sample was transferred into polyethylene Ziploc bags and amended with Milli-Q water to achieve 60% WHC. The amended soils were kept in a thermal incubator at 15 °C for seven days to achieve moisture equilibration. Soil bags were opened regularly for aeration and water lost was added. These soils were then weighed into 50 ml polypropylene vials in several replicates one day before starting the isotope pool dilution (IPD) experiments and exposed to 5, 15 and 25 °C, respectively, allowing for estimations of labile N pools (Nmic, FAA, NH4+ and NO3−), microbial N processes (microbial N growth, gross N mineralization, gross NH4+ immobilization, and gross nitrification) and potential enzyme activities (β-glucosidase, leucine-amino peptidase and acid phosphatase) at three levels of temperature.</p><p>The moisture/O2 experiment was performed two weeks later. Soils were prepared in triplicate in 50 ml polypropylene vials and the water contents were adjusted to 30, 60 and 90% WHC respectively by gentle drying in ambient air or by water addition. The adjusted soils were incubated at 20 °C at two O2 levels: soils with 30 and 60% WHC were exposed to normal air at 21% O2 while soils with 90% WHC were incubated in a suboxic chamber at 1% O2. The aliquots were then kept under the respective conditions for one week before starting measurements and aerated every second day. Soil N pools, gross N processes, and soil enzymes were measured consecutively.</p><!><p>Gross N mineralization, NH4+ immobilization and nitrification were estimated by isotope pool dilution techniques (15N-IPD) as modified by Wanek et al. (2010). In principle, the assays were conducted by labeling the NH4+ and NO3− pool respectively with enriched 15N tracers and the gross rates of influx to and efflux from the target pool were calculated based on the differences in isotopic composition and pool size between two time points. Considering the stimulatory effects caused by addition of the 15N tracer on immobilization rates, low amounts of inorganic 15N were applied which amounted to less than 20% of the initial pool sizes. Initial pool sizes of NH4+ and NO3− were measured the day before starting the 15N-IPD experiments by colorimetric methods (Hood-Nowotny et al., 2010).</p><p>To determine gross N mineralization, aliquots of 4 g fresh soils were weighed into 50 ml polypropylene vials in duplicate and labelled dropwise with 200 μl of (15NH4)2SO4 (98 atom% 15N, Spectra and Cambridge Isotope Laboratories, Europe, Radeberg, Germany). The concentrations of tracer solutions were adjusted according to preceding measurements of soil NH4+ to increase isotopic enrichment to<20 at %. Labelled soils were shaken vigorously to distribute the tracer solution homogeneously and then incubated under the specific conditions of temperature and moisture/O2. The soil incubations were terminated after 4 h (t1) and 24 h (t2) by extraction with 20 ml of 1M KCl (60 min at room temperature, filtration through ash-free filter paper). Soil NH4+ contents were quantified directly after extraction (Hood-Nowotny et al., 2010). Prior to 15N isotopic analyses, NH4+ in the extracts was isolated using a micro-diffusion method (Brooks et al., 1989; Zhang et al., 2015). For this, soil extracts were pipetted into 20 ml polypropylene scintillation vials and diluted with 1M KCl if necessary to reach a final volume of 10 ml with NH4+ concentrations less than 50 μM. The vials were capped after addition of each an acid trap (4mm diameter of cellulose filter paper disks soaked with 4 μl 2.5M KHSO4, wrapped in Teflon tape) and MgO powder (100 mg, to raise the pH > 9.5) and shaken slowly at room temperature for three days. The acidified cellulose disks with the collected NH4+ were picked out from the acid traps (i.e. the Teflon tapes) after drying and dissolved in 2 ml of Milli-Q water by shaking for 30 min. The NH4+ concentrations were then diluted with Milli-Q water to reach 20 μM and the isotopic composition of NH4+ was measured by a chemical conversion procedure (Zhang et al., 2015). Briefly, aliquots of 1 ml of the acid trap (Milli-Q water) extract were added to 12 ml screw cap exetainers. NH4+ was oxidized to nitrite (NO2−) by addition of 0.1 ml of sodium hypobromite (BrO−) solution and the remaining BrO− was deactivated by adding 50 μl sodium arsenate (0.51 g NaAsO2 in 10 ml of Milli-Q water) after reacting for 30 min. The BrO− reagent was prepared as described in Zhang et al. (2007). Then, 0.15 ml of buffered sodium azide (1M NaN3 in 50% acetic acid) was injected into the tightly sealed exetainers to reduce the produced NO2− to nitrous oxide (N2O) at room temperature, and 0.1 ml of 10M NaOH solution was injected to terminate the reaction after 60 min.</p><p>To determine gross nitrification rates, aliquots of 4 g fresh soils were weighed into 50 ml polypropylene centrifuge tubes and labelled dropwise with 200 μl of K15NO3 (98 atom%, Isotec-Sigma Aldrich, Vienna, Austria). The concentrations of the tracer solutions were adjusted according to preceding measurements of soil NO3− to increase isotopic enrichment to < 20 at%. A similar workflow was carried out as mentioned for gross N mineralization, with two time points for assay termination (4 and 24 h). NO3− contents were measured directly after extraction with 1 M KCl (Hood-Nowotny et al., 2010). The isotopic ratios and concentrations of NO3− were then measured in 1 ml aliquots of each soil extract by chemical conversion of NO3− to NO2− and further to N2O by injections of buffered NaN3 and acidic VCl3 (McIlvin and Altabet, 2005; Lachouani et al., 2010). For this, first 200 μl buffered sodium azide (1M NaN3 in 10% acetic acid) and then 1 ml of acidic VCl3 (50mM VCl3 in 1M HCl) were added with gas tight syringes and allowed to react for 18 h at 37 °C, and then stopped by injecting 300 μl 6M NaOH.</p><p>The produced N2O from both, NH4+ and NO3−, was isotopically characterized by a purge-and-trap IRMS (PT-IRMS) consisting of a cryofocusing unit on a Gasbench II headspace analyzer (Thermo Fisher, Germany) coupled to a Finnigan Delta V Advantage IRMS (Thermo Fisher, Germany) (Lachouani et al., 2010). A series of natural abundance and 15N-labelled standards of NH4+ and NO3− were prepared in the same matrix (1M KCl) and run with each batch of samples to allow determination of concentrations and isotopic ratios of NH4+ and NO3−.</p><!><p>Here, a new approach to measure microbial NUE is proposed based on concurrent measurements of microbial growth and gross N mineralization rates. Microbial NUE is calculated as the ratio of microbial growth N rate (Ngrowth) over microbial organic N uptake rate (Nuptake). Microbial organic N uptake represents the sum of microbial growth and gross N mineralization (MN). Microbial NUE is dimensionless and ranges between 0 and 1. MicrobialNUE=NgrowthNuptake=NgrowthNgrowth+MN</p><p>Microbial growth N rates were determined by 18O–H2O incorporation from soil water into double-stranded DNA (dsDNA) of dividing soil microorganisms (Spohn et al., 2016; Zheng et al., 2019). In this method both growth and mineralization rates are measured over 24 h, thus providing the same time integral for anabolic and catabolic processes, and allowing for more robust representations of microbial organic N uptake (the sum of microbial growth N and N mineralization rates) and growth allocation.</p><p>For this method aliquots of 0.4 g fresh soil were weighed into 2 ml plastic tubes with caps in duplicates. Half of the soil aliquots were labelled with 18O–H2O (97.0 at%, Campro Scientific) reaching a final 18O enrichment of 20.0 at% in soil water. In parallel, the other half was amended with the same amount of natural abundance H2O serving as controls. All tubes were transferred into serum bottles (35 ml glass vials with crimp caps) and incubated after capping for 24 h. Then the soil samples were retrieved, frozen in liquid nitrogen and stored at −80 °C before further analysis. Microbial dsDNA was extracted from the frozen soils with a kit (FastDNA™ SPIN Kit for Soil, MP Biomedicals, Germany) and quantified by the Picogreen fluorescence assay (Quant-iT™ PicoGreen® dsDNA Reagent, Thermo Fisher, Germany) following the manufacturers' instructions. The oxygen content and oxygen isotope composition (18O/16O) of dsDNA was analyzed after drying aliquots of the DNA extract in silver capsules by a Thermochemical Elemental Analyzer coupled to an Isotope Ratio Mass Spectrometer (TC/EA-IRMS, Delta V Advantage, Thermo Fisher, Germany). Microbial biomass N was determined at the same time as microbial growth rates were measured, to allow calculations of the ratios of microbial biomass N to soil DNA content (fDNA-N factor). Calculation of microbial growth was described shortly below and in detail in Zheng et al. (2019), and microbial N growth (Ngrowth) was computed by multiplying DNA produced by fDNA-N. In this study, microbial growth therefore represents the microbial growth rate on a biomass N basis.</p><p>Total dsDNA produced (μg) during the 24 h incubation period was calculated according to differences in 18O measurements between labelled and unlabelled DNA samples. DNAproduced=Ototal×at%excess100×100at%label×10031.21 where Ototal is the total O content of the dried DNA extract (μg O), at %excess is the at% 18O of the labelled sample minus the mean at% 18O of control samples. The average weight% of O in DNA is 31.21, according to the average formula (C39H44O24N15P4). The conversion factor (fDNA-N) was calculated as the ratio of soil Nmic (mg N kg−1) to soil DNA content (mg N kg−1), which was measured and calculated for each individual soil sample. The specific fDNA-N values were then applied to each soil replicate individually which multiplied by the DNA production rate, enabled calculating microbial growth rates based on dry soil mass (Ngrowth, mg N kg−1 d−1). Ngrowth=fDNA−N×DNAproducedDW×t where DW is the soil dry mass and t is the incubation time in days.</p><p>The initial approach to quantify microbial NUE is constrained by an indirect assessment of microbial growth using free amino acid consumption rates (AAuptake; as a proxy of organic N uptake) minus gross N mineralization rates (MN) as a proxy of growth rates (Wanek et al., 2010; Mooshammer et al., 2014a). MicrobialNUE=AAuptake−MNAAuptake</p><p>The method quantifies the contribution of amino acids to N assimilation, but does not include other potential N source such as amino sugars that become available to microbes by microbial necromass decomposition in soils (Hu et al., 2017, 2018). By this the initial approach might underestimate organic N uptake and therefore microbial NUE, and better is termed microbial amino acid use efficiency (Andresen et al., 2015). Moreover, amino acid consumption can potentially be stimulated by adding 15N-labelled substrates to label the free amino acid pool, though addition levels are usually restricted to ∼20% of the native pool size therefore having little impact on rate measurements. Finally, amino acid uptake rates are determined over a time period of 1 h only while gross N mineralization rates are measured over 24 h, causing the process rates contributing to calculations of microbial NUE to be out of balance in terms of time integral. Despite the few but important studies on microbial NUE conducted using the isotope pool dilution technique, alternative methods are therefore clearly required to obtain more robust estimations of microbial NUE across a large range of soils.</p><!><p>β-Glucosidase (BG), leucine amino peptidase (LAP) and acid phosphatase (PHO) were determined fluorimetrically with artificial substrate additions (Kaiser et al., 2010). 4-Methylumbelliferyl (MUF) based substrates were used to assay the enzyme activities of BG and PHO. An L-Leucine-7-amido-4-methylcoumarin (AMC) based substrate was used to measure LAP. Soil slurries (1:100 (w:v)) were prepared in 50 mM sodium acetate buffer (pH = 5) and appropriate substrates were pipetted into wells of black microplates in triplicates and incubated in the dark for 30–180 min with repeated measurements of sample fluorescence. Microplates were read by a TECAN Infinite® M200 spectrophotometer at an excitation/emission wavelength of 365/450 nm. The fluorescence of samples was corrected for quenching and the concentrations of released MUF and AMC by enzymatic cleavage were calibrated by respective standards.</p><!><p>Gross production and immobilization rates of NH4+ and NO3− were calculated using the 15N-IPD equations developed by Kirkham and Bartholomew (1954). NO3− immobilization rates were negative in some soils due to low NO3− assimilation processes or due to the heterogeneity between replicates and thus were not further considered.</p><p>The temperature sensitivity (Q10) of a specific process is its response to an increase in temperature of 10 °C (Kirschbaum, 1995). In our study, Q10 values of microbial N processes, microbial NUE and soil enzymes were calculated with a linear regression model after logarithmic transformations (Hu et al., 2018) across all three incubation temperatures. LN(R)=LN(Q10)10×T+b where R is the measured process including gross N mineralization, NH4+ immobilization, gross nitrification, microbial growth, microbial N uptake, and potential enzyme activities, Q10 is the temperature sensitivity of the parameters mentioned above, T is the incubation temperature (°C) and b is a fitted coefficient.</p><p>All statistical analyses were performed using R 3.1.3 (R Development Core Team, 2015). Normal distribution and homoscedasticity were tested with Shapiro-Wilk test and Levene test. Data were transformed if necessary. Two-way ANOVA was conducted to test the effects of bedrock and land use followed by Tukey's HSD post-hoc test. Pearson correlations were used to assess the relations between soil properties with microbial N processes as well as with microbial NUE. Path analyses were run to discriminate the direct and indirect factors influencing microbial NUE at 15 °C and 60% WHC. The function of sem () was performed in the R language package of "lavaan". Three-way ANOVA was conducted to check for the main effects of bedrock, land use, and temperature or soil moisture/O2, and their interactions. The relative importance of the different main factors and their interactions was further calculated as the percentage of variance (sum of squares (SS)) contributed by each factor and their interactions to the total variance in three-way ANOVAs. In addition, nonlinear models were performed to investigate general relationship between microbial growth and microbial NUE using the nls() function in the "lme4" package of R language, and applied linear regression models to explore the effect of microbial NUE on gross N mineralization with temperature and moisture/O2 treatments. Significance levels were set to P < 0.05 and all values are presented as mean ± 1 standard error of four replicates.</p><!><p>Soil physicochemical properties were strongly affected by land use and bedrock (Table 1, Fig. 1). Soil pH was lower in silicate soils (pH 4.2 to 5.9) than in limestone soils (pH 6.1 to 8.2), and across both bedrocks decreased in the order cropland > grassland > forest. Fine soil particle contents (clay and silt) were higher in limestone than in silicate soils while sand contents behaved inversely. Oxalate and dithionite extractable Al and Fe oxides (except dithionite-extractable Fe) differed between bedrocks, with higher contents in silicate soils, but were not affected by land use. Soil TN and NO3− were higher in limestone soils, while DOC and free amino acids were higher in silicate soils. Microbial biomass (Cmic and Nmic) was higher in limestone soils than in silicate soils, and the maximum values were found in grassland soils on both bedrocks.</p><!><p>Here land use and bedrock effects were tested on microbial processes measured under standard conditions in the temperature experiment (15 °C and 60% WHC). Microbial growth rates were in the range of 0.67–3.59 mg N kg−1 d−1 and were not affected by land use and bedrock (Table 2). Microbial NUE varied between 0.22 and 0.84, and was neither affected by land use or bedrock, but strongly by their interaction. This became clearly evident from the inverse land use effects in silicate (cropland < grassland < forest) and in limestone soils (cropland > grassland > forest). Gross N mineralization rates ranged from 0.53 to 2.62 mg N kg−1 d−1 but did not show a consistent pattern across land uses and bedrocks. Gross nitrification rates were much lower than gross N mineralization rates. Gross nitrification was lowest in silicate forest soils and highest in grassland soils.</p><p>Soil N processes and microbial NUE were related to several soil physicochemical properties (Table 3). Microbial growth rates were related to microbial biomass (Cmic and Nmic), DON, soil texture (sand and silt) and oxalate- and dithionite-extractable Al contents. Gross N mineralization rates were negatively related to clay, SOC, soil C/N and microbial C/N. Microbial NUE increased with SOC and Cmic, as well as with soil C/N and microbial C/N. No significant correlations were found between microbial NUE and any of the measured soil N pools. The result of path analysis (Fig. 2) indicated that SOC, microbial C/N, clay content and dithionite-extractable Al content were major factors influencing microbial growth, gross N mineralization and thereby microbial NUE.</p><!><p>Microbial growth rates and enzyme activities were accelerated by increasing temperature from 5 to 25 °C (Fig. 3A, Fig. S1 A-C), while gross N mineralization rates remained almost constant with temperature (Fig. 3B, Table 4). Both microbial growth and gross N mineralization rates were low at 5 °C while microbial growth rates increased rapidly at higher temperatures in contrast with gross mineralization rates (Fig. 5). Microbial growth became predominant over organic N uptake and microbial NUE increased with temperature in all soils (Fig. 3D), with a greater contribution of microbial growth than gross N mineralization to explain this response (Fig. S3 A and B).</p><p>Temperature sensitivities (5−25 °C) of soil N processes and microbial NUE differed between soils (Table 5). Microbial growth showed the highest temperature sensitivity (Q10) among all studied processes, ranging from 2.28 ± 0.56 to 3.53 ± 0.22. Microbial NUE was sensitive to temperature changes with Q10 values ranging from 1.30 ± 0.05 to 2.48 ± 0.34. Q10 values of NUE were lower at higher SOC and clay content (Table S2). NH4+ immobilization and nitrification exhibited a higher temperature sensitivity compared to gross N mineralization.</p><!><p>No significant difference was detected in any measured soil N process as well as in microbial NUE between 30 and 60% WHC under oxic conditions (Fig. 4, Table S3). Microbial growth rates were apparently higher than gross N mineralization rates, and microbial NUE was more than 0.5 in most soils (Fig. 5). In contrast, changing moisture/O2 conditions to 90% WHC and 1% O2 content dramatically decreased microbial growth and organic N uptake rates (Fig. 4 A and C), while gross N mineralization rates increased markedly (Fig. 4B). Microbial NUE declined under suboxic conditions caused by the depression of microbial growth rates (Fig. 5). Across overall moisture/O2 treatments, microbial NUE was found largely driven by microbial growth while gross N mineralization responded to a smaller extent (Fig. S3 C and D). NH4+ immobilization increased, particularly in silicate forest soils and in all limestone soils (Fig. 4E). Nitrification was severely depressed in most soils, except in acidic forest soils, under suboxic conditions (Fig. 4F). Concurrently, NH4+ concentrations increased by an order of magnitude, whereas NO3− concentrations declined to levels close to the detection limit (Fig. S2 H and I).</p><!><p>In this study we for the first time demonstrate that microbial NUE responds highly sensitively to short-term fluctuations in soil temperature and soil moisture/O2 regime, and therefore strongly affects inorganic N cycling and microbial N retention. In contrast bedrock and land use effects were mediated by changes in soil physicochemistry and soil biology affecting C and N availability. We moreover demonstrated that NUE is mainly driven by factors promoting microbial growth, and that changes in microbial NUE adversely impacted microbial N mineralization. Our view of gross N mineralization has greatly changed in the recent years from an extracellular process controlling soil N cycling to an intracellular process driven by the balance between microbial N demand and supply (Mooshammer et al., 2014a). Microbial N metabolism is central to the metabolic functioning of microbes and therefore was shown to be strongly controlled (Kingsbury et al., 2006; Shimizu, 2013). The balance between anabolic processes (such as nucleic acid and protein biosynthesis, and growth) and catabolic processes (that eventually trigger exudation of catabolic products in excess of microbial N demand) in microbes is therefore highly regulated (Mooshammer et al., 2014a), and microbial NUE is regulated independent of microbial C metabolism, though some cross-talk between microbial C and N metabolism is expected (Kingsbury et al., 2006; Shimizu, 2013).</p><!><p>As an important component of microbial NUE, microbial growth triggers the assimilation of organic N and thus promotes the efficiency of retaining N in microbial biomass. Microbial growth is assumed to be highly dependent on substrate quantity and quality (Rousk and Bååth, 2011), while it was not directly correlated with available C, N and soil C/N (except DON) in this study. In contrast, microbial growth was strongly affected by soil texture and Al (hydr)oxides through changing substrate availability. Soil organic matter (SOM) can be stabilized by organo-mineral interactions with clays and silts (Jensen et al., 1989; Lützow et al., 2006) and Fe- and Al (hydr)oxides (Wiseman and Püttmann, 2005), further reinforced by high Ca2+, CEC, BS, and calcium carbonate (Six et al., 2004; Doetterl et al., 2015; Minick et al., 2017; Rowley et al., 2018). However, this organic matter is likely more resistant against decomposition due to stronger mineral-organic matter binding, causing eventual nutrient limitation. In addition, microbial growth was found to be positively linked to microbial biomass (Cmic and Nmic), indicating the co-dependency of microbial biomass and growth during nutrient utilization. It is interesting to note that microbial growth was not affected by soil pH though soil reaction is a major driver of microbial community composition (Fierer and Jackson, 2006; Lauber et al., 2009) and nutrient availability (Bleasing, 2012; Augusto et al., 2017) and differentially affects bacterial and fungal growth (Rousk et al., 2009; Rousk and Bååth, 2011).</p><p>Compared to microbial growth, gross N mineralization seems less informative when interpreting the allocation of organic N to biosynthesis and growth as it is regulated by microbial NUE releasing organic N in excess. However, changes in gross N mineralization indicate a nutrient imbalance to microbial growth, which is useful to investigate the decoupling of organic N uptake and actual N demand decreasing microbial NUE. In this study, gross N mineralization was found to increase with decreasing SOC and microbial C/N as well as with clay content, as a consequence of enhancement in relative C limitations. Under C limiting conditions organic compounds containing C and N such as amino acids and other DON compounds are utilized by microbes for growth and for energy metabolism (Schimel and Bennett, 2004; Geisseler et al., 2010), and with increasing C limitation a larger fraction of organic N taken up is not retained or allocated to growth but mineralized and exuded as NH4+, causing increases in gross N mineralization. Therefore we found that with increasing C availability microbial NUE increased and gross N mineralization decreased. In their data synthesis Booth et al. (2005) also found that gross N mineralization was negatively related to soil C/N, after correcting for effects of SOC quantity. Although microbial biomass N (Wang et al., 2018) and soil pH (Cheng et al., 2013) were reported also as important effectors of gross N mineralization, these were not confirmed by our dataset. Overall, microbial growth and gross N mineralization are crucial drivers and outputs of microbial NUE but they correlate with microbial NUE in opposite directions (Table S1). According to the result of path analysis, microbial NUE was primarily affected by enhanced N limitation (higher SOC and wider microbial C/N) and greater physical protection of SOC (higher clay content and Al (hydr)oxides). Furthermore, microbial NUE was almost equally related to microbial growth and gross N mineralization, but growth and N mineralization were independent and not correlated (Table S1). This means that factors promoting anabolic processes and growth did not or negatively affect catabolic processes such as N mineralization. Again this highlights the central role that microbial NUE plays in regulating the partitioning of organic N between anabolic and catabolic processes.</p><!><p>As expected, microbes became active and turned to grow as temperatures increased. The activities of enzymes functioning in soil organic C, N and P mobilization were also promoted across all soils (Fig. S1 A-C), likely increasing the substrate availability for microbial assimilation. Consistent with previous studies (Pietikäinen et al., 2005; Barcenas-Moreno et al., 2009), microbial growth rates were stimulated by increasing temperatures, and growth increased exponentially from 5 to 25 °C in all studied soils. Contrasting to our hypothesis, gross N mineralization rates remained invariant to temperature which also was different from the findings of most other related studies (Schütt et al., 2014; Cheng et al., 2015). Microbial N uptake rates increased dramatically with temperature at the same time, but N requirements for growth were more strongly enhanced causing an increase in microbial NUE and a decline in the proportion of organic N released as NH4+. Therefore, microbial growth rates and gross N mineralization rates exhibited different responses to increasing temperature in all soils, while microbial NUE increased obviously regulated by enhanced growth N demands (Fig. S3 A and B).</p><p>In contrast, microbial growth was more strongly limited than N mineralization at low temperatures (5 °C). Though microbial NUE varied in a wide range from 0.05 to 0.75 in soils, catabolic processes were more strongly negatively related to microbial NUE at low temperatures, with a relatively flat growth-NUE relationship. This indicates that growth (proliferation, cell division) is more cold sensitive than catabolic processes, i.e. gross N mineralization. In contrast, at higher temperatures the relationship between NUE and N mineralization changed little while the slope of the relationship between growth and NUE increased markedly, indicating a greater control of anabolic processes at intermediate to high soil temperatures and therefore increasing microbial N stabilization in warmer soils.</p><p>Microbial NUE regulates the proportion of organic N routed to mineralization, providing the substrate to nitrification. Nitrification was found to increase with temperature (Van Schöll et al., 1997; Lang et al., 2010), however, it was not related to microbial NUE here (data not shown) due to other factors more strongly affecting nitrification, for example low soil pH in silicate forests. Likewise, gross NH4+ immobilization responded positively to temperature increases, while nitrification accounted only for a minor fraction of NH4+ immobilization in most soils.</p><!><p>Despite the wealth of studies on gross N transformation processes in unsaturated soils much less is known of their response to soil water saturation and O2 limitation. Microbial activity was high under oxic conditions (21% O2), conditions that prevailed at 30 and 60% WHC. This indicates that most soil N processes catalyzed by aerobic microbes were not stimulated by increases in soil water content from suboptimal to optimal levels. Substrate accessibility and diffusivity are supposed to increase with soil water content i.e. also from 30 to 60% WHC, due to faster diffusion rates of substrates and extracellular enzymes along thicker and better connected water films coating the soil particles (Linn and Doran, 1984; Or et al., 2007; Manzoni et al., 2012). Furthermore, more labile substrates might become dissolved into the growing soil water films as well as be released from weak interactions with mineral surfaces, further stimulating microbial activity. One possible explanation for this unresponsiveness is the wide optimal moisture range of soil microbial metabolism. This is supported by maximum bacterial growth rates being reached at relatively low soil moisture levels, i.e. at around 20% water content on a dry matter basis which translates to approximately 33% WHC (Iovieno and Bååth, 2008). Extrapolating the (absent) response of microbial N processes to soil moisture levels between 30 and 60% WHC to 90% WHC implies that one would also not expect that changes induced at highest WHC would affect substrate/enzyme diffusion to an extent that triggers major changes in soil N processes. However, the highest WHC level also switched the microcosms to a suboxic regime (1% O2), causing microbial O2 limitation reinforced by slowed O2 transport through the water filled pores.</p><p>Potential enzyme activities declined or remained constant (Fig. S2 A - C), implying a slowing down of microbial metabolic activity (Zaman et al., 1999). Microbial growth declined remarkably under suboxic conditions, compared to rates measured at 30 and 60% WHC at 21% O2. The slow growth rates can either be attributed to substrate limitation or to energy limitation of the aerobic microbial community. Substrate availability seemed not to have become limiting after one week of incubation under suboxic conditions, as the concentrations of free amino acids as labile energy-rich substrates increased at 90% WHC (Figs. S2 and G). In contrast, soil microbes were severely inhibited by O2 limitation, with indications of a switch from aerobic respiration to anaerobic respiration based on the use of alternative electron acceptors such as NO3− by denitrification (Figs. S2 and I). Nitrate levels declined to levels below the limit of quantification after one-week suboxia, high-lighting denitrification as alternative to aerobic respiration for energy production. While nitrification decreased due to O2 limitation, as also reported by Pett-Ridge et al. (2006), gross N mineralization was promoted dramatically in all soils. This implies that N was not the limiting element under suboxic conditions, but that microorganisms were restricted by C (energy) and therefore took up and utilized free amino acids for energy production rather than for N. Other studies also found that under anoxic conditions soil microbial metabolism produces energy with a lower efficiency by anaerobic respiration than under oxic conditions, also negatively affecting organic matter decomposition rates and microbial growth (Boyd, 1995; Bastviken et al., 2001; Schädel et al., 2016). Moreover, microbial NUE was found to decline strongly at 90% WHC, supporting microbial limitation by energy (C), and henceforth gross N mineralization as an N overflow metabolism increased as organic N uptake became in excess of microbial N demand for growth. Therefore, shifting the soil O2 status from oxic to suboxic was likely to reverse microbial N limitation to C limitation, thus turning the balance from predominant anabolic processes to predominant catabolic processes.</p><!><p>In this study microbial NUE was driven by microbial growth but in turn regulated microbial N mineralization. Across the studied soils differing in land use and bedrock, microbial NUE was primarily affected by C availability, microbial C/N relations, and organic matter sorption (clay content and Al (hydr) oxides), implying the important regulation by C–N imbalances. Consistent across all soils microbial NUE increased with temperature, and decreased with O2 depletion. Future experiments shall expand on these findings and explore the long-term effects of environmental factors on microbial growth and NUE. This will greatly advance the understanding of the soil inorganic N cycle and has strong ecosystem repercussions such as for soil N conservation versus soil N leakage by gaseous or hydrological pathways.</p><!><p>Supplementary data to this article can be found online at https://doi.org/10.1016/j.soilbio.2019.05.019.</p>

PubMed Author Manuscript

A Bidirectional System for the Dynamic Small Molecule Control of Intracellular Fusion Proteins

Small molecule control of intracellular protein levels allows temporal and dose-dependent regulation of protein function. Recently, we developed a method to degrade proteins fused to a mutant dehalogenase (HaloTag2) using small molecule hydrophobic tags (HyTs). Here, we introduce a complementary method to stabilize the same HaloTag2 fusion proteins, resulting in a unified system allowing bidirectional control of cellular protein levels in a temporal and dose-dependent manner. From a small molecule screen, we identified N-(3,5-dichloro-2-ethoxybenzyl)-2H-tetrazol-5-amine as a nanomolar HALoTag2 Stabilizer (HALTS1) that reduces the Hsp70:HaloTag2 interaction, thereby preventing HaloTag2 ubiquitination. Finally, we demonstrate the utility of the HyT/HALTS system in probing the physiological role of therapeutic targets by modulating HaloTag2-fused oncogenic H-Ras, which resulted in either the cessation (HyT) or acceleration (HALTS) of cellular transformation. In sum, we present a general platform to study protein function, whereby any protein of interest fused to HaloTag2 can be either degraded 10-fold or stabilized 5-fold using two corresponding compounds.

a_bidirectional_system_for_the_dynamic_small_molecule_control_of_intracellular_fusion_proteins

Introduction<!>The stabilization of HaloTag2 with 22 point mutations provides the rational to screen for small molecule stabilizers of HaloTag2<!>The bidirectional HyT/HALTS system allows temporal and dose-dependent control of HaloTag2 fusion protein levels in cells<!>HALTS1 provides thermal stability to HaloTag7 by binding to it directly with a dissociation constant of 143 nM<!>The HyTs/HALTS system modulates HaloTag2 fusion protein levels in cells through Hsp70 and the Ubiquitin-Proteasome System<!>HALTS1 binding to the active site of HaloTag2 induces conformational changes that may be involved in stabilization<!>Use of the HyT/HALTS system to study the physiological effects of either decreasing or increasing in activity of therapeutically relevant proteins<!>Conclusions<!>Cell Culture and Materials<!>Immunoblotting<!>HaloTag2 Aggregation Assays<!>Small Molecule HaloTag2 Stabilization Screen<!>Recombinant Protein<!>Thermal Shift Assays<!>Isothermal Titration Calorimetry<!>Immunoprecipitation<!>Focus Formation Assay

<p>The use of small molecules to manipulate protein activity has proven helpful for delineating protein function and treating human disease1. Although great strides have been made in developing small molecule inhibitors for certain classes of proteins (e.g. kinases and G protein-coupled receptors), the great majority of proteins lack a potent small molecule inhibitor2. The cost and effort of developing potent small molecules often precludes many attractive drug targets from drug development programs. Therefore, a need exists for a streamlined method to establish whether a putative protein is critical for disease progression and thus could serve as a drug target.</p><p>An additional challenge in therapeutic development is that many drugs have off-target effects that can derail their further development. Thus, validating the on-target effects of new drugs in animal models is necessary. Although RNA interference3; 4 and genetic knockout have played a crucial role in drug target discovery and validation, there are many instances in which a small molecule-based target validation approach would provide utility5. For example, some of the more attractive features of small molecules include: high bioavailability, direct action on targeted proteins, precise dose dependence, rapid temporal control, and reversibility. Therefore, a small molecule approach would provide significant utility as an additional tool in drug target validation efforts.</p><p>Significant effort in the chemical biology community has focused on developing systems that induce the destabilization of a target fusion protein using small molecules (reviewed previously6). These generalized strategies should, in principal, allow researchers to assess the effects of small molecule inactivation of a target protein before a specific inhibitor has been developed. Recently, two methods were developed where the addition of a single small molecule is sufficient to induce the degradation of any target fused to a specific tag. Both techniques enable researchers to perform chemical knockdown of a protein of interest by fusing it to a single degradation-conferring protein. The first system fuses target proteins to FKBP12 containing a cryptic degron, which upon binding to a small molecule FK506 analog, leads to destabilization of the fusion protein7; 8. Recently, we developed a system whereby a small molecule is employed to append hydrophobic tag (HyT) compounds to the surface of the HaloTag2 protein9, thereby mimicking a misfolded protein state and inducing proteasome-mediated degradation of HaloTag2 fusion proteins10; 11. This system allows for degradation of approximately 90% of the fusion protein and has been shown to work in cell culture, zebrafish, and murine disease models.</p><p>Although the HaloTag2 system is an effective tool to degrade proteins of interest, it lacked a complementary method to evaluate the stabilization of target proteins. As evidenced by the recent development of sirtuin activators for diabetes and obesity12, protein kinase C activators for acute myeloid leukemia13, and p53 stabilizers for various cancers14, many studies require the ability to increase the activity of a protein of interest in disease models. In fact, a method to stabilize FKBP12 derivative fusion proteins with a FK506 analog has been shown to work in many systems, complementing the degradation system that employs the same small molecule but a different variant of FKBP128. However, an ideal system for the small molecule control of fusion proteins would employ one domain for both stabilization and destabilization so that the activity of a single fusion protein can be either increased or decreased via the use of two different small molecules.</p><p>Here, we describe a small molecule stabilizer of the HaloTag2 protein. This HaloTag2 Stabilizer (HALTS1) binds the HaloTag2 protein with nanomolar affinity, conferring a large thermodynamic stability to HaloTag2-fusion proteins. HALTS1-stabilized HaloTag2 shows drastically reduced interaction with heat shock protein Hsp70, resulting in reduced ubiquitination and reduced aggregation upon proteasome inhibition. Moreover, we demonstrate the functional utility of the combined HyT/HALTS system in probing the physiological role of putative drug targets by modulating the activity of an oncogenic HRas-HaloTag2 fusion protein in a focus formation assay. In summary, we have developed a system whereby any protein of interest fused to a single tag can be targeted with a small molecule to be degraded 10-fold or stabilized 5-fold in a reversible, dose-dependent manner, thereby providing new opportunities in drug target validation.</p><!><p>HaloTag2 is a mutant bacterial dehalogenase9 with limited stability (t1/2 ~9h, Supplementary Information Figure S1) when expressed in eukaryotic cells. To increase its stability, 22 mutations have previously been introduced to create the more stable HaloTag7 protein for recombinant protein work15; 16. We hypothesized that these mutations would also provide increased stability in mammalian cells. In fact, when HEK293 FlpIn cells were used to insert HaloTag7- and HaloTag2-enhanced green fluorescent protein (EGFP) fusion proteins into identical genetic loci, we observed ten times lower levels of HaloTag2-EGFP, suggesting decreased intracellular stability for HaloTag2 (Figure 1a). Consistent with its decreased intracellular stability, we also observed aggregation of HaloTag2-EGFP upon treatment with a proteasome inhibitor, YU101, whereas no aggregation was observed with the more stable HaloTag7 (Figure 1b). These data suggest that HaloTag2 can render instability to fusion proteins and that it would be possible to stabilize HaloTag2 using a small molecule.</p><p>To find small molecule stabilizers of HaloTag2, we screened a 35,000 compound library of small molecules for compounds that prevent the degradation of HaloTag2-luciferase in response to the HaloTag2 destabilizer, HyT1310. The rationale for including HyT13 in the assay was to expedite the degradation of the HaloTag2 fusion protein, thereby enabling the cell-based screen to be completed in 24 hours. The top hits from the screen were largely able to prevent the HyT13 mediated degradation, and notably, a clear pharmacophore emerged among the top hits (Figure 1c). Three aminotetrazoles, containing mono- or dichloro-substituted benzyl amines were among the top four hits. After evaluating the efficacy and solubility of our top hits in follow-up assays, we decided to investigate further N-(3,5-dichloro-2-ethoxybenzyl)-2H-tetrazol-5-amine, hereafter referred to as HaloTag Stabilizer1 (HALTS1).</p><!><p>To assess the kinetics of HaloTag2 stabilization, we treated HA-EGFP-HaloTag2 expressing cells with 1 μM HALTS1 for increasing lengths of time. Initial HALTS1-induced stabilization was observed by 6 hours and a new equilibrium was reached at 72 hours, at which point the fusion protein level had increased 4-fold (Figure 2a). Similarly, we treated the same cells with various concentrations of HALTS1 for 48 hours. Fusion protein stabilization was observed beginning at 50 nM with a maximum reached at 2.5 μM, at which point an almost 5-fold increase in protein level was observed (Figure 2b). From these cell culture studies, we estimated that the EC50 for HALTS1 is approximately 100 nM. No toxicity was observed with HALTS1, which was tested in an alamarBlue Cell Viability Assay with a 24 hour treatment of up to 200 μM HALTS1 (Supplementary Information, Figure S2). Given the ability of the HaloTag2 destabilizing ligand HyT36 to reduce HaloTag2 levels by 10-fold and the ability of HALTS1 to increase the protein levels ~5-fold, the total dynamic range for small molecule-control of HaloTag2 fusion proteins is expected to be about 50-fold. To obtain a more precise measure of the dynamic range, we analyzed EGFP-HaloTag2 expressing cells by flow cytometry after treatment with 10 μM HyT36 and 10 μM HALTS1. We confirmed that the dynamic range is 42-fold (Figure 2c). Additionally, after using HALTS1 to upregulate EGFPHaloTag2 levels 2.2-fold, HyT36 treatment was able to return levels to baseline within approximately 3 hours (Supplementary Information, Figure S3). As a secondary confirmation of the stabilization of HaloTag2 by HALTS1, we also showed that HALTS1-treated cells have higher EGFP-HaloTag2 levels in the presence of cycloheximide (Supplementary Information, Figure S4). These data demonstrate that HALTS1 is able, in a time- and dose-dependent manner, to stabilize HaloTag2 fusion proteins and thereby increase their levels in cells. Furthermore, the combination of HyTs and HALTS1 results in a HyT/HALTS system for bidirectional control of a single fusion protein.</p><!><p>To test our hypothesis that HALTS1 increases cellular HaloTag2 levels by binding to HaloTag2 directly and stabilizing it, we turned to differential scanning fluorimetry17. The technique involves incubating recombinant protein with the dye SYPRO Orange, which exhibits increased fluorescence upon interaction with hydrophobic residues that are exposed as proteins unfold. By measuring the amount of increased dye fluorescence as a product of increased solution temperature, the melting temperature of the protein can be determined. If HALTS1 binds directly to the target protein and stabilizes it, we would expect to observe an increase in the protein melting temperature, i.e. a thermal shift. Given our inability to obtain a baseline melting curve for the HaloTag2 protein in this in vitro binding assay, presumably due to its inherent instability, we instead used HaloTag7 because of its increased stability at the temperatures required for the study15. As shown in Figure 3, HaloTag7 exhibited a melting temperature of 59°C, and 333 μM HALTS1 increased its melting temperature to 74°C, resulting in a total 15°C thermal shift. In contrast, a simple chloroalkane, HyT32, binds HaloTag7 covalently but does not induce a detectable thermal shift (Supplementary Information, Figure S5; see also past work with hydrophobic tags11). We next used isothermal titration calorimetry (ITC) to determine the precise binding constant and physical parameters of this interaction. We found that HALTS1 binds HaloTag7 in vitro with a Kd of 143 nM, an enthalpy of −10.7 kcal/mol, and entropy of −4.8 cal/mol/deg. Since HaloTag2 protein is less stable than HaloTag7, we expect the HALTS1 interaction with HaloTag2 to have an even more robust cellular response than with HaloTag7. We further confirmed that HALTS1 binds to and stabilizes HaloTag2 by performing a cellular thermal shift assay as described recently by Molina and colleagues18 (Supplementary Information, Figure S6). In this cellular thermal shift assay, HALTS1 bound to HaloTag2 in cell lysate and stabilized it, increasing the temperature required for HaloTag2 to precipitate out of solution.</p><!><p>We have previously shown that appending hydrophobic moieties to HaloTag2 leads to its degradation in cells by the proteasome10; 11. It is possible that HALTS1 prevents the intracellular turnover of HaloTag2 proteins by protecting HaloTag2 from the Ubiquitin-Proteasome System (UPS). Heat Shock Protein 70 (Hsp70) is a central chaperone involved in determining whether a protein is refolded or targeted for degradation by the UPS19. To study the potential interaction between Hsp70 and HaloTag2, we immunoprecipitated HA-EGFPHaloTag2 from cells treated for 2 hours with either HyT36 (10 μM) or HALTS1 (10 μM). Probing for Hsp70 in these immunoprecipitates confirmed that HyT36 leads to increased binding of Hsp70 to the fusion protein, whereas HALTS1 diminishes this interaction (Figure 4a). We did not detect any Hsp90 in the immunoprecipitates, suggesting that only the Hsp70 component of the chaperone system is involved in HaloTag2 stability surveillance. To confirm further that Hsp70 is responsible for the degradation of HaloTag2 fusion proteins, we treated cells with an Hsp70 activator, SW0220, with the expectation that the compound would enhance fusion protein degradation. We made three observations from these data: first, Hsp70 activation led to 30% degradation of the HaloTag2 fusion protein, demonstrating that Hsp70 is involved in the normal turnover of HaloTag2 proteins (Figure 4B). Second, whereas a low concentration of HyT36 led to a 50% degradation of HaloTag2 fusion protein, the combined effect of SW02 and HyT36 is additive, leading to 70% degradation. Hence, overactive Hsp70 enhances HyT36-mediated degradation. Third, even in SW02 treated cells, HALTS1 still exhibited a stabilizing effect on HaloTag2, further demonstrating that the small molecule acts upstream of Hsp70. Identical results were obtained when we increased Hsp70 levels using geldanamycin (Supplementary Information, Figure S7).</p><p>Interestingly, in the immunoprecipitates shown in Figure 4A, we noticed higher molecular weight HA-EGFP-HaloTag2 bands that appeared to correlate with increased (in the case of HyT36 treatment) or decreased (in the case of HALTS1) ubiquitination. To confirm that HALTS1 prevents HaloTag2 ubiquitination, we used the proteasome-specific inhibitor YU10121 and observed the appearance of additional higher molecular weight bands, characteristic of polyubiquitinated proteins. Inclusion of HALTS1 in the media completely blocked this ubiquitination (Figure 4c). Further, the aggregation of HaloTag2 fusion proteins in cells upon proteasome inhibition can be completely prevented by HALTS1 (Figure 4d). These results demonstrate that in cells, the thermodynamic stabilization of HaloTag2 protein by HALTS1 prevents the recruitment of Hsp70 and thus protects the protein from the UPS.</p><!><p>We next determined the crystal structure of HaloTag2 alone and in complex with HALTS1 at 2.3 Å resolution (Table S1) to provide direct, molecular evidence for its mechanism of action. We found that the compound is bound in the active site of the enzyme, in a deep tunnel between the α/β hydrolase core and the helical cap domain (Fig. 5a). In fact, the compound is completely shielded from the solvent and is not visible from the surface of the protein. The tetrazole ring of HALTS1 is bound deepest in the tunnel, and also has hydrogen-bonding interactions with the side chains of residues Asp117 and Trp118 (Fig. 5b). The phenyl ring of the compound has π-stacking interactions with Phe283, which is a histidine residue in the wild-type haloalkane dehalogenase. The cap domain contributes several hydrophobic, especially aromatic, side chains to the binding site.</p><p>Compared to the structure of HaloTag2 alone, conformational changes in the main chain and side chain of many residues in the cap domain are clearly visible in the HALTS1:HaloTag2 complex (Fig. 5c). Smaller conformational changes in several of the loops in the α/β hydrolase core that form the binding tunnel are also observed. These conformational changes upon HALTS1 binding, as well as the fact that the HALTS1 fills an internal cavity of HaloTag2, may explain the observed thermal stabilization of the structure. The conformation of free HaloTag2 is more similar to that of the wild-type Rhodococcus haloalkane dehalogenase22.</p><!><p>Given the robust destabilization and stabilization of HaloTag2 fusion proteins by HyT36 and HALTS1, respectively, we sought to demonstrate the functional utility of modulating levels of HaloTag2-fusion proteins and used oncogenic Ras as a model. NIH-3T3 cells expressing HRasG12V undergo cellular transformation and as a result, these cells begin to grow as multilayered foci. We infected NIH-3T3 cells with the HaloTag2-HRasG12V construct at low multiplicity of infection (MOI) so that the cells form only limited foci. The levels of HaloTag2-HRasG12V in these cells can be readily increased by HALTS1 and decreased by HyT36 (Figure 6A). Correspondingly, the elevated levels of oncogenic Ras lead to qualitative appearance of multilayered foci, whereas HyT36 treatment leads to disappearance of foci and reversion to an untransformed phenotype (Figure 6B). These results demonstrate that modulating HaloTag2 fusion protein levels with HyT36 or HALTS1 can lead to robust phenotypic consequences.</p><!><p>Functional validation of putative disease genes remains a significant challenge. RNAi technology provides great utility, but systemic delivery of RNAi species has been difficult to achieve. Transgenic animals with inducible shRNA or inducible genetic knockouts (e.g. Cre-LOX) can overcome many of the hurdles associated with delivery, but the timescale of protein knockdown and the catalytic nature of RNAi can make functional assignment difficult. Furthermore, adaptations associated with long-term genetic ablation may confound the true effect of gene disruption. Small molecule approaches, with a different profile of advantages and disadvantages, may therefore provide significant utility as an additional tool in drug target validation.</p><p>Given the impracticality of developing a potent and specific inhibitor/activator pair for every protein of interest, there exists an unmet need for a general small molecule-based method to modulate target proteins. We have recently demonstrated that a number of proteins of interest fused to HaloTag2 can be readily degraded by destabilizing the HaloTag2 portion of the protein with HyT molecules. We believe the ability to stabilize HaloTag2 fusion proteins with HALTS1 will greatly expand the utility of the HaloTag2 system. Hence, HALTS1 complements our previous HyT molecules, resulting in a system to manipulate protein levels in a rapid and dose dependent manner and thereby study physiological effects inaccessible to traditional approaches (Figure 6C). Thus, to functionally characterize a protein of interest, one need only generate a single HaloTag2 fusion protein, making the system particularly useful in studying transgenic animal models. For instance, in a single animal model one can test the functional consequences of inactivating and overexpressing a protein of interest. Additionally, this system can theoretically be combined synergistically with others, e.g. the FKBP12-based destabilization method mentioned earlier, to conduct more intricate experiments. For example, one could control multiple proteins in a signaling pathway in a temporal and dose-dependent manner.</p><p>In our previous studies we have demonstrated that the proteasome is responsible for degrading HaloTag2 proteins upon hydrophobic tagging. Here we extend the dissection of molecular mechanisms dictating small molecule control over protein stability. Hydrophobic tags, which reduce the thermal stability of HaloTag2, lead to increased binding by the quality control chaperone Hsp70. As refolding of the HaloTag2 protein fails, the HaloTag2 protein is ubiquitinated, presumably by the E3 ubiquitin ligase CHIP or Parkin, which are known to associate with Hsp70. Increased ubiquitination leads to enhanced binding to the proteasome and ultimately degradation of the fusion protein. HALTS1, on the other hand, stabilizes the HaloTag2 protein, rendering it less likely to interact with Hsp70, and as a result, HaloTag2 is protected from degradation via the ubiquitin-proteasome system. These observations could inform future studies on how therapeutic inhibitors that also stabilize or destabilize target proteins exert their effect. For instance, anti-androgen bicalutamide and anti-estrogen fulvestrant both lead to the degradation of their respective nuclear hormone receptors via the UPS23; 24. However, the molecular mechanisms leading to this degradation are unclear. Similar to our results, irreversible Her2 inhibitor, Canertinib, has been shown to promote the degradation of Her2 protein via increasing the binding of Her2 to Hsp7025. Given the direct role for Hsp70 in modulating HaloTag2 and Her2 stability in response to small molecules, we propose that Hsp70 client proteins could be particularly attractive targets for a hydrophobic tagging approach.</p><p>The control over HaloTag2 stability with HyT and HALTS1 molecules may allow for future modifications that add additional customization to assay design. Akin to reports with FKBP12, one could employ mutagenesis to skew the dynamic range of HaloTag2 stabilities with the two molecules. For instance, mutagenesis could yield a HaloTag2 mutant that would exhibit an even larger dynamic range. We envision that in addition to the HaloTag2 version reported here, one could design additional HaloTag2 variants for specific purposes.</p><p>In summary, we present here a system that allows researchers to modulate the levels of a protein of interest with a small molecule to either increase (HALTS1) or decrease (HyT) the protein's abundance in cells. The result is a general method to probe the physiological role of any target protein using small molecules.</p><!><p>HA-EGFP-HaloTag2, HA-EGFP-HaloTag7 and HA-EGFP were stably incorporated into FRT recombination-based HEK 293 Flp-In cells according to manufacturer's protocol (Invitrogen). All cells were grown at 37 °C in DMEM and supplemented with 10% FBS, 100 U/ml carbenicillin and 100 μg/ml streptomycin. HRASG12V was obtained from Addgene plasmid 9051 (provided by R. Weinberg) and cloned into a retroviral pEYK3.1 vector by excising EGFP. Retrovirus was generated in GP2-293 cells (Clontech) with a pVSV-G vesicular stomatitis virus retroviral vector and a corresponding pEYK plasmid, and NIH-3T3 cells were infected. Rabbit HA-specific (C29F4) and Hsp70 (D69) antibodies were purchased from Cell Signaling, and β-actin–specific antibody was purchased from Sigma (clone AC-74). HALTS1 was purchased from ChemBridge (9074451), YU101, HyT13 and HyT36 have been synthetized in our laboratory, and geldanamycin was purchased from InvivoGen. HyT36, HyT13, HALTS1, YU101, SW02 and geldanamycin were stored and aliquoted in DMSO as 1000x stock solutions.</p><!><p>The indicated cells were washed twice with cold PBS, and the cells were lysed in lysis buffer (1× PBS, 1% NP-40, 1 mM EDTA, 40 mM HEPES) with protease inhibitors. The lysates were cleared by centrifugation at 10,000g for 5 min. The total protein concentration was determined by Bradford assay, and 50 μg of protein was loaded onto an 8% Bis-Tris gel. To solubilize polyubiquitinated and aggregated proteins upon proteasome inhibition, samples were lysed with an SDS lysis buffer (1× PBS, 1% NP-40, 1% SDS, 1% sodium deoxycholate, 1 mM EDTA, 40 mM HEPES) with protease inhibitors and sonicated. The blots were processed by standard procedures with the indicated antibodies, and the band intensities were quantified by ImageJ (US National Institutes of Health).</p><!><p>HEK293 KG, KGH2, or KGH7 cells were seeded overnight on coverslips. They were then treated for 16 hours with media containing 10 μM HALTS1, 10 μM YU101, both, or vehicle alone (0.2% DMSO). Cells were then fixed with 4% paraformaldehyde, mounted on slides, and visualized for EGFP fluorescence.</p><!><p>The screen was carried out at the Yale Center for Molecular Discovery. HEK293 cell line stably expressing HaloTag2-luciferase was plated in 384-well plates at 10K cell/well. Twenty four hours later, a 15K ChemBridge DIVERSet or 20K Maybridge Diversity compound libraries were pinned onto plates at a final concentration of 10 μM. One hour later, 10 μM HyT13 was added. Luciferase assays with Steady Glo (Promega) were performed 24 hours later. For each plate, a control set of DMSO and HyT13 alone was included. The fraction of HaloTag2-luciferase remaining for each well was calculated as a ratio of luciferase activity for each well and the difference between DMSO and HyT13 treated wells. This ratio is plotted in Figure 1C as a percentage for the top 30 compounds preventing degradation.</p><!><p>HaloTag7 was amplified from pFN18A vector (Promega) and cloned into the pET-28a(+)vector (Invitrogen). The resulting plasmid was then transformed into BL21 (DE3) PLysS chemically competent E. coli (Invitrogen). Two liters of these bacteria were induced in exponential phase with IPTG (1 mM, 16 hr) at room temperature. The bacteria were resuspended in equilibration buffer used with the HisPur™ Cobalt Purification Kit (Thermo Scientific), spiked with 0.1% NP-40 and protease inhibitors. The bacteria were lysed with Branson Sonifier 450, and cellular debris was removed by centrifugation at 10,000g for 10 min. The recombinant protein was purified using the HisPur™ Cobalt Purification Kit (Thermo Scientific), then dialyzed 1:100 against PBS (Gibco) twice. Elutions and dialyzed proteins were assayed for purity using a Coomassie stain, and protein levels were quantified using the Bio-Rad Protein Assay.</p><!><p>Purified HaloTag7 protein was diluted in PBS to either 30 μM or 10 μM, and SYPRO Orange Protein Gel Stain stock (Sigma Aldrich) was diluted 1:1000 into the protein solution. A 1 μL volume of various compound concentrations was then placed into each well of a Low 96-well Clear Multiplate PCR Plate (Bio-Rad), 30 μL of the protein-SYPRO Orange mixture was placed into each well, and the plates were covered with Microseal 'B' Film (Bio-Rad). Real-time thermal cycling was then performed using the CFX96 Real-Time System (Bio-Rad). The reaction mixture was first equilibrated at 25 °C for 5 minutes. Thereafter, the temperature of the solution was increased by 1 °C and incubated for 5 sec repeatedly to generate a melting curve from 25-100 °C. Protein thermal stability was then quantified in terms of Hex fluorescence at each temperature of the melting curve. Raw data files were exported from the CFX software and formatted in Microsoft Office Excel. Data were trimmed to include the lower limit and upper limit values flanking the melting temperature and Tm values were calculated as the inflection point of the resulting sigmoidal curve using PRISM™ (GraphPad Software, Inc.).</p><!><p>Recombinant HaloTag7 was dialyzed 1:3000 into buffer containing 50 mM Tris (pH = 7.4) and 200 mM NaCl. HaloTag7 was then diluted with buffer, and DMSO was added to 2%. HALTS1, dissolved in DMSO, was added to buffer to a final concentration of 2% DMSO and 100 μM. Both solutions were filtered and degassed before the assay. An ITC200 calorimeter (MicroCal) was used to titrate compound into a sample well containing 200 μL of protein. Data was analyzed using Origin software.</p><!><p>Indicated cells were grown in 10-cm plates to 80% confluency and treated for 2 hours with vehicle, 10 μM HyT36 or 10 μM HALTS1. The cells were washed with cold PBS and lysed in HENG buffer (50 mM HEPES-KOH [pH 7.5], 150 mM NaCl, 20 mM Na2MoO4, 2 mM EDTA, 5% glycerol, 0.5% Triton X-100). The lysates were cleared by centrifugation at 10,000g for 5 min. Five hundred micrograms of protein was immunoprecipitated with 30 μL of anti-HA Affinity Gel (Sigma) at 4°C for 3 hours. The beads were washed 3 times with HENG buffer, and the protein complexes were released by boiling in sample buffer.</p><!><p>One hundred thousand NIH-3T3 cells infected with HA–HaloTag2–Hras1G12V at MOI=0.1 were plated onto 10-cm cell culture plates in 10% FBS with DMEM. The next day, the medium was replaced with DMEM containing 1% FBS and the cells were administered either vehicle, 1 μM HyT13 or 1 μM HALTS1. On day 4, the plates were photographed and the cells were harvested for immunoblotting.</p>

PubMed Author Manuscript

Root Herbivore Effects on Aboveground Multitrophic Interactions: Patterns, Processes and Mechanisms

In terrestrial food webs, the study of multitrophic interactions traditionally has focused on organisms that share a common domain, mainly above ground. In the last two decades, it has become clear that to further understand multitrophic interactions, the barrier between the belowground and aboveground domains has to be crossed. Belowground organisms that are intimately associated with the roots of terrestrial plants can influence the levels of primary and secondary chemistry and biomass of aboveground plant parts. These changes, in turn, influence the growth, development, and survival of aboveground insect herbivores. The discovery that soil organisms, which are usually out of sight and out of mind, can affect plant-herbivore interactions aboveground raised the question if and how higher trophic level organisms, such as carnivores, could be influenced. At present, the study of above-belowground interactions is evolving from interactions between organisms directly associated with the plant roots and shoots (e.g., root feeders - plant - foliar herbivores) to interactions involving members of higher trophic levels (e.g., parasitoids), as well as non-herbivorous organisms (e.g., decomposers, symbiotic plant mutualists, and pollinators). This multitrophic approach linking above- and belowground food webs aims at addressing interactions between plants, herbivores, and carnivores in a more realistic community setting. The ultimate goal is to understand the ecology and evolution of species in communities and, ultimately how community interactions contribute to the functioning of terrestrial ecosystems. Here, we summarize studies on the effects of root feeders on aboveground insect herbivores and parasitoids and discuss if there are common trends. We discuss the mechanisms that have been reported to mediate these effects, from changes in concentrations of plant nutritional quality and secondary chemistry to defense signaling. Finally, we discuss how the traditional framework of fixed paired combinations of root- and shoot-related organisms feeding on a common plant can be transformed into a more dynamic and realistic framework that incorporates community variation in species, densities, space and time, in order to gain further insight in this exciting and rapidly developing field.

root_herbivore_effects_on_aboveground_multitrophic_interactions:_patterns,_processes_and_mechanisms

Introduction<!><!>Impact of Root-Feeding Insects on Foliar Herbivores<!><!>Impact of Root-Feeding Nematodes on Foliar Herbivores<!>Interactions via Changes in Herbivore Induced Plant Volatiles<!>Interactions via Changes in Host Quality and Consequences for Parasitoid Behaviour<!><!>Interactions via Changes in Host Quality and Consequences for Parasitoid Behaviour<!>Incorporating Community Variation in Species, Densities, Space, and Time<!>Time of Arrival of Root and Shoot Herbivores<!>Spatial Distribution of Root Feeders<!>Herbivore and Parasitoid Preferences<!>Parasitoids and Effects Through Changes in the Habitat<!>Plant-Mediated Aboveground-Belowground Interactions in the Field<!>Belowground Influences of Aboveground Induced Defenses in the Field<!>Concluding Remarks<!>Open Access

<p>A central subject in terrestrial ecology is to understand the driving forces underlying the assemblage and functioning of plant-based communities. Within this field, the study of plant-insect interactions has played a pivotal role. Plant-insect interaction studies traditionally have focused on organisms that share a common domain, mainly aboveground. Aboveground herbivorous insects are the most speciose animal group on earth, and the intricate interactions with their host plants have fascinated ecologists for decades. In response to herbivory, plants often are defended by the production of or increase in the production of secondary plant compounds, phytotoxins, which impact the herbivore's feeding activity and/or development. These plant defense responses often result in increased mortality, reduced growth rates and fitness of the attacker (Schoonhoven et al., 2005). Herbivorous insects, on the other hand, have evolved ways that detoxify such deleterious plant chemicals. Increased plant resistance in response to herbivory is called induced direct plant defense. Concentrations of plant defense compounds do not only occur locally in the leaf subjected to herbivory, but often increase in other leaves as well. Such a systemic response enables the protection of the still undamaged leaves from the herbivore. As a consequence, this response also can influence the performance of other organisms that are feeding from the same plant, but at other locations or later in time. In response to herbivory and egg deposition, plants also emit volatile secondary metabolites, which can be used by natural enemies of the herbivores, for example insect parasitoids, to locate their hosts (Dicke and Sabelis, 1988; Turlings et al., 1990; Vet and Dicke, 1992; De Moraes et al., 1998; Dicke, 1999; Fatouros et al., 2008). This response, known as induced indirect plant defense, is beneficial for parasitoids, because these detectable plant cues can indicate the presence of their 'hard to detect' hosts (Vet et al., 1991). The plants subsequently benefit from reduced levels of herbivory due to increased top-down control. The phytotoxins consumed by herbivores often accumulate in tissues such as fat body and hemolymph, and via this mechanism plants may also negatively affect the fitness of the developing parasitoid larvae that consume the host herbivore. This exemplifies how plant defenses can cascade up trophic chains in complex ways (Harvey et al., 2003). Because herbivore-induced direct and indirect plant defenses mediate interactions between species within and between trophic levels, across space and time, they are considered a central force in assembling plant-based communities (Kaplan and Denno, 2007).</p><p>In the field, plants also are exposed to belowground consumers. In many terrestrial ecosystems, root-feeding nematodes and insects are the dominant belowground attackers. In the early 1990's, Masters et al. (1993) were among the first to report that root feeders can significantly alter interactions between plants and aboveground herbivores. This awareness of plant-mediated above-belowground interactions has brought a new level of complexity to the field of plant-insect ecology (Van der Putten et al., 2001; Bardgett and Wardle, 2003; Wardle et al., 2004). Interactive effects between plant consumers across domains have been explained by various induced plant responses, and a number of more recent studies indicate that these interactions often are mediated by herbivore induced plant defenses (reviewed in Bezemer and van Dam, 2005; Kaplan et al., 2008a; van Dam, 2009). In the early 2000's, the question was raised whether and how changes within the plant induced by root herbivores could cascade up influencing parasitoids of foliar herbivores (Bezemer et al., 2005; Soler et al., 2005; White and Andow, 2006; Rasmann and Turlings, 2007). Other studies focussing on the effects of soil-dwelling plant mutualists have shown that, for example, arbuscular mycorrhizal fungi, plant growth-promoting rhizobacteria, and decomposers also can affect the growth and development of foliar herbivores and their level of parasitism (Masters et al., 2001; Van der Putten et al., 2001; Gange et al., 2003; Wurst and Jones, 2003; Guerrieri et al., 2004; Hempel et al., 2009; Pineda et al., 2010; 2012).</p><p>In the present review, we focus on the impact of root-feeding insects and nematodes on aboveground insect herbivores and their parasitoids; the effects of belowground symbionts are reviewed elsewhere in this issue (Jung et al. 2012, this issue). We first discuss the conceptual models that have been put forward to explain plant-mediated effects of root herbivores on aboveground insect herbivores; changes in plant nutritional quality and in secondary chemistry, from altered concentrations of foliar phytotoxins to defense signaling. The effects of root herbivory on higher trophic levels aboveground are comparatively less explored, and because general patterns cannot yet be drawn we discuss cases that exemplify the magnitude of these effects. We end by proposing that a way to advance this field is to study above-belowground interactions within a more dynamic and complex spatial-temporal approach that includes insect mobility and spatial and temporal aspects in experimental designs. A new approach that goes beyond the relatively static interactions between pairs of organisms forced to feed on the same plant at a single density and time.</p><!><p>Plant-mediated effects of root-feeding insects on aboveground leaf chewers and phloem feeders. The aphid (left) represents aboveground phloem feeders, and the white caterpillar (right) represents leaf chewers. The grey caterpillar represents root-feeding insects. Effects of root herbivory can be positive (+) or negative (−) for overall aboveground insect performance, relative to insects on undamaged plants. Mechanisms that have been put forward to explain these plant-mediated effects are induced changes in shoot nutritional quality (1), shoot secondary chemistry (2 and 3), and hydraulic leaf changes (4). Numbers indicate each of the proposed hypotheses discussed in the text</p><!><p>Other studies that have examined the effects of root herbivores on aboveground leaf chewers have reported negative effects, showing that besides facilitation, plant-mediated competition also is common in aboveground-belowground interactions (Tindall and Stout, 2001; Bezemer et al., 2003; van Dam et al., 2003, 2005; Soler et al., 2005; Staley et al., 2007). The frequently observed negative impact of root herbivory on leaf chewer fitness has been explained by the 'Defense Induction Hypothesis' (Bezemer et al., 2003). This hypothesis states that above- and belowground insect herbivores influence each other via induced changes in secondary plant compounds (Fig. 1, ②). Insects that feed from the phloem are less exposed to secondary plant compounds, since phytotoxins generally are stored in cells (Larsson, 1989). This can explain why aboveground aphids usually are not negatively affected by root herbivory. In this view, root-chewing insects induce an increase in foliar secondary plant compounds, which negatively affects the performance of leaf chewers without affecting phloem feeders (reviewed in Bezemer and van Dam, 2005; Johnson et al., 2008; Kaplan et al., 2008a; van Dam and Heil, 2011).</p><p>There has been a significant development in the understanding of the molecular mechanisms underlying local and systemic induced plant defenses triggered by pathogens and insects aboveground (Kessler and Baldwin, 2002; Zheng and Dicke, 2008; Pieterse et al., 2009). This has enabled the exploration of induced plant defenses beyond measuring changes in nutrients and phytotoxins, thus providing a basis to mechanistically understand plant-mediated interactions. Generally, leaf-chewing insects such as caterpillars cause a response in the plant that triggers the jasmonic acid (JA) signaling pathway, while phloem-feeding insects such as aphids induce the salicylic acid (SA) signaling pathway. Although the majority of studies have focused on signaling responses in the foliage in response to shoot attack, these responses also occur in the roots (reviewed in Erb et al., 2009a). It has been shown that jasmonates can be transported from shoots to roots (Baldwin et al., 1994), showing how long distance defense signaling can occur across roots and shoots. The transport of jasmonates from roots to shoots can explain why root-feeding insects may negatively impact the performance of foliar insect herbivores, because JA in the roots is transported to/activated in the shoots (Fig. 1, ③).</p><p>Jasmonic acid and salicylic acid often act antagonistically, and increases in the levels of one of the phytohormones can interfere with the activity of other phytohormones (Pieterse and van Loon, 1999; Engelberth et al., 2001; Kessler and Baldwin, 2002; Koornneef et al., 2008; but see e.g., Schenk et al., 2000; Van Wees et al., 2000 that report synergistic interactions). If this so-called cross-talk between pathways (Pieterse et al., 2009) also occurs across plant organs, root herbivory can cause a reduction in SA- related defenses in the foliage by inducing JA-related defenses as proposed by Van der Putten et al. (2001). This can provide an alternative explanation for the frequently observed increased performance of phloem feeders on plants previously attacked by root-feeding insects. However, in Zea mays (maize) plants, neither JA nor SA were found to be induced in the shoots by the rootworm Diabrotica virgifera (Erb et al., 2009b). Interestingly, leaves of root-infested maize plants had reduced leaf water contents and increased levels of abscisic acid (ABA) (Erb et al., 2011a).</p><p>Reduced resistance to leaf chewers has been reported on ABA-deficient plants (Thaler and Bostock, 2004; Bodenhausen and Reymond, 2007), leading the authors to hypothesize that, in Z. mays, increased resistance to leaf chewers in plants with root herbivory is due to induced ABA signaling and/or hydraulic changes in the leaves (Erb et al., 2011a). Abscisic acid is involved in a number of physiological adaptations of plants to drought stress, and it can act as a chemical signal that controls the opening and closing of stomata. It might be difficult then to disentangle the effects of changes in ABA and leaf water content on foliar herbivores. Interestingly, the negative effects on the leaf chewer were still observed after ABA signaling was inhibited. More studies that explore defense signaling that cross the border between the below- and aboveground domains are needed to understand the mechanistic basis that mediate these interactions (Erb et al., 2009a).</p><p>Knowledge about the molecular mechanisms underlying plant defenses is derived from a limited number of model plants species from genetic and molecular biology (Felton and Korth, 2000; Stout et al., 2006; Wang et al., 2008; but see Wu and Baldwin, 2010; Broekgaarden et al., 2010), and often herbivory is simulated by using exogenous applications of JA and SA (e.g., Spoel et al., 2003; Koornneef et al., 2008; Leon-Reyes et al., 2010; but see e.g., Kessler et al., 2004). Consequently, extrapolations into ecologically representative scenarios have to be taken with caution. Studies with natural communities are needed to determine the full ecological and evolutionary consequences of above-belowground multitrophic interactions.</p><!><p>Plant-mediated effects of root-feeding nematodes on aboveground aphids. The aphid represents aboveground phloem feeders, and the black circles and curved lines represent ecto- and migratory endoparasitic nematodes and root-knot or cyst-forming nematodes, respectively. Effects of herbivory by nematodes on aphid fitness are mostly negative (−) relative to that on undamaged plants. Mechanisms that have been put forward to explain these negative effects are induction of common defense signaling (1), competition for assimilates in the phloem (2), and reduced amino acid concentration in the phloem (3). Numbers indicate each of the proposed hypotheses discussed below</p><!><p>In contrast to root feeding by insects, which often facilitate the growth and development of aphids, studies on feeding by nematodes consistently report negative effects on aphid performance (Bezemer et al., 2005; Wurst and Van der Putten, 2007; Kaplan et al., 2009, 2011; Hol et al., 2010; Vandegehuchte et al., 2010; Kabouw et al., 2011). Nematode-caterpillar interactions are less well-studied, and positive (Alston et al., 1991; Kaplan et al., 2008b), neutral (Wurst and Van der Putten, 2007), and negative effects (van Dam et al., 2005) have been reported. We will, therefore, focus on the mechanisms that have been proposed to link the consistent negative impact of nematodes on aphid fitness. The first proposed explanation was that nematodes and phloem feeders trigger a common defense signaling pathway (Kaplan et al., 2009). This hypothesis is based on studies that showed that in Solanaceae, the defense gene Mi-1 mediates resistance to both root-knot nematodes and aphids (Li et al., 2006; Bhattarai et al., 2007). Thus, aboveground phloem feeders and root-feeding nematodes might be inducing similar defense pathways in plants (Fig. 2, ①). Subsequent studies have shown that although Mi-1 mediates resistance to both nematodes and phloem feeders/sap suckers, it is involved in the activation of distinct signaling pathways. Therefore, the Mi-1 defense gene may contribute differently to the resistance to aphids and nematodes (Mantelin et al., 2011). There is no empirical evidence yet that links the reduced performance of phloem feeders on plants exposed to nematodes with changes in levels of phytohormones or defense marker genes.</p><p>More recently, Kaplan et al. (2011) empirically tested the 'Sink Competition Hypothesis', which proposes that aboveground phloem feeders and root-feeding nematodes compete for assimilates in the phloem. Root-knot nematodes and aphids feed from vascular tissues and attract photoassimilates to their feeding site. Therefore, the pressure-driven transport in the phloem sieve elements can be re-directed towards root-feeding nematodes or aphids, and thus both can act as a nutrient sink for the plant (Guerrieri and Digilio, 2008). Thus, when nematodes colonize the roots of the plant earlier than aphids, the sink created by nematodes in the roots may compete with the subsequent sink that aphids will initiate in the shoots (Fig. 2, ②). Empirical evidence for this potential mechanism is lacking (but see Inbar et al., 1995; Larson and Whitham, 1997 for evidence supporting the hypothesis in aboveground plant-herbivore interactions). Especially cyst- or gall-forming species are able to feed from the phloem, which makes them potential competitors of aphids. It is noteworthy that aphids also perform suboptimally on plants infested by migratory endoparasitic species that do not create nutrient sinks within the plant (e.g., Wurst and Van der Putten, 2007). The concentration of amino acids in the phloem of plants infested by root-feeding nematodes also has been reported to be lower than on plants without nematodes, and this change correlated with the reduced aphid fitness that was observed (Bezemer et al., 2005). More studies are needed to confirm how widespread this mechanism is.</p><!><p>In the early 2000's, the question was raised whether soil-dwelling organisms also could affect parasitoids of aboveground herbivores. The first studies focused on parasitoid host-plant preferences, and all reported that the level of attraction of female parasitoids was increased when plants were exposed to soil-dwelling organisms, independently of the soil functional group triggering the effect. Therefore, it was proposed initially that soil organisms, independent of whether they were root antagonists or plant beneficials, would all benefit host-parasitoid interactions (e.g., Masters et al., 2001; Gange et al., 2003; Wurst and Jones, 2003; Guerrieri et al., 2004). However, a potential mechanism responsible for the increase in host plant preference was not provided in these studies. Considering that in aboveground systems, parasitoid host-searching is guided primarily by volatile cues that are produced by the host-infested plant (Dicke, et al., 1990; Turlings, et al., 1990; Vet and Dicke, 1992), herbivore-induced plant volatiles were a primary candidate to test. Subsequent studies have shown that the composition of the volatile blend induced by foliar herbivores can be affected by root-feeding insects. The result is that the plant becomes less attractive to female parasitoids foraging for hosts (Rasmann and Turlings, 2007; Soler et al., 2007a). In these studies, root-feeding by insects clearly interfered with host-parasitoid interactions. Other studies also have shown that volatiles emitted by plants exposed to both foliar- and root-feeding insects can be quantitatively and qualitatively different from blends emitted by plants exposed to each herbivore in isolation (Olson et al., 2008; Pierre et al., 2011). It is well-established that specialist parasitoids can distinguish between plants attacked by their hosts and plants attacked by non-hosts by exploiting differences in induced plant volatiles (de Moraes et al., 1998). It is less clear, however, what can happen when the same plant is exposed to multiple host and non-host herbivores of the parasitoid (but see Shiojiri et al., 2001, 2002; Vos et al., 2001; Rodriguez-Soana et al., 2002; 2005; Zhang et al., 2009; Dicke et al., 2009; Erb et al., 2010), especially when these herbivores feed from roots and shoots.</p><!><p>Parasitoid larvae are highly susceptible to changes in the quality of the internal biochemical environment provided by their hosts, and thus are tightly linked to host development (Harvey, 2005). As root herbivores can influence the growth and development of aboveground insect herbivores via induced changes in foliar secondary chemistry, these effects also could affect the developing parasitoid larvae. A number of studies have shown that root herbivore effects can even be stronger for the developing parasitoid larvae than for the herbivore itself (Bezemer et al., 2005; Soler et al., 2005, but see Kabouw et al., 2011 where no effects were observed). These effects can cascade up to at least the fourth trophic level influencing hyperparasitoid fitness (Soler et al., 2005).</p><!><p>Effects of root feeding insects (a) and nematodes (b) on parasitoid performance, behavior, and/or changes in plant volatiles. Rch: root-chewer, rk: root-knot, sp: seed predator, lch: leaf chewer, and pf: phloem feeder</p><p>Root-feeding insects and aboveground parasitoids. A case study. a Percentage of Brassica nigra plants with foliar-feeding Pieris brassicae hosts selected for oviposition by females of the parasitoid Cotesia glomerata. The size of the parasitoid reflects its relative performance on plants without (white bars) and with (grey bars) Delia radicum root-feeding larvae. b Glucosinolate (sinigrin) level in young leaves of B. nigra plants (white dotted squares) and plants infested by D. radicum (grey squares). c Canonical discriminant plot showing sample scores based on volatile blends of B. nigra plants (1) without herbivores (2) with Pieris brassicae larvae, (3) with Delia radicum larvae and (4) with both herbivores. Each circle represents a sampled plant. Beta-farnesene and dimethyl-nonatriene are known attractant compounds (white arrows) for insect parasitoids, while sulfides are known repellent volatiles (grey arrows) for insects; the size of the arrows represents the relative amount of the compounds in the blends of the plants with root- and foliar-feeding insects. Summary from R. Soler PhD Thesis, Netherlands Institute of Ecology, 2007 (reprints of the thesis can be requested by e-mail)</p><!><p>Innate responses of foraging parasitoids to plant odors can change with experience, leading to local or temporary specialization and enhancement of foraging success (Turlings et al., 1990; Vet et al., 1995). Parasitoids have the ability to learn to distinguish between volatile blends emitted by plants infested by their hosts versus plants infested with their hosts and root-feeding insects (Rasmann and Turlings, 2007). Therefore, they could regain attraction for hosts feeding on root-infested plants with experience (Rasmann and Turlings, 2007). Yet, the effects of parasitoid learning in this process need to be explored. The role of parasitoid learning in dealing with natural variation in plant and host quality and plant volatiles induced by root herbivory remains largely unstudied.</p><!><p>Thus far, the majority of above-belowground interaction studies that involve plants, insects, mutualistic symbionts, and natural enemies have encompassed relatively little variation in number of players and in environmental conditions. Here, we review studies that are extending this scope by bringing in effects of time, space, behavior, and habitat conditions. We identify this as the direction of future studies in the area of above-belowground multitrophic interactions.</p><!><p>The sequence of arrival of above- and belowground herbivores on a plant can greatly affect the outcome of the interaction (Maron, 1998; Blossey and Hunt-Joshi, 2003). The leaf chewer Spodoptera fugiperda, for example, had a significant negative effect on the colonization of the root chewer Diabrotica virgifera when first colonizing the plant, but the aboveground herbivore did not influence the performance of the root feeder when arriving later than the root herbivore (Erb et al., 2011b). The sequence of arrival also has been shown to be an important determinant of plant responses at the gene level. Transcriptional changes, for example, have been shown to differ significantly for sequential and simultaneous attack of aboveground leaf chewers and phloem feeders (Voelckel and Baldwin, 2004). Similarly, the expression of SA- and JA-related genes has been found to differ in response to individual and simultaneous shoot attack by insect herbivores from contrasting feeding-guilds (Zhang et al., 2009; Soler et al., 2012). Aboveground insect herbivores that feed on a plant already infested by root feeders are expected to be inevitably confronted with higher levels of phytotoxins, and thus potential fitness costs (Bezemer and van Dam, 2005). This idea is based on studies with Gossypium herbaceum, cotton plants, that showed that in response to root herbivory levels of secondary compounds increased along the entire shoot (Bezemer et al., 2004). However, it is not clear how widespread this response can be. For example, a subsequent study in which B. nigra plants were exposed to root herbivory showed that levels of secondary compounds were increased only in young leaves in response to root feeding, but that they did not change in mature and old leaves (Soler et al., 2005). More studies that record changes in secondary chemistry in response to root herbivory that compare both young and old leaves are needed to determine how common this phenomenon is.</p><!><p>Besides the mere presence or absence of root feeders on the plant, the spatial distribution of root-infested plants in a habitat can be of crucial importance. Evidence for this assumption is provided by a field study where the specialist aphid Brevicoryne brassicae preferred to feed and reproduce on B. nigra plants without root herbivores over plants infested by the root herbivore D. radicum. This preference was observed only when plants with root herbivores were grouped in clusters. When the plants with and without root herbivores were placed in a mixed design, aphids no longer differentiated (Soler et al., 2009). This shows that the spatial arrangement of root herbivores in the field also can be an important factor determining the amount of aboveground herbivory. However, as discussed in the previous section, it remains unknown whether root feeders uniformly influence the secondary chemistry of the entire shoot or if these changes are restricted to certain parts of the shoot. In response to aboveground insects, for example, phytotoxins often increase in certain tissues, e.g., young leaves, rather than uniformly along the shoot, thus allowing secondary attackers to scape potential fitness costs by avoiding feeding on theses leaves (Stout et al., 1996). When root induced plant responses are expressed only in certain parts of the shoot, only the aboveground herbivores that feed on these parts are expected to be influenced by root feeders (Kaplan et al., 2008c).</p><!><p>Most above-belowground studies are based on non-choice experiments where the survival, growth, and development of caterpillars or aphids on plants with or without root herbivores are compared. Foliar herbivores, however, can precisely select plants for oviposition and feeding. Where free choices can be made, aboveground insect herbivores can avoid or prefer plants that are already colonized by root feeders. Optimal oviposition theory predicts that females of herbivorous arthropods with offspring with limited mobility, such as butterflies, will evolve to select those host plants for oviposition on which their offspring perform best thus maximizing their fitness (Jaenike, 1990). Considering that plants attacked by root-feeding insects often represent a suboptimal food source for leaf chewers, butterflies should avoid plants with root herbivores and select uninfested conspecifics if these represent fitness costs (Soler et al., 2010). When such avoidance occurs, this also will be beneficial for the plant by reducing the probability of root-damaged plants being simultaneously attacked belowground and aboveground. The same approach might apply belowground, and there are studies, for example on root-feeding nematodes, where the presence of potential enemies may direct attackers away from potential feeding sites (Piskiewicz et al., 2009).</p><p>Adding effects on the reduced preferences that natural enemies of herbivores can show for hosts feeding on plants also attacked by root herbivores (Rasmann and Turlings, 2007; Soler et al., 2007a) will show the complex dimensions of the ecological 'dilemma' for leaf-chewing insects with respect to root-infested host plants. The evolutionary choice would be between growing more slowly and/or attaining a smaller size but benefitting from a smaller probability of being found by natural enemies on root-infested plants, or optimizing performance at the cost of running a higher risk of parasitism or predation on root-uninfested healthy plants. From the plant's point of view, the benefits of acting as a communication channel between root- and foliar-feeding herbivores that attenuates simultaneous infestations is then counterbalanced by interferences with the indirect defense system of the plant that reduces the attraction of natural enemies of the herbivore. If and how above- and belowground herbivores may integrate all this information in their "decision-making" remains to be elucidated.</p><!><p>Interactions between root feeders and parasitoids are not restricted to interactions on a single plant. For example, root herbivores can influence host-parasitoid interactions aboveground via their effects on changes in the structure of the plants. In Z. mays, the percentage of parasitism of the European corn borer, Ostrinia nubilalis, by its specialist parasitoid Macrocentrus grandii was significantly reduced in the presence of the corn rootworm Diabrotica virgifera in the habitat (White and Andow, 2006). Plant height and density were reduced in habitats where the rootworm was present, resulting in more open habitats that are less preferred by female parasitoids of this species. Interestingly, this positive indirect interaction, known as associational resistance, in which one species gains protection from its consumer by association with a competitor, has been widely documented in plants (Andow, 1991), but not among insects. Root herbivores also can influence host-parasitoid interactions aboveground via changes in the quality of the surrounding environment triggered by belowground insects. Females of the parasitoid Cotesia glomerata found their hosts on focal plants much faster in situations when neighboring plants were exposed to root herbivory, than when neighboring plants were kept undamaged (Soler et al., 2007b). In that study, the microhabitat was composed of root-damaged and root-undamaged plants of the same species that all had similar size and height, which minimizes the influence of physical plant characteristics on the foraging wasps (McCann et al., 1998; Gols et al., 2005).</p><!><p>A number of studies have shown that the abundance or preference of aboveground organisms, such as herbivores, pollinators, predators, or parasitoids, on plants growing in natural or agricultural systems can be affected by whether the plant is also exposed to root herbivory (e.g., Masters, 1995; Poveda et al., 2003; Hunt-Joshi and Blossey, 2005; Staley et al., 2007; Wurst et al., 2008; Kaplan et al., 2009; Soler et al., 2009). Most of these studies have used potted plants with or without root herbivory that are placed in the field (e.g., Poveda et al., 2003; Wurst et al., 2008; Soler et al., 2009). However, several studies have manipulated aboveground and belowground herbivory in the field that show that root herbivory by insects or nematodes can affect aboveground multitrophic interactions under natural conditions (e.g., Blossey and Hunt-Joshi, 2003; White and Andow, 2006; Kaplan et al., 2009), while others have not detected a significant effect (Hladun and Adler, 2009; Hong et al., 2011; Heeren et al., 2012). Interestingly, two recent independent studies report that there are no significant interactions between soybean cyst nematodes and aphids in soybean fields (Hong et al., 2011; Heeren et al., 2012). In contrast, greenhouse studies with soybean plants have reported that the performance of soybean aphids is significantly influenced by cyst nematodes (e.g., Hong et al., 2010). These results indicate that care needs to be taken when extrapolating results from greenhouse and common garden experiments to real field situations, and emphasize the urgent need for more realistic above-belowground studies.</p><!><p>Another issue that remains largely unresolved is how important the effects of root herbivory on aboveground induced plant defense responses are for plants that are growing in the field and are interacting with multiple antagonists, mutualists, decomposers, and other plants simultaneously. Most field studies that examine root herbivore effects on aboveground plant-insect interactions do not report effects on secondary plant compounds or emission of volatiles. However, a recent study by Megias and Muller (2010) shows that exposure to root herbivory in field-grown brassicaceous plants (Moricandia moricandioides) led to significant changes in aboveground glucosinolate profiles, and that these differences correlate with changes in the composition of the aboveground food web on these plants. This study shows clearly that root induced changes in aboveground plant secondary compounds can be of significant importance in the field. Similarly, Hladun and Adler (2009) showed that Cucurbita moschata plants, butternut squash, grown in the field had increased floral nectar concentrations when exposed to root herbivory. This can subsequently affect pollinators, but also parasitoids and predators in the field. As there is now a considerable number of studies that have shown that levels of parasitism and predator abundance in the field can be affected by root herbivory (e.g., Masters et al., 2001; White and Andow, 2006; Soler et al., 2009), it is quite possible that root herbivory indeed affects aboveground indirect induced defense responses in the field. Further field-based studies are needed in order to determine how these interactions can influence, or are influenced by, species diversity and community structure. How important indirect plant defense responses can be in the field (Obermaier et al., 2008), and how this is affected by root herbivory remains to be explored.</p><!><p>It is evident that root feeders can be important players in aboveground plant-based communities, via their effects on direct and indirect defenses of plant shoots that can cascade up to at least the fourth trophic level. Knowing this, the new challenge is to study above-belowground interactions under more realistic conditions. This will bring us closer to the detection of mechanisms with evolutionary potential and patterns that can be used in practice, for example when attempting to enhance sustainable pest control. It is puzzling why root-feeding insects and nematodes are still playing a minor role in the studies of contemporary community, behavioral, chemical, and molecular ecology. Currently, the notion of 'out of sight, out of mind' is no longer a valid argument for leaving out root feeders!</p><!><p>This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.</p>

PubMed Open Access

Site-Specific Glycosylation Patterns of the SARS-CoV-2 Spike Protein Derived From Recombinant Protein and Viral WA1 and D614G Strains

The SARS-CoV-2 spike protein is heavily glycosylated, having 22 predicted N-glycosylation sites per monomer. It is also O-glycosylated, although the number of O-glycosites is less defined. Recent studies show that spike protein glycans play critical roles in viral entry and infection. The spike monomer has two subdomains, S1 and S2, and a receptor-binding domain (RBD) within the S1 domain. In this study, we have characterized the site-specific glycosylation patterns of the HEK293 recombinant spike RBD and S1 domains as well as the intact spike derived from the whole virus produced in Vero cells. The Vero cell-derived spike from the WA1 strain and a D614G variant was analyzed. All spike proteins, S1, and RBDs were analyzed using hydrophilic interaction chromatography (HILIC) and LC-MS/MS on an Orbitrap Eclipse Tribrid mass spectrometer. N-glycans identified in HEK293-derived S1 were structurally diverse. Those found in the HEK293-derived RBD were highly similar to those in HEK293 S1 where N-glycosites were shared. Comparison of the whole cell-derived WA1 and D614G spike proteins revealed that N-glycosites local to the mutation site appeared to be more readily detected, hinting that these sites are more exposed to glycosylation machinery. Moreover, recombinant HEK293-derived S1 was occupied almost completely with complex glycan, while both WA1 and D614G derived from the Vero E6 cell whole virus were predominantly high-mannose glycans. This stands in stark contrast to glycosylation patterns seen in both CHO- and HEK cell-derived recombinant S1, S2, and the whole spike previously reported. Concerning O-glycosylation, our analyses revealed that HEK293 recombinant proteins possessed a range of O-glycosites with compositions consistent with Core type 1 and 2 glycans. The O-glycosites shared between the S1 and RBD constructs, sites T323 and T523, were occupied by a similar range of Core 1 and 2 type O-glycans. Overall, this study reveals that the sample nature and cell substrate used for production of these proteins can have a dramatic impact on the glycosylation profile. SARS-CoV-2 spike glycans are associated with host ACE2 receptor interaction efficiency. Therefore, understanding such differences will serve to better understand these host–pathogen interactions and inform the choice of cell substrates to suite downstream investigations.

site-specific_glycosylation_patterns_of_the_sars-cov-2_spike_protein_derived_from_recombinant_protei

Introduction<!>Recombinant Proteins and Intact Viruses<!>Protein Digestion<!>Glycopeptide Enrichment by the HILIC Resin<!>Reversed-Phase HPLC Fractionation<!>Site Occupancy Analysis<!>Mass Spectrometry Analysis<!>Data Analysis Using Byonic and Manual Verification<!>Model Construction<!>Mapping Glycosylation on Recombinant RBD Proteins<!><!>Mapping Glycosylation on Recombinant RBD Proteins<!>Site-Specific Microheterogeneity of Spike Glycosylation in Recombinant S1 Proteins<!><!>Site-Specific Microheterogeneity of Spike Glycosylation in Recombinant S1 Proteins<!><!>Site-Specific Microheterogeneity of Spike Glycosylation in Recombinant S1 Proteins<!><!>Site-Specific Microheterogeneity of Spike Glycosylation in Recombinant S1 Proteins<!><!>Site-Specific Microheterogeneity of Spike Glycosylation in Recombinant S1 Proteins<!>Site-Specific Glycosylation of the Spike From the WA1 Strain and the D614G Variant<!><!>Site-Specific Glycosylation of the Spike From the WA1 Strain and the D614G Variant<!><!>Site-Specific Glycosylation of the Spike From the WA1 Strain and the D614G Variant<!>Conclusion<!>Author Summary

<p>SARS-CoV-2 is an enveloped, positive single-stranded RNA beta coronavirus expressing four main structural proteins, which include nucleocapsid, spike, membrane, and envelope proteins (Lu et al., 2020; Singh and Yi, 2021). The trimeric spike protein is the major surface protein of the SARS-CoV-2 virus and serves as an entry protein for host cell infection (Shang et al., 2020). To facilitate the fusion of the viral membrane with the infected cells, the spike proteins are cleaved into S1 and S2 subunits by cellular proteases, such as furin (Hoffmann et al., 2020a; Hoffmann et al., 2020b; Xia, 2021). The S1 subunit contains the N-terminal and the receptor-binding domain (RBD) (Xia, 2021), and recombinant RBD binds to the human angiotensin converting enzyme 2 (ACE2) present as a surface receptor on host cells. The S2 domain serves the function of membrane fusion, which contains a fusion peptide (FP), an internal fusion peptide (IFP), two heptad repeat domains (HR1 and HR2), a transmembrane domain, and a C-terminal domain (Coutard et al., 2020; Xia, 2021).</p><p>Enveloped viruses have evolved to take advantage of many host cell processes including glycosylation (Watanabe et al., 2019). Viral protein glycosylation functions in a number of ways in the viral lifestyle including viral particle stability, mediating viral infection, host immune response, and immune evasion (Walls et al., 2016; Watanabe et al., 2019). Viral glycosylation of key envelope glycoproteins can be dynamic over time as the virus propagates through the host population, allowing immune avoidance to evolve over time (Li et al., 2021). Recent cryo-EM studies reported that the recombinant SARS-CoV-2 spike protein is extensively glycosylated (Grant et al., 2020; Wrapp et al., 2020). Using recombinant proteins, earlier studies reported glycosylation of the 22 predicted N-linked glycosites in the spike protein at high occupancy and lower glycosylation occupancy on O-linked glycosites (Watanabe et al., 2020a; Watanabe et al., 2020b; Shajahan et al., 2020; Zhao et al., 2020; Zhou et al., 2021). A recent study reported that glycosylation is essential for SARS-CoV-2 viral entry and infection (Yang et al., 2020). Since glycans are produced through a stochastic process that is dependent upon glycosylation, enzyme expression, location, concentration, and the particular glycoprotein's sequence and structural characteristics, it can be altered under selective pressure. During viral evolution, with passage through the human population, glycosites are added and deleted often, leading to an increased number of sites and glycan complexity. The overall glycosylation characteristics such as composition, subclass, heterogeneity, and density over the surface of the protein can have dramatic effects on viral survival, transmission, and immune evasion (Vigerust and Shepherd, 2007; Watanabe et al., 2019; Li et al., 2021). Spike glycoproteins are often the major target for vaccine design and antivirus drug development. Understanding the glycosylation microheterogeneity of the spike protein can facilitate the process.</p><p>Here, we characterize site-specific glycosylation on recombinant RBD and the S1 domain of the spike protein produced in HEK293 cells to understand the glycosylation microheterogeneity produced using this cell substrate. The question remains open: whether the glycosylation of these recombinant proteins differs from that of the native spike produced in the whole virus. Thus, we compare the glycosylation of recombinant RBD and S1 to two intact viruses, the WA1 strain and a D614G variant, both produced in Vero E6 cells. The SARS-CoV-2/USA-WA1/2020 (USA-WA1) viral strain was isolated from the specimen of the first confirmed case in the United States (Harcourt et al., 2020; Wang et al., 2021a). Whole genome sequencing confirmed that this strain contained D614 as the original form of the SARS-CoV-2 virus. SARS-CoV-2/Massachusetts/VPT1/2020 (MA/VPT1), containing the D614G mutation, was isolated in Vero E6 cells from a nasopharyngeal specimen collected in April 2020 (Wang et al., 2021a). The D614G mutation, which appeared in early 2020 (Korber et al., 2020), has become dominant worldwide. The D614G mutation is also carried by the more recent and concerning SARS-CoV-2 variants, including B.1.1.7, B.1.351, P.1, and B.1.617 (https://www.cdc.gov/coronavirus/2019-ncov/variants/). Compared to strains containing the original D614, viruses with the D614G mutation have significantly higher infection titers as well as faster transmission but are less sensitive to spike-based SARS-CoV-2 vaccine sera produced in mice, non-human primates, and humans (Hou et al., 2020; Korber et al., 2020; Yurkovetskiy et al., 2020). In addition, structural analysis demonstrates that the G614 spike is in a more open conformation with extended RBDs (Yurkovetskiy et al., 2020). Given this conformational shift, it is of interest to examine glycosylation for possible changes in the D614G spike compared to its close progenitor, the WA1 strain, while keeping the viral propagation cell platform the same. Therefore, in addition to analysis of recombinant spike constructs, we report the glycosylation patterns of spikes in WA1 and D614G strains produced by the whole virus in Vero E6 cells. Our results may aid in interpretation of experimental data concerning spike interactions with the host and surrogates as well as the development of therapeutics and vaccines against the SARS-CoV-2 virus.</p><!><p>Recombinant protein SARS-CoV-2 Spike S1 and RBD proteins expressed in HEK293 cells were purchased from Sanyou Bio (China). The whole virus of the WA1 strain was from the first patient of SARS-CoV-2 virus infection in the United States (Harcourt et al., 2020). The virus was isolated from nasopharyngeal and oropharyngeal specimens from this patient, and the viral sequence was confirmed (Wang et al., 2021a). This strain, SARS-CoV-2/USA-WA1/2020 (USA-WA1), is the original form of the SARS-CoV-2 virus without mutation at the 614 amino acid (Li et al., 2021). The D614G variant carrying the spike protein amino acid change at 614D to G, SARS-CoV-2/Massachusetts/VPT1/2020 (MA/VPT1), was isolated from Vero E6 cells from a nasopharyngeal specimen collected in April 2020 (Wang et al., 2021a). Both viruses were grown in Vero E6 cells, and the supernatant of the passage 4 stock of each virus was collected by centrifugation. After the viruses were frozen to −80°C at least overnight, the viruses were inactivated by gamma irradiation.</p><p>Chemicals, reagents, and TSKgel amide 80 particles were purchased from Tosoh Bioscience LLC (Montgomeryville, PA). Sep-Pak C18 cartridges were purchased from Waters (Milford, MA, United States). Sequencing grade-modified trypsin, chymotrypsin, and Glu-C were purchased from Promega Corp. (Madison, WI). PNGase F was purchased from New England BioLabs, Inc (Ipswich, MA). Iodoacetimide, dithiothreitol (DTT), trifluoroacetic acid (TFA) (≥99%), and other chemicals were purchased from Sigma-Aldrich (St. Louis, MO, United States). Solvents were of high-pressure liquid chromatography (HPLC) grade or higher and purchased from Thermo Fisher Scientific (Waltham, MA). All other reagents were of ACS grade or higher.</p><!><p>Recombinant proteins (200 µg) were dissolved in 50 mM ammonium bicarbonate at a concentration of 2 µg/µL. DTT was added to reduce the disulfide bonds at a final concentration of 5 mM for 30 min at 60°C. The samples were cooled to room temperature (RT), and iodoacetamide (IAA) was added to alkylate the reduced cysteine residues at a concentration of 15 mM for 30 min in the dark at RT. DTT was added to 25 mM to neutralize the remaining IAA. Trypsin or chymotrypsin was added (enzyme/protein, 1:50, w/w), and the samples were incubated at 37°C overnight.</p><p>For intact viruses, approximately 500 µg of the proteins was reduced with 5 mM DTT in 6M urea and 50 mM NH4HCO3 for 1 h at 37°C and subsequently alkylated with 15 mM iodoacetamide for 30 min at RT in the dark. To neutralize the remaining IAA, DTT was added to 25 mM. Samples were diluted 6-fold with 50 mM NH4HCO3 and 1 mM CaCl2 and digested with sequencing grade-modified trypsin or chymotrypsin at 1:50 (enzyme/protein, w/w) overnight at 37°C. Glu-C was added to the tryptic digest at 1:50 (enzyme/protein, w/w) and incubated overnight at 37°C.</p><!><p>Intact glycopeptides were enriched by solid-phase extraction using the TSKgel amide 80 hydrophilic interaction chromatography (HILIC) resin according to our previous report (An and Cipollo, 2011). Briefly, 200 mg (400 µL of the wet resin) of the amide 80 resin was placed into a Supelco fritted 1 ml column and washed with 1 ml of 0.1% trifluoroacetic acid (TFA)–water solution. The column was conditioned with 1 ml of 0.1% TFA–80% ACN. The peptides were suspended in 0.1% TFA–80% ACN and slowly loaded to the column. The hydrophobic species were washed away with 3 ml of 0.1% TFA–80% ACN. For recombinant proteins, the glycopeptides were eluted with 1 ml of 0.1% TFA–50% ACN and 1 ml of 0.1% TFA–25% ACN. The eluents were combined and vacuum-dried. For whole viruses, the glycopeptides were eluted sequentially with 1 ml of 0.1% TFA–65% ACN, 0.1% TFA–60% ACN, 0.1% TFA–50% ACN, and 0.1% TFA–25% ACN. Each eluent was vacuum-dried and analyzed by mass spectrometry separately.</p><!><p>The PNGase F-treated WA1 and D614G peptides were dried by speed vacuum and resuspended in 20 μl of 10 mM TEAB. The fractionation of the peptide samples is carried out using an Agilent Poroshell 120 Column (2.7 μm, 2.1 × 150 mm) and an Agilent UHPLC 1290 system. The separation was performed by running a gradient of Solvent B (10 mM TEABC, pH 8.0, 90% ACN) and Solvent A (10 mM TEAB, pH 8.0) at a flow rate of 200 μl/min in a 150 min run. The elute fractions are collected into a 96-well plate using a 1260 series auto-sample fraction collector. The 96 elute fractions were further combined into 12 fractions according to the collection time (combined per column into one fraction, 12 column 12 fractions). Each fraction was dried by rotary evacuation.</p><!><p>Digested peptides were deglycosylated with PNGase F in 50 mM NH4HCO3. PNGase F cleaves between the innermost N-linked core GlcNAc and the Asn residue to which it is covalently linked. PNGase F deamidates the N-linked Asn producing an Asp residue, with a resulting increase of 0.984 Da in molecular weight (Gonzalez et al., 1992). PNGase F-treated peptides were desalted by C18 cartridge solid-phase extraction. The percent occupancy for each site is calculated by comparing the extracted chromatographic area under the curve of peptides with Asn to those with Asp using Byonic software (Version 3.10; Protein Metrics Inc.).</p><!><p>The peptides were reconstituted in 0.1% formic acid–water solution and analyzed on an Orbitrap Eclipse Tribrid mass spectrometer equipped with a nanospray ion source and connected to a Dionex binary solvent system (Thermo Fisher Scientific). Peptides were separated using an Acclaim™ PepMap™ 100 C18 Column (75 μm × 15 cm). A trapping column (PepMap 100 C18 3 μM 75 μM × 2 cm) was used in line with the LC prior to separation with the analytical column. The solvent system consisted of solvent A (100% water/0.1% formic acid) and solvent B (100% ACN/0.1% formic acid). The LC conditions were as follows: 5–35% of solvent B for 165 min, 90% of solvent B for 5 min, and 1% of solvent B for 5 min. The flow rate was set to 300 nl/min. The spray voltage was set to 2.7 kV, and the ion-transfer tube temperature was set to 275°C. The full MS scan range was 400–2000 m/z. Precursor masses were detected in the Orbitrap at resolution (R) = 120,000 (at m/z 200). Stepped HCD (higher-energy collisional dissociation) spectra (HCD energy at 15, 25, and 35%) were recorded for the top 15 most abundant precursors with the standard mode of the AGC target. Dynamic exclusion was set at 30 s. If at least one typical glycan fragment ion abundance (m/z 204.0867 and 366.1396 Da) was observed within the top 15 most abundant fragments and within a 15 ppm mass accuracy, an EThcD [electron-transfer dissociation (ETD) followed by supplemental HCD collision energy at 25%] spectrum of the same precursor would be recorded in the Orbitrap at R = 15,000. The ETD reaction time was set to use calibrated charge-dependent ETD parameters. The glycopeptides of the intact virus were analyzed by stepped HCD fragmentation and HCD-triggered EThcD fragmentation to analyze N-linked glycans and O-linked glycans, respectively. Deamidated peptides were analyzed only by stepped HCD fragmentation.</p><!><p>The LC-MS/MS spectra were searched against the FASTA sequence of the spike protein of the SARS-CoV-2 original virus or the D614G variant using Byos™ (Version 3.10; Protein Metrics Inc.). The searching parameters were specified as follows: fully specific digestion specificity, 2 missing cleavage sited allowed, carbamidomethyl at cysteine as a fixed modification, and oxidation at methionine as a common modification. The precursor ion mass tolerance was set at 6 ppm, and that for fragment ions was at 20 ppm. A 1% false discovery rate (FDR) was applied. The results were filtered with PEP 2D < 0.01, score ≥100, and Delta Mod. Score ≥10. The glycopeptide fragmentation data were evaluated manually for each glycopeptide; the peptide was confirmed when the b and y fragment ions were observed along with oxonium ions corresponding to the glycan identified. A minimum of 3 b/y ions were required. The relative amounts of each glycan at each site were determined by comparing the extracted chromatographic area under the curve. All charge states for a single glycopeptide were summed.</p><p>Glycans were categorized to subtypes according to the composition detected: Hex (4–10)HexNAc(2) was classified as high mannose, NeuAc (0–1)dHex (0–1)Hex (3–7)HexNAc(3) was classified as Hybrid, and NeuAc (0–5)dHex (0–3) Hex (3–8)HexNAc(4–7) was classified as complex-type glycans. Any glycan containing at least one fucose or sialic acid was counted as fucosylated or sialylated, respectively.</p><!><p>Monomeric structural models of N-linked glycan presentation on SARS-CoV-2 were created using electron microscopy structures (PDB ID: 6ZGG), which were visualized with CCP4MG. Glycan cartoon structures are inferred from knowledge of common glycans as identification was done solely by mass. A trimeric structural model of SARS-CoV-2 was created from the electron microscopy structure (PDB ID: 7A96) and visualized with CCP4MG. The antigenic epitopes were predicted using NetCTL-1.2 (Larsen et al., 2007).</p><!><p>Recombinant RBD proteins expressed in HEK293 cells were trypsin-digested. Seventy-five percent of the digest was used for glycopeptide enrichment using HILIC separation, and 25% was deglycosylated in preparation for glycosylation site occupancy analysis. The HILIC-enriched intact glycopeptides were analyzed by LC-MS/MS using HCD-triggered EThcD fragmentation. The deglycosylated peptides were analyzed by LC-MS/MS with HCD fragmentation. The LC-MS/MS raw files were analyzed using Byonic software. The Byonic results were filtered with a 1% false discovery rate and other parameters to achieve high confidence identifications (see the Method section). All the spectra were manually verified.</p><p>The RBD has two potential N-linked glycosylation sites at amino acid positions 331 and 343 relative to the WA1 spike protein. Our data show that both sites are heavily glycosylated with greater than 99% occupancy (Figure 1 and Supplementary Table S1). We observed a high degree of fucosylation at the two N-glycosites, and Man5GlcNAc2 (Man5) is highly abundant at both sites (Figure 1 and Supplementary Table S1). Glycans identified at N331 included high mannose; short complex, paucimannose; and highly abbreviated forms (Figure 1; Supplementary Table S1 and Supplementary Figure S1). The reason is not fully understood but may be related to prompt fragmentation or degradative processes incurred during RBD production and/or purification. Prompt decay is unlikely as no other glycosites demonstrated this pattern.</p><!><p>Glycosylation profile on the recombinant RBD protein. Two N-glycosites and two O-glycosites were identified. Glycan cartoon structures are inferred from knowledge of common glycans as identification was done solely by mass. The bar graphs represent the glycan abundance and unoccupied percent based on total ion abundances at each site.</p><!><p>We also identified two O-glycosylation sites at residues T323 and T523 with a diverse range of glycan compositions. Interestingly, most glycans at the two O-glycosites contain sialic acid (Figure 1 and Supplementary Table S2). Glycosylation of T323, but not T523, has been previously reported. Therefore, we carefully examined the spectra and observed strong evidence of glycosylation at T523 (Supplementary Figure S2). Previous studies reported O-glycosylation at T325 (Shajahan et al., 2020; Zhao et al., 2020), although the occupancy was estimated to be low (Zhao et al., 2020). However, our data did not show direct evidence of fragment ions which can confirm that T325 is glycosylated.</p><!><p>The recombinant S1 protein expressed by HEK293 cells was treated according to the same protocol as the RBD protein (see above), except that two enzymes were used for digestion to facilitate glycoproteomic coverage of the protein. These two enzymes were trypsin and chymotrypsin, used in separate digestion. Byonic search parameters and filters were also the same as for the RBD protein.</p><p>The gene encoding the S1 domain has 13 possible sites of N-glycosylation. Twelve of the 13 predicted N-glycosites were found to be extensively glycosylated (Supplementary Figure S3 and Supplementary Table S3). The one missing glycosite, N17, was detected glycosylated, but it did not meet our criteria due to low confidence scores. Although the scores are low, many hybrid and complex-type glycans were detected at N17 with at least two technical replicates. The site occupancy for 10 glycosites is greater than 90%. Sites N149 and N657 had a site occupancy rate of 25 and 58%, respectively (Supplementary Figure S3 and Supplementary Table S3).</p><p>We observed a diverse range of glycan compositions across the N-linked glycosylation sites. Glycosites N331, N343, N603, and N616 had less glycan variety, while those at N122, N165, N234, N282, and N657 were more diverse (Supplementary Figure S3 and Supplementary Table S3). In addition to the site-specific glycan compositions, overall trends in glycosylation across sites were examined. The results revealed that the three most common types of N-glycans were Man5GlcNAc2 (Man5), HexNAc4Hex3Fuc1, and HexNAc5Hex3Fuc1 (Supplementary Figure S3). Man5 has also been reported by others as a predominant high-mannose glycan composition across all N-glycosites on the SARS-CoV-2 spike protein when produced in HEK293 cells but interestingly not in CHO cells (Zhao et al., 2020; Wang et al., 2021b). The relative abundance of complex-type glycans and the level of fucosylation and sialylation for each site were examined. As shown in Figure 2, the N-linked glycans on the S1 protein were both highly fucosylated (∼80%) and sialylated (∼30%) with overlap where both substitutions were observed on some glycans.</p><!><p>Fucosylated and sialylated N-linked glycosylation of the recombinant S1 protein. Twelve of 13 potential N-glycosylation sites were found occupied, and these N-linked glycans are highly fucosylated.</p><!><p>We also identified 14 O-linked glycosites on the recombinant S1 protein, including the two sites, T323 and T523, which were identified in the recombinant RBD protein. O-glycosylation has been reported to function in immunological shielding, protein stability, and regulation of conformational changes (Casalino et al., 2020a). About half of the 14 sites have not been reported before, and most glycosites display a variety of glycan modifications (Figure 3 and Supplementary Table S4). Three O-glycosites, S673, T678, and S686, are located in the furin cleavage region. Thus, such glycans may modulate the SARS-CoV-2 entry (Andersen et al., 2020). Of the three hypothesized O-glycosites, T678 was identified in this study.</p><!><p>O-linked glycosylation of the recombinant S1 protein. Glycan cartoon structures are inferred from knowledge of common glycans as identification was solely based on mass.</p><!><p>The experimentally observed glycosylation sites are illustrated on the monomeric SARS-CoV-2 spike glycoprotein (PDB code 6ZGG) (Figure 4). To convey the main processing features at each site, the abundances of each glycan were summed by glycan subtype and displayed as a pie chart next to each site. We observed a combination of high-mannose, hybrid, and complex-type N-glycans for most of the sites. Overall, all glycosites were dominated by complex-type glycans when tabulated by subtype. N74 displayed more hybrid-type glycans (30%). N343, in the RBD region, displayed a higher relative amount of mannose-type glycans (28%). This observation aligned with our observations in the recombinant RBD protein (see Figures 1, 4).</p><!><p>Structure-based mapping of glycans on the recombinant SARS-CoV-2 spike protein. The modeling of experimentally observed glycosylation site compositions is illustrated on the monomeric SARS-CoV-2 spike glycoprotein (PDB code 6ZGG). The S1 subunit is colored light blue and peach. The S2 subunit is gray. N- and O-linked glycosylation sites are indicated by green balls and purple balls, respectively. Most abundant glycans at each site are shown. Pie charts show the percentage of glycan subtypes at each site. The boxed area shows the predominant glycans and the N-linked glycosylation subtype distribution for the glycans identified in the recombinant RBD sample. *N74 and T678 are not in the structure.</p><!><p>To illustrate the possible impact of the glycosylation microheterogeneity on the virus antigenicity, we mapped the N-glycosites with antigenic sites and the receptor-binding motifs to the SARS-CoV-2 trimer using a 3D model previously determined by electron microscopy (PDB code 7A96) (Figure 5). The data show extensive microheterogeneity across the glycosites. The number of identified glycoforms at each site ranged from 12 to 83. The antigenic epitopes were predicted using NetCTL-1.2 (Larsen et al., 2007) (Supplementary Table S5). We found that many occupied glycosites are close to, or even overlap with, the antigenic epitopes. Those that overlapped with antigenic sites included N165, N343, N616, and N657, which display substantial glycan diversity (Figure 5). The 3D model has one open RBD bound to the ACE2 protein. The shielding of receptor-binding sites by glycans is a common feature of viral glycoproteins and has been observed for the SARS-CoV spike, the HIV-1 envelope, and influenza hemagglutinin (Bonomelli et al., 2011; Stewart-Jones et al., 2016; An et al., 2019; Zhao et al., 2020).</p><!><p>3D structural modeling of spike glycosylation microheterogeneity. The N-glycosites are mapped on the SARS-CoV-2 trimer structure (PDB code 7A96). Blue indicates the receptor-binding motif in the RBD region. Yellow indicates the predicted MHC antigenic sites. The glycosites are colored according to the mannosylation percentage. The number of glycoforms at each site from less to more heterogeneous glycoforms detected is colored by light to dark, and the number is also listed.</p><!><p>There are two states of RBD: the "down" conformation and the "up" conformation, corresponding to the receptor-inaccessible state and receptor-accessible state, respectively (Gui et al., 2017; Walls et al., 2019; Wrapp et al., 2020). The modeling reveals that N343, N234, and N165 are near to the receptor-binding motif [limited to amino acids 438–506 (Zhou et al., 2021)]. Previous structure analysis revealed that in the RBD "down" state, the RBD region is shielded by the glycans at N343, N165, and N234 (Casalino et al., 2020b). Besides shielding, the glycans at N165 and N234 have also been reported to stabilize the RBD in the "up" conformation (Casalino et al., 2020b). Sztain et al. revealed that the receptor-binding motif is consistently shielded by the glycans at N165 and N234, but RBD opening decreases shielding by the glycans at N343 (Sztain et al., 2021). The N343 glycan may play the role of the "glycan gate" by facilitating conformational shift of the RBD from the "down" to the "up" conformation by interacting with the residues of the ACE2-binding motif (Sztain et al., 2021).</p><!><p>To determine the differences and similarities in glycosylation between the recombinant S1, produced in HEK293 cells, and that of the spike produced in the virus, we examined the spike derived from the intact virus from two strains, the WA1 strain and D614G, propagated in Vero E6 cells. The WA1 strain was from the first patient in the United States who was diagnosed with SARS-CoV-2 viral infection. This case was declared by the state of Washington and CDC on January 20, 2020 (Harcourt et al., 2020). This viral identity was confirmed by whole genome sequencing (GenBank accession no. MN985325), and it did not have mutation at the 614 amino acid. The D614G variant contains the spike protein amino acid change at 614 from D to G, which is more infectious and transmissible and has become the most prevalent form in the global pandemic since March 2021 (Hou et al., 2020; Korber et al., 2020). Both viruses were grown in Vero E6 cells, and the supernatant of the passage 4 stock of each virus was collected, inactivated by gamma irradiation, and analyzed by our glycoproteomics approach.</p><p>Of the 13 predicted N-linked glycosites in the S1 domain, 10 N-glycosites were identified in the WA1 strain (Figure 6A and Supplementary Table S6). The two N-glycosites, N603 and N616, were identified with several high-mannose-type glycans (Man7GlcNAc2, Man8GlcNAc2, and Man9GlcNAc2 at N603 and Man8GlcNAc2 at N616) in a single replicate of the WA1 sample, which did not meet our criteria where two replicates were required to achieve confident identification. Therefore, glycosylation at these sites was considered tentative and not considered further. In contrast, 12 of 13 N-linked glycosites were identified in the S1 domain of the D614G variant (Figure 6A and Supplementary Table S7). Site occupancy identified by the PNGase F deglycosylation methodology revealed that 10 S1 N-glycosites from the WA1 strain (N61, N122, N165, N234, N282, N331, N343, N603, N616, and N567) and 9 N-glycosites from the D614G variant (N61, N149, N165, N234, N331, N343, N603, N616, and N657) were almost 100% glycosylated (Supplementary Tables S8, S9). Three N-glycosites, N74, N122, and N282, were only identified in D614G with a single replicate in the site occupancy study; therefore, we were not able to determine site occupancy at these three sites. (Supplementary Table S9). We do note, however, that these peptides are at least partially occupied by high-mannose glycans based on our glycopeptide analysis (Figure 6A and Supplementary Table S7). Likewise, some N-glycosites were identified with highly diverse glycan compositions upon glycoproteomics analysis of intact glycopeptides. However, occupancy analysis at sites such as N149 in the WA1 strain and N122 in D614G strains did not meet our criteria. Conversely, no glycopeptides were identified at N603 and N616 of the WA1 strain, but these two sites were identified as occupied based on detection of Asp in place of Asn subsequent to PNGase F digestion, which supports that the two sites were glycosylated. Estimated occupancies were between 96.4 and 100%, making it unlikely that spontaneous deamination of unoccupied sites was solely responsible for the Asp presence at the site. The most likely reason for the discrepancy is due to the random sampling issue of mass spectrometry or incomplete enrichment for these very complex samples. The samples of intact viruses contained less than 5% of spike protein abundance according to the protein quantitation by Byonic software. High-abundance glycopeptides of host cells were also enriched using the HILIC column, which resulted in the high complexity of the glycopeptide pool in this experiment.</p><!><p>Comparison of the glycosylation pattern on the spike protein from recombinant S1, the WA1 strain, and the D614G variant. The most abundant glycoforms detected in Vero E6 cell-derived WA1 and D614G strain spike N-glycosites were comparable but different from that detected in recombinant S1 produced in HEK293. (A) N-linked glycan subtypes in the S1 domain. (B) N-linked glycan subtypes in the S2 domain.</p><!><p>When comparing the mutant form, D614G, with the original form, WA1, we observed a similar glycosylation pattern for most N-linked glycosites in both S1 and S2 domains (Figures 6A,B). The most abundant glycoform at each N-glycosite was comparable between WA1 and D614G (Figure 7). Man7GlcNAc2 and Man8GlcNAc2 were the most abundant glycoforms for the majority of the sites, except for N343. There are several sites showing different glycan contents between the two strains, such as N331, N343, and N1074. The D614G variant presents more complex-type glycans at N331 but less complex-type glycans at N343 compared to the WA1 strain. As mentioned earlier, the N343 glycans significantly affect the RBD "up" conformation (Sztain et al., 2021). The glycan changes at N343 in D614G compared to WA1 could, at least partially, account for D614G phenotype changes if similar shifts in glycosylation occur in nature.</p><!><p>N-glycosylation map of Vero E6-derived viruses. The most abundant glycoform detected in Vero E6 cell-derived WA1 and the D614G strain spike N-glycosite was comparable but different from that detected in recombinant S1 produced in HEK293.</p><!><p>In addition, four glycosites, N603, N616, N1158, and N1194, were not identified in the WA1 strain, while all four were identified in the D614G variant (Figures 6A,B, Supplementary Tables S6 and S7). This may not mean that these sites are not glycosylated in WA1. Their absence may have resulted from sample complexity, random sampling, and limitations of our enrichment strategy as discussed above. This hypothesis is supported by the site occupancy analysis where WA1 spike N603, N616, and N1194 sites were clearly occupied (Supplementary Table S8).</p><p>Interestingly, in the recombinant S1 protein, most N-linked glycosites are dominated by complex-type N-glycans, while most glycosites in both the WA1 strain and the D614G variant produced in the virus were dominated by high-mannose-type glycans (Figure 6A). No O-linked glycosites were identified in the virus-derived spike from WA1 and D614G. The reason for the observed dramatic differences in the glycosylation pattern detected between recombinant spike S1 and the virus-derived spike is not clear. There are several possibilities: first, protein structural differences may influence access to glycosylation machinery. Second, the expression of glycosylation enzymes may differ between cell substrates. Third, the secretory location of glycosylation enzymes may differ between cell substrates under conditions of expression and/or virus propagation. WA1 and D614G strains were grown in Vero E6 cells which are derived from the African green monkey kidney, while the recombinant S1 protein was expressed by HEK293 cells which are derived from human embryonic kidney cells.</p><p>It is clear that S1 and the whole spike differ structurally as the former is without the S2 subunit. Several publications reported differences in glycosite occupancy and glycan composition between the intact spike protein and individually expressed S1 and S2 proteins (Watanabe et al., 2020a; Watanabe et al., 2020b; Shajahan et al., 2020; Zhao et al., 2020). The spike expressed in FreeStyle293F cells was found to be partially expressed as the S0 form, without S1/S2 or S2′ cleavage. The S0 form was found to primarily contain high-mannose glycans (Watanabe et al., 2021). We searched our data for evidence of S1/S2 cleavage. We did detect cleavage of the S1/S2 furin cleavage site in the chymotrypsin digest. However, the intensity of these peptides was low, suggesting that significant amounts of the uncleaved S0 forms were present. We must also note that no peptides representing the uncleaved site were detected. It has been reported that Vero E6 cells do not produce high-abundance furin and cleavage of the S glycoprotein in SARS-CoV-2-infected Vero E6 cell lysates was reported to be inefficient (Klimstra et al., 2020). One study, using human serum to detect SARS-CoV-2 proteins produced in infected Vero E6 cell lysates, showed mainly an uncleaved S protein (Haveri et al., 2020). Additionally, in our case, there was also no evidence for TMPRSS2 or cathepsin cleavage. Our observations, however, may have been due to low abundance of the spike in our samples; thus, peptides specifically containing these cleavage sites may not have been detected.</p><p>In general, Vero cells are not known to limit glycan processing primarily to high-mannose glycans. Vero cells have been used as a cell substrate for propagation of influenza and recombinant proteins without report of bias toward high-mannose glycans (Gornik et al., 2012; Rödig et al., 2013). The nascent capacity of Vero cell expression and the secretory location of glycosylation enzymes should not be an issue. Therefore, our observations are not likely to be due to inherent limitations of Vero cells in terms of glycosylation processes. However, our data show that high-mannose-type glycans represent a large portion of total glycans displayed on the Vero E6 host's glycoproteins (Supplementary Figure S4). A range of complex glycans were also identified, albeit with far less abundance. Therefore, the high percent of high-mannose-type glycans on WA1 and D614G grown in Vero E6 cells was not limited to the SARS-CoV-2 spike.</p><p>The proper location of glycosylation enzymes is a complex process involving Rab GTPases, coiled-coil tethers termed golgins, and the multi-subunit tethering complex known as the conserved oligomeric Golgi (COG) complex (Fisher and Ungar, 2016). These factors contribute toward anterograde and retrograde transport of glycosylation active enzymes and other necessary proteins involved in glycoprotein production. Regulation of these processes is essential for appropriate localization and sequential activities of glycosylation active enzymes (Starr et al., 2010). In our studies of the Vero cell-propagated SARS-CoV-2 spike, we noted a low amount of glycosylation processing beyond ER mannosidase I (Moremen et al., 2012) and other mannosidases which are normally present in the cis/medial cisternae of the Golgi (Moremen et al., 2012). This was evidenced by the dominant presence of primarily Man5-8GlcNAc2. There were only low abundances of Man3-5GlcNAc3, also suggesting little processing by cis/medial Golgi located N-acetyglucosaminidase I (Gnt1) (Sztain et al., 2021; Moremen et al., 2012). Therefore, one possibility is that under conditions of viral propagation, the Golgi COG system-mediated anterograde/retrograde system is shifted or viral packaging and routing differs from normal secretion, resulting in an altered distribution of glycosylation active enzymes or proper sequential exposure of these enzymes to nascent glycoproteins. Notably, virus-like particles and the SARS-CoV-2 virus have been localized to the endoplasmic reticulum–Golgi intermediate compartment (ERGIC), a site of secretory sorting between the ER and Golgi, and it has been hypothesized that SARS-CoV-2 exits the cell via lysosomal exocytosis, suggesting little exposure to Golgi enzymes (Mendonça et al., 2021; Plescia et al., 2021). We note that among 25 cell lines tested, Vero E6 produced among the highest viral titers including all those expressing the human ACE2 receptor. Therefore, the high-mannose glycan distribution does not appear to significantly negatively affect viral propagation in the Vero E6 cell line compared to alternative cell substrates typically used in the SARS-CoV-2 viral study (Wang et al., 2021a).</p><p>Watanabe et al. (2021), who also noted unprocessed glycans on the spike, albeit produced in HEK cells expressing the ChAdOx1 vaccine vector, hypothesized that these high-mannose-bearing spike proteins represented those in transit through the secretory system and suggested that furin protease is located in the later trans-Golgi stacks. In our case, this is unlikely as the majority of the virus isolated formed mature viral particles. Significantly, both HEK293 and Vero cells produce predominantly high-mannose glycosylation patterns on the SARS-CoV-2 spike under certain circumstances. The exact reason for this remains an open question.</p><p>Overall, we have found similar glycosylation site-specific N-glycan distributions in S1 and RBD to those previously reported that were produced in HEK293 cell lines. We also report here previously unreported O-glycosylation site occupancy including T523 and confirm the presence of 14 total O-glycosylation site occupancies including T678, which appears in the furin cleavage domain. Significantly, we also report that the native spike produced in SARS-CoV-2/USA-WA1/2020 (USA-WA1) is substituted with primarily high-mannose glycans that do not appear to effect viral propagation in Vero E6 compared to alternative cell substrates (Wang et al., 2021a).</p><!><p>In this study, we characterized the site-specific glycosylation of the spike protein from recombinant RBD and S1 domains and from two intact viruses, the WA1 strain and the D614G variant. Glycosylation was found to be of high occupancy in all samples examined and highly heterogeneous across the majority of glycosites in the HEK293-derived S1 and RBD. Glycan modification at most N-glycosites is very similar between WA1 and D614G and primarily high-mannose, with significant differences at N343. Our results also revealed different patterns of glycan modification among the recombinant S1 protein, recombinant RBD, and the WA1 and D614G strains, which implies that these spike proteins may perform differently in vitro and in vivo. Therefore, the origin of spike glycosylation should be put in consideration for vaccine design and drug development.</p><!><p>The SARS-CoV-2 virus spike protein binds to host cells, fuses with the host cell membrane, and enters the cell. It is heavily glycosylated, and recent studies revealed that glycan modification is essential to modulate spike conformation and host cell invasion. In this study, we analyzed the glycan modification of recombinant spike protein subunit RBD and the S1 domain, both of which function to bind host receptor ACE 2. We also analyzed the glycan modification of whole viruses, the WA1 strain, and the D614G variant. The WA1 strain was isolated from the first case of COVID-19 in the United States. The D614G variant, carrying the protein amino acid change at 614 from aspartate(D) to glycine(G), is now prevalent in the circulating SARS-CoV-2 virus and is carried by all recently identified and highly concerning SARS-CoV-2 variants. We found different patterns of glycan modification among the recombinant S1 protein, recombinant RBD, and WA1 and D614G strains. Glycan modification at most N-glycosites is very similar between WA1 and D614G, with significant differences at N343. This recombinant S1 and RBD glycosylation patterns differ dramatically from the whole virus produced in Vero cells and implies that these spike proteins may perform differently in vitro and in vivo, which could have implications for viral studies, vaccine design, and drug development.</p>

PubMed Open Access

PDMS Elastic Micropost Arrays for Studying Vascular Smooth Muscle Cells

This paper describes the design, modeling, fabrication and characterization of a micromachined array of high-density 3-dimensional microposts (100\xc3\x97100) made of flexible material (silicone elastomers) for use to measure quantitatively the cellular traction force and contractile events in isolated vascular smooth muscle cells (VSMCs). The micropost array was fabricated with diameters ranged from 3 to 10 \xce\xbcm, with edge to edge spacing of 5, 7 and 10 \xce\xbcm, and with a height to diameter aspect ratio up to 10. VSMCs exerted larger basal traction forces when they were grown on stiffer micropost arrays. These basal traction forces were 80% larger in control VSMCs than in VSMCs in which integrin linked kinase (ILK) was knocked down using shRNA. The addition of Angiotensin II (ANGII) led to VSMC contraction as evidenced by an increased traction force exerted on the microposts under the cell. This ANGII induced contractile response and change in traction force on the microposts was not observed in VSMCs lacking ILK. Following treatment of VSMCs with Cytochalasin D to depolymerize the actin cytoskeleton, the VSMCs exhibited relaxation that was apparent as a significant reduction in the measured traction force exerted on microposts under the cell. Overall, this study demonstrates the usefulness of micropost arrays for study of the contractile responsiveness of VSMC and the results indicate that ILK plays a critical role in the signaling pathways leading to the generation of substrate traction force in VSMC.

pdms_elastic_micropost_arrays_for_studying_vascular_smooth_muscle_cells

1. Introduction<!>2. Micropost design and modeling<!>3. Device fabrication<!>4.1 Measurements of Young\xe2\x80\x99s Moduli and Spring Constants<!>4.2 Preparation for Cell Study<!>4.3 Description of Cell Study<!>5. Results and Discussion<!>6. Conclusion<!>

<p>Vascular smooth muscle cells (VSMCs) comprise the mechanically active component of the blood vessel wall endowing it with the ability to constrict and dilate. This vasoregulation occurs through the transmission of mechanical force from the cells to the external environment. Thus the processes of force generation and transmission play an important function role in defining the mechanical efficiency of VSMCs in the vascular wall [1,2]. Techniques for quantitative assessment and measurement of force generation at the level of single cells and at focal adhesion sites where cells interact with their substrate have been limited by technical difficulties. However, recent technical advances in micropost fabrication have permitted development of nanostructured devices that allow questions concerning cell force generation and transmission to its environment to be quantitatively investigated.</p><p>Many techniques have been developed to detect the traction forces generated by cells on a flexible continuum substrate. One example is the use of a thin silicone substrate such as polydimethylsiloxane (PDMS) where cells cultured on the substrate cause it to deform by applying force to the substrate and thereby produce wrinkled pattern on it [3]. Another approach employs the use of fluorescent beads-embedded polyacrylamide gels where the cultured cells generate traction forces that deform the gel structure. The deformation is then be detected by observing and tracking the movement of the embedded beads. These techniques have led to significant improvements in our understanding of spatial and temporal aspects of force generation in cells. However, a limitation of these approaches is that they cannot determine precisely the location, magnitude and directional vector of point forces [3–8]. To overcome this limitation, other techniques have been developed to quantitatively measure cell traction forces including cantilever beams which measure the cell traction forces, micro-fabricated microgrippers, atomic force microscope (AFM) tips or micro-needles in conjunction with photodiode, carbon fiber beams or carbon nanotubes, magnetic beads and optical tweezers [9–16]. These approaches, although highly quantitative are limited to measuring force at a single site on a cell. These limitations were overcome by the development of high-density elastomeric micropost arrays [17–24]. Using the micropost array approach, cells attach and spread across the top of regularly ordered microposts. Since each post is discrete, it can act as an independent cantilever and detect the cell traction force independently at the site where it contacts the cell. Analysis requires only a spring constant and a measurement of the micropost deflection from acquired images of the micropost array. Currently, this technique appears to be among the most effective for quantitative measurement of force at multiple sites between a cell and its substrate. Micropost array technologies have been used to analyze cells such as cardiac fibroblasts, cardiac myocytes, smooth muscle cells, and cell monolayers and detect significantly large contractile forces at the edge of the cell and cells such as tendon fibroblast with weaker contractile force [7,14,20,17,12].</p><p>The objective of our work is to improve the micropost array fabrication process in order to produce high aspect ratio microposts that could be used to measure small changes of traction force when exposed to AngII- a factor in human body that has a vessel constriction effect, and is commonly known to induce VSMC contraction, and compared it with that of the VSMC relaxation effect by Cytochalasin D, and cells with ILK expression reduced. In present work, we successfully fabricated high aspect ratio (up to 10) PDMS micropost arrays. Using these micropost arrays we have shown that VSMCs that grown on a stiffer micropost array substrates generate larger traction forces than those grown on a softer micropost array. In addition, the highly sensitive high-aspect ratio micropost arrays were able to pick-up the change of VSMC contraction forces that occur in an in vivo situation. We have demonstrated that ANGII caused a significant increase in VSMC cell traction force that was not observed in cells with ILK expression reduced. Depolymerization of the actin cytoskeleton with Cytochalasin D reduced traction force exerted by the VSMC on the micropost array and confirmed the involvement of the cytoskeleton in our system.</p><!><p>The device was designed with array of high-density 3-D microposts (100×100) made of flexible material (silicone elastomers) with known physical and chemical properties. The 3-D flexible environment allowed the study of the cell-substrate distribution of traction forces exerted by VSM cells on the micropost array. This device was used to measure the traction forces generated by VSM cells attached to them. It is important to note that the stiffness of the micropost depends on the geometry and material used. Therefore PDMS was used to create the micropost with appropriate mechanical properties and biocompatibility for cells to favorably grow on them. Initially, the micropost's top surfaces were treated and coated with extracellular matrix proteins (collagen or fibronectin) in order to enhance the surface properties and enable the cells to attach to them. Once the cells start moving (expand or contract), the posts will bend. This post-deflection was measured using phase contrast microscopy and was translated into traction force using beam-bending theory. In the linear regime, the post behaves similar to a spring such that the deflection is directly proportional to the force applied by the attached cells. Hence the traction forces can be quantified by determining the deflection of each post.</p><p>The micropost was treated as a cylindrical cantilever beam, one end was fixed to the substrate and the other end was free. The relationship between force, F, and free end displacement, x, for a cylindrical beam can be determined using the theory of cantilever beam bending (Euler-Bernoulli Beam Theory) [6,10,12,25].</p><p> (1)F=Kx=(Kt+K)xd (2)K=[xd/(x-xd)]Kt=(3πED4/64L3) where E, D, L, K, Kt, and x are the Young's modulus, the diameter, the height, the spring constant and the displacement of the micropost, respectively. Kt and xd are the spring constant and displacement of the AFM cantilever beam. Therefore, the Young's modulus of PDMS can be calculated after measuring the spring constant of the micropost and it can then be used as a calibrated value for determining the force.</p><p>Finite element analysis (FEA) using Coventorware simulation package were employed to determine the micropost geometry, the deflection as a function of force applied on the top surface of the micropost, and hence provided an accurate prediction of their performance. Three models were studied in this paper. In these models, the micropost's base was assumed to have a rectangular shape with thickness and area of 500 μm and 250×250 μm2, respectively. The experimentally measured Young Modulus was used in all models (1.38 MPa), and the simulation was performed at the same temperature as the cell environment (23°C) [26]. In addition, the mesh size in the simulation ranged from 10×50 to 70×360 that was sufficient to obtain accurate results. The convergence was accomplished when mesh density increased and a mesh size of 45×230 was chosen to model the behavior of the PDMS micropost. In the first model, the force-deflection relationship was investigated. The micropost diameters and heights were fixed at 5 μm and 25 μm, respectively. The bottom of the micropost's base was fixed and a shear force was applied to the center of its top surface and ranged between 0 to 60 nN. The deflections corresponding to different levels of shear forces were computed as shown in figure 1a. For example, a 15 nN force was able to cause a noticeable deflection to the post by 2.55 μm. It should be noted that the post was deflected linearly when the traction force was between 0 – 40 nN. In second model, the height-deflection relationship was determined. The height of the micropost was varied between 0–25 μm, and the diameter and applied force were fixed to 5 μm, and 15 nN, respectively as shown in figure 1b. A deflection of 2.52 μm was achieved for a post with a height of 25 μm (figure 1c). In the third model, diameter-deflection relationship was determined by varying post diameter from 3 to15 μm and fixing post height and applied force to 25 μm and 15 nN, respectively. The simulation result is shown in figure 1d. The micropost was also simulated with a rigid anchor (without base) and compared with the one with a base. The results showed a slight difference of 2.5%. For example, the deflection of a micropost with diameter and height of 5 μm, and 25 μm, respectively and under a constant force of 15 nN with and without base were 2.55 μm and 2.60 μm, respectively. A mask was designed and fabricated based on the modeling results, the following dimensions were included in the mask: the post diameters were 3, 5, 7, and 10 μm, the spacing were 5, 7, and 10 μm.</p><!><p>The micropost arrays were fabricated using standard microlithography and replica-molding techniques in the following sequence: 1) a photoresist layer (Shipley 1813 and 1827) was first patterned on a 3″ silicon wafer (figure 2a), which was initially cleaned with piranha solution for 15 minutes, to form openings at locations corresponding to the microposts; 2) silicon micromold was formed by etching micro-holes with high aspect ratio using a Deep Reactive Ion Etching System (Alcatel DRIE AMS-100) (See figure 2b); 3) the wafer was cleaned again with Pirhana solution and the natural oxide layer was removed using Hydrofluoric acid (HF); 4) the wafer was treated with a vapor of hexamethyldisilazane reagent in vacuum desiccators for 15 minutes.</p><p>This step proved to be crucial for preventing the PDMS from sticking to the substrate due to reduction of PDMS's surface tension to the silicon mold and hence facilitated peeling off the posts from the mold after being cured; 4) the wafer was diced into 2×2 cm2, each containing 4 arrays and was placed in a Petri dish and a PDMS, prepared by mixing the resin (Sylgard 184 Silicon Elastomer Kit, Dow Corning) with curing agent (10:1), was poured over them. The PDMS - wafer mold were placed inside a vacuum at 65°C for 24 hours (figure 2c). The long baking time was needed to completely cure the PDMS in the high aspect ratio mold. This step was necessary for successfully removing the micropost from the mold. 5) After the PDMS was cured, it was peeled off manually creating the micropost arrays (figure 2d). Each of the removed samples was cut into 4 arrays and they were ready for cell culturing experiment. The microposts were fabricated with diameter ranged between 3–10 μm, height between 5–40 μm and spacing of 5–10 μm. The best aspect ratio achieved was 10. In this case the post diameter and height were 3 μm and 30 μm, respectively. Scanning electron micrographs of the fabricated arrays are shown in figure 3.</p><!><p>We have measured the Young's Modulus and stiffness of PDMS and used them as a calibrated value for determining the traction forces of VSMCs. The measurements were performed on over 50 microposts divided between four set with small and large dimensions that were cured at room temperature or 65 °C for the same duration (24 hours). The small dimension and large dimension microposts are referred to microposts with diameter (D) and height (H) of 3.3–9.2 μm, 10–27 μm, 16.1–26.7 μm and 30–42.9 μm, respectively. A range of forces was applied on the surface of the micropost using an AFM cantilever beam with a spring constant of 0.06 N/m (Veeco Company). The applied forces and the corresponding micropost displacements were recorded in real time by computer and their relation was obtained by using Matlab program. This step was repeated several times to remove the system noise and measurement error. The calibration of the microposts starts by measuring the sensitivity of AFM against a glass substrate, and measuring the indentation of PDMS substrate by applying force between 0–60 nN. In the former case, there was no indentation during the measurement. The deflection of the micropost was determined by applying force using AFM on its free end and by subtracting the PDMS indentation (σ) and the deflection of AFM cantilever beam (d) from Z-travel of the cantilever beam's base (z). The deflection is given by x=z−d−σ. For this analysis, we assumed the microposts have uniform material properties such that the deflection is equivalent to the corresponding traction force divided by the spring constant. The calibrated results show that a low value Young's Modulus of PDMS was found in small-scale and large-scale microposts that were cured at room temperature (0.936±0.037 MPa and 0.543±0.032 MPa, respectively), while microposts that were cured at 65 °C had higher Young's Modulus values of 1.378±0.162 MPa and 1.090±0.090 MPa, respectively. The p-values of all data sets were much less than 0.001 which verifies that there is significant variation between compared data sets. The results indicate that the posts cured at higher temperature were stiffer than posts cured at room temperature. The comparison between small and large microposts that were cured at the same temperature indicates that the speed of curing of the micropost was influenced by the dimensions of micropost. The large size of a given micropost will slow the curing process in the silicon mold compared to those of smaller size. The exact diameter and height of the micropost was measured using scanning electron microscopy (SEM), and the same size and fabrication regime for micropost, used in the cell experiments, was used to determine the Young's Modulus. For example, the spring constants of the microposts with diameter, height and spacing of 5 μm, 25 μm, and 7 μm; and 3 μm, 25 μm 5 μm, respectively are 8.12 nN/μm and 1.05 nN/μm, respectively. The Young's Modulus in both cases was 1.378±0.162 MPa.</p><!><p>We have successfully cultured VSMCs on the fabricated high aspect ratio microposts with two different micropost geometries of diameter, height and spacing of 5 μm, 25 μm, and 7 μm, and 3 μm, 25 μm, 5 μm, respectively. Figure 4 shows an SEM micrograph of VSMC grown on two dimensional micropost array. This micropost array was first treated with oxygen plasma for 45 seconds and then exposed with UV light for 30 minutes to sterilize them before using the arrays for cell culturing. VSMCs enzymatically isolated from skeletal muscle arterioles from male Sprague-Dawley rats (250–350 g) were cultured on top of the micropost arrays and incubated for 2 days before image acquisition experiment. The cells on arrays were imaged using an Olympus Fluoview 1000 confocal microscope with 60X water objective. A series of wide field optical images were captured by scanning the micropost array from the micropost bottom to the micropost top. A single pixel in the wide field optical image represents a square area of 400×400 nm2 which corresponds to minimum measured traction force for the two types of microposts of 1.62 nN and 0.21 nN, respectively, which represents the resolution of our force measurement system using these microposts. The largest measurable traction force was limited by the spacing between adjacent posts. It was 97.08 nN, and 8.73 nN for the two micropost geometries, respectively. The microposts with no cells attached maintained their erect position and were unchanged when emerging in the media solution through the cell experiment (See Figure 5). In order to verify that the micropost without cells maintained a stable position during the experiment, the top and bottom positions of the same post were compared initially. It was found that there was no position change of the micropost without cell contact in these images that proved that the microposts with such geometry were able to stand stably and vertically in the solution. For example, Figure 5 shows the micropost array (with diameter, height and spacing of 5 μm in 25 μm, and 7 μm, respectively) used in cell study experiment. By calculating the center position of each micropost in both bottom and top images, the average deflection of posts with no cell is 0.348 pixel which is less than the minimum detectable displacement. Therefore, motion noise contributions from deflections of microposts with no cells on top were negligible. The microposts at the cell's edge had an average deflection of 16.26 pixels (with standard error of 2.17) while the ones at the center of the cell have the average deflection of 2.34 pixels (with standard error of 2.17). This experiment demonstrated that the cells were exerting higher traction forces at the cell margin the in the cell center region and shown that the micropost array was able to measure the traction force of VSMCs effectively and accurately.</p><!><p>The effects of Integrin-linked kinase (ILK), a serine/threonine protein kinase implicated in signaling pathways involved in VSMCs adhesion, proliferation and migration were analyzed by measuring the traction forces exerted by VSMCs on the microposts [27]. Two groups of VSMCs were studied, the first was a control group, while the second group consisted of the VSMC cells in which ILK expression was reduced using a shRNA approach and designated as VSMC-ILK cells. The control and ILK cells were grown on top of PDMS micropost arrays. The cells were grown on separate micropost arrays with the same geometry in a 35 mm culture dish for 2 days. The optical images of the cells were recorded using Olympus Fluoview1000 confocal microscope. These images were used to determine the magnitude of the micropost deflection and map the distribution traction forces. Figure 6 shows the wide field optical images of a single control VSMC and a VSMC-ILK grown on separate micropost arrays with same geometry, and the corresponding force distributions maps. The arrows (force vectors) in both figure panels (a and b) point towards the direction of deflection to display the force vector and the arrow lengths are proportional to the magnitude of the traction force applied to the corresponding micropost. The results demonstrate that the traction forces of VSMCs cell point toward the cell's physical center. The analysis of traction forces also shows that the forces are small around the center region of the cell compared to higher forces near the cell edge. It was also found that there was no notable difference in general morphology between control VSMC and VSMC-ILK.</p><p>In the second part of the experiment, Angiotensin II (ANG II), a potent vasoconstrictor hormone, was diluted to 10−5 M/ml by Dulbecco's Phosphate-Buffered Saline buffer (DPBS) and then added to both groups of cells. The effect of ANGII on the cells was analyzed by measuring the subsequent change in micropost deflections. This process was recorded by camera at 10 frames/min scanning rate for 30 minutes beginning immediately after addition of ANGII. After 30 minutes, Cytochalasin D (CYTO), an agent that depolymerizes the f-actin cytoskeleton, was diluted to 10−7 M/ml by DPBS and added to the cells. Images were continuously recorded for 40 minutes after application of Cytochalasin D. The responses of two groups of VSMCs with respect to ANG II and CYTO were obtained and compared by recording and analyzing the images to quantify the deflection of the microposts.</p><!><p>Control VSMCs and VSMCs-ILK have been grown on micropost arrays with different geometries. Two examples are shown in figure 7 and figure 8 respectively. It should be noticed that the arrows' length in figure 7a and 8a has been magnified 8 times to make those arrows long enough to been seen. The same experiment was repeated 5 times to verify the repeatability. Table I shows the experimental results of average deflection of the microposts and the corresponding traction force generated by control VSMCs and VSMCs-ILK. The results demonstrate that control VSMCs and VSMCs-ILK tend to exert larger traction force when they grow on stiffer micropost array. However, the traction forces generated by control VSMCs are more than 80% larger than that generated by VSMCs-ILK when the same size of micropost array was used. The P-values of controlled VSMCs and ILK cells grown on 3 μm and 5 μm micropost arrays are 9.75×10−06 and 6.24×10−06 which shows the extinguished traction force from these two types of cells.</p><p>The dynamic change and traction force distribution map of a single control VSMC after adding ANG II followed by CYTO, are shown in Figure 9a–9e. The optical image recorded before adding ANGII to the control VSMCs (figure 9a) shows that control VSMCs generates a basal level traction force on the micropost array. The images recorded 3 minutes, 6 minutes and 20 minutes after adding ANGII are shown in Figure 9b to 9d, and in Figure 9e, 40 minutes after adding CYTO. Micropost labeled as 1 (in figure 9a) shows the deflection equivalent to the traction force of 63.4 nN (before adding ANGII). This force has increased to 69.4 nN (3 minutes after adding ANGII) and to 79.3 nN (6 min after adding ANGII). The traction force has decrease to 71.6 nN (20 minutes after adding ANGII). The results show that ANGII leads to significant contraction in the control VSMC. The cell contraction continued for 6 minutes. Then, the contraction subsided slightly. A pronounced relaxation (expansion) occurred 15 minutes after adding the CYTO to control VSMC and it persisted for 30 minutes. Figure 9k shows the traction force of control VSMC applied on post 1 as a function of observation time. The force applied to the post before adding ANGII (63.4 nN) reflects the steady state level. The contraction was noted as when the traction force applied at that time is larger than the steady state level and relaxation occurred when the force level is lower than the steady state level. The exact experiment has been repeated by 5 VSMCs.</p><p>The dynamic changes and traction force distribution map of a single VSMC-ILK growing on a PDMS micropost array after adding ANG II and CYTO are shown in Figure 9f–9j. Similarly, the optical image recorded right before adding ANGII to the ILK cell (Figure 9f) show that VSMC-ILK also generates a basal traction force on the micropost array. The image recorded 3 minutes, 6 minutes and 20 minutes after adding ANGII to VSMC-ILK are shown in Figure 9g to 9i, and 40 minutes after adding CYTO to VSMC-ILK is shown in Figure 9j. Micropost labeled by 1 shows traction force of 15.32 nN (before adding ANGII). This force has been reduced to 12.39 nN (3 minutes after adding ANGII) and to 10.86 (6 min after adding ANGII). The traction force has been decrease to 9.28 nN by 20 minutes after adding ANGII). The results show that the reaction from VSMC-ILK is opposite to that of CK4 cell during this period. The VSMC-ILK continues relaxing during the remainder of the observing process. The relaxation (expansion) occurs 15 minutes after adding the CYTO to VSMC-ILK and persisted approximately 30 minutes. Figure 9i shows the traction force of VSMC-ILK applied on post 1 as a function of observation time. The force applied to the post before adding ANGII (12.39 nN) is the basal level. Relaxation occurred at all time points since the force levels are lower than the steady state level. The experiments described in figure 9 were repeated on 6 VSMCs-ILK. Figure 10 shows the average forces in those 6 experiments applied by control VSMCs and VSMC-ILK on microposts, respectively, as a function of observation time. The error bar in the curve represents the standard error of the data set. Table II shows the average traction forces of the 6 repeated experiments and the corresponding standard errors. The solid and dot curve in figure 10 represents the dynamic change of average traction force applied on micropost by control VSMC and VSMC-ILK, respectively. The traction force applied by control VSMC continued to increase after adding ANGII to the cell bath and reached maximum at 6 minutes after ANGII. While there was no contraction found all through the dynamic study of VSMCs-ILK, suggesting that ILK plays an important role in mediating force generation in VSMC and in mediating the ANGII induced VSMC contraction.</p><p>ILK has been shown to directly interact with the cytoplasmic domain of integrins β1 and β3 [28], and has been implicated to play an important role as both a kinase and as a scaffolding protein in the focal adhesion (FA) structure [29,30]. Previous studies from our laboratory [31] have shown that knock-down of ILK expression level using an shRNA approach, enhanced integrin-mediated adhesions and FA formation but also resulted in a significantly lowered cortical stiffness of VSMC. This suggests that there may be lower mechanical force transmitted from the cell across the FA by way of the VSMC stress fibers. This observation is consistent with the current findings that ILK-knock-down in VSMC displayed a reduced traction force on the micro-posts. These findings raised the possibility that ILK silencing could weaken the VSMC contraction and their ability to apply stress to the external mechanical environment. Earlier studies in search of a target of the ILK's kinase activity have suggested a role in calcium-independent myosin dephosporylation in WSMC [32,33], and in phosphorylation of the myosin phosphatase target subunit, thereby inhibiting the activity of myosin phosphatase [34]. These mechanisms provide possible explanation for the low traction forces observed in these studies and the reduced contractile response to AngII stimulation in the ILK-knock-down VSMCs. Together, these data provide further evidence to support a functional role of ILK in promoting transmission of force through Fas during contraction in VSMC.</p><!><p>The high aspect ratio PDMS micropost array was utilized for a traction force study of VSMCs. The smallest spring constant of the micropost is 1.05 nN/μm and the corresponding minimum detectable traction force is 0.21 nN. The control VSMCs cells and VSMCs-ILK were successfully grown on microposts array with two different geometries. It was found that both cell groups exert larger traction force when growing on stiffer micropost arrays. By comparing these two groups of VSMCs using the micropost arrays with the same size, it was observed that control VSMCs generate larger traction forces than VSMCs-ILK. It has also been found that VSMCs-ILK loses the ability to increase traction force following stimulation with ANGII when compared to control VSMCs. This demonstrates that ILK is important for the VSMC to develop and maintain substrate traction force and to alter traction force in response to agonists that activate the contractile apparatus.</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>Qi Cheng (S'09) received the B.Sc. and M.Sc. degrees from Nanjing University of Science and Technology, Nanjing, China, in 2000 and 2003, respectively, and the M.Sc. degree from the Royal Institute of Technology (KTH), Sweden in 2006, all in electrical engineering. He completed his Ph.D in 2011 in electrical engineering at the University of Missouri, Columbia. His research interests include optimization of Si-Ge-O sensing material for uncooled infrared detection; optical, thermal, and mechanical design and modeling of uncooled tunable microbolometers; and micropost array biosensor for cardiovascular cell traction force studies. He is currently working at Intel Corporation, Hillsboro, Oregon.</p><p>Zhe Sun received Ph.D. degree in Engineering from Department of BioEngineering, The University of Toledo, U.S.A., in 2000. He now serves as Assistant Research Professor in the Dalton Cardiovascular Research Center, University of Missouri-Columbia, U.S.A. His current research interests include N-cadherin function as a mechano-sensor in Vascular Smooth Muscle Cells and the regulation of cerebral artery myogenic tone.</p><p>Gerald A. Meininger serves as Director of the Dalton Cardiovascular Research Center at the University of Missouri. He is a Margaret Mulligan Professor in Medical Research and appointed in the Department of Medical Pharmacology with adjunct appointments in the Departments of Biomedical Sciences and Biological Engineering. Prior to MU, Dr. Meininger was at Texas A&M University, for 25 years and was a Regents Professor, Director of the Division of Vascular Biology at the Cardiovascular Research Institute and Associate Department Head of the Medical Physiology. His research interests are in mechanobiology of the vascular wall and cell interactions with extracellular matrix.</p><p>Mahmoud Almasri received PhD in Electrical Engineering from Southern Methodist University in 2001. He is currently an associate professor with the Department of Electrical and Computer Engineering, University of Missouri, Columbia. From 2001 to 2002 he was a research scientist with General Monitors, Lake Forest CA. From 2002 to 2003 he was with Albanynanotech of Albany, NY as a post doctoral research associate, and from 2004 to 2005 he was with Georgia Institute of Technology as a post doctoral fellow. His research includes biosensors, capacitors for power harvesting, infrared detectors, MEMS Coulter counter, and uncooled infrared material and detectors.</p><p>The relationship between a) the micropost deflection and the lateral force (diameters and heights were fixed at 5 μm and 25 μm, respectively), b) the micropost deflection and the micropost height (the diameter and applied force were fixed to 5 μm, and 15 nN, respectively), c) the micropost deflection and the micropost diameter (post height and applied force were fixed to 25 μm and 15 nN, respectively), d) schematic of FEA model for a micropost with a lateral force exerted on its top. The post's diameter, height, and force are 5 μm, 25 μm, 15 nN, respectively. The simulations were performed using Coventorware finite element tool.</p><p>Micropost array fabrication process, a) after photolithography and patterning circles which corresponds to post diameter, b) after deep Si etching to create Si mold with high aspect ratio, c) after pouring and curing PDMS, d) after peeling off cured PDMS from the mold.</p><p>Scanning Electron Micrographs (SEMs) of microposts with diameter, height, and spacing of a) 10 μm, 30 μm, 7 μm, b) 3 μm, 20 μm, 7 μm; c) 3 μm, 25 μm, 7 μm c), 3 μm, 30 μm, 5 μm; respectively.</p><p>Single VSMCs grown on two micropost arrays with diameter, height, and spacing of a) 7 μm, 25 μm, and 7 μm. The dehydration (freezing) process of the VSMC on the microposts causes them to squeeze together due to shrinkage of the cell. We were not able to capture good images of cells on high aspect microposts.</p><p>Wide field optical images of single VSMC grown on micropost array with 5μm in diameter, 25μm in height and 7μm in spacing, (a) recorded at the top of micropost array (b) recorded at bottom of the micropost array, respectively.</p><p>Wide field optical images of single a) control VSMC and b) VSMC-ILK grown on micropost array with diameter, height and spacing of 5 μm, 25 μm and 7 μm, respectively.</p><p>Wide field optical images of single control VSMC grown on micropost array with a) 3 μm in diameter, 25 μm in height and b) 5μm in diameter, 25 μm in height, respectively. The arrow's length in figure 7a is magnified 8 times to make it long enough to be seen.</p><p>Wide field optical images of single control VSMC-ILK grown on micropost array with a) 3 μm in diameter, 25 μm in height and b) 5 μm in diameter, 25 μm in height, respectively. The arrow's length in figure 8a is magnified 8 times to make it long enough to be seen.</p><p>Wide field optical images of single control VSMC (a–e) and VSMC-ILK (f–j) cell recorded in real time: (a) and (f) right before adding ANGII, (b)–(d) and (g)–(i) 3 minutes, 6 minutes and 20 minutes after adding ANG II, (e) and (j) 40 minutes after adding CYTO. The arrows' length in b, c, g and h has been magnified 6 time times; the arrows' length in d, e, i and j has been magnified 2 times to make arrows observable. (k) The traction force of control VSMC applied on post 1 (figure 9a) as a function of observation time. (l) The traction force of ILK cell applied on post 1 (Fig. 9f) as a function of observation time.</p><p>The average traction force of control VSMC and VSMC-ILK applied on posts as a function of observation time. (The error bar represents the standard error of the measured traction forces)</p><p>Average deflection of the micropost and corresponding traction force generated by cells on micropost array. D is the deflection of the micropost, F is the traction force applied on micropost. C and I represent the control VSMCs and VSMC-ILK, respectively.</p><p>Traction force applied on post 1, 2, 3 in Figure 9f</p>

PubMed Author Manuscript

Reformulating a Pharmacophore for 5-HT2A Serotonin Receptor Antagonists

Several pharmacophore models have been proposed for 5-HT2A serotonin receptor antagonists. These typically consist of two aromatic/hydrophobic moieties separated by a given distance from each other, and from a basic amine. Although specified distances might vary, the models are relatively similar in their general construction. Because our preliminary data indicated that two aromatic (hydrophobic) moieties might not be required for such action, we deconstructed the serotonin-dopamine antipsychotic agent risperidone (1) into four smaller structural fragments that were thoroughly examined in 5-HT2A receptor binding and functional (i.e., two-electrode voltage clamp \xe2\x80\x93 TEVC \xe2\x80\x93 and intracellular calcium release) assays. It was apparent that truncated risperidone analogs behaved as antagonists. In particular, 6-fluoro-3-(1-methylpiperidin-4-yl)benzisoxazole (4) displayed high affinity for 5-HT2A receptors (Ki ca 12 nM) relative to risperidone (Ki ca 5 nM) and behaved as a potent 5-HT2A serotonin receptor antagonist. These results suggest that multiple aromatic (hydrophobic) moieties are not essential for high-affinity 5-HT2A receptor binding and antagonist activity and that current pharmacophore models for such agents are very much in need of revision.

reformulating_a_pharmacophore_for_5-ht2a_serotonin_receptor_antagonists

Introduction<!>Synthesis<!>Binding<!>Functional assays<!>Synthesis<!>4-(4-(6-Fluorobenzisoxazol-3-yl)piperidin-1-yl)-1-(piperidin-1-yl)butan-1-one Hydrochloride (2)<!>6-Fluoro-3-(1-(4-piperidin-1-yl)butyl)piperidin-4-yl)benzisoxazole Hydrochloride (3)<!>6-Fluoro-3-(1-methylpiperidin-4-yl)benzisoxazole Hydrochloride (4)<!>4-Chloro-1-(piperidin-1-yl)butan-1-one (12)<!>Transient Transfection of HEK293 cells<!>Radioligand binding<!>Expression of Recombinant Proteins in Xenopus oocytes<!>Two-Electrode Voltage-Clamp Recording and Analysis<!>Measurement of Intracellular Ca2+

<p>Over the past 25 years, several pharmacophore models have been described for 5-HT2A serotonin receptor antagonists.1–7 To some extent, the various models might reflect the nature of the antagonists examined in the individual studies. Nevertheless, there now seems agreement that multiple binding modes are possible for competitive 5-HT2A receptor antagonists,2,4–8 and this could account for the multiple models. Typically, the pharmacophore models consist of two aromatic (hydrophobic) moieties separated by a given distance from each other, and from a basic amine. Although the distances might vary from model to model,2–5 these three features are common to nearly all the agents examined. When the two aromatic features flank a third (i.e., central) ring, as in a tricyclic ring system, even the fold-angle of the central ring is thought to play a role.9</p><p>Two of the few 5-HT2A receptor antagonists that have found clinical application are the SDA (serotonin-dopamine antagonist) antipsychotic agents risperidone (1; Figure 1) and its 9-hydroxy metabolite, paliperidone. Risperidone (1) has been considered only in a few pharmacophore studies.10,11 In one study, the authors noted that their model did not necessarily predict structural features required for binding at 5-HT2A or dopamine D2 receptors, but simply identified features important for 5-HT2A/D2-associated antipsychotic action.10 In another study, multiple pharmacophoric features for 5-HT2A antagonist action were identified including the two aromatic (hydrophobic) moieties.11</p><p>Some pharmacophore studies might have been limited by the nature of the agents examined. Indeed, most 5-HT2A receptor pharmacophore studies have examined fairly large molecules and, as already mentioned, nearly all contained at least two aromatic (hydrophobic) centers. The structural requirements for risperidone (1) to act as a 5-HT2A receptor antagonist have not been extensively examined. Hence, it was of interest here to determine how much of the risperidone (1) structure is actually required in order to retain this action. Consequently, four deconstructed analogs (i.e., partial structures) of risperidone (1), 2–5 (Figure 2), were prepared and investigated. Compound 2 retains the amide terminus of risperidone (1) whereas 3 possesses an amine terminus. The terminal chain of 2/3 was truncated to the N-methylpiperidine 4 and piperidine analog 5.</p><p>Preliminary evaluation of 2–5 (at a single concentration of 10 µM) revealed 5-HT2A antagonist properties. To determine if the effect involved the orthosteric binding site of 5-HT2A receptors, their binding affinity was measured. Their functional activity was then compared with that of risperidone (1) using a two-electrode voltage clamp (TEVC) assay as well as an intracellular calcium release assay.</p><!><p>The deconstructed analogs of risperidone, 2–5, were synthesized as outlined in Scheme 1. Compound 5 was prepared according to a literature procedure12 and served as a key intermediate in the preparation of 2–4. N-Alkylation of 5 by 4-chloro-1-(piperidin-1-yl)butan-1-one (12) utilizing a Finkelstein reaction yielded the desired amide 2 (Scheme 1). Reduction of 2 with diborane·THF complex resulted in 3. The N-methyl analog of 5, 4, was obtained using an Eschweiler-Clarke N-methylation reaction. All compounds were prepared as water soluble hydrochloride salts.</p><!><p>Competition binding assays were performed in plasma membrane preparations of human embryonic kidney (HEK293) cells transiently transfected with a construct encoding 5-HT2A receptors for determining the affinity of risperidone. Risperidone displaced [3H]ketanserin binding (Supporting Information, Figure SI-1) with a Ki value of 5.29 nM, consistent with its previously reported high affinity for 5-HT2A receptors.13</p><p>We examined the binding affinity of the risperidone derivatives much in the same way we did for risperidone. Compounds 2 (Ki = 39.81 nM) and 3 (Ki = 34.83 nM) retained high binding affinity to 5-HT2A receptors, even though they exhibited lower affinity (7.4- and 6.6-fold lower affinity, respectively; Figure 3) than risperidone (1). The N-methylpiperidine derivative (4; Ki = 12.27 nM) showed the highest binding affinity (half that of risperidone), whereas the piperidine derivative (5; Ki = 71.41 nM) showed the lowest binding affinity of all derivatives tested (13.5-fold lower than risperidone). Data for the binding of risperidone (Supporting Information, Figure SI-1) and 2–5 are shown in Figure 3 and are summarized in Table 1. These results suggested that all derivatives tested bind competitively at 5-HT2A receptors with relatively high affinity.</p><!><p>Ion channels can serve as sensitive reporters for G protein-coupled receptor (GPCR) activity.14,15 We utilized the Xenopus laevis oocyte system to heterologously express 5-HT2A receptors and the G protein-gated inwardly rectifying K+ (GIRK4-S143T or GIRK4*) reporter, a channel activated by Gβγ associated with PTX-sensitive Gα subunits.16,17 When 1 µM serotonin (5-HT) was perfused in the bath in a two-electrode voltage clamp (TEVC) experiment, two effects became apparent: activation of a transient outwardly rectifying (larger outward than inward) current, followed by inhibition of the inwardly rectifying (larger inward than outward) GIRK4* current (Figure 4A). The transient current reflects activation of a calcium-activated chloride channel (ICa-Cl) endogenous to Xenopus oocytes, providing functional evidence that 5-HT2A receptor signaling occurred (i.e. Gq activation → PLCβ1 activation → hydrolysis of PIP2 to DAG and IP3 generation → release of Ca2+ from ER stores).e.g. 18 The ensuing inhibition of the GIRK4* current is due to phosphoinositide hydrolysis and, thus, a decrease in the plasma membrane concentration of PIP2, as interactions of this and most ion channels with PI(4,5)P2 are essential to keep the channel gates open.19 In the presence of 3 µM risperidone (1), 5-HT-mediated current inhibition was greatly attenuated. Some ICa-Cl could be seen only in the outward direction, while the inhibition of the GIRK4* current was abolished (Figure 4B).</p><p>A concentration-response of risperidone antagonizing the action of 5-HT (1 µM) was performed and the results showed significant effects at concentrations of 100 nM or greater (Figures 4C and D). The apparent risperidone IC50 value was estimated by this assay at 55.7 nM, ~10-fold lower than its binding affinity (see Supporting Information, Figure SI-1).</p><p>Before proceeding with similar functional characterization of antagonist action of the deconstructed risperidone analogs, we examined their possible agonist effects. All compounds, except compound 5, yielded significant current inhibition at concentrations of 50 µM or higher (Supporting Information, Figure SI-2A-D). Compound 3 seemed to cause significant current inhibition at concentrations as low as 5 µM. When we compared the effects of the risperidone derivatives at 50 µM or higher in oocytes expressing GIRK4* alone, versus GIRK4* and 5-HT2A receptors together, we found that the effects of risperidone and simplified derivatives, 2 and 3, showed no significant differences between the two groups, suggesting that high concentrations of risperidone and 2 and 3 inhibited the GIRK4* channel directly (Supporting Information, Figures SI-3A–C). Unlike compounds 1–4, compound 5 did not exhibit significant agonism at the concentrations tested up to 50 µM (data not shown). Overall, these data revealed a limitation as to how high in concentration we could test the effects of risperidone derivatives using GIRK4* as a reporter for 5-HT2A receptor activity.</p><p>Concentration-response experiments aiming to antagonize 5-HT responses revealed that significant inhibition could be achieved at the highest concentrations tested (10 µM) for compounds 2, 4, and 5, while for compound 3 no significant effects could be achieved, possibly because at as low as 5 µM this compound inhibited the GIRK4* channel directly (Figure 5B and Supporting Information, Figures SI-2C and 3C).</p><p>The concentration-response experiments in the oocyte system using GIRK4* channel currents as the reporter of the action of risperidone analogs suggested that at least three of the four analogs functioned as antagonists of 5-HT responses. Yet, the direct action of compounds 1–4 on the reporter channel itself limited full characterization of these compounds.</p><p>These results prompted us to consider a complementary functional assay to report on the action of risperidone and its deconstructed analogs. Figure 6 shows epifluorescence experiments in HEK 293 cells stably expressing 5-HT2A receptors and using the Fura 2 dye to report changes in intracellular Ca2+ due to risperidone and its deconstructed analogs. First we applied compounds 1–5 by themselves (up to a concentration of 100 µM) and recorded no changes in intracellular Ca2+ concentration (data not shown). Application of 1 µM 5-HT produced robust Ca2+ transients that could be blocked progressively by increasing concentrations of risperidone and its derivatives (Figure 6). Thus, based on apparent IC50 values in the [Ca2+]i assay the order of potency was as follows: risperidone (5.59 µM) > 4 (7.40 µM) > 2 (16.65 µM) > 5 (20.12 µM) > 3 (43.88 µM) (summarized in Table 1).</p><p>According to existing pharmacophore models, it might not have been expected that compounds lacking two aromatic (hydrophobic) moieties would be effective 5-HT2A receptor competitive antagonists. Clearly, the entire structure of risperidone (1) is shown here to be unnecessary for this action. For example, compounds 2, 3, and 4 bind with only about 2- to 8-fold lower affinity than 1, and are only about 2- to 3-fold less potent than risperidone as antagonists in the Ca2+ imaging assay.</p><p>Although changes in intracellular Ca2+ proved to be a three-orders of magnitude lower sensitivity assay than the GIRK4* channel reporter, results from the two methods were consistent in showing that risperidone (1) was the most potent antagonist, whereas compound 3 was the least potent antagonist of 5-HT responses (compare Figures 4C and 6B versus Figures 5B and 6D). Moreover, the [Ca2+]i assay proved cleaner than the channel reporter assay since risperidone (1) and its derivatives did not change basal [Ca2+]i levels and allowed a complete ranking of these compounds in terms of their functional potency.</p><p>In summary, it would appear from the present results that truncated versions of risperidone (1) bind with nanomolar affinity and retain 5-HT2A antagonist character. In fact, we reported some time ago that compound 5 binds only with 16-fold lower affinity than risperidone (1) at [3H]ketanserin-labeled 5-HT2A receptors (rat brain homogenates) and acted as a potent 5-HT2 receptor antagonist in vitro (5-HT-induced inositol phosphate production) and in vivo (rats trained to discriminate the 5-HT2 receptor agonist 1-(2,5-dimethoxy-4-methylphenyl)-2-aminopropane from vehicle) without producing any agonist action when examined alone.13 Taken together with the present findings, future pharmacophore models for 5-HT2A receptor antagonists will need to consider this information. Although most 5-HT2A serotonin receptor antagonists commonly possess more than a single aromatic/hydrophobic feature, this would not appear to be an essential requirement.</p><!><p>Melting points were taken on a Thomas-Hoover melting point apparatus in glass capillary tubes and are uncorrected. 1H NMR spectra were recorded with a Bruker ARX 400 MHz spectrometer with tetramethylsilane (TMS) as an internal standard. Peak positions are given in parts per million (δ). Infrared spectra were obtained on a Nicolet iS10 FT-IR spectrometer. Elemental analyses were performed by Atlantic Microlab Inc. (Norcross, GA) for the indicated elements and results are within 0.4% of calculated values. Reactions were monitored by thin-layer chromatography (TLC) on silica gel GHLF plates (250 µ, 2.5 × 10 cm; Analtech Inc., Newark, DE).</p><!><p>Compound 12 (0.22 g, 1.14 mmol) and free base of 5 (0.25 g, 1.14 mmol) were added to a stirred solution of K2CO3 (0.31 g, 2.26 mmol) and KI (few crystals) in anhydrous MeCN (5 mL). The reaction mixture was allowed to stir in a screw-cap vial at 88 °C for 16 h. The solvent was evaporated under reduced pressure; the residue was suspended in H2O and extracted with CHCl3 (3 × 15 mL). The combined organic portion was washed with H2O (3 × 10 mL), dried (Na2SO4), and evaporated under reduced pressure to yield an ivory-colored solid. The solid was dissolved in EtOH (5 mL) and the solution was cooled to 0 °C. A saturated solution of gaseous HCl in Et2O (2 mL) was added and the mixture was allowed to stand at room temperature for 3 h. The solvent was evaporated to yield a white solid that was recrystallized from EtOH to yield 0.01 g (3%) of 2 as a tan-colored solid: mp 216–218 °C 1H (DMSO-d6): 1.43–1.69 (m, 4H, CH2), 1.82–2.10 (m, 6H, CH2), 2.21–2.24 (m, 2H, CH2), 2.82–2.88 (m, 2H, CH2), 2.82–2.88 (m, 1H, CH), 3.23–3.26 (m, 2H, CH2), 3.40–3.52 (m, 2H, CH2), 3.63–3.66 (m, 2H, CH2), 3.98–4.02 (m, 2H, CH2), 4.48–4.51 (m, 2H, CH2), 7.29–7.38 (m, 1H, ArH), 7.70–7.76 (m, 1H, ArH), 8.05–8.18 (m, 1H, ArH), 10.17 (br s, 1H, NH+) Anal Calcd for (C21H28FN3O2·HCl·0.5H2O) C, 60.21; H, 7.02; N, 9.77. Found: C, 60.32; H, 7.22; N, 10.03.</p><!><p>A 1M solution of BH3 in THF (1.41 mL) was added to a stirred solution of 2 (0.13 g, 0.35 mmol) in anhydrous THF (4 mL) at 0 °C (ice-bath). The reaction mixture was allowed to warm to room temperature and stirring was continued for 10 h. The reaction mixture was carefully quenched with 6M aqueous HCl (0.40 mL) and then heated at reflux for 1 h. The mixture was allowed to cool to room temperature and basified with 1N aqueous NaOH (2 mL). Water (10 mL) was added and the aqueous portion was extracted with EtOAc (3 × 15 mL). The organic portions were combined, dried (Na2SO4), and solvent was removed under reduced pressure to give an oily residue that was dissolved in absolute EtOH, and HCl-anhydrous Et2O was added to afford a solid. Recrystallization from absolute EtOH/anhydrous Et2O gave 0.04 g (26%) of 3 as yellow crystals: mp 270–273 °C; 1H-NMR (DMSO-d6: salt) δ 1.37 (m, 1H, CH2), 1.68–1.78 (m, 10H, CH2), 2.17–2.20 (m, 2H, CH2), 2.38–2.53 (m, 2H, CH2), 2.83 (m, 2H, CH2), 3.01–3.10 (m, 7H, CH, CH2), 3.36–3.46 (m, 1H, CH2), 3.58–3.61 (m, 2H, CH2), 7.32 (td, J = 9.1 2.0 Hz, 1H, ArH), 7.71 (dd, J = 9.0, 1.9 Hz, 1H, ArH), 8.25 (dd, J = 8.6, 5.4 Hz, 1H, ArH), 10.33 (br s, 1H, NH+), 11.06 (br s, 1H, NH+). Anal. Calcd for (C21H30FN3O·2HCl· 0.25H2O) C, 57.73; H, 7.50; N, 9.62. Found: C, 57.36; H, 7.21; N, 9.30.</p><!><p>The free base of 6-fluoro-3-(piperidin-4-yl)benzisoxazole (5; 0.20 g, 0.90 mmol) was added to a stirred solution of HCOOH (0.2 mL, 5.45 mmol) and HCHO (0.2 mL, 5.45 mmol) in EtOH under an N2 atmosphere. The reaction mixture was heated at reflux for 10 h and the solvent was evaporated to yield a white solid. The solid was dissolved in EtOH (5 mL) and cooled to 0 °C. A saturated solution of gaseous HCl in Et2O (2 mL) was added and the mixture was allowed to stand at room temperature for 3 h. The solvent was evaporated and the crude solid was recrystallized from EtOH to yield 0.01 g (4%) of 4 as a white solid: mp 218–220 °C, 1H (DMSO-d6): 2.20–2.25 (m, 4H, CH2), 2.80 (s, 3H, CH3), 3.09–3.19 (m, 2H, CH2), 3.53–3.56 (m, 2H, CH2), 7.33–7.38 (m, 1H ArH), 7.73–7.75 (dd, J= 2, 8 Hz 1H, ArH), 8.15–8.18 (m, 1H, ArH), 10.64 (br s, 1H, NH+). Anal Calcd for (C13H15FN2O·HCl·0.5H2O) C, 55.82; H, 6.13; N, 10.01. Found: C, 55.50; H, 6.08; N, 9.69.</p><!><p>Piperidine (11; 1.2 mL, 11.54 mmol) was added in a dropwise manner to a stirred solution of Et3N (1.2 mL, 11.54 mmol) in CH2Cl2 (25 mL) and the reaction mixture was cooled to 0 °C under an N2 atmosphere. A solution of 4-chlorobutyryl chloride (1.5 mL, 11.74 mmol) in CH2Cl2 (5 mL) was added and the reaction mixture was allowed to stir at room temperature for 75 h and washed with H2O (2 × 25 mL). The organic portion was dried (Na2SO4) and evaporated under reduced pressure to yield a yellow-colored oil, which was purified by vacuum distillation to yield 1.00 g (45%) of 12 as a pale yellow-colored oil: bp 103 °C at 0.02 psi. The oil was used without further characterization in preparation of 2.</p><!><p>Human embryonic kidney (HEK293) cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum at 37 °C in a 5% CO2-humidified atmosphere. Transfection was performed using Lipofectamine 200 reagent (Invitrogen) according to the manufacturer's instructions. The N-terminally c-Myc-tagged form of wild type human 5-HT2A receptor (pcDNA3.1-c-Myc-5-HT2A) has been described previously.21</p><!><p>Radioligand binding assays were performed as described previously21 with minor modifications. Briefly, HEK293 cell pellets were homogenized using a Teflon-glass grinder (10 up-and-down strokes at 1,500 rpm) in 1 mL of binding buffer (see below), supplemented with 0.25 M sucrose. The crude homogenate was centrifuged at 1,000 × g for 5 min at 4 °C, and the supernatant re-centrifuged at 40,000 × g for 15 min at 4 °C. The resultant pellet (P2 fraction) was washed twice in homogenization buffer and re-centrifuged in similar conditions. Aliquots were stored at −80 °C until assay. Protein concentration was determined using the Bio-Rad protein assay.</p><p>[3H]Ketanserin binding was measured at equilibrium in 100-µl aliquots (50 mM Tris-HCl, pH 7.4) of membrane preparations (~ 10 µg of total protein) that were incubated at 37 °C for 60 min. Competition curves were carried out by incubating the tested compound (10−10 – 10−4 M; 14 concentrations) in binding buffer containing 4 nM [3H]ketanserin. Non-specific binding was determined in the presence of 10 µM methysergide. Incubations were terminated by dilution with 200 µl ice-cold incubation buffer and free ligand was separated from bound ligand by rapid filtration under vacuum through GF/C glass fiber filters using a microbeta filtermat-96 harvester (PerkinElmer). These filters were then rinsed twice with 200 µl ice-cold incubation buffer, air dried, and counted for radioactivity by liquid scintillation spectrometry, using a MicroBeta2 detector (PerkinElmer). Radioligand binding data were analyzed using non-linear curve-fitting software (GraphPad Prism).</p><!><p>Oocytes were isolated and microinjected with equal volumes (50 nl) as previously described.14 In all two-electrode voltage-clamp experiments (TEVC), oocytes were injected with 2 ng of 5-HT2A receptor and 2 ng of GIRK4* and were maintained at 18 °C for 1–4 days before recording.</p><!><p>Whole-cell currents were measured by conventional TEVC with a GeneClamp 500 amplifier (Axon Instruments, Union City, CA), as previously reported.14 A high-potassium (HK) solution was used to superfuse oocytes (96 mM KCl, 1 mM NaCl, 1 mM MgCl2, 5 mM KOH/HEPES; pH 7.4) to obtain a reversal potential for potassium (EK) close to zero. Inwardly rectifying potassium currents through GIRK4* were obtained by clamping the cells at −80 mV. Basal GIRK4* currents were defined as the difference between inward currents obtained at −80 mV in the presence of 3 mM BaCl2 in HK solution and those in the absence of Ba2+ and measured for each trace. Current inhibition to 5-HT was measured and normalized to basal current to compensate for size variability in oocytes. Current inhibition of risperidone derivatives was normalized to 1 µM 5-HT to compensate for variability in basal currents.</p><!><p>The 5-HT2A receptor cDNA was subcloned into the pcDNA5/FRT/TO plasmid and stable inducible cell lines were produced using the single site recombination T-Rex Flp-In system as previously described.20 5-HT2A receptor expression was induced adding 1 µg/mL doxycycline to the culture media at least two days before the experiment, and cells were plated on 96-well dishes one day before the experiment. The day of the experiment the cells were switched to serum-free medium for about 3–4 h, before being loaded with 5 µm Fura2-AM (Molecular Probes, OR) in Imaging Solution (125 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgSO4, 6 mM glucose, and 25 mM Hepes/Tris, pH 7.4). Following incubation for 30 min at 37 °C, cells were washed with Imaging Solution and kept at room temperature for about 15 min before being placed on the stage of an epifluorescence microscope, coupled to an automatic perfusion system and controlled by the Live Acquisition Software from Till Photonics (see reference 20 for details). The Fura-2 signal was acquired at 510 nm by switching the excitation wavelength between 340 nm and 380 nm. Baseline was recorded for 30 s before perfusion of 5-HT (1 µM) in the presence and absence of various concentrations of the drugs studied for another 45 s, followed by perfusion of Imaging Solution for 30 s to wash out the drugs. Intracellular calcium concentration was reported as fluorescence ratio (F 340/F380) and values were normalized to the basal F340/F380 ratio level before perfusion of the drugs. Data obtained for each drug in the presence of 5-HT were further normalized to the mean responses elicited by 5-HT alone in the same experiment. Results are expressed as mean ± SEM. Dose-response curves, curve fitting and IC50 values were obtained by analysis using GraphPad PRISM 6 software.</p>

PubMed Author Manuscript

Development and investigation of a site selective palladium-catalyzed 1,4-difunctionalization of isoprene using pyridine–oxazoline ligands

Palladium-catalyzed 1,4-difunctionalizations of isoprene that produce skipped polyenes are reported.Complex isomeric product mixtures are possible as a result of the difficult-to-control migratory insertion of isoprene into a Pd-alkenyl bond, but good site selectivity has been achieved using easily accessible pyrox ligands. Mechanistic studies suggest that the control of insertion is the result of the unique electronic asymmetry and steric properties of the ligand.

development_and_investigation_of_a_site_selective_palladium-catalyzed_1,4-difunctionalization_of_iso

Introduction<!>Results and discussion<!>Conclusions

<p>Terpenoid natural products exhibit a broad range of important physiological effects. In such molecules, isoprene units are oen paired with diverse stereodened di-and tri-substituted alkenes to give rise to a larger class of molecular frameworks known as skipped polyenes (Fig. 1a). [1][2][3][4][5][6][7][8][9][10] Unfortunately, the rapid assembly of stereochemically-dened skipped polyenes remains a signicant challenge in modern synthetic chemistry.</p><p>We have recently reported a Pd-catalyzed 1,4-difunctionalization reaction of 1,3-butadiene with alkenyl triates, and aryl or alkenylboronic acids that enabled the rapid assembly of diverse skipped polyene frameworks. 11 In an effort to advance this strategy to directly access skipped polyene-containing terpenoid fragments in a single step, we sought to utilize isoprene as the 1,3-diene substrate (Fig. 1b). The effective use of isoprene in such a three-component coupling reaction would require us to address the added challenge arising from the use of a diene substrate containing two similar, yet distinct alkenes. As others have reported, 12 site selective 1,4-addition (as opposed to 4,1-addition, Fig. 1c) to isoprene is dependent on a difficultto-control alkene migratory insertion into the cationic Pdalkenyl intermediate A. The desired 1,4-difunctionalization product D is hypothesized to be accessed upon insertion of the less-substituted alkene of isoprene to give a cationic p-allyl palladium intermediate (B). This is followed by transmetallation with an alkenylboronic acid, and reductive elimination to yield D (or the 1,2-addition product E). Isomeric mixtures derived from difunctionalization reactions of simple 1,3-dienes give 1,2-products typically as the minor isomer, with some notable exceptions. 13 Additionally, formal 4,3-addition products (not shown) are not frequently observed as a result of the opposite alkene insertion pathway (A / C / F), likely due to a high barrier for reductive elimination of a quaternary center from palladium. In total, ve distinct constitutional and stereoisomers [11][12][13][14] can be formed in this three-component coupling reaction through likely energetically similar pathways. The cationic nature of the palladium catalyst, resulting aer oxidative addition of an alkenyl triate, is proposed to account for the high selectivity of three-component coupling products rather than Heck or Suzuki products (not shown). 11,14 Herein we report a Pd-catalyzed 1,4-difunctionalization, which utilizes pyrox ligands to afford a site preference for isoprene migratory insertion.</p><!><p>Starting with the reaction conditions we reported for the Pdcatalyzed 1,4-difunctionalization of 1,3-butadiene, 11 only minor adjustments were required to increase yield and selectivity for the formation of the (E)-4a. 15 Most signicantly, an increase in the amount of isoprene (7.0 equivalents) was required to achieve a good yield of alkene difunctionalization products when coupling to cyclohexenyl triate (1a) and styrenylboronic acid (3a) with isoprene. Under the indicated optimal conditions, good yield and selectivity for the formation of (E)-4a was observed (Table 1). Of note, all of the isomeric products can be separated by HPLC and a 61% yield of the desired isomer is achieved.</p><p>The assessment of alkenyl triates yielded products incorporating a tetrahydropyranyl (4b), a N-protected piperidinyl (4c) heterocycles and a simple aliphatic (4d), which revealed similar selectivity for the formation of (E)-4 relative to (E)-6 (>6 : 1). Under the standard optimized conditions, lower yields were observed of skipped polyene products when using a conjugated ester derived boronic acid to yield 4b. We hypothesize that this could be due to either catalyst inhibition by the boronic acid or the product through competitive binding to the catalyst. By simply increasing the concentration of the alkenyl triate, signicantly improved yields were observed (4b-e). 16 Next, we varied the alkenylboronic acid component, specically coupling styrene-(4f-h), N-heterocycle-(4i), and alkyl halide-containing (4j) boronic acids to an electron-decient alkenyl triate (1e) and isoprene. Triate 1e was evaluated because the vinylogous lactone is an attractive synthetic handle for further elaboration. While the boronic acids minimally inuenced the product yields and isomeric ratios, alkenyl tri-ate 1e proved to impact selectivity greatly compared to other alkenyl triates (4f-j). Specically, the ratio of (E)-4 : (E)-6 was reduced to nearly 1 : 1 in these examples. This shortcoming is addressed below.</p><p>Other important observations regarding the Pd-catalyzed three-component difunctionalization of isoprene can be discerned from Table 1. For example, (E)-4h was formed in adequate yield from 1-phenylvinylboronic acid. This is somewhat surprising since 1,1-disubstituted terminal alkenes are known to undergo migratory insertion into Pd-alkenyl bonds like that of A (Fig. 1c), 12,13a,17 but are tolerated under our reaction conditions. Additionally, higher selectivity for (E)-4h and (E)-4i over 1,2-addition products 5h and i (8.8 : 1 and 11 : 1 respectively) is observed for nonlinear boronic acids coupling partners (compared to other alkenylboronic acids, all of which are linear). This may be due to an added steric inuence on the Pdp-allyl (B in Fig. 1c), thereby promoting 1,4-addition as compared to 1,2-addition. 18 Interestingly, the relative abundance of the (E)-stereoisomers of 4 and 6 were only modestly dependent on the coupling partners used in these reactions: (E)-4 consistently predominated over (Z)-4, while (E)-6 and (Z)-6 were consistently formed in nearly equal amounts. An explanation for this observation stems from the steric environment about the relevant s-allylstabilized palladium intermediates (Fig. 2a, compare G to H and I to J), which can interconvert through a s / p / s process. [11][12][13][14]18 A simplied example of this is shown in Fig. 2b, wherein s-allyl palladium intermediate K (precursor to the (Z)stereoisomer) can form s-allyl palladium P (precursor to the (E)stereoisomer) aer a key bond rotation that occurs between intermediates M and N (Fig. 2). Similar processes occur for isomer formation in the 4,1-addition scenario through intermediates I and J to form (Z)-6 and (E)-6, respectively. The comparable steric environment about s-allyl intermediates that lead to (E)-6 or (Z)-6 accounts for the nearly equal amounts of the 4,1-stereoisomers.</p><p>Table 1 Initial scope of the Pd-catalyzed 1,4-difunctionalization of isoprene a a Yields are reported as a combination of isomers of reactions performed on a 0.5 mmol scale. Structures of isomers were conrmed by separation using HPLC and NMR analysis. Isomeric ratios were determined by either 1 H NMR or HPLC. b Reaction performed with 3.0 equivalents of 1 and 1.0 equivalents of 3. c (Z)-4d and (Z)-6d were inseparable by HPLC and 1 H NMR signals overlapped. Thus, values are reported as a mixture.</p><p>The aforementioned undesired 1 : 1 ratio of (E)-4 : (E)-6 suggests indiscriminate alkene insertion into the Pd-alkenyl intermediate A (Fig. 1c). From an electronic prospective, tri-ates 1a-d are comparatively electron-rich (unstabilized) as opposed to 1e, which may play an important role in selective alkene insertion into the Pd-alkenyl intermediate A (i.e., (E)-4a : (E)-6a 7.3 : 1 compared to (E)-4e : (E)-6e 1.5 : 1). Due to the potential utility of electron decient (stabilized) alkenyl tri-ates, we sought to address the challenge of indiscriminate alkene insertion, and, in doing so, potentially regain control over the selectivity.</p><p>Electronically-stabilized alkenyl groups would be expected to render the cationic Pd-intermediate A (Fig. 1c) more sensitive to the differential nature of the two alkenes in isoprene, with insertion of the more electron-rich disubstituted alkene of isoprene resulting in increased formation of the undesired 4,1products. In this scenario, a ligand might easily override the inherent electronic bias of substrate insertion. Fortunately, in the course of our initial reaction optimization to produce 4a, we observed that several ligands were tolerated, although their use led to similar product distribution and generally lower yields as compared to the "ligandless" conditions ultimately employed in Table 1. 15 Thus, we sought to evaluate the propensity of ligands to override the 1 : 1 selectivity of (E)-4 : (E)-6 products observed for 4e-j.</p><p>To explore this possibility, we evaluated the ability of ligands to empower the selective formation of (E)-4f (Table 2). 19,20 The use of monodentate ligands, 4-dimethylaminopyridine (DMAP) or a simple oxazoline (L1 and L2), afforded little enhancement in the isomeric ratios as compared to the "ligandless" conditions. The bidentate quinolone-oxazoline (quinox) ligand (L3) led to a two-fold increase in selectivity of (E)-4f over (E)-6f (2.1 : 1), albeit in low yield. A related ligand, chiral pyridineoxazoline (pyrox) L4, which has been used with success in recent Pd-catalyzed redox-relay Heck reactions in our lab, 21 signicantly enhanced selectivity between (E)-4f and (E)-6f (7.7 : 1). Evaluation of L5 reveals the importance of one unhindered catalyst face, as a low yield and modest selectivity is observed when a geminal dimethyl-substituted ligand is used. Other similar N,N-type ligands were examined including the CF 3 -substituted quinox L6 and the 6-CF 3 -substituted pyrox L7, both of which afford low selectivity. Lastly, an interesting result is observed when ligand L8 is employed: selectivity between the three major regioisomers favor the formation of 5f (1,2-addition product), albeit in reduced yield. Unfortunately, the selectivity between the 1,4-addition and 1,2-addition products (4f and 5f) did not exceed 4 : 1 with any of the ligands that were evaluated.</p><p>We next sought to investigate the mechanistic basis for the putative site-selective migratory insertion that occurs in the presence of L4. Using the same reactants and conditions found in Table 2, a library of easily-accessible pyrox ligands was evaluated with varying oxazoline substitution. From these experiments, a trend in alkene insertion was observed based on the size of the R-substituent on the oxazoline portion of the ligand (Fig. 3a). Of the ligands evaluated, those featuring smaller Rgroups afforded diminished selectivity for the formation of the 1,4-and 1,2-products compared to the (E)-4,1-product (3.6 : 1 for R ¼ H, compared to 15 : 1 for R ¼ t-Bu). Indeed, a correlation of the logarithm of product selectivity (corresponding to the presumed relative rate of insertion) versus Sterimol B 1 values 22 (minimum radius corresponding to the pyrox ligand R-substituent) is observed. This suggests that the oxazoline's steric environment is partially responsible for the observed site selection.</p><p>Table 2 Ligand evaluation for the Pd-catalyzed difunctionalization of isoprene a a Yields are reported as a combination of isomers for reactions performed on 0.2 mmol scale. Yields and isomeric ratios as determined by 1 H NMR using an internal standard. Satised with our observation of this clear ligand steric effect, we turned to our recently-developed methodology of combining design of experiments with multi-parameter ligand modulation. 23 In this case, we evaluated both electronic effects on the pyridine ring and steric modications on the oxazoline simultaneously. By using such an experiment, we hoped to ascertain which of these factors was most inuential in the observed selectivity for 1,4-and 1,2-addition versus 4,1-addition. To analyse the results, ratios between 4f, 5f, and (E)-6f were normalized and plotted by way of a ternary plot as a means to identify general trends in the isomeric distribution (Fig. 3b). As illustrated, the highest selectivity for 1,4-addition (4f) is observed with ligands combining bulky R-substituents and electron-decient pyridine rings. To delineate these observed steric and electronic effects on the reaction outcome, a precise mathematical model was needed; Sterimol and Hammett values were chosen as our respective descriptors. We then employed a standard stepwise linear regression algorithm to expedite statistical exploration of the relationship between these parameters and DDG ‡ (experimentally-derived, and equalling ÀRT ln(4f + 5f : (E)-6f), where R is the ideal gas constant and T is temperature). The resultant normalized equation and a plot of measured versus predicted DDG ‡ values is depicted in Fig. 3c and d. The relatively high R 2 -value as well as the slope nearly equal to one, validates the strength of the model.</p><p>To evaluate the inuence of each specic parameter, the coefficients can be compared. As observed in Fig. 3a, the largest coefficient belongs to the Sterimol B 1 parameter (minimum radius of the oxazoline R-substituent), again suggesting a strong inuence of substituent size on site selection. The coefficient relating to Hammett s-values gives us information that the electronics of the pyridyl ring are important to inuencing selectivity as well. As articulated in previous studies 21, 24 from our lab, these results suggest that the electronic asymmetry of the pyrox ligands likely impart control of the catalyst coordination sphere, thus inuencing insertion site selection, and therefore product outcome. Furthermore, virtual extrapolation of the model to predict a better catalyst was investigated, but no reasonable improvements could be identied following intuitive and accessible changes to the ligand structure.</p><p>Based on the above-presented studies, a mechanistic model is proposed for the observed ligand-control over alkene insertion into a Pd-alkenyl bond. Previously reported computations on Heck reactions using L4 suggest the coordination environment of a cationic Pd-aryl intermediate following oxidative addition is rapidly isomerizing and, as such, the coordination environment would be dictated by the relative energies of the different alkene insertion transition states. 25 In that report, the transition states for insertion were computed to be more favourable by nearly 1 kcal mol À1 for the complexes wherein the oxazoline and the arene are oriented trans to one another. For our purposes, this type of orientation could also explain the signicant inuence of the oxazoline substituent on the observed selectivity of isoprene insertion into a Pd-alkenyl intermediate. The steric effect generated by the tert-butyl group of the ligand and the highly electrophilic nature of palladium (aided by the electron-decient pyridyl group of the ligand) likely promotes rapid alkene association/dissociation, such that each of the corresponding coordination complexes to the proposed transition states is in equilibrium (Fig. 4). Thus, we propose that the transition state of the selectivity-determining step is controlled by the inuence of the tert-butyl group on isoprene. Assuming coordination of the alkene trans to the pyridine ring, the preferred insertion should occur through A ‡ , wherein the likely steric interactions of isoprene insertion are minimized, consistent with the correlative information obtained. This mechanistic model contrasts the original hypothesis (Fig. 1c), wherein isoprene binds in a cisoid-type coordination mode.</p><p>Further evaluation of the raw data obtained in the ligand survey revealed an additional relationship between the formation of the 1,4-and 1,2-addition products. These products arise from a common intermediate and are only differentiated by the reaction of the p-allyl-stabilized palladium intermediate with a transmetallation partner prior to reductive elimination (Fig. 1c, B / D + E). The observed correlation shows a modest electronic effect on the ratio of products, which can be quantied using Hammett s-values of the pyridyl substituent (Fig. 5). Specically, greater amounts of the 1,4-addition product are formed as the catalyst becomes more electron decient, although overall yield sharply decreases for pyrox ligands bearing either a 5-CN or a 5-NO 2 group (not shown). While the origin of this effect is not clear and will require further investigation, the discovery of electronic control will likely impact future ligand design.</p><p>Aer evaluating a diverse group of pyrox ligands (Fig. 3b), L4 remained the most selective and also provided modest yields of the desired skipped polyene products. A brief re-optimization of the reaction conditions resulted in lowering the stoichiometry of L4 as well as increasing the reaction temperature. 26 Select reactions presented in Table 1 were repeated under the new conditions to evaluate the utility of L4 toward a more selective process (Table 3). We were pleased to see selectivity improve throughout, and especially for the "stabilized" alkenyl triates. In conjunction with enhanced selectivity, yields were also similar when using L4. However, the formation of the 1,2addition product still accounts for a considerable amount of the mass-balance, reducing the ability to access the desired product, (E)-4, in higher yields.</p><!><p>In the course of developing this Pd-catalyzed difunctionalization reaction of isoprene, a major inuence of coupling partner electronics upon the selectivity of alkene migratory insertion was identied. Through a series of studies, we were able to identify L4 as a ligand that could enhance control of site selective isoprene insertion into a palladium-alkenyl intermediate. A library of pyrox ligands were examined in order to provide information that ultimately led to a mechanistic rationalization of the role that L4 plays in controlling alkene insertion. The re-examination of reactions in the presence of ligand afforded better selectivity for the desired 1,4-addition in generally reasonable yields. A limitation in this chemistry remains the confounding formation of 1,2-addition products as a result of a p-allyl-stabilized cationic Pd-intermediate, regardless of the ligand used. Understanding and controlling reaction outcomes of such Pd-p-allyl intermediates in alkene difunctionalization reactions is a focus of on-going studies in our lab.</p>

Royal Society of Chemistry (RSC)

Recognition of dextran-superparamagnetic iron oxide nanoparticle conjugates (Feridex) via macrophage scavenger receptor charged domains

Dextran-coated superparamagnetic iron oxide nanoparticles (dextran-SPIO conjugates) offer the attractive possibility of enhancing MRI imaging sensitivity so that small or diffuse lesions can be detected. However, systemically injected SPIO are rapidly removed by macrophages. We engineered embryonic cells (HEK293T) to express major macrophage scavenger receptor (SR) subtypes including SR-AI, MARCO, and endothelial receptor collectin-12. These SRs possess a positively charged collagen-like (CL) domain and they promoted SPIO uptake, while the charge neutral lipoprotein receptor SR-BI did not. In silico modeling indicated a positive net charge on the CL domain, and a net negative charge on the cysteine-rich (CR) domain of MARCO and SR-AI. In vitro experiments revealed that CR domain deletion in SR-AI boosted uptake of SPIO 3-fold, while deletion of MARCO\'s CR domain abolished this uptake. These data suggest that future studies might productively focus on the validation and further exploration of SR charge fields in SPIO recognition.

recognition_of_dextran-superparamagnetic_iron_oxide_nanoparticle_conjugates_(feridex)_via_macrophage

Introduction<!>Preparation and storage of nanoparticles<!>Nanoparticle size determination<!>Nanoparticle structure determination<!>Scavenger Receptor Gene Cloning, Amplification, and Expression in HEK293T cells<!>SR MARCO<!>SR-AI (CR domain +) and SR-AII (CR domain \xe2\x88\x92)<!>SR-BI and Collectin-12 (endothelial)<!>Transfection of HEK293T cells<!>Determination of receptor expression on the HEK293T cell surface<!>SPIO uptake and binding experiments<!>Uptake measured by optical absorbance<!>Microscopy-based measurements of SPIO uptake<!>SPIO binding to type Icollagen<!>In silico modeling of SR domains<!>SPIO Nanoparticle Size and Structure<!>SPIO uptake by macrophages is mediated via SRs and blocked by charged polymers<!>Charged SRs exhibited greater SPIO uptake<!>SPIO binds to type I collagen<!>MARCO and SR-AI exhibited different uptake<!>Predicted charge differences between SRs<!>CL/CR domain effects on SPIO uptake in SR subtypes<!>Discussion

<p>Magnetic resonance imaging (MRI) is a modality that has long attracted considerable interest for early disease detection and staging. However, to be useful for small, indistinct lesions MRI often requires enhancement (1). FDA approved superparamagnetic iron oxide (SPIO) contrast agents are the most detectable and effective medium for enhancing contrast in MRI acquisitions (2–6). A major difficulty with current SPIO formulations, viz., dextran coated SPIO nanoparticles, is that they are avidly taken up by macrophages, which results in premature clearance, reduced effectiveness of imaging, and toxicity (7–10). The mechanism of macrophage SPIO uptake is incompletely characterized, and we along with others (11–13) have sought to further elucidate this process to identify possible inhibitory strategies.</p><p>Much remains to be learned about the key specifics of particle uptake by macrophages, although it is known that uptake is opsonin-independent for many types of nanoparticles, including gold, silica (14–16), polystyrene (17) and liposomes (18). SPIO-dextran uptake has recently been shown by us to be complement and IgG-independent in mouse knockout models (19). Although several classes of macrophage receptors could conceivably mediate SPIO recognition, multiple studies have collectively provided convincing data that SPIO uptake is coordinated by scavenger receptors (SRs) (11, 12).</p><p>What remains to be elucidated is the specific mechanism of SPIO recognition by SRs. This is important as it may provide an opportunity for selectively inhibiting or manipulating SPIO nanoparticle uptake by macrophages. Current evidence reveals that often it is polyanionic pathogen-associated molecular patterns that are eliminated by SRs (12, 15, 17, 20–22). Moreover, the major macrophage SR AI subtype (SR-AI) recognizes negatively charged surfaces via a positively charged collagen-like domain (CL-domain) (12, 17, 20, 23). Therefore we hypothesized that SPIO-dextran nanoparticles, which are weakly anionic, are recognized by the CL-domain of SRs. We further hypothesized that this mode of recognition might be differentially modulated between various SR subtypes.</p><p>Our hypotheses were addressed using in silico modeling of the SR domain charge field, together with a defined system in which human embryonic kidney cells (HEK293T) were engineered to individually express the major macrophage SR subtypes SR-AI, MARCO, SR-PSOX, SR-BI, and the primary endothelial SR collectin-12 (CL-P1). The experiments indicated that the net positively charged collagen-like domain mediated SPIO uptake by SR-AI, and that deletion of a net negatively charged cysteine rich (CR) domain adjacent to the CL domain differentially affected SPIO binding to SR-AI versus MARCO.</p><p>The results of this modeling and in vitro study provide an essential step for followup investigations of SR mediated macrophage nanoparticle recognition, and a comprehensive validation of various SR domains according to the fine structure of charge fields. This will lead to strategies for (i) inhibiting nanoparticle clearance, (ii) minimizing undesired labeling of macrophages, and (iii) targeting specific subpopulations of macrophages,</p><!><p>Commercial Feridex I.V.® nanoparticles were used for this study and were obtained from a commercial source on our behalf by the UCSD Department of Radiology. Feridex consists of a suspension of SPIO-dextran composites. Each composite is 50–160 nm across and contains multiple SPIO particles approximately 5–6 nm in diameter embedded in a meshwork of linear dextran (10 kDa, T-10). Particles were resuspended in PBS at 1–2 mg (Fe)/ml, filtered through a 0.2 μm membrane filter, and stored at 4°C.</p><!><p>The size distribution and z (zeta)-potential of diluted aliquots of the nanoparticle suspension was measured with a Zetasizer Nano (Malvern, UK). To determine any effects of adherent plasma proteins on nanoparticle size, SPIO was mixed with citrated mouse plasma (1:3 volume ratio), incubated for 10 minutes, and applied to a MINI magnetic column (Miltenyi Biotech), then eluted and sized.</p><!><p>Nanoparticle structure was confirmed using transmission electron microscopy; 5 μl of 0.5 mg/ml of SPIO-dextran in double distilled water was applied to Formvar/carbon coated grids (Ted Pella, Redding, CA). After drying grids were viewed using a JEOL 1200EX II (JEOL, Peabody, MA) transmission electron microscope at 75 keV and photographed using a Gatan digital camera (Gatan, Pleasanton, CA).</p><!><p>In order to provide a general characterization of SPIO recognition according to major SR subtypes, we transfected HEK293T cells using equal amounts of constructs coding for the following receptors: SR-AI which is expressed on macrophages and monocytes (24); MARCO which is a Macrophage Receptor with Collagenous structure expressed on macrophages resident in the lung alveoli, in the spleen, and in the liver (Kupffer cells) (25); lectin SR Collectin-12 (collectin placenta 1 or CL-P1, expressed on endothelial cells) (26); chemokine SR for phosphatidylserine and oxidized lipoproteins (PSOX/CXCL16, expressed on dendritic cells and atherogenic macrophages (27)); and ubiquitous lipoprotein receptor SR-BI (28). The details of cloning, amplification and expression are given below.</p><!><p>A pCMV6-AC plasmid carrying full-length human macrophage receptor with collagenous structure (MARCO) as transfection-ready DNA (Catalog SC319619) was purchased from OriGene (Rockville, MD). Human MARCO with the truncated cysteine domain and MARCO with charged collagen were designed with flanking BamHI and XhoI restriction sites and synthesized by Epoch Life Sciences (Missouri City, TX); see supplement for sequences of the construct and primers. The inserts were cloned into pCDNA3.1+(Zeo).</p><!><p>Full-length mouse cDNA of SR-AI (splicing variant A, NM_001113326) was amplified from mouse liver mRNA by RT-PCR and cloned into a pCDNA 3.1+ plasmid using the BamHI and XhoI restriction sites. pCMV-SPORT6 plasmid encoding mouse scavenger receptor SR-AII (splicing variant B; without the cysteine-rich domain) was obtained from ATCC (Catalog MGC-6140), PCR amplified, and cloned into pCDNA 3.1. All inserts were verified by forward and reverse sequencing. Full details for the primers are provided in the Supplemental Methods.</p><!><p>A transfection-ready clone encoding human macrophage scavenger receptor SR-BI was obtained from ATCC in pCMV-SPORT6 cassette. A transfection-ready human Collectin Placenta 1 (CL-P1) or Collectin-12 in pCMV6-XL4 cassette was obtained from OriGene.</p><!><p>HEK293T (ATCC) were transiently transfected using Lipofectamine 2000 (Invitrogen) with 0.5 μg of receptor plasmids or empty vector pCDNA 3.1 (Invitrogen) per 1×106 cells in 24 well plates. The mRNA expression of SR-AI, MARCO, CL-P1 and SR-BI was compared via quantitative PCR as described in Supplemental Methods.</p><!><p>Comparisons of SPIO binding between receptor subtypes required equal SR densities on the cell surface to be valid, or a measure of relative density by which to normalize binding data between receptor subtypes. Our assessment of receptor expression did not rely on one type of measure alone, and was determined using three different approaches for cross validation. First, we used immunostaining with MARCO rabbit anti-human polyclonal AP9891a, SR-BI rabbit anti-human monoclonal AJ1734a (both from Abgent, San Diego, CA), rat-anti mouse SR-AI (R&D). The second approach involved Western blotting using rabbit anti-human MARCO (Abgent) and rat anti-mouse SR-AI (AbD Serotec MAB 1322). Cells were transfected with SR-AI and MARCO and fractionated with Mem-PER membrane isolation kit (Thermo). The quality of fractionation was determined with anti-HSP90 antibody (Cell Signaling Technology, Danvers, MA) as a cytoplasmic marker and anti-alpha1 sodium-potassium ATPase antibody (Abcam, Cambridge, MA) as a membrane marker. The level of SR-AI and MARCO in each fraction was compared with western blotting using the above-described antibodies. The final measure of cell receptor expression was based on real time PCR of the receptor transcripts (details in the Supplemental data).</p><!><p>Nanoparticle uptake experiments in receptor-transfected HEK293T cells and in J774A.1 cells were performed similarly.</p><!><p>Cells in 24 well plates were incubated with 0.1 mg/ml SPIO-dextran nanoparticles for 2 h in complete medium at 37° C. For the ligand inhibition experiment, polyinosinic acid, dextran sulfate 500-kDa, fucoidan, dextran or gelatin (all from Sigma) were incubated with J774A.1 cells for 15 min prior to the addition of nanoparticles. SPIO uptake was quantified by adding 200 μl of QuantiChrom Iron Assay reagent (BioAssay Systems), overnight incubation, and measurements of relative absorbance (570 nm). For cell binding experiments, SPIO was added to cells at 4°C at 3-fold higher concentration (0.3mg/ml) for 15min, washed and assayed as above.</p><!><p>Cells were seeded into 8-well chamber slides (NalgeNunc) and incubated for 2 h with 0.1 mg/ml SPIO, washed, fixed with 4% formaldehyde, and stained with Prussian blue dye (29) to visualize iron inside the cells.</p><!><p>To determine whether SPIO does in fact bind to collagen, which comprises the CL domain of SRs, microwell 96-well plates (Costar) were coated with either calf skin type I collagen (100 μg/ml 0.1 M acetic acid/PBS; Sigma, C3511) or BSA/PBS (controls) and blocked with 1% BSA/PBS. SPIO solution (3, 10, 30 or 100 μg/ml of Fe in PBS) was added and bound iron quantified using the QuantiChrom Iron Assay as described previously.</p><!><p>We sought to predict whether SR domains might possess charge differences that form the basis of differential recognition between SR receptor subtypes. Therefore, modeling of the SR-AI and MARCO electrostatic charge fields was performed using the homology module of Insight II software (Accelrys, San Diego) on the San Diego supercomputer at UCSD. Collagen-like domains were modeled based on the type I collagen (pdb ID 3HQV). The cysteine-rich domain models were based on the crystal structure of the monomeric cysteine-rich domain of mouse MARCO (pdb ID 2OY3). Equipotential maps of receptor charges were plotted using the NAMD program (30). For human SR-BI, the charge was estimated as the number of charged ionizable residues (positive, ARG and LYS (+1); negative, GLU and ASP (−1)) per 10 residues at pH 7.5.</p><!><p>We used the commercial SPIO MRI contrast agent Ferridex I.V., which is essentially a SPIO-dextran nanoparticle, to study SPIO recognition by macrophage SRs. The measured diameter of the SPIO dextran particles (composites) was between 80–150 nm (average 112 nm), and zeta potential was slightly negative, −13 mV (Fig. 1A–B). TEM confirmed our expectation that the particles consisted of several 5 nm cores of crystalline iron oxide embedded in a meshwork of dextran. The clusters exhibited a worm-like shape, although irregular aggregates were also visible (Fig. 1C).</p><!><p>Previously reported SR-dependent recognition/uptake of Feridex by J774a.1 macrophages was confirmed by inhibition experiments using various SR ligands (Fig. 2). Addition of the polyanionic scavenger receptor ligands fucoidan (10 μg/ml) or dextran sulfate (3 μg/ml) to J774A.1 macrophages prior to the addition of SPIO produced up to 80% inhibition of SPIO uptake (P value=0.0003 for dextran sulfate). The addition of positively charged gelatin (hydrolyzed collagen) at 1 mg/ml also inhibited the uptake, whereas branched 20kDa dextran at 1 mg/ml had no effect.(data not shown) The suppression of Feridex uptake by known SR polyanionic ligands confirms that in J774A.1 macrophages recognition is mostly SR-dependent, and inhibition by both positively and negatively charged polymers provides a preliminary indication that charge interactions may play an important role in the uptake.</p><!><p>Transiently-transfected MARCO, SR-BI, CL-PI showed similar levels of mRNA expression based on quantitative PCR, while SR-AI showed about 50% lower mRNA expression (Supplemental Fig. S1A–B). For SR-AI, SR-BI and MARCO, we also verified the expression with western blotting and confirmed membrane expression with immunostaining (see Supplemental Fig. S2 and data below).</p><p>According to the Prussian blue staining and iron quantification (Fig. 3A–B), SR-AI, CLP1, SR-PSOX and MARCO, which have positively charged SRs, significantly promoted the binding and uptake of SPIO by HEK293T cells (5–20-fold increase compared to vector-transfected), while the charge neutral SR-BI did not show any uptake despite being expressed on the cell surface (Supplemental Fig. S2). According to charge calculations, the SR-BI extracellular domain has a neutral charge of +0.05 and no charged domains were detected in the sequence.</p><!><p>SR-AI, MARCO and CL-P1 each possess a positively charged CL domain (Fig. 4A) and we sought to determine if the SPIO-dextran nanoparticle could bind collagen alone. Collagen type I has a cationic heparin-binding site with affinity of 150nM (31, 32). We tested the binding of Feridex to a collagen type I-coated plate. There was a concentration-dependent binding that was completely inhibited by addition of polyanionic dextran sulfate 500kDa to the plate prior to the addition of nanoparticles (Fig. 4B, Supplemental Fig. S3).</p><!><p>SPIO uptake by Prussian blue (Fig. 5A) showed major differences between SR-AI and MARCO (quantitatively 8-fold difference, Fig. 3A). We tested whether this difference is due to the unequal cell surface expression. Immunostaining of the cell surface receptors (Fig. 5A bottom panel) and western blot analysis of the membrane fraction (Fig. 5B) showed that both receptors are expressed on the cell surface at the approximately same level. The binding of SPIO to receptor-transfected cells showed about 5-fold higher binding to SR-AI than to MARCO (Fig. 5C).</p><!><p>We sought to determine whether differences between SR-AI and MARCO are due to differences in SR domains. As an initial strategy we modeled charge, which is the most obvious, but not necessarily the only, possible basis for any differential domain effects on SPIO binding. After predicting possible differences between the domains, according to charge, we were prompted to follow up with uptake experiments in which the CR domain was either added or deleted.</p><p>Using well-validated and widely used structural modeling approaches and software (see Methods), 3D crystal structure-based models of the extracellular region and its electrostatic profile were built for human SR-AI and for murine MARCO (Fig. 6). The crystal structure of human MARCO is not available, but murine and human MARCO share >70% sequence identity (33) and 96% charge amino acid identity (Supplemental data). According to the 3D equipotential (from +1.8 to −1.8) surface charge map for both SRs (Fig. 6), the CL domain positive charge field extends well outside the protein backbone. In contrast the C-terminal CR domain of both SRs had a negative charge field (Fig. 6). Interestingly the net positive charge field of MARCO was lower than SR-AI, suggesting the possibility of differential effects on recognition by the CR domain.</p><!><p>Functional tests were made to determine in vitro whether the CL versus thr CR domain have differential effects on SR - SPIO binding. SR-AI and MARCO were compared since they are structurally very similar and we made the working assumption that uptake differences between them would be more influenced by the CL and CR domains rather than overall SR structure.</p><p>We prepared shortened isoforms of SR-AI (known as SR-AII) and MARCO in which the entire CR domain was deleted (Fig. 7A). In addition, we prepared a chimeric MARCO/SR-AI receptor by fusing a highly charged collagen fragment of SR-AI to the C-terminal part of MARCO's CL domain (Fig. 7A). This charged collagen fragment is known to mediate the binding of acetylated LDL to SR-AI and to inhibit ligand binding (23). The expression of the constructs on the cell surface was verified with western blotting and immunostaining (Fig. 7B–C, Supplemental Fig. S4).</p><p>The deletion of the CR domain in SR-AI boosted SPIO uptake almost 3-fold (Fig. 8A). The cell binding studies performed at 4°C showed similar effect of CR domain deletion on SR-AI mediated uptake (Fig. 8B). Deletion of MARCO's CR domain abolished the uptake (Fig. 8C), whereas fusion of the SR-AI collagen fragment to MARCO increased MARCO uptake 2-fold (Fig. 8C). These results suggest that the CL and CR domains on different SRs may produce a different net effect, and future experiments will examine in greater detail the mechanims of iron nanoparticle recognition in the context of domain charge fields in different SRs.</p><!><p>SPIO nanoparticle-based contrast media for MRI are recognized and scavenged by macrophages. This clearance of circulating contrast agent reduces the amount of MRI label available for target tissue contrast enhancement. Moreover interpretation of MRI volumes becomes complicated because the scavenging macrophages end up being labeled and detected. This obstacle to effective MR imaging may potentially be addressed by selective macrophage SR inhibitors, but currently available polyanionic SR antagonists are not selective and are toxic. The present study was intended to contribute to our understanding of SPIO recognition to provide a basis for further work leading to selective inhibitors of SR mediated SPIO recognition.</p><p>Our data confirmed previous observations that that SPIO uptake by macrophages is mediated by SRs (11, 12), and demonstrated that physical differences in collagen-like (CL) and cysteine-rich (CR) domains, very likely charge related, may provide a basis for differential recognition and uptake of SPIO and other nanoparticles according to SR subtype. These results provide an initial point of departure more detailed investigations of SR recognition and specificity, and the precise contribution of charge field fine structure.</p><p>Our test platform was the embryonic kidney cell (HEK293T) expression system and we were able to acquire consistent results and successfully confirm previous reported findings for SRs and SPIO. SPIO nanoparticles were recognized by two major and well defined SRs, SR-AI and MARCO, by the SR Collectin-12, which is expressed on endothelial cells (26), and by SR-PSOX, which is mainly expressed on dendritic cells and macrophages in atherosclerotic lesions.</p><p>According to our in vitro uptake data, and in agreement with our hypothesis that the CL domain plays a key role in SPIO uptake, the CL domain appears to involve mediate recognition of SPIO nanoparticles by SR-AI. We found that the common feature of tested receptors that bound and internalized SPIO-dextran was the presence of a positively charged CL domain, and in agreement with this observation it has been reported that the positively charged fragment of the SR-AI CL domain mediates the binding of polyanionic ligands and negatively charged, acetylated LDL (23).</p><p>Importantly, there was a distinct difference in the uptake mechanism between SR-AI and MARCO despite their similar overall structures. Differences in recognition between these two receptor subtypes is suggested by previous reports indicating that MARCO recognizes other anionic nanoparticles, and recognizes larger iron oxide particulates4,22. These was no evidence of iron aggregation, which might favor MARCO, in our experiments, and transiently transfected MARCO in fact had lower uptake than SR-AI, despite the two having similar levels of cell surface expression. The reason for this difference is not clear but could be related to the difference in charge properties of the receptor domains. Deletion of the CR domain from SR-AI boosted uptake, and a similar effect was observed by Doi et al. (23) for the uptake of acetylated lipoprotein. However, deletion of MARCO's CR domain abolished uptake, indicating a differential effect of the CR domain depending on SR subtype.</p><p>Our in silico models show a net negative charge for the CR domain, and other studies have reported that MARCO's CR domain has a fine structure comprised of positive and negative subdomains (34). This could influence nanoparticle recognition in subtle ways. Bacteria and nanoparticle uptake studies of MARCO and SR-AI have uncovered different SR specificities (25, 34, 35), and deletion of the MARCO CR domain abolishes the uptake of bacteria (34). Since MARCO may recognize SPIO-dextran and other particles through a somewhat different binding mechanism than SR-AI, there is a need for future, specific in vitro experiments that would be conceptually linked with what is known about SR behavior and with further, detailed modeling of the fine structure of charge domains and charge interactions between closely spaced receptors. For example, SRs have been reported to form clusters on the cell surface (36), and such receptor clustering could have important implications on local charge fields and on SPIO binding.</p><p>In conclusion, we demonstrated that a variety of SRs are able to promote the uptake of dextran-coated SPIO (Feridex) in vitro, with the positively charged CL domain largely responsible for uptake by SR-AI. The precise role of the CR domain and charge fields in SPIO uptake by different SR subtypes remains to be elucidated. Specific further investigations suggested by the present study may provide a basis for defined strategies by which to selectively inhibit macrophage uptake of SPIO.</p>

PubMed Author Manuscript

Soft-photoconversion using floating self-assembled crystalline films of porphyrin nanostructures

One of many evolved functions of biological cell membranes is to induce and regulate selfassembly of photoactive molecules into efficient light harvesting nanomaterials. Synthetic molecular assemblies at soft interfaces exhibit macroscale long-range order and so provide routes to biomimetic analogues that minimise concentration quenching. Here, we report the facile assembly of free-standing layered crystalline films of zinc(II) meso-tetrakis(4carboxyphenyl)porphyrin nanostructures that exhibit significant photocurrents in situ at an electrified liquid | liquid interface. This methodology does not require acidic conditions, specialised amphiphilic porphyrins, or the use of additives or external stimuli. The assembly process is driven by an interplay between the hydrophobicity gradient at an immiscible aqueous | organic interface and optimised hydrogen bonding in the formed nanostructure.Highly-ordered interfacial nanostructures may provide a new paradigm for realisation of light-harvesting antennae in artificial photosynthetic technologies.Photosynthetic organisms universally exploit antenna systems to capture high energy photons and funnel this excitation energy in the form of excitons toward a coupled reaction centre. 1,2 There, the excitation energy is transformed into chemical potential in the form of a charge separated state. 1,2 Self-assembled molecular antennae consisting of multi-layers of chromophores, such as porphyrins, can potentially function as high efficiency light harvesters due to their exceptionally high molar absorption coefficients (10 5 cm -1 ⋅M -1 ). 3 However, to mimic the evolved nanomachinery in photoautotrophs and avoid "concentration quenching" of the excited state at disordered trap sites, 3 the supramolecular packing of the individual chromophores within the antenna nanostructure must be precisely controlled and show longrange molecular order.

soft-photoconversion_using_floating_self-assembled_crystalline_films_of_porphyrin_nanostructures

<!>Results and discussion<!>This pH dependency of Por-IN formation indicates that cooperative H-bonding is prerequisite<!>Adsorption isotherms indicate that assembly is driven by hydrogen-bonding and π-π<!>Enhancing photoconversion by an order-of-magnitude at electrified aqueous<!>Conclusions<!>Experimental Methods

<p>Molecular self-assemblies at "soft" liquid | air or immiscible liquid | liquid interfaces can exhibit the required macroscale long-range order. 4 These soft interfaces are exceptionally smooth, and have no inherent defects leading to an unrivalled macroscale uniformity in molecule-interface interactions. 5,6 By contrast, the grain boundaries, step defects and edge sites always present at solid | liquid interfaces can impede diffusion of adsorbed molecules, trapping them in local energy minima as molecules stick to defect sites. 7,8 The uniform templating of adsorbed molecules at liquid | air or immiscible liquid | liquid interfaces has been exploited to create a variety of ordered porphyrin nanostructures, 9-14 yet a facile and robust route to create films of layered crystalline porphyrin nanostructures has not yet been reported to the best of our knowledge.</p><p>Here we describe light harvesting nanomaterials made using readily available, hydrophilic symmetrically substituted porphyrins that work efficiently in mild pH conditions. Thus, we avoid acidic conditions (that would lead to the expulsion of the central metal ion), 15 as well as the use of synthetically challenging and relatively expensive amphiphilic porphyrin molecules, 16 and more complicated routes using additives (e.g., divalent cations or surfactants) [17][18][19] or external stimuli (e.g., electric fields). [20][21][22] The key design criteria for the porphyrin molecule is that it must possess a meso-substituent capable of hydrogen bonding, and at pH = pKa conditions either the protonated or deprotonated form of the porphyrin (each present in a 1:1 ratio) must have a molecular charge of zero and adsorb at the liquid | liquid interface.</p><p>We chose zinc(II) meso-tetrakis(4-carboxyphenyl)porphyrin (ZnTPPc) as a prototypical model system (pKa COOH = 5.8) 23 to demonstrate this new means of self-assembly at an immiscible aqueous | organic interface. Simply contacting aqueous ZnTPPc solutions prepared in citrate buffer at pH 5.8 with a neat, immiscible organic phase of α,α,α,trifluorotoluene (TFT) lead to the immediate formation of free-standing films of porphyrin nanostructures. We rationalise our findings using atomistic computer simulations of the hydrophobicity gradient at the immiscible aqueous | organic interface that show that the carboxylic acid-carboxylate hydrogen bonding in the formed nanostructure is maximised under the mild pH = pKa COOH conditions. The presence of multi-layers with strong visible light absorption (due to the extended π-π conjugated electronic structure in porphyrin assemblies), and the crystalline macroscale long-range molecular order in the porphyrin nanostructure, suggested these films as ideal light-harvesting antennae in artificial photosynthetic technologies. To confirm this, we employed these nanostructures in a photoinduced interfacial electron transfer (PIET) arrangement between electron donor and acceptor species confined to the organic and aqueous phases, respectively. The energy absorbed by the nanostructure was used to create a highly energetic excited state that mediated electron transfer between the donor and acceptor molecules. In situ photocurrent transient measurements at a controllably electrified and LED illuminated liquid | liquid interface were as high as 20 µA•cm -2 , one order of magnitude higher than in previous studies. 24 Our findings clearly demonstrate that the ordered three-dimensional spatial arrangement of the individual porphyrin molecules in the nanostructure diminishes concentration quenching.</p><!><p>Triggering the formation of interfacial nanostructures. The selective formation of ZnTPPc nanostructures at the interface between water and TFT was observed upon contacting the ZnTPPc aqueous solution (at pH 5.8) with neat TFT. A yellow/green colour was observed at the water | TFT interface within minutes, easily distinguishable from the purple colour of the bulk ZnTPPc aqueous solution, and associated with the formation of porphyrin interfacial nanostructures (Por-INs) (Fig. 1a,b and Supplementary Fig. 1). Self-assembly was observed in a very narrow pH range around the pKa of the carboxylic substituents (pKa COOH = 5.8). 23 At this pH, ZnTPPc in solution exists in a ratio of 1:1 between the fully protonated (H 4 [ZnTPPc]) and fully deprotonated ([ZnTPPc] 4-) species. Molecular dynamics (MD) computer simulations indicate that the hydrophobic, neutral H 4 [ZnTPPc] species accumulates either at the interface, or even on the oil side of the interface (Fig. 1c,d and Supplementary Fig. 2), driven by the hydrophobicity gradient at the immiscible liquid | liquid interface. This interfacial layer acts as a template structure for the hydrophilic, anionic</p><p>[ZnTPPc] 4species to adsorb via carboxylic acid-carboxylate hydrogen-bonding and π-π interactions (Fig. 1c). In this manner, a highly crystalline film of ZnTPPc nanostructures, stabilised by hydrogen bonding and π-π interactions, builds up layer-by-layer at the interface.</p><p>Akin to the clathrate crystals of ZnTPPc developed by Goldberg and coworkers, [25][26][27][28][29] the strength of each individual hydrogen bond or π-π interaction may be insubstantial, but the cooperative effect allows the net enthalpies of these multivalent interactions to cumulatively rival the strength of a covalent bond and stabilise the Por-IN. Ex situ scanning electron microscopy (SEM) images revealed the thickness of the floating film of ZnPor-INs to be approximately 135 (± 5) nm (Fig. 1e). Control experiments and MD simulations demonstrated that the ZnTPPc and free-base H 2 TPPc molecules are kinetically stable in solution at pH 5.8, and do not undergo spontaneous bulk aggregation in the concentration range studied (Supplementary Fig. 3). This indicates that the Por-INs form only by self-assembly in situ at the water | TFT interface. Furthermore, to achieve selective Por-IN formation, the pH of the aqueous buffer solution must be controlled across a narrow pH range 5.1 ≤ pH ≤ 6.1 (Fig. 1f). More alkaline conditions inhibit formation of the Por-INs because of electrostatic repulsion between tetraanionic porphyrins. MD simulations show that the [ZnTPPc] 4species prefer to sit in water at a distance of ~3-5 Å above the interface due to their hydrophilic nature (Fig. 1d). Meanwhile, at more acidic conditions, spectroscopic measurements reveal simultaneous formation of a host of randomly structured aggregates in the bulk aqueous phase (Supplementary Fig. 4).</p><!><p>for self-assembly of an ordered film. Por-IN formation is not restricted to a single immiscible biphasic system, with control experiments showing that ZnPor-INs also form selectively at immiscible interfaces between water and 1,2-dichloroethane (DCE, Supplementary Fig. 5).</p><p>TEM and XRD analysis confirm the crystalline and layered nature of the porphyrin interfacial nanostructures. Ex situ transmission electron microscopy (TEM) images with corresponding selected area electron diffraction (SAED) analysis, and X-ray diffraction (XRD) patterns from ZnPor-INs were acquired after immobilization of the film on an amorphous hydrophilized glass substrate (Fig 2a-d). The diffraction pattern for ZnPor-INs (Fig. 2d) bears a striking resemblance to NAFS (nanofilm on a solid surface) structural models for liquid phase interfacial synthesis of highly ordered molecular nanosheets. 30 SAED and XRD estimate a 0.55-0.57 nm interplanar spacing between {110} planes, with a ~1.1 nm periodicity measured from HRTEM lattice fringes (Fig. 2a). The intense 220 reflection is consistent with a lack of layer-on-layer stacking order (due to non-registered stacking), but a highly crystalline nanosheet material is formed. The (200) reflections indicate strong axial coordination that is a distinct feature of layering along the c-axis normal to the support.</p><p>Clearly discernible (hk0) reflections are consistent with a tetragonal unit cell with preferred growth orientation along the plane of the liquid | liquid interface (Fig. 2b,c).</p><!><p>interactions between the tightly spaced porphyrins in the nanostructure. The interfacial concentration of porphyrin adsorbed to form the Por-INs was measured at equilibrium as a function of the solution concentration of ZnTPPc and H 2 TPPc, respectively. Over the concentration range studied (0.5-100 μM), ZnTPPc adsorption followed a Brunauer-Emmet-Teller (BET) isotherm behaviour, whereas H 2 TPPc adsorption followed a linear isotherm behaviour (Fig. 3). Using the BET model for liquid phase adsorption reported by Ebadi et al., 31 the isotherm obtained for ZnTPPc adsorption was fit to Eqn. 1: (assuming the area of a single ZnTPPc molecule as 3.14 nm 2 , calculated from its molecular surface). Experimentally the value obtained from the ZnTPPc absorption isotherm is an order of magnitude greater at 4.76 x 10 15 molecules⋅cm -2 . Therefore, the experimental surface coverage suggests that the monolayer surface concentration ( ) corresponds to a layer of interdigitated molecules. symmetry typical of single H 2 TPPc molecules. In solution, one main peak centred at 422 nm was deconvoluted from the experimental spectrum of ZnTPPc (Fig. 4a). This peak represents the B(0,0) transition, indicating the monomeric nature of ZnTPPc under these experimental conditions. The latter is consistent with the adherence of ZnTPPc in solution to the Beer-Lambert law across an extensive concentration range (Supplementary Fig. 3), and with previous studies. 33 Best fits in all other instances required multiple peaks, each related to different vibronic transitions, revealing the co-existence of monomers and dimers in solution Since ZnTPPc self-assembles in mild pH conditions, expulsion of Zn 2+ was avoided, as confirmed by analysis of the effect of nanostructure formation on the vibrational modes of the ZnPor-and H 2 Por-INs by ex situ Raman spectroscopy (Fig. 4e). Prominent differences between the two are entirely consistent with previous comparisons of metallo-and free-base 4-carboxyphenyl-substituted porphyrin Raman spectra (see Supplementary table 2). 34,35 The retention of Zn 2+ increases the inter-system crossing (ISC) rate constant, k ISC , due to the heavy atom effect, increasing the probability of the forbidden S 1 → T 1 transition. From the T 1 state, relaxation may occur via phosphorescence or charge transfer. The long-lived (up to millisecond) excited triplet state lifetimes provides sufficient time for the excited state to efficiently interact with ground state quencher molecules, 36 and is exploited herein to achieve photoconversion at a controllably electrified and LED illuminated liquid | liquid interface as described below. μA⋅cm -2 at 0 V and 20 μA⋅cm -2 when the aqueous phase was polarised 0.5 V with respect to the organic phase. The magnitude of the photocurrents increase substantially at 0.5 V due to the increase in the rate of interfacial electron transfer and the suppression of shunt resistances associated with the competitive transfer of cationic by-products (decamethylferrocenium cations) across the interface. The photocurrents generated with the ZnPor-INs of ~20 μA⋅cm -2 are an order of magnitude greater than those under similar conditions (< 1 μA⋅cm -2 ) with</p><!><p>ZnTPPc molecules adsorbed at the interface. 24,37 This step-change in photoconversion efficiency demonstrates the feasibility of using these highly ordered molecular assemblies in photosynthetic technologies.</p><!><p>The defect-free nature of the water | organic interface provides an ideal platform to selfassemble interfacial nanostructures with unique structural arrangements. In this Article, we report the self-assembly of free-floating interfacial nanostructures of zinc meso-tetrakis(4carboxyphenyl)porphyrin. The nanostructures are stabilised by cooperative hydrogen bonding and, due to the templating interaction of the interface with adsorbed porphyrin molecules, possess a crystalline, layered structure. This approach uniquely harnesses the difference in hydrophobicity between the neutral protonated and tetra-anionic non-protonated versions of the porphyrin at pKa conditions, combined with the introduction of a hydrophobicity gradient to trigger interfacial self-assembly. We open a new avenue to porphyrin nanostructure formation under mild experimental conditions without the need for acidic pH, designer amphiphilic porphyrin molecules, aggregation-inducing additives or external triggers. The macroscale long-range molecular order and diminished concentration quenching of these photoactive nanostructures facilitates their application in next generation photocatalytic, photovoltaic, opto-electronic and sensor devices. In this regard, the feasibility of using such nanostructures for light collection and harvesting was demonstrated in situ by measuring photocurrents associated with photoinduced interfacial electron transfer across the water | TFT interface, with an order-of-magnitude increase in photoconversion efficiency achieved in comparison to previous studies. The ability to isolate the photoactive film at the liquidliquid interface provides an elegantly simple system for future exploration of photo-induced electron transfer at electrified soft interfaces to gain the fundamental insights necessary to realise a new approach to solar energy conversion entirely based on a self-assembled system.</p><!><p>Chemicals. All reagents were used as received without further purification. Meso-tetrakis(4carboxyphenyl)porphyrin (H 2 TPPc, ≥98%) and its zinc(II) derivative (ZnTPPc, ≥98%) were obtained from Porphychem. Lithium hydroxide (LiOH, ≥98%), citric acid (H 3 Cit, ≥99.5%), decamethylferrocene (97%), and 1,2-dichloroethane (DCE, ≥99.0%) were purchased from Sigma-Aldrich, and α,α,α-trifluorotoluene (TFT, ≥99%) from Acros Organics. All aqueous solutions were prepared using Milli-Q® deionized water (18.2 MΩ). Aqueous solutions of ZnTPPc were prepared by directly dissolving the solid in the lithium citrate buffer preadjusted to the desired pH, followed by sonication of the solution for three minutes. Initially, H 2 TPPc was insoluble in the buffer at neutral pH. Therefore, the solid was dissolved first in LiOH and the pH subsequently adjusted with H 3 Cit. The ionic strength of each lithium citrate buffer solution containing either porphyrin was maintained at 10 (±2) mM.</p>

ChemRxiv

Development of CBAP-BPyne, a probe for \xce\xb3-secretase and presenilinase

\xce\xb3-Secretase undergoes endoproteolysis of its catalytic subunit, presenilin (PS), to form PS N-terminal and C-terminal fragments (PS1-NTF/CTF), which generate the active site. PS endoproteolysis, catalyzed by presenilinase (PSase), remains poorly understood and requires novel chemical approaches for its mechanistic study. CBAP is a dual inhibitor that suppresses both \xce\xb3-secretase and PSase activities. To probe \xce\xb3-secretase and PSase activity in cells, we have synthesized the clickable photoaffinity probe CBAP-BPyne. We found that CBAP-BPyne specifically labels PS1-NTF and signal peptide peptidase (SPP). CBAP-BPyne is a valuable tool to directly study the mechanism of endoproteolysis.

development_of_cbap-bpyne,_a_probe_for_\xce\xb3-secretase_and_presenilinase

<p>γ-Secretase is an aspartyl protease that belongs to the I-CLiPs family (intramembrane-cleaving proteases), a class of membrane-embedded enzymes that performs transmembrane (TM) hydrolysis on its substrates.1 γ-Secretase cleaves a wide array of type-1 TM substrates that have undergone ectodomain shedding. Some important γ-secretase substrates include amyloid precursor protein (APP), Notch, and E-cadherin. γ-Secretase plays a pivotal role in Alzheimer's disease (AD) and cancer and is an important target for prospective drug development.2,3</p><p>γ-Secretase is composed of at least four subunits: PS, nicastrin, Aph-1, and Pen-2.4 PS is the catalytic subunit of γ-secretase.5-8 The assembly, stabilization, trafficking, and maturation of the γ-secretase complex are tightly controlled and well regulated. The final step of γ-secretase activation occurs via Pen-2-mediated endoproteolysis of PS.9-11 Specifically, PS is translated as a single polypeptide chain and then, upon Pen-2 insertion into the complex, processed into two fragments, PS1-NTF and PS1-CTF. The two fragments of PS form a stable heterodimer, with each fragment contributing an aspartate residue to generate the active site of γ-secretase (Figure 1).</p><p>The enzyme responsible for the endoproteolytic cleavage of PS is termed PSase. Current evidence suggests that PSase is actually PS itself, and endoproteolysis is an autocatalytic cleavage event. This is illustrated by the following observations: First, mutation of PS's catalytic aspartate residues not only blocks γ-secretase activity, but also PSase activity.5 Second, pepstatin A, an aspartyl protease inhibitor, suppresses PSase activity, further suggesting that PSase is an aspartyl protease.12 However, the coexpression of WT PS1 with PS1 D257A (a γ-secretase and PSase deficient mutant) does not restore endoproteolysis of the mutant, indicating that endoproteolysis occurs in cis and is an autocatalytic event.13 Finally, an in vitro reconstitution study showed that bimolecular interaction of PS1 and Pen-2 is necessary and sufficient for PS1 endoproteolysis.8 Collectively, these studies strongly indicate that PS has PSase activity. Notwithstanding findings that PS possesses γ-secretase and PSase activities, it has been a formidable challenge to characterize both activities and understand their differences due to their complex interdependence. While many γ-secretase active site-based inhibitors exist to directly probe γ-secretase, no successful PSase-directed probes exist to date. CBAP (Figure 2A) is a γ-secretase inhibitor that also causes a "pharmacological knock-down" of PS1 NTF/CTF with a concomitant accumulation of full-length PS1 (PS1-FL) in the cell.14 However, the mechanism of action of CBAP in γ-secretase and PSase remains to be investigated. We have synthesized CBAP-BPyne, a clickable, photoreactive form of CBAP, as a tool to understand the mechanism of PSase (Figure 2A).</p><p>The CBAP intermediate TBS-protected alcohol (4) was synthesized by coupling amino benzodiazepinone 3 to carboxylic acid 1 as previously reported.14 To synthesize CBAP-BPyne, we initially investigated the selective removal of the NHBoc group from 4, but all conditions examined resulted in poor product formation where removal of the silyl and Boc protecting groups occurred at competitive rates. It was determined that selective Boc group removal or one-pot global deprotection strategies were not viable to produce the CBAP-BPyne in sufficient yields and purity. CBAP-BPyne was ultimately synthesized by removing the silyl protecting group in 4 with TBAF to yield CBAP followed by a rapid deprotection of the NHBoc group at 0 °C in dilute TFA to produce the fully deprotected scaffold. The crude amino alcohol was then immediately coupled with the NHS ester of propargyl benzophenone 2 to afford CBAP-BPyne.15‡</p><p>CBAP-BPyne contains a photophore for photoaffinity labeling and an alkyne for click chemistry (copper catalyzed azide-alkyne cycloaddition). This clickable probe approach facilitates the design of functional probes that can selectively label and detect proteins in complex cellular systems with minimal modification to the original compound.16-19 First, using our in vitro γ-secretase activity assay with recombinant APP and Notch1 substrates,20-24 we determined that both CBAP and CBAP-BPyne are equipotent γ-secretase inhibitors. Specifically, both compounds potently inhibit γ-secretase activity for both the production of Aβ40 and Notch1-NICD (Figure 3A). Next, we examined CBAP and CBAP-BPyne's cellular activity in inhibiting PS1 processing. HeLa cells were treated for four days in a 12-well plate with CBAP, CBAP-BPyne, L685,458 (an active site-directed γ-secretase inhibitor), or vehicle control at concentrations of 1, 3, or 10 μM. Cells were lysed with RIPA buffer (50 mM Tris base pH 8.0, 150 mM NaCl, 0.1% SDS, 1% NP-40, 0.5% deoxycholate) and 15 μg of cell lysate was separated on a 12% Bis-Tris gel in MES buffer. Proteins were transferred to Immobilon-FL PVDF membrane, probed with anti PS1-NTF antibody, and imaged on Odyssey (LI-COR Biosciences). CBAP and CBAP-BPyne caused a much greater accumulation of PS1-FL at the expense of PS1-NTF/CTF than did L685,458 (Figure 3B for PS1-NTF and data not shown for PS1-CTF). This indicates that CBAP and CBAP-BPyne are capable of inhibiting PSase activity, while L685,458 has nearly no effect on PSase activity under these conditions. CBAP is a more potent PSase inhibitor than CBAP-BPyne as seen by the nearly complete depletion of PS1-NTF in CBAP treated cells compared to the incomplete PS1-NTF depletion in CBAP-BPyne treated cells (Figure 3B). This effect is not due to a difference in the ability of the compounds to permeate the cell membrane because CBAP and CBAP-BPyne were equipotent in a cell-based γ-secretase activity assay (Aβ42 IC50 = 28 nM and 20 nM, for CBAP and CBAP-BPyne, respectively, in CHO-APP cells).</p><p>Finally, we determined that CBAP-BPyne is a functional probe, as it specifically labels PS1-NTF (Figure 4A). Briefly, 600 μg of HeLa cell membrane, diluted with PBS to a volume of 500 μL in a 12-well plate, was incubated with either 2 μM CBAP or vehicle control for 15 min at 37 °C. 20 nM CBAP-BPyne was added for 1 hour at 37 °C followed by UV irradiation (350 nm) for 45 min to promote benzophenone-protein crosslinking. Membrane was pelleted by centrifugation at 100,000 ×g for 30 min and resuspended in PBS using Qiagen TissueLyser. Click chemistry reagents [1 mM tris(2-carboxyethyl)phosphine, 1 mM CuSO4, 0.1 mM tris-(benzyltriazolylmethyl)amine, and 0.1 mM biotin azide in 5% t-butyl alcohol with 1% DMSO] were added and the mixture was rotated for 1 hour at room temperature. Membranes were pelleted by centrifugation at 100,000 ×g for 30 min, resuspended in 500 μL PBS and solubilized with the addition of RIPA buffer. Samples were centrifuged at 13,400 ×g and supernatant was added to Pierce Streptavidin Plus UltraLink Resin and rotated overnight at 4 °C. Proteins were eluted with 2 mM biotin in SDS sample buffer at 70 °C for 10 min, separated on a 12% Bis-Tris gel or a 4-20% TGX gel, transferred to Immobilon-FL PVDF, probed with the relevant antibody, and visualized on Odyssey (LI-COR Biosciences). CBAP-BPyne was also found to label SPP, a protein structurally similar to PS (Figure 4A). CBAP-BPyne does not label PS1-CTF or any of the other three subunits of γ-secretase (data not shown). Photoaffinity labeling studies followed by click chemistry with TAMRA-azide confirmed the specific labeling of PS1-NTF and SPP, and showed that CBAP-BPyne binds additional proteins, although PS1-NTF is the primary target (Figure 4B). The additional unidentified proteins that are specifically labeled (denoted with a star) may play a role in endoproteolysis and will be studied further for their identity and function. Whether bands that migrated in the range of high molecular weight represent aggregated PS1-NTF, SPP or novel proteins also remains to be investigated.</p><p>CBAP-BPyne is the first clickable, photoreactive probe that inhibits both γ-secretase and PSase activities. Of note, based on current clinical investigation of non-selective γ-secretase inhibition (i.e. the case of semagacestat), PSase may not be a viable drug target for the treatment of AD since PSase inhibition also blocks γ-secretase activity, leading to toxicity.25 However, PSase could serve as a target for cancer therapy. Furthermore, this probe can be used to investigate PSase and γ-secretase activation, which appear to play a role in disease states, as evidenced by reports that some familial AD PS1 mutations affect PSase activity.26-28 CBAP-BPyne provides a novel means to investigate the mechanism of PSase as it has the capacity not only to bind and inhibit γ-secretase, but also to inhibit the endoproteolysis of PS1-FL, a novel function not observed in other γ-secretase probes. CBAP-BPyne may aid in the identification and characterization of PSase, revealing the mechanism of γ-secretase activation and uncovering PSase as a potential target in cancer therapy.</p>

PubMed Author Manuscript

Simple Method for De Novo Structural Determination of Underivatised Glucose Oligosaccharides

Carbohydrates have various functions in biological systems. However, the structural analysis of carbohydrates remains challenging. Most of the commonly used methods involve derivatization of carbohydrates or can only identify part of the structure. Here, we report a de novo method for completely structural identification of underivatised oligosaccharides. This method, which can provide assignments of linkages, anomeric configurations, and branch locations, entails low-energy collision-induced dissociation (CID) of sodium ion adducts that enable the cleavage of selective chemical bonds, a logical procedure to identify structurally decisive fragment ions for subsequent CID, and the specially prepared disaccharide CID spectrum databases. This method was first applied to determine the structures of four underivatised glucose oligosaccharides. Then, high-performance liquid chromatography and a mass spectrometer with a built-in logical procedure were established to demonstrate the capability of the in situ CID spectrum measurement and structural determination of the oligosaccharides in chromatogram. This consolidation provides a simple, rapid, sensitive method for the structural determination of glucose oligosaccharides, and applications to oligosaccharides containing hexoses other than glucose can be made provided the corresponding disaccharide databases are available.

simple_method_for_de_novo_structural_determination_of_underivatised_glucose_oligosaccharides

<!>The approach of de novo structural determination<!>Dissociation mechanism. Dehydration and cross-ring dissociation mainly occur on the reducing side of sodiated oligosaccharides.<!>Procedures of structural determination.<!>Results and Discussion<!>β-Glc-(1→3)-β-Glc-(1→4)-Glc.<!>HPLC-ESI-MS n .<!>ESI-MS n .

<p>Carbohydrates (or saccharides) are the most abundant biological compounds on Earth 1 . They play crucial roles in many biological processes, such as immune responses, molecular recognition, signalling, and cellular communication 2 . To elucidate the chemical and biological properties of carbohydrates, relevant structure-function relationships must be studied. However, the identification of carbohydrate structure remains difficult 3 because of the presence of a high number of carbohydrate isomers in a given chemical formula. For example, an oligosaccharide containing six hexoses has more than 10 12 isomers 4 , and the differentiation of such a high number of isomers by using a single simple analytical method is difficult.</p><p>The most frequently used carbohydrate analysis techniques are liquid chromatography (LC) 5 , capillary electrophoresis (CE) 6,7 , nuclear magnetic resonance spectroscopy (NMR) 8 , and mass spectrometry (MS) 9 . The structures of carbohydrates cannot be determined directly by LC and CE. They are determined by the comparison of chromatography to that of databases which the structure of carbohydrates are characterized by other methods, e.g., NMR and MS. NMR and MS are widely applied in the structural analysis of molecules. However, MS requires much less sample than that of NMR, that makes MS particularly useful for the structural analysis of limited amount of available samples, e.g., the carbohydrates extracted from biological systems. Although MS is widely used in the protein structural determination, application of MS to the identification of carbohydrate structures remain challenging due to the low ionisation efficiency of carbohydrates in a mass spectrometer [10][11][12] , the large amount of carbohydrate isomers, and the similarity of mass spectra between isomers.</p><p>Collision-induced dissociation (CID) tandem MS is one of the major methods of determining the structure of carbohydrates 9,13,14 . Several oligosaccharide fragment databases [15][16][17] and empirical fragmentation patterns of carbohydrate cations [18][19][20] , anions [21][22][23][24][25][26] , and derivatised carbohydrates 13,[27][28][29][30][31] have been used for structural 1 Institute of Atomic and Molecular Sciences, Academia Sinica, P. O. Box 23-166, Taipei, 10617, Taiwan. 2 Department of Chemistry, National Taiwan Normal University, Taipei, 11677, Taiwan. 3 Department of Chemistry, National Tsing Hua University, Hsinchu, 30013, Taiwan. Correspondence and requests for materials should be addressed to C.-K.N. (email: ckni@po.iams.sinica.edu.tw) determination. However, these patterns can determine only a part of the structure or are mainly applied to the well-characterised disaccharides and oligosaccharides in databases.</p><p>Several advancements in the de novo structural identification of monosaccharides and oligosaccharides have been demonstrated recently. Nagy et al. reported a fixed-ligand kinetic method for the determination of monosaccharide absolute configuration 32 . However, this method was not used for the determination of the linkage positions, anomeric configurations, sequences, and branch locations of oligosaccharides. Konda et al. reported that the CID of anion m/z 163 exhibited distinct fragmentation fingerprints corresponding to linkage positions. They applied the CID of anion m/z 163 to the linkage determination of 18 O-labelled linear oligosaccharides 33 . Bendiak et al. demonstrated that anion m/z 221 can be used to identify the stereochemistry and anomeric configuration of hexose in oligosaccharides [34][35][36] . This method has the following limitations: The reducing end must be derivatised, resulting in the structure of two hexoses on the reducing side cannot be determined, anion intensities are usually low that it may take several hours to obtain a mass spectrum with a good signal-to-noise ratio, a complicated mass spectrometer is required, and this method is currently used for only linear oligosaccharides.</p><p>Recently we have proposed a new de novo method for determining the entire structure of underivatised oligosaccharides through CID tandem MS of sodium ion adducts 37 . In this study, the structural determination of glucose trisaccharides and tetrasaccharides was demonstrated. This method can be extended to larger oligosaccharides and oligosaccharides containing hexoses other than glucose.</p><!><p>The new method involves the sequential low-energy resonance excitation CID experiments of sodiated oligosaccharides in a typical ion trap mass spectrometer. However, the sequence of the tandem CID experiment is specially designed according to the dissociation mechanism we found recently 38,39 , and the measured CID spectra are compared with our specially prepared database. In the following sections, we first describe the dissociation mechanism, followed by a description of the database and the scheme of the procedure to determine the structures of oligosaccharides.</p><!><p>Recent high-level quantum chemistry calculations 12,38,39 have indicated that glycosidic bond cleavage, cross-ring dissociation and dehydration reactions on the reducing sugar are dissociation channels with low barrier heights for sodiated carbohydrates. By contrast, cross-ring dissociation and dehydration reactions on the nonreducing sugar are dissociation channels with high barrier heights. Coincidently, the sodiation energy of carbohydrates is near or just slightly higher than the dissociation barrier heights of cross-ring dissociation and dehydration reactions on the reducing sugar.</p><p>A mass spectrometer with low-energy CID and resonance excitation is used in this method. We selected sodium ion adducts because sodium ions are an efficient dissociation channel discriminator in CID because of their appropriate sodiation energy and the loose transition state property of desodiation. In the process of low-energy CID using resonance excitation in an ion trap, only the parent ions in an ion trap are excited by the dipolar frequency corresponding to the parent ion motion frequency. They accumulate internal energy from collisions with neutral gases. In each collision, only a small amount of translational energy is transferred to the vibrational energy of trapped ions. It requires many collisions for a trapped ion to obtain enough energy and undergo dissociation. Because the amount of energy transferred in each collision is not fixed, the internal energy distribution of parent ions becomes very broad after many collisions. When the accumulated internal energy is larger than the dissociation threshold, dissociation occurs. Dissociation mainly occurs on the channels with low barriers because the energy transfer is slow. However, if the internal energy is large, sodium cations are eliminated before the occurrence of reactions with large dissociation barrier heights 31 (e.g., cross-ring dissociation and dehydration reactions on the nonreducing sugar). The product ions are trapped in the ion trap, but they cannot be excited by the parent ion motion frequency. Most of the product ions do not have enough excess of internal energy to undergo consecutive cleavages, thus fragment ions of the secondary dissociation are minimized. The combination of low-energy CID, resonance excitation, and sodium adducts guarantees the occurrence of most cross-ring dissociation and dehydration reactions on the reducing sugar.</p><p>Linkages of the reducing sugar can be determined from the fragmentation patterns of dehydration and cross-ring dissociation. The fragmentation patterns of dehydration and cross-ring dissociation found in sodiated and lithiated disaccharides in previous studies [40][41][42][43] can be extended based on the dehydration mechanism 38 and the retro-Aldol reaction 38,40 . These fragmentation patterns can be used for linkage determination of the reducing sugar of oligosaccharides. For linear oligosaccharides, these patterns include 0,2 A n for 1→4 and 1→6 linkages; 0,3 A n (or 0,3 X 0 ) for 1→3 and 1→−6 linkages; 0,4 A n for a 1→6 linkage; and dehydration for 1→3, 1→4, and 1→6 linkages. When a branch is located on the reducing sugar, dehydration occurs for (1→6, 1→4), (1→6, 1→3), and (1→4, 1→3) linkages, and cross-ring dissociation on the reducing sugar follows the retro-Aldol reaction; that is, 0,2 A n (or 0,2 X 0 ) for (1→6, 1→4), (1→6, 1→2), and (1→4, 1→2) linkages and 0,3 A n (or 0,3 X 0 ) for (1→6, 1→3) and (1→3, 1→2) linkages. The details of the fragmentation patterns of trisaccharides are listed in Supplementary Table S8. In general, different linkages result in different fragmentation patterns, except for a linear oligosaccharide with a (1→4) linkage and a branched oligosaccharide with (1→6, 1→4) linkages. Oligosaccharides with these two linkages can be differentiated using subsequent CID spectra.</p><p>Disaccharides from the nonreducing side can be generated from the products of dehydration and cross-ring dissociation. Part of the structural determination process includes the dissociation of oligosaccharides into disaccharides by CID in a mass spectrometer; these disaccharides are then in situ fragmented into their corresponding fingerprint fragments. To simplify structural determination, a disaccharide is generated only from one side (the reducing or nonreducing side) of parent or fragment ions in each step of CID. Similar concept has been used in negative mode with derivatized oligosaccharide 36 . Derivatization of oligosaccharide on the reducing sugar is crucial in those studies 36 in order to ensure that the disaccharide fragment is produced only from the nonreducing side of the oligosaccharide during CID. However, the derivatization processes are tedious and time-consuming. Most important, the derivatization changes the structure of the disaccharide on the reducing side. Consequently, the linkage and the anomeric configuration of the disaccharide on the reducing side of oligosaccharide cannot be determined.</p><p>To simplify the structural determination process and determine the structure of entire oligosaccharide, we introduce a simple approach to obtain the desired disaccharide fragments from underivatized oligosaccharide unambiguously. We found that generating a disaccharide from only the nonreducing side is easy if dehydration or cross-ring dissociation products are selected as precursor ions. Because dehydration or cross-ring dissociation of sodiation oligosaccharides mainly occurs on the reducing side, the disaccharide generated from parent or fragment ions after dehydration or cross-ring dissociation is mainly the disaccharide from the nonreducing side.</p><p>Structures of disaccharides are determined according to rules. After the generation of a disaccharide from an oligosaccharide, the structure of the disaccharide can be determined according to the following rules. First, the fragmentation patterns of dehydration and cross-ring dissociation found in sodiated and lithiated disaccharides in previous studies [40][41][42][43] can be directly used in linkage determination. Second, our recent theoretical calculations revealed that dehydration is mainly related to the cis or trans configuration of the O1 and O0 atoms of the reducing sugar 38 . Therefore, the anomeric configurations of the reducing sugar in a disaccharide can be determined using the ratio of dehydration to any dissociation channel that is not related to dehydration. Third, the dissociation mechanism of glycosidic bond cleavage is analogous to that of dehydration; that is, it is related to the cis and trans configuration of the O1 and O0 atoms of the nonreducing sugar of a disaccharide. The anomeric configuration of the glycosidic bond in a disaccharide can be determined using the ratio of glycosidic bond cleavage to any dissociation channel that is not related to glycosidic bond cleavage.</p><p>For practical applications, the CID spectra of disaccharides with various linkages and anomeric configurations are measured in advance and prepared as a database. The structures of these disaccharides are determined according to the aforementioned rules. The structures of disaccharides produced from the dissociation of an oligosaccharide are then determined by comparing the measured CID spectra with the database.</p><p>Specially prepared database of disaccharide CID spectra. The α and β anomeric configurations of the sugar at the reducing end typically coexist in a solution. The ratio of these two anomers depends on the solvent, pH, and temperature of the solution. Therefore, the CID spectrum of a given disaccharide depends on sample preparation if these two configurations are not separated before the CID spectrum measurement. Because these two configurations of a given disaccharide were not separated before the CID spectrum measurement in previous studies [40][41][42][43] , those CID spectra can only be used for determining linkage positions, not anomeric configurations.</p><p>We constructed our special disaccharide database by separating the two anomeric configurations prior to CID spectrum measurement. Figure 1(A) presents the total ion count (TIC) chromatogram of maltose obtained using the online coupling of high-performance liquid chromatography (HPLC) with electrospray ionisation (ESI) MS (ESI-MS). Two configurations of maltose, namely α-Glc-(1→4)-α-Glc and α-Glc-(1→4)-β-Glc, coexist in a solution. Because mutarotation between configurations is typically slower than separation time in HPLC 44,45 , the separation of these two configurations can be clearly observed on chromatograms. The CID spectra of the two peaks in the chromatogram are presented in Fig. 1(B) and (C), respectively.</p><p>A major difference between Fig. 1(B) and (C) is the relative intensity of ion m/z 347, representing the dehydration (elimination of H 2 O) from the reducing end of maltose. This intensity difference of ion m/z 347 can be explained by the dissociation mechanism discovered from high level quantum chemistry calculations 38 . Water elimination mainly occurs through the transfer of the H atom from the O2 atom of the reducing glucose to the O1 atom of the same glucose, followed by C1-O1 bond cleavage. The O1 and O2 of the reducing glucose of α-Glc-(1→4)-α-Glc are in a cis configuration, in contrast to the trans configuration of those of α-Glc-(1→4)-β-Glc. Calculations reveal that the water elimination barrier of the cis configuration is substantially smaller than that of the trans configuration 38 . Therefore, the CID spectrum with a high intensity of ion m/z 347 [Fig. 1(C)], which represents a large branching ratio of the water elimination, is assigned to α-Glc-(1→4)-α-Glc. Similar assignments can be made for the spectra of other disaccharides, except for kojibiose and sophorose, in which the dehydration reaction is a minor channel and mechanism is different.</p><!><p>A schematic of the procedure for the structural determination of oligosaccharides is presented in Fig. 2(a); linear pentasaccharides are used as an example. The fragments in Fig. 2(a) are generated according to the aforementioned dissociation mechanism; that is, the cleavage of the glycosidic bond to generate B, C, Y, and Z ions can occur between any two adjoining monosaccharides, but dehydration and cross-ring dissociation only takes place on the reducing sugar. Not all possible fragments are plotted in Fig. 2(a); only structural decisive fragments are shown. The structural determination procedure is as follows. (1) The linkage of the reducing sugar is determined by the aforementioned fragmentation patterns in step 1 (MS2).</p><p>(2) A disaccharide at the nonreducing side of an oligosaccharide comprising monosaccharides labelled with 4 and 5 is generated from the CID of dehydration or cross-ring dissociation products in step 2. The linkage and anomeric configuration of the glycosidic bond between monosaccharides 4 and 5 and the anomeric configuration of the reducing sugar (monosaccharide 4) are determined in step 3 by matching with the database. Interestingly, the anomeric configuration of the reducing sugar (monosaccharide 4) of a disaccharide comprising monosaccharides 4 and 5 also represents the anomeric configuration of the glycosidic bond between monosaccharides 4 and 3, which can be determined separately from the disaccharide comprising monosaccharides 3 and 4 (in step 5). These two independent approaches provide a crosscheck of the anomeric configuration between monosaccharides 3 and 4. (3) The other disaccharides can be produced by CID from the nonreducing end of various fragment ions, as shown in steps 4 and 6. The linkage and anomeric configurations of corresponding disaccharides can be determined in steps 5 and 7, respectively. (4) Disaccharides from both the reducing (comprising monosaccharides 1 and 2) and nonreducing ends (comprising monosaccharides 4 and 5) of an oligosaccharide are produced in step 1. The CID spectrum of these ions is the sum of these two disaccharide CID spectra. If the structure of one disaccharide is determined, the structure of the other disaccharide can be determined using the CID spectra obtained in step 8 after subtracting the CID spectrum of the disaccharide whose structure is determined. This provides additional information for crosscheck. (5) The structure of the entire oligosaccharide is then determined by the combination of structural information obtained individually from various disaccharides, as illustrated in Fig. 2(b). Most of the anomeric configurations can be determined using more than one approach. For example, the anomeric configuration of monosaccharide unit 4 can be determined by steps 3, 5, and 8. Multiple approaches increase the reliability of this method.</p><p>A similar procedure can be employed for the structural determination of branched oligosaccharides. The procedure becomes complicated when both linear and branched oligosaccharides are considered. However, once the procedure is established, the sequence of CID spectrum measurement and the structural determination are straightforward. Figure 3 presents a schematic of the procedure for the structural determination of branched and linear trisaccharides. The procedure includes all possible disaccharides that can be generated by CID and the necessary CID spectrum measurement for structural determination. The details of the fragmentation patterns are listed in Supplementary Table S8. The applications of the scheme for structural determination are demonstrated in the next section. A similar scheme for oligosaccharides containing more than three monosaccharides can be developed using the same concept.</p><!><p>Application for the structural determination of oligosaccharides. Panose. The CID spectrum of sodiated panose is displayed in Fig. 4(a). The ions m/z 509 and 467 indicate that the carbohydrate is a linear trisaccharide with a 1→4 linkage or a branched trisaccharide with 1→4 and 1→6 linkages on the sugar of the reducing end according to the fragmentation patterns shown in Supplementary Table S8. Ion m/z 365 in the CID spectrum of 527→509→fragments [Fig. 4(d)] indicates that the carbohydrate is a linear trisaccharide according to the scheme in Fig. 3. The CID spectrum of the disaccharide from the nonreducing side, 527→509(B 3 )→365 (C 2 /B 3 )→fragments, is presented in Fig. 4(h). A comparison of this spectrum with that in Fig. 1 suggests that this disaccharide is α-Glc-(1→6)-α-Glc. Therefore, the trisaccharide is determined to be α-Glc-(1→6)-α-Glc-(1→4)-Glc. S8.</p><p>Dehydration is a minor dissociation channel of the disaccharide with a 1→6 linkage. If noise happens to appear near ion m/z 347, it may affect the identification of the anomeric configuration of the reducing sugar. Here, we crosschecked the anomeric configuration of the 1→6 linkage by using a different approach. The sodiated disaccharide ion m/z 365 produced from 527→365 could be the disaccharide on the reducing side (Y 2 ion) or that on the nonreducing side (C 2 ion). The CID spectrum of 527→365→fragments [Fig. 4(b)] is the sum of the spectra of these two disaccharides, weighted by the percentage of each disaccharide produced in CID. These two spectra include one spectrum from the disaccharide with a 1→6 linkage [i.e., Fig. 4(h)] and the spectrum from the disaccharide with a 1→4 linkage [i.e., one of Fig. 1(B,C,E An alternative approach for structural determination involves 527→467( 0,2 A 3 )→365(C 2 / 0,2 A 3 )→fragments. This approach shares the same first step of the aforementioned method, i.e., the ions m/z 509 and 467 in the CID of 527→fragments [Fig. 4(a)] indicate that the carbohydrate is a linear trisaccharide with a 1→4 linkage or a branched trisaccharide with 1→4 and 1→6 linkages on the reducing sugar according to the fragmentation patterns shown in Supplementary Table S8. The difference of this alternative approach is the use ion m/z 467 instead of m/z 509 from the CID of 527→fragments for subsequent CID. Ion m/z 347 in the CID spectrum of 527→467→fragments, [Fig. 4(c)], indicates that the carbohydrate is a linear trisaccharide with a 1→4 linkage at the reducing sugar according to the scheme in Fig. 3. A comparison of Fig. 1 with the disaccharide CID spectrum produced from 527→467( 0,2 A 3 )→365(C 2 / 0,2 A 3 )→fragments [Fig. 4(f)] suggests that the disaccharide on the non-reducing side is α-Glc-(1→6)-Glc. Subtraction of the CID spectrum in Fig. 4(f) from the CID spectrum in Fig. 4(b) yields the spectrum in Fig. 4(e), from which the carbohydrate can be unambiguously identified as α-Glc-(1→4)-α-Glc [Fig. 1(B)] or α-Glc-(1→4)-β-Glc [Fig. 1(C)]. Consequently, the structure of this trisaccharide can be determined as α-Glc-(1→6)-α-Glc-(1→4)-Glc. The spectrum matching by using calculations of similarity for these three different approaches are presented in Supplementary Table S4.</p><!><p>The structural determination procedure for this trisaccharide is similar to that for panose. Ions m/z 509 and 467 in the CID spectrum of ion m/z 527 [Fig. 5(a)] suggest that the carbohydrate is a linear trisaccharide with a 1→4 linkage or a branched trisaccharide with 1→4 and 1→6 linkages on the reducing sugar, and ion m/z 365 in the CID spectrum of 527→509→fragments [Fig. 5(d)] indicates that the trisaccharide is linear according to fragmentation patterns shown in Supplementary Table S8. Comparing Fig. 1 with the CID spectrum of the disaccharide at the nonreducing side, 527→509(B 3 )→365(C 2 /B 3 ) →fragments [Fig. 5(h)], suggests that the disaccharide is β-Glc-(1→3)-β-Glc. Therefore, the trisaccharide can be identified as β-Glc-(1→3)-β-Glc-(1→4)-Glc. Analogous to panose, alternative approaches can be used for this trisaccharide to crosscheck the structure. The alternative approaches include using (1) 527→467( 0,2 A 3 ) →fragments and 527→467( 0,2 A 3 )→365(C 2 / 0,2 A 3 )→fragments or (2) 527→365→fragments. The CID spectra for the alternative approaches are illustrated in Fig. 5(b,c,e and (f). Both alternative approaches obtain the same results. The spectrum matching by using calculations of similarity is presented in Supplementary Table S5.</p><p>Isopanose (a branched trisaccharide). Branched oligosaccharides have more than one non-anomeric carbons (C2, C3, C4, or C6) of a given monosaccharide connected to another sugar. Most of the present de novo structural determination methods are applied to linear oligosaccharides. Structural identification of branched oligosaccharides remains challenging. The CID spectrum of sodiated isopanose, α-Glc-(1→4)-[α-Glc-(1→6)]-Glc, a branched trisaccharide, is illustrated in Fig. 6(a). The ions m/z 509 and 467 indicate that the carbohydrate is a linear trisaccharide with a 1→4 linkage or a branched trisaccharide with 1→4 and 1→6 linkages, according to the fragmentation patterns shown in Supplementary Table S8. Ions m/z 365 produced from 527→509→fragments [Fig. 6(b)] or ions m/z 347 and 365 produced from 527→467→fragments [Fig. 6(c)] were not observed. According to the scheme in Fig. 3, this finding indicates that the carbohydrate is a branched trisaccharide.</p><p>The CID spectrum of 527→365→fragments [Fig. 6(d)] is the sum of the spectra of disaccharides with 1→6 and 1→4 linkages. Subtraction of the spectrum in Fig. 1(W) or (X) [β-Glc-(1→6)-Glc] from that in Fig. 6(d) yields a spectrum of large intensity of ion m/z 203 and near zero intensities of ions m/z 245, 275, and 305, which does not match with any spectrum of disaccharides with 1→4 linkage. By contrast, subtraction of the spectrum in Fig. 1(T) or (U) [α-Glc-(1→6)-Glc] from that in Fig. 6(d) yields a spectrum of near zero intensity for all ions, indicating that the disaccharide produced from 527→365 is mainly α-Glc-(1→6)-Glc and almost no disaccharide with 1→4 linkage is produced. At this moment, we can only determine the linkages of both branches and the anomeric configuration of one branch. The spectrum matching by using calculations of similarity is presented in Supplementary Table S6.</p><p>If the quantities of disaccharides with 1→6 and 1→4 linkages produced through CID of parent ions are not very different, the anomeric configurations of both disaccharides can be determined using the method similar to the structural determination procedure of panose. Unluckily, this does not occur in isopanose. However, there is no reason that all branched oligosaccharides break only the glycosidic bond of one branch without breaking that of the other branch. The determination of only one anomeric configuration in isopanose is not some kind of inherited problems of this method when sequencing branched sugars. Unfortunately, isopanose is the only branched glucose-trisaccharide commercially available at this moment. We do not have another branched glucose-trisaccharide available to test our method.</p><p>Cellotetraose. There are four types of tetrasaccharides, namely the linear tetrasaccharide (I), branched tetrasaccharide on the reducing sugar (II), branched tetrasaccharide on the nonreducing sugar (III), and tetrasaccharide with two branches on the reducing sugar (IV). The scheme used to differentiate these four types of tetrasaccharides is illustrated in Fig. 7.</p><p>Ions m/z 365 and 347 produced through the CID of ions m/z 689 [Fig. 6(e)] and ions m/z 527 and 509 produced through the CID of 689→629→fragments [Fig. 6(f)] suggest that the tetrasaccharide is linear according to the scheme illustrated in Fig. 7. The generation of ions m/z 671(B 4 ) and 629( 0,2 A 4 ) from the sodiated tetrasaccharide ion m/z 689 [Fig. 6(e)] indicates a 1→4 linkage between the first two monosaccharides on the reducing side. The CID spectrum of 689→629( 0,2 A 4 )→365 (C 2 / 0,2 A 4 )→fragments [Fig. 6(h)] suggests that the structure of the first two monosaccharides on the nonreducing side of the tetrasaccharide is β-Glc-(1→4)-β-Glc, and the CID spectrum of 689→629( 0,2 A 4 )→467(Y 3 / 0,2 A 4 )→365 →fragments [Fig. 6(g) and (i)] suggests that the structure of the two monosaccharides at the centre of the tetrasaccharide is β-Glc-(1→4)-β-Glc. A combination of these spectra can be used to identify the structure as β-Glc-(1→4)-β-Glc-(1→4)-β-Glc-(1→4)-Glc. The spectrum matching by using calculations of similarity is presented in Supplementary Table S7.</p><p>In situ CID spectrum measurement. Common commercially available mass spectrometers are equipped with sophisticated software to perform experiments automatically. However, if no appropriate guidance is established for the selection of daughter ions for tandem mass spectrum measurement, time and precious samples are wasted because many CID spectra do not provide the structural information necessary for structural determination. The situation becomes critical when the amount of sample is limited, which is common when carbohydrates are extracted from biological samples. By contrast, structurally decisive fragments can be identified according to the schemes shown in Fig. 3. The entire structural determination can be considerably simplified by measuring only the CID spectra of these fragments.</p><p>A mass spectrometer with built-in logical procedures was established for in situ CID spectrum measurement and structural determination. Figure 8(a) shows the TIC chromatogram of a trisaccharide, panose, obtained through the online coupling of HPLC with ESI-MS. While the oligosaccharide was passed through liquid chromatography, the mass spectrometer in situ performed all the necessary CID spectrum measurements. Figure 8(b-e) shows the CID spectra obtained during the appearance of a peak in the chromatogram. The results are the same as the CID spectra shown in Fig. 4.</p><p>Although the apparent duration of the peak in the chromatogram is less than 30 seconds, structural determination procedures are simple and sodiated ions are abundant, such that the entire CID spectrum measurement with a favourable signal-to-noise ratio can be performed three times within 30 seconds. The detection limit of our method for trisaccharide was estimated to be lower than 0.33 nmole from the amount of panose used in Fig. 8 (10 μl injection of 10 −4 M solution for three times of MS n spectrum measurements). The successful structural determination in this study indicates the high capability of this method for the in situ structural determination of oligosaccharides through chromatograms.</p><p>Comparison to other MS methods. Currently, the determination of linkages, anomeric configurations, and the branching location of oligosaccharides represents a major limitation in carbohydrate research. Several mass spectrum approaches have been developed to determine the structures of carbohydrates. A commonly used method is the mass spectra of permethylated carbohydrates 13 . This method requires the permethylation of carbohydrates prior to mass spectrum measurement. A fraction of sample may be lost due to the incomplete permethylation and during the extraction of the permethylated carbohydrates. This method only provides the information of linkages.</p><p>Another method is the current de novo structural determination for oligosaccharides developed by Bendiak et al. [34][35][36] . It provides the information of monosaccharide constitute, linkages, and anomeric configuration. However, the demonstrating experiment shows that it takes 11 hours to obtained good signal-to-noise ratio mass spectra from a tetrasaccharide 36 . These spectra only provide the information of two monosaccharide constitutes at the nonreducing side, and the linkages and anomeric configurations of two glycosidic bonds. One glycosidic bond and two monosaccharide constitutes at the reducing side cannot be determined. In addition, Bendiak's method only works for linear oligosaccharides.</p><p>The third method is to build glycan mass spectrum libraries. Mass spectra of unknown sample are compared to the spectra in glycan library for structural identification. However, building a complete glycan MS library is time consuming, considering that the high number of carbohydrate isomers (e.g., 10 12 isomers for an oligosaccharide containing six hexoses) of a given chemical formula. Most of these isomers are not available at this moment and the synthesis of each isomer takes weeks for an experienced chemist. Even if all isomers are available, they are not likely to be distinguishable from each other by a single mass spectrum. Mass spectra obtained from multiple-stage tandem mass spectrometry are necessary in the structural determination. If each stage generates 10 fragments, there are more than 100 spectra in a MS 4 experiment. The measurement of 100 spectra for a given isomer is impractical, not to mention that most of the spectra are similar or identical which are not useful in the structural identification. Hence, a guild line to choose the critical fragments for MS n measurement is needed.</p><p>In this study, we demonstrated a simple and rapid method with high sensitivity for the structural determination of underivatised glucose oligosaccharides. Our method can determine the linkages, branch location, and anomeric configurations for both linear and branched oligosaccharides. Currently, other methods can only determine the molecular weights and part of the structures during the short appearance period of each oligosaccharide in liquid chromatography. In our method, the CID spectra with good S/N ratio can be obtained within a very short period of time, and the number of CID spectra requires for the structure determination is minimized by the logical procedure we developed. These advantages enable us to in situ determine the structure of each oligosaccharide separated from liquid chromatography.</p><p>Our method provides a simple logical procedure to determine the structural decisive fragments for MS n measurement and only the disaccharide database is required. It greatly reduces the effort in building the glycan mass spectrum library for structural identification. The concept of this method can be extended to larger glucose oligosaccharides. Because the structural similarity of galactose, mannose, and glucose, the same concept can be applied to galactose-and mannose-oligosaccharides if the corresponding disaccharide database is available. The applications to mannose are demonstrated in a separate report 46 .</p><p>The drawback of our method is that it does not work for mixture of glycans if the glycans happen to have the same molecular weight. This is also the limitation of current structural determination methods using MS. The probability of the coincidence that glycans with the same molecular weight happen to be in a single CE or HPLC eluent peak is small. Consequently, combination of CE or HPLC with mass spectrometer in our method can conquer most of the difficulty in the analysis of mixture, although such combination may require larger amount of sample.</p><!><p>The CID spectra of disaccarhides in the database were measured by using a heated electrospray ionization (HESI-II) probe with an Ion Max housing and a linear ion trap mass spectrometer (LTQ XL, Thermo Fisher Scientific, Waltham, MA USA) coupled with an HPLC system (Dionex Ultimate 3000, Thermo Fisher Scientific, Waltham, MA USA) in the positive mode. The entire HPLC and mass spectrometer system is controlled by using Dinoex Chromatography MS Link 2.14, Chromeleon Version 6.80 SR13, LTQ Tune Plus Version 2.7.0.1103 SP1, and Thermo Xcalibur 2.2 SP1.48 software from Thermo Fisher Scientific. No customization of these instruments was made.</p><p>Liquid chromatography separation of all disaccharides was achieved using a Hypercarb (100 × 2.1 mm 2 , Thermo Fisher Scientific, Waltham, MA USA) column with a particle size of 3 µm operated in the multistep gradient mode at 25 °C. The mobile phase comprised (A) 0.1% (v/v%) aqueous formic acid containing 1 × 10 −4 M NaCl and (B) HPLC-grade acetonitrile. The multistep gradient mode conditions were as follows: t = 0 min, A: 100%, B: 0%; t = 1 min, A: 100%, B: 0%; t = 21 min, A: 90%, B: 10%; t = 21.1 min, A: 100%, B: 0%. For laminaribiose, the mobile phase gradient was as follows: t = 0 min, A: 95%, B: 5%; t = 1 min, A: 95%, B: 5%; t = 26 min, A: 94%, B: 6%. Samples were prepared in ultrapure water at a concentration of 1 × 10 −4 M. The injection volume of the sample was 10 µL, and the mobile phase flow rate was 300 µL/min. The column eluate was directly infused into the ESI source without any postcolumn addition. The MS conditions were optimised using the built-in semiautomatic tuning procedure in the Xcalibur software. The ESI source was operated at a temperature of 280 °C with 30 units of sheath gas flow and 10 units of auxiliary gas flow. The ion spray voltage was 4.00 kV, and the transfer capillary temperature was 280 °C. The capillary voltage was 80 V, and the tube lens voltage was 150 V. Helium gas was used as a buffer gas for the ion trap as well as a collision gas in CID. The pressure of He gas at the output of regulator connected to gas cylinder was set at the specification (40 psi). The pressure measured by the ion gauge in the vacuum chamber of mass spectrometer was 0.9 × 10 −5 Torr. The MS n experiments were performed at an activation Q value of 0.25, an activation time of 30 ms, normalised collision energy 25% for standard spectra of disaccharide database, and normalised collision energy 20-100% with 10% increment for test spectra. The numbers of standard and test spectra taken for database and the calculations of uncertainty were described in Supplementary Information. The number of ions was regulated by injection time (5 ms) or automatic gain control (1 × 10 5 for full scan, and 1 × 10 4 for MS n ). The precursor ion isolation width was set to 1 or 2 u. No difference in spectra was observed for the change of isolation width. MS n of panose in Fig. 8 were measured using the same conditions of disaccharides. Spectra were analysed by in-house code. Threshold of spectra was set to be 0.01.</p><!><p>All oligosaccharide MS n (except Fig. 8) were obtained using the same mass spectrometer under the same operation conditions as for disaccharides, except that HPLC was not used, the ESI source was operated at a temperature of 35 °C, and CID was performed only at normalised collision energy of 30%. Samples were prepared in 50% (v/v%) HPLC-grade methanol and ultrapure water at a concentration of 1 × 10 −4 M. Sodium chloride was added to the sample solution at a concentration of 1 × 10 −4 M. A total of 1 or 2 minutes spectral acquisition time were accumulated for each MS spectrum in Figs 4-6. They are the average of 250 or 500 microscans.</p>

Scientific Reports - Nature

Benchmarking Adaptive Variational Quantum Eigensolvers

By design, the variational quantum eigensolver (VQE) strives to recover the lowest-energy eigenvalue of a given Hamiltonian by preparing quantum states guided by the variational principle. In practice, the prepared quantum state is indirectly assessed by the value of the associated energy. Novel adaptive derivative-assembled pseudo-trotter (ADAPT) ansatz approaches and recent formal advances now establish a clear connection between the theory of quantum chemistry and the quantum state ansatz used to solve the electronic structure problem. Here we benchmark the accuracy of VQE and ADAPT-VQE to calculate the electronic ground states and potential energy curves for a few selected diatomic molecules, namely H2, NaH, and KH. Using numerical simulation, we find both methods provide good estimates of the energy and ground state, but only ADAPT-VQE proves to be robust to particularities in optimization methods. Another relevant finding is that gradient-based optimization is overall more economical and delivers superior performance than analogous simulations carried out with gradient-free optimizers. The results also identify small errors in the prepared state fidelity which show an increasing trend with molecular size.

benchmarking_adaptive_variational_quantum_eigensolvers

1. Introduction<!>2. Variational Quantum Eigensolver<!>2.1. ADAPT-VQE<!>2.2. Gradient Estimate and Classical Optimization in ADAPT-VQE<!>3. Computational Details<!>4. Results and Discussion<!>4.1. Potential Energies Curves<!><!>4.1. Potential Energies Curves<!><!>4.1. Potential Energies Curves<!><!>4.2. Optimization Strategies<!><!>4.2. Optimization Strategies<!><!>4.3. State Fidelities<!><!>4.3. State Fidelities<!><!>4.4. Resource Estimation<!><!>4.4. Resource Estimation<!><!>4.4. Resource Estimation<!>5. Conclusion<!>Data Availability Statement<!>Author's Note<!>Author Contributions<!>Conflict of Interest

<p>Quantum mechanics naturally lends itself to the description of phenomena at the atomic and molecular scale, including problems of chemical interest, which has culminated in the field of research known as quantum chemistry. Despite the formal impediments to achieve exact, closed-form solutions to quantum chemistry problems, there is a wide array of possible approximations, such as coupled cluster (CC) theory (Shavitt and Bartlett, 2009), which have elevated quantum chemistry to good standing in the scientific community due to their reliability.</p><p>In practice, CC faces two main difficulties that have hindered a more widespread adoption. One is that most of the success it has garnered over the years is due to its superior performance in the weak electron correlation regime, for which single-reference (SR) CC remains unchallenged. This success is justified because many problems in chemistry, such as thermochemistry, can be adequately treated as being largely weakly correlated. Yet, many other problems of interest involve molecules and materials that do not comply with this assumption, and for these instances, SR-CC breaks down. Despite multi-reference (MR) CC being an active area of research (Jeziorski and Monkhorst, 1981), theoretical and computational challenges currently curb the applicability of MR-CC (Lyakh et al., 2012).</p><p>A second obstacle to a more extensive use of CC theory is its pronounced computational cost. Reliable SR-CC methods, such as the so-called "gold standard" of quantum chemistry, coupled cluster singles and doubles (and perturbative triples), aka CCSD(T) (Urban et al., 1985; Raghavachari et al., 1989; Watts et al., 1993), scale unfavorably with one-particle basis spanning the Hilbert space that houses the electronic wave function, which largely constrains the application of CCSD(T) to relatively small molecular systems. It is important to note that some of these limitations can be mitigated with methods such as configuration interaction (CI) in its MR formulation and the density matrix renormalization group (DMRG) which have in turn their own shortcomings, such as lack of size-extensivity and exactness contingent upon the dimensionality of the problem.</p><p>Concurrent with developments in CC theory has been the increase in performance of computing technologies, which broadens the reach of computational chemistry techniques. Presently, this trend is continuing with the adaptation of chemistry methods, including CC, to the new technology paradigm of quantum computing (Britt and Humble, 2017; Humble et al., 2019). Because of the shared foundation in quantum mechanics, one of the most immediate applications for quantum computers is quantum chemistry (McArdle et al., 2020). Recent advances have reformulated conventional problems in electronic structure for currently available quantum computing platforms (Cao et al., 2019). In particular, these efforts have led to a resurgence of the unitary coupled cluster (UCC) theory (Bartlett et al., 1989; Kutzelnigg, 1991; Taube and Bartlett, 2006; Romero et al., 2018), which can be employed in investigations where strong correlation is dominant. Quantum computing hardware appears to be well suited for building the states described by UCC, as this hardware can efficiently implement unitary operations to construct physical representations of the quantum state. Moreover, the intrinsic nature of the quantum computing logic can be exploited in order to propose new ansatze that, despite lacking a close connection to the underlying chemical intuition lent by UCC, are prone to a more efficient implementation, such as the so-called hardware efficient ansatz (Kandala et al., 2017).</p><p>It is in the context of noisy intermediate-scale quantum (NISQ) (Preskill, 2018) devices that the variational quantum eigensolver (VQE) (Peruzzo et al., 2014) has emerged as a promising method for testing the preparation and measurement of quantum states including those that represent the electronic eigenstates described by UCC (Quantum et al., 2020). Several variants of VQE are available (Parrish et al., 2019b; Chivilikhin et al., 2020), but all build on the variational principle from quantum mechanics, which constrains the quantum states that can satisfy the electronic eigenvalue problem (McClean et al., 2016). While the initial VQE proposal assumes a predefined ansatz, this constraint has been relaxed, opening the door to adaptive approaches (Grimsley et al., 2019; Ryabinkin et al., 2020), by which the preparable quantum states are driven by the problem at hand. In particular, the adaptive derivative-assembled pseudo-trotter (ADAPT) ansatz, which finds support on the recently coined "disentangled" UCC (Evangelista et al., 2019) and starts from an exact UCC representation of the electronic ground state to construct an approximate prepared state based on the dominant contributions. Early studies demonstrated this as a promising avenue for developing ansatze for specific molecules and constraints, such as highly accurate energetics or shallow circuits.</p><p>Here we benchmark adaptive VQE prescriptions, ADAPT-VQE in particular, by comparing the prepared quantum states with the conventional solutions obtained from exact diagonalization of the full configuration interaction Hamiltonian. We track the energy of the minimized expectation value as well as the fidelity of the corresponding prepared state using multiple ansatz, optimization methods, and molecular Hamiltonians. We calculate infidelity as a measure of error for the prepared quantum state relative to the expected, exact result from quantum chemistry using frozen-core Hamiltonians. Across these examples, we find that ADAPT-VQE is the more robust method due mainly to its performance with respect to optimization methods. While all methods lead to small errors as measured by the infidelity, these errors are found to grow with molecular size.</p><p>This presentation is structured as follows. In section 2, we provided an overview of the ingredients in the VQE approach relevant to our purposes, followed by a short exposition of the underpinnings of ADAPT-VQE (section 2.1) and a brief discussion on implementation of gradients and optimization in ADAPT-VQE (section 2.2). The computational details permeating the reported simulations are exposed in section 3. The main results are presented and discussed in section 4 and several conclusions are drawn in section 5.</p><!><p>This section serves to illustrate the pertinent fundamentals of the VQE algorithm and to motivate the following exposition of adaptive ansatz construction. We start by recalling the variational principle, which is at the heart of VQE, and given as</p><p>where |Ψ〉 is normalized trial wave function for which Equation (1) becomes an equality when Ψ is constructed from a basis that spans the single-particle Hilbert space of all possible occupation numbers (the underlying Fock space) and the electronic Hamiltonian Ĥ for a molecular system is given as</p><p>The central problem in modern electronic structure theory is the description and quantification of the electron correlation from an un-entangled, mean-field wave function |0〉 whose preparation can be carried out in classical hardware in a timely fashion, e.g., Hartree-Fock (HF). In analogy with quantum chemistry, we can expect that there exists an operator that, once applied to |0〉, will account for the missing electron correlation. Bearing in mind that quantum computers manipulate quantum states in a well-defined Hilbert space, this configures a generic unitary operator Û(θ→) whose main purpose is to build entanglement from an un-entangled reference function |0〉. The set of scalars θ→ are parameters variationally varied in order to minimize the expectation value in Equation (1). With that, we recast Equation (1):</p><p>In order to ensure that Equation (3) meets the requirements of quantum hardware, the fermionic, second-quantized operators found in the formulation of electronic structure problem, such as those in Equation (2), are brought to a qubit (spin) representation, with the additional constraint of fermionic anti-symmetry. Our approach uses the Jordan-Wigner transformation (Jordan and Wigner, 1928), but others exist, and such a transformation yields ĤP from Ĥ, that is, the Hamiltonian in terms of strings of Pauli operators. Starting from the UCC ansatz, the unitary Û(θ→) can be written as:</p><p>with the T^k representing the usual cluster operators in CC theory and τk=T^k-T^k†, ensuring the anti-Hermiticity of the operators, which is a necessary condition for their utilization in quantum computing.</p><p>Once in possession of all ingredients in Equation (3), the tasks of preparing the state Û(θ→)|0〉 and measuring the terms in ĤP are delegated to the quantum hardware, and Û(θ→) is varied variationally with the aid of a classical optimization routine until 〈ĤP〉 reaches its minimum, which is dependent on the chosen optimizer and is taken as a good approximation to the sought ground state energy. Due to the isomorphic property of the qubit mappings, 〈ĤP〉 = 〈Ĥ〉, yielding the lowest energy eigenvalue of the molecular Hamiltonian in Equation (2).</p><!><p>An important choice in the specification of the VQE method is the functional form of the ansatz Û(θ→). Even for a relatively small Hilbert space, with a moderate number of cluster operators T^i, the ansatz Û(θ→) gives rise to a unitary that translates into multi-qubit gates and thus cannot be efficiently implemented in an actual quantum processor. Borrowing from the dynamics community, this can be alleviated by resorting to the Trotter-Suzuki decomposition, or Trotterization for short:</p><p>which here is limited to first-order.</p><p>The Adaptive Derivative-Assembled Pseudo-Trotter ansatz Variational Quantum Eigensolver (ADAPT-VQE) (Grimsley et al., 2019) takes advantage of Equation (5) to propose an iterative ansatz construction whereby only the perceived most relevant operator for energy lowering is added to the ansatz. A set of operators the algorithm can choose from needs to be provided, which in this work is comprised of the fermionic spin singlet adapted single and double excitations, borrowing from the usual UCCSD formulation, and subsequently mapped into the appropriate tensor products of Pauli operators via the Jordan-Wigner transformation. In principle, one could envision explicit enforcement or relaxation of other types of symmetry, and the effect of such choices on the performance of ADAPT is certainly a topic worth exploring. Moreover, the ADAPT algorithm has also recently been reported to perform well with other choices of operators, including a more economical pool of qubit operators (Tang et al., 2020), and has been applied to variational algorithms other than VQE (Zhu et al., 2020).</p><p>From a practical standpoint, at the i-th iteration of the algorithm, the energy gradient vector (G) with respect to all {θk} in Equation (5) is computed from measurements on the circuit that prepares the state optimized in the previous iteration, represented by |ψi−1〉, with |ψ0〉 = |0〉. Labeling the energy at the current iteration Ei, we have:</p><p>and if the norm of this vector falls below a set threshold, the algorithm is deemed converged and the ansatz-growing loop is exited. Otherwise, the operator associated with the largest absolute component of G is selected to increment the ansatz:</p><p>where 〈[H,τk]〉i−1 means this commutator was computed from observations in the circuit obtained from the previous iteration. With the selection of a new operator, the new ansatz is subject to the usual VQE routine and the corresponding energy minimum is obtained.</p><!><p>From a quantum computing standpoint, ADAPT-VQE improves on VQE by potentially offering a more tractable circuit. However, this may come at the expense of a much larger number of measurements, as the evaluation of all [H, Âk] is performed at each iteration of the ADAPT loop, on top of the expected energy evaluations. In order to reduce the number of measurements associated with ADAPT-VQE simulations, adoption of a gradient estimate strategy can help improve the classical optimization step by reaching the sought minima with fewer calls to the hardware backend.</p><p>To motivate the discussion, we start by invoking the gradient expression as introduced in the original formulation of ADAPT-VQE:</p><p>where ∏​ieθiτi|0〉=|ϕ〉.</p><p>Equation (8) can be further simplified into a recursive formula:</p><p>Before moving further in the discussion regarding the use of gradients to support the classical optimizer, let us clarify a potential source of confusion. At a certain ADAPT-VQE iteration, the circuit previously optimized is implemented to prepare the state from which the current iteration builds upon. The gradient vector G is then computed upon the necessary measurements for all τk in the chosen operator pool (Equation 6), and the operator that has the largest commutator (in absolute value) is selected. And this the extent to which the gradient is employed at this stage. On the other hand, we now have a new ansatz, which is composed of all previously added operators that enable preparation of |ψi−1〉, along with the newly added operator from Equation (7). Each of these operators have a corresponding variational parameter, which in the following VQE step need to be re-optimized. It is in this optimization that we would employ the gradients as written in Equations (8) and (9), and whose magnitude needs to be minimized in order to signal the finding of an extremum (minimum in this case). For an operator pool containing N elements, at each ADAPT-VQE iteration, all N elements of G need to be evaluated, but the magnitude of this vector is not directly minimized by varying the circuit parameters, only indirectly by the addition of enough operators in the ansatz. On the other hand, for optimization purposes, at the i-th iteration, only i gradient elements are considered, and the search for the energy minimum is guided by the minimization of the magnitude of this i-th dimensional gradient vector. Finally, another crucial point worth pointing out is that the commutators in Equation (6) are equivalent to those in Equation (9) only for the operator most recently added, i.e., τi in Equation (7).</p><p>For the purposes of an economical quantum resource utilization, it is desirable to deploy only one circuit to be used in both energy and gradient estimates (the same circuit is implemented many times, one for each term in the Hamiltonian/gradient). Even though the recursive formula in Equation (9) could, in principle, satisfy this requirement, this commutator cannot be measured (Mitarai et al., 2018). As originally proposed, the gradient is no longer given in an expectation value form, requiring an auxiliary state to be prepared via introduction of ancilla qubits, which deviates from our requirement of saving quantum resources. For that reason, we resort to numerical finite differences as means of carrying out gradient-based optimizations in the current work.</p><p>In terms of resource estimation, for a circuit depth of O(N), forward or backward finite differences are akin to introducing a single Rz(h), where h is the step size, leading the a circuit depth of O(N+1), while the use of central differences, thus, has circuit depth of O[2(N+1)], the former being used here due to its superior convergence properties. This is the cost incurred in the numerical gradient estimate for each parameter being optimized and a detailed discussion is provided in section 4.4. Such an estimate may be improved with strategies such as the quantum natural gradient (Stokes et al., 2020) or exploiting partial tomography (Parrish et al., 2019a). These ramifications are worthy of a separate study, and will not be further investigated here.</p><!><p>The quantum simulations detailed in this manuscript were performed using the VQE and ADAPT-VQE algorithms and numerical gradient strategies as implemented in the XACC hybrid quantum-classical computing framework (McCaskey et al., 2018b, 2020), with the latter algorithm leveraging a convergence criterion of ||G||≤10-2. We emphasize that this parameter can be of substantial impact on the results, as it controls the size of the obtained ansatz. In light of the findings in Grimsley et al. (2019), the adopted value in this paper is believed to strike a satisfactory balance between accuracy and circuit depth. The resulting circuits were simulated via the TNQVM (tensor-network quantum virtual machine) (McCaskey et al., 2018a) XACC simulation backend and employed a noiseless, matrix product state (MPS) wave function decomposition for the quantum circuit with the aid of the ITensor library (Fishman et al., 2020). XACC provides other simulation backends, as well as physical backends targeting QPUs from IBM and Rigetti. For the size of the problems studied in this work, there may not be perceived benefits from choosing TNQVM over other XACC simulation backends like Aer (Abraham et al., 2019) or QPP (Gheorghiu, 2018). TNQVM is expected to be advantageous over other simulation approaches for problems requiring more qubits (McCaskey et al., 2018a), but we leave this to future work and do not investigate it here.</p><p>The COBYLA (Powell, 1994) algorithm was used as a gradient-free optimizer, while gradient-based optimizations were carried out with the L-BFGS algorithm (Nocedal, 1980; Liu and Nocedal, 1989), with all parameters being initialized at 0 at each optimization cycle for both optimizers. Other approaches have been reported in the literature, such as random initialization (Grimsley et al., 2020), or as in the original implementation of ADAPT-VQE (Grimsley et al., 2019) where the new parameter is initialized at 0, while the previous parameters are initialized from the optimal values obtained in the previous ADAPT iteration. XACC offers both optimizers via its interface with NLOpt (Johnson).</p><p>The potential energy curves (PEC) of NaH and KH, were generated by imposing the frozen-core approximation, reducing the number of configurations to only those arising from one σ orbital and its σ* counterpart, that is, a two electrons in two orbitals complete active space [CAS(2,2)] problem. The one- and two-electron integrals necessary for the construction of the Hamiltonians and the corresponding references CAS energies were obtained with PySCF (Sun et al., 2017), with all calculations performed with the STO-3G basis set (Hehre et al., 1969, 1970; Pietro et al., 1980).</p><p>The quality of the output circuits in preparing the desired states is assessed via the fidelities computed with respect to the ground state full configuration interaction (FCI) wave function. This corresponds to the lowest energy eigenvector from exact diagonalization in the 2N Hilbert space, with orbital occupation determined by the number of electrons. In possession of the circuits from the quantum simulations, the respective state vector representation is obtained using the XACC interface to the Qiskit Aer simulator (Abraham et al., 2019).</p><!><p>Typically, the quality of the state obtained from the variational optimization of the gate parameters is probed indirectly by comparison of the computed energies with trustworthy references values or the exact lowest energy eigenvalue whenever computationally feasible. Thus, we start by investigating the energy profile along the atomic displacement, and subsequently contrast these findings with the analysis of the appropriateness of the corresponding states via evaluated fidelities relative to the vector corresponding to the lowest eigenvalue in the active space.</p><!><p>We start investigating the behavior of the energy by studying the H2 molecule. This example has been extensively approached in quantum computing, and hardly poses any difficulty, at least from the standpoint of numerical simulations, as opposed to deployment to actual hardware. However, it serves as a baseline for the following discussion, as the orbital spaces in the other molecules are reduced to an active space with the goal of resembling the H2 molecule. Results with the VQE and ADAPT-VQE ansatze are plotted in Figure 1, along with FCI results.</p><!><p>(Top) Potential energy curves of H2 computed with the STO-3G basis set for FCI (green solid line), VQE (blue circles), and ADAPT-VQE (orange diamonds) with the COBYLA optimizer. (Bottom) Absolute error in the minimized energy for VQE (blue) and ADAPT-VQE (orange) relative to the FCI reference value.</p><!><p>Unsurprisingly, there is a remarkable agreement between simulated and exact values, both qualitatively and quantitatively. Absolute errors from FCI are found in the sub-miliHartree range throughout the energy scan, and with either choice of ansatz, the observed errors would be inconsequential when taking into account the scale of the errors introduced by noise in the operation of quantum devices. The impression that some points are "missing" from the bottom plot of Figure 1 is explained by these values being numerically identical to the FCI values (to seven decimal places), hence not being plotted in the logarithmic scale.</p><p>The results from the potential energy curve from simulations on the NaH molecule are presented in Figure 2.</p><!><p>(Top) Potential energy curves of NaH computed with the STO-3G basis set for FCI (green solid line), VQE (blue circles), and ADAPT-VQE (orange diamonds) with the COBYLA optimizer. (Bottom) Absolute error in the minimized energy for VQE (blue) and ADAPT-VQE (orange) relative to the FCI reference value.</p><!><p>Visual inspection of the top plot reveals that the choice between the two ansatze being considered here yield energies that track one another very well, but because of the energy scale of this plot, it begs a closer look. The bottom plot displays the absolute errors between VQE and ADAPT-VQE with respect to FCI. The errors here are still within chemical accuracy (<1 kcal/mol), and are unlikely to be of much relevance in the total error if such simulations are executed in a quantum computer. However, there is a clear trend of increase in the magnitude of the computed deviations when compared to the hydrogen molecule, whose results are in Figure 1.</p><p>In Figure 3, we again observe some of the patterns that follow from the analysis of Figures 1, 2. The energy scale here is much too large to able to reveal relatively minor inadequacies, even though qualitative discrepancies, such as those arising from symmetry breaking or the crossing of lines of different states, would be evident had they been present. The bottom plot, exhibiting the energy differences from FCI, offers a more reliable evidence, allowing us to infer that ADAPT-VQE is overall superior, with smaller errors for the vast majority of points (the exception being 1.4 Å). Perhaps more importantly, we observe a general trend of the points from simulations with the plain VQE ansatz approaching the 1 mHartree, with the distances of 2.9 and 3.9 Å now found more than 1 kcal/mol above the respective FCI energy.</p><!><p>(Top) Potential energy curves of KH computed with the STO-3G basis set for FCI (green solid line), VQE (blue circles), and ADAPT-VQE (orange diamonds) with the COBYLA optimizer. (Bottom) Absolute error in the minimized energy for VQE (blue) and ADAPT-VQE (orange) relative to the FCI reference value.</p><!><p>The potential energy curves presented and discussed in section 4.1 are based upon gradient-free optimization carried out with the COBYLA optimizer. We report that analogous simulations were performed with the Nelder-Mead optimizer, which is also a gradient-free alternative, but preliminary investigations pointed to COBYLA being a superior choice, at least for the chosen molecules. To contrast the performance of gradient-free optimization in the current context, we use the L-BFGS optimizer for parameter update, as implemented in NLOpt, with gradient estimated via central numerical finite differences. To assess the relative performance of these two approaches as the bond in the current diatomic molecules is stretched, we plot the difference between energies obtained with the COBYLA optimizer and those with L-BFGS+finite differences, that is, E(COBYLA) − E(L-BFGS). That way, positive energy differences indicate there is an improvement by turning to a gradient-based optimization, while the opposite signals that the current gradient-free method reached a lower energy.</p><p>We observe compatible energies for the H2 case, regardless of the underlying optimization strategy, for the entirety of Figure 1. In order to maintain consistency, we plot the energy difference between the two optimization prescriptions in a miliHartree scale, and the spike in E(COBYLA) − E(L-BFGS) in 1.7Å, when rationalized with the scale in mind, shows a deviation in the μHartree range. Due to the presence of all the many-body operators necessary for exactness (Evangelista et al., 2019), we expect and in fact observe results on par with the numerical precision imposed by the employed optimizers (10−6 Hartree in relative energy).</p><p>While most of the PEC for H2 showed no major dependence on the adopted optimization procedure, according to Figure 4, the picture is significantly different in the case of NaH, as portrayed in Figure 5. Even though the values for E(COBYLA)−E(L-BFGS) are still rather small, in the sub-miliHartree range, noticeable differences are more frequent here. Albeit of μHartree in magnitude, we also observe cases where COBYLA provides a lower energy than L-BFGS, most notably for ADAPT-VQE in the 1.4 and 2.5 Å interatomic distances. On the other hand, in an overall assessment of the performance between VQE and ADAPT-VQE, the latter displays a more pronounced insensitivity with respect to the choice of optimization scheme.</p><!><p>Difference between the energies from COBYLA and L-BFGS optimization with central finite differences for the H2 potential energy curve.</p><p>Difference between the energies from COBYLA and L-BFGS optimization with central finite differences for the NaH potential energy curve.</p><!><p>An even more drastic contrast is found from inspection of Figure 6, where E(COBYLA) − E(L-BFGS) are plotted for the KH molecule. Some of the qualitative assertions pointed out in Figure 5 hold, namely that the performance of VQE is much more influenced by the choice of optimization strategy than ADAPT-VQE. Not only that, but ADAPT-VQE is largely unaffected by employed optimizer, at least between the two alternatives in consideration. Here again, the differences seen for VQE correlated well with the deviations from FCI reported in Figure 3, further corroborating the claim that a gradient-based optimization, given the current conditions, is a more robust for approaching the lowest energy eigenvalue of molecular Hamiltonians.</p><!><p>Difference between the energies from COBYLA and L-BFGS optimization with central finite differences for the KH potential energy curve.</p><!><p>As previously stated, energy values can be used as valuable metric of the adequacy of a given set of variational parameters and trial state. However, the energy alone may not be indicative of the quality of the corresponding state and even acceptable energy values do not guarantee equally satisfactory values for other properties. The usual electronic Hamiltonian, as shown in Equation (2), transforms as the most symmetric irreducible representation for a given point group, therefore yielding the same energy in the case of degenerate states. Other operators, however, such as the terms in the multipole expansion of the electric potential, do not display this feature, meaning that degenerate states may yield different expectation values for such operators.</p><p>In order to examine the state prepared by the two circuit approaches considered here, we compute their "infidelities" with respect to the exact FCI state within the aforementioned active spaces, which is mathematically represented by 1-|〈ΨFCI|Û(θ→)|0〉|, where θ→ here are the set of optimal values also utilized for the energy computations in sections 4.1 and 4.2. We acknowledge that, while this provides direct inroads in the state being output at completion of the state preparation, it cannot be experimentally realized. However, in the case of moderately sized molecules for which the exact diagonalization of the Hamiltonian is feasible, this can provide valuable insights.</p><p>The energy differences discussed in the case of the hydrogen molecule in sections 4.1 and 4.2 are quite small when considering the magnitude of the other potential sources of error that can arise in the presence of noise, either through a model or in the operation of an actual quantum device. Due to the simplicity of the electronic structure of this molecule the state prepared according to the two ansatze construction prescriptions investigated here yield infidelities that are below the numerical thresholds employed here, and certainly would be unnoticeable for realistic purposes. Because they offer little insight, we abstain from plotting the infidelity results for this molecule here.</p><p>Before delving into the particularities of each curve in Figure 7, we bring the reader's attention to the scale of the plots, signaling a remarkable agreement between the state prepared and the one expected (FCI). It should come as no surprise that the largest infidelities are found in the vicinity of the Coulson-Fischer point, the most demanding region in the energy landscape, and subsequently approach zero as the atoms are moved far apart. The infidelities for the VQE ansatz follow a smooth progression when employed in conjunction with the gradient-based optimizer L-BFGS, whereas the same is not true for the other combinations of ansatze and optimization. This is likely a compound effect, explained by the former being a fixed circuit, where only the associate θ→ changes throughout the energy scan. The latter, however, can assume a different composition, changing according to the demands of the electronic structure at each bond length. This works along with the fact that gradient-based optimization, at least in the current study, provides a tighter, more reliable solution. For the NaH and KH cases, we plot the number of operators in the ansatz in Figure 8.</p><!><p>State infidelities for VQE and ADAPT-VQE using COBYLA (solid line) and L-BFGS (dashed line) optimization with central finite differences for the NaH potential energy curve.</p><p>Number of operators in the ADAPT-VQE ansatze using the COBYLA (solid line) and L-BFGS (dashed line) optimizers. The corresponding VQE ansatz has two operators.</p><!><p>Once again, there is a clear advantage in turning to gradient-based optimization, as it renders ansatze with fewer operators. For some internuclear distances, the ADAPT-VQE ansatze, even when optimized with L-BFGS, contain more operators than the corresponding VQE ansatz. This is not necessarily in contradiction with the some of the findings from Grimsley et al. (2019) because those results were obtained for different molecules and using different optimization implementations. Yet, we would expect that when comparing against a larger VQE problem, such as those investigated in that paper, we would see similar trends. We also speculate that another variable that can contribute to the observed behavior is the tolerance that controls how tight the optimization should be. Because we are using a default 10−6 threshold in relative energy as the tolerance and there is no clear connection between the quality of the energies and the respective prepared states, the absolute energies values may fall in a scale that may have a small, but non-negligible effect on the fidelities, which is also evidence of the effect it can have in the output state, further corroborated by the number of operators found in the respective ansatze, yet not enough to alter any of the main conclusions drawn from the results presented here.</p><p>Many of the main inferences from the analysis of the Figure 7 hold for the KH molecule, whose infidelities are shown in Figure 9. Firstly, the infidelities, though still quite small, are about an order of magnitude larger. The smoothness and overall profile observed for the VQE UCCSD is retained, but the behavior of the ADAPT-VQE infidelities is much more erratic. Secondly, while the ADAPT-VQE ansatz for NaH around the Coulson-Fischer point is mostly the same, but the larger number of variational parameters make it more vulnerable to the optimization inconsistencies discussed above, here the large oscillations are due to ansatze of alternating operator compositions. Because the ADAPT-VQE convergence criterion depends upon a fixed numerical threshold, sometimes the ansatz at a given iteration may already be close to convergence, but still not quite below the gradient norm threshold, and upon the addition of an extra operator, the state may be improved significantly in the scale of the plots seen in this section.</p><!><p>State infidelities for VQE and ADAPT-VQE using COBYLA (solid line) and L-BFGS (dashed line) optimization with central finite differences for the KH potential energy curve.</p><!><p>One of the main motivations behind the present work is to serve as the baseline for following studies focusing on the investigation of the electronic structure of molecules carried out in NISQ devices. With this in mind, it is important to develop some intuition on the resource demands involved in such tasks.</p><p>First we analyze the circuit proposed by VQE and ADAPT-VQE to prepare the states whose energies and fidelities were shown in sections 4.1–4.3 in terms of total gate count and circuit depth, plotted in Figure 10.</p><!><p>Gate counts (left axis) and circuit depth (right axis) from ADAPT-VQE circuits optimized with the COBYLA and L-BFGS optimizers. The black solid and dashed lines are the gate counts and circuit depths from the VQE ansatz, respectively.</p><!><p>Let us first compare the ADAPT-VQE results on the basis of the two optimizers. As we move from H2 to NaH and KH, we see a more intricate picture of how these optimizers impact the final circuit. Qualitatively, L-BFGS has an overall advantage as it provides circuits that are shallower and with fewer gates. While there are a few points along the potential energy scans where the circuits generated based on L-BFGS are not as efficient as those from a COBYLA optimization, the scales of the plots are determined solely by the latter. We noticed that in several points, the simulations with the COBYLA optimizer would produce states with two instances of the same operator adjacent to each other. If the actual minimum value had been achieved in a certain iteration of ADAPT-VQE, the commutator of the same operator in the next iteration would have been zero. Because this procedure is accomplished numerically, the magnitude of this commutator is related to how close the determined minimum is from the actual one. It turns out that the default threshold in relative energy (10−6) is found not to be stringent enough, which incurs a commutator whose deviation from the expected zero is non-negligible, resulting in the same operator being added in successive iterations. Another factor that accounts for the displayed circuit figures is the fixed gradient norm threshold in ADAPT-VQE. In some iterations, this quantity is above, but already quite close to the pre-defined 10−2, and one extra iteration is performed, with only marginal energy improvement. To illustrate this, the ADAPT-VQE simulation for NaH with internuclear separation of 1.8Å converges to ansatz with three operators, with E = −160.3146751 Hartree and ||G||=0.001. Had the ADAPT cycle been stopped in the second iteration, we would have ||G||=0.013, with E = −160.3146492 Hartree, that is, the energy improvement was in the μHartree range, yet at the expense of a deeper circuit, which calls for a more flexible operator selection in ADAPT-VQE.</p><p>These resource estimation parameters in Figure 10 are comparable between the two ansatz strategies. In general terms, the circuits optimized upon L-BFGS are more affordable than the corresponding VQE ones, while using COBYLA tends to yields circuits that are deeper and need to implement more gates. We bring attention to the fact that there is not a one-to-one correspondence between the present analysis and that in the Figures 2C,F,I in Grimsley et al. (2019). This is because the latter refers to the number of parameters/operators in the ansatz. A circuit with more parameters/operators does not readily translate into a more complex circuit, which depends on the number of qubits in a given operator and the operator locality and placement. This means these results are not at odds with what was previously reported, which were obtained for a distinct set of molecules, but can be seen as complementary.</p><p>Another important metric when estimating the necessary resources for implementation and deployment of the simulations discussed here is the number of measurements. To complement the end of the last paragraph, it is important to mention that in this context the rationalization in terms of number of operators increases in relevance. In Figure 11, we plot the total number of measurements to achieve the results reported in section 4.1.</p><!><p>Total number of measurements for final energy evaluation for the VQE and ADAPT-VQE ansatze using the COBYLA (solid line) and L-BFGS (dashed line) optimizers.</p><!><p>As pointed out in Grimsley et al. (2019), the ansatz put forth by ADAPT-VQE offers a trade-off between circuit depth and number of measurements. We can readily confirm by visual inspection of Figure 11 that ADAPT-VQE incurs a much larger number of measurements. These figures account for all measurements involved in computing the commutators in Equation (6), the energy evaluations at each optimization iteration, and the computations necessary to minimize the gradients when L-BFGS is employed. The measurement burden in ADAPT-VQE reported here can be partially alleviated by employing a better parameter initialization, such as starting the VQE optimization at each iteration with the previously optimized parameters and initializing just the newly added parameter at zero. This demand is also expected to be greatly relieved by resorting to a different set of operators, such as those introduced in the qubit-ADAPT-VQE variant (Tang et al., 2020), which can still span the underlying Hilbert space, yet with linear growth in the number of qubits. This approach would require much fewer commutator computations at each iteration, but would likely be of noticeable advantage for operator pools larger than those in question here. These results are also contingent upon the choice of optimizer, and there may exist better suited choices than those investigated here. Yet, we do not believe this would dramatically change the overall qualitative picture drawn in Figure 11.</p><p>Another key outcome from the analysis of Figure 11 is the fact that, even though the gradient computation with L-BFGS requires more measurements per iteration, it is overall much more economical than the gradient-free optimization, represented here by COBYLA. This furthers strengthens the case for gradient-based optimization in VQE, as it not only results in smaller errors/better convergence with respect to the sought ground state, but it is also much less demanding from a resource standpoint.</p><!><p>For a broader adoption of adaptive methods for ansatz construction in the realm of quantum chemistry, and perhaps, for many-body methods in general, many aspects still needs to be explored and their underpinnings better understood. This work provides a contribution toward this goal by showing a comprehensive study of potential energy curves of a selection of molecules of the general formula XH (X = H, Na, K). Despite their simplicity, they serve to shed light on some of the mentioned characteristics, and deliver a baseline for feasible studies involving actual quantum hardware.</p><p>Even a relatively conservative gradient norm threshold of 10−2 in ADAPT-VQE is sufficient to provide overall better energetics than corresponding fixed ansatz approach embodied by the ordinary VQE, which is in agreement with the initial ADAPT-VQE proposal. Due to its iterative nature, ADAPT-VQE has an extra layer of tunability which can be controlled via the threshold on ||G||. This means that the errors observed with ADAPT-VQE might have been reduced had ||G|| been made tighter, which could in turn increase the depth of the circuits, and even having to cope with more necessary measurements than those of UCCSD, as suggested with ||G||=10-3 in Figure 2i by Grimsley et al. (2019) However, upon a simple choice of gradient strategy motivated by the constraints of quantum hardware, we report that ADAPT-VQE is fairly resilient with respect to the employed optimization strategy and that encouraging improvements in performance by adopting a gradient-based approach in the search of the parameter set that minimizes the objective function can be mostly beneficial in the case of VQE. These findings call for a follow-up study on the role of optimizer in conjunction with ADAPT-VQE, extending the analysis to a larger selection of optimizers and gradient strategies.</p><p>The ongoing development of VQE methods, including ADAPT-VQE, must also address the noise that is intrinsic to the operations implemented in experimental quantum computers. The above benchmarks of infidelity and energy error place lower bounds on the expected accuracy for VQE methods using noiseless numerical simulation. However, we anticipate that the introduction of noise will substantially affect the accuracy with which the prepared ansatz state approaches the pure state expected from conventional quantum chemistry theory. However, if the state infidelity grows with increasing molecular size, as indicated by our short series of examples, then lower bounds on ansatz accuracy may become a non-trivial contribution to observed errors in experimental measurements.</p><!><p>The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation. XACC and TNQVM are open-source and can be found at https://github.com/eclipse/xacc and https://github.com/ornl-qci/tnqvm.</p><!><p>This manuscript has been authored by UT-Battelle, LLC, under Contract No. DE-AC0500OR22725 with the U.S. Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for the United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan.</p><!><p>DC implemented the ADAPT-VQE algorithm, ran some of the simulations, and wrote the manuscript. JW ran some of the simulations, wrote the code to compute state fidelities, and generated the plots. AM implemented the VQE algorithm and oversaw the ADAPT-VQE implementation. TH designed the research and helped writing the manuscript.</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

Recyclable Thermoresponsive Polymer-Cellulase Bioconjugates for Biomass Depolymerization

Here we report the construction and characterization of a recoverable, thermoresponsive polymer-endoglucanase bioconjugate that matches the activity of unmodified enzymes on insoluble cellulose substrates. Two copolymers exhibiting a thermoresponsive lower critical solution temperature (LCST) were created through the copolymerization of an aminooxy-bearing methacrylamide with N-isopropylacrylamide (NIPAm) or N-isopropylmethacrylamide (NIPMa). The aminooxy group provided a handle through which the LCST was adjusted through small-molecule quenching. This allowed materials with LCSTs ranging from 20.9 \xc2\xb0C to 60.5 \xc2\xb0C to be readily obtained after polymerization. The thermostable endoglucanase EGPh from the hypothermophilic Pyrococcus horikoshii was transaminated with pyridoxal-5\xe2\x80\x99-phosphate to produce a ketone-bearing protein, which was then site-selectively modified through oxime linkage with benzylalkoxyamine or 5 kDa-poly(ethylene glycol)-alkoxyamine. These modified proteins showed activity comparable to the controls when assayed on an insoluble cellulosic substrate. Two polymer bioconjugates were then constructed using transaminated EGPh and the aminooxy-bearing copolymers. After twelve hours, both bioconjugates produced an equivalent amount of free reducing sugars as the unmodified control using insoluble cellulose as a substrate. The recycling ability of the NIPAm copolymer-EGPh conjugate was determined through three rounds of activity, maintaining over 60% activity after two cycles of reuse and affording significantly more soluble carbohydrates than unmodified enzyme alone. When assayed on acid-pretreated Miscanthus, this bioconjugate increased the amount of reducing sugars by 2.8-fold over three rounds of activity. The synthetic strategy of this bioconjugate allows the LCST of the material to be changed readily from a common stock of copolymer and the method of attachment is applicable to a variety of proteins, enabling the same approach to be amenable to thermophile-derived cellulases or to the separation of multiple species using polymers with different recovery temperatures.

recyclable_thermoresponsive_polymer-cellulase_bioconjugates_for_biomass_depolymerization

Introduction<!>Materials and Methods<!>Synthesis of LCST Copolymers (1a and 1b)<!>Small-Molecule Modification of 1a and 1b<!>LCST measurements<!>Expression and Purification of AKT-EGPh<!>Small-Molecule Modification of EGPh<!>Construction of EGPh-polymer Bioconjugates<!>Protein Quantification<!>Activity of Modified EGPh<!>Activity of Recycled Polymer-EGPh Conjugates<!>Analysis of Soluble Reducing Sugar<!>Results and Discussion<!>Conclusion<!>

<p>Lignocellulosic biomass is a highly heterogeneous material composed of lignin, hemicellulose, and cellulose in a complex structure. The ability to convert the cellulose component of this material into fermentable sugars for biofuel production requires the cooperative action of the three cellulase enzymes endoglucanase, exoglucanase, and β-glucosidase. Collectively, these cellulases have been isolated from a broad assortment of organisms and exhibit a range of temperature, pH, and substrate optima.1 While the exact enzyme cost in the production of lignocellulosic biofuels depends on many factors – including the type of feedstock, enzyme loading, and overall biofuel yield – it is widely recognized that enzyme costs are a significant portion of biofuel prices and pose a key barrier to economically viable fermentation processes.2,3</p><p>One approach to reducing enzyme costs is to develop methods to collect and reuse enzymes through multiple rounds of processing. With this goal in mind, much work has been done in the field of immobilizing cellulases, including their covalent attachment or adsorption onto substrates such as silicon dioxide wafers,4 silica,5 glass beads,6 calcium alginate beads,7 and magnetic nanoparticles.8,9 By rendering the enzymes insoluble, however, access to the crystalline cellulose is potentially impeded and enzyme recovery can be difficult if any insoluble cellulose remains after the maximal hydrolysis yield has been reached. An alternative strategy is the use of stimuli-responsive or "smart" polymers, which are materials that undergo solubility changes in response to external stimuli such as alterations in pH or temperature. There has been some development of reversibly soluble-insoluble polymer-cellulase materials, most commonly utilizing pH sensitive polymers such as Eudragit L-100 or methacrylic acid polymers.10,11,12 These materials limit industrial processes to a fairly narrow pH range and require multiple pH adjustments to recover and reuse the enzyme.</p><p>As a more generalizable alternative, one can envision the attachment of cellulases to poly(N-isopropylacrylamide) (pNIPAm), which has been well-studied in biotechnology applications because of its thermal responsiveness.13,14 pNIPAm exhibits a highly reversible phase transition at its lower critical solution temperature (LCST) of 32 °C. The polymer is freely soluble below the LCST, but above the LCST the polymer chains undergo a spontaneous coil-to-globule transition, expelling water and precipitating (Figure 1). Several groups have utilized pNIPAm to develop a range of thermoresponsive polymer-biomolecule conjugates, typically by attaching the polymers or polymerization initiators to lysine side chains15 or introduced cysteines.16 In some cases, small molecule binders such as biotin17 or glutathione,18 have been used for protein introduction. As one particularly compelling example, the Hoffman and Stayton groups have conjugated a thermoresponsive polymer to a unique cysteine near the active site of an endoglucanase, with the goal of modulating enzyme activity through polymer collapse above the LCST.16 They found that the activity could indeed be controlled for the hydrolysis of a soluble substrate simply by changing the reaction temperature. This would suggest that similar LCST-based strategies could lead to recoverable enzymes after cellulose depolymerization, but to date these materials have not been studied for their recycling potential after significant levels of conversion, or for their use with insoluble cellulosic substrates.</p><p>Here we report a thermoresponsive endoglucanase bioconjugate that can match and even exceed the activity of unmodified enzymes on insoluble cellulosic substrates. We further demonstrate that this material can be recovered and used for several rounds of cellulose depolymerization, leading to substantially more glucose and cellobiose than can be produced by non-recoverable enzymes. The bioconjugation strategy allows the introduction of the polymer chains at the N-terminus, which is a position that is remotely disposed from the active sites of most cellulases.19 Finally, the synthetic strategy also allows the LCST of the polymer to be changed readily, enabling the same approach to be amenable to thermophile-derived cellulases or to the separation of multiple species using polymers with different recovery temperatures.</p><!><p>Unless otherwise noted, all chemicals and solvents used were of analytical grade and were used as received from commercial sources. Room temperature and 4 °C centrifugations were conducted either with a Sorvall RC 5C Plus (Sorvall, USA) for volumes greater than 50 mL, a Sorvall LEGEND Mach 1.6R for volumes between 1 and 50 mL, or an Eppendorf Mini Spin Plus for volumes less than 1 mL (Eppendorf, USA). Centrifugations above room temperature were performed on a Hettich Rotofix 46 H (GMI, Ramsey, MN). Samples were lyophilized using a LAB CONCO Freezone 4.5 (Lab Conco, USA). UV-Vis spectroscopic measurements were conducted in a Varian Cary 50 spectrophotometer (Agilent, USA). Fluorescence measurements of 96 well plates were obtained on a SpectraMax M2 (Molecular Devices, Sunnyvale, CA). Tryptophan fluorescence measurements were obtained using a 50 µL quartz cuvette on a Photon Technology International Quanta Master 4 L-format scanning spectrofluorometer (Lawrenceville, NJ) equipped with an LPS-220B 75-W xenon lamp and power supply, A-101013 lamp housing with integrated igniter, switchable 814 photon-counting/analog photomultiplier detection unit and an MD5021 motor driver. Unless otherwise noted, all buffers are 50 mM sodium acetate (NaOAc), pH 4.5.</p><!><p>The synthesis of tert-butyl 2-(3-(2-methylprop-2-enamido)propylamino)-2-oxoethoxycarbamate (MEPO) was adapted from a previously published procedure.20,21 Azobisisobutylonitrile (AIBN) was recrystallized once from pure methanol (MeOH) and N-isopropylacrylamide (NIPAm) was recrystallized twice from hexanes and toluene before use. Polymerization was conducted using a 1:9 molar ratio of MEPO:NIPAm and weight percent 11.1:0.6:88.3 for monomers:AIBN:CH3OH. MEPO (378.5 mg, 1.2 mmol), NIPAm (1.22 g, 10.78 mmol), and AIBN (80 mg, 0.49 mmol) were added to a clean scintillation vial. The vial was purged and refilled N2. MeOH (12.67 g, 395.4 mmol) which had been previously sparged with N2 for 1 h was added and the components were dissolved under N2. The mixture was divided into six clean scintillation vials, a stream of N2 was bubbled through the solution in each vial for 10 min, and the vials were sealed under N2 and placed in a 60 °C oil bath for 6 h. The polymer was recovered from the reaction mixture by one precipitation from MeOH into cold diethyl ether followed by centrifugation. It was dissolved in 1:1 CH2Cl2:trifluoroacetic acid for 1 h, concentrated in vacuo, then neutralized using 5 M NaOH. The polymer was purified through ultrafiltration (10 kDa molecular weight cut-off [MWCO]) and lyophilized to afford the final poly(MEPO-co-NIPAm) (1a). The molar ratio of MEPO:NIPAm was calculated to be 1:10.5, or 8.5% incorporation of MEPO, through 1H-NMR analysis (see Figure S1). SEC analysis using PMMA standards indicated Mn=85,127, Mw=155,321, and PDI=1.82. The same general procedure was followed to make a copolymer with MEPO and N-isopropylmethacrylamide (NIPMa) (1b). From analysis of the 1H-NMR spectrum, the molar ratio of MEPO:NIPMa was calculated to be 1:10.36, or 8.8% incorporation of MEPO (see Figure S2). SEC analysis using PMMA standards indicated Mn=11,144, Mw=14,037, and PDI=1.26.</p><!><p>A 20 mg/mL stock solution of 1a was made in pH 4.5 buffer. A series of 1 mL, 120 mM stock solutions of formaldehyde, acetone, 3-fluoroisonicotinaldehyde (with 10% DMSO), and 4-hydroxy-2-butanone, and 1.2 M stock solutions of D-(+)-mannose and D-(+)-dextrose were made in pH 4.5 buffer. 750 µL of the polymer solution was mixed 1:1 with each of the small molecule solutions in 4 mL glass dram vials. The reactions were incubated at rt for 24 h, then excess small molecules were removed and the polymers buffer exchanged into pure water through 8 rounds of ultrafiltration (10 kDa MWCO) at 4 °C. They were lyophilized, analyzed by NMR spectrometry to confirm modification, and the LCST was determined. The same procedure was followed for small-molecule modification of 1b.</p><!><p>Polymer samples were dissolved in pH 4.5 buffer at a concentration of 1 mg/mL and mixed for 30 min to ensure complete dissolution. They were transferred to a cuvette with a stir bar and warmed at a rate of 0.5 °C/min while stirring in a Horiba Scientific F-3004 Peltier device (Kyoto, Japan) controlled by a LFI3751 5A digital temperature control instrument (Wavelength Electronics, Bozeman, MT). The cuvettes were quickly removed every 0.5 °C and the absorbance at 600 nm was measured, then the cuvette was returned to the Peltier device. The LCSTs reported here are the temperature at 10% of the maximum absorbance for each sample.</p><!><p>BL21starDE3 E. coli cells containing an EGPh-pet24b plasmid were obtained from the Douglas Clark lab at UC Berkeley. 22 The plasmid initially contained both N-terminal and C-terminal His6 tags, so site-directed mutagenesis and restriction digestion were used to remove the N-terminal His6 tag and to install an AKT sequence at the N-terminus to maximize transamination yield (see Supporting Information). AKT-EGPh plasmids were transformed into One Shot® BL21 (DE3) E. coli cells (Invitrogen) via heat shock and plated on Luria broth (LB) agar plates containing kanamycin (50 µg/mL). Cultures were grown in 1 L of LB containing kanamycin (50 µg/mL) at 37 °C until an optical density (OD) of 0.5 was observed at 600 nm. Expression of AKT-EGPh was induced by the addition of 0.1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). Cultures were grown for 12 h at 25 °C, then spun down at 7,000 rcf, 4 °C for 40 min to pellet the cells. The cells were purified using nickel-nitrilotriacetic acid (Ni-NTA) agarose resin following the reccommended protocol (Qiagen). The purified protein was buffer exchanged into 50 mM NaOAc buffer (pH 4.5) through ultrafiltration (10 kDa MWCO) to yield 85 mg of purified protein per L of culture.</p><!><p>Transamination of the EGPh N-terminus was performed following a previously reported method.23 EGPh at a concentration of 50–60 µM in pH 4.5 buffer was mixed 1:1 with a pH 4.5 solution of 200 mM pyridoxal-5'-phosphate (PLP). Samples were reacted for 1 h at 37 °C, then excess PLP was removed by eight rounds of ultrafiltration (30 kDa MWCO) at 4 °C. Controls were conducted following the same procedure but without PLP. To modify the transaminated protein with benzylalkoxyamine (BnONH2), 125 µL of a 250 mM solution of BnONH2 (pH adjusted to 5.5) was added to 625 µL of transaminated EGPh (30 µM) in pH 4.5 buffer and incubated at rt for 42 h. Excess BnONH2 was removed through ultrafiltration (30 kDa MWCO). A control was conducted following the same procedure but with non-transaminated EGPh. A portion of both samples were submitted for LC-MS analysis to determine the level of modification achieved (Figure S3). To modify transaminated protein with PEG, a stock solution of 1 mM 5 kDa alkoxyamine-poly(ethylene glycol) (PEG)24 in pH 4.5 buffer was mixed 1:1 with 50 µM transaminated EGPh and incubated at rt for 42 h. Excess PEG was removed through ultrafiltration (30 kDa MWCO). A control was conducted following the same procedure but with non-transaminated EGPh. A portion of both samples were visualized by SDS-PAGE (see Figure 2).</p><!><p>1 mL of a 40 mg/mL stock solution of 1a in pH 4.5 buffer was combined with 1 mL of 50 µM transaminated EGPh in a 4 mL glass dram vial and pipetted vigorously to mix. The solution was incubated at rt for 24 h, then 2 mL of 1.2 M D-(+)-mannose in pH 4.5 buffer was added, mixed via pipet, and the new mixture was incubated an additional 24 h at rt. The solution was transferred to a 15 mL Falcon tube and heated for 10 min in a 55 °C water bath to precipitate the polymer, then centrifuged for 5 min at 55 °C and 2,000 rpm. The supernatant was removed and the pelleted polymer was resuspended in the same volume of rt buffer (pH 4.5). This procedure was repeated three more times for a total of 4 precipitation cycles, then any excess mannose was removed through three cycles of ultrafiltration at 4 °C (10 kDa MWCO). The concentrated polymer-EGPh conjugate was transferred to an Eppendorf tube and stored at 4 °C. A small portion of the purified conjugate was buffer exchanged into pure water using ultrafiltration (30 kDa MWCO), lyophilized, and analyzed for protein concentration using tryptophan fluorescence. Attachment of 1b was performed using a similar procedure. A control with 1a was conducted following a similar procedure but with non-transaminated EGPh.</p><!><p>Unmodified, transaminated, PEGylated, and BnONH2-modified EGPh concentrations were measured using a NanoDrop 1000 spectrophotometer (Thermo Scientific), with an extinction coefficient of 139,020 M−1cm−1 and molecular weight of 49,023 Da. Protein concentration of the polymer conjugates was determined using tryptophan fluorescence. Buffered standards at pH 4.5 were prepared in triplicate containing 5 mg/mL of mannose-quenched copolymer and 7.06, 5.01, 3.0, 1.0, and 0 µM of EGPh. Triplicate 5 mg/mL samples of lyophilized polymer-EGPh conjugate were made in pH 4.5 buffer. The fluorescence spectrum of each of the standard and experimental samples was collected from 290 nm to 400 nm, with excitation at 280 nm. The maximum fluorescence intensity of each standard set was plotted versus the protein concentration and a linear fit was applied to the data points (all R2 > 0.99). This linear fit was used to calculate the protein concentration in the lyophilized samples using their fluorescence maxima. Serial dilutions of the reserved, non-lyophilized protein-polymer conjugate were prepared and their fluorescence intensities were measured to determine the dilution level that matched that of the 5 mg/mL lyophilized samples. From these data, the protein and polymer concentrations of the reserved polymer-EGPh conjugates were determined. The protein concentrations in µM EGPh/mg material were 0.520 µM/mg for 1a-EGPh, 0.177 µM/mg for 1b-EGPh, and 0.066 µM/mg for a control of 1a combined with non-transaminated EGPh.</p><!><p>All protein samples were assayed in triplicate in 1.5 mL Eppendorf tubes containing stir bars at 40 °C, using a 1% (w/v) suspension of Sigmacell cellulose powder (Sigma-Aldrich) in pH 4.5 buffer and 0.2 µM protein. Additional mannose-quenched 1a or 1b was added to the polymer-EGPh bioconjugate assays to a total concentration of 2 mg/mL polymer. To measure the reactions, each tube was shaken vigorously to ensure even distribution of the substrate and protein and a 100 µL aliquot was immediately removed and transferred to a clean, empty Eppendorf tube. This aliquot was centrifuged for 1 min at 13.2k rpm, and then the clarified supernatant was transferred to a 0.6 mL Eppendorf tube and immediately frozen in dry ice. The supernatant aliquots were stored at −20 °C until analysis for the amount of soluble reducing sugar.</p><!><p>In both recycling assays, a 100 µL aliquot was taken at t=0 and a 50 µL aliquot was removed at the end of 12 h to measure the amount of reducing sugar. The stir bars were removed and the polymer-containing tubes were heated at 55 °C for 5 min to precipitate the polymer, then centrifuged at 2k rpm for 10 min at 55 °C to pellet the polymer. All tubes were then centrifuged at 13.2k rpm for 1 min at rt to pellet the substrate. The cleared supernatant was removed and replaced with ice cold buffer, clean stir bars were added, and the tubes shaken vigorously for 3 min to ensure an even suspension of substrate and polymer. The assay procedure was then repeated twice more, beginning with removing a 100 µL aliquot to measure initial reducing sugar, for a total of two precipitation events and three 12 h assay periods.</p><p>Acid and steam-pretreated Miscanthus (Miscanthus giganteus) was obtained from the lab of Prof. Douglas Clark at UC Berkeley. The substrate had been cut into approximately 1-inch pieces and then subjected to 1.5% (w/w) sulfuric acid, 25% biomass loading (w/w), at 190 °C for approximately 1 min. It underwent a steam explosion step and then the solids were pressed to remove liquids. It was then washed extensively with deionized water until the filtrate had a neutral pH and no detectable glucose. The material was dried for 24 h at 104 °C, then ground into a fine powder with a mortar and pestle.</p><!><p>This procedure was performed following a previously reported method, using the glucose oxidase-peroxidase assay with OxiRed as the substrate. 22 Analysis was performed in clear-bottom plastic 96 well plates, with each sample analyzed in triplicate. Internal standards of 300, 200, 100, 50, 25, and 0 µM glucose, and 150, 100, 50, 25, and 12.5 µM cellobiose in pH 4.5 buffer were included in each plate. Frozen aliquots from the activity assays were thawed on ice and then diluted 0- to 20-fold with cold buffer, then 8 µL of the solution was incubated with 8 µL of β-glucosidase (5 mg/mL in 10 mM NaOAc pH 4.6) for 60 min at 37 °C to convert all of the cellobiose to glucose. The amount of glucose present was then measured by adding 65 µL of glucose oxidase (1.25 U/mL), horseradish peroxidase (1.25 U/mL), and OxiRed (60 µM) in 125 mM phosphate buffer (pH 7.45) and incubating at rt for 10 min in the dark. The amount of Resorufin formed was measured on an optical plate reader with excitation at 535 nm and emission detection at 590 nm. The amount of Resorufin formed corresponded to the amount of glucose present. Linear standard curves were made from the internal standards in each plate (all r2 > 0.97), which were then used to calculate the amount of glucose equivalents present in each activity assay sample. The triplicate measurements of each supernatant sample were averaged, and then the measurements of the triplicate activity assay samples were averaged to calculate each data point.</p><!><p>The key synthetic requirement for these studies is the site-selective attachment of LCST polymers to cellulase enzymes. Although the attachment of polymer chains is a common bioconjugation practice, there are relatively few strategies for doing so that are site-specific. In previous work, we have reported the attachment of PEG chains to ketone and aldehyde groups on protein surfaces through oxime formation.24 This method provides a hydrolytically stable linkage and can be formed under mild pH and temperature conditions. We have also used this strategy to incorporate protein crosslinks into methacryl hydrogels,20,21 and other labs have used it to immobilize proteins on polymer films.25 To introduce multiple copies of the requisite aminooxy handles into the temperature-responsive polymer, free-radical copolymerizations were performed with a Boc-aminooxy methacrylamide and N-isopropylacrylamide (NIPAm) (yielding 1a) or N-isopropylmethacrylamide (NIPMa) (yielding 1b) using AIBN as a radical initiator (Figure 2a). A 1:9 starting molar ratio of aminooxy monomer:NIPAm or NIPMa was used in methanol at 60 °C for 6 h. Following ether precipitation, the copolymers were deprotected with trifluoroacetic acid in dichloromethane. Characterization by 1H-NMR spectroscopy showed an aminooxy monomer incorporation of 8.6% for 1a and 8.8% for 1b (Supporting Information Figures S1 and S2). Polymer size was determined by size exclusion chromatography (SEC) using poly(methyl methacrylate) standards, with number-average molecular weights (Mn) of 85,127 and 11,144 Da and polydispersity indices (PDI) of 1.82 and 1.26 for 1a and 1b, respectively. Polymer that was subjected to thermal precipitation, isolated by filtration, and redissolved provided identical 1H-NMR spectra, suggesting that the aminooxy monomers were incorporated throughout the material.</p><p>A well-studied phenomenon of LCST polymers is the ability to adjust the thermoprecipitation point by changing the hydrophilicity of the material.26,27 To determine the LCST of 1a and 1b, 1 mg/mL polymer samples in sodium acetate buffer (50 mM, pH 4.5) were warmed at a rate of 0.5 °C/min while stirring, and the absorbance at 600 nm was measured every 0.5 °C. Incorporation of the hydrophilic aminooxy monomomer increased the LCST of the pNIP-Am copolymer to 42.5 °C from 32 °C for the homopolymer, and the LCST of the pNIPMa copolymer increased to 58.1 °C from 43 °C.28</p><p>In addition to providing a potential method for protein attachment, the aminooxy group also provides a handle through which the LCST can be adjusted through small-molecule quenching (Figure 2). This allows a wide range of LCSTs to be accessed starting from a common supply of copolymer. Quenching also prevents the aminooxy functional groups from reacting with adventitious aldehydes, such as those of the glucose molecules produced during cellulose depolymerization. A 10 mg/mL solution of 1a or 1b was reacted with six different small molecules (60 mM or 600 mM) for 24 h, then purified by ultrafiltration and lyophilized. Using this strategy and starting from only two copolymers, materials were obtained with LCSTs ranging from 20.9 °C to 60.5 °C (Figure 2c). From these possibilities, we decided to use mannose to quench the protein-polymer bioconjugates because of its economical cost, low likelihood to affect enzyme activity, and compatibility with our method of glucose quantification in further experiments.</p><p>Following the construction of LCST copolymers with tunable precipitation temperatures, we next focused on the modification of the enzyme. We chose to use a hyperthermophilic endoglucanase from the deep-sea archaeon Pyrococcus horikoshii (EGPh). This family 5 cellulase was discovered in 2002, and its ability to hydrolyze a variety of cellulosic substrates and stability at temperatures above 97 °C makes it a promising candidate for industrial applications.22,29 Our method of protein modification involved a previously reported site-selective transamination, in which pyridoxal 5'-phosphate (PLP) is used to oxidize the N-terminus of proteins to yield a ketone group (Figure 3a). In a subsequent step, the N-terminus can be modified selectively through oxime formation with aminooxy-functionalized small molecules or the aminooxy-substituted LCST polymers described above.30,31</p><p>We have previously identified that installing an alaninelysine motif at the N-terminus of proteins leads to optimal transamination levels.23 An ala-lys-thr sequence was inserted at the N-terminus of EGPh and the new construct was expressed in Escherichia coli. Using a C-terminal His6 tag for purification with Ni-NTA chromatography, an average yield of 85 mg of purified protein per L of expression media was obtained (Figure 3b, lane 1). The N-terminus of the protein was then site-selectively transaminated by reacting with PLP (100 mM) for one hour. To establish the level of transamination and confirm the site-selectivity of the reaction, the transaminated protein was incubated with benzylalkoxyamine (42 mM, pH 4.5) or 5k-aminooxy-poly(ethylene glycol) (PEG)24 (500 µM, pH 4.5) for 42 h, followed by ultrafiltration. Samples were analyzed by LC-MS to determine a small-molecule modification yield of 84%, and SDS-PAGE followed by Coomassie staining and densitometry showed modification with a single PEG chain in 53% yield (Figure 3b and Supporting Information Figure S3).</p><p>Next, assays were performed to ascertain whether the enzyme modification site or the conditions used negatively affected its catalytic activity. Each sample was evaluated in triplicate, using 0.2 µM EGPh and an insoluble cellulose substrate (Sigmacell, 1% w/v). Both benzylalkoxyamine and 5 kDa-PEG-alkoxyamine modified EGPh were assayed, along with two controls in which the protein was not transaminated but was still incubated with the alkoxyamines for 42 h. These controls showed no modification by LC-MS and SDS-PAGE analysis (Figure 3b, lane 2). In addition, an unmodified, non-transaminated control was included, as well as a sample of EGPh that had been incubated with mannose (600 mM) for 24 h to ensure protein activity would not be affected by the quenching step used for the LCST polymers. Each protein reaction was assayed in a 40 °C water bath and 100 µL aliquots were removed at 2, 4, 6, 8, and 12 h. The supernatants were analyzed for the amount of soluble sugar released using the glucose oxidase-peroxidase assay with OxiRed as the fluorescent substrate.22</p><p>As shown in Figure 3c, the activities of the two modified endoglucanases were slightly reduced relative to the unmodified control, but the differences were generally within the standard deviations of the assays. The activities of unmodified enzyme samples that had been exposed to benzylalkyoxyamine, PEG-alkoxyamine, and mannose were also slightly lower, but within one standard deviation of the activity of the unmodified control at each time point. From these experiments, we concluded that neither the N-terminal modification itself nor the reaction conditions used had a significant effect on endocellulase activity.</p><p>To generate the thermoresponsive material, ketone-bearing EGPh was covalently attached to the aminooxy-substituted LCST polymers. Transaminated EGPh was combined with polymer 1a or 1b and allowed to react at room temperature for 24 h. Any remaining aminooxy groups on the polymers were then capped by the addition of mannose (600 mM), followed by further incubation at room temperature for 24 h. To remove any unmodified protein and free mannose, the mixture was heated to 7 °C above the LCST, centrifuged at that temperature to pellet the precipitated polymer bioconjugate, and the supernatant was removed and replaced. This was repeated for a total of 4 cycles. To determine the amount of protein that could be non-covalently adsorbed onto the polymer, this same procedure was also performed using non-transaminated EGPh and polymer 1a. A small portion of each material was then lyophilized to measure protein attachment.</p><p>The protein-polymer ratio was determined by measuring the tryptophan fluorescence of the weighed lyophilized sample compared to standards that contained mannose-quenched 1a or 1b and EGPh. This analysis indicated protein:polymer ratios of 25.5 mg/g for the NIPAm bioconjugate (corresponding to 0.044 proteins/polymer chain) and 8.7 mg/g for the NIPMa bioconjugate (corresponding to 0.002 proteins/polymer chain). The NIPAm copolymer+unmodified EGPh control had a ratio of 3.2 mg/g. For enzyme activity experiments, serial dilutions of the reserved, non-lyophilized protein-polymer conjugates were analyzed to find the dilution levels with fluorescence values that matched those of the lyophilized samples.</p><p>To assess any change in activity as a result of polymer attachment, an activity assay was performed using the unmodified, non-transaminated EGPh, the 1a-EGPh conjugate, and the 1b-EGPh conjugate. Additional mannose-quenched 1a or 1b was added to the 1a-EGPh and 1b-EGPh assays to bring the total polymer concentrations up to 2 mg/mL, so the conditions would be comparable to subsequent recycling experiments in which additional polymer was added to enhance bioconjugate aggregation. It was observed that both 1a-EGPh and 1b-EGPh displayed about half the endoglucanase activity of the free EGPh after 2 h, but at later time points the enzymatic activity in the samples converged (Figure 4a). After 12 h, any initial differences in activity had subsided and the differences in the total concentration of reducing sugars was statistically insignificant. Other studies of endogluconases have observed similar decreases in activity over time. The mechanisms for this inactivation are complex, and could involve product inhibition, protein adsorption on the substrate, and denaturation. 32–34 In the case of EGPh, we have found that product inhibition by cellobiose is unlikely to be the cause of this activity loss, as added beta-glucosidase does not lead to increases in the overall activity of the unmodified enzyme (as measured at 12 h, Supporting Information Figure S4). Regardless of the cause, this deactivation places an upper boundary on the total amount of product that can be obtained, allowing the polymer-enzyme conjugates to reach similar conversion levels.</p><p>The activity of the polymer bioconjugates was equivalent to free protein after 12 h, but if the materials could be precipitated, collected, and reused, the total amount of reducing sugar produced over the lifetime of the material would potentially be much greater than that possible using only free enzyme. To test the recycling potential of the bioconjugates, a free unmodified EGPh control and the 1a-EGPh bioconjugate were prepared as described for the previous assay and allowed to react for 12 h at 40 °C. The bioconjugate was then heated to precipitate the protein-polymer material and centrifuged to pellet the aggregated polymer and the cellulosic substrate. The EGPh control was also centrifuged. The supernatant of all samples was removed and replaced with an equal volume of fresh, ice-cold buffer to ensure rapid bioconjugate resolubilization, and the mixtures were again allowed to react for 12 h at 40 °C. This was repeated once more, for a total of 2 precipitations and 3 cycles. Aliquots of each reaction were removed at the beginning and end of each cycle to measure the concentration of additional reducing sugar produced during each 12 h.</p><p>As shown in Figure 4b, after the initial 12 h of reaction, the 1a-EGPh bioconjugate produced reducing sugars at 86% of the level of the control. However, it retained 68% and 63% of its initial activity over two cycles of precipitation and recovery of the material. The free enzyme, in contrast, retained only 11% and 4% of its initial activity. While the supernatant was removed from all samples, only the polymer bioconjugate sample had the majority of the protein precipitated in the substrate pellet. Any subsequent activity for the free control was due to enzyme carryover contained in the pelleted insoluble cellulose. Over three cycles, the polymer-protein bioconjugate was able to produce a 1.7-fold increase in the amount of free reducing sugar over that produced by the control (Figure 4c).</p><p>While it is clear the polymer can be recycled, we also wanted to compare its activity over three 12 h reaction cycles to the activity of free protein over one 36 h reaction to determine if recycling the enzyme actually provided any benefit over simply letting free enzyme react for an extended period of time. Free, unmodified EGPh was allowed to react with Sigmacell for 36 h at 40 °C, with free reducing sugar measured at 0 h, 12 h, 24 h, and 36 h. Consistent with the results above, the free enzyme reached an upper activity limit by 12 h, with only modest increases in reducing sugar achieved in the next 24 h (Figure 4d). At 12 h, the polymer-protein bioconjugate was 91% as active as the free control; by recycling the enzyme, however, 78% more reducing sugars were produced over 36 h than if the free enzyme was allowed to react.</p><p>We next decided to characterize the bioconjugate using a more realistic substrate that included hemicellulose and lignin. We chose the perennial grass Miscanthus (Miscanthus giganteus), as this plant has been extensively studied for its potential as a large scale feedstock for biofuels production.36 Acid and steam-pretreated Miscanthus was washed, dried, and ground into a powder. The activity and recycling ability of the 1a-EGPh bioconjugate was then assessed following the same procedure that was used for Sigmacell. As seen in Figure 5, the activity of the unmodified enzyme control was significantly lower for Miscanthus that it was for Sigmacell. This drop is not unexpected, as Miscanthus has much greater chemical and structural heterogeneity than substrates consisting of isolated cellulose. In addition, protein adsorption onto lignin and subsequent deactivation is a known challenge with unrefined biomass. 22,34,35 The activity of the bioconjugate also decreased for this substrate, but to a lower extent than the control. At the same enzyme loading, the bioconjugate produced substantially more reducing sugars after the first cycle. This activity difference between the bioconjugate and the control on Miscanthus is similar to that seen upon addition of certain surfactants and polymers to the enzymatic hydrolysis of lignocellulic biomass. 34,35 Specifically, the addition of surfactants such as Tween and Tiron, or polymers such as PEG, has been shown to increase saccharification of lignocellulosic biomass (albeit only by increases of <20% in most cases).34,36 The mechanism behind this increased activity has not been completely elucidated, but hypotheses include interactions between the additives and lignin that prevent enzyme adsorption onto the hydrophobic surface or enhance its subsequent desorption. Either mechanism would increase the concentration of active enzyme in solution. It appears that the polymer also shows this effect.</p><p>In subsequent cycles of reuse, there was a drop in activity of the bionconjugate, presumably because some protein is still adsorbed onto the biomass and ultimately deactivated. However, the bioconjugate remained 104% and 83% as active as the initial control in cycles 2 and 3. The ability to recover the bioconjugate, combined with the added activity due to the surfactant effect, increased the amount of reducing sugars by 2.8-fold over three rounds. We are currently investigating the mechanism through which the activity enhancement occurs, and we are further optimizing the polymer component to minimize adsorption further.</p><!><p>Through these studies, highly adaptable thermoresponsive polymer-protein bioconjugates have been developed through the copolymerization of NIPAm or NIPMa with an alkoxyamine-bearing methacrylamide. The two copolymers exhibited LCSTs of 42.5 °C and 58.1 °C, respectively, but small-molecule quenching of the alkoxyamine pendant groups of those two copolymers allowed materials with LCSTs ranging from 20.9 °C to 60.5 °C to be obtained. To allow polymer attachment, the hyperthermophilic endoglucanase from Pyrococcus horikoshii (EGPh) was site-selectively transaminated using pyridoxal 5'-phosphate. Compared to free enzyme, protein modified with the small molecule benzylalkoxyamine or the polymer 5 kDa-PEG-alkoxyamine showed insignificant decreases in producing soluble reducing sugars from hydrolysis of an insoluble cellulosic substrate after 12 h. Protein modified with either of the two NIPAm or NIPMa copolymers exhibited a decrease in activity initially, but the levels of soluble reducing sugars were comparable to those of the unmodified control protein after 12 h. The NIPAm copolymer-protein conjugate retained over 60% of its initial activity following two cycles of thermal recovery, and a total of 1.7-fold more soluble reducing sugars were produced over three cycles compared to the unmodified control. When applied to a sample of Miscanthus, the enzyme bioconjugate provided both an increase in overall activity and the capability of recycling to result in a 2.8-fold increase in depolymerized product.</p><p>In this way, a method for the recovery and reuse of cellulase enzymes was achieved that could be applied to the wide range of cellulases currently being studied for industrial applications, and could be adjusted to be compatible with the different temperatures at which these enzymes are used. It is difficult to obtain cost estimates of cellulase enzymes on industrial scale, as research in biomass depolymerization is rapidly developing. 2 However, we currently estimate that both the enzymes and the LCST polymers can be obtained on bulk scale for $50/kg or less. When the advantages of recycling are taken into account, these figures suggest that this concept could be economically feasible if a 1:1 ratio of protein to polymer can be used. While we consider the bioconjugation methods used herein to be practical and inexpensive, our current bioconjugate loadings are admittedly far below this number due to the high concentrations of functional groups that are required for oxime formation. Unfortunately, there are currently few alternative strategies for the site-specific, stoichiometric, and scalable attachment of polymers to proteins. Thus, while these results serve to validate the recycling concept, they also underscore the need for new, highly efficient bioconjugation reactions that can be carried out on process scale. The development of such methods serves as a major objective in our research, as they will be of key importance to increasing the practicality of these and other bioconjugates for materials applications.</p><!><p> ASSOCIATED CONTENT </p><p>Additional procedures and supporting figures. This material is available free of charge via the Internet at http://pubs.acs.org.</p><p>The authors declare no competing financial interest.</p>

PubMed Author Manuscript

Exploration of the Synergy Between 2D Nanosheets and a Non-2D Filler in Mixed Matrix Membranes for Gas Separation

Dual-filler MMMs have attracted special interests in recent years because of the possibility of producing synergetic effect. This study is aimed at exploring the underlying synergy between two-dimensional (2D) nanosheets and a non-2D filler in mixed matrix membranes for gas separation. MXene or graphene oxide (GO) as typical nanosheet filler is selected to be in pair with a non-2D filler, SiO2 or halloysite nanotubes (HNTs), with Pebax as the polymer matrix. In this way, four pairs of binary fillers are designed and the corresponding four groups of MMMs are fabricated. By tuning the mass ratio of binary fillers, synergetic effect is found for each group of MMMs. However, the two 2D fillers found different preferential non-2D partners. GO works better with HNTs than SiO2, while MXene prefers SiO2 to HNTs. To be noted, GO/HNTs renders the membranes the maximum enhancement of CO2 permeability (153%) and CO2/N2 selectivity (72%) compared to Pebax control membrane, while each of them as single filler only brought about very limited enhancement of CO2 separation performance. The possible mechanisms are thoroughly discussed in terms of filler dispersion, nanosheet flexibility, and the tortuosity and connectivity of the surface diffusion pathways along nanosheets.

exploration_of_the_synergy_between_2d_nanosheets_and_a_non-2d_filler_in_mixed_matrix_membranes_for_g

Introduction<!><!>Materials<!>Synthesis of SiO2<!>Modification of HNTs<!>Synthesis of GO<!>Synthesis of MXene<!>Membrane Preparation<!>Characterization of Fillers and Membranes<!>Gas Permeation Experiments<!>The Physical Structure of SiO2, HNTs, GO, and MXene<!><!>The Chemical Structures of SiO2, HNTs, GO, and MXene<!><!>Filler Dispersion in Membranes<!><!>Filler Dispersion in Membranes<!>Gas Transport Properties of Membranes<!><!>Gas Transport Properties of Membranes<!><!>Gas Transport Properties of Membranes<!>Conclusions<!>Data Availability Statement<!>Author Contributions<!>Conflict of Interest

<p>Mixed matrix membranes (MMMs), containing a continuous polymer phase and a dispersed inorganic filler phase was introduced by Kulprathipanja in 1980s (Kulprathipanja et al., 1988). The investigation of MMMs has been increasingly focused on solving the permeability-selectivity tradeoff of original polymer membranes since it aims to combine the advantages of inorganic materials with superior gas transport and good mechanical property, but also those of the polymer with the economic applicability and good machining performance (Li et al., 2013b; Rezakazemi et al., 2014; Vinoba et al., 2017; Wang et al., 2019b). Although MMMs have been introduced for many years, there is still plenty of room for development, because of its unique "4 M" characteristics including multiple interactions, multiscale structures, multiphase, and multiple functionalities (Li et al., 2013b), which have revealed infinite possibilities in designing and tuning the structure of membranes.</p><p>Currently, the research on MMMs mainly centers on the development and use of materials, the exploration of the membrane fabrication method and the study of the theoretical model for predicting the gas separation performance of MMMs (Vinh-Thang and Kaliaguine, 2013). Among these topics, the innovation of the materials and fabrication method mainly depends on the advent of new fillers with high selectivity, their distribution, and adhesion to the polymer matrix. The reported strategies in filler development can be categorized into the following four types: (i) Exploration of new fillers. During the past 20 years, various types of filler materials such as zeolites, metal organic frameworks (MOFs), covalent organic frameworks (COFs), SiO2, carbon nanotubes, graphene, etc., have been developed (Zornoza et al., 2011; Xin et al., 2015; Kim et al., 2016a; Vinoba et al., 2017; Idris et al., 2019). Besides the composition of the filler materials, their shape/morphology is important. In recent years, two-dimensional (2D) nanosheets such as graphene, graphene oxide (GO), MXene, molybdenum disulfide (MoS2), and graphitic-phase carbon nitride (g-C3N4) have been gaining increasing attentions due to the capacity of forming long-range tortuous channels in membrane, which hinders the diffusion of larger molecules but permits the transport of smaller ones (Smith and Freeman, 2014; Dong et al., 2016b; Zhang et al., 2019). (ii) Chemical functionalization of existing fillers. It is an extensively applicable strategy to overcome the poor interfacial compatibility between polymer and filler (Zhang et al., 2019), or to directly impart more efficient transport mechanisms—such as surface diffusion and facilitated transport—to membranes. (iii) Creation of nanoscale morphologies on the filler surface. Different from Chemical functionalization, this strategy was proposed to enhance the interfacial adhesion at the nanometer scale rather than molecular level, which is expected to reduce the possibility of interfacial rigidification. This strategy has proved valid for zeolites and other silicate fillers (Shu et al., 2007a,b; Bae et al., 2009), and it needs more attention when other molecular sieves are used as fillers. (iv) Integration of dual fillers. This strategy is usually easy to operate. The interaction between dual fillers and the matrix might improve their dispersion, providing different functional domains within a membrane. They might also provide a unique way to control the morphology of permeation channels (Wang et al., 2019a). As such, a synergy is likely to occur between dual fillers, and hence significantly improve the membrane performance.</p><p>Early studies involving dual fillers combined MOF (HKUST-1) and zeolite. The author demonstrated that the different natures of fillers could improve the dispersion and further increase the membrane performance due to a synergetic effect (Zornoza et al., 2011). In recent years, more and more dual-filler MMMs have emerged with fascinating phenomena (Tang et al., 2008; Zornoza et al., 2011; Hu et al., 2012; Galve et al., 2013; Valero et al., 2014; Li et al., 2015b, 2018; Ahmad et al., 2017; Jamil et al., 2019; Wong et al., 2019). Cornas' group has carried out a series of experiments to investigate the synergetic effect of two fillers with different natures (Galve et al., 2013; Valero et al., 2014). The silica-based MCM-41 as well as NH2-MIL-51 MOF was integrated into polysulfone or polyimide matrix. The resulting MMMs possessed enhanced permeability due to the mesoporosity of MCM-41, while the enhanced gas selectivity was originates from the microporosity and flexibility of MOF (Valero et al., 2014). Besides, the dispersion of MCM-41 was found significantly improved when MCM-41 was in combination with 2D JDF-L1 sheets, which was ascribed to the strong steric effect of JDF-L1 sheets. Meanwhile, the enhanced gas selectivity was interpreted by the preferential horizontal orientation of JDF-L1 sheets which hinder the gas transport of large gas molecules (Galve et al., 2013). Wu et al. incorporated carbon nanotubes (CNTs) and GO blesheets into Matrimid® matrix. The high aspect ratios and smooth walls of CNTs were thought to furnish fast gas permeation pathways, and the GO sheets were perceived as a selective barrier on account of the horizontal orientation and functional groups on the surface of GO. As a result, the MMMs exhibited super characteristics of both CNTs and GO in gas separation (Li et al., 2015b). As mentioned above, dual-filler MMMs can not only combine their advantages, but even lead to a synergy to acquire non-linear effects. Nevertheless, now only a handful of such studies can be found in the literature, demonstrating that dual-filler MMMs are still in the initial stage of development and the origin of synergetic effect needs further exploration.</p><p>In this study, we explored the synergy between 2D nanosheets and a second filler (1D or 3D) in MMMs. The 2D sheets were selected as the major filler, because they have distinct advantages in paving selective molecular pathways, but suffer from agglomeration and extra transport resistance (Zhang et al., 2019). GO and MXene were chosen as two representative 2D fillers for comparison with known differences of rigidity and surface functional groups (Jeon et al., 2016; Wang et al., 2018). SiO2 (3D) and HNTs (1D) were selected as the second fillers because of the highly controllable morphology and availability. GO/MXene was paired with SiO2/HNTs, resulting in four systems of dual fillers (Figure 1). The different matches were compared in terms of structures and gas transport properties, with the purpose of revealing part of the rules that could guide future work in this field.</p><!><p>Scheme of the fabrication of dual-filler mixed matrix membranes.</p><!><p>Commercial Pebax®1657 (consisting of 60 wt% PEO and 40 wt% PA6) was purchased from Arkema Inc. A mixture of ethanol and water (70/30 wt%) was used as a solvent for Pebax®1657. Ammonia water (NH3·H2O, 25%) and tetraethyl orthosilicate (Si(OEt)4, TEOS, 98%) were supplied by Feng chuan Chemistry Co., Ltd. (Tianjin, China) and Aldrich, respectively. Pristine HNTs were provided by Henan Xianghu Environmental Protection Technology Co., Ltd. (Henan province, China). Polystyrene sulfonate sodium salt (PSS, MW = 70,000) was provided by Sigma-Aldrich. Hydrofluoric acid (HF) and Ti3AlC2 powder were purchased from Sigma-Aldrich. Ethanol, hydrogen peroxide, hydrochloric acid, phosphoric acid, dimethyl sulfoxide (DMSO), sulfuric acid, and KMnO4 were provided by Kewei Chemistry Co., Ltd. (Tianjin, China). Deionized water was used throughout the experiment. The polymer and other chemicals were used as received without further treatment.</p><!><p>Silica sub-microspheres were prepared according to the classical Stöber method (Chen et al., 2016): 2 mL of TEOS was added to the mixture of 200 mL ethanol, 20 mL deionized water and 15 mL aqueous solution of 25% ammonium with vigorous stirring at room temperature and the reaction was continued further for 24 h with stirring. The resultant silica particles were purified by three cycles of centrifugation, decantation, and resuspension in ethanol with ultrasonic-bathing. The silica particles were dried in a vacuum oven at 60°C till constant weight.</p><!><p>Before modification, the pretreatment for pristine halloysite was required to obtain HNTs with uniform size. The pristine halloysite was mashed mechanically and soaked in deionized water for 2–3 days. Then the obtained slurry was filtered and dried in 50°C. Afterwards, the powder was grinded using mortar, and the HNTs were obtained after filtering through a 300-mesh sieve.</p><p>The HNTs was modified with PSS to improve their dispersion in Pebax®1657 (Qin et al., 2016; Zhang et al., 2018) 2 g PSS was dispersed in 100 mL deionized water in a flask, followed by 30 min agitation to form homogenous suspension. The resulting HNTs (2 g) were gradually added under continuous magnetic stirring for 48 h at ambient temperature and then left standing for 30 min to precipitate aggregates. The supernatant dispersion was collected and centrifuged at 5,000 rpm for 10 min and washed 3–4 times with deionized water until it became neutral. Finally, the obtained solid (PSS-HNTs) was dried in vacuum drier for 24 h and then ground into powder for use.</p><!><p>The graphene oxide was synthesized through the improved Hummer's method as the literature reported (Zhang et al., 2019). Firstly, the suspended mixture solution of concentrated H2SO4/H3PO4 (540 mL/90 mL) was prepared in a 1,000 mL three-necked bottle, and then 4.5 g graphite powder and 27 g potassium permanganate were added into the mixture solution and stirred under 50°C for 24 h. The unreacted permanganate and manganese dioxide was transferred into soluble sulfates with 1,200 mL ice solution containing 10 mL H2O2 solution (30 wt%). The resulting suspension was re-dispersed by ultrasonic treatment and then centrifuged to separate the sediment, which was washed with the mixture solution of HCl/H2O (400 mL, 150 mL). The obtained suspension was stirred for 12 h washed with water until neutral, and was finally washed with ethanol followed by drying in the vacuum oven for 24 h.</p><!><p>MXene was synthesized following a previously reported method (Jeon et al., 2016). The Ti3AlC2 powder was etched with 49% HF aqueous solution under 60°C for 72 h to obtain the Ti3C2Tx sheets, which was added into DMSO solution for 48 h stirring to enable intercalation. With the purpose of exfoliation, plenty of water was added into the as-prepared solution and then centrifuged to separate the sediment. Finally, the obtained sediment was re-dispersed into water with a weight ratio of 1:500, followed by ultrasonic treatment and centrifugation to obtain the supernatant of MXene.</p><!><p>MMMs were prepared by a physical blending method. Firstly, a certain amount of Pebax®1657 was dissolved in ethanol/water mixture (70/30 wt%) with reflux under mild mechanical stirring at 80°C for 2 h to obtain 3 wt% homogeneous solution and cooled the solution to ambient temperature. Secondly, a certain amount of filler was fully dissolved into deionized water and added to the previously prepared polymer solution. After 30 min ultra-sonication treatment and 12 h stirring, the mixed homogeneous casting solution was poured onto Teflon Petri dishes. Eventually, the membranes were obtained after removing the residual solvent by drying at ambient temperature for 24 h. The thickness of membrane is in the range of 80–100 μm. The membranes were denoted as Pebax-A-X (A: GO or MXene) or Pebax-A/B-X/Y (B: SiO2 or HNTs), where X (0, 0.2, 0.5, 0.8, 1, 2.5, 5) denotes the wt% of filler A to matrix, and Y denotes the wt% of filler B to matrix.</p><!><p>The morphology was investigated by scanning electron microscopy (SEM) on Zeiss/Auriga FIB equipment, the membrane was broken in liquid nitrogen atmosphere, and all of the samples were cover with gold before observation. Besides, the transmission electron microscopy (TEM) was also used to investigate the morphology of membrane on a FEI TalosTM F200S microscope. The chemical analysis was performed by Fourier Transform Infrared (FTIR) spectroscopy on a FTLA 2000 spectrometer in the 4,000–400 cm−1 scan range with resolution of 1.93 cm−1. The positron annihilation spectroscopy (PALS) analysis, which used 50 mCi of 22Na as the positron source, was used to measure the free volume of membranes. A GORTEC fast-fast coincidence system was used with the resolution of 201 ps.</p><!><p>Membrane transport properties were measured by time-lag method. All the measurements were conducted under 2 bar and 30°C. The permeability coefficient P [Barrer, 1 Barrer = 10−10 cm3 (STP) cm cm−2 s−1 cmHg−1], diffusivity coefficient D, and solubility coefficient S of gas "i" were calculated by the following equations:</p><p>where Vp represents the constant permeate volume, l represents for the membrane thickness, Δt is the time during which the permeate pressure increases from pp1 to pp2, A is the effective membrane area, and θ is the so-called time lag. The error of gas measurement is <10% for gas permeability and 20% for diffusivity coefficient, respectively.</p><!><p>The physical morphology of the fillers was characterized by TEM, as shown in Figure 2. It's noticeable that HNTs has the inherent structure of hollow tubular and end-open structure. The tube length is 450–950 nm, the inner and external diameter is 20 and 44 nm, which is consistent with the literature (Qin et al., 2016; Liu et al., 2018). Figure 2B shows the relatively uniform and narrow particle size distribution with an average size of 200 nm for SiO2 microspheres. Both GO and MXene have the inherent sheet morphology of 2D-materials. Besides, it is obvious that GO and MXene are very thin. The lateral dimension and thickness of GO are 500–1,000 nm and 1.5–2.0 nm, respectively; for MXene, 1–2 μm and 1–2 nm. The TEM of GO sheets show wrinkles and folding edges, which are common in flexible GO sheets. On the contrary, the TEM of MXene sheets shows almost no wrinkles, which is expected in rigid MXene sheets (Shen et al., 2015).</p><!><p>TEM images of (A) m-HNTs, (B) SiO2, (C) GO, and (D) MXene.</p><!><p>The chemical characteristics of the fillers were record by FTIR spectrum (Figure 3). In Figure 3A, the characteristic bands at 3,696 and 3,619 cm−1 correspond to the stretching vibration of hydroxyl groups of HNTs. The strong band at 1,021 cm−1 is assigned to the asymmetric flexible vibration of Si–O bond arising from the abundant O–Si–O groups in HNTs. Compared to the FTIR spectrum of pristine HNTs, the new and strengthened peaks at 1,228 and 594 cm−1 can be assigned to the asymmetric and symmetric vibration of S=O groups of –SO3Na (Qin et al., 2016). Figure 3B shows the FTIR spectrum of the SiO2 particles. The characteristic bands at 1,100 and 950 cm−1 are ascribed to the asymmetric stretching vibration of Si–O–Si and the stretching vibration of Si–OH, respectively (Li et al., 2014; Shi et al., 2019). Besides, the peaks at 1,640 and 3,400 cm−1 are the bending and stretching vibrations of water molecules bond to –OH groups of SiO2 (Kim et al., 2016b). In the spectrum of GO (Figure 3C), the band at 1,628 cm−1 corresponds to the sp2-hybrid carbon atoms. The bands at 1,221, 1,720, and 3,376 cm−1 are assigned to the C–O, C=O and C–OH, respectively, indicating that there are abundant oxygen-containing groups on the surface of GO (Quan et al., 2017). The MXene spectrum (Figure 3D) shows bands at 1,640 and 3,430 cm−1, which are assigned to the carbonyl group at the edge of MXene sheet and –OH stretching vibration, respectively (Gong et al., 2018).</p><!><p>FTIR spectrum of (A) HNTs, (B) SiO2, (C) GO, and (D) MXene.</p><!><p>The cross-section morphology of single filler MMMs, and dual-fillers Pebax-GO/HNTs and Pebax-MXene/SiO2 membranes, are shown in Figure 4 (with 1 and 5 wt% total filler content). It's noticeable that the dispersion of the high concentration (5 wt%) of 2D materials is not homogeneous (see aggregation in Figures 4A,D; Figure S1). Also, high concentration of HNT in Pebax suffers from severe agglomeration (Figure 4B) due to the high surface area. By comparison, the filler dispersion in dual-filler MMMs appears much better than each single-filler one. TEM image (Figure S2) confirms the good dispersion of both MXene and SiO2 in dual-filler membranes, especially at low total filler content (1 wt%). Meanwhile, compared to Pebax-GO membrane, the interfacial boundary between MXene and Pebax is more obscure, revealing that the dispersion of MXene might be more homogeneous than that of GO in membrane. This may results from the differences in functional groups of GO and MXene. MXene possesses higher density of functional groups (O, OH, and/or F) with more evenly distribution, which benefits an effective interaction between MXene and the Pebax matrix (Jeon et al., 2016). As for the cross-section image of Pebax-SiO2 membrane, the SiO2 achieves excellent dispersion in matrix at both 1 and 5 wt% content because of the small and uniform size, as illustrated in Figure 4E. However, when it comes to Pebax-HNTs membrane (Figure 4B), the HNTs exhibit inferior dispersion than SiO2, especially at 5% content, as a result of the extremely high aspect ratio and the strong van de Waals forces between HNTs (Wong et al., 2019). The dispersion is however better than for GO, probably due to some favorable interaction with the polyamide block of Pebax, as claimed before (Zhang et al., 2018).</p><!><p>Cross-section SEM images of membranes filled with (A) GO, (B) HNTs, (C) GO+HNTs, (D) MXene, (E) SiO2, (F) MXene+SiO2. Circled area with arrows indicates aggregates.</p><!><p>Compared to Pebax-HNTs and Pebax-GO membrane, the Pebax-GO/HNTs membrane has an obscure HNT-polymer interface (Figure S3), indicating that there is synergetic effect between the mixed GO and HNTs that improve the dispersion of HNTs. Analogously to GO composites with carbon nanotubes, the flexible GO sheets could encase HNTs to facilitate the dispersion (Tian et al., 2010; Hu et al., 2012; Li et al., 2015b). On the other hand, the incorporation of HNTs would prevent the restack of GO sheet, which improves the dispersion of GO itself in matrix (Li et al., 2015b). Furthermore, there is a strong interaction between the surface functional group of GO and HNTs modified with sulfate. In addition to this, the membrane of Pebax-MXene/SiO2 also exhibits better interface morphology than Pebax-MXene membrane, revealing that the addition of SiO2 improves the dispersion of MXene sheets in the matrix. At the same time, the dispersion of SiO2 is not destroyed in dual-filler MMMs. Besides, there is also a strong hydrogen bond interaction between the hydroxyl groups of MXene and SiO2 and this further improves the filler dispersion (Hu et al., 2012).</p><!><p>The gas separation performance of dual-filler MMMs based on GO and MXene was tested and shown in Figures 5, 6. CO2 permeability of 106 Barrer and CO2/N2 selectivity of 41 were measured for the pristine Pebax membrane. The effects of both overall loading (1 and 5 wt%) of fillers and the relative content of 2D fillers can be clearly seen (permeation data of membranes based on other overall loading can be found in Figure S4, which shows 1 wt% is the optimal overall loading). From a general view, we find a maximum in performance in different dual-filler MMMs, demonstrating the occurrence of synergetic effect. It is also notable that 1 wt% is usually the better overall filler loading than 5%, which accords well with the reported single-filler Pebax-based MMMs in the literature (Li et al., 2013a; Dong et al., 2016a). This also can be understood by considering the morphological observation in Figure 4. When a higher concentration of filler is added, aggregation occurs and the expected improvement is partially lost. Only a homogeneous distribution can lead to a significant performance enhancement. Since Pebax belongs to a solubility-controlled class of polymer, the major factor that determines the separation performance is the content and distribution of PEO domains, which preferentially interacts with CO2. Relatively small content of filler could disturb the PEO crystallization and increase the content of the amorphous PEO phase segments, making them more available to interact with CO2 (Yave et al., 2010), while excessively increasing the filler content would reduce the PEO content of the whole membrane with an undesirable aggregation, which does not effectively affects the crystallinity (Table S2). As a result, the series with overall loading of 1 wt% generally show more pronounced synergetic effect than those with 5% loading. This phenomenon is not as distinct for MXene-containing MMMs (Figure 6), which can be ascribed to a better dispersion of MXene than GO at high loading.</p><!><p>The enhancement of (A,B) CO2 permeability and CO2/N2 selectivity and (C,D) CO2 diffusivity of Pebax-GO/HNT and Pebax-GO/SiO2 membranes.</p><p>The enhancement of (A,B) CO2 permeability and CO2/N2 selectivity and (C,D) CO2 diffusivity of Pebax-MXene/HNT and Pebax-MXene/SiO2 membranes.</p><!><p>More interesting results can be found by comparing the four pairs of dual fillers. First, GO is selected as the common filler and the effects of HNTs and SiO2 are compared. The results show that Pebax-GO/HNTs membranes has both higher CO2 permeability and CO2/N2 selectivity than Pebax-GO/SiO2 membranes, especially at 1 wt% overall filler loading. The Pebax-GO/HNTs-0.5/0.5 membrane exhibits optimal gas separation performance. The CO2 permeability and CO2/N2 selectivity achieve 131 and 74% enhancement, respectively, compared to the Pebax control membrane. On one hand, the enhanced CO2 transport originates in part from the ameliorative filler dispersion in the dual-filler MMMs, which increases the effective area of filler and thus leads to an enhanced CO2 transport of membrane. Besides, as shown in Table S1, Pebax-GO/HNTs exhibits higher gas diffusivity but lower gas solubility than Pebax-GO/SiO2 membrane, demonstrating that the SiO2 particle is much more CO2-philic than HNTs. Owing to the relatively lower diffusion resistance, Pebax-GO/HNTs (0.5/0.5) membranes exhibit 33% higher gas permeability than Pebax-GO/SiO2 (0.5/0.5) membranes. According to Figures 5C,D, the enhanced permeability compared to the pristine Pebax membrane is mainly based on the increase of diffusivity, particularly to CO2. The fillers are not enhancing the CO2 solubility, which in Pebax is already quite high and responsible for its excellent performance for CO2/N2 separation. The addition of fillers increased the CO2 diffusivity up to 364%, particularly in the case of the GO/HNTs dual filler system. Simultaneously, a maximum enhancement of CO2/N2 diffusivity selectivity considering diffusivity changes for both gases was higher than 130%. The enhancement of CO2/N2 selectivity can be understood from Figure 7, which clearly shows that the overall selectivity enhancement is determined by diffusivity selectivity, and the solubility selectivity of each MMM is lower than the pristine Pebax membrane. The increment of diffusivity selectivity is acquired at the cost of sacrificing solubility selectivity, and similar phenomena can be found in the literature, where GO was incorporated into Pebax membrane (Li et al., 2015a). It was reported that the complexation enthalpy of CO2-dimethyl ether complex is approximately 8 kJ mol−1 (Van Ginderen et al., 2003), reflecting the strength of typical dipole–quadrupole interactions. However, the hydrogen bonding energy for O–H…O was reported to be much higher (20~30 kJ mol−1) (Zhao et al., 2017). Therefore, it is reasonable to speculate that the hydrogen bonding between PEO and hydroxyl group-containing fillers will affect the formation of CO2-ether complex, which may decrease both the solubility coefficient and solubility selectivity. Another possible reason is that the presence of HNTs might induce the horizontal orientation for both GO sheets and HNTs, which is known to create a tortuous path to transport (Wong et al., 2019). In this way, GO/HNTs pair can effectively improve the tortuosity of gas transport channel, and hence the CO2/N2 selectivity (Li et al., 2015b). Decisive evidences can be found from the PALS data (Table S3). The notable decrease of r3 from 0.316 nm (Pebax-GO-1) or 0.317 nm (Pebax-HNTs-1) to 0.311 nm (Pebax-GO/HNTs-0.5/0.5) clearly reveals that the co-existence of GO and HNTs produces synergetic effect that increases the chain rigidity and diffusivity selectivity. This effect can be better understood though the interfacial morphology theory. Since HNT is mesoporous filler with lumen size up to 20 nm, there is no doubt that partial pore blockage by polymer chains will occur. Furthermore, in this study, the HNT was modified with PSS, which further enhanced the interactions between HNT and Pebax due to the favorable interactions between PSS and PEO chains (Wang et al., 2005; Mcdonald and Hammond, 2018). On the other hand, HNT can significantly enhance gas diffusivity due to the presence of broad internal channel, and Pebax-HNTs-1 membrane shows the highest CO2 diffusivity coefficient among all the membranes prepared in this study (Table S1). It is not surprising that the large inner diameter of HNT does not bring any decrease of permeability, considering the potential pore blockage interfacial morphology. According to the updated morphological diagram proposed by Ismail's group (Hashemifard et al., 2011), the best interfacial morphology for mesoporous filler-based MMMs is often "rigidification" or "pore blockage," rather than the ideal case. The chain rigidification effect caused by fillers can be also seen from the transition temperature data shown in Figure S5. In this way, HNTs as fillers have the potential to simultaneously enhance permeability and selectivity. Although the concentration of HNT is rather low (1 wt%), the nanotubes are well-dispersed, especially at the presence of GO. The good dispersion of HNTs are beneficial to make full use of their benefits. When increasing the overall filler content up to 5%, the membranes exhibit decreased gas separation performance, which probably derived from the filler agglomeration that decreased the property of membrane.</p><!><p>The enhancement of diffusivity selectivity and solubility selectivity of (A) Pebax-GO/HNT, (B) Pebax-GO/SiO2, (C) Pebax-MXene/HNT, and (D) Pebax-MXene/SiO2 membranes. The overall loading is fixed at 1 wt%.</p><!><p>In contrast to GO, MXene works better with SiO2 than with HNTs in matrix. The difference between Pebax-MXene/HNTs and Pebax-MXene/SiO2 is not as sharp as that between Pebax-GO/HNTs and Pebax-GO/SiO2, but the highest values of CO2 permeability and CO2/N2 selectivity of Pebax-MXene/SiO2 membrane are obviously higher than those of Pebax-MXene/HNTs membrane at 1 wt% overall loading. The Pebax-MXene/SiO2-0.2/0.8 membrane shows optimal gas separation performance with 104 and 49% enhancement of CO2 permeability and CO2/N2 selectivity based on pure Pebax membrane, or 129 and 119% as high as those of Pebax-MXene/HNTs-0.2/0.8 membrane, respectively. By comparing Figures 6C,D, there is continuous decline of CO2 diffusivity with the increase of relative content of MXene/HNT dual fillers, but the MXene/SiO2 dual fillers result in a maximum CO2 diffusivity higher than that of each of the corresponding single-filler MMM. That is, MXene/SiO2 dual fillers produce synergic effect while the MXene/HNT ones do not. Since SiO2 microspheres synthesized by Stöber method is known to achieve mono-dispersion, the dispersity of SiO2 is expected to much better than HNT, and the former is therefore envisaged to better interrupt the stacking of MXene nanosheets into thicker ones.</p><p>If we keep the HNTs concentration constant and compare the two 2D fillers, we can find that Pebax-GO/HNTs membrane shows much higher separation performance (especially CO2/N2 selectivity) than Pebax-MXene/HNTs membrane with the overall loading at both 1 and 5 wt%. Notably, the CO2 permeability and CO2/N2 selectivity of Pebax-GO/HNTs-0.5/0.5 membrane are 48 and 69% higher than those of Pebax-MXene/HNTs-0.5/0.5 membrane, respectively. Such results stem from the difference in rigidity between GO and MXene. In single-filler MMMs, HNTs tend to agglomerate in the matrix because of the high aspect ratio and strong inter-particle Van de Waals forces, thus cause sharp performance degradation (Wong et al., 2019), although after modification with PSS, a considerable improvement has been observed (Zhang et al., 2018). When it comes to dual-filler MMMs, the flexible GO sheets are known to be able to wrap the nanotubes and thus retard their agglomeration (Meng et al., 2012). Despite the lack of clear evidence of GO-wrapped nanotubes, the steric effect arising from GO sheets and the hydrogen interaction of surface functional groups between GO and HNTs can also promote the dispersion of HNTs, therefore improving the effective surface area of the fillers to furnish gas transport pathways. Furthermore, the preferential horizontal orientation GO and HNTs improve the tortuosity of gas transport, which increase the CO2/N2 selectivity. For Pebax-MXene/HNTs membrane, since MXene is more rigid than GO and more difficult to be dispersed as single sheets, the amelioration of HNTs dispersion is not pronounced.</p><p>From another perspective, MXene outperforms GO when SiO2 incorporation is fixed. Pebax-MXene/SiO2 membrane shows superior CO2 permeability and CO2/N2 selectivity compared to Pebax-GO/SiO2 membrane, which is distinct at 1 wt% filler content. Especially, the CO2 permeability and CO2/N2 selectivity of Pebax-MXene/SiO2-0.2/0.8 membrane is 104 and 49% enhanced compare to the pristine Pebax. Again, according to Figures 6D, 7 the enhancements are due to increase in diffusivity, since the CO2 solubility in the dual filler MMM is slightly smaller than in the pure polymer. Herein the r3 values from PALS data do not reveal the same chain rigidification effect as shown in Pebax-GO/HNT membrane. This phenomenon is reasonable because there is no pore blockage and PSS modification around SiO2 surface. The enhancement of diffusivity selectivity can be only interpreted by the tortuosity of gas transport channel. For the dual filler MXene/SiO2-0.2/0.8, there are two advantages for acquiring good MXene dispersion: the very low MXene concentration and the presence of highly disperse SiO2 microspheres. In this case the MXene nanosheets can create more diffusion obstacles and prolong the molecular diffusion routes, so as to effectively enhance diffusivity selectivity.</p><p>The effect of temperature and pressure on membrane performance was also investigated. As shown in Figure S6, each membrane displays a substantial increment when the operation temperature increases from 30 to 60°C, which typically represent the increase of gas diffusivity and polymer chain mobility. Interestingly, the decline of selectivity is not as distinct as the increase of permeability. Since solubility selectivity is very sensitive to temperature change, this fact further supports the diffusivity-dominated selectivity mechanism, and indicates the adequate polymer-filler interactions within the temperature range. In addition, the dependence CO2 permeability on temperature can be further correlated according to Arrenius relationship (Figure S7). The slope of each straight line is known to reflect the activation energy of CO2 permeation. As such, Pebax-GO/HNT-0.5/0.5 shows the lowest CO2 permeation activation energy, in good accordance with the high diffusivity coefficient at room temperature. Figure S8 shows that CO2 permeability as well as CO2/N2 selectivity only changes little in the pressure range from 1 to 5 bar, which accords well with the Pebax-based MMMs reported in the literature (Duan et al., 2019). It demonstrates that the absorption of CO2 of the membranes within such pressure range almost follows Henry's Law, and no effect due to compaction is observed.</p><!><p>In summary, we fabricate a series of dual-filler MMMs by matching two non-2D fillers, SiO2 and HNTs, with two 2D fillers, GO and MXene, respectively. All dual-filler MMMs exhibit superior gas separation performance compared to the corresponding single filler MMMs, revealing the existence of synergetic effect between each pair of fillers. Such effect at low overall loading (1 wt%) is more notable than that at high loading (5 wt%), arising from the better dispersion of samples with 1 wt% filler. Interestingly, GO and MXene are found to meet different preferential partners due to their differences. On one hand, GO/HNTs proves to be a better pair than GO/SiO2, since HNTs are known to be wrapped by the flexible GO sheets so as to promote the dispersion of nanotubes. In turn, HNTs are deemed to hinder the restacking of GO sheets because of the strong steric effect. Compared to Pebax-HNTs and Pebax-GO membranes, the Pebax-GO/HNTs-0.5/0.5 membrane has optimal CO2 permeability with the enhancement of 107 and 100%, respectively. On the other hand, MXene works well with SiO2 rather than HNTs. In particular, the Pebax-MXene/ SiO2-0.2/0.8 membrane achieves 33% and 58% enhancement of CO2 permeability and CO2/N2 selectivity compared to Pebax-MXene membrane.</p><!><p>All datasets generated for this study are included in the article/Supplementary Material.</p><!><p>YL designed the study. JS and FS prepared and characterized fillers and membranes. XC performed the PALS analysis. FS, YL, SN, SW, ML, ZY, and JW conducted data analysis, figure drawing, and writing.</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

Comparison of mRNA Splicing Assay Protocols across Multiple Laboratories: Recommendations for Best Practice in Standardized Clinical Testing

Background Accurate evaluation of unclassified sequence variants in cancer predisposition genes is essential for clinical management and depends on a multifactorial analysis of clinical, genetic, pathologic, and bioinformatic variables and assays of transcript length and abundance. The integrity of assay data in turn relies on appropriate assay design, interpretation, and reporting. Methods We conducted a multicenter investigation to compare mRNA splicing assay protocols used by members of the ENIGMA (Evidence-Based Network for the Interpretation of Germline Mutant Alleles) consortium. We compared similarities and differences in results derived from analysis of a panel of breast cancer 1, early onset (BRCA1) and breast cancer 2, early onset (BRCA2) gene variants known to alter splicing (BRCA1: c.135-1G>T, c.591C>T, c.594-2A>C, c.671-2A>G, and c.5467+5G>C and BRCA2: c.426-12_8delGTTTT, c.7988A>T, c.8632+1G>A, and c.9501+3A>T). Differences in protocols were then assessed to determine which elements were critical in reliable assay design. Results PCR primer design strategies, PCR conditions, and product detection methods, combined with a prior knowledge of expected alternative transcripts, were the key factors for accurate splicing assay results. For example, because of the position of primers and PCR extension times, several isoforms associated with BRCA1, c.594-2A>C and c.671-2A>G, were not detected by many sites. Variation was most evident for the detection of low-abundance transcripts (e.g., BRCA2 c.8632+1G>A \xce\x9419,20 and BRCA1 c.135-1g>t \xce\x945q and \xce\x943). Detection of low-abundance transcripts was sometimes addressed by using more analytically sensitive detection methods (e.g., BRCA2 c.426-12_8delGTTTT ins18bp). Conclusions We provide recommendations for best practice and raise key issues to consider when designing mRNA assays for evaluation of unclassified sequence variants.

comparison_of_mrna_splicing_assay_protocols_across_multiple_laboratories:_recommendations_for_best_p

<!>Materials and Methods<!>Results<!>Discussion

<p>Germline mutations in the breast cancer susceptibility genes breast cancer 1, early onset (BRCA1)26 and breast cancer 2, early onset (BRCA2) (OMIM #113705 and #600185, respectively) are associated with a significantly increased risk of breast and other cancers (1). Although many thousands of disease-associated mutations have been identified in these genes, many DNA sequence changes found during genetic screening fall into the category of unclassified variants because their functional and clinical significance is not immediately clear. Such unclassified variants pose a challenge for clinical management of variant carriers.</p><p>Unclassified variants have the potential to alter protein function by changing the coding sequence of a transcript, or the level or structure of the gene transcript, and by disrupting regulatory regions in promoters, untranslated regions, exons, or introns (2-5). Such regulatory variants include those affecting normal splicing of BRCA1 and BRCA2, many of which have been shown to be clinically significant by use of cDNA studies and multifactorial likelihood analysis methods that combine bioinformatic, pathologic, and clinical information (6-8). These variants include those that affect splicing by disrupting or weakening the motifs at intron-exon boundaries, introducing de novo splice acceptor or donor sites, activating cryptic splice sites, or disrupting enhancer and silencer sequences. Several studies have shown that bioinformatic prediction tools can be used to prioritize variants for splicing assays (9-14).</p><p>To date, a total of 82 studies have reported findings related to splicing in BRCA1 or BRCA2 (15). The majority of these used reverse transcriptase PCR (RT-PCR)27 analysis of RNA extracted from blood of variant carriers or alternatively, minigene constructs containing the variant and assayed in non-patient-derived cell lines. The interpretation of splicing results for variant carriers can be complicated by the detection of normal alternatively spliced transcripts that occur in healthy individuals—an issue that has yet to be extensively addressed in the literature. The effect of the range of variables found in protocols used in research and clinical testing laboratories, including the PCR assay design, reagents used, and tools for visualizing and characterizing transcripts identified by PCR on assay result interpretation, is also unclear.</p><p>There are 4 instances of inconsistent or conflicting splicing results (6, 8, 14, 16-19). These include BRCA1 c.212+3A>G, c.670+8C>T, and c.736T>G and BRCA2 c.517-19C>T (4, 19-25). Reports of splicing results from a further 7 variants differed in the number of aberrant bands found in each study. The potential clinical implications of such inconsistencies highlight the need to establish the advantages and limitations of the various techniques in practice.</p><p>Guidelines for clinical interpretation and reporting of unclassified variants analyzed using splicing assays are available in the UK and Netherlands via the UK Clinical Molecular Genetics Society (http://www.cmgs.org/BPGs/Best_Practice_Guidelines.htm) and Dutch Society of Clinical Genetic Laboratory Specialists (http://www.vkgl.nl/). In addition, a range of in silico approaches have been compared with one another, and with transcript analysis, by the splice network of the French BRCA diagnostic testing laboratories, recently reported by Houdayer et al. (11). In this study (11), Houdayer et al. investigated the value of combining Splice-site Finder and MaxEntScan prediction tools and showed that major splice defects were consistently identified across a number of different laboratories. The authors did find some discrepancies with results previously reported in the literature and recommended a large cross-validation study as a future priority.</p><p>The Evidence-Based Network for the Interpretation of Germline Mutant Alleles (ENIGMA) consortium was established in 2009 with the purpose of sharing data, methods, and resources to facilitate classification of unclassified variants (21). To date, a total of 3286 unique BRCA1 and BRCA2 variants considered to be of uncertain clinical significance have been submitted to ENIGMA from more than 43 sites in 19 countries. The consortium has established several working groups, including one dedicated to examining variants that potentially alter RNA splicing.</p><p>Here we describe the outcome of an ENIGMA Splicing Working Group study to assess the importance of various mRNA assay components on consistency of results. We identified a variety of differences in protocols from 23 laboratories, the majority of which conduct routine clinical assays (see Table 1 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol60/issue2). We report the critical elements on assay design that should be considered in the analysis of variants that may impact RNA splicing.</p><!><p>Each participating laboratory submitted information about the mRNA splicing protocol in use at their site. These protocols were then compared on the basis of the source of biological material; the use of a nonsense mediated decay (NMD) inhibitor, RNA extraction, or removal of contaminating genomic DNA; the choice of cDNA synthesis primer, reverse transcriptase, and DNA Taq polymerase; the method of PCR product detection; and whether products were isolated, subcloned or sequenced (see online Supplemental Table 1).</p><p>To compare the assays used by laboratories within the ENIGMA consortium, 23 sites were sent aliquots of samples from the same lymphoblastoid cell lines (LCLs) that had been generated by the Kathleen Cuningham Consortium for Research into Familial Breast Cancer (kConFab) from 9 carriers of BRCA1 or BRCA2 variants known to be associated with splicing defects (Fig. 1) and from 11 controls. Four LCLs carried variants that produce unequivocal splicing aberrations resulting in a clear exon-skipping event. Five LCLs carried variants considered to produce equivocal splicing aberrations, based on the observation that they confer more subtle and variable effects, such as altering the availability of naturally occurring isoforms to avariable extent and/or producing a large and variable number of uncharacterized splicing products.</p><p>The project was conducted in 2 phases. In the initial phase (phase 1), 16 sites used an mRNA splicing protocol they routinely use in their laboratory (see online Supplemental Table 1), summarized their results, and submitted these to PJW and MAB. for collation. Following an analysis of phase 1 results, phase 2, informed by the phase 1 findings, was initiated, during which some sites repeated each assay using a standard set of PCR primers and cycling conditions (see online Supplemental Table 2). All other components of the protocol were per phase 1, apart from site 8, which used a Bioanalyzer in phase 1 and capillary electrophoresis (CE) in phase 2. Seven sites that participated in phase 1 repeated the assays under the controlled conditions of phase 2. An additional 3 sites joined the study to assay all variants for phase 2. A further 3 sites joined phase 2 to specifically assay BRCA1 c.671-2A>G, following the finding that this equivocal variant gave rise to the greatest range of alternatively spliced transcripts.</p><!><p>The initial comparison of protocols used by participating laboratories revealed that cycloheximide or puromycin was sometimes used for NMD treatment, with incubation times between 4 and 8 h and concentrations between 100 and 250 μg/mL, the use of 8 cDNA synthesis kits, 12 different DNA polymerases, and transcript isolation strategies that included band excision, subcloning, and sequencing. The majority of laboratories used agarose gel electrophoresis for visualizing transcripts, but several used digital visualization strategies.</p><p>In phase 1 of the study, all sites detected the fulllength transcript for each of the 4 unequivocal variants (Table 1). All sites also detected the most prominent single-exon skipping events not seen in controls for each of the unequivocal variants, apart from site 4 and 14, which did not detect the Δ20 transcript for the unequivocal variant BRCA2 c.8632+1G>A. Not all sites detected all of the less abundant transcripts from this variant, however, with only 3/16 sites detecting the Δ19&20 transcript and only 6/16 sites detecting the ins i20 transcript. For the unequivocal variant BRCA1 c.135-1g>t, which has been associated with multiple splice isoforms (22), only 3/16 detected the Δ5q transcript, and only 1 site detected the Δ3 transcript.</p><p>Detailed analysis of each of the protocols and resulting data revealed that the range of PCR design strategies contributed to the variation in detection of transcripts, in particular PCR primer design and PCR cycling conditions. For example, 11 out of 16 sites that analyzed equivocal BRCA1 c.671-2A>G were unable to detect all of the transcripts because primer position did not allow some, clearly unanticipated, fragments to be amplified (Table 1). Forward primers positioned in exon 9 or 10 were unable to amplify isoforms lacking those exons, including Δ9/10, Δ9/10/11, or Δ9/10/11q isoforms seen in controls, or the Δ9/11 or Δ10/11 variant-associated isoforms detected by other sites.</p><p>The length of extension time during PCR amplification was also found to be a contributing factor, with several protocols using times that were likely to be too short to detect the longer PCR products amplified from some splice isoforms. For example, 5 sites used elongation times of 3 min or less and were unable to amplify full-length transcripts or transcripts containing exon 11 (Δ9 or Δ9/10) for BRCA1 c.671-2A>G, which are longer than 3 kb. As for the results observed for unequivocal variants, an additional explanation for variation in detection of transcripts was the low abundance of some transcripts, including those identified in the variant carrier only (e.g., Δ9/11, Δ10/11, and Δ>3kb exon 11 transcripts), which is known to lead to variable PCR amplification. PCR cycle number was also important, with site 23 detecting only a limited number of transcripts (Table 1), likely reflecting the use of only 25 cycles (see online Supplemental Table 1).</p><p>Given that phase 1 showed that many transcripts were not observable due to the positioning of primers or elongation time, phase 2 of the study was initiated. Phase 2 included assays conducted by 10-12 sites (depending on the variant analyzed) using a standard set of primers and elongation times appropriate for the expected lengths of the transcripts (see online Supplemental Table 2). The outcome was a much greater analytical sensitivity and consistency of results (Table 2). For example, all sites were now able to detect relatively high-abundance isoforms or variant-associated transcripts reported in previous studies, but not consistently reported in phase 1 [Δ17,18 for BRCA2 c.7988A>T, Δ20 for BRCA2 c.8632+1G>A, Δ5 for BRCA2 c.426-12_8delGTTTT, and Δ10 for BRCA1 c.594-2A>C (5, 7, 8)]. Importantly, unlike phase 1, in phase 2 all study sites were able to detect at least 1 aberrant band (cf. controls) and thus may have been able to better classify the variant using the IARC (International Agency for Research on Cancer) 5-tier classification scheme.</p><p>There remained some inconsistencies in the phase 2 data. Further comparison of protocols suggested that the method of PCR product detection was likely to be a contributing factor. Sites 2 and 8 in phase 2 were the only sites to use CE exclusively for detection of transcripts. Site 2 had higher overall detection compared to the other sites. Indeed, 10 of the 23 transcripts (43.5%) identified across all sites in the phase I analysis of unequivocal variants were detected only by CE, demonstrating it to be a comparatively more analytically sensitive detection method. This trend continued for equivocal variants analyzed in phase 2, with 12 of the 49 (24.5%) transcripts detected only by capillary CE. The sites employing a Qiaxcel visualization, Bioanalyzer, or MultiNA systems demonstrated that these systems were often more analytically sensitive than gel electrophoresis. For example, sequencing of transcripts identified by Qiaxcel analysis of BRCA2 c.426-12_8delGTTTT (site 9, phase 2) showed that it was the only system to discriminate the small insertion of 18 nucleotides from the full-length transcript (Table 2; also see online Supplemental Fig. 1).</p><p>Analysis of BRCA1 c.594-2A>C in phase 2 identified 11 different transcripts. Excising bands from agarose gel or sequencing PCR products directly enabled detection of 3-6 transcripts (sites 3, 4, 17, and 18). Cloning PCR products followed by sequencing detected 6-7 transcripts (sites 1 and 16), and CE detected 10-11 transcripts (sites 2 and 8). This showed that cloning PCR products improved analytical sensitivity, and visualization by the Qiaxcel system or capillary CE together with sequence analysis is optimal to identify and characterize transcripts. The number of clones sequenced also appeared to improve analytical sensitivity; screening 40 clones (site 16) in comparison to 24 clones (site 1) enabled the detection of 1 additional transcript.</p><p>Finally, we examined the effect of using different reverse transcriptase enzymes with the same RNA, cDNA synthesis primers, and PCR primers, enzymes, and conditions. As shown in Fig. 2, the amplification of the longest transcripts was not possible with GoScript; with M-MuLV we missed in the patient with the c.671-2A>G variant the wild-type transcript; only Superscript II allowed amplification of the longest transcript in both controls and variant carriers.</p><p>It is important to note that all transcripts shown in Tables 1 and 2 were the outcome of results by scorers who were blind to the transcripts identified by other participants, to avoid biasing the interpretation and thus the value of each approach. Once the full range of transcripts was known, however, it was possible to find some missing transcripts, demonstrating the importance of prior knowledge in both the design of the assays and the interpretation of results.</p><p>There was no clear evidence of any differences as a result of using (a) cycloheximide vs puromycin treatment for NMD inhibition; (b) differing RNA extraction methods; (c) oligo d(T) and random hexamers vs gene specific primers; (e) various methods of DNase treatment; and (f) a particular type or brand of Taq polymerase.</p><p>A summary of the recommendations arising from this study is provided in Table 3.</p><!><p>RNA splicing assays are commonly used in diagnostic and research settings to assess the potential effects of unclassified variants in multiple genes, including BRCA1 and BRCA2. There are a multitude of differing protocols used in clinical and research laboratories, including those within the ENIGMA consortium, and this prompted a study aimed at establishing assay guidelines.</p><p>This study shows that prior knowledge of the expected transcripts, including naturally occurring isoforms and aberrant transcripts predicted to occur in variant-carrying samples, is important for assay design. Phase 1, followed by phase 2, demonstrated that the selection of primers used to amplify exons and the design of cycling conditions appropriate for that primer design explain the vast majority of the differential success of detecting some isoforms. In phase 2 of the project, during which primer design and extension time were controlled, all sites detected the fulllength transcript and the predominant alternative transcripts, suggesting that high-abundance aberrant transcripts will be detectable regardless of assay protocol, which is consistent with the conclusions of Houdayer et al. (12).</p><p>Variability in overall detection increased as the apparent abundance of individual transcripts in a sample decreased, and thus detection became more dependent on the sensitivity of the method of analysis. This variability is also likely to occur between replicates done in a single laboratory, in addition to that between different laboratories. A controlled comparison of different reverse transcriptases showed that Superscript is much better able to copy longer transcripts (Fig. 2). It is also possible that the maximum span length of some PCR polymerases contributed to the ability of some groups to detect longer transcripts. Furthermore, primer pairs that selectively amplify disease-associated isoforms rather than naturally occurring isoforms could increase assay sensitivity.</p><p>Sites that used gel electrophoresis visualization alone were unable to detect some bands because of the inherent insensitivity of this technique, combined with the stochastic nature of PCR when analyzing low levels of target (26). An example of this is site 1, which when analyzing the equivocal variant BRCA2 c.8632+1G>A detected the ins i21bp intron 9 in phase 1 but not in phase 2, despite using the same primers and PCR conditions.</p><p>Some sites sequenced PCR products. Sites that directly sequenced the products of PCR reactions experienced some challenges in determining the sequence of low-level transcripts. An accurate assessment of transcript sequence was also confounded by the presence of multiple (3 or more) PCR products of similar lengths. In these instances, adjustments to the concentration of agarose and running times of electrophoresis may improve analytical sensitivity. However, it appears that this may be less relevant if CE systems are adopted (see below). Cloning single PCR products into a vector system is a useful alternative for isolating transcripts and appears to improve sensitivity over band excision and sequencing alone. Furthermore, by increasing the number of clones screened it is possible to marginally increase the number of transcripts detected. However, to identify low-abundance transcripts, analysis of very large numbers of clones (100s or 1000s) or next generation sequencing would be necessary.</p><p>Of all the detection methods used, CE was shown to be the most analytically sensitive. For example, site 8 showed an increase in sensitivity from phase 1 to phase 2 after switching from using a Bioanalyzer to using CE. In addition to analytical sensitivity, the CE system has the added advantage of a greater resolution (1–2 bp) compared to Qiaxcel (3–5 bp). However, the limitation with both the Bioanalyzer and CE is the inability to harvest and thus perform sequence analysis of the PCR product. Also, CE relies on a prediction of the splicing event based on the length of the product observed, which can be limited by the inaccuracy of size standards, so a secondary set of primers may be required. It is also worth noting that very long full-length (or alternative) transcripts (like those involving BRCA1 exon11 and BRCA2 exons 10 and 11) cannot be analyzed by CE.</p><p>The results presented here represent each laboratory's initial assessment of each variant. Each site had the opportunity to reassess their results after the data from all sites were released to the group and several sites reported that they detected additional transcripts in addition to (and thus not shown) the initial conclusions reported in Tables 1 and 2. This finding suggests that a prior knowledge of all potential splice transcripts related to variant carriers, from studies such as these, as well as those that occur as naturally occurring isoforms in healthy controls, is essential not only to design detection strategies (see above) but to interpret results.</p><p>The use of analytically sensitive PCR product detection (CE and Qiaxcel in phases 1 and 2, Bioanalyzer in phase 1) enabled the identification of several novel low-abundance transcripts, in both normal controls and variant carriers. This raises the question of which detectable transcripts are functional and thus relevant for determining the pathogenicity of clinically identified unclassified variants, and whether or not low abundance transcripts are of biological or pathological significance in vivo. It is generally accepted that variants resulting in single major transcripts that lack an open reading frame will be deleterious (27). However, it is much less clear whether changes in the levels of low-abundance alternative splicing events will have an impact either directly or through altering the function or levels of endogenous transcripts including fulllength mRNA.</p><p>It is possible, for example, that a reduction in the full-length expression will have a deleterious effect on known BRCA1 functions (DNA repair, cell cycle control) (28). A quantitative analysis of the range of naturally occurring isoforms relative to full-length expression and relative to other BRCA1 or BRCA2 isoforms is required, as is a comprehensive analysis of the functional role of each of these isoforms in both the healthy functioning of BRCA genes and the consequences of sequence variation on this process (29). It will also be important to extend this investigation to breast and ovarian tissue, to gain a broader understanding of the tissue-specific nature of splice-isoform regulation. Importantly, this information will be essential to determine whether knowing the full complement of transcripts has the potential to have an impact on the final classification of the variant as pathogenic or otherwise. For example, does the expression profile of the 16 alternately spliced transcripts detected in BRCA1 c.671-2A>G carriers change at different tissue sites, and will this new information influence the classification of the variant?</p><p>In summary, we have shown that primer design, PCR conditions, and PCR product detection methodology, together with prior knowledge of potential transcripts, are important contributors to the analytical sensitivity of PCR-based assays for detecting alternatively spliced RNA transcripts from variant carriers and from wild-type sequences. These factors must be considered when designing assays, particularly when they form the basis of clinical decision-making. Furthermore, the formulation of standard assay design and detection methods is indicated for all variants, but particularly for those that may impact on isoform expression.</p>

PubMed Author Manuscript

A More Reactive Trigonal Bipyramidal High-Spin Oxoiron(IV) Complex with a cis-Labile Site

The trigonal bipyramidal high-spin (S = 2) oxoiron(IV) complex [FeIV(O)(TMG2dien)(CH3CN)]2+ (7) was synthesized and spectroscopically characterized. Substitution of the CH3CN ligand by anions, demonstrated here for X = N3\xe2\x88\x92 and Cl\xe2\x88\x92, yielded further S = 2 oxoiron(IV) complexes of general formulation [FeIV(O)(TMG2dien)(X)]+ (7-X). The reduced steric bulk of 7 relative to the published S = 2 complex [FeIV(O)(TMG3tren)]2+ (2) was reflected by enhanced rates of intermolecular substrate oxidation.

a_more_reactive_trigonal_bipyramidal_high-spin_oxoiron(iv)_complex_with_a_cis-labile_site

<p>Non-heme monoiron oxygen activating enzymes perform a remarkably diverse array of highly selective oxidative transformations, 1 Most have iron centers with a 2-His-1-carboxylate facial triad structural motif, and their catalytic cycles often involve oxoiron(IV) intermediates as oxidants. Within the past several years, such oxoiron(IV) species have been trapped and spectroscopically characterized in several enzymes and found in all cases to be high spin (S = 2).2 In contrast, the overwhelming majority of existing synthetic oxoiron(IV) complexes have S = 1 ground states.3 To date the only published examples of S = 2 oxoiron(IV) complexes are [FeIV(O)(H2O)5]2+ (1),4 [FeIV(O)(TMG3tren)]2+ (2, TMG3tren = 1,1,1-tris{2-[N2-(1,1,3,3-tetramethylguanidino)]-ethyl}amine))5 and [FeIV(O)(H3buea)]− (3, H3buea = tris[(N′-tert-butylureaylato)-N-ethylene]amine trianion).6 The crystallographically characterized trigonal bipyramidal (TBP) complex 2 was found to react rapidly via intramolecular ligand hydroxylation (t1/2 at 25 °C = 30sec), but reacted with external hydrocarbon substrates at rates comparable to those of existing S = 1 complexes. Given that many DFT studies predict more facile H-atom abstraction by S = 2 oxoiron(IV) centers than their S = 1 counterparts,7 the intermolecular reactivity observed for 2 was disappointingly sluggish, a fact attributed to a steric retardation of reaction due to the bulk of the tetramethylguanidine donors.5a,8</p><p>In an effort to assess and rationalize the inherent reactivity of the S = 2 oxoiron(IV) center in 2, we sought to (i) reduce the steric bulk of the supporting ligand and (ii) expand the palette of existing high-spin oxoiron(IV) complexes. Both aims are easily accommodated by replacement of one arm of the tripodal TMG3tren ligand with a methyl group to yield the tridentate TMG2dien ligand (Figure 1A).</p><p>The iron(II) starting material used in this study, [FeII(TMG2dien)(OTf)2] (4), was prepared by thallium(I) salt metathesis of the chloride ligands in [FeII(TMG2dien)(Cl)2] (5), which was itself generated by combination of equimolar quantities of TMG2dien and FeCl2. The high-resolution X-ray structures of 4 and 5 (Figures 1B and S1, Tables S1 and S2) reveal 5-coordinate complexes with a geometry that is intermediate between square pyramidal (SP) and TBP (τ = 0.64 and 0.56, respectively). 9 This contrasts with the strictly TBP geometry of [FeII(TMG3tren)(OTf)](OTf) (6, τ = 0.96), the iron(II) starting material used in the generation of 2.5a</p><p>Treatment of a CH3CN solution of 4 with 2-(tert-butylsulfonyl) iodosylbenzene (2-(tBuSO2)C6H4IO)10 yields an orange-brown species 7 (t1/2 at −30 °C ≈ 0.5 h vs. 4.5 h for 2) whose electronic spectrum is reminiscent of 2 (Figure 2, Table 1), with a weak near IR (NIR) band centered at 805 nm (270 M−1 cm−1) and an intense UV absorption having a shoulder at ca. 380 nm (8200 M−1 cm−1). Notably, maximization of the intensity of the aforementioned NIR band requires addition of 2.5 – 3 equiv of oxidant. Mössbauer spectroscopy revealed that reaction with a single equivalent of oxidant leads to substoichiometric yields of 7 (45 %), with the remainder of the iron content being associated primarily with unreacted 4 (Figure S2). The absence of other iron products is consistent with non-productive reaction of 2-(tBuSO2)C6H4IO due to metal-catalyzed disproportionation, a process that has been documented for this oxidant.10 Additionally, the 19F NMR spectrum of 7 in CD3CN displayed a single peak at −80 ppm, which is indicative of free triflate and suggests that the potentially cis-labile site is filled by a solvent ligand. This observation, combined with the other data detailed herein, leads to formulation of 7 as [FeIV(O)(TMG2dien)(CH3CN)]2+.</p><p>One of the primary motivations for the development of the chemistry of the TMG2dien ligand was to generate a TBP high-spin oxoiron(IV) complex incorporating a labile site. To test the viability of our approach, tetraalkylammonium azide and chloride salts were added to pre-formed solutions of iron(IV) complex 7. This led to near instantaneous UV-Vis spectral changes (Figure 2, Table 1), consistent with substitution of the CH3CN ligand and formation of new complexes formulated as [FeIV(O)(TMG2dien)(X)]+ (7-X, X = N3, Cl), with the resultant spectra retaining the same general features as the parent complex 7 (i.e. a weak NIR band and an intense UV feature that trails into the visible region). Notably, 7-N3 is of comparable stability to 7, but 7-Cl undergoes self-decay at a significantly accelerated rate (t1/2 at −30 °C ≈ 34 and 2 min. for 7-N3 and 7-Cl, respectively).</p><p>The presence of a terminal Fe=O unit in complexes 7, 7-N3, and 7-Cl was confirmed by resonance Raman spectroscopy (Figure S3), which yielded ν(Fe=O) vibrational modes at 807, 833 and 810 cm−1, respectively. In close agreement with our expectations based upon Hooke's Law (∆ν theoretical ≈ 36 – 37 cm−1), these bands shifted upon 18O-labelling to 773, 795 and 775 cm−1, respectively.</p><p>Mössbauer spectroscopy confirms that complexes 7, 7-N3, and 7-Cl all contain S = 2 oxoiron(IV) centers (Figures 3 (left panel) and S4–S7), with zero-field spectra each exhibiting a doublet with isomer shift δ ≈ 0.1 mm s−1 (Table 1). Their distinct quadrupole splittings, ∆ EQ, confirm formation of new anion complexes. The three species were obtained in high yield, with 7, 7-N3, and 7-Cl accounting for 88, 80 and 87 % (all ca. ±4%) of the Fe present, respectively. Minor iron(III) impurities account for the remainder. Fitting Mössbauer spectra observed at variable applied fields, B, to an S = 2 spin Hamiltonian (see SI) yielded zero-field splitting (ZFS) parameters (D, E/D) and hyperfine parameters that compare well with other S = 2 oxo-iron(IV) species.</p><p>The high-spin ground state of 7, 7-N3, and 7-Cl and the D and E/D values obtained by Mössbauer spectroscopy were confirmed by parallel mode EPR (Figure 3 (right panel) and S8–S10). The X-band spectra of 7, 7-N3 and 7-Cl all displayed broad resonances at g ~ 8 – 13 that originate from the excited state MS = ± 2 quasi-doublet of an S = 2 multiplet. As the intensity of such signals is proportional to (E/D)4, quantitative simulations allowed quite accurate determinations of E/D (Table 1).11</p><p>Consistent with their assignment as oxoiron(IV) complexes, 7, 7-N3 and 7-Cl exhibit XAS edge energies (E0) of ~7124 eV (Fig. 4, Table 1) that are in the range observed for other oxoiron(IV) complexes and blue-shifted by ≥ 1.7 eV relative to the iron(II) starting material 4 (E0 = 7121.9 eV). Unlike S = 1 oxoiron(IV) complexes, which exhibit a single symmetrical pre-edge feature,12 7, 7-N3 and 7-Cl all exhibit pre-edge features composed of two discernible peaks (Figure 4 top and Table S3), originating from 1s→3d electronic transitions. Consequently, two Gaussians are required to model them successfully, as predicted by DFT for high-spin iron(IV) complexes.13</p><p>Analysis of the EXAFS data for both 7 and 7-Cl furnished best fits (Tables S4 and S5, Figures 4 and S11) with an O/N scatterer at ~1.65 Å that corresponds to the Fe=O unit. (Thus far, we have been unable to collect high-quality EXAFS data for 7-N3 due to rapid photo-decomposition.) These distances are comparable to the Fe=O lengths observed in the X-ray structures of 2 and 3 (1.661(2) and 1.680(1) Å, respectively),5a,6 the EXAFS of enzymatic oxoiron(IV) intermediates,2e,14b,15 and the plethora of existing S = 1 oxoiron(IV) complexes.3,16 Both 7 and 7-Cl also have a shell of O/N scatterers at ca. 1.94 Å, four in the former case and three in the latter, assigned to the N-donors of the supporting ligands. This Fe-N distance is shorter than that found for 2 by EXAFS (1.99 Å),5a reflecting the lower steric constraints of the TMG2dien ligand. Lastly, 7-Cl has a Cl scatterer at 2.27 Å, a distance that is very similar to the 2.31 Å Fe–Cl distance obtained by EXAFS for the chloroferryl intermediate of the α-ketoglutarate-dependent aliphatic halogenase SyrB2.2e</p><p>DFT calculations for 7, 7-N3 and 7-Cl further support our S = 2 spin state assignment for these three complexes (Tables S6– S13). Complex 7, 7-N3, and 7-Cl all have a 5A ground state with four d electrons located in two half-filled E levels (Table S7). Spin populations calculated for the iron and the oxo atoms, respectively, are +3.0(1) and +0.6(1) (Tables S9 and S10), similar to the values obtained for 1, 2, and TauD-J.4a,5a,14c The DFT geometry-optimized structures of 7, 7-N3 and 7-Cl (Figures 1C and S12, Table S8) exhibit geometries that are best described as TBP (τ = 0.79, 0.83 and 0.72, respectively)9 and have Fe=O bond lengths of ~1.65 Å, in close agreement with values obtained from EXAFS. In contrast, the Fe-Cl distance of 2.35 Å calculated for 7-Cl is somewhat longer than the value of 2.27 Å determined by EXAFS. Lastly, the DFT-calculated spin-dipolar contribution to the 57Fe A-tensor is in good agreement with the experimental data (Table S6), indicating that the z axis of the spin Hamiltonian (determined by the ZFS tensor) is oriented along the Fe-O bond (within about 5°).</p><p>In addition to creating a S = 2 oxoiron(IV) complex with a cis-labile site, it was anticipated that removing one of the bulky tetramethylguanidinyl donors of the TMG3tren ligand to give TMG2dien would provide substrates greater access to the FeIV=O unit, thereby allowing the inherent reactivity properties of the S = 2 oxoiron(IV) center to be manifested. Consistent with these expectations, the oxo-transfer reaction of 7 to PPh3 proceeded so rapidly that we were unable to accurately measure the associated rate constants at −30 °C for comparison with published data for other oxoiron(IV) complexes listed in Table 2. Additionally, H-atom abstraction from 1,4-cyclohexadiene (CHD) and 9,10-dihydroanthracene (DHA), substrates with similarly weak C–H bonds but differing steric profiles, proceeded at comparable rates, with respective second-order rate constants 15 and 630 times larger than for the more sterically hindered 2 (Table 2, Figures S13–S15). </p><p>Notably, 7 exhibits reactivity more than one and three orders of magnitude greater than the S = 1 complexes [FeIV(O)(N4Py)]2+ (8) and [FeIV(O)(TMC)(CH3CN)]2+ (9) (Table 2), respectively, which would appear to support DFT-based predictions of a more reactive S = 2 FeIV=O center relative to a S = 1 FeIV=O center.7 However, 7 is an order of magnitude less reactive than the recently reported S = 1 complex [FeIV(O)(Me3NTB)]2+ (10).17 This fact serves to highlight the difficulty of making such comparisons without consideration of the thermodynamic and steric consequences of the differing ligand environments of the various complexes. Thus far, there is only one pair of closely related complexes that have identical ligand environments, namely [(HO)(L)FeIII/IV–O–FeIV(L)(O)] where L = tris(3,5-dimethyl-4-methoxypyridyl–2-methyl)amine, but differ in having a S = 1 or S = 2 oxoiron(IV) unit.18 Remarkably, the high-spin FeIIIFeIV complex was found to be a thousandfold more reactive than the low-spin FeIVFeIV complex, thereby providing support for the DFT-based predictions.</p><p>In summary, we have described the synthesis of the high-spin oxoiron(IV) complex [FeIV(O)(TMG2dien)(CH3CN)]2+ (7), which is related to the S = 2 complex [FeIV(O)(TMG3dien)(CH3CN)]2+ (2) by replacement of one of the tetramethylguanidinyl arms of the TMG3tren ligands by a methyl group and inclusion of a solvent ligand in its place. This modification provides greater access to the FeIV=O subunit, eliminating the selectivity for smaller substrates exhibited by 2 and resulting in a significant increase in the rates of intermolecular reactions. Furthermore, the introduction of CH3CN as an equatorial ligand in 7 provides a means to access a series of closely related anion substituted S = 2 oxoiron(IV) complexes that are highly amenable to characterization, as illustrated here for [FeIV(O)(TMG2dien)(X)]+ (7-X, X = N3, Cl). This offers the promise of elucidating spectroscopic and reactivity trends as a function of the electronic properties of an S = 2 oxoiron(IV) center and may provide answers to specific bio-relevant questions, such as the reason for the omnipresence of carboxylato ligands in non-heme enzymes1 and for the inherent preference for halogen versus oxygen atom rebound in non-heme iron halogenase enzymes. 19</p>

PubMed Author Manuscript

Design and Synthesis of (+)-Discodermolide-Paclitaxel Hybrids Leading to Enhanced Biological Activity1

Potential binding modes of (+)-discodermolide at the paclitaxel binding site of tubulin have been identified by computational studies based on earlier structural and SAR data. Examination of the prospective binding modes reveal that the aromatic pocket occupied by the paclitaxel side-chain is unoccupied by (+)-discodermolide. Based on these findings, a small library of (+)-discodermolide-paclitaxel hybrids have been designed and synthesized. Biological evaluation reveals a two- to eight- fold increase in antiproliferative activity compared to the parent molecule using the A549 and MCF-7 cancer cell lines.

design_and_synthesis_of_(+)-discodermolide-paclitaxel_hybrids_leading_to_enhanced_biological_activit

<!>The Solution Conformations of (+)-Discodermolide<!>Docking Studies<!>Design of Hybrid Analogs<!>Synthesis of the Designed Analogs<!>Biological Evaluations of the (+)-Discodermolide/Paclitaxel Hybrids<!>Experimental details for obtaining IC50 values<!>Experimental details for tubulin polymerization studies<!>Experimental details for synthesis and characterization<!>Synthesis of (+)-3a<!>Synthesis of (+)-3

<p>(+)-Discodermolide [(+)-1], a potent antitumor polyketide natural product, was first isolated in the early 1990's from extracts of the Caribbean marine sponge Discodermia dissoluta by Gunasekera and coworkers.1 Confirmation of the assigned structure, as well as the absolute stereochemistry, was subsequently achieved by Schreiber and coworkers via total syntheses of the (−) enantiomer (1993), the natural (+) enantiomer (1994) and a number of novel analogs.2,3 Initially reported to be a potent immunosuppressive agent, (+)-discodermolide was later recognized to possess significant antiproliferative activity both in vitro 4a across the NCI panel of human cancer cell lines,5 including paclitaxel resistant cell lines,6 and in vivo.4b Like paclitaxel (2), (+)-discodermolide stabilizes tubulin thereby modulating microtubule dynamics that leads to cell arrest at mitosis.</p><p>As a result of the observed antitumor activity, (+)-discodermolide generated significant interest in the synthetic community: 13 total syntheses have been reported to date.7 Of particular significance, as recently recognized,8 our four generation strategies exploited an early example of the Negishi coupling protocol to access the C14–C15 olefin. 7e,9 Subsequently Novartis employed a similar cross-coupling tactic (i.e. Suzuki) to provide (+)-discodermolide for their Phase I clinical trials.10 Also early on, we both proposed 11 a solution conformation of (+)-discodermolide based on a combination of NMR data and the solid state X-ray structure and carried out a comprehensive structure activity relationship (SAR) study with Kosan Biosciences, Inc. 12,13</p><p>More recently we demonstrated that (+)-discodermolide both causes accelerated cell senescence,14 and binds to a region on β-tubulin close to the paclitaxel binding site.15 Of interest here (+)-discodermolide and paclitaxel act synergistically, not only in cell culture16 but also in an ovarian xenograft tumor model in nude mice.17 Subsequent studies, utilizing hydrogen/deuterium exchange and mass spectrometry, indicated that paclitaxel and discodermolide induce stability on opposite sides of the microtubule interface, suggesting a possible explanation for the synergy observed between these two drugs.18 Finally to decrease the pneumotoxicity of (+)-1, the presumptive cause leading to the failure of the Novartis Phase I clinical trial,10 possibly due to metabolites,19 we extended our analog program, wherein the metabolically less stable lactone portion of the molecule 13a–c and C14–C15 olefin,13d were either individually or together, replaced with isosteres. Pleasingly these analogs preserved the cytotoxicity at a level similar to (+)-discodermolide (1).</p><p>However despite the availability of extensive SAR data,12 as well as structural,1 computational and NMR studies,11,20 the binding mode of (+)-discodermolide on tubulin remains unclear. Towards this end, we report here two possible binding modes of (+)-discodermolide in conjunction with the design, synthesis and biological evaluation of (+)-discodermolide/paclitaxel hybrid molecules, possessing photoaffinity labels.</p><p>In our SAR studies, we demonstrated the following: (1) that changes in the substitution pattern of C(1), C(3) and C(7) centers of (+)-discodermolide are favorable to retention of potent cytotoxicity; (2) that variations at the C(2), C(14) and the carbamate carbons are tolerable; and (3) that changes within the remaining portions of (+)-discodermolide significantly decrease the cytotoxicity.12,13 Based on these observations, in conjunction with our earlier photoaffinity labeling study,12e,f we reasoned that the diene terminus of discodermolide resides in a hydrophobic/aromatic pocket, while the lactone moiety either occupies a binding pocket that can accommodate large groups, or more likely, is partially interfaced with the solvent. Further information gained from our earlier photoaffinity labeling study revealed that attachment of a large aromatic system (cf. benzophenone) and/or alkyl groups to the carbamate nitrogen did not significantly alter the observed activity.12f In fact, in two cases, comprising attachment of a p-dimethylaniline or p-benzophenone substituent at the carbamate nitrogen, a significant increase in tumor cell growth inhibitory activity was observed. We emphasize however that our earlier study was directed specifically at the preparation of photoaffinity labeled (+)-discodermolide analogs, with little or no attention focused on activity changes, except for the possibility of increased hydrophobicity and/or conformational preferences. A closer examination of the earlier carbamate analogs however reveals a strong activity dependency both on the nature of the aromatic rings and linker length. Based on these observations, we have now examined the scope and molecular basis for the observed increase in cytotoxicity of these analogs exploiting detailed computational analysis.</p><!><p>In 2001, we reported that the preferred major solution conformation of (+)-discodermolide (1) was similar to the solid state structure,1 notwithstanding the 15 rotatable σ bonds.11 In particular, variable temperature and co-solvent titration NMR studies provided evidence of intramolecular hydrogen bonds between the C(7)-OH and the C(1)-carbonyl, in conjunction with interactions between the carbamate nitrogen and the C(3) and C(7) hydroxyl groups. We postulated that syn-pentane interactions21 were responsible for the proposed turn geometry of the backbone. Importantly, biological studies of epimers bearing inverted stereogenicities at the C(7), C(11), C(16) and C(17) revealed significantly lower activities, which led us to suggest that the conformation of the molecular segment between C(7)–C(17) is conserved upon binding. These conclusions were later confirmed by studies from other research laboratories.20</p><p>To understand more fully the solution behavior of (+)-discodermolide, we recently initiated computational studies exploiting molecular dynamics (MD) simulations. Monitoring backbone torsional angles over a 1 microsecond simulation period in water revealed the flexible torsional angles. Of considerable interest, the lactone ring and the diene side chain moieties, were found to display more flexibility compared to the rest of the molecule, when the turn geometry [i.e., C(7)–C(16)] was maintained throughout the simulation.</p><p>We next turned to NMR experiments to refine the solution conformation. The DISCON (Distribution of Solution Conformations) program developed in our laboratory22 was employed to obtain a distribution of conformers based on NMR data. Earlier studies had revealed that (+)-discodermolide is more flexible in organic solvents.20a, 20b Based on this observation, acetonitrile was selected as the solvent for the NMR experiments, given that the aim of this study was to identify the "flexible" torsional angles and possible diverse conformational families for our pending docking studies. Proton-proton coupling constants and NOE derived distances reported earlier were utilized.11 Deconvolution of the NMR observables over an ensemble of structures obtained by Monte Carlo conformational searches, followed by the clustering analysis that is embedded in DISCON software, led to an ensemble of four principal conformers, that in combination fit the NMR derived distance and torsional angle data better than any single conformation. Two conformers (Figure 2, A and B), similar to the solid-state structure with differences in the lactone ring orientation, were identified as the major solution conformations (A: 35% and B: 32%), an observation fully consistent with our earlier solution structure.11 Further examination of members of the major conformational families revealed (+)-discodermolide can adopt different conformations by adjusting the torsional angles around C(11)–C(12) and C(7) and C(8) bonds. In addition, the lactone ring can exist in equilibrium between chair, half chair and skew boat conformations. The next most populated conformer (C: 26%) comprised an inverted orientation at the diene region, wherein a different gauche orientation is preferred around the C(16)–C(17) and C(18)–C(19) bonds, with the remaining regions of (+)-discodermolide identical to the solid state structure. The remaining major family (7%) comprises both of these conformational changes, and thereby forces the molecule to occupy an extended conformation. The distribution of the conformations can be rationalized by the observed 3JH16–H17 and 3JH11–H12 values of 6.2 Hz and 6.6 Hz respectively, corresponding to torsional angle averaging and by the observed NOE correlation variations from the solid state structure, in particular those between H(7)–H(9), H(8)–H(12), H(9)–H(11), and H(12)–Me(18). Similar analyses reported earlier by the Jiménez-Barbero20b and the Snyder Laboratories20c are in accord with these observations, wherein differences in distributions were observed, presumably due to the differences in the solvents employed in the NMR experiments, as well as to the force fields and clustering methods employed.</p><p>In summary, (+)-discodermolide in solution displays considerable flexibility in the orientation of the lactone and the diene/carbamate termini, with conservation of the turn region of the conformation comprising the C(7)–C(16) centers. For docking studies, we reasoned that the solution structures hold considerable importance, since the bioactive conformation is more likely to exist in solution given that a lower energy penalty is paid for ligand reorganization upon binding (i.e., pre-organization).23 Stated differently, the 3D binding pharmacophore of the ligand is likely to be found in the conformation(s) that exist in solution, thus permitting easy recognition.24</p><!><p>Having an understanding of the available major ensemble of solution conformations, we turned to define the binding mode of (+)-discodermolide on microtubules. In general, docking studies of microtubule stabilizing agents utilize the heterodimer α-β tubulin structure (PDB code: 1JFF) obtained by high resolution electron microscopy from the Zn induced crystalline sheets of tubulin in the presence of paclitaxel.25 Although the tubulin structure in the crystal was shown to match the microtubule structure, it remains only an approximate model for docking studies for the following reasons: (1) the protein structure is distorted by the bound paclitaxel, and thus does not necessarily represent the unbound microtubule;26 (2) in vivo orientations and conformations of tubulin in microtubules and Zn-induced tubulin crystal sheets are different;27 (3) the flexible M-loop region, at the taxane binding site is not well resolved;25 (4) the lateral contacts between protofilaments, which are important for microtubule formation and tubulin dynamics28 are missing; and (5) in humans the tubulin isotypes have different sequences than found in the reported protein structure.29 Equally significant, the majority of reported (+)-discodermolide docking studies employ either the solid state conformation of (+)-discodermolide as a rigid scaffold, or (+)-discodermolide is treated as a fully flexible molecule without regard to the accessible solution conformations.15,18,30 Finally, earlier docking studies did not take into account the reported SAR data,19 in particular the potent carbamate analogs.13f</p><p>More recently Carlomagno20a and then Jiménez-Barbero et al. 20b concluded, based on NMR transfer NOE studies of (+)-discodermolide and tubulin, that the binding conformation of (+)-discodermolide is identical to the solid state conformation with differences in the lactone ring portion, and that (+)-discodermolide binds at the paclitaxel binding site. Paterson and coworkers in turn, based on these proposals, prepared a series of discodermolide-dictyostatin-paclitaxel hybrids which upon biological evaluation revealed a 20–35 fold decrease in tumor cell growth inhibition activity, employing the PANC-1 human cancer cell lines.31 Notwithstanding these disappointing observations, we undertook docking studies employing the solution conformations identified by DISCON (vide infra).</p><p>Our 2009 hydrogen-deuterium exchange experiments also revealed that (+)-discodermolide binds to a region quite close to the paclitaxel site on β-tubulin, however the stabilized microtubule geometries were different as evidenced by the observed solvent exposure of the lateral sites,18 with the most important interactions observed in the M-loop region of β-tubulin. Based on this information, we also chose to employ the paclitaxel site for docking studies. Beginning with the paclitaxel bound protein structure (PDB Code: 1JFF), paclitaxel was removed and the remaining protein structure was minimized with the OPLS-2005 force field to remove possible steric clashes. The binding site was not altered excessively, as we did not want to dock to a computationally generated binding pocket. Sets of (+)-discodermolide conformations, belonging to the families identified in solution (A–D, Figure 2), were then docked, employing the Glide32 and Molegro33 software packages, treating the (+)-discodermolide conformations as rigid structures and the protein structure as flexible. Docking modes inconsistent with the SAR data were eliminated by visual inspection. We reason that this approach provides more valid criteria for finding a working binding hypothesis than calculated docking scores and energies, due to the uncertainties (vide infra) in the protein structure (cf. Supporting Information for elimination criteria). Two binding poses for (+)-discodermolide were identified to be consistent with the SAR data.</p><p>The first pose (Figure 3B) orients the lactone end of (+)-discodermolide in a region occupied by the baccatin core of paclitaxel. The turn structure of (+)-discodermolide fits tightly in a region under the M-loop (β212–230) with the C(3) and C(7) hydroxyl groups forming hydrogen bonds to Pro272 and Thr274. In this pose, the carbamate moiety is directed towards the aromatic pocket that accommodates the paclitaxel tail (Figure 3A). The diene side chain and the lactone carbonyl group are directed towards the solvent, where additional substituents would appear to be tolerated. Rescoring of this pose via the MM/GBSA protocol34 provided a favorable ΔG value of −12 kcal/mol. The second pose (Figure 3C), similar to pose 1, utilizes the (+)-discodermolide turn scaffold to fit in the hydrophobic pocket under the M-loop. This pose however orients the lactone moiety to the region where the paclitaxel C(2) benzoate group binds. This region would also appear to be able to accommodate larger groups, while maintaining the activity of (+)-discodermolide analogs. That is, replacement of the lactone ring with other moieties might prove feasible. In this pose the known decrease in activity of (+)-discodermolide analogs upon acetylation at the C(11) hydroxyl group can be rationalized by the loss of a hydrogen bond to Thr274. This pose also orients the diene side-chain in line with the paclitaxel benzamide group, with the (+)-discodermolide carbamate group directed to the C(3′)-phenyl pocket. Rescoring employing the MM/GBSA protocol again provided a favorable ΔG value of −18 kcal/mol. Noteworthy, the Jiménez-binding mode differed significantly from the poses presented here, and was found to be inconsistent with our SAR criteria.</p><!><p>Based on the (+)-discodermolide docking modes, we designed a series of analogs that we reasoned would hold the promise of not only confirming our binding hypothesis, based initially on enhanced binding to tubulin and in turn enhanced cytotoxicity, but also defining the true binding pose. Importantly, the design criteria were orchestrated to accommodate incorporation of photoaffinity labels for future structural studies.</p><p>Utilizing docking poses 1 and 2, we envisioned filling the aromatic pocket in the vicinity of the carbamate group, which in the case of paclitaxel is occupied by C(3′)-Ph group. In silico docking of our earlier carbamate analogs (vide infra), employing the identified (+)-discodermolide poses, revealed that the earlier analogs did not fully occupy the proposed pocket due to the shorter tether lengths. To test the validity of this scenario, analogs having aromatic groups attached to the carbamate group with differing tether lengths were generated. We hypothesized that if the binding pocket is valid, the cytotoxic activities should vary, with the analog possessing the most suitable linker length displaying the highest tumor cell growth inhibition activity. A second amide linkage was incorporated onto the tether to participate in a possible hydrogen bond, with Asp26 and His229 of β-tubulin, similar to the paclitaxel amide linkage. Finally, we envisioned that photoaffinity tags could be appropriately introduced without deleterious effects on the cytotoxicity. To this end (+)-discodermolide/paclitaxel hybrid structures with a p-azidophenyl substituent (3–5, Figure 4) were generated. Docking studies were then conducted similar to the procedure applied for (+)-discodermolide. In these studies the newly introduced paclitaxel C(3′)-phenyl group surrogates were permitted to be flexible, whereas the (+)-discodermolide core was kept rigid. Examination of the docking scores identified hybrid 4 to fill the aromatic pocket. Two additional hybrids 6 and 7 possessing different photoaffinity tags and optimal linker lengths were also designed, to explore the potential effects of the hybrid structure motif, as well as the photoaffinity groups on the cytotoxicity.</p><!><p>The requisite hybrids, carrying the photoaffinity tags, were envisioned to arise via union of known (+)-discodermolide intermediate 19,7e prepared in connection with our gram-scale synthesis,11 with the photoaffinity units bearing the suitable amine linkers (Scheme 1). One of the attributes of developing preparative scale synthesis of architecturally complex natural products possessing significant bio-regulatory properties is the subsequent ready availability of advanced intermediates for SAR and photoaffinity labeling studies.</p><p>To permit facile formation of carbamate linkage, the secondary alcohol on (+)-discodermolide would be activated as a carbonate ester (cf. 20). The amine coupling partners were envisioned to arise via condensation reactions of the photoaffinity tags possessing carboxylic acid functionality with the diamine tethers of suitable length. The photoaffinity tags for this study were prepared following known synthetic routes,35,36,37 wherein acids 8 and 13 were activated as hydroxysuccinamide esters,38 followed by condensation with the alkyl diamine linkers (Scheme 2). For the aryl-(trifluoromethyl)diazirine component 18, the mono-Boc protected amine linker was employed with N,N′-Carbonyl-diimidazole (CDI)39 to reduce possible side reactions; the Boc group was subsequently removed under acidic conditions.</p><p>In the event, construction of (+)-discodermolide/paclitaxel hybrids 3–7 possessing the phototags entailed a two-step protocol, wherein the secondary alcohol in (+)-19 was converted to a p-nitrophenylcarbonate ester, followed by union with amines 10–12 and 15 and 18.40 Global deprotection under acidic conditions provided (+)-3–(+)-7 in good to excellent yields. Importantly, the synthetic route not only provides facile access to the designed (+)-discodermolide/paclitaxel hybrids possessing the photoaffinity tags, but was also designed to permit ready access to either 3H or 14C analogs for future labeling experiments.</p><!><p>The designed (+)-discodermolide/paclitaxel hybrids possessing the photoaffinity labels were tested for antiproliferative activity against human lung (A549) and breast (MCF-7) cancer cell lines. Discodermolide (−1) and Taxol (−2) served as controls. In accord with our design strategy, the analogs revealed improved activity in a trend changing with the tether length (Table 1).</p><p>Given the similarity of the analogs, the up to an eight-fold increase in potency is believed to be associated with an interaction with the proposed aromatic pocket, and not simply to an increase in hydrophobicity or other factors. Also of importance, (+)-discodermolide and each of the six analogs increased the amount of microtubule polymerization compared to the control that had no drug added (Fig. 6). Taken together these observations strongly suggest that all of the discodermolide molecules have the same mechanism of action. In summary, we have defined two possible binding modes for (+)-discodermolide in the paclitaxel binding pocket. To test the validity of the proposed modes, we designed and synthesized a series of hybrid molecules, wherein the aromatic ring of paclitaxel was attached to (+)-discodermolide, exploiting a tether of suitable length. Biological testing demonstrated that hybrid congeners possess increased potency compared to (+)-discodermolide. Given the availability of the already incorporated phototags, the discodermolide/paclitaxel hybrids hold the promise of defining the (+)-discodermolide binding site and mode of action, as well as determining whether the drug demonstrates selectivity for any of the 7 β-tubulin isotypes. Since certain isotypes, particularly βIII-tubulin, have been associated with paclitaxel resistance,41 an investigation of the interaction between drugs and specific tubulin isotypes becomes an important goal, and in particular definition of human tubulin isotypes that bind (+)-discodermolide. Such studies in conjunction with radiolabeling experiments are ongoing in our laboratories.</p><!><p>Cells lines were obtained from the American Type Culture Collection (ATCC) and maintained in RPMI medium supplemental with 10% fetal bovine serum. Cells (800 A549 or 2000 MCF-7) were added to each well of a 96 well plate. Increasing concentrations of the indicated drugs were added 18 h after plating. IC50 values, the concentration of drug that inhibits cell growth by 50%, were determined after 72 h or 96 h incubation at 37 °C for A549 or MCF-7 cells, respectively, using the SRB method (40). The two cell lines were fixed and stained at different times to take into account their different growth rates.</p><!><p>Bovine brain tubulin (2.5 μM) from Cytoskeleton Inc. in a buffer containing 0.1 M MES, 1 mM EGTA, 1 mM MgCl2 was incubated with 3 μM of the indicated drugs, in the presence of 2 mM GTP, at 37 °C for 30 min. Samples were centrifuged at 120,000 × g at 37 °C for 1 h and the pellet that contains the polymerized tubulin was dissolved in SDS sample buffer and analyzed by SDS-polyacrylamide gel electrophoresis. Proteins were then transferred to nitrocellulose and stained with Ponceau S.</p><!><p>Except as otherwise indicated, all reactions were run under an argon atmosphere in flame- or oven-dried glassware, and solvents were freshly distilled. The argon was deoxygenated and dried by passage through an OXICLEAR™ filter from Aldrich and Drierite tube, respectively. Diethyl ether (Et2O), tetrahydrofuran (THF) and dichloromethane (CH2Cl2) were purchased from Aldrich (HPLC purity) and further purified by Pure Solve™ PS-400. All other reagents were purchased from Aldrich or Acros and used as received. Reactions were monitored by thin layer chromatography (TLC) either with 0.25 mm Silicycle or 0.25 mm E. Merck (Kieselgel 60F254, Merck) pre-coated silica gel plates. Silica gel for flash chromatography (particle size 0.040–0.063 mm) was supplied by Silicycle or Sorbent. Yields refer to chromatographically and spectroscopically pure compounds unless otherwise noted. 1H and 13C spectra were recorded on a Bruker AMX-500 spectrometer. Chemical shifts are reported as δ values relative to internal chloroform (δ 7.26) or benzene (δ 7.15) for 1H and either chloroform (δ 77.0) or benzene (δ 128.0) for 13C. Infrared spectra were recorded on a Jasco FTIR-480plus spectrometer. Optical rotations were measured on a Jasco P-2000 polarimeter in the solvent indicated. High resolution mass spectra were measured at the University of Pennsylvania Mass Spectrometry Center on either a VG Micromass 70/70H or VG ZAB-E spectrometer. Analytical reverse-phased (Sunfire C18; 4.6mm 50mm, 5mL) high-performance liquid chromatography (HPLC) was performed with a Waters binary gradient module 2525 equipped with Waters 2996 PDA and Waters micromass ZQ. All final compounds were analyzed employing a linear gradient from 10% to 90% of acetonitrile in water over 8 min and a flow rate of 1 mL/min, and unless otherwise stated, the purity level was >95%.</p><!><p>Compound (+)-20 (80 mg, 0.0725 mmol) was dissolved in CH2Cl2 (1 mL) and MeOH (1 mL). N-(4-Azido-benzoyl)-1,2-ethanediamine (10) (45 mg, 0.218 mmol) in MeOH (1 mL) and Et3N (40 μL, 0.290 mmol) were added to the above solution. The resultant solution was stirred at room temperature for 6 days. Water (50 mL) was added to the reaction mixture and extracted with CH2Cl2 (3 × 50mL). The combined organic layers were washed with brine (50 mL), dried over MgSO4, filtered and concentrated. Flash chromatography (11 to 25% EtOAc in hexane) provided compound (+)-3a (75.1 mg, 89% yield) as a colorless amorphous solid. [α]22D = +36.0° (c = 0.65, CHCl3). IR (film, NaCl): 3336, 2957, 2929, 2857, 2123, 1718, 1653, 1604, 1539, 1500, 1254, 1044, 836, 775 cm−1. 1H NMR (500 MHz, CDCl3) δ: 7.81 (d, J = 8.3 Hz, 2H), 7.33 (brs, 1H), 7.03 (d, J = 8.4 Hz, 2H), 6.54 (ddd, J = 16.7, 10.6, 10.6 Hz, 1H), 5.88 (apparent t, J = 10.9 Hz, 1H), 5.35 - 5.13 (m, 5H), 5.11 (d, J = 10.2 Hz, 1H), 5.02 (d, J = 9.9 Hz, 1H), 4.81 (apparent t, J = 9.1 Hz, 1H), 4.74 (apparent t, J = 5.8 Hz, 1H), 4.59 (ABq, JAB = 6.9 Hz, ΔAB = 28.4 Hz, 2H), 4.50 (apparent t, J = 10.3 Hz, 1H), 3.63 (brs, 1H), 3.56 - 3.44 (m, 3H), 3.41 (apparent t, J = 4.0Hz, 1H), 3.39 - 3.30 (m, 1H), 3.34 (s, 3H), 3.06 (apparent t, J = 5.6 Hz, 1H), 2.99 - 2.91 (m, 1H), 2.75 - 2.66 (m, 1H), 2.60(qd, J = 7.5, 2.8 Hz, 1H), 2.55 - 2.47 (m, 1H), 2.06 (apparent t, J = 12.7 Hz, 1H), 1.91 - 1.76 (m, 3H), 1.75 - 1.66 (m, 2H), 1.63 - 1.54 (m, 1H), 1.57 (s, 3H), 1.23 (d, J = 7.6 Hz, 3H), 0.96 (d, J = 6.7 Hz, 3H), 0.95 - 0.84 (m, 12H), 0.91 (s, 9H), 0.87 (s, 9H), 0.85 (s, 9H), 0.72 (d, J = 6.7 Hz, 3H), 0.08 (s, 3H), 0.06 (s, 3H), 0.054 (s, 3H), 0.048 (s, 3H), 0.04 (s, 3H), 0.03 (s, 3H). 13C NMR (125 MHz, CDCl3) δ: 173.6, 166.7, 158.8, 143.4, 133.9, 133.4, 132.3, 132.1, 132.0, 131.1, 130.8, 130.0, 129.0, 119.1, 118.5, 97.5, 86.4, 79.5, 77.3, 77.0, 74.9, 64.9, 56.2, 44.2, 42.7, 42.5, 40.5, 37.9, 36.1, 35.8, 35.7, 34.6, 34.4, 34.3, 26.4, 26.1, 25.9, 23.3, 18.7, 18.3, 18.1, 17.7, 16.8, 16.7, 16.5, 14.4, 14.2, 10.2, −3.2, −3.5, −4.2, −4.3, −4.66, −4.68. High resolution mass spectrum (ESI+) m/z 1190.7408 [(M+Na)+; calcd for C62H109N5O10Si3Na: 1190.7380].</p><!><p>To a solution of compound (+)-3a (67 mg, 0.0573 mmol) in MeOH (5 mL) was added aqueous hydrochloric acid (4M, 4 mL) in 100 – 250 μL portions over 7 h at a rate which minimized precipitation (ca. 8 to 30 min intervals) and the sides of flask were rinsed with MeOH (1 mL). The reaction mixture was stirred at room temperature for 18 h, diluted with EtOAc (70 mL). The resulting solution was neutralized with NaHCO3 (1.35 g in H2O 18 mL) at 0 °C. Phosphate buffer pH7 (1M, 30 mL) and NaCl (17 g) were added and the aqueous layer was extracted with EtOAc (3 × 70 mL). The combined organic layers were washed with brine (100 mL), dried over MgSO4, filtered and concentrated. The residue was purified three times by flash chromatography (80% EtOAc in hexane or 3 – 6% MeOH/CH2Cl2) to afford compound (+)-3 (31.1 mg, 69% yield) as a colorless amorphous solid. [α]22D = +17.4° (c = 0.10, CHCl3). IR (film, NaCl): 3359, 2968, 2123, 1699, 1639, 1604, 1543, 1501, 1457, 1284, 1120, 1032, 731 cm−1. 1H NMR (500 MHz, CDCl3) δ: 7.81 (d, J = 8.5 Hz, 2H), 7.26 (br, 1H), 7.06 (d, J = 8.4 Hz, 2H), 6.55 (ddd, J = 16.6, 10.6, 10.6 Hz, 1H), 5.84 (apparent t, J = 11.0 Hz, 1H), 5.54 - 5.41 (m, 3H), 5.26 (t, J = 10.4 Hz, 1H), 5.17 - 5.05 (m, 3H), 4.76 - 4.66 (m, 2H), 4.61 (apparent t, J = 9.3 Hz, 1H), 3.70 (br m, 1H), 3.57 - 3.44 (m, 2H), 3.43 - 3.31 (m, 2H), 3.29 - 3.16 (m, 3H), 3.00 - 2.91 (m, 1H), 2.83 - 2.74 (m, 1H), 2.68 (qd, J = 7.3, 4.6 Hz, 1H), 2.63 - 2.51 (m, 2H), 2.40 - 2.20 (br, 2H), 2.00 - 1.74 (m, 6H), 1.69 - 1.62 (m, 1H), 1.58 (s, 3H), 1.28 (d, J = 7.3 Hz, 3H), 1.06 (d, J = 6.9 Hz, 3H), 1.01 (d, J = 6.8 Hz, 3H), 0.93 (apparent d, J = 6.8 Hz, 6H), 0.90 (d, J = 6.6 Hz, 3H), 0.82 (d, J = 6.5 Hz, 3H). 13C NMR (125 MHz, CDCl3) δ: 174.6, 167.2, 158.4, 143.6, 134.5, 133.9, 133.6, 133.0, 132.3, 130.6, 130.0, 129.6, 129.1, 119.2, 118.3, 79.4, 78.9, 77.4, 75.2, 73.1, 64.4, 43.3, 42.0, 41.2, 40.5, 37.4, 36.2, 36.0, 35.94, 35.91, 34.9, 33.0, 23.4, 18.8, 17.5, 16.7, 15.7, 14.2, 12.8, 9.3. High resolution mass spectrum (ESI+) m/z 804.4495 [(M+Na)+; calcd. for C42H63N5O9Na: 804.4523].</p>

PubMed Author Manuscript

Muraymycin nucleoside-peptide antibiotics: uridine-derived natural products as lead structures for the development of novel antibacterial agents

Muraymycins are a promising class of antimicrobial natural products. These uridine-derived nucleoside-peptide antibiotics inhibit the bacterial membrane protein translocase I (MraY), a key enzyme in the intracellular part of peptidoglycan biosynthesis. This review describes the structures of naturally occurring muraymycins, their mode of action, synthetic access to muraymycins and their analogues, some structure–activity relationship (SAR) studies and first insights into muraymycin biosynthesis. It therefore provides an overview on the current state of research, as well as an outlook on possible future developments in this field.

muraymycin_nucleoside-peptide_antibiotics:_uridine-derived_natural_products_as_lead_structures_for_t

<!>Introduction<!>Structures of naturally occurring muraymycins<!><!>Structures of naturally occurring muraymycins<!><!>Structures of naturally occurring muraymycins<!><!>Structures of naturally occurring muraymycins<!>Mode of action<!><!>Mode of action<!><!>Mode of action<!><!>Mode of action<!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!><!>Synthetic access<!>Structure–activity relationship studies<!><!>Structure–activity relationship studies<!><!>Structure–activity relationship studies<!><!>Structure–activity relationship studies<!><!>Structure–activity relationship studies<!><!>Structure–activity relationship studies<!><!>Structure–activity relationship studies<!>Biosynthesis<!><!>Biosynthesis<!><!>Biosynthesis<!><!>Biosynthesis<!>Conclusion

<p>This article is part of the Thematic Series "Natural products in synthesis and biosynthesis II" and is dedicated to Professor Wittko Francke on the occasion of his 75th birthday.</p><!><p>The treatment of infectious diseases caused by bacteria is a severe issue. With multiresistant bacterial strains rendering well-established therapeutic procedures ineffective, the exploration of novel antimicrobial agents is of growing significance. The discovery of penicillin [1] and the proof of its in vivo efficacy [2] marked the starting point for the research on antibacterial drugs during the so-called "golden age" of antibiotics. Despite the early occurrence of first resistances [3–5], an innovation gap followed from the 1960s onwards, during which only few antibiotics were introduced into the market. Most of them were modifications of established substances already in clinical use. Current and future developments will have to consider these improved 2nd and 3rd generation antibiotics [6] alongside the search for completely unknown structures. For such novel agents, natural products appear to be a promising source [7–9].</p><p>Bacteria deploy different mechanisms to escape the toxic effect of an antibacterial drug [10–12]. These include the structural modification and degradation of a drug, as it is reported for aminoglycoside-modifying proteins [13], and alteration of the drug target, as can be found in macrolide-resistant bacteria that contain mutations in the bacterial ribosome [14]. Further mechanisms are an increased efflux [15] and a change in permeability of the cell wall [16–17]. Due to the evolutionary pressure exerted by antibiotics, bacteria featuring the aforementioned mutations survive, proliferate and may even develop resistances against multiple drug classes. Excessive application of antibiotics fuels the emergence of multiresistant strains such as hospital and community-associated methicillin-resistant Staphylococcus aureus (MRSA) [18–19] and vancomycin-resistant Enterococcus (VRE) [20]. This development raises the demand for antibiotics exploiting yet unused modes of action. Potential targets within bacteria include peptidoglycan biosynthesis, protein biosynthesis, DNA and RNA replication and folate metabolism [21].</p><p>Promising candidates meeting the requirements for new drugs are nucleoside antibiotics, i.e., uridine-derived compounds that address the enzyme translocase I (MraY) as a novel target, thereby interfering with a membrane-associated intracellular step of peptidoglycan biosynthesis. This review will focus on muraymycins as a subclass of nucleoside antibiotics, covering their mode of action, synthetic approaches as well as SAR studies on several derivatives. Furthermore, first insights into the biosynthesis of these Streptomyces-produced secondary metabolites will be discussed.</p><!><p>The muraymycins were first isolated in 2002 from a broth of a Streptomyces sp. [22]. McDonald et al. discovered and characterised 19 naturally occurring muraymycins (Figure 1). These compounds belong to the family of nucleoside antibiotics which have a uridine-derived core structure in common. Their antibiotic potency is based on the inhibition of MraY, thereby blocking a membrane-associated intracellular step of bacterial cell-wall biosynthesis. The structure elucidation was carried out using one- and two-dimensional NMR experiments as well as FT mass spectrometry [22].</p><!><p>Structures of the naturally occurring muraymycins isolated by McDonald et al. [22].</p><!><p>Muraymycins have a glycyl-uridine motif, which is connected via an aminopropyl linker to a urea peptide moiety consisting of L-leucine or L-hydroxyleucine, L-epicapreomycidine (a non-proteinogenic cyclic arginine derivative) and L-valine. The uridine structure is glycosylated in its 5'-position with an aminoribose unit and in some cases a lipophilic side chain is attached to the hydroxyleucine residue. The 19 compounds are divided into four different series (A–D) which mainly vary in the leucine residue and the lipophilic side chain or the amino sugar (Figure 1). The aminoribose is missing in muraymycins A5 and C4, which may eventually be hydrolysis products. The series A and B have lipophilic side chains with varying chain lengths, which are either ω-functionalised with a guanidino or hydroxyguanidino-function in case of series A or unfunctionalised but terminally branched in case of series B. Muraymycins of series C contain unfunctionalised L-hydroxyleucine while in series D proteinogenic L-leucine occurs instead.</p><p>Muraymycin A1 is one of the most active members of this family and shows good activity mainly against Gram-positive (Staphylococcus MIC: 2–16 μg/mL, Enterococcus MIC: 16–64 µg/mL) but also a few Gram-negative bacteria (E. coli MIC: down to 0.03 μg/mL). Since the activity against wild-type E. coli is clearly lower (MIC > 128 μg/mL) [22], it is assumed that this might be an effect resulting from low membrane permeability.</p><p>There are other naturally occurring nucleoside antibiotics which address the same biological target, thereby inhibiting peptidoglycan biosynthesis. Figure 2 shows the structures of selected other classes of nucleoside antibiotics, with structural similarities being highlighted. A broad overview of antimicrobial nucleoside antibiotics blocking peptidoglycan biosynthesis is given by Bugg et al. in two review articles [23–24] and by Ichikawa et al. in a recent review [25].</p><!><p>Structures of selected classes of nucleoside antibiotics. Similarities to the muraymycins are highlighted in different colours.</p><!><p>Representing the first discovered nucleoside antibiotics, the tunicamycins were isolated in 1971 from Streptomyces lysosuperficus nov. sp. by Takatsuki and Tamura et al. [26–28]. They contain a uridine moiety, two O-glycosidically linked sugars, the so-called tunicamine and a fatty acid moiety, which typically is terminally branched and unsaturated. Two closely related nucleoside antibiotics were isolated later on and named streptoviridins (isolated in 1975 from Streptomyces griseoflavus subsp. thuringiensis [29–31]) and corynetoxins (isolated in 1981 from Corynebacterium rathayi [32]). These classes have merely the uracil nucleoside core structure in common with the muraymycins and the terminally branched lipophilic side chain resembles the acyl moiety in muraymycins of group B.</p><p>Capuramycin, a nucleoside antibiotic isolated in 1986 from Streptomyces griseus, shares the uracil-derived nucleoside moiety with the muraymycins [33–34]. The antibiotic FR-900493, which is structurally closely related to muraymycins, was isolated from Bacillus cereus and characterised in 1990 [35]. In comparison to the muraymycins, only the urea peptide moiety and the lipopeptidyl motif are absent.</p><p>The mureidomycins [36–38] and pacidamycins [39–41], both reported in 1989, the napsamycins (1994) [42] and the sansanmycins (2007) [43–44] are structurally closely related. They consist of a 3'-deoxyuridine unit with a unique enamide linkage and the non-proteinogenic N-methyl-2,3-diaminobutyric acid, which branches into two peptide moieties. They differ in the amino acid residues AA2, AA4 and AA5, with AA2 and AA5 being aromatic in all four classes. The amino acid residue AA4 is either methionine for mureidomycins, napsamycins and sansanmycins or alanine in case of pacidamycins. Remarkably, these natural products share a urea peptide motif with the muraymycins. They are mainly active against Gram-negative bacteria, which is a noteworthy difference to the muraymycins and other related nucleoside antibiotics.</p><p>The liposidomycins (isolated in 1985) [45] and the related caprazamycins (isolated in 2003) [46–47] have a unique diazepanone ring, and in case of the caprazamycins a permethylated rhamnose residue. They resemble the muraymycins in their uridine-derived core structure, which is also glycosylated in 5'-position with an aminoribose unit, and they contain a fatty acid moiety as well. Caprazamycins also display noteworthy antimicrobial activity against M. tuberculosis as well as most Gram-positive bacteria (Table 1) [46,48].</p><!><p>Comparison of the antimicrobial activities of selected representative compounds of different classes of nucleoside antibiotics against selected bacterial species.a</p><p>a++: good activity (MIC < 10 μg/mL), +: moderately active (10 μg/mL < MIC < 32 μg/mL), −: no notable activity (MIC > 32 μg/mL), n.r.: not reported. bNot active against wild-type E. coli. cActive against M. phlei.</p><!><p>All aforementioned nucleoside antibiotics address the same biological target and most likely have the same mode of action by inhibiting MraY (see below), but their in vitro activity differs significantly. It is important to notice that a comprehensive comparison of minimum inhibitory concentrations (MIC values) is difficult because naturally occurring nucleoside antibiotics have been tested against different bacterial strains. However, synthetic analogues of the nucleoside antibiotics listed in Table 1 have been tested against some of the listed bacterial species. It can therefore be assumed that the parent natural products display similar activities even though there are no data available. Furthermore, the activity of a compound against different strains of a bacterial species can vary. Nonetheless, there are certain trends and differences that can be observed. Muraymycin A1 is mainly active against Gram-positive bacteria such as S. aureus or E. faecalis, but also against some Gram-negative E. coli strains [49]. Tunicamycin, capuramycin and FR-900493 only show antimicrobial activity against Gram-positive strains. For mureidomycin C (R5 = Gly, AA2 = AA5 = m-Tyr, AA4 = Met, B = uracil, see Figure 2) as a representative compound, no activity against Gram-positive bacteria was observed, but it displayed pronounced antibacterial activity against P. aeruginosa. This remarkable finding distinguishes the mureidomycins, pacidamycins, sansanmycins and napsamycins from other nucleoside antibiotics. On the other hand, caprazamycin B shows good activity against Gram-positive bacteria, Pseudomonas and M. tuberculosis [48]. The related liposidomycins display good activity against M. phlei, while they are not active against a range of other bacteria [45].</p><!><p>To develop an effective antibiotic one needs to choose a target that is essential for bacterial survival or growth and offers selectivity to strike only bacterial cells (without cytotoxicity to human cells). There are mainly four classical target processes for antibiotics: bacterial cell wall biosynthesis, bacterial protein biosynthesis, DNA replication and folate metabolism [21]. Novel approaches that differ from these established modes of action are under investigation, but many new compounds in development still address bacterial cell wall biosynthesis. They are accompanied by a rich variety of prominent antibiotics in clinical use such as the penicillins [23,50–51]. All bacteria, i.e., Gram-positive and Gram-negative congeners, have a cell wall as part of their cell envelope. While its thickness differs among bacteria – Gram-positive strains usually have a thicker cell wall relative to Gram-negative ones – the principle molecular structure remains identical: Bacterial cell walls consist of peptidoglycan, a heteropolymer with long chains of alternating units of N-acetylmuramic acid (MurNAc) and N-acetylglucosamine (GlcNAc) that are cross-linked through peptide chains attached to the muramic acid sugar (Figure 3) [52].</p><!><p>Structure of peptidoglycan. Long chains of glycosides (alternating GlcNAc (green) and MurNAc (blue)) are cross-linked through the MurNAc peptide chain. The exact composition of the peptide chain varies among different bacterial species.</p><!><p>The biosynthesis of peptidoglycan is illustrated in Figure 4 and has been described in detail in several reviews (e.g., [51,53–57]). It can be divided into three parts: first, the formation of the monomeric building blocks in the cytosol (Figure 4, step A); second, the membrane-bound steps with the attachment to the lipid linker, transformation to a disaccharide and transport to the extracellular side of the membrane (Figure 4, steps B, C); finally, polymerisation to long oligosaccharide chains and cross-linking occur (Figure 4, steps D, F).</p><!><p>Schematic representation of bacterial cell wall biosynthesis.</p><!><p>In the cytosol, uridine diphosphate-N-acetylglucosamine (UDP-GlcNAc), that is formed from fructose-6-phosphate in four steps, is transformed into UDP-MurNAc-pentapeptide in a number of enzyme-catalysed reactions (Figure 4, step A). The exact composition of the peptide chain varies in different organisms. Examples given in Figure 3 are frequently occurring ones and a more comprehensive list has been reported elsewhere [52].</p><p>The membrane-associated steps commence with the transfer of UDP-MurNAc-pentapeptide to the lipid carrier undecaprenyl phosphate, catalysed by translocase I (MraY), to give lipid I (Figure 4, product of step B). The glycosyltransferase MurG attaches a GlcNAc sugar to furnish lipid II (Figure 4, product of step C). This building block is then transported to the extracellular side of the membrane. It is speculated that there might be some kind of 'flippase' involved but this particular step is still unclear and requires further investigation [55]. On the extracellular side of the membrane, the building blocks are connected by transglycosylases to form long chains (Figure 4, step D) and then are cross-linked by transpeptidases (Figure 4, step E). Both enzymes are members of the family of penicillin-binding proteins [23].</p><p>As mentioned above, there are many antibiotics in clinical use that target at least one step of bacterial cell wall biosynthesis. Prominent examples besides penicillins are cephalosporins, cycloserine, vancomycin, fosfomycin and daptomycin [9]. All of them (except fosfomycin and cycloserine) inhibit late, extracellular steps of cell wall formation. Thus, there are still many steps not addressed by clinically used drugs, which implies that cell wall biosynthesis still offers promising novel targets for the development of antibiotics with new modes of action. Muraymycins and other nucleoside antibiotics target translocase I (MraY) that represents such a potential novel molecular target [22].</p><p>Overexpression of the mraY gene, identified in an mra (murein region A) cluster, led to an increase of UDP-N-acetylmuramoyl-pentapeptide: undecaprenyl phosphate phospho-N-acetylmuramoyl-pentapeptide transferase activity [58]. Gene knockout experiments revealed the MraY-catalysed reaction in cell wall biosynthesis to be an essential process for bacterial viability and growth [59–63].</p><p>The chemical transformation catalysed by MraY is shown in Figure 5. The cytosolic precursor UDP-MurNAc-pentapeptide is linked to undecaprenyl phosphate, a C55-isoprenoid lipid carrier that is located in the cellular membrane. With concomitant release of uridine monophosphate (UMP), this furnishes a diphosphate linkage between the two substrates. The reaction is reversible and MraY accelerates the adjustment of the equilibrium state. Whereas this reaction was known for a long time [64–65], the structure of the MraY protein remained unclear. The mechanism of the MraY-catalysed reaction was investigated by kinetic studies by Heydanek, Neuhaus et al. in the 1960s. They proposed a two-step mechanism for lipid I formation that was later revised (Figure 6A) [55,66–69].</p><!><p>Translocase I (MraY) catalyses the reaction of UDP-MurNAc-pentapeptide with undecaprenyl phosphate towards lipid I.</p><p>Proposed mechanisms for the MraY-catalysed reaction. A: Two-step mechanism postulated by Heydanek et al. [66]; B: one-step mechanism postulated by Bouhss et al. [69].</p><!><p>The identification of the mraY gene [58] facilitated the alignment of MraY homologue sequences by van Heijenoort et al. and resulted in a two-dimensional topology model of MraY from E. coli, among others [70]. Bugg et al. identified three conserved residues with nucleophilic side chains within the superfamily of polyisoprenyl-phosphate N-acetyl hexosamine 1-phosphate transferases (PNPT). Mutation of these three aspartate residues (D115, D116 and D267 in the E. coli protein) resulted in a complete loss of catalytic activity. This led to a proposed model for the active site of MraY in accordance with previous findings [66]: D115 and D116 bind a Mg2+-cofactor, UDP-MurNAc-pentapeptide also binds the Mg2+-cofactor and D267 acts as a nucleophile within the proposed two-step mechanism (Figure 6) [68]. In a study with purified MraY from B. subtilis, Bouhss et al. found small remaining activity in the D231N mutant (corresponding to D267 in MraY from E. coli). They assumed that this would contradict the two-step mechanism as a nucleophilic residue is essential for the previously proposed mechanism. They found D98 to be crucial for activity and proposed its role to deprotonate undecaprenyl phosphate. This was speculated to be followed by a one-step nucleophilic attack of the C55-alkyl phosphate at the UDP-MurNAc-pentapeptide (Figure 6B) [69].</p><p>In 2013, Lee et al. reported an X-ray crystal structure (3.3 Å resolution) of MraY from Aquifex aeolicus (MraYAA) as the first structure of a member of the PNPT superfamily. MraYAA crystallised as a dimer and additional experiments showed that it also exists as a dimer in detergent micelles and membranes [71]. The previously proposed models are in agreement with the solved structure showing ten transmembrane helices and five cytoplasmic loops. The authors identified a cleft at the cytoplasmic side of the membrane that showed the highest conservation in sequence mapping. Furthermore, it is also the region where most of the previously identified, functionally important residues [69] are located [71]. The location and binding mode of the Mg2+ ion in the crystal does not support the proposed model for a two-step mechanism [68]. In experiments with Mn2+ exchange no interaction of the metal with D117 and D118 could be detected. Surface calculation of MraYAA showed an inverted U-shaped groove that could harbour the undecaprenyl phosphate co-substrate. The locations of this groove, the Mg2+ and D265 do at least not contradict the proposed one-step mechanism. Nevertheless, there is still a need for further studies to fully understand the MraY-catalysed reaction at the molecular level [71].</p><p>In the context of a different MraY inhibitor, i.e., lysis protein E from bacteriophage X174, Bugg et al. reported a different site of inhibition in pronounced distance to the proposed active site. It has been demonstrated before that mutation of phenylalanine 288 (F288L) in helix 9 of MraY caused resistance against lysis protein E [72–73]. An interaction between F288 and glutamic acid 287 (E287) with the peptide motif arginine-tryptophan-x-x-tryptophan (RWxxW, x represents an arbitrary amino acid) was found. Mutants F288L and E287A showed reduced or no detectable enzyme inhibition, thus indicating a secondary binding site for potential MraY inhibitors. Nevertheless, it remains unclear how binding at helix 9 can inhibit MraY function and further studies are probably inevitable [74].</p><p>In order to investigate the biological potencies of MraY inhibitors such as the muraymycins, in vitro assay systems are needed. A widely used and universal method to evaluate the in vitro activity of potential agents against certain bacteria is the determination of minimum inhibitory concentrations (MIC). MICs are defined as the lowest concentration at which a potential antimicrobial agent inhibits the visible growth of a microorganism [75]. They are easily determined and reflect several effects such as target interaction, cellular uptake and potential resistance mechanisms of the microorganism. MIC values are therefore widely used, also in studies on muraymycin analogues (e.g., [22,76–78]) and have been the basis of many structure–activity relationship studies (see below).</p><p>This bacterial growth assay, however, does not elucidate the inhibitory potency of the potential antimicrobial solely against the target protein MraY. Thus, another assay system that is not based on the interaction with whole cells but only with the target protein is required. For MraY, there are three different assays available that provide such inhibition data: i) a fluorescence-based and ii) a radioactivity-based assay as well as iii) a relatively new Förster resonance energy transfer (FRET)-based method.</p><p>The fluorescence-based assay was developed by Bugg et al. [79–80] and uses a fluorescently labelled (dansylated) analogue of the MraY substrate UDP-MurNAc-pentapeptide. The reaction of this substrate analogue with undecaprenyl phosphate leads to an increase in fluorescence intensity that can be used as a measure for enzymatic activity (e.g., [74,78]). The assay reported by Bouhss et al. [81] uses a radioactively labelled UDP-MurNAc-pentapeptide and thin layer chromatography (TLC) separation of undecaprenyl-linked MurNAc-pentapeptide from unreacted substrate (e.g., [77,82]). The third assay was introduced in 2012 by Shapiro et al. and uses a FRET system with the FRET donor attached to the UDP-MurNAc-pentapeptide and the FRET acceptor in a detergent or detergent/lipid micelle that also hosts the MraY protein [83].</p><p>The overexpression and purification of the transmembrane protein MraY is challenging. MraY from different bacterial strains was heterologous overexpressed in E. coli and was used in assays mentioned above as a crude cellular membrane preparation or as a detergent-solubilised membrane protein mixture [79,84]. A purification to homogeneity was reported for MraY from B. subtilis by Bouhss et al. in 2004 [81] and for the congener from Aquifex aeolicus by Lee et al. in 2013 [71]. Wang, Bernhard et al. achieved a cell-free production of MraY from B. subtilis and E. coli, also experiencing the need of pronounced adjustments in expression conditions [85].</p><!><p>Following the isolation of muraymycins [22], a group of scientists from Wyeth reported the semisynthetic access towards 16 derivatives of muraymycin C1 for structure–activity relationship (SAR) studies [86]. At the same time, a first set of fully synthetic structurally simplified muraymycin analogues was described [76]. Starting from uridine (1), protected uridine-5'-aldehyde 2 was prepared in four steps (Scheme 1) [87–88]. This was followed by an aldol reaction of aldehyde 2 with N,N-dibenzylglycine tert-butyl ester (3) [89] and LDA as a key step of the synthesis (Scheme 1). The resultant products were the two 5'-epimers 4 (5'R,6'S) in 31% yield and 5 (5'S,6'S) in 14% yield, which could be separated by column chromatography. After debenzylation, the resultant primary amines were connected with amido aldehydes 6 substituted with different moieties R and R' by reductive amination with R being either a hydroxy group or a hydrogen and R' representing an alkyl, allyl, ester or a protected amino moiety. This led to many truncated muraymycin analogues based on the structures 7 and 8 [76]. Cbz deprotection and subsequent peptide coupling with the L-arginine-L-valine-derived urea dipeptide 9 gave various full-length muraymycin analogues 10 and 11 [76]. Some of the truncated and the full-length compounds were able to inhibit lipid II formation. These active compounds are discussed in the section on structure–activity relationship (SAR) studies.</p><!><p>First synthetic access towards simplified muraymycin analogues as reported by Yamashita et al. [76].</p><!><p>In 2005, Ichikawa, Matsuda et al. reported the synthesis of (+)-caprazol [90–92] which contains the same uridine-derived core structure as the muraymycins. The latest and optimised synthesis is shown in Scheme 2 [92]. Oxidation of the isopropylidene-protected uridine 12 to the 5'-aldehyde and a Wittig reaction [93] gave olefin 13. The key step was a subsequent asymmetric Sharpless aminohydroxylation [94] furnishing (5'S,6'S)-nucleosyl amino acid 14 in 96% yield (98% de) [92,95]. A novel β-selective glycosylation of the 5'-hydroxy group was also established. Thus, 14 was reacted with the ribosyl fluoride 15 and BF3·Et2O, which afforded the glycosylated product 16 in 77% yield and with a β/α-selectivity of 24:1 [91–92]. This reaction was followed by an azide reduction, Boc protection, saponification of the ester, peptide coupling with the amino acid 17, oxidative cleavage of the double bond to give 18 and an intramolecular reductive amination in order to construct the seven-membered ring. Methylation with subsequent acidic global deprotection led to the target compound (+)-caprazol (19) [90,92].</p><!><p>Synthesis of (+)-caprazol (19) reported by Ichikawa, Matsuda et al. [92].</p><!><p>For the synthesis of muraymycins, Ichikawa, Matsuda et al. furthermore developed a new route towards the epicapreomycidine-containing urea dipeptide unit via C–H activation (Scheme 3) [96–97]. For this purpose, the commercially available δ-N-Boc-α-N-Cbz-L-ornithine (20) was transformed into sulfamate 21. Subsequently, the C–H insertion representing the key step of this synthesis was examined with two different catalysts and different reaction conditions. Despite different ratios in the outcome of the C–H insertion in favour of the unwanted diastereomer 22, the synthesis was finished with the desired minor component 23. Boc deprotection followed by reaction with guanidinylation reagent 24 gave bicyclic compound 25. The next steps included a desulfonylation and the reaction with 26 leading to protected epicapreomycidine-containing urea dipeptide 27 [96–97].</p><!><p>Synthesis of the epicapreomycidine-containing urea dipeptide via C–H activation [96–97].</p><!><p>Starting from the uridine derivative 28 used in the synthesis of (+)-caprazol, Ichikawa and Matsuda built up muraymycin D2 and its epimer (Scheme 4). They used an Ugi four-component reaction with an isonitrile derivative 29 obtained from the uridine-derived core structure 28, aldehyde 30, amine 31 and the urea dipeptide building block 27. A two-step global deprotection then gave the desired muraymycin D2 and its epimer which could be separated by HPLC [96–97].</p><!><p>Synthesis of muraymycin D2 and its epimer reported by Ichikawa, Matsuda et al. [96–97].</p><!><p>In 2012, Kurosu et al. also reported the synthesis of potential key intermediates for the total synthesis of muraymycins (Scheme 5) [98]. A fully protected ureidomuraymycidine tripeptide was prepared through lactone opening followed by urea formation and a final Mitsunobu ring closure as key steps. A Strecker reaction of the benzylimine 34 followed by several steps afforded the alcohol 35. A thermal lactonisation as a first key step of the synthesis led to a 1:1 mixture of the two epimers 36 and 37, and the undesired lactone 37 could be epimerised and converted into 36 by treatment with DBU [98]. Epimerisation and simultaneous lactone opening could be achieved in another key step using L-valine tert-butyl ester. Acetylation of the thus formed primary alcohol resulted in compound 38. This was followed by benzyl and Cbz deprotection and the subsequent urea formation with the imidazolium salt 39 to furnish tripeptide 40. After Boc deprotection, the resultant amine was guanidinylated using isothiourea 41. The thus obtained precursor 42 was treated with DIAD and PPh3 in a final step for an intramolecular Mitsunobu ring closure to finish the synthesis of the fully protected ureidomuraymycidine 43 (Scheme 5) [98].</p><!><p>Synthesis of the urea tripeptide unit as a building block for muraymycins reported by Kurosu et al. [98].</p><!><p>In 2010, Ducho et al. reported an alternative synthesis of the naturally occurring uridine-derived muraymycin core structure (Scheme 6) [78,99]. The key step of their route was a sulfur-ylide reaction with high substrate-controlled diastereoselectivity [100–102]. This epoxide-forming sulfur-ylide reaction had been established before by Sarabia et al. [103–104]. After some initial confusion regarding the stereochemical configuration of the epoxide product, it could be unambiguously proven that the transformation of uridine-5'-aldehyde 44 with sulfonium salt 45 under basic conditions furnished epoxide 46 with high diastereoselectivity (Scheme 6). Subsequent ring opening of this epoxide with Bu4NBr resulted in bromohydrine 47, followed by levulinyl (Lev) protection of the hydroxy group (product 48). Nucleophilic substitution at the 6'-position with Bu4NN3 gave the naturally occurring (5'S,6'S)-stereochemistry of the uridine core structure in a double inversion manner [78,99]. DDQ oxidation then provided indolamide 49. Hydrolysis of the amide, formation of the synthetically more versatile tert-butyl ester, azide reduction and final Cbz protection resulted in the uridine-derived building block 50 for the synthesis of naturally occurring muraymycins (Scheme 6). Furthermore, 5'- and 6'-epi analogues of muraymycins were also synthesised via suitable epoxide precursors by Ducho et al. [105].</p><!><p>Synthesis of the uridine-derived core structure of naturally occuring muraymycins reported by Ducho et al. [78,99].</p><!><p>Ducho's synthesis of epicapreomycidine (Scheme 7) started from the (R)-configured Boc-protected Garner aldehyde 51 [106], which was transformed into the N-benzylimine 52. The latter was then diastereoselectively converted with a Grignard reagent into the amine 53 as a key step of the synthesis [78]. Cbz protection followed by ozonolysis with subsequent reductive amination and hydrogenolysis led to the 1,3-diamine 54. The cyclisation to the guanidine functionality was achieved with the novel guanidinylation reagent 55. With the protected epicapreomycidine precursor 56 in hand, the Boc and acetonide protecting groups were removed. Urea formation with the valine derivative 57 with final oxidation of the primary hydroxy function afforded the desired dipeptide 58 [78] (Scheme 7).</p><!><p>Synthesis of the epicapreomycidine-containing urea dipeptide from Garner's aldehyde reported by Ducho et al. [78].</p><!><p>Furthermore, Ducho et al. synthesised the hydroxyleucine moiety found in naturally occurring muraymycins of classes A to C (Scheme 8) [107]. Adapting a strategy developed by Zhu et al., D-serine (59) was stereoselectively converted into the protected amino alcohol 60 [108]. Key intermediate 60 was then Cbz- and acetonide protected to give 61. A sequence of desilylation and oxidation furnished the acid 62. Peptide coupling with amine 63 and acidic deprotection then afforded the desired aldehyde 64, which already contained the muraymycin linker unit (Scheme 8) [107]. Together with the uridine core structure 50 and the urea dipeptide 58, the aldehyde 64 was the third building block of Ducho's envisioned stereocontrolled tripartite route towards muraymycins, in contrast to Ichikawa's and Matsuda's modular multicomponent, but non-stereocontrolled approach (see above).</p><!><p>Synthesis of a hydroxyleucine-derived aldehyde building block reported by Ducho et al. [107].</p><!><p>This novel tripartite approach was then used by Ducho et al. to synthesise the structurally simplified natural product analogue 5'-deoxy muraymycin C4 (65), which formally differs from the parent natural product only by absence of one oxygen atom (Scheme 9) [78,109–110]. Starting from protected uridine-5'-aldehyde 44, the first key step of the synthesis was a (Z)-selective Wittig–Horner reaction with phosphonate 66 [111] in order to obtain the didehydro amino acid 67. The next important step of this route was an asymmetric catalytic hydrogenation [112–113] with the chiral Rh(I)–DuPHOS catalyst 68 to prepare the (6'S)-configured product 69 [109–110]. Subsequent hydrogenolytic cleavage of the Cbz group gave the nucleosyl amino acid 70. To complete the tripartite approach, the reductive amination with the aldehyde 64 furnished 71, and Cbz deprotection and peptide coupling with the epicapreomycidine-containing urea dipeptide 58, followed by acidic global deprotection, gave the desired 5'-deoxy muraymycin C4 (65) (Scheme 9) [78].</p><!><p>Synthesis of 5'-deoxy muraymycin C4 (65) as a closely related natural product analogue [78,109–110].</p><!><p>In addition to the described synthetic routes, a range of other muraymycin analogues has been prepared. In the interest of conciseness, this synthetic work is not discussed here, but the biological properties of such analogues will be summarised in the following section on SAR studies.</p><!><p>With various structurally diverse compounds at hand, the stage has been set for SAR studies on muraymycins. The antimicrobial activities found by McDonald et al. introduced muraymycins as a promising subject of study [22]. The naturally occurring muraymycins isolated from Streptomyces guided first insights into the structural features essential for MraY inhibition. For the most active member of the family, i.e., muraymycin A1, antibiotic activity could be found against various bacteria ranging from Staphylococci with MIC values of 2 to 16 μg/mL, Enterococci with 16 μg/mL and higher to some Gram-negative bacteria (8 μg/mL). Against an E. coli mutant with increased membrane permeability, an MIC value below 0.03 μg/mL was obtained, suggesting that inhibition is a matter of cellular uptake of the compound. In vivo efficacy was demonstrated for muraymycin A1 with an ED50 of 1.1 mg/kg in Staphylococcus aureus-infected mice.</p><p>Five of the 19 naturally occurring compounds (i.e., muraymycins A1, A5, B6, C2 and C3) were capable of inhibiting both MraY and peptidoglycan synthesis at the lowest concentration tested (IC50 = 0.027 μg/mL), which represented activities comparable to those of liposidomycin C (0.05 μg/mL) and mureidomycin A (0.03 μg/mL). As a general trend, higher antimicrobial activities were found for acylated compounds, in particular with longer and functionalised fatty acid side chains.</p><p>Lin et al. employed a semisynthetic approach for modifications of muraymycin C1 as starting point of their SAR studies (Figure 7) [86]. In accordance with the results reported by McDonald et al., their work was based on the assumption that the cellular uptake required for MraY inhibition is mainly dependent on fatty acids connected to the hydroxyleucine moiety. The attachment of lipophilic groups on either the primary or both the primary and secondary amino function was supposed to have similar effects. The muraymycin derivatives 72–86 were thus evaluated against the target in a coupled MraY–MurG in vitro assay employing radiolabelled UDP-N-acetylglucosamine. Disubstituted analogues were not active at the concentrations tested, suggesting that one free amino group is vital for activity. Hydantoin-derived compounds 79 with C12H25 and 80 with PhCH2 as residues R at the hydantoin moiety gave the best results with inhibition of lipid II formation at 6.25 μg/mL, which is comparable to muraymycin C1. Good activity was also found for hydantoin derivative 77 with the 4-FC6H4 substituent, showing inhibition of lipid II formation at 25 μg/mL. The only N-alkylated derivative inhibiting in the same order of magnitude was 83 with n-C11H23 substitution. However, activities of the other compounds within this group also coincided with the previous observation that lipophilic compounds were more active. Overall, the tested monosubstituted hydantoin derivatives confirmed the assumed correlation between inhibitory activities and lipophilicity of the substituent.</p><!><p>Summary of modifications on semisynthetic muraymycin analogues tested by Lin et al. [86]. Most active compounds are highlighted (in vitro inhibition of lipid II formation at 6.25 μg/mL: orange; 25–50 μg/mL: green).</p><!><p>Yamashita et al. studied truncated muraymycin analogues lacking the lipophilic side chain as described in the section on synthetic access (compounds of type 7, 8 and 10) [76]. The activities measured in a soluble peptidoglycan assay indicated a stereochemical preference for the (5'S)-configuration, contrary to the results of MIC value determination. Further studies were then carried out with (5'R)-derivatives only, i.e., with 5'-epimers of the parent natural products. The influence of protecting groups was examined applying a strategy of stepwise deprotection. This led to the observation that fully protected compounds were not active at all, as well as the completely deprotected analogues. Remarkably, some partially protected congeners 87–90 with the free terminal amino group were found to show good inhibition (MIC = 1–16 μg/mL) of the growth of Gram-positive bacteria including S. aureus and E. faecalis strains, with best results obtained for 88 (Figure 8). Evaluation of the inhibition of lipid II formation revealed the importance of the substitution pattern of the terminal amino acid.</p><!><p>Bioactive muraymycin analogues identified by Yamashita et al. [76].</p><!><p>In 2010 and 2011, Ichikawa, Matsuda et al. published SAR studies with a range of synthetic muraymycin analogues [77,114]. The IC50 values were measured in an in vitro assay mentioned above to examine the inhibitory activity of the prepared analogues against the target enzyme. MIC values were determined against several bacterial strains. The inhibitory activities of the synthesised muraymycin D2 33 (with an L-leucine unit) and its epimer (with a D-leucine unit) on the purified MraY enzyme from B. subtilis were determined. Both compounds showed good inhibitory activities with IC50 values of 0.01 μM and 0.09 μM, respectively. However, their antibacterial activities against several Gram-positive bacteria (S. aureus, E. faecalis, E. faecium) were low (MIC values up to 64 μg/mL). In comparison to the analogues of the A and B series, which showed good antibacterial activities (see above), muraymycin D2 (33) and its epimer lack the hydrophobic side chain at the leucine moiety [22]. It was postulated that this lipophilic side chain may not be necessary for target inhibition, but for cellular uptake through the lipid bilayer of the cytoplasmic membrane, as an increased lipophilicity is advantageous for this [77,114].</p><p>Consequently, several lipophilic derivatives 91a–d were prepared (Figure 9). Long-chain lipophilic amino acids were incorporated into the muraymycin core structure as a simplified replacement of the O-acylated hydroxyleucine moiety. Compound 91a (highlighted in orange) with the pentadecyl side chain showed the best activity as an MraY inhibitor (IC50 = 0.33 μM (with L-leucine moiety), IC50 = 0.74 μM (with D-leucine moiety)), but relative to muraymycin D2 and its epimer, this implied a 33-fold and 8-fold, respectively, decrease of inhibitory activity. In bacterial growth assays, the analogue 91a exhibited the best MIC values ranging between 0.25 μg/mL and 4 μg/mL (see Table 2). These values were comparable to those of the naturally occurring congeners of the A and B series [22]. Generally, derivatives with the naturally occurring L-configuration in the leucine moiety showed slightly better activities. These lipophilic analogues were also tested for cytotoxicity towards Hep G2 cells and showed no cytotoxicity (IC50 > 100 μg/mL) [114].</p><!><p>Muraymycin D2 and several non-natural lipidated analogues 91a–d [77,114].</p><p>Inhibitory (against MraY) and antibacterial activities of non-natural lipophilic muraymycin analogues [77,114].</p><p>aInhibitory activities were determined against purified MraY enzyme from B. subtilis [77]; bMIC values were determined for different strains of S. aureus, E. faecalis and E. facium including some multiresistant strains [77]; n.d. = not determined.</p><!><p>In another series of analogues with different peptide units, the pentadecyl side chain of 91a was kept. The L-epicapreomycidine (L-epi-Cpm) unit of 91a was replaced by L-capreomycidine (L-Cpm, 92a), L-arginine (L-Arg, 92b) and L-ornithine (L-Orn, 92c) in order to investigate the role of the cyclic guanidine functionality (Figure 10) [77].</p><!><p>Non-natural muraymycin analogues with varying peptide structures [77,114].</p><!><p>These compounds were all active against MRSA and VRE with varying MIC values (Table 2). The most active analogues of this series were 92a and 92b (Figure 10, highlighted in orange) with MIC values between 1 μg/mL and 4 μg/mL. Derivatives with unnatural D-stereochemistry in the pentadecyl glycine motif possessed a similar antibacterial activity (potency within factor 2). Truncated analogues lacking the L-valine urea terminus (Cbz-protected 92d and N-terminally unprotected 92e) showed only a minor loss of activity (MIC = 4–8 μg/mL) (Table 2). These results indicated that the guanidine motif of analogues 91a, 92a and 92b (MICs between 0.25 μg/mL and 4 μg/mL) is preferred, but that amino analogues 92c and 92f still show good activity (MICs between 2 μg/mL to 8 μg/mL). The different stereochemistry at the central leucine unit and the terminal truncation had no crucial effects on the antibacterial activity (Table 2). Truncated derivatives 92f–h (Figure 10) without the L-valine urea terminus contained L-ornithine (L-Orn, 92f), L-arginine (L-Arg, 92g) and L-methionine (L-Met, 92h), respectively. They were also tested and showed reasonable activity against some bacterial strains (MIC = 4–8 μg/mL), which further indicated that significant variations in the peptide moiety are tolerated. The truncated analogue 93 (Figure 10) only consisted of the N-alkylated nucleoside core structure. Its inhibitory activity was 6 to 12-fold reduced (IC50 = 5 μM) and the antibacterial activity decreased with MIC values between 32 μg/mL and 64 μg/mL. In summary, these systematic SAR studies demonstrated the importance of the lipophilic side chain for the antibacterial activity. The urea dipeptide motif is important for antibacterial activity as well, but it could be diversified with simpler amino acids as well as being truncated in order to provide bioactive analogues. A graphical summary of these results is provided in Figure 11.</p><!><p>SAR results for several structural variations of the muraymycin scaffold.</p><!><p>In 2014, Ichikawa, Matsuda et al. continued their SAR studies with respect to urgently needed anti-Pseudomonas agents [115]. These Gram-negative bacteria possess an outer membrane which acts as an additional permeability barrier, making them generally less sensitive to antibacterial agents. In this context, the aforementioned muraymycin analogues (91a, 92a–h) were tested for MraY inhibitory activity again, with MraY enzyme from S. aureus (Table 3). However, antibacterial activities against several Pseudomonas strains were moderate to low with MICs between 8 μg/mL and >64 μg/mL. Analogue 92g was the most active congener in this series with MIC values between 8 μg/mL and 32 μg/mL. Compounds 92e and 92f showed nearly no activity (MIC = 32 to >64 μg/mL). More lipophilic truncated analogues 94 without the urea dipeptide unit (Figure 12) were synthesised and tested, but they all showed nearly no activity.</p><!><p>Inhibitory (against MraY) and antibacterial activities of non-natural muraymycin analogues against Pseudomonas aeruginosa [115].</p><p>aInhibitory activities were determined against MraY enzyme from S. aureus [115]; bMIC values were determined for several P. aeruginosa strains [115].</p><p>Muraymycin analogues designed for potential anti-Pseudomonas activity (most active analogues are highlighted in orange) [115].</p><!><p>These results indicated the importance of the presence of a guanidine residue and a lipophilic side chain for potential antibacterial activity against Pseudomonas strains. Hence, several derivatives were prepared in which the positions and numbers of the guanidine groups and the lipophilic side chains were varied in order to optimise their relative orientation for best biological activity. This strategy resulted in the bioactive analogues 95–98 (Figure 12). Analogue 95 with an interconversion of the lipid side chain and the guanidine group had a slightly reduced activity compared to lipidated analogue 92g. Analogue 96 showed an increased antibacterial activity towards some of the tested Pseudomonas strains. Analogue 97 is an interconverted version of 96 and displayed a comparatively poor activity. The most active analogue was compound 98 which is a hybrid type of the aforementioned analogues 95–97. The results indicate that a lipophilic side chain and guanidine groups are necessary for antibacterial potency. Compounds 95–98 showed antibacterial activity, with the branched-type compound 96 (MIC values between 8 μg/mL and 16 μg/mL) and the hybrid-type compound 98 (MIC between 4 μg/mL and 8 μg/mL) being the most active congeners. A limitation of both analogues 96 and 98 is their increased cytotoxicity against HepG2 cells with IC50 values of 4.5 μg/mL and 34 μg/mL, respectively. Further, the metabolic stability was studied in vitro for the analogues 95, 96 and 98 using human or rat liver microsomes and all of them proved to be reasonably stable [115].</p><p>In 2014, Ducho et al. reported the synthesis of 5'-deoxy muraymycin C4 (65, see above) [78]. Biological assays revealed that 65 inhibited the MraY enzymes of E. coli and S. aureus with potencies in the range of tunicamycins. The antibacterial activity of 65 was tested against some selected E. coli and S. aureus strains although the lack of a lipophilic moiety indicated that the compound should not be a potent antibiotic. However, an unexpected moderate activity against E. coli DH5 alpha was observed, whereas 65 was weakly active against E. coli strain ΔtolC but not active against the S. aureus Newman strain. Further studies indicated excellent plasma and metabolic stability and no cytotoxicity. Overall, the structurally simplified 5'-deoxy muraymycin scaffold 65 may therefore be useful for further antibacterial development. It should also be noticed that it has inspired the design of a novel oligonucleotide backbone modification [116–117].</p><!><p>So far, there are only limited insights into muraymycin biosynthesis. The biosynthetic gene cluster for the formation of muraymycins in Streptomyces sp. NRRL 30471 has been identified by Chen, Deng et al. in 2011 [118]. The sequence analysis revealed the cluster to contain 33 open reading frames (ORFs) with 26 of them being involved in muraymycin formation. Based on their elucidation of the gene cluster and sequence homologies, Chen, Deng et al. proposed an outline pathway for muraymycin biosynthesis (Scheme 10).</p><!><p>Proposed outline pathway for muraymycin biosynthesis based on the analysis of the biosynthetic gene cluster by Chen, Deng et al. [118]. MTA = 5'-deoxy-5'-(methylthio)adenosine.</p><!><p>According to this biosynthetic proposal, uridine (1) is enzymatically oxidised to give uridine-5'-aldehyde 99. Aldehyde 99 then supposedly undergoes an aldol addition with glycine 100 as the enol(ate) component, thus furnishing the amino acid–nucleoside hybrid 5'-C-glycyluridine (GlyU, 101). Alkylation of the 6'-amino group is then achieved by reaction with S-adenosyl methionine (SAM), and the resultant intermediate 102 is decarboxylated to provide diamine 103. Attachment of the aminoribosyl moiety (which is supposedly also derived from uridine (1) over several enzymatic steps) finally affords the aminopropyl-substituted 5'-O-aminoribosylated GlyU core structure 104. Transformation of 104 with the thioester-activated peptide moiety 105 then gives muraymycin C2 (Scheme 10), which is speculated to serve as an intermediate en route to other muraymycins, in particular towards O-lipidated congeners of the A and B series (see Figure 2).</p><p>A fragmented non-ribosomal peptide synthetase (NRPS) system appears to be responsible for the assembly of the urea tripeptide building block 105. However, the non-proteinogenic amino acids need to be formed first. It has been proposed that L-arginine (106) undergoes 3-hydroxylation (giving 3-hydroxy-L-arginine (107)) and subsequent ring closure to furnish L-epicapreomycidine ((2S,3S)-capreomycidine, 108), that is then activated as thioester 109 (Scheme 10). This proposal is based on the elucidated formation of the epimeric amino acid L-capreomycidine ((2S,3R)-capreomycidine) as part of viomycin biosynthesis in Streptomyces vinaceus. In this producing organism, L-arginine is diastereoselectively hydroxylated to afford (3S)-3-hydroxy-L-arginine. The ring-closure reaction then occurs with formal inversion of the β-stereocenter (but quite likely through an aza-Michael addition to the α,β-unsaturated intermediate) [119–121]. The exact stereochemical course of epicapreomycidine formation in muraymycin biosynthesis is unclear though as the stereochemical configuration at C-3 of the intermediate 3-hydroxy-L-arginine (107) has not been identified yet. It cannot be ruled out that an epimerisation reaction might be involved in the biosynthesis of 108, in particular with respect to other epimerisation steps in bacterial biosynthetic pathways [122]. Consequently, synthetic routes towards both 3-epimers of 3-hydroxy-L-arginine have been developed which would also enable the preparation of isotopically labelled congeners for biosynthetic studies [123–124]. It should also be noted that a biomimetic domino guanidinylation–aza-Michael-addition reaction for the synthesis of the capreomycidine scaffold has been developed, which only furnished the target structures as stereoisomeric mixtures though [125].</p><p>The epicapreomycidine-derived thioester 109 is proposed to be converted into the urea dipeptide motif with valine derivative 110 and possibly hydrogen carbonate as a C1-building block for urea formation, thus furnishing 111. The 3-hydroxy-L-leucine moiety might be obtained by stereoselective enzymatic β-hydroxylation of thioester-activated L-leucine 112, which leads to the formation of 113. Finally, peptide formation by condensation of 111 with 113 affords the complete thioester-activated urea tripeptide unit 105 (Scheme 10). One interesting aspect of this biosynthetic proposal by Chen, Deng et al. is that they assume the putative dioxygenase Mur16 to catalyse β-hydroxylations of two structurally distinct amino acid substrates, i.e., L-arginine (106) and thioester-activated L-leucine 112.</p><p>As pointed out, there is a lack of experimental insights into muraymycin biosynthesis beyond the elucidation of its gene cluster. However, Van Lanen et al. have studied the early steps of the biosynthesis of A-90289 nucleoside antibiotics in detail (Scheme 11) [126]. The A-90289 subclass is structurally closely related to caprazamycins and liposidomycins, and its aminoribosylated nucleoside core is identical to that of muraymycins (Figure 2). This supports the assumption that the early steps of the biosynthesis of all these subclasses are probably highly similar, if not identical. For the A-90289 nucleoside antibiotics, Van Lanen et al. have demonstrated that uridine monophosphate (UMP, 114) is the actual source of uridine-5'-aldehyde 99, which is furnished in an oxidative transformation of UMP 114 with the 2-oxoglutarate (2-OG)-dependent non-haem Fe(II)-oxygenase LipL [127]. This result challenges the proposal by Chen, Deng et al. that aldehyde 99 might be formed by oxidation of uridine (1) in muraymycin biosynthesis. Aldehyde 99 then undergoes the aforementioned aldol-type transformation to GlyU 101, catalysed by the enzyme LipK. However, aldehyde 99 also serves as a source of the aminoribosyl moiety. Thus, it is converted into 5'-amino-5'-deoxyuridine (115) in a transamination reaction mediated by LipO. This is followed by the LipP-catalysed displacement of the uracil with a phosphate moiety to afford 5-amino-5-deoxyribose-1-phosphate (116). The LipM-mediated reaction of ribosyl phosphate 116 with a nucleoside triphosphate (NTP) then yields nucleoside diphosphate (NDP)-aminoribose 117. Finally, aminoribosylation of 101 with glycosyl donor 117, catalysed by glycosyltransferase LipN, furnishes the complete nucleoside core structure 118 (Scheme 11). The order of 6'-N-(3-aminopropyl) attachment and 5'-O-aminoribosylation is not fully clear yet, i.e., it is not elucidated if 101 or 6'-N-aminoalkyl intermediate 103 (see Scheme 10) act as the glycosyl acceptor in the aminoribosylation step.</p><!><p>Biosynthesis of the nucleoside core structure of A-90289 antibiotics (which is identical to the muraymycin nucleoside core) according to the studies of Van Lanen et al. [126]. 2-OG = 2-oxoglutarate.</p><!><p>Van Lanen et al. then studied the LipK-catalysed aldol-type formation of GlyU 101 in more detail [128]. Surprisingly and in contrast to Chen's and Deng's proposal, L-threonine (119) turned out to be the source of the enol(ate) component instead of glycine (100). Hence, LipK was revealed to be a transaldolase mediating a retro-aldol reaction of L-threonine (119) towards the enol(ate) and acetaldehyde (120), followed by a stereoselective aldol addition of the former to uridine-5'-aldehyde 99 (Scheme 12). Using synthetic reference compounds, it could be proven that (5'S,6'S)-GlyU 101 is the stereoisomer furnished in this reaction, so that no epimerisation at a later stage of the biosynthetic route is required for the formation of the A-90289 nucleoside antibiotics.</p><!><p>Transaldolase-catalysed formation of the key intermediate GlyU 101 in the biosynthesis of muraymycin-related A-90289 antibiotics [128].</p><!><p>Based on the elucidation of the LipK-mediated reaction, Van Lanen et al. then performed a PCR-based screening of a collection of ≈2500 actinomycete strains for similar transaldolase-encoding genes [129]. They could identify the gene sphJ from a Sphaerisporangium sp., which encoded the transaldolase SphJ having 51% amino acid sequence identity with LipK. Following detailed characterisation of this enzyme, the sphJ gene was employed as a probe to clone the entire genetic locus consisting of 34 putative ORFs. The expression of three selected genes (including sphJ) was monitored under different growth conditions. Under the thereby identified optimal conditions, the actinomycete produced a set of four unprecedented MraY-inhibiting nucleoside antibiotics named sphaerimicin A to D [129]. Hence, detailed studies on LipK-like transaldolases led to the discovery of novel antimicrobially active secondary metabolites.</p><p>It remains to be proven that the results obtained for the early steps of A-90289 and sphaerimicin biosynthesis are also valid for the biosynthetic formation of muraymycins. Bioinformatic analyses of the biosynthetic gene clusters of A-90289 antibiotics, caprazamycins and muraymycins revealed six shared ORFs overall [128]. A sequence comparison of a range of transaldolases gave 47% identity and 78% similarity of Mur17 with LipK [129]. Overall, these insights suggest that the formation of the GlyU intermediate 101 and very likely also of the whole aminoribosylated nucleoside core structure occur in a conserved manner. Further studies on muraymycin biosynthesis are still pending.</p><!><p>In summary, this review describes a promising class of antimicrobially active natural products, the uridine-derived muraymycins. Muraymycins are one subclass of nucleoside antibiotics inhibiting the membrane protein translocase I (MraY), a key enzyme in the intracellular part of peptidoglycan formation. Synthetic methodology for the preparation of muraymycins and their analogues has been established, and first SAR insights revealed that the design of structurally simplified, biologically active muraymycin analogues is an auspicious approach. However, further SAR studies as well as investigations on the interplay of target inhibition and cellular uptake for the antibiotic activity are surely desirable. Studies on muraymycin biosynthesis may not only be of academic interest, but could also lead to semi- or mutasynthetic methodology for the preparation of novel muraymycin analogues. Several laboratories around the world currently perform research on muraymycins and other uridine-derived nucleoside antibiotics. Hopefully, this work will contribute to the development of urgently needed novel antimicrobial drugs.</p>

PubMed Open Access

Universal kinetics of imperfect reactions in confinement

Chemical reactions generically require that particles come into contact. In practice, reaction is often imperfect and can necessitate multiple random encounters between reactants. In confined geometries, despite notable recent advances, there is to date no general analytical treatment of such imperfect transport-limited reaction kinetics. Here, we determine the kinetics of imperfect reactions in confining domains for any diffusive or anomalously diffusive Markovian transport process, and for different models of imperfect reactivity. We show that the full distribution of reaction times is obtained in the large confining volume limit from the knowledge of the mean reaction time only, which we determine explicitly. This distribution for imperfect reactions is found to be identical to that of perfect reactions upon an appropriate rescaling of parameters, which highlights the robustness of our results. Strikingly, this holds true even in the regime of low reactivity where the mean reaction time is independent of the transport process, and can lead to large fluctuations of the reaction time -even in simple reaction schemes. We illustrate our results for normal diffusion in domains of generic shape, and for anomalous diffusion in complex environments, where our predictions are confirmed by numerical simulations.

universal_kinetics_of_imperfect_reactions_in_confinement

<!>Results and discussion<!>Conclusions

<p>T he First Passage Time (FPT) quantifies the time needed for a random walker to reach a target site [1][2][3][4][5][6][7][8][9][10] . This observable is involved in various areas of biological and soft matter physics and is particularly relevant in the context of reaction kinetics, because two reactants have to meet before any reaction can occur [11][12][13] . When the reaction is perfect, i.e. when it occurs for certain upon the first encounter, its kinetics is controlled by the first passage statistics of one reactant, described as a random walker, to a target site. Of note, earlier works have determined the mean 2,8,14 and the full asymptotic distribution 15,16 of first passage times in confinement for broad classes of transport processes.</p><p>While most of the literature focuses on perfect reactions, the case of imperfect reactions (i.e. which do not occur with certainty upon the first encounter) arises in a variety of contexts (see 17 for a recent review), if a reaction occurs only when reactants meet with prescribed orientations 12 or after crossing an energy 18 (or entropy 19 ) activation barrier, if the target site is only partially covered by reactive patches 20 , or in the case of gated reactions where the target (or the reactant) switches between reactive and inactive states 21,22 .</p><p>The formalism to calculate the rate of imperfect reactions between diffusive spherical particles in the dilute regime (thus in infinite space) is well established 12,[23][24][25] . However, geometric confinement has proved to play an important role in various contexts, such as reactions in microfabricated reactors or in cellular compartments. Yet, the kinetics of imperfect reactions in a confined volume is still only partially understood: existing methods are restricted to (i) diffusive (or amenable to diffusive) transport processes [26][27][28][29] and most of the time (ii) specific shapes of confining volume [30][31][32][33] (spherical or cylindrical). In fact, a general theoretical framework to quantify the kinetics of imperfect reactions involving non-Brownian transport (such as anomalous diffusion in complex environments 34 ) in general confined domains is still missing.</p><p>Here, we propose a formalism that determines the full kinetics of imperfect reactions in confinement for general Markovian processes in the large confining volume limit (see Fig. 1). This allows us to answer the following questions: (i) Is reaction limited by transport or reactivity? (ii) What is the magnitude of the fluctuations of the reaction time? In particular, is the first moment sufficient to fully determine reaction kinetics? (iii) Do reaction kinetics depend on the choice of model of imperfect reactivity-namely partially reflecting (Robin) conditions 23,35,36 or sink with locally uniform absorption rate 24,37 in continuous models, or finite reaction probability in discrete models?</p><!><p>Discrete model of imperfect reactions. A first straightforward definition of imperfect reactivity is based on the statistics of encounter events between reactants, and thus requires a discrete description of the dynamics. We therefore start by considering a Markovian random walker moving on a discrete space (or network) of N sites. We consider a continuous-time dynamics with exponentially distributed waiting times on each site, where ν i denotes the jump rate from site i to any neighboring site. The reactive site is denoted i = 0. Imperfect reactivity is then naturally defined as follows: each time the walker visits the reactive site, the reaction occurs with probability p, and the random walk continues without reaction with probability 1 − p. We call T r (p) the reaction time and F(T|r, p) its probability density function (PDF) for a random walker starting from r. Next, we call τ r the first passage time to the reactive site starting from r (including the residence time on the reactive site), and we call F * (τ r |r) its PDF. We also introduce the first return time to the target τ 1 (i.e., the first passage time to the target starting from any site at distance 1 from the target) and F à 1 ðτ 1 Þ its PDF. The probability that a reaction happens after exactly n visits to the target is given by p(1 − p) n−1 , in which case T r (p) is the sum of the first passage time (starting from r) and of n − 1 independently distributed first return times (see Fig. 1). Hence, partitioning over the number of visits n yields</p><p>where τ ðkÞ 1 represents the return time after k − 1 visits to the reactive site. This exact equation is conveniently rewritten after Laplace transform (denoted f ðsÞ ¼ R 1 0 dtf ðtÞe Àst for any function f):</p><p>(see Supplementary Note 1 for details). In the small s limit, the property Fðsjr; pÞ ' 1 À shT r ðpÞi can be used to obtain an exact expression of the mean reaction time as a function of the mean first passage and the mean return time:</p><p>Of note, expression (3) [as well as (2)] is a straightforward consequence of well-known results on random sums 38 , bearing here a clear interpretation because (1 − p)/p is the mean number of encounter events. Below, we make this result fully explicit by determining 〈τ r 〉 and 〈τ 1 〉.</p><p>The mean return time 〈τ 1 〉 can be obtained exactly from the knowledge of the stationary probability density q i for the random walker to be at site i in absence of target; this exact result is known as Kac theorem 39 and yields</p><p>where we have chosen a uniform stationary distribution q i = 1/N, which is realized when the waiting time 1/ν i at each site is inversely proportional to the number of neighbors 40 .</p><p>To gain explicit insight of the behavior of the first reaction times, we next determine 〈τ r 〉 and make use of the scale invariance Fig. 1 Imperfect reaction kinetics in confinement. (a) In the case of imperfect reactions, multiple random interaction events between reactants are typically required before reaction occurs. The reaction time T r for a random walker starting from r with a target (red dot) can then be written</p><p>, where τ r is the first passage time (FPT) to the target, n is the total number of visits to the target before reaction, and τ ðkÞ 1 is a first return time to the target. (b) In the case of discrete space models, imperfect reactivity is parametrized by the probability p that reaction occurs at each visit of the random walker to the target. In the case of continuous space models, imperfect reactivity is modeled either by (c) a reaction rate k(r) when the random walker is within the reactive volume that defines the target, or (d) partially absorbing boundary conditions (parametrized by κ) at the target boundary.</p><p>property observed for a broad class of random walks, for which one can define a fractal space dimension d f (defined such that the characteristic size R grows as R / N 1=d f ) and a walk dimension d w such that the mean square displacement of a random walker scales as hr 2 ðtÞi / t 2=d w (without absorption, in unconfined space). Here, we make use of the chemical distance r, defined as the minimal number of links between two sites. The first passage kinetics is known to strongly differ for compact walks (d w > d f , for which the random walker explores densely its surrounding space and the probability to visit a site in infinite space is one) or noncompact walks (d w < d f , for which an infinite trajectory typically leaves a fraction of unvisited sites which is almost surely one). We shall prove here that the effect of imperfect reactivity is markedly different in these two cases as well.</p><p>We first focus on the compact case d w > d f , for which it was shown 41 that hτ r i ' αNr d w Àd f for large r and large volume N, where α is a constant independent of N and r. Following 41 , we assume that this scaling relation holds up even for r = 1. Making use of the above determination of 〈τ 1 〉, this yields α = 1/ν 0 . This leads to the following fully explicit determination of the mean reaction time:</p><p>As expected, the reaction time is thus the sum of a diffusioncontrolled (DC) time 〈τ r 〉, obtained when p = 1, corresponding to the time needed for the reactants to meet, and a reaction controlled (RC) time 〈τ 1 〉(1 − p)/p, which dominates when p → 0, corresponding to the sequence of returns to the target needed for the reaction to occur. These two times are equal when r ≃ l c , where the characteristic distance l c is given by</p><p>For this compact case, we can therefore split the confining domain into a region where the reaction time is reaction controlled (RC, for r < l c ), and another one where it is diffusion controlled (DC, r > l c ), see Fig. 2(b). Remarkably, we note that DC region disappears only when the size R of the confining volume becomes of the order of l c , i.e., when p ( 1=R d w Àd f ; this means that even for very small values of the intrinsic reactivity there will exist DC regions for large enough volumes.</p><p>To quantify reaction kinetics at all timescales, the full distribution of the reaction time, or equivalently the survival probability S(t|r, p), defined as the fraction of walkers that have not reacted up to time t, is needed. We show in Supplementary Note 2 how to determine S(t) by evaluating the leading order behavior of all the moments hT n r ðpÞi in the large volume limit (defined by N → ∞ with all other parameters fixed). Using the additional hypothesis that the scaling behavior of of all moments hτ n r i $ r d w Àd f R d f þðnÀ1Þd w holds up to r = 1, this leads to an explicit determination of the survival probability</p><p>where 〈τ〉 G is the global mean first passage time, i.e., the average of 〈τ r 〉 overall starting positions of the random walker and is independent of p. Here, Φ ν is a universal function depending only on ν = d f /d w , which was obtained 15 for the first passage problem by relying in the O'Shaughnessy-Procaccia operator 42 (which is known to provide accurate expressions for propagators for nottoo-large distances 43 ):</p><p>Here Γ is the gamma function, J is the Bessel function of the first kind, and α 0 < α 1 < . . . are the zeros of the function J −ν . Several comments are in order. (i) This main result shows that the functional form of the survival probability is exactly the same as that of the first passage time to the target (obtained for p = 1), with a rescaled prefactor 〈T r (p)〉 that encompasses all the dependence on the reactivity parameter p. It generalizes the result obtained for perfect reactions 15 . (ii) Importantly, the full distribution can be obtained from the knowledge of the first moment 〈T r (p)〉 only, which makes the mean the key quantity to determine reaction kinetics. (iii) Remarkably, the shape of the reaction time distribution is the same as that of first passage times even in regions of the domain where the mean reaction time is reaction controlled and seemingly independent of the dynamics. As stressed above, the dependence on p lies only in the prefactor of the survival probability. This implies that the property of broadly distributed reaction times (nonexponential), characteristic of first passage times for compact transport, is maintained even for low intrinsic reactivity in large networks. (iv) The importance of fluctuations can be quantified by the ratio hT 2 i=hTi 2 $ R d w =hTi ) 1, which is large in the large volume limit that we consider.</p><p>In order to test these predictions for compact processes, we have performed numerical calculations on different examples of both disordered and deterministic fractal networks: the 2-dimensional critical percolation cluster, the Vicsek fractals and the dual Sierpinski gasket, see Fig. 2(a). This enables us to test different values of d f , d w . This class of models has been used to describe transport in disordered media for example in the case of anomalous diffusion in crowded environments like biological cells [44][45][46] as a first step to account for geometrical obstruction and binding effects involved in real crowded environments. Our calculations of the reaction times in the case of deterministic fractals are based on a recursive construction of the eigenvalues and eigenfunctions of the connectivity matrix (see Supplementary Note 3 for details) and enable us to obtain exact forms for the Laplace transform of S(t) for volumes up to N ~10 6 sites. As seen on Fig. 2, these numerical results confirm our predictions for the evaluation of the mean first passage time and the rescaled form of the survival probability. These results indicate that our approximations (i.e. the use of the O'Shaughnessy-Procaccia operator, the hypothesis that scaling of all moments hold up to r = 1, large volume limit) lead to accurate predictions for the mean reaction time and its full distribution. Of note, even for small values of the reaction probability (p = 0.05) the shape of reaction time distribution is exactly the same as that of first passage times, as we predict. In the limit of small p (at fixed volume), one expects that the reaction becomes much slower than the transport step, with an exponentially distributed reaction time. However, this exponential regime appears when the length l c in Eq. ( 6) becomes comparable to the size R of the fractal, i.e. when p ( p à 1=N d w =d f À1 . Since p * vanishes for large N, this means that the distribution of first passage times remains broadly distributed, with no well-defined reaction rate even for very low values of p.</p><p>Noncompact case. We now focus on noncompact processes (d w < d f ). In this case, we make use of the asymptotic FPT distribution, which can be written 15 as</p><p>where 〈τ〉 G has been defined above. The term δ(t) accounts for the FPT density restricted to trajectories that do not reach the boundary before finding the target, the shape of the function approximated by this δ-function does not modify the value of the moments of the distribution in the large volume limit. Now, we make use of this separation of timescales in the FPT distribution and obtain finally the distribution of the reaction time by (i) taking the Laplace transform of ( 9), (ii) inserting the result into (2) and (iii) taking the inverse Laplace transform. The result of this procedure for the survival probability is</p><p>where the mean reaction time 〈T r (p)〉 is deduced from Eqs.</p><p>(3),( 4), and 〈T(p)〉 G = 〈T r=∞ 〉 is the global (indexed by G) mean reaction time, i-e averaged over all starting positions. This result has important consequences. (a) Similarly to the compact case, the shape of the survival probability for imperfect reactions is the same as that of first passage times, with renormalized parameters; in particular the mean gives access to the full distribution and is thus the key quantity to quantify reaction kinetics, as in the compact case. (b) Because the mean FPT scales as 〈τ〉 ~(N/ν 0 ) g(r), where g is a bounded function of r, the mean reaction time is dominated by the reaction limited step in the full domain as soon as p ≪ 1 for any domain size, in contrast to the compact case. (c) Note that, in Eq. ( 10) one has S(t → 0) < 1, which means that it does not take into account the events whose duration does not -2, and Supplementary Table 1. In (d) the error bars represent 95% confidence intervals.</p><p>scale with R; the survival probability for these events was identified to the survival probability in infinite space 15,32 . Nevertheless, Eq. ( 10) can be used to calculate all the moments hT n r ðpÞi with n ≥ 1 in the large volume limit.</p><p>Continuous models: imperfect extended targets. We now aim at discussing alternative microscopic models of imperfect reactivity, which are naturally defined in continuous space. We consider a d-dimensional Brownian diffusive particle of diffusion coefficient D and analyze two classical models of imperfect targets (see Fig. 1) : (i) a sink region V r (in which the reaction happens with rate k(r) and vanishes elsewhere) and (ii) a target region S r with partially reactive impenetrable boundary. Our above results for discrete models show that the full distribution of reaction times (f) and (g) Rescaled survival probabilities for 2D/3D simulations, all parameters are in legend except for R/a = 6 and ka 2 /D = 1 (for sink reactivity) and κa/D = 1 (for surface reactivity). In 3D we evaluated 〈T〉 G = Vϕ(∞). In 2D, we used 〈T〉 G = Vϕ(1) + 〈τ〉 G where 〈τ〉 G was evaluated numerically for each domain. In all simulations we used a time step Δt = 10 −4 a 2 /D. For surface reactivity we implemented our simulation algorithm by using ref. 49 . Error bars (95% confidence intervals) are smaller than symbols. can be obtained in the large volume limit from the first moment of the reaction time only; we conjecture and verify numerically that this holds also for continuous models. We are thus back to determining the mean reaction time in both cases (i) and (ii). In case (i), the mean reaction time 〈T(r)〉 starting from the position r satisfies the following backward equation 47 :</p><p>We next define ΦðrÞ ¼ lim V!1 hTðrÞi=V, and obtain from (11) (and ( 11) integrated over the volume) :</p><p>which fully determines Φ for all r 2 R d . This formalism can be adapted to the case (ii) of partially reactive target of surface S r characterized by a surface reactivity κ that interpolates from perfect reaction (κ → ∞) to complete absence of reaction (κ → 0) 17,23,35,36 ; we obtain</p><p>These equations ( 12) and ( 13) generalize the formalism of Ref. 48 to the case of imperfect reactions. Note that (i) they can be extended to general Markovian transport operators, for both compact and noncompact cases, and (ii) they are valid for any shape of the confining volume.</p><p>To illustrate our formalism, we give solutions for diffusive transport for dimensions d = 1, 2, 3 for a spherical target of radius a. For the case (i) of a sink region k(r) = kθ(a − r) with θ the Heaviside step function, the MRT outside the sink region (r > a)</p><p>where K = ka 2 /D and I 0 , I 1 are modified Bessel functions of the first kind. In the case (ii) of a partially reactive impenetrable spherical target we obtain hTi</p><p>Of note, in d = 3, the MRT at r = ∞ in Eqs. ( 15), ( 14)) is the inverse of effective reactions rates calculated in 23,24 , and in fact the expression (10) then corresponds to the survival probability at long times for r/a ≫ 1 identified in Ref. 26 . Finally, our results show that both models are equivalent in the low reactivity limit upon the identification κS r = kV r ; however, in the high reactivity limit, the RC timescales as κ −1 for surface reactivity, while for sink absorption the RC time displays a non-trivial scaling ∝ k −1/2 , due to the fact that most reaction events occur in a small penetration length from the target surface. These results for both models (i) and (ii) have been confirmed by numerical simulations for confining volumes of various shapes (see Fig. 3), which have been chosen as representative of nonspherical volumes, displaying anisotropy (A) or protrusions (B). Importantly, numerical results confirm our prediction that for d ≥ 2 the full distribution is still given by Eq. ( 10) for both continuous models (where the case d = 2 is considered as noncompact) [Fig. 3(f-g)].</p><!><p>We have provided a general formalism to determine the reaction time distribution for imperfect reactions involving the broad class of diffusive and anomalously diffusive Markovian transport processes in confinement. We have investigated several representative mechanisms of imperfect reactivity to test the robustness of our conclusions. Importantly, our results show that the first moment alone, although not representative of typical reaction times, gives access to the full distribution in the large volume limit, which allows to quantify reaction kinetics at all timescales. Thanks to this property, our formalism can be adapted in principle to refined mechanisms of imperfect reactivity (gating, orientational constraints...), as soon as the mean first passage time can be asymptotically determined. Remarkably, and counterintuitively, we find that in the large volume limit the reaction time distribution is identical to that of the first passage time upon an appropriate rescaling of parameters. This implies that for compact transport processes, the reaction time distribution is broadly distributed with large fluctuations even in the reaction controlled regime where the mean reaction time is independent of the transport process. This is in striking contrast with the naive prediction of exponentially distributed reaction times for firstorder kinetics, which in fact is valid only for extremely low reactivity. This unexpected property could lead to large fluctuations of concentrations-as observed in the context of gene expression-even in simple reaction schemes, and even for low reactivity. We expect that the main effect identified here, i.e. that complex first passage properties due to compact transport do not disappear for imperfect reactivity, could be extended to more general processes that are more complex than scale-invariant Markovian processes, and to more complex reactions schemes potentially involving competitive reactions. This will be the subject of future works.</p>

Nature Communications Chemistry

Palladium-Catalyzed Asymmetric Allylic Alkylations with Toluene Derivatives as Pronucleophiles

The first two highly enantioselective palladium-catalyzed allylic alkylations with benzylic nucleophiles activated with Cr(CO)3 have been developed. These methods enable the enantioselective synthesis of \xe2\x80\x9c\xce\xb1-2-propenyl benzyl\xe2\x80\x9d motifs, which are important scaffolds in natural products and pharmaceuticals. A variety of cyclic and acyclic allylic carbonates are competent electrophilic partners furnishing the products in excellent enantioselectivity (up to 99% ee and 92% yield). This approach was employed to prepare a nonsteroidal anti-inflammatory drug analogue.

palladium-catalyzed_asymmetric_allylic_alkylations_with_toluene_derivatives_as_pronucleophiles

<p>Transition-metal-catalyzed allylic substitution reactions have emerged as a powerful method to construct C−C bonds.[1] Of these, the palladium-catalyzed asymmetric allylic alkylation (AAA) reaction has attracted the greatest interest due to its applications in total synthesis and the preparation of bioactive compounds.[2] A wide variety of stabilized or "soft" nucleophiles (originally defined as pronucleophiles with pKa < 25[3]) have been successfully employed in palladium-catalyzed AAA reactions, including malonates,[4] imides[5] and many others.[6] In contrast, few palladium-catalyzed allylic substitution reactions with "hard" nucleophiles (generally defined as pronucleophiles with pKa > 25) have materialized.[1a] Recent advances in palladium-catalyzed AAA with "hard" nucleophiles have been reported by Morken and co-workers with allylboronates in allyl-allyl coupling reactions[7] (Scheme 1A) and Maulide and co-workers with the use of dialkylzinc nucleophiles[8] (Scheme 1B).</p><p>The most significant difference between "soft" and "hard" nucleophile classes in the Tsuji-Trost allylic substitution is their reaction pathways: "soft" nucleophiles react via a double inversion mechanism whereas "hard" nucleophiles are proposed to undergo single inversion,[9] wherein the nucleophile transmetallates to the palladium catalyst. Both reactions in Scheme 1 were shown to proceed by the "hard" nucleophile pathway. As a valuable complement, Fletcher and co-workers reported copper-catalyzed AAA reaction of alkylzirconium reagents (generated in situ from alkenes by hydrometallation) via dynamic kinetic asymmetric transformation.[10] The scope of this reaction, however, does not include simple benzylic nucleophiles.</p><p>To broaden the synthetic utility of the palladium-catalyzed AAA, researchers have focused on expanding the classes of nucleophiles that can undergo the more controllable double inversion or "soft" nucleophile reaction pathway. This approach entails "softening" of hard nucleophiles by addition of activating agents to stabilize the resulting anionic charge. In pioneering studies, Trost and co-workers reported the highly enantioselective palladium-catalyzed AAA with 2-methylpyridine derivatives (Scheme 2A, LG=leaving group).[11] Key to success of this approach was addition of 1.3 equiv BF3 to bind the nitrogen and acidify the sp3-hybridized C-H's of 2-methylpyridine (pKa ~ 34[12]). A different strategy is necessary for pronucleophiles bearing less acidic C–H's in the absence of Lewis basic heteroatoms, such as toluene derivatives (pKa ~ 43[13]). AAA with benzyl anion nucleophiles could potentially furnish benzylated substructures, which are important scaffolds in natural products and pharmaceuticals, such as FR181157,[14] Sophoraflavanone G,[15] and Dactylosponol[16] (Figure 1).</p><p>Herein we report the first palladium-catalyzed AAA reactions of toluene-based pronucleophiles. By activation of toluene derivatives with η6-tricarbonylchromium, both cyclic and acyclic allylic carbonates are benzylated with enantioselectivities as high as 96 and 99%, respectively (Scheme 2B).</p><p>Recently, our team introduced a strategy to employ (η6-C6H5CH2R)Cr(CO)3 complexes as cross-coupling partners to produce di- and triarylmethanes[17] and enantioenriched diarylmethylamines[18] via direct arylations.[19] Using these pronucleophiles, we also explored palladium-catalyzed allylic substitution of diverse cyclic and acyclic electrophiles to give racemic products.[20] Despite significant progress in palladium-catalyzed Tsuji-Trost reactions, highly enantioselective processes with benzylic nucleophiles (toluene derivatives) remain a limitation of this method. Our prior demonstration of a double inversion reaction pathway with Cr(CO)3 activated benzylic nucleophiles to afford racemic products inspired us to pursue palladium-catalyzed AAA reactions.[20]</p><p>Given the known difficulty of identifying chiral ligands that can moderate all of the steps of a complex catalytic cycle and provide useful enantioselectivity, we initially examined over 140 diverse enantioenriched mono- and bidentate phosphine ligands (see Supporting Information for full ligand structures and results). We employed 1 equiv (η6-C6H5CH3)Cr(CO)3 (1a) as the pronucleophile, 2 equiv tert-butyl cyclohex-2-enyl carbonate (2a), 3 equiv LiN(SiMe3)2, 1 equiv PMDTA (pentamethyldiethylenetriamine) additive,[21] 10 mol % Pd(COD)Cl2, and 10 mol % chiral bidentate ligand or 20 mol % monodentate phosphine in THF at room temperature for 12 h (Table 1). Surprisingly, only the Ph-Taniaphos ligand[22] out of 140 ligands, provided significant turnover and high enantioselectivity, highlighting the challenging nature of this reaction. Translation of this lead to laboratory scale with Ph-Taniaphos at 0 °C afforded 3a in 50% assay yield (AY), with 85% ee (entry 1). Solvent 2-MeTHF resulted in improvement to 61% AY, while maintaining the ee (85%, entry 2). Next we screened the impact of the temperature on the reactivity and enantioselectivity in 2-MeTHF. Conducting the reaction at room temperature resulted in a drop in AY (48%), and ee (79%, entry 3). As the temperature was decreased from 0 °C to −20 °C, the AY increased to 67% at −10 °C (86% ee), then decreased to 56% at −20 °C (87% ee) (entries 4–5). We next examined the solvent composition.[18] Use of 30% toluene as cosolvent resulted in an increase in AY to 79% (86% ee, entry 6). Changing the ratio of the toluene to 2-MeTHF had a dramatic impact on the yield but maintained the ee (see Supporting Information for details). Increasing the concentration from 0.1 M to 0.2 M had a detrimental impact (entry 7) while decreasing to 0.05 M led to an increase in assay yield (84%, 86% ee, entry 8).</p><p>Palladium precursors can impact both activity and enantioselectivity in AAA.[3] Varying the palladium source from Pd(COD)Cl2 to Pd(OAc)2, Pd(dba)2, Pd(NCPh)2Cl2, [Pd(ally)Cl]2 (entries 9–12) revealed that Pd(OAc)2 resulted in the highest enantioselectivity (89% ee, 79% AY, entry 9). With Pd(OAc)2 we revisited the reaction temperature. As the temperature was decreased from −10 °C to −40 °C, the enantioselectivity increased to 91% at −30 °C (87% AY, entry 14). When the additive PMDTA was decreased to 0.5 equiv at −30 °C, the AY dropped to 63% (91% ee, entry 16). In contrast, the AY increased to 92% (92% ee) with 1.5 equiv. PMDTA (entry 17). The role of the PMDTA is likely to decrease the aggregation state of the LiN(SiMe3)2, facilitating the deprotonation of the arene complex.[23] The optimized reaction conditions in entry 17 afforded the AAA product in 92% yield and 92% ee, favoring the R enantiomer.[24]</p><p>With the optimized reaction conditions in hand, we next examined the scope of the nucleophiles in the AAA with cyclohexenyl-OBoc and cycloheptenyl-OBoc (Scheme 3). In addition to the toluene complex (3a), a variety of substituents on the η6-arene were well tolerated. To compare the selectivity of the (η6-C6H5CH3)Cr(CO)3 with an unactivated tolyl group, the 4,4'-dimethylbiphenyl complex was examined. The AAA °Ccurred exclusively at the chromium-activated position giving the product in 86% yield with 94% ee (3b). Substrates bearing 2-pyridyl, 2-thiophene, and N-pyrrolyl groups on the 4-position of the η6-arene underwent AAA reaction affording the desired product in 80–84% yield with excellent enantioselectivities (94–96%, 3d, 3f, and 3h). The aryl chloride containing substrate was also a good partner, giving the desired product in 83% yield and 94% ee (3j). The trifluoromethyl containing biaryl substrate was compatible with this AAA reaction, furnishing the product with 84% yield and 94% ee (3k). Chromium complexes of aryl chlorides exclusively give the AAA products in excellent enantioselectivity (94% ee, 3l).[25] In addition to the cyclohexenyl-OBoc derivative, seven-membered allylic carbonate (Scheme 3, n=2) was also a competent electrophile. With the same set of pronucleophiles, the corresponding products were generated in 55–85% yields with 80–87% ee (3c, 3e, 3g, and 3i).</p><p>To demonstrate the versatility of this approach, a two-step one-pot procedure to afford the chromium-free products was explored. [η6-(3-Chlorotoluene)]Cr(CO)3 gave the demetallated product in 76% yield and 91% ee (3m).[25] The (η6-3-methylanisole)Cr(CO)3 complex gave 72% yield with 91% ee (3n). The 4-(2-pyridyl) containing toluene complex was also a good substrate, providing the product in 74% isolated yield with 94% ee (3o). Comparison of metallated (3d) and demetallated (3o) products indicates that demetallation °Ccurs with loss of 6% yield and only 2% erosion in the ee. Diarylmethane derivatives are important motifs in pharmaceuticals and have found wide application in material science.[26] The diphenylmethane complex (η6-C6H5CH2Ph)Cr(CO)3 gave the AAA/demetalated product in 98% yield with 92% ee (3p). We also examined the benzyl amine complex, which afforded the allylation product as ~ 1:1 ratio of diastereomers in 96% yield with 83 and 93% ee (3q). These different ee's suggest that the AAA products (3q) do not epimerize under the reaction conditions.</p><p>We next investigated palladium-catalyzed allylic substitutions with acyclic allylic substrates, starting with (E)-tert-butyl (1,3-diphenylallyl) carbonate (Scheme 4). Given the significant difference between acyclic and cyclic substrates, it is not surprising that Ph-Taniaphos did not give high enantioselectivity. After rescreening a subset of ligands (see Supporting Information) we found that (R)-CTH-JAFAPHOS[27] (Scheme 4) was the leading hit. (E)-tert-Butyl (1,3-diphenylallyl) carbonate underwent AAA to furnish the corresponding products 3r (72% yield, >99% ee) and 3s (65% yield, 94% ee). Substrates containing 4-CF3, 4-Cl and 4-F were also compatible with the AAA affording the allylated products with excellent enantioselectivities (3t–3v, 92 to >99% ee). (E)-tert-Butyl (1,3-diethylallyl) carbonate exhibited lower enantioselectivity (3w, 63% yield, 64% ee).</p><p>To determine if our organolithium nucleophiles react via the "double inversion" pathway, the toluene complex was coupled with the cis-disubstituted stereoprobe (Scheme 5). The product (3x) was obtained in 64% yield with 64% ee (unoptimized). Comparison of the 1H NMR coupling constants with related compounds led to the conclusion that this reaction proceeded by the "soft" nucleophile pathway.[20]</p><p>To demonstrate the utility of our protocol, a nonsteroidal anti-inflammatory drug analogue (NSAIDs)[28] was prepared in two steps (Scheme 6). Beginning with the pyridyl-containing toluene complex 1c, AAA/demetalation as outlined in Scheme 3 gave compound 3y in >99% ee. The allylated product 3y was then converted to enantioenriched α-arylalkanoic acid 4 in 61% yield with >99% ee (Scheme 6).</p><p>In conclusion, we have successfully developed the first two catalysts for the palladium-catalyzed AAA employing toluene-derived pronucleophiles. This method provides ready access to the enantioenriched "α-2-propenyl benzyl" motifs and expands the classes of nucleophiles that can be employed in AAA reactions. It is noteworthy that the optimized catalysts for both AAA reactions advanced herein, and for our palladium(Cy-Mandyphos) catalyzed enantioselective arylation of (η6-C6H5CH2NR2)Cr(CO)3 complexes,[18] all contain strongly coordinating tertiary amine or amide groups on the ligands that can serve to bind the lithium counterions of the nucleophiles. This observation suggests a possible design feature for related enantioselective processes with strongly basic lithiated nucleophiles.</p>

PubMed Author Manuscript

Max-E47, a Designed Minimalist Protein that Targets the E-Box DNA Site In Vivo and In Vitro

Max-E47 is a designed hybrid protein comprising the Max DNA-binding basic region and E47 HLH dimerization subdomain. In the yeast one-hybrid system (Y1H), Max-E47 shows strong transcriptional activation from the E-box site, 5\'-CACGTG, targeted by the Myc/Max/Mad network of transcription factors; two mutants, Max-E47Y and Max-E47YF, activate more weakly from the E-box in the Y1H. Quantitative fluorescence anisotropy titrations to gain free energies of protein:DNA binding gave low nM Kd values for the native MaxbHLHZ, Max-E47, and the Y and YF mutants binding to the E-box site (14 nM, 15 nM, 9 nM, and 6 nM, respectively), with no detectable binding to a nonspecific control duplex. Because these minimalist, E-box-binding hybrids have no activation domain and no interactions with the c-MycbHLHZ, as shown by the yeast two-hybrid assay, they can potentially serve as dominant-negative inhibitors that suppress activation of E-box-responsive genes targeted by transcription factors including the c-Myc/Max complex. As proof-of-principle, we used our modified Y1H, which allows direct competition between two proteins vying for a DNA target, to show that Max-E47 effectively outcompetes the native MaxbHLHZ for the E-box; weaker competition is observed from the two mutants, consistent with Y1H results. These hybrids provide a minimalist scaffold for further exploration of the relationship between protein structure and DNA-binding function and may have applications as protein therapeutics or biochemical probes capable of targeting the E-box site.

max-e47,_a_designed_minimalist_protein_that_targets_the_e-box_dna_site_in_vivo_and_in_vitro

INTRODUCTION<!>RESULTS<!>The Max-E47 hybrid series targets the E-box in the yeast one-hybrid assay<!>All three Max-E47 hybrids exhibit strong binding affinities to the E-box in in vitro fluorescence anisotropy titrations<!>The Max-E47 hybrids effectively compete with native MaxbHLHZ for the E-box site in the modified yeast one-hybrid system<!>The Max-E47 hybrids do not interact with the c-MycbHLHZ in the yeast two-hybrid assay<!>The Max-E47 hybrids inhibit native MaxbHLHZ in a dominant-negative fashion<!>Structural basis of the Max-E47:E-box interaction<!>Differences between in vivo and in vitro measurements were also observed in other designed systems<!>Conclusions<!>EXPERIMENTAL SECTION<!>Construction of HIS3 and lacZ reporter strains<!>Construction of genes<!>Transformation of yeast cells<!>Construction of genes<!>3-AT titration test<!>Construction of genes<!>Western blot<!>Fluorescence Anisotropy Titrations<!>Determination of Kd Values<!>Supporting Information Available<!>

<p>The basic-region/helix-loop-helix/leucine zipper (bHLHZ) transcription factor (TF) family includes the ubiquitous Myc/Max/Mad network involved in 50% or more of known cancers and tumors.1,2 Heterodimeric Myc/Max is a transcriptional activator that binds the Enhancer box (E-box) sequence 5'-CACGTG, thereby regulating the expression of target genes critical for normal cell proliferation and differentiation.3–6 Myc proteins, which are proto-oncogenic and contain activation domains, do not homodimerize and do not bind to DNA unless partnered with Max, which does not possess an activation domain and serves to regulate Myc activity.5 In contrast, Max can homodimerize and bind the E-box.3 Thus, proteins that interfere with Myc/Max dimerization or its recognition of the E-box site may interfere with Myc's disease-promoting activities.</p><p>In recent years, use of a dominant-negative (DN) system to inhibit protein function has become increasingly popular. A number of different DN inhibitors of dimeric TFs has been described and mainly divided into two general classes.7,8 The first class of DN binds the same DNA target but lacks an activation domain (AD).9,10 Such a DN binds DNA but fails to activate transcription, and thus functions as a competitive inhibitor of the target protein for its cognate binding site.8,11,12 The Max homodimer belongs to this group of competitive inhibitors, acting as a dose-dependent antagonist of Myc function.13 The second class of DN lacks a DNA-binding domain14; such DNs heterodimerize with target proteins and prevent their DNA binding. For instance, Nasi and coworkers created a c-Myc derivative, Omomyc, capable of homodimerization as well as heterodimerization with c-Myc and Max; Omomyc sequesters c-Myc in complexes with poor DNA-binding ability and prevents heterodimerization with Max.15 For both classes of DN, overexpression is often necessary for efficient inhibitory activity.8,16</p><p>Owing to the importance of E-box regulation, we applied our minimalist strategy toward design of the first class of DN based on the protein α-helix, a straightforward molecular-recognition scaffold that targets the E-box and allows manipulation of gene expression at the level of the protein:DNA recognition event (Fig. 1). The Myc/Max network provides an excellent starting point for molecular design, as there exists much experimental data including high-resolution structures17–19; therefore, it serves as an ideal proof-of-principle to test our minimalist design strategy. Our aim is to generate smaller proteins of simplified structure compared to their native counterparts, while still retaining DNA-binding function. The Max-E47 hybrids comprise 66 amino acids, proteins easily accessible by either chemical synthesis or bacterial expression.</p><p>We created the hybrid Max-E47 by fusing the basic region of bHLHZ protein Max and HLH subdomain of bHLH protein E47: hence, we exchanged the DNA-binding regions and dimerization elements between two different protein families toward design of hybrid proteins that target the E-box. These hybrids were assayed for E-box binding function both in vivo and in vitro by yeast genetic assays and thermodynamic fluorescence anisotropy titrations. The yeast assays demonstrate that the Max-E47 series of hybrids is capable of E-box-responsive reporter gene activation and can compete with native MaxbHLHZ for the E-box target. Strong, specific binding of all three hybrids to the E-box site was measured by fluorescence anisotropy. Hence, the Max-E47 series of hybrids has the potential to serve as the first class of dominant negatives and inhibit the expression of E-box-responsive genes targeted by transcription factors such as the c-Myc/Max heterodimer.</p><!><p>Two protein mutants were obtained while cloning the gene encoding Max-E47: Max-E47Y contains a Val-to-Tyr mutation at position 385, and Max-E47YF contains an additional Val-to-Phe mutation at position 393 (Fig. 2). Experiments paralleling those for Max-E47 were performed for these fortuitous mutants as well.</p><!><p>We tested our design strategy in the yeast one-hybrid (Y1H) assay with the HIS3 and lacZ reporters. Activation of the HIS3 reporter was confirmed by growth on medium lacking histidine, whereas activation of the lacZ reporter was detected by two colorimetric assays: qualitative X-gal colony-lift filter assay and quantitative ortho-nitrophenyl-β-galactoside (ONPG) liquid assay.</p><p>Max-E47 showed comparably strong transcriptional activation from the E-box as the positive control native MaxbHLHZ: colonies appeared at 2 days on test plates in the HIS3 assay (Fig. 3A). Despite variability in colony numbers between plates, colony sizes were comparable to native MaxbHLHZ. Max-E47Y showed strong colony growth as well: colonies appeared 4 hours later. Max-E47YF activated from the E-box more weakly: colonies appeared at 4 days. In comparison, the truncated native MaxbHLH, without leucine zipper, shows no activation from the E-box in any of the yeast assays performed (data not shown). For all transformations, plasmids were extracted from positive transformants, subjected to restriction enzyme digest analysis, and sequenced. Both gel analysis and sequencing confirmed the correct identity of plasmids, which were retransformed into the YM4271[pHisi-1/E-box] strain. The same growth was observed, confirming the initial positive results.</p><p>For further confirmation, plasmids were transformed into strain YM4271[pLacZi/E-box], which employs reporter lacZ. Both the X-gal colony-lift filter assay and quantitative ONPG liquid assay confirmed the positive results from the HIS3 selection. In the colony-lift assay, blue color appeared for both Max-E47 and Max-E47Y at 20 minutes and turned bright blue at 2 hours, similar to native MaxbHLHZ (Fig. 3B). The color for Max-E47YF was much fainter but clearly above background (negative control pGAD424).</p><p>The ONPG assay quantitatively confirms the trend in transcriptional activation from the E-box of the three hybrids (Fig. 3C). The ONPG value for Max-E47 is 153.9±17.7, comparable to native MaxbHLHZ (147.4±7.3). Max-E47Y is somewhat lower at 101.0±5.5, but still gives a high value. Max-E47YF gives a much lower reading of 13.3±0.5; for comparison, the pGAD424 value is 7.0±0.5. The ONPG assay confirms that Max-E47 and Max-E47Y are strongly capable of transcriptional activation from the E-box.</p><!><p>Yeast genetic assays measure the E-box-responsive activity of our hybrids in a physiologically relevant, in vivo environment. However, these reporter assays rely on indirect means for detection of protein:DNA interactions, and the ONPG assay is not linear or stringently quantitative.21 Thus, we conducted quantitative fluorescence anisotropy titrations to measure free energies of protein:E-box complexation. Native MaxbHLHZ, Max-E47, Max-E47Y, and Max-E47YF were assayed with fluorescein-labeled 24-mer DNA duplexes (Fig. 4); no binding by any protein was detected with the nonspecific DNA control, even at 2 µM monomeric protein concentration (data not shown).</p><p>In contrast to the in vivo Y1H that shows the E-box-activation trend of Max-E47, Max-E47Y, and Max-E47YF from strongest to weakest, thermodynamic analysis gave comparable low nM Kd values for all four proteins binding to the E-box: native MaxbHLHZ at Kd 14.3±7.9 nM, Max-E47 at 15.3±1.6 nM, Max-E47Y at 8.7±3.3 nM, and Max-E47YF at 6.4±0.5 nM. Hill Coefficient analyses of binding isotherms show that all three Max-E47 hybrids have similar values to that of MaxbHLHZ; this indicates that the Max-E47 hybrids, like MaxbHLHZ, likely form dimeric structures for cooperative binding to the E-box site. These dissociation constants compare well with those reported by three different labs using electrophoretic mobility shift assay (EMSA),22 fluorescence anisotropy,23 or calorimetry24 for measurement of the MaxbHLHZ domain bound to the E-box, with Kd values in the 1–3 nM range. Thus, by fluorescence anisotropy, strong and specific binding to the E-box was measured by all four proteins, with no pronounced differences in DNA-binding function, as observed in the Y1H.</p><p>The Max proteins were difficult to manipulate in quantitative titrations. We suspect protein misfolding, and possible formation of soluble aggregates, may lead to nonfunctional protein. Therefore, we varied buffers and conditions to find a reliable environment for obtaining quantitative information. Protein misfolding and aggregation was also reported by those groups that measured the binding affinities of the MaxbHLHZ with the E-box,22–24 as well as in our own studies with other bHLH derivatives.25 Such intractability appears to be prevalent with the DNA-binding domains of the bHLH superfamily of transcription factors.</p><p>In our fluorescence analysis, the buffer system is believed to play an important role in maintaining stably folded protein structure leading to DNA-binding function. However, the chosen in vitro conditions may not allow discrimination of fine structural and functional differences between the three hybrids. Additionally, the sequences flanking the E-box site on the FA titration probes are not identical to those in the yeast reporter assays, and these differences likely affect the structure of the DNA ligand targeted by our hybrids. It is also possible that in the yeast system, in which proteins were expressed at low levels, the proteins were properly folded and stable; therefore, differences in DNA-binding activity could be distinguished. A detailed discussion of the discrepancy between in vivo and in vitro results is provided in the Discussion section.</p><!><p>We have proven that the Max-E47 hybrids can activate transcription from the E-box in the Y1H, and that they bind strongly and specifically to the E-box site by quantitative fluorescence analysis. The following results from the modified yeast-one hybrid (MY1H) assay demonstrate that the Max-E47 hybrids can potentially serve as competitive inhibitors of c-Myc/Max binding to the E-box.</p><p>Ideally, our yeast system would detect a positive signal from transcriptional activation from the E-box by the c-Myc/Max heterodimer; upon addition of Max-E47, this positive signal would be reduced. However, such a system is complicated, for it would involve protein/protein and protein:DNA interactions between three different proteins nonnative to yeast. A logical alternative is to test only two proteins: whether the Max-E47 hybrids can compete with the MaxbHLHZ homodimer for E-box binding. If the MaxbHLHZ homodimer's ability to activate from the E-box decreases after a Max-E47 hybrid is added, then by extrapolation, this hybrid is also likely to inhibit activation from the E-box site by the c-Myc/Max heterodimer.</p><p>In the traditional Y1H, only one protein can be expressed. We developed a MY1H that enables expression of two different proteins from the same plasmid with concomitant detection of transcriptional activation. We demonstrated that our MY1H is an ideal system for testing a second coexpressed protein's ability to inhibit the gene-regulatory activity of the first protein.26 In the MY1H, plasmid pCETT contains two multiple cloning sites (MCS): the gene cloned into MCSI is expressed as a fusion to the GAL4 AD, while the gene in MCSII has no AD. Both genes are governed by independent truncated ADH1 promoters, with low expression levels expected. Therefore, by use of pCETT in the MY1H system, the effects of a new protein, or mutant versions of a protein, on activation by a transcription factor can be readily examined.</p><p>The MY1H provides the most interesting test of the utility of our minimalist design, for direct competition between a Max-E47 hybrid and the native MaxbHLHZ simultaneously vying for the E-box target site can be assessed. The MaxbHLHZ gene was inserted into MCSI of pCETT: hence, Max is now a transcriptional activator. The genes for the Max-E47 hybrids were cloned into MCSII: hence, the expressed hybrids are repressors. If only AD+MaxbHLHZ is expressed, the homodimer's strong activation from the E-box will be visualized as a positive signal. If a Max-E47 hybrid is coexpressed with AD+MaxbHLHZ, it will compete for the E-box target: hence, repression of MaxbHLHZ:E-box interactions by a Max-E47 hybrid. Thus, the competitive binding of two proteins targeting the same DNA site can be detected based on the outcome of reporter transcription in the MY1H.</p><p>First, we note that all controls in the MY1H functioned properly. Positive control pCETT expressing AD+MaxbHLHZ showed strong activation: in the HIS3 reporter assay, colonies appeared at 2 days (Fig. 5A). In contrast, even with the same transformation efficiency, negative controls pCETT expressing only Max-E47, Max-E47Y, or Max-E47YF gave very small colonies at 6 days (data not shown). Theoretically, there should be no growth, as these hybrids have no AD. However, small colony growth might arise from interactions between the hybrids and endogenous proteins possessing activation domains.</p><p>For pCETT/AD+MaxbHLHZ//Max-E47YF (i.e. AD+MaxbHLHZ coexpressed with Max-E47YF), colony growth was strong, and sizes were similar to that of pCETT/AD+MaxbHLHZ, indicating no detectable reduction in transcriptional activation upon expression of Max-E47YF (Fig. 5A). (Note: the gene after "/" was inserted into MCSI, and the gene after "//" was inserted into MCSII.) For pCETT/AD+MaxbHLHZ//Max-E47Y, except for a few medium-sized colonies, much smaller colonies were observed. For pCETT/AD+MaxbHLHZ//Max-E47, only tiny colonies can be seen, and the colony number was greatly reduced. These results demonstrate that activation from the E-box by native MaxbHLHZ is strongly inhibited by Max-E47 and Max-E47Y. All plasmids had relatively equal transformation efficiencies (Fig. 5A), ruling out the possibility that less growth from coexpression with Max-E47 or Max-E47Y was caused by low transformation efficiencies. In addition, both proteins competing for the DNA target (i.e. AD+MaxbHLHZ and either Max-E47 or Max-E47Y or Max-E47YF) are expected to be produced in comparable amounts in the cells, despite the fact that the levels of protein expression are too low to be detected in the western blot analysis; this interesting observation indicates that a large excess of Max-E47 (or even Max-E47Y) is not required for efficient inhibition of native MaxbHLHZ binding to the E-box site. More discussion is provided in text below.</p><p>We titrated these transformants on plates containing inhibitor 3-AT (3-aminotriazole) to test their ability to inhibit activation from the E-box by native MaxbHLHZ. Colonies expressing AD+MaxbHLHZ grew well even on 80 mM 3-AT and could only be inhibited on 100 mM 3-AT (all data in this paragraph are shown in Fig. S2, Supporting Information). When Max-E47 was coexpressed with AD+MaxbHLHZ, 20 mM 3-AT was enough to inhibit colony growth completely. When Max-E47Y was coexpressed, some colony growth on 30 mM 3-AT was observed, but nothing can be seen on 40 mM 3-AT. When Max-E47YF was coexpressed, cells grew well even on 60 mM 3-AT, and 80 mM 3-AT was required for total inhibition of growth. Therefore, with increased concentrations of 3-AT, the growth of cells transformed with the different hybrids decreased as expected. These results support the conclusion that our Max-E47 hybrids competitively inhibit transcriptional activation by native MaxbHLHZ from the same E-box target. Though not quantitative, these results demonstrate the extent that the Max-E47 hybrids can inhibit native MaxbHLHZ.</p><p>Both the qualitative X-gal colony-lift filter assay and the quantitative ONPG liquid assay confirm the HIS3 assay results above in the MY1H. In the X-gal assay, pCETT/MaxbHLHZ turned blue rapidly (approximately 20 min) and became bright blue in 2 hours (Fig. 5B). Transformants coexpressing AD+MaxbHLHZ and Max-E47 or Max-E47Y displayed very faint blue color, suggesting that both hybrids inhibited activation by native MaxbHLHZ from the E-box to a high degree. The transformant containing pCETT/AD+MaxbHLHZ//Max-E47YF showed fairly strong blue color, indicating weak repression, consistent with Max-E47YF being the weakest activator from the E-box.</p><p>The ONPG assay quantitatively confirms these results. When only AD+MaxbHLHZ was expressed in pCETT, the ONPG value was 155.8±20.5 (Fig. 5C). This value is strongly consistent with that obtained for pGAD424/MaxbHLHZ (147.4±7.3) in the traditional Y1H (Fig. 3C), where a different plasmid was used with the same promoter. When Max-E47YF or Max-E47Y was coexpressed, the ONPG values decreased to 42.5±2.6 or 33.4±5.4, respectively. When Max-E47 was coexpressed, the value was even lower at 14.8±0.8: in comparison, background is 5.0±1.8 (pCETT). These quantitative measurements are consistent with the same trend observed in the qualitative HIS3 and X-gal assays, and confirm that Max-E47 is a stronger inhibitor of native MaxbHLHZ activation from the E-box than either Max-E47Y or Max-E47YF.</p><p>We emphasize that all of our yeast assays corroborate each other well and repetitively show consistent trends: these include the HIS3 selection assay and both lacZ-based assays (colony-lift and ONPG) in the traditional Y1H, and these same three assays in our MY1H. In the Y1H, Max-E47, Max-E47Y, and Max-E47YF showed E-box-responsive reporter gene activation as listed from strongest to weakest. In the MY1H, Max-E47 and the Y and YF mutants showed the same relative ability to compete with native MaxbHLHZ for occupying the E-box site. Thus, the same ordering of strongest to weakest ability to activate transcription from the E-box in the Y1H and to inhibit native MaxbHLHZ activation from the E-box correlate well between the two yeast systems.</p><!><p>To further investigate whether the Max-E47 hybrids can serve as competitive inhibitors of c-Myc/Max binding to the E-box, the yeast two-hybrid (Y2H) assay was used to test for protein/protein interactions between the Max-E47 hybrids and the MaxbHLHZ or c-MycbHLHZ domain. We chose to examine c-Myc, for it is the most highly characterized of the three Myc isoforms (c-, N-, and L-Myc), including high-resolution structural information.19</p><p>We initially tried to test the Max-E47 series' interactions with MaxbHLHZ by expressing the Max-E47 hybrids as GAL4 AD fusions and MaxbHLHZ as the GAL4 DBD (DNA-binding domain) fusion. However, cells transformed with recombinant pGADT7 plasmids containing the Max-E47 hybrids' genes or MaxbHLHZ gene died or grew very slowly. The reason is unclear. However, this phenomenon in the Y2H has been observed by many other researchers, as reviewed by Vidal and Legrain.27</p><p>In contrast, interactions with the c-MycbHLHZ domain were successfully measured by expressing c-MycbHLHZ as the GAL4 AD fusion, and native MaxbHLHZ and Max-E47 hybrids as GAL4 DBD fusions. Heterodimerization between AD+MycbHLHZ and DBD+MaxbHLHZ functioned as a positive control, as their association should reconstitute a functional transcriptional activator.</p><p>The Y2H demonstrated that neither Max-E47 nor Max-E47Y interacts with the c-MycbHLHZ (Fig. 6; Max-E47YF was not tested). On SD/-L/-W/-A/-H and SD/-L/-W/-A/-H/X-α-gal plates, the transformant that coexpresses DBD+MaxbHLHZ and AD+MycbHLHZ showed colony growth and blue color as expected, demonstrating that native MaxbHLHZ interacts with c-MycbHLHZ strongly and specifically. In contrast, the transformant that coexpresses DBD+MycbHLHZ and AD+MycbHLHZ resulted in no colony growth and appeared colorless, verifying that c-MycbHLHZ cannot homodimerize. Samples 8 and 9 in Fig. 6A show no cell growth or blue color, demonstrating no interaction between Max-E47 or Max-E47Y and c-MycbHLHZ. In this case, the HLH subdomain does not heterodimerize with the HLHZ subdomain. Fig. 6C summarizes all test results from the Y2H.</p><p>To exclude the possibility of no or low expression, western blot analysis was performed. Comparable expression levels were observed for all proteins (Fig. 6D and Fig S1, Supporting Information). SDS-PAGE analysis affirmed that all lanes in the western blot were comparably loaded (data not shown). Therefore, the lack of signal in the Y2H cannot be attributed to poor protein expression. These Y2H data show no interaction between Max-E47 (or Max-E47Y) and the c-MycbHLHZ, and by extension, no interaction between Max-E47 (or Max-E47Y) and the c-Myc/Max heterodimer. This result is consistent with no reported protein/protein interactions occurring between the bHLH and bHLHZ families.28,29</p><p>Both the Y1H and fluorescence analysis show that the Max-E47, Y, and YF homodimers are capable of binding the E-box. The Y2H demonstrates no protein/protein interactions between the c-MycbHLHZ domain and the Max-E47 series (bHLH). Thus, we conclude that inhibition of transcriptional activation from the E-box in the MY1H is likely due to the Max-E47 hybrids outcompeting native MaxbHLHZ for binding to the E-box site.</p><!><p>The in vivo assays in the traditional Y1H (HIS3 assay, X-gal colony-lift filter assay, and ONPG liquid assay) and the in vitro fluorescence analysis showed that Max-E47 activates transcription from the E-box site as strongly as does the MaxbHLHZ. MY1H assays demonstrated that Max-E47 effectively repressed E-box binding by MaxbHLHZ. These results consistently support the conclusion that Max-E47 is a strongly competitive dominant-negative inhibitor of MaxbHLHZ binding to the E-box site. In addition, Y2H assays showed no interaction between Max-E47 and c-MycbHLHZ. Therefore by extension, Max-E47 should be able to serve as a DN inhibitor of the native c-Myc/Max heterodimer that targets the E-box and regulates transcriptional activation.</p><p>Moreover, DN proteins are often expressed in excess relative to their targets for efficient inhibition; for example, Vinson and coworkers used a protein:DN ratio of 1:15.8 In our case, although the level of protein expression driven by the truncated ADH1 promoter is too low to be detected in the western blot analysis (as discussed in the Yeast Protocols Handbook, Clontech, 2001), both proteins competing for the DNA target are independently expressed from truncated ADH1 promoters, which are exactly same. In addition, we have demonstrated that in the MY1H system, the expression of the AD-fusion protein from MCSI is not affected by a nonsense control protein expressed from MCSII.26 Therefore, similar concentrations of both expressed proteins are expected in the cells. We find it highly noteworthy that even under these conditions, Max-E47 (and even Max-E47Y) can efficiently outcompete native MaxbHLHZ for binding to the E-box site. In addition, it is likely that Max-E47 can also inhibit N-Myc/Max and L-Myc/Max heterodimers as well, as these highly conserved Myc isoforms bind to the E-box by heterodimerization with Max.30,31</p><p>We observed strong correlation between the Y1H and MY1H systems: the strength of activation by the hybrid from the E-box in the Y1H correlated with that hybrid's ability to compete with native MaxbHLHZ for the E-box in the MY1H. The thermodynamic titrations support the conclusion that the hybrids are capable of strong, specific binding to the E-box, but the E-box binding trend so clearly observed in all the yeast assays was not replicated in the fluorescence analysis. It is not uncommon to find that results from in vivo and in vitro experiments are not consistent (discussed further below), although we emphasize that the fluorescence analysis corroborates our observations in the Y1H and MY1H that all three hybrids effectively target the E-box.</p><p>The Max-E47 series was obtained through swapping subdomains of the DNA-binding domains between the bHLHZ and bHLH families. Although subdomain swapping within the same protein family has been successful,32–35 it was unknown whether subdomain swapping between different families would lead to functional hybrids. While this work was in progress, Chapman-Smith and Whitelaw reported a subdomain swap to generate a hybrid comprising the bHLH domain from the Arnt bHLH/PAS protein and the leucine zipper from Max; their hybrid was shown to bind to the E-box by EMSA, but no quantitative binding assessment or in vivo work was reported.36 Our work provides another trial to explore the feasibility of subdomain swapping between different families, and both in vitro and in vivo results confirm the binding event.</p><!><p>All of the assays firmly validate that Max-E47 targets the E-box efficiently, thereby demonstrating the successful design of this hybrid protein. The bHLH requires two basic regions to bind DNA, which is achieved by dimerization.37,38 Therefore, any DNA-binding activity of Max-E47 depends upon dimerized structure via the E47 HLH subdomain, which we chose to use as the E47 HLH strongly homodimerizes. In contrast, most bHLH and bHLHZ proteins do not homodimerize, but rather heterodimerize. Thus, both Max and E47 serve comparable roles within their protein families, as they both homodimerize and heterodimerize to regulate partner protein activities.</p><p>The unique ability of E47 to homodimerize is tightly correlated with its structure. Ellenberger and coworkers compared their structure of E47 bound to 5'-CACCTG with the native Max bHLHZ:E-box structure17; their findings suggest that the HLH subdomain from bHLH proteins, as represented by E47, and the HLH subdomain from bHLHZ proteins, as represented by Max, have distinct structural features.20 Helix 1 of the E47 HLH is one turn longer than the analogous helix in Max and USF.17,39 This extra helical turn provides more dimer contact surface between E47 subunits by allowing salt bridge formation between His366 (near the C-terminus of Helix 1) and Glu390 (at the C-terminus of Helix 2'). The authors observe that this His/Glu pair is present in all E proteins (E47, E12, HEB, and Da) that homodimerize efficiently without a zipper. Another unique feature of E proteins is the triad of glutamines in the HLH. Gln373 participates in hydrogen bonds with the carbonyl oxygen of Gly360 (Helix 1) and side chains of Gln364 (Helix 1) and Gln381 (Helix 2). The authors emphasize that "this network of hydrogen bonds stabilizes the conformation of the loop as well as the orientation of Helices 1 and 2 within each subunit." These distinctive structural features contribute to the markedly enhanced stability of E47 homodimers.20</p><p>Interestingly, the in vivo yeast assays allow the differential activities between the three hybrids to be clearly distinguished. The transcriptional activation trend from the E-box is strongly consistent in all the yeast assays, even with different reporters and plasmids. These in vivo results can be explained by the E47 structure. Max-E47Y has one mutation, V385Y. According to Ellenberger and coworkers, this Val protrudes from Helix 2 and packs against the C-terminus of Helix 1.20 In their comparison of E47 with Max, they note that Val is relatively small, so it allows increased length of α-helical structure of Helix 1 in the E47 dimer, and the additional turn permits the His366-Glu390' salt bridge. However, the larger side chain of Tyr in Max-E47Y could distort this three-residue extension of Helix 1; incidentally, there is a Tyr at this same position in Helix 1 of Max. Therefore, this replacement likely alters the structure of the dimer by interfering with the His366-Glu390' interchain interaction, thereby affecting DNA-binding activity.</p><p>Max-E47YF has the additional V393F mutation. This mutation is only three residues from Glu390 and approximately on the same face of the α-helix. The large aromatic side chain of Phe likely interferes with the His366-Glu390' interaction as well. Moreover, this mutation occurs at the dimer interface and likely affects formation of the hydrophobic core. Val393 is also very close to the crossover point of Helices 2. At this critical junction, a mutation would be expected to affect dimerization ability negatively, consequently lowering DNA-binding activity.40 Thus, it is not surprising that Max-E47YF's E-box-responsive activity dropped measurably in the yeast assays. The E47 crystal structure has not been deposited in the Protein Data Bank, and therefore, determination of the precise positions of amino acids and their effects on protein structure is difficult to ascertain. This adds more challenge and risk to our design, but our results confirm the basic E47 structure as elucidated.20</p><!><p>That the yeast assays clearly delineate the differences in transcriptional activation capability between the three hybrids can be explained by the structural analysis above. In comparison, the thermodynamic titrations do not discriminate these differences in binding function, although they do affirm that the Max-E47 hybrids are high-affinity, sequence-specific binders of the E-box.</p><p>In fact, it is not uncommon that the levels of reporter gene activation by artificial transcription factors measured in cells do not correlate with their DNA-binding affinities measured in vitro, as shown by two examples involving artificial zinc-finger TFs. In their design of Zn-finger TFs for regulation of the endogenous human ERBB-3 gene, Barbas and coworkers found that the six-finger protein pE3Z, with the strong target-site binding affinity of 2 nM, was incapable of altering gene expression; in contrast, the six-finger protein pE3Y, which showed slightly weaker target-site binding affinity than pE3Z, was able to activate gene transcription.41</p><p>The authors offered that this discrepancy may be due to many factors, such as competition with cellular factors that bind to the same site and/or orientation of the Zn-finger fusion protein with respect to DNA. Similarly, in their study of VEGF gene regulation by Zn-finger TFs, Kim and coworkers observed no strong correlation between the levels of gene expression in their Y1H system and the Zn-finger:DNA binding affinities measured in vitro.42 The authors suspected that binding of another protein at the target site or the local chromatin structure may have rendered the target site inaccessible to the Zn-finger TF, which caused inconsistency between the in vivo and in vitro results. More recently, our group observed that the GAL4 AD fusion of the bHLH domain of bHLH/PAS protein Arnt did not activate E-box-responsive reporter gene expression in the Y1H, while fluorescence anisotropy showed that the same ArntbHLH domain bound to the E-box with Kd 40 nM.25 Misfolding of the ArntbHLH domain in the yeast cellular environment is the likely reason for its inability to activate reporter gene transcription, as circular dichroism showed little intrinsic structure for the ArntbHLH domain, which also proved to be highly insoluble during fluorescence titrations; likewise, an optimized buffer system was also believed to play an important role in improving and maintaining the protein fold.25</p><p>Native transcription factors can also show different activities in vivo and in vitro. Daignan-Fornier and coworkers showed that single-site mutants of the Bas1p DNA-binding domain discriminated between different promoter sequences in yeast, but bound equally well to the same promoters when evaluated by EMSA.43 The authors speculate that the mutations may affect promoter-specific interactions in vivo, and note that Bas1p and Bas2p (Pho2p) may need to interact cooperatively in order to activate transcription; the possibility that the mutations affected the concentrations of proteins in the yeast cells was shown unlikely by western blot analysis, which similarly showed that our Max-E47 series of proteins were also present in comparable concentrations in our Y2H experiments. Mutants of the N-terminal arm of the DNA-binding homeodomain of Bas2 can also distinguish among different promoter sequences in yeast and EMSA; Vershon and coworkers showed that their in vivo and in vitro binding studies generally correlated except with one mutant, and they suggest that other factors may contribute in the yeast assay.44 Although these are the closest examples we could find in the literature, the cases above of the designed Zn-finger are TFs are somewhat different from our system; the Zn-finger TFs are targeting different sites in a promoter, whereas ours are different TFs (the Max-E47 series or the Arnt derivatives25) targeting the same site. The examples involving Bas proteins may be complicated by additional interactions involved in activating transcription, which should not be an issue in our system. However, these examples demonstrate that in vivo and in vitro measurements are not always consistent, and that the reasons for discrepancy are unclear.</p><p>As for Max-E47 and its two mutants, we suspect two main reasons for the in vivo and in vitro differences in activity. One possibility is the differences in DNA sequences assayed: sequences flanking the E-box site on the FA titration probes were chosen to minimize potential for unintended 2° structure formation, like hairpins, and to minimize resemblence to the E-box sequence (i.e. to minimize binding at fortuitous E-box-like sequences). In comparison, the sequences targeted in the yeast assays comprised four tandem E-box sites cloned upstream of the HIS3 or lacZ reporter genes (multiple target sites are commonly integrated into the genome in yeast reporter assays; the manufacturer, Clontech, recommends three-six target sites; see Experimental Section and Supporting Information). Although the basic regions of our hybrids are unlikely to have direct contact with bases outside the E-box, such differences in flanking sequences can critically affect the structure of the DNA ligand recognized by protein; we note that for the Bas proteins and mutants, both in vitro and in vivo studies on target sites embedded within different promoters had significant effects on binding and transcription activities.44–46 A second possibility is that the low protein expression levels and the in vivo yeast environment provided conditions in which the fine structural differences, which lead to subtle differences in ability to activate transcription from the E-box, among the three hybrids could be distinguished. We emphasize, however, that our Max-E47 hybrids consistently displayed specific E-box-targeting activity both in vivo and in vitro, and that all our data confirms the conclusion that these hybrids are Class 1 DN inhibitors.</p><p>Although the three hybrids bind to the E-box with comparable strengths as measured by fluorescence anisotropy, DNA-binding affinity is only one of many factors that affects reporter activation. In other words, these one or two mutations in Max-E47 may alter protein structure and stability or accessibility of the protein to its DNA target in vivo; a strong DNA-binding affinity (thermodynamics) does not necessarily mean that the transcription factor stays on its DNA target long enough to trigger reporter gene activation (kinetics). In addition, interference from endogeneous proteins might also cause the functional discrimination between the three hybrids. It is also possible that in the yeast system, where proteins were expressed at low levels, the proteins were properly folded and stable; therefore, differences in reporter gene activation could be distinguished. However, the in vitro conditions of the fluorescence measurements, chosen to maintain protein solubility during the lengthy titration (each data point, after addition of protein aliquot, required overnight incubation to maintain solubility), may not have been optimal for distinguishing differences in DNA-binding function, perhaps by diminishing the fine structural differences between the three hybrids. This fact, again, proves the necessity of performing both in vitro and in vivo measurements in the study of DNA-binding proteins.</p><p>According to the Max bHLHZ:E-box crystal structure, Lys57 in the loop nonspecifically contacts the DNA phosphodiester backbone.17 Burley and coworkers note that this interaction is significant for the Max:E-box complex, but this interaction does not exist with the E47 loop.20 However, Max-E47 still targets the E-box as well as does native MaxbHLHZ in the Y1H, MY1H, and quantitative fluorescence analysis, despite loss of this important interaction. Additionally, the Max-E47 hybrid, which lacks a leucine zipper, can target the E-box site as efficiently as does native MaxbHLHZ, which requires its zipper for dimerized structure and DNA-binding function; the truncated MaxbHLH is not functional in the Y1H. Therefore, Max-E47 provides a useful scaffold for further exploration of the relationship between protein structure and DNA-binding function.</p><!><p>These hybrids of Max and E47 are part of our effort to generate minimalist proteins with desired DNA-recognition capabilities from an α-helical molecular recognition scaffold—hence, protein-based tools for recognition of desired DNA targets. Our minimalist design strategy provides a launching point for generation of artificial transcription factors based on native proteins that are likely to be easier to express or synthesize than their native counterparts, more tractable for high-resolution studies, and may have further applications in fields other than protein design, including drug discovery and functional genomics.47–49 Already, artificial Zn-finger transcription factors have been reported.42,50–53</p><p>We chose to apply our minimalist design strategy to the Max/Myc:E-box network, given its broad involvement in normal cellular function, as well as the etiology of cancers and tumors. Although some of Myc's normal cellular activities have recently been found to be independent of Max and cannot be explained by activation from E-box-responsive Myc targets,54 the short, simplified Max-E47 hybrids may find utility as DN inhibitors of undesirable transcriptional activation from the E-box. Very recently, the designed dominant-negative Omomyc, discussed in the Introduction, was shown to target Myc specifically in a transgenic mouse model of cancer.55 Omomyc shows the promise of protein-based drugs against disease. Expression of Max-E47 in mammalian cells could provide a bHLH protein capable of interfering with c-Myc's transactivation potential by targeting the E-box DNA site, and these next-generation experiments are being explored. These hybrid proteins may serve as leads for the design of smaller proteins or peptidomimetics with desirable pharmacological properties.</p><!><p>More experimental details for all procedures are provided in the Supporting Information.</p><!><p>Four tandem copies of the E-box target sequence (5'-CACGTG) were cloned into the pHISi-1 integrating reporter vector at the his3-200 locus of S. cerevisiae YM4271 (Matchmaker™ One-Hybrid System, Clontech, Mountain View, CA). 10 mM 3-AT (3-aminotriazole, Bioshop, Burlington, ON) was sufficient to suppress background due to leaky His3 expression in reporter strain YM4271[pHISi-1/E-box]. Similarly, reporter strain YM4271[pLacZi/E-box] was constructed with four copies of the E-box site upstream of the lacZ reporter gene.</p><!><p>DNA oligonucleotides were purchased from Operon Biotechnologies (Huntsville, AL). The genes for expression of native MaxbHLHZ (92 aa, residues 22–11317), c-MycbHLHZ (87 aa, residues 22–107, numbering from ref. 17 according to Ziff and coworkers,56) native MaxbHLH, and Max-E47 were synthesized in a single PCR reaction.57 Amplified gene inserts were inserted into vector pGAD424 (Matchmaker One-Hybrid System, Clontech). Recombinant plasmids were transformed into E. coli strain DH5α and sequenced.</p><!><p>For the HIS3 assays, we developed an electroporation protocol based on the methods of Suga and Hatakeyama.58,59 After electroporation, cells were plated on minimal selective medium (SD, Synthetic Dropout) lacking Leu and His with the appropriate amount of 3-AT to suppress background (Matchmaker™ One-Hybrid System). For the assays using the lacZ reporter, plasmids were transformed into the integrating reporter strain YM4271[pLacZ/E-box] by the TRAFO method.60 Protein:DNA interactions were detected by X-gal colony-lift filter assay and ONPG liquid assay. These protocols are provided in the Yeast Protocols Handbook (Clontech).</p><!><p>The genes for expression of native MaxbHLHZ and the hybrids were inserted into vectors pCETT and pCETF.26</p><!><p>One colony was resuspended in 1 mL sterile H2O and vortexed vigorously to disperse the cells. For testing of inhibitory activity, 10 µL of cells were pipetted on SD/-L/-H plates containing different concentrations of 3-AT (0–200 mM).</p><!><p>The genes amplified from recombinant pGAD424 plasmids were inserted into vectors pGADT7 (for MaxbHLHZ) or pGBKT7 (for Max-E47 and Max-E47Y). Cotransformation of the recombinant pGBKT7 and pGADT7 plasmids into strain AH109 was performed by the TRAFO method.60 After 3–4 days growth, one colony was resuspended in 1 mL sterile H2O and vortexed vigorously. A sterile inoculating loop was dipped into the cell dispersion, and cells were spread on SD/-L/-W plates to confirm healthy cell growth and SD/-L/-W/-H/-A and SD/-L/-W/-H/-A/X-α-gal plates for testing.</p><!><p>AH109 cells were transformed with the recombinant pGADT7 and pGBKT7 and grown to exponential phase in YPDA media. Cells were lysed by sonication and separated by SDS-PAGE. Immunodetection was performed with anti-c-Myc or anti-HA antibody (Covance Inc., Princeton, NJ) and visualized by fluorescence on a Molecular Dynamics Storm 840 PhosphorImager.</p><!><p>The genes for MaxbHLHZ and the hybrids were reconstructed in codons preferred for bacterial expression, cloned into pET-28A(+) (Novagen, Mississauga, ON), expressed from BL21(DE3)pLysS (Stratagene, La Jolla, CA), purified by TALON metal ion affinity chromatography (Clontech) and reversed-phase HPLC (Beckman, Fullerton, CA), and identities confirmed by ESI-MS (see refs. 61, 62 for detailed protocols). The 6-carboxyfluorescein label (6-FAM) was incorporated at the 5' end of the labeled oligonucleotides (Operon Biotechnologies, Huntsville, AL), and all oligonucleotides were purified by HPLC. Protein concentrations were assessed (Beckman DU 640 UV/vis spectrophotometer) by Tyr absorbance for MaxbHLHZ, Max-E47Y, and Max-E47YF (absorbance maximum 275–280 nm, e275 = 1405 M−1•cm2212;1 per tyrosine) or by measurement at 205 nm and 280 nm by the method described by Scopes for Max-E47.63</p><p>Fluorescence was measured on a JY Horiba Fluorolog-3 spectrofluorimeter (University of Toronto). The cell (Starna, Atascadero, CA) contained 1 nM DNA duplex in 4.3 mM Na2HPO4, 1.4 mM KH2PO4, 150 mM NaCl, 2.7 mM KCl, 1 mM EDTA, 800 mM urea, 20% glycerol, 0.1 mg/mL acetylated BSA, and 100 µM bp calf thymus DNA. Several other buffers, including Tris- and HEPES-based buffers, were explored but did not maintain functional protein. The volume change was maintained at <5% of total volume. For each data point, the sample was incubated at 4 °C overnight followed by at least 20 min at room temperature; such extensive incubation was necessary to minimize protein misfolding and aggregation.</p><!><p>The polarization values were used to calculate apparent dissociation constants using Kaleidagraph 3.6 (Synergy software) and Eqn. (1)64: (1)P=((Pbound−Pfree)[M]/(Kd+[M]))+Pfree where Kd corresponds to the apparent monomeric dissociation constant, M is the concentration of monomeric protein, Pfree is the polarization for free DNA, and Pbound is the maximum polarization of specifically bound DNA; two independent titrations (R values >0.950) were performed for each Kd value ± SEM (standard error of the mean).</p><!><p>Detailed experimental protocols for all yeast work, fluorescence spectroscopy, and binding analysis; yeast data, including Y2H western blot and HIS3 titration. This material is available free of charge from the JACS home page (http://pubs.acs.org/JACS).</p><!><p>Schematic of the minimalist design strategy. The c-Myc transcriptional activator must heterodimerize with Max in order to bind to the E-box site; c-Myc is proto-oncogenic, so activation at the E-box can lead to disease. Our Max-E47 series of hybrids can effectively compete for binding at the E-box and may serve as dominant negative competitors of native c-Myc/Max, thereby inhibiting activation from the E-box.</p><p>Sequences of the Max-E47 hybrids. The three highly conserved basic-region residues that make sequence-specific contacts to DNA major groove bases are in bold. The mutated amino acids are in bold, underlined. The numbering follows that used by Ellenberger et al.20</p><p>Max-E47 hybrids activate transcription from the E-box (Y1H). (A) The HIS3 assay of Max-E47 hybrids expressed in pGAD424. SD/-H/-L + 10 mM 3-AT plates were incubated at 30 °C for 6 days. a. pGAD424 (negative control); clean. b. pGAD424/native MaxbHLHZ (positive control). Note the colonies were too crowded to grow larger. Same amounts were plated on all plates for comparison. c. pGAD424/native MaxbHLH. d. pGAD424/Max-E47. e. pGAD424/Max-E47Y. f. pGAD424/Max-E47Y. (B) The X-gal colony-lift filter assays of Max-E47 hybrids. All of the SD/-U/-L plates were incubated at 30 °C for 4 days before testing. Photos were taken after 2 hours incubation. a. pGAD424 (negative control); very faint blue. b. pGAD424/native MaxbHLHZ (positive control); vivid blue at 20 minutes. c. pGAD424/Max-E47; vivid blue at 20 minutes. d. pGAD424/Max-E47Y; vivid blue at 20 minutes. e. pGAD424/Max-E47YF; faint blue. (C) Histogram comparing the binding strengths of Max-E47 hybrids to E-box. All values are averages of 9–12 measurements (± standard deviation) from 3–4 separate cell-growth cultures.</p><p>(top) DNA duplexes used in fluorescence anisotropy titrations. "6-FAM" is 6-carboxyfluorescein, and the Max-preferred E-box is underlined (core E-box is CACGTG). (bottom) Representative equilibrium binding isotherms for native MaxHLHZ (●, blue line), Max-E47 (▲, red line), Max-E47Y (▼, green line), and Max-E47YF (△, black line) targeting the E-box. Each isotherm was obtained from an individual titration, and each Kd value is the average of two individual titrations ± SEM.</p><p>Max-E47 hybrids inhibit native MaxbHLHZ activation from the E-box (MY1H). (A) The HIS3 assay of the inhibition of native MaxbHLHZ by the Max-E47 hybrids. Plates a–d are transformations plated on SD/-H/-L + 10 mM 3-AT plates, which were incubated at 30 °C for 6 days. a. pCETT/native MaxbHLHZ. b. pCETT/native MaxbHLHZ//Max-E47. c. pCETT/native MaxbHLHZ//Max-E47Y. d. pCETT/native MaxbHLHZ//Max-E47YF. Plates a'–d' are the corresponding SD/-L efficiency plates, which were incubated at 30°C for 4 days. (B) The X-gal colony-lift filter assay. All SD/-U/-L plates were incubated at 30 °C for 4 days. Photos were taken after 2 hours incubation. a. pCETT/native MaxbHLHZ. b. pCETT/native MaxbHLHZ//Max-E47. c. pCETT/native MaxbHLHZ//Max-E47Y. d. pCETT/native MaxbHLHZ//Max-E47YF. (C) Histogram comparing the Max-E47 hybrids inhibition of native MaxbHLHZ activation from the E-box. All values are averages of 9–12 measurements from 3–4 separate cell-growth cultures.</p><p>Max-E47 hybrids do not interact with c-Myc. (A) Y2H assay. For the same sample, the same cell density was plated on all three plates. The plates were incubated at 30 °C for 6 days. a. SD/-L/-W plate; colonies appeared at 2 days. b. SD/-L/-W/-H/-A plate; colonies appeared at 3 days. c. SD/-L/-W/-H/-A/X-α-gal plate; the blue color developed at 4 days. (B) Sample alignment on each plate in Fig. 4A. (C) Summary of Y2H results. The indicated pGBKT7- and pGADT7-encoded proteins were coexpressed in yeast and then tested for adenine and histidine auxotrophy, as well as expression of α-galactosidase after 6 days incubation. The main results are highlighted. (D) Western blot of Y2H. Lane 1: pGADT7 (=GAL4AD) + pGBKT7 (=GAL4DBD), Lane 2: pGADT7/c-MycbHLHZ (=GAL4AD+c-MycbHLHZ) + pGBKT7/c-MycbHLHZ (=GAL4DBD+c-MycbHLHZ), Lane 3: pGADT7/c-MycbHLHZ (=GAL4AD+c-MycbHLHZ) + pGBKT7/MaxbHLHZ (=GAL4DBD+MaxbHLHZ), Lane 4: pGADT7/c-MycbHLHZ (=GAL4AD+c-MycbHLHZ) + pGBKT7/Max-E47 (=GAL4DBD+Max-E47), Lane 5: pGADT7/c-MycbHLHZ (=GAL4AD+c-MycbHLHZ) + pGBKT7/Max-E47Y (=GAL4DBD+Max-E47Y). (The HA-tagged western, Fig. S1, which proves comparable c-MycbHLHZ expression levels in all samples, is provided in the Supporting Information.)</p>

PubMed Author Manuscript

Dissect interaction kinetics through single-molecule interaction simulation

The ability to extract kinetic interaction parameters from single-molecule fluorescence resonance energy transfer trajectories without the need for solving complex single-molecule differential equations has the potential to address some of the critical biophysical questions. Here, we provide a single-molecule interaction simulation (SMIS) tool to give the expected dwell-time distributions and relative populations of each FRET states based on the assigned kinetic model and to dissect kinetic interaction parameters from single-molecule FRET trajectories. The method provides the expected dwell-time distributions, averaged transition rates, and relative populations of each FRET states based on the assigned kinetic model. Comparing extensive simulated data with experimental data enables the quantification of the kinetic rate and equilibrium constants. We have also demonstrated that SMIS is useful to quantify the interaction kinetic rate constants that were originally unobtainable through the traditional single-molecule analytical solution approach.

dissect_interaction_kinetics_through_single-molecule_interaction_simulation

<!>EXPERIMENTAL SECTION<!>Derive analytical solutions of kinetic properties of the kinetic model<!>Simulate kinetic properties of the kinetic model using single-molecule interaction simulation (SMIS)<!>Define the kinetic model and transition probability from the rate constants (Step 1).<!>Build a sequence of dwell time for each species (Step 2).<!>Associate species with FRET states and generate a single-molecule FRET trajectory (Step 3).<!>Generate the probability density function of dwell time PDF(\xcf\x84), average transition rates and, relative populations.<!>Validation of SMIS using Michaelis-Menten enzyme kinetic model<!>Application of SMIS to kinetic model without analytical solutions<!>CONCLUSION

<p>Single-molecule fluorescence resonance energy transfer (or smFRET) is a powerful biophysical technique to dissect stochastic interactions of protein-substrate, [1–3] protein-protein, [3–5] and protein-DNA complexes,[6–10] as well as the folding behaviors of DNA[11,12], RNA[13], and protein[14–17]. smFRET typically quantifies the number of FRET states and the waiting times of specific state transitioning to other states. These pieces of information provide microscopic insight into the interaction kinetics. Many methodologies have been created to reliably distilled out the FRET state identifications and transitions[18–22]. These approaches enable the experimental determination of the probability density function of waiting times PDF(τ), which is typically obtained from the histogram of microscopic dwell times. It represents the overall lifetime of species within a FRET state. Potentially, analyzing the PDF(τ) enables the quantification of rate constants[23], which reveals the mechanism of the interaction process.</p><p>Extracting out the interaction rate constants from the PDF(τ) is typically achieved via assigning a kinetical model and fitting the PDF(τ) by the analytical probability density function f(τ). For example, Jain's laboratory studied the kinetics of self-assembled monolayer formation on individual nanoparticles and extracted the formation rate constants of self-assembled thiol monolayer[24]. Chen's laboratory discovered how transcription factors regulate transcription process in vitro and in cells[25–27]. They also provided a detailed single-molecule kinetic theory for heterogeneous catalysis[28–30]. Scherer's laboratory revealed the kinetics and mechanism of the physical passing of particles in an optical ring trap with an adjustable driving force[31]. Landes' laboratory reported the multistep desorption kinetics of α-lactalbumin from Nylon, which provided insight into the mechanisms driving protein-polymer interactions[32].</p><p>However, the determination of the kinetical model and the derivation of analytical solutions for kinetic parameters may not always be straightforward. In general, the process begins with the selection of the kinetic model based on the features of observed FRET states. This kinetic model generates single-molecule rate equations. With sufficient boundary conditions, one can solve the rate equations to obtain the analytical probability density function f(τ) for transitions between different FRET states, as well as the analytical solutions of relative populations of each FRET state. Unfortunately, having sufficient boundary conditions could be difficult due to too many species co-existing in the same FRET or the kinetic model containing repeated differential equations. The difficulties getting the f(τ) and relative populations of each FRET state hinder the mechanistic study of interaction processes. One possible alternative way to address this issue is to compare the experimental PDF(τ) with the corresponding simulated results of various kinetic models with wide-range rate constants. This approach does not require any boundary conditions and thus can broadly apply to dissect stochastic interaction kinetics. This idea has been demonstrated at the ensemble level. Software such as Berkeley Madonna has been developed for modeling and visualization of chemical reactions[33–36]. However, a systematic way to simulate these kinetic parameters using single-molecule interaction FRET events is still lacking.</p><p>Here, we provide a single-molecule interaction simulation (SMIS) tool to provide the expected kinetic characteristics of each FRET state based on the assigned kinetic model, and to dissect interaction kinetics from single-molecule FRET trajectories. In the EXPERIMENTAL SECTION, using the two-state kinetic model, we derived the analytical probability density function and the relative equilibrated populations of each FRET state. We then introduced SMIS and provided a step-by-step simulation of the single-molecule interaction FRET events, the probability density of dwell times, average transition rates, and the relative population of each FRET state. In the RESULTS and DISCUSSIONS, the general solutions from the two-state kinetic model were then converted into the analytical solutions of the well-known Michaelis-Menten enzyme kinetic model to validate the SMIS. Finally, we further demonstrated a successful application of SMIS to quantified kinetic rate constants that were originally unobtainable through the traditional single-molecule analytical solution approach.</p><!><p>The FRET data directly associates the interacting species in the proposed kinetic model with the observed FRET states. Using a kinetic model containing three interacting species in two FRET states (Figure 1a) as an example, we demonstrated the derivation of dwell-time distributions, average transition rate, and relative populations of each interacting species through the analytical approach or the simulation approach.</p><!><p>Figure 1a shows the kinetic model describing an enzyme existing as one of the three interacting species (E, ES*, and ES) associated with the high (IH) and low (IL) FRET states. The substrate (S) bound to the enzyme (E) to form the interacting complex ES through an intermediate ES* with the forward (k1, k2, and k3) and reversed (k−1, k−2, and k−3) rate constants annotated. Assuming in the single-molecule FRET experiment, we observed the highly fluorescent enzyme E interacts substrates, forms a fluorescent intermediate ES*, and eventually generate a weakly-fluorescent product ES. Under constant laser illumination, one can detect these steps at the single-molecule level in real-time. Figure 1b shows the typical single-molecule trajectory that reflects these processes through the stochastic on-off burst like signals. Each FRET-efficiency increase marks the presence of E or ES*; each decrease marks a formation of ES. The τHL, the dwell time on the IH state before transitioning to the IL state, reports the microscopic dwell times for completing steps that involve k1° (i.e., k1[S]), k−1, k2, and k3° (i.e., k3[S]). τLH, the dwell time on the IL state before transitioning to the IH state, reports the microscopic dwell times for completing steps that involve k−2 and k−3. The distributions of these two stochastic observables (τHL and τLH) provide valuable kinetic information and molecular insights to the interaction mechanism.</p><p>To obtain the kinetic rate constants analytically, we derived the probability density function of dwell times f(τ) by solving the single-molecule rate equations under proper initial conditions. For example, to derive the probability density function of τHL, fHL(τ), we wrote out the single-molecule kinetic equations based on the processes (Figure 1a middle) that lead to the transition from the IH to IL state. Entering the IL state can occur either through the E to ES (pathway involves k3°) or ES* to ES (pathway involves k2) and relevant to PE(t)1, PES*(t)1, PES(t)1, PE(t)2, PES*(t)2, and PES(t)2. The overall probability function of time is a linear combination of both pathway Pi(t) = C1Pi(t)1 + C2Pi(t)2 (i ∈ [E, ES*, ES]), which C1 and C2 are the probability coefficients for two different initial conditions (Supporting Information S1.1). We evaluated the probability density function fHL(τ) from the Pi(t). fHL(τ) dictates the probability density of finding a dwell time τHL for the transition from the IH to the IL state. The probability for finding a particular dwell time τHL is fHL(τ)Δτ, which equals the probability of switching from IH to IL between t = τ and τ + Δτ. Since the transitions from the IH to the IL state only occur via E to ES or ES* to ES pathways, fHL(τ)Δ τ can be estimated from the overall ΔPES(τ). In the limit of infinitesimal Δτ, the analytical expression of overall probability density function fHL(τ) is fHL(τ)=C1dPES(τ)1dτ+C2dPES(τ)2dτ (Supporting Information S1.2). Eq 1 shows the final analytical expression of fHL(τ). (1)fHL(τ)=D1e(B+A)τ+D2e(B−A)τ Where A=(k−1+k3°+k2+k1°)2−4(k3°k−1+k2(k1°+k3°))2, B=−(k3°+k2+k1°+k−1)2, D1=(k2(A+B+k1°+k3°)+k3°k−1)k−2+(k1°k2+k3°(A+B+k2+k−1))k−32A(k−2+k−3), and D2=(k2(A+B+k2+k−1)−k3°k−1)k−2+(−k1°k2+k3°(A+B+k1°+k3°))k−32A(k−2+k−3).</p><p>Finally, with the fHL(τ), we evaluated the average transition time ⟨τHL⟩ by 〈τHL〉=∫0∞τfHL(τ)dτ, whose reciprocal value reports the average transition rate (Eq 2). (2)〈τHL〉−1=(k−2+k−3)(k1k2+k3(k2+k−1))[S]k−3(k2+k−1+k1[S])+k−2(k−1+(k1+k3)[S])</p><p>Similarly, we evaluated the PLH(t), fLH(τ), and ⟨τLH⟩−1 for the transitions from the IL to IH state (Supporting Information S2). Eq 3 and Eq 4 summarize the final analytical expression of fLH(τ) and ⟨τLH⟩−1, respectively. (3)fLH(τ)=(k−2+k−3)e−(k−2+k−3)τ (4)〈τLH〉−1=(∫0∞τfLH(τ)dτ)−1=k−2+k−3</p><p>With these analytical solutions, we can predict the fHL(τ) and fLH(τ) and test the effect of substrate concentration, [S], on the ⟨τHL⟩−1 and ⟨τLH⟩−1. Figure 1d–f show these predictions for the below conditions (Figure 1c): (1) all rate constants are similar; (2) ES* existing as a transient complex; (3) ES* existing a stable complex; (4) ES existing as a transient complex; and (5) ES existing as a stable complex. Since most of the rate constants range from 10−2 to 102, we used rate constants of 1 s−1 to predict the kinetic properties. We then varied the rate constants to simulate ES* and ES existing as the transient or stable complexes. For the stable complex condition (i.e., formation rate constant is larger than dissociation rate constant), we set the formation rate constants to ~ 102 s−1 and the dissociation rate constants to ~ 10−2 s−1. On the other hand, formation rate constants are ~ 10−2 s−1 and the dissociation rate constants to ~ 102 s−1 for the transient complex condition. Figure 1d and 1e top panels show the fHL(τ) and ⟨τHL⟩−1 for these conditions. The fHL(τ) predicts the multiple exponential binding behaviors and the ⟨τHL⟩−1 predicts the hyperbolic dependence of [S] on the ES formation rate. On the other hand, Figure 1d and 1e bottom panels show the fLH(τ) and ⟨τLH⟩−1. Since the transitions from the IL to IH state represent the dissociation of ES, the log-log plot of fLH(τ) predicts that the dissociation of ES follows the unimolecular dissociation mechanism.</p><p>Thermodynamic parameters such as the relative subpopulation of each species were also derived. PE([S]), PES([S]), and PES*([S]) are the relative subpopulation of E, ES*, and ES, respectively. Using the two-state kinetic model (Figure 1a), the single-molecule rate equation of each species were drafted based on the generation and consumption of the species (Supporting Information S1.3). When reaching equilibrium, the relative population of each species remains constant. This fact means that dPi(t)dt=0 (i ∈ [E, ES, ES*]). Furthermore, the sum of the relative subpopulation of all species is always equal to 1. By setting up experiments on different [S], together with these boundary conditions, we obtained the populations of E, ES*, and ES. One can expect the system reaches equilibrium by considering the time-dependent populations at t = ∞. We summarized PE([S]), PES*([S]), and PES([S]) in Eq S36–Eq S38. Since E and ES* contribute to the IH while ES to the IL, we also obtained the PHL([S]) and PLH([S]) as shown in Eq 5 and Eq 6, respectively. Figure 1f summarized the [S] dependent PHL([S]) and PLH([S]). (5)PHL([S])=k2k−3+k3k−2[S]+(k−2+k−3)(k−1+k1[S])k2(k−3+(k1+k3)[S])+k3k−2[S]+k1(k−2+k−3)[S]+k−1(k−2+k−3+k3[S]) (6)PLH([S])=(k1k2+k3(k2+k−1))[S]k2(k−3+(k1+k3)[S])+k3k−2[S]+k1(k−2+k−3)[S]+k−1(k−2+k−3+k3[S])</p><!><p>To obtain the kinetic properties through simulation, we created the single-molecule interaction simulation (SIMS) tool. We simulated single-molecule interaction trajectories in MATLAB by following the procedures below (Figure 2):</p><!><p>We first define the kinetic model by assigning the number of interacting species in each FRET state and assign rate constants for transitions between interacting species. For example, E, ES*, and ES specify the species in the model (Figure 2a). The kI,J represents the rate constant for interconversion from the state I to J (J ≠ I; and I, J ∈ [E, ES*, ES]). In other words, kE,ES* is the k1°, kES*,ES is the k2, kES,E is the k−3, and vice versa. The transition from species I to a J follows the relative probability kI,J/ΣI kI,J.</p><!><p>With the kinetic model defined, we can define the sequence of species (e.g., E→ES*→E→ES*→ES→E→ES*→E→ES*→ES) based on the transition probability. The dwell times of each species follows ΣI kI,J exp (−ΣI kI,Jt), where the ΣI kI,J is a rate-constant sum of all competing pathways leaving from species I to J (J ≠ I). We randomly sample one dwell time from the dwell-time distribution of ΣI kI,J exp (−ΣI kI,Jt) and assign it to each species and generated the sequence of dwell time (e.g., τ→τES*→τE→τES*→τES→τE→τES*→τE→τES*→τES).</p><!><p>The dwell time for each species in the sequence is associated with FRET states. For example, by assigning E and ES* in the high; ES in the low FRET state, and combining dwell times belonging the same FRET state, we can convert the τE→τES*→τE→τES*→τES→τE→τES*→τE→τES*→τES sequence into single-molecule FRET trajectory of τHL→τLH→τHL→τLH.</p><!><p>With single-molecule trajectories, we can extract microscopic dwell times (i.e., τHL and τLH). Normalizing the histograms of the dwell times by the overall area generates the probability density function of dwell times. The average FRET-state transition time can be calculated from 〈τi〉=∑τiNi (i ∈ [HL, LH], Ni,: number of dwell-times), whose reciprocal value reports the average transition rate between FRET states. Using single-molecule FRET trajectories, we can generate the relative population of each species Pi (i ∈ [E, ES*, and ES] ). Pi, is calculated by dividing the sum of microscopic dwell times (i.e. ΣτE, ΣτES* or Σ τES) with the length of the trajectory (i.e., Pi=∑τi∑τ). Similarly, we can also generate PHL and PLH (i.e., PHL=∑τHL∑τ, PLH=∑τLH∑τ).</p><!><p>We validated the SMIS by comparing our simulation with 200,000 dwell times to the well-known Michaelis-Menten enzyme kinetic model with rate constants varies from 0.01 to 100 s−1. Michaelis-Menten model is a simplified condition of our two-state model where k−2 and k3 equal to zero. By replacing k−2 and k3 in Eq 1 to Eq 6, we obtained the fHL(τ), fLH(τ), PHL, PLH, ⟨τHL⟩−1, and ⟨τLH⟩−1 of the Michalis-Manton model in Eq 7 to Eq 12, respectively. (7)fHL(τ)=k1°k22A(e(B+A)τ−e(B−A)τ) Where A=(k−1+k2+k1°)2−4k2k1°2, B=−(k2+k1°+k−1)2 (8)fLH(τ)=k−3e−k−3τ (9)PHL([S])=k2k−3+k−1k−3+k1k−3[S]k2(k−3+k1[S])+k1k−3[S]+k−1k−3 (10)PLH([S])=k1k2[S]k2(k−3+k1[S])+k1k−3[S]+k−1k−3 (11)〈τHL〉−1=k1k2[S]k2+k−1+k1[S] (12)〈τLH〉−1=(∫0∞τfLH(τ)dτ)−1=k−3</p><p>Our simulation accurately recovers the characteristics predicted by the Michaelis-Menten equations. We first compared the analytical and simulated results using a model where E, ES*, and ES are equally stable (i.e., condition 1 in Figure 1c). The other conditions mimic the ES* and ES existing as transient and stable complexes are summarized in Supporting Information S2. Figure 3b–d shows the comparison of dwell-time distributions (Figure 3b), average transition rates (Figure 3c), and relative populations (Figure 3d) of each species. In all conditions, the simulations nicely overlap with the predictions, validating that SMIS successfully generate single-molecule trajectories for the target kinetic model.</p><p>To determine how many dwell times are needed to robustly recovery the rate constants, we compared how the extracted rate constants deviate from the input. Here, we used the transition from IL to IH state as the testing model. Using k−3 with the input value of 5 s−1, we simulated the distribution of τLH with the number of transition varies from 30 to 10,000. Figure 3e shows the percent error (% ERR) of extracted rate constants comparing to the input value of 5 s−1 as a function of the number of dwell times. The % ERR decreases and gets below 10% once the number of dwell time is larger than 300.</p><!><p>f(τ) plays a crucial role in quantifying kinetic rate constants in the typical single-molecule FRET approach. However, the derivation of f(τ) requires a system of interest to fulfill many criteria. General procedures to derive f(τ) involve (1) formulating the kinetic model from experimental results; (2) dissecting kinetic model to specific transitions between FRET states; and (3) solving the differential equations with proper initial conditions (see example in Supporting Information S1). However, solving f(τ) could be difficult or impossible when there are repeated differential equations, which results in a deficit of useful equations for the variables (i.e., number of useful equations is less than the number of variables). Difficulty can also be originated from undefined initial conditions, such as multiple species co-existing in the same FRET state with relative population undefined. Figure 4a shows one example whose analytical solutions are unobtainable. The kinetic model describes an enzyme existing as one of the four interacting species (E, ES*, ES**, and ES) where E and ES* associates with the FRET high (IH) state, and ES** and ES with the low (IL) state. The substrate S binds to the enzyme E to form the interacting complex ES through different configurational intermediates ES* and ES** with the forward (k1, k2, k3, and k4) and reversed (k−1, k−2, k−3, and k−4) rate constants annotated.</p><p>Even though the analytical solutions are not available, in principle, one may still extract the rate constants and species population through simulations with a range of rate constants. To test how effective SMIS can extract rate constants, we applied SMIS to a kinetic model without analytical solutions. We randomly selected a set of rate constants as specified in Figure 4b. With the number of transition set to 500,000 (equivalent to the number of dwell time of 217,000), we generated the PDFHL(τ), PDFLH(τ), PHL, and PLH under three substrate concentrations ([S] = 2, 10, and 50 μM) as shown in Figure 4c. This set of data serves as the experimental data, on which we applied SMIS to extract the rate constants.</p><p>To extract out the rate constants, we performed an extensive simulation using SMIS and search for most probable rate constants by minimizing the average of percent residue. We adapted SMIS to simulate the PDFHL(τ), PDFLH(τ), PHL, and PLH under [S] = 2, 10, and 50 μM with all eight rate constants vary from 0.2, 0.6, 2, 6, and 20 s−1 in the first search. Since there are eight rate constants for each simulation, this step creates 32768 (85) simulations. Deviation of each simulation from the experimental data was individually estimated through the averaged % RSD, %RSD¯. For each simulation, deviations of simulated PDFHL(τ), PDFLH(τ), PHL, and PLH from the experimental data were first calculated by the ratio of residue to the total area under the curve (Figure 4d). Averaging of all deviations gave the averaged % RSD, %RSD¯, which serves as the goodness of simulation for each input rate constant set. Since the simulation has a ~10% error, we consider the simulations with %RSD¯<10% are equally accurate. We thus sorted and picked the simulations with %RSD¯<10% (Figure 4e). With the selected simulations, we generated the histogram for each rate constant to identify the most probable rate constants (Figure 4f).</p><p>To search for the most probable rate constants, we applied the repeated bisection method (Figure 4g) to each rate-constant histogram. In the simulation step, we used SMIS to generate simulations at five different rate constants, used %RSD¯ to select the top 10% simulations, and generated the histogram of each rate constant (Figure 4f). Using the histogram, we bisected the most probable range for each rate constant when the sum of the possibility of simulated values (starting from high to low) is more than 50%. Take k2 as an example; we picked the second and the first bin (red bars) to define a new range for the most probable k2. This approach provides the boundaries of rate constants for the subsequent screening. In the subsequent screening, the rate-constant space between boundaries was further divided into five zones to repeat the searching process (Figure 4g). Figure 4h shows the searching results for each rate constant after seven searches. With this approach, we extracted most rate constants (k1, k2, k3, k−1, k−2, and k−3) which were originally unobtainable (Figure 4b). Unfortunately, SMIS still can not recover k4 and k−4. This is most likely due to the lack of useful data since only PDFLH(τ) contains the k4 and k−4 information. In contrast, from the [S] dependent PDFHL(τ), PDFLH(τ), PHL, and PLH, we identified the k1, k2, k3, k−1, k−2, and k−3 with only seven-round simulations (extracted values were summarized in Figure 4b).</p><!><p>The single-molecule interaction simulation (SMIS) opens new possibilities for objective characterization of interaction kinetics based on the kinetic model of interest, regardless of whether the analytic solutions are available or not. With the two-state model, we derived, as well as used SMIS to simulate, the probability density function (i.e., fHL(τ), fLH(τ)), average transition rates (⟨τHL⟩−1 and ⟨τLH⟩−1), and the relative populations of high and low states (i.e., PHL and PLH). These derived analytical solutions justified the feasibility of SMIS to recovery important kinetic distributions using the Michaelis-Menten enzyme kinetic model. To test how effective SMIS can extract rate constants, we applied SMIS to a kinetic model without analytical solutions. We extracted most rate constants that were originally unobtainable. These results indicate the SMIS is useful in providing important characteristics of kinetic parameters for an assigned kinetic model. Comparison between the experimentally determined distributions of kinetic parameters and the simulations crossing a wide range of rate constants can robustly quantify the rate constants. Our findings here contribute to the quantitative analysis of smFRET data, which is an essential step toward understanding biophysical problems using the smFRET approach.</p>

PubMed Author Manuscript

Synthesis of Thiomorpholine via a Telescoped Photochemical Thiolene/Cyclization Sequence in Continuous Flow

A procedure for the continuous flow generation of thiomorpholine in a two-step telescoped format was developed. The key step was the photochemical thiol-ene reaction of cysteamine hydrochloride and vinyl chloride as low-cost starting materials under highly concentrated (4 M) conditions, leading to the corresponding half-mustard intermediate in quantitative yield. Thiomorpholine was subsequently obtained by a base-mediated cyclization. The robustness of the process was demonstrated by performing the reaction for 7 h (40 min overall residence time) isolating the desired thiomorpholine via distillation.

synthesis_of_thiomorpholine_via_a_telescoped_photochemical_thiolene/cyclization_sequence_in_continuo

<!>Scheme 1. First Reaction Step Toward Sutezolid/Linezolid<!>Scheme 2. Syntheses Toward Thiomorpholine.<!>Scheme 3. Thiol-ene Reaction of Cysteamine Hydrochloride with Vinyl Chloride in Batch<!>EXPERIMENTAL SECTION<!>CAUTION:<!>ASSOCIATED CONTENT Supporting Information

<p>The thiomorpholine moiety is an important structural motif that is incorporated into a variety of active pharmaceutical ingredients (APIs) because of its interesting pharmacological profile, including antimalarial, antibiotic, antioxidant or hypolipidemic activity. 1 A prominent example is the oxazolidinone antibiotic sutezolid that is currently in phase 2 clinical trials for the treatment for multidrugresistant tuberculosis (MDR-TB). Due to its improved therapeutic potential, it is considered to be a promising replacement for the FDA-approved, first-generation drug linezolid, the morpholine analog of sutezolid (Scheme 1). 2 However, the Medicines for All Institute (M4ALL) conducted techno-economic analyses of the routes toward sutezolid (see Scheme S1) 3,4 that identified thiomorpholine as the most significant cost driver. In order to be cost competitive with linezolid, and therefore accessible to low-and middle-income countries, a scalable route to generate thiomorpholine from lowcost starting materials is highly desirable.</p><!><p>Approaches toward the synthesis of thiomorpholine (1) are displayed in Scheme 2a and include the transformation of diethanola-mine to 1 via generation of an amino-mustard species and its cyclization by treatment with sodium sulfide (routes 1 and 2). 5,6 Starting from ethyl mercaptoacetate and aziridine, 1 can be obtained by LiAlH4 reduction of the generated thiomorpholin-3-one (route 3). 7 Another strategy involves the reaction of 2-mercaptoethanol with aziridine and further conversion to 2-(2-chloroethylthio)ethylamine hydrochloride, which is then cyclized with Et3N to 1 (route 4). 8 These procedures are rather time consuming (2−54 h) and isolation of thiomorpholine is achieved in 44−81% overall yield after either a distillative work-up [6][7][8] or crystallization as HCl salt. 5 Although most of the reported routes use low-cost starting materials, they also generate nitrogen-or half-mustards, respectively, and thus producing these molecules on scale in a standard laboratory environment would be a safety challenge.</p><!><p>We were inspired to develop an alternative, time-and atom-efficient continuous flow route toward thiomorpholine based on the thiol-ene reaction of cysteamine and vinyl chloride (VC). Thiol-ene reactions fall into the category of click chemistry due to their high yield, solvent and oxygen tolerance, and absence of byproducts. [9][10][11] They proceed via a free-radical mechanism, and initiation is typically achieved either with UV irradiation or by thermolysis of a chemical additive. This strategy would lead to the same half-mustard 2-(2-chloroethylthio)ethylamine hydrochloride intermediate reported by Asinger et al. (route 4, Scheme 2a) 8 in one step, which is then further cyclized without isolation under basic conditions to 1 (Scheme 2b). Both reagents are considered as low-cost bulk materials, with cysteamine itself being a high-volume FDA approved drug, ensuring a stable supply. VC on the other hand, is one of the world's most important commodity chemicals with an annual production of ca. 13 million metric tons.</p><p>However, since VC is supplied as a compressed, liquefied gas, is highly toxic, flammable, and a Group 1 human carcinogen, 12 and the generated intermediate is a half-mustard, the process is best conducted in a telescoped continuous flow format. 13,14 Low reactor volumes ensure that only a small amount of material is present at any given time, thus hazardous materials can be safely handled in continuous flow. [15][16][17][18] In addition, head space issues are eliminated, gaseous reagents can be dosed precisely and processes can be seamlessly translated from lab to industrial scale. 19,20 Photochemical reactions typically can be improved when performed within the narrow channels of a microreactor combined with state-of-the-art LED irradiation technology. [21][22][23] Preliminary studies on the thermal and photochemical thiol-ene reaction were performed employing vinyl acetate as a more convenient and readily available alkene precursor on lab scale, compared with VC (see Supporting Information). These studies revealed that cysteamine as free base furnished 2-methyl-1,3-thiazolidine as major product (Figures S1, S2), but as HCl salt the desired intermediate 2-aminoethylthioethyl acetate was obtained (Figures S3−S5). Methanol proved to be the solvent of choice, as cysteamine showed poor solubility in other solvents (e.g., MeCN, THF, toluene, DCM). With this information in hand, we next turned our attention to VC, which is expected to be the superior alkene building block with respect to atom-economy, material availability and reactivity.</p><p>For batch optimizations, 1.1 equiv. VC (bp −13.4 °C) were condensed into a 1 M solution of cysteamine hydrochloride (2) in MeOH (Figure S11). This procedure allowed more accurate dosing of VC compared to simply sparging the solution with VC, which results in a higher amount of VC in the headspace and thus less conversion to product. A quick comparison of the thermal (80 °C, 30 min, 5 mol% AIBN) and photochemical (rt, 30-60 min, 365 nm) reaction provided similar results as with vinyl acetate: The photochemical route proved to be highly selective, and a quantitative yield of 2-(2-chloroethylthio)ethylamine hydrochloride (3) by NMR was obtained, versus 83% thermally (Scheme 3).</p><!><p>We moved forward to optimization studies in continuous flow using a commercial plate-based flow photoreactor (Corning AFR, Lab Photo Reactor) 24 and the set-up shown in Table 1. VC was fed via a calibrated mass flow controller (MFC), which allowed accurate dosing of the gas. Since VC is provided as a compressed, liquefied gas at 3 bar, the maximum outlet pressure was limited to 1 bar, which prevented performing the reaction at gas flow rates higher than ca 12 mln/min and thus higher throughput. In addition, the integration of a back pressure regulator (BPR) was not suitable, most likely resulting in less VC dissolved in solution and a lower residence time than calculated. Employing a 1 M solution of 2 in MeOH, a calculated maximum residence time 25 of 10 min and irradiation at 365 nm, provided NMR yields of only 53-58% (Table 1, entries 1 and 2). By lowering the temperature from 20 °C to 6 °C, the yield dropped to 18% (entry 3). The poorer performance in flow was not unexpected, since an extremely low absorption coefficient of 2 was determined ( = 0.025 Lmol −1 cm −1 at 363 nm), resulting in an absorption of only about 1% of the incident light at a concentration of 1 M due to the short path length of 0.04 cm (plate´s channel size). More details can be found in the Supporting Information. Therefore, it appeared advantageous to employ a photocatalyst, a common practice for this photochemical reaction. For thiol-ene reactions in continuous flow, 2,2-dimethoxy-2-phenylacetophenone (DMPA) as photoinitiator or Ru(bpy)3(PF6)2 as photocatalyst have been reported. [26][27][28] While Ru(bpz)3(PF6)2 has been reported to catalyze the thiol-ene reaction of similar thiol substrates such as cysteine methyl ester via a single electron transfer (SET) mechanism, 29 it was important to avoid expensive metal catalysts in our approach. Therefore, to accelerate this reaction, we envisioned employing 9-fluorenone (9-FL) as inexpensive photocatalyst. 30 Although the oxidation potential of the T 1 exited state of 9-FL (+0.96 V vs SCE) 31 would suffice to oxidize cysteamine (+0.92 V vs SCE) 32 it might also act via energy transfer catalysis. [33][34][35] However, it is difficult to distinguish the two pathways experimentally.</p><p>As expected, the addition of 5 mol% of 9-FL increased the yield to 87% (entry 4). A further improvement in yield could only be accomplished by increasing the concentration of the liquid substrate feed (entries 5 and 6, see also Figure S10). The maximum achievable concentration was 4 M, providing 3 in quantitative yield. It has to be noted that for such highly concentrated solutions of 2, dissolution is aided by sonication and that crystallization occurs upon cooling below room temperature. Although the thiol-ene reaction is rather insensitive toward O2, the photocatalyst might be quenched by O2. Hence, the liquid substrate feedcontaining 2, methyl benzoate as internal standard and 9-FLwas additionally degassed by sparging with argon for ca. 1 min. Interestingly, no change in reactivity was observed whether the liquid feed was degassed or not (entry 6). Further optimizations revealed that the sensitizer concentration can be reduced to 0.1 mol% (720 mg/L) without compromising the yield (entries 7 and 8). Remarkably, at such a high concentration the reaction proceeded to 98% yield (vs 58% at 1 M) even without sensitizer (entry 9). Under optimized conditions − 4 M solution of 2, 1 equiv. VC, 0.1−5 mol% 9-FL, 20 min residence time − intermediate 3 was generated with a throughput of 5.9 g/h.</p><p>Interestingly, gas formation was observed at the reactor outlet, although only an equimolar amount of VC was employed. Therefore, an isolation experiment was performed. As the isolated yield was in agreement with the 1 H NMR yield (see Supporting Information and Figures S12 and S13), no further investigations were conducted. It also has to be noted that over the course of the reaction the volume of the liquid stream increased by 17%, resulting in a reduced concentration of intermediate 3 of 3.42 M.</p><p>For the cyclization of intermediate 3 to thiomorpholine, a base screen was first performed in batch (Table S2). Full conversion of 3 was achieved and thiomorpholine was obtained with an NMR yield of 86−89% after 5 min at 100 °C by employing 2 equiv. of either Et3N, DIPEA or DBU. Et3N was successfully used for this reaction by Asinger et al., 8 however, since precipitation was observed, this base was unsuitable for a flow protocol. With respect to cost efficiency, we decided to explore the telescoped thiolene/cyclization reaction with DIPEA.</p><p>Since the cyclization proceeds faster at temperatures above the boiling point of MeOH, a back pressure regulator (BPR) set to 3 bar was installed. Due to the outlet pressure limitation of 1 bar of the VC cylinder, a hold vessel needed to be introduced to collect the exit stream of the thiol-ene reaction, which was then further pumped to be mixed with 2 equiv. of neat DIPEA (Scheme 4 and Figure S17B). This hold vessel also functioned as a gas separator. The thiol-ene reaction was performed under the conditions depicted in entry 8 in Table 1 using 0.1 mol% of 9-FL. We realized that a simple T-mixer did not provide efficient mixing resulting in lower thiomorpholine yields. When introducing a coil filled with glass beads (PFA, 1.6 mm ID, 3.2 mm OD, 0.5 mL void volume when filled) after the T-mixer, which functions as both a mixing and reaction unit, the same outcome as in batch was observed: an 87% NMR yield could be achieved at 100 °C and 5 min residence time (Figure S14). Finally, we performed a long run to demonstrate the robustness of this process. For this purpose, the concentration of 9-FL was increased to 0.5 mol% to ensure a stable performance over a multihour run. After experiencing clogging issues with the above-mentioned set-up at process times >1 h, a 7.5 mL coil (0.8 mm ID, 1.6 mm OD) that was immersed in an ultrasonic bath 36 was used as residence time unit at a temperature of 76−78 °C. With this set-up, the process was constant for 7 h after reaching steady state (Scheme 4 and Figure 1). Yields of intermediate 3 and thiomorpholine of ≥98% and 84%, respectively, by NMR were achieved (Figure 1), which matches well with previous optimizations. After distillation, 12.74 g (54% overall) of thiomorpholine was isolated, which corresponds to a throughput of 1.8 g/h. The difference between isolated and NMR yield is related to losses during distillation (see the Supporting Information), which has not been fully optimized at this small scale. However, we expect these losses to be minimal when an improved work-up procedure is employed on larger scale. In conclusion, we have developed a continuous flow process for the atom-and time-efficient generation of thiomorpholine by using readily available cysteamine (as its hydrochloride salt) and vinyl chloride as bulk materials. The telescoped photochemical thiol-ene/ cyclization sequence furnished thiomorpholine at laboratory scale in 54% overall isolated yield (84% NMR yield) after distillation, which was comparable to most of the reported procedures (44-81%). [5][6][7][8] Key to this telescoped sequence was the continuous photochemical thiol-ene reaction, which proceeded under highly concentrated conditions (4 M solution of 2), low amounts of 9-fluorenone as photocatalyst (≤0.5 mol%) and quantitative yield of 3. Due to the low pressure of the VC gas cylinder, the throughput on lab scale was limited to a maximum of 5.9 g/h of intermediate 3, and in turn of 1.8 g/h of thiomorpholine. To achieve higher production capacity, VC is best processed in liquid form, as is common in polyvinyl chloride (PVC) production. On lab scale, this technique is too high risk and impractical, however, with the appropriate equipment, this reaction has the potential to be safely scaled in a continuous flow format at production scale. Several photochemical reactions have been demonstrated in continuous flow on production scales, 23,[37][38][39] which supports this assessment.</p><!><p>General Methods. All materials were purchased from commercial sources (TCI, Sigma-Aldrich, AirLiquide) and used without further purification. 1 H NMR spectra were recorded on a Bruker 300 MHz instrument. 13 C NMR spectra were recorded on the same instrument at 75 MHz. Chemical shifts (δ) are expressed in ppm downfield from TMS as internal standard. The letters s, d, t, q, m and brs are used to indicate singlet, doublet, triplet, quadruplet, multiplet and broad singlet. GC-FID chromatography was performed using a Shimadzu GC FID 230 gas chromatograph with a flame ionization detector (FID). Helium, used as the carrier gas (40 cm sec -1 linear velocity), goes through a RTX-5MS column (30 m × 0.25 mm ID × 0.25 μm). The injector temperature is set to 280 °C. After 1 min at 50 °C, the column temperature is increased by 25 °C min -1 to 300 °C, then held for 4 min at 300 °C. The gases used in the detector for flame ionization are hydrogen and synthetic air (5.0 quality). GC-MS analysis was performed on a Shimadzu GCMS-QP2010 SE coupled with a DSQ II (EI, 70 eV). A fused silica capillary column Rtx-5MS column (5% diphenyl, 95% dimethylpolysiloxane, 30 m × 0.25 mm × 0.25 μm) was used. The injector temperature was set at 280 °C. After 1 min at 50 °C, the oven temperature was increased by 25 °C/min to 300 °C and kept at 300 °C for 3 min. As a carrier gas, helium at 40 cm s −1 linear velocity was used. MS conditions were: ionization voltage of 70 eV, acquisition mass range 50-450 m/z. Mass spectral libraries (Wiley Registry of Mass Spectral Data 11th Edition, NIST/EPA/NIH Mass Spectral Library 14) were searched with NIST MS Search software. LC-MS analysis was carried out on a Shimadzu instrument using a C18 reversed-phase (RP) analytical column (150 mm × 4.6 mm, particle size 5 μm) using mobile phases A (H2O/MeCN 90:10 (v/v) + 0.1% HCOOH) and B (MeCN + 0.1 % HCOOH) at a flow rate of 0.6 mL/min. The following gradient was applied: hold at 5% solvent B until 2 min, increase to 20% solvent B until 8 min, increase to 100% solvent B until 16 min and hold until 22 min at 100% solvent B. Low resolution mass spectra were obtained on a Shimadzu LCMS-QP2020 instrument using electrospray ionization (ESI) in positive or negative mode. Melting points were obtained on a Stuart melting point apparatus in open capillary tubes. Batch reactions above the boiling point of MeOH were performed in an Initiator+ single-mode microwave reactor from Biotage, using 2.5 mL Pyrex vials. The reaction temperature was controlled by an external infrared sensor. Reaction times refer to hold times at the temperature indicated. UV/vis spectra were recorded using a fiber-coupled Avantes Starline AvaSpec-2048 spectrometer and were processed using Avasoft 8.7 software. A commercial continuous flow photoreactor (Corning Advanced-Flow Lab Photo Reactor) was used.</p><!><p>Vinyl chloride is a highly toxic, flammable, and carcinogenic gas. Laboratory personnel working with vinyl chloride must familiarize themselves with the potential hazards and prevention measures. It is recommended to use a dedicated gas detector.</p><p>Continuous Flow Procedure for the Photochemical Thiol-ene Reaction of Cysteamine Hydrochloride with Vinyl Chloride (Table 1): The liquid feed solution was prepared by dissolving cysteamine hydrochloride, 9-fluorenone, and methyl benzoate as internal standard in a volumetric flask (25 mL) in MeOH. The solution was degassed by sparging with Ar using a balloon and needle. The thermostats were set to the desired temperature beforehand (respective temperature for the reaction, 15 °C LED-cooling). The liquid feed was directly pumped from the volumetric flask using a syringe pump (Syrris-Asia) at maximum flow rate (2.5 mL/min) until the reactor was filled with the substrate solution. Then, the flow rate was reduced to the desired value, the LEDs were turned on (365 nm, 100% intensity) and the MFC was set to deliver the desired amount of VC. After reaching steady state (about 20 minutes), a sample was collected. 100 µL of this sample were diluted with 500 µL of MeOH-d4 and analyzed by 1 H NMR (300 MHz).</p><p>Telescoped Continuous Flow Procedure for the Synthesis of Thiomorpholine: The liquid feed solution was prepared by dissolving cysteamine hydrochloride (45.44 g, 0.4 mol), 9-fluorenone (0.36 g, 2 mmol, 0.5 mol%) and diphenyl ether (7.264 g, 0.04 mol) as internal standard in a volumetric flask (100 mL) in MeOH. Dissolution was aided by sonication. The thermostats were set to the desired temperature beforehand (20 °C reaction, 15 °C LEDcooling). The liquid feed was directly pumped from the volumetric flask using a syringe pump (Syrris-Asia) at maximum flow rate (2.5 mL/min) until the reactor was filled with the substrate solution. Then, the flow rate was reduced to the desired value (0.139 mL/min), the LEDs were turned on (365 nm, 100% intensity) and the MFC was set to deliver the desired amount of VC (12.1 mLn/min, p = 0.8−0.9 bar). After reaching steady state (about 20 minutes) the output of the reactor was connected to the gas separator/hold vial. About 15 min later, the two pumps delivering the thiol-ene mixture and DIPEA (neat) were turned on and set to the corresponding flow rates (see Scheme 4). The sonication was turned on, the ultrasonic bath was set to the desired temperature (80 °C) beforehand. The temperature of this water bath equilibrated between 76-78 °C and was monitored by a K-type thermometer. 50 min later the cyclization reaction had reached steady state and the reactor output was collected for 7 h (7 fractions of 1 h each). During this time, the thiol-ene mixture in the hold vial was sampled every 1 h (100 µL diluted with 500 µL of MeOH-d4 and analyzed by 1 H NMR (300 MHz)). The output of the cyclization reaction was collected every 30 min (100 µL diluted with 500 µL of MeOH-d4 and analyzed by 1 H NMR (300MHz)). To the combined fractions 1 M HCl (140 mL) and EtOAc (300 mL) were added. After separation of the phases, the organic phase was washed with 1 M HCl (3×25 mL) until no more thiomorpholine could be detected in the organic phase by LC-MS. Next, ~4 M NaOH was added to the combined aq. phases until pH >13 and extracted 3 x with DCM. Additional NaOH was added, because the pH dropped to ~12. The aqueous phase was further extracted with DCM until no more thiomorpholine could be detected in the aqueous phase by LC-MS. The combined organic fractions were dried over Na2SO4, filtered and the solvent was removed by evaporation (100 mbar at 40 °C water bath). After vacuum distillation, 12.74 g (54% overall) thiomorpholine (1) was obtained as a colorless oil. Bp: 58-64 °C at 20 mbar. 1 H NMR (300 MHz, CDCl3) δ 3.09-3.05 (m, 4H), 2.57-2.53 (m, 4H), 1.52 (brs, 1H). 13 C NMR (75 MHz, CDCl3) δ 47.9, 28.3.</p><p>The data are in agreement with previously published values. 6,8</p><!><p>The Supporting Information is available free of charge on the ACS Publications website.</p><p>Additional experimental details, photographs of reactor set-up, further optimizations studies, 1 H NMR and 13 C NMR spectra (PDF)</p>

ChemRxiv

Enantio- and Regioselective CuH-Catalyzed Hydroamination of Alkenes

A highly enantio- and regioselective copper-catalyzed hydroamination reaction of alkenes has been developed using diethoxy(methyl)silane (DEMS) and esters of hydroxylamines. The process tolerates a wide variety of substituted styrenes, including trans-, cis-, and \xce\xb2,\xce\xb2-disubstituted styrenes to yield \xce\xb1\xe2\x80\x93branched amines. In addition, aliphatic alkenes coupled to generate exclusively the anti-Markovnikov hydroamination products.

enantio-_and_regioselective_cuh-catalyzed_hydroamination_of_alkenes

<p>Hydroamination, the direct formation of a C-N bond by the formal addition of an amine to an alkene, is a powerful synthetic procedure with the potential to gain access to amine products which are widely featured in pharmaceutically active compounds.1 Although great progress has been made in the field of late transition metal-catalyzed hydroamination,2 several challenges still exist. For example, the intermolecular process requires activated alkenes such as vinyl arenes.2a,i,h or acrylic acid derivatives,2c while asymmetric variants are limited to the addition of aryl amines to simple β-unsubstituted styrene derivatives and achieve only moderate levels of enantiomeric excess.2a,3 In addition, there are limited methods available to obtain the anti-Markovnikov product in hydroamination reactions of aliphatic amines.4 Thus, there remains a need for the development of asymmetric hydroamination reactions that tolerate a wide variety of substitution patterns on the alkene component and proceed with high regio- and enantioselectivity.</p><p>Over the last decade, our laboratory has reported several examples of asymmetric reactions involving copper-hydride (CuH) intermediates.5a.e We postulated that this CuH strategy could serve as a platform for the hydroamination of alkenes (Eq. 1). In our approach for asymmetric intermolecular hydroamination, we propose that insertion of an alkene (1, 4) into a chiral ligand-bound LCu(I)H species (I) would form an alkyl-copper complex (II) (Figure 1).6 Sub-sequent oxidative addition of an electrophilic amine source, such as a hydroxylamine 2,7 followed by reductive elimination, would form the C-N bond enantioselectively. The copper (I) species generated would then undergo transmetalation with an external hydride-transfer reagent to reform I.10 This mechanism (Figure 1) comes in a straightforward manner from a combination of our previous work in two are-as.5a,11 Herein, we report a mild copper-catalyzed hydroamination strategy using a chiral copper catalyst with a broad substrate scope. We note that toward the end of our work, a paper describing a method similar to the first portion (asymmetric) of this chemistry by Hirano and Miura was reported. 2a</p><p>We began our investigation by attempting the hydroamination of styrene (1a) using readily available Cu(OAc)2 and easily accessible O-benzoylhydroxylamine 2a (Table 1). Various ligands and hydride-transfer reagents were tested. We were able to achieve the desired cross-coupled products in up to 74% ee using polymethylhydrosiloxane (PMHS) or diethoxymethylsilane (DEMS) in conjunction with the commercially available ligand BINAP (L1) (entries 2-3). DEMS generated the desired product in the highest yield (entry 3), and thus was chosen as the hydride transfer reagent of choice in the examination of other chiral ligands (entries 4-8). We were able to realize up to 97% ee when using (R)-DTBM-SEGPHOS (L5) as the ligand (entry 7). Further optimization revealed that the reaction proceeds with low catalyst loading (2 mol%) at 40 °C (entry 8), without diminishing the yield or enantioselectivity. The reaction exclusively generated an α-branched amine, which is consistent with the proposed catalytic cycle (Figure 1) because the hydride migration from the copper catalyst to the alkene would generate the more stable α-bond Cu species.12</p><p>With an optimized protocol in hand, we then explored the substrate scope with respect to the styrene component (Table 2). This hydroamination tolerates a variety of substituents on the aryl ring of styrene (3b-g). The reaction also works efficiently with both trans- and cis-β-substituted styrenes (3h-o). Even hindered β,β-disubstituted styrenes undergo hydroamination in high yield and ee in this reaction (3p-q). Notably, the hydroamination of β,β-disubstituted styrene 1q gave the product 3q as a single diastereomer.</p><p>We next explored the use of other amine electrophiles in this reaction. We found that this reaction is applicable to several alkyl- and dialkyl-N-OBz amines (Table 3). N-(OBz)azepane and other heterocyclic-N-OBz amines also furnished the respective hydroamination products in high yields and enantioselectivities.</p><p>Since hydroamination of unactivated alkenes remains a challenge, we examined whether the developed protocol would be applicable with aliphatic alkenes.4b We found that terminal aliphatic alkenes could be effectively hydroaminated under the same conditions (Tables 4 and 5). In every case, the reaction exclusively produces the anti-Markovnikov products. Like the reaction with styrene, this protocol tolerated alkenes containing a primary alkyl bromide (5c), an epoxide (5g), and was compatible with alkenes containing a tosylamine (5d), an amide (5e), a pyridine (5f), a tert-butyldimethylsilyl ether (5i), and ones with geminal substituents (5h-i). Additionally, a number of amine electrophiles, including the sterically hindered tetramethylpiperidine N-OBz (5m), cross-coupled efficiently. Our hypothesis for the observed selectivity for the anti-Markovnikov products is that the hydride migration from the copper catalyst proceeds to form the less sterically crowded terminal copper intermediate (Scheme 1); here there is no electronic advantage as for styrenes to form the 2°-alkyl-Cu intermediate. Oxidative addition of the hydroxylamine and subsequent reductive elimination would generate the un-branched tertiary amines.</p><p>As a demonstration of the robustness and practicality of this method, it was carried out at 10 mmol scale (Scheme 2) using the β-substituted styrene ((E)-(3-methoxyprop-1-en-1-yl)benzene) as β-substituted styrenes are known to be difficult substrates in asymmetric hydroamination reactions.3 We were able to lower the catalyst loading to 1 mol% with no decrease in the yield or enantioselectivity.</p><p>In summary, we have reported a mild method for synthesizing chiral tertiary amines by employing an asymmetric copper-catalyzed hydroamination. Substitution occurs in a regioselective manner to generate a C-N bond at the α-position of styrene derivatives. This method has been shown to be compatible with various substituted styrene derivatives, and styrenes with β–substitution. Additionally, this method allows the development of copper-catalyzed anti-Markovnikov hydroaminations of terminal aliphatic alkenes. We are currently investigating the asymmetric version of internal aliphatic alkene hydroamination, which will be reported in due course.</p>

PubMed Author Manuscript

Building an emission library of donor–acceptor–donor type linker-based luminescent metal–organic frameworks

Luminescent metal-organic frameworks (LMOFs) have been extensively studied for their potential applications in lighting, sensing and biomedicine-related areas due to their high porosity, unlimited structure and composition tunability. However, methodical development in systematically tuning the emission properties of fluorescent organic linker-based LMOFs to facilitate the rational design and synthesis of target-specific materials has remained challenging. Herein we attempt to build an emission library by customized synthesis of LMOFs with targeted absorption and emission properties using donor-acceptor-donor type organic linkers. By tuning the acceptor groups (i.e. 2,1,3-benzothiadiazole and its derivatives), donor groups (including modification of original donors and use of donors with different metal-linker connections) and bridging units between acceptor and donor groups, an emission library is developed for LMOFs with their emissions covering the entire visible light range as well as the near-infrared region. This work may offer insight into well controlled design of organic linkers for the synthesis of LMOFs with specified functionality.

building_an_emission_library_of_donor–acceptor–donor_type_linker-based_luminescent_metal–organic_fra

Introduction<!>Tuning acceptor groups<!>Modifying original donor groups<!>Tuning donor groups with different types of M-L bonds<!>Tuning bridging units<!>Conclusion

<p>Organic linker-based luminescent metal-organic frameworks (LMOFs) 1,2 have shown great potential for applications in various areas, such as solid-state-lighting, [3][4][5][6][7] sensing, [8][9][10][11][12] and bioimaging 13,14 due to their highly tunable structures and compositions by varying organic linkers and metal nodes and via crystal engineering under various conditions. To tune the emission properties of LMOFs in a broad energy region (including the visible light and near-infrared range) and to study the structural effect on their emission behaviors, it is essential to build design principles based on the structureproperty relationship, which will facilitate the customized synthesis and target-specic applications of LMOFs. However, tuning the emission behaviors of organic linker based LMOFs in a fully controllable manner remains a challenging task.</p><p>In our previous work, we achieved full-color emissive LMOFs using 2,1,3-benzothiadiazole and its derivative-based dicarboxylic acids and tetratopic carboxylic acids as organic linkers. 15,16 In these donor-acceptor-donor (D-A-D) based linkers, the emission tunability relies on the changeable electron-withdrawing capacity of acceptor groups. While these acceptors have been used in different D-A-D linkers to prepare LMOFs for various applications, 10,[17][18][19][20][21][22][23] a systematic study to tune the emission properties of D-A-D linker based LMOFs and to develop a related design principle is still lacking.</p><p>For D-A-D type molecular linkers, three common strategies can be utilized to tune the emission behavior of LMOFs: (1) tuning the electron density of the acceptor groups; (2) tuning the electron density of the donor groups 15,16 and (3) tuning the bridging units between donor and acceptor groups 24,25 (Scheme 1). There are usually two approaches to changing the electron density of the donor groups: (2.1) modication of the original donor, such as addition of functional groups (i.e. NH 2 , OH, and CF 3 ) with strong electron donating or electron withdrawing capacity; (2.2) use of totally different donor groups, for which metals may link to different binding sites, e.g. to oxygen when using carboxylic acid-based donor groups or nitrogen when using pyridine or azole-based donor groups (Scheme 1). Note that changing the electron density of acceptor groups usually has little or no effect on the topology of the resultant MOFs. 15,16 However, changing the electron density of donor groups, which coordinate to the metal-containing clusters or ions to eventually form MOFs, oen leads to structure variations.</p><p>Based on these strategies, we carry out a systematic study in the present work to tune the emission properties of D-A-D type linker based LMOFs by tailoring the acceptor groups, donor groups and the bridging units, where 2,1,3-benzothiadiazole and its derivative are utilized as the acceptor groups. Among the 23 organic linkers used in this study, 13 are newly synthesized or used to prepare MOFs for the rst time, and 8 new MOF structures are obtained. For MOFs with the same structure as UiO-68, we name them UiO-68-L, in which L is the abbreviation of the organic linker; and for MOFs with new structures, we name them HIAM-N (HIAM ¼ Hoffmann Institute of Advanced Materials), where N is the number for each new MOF. This relatively large database allows us to build an emission library with a broad range of emission energies, including both visible and near infrared (NIR) regions.</p><!><p>As reported in our previous work, 15 1a). This strategy can be extended to other linker systems, such as tetratopic carboxylic acids, not only to realize tunable emission from blue to red, but also to increase the structural diversity of the resultant LMOFs with target properties. 16 In an attempt to expand the emission library, two more acceptor groups of 1H-benzo 1b), were prepared according to the reported method. 15 The PXRD patterns of these two LMOFs are almost identical to those of the simulated UiO-68-BTMB (Fig. 1c), conrming their isoreticular nature and high purity. UiO-68-BIMB and UiO-68-BOMB exhibit bright deep blue and green emission with the peak maximum at 417 nm and 530 nm, respectively. The corresponding photoluminescence quantum yield (PLQY) values are 69.5% and 5.9% under 365 nm excitation.</p><p>By tuning the acceptor groups, the emission energies of the resultant UiO-68-type LMOFs cover the entire visible light range Scheme 1 The schematic diagram illustrating the structure and multiple strategies to tune the emission properties of D-A-D type organic linkers using 2,1,3-benzothiadiazole and its derivative as acceptors.</p><p>Fig. 1 D-A-D type organic linkers with tunable acceptor groups. (a) The molecular structures of linkers with different acceptor groups. Among them, BAMB, MBTB, BTMB, BSMB, NTMB and NSMB were previously reported 15,21 (417, 445, 470, 520, 530, 545, 585 to 637 nm). These results further demonstrate that highly tunable emissions are achievable for D-A-D type organic linker based LMOFs by simply changing 2,1,3-benzothiadiazole and its derivative based acceptor groups.</p><!><p>As mentioned earlier, two strategies can be used to tune the donor groups: 2a), where the order of the electron donating capacity is NH 2 > OH > OCH 3 > CF 3 . As expected, four UiO-68 type MOFs, UiO-68-BTTB, UiO-68-BTMB, UiO-68-BTHB and UiO-68-BTAB, were obtained, conrmed by the PXRD analysis (Fig. 2b). The solid-state emission of these LMOFs also covers the whole visible light spectrum from blue to red with the peak maximum at 438 nm, 520 nm, 575 nm and 650 nm for UiO-68-BTTB, UiO-68-BTMB, UiO-68-BTHB and UiO-68-BTAB, respectively (Fig. 2c). The corresponding PLQYs are 40.8%, 30.6%, 1.0% and 0.1%. It should be noted that a remarkable blue-shi, compared with UiO-68-BTMB, was observed when adding a group (such as CF 3 ) with strong electron withdrawing capacity and a signicant bathochromic shi was realized when using strong electron donating groups (i.e. OH and NH 2 ). 26 These results indicate that single-site modication of original donor groups is indeed a useful strategy to tune the emission properties of the LMOFs without changing the crystal structure.</p><!><p>Compared with carboxylic acid-based LMOFs, the utilization of donor groups with various metal-linker bonds will not only increase the structural diversity, but also introduce new or unprecedented properties to the resultant LMOFs. Herein, we prepared two series of organic linkers using pyridine and pyrazolate as the donor groups and investigated their effects on the emission behavior of the resultant LMOFs.</p><p>When pyridine was chosen as the donor group, three linkers 3a). The emission wavelengths of PBT, PNT and PNS are 452 nm, 567 nm and 611 nm, respectively. Compared to their carboxylate counter parts BTMB (500 nm), NTMB (566 nm) and NSMB (610 nm) with the same acceptor groups, a signicant blue- shi was observed for PBT (Fig. S2 †), for which the acceptor group has a relative weak electron withdrawing capacity, like benzo[c][1,2,5]thiadiazole. However, for acceptor groups with strong electron withdrawing capacity, almost no energy shi was observed between NTMB and PNT, and NSMB and PNS.</p><p>A typical synthesis for PBT-, PNT-and PNS-based Zn-LMOFs, HIAM-300X (HIAM ¼ Hoffmann Institute of Advanced Materials; 30 ¼ zinc; X ¼ 1 for PBT, X ¼ 2 for PNT and X ¼ 3 for PNS) is as follows: a 5 mL vial containing 0.1 mmol Zn(NO 3 ) 2 $6H 2 O, 0.1 mmol 1,4-dicarboxybenzene, 0.1 mmol designed linker, 1 mL DMF and 3 mL water was placed in a preheated oven at 120 C for 3 days. Aer cooling down to room temperature, the corresponding single crystals were obtained.</p><p>The pillar-layered structure of HIAM-3001 has been reported to have a pcu topology and belongs to an orthorhombic crystal system with a Pbca space group. 19 As shown in Fig. 3b, two equivalent Zn(II) are bridged by four carboxylate groups from four BDC ligands to form a binuclear "paddle-wheel" Zn 2 (COO) 4 . These 6-c second building units (SBUs) are connected by the BDC ligands to give (4, 4) layers which are further extended by PBT ligands to form the 3D network with a pcu The PXRD patterns of the as-synthesized LMOFs exhibit excellent agreement with the simulated ones, indicating the high purity of the obtained bulk samples (Fig. 3d). The peak maxima of solid-state emission of HIAM-3001, HIAM-3002 and HIAM-3003 are 499 nm, 592 nm and 678 nm, respectively (Fig. 3e), with PLQYs of 2.8%, 0.6% and 1.0% under 365 nm excitation. The lower PLQY could be attributed to the fact that the p-p stacking is much stronger in pillar-layered structures, which will cause severe non-radiative decay. The gradual red-shi was also observed in the UV-vis absorption spectra from HIAM-3001 to HIAM-3003 (Fig. 3f). Compared with UiO-68 type MOFs, HIAM-3001, 3002 and 3003 show high stability in aqueous solutions aer treatment at pH ¼ 2 to 12 for one day, conrmed by the nearly identical PXRD patterns (Fig. S4 and S5 †). HIAM-300X (X ¼ 1-3) also exhibits high resistance to heat and stability up to 350 C for HIAM-3001 and 400 C for HIAM-3002 and HIAM-3003, respectively (Fig. S6 †). The above results conrmed our hypothesis that changing the donor group to form a different metal-linker bond model will not only give rise to tunable emission behavior, but also contribute to structural diversity of the resultant LMOFs.</p><p>In recent years, pyrazolate-based MOFs have received considerable attention due to the pyrazolate-metal bond induced high stability and unique properties. 4a). Compared with MBTB (460 nm), BTMB (500 nm) and BSMB (530 nm), the emission peaks of DDPBT, DPBT and DPBS are 557 nm, 585 nm and 632 nm in DMF solution under 365 nm excitation (Fig. S7 †). These results demonstrate that pyrazolate is a stronger electron donating group compared with carboxylic acid-based linkers adapted to the same acceptors, which can induce a strong bathochromic shi.</p><p>The synthesis conditions for pyrazolate-based HIAM-300X (X ¼ 4 for DPBT, X ¼ 5 for DPBS and X ¼ 6 for DDPBT) are similar to those used to synthesize HIAM-3001 but without addition of 1,4-dicarboxybenzene. Single-crystal X-ray diffraction analysis reveals that HIAM-3004 crystallizes in the tetragonal crystal system with an I4 1 space group (Fig. 4b and S8 †). The Zn(II) cation is fully coordinated in a tetrahedral geometry with four nitrogen atoms from four DPBT ligands. Each DPBT ligand is fully coordinated with four Zn(II) cations, in which each pyrazolate group of the ligand connects two adjacent Zn(II) cations. The alternative connection of Zn atoms and pyrazolate groups results in an innite 4 1 helical chain along the c axis (Fig. S9 †). These screw chains are further extended by the DPBT ligand to give the 3D framework with 3D channels. Strong p-p interactions are found between the benzo[c][1,2,5]thiadiazole rings in HIAM-3004 (centroid-to-centroid distance is 3.5822(3) Å). An identical crystal structure was formed when DPBS was employed as the luminescent linker. It should be noted that a similar MOF to HIAM-3004 was reported for photocatalytic aerobic oxidation when we prepared our manuscript, which further demonstrates the promising applications of these MOFs. 34 A totally different crystal structure was obtained when the acceptor group was changed from benzo[c] [1,2,5]thiadiazole to 5,6-dimethylbenzo[c][1,2,5]thiadiazole for pyrazolate-based linkers. For HIAM-3006, a similar connection model to HIAM-3004 was observed where the Zn(II) cation is fully coordinated in a tetrahedral geometry with four nitrogen atoms from four DDPBT ligands. Each DDPBT ligand is fully coordinated with four Zn(II) cations, in which each pyrazolate group of the ligand connects two adjacent Zn(II) cations to give a 1D Zn chain along the a-axis. The Zn chains are further extended by the DDPBT ligand to yield a 3D framework with 1D channels along the aaxis (Fig. 4c and S10 †).</p><p>The PXRD patterns of the as-synthesized HIAM-3004 and HIAM-3005 show excellent agreement with the simulated HIAM-3004 pattern (Fig. 4d), indicating the phase purity and isoreticular nature of HIAM-3004 and HIAM-3005. HIAM-3006 also exhibits an essentially identical PXRD pattern to the simulated one. As shown in the emission spectra in Fig. 4e, compared with the emission peak maxima of UiO-68-MBTB (470 nm), UiO-68-BTMB (520 nm) and UiO-68-BSMB (545 nm), strong bathochromic shis were recorded for HIAM-3006, HIAM-3004 and HIAM-3005 at 534 nm, 563 nm and 616 nm with the corresponding PLQYs of 8.5%, 0.6% and 28.9% under 365 nm excitation (Fig. 4e). A signicant red-shi was also observed for the light absorption from HIAM-3006 to HIAM-3004 (Fig. 4f). HIAM-300X (X ¼ 4-6) also exhibits high stability in aqueous solutions from pH 2 to 12 and thermal stability up to 500 C (Fig. S11 and S12 †).</p><p>By comparing the emission properties of UiO-68-BTMB (520 nm), HIAM-3001 (499 nm) and HIAM-3004 (563 nm), it is clear that introducing a pyridine group as the donor will induce a blue-shi, while a red-shi is observed when using pyrazolate as the donor. More importantly, with the tunable donor groups, abundant structural diversity can be realized, which might lead to new properties and applications.</p><!><p>It has been reported previously that different bridging groups show signicant effects on the light absorption and emission behavior of D-A-D type compounds. For example, an emission shi from 610 nm, 622 nm to 647 nm was observed upon adding an ethynyl or a vinyl group between the donor and acceptor. 24,25 Incorporating thiophene groups into uorophores could also increase the light absorption and emission. [35][36][37] In our previous work, we found that when the acceptor groups change from benzo[c] [1,2,5]thiadiazole to naphtho[2,3-c][1,2,5] thiadiazole, the emission shis from 520 nm to 585 nm. 15 According to these results, we believe that organic linkers with a signicant red-shi might be synthesized if these two features are combined in one structure. On the other hand, benzo[c] [1,2,5]thiadiazole and naphtho[2,3-c][1,2,5]thiadiazole based carboxylic compounds have been utilized to prepare LMOFs with tunable emissions, and thus the biggest challenge is how to synthesize carboxylic compounds with a bathochromic shi in their emissions, which may provide the opportunity to obtain organic-linker-based NIR LMOFs.</p><p>To prove our hypothesis, six organic linkers were utilized in this section (Fig. 5a). Two of them have been reported, 4,4 0 -(benzo[c][1,2,5]thiadiazole-4,7-diyl)dibenzoic acid (BTBA) 10 and 4,4 0 -(benzo[c][1,2,5]thiadiazole-4,7-diylbis(ethyne-2,1-diyl)) dibenzoic acid (BTTBA). 22 And the other four are newly synthesized linkers, namely 4, [1,2,5]thiadiazole-4,7-diyl)bis(thiophene-5,2-diyl))dibenzoic acid (BTTD). BTBA, BTTBA and BTEBA were chosen to conrm the effect of bridging groups of ethynyl and vinyl. NTEBA was designed to achieve NIR emission. BTTB and BTTD were prepared to investigate the effect of thiophene and the substitution sites on the emission behaviors. The molecular orbitals of BTBA, BTTBA, BTEBA and NTEBA were calculated using density functional theory (DFT). As depicted in Fig. 5b, a signicant increase in the highest occupied molecular orbital (HOMO) energies was observed along with a decrease in the lowest unoccupied molecular orbital (LUMO) energies when an ethynyl or a vinyl group was added. As a result, the HOMO-LUMO energy gap decreased from 3.399 eV to 2.792 eV and 2.602 eV for BTBA, BTTBA and BTEBA, respectively, indicating that the bridging groups indeed can be used to effectively tune the electronic structures. More importantly, a further decrease was obtained from 2.602 eV to 1.987 eV, when the acceptor group was changed from benzo[c] [1,2,5]thiadiazole to naphtho [2,3-c] [1,2,5]thiadiazole.</p><p>The UV-vis absorption and steady-state emission spectra of the four organic linkers in DMF solution were then measured and plotted in Fig. 5c. As expected, the emission energy varies from blue (BTBA) to NIR (NTEBA), which is consistent with the red-shis of the corresponding absorption spectra. The maximum emission peaks appear at 480, 500, 565 and 690 nm for BTBA, BTTBA, BTEBA and NTEBA, respectively. The corresponding colors of the samples under daylight and 365 nm excitation are shown in Fig. 5d. A closer look at the molecular structures reveals an interesting structure-emission correlation. (i) Compared with BTBA, when a vinyl group is added between the donor and acceptor groups, an 85 nm bathochromic shi was observed for BTEBA. The remarkable effect of vinyl groups on the emission behavior could also be proven by comparison of the emission wavelength of NTEBA and the compound without vinyl groups, 4,4 0 -(naphtho[2,3-c][1,2,5]thiadiazole-4,9diyl)dibenzoic acid (NTB), in which a bathochromic-shi of 100 nm was recorded aer the addition of vinyl groups to NTB to form NTEBA. (ii) When acceptors were changed from benzo[c] [1,2,5]thiadiazole to naphtho[2,3-c][1,2,5]thiadiazole, a 125 nm red-shi was realized. These results demonstrate that the combination of these two strategies together indeed offers a powerful approach to tune the emission properties and to achieve dicarboxylic acid-based NIR emissive organic linkers.</p><p>The single crystals of BTBA and BTTBA based Zr-MOFs (1 and Zr-L7) were synthesized according to literature procedures (Fig. S13 †). 10,22 A typical synthesis of BTEBA and NTEBA based Zr-MOFs, HIAM-400X (HIAM ¼ Hoffmann Institute of Advanced Materials; 40 ¼ zirconium; X ¼ 5 for BETBA and X ¼ 6 for NTEBA) is as follows: a 5 mL vial containing 54.0 mg L-proline and 63.0 mL was placed in a preheated oven at 100 C for 30 minutes; aer cooling down to room temperature, 3 mL DMF, ZrCl 4 (22.0 mg, 0.094 mmol), and an organic linker (0.094 mmol) were added to the vial, which was placed in a preheated oven at 120 C for 48 hours. Orange (HIAM-4005) and dark red (HIAM-4006) octagon-shaped single crystals were obtained (Fig. 5e). Single-crystal X-ray diffraction analysis indicates that HIAM-4005 and HIAM-4006 adopt the typical cubic structure and crystallize in the space group Fd 3m. The structure consists of two sets of independent and mutually interpenetrating UiO-type frameworks (Fig. 5f and g). Therefore, the structure and connectivity of the SBUs in HIAM-4006 are the same as in UiO-type MOFs. The two new LMOFs show nearly identical powder X-ray diffraction (PXRD) patterns to the simulated one (Fig. 5h), indicating that these LMOFs made of linkers with different acceptor groups belong to the same isoreticular series.</p><p>The solid-state photoluminescence and UV-vis absorption spectra were measured. As shown in Fig. 5i, the emission maxima at 522 nm, 560 nm, 607 nm and 747 nm were recorded for 1, Zr-L7, HIAM-4005 and HIAM-4006 with the corresponding PLQYs of 7.8%, 5.6%, 8.8% and 0.5% under 365 nm excitation, respectively. The NIR emissive LMOF (HIAM-4006) was obtained by the combination of an acceptor with strong electron withdrawing capacity and an optimized bridging unit. A gradual red-shi was also observed in the absorption spectra from 1 to HIAM-4006, which is consistent with their corresponding emission energies (Fig. 5j). The lowest energy absorption edge is close to 750 nm, which may be suitable for applications in photocatalytic related areas, such as photocatalytic hydrogen generation and carbon dioxide reduction.</p><p>In addition to ethynyl and vinyl bridging units, thiophene was also utilized to link the donor and acceptor groups. Compared with BTBA (480 nm), BTTBA (500 nm) and BTETA (565 nm), the maximum emission peaks of BTTB and BTTD are 553 nm and 638 nm in DMF solution under 365 nm excitation, respectively (Fig. S14 †). This result indicates that (i) thiophene is a useful bridging unit to achieve the bathochromic-shi; and (ii) the substitution position has a signicant effect on the emission properties of the resultant compounds. However, it is very difficult to obtain the crystal structure of the corresponding LMOFs, which might be attributed to the distorted molecular structures.</p><p>Based on the aforementioned results, it is clear that the emission properties of D-A-D type organic linker-based LMOFs can be systematically tuned by changing the acceptor groups, modication of the original donor groups, using different donor groups and choosing various bridging units. Therefore, an emission library can thus be built by applying different strategies described in this work (Fig. 6).</p><!><p>In conclusion, four strategies have been used to methodically tune the emission behaviors of D-A-D type organic linkerbased LMOFs, from which a large emission library can be built in a systematic manner. By precisely controlling the acceptor groups, donor groups and the bridging units, emissions of the resultant LMOFs covering the entire visible light spectrum as well as the NIR region can be achieved with abundant structural diversity. A large number of organic linkers can be designed and synthesized with varying emission behaviors from deep blue to the NIR range by gradually decreasing the electron density of acceptor groups or increasing the electron density of donor groups. This work may not only serve as a toolbox to facilitate the development of design principles for the rational design of organic linkers and customized synthesis of target-specic LMOFs but also provide a useful platform to explore NIR emissive LMOFs for biosensing and bioimaging applications.</p>

Royal Society of Chemistry (RSC)

MEPicides: \xce\xb1,\xce\xb2-Unsaturated Fosmidomycin Analogues as DXR Inhibitors against Malaria

Severe malaria due to Plasmodium falciparum remains a significant global health threat. DXR, the second enzyme in the MEP pathway, plays an important role to synthesize building blocks for isoprenoids. This enzyme is a promising drug target for malaria due to its essentiality as well as its absence in humans. In this study, we designed and synthesized a series of \xce\xb1,\xce\xb2-unsaturated analogues of fosmidomycin, a natural product that inhibits DXR in P. falciparum. All compounds were evaluated as inhibitors of P. falciparum. The most promising compound, 18a, displays on-target, potent inhibition against the growth of P. falciparum (IC50 = 13 nM) without significant inhibition of HepG2 cells (IC50 > 50 \xce\xbcM). 18a was also tested in a luciferase-based Plasmodium berghei mouse model of malaria and showed exceptional in vivo efficacy. Together, the data support MEPicide 18a as a novel, potent, and promising drug candidate for the treatment of malaria.

mepicides:_\xce\xb1,\xce\xb2-unsaturated_fosmidomycin_analogues_as_dxr_inhibitors_against_malaria

INTRODUCTION<!>Synthesis.<!>Evaluation of 12a\xe2\x80\x93c, 16d\xe2\x80\x93g as DXR Inhibitors.<!>In Vitro Effects on Pathogen Growth by 12a\xe2\x80\x93c, 16d\xe2\x80\x93g, 18a\xe2\x80\x93c, 19e\xe2\x80\x93g.<!>The cLogP, Cytotoxicity, and Selective Indices of 12a and 18a.<!>Compound 18a Is Rapidly Converted to 12a in Vitro and in Vivo.<!>Compounds 12a and 18a Inhibit Isoprenoid Synthesis in P. falciparum.<!>In Vivo Evaluation of 18a in a Mouse Model of Malaria Infection.<!>CONCLUSIONS<!>General.<!>General Procedure for Synthesis of Amides 10a\xe2\x80\x93c and 13.<!>General Procedure for Synthesis of 11a\xe2\x80\x93c, 14, and 18a\xe2\x80\x93c.24<!>General Procedure for Synthesis of 12a,b and 16d\xe2\x80\x93g.<!>General Procedure for Synthesis of 17a\xe2\x80\x93c.<!>Diethyl [(1E)-3-[N-(Benzyloxy)formamido]prop-1-en-1-yl]-phosphonate (10a).<!>Diethyl [(1E)-3-(N-Hydroxyformamido)prop-1-en-1-yl]-phosphonate (11a).<!>Sodium Hydrogen [(1E)-3-(N-Hydroxyformamido)prop-1-en-1-yl]phosphonate (12a).<!>({[(1E)-3-[N-(Benzyloxy)formamido]prop-1-en-1-yl]({[(2,2-dimethylpropanoyl)oxy]methoxy})phosphoryl}oxy)methyl 2,2-Dimethylpropanoate (17a).<!>Optimized Synthesis of ({[(1E)-3-[N-(Benzyloxy)formamido]prop-1-en-1-yl]({[(2,2-dimethylpropanoyl)oxy]methoxy})phosphoryl}oxy)methyl 2,2-Dimethylpropanoate (17a).<!>[({[(2,2-Dimethylpropanoyl)oxy]methoxy}[(1E)-3-(N-hydroxyformamido)prop-1-en-1-yl]phosphoryl)oxy]methyl 2,2-Dimethylpropanoate (18a).<!>({[(1E)-3-[(Benzyloxy)amino]prop-1-en-1-yl]({[(2,2-dimethylpropanoyl)oxy]methoxy})phosphoryl}oxy)methyl 2,2-Dimethylpropanoate (24).<!>P. falciparum DXR Enzyme Inhibition Assay.<!>P. falciparum Growth Inhibition Assays.46<!>In Vivo Studies in a Mouse Model of Malaria Infection.

<p>Malaria is a severe, life-threatening infectious disease with high mortality and morbidity rates.1 In 2016, there was a substantially high incidence rate of malaria with 216 million new malaria cases as well as 0.4 million deaths, 64% of which are children under 5 years of age.1 Malaria is caused by a group of Plasmodium parasites, with Plasmodium falciparum causing the majority of deaths and severe infections.2 Parasites are transmitted to humans via the bites of female Anopheles mosquitoes.3 After growing and replicating initially in human liver cells, the parasites reach the blood and cause malarial symptoms such as fever, headache, chills, or even death.4 Artemisinin-based combination therapy (ACT) is currently the best treatment for malaria and is typically highly effective.3 Resistance to artemisinin, however, has already spread in the Greater Mekong subregion.5 Thus, there is a pressing need for new therapeutics for malaria with novel modes of action that could provide alternate chemotherapies to combat sensitive and drug-resistant parasites.</p><p>P. falciparum uses the methyl erythritol phosphate (MEP) pathway for the biosynthesis of isopentenyl pyrophosphate (IPP) and its isomer dimethylallyl pyrophosphate (DMAPP), the C5 precursors of isoprenoids (Figure 1).6,7 Humans, however, use an alternate acetate/mevalonate pathway exclusively to synthesize these C5 isoprene building blocks.8 Blocking the MEP pathway terminates the biosynthesis of such important metabolites and results in cell death of P. falciparum.3,9 Thus, many enzymes in the MEP pathway could become promising drug targets to design MEPicides that work against malaria through a new mechanism of action and one not found in humans.10</p><p>DXR, 1-deoxy-D-xylulose 5-phosphate reductoisomerase, is the second enzyme in the MEP pathway and catalyzes the first committed step.11 DXR converts DOXP to MEP with the assistance of cofactor NADPH as well as a divalent metal cation (Figure 1).12 DXR is essential for the viability of several pathogens and is validated as a promising drug target for developing antitubercular agents13 and antimalarials.14 Fosmidomycin (Figure 2, 1a), isolated from Streptomyces lavendulae,15 is a potent inhibitor of P. falciparum DXR (IC50 = 0.034 μM).16 FR-900098 (Figure 2, 1b), the N-acetyl analogue of fosmidomycin isolated from Streptomyces rubellomurinus,15 is roughly equipotent to fosmidomycin (P. falciparum DXR IC50 = 0.024 μM).16 While these two natural products have submicromolar inhibition of P. falciparum growth (IC50 = 0.09–0.35 μM),7 their use as a single drug therapy is limited by low bioavailability, short serum half-life, and malaria recrudescence.17 Despite these disadvantages, fosmidomycin is a remarkably safe drug as proven by several clinical trials and a promising candidate for treating uncomplicated P. falciparum malaria in combination therapies.18,19 Thus, we selected fosmidomycin as the parent structure from which to design analogues that would effectively inhibit Pf DXR, have improved pharmacokinetic properties and lead to promising drug candidates against malaria.</p><p>We and others have previously evaluated the structure—activity relationships (SAR) of fosmidomycin and FR-900098 as inhibitors of several DXR homologues as well as various microbial pathogens.20–29 Fosmidomycin binds to DXR competitively with substrate DOXP and uncompetitively with cofactor NADPH.30 SAR studies on fosmidomycin analogues reveal that the retrohydroxamate or hydroxamate moiety should be retained to mimic the crucial interaction of fosmidomycin with the divalent metal cation.21,24,25,27–29,31 Similarly, the phosphonate moiety should be retained as it forms numerous hydrogen bonds with neighboring amino acid residues.32–34 A three-carbon linker between the two moieties is also found to be crucial for DXR inhibition.24 As we reported earlier, the unsaturated FR-900098 analogue (Figure 2, 2) gained a 2-fold increase in potency against Mycobacterium tuberculosis (Mtb) DXR (IC50 = 1.07 μM) compared with parent compound FR900098.24 A prodrug strategy was applied to this structure, and the corresponding pivaloyloxymethyl (POM) phosphonate was synthesized (Figure 2, 3). Compound 3 displays an Mtb MIC99 value of 9.4 μg/mL (22 μM), thus gaining the needed lipophilicity to penetrate the Mtb cell wall.24 This compound likely regenerates 2 inside the bacteria, and this acid inhibits DXR.24,27 Prodrug 3 also shows potent inhibition against P. falciparum growth with an IC50 value of 18.3 nM,35 nearly as potent as artemisinin (P. falciparum IC50 = 10.4 nM),35 a current first-line antimalarial drug. As expected, prodrug 3 displays potent in vivo antimalarial activity.35</p><p>Since it was found that the NADPH-binding pocket of DXR is druggable,36 and because this pocket is adjacent to the cavity where the retrohydroxamate moiety of fosmidomycin binds,37 we previously synthesized analogues with extended aromatic groups on the N-alkoxy group of FR-900098 to develop improved DXR inhibitors as antimicrobials (Figure 2, 4, 5).27,28 These compounds are designed to act as bisubstrate inhibitors that could bind to both the DOXP and NADPH binding sites.27,28 The binding mode of compound 4 was then determined using classical Lineweaver—Burke double reciprocal plots. These experiments showed that compound 4 is competitive with DOXP and NADPH, confirming bisubstrate binding behavior.27 The POM prodrug of 4 was also synthesized (Figure 2, 5), showing effective Mtb growth inhibition (MIC99 = 18.75 μg/mL or 33.2 μM).27 The bisubstrate strategy increases the overall lipophilicity of the analogues, which is likely beneficial for penetration into several pathogens.</p><p>The purpose of the current work is to synthesize a series of α,β-unsaturated N-acyl (Figure 3A) and N-alkoxy (Figure 3B) fosmidomycin analogues and evaluate them as antimalarial agents that work via DXR inhibition. For the N-acyl analogues, we studied the influence of electronic effects on DXR inhibition via introduction of CF3 and OCH3 groups compared with the CH3 group of compound 2. The α,β-unsaturated N-formyl analogue was also made to assess the effect of size and lipophilicity compared with the methyl group in 2. To achieve the bisubstrate inhibitors that compete with both DOXP and NADPH, several aromatic groups (with a linker of 1 carbon atom) were selected as N-alkoxy analogues based on previous results.27,28 POM prodrugs of the initial hits that showed effective DXR inhibition, and which were synthetically accessible, were synthesized. These prodrug compounds were then evaluated as inhibitors of P. falciparum.</p><!><p>N-Formyl analogue 12a and N-acyl analogues 12b and 12c were prepared in seven steps from commercially available starting materials shown in Scheme 1. First, allyl phosphonate 6 was synthesized from triethylphosphite and allyl bromide via a Michaelis—Arbuzov reaction.38 Subsequent addition of bromine to compound 6 resulted in production of dibromide 7.39 O-Benzylhydroxylamine hydrochloride was neutralized in situ and protected using di-tert-butyldicarbonate to yield Boc-protected 8,40 which was then reacted with compound 7 and two equivalents of NaH to prepare compound 9.24 The first equivalent of NaH deprotonated 8 and generated a nucleophile to attack the primary bromide of compound 7. The second equivalent of NaH was used to eliminate the β-bromide to furnish α,β-unsaturated phosphonate 9. Under acidic conditions, compound 9 was hydrolyzed and generated the deprotected amine in situ that acted as a building block to be acylated with an acyl chloride or anhydride to synthesize N-acyl intermediates 10b,c. For compound 10a, the electrophilic reagent N-formylimidazole was prepared using formic acid and 1,1′-carbonyldiimidazole.41 Removal of the benzyl group using BCl3 yielded retrohydroxamic acids 11a–c.27 Treatment of these acids with TMSBr, followed by either NaOH or NH3, gave monosodium salts 12a,b or diammonium salt 12c, respectively.21,27 In the case of 11c, synthesis of the sodium salt resulted in an unstable compound, leading us to make the ammonium salt.</p><p>Scheme 2 shows the four-step synthesis used to prepare the N-alkoxy analogues, starting from intermediate 9. Acetyl chloride was used to synthesize compound 13 following the method described above. Subsequent debenzylation of 13 by BCl3 generated 14, the diethyl ester of FR-900098.24 This compound acted as a core intermediate from which to synthesize the series of the N-alkoxy analogues. The synthesis of 15d–g was initially attempted using Williamson ether synthesis with either NaH or sodium tert-butoxide in polar solvents such as THF and DMF.27 Unfortunately, these reactions failed because of the instability of 14 in the presence of such harsh bases. Weaker bases, such as Na2CO3 or Et3N, were then applied to the reaction, and yet an overwhelming amount of side products were generated concomitantly, possibly due to the high polarity of the solvents. The reaction conditions were optimized using Na2CO3 with CH2Cl2 as a relatively nonpolar solvent. The reaction was carried out in a sealed tube and heated at 60 °C for 48 h to yield 15d–g, which were then converted to monosodium salts 16d–g in a manner similar to that used to prepare 12a–c.</p><p>The prodrugs of selected analogues were made (Scheme 3). To obtain the diPOM esters for N-acyl analogues (17a–c), compounds 10a–c were treated with TMSBr and then transformed to the esters using POM chloride and Et3N.27 Because of the low yield of 17b, only 17a and 17c were carried to the next reaction. Removal of the benzyl group using BCl3 gave the diPOM prodrugs 18a and 18c. Similarly, the diPOM esters of N-alkoxy analogues 19e–g were prepared from 16e–g via treatment with POM chloride and Et3N.</p><p>Because of the low yields of N-acyl analogue 18a (overall yield 2.1%) and 17b (overall yield 0.42%), an optimized synthesis for N-acyl prodrugs was developed (Scheme 4). This route to synthesize diPOM esters was brought forward because the stability of compound 6 enabled a longer reaction time. By this method, compound 20 was obtained in higher yield (52%) compared with previous esterification reactions, and the synthetic attrition was greatly reduced. Bromine addition of the resulting product 20 yielded dibromide 21. Because of the instability of the POM group to acidic conditions, use of Boc-protected O-benzylhydroxylamine was inadvisible due to the acidic conditions that would be required to cleave the Boc group in the future steps. Thus, free amine 23 was prepared to react with monobromide 22, the eliminated product from 21 and NaH, to yield 24. This conversion enabled a variety of diPOM N-acyl analogues. Compound 24 was converted to acylated compounds 17a and 17b, with the latter compound obtained in a 17-fold increase in yield (overall yield 7.0%) versus the previous synthetic route (overall yield 0.42%). The final N-acyl prodrugs 18a (overall yield 3.9%) and 18b were subsequently obtained via a debenzylation reaction as in Scheme 3.</p><!><p>The phosphonic acid salts were evaluated as inhibitors of DXR from P. falciparum, and the results are shown in Table 1. Initially, the percent remaining enzyme activity was measured by treating the enzyme with each compound at a single concentration of 100 μM. This data shows the intrinsic activity of the compounds. Half-maximal inhibitory concentrations (IC50 values) were determined for compounds showing greater than 75% inhibition of DXR.</p><p>Of the new compounds, the most potent P. falciparum DXR inhibitor is 12a, with an IC50 value of 92 nM, slightly more potent than parent unsaturated compound 2. Within the N-alkoxy series of bisubstrate inhibitors 16d–g, compound 16e (4-iprPh) displays the most potent inhibition of the enzyme with an IC50 value of 2.11 μM. Interestingly, the N-alkoxy substituent of 16e was also found among the more active substituents in the saturated series.27 Compounds 16d and 16g with phenethyl and biphenyl substituents, respectively, were also active against the enzyme, displaying low μM inhibition.</p><p>Compounds 12b,c explore the influence of electronic effects in DXR inhibition, comparing an electron-withdrawing group (CF3, 12b) and electron-donating group (OCH3, 12c) with parent formyl analogue 12a and acetyl analogue 2. Compounds 12b and 12c display only weak and moderate inhibition of P. falciparum DXR, respectively. This result shows that electronic effects on the N-acyl group do not sway DXR inhibition as neither the CF3 group (12b) or OCH3 group (12c) improved the activity compared with the CH3 group of 2, which is a potent inhibitor of DXR.</p><!><p>POM prodrugs of selected analogues were synthesized in an effort to improve their cellular activity (and possibly bioavailability). All target compounds were tested for growth inhibition against P. falciparum following reported procedures (Table 2).35 This data indicates the inhibitory concentration of compound required to decrease growth of P. falciparum by 50% (Pf IC50).</p><p>Because of the penetrable cell membrane of eukaryotic parasite P. falciparum, as well as the significant remodeling of host cell membranes by malaria parasites, we expect substantial cellular uptake into P. falciparum.42 In Table 2, the polar phosphonic acid salts show significant activity against P. falciparum parasites. Compound 12a is the most active compound of the phosphonate salts, with an activity surpassing that of parent compound (and clinically evaluated candidate) fosmidomycin (1a). The data also shows that the inhibition of P. falciparum growth corresponds well to the activities of these compounds against the enzyme target P. falciparum DXR. Of the salts, compounds 12a and 16e were the most active DXR inhibitors. These compounds are also the most active inhibitors of P. falciparum growth among the salts.</p><p>The cellular activity of the POM prodrugs is also shown in Table 2. As was the case with the phosphonic acid salts, several of the POM prodrugs are highly active against P. falciparum. Of the N-acyl series, compound 12a was the most potent P. falciparum DXR inhibitor. Its prodrug, compound 18a, is the most potent prodrug inhibitor of P. falciparum (IC50 = 13 nM) from the POM series. In the N-alkoxy series of compounds, compound 16e was the most potent DXR inhibitor and also shows the highest potency (IC50 = 1.1 μM) against P. falciparum parasites. Interestingly, addition of the prodrug did not improve the activity of this compound.</p><p>As is evident from the data in Table 2, several compounds show potent antimalarial activity. Much of our work focuses on analogues of fosmidomycin, which is itself a reasonably potent inhibitor of P. falciparum growth (1a, IC50 = 1.087 μM). Modification of fosmidomycin with the sole change of added α,β-unsaturation yields compound 12a as a highly potent P. falciparum inhibitor with an IC50 value of 19 nM. Its prodrug 18a also potently inhibits P. falciparum with an IC50 value of 13 nM. This value is comparable to the inhibitory activity of current first-line antimalarial drug artemisinin (P. falciparum IC50 = 10.4 nM).35</p><!><p>Because of the remarkable activities of compounds 12a and 18a, additional studies were pursued. The computed cLogP,43 inhibition of HepG2, and selectivity indices (ratio of the antimicrobial activity to the human cell toxicity) for compounds 12a and 18a are shown in Table 3. As expected, compound 12a has a low cLogP value of −5.7. The prodrug strategy significantly increased the lipophilicity of the compound, yielding 18a with a cLogP of 0.89 (an increase of over 6 orders of magnitude). Neither the phosphonic acid salt 12a nor the prodrug 18a show toxicity against HepG2 cell lines, with IC50 values >50 μM. Thus, these compounds have excellent selectivity indices of 2632 and 3846 for 12a and 18a, respectively, against P. falciparum. These compounds show promise as safe drug candidates for malaria.</p><!><p>Compound 18a was designed to be a prodrug for 12a. To determine the rate of hydrolysis by plasma and hepatic esterases, compounds 12a and 18a were incubated in mouse liver microsomes (MLM) and in mouse plasma (Table 3). Compound 12a is stable in plasma (t1/2 > 120 min) and in mouse liver microsomes (t1/2 > 60 min). In contrast, POM prodrug 18a is very rapidly converted to compound 12a in plasma and microsomes (t1/2 < 5 min for both).</p><!><p>Because of their significant activities against P. falciparum growth, we examined compounds 12a and 18a in further assays to determine their intracellular mechanism of action. First, we asked if parasites treated with these inhibitors could be rescued by MEP pathway product IPP supplementation. If the inhibitors target the MEP pathway, parasite growth should be restored if exogenous IPP were added to the growth media.35 As is shown in Figure 4, addition of exogenous IPP effectively rescues growth of P. falciparum treated with 12a or 18a. This pattern is similar to the restoration effect observed in fosmidomycin (1a)-treated parasites supplied with IPP.35 In addition, P. falciparum growth inhibited by the N-alkoxy analogue 16e is also restored by IPP supplementation (Supporting Information, Figure S3). This data hints that these α,β-unsaturated fosmidomycin analogues might act on target to inhibit parasitic growth by blocking the MEP pathway, the targeted intracellular pathway.</p><p>Similarly, the proposed DXR inhibitor 18a should deplete MEP pathway intermediates beyond DXR from the treated P. falciparum. Thus, we employed a mass spectrometry-based method to quantify the MEP metabolites from untreated P. falciparum versus parasites treated with this compound (Figure 5).45 These metabolites include the DXR substrate (1-deoxy-D-xylulose 5-phosphate, DOXP), the DXR product (2-C-methylerythritol 4-phosphate, MEP), and the downstream metabolites (4-diphosphocytidyl-2-C-methylerythritol (CDP-ME) and 2-C-methyl-D-erythritol 2,4-cyclopyrophosphate (MEcPP)). Such metabolite levels were measured by LC-MS/MS after the 3D7 parasites were treated ± 18a after 10 h. The DOXP levels did not show significant difference between the untreated and treated parasites, while the metabolites downstream of DXR displayed reduced levels in treated parasites. This metabolic profiling data suggests that 18a inhibits DXR, the first committed enzyme in the MEP pathway from P. falciparum.</p><p>To further elucidate the mode of action of these analogues, compounds 12a and 18a were evaluated for efficacy against a unique mutated P. falciparum strain by using reported procedures.46 Because of a mutation in the metabolic regulator HAD1 (PF3D7_1033400), this P. falciparum strain produces high levels of DOXP, the substrate for DXR. Increased DOXP levels make it more difficult to inhibit DXR. For example, using this mutant, the (competitive) inhibition of DXR by fosmidomycin (1a) is impeded, resulting in fosmidomycin (1a)-resistant parasites (had1 parasites).35 As shown in Figure 6, had1 parasites (had1; open shapes, black line) were 3.4-fold and 2.5-fold more resistant to 12a and 18a, respectively, when compared with wild-type P. falciparum (3D7; closed shapes, gray line). Notably, the sensitivity of 12a and 18a were restored if the mutant strain was supplied with a wild-type copy of HAD1 (had1 + HAD1-GFP; closed shapes, black line). Similar results were observed with N-alkoxy analogue 16e, where had1 parasites were more resistant to these compounds than wild-type parasites (Supporting Information, Figure S2). These data corroborate earlier findings that these α,β-unsaturated fosmidomycin analogues inhibit P. falciparum growth via inhibition of the DXR enzyme in the MEP pathway.35</p><!><p>Compound 18a stands out among these analogues and appears to be a promising antimalarial agent due to its potent inhibition of P. falciparum growth. Thus, compound 18a was further evaluated for in vivo efficacy in a Plasmodium berghei-infected mouse model of malaria using reported procedures.35 Groups of mice were infected with luciferase-based blood-stage P. berghei ANKA by intraperitoneal (ip) injection. After being infected for 2 days, the mice were treated daily for 5 days with vehicle, 20 mg/kg chloroquine, 20 mg/kg 18a, or 50 mg/kg 18a via ip injection. Seven days after infection, intensity of the luciferin signal was measured, correlating to the parasitemia burden. As shown in Figure 7, parasitemia was greatly reduced in mice treated with the control drug chloroquine (20 mg/kg), lowering the luciferin signal intensity to 2.13 × 103. Interestingly, mice treated with 18a at the same dose showed a similar result as chloroquine, with a 3-log drop in luciferin signal intensity (2.87 × 103) when compared to the vehicle (2.62 × 106). When a higher dose of 18a was administered, the average luciferin signal intensity of the P. berghei-infected mice is 2.34 × 103, not significantly different from the result with the lower dose. Additionally, 18a was well tolerated in mice at these dosages, as no adverse effects were observed, corroborating our results with the HepG2 cells.</p><p>On the basis of our in vitro stability studies, we anticipated that 18a would be rapidly converted to 12a in the in vivo study. We designed a compound exposure study to determine plasma concentration of 12a under similar conditions to the efficacy study. Compound 18a was dosed at 20 mg/kg ip in Swiss Webster mice (n = 3). and plasma samples were removed at select time points over the course of 8 h (Figure 8). Over the course of 8 h, 12a was observed at high concentrations, with a concentration of 485 ng/mL (2.8 μM) at 8 h which is approximately 200-fold above the Pf IC50 of 18a in vitro.</p><!><p>With a growing number of drug-resistant cases of malaria, there is an urgent demand for new antimalarial agents with novel modes of action. DXR, an essential enzyme in the causative pathogens of malaria, while completely absent in humans, appears to be a promising drug target for the treatment of this disease. We aim to develop compounds that selectively inhibit this crucial enzyme to eliminate P. falciparum while avoiding toxicity in humans. Thus, several α,β-unsaturated analogues of fosmidomycin were designed and synthesized, including the N-alkoxy and N-acyl series. Our lead compound, 18a, stands out as a superior antimalarial agent. The excellent safety profile, transparent mode of action, and encouraging in vivo efficacy of 18a in eliminating Plasmodium spp. parasites establish 18a as a promising drug candidate for the treatment of malaria. Encouraging activities of these N-alkoxy and N-acyl fosmidomycin analogues suggests a new direction for the development of more prospective drug candidates in the fight against this important human pathogen.</p><!><p>1H and 13C NMR spectra were recorded in CDCl3, CD3OD, or D2O on Agilent spectrometer at 400 and 101 MHz, respectively, with TMS, H2O, or solvent signal as internal standard. Chemical shifts are given in parts per million (ppm). Spin multiplicities are given with the following abbreviations: s (singlet), br s (broad singlet), d (doublet), dd (doublet of doublets), ddd (doublet of doublets of doublets), t (triplet), dt (doublet of triplets), ddt (doublet of doublet of triplets), q (quadruplet), qt (quintuplet), and m (multiplet). Mass spectra were measured in the ESI mode on an HPLC-MS (Agilent 1100) or in the EI mode on an GC-MS (Shimadzu GCMS-QP2010S). Thin layer chromatography (TLC) was performed on Baker-flex Silica Gel IB2-F silica plates, and flash column chromatography was carried out using SiliCycle SiliaFlash P60 silica gel (40–63 μm). All reagents were purchase from commercial suppliers and used without further purification. Anhydrous solvents were purified by MBRAUN MB-SPS solvent purification system before use. All air sensitive reactions were carried out under nitrogen atmosphere. The purity of synthesized compounds (>95%) was determined by 1H/13C NMR in combination with HPLC-MS (Agilent 1100). Column: Thermo Fisher Scientific Hypersil GOLD aQ C-18 3 μm particle (250 mm × 4.6 mm). Mobile phase (containing 0.1% formic acid as the additive): linear gradient of acetonitrile (50%–100%) in water at a flow rate of 0.8 mL/min over 12.5 min, followed by 100% acetonitrile that was maintained for another 12.5 min. The UV detection wavelength was 210 and 254 nm. High-resolution mass spectroscopy spectra (HRMS) were recorded in positive or negative ESI mode on a Waters Q-TOF Ultima mass spectrometer (UIUC Mass Spectrometry Laboratory) or in positive FAB mode on a VG Analytical VG70SE magnetic sector mass spectrometer (JHU Mass Spectrometry Facility).</p><!><p>To a solution of MeOH (10.1 equiv) in dry CH2Cl2 (1 M) under N2 was added acetyl chloride (10 equiv) dropwise at room temperature, and the mixture was stirred for 10 min. The reaction mixture was then added a solution of 9 (1 equiv) in dry CH2Cl2 (1 M) and stirred at room temperature for 30 min. After the completion of deprotection, dry Na2CO3 (12 equiv) was added at 0 °C and the mixture was stirred at the same temperature for 10 min. The reaction mixture at 0 °C was added dropwise RCOCl, (CF3CO)2O, or N-formylimidazole* (2 equiv). The mixture was then warmed up to room temperature and stirred for 30 min to 24 h, quenched with saturated NaHCO3 (aq), and extracted with CH2Cl2 (3×). The combined organic layers were dried with anhydrous Na2SO4, filtered, and concentrated under reduced pressure. The crude concentrate was then purified by column chromatography on silica gel using EtOAc and CH2Cl2 (with a ratio from 1/10 to 2/1) to give the pure title compound.</p><!><p>To a solution of 10, 13, or 17 (1 equiv) in dry CH2Cl2 (0.1 M) under N2 was added boron trichloride (1 M in CH2Cl2, 4 equiv) at −78 °C dropwise. The reaction mixture was stirred at −78 °C for 30 min to 3 h, quenched with saturated NaHCO3 (aq), and extracted with EtOAc (5×). The combined organic layers were dried with anhydrous Na2SO4, filtered, and concentrated under reduced pressure. The crude residue was then purified by column chromatography on silica gel using EtOAc and MeOH (EtOAc and CH2Cl2 for 18a–c) to give the pure title compound.</p><!><p>To a solution of 11a,b or 15d–g (1 equiv) in dry CH2Cl2 (0.1 M) under N2 was added TMSBr (10 equiv) dropwise at 0 °C. The reaction mixture was warmed to room temperature, stirred overnight, and then concentrated under reduced pressure. The mixture was dissolved in CH2Cl2, evaporated under reduced pressure, and dried under vacuum. The crude residue was then stirred in 0.5 M NaOH (1 equiv) in H2O at room temperature for 1 h, washed with Et2O three times, and lyophilized to give the title compounds.</p><!><p>To a solution of 10a–c (1 equiv) in dry CH2Cl2 (0.1 M) under N2 was added TMSBr (10 equiv) dropwise at 0 °C. The reaction mixture was warmed to room temperature, stirred overnight, and then concentrated under reduced pressure. The mixture was dissolved in CH2Cl2, evaporated under reduced pressure, and dried under vacuum. The crude residue was then stirred in 0.5 M NaOH (2 equiv) in H2O at room temperature for 1 h, washed with Et2O (3×), and lyophilized to give disodium salts as white solids. The crude solid was then dissolved in dry DMF (0.1M), and TEA (6 equiv), chloromethylpivalate (6 equiv), and NaI (0.1 equiv) were added. The reaction mixture was stirred at 60 °C for 24 h, quenched with H2O, and extracted with Et2O (3×). The combined organic layers were dried with anhydrous Na2SO4, filtered, and concentrated under reduced pressure. The crude product was then purified by column chromatography on silica gel using hexanes and EtOAc or CH2Cl2 and EtOAc to give the pure title compound.</p><!><p>Light-yellow oil (594 mg, 73%). 1H NMR (400 MHz, CDCl3) δ 8.24 (s, 1H), 7.41–7.29 (m, 5H), 6.72–6.57 (m, 1H), 5.82 (ddd, J = 12.4, 10.0, 6.3 Hz, 1H), 4.84 (s, 2H), 4.30–4.20 (m, 2H), 4.13–3.98 (m, 4H), 1.32–1.27 (m, 6H). 13C NMR (101 MHz, CDCl3) δ 163.3, 144.40 (d, J = 5.7 Hz), 134.1, 129.4, 129.2, 128.8, 120.8 (d, J = 188.7 Hz), 78.3, 61.9 (d, J = 5.6 Hz), 47.2 (d, J = 26.7 Hz), 16.3 (d, J = 6.3 Hz). LC-MS (ESI+): 328.2 m/z [M + H]+, 655.2 m/z [2M + H]+.</p><!><p>Light-yellow oil (106 mg, 55%). 1H NMR (400 MHz, CDCl3) δ 10.06 (s, 1H), 8.32 (s, 1H), 6.75–6.55 (m, 1H), 5.99–5.74 (m, 1H), 4.32–4.25 (m, 2H), 4.09–3.92 (m, 4H), 1.25 (t, J = 7.1 Hz, 6H). 13C NMR (101 MHz, CDCl3) δ 172.5, 162.78, 146.3 (d, J = 5.5 Hz), 118.7 (d, J = 189.1 Hz), 62.3 (d, J = 5.7 Hz), 48.</p><!><p>Light-yellow solids (29 mg, quantitative yield). 1H NMR (400 MHz, CD3OD) δ 8.32 and 7.99 (s, 1H), 6.52–6.30 (m, 1H), 6.11–5.93 (m, 1H), 4.32–4.17 (m, 2H). 13C NMR (101 MHz, CD3OD) δ 162.6, 137.5 (d, J = 5.0 Hz), 127.3 (d, J = 178.0 Hz), 48.7 (d, J = 23.6 Hz). LC-MS (ESI−): 361.0 m/z [2M-2Na + H]−, 542.2 m/z [3M-3Na + 2H]−. HRMS (ESI−) calculated for C4H7NNaO5P 202.9960, found 180.0065 [M − Na]−.</p><!><p>Light-yellow oil (38 mg, 9%). 1H NMR (400 MHz, CDCl3) δ 8.23 (s, 1H), 7.43–7.30 (m, 5H), 6.72 (ddt, J = 22.4, 17.2, 5.1 Hz, 1H), 5.96–5.82 (m, 1H), 5.66 (dd, J = 13.1, 0.8 Hz, 4H), 4.84 (s, 2H), 4.30–4.20 (m, 2H), 1.21 (s, 18H). 13C NMR (101 MHz, CDCl3) δ 176.7, 163.2, 146.0 (d, J = 6.0 Hz), 129.9, 129.5, 129.3, 128.8, 119.4 (d, J = 193.0 Hz), 81.5 (d, J = 5.4 Hz), 78.3, 47.0 (d, J = 25.8 Hz), 38.7, 26.8. LC-MS (ESI+): 500.2 m/z [M + H]+, 999.2 m/z [2M + H]+.</p><!><p>To a suspension of 1,1′-carbonyldiimidazole (688 mg, 4.3 mmol, 2 equiv) in CH2Cl2 (3 mL) was added formic acid (0.16 mL, 4.3 mmol, 2 equiv) dropwise at room temperature. The mixture was stirred at room temperature for 5 min, and then transferred dropwise to a solution of 24 (1.0 g, 2.1 mmol, 1 equiv) and TEA (0.85 mL, 6.4 mmol, 3 equiv) in CH2Cl2 (15 mL) at 0 °C. The reaction mixture was stirred at 0 °C for 30 min, quenched with saturated aqueous NaHCO3 (30 mL), and extracted with CH2Cl2 (3 × 50 mL). The combined organic layers were dried with anhydrous Na2SO4, filtered, and concentrated under reduced pressure. Chromatographic separation on silica gel (hexanes/EtOAc = 2/1 to 1/3) gave the title compound as a light-yellow oil (862 mg, 82%). 1H NMR (400 MHz, CDCl3) δ 8.23 (s, 1H), 7.43–7.30 (m, 5H), 6.72 (ddt, J = 22.4, 17.2, 5.1 Hz, 1H), 5.96–5.82 (m, 1H), 5.66 (dd, J = 13.1, 0.8 Hz, 4H), 4.84 (s, 2H), 4.30–4.20 (m, 2H), 1.21 (s, 18H). 13C NMR (101 MHz, CDCl3) δ 176.7, 163.2, 146.0 (d, J = 6.0 Hz), 129.9, 129.5, 129.3, 128.8, 119.4 (d, J = 193.0 Hz), 81.5 (d, J = 5.4 Hz), 78.3, 47.0 (d, J = 25.8 Hz), 38.7, 26.8. LC-MS (ESI+): 500.2 m/z [M + H]+, 999.2 m/z [2M + H]+.</p><!><p>Light-yellow oil (103 mg, 55%). 1H NMR (400 MHz, CDCl3) δ 9.41 (s, 1H), 8.42 and 7.94 (s, 1H), 6.87–6.69 (m, 1H), 6.07–5.88 (m, 1H), 5.71–5.59 (m, 4H), 4.37–4.32 (m, 2H), 1.21 (s, 18H). 13C NMR (101 MHz, CDCl3) δ 177.1, 163.0, 157.3, 147.3 (d, J = 6.1 Hz), 118.6 (d, J = 191.8 Hz), 81.6 (d, J = 4.9 Hz), 48.8 (d, J = 26.1 Hz), 38.7, 26.8. LC-MS (ESI+): 410.2 m/z [M + H]+, 819.2 m/z [2M + H]+. HRMS (ESI+) calculated for C16H28NO9P 409.1502, found 432.1388 [M + Na]+.</p><!><p>To a solution of 22 (2.8 g, 6.5 mmol, 1 equiv) in dry THF (50 mL) was added 23 (980 mg, 8 mmol, 1.2 equiv) and TEA (1.73 mL, 13 mmol, 2 equiv). The reaction mixture was stirred at reflux for 3 h and then concentrated under reduced pressure. Chromatographic separation on silica gel (hexanes/EtOAc = 2/1 to 1/2) gave the title compound as a colorless oil (1.0 g, 34%). 1H NMR (400 MHz, CDCl3) δ 7.41–7.26 (m, 5H), 6.87 (ddt, 1H),6.02–5.85 (m, 1H), 5.70–5.62 (m, 4H), 4.68 (s, 2H), 3.67–3.63 (m, 2H), 1.21 (s, 18H). 13C NMR (101 MHz, CDCl3) δ 176.8, 150.1 (d, J = 5.3 Hz), 137.5, 128.4, 128.4, 128.0, 117.8 (d, J = 192.1 Hz), 81.5 (d, J = 5.5 Hz), 76.4, 54.0 (d, J = 24.2 Hz), 38.7, 26.8. LC-MS (ESI+): 472.2 m/z [M + H]+, 943.2 m/z [2M + H]+.</p><!><p>P. falciparum DXR activity was assayed at 37 °C by spectrophotometrically monitoring the enzyme catalyzed oxidation of NADPH upon addition of 1-deoxy-D-xylulose 5-phosphate (DOXP; Echelon Biosciences, Salt Lake City, UT) to the assay mixture, as described previously.47,48 Briefly, the assay system contained 100 mM Tris pH 7.8, 25 mM MgCl2, 0.86 μM Pf DXR, and 150 μM NADPH. The reaction was initiated by adding 144 μM DOXP to the complete assay mixture. One unit of P. falciparum DXR activity is defined as the amount of enzyme that catalyzes the oxidation of 1 μM NADPH per min. The oxidation of NADPH was monitored at 340 nm using an Agilent 8453 UV—visible spectrophotometer equipped with a temperature regulated cuvette holder. All assays were performed in duplicate.</p><!><p>Asynchronous P. falciparum cultures (strain 3D7, MR4/ATCC) were diluted to 1% parasitemia and treated with inhibitors at concentrations ranging from 1.2 nM to 492.4 μM. Growth inhibition assays were performed in opaque 96-well plates at 100 μL culture volume. After 3 days, parasite growth was quantified by measuring DNA content using PicoGreen (Life Technologies) as described.46 Fluorescence was measured on a FLUOstar Omega microplate reader (BMG Labtech) at 485 nm excitation and 528 nm emission. Half-maximal inhibitory concentration (IC50) values were calculated by nonlinear regression analysis using GraphPad Prism software. For isopentenyl pyrophosphate (IPP) (Echelon) rescue experiments, 125 μM IPP was added to the appropriate wells for the duration of the experiment. Experiments were performed in triplicate.</p><!><p>All animal procedures were conducted in compliance with the New York University Institutional Animal Care and Use Committee under Protocol No. 160720. Female Swiss Webster mice, weighing 25–30 g, were infected via ip injection with 103 transgenic P. berghei expressing luciferase (PMID: 16051702). Two days later, mice groups of five mice were injected ip with vehicle (2% methylcellulose, 0.5% Tween 80) or treatments (18a at 20 and 50 mg/kg). As control, two infected mice were treated with chloroquine (ip 20 mg/kg). All mice were injected daily for 5 days. One day after 5 days of treatment (day 7 after infection), the mice were anesthetized by inhalation of isofluorane (controlled flow of 2.5% isofluorane in air was administered through a nose cone via a gas anesthesia system). Mice were injected ip with 150 mg/kg of D-luciferin potassium salt (Goldbio) dissolved in PBS. Mice were imaged 5–10 min after injection of luciferin with an IVIS 100 (Xenogen, Alameda, CA). Data acquisition and analysis were performed with the software LivingImage (Xenogen).</p>

PubMed Author Manuscript

Valence stabilization of polyvalent ions during gamma irradiation of their aqueous solutions by sacrificial protection. III-Valence stabilization of Fe(II) ions by organic additives

Valence stabilization of polyvalent ions in gamma irradiated aqueous solutions is sometimes necessary in some chemical operations. In previous publications, valence stabilization of some polyvalent ions in solution upon gamma irradiation was achieved by using inorganic additives capable of interacting with the oxidizing or reducing species formed during water radiolysis. The results showed that the nature and duration of valence stabilization of Fe(II) depend on the concentration of the inorganic additives used. In the present work, a series of some organic additives has been used to investigate their capability in inducing valence stabilization of polyvalent iron ions, taken as an indicator, in aqueous acidic solutions when subjected to extended gamma irradiation. The results showed that the efficiency of valence stabilization depends on the amount and chemical structure of the organic additive used.

valence_stabilization_of_polyvalent_ions_during_gamma_irradiation_of_their_aqueous_solutions_by_sacr

Introduction<!>Experimental<!><!>Chemicals and materials<!>Equipments<!>Preparation of solutions<!>Preparation of iron solutions<!>Preparation and irradiation of samples<!>Analysis of irradiated solution<!>Reaction rate constants<!>Results and discussions<!><!>Valence stabilization of Fe(II) ions during extended gamma irradiation in presence of organic acids<!><!>Valence stabilization of Fe(II) ions during extended gamma irradiation in presence of aliphatic aldehydes<!><!>Valence stabilization of Fe(II) ions, during extended gamma irradiation, in presence of aliphatic alcohols<!><!>Radiolysis of Fe(III) acidic solutions in presence of organic additives<!><!>Effect of concentration of the organic additives and the resultant valence protection of Fe(II) during extended gamma irradiation<!>

<p>Protective effects against structural and ionic changes induced by traces of some chemicals during irradiation by X- or γ-radiations have been treated by several authors [1–5]. It was suggested that protection of irradiated systems containing aqueous polyvalent ions is a phenomenon that is greatly related to competitive reactions. Competition depends on the reaction rates and also on the equilibrium constant of the oxidation reduction reactions occurring in gamma irradiated aqueous systems containing polyvalent ions.</p><p>The protective effect of additives in some aqueous irradiated systems was observed long time ago. Fricke et al. [6] showed that addition of formic acid to X-irradiated aqueous acetic acid solution highly reduced the oxidation of acetic acid even if the concentration of the protective agent is 100 times lower than the concentration of acetic acid.</p><p>Other authors reported on the existence of protective effects in case of methylene blue decolorization by α rays, in presence of formic, malonic acids or gelatin [7]. Saturated compounds act more weakly than unsaturated compounds [8].</p><p>Stabilization of the oxidation states of certain polyvalent ions when present in a strong irradiation field is an important problem in applied radiation chemistry.</p><p>It is well known that in aqueous irradiated systems the nature of reactions between solutes and primary products of water radiolysis are different. The reactions occurring in presence of transition metal ions generally occur by electron transfer (Oxidation–Reductions reactions) while in systems containing organic solutes the reactions predominantly occur by hydrogen abstraction or addition reactions [9]. It is therefore possible to expect that in aqueous systems containing both transition metal ions and organic additives as solutes both type of reactions can occur. Many studies have been carried out to investigate the role of some inorganic additives in valence protection of some multivalent ions during radiolysis of their aqueous solutions. The results showed that valance stabilization continues for periods dependant on the concentration of the additive used and is most probably due to the competition reactions of the multivalent ions and the additives for the oxidizing or reducing species formed in the systems due to water radiolysis [10–14].</p><p>It is well known that aqueous Fe(II) solutions are rapidly oxidized under the effect of gamma radiations. The present work aims at investigating the possibility of protecting the divalent state of iron ions during extended gamma irradiation by using a series of some organic additives. The results showed that the efficiency of valence stabilization of Fe(II) during extended gamma irradiation depends on the amount and chemical structure of the organic additive used.</p><!><p>In the present work, extended gamma irradiation of aqueous acidic iron ions solutions in presence of different types of organic additives comprising alcohols, aldehydes or organic acids, has been undertaken. The effect of additive type and concentration on the prevailing reactions of the polyvalent ions in the irradiated systems has been particularly treated.</p><!><p>Extra pure ferrous ammonium sulphate (FeSO4(NH4)2SO4·6H2O), ferrous sulphate (FeSO4·7H2O), ferric sulphate (Fe2(SO4)3·9H2O) were obtained from May and Baker (M & B) Co. LTD., and the British Drug Houses (B.D.H.) England.</p><p>Chemically pure methanol, ethanol, n-propanol and n-butanol were obtained from Camberian Chemicals, England. Acetaldehyde, propionaldehyde and butyraldehyde were supplied from Prolabo Co., Paris. Formic acid, chemically pure 100 %, sp.g. 1.231 was also obtained from Prolabo, France, acetic acid for analysis, 99–100 % sp.g. 1.055–1.050 was obtained from Fein Chemie K–H. Kallies KG, Germany. Propionic acid, 99 % was obtained from B.D.H. Co. England. All these chemicals except alcohols were used without further purification. All alcohols were distilled twice over freshly ignited and cooled CaO.</p><p>Analytical grade chemicals such as 1.10, phenanthroline (M.W. 180.21) were obtained from B.D.H. England. Sulphuric acid (98 %); sp.g. 1.84, hydrochloric acid (35–37 %), sp.g. 1.18, were also obtained from B.D.H. England.</p><!><p>All other chemicals were of the analytical grade reagents and were used without further purification.</p><!><p>All pH measurements were carried out using an Orion Research pH meter model 616 A digital ion analyser with a combined glass-calomel electrode</p><p>Spectrophotometric measurements, were carried out using a Shimadzu UV–Vis double beam spectrophotometer type UV-2l0A. Glass or quartz cells were used whenever necessary.</p><p>Potentiometric titrations were carried out using a Radiometer type PO3 pH meter with Pt and saturated calomel electrodes.</p><!><p>All solutions were prepared using double distilled water. The water was boiled then cooled and stored in stoppered glass flasks.</p><!><p>Preparation of Fe(II) solution A.R. FeS04·7H2O crystals were washed twice with double distilled water followed by A.R. acetone and then dried by heating at 50 °C for a few minutes. Exactly 2.7803 g of the purified ferrous sulphate were weighed and dissolved in about 300 ml of freshly boiled and cooled bidistilled water. The solution was quantitatively transferred to a 1 l flask together with 22.2 ml cone. H2SO4, after being diluted with bidistilled water in 400 ml water. The resultant solution was then completed to the mark to give ~0.01 N Fe(II) solution in 0.8 N H2SO4.</p><p>The exact ferrous ion concentrations was determined titrimetrically with a standard (exactly about 0.1 N) potassium permanganate solution prepared as described in detail elsewhere. The end point was detected potentiometrically</p><p>Preparation of Fe(III) solution About 0.01 N Fe(III) solution was prepared by dissolving 3.999 g Fe2(SO4)3·9H2O in hot bidistilled water. The resultant solution was filtered and the filtrate was introduced into a 1 l volumetric flask together with 22.2 ml conc. H2SO4 previously diluted to 400 ml and the resultant solution was then completed to the mark with double distilled water to give a solution containing 0.8 N H2SO4. The exact concentration of ferric ions in the solution was titrimetrically determined against a standard EDTA solution using tiron indicator at 40–50 °C. At the equivalence point the solution turned from green to yellow.</p><!><p>The irradiated samples were prepared by taking 5 ml of 10−2 M Fe2+ or Fe3+ in 0.8 N H2SO4 together with the necessary amounts of different organic additives and the solutions were completed to the mark in 50 ml volumetric flasks. The resultant solutions were introduced into glass irradiation tubes (15 cm long and 2.5 m diameter) provided with a neck 1 cm in diameter ending with a ground glass stopper.</p><p>Irradiation of samples was carried out using a Canadian Co60 gamma cell-220 for extended time periods. The irradiation dose rate of the gamma cell was around 0.43 KGy per hour. Samples were withdrawn from the irradiated solutions at intervals and were analyzed. The irradiation dose of the irradiator was occasionally checked by the well known ferrous sulphate method.</p><p>During irradiation care was always taken to keep the position of the irradiation tubes unchanged along the whole irradiation time.</p><!><p>The concentration of existing Fe(II) ions in the irradiated solutions was followed spectrophotometrically at intervals by measuring the absorbance of the orange red complex formed with 1,10-phenanthroline against a reagent blank at 510 nm [15]. The molar absorptivity is 1.16 × 104 [10]. The unknown concentrations of iron in the analysed samples were determined by calibration curves constructed within the concentration range 1.0 × 10−5–1.5 × 10−4 M of Fe(II).</p><!><p>In the present work, all reaction rate constants (k) in dm3 mol−1 s−1, were used from the work of Anbar and Neta [16, 17].</p><!><p>In air free aqueous irradiated systems the following highly reactive primary water radiolysis products are formed. The corresponding G values, at pH 7 are as follows [18]:</p><p>In presence of small amounts of oxygen, H and OH radicals are rapidly scavenged as follows:2\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{O}}_{2} + { ext{H}} o { ext{HO}}_{2}^{ ullet } \quad K = 2.1\; imes \;10^{10} $$\end{document}O2+H→HO2∙K=2.1×1010 3\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{O}}_{2} + { ext{e}}^{ - } { ext{aq}} o { ext{O}}_{2}^{ ullet - } \quad K = 1.9 imes 10^{10} $$\end{document}O2+e-aq→O2∙-K=1.9×1010</p><p>Consequently, in aerated aqueous systems the most important reaction is the attack of OH radicals and to a lesser extent the perhydroxyl radicals (\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{HO}}_{ 2}^{ ullet } $$\end{document}HO2∙) and perhydroxide radical anions (\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{O}}_{2}^{ ullet - } $$\end{document}O2∙-) on the prevailing species in the irradiated systems [19].</p><p>It has been reported before that a 10−3 M Fe(II) acidic solution (0.08 N H2SO4) was completely oxidized when irradiated for about 2 h, using a gamma dose rate of 310 Gy/h i.e. after absorbing about 620 grays [13].</p><p>In the present work, valence stabilization of Fe(II) during extended gamma irradiation, in presence of different aliphatic organic acids, aldehydes or alcohols, has been investigated. Thus, increasing amounts of some aliphatic acids, aldehydes or alcohols were added to a certain concentration of Fe(II) ions in acidic aqueous solutions and the concentration of existing Fe(II) was followed spectrophotometrically during the continued irradiation of the systems. In the following sections the obtained results are discussed.</p><!><p>Percent existing Fe2+ in γ-irradiated 10−3 M Fe2+ solutions (0.08 N H2SO4) containing a Formic acid, b acetic acid c propionic acid; at various concentrations: 1—1.6 × 10−3 M (filled circle) 2—3.2 × 10−3 M (times) 3—8.0 × 10−3 M (open circle) 4—16.0 × 10−3 M (open triangle). Dashed line 100 % protection line, solid line actual protection line</p><!><p>Thus, in presence of acetic or propionic acid additives, oxidation of Fe(II) ions occurs very rapidly by OH radicals as follows:4\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{Fe}}^{ 2+ } + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{Fe}}^{ 3+ } + { ext{OH}}^{ - } \quad K = 1. 3 imes 10^{ 9} $$\end{document}Fe2++O∙H→Fe3++OH-K=1.3×109</p><p>At that stage, the organic acid additive probably does not interfere due to their relatively lower reaction rates as compared to the reaction rate of OH with Fe(II).5\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{HCOOH}} + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{H}}_{ 2} { ext{O}} + \mathop { ext{C}}\limits^{ ullet } { ext{OOH}}\quad K = 1. 6 imes 10^{ 8} $$\end{document}HCOOH+O∙H→H2O+C∙OOHK=1.6×108 6\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{CH}}_{ 3} { ext{COOH}} + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{H}}_{ 2} { ext{O}} + \mathop { ext{C}}\limits^{ ullet } { ext{H}}_{ 2} { ext{COOH}}\quad K = 1. 8 imes 10^{ 8} $$\end{document}CH3COOH+O∙H→H2O+C∙H2COOHK=1.8×108 7\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{C}}_{ 2} { ext{H}}_{ 5} { ext{COOH}} + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{H}}_{ 2} { ext{O}} + { ext{CH}}_{ 3} \mathop { ext{C}}\limits^{ ullet } { ext{HCOOH}}\quad K = 2.0 imes 10^{ 8} $$\end{document}C2H5COOH+O∙H→H2O+CH3C∙HCOOHK=2.0×108</p><p>When most of the iron ions are present in the Fe(III) state gradual reduction starts to take place by the action of H radicals as follows:8\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{Fe}}^{ 3+ } + { ext{H}}^{ ullet } o { ext{Fe}}^{ 2+ } + { ext{H}}^{ + } \quad K = 1. 3 imes 10^{ 7} $$\end{document}Fe3++H∙→Fe2++H+K=1.3×107</p><p>Organic additives are only slightly capable of affecting the reduction reaction of Fe(III) by H radicals since competition between the organic acids and Fe(III) for H radicals occurs in favor of the H radical reaction with Fe(III), as could be deduced from the lower rate values of the following reactions:9\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{HCOOH}} + { ext{H}}^{ ullet } o { ext{ H}}_{ 2} + \mathop { ext{C}}\limits^{ ullet } { ext{OOH}}\quad K = 2 imes 10^{ 6} $$\end{document}HCOOH+H∙→H2+C∙OOHK=2×106 10\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{CH}}_{ 3} { ext{COOH}} + { ext{H}}^{ ullet } o { ext{H}}_{ 2} + \mathop { ext{C}}\limits^{ ullet } { ext{H}}_{ 2} { ext{COOH}}\quad K = 1. 3 imes 10^{ 5} $$\end{document}CH3COOH+H∙→H2+C∙H2COOHK=1.3×105 11\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{CH}}_{ 3} { ext{CH}}_{ 2} { ext{COOH}} + { ext{H}}^{ ullet } o { ext{H}}_{ 2} + { ext{CH}}_{ 3} \mathop { ext{C}}\limits^{ ullet } { ext{HCOOH}}\quad K = 3. 2 imes 10^{ 6} $$\end{document}CH3CH2COOH+H∙→H2+CH3C∙HCOOHK=3.2×106whereby, \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ k_{ 8} /k_{ 9} = { 38}, \, k_{ 8} /k_{ 10} = 6 1 5, \, k_{ 8} /k_{ 1 1} = { 13}. 5 $$\end{document}k8/k9=38,k8/k10=615,k8/k11=13.5</p><p>Moreover, organic carboxylate radicals, now present at higher concentrations, due to H and OH radicals reactions with acids by reactions 5–7 and 9–11 can also contribute to the reduction process of Fe(III) as follows [20]:12\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ \mathop { ext{C}}\limits^{ ullet } { ext{OOH}} + { ext{Fe}}^{ 3+ } o { ext{Fe}}^{ 2+ } + { ext{CO}}_{ 2} + { ext{H}}^{+ } $$\end{document}C∙OOH+Fe3+→Fe2++CO2+H+ 13\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{R}}\mathop { ext{C}}\limits^{ ullet } { ext{HCOOH}} + { ext{Fe}}^{ 3+ } o { ext{Fe}}^{ 2+ } + { ext{R}}\mathop { ext{C}}\limits^{ ullet } { ext{HCOO}} + { ext{H}}^{ + } $$\end{document}RC∙HCOOH+Fe3+→Fe2++RC∙HCOO+H+</p><p>This could be further clarified by referring to the standard reduction potentials [21] of the following reactions:\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ egin{aligned} 2 { ext{Fe}}^{ 3+ } + 2 { ext{ e}} o 2 { ext{Fe}}^{ 2+ } \quad E^{ ext{o}} & =+0. 77 { ext{ v}} \ { ext{CO}}_{ 2} + 2 { ext{H}}^{ + } + 2 { ext{e}} o { ext{HCOOH}}\quad E^{ ext{o}} & = - 0. 20 { ext{ v}} \ \end{aligned} $$\end{document}2Fe3++2e→2Fe2+Eo=+0.77vCO2+2H++2e→HCOOHEo=-0.20v</p><p>Using these standard reduction potential values, it is possible to find out that the equilibrium constant of the reaction [22]\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ 2 { ext{Fe}}^{ 3+ } + { ext{ HCOOH}} o { ext{ 2 Fe}}^{ 2+ } + 2 { ext{H}}^{ + } + { ext{ CO}}_{ 2} $$\end{document}2Fe3++HCOOH→2 Fe2++2H++CO2is equal to 7.6 × 1032 which shows that the reaction is very favorable.</p><p>At that stage, reduction of Fe(III) continues until most iron ions were reduced to Fe(II) ions.</p><p>It could also be observed in Fig. 1c that in presence of propionic acid, reduction of ferric ions occurred more slowly than in case of acetic acid. This is probably due to the fact that propionic acid competes more effectively for H radicals than acetic acid. This is further confirmed by the decrease in the rate of Fe(III) reduction when greater propionic acid concentrations were used .</p><p>At the end of Fe(III) reduction stage the concentration of the formed Fe(II) ions remained almost constant during continued gamma irradiation for durations depending on the amount of organic acid used. During that stage OH radicals actively interact with the organic acid existing in the medium (by reactions 5, 6, 7) being present in greater excess than Fe(II) ions. At the end of the valence protection stage, the organic acid concentration gradually decreases and consequently OH radicals gradually interact with Fe(II) ions until most irons are transformed to the trivalent state.</p><!><p>Percent existing Fe2+ in γ-irradiated 10−3 M Fe2+ solutions (0.08 N H2SO4) containing a Acetaldehyde, b Propionaldehyde, c Butyraldehyde; at various concentrations: 1—1.6 × 10−3 M (filled circle) 2—3.2 × 10−3 M (times) 3—8.0 × 10−3 M (open circle) 4—16.0 × 10−3 M (open triangle). Dashed line 100 % protection line, solid line actual protection line</p><!><p>It has been reported before that at the beginning of irradiation Fe(II) ions are oxidized very rapidly in presence of aliphatic aldehydes and after a very short steady state Fe(III) is rapidly reduced to Fe(II) ions [12]. On continued irradiation Fe(II) ions survived until the end of the first protection period at the end of which rapid oxidation of Fe(II) to the trivalent state occured. The second protection period starts by the gradual reduction until most of Fe(III) is transformed to Fe(II), the concentration of which remained almost stable for durations depending on the aldehyde concentration used.</p><p>It is possible to assume that the stability of Fe(II) ions during the first protection period is very probably due to aldehydes acting as OH radical scavengers according to the following reactions:14\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ egin{aligned} & { ext{CH}}_{ 3} { ext{CHO}} + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o \mathop { ext{C}}\limits^{ ullet } { ext{H}}_{ 2} { ext{CHO}} + { ext{H}}_{ 2} { ext{O}}\quad K = 7 { } imes { 1}0^{ 8} \ & { ext{CH}}_{ 3} { ext{CH}}_{ 2} { ext{CHO}} + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{CH}}_{ 3} \mathop { ext{C}}\limits^{ ullet } { ext{HCHO}} + { ext{H}}_{ 2} { ext{O}} \ \end{aligned} $$\end{document}CH3CHO+O∙H→C∙H2CHO+H2OK=7×108CH3CH2CHO+O∙H→CH3C∙HCHO+H2O 15\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{CH}}_{ 3} { ext{CH}}_{ 2} { ext{CH}}_{ 2} { ext{CHO}} + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{CH}}_{ 3} { ext{CH}}_{ 2} \mathop { ext{C}}\limits^{ ullet } { ext{HCHO}} + { ext{H}}_{ 2} { ext{O}}\quad K = { 2}. 3 { } imes { 1}0^{ 9} $$\end{document}CH3CH2CH2CHO+O∙H→CH3CH2C∙HCHO+H2OK=2.3×109While4\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{Fe}}^{ 2+ } + \mathop { ext{O}}\limits^{ ullet } { ext{H}} o { ext{Fe}}^{ 3+ } + { ext{OH}}^{ - } \quad K = 1. 3 imes 10^{ 9} $$\end{document}Fe2++O∙H→Fe3++OH-K=1.3×109</p><p>Aldehydes being present in a great excess as compared to Fe(II) concentration can effectively remove OH radicals. During the first protection period continued gamma irradiation continuously changes aldehydes to the corresponding acids through their transformation to the hydrates followed by their interaction with H or OH radicals [19].</p><p>The formed organic radicals can easily change to the corresponding acids:20\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ 2 { ext{H}}\mathop { ext{C}}\limits^{ ullet } ( { ext{OH)OH }} o \, 2{ ext{HCOOH }} + { ext{ H}}_{ 2} $$\end{document}2HC∙(OH)OH→2HCOOH+H2 21\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ 2 { ext{CH}}_{ 3} \mathop { ext{C}}\limits^{ ullet } ( { ext{OH)OH }} o \, 2{ ext{CH}}_{ 3} { ext{COOH }} + { ext{ H}}_{ 2} $$\end{document}2CH3C∙(OH)OH→2CH3COOH+H2 22\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{H}}\mathop { ext{C}}\limits^{ ullet } ( { ext{OH)OH }} + { ext{ H}}_{ 2} { ext{O}}_{ 2} \, o { ext{ HCOOH }} + { ext{ H}}_{ 2} { ext{O }} + \, \mathop { ext{O}}\limits^{ ullet } { ext{H}} $$\end{document}HC∙(OH)OH+H2O2→HCOOH+H2O+O∙H</p><p>At the same time, oxidation of Fe(II) by OH radicals is rendered ineffective. Oxidation of aldehydes to the corresponding acids can also occur by interaction of oxygen liberated from water radiolysis by the following reaction [23].</p><p>It is therefore possible to assume that aldehydes are continuously changed during irradiation to the corresponding acids.</p><p>At the beginning of the second protection period i.e. during the Fe(III) reduction by H radicals, aldehydes can enhance the Fe(III) reduction to Fe(II) being themselves changed to the corresponding acids [24, 25].24\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ 2 { ext{Fe}}^{ 3+ \, } + { ext{HCHO}} \mathop{\longrightarrow}\limits^{{ m H}_{2}{ m O}} 2 { ext{Fe}}^{ 2+ \, } + { ext{ HCOOH }} + 2 { ext{H}}^{ + } $$\end{document}2Fe3++HCHO⟶H2O2Fe2++HCOOH+2H+ 25\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ 2 { ext{Fe}}^{ 3+ \, } + { ext{CH}}_{ 3} { ext{CHO}} \mathop{\longrightarrow}\limits^{{ m H}_{2}{ m O}} 2 { ext{Fe}}^{ 2+ } + { ext{CH}}_{ 3} { ext{COOH}} + 2 { ext{H}}^{ + } $$\end{document}2Fe3++CH3CHO⟶H2O2Fe2++CH3COOH+2H+This could be further clarified by calculating the equilibrium constant of these reactions using the standard reduction potential values [22] of the half reactions involved. The equilibrium constant values obtained are 2.8 × 1026 and 1.4 × 1030 for reactions 24 and 25 respectively showing that these reaction are very favorable.</p><p>When all iron present is reduced to the divalent state, continued survival of Fe(II) ions during the second protection period very probably occurs by continued interaction of OH radicals with the existing acids. That continues until the complete exhaustion of the formed acids at the end of the second protection period.</p><!><p>Percent existing Fe2+ in γ-irradiated 10−3 M Fe2+ solutions (0.08 N H2SO4) containing a Methanol, b Ethanol, c Propanol, d Butanol; at various concentrations: 1—3.2 × 10−3 M (filled circle) 2—4.8 × 10−3 M (times) 3—8.0 × 10−3 M (open circle) 4—14.0 × 10−3 M (open triangle). Dashed line 100 % protection line, solid line actual protection line</p><!><p>It is possible to observe the existence of two protection periods separated by an oxidation then reduction stages. It is interesting to note that the first protection period in case of alcohols is much more developed than in case of aldehydes, while the second protection period is almost similar to that obtained in case of aldehydes and acid additives.</p><p>According to Broszkiewicz [11] at the beginning of irradiation of 10−4 M Fe(II) in presence of aliphatic alcohols in 0.1 N H2SO4 at a dose rate of 2.97 × 1016 eV/ml min−1, organic peroxides are formed capable of oxidizing several Fe(II) ions per OH radical as follows [26]:</p><p>That probably explains the rapid oxidation of Fe(II) at the beginning of irradiation. This was followed by a steady state whereby after the absorption of 440 Grays reduction of the formed trivalent iron occurred gradually. In general, after about 1 h of irradiation complete reduction of iron ions to Fe(II) was observed.</p><p>In the present work, when the initial oxidation, steady state then reduction were over, divalent iron ions remained protected upon continued irradiation for periods dependant on the alcohol concentration used until the end of the first protection period. This probably shows that alcohols act as active scavengers for the oxidizing radicals during that period as follows:26\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{CH}}_{ 3} { ext{OH }} + \mathop { ext{O}}\limits^{ ullet } { ext{H }} o \, \mathop { ext{C}}\limits^{ ullet } { ext{H}_{2}} { ext{OH}}+ { ext{H}}_{ 2} { ext{O}}\quad K = 0. 5 3 imes 10^{ 9} $$\end{document}CH3OH+O∙H→C∙H2OH+H2OK=0.53×109 27\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{C}}_{ 2} { ext{H}}_{ 5} { ext{OH }} + \mathop { ext{O}}\limits^{ ullet } { ext{H }} o { ext{ CH}}_{ 3} \mathop { ext{C}}\limits^{ ullet } { ext{HOH}} + { ext{H}}_{ 2} { ext{O}}\quad K = 0. 8 7 imes 10^{ 9} $$\end{document}C2H5OH+O∙H→CH3C∙HOH+H2OK=0.87×109 28\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{C}}_{ 3} { ext{H}}_{ 7} { ext{OH }} + \mathop { ext{O}}\limits^{ ullet } { ext{H }} o { ext{ C}}_{ 2} { ext{H}}_{ 5} \mathop { ext{C}}\limits^{ ullet } { ext{HOH }} + { ext{ H}}_{ 2} { ext{O}}\quad K = 0. 6 8 imes 10^{ 9} $$\end{document}C3H7OH+O∙H→C2H5C∙HOH+H2OK=0.68×109 29\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{C}}_{ 4} { ext{H}}_{ 9} { ext{OH }} + \mathop { ext{O}}\limits^{ ullet } { ext{H }} o { ext{C}}_{ 3} { ext{H}}_{ 7} \mathop { ext{C}}\limits^{ ullet } { ext{HOH}} + { ext{H}}_{ 2} { ext{O}}\quad K = 4. 6 imes 10^{ 9} $$\end{document}C4H9OH+O∙H→C3H7C∙HOH+H2OK=4.6×109 30\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{Fe}}^{ 2+ } + \mathop { ext{O}}\limits^{ ullet } { ext{H }} o { ext{ Fe}}^{ 3+ } + { ext{OH}}^{ - } \quad K = 1. 3 imes 10^{ 9} $$\end{document}Fe2++O∙H→Fe3++OH-K=1.3×109</p><p>Although the rate constants of the \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ \mathop { ext{O}}\limits^{ ullet } { ext{H}} $$\end{document}O∙H reactions with alcohols are slightly lower than the rate constant of \documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ \mathop { ext{O}}\limits^{ ullet } { ext{H}} $$\end{document}O∙H reaction with Fe(II) yet, the greater concentration of the organic alcohols enhances their interaction with OH radicals and consequently the valence of Fe(II) ions remains rather stable. Moreover, alcohol radicals formed can further reduce any formed Fe(III) ions being themselves transformed to the corresponding aldehyde [27].31\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{RCH}}_{ 2} { ext{OH }} + { ext{ OH }} o { ext{ R}}\mathop { ext{C}}\limits^{ ullet } { ext{HOH}} + { ext{ H}}_{ 2} { ext{O}} $$\end{document}RCH2OH+OH→RC∙HOH+H2O 32\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{R}}\mathop { ext{C}}\limits^{ ullet } { ext{HOH}} + { ext{OH }} o { ext{ RCHO}} + { ext{ H}}_{ 2} { ext{O}} $$\end{document}RC∙HOH+OH→RCHO+H2O 33\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ { ext{R}}\mathop { ext{C}}\limits^{ ullet } { ext{HOH }} + { ext{ Fe}}^{ 3+ } o { ext{RCHO}} + { ext{Fe}}^{ 2+ } + { ext{H}}^{ + } $$\end{document}RC∙HOH+Fe3+→RCHO+Fe2++H+</p><p>Therefore, during the first protection period alcohols are continuously transformed to the corresponding aldehydes. Actually, the first protection period in case of alcohols is much more developed than that occurring in aldehydes. It is probably possible to assume that the first protection period in presence of alcohols is due to the protective effect occurring during alcohol transformation to aldehydes and aldehydes transformation to the corresponding acids. This probably is confirmed by the fact that the rate constants of OH reaction with alcohols (reactions 26–29) are comparable to the rate constants of OH reactions with the corresponding aldehydes (reactions 14, 15).</p><p>The second protection period in case of alcohol additives is similar to that observed in case of aldehyde additives. Valance stability of Fe(II) ions is very probably due to the scavenging effect of the formed organic acids on OH radicals. This occurs until all the formed organic acids are exhausted leading to the restoration of the oxidative effect of OH radicals on Fe(II) ions (reaction 4).</p><!><p>Percent existing Fe2+ in γ-irradiated 10−3 M Fe3+ solutions (0.08 N H2SO4) containing 16.0 × 10−3 M of: 1—Acetic acid, 2—Acetaldehyde, 3—Ethanol</p><!><p>On using aldehyde or alcohol additives, rapid reduction of iron ions to the divalent state occurred most probably by reaction 8. Then, two protection periods were observed.</p><p>In case of aldehyde additive, the first period most probably involves aldehyde protection accompanied with oxidation of the aldehyde to the corresponding acid. When the aldehyde is consumed Fe(II) decays gradually to the trivalent state. This is followed by a reduction stage until iron ions are almost completely reduced to the divalent state whereby the second protection period started and continued until the complete decay of the formed acid.</p><p>In case of alcohol additive two protection periods also exists, as in case of aldehyde additive. However, the first protection period is much more developed than in case of aldehyde additives as has been also observed in the Fe(II) systems discussed before. This has been attributed to the consecutive alcohol and aldehyde protection of existing divalent iron in the first protection period. At the end of the first protection period Iron ions are oxidized to the trivalent state. The second protection period started by the gradual reduction of Fe(III) ions to the divalent state. This is followed by a valence stable stage, for Fe(II) until the formed acid is completely exhausted whereby final oxidation of iron(II) occurred. This conforms with the previous discussions on the prevailing reactions in case of Fe(II) systems.</p><!><p>Relationship between the organic additives concentration and the area under the protection curves (in cm2) 1—Acetic acid, 2—Propionic acid 3—Acetoldehyde, 4—Propionaldehyde, 5—Butyraldehyde, 6—n-Butyl alcohol, 7—n-Propyl alcohol, 8—Ethyl alcohol, 9—Methyl alcohol</p><!><p>In order to evaluate the overall protection capacity occurring in different systems, the percent of total protection was determined as follows:\documentclass[12pt]{minimal} sepackage{amsmath} sepackage{wasysym} sepackage{amsfonts} sepackage{amssymb} sepackage{amsbsy} sepackage{mathrsfs} sepackage{upgreek} \setlength{\oddsidemargin}{-69pt} egin{document}$$ \% { ext{ Total protection }} = \, \left( {A_{ ext{p}} /A_{ ext{t}} } ight) \, imes { 1}00 $$\end{document}%Total protection=Ap/At×100where A p is the area under the protection line and A t total area under the 100 % protection line.</p><!><p>Percent protection of 10−3 M Fe(II) solution during extended gamma irradiation in presence of different organic additives</p><p>aTotal absorbed dose till the complete decay of the used amount of additive</p><p>bPercent taken as a measure of actual protection</p><p>Organic acids, aldehydes and alcohols can effectively protect divalent iron ions against oxidation under the effect of continued gamma irradiation.</p><p>The extent of protection depends on the type and amount of the additives used.</p><p>Protection could be attributed to the competition reactions of the divalent iron ions and organic additive for the oxidizing radicals, formed during water radiolysis.</p>

PubMed Open Access

The Aggregation Pheromone of Phyllotreta striolata (Coleoptera: Chrysomelidae) Revisited

Aggregations of the striped flea beetle Phyllotreta striolata on their crucifer host plants are mediated by volatiles emitted from feeding males. The male-specific sesquiterpene, (6R,7S)-himachala-9,11-diene (compound A), was shown previously to be physiologically and behaviorally active, but compound A was attractive only when combined with unnaturally high doses of the host plant volatile allyl isothiocyanate (AITC) in field trapping experiments. This indicated that our understanding of the chemical communication in this species is incomplete. Another male-specific sesquiterpenoid, (3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one (compound G), has been reported from an American P. striolata population. We confirmed the presence of compound G, and investigated its interaction with compound A and AITC in a P. striolata population in Taiwan. Compound G was attractive to Taiwanese P. striolata in laboratory bioassays, but significantly more beetles were attracted to a blend of compounds A and G. Under the same conditions, P. striolata showed no preference for the blend of A and G combined with a range of doses of AITC over the sesquiterpenoid blend alone. The sesquiterpenoid blend was tested further in field trapping experiments and attracted significantly more beetles than traps baited with compound A and ecologically relevant amounts of AITC. We conclude that A and G are components of the male-specific aggregation pheromone of P. striolata in Taiwan, and that the attractiveness of the pheromone is not reliant on the presence of AITC. Our results further indicate that the male-specific sesquiterpenoid blends differ qualitatively between the Taiwanese and American populations of P. striolata.

the_aggregation_pheromone_of_phyllotreta_striolata_(coleoptera:_chrysomelidae)_revisited

Introduction<!>Insects and Plants<!>Volatile Collections<!>Gas Chromatography-Mass Spectrometry (GC-MS)<!>Synthesis of Compounds A and G<!>Two-Choice Laboratory Bioassay<!>Experiment 1<!>Experiment 2<!>Experiment 3<!>Field Trapping Experiment<!>Data Analysis<!><!>Identification of Compounds G and I in Volatile Collections from P. striolata Males<!><!>Quantification of AITC in Headspace Samples<!>Two-Choice Laboratory Bioassay<!><!>Field Trapping Experiment<!><!>Discussion

<p>Phyllotreta flea beetles (Coleoptera: Chrysomelidae) aggregate on their host plants, which almost exclusively belong to the order Brassicales. These host plants include many economically important crops such as cabbage, mustard, and canola, and several Phyllotreta species are important pests of Brassica crops (Andersen et al. 2005, 2006; Lamb 1989). The beetles' typical "shotgun" feeding damage on cotyledons and leaves can cause considerable crop loss in the seedling stage (Westdal and Romanow 1972) and reduce the marketability of vegetables.</p><p>The aggregation behavior of Phyllotreta spp. is mediated by volatiles emitted from feeding males (Beran et al. 2011; Peng et al. 1999) and facilitates rapid mass infestations in the field. Comparative headspace analyses from feeding males and females revealed a number of male-specific compounds identified as γ-cadinene and himachalene-type sesquiterpenoids (Bartelt et al. 2001, 2011; Beran et al. 2011; Tóth et al. 2012). For example, male Phyllotreta cruciferae emit six sesquiterpenoids of which three elicit electrophysiological responses from beetle antennae (Tóth et al. 2005). Field tests showed that (6R,7S)-himachala-9,11-diene (compound A) is the key aggregation pheromone of P. cruciferae (Tóth et al. 2005). The synthetic sesquiterpenoid alone attracted only few adults in the field, but synergistically enhanced the attractiveness of the plant volatile allyl isothiocyanate (AITC), a known attractant for many Phyllotreta species (Pivnick et al. 1992; Soroka et al. 2005; Tóth et al. 2005, 2007). Several other Phyllotreta species were caught together with P. cruciferae (Tóth et al. 2005) suggesting similarities in their chemical communication. Indeed, compound A also was identified as a component of the aggregation pheromone of Phyllotreta vittula and Phyllotreta striolata (Beran et al. 2011; Tóth et al. 2012). The presence of volatile isothiocyanates (ITCs) was crucial for the behavioral response of P. vittula and P. striolata to component A in the field (Beran et al. 2011; Tóth et al. 2012). However, the AITC doses required to attract beetles greatly exceeded emission rates from host plants (Najar-Rodriguez et al. 2015; Pivnick and Jarvis 1991) indicating that our understanding of how aggregations occur in these species is still limited.</p><p>Isothiocyanates are characteristic defense compounds of plants in the order Brassicales formed via enzymatic hydrolysis of glucosinolates (Halkier and Gershenzon 2006). In intact plant tissue, the corresponding plant enzyme myrosinase, a β-thioglucosidase, is spatially separated from glucosinolates. Herbivore feeding triggers glucosinolate hydrolysis in the damaged plant tissue and non-adapted herbivores are deterred or poisoned by the hydrolysis products, whereas adapted herbivores prevent ITC toxicity using different strategies (Winde and Wittstock 2011). Interestingly, P. striolata adults possess an endogenous myrosinase and release small quantities of volatile ITCs derived from glucosinolates they sequestered from their food plants (Beran 2011; Beran et al. 2014). However, these amounts are much less than the doses required for attraction.</p><p>We previously identified six male-specific sesquiterpenoid compounds in volatiles from a Taiwanese P. striolata population, and found compound A to be physiologically and behaviorally active. Intriguingly, Bartelt et al. (2011) detected a novel male-specific sesquiterpenoid, (3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one (compound G), as major compound in volatile collections from an American P. striolata population. Compound G elicited electrophysiological responses from beetle antennae, but behavioral responses were not assessed. Additionally, (1S,2R)-2,6,6-trimethylbicyclo[5.4.0]undec-7-en-9-one (compound H) and (6R,7S)-2,2,6-trimethyl-10-methylene-bicyclo[5.4.0]-undec-1,11-ene (compound I) were detected as minor compound in volatile emissions from American P. striolata, but these elicited no electrophysiological activity (Bartelt et al. 2011).</p><p>It is assumed that P. striolata has been introduced from Eurasia to North America (Bain and LeSage 1998; Smith 1985); however, an analysis of cytochrome oxidase I (COI) revealed 3.3 to 5.7 % sequence divergence between populations from Eurasia and Canada (Beran 2011), indicating that these populations have been separated for at least one million years (Farrell 2001; Juan et al. 1995). With this background, we asked whether population-specific chemical profiles explain the different results obtained in previous studies (Bartelt et al. 2011; Beran et al. 2011). We reassessed the male-specific volatiles from the Taiwanese P. striolata population and determined the behavioral responses of P. striolata to compounds found, alone and in combination with ecologically relevant amounts of AITC.</p><!><p>Phyllotreta striolata adults were collected from crucifer fields at AVRDC-The World Vegetable Center in Shanhua, Taiwan, and shipped to the Max Planck Institute for Chemical Ecology in Jena. The import authorization to Germany was obtained under Directive 2008/61/EC. Adults were maintained in mesh cages (MegaView Science Co., Ltd., Taichung, Taiwan) on potted 3–4-wk.-old Brassica juncea cv. Bau Sin plants in a controlled-environment chamber at 24 °C, 65 % relative humidity, and L14:D10 h. Seeds of B. juncea were purchased from Known-You Seed Co. LTD, Kaohsiung, Taiwan.</p><!><p>Volatiles were collected from groups of 14 to 20 male P. striolata adults feeding on cut leaves of B. juncea for one day or for three consecutive days in the laboratory under ambient conditions. Compressed air purified by activated charcoal was passed through a 100 ml glass bottle containing beetles and leaf material at a flow rate of 50–100 ml/min. Volatile compounds were trapped on SuperQ filters (25 mg; ARS Inc. Gainsville, FL, USA), which were afterwards eluted with 100 μl hexane containing 10 ng/μl bromodecane (Sigma-Aldrich) as internal standard. For comparison, collections also were made using activated charcoal filters (1.5 mg CLSA filter, Gränicher & Quartero, Daumazan sur Arize, France), and volatiles were eluted with 30 μl hexane containing 10 ng/μl bromodecane as internal standard (Beran et al. 2011).</p><!><p>Collections of volatiles were analyzed using an Agilent 6890 N gas chromatograph (GC; Waldbronn, Germany) equipped with a ZB-5MSi capillary column (30 m × 0.25 mm ID, 0.25 μm film thickness; Phenomenex, Aschaffenburg, Germany) coupled to an Agilent 5973 quadrupole mass spectrometer (Agilent). The carrier gas was helium at constant flow (1 ml/min). One microliter from each sample was injected in splitless mode into the inlet at 220 °C. The oven program started at 40 °C for 3 min, increased at 10 °C/min to 270 °C, and then with 50 °C/min to 300 °C and held for 2 min. MS conditions were electron impact mode (70 eV), and scan mode 33–250 amu (amu). Male-specific sesquiterpenoids were identified as described in Beran et al. (2011). The identity of compounds G and I detected in volatile collections from P. striolata was confirmed by comparing the mass spectra and retention time to reference compounds obtained from Dr. Allard Cossé (USDA-ARS) on the non-polar ZB-5MSi column (conditions as described above) and on an enantiospecific Cyclosil-B column (30 m × 0.25 mm ID, 0.25 μm film thickness, Agilent). The inlet temperature was set to 150 °C to avoid thermal rearrangement of compound G upon injection (Bartelt et al. 2011). The carrier gas was helium at constant flow (1.1 ml/min). The oven program started at 70 °C for 3 min, increased at 1 °C/min to 170 °C, and then at 50 °C/min to 240 °C held for 5 min. MS conditions were electron impact mode (70 eV), and scan mode 33–250 amu.</p><!><p>The synthesis of compound A was performed as described in Jimenez-Aleman et al. (2012), with one modification. We used silica impregnated with CuSO4 instead of AgNO3 for purification of compound A (Szumilo and Flieger 2001). Due to the higher stability of copper ions relative to silver, CuSO4 -modified silica gel can be prepared in advance, handled in the presence of light, stored for several months without appreciable loss of activity and gave separations comparable to those obtained with silica gel impregnated with AgNO3.</p><p>Plates for thin layer chromatography (TLC) were prepared by dissolving 25 g of CuSO4.5H2O in 100 ml of water and dipping the plate in the solution. TLC plates were dried and activated for 2 h at 120 °C. For column chromatography, 125 g of silica gel were added to a CuSO4.5H2O solution prepared as stated above, the water was evaporated at reduced pressure, and the CuSO4 -impregnated silica was activated for 2 h at 120 °C. Compound G was synthesized from compound A as described in Bartelt et al. (2011). The purity of both products A and G was >90 % according to the GC-MS total ion chromatogram (TIC).</p><!><p>Behavioral responses of P. striolata adults to the synthetic compounds A and G were determined in a two-choice experiment as described in Beran et al. (2011). Briefly, two traps containing the test compounds or pure solvent applied onto a piece of filter paper were placed in a plastic container with 50 beetles. After 24 h, the number of beetles in each trap was counted. Compound A (5 μg) diluted in hexane and compound G (5 μg) diluted in acetone and dispensed from filter paper were used per trap. To determine the emission of both compounds within 24 h from filter paper, volatile collections were carried out as described above. SuperQ filters were eluted with 200 μl solvent containing 10 ng/μl bromodecane as internal standard.</p><!><p>Attraction of beetles to synthetic compound A, synthetic compound G, and compounds A and G combined, were each compared to the corresponding solvent control.</p><!><p>Adults were given the choice between compound A and compound G, compound A and the blend of A + G, and compound G and the blend. For Experiments 1 and 2, each combination was replicated 30 times and the position of traps was alternated each time.</p><!><p>The interaction of the blend of A + G with AITC (Sigma-Aldrich, Munich, Germany) was investigated by comparing the attractiveness of traps baited with the blend alone and combined with different doses of AITC (10 ng; 100 ng; 1 μg; 10 μg; 100 μg; 1 mg). Each combination was tested 15 times.</p><!><p>A trapping experiment was carried out at AVRDC-The World Vegetable Center in Shanhua, Taiwan (23°07′04.9″N 120°17′42.1″E). Compounds were applied to dental cotton rolls as dispensers (Lohmann & Rauscher International GmbH & Co. KG, Rengsdorf, Germany) and placed in a wing trap (Jackson Traps, Jhen Yong Company, Taiwan). The following treatments were compared: 100 μg of compound A, 100 μg of AITC, 100 μg of compound A and 100 μg of compound G, 100 μg of compound A and 300 μg of compound G, 100 μg of compound A and 300 μg of compound G, and 100 μg of AITC. Dental rolls treated with pure solvent served as control. Traps were placed in a leafy radish field (variety Mei Lu Cai, Taipei Agricultural Products Marketing Co., Taiwan) at a height of 50 cm and arranged in a randomized complete block design with 6 replicates. Blocks were 10 m apart, and the distance between the traps in each block was 5 m. The experiment was performed from 17 to 20 April 2012 and repeated with new randomization in each block and new lures from 24 to 27 April 2012. Beetle numbers in each trap were counted at the end of each experiment. Since significantly more beetles responded to the combination of A and G compared to the individual compounds in laboratory experiments, and supply of synthetic compound G was limited, this compound was not tested individually in the field.</p><!><p>Trap count data from two-choice assays were analyzed by Wilcoxon matched-pairs signed-ranks test using the software SigmaPlot version 11. Field trap catches were transformed to log(n + 1) and analyzed by analysis of variance (ANOVA) and post-hoc Tukey's HSD test in SAS Version 9.1.</p><!><p>Male specific sesquiterpenoids detected in volatile collections of feeding Phyllotreta striolata</p><p>aMinor compounds D and I could not be quantified in samples; means were calculated from 10 (1d) and 9 (3d) volatile collections, respectively</p><p>bRelative abundance was calculated based on the peak area of single compounds compared to the total peak area of compounds</p><p>GC-MS analysis (total ion chromatogram) of volatiles collected from feeding male Phyllotreta striolata on SuperQ adsorbent. IS, internal standard; A, (6R,7S)-himachala-9,11-diene); B, α-himachalene; C, trans-α-himachalene; D, β-himachalene; E, γ-cadinene; F, (R)-ar-himachalene; G, (3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one; G*, thermal rearrangement product of compound G; I, (6R,7S)-2,2,6-trimethyl-10-methylene-bicyclo[5.4.0]-undec-1,11-ene</p><!><p>(6R,7S)-2,2,6-trimethyl-10-methylene-bicyclo[5.4.0]-undec-1,11-ene (compound I) also was detected by comparison with the mass spectrum and retention time of a reference compound provided by Dr. Allard Cossé using GC-MS. The amounts of compounds D and I were very low, and both compounds often co-eluted with contaminants so they could not be quantified. (1S,2R)-2,6,6-Trimethylbicyclo[5.4.0]undec-7-en-9-one (compound H) could not be detected in GC-MS analyses of any of the collections by comparison with a reference sample.</p><!><p>GC-MS analyses (total ion chromatograms) of volatile collections from feeding male Phyllotreta striolata using activated charcoal filters. Two example chromatograms are shown to demonstrate the low abundance of compound G. IS, internal standard; A, (6R,7S)-himachala-9,11-diene; B, α-himachalene; C, trans-α-himachalene; D, β-himachalene; E, γ-cadinene; F, (R)-ar-himachalene; G, (3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one; G*, thermal rearrangement product of compound G; I, (6R,7S)-2,2,6-trimethyl-10-methylene-bicyclo[5.4.0]-undec-1,11-ene</p><!><p>Allyl glucosinolate is the major glucosinolate of B. juncea leaves (Beran et al. 2014), which were used as food in volatile collections with P. striolata males. The amount of the glucosinolate hydrolysis product AITC in our samples corresponded to 4.4 ± 1.9 ng (N = 10, ±SD) per beetle per day. Whether AITC detected in the sample is released from the plant due to feeding damage or from the beetle cannot be determined in this system.</p><!><p>The amounts of compounds A and G emitted from filter paper over 24 h were 0.57 ± 0.11 μg and 1.49 ± 0.99 μg (N = 5; mean ± SD), respectively.</p><!><p>Behavioral responses of male and female Phyllotreta striolata to synthetic male-specific sesquiterpenoids in two-choice experiments. a Attractiveness of the individual compounds A ((6R,7S)-himachala-9,11-diene) and G ((3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one), and the blend of A and G combined compared to the corresponding solvent control. b Attractiveness of compound A versus compound G, and the individual compounds versus the blend. Significant differences between the treatments are indicated by * P < 0.05; ** P < 0.01; *** P < 0.001; n.s. P > 0.05 (N = 30; Wilcoxon matched-pairs signed-ranks test)</p><p>Effect of the glucosinolate hydrolysis product allyl isothiocyanate (AITC) on the attractiveness of the sesquiterpenoid blend to male and female Phyllotreta striolata adults in two-choice experiments. Different doses of AITC per trap were tested. Significant differences between the treatments are indicated by ** P < 0.01; n.s. P > 0.05 (N = 15; Wilcoxon matched-pairs signed-ranks test). A, (6R,7S)-himachala-9,11-diene; G, (3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one</p><!><p>The data from both trapping periods were combined because time (April 17–20 and 24–27) and the interaction treatment × time were not significant (Two-Way ANOVA, time: F = 0.08, P = 0.783; treatment × time: F = 0.72, P = 0.634).</p><!><p>Response of Phyllotreta striolata adults to synthetic male-specific sesquiterpenoids and allyl isothiocyanate (AITC) in a field trapping experiment in Taiwan in April 2012. Different letters indicate significant differences between treatments (N = 12, ANOVA followed by Tukey's HSD test; F = 8.71, P < 0.001). A, (6R,7S)-himachala-9,11-diene; G, (3S,9R,9aS)-3-hydroxy-3,5,5,9-tetramethyl-5,6,7,8,9,9a-hexahydro-1H-benzo[7]annulen-2(3H)-one; +, 100 μg per trap; +++, 300 μg per trap; −, not present</p><!><p>Previously, we identified six male-specific sesquiterpenoids (A-F) in volatiles from males of a Taiwanese P. striolata population collected on activated charcoal, and of these, compound A was the only physiologically and behaviorally active compound (Beran et al. 2011). Bartelt et al. (2011) identified three additional compounds (G, H, I) in headspace samples from an American P. striolata population collected on SuperQ adsorbent, and showed that compound G elicited electrophysiological responses from beetle antennae. In this study we detected compound G as the major male-specific sesquiterpenoid in the headspace of feeding Taiwanese male P. striolata collected on SuperQ. Activated charcoal is more retentive than SuperQ and can catalyze rearrangement and/or oxidation of sensitive compounds (Tholl et al. 2006), which most likely explains the low recovery of compound G from activated charcoal using the non-polar solvent n-hexane in our initial study.</p><p>While traces of compound I were present in volatile collections from males of our Taiwanese population, we could not detect compound H in any of our samples. On the other hand, compounds B and D were not identified in samples from the American P. striolata population (Bartelt et al. 2011). These results indicate that the male sesquiterpenoid blends produced by the genetically divergent Taiwanese and North American P. striolata populations (Beran 2011) differ qualitatively. However, the physiologically active compounds A and G appear to be conserved. It remains to be determined whether the blend of A and G is behaviorally active in the North American population as well.</p><p>Both laboratory assays and field trapping experiments demonstrated that the male-specific compounds A and G mediate attraction in P. striolata. In the field, a 1:3 blend of A and G was significantly more attractive than compound A combined with ecologically relevant amounts of AITC. Combining the pheromone mixture with AITC did not increase trap catches in the field, consistent with two-choice laboratory assays, thus demonstrating that the sesquiterpenoid blend could be combined with a wide range of AITC doses without being more attractive than the sesquiterpenoid blend alone. In fact, at the highest dose tested (1 mg AITC per trap), beetles significantly preferred the pheromone blend compared to the blend with AITC. In our earlier study, compound A, the only pheromone component identified at that time, required AITC at supranatural doses to elicit aggregation, thus indicating that other plant volatiles or additional pheromone components mediate aggregation behavior in the field (Beran et al. 2011). The results presented here clarify that the aggregation pheromone emitted by male P. striolata consists of at least two components (A and G), and is not reliant on the presence of AITC to attract adult P. striolata.</p><p>The behavioral response of insects to pheromones may be influenced or even depend on the presence of host plant volatiles (Landolt and Phillips 1997; Reddy and Guerrero 2004; Seybold et al. 2006), and this is likely also the case for the aggregation pheromone of Phyllotreta spp. Generally, only few Phyllotreta spp. were attracted to synthetic sesquiterpenoids alone in field trapping experiments, whereas high trap catches were achieved when these compounds were combined with ITCs (Soroka et al. 2005; Tóth et al. 2005). In a comparative study, P. vittula preferred the combination of pheromone component A and 3-butenyl ITC over the combination with AITC, while significantly more P. cruciferae were caught in traps baited with component A and AITC compared to the combination with 3-butenyl ITC (Tóth et al. 2012).</p><p>High amounts of ITCs are required to attract Phyllotreta spp. For example, AITC release rates of several mg per trap per day were necessary to attract high numbers of P. striolata and P. cruciferae in the field (Pivnick et al. 1992), but the amounts of AITC detected in volatile collections of P. striolata males feeding on leaves of a plant that contains rather high foliar allyl glucosinolate concentrations were relatively low. This was surprising at first but correlates well with P. striolata accumulating intact glucosinolates in their bodies (Beran et al. 2014). Apparently, feeding beetles may, at least to certain extent, avoid glucosinolate hydrolysis by the plant myrosinase. The addition of ecologically relevant amounts of AITC to the synthetic sesquiterpenoid blend had no influence on the behavioral response of P. striolata adults in laboratory or field experiments. Rather, excessive AITC doses led to a significant preference for the pheromone. This suggests that non-natural stimuli such as individual compounds may be attractive at non-natural concentrations, while ecologically relevant odorant blends may require close to natural ratios and release rates. Together these results indicate that AITC is not an essential cue for host plant location or aggregation because 1) allyl glucosinolate is not generally present in host plants, and 2) the emission rates from traps required for strong attraction of beetles in the field clearly exceed natural emissions rates from individual or small groups of plants (Najar-Rodriguez et al. 2015; Pivnick and Jarvis 1991).</p><p>In field experiments presented here and earlier (Beran et al. 2011), catch ratios ranged from about 1:10 to 1:4 for controls compared to the most attractive natural (beetles feeding on plants) and synthetic stimuli, respectively. Trap design and timing of experiments during different field seasons may have contributed to this variation. However, it also is possible that other plant volatiles interact with the pheromone and enhance the behavioral response of P. striolata adults to the pheromone. For example, the common monoterpenes (+)-sabinene and (E)-β-ocimene, and the green leaf alcohols 1-hexanol and (Z)-3-hexen-1-ol were shown to attract P. cruciferae adults in an olfactometer (Gruber et al. 2009). We found no attraction of P. striolata adults to 1-hexanol or (Z)-3-hexen-1-ol in field trapping experiments (Beran 2011), but their potential interaction with the synthetic two-component aggregation pheromone blend has not yet been assessed. Further research is required to establish more comprehensively the role of host plant volatiles in the aggregation behavior of P. striolata and other Phyllotreta species.</p>

PubMed Open Access

AKAP95 organizes a nuclear microdomain to control local cAMP for regulating nuclear PKA

Summary Contrary to the classic model of PKA residing outside of the nucleus, we identify a nuclear signaling complex that consists of AKAP95, PKA, and PDE4D5 and show that it forms a functional cAMP signaling microdomain. Locally generated cAMP can accumulate within the vicinity of this complex; but when cAMP is generated at the plasma membrane, PDE4 serves as a local sink and PDE3 as a barrier to prevent accumulation of cAMP within the microdomain as a means of controlling activation of tethered nuclear PKA.

akap95_organizes_a_nuclear_microdomain_to_control_local_camp_for_regulating_nuclear_pka

Introduction<!>AKAP95 forms an endogenous complex with PKA and PDE<!>The AKAP95 microdomain tightly regulates cAMP<!>PDE4 and PDE3 have distinct roles in regulating cAMP around AKAP95<!>Discussion<!>CONTACT FOR REAGENT AND RESOURCE SHARING<!>Cell Culture<!>Nuclear Fractionation<!>Co-Immunoprecipitation<!>Immunofluorescence Imaging<!>Proximity Ligation Assay<!>Computational Modeling<!>Cloning<!>Live-cell Imaging<!>FRAP Imaging<!>FRET Imaging<!>QUANTIFICATION AND STATISTICAL ANALYSIS

<p>Protein kinase A (PKA) is a ubiquitous serine/threonine kinase involved in regulating multiple cellular processes. The PKA holoenzyme is a tetramer composed of a specific regulatory subunit isoform (RIα, RIβ, RIIα, or RIIβ) homodimer and two catalytic subunits, where each catalytic subunit is bound to a regulatory subunit in the dimer. Activation of PKA occurs when cAMP binds the regulatory subunits, unleashing the active catalytic subunits. In the classic understanding of the pathway, PKA signaling in the nucleus was thought to occur by activation of extranuclear PKA and diffusion of the catalytic subunit into the nucleus. Recent studies raise some questions as to whether or not the catalytic subunit of the PKA holoenzyme completely dissociates under physiological conditions (Smith et al., 2013, 2017). In the absence of translocated catalytic subunit, nuclear PKA signaling would need to be activated and regulated through alternative mechanisms. Our previous work suggests that a pool of activatable PKA holoenzyme exists in the nucleus, which can be activated by local cAMP generated in the cytosol or nucleus (Sample et al., 2012). Consistent with our findings, a recent study showed that high cAMP production resulting from synergistic actions of two GPCRs at the endosome can lead to the activation of nuclear PKA (Jean-Alphonse et al., 2016). However, the molecular components that localize this pool of PKA to the nucleus and the mechanism by which nuclear PKA is controlled locally remain to be identified.</p><p>AKAP95, a PKA RII binding protein, has been shown to localize to the nucleus. It has been associated with multiple functions, including chromatin condensation, RNA processing, and gene expression, and recently its upregulation has been correlated with increased cyclin D1 and E1 expression in multiple types of cancer (Akileswaran et al., 2001; Arsenijevic et al., 2004, 2006; Coghlan et al., 1994; Collas et al., 1999; Eide et al., 2002, 2003; Jiang et al., 2013; Marstad et al., 2016; Zhao et al., 2015). We hypothesized that AKAP95 is a nuclear AKAP that anchors a specific pool of PKA holoenzyme within the nucleus.</p><!><p>To determine if AKAP95 forms a complex with PKA holoenzyme in the nucleus, we isolated nuclear proteins from HEK293T cells. These nuclear lysates endogenously contain AKAP95 (Figure 1a). The western blot results were further supported by immunofluorescence studies that detected AKAP95 in the nucleus (Figure 1b).</p><p>To better characterize AKAP95 in the nucleus we performed fluorescence recovery after photobleaching (FRAP) experiments using GFP-tagged AKAP95. On average, bleached areas showed a half recovery time of 25.7 ± 5.7 s (Supplemental Figure 1a) (N=11 cells). Based on these experiments, the majority of AKAP95 is mobile in the nucleus, with an average mobile fraction of 65.8 ± 11.2%. Interestingly, the observed AKAP95 puncta do not easily recover after photobleaching (Supplemental Figure 1b and 1c), suggesting that AKAP95 within these puncta is less mobile.</p><p>In addition to AKAP95, endogenous RIIα and the catalytic subunit of the PKA holoenzyme (PKAcat) were also detected in the nucleus with nuclear fractionation followed by western blotting (Figure 1a) as well as immunofluorescence (Figure 1b). For RIIα and PKAcat, the nuclear staining is significantly less than the cytosolic staining, but is clearly above background (Supplemental Figure 1d, e). Previous studies suggest PDEs, specifically PDE4D isoforms, are directly involved in regulating nuclear PKA activity (Sample et al., 2012). We found evidence of PDE4D, specifically isoform PDE4D5, in the nucleus (Figure 1a) where PDE4D5 appears to form puncta (Figure 1b).</p><p>To determine if the different components can form a complex in the nucleus, co-immunoprecipitation studies on the nuclear lysates were performed. We first pulled down endogenous AKAP95 and found that it associates with endogenous PKA RIIα, PKAcat, and PDE4D5 (Figure 2a), which was further confirmed via reciprocal co-immunoprecipitation experiments in which we isolated RIIα or PDE4D5 and probed for the remaining components (Figure 2a). The signals were detected above the background from normal IgG controls, indicating the specificity of these interactions.</p><p>Further support for the endogenous nuclear AKAP95 complex containing PKA and PDE4D comes from proximity ligation assay (PLA) experiments. In these experiments complexes containing AKAP95 and RIIα were detected in intact, fixed cells (Figure 2b). On average 1.65 ± 0.17 dots per nuclei (N=66 cells) were observed when measuring the AKAP95 and RIIα interaction, which is significantly (p<0.0001) above the background signal observed with AKAP95 (Figure 2b) or RIIα (Supplementary Figure 1f–g) antibody alone (0.40 ± 0.09 dots/nuclei, N=53 cells). Interestingly, although by immunofluorescence AKAP95 and RIIα are mainly detected in the nucleus and cytoplasm, respectively, PLA signal was observed in both the nucleus and cytoplasm (Supplemental Figure 1f–g). Therefore, although there are limited amounts of AKAP95 and RIIα in specific subcellular compartments, they are still interacting in a significant way. PDE4D5 and AKAP95 were also observed to interact in the PLA experiments with an average of 3.80 ± 0.33 dots per nuclei (N=79 cells), which is significantly above the background signal produced by AKAP95 antibody alone (p<0.0001). These results suggest that AKAP95 forms a complex with a resident pool of nuclear PKA, as well as PDE4D5.</p><!><p>The identification of AKAP95 as a specific nuclear scaffold protein that binds to PKA and PDE allows us to directly test the hypothesis that PDEs co-anchored to the same AKAP complex help regulate the activation of this PKA pool by reducing local cAMP levels. First, we employed our previously developed mathematical model (Sample et al., 2012) to generate predictions of the localized cAMP dynamics within the nuclear AKAP compartment. This ordinary differential equation model was developed to test hypotheses for differences in cAMP dynamics between the cytosol and nucleus. We extended the model to include a FRET-based cAMP biosensor, ICUE3 (DiPilato and Zhang, 2009), in the nuclear AKAP compartment and used the previously developed parameters and initial conditions to predict how cAMP dynamics will differ between the general nucleus and AKAP95 compartments (Figure 3a, b). This model predicts that when a modest amount of cAMP is generated at the plasma membrane, the nucleus experiences increased cAMP accumulation, but the nuclear AKAP compartment does not. Furthermore, the model predicts that higher cAMP production can overcome PDE-mediated inhibition in the AKAP compartment, such that cAMP accumulates in both the general nucleus and nuclear AKAP compartments.</p><p>To test these predictions experimentally, we targeted the FRET-based cAMP biosensor ICUE3 to AKAP95 to measure cAMP levels around AKAP95. This biosensor contains a cAMP binding domain sandwiched between a FRET pair and reports cAMP dynamics via changes in FRET. Targeting of this biosensor to AKAP95 was achieved using a chemically induced dimerization approach based on FK506 binding protein (FKBP) and FKBP-rapamycin binding domain (FRB), two proteins that bind specifically and tightly in the presence of rapamycin (Banaszynski et al., 2005). This chemically inducible targeting system was chosen to avoid concerns of expressing a large fusion protein in the nuclear compartment. Although not utilized here, this system could offer some flexibility for experimental designs to evaluate nuclear and AKAP95 compartments in the same cells before and after rapamycin. FKBP was fused to the N-terminus of nuclear-localized ICUE3 (FKBP-ICUE3-NLS), and FRB was fused to the C-terminus of human AKAP95 (AKAP95-FRB) (Figure 3c and Supplemental Figure 2a). When HEK293T cells co-expressing FKBP-ICUE3-NLS and AKAP95-FRB were treated with rapamycin (100 nM), the diffuse pattern of reporter fluorescence became more punctate within the nucleus within minutes, indicating the translocation and recruitment of ICUE3 to AKAP95 (Supplemental Figure 2b). HEK293T cells co-expressing these same two constructs were treated with rapamycin or DMSO (control) and harvested for co-immunoprecipitation experiments. Immunoprecipitation using anti-AKAP95 antibody followed by western blotting for GFP revealed that ICUE3 interacted with AKAP95 in cells treated with rapamycin, confirming the localization of the reporter to AKAP95 (Supplemental Figure 4c). To ensure that targeting ICUE3-NLS to AKAP95 did not affect the dynamic range of the reporter, the maximum responses of cells expressing either FKBP-ICUE3-NLS alone (N=39 cells) or both FKBPICUE3-NLS and AKAP95-FRB (N=35 cells) were compared (Supplemental Figure 4d). In both cases, cells were treated with rapamycin, and the maximum response was stimulated by the addition of forskolin and IBMX to activate adenylyl cyclases and inhibit PDEs, respectively. After rapamycin treatment, the biosensor responded similarly to maximal cAMP stimulation regardless of AKAP95-FRB co-expression, indicating that the dynamic range of the biosensor is not affected by its localization to AKAP95.</p><p>To determine what effect the site of cAMP generation has on the cAMP levels within the AKAP95 compartment, we combined live-cell imaging using the localized cAMP biosensor with spatiotemporal manipulation of cAMP using soluble adenylyl cyclase (sAC), or SMICUS, which is a method previously developed to generate local pools of cAMP using subcellularly targeted sAC (Sample et al., 2012). sAC is pharmacologically distinct from the transmembrane adenylyl cyclase (tmAC). Instead of being activated by GPCRs, it is activated by bicarbonate, ATP, or calcium (Zippin et al., 2013). We can therefore use exogenously expressed mCherry-tagged sAC as a tool to generate cAMP in specific subcellular locations in a dose-dependent manner in HEK293T cells by the addition of sodium bicarbonate to the culture medium (Sample et al., 2012). To this end, mCherry-tagged sAC was localized to either the plasma membrane (PM-sAC) or the nucleus (sAC-NLS) (Figure 3c). PM-sAC generates cAMP at the plasma membrane, mimicking the production of cAMP by tmACs. HEK293T cells expressing PM-sAC and either FKBP-ICUE3-NLS alone for general nuclear targeting, or FKBP-ICUE3-NLS plus AKAP95-FRB for AKAP95 targeting, were first treated with rapamycin (100 nM) (Figure 3d). Cells were then treated with a low dose (2.5 mM) of sodium bicarbonate to activate targeted sAC. Although cAMP production at the plasma membrane by sub-maximal sAC activation led to a clear response in the general nuclear compartment within 15 min, with an average emission ratio change of 11.3 ± 1.5% (N=39 cells), cAMP was not detected within the AKAP95 microdomain (mean emission ratio change −1.9 ± 0.8%, N=35 cells), suggesting that cAMP accumulation is tightly regulated in the area surrounding the AKAP95 complex, which represents a functional microdomain with distinct cAMP dynamics. Subsequent treatment with a saturating dose (15 mM) of bicarbonate to maximally activate sAC resulted in increased FRET responses from both nuclear-localized and AKAP95-targeted ICUE3 (Figure 3d). Hence, although some cAMP generated at the plasma membrane is able to diffuse into the nucleus, an added level of regulation limits the accumulation of cAMP within the AKAP95 microdomain. While there are differences in kinetics, these data qualitatively agree with the model predictions of an ablated cAMP response to the low stimulation condition in the AKAP95 compartment but a response to the high stimulation condition.</p><p>Next, the cAMP responses within the AKAP95 compartment were compared when cAMP was generated at the plasma membrane or within the nucleus. When cAMP was generated in the nuclei of cells expressing sAC-NLS, even a low dose of sodium bicarbonate resulted in clear cAMP responses of 4.9 ± 0.8% within the AKAP95 microdomain (N=18 cells) compared to the lack of a response within the microdomain when cAMP was generated at the plasma membrane (−0.9 ± 0.8%, N=18 cells) (Figure 3e). Maximal activation of sAC in the nucleus or at the plasma membrane further increased the cAMP levels in the AKAP95 microdomain.</p><p>These findings further support the existence of a functional microdomain that is distinct from the general nuclear environment in the immediate vicinity of AKAP95, where cAMP accumulation is limited unless cAMP is generated more locally or in excess levels from the plasma membrane. These results support the hypothesis that AKAP95 is able to create a distinct microdomain that limits cAMP concentrations, thereby providing a way to tune the activation threshold of the AKAP-anchored pool of PKA.</p><!><p>To test the role of different PDE classes in regulating the microdomain, we used inhibitors specific to either PDE4 or PDE3. HEK293T cells expressing PM-sAC and rapamycin-induced AKAP95-targeted ICUE3 were treated with rolipram (PDE4 inhibitor), milrinone (PDE3 inhibitor) or vehicle control. Cells were subsequently treated with a low dose (2.5 mM), followed by a high dose (15 mM), of sodium bicarbonate to activate PM-sAC.</p><p>The initial response after the addition of PDE inhibitor indicates the basal level of PDE activity at the microdomain. As shown in Figure 4a and b, cells treated with the PDE4 inhibitor rolipram (1 μM) exhibited an immediate increase in cAMP levels in the AKAP95 microdomain (9.4 ± 1.9% emission ratio change, N=29 cells), suggesting that co-anchored PDE4 is basally active in restricting cAMP levels within the AKAP95 microdomain. Subsequent addition of low-dose sodium bicarbonate following PDE4 inhibition further increased cAMP levels in the microdomain (10.0 ± 0.8%, N=29 cells), in sharp contrast to the limited response (4.9 ± 0.8%, N=39 cells) in the absence of any inhibitor (Figure 4c). Furthermore, a dominant-negative mutant of PDE4D5 (dnPDE4D5) was used as an alternative to global PDE4 inhibition by rolipram to more specifically investigate the role of PDE4D isoforms at the microdomain. This dominant-negative mutant is inactive and does not directly inhibit endogenous PDEs, but instead competes with and displaces them from binding to scaffolding proteins such as AKAP95 (Terrin et al., 2006). We can therefore specifically interrogate the role of the scaffolded PDE4D5 in regulating the AKAP95 microdomain. HEK293T cells expressing AKAP95-targeted ICUE3, PM-sAC, and a dominant-negative mutant of PDE4D5 were treated with a low dose followed by a high dose of sodium bicarbonate. Cells expressing the mutant PDE4D5 exhibited a robust cAMP response to submaximal activation of PM-sAC (16.5 ± 1.9%, N=24 cells) compared with control cells that were not transfected with dominant-negative PDE4D5 (−1.9 ± 0.8%, N=35 cells) (Figure 4d). These data suggest that basally active PDE4 restricts cAMP accumulation in the AKAP95 microdomain under both basal state and stimulated conditions.</p><p>In contrast, addition of the PDE3 inhibitor milrinone (10 μM) induced no response within the microdomain (2.0 ± 1.3%, N=19 cells) (Figure 4a–b), suggesting PDE3 does not play a role in controlling the basal level of cAMP within the AKAP95 microdomain. Interestingly, when PDE3 is inhibited with milrinone, submaximal activation of PM-sAC led to increased cAMP levels in the microdomain (13.4 ± 1.8%, N=19 cells) (Figure 4c), indicating that PDE3 plays a role in preventing low levels of plasma membrane-generated cAMP from accessing the microdomain.</p><!><p>Here we presented the identification of a nuclear signaling complex consisting of AKAP95, PKA, and PDE4D5. AKAP95 has previously been found to bind RIIα during mitosis after the nuclear envelope has dissolved (Collas et al., 1999). Using both nuclear fractionation studies and whole-cell immunofluorescence experiments we demonstrate that nuclear AKAP95 interacts with the regulatory subunit of PKA during interphase. These findings suggest that AKAP95 can anchor PKA in the nucleus irrespective of cell cycle stages. Previous studies have found that AKAP95 can associate with PDE4A in T-lymphocytes, but the functional role of anchored PDE4A was not tested (Asirvatham et al., 2004). Here we show that anchored PDE4D5 plays a critical role in the functional cAMP signaling microdomain assembled by AKAP95.</p><p>The roles played by different PDE isoforms are complex. Although it is widely accepted that cAMP diffusion is controlled and limited, the mechanisms underlying this compartmentalization are debated. One model for cAMP regulation suggests the existence of a barrier to prevent cAMP diffusion away from a particular location (e.g., the plasma membrane). An alternative model posits that PDEs concentrate in certain locations to prevent cAMP accumulation at those locations. Our results suggest these two mechanisms could co-exist in a PDE isoform-specific manner. PDE4 isoforms are part of a complex with AKAP95 and create a local sink to limit the accumulation of cAMP within the AKAP95 microdomain. Our data suggest that PDE3 is not a part of the AKAP95 complex but could act as a gate to prevent free diffusion of cAMP from the plasma membrane, the specific mechanisms and functional roles of which will be further investigated. The intricate interplay between these two different PDE isoforms and their effects on local signaling expands the concept of localized cAMP compartmentalization control to include proximate and distal PDE isoforms working in concert. cAMP is thus tightly regulated within this microdomain to prevent accidental nuclear PKA activation and reserve this pool of scaffolded PKA holoenzyme for transducing signals that generate specific, local sources of cAMP.</p><p>What are the specific signals that can lead to the activation of this pool of nuclear PKA holoenzymes? Endogenous sAC has been shown to localize to subcellular locations such as the cytosol, nucleus, and mitochondria and respond to changes in bicarbonate, calcium, and ATP. sAC has been shown to act as a sensor for changes in pH and metabolism in different cell types (Tresguerres et al., 2010, 2011, Zippin et al., 2003, 2010, 2013) and could generate cAMP proximal to this pool of nuclear PKA. In addition, it has been demonstrated that endocytosed GPCRs continue signaling at the endosome and that this endosomal signaling elicits distinct transcriptional responses (Irannejad et al., 2013; Tsvetanova and von Zastrow, 2014). cAMP generated at the endosome could provide a local source of cAMP that activates the nuclear PKA pool. Supporting this hypothesis, Jean-Alphonse et al (Jean-Alphonse et al., 2016) recently showed that a synergistic effect between PTHR and β2-AR signaling induces pronounced endosomal cAMP production that corresponds with activation of nuclear PKA and increased levels of phosphorylated CREB, a PKA substrate. Given that GPCR internalization and endosomal cAMP production vary by receptor and ligand, the strength and extent of nuclear PKA activation by internalized GPCRs may be receptor specific. The tight regulation of cAMP, and by extension nuclear PKA, within the AKAP95 microdomain shown here provides a new example of signaling specificity arising from precise spatiotemporal control.</p><!><p>Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Jin Zhang (jzhang32@ucsd.edu).</p><!><p>HEK293T cells (human embryonic kidney, female) were maintained in DMEM (10% (v/v) FBS and 1% (v/v) penicillin-streptomycin at 37°C with 5% CO2.</p><!><p>Sucrose-gradient centrifugation was used to isolate and collect nuclei. All steps were performed on ice. HEK293T cells were collected in DPBS and pelleted at 470 × g for 10 min. Gentle vortexing was used to resuspend the cell pellet in ice-cold Sucrose Buffer I (0.32 M), and cells were lysed by 15 strokes of a Dounce homogenizer B pestle. Lysates were gently mixed with Sucrose Buffer II (2.0 M) in a 50 mL tube. The lysate-sucrose mix was layered on top of Sucrose Buffer II in a polyallomer SW 40.1 ultracentrifuge tube. Any space remaining in the tube was filled with Sucrose Buffer I. A Beckman centrifuge with a SW41 Ti rotor was used to separate the nuclei at 30,000 rpm for 45 min, at 4°C. Whole-cell, non-nuclear, and nuclear fractions were probed with antibodies against CREB (Cell Signaling 9197S) and β-tubulin (Cell Signaling 2146S) to ensure clean fractionation. Antibodies used on fractionation samples: AKAP95 (R-146) (Santa Cruz, sc-10766), RIIα (H-12) (Santa Cruz, sc-137220), PKAcat (A-2) (Santa Cruz, sc-28315), PDE3B (H-300) (Santa Cruz, sc-20793), and PDE4D5 (gift from Dr. George Baillie).</p><!><p>Protein A/G PLUS-Agarose beads (Santa Cruz sc-2003) or Protein A/G Magnetic Beads (bimake.com B23202) were equilibrated in Tris buffer. Pre-clearing was achieved by incubation of agarose beads with 1 mg of nuclear protein and 5 μg rabbit IgG (Santa Cruz sc-2027) for 1 h at 4°C to remove proteins that nonspecifically bind the beads or IgG. Pre-cleared lysate was incubated with the immunoprecipitating antibody for 1 h at 4°C. The lysate-antibody mix was incubated with agarose beads overnight. After 4 washes with ice-cold DPBS, bound protein was eluted by boiling for 5 min in SDS sample buffer. Antibodies used for Co-IP: PKA IIα reg (C-20) (sc-908), PDE4D5 (gift from Dr. George Baillie's Lab), AKAP95 (R-146) (sc-10766), Clean Blot (Thermo Fisher Scientific 21230).</p><!><p>Cells were rinsed in HBSS and then fixed in 4% PFA for 30 min, followed by permeabilization for 1 h at room temperature. Cells were then incubated overnight at 4°C or for 30 min at 37°C with primary antibody. Alexa Fluor 488-conjugated and Alexa Fluor 568-conjugated secondary antibodies (Thermo Fisher Scientific A-11034, A-11001, A-11011, or A-11004) were incubated at room temperature for 1 h or 30 min at 37°C in the dark. Cells were incubated with Hoescht 33342 stain at room temperature for 30 min before imaging. Cells were imaged on a Zeiss LSM 880 Confocal with Airyscan processing using a 40x/1.2NA objective. Images were acquired with the suggested settings using 405 nm, 488 nm, and 561 nm lasers. Antibodies used for IF: AKAP95 (F-11) (Santa Cruz, sc-390335) or AKAP95 (R-146) (Santa Cruz, sc-10766), RIIα (C-20) (Santa Cruz, sc-908), PKAcat (BDBiosciences Clone 5B cat#610980), PDE4D3 (gift from Dr. George Baillie), and PDE4D5 (gift from Dr. George Baillie).</p><!><p>PLA experiments were performed using the Duolink® in situ red starter kit for proximity ligation assays (Sigma Aldrich, DU092101) according to the provided protocol. The only protocol modification was to extend the amplification time by 50 min. Briefly, cells were fixed and permeabilized as in the immunofluorescence experiments before incubation with primary antibody (same used for IF experiments), then with the provided secondary antibody (conjugated with nucleotides) for 30 min at 37°C each with washes after each step. Ligation of the nucleotides and amplification of the strand occurred sequentially by incubating cells with first ligase then polymerase and detection solution. PLA experiments with AKAP95 antibodies from different species were used as positive controls, and experiments withjust one primary antibody (AKAP95) provided our negative control. Images were acquired on a Zeiss LSM 880 Airyscan confocal as described above. A cross section of the nucleus (3.6–5 μm) was acquired and the number of dots within that cross section counted both inside and outside the nucleus.</p><!><p>A previously developed computational model of compartmentalized cAMP and PKA dynamics in HEK293 cells (Sample et al., 2012) was modified to include ICUE in the nuclear AKAP compartment and evaluated in MATLAB (Mathworks). Briefly, the "nucAKAP Model" is a compartmental ordinary differential equation (ODE) model that consists of plasma membrane, cytosol, nucleus and nuclear AKAP compartments. This model consists of equations describing cAMP production by adenylyl cyclases (both endogenous and over-expressed sAC), cAMP degradation by PDEs, and cAMP activation of PKA. Additionally, the cAMP biosensor ICUE and PKA biosensor AKAR were included in all the compartments except for the nuclear AKAP compartment. We extended this "nucAKAP Model" to include ICUE in the nuclear AKAP compartment by making the following modifications.</p><p>A new second-order reaction of cAMP binding to ICUE in the nuclear AKAP compartment: dICUEcAKAPdt=kf,ICUE⋅cAMPAKAP⋅ICUEAKAP−kr,ICUE⋅ICUEcAKAP</p><p>And an additional conservation of mass equation for total ICUE in the nuclear AKAP compartment: ICUEAKAP=ICUEtotAKAP−ICUEAKAP</p><p>And fmally, the conservation of mass for free cAMP in the nuclear AKAP compartment was updated to account for the cAMP that is bound to ICUE.</p><p>No new kinetic parameters were introduced, as the rates of cAMP association and dissociation with ICUE are assumed to be consistent across compartments. Total ICUE in the AKAP compartment, ICUEtotAKAP, was assumed to be equal to the total amount ofPKA in the AKAP compartment, which assumes that each AKAP has both PKA and ICUE bound to it. All other equations and parameters are as described in the supplement of Sample et al 2012 (Sample et al., 2012).</p><p>This model was run using the initial conditions from the published model and allowed to equilibrate by running for 109 s. Using the simulated plasma membrane-targeted sAC built into the model (EsAC pm basal = 1.663 · 10−1 mM/s), cAMP production was stimulated with a low dose ofNaHCO3 (increase EsAC,pm,basal by 1×1.726 · 10−1) for 15 min and then subjected to a high dose ofNaHCO3 (increase EsAC,pm,basal by 6×1.726 · 10−1) for another 15 min.</p><!><p>Full-length human AKAP95 was kindly provided by Dr. John D. Scott (University of Washington). FKBP and FRB constructs were kindly provided by Dr. Takanari Inoue (Johns Hopkins University). FRB or GFP were tagged to the C-terminus of AKAP95 using BamHI and EcoRI or NotI restriction sites. FKBP with a flexible linker was added to the N-terminus of ICUE3 (DiPilato and Zhang, 2009) using HindIII and BamHI restriction sites.</p><!><p>HEK293T cells were transfected with Lipofectamine 2000 (Invitrogen) when cells were 50–60% confluent and imaged 16–24 h later. Cells were washed twice with and maintained in Hank's balanced salt solution buffer (HBSS). Cells were imaged at room temperature.</p><!><p>FRAP experiments were performed on a Leica TCS SP8 confocal microscope using its associated FRAP software, a 63x/1.4NA objective, and a DD488/552 beamsplitter. A total of 5 images were acquired before bleaching a rectangular cross-section or circular area of the nucleus with 25% laser power. Post-bleaching images were acquired every 5 or 10 s for at least 5 min. Images were analyzed using the Jython script found at: http://imagej.net/Analyze_FRAP_movies_with_a_Jython_script.</p><!><p>Prior to imaging, cells were equilibrated in HBSS for 10 min in a CO2-independent incubator. Cells were treated with NaHCO3 (J.T. Baker, purity=100%), which induced a small change in FRET for the FKBP-tagged biosensor with or without co-expression of AKAP95-FRB (seen in Figure 3), forskolin (Calbiochem, purity ≥99% by HPLC), IBMX (Sigma, purity ≥98% by TLC), rolipram (Alexis, purity ≥98% by TLC), or milrinone (Alexis, purity ≥97% by TLC) as indicated. Cells were imaged on a Zeiss Axio Observer Z1 microscope equipped with a 40x/1.3 NA objective and Photometrics Evolve 512 EMCCD, using METAFLUOR 7.7 software (Molecular Devices) to control dual-emission-ratio imaging acquisition every 30 s. A 420DF20 excitation filter, a 450DRLP dichroic mirror, and two emission filters (475DF40 for CFP and 535DF25 for YFP) alternated by a Lambda 10–2 filter-changer (Sutter instruments) were used for the FRET measurements. The microscope settings for RFP (sAC) fluorescence acquisition were as follows: 555DF25 excitation filter, 568rdc dichroic mirror, and 650DF100 emission filter. Exposure times for all channels ranged from 50–500 ms. Emission intensities of individual cells were background-subtracted, and the ratio between the CFP and FRET channel was normalized to the timepoint just before addition of low-dose NaHCO3 (t = 0 min). Because overexpressed sAC has some basal activity prior to stimulation, the starting FRET ratio for AKAP95-localized ICUE3 was higher in cells expressing sAC-NLS compared to PM-sAC. Therefore, a moderate range of overlap in starting ratios between these two experiments was determined, and cells were selected for comparison based on starting ratio within this range to control for basal activity differences. On the other hand, the expression of PM-sAC was controlled to be within a defined range, and this range was held constant between different experimental conditions; cells outside this range were excluded from analysis.</p><!><p>2–3 biological replicates were done for Co-IP and PLA experiments, and at least 3 biological replicates performed for all other experiments. Unpaired two-tailed t-tests with Welch's correction were used for statistical analyses, done in Graphpad Prism. *** indicates a p-value<0.0001. N values, as indicated in figure legends and the main text, represent number of cells. All error bars indicate standard error.</p>

PubMed Author Manuscript

Monodisperse Pt-Co/GO anodes with varying Pt: Co ratios as highly active and stable electrocatalysts for methanol electrooxidation reaction

electrocatalysts for the methanol electro-oxidation reaction as a direct anodic reaction of methanol fuel cells 29,30 . Therefore, various studies have been conducted to investigate bimetallic catalysts such as PtSn 31,32 , PtPb 33 , PtNi 34 , PtAu 35,36 , PtMo 37 . Catalytic performance and utilization efficiency may vary greatly depending on the performance of the platinum-based electrocatalyst in the fuel cell depending upon the type of support and metals [38][39][40] . Supporting and/or stabilizing agents are also very important materials in catalytic systems. They result in increasing dispersity, electron transfer, long-term stability and transport of materials in fuel cell electrodes. Nowadays, many scientists have been much interested in graphene and graphene-based supports [41][42][43] . In this context, various studies have been made by using Platinum-based NPs with graphene due to the very good electrical conduction and good performance of graphene [44][45][46][47] . Because graphene oxide (GO) and graphene-based materials are highly dispersible in H 2 O and a few organic solvents [48][49][50][51][52][53] . Over the years many PtCo catalysts with various ratios and morphology has been reported for catalyzing the electrooxidation of methanol [54][55][56] . For this purpose, in our laboratory, many studies have been carried out in order to increase the efficiency of the catalyst with the help of the addition of second metal and using of different supporting agent 53,57 . In this context, herein, GO-supported platinum-cobalt NPs were synthesized in various ratios by using the double solvent reduction method in this work. All prepared catalysts have been characterized by XRD, XPS, TEM, HR-TEM etc. in order to reveal the morphology and structure of the prepared catalysts. Further, they have been tested for their electrocatalytic efficiency towards methanol oxidation reactions.

monodisperse_pt-co/go_anodes_with_varying_pt:_co_ratios_as_highly_active_and_stable_electrocatalysts

Monodisperse pt-co/Go anodes with varying pt: co ratios as highly active and stable electrocatalysts for methanol electrooxidation reaction Hakan Burhan, Hasan Ay, esra Kuyuldar & fatih Sen *<!>Results and Discussion<!>conclusions

<p>The intense demand for alternative energy has led to efforts to find highly efficient and stable electrocatalysts for the methanol oxidation reaction. for this purpose, herein, graphene oxide-based platinum-cobalt nanoparticles (pt 100−x co x @GO NPs) were synthesized in different ratios and the synthesized nanoparticles were used directly as an efficient electrocatalyst for methanol oxidation reaction (MoR). the characterizations for the determination of particle size and surface composition of nanoparticles were performed by transmission electron microscopy (TEM), X-ray diffraction (XRD) and X-ray photoelectron spectroscopy (XpS). the structure of the catalysts was detected as facecentered cubic and the dispersion of them on graphene oxide was homogenous (distributed narrowly (4.01 ± 0.51 nm)). Cyclic voltammetry (CV) and chronoamperometry (CA) was utilized for testing electrocatalytic activities of all prepared nps for the methanol oxidation reaction. it was detected that the newly produced NPs were more active and stable than commercially existing Pt(0)/Co nanomaterial in methanol electro-oxidation in acidic media.</p><p>The researches for environmentally friendly and reusable power have been enhancing due to the gradually increasing pollution issues of the earth. Fuel cells, which are one of the alternative energy sources, are electrochemical cells that convert chemical energy directly into electrical energy with high conversion efficiency and low environmental pollution. In recent years, several studies have been conducted directly on alcohol fuel cells [1][2][3][4][5] . Direct alcohol fuel cells (DAFCs) are used in portable electronic devices because they are more ergonomic as energy sources. DAFCs have become popular among fuel cells because of their easy supply of alcohol derivatives, high energy densities, and relatively easy storage and transport processes. Among the direct alcohol fuel cells, direct methanol fuel cells (DMFCs) are among the most widely studied alcohol fuel types. Because methanol has a structurally simple, easily available and promising electrochemical activity. Due to these properties, many studies have been carried out on electrooxidation [6][7][8] . Besides, methanol has many advantageous over pure hydrogen as fuel, such as transportability, storage, cost-effectiveness and high theoretical energy density [9][10][11] . As in all fuel cells, the reaction mechanisms of alcohol with a catalyst that affect the efficiency of these mechanisms are one of the most important parameters in alcohol fuel cells. So far, many studies have been carried out on these catalysts [12][13][14] . Some of the main problems in the application of DMFCs can be listed as follows: Poisoning caused by adsorption of CO and similar intermediates formed during dehydrogenation of the methanol by catalysts (eg platinum and platinum-based catalysts). This poisoning reduces the efficiency of methanol oxidation kinetics by inhibiting the operation of active sites of catalysts. Another problem is that the platinum catalysts used are very expensive to be commercially available, as the fuel cells show superior performance in the anode section. In order to make it commercially viable, platinum was modified with different metals and more economical alloys or mixtures [15][16][17][18][19][20][21][22][23][24][25][26] . The catalytic activity and stability of the present electrocatalysts are not efficient enough to directly commercialize methanol fuel cells and make them widely available 7,27,28 . Within these catalysts, Pt+second metal bimetallic nanoparticles bonded onto the graphene oxide support were considered to be one of the most attractive experimental</p><p>The chemicals, instruments and the experimental details are given in supporting information. Graphene powder from graphite powder (GO) was synthesized using the Hummer's method 58,59 . All prepared Pt 100−x Co x @GO NPs were synthesized by a double solvent reduction method. In shortly, for the preparation of NPs with a various atomic ratio of Pt and Co (1:0, 1:1, 1:3, 3:1 ratio), the calculated amount of PtCl 4 and CoCl 2 as precursor materials were dissolved in tetrahydrofuran to be able to prepare the NPs and then. The prepared GO was added to the medium as stabilizing and supporting agents with a 1:1 ratio of platinum-cobalt nanoparticles. Ethanol and super hydride (Li(C 2 H 5 ) 3 BH) were added up to the complete reduction of Pt and Co metals. The formation of Pt 100−x Co x NPs is understood by the observation of brown-black color in solution. Finally, the resulting solid Pt 100−x Co x NPs were dried under a vacuum.</p><!><p>All prepared catalysts were characterized by XRD, XPS, TEM, and RAMAN spectroscopy methods. XRD (X-ray diffraction) was used to determine the size of the crystallites and the crystalline structure of the prepared catalysts. In all XRD patterns, Pt ( 111 220) and (311), respectively (Fig. 1a) (JCPDS 87-0646). Specifically, the Pt (111) peak shifts from 39.9 o (pure Pt) to 40.13 o , 40.21 o , and 40.30 o , for Pt 75 Co 25 @ GO, Pt 50 Co 50 @GO, and Pt 25 Co 75 @GO, respectively. This case indicates the alloy formation in all prepared bimetallic nanoparticles.</p><p>The characteristic plane of Pt (111) shows the crystalline structure of the nanoparticle and a shift of 2θ degree occurs with increasing cobalt concentration. This can be explained by the formation of cage shrinkage due to the integration of cobalt which is a smaller atom than Pt. The average crystallite size was calculated for Pt 100−x Co x @ GO using Scherrer formula.</p><p>where; k = a coefficient (0.9) λ = the wavelength of X-ray used (1.54056 Ǻ) β = the full-width half-maximum of respective diffraction peak (rad) θ = the angle at the position of peak maximum (rad) A decrease in the lattice parameter is observed as a result of the increase in Co concentration in the PtCo nanocatalyst (Table 1). This is considered to be the case that Co fills the internal spaces between Pt atoms. Raman spectroscopy was also used to visualize the changes after PtCo addition on the GO surface. Figure 1b shows the values of the D and G bands related to GO, Pt@GO and PtCo@GO. As shown in this figure, here, the G band shows the E 2g construction mode of the carbon atoms bound to sp 2 , while the D band shows the A 1g breathing mode of an irregular graphite structure. The ratios of D-and G-band (I D /I G ) intensities for GO, Pt@GO, and Pt 75 Co 25 @GO were 1.10, 1.25, and 1.32, respectively. The increasing ratio of D/G band means the functionalization and/or increasing irregularity of the graphene oxide surface after the addition of PtCo NPs. The size, composition, and morphology of Pt 75 Co 25 @GO NPs are shown in Fig. 2 as a model catalyst. The morphology of the prepared catalyst is also shown in Fig. 2b with a high-resolution electron micrograph (HRTEM). From the HRTEM image, it can be said that the particles are generally spherical and do not agglomerate in the synthesized catalyst. Furthermore, the atomic lattice fringes are seen by the HRTEM image of the monodisperse Pt 75 Co 25 NPs as shown in Fig. 2b. As a result of these fringes, a Pt (111) plane was observed on the prepared catalyst in a range of 0.22 nm; which is similar to a nominal Pt (111) range of 0.23 nm. In addition, the mean particle size of Pt 75 Co 25 @GO NPs was found to be 4.01 ± 0.51 nm (Fig. 2c) which is in good agreement with XRD results. Further, TEM-EELS mapping of Pt 75 Co 25 @GO NPs was also performed in order to see the structure of the catalyst and it's seen that Pt and Co co-exist together which also confirms the alloy formation of prepared catalyst. In addition, the formation of an alloy composition with uniformly distributed platinum and cobalt throughout the entire nanoparticle was shown in this figure . The surface composition and chemical oxidation states of Pt and Co in monodisperse Pt 100−x Co x @GO NPs were investigated using X-ray photoelectron spectroscopy (XPS). As a result of this, the Pt 4 f and Co 2p regions of the spectrum were evaluated by the Gaussian-Lorentzian method and the relative density of the species was estimated by calculating the integral of each peak after Shirley background subtraction. The correct binding energies (± 0.3 eV) in the shaped background XPS spectrum were determined with reference to the C1s peak at 284.5 eV (Fig. S1). As shown in Fig. 3, XPS spectra show that the surface Pt and Co are found to be mostly metallic and a small amount of oxides (Fig. 3a,b). Though platinum is mainly metallic form in Pt 100−x Co x @GO, from the images, the presence of PtO and PtO 2 , indicative of the oxidation of surface was understood from the existence of 2+ and 4+ species. Table S1 represents BEs of the 4f 7/2 data for Pt 100−x Co x @GO and Pt@GO and their comparative densities. It is illustrated in Table S1 that the highest amount of platinum (0) is shown in Pt 75 Co 25 @ GO compared to all other prepared ones. From Table S1 and Fig. 3a,b, binding energy (for 4f 7/2 peak) of platinum cobalt/graphene-oxide nanomaterials is 0.1-0.2 eV higher comparing to bulk platinum ones. The cause of positive change is the interaction between the final state of relaxation and platinum/cobalt-graphene-oxide. Table S1 also shows the relative intensities of metallic species in all prepared catalysts and the higher platinum (0) content (83.1%) was shown in Pt 75 Co 25 @GO compared to the other prepared Pt 100−x Co x @GO. When Co 2p peaks are examined, it's seen that cobalt is in mostly zero oxidation state at about 780 eV and in a small amount of oxidized species at about 786 eV.</p><p>After full characterization of prepared catalysts, the electrocatalytic performance of Pt 100−x Co x @GO was employed towards methanol oxidation reaction. For this purpose, the prepared electrodes with the help of nanomaterials were dipped into 0.5 M sulfuric acid in order to prepare electrocatalyst for the measurements. To obtain Table 1. The comparison of particle size obtained from (a) XRD, (b) TEM.</p><p>consistent results in cyclic voltammetry (CV), electrodes were processed among −0.2 and 0.8 V at 50 mV s −1 . The cyclic voltammograms of all prepared Pt 25 Co 75 @GO, Pt 50 Co 50 @GO, Pt 100 Co 0 @GO and Pt 75 Co 25 @GO electrodes are illustrated in Fig. 4a. A strong oxidation peak is observed in all prepared catalysts. It is also seen from Fig. 4a that the methanol oxidation peak is situated at about 0.38 V for Pt 75 Co 25 @GO. As shown in Fig. 4a, the best catalytic performance was seen in Pt 75 Co 25 @GO electrodes which have 1.27, 1.44, 1.54 and 2.94 higher performance than Pt 100 Co 0 @GO, Pt 50 Co 50 @GO, PtRu (20%) E-TEK, and Pt 25 Co 75 @GO catalysts, respectively. The prepared best catalyst have also been compared with PtRu (ETEK), PtCo@C, PtCo@GC and it was seen that Pt 75 Co 25 @GO showed better catalytic performance compared to the others as shown in Fig. S2 and S3. www.nature.com/scientificreports www.nature.com/scientificreports/ In order to explain the higher performance of Pt 75 Co 25 @GO electrodes, electrochemical surface area (ECSA), chemical surface area (CSA) and metal utility % (ECSA/CSA) were calculated as shown in Table S2. In this table, the comparison of crystalline particle size, ECSA, CSA and metal utilization (%) for the prepared catalysts are given in detail 60 . These data show that Pt 75 Co 25 @GO NPs have the highest metal utility (89.38%) compared to the other prepared ones. This case explains very well the higher performance of Pt 75 Co 25 @GO NPs than the other prepared ones. In addition, higher methanol oxidation reaction performance of Pt 75 Co 25 @GO NPs can be explained by several ways correlating with XPS, TEM, etc. X-ray photoelectron spectroscopy data in Table S1 reveals that platinum is in the more metallic state in Pt 75 Co 25 @GO NPs (83.1%) than the other prepared ones and consequently causes a more effective activation of adsorbed CH 3 OH. The higher performance of Pt 100−x Co x @GO is achieved when the more metallic state of platinum is observed in the catalyst because of mixing with cobalt. Besides, the great NPs distribution on graphene oxide leads to an extensive increase in the catalytic performance of Pt 100−x Co x @GO and is also the demonstration of the positive impact of utilizing graphene oxide. When all prepared catalysts were compared, Pt 75 Co 25 @GO NPs exhibited higher performance catalytically in the methanol oxidation reaction, that's why the examination to reveal long-time stability was performed by using 0.5 M methanol and 0.5 M sulfuric acid mixture by chronoamperometry (CA) as shown in Fig. 4b. It can be concluded from the figure (Fig. 4b) that for all catalysts, the peak currents decrease during the time. However, after 3600 seconds, Pt 75 Co 25 @GO NPs have still higher catalytic activity and stability compared to the other prepared Pt 100−x Co x @GO ones. Catalytic lifetime measurements of Pt 75 Co 25 @GO NPs (the best catalyst in prepared ones) are performed in a nitrogen saturated solution of 0.5 M H 2 SO 4 containing 0.5 M CH 3 OH at a scan rate of 50 mV s −1 at a 1 st and 1000 th cycle (vs. Ag/AgCl). As shown in Fig. S4, when it's compared to the current between the 1 st and 1000 th cycle, it can be said that there is only a 12.48% decrease of the initial performance of Pt 75 Co 25 @GO NPs which shows the high stability and durability of the current catalyst. It is important to investigate the high efficiency and stability of Pt-based electrocatalysts. However, it has been found that small Pt nanoparticles can easily be separated from carbon supports and their stability is discussed 21 . Stability problems in PtRu-based catalysts which are considered active catalysts for MOR activity prevent commercial use 61 . Therefore, studies are carried out to protect the catalytic stability at variable molar concentrations and the stability problem is avoided 62 . Similarly, the stability of the Pt 75 Co 25 @GO NPs, in this study is much better than catalysts formed with other molar concentrations. The Pt 75 Co 25 @GO NPs displayed good reactivity in the potential (−0.2 V to 0.8 V) with various scan rates from 50 to 250 mV/s (Fig. S5). In addition, electrical conductivity was determined by performing EIS analysis of the support material (Fig. S6). The increase in the current density with the increase in the potential scan rate is attributed to the excitation signal caused during the charging of the interface capacitance by the charge transfer process. Besides, cyclic voltammetry results showed that when the scan rate was increased, the peaks didn't change which shows the very good electrochemical reversibility and high rate performance. Furthermore, the prepared anodic material indicates high current density and capacitance depending upon its morphology and good conductivity.</p><!><p>As a conclusion, the synthesis and characterization of graphene oxide supported platinum-cobalt nanoparticles with different ratios (Pt 100−x Co x @GO NPs) were performed. The simple double solvent reduction method was employed to produce NPs as a facile method. The structure of the catalysts was detected as face-centered cubic and the dispersion of them on graphene oxide was homogenous (distributed narrowly (4.01 ± 0.51 nm)). The altered promoter served quite dispersed metal holding parts for the nucleation of NPs on the graphene oxide's surface, allowing a monodisperse and homogeneous distribution of Pt 100−x Co x @GO NPs. The synthesized nanoparticles were used directly as anode material in direct methanol fuel cells (DMFCs). Cyclic voltammetry (CV) and chronoamperometry (CA) was utilized for testing electrocatalytic activities of all prepared NPs for www.nature.com/scientificreports www.nature.com/scientificreports/ the methanol oxidation reaction. The best catalytic performance was seen in Pt 75 Co 25 @GO electrodes which have 1.27, 1.44, 1.54 and 2.94 higher performance than Pt 100 Co 0 @GO, Pt 50 Co 50 @GO, PtRu (20%) E-TEK, and Pt 25 Co 75 @GO catalysts, respectively. Further, Pt 75 Co 25 @GO NPs have the highest metal utility (89.38%) compared to the other prepared ones. Besides, platinum is in the more metallic state in Pt 75 Co 25 @GO NPs (83.1%) than the other prepared ones and consequently causes a more effective activation of adsorbed CH 3 OH. The higher performance of Pt 100−x Co x @GO is achieved when the more metallic state of platinum is observed in the catalyst because of mixing with cobalt. Moreover, there is only a 12.48% decrease in the initial performance of Pt 75 Co 25 @GO NPs which shows the high stability and durability of the current catalyst. Besides, cyclic voltammetry results showed that when the scan rate was increased, the peaks didn't change which shows the very good electrochemical reversibility and high rate performance. Furthermore, the prepared anodic material indicates high current density and capacitance depending upon its morphology and good conductivity. We believe that Pt 75 Co 25 @GO NPs will be strategically used in future studies to be used in methanol oxidation reactions. In the near future, these types of materials can also be used in various applications due to their superior properties.</p>

Scientific Reports - Nature

Structure-Activity Relationships of Pyrazole-4-carbodithioates as Antibacterials against Methicillin\xe2\x88\x92Resistant Staphylococcus aureus

Methicillin-resistant Staphylococcus aureus (MRSA) is a major cause of serious hospital-acquired infections and is responsible for significant morbidity and mortality in residential care facilities. New agents against MRSA are needed to combat rising resistance to current antibiotics. We recently reported 5-hydroxy-3-methyl-1-phenyl-1H\xe2\x88\x92pyrazole-4-carbodithioate (HMPC) as a new bacteriostatic agent against MRSA that appears to act via a novel mechanism. Here, twenty nine analogs of HMPC were synthesized, their anti-MRSA structure-activity relationships evaluated and selectivity versus human HKC-8 cells determined. Minimum inhibitory concentrations (MIC) ranged from 0.5\xe2\x88\x9264 \xce\xbcg/mL and up to 16-fold selectivity was achieved. The 4-carbodithioate function was found to be essential for activity but non-specific reactivity was ruled out as a contributor to antibacterial action. The study supports further work aimed at elucidating the molecular targets of this interesting new class of anti-MRSA agents.

structure-activity_relationships_of_pyrazole-4-carbodithioates_as_antibacterials_against_methicillin

<p>Methicillin-resistant Staphylococcus aureus (MRSA) is the most common cause of hospital-acquired infections1,2 and a frequent source of skin and soft tissue infections in the North and Latin Americas, Europe and Asia.3,4 In the USA, MRSA accounts for almost 60% of clinical S. aureus strains isolated from intensive care units5 and it is widespread in residential care facilities.6,7 Worryingly, pathogenic strains are also becoming more prevalent in the community (i.e. community-acquired MRSA).8 Vancomycin has been used extensively for several decades to treat complicated S. aureus and other Gram-positive infections. However, concerns have been growing over the increasing minimum inhibitory concentration (MIC) of the drug against MRSA isolates,9 and its use has other well-known shortcomings (e.g. nephrotoxicity, complex pharmacokinetics, requirement for slow intravenous infusion).10</p><p>It is evident that vancomycin's effectiveness will continue to wane in the coming years, thus compelling the discovery of effective new antibiotics against this major human pathogen.</p><p>We recently reported the discovery of 5-hydroxy-3-methyl-1-phenyl-1H−pyrazole-4-carbodithioate (HMPC) 1 (Scheme 1) as a new bacteriostatic agent against MRSA.11 The compound showed the same minimum inhibitory concentration (MIC 4 μg/mL) against MRSA MW2 (lab strain) and six recent clinical isolates and was able to rescue Caenorhabditis elegans from an MRSA infection. Whole-genome sequencing of mutants resistant to 1 highlighted a role for the global defense regulator MgrA12 in its mechanism and the compound displayed a S. aureus promoter-lux array luminescence profile distinct from all major classes of antibiotics. HMPC 1 appears to exert anti-staphylococcal effects through a novel, uncharacterized mechanism that involves an MgrA-mediated defense response. In the current report, we explored the structural requirements for anti-MRSA activity and eukaryotic cell selectivity in the pyrazole-4-carbodithioate class.</p><p>Initial efforts sought to understand the role of the 4-carbodithioate by replacing the group with amide 2, thioamide 3 and ester 5 isosteres (Scheme 1). Analogs 2 and 3 were formed by quenching the enolate of pyrazolin-3-one 7 (generated using K2CO3) with n-propyl isocyanate and n−propyl isothiocyanate, respectively. Ester 5 was obtained by first generating ketene dithioacetal 4 from 7 using K2CO3, CS2 and excess CH3I via the reported method.13 Subsequent reaction of 4 with sodium in n-PrOH at 80 °C delivered 5. Compounds 2, 3, 5 and intermediate 4 all showed no activity against MRSA MW2 (MIC > 64 μg/mL), establishing the critical importance of the 4-carbodithioate function. The role of the neighboring 3-OH substituent was investigated next but attempting to prepare the O-Me ether of 1 by treatment with K2CO3 and excess CH3I instead delivered ketene dithioacetal 6 (44% yield). Compound 6 was also found to be inactive against MRSA (MIC > 128 μg/mL).</p><p>Replacement of the carbodithioate n-propyl chain with Me, Et, allyl, n-Bu, Bn and 3-pyridyl groups was explored next. Each of these was obtained by forming the enolate of 7 with n-butyllithium and successively quenching with CS2 and the appropriate alkyl halide (Figure 1(a)). Shortening the chain to one carbon 8a led to a 4-fold drop in potency, while Et derivative 8b produced a 2-fold loss. Introduction of an alkene 8c maintained or slightly reduced potency and extension by 1 carbon 8d increased potency 2-fold. Addition of steric bulk and hydrophobicity with a benzyl group 8e resulted in a 2-fold increase in activity, but a significant drop in potency occurred when a nitrogen atom was introduced into the benzylic substituent 8f (MIC 16–32 mg/mL, Figure 1(b)).</p><p>In our previous report,11 we showed that HMPC 1 does not cause hemolysis of human red blood cells at concentrations up to 64 μg/mL but is cytotoxic towards eukaryotic HKC-8 and HepG2 cells at concentrations around its MRSA MIC. Here, cytotoxicity of analogs 8a-e was tested in HKC-8 cells and selectivity indices were calculated (SI = HKC-8 IC50/MRSAMW2 MIC, Figure 1(b)). No MRSA selectivity was observed for 8a, 8b or 8d and modest selectivity (2–4 fold) was seen with allyl derivative 8c. Benzylic derivative 8e delivered the highest selectivity (8–16 fold) in this series.</p><p>A variety of halo, electron donating and electron withdrawing substituents were added to the 4-position of the pyrazole N−phenyl group. Commercially available 4-substituted phenylhydrazine.HCl salts 9a-g were condensed with ethyl acetoacetate to form pyrazol-3-one intermediates 10a-g in 25–87% yield. Treating ketones 10a-g with n-butyllithium and quenching the enolates with CS2 followed by n-bromopropane gave targets 11a-g in 32–75% yield (Figure 1(a)).</p><p>Addition of a Me group 11a gave no change in activity (relative to 1) while introducing an electron donating methoxy group 11b produced a 2-fold loss. Halogen substituents 11c-e gave slight increases in potency (2–4 fold), as did electron withdrawing cyano and nitro groups, with the p−NO2−substituted analog 11g showing the highest activity (MIC 0.5–1 μg/mL). Corresponding increases in eukaryotic cell cytotoxicity were observed though, with 11g showing no MRSA selectivity (Figure 1(b)).</p><p>Replacement of the 3-methyl group of 1 with a phenyl ring was explored next. Condensation of phenylhydrazine.HCl with ethyl benzoylacetate 12a produced pyrazol-3-one 13a, which upon base treatment and successive quenching with CS2 and n-bromopropane afforded 3-phenyl derivative 14a in 90% yield. Compound 14a showed a 2-fold increase in MRSA potency and a 16-fold increase in selectivity. The promising selectivity obtained upon addition of the phenyl ring at the 3-position led to exploration of para−substituted analogs carrying halo, electron donating and electron withdrawing substituents. Condensation of p-substituted ethylbenzoylacetates 12b-i with phenylhydrazine.HCl and appending n-propyl dithioate groups to the resulting ketones 13b-i delivered target analogs 14b-i. Addition of the Me group 14b led to a slight increase in potency (MIC 1–2 μg/mL) relative to 1 but reduced selectivity. The methoxy group 14c did not change activity against MRSA but selectivity was reduced. Larger increases in antibacterial potency were achieved with halo groups 14d-f (MIC 0.5–1 μg/mL) but no improvements in selectivity were seen. A large drop in potency was observed with the CF3 group 14g (MIC 32–64 mg/mL), while other strongly electron withdrawing cyano 14h and nitro 14i substituents maintained activity but reduced selectivity.</p><p>We previously showed that treatment of a S. aureus promoter-lux array with HMPC 1 produces a unique luminescence profile (suggesting a unique mechanism of action), but some similarities to DNA-damaging agents and/or DNA replication inhibitors were noted.11 This led to speculation that the anti-MRSA and apparent general cytotoxicity of 1 might arise from DNA binding. However, UV/vis experiments measuring the binding of 1 to calf thymus DNA and zone of growth inhibition disk measurements performed with 1 in the presence/absence of calf thymus or S. aureus genomic DNA appeared to rule this out. Nevertheless, the similar levels of HKC-8 toxicity observed in the current study with the majority of analogs of 1, combined with the absolute requirement of a carbodithioate function for anti-MRSA activity, led us to examine more closely whether 1 (and hence the class) may exert effects through non-specific nucleic acid and/or protein reactivity.</p><p>The DNA-reactivity of 1 was probed by electrospray ionization mass spectrometry (ESI-MS) using a panel of single stranded, double stranded and G-quadruplex DNA oligonucleotides. When incubated with up to 10-fold excesses of HMPC 1 under a variety of conditions, no evidence for any DNA:1 adducts was observed (Supporting Information Figure S1). Similarly, no adducts were observed by ESI-MS when 1 was incubated with RNA oligonucleotides (data not shown).</p><p>The effects of HMPC 1 on DNA replication in vitro were explored next. In this assay, all of the enzymes, ancillary proteins, nucleotide precursors, DNA template and other molecular components required for duplication of circular bacterial DNA are present and able to effect replication in a cell-free environment.16 Covalent reactivity with any of the reaction components would be expected to read out as inhibition of replication. However, only slight inhibition of replication was observed with 1 at concentrations > 160 μM, well above its MIC against MRSA (Supporting Information Figure S2). In vitro RNA transcription assays similarly showed no inhibition by 1 at relevant concentrations (data not shown). The absence of effects for 1 in these assays rules out non-specific nucleic acid or protein reactivity as a contributor to the anti-MRSA mechanism of the pyrazole-4-carbodithioate class.</p><p>In summary, pyrazole-4-carbodithioates are a new class of anti-MRSA agents that require the 4-carbodithioate function for activity. Non-specific covalent reactivity appears not to be part of the mechanism but intrinsic reactivity of the 4-carbodithioate may still play a role. We showed previously that MgrA-mediated defense responses are triggered by 1.11 MgrA is an oxidation-sensing mechanism used by MRSA to counter challenges of reactive oxygen and nitrogen species. Upon detecting these species, a unique cysteine residue (Cys12) located at the dimer interface of the protein is oxidized to cysteine sulfenic acid, causing dissociation of MgrA from DNA and initiation of signalling pathways that turn on antibiotic resistance.17 We speculate that intracellular redox reactivity of the 4-carbodithioate function (aided by the neighboring OH group) triggers oxidative stress that leads to MgrA activation. Alternatively, metal chelation by the 4-carbodithioate and neighboring OH group may be involved.18 While limited selectivity (maximum 16-fold) for MRSA over eukaryotic HKC-8 cells was achieved with the analogs explored here, further increases seem possible with a larger analog set. Studies to fully elucidate the anti-MRSA mechanism and identify the discrete intracellular targets using such selective analogs would undoubtedly prove insightful.</p>

PubMed Author Manuscript

Assessment of reproductive and developmental effects of graphene oxide on Japanese medaka (Oryzias latipes).

Due to its unique properties, graphene oxide (GO) has potential for biomedical and electronic applications, however environmental contamination including aquatic ecosystem is inevitable. Moreover, potential risks of GO in aquatic life are inadequately explored. Present study was designed to evaluate GO as an endocrine disrupting chemical (EDC) using the model Japanese medaka (Oryzias latipes). GO was injected intraperitoneally (25\xe2\x80\x93200 \xce\xbcg/g) once to breeding pairs and continued pair breeding an additional 21 days. Eggs laid were analyzed for fecundity and the fertilized eggs were evaluated for developmental abnormalities including hatching. Histopathological evaluation of gonads, liver, and kidneys was made 21 days post-injection. LD50 was found to be sex-dependent. Fecundity tended to reduce in a dose-dependent manner during early post-injection days; however, the overall evaluation showed no significant difference. The hatchability of embryos was reduced significantly in the 200 \xce\xbcg/g group; edema (yolk and cardiovascular) and embryo-mortality remained unaltered. Histopathological assessment identified black particles, probably agglomerated GO, in the gonads of GO-treated fish. However, folliculogenesis in stromal compartments of ovary and the composition of germinal elements in testis remained almost unaltered. Moreover, granulosa and Leydig cells morphology did not indicate any significant EDC-related effects. Although liver and kidney histopathology did not show GO as an EDC, some GO-treated fish accumulated proteinaceous fluid in hepatic vessels and induced hyperplasia in interstitial lymphoid cells (HIL) located in kidneys. GO agglomerated in medaka gonads after 21-days post-injection. However, gonad histopathology including granulosa and Leydig cells alterations were associated with GO toxicity rather than EDC effects.

assessment_of_reproductive_and_developmental_effects_of_graphene_oxide_on_japanese_medaka_(oryzias_l

Introduction:<!>Synthesis of Graphene Oxide (GO):<!>Culture and maintenance of Japanese medaka fish.<!>Histopathological evaluation:<!>Statistical analysis:<!>Characterization of the GO synthesized in the laboratory:<!>Evaluation of the reproductive activity:<!>Ovary:<!>Testis:<!>Liver:<!>Kidney:<!>Discussion:<!>Conclusion:

<p>In recent years, graphene, a sp2-bonded carbon nanosheet, has been a promising biomaterial for diagnostic and therapeutic applications. The graphene family of nanomaterials including few-layer-graphene, ultrathin graphite, GO, reduced graphene oxide (rGO), graphene nanoplatelets, graphene nanosheets and others, varied in layer numbers, layer dimensions, surface chemistry, density or purity (Novoselov et al., 2012; Bianco, 2013; James and Tour, 2013). Because of their high water solubility (6.6 μg/mL in de-ionized water; Konios et al., 2014; Johnson et al., 2015), and use as precursors of chemically generating large-scale graphene structures, GO nanosheets have the potential to become a biomaterial for application both in vivo and in vitro models. As a graphene derivative, GO has a large two-dimensional plane, which provide ultrahigh, specific surface area to load drugs through surface adsorption, π-π interaction, hydrogen bonding, and others. GO is biocompatible and nontoxic which makes it a promising candidate for construction of drug carriers. However, several published reports indicate that GO is able to produce toxic effects in various animal models (reviewed by Ema et al., 2016) including fish (reviewed by Dasmahapatra et al., 2019). Moreover, even though GO could yield stable suspensions in water initially after being exfoliated into monolayer sheets, it may aggregate immediately in biomedia (blood or tissue fluids), especially after loading with anticancer drugs. Hence there is a need to investigate the toxic effects of GO considering all possible ways including various exposure routes, and to ensure that the methods used for its application have no toxic impact on animals (Kurantowicz et al., 2015).</p><p>Aquatic environments are the ultimate sink for many contaminants, either due to direct discharge or by hydrological and atmospheric processes. In recent years, increasing amount of carbon-based nanomaterials (CNM) have been observed in aquatic systems. However, the concentration of CNM in the aquatic environment is very low (ng/L), even lower than the lowest observed effect concentrations (LOEC) reported on different aquatic organisms (Freixa et al., 2018). Moreover, CNM interact with other micro-pollutants, thereby modifying the original toxicity of contaminants. Since fish occur in virtually all aquatic environments and play a major role in aquatic food webs, they are useful models for monitoring pollutants in aquatic environment. Small laboratory fish models, including zebrafish (Danio rerio) and Japanese medaka (Oryzias latipes), are widely used as experimental models for monitoring pollutants in the environment.</p><p>The toxicity of GO in fish, especially in zebrafish (Danio rerio) was found to be concentration-dependent (Bageppagari et al., 2019; Dasmahapatra et al., 2019; Yang et al., 2019), and did not affect embryonic development at lower concentrations. However, higher concentration showed significant impact on embryo mortality, hatching, and disorder in cardiovasculature (Bageppagari et al., 2019). Exposure of zebrafish embryos to GO for five consecutive days did not induce acute developmental toxicity, however, hatching and locomotor behavior were altered (Yang et al., 2019). Moreover, genes related to nervous and immune systems were upregulated (Yang et al., 2019). Proteomic analysis in zebrafish larvae identified down-regulation of several oxidative stress-related proteins by GO (Zou et al., 2018; Dasmahapatra et al., 2019). In tilapia fish, oral exposure of GO did not alter behavior, body weight, and the intestinal bacterial population after 30 days of treatment. However, the expression of several oxidative stress-related enzymes was down regulated significantly in liver, but remained unaltered in spleen, gill, intestine, and muscle (Ma et al. 2016). In Geophagus iporangensis, GO exposure decreased metabolic rate compared to control fish (Medeiros et al., 2019). Several oxidative stress-related enzymes were increased in the climbing perch (Anabas testudineus) after 24h of single intraperitoneal (ip) injection of GO (Paital et al., 2019).</p><p>Although in fish models, evaluation of GO was mainly focused on toxicological endpoints mediated probably by oxidative stress-related enzymes and pathways, the role played by GO as an EDC needs to be investigated. A recent report has indicated the potential of GO as a thyroid endocrine disruptor in tadpoles of Xenopus laveis (Li et al., 2019). Due to limited data, the effects of different nanoparticles on endocrine system need to be interpreted carefully (Iavicoli et al., 2013). Previously, our laboratory evaluated the toxic potential of GO administered orally to Sprague-Dawley rats using liver and kidney tissues. We observed that oral administration of higher doses of GO (20–40 mg/kg, five consecutive days) induced cellular damages in liver and kidneys as well as enhanced enzyme activities related to oxidative stress (Patlolla et al., 2016; 2017).</p><p>The aim of the present study was to assess the toxic potential of GO using Japanese medaka fish as an aquatic animal model. In addition, we also evaluated on the endocrine disrupting effects of GO targeting testis and ovary. Leydig cells in the testis produced male hormones (testosterone) while the perifollicular cells, especially theca and granulosa cells, in the ovary of medaka are responsible for female hormones (estrogen). This fish species is a recognized model to evaluate EDCs in the environment and several Test Guidelines (TG) are included in publications made by Organization for Economic Cooperation and Development (OECD, 2018). We used sexually mature, reproductively active male and female intact Japanese medaka fish maintained in glass aquaria under standard laboratory conditions (25±1°C; 16L: 8D light cycle). We focused on fecundity and histopathology of gonads (testis and ovary) as toxicological and reproductive (endocrine) endpoints 21 days after single ip injection of GO (25–200 μg/g). In fish, fecundity can be characterized by a wide spectrum of disorders including endogenous hormone levels, activities of gonads, as well as the behavior of the fish during mating and external fertilization. However, fecundity in medaka is highly variable and reproducibility is questionable (Hutchinson et al., 2006) even though high concentrations of EDCs such as 17 alpha-ethinylestradiol (EE2) or 17 beta-trenbolone (TB) significantly reduced fecundity (Park et al., 2009). Moreover, in addition to histopathology of gonads, and because of the significant role played by liver and kidneys in fish reproduction (OECD, 2010, 2018), we conducted histopathological evaluation of these two organs focusing on EDC effects. To our knowledge, the toxicity and EDC effects of GO in reproductively active Japanese medaka adults remained unexplored.</p><!><p>The GO we used in the experiments was synthesized either in the laboratory, or obtained from a commercial source (Sigma-Aldrich, St. Louis, MO). The synthesis of GO in the laboratory from natural graphite powder was made by following modified Hummer's Method (Hummers and Offeman, 1958; Viraka Nellore et al., 2015). For details of synthesis, please see the supplementary data (Supplementary Figure (SF) 1A–B, SF2A–C; Supplementary Table 1).</p><p>Before ip injection, GO was dissolved to desired concentration (1–2 mg/mL) in nanopure water and sonicated for 5 min (2s on-1s off pulse, 225 W) by a probe sonicator (ultrasonicator LPX 750, Cole Parmer, Chicago, IL, USA). The particle diameter and zeta potential of the synthesized GO after sonication were determined by Zeta Sizer Nano ZS (Malvern Instrument, MA, USA) (Supplementary Table 1).</p><!><p>The Institutional Animal Care and Use Committee (IACUC) of the Jackson State University, Jackson, MS, approve all experimental protocols.</p><p>Adult Japanese medaka fish (orange red variety), obtained from the University of Mississippi, Oxford, Mississippi, USA, by a protocol transfer agreement, and a breeding colony was set up at the Jackson State University, Jackson, MS, USA. For details of fish culture, please see the supplementary data file. Briefly, adult fish (6 females and 4 males) used as breeders were maintained in 25 L balanced salt solution (BSS, pH7.4) in glass tanks under standard laboratory conditions. The collection of eggs was generally made within 1–3 h after the light was turned on.</p><p>For GO experiments, sexually mature and reproductively active fish (one male and one female) were maintained for one week in 500 mL BSS in 1L glass jars. One week prior to GO (25, 50, 100, and 200 μg/g) and vehicle (water) injections, the eggs were collected from each individual pair, and the fertilized eggs cultured in embryo-rearing medium (ERM, pH 7.4). The embryos were allowed to hatch spontaneously until 14 days post fertilization (dpf) in 20 mL glass vials containing 10 mL ERM. After one successful week of egg collection, GO was injected ip to the fish (both males and females) and the controls (both males and females) were injected with nanopure water (1 μL/10 mg body weight). Some fish (4 pairs for both 50 and 100 μg/g and 9 pairs for 200 μg/g doses) were injected with GO obtained from commercial source and the collected data were pooled with the data generated from the fish injected with GO synthesized in the laboratory. After injection of either GO or vehicle, both male and female fish were returned to the respective tanks for continuation of breeding (same partners). The fish mortality data (died within 21 days post-injection period) were analyzed by two way ANOVA followed by post-hoc Tukey's multiple comparison test. The mortality of individual males, or as a breeding pair injected with 200 μg/g GO showed a significant difference (p<0.05) only with controls (Figure SF3A). Although the r2 is low, LD50 was calculated by probit analysis (Finney 1971) which showed males are more GO-sensitive than females (Figure SF3B).</p><p>The egg collection was continued in control and GO-injected fish during 21 days post-injections days (21 days after injection), and the fertilized eggs were allowed for spontaneous hatching in ERM. The hatched embryos (larvae; stage 40; Iwamatsu, 2004) were used for morphological evaluation (complete or incomplete hatching, edema: both cardiac and yolk sac edema were considered as edema). The daily fecundity either as total eggs (both fertilized and unfertilized eggs) or as fertilized eggs during post-injection periods of each breeding pair was calculated by using the following formula: (i)Daily fecundity=(number of eggs produced by a breeding pair/average number oftotal eggs produced by the same breeding pair during pre-injection periods)*100. [The number of eggs indicates either total eggs (fertilized+ unfertilized) or only fertilized eggs].</p><!><p>After 21 days post-injection, surviving fish were anesthetized in MS222 (Readmann et al., 2017), weighed to the nearest mg, and quickly cut into three small pieces excluding post anal tail and fixed in 4% paraformaldehyde (PFA) containing 0.05% Tween 20 (Sigma-Aldrich, St. Louis, MO, USA). For detail of histopathological evaluation, please see supplementary data file. Briefly, all tissues were paraffinized by following standard procedure and the serial sections of 5-μm thickness were cut in a manual rotary microtome (Olympus cut 4055). The sections were stained in hematoxylin/eosin (HE), following standard protocol. Some of the sections (testis) were stained in Masson's trichrome stain (Sigma-Aldrich, St. Louis, MO). The photomicrographs of the sections were taken either in an Olympus B-max 40 microscope attached to a camera with Q-capture Pro 7 software (Media Cybernetics, Inc, Rockville, MD) or in Nikon Eclipse 50i microscope attached to Nikon DS-Fi1 camera (Nikon Inc, Melville, NY). Those sections of the testis, ovary, liver or kidney were used for histopathological evaluation following OECD guidelines (OECD, 2010). Six types of follicles distributed in four different regions of the ovary (anterior, upper middle, lower middle and anal) in stromal compartments were considered for counting; four of them as differentiating follicles (perinucleolar oocytes, PNO; cortical alveolar oocytes, CAO; early vitellogenic oocytes, EVO; and late vitellogenic oocytes, LVO) and two of them (atretic and postovulatory) as degenerating follicles. The specific features of these follicles (differentiating and degenerating) are listed in Supplementary Table 2. The following formula are used for calculation (%) of the distribution of different types of differentiating and degenerating follicles in a particular section (5 μm thick) cut through in any of the four different regions (anterior, upper middle, lower middle, and anal) of the ovary. (ii)Any differencing stromal follicles in %=[number of that particular follicle/ sum of differentiating follicles]*100;where any follicle indicates either PNO,CAO,LVO or EVO. (iii)Degenerating follicles in %=[(sum of atretic+post ovulatory follicles)/(sum of differentiating+degenerating follicles)]*100 The ratio of degenerating and differentiating (degenerating: differentiating) follicles found in that particular section of the ovary were also determined.</p><p>Although among 21 surviving breeding pairs as controls, we used only 7 pairs for histological analysis. However, all the surviving breeding pairs of 25–200 μg/g groups (5 pairs for 25 μg/g, 4 pairs for 50 μg/g, 7 pairs for 100 μg/g, and 3 pairs for 200 μg/g) were used for histological analysis. For testis, the visual assessment of the gametogenic precursors (Spermatogonia, spermatocytes and spermatids), mature gametocytes (sperms), the Sertoli cells and Leydig cells, was made on HE stained slides. In addition, for careful observations of Sertoli cells and Leydig cells, Masson-Trichrome (MT)-stain (Sigma-Aldrich, St. Louis, MO) was used. For liver, the observations were mainly focused on hepatocytes distributed in the central vein regions, while for the kidneys, they were focused on the glomerulus, renal tubules, and interstitial tissues associated to glomerulus regions.</p><!><p>Data were analyzed by either one- or two-way ANOVA followed by post-hoc Tukey's multiple comparison test and expressed as mean ± SD or SEM (GraphPad prism version 7.04 software; GraphPad Prism, San Diego, CA); p<0.05 was considered as significant.</p><!><p>We synthesized GO using the standard protocol (Hummers and Offeman, 1958). After purification, synthesized GO was extensively characterized using several microscopic and spectroscopic procedures. (For details, please see the Supplementary Figures, SF1A–B, SF2A–C).</p><!><p>The physical conditions we provided during the experimental period allowed both control and GO-treated fish to maintain normal growth (Figure SF4), successful breeding, and laying eggs (both fertilized and unfertilized) into the environment (Figures 1A–B). The body weight of the fish remained unaltered except for the females of 200 μg/g group where it was increased significantly (p<0.05) compared to the fish in pre- and post-injected vehicle-treated (control) fish (Figure SF4).</p><p>Reproductive activities of the experimental fish were evaluated as % eggs laid into the external environment by the female (total eggs) or the laid eggs (%) fertilized by the males (fertilized eggs) of each breeding pair (external fertilization) per day before and after GO or vehicle (water) administration (Figures 1A–B). Further, the fertilized eggs were evaluated for mortality, hatchability (complete and incomplete), and the hatched larvae were evaluated for edema (either yolk or cardiac or both) (Figures SF5A–F). Our data indicate that the breeding pairs surviving during the entire period of experiment (7-days before injection and 21-days after injection), were able to breed successfully and the calculated fecundity (%) during post-injection periods were initially low and then gradually returned to the status (both as total and fertilized) observed in pre-injection days. The data further indicate that in controls, and fish that received lower doses of GO (25 and 50 μg/g), steady-state increase (both total and fertilized eggs) was observed after 1 day post-injection. While higher doses (100 and 200 μg/g) required at least one week or longer to achieve the steady-state fecundity (Figures 1A–B). Further analysis of the data showed that the fecundity observed as total eggs in vehicle-injected control fish (n=21), on day 1 post-injection, showed a significant difference with the control fish on post-injection days 3 and 4 and continuously from 11–21 days (p<0.05). No significant difference was observed in total fecundity in control fish in 2–21 post-injection days (Figure 1A). Moreover, the total fecundity in day 1 post-injected fish received 25 μg/g GO (n=5) established significant difference with the same group only on day 18 post-injection (p<0.05). However, those fish that received 100 μg/g GO were significantly different with days 2, 3, and 4 fish only on post-injection day 18 (p<0.05). The total fecundity of the fish that received 50 μg/g (n=4) or 200 μg/g (n=3) GO, maintained equal status in post-injection days 1–21 (not significantly different in any days).</p><p>The fecundity data (%) when expressed as fertilized eggs remained almost identical with the data expressed as total eggs (Figure 1B). In controls (n=21), fecundity as fertilized eggs on day 1 showed significant difference with the fecundity (fertilized) of the control fish 11–21 days post-injection (p<0.5). Like total fecundity, in controls, the fecundity as fertilized eggs remained unaltered between post-injection days 2–21. In GO-treated fish, the fecundity as fertilized eggs of 25 μg/g group (n=5) on day 1 post-injection showed significant difference (p<0.05) with post-injection days 17, 18, and 21. In 100 μg/g group (n=7) post-injection days 2 and 3 were able to show significant difference (p<0.05) only with days 17 and 18. Moreover, as in total eggs, the fecundity as fertilized eggs maintained equal status in 1–21 days post-injection periods in fish that received either 50 (n=4) or 200μg/g GO (n=3) (Figure 1B). We further observed that the laid eggs fertilized by the males in vehicle-treated controls or the fish received GO (50 μg/g and 200 μg/g) were found to be significantly less (p<0.05) than the eggs fertilized by the male fish before ip injection. However, other GO-treated groups (25- and 100 μg/g) did not show significant difference (p<0.05) either with the pre-injected or post-injected control fish (Figure SF5A).</p><p>Reproductive activities were further verified on the hatchability of the fertilized eggs laid by the female fish before and after GO injection (Figure SF5B). During this period, 6–13% of the embryos died (Figure SF5F) though the data were unable to show significant difference with pre- and post-injected controls as well as in different GO-treated groups (25–200 μg/g). Further, our evaluation as dead, hatched, unhatched, incompletely hatched larvae indicates that 24–61% of the fertilized embryos were successfully hatched (Figure SF5B) and 16–55% of the fertilized eggs collected before (7 days) and after injection (21 day) periods were unable to hatch within 8–14 dpf time limits (remained within the chorion, showing heartbeats) (Figure SF5C). Moreover, the data of 200 μg/g group (n=3) show a significant difference with pre-and post-injected control fish (Figure SF5B) and only with pre-injected control fish with regard to unhatched conditions (Figure SF5C). Three to 8% of the larvae show edema (cardiac or yolk) which are also not significantly different between pre-injected or post-injected controls or GO-treated (25–200 μg/g) parents (Figure SF5D). Data on incomplete hatching (1–8%) were significantly different only in 25 μg/g group with the vehicle-injected control fish (Figure SF5E).</p><!><p>During development, medaka have a pair of ovaries, separated as right and left ovaries. As development proceeds, formation of an ovarian cavity on the dorsal side and the fusion of the paired ovaries into a single organ occurred (Nakamura et al., 2018). The mature ovary of medaka is covered with the ovarian epithelium on both dorsal and ventral sides, consisting of two major compartments. The ovarian cavity on the dorsal side to which the mature eggs are ovulated (Nakamura, 2018), and the stromal compartment on the ventral side, in which folliculogenesis, oocyte maturation, and production of steroid hormones by perifollicular cells (PFC; surface epithelial cells, theca and granulosa cells) occurred (Nakamura et al., 2011). The germinal epithelium (Figures 2A1, 2B1, 2C1, SF6A–B), which is equivalent to the surface epithelium in mammals, has basement membrane on the ventral side, which lies between stroma and ovarian cavity (Figures 2A1, 2B1, 2C1, SF6A–B). Germinal epithelium formed germinal cradles (Nakamura et al, 2011) containing oogonia, prefollicular and prethecal cells, epithelial cells, and occasionally small chromatin nucleolar oocytes (Norberg et al., 1999; Parenti and Grier, 2003; OECD 2010). In stromal compartments, PNOs, CVOs, EVOs, and LVOs are distributed within a variably apparent extravascular space (Figures 2A1, 2B1, 2C1, SF6A–B) (OECD, 2010). Moreover, degenerating follicles (atretic and postovulatory follicles) also exist in the stromal compartments of the ovary (Figures 2A, 2B, 2C, SF6A–B).The ovarian lumen, which is also a part of ovarian cavity, covered by the luminal epithelium, where the mature eggs are ovulated from the stromal compartment through the germinal epithelium (Nakamura et al., 2011).</p><p>In order to evaluate the effects of GO in the folliculogenesis of medaka ovary, we observed ip-injected GO accumulated and agglomerated in ovarian compartments in a dose-dependent manner (more in higher doses) and damaged near-by follicles probably by necrosis (Figures 2B2, 2B4, 2C1–3, SF6A–E). However, the histological structure of the rest of the ovary, where GO agglomerated were lacking, remained unaffected and probably continued normal folliculogenesis (Figures 2C1 and SF5C). In some areas of the ovary, interstitial fibrosis might occur that increased the formation of perivascular smooth muscles (Figures 2C1–2). Distributions of PNOs, CAOs, EVOs, and LVOs in the stromal compartments of the ovary of control fish (n=7) indicate that approximately 65% of the differentiating follicles are PNOs, 20% are CAOs, and 10–15% are vitellogenic (EVO and LVO) oocytes. Except PNO population in 50 μg/g group, individual counts of differentiating follicles are unable to show significant differences between control and GO-treated (25, 100, and 200 μg/g) fish after 21-days post injection (Figure 2D). In 50 μg/g group, the PNO population in the stromal compartments of the ovary showed significant reduction (p<0.05) when compared with control or 100 μg/g groups, however, remained at the same level as in 25 and 200 μg/g groups (Figure 2D). Our data further indicate that in controls, approximately 90% of the stromal follicles are the differentiating follicles and 10% of the follicles are in degenerating stages (Figure 2E). GO at 25, 100 and 200 μg/g are unable to show significant differences with controls either as degenerating follicles (Figure 2E) or as a ratio of the degenerating and differentiating follicles (Figure 2F). In 50 μg/g group, the distribution of degenerating follicles increased significantly than in control and 100 μg/g groups, and remained at the same level as in 25 and 200 μg/g groups. Consequently, the ratios of degenerating and differentiating follicles (Figure 2F) are found to be significantly different (increase) in 50 μg/g group than control and 100 μg/g groups (p<0.05) and at the same level as in other two groups (25 and 200 μg/g) (Figure 2F).</p><p>The follicles are covered by chorion (zona radiata or vitelline envelope); in HE stain, the chorion appears to be a pale to dark eosinophilic outer membrane that surround the ooplasm of the oocytes (Figures 2A2–3, 2B2–3, 2C2–3, SF6D–E). Although a chorion-like structure exists in PNOs, in light microscopic observations, they become clearly visible in CAOs and gradually thicken as the oocytes mature to vitellogenic stages (Figures 2A2–3, 2B2–3, 2C2–3, SF6D–E). Next to chorion, the granulosa cell layer surrounds the chorion (Figures 2A2–3, 2B2–3, 2C2–3, SF6D–E) followed by theca cells; the surface epithelial cells remain external to theca layer. Among these cell layers, theca and granulosa cells secret steroid hormones including estrogen. In medaka, the GCs in PFC of oocytes arranged as single cell layer with cuboidal to columnar appearance (Figures 2A2, 2B2, SF6D) even though remained thicker (GCs) at the vegetal hemisphere than at the animal hemisphere (Iwamatsu et al., 1988). The oocytes under spontaneous atrophy showed multiple layers in PFC (Figure 2A3). In atretic follicles, they become stratified (Figures 2B3, 2C3, and SF6E). In post-ovulatory follicles, the granulosa cells multiply and become hypertrophic (Figures 2A4, 2B4, and 2C4), if the fish are exposed to EDCs. Our observations indicate that follicular atresia can occur either in chorion or in granulosa cells, or both. It becomes spontaneous to the follicles probably when they are detracted from maturational processes. We observed a considerable number of atretic follicles in both control and GO-treated (25–200 μg/g) fish. Agglomerated GO can induce atresia in differentiating follicles of the ovary when it is in direct contact either with chorion or granulosa cells, or both (Figures 2B3, 2C2–3, SF6E). However, possibilities of spontaneous atresia cannot be ruled out in these regions. In postovulatory follicles, the granulosa cells become hypertrophic and multiplied (Figures 2A4, 2B4, 2C4); however, when they are associated with GO, it is very difficult to determine their identity (Figure SF6C). Moreover, like differentiating follicles, they appear to be normal in stromal compartments where GO agglomerates are lacking (Figures 2B, SF6C).</p><!><p>The testis of reproductively active medaka is a semi-translucent, elongated structure consisting seminiferous cysts (Satoh, 1974). Histologically, the testes of adult medaka comprise a lobular restricted type structure (Grier 1976: Parenti and Grier 2004) in which spermatogonia are found at the distal ends (periphery) of the lobules (Figures 3A1, 3B1, 3C1, SF7A–C), end blindly at the periphery of the organ (Miller et al., 2012). The central region of the testis consists of efferent ducts (ED) and the lumens, which are flexible in diameter, filled with dark basophilic spermatozoa (Figures 3A1, 3B1–3, 3C1–3, SF7A–B). The flexibility of the lumen can increase or decrease the inner area, which is probably dependent on the invasion or evasion of the ductal walls. Sertoli cells (Figures 3A1–2, 3B1–2, 3C1–2, SF7A–C), compared to germinal cells, present in low numbers, usually as single cell, located adjacent to lobular septa. For histological assessment, the testis of medaka can be divided into two major compartments; the lobular and the interstitial. The lobular compartment consists of germinal components (spermatogonium, spermatocytes, spermatids), sperms (Figures 3A1, 3B1, 3C1, SF7A–C) and Sertoli cells (Figures 3A1, 3B1–2, 3C1–2 and SF7A–C). The interstitial compartment consists of Leydig cells (Figures 3A1, 3A3, 3B1, 3B3, 3C1, 3C3 and SF7A–C), peritubular myloid cells, connective tissue, efferent duct epithelial cells, endothelium of blood vessels, and circulating blood cells. The Leydig cells remain in small clusters in the stroma surrounding the seminiferous cysts and at the smaller terminal divisions of the ED (Gresik et al., 1973; Satoh, 1974). They have dense, dark round or oval nuclei with faintly vacuolated cytoplasm. Compared to germinal cells, interstitial cells are usually present in low numbers, either as single cell or as small aggregates within the interlobular interstitium, and have no close association with other cell types of the interstitium except with other Leydig cells (Gresik et al., 1973). In the lobular compartments, the germ cells have the ability to divide whereas the Sertoli cells, which are also known as the nurse cells, form spermatocyst with spermatogenic (Spermatogonia, spermatocytes, or spermatids) cells, provide nourishment to the testicular components, and remain undivided when spermatogenesis begins. The interstitial components also remain undivided during spermatogenesis. The Leydig cells secret steroid hormones, which are necessary for gonadal maturation and function.</p><p>Our observations indicate that the ip-injected GO agglomerated on the peritoneal membrane or near the testis, and is likely to be associated with the periphery (Figure 3C1), gradually cross the basement membrane and incorporates into the testis in a dose-dependent manner (Figure SF7C). Germinal components (spermatocytes and spermatids) of the lobular compartments of medaka testis indicate that in both control and GO-treated fish (25–200 μg/g), within a particular cyst (lobule), all germinal elements were at the same stage of development. However, the stages of development of germinal elements differ from cyst to cyst (Figures 3A1, 3B1, 3C1, SF7A–C). Moreover, with current light microscopic studies, we did not observe any remarkable morphological variations in germinal cells (spermatogonium, spermatocytes, spermatids) or sperms and Sertoli cells (Figures 3A1, 3B1, 3C1, SF7A–B). However, apoptotic cells (identified by larger size and deep stained basic pycnotic nuclei) were observed in the ED system of the testis of both control and GO-treated (25–200 μg/g) fish (Figures 3A1, 3B1, 3C1, SF7A–B). The Sertoli cells with deep black nuclei and long cytoplasmic processes (Figures 3A1–2, 3B1–2, 3C1–2 SF 7A–C) also remained unaltered after GO treatment (25–200 μg/g) with respect to their shapes and sizes. We also observed that close association or agglomeration of GO inside the testis enhanced the spermatogonial lobules (cysts) in the periphery (Figure 3C1), or moved them toward ED of the testis (Figure SF7C). However, distribution of Leydig cells within the interstitial compartments are identical in control and GO-treated fish (25–200 μg/G); either present as single or aggregates with no significant major alteration in histological structures or shapes (Figures 3A1, 3A3, 3B1, 3B3, 3C1, 3C3, and SF7A–C).</p><!><p>The hepatocytes of medaka are polyhedral in shape, with clear edges, broad cytoplasm with different degrees of hepatocellular intracytoplasmic vacuolation (Figures 4A1 and 4A2). Hepatocytes were radiating from the central vein creating a tubulo-sinusoidal arrangement in which hepatocytes arranged as cellular duplex surrounding a sinusoidal capillary (Hinton et al., 1984). Compared to males (Figure 4A1), cytoplasm of hepatocytes in female fish is basophilic in nature (Figure 4A2). Blood and blood cells are frequently seen in hepatic blood vessels. The muralium duplex suggested a possible mechanism for facilitating the nutrition of hepatocytes and releasing their products conveniently into the bloodstream (de Oliveira et al., 2016). The white vacuoles observed in HE stain indicate the existence of glycogen in hepatocytes. Administration of GO (25–200 μg/g) to male and female fish was unable to document GO agglomerates in the hepatocytes of liver, or induction of any noticeable changes in the histopathology of medaka liver after 21-day post injection. However, some GO-treated fish showed the accumulation of proteinaceous fluid in hepatic blood vessels (Figures SF8A1–2, and SF8B2). Hepatocyte basophilia, which is related to the synthesis of vitellogenin (the egg yolk precursor protein) or other proteins in the hepatocytes, compared to controls (male Figure 4A1, female Figure 4A2) also remained unaltered in GO-treated male (Figures 4B1, 4C1, SF8A1 and SF8B1) or female fish (Figures 4B2, 4C2, SF8A2 and SF8B2). Only one male fish injected with 100 μg/g GO (Sigma-Aldrich, St. Louis, MO) showed lesion in several sites of liver which can be compared with spongiosis hepatis (cystic degeneration) (Figure SF8C), a nonneoplastic lesion observed in Japanese medaka (Boorman et al., 1997; Ding et al., 2010). The spongiosis hepatis formation replaced large areas of liver parenchyma (Figure SF8C) that may affect the normal functions of liver. However, other control or GO-treated male and female fish (25–200μg/g) used in these experiments did not show similar cystic degeneration in the liver.</p><!><p>The kidneys of medaka are mesonephric, both the cranial (head) and trunk portions composed of glomeruli, renal tubules, and interstitial lymphomyeloid tissues (Figures 5A1, male; 5A2, female). Structurally, the glomerulus is divided into vascular and epithelial regions. The vascular region is the core structure of glomerulus and consists of capillary network and mesangium. The vascular region is surrounded by a flat epithelium of Bowman's capsule. There are at least two types of tubular segments identified by PAS staining. The proximal convoluted segments with narrow lumen are lined with tall columnar epithelial cells with PAS positive brush border on apical surfaces. The distal convoluted segments with round and wide lumen, lined by low columnar epithelial cells, and the apical surface is PAS negative (Mochizuki et al., 2005). Interstitial lymphomyeloid cells have large nuclei, stained positive with hematoxylin, and filled the gaps of renal tubules. Brown and blackish pigments, probably the melanomacrophages and melanomacrophage centers (MMC) are occasionally seen in interstitial lymphoid tissues (Mochizuki et al., 2005; Steinel and Bolnick, 2017; Stosik et al., 2019). Compared to controls, administration of GO (25–200 μg/g) was unable to induce any significant alteration in the histopathology of glomerulus, and renal tubules either in males (Figures 5B1, 5C1, SF9A1–2) or females (Figures 5B2, 5C2, SF9B1–2) as observed by low microscopy. Moreover, we have observed hyperplasia in interstitial lymphoid tissues (HIL) mostly at the cranial region of kidneys of some male and female fish injected with GO (Figures SF9C1–2). The kidneys showing HIL became basophilic due to increase in interstitial cell populations. Moreover, the lumen of the tubules in HIL region became compact and in some cases, the cilia of the epithelial cells probably detached from the cell surfaces and clumped at the center that reduced the tubular passages, even though they (cilia) remained undetectable in HE staining (SF9C1–2). In some fish, the renal tubules were tightly packed, seldom showing any interstitial lymphomyeloid cells in between them. However, the reproductive activities of these fish having HIL were identical with other GO-injected fish (25–200 μg/g).</p><!><p>As a sequel of our previous investigations on evaluation of toxic effects of GO in mammalian model (Patlolla et al., 2016; 2017), we continued our studies in a fish model and evaluated the toxicity and the EDC effects of GO. We used Japanese medaka fish, a recognized model for characterization of EDCs (Hinton et al., 2005; Urushitani et al., 2007; Scholz and Mayer 2008; Dang, 2016; OECD 2018), for the research. Several physiological and molecular markers in medaka including hepatic vitellogenin, testis-ova, altered morphology of anal and dorsal fins, genetic sex identification by polymerase chain reaction (PCR), are useful parameters which can be considered for identification and characterization of EDCs disposed in the aquatic environment (Urushitani et al., 2007; OECD 2018; Dang and Kienzler, 2019).</p><p>In our previous studies on rats, the GO we used was obtained from commercial source, and we administered multiple doses of GO orally and repeatedly to the experimental animals (Patllola et al., 2016; 2017). However, in the present experiment, we synthesized GO in our laboratory by modified Hummer's method (Viraka Nellore et al., 2015), injected ip to the sexually mature, reproductively active, intact adult male and female fish at the doses of 25–200 μg/g of body weight (some fish were also injected with GO obtained from commercial source). The experimental fish as a pair (one male and one female) bred successfully for 7 days (one week) before injection and continued breeding for another 21 days(three weeks) after injection. The collected eggs were used for the estimation of fecundity (Figures 1A–B) and the developmental disorders (Supplementary Figures SF5A–F). After 21-days post-injection, the survived fish were sacrificed and the gonads (testis and ovary), livers, and kidneys were used for histolopathological analysis.</p><p>The physicochemical properties of the synthesized GO as verified by SEM,TEM, AFM FTIR spectra and zeta potentials (SF1A–B, SF2A–C; Supplementary Table 1) showed excellent similarities with the results reported previously by us (Patlolla et al., 2016; 2017) as well as by other investigators (Chen et al., 2016; Paital et al., 2019). We observed after sonication GO was agglomerated in the water over time, however, because of the negative value of zeta potential (−30.5 mV) the solution was considered stable (Patlolla et al., 2017). Although the conventional method of exposure to nanomaterials in fish is immersion, direct administration of GO or EDCs by ip-injection(s) have also been reported (Zhang et al., 2002; Kiparissis et al., 2003; Dang 2016; Dorelle et al., 2017; Movahedinia et al., 2018;Dang and Kienzler, 2019; Paital et al 2019). To avoid agglomeration of GO in the environment, we preferred to use ip injection as an alternative method, because the route of administration of nanomaterials has significant impact on the retention and biological function of the compounds (Ema et al., 2016). The ip-route of GO was inadequately investigated in fish but seems to be safe, convenient, rapid, and less stressful than intravenous (IV) injections used in mammals (Kurantowicz et al., 2015), which is not very convenient in a small laboratory fish like medaka.</p><p>In our research, the fish injected with GO were able to maintain normal breeding and produced both fertilized and unfertilized eggs (Figures 1A–B) with dose-dependent increase of body weight in female and remained unaltered in males (Figure SF4). Moreover, with these limited data, we have documented that the males are more sensitive to GO than females (calculated LD50 for males is 175.39 and for females are 2,901.20 μg/g) (Figures SF3A–B). Our observations on body weights (Figure SF4) are similar with the observations made by Ma et al (2016) who have exposed adult tilapia fish to GO (total 4 mg GO fed for 30 days) and observed non-significant increase in body weight of the experimental fish (p=0.668, n=6) without affecting the growth and behavior of the fish. Further, the accumulated GO has the potential to agglomerate in the body fluids that can affect the incorporation of GO into specific organs. However, we have observed noticeable accumulation of unidentified black particles (probably GO or its metabolites) in the proximity of the injection site, especially between the connective tissues of the abdominal skin, muscles, peritoneal membrane, and in gonads (ovary and testis) after 21 days post-injection (Figures 2B2, 2C1–3, SF6A–C and SF7). In Wistar rats, ip-injection of GO nanoparticles retained in the body, mostly as agglomerates, and the largest agglomerate appeared to be 10 mm in diameter, which are mostly seen at the proximity of the injection site (Kurantowicz et al., 2015). We therefore think that the animal model (Japanese medaka) we used, and the mode of administration (ip injection) we followed, provide a new insight into the toxicological and endocrine disrupting potential of GO.</p><p>Our data on fecundity apparently show a dose-dependent reduction by GO at least in the early days of post-injection (Figures 1A–B). However, the statistical analysis of the data (fecundity) was unable to show significant difference between the water-injected controls and various GO-treated groups at all-time points; either as total eggs (fertilized and unfertilized) or as fertilized eggs (25–200 μg/g, single ip injection). Considering the impact of GO on gonadal activity, we hypothesize that the egg laying activity of the fish is a functional endpoint of the ovarian function/activity of the female fish (Figure 1A) and the fertilized eggs (Figure 1B) found in post mating period represent the testicular activity/function of the male fish. Although our data are unable to establish a significant difference, available reports indicate that in other models such as Caenorhabditia elegans chronic exposure to GO was able to reduce fecundity, and size of the animals; however, acute exposure failed to do that (Rive et al., 2019). In Spodoptera frugiperda, the diet containing 1mg/g GO during larval rearing, reduced fecundity and fertility of the insects (Martins et al., 2019). In house cricket (Acheta domesticus) multigenerational exposure to GO reduced fecundity and induced cellular damage (Dziewiecka et al., 2018). Moreover, our data on Japanese medaka indicate that fecundity increased gradually with time even though it did not reach up to the same level observed during pre-injection periods (Figures 1A–B). We expect that despite the deposition of GO agglomerates in the different organs of medaka especially in gonads (Figures 2C1–3, 3C1, SF6B–E, SF7B–C), there is a possibility of depuration of GO over time as observed in other EDC studies (Miller et al., 2012). The fertilizing activity of the male medaka reduced significantly during 14 days exposure to EE2, a xenoestrogen and a known EDC disposed in the environment, and the fish were unable to recover completely to the pre-exposure period during depuration probably due to the impairment of the male reproductive capacity by the xenoestrogen (Miller et al., 2012). Although our data on fecundity were unable to document GO as an indicator of EDC, we hypothesize that the injected GO may have the ability to impair gonadal functions and will be considered as a potential EDC disposed in the environment.</p><p>Reproductive activities were further verified on the hatchability of the fertilized eggs laid by the female fish before and after GO injection (Figure SF5B). We hypothesize that the injected materials (water or graphene oxide) are adsorbed into the ovary/testis and consequently into the gametes (eggs/sperms) that can affect either the fertilizing ability (remained unfertilized) of the gametes (sperms and ovum), or the morphogenesis of the embryos during post-fertilization periods, possibly due to transgenerational inheritance (Pooma et al., 2014). If the gametes are unable to be released into the environment for external fertilization, they will be degenerated within the gonads (follicular atresia or degeneration of germinal components). Although we are unable draw a definite conclusion on the effects of GO on embryonic development of medaka exposed through parenteral administration, available reports indicate that in house cricket exposed to GO throughout life cycle show a significantly decreased their reproductive capabilities. Moreover, there are possibilities to produce multigenerational harmful effects on this species (house cricket) by GO (Dziewiecka et al., 2018). In zebrafish embryos or larvae, exposed to GO either by microinjection or waterborne, remained unresponsive or induced several morphological disorders including yolk sac and pericardial edema, abnormal tail flexures, malfunction of heart and cardiovascular system, hatching delay, enhanced mortality and others in a concentration dependent manner (Dasmahapatra et al. 2019). In mammals, oral administration of GO to lactating mothers induced growth retardation of offspring; moreover, intravenous injection of rGO during late gestation period caused maternal death and abortion in mice (Fu et al., 2015; Xu et al., 2015; Ema et al., 2016). However, no adverse effects of S-GO (55 nm) or L-GO (238 nm) were observed on male reproduction in mice injected either intravenously or intraperitoneally (Liang et al., 2015). We injected GO once to both male and female fish, which make difference with others. Our observations in the reproductive activities of medaka and on developing embryos and larvae are found to be very minimal compared to the observations made by others in other animal models like insects (Dziewiecka et al., 2018), zebrafish (Dasmahapatra et al., 2019) or in mammals (Fu et al., 2015; Liang et al 2015; Ema et al., 2016).</p><p>Our studies further extend to evaluate the histopathological changes in gonads (ovary and testis), liver and kidney of the experimental fish. Histopathologic assessment of the gonads is a valuable endpoint in EDC studies (Hutchinson et al., 2006). Generally, estrogenic (EE2) or androgenic (TB) chemicals have the potential to induce sex skewing and producing intersex phenotypes targeting estrogen, androgen or steroidogenic pathways (Dang 2016; Dang and Kienzler, 2019). In medaka spontaneous induction of intersex gonad is a very common phenomenon depending on culture conditions, breeding colonies used, and other factors (Grim et al., 2007; Miller et al., 2012; Cheung et al., 2014); however, despite severe intersex gonads, medaka males are still capable of successful reproduction (Flynn et al., 2017) and probably show no impact on fecundity. EDC can also cause malformation of accessory structures such as oviducts in females or efferent ducts in the testis of males that can adversely affect reproduction (Flynn et al., 2017).</p><p>In the present experiment, histopathological evaluation of the ovary or testis of the experimental fish after 21 days post-injections of different doses of GO did not show any testicular tissues in ovary or ovarian components in testis (sex reversal) of the experimental fish. Therefore, we conclude from our data that the medaka colony we maintained in the laboratory, or the experimental protocol we followed to evaluate GO as an EDC, are unable to induce sex reversal in medaka, either spontaneously (Figure 2A1 and 3A1) or after GO injections (25–200 μg/g) (Figures 2B1, 2C1, 3B1, 3C1, SF6A–B and SF7A–B). The deposition of unidentified black particles (probably the agglomerated GO), 21-days post-injections on gonads, ovarian follicles, or in the adjacent periphery of the germinal epithelium of ovary and testis, indicate that the injected GO agglomerates into the gonadal tissues of medaka in a dose-dependent manner (Figures 2C1, 3C1, SF6B–E, SF7B–C). Our data further indicate that despite the presence of GO inside the gonads, the gametogenesis process is continuing in ovary (Figure SF6C) and testis (Figure SF7C). Moreover, the existence of both differentiating and degenerating follicles (Figures 2B1, 2C1, SF6A–B) in the stromal compartments of the ovary, and substantive number of Sertoli and Leydig cells in the testis (Figures 3A–C, SF7A–C), indicate that any histological damages induced by GO are limited to the regions where GO agglomerates are accumulated within the organs. Alternatively, the organs are able to recover from GO-induced damages, if any, within 21-days post injection periods. Our observations on fecundity (Figures 1A–B) indicate that the fish received higher doses of GO (100–200 μg/g) are able to produce fewer eggs (both as total and fertilized) in the early part of post-injections (first 7 days or more) than the eggs produced in the late part of the experiment (14–21 days). Moreover, our data on follicular counts (Figures 2D and 2E), except 50 μg/g groups, are also unable to establish significant differences between vehicle-treated controls and GO-treated (25, 100, and 200 μg/g) fish. Furthermore, we observed that the ovaries themselves, tended to eliminate agglomerated GO from the ovarian compartments probably by an invagination of dorsal ovarian epithelium (Nakamura, 2018), originated from the intact ovarian structures (Figure SF6C). Therefore, we predict that the stem cells in the germinal epitheliums of ovary and testis of medaka are still active in the fish treated with GO, and the presence of GO inside the gonads are unable to prevent their functions, completely. Reports document that structural alteration of the seminiferous tubules and interstitium observed in male Wistar rats exposed to a high dose of GO (10 mg/kg BW, every alternative day for 15 days) (Nirmal et al., 2017). However, in ICR-strain male mice nanoscale GO showed no toxicity with regard to histopathological changes in testis and epididymis (Liang et al., 2015). No histopathological changes were observed in the testis of BALB/c mice after single intravenous injection of GO (Qu et al., 2013). Moreover, single-walled carbon nanotubes (SWCNT-COOH) and rGO did not affect human sperm viability in vitro even though SWCNT-COOH generated significant super oxide species (Asghar et al., 2016). However, the metabolites of TB (androgenic in nature) reduced the primary ovarian follicles (similar to PNOs in our studies) and increased vitellogenic follicles (similar to EVOs and LVOs in our studies) in medaka ovary (Forsgren et al., 2014). Despite the observations made by other investigators, with our present experimental data, we have reason to believe that the destructive effects of GO (probably by necrosis), if any, was either restricted only in the GO accumulated sites of the organ, or the fish were able to recover the damages induced by GO agglomerates within 21 days of depuration.</p><p>In teleost, PFC of the ovarian follicles (theca and GC) are the hormone producing cells (estrogen and progesterone), and the Leydig cells in the testes secrete male hormones (testosterone). Exposure to an EDC (TB) has been associated with an increase in the height and number of GCs, which give PFC a pseudostratified columnar appearance (OECD 2010). Moreover, hyperplasia in Leydig cells and hypertrophy of Sertoli cells are also significant indicators of EDC effects in testis of fish (OECD 2010). GO has been documented as an inducer of size-dependent toxicity on Sertoli cells (TM4) and Leydig cells (TM3) in vitro (Gurunathan et al., 2019), and atrophy of seminiferous tubules and damaged interstitium by nano GO in Wistar rats in vivo (Nirmal et al, 2017). Although we did not measure the endogenous hormone levels (estrogen or testosterone), from the cellular morphology of GC layers (Figures 2B2–3, 2C2–3, SF6A–B) and Leydig cells (Figures 3B1, 3B3, 3C1, 3C3 SF7A–C), by light microscopy studies, it is very difficult to predict the hormonal status of the fish after GO-injections. The morphology of GC in the follicles adjacent to GO agglomeration sites were altered (multiple layer) expecting a hormonal alteration, however, the data on follicular counts (Figure 2E), except 50 μg/g, did not support the hypothesis. We predict that steroid hormones produced by GC or Leydig cells of Japanese medaka remained unaltered or at the same level after GC injections as in controls. Alternatively, any alterations in GC or Leydig cell morphology, induced by GO in the earlier phase of treatment, could recover during the late part of the treatment either by depuration or by other means.</p><p>The toxic effects of GO on liver and kidney of fish were reported by many investigators (Chen et al., 2016; Souza et al; 2017; Dasmahapatra et al., 2019 (review); Paital et al., 2019). To our knowledge, it is unknown to us, whether GO is able to induce any EDC effects in liver and kidneys of medaka. In the present study, we did not observe any histopathological variation in the liver (Boorman et al., 1997; OECD 2010) or kidney (OECD 2010) of male and female medaka after 21 days post-injection of GO. The hepatocytes in male fish due to the activation of vitellogenin gene by estrogenic EDCs, can turn the cytoplasm of the hepatocytes basophilic. Consequently, EDCs are able to modulate the basophilicity of the hepatocytes in female fish, possibly by enhancing the expression of vitellogenin or other genes. Moreover, alteration in cytoplasmic vacuolization, and accumulation of proteinaceous fluid in the hepatic blood vessels were significant indicators of EDC activities, especially in fish (OECD, 2010). In the present experiment, we did not measure the vitellogenin or other protein contents of medaka in the liver, or serum. However, our histopathological observations are able to document a marked difference in basophilia of the hepatocytes between male (Figure 4A1) and female fish (Figure 4A2), and unable to establish remarkable difference between control and GO-treated male (Figures 4B1, 4C1, SF8A1, SF8B1) or female (Figures 4B2, 4C2, SF8A2 and SF8B2) fish. The accumulation of proteinaceous fluid in some male and female fish (Figures SF8A1–2, and SF8B2) are probably the blockage of the hepatic vessels by GO.</p><p>EDCs, which are estrogenic in nature, are able to induce renal impairments in kidneys of fish including swelling of tubular epithelial cells and necrosis, dilution of Bowman's capsule, interstitial fibrosis, casts, and hyaline droplets in tubules or glomerulus (OECD, 2010). However, our light microscopy studies are unable to document any of these changes in the glomerulus, tubules, either in vehicle-injected control male (Figure 5A1) or female (Figure 5A2) or in GO-treated male (Figures 5B1, 5C1, SF9A1–B1), and female (Figures 5B2, 5C2, SF9A2–B2) fish. Although we are unable to claim the incorporation/agglomeration of GO nanoparticles in the kidneys, we observed substantive number of black particles in the head kidneys of both control and GO-treated fish, probably MMC, within the interstitial lymphomyleyoid cells found in medaka kidneys (Mochizuki et al., 2005). MMCs are aggregates of pigmented phagocytes found primarily in poikilotherm lymphoid tissues including liver, spleen and kidneys. We did not see similar MMC-like components in liver of both male and female medaka; however, we observed substantive number of MMC-like structure in spleen of GO-injected male and female medaka (data not shown). MMCs play significant role in mediating immunologic disorders, and we are confused with the GO (which were also black in color), we did not pay any attention to MMCs we observed in kidneys. Moreover, the presence of HIL in kidneys suggests a probable immune-related mechanism associated with GO. Immunomodulatory effects of GO was studied in both mammalian models (Ema et al., 2016) as well as in adult zebrafish (Dasmahapatra et al., 2019, review). Expect HIL, the histopathological studies in kidneys of both male and female fish do not indicate any functional overload of vitellogenin, or sppigin (an EDC-sensitive renal glue protein found in male three-spined stickleback, Gasterosteus aculeatus, during breeding season, Hahlbeck et al., 2004, Jolly et al., 2009), or any other proteins (seen in livers of some GO-injected fish) induced by GO.</p><!><p>Taken together, our data indicate that single ip injection of GO (25–200 μg/g) to reproductively active intact adult medaka male and female show significant accumulation of GO agglomerates in the gonads after 21 days post-injections. However, as observed by light microscopy, administration of GO, ip (25–200 μg/g), is unable to disrupt significant reproductive activities of medaka as evidenced by everyday fecundity and from the histopathological evaluation of gonads (ovary and testis), liver and kidneys. Although we did not study the oxidative-stress mediated mechanisms, the damaging effects of GO we observed in gonads and livers are probably mediated through necrosis. HILs observed in kidneys indicate the immunological response of the organ to GO. To evaluate the EDC activities of GO, more studies at the cellular levels including biochemical, immunological and molecular biological approaches are necessary. Moreover, transgenerational studies will be an important endpoint for the evaluation of EDC effects of GO in aquatic models such as fish.</p>

PubMed Author Manuscript

Synthesis and characterisation of thiobarbituric acid enamine derivatives, and evaluation of their α-glucosidase inhibitory and anti-glycation activity

AbstractA new series of thiobarbituric (thiopyrimidine trione) enamine derivatives and its analogues barbituric acid derivatives was synthesised, characterised, and screen for in vitro evaluation of α-glucosidase enzyme inhibition and anti-glycation activity. This series of compounds were found to inhibit α-glucosidase activity in a reversible mixed-type manner with IC50 between 264.07 ± 1.87 and 448.63 ± 2.46 µM. Molecular docking studies indicated that compounds of 3g, 3i, 3j, and 5 are located close to the active site of α-glucosidase, which may cover the active pocket, thereby inhibiting the binding of the substrate to the enzyme. Thiopyrimidine trione derivatives also inhibited the generation of advanced glycation end-products (AGEs), which cause long-term complications in diabetes. While, compounds 3a–k, 5, and 6 showed significant to moderate anti-glycation activity (IC50 = 31.5 ± 0.81 to 554.76 ± 9.1 µM).

synthesis_and_characterisation_of_thiobarbituric_acid_enamine_derivatives,_and_evaluation_of_their_α

Introduction<!>Synthesis of the target compounds<!><!>Biological activity<!><!>Biological activity<!><!>Molecular docking studies<!><!>Molecular docking studies<!><!>Molecular docking studies<!><!>Conclusions<!>General methods<!>General procedure for the synthesis of 3a–k, 4a–d, 5, and 6<!>1,3-Diethyl-5-(((2-morpholinoethyl)amino)methylene)-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3a)<!>1,3-Diethyl-5-(((pyridin-2-ylmethyl)amino)methylene)-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3b)<!>1,3-Diethyl-5-(((4-methylpyridin-2-yl)amino)methylene)-2-thioxodihydro pyrimidine-4,6(1H,5H)-dione (3c)<!>4-(((1,3-Diethyl-4,6-dioxo-2-thioxotetrahydropyrimidin-5(2H)-ylidene)methyl)amino)-N-(pyrimidin-2-yl)benzenesulfonamide (3d)<!>1,3-Diethyl-2-thioxo-5-((p-tolylamino)methylene)dihydropyrimidine-4,6(1H,5H)-dione (3e)<!>5-(((4-Chlorophenyl)amino)methylene)-1,3-diethyl-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3f)<!>5-((Cyclohexylamino)methylene)-1,3-diethyl-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3g)<!>5-(((2-Chlorophenyl)amino)methylene)-1,3-diethyl-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3 h)<!>1,3-Diethyl-5-(((2-iodophenyl)amino)methylene)-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3i)<!>2-(((1,3-Diethyl-4,6-dioxo-2-thioxotetrahydropyrimidin-5(2H)-ylidene)methyl) amino)benzenesulfonic acid (3j)<!>1,3-diethyl-5-((pyridin-2-ylamino)methylene)-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (3k)<!>5-(((4-Chlorophenyl)amino)methylene)-1,3-dimethylpyrimidine-2,4,6(1H,3H,5H)-trione (4a)<!>5-(((2-Chlorophenyl)amino)methylene)-1,3-dimethylpyrimidine-2,4,6(1H,3H,5H)-trione (4b)<!>5-(((2-Iodophenyl)amino)methylene)-1,3-dimethylpyrimidine-2,4,6(1H,3H,5H)-trione (4c)<!>1,3-Dimethyl-5-((pyridin-2-ylamino)methylene)pyrimidine-2,4,6(1H,3H,5H)-trione (4d)<!>5-(((4-(((1,3-Dimethyl-2,4,6-trioxotetrahydropyrimidin-5(2H)-ylidene)methyl) amino)benzyl)amino)methylene)-1,3-dimethylpyrimidine-2,4,6(1H,3H,5H)-trione (5)<!>5-(((4-(((1,3-Diethyl-4,6-dioxo-2-thioxotetrahydropyrimidin-5(2H)-ylidene)methyl)amino)benzyl)amino)methylene)-1,3-diethyl-2-thioxodihydropyrimidine-4,6(1H,5H)-dione (6)<!>Protocol for in vitro α-glucosidase inhibition assay<!>Protocol for anti-glycation assay40,41<!>Calculation of inhibitory activity<!>

<p>Diabetes mellitus (DM) is a disease which caused by a breakdown of carbohydrate metabolism, which plays a significant role in the development of long-term diabetic complications. According to the International Diabetes Federation, 693 million people will suffer from this condition by 20451,2. DM can be categorised into three types: Type I (T1DM); Type II (T2DM); and gestational (GDM). About 80 − 90% of all DM patients are Type II (T2DM). Drug treatments of T2DM aim to decrease hepatic glucose production, enhance insulin action, and boost insulin secretion from β-pancreatic cells, or block α-glycosidase enzyme (carbohydrate digestive enzymes)3–6. Therapeutic in individuals with this disease may lead to various complications, including kidney disease, disorders of the nervous system, leg amputation, heart disease and severe retinopathy up to blindness7.</p><p>Carbohydrate digestive enzymes are found in the brush border of the intestine. They catalyse the breaking down long-chain polysaccharides into absorbable monosaccharide units. Of these enzymes, α-glucosidases, which play a key role in the digestion and absorption of complex carbohydrates, and has emerged as target to maintain postprandial blood glucose control. α-Glucosidase inhibitors currently used to treat T2DM include acarbose (Precose), voglibose, and miglitol8. However, these drugs are associated with several side effects, such as flatulence, stomach-ache, diarrhoea, and liver damage9. Therefore, an increasing interest in exploring new drug candidates for glycosidase inhibition is needed10–12.</p><p>Barbituric acid (BA) derivatives have been reported to have potential anti-hypertensive13, anti-cancer14, anti-convulsant15, anti-inflammatory16, anti-psychotic17, and antitumor properties18–21. Recently, these derivatives have also been reported as anti-diabetic agents22. On the other hand, thiobarbituric acid (TBA) analogues has been described to exert anti-inflammatory16,23, immunotropic24, anticonvulsant25, and anti-hypnotic25,26, anti-neoplastic27, and antitumor activities28. De Belin et al.29 reported a number of TBA derivatives as inhibitors of hypoxia-inducible factor 1 (HIF-1). Recently, Barakat et al.30 described the synthesis of a new series of diethylammonium salts of aryl substituted TBA derivatives as α-glycosidase inhibitors. Therefore, given the relevance of TBA derivatives in medicinal chemistry, the design of new molecules containing the thiobarbituric moiety is an inspiring goal.</p><p>In continuation of our studies on the synthesis of biologically active compounds22,30,31, herein, we synthesised 1,3-diethylthiobarbiturate enamine derivatives and evaluated their in vitro α-glucosidase inhibitory and anti-glycation activities. In addition, molecular docking studies were performed to study the interactions of the compounds with the catalytic site of the enzyme using acarbose and evaluated their α-glucosidase inhibition capacity and the anti-glycation properties.</p><!><p>Enamine derivatives 2a32 and 2b33 were prepared by reacting the commercially available compounds, 1,3-diethylthiobarbituric acid 1a or 1,3-dimethylbarbituric acid 1b with DMF in the presence of acetic anhydride as solvent for 2 h at 90 °C to afford 2a and 2b, respectively as a yellow crystalline solid in good yields. Compounds 2a32 and 2b33 were reacted with different amines in ethanol at RT to afford the target products 3a–k and 4a–d, respectively (Scheme 1) in excellent yields and purities, as observed from their spectral data. The reaction of 2a (2 equiv.) or its analogues 2b (2 equiv.) with the commercially available material 4-(aminomethyl)aniline (1 equiv.) under the same conditions described above gave the dimeric products 6 and 5, respectively as shown in Scheme 1. The structures of the products obtained were deduced by 1H- and 13C-NMR spectra (Supplementary material).</p><!><p>Synthetic route for the synthesis of 3a–k, 4a–d, 5, and 6.</p><!><p>All the synthesised derivatives of TBA (3a–k), BA (4a–d), and the dimeric analogues 5 and 6 were evaluated for their capacity to inhibit α-glucosidase and protein glycation in vitro in comparison to acarbose (IC50 = 875.75 ± 2.08 µM) and rutin (IC50 = 54.59 ± 2.20 µM), as standard tested compounds (Table 1).</p><!><p>Result of in vitro α-glucosidase enzyme inhibitor and anti-glycation activities.</p><p>Significant activity.</p><p>Moderate activity.</p><p>ND: not determined; NA: not active.</p><!><p>The results summarised in Table 1 indicated that all the N,N′-dimethylbarbituric-based enamine acid derivatives 4a–d were completely inactive, while compounds 3a–k, 5, and 6 showed moderate to significant activity against protein glycation (IC50 = 554.76 ± 9.1 to 31.5 ± 0.81 µM). The dimeric moiety of TBA 6 via diaminobenezene linkage (IC50 = 31.5 ± 0.81 µM, Table 1) was the most protein glycation inhibitor in this series of compounds, and showed more activity than the standard rutin (IC50 = 54.59 ± 2.20 µM). While, the dimeric analogues of BA 5 (IC50 = 554.76 ± 9.1 µM) was the least active.</p><p>On the other hand, substituted phenyl with an electron-withdrawing group such as a chlorine atom (a weak deactivating group) at the ortho position, showed a better anti-glycation activity than the same atom at the para position. Therefore, the change in the position had a remarkable effect on the anti-glycation activity34 (3 h vs 3f, Table 1). Halogen with a higher atomic weight and weaker electron-withdrawing effect, such as iodine at the ortho position, decreased the activity as compared to chlorine at the same position (3i vs 3 h). This observation could be attributed to the negative inductive effect35,36. In contrast, a strong electron-withdrawing group, such as sulphonic acid at ortho position, decreased the activity compared to chlorine and iodine in the ortho position (3j vs 3 h). Electron-donating group such as methyl (a weak donating group) at the para position yielded slightly better and a moderate activity as compared to the chlorine at the same position (3e vs 3f). On the other hand, replacing the 4-methylphenyl 3e by 2-pyridylmethylene 3b or 3-methylpyridyl 3c decreased the anti-glycation activity, and showed a comparable activity to compounds 3g and 3f as shown in Table 1. Compound with pyrimidine benzenesulfonamide 3d moiety decreased the activity, which is consistent with the result obtained for 3j with a strong withdrawing group. While, compounds with 2-morpholinoethyl 3a and cyclohexyl 3g moieties showed moderate activity against protein glycation.</p><p>The results summarised in Table 1 indicated, once again, that none of the BA enamine derivatives showed any activity, while 3g, 3i, 3j, and 5 exerted a significant activity against α-glucosidase (IC50 = 264.07 ± 1.87 to 448.63 ± 2.46 µM). Of the series of compounds, thiopyrimidine trione derivative with higher atomic weight halogen, such as iodine at the ortho position, was the most active, exhibiting 3.3-fold higher activity than the standard acarbose. Compounds with a cyclohexyl ring 3g, sulphonic acid 3j, and the dimeric analogue of BA 5 showed twice the activity of the standard drug. The rest of the compounds did not show any activity.</p><p>Finally, the most two active compounds from the series are shown in Figure 1. In conclusion, this work has demonstrated that the core of TBA-based enamine derivatives is a privileged structure for anti-glycation and α-glucosidase inhibition and thus deserves further investigation.</p><!><p>Lead compounds 3i and 6 with promising activities.</p><!><p>Molecular docking provides significant insight into ligand-protein binding modes and mechanisms. Here, molecular docking studies were carried out to explore the binding modes of TBA derivatives with a notorious α-glucosidase, such as that of Baker's yeast (Saccharomyces cerevisiae). We used our previously built homology model of α-glucosidase from the template (PDB ID: 3A4A)30. Initially, the 3 D structures of all the ligands were built, protonated, and minimised by means of the MMFF94x force field37, and using the molecular operating environment (MOE)38 2018.04. All recently synthesised TBA derivatives and a reference inhibitor (acarbose) were docked into the active site of the receptor using the default parameters in MOE. Each complex was visually analysed for ligand–protein interactions, and their images were prepared using UCSF chimaera software39.</p><p>The top ranked conformer of TBA derivatives and standard (acarbose) were selected based on docking score. The docking score of the ligands 3g, 3i, 3j, and 5 and acarbose were −3.081, −4.909, −5.19, −5.642, and −4.382, respectively. The docking study revealed that the acarbose, and all the ligands accommodated into the binding pocket of the C-terminal domain of α-glucosidase. The clustering of standard and synthetic compounds at the allosteric site of the C-terminal domain is shown in Figure 2.</p><!><p>Binding mode of thiobarbituric acid derivatives into the α-glucosidase binding cavity. For clarity, acarbose is shown in cyan. Compounds 3g, 3i, and 3j are indicated in pink, and 5 in green. The part of the enzyme in the background is shown as surface model.</p><!><p>Acarbose occupied a large cavity in the binding sites of α-glucosidase due to its larger size, as compared to the synthetic compounds. The oxygen functionality of acarbose formed two hydrogen bonds with the active site residues, Arg212 and Arg439. Ring structures were involved in the π–π interactions with Phe177, His239, and Pro309. Moreover, residues Glu276, Glu304, and Asp349 interacted hydrophobically with the ligand (Figure 3).</p><!><p>Interactions of acarbose with crucial residues of α-glucosidase.</p><!><p>The carbonyl oxygen of the thiobarbituric ring of 3g, 3i, and 3j showed hydrogen bond interactions with crucial residue Arg212. Another hydrogen bond was observed between the nitrogen atoms of 3g with Thr215. These compounds were further stabilised through π–π interactions with the crucial active site residues Tyr71, Phe157, and Phe177. Additionally, π–π interactions were observed with the thiol ring of 3j through Phe177 and Tyr71. In the case of compound 3i, Phe157 was involved in forming halogen–π interactions. Moreover, hydrophobic interactions with the active site residues Phe157, Thr215, Leu218, and Arg349 stabilised these compounds. In the case of 5, the 2,4,6-trione ring-bearing oxygen atom formed hydrogen bonds with His111 and Arg212. Meanwhile, the amine functionality of the ligand also formed a hydrogen bond with residue Arg312. The benzene ring was involved in π–π interactions with Phe157 and Phe300. The hydrophobic interaction with crucial residue Arg349 also contributed to the binding of 5 with α-glucosidase. The interaction diagrams of all the ligands are shown in Figure 4. The docking results of 5 were in good agreement with experimental results, thereby indicating that it could be a good candidate as α-glucosidase inhibitor.</p><!><p>The predicted binding interactions of compounds 3g, 3i, 3j, and 5 in the active site.</p><!><p>Several derivatives of barbitutic and thiobarbituric enamine derivatives were synthesised, characterised, and screened for in vitro evaluation of α-glucosidase enzyme inhibition and anti-glycation activity. The results reveal that the four monomeric compounds 4a–d derived from N,N′-dimethylbarbituric enamine derivatives showed no anti-glycation activity, while compounds derived from N,N′-diethylthiobarbituric enamine derivatives 3a–k exhibited moderate activity against protein glycation with IC50 in the range 70–550 µM. The most potent anti-glycation activity was showed by the dimeric product from N,N′-diethylthiobarbituric enamine 6 with an IC50 of 31.5 µM, while the dimeric analogue of N,N′-dimethylbarbituric enamine 5 showed less activity with an IC50 of 554.8 µM. The reported series of compounds were found to inhibit α-glucosidase activity in a reversible mixed-type manner with IC50 between 264 and 448 µM. The type and position of substituent on phenyl ring (enamine moiety) has great impact on the biological activity. In this regard, the moderate electron-withdrawing group, such as a chlorine atom at the ortho position 3 h showed greater activity compared to the same atom at the para position 3f. On the other hand, the presence of iodine at the ortho position decreased activity compared to chlorine in the same position (3i vs 3 h). The strong electron-withdrawing group, such as sulphonic acid showed decrease in activity compared to weak electron-donating group like methyl (3e).</p><p>The dimeric thiobarbituric derivative 6 showed better anti-glycation activity compared with the standard rutin, while thiobarbituric ortho-iodo-enamine derivative 3i showed a positive effect as α-glucosidase inhibitor compared to the standard acarbose.</p><p>Molecular docking studies indicated that compounds of 3g, 3i, 3j, and 5 are located close to the active site of α-glucosidase, which may cover the active pocket, thereby inhibiting the binding of the substrate to the enzyme.</p><p>This work has confirmed that the core of (thio)barbituric-based enamine derivatives are a privileged structure, because in addition of the previous described biological activity, they have shown activity for anti-glycation and α-glucosidase inhibition.</p><!><p>All melting points were determined using Mel-Temp apparatus and are uncorrected. Thin layer chromatography (TLC) was performed on silica gel (Kiesel gel G, Merck) and spots were detected under UV light at 254 nm. FTIR Spectra were recorded in a KBr matrix on a Bruker Tensor 37 FTIR spectrophotometer. 1H-NMR spectra were recorded with a JEOL 400 MHz, 13C-NMR were recorded using the JEOL spectrophotometers, and the chemical shifts (δ) are given in ppm.</p><!><p>A solution of 2a32 or 2b33 (1 equiv.) was mixed with different amines (1 equiv.) in MeOH (10 ml) and stirred at room temperature for 10–120 min (TLC 20% EtOAc/n-hexane). The solvent was evaporated slowly, providing the corresponding solid products in excellent yields and purities.</p><!><p>Compound 3a was synthesised from 2a and 4–(2-aminoethyl)morpholine following the general procedure, affording the product as a yellow powder in 81% yield; mp 135 °C; IR (KBr, cm−1): 3420, 2999, 2960, 2908, 2870, 1624, 1591, 1456; 1H-NMR (CDCl3, δ, ppm): 10.60 (brs, 1H, NH), 8.23 (d, 1H, J = 14.8 Hz, CH=), 4.50 (m, 4H, 2CH2), 3.72 (q, 2H, CH2), 3.54 (m, 4H, 2CH2), 2.60 (m, 2H, CH2), 2.50 (m, 4H, 2CH2), 1.25 (m, 6H, 2CH3); 13C-NMR (CDCl3 δ, ppm): 179.1, 163.0, 161.2, 160.6, 93.0, 66.9, 57.6, 53.6, 47.1, 43.0, 42.3, 12.5, 12.4; LC/MS (ESI): 341.44 [M + 1]+; Anal. Calcd for C15H24N4O3S: C, 52.92; H, 7.11; N, 16.46; Found: C, 53.01; H, 7.25; N, 16.59.</p><!><p>Compound 3b was synthesised from 2a and 2-picolylamine following the general procedure, affording the product as a pink powder in 83% yield; mp 154 °C; IR (KBr, cm−1): 3215, 3045, 2958, 2908, 2866, 1614, 1598, 1544,1463; 1H-NMR (CDCl3 δ, ppm): 11.01 (brs, 1H, NH), 8.63 (d, 1H, J = 5.2 Hz, Ar-H) , 8.39 (d, 1H, J = 14.0 Hz, CH=), 7.72 (t, 1H, J = 8.8 Hz, Ar-H), 7.27 (d, 1H, J = 8.8 Hz, Ar-H), 7.23 (m, 1H, Ar-H), 4.76 (d, 2H, J = 3.6 Hz, CH2), 4.56 (m, 4H, 2CH2), 1.28–124 (m, 6H, J = 16.4 Hz, 2CH3); 13C-NMR (CDCl3 δ, ppm): 179.1, 163.0, 161.2, 160.7, 154.1, 150.3, 137.3, 123.5, 121.7, 93.5, 55.1, 43.0, 42.3, 12.5, 12.4; LC/MS (ESI): 319.40 [M + 1]+; Anal. Calcd for C15H18N4O2S: C, 56.59; H, 5.70; N, 17.60; Found: C, 56.72; H, 5.81; N, 17.78.</p><!><p>Compound 3c was synthesised from 2a and 2-amino-4-picoline following the general procedure, affording the product as a light yellow powder in 87% yield; mp 175 °C; IR (KBr, cm−1) 3215, 3157, 3045, 2958, 2908, 2866, 1614, 1598, 1544, 1463; 1H-NMR (CDCl3 δ, ppm):12.25 (d, 1H, J = 13.2 Hz, NH), 9.40 (d, 1H, J = 13.2 Hz, CH=), 8.27 (d, 1H, J = 5.2 Hz, Ar-H), 6.98 (d, 1H, J = 5.2 Hz, Ar-H), 6.86(s, 1H, Ar-H), 4.55 (m, 4H, 2CH2), 2.38 (s, 3H, CH3), 1.29 (m, 6H, 2CH3); 13C-NMR (CDCl3 δ, ppm):179.1, 163.3, 160.7, 152.5, 150.7, 149.6, 149.0, 122.9, 113.6, 95.8, 43.2, 42.5, 21.2, 12.5, 12.4; LC/MS (ESI): 319.40 [M + 1]+; Anal. Calcd for C15H18N4O2S: C, 56.59; H, 5.70; N, 17.60; Found: C, 56.81; H, 5.78; N, 17.79.</p><!><p>Compound 3d was synthesised from 2a and sulphadiazine following the general procedure, affording the product as a yellow powder in 85% yield; mp 204 °C; IR (KBr, cm−1): 3421, 3116, 2958, 2860, 1618, 1591, 1508, 1440; 1H-NMR (DMSO-d6, δ, ppm): 12.20 (d, 1H, J = 14.0 Hz, NH), 8.72 (d, 1H, J = 14.0 Hz, NH), 8.52 (d, 1H, J = 4.4 Hz, CH=), 8.47 (d, 1H, J = 8.8 Hz, Ar-H), 8.0 (d, 1H, J = 8.8 Hz, Ar-H), 7.79 (d, 2H, J = 8.8 Hz, Ar-H), 7.10 (m, 1H, Ar-H), 6.57 (d, 2H, J = 8.8 Hz, Ar-H), 4.42 (m, 4H, 2CH2), 1.21 (m, 6H, 2CH3); 13C-NMR (DMSO-d6, δ, ppm): 178.9, 162.2, 160.5, 158.8, 157.8, 157.3, 154.3, 153.6, 142.4, 130.4, 129.8, 125.4,119.8, 116.1, 112.7, 95.7, 42.9, 42.4, 12.8, 12.7; LC/MS (ESI): 461.53 [M + 1]+; Anal. Calcd for C19H20N6O4S2: C, 49.55; H, 4.38; N, 18.25; Found: C, 49.66; H, 4.50; N, 18.41.</p><!><p>Compound 3e was synthesised from 2a and 4-methylanline uracil following the general procedure, affording the product as a yellow powder in 89% yield; mp 139 °C; IR (KBr, cm−1): 3448, 3215, 3169, 2953, 2866, 1595, 1570, 1554, 1476, 1435, 1440; 1H-NMR (CDCl3, δ, ppm): 12.32 (d, 1H, J = 13.8 Hz, NH), 8.70 (d, 1H, J = 14.0 Hz, CH=), 7.27 (d, 2H, J = 8.0 Hz, Ar-H), 7.22 (dd, 2H, J = 8.0 Hz, Ar-H), 4.60(m, 4H, 2CH2), 2.36 (s, 3H, CH3), 1.30 (m, 6H, 2CH3); 13C-NMR (CDCl3, δ, ppm): 178.8, 163.2, 160.9, 152.9, 137.3, 135.5, 130.9, 118.2, 94.6, 43.2, 42.5, 12.4, 12.2; LC/MS (ESI): 317.41 [M + 1]+; Anal. Calcd for C16H19N3O2S: C, 60.55; H, 6.03; N, 13.24; Found: C, 60.32; H, 6.00; N, 13.43.</p><!><p>Compound 3f was synthesised from 2a and 4-chloroanline following the general procedure, affording the product as a yellow powder in 78% yield; mp 215 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (CDCl3, δ, ppm): 12.32 (d, 1H, J = 14.0 Hz, NH), 8.65 (d, 1H, J = 14.0 Hz, CH=), 7.38 (d, 2H, J = 8.0 Hz, Ar-H), 7.25 (d, 2H, J = 8.8 Hz, Ar-H), 4.55 (m, 4H, 2CH2), 1.30 (t, 6H, J = 7.9 Hz, 2CH3); 13C-NMR (CDCl3, δ, ppm): 178.9, 163.2, 160.8, 152.8, 136.6, 132.6, 130.4, 119.4, 95.3, 43.2, 42.5, 12.5, 12.3; LC/MS (ESI): 338.82 [M + 1]+; Anal. Calcd for C15H16ClN3O2S: C, 53.33; H, 4.77; N, 12.44; Found: C, 53.54; H, 4.80; N, 12.63.</p><!><p>Compound 3g was synthesised from 2a and cyclohexylamine following the general procedure, affording the product as a white powder in 84% yield; mp 105 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (DMSO-d6, δ, ppm): 10.61 (brs, 1H, NH), 8.25 (d, 1H, J = 15.6 Hz, CH=), 4.52 (m, 4H, 2CH2), 3.40 (m, 1H, CH), 1.98 (m, 2H, CH2), 1.83 (m, 2H, CH2), 1.78 (m, 2H, CH2), 1.47 (m, 2H, 2CH2), 1.28 (m, 2H, 2CH3); 13C-NMR (DMSO-d6, δ, ppm): 179.0, 163.2, 161.3, 158.4, 92.6, 59.3, 42.9, 42.3, 33.5, 24.9, 24.3, 12.5, 12.4; LC/MS (ESI): 310.43 [M + 1]+; Anal. Calcd for C15H23N3O2S: C, 58.23; H, 7.49; N, 13.58; Found: C, 58.39; H, 7.53; N, 13.78.</p><!><p>Compound 3h was synthesised from 2a and 2-chloroanline following the general procedure, affording the product as a yellow powder in 89% yield; mp 160 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (CDCl3, δ, ppm): 12.62 (d, 1H, J = 13.2 Hz, NH), 8.73 (d, 1H, J = 14.0 Hz, CH=), 7.49 (m, 2H, Ar-H), 7.39 (t, 1H, J = 7.9 Hz, Ar-H), 7.21 (t, 1H, J = 8.0 Hz, Ar-H), 4.56 (m, 4H, 2CH2), 1.30(m, 6H, 2CH3); 13C-NMR (CDCl3, δ, ppm): 178.9, 163.0, 160.9, 152.1, 135.3, 130.7, 128.5, 127.3, 125.0, 116.9, 96.1, 43.2, 42.6, 12.5, 12.4; LC/MS (ESI): 338.82 [M + 1]+; Anal. Calcd for C15H16ClN3O2S: C, 53.33; H, 4.77; N, 12.44; Found: C, 53.53; H, 4.92; N, 12.60.</p><!><p>Compound 3i was synthesised from 2a and 2-iodoanline following the general procedure, affording the product as a yellow powder in 83% yield; mp 175 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (CDCl3, δ, ppm): 12.42 (d, 1H, J = 13.2 Hz, NH), 8.67 (d, 1H, J = 14.0 Hz, CH=), 791 (d, 1H, J = 7.2 Hz, Ar-H), 7.45 (t, 1H, J = 7.9 Hz, Ar-H), 7.36 (d, 1H, J = 8.0 Hz, Ar-H), 6.98 (t, 1H, J = 7.2 Hz, Ar-H), 4.53 (m, 4H, 2CH2), 1.30 (m, 6H, 2CH3); 13C-NMR (CDCl3, δ, ppm): 179.0, 162.8, 160.9, 153.0, 139.9, 130.0, 128.2,125.1, 117.7, 95.9, 89.9, 43.2, 42.5, 12.6, 12.4; LC/MS (ESI): 430.28 [M + 1]+; Anal. Calcd for C15H16IN3O2S: C, 41.97; H, 3.76; N, 9.79; Found: C, 41.88; H, 3.81; N, 10.01.</p><!><p>Compound 3j was synthesised from 2a and 2-aminobenzenesulfonic acid following the general procedure, affording the product as a yellow powder in 80% yield; mp 243 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (DMSO-d6, δ, ppm): 13.01 (d, 1H, J = 14.8 Hz, NH), 8.62 (d, 1H, J = 14.4 Hz, CH=), 7.78 (d, 1H, J = 7.9 Hz, Ar-H), 7.64 (d, 1H, J = 8.0 Hz, Ar-H), 7.49 (t, 1H, J = 8.6 Hz, Ar-H), 7.30 (t, 1H, J = 8.4 Hz, Ar-H), 4.43 (m, 4H, 2CH2), 1.20 (m, 6H, 2CH3); 13C-NMR (DMSO-d6, δ, ppm): 178.9, 161.1, 160.9, 153.9, 138.6, 135.9, 130.7, 128.1, 124.6, 118.3, 95.3, 42.9, 42.2, 12.8. LC/MS (ESI): 384.44 [M + 1]+; Anal. Calcd for C15H17N3O5S2: C, 46.99; H, 4.47; N, 10.96; Found: C, 47.09; H, 4.53; N, 11.13.</p><!><p>Compound 3k was synthesised from 2a and pyridin-2-amine following the general procedure, affording the product as a yellow powder in 83% yield; mp 243 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (DMSO-d6, δ, ppm): 13.03 (d, 1H, J = 14.8 Hz, NH), 8.62 (d, 1H, J = 14.4 Hz, CH=), 7.76 (d, 1H, J = 7.6 Hz, Ar-H), 7.63 (d, 1H, J = 8.8 Hz, Ar-H), 7.49 (t, 1H, J = 9.6 Hz, Ar-H), 7.28 (t, 1H, J = 8.4 Hz, Ar-H), 4.46 (m, 4H, 2CH2), 1.22 (m, 6H, 2CH3); 13C-NMR (DMSO-d6, δ, ppm): 178.4, 160.5, 159.2, 138.1, 135.6, 130.7, 118.3, 94.6, 42.4, 41.7, 12.2. LC/MS (ESI): 305.35 [M + 1]+; Anal. Calcd for C14H16N4O2S: C, 55.25; H, 5.30; N, 18.41; Found: C, 55.38; H, 5.41; N, 18.59.</p><!><p>Compound 4a was synthesised from 2b and 4-chloroanline following the general procedure, affording the product as a white powder in 87% yield; mp 197 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409;1H-NMR (CDCl3, δ, ppm):12.00 (d, 1H, J = 13.6 Hz, NH), 8.60 (d, 1H, J = 14.0 Hz, CH=), 7.36 (d, 2H, J = 8.8 Hz, Ar-H), 7.16 (d, 2H, J = 8.8 Hz, Ar-H), 3.32 (s, 6H, 2CH3); 13C-NMR (CDCl3, δ, ppm): 165.1, 162.6, 151.8, 136.8, 132.1, 130.3,119.2, 93.4, 28.1, 27.4; LC/MS (ESI): 294.71 [M + 1]+; Anal. for C13H12ClN3O3; Calcd: C, 53.16; H, 4.12; N, 14.31; Found: C, 53.15; H, 4.12; N, 14.33.</p><!><p>Compound 4b was synthesised from 2b and 2-chloroanline following the general procedure, affording the product as a white powder in 87% yield; mp 202 °C; IR (KBr, cm−1): 3637, 3423, 3197, 2960, 2935, 1583, 1570, 1510, 1462; 1H-NMR (CDCl3, δ, ppm): 12.44 (d, 1H, J = 12.8 Hz, NH), 8.73 (d, 1H, J = 13.2 Hz, CH=), 7.48 (m, 2H, Ar-H), 7.36 (t, 1H, J = 7.2 Hz, Ar-H), 7.18 (m, 1H, Ar-H), 3.38 (s, 3H, CH3), 3.36 (s, 3H, CH3); 13C-NMR (CDCl3, δ, ppm): 164.9, 162.7, 151.9, 151.1, 135.4, 130.6, 128.5, 126.9, 124.7, 116.6, 94.4, 28.2, 27.5; LC/MS (ESI): 294.71 [M + 1]+; Anal. for C13H12ClN3O3; Calcd: C, 53.16; H, 4.12; N, 14.31; Found: C, 53.17; H, 4.11; N, 14.29.</p><!><p>Compound 4c was synthesised from 2b and 2-iodoanline following the general procedure, affording the product as a white powder in 89% yield; mp 285 °C; IR (KBr, cm−1): 3086, 3053, 2953, 2885,1598, 1560, 1516, 1463, 1371; 1H-NMR (CDCl3, δ, ppm): 12.77 (d, 1H, J = 14.8 Hz, NH), 8.55 (d, 1H, J = 14.0 Hz, CH=), 7.77 (d, 1H, J = 7.2 Hz, Ar-H), 7.57 (d, 1H, J = 8.0 Hz, Ar-H), 7.46 (t, 1H, J = 7.6 Hz, Ar-H), 7.26 (t, 1H, J = 7.2 Hz, Ar-H), 3.20 (s, 6H, 2CH3); 13C-NMR (CDCl3, δ, ppm):163.3, 162.9, 152.2, 151.9, 138.3, 136.4, 131.3, 128.1, 125.9, 118.5, 93.9, 28.2, 27.6; LC/MS (ESI): 386.16 [M + 1]+; Anal. for C13H12IN3O3; Calcd: C, 40.54; H, 3.14; N, 10.91; Found: C, 40.55; H, 3.15; N, 10.90.</p><!><p>Compound 4d was synthesised from 2b and 2-aminopyridine following the general procedure, affording the product as a white powder in 85% yield; mp 287–290 °C; IR (KBr, cm−1): 3420, 2999, 2960, 2908, 2870, 1624, 1591, 1456; 1H-NMR (CDCl3, δ, ppm): 12.11 (brs, 1H, NH), 9.41 (d, 1H, J = 13.2 Hz, CH=), 8.43 (d, 1H, J = 4.4 Hz, Ar-H), 7.75 (dd, 1H, J = 8.0, 2.4 Hz, Ar-H), 7.16 (t, 1H, J = 7.6, Hz, Ar-H), 6.99 (d, 2H, J = 8.0 Hz, Ar-H), 3.36 (s, 6H, 2CH3); 13C-NMR (CDCl3, δ, ppm): 165.3, 162.6, 152.0, 151.3, 149.7, 149.3, 138.9, 121.4, 112.7, 94.3, 28.2, 27.5; LC/MS (ESI): 261.25 [M + 1]+; Anal. for C12H12N4O3; Calcd: C, 55.38; H, 4.65; N, 21.53; Found: C, 55.38; H, 4.64; N, 21.51.</p><!><p>Compound 5 was synthesised from 2b (2 equiv.) and 4-aminobenzylamine (1 equiv.) following the general procedure, affording the product as a white powder in 90% yield; mp 195 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (DMSO-d6, δ, ppm): 10.48 (brs, 2H, NH), 8.25(d, 2H, J = 14.0 Hz, 2CH=), 7.01 (d, 2H, J = 8.0 Hz, Ar-H), 6.55 (d, 2H, J = 9.0 Hz, Ar-H), 4.50 (d, J = 6.4 Hz, 2H, CH2), 3.12 (s, 12H, 4CH3); 13C-NMR (DMSO-d6, δ, ppm): 168.1, 165.5, 160.6, 151.9, 134.4, 118.7, 145.2, 95.3, 28.1, 27.4; LC/MS (ESI): 455.44 [M + 1]+; Anal. Calcd for C21H22N6O6: C, 55.50; H, 4.88; N, 18.49; Found: C, 55.65; H, 4.93; N, 18.70.</p><!><p>Compound 6 was synthesised from 2a (2 equiv.) and 4-aminobenzylamine (1 equiv.) following the general procedure, affording the product as a yellow powder in 86% yield; mp 247 °C; IR (KBr, cm−1): 3302, 2958, 2908, 2866, 1614, 1587, 1545, 1504, 1438, 1409; 1H-NMR (CDCl3, δ, ppm): 12.34 (d, 1H, J = 13.6 Hz, NH), 10.77 (m, 1H, NH), 8.68 (d, 1H, J = 14.0 Hz, CH=), 8.29 (d, 1H, J = 13.6 Hz, CH=), 7.39 (m, 4H, Ar-H), 4.65 (d, 2H, J = 6.4 Hz, CH2), 4.46 (m, 8H, 4CH2), 1.35 (m, 12H, 4CH3); 13C-NMR (CDCl3, δ, ppm): 179.0, 178.8, 163.3, 163.2, 161.0, 160.7, 160.4, 152.7, 138.4, 133.6, 129.7, 129.6, 118.9, 118.8, 95.4, 93.6, 53.6, 43.2, 43.0, 42.5, 42.4, 12.5, 12.4, 12.3, 12.2; LC/MS (ESI): 543.67 [M + 1]+; Anal. Calcd for C25H30N6O4S2: C, 55.33; H, 5.57; N, 15.49; Found: C, 55.54; H, 5.69; N, 15.66.</p><!><p>The assay protocol for was performed spectrophotometrically following the reported method22, where α-glucosidase from S. cerevisiae (G0660-750UN, Sigma Aldrich) was dissolved in phosphate buffer (pH 6.8, 50 mM). Test compounds were dissolved in 70% DMSO. 20 μL of test sample, 20 μL of enzyme and 135 μL of buffer were added to 96-well plates and incubated for 15 min at 37 °C. After incubation, 25 μL of p-nitrophenyl-α-d-glucopyranoside (0.7 mM, Sigma Aldrich) was added and changes in absorbance were monitored for 30 min at 400 nm. The test compound was replaced by DMSO (7.5% final) as control. Acarbose (Acarbose, Sigma Aldrich) was used as a standard inhibitor.</p><!><p>The assay was performed following Gutierrez. R. M. P, with slight modifications. In brief, Bovine Serum Albumin solution (10 mg/mL) was prepared in 100 mM of phosphate buffer pH 7.4 containing 3 mM sodium azide as antimicrobial agent. A methylglyoxal solution of 14 mM was also prepared in the same buffer. 1-mM concentrations of the test compounds and standard inhibitor were prepared in dimethyl sulfoxide (DMSO). Each well of a 96-well plate contained 20 µL of inhibitor, 50 µL of BSA, 50 µL of methylglyoxal and 80 µL of phosphate buffer, while the control contained 20 µl of DMSO instead of test compound. The total reaction volume was 200 µL. The reaction mixture was then incubated for 9 days at 37 ° C. After incubation, each sample was examined for the development of specific fluorescence (excitation 330 nm; emission 420 nm) against a blank on a microplate reader (Spectramax M2 Devices, CA, USA).</p><!><p>The percentage inhibition of advanced glycation end (AGEs) products formation by the test sample versus control was calculated using the following formula: The % inhibition  of  AGE  formation=1−[(fluorescence  of  the  test  groupfluorescence   of   the   control  group)]×100%</p><!><p>Click here for additional data file.</p>

PubMed Open Access

Phosphatidylserine directly and positively regulates fusion of myoblasts into myotubes

Cell membrane consists of various lipids such as phosphatidylserine (PS), phosphatidylcholine (PC), and phosphatidlethanolamine (PE). Among them, PS is a molecular marker of apoptosis, because it is located to the inner leaflet of plasma membrane generally but it is moved to the outer leaflet during programmed cell death. The process of apoptosis has been implicated in the fusion of muscle progenitor cells, myoblasts, into myotubes. However, it remained unclear whether PS regulates muscle cell differentiation directly. In this paper, localization of PS to the outer leaflet of plasma membrane in proliferating primary myoblasts and during fusion of these myoblasts into myotubes is validated using Annexin V. Moreover, we show the presence of PS clusters at the cell-cell contact points, suggesting the importance of membrane ruffling and PS exposure for the myogenic cell fusion. Confirming this conclusion, experimentally constructed PS, but not PC liposomes dramatically enhance the formation of myotubes from myoblasts, thus demonstrating a direct positive effect of PS on the muscle cell fusion. In contrast, myoblasts exposed to PC liposomes produce long myotubes with low numbers of myonuclei. Moreover, pharmacological masking of PS on the myoblast surface inhibits fusion of these cells into myotubes in a dose-dependent manner

phosphatidylserine_directly_and_positively_regulates_fusion_of_myoblasts_into_myotubes

1. Introduction<!>2.1. Cell culture of mouse primary myoblasts<!>2.2. Preparation of SUV (small, unilamellar vesicles) liposomes<!>2.3. Immunocytochemistry<!>2.4. Differentiation of mouse primary myoblasts<!>2.5. Statistical analysis<!>3.1. Phosphadidylserine is enriched at cell-cell contact regions during myoblast differentiation into myotubes<!>3.2. PS treated myoblasts form robust myotubes<!>3.3. Myoblast fusion index is decreased by masking PS with Annexin V or PS-specific antibody<!>4. Discussion<!>

<p>The cell plasma membrane largely consists of a phospholipid bilayer containing phopsphatidylcholine (PC), phosphatidylserine (PS) and phosphatidylethanolamine (PE) [1,2]. PS is a hallmark of apoptosis because in normal healthy cells it is localized to the inner lipid layer, or leaflet, of the plasma membrane, but during early apoptosis PS localizes to the outer leaflet by phospholipid(s) scramblase [3,4]. The exposed PS recruits various immune cells and in particular is one of the 'eat me' signals present on the surface of apoptotic cells signaling macrophage engulfment [5].</p><p>During muscle maintenance and repair, muscle stem or satellite, cells activate, proliferate and give rise to myoblasts, which can both proliferate and differentiate into myotubes. Myotubes are the final product of the muscle lineage, where each cell is post-mitotic and multinucleated and is produced by the fusion of many myoblasts [6]. Some growth factors [7–9], macrophages [10,11], and leukocytes in general [12] can control muscle regeneration and specifically, cell fusion. However, the molecular nature of fusogen(s) remains unknown and it is yet to be determined what specific molecules in the plasma membrane of myoblasts play a crucial role during fusion of these cells into myotubes or myofibers. To date, several molecules that are thought to control myoblast fusion, such as caveolin-3 [13,14], myoferin [15] and nephrin [16], were reported to cluster at the cell membrane, specifically localizing at the cell-cell contact regions during myotube formation. However, none of these molecules were demonstrated to directly regulate the membrane fusion process. Additionally, damaging stimuli that are known to promote apoptosis as well as direct experimental induction of apoptosis have been shown to enhance myotube formation [17,18]. However, no molecular mechanism connecting apoptosis with myoblast fusion was revealed, until now.</p><p>In this paper, we demonstrate that one of the key membrane-bound regulators of myoblast fusion into myotubes is PS, which provides the first molecular identification of a fusogen and gives an explanation as to why induction of apoptosis promotes terminal muscle differentiation.</p><!><p>Primary myoblasts were isolated from young (2–4 month) C 57/Bl6 mice (Jackson Lab., ME) as described [20]. Briefly, tibialis and gastrocnemius muscles were dissociated into myofibers by a digest for 1 hr at 37°C in DMEM with 250 units/ml Collagenase Type II (Sigma-Aldrich, MO), containing 1% Penicillin-Streptomycin. Digested muscles were washed with PBS two times and triturated into myofibers in Ham's F10 (Mediatech, VA), 20% Bovine growth serum (BGS; Hyclone, IL) and 1% Penicillin-streptomycin. In order to get primary myoblast from myofibers, myofibers were incubated in Ham's F10 including 1% Pnicillin-Streptomycin and 2 U/ml Dispase for 1 hr at 37°C incubator with shaking. Isolated satellite cells were cultured in growth media (Ham's F10, 20% BGS, 5 ng/ml bFGF (Invitrogen, CA) and 1% Penicillin-Streptomycin). Satellite cells gave rise to primary myoblasts in growth medium. Primary myoblasts were used for no more than 10 passages.</p><!><p>Phosphatidylserine and phosphatidylcholine were purchased from Avanti polar lipids (Alabaster, AL). PS and 50:50% PS:PC SUV liposomes were made as described [21]. Briefly, 2.6 µmole each phospholipid (PL) was prepared in a glass test tube (13 × 100 mm), in chloroform and then dried under a nitrogen stream in the hood, followed by vacuum drying for 1 hr. Dried PL was resuspended with 2.6 ml HBS solution for 1 hr at room temperature and then vortexed to mix. Resuspended PL was sonicated until clear using a bath sonicator (75T, VWR, PA) for 5–10 min. PC, PS and PC:PS liposomes were added at 40, 80, and 80 µM concentrations to myoblasts cultured in DM for 48 hours.</p><!><p>Immunostaining was performed as described [19]. Briefly, differentiated myotubes were fixed in 70% EtOH over night at 4°C. Cells were permeabilized with PBS + 1% FBS + 0.25% Triton X-100, incubated with 1 µg/ml of primary antibodies for 1 hr at room temperature, washed in staining buffer (PBS + 1% FBS), incubated with fluorochrome conjugated secondary antibodies and Hoechst, washed again and mounted.</p><!><p>Primary myoblasts were seeded onto 6 well plates at a density of 5 × 105 cells/well in growth medium. Culture medium was switched to differentiation medium (DMEM containing 2% Horse serum and 1% Penicillin-Streptomycin), for 48 hrs at 37°C under 5% CO2. Exposed PS was detected using Alexa Fluor® 488 conjugated Annexin V (Invitrogen, CA), as recommended by the manufacturer [23].</p><!><p>Quantified data are expressed as mean ± sd. Significance testing was performed using T-test of variance to compare data from different experimental groups. Three independent experiments were conducted to obtain the means. P-values of < 0.05 were considered statistically significant.</p><!><p>To determine whether phophatidylserine (PS) is exposed on the outer leaflet of the plasma membrane during fusion of primary myoblast into de-novo myotubes, mitogenic growth medium (GM) was replaced by the mitogen-low differentiation medium (DM) where myoblasts normally form myotubes by 48 hrs, and these cell cultures were stained by Alexa Fluor 488 conjugated-Annexin V (Fig. 1). H2O2 treated myoblasts were used as a positive control for PS translocation and Annexin V staining, as well as for propidium iodide (PI) staining. All cells were co-stained by PI in order to clearly distinguish membrane permeable (apoptotic) cells from live, fusing cells. H2O2 treated cells were easily detected by PI and Annexin V, indicating they had either localized PS to the outer surface of their membranes (Fig. 1a, b) or their membranes were antibody permeable. Some Annexin V staining was seen in live, PI-excluding myoblasts in GM (Fig. 1c, d), indicating these cells had localized PS to the outer leaflet of the plasma membrane. In DM, however, fused myotubes which excluded PI and thus were not apoptotic, had significantly more Annexin V staining than myoblasts cultured in GM, and furthermore, the highest presence of PS clusters was identified at the cell-cell contact regions of apparently fusing myoblasts (Fig. 1e, f). These results reveal that re-location of PS from the inner to the outer leaflet of the plasma membrane in myoblasts and myotubes is not caused by the process of apoptosis and might be specific to cell-cell fusion.</p><!><p>In order to address whether PS just correlates with or actually causes fusion of primary myoblasts into multinucleated myotubes, we generated PS liposomes (as well as a negative control PC liposomes), and added them to myoblasts that were cultured for 48 hrs in DM. Myoblast treated with PS liposomes and a 50:50% mixture of PS:PC liposomes, but not with PC liposomes alone, displayed greatly enhanced fusogenic properties, based on the quantification of the width of the de-novo formed, eMyHC expressing myotubes which have more than 2 nuclei, and on the counts of the myonuclei in these myotubes (Fig. 2A, quantified in B). Myoblasts treated with PC liposomes alone formed long narrow myotubes with a low number of myonuclei, suggesting a defect in myogenic cell fusion (Fig. 2 C). In contrast, the width and the number of myonuclei (fusion index) of de-novo myotubes were enhanced by adding PS or PS:PC liposomes. PS:PC liposomes increased the width, length and fusion index of the myotubes, suggesting that while PS directly and specifically enhanced the myoblast fusion, the high concentration of liposomes (80 uM) in PC and PS:PC cultures could have an indirect effect on myotube length. These results strongly suggest that PS liposomes on the myoblast cell surface directly enhance cell-cell contacts and promote myogenic fusion into multinucleated myotubes of larger width and with more myonuclei.</p><!><p>To confirm and extrapolate these data, we blocked PS on the cell surface of myoblasts cultured in DM by treating these cells with Annexin V or with PS-specific antibody, and assayed the ability of these myoblasts to fuse into myotubes. Indeed, the myoblast fusion index was significantly decreased in a dose-dependent manner when PS was experimentally masked (Fig. 3A, C, quantified in B, D). These data establish that translocation of the PS to the outer leaflet of the plasma membrane directly and positively regulates the fusion of primary myoblasts into multinucleated myotubes.</p><!><p>Phosphatidylserine (PS) is a hallmark of eukaryotic cell apoptosis and its translocation from the inner to outer leaflet of the membrane is well known [3,4]. While previous work suggested that some membrane bound proteins may regulate myoblast fusion and that apoptosis generally promotes myotube formation, the molecular regulator of this event remained unknown and the mechanism by which apoptosis promotes myotube generation was undefined. This work is the first to provide a molecular explanation to both of these questions, as it occurs in primary mouse myoblasts that were freshly generated from muscle stem cells. Specifically, we show that PS is moved to the outer leaflet of the plasma membrane in primary myoblasts and that this process plays a crucial causal role during myoblast fusion. Even though PS was localized to the outer leaflet in myoblasts fusing into myotubes, these cells were not undergoing apoptotic pathway. Hence, these data uncover that membrane ruffling occurs independently of apoptosis, which changes the dominant paradigm. As shown in figure 1, some of PS is present on the outer membrane leaflet of proliferating myoblasts [24]. The function of exposed PS during myoblast cell proliferation is unknown, but we propose that it predisposes these cells for the fusion events. Interestingly, according to a previous report, the presence of PS at the cell-cell contact regions in C2C12 and H9C2 cell lines was only transient [25]. However, our data in primary myoblasts suggest that PS is present in the outer membrane leaflet broadly and persistently, particularly when these myogenic cells are cultured for 48 hours in DM that is known to promote myoblast differentiation into myotubes. Thus, our work not only validates but also greatly clarifies the localization of PS in proliferating and differentiating muscle progenitor cells.</p><p>Importantly, when PS liposomes were added exogenously, fusion of myoblasts into multinucleated myotubes was significantly and specifically enhanced; conversely, masking of PS with Annexin V or an antibody, inhibited the fusogenic properties of myoblasts (Fig. 2). Therefore, PS localization to the outer membrane that is unrelated to cell-death, directly promotes the physiological fusogenic properties of primary myoblasts. Future identification of the PS receptor in myoblasts will further enhance our understanding of the cell-cell fusion process and of the molecular regulation of muscle regeneration.</p><!><p>PS broadly and persistently trans-locates to the outer leaflet of plasma membrane duirng myoblast fusion into myotubes. Robust myotubes are formed when PS liposomes are added exogenously. PS increases the width of de-novo myotubes and the numbers of myonuclei, but not the myotube length. Annexin V or PS antibody inhibits myotube formation by masking exposed PS.</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>

PubMed Author Manuscript

Shallow Representation Learning Via Kernel PCA Improves QSAR Modelability

Linear models offer a robust, flexible, and computationally efficient set of tools for modeling quantitative structure activity relationships (QSAR), but have been eclipsed in performance by non-linear methods. Support vector machines (SVMs) and neural networks are currently among the most popular and accurate QSAR methods because they learn new representations of the data that greatly improve modelability. In this work we use shallow representation learning to improve the accuracy of L1 regularized logistic regression (LASSO) and meet the performance of Tanimoto SVM. We embedded chemical fingerprints in Euclidean space using Tanimoto (aka Jaccard) similarity kernel principal components analysis (KPCA), and compared the effects on LASSO and SVM model performance for predicting the binding activities of chemical compounds against 102 virtual screening targets. We observed similar performance and patterns of improvement for LASSO and SVM. We also empirically measured model training and cross validation times to show that KPCA used in concert with LASSO classification is significantly faster than linear SVM over a wide range of training set sizes. Our work shows that powerful linear QSAR methods can match nonlinear methods, and demonstrates a modular approach to non-linear classification that greatly enhances QSAR model prototyping facility, flexibility, and transferability.

shallow_representation_learning_via_kernel_pca_improves_qsar_modelability

INTRODUCTION<!>CONCLUSIONS<!>Step 1: Compute Tanimoto similarity matrix<!>Step 2: Compute its eigenvalue decomposition (or SVD)<!>Step 3: Use the eigenvectors to project data into Tanimoto space<!>Software<!>Classification Datasets<!>Model Training and Classification Accuracy Assessment<!>LASSO Classification<!>SVM Classification<!>Comparison between LASSO and SVM<!>Computation Time<!>Tanimoto KPCA embedding improves LASSO and Linear SVM performance<!>Effects of KPCA embedding depend on the properties of the underlying representation<!>LASSO models trained with embedded fingerprints are comparable to Tanimoto SVM<!>KPCA/LASSO classification is significantly faster than linear SVM for ADR target datasets<!>Implications for deep and shallow learning<!>Considerations for KPCA embedding

<p>Since its introduction over 50 years ago, quantitative structure activity relationship (QSAR) modeling has become an indispensible tool for drug development.1, 2 A powerful application of QSAR is ligand based virtual screening, where predictive models are built from experimental data and used to mine chemical libraries for promising lead compounds.3, 4 Common screening objectives include molecular properties that influence ADME profiles,5 activity against desired therapeutic targets, and liability against targets that contribute to adverse drug reactions (ADRs).6, 7 The diversity of targets, objectives, and constraints make virtual screening challenging and complex.</p><p>The most straightforward approach is similarity-based virtual screening.8 Library compounds are compared to a reference set of known active molecules, and those similar above some threshold are selected as putative leads. 2D chemical fingerprints and Tanimoto similarity are the most popular representations for assessing similarity.9 These methods are (1) versatile-- because they admit any representation that supports comparison, and (2) efficient--because they are non-parametric and do not require fitting models or hyper parameters. However, they are limited because they summarize similarity as a scalar quantity, and do not have sufficient granularity to capture the precise features that drive critical similarities and differences.10</p><p>Linear models give granular insight into structure activity relationships by learning quantitative rules that relate specific substructural features to biological activities.11 Since their early application in QSAR, linear modeling techniques have been refined and optimized extensively.12–14 Today many robust implementations are available for popular models such as logistic regression, principal components analysis (PCA),15 and partial least squares (PLS).16 A limitation of linear methods for QSAR is their use of the dot product, which is less discriminative than non-linear Tanimoto similarity for sparse bit vectors like chemical fingerprints.17 Consequently, non-linear methods such as support vector machine (SVM),18, 19 neural networks,18, 20, 21 random forests,22 and influence relevance voting (IRV),23 often eclipse the performance of linear QSAR methods and make them appear uncompetitive, despite the availability of efficient, accessible, mature techniques and associated software implementations.</p><p>In particular, Support Vector Machines (SVMs) are frequently used for non-linear QSAR.19 They combine the strengths of similarity search and linear models by using non-linear similarity functions to map data into high dimensional vector spaces.24 A key advantage of SVM is that task-specific prior information can be used to choose a kernel and representation that gives a favorable non-linear abstraction (e.g. Tanimoto similarity for 2D fingerprints).25 However, SVMs also have a number of limitations: models may take a long time to train or have convergence issues, hyperparameter adjustments can be difficult and time consuming, and limited custom kernel support complicates workflows and limits the transferability of the methods.26, 27</p><p>Kernel principal components analysis (KPCA) is a nonlinear embedding technique closely related to SVM.24, 28 KPCA takes a set of data examples and a positive definite measure of similarity, and returns a set of continuous vector representations whose dot products give minimum error regularized estimates of the similarity measure. Of course, it shares the qualities of classical PCA: features of the transformed data are uncorrelated and ordered by their explanatory power, and new data can be projected into the subspace spanned by the training examples.</p><p>Our work is motivated by the hypothesis that KPCA allows us to access the non-linear abstraction of our data in the hidden layer and extend it to other machine learning models.25 We train sparse logistic regression models on KPCA embedded chemical fingerprints and show significant improvements in accuracy. We use KPCA embedded vectors to train Tanimoto SVM models with an implementation of that does not ordinarily permit use of non-standard kernels. We show that KPCA embedding is surprisingly efficient over a broad range of virtual screening dataset inputs. Our work shows that KPCA embedding makes non-linear QSAR more flexible and transferable, and provides a role for linear models within the broader context of representation learning.29</p><!><p>We have demonstrated the potential for modularity in non-linear classification enabled by shallow, unsupervised representation learning: we provide nonlinear modeling capabilities in the context of a linear modeling formalism. Any learning architecture must provide two functions: (1) the creation of a useful abstraction of the data, and (2) a method for fitting a model using the data abstraction. Often the embedding step is done implicitly, but this can limit transferability and applicability.2, 51 We show that separating these functions into isolated modules (for QSAR modeling, and in the context of linear models) can greatly enhance speed, facility, and flexibility.29, 52</p><!><p>Given a set of data molecules, X= {x1,x2,…,xn} represented as sets of sparse binary indicator features, we can compute the Tanimoto similarity T(xi,xj) using the formula above where 〈xi,xj〉 denotes the dot product of fingerprints xi and xj.</p><p>The heart of the "kernel trick" is Mercer's theorem, which relates the output of certain types of non-linear similarity functions to dot products of vectors ϕ(xi) and ϕ(xj) in a high dimensional space.</p><!><p>We factor the Tanimoto similarity matrix with the eigenvalue decomposition (or SVD), and then multiply the eigenvectors (Q) by the singular values (Λ) to return the embedded data Φ(X).</p><p>The new features are orthogonal and ordered by variance, and dot products of molecule vectors approximate Tanimoto similarities. We can also smooth the data by discarding low variance features. In general, the first k eigenvalues and eigenvectors give a minimum error rank k approximation of the data. Here ||·||2 is the spectral norm, and Λ(k) denotes the first k singular values of T(X,X).</p><!><p>To embed new data points we compute similarity to the points in our training data, and than right multiply by the inverse of our embedded training data.</p><!><p>We built our software in R 3.3.0. We used the glmnet package for LASSO classification,30 and the e1017 (LibSVM) and kernlab packages for SVM.26, 31, 32 We used the caret package to do stratified sampling for cross validation,33 and the ROCR package to compute performance measures.34</p><!><p>We tested classification performance using representation-benchmarking dataset assembled by Heikamp et al.35, 36 The active set consists of 75,251 compounds that affect 102 therapeutic targets (ChEMBL IC50 ≤ 10 μM). A number of therapeutic target families (5HT transporters/receptors, carbonic anhydrases, kinases), and exclusionary targets (HERG, CYP family) are represented. The data are grouped into Easy, Intermediate, and Difficult classes based on maxSIM recovery rates. Summary statistics describing the size and diversity of activity class datasets are shown in Table 1. An additional 10,000 compounds from ZINC were also provided as decoys. The data are encoded as ECFP4 and MACCS fingerprints to give contrasting examples of high and low fidelity representations.37, 38</p><!><p>We evaluated the classification performance of LASSO and SVM models using AUC, F1-score, and Matthew's correlation coefficient (MCC). Because the thresholding process used to compute AUC yielded a range of values for F-score and MCC, we report their maxima. We evaluated changes in classification performance (Δp/p0) and error (Δε/ε0) for models trained on KPCA vectors relative to the baseline of linear models trained on 2D fingerprints. We computed the performance measures using a 10-fold cross validation in scheme in which 90% of the data was used to train each fold-specific model, and the remaining 10% was used to evaluate prediction accuracy. For models trained with embedded molecule vectors, KPCA embedding was included in the 10-fold cross validation. We used nested 10-fold cross validation on the training data to tune model hyperparameters. We fit lasso models using the glmnet package with a lambda min ratio of 0.001 and selected the lasso penalty parameter using the one standard error criterion with AUC as the performance metric. For SVM, we used Platt scaling and evaluated the C parameter over six orders of magnitude [10−3,103]. We measured the total training and cross validation times and fit polynomial models for extrapolation.</p><!><p>Figure 1 (Top) shows the effect of KPCA embedding on LASSO classification model performance for each of the 108 ChEMBL activity classes. In aggregate, changes in performance were biased toward improvement, but within the expected variation for both MACCS (z-score=0.76) and ECFP4 (z-score=0.26) fingerprints. However, the change in performance was unevenly distributed across virtual screening targets, favoring intermediate and difficult target activity models trained on MACCS fingerprints (Table 2a). Of the ten most significantly affected activity class models (Table 2b), all except Glucocorticoid receptor (GCR) were difficult targets. In general, the relatively small changes in performance (Δp) we observed represented large portions of the maximum possible improvement (Δε). No significant decrease in accuracy (z-score < −1) was observed for any activity class. Performance tables for all individual activity classes and are listed in the Supporting Material.</p><!><p>Figure 1 (bottom) shows the effect of KPCA embedding on SVM classification performance for each of the 108 ChEMBL activity classes. Overall, they were similar to those observed for LASSO classification. The aggregate change in performance was larger for MACCS (z-score= 0.81) than ECFP4 (z-score=0.11) fingerprints, but within the expected variation for both fingerprint types. Improvements again favored intermediate and difficult activity classes (Table 2c). Small changes in performance represented large portions of the maximum possible improvement. 8 of the top 10 most significantly affected model also showed the biggest improvement for LASSO classification (Table 2d). We did not observe significant differences in performance (z-score < −1) for any activity class. Performance tables for all individual activity classes and are listed in the Supporting Material.</p><!><p>Figure 2 shows a direct comparison of LASSO and SVM classification performance for native and embedded fingerprints. SVM generally performed better than the LASSO, however differences in performance were dependent on the representation used to train the models. Native ECFP4 fingerprints gave the largest difference in accuracy (z-score=0.76). The effect was most pronounced for difficult targets (Table 3a), and a number of individual models showed differences in performance above the expected variation (Table 3b). The difference in performance was much smaller for native MACCS fingerprints (z-score=0.21). Here the effect was more evenly distributed across easy, intermediate, and difficult targets (Table 3a), and no individual target had a z-score > 1. Tanimoto KPCA embedding had the effect of reducing the gap in performance between SVM ad LASSO models for ECFP4 (z-score=0.57), while doing the opposite for MACCS (z-score=0.31). In both cases, the effect was concentrated in the intermediate and difficult classes (Table 3c,d). We also compared F1-scores for SVM and LASSO classification models trained on KPCA embedded fingerprints with the results for Tanimoto SVM reported by Balfer et al on Dopamine-D2 receptor (72), Cannabinoid CB2 receptor (259), and MAP Kinase P38 receptor (10188) activity classes.17 An F-test showed no significant difference (p=0.96).</p><!><p>Figure 3 shows the activity class size and total cross validation times for non-linear LASSO and linear SVM. KPCA embedding combined with lasso training was faster than linear SVM trained on 2D fingerprints for all input sizes and fingerprint types. The speed up ranged from negligible for smaller inputs to an order of magnitude for the largest inputs tested. We observed that the computation time of linear SVM scaled much better with ECFP4 than MACCS fingerprints. Fingerprint type did not significantly impact the compute times for KPCA embedded models. Extrapolation from polynomial models (Table 4) predicted that linear SVM would overtake non-linear LASSO at and input size of 19,077 compounds for ECFP4, and 484,673 compounds for MACCS.</p><!><p>Our results (Figure 1) show that non-linear embedding of chemical fingerprints using KPCA improves LASSO (Table 2) and SVM (Table 2c) classification performance for a number of important virtual screening targets. The magnitude of the performance gains we observed were relatively small, but meaningful when considered in the context of the large chemical libraries, which order on the number of tens of millions of molecules,39 where small reductions in error can equate to hundreds or thousands of fewer misclassified compounds. The targets in our dataset fall into a number of families that commonly participate in ADRs, such as ion and neurotransmitter transporters (HERG, 5HTT, NET); G-protein coupled receptors (CB1, CB2, H3); nuclear receptors (GCR); and enzymes (HSD1, VEGFR2).7 Among these, the difficult targets are critical because they are highly promiscuous, and thus the most likely causes of safety related attrition from off target effects; and it was here that non-linear embedding delivered the greatest improvements in performance (Tables 2b, 2d). Furthermore, while Tanimoto KPCA embedding did not improve classification performance on all targets, it never significantly hurt performance.</p><!><p>The smaller effect size for ECFP4 fingerprints can be explained by the underlying hashing scheme employed in the fingerprint generation process. Often referred to as "feature hashing" or the "hashing trick", the practice of using hash table values to efficiently represent extremely sparse, high-dimensional data is a common optimization for natural language models, SVMs in particular.40, 41 Thus, ECFP4 fingerprints can be considered to be highly optimized for linear models, and almost uniquely tailored for linear SVMs. KPCA embedding brought the performance of models trained on MACCS closer to those trained on ECFPs. This implies that the high dimensional feature space of ECFPs may not capture significantly more classification relevant information than the 166 structural keys of MACCS, so much as represent it in a way such that relationships between data are linear. The improvement from KPCA embedding observed for MACCS can similarly be attributed to representing the data in a way such that similarity relationships between molecules are linear. Thus, for ECFP4 fingerprints, which are already highly optimized for linear properties, the effects were limited. Our results suggest that KPCA embedding is most useful in combination with compact, "lo-fi", SMILES based representations that are not highly optimized for linear classification, such as MACCS and LINGOS.42, 43</p><!><p>Our results show that embedding fingerprints with Tanimoto KPCA improves LASSO performance such that it is virtually equivalent to Tanimoto SVM (Figure 2). While performance was biased in favor of SVM models, the differences were within the standard error for all but a few cases (Table 3). For those, SVM models already significantly outperformed LASSO models when trained on native fingerprints, and Euclidean embedding of chemical structures narrowed the gap in performance. The marginally better performance of SVM relative to comparable LASSO models can be explained by the difference in how each fits a separating hyperplane. LASSO classification selects a subset of features, and considers all of the data examples, even those far from the decision boundary, when computing an optimally separating hyperplane. SVM maps the data to a high dimensional feature space, and considers only a subset of examples near the class boundary. Thus, SVM models tend to perform better near class boundaries, but the overall contribution to accuracy is small because points are unlikely to fall in the affected regions.</p><!><p>Our observation that LASSO classification with KPCA embedded fingerprints was much faster than linear SVM (Figure 3) is counterintuitive given the theoretical complexity of KPCA O(n3) is greater than SVM O(n2). This can be explained by two factors: the absence of KPCA hyperparameters, and the size of the datasets. KPCA has no hyperparameters to optimize, thus it is a one-time cost. Grid search for tuning SVM hyperparameters can require on the order of hundreds of model fits, which inflates the total computational cost up to an equivalent factor. The result is a range of inputs where KPCA is faster than training an accurate SVM model, even when it is much slower than fitting a single unoptimized SVM model. We expect SVM to overtake KPCA/LASSO at large enough scale, however it is unclear when that will happen. Our polynomial models (Table 4) provide rough estimates, but like most extrapolations they are subject to wide confidence intervals. In the limit of largest most diverse activity class in our dataset (HERG), KPCA was still two orders of magnitude faster than linear SVM. Our results highlight an important caveat to keep in mind when considering theoretical guarantees: they may omit non-trivial application details that significantly affect empirical performance measurements, thus asymptotic limits may not apply to a range of relevant problem sizes.</p><!><p>Much of the recent interest in data embedding (particularly in the deep neural network learning community) stems from the idea that deep and shallow learning architectures generate their effects by learning new distributed representations of the data in their hidden layers.25 Neural network embedded chemical fingerprints have shown promise, but share limitations common to deep learning: Models are computationally expensive and difficult to train, and hyperparameters like the learning rate, smoothing parameters, and model architecture must be tuned for each application.20, 44, 45 Shallow learning offers a simpler and more robust alternative, but with limitations. Restricting network depth makes training easier and more efficient, but limits the expressiveness of the range of nonlinear representations that can be learned. While the added expressiveness of the representations learned by deep neural networks generally accounts for their superiority in complex learning tasks, selection of an appropriate kernel using prior knowledge can allow us to obtain favorable tradeoffs in efficiency and expressivity. Thus, "shallow" learning as embodied by our methods and SVM may be preferred tasks for which the user has specific a priori understanding of the feature space; and deep learning may be preferable for tasks involving extremely large datasets, for which the user lacks confidence about the most appropriate representation or similarity function.</p><!><p>The choice of the kernel and underlying data representation are the two most important metaparameters for KPCA, and most effectively selected on the basis of prior knowledge and intuition. For this work, we used Tanimoto similarity because of its recognized utility as similarity measure for 2D chemical fingerprints. Alternatively if the data were continuous and real valued, as in the case of whole molecule descriptors, a radial basis function or polynomial kernel would be more appropriate. For distance based modeling approaches such as k-means clustering, we might use classical multidimensional scaling (MDS), a particular type of KPCA that preserves distances.28, 46 It should be noted that while classical MDS and PCA are equivalent up to an orthogonal rotation around the origin, this is not the case for other distance/similarity metric pairs. The biggest limitations of KPCA are the memory and time requirements, which are O(n2) and O(n3) respectively. We have shown that for the range of problems we have addressed, these are not prohibitive; however for much larger problems a number of optimizations such as matrix sketching,47 non-random sampling,48 and ensemble approaches49 are available as well as distributed and streaming versions of KPCA.50</p>

PubMed Author Manuscript

Electrocatalytic C-N Coupling via Anodically Generated Hypervalent Iodine Intermediates

Development of new electrosynthetic chemistry promises to impact the efficiency and sustainability of organic synthesis. Here we demonstrate that anodically generated hypervalent iodine intermediates effectively couple interfacial electron transfer with oxidative C-H/N-H coupling chemistry. The developed hypervalent iodine electrocatalysis is applicable in both intraand intermolecular C-N bond-forming reactions. Available mechanistic data indicates that anodic oxidation of aryl iodides generates a transient I(II) intermediate that is critically stabilized by added acetate ions. This report represents the first example of metal-free hypervalent iodine electrocatalysis for C-H functionalization and provides mechanistic insight that we anticipate will contribute to the development of hypervalent iodine mediators for synthetic electrochemistry.Electrochemistry is an attractive approach to sustainable synthesis that obviates the need for stoichiometric redox reagents and thus, generation of the attendant waste streams. 1 Due to its inherent tunability and scalability, electrosynthesis should impact many of the enormous variety of organic transformations in which electrons are added to, or removed from, substrates. In practice, challenges such as 1) the sluggish interfacial electron transfer rates for many organic molecules, which necessitates application of substantial overpotential to achieve practical current densities, 2 and 2) the need to couple the single-electron events that are typical of electrochemistry with the multi-electron events required for bond-breaking and -making in organic reactions, can

electrocatalytic_c-n_coupling_via_anodically_generated_hypervalent_iodine_intermediates

<p>limit direct electrosynthesis (Figure 1). 3 Indirect electrocatalysis, in which small molecule electrocatalysts that display well-behaved electrochemistry convey applied potential from the working electrode to the bulk reaction medium, has emerged as an important strategy in selective organic electrocatalysis. 4 Redox catalysts can both facilitate electron transfer and couple a diverse array of substrate functionalization mechanisms to the electrochemical stimulus. Important methods based on quinone-, 5 amine-, 6 nitroxyl radical-, 7 and transition metal-redox catalyst 8 have been disclosed. Hypervalent iodine reagents are a class of organic oxidants that have been deployed in a wide variety of substrate functionalization reactions. 9 Electrochemical oxidation of aryl iodides typically requires substantial overpotential, 10 thus hypervalent iodine electrochemistry has largely been limited to ex cell applications, 11 in which aryl iodides are electrolyzed in the absence of substrate and subsequently used as stoichiometric reagents, or implemented in the context of flow systems. 12,13 During the development of aerobic hypervalent iodine catalysis, 14 we discovered that the aerobic generation of hypervalent iodine compounds proceeded through the intermediacy of acetate-stabilized iodanyl radicals (i.e. I(II) species; Figure 1). 15 We speculated that electrolysis of aryl iodides in the presence of acetate sources may provide access to the same iodanyl radicals and thus enable hypervalent iodine electrocatalysis. Here, we report that facile acetate-dependent anodic oxidation of aryl iodides enables hypervalent iodine electrocatalysis of both intra-and intermolecular C-H / N-H coupling reactions.</p><p>We initiated our investigations of hypervalent iodine electrocatalysis by examining the intramolecular C-H/N-H coupling with N-([1,1'-biphenyl]-2-yl)acetamide (1a) to afford Nacetylcarbazole (2a) in the presence of stoichiometric amounts of aryl iodide mediators (Figure 2). 16 Based on the relative onset potentials for oxidation of various substituted aryl iodides (Table S1), we selected 4-iodoanisole (3a) as an initial catalyst. Constant potential electrolysis (CPE) of a mixture of 1a and 3a in 1,1,1,3,3,3-hexafluoroisopropanol (hfip) with 0.2 M [TBA]PF6 as supporting electrolyte resulted in no desired C-N bond forming chemistry and partial decomposition of the starting material (87% of 1a was recovered following electrolysis; TBA = tetrabutylammonium). Based on the hypothesis that acetate ligands can stabilize initially generated iodanyl radicals, 15 we examined the impact of added [TBA]OAc on the electrolysis of 3a and found that 2.0 equivalents of [TBA]OAc, with respect to substrate, promotes electrochemical C-N coupling in 61% yield (see Table S2 for results from varying [TBA]OAc loading). No C-N coupled product was obtained in absence of aryl iodide. The loading of 4-iodoanisole can be lowered; we find the 25 mol% catalyst loading gives the best yield of 76% and that further reduction of the catalyst loading leads to attenuation of the reaction efficiency (Table S3).</p><p>Examination of other solvents, N-protecting groups, reaction temperatures, and electrode materials did not result in substantively better reaction efficiency (Tables S3-S5). Redox balance in the observed chemistry is achieved by proton reduction (presumably of hfip) to generate H2, which was observed by GC analysis of the reaction headspace (Figure S1). With the optimized conditions in hand, we evaluated the scope of intramolecular C-N bondforming chemistry (Figure 2). We specifically examined the impact of substituents in the 4-and 5'-positions of the [1,1'-biphenyl]-2-acetamide scaffold (i.e. 1) under constant potential conditions with 3a as catalyst. We found that both 4-and 5'-halogenation are well tolerated (2b-2g), as is the introduction of weakly electron withdrawing groups like 4-aldehyde (2h) or 5'-phenyl (2i). Under these conditions, substrates with more electron-withdrawing substituents, such as nitro (2j) and ester (2k), did not afford the expected carbazole. Based on the hypothesis that these more electron-</p><p>deficient substrates may require a more oxidizing hypervalent iodine catalyst, we employed 2,2'diiodo-4,4',6,6'-tetramethyl-1,1'-biphenyl (3b) 17 as catalyst (onset potential for oxidation is 1.78 V vs Ag + /Ag for 3b compared to the onset potential of 3a which is 1.43 vs Ag + /Ag). The more oxidizing conditions allowed both 3j and 3k to be accessed (43% and 71% yields, respectively).</p><p>Electron donating groups were tolerated in the 5'-position (i.e. 2l and 2m). In contrast, introduction of methyl and methoxy groups at the 4-position (i.e. 1n and 1o) led to trace amount of carbazole and starting material decomposition. We speculate that the presence of electron donating substituents at the 4-position decreases the onset potential for direct substrate oxidation below that of the aryl iodide catalyst and thus leads to direct substrate activation (for CV analysis, see Figure S2-S4). Consistent with this hypothesis, CPE of 4-t-butyl acetamide (1p) in the absence of aryl iodide catalyst afforded 2p in 65% yield (none of the other substrates in Figure 2 participate in C -N coupling chemistry in the absence of 3). 18 In general, the broader tolerance for substitution in the 5'-position than the 4-position is consistent with the smaller impact of substituents in this position on the onset potential for direct substrate oxidation (Figure S2). Finally, C-N bondformation can be accomplished in multifunctional substrates, as highlighted by the synthesis of 2p, a precursor to anti-HIV natural product clauszoline-K. 19 Hypervalent iodine electrocatalysis can also be applied to intermolecular C-N bond-forming chemistry (Figure 3). Catalyst 3b was used as the aryl iodide mediator due to its previously reported success in furnishing intermolecular C-H amination reactions (see Table S6 for analysis of other aryl iodide catalysts). 20 CPE in the presence of 1 equivalent of 3b affords 81% yield of Nphenylated compound 5a and the loading can be decreased to 25 mol% without significantly depressing the yield (71% of isolated product). Similar to the above-described intramolecular C-N bond-forming chemistry, no C-N coupling products are observed in the absence of either aryl iodide or [TBA]OAc. The intermolecular C-H amination reactions with halogenated aryl group were accomplished in 35-82% yields (5b-f). Electron-rich hydrocarbons like toluene, xylene, and naphthalene were not viable substrates in the intermolecular N-H arylation. Compound 5a can be elaborated to the corresponding arylhydrazine (7), which are useful precursors to various heterocyclic compounds, by treatment with hydrazine (Eq 1). 21 Alternately, hydrogenolysis of the N-N bond leads to N-acyl aniline derivatives (6), which can be challenging to synthesize by transition-metal-catalyzed cross-coupling reactions due to the stability of ammonia adducts of many transition metals. 22 (1)</p><p>Hypervalent iodine catalyzed C-N bond forming chemistry is most often accomplished with peracid terminal oxidants. 23 While recently developed aerobic hypervalent iodine chemistry proceeds through a distinct one-electron mechanism, the autoxidation chemistry used to couple O2 reduction to aryl iodide oxidation, produces a significant steady state concentration of peracid. For this reason, the developed aerobic oxidation conditions display substrate scope limitations similar to those displayed by peracid conditions when assayed by robustness analysis. 15,24 We were interested in evaluating the robustness of the developed hypervalent iodine electrocatalysis to evaluate if a broader functional group tolerance may be achieved by avoiding the use or evolution of peracids. Figure 4 displays both the impact of a variety of small-molecule additives on the yield of intramolecular C-H/N-H coupling as well as the amount of recovered additive following the electrochemical reaction (see also Figure S5). The efficiency of electrocatalytic C-N coupling is superior to aerobic conditions for all additives and similar to that of peracetic acid. The electrochemical conditions display higher additive recovery, in particular when challenged against oxidatively labile functional groups, such as alkynes and olefins (96% and 92%, respectively). We envision two potential limiting mechanisms for the electrosynthesis of hypervalent iodine compounds (Figure 5). The initial interfacial electron transfer could be between the working electrode and the aryl iodide to generate an I(II) intermediate ( 8), which would then be trapped by exogenous acetate ion to generate acetoxy iodanyl radical 9. Subsequent oxidation chemistry would ultimately lead to I(III) (10). Alternatively, the observed acetate-dependent chemistry might arise from an interrupted Kolbe electrolysis 25 in which initially formed acetoxy radicals add to aryl iodides to generate iodanyl radical intermediates (9), which would subsequently undergo further oxidation to I(III) (10). Available evidence, summarized below, is most consistent with the former mechanistic scenario. Examination of the cyclic voltammogram (CV) of iodoarenes as a function of scan rate indicate that while the oxidation is irreversible at low scan rates (i.e. < 100 mV/sec), chemical reversibility emerges at higher scan rates (>250 mV/sec, Figure S6-S7). 26 Addition of [TBA]OAc to a CV experiment of either 4-iodoanisole or 4-ioodtoluene results in both the loss of reversibility and the substantial increase in the anodic current (Ipa), indicating that the electrochemically generated species is trapped by added acetate (Figure S8-S10). The measured peak potential is linearly correlated with the square root of scan rate, which indicates electron transfer from a solutionbound, not surface adsorbed, species (Figure S11-S12). 27 Regarding the potential that the reported hypervalent iodine chemistry arises from an interrupted Kolbe electrolysis, we observed that electrolysis of CH3CN solutions containing [TBA]OAc and [TBA]PF6 results in the products expected of Kolbe electrolysis: ethane, methane, 1 OAc AcO (a) 1) -e - and CO2 (observed by GC analysis of reaction headspace, Figure S13). In contrast, acetate oxidation is suppressed in hfip, the solvent in which the chemistry is (uniquely) effective (no volatiles are observed in headspace analysis as well as no oxidation peak in the CV (Figure S14-S15)). The suppression of Kolbe electrolysis is consistent with strong hydrogen bonding of acetate to the acidic O-H of hfip (pKa = 9.3). 28 Consistent with this hypothesis, NMR analysis of solutions containing both hfip and [TBA]OAc reveals a significant downfield shift in the methine resonance (Figure S16). Job analysis indicates a 1:1 adduct is formed between these two species (Figure S17 and Table S7) and NMR analysis provides an equilibrium constant for association of 0.767 (Figure S18-S20 and Table S8).</p><p>In summary, we report the first example of hypervalent iodine electrocatalysis for C-H amination chemistry. The developed chemistry is applicable to both intra-and intermolecular C-</p>

ChemRxiv

A simple methodological validation of the gas/particle fractionation of polycyclic aromatic hydrocarbons in ambient air

The analysis of polycyclic aromatic hydrocarbons (PAH) in ambient air requires the tedious experimental steps of both sampling and pretreatment (e.g., extraction or clean-up). To replace pre-existing conventional methods, a simple, rapid, and novel technique was developed to measure gas-particle fractionation of PAH in ambient air based on 'sorbent tube-thermal desorption-gas chromatograph-mass spectrometer (ST-TD-GC-MS)'. The separate collection and analysis of ambient PAHs were achieved independently by two serially connected STs. The basic quality assurance confirmed good linearity, precision, and high sensitivity to eliminate the need for complicated pretreatment procedures with the detection limit (16 PAHs: 13.1 ± 7.04 pg). The analysis of real ambient PAH samples showed a clear fractionation between gas (two-three ringed PAHs) and particulate phases (five-six ringed PAHs). In contrast, for intermediate (four ringed) PAHs (fluoranthene, pyrene, benz[a]anthracene, and chrysene), a highly systematic/gradual fractionation was established. It thus suggests a promising role of ST-TD-GC-MS as measurement system in acquiring a reliable database of airborne PAH.The presence of polycyclic aromatic hydrocarbons (PAHs) in ambient air is mostly due to anthropogenic processes, particularly incomplete combustion of organic fuels like coal, oil, and gas 1 . The International Agency for Research on Cancer (IARC) has designated some PAHs (e.g., benzo[a]pyrene (BAP), and benz[a]anthracene (BAA)) as known human carcinogens 2 . In particular, BAP is listed as a "level 1 carcinogenic substance" 2-4 . PAHs are present in the atmosphere in both gaseous and particulate phases. Because the concentrations of PAHs in ambient air are low, e.g., at or below several ng•m −3 levels, the experimental procedures for their sampling and analysis are sufficiently complicated to suffer from large uncertainties due to low recovery despite large sampling volume 5 .In order to accurately quantify trace level PAHs in ambient air, a highly sensitive analytical system must be employed, even after acquisition of a large sampling volume. Many researchers have relied on protocols such as the US Environmental Protection Agency (EPA) method to calculate quantities from ambient air samples [6][7][8][9] . Accordingly, atmospheric samples should be collected using high-volume air samplers (HVAS), which allow selective collection of gaseous and particulate PAHs on a quartz (or glass) filter and polyurethane foam (PUF) sampler, respectively. In such a sampling system, a flow rate of approximately 0.1 to 1 m 3 min −1 is recommended with a sampling time of about 0.5 to 3 days. The PAHs collected by the filters and PUF sampler are subjected to Soxhlet extraction with appropriate solvent, and the extract is concentrated by an evaporator. Then, the extracts are repeatedly passed through a silica column for purification (over 24 hours). The re-concentrated extracts are analyzed by gas chromatography (GC) coupled with mass spectrometry (MS) or by high-performance liquid chromatography

a_simple_methodological_validation_of_the_gas/particle_fractionation_of_polycyclic_aromatic_hydrocar

<!>Results and Discussion<!>Detectability of PAHs in terms of LOD and MDL.<!>Comparison of detectabilities associated with sampling and pretreatment procedures.<!>Test of breakthrough volume of PAHs on the ST.<!>Method<!>Conclusions<!>Methods<!>Experimental approaches.

<p>(HPLC) equipped with a fluorescence detector (FLD) 6,7,10 . In addition to this complicated pretreatment procedure, the US EPA method further requires the use of internal standards to balance low recovery stemming from complicated extraction procedures.</p><p>In this study, a highly simplified technique for the analysis of PAH in ambient air was developed and validated using sorbent tube (ST) sampling and thermal desorption (TD)-GC-MS analysis (Table 1). We were able to accurately describe the distribution of PAH in different phases with reliable QA data and to eliminate complicated sampling and pretreatment procedures. To this end, the minimum range of air sampling volume (e.g., less than several m 3 ) was first investigated in order to create optimal conditions for the PAH analysis by ST/TD-GC-MS. The reliability of each process was examined further through the assessment of basic QA parameters (detectability, recovery, and breakthrough volume). This method was employed to explore the fractionation patterns between gaseous and particulate phases from a series of real ambient samples.</p><!><p>Results of PAH calibration experiments based on the ST method. In order to assess the reliability of our ST method for PAH analysis, the calibration results of 16 target PAHs were obtained by analyzing the liquid-working standard (L-WS) using a TD-GC-MS system (Table S1). All target PAHs had fairly good linearity with R 2 values > 0.99 (mean R 2 ± SD (n = 16) = 0.9987 ± 0.0015). The analytical precision of PAH determination was also assessed in terms of relative standard error (RSE, %) using triplicate analyses of the third calibration point of the L-WS (BAP = 4.98 ng•μ L −1 ; analytical volume = 1 μ L). The</p><!><p>In this study, the detection limits (DL) of target PAHs were determined based on the ST method using the L-WS in two common ways: (1) method detection limit (MDL) and (2) limit of detection (LOD) (Table S2). Although the MDL is more representative of actual or practical detection limits than the LOD, the LOD values were commonly reported as the ultimate limit of detection (as their DL) in many previous studies of PAHs. It should however be noted that the use of the LOD values can be misleading in a practical sense, as they tend to be significantly lower than the MDL values 11,12 . In this research, to assess the gaps between the two contrasting DL concepts, the MDL values were also assessed according to the relevant US EPA guidelines. To this end, seven repetitive analyses were completed using a diluted L-WS (BAP = 49.8 pg•μ L −1 : 25-fold dilution of the lowest concentration standard) in order to determine the standard deviation (SD) in peak areas. The resulting SD values were then multiplied by 3.14 (Student's t-value at the 99.9% confidence interval) and divided by the response factor (RF) to yield the MDL in mass quantity (pg). On the other hand, LOD values were determined as three times the standard deviation of background noise (n = 7). The MDL values for 16 target PAHs were found to range from 5.19 pg (BKF) to 27.0 pg (ACL) (mean 13.1 ± 7.04 pg), while the LOD values were in the range of 0.57 pg (DBA) to 1.23 pg (FLR) (mean = 0.76 ± 0.20 pg). A wide range of DL values of PAHs have been reported, from several pg to ng levels. As shown in Table 2, the DL values of this study are considerably lower than those in many previous studies. For example, Ma et al. 13 analyzed PAHs using the US EPA method (HVAS sampling, extraction-cleanup-concentration (E-C-C) pretreatments, and GC-MS detection) and reported MDL values of 9.21-25.3 pg, which are similar to those of the present study (the range of MDL for 16 PAHs: 5.19-27.0 pg). In contrast, using the same analytical method as Ma et al. 13 , others have reported MDL values one to two orders of magnitude higher (66.2-497 pg and 200-1,000 pg, respectively) 14,15 . Interestingly, some researchers who employed an HPLC-FLD system after the E-C-C pretreatment procedures presented LOD values of 0.58-7.99 pg and 0.70-4.30 pg, respectively 16,17 . As such, these HPLC-based LOD values appear to be highly comparable to our results (the range of LOD for 16 PAHs: 0.57-1.23 pg). All types of methodological options involved in sampling/pretreatment/detection are summarized in Fig. 1. In Table 3, the different detection limits are also compared using the clarification criteria introduced in Fig. 1.</p><!><p>It is interesting to consider explanations for the observed differences in detectability of PAHs between different studies. The DL values for PAHs in this study were similar to or somewhat lower than those in many previous studies, probably due to differences in total sampling volume vs. the amount actually delivered into the detection system. Note that most previous studies generally relied on large sampling volumes (e.g., exceeding 1,000 m 3 ) to quantify ambient PAHs at sub ng•m −3 levels (Table 3). If PAH samples are 1).</p><p>subject to the E-C-C pretreatment procedures (US EPA method), the actual mass for quantification is substantially reduced as follows. First, only certain fractions of the filter used for PAH collection are included due in the extraction. Hence this approach is subject to dual loss effects (only a small fraction of the sample is used for the extraction and subsequent loss due to treatment) that lead to a noticeable reduction in MDL under the ideal recovery conditions for PAHs at 0.01% to 2% (Table 2).</p><p>To obtain sufficient quantities of analytes (PAHs), a high-volume air sampler has commonly been employed to collect air samples of 100 to 1,000 m 3 levels at high flow rates (> 100 m 3 min −1 ) over one to two days. In the present research, the volume size of PAH samples was reduced dramatically to 1.44 m 3 (flow rate of 2 L min −1 ) by employing a low flow rate mini-vacuum pump. Despite this small sample volume, the absolute mass collected for each sample amounted to 17.8 pg (BKF) ~ 161 ng (NAP), which is still about 3 to 1,500 times larger than the MDL values. We are currently extending our efforts to improve ST sampling more efficiently in order to considerably shorten the sampling time for rapid monitoring of PAHs.</p><p>Bates et al. 18 analyzed PAHs using a procedure similar to our TD-GC-MS method. Although they used a similar system for PAH analysis, their methods were limited in that the media used for collecting air samples (filter) and for analyzing standards (sorbent tube) were different from each other. They relied on LVAS to collect 24 m 3 of air on a quartz filter. However, as they were unable to establish the optimal conditions of ST-TD-GC-MS (e.g., sufficiently high temperature for sample transfer in a TD system), their DL values for the PAH analysis are considerably higher than ours (mean LOD: (1) Bates et al. 18 = 122 ± 69.0 pg vs. (2) This study = 0.76 ± 0.20 pg). The use of ST-TD-GC-MS system for the analysis of airborne PAHs can also be found from some other previous studies 19,20 . In those studies, the ST packed with Polydimethylsiloxane foam filter and Tenax TA was used to collect the PAH samples in air. In the case of Wauters et al. 19 , the DL values for PAHs were significantly low with mean 1.86 ± 0.79 pg (LOD, n = 16), but they did not present the MDL values. In addition, their quantitation was not made separately for each of particle and gaseous phase, as a single tube (packed with PDMS and Tenax TA adsorbent) was used to collect PAH in both phases for the TD-based analysis.</p><!><p>In this study, ambient PAHs were collected on the ST using a small vacuum pump. A total of 1.44 m 3 of air was drawn for 12 hours at an air flow rate of 2 L min −1 . The required sample volume for PAH analysis is quite small (1.44 m 3 ) compared to those in most previous studies, but the breakthrough volume (BTV) of PAHs on the ST sampler needs to be assessed for accurate quantification. To examine the BTV of PAHs on the ST, N 2 gas was purged to the ST with six different volumes after loading the L-WS.</p><p>Table S3 shows the mass recovery of PAHs in the CC tube with different purge volumes in order to test the BTV. The PAH mass recovery was calculated using the RF values (ng −1 ) obtained by L-WS analysis: (1) Measured mass (ng) = Peak area / RF value (ng −1 ) and (2) Relative recovery (%) = Measured mass (ng) / Injected mass (ng) *100. For the total purge volume of 1 L, the recovery of PAH averaged 72.6 (± 3.89%: SD). Likewise, at the low purge volume of 1 L, adsorption-partitioning equilibria were not attained between analytes and sorbents in the CC tube, resulting in poor recovery. However, if the purge volume increased above 9 L, all target PAHs had sufficiently high recoveries (> 99%; mean recovery = 99.5 ± 0.50%). These high PAH recoveries were maintained up to the maximum tested purge volume of 2.52 m 3 (mean recovery = 99.1 ± 1.29%). The results of our BT point test did not directly identify the BT but confirmed the importance of a purge step to ensure optimal recovery. As a result, we were able to predict that the BT of PAHs on the ST should not occur during routine sampling (e.g., up to 1.44 m 3 of sample volume). The partitioning behavior of PAHs in air between gas and particulate phases. As a means to demonstrate the feasibility of our ST method for PAH analysis, ambient PAH samples were continuously measured using the ST method. The PAH sampling was conducted on the seventh floor of the Jae Sung Engineering Building, HanYang University, for five successive days in Oct. 2014. The QC sampler (QW + CC tubes) was used as sampling media to separately collect the particulate and gaseous PAHs from outdoor air. In addition, triplicate samples of ambient PAHs were also simultaneously collected and analyzed using three QC samplers in order to test the reproducibility (or compatibility) of the QC sampling method. ). b Below method detection limit. c Not available data: PAHs were not detected from one or both phases (particle and gas) in air.</p><!><p>data exhibited roughly four-fold variation during these five days. When the relative proportions of individual PAHs were compared against their total concentration, NAP demonstrated the highest value, with a mean of 66.8% ± 7.82% (Figure S1). Thus, the total PAH concentration was most strongly influenced by the NAP concentration. In contrast, five-to six-ringed PAHs had very low concentrations, typically below 1 ng•m −3 . The results of triplicate analyses of ambient PAHs confirmed that NAP was predominant (the relative proportion of NAP = 59.1 ± 3.34%) (Table 4): total PAH concentrations of triplicate analyses (A, B, and C) were 110, 88.4, and 77.6 ng•m −3 , respectively (mean: 91.9 ± 16.2 ng•m −3 ). Although ambient PAHs were sampled at the same time, the total PAH values varied moderately between different samples (RSE = 10.2%). However, if the NAP from these triplicate analyses was excluded from the total concentration, the compatibility between triplicates increased greatly to 40.6 (A), 37.9 (B), and 33.2 (C) ng•m −3 (RSE = 5.76%) (Table S4). The chromatograms of 16 target PAHs detected from outdoor air are presented in Fig. 2 using the results from the fourth day (10 Oct. 2014).</p><p>The target PAHs showed a clear partitioning pattern between gas and particulate phases, especially based on such simple criteria as the number of aromatic rings and/or molecular weight (Figure S2). The two-and three-ringed PAHs existed mainly in the gas-phase (mean gas fraction = 96.6% ± 4.01%). In the case of four-ringed PAHs, systematic fractionation was established across the particle/gas boundary. For instance, in the particle fraction, FLT and PYR (molecular weight = 202 g/mole) remained at 40.3% ± 6.90%, while BAA and CHY (molecular weight = 202 g/mole) were much more abundant (mean 77.4% ± 5.34%). All five-and six-ringed PAHs were detected predominantly in the particulate phase (mean particle fraction = 87.4 ± 4.71%). As such, the particle/gas partitioning ratio increased consistently and systematically with increasing molecular weight.</p><p>The results of particle-gas partitioning patterns in this study are thus very similar to those reported from many previous studies based on conventional methods (e.g., US EPA methods) (Table 5). Simcik et al. 21 quantified gaseous and particulate PAHs in outdoor air (sampling sites: (1) Chicago (southwest and north winds) and (2) Lake Michigan (southwest and north winds, USA)) using a glass fiber filter and PUF sampler, respectively. Accordingly, the fraction of three-ringed PAHs (FLR, PHN, and ANT) in the particulate phase averaged 3.03% ± 3.01%, while those of five-and six-ringed PAHs (BBF, BKF, BAP, BGP, and benzo[e]pyrene) comprised a high proportion (mean 88.9% ± 8.90%). In the case of four-ringed PAHs, the particulate fractions of 202 g/mole (FLT and PYR) and 228 g/mole (BAA and CHY) were measured as 15.1% ± 10.4% and 60.1% ± 12.1%., respectively. Despite differences in experimental approaches compared to our study, the results of Simcik et al. 21 also showed a systematic fractionation of PAHs in air to be strongly associated with molecular weight. This type of particle-gas partitioning pattern of ambient PAHs was in fact observed not only in Simcik et al. 21 , but also in many other previous studies. Ma et al. 13,22 analyzed ambient PAHs at Harbin and Beijing, China using an analytical method comparable to that of Simcik et al. 21 and reported that five-to six-ringed PAHs had high fractionation in the particulate phase (> 99%). In contrast, the particle fractions of two-and three-ringed PAHs were relatively low, with means of 17.4% and 5.93%, respectively. Albinet et al. 17 were unable to detect heavy PAHs with more than five aromatic rings in the gas fraction.</p><!><p>In order to analyze ambient PAHs at sub ng•m −3 levels, large sampling volumes (and long sampling times of up to a few days) and complicated pretreatment procedures (such as extraction, clean-up, and concentration) are required. In addition, the use of an internal standard is required in order to balance low recovery stemming from the loss of analytes due to the complicated pretreatment. In this study, a novel technique for PAH analysis was developed using ST sampling coupled with a TD-GC-MS system and was validated against real ambient air samples. To this end, basic calibration and QA data for PAH analysis were acquired by ST-based analysis. Then, ambient PAH samples were collected continuously over a five day period in October 2014 using the ST and analyzed using a TD-GC-MS system. In addition, to remove the carry-over effect of the ST-based analysis, conditioning of ST was carried out in all stages of PAH analysis.</p><p>All 16 target PAHs had fairly good linearity (R 2 > 0.99) and reproducibility (RSE < 1%) according to the ST-based analysis of the liquid-phase PAH standards. In addition, the MDL values of all PAHs as determined by the ST method were very low, with a mean of 13.1 pg. In addition to the simplicity of the ST method (without pretreatment procedures or an internal standard), it is possible to accurately quantify ambient PAHs (at sub ng•m −3 levels) at sufficiently low sampling volume (1 m 3 level). For a daily ambient sample of 1.44 m 3 , the total concentration of target PAHs averaged 78.8 ng•m −3 over a five day period. Light PAHs were detected predominantly in the gas phase (sampled by the Carbopack C tube), while heavy PAHs existed mainly in the particulate phase (collected by a quartz wool tube). In the case of BAP, the mean of 0.21 ng•m −3 was detected from the ambient samples, representing a particle fractionation of 86.5%. The four-ringed PAHs showed dynamic fractionation between gas and particulate phases. This study thus successfully demonstrated the feasibility of ST-based sampling and TD-based analysis Method type-Phase/Sampler-Pretreatment-Separation/Detection (refer to the Fig. 1).</p><p>for accurate and reliable quantification of PAHs. In addition, for practical application of the ST method, we confirmed that a reasonably small (1 m 3 ) sample volume is sufficient. Due to the thermal desorption procedure employed for PAH analysis in this study, we were able to simplify the pretreatment procedures for the optimum recovery of the PAHs. This ST method can thus be used to establish a routine monitoring system for PAH and to replace the methods or procedures based on conventional systems.</p><!><p>Preparation of liquid PAH standards. A total of 16 PAHs promulgated as priority pollutants by the US EPA were selected as the target analytes in this research (Table S5): S6). Preparation of sorbent tubes. The ST for the collection and analysis of 16 target PAHs was prepared to calibrate L-WS and to quantify real ambient samples. The feasibility of the ST method for PAH analysis has been explored in many previous studies 18,23 . However, the use of ST in those studies was confined to calibration only. Earlier researchers encountered some technical limitations due to the difficulty of increasing sampling volume with an ST sampler, although it is crucial to acquire a sufficient quantity of analytes for detection. Hence, the use of glass or quartz filters adaptable to large volume sampling (> tens of m 3 ) was used in order to expand the sampling capacity of ambient air to the maximum level. Consequently, the application of ST in the TD-based analysis was confined to standard calibration experiments rather than for actual sampling 18 . In light of the physical differences in media used for environmental sampling (filter) and standard calibration (ST), the objectivity of the QA data in previous studies is somewhat questionable 18,24,25 . In the present study, to overcome diverse technical or practical problems encountered in many previous TD-based analyses, we designed a new approach to maintain the compatibility of the sampling methods by employing the same ST for both standard calibration and sample analysis.</p><p>In this study, for the collection and analysis of all target PAHs in gas and particle fractions, we prepared two types of ST. The first tube was filled with 50 mg of Carbopack C (60/80, Supelco, USA) applied to 10 mg of quartz wool (Supelco, USA) (the "CC tube"). Carbopack C was selected as the sorbent in order to induce optimal adsorption of gas-phase PAHs 26,27 . The second tube was packed solely with 25 mg of quartz wool (QW tube) and was used to capture particulate PAHs.</p><p>Calibration and QA experiments were performed using the CC tube loaded with known quantities of the L-WS containing PAHs. For analysis of real PAH samples in ambient air, collection was performed using serially connected QW and CC tubes (Fig. 3). The resulting tube was called the "QC" tube in order to represent a combination of "QW and CC" in the sampling stage. The front and back fractions of QC were thus used to collect the particulate and gaseous PAHs, respectively (Fig. 4).</p><p>Instrumental system. The analysis of PAH samples in this work was carried out using a GC (model: GC-2010, Shimadzu, Japan) connected to an MS (model: GCMS-QP2010 ultra, Shimadzu, Japan) and a thermal desorber (model: TD-20, Shimadzu, Japan). The schematic and operational conditions of the TD system were set to maximize PAH recovery by virtually eliminating the long transfer line for carrying thermally desorbed PAH from the TD to the GC-MS. The PAHs loaded on the ST were thermally desorbed at 290 °C (7 min) at a reverse flow of 100 mL•min −1 of helium (> 99.9999%) carrier gas. The desorbed analytes were swept into the cold trap (held at 5 °C) in the stream of carrier gas. The cold trap packed with quartz wool (10 mg) and Tenax TA (50 mg) in a Silcosteel holder (Shimadzu, Japan) was then rapidly desorbed (300 °C for 5 min) in a reverse flow of carrier gas in order to transfer (inject) the target PAHs into the column (DB-5ms -length: 30 m, diameter: 0.25 mm, and thickness: 0.25 μ m, Agilent, USA). The transfer/injection of analytes from the cold trap into the GC column was carried out by splitting the flow between the column (2 mL•min −1 ) and the split vent (10 mL•min −1 ). The oven temperature was initially set at 80 °C (for 5 min), ramped at 20 °C•min −1 to 300 °C, and held at this temperature for 24 min (a total run time of 40 min).</p><p>To detect all 16 target PAHs, the interface and ion source temperatures were set relatively high (e.g., at 280 °C) in order to prevent contamination in the MS system. The PAHs were initially analyzed in total ion chromatographic (TIC) mode over a mass range of 35 to 600 m/z. Extracted ion chromatographic (EIC) mode was subsequently applied to minimize interference and to maximize the sensitivity using significant ions identified from the spectrum of each PAH (Table S5). Detailed information on the instrumental system is included in Table 1.</p><!><p>For the calibration and QA-related experiments, the L-WS containing 16 target PAHs was injected directly into the CC tube and analyzed using the TD-GC-MS system (Figure S3). The CC tube is stronger adsorbent than QW tube. As the analysis of CC tube is expected to show the maximum recovery of PAH, our calibration exp was conducted by CC tube only 27 . The inlet of the CC tube was connected to a gas cylinder containing ultra-pure nitrogen (> 99.999%). A Teflon tube was used to connect the ST and the gas cylinder. Then, 1 μ L of the L-WS was injected onto the CC tube via a temporary injection port made from the Teflon tube that connected the inlet of the CC tube and the gas cylinder. The nitrogen gas in the gas cylinder was then delivered to the CC tube (flow rate = 3 L•min −1 for 3 min). The PAHs loaded on the CC tube were then desorbed using the TD system prior to separation by the GC and final detection by the MS.</p><p>The collection of PAH in real ambient air samples was carried out by two-stage sampling with the aid of the QC sampler. This sampler was created as a combination of serially connected QW (front) and CC (back) tubes for collecting particulate and gaseous PAHs, respectively (Fig. 3). The particulate PAHs were first introduced into the QW tube placed at the front, and the gaseous PAHs penetrating the QW tube were then collected by the CC tube (Fig. 4). The outlet of each QC tube was connected to the vacuum pump that interfaced with a mass flow controller (MFC) (Sibata Σ MP-300, Japan). To measure the gas/particle fractionation of PAHs in air, two types of ST (QW and CC) were used simultaneously as PAH sampling media. However, they were desorbed individually for the analysis of PAH partitioned to each individual fraction. The PAH sampling from outdoor air was conducted on the seventh floor (about 21 m) of Jae Sung Engineering Building (HanYang University, Seoul, Korea) for a period of five consecutive days (7 to 11 Oct. 2014). The collection of PAH samples continued for 12 hours starting at midnight (flow rate = 2 L•min −1 and total sampling volume = 1.44 m 3 ) (Fig. 3). For QC tubes, five code numbers, 1, 2, 3, 4, and 5, were assigned to samples obtained for each of five days. The sample codes were further categorized by assigning a number (order of day) and tube type (QW or QC) such as QW (QW-1, QW-2, QW-3, QW-4, and QW-5) and CC (CC-1, CC-2, CC-3, CC-4, and CC-5). In addition, to test the reliability of QC sampling, triplicate samples of ambient PAHs were simultaneously collected using three QC samplers (Set code: A, B, and C) on 7 Sept. 2014 (sample codes: QW (QW-A, QW-B, and QW-C) and CC (CC-A, CC-B, and CC-C)).</p><p>In ST analysis, the breakthrough (BT) volume is one of the key criteria for accurate quantification. In this study, a total volume of 1.44 m 3 ambient sample was collected at 2 L•min −1 for 12 hours. The occurrence of BT on the CC tube was examined by increasing the sampling volume of ambient PAH. To this end, the third calibration point of the L-WS (BAP = 4.98 ng•μ L −1 ) was injected onto the CC tube and purged with nitrogen gas up to 2,520 L at a fixed flow rate of 3 L•min −1 (six volumes tested between 1 and 2,520 L). The procedures for BT test using the L-WS were adopted from those used for the analysis of the L-WS but at varying purge volumes (by controlling the purge times). After this purge stage, the CC tubes were analyzed using a TD-GC-MS system.</p>

Scientific Reports - Nature

Structure of the active pharmaceutical ingredient bismuth subsalicylate

Structure determination of pharmaceutical compounds is invaluable for drug development but is challenging for those that form as small crystals with defects. Bismuth subsalicylate (BSS), among the most commercially significant bismuth compounds, is an active ingredient in over-the-counter medications such as Pepto-Bismol, used to treat dyspepsia and H. pylori infections. Despite its century-long history, the structure has remained unknown. Three-dimensional electron diffraction and hierarchical clustering analysis were applied on select data from ordered crystals, revealing a layered structure. In other less ordered crystals, high-resolution scanning transmission electron microscopy revealed variations in the stacking of layers. Together, these modern electron crystallography techniques provide a new toolbox for structure determination of active pharmaceutical ingredients and drug discovery, demonstrated by this study of BSS.

structure_of_the_active_pharmaceutical_ingredient_bismuth_subsalicylate

<p>The physical, chemical and therapeutic properties of active pharmaceutical ingredients (APIs) are governed by their molecular structures and intermolecular interactions. Many APIs are crystalline substances, with periodic arrangements of their constituent molecules or ions. The specific arrangement of molecules and their intermolecular interactions affect the APIs stability and solubility, which in turn influences bioavailability, efficacy, and dosage. Therefore, determining the structures of pharmaceutical compounds is an integral part of drug formulation. Traditionally the method of choice for crystal structure determination has been single-crystal X-ray diffraction (SCXRD). However, the technique requires large specimens and is not readily applicable for sub-micrometer-sized crystals. While structure determination of small crystallites can often be performed by powder X-ray diffraction (PXRD), the technique can at times struggle with complicated and disordered structures.</p><p>These reasons among others have previously prevented structure determination of the API bismuth subsalicylate (BSS), a crystalline compound made from Bi 3+ cations and salicylic acid (H2sal, Figure 1A). It is administered in its crystalline state, and is the API of popular over-the-counter medications, such as Pepto-Bismol, commonly used to treat gastrointestinal disorders such as dyspepsia and diarrhea. Numerous studies have confirmed the efficacy of BSS as an antimicrobial, anti-inflammatory, and antacid agent (1-3). Recent studies have even demonstrated that bismuth compounds can even combat antibiotic resistance in bacteria (4) and suppress SARS-CoV-2 replication (5).</p><p>BSS formulations were first developed in 1900 to treat Campylobacter infections, a major cause of infant deaths at the time (6). Since the discovery in the 1980s by Nobel laureates Barry Marshal and Robin Warren (7) of Helicobacter pylori, a bacterium harbored by 60% of the global population, bismuth compounds including BSS have been used to effectively treat peptic ulcer disease (8). In 1990, a report from Procter & Gamble (P&G) estimated that over 10 billion doses of Pepto-Bismol had been consumed and that it could be found in approximately 60 % of U.S. households (6).</p><p>In 2019 overall sales of more than 20 million units grossed over $100 million in the U.S. alone, making it the most-sold stomach remedy in the country (9). Despite its century-long history and continuing widespread use, the structure of BSS has remained unknown and only a limited understanding has been established of its mechanisms of action. Speculation on the formula and structure of BSS has been published in many chemical and pharmaceutical databases, websites, patents, textbooks, and articles, where BSS is often represented as a simple metal complex. Although efforts have been made to determine its crystal structure, obtaining sufficiently large specimens of BSS for SCXRD has not been possible -likely due to its poor solubility in water. Due to the difficulties in characterizing BSS, several model bismuth compounds have been developed through various approaches, including the synthesis of bismuth thiosalicylates (10), the incorporation of water or organic solvent molecules into the crystal structures (11)(12)(13), or by altering the Bi:Hsal stoichiometry (14,15), which is 1:1 for BSS.</p><p>To finally uncover the structure of BSS, we turned to two modern transmission electron microscopy (TEM) techniques: three-dimensional electron diffraction (3DED) (16) and high-resolution scanning transmission electron microscopy (STEM). 3DED techniques, such as cRED, ADT, fast-EDT, and MicroED, can be applied to obtain single-crystal diffraction data on crystallites that are even smaller than 200 nm for the determination of their average ordered structures. This has been facilitated through the development of methodology and hardware in recent years (17)(18)(19)(20)(21)(22) and has allowed for faster and higher quality measurements for the structure determination of proteins (18,20,23), inorganics (24), and organics (25,26), including pharmaceuticals (27,28) as well as various bismuth compounds (29)(30)(31)(32). Concurrently, aberration-corrected STEM has evolved into an essential technique for atomic-scale structural investigations, particularly of local disorder. Recent development of STEM techniques such as integrated Differential Phase Contrast (iDPC) has allowed for studies on beam-sensitive specimens, including studies of organic molecules inside inorganic framework materials (33,34). To identify appropriate BSS crystals for detailed investigations, five samples from different suppliers or formulations were screened by PXRD (Figure S1). All samples were crystalline and had characteristic PXRD patterns of BSS. Based on the quality of the PXRD patterns and commercial significance, our investigation narrowed its focus to two samples: BSS purchased from the chemical provider Sigma-Aldrich (BSS-SA) and BSS isolated from Pepto-Bismol original liquid formulation (BSS-PB).</p><p>Inspection of both BSS-SA and BSS-PB samples by scanning electron microscopy (SEM) and TEM imaging revealed crystals with a plank-shaped morphology that appeared homogeneous with no obvious indications of impurities (Figures 1B, S2, and S3). Despite efforts to further elucidate the structure by synchrotron SCXRD, the crystals proved too small and agglomerated. Structure determination by PXRD was not successful due to the preferred orientation of the crystals causing both over and underemphasized intensities, but also due to diffuse scattering and a rather complicated crystal structure, as disclosed by 3DED experiments. 3DED datasets were collected at 98 K on crystals from BSS-PB which had been centrifuged out of suspension and washed with water. The quality of datasets suffered from a variety of problems including inadequate data resolution, irregular peak shapes, twinning, and in some cases diffuse scattering (Figure S4, sample preparation, data collection, and data processing for all materials are described in the supplementary material). A few of the datasets could be indexed to a triclinic unit cell, however, no reasonable structure model was obtained. Higher quality 3DED data were acquired from crystallites of BSS-SA (Figure S5). Datasets from 18 crystallites could be indexed with a triclinic unit cell (a = 8.35 Å, b = 12.17 Å, c = 18.09 Å, α = 77.9°, β = 83.2°, γ = 76.7°). Initial structure solution was attempted on individual 3DED datasets but was unsuccessful. This was attributed to the low completeness of the individual datasets (≤ 50 %) caused by the low symmetry of the crystals and the limited tilt range intrinsic to the TEM. To improve data completeness, hierarchical clustering analysis (HCA) was performed to merge datasets that were most similar in terms of measured reflection intensities (35). A distance-metric based on the correlation coefficient (CC) of overlapping data was generated for all possible pairs of datasets. This resulted in two separate clusters with a CC of at least 0.90, corresponding to a distance metric of 0.44 (Figure 1D). Structure solution by direct methods using the data of the large cluster, composed of 12 individual datasets with an overall completeness of 84.6 % (Figure 1D), resulted in a model with all non-hydrogen atoms appropriately located in the crystal structure with the space group P-1 (Table S1). As revealed by 3DED, BSS proved to be a coordination polymer with a layered structure (Figures 2A and S6). The crystal structure and its asymmetric unit of Bi4O4(Hsal)4 is in accordance with the commonly presented empirical formula of BiC7H5O4, as well as with elemental and thermogravimetric analyses (Figure S7). In BSS, Bi 3+ cations are bridged by O 2-anions into bismuth-oxo rods which extend along the a-axis and form the inorganic building unit (IBU) of the structure (Figure 2C). Along the center of the rods, O 2-anions bridge alternatingly three (µ3) and four (µ4) Bi 3+ cations. The IBU of BSS is nearly identical to one found in a previously reported bismuth-biphenyltricarboxylate coordination polymer (31). There are two types of salicylate anions (Hsal -) in the BSS structure. One type of Hsalcoordinates via the carboxylate group only to Bi 3+ cations of a single rod, while the phenol group does not coordinate to any Bi 3+ cations. The other type also coordinates to Bi 3+ cations through the carboxylate group; however, the phenol group coordinates to Bi 3+ cations in adjacent rods, essentially linking the rods along the b-axis into centrosymmetric layers in the ab-plane. These layers stack along the c-axis and only weakly interact with one another via dispersion forces. As the unit cell is only one layer thick, IBUs in neighboring layers are oriented in the same direction in the ordered crystal. The protonation, as assigned in Figure 2D, results in a charge-balanced material with phenol groups still protonated and carboxylic acid groups deprotonated and coordinating to the Bi 3+ cations. The carboxylate oxygen atoms form relatively shorter Bi-O bonds (2.6-2.8 Å) compared to the phenol oxygen atoms (2.8-2.9 Å), which suggest deprotonation of the carboxylic acid of salicylic acid rather than the phenol group, which is also supported by IR spectra (Figure S8). Similar protonation assignment of the Hsalligands has been reported in structures such as [Bi4O2(Hsal)8]•2MeCN/MeNO2 and [Bi(Hsal)3(H2O)] (13,36).</p><p>Due to the poor quality of the 3DED data on BSS-PB crystals, PXRD data were instead utilized to investigate the structure of BSS-PB. Structure refinement of BSS-PB showed overall good agreement with the BSS structure obtained by 3DED on BSS-SA (Figure S9 and Table S2). However, diffuse scattering and asymmetric peak shapes in the PXRD pattern (Figure S10) suggested the presence of disorder in the BSS-PB samples.</p><p>To further validate the structure of BSS-PB and investigate structural disorder, aberration-corrected scanning transmission electron microscopy (STEM) imaging was applied. Imaging was performed using both annular dark-field (ADF) as well as integrated Differential Phase Contrast (iDPC) signals. The ADF contrast scales rapidly with the atomic number, thus highlighting heavier elements such as Bi. The iDPC contrast, on the other hand, scales linearly with atomic number, thereby emphasizing lighter elements when compared to ADF (37). Crystals of BSS-PB consisted of large ordered domains in projection consistent with the structure of BSS-SA determined by 3DED. Images along the [100] direction of BSS-PB revealed a similar orientation of the IBUs (Figure 3). The ADF contrast shows well-resolved projected positions of the Bi 3+ ions of the IBU (Figure 3A), whereas the iDPC contrast, in addition, showed enhanced contrast in locations consistent with the positions of the salicylate anions, although not as well resolved (Figure 3B).</p><p>However, upon inspection of other sections of the BSS-PB crystals, iDPC-STEM images revealed different types of disorder, particularly an inconsistency in the stacking of layers (Figures 4 and S11). It can be seen that the orientation of the layers vary, which can be caused by a two-fold rotation of the layers around the b-axis or perpendicular to the ab-plane. In some domains it was evident that the unit cell was doubled along the c-axis and the unit cell angle α changed due to a periodic alternation of the layer orientation. In other domains the orientations of the layers appear to be random and the disorder is observed as diffuse features in the Fast Fourier transform (FFT) of the image, as shown in Figures 4D-4F. The fact that domains of disordered sequences are observed explains the occurrence of inconsistent peak shapes and diffuse scattering in the PXRD pattern, as well as the initial difficulties in obtaining a structure model.</p><p>As such, the material appears to have (1) ordered domains with a c-axis of 17 Å with a single layer orientation (Figure 4A, area 1, and 4G), (2) ordered domains with a doubled c-axis of 34 Å and alternating layer orientation for adjacent layers (Figure 4A, area 3, and 4H), (3) domains of disordered stacking of the layers (Figures 4A, area 2, and 4C), as well as (4) defects where the orientation changes within an individual layer (Figure 4B). Elucidation of the structure of commercial BSS provides a major step towards understanding the properties of one of the most commercially significant bismuth compounds. The fact that BSS is practically insoluble in water and the hydrophobic properties of the powder can be partly attributed to the continuous structure of the coordination polymer, where the less polar section of the salicylate anions form the outer surfaces of the layers, while all ionic and hydrophilic components, such as the phenol, carboxylate, µ3-O 2-, µ4-O 2-and Bi 3+ , are contained within the layers. This hydrophobic character is also in alignement with the fact that BSS embedded in hydrophobic resin starts to dissolve from the (010) facets (Figure S12). In addition to its high stability in water, BSS also demonstrated decent stability in aqueous solutions of HCl (Figure S13). No changes were observed in the PXRD patterns of BSS treated at a pH of 3 or higher. At a pH of 2, a small proportion of BSS converted into bismuth oxychloride, and was fully converted at a pH of 1. Considering this, interactions between the hydrophobic exterior of the BSS crystals and the gastric lining could to some extent govern the pharmacodynamics of this long-used formulation (38). The various forms of analysis, including PXRD, 3DED, and STEM imaging, indicated that the two investigated samples, BSS-PB and BSS-SA, were built from the same layers but that the samples differed in terms of the degree of crystallinity and disorder in their structures. As such, it is expected that the stacking of the layers is a consequence of the synthesis conditions utilized by various producers. Understanding the differences in local ordering of BSS opens opportunities to develop new analogs containing unique stacking sequences or higher degrees of exfoliation, which may influence the properties and the efficacy of the API. Considering this, the combination of electron crystallography tools used, 3DED and STEM, is expected to become an important part of drug discovery and structure determination of active pharmaceutical ingredients.</p>

ChemRxiv

Regioselective synthesis of heterocyclic N-sulfonyl amidines from heteroaromatic thioamides and sulfonyl azides

N-Sulfonyl amidines bearing 1,2,3-triazole, isoxazole, thiazole and pyridine substituents were successfully prepared for the first time by reactions of primary, secondary and tertiary heterocyclic thioamides with alkyl- and arylsulfonyl azides. For each type of thioamides a reliable procedure to prepare N-sulfonyl amidines in good yields was found. Reactions of 1-aryl-1,2,3-triazole-4-carbothioamides with azides were shown to be accompanied with a Dimroth rearrangement to form 1-unsubstituted 5-arylamino-1,2,3-triazole-4-N-sulfonylcarbimidamides. 2,5-Dithiocarbamoylpyridine reacts with sulfonyl azides to form a pyridine bearing two sulfonyl amidine groups.

regioselective_synthesis_of_heterocyclic_n-sulfonyl_amidines_from_heteroaromatic_thioamides_and_sulf

Introduction<!><!>Introduction<!><!>Introduction<!>1-Alkyl-1,2,3-triazole-N-sulfonyl amidines<!><!>1-Alkyl-1,2,3-triazole-N-sulfonyl amidines<!><!>1-Alkyl-1,2,3-triazole-N-sulfonyl amidines<!><!>1-Alkyl-1,2,3-triazole-N-sulfonyl amidines<!><!>1-Alkyl-1,2,3-triazole-N-sulfonyl amidines<!>5-Arylamino-1,2,3-triazole-N-sulfonyl amidines<!><!>5-Arylamino-1,2,3-triazole-N-sulfonyl amidines<!><!>5-Arylamino-1,2,3-triazole-N-sulfonyl amidines<!><!>2-Aminothiazole-4-N-sulfonyl amidines<!><!>3-Methyl-5-phenyl-isoxazole-4-N-sulfonyl amidines<!><!>3-Methyl-5-phenyl-isoxazole-4-N-sulfonyl amidines<!>2,5-Bis(N-sulfonylamidino)pyridines<!><!>2,5-Bis(N-sulfonylamidino)pyridines<!><!>2,5-Bis(N-sulfonylamidino)pyridines<!>Conclusion<!>X-ray diffraction study<!>

<p>The biological activity, rich chemistry and technically useful properties of heterocyclic compounds have made them a focal point of science and industry over the years. Heterocyclic compounds including azoles and azines have been found in natural products, and they are included in the structures of nucleic acids, vitamins, antibiotics and in many types of synthetic drugs [1–12]. N-Sulfonyl amidines have received considerable attention because they exhibit various types of pharmaceutical properties and biological activities [13–21] and also have been used as interesting building blocks in organic synthesis [17–20] (Figure 1).</p><!><p>Examples of biological activity and interesting chemical reactivity of N-sulfonyl amidines.</p><!><p>An N-sulfonyl amidine was recently found to be a key group in acid–base-induced rearrangements of 1,2,3-triazoles and thiadiazoles [22]. A variety of methods have been developed for the synthesis of N-sulfonyl amidines. The most commonly used methods to prepare these compounds include the Cu-catalyzed multicomponent reaction of alkynes, sulfonyl azides and amines [23–31], the reaction of thioacetamide derivatives and cyclic thioamides with sulfonyl azides [22,32–33], the chlorophosphite-mediated Beckmann reaction of oximes with p-toluenesulfonyl azide [34], the sulfonyl ynamide rearrangement by treatment with amines [35], the sodium iodide catalyzed reaction of sulfonamide with formamide [36], and the condensation of sulfonamide derivatives with DMF–DMA [37].</p><p>A few representatives of N-sulfonyl amidines of heteroaromatic acids have been prepared and applied [22,32,38–40]. However, no efficient and general method to prepare a series of heterocyclic N-sulfonyl amidines has been elaborated so far. A new approach to N-sulfonyl amidines has been published recently, based on the reaction of thioamides with sulfonyl azides [33,41–42] (Figure 2).</p><!><p>Data on the synthesis of N′-sulfonylazole-4-carboximidamides.</p><!><p>This method was used successfully for the synthesis of N-sulfonyl amidines of aliphatic acids and benzoic acid, including biologically active compounds. On the other hand, reactions of thioamides with electrophilic reagents have often been used for the synthesis of various types of sulfur containing heterocyclic compounds [43–47]. This gives some promise to the development of a general and efficient method for the synthesis of N-sulfonyl amidines of heteroaromatic acids based on the reaction of heterocyclic thioamides with higly electrophilic sulfonyl azides.</p><p>With the purpose of the synthesis of heterocyclic N-sulfonyl amidines bearing various heteroatoms in the ring, namely nitrogen, sulfur and oxygen atoms, we have studied reactions of thioamides of 1,2,3-triazole-, isoxazole-, thiazolecarboxylic acids and 2,5-dithiocarbamoylpyridine with sulfonyl azides. Due to the high dipole moment, the presence of electronegative heteroatoms bearing electron lone pairs, one could propose alternative reactions which might make it difficult to find a general regioselective procedure for the synthesis of the target molecules in good yields. To the best of our knowledge, there are no examples for the synthesis of N-sulfonyl amidines of heteroaromatic acids through this reaction so far.</p><!><p>Since 1,2,3-triazole derivatives exhibit valuable biological and technical properties, and take part in various ring transformations and rearrangements [48–51], we decided to study reactions of 1-alkyl-1,2,3-triazole-4-carbothioamides 1a–d with aryl- and alkylsulfonyl azides 2a–f with, with the goal of affecting a "iminosulfonylation" (Scheme 1).</p><!><p>Synthesis of 1-alkyl-N-phenyl-N'-(sulfonyl)-1H-1,2,3-triazole-4-carboximidamides 3.</p><!><p>The thioamides 1a–d were prepared from the corresponding amides 4a–d by treatment with phosphorus decasulfide (Scheme 1). It is worth noting that amides of 1-alkyl-1,2,3-triazole-4-carboxylic acids are poorly represented in the literature and the methods of their preparation require the addition of alkyl azides to acetylene carboxylic esters and reactions of 2-diazomalonates with aliphatic amines [40,52]. The first approach leads to a mixture of two regioisomers and the second method involves the use of explosive diazo compounds. Therefore, such compounds are better prepared by a recently found method in our laboratory which includes the reaction of 4-acetyl-1,2,3-triazole 5a–d with aniline followed by a Cornforth rearrangement of the 1,2,3-triazole ring [52]. Alkyl- (2a,b) and arylsulfonyl (2c–g) azides were prepared, respectively, from the corresponding sulfonyl chlorides and sodium azides according to published methods (Figure 3) [53].</p><!><p>Starting compounds.</p><!><p>We have found that 1-butyl-1,2,3-triazole-4-carbothioamide (1c) reacts well with benzenesulfonyl azide (2c) in various solvents to form the desired 1-butyl-1,2,3-N-sulfonyl amidine 3n in diverse solvents such as n-butanol, n-propanol, toluene, ethanol, water and even under solvent-free conditions (see Table 1 for the yields and other circumstances).</p><!><p>Optimizations of the reaction conditions for the reaction of thioamide 1с with phenylsulfonyl azide 2ca.</p><p>aReaction conditions: 0.18 mmol of 1c, solvent (1 mL); bisolated yield.</p><!><p>From these data we can conclude that the yield of the final product is optimal for the reaction under solvent-free conditions. 1-Butyl-1,2,3-triazole 1с reacts faster than 1,2,3-triazole-4-carbothioamide 1f while using a lower amount of a sulfonyl azide (Table 1, entry 11 and Table 2, entry 14). Thus solvent-free conditions, a temperature of 88 °C and a thioamide/azide ratio of 1:2.5 are optimal to prepare N-sulfonyl amidine 1c (entry 11, Table 1).</p><p>Next, these optimized conditions were used for the synthesis of a small library of 1-alkyl-1,2,3-triazoles 3a–s (Scheme 2).</p><!><p>Scope for the reaction of 1-alkyl-1,2,3-triazole-4-carbothioamides 1a–d with azides 2a–f.</p><!><p>The reaction can be applied without problems to various alkyl substituents in position 1 of the 1,2,3-triazole ring from methyl to decyl and benzyl, goes well with alkylsulfonyl azides and arylsulfonyl azides that were 4-substituted with both electron-withdrawing and electron-donating substituents.</p><!><p>To further expand the scope of the reaction we continued studying the reaction of 1-aryl-1,2,3-triazole-4-carbothioamides 1e–h with aryl- and alkylsulfonyl azides 2a,c,f.</p><p>We have found that thioamide 1e did react with benzenesulfonyl azide (2c) neither in water, ethanol nor in the absence of a solvent, conditions that were successfully used in the synthesis of 1-alkyl-1,2,3-triazole-4-N-sulfonylimidamides 3a–s (Scheme 2). On the other hand, we have found the formation of a new product 3t in low yield together with the starting compound 1e and the product of its rearrangement to 5-(4-nitrophenyl)aminotriazole 1j [54], when the reaction was carried out in n-butanol at 105 °C (Table 2). Therefore, we can conclude that compound 3t was the product of a tandem reaction involving first the rearrangement of thioamide 1e to 1j followed by iminosulfonylation of the latter to form amidine 3t (Table 2).</p><!><p>Synthesis and optimization of the reaction conditions for the reaction of thioamide 1j with phenylsulfonyl azide (2c)a.</p><p>aReactions conditions: 0.45 mmol of thioamide 1j, solvent (3 mL); bIsolated yield.</p><!><p>To obtain higher yields of sulfonyl amidines we decided to prepare 5-arylamino-1,2,3-triazole-4-carbothioamide 1j by rearrangement of triazole 1e [54] and carried out an optimization with variations of the solvent, temperature and various additives (Table 2). We have shown that optimal conditions include the use of n-propanol, a temperature of 97 °C and a ratio of thioamide 1j and azide 2c of 1:7 which allowed to prepare the desired compound 3t in 78% (Table 2).</p><p>With the optimal conditions in hand we prepared a series of N-sulfonyl amidines 3t–aa in good yields (Scheme 3). Thus, a library of N-sulfonyl amidines bearing differently substituted 1,2,3-triazoles was successfully prepared. Among them are compounds bearing an NH-unsubstituted 1,2,3-triazole ring which gives extra possibilities for the modification of the molecules by the reaction with electrophilic reagents to prepare new compounds of this series [55] (Scheme 3).</p><!><p>Scope of the reaction of 5-arylamino-1,2,3-triazole-4-carbothioamides 1i–l with azides 2a,c–f.</p><!><p>To show the practical convenience of the developed method we tried to synthesize these compounds in a one-pot procedure starting from readily available 1-aryl-1,2,3-triazoles 1f,g,t and sulfonyl azides 2c,f (Table 3). Thioamides 1f,g,t were converted to 5-arylamino-1,2,3-triazoles 1i–k by heating at reflux in n-propanol in the presence of DBU and these rearranged thioamides were then treated with sulfonyl azide 2c,f and kept at the same temperature for 17‒31 h. After flash column chromatography, pure N-sulfonyl amidines 3t,u,x were isolated in 41‒65 % yield. The data of Table 3 demonstrates that the yields of sulfonyl amidines 3t,u,x are higher when we used the one-pot protocol in comparison with the two-step method. Furthermore, the one-pot procedure is obviously more simple and less time consuming.</p><!><p>Yields of triazoles 3t,u,x following a one-pot procedurea compared to the yields involving the isolation of 5-arylamino-1,2,3-triazoles 1i–k.</p><p>a1 (0.60‒0.65 mmol), DBU (0.63‒0.65 mmol), 2 (3.56‒4.0 mmol), HOAc (1 mL).</p><!><p>In spite of the presence of a nucleophilic amino group capable to react with sulfonyl azide to form an azide group, the reaction of azides 2 occurred selectively to the thioamide group of compound 1m.</p><p>Thus, similar to the reaction of 5-arylamino-1,2,3-triazole-4-carbothioamides 1i–l, the reaction of the primary thioamide of 2-aminothiazole-4-carboxyamide (1m) with sulfonyl azides 2a,c is succesful in n-propanol at reflux temperature, to afford N-sulfonyl amidines 3ab and 3ac bearing a 2-aminothiazole ring in very good yields (Scheme 4).</p><!><p>Synthesis of 2-aminothiazole-4-N-sulfonyl amidines.</p><!><p>The primary thioamide 1n containing an isoxazole ring was shown to react with mesyl azide or arylsulfonyl azides in n-propanol at reflux temperature to form the N-sulfonyl amidines 3ad–ag in 49‒76% yields (Scheme 5).</p><!><p>Synthesis of N-sulfonyl amidines of isoxazolylcarboxylic acid.</p><!><p>The reaction takes place also in the absence of a solvent, albeit in lower yields. We have found that secondary thioamide 1o does not react with sulfonyl azides 2a,c either in n-propanol or in the absence of a solvent. On the other hand, we have found that the reaction can occur in n-butanol at 118 °C to form compounds 3ah–ai in low yields (38‒45%) accompanied with the formation of tar-like products.</p><!><p>Bis(thioamide) 1p containing a pyridine ring was found to react with sulfonyl azides 2a,c–f either in boiling propanol or in the absence of a solvent to form compounds 3aj–an bearing two N-sulfonyl amidine fragments connected to a pyridine linker. The solvent-free protocol includes the use of a lower amount of azide 2d,c,f (2.5 equiv) in comparison with the reaction in n-propanol (4 equiv of azide) to afford the desired products in the same yield and therefore was selected as the method of choice for the synthesis of 3aj–an (Scheme 6). The synthesis of complexes of bis(sulfonyl amidines) 3aj–an with metals is in progress.</p><!><p>Synthesis of bis(sulfonyl amidines) 3aj–an.</p><!><p>1H and 13C NMR spectra including 2D HMBC and HSQC experiments of compounds 3a–an, as well as high-resolution mass spectra are consistent with the proposed structures. Carbon signals of the amidine groups of compounds 3 appear at 154.1‒159.7 ppm which is close to 156 ppm which is the value found for N-sulfonyl amidines of 1,2,3-thiadiazole-4-carboxylic acid prepared by another method [22] and was clearly different from the thioamide carbon signal at 185‒187 ppm in the 13C NMR spectra of starting materials 1. A final proof of the structures of the prepared compounds comes from the X-ray data for 3e,t,ag (Schemes 2, 3, and 5). Moreover, the X-ray data reveal the existence of N-sulfonyl amidines 3e,t in the E-isomeric form and N-sulfonyl amidine 3ag in Z-isomeric form. The existence of the latter in the Z-isomeric form can be explained by steric hindrance between the phenyl and the arylsulfonyl groups.</p><p>Because of the observed evolution of nitrogen and sulfur in every reaction of heterocyclic thioamides and sulfonyl azides it is logic to propose the formation of a thiatriazole ring via [3 + 2] cycloaddition of the azide group and the C=S moiety of the thioamide group (Scheme 7).</p><!><p>Plausible mechanism for the reaction of heterocyclic thioamides with sulfonyl azides.</p><!><p>The formation of nitrene-like products was excluded because of the high selectivity of the process, where only the thioamide group takes part, even with heterocyclic rings that contain other nucleophilic centers, and in one case, an amino group. Thiatriazoles are known to be unstable compounds that readily evolve nitrogen and sulfur upon heating [56].</p><!><p>We have shown that the reaction of sulfonyl azides with thioamides can serve as the basis for a general and efficient method for the regioselective synthesis of N-sulfonyl amidines of azolyl and pyridine carboxylic acids. The most promising aspect for organic synthesis and green chemistry is a solvent-free process which was successfully applied to prepare sulfonyl amidines containing pyridine and isoxazolyl rings and 1-alkyl-1,2,3-triazole-4-N-sulfonylamidino-1,2,3-triazoles. The 1-alkyltriazole thioamides are the most active in the solvent-free method due to their low melting points and good solubility in alkyl- and arylsulfonyl azides. Conversely, thioamides containing 5-arylamino-1,2,3-triazole and 2-aminothiazole rings are not soluble in sulfonyl azides and could be transformed to the corresponding N-sulfonyl amidines by reactions in 1-propanol via two- or one-pot procedures. Pyridine-2,6-dithioamide was shown to react with mesyl and arylsulfonyl azides to form pyridine derivatives bearing two N-sulfonyl amidine moieties in excellent yield. Depending on the structure of the heterocycle the N-sulfonyl amidines exist in either E- or Z-isomeric forms.</p><!><p>X-ray analyses were accomplished on an Xcalibur 3 diffractometer using the standard procedure (graphite-monochromated Mo Kα irradiation, ω-scanning with step 1o, T = 295(2) K (see Supporting Information File 1). Using Olex2 [57], the structures were solved with the Superflip [58] structure solution program using charge flipping and refined with the ShelXL [59] refinement package using least squares minimization. Deposition numbers for compounds 3e (2020829), 3t (2020831) and 3ag (2020830), contain the supplementary crystallographic data for this paper. These data can be obtained free of charge from the Cambridge Crystallographic Data Centre via http://www.ccdc.cam.ac.uk/data_request/cif.</p><!><p>Full experimental details and characterization data of all new compounds, crystal data and structure refinement for 3e, 3t, and 3ag.</p><p>Copies of NMR spectra of all new compounds.</p><p>Crystallographic information files for compounds 3e, 3t and 3ag.</p>

PubMed Open Access

DGCR8-dependent efficient pri-miRNA processing of human pri-miR-9-2

Microprocessor complex, including DiGeorge syndrome critical region gene 8 (DGCR8) and DROSHA, recognizes and cleaves primary transcripts of microRNAs (pri-miRNAs) in the maturation of canonical miRNAs. The study of DGCR8 haploinsufficiency reveals that the efficiency of this activity varies for different miRNA species. It is thought that this variation might be associated with the risk of schizophrenia with 22q11 deletion syndrome caused by disruption of the DGCR8 gene. However, the underlying mechanism for varying action of DGCR8 with each miRNA remains largely unknown. Here, we used in vivo monitoring to measure the efficiency of DGCR8-dependent microprocessor activity in cultured cells. We confirmed that this system recapitulates the microprocessor activity of endogenous pri-miRNA with expression of a ratiometric fluorescence reporter. Using this system, we detected mir-9-2 as one of the most efficient targets. We also identified a novel DGCR8-responsive RNA element, which is highly conserved among mammalian species and could be regulated at the epi-transcriptome (RNA modification) level. This unique feature between DGCR8 and pri-miR-9-2 processing may suggest a link to the risk of schizophrenia.

dgcr8-dependent_efficient_pri-mirna_processing_of_human_pri-mir-9-2

<!>Fluorescence-based live-cell pri-miR-9-1 processing reporter system<!><!>Pri-miR-9-2 is processed by the canonical microprocessor complex<!><!>Identification of the DGCR8-responsive RNA element in pri-miR-9-2<!><!>Exploration of pri-miRNA candidates possessing DRE<!><!>Role of DRE in DSCR8-dependent microprocessor activity<!>Discussion<!>Vector construction and plasmid preparation<!>Cell culture and transfection<!>Measurement of live-cell pri-miRNA processing activity<!>siRNA transfection<!>Quantification of miRNA, pri-miRNA, and mRNA by quantitative RT-PCR<!>Western blotting with the Wes system<!>miRNA profiling with an nCounter miRNA analysis system<!>Genomic Evolutionary Rate Profiling<!>CLIP-RT-qPCR assay<!>Statistical analyses<!>Data availability<!>Supporting information<!>Conflict of interest<!>Supporting information

<p>Edited by Ronald Wek</p><p>MicroRNAs (miRNAs) are small noncoding RNAs that regulate gene expression through specifically targeting mRNAs for degradation and translation inhibition (1, 2, 3). miRNAs are initially produced as long primary transcripts (pri-miRNAs), which are processed by the microprocessor complex composed of ribonuclease III, Drosha, and the RNA-binding protein DiGeorge syndrome critical region gene 8 (DGCR8) (4, 5, 6, 7, 8, 9, 10). Precursor miRNAs (pre-miRNAs) are exported from the nucleus into the cytoplasm by exportin-5, processed again by Dicer, and then loaded into the RNA-induced silencing complex (11, 12).</p><p>Several RNA-binding proteins, including DGCR8, function in pri-miRNA processing and are involved in neuronal development and diseases such as DiGeorge syndrome (1, 13, 14). The human nervous system expresses 70% of known miRNAs, and several brain-specific miRNAs have recently been identified (15). Among these, miR-9 is expressed specifically in neurogenic regions of the brain during neural development and adulthood (16, 17, 18, 19, 20, 21). miR-9 is encoded by three pri-miR-9 genes, pri-miR-9-1, pri-miR-9-2, and pri-miR-9-3, in human and mouse genomes (20). In humans, pri-miR-9 can be transcribed from chromosomes 1 (pri-miR-9-1), 5 (pri-miR-9-2), and 15 (pri-miR-9-3); the mature miRNA sequences generated from all three loci are identical. Only pre-miR-9-2 is expressed in neural stem cells differentiated from human induced pluripotent stem cells (22), and neither miR-9-1 nor miR-9-3 show robust expression in the developing human brain; miR-9-2 expression peaks by postgestation 16 weeks (23). A recent genome-wide association study (24), Schizophrenia Working Group of the Psychiatric Genomics Consortium, showed that the miR-9-2 gene is located near a single nucleotide polymorphism locus associated with schizophrenia with genome-wide significance (24). Moreover, a gene set enrichment analysis using summary statistics from the Psychiatric Genomics Consortium found an enrichment of predicted miR-9 target mRNAs among schizophrenia-associated genes (24). These results suggest that genetic variants in both miR-9 and its targets are associated with an increased risk of schizophrenia.</p><p>DiGeorge syndrome is caused by the chromosomal deletion 22q11.2, where the DGCR8 gene is located, and patients often show cognitive and behavioral impairment (25, 26); in fact, this deletion is one of the well-established risk factors for the development of schizophrenia (27, 28). The mouse model for haploinsufficiency of the Dgcr8 gene shows abnormal miRNA biogenesis caused by decreased Dgcr8 gene expression, schizophrenia-like deficits, and decreased neurogenesis in the adult hippocampus (29, 30). Of interest, this model also revealed that only a subset of pri-miRNA genes is upregulated and that a smaller subset of mature miRNAs is downregulated (29). pri-miR-9-2 is one of the most upregulated pri-miRNAs in the adult hippocampus of model mice, supporting the fact that DGCR8 more efficiently processes miR-9-2 than other miRNA species. Considering that miR-9 is a critical miRNA for the differentiation of neural precursor cells (21), the DGCR8–miR-9-2 axis is thought to be involved in neurogenic differentiation and a cause of schizophrenia.</p><p>In this study, we address the underlying mechanism of how DGCR8 efficiently processes pri-miR-9-2 to contribute to the generation of miR-9. To analyze this efficiency, we performed a cellular optimized pri-miRNA processing assay using a ratiometric fluorescence reporter, based on a similar system reported previously (31). Using this assay in combination with various types of miRNAs and pri-miR-9 mutant reporters, we provide clear evidence that pri-miR-9-2 has a novel DGCR8-responsive RNA element (DRE) that is well conserved among mammalian species and promotes DGCR8-dependent pri-miRNA processing activity.</p><p>Live-cell pri-miRNA processing reporter assay.A, schematic of fluorescent reporter vector construction. Venus and tdTomato mRNAs were transcribed under the Tet-On-responsive bidirectional promoter (PTight-BI). The 300-nt cDNA coding human pri-miR-9-1 was subcloned into a multicloning site (MCS) in the 3′-UTR of tdTomato mRNA. A microprocessor complex including Drosha and DGCR8 cleaves pri-miRNA to produce pre-miRNA from the 3′-UTR of tdTomato mRNA, which is destabilized because the poly(A) sequence is removed from the 3′-UTR. pA, poly(A) signal sequence. B, After transfecting the fluorescent reporter vector into HeLa Tet-On 3G cells, nuclear expression of Venus and tdTomato was observed by Opera Phenix, a high content cell imaging analyzer. The scale bar represents 200 μm. C, After transfecting the fluorescent reporter control, the pri-miRNA or pri-miR-9-1 reporter vector was transfected with pcDNA3.1 or the FLAG-DGCR8 expression vector into HeLa Tet-On 3G cells. The sums of the Venus fluorescent signal intensity and tdTomato fluorescent signal intensity in selected nuclei were shown in the graph. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated for each well and relative values are shown. Error bars show the standard deviation (n = 3). D, fluorescence pri-miRNA processing reporter assay. Control, pri-miR-9-1, and pri-miR-9-1M were transfected with pcDNA3.1 or FLAG-DGCR8 expression vectors into HeLa Tet-On 3G and fluorescent signals were monitored from each cell. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated in each well and shown. The error bar shows the standard deviation (n = 3). E, has-miR-9-5p was quantified by qRT-PCR with total RNA purified from HeLa Tet-On 3G cells transiently transfected with control, pri-miR-9-1, pri-miR-9-1M reporter, and FLAG-DGCR8 expression vectors. The error bar shows the standard deviation (n = 3). The asterisk indicates significant change (t-test p < 0.001); NLS, nuclear localization signal; PEST, a peptide sequence that acts as a signal peptide for protein degradation.</p><!><p>In this study, we used the reciprocal Venus:tdTomato ratio as a positive indicator of pri-miRNA processing efficiency, as described previously (31). First, we engineered 300 nucleotides (nt) of human pri-miR-9-1 into our fluorescent reporter. HeLa Tet-On 3G cells transiently transfected with control or pri-miR-9-1 fluorescent reporter showed nuclear expression of Venus and tdTomato because of the presence of an N-terminal nuclear localization signal tag. The Venus and tdTomato signal intensity was monitored, and nuclei with a Venus signal were selected (Fig. 1B). Then, Venus signals were calculated relative to the tdTomato signal for each well as an indicator of pri-miRNA processing efficiency (Fig. 1C). Ectopically overexpressing N-terminal FLAG-tagged DGCR8 (FLAG-DGCR8) promoted pri-miRNA processing activity in the fluorescence-based live-cell pri-miRNA processing reporter assay, as reported previously (31).</p><p>The C-terminal tail of DGCR8 was previously shown to be required for pri-miRNA processing (31, 32, 33), and a mutant with the C-terminal tail deleted did not promote pri-miRNA processing activity, as reported previously. We also constructed a mutant pri-miR-9-1 reporter pri-miR-9-1M, which did not respond to ectopic FLAG-DGCR8 expression because of four point mutations within the pri-miR-9-1 cleavage site (Figs. 1C and S1). The fluorescent miRNA processing assay was performed with the pri-miR-9-1M reporter vector. Ectopically expressing FLAG-DGCR8 promoted pri-miRNA processing activity in cells transfected with the pri-miR-9-1 reporter but not in those transfected with the pri-miR-9-1M reporter (Fig. 1D). Moreover, the production of miR-9-5p was decreased in cells transfected with the pri-miR-9-1M reporter (Fig. 1E). Taken together, these results indicate that our fluorescent reporter system was able to monitor pri-miR-9-1 processing activities in an ectopically expressing DGCR8-dependent manner, as reported previously.</p><!><p>pri-miR-9-2 is processed by the microprocessor complex.A, the HeLa Tet On 3G clone stably expressing the pri-miR-9-2 reporter can monitor the cellular microprocessor complex activity. Venus and tdTomato proteins were induced by doxycycline treatment in a dose-dependent manner. Negative control siRNA (siNC#1) and DGCR8 siRNA (siDGCR8#1) were transfected into the cells, and nuclear expression of Venus and tdTomato was observed by the cell imaging analyzer. B, negative control siRNAs (siNC#1 and #2) and DGCR8 siRNAs (siDGCR8#1–#3) were transfected into HeLa Tet On 3G cells with pri-miR-9-2 reporter or control reporter vectors. The amount of endogenous miR-9-5p was quantified with the specific TaqMan qRT-PCR system. miR-9-5p production was suppressed by transfecting DGCR8 siRNAs. DGCR8 proteins were knocked down by DGCR8 siRNAs. C, HeLa Tet On 3G cells stably expressing the pri-miR-9-2 reporter were treated with siNC and siDGCR8, and Venus and tdTomato expression was observed by the cell imaging analyzer (left). The scale bar represents 500 μm. The fold-change of the tdTomato fluorescence signal intensity is shown in the graph. D, hsa-miR-9-5p was quantified by qRT-PCR with total RNA purified from human cell lines Daoy, U251MG, U251MG(KO), HeLa Tet-On 3G,U2OSTteOn, and HEK293TetOn3G. The error bar shows the standard deviation (n = 3). E, negative control siRNAs (siNC#1 and #2) and DGCR8 siRNAs (siDGCR8#1, #2, and #3) were transfected into U251MGKO cells, and the amount of endogenous miR-9-5p was quantified. F and G, the amount of endogenous pri-miR-9-2 was quantified in U251MGKO cells treated with DGCR8 siRNAs (F) and Drosha siRNAs (G). H, negative control siRNAs (siNC#1 and #2), DGCR8 siRNAs (siDGCR8#1, #2, and #3), and Drosha siRNAs (siDrosha#2 and #3) were transfected into U251MGKO cells, and DGCR8 and β-actin protein expression was analyzed with the Wes protein analysis system. The asterisk indicates significant change (∗ p < 0.001, ∗∗ p < 0.01)</p><!><p>Next, we searched for human cell lines that express endogenous miR-9-5p using Daoy, U251MG, U251MG (KO), HeLa Tet-On 3G, U2OSTteOn, and HEK293TetOn3G cells. Among these, human astrocytoma U251 MG and U251 MG (KO) cells expressed high levels of miR-9-5p (Fig. 2D). Of importance, miR-9-5p expression levels in U251MG (KO) cells were suppressed by treatment with DGCR8 siRNAs (Fig. 2E), while pri-miR-9-2 was greatly increased (Fig. 2F). Hence, miR-9-5p production from endogenous pri-miR-9-2 occurred in a DGCR8-dependent manner. Similarly, Drosha siRNAs effectively knocked down Drosha cellular protein levels and promoted the accumulation of pri-miR-9-2 (Fig. 2, G and H). These findings suggest that the processing of both endogenous pri-miR-9-2 and pri-miR-9-2 fluorescent reporter depended on the canonical microprocessor complex.</p><!><p>Higher DGCR8 sensitivity of pri-miR-9-2 processing.A, fluorescent pri-miRNA processing reporter assay. Control, pri-miR-9-1, pri-miR-9-2, and pri-miR-9-3 were transfected with pcDNA3.1 or FLAG-DGCR8 expression vectors into HeLa Tet-On 3G cells, and fluorescent signals were monitored. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated and shown in the graph. The error bar shows the standard deviation (n = 3). B, has-miR-9-5p was quantified by qRT-PCR with total RNA purified from HeLa Tet-On 3G cells transiently transfected with control, pri-miR-9-1, pri-miR-9-2, pri-miR-9-3 reporter, and FLAG-DGCR8 expression vectors. The error bar shows the standard deviation (n = 3). C, control, pri-miR-9-1, and pri-miR-9-1x2 (containing twice tandem repeat of pri-miR-9-1) reporter vectors were transfected with pcDNA3.1 control or DGCR8 expression vectors into HeLa Tet-On 3G cells and fluorescent signals monitored. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated and is shown in the graph. The error bar shows the standard deviation (n = 3). D, control, pri-miR-9-2, and pri-miR-9-2x2 (containing twice tandem repeat of pri-miR-9-2) reporter vectors were transfected with pcDNA3.1 control or DGCR8 expression vectors into HeLa Tet-On 3G cells and fluorescent signals were monitored. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated and is shown in the graph. The error bar shows the standard deviation (n = 3). E and F, the sum of the Venus fluorescent signal intensity in selected nuclei was calculated and is shown in the graph. The error bar shows the standard deviation (n = 3). The asterisk indicates significant change (t-test ∗p < 0.001).</p><p>pri-miR-9-2 has a DGCR8-responsive element in the 3' wing region of pri-miR-9-2.A, schematic of deletion mutant reporters used in this study. The wing region around the stem–loop structures coding pre-miRNA was removed in the deletion mutant reporters. B, fluorescent pri-miRNA processing reporter assay. Control, pri-miR-9-2, pri-miR-9-2-200, pri-miR-9-2-100, pri-miR-9-2-200-1, and pri-miR-9-2-200-2 reporter vectors were transfected with pcDNA3.1 or FLAG-DGCR8 expression vectors into HeLa Tet-On 3G cells and fluorescent signals were monitored. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated and is shown in the graph. The error bar shows the standard deviation (n = 3). C, schematic of deletion mutant reporters used in this study. The wing region around the stem–loop structures coding pre-miRNA was removed in the deletion mutant reporters. D, fluorescent pri-miRNA processing reporter assay. Control, pri-miR-9-1, pri-miR-9-1-200, and pri-miR-9-1-100 reporter vectors were transfected with pcDNA3.1 or FLAG-DGCR8 expression vectors into HeLa Tet-On 3G cells and fluorescent signals were monitored. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated and is shown in the graph. The error bar shows the standard deviation (n = 3). E, alignment of human pri-miR-9-1 (300 nt) and pri-miR-9-2 (300 nt) to 35 other mammalian species by Genomic Evolutionary Rate Profiling on the UCSC Human Genome Browser. The red asterisk indicates the highly conserved region in the 3'-end wing of pri-miR-9-2. The asterisk indicates significant change (∗ p < 0.001 ∗∗ p < 0.005).</p><!><p>We next performed genome alignment analysis of human pri-miR-9-2 to the sequences of 35 mammalian species from the UCSC Human Genome Browser. A highly conserved region was identified in the 3′-end wing of pri-miR-9-2, but not in the wing of pri-miR-9-1 (Fig. 4E). For example, a 50-nt RNA sequence in the 3′-end wing of pri-miR-9-2 was found to be 98% identical between humans and mice, compared with only 65.4% identity for a 50 nt RNA sequence in the 3′-end wing of pri-miR-9-1. This highly conserved region supports the idea that the 3′-end wing of pri-miR-9-2 contains a DRE.</p><!><p>pri-miR-9-1 also has a DGCR8-responsive element in the region near the pre-miR-9-1.A, alignment analysis of human pri-miR-9-1-100 and human pri-miR-9-2-100 by the Clustal W method. The 5′-end (black dashed box), 3′-end (blue dashed box), and loop between miR-9-5p and miR-9-3p (red dashed box) are unique. The box residues match the consensus/majority exactly. B, schematic of chimera reporters used in this study. Unique sequences of pri-miR-9-1-100 and pri-miR-9-2-100 are shown in gray and red, respectively. pri-miR-9-1/2-102 (including pre-miR-9-1, and 5′- and 3′-ends of pri-miR-9-2-100) and pri-miR-9-1/2-98 (including pre-miR-9-2, and 5′- and 3′-ends of pri-miR-9-1-100) were constructed. C, fluorescence pri-miRNA processing reporter assay. Control, pri-miR-9-1, pri-miR-9-2, pri-miR-9-1-100, pri-miR-9-2-100, pri-miR-9-1/2-102, and pri-miR-9-1/2-98 reporter vectors were transfected with pcDNA3.1 or FLAG-DGCR8 expression vectors and fluorescent signals were monitored. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity was calculated and is shown in the graph. The error bar shows the standard deviation (n = 4). The asterisk indicates significant change (t-test p < 0.001).</p><p>DGCR8-responsive RNA elements in human pri-miR9-1 and pri-miR9-2. DGCR8-responsive RNA elements (DREs) were identified in this study. The DRE of pri-miR-9-1 is in the vicinity of pre-miR-9-1, and the DRE of pri-miR-9-2 is in the 3' wing region. DRE promotes pri-miR-9 processing activity in an ectopically expressed DGCR8-dependent manner.</p><p>Exploration of pri-miRNA candidates possessing DRE.A, miRNA profiling of U251 MG (KO) cells treated with siNC#1, siNC#2 siDGCR8#1, and siDGCR8#2 was performed with the nCounter miRNA analysis system. Expression levels of top20 ranked miRNAs are shown in the graph. The bar graph indicates the average in each two technical replicates (n = 1). B, pri-miRNA expression levels of hsa-let-7b-5p, hsa-miR-99a-5p, hsa-miR-15a-5p, and hsa-miR-100-5p in U251 MG (KO) cells treated with siNC#1, siNC#2 siDGCR8#1, and siDGCR8#2 were quantified by qRT-PCR. The error bar represents SD using four biological replicates. C, pri-miRNA expression levels of hsa-let-7b-5p, hsa-miR-99a-5p, hsa-miR-15a-5p, and hsa-miR-100-5p in HeLa Tet-On 3G cells treated with siNC#1, siNC#2 siDGCR8#1, and siDGCR8#2 were quantified by qRT-PCR. The error bar represents SD using four biological replicates. D, pri-miRNA processing reporters containing pri-miR-9-2 (300 nt), pri-let-7b (300 nt), pri-miR-99a (300 nt), pri-miR-15a-16-1 (400 nt), and pri-miR-100 (300 nt) were constructed and the processing assay was performed with HeLa Tet-On 3G cells. The error bar represents SD using three biological replicates. E, pri-miRNA processing reporters containing pri-miR-9-2 (300 nt), pri-miR-9-2M (300 nt), pri-miR-17/92 (887 nt), pri-miR-409-412-369-410 (800 nt), and pri-miR-137 (500 nt) were constructed, and the processing assay was performed with HeLa Tet-On 3G cells. The error bar represents SD using three biological replicates. Asterisks indicate significant change (∗p < 0.01).</p><!><p>Schizophrenia and 22q11.2 deletion syndrome–related miRNAs, miR-17-92 cluster pri-miRNA, miR-409-412-369-410 cluster pri-miRNA, and miR-137 were also investigated (34, 35, 36, 37). miR-17-92 cluster pri-miRNA, which contains six different pre-miRNAs, and pri-miR-137 reporters showed less efficient processing than the pri-miR-9-2 reporter (Fig. 7E). The miR-409-412-369-410 cluster pri-miRNA reporter showed less efficient processing than the pri-miR-9-2×2 reporter, even though it contained four different pre-miRNAs (Fig. 7E).</p><!><p>Role of DRE in DGCR8-dependent microprocessor activity.A, pri-miRNA processing reporters containing pri-miR-99a (300 nt) and pri-miR-99a+DRE (300 nt) were constructed, and the processing assay was performed with HeLa Tet-On 3G cells. The graph indicates plots for each cell obtained from Venus and tdTomato fluorescence signals. Slopes from linear regression are shown in each graph. The bar graph indicates slopes with or without DGCR8. The asterisk indicates significant change (t-test p = 0.026). B, RNA expression levels of METTL3 and unprocessed pri-miR-9-2 in U251 MG (KO) cells treated with siNC#1, siNC#2 siMETTL3#1, and siMETTL3#2 were quantified by qRT-PCR. The asterisk indicates significant change (t-test p < 0.01) C, Quantification of DGCR8-bound pri-miRNA was analyzed by CLIP-qRT-PCR assay. The asterisk indicates significant change (t-test p < 0.05). D, Model of DRE in DGCR8-dependent microprocessor activity. DRE, DGCR8-responsive RNA element; CLIP, UV cross-linking and immunoprecipitatio.</p><!><p>Taken together, our findings show that the pri-miR-9-2 reporter had the most efficient processing of all pri-miRNA reporters investigated in this study and that the DRE identified in pri-miR-9-2 might be a unique RNA element for potentiating DGCR8-dependent processing.</p><!><p>In the present study, we found that pri-miR-9 is the transcript most efficiently processed out of other miRNA species by the canonical microprocessor complex. We also identified novel sensitive RNA elements using a modified version of the previously described in vivo pri-miRNA processing fluorescence reporter system (31). This reporter system enabled the investigation of DGCR8- and Drosha-dependent microprocessor activity and efficiency by observation of a ratiometric fluorescence color reporter, which mimicked the cleavage process from pri-miRNA to pre-miRNA and mature miRNA.</p><p>Previous studies reported that human pri-miR-9-2 is efficiently processed in a DGCR8-dependent manner in 22q11.2 deletion syndrome model mice and Dgcr8-deficient mice (29, 34). These mice also displayed a psychiatric phenotype and showed an accrual of pri-miR-9-2, which was one of the most accumulated pri-miRNAs in the prefrontal cortex and hippocampus. Therefore, miR-9 is thought to be a strong candidate to understand the molecular etiology of schizophrenia and other neurological diseases. Indeed, miR-9 regulates neurogenesis in the mouse telencephalon by targeting downstream mRNAs, as evidenced by a study of miR-9 knockout mice (39). miR-9/9∗ are also able to directly convert adult human fibroblasts to neurons through the control of chromatin accessibility by inhibiting neuron-restricted silencer factor (40). Given the two-hit hypothesis for schizophrenia, including a combination of genetic and environmental factors (41) and quality control of neurogenic factors, miR-9 must be critically involved at multiple RNA steps.</p><p>In this study, we provide evidence that pri-miR-9-2 is processed by the canonical microprocessor complex including DGCR8 and Drosha, that the pri-miR-9-2 processing efficiency is relatively high, and that a mediating DRE sequence is present in the 3′-end wing of pri-miR-9-2. Of interest, this DRE is highly conserved among mammals, supporting the idea that it is a critical quality control element. We also found other DRE sequences near both ends of pre-miR-9-1 that are important for DGCR8 sensitivity. Of importance, conservation of these RNA sequences between humans and mice is 100% identical, while the comparable region in pri-miR-9-2-100 is not (Fig. 6). Recently, an RNA modification with N6-methyladenosine functions was reported as a marker that is efficiently recognized by a microprocessor complex (38, 42). It is possible that the DREs in the present study have the potential to be similarly methylated. In fact, a recent m6A HIT-CLIP result derived from mouse brain shows two independent m6A methylation sites in the 3' wing region on pri-miR-9-2 (43) (Fig. S4). Given the high interspecies conservation of this 3' wing region, this DRE might be deeply involved with DGCR8 sensitivity.</p><p>Among the pri-miRNA candidates we explored for possessing DRE to potentiate pri-miRNA processing, the pri-miR-9-2 reporter showed the most efficient processing. This indicated that the DRE RNA element of pri-miR-9-2 may be unique for potentiating DGCR8-dependent processing and that it might be required for pursuing the link to schizophrenia in the future.</p><!><p>Vector construction was performed by GENEWIZ. Synthesized FLAG-human-DGCR8 and FLAG-tagged DGCR8 mutant with the C-terminal tail (701–773 aa) deleted were subcloned into the HindIII/NotI site of pcDNA3.1(+) vectors (GenBank:CAK54796.1). pri-miRNA processing reporters were engineered based on the bidirectional tetracycline-inducible vector, pTRE-Tight-BI. Synthesized Venus cDNA with an N-terminal nuclear localization signal and C-terminal PEST cDNA were subcloned into the EcoR I/Xba I site in MSC-II of the pTRE-Tight-BI vector, and tdTomato cDNA with an N-terminal nuclear localization signal, C-terminal PEST, and pri-miRNA cDNA were subcloned into the Xma I/Cla I site in MCS I of the pTRE-Tight-BI vector. The detailed sequences are shown in Table S1.</p><!><p>HeLa Tet-On 3G, U2OS Tet-On, and HEK293 Tet-On 3G cell lines were purchased from Takara-Clontech. Daoy, U251 MG, and U251 MG (KO) cell lines were purchased from JCRB Cell Bank. HeLa Tet-On 3G cells were cultured in FluoroBrite Dulbecco's modified Eagle's medium (Life Technologies) with 10% fetal bovine serum (Life Technologies) in 5% CO2 at 37 °C for maintenance. For the fluorescence reporter assay, HeLa Tet-On 3G cells were cultured with a 10% Tet-system–approved fetal bovine serum (Clontech) in 5% CO2 at 37 °C. Cells were transfected with Lipofectamine3000 (Life Technologies) and Opti-MEMI (ThermoFisher Scientific) following the manufacturer's instructions. For transfection of the reporter and/or expression vectors, 1× 104 cells were seeded in CellCarrier-96 plates (PerkinElmer) or 1.6 × 105 cells were seeded in 6-well plates (Corning). Three hours after transfection, the culture medium containing transfection reagents was changed to fresh medium containing 10 ng/ml doxycycline. Nonconfocal images of the cells were obtained 18 to 24 h later by the Opera Phenix high-content screening system with a 10× Air, NA0.3 objective lens (PerkinElmer).</p><!><p>Venus and tdTomato signal intensities were obtained for 20 and 5 ms, respectively. Images were obtained from eight fields of view for each well, and nuclei with Venus signals were selected by the NEW Opera Phenix HCS System (PerkinElmer). The sum of the Venus and tdTomato signals in each well was calculated from the Venus and tdTomato signal intensities of each cell in the well. The relative sum of the Venus signal intensity to the sum of the tdTomato signal intensity in each well was calculated as the pri-miRNA processing activity and relative values are shown in the graph. Alternatively, the cell images were obtained by KEYENCE BZ-X810. Quantification of Venus and tdTomato fluorescence signal from single cells were automatically quantified using the Hybrid Cell Count Module BZ-H4C (KEYENCE). The slope of the Venus and tdTomato signals as the pri-miRNA processing activity is shown in the graph.</p><!><p>For siRNA transfection, HeLa Tet-On 3G cells (Takara Bio) or U251MGKO cells (JCRB Cell Bank) were seeded in CellCarrier-96 plates (PerkinElmer) or 6-well plates (Corning). siRNAs were transfected with Lipofectamine RNAiMAX Transfection Reagent (ThermoFisher Scientific) in Opti-MEMI following the manufacturer's instructions. siRNAs purchased from Invitrogen in this study are listed below.</p><p>siNC#1 as a negative control siRNA in human, mouse, and rat cells;</p><p>siNC#2 as a negative control siRNA in human, mouse, and rat cells;</p><p>siDGCR8#1, a human DGCR8 siRNA</p><p>sense sequence (5′–3′), CCCUGUCUAUAAUUUCUUUtt</p><p>antisense sequence (5′–3′), AAAGAAAUUAUAGACAGGGcg;</p><p>siDGCR8#2, a human DGCR8 siRNA</p><p>sense sequence (5′–3′), GGAUCAUGACAUUCCAUAAtt</p><p>antisense sequence (5′–3′), UUAUGGAAUGUCAUGAUCCac;</p><p>siDGCR8#3, a human DGCR8 siRNA</p><p>sense sequence (5′–3′), GGUUCACGGCUAAAGCAAUtt</p><p>antisense sequence (5′–3′), AUUGCUUUAGCCGUGAACCcg;</p><p>siDrosha#1, a human Drosha/RNASEN siRNA</p><p>sense sequence (5′–3′), GCUCUGUCCGUAUCGAUCAtt</p><p>antisense sequence (5′–3′),</p><p>UGAUCGAUACGGACAGAGCtt;</p><p>siDrosha#2, a human Drosha/RNASEN siRNA</p><p>sense sequence (5′–3′), GACCAGACUUUGUACCCUUtt</p><p>antisense sequence (5′–3′), AAGGGUACAAAGUCUGGUCgt;</p><p>siDrosha#3, a human Drosha/RNASEN siRNA</p><p>sense sequence (5′–3′), CACUUAACUUUGUUGCGAAtt</p><p>antisense sequence (5′–3′), UUCGCAACAAAGUUAAGUGtc</p><p>siMettl3#1, a human Drosha/RNASEN siRNA</p><p>sense sequence (5′–3′), GAUCCUGAGUUAGAGAAGAtt</p><p>antisense sequence (5′–3′),</p><p>UCUUCUCUAACUCAGGAUtg;</p><p>siMettl3#2, a human Drosha/RNASEN siRNA</p><p>sense sequence (5′–3′),</p><p>GCAGUUCCUGAAUUAGCUAtt</p><p>antisense sequence (5′–3′),</p><p>UAGCUAAUUCAGGAACUGCtg;</p><!><p>Total RNAs were extracted using the miRNeasy Mini kit (Qiagen) and quantified by Qubit with the Qubit RNA HS Assay Kit (Life Technologies). Quantification of hsa-miR-9-5p and hsa-miR-9-3p were performed with a TaqMan microRNA RT Kit (Life Technologies, TM/RT000583 and TM/RM002231, respectively), TaqMan microRNA Assays (Life Technologies), and TaqMan Universal PCR Master Mix II with UNG (Life Technologies) following the manufacturer's instructions. Five nanograms of total RNA was used in each reverse transcription reaction. To quantify human DGCR8 mRNA, human Drosha mRNA, and human β-actin mRNA expression levels, real-time PCR was performed using the TaqMan Gene Expression assay (Life Technologies, DGCR8; Hs00256062_m1 and Hs00987085_m1, Drosha; Hs00203008_m1, β-actin; Hs01060665_g1, pri-let-7b; Hs03302548-pri, pri-miR-100; Hs03302731-pri, pri-miR-15a; Hs03302582-pri, pri-miR-9-1; Hs03303201_pri, pri-miR-9-2; Hs03303202_pri, pri-miR-9-3; Hs03293595_pri, pri-miR-99a; Hs03302729-pri), TaqMan Gene Expression Master Mix (Life Technologies), and ViiA7 Real Time PCR or StepOnePlus system (Life Technologies) following the manufacturer's instructions. Quantitative RT-PCR analysis was used with at least three biological replicates.</p><!><p>Samples prepared for western blotting underwent simple western analysis with the Wes system and 12–230 kDa Wes Separation Module (Protein Simple) using rabbit polyclonal anti-DGCR8 (PGI Proteintech Group), rabbit polyclonal anti-Drosha (Bethyl Laboratories), and anti-β-actin (Novus Biologicals) antibodies. Data analysis was performed with Compass software (Protein Simple).</p><!><p>miRNA profiling analysis of total RNA was performed with an nCounter Analysis System and a Human miRNA Assay Kit Version 3.0 (NanoString Technologies) using the FOV max mode. Data analyses were performed with nSolver analysis software version 2.0 equipped with the nCounter Analysis System. Normalization of miRNA profiling data was performed with housekeeping gene expression levels. This experiment was carried out in two technical replicates from one each sample.</p><!><p>The Genomic Evolutionary Rate Profiling score defined the reduction in the number of substitutions among the multispecies sequence alignment, using 35 other mammalian species to human genome (44,45). All these data analyses were done using the publicly available tool, the UCSC Genome Browser (https://genome.ucsc.edu).</p><!><p>The CLIP-RT-qPCR assay was performed as described (46).U251 cells were UV cross-linked at 254 nm (UV-B) with 200 mJ/cm2 three times. Lysates were subjected to immunoprecipitation and qRT-PCR (47). CLIP-qRT-PCR enrichments were normalized by quantifying relative levels of immunoprecipitated pri-miRNA to input from lysates. The detailed primer sequences are shown in Table S2.</p><!><p>All experiments were carried out using at least three biological replicates. Statistically significant differences were calculated by two-tailed Student's t-test and presented as the mean and standard deviation.</p><!><p>All data are contained within this article and supplemental information.</p><!><p>This article contains supporting information.</p><!><p>H. O. is a paid member of the Scientific Advisory Board of San Bio Co, Ltd and K Pharma, Inc M. Y. is a scientific advisor of K Pharma, Inc.</p><!><p>Figures S1–S4Tables S1 and S2</p>

PubMed Open Access

Error quantification of phase transition quantities from cluster weighting calculations

In this work, we investigate how uncertainties in experimental input data influence the results of quantum cluster equilibrium calculations. In particular, we focus on the calculation of vaporization enthalpies and entropies of seven organic liquids, compare two computational approaches for their calculation and investigate how these properties are affected by changes in the experimental input data. It is observed that the vaporization enthalpies and entropies show a smooth dependence on changes in the reference density and boiling point. The reference density is found to have only a small influence of the vaporization thermodynamics, whereas the boiling point has a large influence on the vaporization enthalpy but only a small influence on the vaporization entropy. Furthermore we employed the Gauss-Hermite estimator in order to quantify the error in the thermodynamic functions that stems from uncertainties in the experimental reference data at the example of the vaporization enthalpy of (R)-butan-2-ol.We quantify the error as 30.95 •10 −3 kJ mol −1 . Additionally we compare the convergence behaviour and computational effort of the Gauss-Hermite estimator with the Monte Carlo approach and show the superiority of the former. By this, we present how uncertainty quantification can be applied to examples from theoretical chemistry.

error_quantification_of_phase_transition_quantities_from_cluster_weighting_calculations

Introduction<!>Quantum cluster equilibrium<!>Error Quantification<!>Computational Details<!>Systems Investigated<!>Results and Discussion<!>Varying the input quantities density and boiling point: Enthalpy<!>Error Quantification

<p>Fast but approximate methods are needed in theoretical chemistry in order to reduce the computational costs while still providing results at acceptable accuracy for the complex and large chemical space, 1,2 i.e. Avogadro's number. Since these approximate methods allow the treatment of larger and thus more realistic systems with reasonable computational effort, 3 they are a necessary step in terms of sustainable research as they can also help to save expensive computer time. Not only will larger and much more complex systems be available to computations, but the reduced computational effort will allow the usage of smaller work stations using less resources. In this respect, many pioneering works were undertaken in quantum chemistry or electronic structure theory. Examples are developments and applications in the context of density functional theory, including low-cost but well performing functionals and composite methods, 4 but also wave function based theories like for instance local coupled cluster methods have made major contributions to this field of research. 5 Similarly, the further development of semiempirical tight-binding methods 6 or socalled embedding methods 7 were large steps forward, in allowing to calculate more realistic problems. From the side of simulations, coarse grained or enhanced sampling techniques, 8,9 but also other linear scaling methods 10 should be mentioned here.</p><p>In spite of these efforts, realistic applications from sustainable chemistry often require complex or involved models of the underlying chemistry and large computing time. Examples are the inclusion of liquids or solvents, for example ionic liquids, [11][12][13][14] but also the appropriate description of interfaces 15,16 or systems from energy devices. [17][18][19] When using such approximate methods, the assessment of the errors included is of high importance. In this article, we use the quantum cluster equilibrium (QCE) method, 20,21 in order to explore the nature of liquids and calculate their properties. The QCE method treats clusters quantum chemically in the mean-field of a liquid and calculates the cluster populations at a given phase point. If the same level of electronic structure is used, QCE provides an approximate method compared to for example ab initio molecular dynamics simulations as it samples a dynamic cluster ensemble.</p><p>As opposed to classical Boltzmann factors which are limited to same sized structures at a time, the QCE approach allows the simultaneous inclusion of differently sized clusters. 22 While in its original form the theory has been developed for neat substances, 20,21,[23][24][25] it has been further extended to binary systems, [26][27][28] enabling the calculation of activity coefficients 29 or concentration dependent acid dissociation. 30,31 Within the QCE method, molecular clusters including the monomers and several oligomers are considered to be in chemical equilibrium. Interactions between the clusters are described by the mean-field of the cluster ensemble. Nonetheless, within the clusters themselves, all interactions are treated explicitly. This allows the application of a variety of quantum chemical methods, 23 enabling the treatment of the condensed phase with low-cost methods 32 but also on the level of highly accurate post Hartree-Fock electronic structure methods such as coupled cluster. 33 For a given ensemble of clusters, the connection between the thermodynamic quantities of the condensed phase and the set of structures is given by the system partition function, for which an approximate but analytical expression is available. Based on this expression, any thermodynamic function of the cluster ensemble is accessible. While the treatment of the liquid phase as thermodynamic equilibrium of clusters can circumvent sampling problems and thus the computational bottleneck of e.g. ab initio molecular dynamics simulations, the combination of the explicit cluster approach with the mean-field keeps the computational effort low. QCE is thus a promising method for calculating properties of liquids and their mixtures at low-cost. Nevertheless, a quantification of the errors due to the approximations made is desirable, especially because QCE computations require experimental input data which are prone for uncertainties.</p><p>Therefore, the focus of this article is to gain a basic understanding of the influence of measurement errors in physical quantities on the results of QCE calculations. To this end, we follow uncertainty quantification approaches of various scientific fields 34 and assume the measurement error as random. As such, the physical quantities affected by these measurement errors and entering the QCE calculations are random as well and are considered as random parameters. This propagates through the QCE calculations such that the QCE results themselves need to be considered as a random quantity. The aim of uncertainty quantification is then to quantify the statistical behaviour of the calculation results, given some information on the statistical behaviour of the measurement errors.</p><p>The most straightforward approach to achieve this aim is the Monte Carlo (MC) method. 35 Drawing random values from the probability distributions of the input parameters, the Monte Carlo method performs a QCE calculation for each of these random parameter sets to compute statistical quantities of interest of the outputs, such as the empirical mean and the empirical standard deviation, for example. Unfortunately, Monte Carlo simulations often require many samples and thus model evaluations to yield sufficiently accurate results, 36 such that they are computationally quite expensive, even when the underlying model can be evaluated at relatively low cost.</p><p>Given a certain amount of a-priori knowledge on how smooth the outcome of the calculations changes with variation of the input parameters, various deterministic methods and acceleration techniques can be applied to speed up computations in uncertainty quantification. Most of them are based on the mathematical representations of mean, variance, etc. as integrals in one or several variables over the range of possible parameter values, 37,38 which are then approximated by a combination of one-dimensional numerical quadrature rules. 37,39 This usually works well for a small number of parameters of significantly less than ten. For a larger number of parameters the curse of dimensionality makes this approach computationally prohibitive and special quadrature rules need to be employed for acceleration, such as quasi Monte Carlo methods, 35,40 sparse grids, 41 or other multi-level techniques. 42,43 The article is structured as follows. First we present two approaches for the calculation of vaporization enthalpies and entropies. Both approaches are based on QCE computations and we apply them to seven organic liquids in order to compare the quality of the results and the computational effort. Afterwards we investigate the dependence of the calculated vaporization enthalpies and entropies as well as of the cluster populations with respect to variations in the experimental reference data that need to be provided for QCE calculations.</p><p>As a last step we use an uncertainty quantification approach in order to asses the error in the vaporization enthalpy that originates from uncertainties in the experimental input data.</p><p>We compare the computational effort and convergence behaviour of the used method with the well known Monte Carlo method and finally provide a quantification of the error that is inherent in the vaporization enthalpy.</p><!><p>In the following, the theory behind the QCE method will be explained in greater detail. For a shortened presentation, the reader is referred to Ref. 22,32,44 The familiar reader may skip this section. The QCE method 20,23 describes the liquid and the gaseous phase of a system in terms of quantum chemically optimized clusters which are in thermodynamic equilibrium with each other. Clusters are an accumulation of multiple molecules (monomers), which arrange in optimal conformations and which are held together by attractive interactions, see also Fig. 1. The thermodynamic equilibrium of a set of clusters can be described as:</p><p>where i C (℘) describes the number of monomers of component C present in a cluster ℘.</p><p>The aim of the QCE method is to find an equilibrium distribution {N ℘ } of the cluster populations N ℘ , for which the free energy A at a given volume V and temperature T becomes minimal. Starting point is the canonical partition function:</p><p>where q tot ℘ denotes the partition function of a cluster ℘. The cluster partition function is the product of the partition functions of the different degrees of freedom:</p><p>The translational, rotational and vibrational partition functions are calculated by standard equations from textbooks that are derived from simple models like the particle in a box, the rigid rotor or the harmonic oscillator: 37,45</p><p>Therein, Λ is the thermal de Broglie-wavelength, h the Planck constant, m the mass of the particle, k B the Boltzmann constant, σ the symmetry-number, Θ rot j and Θ vib j the characteristic rotational and vibrational temperatures, I j the moments of inertia and ν i the frequencies of the different vibrational modes. The variable x in the sum of the vibrational partition function is equal to 5 for linear molecules and 6 for non-linear molecules. A detailed description can be found in physicochemical textbooks. 37 With the electronic ground state energy elec 0 and the degeneracy of the ground state g 0 , the electronic partition function can be expressed by the following equation:</p><p>All input data for those partition functions can be obtained from the quantum chemical calculations of the clusters.</p><p>The translational partition function as presented above is problematic, since it describes an ideal gas where the molecules do not have a volume themselves. To counteract this, a correction is introduced in the spirit of the van der Waals equation. The new parameter b xv scales the volume which is not available for translation since it is occupied by the clusters.</p><p>This exclusion volume V ex is given as:</p><p>with β xv being the exclusion volume expansion coefficient and b 0 xv is the base of the intercept. 46 The resulting translational partition function is now given as follows:</p><p>The second parameter a mf is introduced in order to scale the mean-field binding energies of the clusters, which is expressed by a density dependent term:</p><p>Together with the adiabatic binding energy:</p><p>the electronic partition function changes to:</p><p>The two parameters introduced in the QCE theory, namely b xv and a mf , are optimized with respect to experimental inputs. In particular, the difference between the calculated results and the reference data is minimized.</p><p>Knowledge of the three independent variables ({N ℘ }, V, T ) in equilibrium allows to obtain thermodynamic data from the canonical partition function solely. A condition is introduced to ensure the conservation of the mass/particle number:</p><p>where I ℘ is the monomer-normalized population of a cluster ℘. As long as the system is in equilibrium and all thermodynamic variables stay constant, the slope of the free energy A in dependence of the cluster populations is zero:</p><p>Here dλ describes the reaction progress. Since this equation is independent of the reaction progress and has to hold for every cluster, it can be rearranged as follows:</p><p>The mathematical connection of the free energy A and the canonical partition function is given by:</p><p>which in combination with equations 2 and 15 leads to:</p><p>By using the Stirling equation and rearranging, one obtains an equation for N ℘ :</p><p>Inserting this equation in Eq. 14 gives the generalized population-polynomial which is one of the central equations of the QCE theory:</p><p>For neat liquids this equation is exactly solvable.</p><p>The second central equation is the so called volume-polynomial. In order to derive this, the pressure of the system is defined as in statistical thermodynamics as the negative volume derivative of the free energy at constant temperature:</p><p>By inserting the expression of the free energy A in terms of the canonical partition function one obtains:</p><p>Since the vibrational and the rotational partition functions do not depend on the volume, only the translational and electronic parts of the partition function have to be taken into account here. By inserting equations 2, 3, 8, 10, 11 and 13, the volume-polynomial is obtained:</p><p>While solving the population-and the volume-polynomial, the volume, the cluster populations and the partition function are independent, meaning those equations have to be solved iteratively. If multiple combinations of the volume and the populations exist, which solve both polynomials, the solution with the lowest Gibbs energy G is taken:</p><p>After the best solution is determined, the absolute enthalpy at a given temperature can be obtained by:</p><p>In order to calculate enthalpies of vaporization ∆ vap H, absolute gas phase enthalpies are needed. This is done by performing another QCE calculation, called QCE 0 , where the parameter b xv is artificially set to 1 and the parameter a mf is set to 0, which enforces a nearly ideal gas phase behaviour. The absolute enthalpies of the two calculations can then be subtracted, yielding an enthalpy required to vaporize the liquid at a given temperature.</p><p>The same procedure can be applied for the calculation of vaporization entropies ∆ vap S, with the absolute entropy given as</p><p>Lastly, low vibrational frequencies can (optionally) be treated as hindered rotations since the vibrational partition function diverges for ν i −→ 0. 47 Within this modified rigid-rotorharmonic-oscillator (mRRHO) approach, a modified vibrational partition function is introduced as:</p><p>where q vib HO is the vibrational partition function derived from the harmonic oscillator discussed above (equation 6) and q vib HR is the vibrational partition function of the hindered rotor:</p><p>where Ī is the average moment of the molecule, µ is the moment corresponding to the normal mode and is the reduced Planck constant. As switching function f (ν i ) between the two partition functions, the Chai-Head-Gordon damping function 48 is used:</p><p>with ν 0 being the user defined rotor-cutoff, which usually is assigned a value of 50 cm −1 or 100 cm −1 .</p><!><p>The QCE computations depend on physical input quantities, for example one input density ρ in and the boiling point T b , whose exact values are unknown and cannot be determined due to unavoidable measurement errors. As these measurement errors can be assumed to be random and normally distributed it is reasonable to model ρ in and T b as normally distributed random variables. To that end, experimental data from the literature are considered as the means µ ρ and µ T of ρ in and T b , and the estimated uncertainties are considered as the standard deviations σ ρ and σ T for the distribution. The probability of ρ in and T b having a certain value s is then given by the probability distributions</p><p>An immediate consequence of considering ρ in and T b as uncertain random variables is that all quantities calculated with the QCE method must now also be regarded as random variables, and are thus subject to an a priori unknown uncertainty which needs to be quantified.</p><p>For this purpose, the mean of the vaporization enthalpy is given by</p><p>and its standard deviation by</p><p>Since the integral eq. ( 31) cannot be analytically determined, it must be numerically approximated. To that end, it should be noted that most deterministic numerical approximation methods for integrals require some kind of "smooth" dependence of the integrand on the integration variable for reliable results. 37,39 However, the behaviour of ∆ vap H(ρ in , T b ) in dependence of ρ in and T b in QCE is quite complicated and, to the best of our knowledge, not yet analyzed. A robust alternative to deterministic methods which does not assume a smooth dependence of the integrand is the Monte Carlo method: 35 given N random sample densities ρ</p><p>in and boiling temperatures T (i) b , i = 1, . . . , N , drawn from the probability distributions in eq. ( 30), the Monte Carlo estimator reads as</p><p>The estimator is an established tool in computational uncertainty quantification if the smoothness behaviour of the integrand is unknown and the standard benchmark for other methods.</p><p>Despite its popularity the Monte Carlo estimator has also a few disadvantages, the obvious one being that its value strongly depends on the randomly drawn samples for ρ in and T b .</p><p>Approximation estimates for the estimator can therefore only be made in the statistical sense that the root mean square error of all possible Monte Carlo simulations will converge to the exact value of the integral for an increasing number of samples. Moreover, the error behaves like O(N −1/2 ) with the number of samples, 35 resulting in roughly O(10 6 ) samples to reach an accuracy of O(10 −3 ) for example. Since each evaluation of ∆ vap H ρ</p><p>requires a QCE computation it should be clear that the Monte Carlo method is computationally quite intensive and time consuming.</p><p>To investigate the potential for more efficient deterministic approximation methods we therefore performed a numerical study on the behaviour of ∆ vap H(ρ in , T b ) in dependence of ρ in and T b . Its detailed results are discussed below but they suggest that ∆ vap H(ρ in , T b ) presumably changes indeed smoothly with varying ρ in and T b , which paves the way for more efficient approximation methods for eq. ( 31). Thus, in a second step, we replaced the Monte Carlo estimator by a Gauss-Hermite quadrature, which is a special case of Gauss quadrature on the real line. That is, for a smooth function f depending on a single variable, a number of quadrature nodes N , and generic mean µ and variance σ 2 , we approximate the integral</p><p>where (ω i , x i ), i = 1, . . . , N , are the quadrature weights and nodes, see also Eq. (25.4.46) and Table 25.10 in Abramowitz & Stegun 38 for their specific values. The deterministic estimator for eq. ( 31) then reads</p><p>Given the presumed sufficient smooth dependence of ∆ vap H on ρ in and T b the error of the Gauss-Hermite quadrature decays exponentially, i.e., like O(e −cN ) for some c > 0, with the number of quadrature nodes N . Thus, the approximation of the integral eq. ( 31) can be achieved with only a few QCE calculations on a standard laptop.</p><!><p>The quantum chemical calculations dealing with the apolar liquids have been carried out in a previous work which provides more details about the computational protocol. 32 The clusters were build with the Ogolem program package 49,50 in combination with the generalized Amber force field. 51 The clusters contain up to ten monomers and several distinct minima of the potential energy surface were selected for a respective cluster size. The geometry optimizations as well as frequency calculations were carried out with GFN2-xTB 6 using the xtb 6.2.1 program package. 52 If not specified otherwise, all QCE computations were carried out at a fixed pressure of 101.325 kPa employing Peacemaker 2, our publicly available software package for neat and binary QCE calculations. 23,26,44 The cluster volumes were assumed as van der Waals volumes with radii from Bondi's compilation. 53 The empirical parameters a mf and b xv were optimized in order to reproduce the boiling point and density at the given temperature as accurately as possible. Throughout this study, b xv is treated as being temperature independent. In the QCE 0 -calculations, the empirical parameters were set to a mf = 0.0 and b xv = 1.0, which results in the neglection of inter-cluster interactions and the unscaled van der Waals volume as the exclusion volume. For all calculations the mRRHO correction with a rotor cutoff value of 100 cm −1 was used.</p><!><p>Within this study, we investigate seven organic substances, namely ethane, ethylene, propane, propylene, butane, methanol and the chiral (R)-butan-2-ol. Ball-and-stick images of the respective molecules as well as some exemplary clusters of ethane and (R)-butan-2-ol are shown in Fig. 1. The first five substances represent non-polar and aprotic hydrocarbons whose intermolecular interactions are mainly dominated by London dispersion forces. In contrast to that, methanol and (R)-butan-2-ol are polar alcohols which are able to form hydrogen bonds, see dotted lines in Fig. 1. Table 1 summarizes the experimental boiling points for all seven substances as well as the experimental densities at the respective reference temperatures.</p><!><p>Calculating Thermodynamic Functions of Vaporization</p><p>As already shown in previous works, 32,56 the vaporization enthalpy of neat liquids at a specific temperature can be calculated by taking the difference between the absolute QCE 0 -enthalpy and the absolute enthalpy obtained from a QCE calculation in which the empirical parameters were optimized, at the respective temperature. Similarly, the vaporization entropy can be obtained by subtracting the absolute QCE entropy from the absolute QCE 0 -entropy.</p><p>Throughout this work we will label this procedure for the calculation of thermodynamic vaporization functions as standard QCE approach.</p><p>However, another approach for the calculation of vaporization enthalpies makes use of the well-known Clausius-Clapeyron relation 37</p><p>which gives the slope of the tangents (dp/dT ) on a coexistence curve in a pressuretemperature (P-T) diagram. Here, ∆H is the enthalpy change and ∆V is the specific volume change for the phase transition. If only transitions from a condensed phase to the gas phase are considered and for temperatures far below the critical temperature, the specific volume change corresponds approximately to the gas phase volume. By using the ideal gas law, the relation shown above can be rewritten to obtain the Clausius-Clapeyron equation</p><p>where R is the ideal gas constant. In practice, ∆ vap H is a function of the temperature, however, for most systems it only varies slightly with the temperature and is thus considered as constant. Applying this approximation allows the integration of the Clausius-Clapeyron equation, which yields:</p><p>Using Eq. 38, the enthalpy of vaporization can be calculated by plotting the natural logarithm of the pressure against the inverse of the boiling temperature (or in case of a solid-gas transition the sublimation temperature) under the respective pressure.</p><p>In order to obtain several data points on the liquid-gas coexistence curve, the boiling points of the respective substance have to be determined at different pressures. Using Peacemaker, 23,26,44 this can be done by first optimizing the two QCE parameters a mf and b xv at a pressure of one atmosphere based on the experimental boiling point and an experimental density. After that, QCE calculations can be performed at different pressures (in this case ranging form 20 kPa to 500 kPa in increments of 20 kPa) by using the optimized parameters a mf and b xv as fixed input. Based on this procedure, the boiling point at a respective pressure corresponds to the temperature point at which the thermodynamic functions of the system experience the steepest slope. In order to identify this steepest slope, many thermodynamic functions can be used, for example the absolute enthalpy H, the absolute entropy S, the molar Volume V or the heat capacity c v . However, note that this artificial phase transition is not always recognizable and does not necessarily correspond to the actual boiling point at the respective pressure. Throughout this work we will label this procedure for the calculation of thermodynamic vaporization functions as Clausius-Clapeyron approach.</p><p>In order to assess the quality of the standard QCE and the Clausius-Clapeyron approach, we calculated vaporization enthalpies for the seven investigated organic liquids. The results are summarized in Table 2 together with the experimental references. The absolute and relative deviations of the calculated enthalpies with respect to the experimental references are summarized in Table S1.</p><p>For the five aprotic systems, the vaporization enthalpies based on the standard QCE approach were already calculated in a previous study. 32 The obtained vaporization enthalpies of these systems were found to be in good agreement with the experimental references, showing approach. Even though this results in a somewhat worse agreement with the experimental references, the largest difference between both approaches is observed for butane, with a deviation of 0.29 kJ mol −1 (< 2 %) between the two calculated values, thus being almost negligible. Consequently, the deviations of the vaporization enthalpies obtained with the Clausius-Clapeyron approach with respect to the experiment are similar compared to the standard QCE approach, namely between 0.67 (4.9 %) and 5.50 kJ mol −1 (24.5 %).</p><p>Besides the investigation of the five aprotic systems, both approaches were applied to two protic substances as well, namely methanol and (R)-butan-2-ol, the last one being simply termed as butanol. In the case of butanol, both approaches result in theoretical vaporization enthalpies which differ from each other by only 0.96 kJ mol −1 and agree almost perfectly with the experiment. The deviation from the experimental reference equals to 0.48 kJ mol −1 respectively, with the standard QCE approach underestimating the experimental value and the Clausius-Clapeyron approach overestimating it. By this, butanol is the only system investigated for which the Clausius-Clapeyron approach yields a larger vaporization enthalpy than the standard QCE approach. For methanol, the standard QCE approach yields a vaporization enthalpy of 43.13 kJ mol −1 while the Clausius-Clapeyron approach results in a smaller value of only 41.60 kJ mol −1 . This system shows the largest difference between both theoretical methods, namely 1.53 kJ mol −1 . The deviations to the experimental reference are 5.73 kJ mol −1 (15.3 %) for the standard QCE approach and 4.20 kJ mol −1 (11.2 %)</p><p>for the Clausius-Clapeyron approach. Methanol is therefore the only system investigated for which both approaches overestimate the experimental vaporization enthalpy. Such an overestimation of the vaporization enthalpy has already been observed before 46 and might be due to the reason that methanol molecules aggregate to small clusters in the gas phase 63 which is not properly sampled by the QCE 0 -calculations.</p><p>To summarize, the deviations from the experimental vaporization enthalpies are within a range of 0.44 to 5.73 kJ mol −1 for the standard QCE approach and within a range of 0.48 to 5.50 kJ mol −1 for the Clausius-Clapeyron approach. The corresponding mean absolute deviations (MADs) equal to 2.86 kJ mol −1 (12.57 %) for the standard QCE approach and 2.79 kJ mol −1 (12.84 %) for the Clausius-Clapeyron approach, which is within chemical accuracy for the calculation of thermodynamic properties and thus an acceptable error. Both approaches can thus be considered to perform equally well for the calculation of vaporization enthalpies.</p><p>Similar to the total enthalpy, the total entropy of a system can be calculated using QCE. Based on that, the vaporization entropy can be obtained in the same manner as the vaporization enthalpy, namely by the difference between the total liquid phase entropy and the total gas phase entropy at a given temperature. 32 Yet another possibility to obtain the vaporization entropy of a specific system is its calculation based on the vaporization enthalpy at the boiling point. By exploiting the Gibbs-Helmholtz equation</p><p>and the fact that ∆G is equal to zero for a phase transition at the boiling point, the vaporization entropy can be expressed as</p><p>Here, ∆G, ∆H and ∆S are the changes in Gibbs energy, enthalpy and entropy, respectively. Consequently the enthalpy and entropy changes at the boiling point correspond to the vaporization enthalpy ∆ vap H and vaporization entropy ∆ vap S. Assuming that the vaporization enthalpy remains constant during temperature changes, vaporization entropies at the boiling point can be calculated based on Eq. 40 as the ratio of the vaporization enthalpy obtained by the Clausius-Clapeyron approach and the boiling temperature. For this reason, we calculated the vaporization entropies solely at the boiling point. According to the Trouton's rule, the vaporization entropy equals to ≈ 88 J mol −1 K −1 for various kinds of liquids at their boiling points. 64,65 This assumption is valid for many liquids like e.g. propylene or butane, however, there are also exceptions, mostly among polar molecules like methanol or (R)-butan-2-ol.</p><p>We employed the standard QCE as well as the Clausius-Clapeyron approach (Eq. 40) to calculate the vaporization entropies at the boiling points for all seven liquids. The results as well as the corresponding experimental references are summarized in Table 3. The absolute and relative deviations of the calculated entropies with respect to the experimental references are summarized in Table S2. Note that for the aprotic substances, the corresponding values obtained with the standard QCE approach were already calculated in a previous study. 32 Comparing the experimentally observed vaporization entropies of the five aprotic liquids with the value of ≈ 88 J mol −1 K −1 predicted by Trouton's rule, a generally good agreement can be observed. Although the experimental references are about 6 to 8 J mol −1 K −1 smaller, the application of Trouton's rule seems to be justified here. Considering the vaporization entropies calculated by the standard QCE and the Clausius-Clapeyron approach, both approaches consistently underestimate the experimental references of the five aprotic compounds. Thereby, the Clausius-Clapeyron approach results in vaporization entropies which are 0.47 to 1.17 J mol −1 K −1 (< 2 %) smaller than the ones obtained with the standard QCE approach. The vaporization entropies obtained with the standard QCE approach differ from the experimental references by 1.97 (2.5 %) to 19.02 J mol −1 K −1 (23.2 %).</p><p>When considering the Clausius-Clapeyron approach on the other hand, the deviations of the obtained vaporization entropies lie within a range of 3.14 (3.9 %) to 19.91 J mol −1 K −1 (24.5 %). Thus, it can be concluded that both approaches perform comparably well for the calculation of vaporization entropies in the case of aprotic substances.</p><p>The vaporization entropies of the protic liquids are overestimated by both approaches.</p><p>The experimental reference of butanol is overestimated by 29.92 (27.4 %) and 25.54 J mol −1 K −1 (23.4 %) when considering the standard QCE and Clausius-Clapeyron approach, respectively. The reference of methanol is overestimated by 20.54 (19.7 %, standard QCE) and 19.20 J mol −1 K −1 (18.4 %, Clausius-Clapeyron). As already mentioned in the last section, this is most likely due to the fact that the gas phase of the polar alcohols is not described properly by a QCE 0 calculation. 46 In summary, when considering the standard QCE approach, the deviations from the experimental vaporization entropies are within a range of 1.97 (2.5 %) to 29.92 J mol −1 K −1 (27.4 %). Considering the Clausius-Clapeyron approach on the other hand results in deviations of 3.14 (3.9 %) to 25.54 J mol −1 K −1 (23.4 %). The corresponding MADs for the calculated vaporization entropies are 15.23 J mol −1 K −1 for the standard QCE approach and 15.01 J mol −1 K −1 for the Clausius-Clapeyron approach. Although these values imply a slightly better performance of the Clausius-Clapeyron approach, the difference between the MADs is negligible and overall both approaches can be considered to perform equally well for the calculation of vaporization entropies. However, as an advantage, the QCE approach allows the calculation of vaporization entropies at every temperature, while the Clausius-Clapeyron approach is restricted to the boiling point. Additionally, the application of the Clausius-Clapeyron approach is computationally much more cumbersome compared to the standard QCE approach. As explained before, the standard QCE approach requires only two QCE calculations, one of them performing an optimization of a mf and b xv , and the other one serving as gas phase reference employing a mf = 0.0 and b xv = 1.0. The Clausius-Clapeyron approach on the other hand requires an optimization of a mf and b xv as well, however, these optimized parameters have to be employed for several further QCE single point calculations at different pressures. Since in the present case, we sample a pressure range from 20 kPa to 500 kPa in increments of 20 kPa, this sampling procedure corresponds to 25 QCE calculations which in total sums up to 26 QCE calculations required for the present application of the Clausius-Clapeyron approach. Although the computing time of QCE calculations with fixed parameters a mf and b xv is only a matter of seconds, the computational superiority of the standard QCE approach becomes relevant if the calculation of a huge amount of vaporization enthalpies or entropies is desired. Since the quantification of the influence of uncertainties in the physical input quantities on the results of QCE calculations requires the execution of many QCE computations, we will base the following discussion on the standard QCE approach.</p><!><p>As mentioned above, the two empirical parameters a mf and b xv are optimized in order to reproduce some experimental reference data as accurate as possible. Since these data are prone for measurement errors within the experiments, the uncertainty of the experimen-tal data will inevitably affect the results of a QCE computation and thus the calculated thermodynamics of vaporization. In order to investigate the dependence of the calculated vaporization thermodynamics with respect to uncertainties and errors in the experimental input data, we intentionally increased and decreased the employed reference densities and boiling points (see Table 1) by 20 %, respectively, using increments of 1 %. Note that the minimum boiling point that can be employed corresponds to the temperature at which the experimental vaporization enthalpy or entropy was measured. Therefore, for the vaporization enthalpies, the boiling points of butanol and methanol could be decreased by 19 % and 11 %, respectively, while this was not possible for the other solvents. Using these artificial input values, we calculated the corresponding vaporization enthalpies for each resulting combination of boiling point and density. This analysis was performed for all seven investigated compounds and the obtained vaporization enthalpies were plotted against the corresponding densities and boiling points. The results for ethane and butanol are shown in Fig. 2 and 3, respectively. The corresponding graphs for the remaining substances can be found in the supporting information, Fig. S1 to S5. Additionally, Table 4 summarizes the vaporization enthalpies obtained with the minimum and maximum densities and boiling points as well as the relative deviations to the vaporization enthalpies obtained with the experimental reference data.</p><p>Generally, Fig. 2 and 3 show that the vaporization enthalpy changes smoothly upon the variation of density and boiling point. This behaviour is observed for the other systems as well (see supporting information, Fig. S1 to S5) which paves the way for the later application of the Gauss-Hermite quadrature rather than the Monte Carlo method for the computation of Eq. 31. Considering the variation of the density in greater detail, both Fig. 2 and 3 as well as the data reported in Table 4 suggest that the variation of the density has only a small influence on the calculated vaporization enthalpies. Considering ethane as an example, when employing its boiling point of 184.55 K combined with densities varying from 0.4357 to 0.6535 g cm -3 , the obtained vaporization enthalpies are within a range of 13.14 to ). Vertical line: reference boiling point (left) 54 and density (right); 54 horizontal lines: experimental ∆ vap H at 298 K. 54 13.76 kJ mol −1 . The deviations with respect to the vaporization enthalpy obtained with the reference density of 0.5446 g cm -3 (13.50 kJ mol −1 ) are equal to 0.36 (decreased density) and 0.26 kJ mol −1 (increased density), respectively. Thus the variation of the density by ± 20 % causes a relative decrease/increase of only 2.6/1.9 % in the vaporization enthalpy of ethane.</p><p>A similar behaviour is observed for butanol as well. Here, a combination of the reference boiling point with densities varying from 0.6930 to 1.0396 g cm -3 results in vaporization enthalpies within a range of 48.81 to 49.56 kJ mol −1 . The deviations with respect to the vaporization enthalpy obtained with the reference density of 0.8663 g cm -3 (49.22 kJ mol −1 ) are equal to 0.41 (decreased density) and 0.34 kJ mol −1 (increased density), respectively, which corresponds to relative deviations < 1 %. Given the fact that the measurement errors of experimental densities can be estimated to be around 0.0001 g cm -3 ( 1 %), it can be expected that these experimental uncertainties will not significantly affect the vaporization enthalpies obtained by the standard QCE approach.</p><p>Nevertheless, some interesting trends are recognizable from the data shown in Table 4.</p><p>It can be observed that for all investigated substances a decrease of the density consistently causes a decrease in the vaporization enthalpy and vice versa. This is reasonable, since a decrease of the liquid phase density results in the liquid phase becoming more similar to the gas phase. However, the aprotic substances seem to be much more sensitive to a density variation than the protic ones, i.e. the variation of the density by ± 20 % induces deviations of 1.9 % to 5.0 % in the vaporization enthalpies of the aprotic liquids while the relative deviations observed for the protic substances are only between 0.4 % to 0.8 %. This is due to the fact that the main interactions in the protic substances are hydrogen bonds, which are not affected by the QCE parameters and experimental inputs.</p><p>The variation of the boiling points has a significantly larger influence on the calculated vaporization enthalpies which is immediately recognizable from Fig. 2 and 3. While in the case of ethane the increase of the density by 20 % caused an increase of the vaporization enthalpy by only 0.26 kJ mol −1 , the increase of the boiling point by 20 % results in an increase of the vaporization enthalpy by 3.07 kJ mol −1 . For comparison, this corresponds to an almost negligible increase of the vaporization enthalpy by 1.9 % upon density variation, but to an increase of 22.6 % upon variation of the boiling point. A similar behaviour is observed for all other investigated substances as well and the increase of the boiling point by 20 % induces an increase of the vaporization enthalpies by more than 20 % for all substances investigated. Since a decrease of the boiling point was only reasonable for methanol and butanol, a detailed investigation on how the vaporization enthalpies change upon a boiling point decrease is not feasible here, however, the results obtained for methanol and butanol suggest that a lowering of the boiling point has a similarly strong effect on the vaporization enthalpies as its increase. In the case of butanol, a decrease of the boiling point by 19 % results in a decrease of the vaporization enthalpy by 11.50 kJ mol −1 (23.4 %) and in the case of methanol a decrease of the boiling temperature by 11 % causes a decrease in the vaporization enthalpy of 5.02 kJ mol −1 (11.6 %).</p><p>It can thus be concluded that the high sensitivity of the vaporization enthalpies with respect to the employed boiling points necessitates the availability of accurate reference data.</p><p>However, considering the expanded uncertainties of the experimental boiling temperatures summarized in Table 1, the estimated measurement errors are typically below 1 K (< 1 %).</p><p>This suggests that typical measurement errors of experimental boiling points will not significantly affect the calculation of vaporization enthalpies when employing the standard QCE approach. Varying the two input quantities density and boiling point: Entropy</p><p>In order to investigate the dependence of the vaporization entropies with respect to changes in the input densities and boiling points, we performed the analysis employed in the previous section for the calculation of vaporization entropies as well. Note that for the calculation of the vaporization entropies, it was not possible to decrease the boiling points. The results for ethane and butanol are shown in Fig. 4 and 5 and Table 5 summarizes the vaporization entropies obtained with the minimum and maximum densities and boiling points as well as the relative deviations to the vaporization entropies obtained with the experimental reference data. The corresponding graphs for the remaining systems can be found in the supporting information, Fig. S6 to S10.</p><p>As already observed for the vaporization enthalpies, the vaporization entropies change smoothly upon the variation of density and boiling point. The influence of density variations on the vaporization entropies is comparable to what has been observed for the vaporization enthalpies and a decrease/increase of the density consistently causes a decrease/increase of the vaporization entropy. In case of the aprotic substances, a 20 % decrease of the densities causes a decrease in the vaporization entropies by 2.6 to 4.9 %. A 20 % increase of the density on the other hand causes an increase of the vaporization entropies by 2.0 to 3.4 %.</p><p>Considering the protic liquids methanol and butanol, the variation of the density has an even smaller influence here and the vaporization entropies change by only 0.7 to 1.6 % when the density is varied by ± 20 %.</p><p>While the variation of the boiling point was found to have a strong influence on the vaporization enthalpies, the vaporization entropies seem to be less prone for changes in the boiling temperature. For instance, the increase of the boiling point of ethane by 20 % causes an increase in the corresponding vaporization entropy from 73.13 J mol −1 K −1 to 75.07 J mol −1 K −1 . This corresponds to an almost negligible increase of 1.94 J mol −1 K −1 or 2.7 %. The remaining systems show a similar dependence on the variation of the boiling point and the relative changes of the vaporization entropies induced by a 20 % increase of the boiling temperature are within a range of 1.4 % to 4.3 %.</p><p>Overall it can be concluded that density and boiling point variations have similar effects on the vaporization entropies. While the density has a similar influence on the vaporization enthalpies and entropies, the boiling point has a much stronger effect on the vaporization enthalpies than on the entropies. This effect can be attributed to the fact that the entropy calculated in the QCE model (see Eq. 26) is not directly dependent on the electronic partition function q elec and thus not directly dependent on the mean-field parameter a mf . Since the main influence of the variation of the boiling point is observed in a strong change of the a mf parameter, as discussed further in the next section, this results in a smaller dependence of the entropy on the input boiling point. Given the fact that the typical measurement errors of densities and boiling points are < 1 %, these uncertainties will mostly not significantly affect the calculation of vaporization entropies with the standard QCE approach. Right: plotted versus the input boiling point, each curve represents a constant density (from 0.4357 to 0.6535 g cm -3 ). Vertical line: reference boiling (left) 54 and density (right). 54 ). Vertical line: reference boiling point (left) 54 and density (right). 54 Varying the two input quantities density and boiling point: Cluster populations</p><p>In order to investigate the dependence of the cluster populations with respect to the input quantities, Fig. 6 and 7 show the respective cluster populations of ethane and butanol obtained with the experimental reference data from Table 1 as well as the cluster populations obtained with the minimum and maximum densities and boiling points. For the sake of clar- ity, we summarized the populations of the monomers, dimers, trimers and larger oligomers (> 3 monomers), respectively.</p><p>For ethane, the larger oligomers are found to be the dominant species at low temperatures. However, with increasing temperature the populations of these clusters decrease at the expense of increasing monomer, dimer and trimer populations. The transition from the condensed to the gas phase manifests as a sharp transition at which the monomer becomes the dominant species. Over the whole liquid range, the dimers are only minorly populated.</p><p>Figure 6 shows that an increase or decrease of the density has only a small influence on the cluster populations (compare vertical plots) which explains the low influence of density variations on the calculated vaporization thermodynamics. Nevertheless, some interesting changes can be observed. First of all, an increasing density causes increasing populations of the larger ethane oligomers, mainly at the expense of trimer populations at low tempera- A smaller sensitivity with respect to the density variation is observed for the butanol populations shown in Fig. 7. Considering the results obtained with the experimental reference temperature, the liquid phase is continuously dominated by the large oligomers. Below 300 K these clusters make up the entire population, whereas their populations slowly decrease above 300 K at the expense of increasing monomer, dimer and trimer populations. The phase transition manifests again as a sharp transition and the gas phase is solely composed of monomers. An influence of the density variation on the populations is only recognizable close to the boiling point, where a decrease/increase of the density causes a decrease/increase in the populations of the larger oligomers. The variation of the boiling temperature has a much more pronounced influence on the cluster populations (compare horizontal plots). In the case of ethane, an increase of the boiling temperature causes a decrease in the monomer population at a given temperature, mainly at the expense of greater oligomer populations. This can be understood in the sense that larger clusters are required to keep the system in the condensed phase over a wider temperature range. Thus, the same effect as observed for the density applies for the boiling temperature: the increase of the respective quantity leads to an increased population of large clusters. When employing the experimental reference boiling point, the liquid phase of butanol is mainly composed of oligomers, whose populations decrease towards the phase transition at the expense of increasing monomer, dimer and trimer populations. The boiling point increase on the other hand leads to increased monomer, dimer and trimer populations in the liquid phase close to the phase transition. However, considering the temperature at which the vaporization enthalpy was calculated, namely 298 K, the variation of the boiling point barely induces any changes in the populations at this particular temperature. Here, the larger oligomers are the dominant species and occupy the entire population, thus, a closer look into the oligomer populations is required in order to identify the reasons for the changes in the vaporization enthalpy. We observe that there are three oligomers which make up most of the oligomer populations, namely a tetramer, an octamer and a nonamer (see Fig. S16).</p><p>At 298 K, a decrease of the boiling point leads to an increase of the tetramer population and a simultaneous decrease of the octamer and nonamer populations. An increase of the boiling point on the other hand causes a decreased tetramer population and increased populations of the octamer and nonamer. These findings are in agreement with what has been observed for ethane. In general it can thus be concluded that a decrease of the boiling temperature induces increased populations of small clusters while an increase of the boiling temperature induces increased populations of larger clusters.</p><p>The observations from the last three sections are supported by the changes in the empirical parameters a mf and b xv in dependence of the experimental inputs. All empirical parameters obtained from the QCE calculations with increased/decreased density and boiling point are given in the supporting information, Table S3 and S4. For the case of butanol, increasing the boiling point by 20 % increases the mean-field parameter a mf from 1.43 to 2.28 J m 3 mol −2 while a decrease of the boiling point decreases the parameter to 0.50 J m 3 mol −2 . This behaviour is in agreement with the prior explanation that stronger interactions between the clusters are required in order to increase the temperature range in which the system stays liquid (and vice versa). The variation of the density influences the mean-field parameter much less, decreasing it to 1.21 J m 3 mol −2 for the increased density and increasing it to 1.77 J m 3 mol −2 for the decreased density.</p><p>However, the b xv parameter, which scales the cluster volumes, is strongly influenced by the density variation. Here an increase of the density by 20 % yields significantly smaller cluster volumes with b xv decreasing from 0.90 to 0.75. The decrease of the density yields a value of 1.12. This is again in agreement with the prior explanations inasmuch as smaller cluster volumes lead to the clusters being closer together, ultimately increasing the density. In contrast to this, b xv seems to be only slightly affected by variations in the boiling temperature, increasing to 0.91 for the increased boiling point and decreasing to 0.84 for the decreased boiling point. These observations additionally point to the fact, that the density mainly influences the scaling of the cluster volumes (b xv ) and the boiling point mainly influences the mean-field (a mf ) in QCE calculations.</p><!><p>Up to now we have examined the sensitivity of the cluster populations and vaporization thermodynamics with respect to changes in the input densities and boiling points. As a last step we want to quantify the error that stems from the inaccuracies of the experimental input data at the example of the vaporization enthalpy of butanol. Therefore we considered the density and boiling point as normally distributed quantities with standard deviations equal to the measurement errors (see Table 1) and with the experimentally measured values corresponding to the mean of the distribution. Based on that, we calculated a realization of the Monte Carlo estimator using different numbers of data points up to N=10 6 as well as the Gauss-Hermite quadrature with up to N=9 quadrature nodes. As a reference we take the Gauss-Hermite estimator with N=10, which, from the numerical results below, seems to be sufficiently accurate. The random input quantities for the Monte Carlo estimator were created using the Box-Muller transform.</p><p>Table 6 summarizes some of the calculated estimators. Generally it can be seen that the Monte Carlo estimator reaches the estimator of the Gauss-Hermite quadrature with N=10 with increasing number of data points. However, the Monte Carlo estimator fluctuates which is due to the random nature of the produced input values. This can be nicely seen when comparing the estimators of MC 1000 and MC 10 4 , as the MC 1000 estimator is closer to the estimator of the Gauss-Hermite quadrature than the MC 10 4 estimator. The difference between the MC 10 and Gauss-Hermite N=10 estimator is only about 0.01 kJ mol −1 , which, from a chemical point of view, is only a minor difference. This suggests that the error introduced by the experimental input quantities is rather small.</p><p>To finally quantify the error introduced by the uncertainties in the experimental reference data, we consider the standard deviation (Eq. 32) computed with the Gauss-Hermite quadrature. The thereby obtained value equals to about 30.95 • 10 −3 kJ mol −1 and corresponds directly to the error introduced by the measurement errors of the experimental reference data. From a chemical point of view this can be considered as a negligible error for physicochemical properties.</p><p>In Fig. 8 we compare the convergence behaviour of the Monte Carlo and Gauss-Hermite estimator. Therefore we plot the deviation of the calculated Monte Carlo estimators with respect to the Gauss-Hermite estimator with N=10 nodes as a function of the data points.</p><p>Additionally Fig. 8 shows the error of the Gauss-Hermite estimators with N=1-9 nodes with respect to the Gauss-Hermite estimator obtained with N=10. Note the logarithmic scale.</p><p>The classical fluctuations of the Monte Carlo estimator, already seen in Table 6, can be observed here as well. With a convergence behaviour of 1/ √ N (shown by the dashed line), the Monte Carlo estimator shows a rather slow convergence compared to the Gauss-Hermite estimator, which already seems to reach numerical computer accuracy at N=4 nodes. The fact that the Gauss-Hermite estimators start to fluctuate for N=4 and larger is directly connected to this fast convergence. Overall it can thus be concluded that the Gauss-Hermite quadrature is an efficient approach that can be used to estimate error propagation with low computational cost as long as the investigated quantity behaves smoothly with respect to the error afflicted input. ters, as well as significantly smaller values of b xv . This is reasonable since larger clusters and smaller cluster volumes are required to model higher densities. A similar behaviour is observed for the boiling point, whose increase leads to higher populations of larger oligomers at the expense of smaller clusters. Additionally the empirical parameter a mf becomes significantly larger, increasing the inter-cluster interactions. These effects can be understood based on the fact that stronger interactions are required for the system to stay in the liquid phase over a broader temperature range.</p><p>Using both the Monte Carlo method and the Gauss-Hermite quadrature, we addressed the influence of the measurement errors in the experimental inputs on the calculated enthalpies of vaporization of the polar (R)-butan-2-ol. As total error introduced by the experimental inaccuracies we obtained a value of about 30.95 • 10 −3 kJ mol −1 . This corresponds to a negligible error in a chemical sense, since most experimental measurement errors are around 4 kJ mol −1 . This means, that experimental reference data can be used to calculate vaporization enthalpies with QCE without the need to consider measurement errors in greater detail. However, this statement is not necessarily transferable to the calculation of other thermodynamic quantities or to methods based on different theories.</p><p>It was also shown that the Monte Carlo estimator slowly converges to the predicted Gauss-Hermite estimator determined with a number of N=10 nodes. The Gauss-Hermite quadrature on the other hand shows a fast convergence and from four nodes on the estimators start to fluctuate as numerical precision is already reached. This undermines the superior performance of the Gauss-Hermite approach compared to the Monte Carlo approach.</p>

ChemRxiv

Short-Range Imbalances in the AMBER Lennard-Jones Potential for (Deoxy)Ribose…Nucleobase Lone-pair…π Contacts in Nucleic Acids

The lone-pair…π (lp…π) (deoxy)ribose…nucleobase stacking is a recurring structural motif in Z-DNA and RNAs that is characterized by sub-van der Waals lp…π contacts (<3.0 Å). It is part of the structural signature of the CpG Z-steps in Z-DNA and r(UNCG) tetraloops. These nucleic acid structures are poorly behaving in molecular dynamics (MD) simulations. Although the exact origin of these issues remains unclear, a significant part of the problem might be due to an imbalanced description of non-bonded interactions including the characteristic lp…π stacking. To gain insights into the links between lp…π stacking and MD issues, we present an in-depth comparison between accurate large-basis-set double-hybrid Kohn-Sham density functional theory calculations DSD-BLYP-D3/ma-def2-QZVPP (DHDF-D3) and data obtained with the non-bonded potential of the AMBER force field (AFF) for NpN Z-steps (N = G, A, C, U). Among other differences, we found that the AFF overestimates the DHDF-D3 lp…π distances by ~0.1-0.2 Å while the deviation between the DHDF-D3 and AFF descriptions sharply increases in the short-range region of the interaction. Based on atom-in-molecule (AIM) polarizabilities and SAPT analysis, we inferred that the DHDF-D3 vs. AFF differences partly originate in the Lennard-Jones (LJ) parameters that are identical for nucleobase carbon atoms despite the presence/absence of connected electron withdrawing groups that lead to different effective volumes or vdW radii. Thus, to precisely model the very short CpG lp…π contact distances, we recommend revision of the nucleobase atom LJ parameters. Additionally, we suggest that the large discrepancy between DHDF-D3 and AFF short-range repulsive part of the interaction energy potential may significantly contribute to the poor performances of MD simulations of nucleic acid systems containing Z-steps. Understanding where, and if possible why, the point-charge-type effective potentials reach their limits is vital for developing next-generation FFs and for addressing specific issues in contemporary MD simulations. Figure 2. a) Top-down view on model systems showing DME…G, 17 DME…A, DME…U and DME…C with DME located above the pyrimidine ring centroid. b) Purine (top) and pyrimidine (bottom) ring atom numberings.

short-range_imbalances_in_the_amber_lennard-jones_potential_for_(deoxy)ribose…nucleobase_lone-pair…π

Introduction<!>Computational Details<!>AMBER FF single point energy computations:<!>Minimum Interaction Energy distance (minIEd) and Interaction Energy (IE) surfaces:<!>Results and Discussion<!>AFF overestimates short-range repulsion for lp…π contacts compared to the DHDF-D3 reference.<!>LJ parameters for nucleobase carbon atoms should not be identical in force fields.<!>Interaction energies for the lp…π DME…nucleobase stacking are equally weak for all the nucleobases.<!>AFF electrostatic and LJ energies show substantial deviations from SAPT especially for compressed structures.<!>Sampling lp…π distances in the repulsive short-range region by MD simulations is problematic due to inaccurate LJ potentials.<!>Conclusion<!>Supporting Information<!>Data and Software Availability

<p>DNA and RNA are key biological molecules whose shapes are primarily defined by nucleobase…nucleobase, 1 nucleobase…phosphate 2 and (deoxy)ribose…nucleobase 3 interactions. A sub-category of the latter interactions involves a T-shaped (deoxy)ribose…nucleobase stacking. 4,5 This interaction has been first identified in Z-DNA CpG steps 6 and was subsequently named lone-pair…π or lp…π interaction 7 while the associated CpG steps were called Z-steps. 4 A typical CpG Z-step involves a T-shaped stacking arrangement of the cytidine (deoxy)ribose with the guanine base 4 (Figure 1). Recent DNA and RNA PDB surveys 4,5 revealed the occurrence of 'Z-like' steps that correspond to dinucleotide sequences involving any combination of the four nucleobases and possessing structural characteristics similar to those of a CpG Z-step, i.e. a 3'-nucleotide in a syn conformation, a 5'-nucleotide (deoxy)ribose with a C2'-endo pucker and a sub-van der Waals (sub-vdW) lp…π contact below 3.2 Å. The recurrence of short T-shaped (deoxy)ribose…nucleobase contacts that occur in Z-like steps or between non-consecutive nucleotides stresses the structural relevance of lp…π motifs in nucleic acids. 4,5,8 Detailed PDB surveys of Z-like steps have shown that in some instances the O4' atom points towards the nucleobase with contact distances ≤ 3.0 Å and sometimes close to 2.8 Å. 4,8 These salient stacking distances, which are among the shortest distances observed between an oxygen atom and a planar ring, raised questions regarding the strength and nature of the (deoxy)ribose…nucleobase stacking and in general of lp…π contacts. 4,5,7,9,10 Do orbital interactions contribute to their stabilization or is the common non-covalent interaction framework sufficient to describe lp…π contacts? [9][10][11][12][13][14][15][16][17] Recently, we performed energy decomposition analysis via high-level symmetryadapted perturbation theory (SAPT) 17 on a deoxyribose…guanine system that showed that the lp…π contact is mainly stabilized by London dispersion and to a lesser extent by electrostatic interactions. Thus, it became clear that the deoxyribose…nucleobase stacking is essentially a common non-covalent interaction. Moreover, we found that the origins of the observed sub-vdW contacts cannot be explained by orbital effects as the "lp…π interaction" terminology would imply. 17 Instead, we established that the sub-vdW lp…π contacts could be rationalized through atom-in-molecule (AIM) α(0) polarizabilities, a measure that provides an estimate of effective atomic volumes. 18,19 The calculated α(0) polarizabilities revealed that atomic volumes of the sp 2 carbon atoms of a guanine nucleobase do vary as a function of their chemical environment and that the carbon atoms connected to electron withdrawing groups are among the smallest of the nucleobase. Indeed, a small effective vdW radius for a sp 2 carbon may explain a lp…π contact distance in the 2.8-3.0 Å range that is much shorter than the 3.3-3.4 Å vertical distance between stacked base pairs in canonical double helical systems, 20 the latter distance being twice the 1.70 Å vdW radius of an sp 3 carbon mentioned in the Bondi tabulation dated from 1964 and still in use. 17,[21][22][23] The smaller vdW radius of some sp 2 carbons may also explain the absence of orbital effects.</p><p>Understanding vdW radii variations has far reaching consequences, one of them being the calibration of biomolecular force fields (FF) used for molecular dynamics (MD) simulations. The AMBER FF (AFF in the following), based on the 1995 Cornell et al. parametrization, 24 is currently one of the most popular nucleic acid force fields. [25][26][27][28][29][30] Most of the recent AFF modifications focused on adapting the backbone dihedral terms, which is one of the most straightforward tuning methods. 25 Additional modifications of non-bonded terms have been suggested, including direct modifications of vdW parameters of phosphate group oxygen atoms, adjustments of Lorentz-Berthelot combination rules for Lennard-Jones (LJ) solute-solute and solute-solvent interactions as well as the addition of simple H-bond interaction tuning terms. [31][32][33][34][35][36][37][38] However, none of these modifications have been widely tested and no attempts have been made to address lp…π parametrizations in nucleic acids.</p><p>Among nucleic acid systems, the Z-steps containing Z-DNA helices and r(UNCG) tetraloops are especially challenging for MD simulations. 32,34,37,[39][40][41][42][43][44][45] Previous studies exploring the behavior of Z-DNA in MD simulations focused mainly on the dynamics of backbone dihedrals and helical parameters. 39,41,[46][47][48] A recent study that analyzed the disruption of the r(UUCG) tetraloop canonical structure occurring in µs long MD simulations focused more precisely on intramolecular interactions within the loop region and traced several potential FF issues, the ribose…guanine lp…π stacking being one of those. 40 To this point, it is unknown to what extent a potentially imbalanced description of lp…π contacts causes MD simulations of Z-step-containing systems to ill-behave.</p><p>The purely non-covalent nature of (deoxy)ribose…nucleobase lp…π contacts indicates that they could be reasonably modelled by the AFF, as previously suggested for base…base stacking. 49,50 However, the (deoxy)ribose…nucleobase stacking includes short lp…π contacts where the simple LJ potential starts to deviate from more rigorous descriptions. 51,52 Thus, the approximate nature of the empirical potential might lead to errors affecting the outcome of the simulations as observed for some specific base…base stacking motifs. 46,50,53 A few wellknown issues arising from the oversimplified non-bonded potential of the pairwise-additive FFs are: i) the use of a r −12 term for the repulsive part of LJ potential that lead to an overrepulsive short-range region 51,52 instead of a more accurate exponential distancedependence as found in a Buckingham potential; 54 ii) missing atomic anisotropicity; iii) missing multi-body effects for the London dispersion component of the LJ term; and iv) missing (explicit) polarization and charge-penetration terms as a consequence of using fixed atomic point charges. 30,[55][56][57][58][59][60] Regarding deficiencies of the fixed point-charge based FFs for short-range electrostatics, various correction schemes were proposed. [61][62][63][64] Notable advances for polarizable FFs have been made. 61,[65][66][67] Recent QM derived FF approximations provided excellent results for the treatment of short-range electrostatics. 68,69 However, these rather computationally demanding FFs are not suited for explicit-solvent MD simulations of large biomolecules.</p><p>Herein, we extend our previous QM characterization of the (deoxy)ribose…nucleobase stacking in CpG Z-steps 17 to all NpN (N = G, A, C, U) Z-like steps by varying the 5' nucleobase 4,8 The represented C3pG4 step is extracted from the ultra-high resolution (0.54 Å) Z-DNA structure (PDB 3P4J). 17,72 In this arrangement, the lp…p stacking contact distance (or O4' atom to guanine plane distance) is 2.85 Å and is represented by a dashed line.</p><p>(the 3' nucleotide can be of any type since it establishes a non-specific contact to the 5' nucleobase through its O4' (deoxy)ribose atom). We use highly-accurate dispersion-corrected double-hybrid density functional (DHDF-D3) calculations to evaluate the performance of the non-bonded potential of the AFF by constructing minimum interaction energy distance (minIEd) surfaces. For that purpose, and as described earlier, 17 we approximated the full (deoxy)ribose by a dimethylether (DME) molecular probe. Rationalization of the DHDF-D3 surface topologies is made by using atom-in-molecule (AIM) polarizabilities. 70,71 To further assess the performance of the AFF, these data are compared against symmetry-adapted perturbation theory (SAPT) computations. Lastly, we analyze the behavior of Z-steps in MD simulations of Z-DNA and r(UUCG) tetraloop systems by employing the latest AFF versions. We found that the massively sampled short O4'…guanine plane distances are likely to cause imbalances in the simulations because of the overestimated short-range repulsion originating from the LJ potential of the AMBER FF. This led us to suggest that the usage of unified LJ parameters for all nucleic acids carbon atoms contribute to a yet unrecognized degree to the AMBER FF errors.</p><!><p>Structural models: As described earlier, 17 a CpG Z-step (see Figure 1) was extracted from a 0.54 Å ultra-high resolution Z-DNA crystallographic structure (PDBid: 3P4J; 72 only the deoxyribose of residue 3 and the guanine of residue 4 were kept). Then, the deoxyribose was changed into dimethylether (DME) while keeping the C-O4'-C atoms fixed in space to preserve the X-ray deoxyribose…guanine orientation (in the following, the DME oxygen atom will be named O4'). The DME orientation was maintained for all nucleobases (note that the DME orientation may differ for lp…p contacts involving non-consecutive nucleotides). 5 To study all (deoxy)ribose…nucleobase stacking types, G was substituted by each of the three A, C or U nucleobases (Figure 2). Most of the computational protocol followed procedures described in earlier work. 17 QM structure optimization: Prior to making rigid-monomer interaction energy scans, the DME-nucleobase systems were gradient-optimized by the hybrid density functional approximation B3LYP 73 with DFT-D3(BJ) 74,75 dispersion correction and the def2-QZVP 76 atomic orbital basis set (B3LYP-D3/def2-QZVP). Harmonic penalty restraints as implemented in our in-house optimizer 77,78 were used to keep the initial DME…nucleobase fragment orientation (restraints are listed in Table S1 of the Supporting Information). Turbomole 79,80 energy and gradient calculations employed an energy change threshold of 10 -7 Eh for the SCF and the total energy convergence. The RI-JK integral approximation (density-fitting) was used with the m4 DFT quadrature grid. 81,82 The XYZ coordinates for the optimized structures are provided in the Supporting Information.</p><p>Reference ab initio CCSD(T) and SAPT calculations: PSI4 83 was employed for density fitted FNO-CCSD(T) 84 and density fitted SAPT2+(3)δMP2 [85][86][87] computations using the frozen core approximation. Extrapolation to complete basis set (CBS) FNO-CCSD(T) computations was automated with PSI4 using the formula: 𝐸 !!"#(%)/!(" = 𝐸 )*/+, + 𝐸 -.//!("(+%,+,) + 𝐸 *12-!!"#(%)/456%7 − 𝐸 -.//456%7 (1) where 𝑎𝑇, 𝑎𝑄 and 𝑗𝑢𝑛𝑇𝑍 represent the aug-cc-pVTZ, aug-cc-pVQZ and jun-cc-pVTZ basis sets, respectively. [88][89][90] MP2/CBS(at,aQ) indicates the standard 2-point extrapolation following Halkier et al. 91 SAPT2+(3)δMP2 calculations also employed the jun-cc-pVTZ basis set. Both, CCSD(T) and the mentioned SAPT theory level are highly accurate QM methods if appropriate basis sets are used and are labelled as 'gold standard' for the evaluation of non-covalent interactions. 87,92 The CCSD(T) data were used to validate the DHDF-D3 approach (see below). SAPT calculations were used to decompose the interaction energy into various components.</p><p>QM DHDF-D3 single point energy computations: Energies were computed using the doublehybrid density functional approximation DSD-BLYP-D3 93 with D3(BJ) 94 correction, the minimally-augmented ma-def2-QZVPP 76,95 basis set (we use the abbreviation DHDF-D3 for DSD-BLYP-D3/ma-def2-QZVPP in following sections) and the ORCA program. 96 The "GRID5" quadrature grid was used for DFT calculations as well as the RI approximation for Coulomb integrals (RI-J) and the COSX (chain-of-spheres exchange) approximation for exchange integrals with "GRIDX6" grid. 97 Auxiliary basis sets were automatically constructed (AutoAux keyword). 98 A tight energy change threshold for SCF convergence was used (TightSCF keyword; 10 -8 Eh). The double-hybrid calculations used the frozen core approximation for the RI-MP2 contribution.</p><!><p>The in-house program bff 99 was used for calculating AFF interaction energies on QM-optimized geometries of DME…nucleobase and deoxyribose…nucleobase models. Note that for interaction energies, only the non-bonded potential of the AFF is needed. The AFF interaction energy was calculated as</p><p>where the first term is the electrostatic potential describing the pairwise Coulomb interactions between all intermolecular atomic pairs 𝑖 and 𝑗 by using 𝑞 8 and 𝑞 4 partial atomic charges and 𝑟 84 interatomic distance. The 332 factor takes care of the conversion to kcal.mol −1 from charges in atomic units and distances in Ångström. The second term is the Lennard-Jones (LJ) potential composed of a repulsive (𝑟 84 @>/ ) and an attractive (−𝑟 84 @? ) part. The LJ parameters are represented by the 𝑅 ;<= (84) optimal interatomic distance and by the 𝜀 84 potential well depth. Partial charges were calculated by using the Antechamber 100 automated RESP 101 procedure at the recommended HF/6-31G* level of theory. LJ parameters for the "parm99bsc0 102 AMBER variant" were taken from the AmberTool library files of AMBER16. 103 These LJ parameters are identical to those of the original Cornell et al. AMBER parametrization 24 and to those used in MD simulations of Z-DNA and r(UUCG) (see AMBER MD simulations and Table S5). The atom types and partial charges used for the AFF calculations are available as text files with '.resp' ending in the Supporting Information.</p><!><p>To compute these surfaces, we have performed vertical interaction-energy scans for a series of positions of the DME probe covering the entire purine and pyrimidine nucleobase rings. We define the "minimum interaction energy distance" (minIEd) as the O4'…nucleobase distance at which the lp…π DME…nucleobase interaction energy is at its minimum. The complementary interaction energy (IE) surface shows the corresponding interaction energy minima to the minIEd surface. Thus, both surfaces use the same geometry for a given xyz coordinate; the IE surface displays the value of the interaction energy minimum and the minIEd surface displays the corresponding lp…π distance. Note that the minIEd distance is not systematically associated with the interaction energy strength as the balance of interactions differs depending on the local environment of the probed point. We found that it is useful to construct both minIEd and IE surfaces for the DHDF-D3 vs. AMBER comparison. The minIEd surface allows to compare the vertical distances while the IE surface shows the energetics of the interaction. A schematic that explains how the minIEd and IE values are derived from vertical scans is provided in Figure S1 of the Supporting Information.</p><p>Note that we changed the previous term closest-contact surface 17 to the minIEd surface, since the former term could be perceived as ambiguous outside the structural biology community.</p><p>The four-dimensional data sets (xyz coordinates and interaction energy) used for the construction of the surfaces were collected as follows: first, the O4' atom of the optimized DME…nucleobase system was placed above the centroid of the pyrimidine ring at a 2.5 Å vertical distance. Then, DME was moved as a molecular probe in horizontal and vertical directions with respect to the nucleobase to sample 29 (horizontal scan) x 20 (vertical scan) = 580 points for purines and 20 (horizontal scan) x 20 (vertical scan) = 400 points for pyrimidines to cover the whole rings (see Figure S1 and Parameters for the vertical scans of the nucleobase surfaces in the Supporting Information).</p><p>One additional horizontal point was added for the DHDF-D3 surfaces of C and U pyrimidine nucleobases outside the ring next to the C2 atom (along the centroid-C2 vector) to localize more precisely the minIEd surface minima (Figure S1). Gnuplot was used to create the surfaces using its internal algorithm for data interpolation. 104 Limitations to this modelling approach are described in the Surface scan limitations section of the Supporting Information.</p><p>Atom-in-molecule (AIM) α(0) polarizabilities calculated with DFT-D4 dispersion theory using Grimme's dftd4 code were used to rationalize DHDF-D3 surface shapes. 17,71,105 AMBER MD simulations: Molecular dynamics (MD) simulations of a Z-DNA [(CpG)3]2 hexamer duplex (PDBid: 3P4J X-ray structure) 72 were performed in AMBER18 106 by employing the AMBER OL15 force field 39 in an NPT ensemble using the Langevin thermostat with a collision frequency of 2 ps −1 and the Berendsen (weak-coupling) barostat. 107,108 A total simulation time of 3 µs (3x1 µs trajectories with randomized initial velocities) was achieved at both 100 K and 300 K temperatures. The CUDA-driven pmemd module (SPFP-implementation) of Amber was used, 109 along with a 2 fs time step and standard SHAKE restraints. 110 A SPC/E 111 water box with a 10 Å buffer region and with an octahedral periodicity was used. Simulations were done in Na + excess salt conditions of 0.2 M NaCl resulting from the addition of 15 Na + and 5 Cl -. 112 Ions were placed using AmberTool's leap module that uses basic electrostatic mapping.</p><p>MD simulations of the r(ggcacUUCGgugcc) tetraloop hairpin were initiated from the NMR structure (PDBid: 2KOC 113 ) and data were taken from our previous work. 40 Simulations were performed with the AMBER OL3 (i.e., ff99bsc0χOL3) 114 RNA force field with modified LJ parameters for phosphate oxygens, 31 associated dihedral adjustments, 115 and the external gHBfix potential. 37 Trajectory snapshots were saved every 10 ps for both Z-DNA and r(UUCG) tetraloop. Equilibration protocols and further details to r(UUCG) production runs are provided in the Supporting Information.</p><p>For Z-DNA, the whole 6 µs trajectories were analyzed. In contrast, we used only the initial 4.5 μs (MD1), 2.4 μs (MD2) and 2.8 μs (MD3) segments extracted from 3x10 μs trajectories of r(UUCG) tetraloop. 40 In these segments the majority of the signature interactions of the r(UNCG) tetraloop is preserved. 40 The remaining MD segments were not analyzed since the r(UUCG) tetraloop denaturates and deviates significantly from the canonical structure. Further, the ribose…guanine lp…π contact is totally lost because the guanine nucleobase flips out of the loop.</p><p>Additional software: Structures were visualized by Molden 116 and VMD 117 ; figures were prepared with PyMOL 118 , Jmol 119 and VMD; graphs were prepared by Gnuplot 104 and Xmgrace, MD analyses were performed using CPPTRAJ 120 obtained from AmberTools18 106 .</p><!><p>Vertical energy scans reveal substantial differences between DHDF-D3 and AFF lp…p descriptions.</p><p>First, we discuss two representative DME…nucleobase vertical energy scans (Figure 3; for deoxyribose…nucleobase scans see DME…G/C versus deoxyribose…G/C calculations and Figure S2 in the Supporting Information). For DME…G, the O4' atom is located above the pyrimidine ring centroid; for DME…C, the O4' atom is positioned above the C2 atom. Both positions are near the global minimum on the respective minIEd surfaces, i.e., the DME positions leading to the closest optimal contact distance (see below). Additionally, Figure 3 displays FNO-CCSD(T)/CBS (abbreviated as "CCSD(T)") reference data to validate the DHDF-D3 results. The excellent agreement between the less computationally demanding DHDF-D3 (in red) and the more accurate CCSD(T) 'gold standard' calculations (in black; Figure 3) supports the use of the DHDF-D3 approximation.</p><p>For DME…G, the DHDF-D3 interaction energy minimum at 2.97 Å versus 3.10 Å for the AFF interaction energy curve displays a clear +0.13 Å shift of the AFF minimum towards longer distances. For DME…C, the DHDF-D3 and AFF curves show a greater distance shift (+0.24 Å) with interaction energy minima at 2.91 Å and 3.15 Å, respectively.</p><!><p>To assess the importance of the AFF limitations and their implications for MD simulations, it is crucial to consider not only the position of the minima, but also the gradients of the energy curves in the regions commonly sampled in MD simulations. The too steep short-range AFF repulsive regions are a source of spurious forces that could affect the outcome of MD simulations. Figure 3 highlights large AFF energy gradients at distances below the energy minima and a sharp increase in the DHDF-D3 vs. AFF difference for intermonomer distances below 2.9 Å, where the LJ approximation for the short-range repulsion starts to break down (see AFF electrostatic and LJ energies show substantial deviations from SAPT especially for compressed structures). For instance, the AFF energies become positive below 2.7 Å, while the QM energies are still around −3.0 kcal.mol -1 at this distance.</p><p>AFF minIEd surfaces are too repulsive and shallow compared to the DHDF-D3 reference.</p><p>DME…purines. The global minimum of the DHDF-D3 minIEd surface for both purine nucleobases is located in the middle of the pyrimidine rings at 2.97 Å (G) and 2.95 Å (A) (Figure 4 and Table S2). The surface remains shallow towards the C2 and C6 atoms. A secondary minimum is located above the imidazole rings at 3.04 Å (G) and 3.01 Å (A). For G, the location of the global minIEd minimum is in agreement with PDB database analysis of NpG Z-like steps that show that the O4' atom is mainly distributed above a region located between the C2 and C6 atoms and the pyrimidine ring centroid. 4,17 The data is also consistent with a PDB database analysis taking into account (deoxy)ribose…nucleobase stacking contacts between nonconsecutive nucleotides. 5 The AFF minIEd surfaces recover some features of the DHDF-D3 surfaces, i.e., the surface minima are located close to the centroid of the pyrimidine ring (Figure 4). However, the minima lie at ≈3.10 Å, a notable deviation of 0.13 Å (A) and 0.15 Å (G) from the DHDF-D3 surface minima (Figure 4 and Table S2). Overall, the AFF minIEd surfaces are too repulsive over the whole surface area and are shallower than the DHDF-D3 minIEd surfaces. This is likely a consequence of the simple form of the AFF non-bonded potential that is unable to capture short-range electronic-structure effects like polarization, charge-penetration, and exchangerepulsion. AFF also fails to correctly describe the area between the two rings and the second minimum within the imidazole ring. The latter point should be of limited practical importance for MD simulations of systems with Z-step-like lp…π stacking since the O4' atom does not occur in positions above the imidazole ring in these structures, in contrast to what is observed for some ribose…nucleobase stacks occurring between non-neighboring residues in RNAs. 4,5 DME…pyrimidines. Z-like steps with pyrimidine instead of purine nucleobases are less frequent. Few cases were reported for RNA and only one for DNA. 4 However, in RNA different, non-Z-step-like ribose…nucleobase stacking types were observed in crystallographic structures with uracil or cytosine as a nucleobase. 5 The DHDF-D3 minIEd surfaces of both pyrimidine nucleobases are similar to each other (Figure 5 left). The minima are located close to the C2 atom and not close to the ring centroid as one would expect from the purine nucleobases data (Figure 4). This dissimilarity may result from the absence of the pyrimidine/imidazole junction that allows for smaller atomic volumes of the purine C4/C5 carbon atoms when compared to the "equivalent" pyrimidine C6/C5 atoms due to the presence of the electron-withdrawing nitrogen atoms of the imidazole ring. For instance, the minimum distance of approach to the C5 atom in pyrimidines is ≈3.22 Å while it is close to 3.15 Å (G) and 3.09 Å (A) in purines (see AIM polarizabilities α(0) explain shapes of minIEd surfaces and Figure 6). Interestingly, our DHDF-D3 results are in agreement with a PDB database study of (deoxy)ribose…nucleobase stacking that shows that the O4' atom is mostly positioned above the C2 atom of a pyrimidine nucleobase. 5 Overall, the agreement between DHDF-D3 and AFF for the pyrimidine systems is poor. The AFF minIEd surfaces of the pyrimidine systems (Figure 5) are barely capturing any of the essential features of their DHDF-D3 counterparts. For DME…C, the AFF surface shows a spread-out minimum covering a large part of the left side of the surface plot and the lowest value is by ~0.21 Å larger compared to the DHDF-D3 minimum (Figure 5 and Table S2). As a reminder, Figure 3b shows the vertical scan for the position above the C2 atom and a 0.23 Å difference between DHDF-D3 and AFF for the minimum. The lowest point of the DME…U surface lies outside the ring near the N3 atom. Adding another nucleobase point outside of was not sufficient to determine an unambiguous minimum. In lieu of a proper minimum the shortest distance of the AMBER minIEd surface is used as a pseudo-minimum for DME...U in Figure 5 and Table S2.</p><p>AIM polarizabilities α(0) explain the minIEd surface shapes. Figure 6 pictures atom-in-molecules (AIM) polarizabilities α(0) of the four nucleobases that we calculated to rationalize the shapes of DHDF-D3 minIEd surfaces (see below). The AIM polarizabilities obtained from DFT-D4 dispersion theory 70,71 approximate atomic volumes 18,19 and thus can serve as a qualitative surrogate measure for the elusive 'real' vdW surface, although atomic anisotropy information is missing. 17 The shape of the guanine minIEd surface fits nicely with the relative magnitudes of the AIM α(0) polarizabilities (Figure 6 and Kruse et al. 17 ), especially for the pyrimidine ring. The C2 and C6 atoms, being the two smallest atoms of the guanine rings, allow for the closest Color represents the 'z' coordinate of the O4' DME atom. The color scale is the same for all plots to ease visual comparison. Black arrows and distance values point to minima of the minIEd surfaces marked by red dots; small black dots mark points where the vertical scans were calculated (see also Figure S1). Note that for DME…U no AFF minIEd minimum is defined as it is outside the ring beyond the N3 atom. In this case, a grey arrow points to the N3 atom and the minIEd distance from vertical scan above the N3 atom is written in the graph. The AFF vs. DHDF-D3 difference range is 0.01-0.23 Å and 0.00-0.20 Å for C and U, respectively. O4'...ring atom contacts. For DME…A, the minIEd surface minimum spreads to the C4 and C6 atoms which are the smallest adenine rings atoms.</p><p>AIM polarizabilities explain also the minIEd surfaces of pyrimidine nucleobases where the smallest C2 and the biggest C5 atom show the shortest and the largest O4'...ring atom contact distances, respectively. Moreover, the significant shift of the minIEd minimum from the center of the six-atom rings for purines to the C2 atom for pyrimidines can be explained by the AIM α(0) polarizability values of C5 and C6 atoms (right side of the ring) that are significantly larger compared to the corresponding C4 and C5 atoms of purine nucleobases (Figure 6).</p><!><p>The large range of DHDF-D3 vs. AFF differences shows that some minima locations on nucleobase surfaces are better described by AFF than others. The largest range is observed for DME…C where the difference between the DHDF-D3 and AFF minIEd distance is negligible for O4'…C(C5), while the deviation is large for O4'…C(C2) (0.23 Å). This appears to be correlated with the differences in AIM polarizabilities between the carbons (small C2, large C5; Figure 6). It suggests that the vdW radii for the nucleobase carbon atoms -currently represented by an identical set of LJ parameters in the AFF (Table S5) -might be too large for the atoms connected to electron withdrawing groups like the guanine C2 atom. 17 Note that the AFF LJ parameters for the nucleobase sp 2 carbons are derived from Monte Carlo simulations on liquid benzene and are thus identical to those of benzene rings devoid of electro-attractive-groups (Table S5). 24 Thus, it can be assumed that the pyrimidine C5 and C6 atoms that display the largest AIM volumes are similar to the benzene sp 2 carbon atoms and that the same AFF LJ parameters can be used. But it is clearly not the case for the carbon atoms connected to electron withdrawing groups and it is timely to add some diversity to the set of nucleobase LJ parameters.</p><p>To support our rationalizations based on AIM polarizabilities, we also calculated SAPT vertical scans with (DME)O4' located above the smallest (C2) and largest (C5) cytosine atom. Results show that exchange-repulsion (essentially Pauli repulsion) as well as the London dispersion (related to atomic polarizabilities) are notably smaller when O4' is located above the C2 atom (Figure S3 in Supporting Information). This agrees with a carbon atom with a smaller effective vdW radius and a smaller α(0) polarizability value.</p><p>The DHDF-D3 and AFF minIEd surface comparison and the different AIM polarizability volumes of the ring carbon atoms indicate that (deoxy)ribose…nucleobase stacking cannot be precisely described at a quantitative level by the current uniform carbon atom AMBER LJ parameters. On one side, FFs should be parametrized with limited number of atom types to avoid over-fitting, as discussed in Schauperl et al. 121 On the other side, describing all nucleobase ring carbons by the same LJ parameters is a very rough approximation as becomes apparent from our data. This may be especially important for systems where the distance of an oxygen atom to a carbon atom is among the shortest known. Thus, we tentatively propose that LJ radii for nucleobase carbon atoms should adhere to the following size-relation: C2<C6<C4~C5 for G, C4~C6<C2~C5 for A, and C2<C4<C6<C5 for C and U and be smaller or equal to those of benzene rings. Further evaluation using either a larger dataset or following a different approach, like obtaining AIM electron density-based LJ parameters unique for each atom prior to MD simulation as proposed by Kantonen et al. 122 , is needed to precisely parametrize LJ radii values for sp 2 carbon atoms. It can be noted that earlier we calculated α(0) polarizabilities for the indole rings 17 and that the values for the two junction carbons are large and close to those calculated for pyrimidine C5 and C6 nucleobase atoms.</p><!><p>Interaction energy (IE) surfaces, defined as the strongest (minimum) interaction energies of the vertical scans, are discussed in detail in the Supporting Information. Table S2 shows that DME…G is the system with the weakest DHDF-D3 interaction energy (−3.9 kcal.mol -1 ) and DME…C the strongest (−4.4 kcal.mol -1 ). A cautious interpretation of the data is that all our systems are very close in energy, i.e., no nucleobase stands out as particularly suitable or unsuitable for (deoxy)ribose…nucleobase interactions and all the interaction energies are weak compared to typical nucleobase...nucleobase interactions (−6.0 kcal.mol −1 ). 53 The AFF reproduces the DHDF-D3 energy range around the surface minima relatively well although it fails to reproduce important structural features and short-range energies (Figure 3). We also note that interaction energies cannot be straightforwardly related to free energies in nucleic acids. 123</p><!><p>It has been suggested that short (deoxy)ribose…nucleobase contacts may be subtly vertically compressed by the surrounding structural context, because lp…π contacts in nucleic acids often occur in locally strained molecular topologies as in Z-DNA and in r(UNCG) tetraloops. 17,124,125 To evaluate the range of potential compression effects, we performed SAPT computations that decompose the interaction energy into electrostatics, London dispersion, induction, and exchange-repulsion at three vertical separations: (i) the minimum of the corresponding minIEd surface (ii) extended by 0.3 Å and (iii) compressed by 0.2 Å. For all three separations SAPT shows dominating London dispersion followed by strong electrostatics (Table 1), in agreement with previous works. 16,17 Although extending or compressing the stacking distance does not change the interaction energy substantially, as typical for non-covalent interactions with naturally flat potentials near their minimum, it does change the relative composition of the interaction energy (Table 1 and Figure 4 of Kruse et al. 17 ). At longer distances the electrostatic interactions contribute less and dispersion becomes more important, while at shorter distances, where the electron clouds overlap more, the electrostatic contribution increases as charge penetration effects come to play. 60 The AFF decomposition shows that the LJ potential dominates the interaction energy curve as the Coulomb FF electrostatics remains flat in the equilibrium region and almost without a contribution to the AFF curve (Table 1 and Figure 3). It is known that the AMBER electrostatics is fairly insensitive for nucleobase…nucleobase stacking vertical separation. 50,52 We show that this is also the case for the lp…π (deoxy)ribose…nucleobase stacking. Pointcharge schemes are known to be deficient at short-range due to the lack of explicit chargecloud penetration effects that become important once the electron densities of both molecules start to overlap ('interpenetrate') and are usually attractive for charge-neutral molecular densities. [56][57][58][59][60] However, most classical non-bonded FF potentials should be understood as effective potentials, where the accuracy of the individual parts is less important than their sum. 57 For example, the exponential distance-dependence of the charge penetration and the short-range exchange-repulsion allow for the systematic cancelling of errors between the missing charge penetration effects and the approximate r −12 distance- dependence of the repulsive part of the LJ potential in AFF. 57 Thus, the short-range chargepenetration part of QM electrostatics is effectively included in the vdW term of AFF. While these schemes work relatively well at typical vdW distances, the erroneous behavior of the too steep r −12 term of the LJ potential dominates the short-range region entirely, as also shown by a significant weakening of AFF interaction energy in the compressed lp…π structures (Table 1). For instance, while the difference for the DME…G model between the SAPT and AFF energy at the minIEd minimum is -0.3 kcal.mol -1 , it is -1.5 kcal.mol -1 when compressed by only 0.2 Å. For DME…C, the AFF interaction energy is even repulsive (+0.8 kcal.mol -1 ) above the minimum with a compression of 0.2 Å. Similar observations of a systematically too repulsive AFF potential in the short-range region of interatomic contacts have been reported before. 40,[50][51][52]57,126 The awareness about the too steep repulsion of the LJ potential goes all the way back to Buckingham. 54 However, the r −12 form (equation ( 2) in Computational details) has been favored in all major biomolecular FFs over the Buckingham exponential form for computational efficiency reasons.</p><!><p>Above, we have shown that AFF overestimates the minIEds in lp…π (deoxy)ribose…nucleobase stacking compared to the reference DHDF-D3 method (Figure 4 and 5). In addition, for the short-range repulsion part, the agreement between AFF and DHDF-D3 worsens due to the steepness of the r −12 LJ repulsive term at small contact distance (Figure 3). This means that the shorter the lp…π distances below the QM minimum, the larger the bias introduced by the AFF approximation. To evaluate the extent to which the short-range regions are sampled by MD simulations, we analyzed the (deoxy)ribose…G contacts in MD simulations of Z-DNA and r(UNCG) tetraloops systems. Z-DNA. It has been shown that MD simulations of Z-DNA using currently available AFFs lead to an overall unsatisfactory description of the double-helical geometry. 39,[46][47][48] Although not fully satisfactory regarding description of the Z-DNA backbone substates, the best performance so far has been obtained with the OL15 version of the AMBER FF. 39,48 We analyzed Z-DNA MD simulations conducted with OL15 at 100 K and 300 K. The 300 K simulations sampled various backbone substates similar to those described by Zgarbová et al. 39 including flipping of the native αg+/γt to the non-native αt/γg+ conformation in CpG steps. The 100 K simulations are mirroring the situation of the crystal structure more closely and retain well the crystal structure conformation. We noticed that the cytidine deoxyribose changes its pucker from C2'-endo to C1'-exo at both temperatures and even moves to O4'endo in the 300 K simulation.</p><p>The average O4'…G-plane distance (Figure 7) analyzed for the middle CpG dinucleotides of Z-DNA strands is 3.03 Å (300 K). If selecting only the snapshots with a preserved canonical CpG Z-step (around 18 % of all the snapshots), the average distance is reduced by 0.07 Å to 2.96 Å (see Table S6). Interestingly, sampling of short deoxyribose…G distances is correlated with a moderate deformation of the deoxyribose ring (the abovementioned shift of deoxyribose pucker to C1'-exo or O4'-endo conformations). The deoxyribose deformation appears to ease the atom-atom steric conflict and leads to extension of the O4'…G distances. While the average O4'…G distance over C3pG4 and C9pG10 in the 3P4J X-ray structure (2.87 Å) 17 is exceeded by the 300 K MD, it is well reproduced by the 100 K simulation (2.91 Å; this distance is reduced by 0.03 Å to 2.88 Å if only the C3pG4 and C9pG10 satisfying the Z-step criteria -36 % of all the snapshots -are considered; see Table S6). At the same time, the average MD distances are smaller than the AFF minIEd values for the model system indicating that the repulsive region of the non-bonded potential for the deoxyribose…G stacking is sampled in both the 100 K and 300 K simulations. However, in the 100 K simulations the Z-DNA conformation remains stable, likely due to the reduced temperature preventing crossing of conformational energy barriers, which is also the reason for the well reproduced averaged O4'…G distance. Also, it agrees with the distribution of the O4' projection points on the G nucleobase (Figure 7). While in the 300 K simulations the O4' projection points form a single cloud spreading from the center of the pyrimidine ring to the ring-atoms, there is a group of six clouds in the 100 K simulations (one for each CpG step) as the O4' occupies a smaller region of space above the nucleobase in the locked Z-DNA geometry. Figure 7 also shows no preference for any of the ring atoms by the O4' atom in 300 K simulation, which is in agreement with the spuriously flat AFF minIEd surface (Figure 4). However, in X-ray data 72 and to some extent in the 100 K simulations, the C2 atom contact is preferred. 17</p><p>The r(UUCG) tetraloop. Tracing the origins of systematic AFF shortcomings from real-system MD simulations is challenging. Inaccuracies in the approximate potentials may accumulate and are often mutually interrelated as highlighted by a recent study of the r(UUCG) tetraloop. 40 Despite efforts involving a true multi-scale approach employing small QM models, large QM/MM systems and long time-scale MD simulations no singular or conceptually simple force-field aspect leading to the observed instabilities could be identified. It was concluded that the difficulties in describing r(UNCG) tetraloops are likely due to multiple force-field imbalances that are magnified by the unusually confined conformational space of the native r(UNCG) free-energy basin.</p><p>Here, we used three 10 μs-long MD simulations of the r(ggcacUUCGgugcc) tetraloop NMR structure 113 taken from Mráziková et al. 40 to analyze O4'…G distances in three simulations segments where the tetraloop structure is preserved (see "Computational details"). 40 We observe again a high sampling of short O4'…G distances (Figure 7 right). The average from all analyzed segments is 2.93 Å, which is shorter by 0.10 Å compared to Z-DNA MD data but close to the ≈2.90 Å experimental O4'…G distances derived from both the NMR (PDBid: 2KOC 113 ) and an ensemble of X-ray structures. 4 In contrast to Z-DNA, the O4' is frequently located directly over the ring and close to the C2 atom, as displayed by the O4' projections on the G nucleobase (Figure 7). It indicates that the r(UUCG) tetraloop is too stiff to ease the steric repulsion between the O4' and the ring-atoms, which eventually leads to disruption of its native conformation. In conclusion, distances below the minIEd surface minima are sampled in both Z-DNA and r(UUCG) tetraloop trajectories. It indicates that the overall Z-DNA and UNCG tetraloop structures tend to compact the (deoxy)ribose…G stacking. Thus, due to thermal fluctuations, the simulations face an exaggerated short-range LJ potential repulsion that inevitably biases the statistical MD sampling and has the potential to destabilize the native state. This may be one of the reasons leading to a problematic description of Z-step containing nucleic acid structures in current MD simulations.</p><!><p>Herein, we evaluated the ability of the AMBER FF (AFF) non-bonded potential to describe the (deoxy)ribose…nucleobase lp…π stacking through comparisons with the highly accurate double-hybrid DFT (DHDF-D3) method. Vertical interaction energy scans calculated using DME…nucleobase (G, A, C, U) models reveal a clear over-repulsiveness of the AFF especially in the short-range region of the interaction. Moreover, the over-repulsiveness of the AFF is non-systematic and varies for different positions of the O4' atom above the nucleobase. For instance, while the difference in DHDF-D3 and AFF minimum interaction energy distances (minIEds) is significant for the C2 atom for all the nucleobases, the AFF behaves correctly for the C5 atoms of pyrimidine nucleobases. This is a consequence of notable differences in atomic volumes of the nucleobase carbon ring-atoms, as indicated by AIM α(0) polarizability calculations. This implies that using the same vdW radius for all ring-carbon atoms is not the best choice for nucleobase systems. Based on our results we suggest that nucleobase carbon atom LJ parameters, which are all identical in the AFF, should follow the size-relationships: C2<C6<C4~C5 (for G), C4~C6<C2~C5 (for A), and C2<C4<C6<C5 (for C/U) and recommend that the LJ parameters of N and O atoms in aromatic rings should also be reevaluated.</p><p>Since the (deoxy)ribose…nucleobase lp…π stacking occurs commonly in conformationally highly strained structures, the AMBER MD simulations may either sample short (deoxy)ribose…nucleobase distances, contributing to a structural strain caused by the over-estimated short-range repulsion, or avoid them, leading to disruption of local conformations. MD distance sampling of a Z-DNA helix and a r(UUCG) tetraloop strongly suggests that simulations of nucleic acid structures with Z-steps are affected by an exaggerated short-range repulsion of the O4'…G contacts. The excessive short-range LJ repulsion prevents further compaction of the O4'…G interaction that the real system would use to achieve an overall balance of the intricate Z-step conformation. Therefore, the too 'large' AFF nucleobase carbon atoms distort any structural sampling of Z-steps due to an increased strain in these highly specific and apparently stiff (and thus also brittle) conformations. The mere modification of the AFF atomic radii would not necessarily eliminate the problem, as the spurious behavior of the AFF short-range repulsion region stems from the oversimplified form of the LJ term. Thus, it is necessary to work on both issues: improving the vdW description of nucleobase atoms and finding a way to flatten the AFF lp…π interaction potential. Current computational means would allow the use of Buckingham like potentials, although this would imply significant reparameterization efforts.</p><!><p>Additional computational details, surface scan limitations, deoxyribose…G and deoxyribose…C calculations, SAPT for DME…guanine with rotated DME, discussion to IE surfaces, MD equilibration and simulations protocols, AMBER atom types and LJ parameters for nucleobase atoms, additional MD analyses, SAPT comparison between O4'…C(C2) and O4'…C(C5) (PDF)</p><p>Optimized geometries of DME…nucleobase and deoxyribose…nucleobase models; input files for bff calculations containing XYZ coordinates of the optimized geometries of DME…nucleobase and deoxyribose…nucleobase models, RESP charges and atom types; surface data containing XYZ files used for single-point energy calculations and files containing O4'…nucleobase distance and the associated interaction energy for both AFF and DHDF-D3 methods and for each nucleobase (ZIP)</p><!><p>The datasets generated during the current study are contained within the manuscript or Supporting Information. Raw trajectory data are available from the authors on request.</p><p>The in-house program used for restrained geometry optimizations can be downloaded from https://github.com/hokru/xopt (Xopt) and the QM software coupled to Xopt used for energy and gradient calculations can be purchased from https://www.turbomole.org/ (Turbomole version 7.3). The software used for DHDF-D3 single-point calculations for surface scans is freely available after registration at https://orcaforum.kofo.mpg.de/app.php/portal (ORCA version 4.0.1), the software used for CCSD(T) calculations and SAPT analysis is freely available at https://psicode.org/installs/v14/ (PSI4 version 1.4). The software used for the electrostatic potential maps calculations further used for generating RESP charges by Antechamber can be purchased from http://gaussian.com/ (Gaussian 09). Antechamber itself, as well as CPPTRAJ used for analysis of MD trajectories, is a part of freely available AmberTools found at https://ambermd.org/GetAmber.php. The MD software AMBER can be purchased from the same website. The in-house program used for single-point AFF calculations for surface scans can be downloaded from https://github.com/hokru/BrnoFF (bff). A program used for calculations of AIM polarizabilities is freely available at https://www.chemie.unibonn.de/pctc/mulliken-center/software/dftd4 (dftd4).</p><p>The used visualization programs are freely available at https://www3.cmbi.umcn.nl/molden/ (Molden), https://www.ks.uiuc.edu/Research/vmd/ (VMD version 1.9.3), http://pymol.org/ (PyMOL version 2.5.0) and http://jmol.sourceforge.net/download/ (JMOL version 14.6.4). The software used for preparing graphs is freely available at http://www.gnuplot.info/ (Gnuplot version 4.6) and http://plasma-gate.weizmann.ac.il/Grace/ (Xmgrace version 5.1.22).</p>

ChemRxiv

Binding-induced, turn-on fluorescence of the EGFR/ERBB kinase inhibitor, lapatinib\xe2\x80\xa0

We report the photophysical properties, binding-induced turn-on emission, and fluorescence imaging of the cellular uptake and distribution of lapatinib, an EGFR/ERBB inhibitor. Lapatinib, a type II, i.e. inactive state, inhibitor that targets the ATP binding pocket of the EGFR family of receptor tyrosine kinases. DFT calculations predict that the 6-furanylquinazoline core of lapatinib should exhibit an excited state with charge transfer character and an S0 to S1 transition energy of 3.4 eV. Absorption confirms an optical transition in the near UV to violet, while fluorescence spectroscopy shows that photoemission is highly sensitive to solvent polarity. The hydrophobicity of lapatinib leads to fluorescent aggregates in solution, however, binding to the lipid-carrier protein, BSA or to the kinase domain of ERBB2, produces spectroscopically distinct photoemission. Confocal fluorescence microscopy imaging of lapatinib uptake in ERBB2-overexpressing MCF7 and BT474 cells reveals pools of intracellular inhibitor with emission profiles consistent with aggregated lapatinib.

binding-induced,_turn-on_fluorescence_of_the_egfr/erbb_kinase_inhibitor,_lapatinib\xe2\x80\xa0

Introduction<!>Electronic structure of lapatinib<!>Optical spectroscopy<!>Binding and aggregation induced emission<!>Confocal microscopy<!>Conclusions<!>Computational methods<!>Materials<!>Optical spectroscopy<!>Cell culture and confocal microscopy

<p>Fluorescent analogues of biomolecules have attracted considerable attention over the past few decades.1–3 While many biomolecules, such as nucleobases, aromatic amino acids and their metabolites are inherently fluorescent, they require high energy excitation between 250 and 300 nm and emit in the UV or blue edge of the visible spectrum. Fluorescent analogues of many biomolecules, especially nucleobases, have been successfully generated by expanding the existing aromatic framework of the parent molecule thereby lowering the HOMO–LUMO gap and pushing excitation and/or emission to longer wavelengths.</p><p>The presence of aromatic cores in many non-natural ligands, such as pharmaceuticals or drug-like compounds, suggests that these molecules should also behave as fluorescent probes, or can readily be modified to generate fluorescent analogues. We recently reported a family of fluorescent quinazolines (general structure A) designed as mimics of EGFR/ERBB-targeted chemotherapies gefitinib, erlotinib and lapatinib (Fig. 1).4 These kinase inhibitors are employed to treat cancers with deregulated ERBB receptors and represent paradigms for alternative modes of kinase inhibition, i.e. type I, active state, and type II, inactive state, inhibitors. They inhibit, to varying degrees, all three ERBB receptors with robust kinase activity, i.e. EGFR, ERBB2 and ERBB4. This makes them especially useful in conjunction with specific receptor directed antibody regiments, such as Trastuzumab for ERBB2, or following the emergence of resistance against the first line of treatment. By extending the quinazoline core common to gefitinib or erlotinib (Fig. 1), we were able to generate fluorescent mimics with excitation and emission wavelengths in the visible region of the spectrum.4a These probes also exhibited 'turn-on' fluorescence induced by binding to the ERBB2 kinase domain. This ON/OFF emission switching is a direct result of their intramolecular charge transfer (ICT) excited states which leads to quenching in polar environments such as water. When bound in the solvent-excluding and relatively apolar ATP-binding pocket of ERBB2, emission is enhanced.</p><p>The structure of lapatinib is somewhat unique amongst the EGFR/ERBB inhibitors in that it possesses a pendant furan ring at the 6-position of the quinazoline core, required for the stabilizing insertion into a hydrophobic pocket that is specific to the inactive state. This structure suggests that lapatinib should have somewhat longer wavelength excitation energies than other members of the quinazoline class of inhibitors, such as gefitinib and erlotinib; these ligands have only solubilizing ether substituents at the 6-position. Solutions of lapatinib appear yellow, indicative of optical transitions in the violet to blue region of the visible spectrum and, under illumination of a UV-lamp, solutions of lapatinib in organic solvents exhibit blue to green fluorescence. As part of our on-going investigations into fluorescent ligands, we decided to study the basic photophysical properties (e.g. λmax, abs, ɛ, λmax, em, ϕem) of lapatinib to determine if it could function as a competent fluorescent reporter. Our results show that lapatinib possesses an S0–S1 transition in the UV to violet region of the spectrum and exhibits environmentally-sensitive fluorescence in the blue to green portion of the spectrum. Most importantly, emission is dramatically enhanced upon exposure to the lipophilic drug-carrying protein, albumin, as well as the ERBB2 kinase domain, but with distinct spectra for the bound species. We were able to image the uptake and distribution of lapatinib in MCF7 cells; we observed, surprisingly, that lapatinib forms fluorescent aggregates in the cytoplasm, but does not show signs of binding-induced emission in ERBB2, even in the presence of highly overexpressed ERBB2.</p><!><p>We first investigated the electronic structure of lapatinib through TD-DFT calculations at the 6-31G* level using the Coulomb-attenuating method, CAM-B3LYP.5 The geometry of EGFR-bound lapatinib (PDB ID: 1XKK)6 was utilized for the TD-DFT calculations, with the non-conjugated atoms, i.e. the amino-sulfone and benzylic fragments, removed (see ESI† for atomic coordinates). Inspection of the frontier molecular orbitals (FMOs) reveals that the HOMO is distributed over both the furan and quinazoline ring systems while the LUMO is largely concentrated on the quinazoline core (Fig. 2). The FMO distribution suggests that there should be a moderate degree of charge redistribution following the one electron excitation to the S1 state. In the gas phase, two closely spaced transitions are predicted at 330 nm and 300 nm. While these values suggest that excitation would be optimal in the UV, a prominent shoulder extends to longer wavelengths (Fig. 3). The moderately polar environment of the ATP binding pocket may also influence the optical properties of lapatinib, shifting the optical transitions to slightly longer wavelengths. Our previous investigations of solvent-excluding binding sites revealed a polarity close to that of THF7 and when this solvent dielectric is employed in the TD-DFT calculations, the S0–S1 transition is predicted to occur at slightly longer wavelengths (λmax, abs of 347 nm) making excitation with a DAPI filter set or 405 nm diode laser feasible.</p><!><p>In the solution UV-vis spectra of lapatinib (Fig. 3) the longest wavelength absorption appears at 367 nm in methanol and 380 nm in THF, in good agreement with the calculated values. The molar absorptivity in both solvents is moderately high with a value near 30000 M−1 cm−1. The closely spaced transitions predicted by TD-DFT appear to be more clearly separated in the actual spectra, with the second prominent transition at approximately 325 nm for both solvents. Broad shoulders extend from the longer wavelength peaks into the violet portion of the visible spectrum. The molar absorptivity at 405 nm is between 10000 in MeOH and 20000 in THF, suggesting that confocal microscopy using this excitation source is feasible.</p><p>In methanol, the emission of lapatinib is largely quenched, with ϕem of only 2.8 × 10−3. As the excited state is expected to exhibit some charge transfer character based on the quantum chemical calculations, this is not unexpected since methanol is a quite polar solvent with a value of 55.5 kcal mol−1 on Reichart's ET(30) scale.8 In a less polar solvent, such as THF, ET(30) = 37.4 kcal mol−1, the quantum yield is 0.14, an increase of 40-fold compared to methanol.</p><!><p>Solution spectroscopy suggests that lapatinib may exhibit binding-induced emission enhancements. In water (ET(30) = 63.1 kcal mol−1)8 as in MeOH, fluorescence should be quenched, while in the less polar and geometrically defined environment of a binding pocket, emission should be enhanced. We attempted to obtain the quantum yield of lapatinib in both water and phosphate buffered saline (PBS), however, the hydrophobic nature of lapatinib led to substantial aggregation, as previously reported by Owen et al.9 In their studies, use of a surfactant (Tween-80) reduced aggregation and restored the antiproliferative activity of lapatinib and other drugs. In the body, carrier proteins such as albumin or α1-acid glycoprotein bind hydrophobic molecules, helping to eliminate aggregation and distribute lapatinib.10 Therefore, in addition to investigating the optical response of lapatinib upon binding to ERBB2, we also explored the optical properties albumin-bound lapatinib, as well as aggregates formed in aqueous solutions.</p><p>Addition of DMSO-solubilized lapatinib to PBS to produce a clinically-relevant concentration of 3 μM,11 results in the formation of aggregates with a hydrodynamic radius of approximately 150 nm (Fig. 4) as determined by dynamic light scattering (DLS). The aggregates increase in size over a 60 min period until a visible precipitate can be observed. The lapatinib aggregates were found to be emissive, with an emission maximum of 464 nm, but a relatively low quantum yield (Table 1). This behaviour is similar to many other dyes that exhibit aggregation induced emission (AIE) and may be the result of limited rotation between the furan and quinazoline systems. Alternately, molecules present inside the aggregates should experience a less polar environment, which could lessen non-radiative decay pathways of charge transfer excited states. In either case, the presence of fluorescent aggregates could significantly impact solution spectroscopy and fluorescence microscopy, however, the use of Tween-80, as reported by Owen et al.,9 or the presence of a lipid-carrier protein should eliminate this issue.</p><p>Aggregation is indeed eliminated in the presence of 100 μM (6.6 mg ml−1) bovine serum albumin (BSA); DLS reveals a mean particle size of 7 nm that is indistinguishable from BSA alone and the particle size remains unchanged; these solutions are stable with no precipitate observed for at least 48 hours. From this result, we conclude that lapatinib must be bound to BSA and likely resides in one of the hydrophobic pockets of the protein. This notion is supported by fluorescence spectroscopy (10 μM BSA, 1 μM lapatinib) as the emission wavelengths and lifetimes differ significantly from those observed for lapatinib aggregates (Table 1). The emission is blue shifted by 41 nm while the lifetime exhibits a biexponential decay with a very short component of 0.06 ns and a longer component of 1.1 ns.</p><p>We next examined the optical response of lapatinib (200 nM) in solutions of the soluble ERBB2 kinase domain fragment, amino acids Q679-V1255, (500 nM). The emission maximum of lapatinib appears at 445 nm, in between that observed for the aggregates and BSA-bound forms. The emission lifetime measurements reveal two components: the longer component of 2.7 ns (79%) is readily assigned to ERBB2-bound lapatinib, while the shorter and less abundant component (21%) of 0.48 ns can be traced to aggregates of unbound lapatinib.</p><p>Taken together these results demonstrate that lapatinib exhibits a turn-on emission response that can be induced either by binding to a hydrophobic pocket, or through aggregation. The emission maxima progress in approximately 20 nm increments from BSA-bound, to ERBB2-bound to the aggregates. While the emission peaks overlap, the emission lifetimes are distinct and allow for resolution of the bound and unbound populations.</p><!><p>The ERBB2 gene is amplified in approximately 30% of breast cancer patients, leading to a substantial overproduction of the receptor. This results frequently in more than 1 million receptors per cell as opposed to 1000 or less in non-transformed cells. However, protein levels often do not correlate well with gene amplification levels, and the responsiveness to ERBB2 targeted antibodies shows often a poor correlation with detected ERBB2 levels. The gold standard for the determination of ERBB2 status of a cancer remains therefore the very labour-intensive analysis by fluorescence in situ hybridization (FISH).12,13 The rationale for this discrepancy is poorly understood, but one limitation is the presence of ERBB2 species with altered antibody epitopes or truncated species that lack the segments being detected altogether. These truncated species have been shown to be even more tumorgenic than the full-length receptor.14–16 The binding-induced emission of lapatinib might serve as an alternative test that uses a simple and stable fluorescent "stain" to detect the presence of ERBB2 kinase domains, regardless of modifications in the remainder of the protein. We therefore decided to examine the optical response of lapatinib when exposed to ERBB2-overexpressing MCF7 and BT474 cells. As previously reported, aggregation of lapatinib will lower the efficacy in cell culture.9 While their study restored drug efficacy through the use of a surfactant, our optical spectroscopy and DLS studies revealed that serum albumin can also inhibit aggregation of lapatinib, and is arguably more representative of the carrier and exchange behaviour in vivo. MCF7 and BT474 cells used for imaging studies, were cultured in media containing 10% fetal bovine serum, which we found to be sufficient to inhibit aggregate formation. To ensure the absence of aggregates, cell culture media containing lapatinib (3 μM) was allowed to stand for 2 hours before being passed through a 0.200 μm filter. These solutions were added to the cells, which were then imaged either immediately or at 24 and 48 hours, using a 405 nm excitation source.</p><p>15 minutes after addition, lapatinib was clearly visible inside the cells and was widely distributed throughout the cytosol (Fig. 6A). With prolonged exposure (Fig. 6B) very bright punctate structures were also observed within the cytosol. The emission spectrum of lapatinib at early time points was red-shifted relative to the emission observed when bound to ERBB2 or BSA and very closely matches the emission spectrum of aggregates obtained in PBS solutions (Fig. 6C). This suggests that the lapatinib is aggregating rapidly upon entering the cell and the vast majority of the observable lapatinib is not bound in a hydrophobic pocket, which should be distinguishable by bluer emission maxima. The emission spectrum of the larger aggregates observed after 24 h is red-shifted further by approximately 60 nm. This may be the result of electronic coupling between lapatinib molecules, leading to delocalized excited states with lower energy emission.</p><p>Interestingly, no emission was observed at the cell membrane, where ERBB2 is localized. The presence of ERBB2 was confirmed through immunofluorescence (see ESI†), but no lapatinib emission was found to colocalize with the labelled antibodies. Previous studies17,18 have demonstrated the ability to image ON/OFF dynamics of fluorescent probes binding to approximately 50000 membrane proteins. One possible explanation is that the ERBB-bound pool of lapatinib is not adequately excited at 405 nm compared with the intracellular aggregates. We recently investigated18c such a case in which a protein bound form of a dye could be selectively imaged by shifting the excitation source from 458 nm to 405 nm. Based on the excitation spectra (Fig. 5), lapatinib is optimally excited near 370 nm. Therefore, we attempted to image ERBB2-bound lapatinib using an epifluorescent microscope equipped with a UV bandpass filter set, then via a two photon confocal microscope with excitation at 765 nm. In neither case were we able to detect lapatinib emission at the membrane.</p><p>Several factors may contribute to the apparent absence of lapatinib fluorescence despite the abundance of binding sites in these ERBB2-overexpressing cells. First, the presence of chaperone proteins may preclude access to the ATP-binding fold. Specifically the abundant chaperon HSP90 is known to associate constitutively with the ERBB2 kinase domain.19 However, given the high affinity of lapatinib to the intracellular segment of ERBB2, estimated from its Ki of 13 nM,20 any interfering protein complex would have to be of significantly higher effective affinity to cause more than just a time delay in saturation.</p><p>The expectation that cellular ERBB2 should be fully saturated by lapatinib is derived from inhibition data. Inhibition data only state that 100% of the signalling competent receptors can be saturated and blocked with inhibitor, however, this inhibited population may represent only a portion of the total receptor pool. Studies on the state of clustered ERBB2 receptors in overexpressing cancer cells, using gold particle-modified antibodies and electron microscopy,21 found that clusters of ERBB2 only contain a small portion of phosphorylated receptors, suggesting a small pool of catalytically active species. Even under conditions of ligand-induced hetro-dimerization and activation at low receptor levels, only a very small population of receptors takes part in dimerization.22 Thus, one possible explanation for the lack of observable lapatinib fluorescence is that only a small pool of autoactivated receptors may be the target for inhibition and saturation by lapatinib. Alternatively, lapatinib may stabilize an inactive state that is subsequently handed over to other inactive state stabilizing complexes that replaces lapatinib. Many of the factors involved are multimeric or clustered and their true affinities may involve large avidity components and are difficult to evaluate.</p><!><p>We have investigated the photophysical properties of the EGFR/ERBB-targeting kinase inhibitor, lapatinib. Optical spectroscopy and DFT calculations demonstrated that lapatinib functions as an environmentally responsive 'turn-on' fluorophore. This allows the detection of aggregates and protein-bound drug through spectroscopically distinct emission. One possible application of this information is the monitoring of dose and time dependent blood concentrations. The lack of an ERBB2-specific signal precludes the use of lapatinib fluorescence to differentiate HER2-positive tumors from other cancer types, however, using the inherent fluorescence of lapatinib, we were able to image the accumulation, distribution and aggregation of the drug within cells. It is not clear if aggregates are present in vivo and if they may possibly serve as a reservoir of drug that may act over longer time periods.</p><!><p>The transition energies and frontier molecular orbitals of the aromatic core of lapatinib (see ESI† for atoms and coordinates) were calculated by TD-DFT using the CAM-B3LYP functional at the 6-31G(d) level using Gaussian'09.23 The geometries utilized in the TD-DFT calculations were either calculated for the ground state in the gas phase or obtained from the crystal structure of ERBB2-bound lapatinib (PDB ID: 1XKK).6</p><!><p>Inhibitors and bioreagents were obtained from commercial suppliers: lapatinib ditosylate (Selleck Chemicals); canertinib and adenosine 5′-(β, γ-imido) triphosphate (Sigma-Aldrich); ERBB2 kinase domain fragment Q679-V1255 (Biaffin); BSA and mouse anti-ERBB2 (Calbiochem); Alexa Fluor546 goat antimouse (Life Technologies).</p><!><p>UV-vis absorption spectra were obtained on a Lambda 35 UV-vis (Perkin-Elmer) spectrometer with a path length of 1 cm. Fluorescence studies were performed on a LS55 Fluorometer (Perkin-Elmer) with a path length of either 1 cm or 3 mm. For determination of Φem, solutions were prepared to an optical density of less than 0.05 in order to minimize inner filter effects. Perylene in cyclohexane was used as a reference for quantum yields.24 Emission lifetime measurements were performed on an EasyLife II (Photon Technology International) using a 340 nm pulsed LED.</p><!><p>MCF7 and BT474 cells were cultured as previously described in sterile T-75 flasks.22 Cells were maintained in RPMI (Cellgro) containing 10% dialyzed FBS (Atlanta Biologicals), penicillin (100 units mL−1) and streptomycin (0.01%) solution (Cellgro) under a humidified 5% CO2 atmosphere. For imaging, cells were seeded at a density of 105 cell cm−2 in 96 microwell plates or on glass coverslips. Cells maintained a normal morphology during the course of the experiments (maximum of 1.5 h) and remained adhered to the imaging plate or coverslip.</p><p>Single photon imaging was performed on a Leica SP5 confocal microscope housed within the UM Biology Imaging Core Facility using 405 nm excitation. Two-photon imaging was performed on a Leica MP/SP5 confocal microscope using a Sapphire multiphoton laser tuned to 765 nm.</p>

PubMed Author Manuscript

Functional expression of the eukaryotic proton pump rhodopsin OmR2 in Escherichia coli and its photochemical characterization

Microbial rhodopsins are photoswitchable seven-transmembrane proteins that are widely distributed in three domains of life, archaea, bacteria and eukarya. Rhodopsins allow the transport of protons outwardly across the membrane and are indispensable for light-energy conversion in microorganisms. Archaeal and bacterial proton pump rhodopsins have been characterized using an Escherichia coli expression system because that enables the rapid production of large amounts of recombinant proteins, whereas no success has been reported for eukaryotic rhodopsins. Here, we report a phylogenetically distinct eukaryotic rhodopsin from the dinoflagellate Oxyrrhis marina (O. marina rhodopsin-2, OmR2) that can be expressed in E. coli cells. E. coli cells harboring the OmR2 gene showed an outward proton-pumping activity, indicating its functional expression. Spectroscopic characterization of the purified OmR2 protein revealed several features as follows: (1) an absorption maximum at 533 nm with all-trans retinal chromophore, (2) the possession of the deprotonated counterion (pK a = 3.0) of the protonated Schiff base and (3) a rapid photocycle through several distinct photointermediates. Those features are similar to those of known eukaryotic proton pump rhodopsins. Our successful characterization of OmR2 expressed in E. coli cells could build a basis for understanding and utilizing eukaryotic rhodopsins.

functional_expression_of_the_eukaryotic_proton_pump_rhodopsin_omr2_in_escherichia_coli_and_its_photo

<!>Results and discussion<!>Methods<!>Gene preparation, protein expression and purification of HEK293T cells. For protein expression<!>Gene preparation, protein expression and ion transport measurements of E. coli cells.

<p>To capture sunlight, organisms use a variety of photoreceptive proteins that are responsible for light-energy conversion and light-signal transduction in nature. Photoreceptive membrane proteins called microbial rhodopsins form large phylogenetic clusters in three domains of life, archaea, bacteria and eukarya 1,2 . Microbial rhodopsins consist of seven-transmembrane α-helices covalently linked to the chromophore all-trans retinal, a derivative of vitamin-A 2,3 . The chromophore retinal covalently binds to a conserved Lys residue located in the 7th helix of the rhodopsin apoprotein through a protonated Schiff base linkage 3 . After photoisomerization of the retinal from an all-trans to a 13-cis configuration, microbial rhodopsins undergo a series of cyclic reactions called a photocycle in which several spectrally distinct photointermediates are sequentially formed and the initial state is recovered along with the conformational changes 3 . During each photocycle, rhodopsins exhibit their biological functions such as ion transport and photosensing [1][2][3] . For instance, outward proton pumps produce the molecular currency adenosine triphosphate (ATP) through the formation of a proton gradient across the cell membrane like photosynthesis, indicating their physiological significance in microorganisms 4 .</p><p>Archaeal and bacterial rhodopsins have been the basis for research of microbial rhodopsins. Historically, bacteriorhodopsin (BR) was first rhodopsin discovered from the halophilic archaea Halobacterium salinarum as an outward proton pump in 1971 5 . After that, halorhodopsin and sensory rhodopsin I and II were identified from the halophilic archaea as an inward chloride pump and a phototaxis sensor, respectively 1 . Many other archaeal OPEN 1 Division of Pharmaceutical Sciences, Okayama University, Okayama 700-8530, Japan. 2 Graduate School of Medicine, Dentistry and Pharmaceutical Sciences, Okayama University, Okayama 700-8530, Japan. 3 Atmosphere and Ocean Research Institute, The University of Tokyo, Chiba 277-8564, Japan. 4 Faculty of Advanced Life Science, Hokkaido University, Sapporo 060-0810, Japan. 5 Global Station for Soft Matter, GI-CoRE, Hokkaido University, Sapporo 001-0021, Japan. 6 These authors contributed equally: Masuzu Kikuchi and Keiichi Kojima. * email: sudo@ okayama-u.ac.jp channelrhodopsins are used to induce neural activation, outward proton pump rhodopsins and anion channelrhodopsins are used to induce neural silencing 21,22 . Therefore, the molecular characterization of eukaryotic ion transporting rhodopsins should provide useful information for their modification and development as new optogenetics tools.</p><p>Although eukaryotic rhodopsins can be characterized by heterologous expression systems using yeast cells, Xenopus oocytes, mammalian cells, insect cells and cell-free systems [23][24][25] , it is generally difficult to achieve their functional expression in E. coli cells. Nonetheless, we previously succeeded in expressing CrChR1 in E. coli cells by truncating the N-and C-termini of the proteins, although the truncated mutants showed constitutive activities that are different from the wild-type proteins 26 . We also succeeded in the functional expression and mutational analysis of GtACR2 in E. coli cells, which could provide a characteristic mutant for new optogenetics tools [27][28][29] . However, we were not able to purify the photoactive protein of GtACR2 from E. coli cells probably due to its denaturation during the solubilization step in detergent micelles. One research group fused the Mistic domain, a membrane-associated protein from Bacillus subtilis, into the N-and C-termini of ARI and CrChR1, to allow their functional expression in E. coli cells and successfully characterized the photochemical properties of the purified proteins 30 . However, the yields of the purified proteins were 0.12 mg for Mistic-fused ARI and 0.04 mg for Mistic-fused CrChR1 per liter of culture medium, which is more than tenfold lower than those of bacterial rhodopsins (2 and 5 mg for TR and Salinibacter ruber sensory rhodopsin I, respectively) 9,15 .</p><p>While organisms usually have several rhodopsin genes in their genome, the eukaryotic dinoflagellate O. marina uniquely shows more than 10 putative rhodopsin genes (Fig. 1B) [31][32][33] . O. marina is a heterotrophic dinoflagellate that is widely distributed on earth 34 . Noteworthy, O. marina shows several important characteristics as follows: (1) it can be isolated from the environment and easily cultured in medium, (2) genetic approaches are available, and (3) it is inexpensive to obtain, maintain and is practical to use. Therefore, O. marina has been widely used as a model for dinoflagellates for over 100 years in various scientific fields including phylogeny, biogeography and ecology 35 . By employing those characteristics, it has been reported that putative rhodopsin genes can be expressed as transcripts and proteins in O. marina, which suggests that those genes encode functional proteins 31,32 . Hartz et al. reported that O. marina can orient to light based on rhodopsins and may use that photosensory response to detect algal prey based on chlorophyll autofluorescence 36 . Based on that background, we assume that O. marina rhodopsins (OmRs) would be good candidates for the functional expression and analysis of eukaryotic rhodopsins. So far, one OmR, OmR1 (Genbank accession number: ABV22426) has been characterized by a heterologous expression system in yeast, but not by the E. coli cell expression system 37 . Thus, for expression in E. coli cells, we focused on the other gene named O. marina rhodopsin-2 (OmR2) (Genbank accession number: AIN36546). OmR2 contains several amino acids that are responsible for its outward proton pump functions such as Asp88, Thr92, Asp99 and Asp209, which correspond to Asp139, Thr143, Asp150 and Asp266 in LR, respectively (Fig. 1B and Fig. S1). Thus, OmR2 should work as an outward proton pump. It is noteworthy that OmR2 is phylogenetically distinct from OmR1 and other eukaryotic rhodopsins (Fig. 1A), where the amino acid identities and similarities of OmR2 with LR (15.8% identity, 32.1% similarity) and OmR1 (17.8% identity, 39.4% similarity) are relatively lower than those between other characterized eukaryotic rhodopsins, suggesting its phylogenetically distinct feature.</p><p>In this study, we characterize the function and molecular properties of OmR2 as a new model for eukaryotic rhodopsins. We show that OmR2 can functionally work in E. coli cells as a recombinant protein with a lightdriven outward proton pump activity, which is confirmed to be consistent with the electrophysiological results obtained with the mammalian expression system. By taking advantage of the E. coli cell expression system, we obtained highly purified photoactive proteins and successfully performed spectroscopic analyses. The results indicate that OmR2 has similar photochemical properties to the well-characterized eukaryotic proton pump rhodopsins, which suggests that OmR2 can be a model for eukaryotic proton pump rhodopsins. Thus, the successful expression, purification and characterization of OmR2 in E. coli cells could build a basis towards understanding the molecular mechanism of eukaryotic rhodopsins.</p><!><p>Absorption spectrum and electrophysiological experiments of OmR2 in mammalian cells. To investigate whether the OmR2 gene encodes a photosensitive protein, we first expressed and purified its recombinant protein using the HEK293 cell expression system, which has been utilized for the functional expression of several eukaryotic rhodopsins from microbes and animal rhodopsins 29,38 . The purified OmR2 protein in detergent DDM micelles was colored purple and its absorption spectrum showed an absorption peak at 533 nm (Fig. 2A), indicating that OmR2 works as a green light sensitive protein.</p><p>We then performed an electrophysiological study to investigate the function of OmR2. For expression in ND7/23 cells, the cDNA for OmR2 and EYFP were inserted downstream of the CMV promoter with the trafficking signal (TS) and endoplasmic reticulum export signal (ER) to enhance the membrane localization. The corresponding gene constructs have been utilized for functional expression in mammalian cells for various kinds of microbial rhodopsins such as sodium pump rhodopsins and halorhodopsin 39,40 . For transfected ND7/23 cells, the yellow fluorescence from EYFP was observed especially at the plasma membrane of the cells (Fig. 2B). This indicates the successful expression and localization of OmR2 in the plasma membrane. We then performed electrophysiological analysis to confirm the ion transport activity of OmR2. A positive photocurrent upon illumination was observed under conditions where the extracellular and intracellular pH were 7.4 and 7.3, respectively, and the holding membrane potential was 0 mV (Fig. 2C). We also measured the peak currents at membrane potentials from − 60 to 60 mV to obtain the current-voltage relationship (I-V curve) (Fig. 2D). The positive peaks were kept at all potentials, which suggests that OmR2 works as an outward cation pump or an inward anion pump. To further identify the substrate ion of OmR2, we measured the photocurrents under www.nature.com/scientificreports/ different extracellular ion compositions. It was first observed that the higher the extracellular pH was set, with an intracellular pH of 7.3, the higher the positive peaks were at all membrane potentials (Fig. 2D). Thus, the amplitudes of peak currents were sensitive to the extracellular pH values. On the other hand, when NaCl in the extracellular solution was replaced by CsCl, KCl or sodium gluconate, no significant change in the amplitudes was observed (Fig. 2E). These results indicate that OmR2 works as a light-driven outward proton pump. The peak photocurrent (~ 100 pA) was comparable to that of archaerhodopsin-3, AR3 (~ 100 pA) 41 and a sodium pump rhodopsin KR2 (~ 100 pA) 40 . The successful expression and robust outward photocurrents of OmR2 suggest its applicability as a neural silencing tool for optogenetics similar to AR3 and KR2 22,40 .</p><p>Functional expression of OmR2 in E. coli cells. So far, the E. coli cell expression system has been widely utilized for various archaeal and bacterial rhodopsins [9][10][11][12][13]39,41,42 . Since OmR2 is phylogenetically distinct from the other characterized eukaryotic rhodopsins, we sought to express the OmR2 recombinant protein using the E. coli cell expression system. We cultured E. coli BL21(DE3) cells harboring expression plasmids of OmR2. To prove that the OmR2 protein works as a photoactive protein, its light-dependent ion transport activity was observed as light-induced pH changes of a suspension of E. coli cells (Fig. 3). Illumination induced a pH decrease in the cell suspension, which would reflect the outward proton movement across the membrane while no pH decrease was observed in E. coli cells harboring the empty vector without the OmR2 gene (Fig. 3). The pH change disappeared in the presence of a proton-selective ionophore, CCCP, which works to collapse the proton motive force across the membrane. These results indicate that OmR2 has a light-dependent outward proton transport activity.</p><p>In other words, OmR2 works as an outward proton pump in E. coli cells, which is consistent with the electrophysiological results for OmR2 expressed in ND7/23 cells (Fig. 2). Therefore, we concluded that the functional expression of OmR2 can be realized as a recombinant protein in E. coli cells. As far as we know, there has been www.nature.com/scientificreports/ no report that successfully obtained recombinant proteins of eukaryotic rhodopsins in E. coli cells, except for a few examples 26,27,30 .</p><p>To discuss what is the factor required for the successful expression of OmR2 in E. coli cells, we compared the amino acid sequences of OmR2 and other typical eukaryotic rhodopsins (Fig. 1). Among the characterized eukaryotic rhodopsins, NR is the phylogenetically closest to OmR2 but cannot be functionally expressed in E. coli cells in amounts sufficient for analysis 43 . It should be noted that the amino acid identity and similarity between them are 19.8 and 41.3%, respectively, and therefore it is difficult to identify which element is essential for the functional expression of OmR2 in E. coli cells from the comparison between them. In addition to OmR2, uncharacterized and phylogenetically distinct microbial rhodopsins have been continuously identified from various eukaryotes. From comprehensive comparisons of amino acid sequences among these possible molecules that can be functionally expressed in E. coli cells, it may be possible to identify which region in eukaryotic rhodopsins is responsible for the functional expression in E. coli cells. To prove this concept, we will identify other molecules that can be functionally expressed in E. coli cells from a comprehensive expression analysis of eukaryotic rhodopsins. The region(s) conserved among them will then be introduced into molecules that could not be expressed in E. coli cells, such as NR, LR and OmR1. That approach should lead to the identification of underlying element(s) in eukaryotic rhodopsins that allow the successful functional expression in E. coli cells.</p><p>In addition to the sequence information, it is known that post-translational modifications (PTMs), such as glycosylation and disulfide bond formation, play important roles in the functional and structural maintenance of membrane proteins while many types of PTMs are deficient in bacteria 44 . We speculate that eukaryotic rhodopsins can be expressed in mammalian cells partially due to PTMs. The PTMs of OmR2 are still unclear, and therefore, further investigation is required in the future.</p><p>Purification and photochemical properties of OmR2. As described in "Methods", the E. coli cells were solubilized in DDM, after which the solubilized OmR2 proteins were purified by Ni-affinity column chromatography and had a purple color similar to the purified OmR2 expressed in HEK293 cells (Figs. 2A and 4A). The absorption spectrum of the purified OmR2 in E. coli cells showed an absorption peak at 533 nm (Fig. 4A), which is consistent with the result from the purified OmR2 in HEK293 cells (Fig. 2A). Thus, OmR2 obtained from E. coli cells forms a photoactive pigment in the detergent micelles without significant denaturation. As far as we know, this is the first demonstration where a eukaryotic rhodopsin was purified in detergent micelles using the E. coli cell expression system, except for the previous report of Mistic-fused ARI and CrChR1 proteins 30 . The yield of purified OmR2 protein was 1 mg per liter of culture medium, which is more than ninefold higher than the yields of Mistic-fused ARI (0.12 mg) and Mistic-fused CrChR1 (0.04 mg) and is comparable to the yield of TR (2 mg) 9 . The absorption of OmR2 at ~ 280 nm represents the absorption of aromatic residues such as Trp and Tyr. From the ratio of absorbance at 280 and 533 nm with the molecular coefficient of microbial rhodopsins (~ 50,000 cm −1 M −1 ), we roughly estimated the purity of the sample as 67.5%.</p><p>Using the purified proteins from E. coli cells, we performed the photochemical characterization of OmR2. We first performed HPLC analysis to determine the retinal configuration (Fig. 4B). The HPLC patterns of the retinal oxime isomers in the dark-and light-adapted states predominantly exhibited the peaks of all-trans isomers. The ratios of all-trans isomers were estimated to be 95 and 99%, respectively, in the dark-and lightadapted states by calculating the area under the peaks considering the molecular coefficient of each isomer as previously described 14,45 . This indicates that OmR2 possesses all-trans retinal regardless of the light environment and functions with all-trans retinal. It is generally known that archaeal proton pump rhodopsins, such as BR and AR3, possess both all-trans and 13-cis retinals whose ratio is dynamically changed according to the light environment 46,47 . On the other hand, bacterial and eukaryotic proton pump rhodopsins, such as PR, TR, LR and ARII, possess all-trans retinal predominantly both in dark-and in light-adapted conditions 9 . Similarly, OmR2 was found to possess all-trans retinal predominantly regardless of the light environment (Table 1). The proton pump rhodopsins are known to transfer the substrate protons through some proton acceptable charged residues such as Asp and Glu inside the proteins via the Grotthuss mechanism 53 . The protonated Schiff base of the retinal chromophore is stabilized by an aspartic acid as a counterion, which accepts the substrate proton from the Schiff base during the photocycle for the proton pumping function. To estimate the pK a value of the counterion of OmR2, we measured its spectral changes upon acidification. When the pH values decreased from 7.1 to 1.7, a large spectral red-shift from 533 to 549 nm was observed (Fig. 5A). Interestingly, when the pH value further decreased to 0.98, OmR2 showed a small spectral blue-shift from 549 to 548 nm. The spectral red-shift can be interpreted as the protonation of the counterion leading to a decrease in the energy gap between the electronic ground and excited state, which is commonly observed in outward proton pump rhodopsins 9,54 . Judging from the sequence alignment between OmR2 and other proton pump rhodopsins, Asp88 (Asp 85 in BR) is assigned to the putative counterion in OmR2 (Fig. 1B and Fig. S1). Since a spectral blue-shift is observed at low pH in BR and is assigned to the protonation of Asp212 as a secondary counterion 55 , the observed small spectral blue-shift at the extremely low pH would be the protonation of Asp209 (corresponding to Asp212 in BR) as a putative secondary counterion in OmR2 (Fig. 1B). The difference spectra showed the increase in absorbance at 591 nm and the concomitant decrease in absorbance at 516 nm. These difference spectra did not show an isosbestic point (Fig. 5B), indicating that the process of spectral changes reflects the transition between more than three states, probably the deprotonated and protonated states of the putative counterions (Asp88 and Asp209). The plots of the difference absorbance at 591 and 516 nm against the acidic pH values (Fig. 5C) were well fitted using the Henderson-Hasselbalch equation assuming two pK a values. From the fitting analysis, the pK a values for the spectral red-shift and blue-shift were estimated to be 3.0 ± 0.04 and 1.5 ± 0.22, respectively. Since both spectral shifts correspond to the protonation process of Asp88 and Asp209, we estimated the pK a of the putative counterion Asp88 in OmR2 as 3.0 ± 0.04, and the value of the secondary putative counterion Asp209 as 1.5 ± 0.22. The pK a of the counterion in ARII was estimated as 2.6, which is a value similar to that of OmR2 (Table 1). To determine the counterion residues of OmR2 and further clarify this issue, mutational analysis of Asp88 and Asp209 will be required as future work. Although we attempted to estimate the pK a of the protonated Schiff base of OmR2 by measuring its spectral changes upon alkalinization, the protein denaturation under alkaline conditions (> ~ pH 10.5) made the estimation impossible.</p><p>To analyze the photocycle of OmR2, we then performed flash-photolysis experiments. Figure 6A shows the flash-induced difference spectra over the spectral range of 380-710 nm. The depletion and recovery of absorbance at ~ 540 nm correspond to the bleaching of the original state, while an increase and decrease of absorbance at ~ 400 and 600 nm were characteristically observed. Figure 6B shows the time courses of the difference absorbance changes at the three wavelengths of 400, 540 and 600 nm. Following the illumination, an absorption increase at ~ 600 nm was observed together with the depletion of the original state. An absorption increase at ~ 400 nm was then observed with a concomitant absorption decrease at ~ 600 nm within 0.1 ms. Considering the temporal and spectral ranges of the absorption changes, the absorbances at 600 and 400 nm were tentatively attributed to the K-and M-intermediates, respectively. The absorbance at ~ 400 nm decreased with the concomitant absorbance increase at ~ 600 nm, which was tentatively assigned as the O-intermediate, within 50 ms. Finally, the absorbance at ~ 600 nm was depleted with recovery of the original state within 1 s. Thus, after the light absorption, OmR2 sequentially forms K-, M-and O-intermediates, and then returns to the original state. To estimate the decay time constants of the intermediates, the temporal absorption changes at 400, 540 and 600 nm were fitted with a triple-exponential function assuming the irreversible sequential model. The decay time constants of the K-, Mand O-intermediates were estimated as 0.015, 6.4 and 30 ms, respectively. Finally, we investigated how proton uptake and release happen during the photocycle since OmR2 exhibits a proton pumping function. We measured the flash-induced absorption change at 450 nm of a pH-sensitive fluorochrome, pyranine, which reflects the solvent pH changes as previously described 14,42 . As a result, the absorbance of pyranine decreased within 10 ms and then increased within 100 ms. The time constants of the absorbance decrease and increase processes were estimated as 1.5 and 47 ms, respectively, which were consistent with the formation and decay time constants 6B). This suggests that the substrate proton was first released from OmR2 upon the O-formation and then taken up from the bulk solution upon the O-decay during the photocycle. Based on the above results, we propose a photocycle model of OmR2 as shown schematically in Fig. 6C. It is generally known that proton pump rhodopsins carry one substrate proton per one photocycle. That means that the strength of proton pumping activities would be proportional to the period of the photocycle, which is predominantly determined by the decay time of the late intermediate such as the O-intermediate. As the decay rate of the O-intermediate of OmR2 was relatively fast (0.03 ms −1 ) compared with AR3 and LR (0.03 and 0.05 ms −1 , respectively) (Table 1), OmR2 can be thought to show an efficient proton pumping activity, which was demonstrated by the electrophysiological analysis (Fig. 2). As future work to clarify the photocycle model of OmR2, structural investigations of each photointermediate, such as vibrational spectroscopic analysis and X-ray crystallographic analysis, will be necessary.</p><p>Comparison of the photochemical properties of OmR2 with other proton pump rhodopsins. The photochemical properties of OmR2 are listed with the well-characterized proton pump rhodopsins, OmR1, LR and ARII as eukaryotic rhodopsins, BR and AR3 as archaeal rhodopsins, and PR and TR as bacterial rhodopsins (Table 1). OmR2 possesses a deprotonated counterion (presumably Asp88 with the pK a value of 3.0) of the protonated Schiff base (Lys213) in the unphotolyzed state and shows an absorption maximum at 533 nm with the all-trans retinal chromophore. These properties are similar to those of OmR1, LR and ARII. During the photocycle, OmR2 sequentially forms the red-shifted K-intermediate, the blue-shifted M-intermediate and the red-shifted O-intermediate, whose decay rates are similar to those of OmR1, LR and ARII. The above comparisons suggest that OmR2 exhibits the typical molecular and photochemical properties present in eukaryotic proton pump rhodopsins.</p><p>OmR2 contains the acidic amino acid residues that are key residues for function in outward proton pump rhodopsins (Fig. 1B and Fig. S1). Asp88 corresponds to the counterion (Asp85 in BR) that works as a proton acceptor from the Schiff base. Additionally, Asp99 corresponds to a proton donor (Asp96 in BR) and Glu191 and Glu201 correspond to the proton releasing group (Glu194 and Glu204 in BR). From the analogy with BR 3 , we propose the proton movement during the photocycle in OmR2 as follows: (i) the proton of the Schiff base is transferred to the counterion Asp88 during M-formation, (ii) the proton is released from the counterion to the extracellular side through the putative proton releasing group (Glu191 and Glu201), and simultaneously the proton of Asp99 is transferred to the Schiff base during M-decay, and (iii) the proton is taken up from the intracellular side to Asp99 during O-decay (Fig. 6D). Noteworthy, our results indicate that the proton release and uptake process correspond to the formation and decay of the O-intermediate, suggesting the structural importance of the O-intermediate of OmR2. The proton uptake from the intracellular side to the proton donor residue (Asp96 in BR) is generally thought to disrupt the hydrogen-bonding network between the proton donor residue and the protonated Schiff base that triggers reisomerization of retinal from the 13-cis to the all-trans form 56 . In fact, a proton is taken up during O-formation in BR and ARII 3,48 . In contrast, a proton is taken up during "O-decay" in OmR2, suggesting its structural difference of the O-intermediate. The structural features of the O-intermediate should be analyzed by structural and vibrational spectroscopic measurements to further understand of the detailed proton pumping mechanism in the future. By taking advantage of the successful expression of OmR2 in E. coli cells, OmR2 will be a good model to analyze the functional mechanism of eukaryotic rhodopsins using structural and spectroscopic measurements in the future.</p><!><p>Construction of the phylogenetic tree of microbial rhodopsins. The protein sequences of eukaryotic rhodopsins, which were previously reported as putative proton pump rhodopsins, and rhodopsins from O. marina were obtained from the Genbank database. The protein sequences for Oxyrrhis marina rhodopsin-1 (OmR1, Genbank accession number; ABV22426), Oxyrrhis marina rhodopsin-2 (OmR2, AIN36546), Oxyrrhis marina rhodopsins (ADY17806, ADY17809, ABV22427, ABV22430, ABV22432, AIN36547, AIN36548, AIN36549) (Fig. 1B), Leptosphaeria maculans rhodopsin (LR, AAG01180) 24 , Phaeosphaeria nodorum rhodopsin-1 (SNOG_00807) 57 , Phaeosphaeria nodorum rhodopsin-2 (SNOG_00341) 57 , Bipolaris oryzae rhodopsin-1 (AB489199) 58 , Bipolaris oryzae rhodopsin-2 (AB489200) 58 , Sclerotinia sclerotiorum rhodopsin-1 (XP_001597420) 59 , Sclerotinia sclerotiorum rhodopsin-2 (XP_001594532) 59 , Botrytis cinerea rhodopsin (BC1G_02456) 60 , Aureobasidium pullulans rhodopsin (KEQ87154) 61 , Acetabularia acetabulum rhodopsin I (ARI, AEF12206) 62 , Acetabularia acetabulum rhodopsin II (ARII, AEF12207) 48 , Chlorella vulgaris rhodopsin (JQ255360) 21 , Coccomyxa subellipsoidea rhodopsin (CsR, XP_005646688) 63 , Neurospora crassa rhodopsin (NR, AAD45253) 17 , Fusarium fujikuroi rhodopsin (OpsA, CAR82401) 64 , Fusarium fujikuroi rhodopsin (CarO, CAD97459) 64 , Pseudo-nitzschia granii rhodopsin (AJA37445) 65 , Prorocentrum donghaiense rhodopsin (KM282617) 66 , Pyrocystis lunula rhodopsin (AF508258) 67 , and Cyanophora paradoxa rhodopsin (ACV05065) 68 were aligned using the MUSCLE algorithm in MEGA-X software (https:// www. megas oftwa re. net/). The phylogenetic tree was inferred using the maximum likelihood method of MEGA-X software. At this time, the substitution model was selected as the LG model, a discrete Gamma distribution was used to model evolutionary rate differences among sites (five categories, + G parameter = 2.32), and the rate variation model allowed for sites to be evolutionarily invariable (1.94% sites). The bootstrap values were given by 100 iterations of the bootstrap test. www.nature.com/scientificreports/ Toru Ishizuka) as previously described 39,40 . In short, enhanced yellow fluorescent protein (EYFP) was fused to the C-terminus of OmR2 as a reporter. Also, EYFP was flanked with a membrane trafficking signal (TS) at the N-terminus and an endoplasmic reticulum export signal (ER) at the C-terminus to improve its expression and plasma membrane localization. The TS and ER signals were "KSRITSEGEYIPLDQIDINV" and "FCYENEV", respectively, derived from the Kir2.1 potassium channel 69 . Furthermore, the WPRE (Woodchuck hepatitis virus Post-transcriptional Regulatory Element) sequence was inserted to stabilize the transcribed mRNA and increase the amount of translated protein. The expression vector encoding OmR2 was prepared with an In-Fusion cloning Kit (Takara Bio, Japan) according to the manufacturer's instructions as previously described 28,70 . Electrophysiological measurements were carried out at room temperature (20-25 °C) with ND7/23 cells. ND7/23 cells were cultured in Dulbecco's Modified Eagle Medium (Gibco, DMEM/F12, Thermo Fisher Scientific Life Sciences, USA) supplemented with 10% fetal bovine serum, 0.0625% (w/v) penicillin and 0.01% (w/v) streptomycin under a 5% CO 2 atmosphere at 37 °C. The expression plasmid was transiently transfected into cells using the calcium-phosphate method 45 . After 6 h incubation of the transfected cells, all-trans retinal (final concentration = 1 µM) was added into the medium. Electrophysiological experiments were conducted 48-60 h after transfection. Transfected cells were identified by the presence of EYFP fluorescence. The fluorescence signals for EYFP were observed using an IX71 inverted microscope (Olympus, Japan) with a fluorescence mirror unit (U-MYFPHQ, Olympus) and a mercury lamp (U-LH100HGAPO, Olympus). Photocurrents were measured using an EPC 10 USB computer-controlled Patch Clamp Amplifier (HEKA Elektronik, Germany) under a whole-cell patch clamp configuration. The data were analyzed with Patch master software (HEKA Elektronik, Germany). The internal pipette solution for whole-cell voltage clamp recordings from ND7/23 cells contained 50 mM HEPES, 140 mM CsCl, 3 mM MgCl 2 , 5 mM Na 2 EGTA and 2.5 mM MgATP, adjusted to pH 7.3 with CsOH. The cells were continuously superfused by an extracellular medium (10 mM HEPES, 138 mM NaCl, 3 mM KCl, 1 mM MgCl 2 , 2 mM CaCl 2 , 0.1 M glucose, adjusted to pH 9.0, 7.4 and 5.0 with NaOH or HCl). To investigate pump activity in the absence of extracellular chloride, sodium and the presence of potassium, NaCl was also substituted by the same amount of sodium gluconate, CsCl, KCl. Current traces were recorded at − 60, − 40, − 20, 0, 20, 40 and 60 mV. The cells were illuminated with a white LED (THORLABS, USA) through a band-pass filter (520 ± 10 nm), where the light intensity was adjusted to 0.98 mW mm −2 .</p><!><p>and purification, the human codon-optimized OmR2 gene was inserted into a CAG promoter-based mammalian expression vector, pCAGGS, as previously described 29 . A hexa-histidine tag was fused to the C-terminus of OmR2. The expression vector encoding OmR2 was prepared using an In-Fusion cloning Kit. For protein expression in HEK293T cells, the plasmid was transfected into the cells using the calcium-phosphate method 29,38 . After 1 day incubation, all-trans-retinal (final concentration = 5 µM) was added to the transfected cells 29,38 . After another day of incubation, the cells were collected by centrifugation and were then solubilized with 1.0% (w/v) n-dodecyl-β-d-maltoside (DDM, DOJINDO Laboratories, Japan). The solubilized fraction was purified by Ni 2+ affinity column chromatography with a linear gradient of imidazole as described previously 29 . The purified protein was concentrated by centrifugation using an Amicon Ultra filter (30,000 M w cut-off; Millipore, USA). The sample was then loaded into and eluted from a PD-10 column (GE-Healthcare, UK) with Buffer A (50 mM Tris-HCl, pH 7.0, 1 M NaCl and 0.05% (w/v) DDM).</p><!><p>The full-length cDNA for OmR2, whose codons were optimized for E. coli codon usage, were chemically synthesized by Eurofins Genomics and inserted into the NdeI-XhoI site of the pET21a(+) vector as previously described 14 . A hexa-histidine-tag was fused at the C-terminus of OmR2, which was utilized for purification of the expressed protein. The procedures for protein expression were essentially the same as previously described 14,42 . E. coli BL21(DE3) cells harboring the cognate plasmid were grown at 37 °C in LB medium supplemented with ampicillin (final concentration = 50 µg mL −1 ). Protein expression was induced at an optical density at 600 nm (OD 600 ) of 0.8-1.2 with 1 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) and 10 μM all-trans retinal, after which the cells were incubated at 37 °C for 3 h. The proton transport activity of OmR2 was measured as light-induced pH changes of suspensions of E. coli cells as previously described 14,42 . In short, cells expressing OmR2 were washed more than three times in 150 mM NaCl and were then resuspended in the same solution for measurements. Each cell suspension was placed in the dark for several min and then illuminated using a 300 W Xenon lamp (ca. 30 mW cm −2 , MAX-303, Asahi spectra, Japan) through a > 460 nm long-pass filter (Y48, HOYA, Japan) for 3 min. Measurements were repeated under the same conditions after addition of the protonophore carbonyl cyanide m-chlorophenylhydrazone (CCCP) (final concentration = 10 μM). Light-induced pH changes were monitored using a Horiba F-72 pH meter. All measurements were conducted at 25 °C using a thermostat (Eyela NCB-1200, Tokyo Rikakikai Co. Ltd, Japan).</p><p>Purification of OmR2 from E. coli cells and spectroscopic measurements of the purified protein. Escherichia coli cells expressing OmR2 were disrupted by sonication for 30 min in ice-cold water in Buffer B containing 50 mM Tris-HCl (pH 7.0) and 300 mM NaCl. The crude membrane fraction was collected by ultracentrifugation and solubilized with 1.0% (w/v) DDM. The solubilized fraction was purified by Ni 2+ affinity column chromatography with a linear gradient of imidazole as described previously 14,42 . The purified protein was concentrated by centrifugation using an Amicon Ultra filter (30,000 M w cut-off; Millipore, USA). The sample media was then replaced with Buffer A by ultrafiltration for 3-times.</p><p>Absorption spectra of purified proteins were recorded using a UV-2450 spectrophotometer (Shimadzu, Japan) at room temperature in Buffer A. The retinal composition in OmR2 was analyzed by high-performance liquid chromatography (HPLC) as described previously 14 . For light-adaptation, the samples were illuminated for 3 min at 530 ± 10 nm, where the light power was adjusted to ~ 10 mW cm −2 . The molar compositions of the retinal isomers were calculated from the areas of the peaks in HPLC patterns monitored at 360 nm using the extinction coefficients of retinal oxime isomers as described previously 14,45 . For pH titration experiments, the samples were suspended in Buffer A. The pH values of the samples were adjusted to the desired acidic values by adding HCl, after which the absorption spectra were measured at each pH value. All measurements were conducted at room temperature (approx. 25 °C) under room light. After the measurements, the reversibility of the spectral changes was checked to confirm that the sample was not denatured during the measurements. The absorption changes at specific wavelengths were plotted against pH values and the plots were fitted to the Henderson-Hasselbalch equation assuming double pK a values as previously described 14 .</p><p>Transient time-resolved absorption spectra of the purified proteins from 380 to 700 nm at 5 nm intervals were obtained using a homemade computer-controlled flash photolysis system equipped with an Nd:YAG laser as an actinic light source 14,42 . By using an optical parametric oscillator, the wavelength of the actinic pulse was tuned at 530 nm for OmR2. The pulse intensity was adjusted to 2 mJ per pulse. All data were averaged to improve the signal-to-noise ratio (n = 30). All measurements were conducted at 25 °C. For these experiments, the samples were suspended in Buffer A. After the measurements, the reproducibility of the data was checked to confirm that the sample was not denatured during the measurements. To investigate proton uptake and release during the photocycle, we used the pH indicator pyranine (final concentration = 100 µM, Tokyo Chemical Industry Co., Ltd, Japan), which has been extensively used to monitor light-induced pH changes in various rhodopsins 14,42 . The pH changes in the bulk environment were measured as the absorption changes of pyranine at 450 nm. The absorption changes of pyranine were obtained by subtracting the absorption changes of samples without pyranine from those of samples with pyranine. The experiments using pyranine were performed in an unbuffered solution containing 1 M NaCl and 0.05% (w/v) DDM (pH 7.0) to enhance the signals. The results of 1000-traces were averaged to improve the signal-to-noise ratio.</p>

Scientific Reports - Nature

Mechanism for the Generation of Robust Circadian Oscillations through Ultransensitivity and Differential Binding Affinity

Biochemical circadian rhythm oscillations play an important role in many signaling mechanisms. In this work, we explore some of the biophysical mechanisms responsible for sustaining robust oscillations by constructing a minimal but analytically tractable model of the circadian oscillations in the KaiABC protein system found in the cyanobacteria S. elongatus. In particular, our minimal model explicitly accounts for two experimentally characterized biophysical features of the KaiABC protein system, namely, a differential binding affinity and an ultrasensitive response. Our analytical work shows how these mechanisms might be crucial for promoting robust oscillations even in suboptimal nutrient conditions. Our analytical and numerical work also identifies mechanisms by which biological clocks can stably maintain a constant time period under a variety of nutrient conditions. Finally, our work also explores the thermodynamic costs associated with the generation of robust sustained oscillations and shows that the net rate of entropy production alone might not be a good figure of merit to asses the quality of oscillations.

mechanism_for_the_generation_of_robust_circadian_oscillations_through_ultransensitivity_and_differen

Special Issue<!>Introduction<!><!>Introduction<!>Methods: KaiABC Oscillator and Model Details<!><!>Methods: KaiABC Oscillator and Model Details<!>Differential Binding of KaiA to KaiC Drives the Phosphorylation Phase<!>Dependence of the Kinetic Rates on the ATP Concentration<!>Dynamics of the Dephosphorylation Phase<!>Ultrasensitive Response of KaiC Phosphorylation to KaiA Concentration<!><!>Ultrasensitive Response of KaiC Phosphorylation to KaiA Concentration<!>Dependence of the Kinetic Rates on the KaiA Concentration<!>Results: Role of Differential Affinity and Ultrasensitivity. Insights from an Analytical Treatment of the Nonlinear Fokker–Planck Equations<!><!>Results: Role of Differential Affinity and Ultrasensitivity. Insights from an Analytical Treatment of the Nonlinear Fokker–Planck Equations<!><!>Results: Role of Differential Affinity and Ultrasensitivity. Insights from an Analytical Treatment of the Nonlinear Fokker–Planck Equations<!><!>Results: Role of Differential Affinity and Ultrasensitivity. Insights from an Analytical Treatment of the Nonlinear Fokker–Planck Equations<!>Increasing Differential Affinity Leads to Oscillations at Low % ATP<!><!>Improving the Ultrasensitive Response Leads to Oscillations at Lower % ATP and Fixed Differential Affinity<!><!>Improving the Ultrasensitive Response Leads to Oscillations at Lower % ATP and Fixed Differential Affinity<!>Metabolic Compensation of Time Period: Insights from the Minimal Markov State Model<!><!>Metabolic Compensation of Time Period: Insights from the Minimal Markov State Model<!>Thermodynamic Costs of Setting Up Oscillations<!><!>Conclusion<!>

<p>Published as part of The Journal of Physical Chemistry virtual special issue "Dave Thirumalai Festschrift".</p><!><p>Most living organisms, ranging from simple single celled organisms like cyanobacteria to multicellular organisms, possess an internal clock which is entrained with the day–night cycle.1−5 The fidelity and robustness of this clock are crucial for the well-being and survival of the organism.6−9 The time period of the internal clock has, for example, been found to be robust with respect to changes in the temperature, nutrient conditions, and pH.10−13 Understanding the biochemical and thermodynamic underpinnings of such robust behavior remains an important challenge given the crucial biological role of the internal clock.</p><p>In this paper, we build on recent experimental and modeling work in ref (17) and show how a particular ultrasensitive switch in the KaiABC biochemical circuit can control the quality and robustness of oscillations. In particular, in ref (17), the authors identify a previously underappreciated ultrasensitive response in the phosphorylation levels of the KaiC proteins as the concentration of the KaiA proteins is tuned. The KaiB proteins play no role in this ultrasensitive response. It was postulated in ref (17) that this ultrasensitive switch plays a central role in ensuring robust oscillations. Specifically, the ultrasensitive switch allows the system to exhibit sustained oscillations even at low levels of the energy rich molecule, ATP.17 Motivated by this work, we build a minimal Markov state model that provides analytical insight for how an ultrasensitive KaiA–KaiC switch can modulate the quality of oscillations. Our minimal model also allows us to analytically study how another biophysical driving force, namely, the differential affinity of the different forms of KaiC to KaiA,10,15,19,22 also controls oscillations. Finally, our minimal Markov state model allows us to comment on the thermodynamic costs associated with setting up robust oscillations in the KaiABC system.</p><p>The KaiABC protein system (see Figure 1) provides a minimal biochemically tractable model to explore the above-mentioned questions. The KaiABC system is found in cyanobacteria S. elongatus where it plays the role of regulating the circadian cycle. The KaiABC system consists of three proteins, KaiA, KaiB, and KaiC.14 In vitro, the system of KaiABC proteins undergoes sustained oscillations as evidenced by the phosphorylation state of the KaiC protein. These oscillations have been shown to have many of the same robust features as those observed in the circadian oscillations they support in cyanobacteria.15,16 The KaiABC model system has been probed in many experimental and theoretical studies.10,14 These have elucidated some of the necessary requirements for the generation of sustained oscillations.10,15,17−21 Despite these advances, understanding the biochemical and biophysical driving forces that are responsible for sustaining robust oscillations remains an open question.10,14,16,19,21,22</p><!><p>KaiC monomer. The following schematic has been inspired from ref (14). The KaiC protein exists as a hexamer, and each monomer consists of 2 domains, CI and CII. The CII domain has two phosphorylation sites, Ser-431 and Thr-432, a KaiA binding site, and a nucleotide binding site (which binds either ATP or ADP). The CI domain binds to KaiB and helps sequester KaiA. Subsequently, the KaiABC complex will be denoted using –/A/BCITP/DP––/ACIITP/DPU/T/S/D. Here, TP/DP denotes ATP/ADP attached to the domain: U denotes that none of the sites in CII are phosphorylated, S means that only the serine site is phosphorylated, T means the threonine site is phosphorylated, and D denotes the doubly phosphorylated form. A attached to CI denotes sequestered KaiA; A attached to CII denotes active KaiA acting as an assistant in phosphorylation. B attached to CI implies the inactive form which will start sequestering KaiA.</p><!><p>The rest of the paper is organized as follows. We first briefly review the salient features of the KaiABC biochemical circuit and then outline our minimal model. This model captures the above-mentioned features of the KaiABC circuit, namely, the differential affinity of KaiC to KaiA binding, and the ultrasensitive response of KaiC phosphorylation levels to changes in KaiA concentration. It also additionally accounts for many other experimentally characterized biophysical forces.14 We then write down a stochastic master equation to describe the dynamics of our model. This stochastic master equation is nonlinear in the probability. The nonlinearity is due to the various feedback mechanisms that are necessary for sustaining oscillations. Interestingly, by solving the nonlinear stochastic master equation, we are able to analytically describe the emergence of global oscillations in response to changing the differential affinity.21 Our model allows us to obtain approximate analytical solutions that provide qualitative insight into how tuning ultrasensitivity tunes the quality of oscillations. Crucially, our results allow us to elucidate how an ultrasensitive switch can support oscillations even at a lower concentration of ATP. Our results also allow us to explain how the time period of oscillations can be robustly maintained even as the concentration of ATP is tuned, a phenomenon known as affinity compensation. Finally, we comment on the thermodynamic costs associated with sustaining robust oscillations.</p><!><p>The KaiC protein, complexed with KaiA, and KaiB proteins, forms the core of the KaiABC oscillator system. The various possible states of the KaiC protein are described in Figure 1. Our minimal model, described in Figure 2b and inspired by refs (14) and (21) (with additional modifications to include features such as ultrasensitivity), can be viewed as a coarse-grained description of the various biochemical states accessed by the KaiABC protein system.14 In the full KaiABC cycle, the KaiABC has two conformations, an active conformation (cyan background in Figure 2) which can phosphorylate the Ser and Thr sites with KaiA as an assistant molecule and an inactive conformation (red background in Figure 2) which sequesters KaiA with the help of KaiB and dephosphorylates the Thr and Ser sites. In our model, the P1 and P3 states correspond to the active conformation and P2 to the inactive conformation.</p><!><p>In penel a, rows are labeled I, II, III, IV, and columns are labeled A, B, C, D. In panel a, the colors in the reaction arrows correspond to those in panel b. Active conformations are denoted using a cyan background, and inactive conformations are denoted using a red background. In our model (panel b), the horizontal axis represents the amount of phosphorylation in the system, with ϕ = 0 and ϕ = 2π corresponding to the completely dephosphorylated state and ϕ = π corresponding to the completely phosphorylated hexamer. The phosphorylation function is a linearly increasing function, 0 at ϕ = 0, 1 at ϕ = π, and then symmetrically decreasing from ϕ = π to 2π. Thus, phosphorylation, . Changes in the phosphorylation levels of the KaiC hexamers give rise to oscillations. KaiA binds to KaiC during the "day" and promotes phosphorylation, whereas at "night", KaiB binds to KaiC and sequesters KaiA, thus leading to dephosphorylation. The horizontal rungs in all the states correspond to the phosphotransfer reactions and the hydrolysis of ATP accompanying it, i.e., the red arrows between IA → IIB, and IB → IIC, purple arrows between IIIC → IIIA, and green arrows between IVD → IVA in Figure 2a. The ratio of the forward and backward rates is given by, γ, γ1, and γ2 which are all less than 1, because of the fact that these describe reactions coupled to ATP hydrolysis which are inherently irreversible. In the model, α > 1 is responsible for differential affinity, corresponds to % ATP, and k1 helps in tuning ultrasensitivity. Free KaiA, Af, provides nonlinearity to the system.</p><!><p>The various biochemical states of the KaiABC protein are summarized in Figures 1 and 2. Below, we briefly recap the various salient features of the KaiABC oscillatory cycle and explain how they are taken into account in our minimal model.</p><!><p>At the beginning of the cycle, most of the KaiC is in the active conformation in the CIDP–CIIDPU form (IIIA in Figure 2a, P1(0) in Figure 2b), and most of the KaiA is free. Depending on the phosphorylation level of active KaiC, it binds differently with KaiA. At low levels of phosphorylation (IIIA, IIIB), KaiC binds very strongly with KaiA. By constrast, the affinity of KaiA for KaiC is low when the KaiC is in a highly phosphorylated state (IIIC, IIID). This phenomena is termed as a differential affinity of KaiC for KaiA dimers.23 Our model captures this effect through the parameter α, where α > 1. Specifically, the rates of P1–P3 exchange are given by kAfAf (where Af is the free KaiA concentration) from P1 to P3 and by kAb,0αϕ−π in the reverse direction. As the phosphorylation level increases with ϕ, the term αϕ−π ensures that the proportion of P1 (KaiA unbounded) states increases. The extent of differential affinity in our model can be tuned by varying the parameter α. Differential affinity ensures that the unphosphorylated IIIA state is primed for KaiA binding at the start of the phosphorylation cycle. Indeed, KaiA binding to the IIIA state transitions the system into the IIA and IA states. Subsequently, KaiA facilitates rapid exchange of nucleotides which lead to the formation of more ATP bound states and pushes the system toward phosphorylation; i.e., it leads to the formation of CITP–ACIITPS, CITP–ACIITPT, and CITP–ACIITPD states (IB, IC, and ID states, respectively, in the schematic).</p><!><p>The concentration of the energy rich molecule, ATP, is an important external condition for the cyanobacteria which affects the KaiABC oscillator. It has been observed that oscillations with almost the same time period are sustained until the % ATP in the system reaches 25% below which oscillations vanish completely.10 Here, % ATP . In our model, the concentration of ATP controls the kinetics of the crucial ATP–ADP nucleotide exchange reaction.14 Since in our minimal model the reaction corresponding to III(A, B, C) → I(A, B, C) is coarse-grained into P1(i) → P3(i), and since the second step in these reactions, i.e., II(A, B, C) → III(A, B, C), is dependent on % ATP, the % ATP in our model is set by the ratio of the rates connecting the P1 to the P3 states:2.1Increasing Kd0 decreases the rate of transitions to the P3 form and thus corresponds to lower % ATP and vice versa.</p><!><p>In the hexamer, the dephosphorylation phase starts even before total phosphorylation of each and every monomer. Specifically, once the number of phosphorylated serine sites becomes larger than the number of threonine sites which are occupied, the KaiA dissociates from the complex, the KaiC transforms into an inactive conformation, and the dephosphorylation phase kicks off. This transition corresponds to ID → IID in the schematic in Figure 2a and to the vertical rungs between P1 and P2 states that are colored magenta in our model in Figure 2b.</p><p>The dephosphorylation phase (IVD → IVA) is relatively simple. It does not require KaiA as an assistant molecule for the reactions. When the proportion of doubly phosphorylated KaiC (ID, IID) is high, KaiB binding to the CI domain of KaiC is triggered, IID → IIID. In our model, the KaiB binding to KaiC is taken into account implicitly during the transition from P1 to P2 states. KaiB bound KaiC, BCIDP–CIIDPD (IIID), sequesters KaiA, i.e., binds to KaiA and makes it unavailable for active use. This is taken into account through the parameter ϵseq in our model which reduces the free KaiA in the system by an amount ϵseq∑P2. The dephosphorylation proceeds through the serine sites and then the threonine sites. Dephosphorylation reactions occur through phosphotransfer.22 This corresponds to the system moving through the P2 states in our model. As the reactions reach the completely dephosphorylated state ABCIDP–CIIDPU (IVB), the KaiABC complex starts dissociating into KaiC and KaiB and releasing free KaiA into the system (IVB → IVA). The connection between P2(0) and P1(0) in our model takes this dissociation step. This prepares the system for the next cycle.</p><!><p>It has been experimentally observed that, in the absence of KaiB in the system, KaiC shows an ultrasensitive response in phosphorylation to KaiA concentration in the system; i.e., the phosphorylation level of the KaiC hexamers changes rapidly within a very narrow range of total KaiA concentration.10,17 This ultrasensitivity was speculated to be an important prerequisite for sustaining robust oscillations, particularly in conditions wherein the concentration of the energy rich molecule, ATP, is low. Our model captures the ultrasensitive response observed in ref (17) and described in Section II, through the introduction of the dephosphorylation rate k1 (see Figure 3). Indeed, a standard way to obtain an ultrasensitive response is through the action of two antagonistic enzymes working at saturation.24,25 Under such conditions, the response of the system changes rapidly over a very narrow range of the enzyme concentration. In the KaiABC system, the roles of the antagonistic enzymes are played by KaiA, which acts as a kinase phosphorylating KaiC and KaiC, which acts as its own phosphatase dephosphorylating itself.10,22</p><!><p>Ultrasensitive response in phosphorylation of KaiC with regard to the total KaiA concentration for Kd0 = 10 and α = 10. The values in the bracket are the Hill coefficients for the response curves (calculated using the method of relative amplification26). Values of other parameters are given in Table S2. These kinetics are in the absence of KaiB and P2 states (ω = ω1 = 0); i.e., they represent only the active form of KaiC in Figure 2b. Thus, there are no oscillations, and the system always settles into a final steady state.</p><!><p>The rate k1 in our model captures this dephosphorylation. Tuning dephosphorylation rates by increasing k1 leads to competition between phosphorylation in the P3 states and dephosphorylation in the P1 states. In the absence of KaiB, which corresponds to setting ω = ω1 = 0 in our model, we consequently observe an ultrasensitive response of phosphorylation level of KaiC to changes in the KaiA concentration (Figure 3).</p><!><p>As has been described above, the rates of transition between the P1 and P3 states in our minimal model depend on the concentration of free KaiA, Af. The amount of free KaiA in turn depends on the concentrations of the P3 and P2 states since the KaiC complex is bound to KaiA in these states. Subsequently, . As the amount of P3 and P2 states increases, the free KaiA concentration decreases. This step gives rise to nonlinearity in the system.</p><!><p>Our minimal model described in Figure 2b and Section II can be represented mathematically using a nonlinear Fokker–Planck equation,, where P⃗ is the probability vector of all the states (P1, P2, P3), and W = W(P⃗) is the rate matrix that is dependent on the state of the system. The nonlinear Fokker–Planck equation is described in full detail in the Supporting Information, Section S1.</p><p>If there were no nonlinearity in the Fokker–Planck equation, the Perron–Fobenius theorem would have ensured that the Fokker–Planck equation has a stable time-independent steady-state solution. The oscillatory solutions of the rate matrix decay with time as they have eigenvalues with a negative real part. Due to the nonlinearity in the Fokker–Planck equation in the Supporting Information, eq S1.5, time-dependent oscillatory steady-state solutions may be possible.</p><p>In this work, we focus on how the solutions of the Fokker–Planck equation change as two specific parameters, namely, α controlling the differential affinity and k1 controlling the ultransensitivity, are varied. In particular, we analytically show how the system can be made to transition from a time-independent steady state, where it cannot function as a biological clock, to a time-dependent steady state, where it can function as a biological clock, as the differential affinity parameter α is tuned. For the case where the ultrasensitivity parameter k1 is tuned, we take inspiration from our solution from tuning α and obtain an approximate solution. Our approximate analytical arguments provide insight into how ultrasensitivity also supports the functioning of the biological clock.</p><p>Finally, as has been reported in many experimental and theoretical studies,10,14,16 oscillations are affected by the concentration of % ATP in the system. In particular, it has been found that the KaiABC system cannot sustain oscillations below a critical ATP concentration. In the next section, we will use our minimal model to show how stronger differential affinity and a better ultrasensitive switch can in fact sustain oscillations even at lower ATP concentrations.17</p><p>We begin our analytical treatment by first considering the case where k1 = 0, i.e., in a model devoid of ultrasensitivity. In this case, a time-independent solution for the nonlinear Fokker–Planck equation can be obtained in the limit when ϵseq = 0 and ϕ0 = π. ϵseq = 0 corresponds to the absence of KaiA sequestration by KaiB bound KaiC states. ϕ0 = π means that the dephosphorylation phase starts only after all the KaiC species have become doubly phosphorylated. Our analytical derivation is discussed in detail in Supporting Information, Section S2A, and leads to the following solutions (Figure 4).3.13.23.3where b = P3(0) can be obtained by solving a quadratic equation as mentioned in the Supporting Information, Section S2, . Even when ϕ0 < π, our solution gives a very good approximation if we set P1(j) ≈ P2(2N – j) ≈ P3(j) ≈ 0 ∀j ∈ [j0, N].</p><!><p>Comparison between numerical and analytical results for the time-independent solution of P1 states (eq 3.2) for different α's. The figure in the inset is a representation of the Markov state network with the P1 states highlighted. In the main figure, gray corresponds to α = 2, red to α = 4, blue to α = 6, and green to α = 8.</p><!><p>As α is increased, this time-independent state becomes unstable giving rise to a oscillatory state. As described in the Supporting Information, Section S3, a linear stability analysis can be performed around the steady state of the system, , to characterize this instability. The linear stability analysis has been detailed in the Supporting Information, Section S3A. This analysis correctly predicts the observed oscillatory behavior. Indeed, in Figure 5, we show that the analytical estimate of the time period of oscillations provides a very good description of the actual observed oscillation periods.</p><!><p>Time period of oscillations for various α and Kd0, i.e., at varying levels of differential affinity and % ATP. k1 = 0. Other parameters are given in Table S1. Since k1 = 0, there is no effect of ultrasensitivity. The figure on the left represents time periods calculated by numerically simulating the FPEs. The figure on the right represents the time periods which were calculated from the imaginary part of the maximum positive eigenvalue of the instability matrix W, for small perturbations around the steady-state probability distribution. As can be seen, the analytical solution provides us with a good approximation of the time period as well as the critical α at which oscillations take place for different Kd0 values. The contours in the figure are for the time period of the oscillations.</p><!><p>In the case of k1 ≠ 0, only an approximate solution for the time-independent steady state can be obtained. In order to obtain this approximate solution, we take inspiration from the solution for the case when k1 = 0 and assume kAfAfP1(ϕ) = kAb0αϕ−πP3(ϕ) for ϕ ∈ [0, ϕ0] (along the P1–P3 connections in Figure 2b) and P1(ϕ) ≈ 0 ≈ P3(ϕ) for ϕ > ϕ0. This assumption is supported by numerical evidence. Under this assumption, we obtain3.43.53.6Here, P3(0) can be obtained numerically, and ϕ0 denotes the place where P1–P2 connections start in Figure 2b. This is described in more detail in Supporting Information, Section S2B. Figure 6 shows a comparison between the numerically obtained steady state with the one constructed using our approximate solution. We also provide approximate analytical arguments to show how a linear instability analysis can again be used to characterize the onset of oscillations as k1 is tuned. The Gershgorin circle theorem provides us with a way to understand where we can find the eigenvalues of any matrix. As k1 is tuned, the negative off-diagonal elements of the rate matrix W increase in magnitude, as do the radii of the Gershgorin discs (see Figure S7), because, for any transition rate matrix, M, ∑iMij = 0. In effect, the Gershgorin discs have a finite area protruding into the positive half-plane. With higher k1, this area increases; thus, there is a higher chance of finding eigenvalues in the positive half-plane. These arguments are explained in more detail in the Supporting Information, Section S3A.</p><!><p>Comparison between numerical and approximate analytical results for the time-independent solution of P3 states for the case when k1 ≠ 0 (eq 3.4). The figure in the inset represents the Markov state network with the P3 states highlighted. In the main figure, gray corresponds to k1 = 0, cyan to k1 = 10–4, violet to k1 = 5 × 10–4, red to k1 = 10–3, blue to k1 = 5 × 10–3, green to k1 = 10–2.</p><!><p>In the next section, we build on these results and show how ultrasensitivity and differential affinity can support oscillations even at a lower ATP concentration. We also use the insight from these analytical arguments to explain how the time period can be stably maintained in a variety of ATP concentrations, a phenomenon known as affinity compensation. Finally, using our minimal model, we also comment on the thermodynamic costs associated with maintaining oscillations.</p><!><p>It has been numerically shown previously in ref (21) that oscillations in a model system similar to ours can be obtained by increasing the value of α, i.e., by improving the differential affinity. α controls the rate of reaction between P1 and P3 states in Figure 2b. Our analytical results explain this numerical observation. Further, our analytical results at k1 = 0 also help predict the required interplay between α and the ATP concentration in order for oscillations to be sustained. Specifically, we find that, at k1 = 0, a higher value of α is required for oscillations to take place at higher Kd0 (or a lower ATP concentration). In Figure 8, we provide estimates of how the critical value of α changes as a function of the Kd0. Our analytical estimates agree very well with those obtained from the numerical calculations.</p><!><p>Instability leading to oscillations when changing k1. The y-axis denotes the maximum eigenvalue of the rate matrix W for the perturbations (refer to Supporting Information). The presence of a positive eigenvalue denotes that the time-independent steady state is unstable. α = 10, and the other parameter values are listed in Table S2.</p><p>Value of α required for the onset of oscillations as a function of Kd0. Estimates have been obtained both from our theory and from numerical simulations. We set k1 = 0 for these calculations.</p><!><p>As mentioned in Section II, it has been speculated that ultrasensitivity plays an important role in sustaining oscillations at low % ATP conditions. Our minimal model captures this role played by ultransensitivity. Indeed, we find that, at a higher value of k1, corresponding to a sharper ultransensitive response (Figure 3), oscillations can be sustained for a larger Kd0 (or a smaller ATP concentration). We describe this trade-off in Figures 7 and 9.</p><!><p>Value of k1 required for the onset of oscillations as a function of Kd0. Since k1 ≠ 0 is only approximately tractable analytically, we have only plotted estimates from numerical simulations.</p><!><p>Our analytical analysis also allows us to provide a phenomenological understanding of the role played by the ultransensitive switch. Ultrasensitivity offers coherence to the traveling wave packet of phosphorylation at the start of every new cycle of oscillation. Phosphorylation is halted until a critical amount of KaiA is present in the system. Just before the beginning of every new phosphorylation cycle, most of the KaiA is sequestered by the P2 states. Only after a certain amount of KaiA is freed from P2 states can the phosphorylation reactions in the P3 states start again. This leads to a build-up of probability density near P2(2π) and P1(0) before the start of every cycle and provides coherence to the system, and oscillations can be sustained.</p><!><p>One of the most important features of the KaiABC oscillator is that the time periods of the oscillations are robust to changes in the % ATP in the system, a phenomenon known as metabolic compensation. Our model shows a similar behavior. Upon increasing Kd0, the time period increases, changing by 10% for an increase from Kd0 = 1 to 11 (see Figures 10, 11, and 12). At Kd0 > 11, oscillations are not supported. This is analogous to losing oscillations when % ATP is below 20% ATP in the real system.10,14</p><!><p>Time period of oscillations for various Kd0 and k1 values, i.e., at different levels of % ATP and ultrasensitivity. The white region denotes the parameter space which does not support oscillations. This is also supported by the plot for the amplitude of oscillations, Figure 11. In order to have oscillations at higher values of Kd0, the system requires a higher value of k1. The contours in the figure are for the time period of oscillations.</p><p>Amplitude of oscillations as a function of Kd0 and k1 at α = 10 and parameters given in Supporting Information, Section S2. The contours in the figure are for the amplitude of oscillations.</p><p>Velocity of phosphorylation wavepacket as a function of average angle for k1 = 0.05, with various Kd0's and other parameters as given in Table S2. Here, the average angle ⟨ϕ⟩ = ∑ϕϕP(ϕ), and velocity . The time period of oscillation for the different cycles is denoted along the curves.</p><!><p>Our minimal model helps provide a simple phenomenological explanation of affinity compensation. In the regime where our model allows oscillations, the speed of the waveform as it traverses the top P1–P3 rungs in Figure 2b from regions of lower ϕ to regions of higher ϕ can be shown to be through a first-passage time analysis (outlined in Supporting Information, Section S4). Thus, it is expected to decrease with Kd0. Simultaneously, 1/Kd0 ≡ kAf/kAb,0 can be expected to control the relative occupancy of the P1–P3 states, and the transitions in the P3 states promote the probability flux toward regions of higher ϕ. Thus, with increasing 1/Kd0, the waveform can be expected to traverse more of the large ϕ states in the P3 rung before transitioning to the P1 and then eventually to the P2 states as it restarts the oscillation. Hence, at higher 1/Kd0 or higher % ATP, the system traverses a larger orbit as described in the "angle-angular velocity" phase space (Figure 12). This is analogous to shifting in the trough and crest in the phosphorylation oscillations observed in the KaiABC system.10 Together, these effects make the time period of oscillations relatively insensitive to % ATP levels (Figure 12). In this way, the KaiABC system can accomplish affinity compensation and maintain a relatively constant time period.</p><!><p>Finally, the stochastic thermodynamics of our minimal model can be readily probed. The total steady-state entropy production rate can be estimated using the probability fluxes along every edge of the model as274.1Here, J+ refers to the flux in the forward direction, and J– refers to the flux in the backward direction. For instance, if A and B are two states of a system with reactions between them given by , then J+ = k1[A] and J– = k2[B]. Since the entire KaiABC system has been coarse-grained into a minimal Markov system, we underestimate the value of actual entropy production in the entire system.28 We use eq 4.1 to estimate the entropy production rate for various values of α, Kd0, and k1. These results are described in Figures 13 and 14. Of particular note, our results show that σ̇ varies continuously through the transition of the system from a stationary to an oscillatory phase. In the case where the ultrasensitivity parameter k1 is tuned (Figure 14), the entropy production rate σ̇ is almost a linearly increasing function of k1. While the entropy production rate σ̇ does indeed increase as oscillations are set up in agreement with previous studies,21 and it does indeed improve the overall quality and coherence of oscillation,29,30 an analysis focused on just the entropy production rate might miss the important and specific roles played by biophysical mechanisms such as the ultransensitivity and differential affinity in promoting and sustaining robust oscillations.31</p><!><p>Entropy production rate vs α for Kd0 = 5, k1 = 0, and other parameters given in Table S1. Oscillations start at α = 21. α = 1 corresponds to the absence of differential affinity. In order to have oscillations, an additional 0.113 units of energy are required. This energy goes into building coherence among the KaiABC oscillator population21</p><p>Entropy production rate vs k1 for α = 10, Kd0 = 8 and other parameters given in Table S2. Unlike the case with changing α in Figure 13 where the entropy production plateaus very quickly with increasing α, in this case, the entropy production increases almost linearly with increasing k1. As expected, decreasing Kd0 and increasing k1 lead to a higher dissipation of energy. Oscillations start at k1 = 0.03. k1 = 0 corresponds to the absence of ultrasensitivity in the system. An additional 0.052 units of energy are dissipated in order to have oscillations. This additional energy goes into improving the ultrasensitive response of the system, eventually leading to coherence.</p><!><p>In conclusion, this work elucidates the role played by biophysical mechanisms such as ultrasensitivity and differential affinity in controlling the quality of circadian oscillations. Our minimal theoretical model also provides a route to explain how biochemical circuits can ensure oscillations with constant time periods, even under a range of experimental conditions. Finally, we show that the net rate of energy dissipation is not a very effective order parameter to gauge the quality of oscillations, particularly in regimes where the ultrasensitivity is important, while our work relies on a very minimal abstraction of the KaiABC system. In future work, we hope to adapt these ideas to more complex and complete models of circadian rhythm oscillators.</p><!><p>Details of the analytical calculations, the numerical simulations, and the methods used (PDF)</p><p>jp1c05915_si_001.pdf</p><p>The authors declare no competing financial interest.</p>

PubMed Open Access

Solvent-Dependent Characterization of Fucoxanthin through 2D Electronic Spectroscopy Reveals New Details on the Intramolecular Charge-Transfer State Dynamics

The electronic state manifolds of carotenoids and their relaxation dynamics are the object of intense investigation because most of the subtle details regulating their photophysics are still unknown. In order to contribute to this quest, here, we present a solvent-dependent 2D Electronic Spectroscopy (2DES) characterization of fucoxanthin, a carbonyl carotenoid involved in the light-harvesting process of brown algae. The 2DES technique allows probing its ultrafast relaxation dynamics in the first 1000 fs after photoexcitation with a 10 fs time resolution. The obtained results help shed light on the dynamics of the first electronic state manifold and, in particular, on an intramolecular charge-transfer state (ICT), whose photophysical properties are particularly elusive given its (almost) dark nature.

solvent-dependent_characterization_of_fucoxanthin_through_2d_electronic_spectroscopy_reveals_new_det

<p>Carotenoids are pigments that play a crucial role in photoprotection1−4 and energy-transfer processes taking place in light-harvesting complexes.5−11 It is clear that to understand the fine details of these processes, an accurate description of the photophysics of these molecules is necessary. Many decades of investigations lead to a good knowledge of the main photophysical and dynamic properties of this crucial family of chromophores.</p><p>The structural feature common to all the carotenoids is a long polyene backbone of N conjugated C=C bonds, a number that determines the main spectroscopic properties of these chromophores.10,12 The main challenge in characterizing the electronic states of carotenoids is connected primarily to the presence of dark states, one of their most intriguing features. In fact, it is well-known that the absorption properties of longer polyenes are dominated by the transition from the ground state (S0) to the second excited state (S2), since the transition to the first excited state (S1) is symmetry forbidden.13−16 Historically, the dark S1 state has been studied, for example, through pump–probe spectroscopy where S1 can be populated indirectly from the S2 state, through ultrafast internal conversion (typically happening in hundreds of femtoseconds),17−19 and then it can be probed thanks to an excited state absorption (ESA) promoted by the transition from S1 to higher excited states (Sn). The ESA signal decays as the S1 population returns to S0, giving information about the S1 lifetime (usually characterized by a picosecond time scale).10,18,20</p><p>In this work, the attention is focused on fucoxanthin (Fx), a carotenoid typically found in diatoms antenna proteins called fucoxanthin-chlorophyll proteins (FCP), that belong to the family of intrinsic light-harvesting complexes. In FCP, fucoxanthin is directly involved in the light-harvesting actions since it serves as a major light-harvesting pigment, transferring excitation energy very efficiently to chlorophyll molecules.21−24 It is classified as a carbonyl carotenoid, a class of carotenoids that presents a keto group conjugated with the polyene chain. This functional group is responsible for peculiar spectroscopic features, mainly associated with the formation of an intramolecular charge-transfer (ICT) state in the excited states' manifold. Carbonyl carotenoids generally display steady-state absorption spectra asymmetrically broadened and devoid of the characteristic vibronic structure. Moreover, their transient absorption spectra are characterized by the presence of an additional ESA band, associated with the ICT → Sn′ transition.25−32 These features appear or are enhanced when the system is dissolved in a polar environment that supposedly stabilizes the ICT state. In these conditions, the ICT can be effectively populated via internal conversion from S2.26 Extensive experimental and computational studies have not succeeded yet to fully clarify the nature and the properties of the ICT state, whether it is coupled with the S1 state, in which case the two states would represent two minima of the same potential energy surface,33−36 or it is a distinct electronic state.28,29,31,35,37 Several models have been suggested, but none of them fully explains all the experimental evidence collected to date.12,38,39</p><p>A better understanding of the nature and the photophysical behavior of the ICT state is crucial considering the key role it might assume in the excitation energy-transfer mechanisms at the base of the photosynthetic process in the antenna complexes, where the efficiency of the transfer between carotenoids and chlorophylls may exceed 90%.21,40−42</p><p>In this work, Two-Dimensional Electronic Spectroscopy (2DES) has been exploited to obtain additional information on the nature of ICT states in Fx, searching for still unidentified spectral features that could contribute to the overall comprehension of the photophysics of this carotenoid and its dark states, a puzzle that still seems to lack some relevant pieces before its completion. 2DES appears to be ideal in this task because of its recognized capability of identifying signatures of dark states through their dynamic coupling with bright states.42−45 Moreover, the 10 fs time resolution achieved by this technique is essential to investigate with better detail the ultrafast dynamics subsequent to S2 excitation, expected to be in the sub-100 fs time range.26,28,46,47</p><p>As a carbonyl carotenoid, Fx displays solvent-dependent spectral features. Therefore, we investigated the spectral properties of this carotenoid in solvents with different polarity: methanol (Me), acetone (Ac), and toluene (To). The steady-state absorption spectra of Fx in the three solvents are shown in Figure 1, together with the laser emission profile used for the 2DES experiments. In apolar solvents, like toluene, the typical well-resolved vibronic progression can be identified, while in more polar solvents this structure is progressively lost, and the spectra become broader (spectra in acetone and methanol). It is important to notice that this broadening is asymmetrical, leading to an increase of the intensity on the low-energy red tail of the spectra, especially in methanol. Previous studies on peridinin, another carbonyl carotenoid similar to Fx, suggested that this feature could be related to the charge-transfer character of the ground state, resulting stabilized in polar solvents.25,26,37 More recently, Kosumi et al.31 proposed that the asymmetric red-shift in methanol is caused by a peculiar conformation of the Fx molecule ("red" form) with strong ICT character.</p><p>Linear absorption spectra of fucoxanthin in methanol (red line), acetone (yellow line), and toluene (green line), together with the laser emission profile (yellow area). The molecular structure of fucoxanthin is reported in the inset.</p><p>The laser profile used in the 2DES measurements covers exactly this spectral region. On the one hand, it is not possible to push the laser spectrum further to blue to cover a bigger portion of the S0 → S2 transition due to bandwidth limitations of the experimental apparatus used to generate the exciting pulses in the visible range (see the SI). The limitations in the exciting wavelengths prevented so far the systematic characterization of carotenoids by 2DES, as witnessed by the few works available in the literature.48−52</p><p>On the other hand, however, this configuration focused the investigations only on the red tail of the absorption spectrum, allowing us to better characterize the formation of the ICT state in Fx, the ensuing relaxation dynamics, and its solvent-dependent properties. Moreover, this spectral window also allowed performing a spectral filtering action and neglecting all the relaxation dynamics involving higher energy vibrational levels within the S2 manifold. This permitted a significant simplification in the interpretation of the complex dynamics of Fx.</p><p>The results of the 2DES experiments, cast into a series of frequency–frequency maps at selected values of population time t2, are summarized in Figure 2 for the three fucoxanthin samples. The 2DES spectra of Fx are dominated by strong Excited State Absorption (ESA) signals, conventionally reported with a negative sign in 2DES spectroscopy.44,53 All the ESA signals are characterized by the same excitation frequency (x-coordinate), at around 18900 cm–1 (= 529 nm). The x-coordinate of the signal reflects what happens to the systems after the first interaction with the laser pulse, i.e., the promotion of the bright S0 → S2 transition. As previously discussed, this frequency value represents only the red tail of the absorption band, as the laser excitation profile does not cover frequencies higher than 19200 cm–1.</p><p>2DES purely absorptive maps of fucoxanthin dissolved in (a) methanol, (b) acetone, and (c) toluene. The square and the circle identify the ESA signals from the S1 state and from the ICT state, respectively. The cross indicates the coordinates of the 2D map where the hot vibrational states of S1 contribute to the ESA signal.</p><p>Fucoxanthin in methanol (FxMe, Figure 2a) displays two distinct ESA signals, one with an emission frequency (y-coordinate) of ∼18300 cm–1 (square) and the other of ∼16500 cm–1 (circle). The presence of two distinct ESA signals in the transient absorption spectra of Fx dissolved in polar solvents has already been captured by pump–probe spectroscopy.10,28 The high-energy ESA signal is common to all carotenoids, and it has been attributed to the S1 → Sn transition.26 The cross in Figure 2a indicates the coordinates where the hot vibrational states of S1 are expected to contribute to the ESA signal through the S2 → hot S1 internal conversion and the hot S1 relaxation processes. The band at 16500 cm–1, instead, can be found only in Fx and some other carbonyl carotenoids. This band is usually attributed to the presence of an ICT state in the excited states manifold indicating an ESA signal related to the ICT → Sn′ transition.10,28,29,38</p><p>The response of Fx in acetone (FxAc, Figure 2b) is very similar to the one of FxMe. The main difference is that the lower energy ESA band is substantially less intense. This behavior can be explained through the destabilization of the ICT state in more apolar solvents.10 In fact, this solvent-dependent trend is confirmed in the latest set of data collected on Fx in toluene (FxTo, Figure 2c), where the low energy ESA band is not even detected. The positive signals recoded at low excitation and emission frequency coordinates are due to nonresonant solvent contributions.</p><p>The evolution of the 2DES maps as a function of the population time has been analyzed through a global complex multiexponential fitting procedure, which allows fitting the dynamic behavior at all the coordinates of the 2D maps simultaneously.54 It was demonstrated that this procedure could disentangle in a very efficient way the different components that contribute to the evolution of the 2DES signal and determine with remarkable robustness the kinetic constants regulating the time evolution.54 The global fitting methodology was applied to the three sets of data after exclusion of the first 20 fs in order to avoid possible artifacts originating by the time overlap of the exciting pulses.</p><p>Figure 3 summarizes the results obtained for the FxAc sample. The results for the other two samples are reported in the SI. The global fitting provides the values of the time constants regulating the relaxation dynamics together with the amplitude distribution of these constants as a function of the excitation and emission frequencies. This amplitude distribution can be visualized in the form of the so-called 2D-DAS (2D-decay associated spectra), associated with each time constant resulting from the fitting (Figure 3a–c).54</p><p>2D-DAS of fucoxanthin in acetone for the three time constants emerged from the fitting and relative traces. The black square (circle) marks the ESA signal from the S1 (ICT) state. The cross pinpoints the coordinates where hot S1 states are mainly contributing. (a) 2D-DAS of the ultrafast time constant of 130 fs associated with the S2 → S1 internal conversion. (b) 2D-DAS of the 310 fs component associated with the hot S1 relaxation. (c) 2D-DAS of the longer time constant (>1 ps) associated with the S1 relaxation. (d) Decay trace showing the t2 evolution of the signal at the coordinates of the ESA-S1 signal (black square). (e) Decay trace at the coordinates of the ESA-ICT signal (black circle).</p><p>In the specific case of ESA signals, characterized by a negative amplitude, a positive peak (red) in a 2D-DAS means that, overall, the signal at those coordinates is becoming more negative. This corresponds to a growth of the population of the state from which the ESA originates, and therefore, we refer to this behavior as a "rising" component.54 The opposite is true for negative peaks (blue signals in the 2D-DAS), associated with the decay of the population of the same state. These trends can be easily verified by inspecting the time traces at the coordinates of the main peaks appearing in the 2D-DAS, as shown in Figures 3d and 3e.</p><p>The 2D-DAS relative to the long-time constant (Figure 3c) captures the main decaying component in both ESA bands, which represents the restoration of the ground state population in the picosecond time scale. The 2D-DAS relative to the shortest time constant (Figure 3a) depicts a rising signal (positive red signal pinpointed by the square), which, considering the sign and the position, can be associated with the S2 → S1 internal conversion.29−32Figure 3b instead is dominated by a decaying component on the lower part of the S1-ESA band (blue signal highlighted by the cross). This signal appears at coordinates already associated with hot S1 states' contributions and can been interpreted as the hot S1 relaxation, in agreement with other studies.32</p><p>The same analysis has been applied also to FxMe and FxTo samples. In FxMe, the overall dynamics appeared to be faster with respect to less polar Ac solvent. In this case, it was not possible to identify signatures attributable to the hot S1 relaxation, and an overall decay component with a time constant of 65 fs has been found, which we attribute to the S2 → S1 internal conversion (Figure S4). In the FxTo sample, similar to FxAc, besides the >1 ps long time component, two time constants of 120 and 250 fs have been found. The sign and the amplitude distribution of these time components (Figure S5) suggest a possible attribution to the S2 → hot S1 and S2 → S1 processes, respectively. For FxAc (and FxMe), the S2 → hot S1 process is probably too fast to be clearly characterized.</p><p>A closer analysis of the dynamic behavior of the signal at coordinates where hot S1 states are expected to contribute (coordinates pinpointed by the cross in Figure 3) highlighted the presence of an additional time component, which could not be reliably distinguished in the global fitting. A local fitting performed at these coordinates (Figure S7) and the comparison with the time decay of FxAc at the same position (Figure S6) lead to the identification of an additional kinetic component with a time constant of 310 fs, which we likely attribute to the hot S1 vibrational relaxation. The global fitting could not fully discriminate this component from the 250 fs one, given the small difference between the associated time constants.</p><p>The time constants found for the FxTo sample are slightly different than what has been previously found by pump–probe experiments in other nonpolar solvents. For Fx in cyclohexane solutions, Kosumi et al. estimated a time constant of 60 and 620 fs for the S2 decay and the S1 vibrational relaxation, respectively.28 This discrepancy can be explained by accounting for the different nature of the solvent used (toluene rather than cyclohexane) and the different time-resolution (∼10 fs vs ∼100 fs). Nonetheless, the important point is the confirmation of the progressive slowing down of the dynamics in less polar solvents.</p><p>Overall, although the relaxation dynamics of S2 has been the object of an extensive investigation by pump–probe experiments, the improved time resolution of ∼10 fs achieved with 2DES experiments and the possibility of inspecting the sign and the amplitude distribution of the components in the 2D-DAS plots allowed a better characterization and a robust interpretation of the early steps of relaxation, which also revealed a clear solvent-dependent trend. The obtained time constants for the three samples are summarized in Table 1.</p><p>The asterisk (*) indicates the time constants that could be discriminated only by a local fitting.</p><p>The time trace extracted at coordinates corresponding to the S1 → Sn ESA (Figure 3d) shows how the signal evolves in the first 1000 fs after the excitation. The rising and the decay of the ESA band can be observed. Instead, the time trace extracted where the ESA band associated with the ICT → Sn′ transition contributes (Figure 3e) does not present any rising component, and it starts to decay immediately after the laser excitation. This behavior is confirmed by the absence of any rising (red) signals in the corresponding portion of the 2D-DAS (Figure 3a, dashed circle). In addition, both traces reveal the presence of a lively beating behavior, due to the activation of vibrational modes of the Fx. The main beating components in the 2DES signal have been identified through Fourier spectra analysis. Their frequencies agree with well-known characteristic vibrational modes of carotenoids (see the SI).</p><p>One of the most important pieces of evidence emerging from the analysis of the three samples is the increase of the rising ultrafast time constant as the polarity of the solvent decreases (Table 1). A similar solvent-dependent trend has been already detected by Kosumi et al.,30 but the limited time resolution (ca. 100 fs) did not allow for a detailed discussion of this behavior.</p><p>An interesting explanation can be attempted based on a model proposed by Wagner et al.38 for peridinin. In this paper, the authors propose the use of an (S1+S2)/ICT state model, where the ICT state arises from a configurational mixing of the lowest two excited singlet states S1 and S2. The idea of a mixed state between S1 and S2 was already proposed in some early works on Peridinin-Chlorophyll Protein (PCP) to explain particular vibrational features.55,56 This ICT state is characterized by an enhanced dipole moment and thus requires a polar solvent for stabilization. In this picture, the ICT and the S1 states are separated by a polarity-dependent barrier of potential, expected to be very small in polar solvents. The application of a similar model also to Fx would explain the solvent-dependent S2 → S1 conversion rates: in apolar solvents, the barrier is higher, leading to a longer internal conversion between S2 and S1.</p><p>This model is useful also to discuss the nature of the ICT state itself, focusing on another feature already glimpsed in the analysis of its trace along t2 (Figure 3d): the absence of a rising time constant related to the ICT-ESA band. In fact, at early delay times, the samples in methanol and acetone share the presence of the ICT-ESA immediately after the laser excitation. The S1-ESA band instead is characterized by a rising component, derived from the internal conversion from the S2 state. This new experimental evidence could only be captured with a high temporal resolution, and it goes together with recent studies that suggest that S1 and ICT are indeed two distinct electronic states, as the temporal evolution of the relative ESA signals is different.35,36 The presence of an ESA from the ICT state already at t2 = 0 suggests an instant population of that state. This evidence has led to the hypothesis that the ICT state could be directly coupled to the S2 state, an idea already presented in the (S1+S2)/ICT state model invoked previously. Indeed, according to Wagner et al.,38 "in terms of most properties, the ICT state is S2-like in character". This picture seems to be also supported by the work of Ghosh et al., who identified a <20 fs nonradiative decay of the S2 of peridinin to an Sx state with a strong ICT character, assigned to a distorted configuration.47,57</p><p>This model refers to peridinin, whose ESA bands are indistinguishable, and it should also be tested for fucoxanthin; but the idea of a strong mixing between the various states of the system could explain this and other controversial features. Moreover, it is important to stress that, even though the S1 and the ICT states are two distinct electronic states, an interplay between the two states is still possible, as recent pump-dump-probe studies have proposed.35,36 In this picture, more polar solvents, besides stabilizing the ICT and lowering its energy, would also promote a better mixing with S2 and thus a lower barrier of potential. Another interesting insight of this model is the interpretation of the asymmetrical increase of the red tail of the linear absorption spectrum in polar solvents: if the ICT state, coupled with the S2 state, can be populated instantly, there has to be a trace of this process also in the linear absorption, and this could be indeed related to the increased absorbance in the red tail. In fact, in apolar solvents, the ICT is not populated, and the red tail absorbance is less pronounced. It must finally be noted that all these results have been obtained with very specific excitation conditions, where only the most red portion of the absorption spectrum could be addressed (Figure 1).</p><p>In conclusion, in this work, the ultrafast relaxation dynamics of fucoxanthin has been characterized with 2DES, a technique scarcely applied to carotenoid systems because of technical limitations. The obtained results permitted the unraveling of subtle details of the kinetics of the system, through the investigation of the time evolution of ESA signals promoted after the photoexcitation of the red tail of the S2 absorption band and the ensuing relaxation to S1 and ICT states.</p><p>Thanks to the 10 fs time resolution achieved in these experiments, the kinetic constants regulating the S2 → S1 internal conversion of Fx in three solvents with different polarity have been determined with an unprecedented level of detail. Indeed, it is known that the relaxation dynamics of the S2 state in carbonyl carotenoids takes place in the sub-100 fs time regime,26 and therefore, conventional pump–probe experiments with a typical 100 fs time resolution so far could only provide rough estimates of the associated kinetic constants. A clear solvent-dependent trend of the S2 → S1 internal conversion rates has been found: the process becomes faster as the polarity of the solvent increases.</p><p>In addition, we found evidence for two distinct ESA signals developing from the S1 and ICT states, whose relative intensities also depend on solvent polarity. The S1-ESA signal rises in the first 100 fs, which clearly indicates that the S1 state is progressively populated via relaxation from S2. Instead, the ICT-ESA signal is instantaneously present immediately after photoexcitation, with no rise component, suggesting an instant population of the ICT state. These features have been interpreted on the basis of a model previously proposed in the literature for peridinin, which identified a strong coupling between the ICT and the bright S2 state (Figure 4).</p><p>Proposed scheme summarizing the main dynamic processes captured by the 2DES measurements on solutions of Fx. Red arrows: excitation; black wavy arrows: nonradiative processes; green arrows: ESA processes. The black markers (square, cross, and circle) pinpoint the processes identified in the 2D maps of Figures 2 and 3. The barrier of potential between S2 and ICT is solvent-dependent (dashed blue arrow). The purple/blue arrow represents the interplay between the S1 and ICT states, leading to picosecond equilibration between the potential minima.</p><p>While the effective applicability of this model also to fucoxanthin needs to be supported by further studies, the instant presence of the ICT-ESA signal at early times and also its peculiar dependence on solvent polarity point toward a new interpretation of the nature of the ICT state.</p><p>These findings represent, in any case, an important piece of information about the nature and the dynamics of dark states in carotenoids. From a broader perspective, we believe that a better characterization of such dynamics is essential for a better comprehension of the ultrafast relaxation dynamics taking place in more complex multichromophoric antenna systems. This preliminary characterization of Fx in vitro will also contribute to a deeper understanding of the excitation energy-transfer processes that involve carotenoids in vivo.</p><p>Details of 2DES setup and pulse characterization; additional 2DES maps and relative fitting results; and beating analysis (PDF)</p><p>jz1c00851_si_001.pdf</p><p>The authors declare no competing financial interest.</p>

PubMed Open Access

Tip-induced nanoreactor for silicate

Nanoscale scientific issues have attracted an increasing amount of research interest due to their specific size-effect and novel structure-property. From macro to nano, materials present some unique chemical reactivity that bulk materials do not own. Here we introduce a facile method to generate silicate with nanoscale control based on the establishment of a confined space between a meso/nanoscale tungsten tip and a smooth silica/silicon substrate. During the process, local water-like droplets deposition can be obviously observed in the confinement between the Si/SiO2 surfaces and the KOH-modified tungsten tip. By the combination of in-situ optical microscopy and Raman spectroscopy, we were able to take a deep insight of both the product composition and the underlying mechanism of such phenomena. It was indicated that such nanoreactor for silicate could be quite efficient as a result of the local capillarity and electric field effect, with implications at both nano and meso scales.Recently, with the rapid development of high-quality small-scaled techniques and the unlimited possibilities of fabricating micro structures, people are moving the spotlights on researches from macro scale to meso/nano-scale. As a science with objects of smallest dimension, a large amount of unique properties that bulk scale do not possess have been observed in the research of nanotechnology. These not only depend on the size, shape, molecular structure, but also the pressure and temperature of the environment. A considerable amount of materials have already been demonstrated that when they decrease to a critical size, their electrical conductivity 1 , melting point 2 , chemical reactivity 3 and magnetic permeability 4 will be completely different from that at macro scale. Meanwhile, in the nano-sized regime, significant effect would be initiated due to closely confined space, such as the quantum-mechanical confinement effect in nanoelectronics, in which the energy of delocalized electrons increases with decreasing size 5 , and the titanate nanotubes (TNTs) 6 , which can offer a special nano-confinement location. Other scientific fields, like biochemistry 7 and tribology 8,9 , also present a great interest and bright foreground in nano-sized issues.In the field of chemistry, a lot of close attention has been paid to nanoscience, particularly the newborn nanometer sized reactor. Over the past years, the so-called nanoreactors 10 , performing a coupling of reactions in time by confining reagents and catalysts within a nanospace, have presented great potential in improving chemical transformation. So far, several advantages have emerged using theses nanoconfined cell, including reducing reaction time, minimizing amounts of reagents and obtaining fresh productions. For catalytic reaction systems, functional nanoreactors are widely used. They are mostly realized basing on hollow spheres, especially mesoporous silica hollow spheres. Song et al. fabricated a composite nanoreactor with mesoporous silica hollow spheres and Pd nanoparticles inside, which shows remarkable activity for Suzuki cross-coupling reactions 11 . Lee et al. introduced an ideal nanoreactor system comprising gold cores and silica hollow shells with empty inner space for the reduction of p-nitrophenol 12 . Nanophase-separated amphiphilic networks also have been demonstrated to possess the ability to stabilize and enhance the catalytic activity of enzymes in organic solvents 13 .However, a bridge between nanoreactors and inorganic chemistry has not been built up because of lacking practical application. Furthermore, these nanoreactor require complex fabrication process and extremely sophisticated manipulation, which limits its development in industry. In fact, with a view to creating and exploring new chemical productions and novel physical phenomena, nano scientific

tip-induced_nanoreactor_for_silicate

<!>Methods<!>Results<!>Discussion<!>Outlook.

<p>effort should be carried out to seek different disciplines. To achieve this goal, a facile method has been developed to create a nanometer-size cell with an electrochemical etching tungsten tip ranging from tens nanometer to hundreds nanometer. The technique first applied nanoreactor to organic materials, inducing a specific chemical reaction without catalyst at room temperature, which generally occurred demanding high thermal energy. Meanwhile, instantaneous springing out of water droplets was observed along with the reaction process.</p><p>Herein, a micro-manipulator was used to control a nanoscale tungsten tip while reciprocating sliding occurring between the tip and substrate. During the experimental procedure, droplets of "water-like" precipitation were observed in the confined space. The compositions of the products and the underlying mechanism were well investigated with various characterization methods. It has been speculated that certain chemical reaction between KOH residual on the tip and Si/SiO 2 surface has been induced and facilitated by the nanoscale reactor between tip and substrates. The major breakout in our work is to produce silicate with nanoscale control even at room temperature. The study made further effort to simplify the concept and fabrication of nanoreactor, introducing it into organic materials. What's more, it also provided a unique sight for silicate products industry. In our future research, large amount of work will be tended to expand the area of nanoreactor application, trying to discovering more similar reactions and phenomena.</p><!><p>The tungsten tips, terminated with a spherical cap of different radius, were prepared by etching electrochemically in 2 M potassium hydrate (KOH) under 15-V alternating voltage for 90 s in the first stage then sharpened under 2-V alternating voltage for various time.</p><p>The substrate samples used in the experiment were N-type (38-50 Ω cm) Si (111) wafers and Si (100) wafers covered by 300 nm SiO 2 after thermal oxidization. Prior to the operation, the substrate samples were immersed into the acetone solution and alcohol solution for 30-60 s separately in order to eliminate organic contaminants, followed by an extensive rinse with ultrapure water. After the cleaning procedure, the samples were completely dried by N 2 gas flow.</p><p>In a typical experiment, the temperature and relative humidity in the environment is two of the key factors. We conducted it in a near-constant temperature and humidity room. The temperature was kept at 20 ± 2 °C and the relative humidity was maintained around 40 ± 5%. The whole manipulation was performed under an optical microscope with a 50× long working distance microscopic objective. The scattered Raman signal was collected in a backscattering configuration through the objective, filtered, and then dispersed onto a liquid-nitrogen-cooled CCD camera through a single grating spectrometer. As shown in Fig. 1, the motion of the tungsten tips was well controlled by a micromanipulator (Kleindiek, MM3A), then it touched slightly on the substrate surface, with a slow sliding rate of 5 μ m/s. The products were observed with scanning electron microscopy (SEM), related characterization was carried out on energy dispersive X-ray (EDX) and Auger Electron Spectroscopy (AES), and the molecular structure was analyzed using Raman Spectroscopy.</p><!><p>The electrochemical etching of tungsten tips. It has been confirmed that the formation of meso/ nano-sized tungsten tips mainly depends on the voltage and etching time during the electrochemical corrosion. To obtain different radius of curvature, etching time under 2-V alternating voltage has been controlled. As shown in Fig. 2, from panel (a) to panel (d), with the etching time increasing gradually, the radius of tip decreased (from hundreds of nanometer to tens of nanometer). In addition, from the SEM images it can be inferred that some residual KOH from the etching solution has adhered on the head of tips after electrolytic corrosion.</p><p>Silicate produced in the nanoreactor. The optical microscopic images (panel a and c) and SEM images (panel b and d) of products were directly shown in Fig. 3. For part 3a and 3b, the substrate was the Si wafer covered with 300 nm SiO 2 , while for part 3c and 3d, the substrate was the N-type Si wafer. The initial status of precipitations were water-like droplets on both substrates, after deposited in air at room temperature for one day, the products dried out. Here, the radius of the tungsten tip was 278 nm for SiO 2 substrate and 85 nm for Si substrate respectively. Here the tip size is one of the key factors for this induced chemical reaction. Tungsten tips with different radius were obtained through electrochemical corrosion in KOH solution, as displayed in Fig. 3. A series of experiments had been carried out with tips of different size on each surface. The droplets deposition can be only observed when the radius of curvature of tips was less than ~500 nm. On the other hand, we also found that when the size of the tips ranged from tens to hundreds nanometers, bigger tips can produce more product.</p><!><p>In order to further investigate the underlying mechanism of this phenomenon, we first characterize the water-like products with Raman spectra and AES. Panel a and b in Fig. 4 shows the results of Raman spectra collected right after the precipitation sprung out. To make a deep and comparative analysis, the Raman spectra of silica and silicon substrate has also been introduced. The inset part in each panel is the local Raman spectra ranging from 700 to 1200 cm −1 . Comparing the spectra of base and products, we found that the peaks near 880 cm −1 , 950 cm −1 , 3400 cm −1 came out for both substrates. As reported in the previous researches, the band near 880 cm −1 is assigned to the Si-O stretching 14 and the band near 950 cm −1 is assigned to the Si-OH stretching 15 , while 3400 cm −1 presented the O-H stretching in liquid water 16 . It was suggested that some kind of silicate had been produced during the chemical reaction between the substrate surface and the KOH left on the tip, which is quite different from macroscopic issues. In general, KOH solution can hardly react with Si or SiO 2 at room temperature. As reported in the past documents, the etching rates between KOH and Si were measured at 3000 ~ 4000 Å/min at 62 °C17 , while at room temperature these rates become unnoticeable. For the chemical reaction between KOH and SiO 2 , the temperature usually was set to 1300 °C. It can be inferred that these chemical reactions have been effectively enhanced by the tip-induced nanoscale space.</p><p>The result of AES spectrum helps us to further confirm the above hypothesis by providing the elements composition of products. To exclude the influence of substrate, we only collect the AES signal from the top surface of each precipitation, which was about 10 nm deep. As shown in Fig. 4c,d, from the AES spectrum, we can understand clearly that elements like K, O, C, W, Si, and N were concluded in the products, indicating that potassium silicate had been generated. The resource of W element was the 3d. Several significant peaks were marked in the graph. The AES spectrum was collected from the top surface of products. Part a presented the signal collected from the precipitation on silica surface displayed in Fig. 3b and part (b) presented the signal collected from the precipitation on silicon surface displayed in Fig. 3d. potassium tungstate remaining on the tip, which has been produced during the electrochemical etching process.</p><p>To illustrate the precipitation procedure in the nanoreactor, we develop a simple analytical model to describe the presence of the water-bridge that forms between the tip and substrate and to analyze the molecular mechanisms of such nanoscale transporter, as displayed in the Fig. 5. In part a, both tip and substrate were assumed to be hydrophilic, they were wetted by a thin water layer in equilibrium with the ambient humidity. Here the relative humidity was set up to be 40 ± 5%. In addition, SiO 2 surface immersed in water is known to own a negative surface charge destiny, primarily through the dissociation of terminal silanol groups 18 . As the tip approaching, a water bridge between the substrate surface and tip was formed. This water meniscus played an important role here as a nanometer-sized cell for chemical reaction. At first stage. it served as a tiny capillary in which KOH can be dissolved into K + and OH − , then these metal ions can be transported from tip to substrate surface, as shown in part b. Once the tip moved into the range of electric double layer of SiO 2 surface, positive charge would be stimulated at the tip top. In this case, an extremely high electric field has been induced. By Gaussian theorem, the electric field intensity is inversely proportional to radius squared. Thus as the decreasing of the tip dimension, higher electric filed energy was assembled in the nanoconfine area. That is the reason why smaller tips here can promote the generation of silicate.</p><p>Once the intensity of the electric field reached to a specific point, it can provide enough power to increase the Gibbs energy, which can help KOH and Si/SiO 2 obtain activation energy required, leading to the reaction in the nanoscale space, even at room termperatue 19 . So the second step was, as displayed in part c, under the effect of electric field, OH − was absorbed onto the surface of substrate, along with the occurrence of chemical reaction and the formation of silicate. For Si substrate, the KOH etching generally involves at least three reactants: OH − , H 2 O and silicon 20 . The reaction may be described by equation (1) 21 :</p><p>Similarly, SiO 2 substrate also occurred a chemical etching, and the reaction can be simplified to equation (2) 22 :</p><p>Here, for both substrates, a nanoreactor was created, in which above reactions have been induced, with Si(OH) 2 O 2 2− as the final products, confirmed by the Raman spectra and EDX, AES results. As a strong water-absorbing silicate, Si(OH) 2 O 2 2− can strongly grasp the water from air. Furthermore, potassium tungstate generated in the etching experiment has been induced via the water bridge. Due to the absorbency property owning by produced potassium silicate and potassium tungstate, ambient water vapor was quickly absorbed, leading to the observed deposition of "water" droplets under optical microscopy.</p><!><p>In conclusion, a simple system aimed at creating confined space induced by a nanoscale tungsten tip has been designed. We are surprised to observe a phenomenon during the experimentsspringing out of big "water-like" droplets. It has been speculated that the H 2 O-bridge between the tip and substrate provides a passage for the hydration of K + ions and OH − ions. In the nanoreactor induced by the tip, an extremely high concentration area filled with electric field and OH − ions has been created, with a local high-pressure and high-temperature area. These conditions would facilitate the reaction between OH − and Si/SiO 2 , leading to generation of silicate, which serves as the source of droplets. Our new findings in this work open up fascinating views to the researches on the nanoscience.</p>

Scientific Reports - Nature

Thermodynamic driving forces and chemical reaction fluxes; reflections on the steady state

Molar balances of continuous and batch reacting systems with a simple reaction are analyzed from the point of view of finding relationships between the thermodynamic driving force and the chemical reaction rate. Special attention is focused on steady state, which has been the core subject of previous similar work. It is argued that such relationships should contain, besides the thermodynamic driving force, also a kinetic factor, and are of a specific form for a specific reacting system. More general analysis is provided by means of the non-equilibrium thermodynamics of linear fluid mixtures. Then, the driving force can be expressed either in Gibbs energy (affinity) form or on the basis of chemical potentials. The relationships can be generally interpreted in terms of forceresistance-flux.

thermodynamic_driving_forces_and_chemical_reaction_fluxes;_reflections_on_the_steady_state

Introduction<!>⁄<!>Steady states of basic open reacting systems<!>Figure 1. Scheme of the continuous stirred tank reactor (CSTR)<!>Steady state as a kinetic approximation<!>An alternative offered by non-equilibrium thermodynamics<!>Chemical potential in the role of the driving force<!>Conclusion

<p>The thermodynamic analysis of chemically reacting systems is still a lively research area which continues to include efforts to find deeper relationships between thermodynamics and chemical kinetics. A typical example is the identification of "thermodynamic driving forces" which could be directly related to reaction rates, usually to rates in both forward and reverse directions, often called fluxes. The driving force is not a precisely defined quantity. The term probably originates with Berthelot (Melle et al., 1993), who related it to the heat released by a reaction and was, for that, soon criticized by Helmholtz. The classical work on the theory of chemical kinetics (Glasstone et al., 1941) also includes a brief mention of the driving force and links it to the strength of the chemical bond formed during a reaction. The energetic nature of the driving force concept has survived to this day (Gamiz-Hernandez et al., 2015;Zhu et al., 2011).</p><p>Perhaps the most common result of the 'driving force-reaction flux' type of relationship is the equation ∆ = − ln( ⁄ )</p><p>(1) where ∆ is the reaction Gibbs energy, or the driving force, and and are the rates of reaction in the forward and the reverse directions, respectively (the forward and reverse fluxes). Eq. ( 1) can be obtained directly combining the reaction isotherm equation known from traditional (equilibrium) thermodynamics and the classical reaction rate equation in the form of the mass action law. A crucial point here is the identification of the true thermodynamic equilibrium constant with the kinetic equilibrium constant -and this is the main source of certain inconsistencies, which probably cannot be fully resolved even under ideal conditions. More details are given in the review by Pekař (2005) and in the relevant book chapter (Pekař, 2011). In short, the two equilibrium constants are conceptually different. The thermodynamic constant is non-dimensional and has no units, whereas the kinetic constant possesses dimension and units. The (value of the) thermodynamic constant is dependent on the selection of the standard state, and its dependence on, or independence of a given parameter, e.g., pressure, also depends on this selection (that is, whether the selected standard state depends on the given parameter, e.g. pressure, or not). This is quite strange for kinetic rate constants, which form the kinetic equilibrium constant. Beard and Qian (2007) tried to derive Eq. ( 1) solely on the basis of the conservation of mass, without invoking any rate law, as a fundamental relation for any chemical process operating in the steady state of an open-system. The steady state condition thus imposes a certain limitation on their approach. Their argumentation is illustrated on the simple general reaction A B. Let us follow it step by step. First, they suppose that in a nonequilibrium steady state the numbers of A and B molecules are held constant by A being pumped into the system, and B out of the system. Second, they suppose that A molecules formed by back-reaction from B molecules can be labeled (A*) but are otherwise identical with A molecules. Thus, the following reactions take place formally:</p><p>(IIb) It is further claimed that in the steady state A* molecules are in equilibrium with B molecules and that the corresponding equilibrium constant is defined according to = *</p><!><p>, where denotes the number of molecules (in the steady state). This, in fact, means that steps IIa and IIb are taken as a single reversible reaction in equilibrium, which is the essence of another claim -the equality of fluxes A* → B and B → A*. This equality is written in Beard and Qian (2007) as</p><p>(2) where Σ is the total number of type A molecules (the sum of A and A*). Upon substitution from the definition of we obtain: ⁄ = Σ ⁄ .</p><p>(3) The thermodynamic definition of the equilibrium constant (Beard and Qian, 2007) gives: ∆ ̅ = − ln (4) and the corresponding reaction Gibbs energy:</p><p>(5) where should be taken generally (not just in a steady state). Combining (3) and (4), we obtain:</p><p>which, however, is not (1) for the exemplified reaction -for this, equation (1) should read: ) . This discrepancy, mixing and , seems to have been unnoticed. In this work, some different -let us say traditional -approaches to steady state kinetics are analyzed from the 'driving force-reaction flux' point of view and supplemented with some results of the non-equilibrium thermodynamics of chemically reacting mixtures.</p><!><p>In chemical kinetics (chemical reaction engineering), two principal models of open systems are used: the well-mixed system with continuous input and output (the continuous stirred tank reactor or CSTR) and the tubular flow-through system with a plug flow regime (the plug flow reactor or PFR).</p><p>Let us analyze the steady states of these systems for the above simple reaction from the point of view of driving force-flux relationships. First, it should be noted that PFR can be modeled as a series (of a sufficient number) of CSTRs; thus, we will focus here only on the CSTR system. As in Beard and Qian (2007), we suppose the validity of the following equation, the reaction isotherm originating in classical (equilibrium) thermodynamics:</p><p>Here, = $ % /$ % is the equilibrium constant (for an ideal system with a standard state of unit concentration; subscript e denotes equilibrium) and ! = $ /$ is the reaction quotient; $ is the concentration of . Obviously, the component reaction rates fulfill the condition: = − = − (all in mol m -3 s -1 ).</p><p>The molar balance of CSTR (Fig. 1) with inputs ' ( (referred to as "pumping", especially in biorelated works) and outputs ' (both in mol m -3 s -1 ) in steady state (no special index is used to denote steady state values) is given by: ' ( + = ' = $ () * ) ⁄ ) = $ /+ (9) where ) * is the volumetric flow rate (m 3 s -1 ) and ), the reactor (system) volume; + is usually called the space time. From the balance (9), the steady state concentrations can be expressed as follows:</p><p>$ = +,' ( + -.</p><p>(10b) The reaction quotient in the steady state is then</p><p>Equilibrium can be defined as a state in which = 0. Then, from (10), $ % = +' ( and $ % = +' ( , and, consequently, = ' ( /' ( . However, the last equation cannot be seen as a definition of the equilibrium constant, but, rather, as a condition on the setting of the input to attain equilibrium. In fact, the input composition is then identical to the equilibrium composition and no reaction occurs in the system (it does not even begin).</p><!><p>Introducing ( 10) into ( 8) we obtain</p><p>where 2 is a measure of the distance (of an actual state described by the value of ∆ ) from equilibrium -properly speaking, the distance from the standard state. However, the standard state is intimately related to the equilibrium descriptor through the definition ∆ = − ln . In fact, 2 is equal to the reaction quotient; however, we will use a special symbol to highlight its definition by the identity in (13), while the reaction quotient is regularly written as a proper fraction with concentrations (activities); for example, see ! = $ /$ above. The reaction rate can be expressed from ( 13) explicitly: = (2' ( − 1)' ( (2 + 1) ⁄ (14) where ' ( = ' ( /' ( and denotes the ratio of the components in the input. Equations ( 13) and ( 14) are the most general relationships which can be derived from the basic equations (balances) describing this system. On the left hand side of (13) we can see the thermodynamic forcing in terms of the distance from equilibrium. On the right hand side we can see the kinetic forcing (the composition of the input) together with the kinetic outcome -the reaction rate. This result should be understood much more like the "bookkeeping" of the (steady state) situation in the system than as some predictor-like equation. Of course, the reaction rate can be expressed from this bookkeeping, as seen in ( 14) -here, again, the rate is related to the thermodynamic forcing (2) together with the (macroscopic) kinetic forcing, which is simply the input composition (' ( ). Selected examples for ' ( = 1 are shown in Fig. 2. It should be remembered that even the thermodynamic forcing is determined by the composition, because ∆ can be expressed in terms of chemical potentials, which, in turn, can be expressed in terms of concentration.</p><p>In equilibrium, ' ( = 1/ and 2 % = , and Eq. ( 14) predicts a zero reaction rate as expected. When there is no B in the input, Eq. ( 13) gives following simplified version of ( 14):</p><p>= 2' ( (2 + 1) ⁄ , (15) which predicts a positive reaction rate, again as expected. 2) shown in the legend; eq. ( 14), ' ( = 1. Left: overall view, right: detailed view It is interesting to compare a flow-through system with a batch (closed) system (reactor). Here, there is no steady state; the reaction reaches equilibrium, where the reaction rate is zero. The general (non-stationary, i.e. out of equilibrium) balance at constant volume is 9$ /9: = . In contrast to the previous CSTR steady state example, the rate is not constant or related directly to concentration. Integrating the balance we obtain:</p><p>$ − $ ( = ;</p><p>($ ( = $ ( /$ ( ). As noted above, here we do not see the actual (instantaneous) reaction rate but its integral (=) up to a specific time. Nonetheless, in (21), we see again the thermodynamic forcing, 2, related to this integral kinetic characteristic, together with the kinetic forcing, which, in this case, is determined by the initial concentrations of both components. If there is no product B present at the beginning:</p><p>In equilibrium, which is the final state here, in contrast to steady state, we have: = = , $ ( − $ ( -( + 1). ⁄ (23) These results arising from the standard balances of a specific reacting system (reactor) can be summarized in the following way. Though a reaction has a unique reaction Gibbs energy, its relationship to kinetics depends on the system in which the reaction occurs. This is due to the concentration dependence of the reaction Gibbs energy and the different equations balancing the concentrations in different systems. Nevertheless, the two "flux-force" relationships ( 14) and ( 21), or ( 15) and ( 22), are analogous -the rate characteristics ( or =) are related to the thermodynamic "distance" (2) and to the initial state, be it the starting time or the input into the reaction system. Both relationships include the history of the reaction up to a specific time point. In the case of the flow-through system, this is (the steady-state point and is) represented just by the steady-state reaction rate, i.e., a single value, which is, however, the result of the previous development of the reacting system before steady state had been achieved. In the case of a batch system, there is no single specific rate value and the history is expressed by the integral from the initial to the actual time. Note that the "flux-force" relationships ( 14) and ( 21), or (15) and ( 22), do not require the separation of the reaction rate into forward and reversed rates, nor do the reaction rates need to be expressed in mass-action form.</p><!><p>The steady state has yet another connotation in chemical kinetics. It is used also as an approximating tool in the description of complex reaction schemes, stating that the concentration of (reactive) intermediates is constant (and usually very low). The analyzed single reaction is too simple for the application of this tool and the easiest extension is the following two-step scheme:</p><p>A + C = AC (R1) AC = B + C (R2) in which AC is the intermediate and C can be viewed, for example, as a catalyst accelerating A to B conversion. Let us start with the batch system, which will give results quickly and easily. The balances are 9$ 9: ⁄ = − ? , (24a) 9$ 9: ⁄ = @ , (24b) 9$ A 9: ⁄ = − ? + @ , (24c) 9$ A 9: ⁄ = ? − @ = − 9$ A 9: ⁄ ,</p><p>where ? is the rate of (R1) and @ the rate of (R2). The steady state approximation for AC gives 9$ A 9: ⁄ = 0, from which it follows that also 9$ A 9: ⁄ = 0 and ? = @ . Thus, = − and the same situation results as for the simple transformation A = B above. Note that without the steady state approximation, ≠ − generally but = − ? , = @ . The nonstationary CSTR balance for AC in the scheme (R1)-(R2) is ' A ( + ? − @ = 9$ A 9: ⁄ + ' A .</p><p>(25) Because AC is an intermediate formed inherently during the reaction, it is quite reasonable (perhaps even necessary) to assume that there is no input or output of it. The steady state approximation then gives ? − @ = 0. Let us denote the common value of the two reaction rates by ; the remaining balances are then:</p><p>' ( − = 9$ 9: ⁄ + ' (26a) ' ( + = 9$ 9: ⁄ + ' (26b) ' A ( = 9$ A 9: ⁄ + ' A (26c) Thus, the accumulation of C is only due to the difference between its input and output. In the reactor steady state, the time derivatives in ( 26) vanish and the analysis given above for the simpler case of A = B remains valid because equation (26c) does not affect equations ( 26a)-( 26b). Note that if also the "catalyst" C is not pumped into or out of the reactor, the steady state approximation is automatically valid for it as well.</p><!><p>Yet another and quite general approach to the "flux-force" topic is provided by continuum nonequilibrium thermodynamics, as developed by Samohýl for chemically reacting mixtures of linear fluids (Pekař and Samohýl, 2014). Linear fluids appear in many systems commonly encountered in chemistry. The theory is explained in the referenced book and, here, only results relevant for the discussed topic and example are given. First, the theory derived what had for a long time been a matter of empirical knowledge in phenomenological kinetics (Pekař, 2010) -the fact that the reaction rate is generally a function of (only) temperature and concentration: C = C( , D).</p><p>(27) Here, C is the vector of reaction rates, whose components are the rates of independent reactions, C = ( ? , @ , … , F ), and D is the vector of molar concentrations. The function is not dependent on the reactor type in which reactions occur nor is it limited to a steady-state. Note that the theory generally does not need the assumption that the reaction rate is the difference between the forward and backward rates. To find some "flux-force" relationship, it is necessary to transform this function to a function of Gibbs energy, or, equivalently, of affinity. Samohýl's method takes two important facts into account. First, the transformation should be mathematically correct; second, the transformation to affinity should conform to stoichiometric constraints.</p><p>The function ( 27) can be easily transformed to a function of chemical potentials if we accept the widely used relationship between concentration and chemical potential (in ideal systems): G = G + ln$ (unit standard concentration used but not emphasized). It should be understood that this is a rather specific formula which simplifies the general finding that the chemical potential of a component is a function of the concentrations of all components. The standard chemical potential G is not dependent on concentration. We thus have: I).</p><p>(28) As shown by Bowen (1968), the linear algebra of reaction stoichiometry results in the conclusion that the vector of chemical potential generally decomposes into two perpendicular vectors:</p><p>where vector J is the vector of chemical affinities or reaction Gibbs energies (its components are</p><p>where R refers to the independent reaction, O MS is the stoichiometric coefficient of the component in reaction R, and T is the number of components 1 ) and vector K is the vector of constitutive affinities (Pekař, 2005). Using the decomposition (29), the final transformation is given by: C = C( , I) → C = C( , J, K).</p><p>(30) Thus, the reaction rate cannot be expressed as a function of chemical affinity only -in other words, as a function of a single affinity (or Gibbs energy). The constitutive affinity reflects the atomic composition of the components of the reacting mixture through its specific combination of chemical potentials (Pekař, 2005;Pekař and Samohýl, 2014). The second equation in (30) can thus be interpreted as pointing to the fact that not only the chemical potentials of the reactants and products (which are combined in J) but also the way in which atoms are combined in the reactants and products (which is reflected in K) participate in the driving force.</p><p>Let us demonstrate the transformation just on the discussed example of a single reaction A B.</p><p>The full derivation of all equations is given in our recent work (Pekař, 2018) and only the results important for this tractate are presented here. There is only one independent reaction with the (first order) rate equation (Pekař, 2018): = U ? ($ − ? $ ).</p><p>(31) Its transformation to the function of affinities is given by: =</p><p>(32) The italicized symbols L and V represent the two affinities (not the two components). Equation (32) shows a rather complex relationship between the reaction rate and the (thermodynamic) driving force in which both affinities are involved. It can be modified to the following "condensed" form:</p><p>= U exp6(L + 2V) 2 ⁄ 7 6exp(−L ⁄ ) − 17 (33) (U = U ? exp(−G ⁄ )). Equations ( 32) and ( 33) are analogs of Eq. ( 1), which could be rewritten in the form ⁄ = exp(−∆ ⁄ ) ≡ exp(−L ⁄ ), and show that both affinities are involved in what could be called the "driving force" in (32) and (33). Besides affinities (thermodynamic factors) these equations also contain the kinetic factor (U ? ) as should be expected. Perhaps the expression in square brackets (the second in the case of (33)) could be called the "principal" or "leading" driving force, because it is this expression which ensures a zero reaction rate in equilibrium (where L = 0) and resembles the expression exp(−∆ ⁄ ) appearing in Eq. ( 1), which was the motivation behind this paper.</p><p>The two affinities in this example are as follows (Pekař, 2018)</p><p>) Thus, L + 2V = 2G and the first exponential in Eq. ( 33) contains only the chemical potential of B (the product). This is natural because this very simple reaction represents some isomerisation and still the same atoms are combined in the isomers. Consequently, the role of the constitutive affinity in the driving force mentioned after Eq.(30) above is not so important here. This approach does not require forward and reversed reaction rates either.</p><!><p>An alternative search for the reaction driving force can focus on chemical potentials and not just on the (Gibbs) energy. The method involving the continuum non-equilibrium thermodynamics of linear fluid mixtures gives the following equation for the discussed reaction:</p><p>The term in the last square brackets can be viewed as the (thermodynamic) driving force, which is simply and naturally rooted in the difference in the (exponentials of the) chemical potentials of the reactants and products. If the chemical potential of reactant (A) is higher than that of the product, the reaction rate is positive as required. When the two potentials are equal, the reaction rate vanishes -this is the equilibrium state. Note that (35) still contains the kinetic factor (U b ). Very similar conclusions were reported for a simple enzymatic reaction (Pekař, 2015). Efforts to formulate the thermodynamic driving force in terms of energy are probably motivated by the general use of energy in the descriptions of the causes and directions of natural process -natural processes are usually described as being "propelled" by a high energy, running in the direction of decreasing energy and settling at the point of minimum energy. Perhaps chemical reactions are better described in terms of chemical potential, which is a (compositional) derivative of energy -and chemical reactions are processes changing the composition.</p><!><p>It seems important to differentiate between the identification of a thermodynamic driving force for a chemical reaction and the finding of a relationship between this driving force and the reaction rate. Such relationships should contain, besides the driving force and general parameters like temperature or the universal gas constant, also some kinetic factor(s).</p><p>The driving force behind a chemical reaction can probably be seen in terms of the difference between the energetic states of its reactants and products. Relating the driving force to the reaction rate ("flux") was achieved by combining the concentration dependencies of both the force and the rate. Thus, there need not be a single universal "flux-force equation" but rather diverse relationships for specific concentration dependencies and reacting systems. In other words, these are not expressions of the cause-effect type but rather the results of the transformations of functions (i.e., the results of changes of independent variables) and express a form of "bookkeeping" in a reacting system.</p><p>The relationships derived from the non-equilibrium thermodynamics of linear fluids, (32) or ( 35 (36)</p><p>The resistance, or, more properly, its reciprocal value, represents the kinetic factor, whereas the driving force represents the thermodynamic (energetic) factor. The thermodynamic factor can be formulated in terms of the reaction Gibbs energy (equivalently, affinities) or the chemical potentials of the reactants and products.</p>

ChemRxiv

A Predictive Model for Additions to N-Alkyl Pyridiniums

Disclosed in this communication is a thorough study on the dearomative addition of organomagnesium nucleophiles to N-alkyl pyridinium electrophiles. The regiochemical outcomes have observable and predictable trends associated with the substituent patterns on the pyridinium electrophile. Often, the substituent effects can be either additive, giving high selectivities, or ablative, giving competing outcomes. Additionally, the nature of the organometallic nucleophilic component was also investigated for its role in the regioselective outcome. The effects of either reactive component are important to both the overall reactivity and site of nucleophilic addition. The utility of these observed trends is demonstrated in a concise, dearomative synthesis of a tricyclic compound shown to have insecticidal activity.

a_predictive_model_for_additions_to_n-alkyl_pyridiniums

<!>39, vide infra).<!>Scheme 2. Concise synthesis of insecticide 3.<!>ASSOCIATED CONTENT Supporting Information

<p>Heterocyclic compounds containing nitrogen atoms are ubiquitous in the realms of natural product isolation, materials chemistry, and drug discovery. [1][2][3][4][5] In particular, as 59% of all marketed small molecule drugs contain a nitrogen heterocycle, 6 direct and concise synthetic access to diverse libraries of these important moieties remains an outstanding challenge. 7 Furthermore, being that the most common nitrogen heterocycles in marketed drugs are piperidine and pyridine (ranked first and second, respectively), easily functionalizing and modifying their structures is of the utmost importance to expediting therapeutic campaigns. [8][9][10] Shown in Figure 1A are a selection of medicinally important natural products and rationally designed therapeutic agents that all contain variously substituted piperidines or tetrahydropyridines. While manzamine A (1) 11 and morphine (2) 12 contain stereochemically dense patterns within and contiguous to their piperidine rings, compounds like 3 13 and Paroxetine (4), 14 while arguably less complex, still pose several synthetic questions regarding their construction. The control of relative and absolute stereochemistry is a consideration for all of these targets (1-4). Furthermore, the regiochemical control of desired unsaturation (e.g. 3) poses an issue, 15 especially when the final product has sensitive functionality. 16 Many strategies exist for the construction of piperidines, some of which are highlighted in Figure 1B. 17,18 Certainly, utilizing nitrogen's innate nucleophilicity dominates most disconnection strategies where substitution or reductive amination tactics are common. Additionally, hydroamination tactics that encourage N-C2 disconnections are also employed, but these approaches are often beset by arduous construction and prefunctionalization of linear precursors. 19 The reduction of lactams and imides in higher oxidation levels also remains a popular avenue for achieving the desired saturation, but controlled reduction of these carbonyls can be challenging in a complex setting, lending need for protecting groups or circuitous redox adjustments. 20,21 Thus, dearomative functionalization arose as an attractive tactic for the conversion of parent pyridine heterocycles to their reduced piperidine analogs. [22][23][24][25][26] This type of disconnection is not without its challenges, however, as derivatization of pyridines requires initial activation to a pyridinium 27 followed by a predictable nucleophilic addition. Prior to this study, organometallic additions to some monosubstituted pyridiniums have been established, but the innate selectivity of addition to multiply-substituted substrates has not been well understood. 28 Notable contributions made by Comins, [29][30][31] Yamaguchi, 32,33 and Donohoe 34 have laid some basic groundwork in this area. Nevertheless, understanding the innate electrophilic regioselectivity of substituted pyridiniums stands as a significant problem towards practitioners harnessing dearomative strategies. In considering factors such as steric, electronic, and chelation effects, this study aims to identify rules and guidelines understanding the regiochemical outcomes of additions to N-alkyl pyridiniums, enabling practitioners to more accurately predict the site of nucleophilic addition to these valuable heterocyclic synthons.</p><p>Activation of pyridines via acylation or alkylation of the nitrogen has been a realm of research for several decades. 15 The study of acyl pyridinium reactivity and regioselectivity has dwarfed the studies of their N-alkyl counterparts, largely due to product stability and/or enhanced reactivity. Many have invoked the acyl/iminoyl group as a non-innocent activating group capable of guiding [35][36][37] and/or sterically controlling [38][39][40] reactivity upon the event of dearomatization. In these examples, the innate reactivity of the pyridinium intermediate is perturbed, giving a more engineered and programmatic outcome. Conversely, with an interest in pursuing targets such as 1-4, exploiting and controlling alkylated pyridiniums would provide a more direct and ideal route for their preparation. Initially, the activation of unsubstituted pyridine and methyl nicotinate, via activation with methyl triflate, was investigated (for ease and solubility purposes) followed by subsequent treatment with 3-butenylmagnesium bromide. By crude 1 H NMR, Grignard addition to the methylated pyridinium gave a near statistical mixture of possible products with a C2:C4 ratio of roughly 2:1. By contrast, executing the same reaction with methyl nicotinate gave the C6 adduct with much higher selectivity of 1:1.6:8.9 of C2:C4:C6 (see Figure 1C). Clearly, the effect of the ester substituent greatly biased the addition and served as a starting point for our study to evaluate the various effects of pyridine substituents on the regioselective preference of nucleophilic addition.</p><p>In Scheme 1A, the results of nucleophilic additions to various pyridiniums are outlined. The general reaction employs first the treatment of a parent pyridine with methyl triflate at rt, followed by the subsequent addition of 3-butenylmagnesium bromide solution at -78 °C (for optimization of conditions, see SI). For simplicity, methyl was chosen as the N-alkyl substituent so that the innate reactivity of the pyridinium was observed with the least steric perturbance, but rationally, many other alkyl substituents can be employed, with similar reactivity trends (e.g.</p><!><p>It is worth noting that the only observed electrophilic sites for addition are the 2, 4, and 6 positions of the pyridine. For monosubstituted pyridines 9-12, various substituent directing effects were observed. Esters and nitriles induced a para-directing effect (see 10 and 11), while the diethyl amide and bromide directed addition to the ortho positions. It is hypothesized that the amide's observed selectivity arises from chelation, while the bromide's effect is electronic in nature. The combined effects of these substituents on regioselective addition were then evaluated. First, bromides 14 and 15 showed preference for C4 owing to an additive directing effect of the C5 alkyl group. Dibromide 16 also showed the same contrasteric preference for C4 addition, albeit with reduced selectivity (C4:C2 = 2.8:1). For 5-bromonicotinitrile (13) and methyl 5-bromonicotinate (22), the combined directing effect resulted in preferential addition at C6. When other nicotinates were evaluated (18-21), the same selectivity was observed as with 22 (C6 preferred). However, the substituent at the 5 position had a drastic effect on the C4/C6 regiomeric ratio. Selectivity was enhanced with more inductively withdrawing substituents (OMe, F, Cl), and, similarly, alkyl substitution (Me, e.g. 18) also served to enhance preference for C6. Diester 17 also displayed reasonable selectivity for C6, despite lacking an ortho directing substituent (vide supra). With regard to nicotinamides, the original ortho directing effect was hypothesized to be a result of organometallic chelation. When evaluated further, 5-bromomicotninamide 23 showed preference for C4 as ester-amide 24 showed marginal selectivity for C6. Diamide 25 showed preference for C2 similar to the observed selectivity for monoamide 12.</p><p>The scope of Grignard addition was largely evaluated using pyridine 22 as a model substrate (see Scheme 1B). Notably, this substrate only gives addition at C4 and C6, but the degree of selectivity is reliant on the Grignard reagent. Initially with 3butenylmagnesium bromide, an exceptional yield (90%) and regiomeric ratio were observed favoring C6 with 8.5:1 selectivity giving 26. 41 A similar reactivity profile was observed when using phenethylmagnesium bromide (see 27). Surprisingly, acetal 28 was produced with a much lower C6 selectivity (2.6:1), which was hypothesized to be a result of internal chelation of the acetal oxygen to the Mg atom in the nucleophile. This cyclic nature of the nucleophile is thought to make it more sterically hindered than the simpler alkyl Grignard reagents used to forge 26 and 27. This steric effect was exacerbated when employing dithianylmagnesium bromide to deliver 29 with nearly equal amounts of the C4:C6 isomers observed. This indiscriminate selectivity was also coupled with poor yield due to the its low reactivity. Decreased yields were also observed using allyl and propargylic 42 nucleophiles (30 and 31), albeit with modest selectivity for C6 in 30 (3.9:1). With sp 2 hybridized Grignard reagents, varying regioselectivities were also observed based on size or substitution pattern. While vinylmagnesium bromide had a high C6 preference (10:1), simply incorporating a methyl group into the nucleophile decreased selectivity substantially. Utilizing 1-Propenylmagnesium bromide gave 33 with 2.2:1 C6:C4 selectivity and the addition of 2-propenylmagnsium bromide gave an even lower selectivity of 1:1. Addition of phenylmagnesium chloride gave a similar ratio of 1.3:1 slightly in favor of C6 (see 35). Again, larger nucleophiles, albeit sp 2 hybridized, gave poorer C6:C4 selectivity. In stark contrast, sp hybridized Grignard reagents gave high selectivities (>20:1) for C6 (36 and 37), which aligns with previously observed reactivity with acylated pyridiniums. 32,33 Small modifications to the electrophilic partner were also evaluated for comparison. Combination of the tert-butyl bromonicotinate and acetal Grignard reagent delivered 38 slightly favoring C6 in 60% yield, with slightly diminished selectivity compared with 28. Activation of pyridine 22 with benzyl triflate led to adduct 39 in 67% yield and a C6:C4 ratio of 2.4:1. This decrease in selectivity for C6 compared to the analogous methylated product 26 was expected due to the increased size of the benzyl substituent. 43 Finally, the Grignard reagent of ethyl 2-iodobenzoate was added in to the methylated pyridinium of methyl nicotinate (11) with exclusive regioselectivity for C6 (>20:1). This substrate was important as it had been previously studied, however, no explanation was giving for the observed regioselectivity trends. 44 Scheme 1. (A) Regioselectivity study for addition of 3-butenylmagnesium bromide to substituted pyridiniums. Unless specified otherwise, the isolated yield is of all respective isomers (C2, C4 and C6). B) Regioselectivity study for the addition of Grignard reagents to alkylpyridiniums of pyridine 22. Reaction conditions: MeOTf (1.0 equiv), Et2O, rt; Grignard reagent (1.0 equiv), THF, -78 °C. The regiomeric ratio was determined by 1 In order to aid practitioners in effectively harnessing pyridinium-based retrosynthetic analysis, a directing group guide and user's roadmap has been delineated in Figure 2 deduced from this study of 3 and 3,5-disubstituted pyridiniums. First, understanding the various directing group effects is paramount to employing a predictive reactivity model (Figure 2A). As unveiled in this study, chelating groups such as amides have an ortho directing effect presumably due to chelation of the organomagnesium nucleophile. 45 Secondly, resonance electron-withdrawing groups like nitriles and esters primarily promote conjugate reactivity to the para position and secondarily to the ortho positions. 28 Lastly, alkyl groups, halides, and heteroatoms (e.g. N,O) promote ortho addition to the pyridinium. 46 It should be noted that the use of other organometallic species (e.g. RCu, R2Zn) are not expected to necessarily follow the above trends as this study concerns the use of Grignard reagents as nucleophiles (See Supporting Information for details).</p><p>An example of how to utilize these additive effects in a stepwise fashion is depicted in Figure 2B. Given bromonicotinamide 23, step 1 involves the identification of C2, C4, and C6 as innate electrophilic sites. Next, step 2 involves identifying chelating directing groups that promote proximal addition to ortho sites. Third, conjugate sites are then determined by the presence of resonance electron-withdrawing groups. Relative to 23, this preferential reactivity trend is not applicable. Lastly, the ortho directors are evaluated, which in this case, the bromine promotes addition at C4 and C6. Adding these effects together reveals a predicted selectivity for C4 being the highest preference with C2 and C6 being minor products of dearomative addition. Indeed, the experimental results indicate this trend in which the N-methylpyridinium of bromoamide 23 gives a 1.4:5.8:1 selectivity for C2:C4:C6. In general, these reactivity trends hold true in the present study, with regard to preferential addition. Variability is seen based on nucleophile identity, where larger nucleophiles seem to be less selective in terms of the product ratios (see 28, 29, and 38). Importantly, while selectivity can be variable in this sense, the major product observed still adheres to the guidelines depicted in Figure 2B. In order to evaluate the synthetic utility of this reactivity study, the construction of 3 was investigated. 13 This compound has shown notable insecticidal activity and has shown affinity towards binding the 5HT7 serotonin receptor. First, methylation of 22 and addition of phenethylmagnesium bromide delivered the dearomatized addition product. This compound was then subsequently reduced with LiAlH4 to give 42 on gram scale in 69% yield over 2 steps. A radical annulation utilizing Bu3SnH and AIBN afforded despyrrole methyl lysergate (43) in 21% yield (43% based on recovered starting material). This 3-step procedure compares favorably to the prior synthesis, which required 7 steps to synthesize this target. [47][48][49][50] Ester saponification with aqueous HCl followed by an amidation protocol with CDI gave 3 in 30% yield over 2 steps and in 5 total steps (previously 13 steps). It is worth noting that the bromide serves both as a directing group for the innate selectivity of the dearomatization and as a handle for C-H annulation. The selectivity of the hydride reduction step is also remarkable, giving dihydropyridine 42, in that very little ester reduction is observed and the reduction regiochemistry is absolute.</p><!><p>In conclusion, this study has shown the innate reactivity trends associated with N-alkyl pyridinium electrophiles. The effects of substitution patterns have revealed outcomes in organometallic additions that are systematic and easily employed in first-pass retrosynthetic analyses. These data serve to delineate reactivity that, prior to this study, were limited and not conclusively understood. In some instances of the current work, the selectivity proved quite high, and could be easily utilized in multistep synthesis, as shown by the concise construction of 3. It is hoped that the lessons learned will be broadly helpful to practitioners ambitious to utilize the inherent advantages of dearomative synthetic approaches. The expansion upon these findings and their application to future synthetic targets is anticipated to be reported by our group in due course.</p><!><p>The Supporting Information is available free of charge on the ACS Publications website.</p><p>Experimental Procedures and analytical data (1H, 13C NMR, MS) for all new compounds (PDF) Crystallographic data (CIF)</p>

ChemRxiv

Ligand Discovery from a Dopamine D3 Receptor Homology Model and Crystal Structure

G-Protein coupled receptors (GPCRs) are intensely studied as drug targets and for their role in signaling. With the determination of the first crystal structures, interest in structure-based ligand discovery has increased. Unfortunately, most GPCRs lack experimental structures. The determination of the D3 receptor structure, and a community challenge to predict it, enabled a fully prospective comparison of ligand discovery from a modeled structure versus that of the subsequently released crystal structure. Over 3.3 million molecules were docked against a homology model, and 26 of the highest ranking were tested for binding. Six had affinities from 0.2 to 3.1\xce\xbcM. Subsequently, the crystal structure was released and the docking screen repeated. Of the 25 compounds selected, five had affinities from 0.3 to 3.0\xce\xbcM. One of the novel ligands from the homology model screen was optimized for affinity to 81nM. The feasibility of docking screens against modeled GPCRs more generally is considered.

ligand_discovery_from_a_dopamine_d3_receptor_homology_model_and_crystal_structure

<!>Prediction of the Dopamine D3\xe2\x80\x93Eticlopride Structure<!>Docking for new D3 ligands<!>Ligand Bias in the Docking Screens<!>Ligand Selectivity<!>Optimization for Affinity<!>Functional Activity of the Docking Hits<!>Discussion<!>Homology Models<!>Molecular Docking Screens<!>Binding affinity and functional activity of the docking-predicted compounds

<p>GPCRs are a large family of membrane proteins that are critical for signal transduction. They have been a major focus of pharmaceutical research and are the primary targets of almost 30% of approved drugs1. All of these drugs were discovered without the aid of receptor structures by classical, ligand-based medicinal chemistry. Accordingly, many of these drugs reflect their origins as mimics of the natural signaling molecules. With the determination of the first drug-relevant GPCR structures in the last four years2-4, the opportunity for structure-based discovery of more novel scaffolds has arisen. Docking screens to these crystal structures have been unusually fruitful, with high hit-rates returning novel and potent ligands5-7. Still, the structures of most GPCRs remain undetermined. There are thought to be just over 360 pharmaceutically relevant GPCRs in the human genome8, and to date only five have had experimental structures determined, all by dint of extraordinary effort and innovation. For structure-based efforts to impact ligand discovery for most GPCRs, certainly in the near term, homology modeling of GPCR structures remains essential.</p><p>In the past, the structure of rhodopsin and, before that, bacteriorhodopsin9, were used to explore GPCR function and ligand recognition10-18. Several efforts to use homology models for ligand discovery, via docking, have also been undertaken19-25. With rare exceptions26,27, such docking screens use a hierarchy of pharmacophore filtering and ligand similarity to focus the molecules being docked. This will typically reduce an "unbiased" library by 10- to 100-fold to one more dominated by precedented chemotypes. Whereas this can be effective, such a combination of filtering and docking perforce removes unexpected chemotypes that a stand-alone, structure-based approach might otherwise find. Interestingly, two of these early studies included work on dopamine receptors, based on rhodopsin as a template20,21. Whereas both screens had high hit-rates, the pharmacophore filtering appears to bias the ligands discovered toward well-established chemotypes, a point to which we will return. More generally, the pharmacophore approach does not address those targets for which ligand information is weak, and does not illuminate how these models compare to what might be achieved with an experimental structure.</p><p>The opportunity to prospectively investigate how homology models compare to experimental structures for ligand discovery, and by extension to investigate what fraction of GPCRs might be exploitable for ligand discovery, emerged recently by way of a community challenge28. After the determination of the structure of the dopamine D329 and CXCR4 GPCRs in complex with antagonists (for D3, eticlopride, 1, Figure 1), the modeling community was asked to predict the structures of each complex before the coordinates were released. This provided an opportunity to not only predict the configuration of the single ligand bound to the complex, but also to use the homology model that emerged to discover new ligands, via structure-based docking screens, before the crystal structure was released. Once released, the same screen was prosecuted against the crystal structure. Since in each case the putative ligands would be tested for affinity, we could compare the two results to illuminate how successful the homology model was compared to the crystal structure in a situation where the predictions were truly blind.</p><p>We thus undertook the following calculations and experiments. Once we had submitted models of the D3/eticlopride complex, we turned to ligand discovery calculations. In these, over 3 million commercially available molecules were screened by docking to identify putative ligands that complemented the structure of the homology model. Before the crystal structure was released, 26 high scoring small molecules were purchased and tested for D3 receptor affinity (compounds 2-27). Finally, when the crystal structure was released several months later, a second docking campaign was prosecuted based on that structure; twenty-five more high scoring small molecules from this second screen were purchased and tested for D3 receptor affinity (compounds 28-52).</p><p>These calculations and experiments enabled the following investigation: Could a homology model—blinded to the (unknown) crystal structure—template the discovery of new ligands? How well would the homology model compare to the subsequent crystal structure, in terms of hit rate and affinity?</p><p>These questions thus directly addressed the possibility of using homology models for at least some of the vast majority of GPCRs whose structures will likely remain undetermined in the near term. More subtly, because the homology model was refined for its ability to enrich known D3 ligands, we wondered if ligands discovered against it would be biased toward known D3 chemotypes, and might therefore be less novel than those discovered using a crystal structure for docking-based discovery? From a chemical biology standpoint, we also wondered how ligand specificity would compare between the two screens. The template for the homology model was the β2 adrenergic receptor, and one might predict that the screened molecules might retain an activity for this target, or might be less specific than those screened against an experimental structure. Similarly, from a chemical probe standpoint, it is important to optimize for affinity, and we were unsure whether a homology model, selected for its ability to recognize general dopaminergic chemotypes, would be competent for such optimization. Here, we explored these questions by undertaking prospective docking screens against first a homology modeled structure of the D3 receptor and subsequently the crystal structure of the same receptor. Since the crystal structure was released after the screen was completed against the model structure, both screens were fully prospective. To our surprise, we found that the hit rates against both the modeled and experimental structures were not only high, but essentially equivalent. Notwithstanding the opportunities for bias toward known chemotypes in optimizing the homology model, both screens returned new scaffolds at a similar rate.</p><!><p>The results of the D3/eticlopride structure prediction and docking challenge have been reported elsewhere 28 but will be briefly summarized here as they influence what follows. We were tasked with predicting the structure of the D3/eticlopride complex, without knowing either. We used the docking enrichment of known ligands among the top scoring molecules, from among a large number of decoys, as a criterion of model accuracy26,30,31. Almost 200,000 homology models were built using the program MODELLER-9v832 templated on the β13 and β22 adrenergic receptor crystal structures, and elastic network models calculated by the program 3K-ENM33 also based on these two structures. The top ranked 2964 of these 200,000 models, judged by MODELLER's internal DOPE score34, were advanced for docking. Up to 1,300 known dopaminergic ligands, along with up to 110,000 property-matched decoys, were docked35. Modeled receptor structures were prioritized for their ability to highly rank the known ligands compared to the decoys in the screens. Whereas this demanded a substantial amount of docking — 98,700,000 complexes calculated overall — it was largely automated. Our top model enriched the known ligands versus the decoys by 32-fold over what is expected at random among the 1% top ranking molecules. This enrichment was substantially higher than that found for docking of dopaminergic ligands against the β1 and β2 adrenergic templates used for the modeling, where the enrichments were 2 and 1, respectively.</p><p>Eticlopride was docked into each of the top models; five were selected for submission to the D3/eticlopride structure prediction and docking28. As was true of the predictions from several groups, our predicted structures showed overall fidelity to the subsequently released crystal complex. Our highest-ranked model had an overall Cα RMSD of 3.4Å, with an eticlopride RMSD of 1.65Å and an RMSD of the orthosteric site residues of 1.65Å (Figure 1). Most of the key ligand interactions14 observed in the crystal structure29 were also observed in this model, including the salt-bridge between the aminergic nitrogen of eticlopride and the recognition Asp1103.32 (Ballesteros-Weinstein numbering36). Similarly, two internal hydrogen bonds in eticlopride were captured by both the model and the crystal structure37. Intriguingly, all of the models that both enriched known ligands and docked eticlopride correctly were based on the templates from elastic network backbones; the higher range of motion explored by such models presumably contributed to the ultimate fidelity of the model to the experimental result. The 3K-ENM model captured backbone movements in several helices (III, IV, V and VI) that influence the shape of the binding site, e.g. a 1Å movement of transmembrane helix III in the region of Asp1103.32.</p><!><p>With this receptor model in hand, we next turned to ligand discovery. Over three million commercially available compounds were screened for complementarity to the receptor model, using DOCK3.638,39. Each molecule was fit into the site in an average of 1170 orientations, relative to the receptor, and for each orientation an average of 789 conformations, (thus over 900,000 configurations in total). Each configuration was scored for van der Waals and electrostatic complementarity, corrected for ligand desolvation39; thus for the 3.1 million compounds screened about 2 trillion complexes were evaluated. Prior to the release of the crystal coordinates, 26 top ranking compounds (2-27), all among the top 0.02% of docking-ranked molecules, were selected for experimental testing. The top scoring docking hits were dominated by mono-cationic molecules that all appeared to ion-pair with the key aminergic recognition residue Asp1103.32; most had overall good van der Waals complementarity to the orthosteric site. In the selecting the particular molecules for experimental testing, we corrected for energetic terms not included in the scoring function such as high ligand internal energy and receptor desolvation (detailed in Supplementary Methods). As far as we know, none had been previously tested for activity against dopaminergic receptors. Six of these, a hit-rate of 23%, bound to D3 measurably, with affinities ranging between 200 nM and 3.1 μM (Table 1, Figure 2, Figure 3 and Supplementary Results, Supplementary Figure 3). The similarity of the new compounds versus dopamine receptor ligands was assessed by calculating the Tanimoto coefficient (Tc) to the 10400 D3 annotations in the ChEMBL database (http://www.ebi.ac.uk/chembl). Four of the active compounds (2, 4, 5, and 7) resembled known ligands, with ECFP_4-based Tc values greater than 0.4. The two other active ligands, compounds 3 and 6, were topologically dissimilar to dopamine receptor ligands in the ChEMBL database (best Tc<0.35 by ECFP_4 fingerprints; novelty was also observed using Daylight fingerprints in Supplementary Table 2). These therefore appeared to be novel chemotypes for the D3 receptor.</p><p>With access to the crystal structure (PDB code 3PBL)29, we then carried out a second docking screen of the 3.6 million lead-like molecules from ZINC40. Unlike the homology model, where side chain positions were optimized to enrich known ligands, the crystal structure heavy atom positions were unmodified. We selected 25 molecules (compounds 28-52) from among the top 0.02% of the crystal structure based screen for experimental testing. Five of these, molecules 28 to 32, were active, with Ki values between 300 nM and 3 μM, a hit-rate of 20% (Table 2, Figure 3, Figure 4 and Supplementary Figure 3). Whereas two of these, 28 and 30, resembled previously known scaffolds, and compound 32 was of intermediate similarity, two others, 29 and 31, represented novel scaffolds. Intriguingly, though 29 explored new substituents distal to the aminergic group, its aryl-amide core resembled that of eticlopride; such a chemotype was not observed among the actives from the homology-model screen. We note that compound 6, chosen from the homology model screen, also scored among the top 0.04% of the docking prioritized molecules from the crystal structure screen.</p><!><p>The homology model had been selected based on its ability to enrich known dopaminergic ligands and, in retrospective calculations it enriched known ligands substantially better than the crystal structure. Among those chemotypes most strongly enriched were phenyl-piperazines, which are characteristic for this target. The difference in enrichments between the screens was reflected in the compounds that were highly ranked in the prospective calculations. Overall, the overlap of the top 1000 docking hits from the two screens was only 90 molecules, two of which were selected for experimental evaluation; both were inactive. Eight of the compounds purchased for experimental testing from the homology model screen closely resembled known dopaminergic ligands, with ECFP_4 Tc values greater than 0.45 to annotated ligands in ChEMBL, but only four of the molecules purchased from the crystal structure screen had this level of similarity. The one instance where we observed a higher bias towards dopaminergic ligands from the crystal structure screen was in similarity to eticlopride itself. Indeed, nine of the molecules selected for testing had the aryl-amide-aminergic chemotype characteristic of eticlopride and its congeners, reflecting the many high-ranking molecules with this feature from the crystal structure screen. Conversely, only one compound from the homology model screen had this aryl-amide-aminergic chemotype. Notwithstanding these apparent biases going into experimental testing, novel chemotypes were ultimately confirmed for both screens.</p><p>The different ligands selected by the two screens reflect differences between the structures of the orthosteric sites in the homology model and the crystal structure. Whereas these two sites only differed by 1.65Å RMSD when superimposed, this was enough, when interrogated at a docking level, to change the identity, if not the nature, of the high-scoring docked molecules. The main difference between the two orthosteric sites is that the homology model is slightly more open and thus larger. For instance, the distance between the Cα atoms of Asp1893.32 and Ser1925.42 grew from 11.9Å in the crystal structure to 12.9Å in the model. More locally, Ile183 differs by 3.6Å between the two structures, while Val1895.39, Phe3456.51 and Phe3466.52 differ by 0.8 to 1.5Å. This overall opening reflects how the model was optimized: we docked ligands of all sizes to the model, and looked for enrichments. The model structure that was chosen could accommodate known ligands across a relatively wide size range, whereas the same site in the crystal structure more tightly encloses eticlopride, a relatively small ligand. Thus, many known phenyl-piperazine ligands that were enriched well by the model would clash with residues such as Val1895.39, Phe3456.51 and Phe3466.52 in the crystal structure.</p><!><p>An important challenge in dopaminergic receptor pharmacology is finding ligands that are specific for the D3 versus the D2 receptor. With few exceptions41, most D3 receptor ligands are also active on D2, making their use as chemical probes problematic. Methodologically, we were interested to learn if the new ligands derived from the homology-model docking retained activity for the β2 adrenergic receptor, from whose template the D3 model was derived. Active ligands were therefore counter-screened against the D2 and β2 receptors (Table 3). None had measurable activity against β2 AR at 10μM, suggesting that no significant template bias remained, and, also, that ligands specific for dopaminergic receptors had emerged. No effort was made to find D3 selective ligands, so achieving selectivity among the dopamine receptor subtypes would be fortuitous. Whereas most compounds showed little selectivity between D3 and D2, a few of the more novel scaffolds did, with affinities 6- and 20-fold better for the D3 over the D2 receptor for compounds 3 and 7, respectively (Table 3).</p><!><p>We were also interested in progressing a novel series for affinity, both as an end in itself and to explore whether model-based approaches could effectively guide this effort. Twenty analogs of compound 3, among the most dissimilar to known dopaminergic ligands, were found that had good complementarity to the D3 modeled structure (Table 4 and Supplementary Table 4, compounds 53 through 72). In its docked pose, the two hydroxyls and the aliphatic amine of compound 3 interacted with Asp1103.32 and Tyr3737.43. As it was difficult for these two residues to optimally interact with all the three ligand donors, we wanted to explore the possibility that the hydroxyls were not crucial for affinity. Commercially available molecules with this feature were extracted from the ZINC database40 and docked to the orthosteric site of the homology model. The docked poses were inspected and a set of analogs representing the diversity found in the database were selected for testing. All of these compounds retained the key Asp1103.32—cationic interaction, but explored variations in the hydroxyl groups and the substituents of the phenyl ring. In particular, we focused on compounds that preserved the meta-substituent on the phenyl ring, which fills a hydrophobic pocket formed by the side chains of residues Phe1643.28, Val1653.29, and Ser1684.57. Eleven analogs had substantially improved affinities, ranging from 4 to 20-fold better than the lead compound 3, with the most active reaching 81 nM (Table 4, Figure 2, Figure 3, and Supplementary Figure 3).</p><!><p>In previous docking screens against the GPCR crystal structures5-7, there has been a close correspondence between the function of the ligand co-crystalized with the receptor, either inverse agonist or neutral antagonist, and that of the docking hits. To explore whether this (presumably structural) bias was present in the D3 screens undertaken here, the docking hits (compounds 2-7 and 28-32) and several analogs of compound 3 (55-57 and 63) were investigated for agonism of the D3 and D2 receptors. With the possible exception of compound 28, which showed very weak partial agonism (Supplementary Figure 4 and Supplementary Table 3), none of these 15 compounds were agonists against either receptor and all appeared to function as antagonists. This corresponds with the known function of eticlopride, with which the receptor was co-crystalized and on whose binding mode the model was predicated.</p><!><p>The determination of the structures of pharmacologically relevant GPCRs2-4 has sparked intense interest42. A crucial question is not only how these structures may themselves be exploited for ligand discovery, but what is the range of homologous targets that they illuminate. An astonishing result of this study was that the docking screen against the homology model was no less effective than that against the crystal structure; we would have been satisfied with the converse answer. The hit rates for both screens were high, at 23% and 20%, and their affinity ranges fully overlapped, with several molecules with 200 to 300 nM affinity from each screen. These predictions were likely right for the right reasons: the ligand poses overlapped with those adopted by eticlopride in its D3 complex, and this complex was itself well-predicted in the original blinded challenge. Homology models of proteins have been previously used for ligand discovery, including for GPCRs19-27. What was unusual, and perhaps unique to this study, was that a docking screen was prosecuted prospectively against a homology model and then, subsequently, the crystal structure. The results were thus doubly unbiased—the crystal structure was unknown at the time of the docking, and what we ultimately compared were new, experimentally-tested ligands.</p><p>A concern we harbored was whether the active molecules from the model-based screen—assuming any would be found—would be highly biased toward known dopaminergic ligands. A criterion for selecting effective models was their ability to enrich known ligands, and it seemed possible, even likely, that any actives that emerged from such a screen against it would simply recapitulate known D3 ligands. Indeed, the high-ranking molecules from the homology-model screen more closely resembled known dopaminergic ligands than did those from the crystal structure screen. Some of this bias can be seen among the experimentally tested molecules: for the model, three aryl-piperazines were confirmed as active, whereas for the crystal screen one eticlopride-like ligand was confirmed, the latter consistent with this structure's own conformational bias. In this sense the concern regarding bias was justified and may affect future studies. In the end, however, the experimentally active molecules were no more biased toward known ligands in one screen than the other. Meanwhile, from each screen emerged two novel scaffolds, four overall, and these not only differed from known ligands but also differed among themselves. Just as compelling, the active molecules from the homology model screen showed no measurable affinity for the β2-adrenergic receptor, which templated its modeling (Table 3). The model thus appeared to not only have captured the broad similarity that exists among aminergic GPCR targets, but also to have represented features specific to the D3 receptor. The lack of overlap among the hits from the two screens, and the observation that many ligands that dock well into the modeled structure did not fit into the corresponding crystallographic site, may reflect the many low energy conformations that GPCRs sample, both active and inactive43. The modeled and the experimental structures may thus represent different but viable low energy D3 receptor conformations, both likely inactive ones. This was also consistent with the satisfactory fidelity of the original D3/eticlopride structure prediction that was the point of departure for this study (Figure 1).</p><p>The relatively high affinities of the docking hits undoubtedly reflected the bias, among even large commercial libraries, toward molecules resembling known aminergic GPCR ligands5,7. It was encouraging that novel chemotypes could nevertheless be discovered. This is, after all, the promise of the structure-based enterprise: that based on complementarity to a protein structure novel ligands can be discovered and from these novel biologies might emerge. In this sense, it was instructive to compare the results of this study, which leveraged complementarity to a modeled structure alone in ligand selection, with an study that used a pipeline of pharmacophore filtering for dopaminergic chemotypes followed by docking20. Whereas this earlier study was in many ways path-breaking, and the hit rates and affinities achieved were high, the molecules discovered were typically much more similar to known dopaminergic ligands than those found here. This can be seen by inspection of the structures and comparison to the previously known ligands (Supplementary Table 5), or more quantitatively by considering the ECFP_4 Tanimoto coefficient values (Tc values). The Tc values averaged 0.55 to the most similar known dopaminergic, whereas, for the molecules discovered here, the average Tc value to the nearest known dopaminergic was 0.42, a large difference for this fingerprint that is born out by visual inspection (Table 1 and Supplementary Table 5).</p><p>Admittedly, the Ki of compound 3, which was among the most novel, was only 1.6 μM, probably too high (poor) to be useful as a probe or lead. As a new chemotype for this target, we wondered if its affinity could be improved. Structure-guided analog exploration led to derivatives of 3 with up to 20-fold improved affinity (Table 4), which may owe to elimination one of ethanolic group and exploration of small groups on the meta-position of the aryl ring. The most potent of these analogs, compound 56, had a Ki 81 nM, much more in the probe range and as potent as many approved dopaminergic drugs. With only 20 non-hydrogen atoms, 56 is only slightly larger than a fragment and is thus far from optimized; its ligand efficiency of 0.49 is promising for further elaboration44. This series may merit further consideration as D3 receptor probes.</p><p>Apart from bias in the ligand libraries, docking hits against GPCRs have previously recapitulated the functional properties of the ligand with which the receptor was co-crystalized, presumably reflecting a bias in the receptor structure used. Thus, earlier structure-based screens against inactive structures of GPCRs found only antagonists and inverse agonists5-7. This was true here too, both the ligands discovered against the homology model and those discovered against the crystal structure were essentially all antagonists. This likely reflected the inactive D3 conformation selected, from among those sampled in solution, by eticlopride in the experimental structure, and the bias toward such a structure from the modeling of a conformation competent to recognize the drug in the homology model. Additionally, most of the known ligands chosen for docking against the homology model were also antagonists.</p><p>GPCRs are central to cell signaling and are key targets for medicinal chemistry. The determination of the structures of pharmacologically-relevant GPCRs illuminates why they are so fruitful for drug discovery—their orthosteric sites are particularly well-suited to accommodate small organic molecules. This, and the substantial bias of chemical libraries towards the ligands of these targets, explains the high-potency of hits emerging from structure-based, and indeed high-throughput screens against them5-7. In this sense, the docking screens against the D3 receptor reprise what we learned from those against the β2-adrenergic and A2A adenosine receptors—hit rates are high, as are the affinities of the hits.</p><p>What was new to this study, in addition to the particular ligands discovered, was the direct, prospective comparison of the ability of homology models of GPCRs to template ligand discovery. The model used here was fully blinded from the crystal structure, but it was ultimately as effective in prioritizing active D3 ligands as judged by the hit-rate, the potency, and the novelty of the new ligands. Whereas we did not expect this result, it was encouraging for the structure-based enterprise against GPCRs. These receptors have advantages for homology modeling: the conservation of the seven trans-membrane helixes, and several strongly conserved residues, e.g. the DRY and NPXXY motifs, allow registry in sequence alignment to be determined with greater confidence than is typically possible. At a conservative cut-off of 35% transmembrane sequence identity, the five structures determined to date resemble 59 other GPCRs45(Supplementary Table 6). Whereas each new GPCR crystal structure will provide a rich vein for ligand discovery, their luster may reflect on a much larger number of exploitable targets.</p><!><p>The initial alignment was generated using PROMALS3D46 using a sequence profile that included all dopamine receptor sequences as well as the β1 and β2 adrenergic receptor sequences (PDB: 2VT4(chain B)3 and 2RH1(chain A)2). The initial alignment was manually refined to correctly align the residues forming the conserved disulfide bonds (C103-C181 and C355-C358). Alternative alignments of the extracellular loop 2 (EL2), which contacts the binding site14-16 were evaluated, resulting in the final alignment (Supplementary Figure 2). All homology models were built with MODELLER-9v832. The models were based on two types of templates: (1) the crystal structures of the β13 and β22 adrenergic receptors and (2) 710 elastic network models produced by 3K-ENM33, based on each of these two crystal structures. This led to almost 200,000 homology models. These were scored using DOPE34 resulting in 2964 models scoring well by modeling criteria, 4 from each 3K-ENM backbone, 64 from each crystal structure backbone.</p><p>The 2964 models were then evaluated for their ability to enrich known ligands among a large number of decoys. The models were ranked based on their adjusted logAUC and the enrichment factor at 1% (EF1) of the database39. Models had to score in the top quartile for logAUC and EF1, and more than 60% of the best scoring ligands had to form the conserved salt-bridge interaction with Asp1103.32, to be considered. The conformational sampling of eticlopride was restricted to conserve the internal hydrogen bonds observed in the Cambridge Structural Database47. Before the release of the crystallographic structure, five modeled eticlopride/D3 structures were submitted to the GPCRDOCK2010 competition28Of these, models #1 and #4 had ligand poses and orthosteric residue positions that resembled that of the crystal structure (to 1.65Å or better).</p><!><p>A version of DOCK3.5.54 with an improved treatment of ligand solvation and with improved speed, DOCK3.638,39 (http://dock.compbio.ucsf.edu/), modified with scripting drawn from DOCK Blaster (http://blaster.docking.org) was used in docking calculations against the homology model and the crystallographic structure of the dopamine D3 receptor (PDB 3PBL29). The flexible-ligand sampling algorithm in DOCK3.6 superimposes atoms of the docked molecule onto binding site matching spheres, which represent favorable positions for individual ligand atoms. Forty-five matching spheres were used; for the crystal structure these were derived from the position of eticlopride while the spheres for the homology models were derived from overlaid docking poses of known ligands. The degree of ligand sampling is determined by the bin size, bin size overlap, and distance tolerance, set to 0.4Å, 0.1Å, and 1.5Å, respectively, for both the matching spheres and the docked molecules. Complementarity of each ligand pose is scored as the sum of the receptor–ligand electrostatic and van der Waals interaction energy, corrected for ligand desolvation39. The best scoring conformation of each docked molecule is then subject to 100 steps of rigid-body minimization. Partial charges from the united atom AMBER force field were used for all receptor atoms except for Ser1925.42, Ser1935.43, and Ser1965.46, for which the dipole moment was increased as previously described5. From the ZINC lead-like set of commercially available molecules, over 3 million compounds were docked. Prior to selecting compounds for experimental testing, the hit list was filtered to remove a previously known high internal energy motif that results in unreasonably favorable docking scores48, using automated scripts. The rankings reported here reflect this filtering (see Supplementary Methods for details).</p><!><p>Affinities for D3, D2 dopaminergic and β2 adrenergic receptors were determined by radioligand competition binding at the NIMH Psychoactive Drug Screening Program49. Briefly, crude P2 (21,000 × g) membrane preparations were prepared from cell lines transiently expressing recombinant human GPCRs at about 50μg protein/microliter of 50mM Tris, 1% BSA, pH 7.4 (assayed by Bradford using a BSA standard). 50μL of membrane suspension were added to the wells of a 96-well plate containing 100μL of binding assay buffer, 50μL of radioligand present at five times its Kd, and 50μL of candidate ligand at a concentration five times that desired in the assay (Supplementary Table 1). Reactions were incubated for 60 to 90 min at room temperature in the dark and then harvested onto 0.3% PEI-treated GF/A filtermats (Wallac). After three washes with ice-cold wash buffer (50 mM Tris, pH 7.4), filter mats were dried in a microwave oven and impregnated with Meltilex scintillant (Wallac). Residual radioligand binding, measured by scintillation using a TriLux microbeta counter (Wallac), was plotted as a function of competitor and regressed using "one-site competition" in Prism4.0 (GraphPad) to obtain IC50 values. Ki values were calculated from the IC50 values using the Cheng-Prusoff approximation.</p><p>To investigate the functional activity of the new ligands (i.e., agonism or antagonism) at D2 and D3 receptors, we measured recruitment of β-arrestin2 to agonist-occupied receptors using the Tango assay50 (summarized here, see Supplementary Methods). HTLA cells were transfected with plasmid encoding either the hD2V2 Tango receptor or the hD3V2 Tango receptor. As a negative control, cells were transfected with pEYFP-N1 (Clontech). Subsequently, the cells were trypsinized, resuspended to 1 × 104 cells/50μl growth medium, and seeded in poly-D-lysine-coated glass-bottom 384-well plates (Costar). The next day, the medium was replaced with serum-free DMEM (Cellgro), and the cells were stimulated with reference agonist (Quinpirole), reference antagonist (Chlorpromazine), or test compounds. Assay concentrations of all compounds ranged from 3pM to 30μM. After an overnight incubation with reference or test compounds, the medium was removed and replaced with 1× BriteGlo (Promega). Luminescence was counted using a TriLux (PerkinElmer) plate reader. Quinpirole (Sigma-Aldrich), Chlorpromazine (Sigma-Aldrich), and the test compounds were all inactive on HTLA cells not expressing a Tango receptor. In additional control experiments, HTLA cells were transfected with a plasmid encoding the human V2 Tango receptor50; Quinpirole and Chlorpromazine had no effect on HTLA cells expressing this receptor.</p>

PubMed Author Manuscript

Peptide length and prime-side sterics influence potency of peptide phosphonate protease inhibitors

Summary The ability to follow enzyme activity in a cellular context represents a challenging technological frontier that impacts fields ranging from disease pathogenesis to epigenetics. Activity-based probes (ABPs) label the active form of an enzyme via covalent modification of catalytic residues. Here we present an analysis of parameters influencing potency of peptide phosphonate ABPs for trypsin-fold S1A proteases, an abundant and important class of enzymes with similar substrate specificities. We find that peptide length and stability influence potency more than sequence composition and present structural evidence that steric interactions at the prime-side of the substrate-binding cleft affect potency in a protease-dependent manner. We introduce guidelines for the design of peptide phosphonate ABPs and demonstrate their utility in a live-cell labeling application that specifically targets active S1A proteases at the cell surface of cancer cells.

peptide_length_and_prime-side_sterics_influence_potency_of_peptide_phosphonate_protease_inhibitors

Introduction<!>2.1 Improvement to the synthesis of peptide phosphonates<!>2.2 Increasing peptide length improves phosphonate inhibition of serine proteases<!>2.3 Structure of MT-SP1 bound to Bz-DPP<!>2.4 Labeling of cell surface proteases by ABPs<!>Discussion<!>Materials<!>Diphenyl-[N-(benzyloxycarbonyl)amino](4-ethylesterphenyl)methane phosphonate hydrochloride (10)<!>Diphenyl amino(4-amidinophenyl)methanephosphonate, trifluoroacetic acid salt (12)<!>Coupling of the biotinylated peptide to 12. General procedure<!>Biotin-PEG-Gln-Arg-Val-Bz-DPP\xc2\xb7TFA (1)<!>Biotin-PEG-Leu-Thr-Pro-Bz-DPP (2)<!>Biotin-PEG-Gly-Ser-Gly-Bz-DPP (3)<!>Biotin-PEG-Glu-Pro-Ile-Bz-DPP (4)<!>Biotin-PEG-Bz-DPP (5)<!>Biotin-PEG-Val-Bz-DPP (6)<!>Biotin-PEG-Arg-Val-Bz-DPP (7)<!>Biotin-PEG-Phe-Thr-Gly-Ser-Gly-Bz-DPP (13)<!>Inhibition assays<!>Crystallization<!>Structure resolution and refinement<!>Western blot labeling analysis<!>Fluorescent substrate synthesis<!>Cell culture and propagation<!>Flow cytometry<!>Microscopy<!>

<p>The creation and implementation of high-throughput nucleic acid analysis techniques have revolutionized medicine and biology. The developments of affinity capture and protein binding arrays have allowed similar analyses of protein levels and interactions. However, such gene and protein profiling data does not always reflect the dynamic milieu that one finds at the cellular level in vivo. Furthermore, enzymatic activity adds an additional layer of complexity in functionality that cannot currently be addressed using available techniques.</p><p>The need to examine enzyme activity is particularly relevant to the study of proteolytic enzymes in biological systems. While synthetic substrates can be used to study individual proteases in vitro, their efficacy in studying multiple proteases in a complex mixture is limited. For example, the caspase family of cysteine proteases have highly overlapping substrate specificities, thus multiple caspases can often cleave the same synthetic substrate (McStay et al., 2008). This complicates functional analyses of individual caspases during apoptosis. However, covalent labeling of active proteases allows one to monitor individual proteases in the context of the whole cell. Technologies based on covalent inhibitors are emerging for activity-based proteomics. Phosphonate inhibitors of serine proteases form the basis of one such technology (Sienczyk and Oleksyszyn, 2009).</p><p>Diisopropylfluorophosphonate (DFP) is one of the earliest irreversible inhibitors described for serine proteases, and radiolabeled DFP was one of the first activity-based probes (ABPs) described (Powers et al., 2002). More recent studies have modified DFP to include a detection tag for purification and/or visualization purposes (Liu et al., 1999). Biotin and fluorophores are the most common tags used, and active enzyme profiles of several types of human tissues and tumor types have been created using such molecules (Jessani et al., 2005). While these efforts are valuable to the field, these studies identify a preponderance of serine hydrolases that are not proteases, and ABPs specific for proteases have been difficult to achieve.</p><p>The specificity of DFP can be improved in two ways. Replacing the highly electronegative fluorine atoms with two phenoxy groups reduces reactivity and increases stability (Powers et al., 2002). Further modifying these functional groups modulates the electrophilicity and shape of the phosphonate reactive center, which can add selectivity to the compound. For example, incorporating different electron withdrawing or donating groups onto the phenyl rings of diphenyl phosphonates (DPPs) results in differential specificity between the urokinase-type plasminogen activator (uPA) and trypsin (Sienczyk and Oleksyszyn, 2006). A second way to reduce DFP promiscuity is to include a short peptide, which helps the inhibitor bind to the active site of the targeted protease. By utilizing substrate cleavage data and combinatorial peptide libraries, the peptide sequence can direct the inhibitor towards a target protease or group of proteases (Lim and Craik, 2009). This methodology was used to design specific ABPs for Granzymes A and B (Mahrus and Craik, 2005).</p><p>Despite the aforementioned tunable properties, the emergence of peptide phosphonates as the premier ABPs for specific serine proteases has yet to occur. One can imagine a platform where peptide phosphonates are incorporated into existing microarray technology to create high-throughput assays to monitor active proteases at the bench or in the clinic. However, such a technology has yet to be developed, and the utility of peptide phosphonates has come into question.</p><p>Here we describe the synthesis, evaluation, and application of peptide phosphonate inhibitors designed to target S1A family proteases. This subfamily of Clan PA proteases contains the trypsin-like enzymes, and it is the largest family of proteases in higher eukaryotes (Rawlings et al., 2010). These proteases play many important roles in biology and disease progression, and thus are of interest to many fields. This work examines the effects of peptide sequence and length on inhibition of two similar S1A proteases, thrombin and MT-SP1/matriptase, via enzymatic, structural, and imaging methods. We find that peptide length and leaving group sterics are large determinants of potency, while sequence composition contributes to a lesser degree. Our findings suggest general guidelines for the design of phosphonate ABPs that are optimized for S1A proteases. By applying these guidelines, we demonstrate that peptide phosphonates can quantitatively label and follow proteases on the surface of cancer cells, a novel use of ABPs that can be applied to many systems.</p><!><p>Several modifications to published protocols have improved the synthesis of the diphenyl phosphonate ester of 4-amidinophenyl-glycine (H-(AmPhg)P(OPh)2, for brevity referred to herein as H-Bz-DPP, 12) and the final biotinylated peptide phosphonate probes (Oleksyszyn et al., 1994). The synthesis of 12 (Figure 1) has been reported previously, though isolation of pure compound has been difficult. Briefly, compounds 9-11 were synthesized as described in (Mahrus and Craik, 2005; Oleksyszyn et al., 1994). Published methods describe isolation of 11 by diethyl ether precipitation, however, under the basic reaction conditions to produce 11, the phosphonate center is not stable. Hydrolysis of one of the phenyl ester groups can occur and be replaced with either an ethyl or methyl ester adduct. Reducing reaction c from 5 to 2 days prevented accumulation of the side product as monitored by LC-MS. Longer incubations increased levels of the undesired product. Direct hydrogenation of 11 followed by HPLC produced pure 12. Alternatively, 11 and 12 were separable by silica gel flash chromatography (10:1 CH2Cl2:MeOH and 5:1 CH2Cl2:MeOH with ninhydrin staining, respectively) at Rf ≈ 0.3.</p><p>Improvements to the synthesis of the final biotinylated peptide diphenyl phosphonate probes include (1) separate construction of the biotinylated peptide moiety and (2) optimization of the coupling reaction of the peptide to 12 to increase yield and reduce reaction time. Instead of building the probe by extending 12 one amino acid at a time as described in (Boduszek et al., 1994) and (Jackson et al., 1998), biotinylated peptides were synthesized on 2-chlorotrityl chloride resin by standard Fmoc chemistry, cleaved under mild acidic conditions to retain the protecting groups, and coupled to the phosphonate. Previous reports indicate this phosphonate coupling reaction to be time consuming (12-48 hours) and low yielding (5-30%)(Oleksyszyn et al., 1994). Optimization of this reaction proceeded with a TFE:CH2Cl2 (2:5) solvent mixture to reduce the reaction time from overnight to 6 hr, where DMF had been used previously. Treatment with the coupling agent EDAC resulted in the greatest yield of product (Table 1) as compared to PyBOP or DCC (~20%). EDAC protects the base-sensitive phosphonate from hydrolysis since it is typically utilized in the pH range of 4-6. Following the coupling reaction, deprotection of the amino acid side chains followed. HPLC purification was applied to both the protected and deprotected biotinylated peptide phosphonates. As a result, a combination of TFE:CH2Cl2 and EDAC resulted in a fast, efficient coupling that protected the integrity of the phosphonate center and improved product yield (Table 1).</p><!><p>Having optimized synthesis of the phosphonate inhibitors, the inhibitors were next characterized in vitro. Previous work using positional scanning, synthetic combinatorial library (PS-SCL) profiling had identified RKSR as the preferred P4-P1 tetrapeptide sequence for MT-SP1 and LTPR as the preferred sequence for thrombin (Bhatt et al., 2007; Harris et al., 2000). Combining this data with other validated substrates and structural information led to the creation of a series of peptide phosphonates that were designed to target MT-SP1 and thrombin (Table 1). Two sequences, QRVBz (1) and LTPBz (2) were rationally designed to maximize specificity for the two S1A proteases. A peptide element designed to be moderately effective against both proteases (GSGBz 3) was also synthesized, as was a fourth sequence (EPIBz 4) containing sub-optimal amino acids for both proteases at each P2-P4 position. Thus, a series of peptides were incorporated to create ABPs that were predicted to cover a range of activities against both MT-SP1 and thrombin.</p><p>IC50 values were calculated for each inhibitor against both MT-SP1 and thrombin, and kinact/KI values were calculated for inhibitors of a representative subset against MT-SP1 (Table 2). For MT-SP1, both the IC50 and kinact/KI values for the tetrapeptide inhibitors trended as expected, with the optimal sequence QRVBz-DPP (1) inhibiting best at 0.37uM, and the non-optimal sequence EPIBz-DPP (4) inhibiting worst at 76uM (Table 2a). These values were in agreement with kcat values measured for synthetic fluorogenic substrates with sequences corresponding to each inhibitor (Supplementary Table 1). This correlation between kcat and inhibitory potency has been previously observed (Drag et al., 2010).</p><p>Each of the peptide phosphonates tested proved to be slow inhibitors of MT-SP1. The fastest inhibitor, QRVBz-DPP (1), was only 200M−1s−1. Previous studies have reported values approaching 37,000M−1s−1 for phosphonate inhibition of serine proteases (Powers et al., 2002). The data shown here indicates that diphenyl phosphonate ABPs did not inhibit MT-SP1 rapidly, and either long incubation times or high phosphonate concentrations were needed to reach complete inhibition. While optimizing the amino acid composition of the peptide improved inhibition, we were unable to gain a large degree of selectivity. The best inhibitor for MT-SP1, 1, was better than optimal inhibitors designed against thrombin or uPA, another S1A protease (data not shown), further evidence suggesting that sequence composition plays a minor role in inhibitory potency for certain proteases.</p><p>The minor contribution due to sequence is reiterated by the ability of the peptide phosphonates to label proteases in vitro. The inhibitors were incubated with MT-SP1 and thrombin at increasing concentrations for 16 hours, and biotin-labeled proteases were detected via a Western blot using avidin-HRP. Figure S1 shows that the ability of the ABPs to label both proteases was also not strongly dependent on sequence composition. GSGBz-DPP (3), designed to confer intermediate potency for MT-SP1 and thrombin, labeled more efficiently than the ABPs with optimal sequences for each protease. This observation is partly explained by the greater stability of 3 over time in aqueous solution (Table S1), indicating that sequence stability was an additional parameter for labeling efficiency. Additionally, QRVBz, the best kinetic inhibitor of MT-SP1, contains a second cleavage site after the P3 Arg. This cleavage removed the biotin from the ABP, rendering it undetectable by avidin-HRP and complicating analysis. This secondary cleavage event is evidenced by the appearance of an 846Da fragment upon post-incubation LC-MS analysis.</p><p>In addition to the ABPs described above that were designed to test the effects of amino acid sequence on inhibition, three additional ABPs were synthesized to test the effects of peptide length on potency. Previous reports have noted a dependence of peptide length on the mechanistic rate constant k2, which suggest a possible effect of length on peptidyl DPP potency (Case and Stein, 2003). A mono-, di-, and tripeptide phosphonate was synthesized for the QRVBz-DPP probe (1), as this sequence proved to inhibit both proteases well (Table 2). A clear correlation was observed between increasing peptide length and improvement in inhibition, with a 10 fold increase between the IC50 of the longest inhibitor (QRVBz-DPP, 1) and the shortest inhibitor (Bz-DPP, 5), from 0.37uM to 3.5uM. A similar trend was seen in the kinact/KI data across the same parameters.</p><p>It was clear that not all positions affected MT-SP1 inhibition equally. The addition of a P3 residue increased potency by an order of magnitude (from 8uM to 0.56uM), while only modest changes were observed by the addition of P2 or P4. This was consistent with previously observed substrate:enzyme binding interactions with serine proteases, in which much of the binding energy is contributed by the P1 amino acid interacting with the S1 pocket of the protease. This also correlated with known MT-SP1 substrate data and with crystallographic observations about the S4 and S3 pockets (Bhatt et al., 2007; Friedrich et al., 2002). Friedrich et al. observed that the deep S3 pocket of MT-SP1 could accommodate a Lys or Arg residue from either P3 or P4 of a substrate while the other amino acid is solvent exposed. This phenomenon could explain the modest increase in inhibition seen when increasing from a tripeptide- to a tetrapeptide phosphonate (Friedrich et al., 2002).</p><p>A similar trend was seen in IC-50 values measured against thrombin (Table 2c) and uPA (data not shown). Again, it is notable that QRVBz-DPP (1), the ideal MT-SP1 inhibitor, was more potent than the ideal thrombin inhibitor LTPBz-DPP (2) by an order of magnitude (0.125uM for 1 compared to 1.1uM for 2). Sequence data alone does not appear to be sufficient to design ABPs that can distinguish between these two proteases, though other proteases with much more stringent substrate binding pockets have shown to be amenable to sequence-derived specificity (Mahrus and Craik, 2005).</p><p>To determine if potency continued to improve beyond P4, a hexapeptide ABP was synthesized and tested against MT-SP1. The preferred P5 and P6 amino acids for MT-SP1 were added to the most stable inhibitor (GSGBz 3) to create 13 (FTGSGBz). 13 is 25% more potent than 3 against MT-SP1, verifying that further increasing length improved potency.</p><p>Interestingly, each of the ABPs synthesized for this study inhibited thrombin at much lower concentrations than MT-SP1. We next sought to obtain a structural basis for the improvement in inhibition observed against thrombin.</p><!><p>The structure of MT-SP1 bound to Bz-DPP phosphonate was solved to 1.19Å resolution (Figure 2, Table 3). The protein was incubated with saturating ABP concentrations overnight at room temperature and purified via size exclusion chromatography. Attempts to crystallize MT-SP1 with Ac-QRVBz-DPP only produced crystals containing Bz-DPP. A possible explanation is that nonspecific hydrolysis of the peptide bond occurred during the several rounds of seeding and crystal growth that were necessary to obtain high-quality crystals, all of which took place in an aqueous environment at room temperature over several months. The truncated inhibitor may have also resulted from incomplete purification of the Bz-DPP (12) starting material from the full-length ABP.</p><p>This is the highest resolution MT-SP1 structure to date, which allows for visualization of all bonds. The phosphonate bound in the solvent exposed binding pocket in the expected conformation, with the phosphorous atom bound to the active site Ser195 and Bz in the S1 pocket (Figure 2a). The positively charged guanidine group of Bz formed two hydrogen bonds with the negatively charged Asp202. The N-terminal end of the inhibitor pointed down the substrate binding pocket, and the phenyl ring is solvent exposed.</p><p>A structure of thrombin (PDB: 1QUR) (Steinmetzer et al., 1999) was overlaid onto the MT-SP1-Bz-DPP structure, and the probe fit similarly into the binding pocket (Figure 2b). Inspection of the placement of the phenyl rings of the phosphonate revealed one important difference in the active site architecture. In MT-SP1 residue Phe99 lay in the region where the leaving group phenyl ring binds during catalysis. In thrombin, a smaller Leu99 occupied this location, forming a larger pocket than seen in MT-SP1. The Phe was only 2Å away from the phosphorous, thus we hypothesized that Phe99 caused steric interference with the leaving phenyl ring, slowing orientation of the inhibitor in the active site and slowing inhibition of MT-SP1 relative to that of thrombin.</p><p>To test this hypothesis, MT-SP1 Phe99 was mutated to Ala to create MT-SP1 F99A, and IC50 measurements with the same series of inhibitors were repeated (Table 2b). The IC-50 values for MT-SP1 F99A fell between that of MT-SP1 and thrombin in each case, indicating the size of the prime side pocket in MT-SP1 influenced phosphonate inhibition.</p><p>When the structure of uPA (PDB: 3KGP) (Zhang et al., 2010), an S1A protease with slow kinetics of phosphonate inhibition similar to those of MT-SP1 was overlaid with thrombin, a similar steric hindrance was observed (Fig 2d). In the prime-side binding pocket for uPA, His99 protruded exactly where Phe99 did in MT-SP1. Thrombin had a larger prime-side binding pocket than either MT-SP1 or uPA, due to the Leu found at residue 99. The smaller pocket in uPA, as with MT-SP1, may explain the slow inhibition of uPA by DPPs. This reinforces the idea that the steric fit of the leaving group into the prime side pocket has an important effect on potency.</p><!><p>Having determined the kinetic and labeling properties of the ABPs in vitro, their potential for labeling active proteases on the surface of live immortalized cancer cells was examined. The most stable inhibitor (13) and the most potent inhibitor (1) were used to label the cell surface proteases of two different cell lines grown in culture. These cell lines represented epithelial cancer types from two different species: PC3 cells, derived from human prostate cancer, and PDAC2.1 cells, derived from a transgenic mouse model of pancreatic ductal adenocarcinoma (Nolan-Stevaux et al., 2009). Live cells were incubated with phosphonate overnight in serum free media at 37°C. Streptavidin conjugated with AlexaFluor488 dye was used to fluorescently tag the ABPs. Cells were examined for the presence of labeled proteases either by fluorescent microscopy or by flow cytometry.</p><p>As seen in Figure 3, green fluorescent labeling was observed in PC3 and PDAC cells that had been exposed to ABPs, indicating the presence of active S1A proteases on the cell surface. PDAC2.1 (Figure 3a) cells labeled discrete foci on the cell surface. In the case of PC3 cells (Figure 3b), these labeled proteases were localized to one end of each cell. In addition to cell surface labeling, punctate labeling was also observed just internal to the plasma membrane. These puncta co-localized with membrane staining dye, indicative of internalization of labeled surface proteases. Internalization was also occasionally observed in PDAC2.1 cells and more frequently observed in MCF7 cells, a human breast cancer cell line (Figure S2) This result indicates that ABPs can visualize and follow active serine proteases on the surface of live cells, a novel use for peptide phosphonates.</p><p>After demonstrating the ability to label cell surface proteases, the ability to quantify the active form of these proteases was tested. Cells were grown, exposed to phosphonate overnight at 37°C, and fluorescently tagged with labeled streptavidin. Fluorescence was quantified using flow cytometry on the live cells. Fluorescence increased relative to background when cells had been incubated with an ABP, and this labeling was reduced in the presence of protease inhibitors (Figure 3e). These results demonstrate that ABPs can be used to quantify active proteases on the surface of live cells, and that labeling can be directly attributed to proteolysis.</p><p>Collectively, the cell-based experiments show that peptide phosphonates can be used to quantitatively label active cell surface proteases on different cell types. Live cell imaging can also be used in conjunction with ABP labeling to determine the localization of active proteases in the context of a cell or group of cells.</p><!><p>This study presents additional guidelines for the synthesis, design, and application of peptide phosphonate ABPs to those described previously (Oleksyszyn and Powers, 1994). Our results show that by focusing on peptide length, stability, and prime side sterics, the lifetime and utility of these ABPs can be improved dramatically. By following these guidelines, a novel protease imaging application using ABPs was developed. These experiments demonstrate that ABPs can be used to quantify, image, and follow proteases on cell surfaces.</p><p>We have improved the utility of phosphonate ABPs by combining synthetic, kinetic, and structural data. Obtaining reagent quantities of material has been a significant barrier to using these molecules in numerous applications. The improved synthesis methodology presented here both increases reproducibility of ABP production and increases yields up to seven fold compared to previous methods. The robust generation of ABPs greatly improves their utility, and synthesizing greater quantities of these probes enabled a systematic study of their potency.</p><p>Three major factors were found to contribute to peptide phosphonate potency: peptide stability, length, and prime side sterics. Peptide sequence is often viewed as a primary determinant of potency, and reports of sequence-based selectivity exist for Granzymes A and M, two S1A proteases with highly selective and differing substrate specificities at P2-P4 (Mahrus and Craik, 2005). However, for the majority of trypsin-fold proteases, much of the substrate binding energy of is contributed via binding of a basic P1 amino acid in the deep S1 pocket of the active site. Because of the strong P1 contribution, the P2-P4 sequence often plays a minor role. Peptide stability, however, was found to contribute significantly to ABP functionality. Inhibitors containing the amino acid Pro were especially unstable and problematic for both synthesis and purification, while long, charged side chains like Arg were found to react with the electrophilic phosphonate, subsequently inactivating the inhibitor. Additionally, the use of basic residues like Arg upstream of P1 can result in a secondary cleavage site for trypsin-like proteases, which removes the probe from the reactive diphenylphosphonate moiety. Poorly fitting residues can have a detrimental effect on inhibition, as seen by the slow kinetics of the EPIBz (4) probe against MT-SP1, thrombin, and uPA (Table 2 and data not shown). Therefore, the optimal peptide element should be composed of the most stable residues tolerated by the protease.</p><p>Peptide length, rather than sequence, had a more noticeable contribution to improving inhibitory capabilities. Increasing length increased potency, most notably at the P3 position. This agrees with earlier findings (Oleksyszyn and Powers, 1991). The data suggests that when designing peptide phosphonate inhibitors, including a longer peptide will improve inhibition more reliably than modifying the sequence. The sequence composition data and stability observations indicate that phosphonate inhibitors of trypsin-fold proteases should contain at least a tetrapeptide composed solely of stable amino acids.</p><p>The third factor contributing to potency is the steric fit of the phosphonate leaving group in the prime side pocket of the protease. Atomic-resolution structural data showed the prime-side pocket is smaller in MT-SP1 than in thrombin. This pocket is adjacent to the catalytic S195, exactly where the leaving group phenyl ring would reside. This smaller pocket in MT-SP1 is due to the presence of a bulky Phe at residue 99. Mutagenesis experiments confirmed that the prime side pocket is an important binding determinant for DPPs and MT-SP1. The data indicates that the leaving group of the phosphonate can have a large effect on inhibition. When designing a phosphonate ABP specific for a protease, multiple leaving groups should be tested on a stable peptide scaffold to find the best inhibitor.</p><p>The data shown here provide important information for the design of peptide phosphonate ABPs with both broad and narrow specificity. We show that the peptide element (in stability and length) and the leaving group (in reactivity and sterics) contribute to inhibitory potency of phosphonate ABPs. When designing a broad-specificity phosphonate ABP, one should start with a scaffold containing at least four stable amino acids and then consider varying the leaving group if improved potency is desired. When designing an inhibitor for a specific S1A protease, substrate cleavage data should be viewed as a starting point. Varying the leaving group may result in larger differences in potency, even for closely related proteases. Interestingly, kinetic data indicates that varying the reaction time may label different sets of proteases, with longer incubations resulting in larger sets of labeled proteases. Therefore, reaction time may also influence selectivity. However, these experiments suggest that only in rare instances will true specificity be engineered by varying peptide composition alone.</p><p>By following the guidelines presented here, several broad-spectrum ABPs for S1A proteases were produced in high yields. We used the most potent (1) and stable (13) ABPs to develop a novel method of imaging proteases on the surface of cancer cells. ABP-labeled proteases can be visualized and quantified with fluorescently conjugated streptavidin. This approach found that levels of protease activity vary by cell type, and that the location of labeled proteases can be tracked through the cell (Figure 3). Protease activity and localization information can be leveraged to obtain new information about protease function.</p><p>The ability to study the localization of active proteases is a novel use of phosphonate activity-based probes, and offers a promising way to examine the biology of cell surface protease activity. Previous studies with cysteine protease ABPs have been successful at labeling and visualizing lysosomal cathepsins and extracellular cathepsins in tumors (Blum et al., 2005; Joyce et al., 2004). The current study extends the use of ABPs to serine proteases on the surface of live cells. Differences in labeling and localization varied between the cell types, a conclusion previously observed with MT-SP1 using specific antibody-based probes (Darragh et al., 2010). These broad-spectrum ABPs may be used to highlight cell-specific differences in global S1A protease function.</p><p>For example, cell surface proteases have been implicated in metastasis, and it is interesting to note that active protease localization was polarized. Notably, high concentrations of enzymes were observed at one edge of the cells distal to cell-cell junctions. In addition, correlating the degree of labeling with a cellular phenotype such as metastatic potential may yield new information about cancer cell biology.</p><p>Though the pattern varied, internalization of proteases was observed in all cell types, as evidenced by the co-localization of membrane and ABP in puncta just inside the cell membrane. We hypothesize two explanations for this observation. First, many proteins undergo natural trafficking to and from the cell surface, modulating cellular interactions with the outside environment. Trafficking has been implicated previously in the regulation of proteases in the cell (Ghosh et al., 2003). ABPs could thus be used to track the movement and function of active proteases in many complex biological systems, including cancer biology, where protease activity is frequently dys-regulated. Alternatively, the ABPs themselves may cause proteases to become internalized. Ligand-dependent endocytosis has been observed with other cell surface receptors (Behrendt, 2004; Ghosh et al., 2003). The molecular mechanism of cell surface protease internalization observed via ABPs invites further investigation.</p><p>In summary, through synthetic, kinetic, and structural insights, we have developed additional guidelines that define and expand the use of peptide phosphonate ABPs. By following these guidelines, we have created a potent pan-S1A protease probe and developed a general methodology for the study of active proteases at the surface of cells. This technology could allow for future insights into the role of proteases in cancer, and has potential applications to other fields in biology.</p><!><p>MT-SP1 was expressed and purified as described previously and stored at −20°C in 50mM Tris pH 8.0, 50mM NaCl, 10% glycerol (Takeuchi et al., 1999). Mutants for crystallography and kinetic studies were expressed, purified, and stored in the same manner. Thrombin was purchased from Sigma and stored at −20°C in 50mM Tris pH 8.0, 50mM NaCl, 0.1 mg/mL BSA. The MT-SP1 substrates spectrazyme-tPA and spectrafluor-tPA were purchased from American Diagnostica and stored at −20 at 10 mM in H20.</p><p>The materials for peptide synthesis including PyBOP, EDAC, HOBt, Fmoc-PEG20atom-OH, and 2-chlorotrityl chloride resin were purchased from NovaBiochem. All other chemicals were purchased from Sigma unless otherwise noted. Solvents including anhydrous ethanol (EtOH), choloroform (CHCl3), 1,4-dioxane, diethyl ether (Et2O), dichloromethane (DCM), trifluroethanol (TFE), trifluoroacetic acid (TFA) and N,N-dimethylformamide (DMF) were used as received. The peptides were synthesized following standard Fmoc-SPPS procedure on 2-chlorotrityl chloride resin.</p><p>Reactions were analyzed by LC-MS performed on a Waters Alliance liquid chromatography system with a Waters Micromass ZQ single-quadrupole mass spectrometer. HPLC purifications were carried out using an Agilent 1200 series system with C18 reversed-phase columns (Waters). Mobile phase consisted of 99.9:0.1% water/trifluoroacetic acid (solvent A) and 95:4.9:0.1% acetonitrile/water/trifluoroacetic acid (solvent B). All final compounds were characterized by Matrix-assisted laser desorption ionization time-of-flight (MALDI-TOF) mass spectrometry using an ABI 4700 MALDI-TOF-TOF mass spectrometer.</p><p>Compound numbers in bold refer to the structures shown in Figure 1. Diphenyl [N-(benzyloxycarbonyl)amino](4-cyanophenyl)methanephosphonate (9) and NHS-biotin were synthesized according to literature procedure.(ref) No attempts were made to resolve the D,L-(4-AmPhGly)P(OPh)2 diastereomers.</p><!><p>A batch of 9 (3.0g, 6.0mmol) was dissolved in 100 mL CHCl3 and placed in a 0° C bath. Next Ar was passed over the solution and under the flow of Ar, 3 mL of anhydrous EtOH (60mmol) and 100 mL of 4 M HCl in dioxane were added. The reaction was stirred under argon at 4°C for 5 days, after which 10 formed as a fine white precipitate formed. The precipitate was filtered, washed with Et2O, and dried under reduced pressure (2.32g, 72%). Mass calcd for C30H29N2O6P: 544.18; found 545.45 (M+H)+.</p><!><p>A batch of 10 (1.0g, 1.8mmol) was dissolved in 25 mL 1,4-dioxane and 25 mL anhydrous EtOH. Next the reaction was purged with Ar and 10mL of 0.5M NH3 in dioxane (5 mmol) was added dropwise. The reaction was stirred under argon at 23° C for two days, and the solvent was removed completely under vacuum to afford the gummy white crude intermediate, diphenyl [N-(benzyloxycarbonyl)amino](4-amidinophenyl) methanephosphonate 11. This intermediate was dissolved in anhydrous EtOH (80mL) and concentrated HCl (305μL, 3.7mmol) and hydrogenated over 10% palladium on activated carbon for 5 h at room temperature. The catalyst was separated by filtration, and the solvent was removed under reduced pressure. The white powder thus isolated was dissolved in HPLC solvent (water with 0.1% TFA) with the aid of DMF and purified by reverse-phase HPLC. Lyophilization of fractions containing product afforded 0.49 g (27%) of 12 as a white powder. Mass calcd for C20H20N3O3P: 381.12, m/z found: 382.01 (M+H)+.</p><!><p>To a DCM/TFE(5:1) solution containing the biotin-PEG-peptide (0.03 mmol), EDAC (0.03 mmol) and HOBt (0.03 mmol), was added H-Bz-DPP 12 (0.03 mmol). After 2 h, an additional eq of EDAC and HOBt was added and the mixture stirred for 8 hr. Next the solvent was removed, the isolated oil was dissolved in MeCN/water (1:3), and purified by reverse-phase HPLC. The fractions with product were lyophilized down to yield the desired product as a white powder. The final probes were dissolved in DMSO and stored at −20° C. The concentration of the solution was determined by HABA biotin quantification kit (Pierce).</p><!><p>Reaction of Biotin-PEG-Gln(Trt)-Arg(Pbf)-Val-OH (0.10 g, 0.11 mmol), with 12 yielded 39 mg (72% yield) of Biotin-PEG-Gln(Trt)-Arg(Pbf)-Val-DPP as a white powder. Mass calcd for C92H119N14O18PS2: 1802.80, m/z found: 1802.10 (M+H)+. Next the powder was resuspended in 95% TFA, 2.5% TIS, 2.5% water and agitated for 2 hours, to obtain the desired probe 1. The product was precipitated into cold ether, pelleted by centrifugation, and purified by reversed-phase HPLC. Lyophilization of fractions containing product afforded 22 mg (79%) afforded 1 as a white powder. Mass calcd for C60H89N14O15PS: 1308.61, m/z found: 656.34 (M+2H)2+.</p><!><p>The protected peptide Biotin-PEG-Leu-Thr(tBu)-Pro-OH was reacted with 12 and resulted in Biotin-PEG-Leu-Thr(tBu)-Pro-Bz-DPP as a white powder (22 mg, 56% yield). Mass calcd for C63H93N10O15PS: 1292.63, m/z found: 1293.29 (M+H)+. Next the powder was resuspended in 95% TFA, 2.5% TIS, 2.5% water and agitated for 2 hours, to obtain the desired probe 2. The product was precipitated into cold ether, pelleted by centrifugation, and purified by reversed-phase HPLC. Lyophilization of fractions containing product afforded 8 mg (36% yield). Mass calcd for C59H85N10O15PS: 1236.57, m/z found: 1237.12 (M+H)+.</p><!><p>Reaction of Biotin-PEG-Gly-Ser-Gly-OH with 12 resulted in 3 as a white powder (22 mg, 65%). Mass calcd for C51H71N10O15PS: 1126.46, m/z found: 1126.92(M+H)+.</p><!><p>Biotin-PEG-Glu(OtBu)-Pro-Ile-OH was reacted with 12 to afford 23 mg (57% yield) of Biotin-PEG-Glu(OtBu)-Pro-Ile-Bz-DPP. Mass calcd for C64H93N10O16PS: 1320.62, m/z found: 1321.43 (M+H)+. Then the powder was resuspended in 95% TFA, 2.5% TIS, 2.5% water and agitated for 2 hours to obtain the desired probe. The product was precipitated into cold ether, pelleted by centrifugation, and purified by reversed-phase HPLC. Lyophilization of fractions containing product afforded 20 mg (90% yield) of 4. Mass calcd for C60H85N10O16PS: 1264.56, m/z found: 1266.31 (M+H)+.</p><!><p>Reaction of Biotin-PEG-OH with 12 yielded 5, which was isolated as white powder. (7mg, 25%). Mass calcd for C44H60N7O11PS: 925.38, m/z found: 926.17 (M+H)+.</p><!><p>A batch of Biotin-PEG-Val-OH was reacted with 12 to afford 6 as a powder after HPLC purification (7mg, 20%). Mass calcd for C49H69N8O13PS: 1024.45, m/z found: 1025.82 (M+H)+.</p><!><p>Reaction of Biotin-PEG-Arg(Pbf)-Val-OH with 12 afforded Biotin-PEG-Arg(Pbf)-Val-OH as a white powder (34 mg, 77%). Mass calcd for C68H99N12O17PS2,: 1450.64, m/z found: 1451.24 (M+H)+. Then the powder was dissolved in 95% TFA, 2.5% TIS, 2.5% water and agitated for 2 hours to obtain the desired probe. The product was precipitated into cold ether, pelleted by centrifugation, and purified by reversed-phase HPLC. Lyophilization of fractions containing product afforded 24 mg (90% yield) of 7. Mass calcd for C55H81N12O14PS: 1180.55, m/z found: 1181.64 (M+H)+.</p><!><p>Reaction of Biotin-PEG-Phe-Thr-Gly-Ser-Gly-OH with 12 afforded 13 as a white powder (45 mg, 40%). Mass calcd for C64H87N12O18PS: 1374.57, m/z found: 1376.27 (M+H)+.</p><!><p>All kinetic fluorescence measurements were taken in duplicate using a SpectraMax Gemini fluorescence spectrometer (Molecular Devices) with an excitation wavelength of 380nm, an emission wavelength of 460 nm, and a 435 nm cutoff filter. A solution of inhibitor was serially diluted over an appropriate concentration range and incubated with enzyme. Substrate was added at the end of 4 hours to initiate the reaction, and IC-50s were calculated. MT-SP1 was used at 0.2nM in a buffer containing 50mM Tris pH 8.0, 50mM NaCl, 0.01% Tween-20, with 200uM spectrafluor-tPA as the substrate. Thrombin was used at 0.5nM in a buffer containing 50mM Tris pH 8.0, 50mM NaCl, 0.01% Tween-20, 0.1 mg/mL BSA, with 200uM Boc-b-benzyl-Asp-Pro-Arg-AMC as the substrate. All reactions were run in duplicate.</p><p>Steady-state kinetics were used to determine the observed rate constants for the inhibition reaction. The inhibitors were serially diluted in a 96 well plate at an appropriate range of concentrations. Enzyme was added in hour intervals over 8-10 hours. The reaction was initiated by the addition of 200uM substrate, and the Vmax recorded on a SpectraMax fluorescence spectrometer. Kobs was determined at each inhibitor concentration by plotting Vmax vs time, and KIapp was determined by plotting Kobs vs [1]. Kinact/KI was determined using the equation k=K2[Io]Io+Ki∗(1+[So]KM)whereKi∗=kinactKI.</p><!><p>Crystals were grown at room temperature by vapour diffusion in hanging drops. A combination of micro and macro seeding was used to grow large single crystals in 4.0 M Na Formate at pH=7.0 and 25 mM FeCl3 as additive. Crystals belong to monoclinic space group C2 with one protease-inhibitor complex in the asymmetric unit corresponding to a solvent content of 50% and diffracted to better than 1.2Å resolution.</p><!><p>Data were collected at beamline 8.3.1 at the Advanced Light Source in Berkeley on a single crystal cryoprotected in mother liquor supplemented with 20% glycerol. The data were indexed, scaled and reduced using Mosflm and Scala in Elves (Holton and Alber, 2004). The structure was solved by molecular replacement using Phaser (McCoy et al., 2007) with the previously solved protease structure (PDB 3BN9) as search probe (Farady et al., 2008). Automatic building and refinement were performed in Phenix (Adams et al., 2002) using Phenix elBow to generate the covalently bound ligand. Manual building was carried out in Coot (Emsley and Cowtan, 2004). The stererochemistry of the final model was validated using MolProbity (Davis et al., 2007).</p><!><p>For recombinant protease labeling, enzyme was combined with varying concentrations of phosphonate inhibitor (3mM-30uM) at room temperature overnight. The reaction was stopped by the addition of SDS loading buffer and boiling for 10 minutes. Western blots were developed using the Vectastain ABC elite kit (Vector Labs).</p><!><p>Substrates corresponding to each inhibitor were synthesized by solid phase peptide synthesis. ACC-Rink-amide resin was obtained from Kimia Corp. The first amino acid was coupled using 5 equivalents each of amino acid, HATU, and collidine in dry DMF under argon for 16 hours with agitation. The full-length peptide was synthesized using a Symphony Quartet peptide synthesizer (Protein Technologies, Inc.), acetylated with 8 equivalents each of acetic anhydride and DIPEA, and cleaved with 95% TFA/2.5% water/ 2.5% triisopropyl silane. Cleaved peptides were precipitated into cold ether, collected by centrifugation, and purified by HPLC.</p><!><p>PC3 cells were obtained from the American Type Culture Collection (ATCC) and propagated in F12K Nutrient mixture with Kaighn's Modification (1X) and L-Glutamine (GIBCO). Media was supplemented with 10% FBS and 1X Penicillin:Streptomycin. MCF7 cells were obtained from the ATCC and propagated in Dulbecco's Modified Eagle Medium with high-glucose (D-ME H21) without phenol red (GIBCO). Media was supplemented with 10% FBS, 10ug/mL Insulin, and 1X Penicillin:Streptomycin. PDAC2.1 cells were isolated from p48-Cre/+, LSL-KrasG12D/+, Trp53F/+ transgenic mice according to (Nolan-Stevaux et al., 2009). PDAC2.1 cells were propagated in D-ME H21 (GIBCO) supplemented with 10% FBS and 1X Penicillin:Streptomycin.</p><!><p>Cells were grown to confluency in 6-well cell culture-treated dishes using complete media appropriate for each specific cell line. Media was then aspirated and replaced with Opti-MEM serum-free medium (GIBCO). All experiments were done in triplicate. FTGSGBz-DPP (13) was added to the appropriate wells at a final concentration of 50mM. To test proteolysis, ABP was added in the presence of 1X Complete Protease Inhibitor Cocktail (Roche) dissolved in Opti-MEM. The cells were then incubated at 37°C for approximately 20hrs. After incubation, ABP-containing media was aspirated and cells were washed 3x with Opti-MEM. Cells were detached from the surface of the wells with Enzyme-free Cell Dissociation Buffer (GIBCO) for 15min. Detached cells were washed 2x with Opti-MEM, resuspended in 200uL, and incubated for 20min with Streptavidin:AlexaFluor 488 conjugate (Invitrogen) at 4°C. Cells were washed 3x with Opti-MEM and resuspended in 500uL Opti-MEM and assayed for fluorescence using a BD FACSCalibur (BD Biosciences). Data was analyzed and population mean fluorescence values were obtained using FlowJo Flow Cytometry Analysis Software (TreeStar, Inc.).</p><!><p>Cells were grown to confluency in 9.5cm2 glass bottom microwell dishes (MatTek) in complete media. Media was then aspirated and replaced with Opti-MEM. Phosphonate was added to each dish to a final concentration of 50mM. FTGSGBz-DPP (13) was incubated for 20hrs at 37°C and QRVBz-DPP (1) was incubated for 3hrs at 37°C, respectively. Cells were washed 3x with Opti-MEM and then incubated for 20min with Streptavidin:AlexaFluor 488 conjugate and tetramethylrhodamine-conjugated wheat germ agglutinin (Invitrogen), simultaneously, at 4°C. Cells were then washed 3x with Opti-MEM and imaged. Fluorescence microscopy was carried out in the wide field using a Nikon Diaphot with a Nikon 60x lens, numerical aperture 1.4, objective and standard interference filter sets (Omega Optical). Images were collected using a 12-bit cooled charge-coupled device camera (Princeton Instruments) interfaced to a computer running Micro-Manager 1.3 software (http://micro-manager.org). Images were processed using Adobe Photoshop to assemble dual color image files. Brightness/Contrast was adjusted, where necessary, to improve image quality and clarity.</p><!><p>Synthesis of peptide phosphonate ABPs has been optimized to increase yield.</p><p>Peptide length is a key determinant of phosphonate inhibition.</p><p>3-D structure reveals an important steric contribution to phosphonate potency.</p><p>ABPs were shown to label active S1A proteases on the surface of live cancer cells.</p>

PubMed Author Manuscript

Diffractometric Detection of Proteins using Microbead-based Rolling Circle Amplification

We present a robust, sensitive, fluorescent or radio label-free self-assembled optical diffraction biosensor that utilizes rolling circle amplification (RCA) and magnetic microbeads as a signal enhancement method. An aptamer-based sandwich assay was performed on microcontact-printed streptavidin arranged in 15-\xce\xbcm-wide alternating lines, and could specifically capture and detect platelet-derived growth factor B-chain (PDGF-BB). An aptamer served as a template for the ligation of a padlock probe and the circularized probe could in turn be used as a template for RCA. The concatameric RCA product hybridized to biotinylated oligonuclotides which then captured streptavidin-labeled magnetic beads. In consequence, the signal from the captured PDGF-BB was amplified via the concatameric RCA product, and the diffraction gratings on the printed areas produced varying intensities of diffraction modes. The detected diffraction intensity and the density of the microbeads on the surface varied as a function of PDGF-BB concentration. Our results demonstrate a robust biosensing platform that is easy to construct and use, and devoid of fluorescence microscopy. The self-assembled bead patterns allow both a visual analysis of the molecular binding events under an ordinary bright-field microscope and serve as a diffraction grating biosensor.

diffractometric_detection_of_proteins_using_microbead-based_rolling_circle_amplification

<!>Materials<!>Construction of PDMS Stamps<!>Sandwich assay with RCA<!>Sandwich assay without RCA<!>Diffractometry Setup<!>Detection of PDGF-BB via RCA<!>Dependence of normalized diffraction mode intensities on PDGF-BB concentration<!>Dependence of bead packing density on PDGF-BB concentration<!>Specificity of PDGF-BB detection with RCA<!>Variations in normalized diffraction intensities<!>Detection of PDGF-BB without RCA<!>Conclusions<!>

<p>Detection of proteins in a sensitive and rapid manner plays an essential role in clinical applications. Numerous studies have been reported on using antibody-based immunoassay systems as recognition elements for detecting proteins.1-3 Antibodies, however, are generally produced in vivo, which generates difficulties in engineering their properties. In contrast, aptamers, generated by an in vitro selection process, are single-stranded oligonucleotides (DNA or RNA) that can specifically bind and recognize a variety of analytes ranging from small organic molecules to proteins, even to whole cells. As a result, aptamers have been used as recognition elements in a number of biosensing platforms.4-7 However, even though aptamers uniquely transduce the recognition of analytes into the generation of readily observable signals, analytes in small quantities are still difficult to detect with aptamers alone, pointing to a need for novel signal enhancement schemes.</p><p>Among many other biosensing platforms, the effectiveness of optical diffraction based biosensors has been demonstrated for recognizing binding events of various biomolecules, which operate based on changes in effective height or refractive index on periodically patterned gratings.8-12 In many studies, in order to detect small amount of biomolecules, additional signal enhancement was necessary.13 The enhancement was accomplished either by microfabrication of solid diffraction gratings or by in situ assembled diffraction gratings that are self-fabricated by nano or micro-size particles. Compared to the microfabrication of diffraction gratings which increases cost, time and in some cases requirement of additional amplification steps, the microbead-based in situ assembled diffraction grating enables rapid, cost effective, and sensitive detection of biomolecular targets. The microbeads, due to their large size compared to target molecules, form self-assembled diffraction gratings, which significantly lowers analyte detection limits.13-15 Nevertheless, the detection limit can be further improved by combination with novel and robust signal enhancement methods.</p><p>Rolling circle amplification (RCA) has been proven to enhance signals for detecting a variety of analytes due to its sensitivity.16-19 Although polymerase chain reaction (PCR) also provides high sensitivity for the detection of various target molecules, RCA has some advantages over PCR. RCA, an isothermal technique, does not require expensive and/or relatively large-scale equipment for thermal cycling or special laboratory conditions to avoid contamination, which makes it adaptable to low-cost and robust biosensing. RCA requires a circular template and a primer with a free 3′ end which is then extended via DNA or RNA polymerase with strand displacement abilities to result in a single stranded concatameric product consisting of tandem repeats of the complement of the circular template. The long concatameric product, thousands nucleotides in length, can be detected by a variety of methods such as hybridization of fluorescent-17, 18 or bead-labeled oligonucleotide probes.20-22 In addition, RCA has been shown as a useful method for chip-based detection because the concatamers can be localized to a given spot on a microarray slide.23-26 Herein we present a microbead-based RCA system as a signal enhancement method to achieve a sensitive self-assembled optical diffraction biosensor.</p><p>In this study, an aptamer-based sandwich assay in combination with RCA was used to detect platelet-derived growth factor B-chain (PDGF-BB) that is known to be related to tumor growth and transformation.27-30. PDGF-BB was captured and sandwiched between two anti-PDGF-BB aptamers (Figure 1). The dimeric nature of the PDGF-BB molecule allowed using two aptamers with identical binding sequences with the secondary aptamer only having a DNA extension to serve as the primer. This extension was formed with considerable ease due to the aptamer itself being already a DNA. The biotinylated capture aptamer was immobilized on the surface where periodic patterns of streptavidin were microcontact printed using a polydimethylsiloxane (PDMS) stamp. The aptamer-primer complex acted as a primer for RCA to localize the RCA product directly to the PDGF. For our purposes, we have utilized a linear padlock probe in which the 15 nucleotides on the 3′ end and 11 nucleotides on the 5′ end hybridized to the ′3 end sequence of the aptamer-primer complex and are ligated together to act as the circular template for RCA. The diffraction grating was then formed by introducing streptavidin-labeled beads that conjugated to biotinylated probes, which bound to RCA-amplified concatamers on the periodic patterns. Illuminating the pattern with a laser light produced diffraction modes with varying intensities that depended on the number of beads present on the patterns. Furthermore, in order to verify effectiveness of RCA on the periodical patterns as a signal enhancement method, we investigated the capability of an aptamer-based sandwich assay without RCA on periodically patterned gratings. Given the small number of beads bound to the surface and hence the lack of observation of diffraction modes, RCA played an essential role in the sensitive detection of PDGF-BB.</p><!><p>The following oligonucleotides were purchased from IDT (Coralville, IA). Capture aptamer labeled with biotin at 3′end was 5′-TACTCAGGGCACTGCAAGCAAT TGTGGTCCCAATGGGCTGAGTATTTTTT-biotin-3′ (The boldface portion is the aptamer sequence). Aptamer-primer complex was 5′-TACTCAGGGCACTGCAAGCAATTGTGGT CCCAATGGGCTGAGTATTTTTTTTGTCCGTGCTAGAAGGAAACAGTTAC-3′ (The boldface portion is the aptamer sequence and the italicized portion is the primer sequence.) and padlock probe was 5′-phosphate-TAGCACGGACATATATGATGGACCGCAGTATGAGTA TCTCCTATCACTACTAAGTGGAAGAAATGTAACTGTTTCCTTC-3′ (The italicized portion is complementary to the italicized sequence of the aptamer-primer complex. Biotinylated oligonucleotide probe labeled with biotin at 5′end was 5′-biotin-GTTTCCTTCTAGCAC-3′. Streptavidin and PBS buffer were obtained from Invitrogen (Carlsbad, CA). PDGF-BB, -AA and -AB were purchased from R&D Systems (Minneapolis, MN). The PDGF proteins were reconstituted in 4mM HCl with 0.2% BSA (Invitrogen, Carlsbad, CA). Phi 29 reagent set and E.coli DNA ligase were purchased from Epicentre (Madison, WI). 750 nm diameter streptavidin-labeled magnetic beads were obtained from Thermo Fisher Scientific Inc. (Waltham, MA). Bovine serum albumin (BSA) was obtained from Sigma-Aldrich (St. Louis, MO)</p><!><p>Polydimethylsiloxane (PDMS) stamp with 15-μm-wide alternating lines was prepared with a master mold having 15 μm alternating stripes. Silicone elastomer base and curing agent (SYLGARD 184 silicone elastomer kit, Dow Corning Corp., Midland, MI) were mixed in a 10:1 ratio and degassed for 1 h. After pouring into the master mold, they were degassed for an additional hour. Curing was carried out at 65 °C overnight. The cured PDMS stamp was carefully peeled off the mold after cooling down to room temperature. The stamp was thoroughly rinsed with nanopure water and cleaned in a sonicator prior to use.</p><!><p>On a gold-coated glass slide (50 nm thick gold with 5 nm thick chrominium, Asylum Research), periodic patterns of streptavidin (1 mg/ml in PBS, 7.4 pH) were microcontact printed (contact time: 5 min) using a polydimethylsiloxane (PDMS) stamp with 15-μm-wide alternating lines. After the streptavidin was microcontact-printed on the gold chip, the receptor-free stripes were blocked with 1 mg/ml of bovine serum albumin (BSA). After washing with 5 ml of PBS for 1 min (0.1 mM Na2HPO4, 1.8 mM KH2PO4, 137 mM NaCl, pH 7.4), 100 μL of 1 μM capture aptamer labeled with biotin at its 3′end in PBSM buffer (PBS, 2.7 mM KCl, 1 mM MgCl2, pH 7.4) was allowed to conjugate to the printed area for 15 min. The aptamer was heat denatured at 95 °C for 3 min and cooled down to room temperature prior to the immobilization. After a 5 ml of PBSM wash for 1 min, 100 μL of PDGF-BB (10 pM-100 nM in PBSM) was introduced and allowed to interact with the bound aptamer for 15 min at 37 °C. The surface was then rinsed with PBSM for 1 min, followed by the addition of 100 μL of 2 μM aptamer-primer complex in PBSM for 15 min at 37 °C. The aptamer-primer complex was also heat denatured at 95 °C for 3 min and cooled to room temperature prior to introduction to the chip. After a 5 ml of PBS wash for 1 min, 100 μL of 2 μM padlock probe in PBS was allowed to hybridize. The padlock probe was designed so that the first 11 of nucleotides on the 5′ end and the 15 of nucleotides on the 3′ end were complementary to the primer moiety found on the aptamer-primer complex. Then, the padlock probe was circularized via ligation by 100 μL of 10 units E.coli DNA ligase in ligation buffer (30 mM Tris-HCl (pH 8.0), 4 mM MgCl2, 10 mM (NH4)2SO4, 1.2 mM EDTA and 0.1 mM b-NAD) for 15 min at 37°C forming the circular template for RCA. The complex was incubated with 50 μL of 50 U phi29 DNA polymerase and 1.0 mM dNTPs in RCA reaction buffer (40 mM Tris-HCl (pH 7.5), 50 mM KCl, 10 mM MgCl2, 5 mM (NH4)2SO4 and 4 mM DTT). The RCA reaction was conducted for 10 min at 37 °C, followed by a 5 ml of PBS rinse step for 2 min. 100 μL of 2.5 μM probe (biotinylated at 5′end) was heat denatured at 95 °C for 3 min and cooled down to room temperature. It was then applied to the surface and allowed to hybridize with the immobilized concatameric RCA products. 10 μL of streptavidin-labeled beads (1 % solid concentration/mL, 0.75 μm diameter) were suspended in PBS, placed in a magnetic separator, washed with 200 μl of PBS for 2 min three times and resuspended in 100 μl of PBS prior to use. The streptavidin-labeled beads were then allowed to conjugate to the biotinylated probes already hybridized to the RCA product. Finally, the chip was washed with 5 ml of PBS for 1 min, followed by a quick (∼10 sec) wash with 5 ml of ammonium acetate buffer (300 mM CH3COONH4, pH 7.0).</p><!><p>After the streptavidin was microcontact-printed on the gold chip, the receptor-free stripes were blocked with 1 mg/ml of bovine serum albumin (BSA). After washing with 5 ml of PBS for 1 min, 100 μL of 1 μM capture aptamer labeled with biotin at its 3′end was conformationally equilibrated by heating to 95 °C for 3 min followed by cooling to room temperature. The aptamer was incubated on the chip for 15 min in PBSM. After a 5 ml of PBSM wash for 1 min, 100 μL of PDGF-BB (1 nM in PBSM) was introduced and allowed to interact with the bound aptamer for 15 min. The surface was then rinsed with PBSM for 1 min, followed by the addition of 100 μL of 2 μM conformationally equilibrated capture aptamer labeled with biotin at its 3′end in PBSM buffer for 30 min at 37 °C. The streptavidin-labeled beads then captured the biotinylated aptamers. Finally, the chip was washed with 5 ml of PBS for 1 min, followed by a quick wash with 5 ml of ammonium acetate buffer (300 mM CH3COONH4, pH 7.0).</p><!><p>A laser beam (He-Ne laser, Newport R-30991, 633 nm, 5 mW) passed through a beam splitter (Thorlabs, BS016) and a convex lens (focal length 60 mm), leading to an incident beam diameter of approximately 150 μm on the gold surface. The reflected beam passed through a beam splitter, focused onto a CCD camera (Thorlabs, DCU223C). The measured signal intensities were recorded by a computer. Since the 0th mode was too intense, the exposure time of the CCD was set to 4.462 ms, and a band pass filter (Thorlabs, FB620-10) was placed between the laser and the beam splitter to decrease the intensity of this mode. On the other hand, a CCD exposure time of 0.398 ms was used for measuring the 1st mode intensity.13, 15</p><!><p>Figure 1 (a) illustrates the sandwich assay procedure, including the capture aptamer, PDGF-BB, the aptamer complex, and the subsequent microbead-based signal enhancement strategy. Periodic patterns of streptavidin were microcontact printed using a PDMS stamp with 15-μm-wide alternating lines. Biotinylated capture aptamers were then adhered to the streptavidin coated lines. PDGF was washed over the surface and bound to the capture aptamer. A secondary aptamer bound to the immobilized protein in a classic sandwich assay format. Both the 3′ and 5′ ends of a padlock probe could hybridize to the 3′ end of the secondary aptamer. Ligation of the padlock probe resulted in a circular template for RCA, with the 3′ end of the secondary aptamer acting as a primer. The requirement for ligation adds an additional layer of specificity, suppresses background, and therefore increases sensitivity, similar to how the proximity ligation assay increases sensitivity.31, 32 The concatamers that accumulated as a result of RCA were thus effectively immobilized on 15-μm-wide alternating lines. Multiple probes could hybridize to the long concatameric product. By conjugating biotinylated probes to streptavidin-coated microbeads it proved possible to obtain visual confirmation of the binding reaction, as well as diffractometric detection. In addition, the fact that a single bead might be bound by multiple repeats of the concatamer meant that we did not have to pre-optimize the binding affinity of the oligonucleotide probe to a given bead. As shown in Figure 1(b), self-assembled microbeads on the RCA-amplified concatamers formed gratings that produced diffraction modes upon illumination with a laser. The self-assembled diffraction gratings were examined under an optical microscope as shown in Figure 2 and were subjected to diffraction measurements.</p><!><p>Illuminating the grating formed by the streptavidin-coated microbeads yielded diffraction modes due to interference of the laser beams that reflected from the beads and the bead-free gold surface. We investigated the dependence of normalized diffraction mode intensities (I1/I0, a ratio of first mode I1 to zeroth mode I0) as a function of applied PDGF-BB concentration. The normalization was performed to suppress the disturbances that are known to influence both modes similarly.13-15 To investigate a worst-case scenario, the measurement was carried out on 5 different areas of each chip to account for variations in bead packing between different areas. The averages and standard deviations were calculated offline. Figure 3 demonstrates the dependence of the normalized diffraction intensity on the concentration of PDGF-BB. As the concentration of PDGF-BB increased from 10 pM to 100 nM so did the number of beads that bound to the printed areas, resulting in a corresponding increase in the normalized modal intensity. Fitting a Langmuir isotherm relation to our dose/response curve yields a dissociation constant of 0.8 nM. This value is comparable to that reported elsewhere (0.1 nM33, 34), but may be slightly higher due to surface effects. The minimum detectable concentration was 10 pM, beyond which the diffraction intensity was too weak to detect.</p><p>This minimum concentration is already better than what we reported in an earlier work that uses fluorescence-based detection scheme (0.4 nM).17 A recent study that investigated protein detection using a combination of aptamers and quantum dot bioconjugates demonstrated a detection sensitivity of 0.4 nM.34 Another study that used a commercial label-free electrochemical sensor in combination with thiolated PDGF aptamers and mercaptohexnol on a gold surface reported PDGF concentration in the range of 1-40 nM.28 Our detection limit is lower by one to two orders of magnitude. On the other hand, the minimum detectable concentration of a commercial electrochemical immunosensor with aptamer-primed polymerase amplification was as low as 0.6 pM35, which is lower than our detection limit by more than an order of magnitude. Our system offers significant advantages in terms of cost, versatility and ease of construction: it can be built from scratch in most laboratory settings with virtually no fabrication (other than a PDMS stamp which can also be purchased) and applied to a wide range of targets. Further, it allows a direct 'visual check' of the resulting reaction by placing the chip under an ordinary bright-field microscope before performing a diffraction measurement.</p><!><p>We also investigated the density of the streptavidin-labeled microbeads as a function of PDGF-BB concentration. This was done by counting the differential number: the number of beads on printed areas minus that on the adjacent non-printed stripes (i.e., non-specifically bound beads). For each investigated area, five printed stripes (15 μm × 150 μm) adjacent to five non-printed stripes (15 μm × 150 μm) were considered. The counting was performed on five different areas (22500 μm2) on the chip. Averages and standard variations were again calculated, and the standard deviations were similar to those in the previous diffraction measurements. The image processing for the bead counting was performed in MATLAB. Figure 4 demonstrates the dependence of the counted bead number on the PDGF-BB concentration in the range of 10 pM – 100 nM. The relationship was similar to that observed for the modal intensities shown in Figure 3. The diffraction measurement, however, was much simpler and quicker since it required measuring only two quantities (I1 and I0), whereas counting beads required image analysis of an entire area.</p><!><p>To validate the specificity of PDGF-BB detection, several control experiments were performed using dimeric isoforms of PDGF (PDGF-AA and AB, 10 nM) as shown in Figure 5. In addition, a negative control was carried out in the absence of PDGF-BB. No significant numbers of beads were obtained in either the case of PDGF-AA or in the absence of PDGF-BB. PDGF-AB yielded slightly more beads than PDGF-AA, but the overall packing was very small in comparison with PDGF-BB. In an earlier study that did not involve sandwich assays, comparable results were obtained with PDGF-AB and −BB.29, 36 Here, we observed much more specificity for PDGF-BB, which may be due to the additional discrimination afforded by the use of two aptamers in a sandwich assay.</p><p>Another negative control experiment (Figure 5) was carried out with 10 nM PDGF-BB, except that no RCA was performed. This data can be directly compared to the bead capture results described in Figure 2(b), but again resulted in a negligible number of beads on the printed area. These results effectively prove that the RCA-amplified concatamers were the templates for microbead capture, and thus that RCA was required to form the diffraction grating used for the detection of PDGF-BB.</p><!><p>To better understand the reasons for the variations in the normalized diffraction intensities, measurements were performed on a single spot for each chip (without moving to a different area on the chip). A reading was made three times on each spot, and averages and standard deviations were calculated. The variations in modal intensities were reduced significantly (compare Figure 3 and Figure 6). Hence, it is likely that a significant portion of the uncertainties in modal intensities are due to variations from one area to another within the same chip, and thus are due to imperfections in stamping or irregularities in the activities of immobilized molecules (i.e., due to surface denaturation). We are currently exploring alternative functionalization schemes involving lithography to improve the quality of the functional lines.</p><!><p>In order to verify that RCA assisted with the sensitivity of optical diffraction methods, a control experiment was performed. Figure 7(a) shows a schematic of a microbead-based aptamer sandwich assay without RCA. After immobilization of the biotinylated aptamers on the periodic patterns of streptavidin, PDGF-BB was captured and biotinylated aptamers and streptavidin-coated magnetic beads were added. As shown in Figure 7 (b), the number of the beads on the periodic pattern with 1 nM PDGF-BB was significantly smaller than was seen with RCA (by a factor of 4 to 5). Furthermore, the diffraction gratings with 1 nM PDGF-BB didn't produce noticeable diffraction modes upon illumination with a laser light. Hence, it was confirmed that RCA can play an important role in signal amplification for optical diffraction sensing.</p><!><p>In conclusion, we demonstrated that rolling circle amplification (RCA) in combination with microbeads can be used as a signal enhancement method to develop a sensitive self-assembled optical diffraction based biosensor. We sandwiched the dimer PDGF-BB with two aptamers and detected the presence of the analyte via RCA-based extension of the secondary aptamer and subsequent bead-based diffraction measurements. We showed that the density of the beads and the normalized diffraction intensity measured on the diffraction grating pattern depended monotonously on the PDGF-BB concentration. Aptamer-based sandwich assay without RCA did not result in significant bead-binding confirming the advantage of RCA. The detection of the RCA product was relatively simple and devoid of fluorescence or radio labels. Having established the proof-of-concept in this study, our future efforts will be focused on detection of disease makers present in various body fluids such as blood or serum, in conjunction with aptamer-based immunomagnetic separation.</p><!><p>(a) Schematic of RCA-based microbead detection assay in combination with aptamers. A biotinylated anti-PDGF-B specific aptamer is immobilized on streptavidin coated periodic patterns. PDGF-BB is introduced and captured by the aptamer. An aptamer-primer complex, with an additional primer sequence binds to the protein. A padlock probe hybridized to the primer tail of the aptamer is ligated and RCA is initiated. Streptavidin conjugated beads bind to elongated concatamers via hybridized biotinylated probes. (b) Self-assembled streptavidin (SA)-coated beads on the RCA-based micropattern form diffraction gratings that yield diffraction modes upon illumination with a laser.</p><p>(a)-(e) Optical micrographs of the self-assembled diffraction gratings formed by streptavidin-labeled beads with varying PDGF-BB concentration (10 pM – 100 nM). (f) Grating with no PDGF-BB.</p><p>Variation of the normalized diffraction intensity (I1/I0) with the PDGF-BB concentration (10 pM to 100 nM). Measurement was performed on 5 different areas, producing an average number of the normalized diffraction intensity (I1/I0).</p><p>Variation of the number of beads with PDGF-BB concentration (10 pM – 100 nM). The counting was performed on five different areas on the chip, yielding an average number of beads.</p><p>Optical micrographs of control chips to demonstrate the specificity of PDGF-BB detection.</p><p>Variation of the normalized diffraction intensity (I1/I0) with the PDGF-BB concentration (10 pM to 100 nM). Measurement was performed on the same area three times where diffraction gratings were well-defined and robust.</p><p>(a) Schematic of microbead-based aptamer sandwich assay without RCA. Biotinylated aptamer was immobilized on streptavidin functionalized periodic patterns. PDGF-BB was introduced, and sandwiched by another biotinylated aptamer. Streptavidin conjugated magnetic beads bound to free biotins of the aptamers that were captured by PDGF-BB. (b) Optical micrograph of the chip with 1 nM PDGF-BB.</p>

PubMed Author Manuscript

Functional Ionic Liquid Modified Core-Shell Structured Fibrous Gel Polymer Electrolyte for Safe and Efficient Fast Charging Lithium-Ion Batteries

Fast charging is of enormous concerns in the development of power batteries, while the low conductivity and lithium ion transference number in current electrolytes degraded the charge balance, limited the rate performance, and even cause safety issues for dendrite growth. Combine inorganic fillers and ionic liquid plasticizer, here in this paper we prepared a core-shell structured nanofibrous membrane, by incorporating with carbonate based electrolyte, a gel polymer electrolyte (GPE) with high conductivity, outstanding Li+ transference number was obtained. Notably, the Li/electrolyte/LiNi0.6Co0.2Mn0.2O2 (NCM622) half-cell with this composite electrolyte delivers a reversible capacity of 65 mAh/g at 20C, which is 13 times higher than that of with Celgard 2325 membrane. It also shows enhanced long-term cycle stability at both 3C and 5C for the suppression of lithium dendrite. This organic-inorganic co-modified GPE guarantees the fast charging ability and safety of LIBs, thus provides a promising method in high performance electrolyte design.

functional_ionic_liquid_modified_core-shell_structured_fibrous_gel_polymer_electrolyte_for_safe_and_

Introduction<!>Materials<!>Preparation of the GPEs<!><!>Characterization of the Membranes and GPEs<!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Morphologies of the Membranes<!><!>Conclusions<!>Data Availability<!>Author Contributions<!>Conflict of Interest Statement<!><!>Supplementary Material<!>

<p>With the increasing demand for energy consumption and the continuous improvement for environmental protection, the development of sustainable and renewable energy is urgently needed currently. Lithium ion batteries (LIBs) have been widely used in the power battery, energy storage applications, portable electronic equipment and other fields, owing to its inherent advantages such as high energy density, high power density, long cycle life, no memory effect, and environmental friendliness (Gao et al., 2009; Goodenough and Kim, 2010; Etacheri et al., 2011; Scrosati et al., 2011; Lu et al., 2014a; Varzi et al., 2016; Tan et al., 2018). Nevertheless, high-capacity and high-power LIBs are facing more and more challenges such as high power output, rapid charge/discharge capability and safety under dynamic conditions (Zhu et al., 2013; Xie et al., 2014; Zhang J. et al., 2014; Li et al., 2018, 2019), and it is of particularly importance to investigate methods to guarantee the safety of LIBs with enhanced fast charge capability (Rui et al., 2016; Tan et al., 2018).</p><p>LIBs generally compose of anode, cathode, electrolyte, and packing material. The electrolyte, which including polymer separator and liquid electrolyte in commercial LIBs, mainly plays two key roles: blocking the direct contact between the electrodes to avoid internal electrical short circuit and acting as the Li+ transport passage (Tarascon and Armand, 2001; Xiao et al., 2009; Huang, 2010; Lee et al., 2014). Therefore, the electrolyte not only directly affects the LIB electrochemical performance, such as cycle stability and rate capability, but also, it has significant influence on the safety of LIBs. Most commercial LIBs adopt polyolefin [such as polyethylene (PE) and polypropylene (PP)] membranes (Huang, 2010; Kim et al., 2015; Shen et al., 2018) and LiPF6/carbonates electrolyte, based on which satisfying cycle performances have achieved, while the rate capability and safety, especially during quick charge, is yet to be enhanced.</p><p>The major issue in fast charging LIBs is the poor kinetics caused by the sluggish ion transport in the cell, which will probably lead to large polarization and low reversible capacity, furthermore, when the anode potential is lower than that of Li plating, dendrite will be generated and may short the cell thus cause safety concerns.</p><p>Electrolyte holds significant influence on both rate capability and safety, on one hand, ionic conductivity and Li+ transference number is the major factor to control the battery kinetics (Diederichsen et al., 2017), on the other hand, the separator, which blocks the direct contact between electrodes, should also work as a physical barrier to suppress the dendrite from shorting the battery (Jana et al., 2015; Shin et al., 2015; Zheng et al., 2018). Therefore, scientist made lots of efforts in electrolyte modification, including surface coating of the separator (Ghazi et al., 2017; Liu et al., 2017a,b; Zuo et al., 2017), doping inorganic nanoparticles in polymer to obtain composite electrolyte (Xiao et al., 2012; Chen et al., 2017; Kim et al., 2017; Wang et al., 2017; Shen et al., 2018), and electrostatic spinning to get nanofibrous membrane (Choi et al., 2003; Wu et al., 2015; Ma et al., 2017; Cheng et al., 2018). Thereinto, the nanofibrous membrane has large specific surface area, three-dimensional (3D) porous structure and high electrolyte uptake ability, on which the rate capability can be enhanced (Xiao et al., 2015; Liang et al., 2016; Park et al., 2017). What's more, functional groups can be easily introduced in the electrospining process, thereby endows the electrolyte with other properties, such as dendrite suppression (Lu et al., 2015a,b; Cheng et al., 2018; Deng et al., 2018), flame retardant (Lu et al., 2017; Jia et al., 2018; Sun et al., 2018), et al. Poly(vinylidene fluoride-co-hexafluoropropene) (PVDF-HFP) is considered as a potential polymer matrix for electrolyte not only as it has a low crystallinity which could promotes rapid ion conduction (Ali et al., 2018; Zhao et al., 2019), but also, it has high dielectric constant and low surface energy that may promote the compact deposition of metal Li (Lopez et al., 2018). It is reported the ionic conductivity of PVDF-HFP-based polymer electrolytes is up to 3.9 mS/cm (Xiao et al., 2012).</p><p>Ionic liquid modified gel polymer electrolytes (IL-GPE) have attracted much attention due to their good thermal stability and mechanical properties. Singh et al. studied imidazolyl ionic liquids and found that the EMIMFSI-based GPEs have excellent electrochemical stability, good compatibility and thermal stability(Singh et al., 2018). Guo et al. prepared the PVDF-HFP-LiTFSI/SiO2/EMITFSI GPE with high thermal stability and good electrode compatibility (Guo et al., 2018). However, IL-GPE also has some shortcomings, such as lower ionic conductivity and lithium ion mobility (Zhou et al., 2015). In this work, a piperidine ionic liquid, 1-methyl-1-propylpiperidinium chloride (PPCl), and Li2SiO3 (LSO) nanoparticles were introduced into the PVDF-HFP matrix via coaxial electrospinning technology. The as prepared membrane has high porosity and electrolyte uptake, remarkable ionic conductivity, and outstanding electrochemical performance especially in quick charging. Commercial NCM622 cathode adopting this IL-GPE delivers a high reversible capacity of 65 mAh/g in 20C rate charge/discharging, which is 13 times higher than that of the cell adopting Celgard 2325 membrane.</p><!><p>Poly (vinylidene fluoride-co-hexafluoropropene) (PVDF-HFP, Mw. ~455,000), 1-methylpiperidine (97%), and (3-chloropropyl) trimethoxysilane (98%) were provided by Sigma-Aldrich. Li2SiO3 (LSO, 99%, Strem Chemicals), N, N-dimethylformamide (DMF, 99.5%, Beijing Chemical Works), LiNi0.6Co0.2Mn0.2O2 (NCM622, Beijing Dangsheng Material Technology Co., Ltd.), Super P (Imerys Graphite & Carbon), Polyvinylidene difluoride (PVDF, Solvay 5130), N-Methyl-2-pyrrolidone (NMP, 99.0%, Sinopharm Chemical Reagent Co., Ltd.), lithium tablet (Li, Tianjin Zhongneng Co., Ltd.), and polypropylene (PP, Japan Ube) were commercially available and used without further purification.</p><!><p>The ionic liquid PPCl was synthesized according to the procedure reported before (Lu et al., 2012; Korf et al., 2014; Cheng et al., 2018), its structure and purity was also testified in our previous work (Xu et al., 2018). The core-shell structured 3D porous nanofiber membrane was prepared by the coaxial electrospinning technique on an ET-2535H machine (Ucalery Tech Inc., China), as shown in Scheme 1. To make composite nonwoven membrane, 80% (wt., similarly hereinafter) DMF was adopted in all the spinning solutions. The slurry for core spinning was prepared by mixing PPCl and PVDF-HFP in DMF solvent, wherein the weight ratio of PPCl: PVDF-HFP: DMF was fixed at 1:19:80. Correspondingly, the slurry for shell spinning was prepared by mixing LSO and PVDF-HFP in DMF solvent with a ratio of 2:18:80. The coaxial electrospinning equipment mainly contained a variable positive voltage of 15 kV and a negative voltage of −2kV, two syringe pumps, a spinneret consisting of two chambers and a collector. The pumped speed of the shell and core solutions supply were fixed at 0.25 and 0.15 mL/h, respectively, the distance between the needle tip and aluminum foil collector was 13 cm. The as prepared membrane has a thickness of 50 ± 5 μm, and was entitled by PHP@PHL. Accordingly, PVDF-HFP, PVDF-HFP-PPCl (PHP), and PVDF-HFP-LSO(PHL) nanofiber membrane was prepared by mixing PVDF-HFP in DMF (20:80), PPCl, and PVDF-HFP in DMF (1:19:80) or LSO and PVDF-HFP in DMF (2:18:80), respectively. The as-prepared nonwoven fiber membranes were cut into disc with a diameter of 16 mm, which were then dried in a vacuum oven at 60°C for 20 h to remove the residual solvent. In the end, the membranes were filled and swollen in a liquid electrolyte, 1.2 M LiPF6 in ethylene carbonate (EC) and ethyl methyl carbonate (EMC) (3:7, weight ratio), for 30 min in an argon filled glove box to obtain the relevant GPEs.</p><!><p>Illustration of the preparation of the PHP@PHL membrane (Zhou et al., 2013).</p><!><p>The thickness of various films was recorded by measuring membrane apparatus (CH1ST, Shanghai Milite Precise Instrument Co., Ltd., China), and morphology of the membrane was studied by field-emission scanning electron microscopy (FE-SEM, JSM-7001F, JEOL, Japan). The field emission transmission electron microscopy (TEM, JEOL, JEM-2100) was used to test the core-shell structure of PHP@PHL nanoporous fiber membrane. The surface chemical composition of PHP@PHL was analyzed by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher Scientifec, America). Differential scanning calorimetry (DSC, Mettler-Toledo, Switzerland) was carried out to analyze the thermal behavior of all kinds of membranes. Samples were put into aluminum pans and the test temperature was set from 50 to 250°C with a heating rate of 5°C/min, under N2 atmosphere. The porosity of various films was measured by soaking n-butanol for 2 h, then calculated using Equation (1): P = (mb/ρb)/(mb/ρb + ma/ρa) ×100%, where ma and mb are the weights of separators and n-butanol, ρa and ρb are the density of separators and n-butanol, respectively (Xiao et al., 2012; Zhou et al., 2013). In an argon filled glove box, the electrolyte uptakes were analyzed by the mass difference of separators before and after soaking in electrolyte for 30 min and then calculated using Equation (2): EU = (W–W0)/W0 ×100%, in which W0 and W are the weights of the films before and after immersing in the liquid electrolyte, respectively. Ten samples are tested to measure the electrolyte uptake and porosity of each membrane. Wettability of the separators were researched by contact angle measurements (DSA 100S, Germany KRUSS).</p><p>The various GPEs' effective ionic conductivities (σ) were calculated using Equation (3): σ = d/(Rb × S), where d is the thickness of the film, S is the area and Rb is the bulk impedance acquired by electrochemical impedance spectroscopy (EIS) in Stainless Steel (SS)/electrolyte/SS symmetric cells. The bulk resistances (Rb) were investigated by a CHI660E electrochemical workstation with a frequency range of 0.01–105Hz and an amplitude of 5 mV at room temperature. The lithium ion transference numbers (tLi+) of different electrolytes were tested using chronoamperometry (CA) and EIS (both by CHI660E), and then calculated using Equation (4): tLi+ = IS(Δv–I0R0)/I0(Δv–ISRS) (Li et al., 2018), where I0 and IS are the initial and steady state currents obtained by CA testing of lithium symmetrical cell, R0 and RS are the interfacial impedance before and after polarization, and ΔV (10 mV) is the applied voltage difference (Yang et al., 2014; Zhang F. et al., 2014). The electrochemical performances were researched in coin cells with lithium foil anode and LiNi0.6Co0.2Mn0.2O2 (NCM622) cathode. The interfacial stability between the electrode and electrolytes was investigated using EIS after standing for 1, 10, 20, and 30 days, respectively. The long term cycling stability, as well as rate performances were tested in a voltage range between 2.8 and 4.4 V with different C rates by LAND battery cycle system (Wuhan Blue Electric Co., LTD, China).</p><!><p>The morphologies of the Celgard 2325 (which is a PP/PE/PP trilayer separator), nanofibrous PVDF-HFP, PHL, and PHP@PHL membranes, as well as microscopic changes after the thermal stability analyze in 130–170°C temperature range, were tested by using SEM, as shown in Figures 1A,D. The Celgard 2325 membrane has tensile holes in the lamellar matrix, but the electrostatic spinning films show 3D porous structure with 80–160 nm diameter nanofiber which interlacing with each other to form interconnected networks. The energy dispersive X-ray spectroscopy (EDS) data shown in Figure S1 gives the elemental distribution in PHL and PHP@PHL complex films. As shown in Figure S1a, Si and O are uniformly distributed in the membrane, which proves that Li2SiO3 is evenly dispersed in the polymer matrix. EDS results of PHP@PHL in Figure S1b shows little signal of N or Cl as the content of PPCl is pretty low (<2%). TEM image of PHP@PHL nanofiber further proved its core-shell structure, and the inorganic nanoparticle Li2SiO3 is dispersed in the out layer of the PVDF-HFP matrix, with the thickness of several to tens of nanometers as shown in Figure 1B. The XPS spectra in Figure S2 detected no Cl element on the surface of PHP@PHL nanofiber, which further proves that the PPCl ionic liquid was fully encapsulated by the out PHL layer.</p><!><p>(A) SEM images of the various separators, Celgard 2325, nanofibrous PVDF-HFP, PHL, and PHP@PHL; (B) TEM micrograph of the PHP@PHL fibers; (C) Picture of Celgard 2325, PVDF-HFP, PHL, and PHP@PHL separators and their macroscopic changes after heat treatment at different temperatures, (D) SEM pictures of Celgard 2325, PVDF-HFP, PHL, and PHP@PHL separators and their morphologies after heat treatment under different temperatures, (E) DSC curves of Celgard 2325, PVDF-HFP, PHL, and PHP@PHL separators.</p><!><p>The thermal stability of various membranes was investigated by storing them in an air circulation oven at a series of temperatures between 130 and 170°C, each for 20 min. Figure 1C shows that Celgard 2325 begins to curl for the melting of PE and molecular tanglement of polyolefin at about 135°C, it turns transparent at 160°C for the melt of PP, meanwhile PVDF-HFP shrink at 150°C and melt at 170°C. However, the PHL and PHP@PHL keeps their porous structure even at 170°C. It can be concluded that the inorganic nano-particle, Li2SiO3, in the shell structure significantly improved the thermal stability of the separator, which may potentially contribute in enhancing the safety of the battery. In order to further clarify the morphology changes of the films after heat treatment, FE-SEM pictures of the four separators are given in Figure 1D for microscopic description. Celgard 2325 was fully shutdown at 160°C, and PVDF-HFP nanofibrous films melt under 170°C, but the PHL and PHP@PHL films still keeps certain amount of pores remain in 170°C (Figure 1D). From the microscopic point of view, the core-shell structure nanofibrous membrane expresses excellent thermal stability, hence makes it have distinguished heat resistance and advance the safety of lithium battery under high temperature condition. According to DSC test results of four separators, membranes based on PVDF-HFP shows better thermal stability than that of Celgard 2325, and the core-shell structured PHP@PHL, released the least heat even on melting (Figure 1E) (Kang et al., 2016; Wang et al., 2019).</p><p>Porosity of the membranes are tested and calculated by Equation (1) as shown in Figure 2A, it's about 54.4, 66.1, 68.5, 72.5, and 74.0% for Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL, respectively. Obviously, fibrous separators have higher porosity than Celgard 2325, which may contribute in electrolyte uptake and thus enhance the rate capability. Contact angle of liquid electrolyte to various membranes were also studied and shown in Figure 2B. The contact angle of Celgard 2325, PVDF-HFP, PHL, PHP and PHP@PHL is 41.30, 28.48, 21.52 13.51, and 15.61°, respectively. All fibrous membranes showed smaller contact angle with electrolyte for the high dielectric constant of PVDF-HFP (Lopez et al., 2018), besides, they also beneficial from the fibrous structure. Smaller contact angle indicated that the PHP@PHL composite nanofibrous membrane has better affinity and is easier to be wetted by the liquid electrolyte, thus better rate capability might be obtained.</p><!><p>Electrolyte uptake and porosity (A), as well as the contact angle images with liquid electrolyte droplet (B) of the Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL separators.</p><!><p>The effect ionic conductivity of the five separators was probed by EIS test at ambient temperature and calculated by Equation (3) as shown in Figure 3A. Rb obtained from EIS data is 2.1105, 3.8473, 2.5165, 3.3936, and 0.6801 Ω for Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL, respectively, and the corresponding effect ionic conductivity are 0.63, 0.64, 1.03, 1.02, and 4.05 mS/cm, respectively. Consistent with the results of porosity and wettability, this data further guaranteed the enhanced rate capability in cells.</p><!><p>(A) Nyquist plot of AC impedance measurements (SS/separator/SS) of the Celgard 2325, PVDF-HFP, PHL, and PHP@PHL separators; Lithium ion transference numbers of (B) the Celgard 2325, (C) PVDF-HFP, (D) PHL, (E) PHP, and (F) PHP@PHL membranes as demonstrated by CA polarization curves and EIS plots before and after polarization.</p><!><p>The lithium ion transference number (tLi+) of Celgard 2325 and the as-prepared membranes with liquid electrolyte was tested and calculated via Equation (4) as shown in Figures 3B–F, where the corresponding tLi+ are 0.42, 0.39, 0.56, 0.45, and 0.62, respectively. tLi+ reduced from 0.42 to 0.39 when Celgard 2325 was replaced by PVDF-HFP, which proves that the polymer itself in fact has little contribute in Li+ transference. The addition of Li2SiO3 enhanced the number to 0.56, which demonstrated that the inorganic nanoparticle may facilitate the desolvation of Li+. While tLi+ of PHP is only 0.45, slightly higher than that of Clegard 2325 and PVDF-HFP, the reason lays on that although PPCl was mixed with PVDF-HPF and formed a membrane, but the Cl anion could still partly dissociated into the electrolyte thus influence the transportation of Li+ (Xu et al., 2018). PHP@PHL shows the highest tLi+ among the five electrolytes for the synergistic effect of LSO and PPCl, encapsulated by the PHL layer, the Cl anion could not influence the Li+ transport but adjust the polarization of the separator polymer matrix. Therefore, PHP@PHL nanofiber membrane prominently improves the tLi+, which may contribute in reduces anion aggregation near the electrode surface, thus reduces the concentration polarization and stabilizes the electrochemical deposition of lithium ions, improving the safety of LIB.</p><p>The film thickness, electrolyte uptake, porosity, ionic conductivity, and lithium ion transference number of the Celgard 2325, PVDF-HFP, PHL, and PHP@PHL separators summarized in Table 1. It is worth noting that the electrolyte uptake of PHP@PHL was about 597%, which is much higher than that of Celgard 2325 (70%), PVDF-HFP (366%), PHL (466%), and PHP (498%) and further contribute in the better rate capability when adopting in the LIBs. PHP@PHL composite membrane demonstrated the highest porosity and electrolyte uptake ability, the smallest electrolyte contact angle, and there are several factors that contribute to its remarkable performances. Firstly, the molecular structure of PVDF-HFP, as well as the 3D nanofibrous morphology facilitated the electrolyte uptake, for its high polarity and the nanoporous structure. Secondly, Li2SiO3 and PPCl reduced the order of molecular arrangement, which produced more chances for the electrolyte to be taken. Moreover, PPCl also worked as a plasticizer which has perfect affinity with the electrolyte, on which the contact angle is also reduced. Therefore, PHP@PHL shows the best physical performances. We also compare the ionic conductivity and the tLi+ with some of the previous publications as listed in Table 2, obviously, by the synergistic effect of LSO and PPCl, lithium ion transport efficiency was strongly enhanced.</p><!><p>The film thickness, electrolyte uptake, porosity, effective ionic conductivity, and lithium ion transference number of Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL separators.</p><p>Ionic conductivity and tLi+ in previous relevant works.</p><!><p>Galvanostatic cycling measurements in Li/electrolyte/Li symmetrical cells can probed into lithium plating/stripping process and analyze the interfacial stability between the electrolyte and lithium electrode. The charge/discharge cycling test was performed at a fixed current density of 0.5 mA/cm2 with a total capacity of 1 mAh/cm2 and the results are shown in Figure 4. As can be seen from the Celgard 2325 voltage-time profile, it exhibits a gradual increase in hysteresis (overpotential between Li deposition and dissolution) as the time increases, and the voltage increases to 0.65 V after 300 h. In other words, the SEI film, or dead lithium is thickening with cycling, which means an out-off-balance lithium plating/stripping. The voltage-time profiles of PVDF-HFP, PHL, and PHP@PHL electrolytes were much stable compared with that of Celgard 2325. In particular, although the tLi+ of PVDF-HFP is lower than that of Celgard 2325, while it turns quite stable in the initial 500 h owing to the fibrous structure of the membrane which could balance the Li+ flux adjacent to the electrode. PHL presents the least hysteresis after 1,000 h plating/stripping, while as it shows several abnormal convex during the cycling, therefore, we consider that the PHP@PHL cell demonstrated the best stability. As discussed before, the enhanced performance of PHP@PHL can be attributed to several reasons, firstly, the nanofibrous structure of PVDF-HFP balanced the Li+ flux and suppressed the mossy deposition of Li; secondly, the synergistic effect of LSO and PPCl reduce the crystallinity of the polymer matrix, thus enhanced the Li+ transportation efficiency; lastly, Cl anion from the IL may partly be involved in the formation of SEI, which further enhance the cycle stability of the symmetric cell. Hence, the formation of Li dendrites is inhibited and the speed of dendrite growth is reduced as well. These results manifest that using 3D porous electrostatic spinning nanofiber membrane could effectively lower the hysteresis, stabilize cycling behavior and elongate the cell lifetime.</p><!><p>Galvanostatic cycling performances of symmetric lithium cells with the Celgard 2325, PVDF-HFP, PHL, and PHP@PHL separators at a current density of 0.5 mA/cm2 at 25°C.</p><!><p>The batteries tested with the above four electrolytes were then disassembled in an argon-protected glove box, the electrodes were cleaned with diethyl carbonate (DEC) and dried strictly, and then put into a vacuum transfer box before FE-SEM analysis. The experimental results are shown in Figure S3. For the symmetrical lithium battery assembled with Celgard 2325 after galvanostatic charge/discharge cycle for 300 h, the FE-SEM diagram indicates that a lot of clavate-shaped lithium dendrite is generated on the surface of the lithium electrode. These dendrites will eventually penetrate the separator, causing short circuit and safety problems. By contrast, the symmetrical lithium battery assembled with the electrostatic spinning films, PVDF-HFP, PHL, and PHP@PHL, even cycled for 1,000 h under the same current density, shows much smoother surface with smooth edges, which was not easy to penetrate the separators, greatly improved the safety of LIBs. The electrospinning separators have large specific surface area, which could balance the local current density on the electrode surface, thus inhibited the formation and growth of lithium dendrite. Therefore, the core-shell nanofiber separator prepared by coaxial electrostatic spinning can significantly improve the safety performance of LIB.</p><p>Linear sweep voltammetry (LSV) was used to investigate the electrochemical windows of the electrolytes with a scan rate of 0.5 mV/s over the range of 2.8–6 V in Li/electrolyte/SS battery, and the results are shown in Figure 5. The oxidation potentials of the Celgard 2325, PVDF-HFP, PHL, and PHP@PHL electrolyte is 4.27, 3.93, 5.5, and 5.45 V (vs. Li+/Li), respectively. As all these membranes are plasticized by the same electrolyte, LiPF6/EC+EMC, therefore, most parasitic reactions should origin from the membrane. Therefore, it can be concluded that both LSO and PPCl could prevent PVDF-HFP from being oxidized and effectively enhance the electrochemical stability of composite membrane, which is pivotal for practical application. No obvious decomposition of Li/PHP@PHL/SS battery was observed below 5.45 V, which manifested PHP@PHL may potentially be able to be coupled with high voltage cathode materials, such as NCM622.</p><!><p>Linear sweep voltammograms of the four electrolytes at a scan rate of 0.5 mV/s (Li/electrolyte/SS).</p><!><p>The interfacial stability between the electrode and the electrolyte is fatal for long-term cycle stability and rate performance, therefore, the separators are adopted and Li/electrolyte/NCM622 half-cells were assembled to further test their compatibility with the electrode materials. Static EIS tests were performed in 1, 10, 20, and 30 days after cell assembled at room temperature and the data was deal with an equivalent circuit fitting (Figure S4), the results are shown in Figure 6 and Table S1. Rb is corresponding to the bulk resistance originated from electrolyte and other cell components, R1 and R2 can be assigned to the interface resistance between electrolyte and anode or cathode, respectively. R2 is smaller than R1 because the battery is never charged after assembly and is left standing, therefore, little reaction would happen in the storage process. The cell with PHP@PHL shows the most stable R1 value among the four cells, it is 18.32 Ω after 30 days of storage, while it is 61.40, 35.45, and 35.42 Ω for Celgard 2325, PVDF-HFP and PHL, respectively, which indicates that PHP@PHL has the best interfacial compatibility with Li among these electrolytes (Cheng et al., 2018). Its smallest increase in interface resistance is mainly attributed to the following reason: firstly, PVDF-HFP has higher dielectric constant than PP or PE, which endows it with better Li compatibility (Lopez et al., 2018); secondly, the introduction of Li2SiO3 partially suppresses the decomposition of LiPF6 and carbonate solvents during long term storage (Fu et al., 2017); and the third, the slow dissociation of PPCl would generate some Cl anion, which would react with Li to form a more stable SEI component, LiCl, thus further stabilized the electrode/electrolyte interface and reduce its resistance (Lu et al., 2014b).</p><!><p>Impedance evolution of the NCM622/Li half-cells with (A) the Celgard 2325, (B) PVDF-HFP, (C) PHL, and (D) PHP@PHL membranes at open circuit potential as a function of storage time at 25°C for 1, 10, 20, and 30 days.</p><!><p>Electrochemical performances of the above mentioned GPEs were tested in Li/electrolyte/NCM622 half cells by different C rates. Figure S5 shows the 0.5C rate performance of Celgard 2325 and PHP@PHL, we can see that the cells delivers quite similar capacity in the initial 300 cycles, both cells kept more than 100 mAh/g reversible capacity in 500 cycles, which proves that the cathode material works well in this electrolyte system. It also can be notice that the Celgard one shows lower coulombic efficiency after about 350 cycles, which might cause by the mossy Li deposition after long term striping/plating. High rate tests were performed to testify the performance of nanofibrous GPEs. Figures 7A,B show the long term cycle performances of the Li/electrolyte/NCM622 half cells at 3C and 5C, respectively. Obviously, the reversible capacity and cycle stability of PHP@PHL system were significantly improved at these C rates. The initial discharge capacities of Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL were 156.6, 160.6, 161.4, 156.5, and 150.6 mAh/g at 3C rate, respectively. It drops sharply to 15.8 mAh/g after 300 cycles for the Celgard 2325 one, while that of PVDF-HFP, PHL, PHP, and PHP@PHL maintained at 85.3, 87.9, 84.1, and 98.5 mAh/g, respectively. The initial discharge capacities of Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL at 5C rate were 156.0, 156.0, 156.2, 158.7, and 160.2 mAh/g, respectively. It declined dramatically to 27.8 mAh/g after 300 cycles for Celgard 2325, and that of PVDF-HFP and PHL were also significantly reduced to 26.3 and 27.2 mAh/g after 400 cycles, respectively. But the one with PHP@PHL remained 78.5 mAh/g even after 500 cycles. The rate performances (from 0.2 to 20C) of Li/electrolyte/NCM622 half-cells are shown in Figure 7C. It can be seen that the difference between the electrospining GPEs and the Celgard separator was not obvious under low current density, but the diversity significantly improved under large current density, especially for PHP@PHL. The Li/electrolyte/NCM622 half-cell with PHP@PHL can still deliver a reversible capacity of 65 mAh/g at 20C, while that of Celgard 2325, PVDF-HFP, PHL, and PHP were 5, 30, 38, and 29 mAh/g, respectively, which means the PHP@PHL delivers 13 times higher capacity than that of Celgard 2325. Charge-discharge profiles of the Li/electrolyte/NCM622 half-cells with Celgard 2325 and PHP@PHL at different C rates were compared in Figure S6. It can be easily conclude from these results that the as prepared core-shell structure PHP@PHL nanofibrous electrolyte has the best electrochemical performances, which can realize safe and efficient quick charging, thus has great potential in the application of high-power density LIBs.</p><!><p>Long-term cycling stability of the NCM622/Li cells with different membranes (Celgard 2325, PVDF-HFP, PHL, PHP, and PHP@PHL) at 3C rate (A), 5C rate (B), and their rate capability (C).</p><!><p>In summary, a core-shell structured nanofibrous membrane with PPCl ionic liquid plasticizer in the core and inorganic nano-particle Li2SiO3 as the filler in shell, PHP@PHL, was prepared by a facile coaxial electrospining method. The electrolyte uptake, porosity, ionic conductivity and lithium ion transference number of PHP@PHL nanofiber membrane were about 597%, 74.0%, 4.05 mS/cm, and 0.62, respectively. By the introduction of PPCl and Li2SiO3, ionic conductivity of the electrolyte increases by an order of magnitude and wettability between the separator and liquid electrolyte also significantly improved compared to commercial Celgard 2325 or even electrospining PVDF-HFP separator. In addition, it has good thermal stability, low interfacial impedance and wide electrochemical window. Symmetrical lithium cells with PHP@PHL demonstrated excellent plating/stripping cycling stability for about 1,000 h without short-circuit, which is 5 times longer than that of Celgard 2325. Moreover, the PHP@PHL electrolyte presents outstanding rate capability, it delivers a reversible capacity of 65 mAh/g at 20C compared to the 5 mAh/g in the case of Celgard 2325. The long term cycle performance was also significantly improved, as demonstrated in 3C and 5C. The core-shell structured nanofibrous membrane, in which different fillers or plasticizers could be used thus different functions can be realized in one single membrane, provides and effective method to enhance the overall performances of LIBs.</p><!><p>The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.</p><!><p>SZ helps to improve ideas and experimental platform. LZ helps to solve some problems of experience. XL helps to do mainly works. YR helps to do any experience.</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><!><p>Funding. This work was financially supported by the National Key Research and Development Program of China (No. 2016YFB0100303), National Natural Science Foundation of China (No. 21706261), Beijing Natural Science Foundation (No. L172045), and the Ford-China University Research Program.</p><!><p>The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2019.00421/full#supplementary-material</p><!><p>Click here for additional data file.</p>

PubMed Open Access

Polarizable Density Embedding for Large Biomolecular Systems

We present an efficient and robust fragment-based quantum-classical embedding model capable of accurately capturing effects from complex environments such as proteins and nucleic acids. This is realized by combining the molecular fractionation with conjugate caps (MFCC) procedure with the polarizable density embedding (PDE) model at the level of Fock matrix construction. The Fock matrix of the core region is constructed using the local molecular basis of the individual fragments rather than the supermolecular basis of the entire system. Thereby, we avoid complications associated with the application of the MFCC procedure on environment quantities such as electronic densities and molecular-orbital energies. Moreover, the computational cost associated with solving self-consistent field (SCF) equations of the core region remains unchanged from that of purely classical polarized embedding models. We analyze the

polarizable_density_embedding_for_large_biomolecular_systems

<p>performance of the resulting model in terms of the reproduction of the electrostatic potential of an insulin monomer protein and further in the context of solving problems related to electron spill-out. Finally, we showcase the model for the calculation of oneand two-photon properties of the Nile Red molecule in protein environments. Based on our analyses, we find that the combination of the MFCC approach with the PDE model is an efficient, yet accurate approach for calculating molecular properties of molecules embedded in structured biomolecular environments.</p><p>Quantum chemical methods enable calculations of a wide range of molecular properties with high accuracy. 1 However, the high computational cost and steep scaling of conventional quantum chemical methods limit practical applications to single molecules or smaller molecular clusters. Fortunately, in many molecular systems, including biomolecular ones, the electrons are fairly localized. For such systems, it becomes feasible to calculate local molecular properties by treating only a smaller part of a composite system with a quantum-mechanicsbased method. For example, a large part of the electronic and vibrational spectrum of a solvated chromophore can be obtained from calculations where it is only the chromophore itself that is described by quantum mechanics. Nonetheless, the environment can substantially alter the properties of the central part and, therefore, cannot be neglected in general.</p><p>Environmental effects can be included effectively through the use of embedding methods.</p><p>Here the total system is partitioned into two subsystems: the core part, which is the system of main interest, and its environment. The central component of an embedding method is the embedding potential produced by the environment, through which the wavefunction of the embedded core part interacts with its environment. Many different embedding methods exist, and these can be roughly divided into three groups according to their description of the environment: structure-less continuum embedding, 2,3 classical atomistic embedding, [4][5][6] and quantum embedding. [7][8][9] Continuum models have proven to be highly successful for capturing bulk solvent effects due to their efficiency and ease of use. However, they are less suited for heterogeneous environments and systems that involve highly directional interactions be-tween the core and its environment, such as hydrogen bonds or π-interactions, which are typical characteristics of biomolecular systems. For such systems, retaining an atomistic environment will perform better. This is the strategy used by quantum mechanics/molecular mechanics (QM/MM)-type embedding models 10 as well as more advanced variants, such as the hybrid quantum mechanics/effective fragment potential (QM/EFP) model [11][12][13][14][15] and the polarizable embedding (PE) model. [16][17][18][19] One of the main limitations of classical atomistic models is the lack of exchange repulsion, which results in an inadequate description of short-range interactions. It manifests itself as electron spill-out and spurious molecular orbitals. [20][21][22][23][24][25][26] To avoid these issues, quantum embedding models, which include exchange repulsion in some form, can be used. However, their applicability to biomolecular systems is limited by a comparably high computational cost; i.e., while such models are very efficient compared to conventional quantum chemistry methods, they are still too expensive compared to the requirements of, e.g., spectroscopic applications on large biomolecular systems, which easily consist of several tens or even hundreds of thousands of atoms and, moreover, require statistical sampling and thus evaluation of a large number of structures. There are thus limits on the size of systems that can be treated and the level of theory that can be used to describe the environment. The main bottleneck stems from the fact that either a quantummechanics-based calculation on the full system is required or, when relying on fragmentation procedures to split the environment into smaller fragments, an iterative procedure is needed in order to include environment polarization. The latter iterative procedure involves successive quantum-mechanics-based calculations on the individual fragments (often referred to as freeze-and-thaw cycles), and, while it is not strictly required, it is often important in order to obtain good accuracy.</p><p>The polarizable density embedding (PDE) model 22 is a fragment-based mixed quantumclassical embedding model where the permanent charge distributions of environment fragments are described quantum mechanically by their electronic densities and nuclear charges, while the induced charge distributions are described classically through an induced-dipole model based on localized dipole-dipole polarizabilities. It was introduced as an extension of the PE model in order to improve electrostatic interactions and to avoid electron spill-out.</p><p>The classical description of polarization mitigates the high cost of pure quantum embedding and provides a relatively accurate description of polarization. Exchange repulsion is modeled by a Huzinaga-Cantu-like projection operator 27 which effectively solves the electron spillout problem. 24 The total embedding operator thus consists of three terms that describe the electrostatic, induction, and exchange-repulsion interactions between the core region and its environment.</p><p>The electrostatic operator (v es ) describes Coulomb interactions between the core region and the permanent charge densities of the environment. It is constructed from the fragmentbased density matrices (D) as well as nuclear charges (Z) and positions (R), together with the corresponding two-electron repulsion integrals (v µν,γδ ) and one-electron nuclear attraction integrals (v µν (R)). The contribution of this operator to the Fock matrix in an atomic orbital (AO) basis is defined as</p><p>where µ and ν, here and in the following equations, are indices of AOs belonging to the core region and γ and δ belong to AOs of fragments in the environment. The first term of eq 1 accounts for the repulsion between electrons in the core region and electrons in the environment, while the second term accounts for the attraction between electrons in the core region and nuclei in the environment. The density matrices are optimized for each fragment in isolation and then kept frozen during the optimization of the core-region electronic density, consistent with the fact that the polarization is described classically (as described below).</p><p>The induction operator (v ind ) takes into account the polarization of the environment through localized dipole-dipole polarizabilities (α) which are calculated for each fragment in isolation. Apart from the polarizabilities, the electronic electric-field integrals (t µν (R))</p><p>are also needed to construct the operator. Its contribution to the Fock matrix is given by</p><p>where the induced dipoles (µ) are determined by the total electric field (F) exerted on the polarizable sites, i.e. µ s = α s F(R s ). Mutual polarization between the core region and environment is included by updating the induced dipoles during the optimization if the core-region electronic density.</p><p>The exchange-repulsion operator (v rep ) is based on a Huzinaga-Cantu-like projection 27 using energy-weighted density matrices (W) and orbital overlaps (S). The corresponding Fock matrix element is written as</p><p>where W γδ is an element of the density matrix of a fragment in the environment weighted by the occupied molecular orbital (MO) energies (ε), i.e., W γδ = N f occ i=1 ε i C γi C δi with C being MO coefficients. This operator models the effect of exchange repulsion and thus effectively prevents the electrons in the core region from penetrating into the environment by modeling the effect of exchange repulsion. It is approximate because we do not solve the Huzinaga-Cantu equations, 27 which would require either a calculation on the full system or an iterative procedure. Instead, the fragment densities are determined in isolation, and the repulsion is dampened by a transferable prefactor. 22,28 Exchange repulsion is a crucial component of the PDE model (and generally in quantum embedding) needed to avoid variational collapse, whereas it can sometimes be neglected when using conventional classical embedding, though with the risk of erroneous results.</p><p>In practice, the quantities in eqs ( 1) and ( 3) are computed in the local molecular basis of the individual fragments (and not the supermolecular basis). The evaluation of these operators therefore scales linearly with the number of fragments in the environment. In the present implementation, the evaluation of the induction operator in eq (2) scales quadratically with the number of polarizable sites. However, this can be reduced to linear scaling by introducing, e.g., a fast multipole method 29,30 for the evaluation of the electric fields from the induced dipoles.</p><p>The PDE model has so far only been considered for solute-solvent systems. [22][23][24][25]28,31,32 When tackling large molecular systems, fragmentation of the environment can be used to circumvent the poor scaling of the underlying methods used to derive the embedding potential. This is the approach used in both the PE and PDE models, as detailed above for the latter. For solute-solvent systems, the selection of individual molecules as fragments is straightforward, but more care is needed when treating typical biological macromolecules such as proteins, which often contain several thousands of atoms. For such systems, the fragmentation method needs to handle the cutting of covalent bonds. In the PE model, inspired by the work of Ryde and coworkers, 33,34 the method of molecular fractionation with conjugate caps (MFCC) 35 was adopted to partition larger molecules into small, computationally manageable fragments. Specifically, Söderhjelm and Ryde 33 used the method to calculate localized properties (multipoles and polarizabilities) of proteins in order to derive a highly accurate force field.</p><p>Zhang and Zhang 35 first introduced the MFCC method in the context of calculating protein-ligand interaction energies 35 and later it was also used for nucleic-acid-ligand interaction energies. 36 In short, the MFCC method proposes the following: In order to calculate protein-ligand interaction energies, the protein is partitioned into a set of amino-acid-based fragments by cutting the peptide bonds (see figure 1). To saturate the open valencies caused by cutting covalent bonds, the fragments are capped by groups that consist of a number of atoms from the neighboring fragments. The procedure thus leads to a set of (partly) overlapping fragments. The capping groups of adjacent capped fragments are combined to create new fragments called conjugate caps (concap). The total protein-ligand interaction energy is then approximated as the sum of the capped-fragment-ligand interaction energies, minus the sum of the concap-ligand interaction energies, thus removing double-counting due to the overlapping groups. The accuracy of the method can be systematically improved by increasing the size of the capping groups. Shortly after the original method was presented, it was extended to (real-space) electronic densities (and thereby total energies), electrostatic potentials, and total dipole moments, 37 and later the method was further extended in many different directions. [38][39][40][41][42][43][44][45][46][47] Order Capped fragments Concap Applying the MFCC procedure for the individual quantities of the embedding potential in the PDE model is not straightforward for the construction of the electrostatic and exchangerepulsion operators (eqs 1 and 3 respectively). Whereas MFCC can be directly applied to derive the localized polarizabilities needed for the induction operator (eq 2), it is more involved for fragment density and overlap matrices, and it is entirely unclear how to handle MO energies and coefficients. However, the operators only contain contributions that depend on one fragment, f , of the environment at a time. Therefore, the fragmentation procedure can be simply realized at the level of Fock matrix contributions, which is the strategy we present here. The electrostatic and exchange-repulsion contributions to the Fock matrix are first calculated fragment-wise, after which all of the contributions are combined according to the MFCC approach. The Fock matrix contributions are thus obtained according to</p><p>where x indicates either the electrostatic or exchange-repulsion contribution, f and c are indices associated with the capped fragments or concaps, respectively. The matrix elements v x,f µν and v x,c µν are calculated according to eqs 1 and 3.</p><p>The fragmentation and application of the MFCC procedure are implemented in the PyFraME 48 package, while the PDE model is implemented in the Polarizable Embedding library (PElib), 49 which is interfaced to the Dalton program. 50 Our implementation relies on the HDF5 file format 51 for storage and transfer of embedding quantities, which is expected to permit relatively easy incorporation of the model into other quantum-chemistry codes.</p><p>The implementations are made available together with the Dalton2020 release.</p><p>Before moving on, we note that this is not the first use of MFCC in the context of embedding. Apart from the PE model and the work by Söderhjelm et al., as mentioned above, the MFCC procedure has also been applied to frozen density embedding (FDE) 52 by Jacob and Visscher 53 who presented a three-partition FDE (3-FDE) model. 53,54 In this model, the density is mapped on a real-space grid, and the MFCC procedure can, therefore, be straightforwardly applied. However, the density can become negative in the cappinggroup regions, which is problematic for a model that employs density-functional exchangecorrelation kernels. To avoid this issue, special constraints are required to ensure positive densities. 53 Benchmarks of the 3-FDE scheme showed good performance in terms of the quality of electronic densities and electrostatic potentials of peptides and proteins. 55 It has also been used to investigate environmental effects in protein systems. 56,57 One of the most important aspects of a classical embedding model is the ability to reproduce the electrostatic potential (ESP) created by the environment. For purely classical embedding, the ESP is the main factor determining the quality of the environment description. 58 Figure 2 shows the errors in the ESP of an insulin monomer relative to a DEC-MP2 59,60 /cc-pVDZ reference 61 for two density-functional approximations and the PE and PDE potentials. The ESP errors are mapped onto a van der Waals (vdW) surface of the protein, with larger errors leading to saturated colors (blue or red). For this example, and in the following, an MFCC order of 2 was adopted (see figure 1), since initial tests revealed that using larger orders did not provide substantial improvement (see the Supporting Information). The PDE-and PE-based ESPs are derived from fragment calculations using CAM-B3LYP 62 /cc-pVDZ, i.e., using the same basis set as was used in the reference MP2 calculation. As reported by Jakobsen et al., 61 it is immediately apparent that large errors in the ESP are found when using the B3LYP functional. The poor reproduction of the ESP occurs due to severe self-interaction errors leading to incorrect delocalization of electrons, which may be exacerbated by the fact that there is not a stabilizing environment surrounding the protein. 55 This highlights the need for using range-separated exchange-correlation functionals when treating large molecular systems. Accordingly, the range-separated CAM-B3LYP functional yields an ESP that is almost identical to the reference MP2-based ESP.</p><p>The ESP errors in Figure 2 Beyond a better reproduction of ESPs, the major advantage of the PDE model over classical embedding models is the inclusion of exchange repulsion. The host-guest complex of 2,3-diazabicyclo[2.2.2]oct-2-en (DBO) chromophore inside cucurbit [7]uril (CB [7]) has previously been shown to suffer from electron spill-out, 26,63 with excitation energies being extremely redshifted when using the PE model to describe CB [7]. For this system, we calculated reference excitation energies based on supermolecular time-dependent Hartree-Fock (TDHF) and compared them to PE-TDHF and PDE-TDHF excitation energies of the DBO molecule in CB [7] (see table 1). When using a non-augmented basis set (cc-pVDZ), both PE-and PDE-TDHF accurately reproduce the two lowest singlet excitation energies, which appear at around 3.5 and 6.5 eV for the supermolecular TDHF calculation. Both PE-and PDE-TDHF slightly overshoot the excitation energies, more so for the second state than the first (0.10 and 0.15 eV using PE, and 0.05 and 0.10 eV using PDE), but are in good quantitative agreement with the supermolecular reference. Upon augmentation of the basis set with diffuse functions (aug-cc-pVDZ 64 ), the supermolecular TDHF excitation energies change only very slightly (less than 0.01 eV). Meanwhile, using PE-TDHF results in two extremely low-lying states at 1.5 and 1.8 eV, which are clearly unphysical. The correct behavior is restored by including exchange repulsion through the PDE model, which gives excitation energies that still are in good agreement with the supermolecular TDHF reference (the errors are the same as for cc-pVDZ). In this case, the problem with the PE model arises only when using diffuse basis functions since compact basis sets simply do not extend far enough into the environment to make the problem apparent. With the introduction of the MFCC method, the PDE model is now applicable to large biomolecular systems. As an illustrative example, we consider the case of the Nile Red chromophore embedded in the binding pocket of the β-lactoglobulin (BLG) protein. This system has been studied previously 65,66 due to the strong solvatochromism of the Nile Red dye. Figure 3 shows simulated one-and two-photon absorption spectra of Nile Red embedded in the protein and in the gas-phase. For the condensed-phase spectra, configurational sampling is included by extracting 125 snapshots from a QM/MM molecular dynamics (MD) trajectory. 66 In all cases, the singlet excitation energies and associated one-and two-photon transition strengths are computed using the CAM-B3LYP functional along with the 6-31+G* basis set. Local-field effects are taken into account through the effective external field (EEF) method. 67 The embedding-potential parameters are computed at the same level of theory.</p><p>The spectra are generated using a Lorentzian broadening function for each transition with a line-width of 1000 cm −1 . The condensed-phase spectra are averaged across 125 snapshots, while the gas-phase spectra are computed on the gas-phase optimized structure of Nile red.</p><p>The one-photon spectrum of Nile Red in the gas-phase has a bright transition at 2.95 eV.</p><p>At higher energies, the gas-phase spectrum shows several distinct peaks with low intensity between 3.5 and 5.3 eV, followed by a major peak (composed of multiple transitions) at 5.7</p><p>eV. Introducing the protein and solvent broadens and redshifts the lowest band (by about 0.2 eV), irrespective of the model used to describe the environment. For this transition, the absorption bands are almost identical. For the transitions at higher excitation energies, the results are also similar, with several low-intensity bands between 4-5 eV, followed by an intense band at 5.5 eV (corresponding to a 0.2 eV redshift compared to the gas-phase result).</p><p>However, unlike for the first transition, there are visible differences between the spectra where the environment is described by PE or PDE. In particular, in the low-intensity regions where the spectrum exhibits broader and less well-defined transition bands when the environment is modeled using PE. This difference can be understood by considering the underlying density of states, as shown in figure 4. The lowest-lying excitations (from 2.5-4.5 eV) are similar in both cases and relatively well-separated. When moving towards higher excitation energies, however, the spectrum based on the PE model becomes plagued by a much higher density of (partially unphysical) states. As a result, the state distribution, when using the PE model, shows a less well-defined structure. Moreover, to reach the same energetic range (i.e., converging the spectrum up to 6 eV) when using the PE model requires the solution of a larger number of states (40) than when using the PDE model (20), which can be seen from the number of color alternations along the y-axis of the histograms in figure 4. Overall, the PE and PDE models show almost identical performance for the low-energy, bright excitation in Nile Red, and also yield quite similar one-photon spectra for higher energies, although the underlying physical description is clearly less satisfactory when using PE to model the environment.</p><p>The simulated two-photon spectra of Nile Red shown in figure 3 reveals larger differences between the PE and PDE models. The spectra are generated from the five lowest singlet excited states. The gas-phase spectrum is composed of three intense transitions located at 1.5, 1.9, and 2.1 eV. Both embedding models predict that the lowest-energy transition is redshifted by 0.1 eV when including the protein and solvent. Moreover, the intensities of this lowest excited state are almost identical. When moving towards higher energies, the spectrum based on the PDE model retains the basic characteristics of the gas-phase spectrum, although (due to broadening) the two peaks from the gas-phase now appear as one peak with a shoulder. In contrast, the spectrum based on the PE model shows no clearly defined structure and instead shows a broad featureless band starting from 1.6-2.1 eV. A correct description of the two-photon transition probability requires an accurate description of not only the initial and final states involved in the transition, but in principle, also all other excited states. 68,69 Thus, the artificial stabilization of higher-lying states and resulting overly large density of states at higher energies (see figure 4), is the root cause of the broader spectra found when using PE to model the environment. Overall, we find that the two-photon absorption spectrum is clearly more sensitive to the environment description than the onephoton absorption spectrum and that the PDE model yields a more physically reasonable result.</p><p>To sum up, we have presented a formulation of the PDE model that is suitable for biomolecular systems, which include large biomolecules such as proteins and nucleic acids.</p><p>This formulation of PDE was realized based on the MFCC scheme at the level of Fock matrix construction, thus avoiding complications and issues related to the application of the fragmentation procedure for fragment quantities such as density matrices and MO energies. We have shown that the PDE potential produced via the MFCC procedure can accurately repro-</p>

ChemRxiv

Investigation of the Charge-Transfer Between Ga-Doped ZnO Nanoparticles and Molecules Using Surface-Enhanced Raman Scattering: Doping Induced Band-Gap Shrinkage

Semiconductor nanomaterial is a kind of important enhancement substrate in surface-enhanced Raman scattering (SERS), and the charge-transfer (CT) process contributes dominantly when they are used as the enhancement substrate for SERS. Doping has significant effect on the CT process of semiconductor nanomaterials. Yet till now, none attempts have been made to explore how doping affects the CT process between the semiconductor and probe molecules. For the first time, this paper investigates the effect of gallium (Ga) doping on the CT process between ZnO nanoparticles and 4-mercaptobenzoic acid (4-MBA) monolayer. In this paper, a series of Ga-doped ZnO nanoparticles (NPs) with various ratio of Ga and Zn are synthesized and their SERS performances are studied. The study shows that the doped Ga can cause the band gap shrinkage of ZnO NPs and then affect the CT resonance process form the valence band (VB) of ZnO NPs to the LUMO of 4-MBA molecules. The band gap of Ga-doped ZnO NPs is gradually narrowed with the increasing doping concentration, and a minimum value (3.16 eV) is reached with the Ga and Zn ratio of 3.8%, resulting in the maximum degree of CT. This work investigates the effects of doping induced band gap shrinkage on CT using SERS and provides a new insight on improving the SERS performance of semiconductor NPs.

investigation_of_the_charge-transfer_between_ga-doped_zno_nanoparticles_and_molecules_using_surface-

Introduction<!>Chemicals<!>Synthesis of ZnO and Ga-Doped ZnO NPs<!>Adsorption of Probing Molecules<!>Sample Characterization<!>Measurement of XRD<!><!>Measurement of XPS and ICP<!>Measurement of TEM Images and UV-vis Spectra<!><!>Measurement of TEM Images and UV-vis Spectra<!><!>Raman Spectra of ZnO and Ga-Doped ZnO NPs<!><!>SERS Spectra of 4-MBA Adsorbed on Pure and Ga-Doped ZnO NPs<!><!>Enhancement Mechanism and the Effect of Doped Ga on CT of ZnO NPs<!><!>Enhancement Mechanism and the Effect of Doped Ga on CT of ZnO NPs<!>Conclusions<!>Data Availability<!>Author Contributions<!>Conflict of Interest Statement<!><!>Supplementary Material<!>

<p>The interfacial charge-transfer (CT) process between a substrate and an adsorbed molecule is an interesting phenomenon which has attached extensive research. The CT process has a wide range of applications in interface chemistry (Osako et al., 2018), catalytic chemistry (Thang et al., 2018), electronic devices (Liu et al., 2018), solar cells (Yadav et al., 2018), photoelectrochemistry (Chen et al., 2018), and so on. Therefore, it is vital to conduct an in-depth investigation of the CT process, which may provide us better understanding and thus expand its applications in various fields.</p><p>Surface-enhanced Raman scattering (SERS) is a forceful method for studying the CT between adsorbed molecule monolayer and the substrate (Wang et al., 2011; Li et al., 2018; Yu et al., 2018). As is well-known, the emergence of SERS phenomenon is inseparable from enhancement substrates. With the development of relevant researches, SERS active nanomaterials have been extended from noble metals, such as Au, Ag, and so on, to semiconductor nanomaterials (NMs) (Fleischmann et al., 1974; Tian et al., 2002; Biju et al., 2008; Shen et al., 2008; Alessandri and Lombardi, 2016). Many semiconductor NMs such as TiO2 (Yamada and Yamamoto, 1983), ZnO (Wang et al., 2009), NiO (Yamada et al., 1982), InAs/GaAs (Quagliano, 2004), CdS (Wang et al., 2007b), ZnS (Wang et al., 2007a) etc., have been demonstrated to exhibit good SERS activities. In comparison with metal substrates, semiconductor NMs have additional optical and electrical properties, which enable them to display remarkable CT enhancement and catalytic abilities (Lombardi and Birke, 2009; Han et al., 2017). As we all know, for the vast majority of semiconductors, their localized surface plasmon resonance is located in the infrared region. That is to say, when semiconductor NMs was used as enhancement substrate, SERS is mainly generated through the CT enhancement mechanism (Lombardi and Birke, 2007), so the SERS performance of semiconductor NMs is a powerful tool to investigate the CT process between the semiconductor NMs and adsorbed molecules.</p><p>The CT process between the adsorbed molecules and the substrate is closely related to the size and morphology of semiconductor NMs (Musumeci et al., 2009; Tang et al., 2012; Lamberti et al., 2015). Moreover, doping is also a very important factor for the CT process when semiconductor NMs are used as the SERS substrate. Our group have reported that the SERS performance of TiO2 nanoparticles (NPs) can be improved by Fe3+, Co2+, Ni2+ doping (Yang et al., 2011, 2014; Xue et al., 2013a), and the improvements were attributed to the formation of abundant doping (defects) levels in the band gap of TiO2 NPs. Dual functional Ta-doped electrospun TiO2 nanofibers with enhanced photocatalysis and SERS activity for detection of organic compounds were synthesized by Singh et al. (2017) The generation of Ti3+ defects from Ta5+ doping, which acted as an intermediate state for TiO2 to methylene blue molecules electron transfer, was used to explain the enhanced SERS activity of their obtained products. Effects of Mn2+, Zn2+, and Mg2+ doping on SERS properties of TiO2 NPs were also studied by our group (Yang et al., 2010; Xue et al., 2012a, 2013b), it was found that an appropriate amount of metal ions dopant enriched the surface states and improved the photo-generated carrier separation efficiency. SERS characteristics of Co2+ doped ZnO (Xue et al., 2012b) and Mn2+ doped CuO (Prakash et al., 2016) were also investigated, the influences of dopant were attributed to the increasing of defects and the ferromagnetic ordering, respectively. To date, most studies about the effect of metal doping semiconductor on SERS (CT process) focused on the change of surface state level or content of defects, however, the impact of doping induced band gap shrinkage on the CT process between semiconductor NPs and molecules have never been reported. For the first time, this paper investigates the effect of doping induced band gap shrinkage on the CT between semiconductor NPs and molecules.</p><p>Due to some unmatched properties in photoelectric and magnetism, ZnO is widely used in medicine and health care, food industry, varistor, antivirus, gas sensitive elements, and photocatalysis (Zhang et al., 2009; Cushen et al., 2012; Nohynek and Dufour, 2012; Hassan et al., 2013; Lang et al., 2014; Wang et al., 2014) etc. ZnO is also an important enhancement substrate in SERS study because of its wide band gap. Numerous SERS studies around ZnO NPs have been carried out. The size-dependent effect of ZnO NPs for SERS signal (Sun et al., 2007) and ZnO/PATP(p-aminothiophenol)/Ag assemblies (Sun et al., 2008) was investigated, the SERS enhancement mechanism for ZnO was attributed to the CT enhancement mechanism. The influence of contact variation in ZnO-molecules-metal system (Mao et al., 2012), fabrication of one-dimensional ZnO/4-MPy/Ag assemblies (Hu et al., 2010), contribution of ZnO to CT induced SERS in Au/ZnO/PATP assembly (Yang et al., 2008) and so on have been reported. Above all, the investigation about CT process between ZnO NPs and adsorbed molecules is necessary in view of the wide applications of ZnO NPs. Here we study the effect of doping induced band gap shrinkage on CT process between ZnO NPs and molecules. The band gap of ZnO NPs is altered by doping gallium (Ga) into ZnO NPs, the ionic radius of Ga3+ ions (0.062 nm) is less than the Zn2+ ions (0.074 nm), so the Ga3+ ions are soluble in the ZnO matrix.</p><p>In this work, a range of Ga-doped ZnO NPs with various ratio Ga and Zn and 4-MBA@ZnO (Ga-doped ZnO) system are obtained. The actual ratios of Ga and Zn in the Ga-doped ZnO NPs are confirmed by ICP measurement. Then it is determined that the degree of crystallinity and particles size of Ga-doped ZnO NPs can be affected by the ratio of Ga and Zn according to XRD, Raman, and TEM characterizations. Moreover, it is found that the band gap of NPs shrinks as the ratio of Ga and Zn increases. The effect of Ga doping on the CT process between ZnO NPs and 4-MBA monolayer is investigated using SERS. The modest amount of doped Ga can enhance the degree of CT between ZnO and 4-MBA monolayer compared to the pure ZnO NPs. The change of CT is mainly due to the size dependence effect and the band-gap shrinkage effect. The doped Ga causes the band gap shrinkage of ZnO and then affects the CT resonance process from the valence band (VB) of ZnO NPs to the LUMO of 4-MBA molecules. This work conducts an in-depth investigation on the effects of doping induced band gap shrinkage on CT using SERS and provides a new insight on improving the SERS performance of semiconductor NPs.</p><!><p>4-Mercaptobenzoic acid (4-MBA) and Gallium(III) nitrate hydrate were purchased from Sigma-Aldrich and used without further purification. All other chemicals were acquired from Beijing Chemical Reagent Factory and used without further purification. The distilled and deionized water from a Milli-Q-plus system with the resistivity >18.0 MΩ was used in aqueous solution.</p><!><p>ZnO and Ga-doped ZnO NPs were synthesized as follows. In short, 40 mL of 0.5 mol/L NaOH solution was slowly added dropwise into 100 mL of 0.1 mol/L Zn(Ac)2 solutions under vigorous stirring in order to produce the Zn(OH)2 precipitate. Subsequently, 1.2 g of NH4HCO3 powder was added. After stirring for 30 min, a semitransparent zinc carbonate hydroxide colloid was obtained. Then the colloid was centrifuged and rinsed three times with purified water and absolute ethyl alcohol in alternation and dried at 80°C. Thus, the precursor of a small crystallite Zn5(CO3)2(OH)6 was formed. Then, the as-prepared precursor was calcined at 550°C for 2 h to obtain the ZnO NPs.</p><p>The synthetic methods of Ga-doped ZnO NPs is similar with what we described above. The difference is that there is an opportune amount of Ga(NO3)3 (0.0003, 0.0005, 0.0007, 0.0009, 0.002, 0.004, 0.006, 0.008, and 0.01 mol/L, respectively) was added to the Zn(Ac)2 solutions to obtain Ga-doped ZnO NPs with difference ratio of Ga and Zn.</p><!><p>ZnO and Ga-doped ZnO NPs surface-modified by molecules were obtained as follows: 20 mg of ZnO and Ga-doped ZnO nanocrystals were dispersed in 20 mL of 4-MBA (1 × 10−3 M) ethanol solution and the mixture was stirred for 6 h. Then, the precipitate was centrifuged and rinsed with absolute ethyl alcohol twice. ZnO and Ga-doped ZnO nanocrystals modified by 4-MBA were obtained.</p><!><p>The crystal structure of ZnO sample was determined by X-ray diffraction (XRD) using a Siemens D5005 X-ray powder diffractometer with a Cu Kα radiation source at 40 kV and 30 mA. X-ray photoelectron spectra (XPS) were obtained by using a Thermo ESCALAB 250 spectrometer with an Mg Ka excitation (1253.6 eV). Elemental analysis was carried out by ICP-AES with an Agilent 725 spectrometer. The UV-Vis DRS spectra were recorded on a Shimadzu UV-3600 spectrophotometer. Transmission electron microscopy (TEM) images were taken using a JEM-2100F high-resolution transmission electron microscopy operating at 200.0 kV. Raman spectra acquired at ambient pressure were obtained by using a Horiba-Jobin Yvon LabRAM ARAMIS system with the resolution of ca. 4 cm−1; The 633 nm radiation from a 20 mW air-cooled HeNe narrow bandwidth laser was used as exciting source. The laser beam was focused onto a spot with a diameter of approximately 1 μm using an objective microscope with a magnification of 50×. The Raman band of the silicon wafer at 520.7 cm−1 was used to calibrate the spectrometer. Data acquisition was the result of two times 30 s accumulations for the 4-MBA molecules absorbed on ZnO (Ga-doped ZnO) NPs.</p><!><p>The XRD patterns of Ga-doped ZnO NPs with different ratio of Ga and Zn are shown in Figure 1. All the diffraction peaks of pure and doped ZnO are the characteristic peak of hexagonal wurtzite ZnO (JCPDS36-1451). No characteristic peaks corresponding to Ga or Ga2O3 are observed in the diffraction patterns, which is because the concentration of gallium is too low for those impurities to be detected by the XRD instrument. Figure S1 demonstrates the relationship between full width at half maximum (FWHM) of the peak (101) and the ratio of Ga and Zn : the intensity of the peak decreases and the FWHM is increased with the ratio of Ga and Zn increasing. All these results indicate that the Ga is doped into ZnO successfully and the crystallinity of Ga-doped ZnO NPs is decreased with the increasing doping ratio. The diameter of Ga-doped ZnO NPs (0–10%) are 35.4, 30.1, 27.4, 26.2, 25.0, 22.6, 19.0, 14.4, 13.5, and 12.9 nm, respectively, calculated using Scherrer's formula (Swamy et al., 2006) based XRD data.</p><!><p>XRD spectra of ZnO and Ga-doped ZnO NPs with different ratio of Ga and Zn.</p><!><p>According to the XPS spectra which is shown in Figure S2A, we can observe that all samples show similar characteristics as the pure ZnO spectra. However, in terms of Figure S2B, the relevant peaks of Ga-dopant (Ga 2p1/2 and Ga 2p3/2 located at 1117.9 and 1144.8 eV, respectively) can be observed when the spectra is magnified in the approximate range of 1110–1150 eV, moreover, the intensities of Ga-dopant-related peaks increases with the Ga concentration increasement. Figure S2C shows the XPS spectra of Zn3/2 and Zn1/2 for pure and doped ZnO NPs, and the two peaks are observed at 1021.7 and 1044.7 eV, respectively, the data is consistent with the binding energy of Zn-O (Sano et al., 2002; Jin et al., 2009).</p><p>All the description about the ratio of Ga and Zn used previously is the initial ratio, the ICP test is conducted in order to determine the actual Ga and Zn ratio of doped ZnO NPs. The results (initial ratio 0–10%) are 0, 0.29, 0.39, 0.56, 0.66, 1.4, 2.7, 3.8, 5.0, and 6.1%, respectively. It is clear that the results of quantitative analysis from ICP is similar with the initial ratio in the low doping ratio and a significant discrepancy appears when the initial ratio is higher than 0.9%. The reason for this phenomenon is that the vast majority of Ga is doped into the lattice of ZnO at low doping ratio, and a portion of Ga is not introduced into ZnO when the content of Ga is too high. Above all, the XPS measurement proves that the Ga is doped into ZnO successfully and the actual ratio of Ga and Zn is determined by ICP.</p><!><p>Figure 2 shows the TEM images of pure and doped ZnO NPs, all the NPs are spherical and the diameters decreases with the increased Ga/Zn ratio. The particle sizes of the NPs are determined as approximate 51.4, 44.3, 35.7, 31.4, 28.6, 25.7, 22.9, 15.7, 14.3, and 12.9 nm, which are a little different from the value calculated from XRD. Such difference occurs due to the fact that the calculation results of XRD are based on the assumption that the materials are all signal crystals, and the measured values from TEM represent the size of nanoparticles which consist of one or multiple single crystal. However, the tendency of size change coincided with the result of XRD, evidencing that the particle diameter decreases as the doping concentration increases.</p><!><p>TEM images of Ga-doped ZnO NPs: (A–J) represent the initial ratio of 0, 0.3, 0.5, 0.7, 0.9, 2, 4, 6, 8, and 10% for the Ga-doped ZnO NPs, respectively. The bar is 50 nm, and it is common to all images.</p><!><p>The effect of doped gallium on the optical properties of ZnO is investigated via UV-vis absorption spectroscopy. The optical absorption spectra are shown in Figure 3A. The steep drop of the absorption at about 378 nm is assigned to the CT process between valence band (VB) and conduction band (CB). Besides, all the absorption edge of Ga-doped ZnO NPs have a red shift compared to the pure ZnO and the maximal absorption edge appeared at the initial ratio of 6%. The phenomenon is attributed to the effect of doping on the carrier density. The band gap of ZnO NPs was calculated according to the absorption edge. Due to the fact that ZnO has a direct inter-band transition (Mahdhi et al., 2015) and on the basis of practical fact and theoretical calculation, the band gap of ZnO can be obtained through the Tuac's relation (Tauc, 1968; You and Hua, 2012):</p><p>where A is the absorbance, hν is the energy of the incident photon and Eg is the band gap. The band gap is determined by plotting (Ahν)2 vs. hν and extrapolating the straight-line portion to the energy axis, the plots are shown in Figure 3B and the numerical values of Eg (0–10%) are 3.24, 3.22, 3.21, 3.20, 3.19, 3.18, 3.17, 3.16, 3.20, and 3.21 eV, respectively. According to this figure, with the increasement in doping concentration, the band gap of Ga-doped ZnO is first shrinked and down to the minimum value at the initial ratio of 6%, then increased. It should be noted that all the band gap of Ga-doped ZnO is less than the pure ZnO. Generally, there are two well-known theories about the changes of band gap of semiconductors: Burstein-Moss (BM) effect (Burstein, 1954; Moss, 1954) and band gap renormalization (BGR) effect (Dou et al., 1997; Jeon et al., 2011). The former is always related to the widening of the band gap and the latter is in connection with the shrinking-effect. The BGR effect is dominant in our system, the electron-impurity interactions, exchange interactions, and electron-electron Coulomb within the CB resulted in the shrinkage of the host band gap.</p><!><p>(A) The optical absorption spectra of pure and doped ZnO NPs, (B) Relationship between (Ahν)2 and photon energy (hν) for Ga-doped ZnO NPs.</p><!><p>The Raman spectra of pure and doped ZnO NPs are shown in Figure 4, all the Raman peaks are the characteristic peaks of hexagonal wurtzite structure. The ZnO NPs with hexagonal wurtzite phase possess C6v4 (P63/mc) space group and are simple single axial crystals. Only the optical phonon in the center of Brillouin zone is related to the first-order Raman scattering for perfect ZnO single crystal. In the point-group theory, the optical phonon modes are classified as Γopt = A1+2B1+E1+2E2, the A1 and E1 are split into transverse optical (TO) and longitudinal optical (LO) on account of they are polar modes. In all the phonon modes, A1 and E1 modes have Raman and infrared active, and that E2 modes only have Raman active, B1 modes are silent. According to the group theory and Figure 4, the peak at 333 cm−1 is the E2(high)-E2(low) mode and this mode is related to the multi-phonon scattering process. The strongest peak located at 438 cm−1 is the E2(high) mode. This is the characteristic peak of the wurtzite phase and its intensity is associated with the crystallinity of ZnO; the weak peak at 583 cm−1 is assigned to the A1(LO)-E1(LO) mode and the E1 vibrational mode is concerned with defects such as oxygen vacancy, zinc interstitial and the complex-defect and so on. The origin of another weak peak at 633 cm−1 is uncertain although it has been reported in ZnO films (Bundesmann et al., 2003; Shinde et al., 2011). Above all, all the peaks together prove the pure and doped ZnO NPs as hexagonal wurtzite phase, moreover, their crystallinity decrease with the ratio of Ga and Zn increasing, judging by the changes of E2(high) modes. All the results are consistent with the results of XRD and TEM.</p><!><p>The Raman spectra of Ga-doped ZnO NPs with various initial ratio of Ga and Zn(from a to j): 0, 0.3, 0.5, 0.7, 0.9, 2, 4, 6, 8, and 10%.</p><!><p>The SERS spectra of 4-MBA adsorbed on pure and Ga-doped ZnO NPs are obtained as shown in Figure 5A, all the peaks are the characteristic peaks of 4-MBA adsorbed on ZnO NPs and coincides with the reported results from literature (Sun et al., 2007). The strong peaks located at 1,594 cm−1 and 1,078 cm−1 can be attributed to υ8a (a1) and υ12 (a1) aromatic ring characteristic vibrations, respectively. It is worth mentioning that there is an irregular change in the peak position of 1,078 cm−1, the reason for this is attributed to the effect of the vibration mode of ZnO at 1,069 cm−1 (see the Figure S3). Other weak peaks such as 1,148 (υ15,b2) and 1,180 (υ9,a1) cm−1 are corresponded to the C-H deformation modes and agree well with the literature data (Sun et al., 2007; Xue et al., 2012b). The Raman shifts and assignments of the peaks are listed in Table 1. Moreover, it can be certified was that the 4-MBA molecules are bonded to the surface of ZnO (Ga-doped ZnO) by sulfhydryl; comparing the Raman spectrum of bulk 4-MBA with the SERS of 4-MBA adsorbed on ZnO NPs, the peaks at 913 cm−1 of 4-MBA powder disappeared when the molecules are adsorbed on the surface of ZnO. The relational graph between the intensity of the 1,594 cm−1 peak and the actual ratio of Ga and Zn is plotted and shown in Figure 5B, and an interesting phenomenon is observed. With the actual ratio increasing, the intensity increases to a first maximum value when the actual ratio is 0.39%, then the intensity decreases and down to the minimum value when the actual ratio is 0.66%. Interestingly, the intensity increases again and the second maximum value appears at the actual ratio of 3.8%. The reason for this phenomenon will be discussed in the next section. Moreover, to estimate and compare the enhancement ability of Ga-doped ZnO NPs with various ratio of Ga and Zn, the magnitude of the enhancement factor (EF) is calculated (see the Supporting Information). The EF of pure and Ga-doped (6%, initial ratio) ZnO NPs are 3.29 × 103 and 8.13 × 103, respectively, obviously demonstrating that doping Ga into ZnO NPs can significantly improve the enhancement ability of ZnO NPs. In other words, the degree of CT between ZnO NPs and 4-MBA monolayer is improved due to the Ga doping into ZnO NPs. The Ga doping induced band gap shrinkage can enhance the CT process between ZnO NPs and 4-MBA. The EF of Ga-doped ZnO NPs with other ratio is also calculated and is shown in the Table S1.</p><!><p>(A) SERS spectra of 4-MBA adsorbed on Ga-doped ZnO NPs with various initial ratio of Ga and Zn(from a to j): 0, 0.3, 0.5, 0.7, 0.9, 2, 4, 6, 8, and 10%. (B) The relationship between the intensity of the 1,594 cm−1 peak and the actual ratio of Ga and Zn for Ga-doped ZnO NPs.</p><p>Raman Shifts and Assignments of 4-MBA molecule adsorbed on ZnO (Ga-doped ZnO) NPs.</p><p>δ = bend or deformation; ν = stretch.</p><!><p>It is well-known that there are two main mechanisms widely accepted to interpret SERS: EM and CM. The former is related to the surface plasmon resonance (SPR) of the substrate, while the CT process is required for the latter. As we mentioned in the Introduction, the CT enhancement mechanism contributes to our system dominantly. In this system, the CT process occurs between ZnO NPs and the adsorbed 4-MBA molecules. All the possible CT resonance process in our ZnO-molecules system are shown in Figure 6A. The numerical values are referenced from literature data, such as: the highest occupied molecular orbit (HOMO) and the lowest unoccupied molecular orbit (LUMO) level of 4-MBA molecule are −8.48 and −3.85 eV (Yang et al., 2009), respectively; the CB and the VB level of ZnO NPs are −1.9 and −5.2 eV (Xue et al., 2012b), respectively. In addition, the Roman numerals in the figure represent different types of CT resonance: I is the exciton resonance of ZnO (Ga-doped ZnO) NPs; II is the molecule resonance of 4-MBA; III, IV, V, VI are the photon induced CT resonance from matches energy level between ZnO NPs and 4-MBA molecules. The excitation energy required when the process of exciton resonance (I) and molecule resonance (II) takes place is 3.3 and 4.63 eV, respectively. However, the excitation energy provided by 633 nm laser is only 1.96 eV. Therefore, the two processes of CT resonance are ruled out due to the provided energy is far less than the amount of energy needed. Likewise, the photon induced CT resonance between the HOMO level of 4-MBA molecules and the CB of ZnO NPs is impossible to happen owing to the process needs 6.58 eV. According to the literature data, the surface state level of ZnO NPs is located at −3.5 eV (Xue et al., 2012b) approximately, the energy needed for the excited transition of electrons in the VB of ZnO NPs to surface state level (V) and the LUMO orbit of 4-MBA molecules (IV) are 1.7 and 1.35 eV, respectively. The laser excitation energy is enough for the two photon-induced CT resonance process. With respect to the process that the electrons in the surface state level of ZnO injected into LOMO level of molecules (VI), it is occurred by oneself due to the LUMO level of 4-MBA molecules was slightly below the surface state level of ZnO NPs.</p><!><p>(A) All the possible CT resonance process in ZnO-molecules system, (B) The relationship between diameter, band gap and the actual ratio of Ga and Zn for Ga-doped ZnO NPs.</p><!><p>When it comes to the effect of doped Ga on CT of ZnO NPs, there are two changes on the properties of ZnO NPs: particle size and band gap. Correspondingly, there are two main factors that can affect the enhance ability of ZnO NPs for SERS: size dependence effect and band-gap shrinkage effect. Figure 6B shows the relationship between diameter, band gap and the actual ratio of Ga and Zn for Ga-doped ZnO NPs. Combined with Figures 5B, 6B, we think the size dependence effect is predominant when the actual ratio of Ga and Zn is below 0.66% and the strongest SERS signal is obtained when the diameter of Ga-doped ZnO NPs is 27.4 nm. The variation of the intensity of SERS signal on the particle size of ZnO derives from the resonance effect between the level of ionized receptor-exciton complex on the surface and the frequency of incident light. Moreover, the result we obtained is consistent with the conclusion that the maximum SERS signal appeared when the crystal size is 27.7 nm for pure ZnO NPs (Sun et al., 2007). When the actual doping ratio increases over 0.66%, the band-gap shrinkage effect becomes the predominant factor to affect the SERS performance (degree of CT) of ZnO NPs. The maximum shrinkage of bang gap is 0.08 eV in all the doping ratios, so the CT resonance process of I, II, III (described in Figure 6A) are still forbidden. However, the photon induced CT resonance (IV, V, VI) can be influenced by the decrease of band gap, thus promoting the CT process. The gap between VB and the surface state level of ZnO is decreased, and the gap between VB of ZnO and the LUMO level of molecules is also reduced. Therefore, the excited transition of electrons in the VB of ZnO to surface state level (V) or the LUMO level of molecules (IV) are much easier, hence the SERS performance (degree of CT) of ZnO is improved and a stronger SERS signal can be obtained. Besides, due to the decreases of gap between surface state level of ZnO NPs and the LUMO of 4-MBA molecules, the efficiency of electrons in the surface state level of ZnO injected into LUMO level of molecules (VI) is raised. Above all, the improvement of SERS signal (degree of CT) is because the photon induced CT resonance (IV, V, VI) is enhanced due to band-gap shrinkage effect. Moreover, the reliability of the schematic diagram of the possible CT resonance process in our ZnO-molecules system can also be testified by the SERS spectra obtained at excitations of 532 and 785 nm, the data is shown in Figure S5.</p><!><p>In summary, we synthesized a series of Ga-doped ZnO NPs with various ratio of Ga and Zn by a simple method and obtained the 4-MBA@ZnO system by modifying ZnO used 4-MBA molecules. The XRD and TEM measurements are carried out, the particle size diminishes gradually with the increasing ratio of Ga and Zn. When Ga is introduced into ZnO, the degree of crystallinity is decreased according to the results from XRD and Raman. Moreover, a Ga doping induced band gap shrinkage occurs: the band gap of Ga-doped ZnO narrows with the ratio of Ga and Zn increasing, and the band gap is down to the minimum value (3.16 eV) when the actual ratio of Ga and Zn is 3.8%. The last but not the least, the SERS performance (degree of CT) of Ga-doped ZnO NPs is investigated, and the doped Ga is found to enhance the intensity of SERS signal because it can cause the band gap shrinkage and then affect the CT resonance process. The band gap shrinkage can promote the photon induced CT resonance process, the electrons in the VB of ZnO NPs were excited transition to surface state level of ZnO NPs and then injected into the LUMO level of molecules, or that the electrons in the VB of ZnO NPs were directly excited transition into the LUMO level of molecules. In conclusion, this work is not only beneficial for in-depth understanding of the effect of doping on CT resonance process between adsorbed molecules and semiconductor, but also provides a new insight on improving the SERS performance of semiconductor NMs.</p><!><p>All datasets generated for this study are included in the manuscript and/or the supplementary files.</p><!><p>PL performed the experiments and analyzed the date with help from BZ, XY, XW, XZ, and LZ. PL wrote and revised the manuscript with input from all authors. All authors read and approved the manuscript.</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><!><p>Funding. The research was supported by the National Natural Science Foundation (Grants 21773080 and 21711540292) of P. R. China; and the Development Program of the Science and Technology of Jilin Province (20190701003GH).</p><!><p>The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2019.00144/full#supplementary-material</p><!><p>Click here for additional data file.</p>

PubMed Open Access

Therapeutic targeting of MMP-12 for the treatment of chronic obstructive pulmonary disease

Chronic obstructive pulmonary disease (COPD) is a lung disorder characterized by progressive airflow obstruction associated with inflammation and emphysema, and it is currently one of the leading causes of death worldwide. Recent studies with genetically engineered mice reported that during pulmonary inflammation, basophil-derived interleukin-4 can act on lung-infiltrating monocytes causing aberrant expression of the matrix metalloproteinase-12 (MMP-12). MMP-12 activity in turn causes the destruction of alveolar walls leading to emphysema, making it potentially a valid target for pharmacological intervention. Using NMR- and structure-based optimizations, the current study reports on optimized novel, potent and selective MMP-12 inhibitors with single digit nanomolar affinity in vitro and in vivo efficacy. Using a murine model of elastase-induced emphysema we demonstrated that the most potent agents exhibited a significant decrease in emphysema-like pathology compared to vehicle-treated mice thus suggesting that the reported agents may be potentially translated in novel therapeutics for the treatment of COPD.

therapeutic_targeting_of_mmp-12_for_the_treatment_of_chronic_obstructive_pulmonary_disease

Introduction<!>NMR- and structure-based approaches to derive novel potent and selective MMP-12 inhibitors.<!>Efficacy of agents 25 and 26 in a murine model of elastase-induced emphysema<!>Discussion and conclusions<!>General Chemistry.<!>Peptide Synthesis<!>Protein expression and purification<!>Nuclear Magnetic Resonance<!>Enzymatic assays<!>Molecular modeling<!>Pharmacokinetics studies<!>Murine Model<!>In Vivo Porcine Pancreatic Elastase Exposure<!>Cell Counts and Cell Differentials<!>Mean Linear Intercept (MLI) Analysis

<p>Chronic obstructive pulmonary disease (COPD) is a lung disease characterized by a progressive and irreversible airflow limitation, and currently ranks among the top leading causes of death worldwide.1 COPD is often associated with an excessive inflammatory response of the lungs to air pollutants, or cigarette smoking. Given the persisting prevalence of cigarette smoking, and the increasing environmental factors and pollutants, the incidence of COPD is expected to continue grow. Moreover COPD has been linked to other sever co-morbidities including cardiovascular disease2 and lung cancer.3 Currently, only a few disease-modifying targeted therapeutics are available for this indication.4–8 Recent studies with a series of genetically engineered mice revealed, however, that the matrix metalloprotease MMP-12 can play a pivotal role in airway inflammation and remodeling.9 Preclinical studies in COPD/emphysema provide experimental support that approaches aimed at blocking MMP-12 could be translated into useful agents for therapeutic intervention. For example, pathological evidence indicates that MMP-12-deficient mice are protected against the development of emphysema induced by cigarette smoke and pollutants.10 However, for these findings to translate in possible therapeutics, novel, effective and selective drug-like MMP-12 agents must be developed. Currently the most potent and selective MMP-12 inhibitor is agent MMP408,11 while previously a dual MMP-9/MMP-12 inhibitor was also reported (e.g. clinical candidate AZD1236).10, 12 We recently reported on the use of the HTS by NMR approach13–15 to derive novel MMP-12 inhibitors.16 Using NMR- and structure-based guided optimizations, this present study reports on a novel series of agents with single-digit nanomolar inhibition against MMP-12. Furthermore, we found that mice exposed to porcine pancreatic elastase but treated with MMP-12 inhibitors (MMP-408, or lead agents compound 25, and 26) exhibited a significant decrease in emphysema-like pathology compared to vehicle-treated mice. Hence, the data demonstrated once again that the HTS by NMR can identify viable hit compounds that can be readily optimized into lead agents for continued drug development. The approach is of general applicability hence in principle can be deployed for the identification of such agents when targeting any metallo-enzymes. Preliminary pharmacological evaluations of compound 25 suggest that the agent could be a suitable candidate for additional efficacy and drug development studies leading eventually to human trials.</p><!><p>We previously reported on a novel general approach, termed the HTS by NMR13–15 that allowed the discovery of initial hit agents targeting MMP-12.16 Briefly, the method consisted first in deriving a focused positional scanning (POS) combinatorial library of peptide mimetics (of approximately 100,000 compounds) where each element of the library contained the metal-chelating moiety hydroxamic acid at the C-terminal (Figure 1A).</p><p>In the implementation of this approach against hMMP-12, we used a combination of 46 natural and non-natural fn aminoacids to generate 46 mixtures fnXX-CONHOH, 46 mixtures XfnX-CONHOH, and 46 mixtures XXfn-CONHOH, where X represents all the 46 aminoacids.16 Hence, each mixture contained approximately 46 × 46 ~ 2100 compounds (Figure 1A). Subsequently, each mixture was tested against the metalloproteinase hMMP-12 using sensitive protein–NMR screening methods.17 In particular, we rank-ordered the mixtures based on chemical shift changes induced by a given fn element using 1D 1H aliphatic and 2D [15N,1H] so-fast HMQC correlation spectra.13–16, 18–19 The synthesis of the library, and the individual agents, was easily attained by solid phase synthesis and an fmoc-hydroxylamine-2-chlorotrityl resin that after cleavage with 94% of trifluoroacetic acid (TFA) delivered the agents with a C-terminal hydroxamate. Representative 1D 1H aliphatic and 2D [15N,1H] so-fast HMQC correlation spectra for positive top ranking fragments in P1, P2, or P3, along with a negative representative mixture as control, are reported in Figure 1.16 Subsequently, top ranking elements from each position were combined, individual compounds were synthesized encompassing the most active fragments at each position, and tested using NMR and also biochemical assays to more accurately quantify potency and selectivity.16</p><p>These studies culminated with the identification of compound 1, that inhibited hMMP-12 with an IC50 value of 29.7 nM (Figure 2A,B).16 Compound 1 did not significantly inhibit MMP-1, MMP-9, MMP-13, and MMP-14 even at 1 μM, while it inhibited appreciably only MMP-3, which is the most closely related metalloproteinase to MMP-12, with ~19% inhibition at 55 nM.16 Given the geometry of the metal chelating moiety, and supported by NMR experimental chemical shift perturbations, a model of the structure of compound 1 in complex with MMP-12 (PDB-ID 5LAB) was generated (Figure 2C,D) that formed the basis for systematic structure-based optimizations of compound 1.</p><p>Given the modular nature of the agent, and to reduce the number of possible agents to be synthesized and tested, we opted to systematically fine-tune each substructure independently. Hence, we first probed modifications in the P1 position by introduction of few bioisosteres of the ethoxymethyl benzene present in compound 1 at that position (Table 1). Similarly, replacements of the P2 position D-homo-phenylalanine (Table 2), and the D-phenylalanine in position P3 (Table 3) of compound 1, respectively, were carried out. In each SAR study, rank ordering of the agents was accomplished by NMR chemical shift perturbations followed by enzymatic assays. As reported previously with agents of this series, NMR titrations of test agents with recombinant MMP-12 resulted in slow exchange in the NMR-time scale, in agreement with the nanomolar affinity of the optimized ligands of this series for the target.16 Hence, we used the methyl resonances of residues L212 and I220 that resonate in a spectral region of the 1D 1H NMR spectrum of the protein that is void from overlap with any other resonances from the protein or from ligand hydrogen nuclei (Figure 3). Upon binding of potent ligands, two new resonances appear, shifted by approximately 50 Hz from their position in the unbound state. At a given protein:ligand ratio, the spectrum of hMMP-12 would display the new resonances at an intensity that is proportional to the binding affinity of the test ligand. Hence, we defined a ΔI value as the average ratio between the peak intensities for the methyl resonances of L212 and I220 in the unbound versus the bound form at 1:1 protein/ligand ratio of 10 μM, as a rough measure to rank order the binding affinity of closely related test agents (Figure 3, Tables 1–3). Subsequently, agents that displayed a greater affinity for hMMP-12 compared the parent compound 1 were tested in an enzymatic assay for hMMP-12 inhibition (Tables 1–3).</p><p>These studies were aimed at the identification of optimized substituents for each of the 3 positons of the backbone of compound 1 (Figure 3). In particular, substituents that were more hydrophobic than the benzyl-ester in P1 resulted more active (compounds 2 and 5) while more rigid analogues were significantly less potent that compound 1 (Table 1). Commercially available amino-acids carrying isosters of the P2 position side chain resulted in agents listed in Table 2, that when tested by the NMR and enzymatic assays resulted less potent than compound 1, with the exception of compounds 9 and 10, that introduced a p-CF3 or p-methoxyl group in the P2 side chain, respectively. However, under the same experimental conditions, compound 10 appeared more active in the NMR-binding assay, hence was selected as the top ranking for the P2 position. Replacement of the D-Phe in position P3 was also probed by synthesizing and testing the agents listed in Table 3. Substitutions in the para position with a methyl group (compound 17), and in the meta (compound 18) or ortho (compound 21) position with a fluorine atom, increased potency of the agents relative to compound 1. However, unfortunately di- or tri-substituted D-Phe such as p-methyl,o-F-D-Phe, p-methyl,m-F-D-Phe, or p-methyl,o-F,m-F-D-Phe were not commercially available. Hence, the side chain of compound 17 was selected as top ranking for the P3 position.</p><p>Therefore, these studies ultimately led to the synthesis of compounds 25 and 26 that encompass the optimal commercially available P1, P2, and P3 substituents (Figure 3). The morpholino group was introduced at the N-terminal of these agents to increase their solubility. Enzyme activity inhibition assay using the SensoLyte® 520 hMMP-12 Fluorimetric Assay Kit (Anaspec) revealed that compounds 25 and 26 were competitive inhibitors for hMMP-12 with remarkable IC50 values of ~ 4 nM (Figure 4A, B). As controls, we also tested the pan-MMP inhibitor GM6001, and the reportedly potent and selective MMP-12 inhibitor compound MMP408.11 When tested side by side, and using the exact same assay protocol and kit (SensoLyte® 520 MMP-12 Fluorimetric Assay Kit) control agents pan-MMP inhibitor GM6001, and MMP-12 selective inhibitor MMP408 displayed IC50 values of ~2.5 nM, and ~19 nM, respectively (Figure 4D, E). Hence, while the IC50 value obtained for GM6001 is in close agreement with that reported in literature for this non-selective MMP inhibitor, in our assay agent MMP408 resulted significantly less potent than previously reported.11 While absolute IC50 values depend on the assay parameters, the relative potency between compounds 25, 26, and MMP408, tested under the same experimental conditions, reveals that our agents are > 5 times more potent than MMP408.</p><p>Similar to compound 1, compounds 25 and 26 did not display significant inhibition of MMP-1, MMP-9, MMP-13, and MMP-14, while appreciable inhibition of the closely related MMP-3 was observed with both agents and with GM6001 (Figure 5).</p><p>Anticipating their use in in vivo efficacy studies, we further conducted a preliminary pharmacokinetic study with compound 25 (Figure 6A). When administered at 30 mg/kg intraperitoneally, we observed peak plasma drug concentration reaching ~ 400 ng/ml corresponding to ~ 580 nM, hence > 20 times above the in vitro determined IC90 value. Intriguingly, the chemical structures of compound 25 and 26 resemble the FDA-approved proteasome inhibitor carfilzomib (Figure 6A),20–25 with the notable difference that the metal chelating hydroxamate in our agents is a reactive epoxide in carfilzomib, that is necessary to irreversibly inactivate the proteasome. Hence, we can speculate that our agents, presenting a similar chemical structure but lacking the potential chemical liability of the epoxide present in carfilzomib, could also be useful as potential therapeutic agents. In summary, these SAR studies on hit compound 1, originally selected out of 100,000 molecules within the combinatorial library and the HTS by NMR approach,16 resulted in viable drug-like lead agents 25 and 26.</p><!><p>Recent studies have identified a role for MMP-12 in promoting the emphysematous lung tissue destruction associated with chronic obstructive pulmonary disease.9–10 To test the efficacy of compound 25 and compound 26 in an in vivo model system, we chose to investigate the impacts of pharmacological MMP-12 inhibition in a well-established murine model of elastase-induced emphysema. In this model, mice instilled with elastase exhibit early markers of inflammation and injury within hours to days following elastase treatment. While initial inflammatory cell influx and heightened airway cytokine release wane in the first week following elastase instillation, lung injury following elastase treatment evolves over several weeks, with dramatic tissue destruction evident within 3–4 weeks following elastase challenge. Using this model, we assessed the impacts of MMP-12 inhibition on the development of tissue destruction at 21 days following elastase treatment. As expected based on our chosen (21-day) time point, there was not a significant difference in total BALF cellularity/influx amongst all groups (Figure 7A), although differential cell analysis did identify a significant impact of elastase treatment on lymphocytes influx, as shown in Figure 7.</p><p>To assess for tissue destruction in saline- or elastase-treated animals, we performed mean linear intercept (MLI) analysis. Based on this analysis, we identified significant lung tissue destruction in murine lungs exposed to porcine pancreatic elastase (PPE) and treated with the vehicle control, compared to murine lungs exposed to the saline control and treated with the vehicle control (Figure 8). Notably, however, mice exposed to PPE but treated with MMP-12 inhibitors (MMP-408, compound 25, or compound 26) exhibited a significant decrease in emphysema-like pathology compared to PPE + vehicle-treated mice, with MLI measurements for mice treated with PPE + MMP-12 inhibitors exhibiting no significant differences compared to saline (no PPE)-treated mice (Figure 8).</p><!><p>The search for possible therapeutic applications of potent and selective metallo-enzyme inhibitors has remained very active in the past two decades. Small molecules with potentially favorable pharmacological properties have been developed,26 including the design of allosteric inhibitors27 or the deployment of novel metal chelating groups.28–32 MMP-12 is involved in the inflammatory response in chronic obstructive pulmonary disease (COPD) in mice,10 while in humans with asthma and COPD, MMP-12 aberrant activation is associated with disease severity.9 In a recent phase II trial, the dual MMP-12/MMP-9 inhibitor AZD1236 was tested in a randomized short trial (6 weeks) on moderate to severe COPD, and it showed an acceptable safety profile, although the therapeutic efficacy could not be demonstrated given the limited duration of the study.12 In addition, FP-025 (Forsee Pharmaceutical) reported on an ongoing phase II trial of their MMP-12 inhibitor, to assess its efficacy on allergen-induced airway inflammation in mild eosinophilic house dust mite allergic asthma (https://clinicaltrials.gov/ct2/show/NCT03858686). Another more recently reportedly potent and selective MMP-12 inhibitor is agent MMP408.11 However, when tested side by side with compounds 25, or 26, MMP408 exhibited significantly less potency than our agents that in turn are as potent as the pan-MMP inhibitor GM6001, yet displayed a greater selectivity for MMP-12. Following a preliminary pharmacokinetic study, compound 25 was administered in mice via intraperitoneal injections and displayed a favorable drug plasma levels and half-life. Interestingly, the chemical structures of optimized agents 25 and 26 structurally resemble the FDA-approved proteasome inhibitor carfilzomib (Figure 5), suggesting potentially a drug-like nature of the agents. Using a well-established animal model of elastase-induced emphysema, we found that mice exposed to porcine pancreatic elastase were protected by treatment with MMP-12 inhibitors (MMP-408, compound 25, or compound 26) from developing emphysema-like pathology compared to untreated mice, indicating that our agents have significant therapeutic potential. This study once again emphasizes the effectiveness of the HTS by NMR approach in deriving novel, potent and selective agents, particularly when an anchoring moiety can be incorporated in the positional scanning combinatorial library, as we have recently reported.13, 15–17, 33 The present SAR studies also further suggest that possible further optimizations of 25 or 26 could include o- and/or m-fluorination of the D-Phe in position P3.</p><p>In summary, we are confident that the identified agents represent innovative, potent and effective MMP-12 inhibitors, hence worthy of continued drug development for their potential translation in novel therapeutics for the treatment of MMP-12-mediated airway inflammatory conditions and potential other co-morbidities associated with COPD.</p><!><p>All common solvent and reagents were obtained by commercial sources. NMR spectra were recorded on Bruker Avance III 700 MHz and these were used both for quality control and to verify the concentration of the stock solutions used for dose response measurements and in vivo studies. An Agilent LC-TOF instrument was used to obtain high-resolution mass spectral data. Purification of all agents was obtained using RP-HPLC on a JASCO preparative system equipped with a PDA detector. The instrument is also equipped with a fraction collector controlled by a ChromNAV system (JASCO). For all agents, a Luna C18 10μ 10 × 250mm (Phenomenex) column was used to purify agents to > 95% purity. For intermediate reagents that were not commercially available, RP-chromatography purification was performed using a CombiFlash (Teledyne ISCO). GM6001 was obtained from Enzo Life science. MMP408 was obtained from EMD Millipore Corp.</p><!><p>Peptides were synthesized by using standard solid-phase synthesis protocols, using an fmoc-hydroxylamine-2-chlorotrityl resin that introduces the hydroxamic acid at the C-terminus of the peptides, after cleavage. For each coupling reaction, 3 eq. of Fmoc-AA, 3 eq. of HATU, 3 eq. of OximaPure, and 5 eq. of DIPEA in 1 ml of DMF were used. The coupling reaction was allowed to proceed for 1 h. Fmoc deprotection was performed by treating the resin-bound peptide with 20% piperidine in DMF twice. Peptides were cleaved from Rink amide resin with a cleavage cocktail containing TFA/TIS/water (94:3:3) for 3 h. The cleaving solution was filtered from the resin, evaporated under reduced pressure and the peptides precipitated in Et2O, centrifuged and dried in high vacuum. The crude peptide was purified by preparative RP-HPLC using a Luna C18 column (Phenomenex) and water/acetonitrile gradient (30% to 70%) containing 0.1% TFA. The final compounds were characterized by HRMS.</p><p>For the synthesis of fmoc-amino acids, 1 eq. of the unprotected amino acid and Na2CO3 were dissolved in THF/H2O (1:1) and cooled to 0°C. 1.1 eq. of Fmoc Chloride was dissolved in THF and added dropwise to the mixture. The reaction was stirred for 2 h at 0°C. The organic solvent was evaporated under reduced pressure and the pH lowered to 0 using concentrated HCl. The aqueous phase was extracted 3 times with AcOEt and the collected organic phase were dried with Na2SO4, filtered and evaporated. The resulting crude was purified using a CombiFlash Rf (Teledyne ISCO) using cyclohexane/Ethyl Acetate (10% to 100%). Fmoc protection was required for 4-methoxy-D-homophenylalanine for the synthesis of both compounds 25, and 26; and for (S)-2-Amino-5-(4-methoxyphenyl)pentanoic acid for the synthesis of compound 26. Detailed experimental procedures and analytical data for compounds 25 and 26 are provided as supplementary information.</p><!><p>The catalytic domain (Gly106-Gly263) of human macrophage metalloelastase (hMMP-12), was expressed by cloning the gene into a pET21 vector (Novagen) using NdeI and BamHI as restriction enzymes and then transfected into E. coli strain BL21 Codon Plus cells. Expression of uniformly 15N-labeled hMMP-12 in M9 minimal media containing 15 mM (15NH4)2SO4was induced with 0.5 mM IPTG at 37 °C for 4 h. Then protein forms inclusion bodies that were isolated and solubilized in a solution of 8 M urea (in 20 mM Tris–HCl, pH 8). The protein was purified in two steps including a first size-exclusion chromatography (Pharmacia HiLoad Superdex 75 16/60) in 6 M urea (in 50 mM sodium acetate). A second cation exchange purification step was carried using a Mono-S column (Pharmacia) and a sodium chloride linear gradient (from 0 to 500 mM). Protein refolding was accomplished by a multiple dialyses into decreasing concentrations of urea (from 4 M up to 2 M; 50 mM Tris–HCl, pH 7.2, 10 mM CaCl2, 0.1 mM ZnCl2, 300 mM NaCl). Further dialysis steps were performed exchanging the protein into the final buffer containing 20 mM Tris–HCl (pH 7.2), 10 mM CaCl2, 0.1 mM ZnCl2, 300 mM NaCl. The final buffer also contained 200 mM of acetohydroxamic acid (AHA), a weak MMP-12 inhibitor, to prevent the self-proteolysis. Recombinant MMP-1 (Cat. # AS-55575-1), MMP-3 (Cat. # 72006), MMP-9 (Cat. # AS-55576-1), MMP-13 (Cat. # AS-72257), and MMP-14 (Cat. # AS-72068) were obtained from AnaSpec.</p><!><p>Solution Nuclear Magnetic Resonance (NMR) experiments were conducted on a 700 MHz Bruker Avance III spectrometer equipped with a TCI cryoprobe. Each samples contained 10 μM of 15N-labeled hMMP-12 catalytic domain in absence or in presence of 10 μM of each compounds, in 20 mM Tris-HCl pH 7.2, 300 mM NaCl, 200 mM acetohydroxamic acid (AHA), 10 mM CaCl2, and 0.1 mM ZnCl2 with 1% final d6-DMSO. For each sample 2D [15N, 1H] so-fast HMQC, and 1D 1H-aliph experiments were acquired. For ranking purpose were measured the ratio between the intensity of the bound peaks (IBND), and the intensity of the apo peaks (IAPO), for the aliphatic residues L212 and I220, that we called ΔI. For a value of ΔI > 1 and producing an effect greater than what observed with compound 1 we gave a rank of ++++, for a value of ΔI > 1 we gave a rank of +++, for a value of ΔI ~ 1 we gave a rank of ++, and for a value of ΔI < 1 we gave a rank of +.</p><!><p>The enzymatic assays to profile the inhibitory effects of all the compounds on hMMP-12, and against a panel of 5 closely related MMPs (MMP-1, MMP-3, MMP-9, MMP-13, and MMP-14) were performed using the SensoLyte® 520 Fluorimetric Assay Kit (AnaSpec) for each MMPs. Assay kit for MMP-12 Cat. # AS-71157, MMP-1 Cat. # AS-71150, MMP-3 Cat. # AS-71152, MMP-9 Cat. # AS-71155, MMP-13 Cat. # AS-71156, and MMP-14 Cat. # AS-72025. The assay was performed according to their protocols. Briefly, 10 ng of each MMPs were incubated on a black flat-bottom 96-well plate (Cat. # 9502867) at RT for 15 minutes in absence or in presence of different concentration of each compound in a total volume of 50 μL. After 15 minutes, 50 μL of each substrate solution were added. The reagents were mixed together by shaking the plate gently for 30 sec. Immediately after the fluorescence was measured using the VICTOR X5 microplate reader (PerkinElmer) every 3 min for 60 min. For SAR purpose, IC50 values for agents in Table 1, 2, and 3, were extrapolated from % Inhibition of MMP-12 in presence of 25 nM of each agent. The IC50 values were calculated from dose-response curves using GraphPad Prism 7.</p><!><p>Compound 1, and 25 were docked using Gold (Cambridge Crystallographic Data Center; www.ccdc.cam.ac.uk) and Protein Data Bank entry 5LAB. The docking preparation for both protein and ligands were performed using MOE 2019.0101 (Chemical Computing Group). The surface representations were prepared using MOE 2019.0101 (Chemical Computing Group).</p><!><p>For these studies, compound 25 was administered i.p. (30 mg/kg) to 5 Balb-C mice. Retrorbital bleeding was used to collect blood samples at times 30 min, 1h, 2h, 4h, 8h, and 24h and the samples analyzed for compound 25 plasma concentration via extraction followed by LC/MS and compared to a standard calibration curve prepared with purified agent. Compound 25 was soluble in 10% DMSO, 40% PEG400, and 10% (2-hydroxypropyl)-Beta-cyclodextrin in PBS. Mice weight varied between 22 g and 26 g, each receiving approximately 200 μL (weight adjusted to administer 30 mg/Kg of compound 25) of formulated agent. The experiments were conducted at the University of California San Diego in vivo pharmacology core facility, according to a UCSD Institutional Animal Care and Use Committee (IACUC) approved protocol.</p><!><p>C57BL/6 mice were purchased from Jackson Laboratories (Bar Harbor, ME) and maintained in a pathogen-free vivarium at room temperature with 12-hour light/dark cycles. All mice were 6–8 weeks and both male and female mice were used for the experiment. All animal-use and euthanasia protocols were approved by the UC Riverside Institutional Animal Care and Use Committee (IACUC).</p><!><p>C57BL/6 mice were given a single intranasal instillation with 50 μl of saline or 0.9 U porcine pancreas elastase (Sigma-Aldrich, St. Louis, MO) diluted in phosphate buffered saline (1X PBS). Two days following the instillation, mice were treated with 200 μl of vehicle control, MMP-408, compound 25, or compound 26 via intraperitoneal injection. Mice were continually given treatment once a day for 7 consecutive days. Twenty-one days following the initial intranasal instillation, mice were sacrificed. Bronchoalveolar lavage fluid (BALF) was collected by making an incision in the trachea and inserting a cannula with syringe (BD Biosciences, CA), then washing with 1 mL of cold PBS three times. BALF was centrifuged at 1200 RPM for 5 minutes, then cell pellets were combined and resuspended in 200 μl of PBS to be used in cell counts and cell differential analysis. The lungs were isolated, slowly filled with 1 mL of 10% formalin then hung in formalin overnight for fixation. The following day, lungs were moved into 70% ethanol and stored at 4°C until paraffin-embedding.</p><!><p>Cell pellets were resuspended in 200 μl of PBS, then 10 μl of the solution was put onto a hemocytometer for total cell counts. For cell differentials, 150 μl of the suspended cell solution was put on a microscope slide using a cytospin. Using a Revolve light microscope (La Jolla, CA), a total of 300 cells were counted per slide, and were differentiated by cell type. The number of each cell type identified was then divided by 300 to get the percentage of each cell type amongst the total cell population, then that percentage was multiplied by the total cell count to represent the total number of each cell type in the BALF.</p><!><p>Mouse lungs were fixed in formalin and paraffin-embedded, then sectioned and stained in hematoxylin and eosin by the University of California Irvine Department of Pathology Experimental Tissue Resource Core Facility. The lungs were imaged on a Revolve light microscope (La Jolla, CA) at 10x magnification. MLI measurements were calculated using an established indirect method on ImageJ software. Briefly, alveolar walls/intersections were counted along a line of known length, then length was divided by the number of intersections to calculate MLI. This process was repeated for a total of five MLI measurements per image, with there being four images per mouse lung. The average MLI measurement for each lung was calculated and statistically analyzed using GraphPad Prism software (San Diego, CA).</p>

PubMed Author Manuscript

C2-Selective Branched Alkylation of Benzimidazoles by Rhodium(I)-catalyzed C\xe2\x80\x93H Activation

Herein, we report a Rh(I)/bisphosphine/K3PO4 catalytic system allowing for the first time the selective branched C\xe2\x80\x93H alkylation of benzimidazoles with Michael acceptors. Branched alkylation with N,N-dimethyl acrylamide was successfully applied to the alkylation of a broad range of benzimidazoles incorporating a variety of N-substituents and with both electron-rich and electron-poor functionality displayed at different sites of the arene. Moreover, the introduction of a quaternary carbon was achieved by alkylation with ethyl methacrylate. The method was also shown to be applicable to the C2-selective branched alkylation of azabenzimidazoles.

c2-selective_branched_alkylation_of_benzimidazoles_by_rhodium(i)-catalyzed_c\xe2\x80\x93h_activation

<!>General Experimental<!>1,2-Bis(bis(3,5-bis(trifluoromethyl)phenyl)phosphino)ethane (dArFpe)<!>1-Phenyl-1H-benzo[d]imidazole (1b)<!>1-((Benzyloxy)methyl)-1H-benzo[d]imidazole (1c)<!>Benzimidazole-5-carboxylic acid ethyl ester<!>Ethyl 1-((benzyloxy)methyl)-1H-benzo[d]imidazole-5-carboxylate (1d)<!>1-Methyl-5-methylsulfonyl-benzimidazole (1f)<!>1-Methyl-6-chloro-1H-benzo[d]imidazole (1g)<!>1-Methyl-1H-benzo[d]imidazole-5-carbonitrile (1h)<!>1-Methyl-7-(trifluoromethyl)-1H-benzo[d]imidazole (1i)<!>1-Methyl-6-(trifluoromethyl)-1H-benzo[d]imidazole (1j)<!>1-Methyl-5-(trifluoromethyl)-1H-benzo[d]imidazole (1k)<!>1-Methyl-4-(trifluoromethyl)-1H-benzo[d]imidazole (1l)<!>1-((Benzyloxy)methyl)-6-methoxy-1H-benzo[d]imidazole (1m) and 1-((Benzyloxy)methyl)-5-methoxy-1H-benzo[d]imidazole (1n)<!>1-((Benzyloxy)methyl)-5,6-dimethyl-1H-benzo[d]imidazole (1o)<!>1-Methyl-1H-imidazo[4,5-b]pyridine (1p)<!>3-Methyl-3H-imidazo[4,5-b]pyridine (1q)<!>6-(Benzyloxy)-3-methyl-3H-imidazo[4,5-b]pyridine (1r)<!>Ethyl 3-methyl-3H-imidazo[4,5-b]pyridine-6-carboxylate (1s)<!>General Procedure for Branched Alkylation<!>N,N-Dimethyl-2-(1-methyl-1H-benzo[d]imidazol-2-yl)propanamide (3a)<!>N,N-dimethyl-2-(1-phenyl-1H-benzo[d]imidazol-2-yl)propanamide (3b)<!>2-(1-((Benzyloxy)methyl)-1H-benzo[d]imidazol-2-yl)-N,N-dimethylpropanamide (3c)<!>Ethyl 1-((benzyloxy)methyl)-2-(1-(dimethylamino)-1-oxopropan-2-yl)-1H-benzo[d]imidazole-5-carboxylate (3d)<!>2-(1-(Dimethylamino)-1-oxopropan-2-yl)-1-methyl-1H-benzo[d]imidazol-5-yl propionate (3e)<!>N,N-Dimethyl-2-(1-methyl-5-methylsulfonyl-benzimidazol-2-yl)propanamide (3f)<!>2-(6-Chloro-1-methyl-1H-benzo[d]imidazol-2-yl)-N,N-dimethylpropanamide (3g)<!>2-(5-Cyano-1-methyl-1H-benzo[d]imidazol-2-yl)-N,N-dimethylpropanamide (3h)<!>N,N-Dimethyl-2-(1-methyl-7-(trifluoromethyl)-1H-benzo[d]imidazol-2-yl)propanamide (3i)<!>N,N-Dimethyl-2-(1-methyl-6-(trifluoromethyl)-1H-benzo[d]imidazol-2-yl)propanamide (3j)<!>N,N-Dimethyl-2-(1-methyl-5-(trifluoromethyl)-1H-benzo[d]imidazol-2-yl)propanamide (3k)<!>N,N-Dimethyl-2-(1-methyl-4-(trifluoromethyl)-1H-benzo[d]imidazol-2-yl)propanamide (3l)<!>2-(1-((Benzyloxy)methyl)-6-methoxy-1H-benzo[d]imidazol-2-yl)-N,N-dimethylpropanamide (3m)<!>2-(1-((Benzyloxy)methyl)-5-methoxy-1H-benzo[d]imidazol-2-yl)-N,N-dimethylpropanamide (3n)<!>2-(1-((Benzyloxy)methyl)-5,6-dimethyl-1H-benzo[d]imidazol-2-yl)-N,N-dimethylpropanamide (3o)<!>N,N-Dimethyl-2-(3-methyl-3H-imidazo[4,5-b]pyridin-2-yl)propanamide (3p)<!>N,N-Dimethyl-2-(3-methyl-3H-imidazo[4,5-b]pyridin-2-yl)propanamide (3q)<!>2-(6-(Benzyloxy)-3-methyl-3H-imidazo[4,5-b]pyridin-2-yl)-N,N-dimethylpropanamide (3r)<!>Ethyl 2-(1-(dimethylamino)-1-oxopropan-2-yl)-3-methyl-3H-imidazo[4,5-b]pyridine-6-carboxylate (3s)<!>2-(1-methyl-1H-benzo[d]imidazol-2-yl)propanal O-methyl oxime (4a)<!>2-(1-methyl-1H-imidazo[4,5-b]pyridin-2-yl)propanal O-methyl oxime (4p)<!>Ethyl 2-(1-((benzyloxy)methyl)-1H-benzo[d]imidazol-2-yl)-2-methylpropanoate (6c) and Ethyl 3-(1-((benzyloxy)methyl)-1H-benzo[d]imidazol-2-yl)-2-methylpropanoate (6c\xe2\x80\xb2)<!>Ethyl 2-methyl-2- (1-methyl-6-trifluoromethyl-1H-benzo[d]imidazol-2-yl)propanoate (6j)<!>Ethyl 2-methyl-2-(3-methyl-3H-imidazo[4,5-b]pyridin-2-yl)propanoate (6q)

<p>The benzimidazole framework is among the most extensively used scaffolds in the pharmaceutical industry,1 and the synthesis and functionalization of benzimidazoles have been the subject of a considerable number of studies.2 Among them, the regioselective transition-metal catalyzed C–H functionalization of benzimidazoles has elicited significant interest because it allows rapid access to elaborated substrates from simple precursors.3 In particular, the direct C–H alkylation of C2-unsubstituted benzimidazoles with alkenes proceeds with complete atom economy.4 Both Rh and Ni catalyst systems have been reported to promote intermolecular benzimidazole C–H alkylations with very high linear selectivity, and depending on the reaction conditions and catalyst used, a vinyl group can be substituted with alkyl,5 aryl6 or electron withdrawing7 groups (Scheme 1, eq 1). In contrast, branched-selective intermolecular alkylation of benzimidazoles is limited to a single report, which employs styrenes with a Ni catalyst system (Scheme 1, eq 2).8,9 We have recently demonstrated that a Rh(I)-precatalyst combined with the electron-poor ligand dArFpe and K3PO4 enables the completely branched selective alkylation of pyridines in high yields.10 Herein, we report that this catalyst system can be extended to the branched selective alkylation of a wide range of benzimidazole and azabenzimidazole derivatives with N,N-dimethylacrylamide (eq 3) and to the first examples of branched-selective C–H alkylation to introduce a quaternary carbon with ethyl methacrylate (eq 4).</p><p>After reevaluation of a broad range of reaction parameters, it was found that the optimal conditions previously reported for the branched alkylation of pyridines were also the most effective for the branched alkylation of N-methylbenzimidazole 1a with N,N-dimethylacrylamide 2. Under these conditions, product 3a was obtained in 71% yield, with exclusive regioselectivity in favor of the C2-position and with complete branched selectivity (Table 1, entry 1). K3PO4 was the base of choice as other inorganic bases either led to complete shut-down of the reaction or drastically reduced yields (Table 1, entries 2–4). The electron-deficient bisphosphine dArFpe ((3,5-CF3Ph)pe) was also a critical element of the catalytic system, and the use of its more electron-rich analog, dppe, only led to a 13% yield (Table 1, entry 5). Reducing the temperature to 100 °C seemed to promote competitive polymerization of the alkene partner (Table 1, entry 6), as did the replacement of acrylamide 2 by tert-butyl acrylate and ethyl acrylate (Table 1, entry 7–8). Mesitylene was evaluated in place of toluene as solvent because its boiling point of 165 °C is well above the reaction temperature of 120 °C, and gave a comparable reaction yield (Table 1, entry 9). However, toluene was employed for evaluating reaction scope because it can be readily purchased in anhydrous form. Other common solvents such as THF could be used in the reaction, but the yield was significantly lower (Table 1, entry 10).</p><p>The optimized conditions proved to be very general, and broad scope was observed for the benzimidazole coupling partner (Scheme 2). Substitution at the N1-position with a methyl, phenyl or benzyloxymethyl (BOM) protecting group led to excellent yields of the corresponding benzimidazoles 3a–c, and scale-up of the alkylation of benzimidazole 1a from 0.4 to 7.6 mmol did not lead to a significant change in yield. Benzimidazoles substituted at the C5, C6, or C7 positions with a wide range of electron-withdrawing groups such as esters (3d–e), sulfonyl (3f), chloro (3g), nitrile (3h) or trifluoromethyl (3i–l) provided alkylation products in good to excellent yields. Only benzimidazole 3l, bearing a trifluoromethyl substituent at the C4-position, was obtained in a low yield of 12%. Electron-donating group such as methoxy or methyl were also very well tolerated when used in conjunction with a BOM-protecting group (3m–o). Because we previously reported that azines could be efficiently alkylated using the present method, it was a surprise for us to see that azine-fused imidazoles 1p and 1q could be selectively alkylated at the C2-position without any undesired alkylation on the azine ring. Electron-withdrawing and electron-donating substituents were well tolerated in these cases as well (3r–s). In very preliminary studies, an N-methyl and an N-BOM substituted imidazole were evaluated under the standard reaction conditions but starting material was primarily recovered (3t and 3u). It is possible that alkylation might occur under different reaction conditions or with imidazoles having alternative substitution patterns.</p><p>Although broad in scope with respect to the benzimidazole substituents, this method is primarily restricted to N,N-disubstituted acrylamides as the Michael acceptor. We therefore demonstrated that the amide products can be easily transformed to the corresponding aldehydes by treatment with Schwartz reagent (Scheme 3).11 Owing to the instability of the aldehyde products to long-term storage, they were isolated and characterized as their O-methyl oximes derivatives 4 after treatment with NH2-OMe•HCl. Given that aldehydes are among the most versatile functional handles in organic chemistry, this method can potentially be used to access a very broad range of branched benzimidazole derivatives.</p><p>The introduction of a quaternary carbon center by Rh(I)-catalyzed C–H bond additions to methacrylate derivatives was also investigated (Scheme 4). While N,N-dimethyl methacrylamide did not provided any desired product, ethyl methacrylate (5) proved to be a suitable alkene partner thus allowing the direct incorporation of a quaternary center (Scheme 4). However, in contrast to what was observed in the scope exploration using N,N-dimethylacrylamide, we found that the branched to linear alkylation selectivity of this reaction is dependent upon the substitution pattern of the nitrogen heterocycle 1. Indeed, while the use of N-BOM-benzimidazole (1c) led to both branched and linear products in 32% (6c) and 34% (6c') yields, respectively, electron-poor heterocycles provided the branched products such as 6j and 6q with high selectively and in excellent yields. We also found that the reaction was very sensitive to any adventitious water, which resulted in low yields of product. The greater sensitivity of the alkylation reaction to water when ethyl methacrylate (5) is used instead of N,N-dimethylacrylamide (2) potentially arises from the higher propensity of ethyl methacrylate to oligomerize under basic conditions.</p><p>In summary, we have developed a Rh(I)-catalyzed C2-selective alkylation of benzimidazole derivatives with N,N-dimethylacrylamide with complete branched selectivity. This transformation complements the previously reported branched alkylation of benzimidazoles by styrenes and the linear alkylation of benzimidazoles by Michael acceptors. Alkylated benzimidazoles featuring a broad range of different electronic properties can be accessed in good to excellent yields, and the amide products can be easily converted to aldehydes as versatile synthetic intermediates, thereby making the reported method useful for drug discovery applications. Moreover, the branched alkylation of benzimidazole derivatives with ethyl methacrylate was achieved for the preparation of products incorporating a quaternary carbon.</p><!><p>Unless otherwise indicated, all reactions were performed under a nitrogen gas atmosphere in oven-dried glassware cooled down under nitrogen gas. Toluene, diethyl ether, dichloromethane and THF were degassed by argon sparging and purified by elution through a column of activated alumina under an argon atmosphere prior to use. Mesitylene was dried and distilled under nitrogen from CaH2 and then was degassed by sparging with N2. 1,2-Bis(bis(3,5-bis(trifluoromethyl)phenyl)phosphino) ethane (dArFpe) was synthesized according to a modified literature procedure. All other catalyst and ligands were purchased from Strem Chemical and used without further purification. Potassium phosphate was dried at 160 °C under high vacuum prior to use. N-Methylbenzimidazole 1a was recrystallized from Et2O/pentane prior to use. All liquid azoles and N,N-dimethylacrylamide 2 were dried by elution through a small column of activated basic alumina and degassed by nitrogen sparging prior to use. Ethyl methacrylate 4 was dried over CaCl2, distilled under reduced pressure under nitrogen and then degassed by nitrogen sparging prior to use. All other reagents were purchased from commercial sources and used without further purification. All microwave-heated reactions were performed in a Biotage Initiator+ microwave reactor, which employs an external IR sensor. Flash column chromatography was performed over Grade 60 silica gel (230–400mesh). Preparative TLC was performed over silica gel-coated glass plates (20x20 cm, 1000 or 2000 μm). Reverse-phase column chromatography was performed on pre-packed cartridges of C18 silica gel using an automated purification system. NMR chemical shifts (δ) are reported in ppm relative to CHCl3 (δ = 7.26 ppm), acetone (δ = 2.05 ppm), H2O (δ = 4.79 ppm), or (CH3)2SO (δ = 2.50 ppm) for 1H-NMR, CDCl3 (δ = 77.16 ppm) or (CD3)2CO (δ = 29.8 ppm), for 13C-NMR, 85% aqueous H3PO4 for 31P-NMR (external reference, δ = 0 ppm) and trifluoroacetic acid for 19F-NMR (external reference, δ = -76.55 ppm). All 13C NMR are proton decoupled. All 19F and 31P NMR are carbon decoupled. The multiplicity and shape of NMR signals are designated by the following abbreviations: s singlet, d doublet, t triplet, q quartet, m multiplet, app apparent. Coupling constants (J values) are reported in Hertz (Hz). High-resolution mass spectra (HRMS) were recorded on a quadrupole time-of-flight mass spectrometer after electrospray ionization. Melting points (Mp) are reported uncorrected.</p><!><p>A freshly titrated solution of iPrMgBr (1.10 M in THF, 26.3 mL, 28.9 mmol, 6.7 equiv)12 was charged in an oven-dried round-bottom flask previously flushed with nitrogen. Additional dry diethyl ether (30 mL) was added, and the resulting solution was cooled down to −10 °C. Neat 1-bromo-3,5-bis(trifluoromethyl)benzene (5.20 mL, 30.2 mmol, 7.0 equiv) was added dropwise over 20 min, the reaction mixture was stirred at −10 °C for 1 h, then allowed to warm-up to room temperature and stirred for 3 h. The resulting clear brown solution was cooled down to 0 °C, and neat 1,2-bis(dichlorophosphino)ethane (0.65 mL, 4.3 mmol, 1.0 equiv) was added dropwise over 5 min. The resulting suspension was allowed to warm-up to room temperature and stirred for 16 h. The resulting suspension was cooled to 0 °C and quenched with an aqueous saturated NH4Cl solution. The organic solvents were removed in vacuo, and water and ethyl acetate were added to the resulting slurry until two clear phases were obtained. The phases were separated, and the aqueous phase was extracted twice with EtOAc. The combined organic phases were dried over Na2SO4 and concentrated in vacuo to give an orange solid. This solid was then washed with small amounts of peroxide-free and BHT-free diethyl ether until a white powder was obtained, which was then dried under high-vacuum to give the desired product (3.96 g, 4.20 mmol, 97% yield). 1H and 19F NMR matched the reported data.13 1H NMR (600 MHz, Acetone-d6) δ 8.11 (s, 8H), 8.05 (s, 4H), 2.73 (t, J = 5.3 Hz, 4H); 19F NMR (150 MHz, Acetone-d6) δ -63.5; 31P NMR (200 MHz, Acetone-d6) δ -9.4.</p><!><p>1H-Benzimidazole (236 mg, 2.00 mmol, 1.0 equiv) and Cs2CO3 (3.26 g, 10.0 mmol, 5.0 equiv) were weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by dry DMA (20 mL) and fluorobenzene (0.88 mL, 9.3 mmol, 4.7 equiv). The vial was sealed with a teflon-lined cap and heated at 190 °C in a Biotage Initiator+ microwave reactor for 16 h. The reaction mixture was poured on water and Et2O, the phases were separated and the aqueous phase was extracted three times with Et2O. The combined organic phases were dried over Na2SO4 and concentrated in vacuo to give the crude product. Purification by flash chromatography over silica gel (60% EtOAc/petroleum ether) gave the desired product 1b as a colorless oil (370 mg, 1.90 mmol, 95% yield). 1H NMR spectra matched the reported data.14 1H NMR (600 MHz, CDCl3) δ 8.13 (s, 1H), 7.93 – 7.84 (m, 1H), 7.62 – 7.50 (m, 5H), 7.50 – 7.43 (m, 1H), 7.39 – 7.29 (m, 2H).</p><!><p>Under N2, at 0 °C, to a solution of 1H-benzimidazole (1.00 g, 8.46 mmol, 1.0 equiv) in dry DMF (20 mL) was added NaH (60% w/w in oil, 340 mg, 8.87 mmol, 1.05 equiv) in one portion. The reaction mixture was stirred at 0 °C for 10 min, then neat BOM-Cl (90% purity, 1.38 mL, 8.90 mmol, 1.05 equiv) was added dropwise. The resulting solution was stirred at rt for 16 h. The reaction mixture was then poured onto a 10% w/w LiCl aqueous solution, EtOAc was added, the phases were separated, and the organic phase was washed three more times with a 10% w/w LiCl solution. The organic phase was dried over Na2SO4 and concentrated in vacuo to give a yellow oil. Purification by flash chromatography over silica gel (60 to 100% EtOAc/petroleum ether) gave the desired product 1c as a white amorphous solid (1.27 g, 5.33 mmol, 63% yield). 1H NMR spectra matched the reported data.15 1H NMR (600 MHz, CDCl3) δ 7.95 (s, 1H), 7.86 – 7.82 (m 1H), 7.56 – 7.53 (m, 1H), 7.38 – 7.31 (m, 5H), 7.30 – 7.26 (m, 2H), 5.60 (s, 2H), 4.47 (s, 2H).</p><!><p>Benzimidazole-5-carboxylic acid (1.00 g, 6.17 mmol, 1.0 equiv) was dissolved in EtOH (15.5 mL), treated with concentrated H2SO4 (0.36 mL, 6.9 mmol, 1.1 equiv), and heated to reflux for 24 h. After cooling to room temperature, the mixture was poured onto ice, and 5 N aqueous NaOH was added until pH~9 was reached. The mixture was extracted with EtOAc (3x30 mL). The combined organic layers were dried over Na2SO4 and concentrated in vacuo. Purification by flash chromatography over silica gel (100% EtOAc) gave the desired product benzimidazole-5-carboxylic acid ethyl ester as a cream-colored amorphous solid (751 mg, 3.95 mmol, 64% yield). 1H NMR matched the reported data.16 1H NMR (500 MHz, DMSO-d6) δ 12.79 (s, 1H), 8.40 (s, 1H), 8.23 (s, 1H), 7.84 (d, J = 7.9 Hz, 1H), 7.67 (s, 1H), 4.32 (q, J = 7.1 Hz, 2H), 1.34 (t, J = 7.1 Hz, 3H).</p><!><p>Benzimidazole-5-carboxylic acid ethyl ester (0.50 g, 2.6 mmol, 1 equiv) was dissolved in DMF (2.5 mL), and the reaction solution was cooled to 0 °C. Sodium hydride (60% w/w in oil, 91 mg, 3.9 mmol, 1.5 equiv) was added, and the reaction mixture was stirred for 15 min followed by the addition of BOM-Cl (90% purity, 0.41 mL, 2.6 mmol, 1.0 equiv). The reaction mixture was then stirred for 45 min at room temperature, and the reaction was quenched with 0.25 mL of a 30% w/w aqueous NH3. The solvents were removed in vacuo and the residue was partitioned between water and dichloromethane. The aqueous phase was extracted with dichloromethane (3x5 mL), and the combined organic layers were dried over Na2SO4, filtered, and concentrated in vacuo. Purification by flash chromatography over silica gel (65% EtOAc/hexanes) gave the desired product 1d as a white amorphous solid (220 mg, 0.71 mmol, 27% yield). mp 82–85 °C; IR (neat) 3066, 2982, 2937, 1709, 1622, 1461, 1278, 1245, 1193, 1058, 748, 702 cm−1; 1H NMR (400 MHz, CDCl3) δ 8.56 (s, 1H), 8.08 (d, J = 8.5 Hz, 1H), 8.02 (s, 1H), 7.55 (d, J = 8.5 Hz, 1H), 7.38 – 7.32 (m, 3H), 7.26 – 7.28 (m, 2H), 5.61 (s, 2H), 4.47 (s, 2H), 4.42 (q, J = 7.1 Hz, 2H), 1.43 (t, J = 7.1 Hz, 3H); 13C NMR (101 MHz, CDCl3) δ 167.0, 144.7, 143.9, 136.8, 136.0, 128.9, 128.6, 128.2, 125.7, 125.3, 123.0, 110.2, 73.6, 70.5, 61.1, 14.6; HRMS (ESI+, m/z) [M+H]+ calcd for C18H19N2O3+: 311.1390, found: 311.1410.</p><!><p>1-Fluoro-4-methylsulfonyl-2-nitrobenzene (1.00 g, 4.57 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (11.4 mL, 22.8 mmol, 5 equiv). The vial was sealed with a teflon-lined cap and the mixture was stirred at room temperature for 30 min. The reaction mixture was then concentrated in vacuo, and the resulting residue was suspended in Et2O and filtered to give crude N-methyl-4-(methylsulfonyl)-2-nitroaniline. This solid was then transferred to a 250 mL Morton flask, to which was added Pd/C (1.06 g, 10 mol %), and the solids were suspended in EtOH (100 mL), placed under an atmosphere of H2, and stirred at room temperature for 2.5 h. The reaction mixture was then filtered through pad of celite, rinsed with MeOH and concentrated to give crude N-methyl-4-(methylsulfonyl)benzene-1,2-diamine. The crude reaction product was transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8cm), to which was added formic acid (90% w/w in water, 17.1 mL, 407 mmol, 90 equiv). The vial was sealed with a teflon-lined cap and heated at 150 C in a Biotage Initiator+ microwave reactor for 1 h. The solution was made basic with 2M NaOH followed by concentrated NaOH until a pH ~9 was reached. The mixture was extracted with CH2Cl2 (3x50 mL), and the organic layers were combined, dried over Na2SO4 and concentrated in vacuo. Purification by flash chromatography over silica gel (100% EtOAc) gave the desired product 1f as a pale orange amorphous solid (711 mg, 3.40 mmol, 68% yield). mp 162–164 °C; IR (neat) 3081, 2995, 2904, 1612, 1501, 1329, 1284, 1259, 1136, 1061, 769, 637 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.42 (s, 1H), 8.04 (s, 1H), 7.90 (d, J = 8.5 Hz, 1H), 7.54 (d, J = 8.5 Hz, 1H), 3.92 (s, 3H), 3.09 (s, 3H); 13C NMR (151 MHz, CDCl3) δ 146.4, 143.5, 138.0, 134.7, 121.9, 121.0, 110.4, 45.3, 31.6; HRMS (ESI+, m/z) [M+H]+ calcd for C9H11N2O2S+: 211.0536, found: 211.0528.</p><!><p>4-Chloro-2-fluoro-1-nitro-benzene (1.00 g, 5.70 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (14.2 mL, 28.5 mmol, 5 equiv). The vial was sealed with a teflon-lined cap and the reaction mixture was stirred at room temperature for 1 h. The resulting suspension was filtered over a sintered glass funnel and the filtrate was concentrated in vacuo to give crude 5-chloro-N-methyl-2-nitroaniline as an orange solid. A slurry of Raney nickel in water was then charged in a 100 mL round-bottom flask, the water was evaporated in vacuo to give ca. 1.8 g of dry Raney nickel (30.7 mmol, 4.8 equiv). Crude 5-chloro-N-methyl-2-nitroaniline was then added to the Raney nickel as a solution in MeOH (50 mL), and the resulting suspension was vigorously stirred under a hydrogen atmosphere for 2 h. The resulting colorless suspension was filtered over a bed of celite (Et2O washings) and then concentrated in vacuo to give crude 5-chloro-N1-methylbenzene-1,2-diamine as a black oil. This oil was then transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8cm), to which was added formic acid (90% w/w in water, 20.4 mL, 486 mmol, 85 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 2.5 h in a Biotage Initiator+ microwave reactor. The reaction mixture was concentrated in vacuo to give a bright pink solid. Purification by reverse-phase flash chromatography over C18 silica gel (20 to 40% MeCN+0.1% TFA/H2O+0.1% TFA) gave the desired product as its TFA salt, which was then treated with an aqueous saturated solution of NaHCO3. Extraction with EtOAc, followed by drying with Na2SO4 and in vacuo concentration led to the desired product 1g as a white amorphous solid (720 mg, 4.32 mmol, 76% overall yield). 1H NMR spectra matched the reported data.17 1H NMR (600 MHz, CDCl3) δ 7.84 (s, 1H), 7.70 (d, J = 8.7 Hz, 1H), 7.38 (d, J = 2.0 Hz, 1H), 7.25 (dd, J = 8.7, 2.0 Hz, 1H), 3.81 (s, 3H).</p><!><p>4-Chloro-3-nitro-benzonitrile (1.00 g, 5.48 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (13.7 mL, 2M in THF, 5 equiv). The vial was sealed with a teflon-lined cap, and the reaction mixture was stirred at room temperature for 30 min. The reaction mixture was then concentrated in vacuo. The resulting residue was suspended in Et2O and filtered to give crude 4-methylamino-3-nitro-benzonitrile, which was then suspended in EtOH/H2O (10:1, 19 mL) under N2 at room temperature. Iron powder (1.00 g, 1.79 mmol, 3.3 equiv) and CaCl2 (0.45 g, 4.1 mmol, 0.7 equiv) were added to the reaction mixture, and the mixture was heated at reflux for 2 h. The reaction mixture was cooled to room temperature and was filtered through a pad of celite. The filtrate was concentrated in vacuo and the residue was re-dissolved in CH2Cl2, washed with water and brine, and dried over Na2SO4. The solvent was removed in vacuo to give crude 3-amino-4-methylamino-benzonitrile. To trimethyl orthoformate (15.8 mL) was added the 3-amino-4-methylamino-benzonitrile and para-toluenesulfonic acid monohydrate (17 mg, 0.089 mmol, 0.016 equiv). The reaction mixture was stirred at 100 °C for 2.5 h, after which time the mixture was cooled and water (60 mL) was added. The solution was partitioned with EtOAc (2x90 mL) and the organic layers were combined, dried with MgSO4 and concentrated. Purification by flash chromatography over silica gel (100% EtOAc) gave the desired product 1h as a light brown amorphous solid (339 mg, 2.15 mmol, 39% overall yield). mp 136–137 °C; IR (neat) 3090, 3030, 2955, 2223, 1614, 1503, 1470, 1332, 1260, 877, 825, 646 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.14 (s, 1H), 8.00 (s, 1H), 7.58 (d, J = 8.4 Hz, 1H), 7.47 (d, J = 8.4 Hz, 1H), 3.90 (s, 3H); 13C NMR (151 MHz, CDCl3) δ 146.0, 143.5, 137.3, 126.5, 125.7, 119.9, 110.7, 105.8, 31.5; HRMS (ESI+, m/z) [M+H]+ calcd for C9H8N3+: 158.0713, found: 158.0719.</p><!><p>2-Chloro-1-nitro-3-(trifluoromethyl)benzene (1.20 g, 5.32 mmol, 1.0 equiv) was weighed in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (12 mL, 24 mmol, 4.5 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 15 min in a Biotage Initiator+ microwave reactor. The resulting suspension was filtered over a sintered glass funnel, and the filtrate was concentrated in vacuo to give crude N-methyl-2-nitro-6-(trifluoromethyl)aniline as a yellow solid. This solid was then transferred to a 250 mL Morton flask, to which was added PtO2 (134 mg, 0.591 mmol, 0.11 equiv) and absolute ethanol (50 mL). This suspension was vigorously stirred under a hydrogen atmosphere for 2 h. The resulting colorless suspension was filtered over a bed of celite (Et2O washings) and then concentrated in vacuo to give crude N1-methyl-6-(trifluoromethyl)benzene-1,2-diamine as an orange oil. This oil was then transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), to which was added formic acid (90% w/w in water, 15.0 mL, 358 mmol, 67 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 10 min in a Biotage Initiator+ microwave reactor. The reaction mixture was concentrated in vacuo to give a red oil. Purification by flash chromatography over silica gel (80% EtOAc/petroleum ether) gave the desired product 1i as a pale yellow amorphous solid (702 mg, 3.51 mmol, 66% overall yield). 1H NMR spectra matched the reported data.18 1H NMR (600 MHz, CDCl3) δ 8.01 (d, J = 8.2 Hz, 1H), 7.94 (s, broad, 1H), 7.64 (d, J = 7.6 Hz, 1H), 7.35 (tapp, J = 7.9 Hz, 1H), 4.00 (q, J = 1.8 Hz, 3H); 19F NMR (470 MHz, CDCl3) δ -55.02.</p><!><p>2-Chloro-1-nitro-4-(trifluoromethyl) benzene (1.00 g, 4.43 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (11.0 mL, 22.2 mmol, 5 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 30 min in a Biotage Initiator+ microwave reactor. The resulting suspension was filtered over a sintered glass funnel, and the filtrate was concentrated in vacuo to give crude N-methyl-2-nitro-5-(trifluoromethyl)aniline as a bright orange solid. This solid was then transferred to a 250 mL Morton flask, to which was added PtO2 (113 mg, 0.500 mmol, 0.11 equiv) and absolute ethanol (50 mL). This suspension was vigorously stirred under a hydrogen atmosphere for 2 h. The resulting colorless suspension was filtered over a bed of celite (Et2O washings) and then concentrated in vacuo to give crude N1-methyl-5-(trifluoromethyl)benzene-1,2-diamine as an orange oil. This oil was then transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), to which was added formic acid (90% w/w in water, 15.0 mL, 358 mmol, 80 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 10 min in a Biotage Initiator+ microwave reactor. The reaction mixture was concentrated in vacuo to give a red oil. Purification by flash chromatography over silica gel (50% acetone/petroleum ether) gave the desired product 1j as a white amorphous solid (745 mg, 3.72 mmol, 84% overall yield). mp 96–98 °C; IR (neat) 3031, 1507, 1338, 1302, 1160, 1093, 1063, 901, 814, 663 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.98 (s, 1H), 7.87 (d, J = 8.5 Hz, 1H), 7.67 (s, 1H), 7.53 (dd, J = 8.5, 1.2 Hz, 1H), 3.88 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 146.0, 145.9, 134.1, 125.4 (q, J = 32 Hz), 124.9 (q, J = 272 Hz), 120.9, 119.2 (q, J = 4 Hz), 107.4 (q, J = 5 Hz), 31.4; 19F NMR (470 MHz, CDCl3) δ -60.67; HRMS (ESI+, m/z): [M+H]+ calcd for C9H8F3N2+: 201.0634, found: 201.0633.</p><!><p>1-Chloro-2-nitro-4-(trifluoromethyl) benzene (1.00 g, 4.43 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (11.0 mL, 22.2 mmol, 5 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 15 min in a Biotage Initiator+ microwave reactor. The resulting suspension was filtered over a sintered glass funnel, and the filtrate was concentrated in vacuo to give crude N-methyl-2-nitro-4-(trifluoromethyl)aniline as a bright orange solid. This solid was then transferred to a 250 mL Morton flask, to which was added PtO2 (113 mg, 0.500 mmol, 0.11 equiv) and absolute ethanol (50 mL). This suspension was vigorously stirred under a hydrogen atmosphere for 2 h. The resulting colorless suspension was filtered over a bed of celite (Et2O washings) and then concentrated in vacuo to give crude N1-methyl-4-(trifluoromethyl)benzene-1,2-diamine as an orange oil. This oil was then transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), to which was added formic acid (90% w/w in water, 15.0 mL, 358 mmol, 80 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 10 min in a Biotage Initiator+ microwave reactor. The reaction mixture was concentrated in vacuo to give a red oil. Purification by flash chromatography over silica gel (50% acetone/petroleum ether) gave the desired product 1k as a white amorphous solid (676 mg, 3.38 mmol, 76% overall yield). 1H NMR spectra matched the reported data.19 1H NMR (600 MHz, CDCl3) δ 8.08 (s, 1H), 7.96 (s, 1H), 7.56 (dd, J = 8.5, 1.2 Hz, 1H), 7.46 (d, J = 8.5 Hz, 1H), 3.88 (s, 3H); 19F NMR (470 MHz, CDCl3) δ -60.60.</p><!><p>1-Chloro-2-nitro-4-(trifluoromethyl) benzene (1.15 g, 5.10 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by powdered silicon carbide (1.15 g) and by a 2 M solution of MeNH2 in THF (12.0 mL, 24.0 mmol, 4.7 equiv). The vial was sealed with a teflon-lined cap and heated at 190 °C for 30 min in a Biotage Initiator+ microwave reactor. The resulting suspension was filtered over a sintered glass funnel, and the filtrate was concentrated in vacuo to give a bright orange solid. This solid was then purified by flash chromatography over silica gel (10% diethyl ether/petroleum ether) to give pure 2-chloro-N-methyl-6-trifluoromethyl-aniline as a bright orange solid (450 mg, 2.15 mmol, 42% yield). This solid was then transferred to a 50 mL Morton flask, to which was added PtO2 (46 mg, 0.20 mmol, 0.1 equiv) and absolute ethanol (16 mL). This suspension was vigorously stirred under a hydrogen atmosphere for 2 h. The resulting colorless suspension was filtered over a bed of celite (Et2O washings) and then concentrated in vacuo to give crude N1-methyl-3-(trifluoromethyl)benzene-1,2-diamine as an orange oil. This oil was then transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), to which was added formic acid (90% w/w in water, 5.50 mL, 131 mmol, 60 equiv). The vial was sealed with a teflon-lined cap and heated at 150 °C for 10 min in a Biotage Initiator+ microwave reactor. The reaction mixture was concentrated in vacuo to give a red oil. Purification by flash chromatography over silica gel (80% EtOAc/petroleum ether) gave the desired product 1l as a pale yellow amorphous solid (208 mg, 1.04 mmol, 20% overall yield). 1H NMR spectra matched the reported data.18 1H NMR (600 MHz, CDCl3) δ 7.99 (s, 1H), 7.58 (ddapp, J = 7.7, 2.3 Hz, 2H), 7.38 (tapp, J = 7.7 Hz, 1H), 3.90 (s, 3H); 19F NMR (470 MHz, CDCl3) δ -60.96.</p><!><p>Under N2, at 0 °C, to a solution of 5-methoxy-1H-benzo[d]imidazole (1.00 g, 6.76 mmol, 1.0 equiv) in dry DMF (15 mL) was added NaH (60% w/w in oil, 405 mg, 10.1 mmol, 1.50 equiv) in one portion. The reaction mixture was stirred at 0 °C for 1 h, then neat BOM-Cl (90% pure, 1.25 mL, 8.10 mmol, 1.2 equiv) was added dropwise. The resulting solution stirred at rt for 16 h. The reaction mixture was then poured onto a 10% w/w LiCl aqueous solution, EtOAc was added, the phases were separated, and the organic phase was washed three more times with a 10% w/w LiCl solution. The organic phase was dried over Na2SO4 and concentrated in vacuo to give a yellow oil. Purification by flash chromatography over silica gel (60 to 100% EtOAc/CH2Cl2) gave, by order of elution, the desired products 1m (325 mg, 1.21 mmol, 18% yield) as a pale yellow amorphous solid and 1n (175 mg, 0.652 mmol, 10% yield) as a pale yellow oil. (1m): mp 67–70 °C; IR (neat) 3099, 2836, 1627, 1587, 1505, 1453, 1361, 1279, 1092, 1022, 955, 817, 733 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.83 (s, 1H), 7.69 (d, J = 8.6 Hz, 1H), 7.38 –7.30 (m, 3H), 7.30 – 7.24 (m, 2H), 6.99 – 6.90 (m, 2H), 5.53 (s, 2H), 4.45 (s, 2H), 3.86 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 157.3, 142.3, 138.6, 136.3, 134.4, 128.8, 128.5, 128.2, 121.1, 112.5, 93.8, 73.4, 70.0, 56.0; HRMS (ESI+, m/z): [M+H]+ calcd for C16H17N2O2+: 269.1285, found: 269.1285. (1n): IR (neat) 2939, 1623, 1593, 1487, 1442, 1349, 1218, 1142, 1071, 941, 829, 697 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.88 (s, 1H), 7.40 (d, J = 8.8 Hz, 1H), 7.37–7.28 (m, 4 H), 7.27–7.24 (m, 2H), 6.98 (dd, J = 8.8, 1.7 Hz, 1H), 5.53 (s, 2H), 4.43 (s, 2H), 3.87 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 156.7, 145.1, 143.4, 136.2, 128.8, 128.4, 128.2, 128.1, 113.8, 110.9, 102.6, 73.6, 70.1, 55.9; HRMS (ESI+, m/z): [M+H]+ calcd for C16H17N2O2+: 269.1285, found: 269.1281.</p><!><p>5,6-Dimethylbenzimidazole (0.500 g, 3.42 mmol, 1.0 equiv) was dissolved in DMF (2.5 mL), and the reaction solution was cooled to 0 °C. Sodium hydride (60% w/w oil oil, 118 mg, 5.13 mmol, 1.5 equiv) was added, and the reaction mixture was stirred for 15 min before the addition of BOM-Cl (90% purity, 0.53 mL, 3.42 mmol, 1.0 equiv). The reaction mixture was stirred for 45 min at room temperature, and the reaction was quenched with 0.25 mL of a 30% w/w aqueous NH3. The solvents were removed in vacuo and the residue was partitioned between water and dichloromethane. The aqueous phase was extracted with dichloromethane (3x5 mL), and the combined organic layers were dried over Na2SO4, filtered, and concentrated in vacuo. Purification by flash chromatography over silica gel (90% EtOAc/hexanes) gave the desired product 1o as a pale yellow amorphous solid (778 mg, 2.92 mmol, 85% yield). mp 98–99 °C; IR (neat) 3089, 2970, 2883, 1704, 1495, 1356, 1215, 1093, 1001, 844, 742, 697 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.83 (s, 1H), 7.59 (s, 1H), 7.38 – 7.26 (m, 6H), 5.52 (s, 2H), 4.43 (s, 2H), 2.40 (s, 6H); 13C NMR (151 MHz, CDCl3) δ 142.8, 142.5, 136.3, 132.9, 132.2, 131.8, 128.7, 128.4, 128.1, 120.5, 110.6, 73.3, 70.0, 20.7, 20.4; HRMS (ESI+, m/z): [M+H]+ calcd for C17H19N2O+: 267.1492, found: 267.1512.</p><!><p>3-Fluoro-2-nitro-pyridine (1.00 g, 7.04 mmol, 1.0 equiv) was weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by a 2 M solution of MeNH2 in THF (17.6 mL, 35.2 mmol, 5 equiv). The vial was sealed with a teflon-lined cap and stirred at rt for 30 min. The resulting suspension was filtered over a sintered glass funnel and the filtrate was concentrated in vacuo to give crude N-methyl-2-nitropyridin-3-amine as a bright yellow solid. This solid was then transferred to a 250 mL Morton flask, to which was added 5% w/w Pd/C (749 mg, 0.352 mmol, 0.05 equiv) and absolute ethanol (50 mL). This suspension was vigorously stirred under a hydrogen atmosphere for 2 h. The resulting colorless suspension was filtered over a bed of celite (Et2O washings) and then concentrated in vacuo to give crude N3-methylpyridine-2,3-diamine as a beige solid. This oil was then transferred into a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8cm), to which was added formic acid (90% w/w in water, 23.6 mL, 564 mmol, 80 equiv). The vial was sealed with a teflon-lined cap and heated at 150°C for 20 min in a Biotage microwave Initiator+ reactor. The reaction mixture was concentrated in vacuo to give a red oil. Purification by flash chromatography over silica gel (100% acetone) gave the desired product 1p as beige crystals (560 mg, 4.21 mmol, 60% overall yield). 1H NMR spectra matched the reported data.20 1H NMR (600 MHz, D2O) δ 8.24 (d, J = 4.9 Hz, 1H), 8.10 (s, 1H), 7.75 (d, J = 8.1 Hz, 1H), 7.26 – 7.07 (m, 1H), 3.72 (s, 3H).</p><!><p>Under N2, at 0°C, to a solution of 1H-imidazo[4,5-b]pyridine (500 mg, 4.20 mmol, 1.0 equiv) in dry DMF (2 mL) was added NaH (60% w/w in oil, 252 mg, 6.30 mmol, 1.50 equiv) in one portion. The reaction mixture was stirred at 0 °C for 10 min, then neat MeI (0.40 mL, 6.30 mmol, 1.5 equiv) was added dropwise. The resulting solution stirred at rt for 16 h. The reaction mixture was then poured onto water, EtOAc was added, the phases were separated, and the organic phase was washed twice more with water. The organic phase was dried over Na2SO4 and concentrated in vacuo to give a red oil. Purification by flash chromatography over silica gel (80% acetone/petroleum ether) gave the desired product 1q as a pale yellow amorphous solid (275 mg, 2.07 mmol, 49% yield). 1H NMR matched the reported data.20 1H NMR (600 MHz, D2O) δ 8.13 (d, J = 5.0 Hz, 1H), 8.09 (s, 1H), 7.87 (d, J = 8.1 Hz, 1H), 7.19 (dd, J = 7.6, 5.0 Hz, 1H), 3.70 (s, 3H).</p><!><p>3-Methyl-3H-imidazo[4,5-b]pyridin-6-ol (250 mg, 1.68 mmol, 1.0 equiv) and K2CO3 (394 mg, 2.85 mmol, 1.7 equiv) were weighted in a 20 mL Biotage microwave vial (#354833, HxD = 8x2.8 cm), followed by MeCN (12 mL). Benzyl bromide (0.30 mL, 2.5 mmol, 1.5 equiv) was added, the vial was sealed with a teflon-lined cap and heated at 80 °C in a Biotage Initiator+ microwave reactor for 5 h. The reaction mixture was filtered over cotton-wool, and the filtrate was concentrated in vacuo. Purification by flash chromatography over silica gel (60% acetone/petroleum ether) gave the desired product 1r as a white amorphous solid (170 mg, 0.710 mmol, 42% yield). mp 119–120 °C; IR (neat) 3079, 2922, 1595, 1512, 1453, 1321, 1221, 1116, 1009, 898, 743, 701 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.26 (d, J = 2.5 Hz, 1H), 7.96 (s, 1H), 7.64 (d, J = 2.5 Hz, 1H), 7.46 (d, J = 7.4 Hz, 2H), 7.39 (t, J = 7.4 Hz, 2H), 7.33 (t, J = 7.4 Hz, 1H), 5.15 (s, 2H), 3.89 (s, 3H); 13C NMR (150 MHz, CDCl3) δ 152.4, 145.0, 142.5, 136.5, 136.0, 135.6, 128.8, 128.3, 127.8, 112.4, 71.5, 30.0; HRMS (ESI+, m/z): [M+H]+ calcd for C14H14N3O+: 240.1131, found: 240.1132.</p><!><p>3-Methylimidazo[4,5-b]pyridine-6-carboxylic acid (0.126 g, 0.710 mmol, 1.0 equiv) was heated in SOCl2 (15 mL) at 80 °C for 2 h. The excess of SOCl2 was removed in vacuo. The remaining residue was dissolved in EtOH (10 mL) and was heated at reflux for 1 h. EtOH was then removed in vacuo. Purification by flash chromatography over silica gel (100% EtOAc) gave the desired product 1s as a white amorphous solid (79 mg, 0.38 mmol, 54% yield). mp 89–90 °C; IR (neat) 3058, 2993, 2928, 1700, 1603, 1403, 1297, 1240, 1180, 1026, 784, 637 cm−1; 1H NMR (600 MHz, CDCl3) δ 9.10 (s, 1H), 8.69 (s, 1H), 8.13 (s, 1H), 4.44 (q, J = 7.1 Hz, 2H), 3.95 (s, 3H), 1.43 (t, J = 7.1 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 166.1, 150.1, 146.7, 146.5, 134.8, 129.7, 121.6, 61.4, 30.1, 14.5; HRMS (ESI+, m/z): [M+H]+ calcd for C10H12N3O2+: 206.0924, found: 206.0930.</p><!><p>In a N2-filled glovebox, an oven-dried Biotage microwave vial (#352016, HxD = 8x1.5 cm) was charged with [Rh(cod)Cl]2 (0.05 equiv), 1,2-bis(bis(3,5-bis(trifluoromethyl)phenyl)phosphino)ethane (dArFpe) (0.125 equiv), potassium phosphate (0.25 equiv) and the azole (1.0 equiv). Toluene was then added (c = 0.6 M), and the reaction mixture was stirred for 5 min to give a yellow-red suspension. The alkene was then added (4.0 equiv). The vial was sealed with a teflon-lined septum, taken out of the glovebox and heated for 24 h at 120 °C in a pre-equilibrated oil bath. The reaction mixture was then filtered over a short pad of silica gel (ca. 2 cm) in a pasteur pipette, followed by washing with acetone until the filtrate came out clear. The filtrate was then concentrated in vacuo, and the products were purified by the indicated chromatography techniques.</p><!><p>0.4 mmol scale: Prepared according to the General Procedure from 1-methyl-1H-benzo[d]imidazole 1a (53 mg, 0.40 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (50% acetone/CH2Cl2), followed by further purification by preparative TLC (5% MeOH/CH2Cl2) gave the desired product 3a as an off-white solid (63 mg, 0.27 mmol, 68% yield). 7.6 mmol scale: Prepared according to the General Procedure from 1-methyl-1H-benzo[d]imidazole 1a (1.00 g, 7.57 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv) using a 10–20 mL Biotage vial (#354833, HxD = 8x2.8 cm). Purification by flash chromatography over silica gel (20% acetone/CH2Cl2+1% aqueous NH4OH) followed by trituration with 10% Et2O/pentane gave the desired product 3a as an off-white amorphous solid (1.32 g, 5.71 mmol, 75% yield). mp 104–105 °C; IR (neat) 2941, 1631, 1495, 1458, 1432, 1390, 1369, 1305, 1283, 1261, 1132, 1072, 769 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.73 (m, 1H), 7.37 – 7.22 (m, 3H), 4.44 (q, J = 7.2 Hz, 1H), 3.80 (s, 3H), 3.01 (s, 3H), 2.98 (s, 3H), 1.68 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.5, 153.2, 142.2, 136.6, 122.7, 122.2, 119.7, 109.3, 39.4, 37.4, 36.4, 30.5, 16.1; HRMS (ESI+, m/z): [M+H]+ calcd for C13H18N3O+: 232.1444, found: 232.1449.</p><!><p>Prepared according to the General Procedure from 1b (78.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (30% acetone/CH2Cl2) gave the desired product 3b as a pale yellow amorphous solid (108 mg, 0.368 mmol, 92% yield). mp 135–137 °C. IR (neat) 3059, 2937, 1642, 1595, 1501, 1454, 1395, 1267, 1148, 1089, 752, 699 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.84 (d, J = 7.9 Hz, 1H), 7.63 – 7.50 (m, 3H), 7.39 – 7.31 (m, 2H), 7.31 – 7.22 (m, 1H), 7.19 (td, J = 7.7, 1.1 Hz, 1H), 7.03 (d, J = 7.9 Hz, 1H), 4.10 (q, J = 7.0 Hz, 1H), 2.80 (s, 3H), 2.66 (s, 3H), 1.66 (d, J = 7.0 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 171.0, 153.8, 142.5, 137.2, 135.8, 130.0, 129.5, 128.1, 123.1, 122.4, 120.0, 110.0, 37.1, 37.0, 36.0, 16.4; HRMS (ESI+, m/z): [M+H]+ calcd for C18H20N3O+: 294.1601, found: 294.1609.</p><!><p>Prepared according to the General Procedure from 1c (95.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (30% acetone/CH2Cl2) gave the desired product 1c as a yellow amorphous solid (129 mg, 0.382 mmol, 96% yield). mp 55–56 °C; IR (neat) 2934, 1648, 1507, 1454, 1394, 1358, 1298, 1137, 1052, 768 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.77 (m, 1H), 7.39 – 7.31 (m, 4H), 7.30 – 7.26 (m, 4H), 5.67 (d, J = 11.0 Hz, 1H), 5.56 (d, J = 11.0 Hz, 1H), 4.51 (s, 2H), 4.44 (q, J = 7.1 Hz, 1H), 2.96 (s, 3H), 2.93 (s, 3H), 1.69 (d, J = 7.1 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.7, 153.9, 142.4, 136.8, 135.9, 128.7, 128.3, 127.9, 123.3, 122.7, 120.0, 109.8, 72.5, 70.6, 38.6, 37.4, 36.3, 16.7; HRMS (ESI+, m/z): [M+H]+ calcd for C20H24N3O2+: 338.1863, found: 338.1869.</p><!><p>Prepared according to the General Procedure from 1d (124 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (30% acetone/CH2Cl2) gave the desired product 3d as a cream-colored amorphous solid (134 mg, 0.330 mmol, 82% yield). mp 93–94 °C; IR (neat) 3064, 2976, 2939, 1700, 1643, 1618, 1420, 1277, 1216, 1112, 1013, 774, 700 cm−1; 1H NMR (500 MHz, CDCl3) δ 8.49 (s, 1H), 8.00 (d, J = 9.5 Hz, 1H), 7.39 – 7.30 (m, 4H), 7.27 – 7.21 (m, 2H), 5.67 (d, J = 11.0 Hz, 1H), 5.55 (d, J = 11.0 Hz, 1H), 4.49 (s, 2H), 4.44 (q, J = 7.2 Hz, 1H), 4.39 (q, J = 7.1 Hz, 2H), 2.98 (s, 3H), 2.93 (s, 3H), 1.71 (d, J = 7.2 Hz, 3H), 1.40 (t, J = 7.1 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 170.6, 167.1, 155.5, 142.0, 139.1, 136.4, 128.8, 128.5, 128.0, 125.3, 124.8, 122.3, 109.5, 72.6, 70.6, 61.0, 38.6, 37.4, 36.3, 16.4, 14.5; HRMS (ESI+, m/z): [M+H]+ calcd for C23H28N3O4+: 410.2074, found: 410.2074.</p><!><p>Prepared according to the General Procedure from ethyl 1-methyl-1H-benzo[d]imidazole-5-carboxylate 1e (82.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (5 to 20% Acetone/CH2Cl2) gave the desired product 3e as an orange amorphous solid (87 mg, 0.287 mmol, 72% yield). mp 108–110 °C; IR (neat) 2982, 1705, 1636, 1618, 1453, 1392, 1296, 1247, 1205, 1102, 1083, 907, 773 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.46 (s, 1H), 8.01 (dd, J = 8.6, 1.2 Hz, 1H), 7.31 (d, J = 8.6 Hz, 1H), 4.44 (q, J = 7.2 Hz, 1H), 4.39 (q, J = 7.1 Hz, 2H), 3.81 (s, 3H), 3.03 (s, 3H), 2.99 (s, 3H), 1.70 (d, J = 7.2 Hz, 3H), 1.40 (t, J = 7.1 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.3, 167.3, 155.0, 141.8, 139.7, 124.8, 124.3, 122.0, 108.9, 61.0, 39.3, 37.5, 36.4, 30.7, 15.9, 14.5; HRMS (ESI+, m/z): [M+H]+ calcd for C16H22N3O3+: 304.1656, found: 304.1673.</p><!><p>Prepared according to the General Procedure from 1f (84.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (0–5% methanol/EtOAc) followed by trituration with 10% Et2O/pentane gave the desired product 3f as a cream-colored solid (105 mg, 0.34 mmol, 85% yield). mp 176–178 C; IR (neat) 3012, 2997, 2933, 1635, 1421, 1292, 1246, 1143, 974, 774, 620 cm−1; 1H NMR (500 MHz, CDCl3) δ 8.34 (s, 1H), 7.87 (d, J = 8.5 Hz, 1H), 7.45 (d, J = 8.5 Hz, 1H), 4.47 (q, J = 7.2 Hz, 1H), 3.85 (s, 3H), 3.07 (s, 3H), 3.07 (s, 3H), 3.00 (s, 3H), 1.72 (d, J = 7.2 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 170.1, 156.5, 142.0, 139.9, 134.4, 121.5, 120.1, 110.1, 45.3, 39.1, 37.5, 36.4, 30.9, 15.8; HRMS (ESI+, m/z): [M+H]+ calcd for C14H20N3O3S+: 310.1220, found: 310.1217.</p><!><p>Prepared according to the General Procedure from 1g (82.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (5 to 20% acetone/CH2Cl2) gave the desired product 3g as a pale yellow amorphous solid (94 mg, 0.354 mmol, 88% yield. mp 139–140 °C; IR (neat) 3065, 2950, 1644, 1502, 1474, 1392, 1270, 1133, 1052, 922, 821, 671 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.62 (d, J = 8.6 Hz, 1H), 7.30 (d, J = 1.9 Hz, 1H), 7.21 (dd, J = 8.6, 1.9 Hz, 1H), 4.40 (q, J = 7.2 Hz, 1H), 3.75 (s, 3H), 3.02 (s, 3H), 2.99 (s, 3H), 1.68 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.3, 154.1, 140.9, 137.2, 128.5, 122.8, 120.5, 109.5, 39.3, 37.4, 36.4, 30.6, 16.0; HRMS (ESI+, m/z): [M+H]+ calcd for C13H17ClN3O+: 266.1055, found: 266.1055.</p><!><p>Prepared according to the General Procedure from 1h (63 mg, 0.40 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv). Purification by reverse-phase flash chromatography over C18 silica gel (10 to 20% MeCN+0.1% TFA/H2O+0.1% TFA) gave the desired product as its TFA salt, which was treated with an aqueous saturated solution of NaHCO3. Extraction with EtOAc, followed by drying with Na2SO4 and in vacuo concentration led to the desired product 3h as a white amorphous solid (51 mg, 0.20 mmol, 50% yield). mp 150 °C (decomp); IR (neat) 3083, 2994, 2945, 2216, 1634, 1615, 1483, 1388, 1297, 1075, 915, 823, 636 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.04 (s, 1H), 7.53 (dd, J = 8.4, 1.5 Hz, 1H), 7.37 (d, J = 8.4 Hz, 1H), 4.44 (q, J = 7.2 Hz, 1H), 3.82 (s, 3H), 3.06 (s, 3H), 3.00 (s, 3H), 1.71 (d, J = 7.2 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 170.0, 156.1, 141.9, 139.3, 126.2, 124.7, 120.1, 110.4, 105.5, 39.9, 37.5, 36.4, 30.8, 15.8; HRMS (ESI+, m/z): [M+H]+ calcd for C14H17N4O+: 257.1397, found: 257.1405.</p><!><p>Prepared according to the General Procedure from 1i (80.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (35% acetone/CH2Cl2) gave the desired product 3i as a pale yellow amorphous solid (101 mg, 0.337 mmol, 85% yield). mp 112–114 °C; IR (neat) 2993, 1641, 1528, 1461, 1388, 1284, 1215, 1169, 1108, 1089, 957, 805, 750 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.94 (d, J = 8.1 Hz, 1H), 7.59 (d, J = 7.7 Hz, 1H), 7.29 (tapp, J = 7.9 Hz, 1H), 4.40 (q, J = 7.2 Hz, 1H), 3.88 (q, J = 1.7 Hz, 3H), 3.023 (s, 3H), 3.02 (s, 3H), 1.72 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.4, 155.6, 144.4, 132.2, 124.3, 124.0 (q, J = 271 Hz), 121.22 (q, J = 6 Hz), 121.18, 113.7 (q, J = 33 Hz), 39.1, 37.4, 36.5, 32.5 (q, J = 6 Hz), 16.0; 19F NMR (470 MHz, CDCl3) δ -53.94; HRMS (ESI+, m/z): [M+H]+ calcd for C14H17F3N3O+: 300.1318, found: 300.1324</p><!><p>Prepared according to the General Procedure from 1j (80.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (35% acetone/CH2Cl2) gave the desired product 3j as a pale yellow oil (111 mg, 0.371 mmol, 93% yield). mp 128–130 °C; IR (neat) 2945, 1647, 1471, 1391, 1341, 1307, 1278, 1154, 1107, 1048, 923, 832, 666 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.80 (d, J = 8.3 Hz, 1H), 7.60 (s, 1H), 7.51 (d, J = 8.3 Hz, 1H), 4.45 (q, J = 7.1 Hz, 1H), 3.84 (s, 3H), 3.03 (s, 3H), 3.00 (s, 3H), 1.71 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.1, 155.9, 144.5, 136.0, 125.0 (q, J = 32 Hz), 124.9 (q, J = 272 Hz), 120.0, 119.2 (q, J = 4 Hz), 107.1 (q, J = 4 Hz), 39.3, 37.5, 36.4, 30.7, 15.9; 19F NMR (470 MHz, CDCl3) δ -60.66; HRMS (ESI+, m/z): [M+H]+ calcd for C14H17F3N3O+: 300.1318, found: 300.1313.</p><!><p>Prepared according to the General Procedure from 1k (80.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (35% acetone/CH2Cl2) gave the desired product 3k as a pale yellow oil (105 mg, 0.351 mmol, 88% yield). mp 148–149°C; IR (neat) 2944, 1646, 1629, 1495, 1391, 1326, 1256, 1228, 1098, 1049, 925, 816, 707 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.01 (s, 1H), 7.53 (d, J = 8.3 Hz, 1H), 7.39 (d, J = 8.3 Hz, 1H), 4.46 (q, J = 7.1 Hz, 1H), 3.83 (s, 3H), 3.04 (s, 3H), 2.99 (s, 3H), 1.70 (d, J = 7.1 Hz, 3H). 13C NMR (150 MHz, CDCl3) δ 170.2, 155.3, 141.7, 138.5, 125.0 (q, J = 272 Hz), 124.8 (q, J = 32 Hz), 119.7 (q, J = 4 Hz), 117.4 (q, J = 4 Hz), 109.7, 39.3, 37.5, 36.4, 30.8, 15.9; 19F NMR (470 MHz, CDCl3) δ -60.66; HRMS (ESI+, m/z): [M+H]+ calcd for C14H17F3N3O+: 300.1318, found: 300.1331.</p><!><p>Prepared according to the General Procedure from 1l (80.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (5 to 20% acetone/CH2Cl2) gave the desired product 3l as a pale yellow oil (15 mg, 0.050 mmol, 12% yield). IR (neat) 2926, 1642, 1464, 1428, 1393, 1318, 1212, 1119, 1101, 936, 754 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.53 (d, J = 7.6 Hz, 1H), 7.50 (d, J = 8.1 Hz, 1H), 7.33 (tapp, J = 7.8 Hz, 1H), 4.65 (q, J = 7.2 Hz, 1H), 3.87 (s, 3H), 3.08 (s, 3H), 2.97 (s, 3H), 1.68 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.2, 154.7, 139.0 (q, J = 1 Hz), 137.8, 124.1 (q, J = 273 Hz), 121.8, 120.7 (q, J = 32 Hz), 119.6 (q, J = 5 Hz), 113.2, 39.8, 37.6, 36.5, 31.0, 16.0; 19F NMR (470 MHz, CDCl3) δ -60.77; HRMS (ESI+, m/z): [M+H]+ calcd for C14H17F3N3O+: 300.1318, found: 300.1324.</p><!><p>Prepared according to the General Procedure from 1m (107 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (20 to 40% acetone/CH2Cl2) gave the desired product 3m as a pale yellow oil (124 mg, 0.337 mmol, 84% yield). IR (neat) 2936, 1644, 1486, 1454, 1393, 1207, 1139, 1074, 1056, 816, 733 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.63 (d, J = 8.7 Hz, 1H), 7.38 – 7.27 (m, 5H), 6.89 (dd, J = 8.7, 2.3 Hz, 1H), 6.79 (d, J = 2.3 Hz, 1H), 5.61 (d, J = 11.0 Hz, 1H), 5.50 (d, J = 11.0 Hz, 1H), 4.50 (s, 2H), 4.40 (q, J = 7.2 Hz, 1H), 3.81 (s, 3H), 2.96 (s, 3H), 2.93 (s, 3H), 1.67 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.8, 157.0, 152.8, 136.8 (2C), 136.5, 128.7, 128.3, 128.0, 120.4, 111.7, 93.8, 72.4, 70.3, 56.0, 38.7, 37.3, 36.3, 16.6; HRMS (ESI+, m/z): [M+H]+ calcd for C21H26N3O3+: 368.1969, found: 368.1970.</p><!><p>Prepared according to the General Procedure from 1n (107 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (20 to 40% acetone/CH2Cl2) gave the desired product 3n as a pale yellow oil (139 mg, 0.378 mmol, 95% yield). IR (neat) 2938, 1643, 1487, 1441, 1394, 1276, 1197, 1151, 1057, 949, 736 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.38 – 7.29 (m, 3H), 7.28 – 7.26 (m, 3H), 7.22 (d, J = 8.8 Hz, 1H), 6.91 (dd, J = 8.8, 2.3 Hz, 1H), 5.62 (d, J = 11.0 Hz, 1H), 5.52 (d, J = 11.0 Hz, 1H), 4.49 (s, 2H), 4.39 (q, J = 7.2 Hz, 1H), 3.85 (s, 3H), 2.95 (s, 3H), 2.93 (s, 3H), 1.68 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.6, 156.3, 154.0, 143.0, 136.5, 130.1, 128.5, 128.1, 127.8, 112.8, 110.0, 102.2, 72.3, 70.3, 55.7, 38.2, 37.1, 36.1, 16.5; HRMS (ESI+, m/z): [M+H]+ calcd for C21H26N3O3+: 368.1969, found: 368.1959.</p><!><p>Prepared according to the General Procedure from 1o (107 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (30% acetone/CH2Cl2) gave the desired product 3o as a cream-colored amorphous solid (85 mg, 0.23 mmol, 58% yield). mp 121–123 °C; IR (neat) 2982, 2948, 2844, 1651, 1466, 1382, 1069, 1026, 852, 743, 697 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.51 (s, 1H), 7.38 – 7.27 (m, 5H), 7.08 (s, 1H), 5.62 (d, J = 11.0 Hz, 1H), 5.51 (d, J = 11.0 Hz, 1H), 4.50 (s, 2H), 4.39 (q, J = 7.2 Hz, 1H), 2.93 (s, 3H), 2.92 (s, 3H), 2.36 (s, 6H), 1.66 (d, J = 7.2 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 170.8, 153.0, 141.0, 136.9, 134.4, 132.3, 131.5, 128.7, 128.3, 128.0, 120.1, 110.1, 72.4, 70.4, 38.8, 37.3, 36.4, 20.7, 20.4, 16.8; HRMS (ESI+, m/z): [M+H]+ calcd for C22H28N3O2+: 366.2176, found: 366.2186.</p><!><p>Prepared according to the General Procedure from 1p (100 mg, 0.751 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.31, 3.0 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (5 to 10% MeOH/CH2Cl2) gave the desired product 3p as a beige amorphous solid (148 mg, 0.637 mmol, 85% yield). mp 141–143 °C; IR (neat) 2940, 1637, 1612, 1506, 1445, 1408, 1389, 1274, 1142, 1081, 781, 755, 581 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.52 (dd, J = 4.8, 1.2 Hz, 1H), 7.61 (dd, J = 7.9, 1.2 Hz, 1H), 7.19 (dd, J = 7.9, 4.8 Hz, 1H), 4.55 (q, J = 7.2 Hz, 1H), 3.83 (s, 3H), 3.07 (s, 3H), 2.97 (s, 3H), 1.71 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.2, 155.6, 155.2, 144.9, 128.9, 117.8, 117.2, 39.7, 37.6, 36.4, 30.8, 15.8; HRMS (ESI+, m/z): [M+H]+ calcd for C12H17N4O+: 233.1397, found: 233.1397.</p><!><p>Prepared according to the General Procedure from 1q (53.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (60% acetone/CH2Cl2) gave the desired product 3q as a pale yellow amorphous solid (90 mg, 0.387 mmol, 97% yield). mp 105–106 °C; IR (neat) 3029, 2977, 1636, 1601, 1509, 1475, 1448, 1399, 1281, 1132, 1076, 806, 782, 702 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.35 (dd, J = 4.8, 1.4 Hz, 1H), 7.99 (dd, J = 8.0, 1.4 Hz, 1H), 7.21 (dd, J = 8.0, 4.8 Hz, 1H), 4.39 (q, J = 7.2 Hz, 1H), 3.86 (s, 3H), 3.02 (s, 3H), 3.00 (s, 3H), 1.71 (d, J = 7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.1, 154.7, 148.9, 143.7, 134.5, 127.1, 118.3, 39.3, 37.4, 36.4, 29.0, 15.7; HRMS (ESI+, m/z): [M+H]+ calcd for C12H17N4O+: 233.1397, found: 233.1406.</p><!><p>Prepared according to the General Procedure from 1r (96.0 mg, 0.400 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (40% acetone/CH2Cl2) gave the desired product 3r as an off-white amorphous solid (125 mg, 0.369 mmol, 92% yield). mp 128–129 °C; IR (neat) 2941, 1636, 1498, 1453, 1393, 1258, 1174, 1028, 869, 731, 692 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.19 (d, J = 2.4 Hz, 1H), 7.56 (d, J = 2.4 Hz, 1H), 7.45 (d, J = 7.2 Hz, 2H), 7.38 (tapp, J = 7.3 Hz, 2H), 7.32 (m, 1H), 5.13 (s, 2H), 4.33 (q, J =7.2 Hz, 1H), 3.81 (s, 3H), 3.00 (s, 6H), 1.67 (d, J =7.2 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 170.1, 155.0, 152.4, 143.8, 136.6, 134.9, 134.6, 128.8, 128.3, 127.7, 112.0, 71.6, 39.3, 37.4, 36.4, 29.0, 15.7; HRMS (ESI+, m/z): [M+H]+ calcd for C19H23N4O2+: 339.1816, found: 339.1821.</p><!><p>Prepared according to the General Procedure from 1s (82 mg, 0.40 mmol, 1.0 equiv) and N,N-dimethylacrylamide 2 (0.17 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (20% acetone/CH2Cl2) gave the desired product 3s as a pale yellow solid (90 mg, 0.30 mmol, 74% yield). mp 137–138 °C; IR (neat) 3041, 2987, 2942, 1715, 1636, 1602, 1481, 1371, 1293, 1197, 1086, 1009, 789, 621 cm−1; 1H NMR (400 MHz, CDCl3) δ 9.04 (d, J = 1.8 Hz, 1H), 8.61 (d, J = 1.8 Hz, 1H), 4.46 – 4.37 (m, 3H), 3.88 (s, 3H), 3.05 (s, 3H), 3.01 (s, 3H), 1.73 (d, J = 7.2 Hz, 3H), 1.42 (t, J = 7.1 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 169.9, 166.2, 156.7, 151.5, 146.0, 133.8, 128.6, 121.6, 61.3, 39.3, 37.5, 36.4, 29.3, 15.5, 14.5; HRMS (ESI+, m/z): [M+H]+ calcd for C15H21N4O3+: 305.1608, found: 305.1624.</p><!><p>In a N2-filled glovebox, a flame-dried Biotage microwave vial (#351521, HxD = 8x1.5 cm) was charged with bis(cyclopentadienyl)zirconium (IV) chloride hydride (62 mg, 0.24 mmol, 1.2 equiv) followed by THF (3 mL). To this suspension was added 3a (46 mg, 0.2 mmol, 1.0 equiv) in THF (2.1 mL), and the vial was sealed with a teflon-lined septum. The mixture was stirred at room temperature for 1.5 h, after which time NH2-OMe•HCl (50 mg, 0.6 mmol, 3 equiv) was added, and the reaction mixture was stirred for an additional 2 h. The reaction mixture was then concentrated in vacuo. An aqueous saturated solution of NaHCO3 was added until a pH ~8 was reached, and the resulting mixture was then extracted with EtOAc (3 x 25 mL). The organic phase was dried over Na2SO4 and concentrated in vacuo. Purification by flash chromatography over silica gel (15% EtOAc/hexanes + 1% Et3N) gave a mixture of oxime isomers as a clear oil (38 mg, 0.176 mmol, 88% yield, 3:1 dr). For characterization purposes, stereoisomerically pure samples of each of the isomers were obtained by chromatography over silica gel (1% acetone/CH2Cl2 + 1% Et3N) followed by a second silica gel column (10% EtOAc/Hexanes + 1% Et3N). (major isomer of 4a): IR (neat) 2991, 2937, 2824, 1615, 1500, 1473, 1279, 1048, 1031, 840, 748 cm−1; 1H NMR (400 MHz, CDCl3) δ 7.76 (d, J = 6.8 Hz, 1H), 7.50 (d, J = 7.6 Hz, 1H), 7.36 – 7.23 (m, 3H), 4.03 (p, J = 7.1 Hz, 1H), 3.84 (s, 3H), 3.78 (s, 3H), 1.68 (d, J = 7.0 Hz, 3H); 13C NMR (151 MHz, CDCl3) δ 153.9, 150.6, 142.5, 136.0, 122.7, 122.2, 119.7, 109.2, 61.8, 33.7, 30.0, 17.3; HRMS (ESI+, m/z): [M+H]+ calcd for C12H16N3O+: 218.1288, found: 218.1310. (minor isomer of 4a): IR (neat) 2976, 2938, 2833, 1616, 1504, 1468, 1279, 1050, 1031, 872, 743 cm−1; 1H NMR (400 MHz, CDCl3) δ 7.76 (d, J = 6.6 Hz, 1H), 7.34 – 7.23 (m, 3H), 6.91 (d, J = 7.4 Hz, 1H), 4.67 (p, J = 7.1 Hz, 1H), 3.94 (s, 3H), 3.74 (s, 3H), 1.61 (d, J = 7.0 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 154.5, 150.5, 142.6, 135.8, 122.6, 122.2, 119.8, 109.3, 62.1, 29.7, 28.8, 16.7; HRMS (ESI+, m/z): [M+H]+ calcd for C12H16N3O+: 218.1288, found: 218.1289.</p><!><p>In a N2-filled glovebox, a flame-dried Biotage microwave vial (#351521, HxD = 8x1.5 cm) was charged with bis(cyclopentadienyl)zirconium (IV) chloride hydride (62 mg, 0.24 mmol, 1.2 equiv) followed by THF (3 mL). To this suspension was added 3p (47 mg, 0.2 mmol, 1.0 equiv) suspended in THF (3.1 mL), and the vial was sealed with a teflon-lined septum. The mixture was stirred at room temperature for 2.5 h, after which time NH2-OMe•HCl (50 mg, 0.6 mmol, 3 equiv) was added, and the reaction mixture was stirred for an additional 2.5 h. The reaction mixture was the concentrated in vacuo. An aqueous saturated solution of NaHCO3 was added until a pH ~8 was reached, and the resulting mixture was extracted with EtOAc (3 x 25 mL). The organic phase was dried over Na2SO4 and concentrated in vacuo. Purification by flash chromatography over silica gel (80% EtOAc/hexanes + 1% Et3N) gave a mixture of E- and Z-isomers as a pale yellow oil (29 mg, 0.131 mmol, 66% yield, 3:1 dr). IR (neat) 2987, 2939, 2819, 1612, 1503, 1454, 1408, 1279, 1050, 888, 778, 731 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.51 (d, J = 4.7 Hz, 1H), 7.62 (d, J = 8.0 Hz, 1H), 7.52 (d, J = 7.6 Hz, 0.75H), 7.19 (t, J = 8.0, 4.8 Hz, 1H), 6.96 (d, J = 7.3 Hz, 0.25H), 4.68 (p, J = 7.0 Hz, 0.25H), 4.03 (p, J = 7.1 Hz, 0.75H), 3.93 (s, 0.75H), 3.83 (s, 2.25H), 3.80 (s, 2.25H), 3.77 (s, 0.75H), 1.69 (d, J = 7.0 Hz, 2.25H), 1.62 (d, J = 7.0 Hz, 0.75H); 13C NMR (151 MHz, CDCl3) δ 156.9, 156.4, 155.4, 155.3, 150.2, 149.9, 144.6, 144.6, 128.1, 127.9, 117.7, 117.6, 116.99, 116.97, 62.0, 61.7, 33.7, 29.9, 29.7, 28.8, 17.1, 16.6; HRMS (ESI+, m/z): [M+H]+ calcd for C11H15N4O+: 219.1240, found: 219.1247.</p><!><p>Prepared according to the General Procedure from 1c (95 mg, 0.400 mmol, 1.0 equiv) and ethyl methacrylate 5 (0.190 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (5 to 10% diethyl ether/CH2Cl2) gave, by order of elution, the desired products 6c as a pale yellow oil (45 mg, 0.128 mmol, 32% yield) and 6c' as a pale yellow oil (48 mg, 0.136 mmol, 34% yield). (6c): IR (neat) 2985, 1728, 1457, 1361, 1247, 1147, 1064, 1025, 909, 728 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.88 – 7.74 (m, 1H), 7.43 – 7.19 (m, 8H), 5.46 (s, 2H), 4.56 (s, 2H), 4.10 (q, J = 7.1 Hz, 2H), 1.81 (s, 6H), 1.13 (t, J = 7.1 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 175.1, 156.1, 141.7, 136.6, 136.4, 128.7, 128.3, 127.9, 123.4, 122.6, 120.0, 110.1, 73.0, 70.8, 61.7, 44.7, 26.1, 14.1; HRMS (ESI+, m/z): [M+H]+ calcd for C21H25N2O3+: 353.1860, found: 353.1851. (6c'): IR (neat) 2979, 1725, 1519, 1456, 167, 1281, 1178, 1066, 906, 736, 697 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.87 – 7.61 (m, 1H), 7.40 – 7.21 (m, 8H), 5.65 (d, J = 11.3 Hz, 1H), 5.54 (d, J = 11.3 Hz, 1H), 4.49 (s, 2H), 4.17 – 4.04 (m, 2H), 3.39 – 3.28 (m, 2H), 2.93 (qapp, J = 9.6 Hz, 1H), 1.33 (d, J = 6.7 Hz, 3H), 1.20 (t, J = 7.1 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 175.7, 153.7, 142.7, 136.7, 135.3, 128.7, 128.3, 127.9, 122.8, 122.5, 119.5, 109.5, 71.9, 70.3, 60.7, 38.3, 30.8, 17.7, 14.2; HRMS (ESI+, m/z): [M+H]+ calcd for C21H25N2O3+: 353.1860, found: 353.1847.</p><!><p>Prepared according to the General Procedure from 1j (80.0 mg, 0.400 mmol, 1.0 equiv) and ethyl methacrylate 5 (0.190 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (10% acetone/CH2Cl2) gave the desired product 6j as a pale yellow oil (108 mg, 0.344 mmol, 86% yield). IR (neat) 2987, 1732, 1509, 1469, 1335, 1279, 1151, 1010, 1049, 822, 732 cm−1; 1H NMR (600 MHz, CDCl3) δ 7.85 (d, J = 8.4 Hz, 1H), 7.59 (s, 1H), 7.52 (d, J = 8.5 Hz, 1H), 4.22 (q, J = 7.1 Hz, 2H), 3.69 (s, 3H), 1.78 (s, 6H), 1.22 (t, J = 7.1 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 174.8, 158.7, 144.0, 136.3, 125.0 (q, J = 32 Hz), 124.9 (d, J = 272 Hz), 120.2, 119.2 (q, J = 3 Hz), 107.0 (q, J = 4 Hz), 62.0, 44.6, 31.1, 25.5, 14.3; 19F NMR (470 MHz, CDCl3) δ -60.63. HRMS (ESI+, m/z): [M+H]+ calcd for C15H18F3N2O2+: 315.1315, found: 315.1310.</p><!><p>Prepared according to the General Procedure from 1q (53.0 mg, 0.400 mmol, 1.0 equiv) and ethyl methacrylate 5 (0.190 mL, 1.6 mmol, 4.0 equiv). Purification by flash chromatography over silica gel (10% acetone/CH2Cl2) gave the desired product 4q as a pale yellow oil (95 mg, 0.384 mmol, 96% yield). IR (neat) 2982, 1731, 1602, 1500, 1464, 1407, 1387, 1283, 258, 1229, 1147, 1021, 774 cm−1; 1H NMR (600 MHz, CDCl3) δ 8.36 (dd, J = 4.8, 1.4 Hz, 1H), 8.03 (dd, J = 7.9, 1.4 Hz, 1H), 7.22 (dd, J = 7.9, 4.8 Hz, 1H), 4.22 (q, J = 7.1 Hz, 2H), 3.75 (s, 3H), 1.78 (s, 6H), 1.22 (t, J = 7.1 Hz, 3H); 13C NMR (150 MHz, CDCl3) δ 174.7, 157.3, 149.3, 143.8, 134.0, 127.2, 118.3, 61.9, 44.8, 29.5, 25.2, 14.3; HRMS (ESI+, m/z): [M+H]+ calcd for C13H18N3O2+: 248.1394, found: 248.1400.</p>

PubMed Author Manuscript

Concerted proton-electron transfer oxidation of phenols and hydrocarbons by a high-valent nickel complex

The high-valent nickel(III) complex Ni(pyalk) 2 + (2) was prepared by oxidation of a nickel(II) complex, Ni(pyalk) 2 (1) (pyalk ¼ 2-pyridyl-2-propanoate). 2 and derivatives were fully characterized by mass spectrometry and X-ray crystallography. Electron paramagnetic resonance spectroscopy and X-ray photoelectron spectroscopy confirm that the oxidation is metal-centered. 2 was found to react with a variety of phenolic and hydrocarbon substrates. A linear correlation between the measured rate constant and the substrate bond dissociation enthalpy (BDE) was found for both phenolic and hydrocarbon substrates. Large H/D kinetic isotope effects were also observed for both sets of substrates.These results suggest that 2 reacts through concerted proton-electron transfer (CPET). Analysis of measured thermodynamic parameters allows us to calculate a bond dissociation free energy (BDFE) of $91 kcal mol À1 for the O-H bond of the bound pyalk ligand. These findings may shed light onto CPET steps in oxidative catalysis and have implications for ligand design in catalytic systems.

concerted_proton-electron_transfer_oxidation_of_phenols_and_hydrocarbons_by_a_high-valent_nickel_com

Introduction<!>Characterization of Ni 3+<!>†).<!>Oxidation of hydrocarbons<!>Thermodynamic analysis<!>Conclusions<!>Conflicts of interest

<p>Reactions in which protons and electrons move in a single, concerted step (concerted proton-electron transfer, or CPET) play a signicant role in many organic, inorganic, and bioinorganic catalytic systems. CPET at high-valent metal centers has been proposed or observed in the catalytic mechanisms ranging from enzymatic reactions 1,2 to water-oxidation catalysis 3,4 to organic synthesis. 5,6 In many of these systems, the proton and electron are transferred to a high-valent metal-oxo species; however, another strategy involves the transfer of the proton to the ligand scaffold instead. This approach is also relevant to some bioinorganic systems. For instance, in nickel superoxide dismutase (NiSOD), a CPET step is proposed to occur at a highly oxidized nickel(III) intermediate in which a coordinated thiol or amide moiety acts as a proton donor. [7][8][9] Understanding the CPET reactivity of high-valent metal centers may give some insight into their reactivity in catalytic systems. In particular, understanding CPET steps in systems in which the proton is transferred to the ligand may provide insight into catalytic systems which do not or cannot go through metal-oxo intermediates, such as copper-or nickel-containing water-oxidation catalysts or NiSOD.</p><p>In our previous studies, we found that the strongly donating ligand 2-pyridinyl-2-propanoate (pyalk) can stabilize metal centers in high oxidation states, including Ir(V). 10 In addition, the pyalk alkoxide has been suggested to act as a proton shuttle in catalytic water oxidation. 11,12 We reasoned that a high-valent metal compound stabilized by the pyalk ligand was a good candidate for fast CPET. We now describe the preparation of a stable Ni(III) species capable of reacting with a variety of O-H and C-H bonds via CPET.</p><!><p>We prepared 1, a new square-planar nickel(II) complex with two pyalk ligands arranged in a trans orientation (Scheme 1), which was characterized by X-ray crystallography, 1 H NMR spectroscopy, and cyclic voltammetry (CV). The CV of 1 shows a reversible redox feature at 0.15 V vs. Fc/Fc + , suggesting that oxidation generates a stable Ni(III) species (Fig. 1).</p><p>In CH 2 Cl 2 , 1 was treated with [NO][BF 4 ] or [NO] [PF 6 ](E 0 ¼ 0.6 V vs. Fc/Fc + ) (Scheme 1). The solution immediately changed from light green to deep blue. UV-visible spectroscopy indicated the presence of two intense features in the absorption spectrum (l max ¼ 340 nm and 610 nm, Fig. S1 †). This change suggested the formation of an oxidized species. Such intense absorption features in the visible and NIR regions are consistent with the data from other Ni(III) compounds, 13 which led us to suspect that the Ni center had been oxidized. In addition, 1 H NMR of 2 gave a broad paramagnetic spectrum, in contrast to the diamagnetic spectrum of 1 (Fig. S3 †). Measurement of the solution magnetic susceptibility by the Evans method 14 at room temperature gave a m eff of 1.87 for 2, consistent with the presence of one unpaired electron, as expected for a square-planar Ni(III) compound.</p><p>Electron paramagnetic resonance (EPR) spectroscopy showed the presence of a single S ¼ 1 2 species (Fig. 2). A ligand centered oxidation would be expected to show an isotropic signal near g ¼ 2.0. However, the observed rhombic spectrum had g values of g x ¼ 2.077, g y ¼ 2.091, g z ¼ 2.274 (Fig. 2, le). These g values, as well as g ave ¼ 2.145, are consistent with a lowspin square-planar d 7 Ni(III) complex. 15 As demonstrated for a variety of iridium complexes, [16][17][18][19] X-ray photoelectron spectroscopy (XPS) provides another method for identifying a metal-centered oxidation. For metal-centered redox events with no change to the ligand set, the binding energy is expected to increase with oxidation state. Between 1 and 2, the Ni 2p binding energy increases by 1.4 eV, consistent with a metal-centered oxidation from Ni(II) to Ni(III) (Fig. 2, right). Electrospray ionization (ESI) mass spectrometry conrmed the molecular formula of 2. A solution of 2 in CH 2 Cl 2 gave a peak at m/z ¼ 330.09 with the expected isotopic pattern for nickel (Fig. S2a †). A mass spectrum of the parent compound, 1, was also taken, which showed a peak at m/z ¼ 331.09 (Fig. S2b †), corresponding to [1 + H + ]. Despite having the same elemental formula, 1 is uncharged, and thus is observed as the positively charged protonated species, whereas 2 bears a positive charge already, and so is observed without any associated ions. This result conrms that 2 is the one-electron oxidation product of 1. The ESI mass spectrum of 2 does show a small signal at m/z ¼ 331.09, which may indicate that some reduction of 2 occurs during the mass spectral analysis or over time in CH 2 Cl 2 .</p><p>Despite our best efforts, we were unable to obtain a crystal structure of 2. While attempting to use pyridine as a co-solvent for crystallization, however, we observed that the solution changed from deep blue to bright orange (l max ¼ 420 nm, Fig. S1 †). Crystals of this new complex, 3, were successfully grown from CH 2 Cl 2 /pentanes (Fig. 3, le). 3 proved to be an octahedral complex, with two equatorial pyalk ligands and two axial pyridines. 3 contains a PF 6</p><p>À anion, and a comparison of Ni-O bond lengths (Table S3 †) indicates that the alkoxide arms of the pyalk ligands on 3 remain deprotonated, suggesting that 3 remains in the Ni(III) oxidation state. The EPR spectrum of 3 (Fig. 3, right) exhibits an axial signal with g x ¼ 2.202, g y ¼ 2.163, and g z ¼ 2.030. A ve-line hyperne pattern is observed on the g z turning point, indicating coupling of the unpaired electron to the pyridyl nitrogens. 20 We found that the pyridine ligands do not remain bound in solution unless excess pyridine is present; when crystals of 3 were dissolved in CH 2 Cl 2 , 3 immediately reverted to 2, as monitored by UV-visible spectroscopy (Fig. S1 †).</p><p>Upon addition of excess pyridine, the absorbance spectrum of 3 was re-established. Remarkably, 2 and 3 are stable at room temperature, both as solids and in solution; this is rare for Ni(III) compounds, which tend to be stable only between À80 C and À20 C. 13,21,22 Oxidation of phenols With the oxidation state of 2 established, we sought to test its oxidative reactivity. We had shown in previous studies that the pyalk ligand could be reversibly protonated while coordinated to rst row transition metals; 23 therefore, we hypothesized that 2 could undergo proton-coupled electron transfer (PCET) through the reaction proposed in Scheme 2.</p><p>To test 2 for PCET reactivity, 2 was treated with 100 equivalents of 2,4,6-tri-tert-butylphenol (TTBP), a common substrate for PCET reactions. When the colorless solution of TTBP was added to the deep blue solution of 2 in CH 2 Cl 2 , the color changed immediately to bright blue. The appearance of the characteristic peaks at 383 nm and 400 nm in the absorbance spectrum of this solution (Fig. 4) indicated that the tri-tertbutylphenoxyl radical had been formed, causing the color change, 24 and the disappearance of the features at 340 nm and 610 nm indicated that 2 had been consumed. In a similar reaction between 2 and 2,6-di-tert-butyl-phenol (DTBP), the radical coupling product 3,3 0 ,5,5 0 -tetra-tert-butyl-[1,1 0 -bi(cyclohexylidene)]-2,2 0 ,5,5 0 -tetraene-4,4 0 -dione was detected by gas chromatography-mass spectrometry (GC-MS, Fig. S4 †). These results suggested that a PCET process was occurring.</p><p>For reactions of highly oxidized metal species with phenols, proton transfer followed by electron transfer (PT-ET), electron transfer followed by proton transfer (ET-PT), and concerted proton-electron transfer (CPET)/hydrogen atom transfer (HAT) mechanisms have all been observed. [25][26][27] Analysis of kinetic measurements can be used to differentiate between these mechanisms. Therefore, the kinetics of the reaction of 2 with a series of para-substituted 2,6-di-t-butylphenols (4-X-2,6-DTBP) was studied by stopped-ow spectrophotometry in order to further investigate the mechanism of oxidation by 2. For these reactions, excess substrate was used to ensure pseudo-rst order conditions (10-100 equivalents 4-X-2,6-DTBP). Representative UV-visible spectra as a function of time can be found in Fig. 5 and representative time traces can be found in Fig. S5 and S6. † Reactions with all substrates under these conditions t well to a single exponential decay at a single wavelength (l ¼ 610 nm), for which global analysis gave k obs values. Plots of k obs vs. initial substrate concentration displayed a linear dependence (Fig. S7-S13 †), allowing us to determine the second-order rate constant (k 2 ) for each substrate (Table S1</p><!><p>Measured k 2 values were plotted against the Hammett parameter s p + (Fig. 6, le). The s p + constant was chosen rather than s p because s p + is suggested to more accurately represent the electronic structure of the transition state for CPET from phenols 28 and a strong negative correlation between s p + and reaction rate is well known for this class of reactions. 29 Hammett analysis showed that k 2 decreases with the electron-withdrawing capabilities of the para-substituent of the substrate (r ¼ À2. 19). This result is inconsistent with a rate-limiting proton transfer followed by electron transfer (PT-ET) mechanism, which would show a positive linear correlation in the Hammett plot. 30 The observed strong negative association is consistent with the formation of the electron decient phenoxyl radical intermediate. 31,32 These results suggest a CPET mechanism for the reaction of 4-X-2,6-DTBP substrates with 2. Similarly, the plot of log(k 2 ) vs. BDE O-H of the phenol substrates also demonstrated a strong linear correlation (Fig. 6, right). 33 In contrast, the plots of k 2 vs. substrate pK a or E 1/2 showed a much poorer correlation (Fig. S14 †). This linear dependence also suggests a CPET mechanism and is consistent with similar plots constructed for other metal-based oxidants. 21,34,35 It should be noted that this mechanism could also be described as hydrogen atom transfer (HAT), which also operates through a concerted mechanism, and is frequently referred to as such in the literature. 36 H/D kinetic isotope effects (k 2 (H)/k 2 (D), KIE) were also used to differentiate between the possible mechanisms. When the phenolic proton of 2,6-DTBP was replaced with a deuterium atom, the measured k 2 value decreased dramatically, and a KIE of $4 was calculated (Fig. 7). A primary KIE of this magnitude implies that the cleavage of an O-H bond occurs in the ratedetermining step, ruling out pathways involving rate-limiting electron transfer. This value for the KIE is also in good agreement with KIEs reported for the reaction of other metal-based oxidants with 2,6-DTBP that react via CPET. 22,31 It should be noted that pK a and BDE values used in our analysis were measured in dimethylsulfoxide (DMSO) rather than CH 2 Cl 2 , due to the lack of an absolute pK a scale in CH 2 Cl 2 . 37 The general linear trend, however, is expected to remain the same regardless of solvent, as relative BDEs and BDFEs (bond dissociation free energies) do not change signicantly with solvent. To test this assumption, the BDFEs in CH 2 Cl 2 of several phenol substrates were estimated by converting from DMSO values using Abraham's empirical model. [38][39][40] Calculating BDFEs in this manner relies on a number of assumptions, which are discussed in more detail in the ESI (p. S20 †), so the BDFE values should be treated as estimates only. Even so, we still see a strong linear correlation between log(k 2 ) and BDFE CH 2 Cl 2 (Fig. S15 †). In a similar fashion, we also used Ingold's kinetic solvent effect relationship to calculate the expected rate constants in DMSO, 41 which we then plotted against known BDFEs and BDEs measured in DMSO (Fig. S16 †). Once again, we saw a strong linear correlation, further supporting our conclusion of a CPET mechanism.</p><!><p>To further probe its CPET reactivity, 2 was also treated with a number of hydrocarbon substrates. Initial reactions of 2 with 1,4-cyclohexadiene and 9,10-dihydroanthracene produced the expected products of benzene and anthracene, as determined by gas chromatography and 1 H-NMR (Fig. S17a †). To determine the extent of reactivity of 2 with hydrocarbons, 2 was treated with a number of substrates having a range of C-H bond strengths. Fig. 4 UV-visible spectrum of 2 before (blue) and after (red) addition of excess tri-tert-butylphenol. The red trace is consistent with the published UV-visible spectrum of the tri-tert-butylphenoxyl radical. 24 As with phenols, the reactions were monitored by stopped-ow UV-visible spectroscopy or by UV-visible spectroscopy for slowly reacting substrates. All reactions were carried out under pseudo-rst order conditions (10-100 equivalents of substrate). Reactions with all substrates under these conditions gave a good t to a single exponential decay at a single wavelength (l ¼ 610 nm), except for reactions with THF, and global analysis was used to nd k obs values. For reactions with THF, the method of initial rates was used to nd k obs values. Plots of k obs vs. substrate concentration gave good linear ts for all substrates, and the slopes of these ts were used to extract k 2 values for each substrate (Fig. S19-S25 †). When appropriate, the data were statistically corrected to account for the number of hydrogen atoms susceptible to oxidation. For kinetic analysis of multiproton/multi-electron reactions, such as the oxidation of 9,10dihydroanthracene to anthracene (Fig. S17b †), the rst concerted proton-electron transfer step was considered to be the ratedetermining step, since the resulting radical species tend to have signicantly lower BDEs than the parent compounds. 38,42,43 2 was found to react with hydrocarbons with C-H bond strengths that ranged from 77-92 kcal mol À1 . For substrates with low C-H bond strengths, 2 reacted at appreciable rateslog k 2 values of 0.89 and 0.74 were found for reactions of 2 with 1,4-cyclohexadiene and dihydroanthracene (DHA), respectively. A plot of log(k 2 ) vs. C-H BDE showed a strong linear correlation (Fig. 8). In contrast, plots of log(k 2 ) vs. E 0 1/2 or pK a showed a very poor correlation (Fig. S26 †). This result strongly suggests a CPET mechanism for hydrocarbon oxidation.</p><p>To further investigate the mechanism, dihydroanthracene-d 4 was prepared, and an H/D kinetic isotope effect was measured. A large H/D KIE of $11 was observed, indicating the involvement of a proton in the rate-determining step (Fig. S27 †). This result, in combination with the linear correlation between log(k 2 ) and C-H bond strength, further supports our assignment of a CPET mechanism.</p><!><p>We sought to identify the nickel-containing products of the reaction of 2 with phenol and hydrocarbon substrates. Based on our results suggesting that a CPET mechanism was at play, we suspected that the nickel center was being reduced and that the pyalk ligand was accepting a proton and transforming the alkoxide ligand to an alcohol, resulting in the formation [Ni(pyalk)(pyalkH)] + . The 1 H NMR spectrum of reaction products of 2 with 2,6-DTBP, however, showed only the presence of the fully deprotonated 1 in solution. A blue precipitate was also identied. This precipitate was dissolved in water and extracted into CH 2 Cl 2 using NaBAr F (BAr F ¼ [B[3,5-(CF 3 ) 2 C 6 H 3 ] 4 ] À ). 1 H NMR analysis of the resulting product indicated the presence of [Ni(pyalkH) 2 ][2(BAr F )] (4) (Fig. 9a and S28 †). This result suggests that, once the nickel metal center has been reduced, the pyalkH proton is labile enough to rearrange. Attempts to crystallize 4 were unsuccessful; however, when treating Ni(OAc) 2 with the pyalkH ligand, Ni(pyalkH) 2 (OAc) 2 (5) was crystallized, demonstrating the binding of the pyalkH ligand to a nickel center (Fig. 9).</p><p>With the reaction products more fully understood, we constructed the square scheme in Scheme 3 to determine the thermodynamics of the reaction. The pK a for the rst deprotonation of the pyalk ligand of 4 (pK a1 , Fig. S29 †) was estimated spectroscopically to be $25 in MeCN using data from the deprotonation of 4 to 1. Using this value, along with the E 0 1/2 value of 0.15 V vs. Fc/Fc + in MeCN, we were able to determine a BDFE for the bound pyalk O-H bond of $91 kcal mol À1 in MeCN using the square scheme shown in Scheme 3 and the following relationship: 38</p><p>It is generally considered more appropriate to use BDFEs to describe the thermodynamics of PCET by transition metal complexes due to non-negligible entropic contributions. 38 However, many reported high-valent metal-oxo or metalhydroxo oxidants report only the BDE of the O-H bond formed upon the reaction with substrate. Therefore, in order to facilitate a comparison between 2 and reported high-valent CPET/ HAT reagents, the BDE of the pyalk O-H bond was also calculated. The BDE of the pyalk O-H bond for the CPET product of 2 was calculated using same thermochemical parameters described above and the following equation:</p><p>For 2, a BDE of $94 kcal mol À1 in MeCN was calculated. This can be compared with a value of 105 kcal mol À1 for tBuO-H, 44 taken as a model compound for free pyalkH, suggesting a modest O-H bond weakening on binding.</p><p>By cyclic voltammetry, 4 also showed a quasi-reversible redox feature in MeCN at E 0 1/2 ¼ 0.58 V vs. Fc/Fc + (Fig. 10). Using the BDFE calculated from eqn (1) and the square-scheme relationship, we can estimate the value for pK a2 to be $18, 7 pK a units below the calculated pK a1 of 25.</p><p>The reactivity of 2 with 2,6-DTBP and DHA compares favorably with other reported high-valent metal-oxo and metalhydroxo complexes capable of CPET or HAT (Table 1). 2 reacts with DHA at a faster rate than several manganese-and iron-oxo complexes. 2 also compares extremely well with high-valent metal alkoxide and carboxylate compounds, including the only other reported Ni(III) systems, which tend to react with C-H bonds at slower rates than their metal-oxo counterparts. We hypothesize that the particularly strong BDE of the ligand O-H bond formed and the strong oxidizing power of 2 contribute to its high reactivity toward O-H and C-H bonds. This result demonstrates that fast CPET can be achieved in high-valent metal systems without necessarily going through a metal-oxo or metal-hydroxo intermediate, a result relevant to several proposed water-oxidation mechanisms in articial photosynthetic systems of cobalt, 49 copper, 11 and nickel. 50 This result also demonstrates that high-valent metal-alkoxide systems are capable of attacking strong C-H and O-H bonds, which is particularly relevant to water-oxidation catalysis. The pyalk ligand thus proves particularly useful for catalytic oxidations of this type.</p><!><p>We have synthesized and characterized a Ni(III)-alkoxide compound capable of reacting with strong C-H and O-H bonds at appreciable rates. A strong linear dependence between the second-order rate constant, k 2 , and substrate bond dissociation enthalpy indicates a CPET mechanism. Large H/D kinetic isotope effects also support this assignment. We attribute the fast reactivity of 2 to the strong O-H BDE of the pyalk/pyalkH supporting ligand and the high oxidizing power of the complex. This result demonstrates that fast PCET can occur in high-valent metal oxidants without a metal-oxo unit, which may be relevant to certain nickel-containing enzymes or wateroxidation catalysts of cobalt, copper, or nickel. This report also provides the rst full thermodynamic analysis of CPET by a high-valent nickel complex. The value of pyalk as a ligand for catalytic oxidations is further supported.</p><!><p>There are no conicts to declare.</p>

Royal Society of Chemistry (RSC)

New Properties of a Bioinspired Pyridine Benzimidazole Compound as a Novel Differential Staining Agent for Endoplasmic Reticulum and Golgi Apparatus in Fluorescence Live Cell Imaging

In this study, we explored new properties of the bioinspired pyridine benzimidazole compound B2 (2,4-di-tert-butyl-6-(3H-imidazo[4,5-c]pyridine-2-yl)phenol) regarding its potential use as a differential biomarker. For that, we performed 1D 1HNMR (TOCSY), UV-Vis absorption spectra in different organic solvents, voltammetry profile (including a scan-rate study), and TD-DFT calculations that including NBO analyses, to provide valuable information about B2 structure and luminescence. In our study, we found that the B2 structure is highly stable, where the presence of an intramolecular hydrogen bond (IHB) seems to have a crucial role in the stability of luminescence, and its emission can be assigned as fluorescence. In fact, we found that the relatively large Stokes Shift observed for B2 (around 175 nm) may be attributed to the stability of the B2 geometry and the strength of its IHB. On the other hand, we determined that B2 is biocompatible by cytotoxicity experiments in HeLa cells, an epithelial cell line. Furthermore, in cellular assays we found that B2 could be internalized by passive diffusion in absence of artificial permeabilization at short incubation times (15 min to 30 min). Fluorescence microscopy studies confirmed that B2 accumulates in the endoplasmic reticulum (ER) and Golgi apparatus, two organelles involved in the secretory pathway. Finally, we determined that B2 exhibited no noticeable blinking or bleaching after 1 h of continuous exposure. Thus, B2 provides a biocompatible, rapid, simple, and efficient way to fluorescently label particular organelles, producing similar results to that obtained with other well-established but more complex methods.

new_properties_of_a_bioinspired_pyridine_benzimidazole_compound_as_a_novel_differential_staining_age

Introduction<!>Materials and methods<!>Synthesis of 2,4-di-tert-butyl-6-(3h-imidazo[4,5-C]pyridine-2-YL)phenol (B2)<!>Physical measurements<!>1H-NMR and Ftir characterization of B2<!>Computational details<!>Cell culture<!>Cellular staining<!>Fluorescence microscopy<!>B2 cellular uptake<!>Cell viability assays<!>Transformation of chemically competent TOP-10 Escherichia coli<!>Cellular transfection<!>Co-localization assays<!>Statistical analysis<!>B2 characterization<!>Theoretical calculations<!><!>Theoretical calculations<!>Cellular studies<!><!>B2 cellular uptake<!><!>B2 cellular uptake<!>B2 is suitable as cellular biomarker<!><!>B2 accumulates in endoplasmic reticulum and golgi apparatus<!><!>B2 accumulates in endoplasmic reticulum and golgi apparatus<!>Blinking and bleaching B2 properties<!><!>Conclusions<!>Dedication<!>Author contributions<!>Conflict of interest statement

<p>The capability to distinguish and identify different subcellular compartments is critical for understanding organelle function, biogenesis, and cell maintenance, as well as for describing protein sorting and intracellular trafficking pathways (Watson et al., 2005). To this end, cell imaging is becoming a powerful tool to reveal particular biological structures, and even molecular mechanisms, unraveling dynamics and functions of many cellular processes. Accordingly, the development of diverse transmitted light microscopy approaches, including fluorescence microscopy, is increasingly contributing to improve this technique. In this context, the research of new, improved fluorescent indicators clearly constitutes a challenge (Sanderson et al., 2014; Wollman et al., 2015). One of the most desired properties of the fluorophores is their ability to differentially interact with discrete structures in the cell; to this end, the understanding of the chemical properties of these fluorophores is crucial in the design of this kind of molecules. Nowadays, most intracellular compartments can be detected by specific labeling through fluorophores conjugated to particular molecules, such as antibodies, which provide differential binding. For instance, endoplasmic reticulum (ER) and Golgi apparatus, two organelles that work together in the secretory pathway and protein sorting of eukaryotic organisms, which can be stained with fluorophores conjugated to either anti-protein disulfide isomerase (PDI) or anti-58K mouse monoclonal antibody, respectively (Bielaszewska et al., 2013, 2017; Cañas et al., 2016). Nevertheless, the use of antibodies is complex because it requires several steps, remarking the need of alternatives, such as transfection of gene fusions. These gene fusions harbor domains that are differentially sorted inside the cell fused with luminescent domains that allow their identifications via fluorescence microscopy. Although extensively used, transfection is a technique that requires a couple of days to be performed (Kingston et al., 2001). On the other hand, differential dyes have been reported for to specifically stain cellular organelles. In this sense, the organic dye DiOC6(3) (3-hexyl-2-(3-(3-hexyl-2(3H)-benzoxazolylidene)-1-propenyl)-iodide), a lipophilic, cationic, green fluorescent compound, has been used to specifically stain the ER (Sabnis et al., 1997, 2009). In addition, fluorescent analogs of ceramide have proved to be especially valuable for specifically labeling the Golgi apparatus, which receives, processes, and sorts newly synthesized proteins exported from the ER (Cooper, 2000). Although these dyes readily accumulate in the ER and in Golgi apparatus of most cell types by a preferential membrane partitioning process, both stains present biocompatibility problems, thereby they must be functionalized with bovine serum albumin (BSA), prolonging and complicating the staining protocol (Sabnis et al., 1997).</p><p>Recently, we demonstrated that 2,4-di-tert-butyl-6-(3H-imidazo[4,5-c]pyridine-2-yl)phenol (B2), a neutral benzimidazole derivate exhibiting an intramolecular hydrogen bond (IHB). B2 shows luminescent emission at room temperature, with a large Stokes shift (i.e., λex = 335 nm; λem = 510 nm in acetonitrile) (Carreño et al., 2016). Furthermore, B2 has proved to efficiently stain free bacteria (i.e., Salmonella enterica and Escherichia coli), biofilms of Lactobacillus kunkei and L. rhamnosus, and epithelial cell lines (SKOV-3 and HEK-293), as assessed by confocal microscopy (Carreño et al., 2016; Berríos et al., 2017). Interestingly, when epithelial cells were observed, we distinguished a punctuated pattern, apparently located in the cytoplasm. These results strongly suggest that B2 is differentially staining a particular structure/organelle in those cells. Thus, in this work we explored the potential use of B2 as a differential, antibody-free fluorophore in epithelial cells. To this purpose, we performed studies aimed to characterize its optical and electrochemical features, to better understand the role of the intramolecular hydrogen bond in the luminescent properties of B2. To support experimental findings, we also performed computational calculations using DFT theory. In addition, we found that B2 is a biocompatible molecule that generates a punctuate pattern in an epithelial cell line (HeLa). Furthermore, we found that B2 uptake can be detected at short incubation times, apparently by passive a diffusion mechanism. Finally, we found that B2 provides a rapid (30 min), simple (no cell permeabilization is required), biocompatible, and efficient way to fluorescently label both ER and Golgi apparatus, producing similar results to that obtained with other well-established methods. We also provide evidence that B2 is a good candidate to be used as a new, differential, antibody-free fluorophore for organelles belonging to the cell secretory pathway, in time-lapse experiments or short videos with continuous exposure, even at low temperatures.</p><!><p>All chemicals and solvents were purchased from Merck or Aldrich and used without further purification. All solvents were stored over appropriate molecular sieves prior to use.</p><!><p>The general procedure for the synthesis was previously reported, (Carreño et al., 2016) obtaining around 40% yield. Melting point: 311–312°C. FTIR (ATR, cm−1): 2961 (νOH), 2904 and 2868 (νNH), 1626 (νCe = N), 1526 (νC = C). 1HNMR (400 MHz, DMSO-d6, ppm): δ = 1.33 [s; 9H; tBu]; 1.43[s; 9H; tBu]; 7.40 [d, J = 1.9 Hz; 1H; H5], 7.69 [d; J = 4.7 Hz; 1H; H2], 8.02 [s; 1H; H4], 8.36 [d; J = 5.3 Hz; 1H; H1], 8.99 [s; 1H; H3], 13.60[s; 1H; O–H]. UV/VIS: (chloroform, room temperature) λ nm (ε mol−1 dm3 cm−1): 332 (13.32 × 103), 294 (19.09 × 103), 284 (15.51 × 103); (acetonitrile, room temperature) λ nm (ε mol−1 dm3 cm−1): 327 (12.91 × 103), 292 (18.67 × 103), 282 (14.88 × 103); (DMSO, room temperature) λ nm (ε mol−1 dm3 cm−1): 332 (11.30 × 103), 294 (15.44 × 103), 284 (12.23 × 103). Rf: 0.38 (ethylacetate as solvent).</p><!><p>NMR spectra were recorded on a Bruker AVANCE 400 spectrometer operating at 400 MHz, at 25°C. Samples were dissolved in deuterated dimethyl sulfoxide (DMSO-d6), using tetramethylsilane as internal standard. FTIR techniques were recorded in an UATR spectrum Two Perkin Elmer spectrophotometer.</p><p>Purity of B2 was checked by TLC using glass plates pre-coated with SiliaPlate TLC Aluminum foil TLC were supplied by Silicycle as stationary phase, and a suitable solvent system was used as mobile phase (ethyl acetate). Spots were visualized with short wave ultraviolet light (λ = 254 nm) using Spectroline LongLife TM Filter. Melting points were determined on a Stuart 10 Scientific melting point apparatus SMP3 (UK) in open capillary tubes.</p><p>For electrochemical experiments, a working solution containing 0.01 mol/L of the respective compound together with 0.1 mol/L tetrabutylammonium hexafluorophosphate (TBAPF6) as supporting electrolyte in CH3CN, was used. Prior to each experiment, the working solution was purged with high purity argon, and an argon atmosphere was maintained during the whole experiment. A polycrystalline non-annealed platinum disc (2 mm diameter) was used as working electrode. A platinum gauze of large geometrical area, separated from the cell main compartment by a fine sintered glass, was used as counter electrode. All potentials quoted in this work were referred to an Ag/AgCl electrode in tetramethylammonium chloride to match the potential of a saturated calomel electrode (SCE), at room temperature. All electrochemical experiments were performed at room temperature on a CHI900B bipotentiostat interfaced to a PC running the CHI 9.12 software that allowed experimental control and data acquisition.</p><!><p>1H-NMR and FTIR techniques were performed as previously described for B2 (Carreño et al., 2016). These techniques were used to confirm the correct synthesis of B2.</p><!><p>All structural and electronic properties were obtained using the Amsterdam Density Functional (ADF) code (Te Velde et al., 2001). All molecular structures were fully optimized by an analytical energy gradient method as implemented by Verluis and Ziegler (Echeverria et al., 2009; Ramírez-Tagle et al., 2010; Alvarado-Soto and Ramirez-Tagle, 2015; Bjorgaard et al., 2015), using the hybrid B3LYP functional and the standard Slater-type-orbital (STO) basis set with triple-ζ quality double plus polarization functions (TZ2P) for all the atoms (Rabanal-Leon et al., 2014; Zhang et al., 2016). Frequency analyses were performed after the geometry optimization to corroborate the minimum and to compare with experimental infrared spectra. Natural bond orbital (NBO) analysis was used to characterize energies of the IHB (Avilés-Moreno et al., 2017; Guajardo Maturana et al., 2017). Time-dependent density functional theory (TDDFT) (Ghane et al., 2012; Fuks et al., 2013; Mosquera and Wasserman, 2015), used at the same level of theory to calculate the excitation energies using in all cases the conductor-like screening model for realistic solvents (COSMO) (Sinnecker et al., 2006; Simpson et al., 2015; Yamin et al., 2016), DMSO to estimate the hydrogen bond stability and to visualize the conformational changes due to the solvent polarity, additionally the calculations were also performed in the gas phase (Tsolakidis and Kaxiras, 2005; Quartarolo and Russo, 2011).</p><!><p>The HeLa cell line (ATCC® CCL-2™) (cervical adenocarcinoma) was grown in 25 cm2 polystyrene bottles in Dulbecco's High Glucose Modified Eagle Medium (DMEM) supplemented with 10% v/v fetal serum bovine (FBS), 1 mM sodium pyruvate and 1% v/v penicillin-streptomycin. Cells were incubated at 37°C and 5% CO2, changing the culture medium every 2–3 days, and propagated when they reached between 80 and 90% confluence.</p><!><p>HeLa cells were seeded in a 24-well culture dish (3 × 105 cells per well) in which a 12-mm diameter coverslip was previously added, and allowed to acclimate for 24 h. Each well was then washed 3 times with sterile 1 × PBS and then the different concentrations of B2 (200, 100, 50, 25, or 12.5 μg/mL) and DMSO vehicle (50, 25, 12.5, 6.125, and 3.0625%) were added to each well and incubated for 15 and 30 min at 37°C with 5% CO2. It is important to underline that permeabilization procedures are not necessary. Subsequently, each well was washed 3 times with sterile 1 × PBS and coverslips were deposited on the slides using 5 μl of Fluoromount® mounting medium. Each slide was left to dry in the dark at room temperature overnight and were sealed with acrylic paint.</p><!><p>The analysis was performed on a Model BX61 Fluorescence Microscope (Olympus Corp., Tokyo, Japan), Spinning Disk Olympus (DSU) system, coupled to an ORCA-R2 camera (Hamamatsu Photonics KK, Japan) and CellSens Dimension software v1.9 (Olympus Corp., Tokyo, Japan). Fluorescence emission was obtained by excitation with a xenon lamp. Emission was collected with a DAPI filter (450 to 500 nm).</p><!><p>HeLa cells (2 × 105 cells/well) were seeded in 24-well plates and incubated for 24 h. To evaluate the cellular uptake of compound B2, concentrations of 25 μg/mL and 50 μg/mL were prepared in culture medium, added to each well and incubated for 15 and 30 min. In addition, to evaluate whether the uptake of B2 depends on energy, cells were incubated with the same treatments at 4°C, a temperature where no endocytic processes occur (Mukherjee et al., 1997).</p><p>After the incubation time, cells were washed 3 times with 1 × PBS and peeled off the culture plate using a 0.2% w/v PBS-EDTA solution and incubated for 20 min at RT. Suspended cells were washed twice with FACS buffer (2% PBS supplemented with fetal bovine serum) and resuspended in 500 μl PBS. Finally, the fluorescence intensity of the cells was quantified through a BD Accuri™ FACSARIA II flow cytometer using the FlowJo 7.6.1 software.</p><!><p>HeLa cells were cultured in Dulbecco's Modified Eagle's Medium (DMEM) containing 10% fetal bovine serum (FBS), 2 mM L-glutamine, 100 units/mL penicillin and 100 μg/mL streptomycin. Cells were maintained in 75 cm2 flasks in a 5% CO2-humidified atmosphere at 37°C. Passage took place every 2–3 days. All cell culture supplies were purchased from Sigma-Aldrich. Toxicity was determined using the 3-(4,5-dimethylthiazol-2-yl)-2,5- diphenyltetrazolium bromide (MTT) cell viability assay after 15 min, 30 min, 1 h, and 24 h of incubation with B2. MTT is a yellow compound that, when reduced by active mitochondria, produces purple formazan crystals that can be measured spectrophotometrically (Low et al., 2016; Sheikh et al., 2016). For this purpose, MTT (Sigma-Aldrich) was dissolved in phosphate buffered saline (PBS) to a concentration of 5 mg/mL and further diluted in culture medium (1:11). Cells were incubated with this MTT-solution for 4 h under normal culture conditions. Afterwards, 100 μL of isopropyl alcohol were added. To completely dissolve the formazan salts, plates were incubated for 10 min on a shaker and quantified by measuring absorbance at 570 nm with an ELISA microplate reader. Cell viability was calculated as percentage of surviving cells compared to untreated control cells.</p><!><p>For propagation of the desired vector, it was introduced into the chemically competent E. coli TOP10 (ThermoFisher)1 (generated according to the manufacturer's instructions). The transformation was performed using 100 μL of the chemo-competent bacteria and 1 μg of circular vector (KDEL-GFP or sialyl transferase signal anchor sequence-RFP). The mixture was incubated on ice for 30 min and a thermal shock was quickly performed at 42°C for 2 min, immediately after, was put on ice for 3 min and 900 μL of LB broth was added and incubated for 1 h at 37°C. Finally, the bacteria were plated on LB agar supplemented with 50 μg/mL kanamycin for the selection of the transforming colonies.</p><!><p>HeLa cells were seeded in a 24-well culture dish (5 × 105 cells per well) in which a glass cover was previously added to each well and allowed to acclimate for 24 h. Subsequently, the entire culture medium was removed, and each well was washed 3 times with sterile 1 × PBS. The different combinations of DNA and Lipofectamine 3000® Transfection Reagent were performed according to the manufacturer's instructions. After all components were added to the cell culture (Lipofectamine 3000® Transfection Reagent, DNA and DMEM medium), it was incubated for 6 h at 37°C with 5% CO2. The content of each well was removed and washed 3 times with 1 × sterile PBS. Then, 500 μL of complete DMEM medium was added to each well and incubated until 48 h of treatment were completed. Finally, cell transfection was checked using an inverted-light microscope and a BX-53 epifluorescence microscope.</p><!><p>HeLa cells were seeded in a 24-well culture dish with 5 × 105 cells per well in which a 12-mm diameter coverslip was previously added and allowed to acclimate for 24 h. Cells were then transfected with the KDEL-GFP and Sialyl-RFP plasmids, which express specific fluorescent peptide of the endoplasmic reticulum and Golgi apparatus, respectively. After the transfection process was completed, each well was washed 3 times with sterile 1 × PBS and then the different concentrations of B2 (25 μg/mL and 50 μg/mL) and the percentages of vehicle DMSO (12.5 and 6.125%) were added for 15 min with 5% CO2 and at 37°C. Each well was then washed 3 times with sterile 1 × PBS and fixed for 5 min with 4% PFA in PBS (4g PFA, 1 M CaCl2, 1 M MgCl2, pH 7.4 adjusted solution), washed 3 times with PBS and each coverslip was mounted on slides using 5 μl of Electron Microscopy Sciences. Each slide was allowed to dry in the dark at room temperature overnight and then sealed with acrylic paint. Finally, all samples were observed using the Olympus BX-61-DSU epifluorescence microscope.</p><!><p>All values of analyzed data are presented as mean standard error (SE) from three biological replicates. Statistical analysis included was one-way ANOVA followed by multiple comparison test (Tukey). Differences among groups were considered statistically significant when p < 0.05.</p><!><p>B2 (Figure S1, see Table S1 for characteristic constants) is insoluble in water, but presents low solubility in chloroform, acetonitrile and methanol, and a good solubility in DMSO at room temperature. B2 synthesis was confirmed by their FTIR (Figures S2, S3), 1H-NMR spectra (for proton numbering, see Figure S4; for 1H-NMR see Figure S5) including TOCSY experiments (Figure S6–S8).</p><p>Since B2 is being tested as a new fluorophore for biological applications, its use in different solvents may be also desirable. In this context, electronic absorption spectra of B2 were measured in different organic solvents: chloroform, acetonitrile, and DMSO, at room temperature. We observed three intense absorption bands. The first two high-intensity absorption can be assigned to n → π* (–C = N–) and π → π* transitions, respectively. No significant shifts were observed for B2 in the different solvents used (see Table S2), suggesting that the IHB in B2 is stable under all the tested conditions, even in presence of DMSO, which can form hydrogen bonds with the solute. Any change in the B2 structure (including dissociation of the IHB), produced by interaction with the solvent, would lead to changes in the absorption spectra. This was not observed for B2 under the tested conditions (see Figure S9 for UV-Vis spectrum).</p><p>To complement the previously reported electrochemical characterization of B2 (Carreño et al., 2016), a scan-rate study was performed at 50, 200, and 400 mVs−1. We found that B2 exhibited a single reversible reduction process [Red(rev)I] at −0.84 V, and two irreversible oxidations, Ox(irr)I and Ox(irr)II, at 0.96 and 1.47 V vs. SCE (saturated calomel electrode), respectively, consistent to previously reported data (Carreño et al., 2016). As inferred from the Figure S10, the control mechanism for all the studied processes depends on the species diffusion from the bulk solution (see Table S3). Diffusional control of red-ox processes has been reported for similar benzimidazoles and other bioinspired compounds (Savarino et al., 1997; Boiani et al., 2006; Moore et al., 2008; Manbeck et al., 2016), but the reduction in those cases has been found to be irreversible. The difference is that B2 possesses an IHB that has been attributed to stabilize the radical form, explaining the reversibility of reduction and emphasizing the importance of the IHB in the B2 features (Benisvy et al., 2006; Moore et al., 2010).</p><!><p>As stated, B2 is a luminescent compound (Carreño et al., 2016). To better characterize this phenomenon, we performed time-dependent density functional theory (TDDFT) calculations to assign the electronic transitions. The first step was to optimize the ground (S0) and the first excited singlet state (S1); in a second step, absorption and emission bands were calculated using TDDFT. All the calculations where performed using COSMO model for solvent with the parameters of DMSO (Liu et al., 2011). We found that the geometry of S0 and S1 exhibited no significant differences, showing, in both cases, that the structure must remain planar, most likely due to the presence of the IHB (see Table S4). To further study the B2 IHB, we evaluated the second-order interaction energy by calculating natural bond orbitals (NBO). We found that the IHB energy is 6.23 kcal/mol for S0, and 6.13 kcal/mol for S1. These values are in agreement with the reported values in similar compounds harboring IHB (Muhammad et al., 2010; Abdel Ghani and Mansour, 2012; Monajjemi, 2012; Carreño et al., 2016; Yankova and Radev, 2016), and support the stability of this interaction (Sosa et al., 2002). In this sense, the experimental UV-vis results (see Table S2) were corroborated by computational methods (see and Figure S9). To further characterize the UV-vis observed transitions, TDDFT calculations were conducted (see Table S5). This calculated transition is composed of a HOMO-2 → LUMO+1 (n → π*), HOMO-1 → LUMO (π → π*), and HOMO → LUMO (π → π*). The band located experimentally at 332 nm (DMSO), theoretically calculated at 339 nm, corresponded to a HOMO → LUMO transition. Both the HOMO and LUMO composition involves the IHB (Figure 1). The isosurfaces provide some suggestions regarding experimental results obtained from UV-Vis studies (see Table S2 and Figure S9). All these results together demonstrate the stability of the IHB in B2.</p><!><p>Molecular orbitals involved in the absorption and emission for B2. S0 corresponds to the ground state. S1 corresponds to the first excited singlet state.</p><!><p>On the other hand, the emission band was calculated using the geometry of the first excited state S1 to perform a TDDFT calculation. The emission was calculated as a π* → π between LUMO and HOMO orbitals at 515 nm, in good agreement with the experimental value reported around 500 nm (data not shown) (Carreño et al., 2016). As stated above, HOMO and LUMO composition involves the IHB, reinforcing the contribution of this interaction in the stability of B2. The rigidity of the B2 structure substantially reduces the vibronic relaxation, explaining the luminescence of this compound, as reported for other molecules (Gopal et al., 1995). Regarding the luminescence of B2, the geometry of the first excited triplet was also calculated and used for a TDDFT calculation using a previously described protocol (Carreño et al., 2017). However the contribution of the triplet to the emission band was less than 0.5%, indicating that the emission can be assigned to fluorescence, as previously suggested (Carreño et al., 2016). Altogether, the combined information obtained by the analyses described above, provides valuable information about B2 structure, and its luminescence.</p><!><p>To further characterize B2 in cellular assays, we stained HeLa cells (epithelial cell line) with B2 (200, 100, 50, 25, or 12.5 μg/mL) for 30 min, prior to fluorescent microscopy. Beforehand, we characterized the B2 staining properties using a confocal microscope, with laser excitation at 405 nm, and emission collected with a long-pass filter in the range of 425 to 525 nm (Carreño et al., 2016). Although confocal microscopy exhibits several advantages, such as the possibility to examine samples through the Z axis, it needs special requirements. For that reason, this time we used a epifluorescence microscope, using a xenon lamp (excitation) and DAPI filter (emission, 450 to 500 nm) (Atale et al., 2014). We found that B2 showed a suitable fluorescence inside cells, were the optimal concentration ranged between 25 and 50 μg/mL of B2 (Figure 2). When 12.5 μg/mL B2 was used, we were unable to observe fluorescence under the tested conditions (data not shown). Interestingly, we observed that B2 produced a punctuate pattern inside cells (Figure 2), strongly suggesting that B2 is not a general, but a differential stain. In addition, it is important to remark that, albeit 30 min of incubation with B2 are optimal, 15 min are sufficient to visualize cells under the fluorescence microscope (data not shown), highlighting the potential of this compound as a relatively quick biomarker, even using common equipment, such as a xenon or mercury lamp, and a DAPI filter. Moreover, it is clear from these experiments that chemical derivatizations are not required for B2 to be used as biomarker. Finally, the B2 staining protocol can be performed without the need of additional permeabilization steps, underlining its simplicity. In this sense, DMSO has been known to enhance cell membrane permeability of drugs or DNA. Studies exploring DMSO applicability to promote plasma membrane permeability using different molecules in living cells, demonstrated that DMSO can increase cell permeabilization, in a both concentration- and time-depending manner (de Ménorval et al., 2012). It has been reported that 10% v/v of DMSO slightly increase cell permeability after 1 h of incubation, but its impact in the entry of polar molecules is marginal, as assessed by the absence of swelling and the limited amount of water that could cross the plasma membrane. Although the presence of small undulations in the plasma membrane were reported in eukaryotic cells treated with 10% DMSO, they were only visible after 1 h in the presence of DMSO (de Ménorval et al., 2012). Considering that be used 25 or 50 μg/ml of B2 to stain cells (involving the presence of 6.25 and 12.5% v/v DMSO, respectively), and shorter times of incubation (i.e., 15 or 30 min), we speculate that the DMSO could contribute to the entry of B2 into epithelial cells, with minimal effects on cell morphology. Nevertheless, the potential impact of the DMSO in cellular compartments, at the concentrations and incubation times proposed in this study for B2 staining protocol, must be explored in future analyses.</p><!><p>Labeling pattern of B2. HeLa cells were seeded on coverslips and incubated with 50 μg/mL B2 for 30 min at 37°C prior being fixed and analyzed by fluorescence microscopy. Bar size 50 μm (A) and 10 μm (B). We found similar results with a shorter (15 min at 37°C) incubation and/or with 25 μg/mL B2; albeit not staining was observed with 12.5 μg/mL B2 (data not shown).</p><!><p>To better understand the B2 potential as biomarker, we characterized its cellular uptake in epithelial cells. For that, we performed flow cytometry assays of cells stained under three different conditions: (1) Two B2 concentrations that allow cell staining (25 and 50 μg/mL); (2) two incubation times with B2: 15 min, suboptimal for staining, and 30 min, optimal for staining; and (3) two temperatures (4 and 37°C). We determined its cellular uptake under the last condition because it is particularly valuable to assess the entry mechanism, since active endocytic processes are inhibited at 4°C (Mukherjee et al., 1997). As shown in Figure 3, B2 cellular uptake is dependent on both incubation time and concentration. Although B2 uptake significantly decreases at 4°C compared to the uptake observed at 37°C, cells still showed considerable B2-dependent luminescence at 4°C, especially when cells were incubated for 30 min. In fact, the decreased uptake at 4°C can be explained by the diminished plasma membrane fluidity, since at low temperatures the diffusion rate is impaired (Cooper and Sunderland, 2000). Since active uptake is inhibited at 4°C, we infer that B2 is internalized by cells through passive diffusion. This property explains why B2 staining protocol does not require a permeabilization step, providing a simplified staining method. Other authors reported Ir (III)-based (d6) compounds as potential biomarkers that were unable to enter cells at 4°C, even after prolonged incubation periods (more than 2 h) (Yin Zhang et al., 2010; Zhang et al., 2010). This fact remarks the B2 advantages in comparison with other fluorophores, even d6 complexes, regarding cell labeling. Thus, some of the advantages exhibited by B2 include its use at low temperatures and low incubation times, valuable properties if considered that not additional permeabilization steps are required.</p><!><p>Quantification of B2 uptake into HeLa cells. HeLa cells were incubated with different concentrations of B2 (25 or 50 μg/mL), at 4°C or 37°C, for 15 min (A) or 30 min (B). The mean fluorescence intensity (MFI) was determined by flow cytometry using a 350-nm excitation laser and a 520-nm detector. The statistical difference is based on the negative control (culture medium alone). A 2-way ANOVA was performed with Tukey's posttest as a statistical analysis. Only relevant differences are depicted in the figure. **p < 0.01, ***p < 0.001, ****p < 0.0001. n = 3 (biological triplicate).</p><!><p>In general, factors determining the entry of fluorophores into cells include size, charge, and hydrophobicity (Juris et al., 1988; Stufkens, 1998; Puckett and Barton, 2007; Yin Zhang et al., 2010; Zhao et al., 2011; Gill and Thomas, 2012). Thus, B2 might permeate inside cells due to its own chemical nature, by the contribution of the DMSO as an incorporated permeabilizer agent in B2 solutions, or by a combination of this features. Nevertheless, since the DMSO contributes to a slight plasma membrane permeabilization only after 1 h of incubation at the concentrations used for staining (de Ménorval et al., 2012), we cannot rule out that B2, itself, is able to penetrate cells by passive transport.</p><p>All these properties must be considered to design efficient chemical compounds as cellular biomarkers. Another desirable property for a biomarker is the differential staining. Differential staining is the ability to specifically stain a particular cell structure (e.g., organelles such as endoplasmic reticulum or Golgi apparatus). Normally, biomarkers can be modified to be used as differential dye through conjugation with antibodies, relatively big and complex molecules that usually preclude cell uptake; thereby, additional permeabilization steps are required. By contrast, B2 provides a simpler method to achieve differential staining without the need of antibodies, as we will discuss below.</p><!><p>A suitable fluorophore for live cell imaging should exhibit three main properties: good and stable luminescence, efficient cellular uptake, and low cytotoxicity (Haas and Franz, 2009). In this context, we evaluated B2 cytotoxicity in epithelial cells by MTT assays. MTT is a yellow compound that, when reduced by functioning mitochondria, produces purple formazan crystals that can be measured spectrophotometrically (Low et al., 2016; Sheikh et al., 2016). We tested different concentrations of B2 (25, 50, and 100 μg/mL) and different incubation times (15 min, 30 min, 1 h, and 24 h). It is important to remark that the optimal staining protocol requires 50 μg/mL of B2, and 30 min of incubation (Figure 2). Figures 4A,B show that, at the staining conditions, B2 presented low cytotoxicity (around 10%) compared with the vehicle alone (i.e., DMSO), values that represent no toxicity for cellular models (Nel et al., 2009). By contrast, more prolonged incubation times, or higher concentrations, produced more pronounced effects (Figures 4C,D). In these cases, cytotoxicity increased by approximately 33% at 1 h incubation time using 25 μg/mL, and at 52% using 50 μg/mL, which can be due to an excessive accumulation of B2 in organelles and to the presence of DMSO (vehicle). DMSO increases cyclic adenosine monophosphate (cAMP), a second messenger involved in cell death by apoptosis (Cho et al., 2014). These results indicate that B2 can be used in living systems as biomarker at relatively short incubation times (15 to 30 min), whereas more prolonged times (i.e., >1 h) are not recommended due to higher cytotoxicity levels, not biocompatible for these kind of studies (i.e., >20%) (Nel et al., 2009). These results show that B2 exhibited low cytotoxicity under the staining conditions proposed in this work, proving its biocompatibility as biomarker.</p><!><p>Quantification of B2 cytotoxicity in HeLa cells. Cytotoxicity assay was performed using the MTT Cell Growth Kit. HeLa cells were incubated with B2 (25 μg/mL or 50 μg/mL) for 15 (A), 30 min (B), 1 (C), and 24 h (D) at 37°C. Percentages are corrected based on the viability control (culture medium alone). The significant difference is based on the negative control (DMSO 100%). A 2-way ANOVA was performed with Tukey's posttest as a statistical analysis. Only relevant differences are depicted in the figure. *p < 0.05, ****p < 0.0001. n = 3 (biological triplicate).</p><!><p>As shown in Figure 2, B2 presented a punctuate staining pattern, with a central unstained area plausibly corresponding to the cell nucleus. This strongly suggests that B2 is accumulated in a discrete, particular cell structure (e.g., an organelle). It has been reported that endoplasmic reticulum (ER) exhibit a similar staining pattern (Li et al., 2016). The ER and Golgi apparatus are two organelles that work together in the secretory pathway of eukaryotic proteins (Hanada, 2017). Thus, we proposed that B2 is being accumulated in these organelles. To test this hypothesis, we performed co-localization experiments using known intracellular fluorescent markers normally used as reference: KDEL-GFP [lysine/aspartate/glutamate/leucine–green fluorescent protein (emission: 503–508 nm)], to stain the ER; and Sialyl-RFP [Sialyl transferase signal anchor sequence—red fluorescent protein (583 nm)], to stain the Golgi apparatus (Tsien, 1998; Dayel et al., 1999; Remington, 2002; Hawes and Satiat-Jeunemaitre, 2005). These reference biomarkers, i.e., KDEL-GFP and Sialyl-RFP, are recombinant fluorescent proteins that accumulates in secretory organelles. As described above, the staining protocol of B2 is simple, consisting mainly in short incubation times (30 min). Unlike B2, KDEL-GFP and Sialyl-RFP need to be expressed directly by the cells, thereby a transfection protocol must be performed. Transfection consists in introducing purified nucleic acids, normally produced in bacteria, into eukaryotic cells to express heterologous proteins, such as KDEL-GFP or sialyl-RFP. Complete transfection protocol can take two or more days.</p><p>As shown in Figure 5, B2 accumulates in subcellular compartments from the secretory pathways (i.e., ER and Golgi apparatus), revealing these organelles as efficiently as KDEL-GFP or Sialyl-RFP, but with a simpler method. B2 accumulation in the ER and in the Golgi apparatus is probably due to its affinity for certain proteins inside these organelles, probably glycoproteins. Zhang et al. demonstrated that two fluorophores based on Ir (III) present high affinity for the Golgi apparatus (Zhang et al., 2010). These Ir (III)-based compounds possess many pyridine groups, which are also present in the B2 structure. Nevertheless, comparing B2 and Ir (III)-based fluorophores, B2 is a simpler molecule exhibiting better quantum yields (φ = 0.21 in acetonitrile) (Carreño et al., 2016), whereas Ir (III)-based fluorophores are complex dendritic cyclometalated compounds exhibiting lower quantum yields (ϕ = 0.036 to 0.14 in acetonitrile, depending on the compound). Most importantly, visualization of Golgi apparatus in HeLa cells requires 2 h of incubation with 2 M of Ir (III) complexes to obtain similar results (Zhang et al., 2010) to those shown in the Figure 5, which needed lower incubation time at lower concentration (155 μM) of B2. Furthermore, although other fluorophores have been reported to stain the ER, such as the ER-tracker (C44H42BClF2N6O7S2), (ThermoFisher1; Diwu et al., 1997) their considerable size mandatorily requires cell permeabilization, a step that is not necessary in the case of the B2 staining protocol.</p><!><p>B2 staining co-localizes with Endoplasmic Reticulum (ER) (A,B) and Golgi apparatus (C), (D). Fluorescence assay was performed using HeLa cells and analyzed under fluorescent microscopy at 48 h post-transfection either with KDEL-GFP gene (A,B) or with Sialyl-RFP (C,D). To stain with B2, respective transfected cells were simply incubated with B2 at 50 μg/mL for 30 min at 37°C. In the case of B2, a pseudo color was used to facilitate its visualization and co localization. White bars represent 10 μm.</p><!><p>On the other hand, it has been reported that DiOC6(3) is able to stain the ER at 10 μg/mL. Nevertheless, at lower concentrations (i.e., 0.1 μg/mL), DiOC6(3) stains mitochondria instead, letting unclear which are the limit concentrations to stain one organelle or another (Sabnis et al., 2009). By contrast, B2 stain ER and Golgi at 50 μg/mL (or more), whereas lower concentrations appear to be unable to stain (as mentioned above), allowing a better identification.</p><p>Thus, unlike other fluorophores which only have cytoplasmic or perinuclear distribution (Puckett and Barton, 2007; Yin Zhang et al., 2010; Zhang et al., 2010), B2 provides a simple and efficient way to differentially stain ER and Golgi apparatus in epithelial cells, showing that B2 can be considered a differential fluorescent dye.</p><!><p>Some fluorophores exhibit unpredictable blinking properties, a clear drawback to obtain high quality images (Michalet et al., 2005; Mahler et al., 2008). On the other hand, photobleaching is also an undesired property of fluorophores when time-lapse experiments or videos under continuous exposure are required (Li et al., 2017; Elisa et al., 2018). For instance, DiOC6(3) exhibits photobleaching similar to that of the rhodamine (Sabnis et al., 2009). To test whether B2 exhibits photobleaching and/or blinking in biological applications, we stained HeLa cells with 50 μg/mL B2 for 30 min at 37°C. Then, stained cells were observed by fluorescence microscopy during 1 h with continuous exposure. As shown in Figure 6, B2 is resistant to photobleaching, and blinking was not observed (see Supplementary Video 1). This phenomenon can be explained by the high stability of B2, as demonstrated above. The IHB, that is stable in different organic solvents (Table S2), contributes to keep the rigidity between the benzimidazole and phenolic ring moieties, substantially reducing the vibronic relaxation due to a coplanar geometry between the rings, and contributing to the fluorescence by minimizing the non-radiative emission. All this evidence suggests that, at least in part, the presence of a stable IHB contributes to strongly decrease photobleaching of B2.</p><!><p>B2 exhibits resistance to photobleaching. HeLa cells were stained with 50 μg/mL B2 for 30 min at 37°C and observed under the fluorescent microscope immediately (A), or 5 min (B), 10 min (C), 15 min (D), 30 min (E), 60 min (F) after the staining protocol in continuous exposure to the exciting light. We obtained similar results with MCF-7 cells (data not shown). Blinking was not observed (see Supplementary Video 1). White bar represents 10 μm.</p><!><p>In this work, we explored new features of the luminescent compound B2 concerning some of its chemical properties, and its use as biomarker for specific cell organelles. We found that B2 exhibits a very stable structure, a feature that in turn contributes to a photobleaching-resistant fluorescence. In addition, B2 provides a rapid (30 min for optimal staining), simple (since no cell permeabilization is required), biocompatible (low cytotoxicity under staining conditions), and efficient way to fluorescently label both ER and Golgi apparatus, producing similar results to that obtained with other well-established methods, but without photobleaching or blinking. Altogether, our results show that B2 is suitable to be used for differential labeling of ER and Golgi apparatus, in time-lapse experiments or short videos with continuous exposure, even at low temperatures.</p><!><p>Dedicated to Professor Juan Manuel Manríquez on the occasion of his retirement.</p><!><p>FL: Cytotoxicity and uptake experiment, discussion of biological experiment and paper writing; JF: Discussion of all the experiments, and paper writing; AC: Synthesis, characterization, discussion of all the experiments, and paper writing; CZ: Optical and other chemical characterizations; DP-H: Theoretical calculations; MG: Electrochemical studies; RP: Discussion of Biological experiments; MP: Discussion of NMR, TOCSY and FTIR experiments; RA-P: Discussion of theoretical calculations; CO: Microscopy and discussion of Biological experiments.</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

Enabling technologies and green processes in cyclodextrin chemistry

The design of efficient synthetic green strategies for the selective modification of cyclodextrins (CDs) is still a challenging task. Outstanding results have been achieved in recent years by means of so-called enabling technologies, such as microwaves, ultrasound and ball mills, that have become irreplaceable tools in the synthesis of CD derivatives. Several examples of sonochemical selective modification of native α-, β- and γ-CDs have been reported including heterogeneous phase Pd- and Cu-catalysed hydrogenations and couplings. Microwave irradiation has emerged as the technique of choice for the production of highly substituted CD derivatives, CD grafted materials and polymers. Mechanochemical methods have successfully furnished greener, solvent-free syntheses and efficient complexation, while flow microreactors may well improve the repeatability and optimization of critical synthetic protocols.

enabling_technologies_and_green_processes_in_cyclodextrin_chemistry

<!>Review<!><!>Review<!><!>Review<!><!>Review<!><!>Review<!><!>Review<!>Ultrasound<!>Monosubstituted CD derivative preparation<!><!>Monosubstituted CD derivative preparation<!><!>Monosubstituted CD derivative preparation<!><!>Monosubstituted CD derivative preparation<!><!>Monosubstituted CD derivative preparation<!>Preparation of persubstituted CD derivatives<!>Preparation of second generation CD derivatives: dimers, and CD hybrids<!><!>Preparation of second generation CD derivatives: dimers, and CD hybrids<!>Preparation of CD-grafted materials and CD-based polymers<!><!>Preparation of CD-grafted materials and CD-based polymers<!><!>Preparation of CD-grafted materials and CD-based polymers<!>Microwaves<!>Preparation of monosubstituted CD derivatives<!><!>Preparation of monosubstituted CD derivatives<!><!>Preparation of persubstituted CD derivatives<!><!>Preparation of persubstituted CD derivatives<!>MW-promoted Cu-catalyzed click reaction for the preparation of second generation CD derivatives and hybrid structures<!><!>Preparation of CD-grafted materials and CD-based polymers<!><!>Preparation of CD-grafted materials and CD-based polymers<!>Ball mill<!><!>Ball mill<!><!>Ball mill<!>Microreactors<!>Conclusion

<p>This article is part of the Thematic Series "Superstructures with cyclodextrins: Chemistry and applications III".</p><!><p>The last decade has witnessed the development of highly efficient alternative synthetic methods which make use of new enabling technologies. The need for a more rational approach to the synthesis of cyclodextrin (CD) derivatives has led to several energy sources been tested for their ability to activate C–C and C–X bond formation. In recent years non-conventional energy sources, such as microwaves (MW), ultrasound (US), ball mills (BM) and microreactors have made access to CD derivatives much simpler, as have heterogeneous catalysts and greener solvents. Besides batch reactors, in the last decade these techniques have been adapted to flow systems, which provide greater efficiency, flexibility and lower energy consumption, or in high-throughput applications. Our experience in process intensification and innovative reactors took advantage from flow-multihorn US systems (Figure 1a) and cavitational turbines (Figure 1b) to optimize mass transfer via intense cavitation [1–2]. Similarly, we have accumulated experiences with mechanochemical conditions that open the way to solventless reactions even on a pilot scale (Figure 1c) [3]. The latest generation of dedicated MW reactors, which enable operators to quickly screen reaction conditions by means of parallel tests across a wide range of operative conditions, has provided outstanding MW-assisted synthesis results (Figure 1d) [4]. While most researchers will most likely be acquainted with the potential of dielectric heating, the specific conditions needed to let react CDs efficiently and selectively are often overlooked.</p><!><p>(a) Multihorn-flow US reactor, (b) Cavitational turbine, (c) Pilot-scale BM, (d) High-pressure MW reactor.</p><!><p>The current trends on CDs' literature and their application in green protocols are clearly depicted in Figure 2. The present literature survey with identical keyword combinations has been done in two major databases [5–6]. The results were partially overlapped only in the full text searches and approximately 4000 records have been found. Further reduction of records, less than 2500, by searching in Title/Abstract/Keyword fields only resulted in more relevant publications. Only 10% roughly of the recently published papers on CDs are dealing with sustainable technologies and only few works are comparing data with conventional synthetic protocols.</p><!><p>Trends in CD papers and CD use in green chemical processes.</p><!><p>Mechanochemical syntheses are typically carried out in BM and also in low-frequency US reactors [7]. This technique has recently developed into a genuine eco-friendly alternative when manufacturing inorganic, organic and metal-organic compounds as well as supramolecular composites, which may differ to those prepared via conventional routes [8]. Higher versatility and selectivity offer a wide range of applications and may facilitate the purification steps [9]. Noteworthy examples are the mechanochemical derivatization of saccharides [10–11], the functionalization of CDs and their complexation with organic molecules [12]. Solid state organic reactions using CD cavities as nanoreactors have also been reported [13].</p><p>Among non-conventional techniques, the largest number of papers is dealing with US-assisted CD solubilization or re-dissolution and in a minor extends CD derivatization. Analogously, ball milling is mostly used in the preparation of CD complexes rather than synthetic preparations. MW-assisted CD chemistry covers 1/4–1/5 of the whole literature as seen in Figure 3, mainly focused on synthetic applications.</p><!><p>Distribution of energy efficient methods in CD publications.</p><!><p>As seen in Figure 4, the cake of document types dealing with CD chemistry under non-conventional techniques shows a similar distribution as observed in general CD publications, namely 70% article, 20% patents and 10% books (including non-journal conference proceedings and dissertations).</p><!><p>Document type dealing with CD chemistry under non-conventional techniques (conference proceedings and dissertations are handled as books).</p><!><p>However, industrial applications of such enabling techniques are a priori restricted to US and BM, owing to safety concerns on big scale MW reactors (Figure 5). Microreactors are a relatively new technologies and the small number of patents may also derive from solubility limitation.</p><!><p>Document type dealing with sustainable technologies in CD publications.</p><!><p>This review highlights the most recent synthetic advances in CDs' chemical modification and some perspectives that make use of non-conventional methods and energy sources. Reaction times and yields have been compared with classic procedures to highlight the huge advantages and potential scalability of these so-called enabling technologies that maximize heat and mass transfer.</p><p>Although many advances have been made during the past decade, the most exciting results in this field are surely yet to come.</p><!><p>US irradiation is an environmentally friendly technique that is well suited to the selective chemical modification of CDs from native α-, β- and γ-CD. The use of this method in heterogeneous phase reactions, such as reductions and "click reactions" [14], is well known, as is its use in full CD derivatization in combination with MW irradiation.</p><!><p>Mono 6I-(p-toluenesulfonyl)-β-CD is the most popular of the CD derivatives because it is a key intermediate in the synthesis of important amino, azido, thio, thiocyanate and halo-derivatives. 6I-(p-toluenesulfonyl)-β-CD was efficiently prepared in an US-assisted procedure by reacting β-CD with tosyl imidazole (TsIm) [15]. Great advantages, in terms of yields, reaction times and product purity, were obtained by using a cavitating tube (40 min, 19.2 kHz, 20 W, yield: 55%).</p><p>Thanks to the fast US-assisted inclusion complex formation between β-CD and TsIm reaction times have been dramatically reduced (10 min vs 1–2 hours, Scheme 1).</p><!><p>Synthesis of 6I-(p-toluenesulfonyl)-β-CD.</p><!><p>More recently, Zheng et al. have described the synthesis of this important intermediate via an US-assisted method in basic water solution [16].</p><p>The synthesis of 6I-amino-6I-deoxy-β-CD was also improved by using non-conventional techniques. The catalytic hydrogenation of 6I-azido-6I-deoxy-β-CD using Pd/C was achieved under US irradiation in MeOH/H2O in 20 min (20.4 kHz, 80 W, yield: 88%); hydrogen was supplied at 1 bar pressure [15].</p><p>Sonochemical metals depassivation in organometallic reactions is well established [17]. A typical example is the Cu(0)-catalysed azide–alkyne cycloaddition (CuAAC) that can be further enhanced by simultaneous US/MW irradiation [18]. The formation of triazole-substituted CDs has been investigated by US irradiation and products can be synthesized in 2–4 hours (Scheme 2) [19].</p><!><p>Example of CuAAC with 6I-azido-6I-deoxy-β-CD and phenylacetylene.</p><!><p>Scondo et al. have reported a preliminary study on sonochemical Staudinger-aza-Wittig tandem reactions [20] proving that isocyanate and urea formation is strongly favored. However, the applied power must be optimised for the best conversions of azido-CD into urea to be obtained and if lower efficiency in the second step is to be avoided. 6I-Benzylureido-6I-deoxy-per-O-acetyl-β-CD was obtained in shorter reaction times and excellent yields as compared to conventional conditions (Scheme 3).</p><!><p>Synthesis of 6I-benzylureido-6I-deoxy-per-O-acetyl-β-CD.</p><!><p>Tosylation of the secondary rim of the CD can be efficiently carried out under US irradiation. This efficient regioselective modification is performed in the presence of tosyl imidazole and molecular sieves under US irradiation. As shown in Table 1, the reaction time was shortened to 2 h for α-CD (yield: 36%), 1 h for β-CD (yield: 40%) and 45 min for γ-CD (yield: 46%) (Scheme 4) [21].</p><!><p>Synthesis of 3I-azido-3I-deoxy-altro-α, β- and γ-CD.</p><p>Selected examples of conventional and non-conventional preparation of monosubstituted CDs.</p><!><p>In Table 1 we compared the preparation of several monosubstituted CDs under conventional condition or under US irradiation. The data show that reaction time were dramatically reduced and the yield was generally slightly increased. Under US irradiation, the 6I-amino-β-CD was obtained by catalytic hydrogenation, while under conventional conditions the reduction of azido β-CD was obtained by a Staudinger reaction or in the presence of hydrazine.</p><p>A new generation of organophosphate scavengers has been obtained by Le Provost et al. [25] in which β-CD was regioselectively monosubstituted at O-2 using a bromomethyl pyridine derivative under US irradiation to avoid polysubstitution.</p><!><p>The complete substitution of all hydroxy groups is difficult because steric hindrance increases upon substitution, the secondary face may be attacked before the last primary hydroxy group has completed the reaction.</p><p>Totally persubstituted products are usually obtained in low yields, whereas significant increases in yields have been achieved in reduced reaction times thanks to our sonochemical protocol (35 kHz bath at 20 °C, 160 W; 20 kHz cooled horn, −20 °C, 600 W). We prepared a series of O-peralkylated β- and γ-CDs which are commonly used as stationary phases in high-resolution gas chromatography or as drug carriers [26].</p><p>CDs and their persubstituted derivatives have recently received a great deal of attention from the field of chromatographic separations. The development of new CD derivatives as important selectors for analytical chiral recognition have been performed [27]. We prepared heptakis(6-O-TBDMS-2,3-O-methyl)-β-CDs with a second CD unit in the 2 position or a (R)-Mosher acid moiety [28].</p><!><p>Bis-CDs and their metal complexes have been extensively studied as versatile receptors for molecular recognition and building blocks for functional materials.</p><p>Due to the binding of two adjacent CD units, bridged bis-CDs display high binding abilities and molecular selectivities compared to native and monosubstituted CDs. A well-organized pseudo-cavity may be provided by the linker that in turn offers additional binding interactions with guest molecules.</p><p>New sonochemical protocols for the preparation of bis(β-CDs) bearing 2-2′ and 3-3′ bridges as new carriers for gadolinium complexes have been reported (Scheme 5) [29]. These new CD dimers were promising candidates for MRI applications because their Gd(III)-adducts endowed with high relaxivities thanks to much larger molecular masses than the contrast agents themselves.</p><!><p>Synthesis of 2-2' bridged bis(β-CDs). Reaction conditions: 1) TBDMSCl, imidazole, dry pyridine, stirring rt, 8 h; 2) 5-bromopentane, LiH, dry THF–DMSO, reflux, 4 h; 3) acetic anhydride, dry pyridine, MW, 50 °C, 1 h; 4) Grubbs' catalyst, Ar, dry CH2Cl2, US, 34 °C; 5) KOH, 2 M, MeOH, H2O; US; 40 °C, 30 min; 6) AcCl 2% in MeOH; CH2Cl2, MW, reflux, 15 min.</p><!><p>Furthermore, the potential use of cyanine/β-CD carrier systems has been evaluated via in vitro experiments on HeLa cells and the monitoring of cell entrance via confocal laser scanning microscopy [30]. Several types of dye moiety/CD derivatives have been suggested as "switch on" or "switch off" fluorescent chemical sensors. In these systems, the complexation with a guest molecule allows to enhance or decrease the fluorescence intensity. Two water-soluble cyanine/β-CD derivatives have been efficiently prepared via CuAAC under simultaneous US/MW irradiation at 75 °C for 2 h (MW 15 W and US 20 W) in good yields (23% and 33%). These dyes were used as versatile carriers for drug delivery and optical imaging.</p><!><p>The reaction of β-CD with diphenyl carbonate (DPC) or hexamethylene diisocyanate (HDI) afforded crosslinked, insoluble polymers. We synthesized these systems and tested as sequestering agents for naringin [31]. These syntheses were carried out under US with shorter reaction times and smaller particle size distribution.</p><p>To investigate the cosmeto-textile applications of CD-grafted materials, a new fabric based on β-CD-grafted viscose loaded with aescin formulations was prepared. This material was designed for the treatment of venous and lymphatic legs. An efficient US-assisted synthetic procedure to graft viscose using a diisocyanate cross-linker was reported (Scheme 6) [32].</p><!><p>Insoluble reticulated CD polymer.</p><!><p>Sonochemical reticulation with HDI was used in the preparation of a new series of solid cross-linked α-, β- and γ-CD-based catalysts containing Cu(I) or Pd(II) [33]. Sonication breaks up intermicellar interaction and may promote the formation of metal nanoparticle clustering. Cu(I)-based system have been used in alkyne/azide [3 + 2] cycloadditions, while Pd(II)-based catalysts have been used in C–C couplings reactions (Scheme 7) [34].</p><p>An example of water-soluble β- and γ-CD/chitosan derivatives have been studied for binding Gd(III) chelates that bear hydrophobic substituents and negative charges [35]. These bio-polymers were easily prepared in two reaction steps by reacting CDs with maleic anhydride followed by activation with carbodiimide to form amide linkages with amino groups of chitosan. The esterification of CD was promoted by MW irradiation, while the chitosan coupling used a water-soluble carbodiimide, N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride, under US.</p><p>A mild sonication at rt using HDI enabled efficient CDs reticulation in the presence of lipases (Scheme 7) whose biocatalytic activity was preserved in the final solid cross-linked β-CD enzyme [36].</p><!><p>CD-HDI cross linked polymers.</p><!><p>Nanosponges are nanostructured materials made of hyper-cross-linked CDs [37]. The capacity of these materials to encapsulate a great variety of substances could be used to design innovative drug carriers, to protect degradable substances and to improve the aqueous solubility of poorly water-soluble molecules. α-, β- and γ-CDs were reacted solventless with diphenyl carbonate or carbonyldiimidazole under US (up to 90 °C). These nanosponges may resolve some active ingredients drawbacks, such as instability, degradation, poor solubility and toxicity, while they can also be used as carriers for inhalation and oral administration treatments [38].</p><p>New hybrid materials have been created from a combination of carbon nanotubes (CNTs) and β-CD [39] affording a peculiar cost-effective fibre. Functionalized β-CD was covalently linked to CNTs and this derivative was immobilized into the wall pores of the hollow fibre under US [40].</p><!><p>A number of general books and reviews discuss in detail the state-of-the-art of MW-assisted organic synthesis and tailor-made MW reactors have been developed for green organic synthesis [41–42]. The most recent generation of professional reactors dramatically increased the applications of MW-assisted organic synthesis thanks to a high power density (up to 1.5 kW/L), high temperature (up to 300 °C) and pressure (up to 200 bar) together with multi-gas inlets. Considering that MW ovens can be interfaced with autosamplers and that new MW reactors can accommodate multiple racks, this technique is well suited for fast optimization of organic protocols and parallel synthesis. The most impressive advantage of the MW technology is the appearance of the kilolab-scale reactors and their special versions that are operating in continuous flow mode [4,43–44].</p><!><p>MW irradiation has been exploited in the synthesis of mono and persubstituted CDs. Several syntheses of CD derivatives have been successfully carried out under MW with higher yield, higher purity, and short reaction time. While US irradiation has found use in the optimization of synthetic protocols for the preparation of versatile intermediates, such 6I-(p-toluensulfonyl)-β-CD from native β-CD, MW irradiation has proved to be extremely efficient in further derivatization, such as the nucleophilic substitution of monohalogenated and monotosylated CDs (Scheme 8) .</p><!><p>Derivatization of 6I-(p-toluenesulfonyl)-β-CD by tosyl displacement.</p><!><p>The 6I-azido-6I-deoxy-β-CD, an extremely versatile intermediate, has been obtained from the displacement of the tosylate group under MW. The reaction time was cut from several hours to 2 min (200 W max, 85 °C) and the formation of side products was reduced [15]. 6I-(p-Toluenesulfonyl)-β-CD was converted to 6I-formyl-β-CD via DMSO oxidation in MW with collidine in 15 min (110 W, 135 °C). MW irradiation promoted the syntheses of 6I-deoxy-6I-thio-β-CD and 6I,6IV-dideoxy-6I,6IV-dithio-β-CD via nucleophilic substitution of the primary tosylate ester in C-6 with thiourea followed by basic hydrolysis. The reaction gave the thiouronium salt after 1 h of irradiation at 100 °C while 20 h heating at 90 °C are required under conventional conditions [21].</p><p>While the previous experiments were performed in a multimode MW oven (MicroSynth-Milestone, Italy), a similar approach was used for the preparation of an ester prodrug of diclofenac and β-CD, but in a monomode MW oven (CEM Discover S-class MW reactor). The reaction was heated at 140 °C for 40 min and the diclofenac β-CD derivative was obtained with a yield of 20% [45]. Analogously, a general MW-assisted procedure for the synthesis of 6I-amino-6I-deoxy-β-CD has been reported by Puglisi et al. The reactions were performed in a MW oven (CEM Explorer) for 30 min at 200 W and 85 °C [46].</p><p>In the Table 2we compared MW vs conventional procedures in the preparation of several monosubstituted derivatives. Besides a slight improvement of formyl and thio derivative yield, the data show a significant reaction rate acceleration.</p><!><p>Selected examples of conventional and MW-assisted preparation of monosubstituted CDs.</p><!><p>Selective permodification refers to a complete derivatization of the hydroxy groups in one side of the CD. The selective full substitution on the primary rim is not a trivial task because of the increase of steric hindrance that makes the secondary face prone to an attack before the last primary hydroxy group was reacted [50–51].</p><p>Pertosylate and perhalogenated derivatives in position 6 can be substituted with different nucleophiles. However, under conventional conditions, the reactions resulted in complicated mixtures with different substitution degree. MW irradiation efficiently afforded pure products. A series of amino derivatives were obtained by displacement of heptakis(6-deoxy-6-iodo)-β-CD (MW reactor 150 W) at 85 °C for 1 h (yield range 52–69%) [52]. Analogously catalytic hydrogenation in a pressure-resistant MW reactor, gave heptakis(6-amino-6-deoxy)-β-CD from a solution of heptakis(6-azido-6-deoxy)-β-CD in methanol/H2O [53]. The desired product was obtained in 90% yield after 3 h of irradiation at 70 °C. Reaction with isocyanates and isothiocyanate gave ureido and thioureido persubstituted β-CD derivatives in a MW oven at 85 °C for 4 h (see Scheme 9).</p><!><p>Synthetic scheme for the preparation of heptakis(6-amino-6-deoxy)-β-CD, heptakis(6-deoxy-6-ureido)-β-CD and heptakis(6-deoxy-6-thioureido-)-β-CD.</p><!><p>A multivalent azido-scaffold such as persubstituted 6-azido-6-deoxy-α-, β- or γ-CD with conformational constraints can be efficiently perfunctionalized in a MW- and ligand-assisted click cluster synthesis. An example of the MW-promoted 'cooperative' click reaction of azido-CDs has recently been reported and offers useful synthetic insights into a specific labelling strategy [54]. The aforementioned reaction afforded a new series of antimicrobial γ-CD derivatives that strongly disrupt bacterial membranes, and a series of persubstituted γ-CD derivatives bearing polyamino groups (77% yield) [55].</p><!><p>The MW-promoted CuAAC between CD monoazides and acetylenic moieties is the most efficient way to functionalize the CD surface [56]. β-CD is able to form a stable sandwich-type complex with Cu(II) ions, where the CDs faced their secondary rims and the use of heterogeneous phase catalysis may overcome the troubles deriving from time consuming purifications [57]. In 2006 Lipshutz et al. demonstrated that the impregnation of charcoal with an aqueous solution of Cu(NO3)2 in US bath, gave copper nanoparticles: an efficient catalyst in CuAAC [58]. Besides the easier work-up of heterogeneous catalysis, Cu(I)/charcoal also gave a higher yield compared to soluble CuSO4/ascorbic acid (76 vs 95% yield, respectively). The reaction was further improved under MW or simultaneous MW/US irradiation [59].</p><p>Recently the preparation of a large number of CD-derivatives by MW-assisted CuAAC regioselective cycloadditions has been described. A selected series of derivatives are depicted in Scheme 10: CD-acryloyl derivative [60–61], β-CD/dye derivatives [31,62–64], CD-ionic liquid hybrids [65–66], CD-based iminosugar conjugates [67], water-soluble CD homo- and heterodimers [68–69], trimers [70–71] and oligomers [72] of α-, β- and γ-CD have all been successfully produced. This wide variety of compounds was obtained in good to excellent yield under MW irradiation (from 20 min to 3 h at 75 °C to 100 °C).</p><!><p>Structure of CD derivatives obtained via MW-assisted CuAAC.</p><!><p>Interest in CD polymers has grown over the last few years. CD-based polymers have a number of applications, as drug delivery systems and toxic compounds scavengers, and have been obtained by grafting CDs into polymeric matrices.</p><p>A multi-carrier for combined diagnostic and theranostic applications was obtained via the functionalization of carbon nanotubes with CD using a MW-assisted 1,3-dipolar cycloaddition. As depicted in Scheme 11, the synthesis generated in situ azomethine ylides which include both a β-CD unit and a DOTAMA tris(t-butyl ester) moiety. The toxicity assessment, cell viability and permeability of single-walled carbon nanotube (SWCNT) platform, was evaluated on five human cell lines. No-toxicity was observed at concentrations up to 333 μg/mL [73].</p><!><p>Preparation of SWCN CD-DOTA carrier.</p><!><p>Separately, a facile and rapid MW-assisted method in water has been used to derivatize graphene nanosheets with (2-hydroxy)propyl-β-CD. The reaction involved the esterification of the HP-β-CD hydroxy groups by the carboxyl groups of graphene oxide (GO) by MW irradiation (450 W) at different temperatures ranging from 50 to 100 °C for 10, 30, 60 and 90 min. After reduction with hydrazine hydrate, this HP-β-CD-RGO modified glassy carbon electrode showed good results in supramolecular recognition a set of six different phenolic organic pollutants and a high electrochemical response [74].</p><p>CuAAC has been successfully used to immobilize molecules on polymers and biopolymers as well as to join sugars to peptides and proteins. CD-polyglycerol dendron amphiphiles (CD-PG) have also been obtained. This derivative showed high encapsulation efficiency, while nanostructure size and shape were regulated according to the structure of the CD-PG dendrons [75].</p><p>CD-based polymers can be easily prepared under MW. Biswas et al. have prepared a number of macromolecular structures from α-, β-, γ-CDs by crosslinking reactions with toluene diisocyanate and methanediphenyl diisocyanate [76]. The authors demonstrated that compared with conventional heating, the reaction was faster (3–10 min) and with higher yields. Analogously, β-CD was grafted onto PEGylated Merrifield resin by reaction with HDI under MW irradiation [77].</p><p>CD nanosponges from anhydrous β-CD and diphenylcarbonate in DMF, have been prepared under MW irradiation (400 W) in 90 min. The optimized method was proven to be a unique opportunity for the large-scale synthesis of CD nanosponges in a high yield and uniform particle size distribution [78].</p><!><p>One of the oldest, cheap, and efficient methods to achieve a homogeneous solid mixture is ball milling. By this method extremely fine powders can be achieved in mineral dressing processes, paints and pyrotechnics, etc. [79]. It is suitable for both batch and continuous operation, it is similarly suitable also for open and closed circuit grinding as well as being applicable for materials of all degrees of hardness.</p><p>Conventional BMs have a cylindrical or conical shell that rotates on a horizontal axis and have an appropriate grinding medium of balls, for example steel, flint or porcelain. The second generation of BMs, which are often called as high-speed ball mills (HSBM), operate in vibrating, mixer or planetary mode. A very simple vibrating BM, consisting of a small milling cup with one or two balls, has been used for a long time in traditional IR spectrometry to homogenize the sample and KBr. Mixer BM are slightly different from the vibrating version and are not only used in IR spectroscopy but also on the preparative scale for homogenization and cracking solid components. The common weakness of these simple accessories is the critical rotation/mixing speed, which can be overcome by a new generation of equipment; planetary BM, that consist of at least one grinding jar arranged eccentrically on a rotating support. The grinding jar moves in the opposite direction to the sun wheel. The difference in speeds between the balls and grinding jars produces an interaction between frictional and impact forces, which releases high dynamic energies for particles size reduction [80]. Detailed descriptions of both operating modes and theoretical considerations can be found and thoroughly discussed in various product brochures.</p><p>An energy efficient method for the preparation of nanocrystalline powders is the high energy ball milling (HEBM) in planetary or vibratory ball mills and HEBM is a common synonym for HSBM [81]. The lower particle size in grinding produces microdeformation in the ground material crystal lattice, while energy is partially spent in creating microstresses, which eventually slow powder grinding. An efficient wet grinding technology can exploit a liquid milling medium.</p><p>The preparation of CD and other complexes with the aid of ball milling is well-known [82–83]. In spite of an easy scale-up of this technology, some disadvantages might occur:</p><!><p>metastable crystalline complexes can recrystallize to an equilibrium state upon storage [84];</p><p>the degradation of mill surfaces and subsequent suspension contamination can be a problem, particularly in the high-energy version [85].</p><!><p>Although, the preparation of complexes or microparticles with ball milling is a common procedure, its use in organic synthesis intensified substantially only recently [86]. Solventless mechanochemical reactions are usually highly efficient and selective, valuable properties exploitable in CD derivatization.</p><p>Nucleophilic substitutions (SN2 reaction) may occur without solvent stabilization because charged species do not need to be formed in the transition state [87]. Solvent effects and ion pair formation are critical to the mechanism of SN1 reactions meaning that this mechanism is usually restricted in HSBM reactions.</p><p>While solid-state intermolecular SN2 reactions depend on contact between interacting particles only, SN1 reactions may show more structure-dependent behaviour, which can be either favourable or unfavourable, because of the solid-state structure.</p><p>Although BM reactions are often said to be solvent-free, some inert solvents can also be used particularly when the reagent mass ratio is very high. A lack of solvent(s) may suggest that ball milling conditions favours SN2 reactions; however it is also true that a solventless environment does not necessarily mean that there is a lack of solution in a liquid phase. Some reaction mixture components can often be liquid, while solvent effects or mixed SN2 and SN1-type reaction mechanisms cannot be excluded. A good example of a mixed reaction mechanism is the glycosylation reported by Tyagi et al. [88], where SN2 glycosylation seems to be dominant, with no neighbouring group participation, which is typical of glycosylation reactions of activated acetylated carbohydrates. A more pure SN2 reaction is described by Patil and Kartha [89], where the preparation of thioglycosides was almost quantitative. Unfortunately, a lack of information on reaction mixture compositions means that the reaction mechanism cannot be completely confirmed because chromatographic purifications and recrystallizations distort the enantiomeric ratio.</p><p>Basically, three major types of HSBM chemical reaction can occur in the presence of CDs:</p><!><p>Preparation of CD complexes and various chemical reactions on the complexed substructures;</p><p>Derivatization of naked, natural CD;</p><p>Reactions of activated CD.</p><!><p>While reactions occur between a complexed molecule and reagent or between host and guest in cases 1) and, usually, 2), reaction type 3) requires a CD derivative that bears a good leaving group and the complexation phenomenon can be disadvantageous here. While type 1) can eliminate usually the less problematic solvents only, the application of BM in types 2) and 3) can reduce or eliminate the polluting environment. Reactions of type 1) are dominant in CD/BM literature; more than 98% of publications report the complexation of one or more components. Mechanochemistry opened a new synthetic pathway to the preparation of numerous fullerene derivatives by dissolving C60 in the amorphous powder obtained from the ball milled reactants and β-CD [90]. Another example that uses the energy transfer of ball milling is the preparation of MnBi/Fe-Co core/shell structured composites. However, no pure chemical reaction is used to prepare rare-earth free ferromagnetic materials by grinding under less-environmentally friendly conditions in this case. The components were prepared using classic methods and the final composite was obtained by ball milling of arc-melted MnBi particles and Fe-Co nanoparticles prepared with the aid of a β-CD/oleic acid complex. The composites obtained showed smooth magnetic hysteresis loops [91].</p><p>SWCNT edge activation can be carried out via co-grinding with β- or γ-CDs [92]. Although chemical bonds are also broken in this case, this preparation is closer to the BM assisted preparation of CD complexes in many ways. Nanosized manganese oxides have also been prepared from CD/Mn complexes [93], however, in this case, the CD was only used to obtain a charrable matrix for the Mn2O3 which was prepared finally at 450 °C.</p><p>The only example of the type 2) method is the regioselective CD derivatization described by Menuel et al. who prepared 2-O-monotosylated α-, β-, and γ-CDs [94]. The further reaction of the prepared compounds resulted in a CD derived cyclic oligosaccharide, which contained one mannose residue, in the form of 2,3-mannoepoxide.</p><p>Type 3) reactions in the further derivatization of regioselectively activated – by sulfonic esters or halogenides – CDs are more important in industrial processes involving important CD derivatives. These activated derivatives are usually less soluble in water and their substitution reactions often require high boiling point dipolar aprotic solvents. The complete removal of these solvents is impossible even in gram scale preparations and so the prepared compounds need further purification steps. Additionally, these environmentally unfavorable solvents present other disadvantages; both in their decomposition and toxicology profile. A study of the nucleophilic displacement of 6-monosubstituted β-CDs and the synthesis of 6I-monoazido-6I-monodeoxy-β-CD in HSBM on a preparative scale (5 mmol, 6.5 g) is described in a recent publication by Jicsinszky et al. [95]. Comparing the yields it can be concluded that in larger scale reactions the yields are getting closer to those of the solution reaction. However, since the removal of a high-boiling solvent is not necessary, the work-up becomes simplifyed.</p><p>It has to be highlighted that the reaction product should not be considered as a CD derivative when the reaction centre is on the secondary rim because the SN1 mechanism is restricted to solution environment only. The secondary carbon substitution results in inversion in the reaction centre which changes the sugar moiety from glucoside to mannoside, altroside or alloside making those derivatives CD-based cyclic oligosaccharides and not CDs.</p><p>The design of green synthetic methods for the bulk preparation of CD thiols and thioethers is an emerging challenge because of the importance of intermediary azido derivatives [96] and favorable aggregation properties in nanomedicines and particularly the antidote Sugammadex [97]. The reaction between 6I-O-monotosyl-β-CD and various nucleophiles opens a new way for the more effective syntheses of per-6-substituted CDs from per-6-bromo- and -iodo-CDs.</p><!><p>The typical lateral dimensions of microreactors, sometimes also called as microstructured or microchannel reactors, are below 1 mm with the most typical form of microchannels [98]. The miniaturized continuous flow reactor, also called microreactor, offers many advantages over conventional scale reactors, including considerable improved energy exploitation, increased reaction speed and yield, safety, reliability, scalability, on-site/on-demand production, etc., and a much finer degree of process control. However microreactors do not tolerate mechanical inhomogeneities. To resolve the problem of microparticles, which often cause clogging, a second generation of microreactors has been developed and called microjetreactor [99]. A typical microreactor is made up of a 'sandwich' of thin metal sheet or plates with fluid (micro)channels that have been etched into both sides. The average size of a single unit is approximately 6 × 4 × 0.5 cm with channel widths and wall thicknesses of 200–300 µm. The reactions occur in every other layer and the other layers are used for heat-exchange fluid flows [100].</p><p>The major use of CDs in this equipment, and also in the selective complexation phenomenon, is rather analytical and CDs' principal role is detection only [101]. This has allowed DNA sequencing to become a relatively cheap method and provided momentum to the discovery of the role of genetics in various diseases [102–103]. Although these reactors exhibit an excellent energy and mass efficacy, their use in CD derivatization is just a curiosity. However, exhausting the complexation ability of various CD derivatives is advantageous in solubilization and stereoselective reactions. Delattre and Vijayalakshmi have pointed out the theoretical use of enzymes in the production of CDs or other cyclic oligosaccharides, like cyclofructan, rather than using a microreactor [104].</p><!><p>Dynamic intrusion of the enabling technologies to the CD chemistry is inevitable and shows exponential growth. Although, approximately 10% of the recently published technical papers in the CD field are dealing with sustainable technologies, the number of publications containing information of comparisons with the classical methods is sporadic. Optimized MW-, US- and BM-assisted protocols are energetically more efficient than the classical synthetic methods because their excellent heat and mass transfer. In all cases the reactions are faster avoiding degradations that may occur during protracted heating and time-consuming purifications. Case by case the technique of choice depends from several factors: the solubility of the starting CD, the reaction mechanism, environmental concerns, and the reaction scale are only a part of all the information required to design successful preparations.</p>

PubMed Open Access

A General Pyrrolidine Synthesis via Iridium-Catalyzed Reductive Azomethine Ylide Generation from Tertiary Amides & Lactams

A new iridium-catalyzed reductive generation of both stabilized and unstabilized azomethine ylides and their application to functionalized pyrrolidine synthesis via [3+2] dipolar cycloaddition reactions is described. Proceeding under mild reaction conditions from both amide and lactam precursors possessing a suitably positioned electron-withdrawing or a trimethylsilyl group, using catalytic Vaska's complex [IrCl(CO)(PPh3)2] and tetramethyldisiloxane (TMDS) as a terminal reductant, a broad range of (un)stabilized azomethine ylides were accessible. Subsequent, regio-and diastereoselective, inter-and intramolecular, dipolar cycloaddition reactions with variously substituted electron-poor alkenes enabled ready and efficient access to structurally complex pyrrolidine architectures. Density functional theory (DFT) calculations of the dipolar cycloaddition reactions uncovered an intimate balance between asynchronicity and interaction energies of transition structures which ultimately control the unusual selectivities observed in certain cases.

a_general_pyrrolidine_synthesis_via_iridium-catalyzed_reductive_azomethine_ylide_generation_from_ter

Introduction<!>Results and Discussion<!>Selectivity of cycloaddition.<!>Mechanistic Investigations<!>Conclusion

<p>Saturated pyrrolidine heterocycles are prevalent in biologically active natural products, 4 and are among the 10 most common ring systems in small drug molecules (Scheme 1A). 1 Accordingly, new broad scope methods for their synthesis remain important. Whilst relatively simple pyrrolidine derivatives are commercially available, polysubstituted pyrrolidines generally require synthetic effort. To this end, [3+2] dipolar cycloadditions of azomethine ylides are synthetically powerful, 5 allowing the direct construction of the saturated five-membered ring system with control over up to four newly-formed stereogenic centres in an atom-economic reaction. Consequently, the synthesis and reactions of azomethine ylides have been the focus of a number of research efforts over the years (Scheme 1B). These dipoles can be prepared from the opening of an aziridine ring, 6a-6i or more commonly, from the activation of an imine (usually accessed from the condensation of an aldehyde and a primary or secondary amine either in or ex situ) and are especially useful for the synthesis of pyrrolidines unsubstituted on the nitrogen atom. 5c, 6j-6s Other methods also exist, requiring the construction of finelytuned precursors. 6t-6aa Notwithstanding these many advances, to date a general reductive strategy for azomethine ylide 1,3-dipole generation from tertiary amides and lactams enabling downstream access to desirable pyrrolidine structures, remains unsolved. 6ab-6ae Towards this end and building on our program on reductive manipulation of amide functional groups, 7 we reasoned that iridium-catalysed hydrosilylation of suitably functionalised tertiary amides and lactams could provide a new entry point. With substrates possessing a suitably positioned electron-withdrawing or a trimethylsilyl group, following partial reduction and subsequent silanoate elimination, concomitant deprotonation or loss of a trimethylsilyl group adjacent to the iminium ion could feasibly generate the synthetically versatile azomethine ylide (Scheme 1C). Subsequent cycloaddition reaction with dipolarophiles would then give access to the decorated pyrrolidine ring in a convenient one-pot process. Such a strategy would potentially provide an avenue for the late-stage synthesis of highly functionalized pyrrolidines from stable and widely abundant amides, under mild conditions, whilst eliminating the need for handling sensitive amine functionalities. Herein, we wish to describe our findings.</p><!><p>Optimization Studies. Proline methyl ester benzoylamide derivative 1a was chosen as a model system to investigate the transformation, alongside oxazolidinone dipolarophile 2a, selected for its previous use in [3+2] cycloadditions and for its easy downstream derivatization. 8 Using 1 mol% IrCl(CO)(PPh3)2 (Vaska's complex) and 2 equivalents of tetramethyldisiloxane (TMDS) for partial amide reduction, plus additional triethylamine as a Brønsted base for the generation of the dipole, we were pleased to isolate the desired product 3a in 50% yield as a single diastereoisomer, indicating the formation and subsequent stereoselective cycloaddition of the azomethine ylide (Table 1, entry 1). Notably, a control experiment revealed that no additional base was required for the reaction to proceed (entry 1 and 2), indicating that the eliminated silanoate was indeed a competent Brønsted base for dipole generation (see Scheme 1C). The use of 2 equivalents of TMDS was optimal (entry 2 and 3), and the reaction could also proceed at higher temperature, albeit in a slightly reduced yield (entry 2 and 4). Therefore, entry 2 was chosen as standard conditions for assessing the scope of the reaction.</p><p>Table 1. Iridium-catalyzed reductive dipole generation and reaction optimization studies. a General conditions: 1a (0.25 mmol scale), IrCl(CO)(PPh3)2 (1 mol%), additive, 1,1,3,3-tetramethyldisiloxane (TMDS), toluene (1 mL), rt, 16 h. b NMR yield calculated with 1,3,5-trimethoxybenzene as an internal standard; isolated yield in parenthesis.</p><p>Scope Development. With the optimal reaction conditions in hand, we explored the scope of the [3+2] cycloaddition reaction with a range of electron-poor alkenes as coupling partners. A 1,1'disubstituted methacrylic acid derivative successfully underwent cyclization to give a cycloadduct possessing a quaternary carbon center (3b) while crotonic and cinnamic acid derivatives gave rise to the corresponding polysubstituted products bearing four adjacent stereocenters in good yields with high diastereoselectivity (3c, 3d). Importantly, the reaction was not limited to N-enoyl oxazolidinone coupling partners; acrylate, furanone, nitroalkene, cinnamate, and vinyl sulfone derivatives (3e-3i, respectively) all produced the desired cycloadducts, with a large diversity of substitution patterns on the pyrrolidine ring, and in good to excellent yields. Interestingly, the regioselectivity of the formation of 3h was reversed when compared to 3d, as confirmed by single crystal X-ray diffraction analysis, while 3g and 3i were formed as a mixture of the two regioisomers (in line with literature precedent). 9 Pleasingly, the reaction sequence was found to be general with respect to the amide substrate. The presence of a methyl ester was found not to be a limiting requirement (3j, 3m), while the azetidine-containing amino acid derivative afforded the corresponding [3,2,0] bicyclic compound in moderate yield (3k). Importantly, this methodology could also be applied to lactam substrates, providing bicyclic cycloadducts containing a tertiary stereocenter adjacent to the nitrogen atom, that would otherwise be inaccessible from other azomethine ylide generation methods (3l-3n). This also demonstrated that the methodology was not limited to the reduction of reactive benzoyl amides, therefore increasing the diversity of scaffolds made accessible. A range of 2-carboxy-substituted pyrrolizidine was thus obtained with a 4carbonyl (3l, 3m) or 3-aryl 4-carbomethoxy (3n) substitution pattern. Cinnamoyl amide 1o also gave rise to the 4-alkenyl trisubstituted pyrrolidine 3o in good yield. Having established that amides and lactams possessing a βester appendage on the nitrogen atom were excellent precursors to stabilized azomethine ylides, we next turned our attention to unstabilized 1,3-dipole generation. Using N-(trimethylsilyl)methyl amides as substrates, and following standard conditions for a Vaska's complex catalyzed hydrosilylation, with substoichiometric amounts of TMSCl as an additive to trigger the desilylation, we were pleased to observe [3+2] dipolar cycloaddition of unstabilized azomethine ylides taking place. This approach is complementary to the one described above, as the resulting pyrrolidine ring of the cycloadduct bears no substituent  to the nitrogen atom. As shown in Scheme 4, the reaction was found to be tolerant of a good range of aryl and heteroaryl amides, although alkyl amides remained difficult substrates. Pleasingly, diastereocontrol was improved by increasing the steric demand of the substituent on the amide nitrogen (R) from methyl (5a) to benzyl (5b). Aryl amides containing both electron-donating (methoxy, methyl) and electron-withdrawing (halides, nitro and nitriles) groups afforded the corresponding pyrrolidines, with good to excellent diastereoselectivity (5c-5k). Heteroaromatic amides also underwent cycloaddition smoothly (5l, 5m). Alkyl amide substrates gave a complex mixture, indicating lack of regio-and diasterecontrol. Several other dipolarophiles could also be used, leading to products derived from tert-butyl acrylate (5n), dimethyl fumarate (5o), N-phenyl maleimide (5p) and phenyl vinyl sulfone (5q), although a reduced diastereoselectivity was observed (Scheme 4B). Pyrrole 5r was obtained in a good yield using dimethyl acetylenecarboxylate as the dipolarophile under standard conditions, followed by the oxidation of the resulting dihydropyrrolidine ring by DDQ. This process can provide substituted pyrroles in a one-pot sequence from inexpensive and robust tertiary amides.</p><p>Intramolecular Cyclization. In order to demonstrate potential applications of our methodology, a linear substrate for an intramolecular cyclization was synthesized and subjected to the reductive cycloaddition using the optimized standard conditions (Scheme 5A). Pleasingly, a remarkably chemoselective reduction of the amide carbonyl 6 was achieved, leading to the formation of the tricyclic [5,5,5] core 7, via the putative azomethine ylide, as a single diastereoisomer, in good yield.</p><!><p>As shown previously in Scheme 3 and in Scheme 5B above, the reductive [3+2] cycloaddition reaction showed a high regio-and diastereocontrol, according to the dipolarophile used. N-Enoyl oxazolidinone and tert-butyl acrylate gave products (3a, 3e) with new C-C bond formation occurring at the carbonyl carbon atom and the  position of the alkene as a single product, whereas methyl cinnamate gave a product (3h) with new C-C bond formation occurring at the carbonyl carbon atom and the  position of the alkene as a single product. These results indicate the selectivity is highly dependent on the dipolarophile, and accordingly we turned our attention to the use of a chiral auxiliary as it can potentially give diastereoand enantiomerically pure pyrrolidines after the removal of the oxazolidinone group. Interestingly and unexpectedly, the cycloaddition with a chiral coupling partner gave two regioisomers of the product in a 1:1 ratio, and a combined 78% yield (8a, 8b, Scheme 5B). These two isomers are fully separable by silica gel chromatography, and each of them is obtained essentially as a single isomer (>20:1 dr). The absolute and relative configuration of 8a was unambiguously determined via single X-ray diffraction analysis.</p><!><p>DFT Study. To further investigate the selectivity involved in the cycloaddition reaction, we turned to state-of-the-art DFT calculations. Our focus was on elucidating the mechanism which resulted in an inversion of selectivity when methyl cinnamate was used as a coupling partner as opposed to N-enoyl oxazolidinone, giving respectively 3h and 3d.The regio-and diastereoselectivities of the 1,3-dipolar cycloaddition involving an azomethine ylide and a dipolarophile are determined by either the strain or the interaction energies of the cycloaddition transition structures depending on the nature of the dipolarophile. The strain energy is decisive for the selectivity when the dipolarophile is methyl cinnamate, whereas the interaction energy controls for the selectivity when the dipolarophile possesses an oxazolidinone group. The origin of this unique divergent behavior depending on the structure of the dipolarophile is quantified and explained below.</p><p>The key cycloaddition transition structures between the in situ generated azomethine ylide and methyl cinnamate or N-enoyl oxazolidinone are provided in Scheme 6. Among the four TSs with methyl cinnamate as the dipolarophile, the energy barrier via TSOMe4 is the most favorable, whereas for N-enoyl oxazolidinone, TSOx2 is the most energetically favorable transition structure. 10 The origin of the kinetic preference for the regio-and diastereomer determining cycloaddition step was quantified using the activation strain model (ASM). 11 The ASM involves the decomposition of the electronic activation barrier (E ‡ ) into two distinct energy terms, namely, the strain energy (E ‡ strain) that results from the deformation of the individual reactants and the interaction energy (E ‡ int) between the deformed reactants. These analyses have revealed that the strain energy controls the selectivity through TSOMe4 with methyl cinnamate, while the interaction energy is decisive for the selectivity through TSOx2 with N-enoyl oxazolidinone. The higher degree of asynchronicity, defined as the difference in the length of two newly forming C-C bonds in the transition structures, in TSOMe4 leads to a less destabilizing strain energy. We recently reported the connection between asynchronicity and strain energy in the related Diels-Alder cycloaddition reaction. 12 A higher degree of asynchronicity leads to one C-C bond to form before the other and results in smaller degree of pyramidalization (sum of angles [SoA] around a carbon atom in °) at the reacting carbon atoms (Scheme 7A). A good linear relationship was found between the normalized sum of angles of pyramidalization at two reacting carbon atoms (720-SoA1-SoA2 [for the dipole fragment] or 720-SoA3-SoA4 [for the dipolarophile fragment]) and the strain energies ΔE ‡ strain of each fragment. Therefore, the TS that minimizes the pyramidalization of the reacting carbon atoms (SoA closer to 360°) benefits from a less destabilizing strain energy by the asynchronicity of the TS, and thus, a lower activation barrier if interaction energies are all similar.</p><p>On the other hand, the interaction energy was found to be operative in controlling the selectivity when the dipolarophile contains an oxazolidinone group and actually overrules the strain energy, which was decisive with the methyl cinnamate substrate, as previously discussed. The prominent role of the interaction energy on the observed reactivity trends stimulated the analysis of various contributions to the interaction using a canonical energy decomposition analysis (EDA). 13 Our canonical EDA decomposed the Eint between the distorted reactants into three physically meaningful energy terms: classical electrostatic interaction (Velstat), steric (Pauli) repulsion (EPauli) which, in general, arises from the two-center four-electron repulsion between the closed-shell orbitals of both reactants, and stabilizing orbital interactions (Eoi) that account, among others, for HOMO-LUMO interactions. Analysis of EDA terms computed on the solution phase geometries in the gas phase 14 revealed that the more stabilizing orbital interactions (ΔE ‡ oi) and electrostatic interactions (ΔV ‡ elstat) for TSOx2 set the trend in ΔE ‡ int. Analysis of the bonding mechanism and frontier molecular orbital (FMO) interactions revealed that the origin of the more stable ΔE ‡ oi associated with TSOx2 originates from both a smaller normal electron demand (NED) HOMOdipole-LUMOalkene energy gap and the larger orbital overlap S compared to the other TSs. These combined effects result in the most stabilizing orbital interactions (S 2 /Δ×10 3 = 9.0) for TSOx2 (Scheme 7B). 15 The more stable ΔV ‡ elstat for TSOx2 can be understood from analysis of the molecular electrostatic potential (MEP) maps (Scheme 7C). Here we see that the two carbon atoms participating in the shorter newly forming C-C bond for TSOx2 benefit from a complimentary "charge match" compared to that of TSOx4 (the next most favorable TS). That is, the negatively charged carbon atom on the dipole and the positively charged carbon atom on the dipolarophile conveniently enter into a stabilizing electrostatic interaction, a feature that is maximized when the electron withdrawing group on the dipole and alkene are positioned opposite of each other, such as in TSOx2.</p><!><p>In summary, we have developed a new, general and highly selective reductive [3+2] cycloaddition reaction of amides and conjugated alkenes for structurally complex pyrrolidine synthesis. This unified strategy, enabled by the use of Vaska's complex and TMDS to reductively generate both stabilized and unstabilized azomethine ylide species, generates after cycloaddition a wide range of highly and diversely substituted pyrrolidines and polycyclic amine products. The reaction proceeds under mild conditions, and our investigations have shown good functional group tolerance. The use of single enantiomer or intramolecular dipolarophiles demonstrates applicability to the synthesis of complex and enantiopure cyclic amine architectures. The origin of high regio-and diastereoselectivity in the cycloaddition reaction was elucidated by means of state-of-the-art density functional theory (DFT) computations. Use of the activation strain model (ASM) in conjunction with the matching canonical energy decomposition analysis (EDA) revealed that the selectivity is determined by either the strain or the interaction energies depending on the substituent on the dipolarophile. Highly asynchronous transition states are energetically preferred and go with a lower strain energy than synchronous one, unless a highly stabilizing interaction energy between the reactants is present, in which the orbital and electrostatic interactions are decisive. Furthermore, successful application of this method to the construction of a complex tricyclic core efficiently from a readily prepared substrate shows the potential to synthesize complex molecules possessing naturally abundant pyrrolidine scaffolds.</p>

ChemRxiv

Electrochemical Mechanism of Tellurium Reduction in Alkaline Medium

A systematic electrochemical study was conducted to investigate the reduction of tellurium (Te) in alkaline solutions. The effect of various parameters, including tellurite ion concentration, applied potential, and pH was investigated by both linear sweep voltammograms (LSVs) and electrochemical quartz crystal microbalance (EQCM). EQCM was essential to understand the reduction of Te(0) to soluble Te22-(-I) or Te2−(-II). The Tafel slopes for two Te reduction reactions [i.e., Te(IV) to Te(0) and Te(0) to Te(-I)] indicated that the electrochemical reduction of Te is strongly dependent on solution pH, whereas it is independent of the concentration of TeO32-. At relatively weaker alkaline solutions (i.e., pH ≤ 12.5), the discharge of Te(OH)3+ was determined to be the rate-limiting step during the reduction of Te(IV) to Te(0). For the reduction of Te(0) to Te(-I), the reaction follows a four-step reaction, which consisted of two discharge and two electrochemical reactions. The second discharge reaction was the rate-limiting step when pH ≤12.5 with the Tafel slope of 120 mV/decade. At a higher pH of 14.7, the Tafel slope was shifted to be 40 mV/decade, which indicated that the rate-limiting step was altered to the second electrochemical reaction. Te(0) deposits were found either on the surface of an electrode or in the solution depending on pH due to the different rate-limiting reactions, revealing that pH was a key parameter to dictate the morphology of the Te(0) deposits in alkaline media.

electrochemical_mechanism_of_tellurium_reduction_in_alkaline_medium

Introduction<!>Materials and Methods<!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!><!>Results and Discussion<!>Conclusion<!>Data Availability Statement<!>Author Contributions<!>Conflict of Interest<!><!>Supplementary Material<!>

<p>Tellurium (Te), due to its unique crystal structure, exhibits many unique properties, such as piezoelectric effect, photoconductivity, catalytic activity, gas sensing, and thermoelectrics (Bradstreet, 1949; Shurygin et al., 1975; Kambe et al., 1981; Fujiwara and Shin-Ike, 1992; Wang et al., 2006; Liu et al., 2010b; Zhang et al., 2012; Lee et al., 2013; Park et al., 2013). Recently, considerable attention has been also directed to tellurides, such as bismuth telluride (Bi2Te3), lead telluride (PbTe), antimony telluride (Sb2Te3), and cadmium telluride (CdTe), revealing remarkable performances in electronic applications (Wuttig and Yamada, 2007; LaLonde et al., 2011; Kumar and Rao, 2014; Guo et al., 2015).</p><p>Electrodeposition, which requires ambient conditions and low capital cost, has been widely studied for a synthesis of tellurium and its alloys, and it has been especially promising for large-scale deposition with higher deposition rate to fabricate the Te and its alloy for a variety of applications (Wu et al., 2014, 2015; Ferrer-Argemi et al., 2019; Kim et al., 2019). For instance, the combination of electrochemical deposition of compound semiconductors (metal tellurides) with standard integrated circuit technique makes fabrication of thermoelectric microdevices possible, which can convert rejected or waste heat into usable electric power. The understanding of electrochemical reactions of Te ion in alkaline conditions would benefit the electrodeposition metal telluride compound for thermoelectric microdevices.</p><p>In order to apply Te or its alloys for a specific application, the morphology, crystal structure, and size must be precisely controlled to exhibit different properties. Thus, it is critical to understand the relationship between the operating conditions of electrodeposition—such as electrolyte concentration, pH, temperature, and agitation rate—and resulting physical properties of the synthesized Te. In addition, because the satisfactory quality of Te electrodeposits is highly hinged on a critical control of a technique with a detailed understanding of deposition mechanism, it has been emphasized to investigate the reduction mechanisms in a wide range of electrodeposition conditions. The electrodeposition of Te has been investigated extensively in both aqueous and non-aqueous solution to incorporate chemical and electrochemical thermodynamics as well as to understand the kinetics of electrochemical reactions of Te. Most of the works demonstrated the Te electrodeposited with various nanostructures, showing the dimensional effect on the physical properties. For example, the piezoelectric property of one-dimensional (1-D) nanowires was enhanced by reducing the diameter of the nanowire due to the flexoelectric effect (Liu et al., 2010a; Wang, 2012). In addition, the light absorption observed in 3-D Te nanostructures was enhanced due to the refractive index gradient from nanostructures (Diedenhofen et al., 2009; Chao et al., 2011; Yan et al., 2014). These works, however, typically employed acidic electrolytes, which have low TeO2 solubility, inevitably leading to low tellurium ion (HTeO2+) concentration. This would allow for the deposition reaction to be under the HTeO2+ diffusion control, and consequently, it would be difficult to macro-structure due to a low deposition rate.</p><p>The higher solubility of TeO2 in alkaline solutions allows the fabrication of thick compact films (Wu et al., 2014, 2015). The high TeO32- concentration in the electrolyte would rapidly replenish the TeO32- near the substrate, improving the deposition rate and uniformity of the electrodeposits. The electrochemical reduction mechanisms of Te in alkaline solutions have been examined by several groups using cyclic voltammetry (CV). Lingane and Niedrach. reported that the TeO32-(IV) was reduced to Te(0), observed by the black deposit on the working electrode, then Te(0) was further reduced to Te22-(-I), correlated to the deep violet color of the solution, which finally changed to colorless, representative of the reduction of Te22-(-I) to Te2−(-II) (Lingane and Niedrach, 1949). Schmidt cited the reaction mechanisms from Lingane's publication, but did not mention the formation of Te22-(-I) (Schmidt, 1962). Shinagawa et al. reported the same reduction reaction mechanisms as Lingane and Niedrach. (Shinagawa et al., 1972). Mishra detected the presence of intermediate species Te22-(-I) by using a rotating ring-disk electrode as Te2−(-II) was oxidized to Te(0) (Mishra et al., 1990). The Tafel slopes of the reduction reaction [i.e., TeO32-(IV) to Te(0) and Te(0) to Te(-I)] were reported by several groups and possible reaction mechanisms were proposed (Awad, 1961, 1962; Komandenko and Rotinyan, 1966, 1967a,b). Our previous publication also discussed the reaction mechanisms for Te in comparison with the literature data (Wu et al., 2014, 2015, 2017). However, a systematic study of Te reactions in a large Te ion centration and pH window using linear sweep voltammograms (LSVs), electrochemical quartz crystal microbalance (EQCM), and Tafel plots has not been reported. While the Tafel plots provide details about reaction kinetics, EQCM provides detailed information to interpret the mechanism for reduction of TeO32-(IV) and Te(0). Among the valence states of Te, TeO32-(IV), Te22-(-I), and Te2−(-II) are solvable species, while Te(0) is solid. Therefore, when the Te(0) is deposited, the mass on the electrode would increase. When the Te(0) is then reduced to either Te22-(-I) or Te2−(-II), the solid Te would dissolve and the mass on the electrode would decrease. This mass change can be recorded by the EQCM to understand the electrochemical reactions of Te.</p><p>In this work, voltammetric investigations of reduction mechanisms from TeO32-(IV) to Te(0) and from Te(0) to Te2− (-I) were systematically conducted in various alkaline solutions. The effect of tellurite concentration and pH on reduction reaction kinetics was studied by measuring changes in current, charge, and mass on the working electrodes. Furthermore, the elementary reactions for each reduction reaction were discussed based on Tafel slopes. Accordingly, applied potential window for successful deposition of Te(0) was determined.</p><!><p>All solutions were prepared by dissolving various amounts of tellurium dioxide (TeO2, 99+%, Acros Organics) in sodium hydroxide solutions (NaOH, 10 N, Fisher Chemical). The pH of each solution was adjusted by NaOH. All the solutions were deaerated by bubbling high-purity N2 (99.999 %) for 40 min. Linear sweep voltammetry was performed in a conventional three-electrode cell using a rotating disk electrode (RDE) [50-μm-thick Te films pre-deposited on gold-coated copper rods (diameter = 6.4 mm) embedded in a cylindrical Teflon holder] as working electrode, platinum-coated titanium stripe as counter electrode, and saturated Ag/AgCl as reference electrode. The scan rate was fixed at 1 mV/s.</p><p>Tafel plots were derived from LSV data. The effect of TeO32- concentration on the Tafel slope for the reduction reaction of Te(IV) was investigated by varying the TeO32-(IV) concentration from 50 to 550 mM with a fixed solution pH of 12.0 at 23°C. Additionally, the effect of pH on the TeO32-(IV) reduction reaction was investigated by varying the pH from 10.2 to 14.7 (calculated value) at a fixed TeO32- concentration of 550 mM at 23°C. The pH effect on the reduction reaction of Te(0) was investigated by varying the solution pH from 10.2 to 14.7, while fixing the TeO32- concentration and temperature at 0 mM and 23°C, respectively.</p><p>A potentiostat/galvanostat (Biologic, SP-200) combined with an EQCM (QCM200, Stanford Research Systems) was used for electrochemical investigation. The AT-cut 5-MHz quartz crystal covered with chromium/gold served as the working electrode. The platinum-coated titanium stripe and saturated Ag/AgCl were used as counter and reference electrodes, respectively. The effect of pH on reduction reaction of Te(IV) was examined in a range from 10.2 to 14.7 with TeO32- concentration of 300 mM at 23°C. To investigate the reduction reaction of Te(0) using EQCM at different pH (i.e., 10.2, 12.5, and 14.7), a thin film of Te with a mass of 70 ± 3.5 μg cm−2 was pre-deposited on the top of the working electrode. Scan rate of the EQCM experiments was 50 mV/s. Based on Sauerbrey equation, the frequency changes, Δf, of the quartz crystal were correlated with the mass changes, Δm (Seo et al., 1994):</p><p>where K is the sensitivity factor for the crystal (i.e., 56.6 Hz μg−1 cm2).</p><p>The Faraday law can be described by the following equation:</p><p>where Q is the charge consumed during the deposition, z is the number of electrons involved in the reaction, Δn is the change in the number of moles of the deposits, and F is the Faraday constant. When Equations (1) and (2) are combined with the equation Δm = M·Δn, in which M is molecular weight, the following equation can be derived (Saloniemi et al., 2000a,b):</p><p>In the correlation of Δm as a function of Q, the slope (S) would then be defined as:</p><p>The number of the electrons involved in the reaction can be calculated by the following equation (Santos and Bulhoes, 2003):</p><!><p>The tellurium known to dissolve with simultaneous formation of TeO32-(IV) and Te22-(-I) ions in alkaline solution is cathodically reduced to give either Te(0) or Te2−(-II) ion. Therefore, a systematic voltammetric investigation using its polarographic behavior is required for a comprehensive understanding of Te reduction in alkaline media. As shown by the slight increase in the current density, the LSV curve (a) in Figure 1, the first reduction reaction started at −0.77 V. The current density continued to increase until the applied potential reached −1.25 V, where it marked a dipping point. The large increase in current density was probably caused by the reduction of TeO32-, which are further discussed later in this paper. The current density at more negative potentials past the dip point started to rapidly decrease, and then, at −1.5 V, slightly increased again until the potential reached −1.8 V. The sharp spike in current density at −1.8 V might be attributed to hydrogen gas evolution reaction (Pourbaix, 1966). The curve (b) in Figure 1 was obtained by cathodic reduction of the Te(0) electrode in a NaOH solution in the absence of TeO32-, in which the reduction reaction of Te(0)/Te22-(-I) started at −0.92 V. At more negative potentials, the Te(0)/Te22-(-I) reduction reaction rate increased, resulting in a sharp increase in the cathodic current density. However, at an applied potential of −1.06 V, the Te substrate was almost completely dissolved, leading to the rapid decrease in current density just beyond this point. The current density increased again at −1.7 V, but this was primarily due to H2 gas evolution (Reaction 7) (Pourbaix, 1966). E0 values of reactions 6–12 were calculated at pH of 12.5.</p><!><p>LSV curve of TeO32- in alkaline solutions at a pH of 14.7. The black solid curve is at a TeO32- concentration of 550 mM (a), and the red dashed curve is at a TeO32- concentration of 0 mM (b). Thick Te films were used as substrate.</p><!><p>The reduction reaction of TeO32-(IV) was further investigated by EQCM (Figure 2). Figures 2A–C show the change in charge and mass as a function of applied potential at different pH where the curve (a) represents the charge change and curve (b) represents the mass change. At a pH of 10.2 (Figure 2A), the mass increased monotonically as electrical charge increased, which meant Te was deposited onto the working electrode continuously. The mass change rate (Figure 2D, curve b) started to increase at −0.66 V and kept increasing monotonically as a function of applied potential until the voltage reached −0.86 V. The mass change rate started to fluctuate past the critical point because two reactions took place simultaneously in addition to Te deposition: hydrogen gas evolution and Te dissolution. In comparison to the Te dissolution observed at −0.97 V (Figure 1, curve b), the hydrogen gas evolution was probably the major reason for this fluctuation. Furthermore, the mass change rate remained positive after the critical point, meaning the rate of Te deposition was faster than that of Te dissolution. At a pH of 12.5 (Figure 2B), the mass of deposits increased with charge in the forward scan (i.e. −0.73 to −1.92 V). According to Figure 2E, the dm/dt started to increase at −0.73 V. At low overpotential, dm/dt (Figure 2E) increased monotonically when applied potential became more negative. At this range, Te(0) dissolution might have happened as well, but the rate of Te(0) dissolution was slower than the deposition rate. When the applied potential further decreased from −1.92 to −2.0 V or reversed from −2.0 to −1.66 V, the mass of Te deposits decreased. As shown in Figure 2E, the dm/dt started to decrease from less negative potentials than −1.92 V and then reached negative values as the dissolution rate exceeded that of the deposition. During the reverse scan, when applied potential is more positive than −1.66 V, the mass started to increase again as the Te deposition rate increased. At a pH of 14.7 (Figure 2C), the reduction reaction of TeO32-(IV) started at an applied potential of −0.92 V. The mass of Te(0) continued to increase until −1.24 V, at which it quickly decreased all the way to zero by the time applied potential reached −1.31 V, caused by the fast dissolution of Te(0). As shown in Figure 2F curve (b), the mass change rate started decreasing at less negative potentials, but the total mass of Te(0) deposit did not reach its maximum until −1.24 V because the deposition rate by reduction of TeO32-(IV) to Te(0) surpassed that of Te(0) dissolution. The curve (a) in Figure 2F also showed a current density dip at −1.34 V and then the current density decreased significantly. As shown by these observations, the formation of intermediate species [i.e., Te22-(-I)] and its behavior as a function of pH differed the competition of dissolution and deposition rate. In alkaline solution, the reduction of Te(0) results in two possible species (i.e., Te22- and Te2−). These two products would react with TeO32-(IV) (reactions 11 and 12), inhibiting its transport to the electrode surface and instead produce black Te particles in solution. The presence of such particles was observed right after re-exposing the gold substrate (Shinagawa et al., 1972; Wu et al., 2014). The curve (b) in Figure 2F showed that the mass change rate (dm/dt) increased slightly and then rapidly decreased, leading to a peak at −1.29 V, which meant that, in high pH, the rates of Te deposition and dissolution were affected dramatically as a function of overpotential. Thus, according to the EQCM data obtained at pH 14.7, the applied potential should be controlled from −0.92 to −1.31 V to deposit the Te on the surface of the working electrode. When the applied potential was from −1.31 to −2.00 V, the Te(0) was formed in the solution as a form of particles.</p><!><p>(A–C) are charge curve (a) calculated from cyclic voltammogram and cyclic gravimetric curve (b) measured by EQCM at different pH: (A) 10.2, (B) 12.5, and (C) 14.7. (D–F) are cyclic voltammogram (a) and mass change rate curve (b) measured by EQCM at different pH: (D) 10.2, (E) 12.5, and (F) 14.7. The experiments were conducted with 300 mM TeO32- and a scan rate of 50 mV/s at 23°C. The working electrode was commercialized Au/Cr crystal.</p><!><p>Based on the EQCM data, Δm was plotted as a function of ΔQ, shown in Figure 3, and the slope (S) (Equation 4) for each curve was extracted to calculate the number of electrons involved in the reaction. To minimize the effect of secondary reaction, slopes were extracted based on the lower region of the curve. According to Equation (5), M/z was calculated to be 26.8, 35.5, and 1.9 g mol−1 at pH of 10.2, 12.5, and 14.7, respectively. Using the atomic number of Te(MTe) of 127.6 g mol−1, z at pH 10.2 and 12.5 was 4.7 and 3.6, respectively. From this calculation, it can be deduced that a four-electron reduction (Reaction 6) occurred and caused the increase of mass. Saloniemi et al. reported the M/z value of 22.8 g mol−1 in the solution of 1 mM TeO32- and 100 mM Na(CH3COO) at a pH of 9 on gold electrode, and they claimed that the reaction was a four-electron reduction of TeO32- to solid Te (Saloniemi et al., 2000b). However, at a pH of 14.7, the z value was much larger than 4, which might be attributed to the fact that the dissolution of Te (Reaction 8 or 9) started shortly after Reaction 6, and thus the relationship between Δm and ΔQ (Figure 3C) was not only reflecting Reaction 6, but also Reaction 8 or 9 or even hydrogen gas evolution.</p><!><p>Mass change as a function of charge change, obtained from the mass variation profiles presented in Figure 2 at pH: (A) 10.2, (B) 12.5, and (C) 14.7, with 300 mM TeO32-.</p><!><p>The reduction reaction of Te(0) to Te22-(-I) or Te2−(-II) was also studied using EQCM at different pH (10.2, 12.5, and 14.7) (Figure 4). Before the EQCM experiments, a thin film of Te with a mass of 70 ± 3.5 μg/cm2 was electrodeposited onto the working electrode, and no TeO32- ions were present in the solution. The mass change was therefore negative, caused by the dissolution of Te(0) to Te22-(-I) or Te2−(-II), which were solvable species. At a pH of 10.2 (Figure 4A), the mass of Te(0) started to decrease at −0.97 V during the forward scan, but the Te thin film was not completely dissolved even when the applied potential reached −2.0 V. The mass change rate (dm/dt) in Figure 4D was relatively small (−0.47 μg cm−2 s−1 at maximum) compared to the rate at a pH of 12.5 (−13.0 μg cm−2 s−1 at maximum) and 14.7 (−34.1 μg cm−2 s−1 at maximum). This suggested that OH− or Na+ in the solution influenced the reaction rate significantly. At a pH of 12.5 (Figure 4B), the mass started to decrease at −1.0 V. When the applied potential was −1.54 V, the mass change was about −70 μg cm−2, which meant that all the pre-deposited Te thin film was dissolved. Figure 4E showed that the mass change rate (dm/dt) decreased to zero at −1.54 V, confirming there was no more Te(0) available to dissolve from the working electrode. The further increase in charge at more negative potentials than −1.54 V was probably caused by hydrogen gas evolution (Reaction 7) and possibly Reaction 10. At a pH of 14.7 (Figure 4C), the pre-deposited Te thin film started to dissolve at −0.95 V and was completely dissolved at −1.19 V. The mass change rate (dm/dt) at this pH was the highest of the three conditions, and it decreased to zero at −1.19 V (Figure 4F).</p><!><p>(A–C) are charge curve (a) calculated from cyclic voltammogram and cyclic gravimetric curve (b) measured by EQCM at different pH: (A) 10.2, (B) 12.5, and (C) 14.7. (D–F) are cyclic voltammogram (a) and mass charge rate curve (b) measured by EQCM at different pH: (D) 10.2, (E) 12.5, and (F) 14.7. The experiments were conducted with 0 mM TeO32- and a scan rate of 50 mV/s at 23°C. The working electrode is commercialized Au/Cr crystal coated with 70 ± 3.5 μg cm−2 Te thin film.</p><!><p>Based on the EQCM data, the Δm vs. ΔQ relationship was plotted in Figure 5. The M/z was calculated using the slope (S) extracted from each curve. The M/z values were 105.5, 126.3, and 76.9 g mol−1 at pH of 10.2, 12.5, and 14.7, respectively. The z values were then calculated using MTe to be 1.21 and 1.01 at pH 10.2 and 12.5, respectively, which confirmed that the dissolution of Te(0) at pH 10.2 and 12.5 was a one-electron reduction of Te(0) to Te22- (Reaction 8). At a pH of 14.7, the z value was 1.66, so both Reaction 8—one-electron reduction—and Reaction 9—two-electron reduction—were possible. Based on the analysis in our previous publication, Te(0) was first reduced to Te22-(-I) and then further reduced to Te2−(-II) in alkaline solution (Wu et al., 2015). Lingane and Niedrach also reported that Te(0) was reduced Te22-(-I) first, then to Te2−(-II) (Lingane and Niedrach, 1949). At a pH of 14.7, these two reactions might have occurred sequentially. Although it was difficult to monitor the color change (Te22- is purple, and Te2− is colorless) due to the EQCM cell setup, the solution was nearly colorless after the reaction finished. Additionally, at a lower pH of 10.2 and 12.5, the Te22- ion would be relatively unstable so that it would split to Te(0) and Te2−(-I) by a disproportionation reaction. Therefore, it would be hard to be recorded as charge. Instead, this was confirmed by black deposits in the solution. As a result, the reduction of TeO32-(IV) in alkaline solution would go through three steps: (1) Te(IV) reduced to Te(0), (2) Te(0) reduced to Te22-(-I), and (3) Te22-(-I) further reduced to Te2−(-II) (Lingane and Niedrach, 1949; Wu et al., 2015).</p><!><p>Mass change as a function of charge change, obtained from the mass variation profiles presented in Figure 4 at pH: (A) 10.2, (B) 12.5, and (C) 14.7, without TeO32-.</p><!><p>The onset potential was extracted from the LSV curves shown in Figures S1–S3. Figure S4 shows that the onset potential of TeO32-(IV) to Te(0) became more positive when the concentration of Te(IV) ion (TeO32-) increased, but became more negative as a function of pH (Figure S5). Additionally, the onset potential of Te(0) to Te22- became more positive as a function of pH (Figure S6). For example, the onset potential of TeO32-(IV) to Te(0) at pH of 12.5 and 14.7 are about −0.71 V and −0.85 V, respectively. The onset potential of Te(0) to Te22- at pH of 12.5 and 14.7 are about −1.02 V and −0.93 V, respectively. Therefore, the window to possibly deposit compact Te film at pH of 12.5 and 14.7 is 0.31 V and 0.08 V, respectively. Because Te22- is a solvable species in alkaline solution, once the reaction of Te(0) to Te22- happened, the morphology of the Te film would be deteriorated significantly. From this observation, the morphology of Te electrodeposits in alkaline bath would be highly dependent on the tellurite ion concentration and pH.</p><p>The Tafel slopes of reduction reactions of Te(IV) to Te(0) (Figures 6, 7) and of Te(0) to Te22-(-I) (Figure 8) were extracted from the Tafel plots (Figures S7–S9). In alkaline solution, it is known that the Te(IV) exists in the form of anion [e.g., TeO32-; Bard, 1975, Te(OH)62-; Komandenko and Rotinyan (1967b)]. According to our previous work, the direct reduction of TeO32-(IV) to Te(0) unlikely proceeded on a negatively charged cathodic electrode surface (Wu et al., 2015). Therefore, a four-step process is proposed through the sequence of Reactions 14–17. Reaction 13 shows that the anions Te(OH)62- are in equilibrium with Te(OH)3+ cations. Once the Te(OH)3+ is adsorbed on the cathodic electrode surface, it obtains an electron from the electrode and is then discharged as Te(OH)3. This cyclic mechanism of picking up an electron from the electrode is repeated in Reactions 15 and 16, yielding Te(OH)32-. The Te(OH)32- obtains a fourth and last electron in Reaction 17 and releases all three hydroxide ions to form Te element. According to the Tafel equation, by assuming a charge transfer coefficient of 0.5, the Tafel slopes are 118, 59, 39, and 30 mV/decade when Reactions 14, 15, 16, and 17 are the rate-limiting steps, respectively (Bard and Faulkner, 2001). Komandenko and Rotinyan reported that under their experimental conditions, Reaction 16 was the rate-limiting step, and the Tafel slopes ranged from 40 to 45 mV/decade (Komandenko and Rotinyan, 1966). Furthermore, they claimed that at low alkali concentration, the Te(OH)3+ would directly obtain two electrons and become Te(OH)3-. Thus, they replaced Reactions 14 and 15 with one reaction (Komandenko and Rotinyan, 1967a).</p><!><p>Tafel slope of TeO32- reduction reaction in alkaline solutions with different [TeO32-]: 50, 100, 300, 400, and 550 mM. The experiments were conducted using Te as a substrate at pH of 12.0 and temperature of 23°C.</p><p>Tafel slope of TeO32- reduction reaction in alkaline solutions with different pH: 10.2, 11.0, 12.2, 12.5, 13.1, and 14.7 (calculated value). The experiments were conducted using Te as a substrate at [TeO32-] of 550 mM and a temperature of 23°C.</p><p>Tafel slope of Te(0) reduction reaction in alkaline solutions with different pH: 10.2, 12.5, and 14.7 (calculated value). The experiments were conducted using Te as a substrate at [TeO32-] of 0 mM and a temperature of 23°C.</p><!><p>Figure 6 shows that when the Te precursor concentration increased from 50 to 550 mM at a pH of 12.0, the Tafel slopes ranged from 90 to 117 mV/decade. The differences between Tafel slopes were attributed to changes in the charge transfer coefficient. This meant that the TeO32- concentration did not influence the rate-limiting step substantially. In the whole concentration range, the rate-limiting step was Reaction 14, which was the discharge of Te(OH)3+.</p><p>Figure 7 indicates that at low pH (i.e., 10.5), the Tafel slope was 61 mV/decade, which meant that Reaction 15 was the rate-limiting step. However, when the pH was varied between 11.0 and 12.5, the Tafel slopes of the reaction ranged from 92 to 103 mV/decade, which correlated to Reaction 14 as the rate-limiting step. When pH increased to 13.1, the Tafel slope shifted to 65 mV/decade, making Reaction 15 the rate-limiting step. When pH further increased to 14.7, the Tafel slope was further reduced to 39 mV/decade, indicating that Reaction 16 was the rate-limiting step.</p><p>The Tafel slopes for reduction of Te(0) to Te22-(-I) are shown in Figure 8. At pH of 10.2, 12.5, and 14.7, the Tafel slopes were about 121, 121, and 38 mV/decade, respectively. The single Tafel slope of Te(0) to Te22-(-I) was first reported to be 100 mV/decade at low NaOH concentration by Awad (1961). At high NaOH concentrations, two linear regions were observed with Tafel slopes of about 40 and 100 mV/decade. They proposed a two-step reaction mechanism, which involved Na+ ions (Reactions 18 and 19). However, the calculated number of electrons transferred for each elementary reaction was surprisingly two electrons for Reaction 18 and four electrons for Reaction 19, which could not be explained by this two-step reaction mechanism.</p><p>Therefore, Awad proposed a four-step reaction mechanism (Reactions 20–23) and introduced diatomic tellurium. In the reactions, M+ represented the cations (e.g., Na+ or Ba2+) in the alkaline solution. Reactions 20 and 21 were discharge reactions; Reactions 22 and 23 were electrochemical reactions. In the two discharge reactions, the reaction rate of the second step (Reaction 21) was relatively slower than that of the first step, which could be the rate-limiting step. On the other hand, in the two electrochemical reactions, the reaction rate of the second step (Reaction 23) was relatively slower than that of the first step, which could be the rate-limiting step. Awad used the energy barrier diagram of these four reactions to prove their assumptions. According to the energy barrier diagram, the first discharge (Reaction 20) and electrochemical (Reaction 22) steps had very low energy barriers compared to the second discharge (Reaction 21) and electrochemical (Reaction 23) steps. When Reaction 21 was the rate-limiting step, the Tafel slope was about 100 mV/decade, which was consistent with the theoretically calculated value of 120 mV/decade, taking into account the mechanism and the differences due to small changes in the charge transfer coefficient. When Reaction 23 was the rate-limiting step, the Tafel slope was about 40 mV/decade at low current density and 100 mV/decade at high current density. Therefore, in weakly alkaline solutions, the second discharge reaction (Reaction 21) was the rate-limiting step throughout the whole current density range with a Tafel slope of 100 mV/decade, while in strong alkaline solutions, the second electrochemical reaction (Reaction 23) was the rate-limiting step with a Tafel slope of 40 mV/decade at low current density and a Tafel slope of 100 mV/decade at high current density (Awad, 1962).</p><p>Another mechanism was suggested by Komandenko and Rotinyan with Tafel slopes of about 40 mV/decade for the reduction of Te(0) to Te22-(-I) at moderately alkaline solutions. Also, they proposed a different mechanism, which involved OH− ions shown in Reactions 24–27 where Reaction 25 was the rate-limiting step (Komandenko and Rotinyan, 1967b).</p><p>Figure 8 shows that at relatively weak alkaline solutions (i.e., pH ≤ 12.5), the Tafel slope was about 120 mV/decade. This could be explained by Awad's mechanism, in which the second discharge reaction (Reaction 21) was the rate-limiting step (Awad, 1962). At strong alkaline solutions (i.e., pH 14.7), the Tafel slope was 38 mV/decade at low current density, which meant that the second electrochemical reaction (Reaction 23) was the rate-limiting step. However, at high current density, the Tafel slope remained at 40 mV/decade instead of 100 mV/decade.</p><!><p>The electrochemical reduction mechanism and kinetics of TeO32-(IV) and Te(0) were investigated by LSV and EQCM. These data indicated that reduction of TeO32-(IV) to Te(0) is a four-electron reaction, while the reduction of Te(0) to Te22-(I) is a single-electron reaction. Furthermore, the applied potential must tune to deposit Te(0) on the electrode surface, and it varied as a function of pH. When the applied potential was too negative, Te(0) nanoparticles were formed in the solution instead of on the electrode surface. As the pH increased, the reduction rate of Te(0) to Te22-(-I) was accelerated. Additionally, investigation of two Te reduction reactions [TeO32-(IV) to Te(0) and Te(0) to Te22-(-I)] showed that although the concentration of TeO32- affected mass transfer and reaction rates, it did not influence the rate-limiting step. Instead, solution pH controlled the reaction mechanism. At relatively weak alkaline solutions (i.e., pH ≤ 12.5), the discharge of Te(OH)3+ was the rate-limiting step. For the reduction of Te(0) to Te22-(-I), the second discharge reaction was the rate-limiting step with the Tafel slope of 120 mV/decade. At strong alkaline solutions (i.e., pH 14.7), the Tafel slope was 40 mV/decade at low current density, which meant that the second electrochemical reaction was the rate-limiting step.</p><!><p>All datasets generated for this study are included in the article/Supplementary Material.</p><!><p>TW, JK, and NM evenly contributed to the conception or design of the work, or the acquisition, analysis, or interpretation of data for the work. Also, drafting the work or revising it critically for experimental results.</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><!><p>Funding. This research was supported by the Future Materials Discovery Program through the National Research Foundation of Korea (NRF), funded by the Ministry of Science, ICT & Future Planning (NRF-2016M3D1A1027836) and Korea Research Fellowship Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Science and ICT (2015H1D3A1066157).</p><!><p>The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2020.00084/full#supplementary-material</p><!><p>Click here for additional data file.</p>

PubMed Open Access

Reversible Oxygenation of α-Amino Acid–Cobalt(II) Complexes

We systematically investigated the reversibility, time lapse, and oxygenation-deoxygenation properties of 15 natural α-amino acid–Co(II) complexes through UV-vis spectrophotometer, polarographic oxygen electrode, and DFT calculations, respectively, to explore the relationship between the coordinating structure and reversible oxygenation of α-amino acid–Co(II) complexes. Results revealed that the α-amino acid structure plays a key role in the reversible oxygenation properties of these complexes. The specific configuration of the α-amino acid group affects the eg 1 electron of Co(II) transfer to the π ⁎ orbit of O2; this phenomenon also favors the reversible formation and dissociation of Co–O2 bond when O2 coordinates with Co(II) complexes. Therefore, the co-coordination of amino and carboxyl groups is a determinant of Co complexes to absorb O2 reversibly. The group adjacent to the α-amino acid unit evidently influences the dioxygen affinity and antioxidation ability of the complexes. The presence of amino (or imino) and hydroxy groups adjacent to the α-amino acid group increases the oxygenation-deoxygenation rate and the number of reversible cycles. Our findings demonstrate a new mechanism to develop reversible oxygenation complexes and to reveal the oxygenation of oxygen carriers.

reversible_oxygenation_of_α-amino_acid–cobalt(ii)_complexes

1. Introduction<!>2.1. Materials<!>2.2. UV-vis Spectrophotometry<!>2.3. Mass Spectrometry<!>2.4. Construction of the Absorption (A)-pH Curves<!>2.5. Determination of the Reversibility of Dioxygen Uptake and Release<!>2.6. Oxygenometry<!>2.7. DFT Calculation<!>3. Results and Discussion<!>3.1.1. UV-vis Spectra<!>3.1.2. IR Spectroscopy<!>3.1.3. Mass Spectrometry Analysis<!>3.2. Determination of the Suitable pH Condition for the Formation of Each α-Amino Acid–Co Complex<!>3.3. Determination of Reversibility for the Uptake and Release of Dioxygen<!>3.4. Oxygenometry Method<!>3.5. Dynamics of Oxygenation-Deoxygenation of the α-Amino Acid–Co Complexes<!>3.6. Comparative Study<!>4. Discussion<!>5. Reversible Oxygenation Mechanism<!>6. Conclusions<!>

<p>Oxygenated complexes should be investigated whether as a model compound of natural oxygen carriers or as an environmentally friendly catalyst [1–6]. The oxygenation, related mechanism, and configuration of oxygenated complexes as a model compound of natural oxygen carriers have been extensively explored [7–13]. Nam synthesized crystals of mononuclear oxygenated complexes and speculated their oxygenation mechanism by systematically investigating their aging process, that is, the process of activation of dioxygen and then the oxidation of the oxygenated complex [1, 14–17]. For example, metalloenzymes activate dioxygen to perform various biological reactions [8, 10, 18–20]. The activation of dioxygen at enzyme active sites occurs through several steps: (1) O2 binds to a reduced metal center; (2) superoxo and peroxo species are then generated; the O–O bond of metal hydroperoxo complexes is cleaved; and high-valent metal-oxo oxidants are formed [2, 3, 21–26]. Studies have demonstrated the mechanisms by which biological enzymes activate oxygen molecules. However, reversible oxygenation complexes, which can be used as a model compound of hemoglobin, have been rarely reported because available oxygenation complexes are unstable; as a result, researchers experience difficulty in characterizing the exact structures of products [27, 28]. Two major problems are encountered in studies involving the oxygenation of cobalt complexes in an aqueous solution: (1) complexes absorb dioxygen immediately when they are formed under ambient conditions and (2) aging phenomenon or the oxidation of the oxygenated complex occurs until the aging process is completed [29, 30]. Although the rates of the aging or oxidation of these cobalt complexes differ, oxygenated complexes exist in varying states. As such, oxygenated compounds should be characterized immediately after these compounds are formed in solutions rather than after they become separated from solutions. UV-vis spectrophotometer and polarographic oxygen electrode are used to monitor oxygenation and deoxygenation online through the continuous alternation of O2 and N2 atmospheres. A mass spectrometer is utilized to characterize complexes and to determine their structures [29–31].</p><p>Schiff base ligands and porphyrins are mostly confined in the reversible oxygenation of oxygenated complexes [11, 24]; however, the reversible oxygenation property of these complexes is poor in aqueous solutions and at room temperature. Burk et al. investigated histidine–Co(II) and demonstrated that oxygenation reversibility occurs in an aqueous solution at room temperature [32]. Considering this previous study, Martell and other researchers examined Co(II) complexes with amino acids and dipeptides, which can reversibly uptake oxygen [33, 34]. In our study, four groups of ligands with different N/O, N/N, and O/O difunctional groups [35], such as amino acids, amino alcohols, polyamines, and multicarboxylic acids, were selected. The oxygenation performances of their Co(II) complexes were comparatively analyzed. The following results were obtained: the O/O difunctional group of the Co(II) complexes does not possess oxygenation properties; the N/N difunctional group of the complexes can uptake oxygen but cannot undergo reversibility; the N/O-type difunctional group of the complexes is relatively different; α-amino acid–Co(II) complexes exhibit evident reversible oxygen performance; and α-amino alcohol–Co(II) complexes uptake oxygen weakly and show no reversibility. However, studies have yet to determine the factor causing α-amino acid–Co complexes to exhibit oxygenation reversibility.</p><p>Among Co complexes, α-amino acid–Co complexes display different oxygenation reversibilities. On the basis of this unusual finding, we supposed that the α-amino acid group in a ligand could be a key group associated with the reversible oxygenation of such complexes. Therefore, we synthesized a series of novel amino acid–Co(II) complexes and found that Co complexes can reversibly uptake dioxygen [28, 35, 36]. Indeed, the α-amino acid unit in a ligand is a key structural component related to the reversible oxygenation property of these complexes.</p><p>To explore the factors and coordinating structure that determine the oxygenation reversibility of α-amino acid–Co complexes, we selected 15 natural water-soluble α-amino acids (Scheme 1) and performed a detailed investigation of their oxygenation properties in aqueous solutions at room temperature. Measurements were conducted at a specific pH range at which these complexes are in their major states. We also conducted a thorough comparative analysis to determine the relationship between the structures of ligands and the dioxygen affinity of Co complexes.</p><p>This study aimed to determine the basic structural unit responsible for the reversible dioxygen uptake. This study also aimed to identify the auxiliary functional groups that improve the reversibility of dioxygen uptake. We believe that our findings would remarkably contribute to elucidating and revealing the oxygenation mechanism of oxygen carriers.</p><!><p>Amino acids and Co(CH3COO)2·4H2O were purchased from Shanghai Aladdin Reagents Co., Ltd. (China). Amino acids and Co aqueous solutions were prepared with distilled water. High-purity (99.99%) O2 and N2 were used. All of the chemicals were used without further purification.</p><p>The following instruments were used in the experiment: UV-vis spectrometer (UV-2450, Shimadzu, Japan), infrared spectrometer (VERTEX70-RAMAN II, Bruker Company, Germany; test conditions: dpi: 4.0 cm−1, number of scans: 100, and ATR with water as background), peristaltic pump (PS19-2, Pgeneral, China) (PP2), portable dissolved oxygen meter (HI 9146, Hanna Instruments, Italy), and pH Meter (PHS-3C, Shanghai Shengci Instrument Co., China).</p><!><p>Table 1 provides a list of the concentrations of the α-amino acids and Co(II) to prepare the corresponding complex solutions. The spectra were recorded at 25.0 ± 0.1°C by using a UV2450 spectrophotometer with a 1 cm cuvette within the spectral range of 200–600 nm or at the maximum absorption peak (λ max) of each complex at a certain pH.</p><!><p>Mass spectrometry was performed with Waters Quattro Premier XE mass spectrometer equipped with an electrospray ionization source (Micromass, Manchester, UK).</p><!><p>For each amino acid, the Co(II) solution with a known concentration was mixed with the α-amino acid solution at a molar ratio of 1 : 3 or 1 : 2, depending on the α-amino acid species (Table 1). The A versus pH curve was constructed in accordance with a previously described method [28], and the suitable pH ranges to examine each complex were selected from the A-pH curves.</p><!><p>The oxygenation and deoxygenation kinetics were determined using a PP2 flow injection apparatus [28]. The reversibility of the oxygenation and deoxygenation of the α-amino acid–Co(II) complexes was identified by recording the changes in the absorbance of O2 and N2 saturated solutions. The absorbance difference (ΔA = A O − A N) between the absorbances in O2  (A O) and N2 atmospheres (A N) was considered to evaluate the ability of the complexes to uptake O2. The number of oxygenation-deoxygenation cycles (C) was obtained to estimate the endurance of each complex to antioxidation.</p><!><p>The concentration of the dissolved O2 in the solution corresponded to the evolution of the oxygenation of the complexes. The concentration of the dissolved O2 in the solution was measured in accordance with a previously described method [28].</p><!><p>Calculations were performed with the Gaussian 03W program package [37]. Full geometry optimization computations were conducted via a B3LYP method. In all of the calculations, a LANL2DZ basis set, along with the corresponding effective core potential, was used for Co metal atoms. The 6-31G(d) basis set was utilized for C, H, N, and O atoms.</p><!><p>The oxygenation and reversible performances of the 15 α-amino acid–Co complexes were investigated in this paper. The Ala–Co complex was used as an example to demonstrate the experimental processes.</p><!><p>Figure S1A (in Supplementary Material available online at http://dx.doi.org/10.1155/2016/3585781) shows the UV-vis spectra of Ala and Co(II) salt solution alone and their mixtures. The spectra of the mixture are distinctly different from those of Co or ligands alone; this result confirmed that the complexes were formed. Similar results were observed in the other amino acid–Co systems, as indicated by the UV-vis spectra. Figure S1 presents the UV-vis spectra of Ser–Co, His–Co, and Lys–Co.</p><!><p>Figure S2 shows the IR spectra of L and Ala–Co(II), Ser–Co, His–Co, and Lys–Co. Clearly, the spectra of these complexes are significantly different from those of amino acids alone, which refers to the formation of the complexes.</p><!><p>All amino acid–Co complexes were determined by MS, and the results exhibited the formation of the complexes. Figures S3 and S4 present the ESI mass spectrum of His–Co and Ala–Co.</p><!><p>Co(II) complexes could be formed at different pH because of the differences in the coordinating abilities of amino acids to Co(II). A-pH curves of all the complexes were recorded by UV-vis spectrophotometry according to the part of experiment. The suitable pH for each complex was selected according to these A-pH curves. Table 2 lists the suitable pH for the formation of oxygenated complexes and λ max.</p><p>Figure S5 presents the figure of A-pH curves for Ala–Co (curve 1), Ser–Co (curve 2), His–Co (curve 3), and Lys–Co (curve 4). The suitable pH for the formation of these complexes was concluded to be as follows: 9.3–1.2 (Ala–Co), 9.5–9.8 (Ser–Co), 7.8–10.3 (His–Co), and 10.3–10.8 (Lys–Co); their compositions were also determined via molar ratio method in corresponding pH, and their formulae are Co(Ala)3, Co(Ser)2, Co(His)2, and Co(Lys)3, respectively. Other amino acid–Co systems were tested in the same method.</p><!><p>In the N2 atmosphere, when Ala and Co(II) salt solutions were mixed at pH 9.5, the Ala–Co complex solution showed a distinct spectrum with two main absorption peaks at 365 and 540 nm in an aqueous solution, thereby indicating the formation of the complex (Figure 1(a), curve 1). When dioxygen was added, the absorption intensity of the Ala–Co complex increased abruptly at 365 nm (Figure 1(a), curve 1′), and the color of the solution rapidly changed from light pink to orange-yellow, hence indicating that Ala–Co could be easily oxygenated in an aqueous solution at room temperature. This spectral change is caused by the charge transfer from oxygen to Co(II) (LMCT) [24]. When the atmosphere was changed from dinitrogen to dioxygen and subsequently back to dinitrogen (defined as one cycle), the spectrum changed regularly according to the change of the gas atmosphere (Figure 1(a), curves 2 and 3 for N2 and 2′ and 3′ for O2). These results confirmed that the oxygenation-deoxygenation reactions of Ala–Co are reversible. Other α-amino acid–Co complexes were tested in the same manner. The spectral changes of Ser–Co, His–Co, and Lys–Co displayed that the oxygenation of these three cobalt complexes was reversible (Figures 1(b), 1(c), and 1(d)). The spectra of other cobalt complexes dropped evidently after three cycles, but the spectra of His–Co remained the same after 15 cycles. These results showed that autoxidation occurred during oxygenation, and His–Co had an excellent reversible oxygenation ability. Table 2 provides the results of reversibility for all the 15 amino acid complexes.</p><!><p>Evolution of the dissolved dioxygen concentration in airtight complex solutions was examined using a dissolved oxygen meter within the pH range of 3–11 and subsequently from 11 back to 3 at 25.0 ± 0.1°C. Figure S6 exhibits the diagrams of the dissolved dioxygen concentration as a function of pH (from 3 to 11 and back to 3) of the Ala–Co, Ser–Co, His–Co, and Lys–Co complexes. These concentration curves of the dissolved dioxygen in oxygenation coincided well with those of deoxygenation, thus indicating that the oxygenation of the complex is reversible.</p><p>The reversibility of oxygenation for other α-amino acid–Co complexes was also examined in the same procedure, and the results are listed in Table 2. The results obtained from oxygenometry also agreed well with those obtained from UV-vis spectrophotometry.</p><p>The 15 α-amino acid complexes have affinities to dioxygen, and 14 of them could reversibly bind dioxygen at suitable pH values; however, Cys–Co could only take oxygen but did not release it (Table 2).</p><!><p>After three oxygenation-deoxygenation cycles, the Ala–Co(II) complex still maintained reversible performance. The time-dependent cycle numbers of oxygenation-deoxygenation of Ala–Co(II) were determined to elucidate the complete oxygenation process and antiaging ability of the oxygenated complex. The time elapsing changes of absorbance were recorded as N2 and O2 were alternately bubbled into the system to observe the evolution of the oxygenation species. The difference of absorbance (ΔA) under N2 and O2 in one cycle was used to identify the oxygenation ability of the complex. Figure 2 represents the oxygenation-deoxygenation kinetics of the Ala–Co(II) complex. The Ala–Co(II) complex took about 84 min to complete one oxygenation-deoxygenation cycle. Oxygenation spent about 28 min, whereas deoxygenation took twice times to complete. Kinetics results (Figure 2(a)) showed that the Ala–Co complex could sustain eight continuous oxygenation-deoxygenation cycles over 10 h in an aqueous solution at room temperature. Ser–Co sustained 27 cycles in 20 hours (Figure 2(b)), and His–Co (Figure 2(c)) and Lys–Co (Figure 2(d)) did 550 in 110 h and 20 in 45 h, respectively. Other α-amino acid–Co complexes were also examined in the same procedures, and the results are listed in Table 3.</p><!><p>All 15 α-amino acid–cobalt complexes, except for Cys, displayed reversible oxygenation properties but exhibited different affinities to dioxygen; therefore, a systematic comparative study was conducted to reveal the relationship between the structures of amino acids and oxygenation properties of complexes. Ala–Co complex only has a methyl connected with α-amino acid group; thus it was used as a reference of other α-amino acid–cobalt complexes. In the comparative study, except for reversibility, the times of oxygenation and deoxygenation and number of reversible cycles were involved as the contrast parameters. Some rules were concluded from the comparative study.</p><!><p>All α-amino acid–cobalt complexes exhibit reversible binding ability to dioxygen, except for Cys–Co(II). Generally, the oxygenation time (t o) of a complex is shorter than the deoxygenation time (t d) for almost all complexes. His–Co yields the minimum t o and t d of 1 and 7.5 min, respectively. Glu–Co and Asp–Co reach the maximum t o and t d of 93 and 100 min, respectively. The ligands of the Co complexes were arranged from the shortest to the longest on the basis of the duration dioxygen uptake saturation (t o: min): His (1), Ser (17), Thr (18), Gly (22), Ala (28), (Pro = Arg) (33), Val (35), Met (37), Lys (42), Asn (50), Asp (83), Gln (87), and Glu (93). Likewise, the ligands of the Co complexes were arranged from the shortest to the longest on the basis of the duration of the complete dioxygen release (t d: min): His (7.5), Ser (17), Thr (25), Pro (33), Val (50), Ala (56), Met (57), (Arg = Lys) (58), Asn (67), Gly (75), Glu (87), Gln (92), and Asp (100). The His–Co complex requires much less time than the other complexes do in the oxygenation-deoxygenation process; this finding suggested that the His–Co complex is an excellent model of oxygen carriers. According to the theoretical calculation results, t d is usually longer than t o of these complexes when an H bond forms between ligands and when O2 binding occurs.</p><p>Another important characteristic parameter to evaluate the oxygenation property of a complex is the number of oxygenation-deoxygenation cycles. Our results suggest that His–Co has the maximum cycle number of 550, whereas Gly displays only 2 cycles. The ligands of the Co complexes were arranged from the highest to the lowest depending on whether they could sustain 5% to 100% of the original oxygenation capacity: His (550), Pro (40), Arg (33), Ser (27), (Glu = Gln) (24), (Val = Lys) (20), Thr (17), Met (16), Asp (12), Asn (11), Ala (8), and Gly (2).</p><p>Tables 2 and 3 reveal the results of the comparative analyses of one cycle time (t T, t T = t o + t d) and cycle numbers (C) of 14 α-amino acid–cobalt complexes.</p><p>Ala–Co complex took 84 min to complete one oxygenation-deoxygenation cycle. This complex could also sustain eight reversible cycles (Table 3). Furthermore, Gly–Co, Val–Co, and Pro–Co (97, 85, and 66 min, resp.) showed a similar cycle time to that of Ala–Co. All of these amino acids have a similar alkyl radical to Ala; thus all these complexes have a similar coordinating structure. Nevertheless, the cycle numbers are decreased in the order of the decrease of numbers of carbon atoms in alkyl chain.</p><p>The oxygenation properties of cobalt complexes of His, Ser, and Thr are improved evidently when compared with Ala–Co; these complexes have cycle times of 8.5, 34, and 43 min, as well as reversible cycle numbers of 550, 27, and 17, respectively. In this study, we suppose that this improvement is because they have a heteroatomic group adjacent to their amino acid group. The presence of one more atom from the heteroatomic group (NH2 or OH) that coordinates with amino acid, together with Co(II), is helpful to form the complexes and enhance the oxygenation ability.</p><p>The Cys–Co complex could bind to dioxygen but shows no reversibility, although it has also one more coordinating atom; this observation is probably because S atom is larger and more basic compared with N and O atoms, which increased the electron density between the metal ions and molecular oxygen, thereby increasing the bond strength of Co–O2 and making it more difficult to release dioxygen. This result is consistent with the report that the coordination ability will be modestly increased for metal complexes when a ligand contains S group [38].</p><p>In contrast to the Cys–Co complex, the Met–Co complex showed reversible oxygenation patterns similar to that of Ala–Co, with nearly the same time for one oxygenation-deoxygenation cycle (95 min and 16 reversible cycles). This finding may be caused by the fact that the aliphatic S atom cannot coordinate with Co(II) because it is far from the amino acid group; instead, only the α-amino acid unit of Met can coordinate with Co(II), and it behaves the same as Ala does. However, the aliphatic S plays a role in the resistance of the complex to autoxidation by increasing its number of reversible cycles to 16.</p><p>Arg and Lys have another –NH2 group that could act as a potential coordinating group for the ligands. However, –NH2 is far from the α-amino acid unit as in Met–Co; hence the coordination between the amino acid and Co(II) is much weaker. Thus, the dioxygen affinity of the Co(II) complexes for Arg and Lys is similar to that of Ala–Co, and their cycle times are also almost the same as Ala–Co. The resistance to autoxidation of Arg–Co and Lys–Co is improved by their –NH2 group, and the numbers of their reversible cycles increased to 33 and 20, respectively.</p><p>Glu and Gln have additional –CONH2 and –COOH groups in their structures, respectively, and they exhibit similar oxygenation abilities. The second carbonyl in Glu and Gln can be used as coordination group; however, it does not coordinate with its amino acid group, together with the same Co(II) ion. Instead, this group coordinates with another Co(II) ion to form linear macromolecule during the formation of the Co(II) complexes. Therefore, the times for one oxygenation-deoxygenation cycle of Glu and Gln are extended to 180 and 179 min, respectively, and the numbers of their reversible cycles are also improved to 24.</p><p>Asn also contains one more –CONH2 group, but it possesses one –CH2 group less than Glu in its chain. The carbonyl groups in Asn can promote its coordination with the Co(II) ion. Hence, one cycle time of Asn is 117 min, which is faster than that of Glu. Asp also behaves as Gln, and its one cycle time remains at 183 min, which is nearly the same as that of Gln. Nevertheless, the reversible cycle numbers of both Asp and Asn are retained at 11 and 12, respectively.</p><p>All of above 14 oxygenated α-amino acid–cobalt complexes have UV-vis absorption. The characteristic absorption peaks of all aliphatic α-amino acid–cobalt complexes are similar to one another and appeared at about 365 nm, thereby revealing that oxygenated species of these α-amino acid–cobalt complexes have considerable similarity in terms of their coordinating structures and patterns.</p><p>Some differences exist between the oxygenated complexes of His–Co and Pro–Co and other oxygenated complexes, and the UV-vis absorption peaks appeared at 374 and 380 nm, respectively. The result is attributed to the fact that both of these amino acids have an aromatic ring that can stabilize the complexes and make the absorption band shift to red waves. Based on these results, the absorption peak at 365 nm would be a characteristic absorption peak for the oxygenated species of the complex. In such case, the UV-vis spectra at 365 nm could be used to characterize the aliphatic oxygenated complexes.</p><p>On the basis of the comparative studies, we proposed that the α-amino acid group is the basic unit responsible for the reversible oxygenation properties in these 14 complexes. Other functional groups can also affect the rate, cycle times, and other oxygenation properties. The rates and cycle times of the reversible oxygenation process are mainly determined by the coordination ability of amino acids. Groups, such as imidazole in histidine that can cooperate with the α-amino acid to form more stable complexes with Co(II), will cause the His–Co complex to exhibit faster oxygenation and deoxygenation rates.</p><p>The presence of additional coordinating groups in the amino acids may also affect the oxygenation abilities of the complexes. The presence of –NH2 (or –NH) and –OH at a position adjacent to the α-amino acid unit could increase both oxygenation-deoxygenation rates and number of reversible cycles. The heteroatom group linked with the chain of the α-amino acid can inhibit oxidation and increase the number of cycles.</p><!><p>The DFT calculations were conducted for the structural models of the studied compounds, and the results of Ala–Co and His–Co have been reported [28, 39, 40]. Based on the theoretical calculation and experimental results, the oxygenation mechanism of the Co(II) complexes is proposed as follows.</p><p>When a complex binds to dioxygen, the d-orbitals of Co(II) are split, and the distribution of the electrons on the 3d orbital is t2g 6eg 1. For the oxygenated complexes, the energy level of eg orbitals of Co(II) is fairly close to the energy level of π ∗ orbitals for dioxygen. Therefore, the electron of eg orbitals can transfer to π ∗ orbitals of dioxygen to form the Co–O2 bond [24].</p><p>The α-amino acid–Co(II) complexes can reversibly bind to O2 depending on the co-coordination of α-amino and carboxyl groups. The electronegativity of N atom is smaller than that of O atom, and its lone pair electrons in N are closer to the central Co(II); as a result, the electron cloud density on Co(II) is increased. This phenomenon is helpful to transfer the eg 1 electron from Co(II) to the π ∗ orbit of O2 and form the Co–O2 bond when O2 coordinates with the Co(II) complex. According to the theoretical calculation, before and after oxygenation, the bond lengths of N–C, C–O, Co–O, and Co–N in the complexes shortened; for Ala–Co [39], Co–O and Co–N have 2.0889–2.266 and 2.2131–2.2137 lengths in the complex, respectively; after oxygenation, Co–O and Co–N become 1.941–1.984 and 2.1368–2.2059, respectively. For His–Co [40], Co–O, Co–N, and Co–N (imidazole) are 2.2723, 2.0113 and 2.0775, 2.2847, and 1.9847, 1.9910, respectively, in the complex; after oxygenation, Co–O, Co–N, and Co–N (imidazole) become 1.9519, 1.9286, 1.9361, 2.0006, and 1.9632, 1.9580, respectively. These results showed that the α-amino and carboxyl groups have conjugation in oxygenation. The conjugation of coordinated carboxyl can make Co–O2 bond more stable where the peroxo complex forms. With the transition of the electron and conjugation, the O2 binding becomes reversible when O2 and N2 atmospheres are alternatively changed. The reversible oxygenation of the complexes would occur, as shown in Figures 3 and 4.</p><!><p>Our study revealed that the structural detail of α-amino acid plays a key role in determining the reversible oxygenation/deoxygenation ability of the complexes formed by Co(II) and amino acid. We observed that the auxiliary groups linked to the α-amino acid group can affect the affinities of the complexes to dioxygen and their abilities to undergo antiautoxidation. In particular, the presence of –NH2 (or –NH) or –OH group at a position adjacent to the amino acid unit enhances the oxygenation-deoxygenation rates and number of reversible cycles. A heteroatom group linked to the chain of the amino acid improves the resistance to oxidation and may increase the number of reversible cycles. Therefore, a reversible oxygenation mechanism of amino acid–Co(II) complexes is proposed, that is, the coaction of the strong electron donor of the amino group, and conjugation of the carboxyl group is an important phenomenon of the reversible oxygenation of these complexes. This strategy may provide a useful basis of novel oxygen carriers.</p><!><p>Supplementary Material containing supporting data of oxygenation properties of natural α-amino acid-Co complexes.</p>

PubMed Open Access

Effect of polymer charge on functional reconstitution of membrane proteins in polymer nanodiscs

Although there is a growing interest in using polymer lipid-nanodiscs, polymer charge pose limitations for studies on membrane proteins. Here, we demonstrate the functional reconstitution of large soluble-domain containing positively-charged ~57-kDa cytochrome-P450 and negatively-charged ~16-kDa cytochrome-b5 in lipid-nanodiscs, and the role of polymer charge for high-resolution studies on membrane proteins.

effect_of_polymer_charge_on_functional_reconstitution_of_membrane_proteins_in_polymer_nanodiscs

<p>Membrane proteins play a plethora of important roles in cellular function. Despite the need for high-resolution structures to fully understand the function, poor solubility and stability of membrane proteins continue to pose challenges.1 The development of lipid-nanodiscs as a membrane mimetic has enabled structural studies of membrane proteins in their near-native lipid bilayer environment.3 Nanodiscs are discoidal shaped and contain a planar lipid bilayer surrounded by an amphiphilic molecular belt.4 Different types of amphiphilic belts such as membrane scaffold protein (MSP),3, 5, 6 peptides7, 8 and polymers,9, 10 have been used to form lipid-nanodiscs. In recent years, polymer based lipid-nanodiscs have been developed and their advantages over the well-established protein based nanodiscs have been demonstrated.11–14 Polymer based lipid-nanodiscs have been shown to not only form nanodiscs with varying lipid compositions, but have also been used to directly extract membrane proteins from their native cellular membrane environment.15, 16 Styrene maleic anhydride/acid (SMA) polymers were the first amphiphilic polymers shown to form nanodiscs (known as SMALP),9, 17 and since their introduction several other polymers 10, 18 have also been synthesized to improve the nanodiscs properties. Low molecular weight SMA derivatives such as styrene maleic acid – ethanol amine (SMA-EA)19 and styrene maleimide quaternary ammonium (SMA-QA)2 were recently developed to improve the size control and pH stability of nanodiscs. Although polymer nanodiscs are increasingly used for studies on membrane proteins, they have been limited to polytopic transmembrane (TM) proteins.20 On the other hand, unlike the TM proteins, large soluble domain containing membrane proteins (such as single-pass or double-pass TM proteins) exhibit heterogeneous aggregation due to differences between soluble and TM domains. Interestingly, these are the proteins that continue to pose major challenges for structural studies; significant difference in the time scales of motions between the soluble and TM domains makes it very difficult to crystallize them and, functional reconstitution remains to be major a challenge.21 Therefore, it is important to demonstrate the capabilities of polymer nanodiscs to reconstitute such membrane proteins to broaden the applications of polymer nanodiscs.</p><p>In this study, we chose the ~57-kDa cytochrome P450 (CytP450) and ~16-kDa cytochrome-b5 (Cyt b5) to demonstrate the feasibility of using polymer nanodiscs because they contain a single TM domain, a large soluble domain, and exhibit their function in a lipid membrane environment.22, 23 The additional complexities due to highly cationic CytP450 and anionic Cyt b5 are also utilized to understand the role of polymer charge on their functional reconstitution. CytP450 and its redox partners are involved in the oxidation of a variety of substrates.24 The active site of CytP450 contains a heme moiety axially coordinated to a cysteine, and this coordination plays a major role in its function.25 When functional CytP450 is reduced and binds to carbon monoxide (CO), it displays a characteristic Soret absorption at 450 nm. CytP450 has an inactive form denoted as P420, which causes a blue shifted Soret absorption peak at 420 nm. Previous studies on using peptide or MSP nanodiscs showed the reconstituted CytP450 to be monomeric, functionally active and more stable as compared to the detergent solubilized protein.7, 26, 27 These studies enabled biophysical and structural characterization of CytP450, whereas in the absence of the nanodiscs the heterogeneous nature of CytP450 makes it difficult.28, 29 In this study, we reconstituted the functional cytochrome P450 2B4 (CYP2B4) in different polymer nanodiscs (negatively charged SMA-EA and SMALP, and positively charged SMA-QA), and demonstrated that CytP450 stability is increased when reconstituted in SMA-QA polymer nanodiscs as compared to that in a lipid-free solution (but in detergent). SMA-EA and SMA polymer nanodiscs were found to inactivate CytP450 during reconstitution. The inactivation of CytP450 was found to be dominated by charge-charge interaction between the negatively charged polymer and cationic CytP450. SMA-EA nanodiscs were found to be effective for Cyt b5 reconstitution, whereas SMA nanodiscs exhibited a strong interaction between Cyt b5 and the polymer belt. On the other hand, SMA-QA nanodiscs successfully reconstituted Cyt b5 only in the presence of a high ionic strength medium. These results demonstrate that the innate chemical properties of the chosen polymer to form nanodiscs can have a large effect on successful functional protein reconstitution.</p><p>The nanodiscs used in this study were prepared using DMPC (1,2-dimyristoyl-sn-glycero-3-phosphocholine) and a SMA, SMA-EA or SMA-QA (Figure 1a and 1b) as explained in the supporting information. SMA was commercially obtained, whereas both SMA-EA and SMA-QA were synthesized and characterized as previously described.2, 19 The dynamic light scattering (DLS) (Fig.1c–e) and size exclusion chromatography (SEC) (Fig.1f–h) profiles showed that all polymer nanodiscs exhibited a similar size (~10–15 nm diameter). More details on the stability of these polymer nanodiscs can be found in the literature.2, 9, 19 Following SEC, the nanodiscs solutions were incubated with CytP450 at room temperature for 12 hrs. The resulting CytP450-reconstituted nanodiscs were reduced using sodium dithionite and the solution was bubbled with carbon monoxide (Figure 2). Figure 2b shows the absorption spectra of ferric (Fe3+) CytP450 and CO bound ferrous (Fe2+) CytP450. The CytP450 reconstituted in SMALP absorbed at 420 nm showing that the protein is in its inactive form, similar to that observed from DPC micelles (see the experimental data in Figure S1 in the supporting information).30 CytP450 reconstituted in SMA-EA nanodiscs displayed two absorbance peaks at 420 nm and 450 nm indicating the partial inactivation. CytP450 incubated with SMA-QA nanodiscs exhibited a peak maximum at 450 nm demonstrating that the protein is reconstituted in its active form. The overall percentage of inactive CytP450 in SMA-EA nanodiscs decreased when CytP450 was reconstituted in SMA-EA nanodiscs in the presence of 500 mM NaCl (Figure 2c–d). This observation suggests that the positively charged CytP450 (+6.9 at pH 7.4)31 interacts with the negatively charged SMA or SMA-EA via electrostatic charge-charge interactions. The high affinity of positively-charged CytP450 for SMA-EA or SMA can be attributed to the significant charge density of the nanodisc's polymer-belt that contain negatively-charged carboxylate-groups (see the supporting information).</p><p>To further investigate the interaction between the polymer nanodiscs and CytP450, SEC was performed. SEC analysis (Figure S2) of the positively-charged P450 and the negatively-charged polymer (SMA-EA or SMA) exhibited a shift in the retention volume of 4.4 (for SMA-EA) or 4.6 (for SMA) ml. This observation suggests that the formation of large-size particles due to the interaction of the oppositely-charged protein and polymer-belt. On the other hand, when the protein and polymer have the same type of charges, SEC profiles revealed the formation of particles that are larger than that of empty-nanodiscs but smaller than that observed for oppositely-charged protein and polymer-belt (mentioned above). For example, the positively-charged P450 and the positively-charged polymer (SMA-QA) exhibited a shift in the retention volume of 2 ml as compared to that of empty-nanodiscs (Figure S2). This observation indicates a successful reconstitution of P450 in the positively-charged SMA-QA-polymer-nanodiscs. This finding is also in agreement with UV-vis experimental results and demonstrates a successful reconstitution of a functionally active P450.</p><p>Reconstitution of functional CytP450 in SMA nanodiscs was not possible as SMA is unstable under high NaCl concentrations; whereas salt is essential to screen the nonspecific electrostatic interactions between the polymer and protein. The complete inactivation of CytP450 by SMA is also partially due to the presence of the hydrophobic domains in SMA polymer as explained in a recently published study.32 Use of salt during SMA-QA reconstitution of CytP450 showed negligible effects on the functional reconstitution efficiency (Figure 2e–f). The ability to monomerize the full-length CytP450 and functionally reconstitute in a planar lipid bilayer using SMA-QA and SMA-EA (in the presence of salt) is remarkable. We believe that this first demonstration should enable high-resolution structural and enzymatic mechanistic studies on the full-length CytP450.</p><p>We further examined the effect of polymer charge on the reconstitution of a negatively charged ~16-kDa rabbit Cyt b5 (-8.4 at pH 7.4) which has an isoelectric point of 6.0.31 The incubation of Cyt b5 with SMA-QA nanodiscs was monitored using static light scattering (SLS) (Figure 3b). Increase of light scattering from SMA-QA nanodiscs upon the addition of Cyt b5 indicated the formation of aggregates in the sample. No change in SLS was observed by the addition of Cyt b5 to SMA-QA nanodiscs in the presence of 500 mM of NaCl, suggesting that the positively charged SMA-QA was forming aggregates with negatively charged Cyt b5 via a nonspecific electrostatic interaction; but the addition of salt suppressed the electrostatic interaction between polymer and protein to avoid aggregation. Further insights on the reconstitution of Cyt b5 and stability of resultant nanodiscs were obtained using 2D 1H-15N transverse relaxation optimized spectroscopy heteronuclear single quantum coherence spectroscopy (TROSY-HSQC) NMR experiments on uniformly 15N-labeled Cyt b5 reconstituted nanodiscs (Figure 3). As seen in NMR spectra in Figure 3g, Cyt b5 reconstituted in SMA-EA nanodiscs showed well dispersed resonances indicating that the protein is well folded as reported previously.7, 19 On the other hand, when Cyt b5 is reconstituted in the presence of the SMA polymer nanodiscs, resonances in the 2D TROSY-HSQC NMR spectrum were clustered around in the 7 to 8.5 ppm region signifying a strong interaction between the polymer and Cyt b5 (Figure 3d). The strong intermolecular interaction is most likely due to the hydrophobic domains present in SMA,32 similar to the interactions observed for CytP450 reconstitution as mentioned earlier. It is interesting that SMA-QA polymer nanodiscs containing Cyt b5 produced well dispersed resonances only in the presence of 500 mM NaCl (Figure 3f), while no resonances were observed in the presence of 100 mM NaCl (Figure 3e). This observation suggests that the negatively charged Cyt b5 and positively charged SMA-QA polymer nanodiscs form large aggregates, supporting the above-mentioned SLS results (Figure 3b).</p><p>In conclusion, we have successfully reconstituted functional CytP450 and Cyt b5 in polymer nanodiscs. We have demonstrated that CytP450's stability is increased when reconstituted in SMA-QA nanodiscs as compared to that in a lipid-free solution (or in detergent), which is similar to when reconstituted in MSP and peptide nanodiscs. Polymer-protein charge-charge interactions were demonstrated to play an important role in the functional reconstitution of proteins and on the stability of resultant nanodiscs. These interactions are attributed to the presence a high charge density on the polymer-belt of the nanodisc due to the presence of a large number of polymer molecules within a limited surface area. Positively charged SMA-EA polymer nanodiscs were found to be effective for the reconstitution of Cyt b5 which contains a large anionic soluble domain. The negatively charged SMA based polymer nanodiscs, however, were incapable of functionally reconstituting CytP450 or Cyt b5 due to the presence of hydrophobic domains in the SMA polymer, even when the charge-charge interactions were removed as in the case of Cyt b5. Thus, our study reveals that the innate chemical properties of the chosen polymer to form lipid-nanodiscs can have a large effect on the successful functional reconstitution of a membrane protein. In particular, the presence of charged residues in the large soluble-domain(s) containing single-pass (like cytochromes P450 and b5) and double-pass membrane proteins can pose challenges due to their direct interaction with the charged polymer belt. As demonstrated in this study, with a judicious choice of a polymer to form lipid-nanodiscs and salt concentration, membrane proteins can be reconstituted for studies using NMR. We expect the results reported in this study to be useful to prepare polymer based nanodiscs for studies using a variety of biophysical techniques including SAXS, cryo-EM and crystallography. While SMA based polymers exhibit unique advantages and are increasingly used for the structural studies of membrane proteins, our findings reported in this study indicate the need for neutral molecules, that can form lipid-nanodiscs, for the structural studies on the most challenging large soluble-domain containing membrane proteins. Unlike other types of nanodiscs, tolerance exhibited by SMA-QA nanodiscs over a broad range of pH and divalent metal ions can be further utilized. A complete characterization of lipid bilayer properties of polymer nanodiscs would be useful in the structural studies of reconstituted membrane proteins.</p>

PubMed Author Manuscript

A Variable Temperature Synchrotron X-ray Diffraction Study of Colossal Magnetoresistant NdMnAsO0.95F0.05

The recent discovery of high temperature superconductivity in Fe arsenides has invigorated research into transition metal pnictides. Colossal magnetoresistance (CMR) has recently been reported for NdMnAsO1-xFx for x = 0.05-0.08, with a maximum magnetoresistance achieved at low temperature (MR 9T (3 K)) = −95%). This appears to be a novel mechanism of CMR, which is as a result of a second order phase transition in field from an insulating antiferromagnet to a semiconducting paramagnet. Here we report a variable temperature synchrotron X-ray powder diffraction study of the CMR oxypnictide NdMnAsO 0.95 F 0.05 between 4 K-290 K. An excellent fit to the tetragonal unit cell with space group P4/nmm is obtained over the entire temperature range, with no change in crystal structure detected down to 4 K. A coupling of the lattice and magnetic order is observed, where subtle discontinuities in the temperature variation of a and the c/a ratio are apparent as the Nd spins order antiferromagnetically and the Mn moments reorient into the basal plane at T SR. The results suggest that very small changes in lattice parameters effect the coupling between lattice, electronic and magnetic degrees of freedom.

a_variable_temperature_synchrotron_x-ray_diffraction_study_of_colossal_magnetoresistant_ndmnaso0.95f

<!>19<!>Results and Discussion<!>Synthesis.<!>Structural Characterisation.

<p>The discovery of high temperature superconductivity in the quaternary oxypnictide LaFeAsO 1 reinvigorated research into transition metal oxypnictides. Superconductivity may be induced in 1111-type pnictides upon substituting oxygen with fluorine [2][3][4] , creating oxygen vacancies 5 or by replacing the rare earth with Th 4+ 6 , with a current maximum T c of 56.3 K achieved in Gd 1−x Th x FeAsO. The LaFeAsO parent compound crystallizes with a tetragonal unit cell in the P4/nmm space group, with a conducting Fe 2 As 2 layer situated between insulating La 2 O 2 planes 7 . A structural distortion from tetragonal to orthorhombic symmetry (space group Cmma) occurs upon cooling with antiferromagnetic (AFM) ordering below T s 8,9 . The superconducting phase appears when the structural and magnetic ground state are suppressed (e.g. by chemical doping 10 ) and can also be induced in some non-doped systems by the application of external pressure [11][12][13] . An empirical relation exists between the distortion of As-Fe-As bond angles and the onset temperature for superconductivity (T c ), as maximum values are achieved when regular tetrahedra are formed in the FeAs 4 lattice 14 , indicating a clear relationship between the crystal structure and superconductivity.</p><p>While the manganese analogues (LnMnAsO, Ln = La, Nd) are not superconducting, they have been shown to exhibit sizeable magnetoresistance (MR) between ∼150 K-380 K, with MR values of up to −24% observed at 200 K for LaMnAsO 15,16 . Magnetoresistance is defined as the change of electrical resistivity, ρ , in an applied magnetic field, H, so that MR = (ρ(H)− ρ(0))/ρ(0); materials exhibiting this effect are important for memory device and magnetic sensor applications. Further studies revealed that colossal magnetoresistance (CMR) is observed upon substituting F − for O 2− in NdMnAsO 1−x F x (x = 0.05-0.08). A maximum MR is achieved in NdMnAsO 0.95 F 0.05 at low temperature (MR 9T (3 K)) = − 95%) 17 .</p><p>In contrast to the Fe superconductors no structural distortion is apparent in NdMnAsO 0.95 F 0.05 down to 4 K 17 ; yet, neutron diffraction studies show that several magnetic transitions exist. Antiferromagnetic ordering of the Mn 2+ spins occurs at 356 K with moments aligned parallel to c, followed by ordering of the rare earth at 23 K K where Nd 3+ spins order antiferromagnetically with moments aligned parallel to the basal plane. At the same time a spin reorientation of the Mn spins occurs, as they rotate from their original alignment along the c axis into the ab plane so that by T SR = 20 K the Mn spins are also aligned parallel to the basal plane. MR is observed below 75 K and increases further below T SR . It has been proposed that the CMR in NdMnAsO 0.95 F 0.05 arises due to a hidden order parameter, resulting in competition between an AFM insulating phase and a paramagnetic semiconductor upon application of a magnetic field 17 . Furthermore, recent high pressure neutron diffraction studies revealed that the AFM ordering of Mn spins in NdMnAsO 0.95 F 0.05 are robust up to pressures of 8.59 GPa and T Mn is enhanced from 360-383 K upon applying an external pressure of 4.97 GPa 18 . NdMnAsO 0.95 F 0.05 is however shown to violate Bloch's rule which would suggest that NdMnAsO 0.95 F 0.05 is on the verge of a localised to itinerant transition 18 .</p><p>Changing the rare earth from Nd to Pr in LnMnAsO 0.95 F 0.05 has a dramatic effect on the structural, magnetic and electronic properties of the manganese materials. Variable temperature synchrotron X-ray results describe a structural transition from tetragonal to orthorhombic symmetry with space group Pmmn below 35 K in PrMnAsO 0.95 F 0.05</p><!><p>. The distortion is the result of ferromultipolar ordering of Pr spins and is associated with a sizeable negative MR (MR 7T (12 K) = − 13.4%), instead of the CMR observed in the Nd analogue 17 .</p><p>In order to further investigate the relationship between the crystal structure and electronic and magnetic properties of the CMR material NdMnAsO 0.95 F 0.05 , we have performed a variable temperature synchrotron X-ray diffraction study between 4 K and 290 K. The results demonstrate that there is no change in crystal structure within the temperature range studied in contrast to PrMnAsO 0.95 F 0.05 and the superconducting Fe analogues. However, subtle discontinuities in the a lattice parameter and c/a ratio are observed at T SR .</p><!><p>The temperature dependence of the 7 T magnetoresistance of NdMnAsO 0.95 F 0.05 is displayed in Fig. 1. As reported previously MR is observed below ∼80 K and its magnitude increases exponentially upon cooling 17 . The magnitude of the MR rises sharply below T SR so that at 4 K MR 7T = − 90%. The field variation of the MR is also displayed in Fig. 1 and reproduces previous results 17 .</p><p>The variable temperature synchrotron X-ray powder diffraction data were analysed using the Rietveld refinement method 20 and the GSAS programme 21 to determine the crystal structure. The backgrounds were fitted using linear interpolation and the peak shapes were modelled using a pseudo-Voigt function. A minor impurity phase MnAs is observed and was modelled giving a volume fraction of 1.05%.</p><p>The Rietveld refinement of high resolution synchrotron X-ray powder diffraction data collected between 4 and 290 K confirmed that NdMnAsO 0.95 F 0.05 crystallises at room temperature with the expected ZrCuSiAs-type tetragonal structure of space group P4/nmm (Fig. 2) where insulating layers of ionic (NdO/F) + are embedded between layers of tetrahedral (MnAs) − . An excellent fit to this space group is obtained at all temperatures (Fig. 3). There is no evidence of peak splitting or superstructure peaks to suggest a change in symmetry upon cooling. The refined values for lattice constants, atomic parameters, selected bond lengths and angles with corresponding agreement indices for the respective variable temperature fits to the data are found in Table 1. There is no evidence of cation or As/O anion disorder. The Nd, Mn and As occupancies refined to within ±1% of the full occupancy and were fixed at 1.0. The O and F occupancies were fixed at 0.95 and 0.05 respectively.</p><p>The temperature dependence of the cell parameters are shown in Fig. 4. A subtle anomaly is observed in the temperature variation of the a cell parameter where a change in slope is detected at 23 K (T SR ). This discontinuity is not present in the temperature variation of the c cell parameter, which exhibits a normal thermal expansion (Fig. 4 (inset)) but is apparent in the c/a ratio (Fig. 4 (inset)). It is also not present in any of the bond lengths or angles upon cooling to 4 K (Table 1). The variation of the Mn-As and Nd-O bond lengths with temperature are shown in Fig. 5. Both bond lengths decrease upon cooling. The As-Mn-As and Nd-O-Nd bond angles do not change significantly with temperature (Table 1). The subtle anomaly at T SR in the a cell parameter is not evident in the parent compound NdMnAsO 22 . However the c/a ratio does evidence a change in slope at T SR evidencing a weak coupling between the lattice and the magnetic order 22 . It appears that a stronger coupling of the lattice and magnetic order is present in NdMnAsO 0.95 F 0.05 , where changes in a and the c/a ratio are much more apparent at T SR. This stronger coupling could be a result of the lattice contraction upon substitution of F − for O 2− (a and c shrink from 4.0503(1) and 8.9150(4) to 4.0500(1) and 8.9040(4) upon increasing x from 0 to 0.05 in NdMnAsO 1−x F x 17 ). In LnMnAsO the Dzyaloshinskii-Moriya (DM) and biquadratic (BQ) exchanges between the Ln 3+ and Mn are strong and control the spin reorientation</p><p>2.3277(4) 2.3285(4) 2.3294(4) 2.3303(4) 2.3281(4) 2.3292(4) 2.3273 (5) Table 1. Refined cell parameters, agreement factors, atomic parameters and selected bond lengths and angles for NdMnAsO 0.95 F 0.05 from Rietveld fits against X-ray synchrotron data at various temperatures. Nd and As are at 2c (¼, ¼, z), Mn at 2b (¾, ¼, ½) and O,F at 2a (¾ , ¼ ,0).</p><p>transition 23 ; the BQ exchange dominates in NdMnAsO. In principle the smaller unit cell in NdMnAsO 0.95 F 0.05 could enhance magnetic exchange between Nd 3+ and Mn 2+ ions which in turn will then augment the spin-lattice coupling at T SR .</p><p>The electronic properties of NdMnAsO and fluorine doped samples, NdMnAsO 1−x F x , are also very different. Above 90 K the electronic behaviour of NdMnAsO 0.95 F 0.05 is dominated by thermally activated charge carriers across a band gap so that ρ = ρ 0 exp (E g /2kT) (E g = 23 meV) 17 . The temperature variation of the resistivity can be modelled by three-dimensional variable range hopping (VRH) 24 of the carriers below 85 K (the resistivity, ρ, is defined as ρ = ρ 0 exp (T 0 /T) 0.25 ). In the variable range hopping mechanism, a localised electron can only move from one localised site to another by phonon assisted hopping, which is a combined thermally active quantum tunnelling process. An electron will only tunnel to another site if the thermal activation energy required for the hop is reduced. Below T SR , in NdMnAsO 1−x F x (x = 0.05-0.08), the spin reorientation of the Mn spins from aligning along c to aligning parallel to a precipitates an electronic transition from three dimensional Mott variable range hopping (VRH) to Efros Shklovskii (ES) VRH 25 . This signifies that the reorientation of Mn spins into the basal plane results in enhanced Coulomb correlations between localized electrons 17 , which results in much higher resistivity below T SR . This transition is not observed in the parent compound 22 . The transition to ES VRH in NdMnAsO 1−x F x (x = 0.05-0.08) is crucial for the appearance of CMR in F − doped materials, as the CMR arises due to a reduction in Coulomb correlations upon application of a magnetic field 17 . A transition from an insulating antiferromagnet to a semiconducting paramagnet is observed upon applying a magnetic field, which results in an electronic transition from ES VRH to Mott VRH.</p><p>It is highly likely that the stronger lattice response to the spin reorientation transition in NdMnAsO 0.95 F 0.05 precipitates the electronic transition to ES VRH, as the a cell parameter suddenly contracts below T SR and Coulomb correlations are enhanced. It has previously been shown that the electronic structure of LnFeAsO systems strongly depends on small changes in interatomic distances 26 . It would appear that the same may be true for the 1111 Mn 2+ analogues and further studies of the electronic structure are warranted.</p><p>In summary we have investigated the temperature dependence of the crystal structure of NdMnAsO 0.95 F 0.05 . There is no evidence of a change in crystal symmetry upon cooling but there is a subtle lattice anomaly at T SR in the temperature variation of the a cell parameter and also the c/a ratio. We propose that this coupling between the lattice and magnetic order results in the electronic transition to ES VRH below T SR so that a coupling between lattice, electronic and magnetic degrees of freedom is evident in the CMR material NdMnAsO 0.95 F 0.05 .</p><!><p>A polycrystalline sample of NdMnAsO 0.95 F 0.05 was synthesised via a two-step solid-state reaction method. Initially, the NdAs precursor was obtained by the reaction of Nd pieces (Aldrich 99.9%) and As (Alfa Aesar 99.999%) at 900 °C for 24 h in an evacuated, sealed quartz tube. The resulting precursor was then reacted with stoichiometric amounts of MnO 2 , Mn and MnF 2 (Aldrich 99.99%), all powders were ground in an inert atmosphere and pressed into pellets of 10 mm diameter. The pellets were placed into a Ta crucible and sintered at 1150 °C for 48 h, again in a quartz tube sealed under vacuum.</p><p>Physical Measurements:. The temperature dependence of the electrical resistance was recorded using a Quantum Design physical property measurement system (PPMS) between 4 and 280 K in magnetic fields of 0 T and 7 T. The field dependence of the electrical resistance was recorded in magnetic fields of ± 9 T.</p><!><p>High resolution synchrotron X-ray powder diffraction patterns of NdMnAsO 0.95 F 0.05 were recorded on the ID31 beamline at the ESRF, Grenoble, France at selected temperatures between 4 K and 290 K with a wavelength of 0.3999 Å. The powder sample was inserted into a 0.5 mm diameter borosilicate glass capillary and spun at ~1 Hz to improve the powder averaging of the crystallites. Diffraction patterns were collected over the angular range 2-45 ° 2θ and rebinned to a constant step size of 0.002° for each scan.</p>

Scientific Reports - Nature

Energy Decomposition Analyses Reveal the Origins of Catalyst and Nucleophile Effects on Regioselectivity in Nucleopalladation of Alkenes

Nucleopalladation is one of the most common mechanisms for Pd-catalyzed hydro- and oxidative functionalization of alkenes. Due to the electronic bias of the \xcf\x80-alkene-palladium complexes, nucleopalladations with terminal aliphatic alkenes typically deliver the nucleophile to the more substituted sp2 carbon to form the Markovnikov-selective products. The selective formation of the anti-Markovnikov nucleopalladation products requires the inherent electronic effects to be overridden, which is still a significant challenge for reactions with simple aliphatic alkenes. Because the interactions between the nucleophile and the alkene substrate are influenced by a complex combination of multiple types of steric and electronic effects, a thorough understanding of the interplay of these underlying interactions is needed to rationalize and predict the regioselectivity. Here, we employ an energy decomposition approach to quantitatively separate the different types of nucleophile-substrate interactions, including steric, electrostatic, orbital interactions, and dispersion effects, and to predict the impacts of each factor on regioselectivity. We demonstrate the use of this approach on the origins of catalyst-controlled anti-Markovnikov-selectivity in Hull\xe2\x80\x99s Pd-catalyzed oxidative amination reactions. In addition, we evaluated the regioselectivity in a series of nucleopalladation reactions with different neutral and anionic Pd catalysts and N- and O-nucleophiles with different steric and electronic properties. Based on these computational analyses, a generalized scheme is established to identify the dominant nucleophile-substrate interaction affecting the regioselectivity of nucleopalladations with different Pd catalysts and nucleophiles.

energy_decomposition_analyses_reveal_the_origins_of_catalyst_and_nucleophile_effects_on_regioselecti

INTRODUCTION<!>COMPUTATIONAL METHODS<!>Mechanisms and Regioselectivity-Determining Step of the Pd-Catalyzed Oxidative Amination<!>The energy decomposition analysis approach to dissect the effects controlling the regioselectivity<!>Analysis of different types of nucleophile-substrate interactions along the reaction coordinate<!><!>Identification of the dominant nucleophile-substrate interaction leading to the Markovnikov regioselectivity<!>Effects of different N- and O-nucleophiles on regioselectivity<!>Catalyst effects on the regioselectivity<!>Summary of the catalyst effects and nucleophile effects on the regioselectivity<!>CONCLUSION

<p>Palladium-catalyzed hydro- and oxidative functionalization of unactivated aliphatic alkenes is an efficient and atom-economical synthetic strategy for new C‒C and C‒heteroatom bond formation.1 One of the most common mechanistic pathways in these reactions involve nucleopalladation, in which a nucleophilic coupling partner attacks a π-alkene-Pd complex via either syn-insertion or anti-attack (also referred to as cis- and trans-nucleopalladation, respectively).2 Controlling the regioselectivity still remains a significant challenge in intermolecular nucleopalladation with sterically and electronically unbiased aliphatic alkenes.3 Most nucleopalladation reactions favor the Markovnikov products, in which the nucleophile is added to the more substituted carbon of the alkene. Although a broad variety of N- and O-nucleophiles have been employed in nucleopalladations,4 examples that override the intrinsic Markovnikovselectivity are rare. A practical challenge in the rational design of catalyst-controlled anti-Markovnikov selective reactions is that the dominant factor controlling the regioselectivity remains ambiguous. A sensible explanation to the regioselectivity control is based on favored electrostatic interactions between the nucleophile and the more substituted internal sp2-carbon to provide the Markovnikov addition products (Figure 1a). In addition, orbital interactions,5 steric,4a,6 and dispersion7 effects are also expected to impact the regioselectivity (Figure 1b). Therefore, a thorough understanding of how the different types of effects contribute to the regioselectivity, and more importantly, how the individual factors can be fine-tuned, is essential for the development of anti-Markovnikov selective transformations that overcome the inherent electronic preferences.</p><p>Dissecting the multiple underlying effects and rationalizing the major factors for the regioselectivity control is also challenging in computational studies. A number of computational studies have been reported for functionalization of aliphatic alkenes via migratory insertion mechanisms, in which steric effects typically dominate the regioselectivity.8 In contrast, computational studies on the regioselective nucleopalladation of alkenes are limited due to the complexity of regioselectivity control.9 The potential competition between cis- and trans-nucleopalladation pathways1c,10 further complicated the computational analysis (Figure 1a). Here, we present a systematic computational approach to quantitatively analyze the contributions of different types of effects to the regioselectivity in cis- and trans-nucleopalladations with different palladium catalysts and a variety of N- and O-nucleophiles with different formal charges, steric hinderances, and nucleophilicities (Figure 1c). Using the distortion-interaction/activation-strain model11 and energy decomposition analysis (EDA)12 methods, the computed regioselectivity is dissected into steric repulsions, electrostatic interactions, orbital interactions, and dispersion interactions between the nucleophile and the substrate. This decomposition approach allows for a straightforward way to reveal the dominant factor for regioselectivity control. Therefore, the origin of regioselectivity in different catalyst systems and in reactions with different nucleophiles can be rationally predicted.</p><p>Here, we report the use of this energy decomposition approach to study the origin of the catalyst-controlled anti-Markovnikov regioselectivity in the Pd-catalyzed oxidative amination of terminal alkenes, developed by the Hull group.13 In this reaction, the addition of Bu4NCl and Bu4NOAc effectively reverses the regioselectivity to favor the anti-Markovnikov amination products (Figure 2a). This is a unique example where complete regioselectivity reversal is achieved by simply changing the neutral Pd(OAc)2 catalyst4d system to a putative anionic Pd catalyst.14 Previously, the Hull group performed detailed mechanistic investigations using selectively deuterium labeled substrates and revealed important mechanistic insights into the C–N bond formation (Figure 2b).13 In the first experiment (eq. 1), 2,2-dideuterohomoallyl benzene was subjected to the reaction conditions; the selective migration of one deuterium to C3 and the second deuterium being at both C2 and C1 is consistent with the reaction occurring via aminopalladation (TS-III or TS-IV, Figure 1a) and eliminates the possibility of an allylic C–H activation. The second experiment (eq. 2), obtaining the monodeuteron product from the reaction with (Z)-2-deuterostyrene supports that the reaction is occurring through a transaminopalladation via TS-IV rather than cis-aminopalladation via TS-III (Figure 1a). In addition, kinetic experiments suggested that an associative ligand exchange of Cl– or OAc– for an olefin is the turnover-limiting step (TLS) and that monomeric [Pd] is the resting state, but that both monomeric and dimeric Pd complexes are active catalysts. While this is key information about the catalytic cycle, it limits the experimental investigations that can be conducted to investigate the anti-Markovnikov selectivity as the C–N bond formation occurs after the TLS. The catalyst-controlled regioselectivity may be resulting from a few possible effects. First, the additives may change the number of available coordination sites on the Pd,10b,15 which will affect the preferred mechanism for aminopalladation. The additional anionic ligand may block the coordination site on the Pd and prevent nucleophile coordination in the cisaminopalladation. Second, the electronic property of the Pd catalyst should affect the partial atomic charges, molecular orbital energies and coefficients of the alkene.16 Therefore, both electrostatics and orbital interactions between the alkene and the nucleophile are expected to be different with the neutral and the anionic palladium catalysts. Lastly, the catalyst may also affect the forming C–N bond distances in the transition states, which could affect the sensitivity to the steric repulsions and/or London dispersion interactions between the nucleophile and the alkene. Here, we demonstrate the energy decomposition approach can quantitatively dissect the contributions from different factors in both cis- and transaminopalladation with different Pd catalysts. While the regioselectivity is always affected by multiple factors, the computational analysis provided a straightforward way to predict which factor is dominant in each nucleopalladation reaction. We expect these chemically meaningful insights into the origin of catalyst and nucleophiles effects on the regioselectivity can be utilized to facilitate the catalyst design of regioselective and regiodivergent alkene functionalization reactions.</p><!><p>Geometry optimizations and single-point energy calculations were carried out using Gaussian 09.17 The geometries of intermediates and transition states were optimized using the B3LYP functional18 with a mixed basis set of SDD for Pd and 6–31+G(d) for other atoms in the gas phase. Vibrational frequency calculations were performed for all the stationary points to confirm if each optimized structure is a local minimum or a transition state structure. Truhlar's quasi-harmonic corrections19 were applied for entropy calculations using 100 cm−1 as the frequency cut-off. Solvation energy corrections and CHelpG atomic charges were calculated in dimethylacetamide (DMA) solvent with the SMD continuum solvation model20 based on the gas-phase optimized geometries. The M06 functional21 with a mixed basis set of SDD for Pd and 6–311+G(d,p) for other atoms was used for solvation single-point energy calculations. The energy decomposition analysis (EDA) calculations were performed to dissect the computed gas-phase activation energy (ΔE‡). In our decomposition approach, the activation energy is first separated into the distortion energy (ΔEdist) of the two reactive fragments to reach their transition state geometries and the interaction energy (ΔEint) between the two fragments using the following equation:11a,11d ΔE‡= ΔEdist+ ΔEint For cis-aminopalladation transition states (9M-TS and 9A-TS), the two fragments include the alkene and the (AcO)2Pd–NPhth complex as the nucleophile. For trans-aminopalladation transition states (6M-TS, 6A-TS, 12M-TS, 12A-TS), the two fragments include the Pd-alkene complex and the phthalimide anion or the diisopropylamine as the nucleophile. Using the second-generation energy decomposition analysis based on absolutely-localized molecular orbitals (ALMO-EDA)22 in Q-Chem 5.0,23 the interaction energy (ΔEint) between the two fragments is dissected according to the following equation: ΔEint= ΔEPauli+ ΔEelstat+ ΔEpol+ ΔEct+ ΔEdisp The ALMO-EDA calculations were performed at the M06/6–311G(d,p)‒LANL2DZ level of theory. The dispersion term is computed using the second-generation ALMO-EDA by calculating the difference of the "frozen interaction" term from a standard exchange-correlation functional and from an auxiliary "dispersion free" exchange correlation functional.22c,22d HF is used as the "dispersion free" exchange correlation functional in the dispersion energy calculations.</p><p>Structures along the reaction coordinates in Figure 6 were obtained from intrinsic reaction coordinate (IRC) calculations. The Markovnikov and anti-Markovnikov transition states usually have different forming C–N bond distances. To minimize the effects of early or late transition states in comparing the differences of each energy term between the two regioisomeric pathways, the ΔΔE reported in all pie charts (Figures 7–9) were computed using the average of ΔΔE values at the two C–N bond distances that correspond to the Markovnikov and anti-Markovnikov transition states, respectively: ΔΔEave= 1/2 (ΔΔEdis1+ ΔΔEdis2)= 1/2 [(ΔEM(dis1)– ΔEA(dis1)) + (ΔEM(dis2)– ΔEA(dis2))] Here, dis1 and dis2 are the forming C–N bond distances at the two regioisomeric transition states. ΔEM and ΔEA are the EDA energy terms at a given C–N bond distance along the Markovnikov and anti-Markovnikov reaction coordinates, respectively. The Complementary Occupied-Virtual Pairs (COVPs)24 calculations were performed using Q-Chem 5.0 at the M06/6–311G(d,p) (LANL2DZ for Pd) level of theory. The ΔEct shown in Figure 7 is the charge transfer energy derived from the most significant COVPs. To be consistent with other EDA calculations, the ΔEct(COVP) values were calculated from the average of two structures with C–N bond constrained to dis1 and dis2, where dis1 and dis2 are the C–N bond distances at the two regioisomeric transition states: ΔEct(COVP) = 1/2 [ΔEct(dis1)(COVP) + ΔEct(dis2)(COVP)] The optimized structures and the COVP orbitals were visualized using CYLview25 and GaussView 6.0.</p><!><p>Prior to applying the energy decomposition analysis to study the origin of regioselectivity, we needed to identify the regioselectivitydetermining step in the Pd-catalyzed oxidative amination reactions. Therefore, we computed the catalytic cycles of the oxidative amination of alkene 1 with both neutral and anionic palladium catalysts.26 The DFT calculations indicated the aminopalladation is irreversible and is the regioselectivity-determining step.27 The computed energy profiles of the neutral Pd(OAc)2-catalyzed aminopalladation step of alkene 1 with phthalimide anion28 are shown in Figure 3. The anti-Markovnikov and Markovnikov trans-aminopalladation of the p-alkene complex 5 (6A-TS and 6M-TS) require 17.0 and 16.7 kcal/mol with respect to 5, respectively. In the cisaminopalladation pathway, a stable anionic complex 8 is formed via the coordination of the phthalimide anion with the Pd center in complex 5. Subsequent anti-Markovnikov and Markovnikov aminopalladations occur through four-membered cyclic transition states 9A-TS and 9M-TS, respectively. Both 9A-TS and 9M-TS are lower in energy than the trans-aminopalladation transition states 6A-TS and 6M-TS. This indicates the most favorable mechanism for the reaction with the neutral Pd catalyst is the cisaminopalladation. This conclusion is consistent with previous mechanistic studies that suggest reactions with strongly coordinating nucleophiles prefer the cis-nucleopalladation.1c,10a</p><p>In the presence of Bu4NCl or Bu4NOAc additives,13 the anionic Cl- or OAc- ligand could coordinate to the neutral palladium catalyst to form palladate complexes.10b,15 Because the chloride and acetate salts can both promote the anti-Markovnikov regioselectivity, we chose acetate additives in the calculations for simplicity (i.e. with Pd(OAc)3 - as the active catalyst).29 As shown in Figure 4, the formation of the palladate complex 10 through coordination of an acetate anion to the neutral [Pd(OAc)2]3 is slightly endergonic by 1.7 kcal/mol. Therefore, this equilibrium is expected to shift to favor the palladate complex at higher additive concentrations. From the palladate-alkene complex 11, the trans-attack of the phthalimide anion requires 23.2 and 26.3 kcal/mol for anti-Markovnikov and Markovnikov additions (12A-TS and 12M-TS), respectively. In contrast, the outer-sphere cis-aminopalladation, in which the nucleophile does not coordinate with the Pd center, requires much higher activation energies (See Figure S5 for details). The inner-sphere cis-aminopalladation requires one of the acetate ligands to be replaced by a phthalimide anion to form complex 8. The cis-aminopalladation transition states from 8 are the same as those with the neutral Pd(OAc)2 catalyst (9A-TS and 9MTS, Figure 3). Therefore, this pathway would lead to the Markovnikov- selective products, rather than the anti-Markovnikov products observed experimentally under these conditions. Thus, this inner-sphere cis-aminopalladation pathway is unlikely. Therefore, the most favorable mechanism with the anionic catalytic system is trans-aminopalladation, which is consistent with the deuterium labeling experiments by the Hull group.13</p><p>Taken together, the DFT calculations suggest the oxidative amination with the neutral Pd catalyst occurs via the cisaminopalladation mechanism,4d while the reaction with the anionic Pd catalyst occurs via the trans-aminopalladation mechanism. The aminopalladation is the regioselectivity-determining step. The computationally predicted regioselectivities in the aminopalladation with the neutral and anionic palladium catalysts agree well with the experimental observation (Figure 2a). With the neutral catalyst, the Markovnikov-selective cis-aminopalladation (9M-TS) is favored by 2.7 kcal/mol. The complete regioselectivity reversal is observed with the anionic catalyst, which strongly favors the anti-Markovnikov-selective trans-aminopalladation pathway (12A-TS) by 3.1 kcal/mol.30</p><!><p>Although the computational results nicely reproduced the experimental regioselectivity trend, it remains a challenge to provide a chemically meaningful explanation to the origin of the selectivity. As discussed in the Introduction, the reversal of regioselectivity may be due to several different effects. To establish a general understanding of the origin of regioselectivity, we performed energy decomposition analysis (EDA) calculations to study the oxidative amination reaction described above and a number of related nucleopalladation processes with different nucleophiles and Pd catalysts. Here, we first demonstrate the detailed steps of using the EDA approach to study the neutral Pd(OAc)2-catalyzed C‒N formation transition states with phthalimide anion (i.e. 9M-TS, 9A-TS, 6MTS, and 6A-TS). The same procedure is then applied to study other nucleopalladation reactions.</p><p>Equation 3 was used to dissect the contributions of different types of nucleophile-substrate interactions in the aminopalladation transition states (Figure 5). First, using the distortion/interaction model,11 the activation energy of each transition state (ΔE‡) is decomposed into distortion energy (ΔEdist) of the two reactive fragments (highlighted in yellow and blue, respectively) and the interaction energy (ΔEint) between these two fragments. Then, using the ALMO-EDA method,22c the interaction energy (ΔEint) is further dissected into Pauli repulsion (ΔEPauli), electrostatic interactions (ΔEelstat), polarization (ΔEpol), charge transfer (ΔEct), and dispersion (ΔEdisp) (see Computational Methods for details). Specifically, ΔEelstat is the Coulombic interactions between the two fragments, ΔEpol is the stabilizing interactions from mixing of occupied and vacant orbitals within each fragment, and ΔEct is the interactions between an occupied orbital on one fragment and a vacant orbital on the other fragment. Among these terms, the sum of distortion energy (ΔEdist) and Pauli repulsion (ΔEPauli) can be considered as the contribution of steric effects (ΔEsteric) to the overall activation energy. The sum of ΔEelstat, ΔEpol, and ΔEct describes electronic effects (ΔEelec). Therefore, the overall activation energy (ΔE‡) is dissected into contributions from steric effects (ΔEsteric), dispersion effects (ΔEdisp), and two different types of electronic effects, namely electrostatics (ΔEelstat) and orbital interactions (ΔEorbital).31</p><!><p>We applied this EDA approach to study the four nucleopalladation pathways described in Figure 5. To analyze how the different types of nucleophile-substrate interactions vary along the nucleopalladation reaction coordinates, the computed energy terms with respect to the forming C–N (nucleophile) bond distances are illustrated in Figure 6. A few general conclusions about the origin of regioselectivity can be derived from the comparison of different pathways.</p><!><p>Steric repulsions (ΔEsteric) are always less prominent in the anti-Markovnikov pathway (blue lines) than in the Markovnikov pathway (red lines). In contrast, both types of electronic effects (ΔEelstat and ΔEorbital) always favor Markovnikov additions. Nucleophile-substrate dispersion interactions (ΔEdisp) also favor Markovnikov addition because dispersion interactions are larger when the nucleophile attacks the more substituted internal carbon of the alkene.</p><p>While the different types of interactions are all stronger at shorter C–N distances, the vertical distances between the red and the blue lines (ΔΔE) remain largely constant in the transition state region (2.0–2.3 Å). This indicates the energy difference between the Markovnikov and anti-Markovnikov pathways, i.e. contribution of each factor to the regioselectivity, is not affected by the location of the transition state along the reaction coordinate.</p><p>The magnitude of each ΔΔE term can be very different among different reactions. For example, the difference of orbital interaction energies (ΔΔEorbital) is −17.1 kcal/mol in the four-membered cyclic cis-aminopalladation pathway (Figure 7a). This indicates the Markovnikov transition state 9M-TS is strongly stabilized by orbital interactions. In contrast, the ΔΔEorbital values are much smaller (c.a. −2~−4 kcal/mol) in the trans-aminopalladation pathways (Figure 7b), indicating much smaller effects of orbital interactions on regioselectivity in these reactions. Therefore, detailed analysis of the relative magnitudes of the different effects can reveal the dominant factor on regioselectivity in each reaction.</p><!><p>The energy decomposition analysis results (Figure 6) are summarized as pie charts shown in Figure 7 to highlight the quantitative contributions of each type of nucleophile-substrate interactions to the overall regioselectivity in the aminopalladation with phthalimide anion. Each pie chart includes four different effects on regioselectivity. Among these, steric repulsions (ΔΔEsteric) are the only effect that promotes anti-Markovnikov addition, while electrostatics (ΔΔEelstat), orbital interactions (ΔΔEorbital), and dispersions (ΔΔEorbital) promote Markovnikov addition. Therefore, because the sum of ΔΔEelstat, ΔΔEorbital, and ΔΔEdisp is greater than ΔΔEsteric in both cis- and trans-aminopalladation of 5, the Markovnikov addition is preferred (Figure 7a and 7b). However, the dominant effects leading to the Markovnikov selectivity in these reactions are distinct from each other. In the cis-aminopalladation (Figure 7a), the dominant effect that promotes the Markovnikov-selectivity (9MTS) is orbital interactions (ΔΔEorbital), while electrostatics (ΔΔEelstat) plays a more significant role in the trans-aminopalladation (6MTS, Figure 7b). On the other hand, dispersion effects are typically small compared with steric and electronic effects. This EDA analysis provides an unbiased and straightforward way to identify the dominant factor on regioselectivity. Therefore, a more in-depth theoretical investigation on the dominant effect can then be performed to provide additional mechanistic insights. To understand why orbital interactions promote Markovnikov selectivity in cisaminopalladation, we performed the Complementary Occupied-Virtual Pairs (COVPs) analysis24 to study the donor-acceptor interactions between the Pd–Nu fragment and the alkene in the cisaminopalladation transition states (Figure 7a).32 The COVP results revealed that the charge transfer between the occupied alkene π orbital and the vacant metal d orbital is the most important orbital interaction for the Markovnikov regioselectivity. This π→d orbital interaction is much more pronounced in the Markovnikov transition state 9M-TS than in the anti-Markovnikov transition state 9ATS (ΔEct(π→d) = −41.9 and −33.1 kcal/mol, respectively). This is a result of the polarization of the occupied π orbital of the terminal aliphatic alkene (Figure 7a).33 The FMO interactions between the HOMO of Pd-Nu and the π* of the alkene are comparable in the Markovnikov and anti-Markovnikov pathways (ΔEct(Pd-Nu→π*)=−21.1 and −19.8 kcal/mol in 9M-TS and 9A-TS, respectively). Therefore, the HOMO(Pd-Nu)/π* interactions have a smaller contribution to the regioselectivity (see Figure S10 for details). On the other hand, the dominant role of electrostatics (ΔΔEelstat) in Pd(OAc)2-catalyzed Markovnikov-selective trans-aminopalladation is evidenced by the calculated CHelpG atomic charge on alkenes (Figure 7b).34 Upon coordination with the Pd(OAc)2 catalyst, the electron density on the C=C double bond becomes more polarized than in the free alkene with a greater amount of partial positive charge residing on the internal carbon.35</p><!><p>We then employed the EDA approach to study the origin of regioselectivity in nucleopalladation of the π-alkene/Pd(OAc)2 complex 5 with a variety of N- and O-nucleophiles bearing different steric hindrances and formal charges. We surmised these investigations would reveal how the steric and electronic properties of the nucleophile affect the different types of nucleophile-substrate interactions. In addition, the differences between neutral and anionic nucleophiles, and between N- and O-nucleophiles are also explored. In agreement with previous experimental observations,1c,4a,4b most of the nucleophiles investigated favor Markovnikov-selective nucleopalladation. Surprisingly, the EDA results indicate that the kinetic regioselectivity of nucleopalladation is not very sensitive to the steric properties of the nucleophile. For example, nucleopalladations with sterically distinct primary amines (NH2Me, NH2iPr, and NH2tBu in Figure 8a–c)4a,36 have comparable contributions resulting from steric effects (ΔΔEsteric = 6.1~6.9 kcal/mol). These nucleopalladations are all kinetically Markovnikov-selective, because the sum of electronic and dispersion effects in these reactions is greater than the steric influences. These results are consistent with previous experimental reports that suggest reactions with relatively bulky N-nucleophiles, such as NHMe2,4a still prefer Markovnikov products. Only in the reaction with an extremely bulky secondary amine nucleophile (NHiPr2 in Figure 8d), the ΔΔEsteric is significantly increased and reverses the regioselectivity to favor the anti-Markovnikov addition products. Similarly, nucleopalladations with MeOH and tBuOH also have comparable ΔΔEsteric terms (7.7 and 8.3 kcal/mol in Figure 8e–f, respectively).37</p><p>The insensitivity of nucleophile steric effects underlines the challenge to achieve anti-Markovnikov selectivity with the neutral Pd(OAc)2 catalyst. It also highlighted the importance of understanding the origin of electronic effects because electrostatic interactions and orbital interactions may be more tunable. Interestingly, the magnitudes of the electrostatic effects are sensitive to the formal charge of the nucleophile. Nucleopalladations with negatively charged nucleophiles (Figure 8g–h and 7b), such as CH3CO2−, CF3CO2−, and PhthN−, have much more negative ΔΔEelstat than those with neutral nucleophiles. These results indicate that anionic nucleophiles tend to lead to greater Markovnikov selectivity due to more favorable nucleophile-alkene electrostatic attraction in the Markovnikov addition transition state (Figure 1a). On the other hand, the nucleophilicity of the nucleophile have a small impact on the regioselectivity. For example, the ΔΔEelstat and ΔΔEorbital values are almost identical for nucleopalladations with CH3CO2− and CF3CO2− (Figure 8g–h), although the former is expected to beslightly more nucleophilic. The similar ΔΔEelstat and ΔΔEorbital values of nucleopalladations with PhthN‒ and acetate anion indicate the N- and O-nucleophiles have similar electronic effects.38 On the other hand, O-nucleophiles, including alcohols and carboxylate anions, have slightly greater ΔΔEsteric terms (by ~1 kcal/mol) than primary amine nucleophiles. This is likely due to the shorter C‒O bonds in the O-addition transition states that make these processes more sensitive to nucleophile-substrate steric repulsions.</p><p>Taken together, these EDA results indicate that in the neutral Pd(OAc)2-mediated nucleopalladations, the different types of nucleophile-substrate interactions have different sensitivity to the steric and electronic properties of the nucleophile. Nucleophilesubstrate steric interactions are largely insensitive to the size of the nucleophile, unless a highly hindered secondary amine is used. Although nucleophile-substrate orbital interactions remain nearly constant in all trans-aminopalladation processes tested, the nucleophile-substrate electrostatic interactions are highly sensitive to the formal charge of the nucleophiles. In reactions with negatively charged nucleophiles, such as CH3CO2−, CF3CO2−, and PhthN‒, electrostatic effects provide the greatest contribution to the Markovnikov selectivity, while in reactions with neutral nucleophiles, including various primary amines and alcohols, contributions from electrostatics and orbital interactions are comparable. The only case where steric effects dominate and thus favor the anti-Markovnikov pathway is with the highly hindered NHiPr2 nucleophile.</p><!><p>The EDA results discussed above indicated the Pd(OAc)2- mediated nucleopalladation with a large majority of nucleophiles favors the Markovnikov pathway. Therefore, the regiochemical reversal with the anionic Pd catalyst to kinetically favor the anti-Markovnikov nucleopalladation is remarkably unique. To investigate the origin of the catalyst-controlled regioselectivity, we performed EDA analysis on the regioselectivity of nucleopalladation with various neutral and anionic Pd catalysts. The use of another neutral Pd catalyst, Pd(H2O)Cl2, a potential active catalyst in Wacker oxidations,1c leads to similar regioselectivities compared to those with Pd(OAc)2 discussed above. Reactions with Pd(H2O)Cl2 are also relatively insensitive to the steric property and nucleophilicity, while more sensitive to the formal charge of the nucleophile (See Figure S20 for details). In contrast, EDA analysis with the anionic palladate Pd(OAc)3− as the active catalyst indicated a substantially different regioselectivity control (Figure 9). In particular, the comparison between Pd(OAc)3−- and Pd(OAc)2-mediated trans-aminopalladation with phthalimide anion as nucleophile (Figures 9 and 7b, respectively) reveals the origin of the different regioselectivity with anionic and neutral Pd catalysts. The complete reversal of regioselectivity to favor the anti-Markovnikov products is mainly attributed to the decrease of electrostatic effects in the reaction with the anionic Pd catalyst (ΔΔEelstat = −2.5 kcal/mol foranionic Pd versus −5.4 kcal/mol for the neutral Pd system).</p><p>These results indicate the electronic properties of the Pd catalyst can alter the polarization of the electron-density of the alkene in the π-alkene complex. This hypothesis is supported by the CHelpG atomic charge34,39 calculations (Figure 9). With the more electronrich anionic palladate, the ligand-to-metal charge transfer40 is less significant, and thus the internal carbon of the alkene in complex 11 becomes more electron-rich than that in 5 (Figure 7b). Therefore, the attractive electrostatic interactions between the nucleophile and the internal carbon becomes less favorable with the anionic palladium catalyst due to the decreased polarization of the C=C double bond in the π-alkene complex 11. Because the electrostatic and orbital interaction effects are both small in this anionic Pd catalyst system, steric effects become the dominant factor and effectively override the effects of orbital interactions, electrostatics, and dispersion to favor the anti-Markovnikov addition. EDA studies with other anionic palladate complexes, e.g. PdCl(OAc)2−, Pd(OPiv)3−, Pd(TFA)3−, also support the same conclusion that aminopalladations with anionic palladate complexes have diminished electronic effects on regioselectivity, and thus steric effects are the dominant factor leading to the anti-Markovnikov selective products (See Figure S14 and S18 for details).</p><!><p>The dominant factors controlling the kinetic regioselectivity in nucleopalladation with different Pd catalysts and nucleophiles are summarized in Figure 10. Although a majority of N- and Onucleophiles favor the Markovnikov-selective nucleopalladation when reacting with a neutral π-alkene/Pd(II) complex, the dominant factor on regioselectivity can be distinct. For nucleophiles that undergo cis-nucleopalladation (TS-e), the favorable π→d orbital interactions are the most important factor that leads to the Markovnikov selectivity. In trans-nucleopalladation with anionic nucleophiles (TS-c), electrostatic effects (ΔΔEelstat) become the most important factor favoring the Markovnikov selectivity. In transnucleopalladation with neutral nucleophiles (TS-d), both orbital interactions and electrostatics are the main effects controlling the regioselectivity. In contrast, anti-Markovnikov-selective nucleopalladation is favored when steric effects (ΔΔEsteric) dominate. Our computational analysis suggests the regioselectivity is relatively insensitive to the steric properties of the nucleophile. Therefore, only highly bulky nucleophiles (TS-b) can override the intrinsic electronic preference and selectively form the anti-Markovnikov products. On the other hand, because the nucleophile-substrate electrostatic interactions are sensitive to the formal charges on the Pd catalyst, the use of anionic palladate catalysts (TS-a) significantly reduces the effects of electrostatic interactions that favor the Markovnikov addition. Therefore, the nucleophile-substrate steric repulsions become the dominant factor in these reaction systems, leading to the catalyst-controlled complete reversal of regioselectivity.</p><!><p>We presented an energy decomposition approach to separate different types of nucleophile-substrate interactions and to rationalize their contributions to the regioselectivity of the nucleopalladation of terminal aliphatic alkenes. The computed regioselectivity (ΔΔE‡) is quantitatively dissected into effects from steric repulsions (ΔΔEsteric), electrostatics (ΔΔEelstat), orbital interactions (ΔΔEorbital), and dispersion (ΔΔEdisp) between the nucleophile and the alkene in the nucleopalladation transition state. Therefore, the major factor controlling the regioselectivity of reactions with different catalysts and nucleophiles can be revealed in a robust and practical fashion.</p><p>In this study, we employed this approach to study the origins of regioselectivity in a series of nucleopalladation reactions with neutral and anionic Pd catalysts and different nucleophiles. The computational results indicated that the regioselectivity is largely insensitive to the steric property of the nucleophile, unless an extremely bulky nucleophile is used. On the other hand, the nucleophilesubstrate electrostatic interactions and orbital interactions can be affected by the electronic properties of the nucleophile and the Pd catalyst, and the cis/trans-nucleopalladation mechanisms. In Pd(OAc)2-mediated cis-nucleopalladation, the Markovnikov selectivity is mainly due to the favorable frontier molecular orbital interactions between the Pd-nucleophile complex and the alkene substrate. In Pd(OAc)2-mediated trans-nucleopalladation, the orbital interaction effects are diminished and the favorable electrostatic interactions with the internal carbon of the alkene becomes more important for the Markovnikov-selectivity. The use of an anionic palladate catalyst decreases the polarization of the π electrondensity of the alkene, and thus the electrostatic effects are diminished. Therefore, the catalyst-controlled anti-Markovnikov selectivity in Hull's oxidative amination reactions is due to the diminished electrostatic effect that makes steric effects the dominant factor on regioselectivity. We expect the interplay of steric, electrostatic, and orbital interaction effects on regioselectivity of alkene nucleometallation revealed in this study can offer unique insights to guide future experimental design of regiodivergent functionalization strategies of simple unactivated alkenes. For example, since the electrostatic effect on regioselectivity appears to be more easily tunable, strategies to suppress the favorable electrostatic interactions in Markovnikov attack may lead to greater levels of anti-Markovnikov selectivity. This may be achieved through further optimization of electronic properties of the active Pd catalyst. Furthermore, the energy decomposition approach described here may be applied to study the origin of reactivity, regio-, and stereoselectivity of other types of transition metal-catalyzed reactions.</p>

PubMed Author Manuscript

The Impact of Minor-Groove N2\xe2\x80\x91Alkyl-2\xe2\x80\xb2-deoxyguanosine Lesions on DNA Replication in Human Cells

Endogenous metabolites and exogenous chemicals can induce covalent modifications on DNA, producing DNA lesions. The N2 of guanine was shown to be a common alkylation site in DNA; however, not much is known about the influence of the size of the alkyl group in N2-alkyldG lesions on cellular DNA replication or how translesion synthesis (TLS) polymerases modulate DNA replication past these lesions in human cells. To answer these questions, we employ a robust shuttle vector method to investigate the impact of four N2-alkyldG lesions (i.e., with the alkyl group being a methyl, ethyl, n-propyl, or n-butyl group) on DNA replication in human cells. We find that replication through the N2-alkyldG lesions was highly efficient and accurate in HEK293T cells or isogenic CRISPR-engineered cells with deficiency in polymerase (Pol) \xce\xb6 or Pol \xce\xb7. Genetic ablation of Pol \xce\xb9, Pol \xce\xba, or Rev1, however, results in decreased bypass efficiencies and elicits substantial frequencies of G \xe2\x86\x92 A transition and G \xe2\x86\x92 T transversion mutations for these lesions. Moreover, further depletion of Pol \xce\xb6 in Pol \xce\xba- or Pol \xce\xb9-deficient cells gives rise to elevated rates of G \xe2\x86\x92 A and G \xe2\x86\x92 T mutations and substantially decreased bypass efficiencies. Cumulatively, we demonstrate that the error-free replication past the N2-alkyldG lesions is facilitated by a specific subset of TLS polymerases, and we find that longer alkyl chains in these lesions induce diminished bypass efficiency and fidelity in DNA replication.

the_impact_of_minor-groove_n2\xe2\x80\x91alkyl-2\xe2\x80\xb2-deoxyguanosine_lesions_on_dna_replicati

<!>Western Blot.<!>ODN Synthesis.<!>Plasmid Construction, Replication, Progeny Isolation, and Replication Product Analysis.<!>RESULTS AND DISCUSSION<!>The Effects of N2-alkyldG Lesions on the Perturbation of DNA Replication Rate in HEK293T Cells.<!>The Effects of N2-alkyldG Lesions on the Perturbation of DNA Replication Accuracy in HEK293T Cells.

<p>Alkylating agents are produced from endogenous metabolism and are ubiquitously found in the environment.1,2 Exposure to these agents leads to the formation of DNA lesions, which may challenge genomic stability by impeding DNA replication and transcription and eliciting mutations.1–3</p><p>A large variety of DNA lesions are formed on 2′-deoxyguanosine (dG) in DNA. Many alkylating agents, apart from conjugating with the N7 and O6 positions of guanine, can react with the N2 of guanine.1,2 In this vein, benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE), formed from metabolic activation of benzo[a]pyrene, reacts predominantly with the N2 of dG.4 N2-dG adducts are also produced by chemotherapeutic agents, including mitomycin C (MC) and its metabolite.5,6 In addition, the N2 of guanine is susceptible to reaction with formaldehyde, which could be induced endogenously, and acetaldehyde, a chemical that can be produced endogenously from ethanol metabolism or lipid peroxidation, and is also present in external sources, including diesel exhaust, cigarette smoke, etc.7,8 Moreover, methylglyoxal, a glycolysis byproduct,9 could modify DNA to yield the stable N2-(1-carboxyethyl)-2′-deoxyguanosine (N2-CEdG).10–12</p><p>To mitigate the adverse consequences of DNA adducts, cells have multiple DNA repair pathways to remove DNA lesions.13 As part of the DNA damage tolerance mechanism, cells also employ translesion synthesis (TLS) DNA polymerases to assist the replication bypass of those lesions that normally block replicative DNA polymerases.14 In this respect, polymerase κ and its orthologs in other organisms have been shown to promote the accurate bypass of various N2-dG adducts, including those carrying simple alkyl groups,15 carboxyalkyl groups,11,16 a furfuryl functionality,17 an N2–N2 guanine interstrand cross-link,18 an acrolein-derived peptide cross-link,19 or a bulky BPDE remnant.20 Moreover, Pol ι, a unique TLS polymerase which can accommodate a non-Watson–Crick base pair into its active site,21 is needed for the error-free replication across N2-CEdG lesions in mammalian cells.16</p><p>In the present study, we, by utilizing a highly robust shuttle-vector-based method,22 investigated systematically how the alkyl group size in the N2-alkyldG lesions affects DNA replication efficiency and accuracy in human cells. We also assessed how replication through these DNA lesions is influenced by genetic ablation of TLS DNA polymerases.</p><p>Human REV1 (Rev1): ATCAGATGCTGCTATGCAGAAGG.</p><p>Human POLK (Pol κ): ATCCATGTCAATGTGCACTATGG.</p><p>Human REV3L (Pol ζ): AATGAGCCAACCTGAGTCACAAG.</p><!><p>The Western blot experiments were conducted using approximately 30 μg of protein lysate. Antibodies for human REV1 and GAPDH and the goat secondary antimouse antibody conjugated with horseradish peroxidase were procured from Santa Cruz Biotechnology, and the dilution factors for the three antibodies were 1:1000, 1:50000, and 1:2000, respectively.</p><!><p>The 12-mer ODNs carrying an N2-MedG, N2-EtdG, N2-nPrdG, or N2-nBudG at a specific site were synthesized following a previously reported convertible nucleotide approach,24,25 where the fluorine atom in 2-fluoro-2′-deoxyinosine in ODNs was substituted with the corresponding alkylamines (Figure 1b). The identities of the 12-mer ODNs containing the four N2-alkyldG lesions were confirmed by liquid chromatography-mass spectrometry (LC-MS) and tandem MS (MS/MS) experiments (Figures S2–S5), following published procedures.26</p><!><p>The methods for the construction of the lesion-containing plasmids were described in detail elsewhere,27 except that the aforementioned 12-mer N2-alkyldG-containing ODNs were employed as the lesion-containing insert. Cellular replication and progeny isolation, PCR amplification, and polyacrylamide gel electrophoressis (PAGE) as well as LC-MS/MS analysis were carried out as previously described,27 with the exception that the primer employed for strand-specific PCR was the same as what was used previously for the replication studies of 8,5′-cyclo-2′-deoxyguanosine-containing DNA.22</p><!><p>We set out to investigate the degrees to which the efficiency and fidelity of DNA replication in human cells are perturbed by N2-alkyldG lesions carrying different sizes of alkyl groups, where we considered four lesions, with the alkyl groups being a methyl (N2-MedG), ethyl (N2-EtdG), n-propyl (N2-nPrdG), and n-butyl (N2-nBudG) (Figure 1). These lesions were chosen on the basis of the fact that exposure to various environmental carcinogens and some anticancer agents can give rise to covalent attachment of different alkyl groups on the N2 position of dG, as described above.</p><p>We first prepared several ODNs housing an N2-alkyldG adduct at a specific site (Figure 1), validated the identities of these ODNs by ESI-MS and MS/MS (Figures S2–S5), and ligated them into double-stranded shuttle vectors (Figure 2). A damage-free competitor genome, which contained three additional nucleotides, was used as an internal reference to correct for the variations in transfection efficiencies among different experiments. After cellular replication, all plasmid genomes were extracted from cells, and the residual parental plasmids were eliminated by DpnI and exonuclease III digestion. The sequence segment of interest in the extracted progeny plasmids was subsequently amplified by strand-specific PCR (Figure 2b).22 The resulting PCR products were subjected to NcoI and SfaNI cleavage (Figure 3a), and the digestion products were analyzed by LC-MS/MS and PAGE (Figures 3 and 4 and Figures S6–S8).</p><!><p>We found that increasing size of the N2-alkyl group elicits subtle differences in the efficiency of replicative bypass in parental HEK293T cells, where all four lesions were bypassed at high efficiencies (60–80%) (Figure 5a). We also investigated how individual or combined ablation of TLS DNA polymerases influences the bypass of these lesions in HEK293T cells. We observed that single ablation of Pol η or Pol ζ did not alter the bypass efficiencies of the N2-alkyldG lesions (Figure 5a). This result suggests that, in the presence of other TLS polymerases, Pol η and Pol ζ do not assume appreciable roles in bypassing the N2-alkyldG lesions. Depletion of Pol ι, on the other hand, caused significant drops (to ~38–50%) in the bypass rates for all of the N2-alkyldG lesions (Figure 5a), and depletion of Pol κ also gave rise to significant diminutions in bypass efficiencies for N2-MedG and N2-EtdG (Figure 5a). To investigate whether Pol ι and Pol κ act cooperatively or independently, we explored the impact of their concurrent ablation on the bypass efficiencies of the N2-alkyldG lesions. Our results showed that their codepletion led to bypass efficiencies that are similar to what were observed for cells depleted of Pol ι alone, suggesting that Pol ι and Pol κ act cooperatively when bypassing these N2-alkyldG lesions.</p><p>We next examined how combined depletion of Pol ζ, which was previously found to function during the extension step of TLS,28,29 modulates the efficiencies of replication across the N2-alkyldG lesions in the HEK293T cells depleted of Pol κ or Pol ι. We observed that simultaneous removal of Pol ζ and Pol ι (Figure S1), or codepletion of Pol κ and Pol ζ, conferred significant reductions in bypass rates (to ~13–29 and 16–33%, respectively). Together, these data show that Pol ζ supports the TLS of the N2-alkyldG lesions in cells depleted of Pol κ or Pol ι.</p><p>Previous studies showed that DNA polymerase Rev1 played an indispensable scaffolding role for the recruitment of TLS polymerases in the Y-family, including Pol η, Pol ι, and Pol κ.30,31 Hence, we generated Rev1-deficient HEK293T cells by CRISPR-Cas9, confirmed the successful knockout of REV1 gene by Sanger sequencing and Western blot analysis (Figure S1), and explored the function of this polymerase in the replication across the N2-alkyldG lesions. As displayed in Figure 5a, we observed a substantial decline in the bypass rates for the N2-alkyldG lesions upon loss of Rev1 (~13–36%), underscoring the important function of Rev1 in supporting the replication through these lesions in human cells.</p><!><p>We determined the identities of the mutagenic replication products based on PAGE and LC-MS/MS analyses of the restriction digestion products of PCR amplicons from the isolated progeny plasmids (Figures 3 and 4 and Figures S6–S8). No mutagenic products were found for the N2-alkyldG lesions in parental HEK293T cells (Figure 3). Depletion of Pol η or Pol ζ did not compromise the fidelity of replication past the N2-alkyldG lesions (Figure S8), which is consistent with the corresponding replication bypass data described above.</p><p>Appreciable rates of G → T and G → A mutations were observed for all N2-alkyldG lesions, in all other polymerase-deficient backgrounds examined (Figure 5b,c, Figure 6). In this vein, G → A transition was detected at frequencies of ~10–25% in Pol ι- or Pol κ-deficient cells, accompanied by low frequencies (~4–15%) of G → T substitution. Thus, dTMP and, to a lesser extent, dAMP are misincorporated opposite the N2-alkyldG lesions. This result also reveals that, when in the absence of Pol κ or Pol ι, other TLS polymerases may assume a back-up role in bypassing the N2-alkyldG lesions, albeit at the expense of poorer fidelity. This finding is in line with the prior observations that purified human Pol κ and Pol ι could catalyze accurate and efficient nucleotide incorporation opposite N2-dG adducts.15,16 Further ablation of Pol ζ in Pol ι- or Pol κ-depleted cells led to significant increases in G → A mutations for all four lesions relative to single ablation of Pol ι or Pol κ (Figure 5b,c, Figure 6b,c). Significant elevations in G → T mutations were also found for N2-nPrdG upon deletion of Pol ζ in Pol κ-depleted cells and for N2-MedG and N2-nBudG upon removal of Pol ζ in Pol ι-depleted cells (Figure 5b,c and Figure 6b,c). Together, Pol κ and Pol ι are likely the major polymerases involved in the accurate nucleotide insertion opposite the N2-alkyldG lesions during TLS in HEK293T cells. Loss of Rev1 also led to substantial rates of G → A transition (to ~11–20%) and G → T transversion (~4–11%) mutations (Figure 5b,c), which perhaps can be attributed to the scaffolding role of Rev1 in assembling Y-family polymerases(i.e., Pol ι and Pol κ) during TLS of the N2-alkyldG lesions.30,31</p><p>In all genetic backgrounds where we observed mutagenic bypass of the N2-alkyldG lesions, the G → A and G → T mutation rates increased with the alkyl group size. For example, the frequencies for the G → A and G → T mutations in Pol ι-depleted cells range from ~10% for N2-MedG to ~17% for N2-nBudG and from ~7% for N2-MedG to ~13% for N2-nBudG, respectively (Figure 5b,c, Figure 6b,c). The elevated mutation frequencies are associated with diminished bypass efficiencies of these lesions. Thus, longer alkyl chains elicit lower bypass rate and decreased fidelity during replication past the N2-alkyldG lesions in human cells.</p><p>Many different types of DNA lesions are induced in cells upon alkylating agent exposure,2,32 and N2-dG adducts are of particular interest owing to the relatively poor efficiency in repair of these minor-groove lesions.33 However, the mutagenicity and cytotoxicity of the alkylated N2-dG lesions or the involvement of TLS DNA polymerases in supporting the replication across the N2-alkyldG lesions in human cells has not been systematically investigated.</p><p>Here, we systematically investigated the cytotoxicity and mutagenicity of N2-alkyldG lesions in HEK293T cells and CRISPR/Cas9-engineered cells with individual or combined depletion of TLS polymerases. We observed that alkyl groups located at the N2 of guanine did not significantly impede DNA replication and no mutagenic products were found in parental HEK293T cells that are proficient in TLS, or the isogenic cells with depletion of Pol ζ or Pol η. We also found that the efficiencies and fidelities of DNA replication across the N2-alkyldG lesions are significantly reduced in cells with Pol ι and Pol κ being ablated individually or in combination (Figures 5 and 6). Our results are consistent with the data from a number of previous in vitro biochemical studies showing the capability of DNA Pol κ in preferentially incorporating the correct dCMP opposite N2-dG adducts.11,15–20 For instance, Choi et al.15 found that the ratios for misincorporating dGMP and dTMP (relative to the insertion of the correct dCMP) opposite N2-MedG and N2-EtdG were less than 0.01. This can be attributed to the unique active site structure of Pol κ, where an opening toward the DNA minor groove could accommodate readily the alkyl groups conjugated with the N2 of the template guanine in the active site.34,35 Hence, our results provide important in vivo evidence to substantiate the conclusion that N2-dG adducts constitute cognate lesions for DinB DNA polymerases.11,17,36</p><p>Pol ι-catalyzed nucleotide incorporation across the N2-CEdG lesions was previously shown to be both accurate and efficient,16 and here we revealed that Pol ι promotes the accurate replicative bypass of N2-alkyldG lesions in human cells. Depletion of Pol ζ in Pol ι- or Pol κ-deficient background induced a further diminution in bypass efficiencies (Figure 5a). Pol ζ is known to extend proficiently from the DNA strand after nucleotide incorporation opposite DNA lesions.28,29 Therefore, the accurate bypass of the N2-alkyldG lesions may proceed through incorporation of the correct dCMP opposite the lesion by Pol ι or Pol κ, followed by extension of the primer from the resulting N2-dG:dC base pair by Pol κ or Pol ζ, thereby acting in a cooperative fashion.</p><p>In cells without Pol ι or Pol κ, the TLS of the N2-alkyldG lesions occurred with lower fidelity and efficiency (Figure 5b,c, Figure 6). TLS mediated by Pol η might lead to mutations at the site of N2-alkyldG lesions, which parallels what was observed recently for the bypass of N2-CMdG and N2-CEdG lesions in MEF cells.16 Along this line, Pol η was found to insert not only the correct dCMP but also the incorrect dAMP and dTMP opposite a number of N2-alkylated dG lesions.16,37 Removal of Rev1 led to appreciable rates of G → A and G → T substitutions (Figure 5b,c), which is in line with the previous notion that Rev1 serves as a scaffold for the recruitment of Y-family polymerases in mammalian cells.30,31</p><p>We also compared the findings made from the current study with what were previously observed for replication past other minor-groove DNA lesions. It was found that O2-alkyldT lesions primarily direct T → A and T → G mutations, even in parental HEK293T cells.27 We found here that the bypass of N2-alkyldG lesions does not evoke mutations in parental HEK293T cells, or the isogenic cells lacking Pol ζ or Pol η. By sharp contrast, both Pol ζ and Pol η are required for the efficient bypass of the minor-groove O2-alkyldT lesions. In addition, Pol ι and Pol κ are primarily involved in the accurate bypass of N2-alkyldG lesions (Figure 5b,c, Figure 6b,c). Replication across 3-deaza-3-methyladenine (3-dMeA), which is a model minor-groove DNA lesion, exhibited similarities and differences from that across the N2-alkyldG lesions.38 In this vein, bypass of the 3-dMeA lesion was mediated by the combined action of Pol κ and Pol ι, where it was proposed that Pol ι incorporated a nucleotide across 3-dMeA, followed by Pol κ-mediated extension from the inserted nucleotide.38 Pol ζ contributed to the extension step of TLS across the 3-dMeA lesion after nucleotide incorporation opposite the lesion catalyzed by other polymerases.38 Different from what we observed for the N2-alkyldG lesions, no mutagenic product was found in TLS opposite 3-dMeA in parental as well as Pol ι-, Rev1-, Pol θ-, or Pol ζ-knockdown cells, indicating highly accurate replication past this lesion in human cells.38</p><p>We would like to note some limitations of the shuttle-vector-based method when investigating how DNA damage modulates cellular DNA replication. First, the replication across DNA lesions situated in episomal plasmids may not fully recapitulate cellular replication where DNA is packaged into chromatin. Second, some translesion synthesis DNA polymerases were previously shown to function in DNA repair.39–41 Hence, we cannot formally exclude the possibility that the augmented ability of the lesions to block DNA replication and induce mutations in some polymerase-deficient cells may also arise in part from the functions of these polymerases in repairing the N2-alkyldG lesions.</p><p>In summary, our comprehensive replication study on a structurally defined group of N2-alkyldG lesions reveals previously unrecognized effects that these minor-groove lesions exert on the efficiency and fidelity of cellular DNA replication. We demonstrate that the degrees of cytotoxicity and mutagenicity imparted by the N2-alkyldG lesions depend on the bulkiness of the alkyl group, and Pol ι, Pol κ, and Rev1 are indispensable for the error-free bypass of these lesions. Additionally, we find that individual depletion of Pol η or Pol ζ does not appreciably affect the rate or fidelity of DNA replication across these lesions. Moreover, we reveal the cooperative roles of TLS polymerases, and the possible redundant role of Pol ζ as an extender, while bypassing these lesions.</p>

PubMed Author Manuscript

Prebiotic feeding elevates central brain derived neurotrophic factor, N-methyl-d-aspartate receptor subunits and d-serine☆

Highlights•Prebiotic feeding elevated BDNF and NR1subunit mRNAs, in the rat hippocampus.•The GOS prebiotic increased cortical NR1, d-serine, and hippocampal NR2A subunits.•GOS feeding elevated plasma levels of the gut peptide PYY.•GOS plasma increased BDNF release from human SH-SY5Y neuroblastoma cells.•BDNF secretion from cells by GOS plasma was blocked by PYY antisera.

prebiotic_feeding_elevates_central_brain_derived_neurotrophic_factor,_n-methyl-d-aspartate_receptor_

Introduction<!>Animals<!>Prebiotic experiments<!>Glucose and gut hormone measurements<!>BDNF analysis<!>Western blotting<!>In situ hybridization histochemistry (ISHH)<!>HPLC analysis<!>Cell culture<!>Data analysis<!>Rat faecal Bifidobacteria after prebiotics<!>The effect of prebiotics on BDNF and NR1 in the rat frontal cortex and hippocampus<!>The effect of prebiotics on NR2A and NR2B subunits in the rat frontal cortex and hippocampus<!>The effect of prebiotics on BDNF and NR subunit mRNAs in the hippocampus<!>Faecal, plasma and brain amino acid concentrations after prebiotics<!>Body weight change, and plasma glucose, BDNF, PYY and GLP-1after prebiotics<!>Correlations<!>The effect of plasma from prebiotic-fed rats on BDNF levels in SH-SY5Y cells<!>Discussion<!>Conclusion<!>Conflict of interest<!>Authorship credit

<p>There is now compelling evidence for a link between the enteric microbiota and brain function. The proliferation of the Bifidobacteria and Lactobacilli strains in the large intestine, have anxiolytic and mnemonic effects in both rodents (Li et al., 2009; Bravo et al., 2011) and humans (Messaoudi et al., 2011a,b; Rao et al., 2009; Cryan and Dinan, 2012). The intake of these bacteria as live cultures (probiotics) alters the expression of genes integral to neurodevelopment and complex behaviours in rodents. For instance, the oral administration of Bifidobacteria to rats elevated hippocampal brain-derived neurotrophic factor (BDNF) (Bercik et al., 2011a; O'Sullivan et al., 2011), which may underlie some antidepressant actions (Kerman, 2012). At present, only several probiotics have been examined, but it seems likely that of the 40,000 species in the gut (Forsythe and Kunze, 2012), there will be others with psychotropic properties. Thus, intuitively, augmenting the growth of intrinsic gut microbiota with prebiotics (nutrients for intestinal bacteria) may afford greater benefits to the brain (Burnet, 2012).</p><p>The prebiotics, fructo-oligosaccharide, (FOS) and galacto-oligosaccharides, (GOS) are soluble fibres which are digested by, and result in the proliferation of, the Lactobacilli and Bifidobacteria in the gut. Increasing the proportion of these bacteria with prebiotics has many beneficial effects on the gut and the immune system (Drakoularakou et al., 2010; van Vlies et al., 2012; Vulevic et al., 2008, 2013), and increase circulating gut peptides such as glucagon-like peptide-1 (GLP-1) and peptide YY (PYY), which benefit metabolism (Delmee et al., 2006; Overduin et al., 2013). However, the central effects of prebiotic administration have not been explored. Interestingly, selective antimicrobials which elevate the levels of intrinsic gut Lactobacilli, also increase brain BDNF concentrations in mice (Bercik et al., 2011a). It is possible therefore, that prebiotic-mediated microbiota proliferation has similar effects, and the measurement of brain BDNF in rodents administered with these compounds would provide the necessary proof-of-principle. Additional evidence suggests that gut bacteria may also influence glutamate neurotransmission in the brain.</p><p>Mice devoid of gut microbiota from birth have reduced levels of N-methyl-d-aspartate receptors (NMDARs), specifically the NR1 and NR2A subunits, in the hippocampus (Sudo et al., 2004), or NR2B subunits in the amygdala (Neufeld et al., 2011; Kiss et al., 2012). To our knowledge, the effect of increasing gut microbiota on brain NMDARs has not been explored, and such information may have therapeutic relevance (Collingridge et al., 2013). Interestingly, germ-free mice also lack circulating d-alanine, a bona fide NMDAR co-agonist which is rich in bacterial cell walls (Konno et al., 1993), and their inoculation with bacteria restored d-alanine concentrations, which were then increased further by an additional administration of a Bifidobacteria. It is reasonable to propose, therefore, that an elevation of central d-alanine, and perhaps other amino acids associated with glutamate neurotransmission, would follow prebiotic administration, and thereby present as a strategy to increase brain NMDAR signalling.</p><p>The three major aims of this study were to: (1) test if prebiotic administration to rats altered brain levels of BDNF; (2) examine whether central NMDARs and associated amino acids were altered by prebiotics; and (3) provide initial evidence for neuroactive blood-borne molecules that may affect central BDNF levels after prebiotic feeding. We orally administered water, FOS or GOS to rats for five weeks and measured BDNF NR1, NR2A and NR2B subunits in the frontal cortex and hippocampus, and encoding mRNAs in the hippocampus. The concentrations of glutamate, glutamine, and serine and alanine enantiomers in the plasma, cortex and hippocampus were also quantified. Finally, we measured the levels of PYY and GLP-1 in plasma from prebiotic-fed rats, and tested their effect on BDNF release from SH-SY5Y neuroblastoma cells.</p><!><p>All rat experiments were carried out in accordance with UK Home Office guide lines and under approved licences. Adult male Sprague Dawley rats (225–250 g) were obtained from Harlan Laboratory, UK, and maintained under controlled 12-h light/dark cycle (lights on 7:00 am), temperature (21 ± 1 °C) and humidity (55 ± 5%), with ad libitum access to drinking water/fluid and food (standard chow pellets). Rats were weighed at the start and end of experiments.</p><!><p>Rats were administered a daily oral administration (gavage) of either water, FOS (3 g/kg) or GOS (4 g/kg), for 5 weeks (n = 8/group). This dosing regimen was based on previous studies (Anthony et al., 2006). Copies of Bifidobacteria spp. genes in DNA extracted from faecal pellets were determined with standard QPCR at the end of the study, as previously described (Ketabi et al., 2011). Twenty-four hours after the last gavage, the animals were sacrificed, their brains removed and trunk blood collected in EDTA-coated tubes. Blood was centrifuged (5000 rpm, 15 min) to obtain plasma which was then stored at −80 °C. The frontal cortex and hippocampus were dissected out from half of the harvested brains. Brain hemispheres and isolated regions were snap-frozen in isopentane on dry-ice and stored with plasmas at −80 °C prior to use. Additional faecal pellets were collected from each animal (n = 8/group), weighed, homogenised in PBS (1:1, w/v), and then centrifuged at 14,000 rpm for 10 min at 4 °C. Supernatants were removed and stored at −80 °C prior to HPLC analysis.</p><!><p>The concentration of blood glucose was measured in all plasma samples using a GlucoMen LX, blood glucose meter (A. Menarini Diagnostics, UK). Commercial ELISA kits were used to measure plasma PYY (Abnova, UK) and GLP-1 (Millipore, UK), and were performed according to manufacturer's recommendations.</p><!><p>Cortex and hippocampus tissue from all groups (n = 8 rats/group) were homogenised in RIPA buffer (1:10 w/v, Sigma–Aldrich, UK) containing protease inhibitors ('Complete-Mini', Roche). Protein concentrations were determined using the Bradford reagent (Sigma, UK). Samples of protein extracts were diluted 1:5 v/v in assay buffer, prior to their analysis with a commercial BDNF ELISA kit (BDNF Emax immunoassay system, Promega, UK). Samples of cell culture medium were first diluted in an equal volume of RIPA, and then further diluted (1:5) in BDNF assay buffer. The BDNF ELISA was performed according to manufacturer's recommendations.</p><!><p>Western blots were performed as previously described (Burnet et al., 2011). Briefly, equal concentrations of protein extracts of cortex or hippocampus (20 μg) from prebiotic and control groups (n = 8 rats/group) were mixed with loading buffer (50 mM 1,4-dithiothreitol and 0.025% bromophenol blue), and fractionated with a molecular weight marker (GE Healthcare, Buckinghamshire, UK) by electrophoresis on pre-cast 7.5% SDS/polyacrylamide gels (Biorad, UK), and trans-blotted onto polyvinyl difluoride (PVDF) membranes (Immobilon-P, Millipore, Watford, UK).</p><p>The membranes were blocked with 5% (w/v) non-fat milk in PBS containing 0.1% Tween20 (PBST) for 45 min, and then incubated for 1 h at room temperature in incubation buffer (PBST with 2% [w/v] milk) containing a primary antibody (diluted 1:1000) against one of three NMDAR subunits: NR1 (AB9864, Millipore, UK), NR2A (AB1555, Millipore, UK) and NR2B (AB15362, Millipore, UK), and β-actin (Sigma–Aldrich, UK, diluted 1:50,000). Membranes were then washed three times for ten minutes in PBST and incubated for 30 min in HRP-linked secondary antibody in blocking buffer. Immunoreactive bands were visualized by chemiluminescence using the ECL-Plus kit (GE Healthcare, Buckinghamshire, UK), and apposing membranes to X-ray film (Kodak BioMax AR film). All antibodies produced a single band of expected molecular weight. The optical densities (OD) of bands were measured using the AlphaImager 3400, and the data expressed as OD ratios of NMDAR subunit:β-actin.</p><!><p>The frozen rat brain hemispheres were coronally sectioned (14 μm) on a cryostat, and every three sections thaw-mounted onto Superfrost-plus slides (Fisher Scientific, UK). All slides were stored at −80 °C prior to use. Slides containing sections of the dorsal hippocampus [−3.2 to −3.8 mm from Bregma, (Paxinos and Watson, 1986)], were pre-treated as described (Eastwood et al., 1995). One slide from each group ('water', 'FOS' and 'GOS', n = 8/group) was used for ISHH analysis.</p><p>Commercially synthesized (MWG, UK) oligodeoxyribonucleotides complementary to: BDNF (bases 883–927, NM001270630.1), NR1 (bases 746–780, NM008169.1), NR2A (bases 1642–1676, NM008170.2) or NR2B (bases 1758–1792, NM010350.2) were used in an establish ISHH method (Eastwood et al., 1995). Oligodeoxyribonucleotide probes were 3′-end labelled with [35S]-dATP using terminal deoxynucleotidyl transferase (Promega, UK). Probes were diluted in hybridization buffer, pipetted onto the tissue sections (1 × 106 cpm/section), cover-slipped and then incubated for >16 h at 34 °C lidded Perspex trays lined with filter paper soaked with 4× SSC/50% formamide.</p><p>Post-hybridization washes included: 2× SSC rinse at room temperature to remove cover-slips; 0.5× SSC, 20 min (3×) at 55 °C; 0.5× SSC 30 min (2×) at room temperature. Slides were rinsed in ddH2O, dried and apposed to X-ray film (Kodak, Biomax MS) for 7–28 days with 14C-microscales. Average grey densities over the dentate gyrus, CA1, and CA3 subfields of the hippocampus in the three sections from each group were measured for each of the mRNAs using computer-assisted image analysis, and were calibrated to 35S nCi/g tissue equivalents using the commercial 14C-microscales and a 14C to 35S conversion factor of 3.0 (Eastwood et al., 1995).</p><!><p>Small fragments of the cortical and hippocampal tissue (50 mg) were individually homogenised in ice-cold methanol (1:10 w/v) and centrifuged at 14,000 rpm for 10 min at 4 °C. Supernatants from faecal homogenates and plasma samples were mixed with 3 volumes of methanol and also centrifuged for 10 min at 4 °C. Supernatants (10 μl) from tissue or faecal homogenates or plasma were subjected to online, pre-column, derivatization (Grant et al., 2006), by injecting them onto a Hewlett–Packard 1100 liquid chromatography (Agilent Technologies, Palo Alto, CA), with an equal volume of derivatizing reagent [o-phthaldialdehyde (2 mg) and Boc-l-cysteine (2 mg) in 0.2 ml of methanol and 0.8 ml of 0.4 M of sodium borate buffer (pH = 9)], for 5 min prior to column separation. Separation was achieved using an Agilent Zorbax Eclipse XDB-C18 column (4.6 × 150 mm, 5 μm) maintained at 30 °C and a separation protocol similar to that of (Morikawa et al., 2001). The mobile phases consisted of acetonitrile (phase A) and 100 mM sodium acetate buffer pH = 6 (phase B) and were pumped through the column at 1.4 ml/min. The following gradient system was used (min/% B): 0/91, 35/84, 65/84. Detection of derivatized amino acids was by fluorescence detection (emission: 443 nm; excitation 344 nm). Eight point calibration curves of the d- and l-amino acids (Sigma–Aldrich, UK) were constructed using authentic standards (0.5–1000 pmol) and in each case were found to be linear with correlation coefficients of >0.995.</p><!><p>The release of BDNF from human SH-SY5Y neuroblastoma cells was investigated using a recently described protocol (Coco et al., 2013). Briefly, cells were cultured in Dulbecco's Modified Eagle Medium (DMEM; Sigma, Poole, UK) supplemented with 10% foetal calf serum (Sigma), 2 mM l-glutamine (Sigma) and 1% non-essential amino acids (Sigma), maintained in a humidified incubator at 37 °C and 5% CO2. Prior to release experiments, 24-well plates were seeded with 1 × 105 cells/well and incubated for 24 h. Plasma (50 μl) from rats fed water, FOS or GOS for 5 weeks (see above), containing either IgG or anti-PYY antisera (1:200, Abnova, UK) were then added to the 0.5 ml of culture media for 4 h. A total of 24 plasma samples (8 plasmas/group) were tested in triplicate. Some cells were incubated with synthetic PYY peptide (20 nM) alone. This experiment was then repeated in the presence or absence of GLP-1 antisera (Millipore, UK) or GLP-1 peptide. All media was removed following incubations, and stored at −80 °C prior to BDNF assays.</p><!><p>All data were expressed as mean ± standard error of the mean (SEM). Statistical comparisons between groups from rat experiments were performed with one-way ANOVA followed by post hoc analysis (Tukey HSD). Cell culture data were analysed non-parametrically (Kruskall–Wallis), followed by post hoc Mann–Whitney U tests.</p><!><p>The numbers of Bifidobacteria in faecal pellets from FOS-fed rats were significantly greater than controls in an ANOVA and post hoc (Tukey HSD) analysis i.e. controls: 2.38 × 109 ± 0.23 × 109 vs FOS: 2.98 × 109 ± 0.22 × 109, p < 0.05), whereas the numbers of Bifidobacteria from GOS-fed animals were significantly greater than both controls and FOS-fed rats i.e. controls: 2.38 × 109 ± 0.23 × 109 vs GOS: 4.28 × 109 ± 0.43 × 109, p < 0.01; and FOS: 2.98 × 109 ± 0.22 × 109 vs GOS: 4.28 × 109 ± 0.43 × 109, p < 0.05. Thus, the percentage increase of Bifidobacteria after FOS and GOS administration relative to water intake were approximately,+25% and +80%, respectively.</p><!><p>The levels of BDNF protein in extracts of frontal cortex did not differ between groups (Fig. 1A). However, BDNF in hippocampal extracts of FOS administered rats were significantly higher than those of control and GOS fed animals. Western blots revealed that GOS-fed rats contained significantly greater levels of NR1 immunoreactivity in the frontal cortex compared to control and FOS animals (Fig. 1B). Analysis of the hippocampus, however, revealed that FOS rats contained significantly more NR1 subunits than the other groups, though an increased trend (p = 0.058) was observed in GOS animals relative to controls.</p><!><p>On Western blots hippocampal, but not cortical, extracts from GOS-fed animals, contained significantly greater NR2A immunoreactivity compared to controls (Fig. 2). The level of NR2B in the frontal cortex and hippocampus, was not affected by prebiotic feeding.</p><!><p>Prebiotic administration increased the abundance of BDNF (Figs. 3A, C, E and 4) and NR1 (Fig. 3B, D and F) mRNAs in the dentate gyrus of the hippocampus, relative to controls. A reduction of BDNF mRNA in the CA3 subfield of GOS-fed rats was also observed (Fig. 3C). Densitometry confirmed significantly greater BDNF and NR1 expression in the dentate gyrus of prebiotic rats (Fig. 4A and B). The administration of GOS resulted in an elevation of NR2A (Fig. 4C), but not NR2B (Fig. 4D), mRNA in the dentate gyrus and CA1 subfield relative to controls and FOS-fed animals.</p><!><p>This study tested whether an elevation of gut bacteria increased central d-alanine concentrations by elevating the amounts of this d-amino acid in the gut and the circulation. The concentrations of free d-alanine in faecal pellets of GOS fed rats were significantly greater than control and FOS animals, with FOS administration resulting in intermediate levels of this d-amino acid (Table 1). Plasma d-alanine levels were significantly higher in GOS-fed rats compared to control animals (Table 1), and a slight, though not significant (p = 0.086), increase was observed in FOS-fed rats. Prebiotic administration did not alter the concentrations of other circulating amino acids (Table 2). Rats fed with GOS had a significantly higher concentration of d-serine in the frontal cortex compared to controls (Table 2), though the levels of all other amino acids remained unchanged after prebiotic feeding.</p><!><p>The administration of GOS significantly increased the concentration of plasma PYY, but neither FOS nor GOS significantly altered circulating GLP-1 (Table 3). An increased trend (p = 0.077) was noted in the plasma of FOS-fed rats. Glucose levels, body weight and plasma BDNF were not significantly changed after prebiotic feeding.</p><!><p>There was a significant correlation between the number of faecal Bifidobacteria and frontal cortex NR1 protein (Pearson's r = 0.713, p < 0.05), in a pooled analysis of control and experimental animals (n = 24). There was also an overall correlation between microbiota numbers and plasma d-alanine (r = 0.54, p < 0.05) and l-alanine (r = 0.52, p < 0.05). No other parameter measured in either plasma or brain correlated with faecal Bifidobacteria numbers. There was an overall significant correlation between the levels of cortical d-serine and NR1 protein (Pearson's r = 0.684, p = 0.01). Individual group analysis revealed that this association was only significant after GOS feeding (GOS: r = 0.96, p = 0.04; FOS: r = 0.68, p = 0.32; water: r = 0.01, p = 0.989).</p><!><p>The addition of plasma from rats fed with GOS, elevated the extracellular concentrations of BDNF compared to controls (Fig. 5A). These levels were similar to those from cells exposed to synthetic PYY. A non-significant increase of BDNF release was also noted after the addition of FOS plasma. In the presence of PYY antisera BDNF secretion by GOS plasma did not reach significance. In another experiment, the presence of GLP-1 antisera did not affect BDNF secretion from cells by GOS plasma (Fig. 5B), though a non-significant reduction of BDNF release after the addition of FOS plasma was noted.</p><!><p>Studies have shown that probiotics have psychotropic effects, but the neurobiological consequences of prebiotic intake have not been explored. The aim of the current study was to provide unequivocal evidence for neurochemical and molecular changes in the rat brain following prebiotic consumption, to prelude future functional analyses. The specific hypotheses tested were, first, that Bifidobacteria proliferation by prebiotics is associated with an increase in brain BDNF, as evinced with probiotics; and second, that prebiotic augmentation of commensal microbiota elevates central NMDAR subunits, given that these receptors are reduced in germ-free rodents. Our data support both suppositions and present substantial grounds for exploring the utility of prebiotics in the modulation of brain function. We have also offered an initial indication for the involvement of PYY, and potentially other gut hormones, in the augmentation of BDNF signalling following prebiotic intake. The interpretations and relevance of our findings are discussed in turn.</p><p>The elevated expression of BDNF and encoded protein in rats fed with FOS, is consistent with the effect of a Bifidobacterium probiotic (Bercik et al., 2011a; O'Sullivan et al., 2011) and the selective proliferation of these species with antimicrobials (Bercik et al., 2011a). Thus, FOS administration may have augmented the colonization of similar psychotropic strains, within the moderate overall increase in Bifidobacteria numbers relative to GOS fed rats. In view of these observations therefore, it was surprising that GOS did not alter the levels of hippocampal BDNF protein and, moreover, by a greater magnitude than FOS.</p><p>Unaltered BDNF protein after GOS feeding may have reflected the reciprocal change in BDNF mRNA in the dentate gyrus and CA3 region of the hippocampus. If these alterations were translated to protein, then it is unlikely that an overall change in BDNF would be detected in whole hippocampal homogenates. The measure of BDNF in regionally dissected hippocampus would test this possibility. Arguing against a causal link between gut bacterial densities and BDNF gene expression, is a similar increase of BDNF mRNA in the dentate gyrus of both FOS and GOS rats, in spite of the fewer numbers of Bifidobacteria in the FOS group. The possibility that prebiotics alter brain signalling independently of the gut microbiota cannot be ruled out, and a direct interaction between oligosaccharides and the gut mucosa has been shown to influence the response of the immune system (Bode et al., 2004; Eiwegger et al., 2010), which may then impact on brain chemistry. The elevation of gut hormones after prebiotic intake might also reflect a direct effect of oligosaccharides on the gut (see below).</p><p>The physiological relevance of a reciprocal change in BDNF mRNA in two regions of the hippocampus after GOS intake is difficult to interpret without functional measures. However, an elevation of BDNF gene expression in the dentate gyrus has been associated with antidepressant action (Kerman, 2012). A similar elevation of BDNF mRNA after FOS and GOS administration is, therefore, in keeping with a potential antidepressant/anxiolytic property of gut bacteria (Bercik et al., 2011a). The concomitant decrease of BDNF mRNA in the CA3 is more difficult to interpret, though one possibility is that we were observing the differential, activity-dependent expression of BDNF mRNA splice variants in each hippocampal subfield (Chiaruttini et al., 2008). Regional molecular and electrophysiological analyses of the rat hippocampus after GOS administration are, therefore, required.</p><p>The potential mechanisms underlying the elevation of cortical NR1 subunits after GOS feeding remains elusive, and although they correlate with Bifidobacteria numbers (see Section 3), elevated NR1 may have also been mediated by the direct physiological response of the gut to GOS such as the release of PYY (see discussion below). Nevertheless, the significant elevation of hippocampal NR1 mRNA after FOS and GOS administration is intuitive in view of data showing reduced NR1 in mice devoid of gut bacteria (Neufeld et al., 2011), though it is not clear why a parallel change in hippocampal NR1 protein occurred only after FOS feeding. Given the potential complexities of prebiotic actions already encountered (cf: reciprocal changes of BDNF mRNA abundance in GOS-fed rat hippocampus), unaltered hippocampal NR1 protein in GOS rats may be authentic, and mechanistically associated with the increase of NR2A subunits, which was not apparent after FOS intake. Altered NR1:NR2 subunit ratios in the rodent hippocampus following pharmacological and genetic manipulations are not unusual (Owczarek et al., 2011; Kato et al., 2012), and in the latter study, an increase of NR2A, but not NR1, subunits is associated with significant functional outcomes. Similarly, unaltered NR2B subunits after prebiotic administration may suggest that Bifidobacteria and/or the metabolic response to GOS ingestion may not affect the neurophysiological processes, such as long-term depression, which are associated with dynamic changes in NR2B subunits (Liu et al., 2004). Of course, the lack of functional information to support our interpretations of these data is the major caveat of this study.</p><p>The demonstration of elevated faecal and plasma d-alanine concentrations after GOS feeding is consistent with studies showing that gut bacteria, including Bifidobacteria, are a source of this d-amino acid (Konno et al., 1993). The significant reduction of blood l/d-alanine ratios confirms that this d-alanine was not directly derived from circulating l-alanine. In spite of this, central d-alanine concentrations remained unaltered after prebiotic feeding, and this may be because, higher plasma concentrations are required to initiate the appropriate kinetics for d-alanine uptake into the brain (Morikawa et al., 2007).</p><p>The current investigation also revealed that GOS-fed rats contained greater amounts of cortical d-serine than controls. Since the same animals also showed greater levels of cortical NR1 subunits which are the d-serine binding moiety, it is reasonable to propose that NMDAR mediated signalling was elevated in the frontal cortex. Of course, only direct electrophysiological analysis of NMDAR signalling after GOS feeding can prove this. Of note, the increase of cortical d-serine did not appear to arise from the plasma but may have been synthesized locally, as suggested by a slight, though not significant, rise of cortical l-serine after GOS feeding. That is, the synthesis of d-serine from l-serine by serine racemase (Sikka et al., 2010), may have been a homeostatic response (preservation of l/d-serine ratios) to either local elevations of the l-enantiomer, and/or a greater demand for d-serine. The possibility that d-alanine is linked to an elevation of central d-serine signalling, is suggested by studies demonstrating augmented d-serine release from neurons in the presence of d-alanine (Rosenberg et al., 2010).</p><p>We have confirmed that the administration of GOS to rats increases circulating concentrations of PYY (Overduin et al., 2013). It is reasonable to suggest, therefore, that the effects of GOS on brain signalling may be mediated by gut hormones. Earlier work has shown that PYY has the potential to affect the brain either directly (Connor et al., 1997; Nonaka et al., 2003) or via vagal nerve activity (Hernandez et al., 1994). Our in vitro data suggest that elevated brain BDNF expression after GOS intake, may have been mediated by plasma PYY. Another study has shown that plasma from mice administered with Bifidobacterium longum, did not change the expression of BDNF mRNA in cultured SH-SY5Y cells (Bercik et al., 2011b), and this was reasoned to be further evidence of a vagal, rather than a 'blood-borne' mediation of probiotic central effects. It is reasonable to assume that our discrepant data reflect the different parameters measured in both studies (i.e. BDNF release versus gene expression) and/or different actions of B. longum and prebiotics on brain BDNF. However, the presence of neuroactive substances in blood does not preclude the involvement of gut-brain vagal circuitry after prebiotic ingestion. The enteric secretion of PYY after GOS intake may directly affect local BDNF signalling in myenteric neurons of the gut (Boesmans et al., 2008), which innately influence vagal nerve activity (Murphy and Fox, 2010). Of course, since circulating PYY can also directly access brain PYY receptors (Hernandez et al., 1994; Nonaka et al., 2003), in all likelihood, the central effects of PYY would involve direct and indirect mechanisms. However, the mechanisms underlying the increase of hippocampal BDNF after FOS intake remain elusive. Whether PYY or, indeed, other factors such as short-chain fatty acids (Jakobsdottir et al., 2013), and/or the immune system (Vulevic et al., 2008, 2013), underlie the changes in BDNF and NMDAR signalling after prebiotics, requires further examination.</p><!><p>Our results have provided the necessary 'proof-of-principle' for the central actions of prebiotic consumption. The increase of hippocampal BDNF after prebiotic intake is consistent with a probiotic effect, and may have been a direct consequence of elevated gut Bifidobacteria numbers. However, an additional effect of gut hormones (e.g. PYY) or other mediators, such as the immune system resulting from direct oligosaccharide-gut interactions, cannot be ruled out. The elevation of NMDAR subunits after prebiotics is intuitive given their reduction in the brains of germ-free animals. Furthermore, the strong correlation between Bifidobacteria numbers and cortical NR1 levels presented in this report, further supports a link between the microbiota and central glutamate neurotransmission. Mechanistic investigations beyond the scope of the present study, are now required to ascertain the systems underlying the observed changes, and will also reveal if vagal nerve modulation is involved. Moreover, behavioural analysis in rats will ascertain if the changes in BDNF after prebiotics impart an anxiolytic action, or that increased NMDAR subunits translate to improved cognitive performance. Importantly, our study has provided sufficient cause to warrant further exploration into the utility of prebiotics in therapies of neuropsychiatric illness and which, by virtue of their ability to proliferate gut bacteria and stimulate neuroendocrine (and other) responses, may even prove to be more potent than probiotics.</p><!><p>Clasado Ltd, UK made a financial contribution towards the study, as part of a BBSRC scheme.</p><!><p>H.M.S., J.P.E.S. and P.W.J.B. made a substantial contribution to the conception and design of the study, and the analysis and interpretation of the data together with G.C., L.C., and H.M. H.M.S. and P.W.J.B. drafted the article, which was critically reviewed for intellectual content by G.T.</p>

PubMed Open Access

Expanding Coverage of the Metabolome for Global Metabolite Profiling

Mass spectrometry-based metabolomics is the comprehensive study of naturally occurring small molecules collectively known as the metabolome. Given the vast structural diversity and chemical properties of endogenous metabolites, biological extraction and chromatography methods bias the number, property, and concentration of metabolites detected by mass spectrometry and creates a challenge for global untargeted studies. In this work, we used Escherichia coli bacterial cells to explore the influence of solvent polarity, temperature, and pH in extracting polar and non-polar metabolites simultaneously. In addition, we explored chromatographic conditions involving different stationary and mobile phases that optimize the separation and ionization of endogenous metabolite extracts as well as a mixture of synthetic standards. Our results reveal that hot polar solvents are the most efficient in extracting both hydrophilic and hydrophobic metabolites simultaneously. In addition, ammonium fluoride in the mobile phase substantially increased ionization efficiency in negative electrospray ionization mode as demonstrated by 2.5-fold and 5.7-fold increases in the number of features detected and the average feature intensity, respectively. The improvement in sensitivity with ammonium fluoride resulted in 3.5 times as many metabolite hits in databases compared to ammonium acetate or formic acid enriched mobile phases and allowed for the identification of unique metabolites involved in fundamental cellular pathways.

expanding_coverage_of_the_metabolome_for_global_metabolite_profiling

INTRODUCTION<!>Materials<!>Growth of Escherichia coli<!>Metabolite extraction methods<!>\xe2\x80\x9cHot MeOH\xe2\x80\x9d<!>\xe2\x80\x9cHot EtOH/Ammonium Acetate\xe2\x80\x9d<!>\xe2\x80\x9cCold EtOH/Ammonium Acetate\xe2\x80\x9d<!>\xe2\x80\x9cBoilingWater\xe2\x80\x9d<!>\xe2\x80\x9cAcetone/MeOH\xe2\x80\x9d<!>\xe2\x80\x9cMPA\xe2\x80\x9d<!>\xe2\x80\x9cYM3\xe2\x80\x9d<!>-Standard Mix 1 and 2<!>LC-MS and MS/MS analysis<!>Data processing<!>Influence of the extraction method on the analysis of polar and non-polar metabolites<!>Exploring different C18 stationary-phases for metabolite profiling<!>Optimizing the mobile phase in negative ionization mode for global metabolite profiling<!>Discussion

<p>Metabolomics is a rapidly growing field focused on the profiling and quantification of small, naturally occurring compounds that collectively constitute the so-called metabolome. Electrospray ionization mass spectrometry coupled to liquid chromatography (LC-ESI MS) provides the most comprehensive technology for metabolomics studies1, 2. Yet, two different mass spectrometry-based approaches have been described in the field: targeted and untargeted metabolomics. In general, targeting a specific metabolite or a small group of distinct metabolites is associated with hypothesis-driven studies3 and involves optimization of chromatographic conditions (i.e., retention times) and selected reaction monitoring (SRM) transitions with pure standards4. Recent studies showed that up to 50–100 metabolites can be quantified using this approach5, 6. Untargeted metabolomics studies, in contrast, are designed to profile simultaneously the largest number of compounds possible and therefore have the capacity to implicate previously unexplored biochemical pathways7, 8. It is essential for the untargeted approach to maximize ionization efficiency of metabolites over a broad mass range (e.g., m/z 80–1000), since this determines the number and intensity (abundance) of the features to be analyzed. A feature is defined as a molecular entity with a unique m/z and retention time value. The number of features can be used as a general metric for the comprehensiveness of a global metabolomics study and thereby reflect overall coverage of the metabolome. The intensity of each feature is also important in that a certain threshold is needed for accurate relative quantification and further identification by MS/MS. The intensity of the precursor ion dictates the signal-to-noise ratio of the tandem MS fragment ions, which is of critical importance in untargeted metabolomics studies because metabolites are identified by comparing the entire fragmentation pattern of a naturally occurring compound to that of a pure standard.</p><p>Given the large number of molecules with chemical and structural diversity constituting the metabolome, the method used to biologically extract metabolites and separate them by using liquid chromatography is fundamentally related to the number of features detected by MS with a sufficient signal-intensity threshold. The basic philosophy of an untargeted metabolomics approach is to detect as many metabolites as possible to maximize the opportunity of identifying compounds that are dysregulated in a particular biological condition. Therefore, studies that comprehensively examine optimal extraction methods and LC/MS conditions to detect the largest number of metabolites simultaneously are important. The isolation of metabolites from tissues, cells, or biofluids requires that proteins are precipitated and that polar and non-polar metabolites are dissolved in solution without degradation. Further, the ideal chromatographic conditions should retain and separate the complex mixture of extracted metabolites using a mobile phase (solvent) that promotes ionization of the largest number of analytes.</p><p>In the present study, we used Escherichia coli as a model organism to optimize metabolite extraction and chromatography conditions coupled to ESI-MS. Specifically, we applied seven extraction methods involving different aqueous/organic solvents, temperature, pH, and molecular weight filters to E. coli cultures. Additionally, we explored the separation of 36 model compounds of varying polarity from a standard mixture using different reverse-phase columns containing unique stationary phases. Finally, we investigated the influence of the chromatographic mobile phase to enhance ionization efficiency of complex mixtures of metabolites in negative ionization mode, which is generally associated with the detection of fewer features relative to positive ionization mode. Overall, our results indicate that polar solvents (e.g., water, ethanol/water) in combination with high temperature are more efficient in extracting both hydrophobic and hydrophilic metabolites compared to less polar solvents such as acetone or methanol. The choice of the chromatographic mobile-phase conditions in negative ionization mode proved to be strikingly significant, and we report that the addition of ammonium fluoride substantially increased the absolute intensity of nearly all compounds analyzed up to 22 fold while also increasing the total number of features in E. coli extracts by 2.50 fold compared to mobile phases enriched with ammonium acetate or formic acid.</p><!><p>All pure standards were purchased from Sigma Aldrich (St. Louis, USA), except Peptide T (GenScript, Piscataway, USA) and LysoPC (Cayman Chemical, Ann Arbor, USA). Peptides Phe-Gly-Phe-Gly and Thymopoietin II fragment 32–36 were custom synthesized. Ammonium acetate, ammonium fluoride, formic acid, and EDTA were purchased from Sigma Aldrich (St. Louis, USA). LC/MS grade methanol, acetonitrile, and water were purchased from J.T. Baker (Phillipsburg, USA). M9 minimal salts (5×) and casamino acids were purchased from Difco (Franklin Lakes, USA), and glucose, MgCl2, CaCl2, and meta-phosphoric acid (MPA) from Sigma Aldrich (St. Louis, USA). The Microcon YM-3 (3,000 NMWL) devices and HPF Millex filters (hydrophilic PTFE, 0.20 μm) were purchased from Millipore (Billerica, MA).</p><!><p>The E.coli strain MC4100 (F-, araD139, Δ(arg F-lac)U169, ptsF25, relA1, flb5301, rpsL 150.λ-) was grown overnight at 37 °C in minimal media containing M9 salts (×1), 2% casamino acids, 0.2% glucose, 1mM MgCl2, and 0.1mM CaCl2.</p><!><p>An overnight batch culture of E.coli (25 mL) was divided in seven aliquots (1.5mL each) and the supernatant was removed by centrifugation (5 minutes at 3,000 rpm, 4 °C). Metabolites were extracted by using one of the following methods:</p><!><p>200 μL of hot (80 °C) methanol (100%) was added to the E.coli pellet, vortexed for 30 seconds, and incubated 1–2 minutes at 80 °C (oven). A heating block or other methods can be used to control temperature. The solution was centrifuged for 15 minutes at 13,000 rpm (4 °C) and the supernatant transferred to a new tube. The same process was repeated with the precipitate using 100 μL of hot methanol. Finally, the supernatants (~300 μL total) were pooled in an HPLC vial.</p><!><p>400 μL of hot (80 °C) 60% ethanol/40% water in 5mM ammonium acetate, 1mM EDTA (pH 7.2) was added to the E.coli pellet, vortexed for 30 seconds, and incubated 1–2 minutes at 80 °C (oven). A heating block or other methods can be used to control temperature. The solution was centrifuged for 15 minutes at 13,000 rpm (4 °C) and the supernatant transferred to a new tube. The same process was repeated twice, and the resulting supernatants were pooled in a 1.5 mL tube. The solution was desiccated with a vacuum concentrator (SpeedVac) at room temperature, and redissolved in 300 μL of 5% ethanol/water 5mM ammonium acetate (pH 7.2). Finally, the sample was centrifuged again for 10 minutes at 13,000 rpm (4 °C) and the supernatant was transferred to an HPLC vial.</p><!><p>300 μL of cold (4 °C) 60% ethanol/ 40% water in 5mM ammonium acetate, 1mM EDTA (pH 7.2) was added to the E.coli pellet, vortexed for 30 seconds, and the sample incubated 1 minute in liquid nitrogen. The sample was thawed at room temperature, and incubated in liquid nitrogen two more times. Next, the sample was incubated 1 hour at −20 °C followed by a 15-minute centrifugation at 13,000 rpm. The resultant supernatant (~300 μL) was transferred to an HPLC vial.</p><!><p>300 μL of LC/MS-grade water in 1 mM HEPES and 1 mM EDTA (pH 7.2) was added to the E.coli pellet, vortexed for 30 seconds, and the sample incubated 1–2 minutes in boiling water. Next, the sample was incubated for 1 minute in liquid nitrogen and thawed at room temperature. The incubation in liquid nitrogen was repeated. Finally, the sample was incubated 1 hour at −20 °C, followed by 15 minutes centrifugation at 13,000 rpm (4 °C). The resultant supernatant was transferred to an HPLC vial.</p><!><p>400 μL of cold (−20 °C) acetone was added to the E.coli pellet, vortexed for 30 seconds, and the sample incubated 1 minute in liquid nitrogen. The sample was thawed at room temperature, and incubated in liquid nitrogen two more times. After 1 hour at −20 °C, the sample was centrifuged at 13,000 rpm for 15 minutes. The resultant supernatant was transferred to a separate vial and the precipitate mixed with 200 μL of cold methanol/water/formic acid (86.5:12.5:1.0). The sample was vortexed for 30 seconds and then sonicated for 10 minutes (4 °C) before leaving the sample 1 hour at −20 °C. Next, the sample was centrifuged 15 minutes at 13,000 rpm (4 °C) and the supernatant pooled with the previous. The solution was dried with a vacuum concentrator (SpeedVac) at room temperature, and redissolved in 300 μL of 95% acetonitrile/5% water. The final solution was then centrifuged for 10 minutes at 13,000 rpm and the supernatant transferred to an HPLC vial.</p><!><p>300 μL of cold (4 °C) solution of 5% meta-phosphoric acid (MPA), 1mM EDTA, 0.1% formic acid was added to the E.coli pellet, vortexed for 30 seconds, and the sample incubated 1 minute in liquid nitrogen. The sample was thawed at room temperature and incubated in liquid nitrogen two more times. Next, the sample was centrifuged 15 minutes at 13,000 rpm (4 °C) and the resultant supernatant (~300 μL) was transferred to an HPLC vial. It should be noted that the MPA buffer was filtered with a HPF Millex filter (Millipore) before mixing with the E.coli pellet.</p><!><p>300 μL of cold (4 °C) LC/MS-grade water in 1mM HEPES and 1mM EDTA (pH 7.2) was added to the E.coli pellet, vortexed for 30 seconds, and the sample incubated 1 minute in liquid nitrogen. The sample was thawed at room temperature and bath sonicated for 5 minutes in cold water. This process was repeated two more times. Next, the sample was centrifuged 10 minutes at 13,000 rpm (4 °C) and the resultant supernatant was transferred to a microcon centrifugal filter unit with a YM-3 membrane (3 kDa NMWCO). The solution was spun down 1 hour (4 °C) and the retentate recovered in an HPLC vial. It should be noted that the membrane was spin-rinsed with deionized water 5 times to remove trace amounts of glycerin before applying the sample.</p><!><p>Each compound was dissolved in 50% methanol/water and prepared at a final concentration of 10 μM and 0.1 μM for LC/MS analysis.</p><!><p>Analyses were performed using an HPLC system (1200 series, Agilent Technologies) coupled to a 6538 UHD Accurate-Mass Q-TOF (Agilent Technologies) operated in positive (ESI+) or negative (ESI−) electrospray ionization mode. Vials containing extracted metabolites using one of the seven methods described above or the standard mixture were kept at −20 °C prior to LC/MS analysis. E.coli extractions and standard mixtures were separated using a Cogent Bidentate C18: 4 μm, 100Å, 150mm × 2.1 mm ID (Cat No. 40018-15P-2), a Waters XBridge C18: 3.5 μm, 135Å, 150mm × 1.0 mm ID (Part No. 186003128), or an Imtakt Scherzo SM-C18: 3 μm, 13nm, 150mm × 2mm ID (Prod. No. SM025) column. When the instrument was operated in positive ionization mode, regardless of the column used, the solvent system was: A= 0.1% formic acid in water, and B= 0.1% formic acid in acetonitrile. When the instrument was operated in negative ionization mode, we used four different solvent systems: A1= 0.1% formic acid in water, B1= 0.1% formic acid in acetonitrile; A2= 1mM ammonium fluoride in water, B2= acetonitrile; A3= 5mM ammonium acetate in water, B3= 5mM ammonium acetate in 90% acetonitrile; A4= 1mM ammonium acetate in water, and B4= 200mM ammonium acetate in 50% acetonitrile. The linear gradient elution used started at 100% A (time 0–5 minutes) and finished at 100% B (35–40 minutes). The injection volume was 8 μL. ESI conditions were: gas temperature 325 °C, drying gas 11 L/min, nebulizer 30 psig, fragmentor 120 V, and skimmer 65 V. The instrument was set to acquire over the m/z range 80–1000 with an acquisition rate of 1.3 spectra/second. MS/MS was performed in targeted mode and the instrument was set to acquire over the m/z range 50–1000, with a default iso width (the width half maximum of the quadrupole mass bandpass used during MS/MS precursor isolation) of 4 m/z. The collision energy was fixed at 20V.</p><!><p>LC/MS data from the E.coli extractions (ESI+ and ESI− modes) were processed using the XCMS software9 (version 1.24.1) to detect and align features. Each metabolite extraction method was compared using the same column (Cogent Bidentate C18) and ionization mode (ESI+). XCMS analysis of these data provided a matrix containing the retention time, m/z value, and intensity of each feature for every extraction method discussed above. Each row in the matrix represented a feature. It is important to note that while the retention time and m/z values for each feature were consistent among extraction methods, the intensities of the features varied. Using the statistical software R, each row of feature intensities was normalized such that the highest value was one. A two-dimensional representation of this matrix was calculated using multidimensional scaling (MDS) as implemented within the software R. In negative ionization mode, experimental blanks were run in triplicate to remove "background" features arising from the mobile phases (e.g., ammonium fluoride and ammonium acetate). "Background" features detected in each of the three blanks were removed from E.coli data run with the same mobile phase. Standards were manually quantified by extracting ion chromatograms and integrating peak intensities with Qualitative Analysis of MassHunter Workstation (Agilent Technologies).</p><!><p>E.coli pellets obtained from the same batch culture were used to extract metabolites with the seven different protocols (see Experimental section). The selected protocols represent examples in which different fundamental conditions for metabolite solubility and stability are varied such as solvent polarity, temperature, pH, and molecular weight cut-off filtering. In brief, in method "Hot MeOH" metabolites were extracted using hot 100% methanol. Method "Hot EtOH/Ammonium Acetate" was modified from Buescher et al.10, with metabolites extracted using hot ethanol/water buffered at pH 7.2. In both methods the solvents were pre-heated and mixed with the E.coli pellet at 80 °C for a short period (~1–2 minutes) to avoid thermal degradation or methyl/ethyl ester formation. Method "Cold EtOH/Ammonium Acetate" used the same solvent as method "Hot EtOH/Ammonium Acetate", but the solvent was pre-cooled in ice and the extraction was performed at low temperature. Method "BoilingWater" was modified from Bhattacharya et al.11, with metabolites extracted using a polar solvent (100% H2O) buffered at pH 7.2 and incubated for a short time in boiling water. In method "Acetone/MeOH", metabolites were extracted at low temperature using a pre-cooled non-polar solvent (100% acetone). Method "MPA" was modified from Rellan-Alvarez et al.12, with metabolites extracted using strong acidic/aqueous conditions (meta-phosphoric acid in water) at low temperature. And, finally, in method "YM3" metabolites were extracted using a cold aqueous solvent buffered at pH 7.2 and isolated using a centrifugal filter unit with a cut-off of 3 kDa. All extractions were run as triplicates under the same LC/MS conditions using the Cogent Bidentate reverse-phase C18 column in positive ionization mode (see Experimental section). Each data set was visualized using a multidimensional scaling plot (Figure 1A) to show similarities in the data. Using feature intensities from the XCMS matrix, the data were scaled such that similar methods are near each other and dissimilar methods are farther apart from each other. The multidimensional scaling plot shows short distances between methods "Hot EtOH/Ammonium Acetate", "Cold EtOH/Ammonium Acetate", and "BoilingWater", highlighting that the number and intensities of features detected by XCMS in each method are similar. "Hot MeOH" is also relatively similar to the previous methods. Three extraction protocols, however, are strikingly different from the others: "Acetone/MeOH", "MPA", and "YM3". We deduced that "Acetone/MeOH" is the most non-polar of the seven extractions, and "YM3" is the most polar.</p><p>The mean intensity of 31 specific metabolites identified in E. coli using MS/MS data are shown in Figure 1B. Metabolites characterized by different polarities and chemical functional groups were extracted with different efficiencies based on the method used. The intensity of a feature as determined by its integrated LC/MS peak area has important implications in that metabolite identification requires MS/MS analysis. Tandem MS in Q-TOF analyzers requires ion isolation and, without sufficient signal to noise, produces ambiguous and unreliable fragment ions. Based on our experimental conditions, we established a threshold for which the intensity of detected metabolites did not allow for reliable MS/MS fragmentation (see black line in Figure 1B). Our results indicate that the "Acetone/MeOH" protocol is the least efficient extraction to profile hydrophilic and hydrophobic metabolites simultaneously. Although this method provided increased detection of 4 phospholipids, only 14 out of the 31 metabolites shown were detected. In contrast, method "YM3" did not efficiently extract the 4 phospholipids or other structurally unrelated metabolites such as FAD, FMN, and CoA. In addition, the intensity of metabolites such as acetyl-CoA, succinyl-CoA, γ-glutamylcysteine, adenylsuccinic acid, and ATP/ADP was below or slightly above the MS/MS threshold level. Using method "Hot MeOH" we detected 29 out of the 31 metabolites shown, with the exceptions being succinyl-CoA and ATP. It should be noted, however, that the intensity of 4 metabolites (acetyl-CoA, FAD, CMP, and adenylsuccinic acid) was below or slightly above the MS/MS threshold. Using method "MPA" all 31 metabolites were detected with intensities above the established MS/MS threshold, but we observed an unexpected high analytical variability for most of the compounds (see error bars in Figure 1B). Methods "Hot EtOH/Ammonium Acetate", "Cold EtOH/Ammonium Acetate", and "BoilingWater" appear to be the most efficient extraction protocols examined in this study, and the similarity of the results from these three methods is consistent with the clustering displayed in Figure 1A (see blue area). With the exception of succinyl-CoA, the rest of the 31 metabolites were detected with high reproducibility in all three extractions. Nineteen of the metabolites showed higher intensity with "Hot EtOH/Ammonium Acetate", as compared to 9 metabolites with "BoilingWater", and only 2 with "Cold EtOH/Ammonium Acetate". Importantly, none of the 31 metabolites formed methyl or ethyl esters due to the short incubation time in bowling water or hot ethanol. Overall, we interpret our results in decreasing order of efficiency in extracting both polar and non-polar metabolites simultaneously as follows: "Hot EtOH/Ammonium Acetate" < "BoilingWater" < "Cold EtOH/Ammonium Acetate" < "MPA" < "Hot MeOH" < "Acetone/MeOH" = "YM3".</p><!><p>Reliable quantification of thousands of chemically diverse features requires optimization of chromatographic separation to reduce ion-suppression effects. Traditionally, untargeted metabolomics analyses have been performed using reverse-phase (RP) C18 columns because they generally result in the detection of more features. RP C18 columns, however, are limited in their capacity to retain hydrophilic compounds and consequently result in ion suppression for polar metabolites in the dead volume, thereby limiting MS coverage of the metabolome. Recent developments in RP C18 technology offer opportunities to improve retention of polar molecules and therefore increase metabolite coverage in metabolomics analysis. We analyzed the ability of three different RP C18 columns with unique properties to retain polar model compounds: (i) the XBridge C18 column characterized by broad pH range stability (pH 1–11), (ii) the Cogent Bidentate C18 column characterized by silicon-hydride (Si-H) groups instead of the common silanol group (Si-OH)13, 14, and (iii) the multi-modal Scherzo SM-C18 column containing cation and anion ligands that allow for reverse-phase separation in addition to anion and cation exchange. We analyzed a standard mixture of 31 model compounds characterized by different polarities (e.g., amino acids, tricarboxylic acids, vitamins, peptides and xenobiotics, see Figure 2 for full list of compounds in Standard Mix 1) using each column with the same gradient and mobile phase (A= water, 0.1% FA; B= acetonitrile, 0.1% FA). With the XBridge column, 14 out of the 31 compounds (45%) co-eluted within the first two minutes (Figure 2A). Similar results were obtained with traditional C18 columns (e.g., Zorbax, Atlantis T3) (data not shown). The Cogent column retained more compounds, with only 7 of them (22%) co-eluting within the first 2 minutes (Figure 2B). With the Scherzo column, we observed a remarkable improvement in the retention of polar compounds with only 3 compounds (10%) co-eluting within the first two minutes (Figure 2C). Overall, all three columns showed good performance in separating the rest of the compounds, although taurocholic acid, coenzyme A (CoA) and acetyl-CoA did not elute from the Scherzo column due to strong ion exchange interactions with the stationary phase. Increasing the ionic strength of the mobile phase by adding 5 mM ammonium acetate resulted in elution of taurocholic acid. Further increasing the ionic strength of the mobile phase with 200 mM ammonium acetate resulted in the elution of CoA and acetyl-CoA.</p><!><p>In electrospray ionization, the composition of the solvent (i.e., mobile phase in LC-ESI MS) influences the gas-phase acid-base processes required for the ionization of the compounds to be analyzed. Global profiling of metabolites in positive ionization mode generally produces more features compared to negative mode, most likely due to higher efficiency of protonation relative to deprotonation. Protonation is facilitated by the addition of an acid to the mobile phase, such as formic acid, acetic acid, or TFA. In negative ionization mode, deprotonation in the gas phase is typically promoted by the addition of ammonium salts such as ammonium formate or ammonium acetate. Given the strong basicity of the fluoride anion (F−) in the gas phase15, we examined the effect of adding ammonium fluoride to the mobile phase in negative-mode analysis. A previous report showed that fluoride anion in the electrospray solvent resulted in increased deprotonation, [M-H]−, of neutral steroids with higher abundances than other anions tested16. Using Standard Mix 2 containing 36 compounds at 10 μM (see Figure 3 for full list of compounds), 3 different mobile phases were compared with the same reverse-phase C18 column (i.e., XBridge) and MS conditions in negative ionization mode: (i) 1 mM ammonium fluoride, (ii) 5 mM ammonium acetate, and (iii) 0.1% formic acid. Mobile phases enriched with ammonium fluoride and ammonium acetate were maintained at pH~ 7, ruling out the possibility that differences in ionization efficiency are related to the pH of the solution.</p><p>Figure 3 shows the intensity values and fold change of each compound, revealing that ammonium fluoride is a superior additive to increase ionization efficiency in negative ionization mode. On average, ammonium fluoride increased the intensity of the compounds analyzed by 5.7 fold, including a 15-, 16-, and 22-fold increase for quinidine, peptide T, and lysophosphatidylcholine respectively. Notably, when the same 36 compounds were prepared at 0.1 μM, 20 compounds (55%) were not detected with 0.1% formic acid, and 18 (50%) were not detected with ammonium acetate. Only 6 compounds (16%) were not detected with ammonium fluoride, and all remaining compounds showed significantly higher intensity values with ammonium fluoride compared to ammonium acetate and formic acid (data not shown). It is worth noting that lysophosphatidylcholine was detected as [M-H]− at 10 μM in negative ionization mode using ammonium fluoride, however no signal of this compound was observed at 0,1 μM. Lysophosphatidylcholine, however, was detected at 0,1 μM in positive ionization mode using 0,1% formic acid.</p><p>The total number of features and their absolute intensity were also compared from E.coli extractions analyzed using ammonium fluoride and ammonium acetate enriched mobile phases. E.coli samples in addition to blank samples were run as triplicates with each mobile phase, and features consistently present in all three blanks were subtracted from E.coli data as "background". Then, only features present in the 3 E.coli replicates with intensity values above 5000 ion counts were considered for quantification purposes. The Venn diagram in Figure 4A shows a total of 4,213 features obtained with the method "BoilingWater" using an ammonium fluoride enriched mobile phase, and 1,647 features from the method "BoilingWater" using an ammonium acetate enriched mobile phase, that is, a 2.5-fold increase. Importantly, 77% of the features analyzed using ammonium fluoride additive were not detected using ammonium acetate additive. A similar distribution of unique features was also found when the method "BoilingWater" and the method "Hot MeOH" enriched with the same mobile phases were compared. A total of 1,008 features were detected using "Hot MeOH" with an ammonium acetate enriched mobile phase, and only 24% of these features overlapped with "BoilingWater" using the same mobile phase (data not shown). The increased sensitivity achieved by using an ammonium fluoride enriched mobile phase was also reflected in the number of putative metabolite identifications made on the basis of accurate mass from the METLIN database. Using ammonium fluoride, 722 features out of 4,213 (17.1%) matched with known compounds in METLIN (error <5 ppm, [M-H]−). Using ammonium acetate, 211 out of 1,647 (12.8%) matched with known compounds in METLIN, reflecting almost a 3.5-fold increase in database hits. It is noteworthy that 81% of the METLIN database hits using ammonium fluoride were not present in the ammonium acetate analysis. Features with high intensity, presumably corresponding to abundant metabolites, were typically detected with both enriched mobile phases; whereas less abundant metabolites were often uniquely detected in ammonium fluoride enriched analyses. To confirm this observation the mean intensity value was determined for various metabolites structurally identified based on accurate mass, retention time, and MS/MS data (Figure 4B). The results were consistent with our previous data using Standard Mix 2 containing 36 compounds. All metabolites identified showed higher intensity using an ammonium fluoride enriched mobile phase, with important cellular metabolites being detected only in the presence of ammonium fluoride.</p><!><p>Global metabolite profiling of biological samples is a challenging task due to the chemical and structural diversity of naturally occurring compounds ranging from polar metabolites such as amino acids and nucleotides to non-polar molecules such as steroids and membrane lipids. As a result of this diversity, methods used for metabolite extraction and metabolite separation significantly influence the number and intensity of compounds detected by ESI-MS analysis. The choice of extraction and chromatography biases the chemical distribution of metabolites detected and is therefore problematic for untargeted metabolomics investigations aimed at accomplishing unbiased profiling. The obvious complexity in performing untargeted studies is that the metabolites of potential interest are unknown and therefore extraction and chromatography methods cannot be tailored towards a specific chemical class of compounds. In this context, it is unclear what evaluation criteria should be used to assess the quality of untargeted metabolomics extraction and chromatography methods. The approach we have developed here involves using our metabolomics software XCMS to analyze the number and intensity of features identified in each of different extraction protocols and LC/MS conditions. We interpret methods leading to the identification of more features of greater intensity to be better suited for untargeted studies.</p><p>Decades of research has provided an extensive library of detailed extraction and chromatography methods for analysis of unique classes of compounds, and it is not our intent to comprehensively survey all of them here. Rather, our study is aimed at providing an overview of easy and rapid extraction and RP LC/MS protocols for the metabolomics scientist to use as a general guideline. Optimization of metabolite extraction has been recently pursued using a two stage approach or biphasic mixtures 17. Although these methods may reduce complexity in the number of metabolites to be separated and ionized, and may ultimately extend metabolite coverage, we have not explored them because they are generally more time-consuming and prone to analytical error. Other studies have shown that many metabolites can be extracted by a broad spectrum of solvent mixtures 18–20, suggesting that there is no specific extraction protocol that should be used in metabolomics. Our data corroborate this point, but highlight that a number of considerations should be taken into account when carrying out global metabolite profiling studies. For example, highly polar solvents such as water, or 100% nonpolar organic solvents such as acetone/methanol are not particularly suitable to extract both hydrophilic and hydrophobic metabolites simultaneously. In particular, solvents with intermediate polarity such as a mixture of ethanol/water are more appropriate for this goal. We also report that incubation of the sample with the solvent at high temperatures (80–100 °C) for short periods of time (1–2 minutes) is more efficient than incubation of the sample at cold temperatures (−20 °C). Water may have a polarity similar to some organic solvents when boiled 21, which might explain the similarity of the method "BoilingWater" with extraction protocols using high percentages of ethanol such as the methods "Hot EtOH/Ammonium Acetate" and "Cold EtOH/Ammonium Acetate". The general utility of the described extraction protocols to other cell types that have heavily weighted distributions of non-polar metabolites (e.g., adipose tissue), possess increased tensile properties (e.g., muscle tissue), or contain functionalities that are not stable to the conditions employed requires further investigation.</p><p>Much attention has been dedicated recently to different chromatographic approaches to improve separation and analysis of naturally occurring compounds. HILIC-like stationary phases22 or the use of ion pairing agents with reverse-phase columns 23–25 improves the retention of polar metabolites, however, each of these approaches has drawbacks for processing global metabolite profiling data. HILIC columns do not retain hydrophobic compounds well, and are generally associated with broader peak shapes. The use of ion pairing agents (e.g., TBA) has lead to inconsistent results, lengthy equilibration, general incompatibility with mass spectrometry, and typically requires a dedicated negative-mode LC stack as some ion pairing agents are difficult to remove from the instrument lines and cause contamination affecting analysis in positive-ionization mode. In contrast, we demonstrate that the stationary phases utilized in our study show unique advantages for global metabolite profiling. We found the flexibility of the Scherzo column to be particularly effective, as it combines a non-polar stationary phase with anion and cation ligands for ion exchange separation. Perhaps the major drawback is the strong ionic interaction of this stationary phase with some metabolites such as coenzyme A. This issue can be resolved by increasing the ionic strength of the mobile phase with 200 mM ammonium acetate, however, it is important to mention that percentages of acetonitrile higher than 50 led to precipitation of the ammonium salt, and low percentages of acetonitrile likely affects the elution of highly hydrophobic compounds from the column.</p><p>Finally, a striking result from our work is the substantial improvement of ionization efficiency in negative ionization mode using an ammonium fluoride enriched mobile phase. To the best of our knowledge, this approach has not been described before for metabolomics studies. The use of ammonium fluoride enriched mobile phase in negative ionization mode is not intended to replace positive ionization mode analysis in metabolomics. Ammonium fluoride is stable in different solvents (e.g., water, acetonitrile, methanol) and led to reproducible results. In addition, no contamination of the HPLC lines was observed after using ammonium fluoride. Importantly, the highest sensitivity was achieved using 1 mM ammonium fluoride, and concentrations of 5 mM or higher introduced significant background in the mass spectra. The source of background was attributed to the impurities of the ammonium fluoride stock solution and to the interaction at pH 7.0 of fluoride anions with the silica of the stationary phase of most columns tested. The XBridge column, however, is stable within a wide pH range and performed excellent with 1mM ammonium fluoride, generating mass spectra with low background. When a polymeric (PSDVB) reverse-phase column was tested with ammonium fluoride, such as the Hamilton PRP-1, background was almost absent (similar to ammonium acetate) with no effect on sensitivity. Overall, the consistency of our data with standard compounds and biological samples suggests that ammonium fluoride should be used as a standard additive to mobile phases in negative ionization mode for global metabolomics studies.</p>

PubMed Author Manuscript

Active Metal Template Synthesis and Thermal Actuation of a Nanohoop [c2]Daisy Chain Rotaxane

Molecules and materials that demonstrate large amplitude responses to minor changes in their local environment play an important role in the development of new forms of nanotechnology. Molecular daisy chains are a type of a mechanically interlocked molecule that are particularly sensitive to such changes where, in the presence of certain stimuli, the molecular linkage enables muscle-like movement between a reduced-length contracted form and an increased-length expanded form. To date, all reported syntheses of molecular daisy chains are accomplished via passivetemplate methods, resulting in a majority of structures being switchable only through the addition of an exogenous stimuli such as metal ions or changes in pH. Here, we describe a new approach to these structural motifs that exploits a multi-component active-metal template synthesis to mechanically interlock two pi-rich nanohoop macrocycles into a molecular daisy chain which we show can be actuated through simple thermal changes.

active_metal_template_synthesis_and_thermal_actuation_of_a_nanohoop_[c2]daisy_chain_rotaxane

<p>When downsizing machinery to the nanoscale, the ability to mimic design principles of macroscopic machinery becomes increasingly difficult. For example, at the molecular scale, objects undergo sporadic, unplanned movements owing to Brownian motion. [1] However, in the macroscopic world, objects are not strongly affected by such random motion and as a result, the planned, orchestrated movement of numerous components in a machine is readily accomplished. Developing strategies to overcome hurdles such as this have therefore emerged as a key focus point for generating new types nanotechnology. Mechanically interlocked molecules such as rotaxanes and catenanes represent a unique class of molecules that display properties of motion common to both micro and macroscopic objects, [2] suggesting structures of this form can provide a platform to merge macroscopic principles with molecular scale systems. Importantly, the specific connectivity of the mechanical linkage dictates what type of motion can be harnessed which, in principle, can be leveraged via a variety of actuation mechanisms such as pH, metal-coordination, and redox-chemistry. [2] With a name originating from the similarity between daisy chain garlands, a less explored, yet particularly powerful mechanically interlocked molecular architecture is the daisy chain rotaxane. [3] The simplest form can be understood as a [c2]daisy chain (Figure 1a), where c refers to cyclic and 2 represents the number of mechanically interlocked components. A prized feature of [c2]daisy chains is the ability to actuate these motifs between extended or contracted states (Figure 1a). This results in muscle-like motion which can give rise to large changes in molecular dimensions [3a] and shape. [3c] Thus, when viewed as components for investigating nanoscale motion through macroscopic actuation, these structures represent especially fascinating targets. To date, a majority of the reported daisy chain architectures are actuated through light, [3d] pH, [3d-g] electrochemical, [3h, i] and metalcoordination based mechanisms, [3a,c,j] with thermally actuated systems being relatively rare. [3d] This likely stems from the typical method of preparation, where interactions such as metal-coordination and radicalradical pairing template the formation of the desired daisy chain. [3] These passive-template approaches result in a mechanically interlocked structure where the employed host-guest interaction "lives-on" in the final product, resulting in strong intercomponent binding energies. While the mechanical bond allows each component to translate, rotate, or contract relative to each other, it is the molecular composition of the underlying components that determine properties such as rate of shuttling, strength of contraction or mechanism of actuation. Structurally distinct from typical interlocked molecules, pi-rich molecular nanocarbons [4] such as cycloparaphenylenes or "nanohoops" [4] have emerged as interesting platforms for modulating and studying motion at the nanoscale. For example, a recent report by Itami et al. showed that a fully conjugated all-benzene trefoil knot displays rapid vortex-like motion at temperatures far below room temperature. [4l] Additionally, work by Isobe et al. has demonstrated that nanohoop macrocycles can provide a pathway to ratchet-free solid-state motion. [4m] Recently, we demonstrated a synthetic method to thread sphybridized fragments (diynes) through small-diameter pi-rich macrocycles via an active metal Cadiot-Chodkiewicz template (AT-CC) synthesis. [4n] Fundamentally different from a passive-template strategy, this AT-CC approach provides a mechanically interlocked molecule that contains no complimentary interaction between macrocycle and thread component in the final interlocked molecule. [5] As a result, weaker interaction energies between thread and macrocycle are observed [5c] -a feature that we anticipated would allow for simple thermal actuation of the resulting [c2]daisy chain. Encouraged by this prospect and the emerging properties of nanocarbons, [4] we report here the formation of a [c2]daisy chain rotaxane bearing two carbon-rich nanohoop macrocycles (Figure 1c) that have been mechanically interlocked via two sp hybridized diyne threads. The key mechanical bonds are forged via AT-CC reaction, providing the final structure in good yield. We show that this structure can rapidly contract and expand via thermal actuation-a rare consequence of both structural composition and the AT-CC method of preparation.</p><p>Our work first began by investigating the feasibility of forming a [c2]daisy chain structure via an AT-CC reaction. We opted to expand on our recent work with 2,6-pyridine-embedded nanohoop macrocycles with the understanding that an AT-CC would likely allow for the formation of the desired structure. Accordingly, we began by first preparing a new functionalized nanohoop macrocycle containing an alkyl-linked terminal acetylene (Figure 2). To avoid introducing chirality into the nanohoop macrocycle, we targeted the meta-pyridine ring as the key point for functionalization. The synthesis began (Figure 2) by first preparing diboronate 1 and pyridine derivative 2 via literature procedures. [4n] Leveraging our previous synthetic strategy, we found that macrocycle 3 was readily accessed in 70% under dilute Suzuki-Miyaura conditions with 1 and 2. Macrocycle 3 was then deprotected with tetrabutylammonium fluoride (TBAF) followed by reductive aromatization with H2SnCl4 to give the corresponding benzylic alcohol 4 in 61% over two steps. Conversion to the alkyne-functionalized macrocycle was accomplished via deprotonation with NaH in the presence of excess propargyl bromide in THF, giving the target macrocycle 5 in excellent yield. It should be noted that the formation of [cn]daisy chain structures requires n (n>1) mechanical bonds to be formed, where n represents the number of units, i.e. bonds formed, that comprise the final structure. While this has been observed in multi-component active metal template syntheses of catenanes [5d] as well in a related molecular knot, [5f] it has not been reported for the targeted daisy chain architecture. Accordingly, with 5 in hand, we then proceeded to explore the formation of the desired [c2]daisy chain via an AT-CC reaction (Figure 3a). We first subjected 5 to our previously reported [4n] AT-CC conditions (Toluene, 80 o C, K2CO3, [Cu(MeCN)4]PF6) with 3,5-diester functionalized bromo-alkyne 6 (Figure 3) under dilute (5 mM) conditions, however, after purification, we observed a minor amount of cross-coupled unthreaded product along with significant amounts of unreacted starting materials. On closer inspection of the crude reaction mixture via 1 H-NMR spectroscopy, we observed a small amount of an unidentified third product that we anticipated was the desired [c2]daisy chain. Isolation of this product revealed a relatively simple 1 H-NMR spectrum (Figure 3c) with several proton resonances appearing noticeably upfield (for example, resonance HD)-evidence that suggested an interlocked molecule (See SI for all spectral data). Mass spectrometry of this product revealed an m/z that corresponded to two macrocyclic units (m/z = 1542.4724), allowing us to tentatively assign the structure to that of 7 (Figure 3a). Intrigued by this small formation of 7, we then screened various conditions in an effort to selectively prepare this product. We found that when the reaction was run under dilute conditions (7 mM) at 60 o C using (N,N-Diisopropylethylamine) DIPEA as base and using a chloroform:ethanol (1:1) solvent system, [5d] the desired product 7 could be isolated in 32% yield via chromatography. This is rather impressive given the demanding nature of the product-four separate molecular components must assemble together followed by the formation of two mechanical bonds. Thus, the formation of 7 further demonstrates the ability of the AT approach to act in multi-component synthesis. With a sufficient quantity (~20 mg) of 7 in hand, we next turned to single crystal X-ray crystallography (SXRC) to confirm the identity of 7, where we found that suitable crystals for diffraction could be grown via vapor diffusion of pentane into a concentrated solution 7 in THF. Accordingly, as revealed by SXRC (Figure 4), 7 is indeed an interlocked structure comprised of two units of macrocycle 7, i.e. a [c2]daisy chain. Unlike typical [c2]daisy chains, 7 is unusually small owing to the macrocycles used in the AT preparation. Additionally, 7 is comprised of all sp 2 -hybridized macrocyclic components, making it an overall very rigid system-a feature which was also reflect-ed in a general lack of solubility in most solvents. Intrigued by the formation of 7, we then investigated the geometrical features observed in the solid-state structure (Figure 4). As can be seen, the macrocyclic fragment is located over the ether-linkage between the diyne and stopper moiety with a distance of 3.34 Å between the nitrogen of the macrocycle and oxygen atom of the thread unit being observed (Figure 4c and 4d). Additionally, various short contacts between the hydrogen atoms of the methyl ester moiety were found, ranging between 2.44 and 2.90 Å from the methyl group to adjacent macrocycle (Figure 4c and 4d). Taken together, these distances highlight the highly crowd-ed environment within the macrocyclic pocket as well confirm a contracted, rather than expanded, solid-state conformation of 7. Upon isolation of 7, we observed strong emission (Figure 5a) with a measured emission maximum centered at 517 nm. It is well known that the parent hydrocarbon cycloparaphenylenes ([n]CPPs) possess bright fluorescent features and red-shifting emissions with decreasing macrocycle size. [4d] As the diameter is reduced, the fluorescence is no longer observed where a CPP comprised of 6 aryl rings (i.e. [6]CPP) is non emissive-a con-sequence of both strain and orbital symmetry. [6a] We recently found that the inclusion of metasubstituted aryl rings breaks orbital symmetry resulting in bright fluorescence from smaller macrocycles [6b] which provides a rationale for the observed emission of 7. This was also observed in our previously reported rotaxanes bearing macrocycles of similar size. [4n] Additionally, we found that the quantum of yield of 7 (ɸ = 0.15) was nearly identical to that of our previously reported [2]rotaxanes, suggesting that the daisy chain architecture does not reduce the emission behavior. Lastly, in the absorption spectrum of 7 (Figure 5a), two main absorptions were observed at approximately 330 and 420 nm which is consistent with our previous report (See SI for additional details on electronic structure). While not investigated here, the emerging interest in nanohoop-based polymers [7a] as well as recent reports of mechanically interlocked fluorescent sensors [7b, c] suggests that interlocking nanohoop fluorophores into daisy chain architectures could serve as excellent starting point for the development of polymeric sensing materials. In the presence of certain stimuli such as metals or protons, [c2]daisy chains can often be switched between either a contracted or expanded muscle-like state. [3] The actuation mechanism is highly dependent on the passive templation method used to prepare the structure. In contrast to most daisy chain structures, 7 has no complimentary interactions between the macrocycle and thread component due to the usage of an AT-CC method. It has been shown that this results in weak intercomponent binding energies which can allow for less common stimuli to actuate mechanical motion. [5c] Encouraged by this observation, we targeted thermal energy as potential source to actuate 7. To investigate this, we first carried out variable temperature (VT) 1 H-NMR on 7 over a temperature range of -60 to 50 o C (Figure 6). We found that the resonance belonging to (See Figure 3 for assignments) HD shifts upfield to nearly 1.3 ppm from 3.0 ppm, indicating stronger shielding at lower temperatures. On the other hand, resonances belonging to HJ and HE are shifted downfield by nearly 1.0 ppm each, demonstrating an overall deshielding at lower temperatures. The effect of increased shielding of HD and decreased shielding of HJ and HE on lowering temperature suggests that as the temperature decreases, the macrocycle of 7 begins to reside pre-dominantly over the proton HD. Taken further, this suggests that at lower temperatures, the conformation of 7 begins to adopt that of a contracted form whereas on heating, the macro-cycle begin to participate in rapid shuttling over protons HJ, HE, and HD. This demonstrates that on the NMR timescale, fast muscle-like expansion and contraction is taking place and the macrocycle position, i.e. daisy chain state, can be controlled via a decrease in temperature. To further confirm the location of the macrocycle at various temperatures, we then probed these dynamics via VT selective 1D ROESY on resonance HB (Figure 3 for assignment). At 25 o C two ROE signals between HJ and HK are observed, suggesting that the ester moiety is close in proximity with both the benzylic proton (HJ) and pyridyl ring of the macrocycle. On heating the same sample to 50 o C, ROE signals between these resonances were still observed, but were significantly attenuated (Figure 6b, right) suggesting a larger distance between each moiety at elevated temperature. This data supports the notion that at lower temperature, the macrocyclic ring resides near the stopper moiety, i.e. a contracted form. It should be noted that temperature changes would be expected to impact the motion of nearly any molecular object, however, a direct consequence of the mechanical linkage in systems such as 7 is that the reduced motion of each bond at lower temperatures ultimately orchestrates a large amplitude change in molecular geometry, i.e. contraction or extension. Taken together, these data illustrate that actuation of 7 can be achieved in an additive-free manner via thermal changes which we attribute directly to the lack of complimentary interactions between macrocycle and thread-a feature inherent the AT-CC method of preparation. In conclusion, our work here provides the first example of an usual type of fluorescent, nanohoop [c2]daisy chain formed without an inherent interaction between each fragment-a structure which appears to be accessible only through at active template approach. While the lack of interaction between each component renders typical stimuliinduced expansion or contraction strategies unamenable, we show that this structure's conformation can be tuned via temperature which may allow for operation under a wider range of environmental conditions. Additionally, based on DFT calculations, we find that the frontier molecular orbitals are located between thread and macrocycle. Specifically, the HOMO is localized over the macrocycle while the LUMO is localized over the stopper moieties. We expect that by increasing the thread length, the distance between these groups can be altered via temperature changes which may allow for modulation of the emission. Additionally, given that a variety of active template reactions as well as macrocycle compositions are now known, more common stimuli-induced strategies such as metal-coordination can likely be leveraged through, for example, as AT copper(I) catalyzed azide-alkyne cycloaddition reaction. This work ultimately highlights a new approach to link small carbon nanostructures with mechanical bonds-a strategy and concept that we expect will provide new momentum in the emerging area of supramolecular carbon nanoscience.</p>

ChemRxiv

Cooperative Effects on Radical Recombination in CYP3A4-Catalyzed Oxidation of the Radical Clock \xce\xb2-Thujone**

The oxidation of hydrocarbons by cytochrome P450 enzymes is commonly thought to involve hydrogen atom abstraction by a ferryl species comparable to that of peroxidase Compound I, followed by radical recombination of the resulting carbon radical with the equivalent of an iron-bound hydroxyl radical.1,2 This radical rebound mechanism, first proposed in 1978, is supported by a variety of experimental results, including (a) rearrangement and inversion reactions prior to the radical recombination step, (b) the large magnitude (up to kH/kD ~ 13) of the intrinsic isotope effect for hydrogen abstraction, and (c) computational modeling of the reaction pathway. However, radical clock substrates, in which the radical undergoes a rearrangement at a known rate prior to radical recombination, have provided conflicting evidence on the radical lifetime. Although several radical clocks support a radical recombination mechanism, so-called ultrafast radical clocks yield radical lifetime estimates more consistent with a transition state than an actual intermediate.3 This discrepancy has led to postulates that hydroxylation may involve concerted insertion into the C-H bond or the involvement of multiple oxidizing species. An alternative explanation is provided by computational studies that invoke a reaction manifold with a radical intermediate that exists in two different spin states.4 A further possible explanation is provided by the observation that ultrafast radical clocks generally involve primary radical rearrangements, whereas slower radical clocks generally involve secondary radical rearrangements. The recombination rates of primary and secondary radicals may be differentially susceptible to modulation by interactions with the active site and the iron-oxo species. However, there is little direct evidence that the radical complex exists in two different spin states, or that the radical recombination rates can be influenced by the active site environment.

cooperative_effects_on_radical_recombination_in_cyp3a4-catalyzed_oxidation_of_the_radical_clock_\xce

<!>Chemicals<!>Enzymes<!>GC-MS Analyses of Thujones and their Metabolites<!>Enzyme Incubations

<p>CYP3A4 is responsible for a majority of all P450-catalyzed drug oxidations in humans. One of the salient features of CYP3A4 catalysis is that it is subject to homotropic and heterotropic cooperativity (allosterism) in that one substrate molecule can alter the oxidation rate and/or regioselectivity of another.5 The molecular basis of P450 allosterism remains incompletely understood. Currently, most results fall into three models: multi-substrate binding, peripheral effector binding, and conformational heterogeneity.6 Progesterone has specifically been shown to function as an allosteric modulator of its own CYP3A4-catalyzed oxidation as well as that of other substrates.7,8 Given that allosterism tunes the CYP active site environment, its effect on the reaction of a radical clock substrate may provide valuable mechanistic information associated with the CYP mechanism. In this communication we report that the oxidation of a radical clock, β-thujone (1, shown as its C4 radical in Scheme 1), by CYP3A4 is subject to a concentration-dependent heterotropic cooperative effect of progesterone binding on the manifold of products that derive from C4 oxidation. The results are consistent with acceleration by progesterone of the radical recombination that traps the C4 radical with concomitant suppression of all the rearrangement pathways.</p><p>We have previously shown that α- and β-thujones function as two-zone radical clocks in that C4 oxidation triggers two concurrent timing reactions, one involving ring opening of the cyclopropyl ring and the other inversion of the stereochemistry of the C4 methyl group (Scheme 1).9 The α- and β-thujone C4-radical ring-opening reaction rates are 4.4 × 107 s−1 and 1.0 × 108 s−1, respectively, both of which are consistent with the lifetimes of radical intermediates. This is also consistent with the observation of a significant extent of methyl inversion in these reactions. In addition to hydroxylated products, this reaction manifold gives rise to two minor desaturated products, one with a C4–C10 double bond and another (6, carvacrol) in which the ring aromatizes to give a phenol product. Formation of these two products was attributed to a second oxidation of the C4 radical intermediate. The latter was previously shown to be generated from a cationic intermediate.10</p><p>In this study the CYP3A4 reaction of α- or β-thujone (100 μM) was carried out in the presence of a concentration of progesterone increasing from 0 to 150 μM. The incubation mixture contains purified, recombinant human CYP3A4, NADPH-cytochrome P450 reductase, and cytochrome b5 in a 1:2:1 ratio, respectively, all in a reconstituted lipid environment.11 The identified metabolites include 7-hydroxyl-α(or β)-thujone, 7,8-dehydro-α-thujone, 2-hydroxyl-α(or β)-thujone, 4-hydroxyl-α-thujone (3), 4-hydroxyl-β-thujone (2), carvacrol (6), and the ring-opened product 5.9 These are all the major metabolites formed in the reaction and account for ~95% of the total conversion yield forα-thujone and ~90% for β-thujone. 4,10-Dehydrothujone was previously identified in the reactions of the thujones with P450cam and P450BM3 but not with mammalian P450 enzymes.9 This metabolite was found to conjugate with glutathione in the CYP3A4 solution and the conjugate could not be extracted with the hydrophobic metabolites (Data not shown). It was reported recently for norcarane that desaturation products and their secondary metabolites might compromise its use to measure radical lifetimes.12,13 This was not observed in the CYP3A4 oxidation of thujones, for which no secondary metabolites of the desaturated products were detected at a significant level (>1%, see Figure S1).</p><p>The distribution and yields of the metabolites in the CYP3A4 oxidation of β-thujone in the presence of various concentrations of progesterone are summarized in Tables 1 and S1. Not unexpectedly, a decrease from 15.4 to 11.7% in the proportion of 7-hydroxylation of β-thujone is observed as the progesterone concentration rises from 0 to 150 μM. A smaller progesterone-dependent decrease is also seen in 2-hydroxylation and desaturation of the isopropyl group. These conventional metabolic shifts reflect repositioning of the substrate prior to the oxidation step and provide no information on the reaction mechanism. However, the alterations in the product distributions that stem from oxidation of C4, the carbon that is pivotal for the radical clocks mechanisms of β-thujone (Scheme 1), are more informative. As shown in Figure 1, the proportion of 4-hydroxy-β-thujone (2), the unrearranged product, increases systematically from 65.6% to 78.3% as the progesterone concentration increases from 0 to 150 μM. In contrast, there is a corresponding systematic decrease in the yields of 4-hydroxy-α-thujone (3), in which the methyl stereochemistry is inverted (from 10.7% to 6%), metabolite 5 in which the cyclopropyl ring is opened (from 0.61% to 0.24%), and carvacrol (6), the aromatic product produced by cyclopropyl ring opening and cation formation (from 3.5% to 1.5%). A direct correlation exists between the decrease in the combined rearrangement products (3 + 5 + 6) and the increase in the unrearranged alcohol (2) (Figure 2). This shows that the loss in the rearranged products is quantitatively correlated with the increase in formation of the unrearranged alcohol (2). While it has been well documented that allosterism alters the reaction rates and regioselectivity of CYP oxidation on the different reaction sites of a substrate, to our knowledge, this is the first report of allosteric modulation of the product distribution profile derived from a single reaction site. This differs from a conventional metabolic shift in that it can only happen after the hydrogen abstraction reaction. As all of these reactions stem from initial C4 oxidation, this result demonstrates that the changes in the product profile depend on progesterone allosteric effects associated with or subsequent to the C4 hydrogen abstraction reaction.</p><p>The oxidation of α-thujone, in contrast to β-thujone, shows a progesterone concentration-dependent increase in 2-hydroxylation compensated by a decrease in 7-hydroxylation, but no systematic changes are observed in the products from C-4 oxidation (Table S2). While the formation of 4-hydroxy-β-thujone (2) and carvacrol slightly decreases with increasing progesterone concentration, the trends in the formation of the other two C4 products are ambiguous.</p><p>The results from the oxidation ofβ-thujone indicate that the alteration of the CYP3A4 active site cavity caused by the simultaneous binding of progesterone decreases the apparent lifetime of radical 1 generated fromβ-thujone by the P450 activated species. The apparent radical recombination rate changes from 1.3 × 1010 s−1 in the absence of progesterone to 3.5 × 1010 s−1 in the presence of 150 μM progesterone (Table 1). These numbers represent the ratio of the C4 products 2, 3, plus 5 over the ring opened product 5 multiplied by the rate of the ring opening reaction.9 The apparent change in the radical lifetime can be rationalized if (a) the binding of progesterone restricts the motion of the β-thujone radical in a manner that favors radical recombination, effectively accelerating the trapping of the unrearranged radical, or (b) reduces the rates of the rearrangements themselves. In either case, the decrease in the formation of all the rearrangement products results in an apparent change in radical lifetime. If the observed radical lifetime reflects the ratio between high spin and low spin Cpd I species, as suggested by one theory,4 this result would indicate that the low spin Cpd I species with lower energy barrier increases in response to progesterone allosterism. However, the observed apparent radical rebound rate change is more to result from allosterically-induced physical changes in the enzyme active site.</p><p>The observation of products derived from cationic intermediates has been used as an indicator of the involvement of alternant oxidizing species in the P450 catalytic cycle, one of which involves a direct insertion of a hydroxyl cation into the C-H bond.3 In the present study the cationic product, carvacrol (6), decreases in a trend very similar to that of the radical rearrangement products 5 and 3 as a result of the progesterone allosteric effect (Figure 1). The coordinated changes we observed among the four C4 products strongly suggests that the cationic product carvacrol shares the same precursor as the other radical products and is a secondary product of the C4 radical. This is consistent with our previous observation that the C4 product distribution profile was not altered by a highly-deuterated medium.9</p><p>In conclusion, we have observed that progesterone exerts heterotropic cooperative effects on the C4 derived oxidation product profile in the CYP3A4-catalyzed oxidation of β-thujone. Our results demonstrate (a) that all the C-4 derived products stem from a common radical precursor, and (b) that the rate of radical recombination to give the unrearranged hydroxylated product is accelerated by the allosteric effector. This is the first demonstration that the apparent timing of a radical clock can be specifically altered by modulation of the active site topology or properties by an allosteric effector.</p><!><p>Synthesis of β-thujone (≥99%) and the metabolite standards have been reported previously.9 L-α-dilauroyl-sn-glycero-3-phosphocholine (DLPC), L-α-diloleoyl-sn-glycero-3-phosphocholine (DOPC), and L-α-dilauroyl-sn-glycero-3-phosphoserine (DLPS) were from Avanti Polar Lipids Inc (Alabaster, AL). These lipids were used to reconstitute the membranes required for CYP3A4 reaction.11</p><!><p>Human CYP3A4, rat cytochrome P450 reductase (CPR), and rat cytochrome b5 were expressed and purified according to published protocols.14–16 Optical spectra were recorded on a CARY UV-visible scanning spectrophotometer (Varian) in 100 mM potassium phosphate buffer (pH 7.4) at 20 °C.</p><!><p>GC-MS instrument and methods used to analyze thujones and their metabolites were previously described.9</p><!><p>The CYP3A4 reconstitution system was prepared according to a published protocol.11 The ratio of CYP3A4, cytochrome P450 reductase, and b5 was 1:2:1. To the reconstituted enzyme solution was added α- or β-thujone (100 μM in CH3CN) and the corresponding amount of progesterone in methanol (5 μL). After incubation at 37 °C for 3 min, the reaction was initiated by addition of NADPH (1 mM). The final reaction volume was 1 mL, containing CYP3A4 (1 μM) and substrate (100 μM). The reaction mixtures were incubated at 37 °C for 1 h and quenched by adding ice-cold ethyl acetate (2 mL). The aqueous phase was saturated with NaCl and further extracted with ethyl acetate (2 mL). Control reactions were carried out without enzymes, and separately, without NADPH. The combined organic extract was dried over anhydrous Na2SO4, filtered, and evaporated to 50–100 μL under a stream of nitrogen at room temperature for GC-MS analysis of a 3.0 μL aliquot.</p>

PubMed Author Manuscript

The Influence of Chemical Modification on Linker Rotational Dynamics\nin Metal Organic Frameworks

The robust synthetic flexibility of metal-organic frameworks (MOFs) offers a promising class of tailorable materials, for which, the ability to tune specific physicochemical properties is highly desired. This is achievable only through a thorough description of the consequences for chemical manipulations both in structure and dynamics. Magic angle spinning solid-state NMR spectroscopy offers many modalities in this pursuit, particularly for dynamic studies. Herein, we employ a separated-local-field NMR approach to show how specific intra-framework chemical modifications to MOF UiO66 heavily modulate the dynamic evolution of the organic ring moiety over several orders of magnitude.

the_influence_of_chemical_modification_on_linker_rotational_dynamics\nin_metal_organic_frameworks

<p>Metal-organic frameworks (MOFs) are constructed from inorganic clusters bridged by organic linkers and possess ideal architectures for gas storage and mixture separation as well as generating recent interest as nano-devices and molecular machines.[1] In comparison to traditional inorganic porous materials, MOFs are far more dynamic/flexible due to the incorporation of organic ligands.[2–4] A key advantage of MOFs is the synthetic ease of chemical modification to the organic moieties without change of framework structure, known as isoreticular synthesis. This allows for many avenues of functional modification including adsorptive selectivity based on favorable ligand-guest interactions[5], changes in pore aperture for separations[6,7], and shifts in optical behavior[8]. An additional feature of MOFs, which has captured much attention, is the prevalence of framework dynamics which can lead to dramatic structural shifts such as breathing, swelling, and subnetwork displacements.[9,10] While deformations in the metal cluster have been implicative of dynamic events, the nature of the linker and its functionality is a major contributor to framework dynamics,[10] and play a significant role in determining macroscopic functions such as gas storage and separation, ferroelectricity, spin crossover and luminescence in a variety of MOFs.[11–17] As such, establishing relationships between local dynamics driven by the organic ligands and their chemical modification is of general interest for materials design in MOFs.[17,18]</p><p>Though the structural picture of MOFs is most informed by analysis with X-ray diffraction, this yields an essentially static picture, failing to capture the dynamic aspects of the framework. However, dynamic measures of crystalline materials are challenging due to the potential of multiple timescales of coordinated motions in the solid-state. Many dynamic studies have employed solid-state NMR spectroscopy (ssNMR) through 2H-lineshape analysis to characterize the rotational motion of simple p-phenylene rings within the MOF framework.[15,19–24] Separated-local-field (SLF) solid-state NMR is an alternative, attractive means of dynamic characterization as it utilizes the heteronuclear dipolar coupling as a proxy for molecular dynamics, does not require isotopic labeling, and can be used to characterize motions over a broad dynamic range with atomic resolution.[25]</p><p>In this communication, a combination of computation and the DIPSHIFT[26,27] (Dipolar chemical SHIFT correlation) SLF methodology is employed to characterize ligand dynamics in MOFs; details on the preparation, characterization and calculations are given in the Supporting Information (Figures S1-S6 and Tables S1-S2). For typical heteronuclear dipole-dipole couplings such as 13C-1H and 15N-1H, the experiment is sensitive to a dynamic range spanning 0.1μs-10ms (see Figure 1).[28] Though a number of MOFs are suitable for this approach, we focus on UiO-66[29] and its derivatives (see Figure 2) due to its wide range of potential applications as well as its high thermal and water stability. We demonstrate that the dynamics of the ligands in the framework are heavily impacted by ligand functionalization; specifically, that the variable dynamic behavior arises from differences in interactions with the local chemical environment and is highly responsive to temperature.</p><p>Terephthalic acid is the most commonly used linker for the construction of MOFs. The phenylene units are symmetric and comprised of four equivalent aromatic 13C-1H bonds making it a suitable candidate for DIPSHIFT measurements. Thus, UiO-66 (Zr6(μ3-O)4(μ3-OH)4(terephthalate)6) (see Figure 2a) was investigated first to understand the dynamic behavior of a simple phenylene unit. Using the 13C-1H bond on the phenylene ring as the local field pair, DIPSHIFT measurements were performed on UiO-66 at 23 °C (Figure 3a) resulting in a heteronuclear dipolar coupling constant (DCH) of 16.4 ±0.2 kHz. The exhibited DIPSHIFT curves feature well-depths that are roughly proportional to the strength of the C-H dipolar coupling and can be fit with standard MAS equations to extract the coupling constants (see Figures 3, and S9-S15).[31,32] The experimentally determined DCH value is smaller than the theoretical value for a rigid 13C-1H bond (~22 kHz), which is due to dynamic averaging resulting from the motional modes, namely ring rotations and librations, available to p-phenylene within the framework.1 Given the timescales detected by this approach, the primary event detected is associated with the ring flip although smaller, more rigorous librational modes are also expected to contribute to a lesser degree. By adopting an appropriate model, a correlation time for the motional ring flip can be estimated, which is found to be approximately a few milliseconds falling in "slow" regime dynamics for UiO-66 at 23 °C (see SI) and is confirmed by higher temperature experiments (vide infra). This result aligns with reports on MIL-53[19] and MOF-5[33] both containing p-phenylene units in a sterically unhindered environment. It is notable that the DIPSHIFT data is symmetric around half the rotor period for UiO-66 at 23 °C, which is a model-free confirmation of slower motional time scales assuming the dynamics are not already in the fast limit. Asymmetries arising from intensity decay due to transverse relaxation effects during the encoding time occur when the molecular motion is accelerated (e.g. temperature induced) to timescales coinciding with the inverse of the coupling constant, known as the 'intermediate motional regime,' as depicted in Figure 1.[28]</p><p>After establishing the effectiveness of DIPSHIFT based characterization in the UiO-66 sample, we next turned our investigation to ligand functionalized UiO-66 MOFs. Methyl and hydroxyl groups are among the most commonly used functionalities and are particularly relevant for applications such as separations and catalysis.[34,35] The bulkiness of the methyl group and the hydrogen bonding potential of the hydroxyl groups make them a diametric pair to study their effects on rotational dynamics. Indeed, experiments on UiO-66-(OH)2 (Zr6(μ3-O)4(μ3-OH)4(2,5-dihydroxyterephthalate)6) and UiO-66-(CH3)2 (Zr6(μ3-O)4(μ3-OH)4(2,5-dimethylterephthalate)6) resulted in dramatically different DIPSHIFT curves from one another (Figure 3). UiO-66-(OH)2 gives a higher DCH value (~19.0 ±0.5 kHz) than the parent UiO-66 (DCH ~16 kHz) indicating reduced mobility; on the other hand, the measured DCH value for UiO-66-(CH3)2 is only 10 kHz. This stark reduction of the C-H dipolar coupling constant for UiO-66-(CH3)2 evidences much faster rotary motion than UiO-66, placing it well into the fast regime (τc < 1 μs). These simple substitutions cause dramatic changes to the ligand framework interaction which induces ~4 orders of magnitude difference in the dynamic rotation. These same trends hold at different temperatures: UiO-66-(OH)2, the least dynamic of the three MOFs, must be heated to ~105 °C (60 °C higher than UiO-66) to achieve rotational times on the order of 100 μs (see SI). In contrast, the 10 kHz value for the methyl substituent is similar to the plateau value observed for UiO-66 in the fast regime (τc < μs) at 155 °C (~ 11 kHz). Indeed, when UiO-66-(CH3)2 is cooled to 6 °C, an asymmetric curve is detected (see Figure S12), suggesting that it is in a comparable dynamic regime to UiO-66 at temperatures ~50 °C lower.</p><p>To complement the experimental results, DFT simulations on a model complex were utilized as a tool to understand the details of rotational dynamics. In the case of UiO-66, a model of Zr6O4(OH)4(COOH)11(COOC6H5) bearing a single phenyl rotor was constructed and its geometry was optimized. A relaxed potential energy surface scan was performed with the phenyl ring rotation through its principle axis. A rotational energy barrier of 43.5 kJ/mol was obtained. Despite the simplicity of the model, the obtained activation energy is similar to the one found for MOF-5[36] and MIL-53[19]. As with UiO-66, Zr-oxo complexes were constructed and calculations were also performed on models with hydroxyl and methyl functional groups. A high activation energy of 75.3 kJ/mol was found for the model Zr6O4(OH)4(COOH)11(COOC6H5-o-OH). This is expected to dramatically slow down the ring dynamics which supports the experimental data. A closer inspection of the model shows that the hydroxyl group hydrogen bonds with the carboxylate group, an interaction that must be broken upon rotation (Figure 4). The hydrogen bonding explains the high energy barrier and suggests that it is the dominant factor leading to the experimentally observed high C-H dipolar coupling constant and slow rotational dynamics in UiO-66-(OH)2. On the other hand, the model for UiO-66-(CH3)2 (Zr6O4(OH)4(COOH)11(COOC6H5-o-CH3)) shows a lower activation energy of 33.5 kJ/mol. This reduced energy barrier is attributed to repulsive interactions between the bulky methyl group and the cluster. The simulated energy barrier results are fully consistent with the experimental observations for rotational frequencies: UiO-66-(o-OH)2<< UiO-66<<UiO-66-(o-CH3)2. The model above, however, does not account for geometric restrictions and electronic effects imposed on the rotors from Zr6 clusters at both ends. To simulate these effects, two spatially fixed clusters held at crystallographic distances and bridged by a linker were constructed; the relaxed potential energy scan was performed with a semi-empirical method (SI Section 2). The energy barrier was then found using DFT single point energy calculations on the highest and lowest energy conformations determined from the relaxed torsional scan. Activation energies with this approach resulted in 65.77 kJ/mol, 23.39 kJ/mol and 91.71 kJ/mol for models of UiO-66, UiO-66-(CH3)2 and UiO-66-(OH)2, respectively. These results are consistent with the results of the simpler model and further support the influence of local sterics and electronic effects in determining linker dynamics.</p><p>UiO-66-Br (Zr6(μ3-O)4(μ3-OH)4(2-bromoterephthalate)6) and UiO-66-NH2 (Zr6(μ3-O)4(μ3-OH)4(2-aminoterephthalate)6) were also synthesized to test the influence of linker dynamics from another bulky substituent other than a methyl group and a potential hydrogen bond donor. The measured DCH for UiO-66-Br is closer to UiO-66 at 15 kHz but exhibits a slight asymmetry in the curve (Figure 3b)2. Despite the higher value than UiO-66-(CH3)2, the temperature behavior seen in Figure S13 of UiO-66-Br also shows that it is in the fast regime limit at room temperature. This suggests that the bromine atom also induces a steric effect which lowers the rotational energy barrier similar to the function brought by methyl groups, but is not as extreme given the higher fast limit value. UiO-66-NH2 shows a DCH value of 15.3 ±0.3 kHz, value close to UiO-66. However, the temperature behavior is quite constant in a similar temperature range (see SI Figure S14). This behavior could be interpreted as a combination of effects: a balance between a steric effect and the hydrogen bonding potential of the NH2 group.</p><p>Changing the length of the linker while maintaining the topology, known as isoreticular synthesis, is an important strategy that has been used to increase the surface area of MOFs without altering the topology.[37] Thus UiO-67 (Zr6(μ3-O)4(μ3-OH)4(biphenyl-4,4'-dicarboxylate)6) was examined for its larger pore window (~9–11 Å) compared to the parent structure UiO-66 (~5–7 Å). The measured DCH for UiO-67 is 12.4 ±0.3 kHz, which is smaller than that obtained from UiO-66. The lowering of the rotational energy barrier is attributed to the junction of the biphenyl ring, which has a much lower frictional constraint vs the Zr cluster interaction. At higher temperature, the depth of the curve decreases substantially, but no intermediate dynamics are detected (Figure S15). This is possible in the fast limit regime when the reorientational angle of motion increases; it is plausible that the increased length and biphenyl junction could contribute additional bending modes to facilitate such behavior.</p><p>In summary, DIPSHIFT solid-state NMR experiments performed on a series of Zr MOFs with various structural modifications showed different ligand dynamics. The presence of functional groups such as methyl or hydroxyl groups can greatly speed or slow down the rotational motion of the linker in the Zr MOFs. Insights from DFT simulation reveals that such dynamics changes arise from the local chemical interactions such as hydrogen bonding or steric repulsion. It is also found that isoreticular structure expansion and factors such as temperature also influence the ligand dynamics. The insights gained are relevant to applications, where dynamics of the ligands are heavily involved, such as gas separations and kinetically limited processes such as molecular sieving.</p>

PubMed Author Manuscript

Physiological responses to known intake of ergot alkaloids by steers at environmental temperatures within or greater than their thermoneutral zone

Two studies separated effects of dietary ergot alkaloids from effects of feed intake or ambient temperature on respiration rate (RR), heart rate (HR), surface temperature (ST), rectal temperature (RT), blood pressure (BP), serum hormone, and plasma metabolite concentrations in beef steers. The balanced, single reversal design for each experiment used 8 beef steers fed tall fescue seed [2.5 g/kg body weight (BW)] with (E+) or without (E−) ergot alkaloids as part of a 60:40 switchgrass hay: supplement diet. Periods were 35 days with 21 days of preliminary phase and 14 days of feeding fescue seed once daily. Measures of dependent variables were collected on d 20, 25, 29, and 35 of each period at 0730 (before feeding), 1230 and 1530. In Experiment 1 steers weighed 286 kg, gained 0.61 kg BW/day, E+ supplied 2.72 mg ergot alkaloids including 1.60 mg ergovaline per steer daily, and mean minimum and maximum daily ambient temperatures were 23.6 and 32.3°C. In Experiment 2 steers weighed 348 kg, gained 1.03 kg BW/day, E+ supplied 3.06 mg ergot alkaloids including 2.00 mg ergovaline daily, and mean minimum and maximum daily ambient temperatures were 11.9 and 17.4°C. Dry matter intake was not affected by fescue seed treatment (P < 0.20) in either experiment. In both experiments, E+ reduced HR (P < 0.01) and increased insulin (P = 0.07). Systolic BP minus diastolic BP decreased (P < 0.05) for E+ in both experiments, due to increased diastolic BP in Experiment 1 (P < 0.03) and decreased systolic BP in Experiment 2 (P < 0.07). In Experiment 1, above the thermoneutral zone, E+ increased (P < 0.05) RR, RT, and left side ST in comparison to E−, but in Experiment 2, within the thermoneutral zone, E+ and E− did not differ (P < 0.18). Ergot alkaloids from fescue seed affect the cardiovascular system of steers separately from effects of feed intake or environmental temperature. Ergot alkaloids interact with ambient temperatures above the steers' thermoneutral zone to exacerbate the symptoms of hyperthermic stress.

physiological_responses_to_known_intake_of_ergot_alkaloids_by_steers_at_environmental_temperatures_w

Introduction<!>Materials and methods<!><!>Materials and methods<!><!>Materials and methods<!>Sample analysis<!>Statistical analysis of data<!>Results<!><!>Baseline data<!>Responses to days of feeding fescue seed<!><!>Responses to days of feeding fescue seed<!><!>Time of day and treatment responses<!><!>Discussion<!>Conflict of interest statement

<p>Consumption of toxic, endophyte-infected, tall fescue results in ingestion and absorption of ergot alkaloids produced by the endophyte, Neotyphodium coenophialum, which causes fescue toxicosis in grazing cattle. Ergovaline, the alkaloid produced in greatest concentration, or total ergot alkaloids have been measured in fescue to describe its potential toxicity. Concentrations of ergovaline increased from 250 to 450–500 μg/kg in leaf blades and from 500 to 800–1300 μg/kg in leaf sheaths from April to May. Seed heads contained the greatest concentration of toxins and reached concentrations as high as 5000 μg/kg in June (Rottinghaus et al., 1991). Total ergot alkaloid concentration showed the same seasonal changes as ergovaline (Hill et al., 2000).</p><p>Studies with steers or heifers consuming different sources of fescue hay (Hemken et al., 1981) or consuming alkaloids from fescue seed (Burke et al., 2001) at differing ambient temperatures in a factorial experimental design indicate a greater response to diets with endophyte-infected vs. endophyte-free fescue (increased respiration and rectal temperatures (RTs) and decreased voluntary intake) during hyperthermic heat stress compared to ambient temperatures within the animals' thermoneutral zone. Adverse responses of cattle consuming endophyte-infected tall fescue, including lower tolerance to ambient temperatures outside the animals' thermoneutral zone, decreased voluntary intake, weight gain, and milk production have been linked to hemodynamic effects of ergot alkaloids in the tall fescue (Strickland et al., 2011). The alkaloids, particularly ergovaline, ergovalanine, and ergonovine administered i.v. (Oliver et al., 1994; Browning and Leite-Browning, 1997; Browning, 2000) or fed at doses comparable to amounts of alkaloids ingested as endophyte-infected fescue seed (Rhodes et al., 1991; Aiken et al., 2007, 2009), decreased heart rate (HR), increased blood pressure (BP), and caused vasoconstriction in steers or heifers. The responses include decreased skin temperature or increased RT and increased respiration rate (RR) at ambient temperatures greater than the animals' thermoneutral zone. Usually, ingestion or administration of alkaloids decreased serum or plasma prolactin concentrations. In a thermoneutral environment, the hemodynamic responses appeared to be muted or not detectable.</p><p>Efforts to delineate potential interactions between hemodynamic effects of alkaloid and ambient temperatures above the animals' thermoneutral zone have been complicated by concomitant changes in voluntary intake when animals consume endophyte-infected tall fescue (Hemken et al., 1981; Boling et al., 1989; Aiken et al., 2007, 2009). Additionally, close human contact required to obtain physiological measures may itself alter the animals' response and contribute to variation in response to dose levels or duration of experimental protocol (Aiken et al., 2007); therefore, most reported studies describe acute responses over a period of hours or a few days. Some researchers have addressed this potential problem by adapting animals to facilities and conditions prior to experiments, e.g., Browning (2000).</p><p>The main objectives of the current experiments were to separate the pharmacological effects of endophyte alkaloids from effects of ambient temperatures above the animals' thermoneutral zone and effects attributable to changes in intake or discomfort due to close human contact.</p><!><p>Two experiments were conducted under the supervision and approval of the university animal care and use committee. Angus steers from the North Carolina State University Beef Education Unit university farm of known pedigree, age, and background were trained to be led by halter and accustomed to close human contact while eating a hay diet devoid of fescue. Experiment 1 was in June to August, 2011, and Experiment 2 was in October to December, 2012. Each experiment used 8 steers. In Experiment 1 mean ± SD steers' age was 247 ± 24 d and in Experiment 2 it was 380 ± 27 d. Steers were housed in individual stalls with a chain attached to their halter. The stalls were 115 × 178 cm, with automatic waterers and rubber mats on concrete floors. The daily protocol was removal of orts (if any), feed supplement at 0830 h, move steers outside for exercise in a common pen for about 1 h, then back to stalls for morning hay feeding by about 1000 h. At 1530 h, the second ration of supplement was fed, followed by the second ration of hay. Lights in the barn were 18 h on: 6 h off each day, with adjustment of on and off times to accommodate the season of the year. Steers were weighed weekly, feed and orts recorded daily. Steers' stall assignment was determined when they were randomly allocated to the treatment protocol.</p><p>All steers were fed sliced switchgrass hay (Table 1) at daily amounts equal to 10 g/kg BW and a supplement, each divided into AM and PM feedings. The hay was stored in rectangular bales, and was pressed through a Van Dale Bale Processor, Model S600 (J-star industries, Ft. Atkinson WI) with knives spaced 12.5 cm apart.</p><!><p>Organic matter (OM), crude protein (CP), neutral detergent fiber (NDF) and acid detergent fiber (ADF) concentrations, g/kg dry matter, in feedstuffs for the 2 experiments.</p><!><p>Each experiment had two, 35-d periods with 21 d of preliminary phase and 14 d of feeding endophyte-infected fescue seed (E+) or endophyte-free fescue seed (E−). During the preliminary phase soybean hulls were added to the supplement instead of fescue seed. During the treatment phase of Experiment 1 steers were fed 2.15 kg supplement DM and 0.62 kg of fescue seed DM daily (Table 1), and in Experiment 2 steers were fed 2.62 kg supplement and 0.69 kg fescue seed DM daily (Table 1), with the total weight of supplement plus fescue divided equally in the AM and PM feedings. In Experiment 1, E+ supplied 2.72 mg ergot alkaloids including 1.60 mg ergovaline per steer daily and in Experiment 2, E+ supplied 3.06 mg ergot alkaloids including 2.00 mg ergovaline daily. The dose chosen was similar to that of Aldrich et al. (1993) who fed diets containing 285 μg/kg of ergovaline from fescue seed. The goal was to produce physiological changes but maintain similar intake during E− and E+ feeding to avoid confounding effects due to intake changes.</p><p>For Experiment 1 the ration was formulated to meet National Research Council (1996) nutrient requirements for CP, TDN, Ca, and P for a steer weighing 272 kg and gaining 0.6 kg/d and for Experiment 2 the ration was formulated to meet those requirements for a steer weighing 318 kg and gaining 0.6 kg/d. The fescue seed passed through a 1.1 cm screen in a hammer mill (Meadow Mills, North Wilksboro, NC) before feeding to partially disrupt the seed coat. All of the fescue seed was fed in the morning, so the amount of AM supplement fed was reduced accordingly. Steers were assigned at random to receive E− (Southern States Cooperative, Inc., Cloverdale, VA) or E+ (EverGreen Seed, LLC, Fuquay-Varina, NC) seed in a single reversal design, 4 steers fed each type of seed in period 1. The treatments were reversed in period 2. The amount of fescue seed fed was gradually increased and the amount of soybean hulls was gradually decreased during the first 3 d when seed was fed in each period (days 22–24), so d 25 was the first day on full treatment.</p><p>After 3 weeks' adaptation to facilities and protocol, steers were fed their assigned diets for 14 d then all steers returned to the adaptation diet. After 21 d, the 14-d treatment period was repeated. As in the first period, the amount of fescue seed fed was gradually increased during the first 3 d when seed was fed, so d 25 was the first day on full treatment. Steers' hair was clipped during adaptation, 5 to 7 d before the first day of feeding fescue seed with electric animal clippers that left 3 mm hair (Aesculap® Econom II, Suhl, Germany). Daily minimum, maximum, and 1200 h barn temperature and relative humidity were measured using a calibrated humidity/thermometer (Fisher Scientific, Pittsburgh, PA). These variables were measured at 3 locations in the barn and an average value was calculated. Additionally, temperature and humidity at sampling times were recorded.</p><p>On d 20, 25, 29, and 35 of each period palmar surface temperature (ST) in the area of the large metacarpal bone of the front legs, plantar ST in the area of the large metatarsal bone of the rear legs, and left side ST were measured by digital infrared thermal imaging. The legs were chosen because of accessibility, cleanliness, and potential relationship to ST changes in response to peripheral constriction. The left side was chosen because of the layout of the animal handling facilities and the ability to consistently measure the same area (Huntington et al., 2012). The digital thermal images were recorded with a Ti45FT (Fluke Corporation, Everett WA) with a 20 mm lens, manual focusing, 30 Hz 160 × 120 pixel focal array, and a vanadium oxide uncooled microbolometer. Emissivity was set at 0.95 and background temperature was set at 20°C. The camera has an accuracy of 2°C in the physiological temperature range, and a sensitivity of <0.1°C. Images were stored in files containing approximately 2.5 Mbyte of information that included date, time, image number, emissivity, background temperature, and visual and infrared images. Software provided by the manufacturer allowed detailed isolation of portions of the image (see Figure 1) and presented minimal, maximal, average, and the standard deviation of the pixels in the selected portion. BP and HR were measured with a 16 to 24 cm BP cuff around the tail head connected to a digital monitor (Lifesource® A&D Engineering, Inc., San Jose, CA), RT was measured with a digital thermometer (Becton, Dickenson and Co., Franklin Lakes, NJ), and RR by rib cage movement was measured for 10 s at 0730 (before collection of orts), 1230, and 1500 (before PM feeding). Data were averaged for front legs and averaged for rear legs before statistical analysis. The steers were restrained in a squeeze chute while blood was removed by jugular venipuncture, starting at 1300 h on sampling days.</p><!><p>Digital infrared images of the left side (A) and the plantar surface in the area of the large metatarsal bone in the rear legs (B) of a steer during the experiment. The human fingers (A) indicate the edge of the scapula. The paralumbar fossa is to the right of the measured area. For both panels, the geometric shapes indicate areas in which temperature data were recorded. The vertical bar on the far right side of each panel indicates the temperature range (°C) and the associated light (warmer) to dark (cooler) gradient within the image.</p><!><p>In each experiment, 2 groups of 4 steers (n = 8 total) were staggered by 1 week to allow collection of physiological data within 30 min on each sampling day. Steers were adapted to the procedures by several practice sessions during the preliminary phases of the experiments. Two people were in the barn to collect BP and HR measures, one steer at a time in their stall, steer in standing position. One minute elapsed between deflation of the cuff and the subsequent measure. The instrumental criteria and personal experience were used to assess validity of each BP and HR measure. The steers' demeanor on a given day and time affected the number of BP and HR measures used in statistical evaluation of treatments. Of the 384 measurement episodes within steer, period, day, and time, 6 contained 2 measures, 264 contained 3 measures, 108 contained 4 measures, and 6 contained 5 measures.</p><p>In Experiment 1, a fecal grab sample was collected between 1100 and 1300 h for 3 d at the end of each fescue period for analysis of alkanes to calculate DM digestibility.</p><!><p>Feed samples were analyzed for nutrient content by a commercial laboratory (North Carolina Department of Agriculture, Raleigh, NC). Concentration of total alkaloids in the fescue seed was analyzed by a commercial laboratory (Agrinostics Ltd. Co., Watkinsville, GA) using an ELISA (Hill and Agee, 1994). Ergovaline concentration in the fescue seed was analyzed by a commercial laboratory (University of Missouri Veterinary Medical Diagnostic Laboratory, Colombia, MO) using HPLC (Rottinghaus et al., 1991, 1993). Serum prolactin (Bernard et al., 1993) and serum insulin (Cartiff et al., 2013) were determined by radioimmunoassay. For prolactin, the intra-assay CV was 6.8% and the inter-assay CV was 7.1%. For insulin, the intra-assay CV was 5.5% and the inter-assay CV was 8.4%. Plasma glucose was analyzed using a glucose oxidase method (Yellow Springs Instruments, Yellow Springs, OH). Concentration of hentriacontane in feed and fecal samples and calculation of DM digestibility in Experiment 1 were determined as described by Chavez et al. (2011).</p><!><p>The Mixed procedure of SAS (SAS institute, Cary NC) was used for statistical analysis of data. The model had main effects of treatment, day, time of day, group, period, and all possible interactions of treatment, time and day. Steer, steer × period, steer × treatment, and steer × time were random effects. Except for measures in Experiment 1 of serum insulin, serum prolactin, and plasma glucose, mean values within steers across times collected on day 20 of each period were used as covariates within periods. The effect of period on baseline values was tested using a model with period, treatment, and period × treatment as the main effect. Data for intake, DM digestibility, plasma glucose, and blood hormones did not have day or time of day in the model.</p><!><p>Feed intake was not affected by treatment in either experiment (Table 2) and therefore treatment responses for E+ compared to E− are independent of intake effects. Steers were fed at a slightly restricted intake in both studies to minimize orts and remove confounding effects of intake from the responses. Steers in Experiment 1 had greater orts as a proportion of hay offered and ate slightly less DM as a proportion of BW than steers in Experiment 2. Dry matter digestibility did not differ (P = 0.76) for E− and E+ and was 0.584 and 0.590 g/g DM (SE 0.014), respectively. Barn temperature, humidity, and calculated temperature-humidity index (THI, Mader et al., 2002) indicate that steers in Experiment 1 were near or above their thermoneutral zone, and steers in Experiment 2 were within their thermoneutral zone during the experiments (Table 3). Daily minimums and maximums were within or close to those recorded during sampling times in Experiment 1 (Figure 2) and daily minimums were close to those recorded during sampling times in Experiment 2 (Figure 3). Maximum values were after sampling times in Experiment 2.</p><!><p>Body weight (BW), dry matter intake (DMI), and orts for steers fed endophyte-free (E−) or endophyte-infected (E+) fescue seed above (Experiment 1) or within (Experiment 2) their thermoneutral zone.</p><p>n = 8.</p><p>Barn temperature, relative humidity, and temperature-humidity index (THI) for steers fed endophyte-free (E−) or endophyte-infected (E+) fescue seed above (Experiment 1) or within (Experiment 2) their thermoneutral zone.</p><p>Data from the last 16 d of each period.</p><p>d 20, 25, 29, and 35 of each period.</p><p>THI = 0.8Ta + [(0.01 RH × (Ta−14.3)] + 46.3 where Ta = ambient temperature, °C (Mader et al., 2002).</p><p>Daily minimum (dashed line) and maximum (solid line) temperature (°C) during days 14 to 35 of period 1 (A) and period 2 (B) in Experiment 1. Day 21 was the first day of seed feeding in each period.</p><p>Daily minimum (dashed line) and maximum (solid line) temperature (°C) during days 14–35 of period 1 (A) and period 2 (B) in Experiment 2. Day 21 was the first day of seed feeding in each period.</p><!><p>In both experiments, there were few differences (P < 0.21) between baseline values collected on d 20 of each period that were used as covariates in the statistical model. In Experiment 1 there were differences (P < 0.05) for left side ST standard deviation (0.38 vs. 0.31°C) and trends (0.05 < P < 0.07) for systolic pressure (117 vs. 103 mm Hg) and diastolic pressure (52 vs. 47 mm Hg). In Experiment 2 there were differences (P < 0.05) for RR (31 vs. 24 breaths/min) and RT (38.2 vs. 38.5°C). In no case was there a period × treatment interaction supporting the fact that there was no carryover between periods for the variables measured.</p><!><p>In Experiment 1 there was a trend (P < 0.10) for increased left side ST, but in Experiment 2 STs decreased (P < 0.05) with days of feeding fescue seed (Table 4). There were trends (P < 0.10) for day × treatment interactions; diastolic BP increased in Experiment 1 for E+ but did not change for E− with days of feeding fescue seed, and there was a decrease in systolic BP in Experiment 2 for E− whereas systolic BP increased for E+ with days of feeding fescue seed. The difference between systolic and diastolic BP decreased for E− but increased for E+ with days of feeding fescue seed in Experiment 2 (Table 4). The day × treatment interaction (P < 0.05) for RR in Experiment 1 was caused by a greater increase for E+ than E− with days of feeding fescue seed (Table 4). The trend (P < 0.10) for the same interaction in Experiment 2 (Table 4) was caused by a lesser decrease in RR for E+ than E− with days of feeding fescue seed. The ST of rear legs tended (P < 0.10) to fluctuate more for E+ than E− with days of feeding fescue seed in Experiment 2, but there was no interaction in Experiment 1 (Table 4).</p><!><p>Blood pressure, heart rate (HR), respiration rate (RR), rectal temperature (RT), and surface temperature for steers fed endophyte-free (E−) or endophyte-infected (E+) fescue seed for 14 d above (Experiment 1) or within (Experiment 2) their thermoneutral zone.</p><p>Periods lasted 35 days; fescue seed was fed for 14 days (d22–d35) of each period; n = 8 per treatment.</p><p>Trt, treatment; S–D, systolic–diastolic blood pressure.</p><p>Surface temperature of rear legs represents the plantar surface in the area of the large metatarsal bone; surface temperature of front legs represents the palmar surface in the area of the large metacarpal bone.</p><!><p>The day × treatment interaction of ST difference between front and rear legs (P < 0.01) in Experiment 2 reflected similar decreased surface ST of front legs for E− and E+ and of rear legs for E− while ST of rear legs for E+ showed greater fluctuation with days of feeding fescue seed (Table 4).</p><p>Serum insulin concentration was greater (P < 0.05, Experiment 1) or tended (P = 0.07, Experiment 2) to be greater for E+ than E− and plasma glucose concentration was greater (P < 0.01) for E+ than E− in Experiment 2 (Table 5). Serum prolactin concentrations were numerically lower for E+ than E− in both experiments, but variation among steers precluded statistical significance. The average prolactin concentration for E+ was 26% of E− in Experiment 1 and 60% of E− in Experiment 2.</p><!><p>Serum insulin, serum prolactin, and plasma glucose concentrations for steers fed endophyte-free (E−) or endophyte-infected (E+) fescue seed above (Experiment 1) or within (Experiment 2) their thermoneutral zone.</p><p>Periods lasted 35 days; fescue seed was fed for 14 days (d22–d35) of each period.</p><p>Trt, treatment.</p><p>Day 20 values were used as a covariate in Experiment 2.</p><!><p>Systolic BP decreased (P < 0.05, Experiment 1) with time of day (Table 6) and tended to decrease (P < 0.10, Experiment 2) with E+. Diastolic BP decreased (P < 0.05) with time of day in Experiment 2, and was greater for E+ than E− in Experiment 1. In both experiments, systolic – diastolic BP difference was lesser for E+ than E−, with a trend for greater decrease with time of day for E+ than E− in Experiment 1 (time of day × treatment interaction, P < 0.10). HR increased with time of day and was lesser for E+ than E− in both experiments (Table 6). RR increased (P < 0.01) with time of day in both experiments. It was greater (P < 0.01) for E+ than for E− in Experiment 1 with a greater increase for E+ than E− with time of day in Experiment 1 (time of day × treatment interaction, P < 0.05). In both experiments, RT and STs of legs and the left side increased (P < 0.05) with time of day, but there were no time of day × treatment interactions (Table 6). In Experiment 1, RT and left side ST were greater (P < 0.05) for E+ than E−.</p><!><p>Blood pressure, heart rate (HR), respiration rate (RR), rectal temperature (RT), and surface temperature at different times of day for steers fed endophyte-free (E−) or endophyte-infected (E+) fescue seed above (Experiment 1) or within (Experiment 2) their thermoneutral zone.</p><p>Trt, treatment; Time, time of day; T × T, treatment × time of day; S–D, systolic–diastolic blood pressure.</p><p>Surface temperature of rear legs represents the plantar surface in the area of the large metatarsal bone; surface temperature of front legs represents the palmar surface in the area of the large metacarpal bone.</p><!><p>Scharf et al. (2011) reported the critical point for ambient temperature to increase core body temperature of growing cattle in feedlots at about 25°C. Hahn (1999) suggested a thermal stress threshold of 25°C for growing cattle fed ad libitum which coincided with decreased feed intake and 21°C as the threshold for increased RR. Scharf et al. (2011) observed also that cattle showed nighttime recovery to decrease core body temperature. In both of the current experiments RT, RR, and STs were at their minimum at the 0730 sampling for both E− and E+. This sampling time had ambient temperature and THI that were close to the minimum values. RT, RR, and ST cycled in all steers and an increase in RR, RT, and left side ST due to E+ was observed only when steers were housed above their thermoneutral zone. The increases due to E+ were associated with THI of around 74–80, above 75 which is the point suggested for using thermal stress-limiting measures (Hahn, 1999).</p><p>Routes of heat flow from the animal to the environment are conduction, convection, and radiation which depend on thermal gradients within the animal and between the animal and the environment; and evaporation which depends on humidity. Skin temperature below 35°C provided a large enough temperature gradient between the body core and the skin to use all 4 routes of heat exchange (Collier et al., 2006). Mechanisms for heat dissipation in response to thermal stress include increased RR, increased peripheral vasodilation, increased skin temperature, and increased sweat rate. Blood flow to the periphery increases to increase heat loss via conduction and convection. Hair coat can reduce heat flow via these two routes. Heat stress increases sweating rate and RR. Evaporation is the major route of heat loss as ambient temperature approaches skin temperature (Hansen, 2004; Scharf et al., 2010). In Experiment 1, steers were housed above their thermoneutral zone, and left side ST was above 35°C which may have reduced the effectiveness of heat transfer from the body core to the skin and resulted in increased RR.</p><p>Scharf et al. (2010) observed increased RR, RT, skin temperature, and sweat rate in Angus or Romosinuano steers housed at temperatures above the thermoneutral zone, cycling from 26°C during the night to 36°C during the day, compared to thermoneutral housing (21°C). Skin temperatures were highly correlated with ambient temperatures. Decreased sweat rate was correlated with increased RT during heat stress for Angus cattle (Scharf et al., 2010). STs in both of the current experiments were consistent with a correlation between skin temperature and ambient temperature.</p><p>Steers receiving E+ had greater left side ST than those receiving E− when housed above the thermoneutral zone suggesting vasodilation and greater transfer of core body heat to the periphery which should increase heat loss by conduction and convection. The maximum ambient temperature was 32.3°C. Previous studies in which heat stress was constant showed no change in ST due to short-term feeding of E+ (Rhodes et al., 1991; Al-Haidary et al., 2001; Koontz et al., 2012) or decreased ST due to a single injection of ergot alkaloids (Browning and Leite-Browning, 1997; Browning, 2000). Changes in ST reflect changes in ambient temperatures, hair coat, and peripheral blood flow. We clipped hair in an effort to minimize hair coat effect among steers or in response to treatment. The thermal imaging camera provides data on minimum, maximum, average and standard deviation of ST. Previous work with the same camera showed an inverse correlation between average and the standard deviation of side ST in bulls not exposed to toxic fescue (Huntington et al., 2012), indicating that thermal imaging may detect variation in ST due to thermal patterns created by vasodilation. In the current experiments correlations between mean and standard deviations of ST within experiments (data not shown) were not statistically significant (P < 0.10). Steer's hair in our study was not clipped as close to the skin as it was for the Angus bulls in Huntington et al. (2012). ST of front legs was consistently greater than ST of rear legs although the difference in temperature between them declined with time of day. Lack of interactions between days of feeding fescue or time of day with treatment indicates that either front or rear legs could be used to evaluate changes in ST.</p><p>RR increased to a greater rate in response to E+ in conditions above the thermoneutral zone in Experiment 1. It is possible that other avenues of heat dissipation were not responding to environmental conditions. Increased RR, sweat rate, and peripheral vasodilation contribute to internal body temperature response to heat stress (Scharf et al., 2010). Decreased skin vaporization was observed in steers housed at 32°C and fed a similar dose of E+ to that fed in the present study compared to steers fed E− (Aldrich et al., 1993). RT was greater for E+ than E− in the current study.</p><p>HR decreased due to E+ in cattle housed at both ambient temperatures. Decreased HR should result in decreased BP if other variables that affect pressure are unchanged (Melbin and Detweiler, 1993). Systolic-diastolic pressure difference, pulse pressure, decreased for steers eating E+ in both environments but for different reasons. Under thermoneutral conditions in Experiment 2, there was a trend for decreased systolic BP whereas under conditions above the steers' thermoneutral zone in Experiment 1 diastolic BP increased. Increased diastolic pressure may reflect increased peripheral resistance (Ganong, 1975) due to the known effects of ergot alkaloids from fescue to promote vasoconstriction in some vascular tissues (Oliver et al., 1998; Aiken et al., 2007, 2009; Klotz et al., 2007). These effects may be more pronounced due to regulatory changes in response to thermal stress which promote increased blood flow to the skin than under conditions of basal skin blood flow. At greater doses of alkaloids than that used in the current experiments, vasodilation in response to thermal stress may be more limited due to alkaloid-induced vasoconstriction.</p><p>Effects of ergot alkaloids in the fescue seed on decreasing HR, increasing diastolic BP, and RR are consistent with other reports in cattle (Browning and Leite-Browning, 1997; Browning, 2000; Koontz et al., 2012) and sheep (McLeay et al., 2002). However, Rhodes et al. (1991) found no effects of consuming 1.14 mg ergovaline/d on HR, BP or skin temperature of small (88 kg BW) Holstein steers housed at 32°C and restricted to intake equal to 25 g/kg BW. Aiken et al. (2007) observed decreased systolic BP, diastolic BP, and HR in response to ergot alkaloids in beef heifers (375 kg BW), housed below their upper critical temperature and consuming 7.65 mg ergovaline/d. Intake of the heifers tended (P < 0.15) to be lesser for those consuming alkaloids (9 kg DM/d) than control heifers (10.7 kg/d). Baseline values for the heifers (measured with a pressure cuff on the tail head) for systolic BP (143 mm Hg), diastolic BP (77 to 86 mm Hg), and HR (106 beats/min) were greater than those measures in steers in the current experiments before (data not shown) or during feeding E− or E+. Aiken et al. (2007) did not describe details of location of heifers during measures, or adaptation to BP procedures. The use of a pressure cuff on the tail head (Browning and Leite-Browning, 1997) in animals accustomed to its use and measured in their usual pens provides credible values for HR and BP similar to those reported for cattle (Rhodes et al., 1991) and sheep (McLeay et al., 2002) with indwelling pressure monitors. Koontz et al. (2012) observed a trend toward short-term increases in diastolic BP in response to toxic fescue seed extract administered to the rumen. The steers in Koontz et al. (2012) were evaluated within and above their upper critical temperature, but feed intake decreased in response to intraruminal dosing of extract from E+ and increased ambient temperature from 22 to 32°C. Our results demonstrate cardiovascular effects from E+ independent of potential effects of DMI at ambient temperatures within or above the animals' thermoneutral zone. Increased serum insulin concentrations in both experiments, and either no change (Experiment 1) or increased plasma glucose concentrations (Experiment 2) in response to E+ indicate changes in homeorhetic control of glucose metabolism which could be linked to insulin resistance and subsequent effects on glucose metabolism when the steers were fed E+.</p><p>Ergot alkaloids from fescue seed affect the cardiovascular system of steers separately from effects of feed intake or environmental temperature. Ergot alkaloids interact with ambient temperatures above the steers' thermoneutral zone to exacerbate the symptoms of hyperthermic stress. Based on the data from these studies, there may be insulin-dependent glucose metabolism changes in response to ergot alkaloids.</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