Datasets:

griffin

/

ChemSum

like 11

Prediction of MAYV peptide antigens for immunodiagnostic tests by immunoinformatics and molecular dynamics simulations

the Mayaro virus is endemic to South America, and the possible involvement of Aedes spp. mosquitoes in its transmission is a risk factor for outbreaks of greater proportions. the virus causes a potentially disabling illness known as Mayaro fever, which is similar to that caused by the chikungunya virus. the cocirculation of both viruses, with their clinical and structural similarities, and the absence of prophylactic and therapeutic measures highlight the need for studies that seek to understand the Mayaro virus. Using approaches in silico, we identified an antigenic and specific epitope (p_MAYV4) in domain A of the E2 glycoprotein of the Mayaro virus. This epitope was theoretically predicted to be stable and exposed on the surface of the protein, where it showed key properties that enable its interaction with neutralizing antibodies. these characteristics make it an interesting target for the development of immunodiagnostic platforms. Molecular dynamics simulation-based structural analysis showed that the PHE95 residue in the E1 fusion loop region is conserved among Alphavirus family members. PHE95 interacts with the hydrophobic residues of the E2 glycoprotein to form a cageshaped structure that is critical to assemble and stabilize the E1/E2 heterodimer. These results provide important insights useful for the advancement of diagnostic platforms and the study of therapeutic alternatives.

prediction_of_mayv_peptide_antigens_for_immunodiagnostic_tests_by_immunoinformatics_and_molecular_dy

<!>Results<!>Molecular dynamics analysis of the MAYV E1/E2 glycoprotein dimer. MD simulations of 150<!>Discussion<!>Sequences of the E1 and E2 glycoproteins of MAYV and CHIKV. The methodological flowchart is

<p>The Mayaro virus (MAYV) is a neglected arbovirus belonging to the Togaviridae family and Alphavirus genus. It has a wildlife cycle that involves transmission between Haemagogus spp. mosquitoes and animal reservoirs 1,2 . MAYV is endemic to South America and has been reported in Central America 3,4 . Imported human cases have been detected in European and North American countries 2 . Climate and environmental changes may have contributed to its silent dispersion throughout Brazil and worldwide 1,[4][5][6][7] . Detection of MAYV infections in dengue, Zika and chikungunya outbreaks, together with possible involvement of Aedes spp. in MAYV transmission, observed under laboratorial conditions, warns of the risk of outbreaks in naive populations 2,3,[8][9][10] .</p><p>MAYV causes an acute and nonspecific febrile illness characterized by short viremia that can be accompanied by prodromal symptoms such as fever, headache, retro-orbital pain, vomiting, diarrhea, maculopapular rash, myalgia and arthralgia 2,11,12 . These symptoms are similar to those of other important arboviral diseases, such as chikungunya, dengue, Mayaro and Zika, suggesting a new term for this arboviral infection: "the ChikDenMaZika syndrome" 2 . More than 50% of patients develop debilitating and persistent joint pain during the chronic phase of the disease, similar to that caused by CHIKV infection 2 . Thus, developing sufficiently accurate diagnostic tests to distinguish infections caused by MAYV would be an important advance in regions where the arboviruses cocirculate.</p><p>The MAYV genome is composed of a positive-strand RNA approximately 11.5 kb in length and two open reading frames (ORFs). The first ORF is located in the genome 5′-end and encodes nonstructural viral proteins (nsP1, nsP2, nsP3, and nsP4) involved in viral replication and pathogenesis. The second ORF, positioned in the genome 3′-end, synthesizes the structural proteins of Capsid (C), envelope glycoproteins 1, 2 and 3 (E1/E2/E3) and a small 6 K protein, which are important for infection and protection of viral genetic material 13 . Structurally, the E1 and E2 glycoproteins have three domains interconnected by β-connectors. The E1 glycoprotein has 436 amino acids and three domains (I, II and III) distributed throughout the protein. Domain II is at the amino-terminal region, domain III is at the carboxy-terminal region and domain I is between domains II and III. The E2 glycoprotein has 422 amino acids and three domains (A, B and C). Domain B is positioned at the amino-terminal region of the protein; domain C is positioned at the carboxy-terminal region; and domain A is positioned between domains B and C [14][15][16][17] .</p><p>The E1/E2 glycoproteins are directly involved in the Alphavirus infectious process. The E2 glycoprotein recognizes and binds to a target receptor on the cell membrane to promote endocytosis [18][19][20][21] . The importance of the E2 glycoprotein was demonstrated by mutation studies in domain B of CHIKV and Semliki Forest virus (SFV) 19 . In Alphavirus, the acidic endosomal environment exposes and inserts the E1 fusion loop into the endosome membrane to trigger infection. This fusion results in viral capsid release into the cell cytoplasm, where it is disassembled and the viral genetic material is released, thereby inducing viral replication and polyprotein translation 15,[18][19][20] . Cryogenic electron microscopy (cryo-EM) images of CHIKV 21,22 , Venezuelan Equine Encephalitis virus (VEEV) 23 , Barmah Forest virus (BFV) 24 and Sindbis virus (SINV) 25 show that E1/E2 glycoproteins form dimeric spicules on the membrane surface with regions exposed to the extracellular medium 24 . Exposure of E1/ E2 glycoprotein residues to the extracellular environment enables their interaction with neutralizing antibodies, making them interesting targets for therapeutic and diagnostic studies 19,[26][27][28][29] .</p><p>The development of a diagnostic platform for MAYV will be helpful in the clinical management of patients and especially for epidemiological screenings of a virus circulation pattern to prevent its spread 1 . The immediate obstacle in the development of this platform is the determination of a specific protein region that is adequate for making an accurate distinction among infectious agents 29 . However, studies have shown success by using small peptide sequences as antigens for diagnosing some Flavivirus and Alphavirus [29][30][31][32] . Bioinformatics and immunoinformatics tools can facilitate the rational search for these small peptide sequences within the viral proteome. Bioinformatics has been used in the rational design of drugs and vaccines to reduce the time and cost of their discovery and in the exploration of diagnostic platform designs for viral infections 33,34 .</p><p>To propose a peptide for application in MAYV immunodiagnostic tests, glycoprotein amino acid sequences from the MAYV and CHIKV E2 glycoproteins were analyzed according to their antigenic properties. Then, the three-dimensional structure of the E1/E2 heterodimer was predicted by molecular modeling and protein-protein docking strategies. Finally, the heterodimer produced was subsequently subjected to molecular dynamics (MD) simulations to analyze the heterodimer stability and understand the properties that could support a rationally designed peptide suitable for immunodiagnostic tests.</p><!><p>MAYV E2 glycoprotein is highly conserved. The full amino acid sequences of MAYV and CHIKV E1 and E2 glycoproteins, isolated from different endemic countries, were retrieved from the Virus Pathogens Resource (ViPR) database (see Supplementary Table S1) and aligned in Mega 7.0 software. The alignment showed that the E1 and E2 glycoprotein sequences of MAYV strains are conserved and that there are important variations in some residues compared to the sequence of the CHIKV E2 glycoprotein. The greatest divergence in the MAYV sequences found in the three domains (3.3%) was obtained in a clinical sample acquired from Haiti in 2014. Representative sequences of MAYV (GenBank Access KM400591) and CHIKV (GenBank Access KP164567) were selected for antigenicity analyses and MD simulations. peptide candidate as a potential antigen suitable for immunodiagnostic tests. Thirteen possible B-cell linear antigenic epitopes were identified in the E2 glycoprotein of MAYV and 11 possible epitopes were found in CHIKV using the Kolaskar and Tongaonkar antigenicity scale. The VaxiJen server identified seven antigenic regions of the MAYV glycoprotein and five antigenic regions of the CHIKV counterpart (Table 1 and see Supplementary Fig. S2).</p><p>Among all antigenic regions predicted for MAYV, six peptides (p_MAYV2, p_MAYV4, p_MAYV5, p_ MAYV6, p_MAYV9, p_MAYV11) were not predicted as antigenic regions in CHIKV. However, the peptide regions p_MAYV5 and p_MAYV6 may cross-react due to the presence of conserved residues in the CHIKV glycoprotein. The p_MAYV2 and p_MAYV11 peptides showed no antigenicity for MAYV according to the VaxiJen server. With the exception of the p_MAYV4 peptide, residues 107-116 in domain A, and p_MAYV9 peptide residues 266-274 in domain C, all the residues predicted for MAYV are conserved in the CHIKV E2 glycoprotein sequence (see Supplementary Fig. S3).</p><p>Thus, we consider that p_MAYV4 peptide (PGEVISVSFV) is a promising target for the development of a specific MAYV immunodiagnostic assay. To improve the antigenicity of the peptide, we hypothesized that the elongation of p_MAYV4 between residues 108-120 (GEVISVSFVDSKN), which showed an antigenicity score of 1.1595 according to VaxiJen. This peptide was named p_MAYV4a. The peptide was analyzed on the BLASTp online server (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins), and no sequence overlap with any other Alphavirus was identified. Residue 107 (PRO) was not searched because its insertion in the sequence decreased the antigenicity score of the peptide (VaxiJen Score: 0.9877).</p><p>Physicochemical properties of the p_MAYV4a peptide. The physicochemical properties of the p_ MAYV4a peptide were predicted using the ProtParam tool (http://web.expasy.org/protparam/). According to www.nature.com/scientificreports www.nature.com/scientificreports/ this prediction analysis, the peptide is 1,33 kDa, has acidic features (pI 4.37) and is probably hydrophobic, even though the index was low (GRAVY score: 0.079). In addition, p_MAYV4a is possibly stable under natural conditions (the instability score was 6.92). The yeast half-life time in vivo exceeded 20 hours.</p><p>Molecular docking of the E1/E2 glycoproteins of MAYV. Due to the absence of resolved MAYV E1/ E2 glycoprotein structures, an initial three-dimensional (3D) model was produced for each protein using the I-TASSER server. The five major templates used for MAYV E1 glycoprotein modeling were VEEV (3J0C), SINV (3J0F), CHIKV (3N42), EEEV (6MUI), and BFV (2YEW), and for E2 glycoprotein, they were VEEV (3J0C), BFV (2YEW), SINV (3J0F), CHIKV (3N40) and CHIKV (2XFB) (see Supplementary Table S3).</p><p>The best models of the E1 and E2 glycoproteins showed C-scores of 2.0 and 1.91 and root mean square deviation (RMSD) values of 2.7 Å and 3.1 Å, respectively. The TM-score for both proteins was 0.99. The E1 glycoprotein monomer had 82.7% residues in favorable regions and 97.7% in allowed regions, while the E2 monomer had 81.9% in favorable regions and 96.0% in allowed regions. The values predicted by these analyses indicate that the models determined for the E1 and E2 glycoprotein monomers have a great chance of representing their expected native structures. Thus, from these monomers, it was possible to obtain the dimeric structure from molecular docking simulations performed by ClusPro 2.0. Among the ten structures produced by the ClusPro 2.0 server, we selected the one that presented the highest similarity with the resolved structures of the Alphavirus CHIKV 13 , VEEV 22 and BFV 23 .</p><!><p>nanoseconds (ns) for the predicted MAYV E1/E2 glycoprotein dimer and its behaviors were assessed by trajectory analysis. RMSD (root mean square deviation) was used to evaluate the deviation of the predicted models from the original states during the simulation (Fig. 1A). High-magnitude RMSD fluctuations throughout the simulation can be an indication of a flexible and mobile natural protein or the adjustment of the force field. According to the results of the analysis, Fig. 1A shows that the RMSD tends to stabilize at approximately 75 ns of the simulation, with approximately 1 nm deviating from the initial structure and a periodic fluctuation of low magnitude on the basis of this value.</p><p>A complementary measure of this parameter is performed by evaluating the fluctuations resulting from movements of each of the residues in a protein, which highlights the most flexible chain segments. Therefore, we evaluated the fluctuations of the residues in two ways: the mean fluctuation throughout the trajectory and that for the duration of each simulation. The root mean square fluctuation (RMSF) analysis of each residue was performed to www.nature.com/scientificreports www.nature.com/scientificreports/ determine which residues may have caused an increase in the RMSD values (see Fig. 1B). Significant fluctuations occurred in some loops of the β-linker region of the E1 glycoprotein (residues 120-175) and E2 glycoprotein (residues ~590), in the entire E2 B domain (residues 610-675) and, especially, in E1 (residues 400-442) and E2 (residues 800-859) transmembrane regions. In addition to the fluctuations determined by the RMSF analysis, a region of instability in E1 domain III (residues 300-390), near the membrane contact sites, was identified in an RMSD-per-residue heat map (Fig. 1B). Cluster analysis with a cut-off value of 0.25 nm showed the formation of eight different clusters (Fig. 1C), but after approximately 80 ns only the cluster number 1 was formed, which prevailed until the end of the simulation. Therefore, the central structure of cluster number 1 was defined in this work as the main structure of MAYV E1/E2, and it was similar to that observed in cryo-EM of other Alphavirus, such as CHIKV, VEEV and BFV.</p><p>The MAYV E1/E2 glycoproteins were formed mainly by conserved β-sheet and α-helix secondary structural components interspersed with small structures of bend, turn and coil in the regions that connected the E1 and E2 domains (see Supplementary Fig. S4A). Validation of the model generated by the MD simulation was performed by the MolProbity server, which presented results in Ramachandran plots (see Supplementary Fig. S4B). This structural analysis performed after the MD simulation revealed that 91.2% of the residues were in favorable regions and 98.9% were in allowed regions, showing refinement and improvement of the structure compared to the initial dimer model previously presented.</p><p>The p_MAY4a peptide is stable and exposed at the surface of the MAYV glycoprotein dimeric structure. The p_MAY4a peptide (PGEVISVSFVDSKN) is located in domain A of the E2 glycoprotein (108-120) (Fig. 1D). It is specific to MAYV and has a secondary structure organized as a β-sheet with short structural segments in bend and coil forms (see Supplementary Fig. S4A). The peptide has a protruding structure in the distal portion, which is exposed to the solvent and theoretically accessible to antibodies (Fig. 2A,B). The analysis www.nature.com/scientificreports www.nature.com/scientificreports/ of the solvent accessible exposure area (SASA) showed that the p_MAYV4a peptide exhibited solvent exposure that ranged from 47.77 to 51.14 nm 2 throughout the simulation trajectory (Fig. 2C). The peptide has an aliphatic feature with two distinct regions: a β-sheet structure (residues 110-115), composed predominantly of hydrophobic side chain residues that establish hydrogen interactions with the neighboring antiparallel structure (residues 560-565, 521 and 523) (Fig. 2D), and a loop structure (residues 116-120) with hydrophilic side chain residues. Constant residue exposure and maintenance of the secondary structure of the peptide throughout the simulation were fundamental for the recognition of the neutralizing antibodies produced during MAYV infection.</p><p>PHE95 and TRP89 residues in the E1 glycoprotein are important for MAYV E1/E2 dimer stabilization. The PHE95 residue in the E1 fusion loop (domain I) is conserved among alphaviruses and interacts with the GLN226, TYR228 and ARG178 residues of E2 glycoprotein (domain B) in the binding crevice (Fig. 3A and Supplementary Video S1). These three residues flank and interact with the hydrophobic residue PHE95 (SASA mean of 5.88 nm 2 ), retaining the fusion loop in the binding crevice between domains A and B (Fig. 3B,C). Finally, we identified the TRP89 residue in the fusion loop of the E1 glycoprotein, which is conserved among Alphavirus species and possibly involved in the formation of the dimer. TRP89 and PHE95 residues interact with the residues on the E2 glycoprotein during dimer formation. As shown in Fig. 3D, the PHE95 residue is buried in an E2 glycoprotein cavity, while the TRP89 residue is positioned externally (SASA mean of 23.70 nm 2 ), thus forming a staple-like structure that fixes the E1 fusion loop to the E2 glycoprotein.</p><!><p>Arboviruses (ARthropod-BOrne VIRUS) are transmitted by arthropod vectors, such as mosquitoes and ticks 28 . These viruses are a challenge to public health authorities in endemic regions 6 . Continuous monitoring is required to avoid the dispersion of these viruses among naïve populations and prevent the emergence of new outbreaks 29 . MAYV is an emerging arbovirus in the Pan-Amazon region that causes restricted and self-limited outbreaks in South and Central America 2 . However, although Haemagogus spp. mosquitoes are the main vectors involved in the transmission of MAYV, Aedes spp. mosquitoes, found in peri-domiciliary areas of large cities in the Americas and Europe, may be involved in viral transmission. This possibility has raised concern with its dissemination as an important arboviral agent 1,2,6 .</p><p>The lack of accurate and simplified diagnostic platforms and prophylactic alternatives are obstacles to overcome 29,[35][36][37] . The rational design of antigenic peptides using prediction tools in silico is a promising option in the development of immunodiagnostic platforms 29,32 . Therefore, this study was developed to fill two gaps related The results from the analyses of MAYV E2 glycoprotein in silico identified two potential antigenic linear peptides, the p_MAYV4 peptide present in domain A (residues 107-116) and the p_MAYV9 peptide inserted into domain C (residues 266-274). The p_MAYV4 peptide is more likely to be recognized by neutralizing antibodies produced during MAYV infection because domain A of the E2 glycoprotein has greater solvent exposure than domain C 17,35 . It is known that the neutralizing antibodies against Alphavirus are directed mainly to targets present in domains A and B of the E2 glycoprotein 21,26,28,35,36 . Therefore, due to the location of the p_MAYV4a peptide in domain A of the E2 glycoprotein, close to the cellular receptor binding site (THR58) 38,39 , we hypothesize that antibody binding in this region is associated with inhibited cell recognition.</p><p>The side chains of the hydrophobic amino acids in the β-sheet region (residues 110-115) of the peptide form hydrogen interactions with the neighboring antiparallel structure and provide stability to the peptide in the protein 40 . This interaction is important to ensure that the loop region of the peptide is continuously exposed to the solvent and to enable access by the antibody 41 . The loop region increased the antigenicity score of the p_MAYV4 peptide according to the VaxiJen server (score of 1.1595). This region is composed of four hydrophilic amino acid residues (ASP117, SER118, LYS119 and ASN120) arranged in a β-turn secondary structure at the amino-terminus. β-Turn secondary structures are more accessible to solvent and are more hydrophilic and flexible than β-sheets or α-helices. In addition, they have greater flexibility potential and hydrophilic scores, which are desired in the construction of an antigenic peptide 42,43 . The four residues were constantly exposed to solvent during the MD simulation. However, the insertion of the segment should be performed with caution since other Alphaviruses, such as CHIKV and VEEV, have a similar secondary structure 14,23,30,44 . Although these viruses have the same spatial arrangement in terms of the E2 glycoprotein structure, the primary sequence of the amino acid residues identified in CHIKV 30 and VEEV is completely different from the one proposed for MAYV (p_MAYV4a).</p><p>The envelope proteins of Alphaviruses (E1/E2 glycoproteins) are fundamental during the infection process 17,20,25 ; therefore, we proposed a structural model of the MAYV heterodimer generated by I-TASSER and submitted it for an MD simulation of 150 ns. The high protein RMSD value along the trajectory indicates that a significant change in the molecule, compared to the original structure, has occurred. However, the stability reached by the protein after 75 ns, persisting at a value of approximately 1 nm, illustrates its stability after the www.nature.com/scientificreports www.nature.com/scientificreports/ conformational adjustment. While RMSD is an analysis of the protein as a whole, the RMSF analysis describes the individual residue behavior along the trajectory. In this case, the RMSF analysis showed fluctuations in important regions of the molecule, which exemplifies the flexibility and mobility of the protein under study. The cluster analysis corroborates the RMSD profile when the stability indicated by the RMSD value is reached, and the time for the number of clusters to decrease is similar. It is noteworthy that, after 75 ns, only three clusters appear, and cluster number 1 is the most prevalent among them. This shows that cluster number 1 represents the most stable protein structures. In view of this finding, the central structure is determined to be cluster number 1, which is defined as the structure with the smallest average RMSD compared to all other structures of the cluster and resulted in a protein similar to the crystal structures of other species of Alphavirus 13,[22][23][24] ; thus, it was used as the MAYV E1/E2 model for this study.</p><p>Structurally, domain B is located at the distal end in relation to the membrane, positioned above domain II of E1 and protecting the fusion loop at neutral pH. The fluctuation observed in this domain may be due to a natural flexibility caused by its long β-linker 15 or by the absence of glycoprotein E3, positioned next to domains A and B. When E3 glycoprotein is present, domain B remains stable and protects the fusion loop, but after cleavage, this domain moves to the membrane fusion 14,22 . The absence of the trimeric structure in the presence of neutral conditions (pH 7.0) during the simulation may have triggered instability in domain B and in the fusion loop. To evaluate this possibility, further studies of MD with the MAYV heterotrimer (E1, E2 and E3) will be carried out in different pH conditions to accurately determine the position and behavior of domain B in the viral particle throughout the infectious stages.</p><p>Cryo-EM studies show that the flexibility of domain B in Alphavirus, such as CHIKV, Sindbis virus and SFV, is necessary for them to bind to cellular receptors during the infectious process. In this scenario, domain B and its β-connectors move to expose the E1 fusion loop. The high values of RMSD and RMSF obtained in this study show equal flexibility of domain B of the MAYV. Our results also indicate that the movement observed in the regions of β-connectors may be useful for the protein to adapt to the environment during the fluctuations of domain B until a moment of stability is found.</p><p>Since there are no vaccines available that neutralize the infection by Alphavirus, the movement described in domain B of the MAYV is an interesting target to be explored for the development of vaccines using neutralizing antibodies. Neutralizing antibodies produced against the epitope described in this work or that recognize other regions of domain B, domain A, or the β-connectors may be sufficient to prevent their movement, exposure of the E1 fusion loop and consequent viral infection. Currently, some works explore the development of antibodies against Alphavirus and describe that antibodies produced against domain B, domain A or β-linkers are highly effective in neutralizing the infection 16,21,25,35,36 . However, these works do not exploit peptide regions of the glycoproteins for the determination of targets for neutralizing antibodies. In this study, we explored an approach in silico that enabled us to characterize the E1/E2 glycoprotein structures, to identify important movements and behaviors of the virus, and to identify an antigenic region in domain A.</p><p>A previous CHIKV study showed that residues HIS29, HIS73 and HIS226 of the E2 glycoprotein contribute to dimer stability 14 . Two of those residues (HIS29 and HIS73) are conserved in MAYV, but not HIS226, substituted by H226Q. Our analysis adds that the amino acid PHE95, present in the E1 fusion loop, is conserved among Alphavirus and is important for dimer fixation when interacting with the amino acids TYR228, ARG178 and GLN226 that are present in domain B and in the β-linker of the E2 glycoprotein. The MD simulation results show that these three residues form a cage-like structure that theoretically imprisons the nonpolar residue PHE95. This result will be evaluated by alanine scanning and experiments in vitro in the future. If the peptide predicted in this work is confirmed, then PHE95 is a promising therapeutic target candidate.</p><p>In summary, our results highlight an antigenic sequence in the MAYV E2 glycoprotein with potential for the development of immunodiagnostic platforms. In addition, they suggest a structural model of the MAYV E1/E2 heterodimer with insights for a better understanding of its structure and behavior, making it a hot topic for further studies involving mutagenesis and drug therapy against MAYV and CHIKV, major arthritogenic Alphaviruses. Finally, we emphasize that these results will be applied in subsequent confirmatory in vitro tests.</p><!><p>shown in Supplementary Fig. S5. Complete E1 and E2 glycoprotein amino acid sequences of MAYV and CHIKV were obtained from the Virus Pathogens Research (ViPR) database (https://www.viprbrc.org) and aligned using Muscle software implemented in the Mega 7.0 program 45 . Representative sequences of MAYV (GenBank KM400591.1) and CHIKV (GenBank KP164567) were selected for further analysis.</p><p>prediction of continuous linear B-cell epitopes. MAYV linear B-cell epitopes were predicted using the Kolaskar and Tongaonkar antigenicity scale 46 (http://tools.immuneepitope.org/bcell/). The Kolaskar and Tongaonkar antigenicity scale is a semiempirical epitope prediction method with more than 75% prediction accuracy 46 . Peptides that reached or crossed the threshold of 1.05 were classified as potential antigenic epitopes 47 . Potential antigenic sequences specific for MAYV as determined through the use of the antigenicity scale of Kolaskar and Tongaonkar were submitted to a second online antigenicity prediction platform, VaxiJen (http:// www.ddg-pharmfac.net/vaxijen) 48 . The VaxiJen server is an alignment-independent antigen predictor with 87% viral epitope prediction accuracy 47 . These antigen prediction methodologies are based on the physicochemical properties of amino acid residues and their frequency of identification in previous experimental studies 45,46 .</p><p>prediction of the physicochemical properties of epitopes. Physicochemical properties of the MAYV antigenic sequences, including half-life, instability index, aliphatic index, theoretical pI and the hydropathicity value, were predicted using the ProtParam online tool 49 (http://web.expasy.org/protparam/). The half-life (2019) 9:13339 | https://doi.org/10.1038/s41598-019-50008-3 www.nature.com/scientificreports www.nature.com/scientificreports/ prediction estimates how long a peptide remains stable in prokaryotic and eukaryotic organisms. A protein is considered stable-when the value obtained is lower than the cut-off value of 40, while the hydropathicity index evaluates the probability of a region being hydrophobic (positive values) or hydrophilic (negative values) 32 . The secondary structure of the peptide was assessed by a graphic representation generated by the MD.</p><p>Prediction of the 3D structure of the Mayaro virus E1 and E2 monomers and glycoprotein dimer docking. Due to the absence of resolved dimer and monomer structures of the MAYV E1/E2 glycoprotein in the PDB (Protein Data Bank), a three-dimensional (3D) structural model was generated using a threading modeling methodology on the I-TASSER online prediction server 50 (https://zhanglab.ccmb.med.umich. edu/I-TASSER/). Based on the primary amino acid sequences of select E1 and E2 glycoproteins (KM400591.1), the I-TASSER determines the 3D structure by iterative simulations of segmentation assembly [50][51][52] .</p><p>The quality of the structural model generated by I-TASSER was evaluated with the MolProbity server 53 (http:// molprobity.biochem.duke.edu/index.php). This server uses two measures to evaluate the quality of the produced model: the clashscore, which evaluates the number of severe steric overlays per 1,000 atoms, and the MolProbity value, which combines the clashscore, rotameter and Ramachandran ratings into a single score that is normalized such that it is on the same scale as the X-ray resolution. These combined measures enable the quality assessment of a geometric model 53 .</p><p>For the formation of the E1/E2 dimeric structure of the MAYV envelope glycoprotein, the best models produced by the I-TASSER were submitted to the ClusPro 2.0 online server 54 (https://cluspro.bu.edu/login.php). The server performs molecular docking according to free energy parameters and produces 10 models ordered according to the structure requiring the lowest free energy for docking. The best output hits from ClusPro were submitted for refinement and analysis of the MD simulation 53 .</p><p>Molecular dynamics simulations of the dimeric E1/E2 glycoprotein of the Mayaro virus. Correct disulfide bonds between cysteine residues in the I-TASSER 3D model were determined by the tleap tool using the AMBER ff14SB force field from the AMBER18 package 55,56 . The sugar residues were fixed using the GLYCAM_06j-1force field 57 (http://glycam.org/) and histidine residues predicted by the H++ server 58 (http:// biophysics.cs.vt.edu/credits.php) were protonated using the AmberTools 55 . The protein format was converted using ACPYPE 59 , and the system was minimized and equilibrated with a TIP3P water model and CL − ions 60 . Transmembrane protein domains at residues 397-436 in the E1 glycoprotein and 346-422 in E2 glycoprotein were restricted throughout the simulation. The MD simulation was performed using GROMACS 5.1.2 software 61 in the AMBER ff99SB-ILDN force field 62 . The system was subjected to 150 ns simulation at 300 k temperature and 1 bar pressure, with restriction confirmed for only the transmembrane domain. The trajectory analysis was performed using RMSD and RMSF, calculated from GROMACS tools package, in the gromos algorithm. The RMSD is a measure of the spatial difference between two static structures, and in the simulation, the calculation was performed on the basis of the initial structure and all succeeding trajectory frames. On the other hand, the RMSF profile calculates the flexibility of a residue based on the fluctuation around an average position among all MD simulations 63 . The g_cluster (GROMACS) program, based on the RMSD profile with a cut-off of 0.25 nm, was used to determine the conformations that were found most frequently along the trajectory. In this case, all structures with RMSD values less than 0.25 nm for any element in a cluster are added to the primary cluster. It is unlikely that a molecule with an RMSD value higher than 0.25 nm from another cluster would be considered a structure. UCSF Chimera 64 and Visual Molecular Dynamics (VMD) 65 were used to visualize protein behavior throughout the simulation. The quality of the proposed model was evaluated using the MolProbity server 51 .</p>

Scientific Reports - Nature

Cationic surface modification of gold nanoparticles for enhanced cellular uptake and X-ray radiation therapy

A challenge of X-ray radiation therapy is that high dose X-ray can damage normal cells and cause side effects. This paper describes a new nanoparticle-based method to reduce X-ray dose in radiation therapy by internalization of gold nanoparticles that are modified with cationic molecules into cancer cells. A cationic thiol molecule is synthesized and used to modify gold nanoparticles in a one-step reaction. The modified nanoparticles can penetrate cell membranes at high yield. By bring radio-sensitizing gold nanoparticles closer to nuclei where DNA is stored, the total X-ray dose needed to kill cancer cells has been reduced. The simulation of X-ray-gold nanoparticle interaction also indicates that Auger electrons contribute more than photoelectrons.

cationic_surface_modification_of_gold_nanoparticles_for_enhanced_cellular_uptake_and_x-ray_radiation

1. Introduction<!>2.1 Materials and chemicals<!>2.2 Synthesis of MTAB<!>2.3 Cell culture and cell viability test<!>2.4 Quantifying number of internalized gold nanoparticles<!>3. Results and discussions<!>4. Conclusions

<p>A challenge of X-ray radiation therapy is that high dose radiation can damage normal cells and cause side effect due to low tumor selectivity.1 A variety of beam techniques have been developed to minimize dose on normal cells or maximize dose on cancer cells, but the methods are still limited by low precision of planning and positioning, low spatial resolution due to patient motion during treatment, and cannot treat hard-reaching tumors or tumors with undefined boundary.2–4 Radiosensitizers including oxygen, blood substitutes carrying oxygen, and radiosensitive drugs have been used to enhance efficacy of a given X-ray dose, but, damages to normal cells remain significant when X-ray dose is sufficient to kill tumors due to few reasons: inadequate delivery of radio-sensitive agents, finite targeting sites at tumor, large distance for free radicals to diffuse from sites of production (outside cell) to sites of action (inside cell), and early termination of free radical chain reactions.5–9 All these factors can cause ineffectiveness of radiation therapy, and therefore high dose X-ray is often required for cancer-killing.</p><p>In radiation therapy, X-ray photons generate photoelectrons and Auger electrons, which cause ionization of water and formation of reactive free radicals (mostly hydroxyl radicals). Free radicals diffuse through chain reactions into cells, and damage DNA in mitochondria and nuclei by extracting hydrogen atoms from ribose sugars, leading to cleavage of polynucleotide backbone.10–18 In normal condition, cells can repair damaged DNA. But, when the damage rate is higher than repair rate, damages are inherited and accumulated through cell division, causing cell to die or reproduce slowly.19–23 A typical diffusion length of hydroxyl free radical in an aqueous solution is ~200 nm (in the presence of scavenger), shorter than the distance from cell membrane to cell nucleus. If radiosensitizers can be placed in cancer cells or nuclei, the distance from site of production to nucleus will be reduced. The amount of free radicals available for DNA damage will be enhanced. The cell membrane penetrating ability of nanoparticle is dependent on the sizes, shapes and surface properties (charge and hydrophobicity).24–29 While neutral groups normally prevent nanoparticle adsorption, charged groups are primarily responsible for internalization in cells via endocytosis.30–32 A large amount of natural or synthetic nanoparticles with cationic surface charges can penetrate membrane, escape endosomes, and enter cytoplasm or nucleus. Nanoparticles modified with cell-penetrating peptides or antibodies can enter cells and chaperon cargoes in cytosol.33–35 But these methods require expensive reagents and multiple steps for modification.</p><p>Gold nanoparticles are considered bio-compatible and promising as radio-sensitizer. It is expected that the cationic modification of gold nanoparticles will enhance attachment of nanoparticles on cell membrane due to electrostatic attraction, which will lead to a higher chance of nanoparticle endocytosis. An issue for cationic modification of gold nanoparticles is that normal thiol chemistry leads to carboxyl terminated monolayers, and several additional operations will have to be taken sequentially to alter the surface charge polarity, where the multiple steps of washing, centrifuging and incubation tend to decrease yield of modification. This paper describes the synthesis and use of a thiol based cationic molecule that can be used to modify gold nanoparticles in a single step to form cationic nanoparticles that can be internalized in cancer cells at high yield. Upon irradiated with X-rays, cancer cells are killed at much lower dose.</p><!><p>Vybrant live/dead viability/cytotoxicity kit is from Invitrogen (Carlsbad, CA). RPMI 1640 media, penicillin, streptomycin, fetal bovine serum (FBS), and Dulbecco's phosphate-buffered saline (D-PBS) and gold nanoparticles with the size of 10 nm at concentration of about 100 nM in 0.1 mM PBS are from Sigma–Aldrich (St. Louis, MO). Ultrapure water (18.2 MΩcm−1) from a Nanopure System (Barnstead, Kirkland, WA) is used throughout our experiments. The fluorescent images and dark field images are taken by a fluorescence microscope from Olympus (BX51M) in fluorescence mode and dark field mode, respectively. Synergy HT multimode microplate reader from Biotek (Winooski, VT) is used for absorbance and fluorescence measurements. In order to image nanoparticles, a suspension droplet of gold nanoparticles is dropped on carbon coated copper grid and allowed to dry at room temperature. A JEOL 1011 transmission electron microscope (TEM) operated at 100 kV is used to image nanoparticles. A Mini-X portable X-ray tube (Amptek, Bedford, MA) with a silver anode operating at 40 kV and 100 mA is used to generate primary X-rays and irradiate cells at a distance of 5 cm.</p><!><p>16 mercapto-hexadecyl trimethylammonium bromide (MTAB) is synthesized according to Figure 1, which is following a literature method.36 3.93 g of triphenylphosphine (Ph3P) is added in 50 ml anhydrous tetrahydrofuran (THF); 2.67 g N-bromosuccinamide (NBS) is added into another 50 ml THF. Both solutions are mixed at 0 °C under vigorous stirring. Then, a solution of hexadecane-1,16-diol (1g) in 25 ml THF is slowly added to the mixture of NBS and Ph3P. The resulting solution is heated at 60 °C and stirred for 4 hours. After removing THF by rotary evaporation, the residue is re-crystallized from ethanol to obtain 1.1 g white powder (70% yield), which is tested by 1H NMR. 1H NMR (CDCl3, 400 MHz): δ 1.26–1.46 (m, 24 H), 1.85 (q, 4H), 3.41 (t, 4H). 1 g of 1,16-dibromohexadecane is dissolved in 40 ml methanol, and degassed in argon for 1 hour. 124 mg of sodium methoxide and 204 mg of thioacetic acid are dissolved in 12 ml anhydrous ice-cold methanol, and refluxed in argon. The content in the flask is slowly added to the solution over 4 hours duration. After reaction, the content of the flask is cooled to room temperature, and methanol is removed at reduced pressure. The yellow oil is purified by column chromatography (20% ethyl acetate in hexane) to obtain 480 mg of 16-bromo-1-hexadecane-thioacetate (50% yield). 4 ml of acetyl chloride is added drop-wise to a stirred solution of 16-bromo-1-hexa-decanethioacetate (400 mg) in 10 ml of methanol, followed by keeping at 50 °C for 4 hours. 200 ml of CH2Cl2 is added to the reaction mixture, and excess acetyl chloride and HCl are removed by extractions with deionized water. Methylene chloride is evaporated at reduced pressure to obtain 284 mg of 16-bromo-1-hexa-decanethiol as colorless oil (80% yield). 1H NMR (CDCl3, 400 MHz): δ 1.26–1.46 (m, 25H), 1.60 (m, 2H), 1.85 (q, 2H), 2.52 (q, 2H), 3.41 (t, 2H). 3 ml of 4.2 M ethanolic solution of trimethylamine is added to a solution of 16-bromo-1-hexadecanethiol (284 mg) in 5 ml of ethyl acetate. The mixture is vigorously stirred in argon for 4 days. The resulting white precipitate is filtered and washed with ethyl acetate to remove excess trimethylamine. The residue is dried in vacuum to obtain 270 mg of MTAB (80% yield). 1H NMR (CDCl3, 400 MHz): δ 1.26–1.46 (m, 25H), 1.60 (m, 2H), 1.85 (m, 2H), 2.52 (q, 2H), 3.5 (s, 9H), 3.55–3.7 (m, 2H). The final yield of MTAB is 22%.</p><!><p>HeLa (CCL-2) cells are from American type culture collection (ATCC, Manassas, VA) and cultured in RMPI 1640 medium, supplemented with penicillin(100 U/ml), streptomycin (100 µg/ml), and 10% FBS, followed by a culture in a 5% CO2 incubator at 37 °C according to the protocol from ATCC. To determine cell cytotoxicity, 200 µl of suspension is seeded in each well of 96-well microplate at concentration of 1×105 cell/ml, followed by overnight culturing in 5% CO2 at 37 °C. Cells are exposed to different concentration (0.05, 0.1, 0.5, 1 nM) of citric acid (CA) or MTAB modified gold nanoparticles incubated for 24 hours. Cell viability after exposing to nanoparticles is determined by Calcein AM/EthD-1 assay, performed as follows. 100 µl of D-PBS is added in each well to wash cells and dilute serum-containing esterase, which can lead to false positive. A 100 µl of dual fluorescence calcein AM/EthD-1 assay reagents is added in each well and incubated for 30 min at room temperature prior to fluorescence measurement. The microplate is readout with Synergy HT multimode microplate reader from Biotek (Winooski, VT), where fluorescence signals are measured at 530 and 630nm, respectively. The backgrounds are subtracted before calculation by measuring a cell-free control. The percentages of live cells and dead cells are derived by dividing the fluorescence intensities of live or dead cells with the values obtained for controls. Each experimental condition is repeated for six times.</p><!><p>The concentration of gold nanoparticles internalized into cells is determined with inductively coupled plasma mass spectroscopy (ICP-MS) (Complete Analysis Laboratories Inc., Parsippany, NJ). Briefly, HeLa cells are plated into a 6-well plate at 1×105 cells per each well. After co-incubation with CA- or MTAB-gold nanoparticles for 24 hours, the medium is removed, and cells are washed with 1× PBS for 3 times to remove nanoparticles that adhered to cell membrane. The washed cells are harvested from plate with trypsin-EDTA and then centrifuged to form pellet. The cell pellets are digested with 500 µl aqua regia for 20 min, and gold concentrations are measured by ICP-MS. Gold nanoparticles with known concentrations are used as standards.</p><p>About 105 cells are treated with 2 ml of medium containing 0.05, 0.1, 0.5 and 1 nM MTAB-nanoparticles. From ICP-MS, the total mass of gold nanoparticles taken by these cells is 0.7 µg/105 cells, or 7×10−12 g gold per cell at nanoparticle concentration of 1 nM. Given gold density of 19.30 g/cm3, the mass of a nanoparticle with 10 nm diameter will be 10−17 g. Thus, each cell contains approximately 6.9×105 gold nanoparticles.</p><!><p>Fig. 1A shows the procedure of making the cationic molecule, 16 mercaptohexadecyl trimethylammonium bromide (MTAB). 1,16-hexadecanediol is converted into dibromide through standard bromination reaction, followed by conversion into monothioester. The resulting 16-bromo-1-hexadecanethiol is methylated to yield the final product. Purified MTAB is water-soluble, which allows ligand exchange with citric acid (CA)-coated gold nanoparticles (10nm) in aqueous medium. 5 mg of MTAB thiol is added into a 5 ml suspension of gold nanoparticles, followed by vigorous stirring in ambient condition for 24 hours. Gold nanoparticles is dialyzed against 0.1 mM PBS buffer solution with 3.5 kD cutoff dialysis membrane, and filtered with 0.2 µm membrane. After ligand exchange, the obtained MTAB-gold nanoparticles are stable in phosphate buffer saline (PBS). Fig. 1B shows a TEM image of MTAB-gold nanoparticles. The size of nanoparticles is ~10 nm, and no aggregation of gold nanoparticles is observed. Fig. 1C shows UV-Vis absorption spectra of gold nanoparticles before (black) and after (red) MTAB exchange, where the plasmonic peak shift is induced by MTAB monolayer formation on surfaces of gold nanoparticles. The zeta potential measurement is carried out to qualitatively describe the charge surrounding a nanoparticle. The zeta potential of the MTAB modified gold nanoparticles is found to be +53 mV, suggesting the cationic quaternary ammonium groups positioned around the gold nanoparticles.</p><p>HeLa cells are obtained from ATCC and grown in RPMI 1640 culture media that contain penicillin (100 Unit/ml), streptomycin (100 µg/ml), and 10% FBS in an incubator with 5% CO2 at 37 °C. After cell monolayer reaches 80% confluence, cells are incubated with CA- or MTAB-gold nanoparticles at concentrations of 0.05, 0.1, 0.5 and 1 nM, respectively. The growth media are removed after 24 hours, and cells are washed 3 times with 1× PBS to remove excess gold nanoparticles physically adsorbed on cell surface. Due to strong light scattering ability, nanoparticles can be observed with dark field optical microscopy. Fig. 2A–C are optical images of control cells, cells treated with CA-gold nanoparticles, and cells treated with MTAB-gold nanoparticles, respectively. Compared with CA modified ones, a large amount of MTAB-gold nanoparticles enter cells. Inductively coupled plasma-mass spectrometry (ICP-MS) that can detect metals at low concentration has been used to quantify the number of gold nanoparticles taken by cells. Fig. 2D shows that each cell uptakes an average of 690,000 gold nanoparticles after incubating with 1 nM suspension of MTAB-gold nanoparticles for 24 hours. The image shows that nanoparticles are clustered inside cells, suggesting that MTAB-gold nanoparticles enter cells via an endosomal pathway.36</p><p>Cells have been irradiated with X-ray (40 kVp, 100 µA) at a dose rate of 0.6 Gy/min. Cell viability is measured immediately after X-ray irradiation using Calcein AM/EthD-1 assay. No obvious cell death is observed when exposure time is less than 10 min. In order to minimize X-ray dose, the X-ray exposure time is set at 5 min. Fig. 3A–B shows the fluorescence images of cells treated with CA-gold nanoparticles and MTAB-gold nanoparticles, and irradiated with X-ray for 5min. The green and red colors show viable and dead cells, respectively. Most cells treated with CA-gold nanoparticles are alive; many cells treated with MTAB-gold nanoparticles are dead. Fig. 3C shows that cell viability depends on the concentration of MTAB-gold nanoparticles at irradiation time of 5 min, where cell viability decreases as the nanoparticle concentration increases. In contrast, MTAB-gold nanoparticles alone (black column), and X-ray alone (see supporting Fig. S1) do not cause much cell death. Thus, internalized gold nanoparticles cause the most cell death in the presence of X-ray.</p><p>In order to understand effects of internalized gold nanoparticles in X-ray radiation based cell killing, an analytical approach is used to derive radio-sensitizing capabilities of MTAB-gold nanoparticles inside cells (dimension of 2×10×10 µm3).37 In this model, the radius of sphere with nanoparticle in center is equal to the range of emitted electrons (both photoelectrons and Auger electrons). The ratio of the dose delivered to cells with and without the nanoparticles is known as the dose enhancement factor (DEF). Fig. 3D shows the calculated dose enhancement factor (DEF) of electrons versus the number of internalized gold nanoparticles (100, 101, 102, 103, 104, 105 and 106) in each cell, respectively. Both of DEFs of Auger electrons and photoelectrons increases with the number of nanoparticles in cells. However, the DEFs of Auger electrons is considerably higher than those from photoelectrons at the same nanoparticle numbers. This is attributed to the short-range (< 1 µm) of Auger electrons, leading to deposition of more energy in vicinity of X-ray irradiated nanoparticles. As a result of near-particle energy deposition, dose enhancement within few hundred nanometers from nanoparticle is dominated by Auger electrons.</p><!><p>This paper describes a new surface chemistry method to enhance radiation killing of cancer cells by internalizing gold nanoparticles into cancer cells. The cationic monolayers around gold nanoparticles will allow a large number of nanoparticles interanlized into viable cells, which can lead to cell killing at much lower X-ray radiation dose. The adoption of this approach in radiation therapy will be of great importance, because 50% of cancer patients will have to use radiation therapy at certain stage of disease.</p>

PubMed Author Manuscript

DYNAMIC MONITORING OF GLUCAGON SECRETION FROM LIVING CELLS ON A MICROFLUIDIC CHIP

A rapid microfluidic based capillary electrophoresis immunoassay (CEIA) was developed for on-line monitoring of glucagon secretion from pancreatic islets of Langerhans. In the device, a cell chamber containing living islets was perfused with buffers containing either high or low glucose concentration. Perfusate was continuously sampled by electroosmosis through a separate channel on the chip. The perfusate was mixed on-line with fluorescein isothiocyanate-labeled glucagon (FITC-glucagon) and monoclonal anti-glucagon antibody. To minimize sample dilution, the on-chip mixing ratio of sampled perfusate to reagents was maximized by allowing reagents to only be added by diffusion. Every 6 s the reaction mixture was injected onto a 1.5 cm separation channel where free FITC-glucagon and the FITC-glucagon-antibody complex were separated under an electric field of 700 V cm\xe2\x88\x921. The immunoassay had a detection limit of 1 nM. Groups of islets were quantitatively monitored for changes in glucagon secretion as the glucose concentration was decreased from 15 to 1 mM in the perfusate revealing a pulse of glucagon secretion during a step change. The highly automated system should be enable studies of the regulation of glucagon and its potential role in diabetes and obesity. The method also further demonstrates the potential of rapid CEIA on microfluidic systems for monitoring cellular function.

dynamic_monitoring_of_glucagon_secretion_from_living_cells_on_a_microfluidic_chip

Introduction<!>Chemicals and Reagents<!>Preparation of FITC-glucagon<!>Instrumentation<!>Isolation and Protocol of Islet Measurements<!>Off-line Mixing Assay<!>On-line Mixing Assay<!>On-line Islet Monitoring<!>Conclusions

<p>Microfluidics has been increasingly used in combination with cell culture for in vitro cell biology studies. Microfluidics allows precise control over cellular environment enabling many experiments not possible with conventional techniques [1–3]. In such experiments, it is often of interest to perform chemical measurements on cells to understand how the cellular environment has affected cell function or biochemistry. Most commonly optical measurements are used to probe cells on chips, but the versatility of cell studies on microfluidic systems can be improved by integrating other chemical measurements. In this work, we demonstrate an approach to measure hormone release from small groups of cells on chips using an electrophoretic immunoassay.</p><p>A variety of chemicals are released from cells, including signaling molecules such as hormones or neurotransmitters, trophic and growth factors, and metabolic products. Temporally resolved measurements of cell releasates are important in studying the regulation of secretion, determining the effect of drugs (or drug candidates) on cell function, and assessing the health of cells used in transplant or tissue engineering. Measurements of chemical release from cells are typically performed by perfusing cells, collecting perfusate in fractions, and then performing offline analysis by immunoassay or other appropriate methods. This process can be automated and miniaturized in microfluidic systems by direct measurement of release using sensors built into the chip or perfusing cells and mixing the perfusate with assay reagents on-line (if necessary) so that the result of the assay reaction can be detected downstream [4–11]. Demonstrated advantages relative to conventional approaches include: 1) automation; 2) better temporal resolution because of the speed of assays [12]; 3) better sensitivity by reduction of volume and dilution; 4) reduction of amount of cells and reagents required; 5) continuous measurements for up to 24 h [13]; and 6) complex environmental control integrated with measurement [14].</p><p>Sensors, enzyme assay, fluorescent assay, and capillary electrophoretic immunoassay (CEIA) have been used for chemical measurement of secretions on chips [4–11]. CEIA is particularly powerful because it can be used for a wide variety of compounds. In this approach, fluorescent antigen (Ag*) and antibody (Ab) are added to the perfusate containing antigen (Ag) [15]. Electrophoretic separation allows detection of the Ag*-Ab complex and free Ag*. The relative size of these peaks is quantitatively related to the unlabeled Ag present in the sample. Such systems are well suited for monitoring applications because they can be fast (on-line separations in seconds) and sensitive. Further, they have been shown to be stable enough to allow over 104 assays in one monitoring session [13, 16].</p><p>Previous work on using CEIA for on-line cell monitoring has been confined to measuring insulin release from single islets of Langerhans [9, 12]. Islets are 75–200 μm diameter spheroid microorgans located in the pancreas that contain 2000–4000 endocrine cells each. Although insulin is the most abundant hormone released from islets, several other hormones such as glucagon are also released from islets to help regulate metabolism. In this work, we have adapted the device to allow for monitoring of glucagon secretion from batches of islets.</p><p>Glucagon is a 29-amino acid peptide hormone that acts as the major counter-regulatory signal to insulin [17]. Glucagon is secreted from α-cells in response to low circulating glucose levels to stimulate hepatic glucose production. It is increasingly realized that abnormal regulation of glucagon may play a role in type 2 diabetes and obesity, therefore research into its secretion is growing [18–20]. Study of glucagon secretion is difficult because α-cells comprise only 10–15% of the total islet cell population resulting in relatively low amounts of hormone per islet [21] and small changes in glucagon secretion upon stimulation [22].</p><p>In addition to adaptation of the immunoassay to glucagon, we have modified a previously described microfluidic system [12] to implement a novel approach to reagent addition that minimizes dilution of sample. These results provide a novel tool for studying glucagon secretion, validate a new method of mixing reagents that minimizes dilution of sample, and demonstrates the versatility of the microfluidic CEIA approach for cell secretion measurements.</p><!><p>Cell culture reagents and fluorescein-5-isothiocyanate (FITC) were from Invitrogen (Carlsbad, CA). Tween-20, collagenase type XI, ethylenediaminetetraacetic acid (EDTA), bovine serum albumin (BSA), glucagon and monoclonal antibody to glucagon (clone K79bB10; Ka = 109 M−1) were from Sigma-Aldrich (St. Louis, MO). All solutions were made from Milli-Q (Millipore, Bedford, MA) ≥ 18 MΩ cm−1 deionized water. All other chemicals were obtained from Fisher (Hampton, NJ) and were of the highest purity available. Glucagon standard solutions were prepared daily. Stock Ab was maintained in aliquots at −32 °C in the manufacturer provided ascites fluid. FITC-glucagon fractions (labeling described below) were stored at −32 °C in HPLC buffer. Bulk solutions were filtered using 0.2-μm nylon syringe filters (Fisher). Glucagon or antibody containing solutions were filtered using 0.2-μm inorganic Anotop 10 syringe filters (Whatman, Florham Park, NJ) pre-conditioned with BSA containing solutions.</p><p>Reagent buffer consisted of 20 mM NaCl, 50 mM tricine, 1 mM EDTA, pH 7.4, supplemented with 1 mg/mL BSA and 0.1% (w/v) Tween 20. Electrophoresis buffer consisted of 150 mM tricine, 20 mM NaCl at pH 7.4. Balanced salt solution consisted of 125 NaCl mM, 5.9 mM KCl, 1.2 mM MgSO4, 2.4 mM CaCl2, 25 mM 4-(2-hydroxyethyl)piperazine-1-ethanesulfonic acid (HEPES), pH 7.4, and 1 mg/mL BSA.</p><!><p>Synthesis of FITC-glucagon was performed under conditions similar to those used previously to obtain singly labeled insulin [23]. The labeling reaction was initiated by addition of 0.1 mL of 3 mM FITC dissolved in acetone to 0.9 mL of 0.3 mM glucagon dissolved in 100 mM phosphate buffer, pH 7.5, while stirring. The reaction proceeded unstirred, in the dark, at room temperature for 24 h. HPLC purification (100-μL injection loop) was performed using a Grace-Vydac C4 column (5-μm dp, 250 × 4.6 mm, 214TP54, Columbia, MD) and a linear gradient of 0 – 40 min (20% to 38% B), where A was 0.1% trifluoroacetic acid in deionized water and B was 0.1% trifluoroacetic acid in acetonitrile, while maintaining 1.5 mL/min flow rate. Three major peaks were observed and identified as shown in Electronic Supplementary Material Fig. S1. Free FITC and unlabeled glucagon were identified by comparison of retention time with standards. FITC-labeled glucagon peak was tentatively identified based on its appearance after the reaction and confirmed by CEIA (see results). Minor peaks were due to FITC impurities and present in individual FITC runs. Concentration of FITC-glucagon in purified fractions were determined by UV absorption at 494 nm in pH 8 phosphate buffer based on the molar absorptivity of fluorescein (ε = 68,000).</p><!><p>Fig. 1 illustrates the design of the microfluidic device used in all experiments. Channel dimensions were 6 μm deep, 32 μm wide at the top and 20 μm wide at the base. Fabrication and operation of the glass microfluidic devices has been described previously [9, 12], with the only modification being the expansion of the islet chamber from 300 μm to 1 mm to accommodate batches of islets.</p><p>Microfluidic reservoirs were purchased from Upchurch Scientific (Oak Harbor, WA). ~100-μL reservoirs (Upchurch N-131) were used for all ports except the perfusion inlet, which used a 1/32″ tubing interconnect port (Upchurch N-124H). A thin-film resistive heater (Minco Products, Inc., Fridley, MN) was taped to the bottom of the chip to maintain islets, immunoassay reagents, and the reaction channel at 37 °C, as monitored by digital thermometer (Fisher). The cell chamber was perfused with BSS using a gas-pressure system to drive flow as described previously [12]. The islet reservoir was open at the top to allow perfusion flow to exit (collected by aspiration). Solution in the islet reservoir was sampled by electroosmotic flow (EOF).</p><p>The device utilized a previously described flow gate method of injecting sample onto the separation channel [9, 12, 24]. During the experiment, the islet chamber was held at ground while the reagent reservoirs were maintained at a floating potential to increase sampling efficiency, as described in the text. Voltage was applied to the waste reservoir via a high voltage power supply (Spellman High Voltage Electronics, Hauppauge, NY). The gate reservoir was connected to ground via a high voltage relay (Kilovac, Santa Barbara, CA). When the relay was opened, sample was allowed to load onto the separation channel. The gate was then returned to ground and separation was performed, after which time another sample was loaded for separation. Laser-induced fluorescence detection occurred 1 cm from the injection point.</p><p>Chips were conditioned daily by flowing 1 M NaOH through all channels, followed by deionized water, followed by the experimental solutions for at least 10 min. A chip could be reused after calibration or islet experiments by re-conditioning following the above procedure. During calibration and islet monitoring, 100 μL of the appropriate solution was placed in all reservoirs (excluding the perfusion inlet) and covered with plastic caps that had access holes drilled into them for insertion of Pt electrodes. Immunoassay reagent concentrations used were either 50 nM FITC-glucagon and 25 nM Ab or 200 nM FITC-glucagon and 100 nM Ab (as discussed in the text) dissolved in reagent buffer. For immunoassays, the applied voltage was 6 kV, with an electric field of 700 V cm−1 in the separation channel. Sample injection time was 0.5 s applied at 5.5 s intervals. Islets were perfused at 1 μL/min by applying the appropriate He pressure (typically 50–100 psi) to the perfusion system.</p><p>Electropherograms were analyzed using previously described software [25]. For islet monitoring, concentrations of glucagon were determined from each electropherogram by comparing FITC-glucagon bound to antibody (B) to free FITC-glucagon (F) peak height ratios to a calibration curve. Calibration curves for off-line experiments (mixing off chip) were obtained by mixing of reagents and glucagon standards in vials and pipetting reaction solutions onto chip reservoirs as discussed in the text. Calibration curves for on-line experiments (mixing in chip) were constructed by pumping different concentrations of standard glucagon in the perfusion media (i.e., balanced salt solution) via the perfusion system into the islet reservoir (Fig. 3b). Calibration were fit to a simple logarithmic function using Prism 3.03 (GraphPad Software, Inc., San Diego, CA).</p><!><p>Islets of Langerhans were isolated from CD-1 mice by a previously described method [26]. To monitor secretion, islets were transferred from culture medium in 4 μL aliquots and placed in a Petri dish with 3 mL of the balanced salt solution containing 15 mM glucose. Ten islets were then placed in the cell chamber using a pipette while observing using a stereomicroscope. Islets quickly settled to the glass surface and adhered. Care was taken to ensure that islets were placed in a single layer at the base of the chamber. The device was then transferred to an inverted fluorescence-microscope workstation for monitoring the electrophoresis immunoassay, where the islets were perfused with 15 mM glucose for at least 20 min, during which time glucagon was monitored to establish basal levels of secretion. The glucose concentration was then lowered to 1 mM, and secretion continued to be monitored. All error bars are ±1 standard error of the mean.</p><!><p>Initial experiments were directed towards optimizing the CEIA. We were guided in this work by prior studies which had demonstrated a CEIA of glucagon performed in capillaries [27, 28]. Assay development required preparation and purification of FITC-glucagon. Analysis of the FITC-glucagon reaction mixture by HPLC revealed three main peaks (see Electronic Supplementary Material Fig. S1), two of which corresponded to unreacted glucagon and FITC. The third peak, believed to be FITC-glucagon, was collected and tested for antibody affinity.</p><p>Tests were performed by incubating a mixture of presumed FITC-glucagon (50 nM based on UV absorbance measurements), 25 nM Ab, and different concentrations of glucagon standards in glass vials for 10 min and then analyzing the mixture by CE on the chip in Fig. 1. For this analysis, the premixed samples were added to the islet, FITC-glucagon, and Ab reservoirs and no perfusion flow was used. (In this way, premixed samples are injected onto the electrophoresis channel with no dilution.) Representative electropherograms collected with 0 and 100 nM unlabeled glucagon in the incubation mixture are compared in Fig. 2a. Without glucagon present, two peaks are observed corresponding to free FITC-glucagon and FITC-glucagon-Ab complex. Addition of 100 nM unlabeled glucagon completely eliminated the complex peak (Fig. 2a) by competition to yield a single peak that corresponds to free FITC-glucagon. (Identity of this peak was confirmed by comparison to samples with no Ab present.) These experiments confirmed that the HPLC peak was FITC-glucagon, the FITC-glucagon had affinity for the antibody, and electrophoresis conditions were suitable for separation of complex and free FITC-glucagon.</p><p>A calibration curve generated by this off-line method shows the classic non-linear curve for competitive immunoassays (Fig. 2b). The limit of detection (LOD) for glucagon was 2 nM. LOD was calculated as the concentration necessary to give a bound to free ratio equal to that of the blank minus 3 standard deviations of the blank. These results demonstrate the viability of the reagents to perform the immunoassay.</p><p>These experiments used tricine as the electrophoresis buffer. Initial experiments used HEPES as the electrophoresis buffer because this had been successful for a chip-based CEIA of insulin [9, 12]. Tricine was preferred because HEPES was found to degrade under an applied voltage to yield a fluorescent product that contributed to background. These experiments also used Ab diluted from ascites fluid because this was a low cost source of Ab. Future work may benefit from purification of the Ab used; although we observed no detrimental effects of this preparation.</p><!><p>For on-line mixing, 50 nM FITC-glucagon and 25 nM Ab (both dissolved in reagent buffer) were placed in their respective reservoirs while glucagon (dissolved in balanced salt solution) was placed in the islet reservoir (Fig. 1). The islet, FITC-glucagon, and Ab reservoirs were grounded so that EOF could be used to drive these solutions into the mixing channel. Surprisingly, it was found that under these conditions the on-line CEIA for glucagon had a much lower sensitivity and higher LOD than the off-line assay (compare Figs. 2b and 3b, "grounded"). Subsequent study showed that the cause of this low sensitivity was poor sampling from the islet chamber resulting in excessive dilution of sample by reagent, i.e. too much reagent relative to the sample entered the reaction channel (Fig. 1).</p><p>Poor sampling efficiency was attributed to low EOF in the sampling channel (see Fig. 1) between the islet chamber and the reaction channel. Although the channel lengths from the islet and reagent reservoirs were equivalent, the high ionic strength of the balanced salt solution that was used to dissolve the sample glucagon (and perfuse islets) caused lower EOF from the islet chamber. EOF in this channel was further reduced by the lower potential drop across the sampling channel because of lower electrical resistance due to high conductivity of the balanced salt solution.</p><p>Floating the potential of the reagent reservoirs (FITC-glucagon and Ab) was found to solve this problem. By removing the potential from the reagent channels, the amount of sample entering the reaction channel from the islet reservoir increased by 100-fold. (This increase in sampling was determined by placing fluorophore only in the islet chamber and monitoring the fluorescent signal in the reaction channel). By floating the reagent reservoirs, the majority of flow entering the reaction channel was by electroosmosis from the islet reservoir. Under this condition, reagents enter the reaction channel only by diffusion and possible viscous drag effects.</p><p>With significantly enhanced sampling from the islet reservoir, we investigated the utility for on-line immunoassay. Reaction channel residence time was determined to be increased from 80 s to 120 s when floating the reagent channels. When performing the immunoassay with zero-field in the reagent channels, the reagent concentrations (50 nM FITC-glucagon and 25 nM Ab) were diluted below the fluorescence detection limit. This was not surprising as the reagents could only enter the reaction channel by passive effects. Increasing the FITC-glucagon and Ab concentration to 200 nM and 100 nM respectively allowed sufficient reagent to enter the reaction channel to yield good signal in electropherograms (Fig. 3a). As illustrated in Fig. 3b, the sensitivity of the CEIA was comparable to the off-line assay (i.e., B/F ratio changed approximately 3-fold for a concentration change from 0 to 100 nM) as was the reproducibility (RSD < 5% for all concentrations tested). As a result, the LOD was 1 nM, comparable to the offline assay.</p><p>Using passive addition of reagents to the reaction channel allowed immunoassays and calibration curves with reasonable sensitivity as long as the reagents were present at sufficiently high concentration. In principle other methods could be used to alleviate the problem of unequal EOF from the sample (i.e., islet) and reagent reservoirs. For example, it would be possible to control the potential on each reservoir separately and therefore tune the flow into the reaction channel. Also, the length of the channels could be adjusted so that the field dropped across each channel gave the desired EOF. The approach of using passive reagent addition is simpler than these alternatives because it relies on a single power supply and fixed channel lengths. Flexibility in reagent addition is retained by control of the concentration in the reservoirs. The method described here has the added advantage of minimizing dilution of sample, which is desirable because of the low amount glucagon secreted from islets.</p><!><p>The device was then applied to monitoring glucose-mediated glucagon secretion from islets. When perfused at 1 μL/min with 15 mM glucose to provide minimal secretion, groups of 10 islets secreted 1.6 ± 0.4 pg min−1 islet−1 of glucagon corresponding to ~5 nM. Stepping down to 1 mM glucose elicited a pulse increase in secretion to 3.8 ± 0.6 pg min−1 islet−1 (n = 4) followed by a decay to 1.9 ± 0.4 pg min−1 islet−1 (Fig. 4). Glucagon secretion measured here is similar to, but somewhat higher than, previous studies [29, 30]. For example, one report shows that glucagon secretion was 0.28 pg min−1 islet−1 at 15 mM glucose and 0.45 pg min−1 islet−1 at 1 mM glucose [30]. The higher values observed here could be due to the use of perfusion in this study instead of quiescent solutions in other studies. Paracrine effects play a significant role in glucagon secretion [19, 29, 30]; therefore, the use of perfusion may wash away substances released by other cells in the islet that inhibit glucagon secretion yielding higher values. Higher values may also result from the use of a microfluidic system where losses of hormone are minimized. Finally, since secretion rates are normalized to islet number, use of different islet sizes may account for the differences.</p><p>These results show the value of dynamic measurements on a microscale for glucagon monitoring. The pulse of glucagon secretion seen with the change in glucose is similar to a first phase of insulin secretion and suggests some dynamics of glucagon secretion that have not previously been observed. Therefore, the temporal resolution of measurement offered by the microfluidic system will enable convenient study of the dynamics of glucagon secretion.</p><p>It may eventually be possible to monitor glucagon secretion from single islets, as has been done for insulin [9, 12]; however, in attempts to monitor glucagon secretion from single mouse islets in smaller chambers (300 μm diameter) it was found that glucagon was below detectable levels at all glucose concentrations tested (0, 1, and 15 mM). Possibly further miniaturization of the chambers, to minimize dilution, or use of tighter binding antibodies, to improve sensitivity, will enable detection from single islets.</p><p>It is interesting to compare the CEIA to a conventional ELISA. Commercial ELISAs have LODs of 0.3 to 10 pM (e.g., assays from BioVendor or RayBiotech). This concentration LOD is far lower than the 1 nM achieved here, making them more suitable for serum measurements, where glucagon concentrations are ~50 pM (but highly variable) [31]. In principle it would be possible to incorporate a preconcentration step to improve the concentration LOD of the CEIA. The microfluidic assay has advantages of reduced sample consumption (less than 1 nL per assay compared to 100 μL for ELISA), speed (5 s compared to 6–24 h for ELISA), and reagent cost (a few cents per assay compared to $4 to $12 for ELISA). These advantages make the microfluidic CEIA well-suited for the in vitro research applications which require many assays at small volumes. The experiments illustrated by Figure 4 involved 720 immunoassays which would be cost prohibitive with commercial ELISAs.</p><!><p>A microfluidic system was demonstrated for on-line chemical monitoring of glucagon secretion from living islets of Langerhans. The automated device allowed for real-time evaluation of islet responses with 6 s temporal resolution based on perfusion, reagent addition, and automated CEIA. Initial results reveal dynamics of glucagon secretion not easily studied by other means. The system alleviates many of the disadvantages of traditional immunoassay methods, specifically off-line analysis and manual sample handling. Use of this method should facilitate investigations of the role of glucagon in obesity and diabetes. Further it helps validate and expand the concept of microfluidic CEIA as a tool for monitoring cell function.</p>

PubMed Author Manuscript

Screening for energetic compounds based on 1,3-dinitrohexahydropyrimidine skeleton and 5-various explosopheres: molecular design and computational study

In this paper, twelve 1,3-dinitrohexahydropyrimidine-based energetic compounds were designed by introducing various explosopheres into hexahydropyrimidine skeleton. Their geometric and electronic structures, heats of formation (HOFs), energetic performance, thermal stability and impact sensitivity were discussed. It is found that the incorporation of electron-withdrawing groups (-NO 2 , -NHNO 2 , -N 3 , -CH(NO 2 ) 2 , -CF(NO 2 ) 2 , -C(NO 2 ) 3 ) improves HOFs of the derivatives and all the substituents contribute to enhancing the densities and detonation properties (D, P) of the title compounds. Therein, the substitution of -C(NO 2 ) 3 features the best energetic performance with detonation velocity of 9.40 km s −1 and detonation pressure of 40.20 GPa. An analysis of the bond dissociation energies suggests that N-NO 2 bond may be the initial site in the thermal decompositions for most of the derivatives. Besides, -ONO 2 and -NF 2 derivatives stand out with lower impact sensitivity. Characters with striking detonation properties (D = 8.62 km s −1 , P = 35.08 GPa; D = 8.81 km s −1 , P = 34.88 GPa), good thermal stability, and acceptable impact sensitivity (characteristic height H 50 over 34 cm) lead novel compounds 5,5-difluoramine-1,3-dinitrohexahydropyrimidine (K) and 5-fluoro-1,3,5trinitrohexahydropyrimidine (L) to be very promising energetic materials. This work provides the theoretical molecular design and a reasonable synthetic route of L for further experimental synthesis and testing.In recent years, energy conversion and storage materials have been a hot area of research in materials science. Metal-ion batteries, solar cells, transition metal dichalcogenides and so on, have significantly enhanced our understanding of hydrogen or solar energy [1][2][3] . In contrast to batteries and hydrogen-storage materials, energetic materials (EMs) that can store and release a large quantity of chemical energy stand out due to their excellent combustion efficiency and high energy releasing rate. EMs generally referring to explosives, propellants and pyrotechnics are extensively used for military and civilian purposes [4][5][6] . Modern EMs are expected to have high density, desirable detonation performance and low mechanical sensitivity for safe handling. Recently, hexahydropyrimidine energetic compounds have drawn renewed attention from energetic researchers. With high density, good oxygen balance and low sensitivity, hexahydropyrimidine energetic compounds have displayed potential as energetic additives for high explosives, rocket propellant formulations, and pyrotechnic ingredients 7,8 . 1,3-Dinitrohexahydropyrimidine was firstly synthesized by Bell and Dunstan in 1966 9 and opened up a new field of cyclic 1,3-dinitramine. Some earlier work has concentrated on the synthesis of nitrohexahydropyrimidine energetic compounds,

screening_for_energetic_compounds_based_on_1,3-dinitrohexahydropyrimidine_skeleton_and_5-various_exp

<!>Results and discussion<!>Heats of formation.<!>Compd<!>Potential candidates for HEDMs.<!>Conclusions<!>Computational method

<p>these studies have provided profound insights into the syntheses and thermal decomposition mechanism of hexahydropyrimidine derivatives, there is still lacking comprehensive investigations on explosive performance and systematic molecular design for hexahydropyrimidine compounds. To meet the continuing demand for improved energetic materials, there is a clear need to continue to design and develop new hexahydropyrimidine energetic compounds. This work focus on investigating the most concerned energetic performance, thermal stability and impact sensitivity for 1,3-dinitrohexahydropyrimidine-based energetic compounds that are believed to be candidates of novel EMs. Owing to difficulties in the synthesis of such molecules, theoretical computation is an effective way to design and screen high-energy density compounds.</p><p>For EMs, energy and safety are the two most important properties, because energy determines the effectiveness of application and safety guarantees the application 14,15 . The heat of formation (HOF) is frequently used to indicate the "energy content" of an energetic material [16][17][18][19] . However, due to the sparsity of experimental data and the lack of systematic theoretical study, HOF values for the title compounds are at present unavailable. It is of considerable importance to estimate HOFs through computational methods. Atomization methods and isodesmic reactions are widely applied to figure out HOFs reliably and rapidly 20,21 . Although these methods often generate some significant errors for various frameworks and groups, the errors are sometimes systematic and can be corrected. As to the thermal decomposition process of EMs, the rupture of "trigger linkage" is believed to be a key factor in detonation initiation [22][23][24] . Many researchers believe that C-NO 2 , N-NO 2 and O-NO 2 bonds are trigger spots in nitro compounds. When it comes to impact sensitivity, Politzer 25 proposed that it relates to the bond dissociation energy (BDE) of the trigger bond and molecular electrostatic potential (ESP). Bates 26 suggested that the ability of substitution groups to attract electrons affects the sensitivity of tetrazole: the stronger the ability is, the more sensitive the compound is. Kamlet and Adolph 27 pointed out that the impact sensitivity of some nitro compounds increases with their enhanced oxygen balance. Zhang advanced a method to assess the impact sensitivity of nitro compounds with Mulliken net charges of nitro groups 28,29 . Accordingly, the above-mentioned methods are referenced to analyze the thermal stability and impact sensitivity of hexahydropyrimidine derivatives and we tried to find an optimal standard to evaluate the impact sensitivity of hexahydropyrimidine compounds.</p><p>In this paper, we reported a systematic study of the geometric and electronic structures, HOFs, energetic properties, thermal stabilities and impact sensitivities of a series of 1,3-dinitrohexahydropyrimidine-based derivatives (molecular numbering as A ~ L) as shown in Fig. 1. The compounds were designed based on 1,3-dinitrohexahydropyrimidine skeleton and mono-or di-substituted at C5-position. Our main purpose here is to investigate the important role of different substituents in the design of efficient high energetic materials.</p><!><p>Molecular geometry. The optimized structural data for the title compounds are listed in Table 1. Supplementary Figure S1 displays the optimized geometric structures and atomic numbering.</p><p>As can be seen, the N-N bond lengths in the title compounds range from 1.381 to 1.425 Å, which is shorter than the usual N-N single bond length (1.45 Å). This may be attributed to the hyperconjugation effects between nitro groups and nitrogen atoms. In terms of C-N bonds, some of them in 1,3,5-trinitrohexahydropyrimidine (A), 5-nitrate-1,3-dinitrohexahydropyrimidine (B), 5-fluoramine-1,3-dinitrohexahydro pyrimidine (D), 5,5-difluoramine-1,3-dinitrohexahydropyrimidine (K) and 5-fluoro-1,3,5-trinitrohexahydropyrimidine (L) are longer than normal C-N single bond length (1.472 Å) as a result of cage strain. The substitution of the substituents on hexahydropyrimidine ring should be responsible for some significant difference between the bond lengths 2 lists the calculated HOMO and LUMO energies and the energy gaps (∆E LUMO-HOMO ) for 1,3-dinitrohexahydropyrimidine and its derivatives. When -NO 2 , -ONO 2 , -NHNO 2 , -NF 2 or -N 3 group is attached onto the hexahydropyrimidine ring, both of the HOMO and LUMO energies are higher than those of -CH(NO 2 ) 2 , -CF(NO 2 ) 2 and -C(NO 2 ) 3 derivatives, and those of -N 3 derivative are the highest. Additionally, the disubstituted (attached to C atom) compounds exhibit lower HOMO and LUMO energies than corresponding monosubstituted ones unpredictably. The energy gap between HOMO and LUMO relates the kinetic stability, chemical reactivity, and optical polarizability of a molecule. Some investigations on the excitonic mechanism of detonation initiation show that the pressure inside the impact wave front promotes the HOMO → LUMO transition within a molecule 30,31 . The smaller the energy gap, the easier is HOMO → LUMO electron transfer, the shorter wavelength is required for electronic excitation and the easier is explosive decomposition of the energetic material. Thus, 5-trinitromethyl-1,3-dinitrohexahydropyrimidine (H) may be the most reactive and K the least reactive among these compounds. Moreover, it is interesting to note that the incorporation of -NHNO 2 or -NF 2 group makes a significant increase of ΔE in comparison with the parent compound 1,3-dinitrohexahydropyrimidine (S).</p><p>Electrostatic potential (ESP) is a real and fundamentally significant physical property of compounds as it provides information about charge density distribution and molecular reactivity 5,32 . Hence, the surface electrostatic potential was taken into account in the analysis of electronic properties of the title compounds. The ESP-mapped surfaces of 1,3-dinitrohexahydropyrimidine-based derivatives are presented in Fig. 2. The ESPs are scaled with color. Blue denotes the most negative potential (− 30 kcal mol −1 ) and red the most positive potential (60 kcal mol −1 ). Some pivotal maxima and minima of ESPs are expressed by orange and cyan spheres, respectively. It is shown that the strength and orientation of weak interactions can be well predicted and explained by analyzing the magnitude and positions of minima and maxima on the surface. All the ESP surfaces were created with Multiwfn program and visualized with VMD suite 33 .</p><p>From Fig. 2, it is clear that the strongly positive ESPs distribute in the central regions of hexahydropyrimidine ring and above N-NO 2 and C-R bonds, while the negative ones concentrate on the edges of molecules, especially on the oxygen atoms of nitro groups. The charge density distribution and the magnitude of minima and maxima on the surface are definitely affected by the substituent groups. The overlap of positive (red) and negative potential (blue) displays white region and suggests interactions between electron-withdrawing groups and heterocycle. It is interesting to find that most of the linkages (except F, G and H) between hexahydropyrimidine ring and the substituents are colored in white, indicating that the electronic charges tend to be neutral and there may be expected to exist considerable intra-and inter-molecular interactions between hexahydropyrimidine ring and the substituent groups. Studies show that the molecular electrostatic potential is related to the impact sensitivity of the energetic compounds 25 . The electrostatic potential maxima were summarized in Table 4 and they were used to evaluate the impact sensitivity of the title compounds, as detailed in subsequent text.</p><!><p>In this paper, atomization approach was adopted to estimate the gas phase heats of formation ΔH f (g, 298 K). The energy and the enthalpy data needed during the calculation were obtained at G2 level, which has been shown to accurately predict the heats of formation for a variety of organic compounds 34,35 . The detailed calculation procedure is shown as follows (Fig. 3):</p><p>The gas state HOF of M at 298 K could be obtained as Eq. ( 1):</p><p>(1) where n i stands for the number of atoms of X i in M, H Xi (0 K) stands for the HOF of X i at 0 K, which can be found from the NIST-JANAF tables 36 , and ΔH(0 K) can be derived from H Xi (0 K). (H M (0 K)-H M (298 K)) and (H Xi (0 K)-H Xi (298 K)) represent the enthalpy correction of the molecule (M) and atom (X i ) between 0 and 298 K, respectively. Since the condensed phase for most energetic compounds is solid, the calculation of detonation properties requires solid-phase HOF (ΔH f,solid ). According to Hess's law of constant heat summation, the gas-phase HOF (ΔH f,gas ) and heat of sublimation (ΔH sub ) can be used to evaluate ΔH f,solid 37 : www.nature.com/scientificreports/ Politzer et al. 38 found that the heats of sublimation can correlate well with the molecular surface area and electrostatic interaction index σ tot 2 of energetic compounds. The empirical expression of the approach is as follows:</p><p>where A is the surface area of electronic density of 0.001 electrons/bohr 3 isosurface of the molecule. The coefficients α, β, and γ were determined by Rice et al. 39 : α = 2.670 × 10 -4 kcal/mol/Å 4 , β = 1.650 kcal/mol, and γ = 2.966 kcal/mol. This approach has been proved to be credible for evaluating heats of sublimation of many energetic compounds. Figure 4 illustrates a comparison of HOFs of the title compounds. Supplementary Table S2 summarizes the calculated gas-phase HOFs (ΔH f,gas ), heats of sublimation (ΔH sub ) and solid-phase HOFs (ΔH f,solid ) of 1,3-dinitrohexahydropyrimidine (S) and its derivatives. As can be seen, all the substituted derivatives possess positive HOFs ranging from 152.173 to 568.252 kJ mol −1 . When the substituent is -NO 2 , -NHNO 2 , -N 3 , -CH(NO 2 ) 2 , -CF(NO 2 ) 2 or -C(NO 2 ) 3 group, the HOF value of its substituted 1,3-dinitrohexahydropyrimidine increases significantly compared with the unsubstituted one (S). Therein, the HOF value of -C(NO 2 ) 3 -substituted 1,3-dinitrohexahydropyrimidine (H) is the largest among all the compounds, and this substitution extremely enhances its HOF. However, for the substituent of -ONO 2 or -NF 2 , the case is the complete opposite. In addition, it is found that the HOF value improves with the increasing number of nitro groups in the substituent, indicating a good group additive effect on the HOFs. It is worthy to note that except for -NO 2 group, -N 3 group is also one of the most energetic functional groups known and its substitution can increase the energy content of a molecule greatly. As shown in Fig. 4, all the disubstituted (attached to C atom) compounds exhibit higher HOFs than corresponding monosubstituted ones, implying that increasing the energetic substituents to hexahydropyrimidine ring is favorable to improve the HOF.</p><p>Energetic performance. Density, detonation velocity and detonation pressure are three important parameters reflecting the energetic performance of an energetic material. With a complete neglect of intermolecular interactions within the crystal, conventional M/V approach to calculate density leads to some large error. An improved method by considering the role played by intermolecular forces in the crystal lattice is shown as follows 40 :</p><p>where M is the molecular mass (g mol −1 ), V is the volume of the isolated gas molecules defined as the space inside a counter of electron density of 0.001 e Bohr −3 using a Monte Carlo integration 41 , ν describes the degree of balance between positive and negative potential on the isosurface, and σ tot 2 is a measure of variability of the electrostatic potential on the molecular surface. The values of the coefficients β 1 , β 2 and β 3 are 0.9183, 0.00278 and 0.0443, respectively. We performed 100 single-point calculations for each optimized structure to get an average volume.</p><p>Detonation velocity (D) and pressure (P) can be estimated using empirical Kamlet-Jacobs equations as follows 42 :</p><p>(2) www.nature.com/scientificreports/ where D is the detonation velocity (km s −1 ), P is the detonation pressure (GPa), N is the number moles of gaseous products per gram of explosive, M is the average molecular weight of gaseous detonation products, ρ is crystal density (g cm −3 ) and Q is the detonation energy (cal g −1 ). N, M and Q are determined from the stoichiometric reactions developed for maximum exothermic principle, using arbitrary H 2 O-CO 2 -N 2 decomposition assumption. Due to the explosive and sensitivity nature of the high energy materials, the experimental determination of their Q and ρ are not very frequent, and theoretical approaches have been found to be a viable option.</p><p>Oxygen balance (OB) is used to indicate the degree to which a compound can be oxidized and to classify energetic materials as either oxygen-deficient or oxygen-rich. The higher the oxygen balance is, the larger the detonation velocity and pressure are and the better the performance of the explosive is. For an energetic compound C a H b O c N d , OB(CO 2 ) was calculated by the following equation 43 :</p><p>Table 3 presents the molecular weights (M W ), OB(CO 2 ) and the predicted densities (ρ), heats of detonation (Q), detonation velocities (D) and detonation pressures (P) of 1,3-dinitrohexahydropyrimidine and its derivatives along with cyclotrimethylenetrinitramine (RDX) and cyclotetramethylene tetranitramine (HMX). Figure 5 illustrates a comparison of energetic properties for the title compounds. As shown in Table 3, the calculated density</p><p>The molecular weights (M W ), OB(CO 2 ) and the predicted densities (ρ), heats of detonation (Q), detonation velocities (D) and detonation pressures (P) of the title compounds along with RDX and HMX. a Data in parentheses are from Ref. 10 . b,c Experimental data from Refs. [58][59][60] . www.nature.com/scientificreports/ and detonation velocity of A, RDX and HMX are approximately close to the measured values in literature 10 , demonstrating the reliability of the calculation method to some extent. According to Kamlet-Jacobs equation 42 , ρ is a key factor to influence D and P. Thus, density is one of the most important physical properties for all EMs.</p><!><p>It is found that the introduction of the detonating groups into hexahydropyrimidine ring makes a significant increase of density compared to the unsubstituted compound (S). However, when incorporating -N 3 , there is an unexpectedly little increase of ρ. The possible reason is that -N 3 group contributes a little to the mass of molecule but enhances the volume markedly and reduces the packing regularity on the other hand. One should note that all the disubstituted (attached to C atom) compounds possess higher densities than corresponding monosubstituted ones. Particularly, K with two -NF 2 substitution groups features such good density that even higher than 1.93 g/cm 3 when compared to RDX (1.816 g/cm 3 ) and HMX (1.902 g/cm 3 ).</p><p>As can be seen from Table 3, the monosubstitution of -CH(NO 2 ) 2 , -CF(NO 2 ) 2 or -C(NO 2 ) 3 group and disubstitution of -NO 2 groups (I) enhance OB values of the title compounds significantly, suggesting that OB is greatly influenced by the number of nitro groups. In addition, except for D, E and K, all the compounds have higher heats of detonation than those of RDX and HMX. Therein, H with -C(NO 2 ) 3 group attached to hexahydropyrimidine ring highlights the highest OB of -17.22 and largest Q of 8284.03 kJ kg −1 .</p><p>The calculated detonation velocities of the title compounds lie in the range between 7.87-9.40 km s −1 , and the calculated detonation pressures are between 26.45 GPa and 40.20 GPa. On the whole, -NO 2 , -ONO 2 , -NF 2 , -N 3 , -CH(NO 2 ) 2 , -CF(NO 2 ) 2 and -C(NO 2 ) 3 are effective structural units to improve the detonation performance of 1,3-dinitrohexahydropyrimidine compounds. There is a particular increase in the D and P values when incorporating -CF(NO 2 ) 2 or -C(NO 2 ) 3 group. It is worth noting that the disubstituted compounds own higher D and P values than corresponding monosubstituted ones, suggesting that the increase of the energetic groups on hexahydropyrimidine ring benefits for desirable energetic performance. From Fig. 5, it can be find that the D and P values of G, H and I are very high and close to or above those of RDX. Meanwhile, H stands out among all the compounds with detonation velocity of 9.40 km s −1 and detonation pressure of 40.20 GPa, even higher than those of HMX. Although the heats of formation of D and K are not outstanding, their high densities compensate the disadvantage, proving again that the detonation performance of an energetic material is affected by both density and heat of formation.</p><p>Thermal stability. Bond dissociation energy (BDE) provides useful information for understanding the thermal stability of the title compounds. To elucidate the pyrolysis mechanism and thermal stability of the title compounds, the BDEs of the relatively weak bonds (ring N-NO 2 , C-R or N-R') were calculated and the results were listed in Supplementary Table S3. A comparison of BDEs of the title compounds is displayed in Fig. 6. For monosubstituted compounds (except B and H) and disubstituted compounds J and L, the BDE of ring N-NO 2 bond is much smaller than that of the C-R or N-R' bond, which shows that the ring N-NO 2 bond cleavage is a possible thermal decomposition path for these compounds. It is the O-NO 2 bond of the substituent -ONO 2 group in B and C-NO 2 bond of -C(NO 2 ) 3 group in H that signify the trigger bond in the compounds. Besides, it is found that the introduction of the substituent(s) onto hexahydropyrimidine ring weakens the N-NO 2 bond with decreased BDE values (except B) when compared to S. In principle, all the compounds are energetic materials with BDE > 80 kJ mol −1 and A, C, D and L suffice the stability requirements of high energy density materials with BDE values over 120 kJ mol −1 . On the contrary, the substitution of -ONO 2 , -N 3 , -CH(NO 2 ) 2 , -CF(NO 2 ) 2 , or -C(NO 2 ) 3 group lowers the stability of the compounds. Meanwhile, the BDEs of disubstituted (attached to C atom) hexahydropyrimidine compounds are lower than those of monosubstituted ones, namely increasing the number of energetic substituents is unfavorable with a view to the thermal stability. However, one should note that the BDEs are simply one piece of evidence for molecular thermal stability, the mechanism for the pyrolysis of compounds is mainly linked to their molecular structure. www.nature.com/scientificreports/ Impact sensitivity. The impact sensitivity of the title compounds are analyzed from several aspects: the Mulliken net charges of nitro group (Q NO2 ), electrostatic potential maxima (V s,max ) and characteristic height (H 50 ), and the results are presented in Table 4. Firstly, Zhang 44 pointed out that the more negative charges the nitro group has, the more stable and insensitive the compound is. It can be seen that the incorporation of the substituents at C5-position induces less negative charges of the nitro groups and higher sensitivity than the parent compound (S). It is found that the disubstituted compounds possess less negative charges than corresponding monosubstituted ones, which gives the conclusion that the increase in the number of detonating groups on hexahydropyrimidine ring is unfavorable in view of the impact sensitivity. However, it is encouraging to find that the number of -NF 2 group acts little on the negative charges of nitro group when compared D with K. It is verified from the nitro group charge calculations that -CH(NO 2 ) 2 , -CF(NO 2 ) 2 and -C(NO 2 ) 3 groups are energetic explosopheres that often lead to higher sensitivity. Secondly, electrostatic potential is an important element to take into consideration for analyzing sensitivity. It was first proposed by Politzer et al. 38 , extensively developed by later researchers that the impact sensitivity of explosives has a positive correlation with the surface potential maxima (V s,max ), namely, the impact sensitivity increases with the more positive value of V s,max . As seen from Table 4, the -ONO 2 and -NF 2 derivatives exhibit lower V s,max values than other derivatives, suggesting that -ONO 2 and -NF 2 groups are desirable insensitive groups to construct new molecule frameworks. However, the substitution of other energetic groups leads to positive charge accumulation and the repulsion between these positive potentials elevates the resistance to shear slide, which are detrimental to the stability of the explosives. Except for K with two -NF 2 group, the increase of the substituents does harm to the stability of the title compounds.</p><p>In addition, the characteristic height (H 50 ) was calculated for the title compounds. It is known that the greater H 50 the compound has, the less sensitive the compound is. The calculated H 50 values of RDX and HMX are approximate to the literature measured values, implying that the calculation method is to some extent reliable to predict the H 50 values of nitramine explosives. It can be seen that B, D and K exhibit good stability with H 50 higher than 45 cm, demonstrating that these compounds are impact insensitive to external stimuli relatively. Therein, D owns the highest H 50 value of 50.19 cm and thereby the most stable among the derivatives. The incorporation of -CF(NO 2 ) 2 or -C(NO 2 ) 3 group is in fact detrimental to the stability as predicted above.</p><p>Overall, it is found that the conclusions from Q NO2 , V s,max and H 50 analyses are consistent roughly despite of some discrepancy. -ONO 2 and -NF 2 substituted compounds possess lower impact sensitivity given the above analyses. The impact sensitivity of 1,3-dinitrohexahydropyrimidine compounds rises drastically with the increase of nitro groups. Therefore, one should be cautious in importing nitro groups in molecular design given the stability requirement.</p><!><p>A potential candidate for HEDM should not only have excellent detonation properties, but also could exist stably. Most of hexahydropyrimidine derivatives possess desirable energetic performance, good thermal stability and low impact sensitivity. It is observed that the derivatives B, F, G, H, I, K and L have equivalent or higher detonation performance (D, P) than RDX. Besides, except for F, G, H and I, all the derivatives possess lower impact sensitivity than RDX. One should note that K and L feature high BDEs for the weakest bonds. On the basis of the above suggestions, K and L may be considered as the potential candidates of HEDMs, prompting further investigation in compound synthesis. Chapman et al. 13 have reported a viable route to synthesize K by the reaction of ketone carbonyl groups with difluoramine or difluorosulfamic acid in the presence of a strong acid. A possible synthetic route was designed for L based on Mannich reaction and oxidation reactions as shown in Fig. 7: www.nature.com/scientificreports/ Although A, I, J and K have been successfully synthesized, some detonation and thermodynamics properties are still lacking. In addition, the syntheses of other energetic compounds have not been reported yet. Thus, further investigations are still needed.</p><p>In summary, the introduction of the substituents onto hexahydropyrimidine ring leads to some significant structure deformation and distortion compared to 1,3-dinitrohexahydropyrimidine. -NO 2 , -ONO 2 , -NF 2 , -N 3 , -CH(NO 2 ) 2 , -CF(NO 2 ) 2 and -C(NO 2 ) 3 are effective structural units to improve the detonation performance of 1,3-dinitrohexahydropyrimidine-based compounds. As mentioned above, the molecule structure is a decisive factor for the stability of a nitramine explosive. For most of the title compounds, N-NO 2 bonds possess the least bond dissociation energy and may be the trigger bonds in these compounds. The substitution of -NHNO 2 , -NF 2 or -F shows relatively higher BDE value and better thermal stability among all the derivatives. The impact sensitivity of the title compounds increases with the increasing number of nitro groups in molecules. The Mulliken net charges of nitro group (Q NO2 ), electrostatic potential maxima (V s,max ) and characteristic height (H 50 ) are significant parameters to evaluate the sensitivity of nitramine explosives. As the factors affecting the impact sensitivity are complicated, these elements should be considered as more to predict the impact sensitivity of an energetic material comprehensively.</p><!><p>In the present study, the geometric and electronic structures, HOFs, energetic properties, thermal stability and impact sensitivity of 1,3-dinitrohexahydropyrimidine derivatives were analyzed. The results show that the substitution of -N 3 , -NO 2 , -NHNO 2 , -CH(NO 2 ) 2 , -CF(NO 2 ) 2 or -C(NO 2 ) 3 group contributes positively to the HOFs of the derivatives. Besides, these groups together with -ONO 2 and -NF 2 are effective structural units to improve the detonation performance. Therein, H features the highest HOF and best detonation performance way above HMX.</p><p>The calculated frontier orbital energy shows that H may be the most reactive and K the least reactive among these compounds. An analysis of the bond dissociation energies for the weakest bonds reveals that for most 1,3-dinitrohexahydropyrimidine-based compounds, the ring N-NO 2 bond cleavage is a possible thermal decomposition path for these compounds.</p><p>It is found that -ONO 2 and -NF 2 are effective energetic groups to construct impact insensitive frameworks. Therein, D owns the highest H 50 value of 50.19 cm. Taking into consideration these results, it is regarded that K and L may be the potential candidates of HEDMs with powerful energetic performance and low sensitivity. It is anticipated that these findings will enhance the future prospects for rational energetic materials design.</p><!><p>DFT-B3LYP has been shown to accurately predict the structural parameters and perform frequency calculations on systems containing C, H, O, N and other elements in the study of energetic materials according to a series of studies by Xiao Heming et al. [45][46][47] . The geometry optimization and frequency analysis of the title compounds were fully performed at B3LYP/6-311G(d,p) level using density functional theory (DFT) embedded in Gaussian 09 program 48 . The vibration frequency and infrared spectra (IR) analyses (Supplementary Figure S2) showed that none of the optimized structures exhibited imaginary frequencies. When all the optimized structures corresponded to the local energy minimum points on the respective potential energy surfaces, the stable structures were obtained. Quantitative analysis of molecular van der Waals (vdW) surface was carried out with Multiwfn program. The 0.001 isosurface of electron density was regarded as vdW surface, since this definition reflects specific electron structure features of a molecule, such as lone pairs, π electrons etc., this is also what the definition used in our analyses. The analysis of ESP on vdW surface has been further quantified to extract more information. Researchers have defined a set of molecular descriptors based on ESP on vdW surface, which are taken as independent variables of general interaction properties function (GIPF) 38,40,49 . GIPF successfully connects distribution of ESP on vdW surface and many condensed phase properties, including density, heat of sublimation, impact sensitivity and so on, which are detailed in the text. Besides, the HOF and BDE were calculated and www.nature.com/scientificreports/ discussed based on the optimized structure. The detonation properties were further estimated based on calculated density and HOF according to Kamlet-Jacobs equations. The strength of bonding, which could be evaluated by bond dissociation energy (BDE), is fundamental to understand pyrolysis mechanism and the thermal stability of the compound 50,51 . The BDE of the molecule corresponding to the enthalpy of reaction A-B(g) → A·(g) + B·(g), which is required for homolytic bond cleavage at 0 K and 1 atm, was calculated in terms of Eq. ( 8) 52 :</p><p>where E 0 is the total electronic energy of the species calculated at B3LYP/6-311G(d,p) level. One should note that DFT/B3LYP method has been found to reasonably predict accurate BDEs.</p><p>The bond dissociation energy with zero-point energy (ZPE) correction can be determined by Eq. ( 9) 18 :</p><p>where ΔE ZPE is the difference between ZPEs of the products and reactants. The Mulliken net charges of nitro group (Q NO2 ) is calculated by the sum of atomic charges on the nitrogen and oxygen atoms of the nitro group as follows 53,54 :</p><p>Experimentally, impact sensitivity is characterized through a drop weight test. The drop height (H 50 ) is defined as the height from which there is a 50% probability of initiating explosion 55,56 . Pospíšil et al. have proposed an empirical formula relating H 50 to the electrostatic potential of an energetic molecule, as given by the following equation 57 :</p><p>where ν describes the degree of balance between positive and negative potential on the isosurface, and σ + is the electrostatic potential for the positive charge. The values of the coefficients α 2 , β 2 and γ 2 are taken from Ref. 57 .</p>

Scientific Reports - Nature

Inhibition of the M. tuberculosis 3\xce\xb2-Hydroxysteroid Dehydrogenase by Azasteroids

The cholesterol metabolism pathway in Mycobacterium tuberculosis (M. tb) is a potential source of energy as well as secondary metabolite production that is important for survival of M. tb in the host macrophage. Oxidation and isomerization of 3\xce\xb2-hydroxysterols to 4-en-3-ones is requisite for sterol metabolism and the reaction is catalyzed by 3\xce\xb2-hydroxysteroid dehydrogenase (Rv1106c). Three series of 6-azasteroids and 4-azasteroids were employed to define the substrate preferences of M. tb 3\xce\xb2-hydroxysteroid dehydrogenase. 6-Azasteroids with large, hydrophobic side chains at the C17 position are the most effective inhibitors. Substitutions at C1, C2, C4 and N6 were poorly tolerated. Our structure-activity studies indicate that the 6-aza version of cholesterol is the best and tightest binding competitive inhibitor (Ki = 100 nM) of the steroid substrate and are consistent with cholesterol being the preferred substrate of M. tb 3\xce\xb2-hydroxysteroid dehydrogenase.

inhibition_of_the_m._tuberculosis_3\xce\xb2-hydroxysteroid_dehydrogenase_by_azasteroids

<!>Materials and Methods

<p>The cholesterol metabolism pathway of Mycobacterium tuberculosis (M. tb2), the causative agent of tuberculosis, has recently attracted interest because of the dual possibilities that this pathway may represent a major source of energy production as well as secondary metabolite production with complex molecular structures within the macrophage.1, 2 Blocking this pathway with small molecule inhibitors presents an attractive avenue for developing new anti-tubercular therapies that are specific to the survival of the organism in the host. 3β-hydroxysteroid dehydrogenase (Rv1106c) is proposed to catalyze the first step of the cholesterol metabolism pathway (Scheme 1).3 However, the physiological substrate specificity has not been fully investigated. Herein, we employ the 6-azasteroid series of inhibitors as a tool set to determine the structure-activity relationship for M. tb 3β-hydroxysteroid dehydrogenase.</p><p>3β-hydroxysteroid dehydrogenases (3β-HSD) catalyze the oxidation and isomerization of Δ5-3β-hydroxysteroids to Δ4-ketosteroids. In mammals, this enzyme is required for the biosynthesis of steroid hormones, and pregnenolone, 17-hydroxypregnenolone, dehydroepiandrosterone, and androst-5-ene-3,7-diol are all substrates of mammalian 3β-HSDs.11 3β-HSD orthologs have been identified in plants, fish, amphibians, viruses, and actinomycete bacteria.12–17 Their respective substrate specificities and metabolic functions are not as well established.</p><p>Notably, in M. tb, the enzyme is proposed to catalyze the first step in cholesterol degradation because the 3β-hydroxysteroid dehydrogenase is required for conversion of cholesterol to cholest-4-en-3-one in cell lysates. However, M. tb 3β-HSD also can oxidize and isomerize dehydroepiandrosterone and pregnenolone to their respective α,β unsaturated ketones with equal efficiency.3 Direct comparison of substrate specificities is difficult in this system because the conditions employed to solubilize the steroids differ, and substrate inhibition by the NAD+ cofactor is observed at millimolar concentrations. Therefore, the relative binding affinities for the enzyme are not readily derived from the kinetic experiments.</p><p>As a consequence, the precise sequence of catalytic events in the M. tb cholesterol metabolism pathway is not known. In the predicted pathway for M. tb cholesterol metabolism, the steroid skeleton can undergo oxidative degradation simultaneous with side-chain truncation (Scheme 1). The substrate preference of each enzyme in the pathway has been explored to a limited extent with the exception of 3-ketosteroid dehydrogenase (KstD) which is suggested to prefer the 5α-androstane-3,17-dione and 5α-testosterone as substrates,4 and the meta-cleavage product hydrolase HsaD for which X-ray crystal structures suggest a preference for 4,5–9,10-diseco-3-hydroxy-5,9,17-tri-oxoandrosta-1(10),2-diene-4-oic acid as the substrate.7 In order to develop antibacterial therapies that target the pathway and to understand the metabolic consequence of blocking enzymes within the pathway, the preferred sequence of reactions must be elucidated.</p><p>Azasteroids (Scheme 2a) are validated, therapeutically-useful compounds that inhibit enzymes in steroid biosynthetic pathways. For example, finasteride and dutasteride inhibit 5α-reductase-catalyzed production of dihydrotestosterone from testosterone. Inhibition of 5α-reductase is indicated for treatment of benign prostatic hyperplasia and some prostate cancers. Investigation of the 6-azasteroid series was undertaken by Frye and coworkers at GlaxoWellcome during SAR work to develop enzyme selectivity.18 However, cross-reactivity with human 3β-HSD could not be eliminated and development efforts were refocused on the 4-azasteroid series.19–21</p><p>M. tb 3β-HSD shares 29% amino acid sequence identity with type I and type II human 3β-HSD (UniProtKB ID P14060 and P26439) and these enzymes catalyze the same reaction. Both the active site catalytic triad S131, Y158, K162 and Rossman fold motif for NAD+ cofactor binding are conserved. As the 6-azasteroid moiety is proposed to act as a transition state mimic of the 3β-HSD-catalyzed reaction,19 we expected that the transition state analogy would apply to the M. tb enzyme (Scheme 2b). Given the low amino acid identity of non-catalytic residues between orthologs, we did not expect inhibitor specificity of the M. tb enzyme to necessarily parallel that of the human enzyme.</p><p>We reasoned that a comprehensive study of 6-azasteroids would provide rapid entry into the structure-activity relationship (SAR) of M. tb 3β-HSD and insight into the true substrate for the enzyme in vivo. Moreover, 6-azasteroids have excellent biodistribution and pharmacokinetic properties in humans.20 Inhibitors of M. tb 3β-HSD are important for targeting the cholesterol metabolic pathway and would require little development before in vivo analysis of enzyme inhibition could be undertaken. Here we report the in vitro inhibition SAR for M. tb 3β-HSD using a family of azasteroids to explore the enzyme specificity.</p><p>Three series of azasteroids were tested to survey the importance and tolerance of substituents at a) the 17-position of the D-ring, b) the 4–7-positions of the A- and B-rings, and c) the 1,2 positions of the A-ring (Table 1). In order to identify the most potent compounds, the IC50's for 21 different azasteroids were measured at the KM of dehydroepiandrosterone (120 μM) and at 2×KM of NAD+ (400 μM). Previously, we had demonstrated that M. tb 3β-HSD follows a compulsory order mechanism in which NAD+ binds first.3 Therefore, we expected that competitive inhibitors of steroid binding would bind to the E-NAD+ complex. However, inhibitors were tested with a less than fully saturating concentration of NAD+ because substrate inhibition occurs at millimolar levels of the cofactor.3 The IC50's were determined at 8 different inhibitor concentrations ranging from 6 nM to 400 μM. The maximum concentration that was used ranged from 50–400 μM due to the limited solubility of some of the azasteroids.</p><p>The mechanism of inhibition was determined for azasteroids 3, 7, and 17, which had IC50's that varied over 2 orders of magnitude. We measured steady-state rates as a function of both DHEA and inhibitor concentrations and globally fit the data to equation (2). All three inhibitors were found to be competitive inhibitors of DHEA. We concluded that modifications of the steroid ring framework that reduced the efficacy of inhibition did not alter the mechanism of inhibition. Kinetic competition with DHEA is consistent with the proposal that the 6-azasteroid binds in the steroid binding site and that the IC50's serve as a valid indicator of relative binding to the steroid binding site.</p><p>Functional groups of varying polarity were employed to interrogate the importance of the 6-azasteroid side chain at C17. Carboxylic acid 1, did not inhibit M. tb 3β-HSD at concentrations below 200 μM. In contrast, incorporation of an amide at the 17-carbon was tolerated. Aromatic amide substituents, e.g., 3–5, were effective inhibitors, whereas, a polar heterocycle 2 obviated efficacy. Further constraint of the amide as oxazole 6 improved inhibition a further 3-fold. Elimination of heteroatoms at C17 and substitution of the 6-azasteroid with the native cholesterol 8-carbon side chain to provide azasteroid 7 increased potency an order of magnitude compared to the oxazole. The binding preference for large hydrophobic substituents indicates that C27 sterols are most likely the preferred substrate for M. tb 3β-HSD.</p><p>Addition of a methyl group at C7 to the sterol framework of 7 to provide 8 reduced potency 10-fold. Next, we explored structure-activity space around the A,B ring system using a simple amide substituent at the C17 position. When the C17 was modified as the t-butyl amide, 9, the inhibition potency was poorer than the aromatic substituents. The diethyl amide substituent was selected for further analysis due to its facile synthesis and to compare inhibition constants in a readily measured concentration range. Substitution of the 6-azasteroid framework with linear alkyl moieties at either C4 or N6, 10–15, reduced potency, and in most cases also reduced solubility. Alkylation at both C4 and N6 in a bridged fashion, 16, restored potency. Interestingly, swapping N6 and C4 to provide the 4-azasteroids 17 (finasteride) and 18 (dutasteride) had a deleterious effect on potency, which is consistent with the 6-azasteroid framework mimicking the intermediate in the isomerization half-reaction (Scheme 2).</p><p>Introduction of unsaturation at C1-C2 in azasteroid 9 was tolerated. However, cyclopropanation across C1-C2 in azasteroids 19 and 20 reduced efficacy at least 5-fold. Therefore, the steric requirements about the A,B ring system are quite stringent for the M. tb 3β-HSD.</p><p>This 6-azasteroid series are potent inhibitors against human adrenal 3β-HSD. Increased bulk at C17, e.g., trifluoromethyl aryl substituents like in azasteroid 3, decreased potency.21 Moreover, the human adrenal 3β-HSD tolerated methylation at N6 and cyclopropanation at C1-C2.19 However, larger N6 substituents, e.g, butyl and hexyl reduced potency against the human enzyme 1000-fold.19 Therefore, future work to develop M. tb specific inhibitors should focus on elaboration of the C17 structure-activity space and introduction of constrained substituents between N6 and C4.</p><p>In our screen of 20 compounds, azasteroid 7, which most closely mimics cholestenone, was the most effective inhibitor. Further inhibition analyses were carried out with azasteroid 7 in which the concentrations of both DHEA and NAD+ were varied. The azasteroid competitively inhibits DHEA binding ten times more effectively than NAD+ binding (Table 2, Figures S1 and S2). Moreover, the azasteroid shows strong uncompetitive inhibition with respect to NAD+. Previously, we demonstrated that catalysis by M. tb 3β-HSD follows a compulsory order binding mechanism in which NAD+ binds first.3 The competitive behavior versus DHEA and uncompetitive behavior versus NAD+ suggests that the azasteroid binds most effectively to the E-NAD+ complex. Therefore, increasing levels of NAD+ cofactor cannot overcome inhibition of 3β-HSD by azasteroids.</p><p>In summary, three series of azasteroids were evaluated for binding and inhibition of M. tb 3β-hydroxysteroid dehydrogenase. 6-Azasteroids with large, hydrophobic side chains at the C17 position are the most effective inhibitors. Substitutions at C1, C2, C4 and N6 were poorly tolerated. Our structure-activity studies indicate that the 6-aza version of cholesterol is the best and tightest binding inhibitor and are consistent with cholesterol being the preferred substrate of M. tb 3β-hydroxysteroid dehydrogenase.</p><!><p>6-azasteroids were provided by GlaxoSmithKline (Research Triangle Park, NC). Frye and coworkers describe the synthesis and purity confirmation by elemental analysis of the 6-azasteroids.19, 20, 22 4-azasteroids were purchased from Sigma-Aldrich. The purity (>98%) and identity of azasteroids were confirmed by LC/UV/MS (Waters UPLC/diode array/SQD) for use in these assays. The LC was performed with a 1.7 μM C18 (2.1 × 100 mm) column and the gradient was from 100% H2O to 100% methanol over 10 minutes. Hazard warning: these 6-azasteroids are potential teratogens to the developing male fetus and appropriate precautions should be taken when handling these compounds.</p><p>Purification and assay of 3β-HSD was carried out as previously described.3 Briefly, reactions were initiated by the addition of 3β-HSD (125 nM) to substrates DHEA (120 μM) and NAD+ (400 μM), with or without inhibitor, in 100 mM TAPS pH 8.5 buffer with 150 mM NaCl2 and 30 mM MgCl2. Inhibitor stock solutions were prepared in DMSO. The final DMSO concentration was held fixed at 2% which does not affect enzyme activity.3 The formation of NADH was monitored at 340 nm for the first 150 s of reaction at 30 °C. Assays were performed in duplicate and IC50 values were determined from 8 concentrations by fitting to equation (1). Mechanism of inhibition was determined by fitting initial velocity data to equation (2) using Grafit 4.0 (Erithacus software, Sussex UK).</p><p> (1)v=Vm/(1+[I]/IC50) (2)v=Vm[S]/{Km(1+[I]/Kic)+[S](1+[I]/Kiu)} where v is the initial velocity, Vm is the maximum velocity, S is the varied substrate, KM is the Michaelis-Menten constant for the varied substrate, and Kic is the competitive inhibition constant and Kiu is the uncompetitive inhibition constant.</p>

PubMed Author Manuscript

Preparation of Titanocene-Gold Compounds Based on Highly Active Gold(I)-N-Heterocyclic Carbene Anticancer Agents. Preliminary in vitro Studies in Renal and Prostate Cancer Cell lines.

Heterometallic titanocene-based compounds containing gold(I)-phosphane fragments have been extremely successful against renal cancer in vitro and in vivo. The exchange of phosphane by N-heterocyclic carbene ligands to improve or modulate their pharmacological profile afforded bimetallic complexes effective in prostate cancer but less effective in renal cancer in vitro. We report here on the synthesis of new bimetallic Ti-Au compounds by incorporation of two highly active gold(I)-N-heterocyclic carbene fragments previously reported derived from 4,5-diarylimidazoles. The two new compounds [(\xce\xb75-C5H5)2TiMe(\xce\xbc-mba)Au(NHC)] (NHC = 1,3-Dibenzyl-4,5-diphenylimidazol-2-ylidene NHC-Bn 2a; 1,3-Diethyl-4,5-diphenylimidazol-2-ylidene NHC-Et 2b) with the dual linker (-OC(O)-p-C6H4-S-) containing both a carboxylate and a thiolate group were evaluated in vitro against renal, and prostate cancer cell lines. The compounds were more cytotoxic than previously described Ti-Au compounds containing non-optimized gold(I)-N-heterocyclic fragments. We present studies to evaluate their effects on cell death pathways, migration, inhibition of thioredoxin reductase (TrRx) and vascular endothelial growth factor (VEGF) in prostate PC3 cancer cell lines. The results support that the incorporation of a second metallic fragment like titanocene to biologically active gold(I) compounds improves their pharmacological profile.

preparation_of_titanocene-gold_compounds_based_on_highly_active_gold(i)-n-heterocyclic_carbene_antic

Introduction<!>Synthesis, characterization and stability studies<!>Cytotoxicity, selectivity and cell death<!>Inhibition of migration by selected compounds<!>Inhibition of Thioredoxin Reductase (TrxR) and Vascular Endothelial Growth Factor (VEGF) by selected compounds<!>Conclusions<!>General information and instrumentation for synthesis, characterization and stability studies of the new compounds<!>Synthesis and characterization<!>General Procedure for [Au(Hmba)(NHC)] (1a, 1b).<!>General Procedure for [(\xce\xb75-C5H5)2TiMe(\xce\xbc-mba)Au(NHC)] (2a,2b).<!>DFT calculations<!>Crystal structure determination<!>Cell lines<!>Cell viability analysis<!>Cell death assay<!>Cell migration analysis<!>Immunohistochemistry<!>Analysis of cell TrxR, VEGF, Phalloidin and DAPI.

<p>Metal-based drugs have experienced a resurgence as prospective cancer chemotherapeutics with a growing number of compounds active against tumors in vivo and some in current clinical trials.[1, 2] Current efforts have been focused in the use on non-conventional metallodrugs (those with a mode of action different from that of cisplatin)[3, 4] as well as in understanding their mode of action,[5] their effects in the immune system[6] and in developing agents that can be photoactivated[7] or delivery systems[8] that may improve their pharmacological profiles. The potential of heterometallic compounds (compounds with two or more different metals) in cancer therapy has been recently highlighted[9]. The hypothesis (first described by Casini and co-workers)[10] is that the incorporation of two different biologically active metals in the same molecule may improve their antitumor activity as a result of metal specific interactions with distinct biological targets (cooperative effect) or by the improved chemicophysical properties of the resulting heterometallic compound (synergism).[16]</p><p>Our group at Brooklyn College has described a variety of compounds based on gold(I) biologically active fragments (either containing phosphanes PR3 or N-heterocyclic carbenes NHC and a second metallic fragment (based on titanocene [TiCp2][11–16] or arene ruthenium(II) [Ru(p-cymene)Cl2(dppm)][17–19] derivatives). We unveiled the potential of these compounds as chemotherapeutics against renal, colorectal and prostate cancers (including mechanistic and in vivo studies). Second generation titanocene-based compounds (SG1 and SG2 in figure 1) based on the bifunctional ligand mba = -OC(O)-p-C6H4-S-(derived from 4-mercaptobenzoic acid H2mba) resulted extremely efficacious in human clear cell renal carcinoma Caki-1 cells[14,15] and xenograft mice[14] models. We demonstrated that these bimetallic compounds were more potent and affected a broader spectrum of molecular targets and cellular behaviors than any single isolate monometallic derivative.[15]</p><p>Third generation titanocene-based compounds (TG1–4)[16] containing [Au(NHC)] fragments (instead of [Au(PR3)]) resulted much less cytotoxic (500–100 times less efficient) on the renal cancer cells than the second generation compounds SG1–2. They however displayed relevant in vitro cytotoxic and apoptotic behavior and antimigratory properties in human prostate cancer (PC3) cell lines.[16]</p><p>We report here on the preparation of modified third generation titanocene-gold bimetallic compounds of the type [(η5-C5H5)2TiMe(μ-mba)Au(NHC)] (2a and 2b in Scheme 1) based on two highly active gold(I) compounds containing N-heterocyclic carbenes ([AuX(NHC)]; NHC = 1,3-Dibenzyl-4,5-diphenylimidazol-2-ylidene NHC-Bn; X = Cl a; 1,3-Diethyl-4,5-diphenylimidazol-2-ylidene NHC-Et; X = Br b) already described.[20, 21] Both gold(I)-NHC compounds had been synthesized on the basis of relevant pharmacological properties of 4,5-diarylimidazoles[22] and the well-known antitumor properties displayed by gold(I) compounds containing either phosphanes[23, 24] or N-heterocyclic carbenes.[25, 26] Compound a (described by Tacke and co-workers[20, 26, 27]) and a variation with the 2',3',4',6'-tetra-O-acetyl-β-D-glucopyranosyl-1'-thiolate showed very good activity against a wide range of human cancer cell lines from the NCI 60 cell line panel, and relevant tumor growth inhibition in vivo for a human clear cell renal carcinoma Caki-1 xenograft mice model. More recently compound a has been bioconjugated to an engineered antibody (Thiomab LC-V2050) via cysteine conjugation. The anti-proliferative activity of this ADC in HER2 positive breast cancer cell line showed a promising moderate improvement as compared to the gold-complex drug.[28] Compound b, described by Gust and co-workers in 2011, was found to be cytotoxic (low or sub-micromolar range) in breast and colon cancer cell lines.[21]</p><p>We hypothesized that the incorporation of these highly cytotoxic gold(I) compounds and titanocene in the same molecule would improve their pharmacological profile. In addition to the synthesis, characterization and study of the stability of the new titanocene-gold compounds (and their monometallic precursors containing the mba linker, 1a and 1b in Scheme 1), we report on their cytotoxicity in human renal (Caki-1) and prostate (PC3) cancer cell lines, and their selectivity. We also report on the type of cell death induction, anti-migratory properties and inhibitory properties of TrRx and VEGF in prostate PC3 cancer cell lines for bimetallic compound 2a and precursor 1a. We include comparisons with the previously described[19] [AuX(NHC)] compound a.</p><!><p>The new compounds [(η5-C5H5)2TiMe(μ-mba)Au(NHC)] (NHC-Bn 2a; NHC-Et 2b in Scheme 1) are synthesized by reaction of precursors [Au(Hmba)(NHC)] (NHC-Bn 1a; NHC-Et 1b) and [(η-C5H5)2TiMe2] (Equation 1) via a procedure already reported.[14–16] All new compounds were obtained as white (1a-b) or yellow (2a-2b) solids. The synthesis and characterization details are provided in the experimental section and NMR, IR and ESI-MS-HR spectra are provided in the SI. The structures for the bimetallic compounds in Scheme 1 are proposed on the basis of analytical and spectroscopic data and by comparison with structurally-related compounds.[14–15] Compounds 2a and 2b are stable in solid state in air and at 5°C for months. ESI -MS-HR spectra shows the parent peak for compound 2b as well as a peak (S17 and S18) for trimetallic species containing the Ti(η5-C5H5)2Me(μ-mba) scaffold and two [Au(NHC-Et)] fragments (Figure S17). The identification of mutimetallic species by MS spectrometry (due to coordination of more than one [AuL] fragments to sulfur atoms under these conditions) is well known.[29] The ESI -MS-HR spectra for 2a only shows the trimetallic species (S15 and S16).</p><p>The diagnostic technique to assess the type of coordination of the carboxylate group from the mba ligand is IR spectroscopy. On the basis of the difference between symmetric and antisymmetric stretching bands in the solid state IR spectra of compounds 2a and 2b it is safe to propose a monodentate bonding for the carboxylate[30.31] as previously described by our group for both phosphine and NHC ligands, which was corroborated by DFT calculations.[14–16] In addition, we have performed more detailed DFT calculations in order to identify the coordination mode of the carboxylic group in bimetallic complexes 2a and 2b (see Figure 2 and in the SI, Table S1 and Figures S29 and S30). The relative position of symmetric and asymmetric stretching bands of the carboxylate groups can distinctly characterize monodentate versus bidentate complexes. Both bands are observed in the experimental IR spectra; the symmetric strong band at 1284 cm−1 and the asymmetric moderate band at 1635 cm−1 for both 2a and 2b complexes. The observed difference of about 350 cm−1 indicates the monodentate coordination for these complexes. Moreover, the monodentate coordination mode was confirmed by DFT calculations where both coordination modes were modeled and compared to experimental spectra (Figures S29 and S30). For monodentate complexes the intense symmetric stretching band was computed around 1288 cm−1, and the asymmetric bands at 1690 cm−1 that agrees with experimental bands and confirms the coordination mode. In contrast, for bidentate complexes the intense symmetric stretching band computed at 1456 cm−1 and the asymmetric band around 1523 cm−1 it is not in accordance with experimental spectra and excludes the bidentate coordination mode for these complexes. The most relevant distances calculated are collected in table S1 (SI).</p><p>Unfortunately, we could not obtain crystals of enough quality for bimetallic compounds 2a and 2b for X-ray diffraction studies.</p><p>The crystal structures of the monometallic precursors 1a and 1b were determined by X-ray (Figure 3 and in SI, Figures S31–S33 and Tables S2 and S3). These structures are relevant due to the biological activity found for these precursors (see biological activity section). We have previously reported on the crystallography of compounds containing the mba linker and phosphane or N-heterocyclic carbene ligands.[14–16] The structure of 1a is depicted in Figure 3A. Its monomeric structure is almost identical to that obtained with a different NHC (ligand 3 in Figure 1) and reported by us recently.[16] The structure of 1b is included in the SI information (S32) as it is extremely similar as well. For 1a the environment of the gold atoms is close to linear [C-Au-S 177.93(5)°] (Figure 3). A crystal structure of an analogue of compound a ((1,3-Di(p-methoxybenzyl)4,5-di(p-isopropyl phenyl)imidazol-2-ylidene)gold(I) chloride) had been already reported by Tacke et al.[20] The distance Au-C(1) of 2.0129(19) Å for 1a is close to that of compound a analogue of 1.988(3) Å. The angle of C(1)-Au-Cl(1) for the Tacke analogue is 180° (linear as in the case of 1a). For 1b (Figure S32 in SI) the angle is 174.33(8) (still close to linearity) and the distortion most plausibly due to the supramolecular arrangement (Figure S33). A crystal structure of an analogue of b ([1,3-Diethyl-4,5-bis(4-fluorophenyl)imidazol-2ylidene]gold(I) bromide) was also reported by Gust et al.[21] In this case the distance Au(1)−C(1) was 1.988(9)Å and the angle C(1)−Au(1)−Br(1) 178.3(2)°, again within the same range as for 1b where the C(1)-Au(1) distance is 2.010(3) Å and the angle a little shorter (174.33(8)°).</p><p>The most relevant characteristic of these compounds in solid state, is their ability to form supramolecular species via hydrogen bonding and in some specific cases via gold-gold interactions. In the case of compound 1a (Figures 3 and S31) the monomers organize into chains that are linked by short O-H bonds (1.782 Å). There are additional interactions between monomers and CHCl3 solvent molecules (interactions between the sulfur atom and one H from the CHCl3 molecule of 2.534 Å) as well as weaker interactions between the sulfur atom in one monomer and one hydrogen from a phenyl group in another monomer (2.874 Å). In the case of compound 1b (Figure S33) there are also short O-H bonds (1.871 Å) linking two different monomeric units and longer interactions between an oxygen from a monomer and one hydrogen from a phenyl group from a different monomer (2.589 Å).In these structures there are no appreciable gold-gold interactions. The shorter Au-Au distance observed for Au···Au 3.659 Å (1a) is a little above the sum of the van der Waals radii of ca. 3.6 Å. The only compound that displayed monomers linked by gold-gold bonds, [(η5-C5H5)2TiMe(μ-mba)Au(PEt3)] (SG2 in figure 1) has been reported recently.[16]</p><p>The stability of compounds 2a and 2b was evaluated by 1H NMR spectroscopy in DMSO and DMSO/PBS (5:1) and by mass spectrometry over time (see SI). NMR experiments were performed in DMSO-d6 and in mixtures of DMSO-d6/PBS-D2O. The stability study of compounds 2a and 2b by H NMR in DMSO-d6 showed half-life values of 18 and 6 hours, respectively, considerably longer than former derivatives containing other NHC (third generation TG1–4 derivatives in Figure 1) with half-lives ranging from 1 to 3 hours. The half-lives of 2a and 2b are also longer or in the same range as second generation compounds SG1–2 (Figure 1). Mass spectrometry further supports the presence of species containing both titanium and gold in 1% DMSO/PBS solution after 24 hours (see SI). We have shown in the past that the cyclopentadienyl ligands are dissociated over time, something we also observe for 2a and 2b.[14–16] In general, for titanocene-gold compounds we have been able to prove the co-localization of both metals (titanium and gold) both in cancer cells[14–16] and in tumors.[32] This fact indicates that the bimetallic compounds are pro-drugs that decompose into biological active species still containing a Ti-Au core.[32]</p><!><p>The cytotoxicity of the bimetallic compounds [(η5-C5H5)2TiMe(μ-mba)Au(NHC)] (NHC = NHC-Bn 2a, NHC-Et 2b), monometallic gold precursors [Au(Hmba)(NCH)] (NHC = NHC-Bn 1a, NHC-Et 1b), the gold cytotoxic compounds ([AuX(NHC)]; NHC = NHC-Bn; X = Cl a "Tacke"; NHC-Et; X = Br b "Gust") already described,[11,12] and monometallic titanocene Y was evaluated. Titanocene Y (a compound described by the group of Tacke[33]) is considered a good benchmark for titanocenes due to its high activity in breast,[34,35] and renal[36] cancer in vitro and in vivo. For comparative purposes, the cytotoxic profile of cisplatin and Auranofin was also determined. In this assay, human clear-cell renal carcinoma Caki-1, human prostate PC3 cells and non-tumorigenic human fetal lung fibroblasts (IRM-90) were incubated with the above described compounds for 72 hours. The compounds were assayed by monitoring their ability to inhibit cell growth using the PrestoBlue™ Cell Viability assay (see Experimental Section). The results are summarized in Table 1.</p><p>The heterometallic compounds 2a and 2b are considerably more toxic to the renal (Caki-1) and prostate cancer cell lines (PC3) than cisplatin and Titanocene Y. They have a cytotoxicity similar to that of Auranofin but their selectivity is better. In this case and while gold compound [AuCl(NHC-Bn)] a had shown a cytotoxicity in the low micromolar range for these cell lines,[21] we found that the IC50 value with our method (PrestoBlue™ Cell Viability assay, 72 hours) was 27.7 ± 0.5 μM (similar to that of cisplatin and Titanocene Y). For gold compound b the cytotoxicity in these renal and prostate cancer cell lines had not been reported. The cytotoxicity of the compound with the mba linker 1a improves considerably with respect to compound a and the cytotoxicity of the bimetallic 2a is similar of that of 1a but its selectivity improves. For b the results are different since this compound resulted highly cytotoxic for both the renal and prostate cell lines while having a very good selectivity (see Table 2). The modification of incorporating the mba linker (compound 1b) decreases the cytotoxicity and selectivity which then improves by incorporation of the titanocene fragment (compound 2b) but still this compound does not result as good as b in terms of cytotoxicity and selectivity combined. It is important to note that the cytotoxicity of the bimetallic compounds 2a and 2b in the renal cancer cell line Caki-1 improves considerably (5–13-fold) when compared to previously described third generation compounds [(η5-C5H5)2TiMe(μ-mba)Au(NHC)] (TG1–4 in Figure 1) which displayed IC50 values in the range of 21–51 μM for this cell lines. For the prostate cancer cell lines the IC50 value of 2b (9.5 μM) is in the range for those described for TG1–4 (9.8–17 μM) and compound 2a has a lower value at 3.9 μM. We choose compounds a, 1a and 2a and the cell line PC3 for further studies for better comparison with third generation compounds TG1–4 (in Figure 1). We also chose Auranofin as a control for our experiments (as we have done previously with other titanocene-gold compounds[14–16]).</p><p>Following the evaluation of the cytotoxicity of the compounds we proceeded to evaluate how the cells died. For this assay PC-3 cells were incubated with monometallic gold compounds a, 1a Auranofin and bimetallic 2a at the IC50 concentration for 72 hours. We observed that all the compounds induce apoptosis at their IC50 concentration (Figure 4). From these data, it can be deduced that compounds a, 1a and 2a induce similar apoptosis in prostate PC3 cancer cell lines after 72 hours (45–50%) while AF is more apoptotic (73%) in the same cell population. It should be highlighted however, that monometallic compound a has a considerably higher IC50 value than bimetallic 2a and thus, the incorporation of the titanoce metallic fragment improves the apoptotic properties.</p><p>Second generation compounds (SG1–2) had also diplayed relevant apoptotic behavior in renal cancer cells.[5–6] Third generation compounds [(η5-C5H5)2TiMe(μ-mba)Au(NHC)] (TG1[4 in Figure 1) displayed apoptosis in PC3 cell lines as the major mode of cell death. However a quantification of apoptotic cells versus viable cells was not performed.[16] We have hypothesized that the apoptotic behavior of bimetallic titanocene-gold species comes from the two different metallic fragments[9,32] since both gold[38,39] and titanocene dichloride[40] are known to induce apoptosis in several cancer cell lines.</p><!><p>Relevant anti-migration[14–16] and anti-invasion[15] properties have been found for titanocene-gold compounds in renal cancer cell lines, especially for the second generation compounds SG1–2. This is very important for the development of new anticancer chemotherapeutics as increased local cell migration and distal invasion are hallmarks of metastasis.[41] Two third-generation bimetallic titanocene-gold compounds TG1 and TG2 (in Figure 1) provided a reduction of migration of 46% and 55% in prostate cancer PC3 cells. The effect of compounds a, 1a, AF and bimetallic 2a on migration (at IC20 concentration) was determined using a wound-healing 2D scratch assay on a collagen-coated plate (Figure 5A). IC20 amounts are chosen because for these type of compounds at those concentrations around 80% of cells are alive and the effect measured is not due to cell death.[15]</p><p>Figure 5B shows that bimetallic compound 2a reduces migration in ca. 92%. This reduction is significantly higher than that of Titanocene-Y (19%)[16] and that of the third generation bimetallic compounds already described TG1 and TG2 (Figure 1).[7] It is also slightly higher that that of AF (85%) in this cell line. For the renal cancer cell line Caki-1 we found that AF had a sightly or moderately higher migration inhibition than bimetallic compounds SG1–2.[15] We also found that the migration produced by bimetallic 2a is higher than that of monometallic precursor 1a or previusly described compound [AuCl(NHC)] a (92% 2a versus 66% 1a and 71% a at their corresponding IC20 values) indicating again that there is an advantage in linking the titanocene fragment to the monometallic gold compound a.</p><p>These results are quite significant since we have been able to correlate anti-migration with anti-invasion properties for titanocene-gold derivatives.[15] This means that these compounds may indeed have the potential to act as promising antimetastatic agents. Singaling molecules linked to migration and metastasis have also been inhibited significantly by second generation Ti-Au in renal cancer Caki-1 cell lines.[15] The potential for chemotherapeutics displaying both cytotoxicity and antimetastatic properties should be underscored.</p><!><p>Modification in the anti-oxidant profile of cancerous cells is characteristic of chemoresistance and is often accompanied by the overexpression of thioredoxin reductase (TrxR).[42,43] Increases in TrxR levels are critical to the mechanism of cisplatin-resistant cells which renders it an attractive drug target.[44–46] Additionally, preventing the supply of nutrients and oxygen to the tumor site will hinder the tumor progession. Vascular endothelial growth factor (VEGF) is a key growth factor in angiogenesis, which potently stimulates the formation of blood vessels which allows tumor growth and progression.[47,48,49] Gold compounds, including, Auranofin have been reported to significantly inhibit TrxR levels[14–16] and VEGF levels[15] in clear cell renal carcinoma Caki-1 and prostate cancer PC3 cell lines.</p><p>Immunocytochemistry is a technique that allows direct visualization of the presence, intensity or absence of a cellular protein of interest through the use of fluorescently labeled antibodies. The fluorescent signal emitted results from the highly specific binding of antibodies to their unique antigens.[50,51] In order to detect all cells, the DNA-labeling dye DAPI is used as it labels in blue each nuclei corresponding each to one cell. DAPI positive spots are used to obtain the total cell count in a given sample. Phalloidin is used to label the actin protein, which makes up most of mammalian cell's cytoskeleton and here emits a red signal. Through phalloidin positive staining cytoskeletal integrity can be assessed. In the experiment described here, we have also used a fluorescent-labeled antibody for TrxR, colored green, and for VEGF, colored yellow. Through this experiment we aim to determine to which extent the treatment with the metallo-complexes being studied inhibit or lead to the overexpression of TrRx and VEGF.</p><p>The results obtained for the inhibition of TrRx experiment (Figure 6) indicate that a ([AuCl(NHC-Bn)] or "Tacke" compound) inhibits only 3% of TrxR in the assay conditions (IC20 concentrations and 24 h incubation time). Compound 1a inhibits 12% of TrxR activity while 2a inhibits TrxR activity by 38% a reflected by the decrease of TrxR positive cells. The inhibition of TrxR by the bimetallic compound 2a is within the same range than that of AF that with this assay affords an ihibition of 46%. The results obtained for the inhibition of VEGF experiment (Figure 7) indicate that a ([AuCl(NHC-Bn)] or "Tacke" compound) reduces 25% of VEGF expression in assay conditions (IC20 concentrations and 24 h incubation time). Compound 1a reduces 13% of VEGF activity while 2a reduces 44% of VEGF activity. Their reduction is reflected by the decrease of VEGF in positive cells (control). The reduction of VEGF by the bimetallic compound 2a is also within the same range that of AF (50% of VEGF reduction).</p><p>It is clear that the incorporation of the titanocene fragments to compound a improves significantly its TrxR and VEGF inhibitory properties as we have already reported for second[14,15] and third generation[16] titanocene-gold compounds containing phosphane and N-heterocyclic carbene ligands. The improvement presented here is however quite remarkable.</p><!><p>In conclusion, we have synthesized bimetallic compounds containing titanocenes and gold(I)-N-heterocyclic carbene fragments which were already known to display relevant anticancer properties in vitro and in vivo. We have demonstrated that the incorporation of the titanocene fragment improves or does not decrease the cytotoxicity in the human cancer cell lines Caki-1 and PC3. We have demonstrated that a selected bimetallic compound derived from monometallic "Tacke" [AuCl(NHC-Bn)] has an improved pharmacological profile in terms of apoptosis, inhibition of migration, and inhibition of thioredoxin reductase and VEGF in prostate cancer cell lines with respect to the monometallic bioactive gold compound. The work presented here supports the idea that bimetallic compounds can indeed be designed to contain two different active metal-based fragments (and allow for synergistic and/or cooperative effects) by judicious choice of the linker.</p><!><p>NMR spectra were recorded in a Bruker AV400 (1H-NMR at 400 MHz and 13C{1H} NMR at 100.6 MHz). Chemical shifts (δ) are given in ppm and coupling constants (J) in Hertz (Hz), using CDCl3, d6-DMSO or PBS-D2O as solvent, unless otherwise stated. 1H and 13C NMR resonances were measured relative to solvent peaks considering tetramethylsilane = 0 ppm. IR spectra (4000–500 cm−1) were recorded on a Nicolet 6700 Fourier transform infrared spectrophotometer on solid state (ATR accessory). Elemental analyses were performed on a Perkin-Elmer 2400 CHNS/O series II analyzer by Atlantic Microlab Inc. (US). Mass spectra electrospray ionization high resolution (MS-ESI-HR) were performed on a Waters Q-Tof Ultima. The theoretical isotopic distributions have been calculated using enviPat Web 2.0.</p><!><p>[AuCl(tht)],[52] [AuX(NHC)] (X = Cl, Br),[20,21] and Cp2TiMe2[53] were prepared as previously reported. Chemicals were purchased as indicated: Cp2TiCl2, H[AuCl4] (STREM Chemicals), 4-mercaptobenzoic acid (H2mba) and Methyl Lithium solution (1.6M) (Sigma Aldrich). Reaction solvents were purchased anhydrous from Fisher Scientific (BDH, ACS Grade) and Sigma-Aldrich, used without further purification, and dried in a SPS machine and kept over molecular sieves (3 Å, beads, 4–8 mesh), otherwise over sodium if necessary. Deuterated solvents were purchased from Cambridge Isotope Laboratories, Inc. and were kept over molecular sieves (3 Å, beads, 4–8 mesh). Celite (Celite 545, Diatomaceous Earth) was purchased from VWR International and used as received.</p><!><p>H2mba (0.061 g, 0.395 mmol) was added to a solution of KOH (0.022 g, 0.395 mmol) in 20 mL of ethanol (16 mL) and water (4 mL) and stirred for 20 minutes at room temperature. [AuX(NHC)] (X = Cl, Br) (0.250g, 0.395 mmol) was added to the previous solution and stirred for 5 hours. Solvents were removed under reduced pressure, and the residue was washed with water (3 × 5 mL) and 10 ml of mixture of diethyl ether /hexane (3:2) to afford 1a, 1b as white solids. 1a (NHC-Bn): 81% yield (0.241 g). Anal. Calcd. for C34H43AuN2O2S: C, 57.60; H, 3.86; N, 3.73; S, 4.27. Found: C, 57.40; H, 3.77; N, 3.66; S, 4.15. H NMR (CDCl3): δ 5.49 (s, 4H, CH2), 7.04 (d, 3JHH = 7.2 Hz, 4H, ArH), 7.10 (dd, 3JHH = 6.7 Hz, 2JHH = 2.5 Hz, 4H, ArH), 7.24–7.29 (m, 10H, ArH), 7.34 (t, 3JHH = 7.3 Hz, 2H, ArH), 7.52 (d, 3JHH = 8.4 Hz, 2H, ArH), 7.68 (d, 3JHH = 8.5 Hz, 2H, ArH). 13C{1H} NMR (CDCl3): δ 52.64 (s, CH2), 123.12 (s, 4-C6H4), 127.24 (s, 2-C6H5), 127.33 (s, 1-C6H5), 128.12 (s, 3-C6H5), 128.63 (s, 3-C6H5), 128.71 (s, 4-C6H5), 129.37 (s, 4-C6H5), 129.44 (s, 3-C6H4), 130.68 (s, 2-C6H5), 131.87 (s, 2-C6H4), 132.10 (s, C-imidazole), 135.91 (s, 1-C6H5), 153.81 (s, 1-C6H4) 172.07 (s, C=O), 182.40 (s, C-carbene). IR (cm−1): 2947 m,br (OH), 2322 m, 1667 s, 1580 s (νasym CO2), 1417 m, 1281 s (νsym CO2), 1168 s, 1084 s, 766 s. MS (ESI): m/z Calcd 750.67. Found: 997.35 [NHC-Au-NHC]+. 1b (NHC-Et): 77% (0.174 g). Anal. Calcd. for C26H25AuN2O2S: C 49.84; H 4.02; N 4.47; S 5.12. Found: C, 49.56; H, 4.02; N, 4.66; S, 4.89. 1H NMR (CD3Cl3): δ 1.35 (t, 3JHH = 7.21 Hz, 6H, CH3), 4.24 (q, 3JHH = 7.27 Hz, 4H, CH2), 7.20–7.22 (m, , 2JHH = 1.85Hz, 10H, ArH), 7.36–7.37 (m, , 2JHH = 1.85 Hz, 3JHH = 5.17 Hz, 2H, ArH), 7.73 (d, 3JHH = 8.54 Hz, 6H, ArH), 7.79 (d, 3JHH = 8.54 Hz, 4H, ArH). 13C{1H} NMR (CD3Cl3): δ 17.23 (s, CH3), 44.33 (s, CH2), 127.93 (s, 4-C6H4), 129.01 (s, 2,4-C6H5), 129.47 (s, 1-C6H4), 129.67 (s, 2-C6H4), 130.63 (s, 2-C6H5 ), 131.29 (s, C imidazole), 132.10 (s, 3-C6H4), 170.88 (s, C = O), 179.76 (s, C-carbene). IR (cm−1): 2953 m,br (OH), 2364 m, 1678 s, 1577 s (νasym CO2), 1399 s (νsym CO2), 1259 m, 1116 m, 1080 m, 764 m. MS (ESI): m/z Calcd 626.52. Found: 749.28 [NHC-Au-NHC]+.</p><!><p>The corresponding monometallic gold precursor (1a, 1b) (0.220 mmol) was dissolved in tetrahydrofuran (15 mL) and added via addition funnel over a solution of Cp2TiMe2 (0.046 g, 0.220 mmol) in toluene (5 mL) under nitrogen. The resulting bright orange solution was stirred for 45 minutes at room temperature. The solution was filtered off and the solvents were then removed under reduced pressure to afford an oily solid that was washed with diethyl ether (3 × 5 mL) and hexane (3 × 5 mL) to yield a yellow solid. 2a: (NHC-Bn): 46% yield (0.120 g). Anal. Calcd. for C47H41AuN2O2STi•1.5H2O: C, 58.21; H, 4.58; N, 2.98; S, 3.31. Found: C, 58.05; H, 4.50; N, 2.76; S, 3.55. H NMR (CDCl3): δ 0.97 (s, 3H, Ti-CH3), 5.46 (s, 4H, CH2), 6.19 (s, 10H, Cp), 7.01 (d, 3JHH = 7.2 Hz, 6H, ArH), 7.06 (d, 3JHH = 6.7 Hz, ArH, 5H), 7.22–7.23 (m, ArH, 5H), 7.27–7.33 (m, ArH, 6H), 7.42 (d, 3JHH = 8.4 Hz, ArH, 2H). 13C{1H} NMR (CDCl3): δ 43.98 (s, Ti-CH3), 52.64 (s, CH2), 114.26 (s, Cp), 127.31 (s, 2-C6H5), 127.37 (s, 1-C6H5), 128.04 (s, 4-C6H4), 128.58 (s, 3-C6H5), 128.64 (s, 4-C6H5), 129.24 (s, 4-C6H5), 129.29, (s, 3-C6H4), 130.69 (s, 2-C6H5), 131.68 (s, 2-C6H4), 132.02 (s, C-imidazole), 135.91 (s, 1-C6H5), 149.69 (s, 1-C6H4), 171.77 (s, C=O), 182 (s, C-carbene). IR (cm−1): 2953 m (Cp), 1632, 1580 m (νasym CO2), 1444 s, 1281 vs (νsym CO2), 1165 m (Cp), 1083 m (Cp). 2b: (NHC-Et): 26% yield (0.046 g). Anal. Calcd. for C37H37AuN2O2STi•½C7H9: 54.29; H, 4.56; N, 3.42; S, 3.92. Found: C, 55.64; H, 5.07; N, 3.21; S, 3.38. H NMR (CDCl3): δ 0.97 (s, 3H, Ti-CH3), 1.35 (t, 3JHH = 7 Hz, 6H, CH3), 4.23 (q, 3JHH = 7.4 Hz, 4H, CH2), 6.20 (s, 10H, Cp), 6.16–7.22 (m, 5H, ArH), 7.35–7.37 (m, 4H, ArH), 7.40 (d, 3JHH = 8.6 Hz, ArH, 2H), 7.63 (d, 3JHH = 8.3 Hz, 2H, ArH). 13C{1H} NMR (CDCl3): δ 17.05 (s, CH3), 43.92 (s, Ti-CH3), 44.15 (s, CH2), 114.27 (s, Cp), 127.81 (s, 4-C6H4), 128.19 (s, 1-C6H5), 128.81 (s, C6H5), 129.23 (s, C6H5), 129.30 (s, 3-C6H4), 130.48 (s, C6H5), 131.06 (s, C-imidazole), 131.73 (s, 2-C6H4), 150.05 (s, 1-C6H4), 171.91 (s, C=O), 182.26 (s, C-carbene). IR (cm-1): 2951 m (Cp), 2338 m, 1635 m, 1581 (νasym CO2), 1441 m, 1279 m (νsym CO2), 1168 s (Cp), 1080 s (Cp).</p><!><p>Calculations of compounds 2a and 2b have been carried out with the hybrid density functional method B3LYP as implemented in Gaussian09.[54] Structures were optimized as in a gas phase using 6–311G(d) basis set for all atoms but Au for which the all-electrons DPZ plus polarization basis set was implemented and obtained from the EMSL Basis Set Library.[55,56] The stability of optimized structures has been verified by performing frequency calculations at the same level of theory. Both mono-dentate 2a and 2b complexes were computed to be stabilized over the corresponding bi-dentate complexes by about 3.9 kcal/mol. All computed frequencies for optimized structures were positive and their computed normal mode displacements were employed to predict infra-red intensities. Computed spectra were simulated by a Lorentzian broadening with a band width of 30 cm−1 on half height of each band.</p><!><p>Colorless single crystals were obtained by slow diffusion of n-hexane into solutions of the complex in CHCl3 (1a −30°C) or THF (1b 0 °C). The diffraction data were collected using graphite monochromatic Mo-Kα radiation with a Bruker APEX-II diffractometer at a temperature of 120 K using the APEX-II software. Structures were solved by Intrinsic Phasing using SHELXT[57] and refined by full-matrix least squares on F2 with SHELXL.[58] The absorption correction was performed using MULTISCAN[59] with the WINGX program suite.[60] All non-hydrogen atoms were assigned anisotropic displacement parameters. The hydrogen atoms were positioned geometrically, with isotropic parameters 1.2 times the Uiso value of their attached carbon for the aromatic and methylene hydrogens and 1.5 times for the methyl groups. For 1a one molecule of CHCl3 was found in the asymmetric unit, giving rise the stoichiometry 1a•CHCl3. The structure of 1b shows some residual peaks greater than 1 e Å−3 in the vicinity of the gold atom, with no chemical meaning. These data can be obtained free of charge from The Cambridge Crystallographic Data Center via www.ccdc.cam.ac.uk/data_request/cif (CCDC 1882280 1a and CCDC 1882281 1b) or in the Supplementary Information.</p><!><p>Human renal clear cell carcinoma line Caki-1 and human prostate adenocarcinoma line PC3 were obtained from the American Type Culture Collection (ATCC) (Manassas, Virginia, USA) and cultured using Roswell Park Memorial Institute (RPMI-1640) (Fisher Scientific, Hampton, NH) media containing 10% Fetal Bovine Serum, certified, heat inactivated, US origin (FBS) (Fisher Scientific, Hampton, NH), 1% Minimum Essential Media (MEM) nonessential amino acids (NEAA) (Fisher Scientific, Hampton, NH), and 1% penicillin-streptomycin (PenStrep) (Fisher Scientific, Hampton, NH). Human fetal lung fibroblast IMR90 cells were purchased from ATCC (Manassas, Virginia, USA) and cultured using Dulbecco's modified Eagle's medium (DMEM) (Fisher Scientific, Hampton, NH) supplemented with 10% FBS, 1% MEM-NEAA, and 1% PenStrep. All cells were cultured at 37 °C under 5% CO2 and 95% air in a humidified incubator.</p><!><p>The cytotoxic profile (IC50) of the compounds were determined by assessing the viability of PC3, Caki-1 and IMR90 cells. Cells were seeded at a concentration of 5 × 103 cells/well in 90 μL of appropriate complete media without phenol red into tissue culture grade 96-well flat bottom microplates (BioLite Microwell Plate, Fisher Scientific, Waltham, MA) and grown for 24 h at 37 °C under 5% CO2 and 95% air in a humidified incubator. The compounds a, 1a, 1b and 2a were dissolved in a 1:1 solution of Triethylglycol and DMSO, while compounds b, 2b, AF and Ti-Y were dissolved in DMSO. Cisplatin was dissolved in H2O. The intermediate dilutions of the compounds were added to the wells (10 μL) to obtain concentration of 0.1 μM, 1 μM, 10 μM, 50 μM and 100 μM, 0.1% DMSO was used as control, and the cells were incubated for 72 h. Presto Blue was used to quantitively measure variations in cell viability of treated cells. Following 72 h drug exposure, 11 μL per well of 10x PrestoBlue (Invitrogen, Carlsbad, CA) labeling mixture was added to the cells at a final concentration of 1x and incubated for 1 h at 37 °C under 5% CO2 and 95% air in a humidified incubator. The optical fluorescence of each well in a 96-well plate was quantified using a BioTek Synergy Multi-mode microplate reader (BioTek Instruments, Inc., Winooski, VT) set at 530/25 excitation nm and 590/35 nm emission. The percentage of surviving cells was calculated from the ratio of absorbance of treated to untreated cells. The IC50 value was calculated as the concentration reducing the proliferation of the cells by 50% and is presented as a mean (±S.EM) of at least two independent experiments each with triplicate measurements.</p><!><p>For the assessment of the cell death in PC3, cells were cultured in 100 mm tissue culture dishes (Fisher Scientific, Hampton, NH) using RPMI phenol red free medium and reach a growth ~75% confluency. The cells were dose using the IC50 of compounds a, 1a, 2a, AF and incubated for 72 h at 37 °C under 5% CO2 and 95% air in a humidified incubator. Then, cells were collected using trypsin (Fisher Scientific, Hampton, NH) to later count with a hemocytometer and prepare 25 ×104 cells/sample. After the samples were prepared, the cells in the samples were centrifuged, supernatant was aspirated from the pallet, and washed gently one time with PBS. The cells were centrifuged again, PBS was aspirated and eBioscience Annexin V-FITC Apop Kit (Invitrogen, Carlsbad, CA) was used to labeled cells as follows, the cells were resuspended in 195 μL of 1x binding buffer and 5 μL of Annexin-V dye was added. The cells were incubated at room temperature for 10 min. After the incubation period finished, cells were washed one time with 1x binding buffer and resuspended in 190 μL of 1x binding buffer with 10 μL of propidium iodide. The dye's fluorescence intensity was detected via flow cytometry using a BD C6 Accuri flow cytometer. 10 × 105 events per sample were recorded. The flow cytometer was calibrated prior to each use.</p><!><p>PC3 cells were allowed to seed in fibronectin-coated 6-well plate (Corning Incorporated, Durham, NC) and grown a monolayer of ~90% confluency. After which, the monolayer was scratched using a 200 μL tip. The complete medium and cells detached due to the scratch were aspirated and replaced with serum-free medium. The antimigratory profiles of bimetallic compound 2a, monometallic compounds a, 1a, and AF was assessed with the IC20 of each compound. The diluting agent (1:1, triethylglycol: DMSO) served as a negative control. Cells were incubated at 37 °C under 5% CO2 and 95% air in a humidified incubator. At 0, 6, 24 and 48 h after the scratch, cells were photographed using a Leica MC120 HD mounted on a Leica DMi1 microscope at 5x magnification. The area invaded was measured in five randomly selected segments from each photo then averaged. Data were collected from two independent experiments.</p><!><p>Cells were seeded in an 8-well millicell slide (Millipore, sigma) at a concentration of 25000 cells/well in a humidified atmosphere of 95% air/5% CO2 at 37 °C. 24h post seeding, cells were treated with the IC20 of bimetallic compound 2a, monometallic compound a, 1a, and as control AF and incubated for 24 h. Cells were fixed using 4% PFA and incubate at room temperature for 15 min. The PFA was removed and the wells were washed three times with PBS. Cells were handled carefully as to maintain cellular integrity. The cells were then blocked with 5% BSA (Fisher Scientific, Hampton, NH), 0.3% Triton-X100 (Acros Organics, Morris Plains, NJ, USA) in PBS for 1h at room temperature. The blocking solution was then removed and the wells were washed three times with PBS. After the washes with PBS, cells were incubated overnight at 4 °C with the respective antibodies. TrxR activity in cells was visualized using an rabbit anti-thioredoxin antibody (Novus biological, Littleton, CO). VEGF activity was obtained by using mouse anti-VEGF antibody (Novus biological, Littleton, CO) and goat anti-mouse secondary antiody (Fisher Scientific, Hampton, NH). Then, the antibodies solution was removed from the slide and anti-phalloidin antibody (Cell Signaling Technology, Danvers, MA) to visualize the cytoskeleton of the cells. After 1 h of incubation at room temperature, cells were washed three times with PBS and one drop per well of DAPI containing ProLong Gold Antifade Mounting Medium (Invitrogen, Carlsbad, CA) was used to visualize the nuclei and mount the slide.</p><!><p>Following immunohistochemical processing all stained samples were imaged at 10x magnification using Fluoview FV10i (Olympus America Inc., Center Valley, PA). Cell were quantified one channel at a time using ImageJ, and the percentage of rabbit anti-thioredoxin antibody positive per DAPI positive cells were calculate per field of view over 5 fields of view.</p>

PubMed Author Manuscript

Development of potent dopamine-norepinephrine uptake inhibitors (DNRIs) based on a (2S,4R,5R)-2-benzhydryl-5-((4-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol molecular template

Current therapy of depression is less than ideal with remission rates of only 25\xe2\x80\x9335% and response rates of 45\xe2\x80\x9360%. It has been hypothesized that a dysfunctional dopaminergic system in the mesocorticolimbic pathway in depressive disorder may lead to development of anhedonia associated with loss of pleasure and interest along with loss of motivation. The current antidepressants do not address dopamine dysfunction which might explain their low efficacy. In this report, we have described an SAR study on our pyran-based triple reuptake inhibitors (TRIs) which are being investigated as the next-generation antidepressants. In the present work we demonstrate that our lead TRIs can be modified with appropriate aromatic substitutions to display a highly potent SSRI profile for compounds 2a and 4a (Ki (SERT); 0.71 and 2.68 nM, respectively) or a potent DNRI profile for compounds 6b and 6h (Ki (DAT/NET); 8.94/ 4.76 and 13/ 7.37 nM, respectively). Compounds 4g\xe2\x80\x934i exhibited potencies at all three monoamine transporters. The results provide insights into the structural requirements for developing selective dual- and triple-uptake inhibitors from a unique pyran molecular template for an effective management of depression and related disorders.

development_of_potent_dopamine-norepinephrine_uptake_inhibitors_(dnris)_based_on_a_(2s,4r,5r)-2-benz

Introduction<!>Chemistry<!>Results and Discussion<!>Conclusions<!>Chemistry<!>(3S,6S)-6-benzhydryl-N-(benzofuran-5-ylmethyl)tetrahydro-2H-pyran-3-amine (2a)<!>(2S,4R,5R)-2-benzhydryl-5-((benzofuran-5-ylmethyl)amino)tetrahydro-2H-pyran-4-ol (4a)<!>(2S,4R,5R)-2-benzhydryl-5-((4-hydroxy-3-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol (4b)<!>(2S,4R,5R)-2-benzhydryl-5-((3-hydroxy-4-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol (4c)<!>(2S,4R,5R)-2-benzhydryl-5-((3-hydroxy-5-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol (4d)<!>5-((((3R,4R,6S)-6-benzhydryl-4-hydroxytetrahydro-2H-pyran-3-yl)amino)methyl)benzene-1,3-diol (4e)<!>(2S,4R,5R)-2-benzhydryl-5-((3-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol (4f)<!>(2S,4R,5R)-2-benzhydryl-5-(((6-methoxypyridin-3-yl)methyl)amino)tetrahydro-2H-pyran-4-ol (4g)<!>(2S,4R,5R)-2-benzhydryl-5-((3-fluoro-4-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol (4h)<!>(2S,4R,5R)-2-benzhydryl-5-((2-fluoro-4-methoxybenzyl)amino)tetrahydro-2H-pyran-4-ol (4i)<!>(2S,4R,5R)-5-((benzofuran-5-ylmethyl)amino)-2-(bis(4-fluorophenyl) methyl)tetrahydro-2H-pyran-4-ol (6a)<!>(2S,4R,5R)-2-(bis(4-fluorophenyl)methyl)-5-((4-hydroxy-3-methoxybenzyl) amino)tetrahydro-2H-pyran-4-ol (6b)<!>(2S,4R,5R)-2-(bis(4-fluorophenyl)methyl)-5-((2,3-dihydrobenzofuran-5-yl)methyl)amino)tetrahydro-2H-pyran-4-ol (6c)<!>(2S,4R,5R)-2-(bis(4-fluorophenyl)methyl)-5-((3-hydroxy-5-methoxybenzyl) amino)tetrahydro-2H-pyran-4-ol (6d)<!>(2S,4R,5R)-5-((benzo[d][1,3]dioxol-5-ylmethyl)amino)-2-(bis(4-fluorophenyl)-methyl)tetrahydro-2H-pyran-4-ol (6e)<!>(2S,4R,5R)-2-(bis(4-fluorophenyl)methyl)-5-((3-methoxybenzyl) amino) tetrahydro-2H-pyran-4-ol (6f)<!>(2S,4R,5R)-2-(bis(4-fluorophenyl)methyl)-5-(((6-methoxypyridin-3-yl)methyl) amino)tetrahydro-2H-pyran-4-ol (6g)<!>(2S,4R,5R)-2-(bis(4-fluorophenyl)methyl)-5-((3-fluoro-4-hydroxybenzyl)amino) tetrahydro-2H-pyran-4-ol (6h)<!>Functional transporter assays<!>

<p>Major depressive disorder (MDD) is a debilitating illness affecting 15–20% of the population in the United States.[1] Depression is characterized by symptoms like insomnia, loss of appetite, and psychomotor agitation. MDD is a severe form of depression defined by multiple episodes of depressed mood that persisting for at least 2 weeks accompanied by at least four other symptoms.[2, 3] According to the WHO by 2020 MDD would be the second-most leading cause of disability worldwide, affecting 121 million people, thus making it a global health problem.[4]</p><p>Modulation of serotonin (5-HT) and noradrenergic (NE) systems is at the core of the "monoamine deficiency" hypothesis of depression that postulates impaired monoaminergic transmission, either due to a deficit of monoamine neurotransmitters in synapses and surrounding extracellular space, or disturbed monoamine receptor signaling.[5, 6] Therefore, most current antidepressants block serotonin and norepinephrine transporters (SERTs and NETs), either as serotonin-norepinephrine reuptake inhibitors (SNRIs) or as selective serotonin reuptake inhibitors (SSRIs).[7–10] Converging evidences suggest that inhibitors targeting additionally NE neurotransmission may be more efficacious than those acting selectively on 5-HT systems in MDD.[11–15] In a meta-analysis, venlafaxine (Figure 1), an SNRI, exhibited greater response and remission rates than an SSRI.[11] Although a plethora of antidepressants are on the market, there still remains a significant unmet need for improved therapy, as large numbers of depressed people are still refractory to the current existing drugs. Thus, current therapy is less than ideal with remission rates of only 25–35% and response rates of 45–60%.[16] Furthermore, slow onset of action of the current therapies along with other associated side effects call for improvements in MDD therapy.</p><p>Dopamine has been linked to depression for quite some time.[17–20] The medial prefrontal cortex has been shown to be associated with depressed mood and sadness and neuroimaging studies indicate deficiencies of neuronal activity in this brain area in depressed subjects.[21] This region receives innervations from all three monoamines; thus, restoration in the imbalanced level of monoamines by antidepressants has been shown to improve symptoms of depression.[22] Since dopamine controls mood and emotion, a dysfunctional dopaminergic system in the mesocorticolimbic pathway may lead to development of anhedonia associated with loss of pleasure and interest along with loss of motivation.[19] The persistence of anhedonia makes it one of the most treatment-resistant symptoms of MDD with a tendency for anhedonia to increase over the course of the disorder.23–25 Addition of dopaminergic activity to an antidepressant with serotoninergic and norepinephrine activity should alleviate these symptoms provided the additional dopamine component does not introduce abuse liability.19 Interestingly, in a recent study DOV 21,947 (amitifadine), a TRI, has been shown to produce little to no abuse liability.26 Be that as it may, any novel monoamine transporters blocker which includes dopaminergic activity such as a TRI, should undergo extensive abuse liability testing.</p><p>It has been hypothesized that inhibitors with ability to block dopamine reuptake in addition to NE reuptake (DNRIs) should also have robust therapeutic effects by addressing dopamine deficit-related anhedonia in depression.27, 28 Given the side effects associated with clinically used SSRIs, exploration of dual dopamine-norepinephrine reuptake inhibitors (DNRIs) is important, and understanding the relationship between dopaminergic and noradrenergic systems is of significant interest in addressing the pathophysiology of MDD.29 In this regard, it is important to mention that there are only few compounds which are known to exhibit a DNRI-type profile. One well known example of this is bupropion, used as an antidepressant agent in the clinic.[17, 30, 31] It should be noted that the mechanism of antidepressant action of bupropion is quite complex and involves other targets in addition to dopaminergic and noradrenergic systems.30 It should be noted that a successful adjunct therapeutic approach involving use of bupropion and an SSRI was found to be more efficacious in patients refractory to SSRI.[17, 20] Nomifensine is another DNRI that has antidepressant activity and is as effective as imipramine.[32, 33] It should be noted that DNRIs, including nomifensine, have also been reported to exert a therapeutic effect on attention-deficit hyperactivity disorder (ADHD), a related mental disorder.[34] In this regard, atomoxetine, a NET inhibitor, is the only non-stimulant drug approved for ADHD.[35] It is perceived that a dual DNRI with a potency order of NET>DAT might offer better alternatives to current ADHD therapies.[36]</p><p>In recent years triple reuptake inhibitors (TRIs) inhibiting all three monoamine transporters have been hypothesized to produce greater efficacy than SSRIs or dual uptake inhibitors. A number of TRIs, e.g. DOV 21,947 (amitifadine), PRC200-SS, JNJ-7925476 and GSK-372,475 have been developed (Figure 1) and have been characterized in animal models for depression.[37, 38] Recently, DOV 21,947, now known as amitifadine, has undergone Phase IIb clinical trial which demonstrated promising efficacy in a treatment-resistant MDD patient group.</p><p>In our effort to address the unmet need in antidepressant therapy, we have been working to develop unique asymmetric pyran based inhibitors of monoamine uptake systems.39–43 Thus, our drug development work led to discovery of novel TRIs which interacted with all three transporters. In this manuscript, we report a structure-activity relationship (SAR) study describing modifications of the TRI profile of our compounds resulting in potent SSRI- and DNRI-type transporter inhibitors.</p><!><p>Amines 1 (disubstituted pyran), 3, and 5 (trisubstituted pyran) were synthesized according to procedures we have published in our earlier reports.[39, 40] Reductive amination of amines 1, 3, and 5 (Schemes 1, 2, and 3, respectively) with appropriate aldehydes in presence of sodium triacetoxyborohydride or sodiumcyanoborohydride and catalytic amount of acetic acid in 1,2-dichloroethane afforded final target compounds 2a, 4a–4i, 6a–6h in appreciable yields.</p><!><p>Our recent endeavors have resulted in the discovery of several TRIs that exhibited potencies at all three monoamine transporters. In this regard, our lead TRIs, D-142 and D-161 (Figure 1) were shown to be efficacious, as evidenced by significant reduction of immobility, in both rat forced swim tests (FSTs) and mouse tail suspension tests (TSTs) that are established models for preclinical testing of potential antidepressants.[44, 45] Recently, we have reported development of an orally active TRI, D-473 (Figure 1), which exhibited efficacious activity in FST and elevated level of all three monoamines in a microdialysis study.46</p><p>In our quest to develop effective therapy for the treatment of MDD, we herein report an extensive SAR study on our trisubstituted pyran derivatives leading to the discovery of novel potent DNRIs. Their functional effect at monoamine transporters was assessed by monitoring inhibition of substrate uptake in synaptosome-enriched fractions from rat brain (striatum for DAT assays and cerebral cortex for SERT and NET assays). Subtle structural modifications of our pyran-based inhibitors, interestingly, led to significantly different uptake profiles. Compound 2a exhibited highly potent and selective serotonin reuptake inhibition (Table 1) with Ki values of 82 nM, 0.71 nM, and 25 nM, at DAT, SERT, and NET, respectively. It should be noted that compound 2a is one of the most potent SERT inhibitors known to date. Compound 2a was 20 times more active at SERT than fluoxetine[31] with DAT, SERT, and NET inhibitory ratios of 115:1:35. Its trisubstituted counterpart, compound 4a, was also selective for SERT and much weaker at DAT (Ki: 234, 2.7, and 34 nM for DAT, SERT, and NET, respectively). The exceptional potency at SERT might be due to aromatic π-stacking interactions of benzofuran moiety in 4a. This notion was supported by the observation of a similar π-π stacking interaction between fluoxetine and Tyr176, in the SERT binding site.[47] Moreover, mutation of Tyr176 also has been shown to significantly affect the binding of fluoxetine indicating the importance of this interaction. Since furan oxygen can also participate as H-bond acceptor, the role of Hbonding interactions at this position was further explored. In this regard compounds 4b–4d were synthesized and biologically evaluated. Compound 4b showed moderate affinity at DAT and SERT but was potent at NET (Ki: 167, 223, and 34 nM for DAT, SERT, and NET, respectively). Compound 4c exhibited appreciable potencies at DAT (Ki: 58 nM) and NET (Ki: 30 nM) and was weakly active at SERT (Ki: 281 nM). Inhibition of SERT uptake was further weakened in compound 4d (Ki: 209, 385, and 44 nM for DAT, SERT, and NET, respectively). Thus, the binding data of compounds 4b–4d suggest that for DAT inhibition methoxy was most favorable at the para-position of the N-benzyl group. For inhibition of norepinephrine uptake hydroxyl and methoxy moieties could be substituted at the meta- and para-positions with retention of activities; However, hydroxyl, methoxy disubstitution led to significant reduction in SERT inhibition. It can be inferred that an H-bond donor particularly in the presence of another potential H-bonding group, resulted in weaker potencies at SERT. In accordance, the dihydroxy substituted compound 4e was inactive at SERT (Ki: 3.0 µM) and only moderately active at DAT (Ki: 259 nM) and NET (Ki: 152 nM). In contrast, a monosubstituted compound 4f, which has a methoxy group at the meta-position, is favored at SERT (Ki: 27 nM) and NET (Ki: 5.8 nM) but detrimental at DAT (Ki: 376 nM) inhibition. Thus, 4f displayed an SNRI profile. Interestingly, compound 4g exhibited a balanced TRI profile (Ki: 44, 42, and 38 nM for DAT, SERT, and NET, respectively) suggesting that phenyl in the N-benzyl moiety could be replaced by a heterocyclic pyridine ring. The methoxy group in compounds 4f and 4g can participate in electronic interactions and can also act as a weak H-bond acceptor. The pyridine ring in compound 4g imparts a polar character which is tolerated well at DAT, SERT and NET. Our next exploration was aimed at understanding the effect of fluorine substitution on the N-benzyl group with respect to inhibition of neurotransmitter uptake by the three monoamine transporters. Compound 4h also exhibited TRI activity (Ki: 48, 9.0, and 42 nM for DAT, SERT, and NET, respectively). The improved activity of 4h at SERT may be a result of polar interactions of electronegative fluorine at the meta-position. Next, the 2-fluoro substituted compound 4i also retained TRI activity (Ki: 52, 64, and 9.3 nM for DAT, SERT, and NET, respectively). It should also be noted that enhancement of SERT activity resulting in production of the TRI profile of compounds 4g–4i may be due to the presence of methoxy substituent at the para-position. Taken together with our earlier SAR studies, trisubstituted unsubstituted benzhydryl pyran-based TRI could be designed by appropriate N-benzyl aromatic substitutions, at least with: (1) substitution of a p-methoxy or hydroxyl group, which suggest a H-bonding interaction, at the para-position; (2) presence of polar or H-bond acceptor groups at the ortho- and meta-positions; (3) accordingly, absence of hydrophobic or H-bond donor moieties at the meta-position and; (4) not more than one Hbond donor or two hydrogen bonding groups on the N-benzyl phenyl ring.</p><p>Difluoro benzhydryl pyran derivatives were recently developed as novel TRIs by our laboratory.[40] In general, we found from our earlier SAR studies that difluoro substituted compounds favorably interact with DAT and tend to show somewhat decreased activity at SERT. In this SAR study we have further explored the effect of introduction of the difluoro substituent on the benzhydryl moiety. Compounds 6e (Ki: 37, 400, and 10 nM for DAT, SERT, and NET, respectively) and 6c (Ki: 17, 54, and 26 nM for DAT, SERT, and NET, respectively) conferred (i) an improved inhibitory potency at DAT and NET, and (ii) a weaker SERT inhibitory activity than their non-fluoro counterparts as we have reported in our earlier study.40 This effect is further exemplified by comparing the inhibitory activities of compounds 6a and 4a. Thus, as predicted compound 6a produced much greater potency at DAT compared to 4a with reduction of potency at SERT (Ki values of 29 nM, 68 nM, and 26 nM for DAT, SERT, and NET, respectively). Compound 6f, as predicted, was weaker at SERT (Ki of 129 nM) compared to non-fluorinated analogue 4f but showed improved DAT potency (Ki: 56 nM). It should be noted that NET activity is generally well tolerated in both non-fluoro and disubstituted fluorinated benzhydryl pyran analogs. Based on such extensive SAR data, as well as results from our previous studies, the next series of compounds was designed aiming for DAT and NET, i.e. as novel DNRIs, a relatively unexplored class of antidepressants. As mentioned before, a DNRI can also be potentially applied in the treatment of ADHD, without the liability of acting as a stimulant if its potency rankorder is NET>DAT. Thus, compounds 6d and 6b were designed as DNRIs by incorporating: (a) difluoro substitution on the benzhydryl group; and (b) two hydrogen bonding substituents, at specific positions, on the N-benzyl moiety. Based on the SAR studies, it was expected that a hydroxyl, methoxy disubstitution on the N-benzyl group coupled with difluoro benzhydryl moiety was highly detrimental for SERT inhibition. In concordance to our design, 6d (Ki: 26, 577, and 4.9 nM for DAT, SERT, and NET, respectively) and 6b (Ki: 8.9, 107, and 4.8 nM for DAT, SERT, and NET, respectively) turned out to be potent DNRIs. From compounds 6b and 4f it appears that the unfavorable effect of a meta-methoxy substituent at DAT in 4f is compensated by two favorable structural features, namely hydroxyl at para-position on the N-benzyl group and the difluoro substituent on the benzhydryl moiety. These structural features on the other hand impart reduction in SERT activity with retention of activity at NET and gain in potency for DAT, thus, producing a DNRI type-effect. Finally, compounds 6h (Ki: 13, 334, and 7.3 nM for DAT, SERT, and NET, respectively) and 6g (Ki: 20, 258 and 28 nM for DAT, SERT, and NET, respectively) also displayed DNRI properties. Further comparison of the binding profiles of compounds 6g and 4h validate our design of modifying a TRI to produce a DNRI. Figure 2 represents an updated model of pyran derivatives developed for TRI, SSRI, and DNRI properties. Thus, in the present series, compounds 6b and 6h were the most potent and selective DNRIs discovered; both compounds exhibited favorable physicochemical properties for brain penetration as shown in Table 2.</p><!><p>This report describes an SAR study on our pyran-based TRIs, investigated as the next-generation antidepressants. In the present work we demonstrate that our lead TRIs could be transformed into compounds with SSRI and DNRI profiles by proper aromatic substitutions. Compounds 2a and 4a were discovered as some of the most potent SSRIs developed to date. Compounds 4g–4i exhibited balanced potencies at all three monoamine transporters, whereas compounds 6d–6g were developed as novel DNRIs. Given the absence of 3D-structures for the three monoamine transporters, and the homology existing between them, it is challenging to exploit subtle differences and similarities in their drug binding sites. The SAR approach taken in the present study provides insights into the structural requirements for developing selective, dual, and triple-uptake inhibitors from a unique pyran molecular template which with optimal pharmacokinetic properties should provide a potential for a more effective management of depression and related disorders than afforded by current drug therapies.</p><!><p>Reagents and solvents were obtained from commercial suppliers and used as received unless otherwise indicated. Dry solvents were obtained according to the standard procedures. All reactions were performed under inert atmosphere (N2) unless otherwise noted. Analytical silica gel-coated TLC plates (silica gel 60 F254) were purchased from EM Science and were visualized with UV light or by treatment with either phosphomolybdic acid (PMA) or ninhydrin. Flash chromatography was carried out on Baker Silica Gel 40 µM. 1H NMR and 13C spectra were routinely recorded with a Varian 400 spectrometer operating at 400 and 100 MHz, respectively. The NMR solvent used was either CDCl3 or CD3OD as indicated. TMS was used as an internal standard. NMR and rotation of free bases were recorded. Salts of free bases were used for biological characterization. Elemental analyses were performed by Atlantic Microlab Inc. and were within ± 0.4% of the theoretical value. Optical rotations were recorded on a Perkin-Elmer 241 polarimeter.</p><!><p>To a stirred solution of amine 1 (60 mg, 0.22 mmol) and 1-benzofuran-5-carbaldehyde (35 mg, 0.24 mmol) in 1,2-dichloroethane (6 mL) was added glacial acetic acid (13 µL, 0.22 mmol). After being stirred for 30 minutes, NaCNBH3 (28 mg, 0.44 mmol) was added portion wise followed by methanol (1 mL). The reaction mixture was stirred for overnight. The reaction mixture was quenched with saturated NaHCO3 solution at 0 °C and extracted with dichloromethane (3 X 75 mL). The combined organic layer was washed with water, brine, dried over Na2SO4, and the solvent was removed under reduced pressure. Crude product was purified by column chromatography using 70% ethyl acetate in hexanes to give compound 2a (D-484) (60 mg, 67%) as thick syrup. [α]D25=(−)78.2° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3): δ 1.25–1.38 (m, 1H), 1.50–1.70 (m, 2H), 1.82–2.02 (m, 1H), 2.68 (br s, 1H), 3.56 (dd, J = 1.6, 12.0 Hz, 1H), 3.80–4.12 (m, 6H), 6.73 (d, J = 1.2 Hz, 1H), 7.14–7.46 (m, 12H), 7.56 (s, 1H), 7.61(d, J = 2.0 Hz, 1H). 13C NMR (100 MHz, CDCl3): δ 25.43, 27.76, 50.46, 50.98, 57.42, 70.40, 79.50, 106.75, 111.37, 120.78, 124.93, 126.47, 126.66, 127.70, 128.54, 128.74, 128.80, 135.11, 142.44, 142.70, 145.44, 154.42. The product was converted into the corresponding hydrochloride salt; mp: 140–142 °C. Anal. (C27H27NO2·HCl·0.7H2O) C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 1-benzofuran-5-carbaldehyde (34 mg, 0.23 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (26 mg, 0.42 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using ethyl acetate to afford compound 4a (D-485) (60 mg, 69%) as a white solid. [α]D25=(−)64.8° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3): δ 1.43 (dt, J = 3.2, 14.0 Hz, 1H), 1.70–1.80 (m, 1H), 2.48 (d, J = 2.4 Hz, 1H), 3.76–3.84 (m, 2H), 3.88–4.16 (m, 4H), 4.51 (dt. J = 2.4, 10.4 Hz, 1H), 6.73 (d, J = 1.6 Hz, 1H), 7.12–7.38 (m, 11H), 7.44 (d, J = 8.8 Hz, 1H), 7.52 (s, 1H), 7.61 (d, J = 2.4 Hz, 1H). 13C NMR (100 MHz, CDCl3): δ 33.68, 51.65, 56.74, 56.78, 65.05, 67.72, 73.85, 106.73, 111.45, 120.82, 124.90, 126.56, 126.76, 127.74, 128.62 (3C), 128.67 (3C), 128.87 (3C), 134.86, 142.27, 142.36, 145.53, 154.48. The product was converted into the corresponding hydrochloride salt; mp: 204–206 °C. Anal. (C27H27NO3·HCl·0.5H2O) C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with vanillin (39 mg, 0.25 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (26 mg, 0.42 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using 5% methanol in ethyl acetate to afford compound 4b (D-501) (65 mg, 73%) as a thick syrup. [α]D25=(−)80.4° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3): δ 1.38–1.48 (m, 1H), 1.68–1.78 (m, 1H), 2.52 (br s, 1H), 3.66 (d, J = 12.8 Hz, 1H), 3.78 (s, 3H), 3.76–3.86 (m, 2H), 3.88–4.04 (m, 3H), 4.52 (dt, J = 2.4, 8.0 Hz, 1H), 6.73 (d, J = 8.0 Hz, 1H), 6.77 (d, J = 8.4 Hz, 1H), 6.86 (s, 1H), 7.12–7.37 (m, 10H). 13C NMR (100 MHz, CDCl3): δ 33.53, 51.18, 56.05, 56.63, 56.68, 64.30, 66.82, 74.02, 111.42, 114.78, 119.20, 121.56, 126.61, 126.82, 128.62 (3C), 128.66 (3C), 128.91 (3C), 142.23, 145.29, 147.09. The product was converted into the corresponding hydrochloride salt; mp: 203–205 °C. Anal. (C26H29NO4·HCl·0.5H2O) C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 3-hydroxy-4-methoxybenzaldehyde (39 mg, 0.25 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (26 mg, 0.42 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using 5% methanol in ethyl acetate to afford compound 4c (D-502) (65 mg, 73%) as a thick syrup. [α]D25=(−)78.4° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3 + MeOH-d4): δ 1.36–1.44 (m, 1H), 1.54–1.64 (m, 1H), 2.45 (d, J = 2.8 Hz, 1H), 2.85 (br s, 1H), 3.58 (d, J = 12.8 Hz, 1H), 3.72–3.78 (m, 2H), 3.81 (s, 3H), 3.84–3.94 (m, 3H), 4.47 (dt, J = 2.4, 10.8 Hz, 1H), 6.69 (dd, J = 1.6, 8.4 Hz, 1H), 6.74 (d, J = 8.0 Hz, 1H), 6.83 (d, J = 1.6 Hz, 1H), 7.08–7.32 (m, 10H). The product was converted into the corresponding hydrochloride salt; mp: 168–170 °C. Anal. (C26H29NO4·HCl·H2O) C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 3-hydroxy-5-methoxybenzaldehyde (38 mg, 0.25 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (26 mg, 0.42 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using 5% methanol in ethyl acetate to afford compound 4d (D-503) (65 mg, 73%) as a thick syrup. [α]D25=(−)75.6° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3): δ 1.34–1.42 (m, 1H), 1.52–1.64 (m, 1H), 2.41 (br s, 1H), 3.52 (d, J = 13.2 Hz, 1H), 3.64 (s, 3H), 3.60–3.75 (m, 2H), 3.78–3.93 (m, 2H), 4.45 (dt, J = 1.6, 10.0 Hz, 1H), 4.56 (br s, 2H), 6.26 (s, 1H), 6.32 (s,1H), 6.33 (s, 1H), 7.10–7.32 (m, 10H). 13C NMR (100 MHz, CDCl3): δ 33.35, 50.85, 55.47, 56.34, 56.51, 63.94, 66.58, 74.14, 101.09, 106.51, 108.32, 126.67, 126.86, 128.54 (3C), 128.71 (3C), 128.94 (3C), 140.79, 142.04, 142.18, 158.00, 161.25. The product was converted into the corresponding hydrochloride salt; mp: 160–162 °C. Anal. (C26H29NO4·HCl·0.8H2O) C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 3,5-dihydroxybenzaldehyde (29 mg, 0.21 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (26 mg, 0.42 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using 7% methanol in dichloromethane to afford compound 4e (D-542) (65 mg, 76%) as a thick syrup. [α]D25=(−)74.2° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3): δ 1.34–1.42 (m, 1H), 1.55 (dt, J = 2.4, 10.8 Hz, 1H), 2.43 (br s, 1H), 3.30 (s, 1H), 3.51 (d, J = 12.8 Hz, 1H), 3.60–3.70 (m, 2H), 3.78–3.90 (m, 3H), 4.31 (br s, 3H), 4.45 (t, J = 8.4 Hz, 1H), 6.16 (s, 1H), 6.21 (s, 1H), 6.22 (s, 1H), 7.02–7.28 (m, 10H). 13C NMR (100 MHz, CDCl3): δ 33.06, 50.68, 55.99, 56.73, 63.94, 65.73, 74.08, 102.18, 107.35, 126.51, 126.72, 128.46 (3C), 128.48 (2C), 128.55, 128.77(3C), 139.97, 141.97, 142.18, 158.26. The product was converted into the corresponding hydrochloride salt; mp: 168–170 °C. Anal. (C25H27NO4·HCl·H2O) C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 3-methoxybenzaldehyde (35 mg, 0.25 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (27 mg, 0.42 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using 3% methanol in ethyl acetate to afford compound 4f (D-523) (65 mg, 76%) as a thick syrup. [α]D25=(−)84.2° (c 0.5, MeOH). 1H NMR (400 MHz, CDCl3): δ 1.38–1.46 (m, 1H), 1.66–1.78 (m, 1H), 2.16 (br s, 2H), 2.44 (d, J = 2.4 Hz, 1H), 3.70 (d, J = 13.6 Hz, 1H), 3.76–3.82 (m, 1H), 3.79 (s, 3H), 3.84–3.98 (m, 4H), 4.50 (dt, J = 2.4, 10.4 Hz, 1H), 6.80 (dd, J = 1.6, 8.0 Hz, 1H), 6.86–6.92 (m, 2H), 7.14–7.38 (m, 11H), 13C NMR (100 MHz, CDCl3): δ 33.40, 51.26, 55.22, 56.50, 56.64, 64.76, 67.30, 73.61, 112.53, 113.64, 120.43, 126.35, 126.55, 128.40 (3C), 128.66 (3C), 129.45 (3C), 141.64, 142.04, 142.10, 159.75. The product was converted into the corresponding hydrochloride salt; mp: 197–199 °C. Anal. (C26H29NO3·HCl) C, H, N.</p><!><p>Amine 3 (68 mg, 0.24 mmol) was reacted with 6-methoxynicotinaldehyde (40 mg, 0.29 mmol), glacial acetic acid (12 µL, 0.21 mmol), and NaCNBH3 (92 mg, 0.43 mmol) in 1,2-dichloroethane (6 mL) using procedure A. The residue was purified by column chromatography using 3% methanol in ethyl acetate to afford compound 4g (D-576) (51 mg, 51%) as a thick syrup, [α]D25=(−)44.7°, c = 1 in MeOH. 1H NMR (500 MHz, CDCl3): δ 8.04 (s, 1 H), 7.55 (dd, J = 2.1, 8.6 Hz, 1 H), 7.06–7.40 (m, 10 H), 6.69 (d, J = 8.6 Hz, 1 H), 4.42–4.55 (m, 1 H), 3.85–4.00 (m, 4 H), 3.74–3.84 (m, 2 H), 3.66 (d, J = 13.1 Hz, 1 H), 2.42 (br s, 1 H), 1.65–1.76 (m, 1 H), 1.37–1.47 (m, 1 H). 13C NMR (125 MHz, CDCl3): δ 163.5, 146.1, 142.0, 141.9, 139.0, 128.6(3C), 128.4 (3C), 128.3 (3C), 126.5, 126.3, 110.7, 73.5, 67.4, 64.7, 56.7, 56.2, 53.4, 48.1, 33.4.The product was converted into the corresponding hydrochloride salt; mp: 190–195 °C. Anal. Calcd for [C25H28N2O3·2HCl·0.2H2O] C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 3-fluoro-4-methoxybenzaldehyde (36 mg, 0.23 mmol), glacial acetic acid (16 µL, 0.27 mmol), and NaCNBH3 (80 mg, 0.36 mmol) in 1,2-dichloroethane (3 mL) using procedure A. The residue was purified by column chromatography by using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 4h (D-580) as colorless syrup (60 mg, 68%). 1H NMR (500 MHz, CDCl3): δ 7.12–7.39 (m, 10 H), 7.07 (dd, J = 12.2, 1.8 Hz, 1 H), 6.97 (d, J = 8.2 Hz, 1 H), 6.87 (t, J = 8.6 Hz, 1 H), 4.49 (dt, J = 10.4, 2.1 Hz, 1 H), 3.87–3.98 (m, 3 H), 3.86 (s, 3 H0, 3.80 (d, J = 13.4 Hz, 1 H), 3.75 (d, J = 11.9 Hz, 1 H), 3.63 (d, J = 13.1 Hz, 1 H), 2.41 (s, 1 H), 1.85 (br s, 1 H), 1.66–1.76 (m, 1 H), 1.37–1.47 (m, 1 H). 13C NMR (125 MHz, CDCl3): δ 153.4, 151.4, 146.6, 146.5, 142.1, 142.0, 133.5, 133.4, 128.6, 128.4, 128.3, 126.5, 126.3, 123.6, 123.5, 115.8, 115.7, 113.3, 73.6, 67.5, 64.8, 56.6, 56.4, 56.3, 50.4, 33.4. [α]D25=(−)53.9°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 190–195 °C. Anal. Calcd for [C26H28FNO3·HCl ·H2O] C, H, N.</p><!><p>Amine 3 (60 mg, 0.21 mmol) was reacted with 2-fluoro-4-methoxybenzaldehyde (36 mg, 0.23 mmol), glacial acetic acid (16 µL, 0.27 mmol), and Na(OAc)3BH (80 mg, 0.36 mmol) in a mixture of 1,2-dichloroethane (3 mL) and methanol (1 mL) by following Procedure A. The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 4i (D-581) as a colorless syrup (60 mg, 68%). 1H NMR (500 MHz, CDCl3): δ 7.11–7.38 (m, 11 H), 6.64 (dd, J = 8.6, 2.1 Hz, 1 H), 6.58 (dd, J = 11.9, 2.4 Hz, 1 H), 4.48 (dt, J = 10.1, 2.4 Hz, 1 H), 3.91–3.98 (m, 2 H), 3.88 (dd, J = 11.9, 2.1 Hz, 1 H), 3.80 (d, J = 13.4 Hz, 1 H), 3.76 (s, 3 H), 3.68–3.75 (m, 2 H), 2.42 (m, 1 H), 1.86 (s, 1 H), 1.65–1.78 (m, 1 H), 1.35–1.46 (m, 1 H). 13C NMR (125 MHz, CDCl3): δ 162.5, 160.6, 160.0, 159.9, 142.1, 142.0, 130.7, 130.6, 128.6, 128.4, 128.3, 128.3, 126.5, 126.3, 119.0, 118.9, 109.8, 109.7, 101.6, 101.4, 73.6, 67.3, 65.0, 56.7, 56.4, 55.5, 44.4, 33.4. [α]D25=(−)54.1°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 170–175 °C. Anal. Calcd for [C26H28FNO3·HCl·H2O] C, H, N.</p><!><p>Amine 5 (50 mg, 0.16 mmol) was reacted with benzofuran-5-carbaldehyde (27 mg, 0.19 mmol), glacial acetic acid (13 µL, 0.21 mmol), and NaCNBH3 (17 mg, 0.27 mmol) in a mixture of 1,2-dichloroethane (4.5 mL) and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6a (D-508) (55 mg, 81%) as a light yellow syrup. 1H NMR (500 MHz, CDCl3): δ 7.61 (d, J = 2.1 Hz, 1 H), 7.51 (s, 1 H), 7.43 (d, J = 8.5 Hz, 1 H), 7.25–7.30 (m, 2 H), 7.23 (d, J = 8.5 Hz, 1 H), 7.11–7.18 (m, 2 H), 6.91–7.03 (m, 4 H), 6.69–6.75 (m, 1 H), 4.34–4.44 (m, 1 H), 3.97–4.03 (m, 4 H), 3.81 (d, J = 12.8 Hz, 2 H), 2.48 (s, 1 H), 1.65–1.77 (m, 1 H), 1.40 (d, J = 14.4 Hz, 1 H). 13C NMR (125 MHz, CDCl3): δ 162.4, 160.5, 154.2, 145.3, 137.7, 137.4, 134.5, 129.9, 129.8, 129.7, 129.6, 127.5, 124.6, 120.5, 115.5, 115.3, 115.2, 115.1, 111.2, 106.4, 73.5, 67.4, 64.9, 56.3, 54.9, 51.4, 33.2. [α]D25=(−)39.3°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 160–165 °C. Anal. Calcd for [C27H25F2NO3·HCl·H2O] C, H, N.</p><!><p>Amine 5 (60 mg, 0.19 mmol) was reacted with 3-hydroxy-4-methoxybenzaldehyde (34 mg, 0.23 mmol), glacial acetic acid (17 µL, 0.28 mmol), and NaCNBH3 (20 mg, 0.32 mmol) in a mixture of 1,2-dichloroethane (4.5 mL)and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6b (D-527) (65 mg, 76%) as a colorless syrup. 1H NMR (400 MHz, CDCl3): δ 7.26 (dd, J = 8.2, 5.6 Hz, 2 H), 7.14 (dd, J = 8.5, 5.6 Hz, 2 H), 6.89–7.20 (m, 4 H), 6.78–6.88 (m, 2 H), 6.71–6.77 (m, 1 H), 4.34–4.46 (m, 1 H), 4.0–4.08 (m, 1 H), 3.76–3.97 (m, 7 H), 3.67 (d, J = 12.9 Hz, 1 H), 2.51 (s, 1 H), 1.64–1.77 (m 1 H), 1.37–1.48 (m, 1 H). 13C NMR (100 MHz, CDCl3): δ 162.7, 160.3, 146.6, 144.9, 137.6, 137.4, 130.6, 129.9, 129.8, 129.7, 129.6, 121.1, 115.6, 115.4, 115.3, 115.1, 114.3, 110.9, 73.6, 66.8, 64.3, 56.3, 55.8, 54.8, 50.9, 33.1. [α]D25=(−)28.7°, c = 1 in MeOH. The product was converted into the corresponding hydrochloride salt; mp: 190–195 °C. Anal. Calcd for [C26H27F2NO4·HCl·0.8 H2O] C, H, N.</p><!><p>Amine 5 (50 mg, 0.16 mmol) was reacted with 2,3-dihydrobenzofuran-5-carbaldehyde (28 mg, 0.19 mmol), glacial acetic acid (13 µL, 0.21 mmol), and NaCNBH3 (17 mg, 0.27 mmol) in a mixture of 1,2-dichloroethane (4.5 mL) and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6c (D-507) (56 mg, 79%) as a light yellow syrup. 1H NMR (400 MHz, CDCl3): δ 7.23–7.31 (m, 2 H), 7.10–7.19 (m, 3 H), 6.89–7.03 (m, 5 H), 6.71 (d, J = 8.1 Hz, 1 H), 4.55 (t, J = 8.8 Hz, 2 H), 4.33–4.47 (m, 1 H), 3.95–4.03 (m, 1 H), 3.85–3.94 (m, 2 H), 3.74–3.83 (m, 2 H), 3.63 (d, J = 12.7 Hz, 1 H), 3.17 (t, J = 8.8 Hz, 2 H), 2.46 (d, J = 2.2 Hz,1 H), 1.63–1.74 (m, 1 H), 1.36–1.46 (m, 1 H). 13C NMR (100 MHz, CDCl3): δ 162.7, 160.2, 159.2, 137.7, 137.4, 131.8, 129.9, 129.8, 129.7, 129.6, 127.9, 127.2, 124.8, 115.5, 115.3, 115.3, 115.1, 108.9, 73.5, 71.2, 67.2, 64.7, 56.2, 54.8, 51.0, 33.2, 29.6. [α]D25=(−)39.8°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 145–150 °C. Anal. Calcd for [C27H27F2NO3·HCl·0.5H2O] C, H, N.</p><!><p>Amine 5 (50 mg, 0.16 mmol) was reacted with 3-hydroxy-5-methoxybenzaldehyde (29 mg, 0.19 mmol), glacial acetic acid (13 µL, 0.21 mmol), and Na(OAc)3BH (125 mg, 0.59 mmol) in a mixture of 1,2-dichloroethane (4.5 mL) and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6d (D-526) (65 mg, 76%) as a colorless syrup. 1H NMR (400 MHz, CDCl3): δ 7.17–7.24 (m, 2 H), 7.02–7.12 (m, 2 H), 6.83–6.98 (m, 4 H), 6.42 (s, 1 H), 6.34 (s, 1 H), 6.26 (s, 1 H), 4.60 (br s, 1 H), 4.37 (t, J = 9.1 Hz, 1 H), 4.07 (s, 1 H), 3.82–3.94 (m, 2 H), 3.71–3.81 (m, 2 H), 3.67 (s, 3 H), 3.63 (d, J = 13.2 Hz, 1 H), 2.56 (s, 1 H), 1.65 (t, J = 11.5 Hz, 1 H), 1.39 (d, J = 14.9 Hz, 1 H). 13C NMR (100 MHz, CDCl3): δ 162.7, 161.1, 160.2,, 157.8, 139.1, 137.3 (2C), 129.7 (4C), 115.6, 115.4 (2C), 115.1, 108.1, 106.6, 101.1, 73.9, 65.7, 63.4, 56.2, 55.2, 54.5, 50.4, 32.9. [α]D25=(−)46.4°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 150–155 °C. Anal. Calcd for [C26H27F2NO4·HCl·0.5H2O] C, H, N.</p><!><p>Amine 5 (40 mg, 0.13 mmol) was reacted with benzo[d][1,3]dioxole-5-carbaldehyde (23 mg, 0.15 mmol), glacial acetic acid (13 µL, 0.21 mmol), and NaCNBH3 (14 mg, 0.22 mmol) in a mixture of 1,2-dichloroethane (4.5 mL) and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6e (D-506) (47 mg, 84%) as a colorless syrup.</p><p>1H NMR (400 MHz, CDCl3): δ 7.23–7.32 (m 2 H), 7.11–7.19 (m, 2 H), 6.90–7.02 (m, 4 H), 6.82 (s, 1 H), 6.67–6.78 (m, 2 H), 5.92 (s, 2 H), 4.29–4.46 (m, 1 H), 3.84–4.0 (m, 3 H), 3.70–3.82 (m, 2 H), 3.61 (d, J = 12.9 Hz, 1 H), 2.42 (d, J = 2.4 Hz, 1 H), 1.99 (br s, 1 H), 1.61–1.73 (m, 1 H), 1.32–1.42 (m, 1 H). 13C NMR (100 MHz, CDCl3): δ 162.7, 160.2, 147.7, 146.5, 137.7, 137.4, 133.9, 129.9, 129.8, 129.7, 129.6, 121.0, 115.5, 115.3, 115.3, 115.0, 108.5, 108.0, 100.9, 73.4, 67.2, 64.8, 56.1, 54.9, 51.0, 33.2. [α]D25=(−)23.9°, c = 1 in MeOH. The product was converted into the corresponding hydrochloride salt; mp: 150–155 °C. Anal. Calcd for [C26H25F2NO4·HCl·0.3H2O] C, H, N.</p><!><p>Amine 5 (126 mg, 0.40 mmol) was reacted with 3-methoxybenzaldehyde (31 mg, 0.23 mmol), glacial acetic acid (16 µL, 0.26 mmol), and NaCNBH3 (20 mg, 0.32 mmol) in a mixture of 1,2-dichloroethane (4.5 mL)and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6f (D-524) (55 mg, 66%) as a colorless syrup.1H NMR (400 MHz, CDCl3): δ 7.10–7.36 (m, 5 H), 6.83–7.03 (m, 6 H), 6.75–6.83 (m, 1 H), 4.33–4.45 (m, 1 H), 3.65–4.01 (m, 9 H), 2.45 (s, 1 H), 1.98 (br s, 1 H), 1.64–1.77 (m, 1 H), 1.39–1.48 (m, 1 H). 13C NMR (100 MHz, CDCl3): δ 162.7, 160.2, 159.7, 141.6, 137.7, 137.4, 129.8, 129.8, 129.7, 129.6, 129.4, 120.3, 115.5, 115.3, 115.3, 115.1, 113.6, 112.3, 73.5, 67.3, 64.9, 56.3, 55.2, 54.9, 51.2, 33.2. [α]D25=(−)77.6°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 190–195 °C. Anal. Calcd for [C26H27F2NO3·HCl] C, H, N.</p><!><p>Amine 5 (60 mg, 0.19 mmol) was reacted with 6-methoxynicotinaldehyde (31 mg, 0.23 mmol), glacial acetic acid (13 µL, 0.22 mmol), and NaCNBH3 (22 mg, 0.34 mmol) in a mixture of 1,2-dichloroethane (4.5 mL)and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6g (D-537) (60 mg, 72%) as colorless syrup. 1H NMR (500 MHz, CDCl3): δ 8.03 (d, J = 2.1 Hz, 1 H), 7.55 (dd, J = 2.4, 8.6 Hz, 1 H), 7.23–7.31 (m, 2 H), 7.09–7.18 (m, 2 H), 6.88–7.02 (m, 4 H), 6.70 (d, J = 8.2 Hz, 1 H), 4.34–4.43 (m, 1 H), 3.85–3.99 (m, 6 H), 3.75–3.83 (m, 2 H), 3.64 (d, J = 13.1 Hz, 1 H), 2.42 (s, 1 H), 2.10 (br s, 1 H), 1.62–1.73 (m, 1 H), 1.35–1.45 (m, 1 H). 13C NMR (125 MHz, CDCl3): δ 163.54, 162.4, 160.5, 146.1, 139.0, 137.6, 137.4, 129.9, 129.8, 129.7, 129.6, 127.9, 115.6, 115.4, 115.3, 115.1, 110.7, 73.5, 67.1, 64.7, 56.1, 55.0, 53.4, 48.0, 33.2. [α]D25=(−)85.0°, c = 1 in CH2Cl2. The product was converted into the corresponding hydrochloride salt; mp: 180–185 °C. Anal. Calcd for [C25H26F2N2O3·2HCl·0.1C4H10O] C, H, N.</p><!><p>Amine 5 (50 mg, 0.16 mmol) was reacted with 3-fluoro-4-hydroxybenzaldehyde (29 mg, 0.20 mmol), glacial acetic acid (16 µL, 0.27 mmol), and NaCNBH3 (17 mg, 0.27 mmol) in a mixture of 1,2-dichloroethane (4.5 mL)and methanol (1.5 mL). The residue was purified by gradient silica gel column chromatography using a mixture of dichloromethane and methanol (100:1 to 6:1) to afford corresponding compound 6h (D-531) (50 mg, 80%) as a colorless semi solid. 1H NMR (500 MHz, CDCl3): δ 7.22–7.32 (m, 2 H), 7.09–7.18 (m, 2 H), 6.84–7.02 (m, 5 H), 6.80 (d, J = 7.9 Hz, 1 H), 6.68 (t, J = 8.5 Hz, 1 H), 5.0 (br s, 1 H), 4.43 (t, J = 9.6 Hz, 1 H), 4.11 (s, 1 H), 3.98 (d, J = 11.3 Hz, 1 H), 3.93 (d, J = 8.9 Hz, 1 H), 3.87 (d, J = 12.2 Hz, 1 H), 3.81 (d, J = 12.5 Hz, 1 H), 3.63 (d, J = 12.5 Hz, 1 H), 2.62 (s, 1 H), 1.62–1.80 (m, 1 H), 1.44 (d, J = 14.3 Hz, 1 H). 13C NMR (125 MHz, CDCl3): δ 162.6, 160.5,, 152.3, 150.4, 143.8,, 137.4, 137.3, 129.9, 129.8, 129.7, 129.6,, 124.9, 118.2, 116.1,, 115.6, 115.4, 115.3, 115.2, 73.8, 65.9, 63.3, 56.2, 54.8, 49.9, 32.9. [α]D25=(−)41.0°, c = 1 in MeOH. The product was converted into the corresponding hydrochloride salt; mp: 150–155 °C. Anal. Calcd for [C25H24F3NO3·HCl·0.2H2O·0.6C4H10O] C, H, N.</p><!><p>The ability of test compounds to inhibit substrate uptake by monoamine transporters in synaptosome-enriched fractions from rat brain was monitored exactly as described by us previously.40, 41 [3H]DA ([ring 2,5,6-3H]dopamine (45.0 Ci/mmol, Perkin-Elmer, Boston, MA, U.S.A) was used for monitoring DAT (rat striatum) and NET (rat cerebral cortex). Regarding the latter, it is worth mentioning that DA is an excellent substrate for NET (for our previous discussion see Santra et al., 2012=ref 40); the use of [3H]DA instead of [3H]norepinephrine for rat NET greatly reduced nonspecific uptake, and control experiments with a number of test compounds did not detect significant differences between Ki values obtained with [3H]dopamine and [3H]norepinephrine. [3H]5-HT ([1,2-3H]serotonin (27.9 Ci/mmol, Perkin-Elmer) was the radioligand for monitoring SERT (rat cerebral cortex).</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>Supporting Information Available: Elemental analysis data for all final targets is available. This material is available free of charge via the internet.</p>

PubMed Author Manuscript

Structure of (5\xe2\x80\xb2S)-8,5\xe2\x80\xb2-Cyclo-2\xe2\x80\xb2-Deoxyguanosine in DNA

Diastereomeric 8,5\xe2\x80\xb2-cyclopurine 2\xe2\x80\xb2-deoxynucleosides, containing a covalent bond between the deoxyribose and the purine base, represent an important class of DNA damage induced by ionizing radiation. The 8,5\xe2\x80\xb2-cyclo-2\xe2\x80\xb2-deoxyguanosine lesion (cdG) has been recently reported to be a strong block of replication and highly mutagenic in Escherichia coli. The 8,5\xe2\x80\xb2-cyclopurine-2\xe2\x80\xb2-deoxyriboses are suspected to play a role in the etiology of neurodegeneration in xeroderma pigmentosum patients. These lesions cannot be repaired by base excision repair, but they are substrates for nucleotide excision repair. The structure of an oligodeoxynucleotide duplex containing a site-specific S-cdG lesion placed opposite dC in the complementary strand was obtained by molecular dynamics calculations restrained by distance and dihedral angle restraints obtained from NMR spectroscopy. The S-cdG deoxyribose exhibited the O4\xe2\x80\xb2-exo (west) pseudorotation. Significant perturbations were observed for the \xce\xb2, \xce\xb3, and \xcf\x87 torsion angles of the S-cdG nucleoside. Watson-Crick base pairing was conserved at the S-cdG\xe2\x80\xa2dC pair. However, the O4\xe2\x80\xb2-exo pseudorotation of the S-cdG deoxyribose perturbed the helical twist and base pair stacking at the lesion site and the 5\xe2\x80\xb2-neighbor dC\xe2\x80\xa2dG base pair. Thermodynamic destabilization of the duplex measured by UV melting experiments correlated with base stacking and structural perturbations involving the modified S-cdG\xe2\x80\xa2dC and 3\xe2\x80\xb2- neighbor dT\xe2\x80\xa2dA base pairs. These perturbations may be responsible for both the genotoxicity of this lesion and its ability to be recognized by nucleotide excision repair.

structure_of_(5\xe2\x80\xb2s)-8,5\xe2\x80\xb2-cyclo-2\xe2\x80\xb2-deoxyguanosine_in_dna

Introduction<!>Synthesis and Characterization of the S-cdG-Modfied Oligodeoxynucleotide<!>NMR Resonance Assignments<!>Deoxyribose Coupling Constants<!>Phosphodiester Backbone Conformation<!>Chemical Shift Perturbations<!>Thermal Stability of the S-cdG Modified Duplex<!>Structural Refinement<!>Structure of the S-cdG-Containing Duplex<!>Molecular Dynamics Calculations in Explicit Solvent<!>Discussion<!>Structure of S-cdG in DNA<!>Thermodynamic Considerations<!>(a) DNA Repair<!>(b) Error-Prone Replication Bypass<!>Conclusions<!>(a) N2-((Dimethylamino) methylene)-2\xe2\x80\xb2-deoxyguanosine (1)<!>(b) N2-DMF-5\xe2\x80\xb2-phenylthio-2\xe2\x80\xb2,5\xe2\x80\xb2-dideoxyguanosine (2)<!>(c) N2-DMF-5\xe2\x80\xb2,8-cyclo-2\xe2\x80\xb2,5\xe2\x80\xb2-dideoxyguanosine (3)<!>(d) N2-DMF-3\xe2\x80\xb2-O-(tert-butyldimethylsilyl)-5\xe2\x80\xb2,8-cyclo-2\xe2\x80\xb2,5\xe2\x80\xb2-dideoxyguanosine (4)<!>(e). N2-Isobutyryl-3\xe2\x80\xb2-O-(tert-butyldimethylsilyl)-5\xe2\x80\xb2,8-cyclo-2\xe2\x80\xb2,5\xe2\x80\xb2-dideoxyguanosine (5)<!>(f). (5\xe2\x80\xb2S)-N2-Isobutyryl-3\xe2\x80\xb2-O-(tert-butyldimethylsilyl)-5\xe2\x80\xb2,8-cyclo-2\xe2\x80\xb2-deoxyguanosine (7)<!>(g). (5\xe2\x80\xb2S)-N2-Isobutyryl-3\xe2\x80\xb2-O-(tert-butyldimethylsilyl)-5\xe2\x80\xb2-O-(4,4\xe2\x80\xb2-dimethoxytrityl)-5\xe2\x80\xb2,8-cyclo-2\xe2\x80\xb2-deoxyguanosine (8)<!>(h). (5\xe2\x80\xb2S)-N2-Isobutyryl-5\xe2\x80\xb2-O-(4,4\xe2\x80\xb2-dimethoxytrityl)-5\xe2\x80\xb2,8-cyclo-2\xe2\x80\xb2-deoxyguanosine<!>(i). (5\xe2\x80\xb2S)-5\xe2\x80\xb2,8-Cyclo-2\xe2\x80\xb2-deoxyguanosine Phosphoramidite Derivative<!>Oligodeoxynucleotides<!>Melting Temperature<!>NMR<!>Distance and Dihedral Angle Restraints<!>Molecular Dynamics Calculations

<p>Hydroxyl radicals cause a variety of damage in DNA, affecting the nucleobases1 or deoxyribose sugars,2 or both,3 as in the case of tandem 8,5′-cyclopurine 2′-deoxynucleoside lesions.4 At 2′-deoxyguanosines in DNA, hydrogen abstraction by a hydroxyl radical at the C5′ position of the deoxyribose followed by attack at the C8 carbon of guanine forms an N7-centered radical, which may be oxidized to produce diastereomeric 8,5′-cyclo-2′-deoxyguanosines (cdG).4–10 The corresponding 8,5′-cyclo-2′-deoxyadenosines (cdA) have also been characterized.3,4,7,9–15 For both cdG and cdA, the diastereomeric ratio at the C5′ position depends on experimental conditions and DNA conformation.5–7,13,16–18</p><p>The 8,5′-cyclopurine-2′-deoxynucleosides are believed to be important contributors to the genetic toxicology of oxidative stress and inflammation.4 They have been detected at the nucleotide level,5,11 in DNA,5,19–21 and cells in vitro,6 in human urine,18 and in vivo.21–23 The formation of 8,5′-cyclopurine-2′-deoxynucleosides might contribute to skin cancer risk in xeroderma pigmentosum complementation group C (XP-C) patients.24 They are also believed to play roles in Cockayne syndrome,21 breast and ovarian cancer,22 and familial Mediterranean fever.25</p><p>In Escherichia coli, S-cdG is a block to DNA replication, is highly mutagenic, and is refractory to repair.26 It induced 34% mutation upon induction of the SOS response. Most mutations were S-cdG → A mutations, though S-cdG → T mutations and deletions of the 5′-neighbor dC at low level also were observed.26 It has been reported in a preliminary study that S-cdG does not block primer elongation by Klenow DNA polymerases, and dATP is preferentially incorporated opposite the lesion.27</p><p>Computational studies predicted that the incorporation of the cdA stereoisomers into DNA would result in helical distortions at the lesion site.28–30 Both the R- and S-diastereomers of the 8,5′-cA ribonucleoside have been crystallized.31,32 Both exhibited the anti conformation about the glycosyl bond with χO4′-C1′-N9-C8 = 29.8° or 27.4°, respectively. The fused six-member ring C8-N9-C1′-O4′-C4′-C5′ adopted the half-chair conformation with the O4′ and C4′ out of plane. The deoxyribose adopted the O4′-exo (0T1) pseudorotation with P = ~289° and τm = ~48°. Molecular mechanics calculations predicted that the cdA diasetereomers maintain the O4′-exo pseudorotation when placed opposite dT in DNA.28 The NMR data and ab initio calculations suggest that incorporation of the S-cdA into di- or tri-nucleotides does not change the O4′-exo deoxyribose pseudorotation.33</p><p>Herein, we report the structure of the S-cdG•dC pair in 5′-d(GTGCXTGTTTGT)-3′•5′-d(ACAAACACGCAC)-3′, containing the DNA sequence of p53 codons 272–275, where X denotes the S-cdG (Scheme 1). The lesion is located in codon 273. The S-cdG remains stacked into the duplex and participates in Watson-Crick hydrogen bonding with the complementary dC. However, the S-cdG deoxyribose shifts to the O4′-exo pseudorotation with P = 280°. This alters the γ and δ backbone torsion angles. Additionally, the β and χ torsion angles are changed from those in B-DNA. The twist and base pair shift helicoidal parameters are perturbed at the C4•G21 and X5•C20 base pairs. The purine ring is anti about the glycosyl bond and the fused six-membered ring adopts the half-chair conformation with O4′ and C4′ out of plane.</p><!><p>The S-cdG -modified oligodeoxynucleotide 5′-GTGCXTGTTTGT-3′, containing the sequence of p53 codons 272-275 in which the lesion was located in codon 273, was synthesized by a modification of the method reported by Romieu (Scheme S1 in the Supporting Information).34 The synthesis of N2-isobutyryl-5′-phenylthio-2′,5′-dideoxyguanosine gave 70% yield from N2-isobutyryl-2′-deoxyguanosine. The yield was improved to 91% when the exocyclic N2-dG amino group was protected with DMF. However, the DMF protection was unstable in the subsequent NaBH4 reduction step. Therefore, after cyclization and TBDMS protection of the 3′-hydroxyl group, it was replaced with an isobutyryl group. The modified oligodeoxynucleotide was synthesized using solid phase phosphoramidite chemistry and characterized by HPLC and MALDI-TOF mass spectrometry.</p><!><p>The non-exchangeable protons of the S-cdG-modified duplex were assigned based upon the sequential connectivity of the base proton H6 or H8 dipolar couplings with H1′ deoxyribose protons (Figure 1).35,36 For the modified strand, the NOE sequential connectivity was observed from G1 to C4. Since the S-cdG nucleotide lacked a proton at the C8 carbon, the sequential connectivity exhibited an interruption at X5. However, the X5 H1′ proton was identified at 6.14 ppm; it exhibited a weak X5 H1′→T6 H6 NOE, suggesting that the distance between these two protons was greater than in B-DNA. The sequential NOE connectivity resumed from T6 to T12. For the modified strand, all of the deoxyibose H1′ protons were observed within a narrow chemical shift window, between 5.8–6.3 ppm. The complete sequential NOE connectivity was observed for the complementary strand.</p><p>The assignments of X5 deoxyribose protons were made by analysis of scalar and dipolar couplings. Figure 2 displays a tile plot derived from a NOESY spectrum obtained at 60 ms mixing time. X5 H1′ exhibited strong dipolar couplings with H2′ and H2″; weak scalar couplings were also observed. H3′ exhibited strong dipolar couplings with H2′, H2″ and H4′, whereas the scalar couplings were unobservable. H4′ exhibited both scalar and dipolar couplings with the single H5′ proton. The geminal H2′ and H2″ protons were assigned from their NOEs to H1′ and H3′. H2′ exhibited a weaker NOE with H1′ than did H2″, whereas it exhibited a stronger NOE with H3′ than did H2″. In B-DNA, H2″ resonances are usually more downfield than H2′ resonances. However, the X5 H2′ resonance was observed at 2.55 ppm, whereas the H2″ resonance was observed at 2.27 ppm. For the remainder of the duplex, the H2′, H2″, H3′, and H4′ deoxyribose resonances were assigned unequivocally. The resonance assignments of the non-exchangeable DNA protons are tabulated in Table S1 of the Supporting Information.</p><p>The resonances of the base imino protons were assigned based on sequential connectivity in NOESY spectra and the assignments were supported by NOEs to the amino protons of Watson-Crick base pairs (Figure 3).37 The NOE sequential connectivity was observed from G1→T2→G3→G21 to X5, and from G7→T8→T9→T10 to G11. At and adjacent to the lesion site, G21 N1H exhibited NOEs with C4 N4H1 and N4H2, and X5 N1H exhibited NOEs with the complementary C20 N4H1 and N4H2. At the 3′-neighbor base pair, the T6 N3H resonance was not observed, but A19 H2 exhibited NOEs to both X5 N1H and G7 N1H, suggesting A19 was still intercalated. Except for the terminal base pairs, the remaining NOE cross-peaks arising from Watson-Crick hydrogen bonding were observed.</p><!><p>Figure 4 displays the expansion of an ECOSY spectrum38 in the region of deoxyribose H1′ correlations with H2′ and H2″. The 3JH1′-H2′ and 3JH1′-H2″ coupling constants were measured from the multiplicities of the cross peaks. The 3JH1′-H2′ and 3JH1′-H2″ values for X5 were 2.6 and 7.0 Hz, respectively. Consistently, the H1′-H2′ cross peak was weak. The 3JH4′-H5′ was 5.4 Hz, whereas the 3JH3′-H4′ was not measureable. Except for the terminal nucleotides, the 3JH1′-H2′ for all other nucleotides were 8–10 Hz and the 3JH1′-H2″ were 5–7 Hz, suggesting that the deoxyriboses adopted C1′-exo or C2′-endo conformations. The 3J coupling constants for the deoxyribose protons are tabulated in Table S2 of the Supporting Information.</p><!><p>The 31P resonances were assigned from a 31P-H3′ HMBC spectrum (Figure S2 in the Supporting Information). Except for X5, each phosphodiester exhibited a heteronuclear coupling with H3′ of the 5′-neighbor. Figure 5 displays the 31P NMR of the S-cdG containing duplex compared with the corresponding unmodified duplex. At the modified nucleotide the 31P resonance shifted upfield, indicating a backbone perturbation at the modified base. The other 31P resonances were clustered within a modest chemical shift range, centered in the spectral region characteristic of B DNA.</p><!><p>Chemical shifts of the non-exchangeable protons between the S-cdG containing duplex and the corresponding unmodified duplex were compared (Figure 6). Significant changes were observed at X5 and the 5′- and 3′-neighboring nucleotides of the modified strand. C4 H6, H1′ and H2″ shifted downfield by 0.21, 0.38, and 0.99 ppm, respectively; X5 H2″ shifted upfield by 0.55 ppm; and T6 H6, CH3, and H1′ shifted downfield by 0.36, 0.31, and 0.22 ppm, respectively. In contrast, the chemical shift perturbations for the complementary strand were small, with the exception of A19 H2′, which shifted upfield by 0.21 ppm.</p><!><p>The thermal melting of the modified duplex containing the S-cdG was monitored using UV spectroscopy in 100 mM NaCl at pH 7.0. It exhibited a melting temperature (Tm) of 46 ± 1 °C, as compared to the unmodified DNA that exhibited a Tm: of 55 °C. (Figure S1 in the Supporting Information). Thus, the incorporation of S-cdG reduced the Tm by 9 °C. Figure 7 displays 1H NMR of the S-cdG containing duplex and the corresponding unmodified duplex at different temperatures. In the modified duplex, the X5 imino resonance exhibited significantly more line broadening at 45 °C than the corresponding G5 imino resonance of the unmodified duplex. For the modified duplex, the T6 N3H resonance was not observed at 5 °C, suggesting that the S-cdG nucleotide also significantly perturbed the 3′-flanking T6•A19 base pair.</p><!><p>A total of 426 distance restraints, including 274 intra-nucleotide and 152 inter-nucleotide restraints were calculated from the intensities of NOE cross-peaks using MARDIGRAS (Table S3 in the Supporting Information).39 A total of 29 NOEs involving the S-cdG protons were used as restraints. A total of 45 empirical distance restraints arising from Watson-Crick base pairing interactions were used, as were 165 empirical torsion angle restraints that were applied to refine the non-terminal nucleotides. These were justified based upon the NMR data, which suggested that structural perturbations for the duplex were localized at and adjacent to the lesion site. No base pair distance restraints were used for the T6•A19 base pair, and no torsion angle restraints were used for the C4•G21, X5•C20 and T6•A19 base pairs. The restraints used for the structure refinement are summarized in Table 1.</p><p>The rMD calculations for the S-cdG containing duplex were performed from initial A- and B-form starting structures. Ten final structures, five each for A- and B-DNA starting structures, with lowest energies, were obtained. All structures converged as indicated by pairwise RMSD comparisons (Table 1). The accuracies of the emergent structures were evaluated by comparison of theoretical NOE intensities calculated by CORMA40 for the refined structure to the experimental NOE intensities to yield sixth root residuals (R1x).41 The overall residuals, as well as the residuals for intra- or inter-nucleotide NOEs, were consistently less than 0.1 (Table 1). R1x values for each nucleotide were less than 0.15 (Figure S3 in the Supporting Information). Thus, the refined structures provided accurate depictions of the NOE data.</p><!><p>The significant perturbations involved the modified strand. Figure 8 shows an expanded view at the lesion site. The S-cdG nucleotide was in the O4′-exo, "west" pseudorotation (Figure 9B), with P = 280.2° and τm = 47.6°. The heavy atoms N9, O3′, and C5′ were axial about the deoxyribose ring. With the exception of the terminal nucleotides, all other deoxyribose pseudorotations were either C1′-exo or C2′-endo. Figure 9A displays the six-membered ring C8-N9-C1′-O4′-C4′-C5′ conformation. It adopted the envelope (half boat) conformation. Helicoidal analysis of the backbone torsion angles showed that at the lesion site, the β (P-O5′-C5′-C4′) angle shifted from the characteristic ~180° to −87°. The γ (O5′-C5′-C4′-C3′) angle shifted from ~ 50° to −67°. Modest perturbations of the δ (C5′-C4′-C3′-O3′) and ζ (C3′-O3′-P-O5′) torsion angles were also observed from ~ 120° to +149° and from ~ −90° to −59°, respectively. There was also a modest change for the glycosyl torsion angle χ from ~ −120° to −157°. C4 H2″ was proximate to the X5 purine ring. In contrast, X5 H2″ was farther from the X5 purine ring compared to the H2″ protons in B DNA.</p><p>Figure 10 shows the base stacking and base pairing at the lesion site. The 5′-neighbor C4•G21 base pair exhibited a shift of −1.0 Å resulting in the displacement of C4 toward the major groove. At the C4→X5 step, an increased twist of 49° with respect to the X5•C20 base pair was evident. In contrast, the helix was underwound at the X5→T6 step. Additionally, the 3′-neighbor base pair T6•A19 exhibited a greater than normal base pair opening of −11.3°.</p><!><p>A molecular dynamics simulation was carried out in explicit water at constant pressure at 300 K, for 5 ns. The distances of the atoms involving in the Watson-Crick hydrogen bonding were measured in the trajectories. Figure 11 shows the distances of guanine N1H → cytosine N3 and the thymine N3H → adenine N1 of some base pairs observed in the trajectories. During this simulation, no changes in the monitored distances were observed for the G•C and C•G base pairs including the damaged X5•C20 base pair. In contrast, at the T6•A19 base pair, an opening occurred at ~0.9 ns, as indicated by the distances of T6 O2 → A19 H2, T6 N3H → A19 N1, and T6 O4 → A19 N6H1 jumping from ~ 3.5 Å to ~ 5.5 Å, ~ 2.0 Å to ~ 3.5 Å, and from ~ 1.8 Å to ~ 2.2 Å, respectively. Other non-terminus T•A base pairs exhibited no remarkable changes.</p><!><p>Interest in the 8,5′-cyclopurine-2′-deoxynucleoside lesions has been piqued by evidence that in mammalian cells 8,5′-cyclo-2′-deoxyadenosine (cdA) diastereomers3,4,7,9–15 are repaired by nucleotide excision repair (NER),42,43 an idea that was suggested earlier,5,6 and not by base excision repair. Also, bacterial DNA N-glycosylases endo III and FpG do not excise S-cdG from DNA, suggesting that like cdA, it also is a substrate for NER.27 Although the repair of S-cdG by the human NER system remains to be determined, Jasti et al.26 demonstrated that in DNA the S-cdG lesion was incised by the UvrABC nuclease of E. coli. The covalent bond between C8 of guanine and C5′ of the deoxyribose in the 8,5′-cyclopurines locks the modified nucleotide in the anti conformation. This is believed to hinder the flipping of the purine ring from the duplex, which is consistent with the observation that the 8,5′-cyclopurine-2′-deoxynucleosides are not repaired by BER.42,43 If not repaired, the S-cdG lesion is mutagenic. In SOS-induced E. coli, a mutation frequency of 34% was observed. Most mutations were S-cdG→A mutations, though S-cdG→T mutation and a deletion of the 5′-neighbor C also was observed.26 Hence, it was of interest to determine the structure of S-cdG in DNA.</p><!><p>The present study reveals that S-cdG remains stacked into the duplex and participates in Watson-Crick hydrogen bonding with the complementary dC. However, the S-cdG deoxyribose shifts to the O4′-exo pseudorotation, as opposed to the "south" pseudorotation (C2′-endo) observed in B-DNA, or the "north" pseudorotation (C3′-endo) in A-DNA.44,45 This corroborates computational studies on 8,5′-cyclopurine-2′-deoxynucleosides.28 Crystal structures of the cA ribonucleoside also exhibited the O4′-exo pseudorotation,31,32 and an NMR and DFT study of di- and tri-deoxynucleotides containing S-cdA indicated the O4′-exo deoxyribose.33 The O4′-exo pseudorotation introduces significant helicoidal perturbation into the modified strand of DNA. This involves changes in the S-cdG phosphodiester backbone torsion angles β (P-O5′-C5′-C4′), γ (O5′-C5′-C4′-C3′), δ (C5′-C4′-C3′-O3′), and ζ (C3′-O3′-P-O5′) from ~ 180° to −87°, from ~ 50° to −67°, from ~ 120° to 149°, and from ~ −90° to −59°, respectively. These changes perturb the helicoidal twist and base pair shift parameters at the C4•G21 and X5•C20 base pairs from ~ 30° to 49° and from ~ 0 Å to −1.0 Å, respectively. These changes are consistent with the upfield shift of the 31P resonance at S-cdG. These conclusions also are consistent with computational studies, which predict that the O4′-exo pseudorotation of the cdA deoxyribose should alter the helical twist parameter for the modified cdA•dT base pair as compared to the flanking base pairs.28 In addition, the modified cdA•dT base pair exhibited an altered base pair shift parameter. The altered ζ backbone torsion angle of S-cdG (−59°) results in the greater than normal base opening of −11.3° for the 3′-neigbhor T6•A19 base pair. (Figure S7 in the Supporting Information). Additionally, the glycosyl torsion angle χ (O4′-C1′-N9-C2) torsion angle is modified from ~ −120° to −157°. This places the six-member ring C8-N9-C1′-O4′-C4′-C5′ into the half-boat conformation. The bond between X5 C8 and C5′ pulls X5 H4′ and H5′ closer to the purine ring as compared to the H4′, and H5″ protons in B-DNA. This is consistent with the downfield chemical shifts of both X5 H4′ and H5′. In contrast, X5 H2″ is farther from the X5 purine ring compared to the H2″ protons in B-DNA, consistent with its upfield shift compared to that in the unmodified duplex.</p><!><p>Energetically, the O4′-exo pseudorotation is disfavored due to the axial orientation of all substituent heavy atoms.28 The helical perturbation of the modified strand associated with the unusual O4′-exo deoxyribose at the lesion site is consistent with the 9 °C decrease in the Tm of the modified duplex as compared to the unmodified control. The destabilization likely involves structural perturbations observed for the modified X5•C20 and 3′-neighbor T6•A19 base pairs, and accompanying base stacking perturbations. Indeed, the X5 imino resonance exhibits increased line broadening at 45 °C as compared to the G5 imino resonance of the unmodified duplex (Figure 7). Exchange-mediated line broadening of DNA imino protons is normally associated with the formation of an open state of the base pair in which the imino proton is freed from its hydrogen bond and is accessible to the base that catalyzes the proton exchange,46–48, 49,50 but S-cdG is locked in the anti conformation about the glycosyl bond and incapable of flipping out of the duplex. It seems possible that if the complementary nucleotide C20 nucleotide flips out, this might facilitate proton exchange by allowing water to enter the duplex to access the X5 imino proton, but more detailed studies of the exchange kinetics of the X5 and neighboring imino protons are warranted.49,50 For the modified duplex, the T6 N3H resonance is not observed, suggesting increased exchange with solvent for the imino proton of the 3′-flanking T6•A19 base pair (Figure 7). This may be a consequence of altered ζ backbone torsion angle of S-cdG, which results in the opening of the 3′-neighbor base pair. While the MD simulations occur on a different timescale than the NMR experiments, in the MD simulations, transient opening of the T6•A19 base pair is predicted (Figure 11). In contrast, the thermal melting experiments (Figure 7) suggest that the 5′-neighbor C4•G21 base pair is more stable with respect to imino proton exchange.</p><!><p>In human global genome NER, the XPC/HR23B complex51–55 is believed to be involved in damage recognition. The XPA protein is also essential for NER. Yang et al.56 reported that it exists as a homodimer either in the free state or as a complex with RPA. For example, it binds to mismatched bubble substrates, including the C8-dG adducts of AF, AAF, and 1-nitropyrene, and the T[6,4]T photoproducts.57 XPA is proposed to be involved in the verification of DNA damage.54,55 It may also recruit repair factors and stabilize repair intermediates since it binds more efficiently to undamaged ds-ssDNA junctions with ssDNA branches,57 intermediate structures found in NER.</p><p>The destabilization of the S-cdG modified duplex and the perturbation of the X5•C20 and T6•A19 base pairs is likely relevant with respect to NER. Thermal destabilization of the duplex is believed to modulate recognition of a diverse group of damages by XPC.54,58–61 From studies of the yeast XPC orthologue Rad4 bound to DNA containing a cyclobutane pyrimidine dimer, Min and Pavletich62 concluded that Rad4 may exploit the destabilization of two base pairs. Interestingly, the 5R-thymine glycol lesion, another substrate for NER, also destabilizes two base pairs in DNA.63 The perturbation of the X5•C20 and T6•A19 base pairs in the S-cdG modified duplex may facilitate extrusion of both C20 and A19 (but not X5) out of the helix, enabling XPC/HR23B to recognize S-cdG prior to recruiting XPA.</p><!><p>The bond between C8 of guanine and C5′ of 2′-deoxyribose locks the N-glycosyl torsion angle of S-cdG in the anti domain. Therefore, during translesion synthesis, an incoming dCTP can form a Watson–Crick base pair, whereas an incoming dTTP might form a wobble pair. The insertion of both dATP and dTTP were noted in pol V-dependent TLS by Jasti et al.26 Significantly, they noted the genotoxicity of the S-cdG lesion, which implied that DNA polymerases have difficulty in bypassing this locked nucleotide. They speculated that accommodation of the S-cdG lesion within the active site of the polymerase likely involves rotational adjustments of the nucleoside around the glycosyl bond. 26 Thus, future structural studies of template•primers containing the S-cdG lesion complexed with error-prone polymerase will be of interest.</p><!><p>The structure of S-cdG has been determined when placed opposite dC in DNA. The S-cdG•dC and the flanking base pairs maintain Watson-Crick hydrogen bonding. However, S-cdG exhibits the O4′-exo deoxyribose pseudorotation in DNA. This introduces significant helicoidal and base stacking perturbations into the duplex. The imino proton of the 3′-neighbor T•A base pair undergoes increased exchange with solvent, whereas the 5′-neighbor C•G base pair is only moderately influenced. Collectively, these structural and thermodynamic perturbations may be important in modulating the recognition of the S-cdG lesion during nucleotide excision repair.</p><!><p>To a suspension of 2′-deoxyguanosine (10 g, 35.06 mmol) in dry methanol (100 mL), N,N-dimethylformamide dimethyl acetal (18.7mL, 140.24 mmol) was added dropwise with vigorous stirring. The mixture was stirred at room temperature under argon for 72 h. The solid product was isolated by filtration, washed with cold methanol and dried. The product was isolated as a white solid in quantitative yield.</p><!><p>N2-DMF-2′-deoxyguanosine 1 (1 g, 3.74 mmol) and diphenyl disulfide (1.63g, 7.48 mmol) were dissolved in 15 mL dry DMF under argon, and PBu3 (1.85mL, 7.48mmol) was slowly added dropwise and the mixture was stirred at room temperature for 6 h. The reaction was monitored by thin layer chromatography (TLC) (90/10 CH2Cl2/MeOH, v/v). The reaction was quenched with 10 mL water and evaporated to a glassy syrupy residue. It was purified by silica gel column chromatography with a step gradient of methanol (0–7%) in DCM as the mobile phase. The product was isolated as white foam (1.42g, yield of 91 %)</p><!><p>Previously crushed N2-DMF-5′,8-cyclo-2′,5′-dideoxyguanosine 2 (1.4g, 3.38 mmol) and triethyl phosphate were added to argon-purged 2 L quartz reactor, dissolved in 1 L dry acetonitrile via sonication. This solution was degassed by bubbling argon for 40 min. The reactor was sealed under argon atmosphere and irradiated at 254 nm UV light for 20 h. The reaction was monitored using TLC (85/15 CHCl3/MeOH, v/v). The solution was evaporated to dryness and the resulting brownish yellow solid was purified by silica gel column chromatography with a step gradient of methanol (0–10%) in CHCl3. The product was isolated as a light yellowish white solid (0.53 g, yield of 52%).</p><!><p>N2-DMF-5′,8-cyclo-2′,5′-dideoxyguanosine 3 (1.4g, 4.6 mmol) and imidazole (1.27g, 18.7 mmol) were dried and dissolved in 20 mL dry DMF. TBDMS-Cl (1.39g, 9.2 mmol) was added to this solution while stirring under nitrogen atmosphere. The reaction mixture was stirred at room temperature for 20 h and monitored by TLC (93/7 CHCl3/MeOH, v/v). The solvent was dried under nitrogen and the resulting semi solid was purified on a silica gel column chromatography with a step gradient of methanol (0–3%) in chloroform. The product was isolated as a white solid (1.35g, yield of 70 %).</p><!><p>N2-DMF-3′-O-(tert-butyldimethylsilyl)-5′,8-cyclo-2′,5′-dideoxyguanosine 4 (1.1g, 2.63 mmol) was dissolved in a mixture of 50 mL methanol and 10 mL 29% aq. ammonia, and stirred overnight at room temperature. The solvents were removed under reduced pressure. The resulting white powder was co-evaporated with 5 mL dry pyridine 3 times. This white solid was dissolved in 12 mL dry pyridine, a few crystals of DMAP was added to it and isobutyryl chloride (0.56mL, 5.26 mmol) was added to it dropwise under nitrogen atmosphere. This reaction mixture was stirred at room temperature for 8 h and monitored by TLC (93/7 CHCl3/MeOH, v/v). The solvent was dried under reduced pressure and the resulting yellow solid was purified by silica gel column chromatography with a step gradient of methanol (0–2%) in DCM. The product was isolated as a white solid (1.0 gm, yield of 88 %).</p><!><p>N2-Isobutyryl-3′-O-(tert-butyldimethylsilyl)-5′,8-cyclo-2′,5′-dideoxyguanosine 5 (1.0 g, 2.3 mmol) was dissolved in 250 mL dry 1,4 dioxane, SeO2 (1.28 g, 11.5 mmol) and the mixture was refluxed for 24 h. The reaction was monitored by TLC (93/7 CHCl3/MeOH, v/v). The hot solution was passed through a celite pad and washed with 20 mL 10% methanol in chloroform. The filtrate was dried under reduced pressure to produce brownish white powder of N2-isobutyryl-3′-O-(tert-butyldimethylsilyl)-5′-oxo-5′,8-cyclo-2′-deoxyguanosine. This product was added to 50 mL methanol and NaBH4 (0.174g, 4.6 mmol) was added to it in 3 portions. This reaction mixture was stirred at room temperature for 1 hr and the reaction was monitored by TLC (93/7 CHCl3/MeOH, v/v). The excess borohydride was neutralized by addition of 1N HCl dropwise to the solution. The solution was passed through a celite pad and evaporated to dryness. The resulting yellow solid was purified by silica gel column chromatography with a step gradient of methanol (0–5%) in chloroform. The product was isolated as a white solid (0.37 g, yield of 36%).</p><!><p>(5′S)-N2-Isobutyryl-3′-O-(tert-butyldimethylsilyl)-5′,8-cyclo-2′-deoxyguanosine 6 (0.4 g, 0.89 mmol) was dissolved in 3 mL dry pyridine and evaporated to dryness. This process was repeated twice. The residual white solid was dissolved in 10 mL dry pyridine and DMT-Cl (0.92 g, 2.7 mmol) and a few crystals of DMAP were added to it. The mixture was heated at 80°C and stirred under nitrogen atmosphere for 8 h. It was monitored by TLC (94/5/1 CHCl3/MeOH/NEt3, v/v). The solution was cooled to ~5°C on an ice bath and quenched with methanol. The solvents were removed under reduced pressure and the resulting yellow solid was purified by silica gel column chromatography with a step gradient of methanol (0–1%) in chloroform containing 1% TEA. The product was isolated as a white solid (0.39 g, yield of 58 %).</p><!><p>(5′S)-N2-Isobutyryl-3′-O-(tert-butyldimethylsilyl)-5′-O-(4,4′-dimethoxytrityl)-5′,8-cyclo-2′-deoxyguanosine 7 (0.3 g, 0.40 mmol) was dissolved in 15 mL of dry THF and a solution of 1M TBAF in THF (.8 mL, 0.8 mmol) was added. The reaction was stirred under nitrogen atmosphere for 5 h and monitored by TLC (92/7/1 CHCl3/MeOH/NEt3, v/v). The solvents were removed under reduced pressure and the resulting yellow solid was purified by silica gel column chromatography with a step gradient of methanol (0–2%) in chloroform containing 1% TEA. The product was isolated as a white solid (0.24 g, yield of 95%).</p><!><p>(5′S)-N2-Isobutyryl-5′-O-(4,4′-dimethoxytrityl)-5′,8-cyclo-2′-deoxyguanosine (from the previous reaction) (0.092 mg, 0.14 mmol) was dissolved in dry dichloromethane and evaporated to dryness. This process was repeated twice. The solid was dissolved in 5 mL of dry dichloromethane and kept under argon. Diisopropylethylamine (51 μl, 0.29 mmol) was added to it, then 2-Cyanoethyl N,N-diisopropylchlorophosphoramidite (34 μl, 0.15 mmol) was added to the stirring solution dropwise. The reaction was checked by TLC (95/4/1 CHCl3/MeOH/NEt3, v/v). After 1 hr the solution was cooled to ~5°C with an ice bath, 51 μl of DIEA and 0.2 mL of methanol was added to it. The solvents were removed under reduced pressure and the resulting light yellow semi solid was purified twice on a silica gel column chromatography with a step gradient of methanol (0–1%) in chloroform containing 1% TEA. The product was isolated as a white solid (0.084 g, yield of 70%).</p><!><p>The 5′-d(GTGCGTGTTTGT)-3′ and 5′-d(ACAAACACGCAC)-3′ were synthesized and purified by anion-exchange chromatography by the Midland Certified Reagent Co. (Midland, TX). The dodecamer containing the S-cdG 5′-d(GTGCXTGTTTGT)-3′, where X represents the S-cdG, was synthesized, purified, and characterized using a slightly amended procedure of the synthesis reported by Romieu et al.34 The purity of the modified oligodeoxynucleotide was assessed by HPLC and mass spectrometry. Oligodeoxynucleotides were desalted by chromatography on Sephadex G-25. The 5′-d(GTGCGTGTTTGT)-3′ or 5′-d(GTGCXTGTTTGT)-3′ were annealed with the complementary strand 5′-d(ACAAACACGCAC)-3′ in buffer containing 10 mM NaH2PO4, 100 mM NaCl, and 50μM Na2EDTA (pH 7.0), respectively. The resulting duplexes were heated to 95 °C for 10 min, and cooled to room temperature. They were purified by DNA Grade hydroxylapatite chromatography using a gradient from 10 to 200 mM NaH2PO4 in 100 mM NaCl, 50 μM Na2EDTA (pH 7.0), and desalted using Sephadex G-25.</p><!><p>Melting temperatures of the DNA duplexes were measured in 10 mM NaH2PO4, 100 mM NaCl, 50 μM EDTA (pH 7.0) by UV/vis spectroscopy at 260 nm. The strand concentration was 10 μM. The thermal scan proceeded from 10 to 80 °C with an interval of 1 °C. The melting temperatures were calculated by differentiating the absorbance profiles.</p><!><p>Samples were at 1.0 mM strand concentration. Samples for the non-exchangeable protons were dissolved in 500 μL in 10 mM NaH2PO4, 100 mM NaCl, 50 μM Na2EDTA (pH 7.0). They were exchanged with D2O and suspended in 280 μL 99.996% D2O. The pH was adjusted with dilute DCl or NaOD. Experiments were performed at 800 MHz. COSY and NOESY spectra were recorded with 512 real data in the t1 dimension and 2048 real data in the t2 dimension. NOESY spectra were zero-filled during processing to create a matrix of 1024 × 1024 real points. NOESY experiments used TPPI quadrature detection64 and mixing times of 60, 150, 200 and 250 ms. The relaxation delay was 1.5 s. The TOCSY mixing time was 80 ms. The temperature was 25 °C. Chemical shifts were referenced to water. Data were processed using TOPSPIN65 and analyzed with the program SPARKY.66 The ECOSY data were recorded with 1024 real data in the t1 dimension and 4096 real data in the t2 dimension.38 The spectrum was zero-filled during process to create a matrix of 2048 × 16384 to increase digital resolution. The temperature was 30 °C. Samples for the observation of exchangeable protons were dissolved in 500 μL of 10 mM NaH2PO4, 100 mM NaCl, 50 μM EDTA, (pH 7.0) containing 9:1 H2O:D2O (v/v) (pH 7.0). Experiments were performed at 500 MHz. The temperature was 5 °C. The Watergate sequence was used for water suppression.67 The mixing time was 250 ms. The 31P-H1 experiments were carried out at the 1H frequency of 600 MHz. 31P -H3′ 3J couplings were applied to determine the phosphodiester backbone conformation.31P chemical shifts were referenced using indirect shift ratios.69</p><!><p>Footprints were drawn around NOE crosspeaks obtained at a mixing time of 250 ms. Their intensities were determined by volume integrations. These were combined as necessary with intensities generated from complete relaxation matrix analysis of a starting structure to generate a hybrid NOE intensity matrix.41,70 The program MARDIGRAS39,40,71 iteratively refined the hybrid intensity matrix and optimized agreement between calculated and experimental NOE intensities. The RANDMARDI algorithm39 carried out iterations, randomizing peak volumes within limits specified by the input noise level.71 Calculations were initiated using isotropic correlation times of 2, 3, and 4 ns. Analysis of these data yielded experimental distance restraints used in rMD calculations (Table S3 in the Supporting Information), and the corresponding standard deviations for the distance restraints.</p><p>The deoxyribose pseudorotational angles (P) were estimated by examining the 3JHH of sugar protons.72 The data were fit to curves relating the coupling constants to the pseudorotation (P), sugar pucker amplitude (φ), and the percentage S-type conformation. The pseudorotation and amplitude ranges were converted to the five dihedral angles ν0 to ν4. Coupling constants measured from 1H-31P HMBC spectra were applied73,74 to the Karplus relationship75 to determine the backbone dihedral angle ε (C4′-C3′-O3′-P), related to the H3′-C3′-O3′-P angle by a 120° shift. The ζ (C3′-O3′-P-O5′) backbone angles were calculated from the correlation between ε and ζ in B-DNA.68 Empirical restraints preserved Watson-Crick hydrogen bonding and prevented propeller twisting between base pairs, except for A6•T19 base pair. Except for the modified, the flanking, and the terminal base pairs, other backbone torsion angle restraints were using empirical data derived from B-DNA.44</p><!><p>Restrained molecular dynamics (rMD) calculations for the modified oligodeoxynucleotide duplexes utilized a simulated annealing approach.76 The partial charges on the cdG nucleotide (Figure S5 in the Supporting Information) were obtained from density function theory (DFT) calculations using a neutral total charge, utilizing the B3LYP/6-31G* basis set and the program GAUSSIAN.77 To obtain the starting structures used for rMD calculations, the cdG-modified duplex was energy minimized using 200 iterations with the conjugate gradients algorithm. The rMD calculations were conducted with AMBER78 and the parm99 force field. The generalized Born (GB) model79 with parameters developed by Tsui and Case80 was used for implicit water simulation. The program CORMA was utilized to calculate the NOE intensities from the structures emergent from rMD calculations.</p><p>Molecular dynamics simulations in explicit water were performed using the AMBER force field. The average structure converged from the simulated annealing rMD calculations was used as the starting structure. This was placed in an 8.0 Å cubic TIP3P water box in each direction.81 The necessary Na+ ions were added to neutralize the duplex. The system was subjected to 1000 iterations of potential energy minimization using steepest descents. The solvent was brought to thermal equilibrium by a MD simulation at constant volume for 10,000 iterations with an integrator time of 1 fs, at 300 K. After equilibration of the system at 300 K, MD calculations were performed at constant pressure for 5 ns with an integrator time of 1 fs. Bond lengths involving hydrogens were fixed with the SHAKE algorithm.82 The particle mesh Ewald (PME) method was used to approximate non-bonded interactions.83,84 The cutoff radius for non-bonded interactions was 8.0 Å. The PTRAJ program from the AMBER package was used to analyze the MD trajectories. Helicoidal analyses were carried out with the programs CURVES85 and 3DNA.86</p>

PubMed Author Manuscript

Glycosaminoglycan Storage Disorders: A Review

Impaired degradation of glycosaminoglycans (GAGs) with consequent intralysosomal accumulation of undegraded products causes a group of lysosomal storage disorders known as mucopolysaccharidoses (MPSs). Characteristically, MPSs are recognized by increased excretion in urine of partially degraded GAGs which ultimately result in progressive cell, tissue, and organ dysfunction. There are eleven different enzymes involved in the stepwise degradation of GAGs. Deficiencies in each of those enzymes result in seven different MPSs, all sharing a series of clinical features, though in variable degrees. Usually MPS are characterized by a chronic and progressive course, with different degrees of severity. Typical symptoms include organomegaly, dysostosis multiplex, and coarse facies. Central nervous system, hearing, vision, and cardiovascular function may also be affected. Here, we provide an overview of the molecular basis, enzymatic defects, clinical manifestations, and diagnosis of each MPS, focusing also on the available animal models and describing potential perspectives of therapy for each one.

glycosaminoglycan_storage_disorders:_a_review

1. Introduction<!>2. Mucopolysaccharidosis I<!>2.1. Hurler's Syndrome (MPS IH)<!>2.2. Hurler-Scheie's Syndrome (MPS IH/S)<!>2.3. Scheie's Syndrome (MPS IS)<!>3. Mucopolysaccharidosis II (Hunter's Syndrome)<!>4. Mucopolysaccharidosis III (Sanfilippo's Syndrome)<!>4.1. Mucopolysaccharidosis IIIA (Sanfilippo A)<!>4.2. Mucopolysaccharidosis IIIB (Sanfilippo B)<!>4.3. Mucopolysaccharidosis IIIC (Sanfilippo C)<!>4.4. Mucopolysaccharidosis IIID (Sanfilippo D)<!>5. Mucopolysaccharidosis IV<!>5.1. Morquio's Syndrome Type A<!>5.2. Morquio's Syndrome Type B<!>6. Mucopolysaccharidosis V<!>7. Mucopolysaccharidosis VI (Maroteaux-Lamy Syndrome)<!>8. Mucopolysaccharidosis VII (Sly's Syndrome)<!>9. Mucopolysaccharidosis VIII<!>10. Mucopolysaccharidosis IX<!>11. Conclusion<!>

<p>The mucopolysaccharidoses (MPSs) are a group of lysosomal storage disorders caused by deficiency of enzymes catalyzing the stepwise degradation of glycosaminoglycans (GAGs) and characterized by intralysosomal accumulation and increased excretion in urine of partially degraded GAGs, which ultimately results in cell, tissue, and organ dysfunction [1].</p><p>Glycosaminoglycans (previously called mucopolysaccharides), with the exception of hyaluronic acid, are the degradation products of proteoglycans that exist in the extracellular matrix and are proteolytic cleaved, giving origin to GAGs, which enter the lysosome for intracellular digestion. There are four different pathways of lysosomal degradation of GAGs, depending on the molecule to be degraded: dermatan sulfate, heparan sulfate, keratan sulfate, and chondroitin sulfate. The stepwise degradation of glycosaminoglycans requires 10 different enzymes: four glycosidases, five sulfatases, and one nonhydrolytic transferase, whose structure, biosynthesis, processing, and cDNA sequence have already been extensively documented. Deficiencies of each one of these enzymes have already been reported and result in seven different MPSs, all of them sharing a series of clinical features, even though in variable degrees (summarized in Table 1) [1, 2].</p><p>Usually, MPSs are characterized by a chronic and progressive course, with different velocities of progression depending on the severity of each one. The typical symptoms include organomegaly, dysostosis multiplex, and a characteristic abnormal facies. Hearing, vision, and cardiovascular function may also be affected. Additionally, joint mobility may also be compromised. The majority of symptoms may be explained by abnormal accumulation of undegraded substrates within the lysosomes. In fact, the continued presentation of GAGs to cell for degradation results in storage, which gives rise to an enlargement of lysosomes. As substrates accumulate, the lysosomes swell and occupy more and more of the cytoplasm. As a consequence of this increased number and size of lysosomes, other cellular organelles may be obscured, and the nuclear outline may be deformed. As the process continues, the enlarged cells lead to organomegally. Abnormalities observed in heart cells and function may also be explained by GAGs accumulation. The increase of storage material within the cells of the heart valves causes an alteration of the cell's outline, changing them from fusiform to round. As a consequence, the valve leaflet and cordae tendinea become thickener and interfere with normal cardiac function, producing valvular stenosis. At corneal level, also, storage of undegraded GAGs results in reflection and refraction of light, leading to the cloudiness which is so typical of these pathologies. Also at the CNS level, swollen neurons and lysosomes may produce lesions that include the development of meganeurites and neurite sprouting (reviewed in [3, 4]).</p><p>Traditionally, MPSs are recognized through analysis of urinary GAGs. Several methods have been devised, to precise qualitative identification and quantitative measurements. These analyses of urinary GAGs allow discrimination between broad classes of MPSs but cannot distinguish subgroups. Definitive diagnosis is usually accessed through enzymatic assays of the defective enzyme in cultured fibroblasts, leukocytes, and serum or plasma (reviewed in [1]). During the last decade; however, dried blood spot technology was also introduced for enzymatic assays, allowing cheaper, easier, feasible diagnosis and opening the possibility for large population screenings (see Section 11 for more details).</p><p>In general, MPSs are transmitted in an autosomal recessive fashion, except for MPS II, which is X-linked.</p><p>This paper provides an overview of the molecular basis, enzymatic defects, clinical manifestations, and diagnosis of each glycosaminoglycan storage disease, focusing also on the respective animal models and describing potential perspectives of therapy which are being tested as well as the ones which are already available (summarized in Table 2).</p><!><p>Mucopolysaccharidosis I is caused by a deficiency of α-L-iduronidase (IDUA; EC 3.2.1.76) and can result in a wide range of phenotypic involvement with three major recognized clinical entities: Hurler (MPS IH; MIM#607014), Hurler-Scheie (MPS IH/S; MIM#607015), and Scheie (MPS IS) syndromes. Hurler and Scheie syndromes represent phenotypes at the severe and mild ends of the MPS I clinical spectrum; respectively, and the Hurler-Scheie syndrome is intermediate in phenotypic expression [20]. It is important to stress that, although MPS I may be subdivided into these three clinically diverse entities, the underlying enzymatic defect is common to all of them, being all caused by mutation in the gene encoding α-L-iduronidase (IDUA).</p><p>Functionally, α-L-iduronidase is essential to the correct metabolism of both dermatan sulfate and of heparan sulfate, hydrolyzing the terminal α-L-iduronic acid residues of the above-referred glycosaminoglycans [1].</p><p>In 1992, Scott and colleagues [21] were able to clone and purify the gene that encodes this enzyme, IDUA, demonstrating that it spans approximately 19 kb and contains 14 exons. The first 2 exons are separated by an intron of 566 bp, a large intron of approximately 13 kb follows, and the last 12 exons are clustered within 4.5 kb. Previously, this gene was mapped to 4p16.3, through unequivocal in situ hybridization and southern blot analysis of mouse-human cell hybrids [22].</p><p>There are, presently, several animal models known for MPS I.</p><p>In 1979, Haskins and colleagues [5] described α-L-iduronidase deficiency in a cat, and, few years later, Shull et al. [6] and Spellacy et al. [23] reported a similar deficiency in the dog. Subsequent studies lead to cloning and characterization of the canine IDUA gene as well as the mutation causing the observed phenotype [24, 25] and proved it to be a good model for study of human MPS I. So, in 1994, Shull and collaborators [26] published the first results of enzyme replacement therapy in the canine model. Through intravenous administration of recombinant human α-L-iduronidase, these authors managed to obtain a remarkable resolution of lysosomal storage in both hepatocytes and Kupffer's cells. In the same year, Grosson et al. [27] mapped the homologous IDUA locus in the mouse to chromosome 5. That knowledge was later used to create a knock-out mouse presenting the characteristic MPS I features [7, 28].</p><p>Currently, both hematopoietic stem cell transplantation (HSCT) and enzyme replacement therapy (ERT) using laronidase (recombinant human α-L-iduronidase, Aldurazyme) are available for MPS I. HSCT is the recommended treatment for patients with severe MPS I, before 2 years of age [29–32]. ERT is recommended for the other cases, and it has been shown to be effective in ameliorating some of the clinical manifestations of MPS disease. Among positive effects are decreased hepatosplenomegaly, improved respiratory and myocardial function and physical capacity [33–35] as well as improvement in active movement followed by enhanced self-care [36]. Recently, several reports have been published trying to evaluate long-term effect of ERT on the natural history of treated patients. From those studies, several conclusions have been reached. Concerning treated patients' growth pattern, it became clear that children with MPS I grow considerably slower than healthy individuals, and differences between healthy and affected children increase with age [37]. Other relevant evidences show that early treatment of attenuated MPS I may significantly delay or prevent the onset of the major clinical signs, substantially modifying the natural history of the disease [38]. Additional investigation is needed to clarify the mechanisms by which improvements are achieved in laronidase-treated patients. Such knowledge may support the development of ERT directly targeting the brain.</p><!><p>Hurler's syndrome is the most severe form of MPS I and has been, over the last decades, the prototype description of MPS. Nevertheless, this may be misleading, since not all MPSs share the same features, and this pathology in particular is not representative of all of them, but only of the most severe end of a broad clinical spectrum (reviewed in [1]). Like all other MPSs, the clinical course of this disease is progressive, with multiple organ and tissue involvement. Hallmark clinical features of Hurler syndrome include coarse facies, corneal clouding, mental retardation, hernias, dysostosis multiplex, and hepatosplenomegaly. Children with Hurler's syndrome appear normal at birth and develop the characteristic appearance over the first years of life [39]. Length is often normal until about 2 years of age when growth stops; by age of 3 years, height is under the third percentile [40]. Cardiac disease and respiratory complications are common. Acute cardiomyopathy associated with endocardial fibroelastosis has been a presenting condition in some infants with MPS I less than 1 year of age [41]. Upper and lower respiratory tract infections are also frequent [42]. Developmental delay is often apparent by 12 to 24 months of age, with a maximum functional age of 2 to 4 years followed by progressive deterioration. Most children develop limited language as a consequence of developmental delay, chronic hearing loss, and enlarged tongue [1]. Dermal melanocytosis may also be found in Hurler patients [43], as well as in patients suffering from other LSDs, such as GM1 gangliosidosis. Nevertheless, Hurler's syndrome is the most common lysosomal storage disease associated with dermal melanocytosis, as revealed by a literature analysis.</p><!><p>MPS IH/S corresponds to a clinical phenotype which is intermediate between the Hurler and the Scheie syndromes. It is characterized by progressive somatic involvement with dysostosis multiplex but little or no mental retardation. First symptoms usually occur between 3 and 8 years. Characteristic features of Hurler's syndrome, such as corneal clouding, joint stiffness, deafness, and valvular heart disease, can also appear in MPS IH/S patients. Nevertheless, the onset of these symptoms occurs much later than that in the severe MPS I type, beginning in the midteens and leading to significant impairment and loss of function. Other clinical features, such as micrognathism, pachymeningitis cervicalis, and compression of the cervical cord due to GAG accumulation in the dura, may also occur. Cardiac and respiratory complications may explain the high clinical mortality (reviewed in [1]).</p><!><p>Scheie's syndrome was earlier thought to be a separate entity designated MPS V, instead of a phenotypical subtype of MPS I [44]. This pathology is characterized by a mild phenotype in which dysostosis multiplex can be present. Joint involvement is marked in the hand with a claw-hand deformity. Patients also have genu valgum, stiff, painful feet, and pes cavus [1]. Cardiac and respiratory complications are much milder than in the Hurler syndrome, with aortic and mitral valvular disease being a common feature [45]. At a respiratory level, Perks et al. [46] have reported two brothers with Scheie's syndrome suffering from sleep apnea, but no other complications are known. Intelligence is normal [1]. Pachymeningitis cervicalis (compression of the cervical cord secondary to glycosaminoglycan accumulation in the dura) may also occur.</p><!><p>Mucopolysaccharidosis II is the sole MPS transmitted in an X-linked manner and is caused by deficiency of the lysosomal enzyme iduronate sulfatase, which is crucial to the correct degradation of heparan and dermatan sulfate, by cleaving their O-linked sulfate. As a result, there is a progressive accumulation of glycosaminoglycans in nearly all cell types, tissues, and organs. Patients with MPS II excrete excessive amounts of dermatan sulfate and heparan sulfate in the urine [20, 47]. Hunter syndrome is caused by mutation in the gene encoding iduronate-2-sulfatase (IDS).</p><p>Although the disease is known since the early 1970s, being the first MPS to be defined clinically in humans, it was not until the 1990s that the IDS was cloned. In 1991, Wilson et al. [48] localized the gene to Xq28. Two years later, Flomen and coworkers [49] described the gene's structure as containing 9 exons and characterized the intron sequences surrounding them. In the same year, Wilson et al. [50] reported the complete sequence of the IDS gene, which spans approximately 24 kb. The potential promoter for IDS lacks a TATA box but contains GC box consensus sequences, which are consistent with its role as a housekeeping gene.</p><p>Curiously, a second IDS gene (IDS2) was identified by Bondeson et al. [51]. It is a pseudogene and is located within 90 kb telomeric region of the IDS gene and involved in a recombination event with the primary IDS gene in about 13% of patients with the Hunter syndrome.</p><p>Traditionally, the Hunter syndrome comprises 2 recognized clinical entities, according to the severity of symptoms: mild and severe. Although largely used, this nomenclature does have its difficulties, since the mild and severe forms represent the two ends of a wide and continuous spectrum of clinical severity. Also, in terms of iduronate deficiency, these forms cannot be distinguished since the enzyme's activity is equally deficient in both (reviewed in [1]). They are, though, separated almost exclusively on clinical grounds, although nowadays mutation analysis may help distinguish them.</p><p>This classification of MPS goes back to 1972, when McKusick distinguished between the severe form (which he called MPS IIA), with progressive mental retardation and physical disability and death before age 15 years in most cases, and the mild form (called MPS IIB) compatible with survival to adulthood and in which intellect is impaired minimally, if at all. He also pointed out the lack of corneal clouding in the X-linked form of MPS as opposed to the autosomal forms.</p><p>Presently, this classification has become obsolete since, in 2008, Wraith et al. [47] stated that MPS II should be regarded as a continuum between the two extremes (severe and attenuated). They noted that, although the clinical course for the more severely affected patients is relatively predictable, there is considerable variability in the clinical phenotype and progression of the more attenuated form of the disease and, so, it would not be correct to consider the milder form as a separate entity but, instead, look at Hunter's disease as a phenotypical continuum, with several possible degrees of severity.</p><p>In 1998, Wilkerson et al. [8] described Hunter's syndrome in a Labrador retriever, with the typical clinical features observed in humans: coarse facies, macrodactyly, corneal dystrophy, progressive CNS deterioration, and positive biochemical diagnosis for MPS through urine analysis.</p><p>After the successful results obtained in improving certain disease manifestations in patients with MPS I, including visceral manifestations and attenuation of neurologic disease progression [29, 52], hematopoietic stem cells transplantation (HSCT) has also been performed in several patients with MPS II. Unfortunately, although the transplantation of hematopoietic stem cells provides some enzymatic reconstruction in many target tissues with decreased excretion of GAGs in urine, decreased liver and spleen volumes, diminished facial coarsening, and improved respiratory function and joint mobility [53, 54], the results at neurological level were disappointing (reviewed in [55]). The additional risk of morbidity and mortality associated to this procedure led investigators to focus their attention in ERT for this pathology, with much better results, as discussed below.</p><p>A knock-out mouse model for MPS II was developed by replacing exon 4 and a portion of exon 5 of IDS with the neomycin-resistance gene [9, 56]. Affected mice exhibit a phenotype with notorious similarities to human disease, both at the biochemical and the clinical levels [9]. Several studies with this knock-out mouse model were done to assess the effect of ERT [56] as well as dose and various dosing regimens of idursulfase in urine and tissue GAG levels [57]. The results of these studies were quite promising, with a marked decrease in urinary GAGs as well as decreased GAG accumulation in several tissues [56] verified for several idursulfase doses and several dosing frequencies [57]. These studies have been used to support the first clinical trial of recombinant IDS in Hunter's syndrome patients. At the moment, both phase I/II [58] and phase II/III [59] clinical studies have proven not only the efficacy but also the safety of idursulfase replacement therapy. Consequently, ERT with recombinant human iduronate sulfatase (Elaprase, idursulfase, Shire Human Genetic Therapies Inc.) was approved in the US (July, 2006) and the European Union (January, 2007) for the treatment and the management of MPS II. The recommended dose is 0.5 mg/kg administered once weekly as an intravenous infusion (reviewed in [55]). As time goes by, additional evidence on the efficacy of ERT for MPS II patients is being published, as long-term treatments are successful. This is the case of a recently published report on the improvements observed in a 7 years and 10 months old child who began a 36 months' treatment with Elaprase at 4 years and 10 months. At the end of the treatment, the child presented normal excretion of GAGs in urine, normal-sized liver and spleen, and significant bone remodeling. Cardiac and neurological development, however, still progressively deteriorated [60]. This year, protective effects of ERT in MPS II patients were also reported for DNA damaging in leukocytes [61] and oxidative stress [62].</p><!><p>The Sanfilippo syndrome, or mucopolysaccharidosis III, is caused by impaired degradation of heparan sulfate [1] and includes 4 subtypes, each due to the deficiency of a different enzyme: heparan N-sulfatase (type A; MIM no. 252900), α-N-acetylglucosaminidase (type B; MIM no. 252920), acetyl CoA: α-glucosaminide acetyltransferase (type C; MIM no. 252930), and N-acetylglucosamine-6-sulfatase (type D; MIM#252940). At a clinical level, the four subtypes are quite similar, with a characteristic severe central nervous system degeneration associated with mild somatic disease. Onset of clinical features usually occurs between 2 and 6 years, severe neurologic degeneration occurs in most patients between 6 and 10 years of age, and death occurs typically during the second or third decade of life. Type A has been reported to be the most severe, with earlier onset and rapid progression of symptoms and shorter survival [63].</p><!><p>General MPS IIIA clinical features include severe mental retardation with relatively mild somatic features (moderately severe claw hand and visceromegaly, little or no corneal clouding, little or no vertebral change). Usually, this pathology is characterized by marked overactivity, destructive tendencies, and other behavioral aberrations.</p><p>MPS IIIA phenotype is caused by mutations in the gene encoding N-sulfoglucosamine sulfohydrolase, also named heparan sulfate sulfatase (SGSH; 605270). This enzyme is specific for sulfate groups linked to the amino group of glucosamine.</p><p>In 1995, the gene encoding N-sulfoglucosamine sulfohydrolase, SGSH, was isolated, sequenced, and cloned [64]. Later, it was shown to contain 8 exons spanning approximately 11 kb [65].</p><p>There are two animal models known for MPS IIIA. The first to be discovered was the canine model when Fischer et al. [10] identified sulfaminidase deficiency in two adult wire-haired dachshund littermates. Subsequently, Aronovich et al. [66] determined the normal sequence of the canine heparan sulfate sulfatase gene and cDNA, through PCR-based approaches. Another model was described in 2001, when Bhattacharyya and collaborators [11] found a spontaneous mouse mutant of MPS IIIA resulting from a missense mutation (D31N) in the murine sulfatase gene. Affected mice die at about 10 months of age, exhibiting notorious visceromegaly, distended lysosomes and heparan sulfate accumulation in urine. Hemsley and Hopwood [67] found that these mice had severe brain involvement, with impaired open field locomotor activity and behavioral changes, suggesting axonal degeneration. Later, Settembre et al. [68] observed increased autophagosomes resulting from autophagosome-lysosome function in these mice. Similar findings were observed in another mouse model of another lysosomal storage disorder (multiple sulfatase deficiency; MSD; MIM no. 272200), reinforcing the recent idea that these diseases are disorders of autophagy, which may be a common mechanism for neurodegenerative lysosomal storage disorders.</p><p>MPS IIIA mice were recently tested for substrate deprivation therapy with both genistein and rhodamine B, two chemicals that inhibit GAG synthesis ([4, 69], reviewed in [70]). Encouraging results were obtained with both compounds, and this therapeutic approach started to be considered for several MPSs (see Section 11 for more details). Other interesting results were also obtained when siRNAs were used to reduce GAG synthesis in MPS IIIA mice. Last year, this approach was tested by Dziedzic et al. [71], who managed to reduce mRNA levels of four genes, XYLT1, XYLT2, GALTI, and GALTII, whose products are involved in GAG synthesis. This decrease of levels of transcripts corresponded to a decrease in levels of proteins encoded by them. Moreover, efficiency of GAG production in these fibroblasts was considerably reduced after treatment of the cells with siRNA. Either way, substrate deprivation therapy seems to be a promising approach for Sanfilippo's syndrome type A.</p><p>Gene therapy approaches are also being tested in MPS IIIA mice. Recently, promising results have been reported by Fraldi et al. [72], who performed experiments with intracerebral adeno-associated-virus- (AAV-) mediated delivery of SGSH gene, together with SMUF1 gene, which exhibits an enhancing effect on sulfatase activity when coexpressed with sulfatases. They observed a visible reduction in lysosomal storage and inflammatory markers in transduced brain regions, together with an improvement in both motor and cognitive functions.</p><!><p>With a phenotype quite similar to MPS IIIA, the Sanfilippo syndrome B is characterized by deficiencies of α-N-acetylglucosaminidase, caused by mutations in the NAGLU gene that encodes this enzyme. α-N-Acetylglucosaminidase is required for the removal of the N-acetylglucosamine residues that exist in heparan sulfate or are generated during lysosomal degradation of this polymer by the action of heparan acetyl-CoA: α-glucosaminide N-acetyltransferase (reviewed in [1]).</p><p>The NAGLU gene was cloned in 1995 by Zhao and colleagues [73]. The deduced 743-amino acid protein has a 20- to 23-residue leader sequence, consistent with a signal peptide, and 6 potential N-glycosylation sites. It contains 6 exons and spans 8.3 kb on chromosome 17q21 [74].</p><p>Similarly to the above-referred MPS III syndrome, there is also a natural occurring mutant for Sanfilippo B. It was described by Ellinwood and coworkers, in 2003, in Schipperke's dogs [12].</p><p>During the last decade, Li et al. [75] created a laboratorial murine MPS IIIB was also constructed trough targeted disruption of the NAGLU gene [76]. With a phenotype quite similar to that of patients with MPS IIIB, this model began immediately to be used for therapeutic approaches as well as for pathogenesis studies. The first studies were done to evaluate the potential of ERT for this pathology [76]. The results, however, were quite disappointing since the recombinant NAGLU produced in Chinese hamster ovary (CHO) cells was not efficiently captured by MPS IIIB cells, either in vitro [77, 78] or in vivo [76]. This difficulty has turned the search for a treatment for MPS IIIB even more challenging. Presently, several therapies are under evaluation for this disease, including cell-mediated therapy, enzyme enhancement therapy, substrate deprivation therapy, and viral gene therapy (reviewed in [79]).</p><p>Promising results are being achieved through gene therapy approaches in MPS IIIB mice, namely, through direct microinjection into the brain of adeno-associated virus (AAV) vectors coding for NAGLU [80–82] and intravenous injections and intracranial gene delivery of lentiviral (LV) vector of NAGLU [83–85].</p><!><p>Sanfilippo syndrome C is, in general, characterized by the same clinical features described to MPS IIIA. Nevertheless, the enzyme deficiency in this pathology is different from the one causing the latter. Type C disease is caused by mutations in the gene encoding heparan acetyl-CoA: α-glucosaminide N-acetyltransferase (HGSNAT; 610453). This is the only known lysosomal enzyme that is not a hydrolase. It catalyzes the acetylation of the glucosamine amino groups that have become exposed by the action of heparan-N-sulfatase (reviewed in [1]).</p><p>The HGSNAT gene was cloned in parallel by two different groups, during the last decade: Fan et al. [86] and Hřebíček et al. [87]. The molecular defects underlying MPS IIIC remained unknown for almost three decades due to the low tissue content and the instability of HGSNAT [88].</p><p>To date, 54 HGSNAT sequence variants have been identified including 13 splice-site mutations, 11 insertions and deletions with consequent frameshifts and premature termination of translation, 8 nonsense, and 18 missense (reviewed in [89]).</p><p>Recently, two independent studies from Feldhammer et al. [88] and Fedele and Hopwood [90] have performed exhaustive functional analysis of the majority of the missense mutations already reported for the HGSNAT gene. Attention was focused in this particular type of mutations since there are several MPS IIIC patients carrying only missense mutations, either homozygous or heterozygous, who present an unexpected severe phenotype. In fact, although splicing and frameshift mutations are usually associated to that type of phenotype, since they give rise to premature termination codons and trigger nonsense-mediated mRNA decay (NMD); missense mutations are traditionally associated to milder disease. Nevertheless, this typical/general pattern is not observed for MPS IIIC. That is why these alterations were specifically cloned, expressed, and analyzed for their folding, targeting, and enzymatic activities. As a result, Fedele and Hopwood [90] have observed that the expression levels and enzymatic activity of most mutants were extremely low or even negligible. Feldhammer and colleagues [88], on the other hand, have observed that those mutations cause a misfolding of the enzyme, which is not correctly glycosylated. As a consequence, HGSNAT is not targeted to the lysosome but, instead, stays in the endoplasmic reticulum (ER). Thus, enzyme folding defects due to missense mutations, together with NMD seem to be the major molecular mechanisms underlying MPS IIIC. This makes MPS IIIC a good candidate for enzyme enhancement therapy, where active site-specific inhibitors are used as pharmacological chaperones to modify the conformation of the mutant lysosomal enzymes usually retained and degraded in the ER, in order to increase the level of the residual activity to a point which is sufficient to reverse the clinical phenotypes [88]. Together with inhibitors of heparan sulphate synthesis, pharmacological chaperones are currently being tested to reduce storage of this polymer in the CNS to levels sufficient to stop neuronal death and reverse inflammation.</p><!><p>Like the previous MPS III subtypes, Sanfilippo's syndrome D presents a phenotype similar to MPS IIIA, with a singular enzyme deficiency underlying it: mutation in the gene encoding N-acetylglucosamine-6-sulfatase (GNS; 607664). The enzyme was originally described as specific for the 6-sulphated N-acetylglucosamine residues of heparan sulphate. However, the early data have been reinterpreted, and given that this sulfatase is in fact able to desulphate the 6-sulphated N-acetylglucosamine present in α- or in β-linkage or even as a free monosaccharide (reviewed in [1]).</p><p>N-Acetylglucosamine-6-sulfatase (EC 3.1.6.14) was purified and characterized by Freeman et al. [91], who identified 4 different forms of the enzyme in liver. Its catalytic properties were studied by Freeman and Hopwood [92]. Afterwards, Robertson et al. [93] assigned the glucosamine-6-sulfatase gene, which they symbolized G6S, to chromosome 12q14 by in situ hybridization of a tritium-labeled G6S cDNA probe. The localization was confirmed by using the cDNA clone in analyses of DNA from human/mouse hybrid cell lines. More recently, that information was completed by the work of Mok et al. [94], who amplified and sequenced the promoter and 14 exons of the GNS gene from a patient with MPS IIID. By analyzing that patient, it was also possible to identify a homozygous nonsense mutation in exon 9, predicted to result in premature termination at codon 355, as well as two common synonymous coding SNPs. At the same time, another group identified a 1-bp deletion in the GNS gene in another affected individual [95].</p><p>A naturally occurring large animal model was described by Thompson et al. [13], who reported type D Sanfilippo's syndrome in a Nubian goat. Later, caprine MPS IIID was used to evaluate the efficacy of ERT in this pathology. Recombinant caprine N-acetylglucosamine-6-sulfatase was administered intravenously to one MPS IIID goat at 2, 3, and 4 weeks of age. As a result, a marked reduction of lysosomal storage vacuoles was observed in hepatic cells, but no amelioration was noticed concerning the CNS lesions. No residual enzyme activity was observed either in brain or liver. Taking this preliminary results into account, it was considered that other treatment regimens will be necessary for MPS IIID [96].</p><!><p>Mucopolysaccharidosis IV, or Morquio's syndrome, is caused by impaired degradation of keratan sulphate. Presently, there are two known enzyme deficiencies causing 2 different subtypes of Morquio's syndrome: deficiency in N-acetylglucosamine-6-sulfatase (causing Morquio's disease type A; MIM no. 253000) and deficiency in β-galactosidase (causing Morquio's disease type B; MIM no. 253010). Both MPS IV subtypes present a wide spectrum of clinical manifestations, but there are some characteristic common features: short trunk dwarfism, fine corneal deposits, spondyloepiphyseal dysplasia. Actually, the predominant clinical features of Morquio's syndrome are the ones related to the skeleton. Most of the times, this severe somatic disease is accompanied by a normal intelligence [1]. Patients with the severe phenotype do not normally survive past the second or third decade of life [97].</p><!><p>Morquio's syndrome A is caused by mutations in the gene encoding galactosamine-6-sulphate sulfatase (GALNS), which plays a crucial role on the degradation of both keratan sulphate and chondroitin sulphate.</p><p>The gene coding for human galactosamine-6-sulphate sulfatase (GALNS), was mapped to chromosome 16q24.3 trough fluorescence in situ hybridization assays [98]. Its structure was described at the same time by independent groups as comprising 14 exons and spanning approximately 40–50 kb [99, 100]. Curiously, the GALNS gene contains an Alu repeat in intron 5 and a VNTR-like sequence in intron 6 [100].</p><p>No natural occurring model is known for either type A or type B Morquio's syndrome. Nevertheless, a laboratorial murine model for type A syndrome was created from an induced disruption in exon 2 of the GALNS gene. Mutants presented no detectable enzyme activity and showed increased GAG levels in urine. GAGs accumulation was also detected in several tissues including liver, kidney, spleen, heart, brain, and bone marrow [14]. These mice were later tested for enzyme replacement therapy and, after a 12-week long treatment with native GALNS or SUMF1-modified GALNS, showed manifest clinical improvement, demonstrated by a marked reduction of storage material in visceral organs, bone marrow, heart valves, ligaments, and connective tissue. The clearance of stored material in brain was dose dependent, and the keratan sulphate blood levels were reduced to normal [101].</p><p>Presently, there is no effective therapy for MPS IVA and care has been palliative, as in the majority of LSDs. Enzyme replacement therapy (ERT) and hematopoietic stem cells therapy (HSCT) have been considered as potential therapeutic approaches for MPS IVA (reviewed in [102]), ERT being, though, the most attractive candidate, since affected patients lack CNS involvement.</p><p>Recently, Rodríguez et al. [103] have produced a recombinant GALNS enzyme in Escherichia coli BL21. To produce sufficient amounts of purified GALNS enzyme, high level expression of GALNS in Chinese hamster ovary (CHO) cells has been established as a source of selectively secreted human recombinant enzyme. This recombinant enzyme has already been tested in the murine knock-out model, with consequent clearance of tissue and blood keratan sulphate [101]. These results provided important preclinical data for the design of GALNS ERT trials, which are now in course.</p><!><p>Although presenting overlapping clinical features, Morquio's syndrome B is genetically distinct from Morquio's syndrome A, being caused by impairments in another enzyme involved in the stepwise degradation of keratan sulphate: β-galactosidase, which is coded by the GLB1 gene. Beta galactosidase hydrolases terminal β-linked galactose residues found in GM1 ganglioside, glycoproteins, and oligosaccharides, as well as in keratan sulphate (reviewed in [1]).</p><p>The GLB1 gene spans 62.5 kb and contains 16 exons [104, 105] and maps to chromosome 3p21.33 [106]. The deduced 677-residue protein has a calculated molecular mass of 75 kD and contains a putative 23-residue signal sequence and 7 potential asparagine-linked glycosylation sites. It may be interesting to refer that the GLB1 gene gives rise to 2 alternatively spliced mRNAs: a major 2.5-kb transcript that encodes the classic lysosomal form of the enzyme of 677 amino acids, and a minor 2.0-kb transcript that encodes a β-galactosidase-related protein (elastin-binding protein, EBP) of 546 amino acids with no enzymatic activity and a different subcellular localization. Exons 3, 4, and 6 are absent in the 2.0-kb mRNA as a consequence of alternative splicing of the pre-mRNA [107–109].</p><p>Presently, there are no known animal models for MPS IVB, either natural or engineered.</p><!><p>The designation MPS V is no longer used. In fact, the phenotype which was first classified as MPS V, was later found to be the milder form of MPS I (Scheie's syndrome), caused by deficiencies in α-L-iduronidase, with the typical stiff joints, clouding of the cornea most dense peripherally, survival to a late age with little if any impairment of intellect and aortic regurgitation [44].</p><!><p>Mucopolysaccharidosis type VI is an autosomal recessive lysosomal storage disorder resulting from a deficiency of arylsulfatase B (N-acetylgalactosamine-4-sulfatase). Clinical features and severity are variable but usually include short stature, hepatosplenomegaly, dysostosis multiplex, stiff joints, corneal clouding, cardiac abnormalities, and facial dysmorphism. Intelligence is usually normal [110].</p><p>Arylsulfatase B is a lysosomal enzyme that removes the C4 sulphate ester group from the N-acetylgalactosamine sugar residue at the nonreducing terminus of dermatan sulphate and chondroitin sulphate, during lysosomal degradation [111]. The gene that codes for this enzyme was first mapped to chromosome 5q11-q13 [112] and is now known to contain 8 exons and span about 206 kb [111].</p><p>In 2002, a 3-year-old Siamese/short-haired European cat was referred for clinical disease characterized by dwarfism, facial dysmorphia, paralysis, small and curled ears, corneal clouding, and large areas of alopecia. X-ray examination showed multiple bone dysplasias. These features lead to suspect from a mucopolysaccharide storage disorder. Subsequent analysis proved it to be a natural occurring form of the Maroteaux-Lamy syndrome [15].</p><p>This MPS VI model has been extensively used over the last years to test ERT for this specific pathology. In 2003, Auclair and colleagues [113] have evaluated the cats' response to infusions of recombinant human N-acetylgalactosamine-4-sulfatase (rhASB) and observed an overall improvement in the disease condition at physical, neurological, and skeletal levels. Later, the same team has demonstrated that a high rate of immunotolerance towards rhASB can be achieved in MPS VI cats with a short-course tolerisation regimen [114], which may help the implementation of such procedures. Another interesting approach was designed, specifically to ameliorate joint disease in MPS IVA, through long-term articular administration of rhASB, leading to a notorious improvement in feline joint disease [115]. These successful results lead to the development of clinical trials in MPS VI patients, and three clinical studies including 56 patients have evaluated the efficacy and safety. As a consequence, enzyme replacement therapy (ERT) became available. The specific ERT for MPS VI, galsulfase (Naglazyme, Biomarin Pharmaceutical) was approved in 2005 by FDA and in 2006 by EMA. Long-term follow-up data with patients treated up to 5 years showed that ERT is well tolerated and associated with sustained improvements in the patients' clinical condition [2, 116].</p><p>Even though presently there is ERT available for these patients, other therapeutic approaches are being tested in animal models for MPS VI. In 2009, the first attempt of successful gene therapy was performed through lentiviral-mediated gene transfer to joint tissues of the rat, with consequent correction of MPS VI cells [117]. This year, another study, involving intravascular administration of adeno-associated viral vectors in MPS VI cats, was published. After gene transfer the authors observed clearance of GAG storage, improvement of long bone length, reduction of heart valve thickness, and improvement in spontaneous mobility [118]. Either way, promising therapeutic strategies for MPS VI patients may be arising.</p><!><p>MPS VII, also known as Sly's syndrome, is characterized by the impossibility to degrade glucuronic acid-containing GAGs, due to impaired function of β-glucuronidase, which removes the glucuronic acid residues present in dermatan sulphate as well as in heparan and chonroitin sulphates (reviewed in [1]). Clinical features are highly variable, with phenotypes ranging from severe fetal hydrops to mild forms allowing survival into adulthood. Typical features include hepatomegaly, skeletal abnormalities, coarse facial features, and variable degrees of mental impairment [119].</p><p>MPS VII was first reported by Sly and collaborators in 1973, in a boy with skeletal changes consistent with MPS, hepatosplenomegaly, and granular inclusions in granulocytes. Additional features included hernias, unusual facies, protruding sternum, thoracolumbar gibbus, vertebral deformities, and mental deficiency. When β-glucuronidase activity was measured in fibroblasts, obtained values were less than 2% of control values. Both parents and several sibs of the mother showed an intermediate level of the enzyme [120].</p><p>In 1990, Miller et al. [121] reported that the gene encoding β-glucuronidase (GUSB) is 21 kb long, contains 12 exons, and gives rise to two different types of cDNAs, through an alternate splicing mechanism. Speleman et al. [122] used fluorescence in situ hybridization to map the GUSB gene to 7q11.21-q11.22. This map position was confirmed by dual-color hybridization of β-glucuronidase and another gene which had been mapped proximal to it: elastin (7q11.23).</p><p>Several pseudogenes, located on chromosomes 5, 6, 7, 20, 22, and Y, were also detected by Shipley et al. [123], when amplifying exons 2–4, 3, 6-7, and 11.</p><p>In 2009, Tomatsu et al. [124] provided a review of mutations in the GUSB gene that cause MPS type VII. Forty-nine different pathogenic mutations have been reported in the literature, with approximately 90% of them being missense mutations. Approximately 40% of the known GUSB mutations occur at CpG sites within the gene. The most common mutation is L176F, which has been found in several populations: American (Caucasian), Brazilian, British, Chilean, French, Mexican, Polish, Spanish, and Turkish ([125–127], reviewed in [124]). Genotype/phenotype analysis indicated that the most severe phenotype was associated with truncating mutations and with mutations affecting either the hydrophobic core or the modification of packing.</p><p>In 1984, mucopolysaccharidosis type VII (Sly syndrome) was described in a mixed-breed dog [16]. Since then, several other affected dogs have been studied, in the animal colony established at the University of Pennsylvania, the School of Veterinary Medicine [128] and, later, in a 12-week-old male German Shepherd dog studied in the same school [129]. All dogs shared the same missense mutation and developed similar phenotypes with skeletal deformities, corneal cloudiness, cytoplasmic granules in the neutrophils and lymphocytes of blood and CSF, and glycosaminoglycans in urine [129]. Another animal model was described as naturally occurring: the gusmps/mps mouse, which has a 1 bp deletion in exon 10 resulting in a progressive degenerative disease that reduces lifespan and causes facial dysmorphism, growth retardation, deafness, and behavioral defects [17]. Nevertheless, opportunities for experimental therapies were greatly expanded by the work of Tomatsu et al., in 2006 [18], who developed a new MPS VII mouse model, which is tolerant to both human and murine GUS, without the characteristic immune responses that complicated evaluation of the long-term benefits of enzyme replacement or gene therapy when the naturally occurring mice were used. Ever since, several therapeutic approaches have been attempted in MPS VII mice, and the results have been encouraging. That is the case of the works by Bosch and collaborators, who have been working on gene therapy for this pathology, in order to correct brain lesions. They have used both adeno-associated virus (AAV) [130] and lentivirus-mediated gene transfer [131] and observed that there was a significant correction of pathology in the brain of affected mice.</p><p>Other therapeutic approaches had already been attempted, but their results were not as promising. In fact, in 1998, allogenic bone marrow transplantation was reported in a 12-year-old Japanese girl with consequent improvement of motor function and daily life activities, decrease of upper respiratory and ear infections, but no improvement at all in cognitive function [132].</p><!><p>The clinical entity once known as MPS VIII was described in a single patient, in the late 1970s. The patient, a 5-year-old child, presented short stature, coarse hair, hepatomegaly, mild dysostosis multiplex, mental retardation, and no signs of corneal clouding. Biochemical analysis of the urine revealed increased excretion of keratan and heparan sulphate [133, 134]. The biochemical findings described by this group lead to suspect the existence of two hexosamine sulfatases and propose the existence of this novel MPS, caused by glucosamine-6-sulfatase [133].</p><p>Nevertheless, subsequent analysis on Diferrante's laboratory brought this idea down, and the designation MPS VIII was abandoned [135].</p><!><p>Mucopolysaccharidosis IX, also known as hyaluronidase deficiency, is caused by mutations in the HYAL1 gene.</p><p>This disease was first discovered by Natowicz et al. [136] in a 14-year-old girl with short stature and multiple periarticular soft-tissue masses. Radiographic analysis showed nodular synovia, acetabular erosions, and a popliteal cyst. Lysosomal storage of hyaluronan (HA) was evident within the macrophages and fibroblasts of biopsied soft-tissue masses, and serum concentrations were elevated 38–90-fold. She was proven to have a storage disease of hyaluronan (hyaluronic acid) due to a genetic deficiency of hyaluronidase. The descriptions of hyaluronidase deficiency in this family are consistent with autosomal recessive inheritance.</p><p>In order to determine the molecular basis of MPS IX, Triggs-Raine et al. [137] analyzed 2 different candidate genes tandemly distributed on chromosome 3p21.3, both encoding proteins with homology to a sperm enzyme with hyaluronidase activity. These genes, HYAL1 and HYAL2, encode 2 distinct lysosomal hyaluronidases with different substrate specificities. When characterizing the patient with hyaluronidase deficiency originally reported in [136], they verified that he was a compound heterozygote for 2 mutations in the HYAL1 gene: a missense mutation (c.1412G>A), which introduced a nonconservative amino acid substitution in a putative active site residue (p.Glu268Lys) and a complex intragenic rearrangement, 1361del37ins14, which resulted in a premature termination codon. Through this work, they have also showed that these 2 hyaluronidase genes, together with a third adjacent HYAL3 gene, had markedly different tissue expression patterns, consistent with differing roles in the metabolism of hyaluronan. These findings allowed this team not only to explain the unexpectedly mild phenotype of MPS IX but also to predict the existence of other hyaluronidase-deficiency disorders.</p><p>Presently, three other hyaluronidase-related genes (HYAL4, HYALP1, SPAM1) have been identified at 7q31.3 [138]. These genes are predicted to encode hyaluronidases, endoglycosidases that initiate the degradation of HA, a large negatively charged GAG found in the extracellular matrix (ECM) of all vertebrate cells [19].</p><p>Since there is only one patient reported to date, the development and characterization of a model of Hyal1 deficiency was the first logical step in understanding the main phenotypic symptoms associated with MPS IX. During this decade, a mouse model for MPS IX has become available and was fully characterized [19]. Overall, it was observed that the murine MPS IX model displays the key features of the human disease. Nevertheless, during the same year, another mutant mouse suffering from a hyaluronidase deficiency was described, this one deficient in HYAL2 [139]. Skeletal and hematological anomalies were described in this model, raising the possibility that a similar defect, defining a new MPS disorder, exists in humans [139].</p><!><p>The elucidation of enzyme deficiencies underlying mucopolysaccharidoses was crucial to unveil the normal pathways of glycosaminoglycan catabolism. In fact, only through the consequences of their absence became the role of several enzymes evident. The majority of these enzyme deficiencies were discovered during the 1970s. Over the last decades the enzyme deficiencies underlying each disease, and the molecular defects causing them have been identified and extensively analyzed and characterized. As a result, six different MPS are known, caused by deficiencies in one of the ten different enzymes necessary to intralysosomal degradation of GAGs through one of the four different degradation pathways.</p><p>Each disease has its own hallmark features. Nevertheless, a common pattern arouses: MPS are usually chronic, with a progressive course and different severity degrees. Organomegaly, dysostosis multiplex, and CNS involvement are common but not necessary features.</p><p>Over the years, several MPS have been recognized in animals as naturally occurring diseases, and others were created by knock-out technology. Most animal colonies have been established from single related heterozygous animals, in such a way that the affected offspring is homozygous for the same mutant allele. All these models present disease pathology similar to that seen in humans, making the animals extremely valuable for both investigation of disease pathogenesis and testing of therapies. Large animal homologues are similar to humans in natural genetic diversity, approaches to therapy and care, and possibility of evaluating long-term effects of treatment. Presently, therapeutic strategies for MPS include enzyme replacement therapy, heterologous bone marrow transplantation, and somatic cell gene transfer, all of which have been tested in animals with some success. During the 80s, transplantation of hematopoietic stem cells was tested for several MPSs. Theoretically, haematopoietic stem cells taken from a normal compatible donor and transplanted into an enzyme-deficient recipient can provide a safe, permanent, and self-replicating source of bone marrow-derived cells. By secreting active lysosomal enzymes, these cells cross-connect nonbone marrow-derived cells. Several animal models for GAGs storage diseases have already been subjected to/undergone BMT. From those experiments, along with human clinical trials already tried, it was possible to verify that there are important variations in therapeutic response among different pathologies with some diseases with CNS pathology which can be successfully treated by BMT (the severe form of MPS I, being the example for the GAG storage disorders; [140]) whereas others cannot (MPS II and III in which BMT was tested with few success; [3, 141]). These variations are usually attributed to the different capacities of secretion, stability, and uptake of each specific enzyme. Nevertheless, important conclusions could be drawn from the collective experience of postnatal transplantation including the idea that the earlier the transplants are performed, the better the clinical response. In the 90s; however, a novel approach started to be tested: ERT. Nowadays, it has been the most tested approach in animal models of GAG storage disorders. Until now, the obtained results have been highlighting the potential of administered recombinant enzyme to reduce GAG accumulation. ERTs are presently available for MPS I (since 2003), II (since 2005), and VI (since 2006). Clinical trials are also in course for MPS IVA treatment through ERT. Nevertheless, this approach is ineffective for the brain since recombinant enzymes are not able to cross the blood-brain barrier (BBB). This is one of the reasons why other therapies are being tested for MPS with CNS involvement. ERT with direct administration of the recombinant enzyme into the brain (intrathecal injections) is also being considered in order to overcome that difficulty. Presently, such approaches are being considered for MPS IIIA and to overcome the cognitive deficit of MPS II and MPS I (reviewed in [142]).</p><p>Somatic cell gene transfer is another possible approach, but a long way needs yet to be travelled towards such a therapy is applicable to patients.</p><p>Finally, substrate deprivation therapy (or substrate reduction therapy) is also being considered for some MPSs. This approach is being tested with both genistein and rhodamine B (reviewed in [70]). Genistein, a chemical from the group of isoflavones, has been shown to inhibit the synthesis of GAGs in fibroblasts of patients with various forms of MPSs, namely, types I, II, IIIA, and IIIB [4, 143]. Similar results were obtained with rhodamine B, an inhibitor with an unknown mechanism of action. Remarkably, in MPS IIIA mice treated with rhodamine B, GAG storage decreased not only in somatic tissues, but also in brain, with improved behaviours of the animals [144, 145]. These encouraging results lead to the development of open-label pilot clinical studies with children suffering from Sanfilippo's syndrome types A and B in which a genistein-rich isoflavone extract (SE-2000, Biofarm, Poland) orally administered for 12 months. After one year of treatment, statistically significant improvement in all tested parameters was demonstrated (reviewed in [70]).</p><p>In order to better quantify and assess the efficacy of these therapeutic approaches, investigators have been trying to indentify suitable biomarkers for MPS, which allow the evaluation of short- and long-term treatment effects. This is also assuming particular importance since early detection of MPS is an important factor in treatment success. Recently discovered biomarkers include heparin cofactor II-thrombin complex and dermatan sulphate:chondroitin sulphate ratio [146]. Other biomarkers and/or therapeutic targets for MPS joint and bone disease recently identified through animal studies include several proinflammatory cytokines, nitric oxide, and matrix metalloproteinases (MMPs; [147]).</p><p>Another hot topic which is recently being discussed refers to the possibility of including some MPSs (particularly type I, IIIA, IIIB, and VI) in neonatal screening programs [148]. The ongoing development of enzyme replacement therapy and other treatments for several LSDs, including MPSs combined with the growing evidence that early commencement of therapy improves outcomes, has increased the pressure for the introduction of newborn screening programs, and a number of pilot studies are ongoing [148–152]. This is only possible thanks to the significant advances that were made in last decade since dried blood spot technology was introduced for enzymatic assays and lysosomal protein profile was developed.</p><p>Overall, there are encouraging results, with some therapeutic approaches already approved and others under development. Either way, it is important to stress that the management of MPS requires lifelong attention to the multisystemic involvement by a team of specialists experienced in dealing with these diseases, since none of the therapeutic options currently available result in complete resolution of morbidity.</p><!><p>Summary table of mucopolysaccharidoses.</p><p>Available therapeutic approaches for mucopolysaccharidoses.</p>

PubMed Open Access

Piezochromism and hydrochromism through electron transfer: new stories for viologen materials

While viologen derivatives have long been known for electrochromism and photochromism, here we demonstrated that a viologen-carboxylate zwitterionic molecule in the crystalline state exhibits piezochromic and hydrochromic behaviors. The yellow crystal undergoes a reversible color change to red under high pressure, to green after decompression, and finally back to yellow upon standing at ambient pressure. Ultraviolet-visible spectroscopy, X-ray photoelectron spectroscopy, electron paramagnetic resonance X-ray diffraction and DFT calculations suggested that the piezochromism is due to the formation of radicals via pressure-induced electron transfer from carboxylate to pyridinium, without a crystallographic phase transition. It was proposed that electron transfer is induced by pressure-forced reduction of intermolecular donor-acceptor contacts. The electron transfer can also be induced by dehydration, which gives a stable green anhydrous radical phase. The color change is reversible upon reabsorption of water, which triggers reverse electron transfer. The compound not only demonstrates new chromic phenomena for viologen compounds, but also represents the first example of organic mechanochromism and hydrochromism associated with radical formation via electron transfer.

piezochromism_and_hydrochromism_through_electron_transfer:_new_stories_for_viologen_materials

Introduction<!>Crystal structure<!>Piezochromism under uniaxial pressure<!>Studies under hydrostatic pressure<!>Further insight into the piezochromic process<!>Hydrochromism<!>Conclusions<!>Experimental<!>Crystal structure determination<!>High-pressure experiments<!>Other physical measurements<!>Computation

<p>Chromic materials that change color in response to external physical or chemical stimuli have been enduring topics of research for a variety of present and potential applications in our daily life and in various high-tech areas. 1 In general, the color change is related to a stimulus-induced change in electronic structure, for example, decreased band-gaps for semiconductors, 2 or modulated ligand-elds for d-block compounds. 3 For organic molecules, the change is oen caused by structural isomerization (such as ring-opening/ closing, 4 cis-trans isomerization, 5 and tautomerism 6 ), freeradical formation 7 or just conformational changes. 8 Various chromic phenomena, named aer the stimuli, such as photochromism, electrochromism, mechanochromism and solvatochromism, have been actively studied. Piezochromism, a subclass of mechanochromism, 9 refers to the reversible color change of a solid in response to external pressure. Recently there have been intense studies on piezochromic luminescence, where pressure induced a change in the photoluminescence color of the material. 10 In this paper, we are focused on innocent piezochromism, where pressure can induce a change in the natural color of the material. The phenomenon has been observed for various materials, including metal oxides/complexes (such as molybdates, 11 Cu(II) complexes, 2a,12 Ni(II) glyoximates 13 and Fe(II) spin-crossover compounds 14 ), polymers (such as polysilanes 15 and polythiophenes 16 ), organic molecules in polymer matrices, 17 and crystalline organic molecular solids. 18 Since covalent bonds are rather resistant to the effects of pressure, organic molecular solids respond to high pressure mainly by reducing intermolecular space and/or changing molecular conformations (if exible). 19 In the presence of appropriate chromophores, the pressure-forced structural changes may lead to a color change through modulation of electronic energy gaps. Some exible organic molecules with p-conjugation or donor-acceptor chromophores have been reported to exhibit piezochromism in crystalline states, where pressure enhances intermolecular p overlap or intramolecular coplanarity and thus leads to bathochromic shis of p-p* or charge-transfer (CT) absorptions. 8 In extreme cases, pressureforced structural changes can result in irreversible inter-/ intramolecular chemical transformations such as polymerization and isomerization, concomitant with irreversible (undesirable) color change. 18 Nevertheless, in a few favorable cases of spiropyrans and analogous compounds, reversible piezochromism has been achieved where pressure induces reversible ring-opening isomerization through C-O bond heterolysis. 18b,20 Piezochromism related to reversible bond homolysis (C-C or S-S), which leads to colored radicals, has also been reported for a few dimeric compounds (bis-dithiazolyl, bis-thioindoxyl and bis-chromenyl, bis-imidazolyl derivatives). 21 Such dimer-radical interconversion is also responsible for mechanochromism by grinding (i.e., tribochromism 9 ) in some dimeric compounds. 7f-j Viologen compounds (1,1 0 -disubstituted 4,4 0 -bipyridiniums, also oen called paraquat derivatives) have long been known for two well-established properties associated with their special electron-decient attribute: 22 the propensity to form chargetransfer complexes with electron-rich species, and the redox capability (Scheme 1) to undergo reversible electron transfer under chemical, electrical or optical stimuli. Thanks to these properties, the viologen unit has been widely used in the eld of supramolecular chemistry as the key building block for the assembly of mechanically interlocked molecules 23 and molecular machines. 24 The reversible redox activity involving brilliantly colored viologen radicals (Vc + ) has also allowed for widespread applications of viologens in the eld of chromic materials. In fact, viologens are the most extensively studied organic electrochromes 1a,22 and they can also be photochromic in the presence of appropriate electron donors. 25 In contrast to the well-known electro-and photochromism, we are unaware of any previous reports of mechanochromism related to radical formation through electron transfer, in either viologens or any other class of compounds. In this paper, we demonstrate reversible piezochromism observed with a zwitterionic viologen compound bearing carboxylate groups. The piezochromism involves radical formation/quenching via reversible electron transfer between carboxylate and viologen, which could be related to pressure-controlled intermolecular donor-acceptor contacts. Notably, the mechanism for radical formation in the viologen compound (electron transfer) is clearly distinguished from that in previous mechanochromic compounds involving radical formation (bond homolysis). 7f-j,21 In addition, we show that the compound behaves as a "chromic sponge": the release and re-absorption of lattice water induce reversible electron transfer and consequent color change. Reversible solid-state hydrochromism related to dehydration-hydration has long been known in transition metal compounds, 26 among the best known examples being copper sulfate pentahydrate and cobalt nitrate hexahydrate. The phenomenon has also been observed in organic species. 27 Very recently, some oxazolidines and oxazines were shown to be hydrochromic owing to water-triggered reversible ring-opening isomerisation, and have been tested for water-jet rewritable printing. 28 To the best of our knowledge, solid-state organic hydrochromism through the electron transfer mechanism is unprecedented.</p><!><p>The compound, N,N 0 -bis(carboxylatophenyl)-4,4 0 -bipyridinium hexahydrate (bpybdc$6H 2 O, 1) was synthesized by a multi-step procedure involving the Zincke reaction of N,N 0 -bis(2,4-dinitrophenyl)-4,4 0 -bipyridinium dichloride and ethyl 4-aminobenzoate (Scheme S1 †).</p><p>The structure was determined by single-crystal X-ray crystallography (Fig. 1). The zwitterion bpybdc is centrosymmetric with the central bipyridinium moiety being strictly planar. The terminal benzoate moiety is quasi-planar with a dihedral angle of 2.0(2) between the benzene ring and the carboxylate group, and the benzene ring is twisted with respect to the bipyridinium Scheme 1 Redox processes of viologens. plane by a dihedral angle of 33.85(4) . Neighboring benzoate moieties from different molecules are nearly parallel [dihedral angle: 0.82(2) ] and held together by extended p-p stacking interactions to form a one-dimensional zipper-like array along the c direction (Fig. 1b). The average atom-to-plane distance is 3.476(1) Å, and the centroid-centroid distance is 3.818( 1 S1 and Fig. S1 †).</p><!><p>Compound 1 was synthesized as yellow crystals. No color change was observed when the crystals were ground. When preparing the KBr disc with a hydraulic press for IR spectroscopy, we unexpectedly obtained a green disc aer release of pressure. The same color change was observed when pure samples of 1 were pressed under a pressure estimated at 2.8 GPa (Fig. 2a). A control experiment in the dark was performed to conrm that light has no contribution to the color change. The piezochromic phenomenon was conrmed by UV-vis diffuse reectance spectra (Fig. 2b). The original sample of 1 displays intense absorption bands below 400 nm, with a shoulder band at about 470 nm. The visible-region absorption is responsible for the yellow color and attributable to the CT transition from the electron-rich carboxylate group to the electron-decient viologen moiety. 22,30 The green sample (1A) obtained aer release of pressure shows much more intense absorption in the visible light region, with four maxima at 480, 610, 670 and 740 nm. The differences indicate that the compressiondecompression process has caused dramatic changes in electronic structure. The powder X-ray diffraction (PXRD) prole of 1A is very similar to that of 1 (Fig. 2c), with only some differences in relative intensity because of changes in preferential orientation. The observation suggests that there are no signicant crystallographic differences between the original and nal states. Furthermore, the IR spectra (Fig. S2 †) of 1 and 1A are essentially identical, so no signicant difference is expected between the molecular structures of the two states. Considering the ability of the viologen moiety to form the Vc + radical via oneelectron transfer, we performed electron paramagnetic resonance (EPR) measurements to probe the piezochromic mechanism of 1 (Fig. 2d). It proved that 1 is EPR silent but 1A presents a single strong signal with g ¼ 2.0035, indicative of the presence of radicals. Therefore, piezochromism of 1 involves the formation of radicals through pressure-induced electron transfer. In other words, the compression-decompression process induces a redox chemical transformation from closed-shell to radical states. The two states under ambient pressure have no detectable differences in molecular and crystal structure but show distinctly different electronic structures. Density functional theory (DFT) calculations of the density of states (DOS) using the crystallographic data of 1 suggest that the top of the valence band and the bottom of the conduction band are overwhelmingly dominated by the carboxylate and pyridinium groups, respectively (Fig. 3a and S3 †). Orbital analysis shows that the highest occupied crystal orbital (HOCO) arises almost solely from the nonbonding p orbitals of the carboxylate oxygen atoms, and the lowest unoccupied crystal orbital (LUCO) is dominated by the p-p* orbitals of the viologen moiety (Fig. 3b). Therefore, it is reasonable to assume that the electron transfer occurs from carboxylate oxygen to pyridinium, generating viologen and carboxylate radicals.</p><p>X-ray photoelectron spectroscopy (XPS) was performed to gain experimental evidence for the direction of electron transfer. As shown in Fig. 4, the original state of the compound shows a single N 1s core-level peak at 402.3 eV due to the unique pyridinium. The carboxylate and water O atoms in the structure lead to a peak at 530.3 eV with a weak shoulder at 532.9 eV. Aer compression, a lower-energy shoulder (400.1 eV) appears in the N 1s spectrum, and the high-energy shoulder of the O 1s spectrum is signicantly increased in intensity. The C 1s signals show a slight decrease in the high-energy shoulder. The 5406 Å), and EPR spectra (d) of 1 and 1A. phenomena are in good agreement with the occurrence of electron transfer from carboxylate to pyridinium. 7a,31 The yellow and green colors can be reversibly switched. The pressure-induced green state is metastable. It completely recovers to yellow upon standing overnight, and the recovery can be accelerated by gentle heating below 90 C (higher temperature causes dehydration of the compound; see the Hydrochromism section). The recovered sample is EPR silent (Fig. 2d), suggesting quenching of the radical. The reversible switching between the two states could be repeated for at least 10 cycles, with no apparent sign of fatigue (Fig. S4 †). Notably, the recovery process is independent of the atmosphere and can occur in air, nitrogen and vacuum, so the possibility that radicals are quenched by molecular oxygen can be precluded. Radical quenching should proceed through spontaneous back electron transfer from viologen to carboxylate.</p><!><p>The piezochromic properties of 1 were further investigated under hydrostatic pressure generated in diamond anvil cells (DACs). The DAC technology can not only provide isotropic gigapascal pressure but also allows us to study the color and spectral changes during compression and decompression.</p><p>With the DAC equipment, we found that the pale yellow crystal turns red gradually as the pressure increases. No green color was observed in the compression process, but the crystal begins to show green tints when decompressed to about 2 GPa, and becomes green upon further decompression. UV-vis absorption spectra were recorded under various pressures using a single crystal in a DAC. As shown in Fig. 5a, under a relatively low pressure (0.09 GPa), the absorption in the visible region is dominated by the CT band at about 470 nm, as observed at atmospheric pressure. Upon compression above 2 GPa, new bands appear in the region of 600-800 nm, indicative of radical formation. As the pressure increases, the bands become more intense and show gradual bathochromic shis. The increase in intensity could be due to the increasing concentration of radical species. The bathochromic shis could be due to the modulating effects of pressure on intermolecular interactions and molecular conformation. 18b During decompression, the bands show a regular blue shi and decrease in intensity. However, even aer complete removal of pressure, the radical signals still remain, with the maxima at essentially the same positions as observed for the green sample obtained aer compression with a hydraulic press. The study indicates that the piezochromism of 1 under hydrostatic pressure also proceeds through an electron transfer mechanism. The color of the radical state is red under high pressure, consistent with the much stronger absorption below 600 nm, and it changes to green aer PXRD and IR studies have suggested that the nal green state obtained aer the removal of pressure shows no signicant differences from the original yellow state in the molecular and crystal structures. To probe the possible structural change during the piezochromic process, in situ PXRD measurements in a DAC were performed using a synchrotron radiation source. As shown in Fig. 6a, under increasing pressure up to 8.4 GPa, the compound remains crystalline, and the diffraction peaks are systematically shied to higher angles, which corroborates a volume contraction upon compression. The patterns have been indexed according to the crystallographic data of 1. The rened unit cell parameters are plotted against pressure in Fig. 6b and c The in situ PXRD study indicates that the piezochromic process does not involve crystallographic phase transitions.</p><p>In addition, in situ Raman scattering spectra were measured under 0-8.32 GPa. As shown in Fig. S5, † all peaks shi to higher frequencies, as is expected for molecules under high pressure.</p><p>No new peak was detected and the spectrum was recovered concomitantly when pressure was removed. The results suggest that the compression/decompression process does not involve bond formation or cleavage in the pressure range investigated, 32 as expected for the redox process of the viologen unit.</p><!><p>Photochromism in crystalline viologen-derived compounds usually involves intermolecular electron transfer, and the pathway for the transfer has been related to a short contact between the pyridinium N atom and the donor atom. 22,33 With carboxylate donors, the N/O distance enabling photo-induced intermolecular electron transfer is usually around 3.5 Å. 7d,29 No unequivocal criterion has been established, perhaps because the process could be inuenced by the nature and various environmental factors of the donor and the acceptor. For piezochromism in viologen compounds, no previous example is available for comparison. The inuencing factors for piezochromism could be different from those for photochromism, because pressure and light are disparate in nature and in the way they interact with solids. Nevertheless, shorter donoracceptor distances should always be favorable for electron transfer, regardless of the stimulus. In this sense, we can envision that pressure could be a more powerful energy tool than light for induction of intermolecular electron transfer (and chromic phenomena) in viologen crystals. Light can be absorbed by individual molecules and have no direct effect on intermolecular distances, so it would not induce electron transfer if the donor and the acceptor were not placed at a favorable distance. Quite differently, pressure works directly on intermolecular distances, so it can proactively reduce the distance between the donor and the acceptor in favor of electron transfer. The fact that compound 1 is piezochromic but not photochromic (vide infra) could be an illustration of the power of pressure to induce intermolecular electron transfer.</p><p>To better understand the pressure-induced electron transfer, DFT geometry optimization using the cell parameters obtained from high-pressure XRD was performed to calculate the molecular packing structure under high pressure. The results show that the zwitterionic molecule is less twisted under pressure, with the angle between the pyridinium and phenyl rings being reduced from 33.8 to 21.9 under 3.62 GPa. The center-tocenter distance between the p-p stacking benzene rings is reduced from 3.82 to 3.54 Å (Fig. 7). Accordingly, the intermolecular N/O distances between viologen and carboxylate moieties change from 3.51 and 3.70 Å at ambient pressure to 3.21 and 3.38 Å under 3.62 GPa. The signicantly reduced donor-acceptor contacts could facilitate the electron transfer. Another possible pathway, which has been occasionally referred to in photochromic study, is the hydrogen bond (C8-H/O2) between the carboxylate oxygen and the pyridinium H atom adjacent to the N atom. 29a The H/O contact is signicantly shortened from 2.29 to 1.91 Å under 3.62 GPa, while the C-H/ O angle undergoes a very minor change (<1 ). It is difficult to denitively decide between inter-and intramolecular electron transfer. The above theoretical study supports intermolecular pathways more than intramolecular ones, because the molecular conformation remains twisted under pressure while the intermolecular donor-acceptor contacts are signicantly reduced. This is consistent with the fact that the intermolecular pathway has been well established in photochromic studies of viologen compounds in both solution and solid phases. We performed further experiments to check whether piezochromism can occur in methyl viologen (MV 2+ ) based systems, where only intermolecular pathways are possible. We obtained a solid by concentrating the mixture solution of [MV]Cl 2 and benzoic acid in ethanol. The solid and a pure sample of [MV]Cl 2 were compressed with a hydraulic press. No color change was observed for [MV]Cl 2 , but the solid obtained from [MV]Cl 2 and benzoic acid turned from white to violet. The violet color and the intense absorption covering the region from 400 to 700 nm are typical of the MVc + radical. 22 Although the structure is unclear, the radical can only be generated by intermolecular electron transfer (Fig. S6 †). The simple experiments give us a clue to speculate that the piezochromism of 1 could proceed through intermolecular pathways.</p><p>The study also demonstrates that pressure is a powerful tool to induce chromic phenomena via the intermolecular electron transfer mechanism. We can envision that piezochromism could be a common phenomenon for viologen-based molecular solids if appropriate donors are present. This is to be demonstrated by extended studies.</p><p>According to the above studies, the piezochromic process of 1 can be proposed as follows. Under high external pressure, the molecules in the lattice are forced to move closer. The intermolecular contacts between the electron-rich carboxylate group and the electron-decient viologen moiety are shortened to favorable distances to induce electron transfer from carboxylate to viologen. This leads to a metastable state with two radical sites per molecule, one delocalized over the viologen moiety, the other over the carboxylate groups. There could be intermolecular radical-radical interactions, which, together with planar conguration of the viologen moiety, should help to stabilize the radical state. It could be assumed that the reverse electron transfer is slow and cannot keep pace with the recovery of the intra-and intermolecular parameters upon decompression. Therefore, although the molecular and crystal structure is completely recovered aer removal of pressure, the radical state remains in a certain proportion. The back transfer process continues upon further standing at ambient pressure to nally regenerate the original state. Here, the key to the piezochromism is the formation of radicals via pressure-induced intermolecular electron transfer. The radical mechanism is different from those where pressure modulates p-p* and CT absorptions by strengthening intermolecular p overlapping, reducing intramolecular twisting; 8 it is also different from the cases in which pressure induces bond cleavage/formation. 20,21 Finally, it should be mentioned that not only electron transfer but also the gradual structural changes contribute to the piezochromic phenomena of 1. Electron transfer leads to new absorption bands, and the gradual shis of the bands during the compression/decompression process should be attributed to the modulation effects of pressure on intermolecular interactions and molecular conformation.</p><!><p>Many viologen derivatives are photochromic. In an initial test, a yellow sample of 1 turned green upon irradiation with a 300 W Xe lamp, which seemed to suggest a photochromic behavior. However, when the sample was cooled with jacket water to avoid heat accumulation from the Xe lamp, no color change was observed even under prolonged irradiation. This precludes photochromism but suggests a thermal effect. Indeed, heating 1 at 100 C led to a green sample (Fig. 8).</p><p>Thermogravimetric analysis (TGA), variable-temperature IR and PXRD measurements were performed to investigate other thermal effects on 1. TGA and IR (Fig. S7 and S8 †) suggest that 1 loses the water of crystallization upon heating up to 100 C. As shown in Fig. 9, the PXRD prole of the anhydrous green phase is completely different from that of 1. The new phase (hereaer denoted as 1B) is crystalline and remains unchanged up to at least 250 C. In addition, the XRD pattern of the green sample obtained by Xe-light irradiation without cooling is identical to that of 1B (Fig. 9), conrming that the color change under the Xe lamp actually corresponds to the loss of water molecules due to the thermal effect of the lamp.</p><p>The UV-vis, EPR and XPS spectra of 1B are very similar to those of 1A (Fig. S9-S11 †), so the dehydration-induced color change from 1 to 1B is also due to the formation of radicals through one-electron transfer from carboxylate to viologen. Although the crystal structure of the anhydrous phase is unclear, it could be assumed that the loss of lattice water leads to closer crystal packing of the zwitterionic molecules so that the distance between electron donors and acceptors is reduced in favor of electron transfer.</p><p>The color change occurs readily by heating in the dark and hence electron transfer can occur without the aid of light stimulus. To clarify the role of heating, we have heated 1 at 110 C in the atmosphere of water steam. No color change was observed, indicating that heating alone cannot induce electron transfer. Therefore, water release is the most essential factor, while heating just acts to facilitate water release. Particularly worth mentioning is that the color change also occurs if the compound is placed in a desiccator containing sodium hydroxide as desiccant for several hours (Fig. S12 †). The phenomenon clearly conrms that dehydration can induce electron transfer even without the aid of heating, though heating can accelerate dehydration.</p><p>Viologen derived radicals can usually be quenched by molecular oxygen in air with concomitant color fading, and the quenching/fading process is usually thermally accelerated. Unexpectedly, the anhydrous radical state 1B shows remarkable stability against temperature and oxygen. The green color persists upon heating up to 250 C in dry air, even in a pure oxygen atmosphere. However, it can readily return to yellow in a moist atmosphere, and the color change occurs immediately if direct contact with liquid water is allowed. The yellow product obtained is EPR silent (Fig. S10 †), and XRD studies indicate that the hydrate phase 1 is recovered (Fig. 9). It is worth noting that the reverse color change occurs without the aid of oxygen and heating, illustrating an innocent hydrochromic process. It can be assumed that insertion of water molecules back into the lattice sets the organic molecules apart from each other. The structural change induces back electron transfer so that radicals are quenched and thus color recovered.</p><p>The dehydration-hydration switched color change can be repeated for at least 10 cycles (Fig. 8b). Finally, it is worth mentioning that the anhydrous green phase 1B does not change color when in contact with other solvents such as methanol, ethanol, acetonitrile and chloroform (either vapor or liquid), so the solvatochromic behavior is highly selective towards water.</p><!><p>In summary, we have demonstrated interesting crystallinestate piezochromic and hydrochromic phenomena with a zwitterionic molecule containing anionic carboxylate as electron donor and cationic viologen as acceptor. The exertion of high pressure or the release of lattice water can induce electron transfer from carboxylate to viologen, generating radicals and thus leading to color change. The process is reversible via back electron transfer upon standing at ambient pressure or upon reabsorbing water. Despite their remarkably different nature, the physical stimuli (compression/decompression) and the chemical stimuli (dehydration/rehydration) have one thing in common: the ability to modulate intermolecular donor-acceptor contacts in favor of electron transfer. This work tells new chromic stories for the old family of viologen compounds well-known for electro-/photochromism, and presents the rst example of organic mechanochromism and hydrochromism associated with radical formation through electron transfer. Most importantly, we demonstrated for the rst time that the physical stimulus of pressure can induce single-electron transfer in viologen species, which used to be triggered by electricity and light. It is hoped that this fundamental nding can evoke further studies to explore the generality of pressure-induced electron-transfer in viologen derivatives and other redox-active organic species and to understand the physical and chemical factors inuencing the processes. Considering the large family of viologen compounds and the diversity of redox-active organic compounds, the studies may open new ground and prospects in piezochromic organic materials and other related elds of piezochemistry. The results may also have implications for supramolecular studies dealing with stimulus-responsive organic systems.</p><!><p>Materials and synthesis 4,4 0 -bipy, 2,4 0 -dinitrochlorobenzene, ethyl p-aminobenzoate and NaOH in AR grade were purchased commercially without further purication. Water was deionized and distilled before use. N,N 0 -Bis(2,4-dinitrophenyl)-4,4 0 -bipyridinium dichloride were prepared according to the literature. 34 [H 2 bpybdc]Cl 2 . The synthetic routes of [H 2 bpybdc]Cl 2 were shown in Scheme S1. † A mixture of N,N 0 -bis(2,4-dinitrophenyl)-4,4 0 -bipyridinium dichloride (1.1 g, 2.23 mmol) and ethyl 4-aminobenzoate (1.3 g, 7.87 mmol) was dissolved in 50% EtOH (30 ml) and reuxed for 24 h. Aer the mixture was cooled to room temperature, the solvent was evaporated and the residue dissolved in H 2 O. Aer ltering off the insoluble solid, the ltrate was washed with Et 2 O three times and evaporated to give a crude product of 1,1 0 -bis(4-ethoxycarbonylphenyl)-4,4 0 -bipyridinium dichloride ([Et 2 bpybdc]Cl 2 ). The products were recrystallized from methanol/ethyl acetate (v/v, 1/2) to give a yellow powder. Yield: 0.35 g (30%).</p><p>[Et 2 bpybdc]Cl 2 (0.35 g) dissolved in concentrated HCl (9 ml) was reuxed for 4 h. Aer cooling to room temperature, the precipitated solid product was ltered out, washed with water and dried under vacuum. [H 2 bpybdc]Cl 2 was obtained as a pale yellow powder. Yield: 0.30 g (91%). IR (KBr, cm À1 ): 3506m, 3401m, 3112m, 3047w, 2572w, 1714s, 1633s, 1602s, 1546s, 1496s, 1444s, 1417s, 1380s, 1305m, 1174m, 1130w, 1106s, 865m, 836m, 798s, 769s, 723s, 696s, 640s, 572w, 545m, 520m, 501m. 3581w, 3122w, 2510w, 1699w, 1635s, 1606s, 1565s, 1542s, 1494s, 1438s, 1392s, 1375s, 1251w, 1230m, 1162w, 1137w, 1120w, 1037m, 1022m, 1004m, 875w, 840w, 777m, 690m, 534m, 470m. 1 H NMR (400 MHz, D 2 O, ppm): 7.80 (d, J ¼ 8 Hz, 4H), 8.08 (d, J ¼ 8 Hz, 4H), 8.73 (s, 4H). Note: the 2,6 H atoms (neighboring to the N atom) of the pyridinium ring were not observed in the 1 H NMR spectrum, owing to the fast H-D exchange with D 2 O under basic conditions. Consequently, the 3,5-H atoms appear as a singlet at 8.73 ppm. 35 The zwitterionic compound is sparsely soluble in common solvents such as water, methanol, acetonitrile, and DMF.</p><!><p>Diffraction intensity data of 1 were collected at 293 K on a Bruker APEX II diffractometer equipped with graphite-monochromated Mo-Ka radiation (0.71073 Å) and a CCD area detector. Empirical absorption corrections were applied using the SADAB program. 36 The structures were solved by the direct method and rened by the full-matrix least-squares method on F 2 using the SHELXL program, 37 with anisotropic displacement parameters for all non-hydrogen atoms. The hydrogen atoms attached to carbon atoms were placed in calculated positions and rened using the riding model. The water hydrogen atoms were located from the difference Fourier map.</p><p>Crystal data for 1.</p><!><p>For high pressure PXRD measurements, a sample was ground into ne powder and loaded into the sample chamber (100 mm in diameter) of a pre-indented stainless steel gasket in a diamond anvil cell (DAC) with 300 mm culets. A ruby sphere was then loaded to conrm pressure of the sample by using the ruby uorescence. 38 Silicone oil was used as a pressure-transmitting medium. In situ PXRD was performed at beamline 15U1 of Shanghai Synchrotron Radiation Facility (SSRF) using a wavelength of 0.6199 Å. 2D Debye-Scherrer diffraction rings from powder measurements were collected on a Mar345 image plate and integrated using the FIT2D soware package. 39 Aer each compression, the samples were allowed to equilibrate in the DAC until ruby uorescence (pressure) was invariant. The pressure was veried again at the end of each diffraction experiment. The PXRD peaks were indexed according to the single-crystal data of 1, and the unit cell parameters were rened using the least-squares method.</p><p>The in situ Raman scattering spectra were obtained in a DAC (500 mm culets) using a Renishaw 1000 spectrometer with a 532 nm excitation laser. The in situ UV-vis absorption measurements under high pressure were performed on an Ocean Optics QE65000 scientic-grade spectrometer with a DAC (300 mm culets) containing a single crystal of 1. A ruby sphere was loaded to determine pressure by using the ruby uorescence, and silicone oil was used as a pressure-transmitting medium.</p><p>All high pressure experiments were conducted at room temperature.</p><!><p>1 H NMR spectra were recorded on a Bruker Avance 400 MHz spectrometer. The FT-IR spectra were recorded in the range 500-4000 cm À1 using KBr pellets on a Nicolet NEXUS 670 spectrophotometer. Elemental analyses were performed on an Elementar Vario ELIII analyzer. Thermogravimetric analyses (TGA) were carried out on a Mettler Toledo TGA/SDTA851 instrument under owing air at a heating rate of 10 C min À1 . Powder X-ray diffraction (PXRD) at ambient pressure was recorded on a Rigaku D/Max-2500 diffractometer at 35 kV, 25 mA for a Cu-target tube and a graphite monochromator. Optical diffuse reectance (UV-vis) spectra were measured using a SHIMADZU UV-2700 spectrophotometer, with BaSO 4 plates as references (100% reection). Electron spin resonance (EPR) spectra were recorded on a Bruker Elexsys 580 spectrometer with a 100 kHz magnetic eld in the X band at room temperature. X-ray photoelectron spectroscopy (XPS) studies were performed with a PHI 5000 Versaprobe spectrometer using Al Ka radiation (l ¼ 8.357 Å). To compensate for surface charging effects, all XPS spectra were referenced to the C 1s neutral carbon peak at 284.6 eV.</p><!><p>Based on the X-ray crystallographic data of 1, DFT (density functional theory) calculations were performed to analyze the density of states (DOS) and the frontier orbitals. The DMol 3 module 40 in the Materials Studio soware package 41 was used with ne accuracy for the numerical integration of the Hamiltonian and a ne (10 À6 eV per atom) tolerance for SCF convergence. The DFT exchange-correlation potential was described by the Perdew-Burke-Eruzerhof (PBE) functional within the generalized gradient approximation (GGA). 42 The Tkatchenko-Scheffler (TS) scheme was applied for dispersion corrections. 43 All electrons were included in the calculations and the DNP (double numerical plus polarization) basis set was used with a ne orbital cutoff quality.</p><p>To investigate the effects of external pressure on molecular geometry and packing, the crystal structure of 1 under 3.62 GPa was optimized by DFT calculations using the CASTEP program. 44 The cell parameters obtained from the high-pressure XRD experiment under 3.62 GPa were used without cell optimization. The optimization applied the GGA-PBE exchange-correlation functional and the TS dispersion-correction scheme. A ne optimization convergence level was selected, with an energy tolerance of 10 À5 eV per atom, a maximum displacement of 0.001 Å, and a maximum force threshold of 0.03 eV ÅÀ1 . The norm-conserving pseudopotentials in reciprocal space were used for the electronic Hamiltonian, with an energy cut-off of 750 eV for the plane wave basis set and a SCF convergence tolerance of 10 À6 eV per atom.</p>

Royal Society of Chemistry (RSC)

A Highly Chemo-, Regio-, and Stereoselective Metallacycle-Mediated Annulation Between a Conjugated Enyne and an Ene-Diyne

Alkoxide-directed metal-centered intermolecular [2+2+2] annulation is shown to chemo-, regio-, and stereoselectively engage two polyunsaturated substrate in productive cyclization chemistry. This annulation process is unique in the field, revealing that it is possible to selectively engage three of five \xcf\x80-systems residing in the coupling partners in initial [2+2+2] reaction, and demonstrating that one of the two remaining \xcf\x80-systems (the TMS-alkyne) can ultimately serve to simply generate a new metallacyclopentene of great potential value in additional metallacycle-mediated coupling chemistry.

a_highly_chemo-,_regio-,_and_stereoselective_metallacycle-mediated_annulation_between_a_conjugated_e

Introduction<!>Results and Discussion<!>Conclusions<!>Synthesis of (3R,4R)-4-((tert-butyldimethylsilyl)oxy)-4-(furan-2-yl-3-methylbutan-1-ol (14):<!>Analytical Data for 14:

<p>Since the pioneering work of Professor Vollhardt that revealed the great value of metal-centered [2+2+2] annulation chemistry for accessing stereodefined carbo- and heterocyclic targets, this area of organometallic chemistry has grown to occupy a unique role in organic synthesis (Figure 1A).1–5 These reactions that may be considered to be inspired by the foundation of chemistry first established by Reppe,6 and later developed by luminaries in the field, often embrace organometallic intermediates of relevance for [2+2+2] annulation as a means to realize a large swath of complex and diverse bond-forming reactions.7–13 While many of these advances have been demonstrated to be of great value in complex molecule or natural product synthesis,14 it is typically the case that intermolecular versions of this type of organometallic chemistry are rather restricted in scope, and therefore of limited value in stereoselective synthesis, including natural product synthesis.14 As illustrated in Figure 1B, controlling intermolecular reactivity in metallacycle-mediated coupling chemistry is challenging due to the possibility of generating products of varied composition, in addition to the clear issues of selectivity that surface when employing: (a) unsymmetrical systems (regioselectivity), (b) alkenes as one of the three reactive π-systems (π-facial selectivity), and/or (c) polyunsaturated substrates where issues of chemoselectivity surface.</p><p>For over ten years, we have been interested in establishing principles that ultimately result in avoiding the common dependence on intramolecularity and/or use of symmetrical substrates in this rather broad area of chemistry. To date, we have reported over thirty distinct and highly selective coupling reactions between a variety of unsymmetrically substituted unsaturated systems, and have demonstrated the utility of these reactions in total synthesis and function-oriented synthesis15 of representative members of diverse families of natural products (alkaloids, terpenoids, and polyketides).16–27 These reactions, termed metallacycle-mediated cross-coupling processes, have recently grown to include [2+2+2] annulation reaction between alkynes and 1,6- and 1,7-enynes as a means to prepare angularly substituted hydrindanes and decalins (Figure 2A).28–32 Here, in efforts that aimed to explore the question of whether it is possible to chemoselectively engage two polyunsaturated substrates in metallacycle-mediated annulative cross-coupling, we have elucidated the first organometallic transformation of its kind, where four of five resident π-unsaturations in the overall reaction are embraced as sites of reactivity in a highly selective annulation reaction (Figure 2B).</p><!><p>To explore the coupling reaction depicted in Figure 2B, efforts were first directed toward the synthesis of the desired ene-diyne 13. As illustrated in Figure 3A, epichlorohydrin was easily advanced to the desired ene-diyne 13 by initial reaction with isopropenylmagnesium bromide and CuI, followed by treatment with base to generate the low molecular weight alkenyl epoxide 16.33 With this intermediate in hand, regioselective opening of the epoxide with the Li-acetylide of diyne 17, promoted by the action of BF3•OEt2, delivered the desired coupling partner 13 in 66% yield.</p><p>Ti-mediated annulation was then pursued as depicted in Figure 3B. Initial chemoselective activation of the commercially available enyne 12 by the action of Ti(Oi-Pr)4 and n-BuLi, followed by addition of the lithium alkoxide of ene-diyne 13 as a solution in PhMe, and quenching of the reaction with aqueous NaHCO3, resulted in the formation of the cross-conjugated triene-containing hydrindane 14, isolated in 51% yield as a single regio- and stereoisomer – no evidence was found for the formation of any other isomer of this coupled product, nor was evidence found for the presence of the TMS-alkyne containing product 15.32</p><p>As depicted in Figure 3C, a mechanistic rationale for the formation of 14 is based on initial alkoxide-directed reaction between the preformed Ti–alkyne complex of 12 (I) with the internal alkyne of ene-diyne 13 that is proximal to the hydroxy group. The resulting fully substituted metallacyclopentadiene (III) can then participate in stereoselective intramolecular [4+2] to deliver the bridged bicyclic metallacyclopentene intermediate IV. We currently believe that this intermediate serves as the source of Ti for ligand transfer to the third reactive alkyne of this substrate combination (the TMS-alkyne of the ene-diyne coupling partner 13), resulting in the intermediacy of a new titanacyclopropene intermediate (VI). Interestingly, this intermediate organometallic does not appear competent to engage the 1,1-disubstituted alkene in carbometalation chemistry en route to the generation of a new carbocyclic system (VI → VII → 15).34 Rather, aqueous quenching of this reaction delivers the tricycylic product 14 that contains a pendent (Z)-vinylsilane.</p><!><p>A unique and highly selective intermolecular [2+2+2] annulation reaction of polyunsaturated substrates has been described. The reaction is believed to be possible due to the power of alkoxide-direction as a force to control the regio- and stereochemical course of metallacycle-mediated coupling reactions. Observations made in the course of this study have led to the hypothesis that a pendent alkyne on the enyne substrate has the potential to engage the bridged bicyclic organometallic intermediate generated in the course of [2+2+2] annulation in a simple ligand exchange reaction, and defines a unique sequence of chemical transformations of potential value in more complex multi-component coupling chemistry. We look forward to exploring the value of this new sequence for addressing challenging problems associated with the synthesis of terpenoid and terpenoid-inspired targets.</p><!><p>To a solution of 2-methyl-4-trimethylsilyl-1-butene (12, 1.288 mL, 7.24 mmol) in anhydrous toluene (45 mL) was added Ti(O-i-Pr)4 (2.145 mL, 7.24 mmol) at room temperature. The mixture was cooled to −78 °C and n-BuLi (2.46 M in hexanes, 5.87 mL, 14.46 mmol) was added drop-wise via syringe. The reaction flask was removed from the cooling bath and the mixture was warmed to room temperature before heating to 50 °C (without a reflux condenser) for 1 h. After this period, the reaction solution was cooled to room temperature and then placed in a −78 °C cooling bath.</p><p>Simultaneously, ene-diyne 13 (0.600 g, 2.41 mmol) was dissolved in anhydrous toluene (15 mL) and treated with n-BuLi (2.46 M in hexanes, 0.981 mL, 2.41 mmol) at −78 °C. The reaction mixture was stirred at −78 °C for 15 minutes, and then warmed to room temperature over 5 minutes. The alkoxide solution was added dropwise, via cannula, to the Ti–alkyne complex and then gradually warmed to room temperature overnight (13 h). The reaction was quenched with saturated aqueous NaHCO3 (40 mL), and the organic and aqueous phases were separated. The aqueous layer was extracted with EtOAc (× 3), and the combined organic phases were dried over anhydrous MgSO4. The supernatant was removed from the drying agent by vacuum filtration through a glass fritted funnel, and the solvents were removed in vacuo to afford the crude product, which was purified by flash column chromatography with 89:11 hexanes–EtOAc to afford 14 (475 mg, 51%) as a clear, colorless, amorphous solid.</p><!><p>TLC (SiO2) Rf = 0.30 (hexanes–ethyl acetate, 87:13); 1H NMR (600 MHz, CDCl3) δ 6.26 (dt, J = 14.0, 7.0 Hz, 1H), 5.46 (d, J = 14.2 Hz, 14H), 4.99 (app s, 1H), 4.80 (br s, 1H), 4.48 (app quintet, J = 6.9 Hz, 1H), 2.83 (dd, J = 17.7, 7.4 Hz, 1H), 2.36 (dd, J = 17.7, 6.3 Hz, 1H), 2.22 (d, J = 15.4 Hz, 1H), 2.17–2.02 (m, 5H), 2.01 (d, J = 15.4 Hz, 1H), 1.77 (br s, 3H), 1.50 (dd, J = 12.3, 7.8 Hz, 1H), 0.84 (s, 3H), 0.10 (s, 9H), 0.09 (s, 9H); 13C NMR (150 MHz, Chloroform-d) δ 149.94, 149.22, 146.17, 145.30, 129.10, 128.47, 127.95, 115.77, 72.19, 51.09, 40.73, 39.57, 38.75, 33.68, 29.82, 24.64, 21.02, 0.44; IR (neat) 3315, 2953, 2924, 1606, 1246, 850, 836 cm–1; HRMS (ES-TOF) m/z [M + H]: calcd for C23H41OSi2 389.2696; found 389.2701.</p>

PubMed Author Manuscript

On‐Surface Synthesis of NBN‐Doped Zigzag‐Edged Graphene Nanoribbons

AbstractWe report the first bottom‐up synthesis of NBN‐doped zigzag‐edged GNRs (NBN‐ZGNR1 and NBN‐ZGNR2) through surface‐assisted polymerization and cyclodehydrogenation based on two U‐shaped molecular precursors with an NBN unit preinstalled at the zigzag edge. The resultant zigzag‐edge topologies of GNRs are elucidated by high‐resolution scanning tunneling microscopy (STM) in combination with noncontact atomic force microscopy (nc‐AFM). Scanning tunneling spectroscopy (STS) measurements and density functional theory (DFT) calculations reveal that the electronic structures of NBN‐ZGNR1 and NBN‐ZGNR2 are significantly different from those of their corresponding pristine fully‐carbon‐based ZGNRs. Additionally, DFT calculations predict that the electronic structures of NBN‐ZGNRs can be further tailored to be gapless and metallic through one‐electron oxidation of each NBN unit into the corresponding radical cations. This work reported herein provides a feasible strategy for the synthesis of GNRs with stable zigzag edges yet tunable electronic properties.

on‐surface_synthesis_of_nbn‐doped_zigzag‐edged_graphene_nanoribbons

<!>Conflict of interest<!>

<p>Y. Fu, H. Yang, Y. Gao, L. Huang, R. Berger, J. Liu, H. Lu, Z. Cheng, S. Du, H.-J. Gao, X. Feng, Angew. Chem. Int. Ed. 2020, 59, 8873.</p><p>Atomically precise graphene nanoribbons (GNRs) have attracted intense interest due to their fascinating electronic and magnetic properties, which enable next‐generation materials for carbon‐based nanoelectronics and spintronics.1 Bottom‐up synthetic strategies, including solution synthesis, chemical vapor deposition (CVD), and surface‐assisted synthesis, represent the most powerful approaches for fabricating GNRs with uniform widths and defined edge structures.2 Compared with the solution‐synthesis and CVD‐growth methods, the on‐surface‐synthesis approach takes advantage of in situ characterization of the prepared GNRs and produces high‐quality GNRs that can serve as ideal objects for fundamental studies of graphene‐based electronic devices.3 In recent years, a series of GNRs with different edge topologies and widths have been achieved through surface‐assisted synthesis. These edge topologies include armchair, chevron, cove, and zigzag.1a, 4 Compared to armchair‐edged or cove‐edged GNRs (AGNRs or CGNRs), zigzag‐edged GNRs (ZGNRs) are predicted to preserve gapless (zero band gap) or metallic band structures5 as well as host spin‐polarized electronic edge states, which render them promising materials in spintronics.4b, 6 Nevertheless, these fascinating properties have barely been observed experimentally due to the lack of a facile synthetic approach and the poor stability of the ZGNRs.7 In 2016, the first atomically precise 6‐ZGNR (Figure 1 a) was synthesized on an Au(111) substrate under ultrahigh vacuum (UHV) conditions based on a U‐shaped dibromo dibenzo[a,j]anthracene monomer (DBBT in Figure 1 a) with a preinstalled zigzag edge and two additional methyl groups at the periphery, which led to the formation of fully zigzag‐edged GNRs by oxidative ring closure of the methyl groups. Scanning tunneling spectroscopy (STS) measurements demonstrated the existence of localized edge states with a large energy splitting at the zigzag edges.4b On the other hand, partial ZGNRs have also raised substantial interest since they can not only enhance the stability but also tune their electronic and magnetic properties, and among them, the most prominent examples exhibit quasi‐one‐dimensional trivial and nontrivial electronic quantum phases.3c, 8 Moreover, heteroatom doping, such as the incorporation of boron and oxygen/nitrogen atoms, at the edges or the central sp2‐carbon frameworks of GNRs has been recently demonstrated to modulate the conductance and valence bands of the ribbons and provide access to stable ZGNRs.9</p><p>a) Structures of dibromo‐dimethyl‐biphenylbenzo[m]tetraphene (DBBT) and 6‐ZGNR.4b b) Structures of pristine fully‐carbon‐based ZGNR1 (PCZGNR1) and NBN‐ZGNR1. c) Structures of PCZGNR2 and NBN‐ZGNR2. d) NBN‐dibenzophenalene (NBN‐DBP) and its radical cation as well as its isoelectronic structures.10 e) On‐surface synthetic routes toward NBN‐ZGNR1 and NBN‐ZGNR2.</p><p>Among all kinds of heteroatom doping strategies, boron–nitrogen (B–N) doping is particularly interesting due to the isoelectronic and isosteric relationship between C=C and B−N units.10 Therefore, substituting a C=C unit in the π‐conjugated system with a B−N subunit can remarkably modulate the electronic structure while maintaining the same conjugated skeleton.10, 11 In addition to single B−N‐unit doping, substituting a full C3 unit on the zigzag edge with a nitrogen–boron–nitrogen (NBN) motif not only provides access to stable zigzag‐edged nanographene (namely, polycyclic aromatic hydrocarbon) but also renders the formation of the radical cation on the NBN edge through selective oxidation, which is the isoelectronic structure of its pristine carbon framework with an open‐shell character (Figure 1 d).12 However, the synthesis of NBN‐doped zigzag‐edged GNRs remains challenging due to the lack of a suitable synthetic strategy.</p><p>Herein, we report the bottom‐up synthesis of the first NBN‐doped ZGNRs (Figure 1 b,c, NBN‐ZGNR1 and NBN‐ZGNR2) by employing two novel U‐shaped bis(para‐iodophenyl)‐substituted NBN‐dibenzophenalene (NBN‐DBP) monomers (M1 and M2, Figure 1 e). The monomers M1 and M2 (with an additional phenyl group at ring A in M1, Figure 1 e) feature the preinstalled zigzag‐edge topologies with two iodo groups at the para position of ring D (Figure 1 e), which enables the surface‐assisted polymerization to form swallow‐shaped polymers through thermally induced aryl–aryl coupling (Figure 1 e). Subsequently, intramolecular cyclodehydrogenation of the polymers enables the successful formation of NBN‐doped GNRs with zigzag‐rich edges in which the zigzag‐edge proportions of NBN‐ZGNR1 and NBN‐ZGNR2 are 36 % and 57 %, respectively (for a detailed assignment of the zigzag edges, see Figure S9 in the Supporting Information).3c High‐resolution scanning tunneling microscopy (STM) in combination with noncontact atomic force microscopy (nc‐AFM) clearly reveals the zigzag‐edge topologies of the resultant GNRs. The electronic band gaps of NBN‐ZGNR1 and NBN‐ZGNR2 are determined by scanning tunneling spectroscopy (STS) to be 1.50 eV and 0.90 eV, respectively, which are substantially higher than those of the corresponding pristine carbon‐based ZGNRs (PCZGNR1: 0.52 eV; PCZGNR2: 0.27 eV. Figure 1 b,c). Notably, DFT calculations predict that the electronic structures of the NBN‐doped ZGNRs can be further modulated into gapless or metallic through the one‐electron oxidation of each NBN unit into the radical‐cation form.</p><p>The synthetic routes toward the well‐designed U‐shaped monomers M1 and M2 are illustrated in Scheme 1. First, Suzuki coupling was performed between biphenyl diboronic acid pinacol ester 1 and 1‐bromo‐2‐iodo‐3‐nitrobenzene, which provided dibromo‐dinitrophenyl‐biphenyl 4 in 65 % yield. Then, 4 was subjected to Suzuki coupling with (4‐(trimethylsilyl)phenyl)boronic acid to produce dinitrophenyl‐bis(trimethylsilane)‐biphenyl 6 in 71 % yield. Afterwards, the trimethylsilyl (TMS) groups in 6 were converted into iodo groups by treatment with excess iodine monochloride (ICl) to afford diiodophenyl‐dinitrophenyl‐biphenyl 8 in 90 % yield. Subsequently, compound 8 was reduced to diiodophenyl‐diamine‐biphenyl 10 at room temperature under hydrogen gas with Pt/C in 100 % (crude) yield. Finally, heating a solution of 10 in o‐dichlorobenzene (o‐DCB) at 180 °C in the presence of boron trichloride (BCl3) with excess trimethylamine (NEt3) gave M1 in 32 % yield. Following a similar synthetic strategy, monomer M2 with an additional phenyl ring at the para position of ring A in M1 was successfully synthesized starting from dibromo‐terphenyl 2 over six steps.</p><p>Synthetic routes toward M1 and M2.</p><p>The targeted U‐shaped monomers M1 and M2 were purified by silica‐column chromatography and then recrystallized in chloroform/methanol (CHCl3/MeOH). Afterwards, M1 and M2 were characterized by high‐resolution matrix‐assisted laser desorption/ionization time‐of‐flight (HR‐MALDI‐TOF) mass spectrometry (Figure 2 a,c). There is only one dominant peak in the respective mass spectrum of M1 and M2, revealing their defined molecular compositions; the isotopic distribution pattern of the mass peak is in good agreement with the calculated pattern (Figure 2 a,c insert). Furthermore, 1H NMR spectra of M1 and M2 displayed well‐resolved peaks that could be fully assigned by 2D NMR analysis (Figures 2 b,d and S38–S51). Notably, the singlet peak (orange highlighted peak in Figure 2 b,d) in the 1H NMR spectra was assigned to the proton that connects with the nitrogen atom in M1 and M2. Additionally, there is one broad resonance at approximately 27.2 ppm in the 11B NMR spectrum of M1 (Figure S41).13, 14</p><p>a) HR MALDI‐TOF mass spectrum of M1. b) 1H NMR spectrum (300 MHz, 273 K, C2D2Cl4) of M1, insert: assignment for each proton. c) HR MALDI‐TOF mass spectrum of M2. d) 1H NMR spectrum (300 MHz, 373 K, C2D2Cl4) of M2, insert: assignment for each proton.</p><p>To obtain NBN‐ZGNR1, monomer M1 was sublimed onto an Au(111) substrate held at room temperature under UHV conditions. Constant‐current STM images of the resulting molecular layer revealed that the intact monomers self‐assembled into linear chains (Figure 3 a). Additionally, the assembled structures formed a tail‐to‐tail pattern caused by I⋅⋅⋅H interactions between two monomers (Figure S1). Annealing the substrate at 200 °C induced aryl–aryl coupling, which resulted in linear swallow‐shaped polymer poly‐ 1 through Ullmann‐type coupling of M1 (Figure 1 e). The polymeric chains adsorbed in the fcc region of the Au(111) reconstruction (Figures 3 b and S2). A closer view of poly‐ 1 indicated bright protrusions along the chain, which resulted from the rotation of the benzene rings (benzene rings D and A in Figure 1 e) due to the steric hindrance between the neighboring rings (Figures 3 b and S2). Complete cyclodehydrogenation was achieved by further treatment of poly‐ 1 at 450 °C, which provided fully planarized NBN‐ZGNR1 with an apparent height of 1.75 Å (Figure 3 c). The longest length of the GNRs was up to 30 nm (Figure S6). Additionally, there were some defects caused by the missing benzene ring A in NBN‐ZGNR1 (Figures S3 and S7). These can be attributed to the steric effect between the rotatable middle benzene ring (A) and the side ring (D, Figure 1 e). Therefore, ring A may detach from the polymer at high temperatures before cyclodehydrogenation. Furthermore, the chemical structure of the resultant NBN‐ZGNR1 was unambiguously confirmed by nc‐AFM measurements using a CO‐functionalized tip. Figure 3 d depicts the resulting constant‐height frequency‐shift image where the periodic aromatic carbon atoms together with the nitrogen and boron atoms are clearly unveiled. In the previously reported B‐AGNRs, B,N‐AGNRs, and oxygen–boron–oxygen (OBO) chiral GNRs,9a, 9c, 9f the boron atoms appeared with a darker contrast (more negative frequency shift) due to a stronger interaction with the gold substrate. These contrasting results can be explained by the same interaction of the nitrogen, boron, and carbon atoms with the gold substrate. Notably, NBN‐ZGNR1 was found to laterally align on the Au(111) substrate (Figure 3 c).</p><p>a) STM image of M1 as sublimed on Au(111). b) STM image of M1 after annealing at 200 °C on Au(111), inducing deiodination and polymerization. c) STM image of M1 after annealing at 450 °C on Au(111). d) nc‐AFM image of NBN‐ZGNR1. e) STM image of M2 after annealing at 450 °C on Au(111), showing the formation of NBN‐ZGNR2. f) nc‐AFM image of NBN‐ZGNR2. Scanning parameters: (a)–(c) and (e) V=−500 mV and I=30 pA. (d,f) amplitude=100 pm.</p><p>Following a similar synthetic procedure, NBN‐ZGNR2 was successfully synthesized from monomer M2 on the surface. M2 was sublimed onto a Au(111) substrate and held at room temperature under UHV conditions, where it formed self‐assembled linear chains (Figure S4). Then, linear swallow‐shaped polymer poly‐ 2 was formed after annealing the substrate at 270 °C through Ullmann‐type coupling (see Figure S4). Finally, NBN‐ZGNR2 was achieved by annealing poly‐ 2 at 450 °C, which exhibited an apparent height of 1.75 Å (Figure 3 e), and the longest length was up to 12.3 nm (Figure S5). Due to the increased steric hindrance between the central benzene rings (B and C) and the side rings (D, Figure 1 e) in the precursor M2 and poly‐ 2, the resultant length of NBN‐ZGNR2 is shorter than that of NBN‐ZGNR1. From a statistical analysis (Figure S6), the length of most NBN‐ZGNR2 (or NBN‐ZGNR2 segments) is less than 7 nm (nine repeating units). Similar to NBN‐ZGNR1, the nitrogen and boron atoms were clearly unveiled in NBN‐ZGNR2 by nc‐AFM measurements (Figure 3 f). Due to the NBN edge doping, these two GNRs featured stable zigzag edges in which no further reaction such as edge fusing was observed after thermal annealing, in contrast to the pristine carbon‐based ZGNRs.3c</p><p>To gain insight into the electronic properties of NBN‐ZGNRs, differential conductance (dI/dV) spectra (Figure 4) based on these two GNRs were probed on Au(111). The conduction band (CB) and valence band (VB) could be affected by the electronic states of the substrate and only appear like onsets in the STS measurement.15 The CB and VB of NBN‐ZGNR1 were identified as the onset of the electronic bands, which are located at 0.50 eV and −1.00 eV (Figure 4 a), respectively. Its corresponding electronic band gap is derived to be 1.50 eV. From the DFT calculations (Figure 4 b), the band gap of NBN‐ZGNR1 is estimated to be 1.53 eV, which is in line with the experimental STS result. Additionally, the dI/dV spectra in Figure 4 c reveal the CB and VB of the NBN‐ZGNR2 segment (six repeating units) to be located at 0.20 eV and −0.70 eV, respectively. Accordingly, the corresponding electronic band gap is 0.90 eV, which is further supported by DFT calculations (Figure 4 d, 0.84 eV). The band gap of NBN‐ZGNR2 (0.90 eV) is clearly smaller than that of NBN‐ZGNR1 (1.50 eV), which can be attributed to the laterally expanded width of the GNR with a large π‐conjugated system. In contrast to the pure‐carbon‐based full ZGNRs, which have a zero band gap,5 the resultant NBN‐doped ZGNRs possess a defined band gap, which is comparable to those of 13‐AGNR (1.40 eV)3b and 15‐AGNR (0.86 eV).16, 17 On the other hand, DFT calculations predict that NBN‐ZGNR1 (1.53 eV) and NBN‐ZGNR2 (0.84 eV) exhibit larger band gaps than their corresponding pristine‐carbon‐based ZGNRs (PCZGNR1: 0.52 eV, Figure 5 a; PCZGNR2: 0.27 eV, Figure 5 b; see the structures in Figure 1 b,c).</p><p>STS and calculated band structures of NBN‐ZGNR1 and NBN‐ZGNR2. a) Differential conductance (dI/dV) spectra taken on NBN‐ZGNR1, CB: conduction band, VB: valence band. b) DFT‐calculated band structure and density of states (DOS) results of NBN‐ZGNR1. c) Differential conductance (dI/dV) spectra taken on NBN‐ZGNR2. d) DFT‐calculated band structure and DOS results of NBN‐ZGNR2.</p><p>a), b) DFT‐calculated band structures of pristine PCZGNR1 and PCZGNR2. The bands of different spins are degenerated. c), d) Chemical structures of NBN‐ZGNR1 and NBN‐ZGNR2 radical cations. e), f) DFT‐calculated band structures of NBN‐ZGNR1 radical cations and NBN‐ZGNR2 radical cations in which each NBN unit loses one electron. The bands of different spins are degenerated.</p><p>As the NBN motif on the zigzag edge can be further oxidized into the corresponding radical cation, which is the isoelectronic structure to its pristine carbon framework (Figure 1 d), this enables a potential chemical tunability of NBN‐ZGNRs. To this end, DFT calculations were carried out to study the electronic structures of NBN‐ZGNRs after one‐electron oxidation for each NBN unit (Figure 5). Interestingly, when NBN‐ZGNR1 and NBN‐ZGNR2 are oxidized into radical cations (Figure 5 c,d, each NBN unit loses one electron), their band structures become gapless (0 eV) and metallic, respectively (Figure 5 e,f). Compared to the pristine carbon‐based PCZGNR1 (0.52 eV) and PCZGNR2 (0.27 eV) (Figure 5 a,b), which are low‐band‐gap semiconductors, NBN‐ZGNRs with radical cations clearly show different electronic structures.</p><p>In summary, we demonstrated the bottom‐up on‐surface synthesis of the first NBN‐doped ZGNRs, which are derived from two novel U‐shaped molecular precursors with preinstalled NBN zigzag edges. The geometric structures of the zigzag topologies of NBN‐doped ZGNRs have been unambiguously characterized by STM and nc‐AFM. STS analysis together with DFT calculations revealed that the NBN units play a significant role in modulating the electronic structures of ZGNRs (NBN‐ZGNR1: 1.50 eV; NBN‐ZGNR2: 0.90 eV) compared with those of their corresponding pristine carbon‐based ZGNRs (PCZGNR1: 0.52 eV; PCZGNR2: 0.27 eV). Moreover, theoretical calculations predicted that the band structures of NBN‐ZGNRs can be further tailored to be gapless or metallic through the selective oxidation of the NBN units into the formation of radical cations. The synthetic strategy established in this work provides an opportunity to fabricate stable GNRs containing zigzag edges and lends credence to their possible applications in graphene‐based nanoelectronic devices. Moreover, the chemical tunability of NBN‐ZGNRs paves the way for investigating the isoelectronic structures of pure‐carbon‐based ZGNRs.</p><!><p>The authors declare no conflict of interest.</p><!><p>As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.</p><p>Supplementary</p><p>Click here for additional data file.</p>

PubMed Open Access

N-terminal phosphorylation of xHes1 controls inhibition of primary neurogenesis in Xenopus

The processes of cell proliferation and differentiation are intimately linked during embryogenesis, and the superfamily of (basic) Helix-Loop-Helix (bHLH) transcription factors play critical roles in these events. For example, neuronal differentiation is promoted by class II bHLH proneural proteins such as Ngn2 and Ascl1, while class VI Hes proteins act to restrain differentiation and promote progenitor maintenance. We have previously described multi-site phosphorylation as a key regulator of tissue specific class II bHLH proteins in all three embryonic germ layers, and this enables coordination of differentiation with the cell cycle. Hes1 homologues also show analogous conserved proline directed kinase sites. Here we have used formation of Xenopus primary neurons to investigate the effects of xHes1 multi-site phosphorylation on both endogenous and ectopic proneural protein-induced neurogenesis. We find that xHes1 is phosphorylated in vivo, and preventing phosphorylation on three conserved SP/TP sites in the N terminus of the protein enhances xHes1 protein stability and repressor activity. Mechanistically, compared to wild-type xHes1, phospho-mutant xHes1 exhibits greater repression of Ngn2 transcription as well as producing a greater reduction in Ngn2 protein stability and chromatin binding. We propose that cell cycle dependent phosphorylation of class VI Hes proteins may act alongside similar regulation of class II bHLH proneural proteins to co-ordinate their activity.

n-terminal_phosphorylation_of_xhes1_controls_inhibition_of_primary_neurogenesis_in_xenopus

<!>Introduction<!>Cloning<!>Xenopus laevis embryo manipulation<!>In situ hybridisation (ISH)<!>Quantitative real-time PCR<!>Western blotting<!>Statistical analysis<!>xHes1 is phosphorylated in vitro and in vivo, regulating its ability to inhibit endogenous primary neurogenesis<!><!>xHes1 is phosphorylated in vitro and in vivo, regulating its ability to inhibit endogenous primary neurogenesis<!><!>xHes1 is phosphorylated in vitro and in vivo, regulating its ability to inhibit endogenous primary neurogenesis<!>xHes1 phospho-status influences inhibition of primary neurogenesis induced by proneural bHLH proteins<!><!>xHes1 phospho-status influences inhibition of primary neurogenesis induced by proneural bHLH proteins<!><!>xHes1 phosphorylation status influences Ngn2 protein stability<!>Discussion<!>Transparency document<!>

<p>xHes1 is phosphorylated in Xenopus embryos on conserved N terminal SP/TP sites.</p><p>In vitro, xHes1 protein can be phosphorylated by Cyclin-dependent-kinases.</p><p>Under-phosphorylated xHes1 has increased protein stability relative to WT xHes1.</p><p>Under-phosphorylated xHes1 has enhanced inhibitory activity during neurogenesis.</p><p>xHes1 reduces Ngn2 expression, protein stability and chromatin binding.</p><!><p>Throughout embryogenesis, cell proliferation, fate choice and differentiation must be closely coordinated in time and space; in many tissues, the evolutionarily conserved superfamily of helix-loop-helix (HLH) transcription factors play central roles in all three of these processes [1]. This superfamily is defined by a conserved HLH motif that mediates protein dimerisation, and additionally, most members contain a DNA binding basic domain (bHLH proteins). However, outside of this region, significant diversity of sequence, structure and function allows identification of distinct sub-groups within the superfamily with differing and sometimes opposing regulatory roles [2].</p><p>For example, early studies on Drosophila neurogenesis established a model for selection of neuronal precursors from otherwise equivalent neuroectodermal cells, based on antagonistic interaction between tissue specific class II activating bHLH proneural proteins such as achaete-scute complex and atonal, and the inhibitory bHLH-Orange class VI proteins of the hairy-and-enhancer-of-split family [3]. Within the mammalian nervous system, the class VI protein Hes1 is essential for maintenance and proliferation of neural progenitor cells, ensuring temporal control of differentiation competency, whilst also being required for boundary formation, structural integrity and playing later roles in gliogenesis and neuronal protection; comprehensively reviewed in Ref. [4]. Given its pleiotropic roles, it is not surprising that Hes1 is tightly regulated at transcriptional, epigenetic and post-translational levels. In particular, as with regulation described for other bHLH proteins, phosphorylation at individual sites of Hes1 regulates some of its context-dependent effects; for example, JNK1-mediated phosphorylation of S262 in the Hes1 C terminus influences synaptic plasticity in rat cortex [5].</p><p>Within the last decade, phosphorylation of tissue-specific class II bHLH proteins has emerged as an important way of restraining differentiation in the context of high cell cycle activity [[6], [7], [8]]. Typically, these multiple phosphorylation events occur on Serine-Proline (SP) or Threonine-Proline (TP) sites in the N and C termini of bHLH proteins, and confer regulation in all three germ layers: multi-site phospho-regulation has been demonstrated for Ngn2, Ascl1, and NeuroD4 in neuroectoderm [6,7,9], MyoD in mesoderm [10] and Neurogenin3 in endoderm [8]. Vertebrate Hes1 homologues of class VI also show a conservation of multiple SP/TP sites across species. Although class VI bHLH proteins are traditionally viewed as transcriptional repressors rather than activators, this raises the intriguing possibility that class VI bHLH proteins may undergo similar multi-site phospho-regulation.</p><p>Xenopus embryos provide a rapid and accessible in vivo model of vertebrate development to study the activity of bHLH transcription factors in differentiation of multiple tissues, for example [[6], [7], [8], [9], [10]]. In particular, the generation of Xenopus primary neurons from the neural plate is an established system for probing the interaction between class II proneural bHLH proteins that promote neuronal differentiation, and class VI Hes proteins that inhibit it, for example [11,12].</p><p>Here, we use Xenopus primary neurogenesis to investigate a role for multi-site phosphorylation in regulation of Xenopus Hes1 activity (known as xHes1 or xHairy1). We find that preventing phosphorylation on N-terminal SP/TP sites in xHes1 enhances the ability of xHes1 to inhibit primary neurogenesis driven by three different class II proneural bHLH transcription factors. Mechanistically, we see that preventing phosphorylation of these sites increases stability of xHes1 protein, reduces Ngn2 transcript expression and also leads to greater destabilisation of Ngn2 protein. Furthermore, xHes1 protein is phosphorylated in neural plate stage embryos, and in vitro kinase assay sensitivity indicates a potential role for cell-cycle phase dependent regulation.</p><!><p>Wild-type (WT) Xenopus xHes1 and Xenopus NeuroD1 were cloned into pCS2 with a single C terminal HA tag. Xenopus Ngn2 and mouse Ascl1 have been described previously [6,7]. 5T/S-A xHes1 and 3T/S-A xHes1 were generated by QuikChangeII Site-Directed Mutagenesis Kit (Agilent Technologies). All primers available on request. Nucleotide and protein sequence alignments were conducted using ClustalW [13].</p><!><p>All work has been carried out under UK Home Office Licence and in accordance with the UK Animals (Scientific Procedures) Act, 1986 and associated guidelines. A description of experiments using ARRIVE guidelines is provided in Ref. [14]. Acquisition of X.laevis embryos, preparation and injection of synthetic mRNA, and staging of embryos were conducted as described [9,14].</p><!><p>ISH was performed using dig-oxigenin-labelled anti-sense probes. Semi-quantitative scoring was conducted for gene expression on the injected side of the embryo relative to the uninjected side; grades were assigned: −3, no expression; −2, marked reduction in expression; −1, mild reduction in expression; 0, no change in expression; +1, increased expression within the neural tube only; +2, additional ectopic expression restricted to the dorsal ectoderm; +3, moderate but patchy ectopic expression spreading over the lateral ectoderm; +4, extensive ectopic expression over the lateral ectoderm in a homogenous pattern.</p><!><p>Whole embryo RNA was extracted, cDNAs prepared and qPCR conducted as described [9,10].</p><!><p>In vitro kinase assay was conducted as described [8]. Protein extraction, lambda protein phosphatase treatment and assessment of cytoplasmic and chromatin-bound proteins was conducted as described [9]. All bHLH proteins were detected by a single HA tag and antibodies used according to Ref. [9].</p><!><p>For western blotting, experiments were performed in independent duplicate or triplicate with representative results shown; protein quantification was conducted using ImageJ as described [9]. For qPCR data, mRNA expression was normalised to housekeeping gene EF1α, and for analysis, mRNA levels in injected categories were calculated relative to stage-matched uninjected controls. Mean values and the standard error of the mean (s.e.m.) were calculated from N independent experiments. Statistical significance was determined using a paired two-tailed Student's t-test with not significant = NS; (p < 0.05) = *; (p < 0.025) = **; (p < 0.0125) = ***. For ISH data, experiments were conducted in independent duplicate or triplicate and the N numbers refer to the range of total numbers of embryos in each injection category.</p><!><p>Supplementary Fig. 1 shows xHes1 is expressed in early neural plate stage embryos at the anterior and lateral borders of the neural plate [blue arrows] and in future neural crest [white arrows], consistent with a previous report [15]. In contrast, Ngn2, a master regulator of primary neurogenesis, is expressed from early neural plate stages in the three bilateral stripes of developing primary neurons [black arrows] and [16]. Thus, xHes1 and Ngn2 show complementary expression patterns at neural plate stage, and the spatial location of xHes1 may contribute to limiting the area of primary neurogenesis.</p><!><p>xHes1 has conserved SP/TP sites and is phosphorylated both in vivo and in vitro. (A) Protein sequence alignment for human, mouse and Xenopus Hes1. The bHLH domain is shown in green with orange domain in orange and WRPW domain in blue. SP/TP sites are highlighted in red. A consensus line is also shown below the alignment to indicate the degree of conservation at each position: Residues may be identical (*), strongly conserved (:) or weakly conserved (.). (B) Western blot analysis of stage 12 embryos over-expressing 500 pg WT xHes1 mRNA and incubated with or without lambda protein phosphatase. (C) In vitro kinase assay showing in vitro translated WT xHes1 protein after incubation with recombinant Cyclin/Cdks as labelled. (D) Schematic representation of WT xHes1 protein and phospho-mutant variants, showing approximate location of SP/TP sites that are mutated to AP in each. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)</p><!><p>To determine whether xHes1 is phosphorylated in vivo in early Xenopus development, we expressed Wild-Type (WT) xHes1 in embryos by mRNA injection. Western blotting of protein extracts from stage 12 embryos reveal that WT xHes1 migrates as more than one band (Fig. 1B; red arrow at uppermost band). These forms are reduced to a faster migrating band (black arrow) in the presence of lambda protein phosphatase, indicating that WT xHes1 is indeed phosphorylated in vivo on at least one site. To test whether xHes1, in common with multiple tissue-specific bHLH factors [[6], [7], [8]], can be phosphorylated by Cyclin-dependent kinases, we incubated in vitro translated WT xHes1 protein with a range of cell-cycle associated Cyclin/Cdk combinations, or with MAPK, another proline directed kinase (Fig. 1C). Significant retardation of migration after separation by SDS-PAGE indicates that WT xHes1 is strongly phosphorylated by CyclinB/Cdk1 on more than one site in vitro, and is also phosphorylated to a lesser extent by CyclinA/Cdk2, but not by CyclinD/Cdk4, CyclinE/Cdk2 or MAPK. This suggests that xHes1 may undergo cell-cycle dependent phosphorylation in the G2/M phase of the cell cycle.</p><!><p>xHes1 phospho-status regulates inhibition of endogenous primary neurogenesis. Embryos were injected with 50 pg xHes1 mRNA and assayed at stage 18 by ISH for expression of N-tubulin (A; [N = 30–44]) or xSox2 (B; [N = 22–41]), with representative images shown in (C); N-tubulin dorsal view, xSox2 rostral view. (D) Embryos were assayed by qPCR at stage 14 [N = 3]; significance relative to uninjected embryos (black lines).</p><!><p>It has previously been demonstrated that Hes1 can repress Ngn2 activity in progenitor neuroblasts [17]. To determine whether xHes1 can inhibit neuronal differentiation but expand the progenitor population in the neural plate of embryos, expression of early neural progenitor marker xSox2 was investigated (Fig. 2B + C). WT xHes1 injection results in no apparent change in the xSox2 progenitor population, while both 3T/S-A and 5T/S-A xHes1 expand the rostral xSox2 domain [white arrows] in 20–30% of embryos. Thus, preventing phosphorylation of xHes1 protein results in enhanced ability to restrain neuronal differentiation, and instead promoting a progenitor state.</p><!><p>Ngn2 drives primary neurogenesis in Xenopus [16], while Hes1 can repress transcription of Ngn2 in mammalian cells [17]. We next compared the effects of WT and phospho-mutant xHes1 on Ngn2 expression in early neural plate stage Xenopus embryos, as measured by qPCR (Fig. 2D). At this developmental stage, WT xHes1 has no effect on Ngn2 expression but 3T/S-A and 5T/S-A xHes1 both reduce expression of Ngn2 and N-tubulin by 30–40%, again indicating enhanced activity of the phospho-mutants relative to WT xHes1. Similar repressive activity of 3T/S-A and 5T/S-A xHes1 indicates that the most important phospho-regulatory sites may reside in the N-terminus of xHes1, as these are mutated in both proteins. Hence, we concentrated on defining the function of the minimally mutated 3T/S-A xHes1 in comparison to the WT protein.</p><!><p>xHes1 inhibits ectopic neurogenesis induced by three different proneural proteins. Embryos were co-injected with mRNA encoding either WT or 3T/S-A xHes1 in the presence of one of three different proneural proteins. ISH scores are shown for Ngn2 (A; [N = 78–84]) and Ascl1 (B; [N = 96–104]). qPCR data is shown for Ngn2 (C; [N = 4]), Ascl1 (D; [N = 3]) and NeuroD1 (E; [N = 4]). See methods for grading system.</p><!><p>Ngn2 induces ectopic N-tubulin expression outside the neural tube in 98% of embryos; in 17% this ectopic expression is restricted to the dorsal ectoderm (grade 2) but in 81% this occurs more widely across the lateral ectoderm (grade 3). WT xHes1 co-expression results in a small reduction in the extent of ectopic neurogenesis, while 3T/S-A xHes1 shows enhanced ability to limit ectopic neurogenesis to the neural tube and dorsal ectoderm.</p><p>Ascl1 is another class II bHLH transcription factor particularly associated with GABAergic and noradrenergic neurogenesis in Xenopus embryos [18,19], while NeuroD1 is a proneural protein acting downstream of Ngn2 in the primary neurogenesis cascade. All three factors are able to induce ectopic primary neurons when over-expressed [7,12]. As with Ngn2, the inhibitory activity of WT and 3T/S-A xHes1 was tested with co-injected Ascl1 or NeuroD1 mRNA (Fig. 3B, D, E). Relative to WT xHes1, preventing N terminal phosphorylation in 3T/S-A xHes1 enhances the inhibitory effects of the xHes1 protein and limits ectopic primary neurogenesis induced by all three proneural transcription factors. Additionally, as NeuroD1 is a direct transcriptional target of Ngn2 [20], the proneural activity of induced NeuroD1 may contribute to the phenotype seen on over-expression of Ngn2. The ability of xHes1 to restrict ectopic neurogenesis induced by both NeuroD1 and Ngn2 may therefore contribute to the inhibitory effects at multiple levels of the Ngn2-induced primary neuron cascade.</p><!><p>Phospho-mutant xHes1 has enhanced protein stability relative to WT and reduces both total Ngn2 and chromatin-bound Ngn2 protein. (A) Western blot analysis of stage 11 embryos over-expressing 500 pg xHes1 mRNA and with xHes1 protein density calculated relative to tubulin in (C) [N = 3]. (B) Western blot analysis of whole embryo extracts from embryos overexpressing 150 pg Ngn2 and 500 pg xHes1 mRNA, and cytoplasmic and chromatin fractions from cross-linked stage 13 embryos injected with 250 pg mRNA of Ngn2 and xHes1; relative protein quantification in (D) [N = 3]; statistics by paired students T-test as described in methods.</p><!><p>In addition to regulation at a gene expression level, many class II proneural transcription factors including Ngn2 have been demonstrated to have short half-lives that can be modulated by post-translational modification including phosphorylation [21,22]. Stability is also controlled by protein-protein interactions, including between class II proneural proteins and their heterodimeric class I E protein partners [21], and dimerisation is required for DNA binding [23]. As Hes1 can sequester E proteins from class II proneural transcription factors [24], we next investigated the effect of WT and 3T/S-A xHes1 on the stability and chromatin binding of Ngn2 in embryos.</p><p>After co-injecting mRNAs encoding WT or 3T/S-A xHes1 along with Ngn2, we determined protein expression levels in stage 11 whole embryo extracts (Fig. 4B; Ngn2, black arrows; xHes1 red arrow). A lower dose of xHes1 mRNA was required to improve embryo survival for chromatin analysis, where embryos were cross linked at stage 13 prior to extraction of cytoplasmic and chromatin fractions (Fig. 4B). When Ngn2 is quantified relative to tubulin in whole embryo extracts (Fig. 4D), Ngn2 protein is reduced to 87% of the control level in the presence of WT xHes1, but shows a more pronounced reduction to 64% of the control level when phospho-mutant xHes1 is co-injected. Additionally, when embryos are fractionated, Ngn2 protein in cytoplasmic extracts shows a corresponding reduction in the presence of WT and 3T/S-A xHes1 that is quantitatively similar to that described for the whole embryo extracts. Thus, in addition to a greater reduction in Ngn2 at transcript level (Fig. 2D), we also see that 3T/S-A xHes1 may be inhibiting Ngn2 via a reduction in Ngn2 protein stability. If xHes1 is destabilising Ngn2 by sequestering E protein partners, one might expect a greater effect on stability of chromatin-bound Ngn2 that is dependent on dimerisation [23], compared to protein in the cytoplasm. Indeed, chromatin-bound Ngn2 is reduced to 50% and 38% of control levels in the presence of WT and 3T/S-A xHes1 respectively; a greater reduction than the corresponding 83% and 61% decreases in cytoplasmic Ngn2 protein.</p><!><p>Using a Xenopus model of neurogenesis, here we report that Hes1 homologue xHes1 is phosphorylated in vivo on highly conserved SP/TP residues in the N terminus of the protein. Preventing modification of these sites increases xHes1 protein stability and enhances its inhibitory effects on both endogenous and ectopic proneural protein induced neurogenesis, indicating that class VI bHLH proteins can be regulated by SP/TP phosphorylation in an analogous way to that previously described for class II proneural bHLH factors [[6], [7], [8], [9], [10]].</p><p>Furthermore, we provide an insight in to the multiple levels of interaction between xHes1 and Ngn2 during primary neurogenesis. xHes1 reduces Ngn2 transcript expression and additionally has post translational effects to destabilise the Ngn2 protein and inhibit proneural activity of downstream factor NeuroD1. Mechanistically, Hes1 has previously been described as a transcriptional repressor with multiple modes of action. For example, direct repression of proneural gene expression via recruitment of the co-repressor Groucho/TLE to N-boxes in the promoter regions of proneural genes [25]; Hes1 binding of E protein partners, titrating them away from proneural bHLHs so reducing their activity [24]; competition between Hes1 homodimers and proneural/E protein heterodimers at certain targets [5]; DNA-independent protein tethering to recruit repressor proteins to target genes bound by proneural/E protein heterodimers [25]; or a mutual degradation mode in Drosophila with a tripartite complex of Hes/proneural/E protein components [26]. Preventing phosphorylation of xHes1 enhances its stability, and is consistent with increasing its repressive activity via all but this last mutual degradation mechanism. Using whole embryo bulk chromatin methods, we have been unable to detect chromatin-bound xHes1, possibly due to the highly unstable properties of the xHes1 protein. Nevertheless, ectopic 3T/S-A xHes1 is particularly effective at reducing both the total amount of Ngn2 protein and the chromatin-bound Ngn2 fraction (Fig. 4). This may support a prominent role for E protein sequestration in the mechanism of action of the phospho-mutant xHes1; an effect that would influence both the stability and chromatin binding of the Ngn2 protein [21,23].</p><p>How might Cdk-dependent phosphorylation of xHes1 and proneural proteins function in co-ordination to control neurogenesis? Interestingly, Ngn2 and Ascl1 are reported to be most sensitive to phosphorylation by Cdk2 [6,7], whereas here we find that xHes1 is highly sensitive to Cdk1, suggesting that cell-cycle phase-dependent phosphorylation may further fine-tune the relative activities of these two classes of proteins. Phosphorylation of xHes1 at G2/M phase and resultant protein destabilisation is also reminiscent of that described for MyoD to ensure release from chromosomes and allowing mitotic progression [27]. Similarly, cdc2 (Cdk1) mediated phosphorylation of Hes1 co-repressor Groucho occurs in G2/M to reduce chromatin association and repressor activity [28]. Thus, Cdk1-mediated phosphorylation of xHes1 may augment this relief of transcriptional repression during mitosis. Furthermore, in mammalian cells, the onset of differentiation is assumed to occur in the G1 phase but with the decision to differentiate may be in the preceding cell cycle [29]. As such, the phosphorylation of xHes1 in the late G2/M phase may relieve inhibition of class II proneurals to allow the transition to differentiation in the next G1.</p><p>It is clear that the interaction between class VI Hes proteins and class II proneural proteins is complex; Ascl1, Ngn2 and Hes1 display dynamic oscillatory antagonism in neuronal progenitors, while a transition to sustained expression (Ascl1/Ngn2) or repression (Hes1) accompanies the switch to neuronal differentiation [30]. It is likely that regulation by multi-site phosphorylation of both class II and class IV bHLH factors acts alongside other regulatory mechanisms (reviewed in Ref. [31]), contributing to this exquisite control and its co-ordination with the cell cycle.</p><!><p>coi_disclosurecoi_disclosure</p><!><p>:Supplementary Figure 1: Endogenous expression of xHes1 and Ngn2.Endogenous expression of Ngn2 and xHes1 in stage 13 and 15 embryos revealed by ISH; DV, dorso-ventral view.</p>

PubMed Open Access

A mucin-specific protease enables molecular and functional analysis of human cancerassociated mucins

Mucins are a class of highly O-glycosylated proteins that are ubiquitously expressed on cellular surfaces and are important for human health, especially in the context of carcinomas. However, the molecular mechanisms by which aberrant mucin structures lead to tumor progression and immune evasion have been slow to come to light, in part because methods for selective mucin degradation are lacking. Here we employ high resolution mass spectrometry, polymer synthesis, and computational peptide docking to demonstrate that a bacterial protease, called StcE, cleaves mucin domains by recognizing a discrete peptide-, glycan-, and secondary structurebased motif. We exploited StcE's unique properties to map glycosylation sites and structures of purified and recombinant human mucins by mass spectrometry. As well, we found that StcE will digest cancer-associated mucins from cultured cells and from ovarian cancer patient-derived ascites fluid. Finally, using StcE we discovered that Siglec-7, a glyco-immune checkpoint receptor, specifically binds sialomucins as biological ligands, whereas the related Siglec-9 receptor does not. Mucin-specific proteolysis, as exemplified by StcE, is therefore a powerful tool for the study of glycoprotein structure and function and for deorphanizing mucin-binding receptors. Significance StatementMucin-domain glycoproteins are a major biological reservoir of O-glycans. These towering structures are found on the surfaces of nearly every cell in the human body, and are important in the immune response to cancer. Due to a dearth of available tools to study mucins, the molecular mechanisms by which mucins perform their many functions remain unclear. Here, we overcome a major hurdle to the study of mucin-domain glycoproteins by characterizing a bacterial protease with specificity for mucins. This mucinase enables selective removal of native mucins from biological samples and cuts them into fragments amenable to analysis by mass spectrometry. Enzymatic de-mucination represents an important step in demystifying mucindomain glycoproteins.

a_mucin-specific_protease_enables_molecular_and_functional_analysis_of_human_cancerassociated_mucins

Introduction<!>StcE has peptide-, glycan-, and secondary structure-based specificity for mucins<!>Based on MS analysis of cleaved peptides (Fig 2b and Supplemental Table<!>StcE improves mass spectrometry analysis of mucin-domain glycoproteins<!>StcE cleaves mucins from biological samples<!>StcE treatment of cultured cells reveals that Siglec-7 binds mucin-domain glycoproteins<!>In vitro StcE activity assays<!>Mass spectrometry<!>Flow cytometry and Western blotting of StcE-treated cells<!>Molecular modeling<!>StcE treatment of patient-derived CA-125<!>Cell culture

<p>Mucins are a class of proteins whose closely-spaced serine-and threonine-bound glycans (Oglycans) enforce a rigid extended structure 1 . In addition to being a major component of mucus, mucins are found on cell surfaces on nearly every cell of the human body, where their towering structures act as physical barriers 2 , glycocalyx stiffening agents 3 , receptor ligands 4 , and mediators of intracellular signaling 5 . Aberrant mucin expression and glycosylation are reliable biomarkers of carcinomas in humans 6 . Indeed, the membrane-associated mucin MUC1 is aberrantly expressed in ~60% of all cancers diagnosed each year in the U.S. 5 , rendering MUC1 one of the most prominently dysregulated genes in cancer. Another mucin, MUC16 (also called CA125), is highly expressed in ovarian cancer and clinically used as biomarker for treatment efficacy and surveillance 7 . Recent work highlighted the functional role of not only the C-terminal mucin signaling domain, but also the heavily glycosylated mucin ectodomain in promoting tumor progression. For example, the MUC1 ectodomain alone can drive tumor progression by enhancing cancer cell survival and promoting proliferation in the metastatic niche 8 . Observations such as these have motivated numerous efforts towards mucin-based vaccines 9 , small molecule 10 and antibody 11 therapies, and, more recently, CAR-T cell therapies 12 . Despite all this attention, we know little about the molecular structures of mucin glycoproteins, i.e., which sites are glycosylated, and with what glycan sequences? Our present knowledge is limited to gross characteristics such as the percent of mass represented by total glycans, or the list of glycans that can be chemically released from a sample mucin. Even from such crude data, intriguing relationships have been identified between glycan composition and cancer progression. For example, many cancer types undergo a global upregulation of mucinassociated sialic acid, and/or present with incomplete mucin glycosylation that results in truncated structures, such as Tn (GalNAcaSer/Thr), sialyl Tn (NeuAca2,6GalNAc), T (Galb1,3GalNAc), and sialyl T (NeuAca2,3Galb1,3GalNAc) 13 . How these altered mucin glycoforms relate to disease progression is an open question of considerable current interest. In order for headway to be made, a holistic view of glycoprotein structures at the molecular level is needed. This is a major goal of glycoproteomics, a field that is rapidly advancing with improved instrumentation in mass spectrometry (MS) 14,15 . However, the field has been almost entirely focused on N-glycosylated proteins, which have predictable glycosylation sites and structures, convenient enzymatic tools for glycan manipulation, and effective software for site and structure assignments. Mucin glycoproteins, on the other hand, defy all these conveniences. An Oglycosylation consensus motif has yet to be found, O-glycans are not predictable in structure, and no enzymatic methods exist for quantitative release of O-glycans in their native form. In addition, due to the presence of tandem repeat domains, MUC1 can be > 1200 amino acids long and 50% glycosylation by mass 16 ; MUC16 can exceed 22,000 residues and 85% glycosylation by mass 17 . The high density of O-glycosylation on these tandem repeats makes them resistant to digestion by workhorse proteases such as trypsin, meaning the majority of the sequence space is often left unanalyzed 18,19 . Systems with truncated forms of glycosylation, such as engineered "SimpleCells" lacking the O-glycan elaboration machinery, can simplify the identification of glycosylation sites. But functionally important glycan structures beyond the initiating O-GalNAc are lost 20 . Recently, an "O-protease" was reported that cleaves N-terminally to glycosylated serine and threonine residues 21 . However, this enzyme requires substrate desialylation for optimal activity, and is not specific for mucins.</p><p>The ubiquitous and highly conserved nature of mucin structures implies an ecological need for enzymes with mucin-specific proteolytic activity. Indeed, recent evidence has pointed to the existence of families of mucin-targeting proteases, comprising hundreds of enzymes largely found in organisms living in mucin-rich host environments [22][23][24] . One such enzyme, secreted protease of C1 esterase inhibitor (StcE), is a zinc metalloprotease of human pathogenic enterohemorrhagic E. coli (EHEC). First discovered by Welch et al. in 2002 25 , StcE promotes EHEC pathogenesis in humans by cleaving the protective mucus layers of the gut. It is reported to cleave densely O-glycosylated proteins, but not N-glycosylated or sparsely O-glycosylated substrates 23 . Given these reported pathogenic properties, we speculated that StcE could be transformed into a research tool to efficiently and specifically cleave human mucins.</p><p>Here, we report that StcE has peptide-, glycan-, and secondary structure-based specificity for mucins. This "mucinase" improves sequence coverage, glycosite mapping, and glycoform analysis of purified mucins by mass spectrometry, and enables specific release of mucins from cell lines and human tissue samples. We used StcE to provide evidence for the existence of professional mucin ligands of Siglec-7, an immune checkpoint receptor. We envision mucinspecific proteolysis will be a valuable addition to the biochemist's toolbox, enabling mucindomain containing proteins to be selectively liberated from biological samples and cut into fragments amenable to analysis (Fig. 1a).</p><!><p>We expressed StcE and its catalytically inactive point mutant (E447D) as 98 kDa soluble Nterminal His-tagged proteins in E. coli, as previously described (SI Appendix, Fig. S1) 26 . Previous reports suggested that StcE activity against a known substrate, C1 esterase inhibitor (C1INH), is detectable in a pH range of 6.1-9.0, in a temperature range of 4-55 °C, in high salt and detergent, and after days of incubation at 37 °C, consistent with its pathological activity in the mammalian gut 27 . In our hands, StcE was amenable to high yield expression (80 mg/L), active against C1INH (SI Appendix, Fig. S2), stable to lyophilization (SI Appendix, Fig. S3), and operative at nanomolar concentrations in all media types tested. We next assessed StcE's activity on clinically relevant mucin-domain glycoproteins. StcE did not cleave glycosylated but non-mucin proteins (bovine serum albumin and fetuin), but cleaved all tested mucin-like glycoproteins (recombinant MUC16, podocalyxin, CD43, PSGL-1, Syncam-1, and CD45), as evidenced by gel shifts to lower molecular weights (glycostain and silver stain, Fig. 1b and SI Appendix, Fig. S4, respectively). Further, StcE's activity was abrogated when its substrates were enzymatically deglycosylated, indicating a glycan requirement for cleavage (Fig. 1c).</p><p>We next asked whether StcE has a preferred sequence or structure recognition motif. The recombinant mucin-domain glycoproteins listed above were digested with StcE, de-Nglycosylated with PNGaseF, trypsinized, and subjected to MS analysis using an optimized protocol (SI Appendix, Fig. S5). Through manual validation of peptides present in the StcE samples but not in the control samples (PNGaseF and trypsin only), we discovered that StcE has a distinct peptide consensus sequence, S/T*-X-S/T, where cleavage occurred before the second serine or threonine and X was any amino acid or, to a lesser extent, absent (Fig. 2a and SI Appendix, Fig. S6). As seen from a table containing N-terminal StcE-cleaved peptides, the P2 (*) position was invariably glycosylated (Fig. 2b; see Supplemental Table 1 for all sequenced peptides). This glycosylation ranged from a single O-GalNAc residue to higher order structures such as a di-sialylated T antigen, indicating that StcE accepts a variety of glycans at the P2 position. Note that StcE cleavage was also permissive to glycosylation at the P1' position. In all cases, neither the peptide sequence nor the glycan alone was sufficient to predict cleavage. 1), the minimum necessary glycoform for StcE cleavage was a single GalNAc residue. To confirm this, we incubated a synthetic glycosylated polypeptide comprising GalNAc-a-O-Ser residues mixed with Lys residues in a random sequence 28 together with StcE. As seen in Fig. 2c, StcE cleaved the glycosylated polymer, and this cleavage was mitigated when the polymer was deglycosylated. We also found that StcE cleaved a synthetic peptide containing a single GalNAc, RPPIT*QSSL, into RPPIT*Q (Fig. 2d). These data confirm that O-GalNAc is the minimum required glycoform on the P2 position serine or threonine residue. GalNAc is the first glycan found on every site of mucin-type O-glycosylation and S/T-X-S/T is commonly found in their characteristic proline, threonine, and serine -rich repeat domains. Therefore, StcE is a true mucinase, a protease that specifically cleaves mucins, but is promiscuous within that family.</p><!><p>The observed insensitivity of O-glycosylated but non-mucinous proteins to StcE activity suggested that secondary structure may be an additional recognition determinant. For example, fetuin was not cleaved by StcE (Fig. 1b), though it exhibits a correctly glycosylated StcE consensus sequence GPT*PSAA (* = sialyl T antigen, among others) 29 . To explore the role of secondary structure in the determination of StcE's specificity, we conducted peptide docking studies with a previously reported crystal structure of StcE 26 and model glycopeptides derived from a StcE-labile podocalyxin sequence Ac-P(GalNAcα-)TL(GalNAcα-)TH-NMe (Fig. 2e and SI Appendix, Fig. S7). When docked using a scaffold consistent with previously reported zinc metalloprotease/peptide co-crystal structures 30 , the ligand made specific contacts with the zinc ion and other residues of the catalytic core as well as residues of a flanking β-strand, forming a combined antiparallel β-sheet. The acetyl groups of the GalNAc moieties frequently formed intramolecular contacts with peptide backbone amides. Previous studies have shown that similar carbohydrate-peptide interactions force O-α-GalNAc glycopeptides into a β-strand-like "mucin fold" 31 . The interactions within the modeled StcE/substrate complex therefore support the view that StcE may partially achieve its selectivity by recognizing a mucin-fold. Importantly, it appears that in this conformation, the glycan moieties of docked glycopeptides are oriented away from the enzyme active site (Fig. 2e), which may enable StcE to cleave glycopeptides with larger glycans.</p><!><p>Given its specificity for mucin domains, we predicted that StcE could be incorporated in common proteomic workflows to facilitate analysis of mucin glycoproteins. Recombinant substrates from Fig. 1b were digested with StcE, treated with PNGaseF to remove N-glycans, trypsinized, and subjected to MS. As seen in Fig. 3a, StcE treatment increased protein sequence coverage by up to 50%, number of glycosites by up to 6-fold, and number of localized glycans by up to 11-fold, with averages of 20%, 3.5-fold, and 4-fold improvement, respectively. The observed gains were likely due to StcE's ability to break up areas of dense O-glycosylation, which generated smaller glycopeptides with higher charge density. This allowed for better electron transfer dissociation (ETD) spectra, which are necessary for glycosite mapping. To illustrate this concept, ETD spectra of three representative CD43 peptides are shown in Fig. 3b. In the untreated sample (bottom panel) site-localization of the three O-GalNAc modifications was not possible, but StcE treatment (top panel) resulted in two peptides covering the same sequence, each with sufficient charge and fragmentation for site-localization of the modification. We note also that in silico searches for peptides with serine or threonine at their N-terminus may aid in database searches of StcE-cleaved samples (SI Appendix, Fig. S8).</p><!><p>Next, we tested StcE's ability to cleave native, human-derived mucins. We found that a commercially available semi-crude preparation of MUC16 from cancer patient ascites fluid was sensitive to StcE cleavage (Fig. 4a and SI Appendix, Fig. S9). The density around 200 kDa in the semi-crude preparation does not originate from full-length glycosylated MUC16, however, which migrates with an apparent molecular weight in the megadalton range 32,33 . To demonstrate cleavage of full-length human MUC16, we treated crude ascites taken from an ovarian cancer patient with StcE. In untreated ascites, we detected density in the stacking gel (arrow, Fig. 4b and SI Appendix, Fig. S10), consistent with a very high molecular weight species. StcE treatment for 1 h at 37 °C resulted in a dose-dependent decrease in apparent molecular weight, demonstrating that StcE is active on human MUC16.</p><p>As noted above, cell surface mucins have been implicated as pathogenic drivers of cancer progression. If StcE could cleave such mucins, it would enable their functional analysis. We first tested cell viability after StcE treatment, and found that it was non-toxic to both adherent and suspension cell lines at all concentrations tested, and did not affect proliferation over days (SI Appendix, Fig. S11). Next, we treated the human breast cancer cell line SKBR3 with StcE and probed for changes in abundance of MUC16. As determined by flow cytometry analysis, StcE depleted MUC16, but had no effect on the highly abundant N-glycosylated but non-mucin HER2 receptor (Fig. 4c). We also tested StcE's effects on the breast cancer-associated mucin MUC1 using an MCF10A cell line ectopically expressing a signaling deficient form of this cell-surface mucin (MUC1∆CT) 34 . StcE readily cleaved glycosylated MUC1∆CT but was inactive on an underglycosylated form of MUC1∆CT (~125 kD) also visible on the Western blot (Fig. 4d and SI Appendix, Fig. S12 and Fig. S13). Further, StcE's activity does not appear to be cell-line dependent, as it digested cell surface MUC1 and MUC16 from cell lines derived from a variety of cancer types (Fig. 4e). These data confirm that StcE retains its activity in cellulo. Furthermore, the supernatants of StcE-treated HeLa cells, but not their vehicle-treated counterparts, stained strongly for MUC16, and the apparent molecular weight of the mucin fragments decreased with increasing enzyme concentration and treatment time (Fig. 4f and SI Appendix, Fig. S14). StcE can therefore be used as a tool to release and solubilize mucins from biological samples. To test whether StcE could be employed to purify mucins from protein mixtures, we conjugated StcE to beads using reductive amidation. A pulldown using a mixture of BSA and C1INH showed enrichment of C1INH in the elution (SI Appendix, Fig. S15), indicating StcE's potential as an enrichment tool.</p><!><p>Given StcE's specificity for mucins, we reasoned that it could be employed as a tool to discover mucin-based ligands of glycan-binding receptors whose physiological binding partners remain unknown. There is growing evidence that so-called glycan-binding proteins can recognize discrete glycoprotein or glycolipid ligands via motifs that encompass both glycan structures as well as elements of their underlying scaffolds. As a landmark example, PSGL-1 was identified as a cell-surface mucin that functions as the chief ligand for P-selectin at sites of inflammatory leukocyte recruitment 35 (notably, PSGL-1 is effectively digested with StcE (Fig. 1b)). The molecular determinants of PSGL-1 that confer P-selectin binding include a specific O-glycan structure combined with a nearby peptide motif 36 . Likewise, the immune modulatory receptor PILRa recognizes a composite mucin-derived sialoglycopeptide epitope on cognate ligands 37 . These examples hint at a rich biology for mucin glycoproteins as ligands for a variety of receptors involved in cell trafficking and immune regulation. We speculated that StcE's ability to destroy these structures on cells might help reveal mucins as binding partners of orphan receptors.</p><p>Sialic acid-binding immunoglobulin-type lectins (Siglecs) are a glycan-binding receptor family whose physiological ligands are largely unknown 38,39 . Individual family members exhibit preferences for sialosides of various linkages to underlying glycan motifs, but the specific glycoproteins or glycolipids they interact with in biological settings are mysterious. Recent work has implicated Siglecs-7 and -9 as inhibitory receptors that function similarly to the immune checkpoints PD-1 and CTLA-4, the targets of several successful cancer immune therapies 40 . Extracellularly, Siglec-7 and -9 each have a sialic acid-binding Vset domain (Fig. 5a). Intracellularly, they resemble PD-1, with C-terminal cytosolic tyrosine-based inhibitory motif (ITIM) and tyrosine-based switch motif (ITSM) domains that mediate inhibitory signaling. Enzymatic removal of sialic acids en masse from cancer cell surfaces enhances immune cell mediated clearance of those cells through loss of Siglec-7 and -9 binding 41 . Despite years of effort, however, professional ligands of Siglec-7 and -9 have not been identified, leading to a prevailing view that these Siglecs have broad and overlapping affinities for a multitude of sialylated cell surface molecules 42 .</p><p>Using soluble Siglec-Fc fusions, we assessed the effects of StcE treatment on Siglec-7 and -9 binding to SKBR3 cells. Analysis by flow cytometry showed that StcE treatment depletes Siglec-7-Fc binding but has no effect on Siglec-9-Fc binding (Fig. 5b, top for flow cytometry histogram and Fig. 5b, bottom for biological replicates). The inactive StcE point mutant E447D had no effect on binding of either Siglec-Fc. We confirmed that Siglec-7-and -9-Fc binding are both dependent on sialic acid via treatment with Vibrio cholerae sialidase (SI Appendix, Fig. S16). These results suggest that Siglec-7 recognizes mucin glycoproteins on SKBR3 cells but Siglec-9 binds structures that are resistant to StcE digestion.</p><p>In order to ensure that StcE did not simply bind to cell surface mucins and block accessibility to Siglec-7-Fc, we stained StcE and E447D treated SKBR3 cells with anti-His antibodies, which should bind the His-tagged enzymes (Fig. 5c). E447D bound cell surfaces more tightly than StcE did, but did not deplete Siglec-7-Fc binding, indicating that StcE's enzymatic activity was required for its effects. Interestingly, periodate-mediated labeling of cell surface sialic acids 43 revealed that StcE treatment had only a minor effect on total cell-surface sialic acid levels (Fig 5d). Thus, StcE removes only a small fraction of total sialosides while depleting a majority of Siglec-7-Fc binding structures. Finally, we tested a panel of cell lines for StcE-mediated depletion of Siglec-7 and -9 ligands. In all other cell lines tested, Siglec-7-Fc binding decreased upon StcE digestion while Siglec-9-Fc binding remained unchanged (Fig 5e and SI Appendix, Fig. 16). Note that previous reports that glycolipids such as GD3 can serve as ligands for Siglec-7 44 are consistent with StcE's ability to substantially, but not completely, abrogate Siglec-7-Fc binding.</p><p>To further support the identification of Siglec-7 as a sialomucin-binding receptor, we employed ldlD Chinese Hamster Ovary (CHO) cells, which are deficient in UDP-glucose/galactose-4epimerase (GALE) 45 . GALE interconverts UDP-glucose and UDP-GlcNAc to UDP-galactose and UDP-GalNAc, respectively. Without active GALE, ldlD CHO cells can still take up glucose from tissue culture media and use it to biosynthesize nucleotide sugars of glucose, mannose, fucose, and sialic acid. However, they cannot initiate or elaborate their glycans with GalNAc or galactose, resulting in truncated cellular glycans. Supplementing the media with 10 µM galactose and 100 µM GalNAc rescues the phenotype, as these undergo conversion to the respective nucleotide sugars within cells 45 . Unrescued ldlD CHO cells exhibited weak binding by both Siglec-7-and -9-Fc (Fig. 5f). Siglec-9-Fc binding increased by approximately the same amount after rescue with galactose alone and with both galactose and GalNAc supplementation, but increased only slightly with GalNAc rescue alone (Fig. 5f, right). These results are consistent with a view that Siglec-9 ligands are predominantly non-mucinous, as GalNAc deficiency should abrogate mucin-type O-glycosylation 46 . Siglec-7-Fc binding was largely unaffected by galactose supplementation alone, but increased with GalNAc supplementation. Rescue with both sugars increased Siglec-7-Fc binding further (Fig. 5f, left). In all conditions, both Siglec-7-and -9-Fc binding were sialidase sensitive, confirming their dependence on sialic acid (SI Appendix, Fig. 16). In addition, StcE treatment had no effect on Siglec-9-Fc binding across any rescue condition, but decreased Siglec-7-Fc binding in cases with GalNAc supplementation (SI Appendix, Fig. 16).</p><p>These data distinguish the specificities of Siglec-7 and Siglec-9 on cell surfaces. In the case of Siglec-7, it appears professional glycoprotein ligands may exist, and that at least a subset are mucin-domain glycoproteins. If specificity for a professional glycoprotein ligand could be defined, it would pave the way for a novel class of immune checkpoint interventions. As noted above, the role mucin-domain glycoproteins play in immunological signaling is not limited to Siglec-7. For example, receptors such as CD45 47 and TIM-3 48 , which are emerging as critical players in healthy immune function and the immune response to cancer, contain prominent mucin domains that are thought to be necessary for their activities. Further, several members of the galectin family, which are pro-oncogenic glycan-binding proteins, are known mucin-binders, but their specificities for discrete glycoproteins have not been fully characterized 49 . Enzymatic de-mucination with StcE could provide a powerful tool for de-orphanizing the receptors and ligands that interact with mucin-domain glycoproteins.</p><p>In conclusion, we introduce the bacterial protease StcE as a robust and efficient tool for the analysis of mucin-domain glycoproteins. After solving its peptide-, glycan-, and secondary structure-based specificity, we demonstrated that this enzyme can be used in glycoproteomic mapping of mucin glycosites and their associated glycoforms, as a method for selective cleavage, release, and enrichment of mucins from cell and tissue material, and in the study of native mucin biology. The ease of purification, stability, and potency of the enzyme make it a reagent easily distributed and utilized by a variety of laboratories. Further exploration and engineering of bacterial mucinase families may yield enzymes with complementary glycan and peptide specificities that would expand the biochemist's toolbox further.</p><!><p>Recombinant and purified mucins were purchased from Molecular Innovations (C1INH), and R&D Systems (MUC16, podocalyxin, CD43, PSGL-1, Syncam-1, and CD45). To test StcE's activity against mucin-like glycoproteins and non-mucins, reaction conditions were as follows: 1:10 enzyme:substrate (E:S) ratio, total volume of 15 µL, buffer 0.1% Protease Max in 50 mM ammonium bicarbonate, 3 hours at 37 °C. A portion of each condition (0.5 µg) was loaded onto 10% Criterion™ XT Bis-Tris precast gels (Bio-Rad) and run with XT-MES (Bio-Rad) at 180 V for 1 h. Each gel was stained with silver stain or Pro-Q Emerald 300 Glycoprotein Gel and Blot Stain Kit® (Thermo Fisher Scientific), according to manufacturer's instructions. Deglycosylation of rhMUC16 was performed according to manufacturer's instructions (Deglycosylation Mix, Promega). Mucin mimetic copolymer consisting of 50% GalNAc-Ser and 50% Lys was synthesized as previously described 28 . Both deglycosylated polymer and untreated polymer were subjected to StcE cleavage and gel staining as described above for recombinant protein substrates. For peptide cleavage assays, four synthetic peptides were subjected to StcE treatment. The peptide sequences were as follows: RPPI(T-GalNAc)QSSL, IPV(S-GalNAc)SHNSL, IPVS(S-GalNAc-Galactose)SHNSL, and DRV(Y-Phosphate)IHPF. StcE was added (1:10 E:S) to 50 µL of a 500 fmol/µL solution containing all four peptides for 3 hours at 37 °C. The solution was subjected to a C18 cleanup and MS analysis as described below and in the SI Appendix.</p><!><p>See SI Appendix for sample preparation. Samples were reconstituted in 10 µL 0.1% formic acid (Fisher Scientific) containing 25 fmol/µL angiotensin (Millipore Sigma) and vasoactive peptide (Anaspec). Samples were analyzed by online nanoflow LC-MS/MS using an Orbitrap Fusion Tribrid mass spectrometer (Thermo Fisher Scientific) coupled to a Dionex Ultimate 3000 HPLC (Thermo Fisher Scientific). A portion of the sample (4 µL) was loaded via autosampler onto a 20 µL sample loop and injected at 0.3 µL/min onto a 75 µm x 150 mm EASY-Spray column (Thermo Fisher Scientific) containing 2 µm C18 beads. The column was held at 40 °C using a column heater in the EASY-Spray ionization source (Thermo Fisher Scientific). The samples were eluted at 0.3 µL/min using a 90 minute gradient and a 185 minute instrument method. Solvent A was comprised of 0.1% formic acid in water, whereas Solvent B was 0.1% formic acid in acetonitrile. The gradient profile was as follows (min:%B) 0:3, 3:3, 93:35, 103:42, 104:98, 109:98, 110:3, 185:3. The instrument method used an MS1 resolution of 60,000 at FWHM 400 m/z, an AGC target of 3e5, and a mass range from 300 to 1,500 m/z. Dynamic exclusion was enabled with a repeat count of 3, repeat duration of 10 s, exclusion duration of 10 s. Only charge states 2-6 were selected for fragmentation. MS2s were generated at top speed for 3 s. HCD was performed on all selected precursor masses with the following parameters: isolation window of 2 m/z, 28-30% collision energy, ion trap or orbitrap (resolution of 30,000) detection, and an AGC target of 1e4 ions. ETD was performed if (a) the precursor mass was between 300-1000 m/z and (b) 3 of 7 glyco-fingerprint ions (126.055, 138.055, 144.07, 168.065, 186.076, 204.086, 274.092, 292.103) were present at +/-0.5 m/z and greater than 5% relative intensity. ETD parameters were as follows: calibrated charge-dependent ETD times, 2e5 reagent target, precursor AGC target 1e4. See SI Appendix for MS data analysis.</p><!><p>See SI Appendix for cell culture materials and methods. Cells were treated with StcE or E447D when plated or after lifting with enzyme-free cell dissociation buffer (Thermo Fisher Scientific). Typical treatment conditions were 5 µg of StcE per 1 million cells in 1 mL of complete media or Hank's Buffered Salt Solution (HBSS) for two hours at 37 °C. After treatment, cells were washed with PBS or HBSS. For flow cytometry, cells were resuspended in cold PBS with 0.5% bovine serum albumin and transferred to a 96-well V-bottom plate. Cells were then resuspended in the probe of interest. See SI Appendix for flow cytometry antibody vendors and treatment conditions. Flow cytometry data was analyzed using FlowJo v. 10.0 (Tree Star). For Western blots, supernatants post-treatment (1 mL volumes) were collected into tubes containing 75 µL of 0.5 M EDTA to quench the reaction, then snap frozen in liquid nitrogen and lyophilized to dryness. Post-treatment cells were washed with enzyme-free cell dissociation buffer, which contains EDTA, to quench the reaction. Cells were then washed two times with PBS, pelleted, and lysed with sample buffer (1x NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) supplemented with 25 mM DTT). Genomic DNA was sheared via probe tip sonication. Lyophilized supernatants were brought up in sample buffer. Both cell lysates and supernatants were boiled for 5 min at 95 °C, spun at 14,000 xg for 2 min, and 30 µL of each was loaded into an 18-well 4-12% Criterion™ XT Bis-Tris precast gel (Bio-Rad). The gel was run with XT-MOPS (Bio-Rad) at 180 V for 1 h. Proteins were transferred to 0.2 µm nitrocellulose using the Trans-Blot® Turbo™ Transfer System (Bio-Rad), at 2.5 A constant for 15 min. Total protein was quantified using REVERT stain (LI-COR Biosciences) or Ponceau-S stain (Millipore Sigma). See SI Appendix for Western blot antibody vendors and treatment conditions, as well as StcE treatment of ovarian cancer patient derived ascites fluid.</p><!><p>Using the 2016 Molecular Operating Environment (MOE) software suite, the X-ray crystal structures of StcE (PDB ID: 3UJZ), astacin (PDB ID: 1QJI), and serralysin (PDB ID: 3VI1) were superimposed using the residues (HEXXHXXGXXH) of their conserved metzincin active site 30 . The individual structures were then prepared by (a) capping any termini with acetyl or NMe groups and (b) adding unresolved atoms (side chains and hydrogens) so that each structure was at its proper valency and charge. In their cocrystal structures, the peptidomimetic/peptidic ligands bind in the active site of astacin/serralysin in similar conformations, with the ligands' P2-P1' residues forming antiparallel β-sheets with the enzymes. These crystallographic ligands were thus used as scaffolds to construct the three different ligands used in our docking studies: Ac-PTLTH-NMe, Ac-P(GalNAcα-)TLTH-NMe, and Ac-P(GalNAcα-)TL(GalNAcα-)TH-NMe, where Pro is the P3 residue and His is the P2' residue. Using the Amber10:EHT forcefield 50 , each of the three ligands underwent a brief dynamics simulation to generate a corresponding library of >15,000 conformers, with the individual conformations varying solely in the arrangement of their side chains and GalNAc moieties. Each conformer and the prepared StcE(E447D) structure underwent induced fit docking, again using the Amber10:EHT forcefield, to yield minimized ligand-enzyme complexes. Docking studies containing the normal catalytic E447 residue and/or solvent molecules did not yield reasonable or reproducible results.</p><!><p>Fresh frozen ovarian cancer patient-derived ascites fluid was rapidly thawed in a room temperature water bath, then centrifuged at 500 xg for 5 min at 4 °C to remove cellular debris. A portion of clarified solution (50 µL) was treated with 5, 0.5, 0.05, or 0.005 µg StcE for 1 h at 37 °C. An aliquot of reaction solution (22.5 µL) was removed to tubes containing 7.5 µL 4x NuPAGE LDS Sample Buffer (Thermo Fisher Scientific) + 100 mM DTT, then boiled for 5 min at 95 °C to quench the reaction. Boiled samples were spun at 14000 xg for 2 min, then 20 µL of each was loaded onto an 18-well 4-12% Criterion™ XT Bis-Tris precast gel (Bio-Rad), and the gel was run with XT-MOPS (Bio-Rad) at 180 V for 1 h. Proteins were transferred to 0.2 µm nitrocellulose using the Trans-Blot® Turbo™ Transfer System (Bio-Rad) at 2.5 A constant for 15 min. Total protein was quantified using REVERT stain (LI-COR Biosciences). Western blotting for MUC16 was performed using anti-MUC16 antibody [X75] (Abcam) according to manufacturer recommendations. IRDye® 800CW Goat anti-Mouse IgG (LI-COR Biosciences) was used according to manufacturer recommendations. Reactions with semi-crude Cancer Antigen 125 (Lee BioSolutions) were performed in the same manner as recombinant substrates (see above) and immunoblotted with anti-MUC16 antibody as was done for patient-derived ascites fluid.</p><!><p>Cells were grown in T75 flasks (Thermo Fisher Scientific) and maintained at 37 °C and 5% CO 2 . BT-20, HeLa, and MDA-MB-453 cells were cultured in DMEM supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. SKBR3, K562, and ZR-75-1 cells were cultured in RPMI supplemented with 10% FBS and 1% penicillin/streptomycin. Ldl-D CHO cells were cultured in 1:1 DMEM/F12 with 3% FBS and 1% penicillin/streptomycin. MCF10A MUC1∆CT cells were cultured in phenol red free 1:1 DMEM:F12 supplemented with 5% New Zealand horse serum (Thermo Fisher Scientific), 20 ng/mL epidermal growth factor (Peprotech), 0.5 μg/mL hydrocortisone (Millipore Sigma), 100 ng/mL cholera toxin (Millipore Sigma), 10 μg/mL insulin (Millipore Sigma), and 1% penicillin/streptomycin. MUC1∆CT was induced with 200 ng/mL doxycycline for 24 h. 26 . The docked peptides Ac-PTLTH-NMe (magenta sticks), Ac-P(GalNAcα-)TLTH-NMe (cyan sticks), and Ac-P(GalNAcα-)TL(GalNAcα-)TH-NMe (green sticks) derived from a StcE-labile peptide sequence in podocalyxin all adopted a common backbone conformation that was consistent with that of ligands bound to homologous metzincin enzymes (PDB IDs: 1QJI and 3VI1) 30 . Fig. S8. StcE increased the total number of peptides with N-terminal T/S. Using the PNGaseF treated samples detailed in Supplemental Table 1, we calculated the total number of peptides whose N-terminus was either serine or threonine ("TS peptides"). In all proteins studied, the total number of TS peptides was higher due to the presence of StcE-cleaved peptides ("StcEconsensus peptides"). The increase in TS peptides may aid in database searches of StcEcleaved samples. . A portion of the beads (50 µL) was incubated with 10 µg BSA and 1 µg C1INH in a total volume of approx. 100 µL PBS. EDTA was added at 25 mM to inhibit cleavage of the substrate. After the reaction, the beads were pelleted and the supernatants were saved ("flow through"). The beads were washed once with 100 µL PBS ("wash 1"), once with 100 µL 1% Tween ("wash 2"), and once with 100 µL PBS ("wash 3"). For the elution, 32 µL 1X NuPage LDS sample buffer (Thermo Fisher Scientific) was added and the beads were boiled. Samples visualized by silver stain.</p>

ChemRxiv

Protective role of testis-specific peroxiredoxin 4 against cellular oxidative stress

Peroxiredoxin (PRDX), a newly discovered antioxidant enzyme, has an important role in hydrogen peroxide reduction. Among six PRDX genes (PRDX1–6) in mammals, PRDX4 gene is alternatively spliced to produce the somatic cell form (PRDX4) and the testis specific form (PRDX4t). In our previous study, PRDX4 knockout mice displayed testicular atrophy with an increase in cell death due to oxidative stress. However, the antioxidant function of PRDX4t is unknown. In this study, we demonstrate that PRDX4t plays a protective role against oxidative stress in the mammalian cell line HEK293T. The PRDX4t-EGFP plasmid was transferred into HEK293T cells; protein expression was confirmed in the cytoplasm. To determine the protective role of PRDX4t in cells, we performed image-based analysis of PRDX4t-EGFP expressed cells exposed to UV irradiation and hydrogen peroxide using fluorescent probe CellROX. Our results suggested that PRDX4t-EGFP expressed cells had reduced levels of oxidative stress compared with cells that express only EGFP. This study highlights that PRDX4t plays an important role in cellular antioxidant defense.

protective_role_of_testis-specific_peroxiredoxin_4_against_cellular_oxidative_stress

Introduction<!>Cell culture and transfection<!>Immunoblot analysis<!>PRDX activity assay<!>Detection of reactive oxygen species (ROS)<!>Statistical analysis<!>Expression and localization of PRDX4t-EGFP in HEK293T cells<!>PRDX4t-expressing cells showed lower ROS levels compared to control cells<!>Cytometrical analysis indicated that PRDX4t plays a protective role against oxidative stress caused by H2O2 treatment or UV-irradiation<!>Discussion

<p>Antioxidant enzymes, such as superoxide dismutase (SOD), glutathione peroxidase (GPx), and catalase, play protective roles against oxidative stress caused by elevated reactive oxygen species (ROS), as well as antioxidants.(1) Peroxiredoxins (PRDXs) have attracted attention in recent years as a new family of thiol-specific antioxidant proteins.(2) Six distinct genes comprise the mammalian PRDX family and have been divided into three PRDX subtypes; four typical 2-Cys PRDX, one atypical 2-Cys PRDX, and one 1-Cys PRDX.(3) The major functions of these PRDXs include thioredoxin (Trx)-dependent peroxidase activity,(2) modulation of intracellular signaling through hydrogen peroxide (H2O2) as a second messenger, and regulation of cell proliferation.(4–6) Although other divergent biological functions have been reported for individual PRDX members, the detailed antioxidant function of PRDX family members remains unknown.</p><p>Among mammalian PRDX family members, PRDX4 demonstrates unique properties. Two types of PRDX4 are alternatively transcribed from the single PRDX4 gene, somatic cell type PRDX4 and testis specific PRDX4t. The two forms of PRDX4 differ only in the N-terminal sequence derived from exon 1 (exon 1 and exon 1t, respectively) and share the sequences encoded by exons 2–7 (the catalytic center is in exon 3).(7) It has been reported that PRDX4 is primarily located in the endoplasmic reticulum (ER)/ Golgi apparatus, in spite of the presence of a secretory signal sequence encoded by Ex1.(8) PRDX4 plays an important role in regulating disulfide bond formation in proteins and protecting cells from ER stress by metabolizing hydrogen peroxide.(7,9) On the other hand, PRDX4t is entirely restricted to testicular cells, with induction that is sexual maturation-dependent. Surprisingly, PRDX4 gene knockout mice have indicated that expression of PRDX4t decreases compared to wild type mice, and that testicular atrophy and increased cell death is due to oxidative stress.(10) Therefore, we hypothesized that PRDX4t protects cells from ROS-induced damage; however, the cellular antioxidant function is poorly understood.</p><p>In this study, we show that PRDX4t contributes to the suppression and scavenging of ROS in mammalian HEK293T cells by using fluorescent microscopy and image-based cytometry. Our findings indicate that PRDX4t actually plays a protective role against oxidative stress in mammalian cells.</p><!><p>HEK293T cells were cultured in Dulbecco's Modified Eagle Medium (DMEM; Thermo Fisher Scientific, Waltham, MA) with 10% fetal bovine serum (v/v; FBS). Cells were incubated at 37°C in a carbon dioxide (CO2) gas incubator (Waken B Tech Co., Ltd., Japan) with 5% CO2. Mouse PRDX4t cDNA (Ensembl Transcript ID: ENSMUST00000130349.2) was subcloned into the pEGFP-N1 vector (Clontech, Takara Bio Inc., Japan) and was named PRDX4t-EGFP. HEK293T cells were transfected with the PRDX4t-EGFP plasmid or the EGFP (pEGFP-N1) plasmid using the Lipofectamine reagent (Invitrogen, Thermo Fisher Scientific) according to the manufacturer's manual. Briefly, 1.5 ml tubes containing 100 µl OPTI (Gibco, Thermo Fisher Scientific), 1 µl plus reagent (Invitrogen), and 2 µg PRDX4t-EGFP plasmid or EGFP plasmid, respectively, were prepared and incubated for 5 min at 25°C. After that, 100 µl OPTI and 2.5 µl lipofectamine agent were added to each tube, and tubes were incubated for 30 min at 37°C. After 24 h incubation, HEK293T cells were washed by OPTI-MEM (Gibco, Thermo Fisher Scientific), and incubated with reagent mixed by 800 µl OPTI-MEM added to the cell dishes. After 5 h incubation at 37°C, dishes were treated with 1 ml DMEM containing 20% FBS (final FBS concentration 10%; v/v) and incubated overnight at 37°C.</p><!><p>HEK293T cells transfected with the PRDX4t-EGFP or EGFP plasmid were washed three times with PBS and lysed in buffer (20 mM Tris-HCl, 2% protease inhibitor cocktail; v/v), followed by centrifugation at 17,000 g for 10 min. Protein concentrations of the supernatant were determined using a BCA protein assay kit (Thermo Fisher Scientific). Cell fractionation was performed using the ProteoExtract subcellular proteome extraction kit (Calbiochem, Merck, Darmstadt, Germany) followed by concentration using a common methanol/chloroform protein precipitation method. SDS-PAGE was perform with 10% polyacrylamide gels (w/v); separated proteins were transferred to polyvinylidene fluoride (PVDF) membranes (AmershamHybond P; GE Healthcare, Little Chalfont, UK), blocked for 2 h in 1% skim milk in TBST (w/v; 0.1% TBS and 0.05% Tween-20), and probed overnight at 4°C with polyclonal anti-rat/anti-mouse PRDX4 antibody.(10) After binding of the appropriate HRP conjugate anti-rabbit IgG antibody (Santa Cruz Biotechnology Inc., Santa Cruz, CA), the ECL plus western blotting detection system (GE Healthcare) was used. Results are shown as one representative experiment.</p><!><p>Harvested cells were washed twice with PBS and homogenized by sonication in tubes with buffer (20 mM Tris-HCl, 2% protease inhibitor cocktail; v/v), followed by centrifugation at 17,000 g for 10 min at 4°C. Supernatants containing proteins were transferred to new tubes and used for experiments as samples. Each sample was measured for protein concentration using the BCA protein assay kit before the extractions. PRDX activity was determined using an indirect assay that links PRDX-mediated oxidation of thioredoxin (Trx) with the recycled reduction of Trxox (-S-S-) to Trxred (-SH) by TrxR (thioredoxin reductase) using NADPH as the reductant. Quantification of the PRDX activity was assayed by measuring the decomposition of NADPH by monitoring absorbance at 340 nm at 37°C for 10 min. The reaction was started by the addition of the reaction buffer containing 200 µM NADPH, 1.5 µM yTrx, 0.8 µM yTrxR, 50 mM Hepes-NaOH buffer (pH 7.0), and 1 mM EDTA to 100 µg total protein following addition of 100 µM H2O2. The PRDX activity was defined as the rate of disappearance of NADPH, and we calculated arbitrary units relative to the value from the control.</p><!><p>ROS were detected using the cell-permeable, peroxide-sensitive probes, CellROX Orange Reagent and CellROX Deep Red Reagent (Invitrogen) according to the manufacturer's instructions. The dye exhibits bright orange fluorescence upon oxidation by ROS. We prepared HEK293T cells transduced with blank, PRDX4t-EGFP plasmid, or EGFP plasmid for 24 h. For H2O2 stress assays, cells were incubated with 5 µM CellROX Orange reagent in PBS for 30 min; 250 µM H2O2 was added after 15 min of treatment. For UV irradiation stress assays, cells were incubated with 5 µM CellROX Orange reagent in PBS for a 5 min period of irradiation with UV-B (312 nm, 5 mJ/cm2; TF-20M; Vilber Lourmat, Marne la Vallée, France) followed by incubation at 37°C for 30 min. The cells were observed using a Leica AF 6000 LX fluorescence microscope system (Leica Microsystems, Leica, Wetzlar, Germany). Fluorescence signal intensity was calculated by ImageJ software (Wayne Rasband, NIH) as previously described.(11,12) Cells were also harvested by trypsin treatment following washing with PBS (two times) for cell cytometry analysis; harvested cells were resuspended in DMEM. The cell samples (25 µl) were loaded into the half moon-shaped sample loading areas of Tali Cellular Analysis Slide (Thermo Fisher Scientific). They were examined by a Tali image-based cytometer (Life Technologies), which is a 3-channel (bright field, green fluorescence, and red fluorescence) benchtop cytometer. CellROX+ ratios in EGFP+ cells were calculated as oxidative damaged cell ratios.</p><!><p>Statistical differences were determined by the two-sided Mann-Whitney's U test. Differences with p<0.05 were considered significant. All data in graphs are presented as means ± SEM.</p><!><p>To determine the localization of PRDX4t in mammalian cells, we first tried to express PRDX4t in the mammalian cell line HEK293T. We prepared a PRDX4t-EGFP fusion construct (pEGFP-N1-PRDX4t) and an EGFP control plasmid (pEGFP-N1) for transfection assays (Fig. 1A). Typically, endogenous mouse PRDX4t has been shown to localize only to the cytosol of testicular cells.(13) Consistent with this previous report, we observed the expression of PRDX4t-EGFP in the cytosol of HEK293T cells using a fluorescence microscope system (Fig. 1B). PRDX4t-EGFP protein localization was also determined by immunoblot after cell fractionation and PRDX4t-EGFP was present only in the cytosolic fraction (Fig. 1C). In addition, we found that the PRDX activity of PRDX4t-EGFP expressed cells was significantly high compared with that of the control cells (Fig. 1D). These results indicate that the PRDX4t-EGFP expressed cells may serve as a model for facilitating the understanding of the antioxidant function of PRDX4t in mammalian cells.</p><!><p>Since it has been reported that the overexpression of antioxidant enzymes protects cells against oxidative stress,(14–16) we examined whether PRDX4t-transfection increases the antioxidant ability of the cells. In this investigation, we performed fluorescence microscopy after the PRDX4t-EGFP-CellROX combination method. As a result, we observed that cells expressing PRDX4t-EGFP demonstrated decreased CellROX fluorescence compared to cells just expressing EGFP (Fig. 2A). Quantification of CellROX fluorescence revealed that PRDX4t expression contributes to the cellular antioxidant ability after oxidative stress, even in untreated control cells (Fig. 2B). Therefore, we were able to demonstrate that PRDX4t can actually play a protective role against oxidative stress in mammalian cells.</p><!><p>We quantified the antioxidant effect of PRDX4t in HEK293T cells using an image-based cytometer after treatment with 250 µM H2O2 or UV (312 nm)-irradiation for 5 min. Cells, which were mainly transfected with PRDX4t-EGFP or the EGFP control plasmid, were located in the right field on the panels (EGFP+ cells; Fig. 3). The histograms indicate oxidatively damaged cell ratios, calculated as average CellROX+ ratios in the EGFP+ cells. Interestingly, the percentage of oxidatively damaged fractions of PRDX4t-EGFP-expressing cells was lower than the percentage of EGFP-expressing cells, even in the untreated condition (Fig. 3A). Consistent with microscopic observations, PRDX4t-EGFP-expressing cells showed high resistance against oxidative stress after H2O2 treatment or UV-irradiation compared to EGFP-expressing cells (Fig. 3B and C).</p><!><p>PRDX4t, a newly described member of the PRDX family, is specifically expressed in testicular cells.(13) PRDX4 gene knockout mice, in which PRDX4t expression is decreased compared to wild type mice, show increased spermatogenic cell death due to oxidative stress.(7,10) However, the antioxidant function of PRDX4t against oxidative stress has not yet been evaluated. In the present study, the overexpression of PRDX4t protected HEK293T cells from H2O2- or UV-induced oxidative stress as determined by image-based cytometer analysis (Fig. 3). These results are compatible with the typical enzymatic function of PRDX in antioxidant defense. This study further evaluated the antioxidant function of PRDX4t by fluorescence microscopy. Consistent with image-based cytometer results, we determined that PRDX4t-expressing cells achieve a higher resistance against oxidative stress than control cells (Fig. 2). In the two image-based analysis, especially in Fig. 3A, we detected that overexpression of PRDX4t suppresses oxidative stress in cultured cells even in control condition. We considered that this was caused by stress accumulation during experiment operation and in vitro culture stress because of ambient 21% oxygen. Moreover, we observed higher PRDX activity in PRDX4t overexpressed cells than in control cells (Fig. 1D). These results indicate that PRDX4t plays a protective role against oxidative stress in mammalian cells.</p><p>Generally, mammalian PRDXs are classified as six isoforms.(3) These PRDXs are thought to have Trx-dependent peroxidase activity, in which H2O2, as well as a wide range of organic hydroperoxides (ROOH), are reduced and detoxified.(2) Interestingly, previous studies have demonstrated that typical 2-Cys PRDXs, which include mammal PRDX1–4, play as regulators of H2O2-sensing cellular signaling.(5,6,17,18) PRDX4t may play another role, such as in the signal regulation of mammalian cells; therefore, further studies are needed to evaluate the function of PRDX4t, which is distinct from the antioxidant behavior in mammalian cells.</p><p>Intrinsically, PRDX4t is specifically expressed in sexually matured testes. In the testes, spermatogenic cells experience dramatic changes in the gene expression, morphogenesis, and redox environment during spermatogenesis. Although the chromatin of somatic cells is constituted by histones in ordinary somatic cells, sperm nuclear histones are replaced by protamines during the spermatogenic process.(19) The protamines of primates and rodents contain multiple cysteine residues that are oxidized to form disulfide bridges (sulfoxidation) that contribute to resistance against oxidative stress in the chromatin of spermatogenic cells; PRDX4t is thought to play an important function as a sulfoxidase during spermatogenesis.(7,20) This sulfoxidase function of PRDX4t is similar to the protein folding function of PRDX4 localized to the ER/Golgi apparatus.(21) In summary, we showed possibility that PRDX4t may not only function as a sulfoxidase but also may have an antioxidant function in spermatogenic cells during spermiogenesis.</p>

PubMed Open Access

The potential association between PARP14 and the SARS-CoV-2 infection (COVID-19)

Understanding the potential association between the poly (ADP-ribose) polymerase member 14 (PARP14) and the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) may aid in understanding the host immunopathological response to the virus. PARP14 has an emerging role in viral infections, and this article considers its potential mechanisms for action in either a pro- or anti-viral manner. It is evident that more experimental work is required; however, PARP14 appears vital in controlling the interferon response to the SARS-CoV-2 infection and has potential roles in balancing the proinflammatory cytokines of the cytokine storm. Furthermore, the SARS-CoV-2 macrodomain can prevent the PARP14-mediated antiviral response, suggesting a more complex relationship between PARP14 activity and SARS-CoV-2 infections.

the_potential_association_between_parp14_and_the_sars-cov-2_infection_(covid-19)

<!>SARS-CoV-2 infection<!>PARP14’s potential role in SARS-CoV-2 infection<!>PARP14 inhibitors<!>Conclusion & future perspective<!>

<p>The year 2020 was dominated by the effects of the 2019 coronavirus disease (COVID-19), and consequently the race to understand its mechanism of infection and uncover effective treatment options has taken precedence for many research groups. Patients with severe COVID-19 often display a unique pattern of immune dysregulation, causing us to consider the potential involvement of the poly (ADP-ribose) polymerase member 14 (PARP14, also known as ARTD8, BAL2 or CoaSt6). This posttranslational modifier enzyme is most commonly known for its involvement in promoting tumorigenesis and allergic airway diseases [1]. However, recently it has been found to have an emerging role in viral infection [1]. In this article we aim to consider the biochemical mechanism of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection and whether PARP14 is known (or is likely) to play a role. If it does have a role, would PARP14 inhibition aid in the treatment of COVID-19, or would it promote viral progression?</p><!><p>COVID-19 is a disease caused by SARS-CoV-2, a member of the Coronaviridae family. SARS-CoV-2 is an enveloped, positive-sense single-stranded RNA virus of approximately 100 nm in circular diameter and is covered in large spiked glycoproteins across the membrane surface [2]. During infection, SARS-CoV-2 utilizes ACE2 as a binding receptor and TMPRSS2 as a cofactor to activate its attachment proteins, leading to its internalization and replication within the host cells [3]. As ACE2 is widely expressed on the nasal and oropharyngeal epithelium, these areas within the upper respiratory tract are the first site of viral infection [4]. In many cases of COVID-19, patients only exhibit mild symptoms; however, a considerable population will develop far more severe symptoms, requiring intensive medical care and placing them at a higher risk of death [4].</p><p>Although the exact pathogenesis of COVID-19 and the SARS-CoV-2 infection is still being elucidated, the virus is observed to cause injury to the immune system [4]. As PARP14 has a well-established role in immune regulation, including its dysregulation of the Th2 immune response during allergic airway diseases, this route presented itself as a potential means of PARP14 involvement in COVID-19 pathogenesis. The SARS-CoV-2 immune injury is closely linked to the development of the systemic inflammatory response syndrome, and if left untreated can culminate in multiple organ failure [2]. This has also been noted in other coronaviruses such as SARS-CoV and MERS-CoV [2]. While a rapid and effective immune response is desirable during infection, an excessive response can itself cause damage. Part of the systemic inflammatory response syndrome observed in SARS-CoV-2 is the 'cytokine storm', a mass secretion of cytokines (e.g., IL-1β, IL-1RA IL-6, IL-7, IL-8), and the severity of this cytokine storm is closely linked to disease severity and mortality [3,5,6]. It is suggested that an early anti-inflammatory intervention can help prevent immune damage and thus reduce any potential injury to the nervous system [6].</p><p>Upstream from the SARS-CoV-2 cytokine storm, macrophages, dendritic cells and monocytes are activated and release IL-6, which binds to the IL-6R, activating intracellular signal transducers such as the JAK/STAT pathway (primarily via STAT1 and STAT3 isoforms), JNK proinflammatory pathway, MAPK pathway and PI3K pathway [7]. During SARS-CoV-2 infection, the high activation of these pathways causes the release of inflammatory cytokines which, in excess, become the cytokine storm [5]. IL-6 is further promoted by TLR4 from the NFKB pathway, steering the viral progression toward an excessive activation of the innate immune response, as observed through the inflammatory induced damage to the pulmonary interstitial arteriolar walls [3]. Many COVID-19 investigations focus on IL-6, claiming it is the prime suspect for inducing the proinflammatory response within the body and correlating it positively with the severity of COVID-19 symptoms [5,6]. Multiple targets have been suggested to halt or suppress this cytokine storm, including the use of tocilizumab – an anti-IL-6R monoclonal antibody which blocks the membrane receptor binding of IL-6 – or anakinra, an IL-1 receptor antagonist [5]. PI3K inhibitors have also showed promise in idiopathic pulmonary fibrosis, along with inhibition of the JNK family impairing the synthesis of H5N1 viral RNA [8]. Future proposals include inhibitors of the JAK-STAT pathway (e.g., baricitinib, which was proposed using artificial intelligence algorithms) [9].</p><p>Furthermore, several SARS-CoVs have been shown both in isolation and in vivo to suppress the cellular interferon response, potentially enhancing the virulence of SARS-CoV. However, the presence of interferon is seen to have a dual role, as demonstrated by Fehr et al.; if CoV-infected murine models lacking IFN-1 signaling (IFNAR-/-) received exogenous IFN-1 prior to the peak virus replication, the mice were completely protected from the disease [10]. These results support the use of interferons as a treatment option for SARS-CoVs and highlight the importance of interferon production in protection against the virus [10]. However, it was also seen that without the initial administration, interferons were rapidly produced later in the infection, recruiting inflammatory monocytes to the lung and causing the production of additional proinflammatory cytokines, eventually leading to lethality. The authors restated that they were unable to confidently predict the role of interferon in IFNAR-/- due to this apparent dual role [10]. Other studies report that when interferons are coadministered in vitro with the antiviral ribavirin, their activity is increased compared with administration of interferons alone [11].</p><p>Another aspect to consider regarding the virulence of SARS-CoV-2 and its potential association with the macrodomain-containing PARP14 is the SARS-CoV macrodomain. Located within the transmembrane nonstructural protein 3 (nsp3), this macrodomain is highly conserved across the Coronaviridae subfamily and is suggested to be vital for the virulence of the virus [10]. Many other viruses also encode a macrodomain which acts to bind and hydrolyze ADPr from proteins, and in SARS-CoVs is regarded as essential for its ADP-ribose-1′-phosphatase activity in vitro [10]. Studies by Fehr et al. in 2016 mutated the D1022A, N1040A, H1045A or G1130V residues of the mouse-adapted virus model MA15 to either limit or eliminate the ADP-ribose-1′-phosphatase activity [10]. They found that the SARS-CoV macrodomain suppressed the early interferon and proinflammatory cytokine response for the host immune system, consequentially promoting lung edema and lethality in the infected mice. However, in mice exposed to the virus that lacked the catalytic activity of the macrodomain (N1040A mutation), the virus induced significantly elevated expression of interferon-stimulated genes and other cytokines. Furthermore, when mice were coinfected with both wild-type and N1040A-mutated viruses, they still had better outcomes and increased interferon and proinflammatory cytokine expression than those infected with the wild-type virus alone [10]. The critical nature of the macrodomain has also been observed using the prototypical coronavirus model, mouse hepatitis virus strain JHMV (MHV), as well as the sindbis viruses, in which mutations in the macrodomain render the virus unable to cause severe hepatitis and encephalitis [10]. These results collectively demonstrate the prominent role the viral macrodomain has in promoting virulence. The Fehr et al. study also noted that, although the macrodomain was associated with dephosphorylating ADR-1″-phosphate to ADR, this intermediate had never been detected in a coronavirus infection, suggesting that the macrodomain was more likely involved in deMARylating or dePARylating target proteins [10]. The specific molecular targets of the coronavirus macrodomains still need to be explored; however, given this de-MAR/PAR activity, it is likely that the coronavirus macrodomain has some opposing role to ADP ribosylation, and likely the PARP family.</p><!><p>Although the above is not intended as a comprehensive review of the SARS-CoV-2 mechanism of action, it does highlight some areas in which PARP14 may have some involvement; in particular, SARS-CoV-2's association with inflammatory cytokines and the interferon response and its potential to reverse the MAR activity of PARP14. As previously mentioned, inflammatory cytokines are a key component of the cytokine storm exhibited in the SARS-CoV-2 infection, in particular those related to STAT1/STAT3 activation and pro-inflammatory cytokines. This expression is in opposition to those associated with PARP14 activity: PARP14 is involved in promoting the STAT6-dependent transcription of the Th2 immune response, whereas SARS-CoV-2 promotion of STAT1 transcription is associated with a Th1 response [1,12], and the predominant cytokine of PARP14 activation is IL-4 compared with SARS-CoV infections with IL-6. This opposition suggests that rather than promoting the release of inflammatory cytokines associated with SARS-CoV-2, PARP14 (if involved) may play a role in the host response to counteract the immune imbalance. A 2020 correspondence by Webb and Saad also suggested that PARP14 may play a role in the host response to SARS-CoV-2 infection by counteracting skewing of the Th1:Th2 cytokine ratio [12]. They support this theory in part due to the 32% structural homology between the macrodomain of PARP14 and SARS-CoV-2, stating that the similarity may be caused by the SARS-CoV macrodomain coevolving with ADPr proteins to counter their activity [12]. However, this potential role of PARP14 counteracting the inflammatory cytokines observed in SARS-CoV-2 infection is purely speculative and will need experimental validation.</p><p>Another potential mode of action for PARP14 in SARS-CoV-2 infection is via the interferon response. A study by Grunewald et al. in 2019 suggested that there was strong evidence showing that PARP14 ADP ribosylation (ADPr) is involved in the stimulation of IFN-1. This is promising evidence for PARP14's role in host defense against SARS-CoV-2; IFN-1 treatment is often proposed as a candidate treatment during viral infection [13]. The sensitivity of SARS-CoV (using the human coronavirus 229E) to IFN-α treatment was also increased when the coronavirus macrodomain was attenuated [12]. Furthermore, an uncontrolled, exploratory study showed that for patients with COVID-19, IFN-α2b therapy appeared to shorten the duration of viral shedding, accelerating viral clearance from the respiratory tract and reducing the concentration of IL-6 [14]. A more controlled study would need to take place to better validate these conclusions.</p><p>This association between PARP14 and its potential interferon response is further explored when investigating the coronavirus macrodomain. There is preliminary evidence to suggest that the SARS-CoV-2 macrodomain interacts with PARP14. As previously mentioned, coronaviruses with mutated or absent macrodomains were associated with reduced viral loads, as well as increased sensitivity to IFN-1 treatment in cell culture, suggesting that the coronavirus macrodomain counters antiviral activities [10,15]. A key PARP14 and coronavirus study by Gruenwald et al. firstly looked at the effect of pan-PARP inhibitors (3-aminobenzamide and XAV-939) at high concentrations against the murine hepatitis virus (MHV), finding a resulting decrease in interferon production which was not observed in the wild-type (WT) virus [15]. They inferred that the ADP ribosyltransferase activity was necessary for PARP14's antiviral activity, as the pan-PARP inhibitors target the catalytic domain; however, this is limited as the pan-PARP inhibitors target all 17 PARP enzymes [15]. To look more directly at PARP14, they used the inhibitor 8k, which is selective toward PARP14 when compared with PARP1 (with IC50 values of 0.78 and 19 μM respectively) and is slightly more selective over PARP10 at 1.4 μM [16]. Using 8k in combination with PARP14−/− bone marrow-derived macrophages (BMDMs), along with human PARP14 knockout (KO) A549 and normal human dermal fibroblast (NHDF) cells, the authors showed that PARP14 was required to induce the heightened IFN-1 production during coronavirus infection. This was consistent with other studies detailing PARP14's role in IFN-1 induction following lipopolysaccharide (LPS) stimulation of RAW 264.7 cells and BMDMs [15].</p><p>The study also details the creation of the recombinant virus N1347A, which contains an alanine mutation within the macrodomain that effectively removes the ADP hydrolase activity of MHV macrodomains. Within murine models, this virus was shown to replicate poorly and was not disease-causing, highlighting the importance of the viral macrodomain to counter antiviral activities [15]. As in the MHV studies, the authors used siRNA knockdown with the PARP14 inhibitor 8k to show that although 8k did not affect the cell viability, metabolism, or global cellular PARylation, it restored the replication of the N1347A virus in BMDMs significantly [15], thereby supporting the role of PARP14 blocking N1347A MHV replication [15]. Other PARPs may also be involved in the antiviral activity, as the pan-PARP inhibitors were able to reduce N1347A IFN-1 levels much more effectively than the partially specific PARP14 inhibitor [15]. Together, this study showed that PARP14 was required to inhibit the replication of the mutated coronavirus and demonstrated its importance in interferon expression. Gruenwald et al. were not the only ones to start investigating PARP14's emerging role in viral defense; among a few preliminary studies, an early 2018 paper investigated the previously uncharacterized association PARP14 has with the regulation of IFN-β response in murine macrophages [17]. Following endotoxin stimulation, PARP14 was shown to bind to a small group of specific interferon-stimulated gene-encoded proteins, enabling their nuclear accumulation. This was further reinforced as following the loss of PARP14, the transcription of IRF3-regulated primary response genes was attenuated, resulting in the reduction of IFN-β and the activation of secondary antiviral response genes [17], again validating PARP14's emerging role in the stimulation and regulation of type 1 interferons during a viral attack.</p><!><p>There are three potential pathways in which PARP14 may have a role in aiding the host immune system against SARS-CoV infections; however, all require further experimental validation. A key aspect of investigating the potential role of PARP14 is the availability of potent and selective PARP14 inhibitors. Currently, there are no clinically available PARP14 inhibitors, but there are some promising small-molecule lead compounds. As discussed earlier, one study utilized the PARP14 inhibitor lead compound 8k, which exhibited sub-micromolar potency against PARP14's catalytic domain while also showing the highest selectivity over PARP1 [16]. However, this compound is limited in its ability to be classified as a selective PARP inhibitor, because its potency has not been assessed against all PARP enzymes. Another promising lead has been H10, as proposed by Peng et al., which was found to inhibit PARP14 activity at an IC50 of 0.49 ± 0.07 μM in vitro [18]. H10 operates as a bidentate inhibitor, targeting both the conserved nicotinamide binding site and the less conserved adenine binding site within PARP14's catalytic domain. To date, H10 is the most potent PARP14 inhibitor; however, like 8k, it requires further assessment against the other PARP family members before we can be confident in its selectivity. Suggestions by Grundewald et al. imply that it is PARP14's catalytic domain which is responsible for its potential antiviral action. However, it would be of interest to investigate whether PARP14's macrodomain also plays a role. This aligns nicely with the current studies that are targeting the macro2 domain of PARP14, in the hope of achieving greater selectivity toward PARP14 over the other PARPs. The most effective macrodomain inhibitors target the macro2 domain and include the carbazole 108, which possesses sub-micromolar activity with an IC50 value of 0.66 ± 0.03 μM [19].</p><p>Furthermore, whilst PARP14 appears to be involved in aiding the host defense against SARS coronaviruses, the earlier DNA-dependent poly (ADP-ribose) polymerases (PARP1–3) are speculated to have an opposing role. A review by Curtin et al. proposes a clinical investigation of DNA-dependent PARP inhibitors against COVID-19 [4], highlighting that PARP1 and PARP2 actively increase and prolong inflammation and that the use of PARP inhibitors can limit this tissue damage, as well as decreasing the levels of proinflammatory cytokines [4]. They continue to pitch a COVID-19 clinical trial using the clinically approved (for primarily BRCA-mutated cancers) PARP inhibitors, justifying that the doses would likely not need to be as high as those used in oncological applications [4]. Overall, it is a very well-done study that condenses the evidence for the use of DNA-dependent PARP inhibitors in coronavirus treatment.</p><!><p>Our original aim was to investigate whether PARP14 plays a role in the pathology of COVID-19 and whether a PARP14 inhibitor would be useful as a treatment option. Overall, there needs to be more experimental evidence to support PARP14's potential role in host defense against SARS-CoV-2. However, we have identified three potential modes of action, including PARP14 counterbalancing the proinflammatory cytokines of the SARS-CoV-2 cytokine storm, PARP14's role in the interferon response to aid in clearing the virus, and the relationship between the deMARylating activity of the coronavirus macrodomain and PARP14's antiviral activity. Further investigation of these hypotheses may give us a better understanding of the host immunopathological response to SARS-CoV-2 and aid in our future design of vaccines and treatments. PARP14 inhibitors may be clinically useful in the treatment of cancers and allergic airway diseases, but for coronaviruses the development of selective PARP14 inhibitors is more likely to be of use in furthering our understanding of the immunopathological response. Finally, we emphasize again that there needs to be more evidence-based data to support the potential role of PARP14 in SARS-CoV-2 infection. Use and continual development of selective PARP14 inhibitors will help to answer some of these hypotheses, and hopefully will aid us in designing better treatments and reducing the disease burden of COVID-19.</p><!><p>Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is the causative virus for the disease COVID-19.</p><p>The SARS-CoV-2 macrodomain is shown to suppress the early interferon and proinflammatory cytokine response.</p><p>When untreated, disease progression is closely linked to the systemic inflammatory response syndrome.</p><p>There are suggestions that PARP14's role in the inflammatory response can help to counteract the Th1:Th2 cytokine ratio imbalance caused by SARS-CoV-2.</p><p>PARP14 stimulation of IFN-1 may aid in activating the host defense to SARS-CoV-2 infection.</p><p>There are no clinically available PARP14 inhibitors.</p><p>Currently the most potent PARP14 inhibitor is H10 as proposed by Pen et al., which inhibits PARP14 with an IC50 of 0.49 ± 0.07 μM in vitro.</p>

PubMed Open Access

Keeping Track of the Electrons

Mechanistic investigation and new reaction development are intertwined. This interdependence presents challenges and opportunities in development of all transformations, particularly for those that employ base metal catalysts. In comparison to precious metal counterparts, these catalysts yield less easily to mechanistic analysis. However, base metal catalysts can provide new modes of reactivity and opportunities for discovery. In this commentary we highlight a developing field: nickel-catalyzed stereoselective alkyl cross-coupling reactions. While key features of the relevant catalytic cycles remain ambiguous, chemical intuition and key mechanistic experiments have provided the stepping stones for discovery of stereoselective transformations.

keeping_track_of_the_electrons

<!>Introduction<!>Rising to the challenge: New tools and strategy<!>Case Study: Nickel catalysis of alkyl cross-coupling reactions<!><!>Outlook

<p>In matters of the intellect, do not pretend that conclusions are certain which are not demonstrated or demonstrable.</p><!><p>Development of new catalytic reactions is a vigorous and vital part of organic and organometallic chemistry. New transformations await discovery just beyond the horizon, with many new advances within and slightly beyond reach. To develop new transformations, different groups take slightly different strategies. All approaches combine phenotypic pattern recognition with potential arrow pushing reaction mechanisms. The mechanisms used as starting points range from well-validated to complete conjecture. In a typical catalytic reaction one or two of the intermediates can be characterized, but the others are fleeting and the arrows are often imagined. Nonetheless, the proposed mechanism serves to guide the experimentalist to envision the next transformation. Even a mechanism that turns out to be invalid can still serve as inspiration to catapult a project forward. We need enough information to make decisions and hypotheses, at the same time recognizing our gaps in knowledge and our assumptions. While chemical intuition can help to identify opportunities, oversimplification can stall progress and prevent transformative breakthroughs. Unlike Monty Python's Sir Bedivere, chemists must avoid the temptation to make leaps based on incorrect assumptions or faulty reasoning. A holy grail in new reaction development is conducting relevant mechanistic experiments that provide compelling support to define the structures and reactivity of key intermediates, and using this mechanistic evidence to identify new modes of reactivity (Figure 1).</p><!><p>There have been major advances in mechanistic evaluation since the "Holy Grails" issue 20 years ago, with new tools being developed and existing techniques sharpened. Traditional methods for kinetic analysis such as NMR and GC continue to be staples.2,3 Other methods have been greatly improved, for example, modern reaction calorimetry is significantly more prevalent and straightforward than it was two decades ago.4 Additionally, new tools such as measurement of kinetic isotope effects using natural abundance 13C have been developed.5,6 DFT calculations have been refined and honed to offer improved accuracy for the minute energy differences between competing transition states.7,8 To complement mechanistic analysis, strategies for high-throughput evaluation of reaction conditions and product distributions have been developed. These strategies build on the lessons learned from combinatorial chemistry efforts and provide rapid access to new reactivity and highly optimized conditions for a given transformation.9,10,11,12 Often high-throughput screens result in identification of a set of reaction conditions, including catalyst precursors, additives, and solvents, that would not have otherwise been predicted by careful analysis of the existing literature. Finally, detailed reaction parameterization is beginning to provide a new physical organic approach to analyze and predict the impact of ligand tuning on catalyst activity and selectivity.13,14,15</p><!><p>As a case study of a field that has been profoundly shaped by mechanistic investigations we examine nickel catalysis of cross-coupling reactions, with a focus on stereoselective cross-coupling reactions of secondary alkyl electrophiles or organometallic reagents. We highlight select examples that illustrate the connections between the proposed reaction mechanisms and rigorous mechanistic investigation to new catalytic advances. Comprehensive reviews provide a full picture of the development of alkyl cross-coupling reactions, including those employing alternative metal catalysts.16,17,18</p><p>In 1995, at the time of the first Holy Grails issue, palladium-catalyzed aryl-aryl coupling reactions were poised to take over medicinal chemistry.19 During the intervening years, aryl cross-coupling has developed into a robust and well-traveled reaction, with participation by designer substrates including heterocycles of all varieties. It has provided access to the chemical space where most new active pharmaceutical agents reside.20 The success of this reaction is due, at least in part, to the reliable and predictable nature of palladium catalysts. Once a typical cross-coupling mechanism was established, these reactions were extrapolated to include a wide range of electrophilic and nucleophilic partners. The palladium catalysts are straightforward to handle and mechanisms are easy to draw on a whiteboard, even for the novice organometallic chemist. While ligand tuning is often required for specialized substrates, most practicing medicinal chemists have access to a sufficient library of the key ligands, so that reactions are nearly guaranteed some measure of success.</p><p>In contrast, alkyl coupling, those reactions that employ an alkyl electrophile or alkylmetal reagent or both, was an uncommon transformation in 1995. Proof of concept for these alkyl couplings had been established at that time, but with limited scope and few applications. Unlike aryl-aryl coupling, the field has had a significant induction period, with first reports using nickel catalysts in 1970's,21 and advances by Knochel in 1995.22,23 The field gained momentum in 2001 when Fu reported the palladium-catalyzed Suzuki coupling of primary alkyl bromides,24 followed by a spring of nickel-catalyzed reactions of secondary substrates that began appearing in 2003.25 Two major factors contributed to the slow maturation of the field. First, the familiar and well-understood palladium complexes are typically not ideal catalysts for cross-coupling reactions with alkyl partners. Nickel complexes more frequently provide access to the desired reactivity, however, these catalysts are not as well-behaved or predictable. Second, in comparison to aryl couplings, alkyl couplings provide a larger array of undesired byproducts. A problematic aryl-aryl Suzuki-Miyaura coupling may provide protodeborylation as the major side product. In contrast, a challenging alkyl coupling can provide mixtures resulting from β-hydride elimination, alkylmetal isomerization, hydrogenolysis, dimerization and unreactive starting materials (Figure 2). We illustrate possible side reactions in the coupling of an alkyl electrophile with an arylmetal reagent. Similar side reactions will occur in reactions employing an alkylmetal reagent and aryl electrophile, or an alkylmetal reagent with an alkyl electrophile. The aggregate challenge of tuning a poorly understood catalyst to shut down a larger number of side reactions likely contributed to the lengthy induction period of the field. Advances after 1995 were bolstered by development of related reactions such as hydroalkylation reactions that proceed through alkylnickel complexes;16,17 lessons learned from those projects lowered the barriers, real or perceived, to development of nickel-catalyzed alkyl coupling reactions.</p><p>After decades of meticulous experiments from many laboratories, major features of the mechanisms of nickel-catalyzed alkyl cross-couplings are still unrefined and, importantly, likely differ from reaction to reaction. As a base metal complex, a nickel precatalyst can often generate several complexes of varying oxidation states in situ. From a mechanistic perspective, this means that the active catalytic species is frequently non-obvious. Nickel catalysts can also participate in single-electron and two-electron reactions,26,27 in contrast to the precious metal catalysts that typically favor two-electron reactions. Therefore, while there is robust mechanistic evidence that palladium-catalyzed cross-coupling reactions proceed through Pd(0) and Pd(II) intermediates, the oxidation states of key intermediates in nickel-catalyzed cross-coupling reactions are still ambiguous. At first glance, based on analogy to palladium and the oxidation state of catalyst precursors, one might assume that Ni(0) and Ni(II) intermediates are most likely. However, in seminal studies in the 1970's, Kochi and co-workers established that reductive elimination likely occurs from Ni(III) intermediates, and not from Ni(II) complexes (Scheme 1a).28,29 Kumada and co-workers incorporated these data into the proposed mechanism for coupling of Grignard reagents with alkyl halides, illustrating established intermediates and avoiding proposed structures for intermediates that lack structural characterization (Scheme 1b).30 Additional experiments have subsequently been reported that also implicate Ni(III) intermediates and, by extrapolation, Ni(I) intermediates.31,32,33 In general, further studies to pin down key intermediates have been hampered by the instability of the intermediates themselves, particularly with catalytically relevant ligands, and rapid redox cycling of nickel complexes in solution. For example, Ni(II) complexes have been isolated, however, whether or not they are true catalytic intermediates or simply catalyst precursors is not always straightforward to determine. Extrapolation from rigorous mechanistic studies to complex cycles must be tentative. The exact identity of the ligand will play a critical role and may alter the mechanisms, for example, in comparisons between catalysts supported by redox-active and redox-innocent ligands.32 The identity of the oxidative addition partners will also impact the mechanism. For example, DFT calculations of aryl-aryl and alkyl-aryl coupling reactions of carbamates are consistent with Ni(0)-Ni(II) catalytic cycles.34,35 These results are consistent with the mechanisms proposed for allylic substitution reactions of ethers.36,37 Current approaches must balance a desire to simplify and rationalize the observed reactivity with the knowledge that little data supports the proposed mechanisms.</p><p>Despite the fact that many key features of the catalytic cycles are incomplete, mechanistic studies have been critical in guiding the development of new alkyl coupling reactions. For example, inspired by Kochi's evidence that oxidation of Ni(II) intermediates facilitates reductive elimination, Knochel and co-workers reported that otherwise sluggish cross-couplings could be accelerated by pendant alkenes and alkynes or additives such as electron-deficient alkenes.22,23 The mechanistic hypothesis that there are likely alkyl radicals formed during oxidative addition of alkyl halides with low-valent nickel complexes lead to the Fu laboratory's discovery that chiral ligands could influence a stereoablative and stereoconvergent transformation. These examples underscore the importance of rigorous mechanistic experiments. While we may never have complete answers, we may learn enough to move forward to new discoveries.</p><p>With the viability of nickel-catalyzed alkyl cross-coupling reactions established, the field has gained momentum and is currently developing rapidly. Cross-coupling of primary substrates has begun to be adopted in synthesis, now that reasonably general catalysts and reaction conditions have been identified.38 The implementation of secondary substrates in these reactions provides the opportunity to generate new stereogenic centers with control of absolute configuration and is currently an area of rapid development. Both stereoconvergent and stereospecific alkyl cross-couplings have been established, using secondary electrophiles and organometallic reagents.17,18 Representative examples are shown in Scheme 2. Stereoconvergent reactions of racemic alkyl halides employ chiral catalysts and rely on the formation of alkyl radical intermediates during the course of the cross-coupling reaction, either during the oxidative addition event or by equilibration of diastereomeric alkylmetal intermediates (Scheme 2a).39,40,41 Stereospecific reactions require enantioenriched starting materials and mechanisms which avoid alkyl radical intermediates. Stereospecific reactions of alkyl ethers and esters have been established (Scheme 2b).42,43 Stereospecific reactions of chiral transmetallating agents, including alkylboranes44 and alkylstannatranes45 build on early reports which demonstrated that transmetallation is typically highly stereospecific (Scheme 2c).16,46 Stereoconvergent reactions of racemic alkylmetal reagents are the least well-developed reactions, although proof of concept was established using Grignard reagents in the 1980's (Scheme 2d).47 The scope of each of these transformations continues to expand and push beyond expected limitations. For example, stereospecific reactions of tertiary esters provide access to enantioenriched quaternary carbon centers.48 Creative strategies to enter the catalytic cycle invigorate the field, including employing photocatalytic conditions for formation of alkylradical intermediates.49,50,51</p><p>Branching away from robust aryl-aryl cross-coupling reactions to explore the unknown territory of alkyl cross-couplings is providing powerful new approaches for stereoselective C–C bond formation. Many challenges and questions remain to be addressed. Will this reaction ever be truly general? Will it be embraced by medicinal chemistry in the same way that aryl cross-couplings have? How much more will we learn about the mechanisms? Development of these reactions has already begun to have long-range impacts on related transformations, for example, in providing the framework for control of stereochemistry in cross-electrophile coupling reactions.52,53,54,55</p><!><p>For a research worker, the unforgotten moments of life are those rare ones, which come after years of plodding work, when the veil over nature's secret seems suddenly to lift and when what was dark and chaotic appears in a clear and beautiful light and pattern.</p><!><p>Development of new catalytic reactions flourishes in the intersection of organometallic and physical organic chemistry. Nickel-catalyzed alkyl cross-coupling reactions provide examples of how chemists employ hard-won and sometimes ambiguous mechanistic data to guide discovery of new transformations. As the field of physical organic chemistry continues to evolve, new experimental methods will provide increased resolution for challenging mechanisms. Continued growth and improved fundamental understanding of transition metal complexes, including base metal catalysts, will provide access to new modes of reactivity with yet-undiscovered applications in organic synthesis. The next transformative advances will continue to be driven by the curiosity of chemists who want to better understand their reactions.</p>

PubMed Author Manuscript

TiO2 ALD Coating of Amorphous TiO2 Nanotube Layers: Inhibition of the Structural and Morphological Changes Due to Water Annealing

The present work presents a strategy to stabilize amorphous anodic self-organized TiO2 nanotube layers against morphological changes and crystallization upon extensive water soaking. The growth of needle-like nanoparticles was observed on the outer and inner walls of amorphous nanotube layers after extensive water soakings, in line with the literature on water annealing. In contrary, when TiO2 nanotube layers uniformly coated by thin TiO2 using atomic layer deposition (ALD) were soaked in water, the growth rates of needle-like nanoparticles were substantially reduced. We investigated the soaking effects of ALD TiO2 coatings with different thicknesses and deposition temperatures. Sufficiently thick TiO2 coatings (≈8.4 nm) deposited at different ALD process temperatures efficiently hamper the reactions between water and F− ions, maintain the amorphous state, and preserve the original tubular morphology. This work demonstrates the possibility of having robust amorphous 1D TiO2 nanotube layers that are very stable in water. This is very practical for diverse biomedical applications that are accompanied by extensive contact with an aqueous environment.

tio2_ald_coating_of_amorphous_tio2_nanotube_layers:_inhibition_of_the_structural_and_morphological_c

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<!><!>Results and Discussion<!>Conclusion<!>Author Contributions<!>Conflict of Interest Statement<!><!>Supplementary Material<!>

<p>Various morphologies of TiO2 with nano scale dimensions have been extensively investigated as photo catalysts for H2 evolution, dye-sensitized solar cells (DSSCs), degradation of organic compounds, methanol oxidation, CO2 reduction, self-cleaning and anti-fogging, and many other applications (Chen and Mao, 2007; Schneider et al., 2014; Wang et al., 2014). Particularly in the last 15 years, the anodic self-organized TiO2 nanotube layers have attracted scientific interests in the mentioned areas. This is mainly attributed to the controllable geometry and large specific surface area of the anodic TiO2 nanotube layers which allow higher reaction activities as well as the one-dimensional (1D) orientation which offers unidirectional charge transport from the tubes to the supporting Ti substrate (Macak et al., 2007; Lee et al., 2014; Wang et al., 2014).</p><p>Generally, the as-anodized TiO2 nanotube layers in the amorphous state are not favored in semiconducting applications such as photo catalysts and DSSCs, primarily due to their low conductivity and a significant number of recombination centers which impede efficient charge transport (Roy et al., 2011; Krbal et al., 2016). As the electronic properties are influenced by the structural quality of the nanotube layers (Tsuchiya et al., 2007), post-thermal annealing in temperature range of 280–600°C for 1–3 h (Varghese et al., 2003; Tighineanu et al., 2010) or hydrothermal treatment (Yu et al., 2010) needs to be carried out to crystallize the nanotube layers.</p><p>For a long time, only crystalline TiO2 nanomaterials have been comprehensively studied, whereas its amorphous counterparts have not received much attention so far. In spite of the strong focus on crystalline TiO2 forms (anatine or rutile) that show higher performance in diverse applications, amorphous TiO2 structures have increasingly showcased its popular role in various semiconductor applications as well (Lu et al., 2008; Ortiz et al., 2008; Djenizian et al., 2011; Xiong et al., 2011; Bi et al., 2013; Wang et al., 2015; Jiang et al., 2016; Liang et al., 2018; Liu et al., 2018). TiO2 has been long recognized as excellent biocompatible material owing to its low cytotoxicity, high stability, and antibacterial properties (Fu and Mo, 2018). Its amorphous state is particularly preferred in biomedical applications, including carrier for magnetic nanoparticles for protein purification (Kupcik et al., 2017), supporting layer for enhanced hydroxyapatite (Hap) deposition in Osseo integration (Kar et al., 2006; Crawford et al., 2007), supporting layer for epithelial cells and fibroblasts viability (Mei et al., 2014) and improved magnetic resonance contrast for molecular receptor targeted imaging (Chandran et al., 2011). In the case of TiO2 nanotube layers, controllable nano-geometry, surface modification, topography, and roughness are crucial for tissue and cell vitality (seeding, spreading, and proliferation) (Park et al., 2007; Peng et al., 2009; Fu and Mo, 2018). On top of its hemocompatibility (Huang et al., 2017), the tubular morphology is an added advantage for genes, drugs, and therapeutic carrier or reservoirs, for example, gentamicin sulfate, chitosan, bone morphogenetic protein 2, and tumor necrosis factor-related apoptosis-inducing ligand (Hu et al., 2012; Feng et al., 2016; Kaur et al., 2016).</p><p>For the mentioned biomedical applications, the TiO2 nanotube layers are frequently used in an aqueous environment. Nevertheless, the soaking of the as-anodized amorphous TiO2 nanotube layers in a water bath transform them to polycrystalline anatase structure via so-called water annealing effect or low-temperature crystallization approach (Liao et al., 2011; Wang et al., 2011; Krengvirat et al., 2013; Lamberti et al., 2015; Cao et al., 2016). These water annealing processes are accompanied by a strong morphological transformation. As a result, the unique tubular morphology cannot be sustained in the case of prolonged soaking. Additional particle-like deposits grow on the surface of the amorphous nanotubes and may completely block them, reducing drastically the accessibility of various species inside the nanotubes and reducing also the overal available surface area. Eventually, in certain cases, the amorphous nanotubes were first transformed to double-wall nanotubes, then to core-shell wires/rods-within-tubes and finally full transformation into crystallized nanowires/rods after different soaking durations took place (Wang et al., 2011; Lamberti et al., 2015; Cao et al., 2016). The water annealed nanotubes or nanowires/rods possess much rougher surface as compared to the amorphous nanotube layers. Interestingly, when the similar soaking experiment was carried out in a cell culture environment, such as in fetal bovine serum and phosphate buffered saline (PBS) media, the amorphous TiO2 nanotube layers did not experience any structural or morphological changes (Cao et al., 2016).</p><p>To incorporate the aforementioned advantages of anodic 1D TiO2 nanotube layers in the biomedical applications, it is crucial to maintain the amorphous state and preserve the tubular morphology. In fact, it is quite common that the addition of a shell (an outer layer) serves as a protective layer for the inner core structure (Yan et al., 2012; Hu et al., 2014; Kwiatkowski et al., 2015). For instance, an ultrathin Al2O3 coating was employed to improve the chemical, mechanical, and thermal stability of TiO2 nanotube layers in extreme environments (Zazpe et al., 2017) such as for Li-ion batteries (Sopha et al., 2017a). On the other hand, a SiO2 insulating layer was utilized to encapsulate TiO2 nanoparticles to inhibit the photo catalytic activity, which undesirably darkens the white pigment of TiO2 (Guo et al., 2017), and also to improve the cell compatibility and photo-killing ability (Feng et al., 2013).</p><p>To achieve ultrathin and continuous coatings that completely enfold a high aspect ratio structure such as TiO2 nanotube layers, ALD technique is viable to provide such homogeneous and conformal coatings due to its self-saturating surface reactions (Leskelä and Ritala, 2002; Leskelä et al., 2007; Zazpe et al., 2016, 2018). Tupala et al. first performed an ALD amorphous TiO2 coating within crystalline TiO2 nanotube layers. It is worth noting that with a 5 nm amorphous TiO2 layer, the conductivity of the coated nanotube layer is substantially increased. (Tupala et al., 2012) We have also demonstrated that additional ALD crystalline TiO2 coatings within crystalline TiO2 nanotube layers passivate defects within TiO2 and enhance the charge carrier separation towards improved photo electrochemical and photo catalytic performance (Sopha et al., 2017b).</p><p>Despite other materials such as ZnO, Fe3O4, and CuO may potentially serve as a protective coating, we deliberately select an identical coating material (TiO2) for the core nanotube layers due to the fact that (i) the biocompatible TiO2 coating is required to be robust in extreme environments, and (ii) the stacking of two different materials (different densities) creates a gradient at the interface between outer and inner layers, which complicates the reactants transfer and interaction process. In the present work, we extend the application of ALD TiO2 coatings as a protective coating of amorphous TiO2 nanotube layers to prevent their morphological changes, known as water annealing effect. The longest soaking duration shown in previous works was up to 7 days (Wang et al., 2011; Cao et al., 2016). We significantly prolong the soaking duration up to 28 days to show the extreme stability of these ultrathin TiO2 coated TiO2 nanotube layers in order to broaden their functional range specifically in the aqueous environments for biomedical applications.</p><!><p>Self-organized TiO2 nanotube layers with thicknesses of ≈5 μm and inner diameters ≈230 nm were produced via electrochemical anodization as described in our previous works (Das et al., 2017). Atomic layer deposition (ALD, TFS200, Benes) of TiO2 was carried out at 150°, 225°, and 300°C using TiCl4 (electronic grade 99.9998%, STREM) and Millipore deionized water (15 MΩ) as the titanium precursor and the oxygen source, respectively. Temperature of both precursors was kept at 20°C. High purity N2 (99.9999%) was the carrier and purging gas at a flow rate of 400 standard cubic centimeters per minute sccm (Standard Cubic Centimeters per Minute). Under these deposition conditions, one ALD growth cycle was defined by the following sequence: TiCl4 pulse (500 mS)-N2 purge (3 s)-H2O pulse (500 mS)-N2 purge (4 s). The as-anodized amorphous TiO2 nanotube layers were coated by TiO2 applying different ALD cycles, NALD = 10, 50, and 150, yielding nominal thicknesses of 0.56, 2.8, and 8.4 nm, respectively. The thickness is obtained according to the growth rate per ALD cycle, evaluated from TiO2 thin layers deposited on Si wafers using variable angle spectroscopic ellipsometry using VASE® ellipsometer, J.A. Woollam.</p><p>For water soaking, the blank and TiO2 coated TiO2 nanotube layers were soaked in deionized water (18 MΩ.cm) and phosphate buffered saline (PBS) for different durations, i.e., 1, 7, 14, or 28 day(s) in a still environment at room temperature. The morphology of the blank, coated, and soaked TiO2 nanotube layers was imaged by field-emission scanning electron microscope (SEM, JEOL JSM 7500F, FEI Verios 460L). The structural evaluation was based on X-ray diffraction (XRD) measured by diffractometer (SmartLab 3kW from Rigaku). The diffractometer was set up in Bragg-Brentano geometry using Cu Kα radiation (λ = 1.54 Å) equipped by 1D-detector Dtex-Ultra. Cu lamp was operated at current 30 mA and voltage 40 kV. Phase analysis was performed based on chemical composition using databases PDF2 and ICSD. The chemical state of the blank and TiO2 coated (NALD = 150 at 300°C) amorphous nanotube layers was examined by X-ray photoelectron spectroscopy (XPS, ESCA2SR, Scienta-Omicron) using a monochromatic Al Kα (1486.7 eV) X-ray source. The survey spectra were acquired using 250 W power of X-ray source with pass energy set to 150 eV. The quantitative analysis was based on sensitivity factors provided by the manufacturer. It is noteworthy to point out that the quantitative analysis was performed in order to provide a relative comparison between a chemical composition of blank and TiO2 coated nanotube layers. The absolute values of the atomic concentration of elements are in great extent affected by the surface sensitivity of XPS.</p><!><p>The blank (as-anodized) TiO2 nanotube layers and TiO2 coated (NALD = 150 at 300°C) TiO2 nanotube layers are imaged by SEM in two regions of the nanotube layer, i.e., the top (water/nanotubes opening interface, Figures 1a,c) and the bottom (bottom of nanotubes/Ti interface, Figures 1b,d). At the top of the tubes, the blank nanotube layer clearly presents an inner diameter of ≈230 nm. It is obvious that the ALD TiO2 (≈8.4 nm) coated nanotube layer shows slightly thicker tube walls compared to the blank nanotube layer. In addition, at the bottom of the tubes, the ALD coated nanotube layer has a smaller inner diameter. These images evidence that the coating is uniform across the entire tube walls. The layer thicknesses are ≈5 μm as shown in Figure 1e. The coating thickness is confirmed by the thickness measurement on the identical TiO2 coating deposited on a flat substrate. It is an utmost challenge to differentiate the TiO2 coating and TiO2 tube wall due to the identical material, as they are of the same mass and contrast. This fact disables microscopists to distinguish them. Nevertheless, in a previous work, it was shown that the walls of 400 ALD cycles (≈22 nm) TiO2 coated TiO2 nanotube layer is visibly much thicker than the blank nanotube layer (Sopha et al., 2017b). Thus, from the ALD principle, the thickness of the present coatings follows the same trend: the higher number of ALD cycles, the thicker is the coating.</p><!><p>SEM images of (a,b) blank and (c,d) TiO2 coated (NALD = 150 ≈8.4 nm) TiO2 nanotube layers. (a,c) Are taken at top of the nanotube layer (water/nanotubes opening interface) (b,d) Are taken at the bottom of the nanotube layer (bottom of nanotubes/Ti interface). (e) Shows the entire nanotube layer with thickness ≈5 μm.</p><!><p>The blank and TiO2 coated TiO2 nanotube layers were then soaked in deionized water for different durations from 1 to 28 days. No modification in physical appearance such as changes in color was noticed. We first proceeded to investigate the structural properties on (i) blank nanotube layers soaked for all durations; (ii) nanotube layers with different ALD coating thickness (NALD = 10, 50, 150) deposited at 300°C and (iii) nanotube layers with NALD = 150 deposited at different temperatures (150°, 225°, and 300°C), the latter two cases were soaked for 28 days. The resulting XRD patterns are depicted in Figure 2. The blank nanotube layers remain amorphous after extensive soaking in deionized water up to 28 days, as only the diffraction peaks of hexagonal Ti (from the substrate) are present in the obtained diffraction patterns as shows in Figure 2A. Similarly, all ALD cycles and temperatures coated nanotube layers are in amorphous state after the identical soaking experiments. The only exception is credited to the TiO2 coating deposited with NALD = 150 at 300°C with polycrystalline anatase structure. The corresponding diffraction peaks of anatase are labeled in Figure 2B. Note that the crystallization is not induced by the water soaking; instead, it occurred during the ALD deposition process because the coating with a thickness of 8.4 nm becomes crystalline due to the deposition temperature of 300°C. This is supported by the XRD data of the amorphous nanotube layer, and the same layer then coated with NALD = 150 at 300°C prior to the soaking experiment, as shown in Figure 2C. The anatase diffraction peaks are very similar before and after soaking. Furthermore, previous studies reported that the initiation of crystallization process is influenced by the ALD deposition temperature and coating thickness (Aarik et al., 2001; Nie et al., 2015). This means that initially an amorphous coating is formed on the substrate until a (thermodynamical) threshold thickness is achieved for the nucleation of crystals. The threshold thickness reduces with the increase of deposition temperature (Aarik et al., 2001; Nie et al., 2015). Certainly, the selection of Ti and O2 precursors is another important factor due to the different activation kinetics for different precursors. Several works have suggested that the crystallization process is initiated at the temperature range of 165–250°C (Aarik et al., 1995, 2013; Saha et al., 2014; Chiappim et al., 2016). For example, an 11 nm thin anatase TiO2 film was obtained at deposition temperature of 225°C when TiCl4 and H2O were employed as precursors (Aarik et al., 2013). The whole set of samples was analyzed by XRD, all others were amorphous except for NALD = 150 at 300°C, and selected patterns are shown in Figure 2. Thus, we can state that all nanotube layers with lower coating thicknesses or at lower deposition temperatures remain amorphous after soaking experiments with the duration up to 28 days.</p><!><p>XRD patterns of (A) blank TiO2 nanotube layer soaked for 1, 7, 14, and 28 day(s), (B) TiO2 coated TiO2 nanotube layers with different coating thicknesses (NALD = 10, 50, 150) and different deposition temperatures (150°, 225°, and 300°C) soaked for 28 days, and (C) blank and TiO2 coated (NALD = 150 at 300°C) TiO2 nanotube layers (both without any soaking).</p><!><p>The soaked blank nanotube layers were further inspected for their morphologies. Figure 3 presents the SEM images of the nanotube layers taken at their top (water/nanotubes opening interface) and bottom (bottom of nanotubes/Ti interface). The larger inner diameter at the top than the bottom featuring a conical shape is a typical type of double-walled nanotube layers (Zazpe et al., 2016). It can be seen that the soaked nanotube layers experience a gradual change. After 1 day of soaking, the blank nanotube layer remains similar to the as-anodized nanotube layer in Figure 1a. When the soaking duration was extended to 7 and 14 days, needle-like particles were observed on the tube walls. A pronounced effect was observed after 28 days of soaking, where the tube walls were completely occupied by the nanoparticles, which is drastically different from the as-anodized nanotube layers. The coalescence of nanoparticles has resulted in rather rough outer and inner tube walls and the nanoparticles were found over the entire nanotubes from the top to the bottom of the tube layers, as visualized in Figure 3 (soakings for 28 days). This confirms that the morphological changes occurred over the entire available nanotube surface that was in contact with the water molecules. And it also confirms very good wettability of these nanotube layers. The final morphology very well-resembles the one reported in the literature (Wang et al., 2011; Liu et al., 2012). The transformation mechanism will be discussed later in this section.</p><!><p>SEM images of blank TiO2 nanotube layers soaked in water for 1, 7, 14, and 28 day(s). Scale bars in the left column are 100 nm, in the right column and respective inset are 500 nm, unless stated otherwise.</p><!><p>Similar water soaking procedures were performed on the TiO2 coated TiO2 nanotube layers, where the coatings were deposited by ALD at different deposition temperatures and different coating cycles. For these coated nanotube layers, careful inspection did not reveal any noticeable morphological change for soakings up to 14 days, as shown in Figure 4. This implies that the nanotube layers were well-protected by the additional TiO2 coatings. Thicker tube walls are observed with the increase of ALD coating cycles (and thickness), but the increase of deposition temperature does not yield any detectable morphological difference in Figure 4.</p><!><p>SEM images of atomic layer deposition (ALD) TiO2 coated TiO2 nanotube layers soaked in water for 14 days, no morphological changes can be observed for these coated TiO2 nanotube layers. All scale bars are 100 nm.</p><!><p>When the soakings were extended to 28 days, a considerable amount of needle-like nanoparticles were grown on the tube walls with low coating thickness (NALD = 10). The amount of nanoparticles gradually decreased with thicker coatings (NALD = 50, 150), as shown in Figure 5. However, in comparison to the blank nanotube layers [Figure 3 (28 days)], the amount of needles grown in the coated nanotube layers is significantly lower. The only exception lies in the nanotube layer with TiO2 coating of NALD = 150 at 300°C, which did not undergo any visible changes (i.e., no needles were grown). This is in good agreement with the XRD analysis in Figure 2B that this ALD TiO2 coating was crystalline. As anatase is thermodynamically stable, the anatase coating completely prevents reaction between water and the TiO2 nanotube wall at room temperature (Wang et al., 2011).</p><!><p>SEM images of ALD TiO2 coated TiO2 nanotube layers soaked in water for 28 days. All scale bars are 100 nm.</p><!><p>Other works on water treated TiO2 nanotube layers suggested that the formation of nanoparticles on the tube walls is closely related to the growth of anatase crystals, due to the structural rearrangement of TiO62- octahedral induced by water (Liao et al., 2011; Wang et al., 2011; Cao et al., 2016). Under extreme conditions (extensive soaking periods) it may eventually result in the transformation of hollow nanotubes to solid nanowires, or completely collapsed nanotube walls. Interestingly, the solid-state growth and dissolution-precipitation mechanism were proposed based on Yanagisawa and Ovenstone's model (Yanagisawa and Ovenstone, 1999), but it took more than 10 years to reveal this effect also for the amorphous anodic TiO2 nanotube layers (Liao et al., 2011; Wang et al., 2011).</p><p>Somewhat surprisingly, in the present work, the morphological changes of blank nanotube layers in Figure 3 are not accompanied by structural modification (amorphous to anatase) as of those reported in the literature. To understand the present phenomena, it is helpful to revisit the formation mechanism of the anodic TiO2 nanotube layer in fluoride-containing electrolytes (Macak et al., 2007; Lee et al., 2014). Briefly, the presence of F− ions enables the formation of the fluoride-complex [TiF6]2− ions, and the formation of tubular TiO2 is a competition between the solvatization of Ti4+ to [TiF6]2− and the oxide formation. However, under the absence of an electric field (driven by an applied voltage), at the oxide/water interface, the reaction turns to a self-induced oxide formation, which translates into the nucleation of needle-like nanoparticles observed in Figure 3 (7 days). Apparently, the dissolution-precipitation mechanism is now inclined to the precipitation process, and the continuous precipitation leads to the copious quantity of nanoparticles as seen in Figure 3 (28 days). Dissimilar to Wang et al. (2011) and Cao et al. (2016) where the dissolution process gradually dissolved the tube walls, in Figure 3 (28 days), distinguished walls are identified even though the nanotubes are covered by the nanoparticles. This further affirms that the process is dominated by a surface precipitation process without structural modification.</p><p>Compositional analyses were carried out on the blank amorphous and TiO2 (NALD = 150 at 300°C) coated TiO2 nanotube layers by XPS. Besides Ti and O, the survey spectra in Figure 6 reveals the presence of F, C, and N in the nanotube layers. It has been pointed out that amorphous nanotube layers contain a considerable amount of F and C species from the anodization performed in the electrolyte consists of ethylene glycol and NH4F, specifically for double-walled nanotube layers (Albu et al., 2008) that were used also in this work. In particular, the presence of F content is associated with the high-field migration of electrolyte anions and the competition between small F− ions and O2− ions migration. For the successful formation of nanotubes, the inwards migration rate of F− is twice to that of O2−, therefore accumulating a fluoride-rich inner layer especially toward the bottom of the tubes (Habazaki et al., 2007; Albu et al., 2008). In addition, an ultrathin fluoride-rich layer is also present at the outer walls, i.e., between individual tube walls, caused by the plastic-flow mechanism which pushes the nanotubes upward from the bottom of nanotubes/Ti interface during the tubes formation, hence promotes the F− along the tube walls (Berger et al., 2011). The fluoride-rich, double-walled morphology is well-documented with the support by EDX, XPS, High Resolution Transmission Electron Microscope (HR-TEM), High Angle Annular Dark Field Scanning TEM (HAADF-STEM) Auger Electron Spectroscopy (AES) and Time-of-Flight Secondary-ion Mass Spectrometry (ToF SIMS) depth-profiling measurements (Albu et al., 2008; Berger et al., 2011; So et al., 2017; Dronov et al., 2018).</p><!><p>XPS survey spectra of blank amorphous and TiO2 coated (NALD = 150 at 300°C) TiO2 nanotube layers. The table in the inset shows the atomic concentration of the elements found in the survey spectra. Spectra were offset for better clarity.</p><!><p>As a result, the as-anodized nanotube layers are usually subjected to thermal annealing at elevated temperatures for crystallization as well as for the removal of the F and C species (Albu et al., 2008). A clear double-walled morphology is visible after appropriate annealing process due to the removal of C in the form of CO2 that leads to the separation of the inner and outer walls, which have been shown in our previous works (Sopha et al., 2017b; Motola et al., 2018) and in other reports (Albu et al., 2008; So et al., 2017; Mohajernia et al., 2018). Without the annealing process, a substantial amount of C is noted in the amorphous nanotube layer as shown in Figure 6. The inner and outer walls remain intact and thus, the double-wall effect cannot be visualized in the images in Figure 1.</p><p>Comparing the blank and coated nanotube layers, the breakdown of the atomic concentration of each element tabulated in the inset of Figure 6 shows that the amount of F, C, and N is significantly reduced as a result of the ALD TiO2 coating. The C contamination can be partially assigned to the adventitious C resulting from the exposure to the air ambient. Whereas, the presence of F and N in the TiO2 coating is related to the diffusion of these two elements from the nanotube walls to the coating during the ALD process at elevated temperature (performed at 300°C). Nevertheless, the traces of F (2.0%) and N (0.7%) are almost negligible. Therefore, without sufficient F− ions on the surface of anatase TiO2 coating (NALD = 150 at 300°C) at the water/coating interface, water annealing effect was not observed even after prolonged soakings. The same conclusion was reached by Cao et al. (2016) where the presence of a higher amount of F− ions accelerated the growth of TiO2 nanoparticles.</p><p>When an amorphous TiO2 coating was added to the nanotube layer, forming nanotube/coating/water configuration, the additional coatings prepared by ALD (without F− ions) served as a protective layer (similar function as the anatase coating discussed above) to separate the tube walls and water. Due to the great adhesion of the coatings to the nanotubes, there is no direct contact between F− ions ([TiF6]2− ions) and water. Hence, we observed in Figure 4 that nanoparticles were not formed on the tube walls up to 14 days of soaking. However, at the nanotube/coating interface, the F− ions gradually attack both sides of TiO2 and the thinner coatings are more prone to F− ions transport across the entire coating, as the F− ions are very small and mobile. The thinner coatings may be eventually consumed by the F− ions, and the tube walls may be partially exposed to the water which may result in higher precipitation and growth of more nanoparticles. As the self-induced precipitation process occurs in a slow manner, an extended duration is required to observe the soaking effect. After 28 days of soaking (Figure 5), the most prominent effect (highest amount of nanoparticles) is credited to the thinnest coatings of NALD = 10, followed by NALD = 50. For these two coating thicknesses, the deposition temperatures did not have a significant effect. However, for NALD = 150, the amorphous coating deposited at 225°C has fewer nanoparticles than that at 150°C.</p><p>Further inspection of the bottoms of the TiO2 coated nanotube layers was carried out. At first, we inspected TiO2 coated (NALD = 10 at 150°C TiO2) nanotube layer, which is the lowest coating thickness deposited at the lowest temperature applied in this work, as a representative one for all coated TiO2 nanotube layers soaked for 14 days. The corresponding SEM image is shown in Figure 7a which confirms that no needles were grown for soakings up to 14 days. Note that the images shown in Figure 7 are representative image based on the extensive analyses on a broad range of nanotube samples produced by the corresponding conditions.</p><!><p>SEM images of bottom parts of the TiO2 nanotube layers taken close to the Ti substrate. (a) NALD = 10 at 150°C soaked for 14 days, NALD = 150 at (b) 150°C and (c) 225°C soaked for 28 days. All scale bars are 100 nm. The circle in (b) shows a needle found in an inner tube wall.</p><!><p>For further verification, we compared the bottoms of the TiO2 coated nanotube layers for NALD = 150 at 150° and 225°C and reached the same conclusion. Limited needles were discovered for 150°C in Figure 7b and almost no needle was detected for 225°C in Figure 7c. This is ascribed to the different densities of the TiO2 coating during the ALD process as higher deposition temperature generally promotes the interconnection between the grains (Aarik et al., 1995; Saha et al., 2014). Thus, for identical thicknesses of NALD = 150, the film density is higher at 225°C and it can better resist the attack of F− ions.</p><p>Similar morphology of TiO2 nanoparticles coated TiO2 nanotube layer in Figure 3 (28 days) is observed for the renowned "TiCl4 treatment" often carried out to decorate the TiO2 nanotube layer by additional TiO2 nanoparticles for DSSCs (Meen et al., 2012; So et al., 2015). Likewise, the as-deposited TiO2 nanoparticles produced via hydrolysis of TiCl4 are amorphous and conventional thermal annealing is required to crystallize the nanoparticles. A major difference between TiCl4 treatment and water soaking is that the growth rate of TiO2 nanoparticles is much slower in the present case. We have presented that the nanotube layer was completely decorated by nanoparticles after 30 min of treatment in a TiCl4 bath (Sopha et al., 2017b), in line with other works (Meen et al., 2012; So et al., 2015). This is ascribed to the very reactive TiCl4 precursor and considerably high reaction (chemical bath) temperature at 70°C, which accelerate the growth process. As for water soaking, the precipitation is a self-induced process and much longer duration is required to accumulate a comparable quantity of nanoparticle deposits. It has also been confirmed that higher soaking temperature and longer reaction time promote the growth rate of TiO2 nanoparticles in a water bath (Krengvirat et al., 2013; Cao et al., 2016).</p><p>Altogether, these results indicate that the thin TiO2 coatings act as a protective layer to maintain the smooth tubular morphology of the as-anodized nanotube layers in the amorphous state for more than 14 days, while the unprotected nanotube layers hardly sustain 7 days of soaking. This generously increases more than twice of the initial lifespan of the smooth amorphous TiO2 nanotube layers which offers a stable platform for cell culturing and drug delivery testing typically carried out in the time scale from 3 to 20 days (Peng et al., 2009; Hu et al., 2012; Feng et al., 2016; Kaur et al., 2016). Moreover, it should be emphasized that a smooth morphology is usually favored for cell culturing, as the cell spreading and the cell survival rate is influenced by the morphology of TiO2 supporting layer (Park et al., 2007; Peng et al., 2009; Tian et al., 2015). In addition to the water soaking experiments, we also performed soakings in PBS with identical conditions (temperature, duration) for all blank and ALD TiO2 coated TiO2 nanotube layers. All these nanotube layers have revealed extreme stability in PBS. As shown in Supplementary Figures 1 and 2, no structural and morphological changes were observed even after 28 days due to the disruption of the precipitation kinetics by the inorganic species in the buffer solution. Although the full mechanism is not well-understood yet, this observation is in accord with Cao et al. (2016).</p><p>Overall, we recommend performing 150 ALD cycles of TiO2 coating, equivalent to 8.4 nm thicknesses at 225°C which is sufficiently thick for effective protection for the nanotube layers whilst keeping the amorphous state. Among all the amorphous coatings after 28 days of soaking, this condition has the fewest nanoparticles on the nanotube walls, evidenced in Figures 5 and 7.</p><!><p>We proposed the utilization of thin ALD TiO2 coatings to protect 1D TiO2 nanotube layers against morphological changes within prolonged water soaking experiments. Thin and conformal TiO2 coating of NALD = 10, 50, and 150 corresponding to 0.56, 2.8, and 8.4 nm in thickness, respectively, were deposited by ALD at temperatures 150°, 225°, and 300°C within 5 μm amorphous TiO2 nanotube layers, which yielded amorphous and anatase coatings. The uncoated nanotube layers underwent significant morphological changes with additional nanoparticles formed on the nanotube walls after extensive soakings up to 28 days. The formation of the nanoparticles was related to the reaction between residual F− ions (present in the nanotube walls) and water in a self-induced precipitation mechanism. The additional TiO2 coatings delayed the soaking effect and preserved the nanotube walls for a minimum of 14 days. Overall, the optimum coating was credited to NALD = 150 (8.4 nm) deposited at 225°C. The combination of identical materials by different preparation techniques sustains the amorphous state and tubular morphology of 1D TiO2 nanotube layers for biomedical applications as an example.</p><!><p>SN carried out the soakings and wrote the manuscript. HS synthesized the nanotubes and helped with the soakings. RZ and JP deposited the coatings. LH and VB carried out the SEM measurements. ZS measured and analyzed the XRD results. FD measured and evaluated the XPS results. JM designed the experiments, advised the results, corrected the manuscript and obtained the funding. All authors discussed, 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. We gratefully acknowledge support from the European Research Council (project No. 638857) and the Ministry of Education, Youth and Sports of the Czech Republic (projects No. LM2015082, LM2015041, LQ1601). Part of the work was carried out with the support of CEITEC Nano Research Infrastructure (MEYS CR, 2016-2019).</p><!><p>The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2019.00038/full#supplementary-material</p><!><p>Click here for additional data file.</p>

PubMed Open Access

Identification of the Biosynthetic Gene Cluster of Thermoactinoamides and Discovery of New Congeners by Integrated Genome Mining and MS-Based Molecular Networking

The putative non-ribosomal peptide synthetase (NRPS) gene cluster encoding the biosynthesis of the bioactive cyclohexapeptide thermoactinoamide A (1) was identified in Thermoactinomyces vulgaris DSM 43016. Based on an in silico prediction, the biosynthetic operon was shown to contain two trimodular NRPSs, designated as ThdA and ThdB, respectively. Chemical analysis of a bacterial crude extract showed the presence of thermoactinoamide A (1), thereby supporting this biosynthetic hypothesis. Notably, integrating genome mining with a LC-HRMS/MS molecular networking-based investigation of the microbial metabolome, we succeeded in the identification of 10 structural variants (2–11) of thermoactinoamide A (1), five of them being new compounds (thermoactinoamides G-K, 7–11). As only one thermoactinoamide operon was found in T. vulgaris, it can be assumed that all thermoactinoamide congeners are assembled by the same multimodular NRPS system. In light of these findings, we suggest that the thermoactinoamide synthetase is able to create chemical diversity, combining the relaxed substrate selectivity of some adenylation domains with the iterative and/or alternative use of specific modules. In the frame of our screening program to discover antitumor natural products, thermoactinoamide A (1) was shown to exert a moderate growth-inhibitory effect in BxPC-3 cancer cells in the low micromolar range, while being inactive in PANC-1 and 3AB-OS solid tumor models.

identification_of_the_biosynthetic_gene_cluster_of_thermoactinoamides_and_discovery_of_new_congeners

Introduction<!>Bioinformatic Analysis<!>Data-Dependent LC-HRMS/MS Analysis<!>MZmine Processing and Molecular Networking<!>Isolation and Structural Determination of Thermoactinoamide D and E<!>Cell Viability Assays<!>Identification of Thermoactinoamide Biosynthetic Gene Cluster<!><!>Identification of Thermoactinoamide Biosynthetic Gene Cluster<!><!>Identification of Thermoactinoamide Biosynthetic Gene Cluster<!>Structure and Biosynthesis of Thermoactinoamide A Congeners<!><!>Structure and Biosynthesis of Thermoactinoamide A Congeners<!><!>Structure and Biosynthesis of Thermoactinoamide A Congeners<!><!>Structure and Biosynthesis of Thermoactinoamide A Congeners<!><!>Structure and Biosynthesis of Thermoactinoamide A Congeners<!>Evaluation of Antiproliferative Activity of Thermoactinoamide A (1)<!><!>Discussion<!>Data Availability Statement<!>Author Contributions<!>Conflict of Interest<!><!>Supplementary Material<!>

<p>Microbial genome mining has represented a revolutionary approach to search for novel secondary metabolites as drug leads. The ground-breaking idea behind this strategy is deciphering genetic information to depict the chemical structure of a natural product, without having it to hand.</p><p>Genome-directed discovery of natural products is pursued through the identification of biosynthetic gene clusters, displaying homology with genes involved in the production of known secondary metabolites. Indeed, although secondary metabolites cover a wide and heterogenous chemical space, the biosynthetic routes for several classes of compounds, such as non-ribosomal peptides and polyketides, are outstandingly conserved across microbial species: at the molecular level, this translates into high sequence similarity of many core biosynthetic enzymes (Ziemert et al., 2016).</p><p>Since the discovery of novel secondary metabolites by David Hopwood and coworkers through the whole genome sequencing of Streptomyces coelicolor (Bentley et al., 2002), genome mining has made great strides and has become a well-established strategy to access metabolomes of macro- and microorganisms (Della Sala et al., 2013; Wilson and Piel, 2013). Indeed, if natural product discovery is usually driven by bioactivity and/or dereplication techniques, the advent of genome mining has accelerated the chemical workflow, as prediction of secondary metabolite biosynthetic pathways may provide a rationale for targeted isolation of natural products from complex crude extracts (Bachmann et al., 2014). In addition, the screening of bacterial genomes unveiled a huge, unexplored biosynthetic potential, revealing that the number of compounds detected through analytical chemistry approaches is lagging well-behind the putative chemical entities identified by genome mining (Challis, 2008). Taken together, these observations explain why genome mining has contributed to the renaissance of natural product chemistry, giving renewed energy to drug discovery from natural sources.</p><p>With the advent of high-quality throughput sequencing methods and the drop of costs, an ever-expanding amount of bacterial genome sequences have become publicly available. The current challenge is handling and mining these sequence data, to detect biosynthetic gene clusters and connect them with the chemical structures of known secondary metabolites or putative novel chemical entities to be isolated from the microbial producer. The emergent need to screen and manipulate a huge amount of sequence data has led to the development of powerful and efficient in silico platforms, such as antiSMASH 4.0 (Blin et al., 2017) and PRISM 3 (Skinnider et al., 2017), which are specifically addressed to genome-wide detection of biosynthetic gene clusters and structural prediction of the relevant secondary metabolites.</p><p>In the frame of our ongoing research program focused on the isolation of antibacterial and antitumor secondary metabolites from marine organisms and microorganisms (Teta et al., 2012, 2019; Schirmeister et al., 2017), the lipophilic cyclopeptide thermoactinoamide A (1) was recently isolated from the Icelandic thermophilic bacterium Thermoactinomyces vulgaris and shown to exert antibacterial activity against Staphylococcus aureus ATCC 6538 with a MIC value of 35 μM (Teta et al., 2017). Dissection of the chemical structure of thermoactinoamide A suggests that it is assembled by non-ribosomal peptide synthetase (NRPS) machinery, which is based on modular mega-synthetases displaying a standard set of catalytic domains, namely condensation (C), adenylation (A), and thiolation domains (PCP, also known as peptidyl carrier protein). The acknowledged collinearity between the modular architecture of NRPS with their encoded metabolites and the high homology within the key biosynthetic NRPS domains, together with the chance to predict in silico substrate selectivity of adenylation domains, has made thermoactinoamide A (1) an optimal candidate for a genome mining approach.</p><p>Herein, we report new insights into the biosynthesis of the bioactive cyclic peptide thermoactinoamide A (1). In light of our observations about the predicted assembly line of 1, we retrieved from GenBank the whole genome of a Thermoactinomyces vulgaris strain (Thermoactinomyces vulgaris DSM 43016) and performed a bioinformatic analysis leading to the identification of a putative operon encoding the biosynthetic pathway of thermoactinoamide A. To support this hypothesis, chemical analysis of the organic extract from Thermoactinomyces vulgaris DSM 43016 revealed this microbial strain to produce thermoactinoamide A. Notably, combining biosynthetic gene cluster analysis with a mass spectrometry- and molecular networking- based investigation of the microbial metabolome, we succeeded in the identification of 10 structural variants of 1 from Thermoactinomyces vulgaris DSM 43016, five of them being new compounds (thermoactinoamides G-K). In addition, in vitro evaluation of antiproliferative activity of 1 against three different cancer cell lines is discussed.</p><!><p>To compare bacterial genomes, average nucleotide identity (ANI) calculations were performed by using the Pairwise ANI tool (Chen et al., 2019), available at the DOE Joint Genome Institute website1 Secondary metabolite biosynthetic pathways were identified by genome mining software programs, namely antiSMASH 4.0 (Blin et al., 2017) and PRISM 3 (Skinnider et al., 2017). For the antiSMASH search, the antiSMASH bacterial version was used, setting detection strictness to "relaxed" and selecting the following extra features: KnownClusterBlast, ClusterBlast, SubClusterBlast, ActiveSiteFinder, Cluster Pfam analysis, and Pfam-based GO term annotation. For PRISM analysis, parameters were set as follows: (a) structure limit was set to 50; (b) window was set to 10,000; (c) as optional searches, searches for all families of biosynthetic domains were enabled; and d) for open reading frame prediction, all methods were selected. Substrate selectivity of NRPS adenylation domains was predicted by NRPSpredictor2 (Röttig et al., 2011), while identification and classification of condensation (C) and epimerization (E) domains were accomplished by NaPDoS analyses (Ziemert et al., 2012). The thermoactinoamide (thd) gene cluster from T. vulgaris DSM 43016 was split on two distinct contigs (contig Ga0070019_105 - accession: REFP01000011; contig Ga0070019_114 - accession: REFP01000005). These two contigs were re-assembled using as a template the contig containing the intact thd gene cluster from Thermoactinomyces AS95 (contig NODE_4; accession: LSVF01000006). Contigs from T. vulgaris DSM 43016 were aligned and re-assembled using the blastn suite (Altschul et al., 1990) and the sequence analysis software SeqMan (DNASTAR v.5.00). Reference-guided assembly led to the integration of contig Ga0070019_114 (accession: REFP01000005) into contig Ga0070019_105 (accession: REFP01000011) (Figure S5). The de novo assembled contig sequence is available as Supplementary Material (denovo_assembled_contig.fasta).</p><!><p>A 500-mL culture of Thermoactinomyces vulgaris DSM 43016 (DSMZ)2 was grown for 24 h at 50°C in CYC-medium (Czapek Dox agar 48.0 g/L, yeast extract 2.0 g/L, casamino acids 6.1 g/L, tryptophan 0.02 g/L, sterile Milli-Q H2O), and thereafter lyophilized after freezing. Then, the freeze-dried culture was rehydrated with 4 mL of distilled water and sonicated for 5 min. The suspension was extracted with a mixture of MeOH/CHCl3 (2:1, 6 mL), paper filtered, and dried to afford 20.6 mg of crude extract, according to a previously reported procedure yielding all low molecular weight metabolites (Costantino et al., 2012). The extract was resuspended in MeOH at a concentration of 10 mg/mL for LC-HRMS/MS (Liquid Chromatography - High Resolution Tandem Mass Spectrometry) analyses. Experiments were performed using a Thermo LTQ Orbitrap XL high-resolution ESI mass spectrometer coupled to a Thermo U3000 HPLC system, which included a solvent reservoir, in-line degasser, binary pump, and refrigerated autosampler. A 5-μm Kinetex C18 column (50 × 2.10 mm), maintained at room temperature, was eluted at 200 μL·min−1 with H2O (supplemented with 0.1% HCOOH) and CH3CN, using a gradient elution. The gradient program was as follows: 30% CH3CN 5 min, 30%−99% CH3CN over 30 min, 100% CH3CN 3 min. Mass spectra were acquired in positive ion detection mode. MS parameters were a spray voltage of 4.8 kV, a capillary temperature of 285°C, a sheath gas rate of 32 units N2 (ca. 150 mL/min), and an auxiliary gas rate of 15 units N2 (ca. 50 mL/min). Data were collected in the data-dependent acquisition mode, in which the three most intense ions of a full-scan mass spectrum were subjected to high resolution tandem mass spectrometry (HRMS/MS) analysis. The m/z range for data dependent acquisition was set between 610 and 800 amu. HRMS/MS scans were obtained for selected ions with CID fragmentation, an isolation width of 2.0, normalized collision energy of 35, Activation Q of 0.250, and an activation time of 30 ms.</p><!><p>Processing of LC-HRMS/MS data was performed using a method similar to the feature-based molecular network described by Nothias et al. (2019). Raw files were directly imported into MZmine 2.53 (Pluskal et al., 2010). The mass detection was performed on raw data and exact masses with mass level 1 and centroided masses with mass level 2, by keeping the noise level at 1,000. Chromatograms were built using an ADAP module (Myers et al., 2017) with a minimum height of 1,000, and m/z tolerance of 0.05 (or 20 ppm). For the chromatogram deconvolution, the local minimum search algorithm was used with the following settings: chromatographic threshold = 5%, minimum retention time range = 0.50 min, minimum relative height = 30%, minimum absolute height = 10,000, minimum ratio of the peak top/edge = 1.3, and peak duration range = 0.0–6.0 min.</p><p>Peak alignment was performed using the Join aligner algorithm (m/z tolerance at 0.05 (or 20 ppm), absolute RT tolerance at 0.6 min). [M+Na–H], [M+K–H], [M+Mg−2H], [M+NH3], [M-Na+NH4], [M+1, 13C] adducts were filtered out by setting the maximum relative height at 100%. Peaks without associated MS/MS spectrum were finally filtered out from the peak list. Clustered data were then exported to.mgf file for GNPS, while chromatographic data including retention times, peak areas, and peak heights were exported to a .csv file. The mass spectrometry data were deposited on MassIVE (accession number: MSV000085201).</p><p>A molecular network (Wang et al., 2016) was generated on GNPS' online platform,3 using the Metabolomics workflow with the following parameters: the parent mass tolerance and MS/MS fragment ion tolerance were set at 0.02 Da and 0.2 Da, respectively, the cosine score at above 0.71, and matched peaks above 6. Spectra were retained only if the nodes appeared in each other's respective top 10 most similar nodes. The spectra in the network were then searched against GNPS spectral libraries using a cosine score above 0.7 and at least 7 matched peaks. Once the molecular network (https://gnps.ucsd.edu/ProteoSAFe/status.jsp?task=a365d81e928f4d0dbc5c76b0dff4515b) was generated, chromatographic data in the .csv file were mapped to the relevant nodes using Cytoscape 3.7.2 (Shannon et al., 2003), which was also used for network visualization and analysis. The fragmentation spectra of Thermoactinoamide A–K were deposited to GNPS library (CCMSLIB0000572020, CCMSLIB00005720245 and CCMSLIB00005720251–CCMSLIB00005720259).</p><!><p>The crude extract of Thermoactinomyces vulgaris DSM 43016 was subjected to reversed-phase HPLC using a 10 μm Kinetex C18 column (250 × 10 mm) [eluent A: 0.1% HCOOH in H2O; eluent B: MeOH; gradient program: 60% B 5 min, 60% → 100% B over 17 min, 100% B 13 min; flow rate 5 mL min−1, wavelength 280 nm], thus obtaining pure thermoactinoamide D (4), (0.063 mg, tR 15.5 min), that was subjected to NMR experiments and Marfey's analyses.</p><p>The fraction eluted at tR 9.5 min containing thermoactinoamide E (5) was further purified by reversed-phase HPLC using a 5 μm Kinetex C18 column (250 × 4.6 mm) using the same elution method [eluent A: 0.1% HCOOH in H2O; eluent B: MeOH; gradient program: 60% B 5 min, 60% → 100% B over 17 min, 100% B 13 min; flow rate 1 mL min−1, wavelength 280 nm]. The amount of thermoactinoamide E was not sufficient to perform NMR experiments.</p><p>Both the compounds were hydrolyzed with 6 N HCl/AcOH (1:1) at 120°C for 12 h. The residual HCl fumes were removed under an N2 stream. The hydrolysate of 4 was then dissolved in TEA/acetone (2:3, 100 μL) and the solution was treated with 100 μL of 1% 1-fluoro-2,4-dinitrophenyl-5-d-alaninamide (d-FDAA) in CH3CN/acetone (1:2) (Marfey, 1984). The vial was heated at 50°C for 2 h. The mixture was dried, and the resulting d-FDAA derivatives of the free amino acids were redissolved in MeOH (200 μL) for subsequent analysis. The hydrolysate of 5 was instead treated with 1-fluoro-2,4-dinitrophenyl-5-l-alaninamide (l-FDAA) in the same conditions as above.</p><p>Authentic standards of l-Tyr, l-Val, l-Leu, l-Ile, and d-allo-Ile were treated with l-FDAA and d-FDAA as described above and yielded the l-FDAA and d-FDAA standards. Marfey's derivatives of 4 and 5 were analyzed by HPLC-ESI-HRMS, and their retention times were compared with those from the authentic standard derivatives. A 2.6 μm Kinetex PFP column (100 × 4.6 mm) maintained at 25°C was used for compound 4. The column was eluted at 200 μL min−1 with 0.1% HCOOH in H2O and MeOH. The gradient program was as follows: 60% MeOH 5 min, 60% → 100% MeOH over 30 min, 100% MeOH 15 min. A 5-μm Kinetex C18 column (50 × 2.10 mm), maintained at room temperature, was instead used for the analysis of Marfey's derivatives of 5. The column was eluted at 200 μL·min−1 with H2O (supplemented with 0.1% HCOOH) and MeOH, using a gradient elution. The gradient program was as follows: 5% MeOH 3 min, 5–70% MeOH over 30 min, 70–100% over 1 min, and MeOH 100% 6 min. Mass spectra were acquired in positive ion detection mode, and the data were analyzed using the Xcalibur suite of programs. NMR experiments were performed on Varian Unity Inova spectrometers (Agilent Technology - Cernusco sul Naviglio, Italy) at 700 MHz in CD3OD; chemical shifts were referenced to the residual solvent signal (CD3OD: δH 3.31, δC 49.00). All 13C chemical shift were assigned using the 2D spectra, therefore, mono-dimensional 13C NMR spectra were not recorded (see Table 4). For an accurate measurement of the coupling constants, the one-dimensional 1H NMR spectra were transformed at 64 K points (digital resolution: 0.09 Hz). The HSQC spectra were optimized for 1JCH = 142 Hz and the HMBC experiments for 2, 3JCH = 8.3 Hz.</p><!><p>In order to evaluate the growth-inhibitory effects of thermoactinoamide A (1), the xCELLigence System Real-Time Cell Analyzer (ACEA Biosciences, San Diego, CA, USA) was used for real-time monitoring of cancer cell proliferation after drug exposure, as previously described (Caso et al., 2019).</p><p>Data from biological assay represent the mean (± standard deviation, SD) of three independent experiments. Two-group comparisons were performed using Student's t-test. p < 0.05 were considered to be statistically significant. Statistical analysis was performed using the GraphPad Prism Software Version 5 (GraphPad Software Inc., San Diego, CA, USA).</p><!><p>The genome of Thermoactinomyces vulgaris strain DSM 43016 was retrieved from GenBank (Accession: NZ_REFP00000000). The published genome from this strain consists of 15 contigs, with an overall size of 2.56 Mb and GC content of 47.9%.</p><p>Interestingly, using ANI (Average Nucleotide Identity) calculations, we discovered three further bacterial genomes that definitely affiliate with Thermoactinomyces vulgaris strain 43016 (ANI value ≥ 99.56%), although were deposited in the GenBank database as belonging to three different type strains, namely Thermoactinomyces AS95 (Accession: LSVF00000000), Thermoactinomyces sp. Gus2-1 (Accession: JPZM00000000), and Thermoactinomyces sp. CDF (Accession: LFJU00000000) (Table S1). Therefore, these bacterial genomes can be assigned to the same species as the ANI value is higher than the cut-off score (95%) currently accepted for taxonomic inference (Figueras et al., 2014).</p><p>The focus of our study was to detect the biosynthetic gene cluster encoding the thermoactinoamide family of compounds. d-Amino acids containing cyclic peptides (such as thermoactinoamides) are either synthesized through the classical NRPS assembly line or produced via post-translational epimerization and cyclization of a ribosomally synthesized precursor peptide (RiPP pathway). Therefore, a preliminary screening was performed to mine the draft genome of T. vulgaris DSM 43016 for secondary metabolite biosynthetic pathways, using the publicly available bioinformatic tools, Antismash (Blin et al., 2017) and PRISM (Skinnider et al., 2017).</p><p>The genome of T. vulgaris DSM 43016 contains 5 putative biosynthetic pathways: a type-III-polyketide synthase gene (contig Ga0070019_104 - accession: REFP01000010); a small NRPS system, consisting of A, PCP, and E proteins, which are encoded by separate genes (contig Ga0070019_101 - accession: REFP01000007); a multi-modular NRPS gene cluster, split between two contigs (contig Ga0070019_105 - accession: REFP01000011; contig Ga0070019_114 - accession: REFP01000005); a lassopeptide; and a siderophore gene clusters (contig Ga0070019_101; accession: REFP01000007). As expected, genome mining of Thermoactinomyces AS95, Gus2-1, and CDF revealed the presence of the same secondary metabolite biosynthetic gene clusters as in T. vulgaris DSM 43016.</p><p>Owing to the collinearity between the NRPS domain architecture and the thermoactinoamide chemical structures, the biosynthesis of thermoactinoamides was predicted to be directed by the multi-modular NRPS pathway, which was designated as "thd." Prior to bioinformatic analysis, we manually assembled the two distinct contigs the NRPS gene cluster was split on, using the contig from Thermoactinomyces AS95 (contig NODE_4; accession: LSVF01000006) containing the intact thd gene cluster as a template. Putative genes identified in the de novo assembled contig from T. vulgaris DSM 43016 are summarized in Table 1.</p><!><p>Putative genes identified on the de novo assembled contig containing the thermoactinoamide gene cluster (highlighted in green and bolded) from T. vulgaris DSM 43016.</p><p>C*, truncated condensation domain, probably inactive.</p><!><p>The putative thd gene cluster is 22,223 bp long and encodes two NRPSs, ThdA and ThdB, each comprising of three modules. As shown in Figure 1, biosynthesis of thermoactinoamide A starts with the loading module of ThdA, which activates an l-Leu residue by adenylation. Selectivity for leucine was predicted based on the specificity-conferring code of A domains proposed by Stachelhaus et al. (1999); in addition, an NRPSpredictor2 search (currently incorporated in Antismash) also predicted leucine specificity based both on the sequence of the A domain and on structural features of the active site (Röttig et al., 2011) (Table 2). After activation and its ligation to the first PCP of ThdA, the l-Leu residue is forwarded to the first extending module of ThdA, displaying a LCL-A-PCP-E domain organization and catalyzing condensation between the l-Leu residue and one l-Ile. The A domain showed an amino acid-specific structural motif selective for isoleucine (Table 2). NaPDoS analyses (Ziemert et al., 2012) as well as detection of the down-seq signatures reported by Caradec et al. (2014), supported identification within this extending module of one LCL (condensation of two l amino acids) domain and one E (epimerization) domain (Table S2; Figures S1, S2). The presence of an epimerization domain suggests conversion of the l-Ile residue to the corresponding d-allo-Ile, which is in agreement with the absolute configuration of this amino acid in thermoactinoamide A (Teta et al., 2017).</p><!><p>Putative biosynthesis of thermoactinoamide A (abbreviations: C, condensation domain; A, adenylation; P, peptidyl-carrier protein; E, epimerase; C*, truncated condensation domain, probably inactive).</p><p>Signature motifs of adenylation domains from the Thd synthetase.</p><p>as reported by Teta et al. (2015).</p><!><p>The last NRPS elongation module of ThdA extends the dipeptidyl intermediate l-Leu-d-allo-Ile with a tyrosine residue, because the ThdA_A3 domain was found to be specific for tyrosine according to NRPS predictor2 (Stachelhaus code match: 90%) (Table 2). As expected, this module starts with a condensation domain featuring signatures of DCL domains, to form a peptide bond between d-allo-Ile and l-Tyr. Then, a second epimerase within the last module of ThdA is consistent with occurrence of an l-tyrosine in thermoactinoamide A (Table S2; Figures S1, S2).</p><p>The growing chain is then transferred to the NRPS enzyme ThdB, which includes three modules each adding one amino acid, namely valine, leucine, and leucine. The first adenylation domain (ThdB_A4) of ThdB was predicted to activate either valine or isoleucine, thus accounting for a relaxed substrate selectivity (Table 2). Residues lining the binding pocket of ThdB_A4 are essentially hydrophobic, in accordance with the lipophilicity of the substrate sidechain of valine or isoleucine residues. Considering that (a) the specificity conferring code partially matches both with Ile and Val activating enzymes and (b) thermoactinoamide variants include either Val or Ile at that position, it can be argued that the A-domain binding pocket allows accommodation of both amino acids.</p><p>The first condensation domain of ThdB was classified as a DCL domain in an NaPDoS search and is involved in the amide bond formation between d-Tyr and l-Val (or Ile) (Table S2). Then, the last two modules of ThdB extend the peptide growing chain by the sequential addition of two l-leucine residues. Indeed, ThdB_A5 and ThdB_A6 share a 100% sequence identity and were predicted as specific to leucine (Table 2). The relevant condensation domains (ThdB_C5 and ThdB_C6) within the last two modules feature conserved LCL motifs as expected. Isomerization of the terminal l-leucine to d-leucine is in charge of ThdB_E3 epimerase (Table S2; Figures S1, S2).</p><p>Unexpectedly, the thd gene cluster lacks a typical thioesterase/cyclase, which is usually required for hydrolytic release and head-to-tail cyclization of the linear peptide (Kopp and Marahiel, 2007). Moreover, neither a trans-acting cyclase, such as SurE in surugamide biosynthesis (Kuranaga et al., 2018), nor a terminal condensation domain (CT), such as those responsible for cyclization in fungal NRPSs (Gao et al., 2012), could be detected within the thermoactinoamide NRPS. Despite the lack of usual chain termination domain, the observation that cyclic peptides are produced by T. vulgaris DSM 43016 implies the existence of unique off-loading and macrolactamization mechanisms. Intriguingly, the NRPS involved in the biosynthesis of acetylaszonalenin, namely AnaPS, lacks a canonical thioesterase/cyclase and terminates with an epimerization (E) domain, which has been proposed to promote a specific conformation of the linear intermediate to undergo cyclization (Gao et al., 2012). Similarly, ThdB_E epimerase could position the N-terminal Leu amine group next to the C-terminal Leu thioester carbonyl, thereby allowing the head-to-tail macrocyclization to take place. Moreover, considering that epimerases and condensation domains share both sequence and structure features (Keating et al., 2002), it could be hypothesized that ThdB E catalyzes the last condensation step, thus resulting in the hydrolytic cleavage and macrolactamization of the final thermoactinoamide.</p><!><p>The actual production of thermoactinoamide A and its congeners by T. vulgaris DSM 43016 was confirmed by chemical analysis. The strain was extracted with a MeOH/CHCl3 mixture and the crude extract was subjected to liquid chromatography high-resolution tandem mass spectrometry (LC-HRMS/MS) on an LTQ Orbitrap instrument. LC-MS data were processed using an implementation of feature-based molecular networking (Nothias et al., 2019). The raw LC-MS data were pre-processed using the MZmine program (Pluskal et al., 2010) and the .mgf MS2 data file generated by MZmine was submitted to the online platform at the Global Natural Products Social Molecular Networking website3. Mapping of chromatographic information exported from MZmine and visualization of the network were performed using the Cytoscape program (Shannon et al., 2003).</p><p>The thermoactinoamide cluster (Figure 2) contains eleven nodes, indicating the presence of ten compounds closely related to thermoactinoamide A (1, m/z 715.48), and is represented so that the area of the nodes is proportional to the amounts of the relevant compounds. This makes it immediately clear that all the congeners of thermoactinoamide A are far less abundant than the parent compound. Because no other related NRPS clusters were present in the genome of T. vulgaris DSM 43016, all these minor thermoactinoamide variants (Table 3) must be produced by the same multimodular NRPS gene cluster that synthesizes thermoactinoamide A.</p><!><p>Molecular cluster of thermoactinoamides. Nodes for thermoactinoamides containing one aromatic residue are labeled in purple, while thermoactinoamides containing two and no aromatic residue are marked in blue and orange, respectively. Node size is relative to the areas of the relevant peaks in the extracted-ion chromatograms from the LC-MS run and edge thickness is relative to the cosine score similarity.</p><p>Thermoactinoamide variants from T. vulgaris DSM 43016.</p><p>Color code relates to the presence of one (purple), two (blue), or no one (orange) aromatic aminoacid residue in the structure.</p><p>All the thermoactinoamides MS/MS spectra have been annotated on GNPS.</p><!><p>The structure of thermoactinoamide A (1) was confirmed by comparison of its HPLC retention time, MS/MS spectrum, and 1H NMR spectrum with those reported (Teta et al., 2017). The high-resolution MS/MS fragmentation pattern of 1 was taken as a model to study in detail all the thermoactinoamide congeners (Figure 3). As observed for most cyclic peptides (Ngoka and Gross, 1999), the typical fragmentation mode of thermoactinoamides consists in the cleavage of two amide bonds with the loss of one, two, or three amino acid residues from the parent ion (called ions α, β, and γ, respectively, in the following discussion). These fragment ions provide information on the amino acid sequence in the peptides and in the previous study (Teta et al., 2017) identified correctly the amino acid sequence of compound 1, except for discrimination of isomeric Leu and Ile and all the stereochemical aspects. As a side remark, elucidation of the latter structural aspects required an extensive spectroscopic and chemical analysis, while a previous knowledge of the biosynthetic cluster of thermoactinoamide could have provided the same information immediately.</p><!><p>High-resolution MS/MS fragmentations of new thermoactinoamide variants. Thermoactinoamides A (1), G (7), H (8) containing one aromatic residue are labeled in purple, while thermoactinoamides containing two (9) and no aromatic residue (10 and 11) are marked in blue and orange, respectively. Fragmentation scheme is depicted for thermoactinoamides A (1), I (9), and J (10), as representative of each class of compounds. Fragment α, β, and γ are the losses of one, two, and three amino acids from the parent ions.</p><!><p>Monoisotopic masses, retention times, and fragmentation patterns of peptides 2–6 identified them as the thermoactinoamides B-F reported in Teta et al. (2017). In that study, compounds 2–6 were not isolated and their structures were putatively assigned based on their MS/MS fragmentations and of the assumption that they differed from thermoactinoamides A (1) only in a single amino acid. Therefore, structures 2–6 were re-examined in the light of the newly acquired knowledge of their biosynthetic cluster.</p><p>Besides fitting MS/MS fragmentation data, structures of thermoactinoamide B (2), C (3), and F (6) fit well with the biosynthetic cluster. They show, respectively, an Ile → Val substitution at aa-2, a Val → Ile substitution at aa-4, and a Tyr → Phe substitution at aa-3. These substitutions are among the most common for minor congeners of NRPS products. Putative isoleucine-activating adenylation domains of NRPSs have been shown to also select valine to a similar extent, and conversely valine-activating enzymes are able to recruit isoleucine as well (Challis et al., 2000). Furthermore, based on the aromatic substrate selectivity of ThdA_A3, it is not surprising that a Phe residue may replace a Tyr residue in thermoactinoamide F (6). In contrast, the reported putative structures of thermoactinoamide D (4) and E (5) did not obviously match up with the modular architecture of the thd gene cluster. Therefore, these two congeners were isolated to determine their structure by spectroscopic and chemical means.</p><p>Thermoactinoamide D (4) displays two Tyr residues, with the additional Tyr located on aa-1, a position that in the other variants is held by a strikingly conserved l-Leu. This would imply an extended substrate specificity to Tyr of the Leu-specific ThdA_A1 loading module. Even though the location of the additional Tyr residue could be unequivocally deduced by MS/MS data, this unexpected substitution prompted us to perform a full 2D NMR characterization of thermoactinoamide D (4), whose results are summarized in Table 4. Consistently with MS data, the HMBC correlations of the l-Tyr CO carbon atoms with the d-allo-Ile H-2 proton and of the d-allo-Ile CO carbon atoms with the d-Tyr H-2 proton indicated that the d-allo-Ile residue lies between the two Tyr residues. Marfey analysis (Figure 4) revealed the configuration of the two Tyr residues of 4 to be l and d, respectively, and confirmed the identity and the configuration of the other amino acids. Finally, the l-Tyr residue was assigned to position 1 according to gene-based stereochemistry prediction. Therefore, the hypothesized structure of thermoactinoamide D (4) was fully confirmed by NMR and Marfey analysis, and the relaxed substrate selectivity of ThdA_A1 therefore confirmed. Indeed, multi-specificity of adenylation domains is not unusual and has been extensively documented in the literature (Meyer et al., 2016). It is worthy to note that the N-terminal truncated *C domain could influence substrate selectivity of the downstream A domain within the starter module of ThdA. It has been demonstrated that C-terminal subdomains at the interface with the A domains exert a gatekeeping function and modulate substrate selection by the A domain (Meyer et al., 2016).</p><!><p>NMR data of thermoactinoamide D (4) (700 MHz, CD3OD).</p><p>Selected HMBC correlations from proton stated to the indicated carbon.</p><p>Overlapped.</p><p>Advanced Marfey's analysis of compound 4 using a pentafluorophenyl (PFP) bonded-phase column. (A) Extracted-ion chromatograms at m/z 384.1514 of the d-FDAA derivatives from the hydrolysis of 4 and of d- and l-FDAA derivatives l-Leu, l-Ile, and d-allo-Ile. (B) Extracted-ion chromatograms at m/z 434.1306 of the d-FDAA derivatives from the hydrolysis of 4 and of d- and l-FDAA derivatives l-Tyr. (C) Extracted-ion chromatograms at m/z 370.1357 of the d-FDAA derivatives from the hydrolysis of 4 and of d- and l-FDAA derivatives l-Val.</p><!><p>The isolated amounts of thermoactinoamide E (5) were too small for 2D NMR-based structure elucidation, but advanced (MS-aided) Marfey analysis could still be performed and allowed for distinction between Leu, Ile, and allo-Ile residues and between their respective d and l enantiomers. This way, thermoactinoamide E (5) was shown to contain 1 × l-Ile, 2 × l-Leu, 1 × d-allo-Ile, 1 × d-Leu, and 1 × l-Val, an amino acid composition that did not agree with the reported putative structure (Figure 5). More interestingly, the determined amino acid composition is different from that of all the other congeners (1–4 and 6) discussed so far, in that only two d amino acids are present in the sequence. Thermoactinoamide E (5) contains only aliphatic amino acids and could derive from cyclo-dimerization of the tripeptidyl unit l-Ile/Val-l-Leu-d-Leu/allo-Ile. Therefore, it can be argued that it is synthesized by the sole ThdB module acting iteratively. ThdB possesses only one epimerization domain and if it functions twice to produce a cyclic hexapeptide, the resulting peptide is expected to contain two d amino acid residues, as thermoactinoamide E (5) does. Based on these considerations, and taking into account the substrate selectivity of the A modules of ThdB discussed in the previous sections, the structure of thermoactinoamide E (5) was revised to cyclo (l-Ile-l-Leu-d-allo-Ile-l-Val-l-Leu-d-Leu).</p><!><p>Advanced Marfey's analysis of compound 5 using a C18 reversed-phase column. (A) Extracted-ion chromatograms at m/z 384.1514 of the l-FDAA derivatives from the hydrolysis of 5 and of d- and l-FDAA derivatives l-Leu, l-Ile, and d-allo-Ile. (B) Extracted-ion chromatograms at m/z 370.1357 of the l-FDAA derivatives from the hydrolysis of 5 and of d- and l-FDAA derivatives l-Val.</p><!><p>In addition to the congeners 2–6 discussed so far, the thermoactinoamide node cluster contained additional nodes of five new compounds, which were named thermoactinoamides G-K (7–11). They were present in even smaller amounts than 2–6, and none of them could be isolated. Therefore, their structures, including absolute configurations, were assigned combining spectral information provided by MS/MS fragmentation with biosynthetic information provided by the bioinformatic analysis of the gene cluster as well as the structures of the congeners 2–6 discussed above.</p><p>The molecular formula of thermoactinoamide G (7) (C39H65N6O6+, m/z 713.4958) contained one additional CH2 and lacked one O compared to thermoactinoamide A (1), suggesting concurrent Tyr → Phe and Val → Ile/Leu substitutions. This was confirmed by the HR-MS/MS spectrum of 7, which contained a fragment ion at m/z 566.4276 (ion α, C30H56N5O5+) composed of five Ile/Leu residues and originating from the loss of a Phe residue.</p><p>Thermoactinoamide H (8) (C37H61N6O6+, m/z 685.4646) lacked one CH2 and one O compared to compound 1, suggesting Tyr → Phe and Ile/Leu → Val substitutions. The tripeptide fragment ion at m/z 346.2123 (ion γ1) showed that the molecule contains three Ile/Leu residues in a row, and a complementary Val2Phe fragment ion at m/z 340.2951 (ion γ1) was also present in the MS/MS spectrum. The fragment ions β1/β1, both at m/z 439.3275, and the absence of a fragment ion from loss of two Val residues showed that Phe is located between the two valine residues.</p><p>The fragmentation pattern of thermoactinoamide I (9) (C42H63N6O8+, m/z 779.4705) closely resembled that of thermoactinoamide D (4), differing only in a CH2, therefore in the substitution of the Val residue (aa-4) with an Ile/Leu residue. As for thermoactinoamide D, the fragment ion at m/z 440.2174 is (Leu/Ile)1(Tyr)2 and specifically is a Tyr-Leu/Ile-Tyr fragment, because there is no Tyr-Tyr fragment, nor a fragment due to the loss of Tyr-Tyr.</p><p>Finally, thermoactinoamide J (10) (m/z 679.5117, C36H67N6O6+) and K (11) (m/z 651.4801, C36H67N6O6+) belong, like thermoactinoamide E (5), to the series of thermoactinoamides entirely composed of aliphatic amino acids and putatively synthesized by ThdB alone. Thermoactinoamide J (10) is composed of six isobaric amino acids and MS/MS fragmentation cannot provide any information on its sequence. However, on the basis of the results of this study, a putative but reliable structure, including absolute configuration, can be proposed for 10, which involves a ValIle substitution at aa-4 compared to 5. Likewise, thermoactinoamide K (11) is a lower homolog of thermoactinoamide E (5) and is expected to differ from the latter only for the IleVal substitution at aa-1.</p><!><p>Potential growth inhibitory effects of thermoactinoamide A (1) were evaluated at a single dose exposure (5 μM) for 72 h against three different cancer cell lines, namely BxPC-3 (pancreatic carcinoma), PANC1 (pancreatic carcinoma), and 3AB-OS (osteosarcoma cancer stem cell line). As previously described (Teta et al., 2019), cell proliferation was monitored by the xCELLigence System Real-Time Cell Analyzer (RTCA), based upon dynamic measurements of electronic impedance alterations which are indicative of cell viability and morphology. At 5 μM concentration, thermoactinoamide A (1) was shown to exert selective, moderate antiproliferative activity against BxPC-3 cancer cells, as inducing a) a significant reduction in the slope of the real-time proliferation curve and b) a concurrent increase in tumor cell doubling time exclusively in this cancer cell type (Figure 6). On the other hand, PANC1 and 3AB-OS carcinoma cells were not sensitive to drug treatment (Figures S3, S4).</p><!><p>Effects of thermoactinoamide A (1) on the proliferation of BxPC-3 cells monitored in real time using the RTCA platform. (A) Normalized cell index (NCI) kinetics of the BxPC-3 cells exposed to 0.5% DMSO vehicle and 5 μM of thermoactinoamide A. Arrow shows the starting point of treatment of the cells. Each cell index value was normalized to this starting point. (B) Decrease of the slope of BxPC-3 proliferation curve describes the rate of change of NCI after 72 h drug treatment. Slope values of NCI curves are relative to controls treated with 0.5 % DMSO vehicle. (C) Doubling times of NCI of BxPC-3 cancer cells after 72 h treatment with 5 μM of thermoactinoamide A and 0.5% DMSO. Doubling time is the time required for a curve cell index value to double. Data are presented as mean ± SD; n = 2. Each experiment was performed in triplicate. *p < 0.05; **p < 0.005.</p><!><p>Thermoactinoamides are lipophilic cyclopeptides, putatively assembled through a non-ribosomal peptide synthetase system consisting of two distinct trimodular NRPSs, namely ThdA and ThdB. Notably, the Thd synthetase lacks a typical thioesterase/cyclase, underlying the existence of unique off-loading and macrolactamization mechanisms to yield the final product.</p><p>Integration of genome mining and LC-MS/MS-based molecular networking has provided a powerful approach for a comprehensive investigation of the NRPS metabolome of T. vulgaris DSM 43016. The LC-MS/MS-based molecular networking represents a robust and fast technique for dereplication of complex extracts (Esposito et al., 2019) as well as for detection of novel variants of known natural products (Teta et al., 2015; Grauso et al., 2019) thereby giving an exhaustive metabolic profiling of a certain bacterial strain (Teta et al., 2016; Paulo et al., 2019). Molecular networking performed on the extracts of T. vulgaris DSM 43016 led to the discovery of eleven different members of the thermoactinoamide family, five of them being new compounds.</p><p>Bioinformatic analysis of the thd gene cluster significantly supported detection and stereo-structural determination of thermoactinoamides. Prediction of substrate selectivity of adenylation domains was useful to infer possible amino acid combinations within the thermoactinoamide cyclo-hexapeptide ring and, therefore, allow targeted detection of novel congeners. Moreover, genome mining allowed for the settling of some doubts about structure and configuration of those compounds either (a) present in amounts too small for structure elucidation by NMR, or (b) showing 2-fold symmetry and isobaric amino acids, hampering unambiguous interpretation of MS2 data. For example, bioinformatic analysis of the thd gene cluster was used to locate l-Tyr and d-Tyr residues in positions 1 and 3, respectively, within thermoactinoamide D (4), as well as to determine the amino acid sequence of thermoactinoamides E (5), J (10), and K (11).</p><p>Combining the relaxed substrate selectivity of some adenylation domains with the iterative and/or alternative use of specific modules, the thermoactinoamide synthetase is able to generate a wide chemical diversity. Detection of fully aliphatic thermoactinoamide congeners, containing only two d-configured amino acids, implies that thermoactinoamide synthetase may deviate from the collinearity rule and adopt an alternative strategy for their assemblage. Indeed, it appears that ThdB works as a standalone, trimodular iterative NRPS to create peptides E (5), J (10), and K (11). However, how Thd NRPS is programmed to assemble these compounds still remains unclear. Two possible biosynthetic models can be proposed for the ThdB-driven biosynthesis of 5, 10, and 11: the parallel model and the combinatorial model. In the parallel model, reported for iterative bacterial NRPS (Yu et al., 2017), such as gramicidin S synthetase (Hoyer et al., 2007), ThdB catalyzes condensation of two tripeptidyl monomer intermediates simultaneously tethered to the enzyme. As TE domains play a crucial role in the cyclooligomerization mechanism by iterative bacterial NRPS (Yu et al., 2017) a homologous enzyme should be present in the thd gene cluster too, but it could not be identified by in silico prediction.</p><p>In the combinatorial biosynthetic model, proposed for the first time for vatiamides biosynthesis (Moss et al., 2019), the ThdB could choose as a cognate partner either a ThdA or another ThdB NRPS via specific intermodule interaction motifs. Therefore, this combinatorial capacity results in the formation of two distinct assembly lines, namely ThdA-ThdB and ThdB-ThdB megasynthases, with the latter being involved in biosynthesis of variants E (5), J (10), and K (11). Decoding the intriguing biosynthesis of thermoactinoamides, as well as elucidating the non-canonical head-to-tail cyclization mechanism of the linear peptide, will be the aim of our future work.</p><p>Beside antimicrobial activity, thermoactinoamide A displays a moderate antiproliferative effect against a pancreatic tumor model, in the low micromolar range. Due to a chemical scaffold endowed with promising biological properties, our efforts will be addressed toward sustainable production of thermoactinoamides by microbial fermentation, in the native strain or in heterologous hosts, aiming to perform a deeper investigation of their pharmacological activity.</p><!><p>All information and datasets for this study are included in the article/Supplementary Material.</p><!><p>RT, GD, and AM: conceptualization. RT and GD: data curation, investigation, and writing—original draft. GD, AM, VC, and RT: funding acquisition. AM, VC, and RT: supervision. AM and VC: writing—review and editing.</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>1https://img.jgi.doe.gov</p><p>2Leibniz Institute DSMZ-German Collection of Microorganisms and Cell Cultures GmbH; https://www.dsmz.de/.</p><p>3Global Natural Products Social Molecular Networking: http://gnps.ucsd.edu//.</p><p>Funding. This research was funded by Ministero dell'Istruzione, dell'Universitá e della Ricerca (PRIN), Project 2015MSCKCE_003, and by Regione Campania, POR FESR 2014-2020, O.S. 1.2, Project Campania Oncoterapie No. B61G18000470007. Biological studies were funded by the Italian Ministry of Health, by Current Research Funds to IRCCS-CROB, Rionero in Vulture, Potenza, Italy.</p><!><p>The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2020.00397/full#supplementary-material</p><!><p>Click here for additional data file.</p><p>Click here for additional data file.</p>

PubMed Open Access

Structural Determinants of Oligomerization of the Aquaporin-4 Channel*

The aquaporin (AQP) family of integral membrane protein channels mediate cellular water and solute flow. Although qualitative and quantitative differences in channel permeability, selectivity, subcellular localization, and trafficking responses have been observed for different members of the AQP family, the signature homotetrameric quaternary structure is conserved. Using a variety of biophysical techniques, we show that mutations to an intracellular loop (loop D) of human AQP4 reduce oligomerization. Non-tetrameric AQP4 mutants are unable to relocalize to the plasma membrane in response to changes in extracellular tonicity, despite equivalent constitutive surface expression levels and water permeability to wild-type AQP4. A network of AQP4 loop D hydrogen bonding interactions, identified using molecular dynamics simulations and based on a comparative mutagenic analysis of AQPs 1, 3, and 4, suggest that loop D interactions may provide a general structural framework for tetrameric assembly within the AQP family.

structural_determinants_of_oligomerization_of_the_aquaporin-4_channel*

Introduction<!>Expression Constructs and Mutagenesis<!>Cell Culture and Transfection<!>Confocal Microscopy<!>Cell Surface Biotinylation<!>BN-PAGE<!>Immunoblotting<!>Statistics<!>Simulations<!>Calcein Fluorescence Quenching<!>Mutation of Loop D of AQP4 Reduces Tetrameric Assembly<!><!>Surface Expression of Non-tetrameric Mutants<!><!>Mutants in the Plasma Membrane Exhibit Reduced Tetramerization<!>Fluorescence Recovery after Photobleaching (FRAP)<!>BN-PAGE of Biotinylated Cell Surface Protein<!>Forster Resonant Energy Transfer (FRET)<!><!>Loop D Mutants Have Wild-type Water Permeability<!><!>Loop D Mutants Do Not Relocalize to the Plasma Membrane in Response to Hypotonicity<!><!>Molecular Dynamics Simulations Suggest a Dynamic Network of Loop D Interactions<!><!>Loop D May Represent a Common Motif for Oligomeric Assembly Across the AQP Family<!><!>Discussion<!>Author Contributions<!>

<p>The aquaporin (AQP)3 family of integral membrane proteins facilitate both osmosis and diffusion of small polar molecules through biological membranes. A wealth of medium-to-high resolution structural data for various family members (there are 48 AQP structures in the Protein Data Bank) all suggest that the AQP homotetrameric quaternary structure is highly conserved despite diversity in solute permeability, subcellular localization, and trafficking responses of individual AQPs. Early biochemical work, using carboxyl-amine fusion dimers (consisting of 1 wild-type unit and 1 unit lacking the cysteine residue required for mercurial inhibition), showed that monomers are the functional AQP units (1); numerous molecular dynamics simulation studies support this (2). Recent work has suggested that isolated AQP monomers are equally capable of facilitating water transport as those incorporated into a tetramer (3). Therefore it is not clear why AQPs retain this tetrameric structure. Regulation of AQP function by the formation of heterotetramers has been suggested for some plant AQPs (4). The fifth, central pore formed at the 4-fold axis of the tetramer has also been suggested to transport carbon dioxide (5) and cations (6), at least in mammalian AQP1. Trigger-induced relocalization of AQP-containing vesicles to the plasma membrane is a well established regulatory mechanism for AQPs; the best studied example of this is relocalization of AQP2 to the apical membrane of the collecting duct in the mammalian kidney in response to arginine vasopressin (also called anti-diuretic hormone). There is also evidence for similar mechanisms for other AQPs including AQP1 (7), AQP5 (8), and AQP7 (9). A naturally occurring AQP2 mutant (R187C) has also been reported that is unable to relocalize in response to arginine vasopressin or to form tetramers (10). Given the ubiquity of these regulatory responses across the AQP family and the conservation of the tetrameric quaternary structure, it may be that these trigger-induced relocalization responses involve interaction with proteins that only recognize the tetrameric form of AQPs.</p><p>Here we demonstrate that intracellular loop D of AQP4 forms vital homomeric interactions between AQP subunits that stabilize the tetrameric quaternary structure. We also show that loss of tetramerization does not affect single channel water permeability. Our data suggest that tetramerization is not required for AQP4 to be trafficked through the endoplasmic reticulum and Golgi to the plasma membrane, but that unlike wild-type AQP4, the non-tetrameric mutants are unable to relocalize to the plasma membrane in response to changes in local osmolality. Finally, based on loss and gain of oligomerization mutants of AQP1 and AQP3, we suggest that loop D-mediated inter-monomer interactions may control formation of the signature quaternary structure of the family.</p><!><p>Human AQP4 cDNA cloned into pDEST47 (Life Technologies) was used as previously described (11). An untagged AQP4 construct was created from this by mutagenesis of the first two codons of the GFP linker peptide to stop codons. These were used as templates for mutagenesis following the QuikChange protocol (Stratagene). Mutagenic primers were synthesized by Sigma. Mutant plasmids were amplified in TOP10 Escherichia coli with 100 ng/ml of ampicillin selection. Plasmid DNA was purified using a Wizard Maxiprep kit (Promega) and diluted to 1 mg/ml for transfection.</p><!><p>HEK293 cells were cultured routinely in DMEM with l-glutamine (Sigma) supplemented with 10% (v/v) fetal bovine serum (Sigma) and without antibiotics in humidified 5% (v/v) CO2 in air at 37 °C. Cells were seeded into either tissue culture treated 6-well plates (Falcon) for Blue Native (BN)-PAGE and biotinylation or 35-mm FluoroDishesTM (World Precision Instruments) for confocal microscopy. Cells were transfected (at 50% confluence) using polyethyleneimine (branched, average Mr ∼25,000, Sigma) as previously described (11). MDCK cells were cultured in the same conditions as HEK293 cells. Stable transfections were done using the neomycin resistance gene on the pDEST47 vector. Cells were transiently transfected as described above, then trypsinized and serially diluted into tissue culture plates after 24 h. Cells were treated with 700 μg/ml of G418 antibiotic for 2 weeks, with medium replaced every third day. GFP-expressing resistant colonies were picked using cloning cylinders (Sigma) and serially diluted. The lowest dilution that grew to confluence was used to generate a stable cell line. Cellular protein was subjected to SDS-PAGE and Western blotting as previously described (11), and the highest expressing clone was chosen for experiments. No endogenous AQP4 was detected in Western blots. Reduced G418 pressure (300 μg/ml) was used to maintain the stable cell lines after colony isolation. All cells were routinely tested for mycoplasma using the EZ-PCR test kit (Biological Industries), and all data reported are from cells that tested negative. HEK293 and MDCK cells both expressed only the M1 isoform of AQP4 from the wild-type AQP4 construct. This was confirmed by Western blotting comparing the wild-type construct to AQP4 constructs in which either the M1 or M23 translation initiation sites were removed, as previously described (11).</p><!><p>AQP4-GFP constructs were imaged in live HEK293 cells using a Zeiss LSM 780 confocal microscope with a ×63 1.4 NA oil immersion objective. GFP was excited using the 488-nm line of an argon laser, Venus using the 514-nm line and mTurquoise2 using a 405-nm diode laser. Hypotonic exposure was performed by adding 3 ml of ddH2O to cells in 1 ml of growth medium (280 to 70 mosmol/kg of H2O). Line profiles across the cell membrane and cytoplasm were extracted using ImageJ as previously described (7). Relative membrane expression was calculated from these profiles (3 profiles per cell and at least 3 cells per image) using in-house Matlab code. Fluorescence recovery after photobleaching (FRAP) was done using a circular bleaching 1-μm area of radius. Recovery curves were fitted to a single phase exponential recovery function and diffusion coefficients were calculated using the approach and equations of Kang et al. (12). Recovery curves were collected from 5 different cells on the same plate per experiment. FRET experiments were done using the sensitized emission methodology with the FRET signal corrected for donor emission in the acceptor channel and direct excitation of the acceptor, following van Rheenen et al. (13) and normalized to the acceptor emission to give an apparent FRET efficiency. The contrast of some images in the figures was adjusted manually using ImageJ to aid the eye. All analysis was performed on raw, unadjusted images.</p><!><p>Cell surface amines were biotinylated using an amine-reactive biotinylation reagent that is not cell permeable (Thermo number 21328, EZ-Link Sulfo-NHS-SS-Biotin), and surface AQPs were detected using a neutravidin-based ELISA as previously described (11).</p><!><p>Transfected cells were lysed in ice-cold BN lysis buffer (1% (v/v) Triton X-100, 10% (v/v) glycerol, 20 mm bis-tris, 500 mm aminohexanoic acid, 20 mm NaCl, 2 mm EDTA, pH 7.0, 250 μl/well). The lysate was centrifuged at 21,000 × g at 4 °C for 10 min to remove insoluble material. The supernatant was collected and diluted 10-fold in Triton lysis buffer. 8% bis-tris-buffered polyacrylamide gels (0.75 mm) at pH 7.0 containing 66 mm aminohexanoic acid were used. Wells were topped up with cathode buffer (50 mm Tricine, 15 mm bis-tris, 0.02% (w/v) Coomassie G-250, pH 7.0). 10 μg of BSA was used as a molecular mass marker, giving bands at 66 and 132 kDa. Gels were run on ice at 100 V until samples entered the gel, then at 180 V until the Coomassie dye front reached the end of the gel.</p><!><p>BN-PAGE gels were destained with 40% (v/v) methanol, 10% (v/v) glacial acetic acid for 30 min, refreshing the destaining solution every 10 min. Gels were soaked for 30 min in 1% SDS in Tris-buffered saline, pH 7.4, at room temperature. Proteins were blotted onto PVDF membrane by wet transfer at 100 V for 1 h. Coomassie-stained BSA marker bands were marked onto the membrane using a felt-tipped pen. Membranes were blocked in 20% (w/v) Marvel-skimmed milk powder in 0.1% PBS-Tween for 1 h. Membranes were incubated overnight at 4 °C on a roller in rabbit anti-AQP4 antibody (Abcam, ab128906) diluted 1:5,000 or rabbit anti-GFP (Abcam, ab6556) diluted 1:10,000, both in 5 ml of 0.1% PBS-Tween. Membranes were washed in 0.1% PBS-Tween and incubated with donkey anti-rabbit HRP (Santa Cruz, sc-2313) diluted 1:10,000 in 20 ml of 0.1% PBS-Tween at room temperature for 1 h. HRP was detected on x-ray film using ECL reagent (Amersham Biosciences).</p><!><p>For multiple comparisons, one-way ANOVA was used, followed by post hoc t-tests with the p values subjected to Bonferroni correction for multiple comparisons. All data are presented as mean ± S.E.</p><!><p>Simulations were done using the GROMOS 53A6 forcefield (14) extended to include lipid parameters (15) in Gromacs version 4.5.5 (16). An AQP4 tetramer was generated according to the biological assembly entry in the AQP4 Protein Data Bank file 3GD8 (17). N and C termini of the protein were truncated in the structure so proteins were simulated with neutral termini. The AQP4 tetramer was embedded into 5 pre-equilibrated 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine bilayers using inflateGRO (18) and hydrated using Gromacs. Na+ and Cl− were added to a final concentration of 100 mm. Equilibration was achieved by steepest gradient energy minimization, 100-ps NPT simulation with 1,000 kJmol−1 nm−1 restraints on protein heavy atoms followed by three 1-ns NVT simulations with 1000, 100, and 10 kJmol−1 nm−1 restraints on protein heavy atoms followed by 30-ns unrestrained simulation. A Nosé-Hoover thermostat (0.5 ps, 310 K) was used to maintain constant temperature and 2 Parinello-Rahmann barostats (2 ps, 1 atm) were used to maintain constant pressure with zero surface tension. 1.4-nm cut-offs were applied for dispersion and short-range electrostatic interactions. Long-range electrostatics were treated using particle mesh Ewald. Hydrogen bonds were identified using the H-bonds plugin of visual molecular dynamics (19) using 3 Å and 20° cut-offs. Hydrogen bond occupancy was calculated according to these cut-offs at 100-ps intervals along the 100-ns trajectories and averaged over the 4 monomers and 5 trajectories.</p><!><p>Plate reader-based calcein fluorescence quenching was done following Fenton et al. (20). MDCK cells were plated into black-walled, clear-bottomed tissue culture treated 96-well plates (Greiner) at 50% confluence 24 h before the experiment. Cells were loaded with 5 μm calcein-AM in growth medium supplemented with 1 mm probenecid (to inhibit dye leakage) for 90 min. Cells were washed twice with HEPES-buffered growth medium supplemented with 1 mm probenecid, then covered with 75 μl of probenecid-supplemented HEPES-buffered medium. Fluorescence was read on a BioTek synergy HT plate reader with injector system. Each well was read continuously (dt = 50 ms) for 5 s, followed by injection of 75 μl of HEPES-buffered medium containing 400 mm mannitol to give a final concentration of 200 mm and an osmotic gradient of 200 mosmol. Fluorescence was read for a further 50 s. Normalized fluorescence values were converted to normalized volumes using a Coulter counter generated standard curve. Single-phase exponential decay functions were fitted and rate constants were taken as proportional to the membrane water permeability.</p><!><p>Using the crystal structure of AQP4 (17), residues likely to form the tetrameric interface were identified, based on the physical distance between residues on adjacent monomers. Alanine substitution mutagenesis was used to investigate the contribution of these residues to oligomeric assembly. The identified residues were clustered into two regions. The first cluster comprises a patch of hydrophobic residues at the interfaces of TM1 and TM2 of one monomer with TM4 and TM5 of the adjacent monomer (Fig. 1A); the second cluster comprises 12 predominantly polar and charged residues forming intracellular loop D and the bottom of TM2 (Fig. 1B). 24 single point alanine-substitution mutants were made and six compound mutants were used to investigate the possibility of synergistic effects of several residues. There are two protein kinase A/C consensus sites in loop D (Ser180 and Ser188) so phosphomimetic mutations were also made (S180D and S188D). All mutants are listed in the left-hand column of Table 1. BN-PAGE followed by immunoblotting was used to assess the oligomeric state of all mutants expressed in HEK293 cells. Representative BN blots are shown in Fig. 2A and the effect of all mutations on oligomeric assembly and surface expression summarized in Table 1 and Fig. 2B. None of the hydrophobic residues in the hydrophobic patches had any effect on tetramer formation, either in isolation or in compound mutants. TM compound mutants consisted of simultaneous mutations of all residues identified within that transmembrane segment (e.g. TM1 denotes the simultaneous mutations I43A/I47A/L50A/L51A/I57A). Of the loop D single alanine mutants, only D179A had an effect on oligomeric assembly in isolation, and this effect was only minimal, with 95 ± 2% of the protein still assembled into tetramers. Unlike the hydrophobic cluster, the compound mutants of loop D had a clear effect on the ability of AQP4 to tetramerize. The two compound mutants, loop D1 (D179A/S180A/K181A/R182A/T183A) and loop D2 (D184A/V185A/T186A/G187A/S188A), caused reductions in oligomeric assembly: only 19 ± 4% of the loop D1 protein assembled into tetramers with both dimers (27 ± 5%) and monomers (53 ± 4%) being present, whereas the loop D2 protein predominantly formed dimers (67 ± 7%) with 33 ± 7% assembled into tetramers (n = 3).</p><!><p>Residues at the AQP4 tetrameric interface. A, we identified hydrophobic residues (orange) in TMs 1, 2, 4, and 5 that formed inter-monomer contacts in the crystal structure and B, polar residues (green) at the bottom of TM2 and in the intracellular loop D. C, ball diagram showing the position of the identified residues in the primary sequences and secondary structural motifs of AQP4. Two regions of loop D were selected for compound mutation, which we denote loop D1 (179DSKRT183) and loop D2 (184DVTGS188). Blue lines represent the approximate position of membrane lipid headgroups. All residues are listed in Table 1.</p><p>Oligomerization state and surface expression of AQP4 mutants</p><p>Oligomerization state of all AQP4 mutants analyzed. Compound mutations are represented by a solid slash between the relevant point mutations, e.g. L72A/L79A. T = tetramer, D = dimer, M = monomer. All data are reported as mean ± S.E., n = 3.</p><p>BN-PAGE and Western blotting of AQP4 mutants. A, representative Western blots following BN-PAGE of Triton X-100-solubilized AQP4 mutants, showing the effect of the loop D1 and loop D2 compound mutations, and a lack of effect of mutations on the transmembrane hydrophobic patch. 66 and 132 denote the positions of BSA molecular weight marker bands. The AQP4-GFP construct, including linker peptide, has a predicted molecular mass of 63.1 kDa. B, percentage of protein assembled into tetramers calculated using densitometry following BN-PAGE and Western blotting. Effective mutations are highlighted with white bars. Data are presented as mean ± S.E. from 3 experimental repeats.</p><!><p>There was no significant difference in the surface expression of the loop D mutants compared with wild-type AQP4. Surface expression was assessed qualitatively by live cell confocal microscopy using GFP-tagged AQP4 mutant constructs and quantitatively by cell surface biotinylation. Fig. 3A shows representative confocal micrographs of HEK293 cells transfected with GFP fusion proteins of AQP4 wild-type, and the loop D1 and loop D2 mutants. Surface expression in transiently transfected HEK293 cells measured by cell surface biotinylation was not significantly different (p = 0.53, one-way ANOVA, n = 3) for either of the loop D compound mutants (D1 and D2) or the single alanine mutants (Fig. 3B). This suggests that the trafficking machinery is able to interact with non-tetrameric aquaporins. The cell surface biotinylation data are summarized for all mutants in the third column of Table 1.</p><!><p>Plasma membrane localization of non-tetrameric mutants. A, representative fluorescence micrographs of HEK293 cells transfected with C-terminal GFP fusions of AQP4 WT and loop D mutants. B, surface expression of AQP4 mutants in HEK293 cells measured by cell surface biotinylation followed by a neutravidin-based ELISA. Loop D compound mutants are highlighted in red. The S52D mutant was used as a negative control for surface expression. n.s., not significant. C, reprsentative blots of AQP4 mutants subjected to BN-PAGE. WC, whole cell lysate; S, surface protein only, isolated by cell surface biotinylation. D, representative FRAP curves from photobleaching AQP4-GFP fusion proteins in HEK293 cells. E, average half-times of fluorescence recovery averaged over fits to 5 curves per experiment and 6 experimental repeats. All data are presented as mean ± S.E.</p><!><p>It was important to confirm the reduction in tetramerization of AQP4 molecules that had been constitutively trafficked to the cell surface and to rule out intracellular retention of dimeric/monomeric species or changes in detergent sensitivity caused by loop D substitutions. Several complementary biophysical techniques were used to address this.</p><!><p>Recovery curves were collected from 5 different cells per experimental repeat (n = 6). Representative fluorescence recovery curves are shown in Fig. 3D, along with the average recovery half-times. From these, diffusion coefficients were calculated: 5.2 ± 0.3 × 10−3 μm2 s−1 (AQP4 WT), 5.9 ± 0.3 × 10−3 μm2 s−1 (loop D1), and 5.8 ± 0.2 × 10−3 μm2 s−1 (loop D2). Both loop D mutant diffusion coefficients were significantly different from the wild-type (D1 p = 0.01 and D2 p = 0.02 by Student's t test following ANOVA, with p values subjected to Bonferroni correction). The FRAP data suggest reduced tetramerization for the surface-localized loop D mutants, although the increased diffusion coefficient could also be explained by the inability of these mutants to form a complex with a third party protein.</p><!><p>To complement these FRAP experiments, biotinylated cell surface proteins were isolated using neutravidin-coated plates, eluted by reducing the S-S bond incorporated into the biotinylation reagent (using 1% β-mercaptoethanol in BN lysis buffer) and subjected to BN-PAGE (representative blots are shown in Fig. 3C). Surface-localized mutant AQP4 molecules subjected to BN-PAGE had the same changes in tetramerization seen in whole cell lysates (n = 3).</p><!><p>To complement the above analyses, AQP4 constructs tagged with Venus (a yellow fluorescent protein (YFP) derivative) and mTurquoise2 (a cyan fluorescent protein (CFP) derivative) were generated and co-transfected into HEK293 cells to form a FRET biosensor for homo-oligomerization in living cells. The wild-type AQP4 constructs gave a robust FRET signal with an average apparent efficiency of 44.2 ± 3.6% (Fig. 4). We were unable to measure any FRET in cells co-transfected with AQP1-Venus and AQP4-Turquoise despite high co-localization (data not shown), suggesting that the FRET interactions occur primarily within the AQP4 tetramers and not between tetramers that are transiently close together in the plane of the membrane. The probability of a particular donor molecule taking part in FRET is dependent on the number of acceptors within the Forster radius and vice versa. For CFP-YFP, the Forster radius is ∼5 nm (21). Based on the AQP4 crystal structure, the monomer-monomer center of mass separations are 2.8 (adjacent monomers) and 3.9 nm (diagonal monomers), respectively, so both would be expected to contribute to the FRET signal (assuming that the average separation of the C-terminal tails is similar). The average number of FRET pairings in a sample of co-transfected cells is therefore dependent on the level of AQP4 oligomerization. Both D1 and D2 compound mutants had a slightly larger than 2-fold reduction in FRET efficiency (to 17 ± 6 and 20 ± 4%, respectively, p = 0.003 and p = 0.005, n = 4) compared with the wild-type (Fig. 5A), suggesting that these constructs have a reduced propensity to oligomerize in live cells, further confirming that the changes seen in the BN-PAGE were not mediated by changes in detergent sensitivity. Furthermore, for the mutants, there was no difference in FRET efficiency between plasma membrane and intracellular membranes (Figs. 5, B and C), suggesting that the oligomerization state of these mutants is the same in all membrane compartments (e.g. Golgi, vesicles, plasma membrane). Taken together, these data suggest reduced tetramerization for the AQP4 loop D mutants in the plasma membrane of living cells.</p><!><p>A FRET biosensor for AQP4 oligomerization. A, fluorescence confocal microscopy of live HEK293 cells transiently transfected with AQP4-Venus alone, AQP4-mTurquoise2 alone, and the two co-transfected. Excitation (Ex) at 405 nm is for Turquoise and 514 is for Venus. The contrast of these images has been manually optimized to aid the eye. All analysis was performed on raw, unadjusted images. Em, emission.</p><p>Reduced FRET from AQP4 mutants. A, average apparent FRET efficiencies for AQP4 wild-type and loop D1 and D2 mutants, calculated by normalizing the corrected FRET intensity to YFP intensity for each pixel. Five different areas on each plate were imaged per experimental repeat, n = 4. B, representative line scan across a cell, passing through membrane and cytoplasm and avoiding the nucleus. Whereas the YFP signal (red) shows clear peaks at the plasma membrane, the FRET efficiency (blue) does not. C, representative processed FRET efficiency images compared with unprocessed FRET and YFP. Whereas both YFP and the raw FRET show clear membrane signals, the FRET efficiency does not.</p><!><p>Both loop D compound mutants (D1 and D2) and wild-type AQP4 were stably transfected into MDCK cells to measure water channel function. Membrane water permeability of MDCK cells, measured by calcein fluorescence quenching (Fig. 6A), was increased 7-fold by stable expression of AQP4. Water permeability due to the transfected AQP was calculated by subtracting the permeability of untransfected cells (Fig. 6B); the resulting permeability was normalized to the surface expression measured by cell surface biotinylation to allow for differences in cellular expression due to differences in the position of chromosomal integration of the stably transfected gene (Fig. 6, C and D). After normalization to surface expression, no significant difference in permeability was observed between the loop D mutants and wild-type AQP4 (one-way ANOVA, p = 0.17, n = 4), suggesting that tetramerization of AQP4 is not required for full water channel activity.</p><!><p>Water permeability of non-tetrameric mutants. A, representative calcein fluorescence quenching curves from stably transfected MDCK cells subjected to a 200 mosmol of mannitol osmotic gradient. B, water permeability of MDCK cells normalized to AQP4 WT-transfected MDCK cells. C, normalized surface expression of AQP4 constructs in the stably expressing MDCK clones used for water permeability measurements, measured by cell surface biotinylation. D, MDCK membrane water permeability normalized to surface expression, to give normalized single channel permeability. All data are presented as mean ± S.E., n = 4.</p><!><p>We recently reported that AQP4 rapidly relocalizes to the plasma membrane from intracellular membranes in response to reduced extracellular tonicity (11) and that this phenomenon is true for other mammalian AQPs (7, 22). Despite wild-type water permeability and constitutive surface expression of the loop D mutants, D1 and D2, this response to hypotonicity was not observed for either. Relative membrane expression of wild-type AQP4-GFP imaged in live HEK293 cells increased from 27.9 ± 3.5 to 67.1 ± 4.5 (p = 0.003, n = 3) upon reduction of the extracellular tonicity to 85 mosmol, whereas the distribution of both the D1 and D2 mutants did not change significantly (p = 0.29 and p = 0.34 respectively, n = 3; Fig. 7).</p><!><p>Tonicity-induced translocation of non-tetrameric mutants. A, representative fluorescence micrographs of HEK293 cells transfected with AQP4-GFP fusion proteins, before and after 30 s of exposure to hypotonic (85 mosmol) medium. B, relative membrane expression of AQP4-GFP fusion proteins before and after exposure to hypotonic medium. At least 4 cells per image were analyzed for each experimental repeat, n = 3. Data are presented as mean ± S.E.</p><!><p>To investigate the contribution of loop D to AQP4 oligomerization we did five independent 100-ns molecular dynamics simulations of a hydrated AQP4 tetramer embedded in a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine bilayer. We found a dynamic network of hydrogen bonds between loop D residues on adjacent monomers and also between loop D residues and both a glutamine (Gln86) and histidine (His90) residue at the bottom of TM2. On average, each monomer was involved in 3.8 ± 1.4 (mean ± S.D.) loop D hydrogen bonds with its two neighboring molecules at any given time, which is equivalent to 7.6 ± 2.8 loop D hydrogen bonds per tetramer. Every loop D residue apart from Ser180 was able to act as a hydrogen bond donor or acceptor with another loop D residue on one of the two adjacent monomers; 7 residues (Arg182, Thr183, Asp184, Val185, Thr186, Gly187, and Ser188) were found to have at least 3 different possible hydrogen bonding interactions. This network of interactions is represented as a heat map in Fig. 8A. Interactions involving Val185 and Gly187 were backbone hydrogen bonds. Both of these residues were able to act as hydrogen bond donors from the backbone amine group (Val185 to an adjacent Val185, and Gly187 to Asp184, Val185 and Thr186) and acceptors at the backbone carbonyl group (Val185 from an adjacent Val185 as well as Thr183, Thr186, Gly187, and Ser188, and Gly187 from Thr186, Ser188, and Gln86). Loop D (as well as loop A) showed relatively large structural fluctuations (Fig. 8, B and C) despite maintaining an average of 3.8 ± 1.4 inter-monomer hydrogen bonding interactions. Representative snapshots showing the most highly occupied hydrogen bonds, His90-Asp179, Ser188-Gly187, Ser188-Ser188, Gln86-Asp184, Gln86-Ser188, and Thr183-Asp184 are shown in Fig. 8F. Loop D could potentially stabilize the AQP4 tetramer in one of two ways: by preventing monomers from drifting apart during rapid association/dissociation of the inter-monomeric TM interfaces, or by providing inter-monomer interactions that stabilize the tetramer in a more permanent way, preventing dissociation in the first place. In simulations we found that the average monomer-monomer center of mass separation was stable with a root mean square fluctuation of only 0.53 Å (c.f. typical carbon-carbon bond length of ∼1.5 Å) about a mean of 28.67 Å, and the inter-monomer buried (solvent-inaccessible) area was also stable, with a root mean square fluctuation of 230 Å2 about a mean of 2960 Å2. Representative traces are shown in Fig. 8, D and E. This suggests that there is no spontaneous dissociation/reassociation of the inter-monomer interfaces, at least on the time scale of these simulations (100 ns).</p><!><p>A dynamic network of loop D hydrogen bonds in simulations of AQP4. A, heat map showing percentage occupancy of loop D hydrogen bonds averaged over 4 monomers comprising a tetramer and over 5 independent 100-ns simulations. B, backbone heavy atom root mean square deviation (RMSD) of AQP4 residues, demonstrating the structural flexibility of loops A and D. RMSDs were calculated independently for each trajectory and averaged. C, close-up view of loop D RMSD. D, representative inter-monomer center of mass distance for a single interface (red) and averaged over the four interfaces of a tetramer (blue). E, representative inter-monomer buried area for a single interface (red) and averaged over the four interfaces of a tetramer (blue). F, snapshots of molecular dynamics trajectories in which the six most highly occupied inter-monomer hydrogen bonds involving loop D residues are occupied.</p><!><p>To investigate whether our AQP4 data could be more widely applicable to other members of the AQP family, we performed a sequence alignment of the loop in all human AQPs, as well as the E. coli AQPs as comparison (Fig. 9A) using the Clustal Omega software (23). AQP1 contains a similar acidic-X-basic-basic motif to AQP4 in the first half of the loop (DRRRR versus DSKRT), so we made the D1 mutant in AQP1 (D158A/K159A/K160A/K161A/K162A). This motif is lacking in AQP3, therefore we introduced it via 2 different mutations, P181S/Y182K/N183R and P181E/Y182K/N183R (to give DSKRN and DEKRN versus the wild-type DPYNN). The D1 compound mutation in AQP1 caused a similar loss of oligomerization to that observed for AQP4. Both AQP3 mutations caused the protein to migrate primarily as a band with a molecular weight consistent with a dimeric species, whereas wild-type AQP3 appeared to migrate primarily as a monomer (Fig. 9B), consistent with previous studies on the oligomerization state of this AQP. Both AQP1 and AQP3 migrated as diffuse bands, consistent with glycosylation. PNGase F treatment was used to attempt deglycosylation. This appeared to be incompatible with the BN-PAGE experiments, because 1-h treatments with PNGase F at 37 °C caused AQP1 and −3 wild-type and mutants to form aggregates that did not migrate beyond the interface between the stacking and separating gels (data not shown). This may be due to increased temperature sensitivity of Triton X-100-solubilized AQPs. Finally, to investigate whether loss of AQP oligomerization is a general feature of the glyceroporin subfamily, we transfected HEK293 cells with AQPs 9 and 10 and subjected Triton X-100-extracted protein to BN-PAGE. Unlike AQP3, we found that AQPs 9 and 10 migrated exclusively as tetramers (Fig. 9B).</p><!><p>Comparison of loop D between different human AQPs. A, all 13 human AQPs as well as AQPZ and GlpF (both from E. coli) were aligned using Clustal Omega (EMBL-EBI). Acidic residues are colored red; basic, blue; and neutral residues able to form sidechain hydrogen bonds, green. B, BN-PAGE and Western blotting of wild-type and mutant AQPs 1 and 3 and wild-type AQPs 9 and 10. 66 and 132 represent BSA molecular weight markers.</p><!><p>Our data show that the plasma membrane abundance of AQP4 in HEK293 cells is unaffected by the oligomeric state of the protein. This strongly suggests that tetrameric assembly is not required for AQP4 to be correctly trafficked through the endoplasmic reticulum and Golgi to the plasma membrane. It is possible that our mutants are trafficked as tetramers and then dissociate once inserted into the plasma membrane; the FRET data suggest that this is unlikely to be the case, although we cannot differentiate between particular intracellular compartments on the basis of this data, leaving open the possibility of differences in oligomerization state between different intracellular compartments. Our diffusion coefficients calculated from the FRAP data compare reasonably with other reports of diffusion of AQP-GFP constructs in mammalian cells (e.g. 5.7 × 10−3 μm2 s−1 for AQP2 (24), 3.1 × 10−3 μm2 s−1 for AQP1 (25)). We found a small, but statistically significant increase in diffusion coefficient for the mutants that were non-tetrameric in BN-PAGE. Although we have not estimated hydrodynamic radii for our constructs, the relationship D ∝ ln(1/R) derived by Saffman and Delbruck (26) suggests a hydrodynamic radius of 1.2 ± 1.0 nm for our non-tetrameric constructs (assuming r = 4 nm for a freely diffusing AQP4 tetramer), which is at least consistent with a loss of tetrameric assembly. We found no difference in the mobile fraction between the wild-type and D1 or D2 mutants (90 ± 6, 89 ± 4, and 92 ± 7%, respectively, n.s. p = 0.28), suggesting that there is no difference in the proportion of the protein involved in membrane anchoring interactions. Although we cannot rule out loss of an AQP4 binding partner as an alternative explanation of these data with absolute certainty, in combination with the cell surface BN-PAGE and FRET data it is highly suggestive of non-tetrameric mutant AQP4 molecules in the plasma membrane. Human AQP3 was reported to exist in all four possible oligomeric states (monomer, dimer, trimer, and tetramer) in the plasma membrane of erythrocytes (27). It may be that some or all members of the AQP family can be trafficked to the plasma membrane independently of their oligomerization state.</p><p>Our molecular dynamics simulations suggest a dynamic network of transient hydrogen bonds between the residues of loop D of adjacent monomers and also with two residues at the bottom of TM2. The fact that almost every residue involved in this network has several hydrogen bonding options may explain why none of the single alanine substitution mutations had an effect on tetrameric assembly. For example, the most highly occupied hydrogen bond that we found was His90-Asp179, which had 66.5% occupancy averaged over the four monomers and five trajectories. When this bond was not occupied, the Asp184 side chain was able to act as a hydrogen bond acceptor from the histidine imidazole NH group, and the Thr186 side chain hydroxyl group was able to act as a hydrogen bond donor to Asp179. These alternative interactions may be able to partially compensate for the loss of the Asp179-His90 interaction in the D179A and H90A single mutations. Asp179 had only one alternative interaction, whereas His90 had three. This may explain why the D179A mutation had a slight effect on oligomerization, whereas the H90A mutant did not. Interestingly, MD simulations have suggested that a histidine residue just beyond the bottom of TM2 in the intracellular loop B (H95) could modulate the channel open probability via spontaneous formation and dissociation of a hydrogen bond with a cysteine residue at the interface between TM4 and loop D (Cys178) (28). In our simulations, we did not observe this hydrogen bond. A recent study combining in silico and in vitro evidence also did not observe this hydrogen bond in simulations and suggested that His95 may indeed act as a channel gating residue, but by formation of a histidine protonation state-dependent salt bridge with a glutamate residue (Glu41), independent of Cys178 (29).</p><p>S-Nitrosylation of AQP11 at a cysteine residue in the extracellular loop E (Cys227) has been suggested to be required for AQP11 oligomeric assembly, with the C227S mutant showing reduced oligomeric assembly in mouse kidney (30). This supports the idea of a role for post-translational modification in AQP oligomerization. Loop D of AQP4 contains two sites predicted to be targets for post-translational modification. These are protein kinase sites at Ser180 and Ser188. Phosphorylation at Ser180 was suggested to reduce AQP4 water permeability via a gating effect in LLC-PK1 cells (31), but this was not supported by molecular dynamics simulations of AQP4-Ser(P)180 (32), and structural studies showed no difference between wild-type AQP4 and a phosphomimetic S180D mutant (33, 34). We used phosphomimetic (S180D, S188D) and phospho-blocking (S180A, S188A) mutations to investigate a potential role for post-translational modification in the AQP4 oligomeric assembly. We found that all four phosphorylation mutants were able to assemble into tetramers, suggesting that phosphorylation of loop D is not involved in tetrameric assembly of AQP4.</p><p>We recently reported that AQP4 in primary rat astrocytes and HEK293 cells rapidly relocalizes to the plasma membrane upon reduction of extracellular tonicity (11); this may involve changes in interactions between AQP4-containing vesicles and cytoskeletal elements as described by others (35, 36). Interestingly, neither of the mutants with reduced ability to tetramerize were able to relocalize in response to a hypotonic extracellular stimulus. It is possible that a binding partner of AQP4 involved in the translocation response recognizes an epitope formed by the interface of several AQP molecules within a tetramer, and that disrupting tetrameric assembly disrupts this epitope. Although the C-terminal PKA phosphorylation site (Ser276), which controls this response is ∼100 residues away in the primary sequence, it may be that phosphorylation of this residue causes a conformational change in the large (∼70 residue) C-terminal tail of AQP4, which allows an AQP4-binding protein to bind to the intracellular face of AQP4 including loop D. Structural data for AQPs are routinely collected after cleaving the C terminus at the bottom of TM6 to aid crystallization and high-resolution structure determination (37). This is true for AQP4 and in the structure that we used to identify mutagenic targets and as input for our simulations, the protein was truncated at the bottom of TM6, at residue 254 of 323 (17). This makes it difficult for us to make any concrete predictions about interactions of the AQP4 C terminus, especially given its large size.</p><p>Based on a sequence alignment of loop D, we made mutants of human AQPs 1 and 3. Wild-type AQP1 existed as a tetramer when extracted from HEK293 membranes using Triton X-100, whereas AQP3 existed primarily as a monomer, both of which are in agreement with previous reports (27, 38). Mutating the first five residues of AQP1 loop D to alanine (which rendered AQP4 primarily monomeric) caused the protein to migrate primarily as a monomer in BN-PAGE. Introducing the conserved acid-X-base-base motif (present in loop D of both AQPs 1 and 4) into AQP3 caused it to migrate primarily as a dimer. It has been suggested that lack of oligomerization may have a role in controlling substrate selectivity of the glyceroporin subfamily (39), which consists of AQPs 3, 7, 9, and 10 in humans. However, we find that, unlike AQP3, both AQPs 9 and 10 both exist exclusively as tetramers when extracted from HEK293 membranes. This suggests that solute permeability of AQPs is not correlated with the oligomeric state.</p><p>Previous mutational analysis of AQP1 found an extracellular motif consisting of an aspartate residue (Asp185) at the top of TM5 that can interact with a lysine residue (Lys51) at the top of TM2 to increase tetramer stability in detergent (40). Whether this had an effect on native protein in live cells was not clear. The S205D mutant of an insect AQP (AQPcic) was shown to exist in a primarily monomeric state when extracted from yeast and Xenopus oocyte membranes, whereas the wild-type was tetrameric (41, 42). This mutant had a complete loss of water channel function, and it is not clear whether this was due to subtle changes in structure leading to a gating effect, gross misfolding, loss of surface expression, or a direct result of the loss of oligomerization through loss of monomer-monomer interactions that stabilize the open state of the pore. This makes interpretation of this result very difficult.</p><p>In summary, we show that loop D of AQP4 forms a hub for a dynamic network of interactions that stabilize the AQP4 tetramer, both when solubilized using non-ionic detergent and in living mammalian cells. Tetrameric assembly is not required for either endoplasmic reticulum-to-Golgi-to-plasma membrane trafficking or water channel activity. Mutants with reduced tetrameric assembly are unable to relocalize in response to tonicity changes, which may reflect a requirement for tetramerization in the regulation of AQP relocalization or key protein-protein interactions mediated by loop D of AQP4. We conclude that loop D interactions may represent a conserved mechanism for controlling oligomerization across the AQP family.</p><!><p>A. C. C., P. K., M. T. C., and R. M. B. designed all experiments. P. K. performed and analyzed all experiments and prepared figures. P. K. and A. C. C. drafted the manuscript. R. M. B. and M. T. C. critically revised the manuscript. All authors approved the final version of the manuscript.</p><!><p>The authors report no conflicts of interest.</p><p>aquaporin</p><p>Forster resonant energy transfer</p><p>fluorescence recovery after photobleaching</p><p>transmembrane domain</p><p>Madin-Darby canine kidney cells</p><p>blue native-PAGE</p><p>2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol</p><p>N-[2-hydroxy-1,1-bis(hydroxymethyl)ethyl]glycine</p><p>analysis of variance.</p>

PubMed Open Access

The non-bilayer lipid MGDG stabilizes the major light-harvesting complex (LHCII) against unfolding

In the photosynthetic apparatus of plants a high proportion of LHCII protein is needed to integrate 50% non-bilayer lipid MGDG into the lamellar thylakoid membrane, but whether and how the stability of the protein is also affected is not known. Here we use single-molecule force spectroscopy to map the stability of LHCII against mechanical unfolding along the polypeptide chain as a function of oligomerization state and lipid composition. Comparing unfolding forces between monomeric and trimeric LHCII demonstrates that the stability does not increase significantly upon trimerization but can mainly be correlated with specific contact sites between adjacent monomers. In contrast, unfolding of trimeric complexes in membranes composed of different thylakoid lipids reveals that the non-bilayer lipid MGDG substantially increases the mechanical stability of LHCII in many segments of the protein compared to other lipids such as DGDG or POPG. We attribute these findings to steric matching of conically formed MGDG and the hourglass shape of trimeric LHCII, thereby extending the role of nonbilayer lipids to the structural stabilization of membrane proteins in addition to the modulation of their folding, conformation and function.

the_non-bilayer_lipid_mgdg_stabilizes_the_major_light-harvesting_complex_(lhcii)_against_unfolding

<!>Results<!>Discussion<!>Methods<!>Reconstitution of LHCII into liposomes. Lipid aliquots of pure POPG (purchased from Avanti Polar<!>SMFS experiments.<!>SMFS data analysis.<!>Structure representation.

<p>The major light-harvesting complex (LHCII) found in the chloroplasts of green plants contains more than half of the chlorophylls (Chl) and is the most abundant membrane protein on earth. LHCII plays a key role in photosynthesis by collecting sunlight and efficiently transferring excitation energy to the reaction centers. At the same time, LHCII serves to protect the photosynthetic apparatus from damage of excessive light and to organize the grana structures of the thylakoid membrane 1 . The 25 kDa LHCII apoprotein contains three transmembrane and two amphiphilic helices, non-covalently binding 8 Chl a, 6 Chl b, 4 carotenoid and 2 lipid molecules in a densely packed arrangement as revealed by crystal structure analysis at 2.5 Å resolution 2,3 . LHCII monomers are assembled into trimeric complexes by the end of the greening process of the thylakoid membrane 4 , although trimerization has been proposed to be reversible upon illumination 5 . Trimerization prevents LHCII from proteolysis under high light conditions 6 and increases its thermal stability 7 . Inspection of the LHCII crystal structure allowed to localize the contact sites between the monomers 2 , in part confirming earlier biochemical data 8,9 . However, these findings alone provide little information about how much individual structural features contribute to the stability of the LHCII trimer or monomer.</p><p>LHCII is embedded in the thylakoid membrane, a membrane with an unusual architecture and composition. The thylakoid membrane contains 80% uncharged glycolipids (30% digalactosyl diacylglycerol (DGDG) and 50% monogalactosyl diacylglycerol (MGDG)), and 20% negatively charged lipids (10% each of phosphatidyldiacylglycerol (PG) and sulfoquinovosyl diacylglycerol (SQDG)) 10 . Although the alkyl chains of DGDG are highly unsaturated (mainly 18:3, 18:3), DGDG still has a cylindrical shape due to sterical compensation by its bulky head group containing 2 galactose rings. By contrast, MGDG comprises only 1 galactose moiety and therefore adopts a conical shape with a tendency to form a hexagonal phase 11 . On the one hand, DGDG as well as the fraction of negatively charged thylakoid lipids, modelled here by 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoglycerol (POPG), form a lamellar phase (Fig. 1a,b). On the other hand, MGDG is a non-bilayer lipid, and even the thylakoid membrane with a MGDG fraction of 50% is unable to adopt a lamellar structure unless proteins are inserted into the membrane 12 . The role of the high content of MGDG in the thylakoid is still unclear. Proposed contributions of MGDG include structural aspects like the mediation of spontaneous curvature, the balance of excess membrane area, and specific tasks in protein functionality [13][14][15][16][17] . The lipid composition of chloroplasts, especially the balance of non-bilayer and bilayer lipids, is very sensitive to various kinds of external stress and a main pathway for plants to deal with changing environmental cues [18][19][20][21] .</p><p>Another intriguing feature of the thylakoid membrane is its high protein density. 70-80% of the membrane area is occupied by proteins, with LHCII comprising more than 70% of the proteins in the grana structures 22 . LHCII in particular was found to be essential for the integration of the non-bilayer lipid MGDG in high quantities into the lamellar thylakoid membrane 12 . Conversely, the presence of non-bilayer lipids in lamellar membranes is known to cause alterations in the lateral pressure profile 23,24 , which in turn can influence the structure and function of transmembrane proteins like LHCII [25][26][27] . Based on energy transfer experiments in liposomes it has been hypothesized that lateral pressure exerted by MGDG facilitates the structural association of LHCII and photosystem II 28 . Spectroscopic measurements revealed that the structure of LHCII is sensitive to the molecular environment surrounding the protein 29,30 , thus one might expect MGDG to have a notable impact on the structure and (1-3). The degree of saturation in the alkyl chain region and the size of the head group determine the shape of lipids. Hence, POPG (blue) and DGDG (orange) adopt a cylindrical shape, forming lamellar membranes (1,2). Due to its conical shape the non-bilayer lipid MGDG (red) causes alterations in the lateral pressure profile upon incorporation into lamellar membranes 23,24 ; here: lipid ratio of DGDG / MGDG = 2:1 (3). (c) Schematic representation of SMFS with LHCII embedded in lipid bilayers with compositions according to (b). Sitedirected unfolding of LHCII is achieved by covalently attaching a gold-coated AFM tip to a cysteine motif (yellow) at the amino-terminus N of the polypeptide. Retraction of the tip applies a pulling force F to LHCII monomers, which get extracted from the membrane out of their trimeric assembly (transparent presentation), thereby inducing stepwise unfolding of the protein; gray: polypeptide, dark grey: pigments (obtained with Chimera and PDB ID code 2BHW). (d) Exemplary force-distance curve recorded during the unfolding process described in (c). Each force peak, representing an unfolding event of LHCII, is fitted with the WLC model (yellow) which provides information about the positions of the stabilized domains (contour lengths) and the forces needed to overcome these barriers along the LHCII polypeptide. stability of LHCII. However, the presence of MGDG in liposomes has not been found to influence the thermal stability of LHCII trimers significantly 31 . In this context, new methodical approaches are required to shed light on a potential effect of lateral pressure stabilizing LHCII.</p><p>Single-molecule force spectroscopy (SMFS) is an excellent technique to study the stability of individual domains of a transmembrane protein by measuring the force needed to unfold them, while pulling them out of the membrane [32][33][34][35][36][37][38][39][40][41][42] . Whereas thermal denaturation experiments address the overall protein structure, SMFS provides information about subdomains by sequentially unfolding them. SMFS has also been demonstrated to be a suitable technique to characterize lipid-protein interactions 36,37 . Recently, Scheuring and coworkers were able to apply SMFS to bacterial LH2 elucidating the free energy of oligomerization 38 . In the present work, SMFS has been used to study the mechanical stability of LHCII inserted in supported membranes as a function of lipid composition and oligomerization state. This allowed us to analyze the relationship between the structural stability of the protein and the physicochemical properties of its membrane environment, with particular emphasis on lateral pressure caused by MGDG (Fig. 1b,c). Furthermore, we investigated both trimeric and monomeric LHCII in order to localize the regions specifically contributing to trimer stability. Our findings indicate that MGDG matches the hourglass shape of LHCII thereby providing stability against unfolding.</p><!><p>Mechanical unfolding of LHCII. For atomic force microscopy (AFM) based unfolding experiments, recombinant LHCII was reconstituted into preformed liposomes, which were spread on a flat mica surface. Topographical AFM images of the supported lipid bilayers revealed connected bilayer patches (Supplementary Figure 1). The glycolipids (pure DGDG or DGDG/MGDG mixtures) showed a surface coverage of approximately 30%, while POPG had an increased surface coverage of 60%. Proteins were visible as elevated objects either ~1 nm or ~1.5 nm higher than the bilayer, approximately corresponding to lumenal and stromal protrusions of LHCII, respectively 3 , with diameters on the order of 10 to 100 nm. Apparent differences in protein organization between different lipid compositions were not observed. A Cys 3 cluster near the N-terminus of the LHCII apoprotein allowed stable and selective attachment to gold-coated AFM cantilever tips suitable for SMFS experiments (Fig. 1c). Insertion of recombinant LHCII in artificial membranes presumably results in random up-down orientation; however, the N-proximal Cys 3 attachment site permits to select those complexes for unfolding that expose their stromal surface 32 . The cantilever tip was brought into contact with the membrane surface and was then retracted to unfold the peptide chain as indicated by the saw tooth signature, characteristic for protein unfolding (Fig. 1d), with each force peak representing unfolding of an individual segment of the polypeptide chain 34,35 . Only force-distance curves showing unfolding of the complete peptide chain were taken into account for further analysis as indicated by a force peak in a 10 nm interval around the contour length of the LHCII monomer estimated to be 88 nm. Each individual unfolding peak was fitted using the worm-like chain (WLC) model assuming a persistence length of 0.4 nm 43,44 to assess the contour length l C of the unfolded segment and the corresponding unfolding forces (Fig. 1d).</p><p>Frequent observation of a particular unfolding peak is usually linked to stable structural elements of the mechanically challenged protein that need not necessarily be identical with secondary-structure components 33 . Interestingly, the histogram of all contour lengths obtained by the WLC fits (Fig. 2a, shown as histograms on the x-axes) revealed a very broad distribution with unfolding events almost evenly spread over the full length of the peptide chain. In contrast to unfolding experiments on numerous other proteins distinct peaks were absent (as illustrated in Supplementary Figure 2) 32,36,37,[39][40][41][42] . The broad distribution might be attributed to the complex structure of LHCII forming a superhelix in combination with a high number of bound auxiliary pigments (see Fig. 1c) 2,3,45 . In contrast to the contour length histograms, the corresponding unfolding force histograms are less broad and show a clear preference for unfolding forces around 50 pN dependent on the lipid matrix and the oligomerization state (Fig. 2a, shown as histograms on the y-axes).</p><p>The common procedure for analysis of unfolding force curves is to map recurring unfolding peaks to stable elements and categorize every detected unfolding event accordingly. Due to the broad distribution of detected contour lengths such an analysis would be very unreliable and prone to false assignment of peaks to structural elements. Therefore, in order to circumvent categorization of unfolding events to structural elements, we estimated the probability to find a certain unfolding force at a given contour length as obtained from the WLC fits by using a multivariate kernel density estimate (Fig. 2a). Analysis of this "stability map" then allowed us to relate the mechanical response of the polypeptide to the stabilization of structural elements depending on different intrinsic and environmental conditions (see below). In order to compare unfolding forces between different experiments we considered the most likely unfolding force as a function of the contour length (colored dots in Fig. 2a,b). To find the most prominent contributions to protein stability, only unfolding events with a likelihood of at least 40% of the most probable unfolding event were selected. This also sorts out unfolding events with unfolding forces above 100 pN, which we attribute to simultaneous unfolding of more than one monomer attached to the cantilever in parallel.</p><p>Impact of LHCII trimerization. LHCII monomers inserted in PG-containing membranes tend to trimerize spontaneously 46 ; therefore, in order to address the impact of trimerization on the mechanical stability of LHCII, a LHCII variant was used with two point mutations in its trimerization motif (positions 16 and 17) that render the protein incapable of trimerization, as described earlier 8 . For comparing LHCII monomers and trimers, the complexes were inserted into membranes consisting of POPG exclusively. The distribution of unfolding forces (Fig. 2a, shown as histograms on the y-axes) reveals that the most frequent unfolding force is increased by around 10 pN from monomers to trimers, demonstrating a slightly stabilizing effect of trimerization on LHCII. This is in agreement with previous thermal denaturation experiments revealing an increased stability of the trimer 7 .</p><p>Comparing the most likely unfolding force at each contour length (colored dots in Fig. 2a top left, top right) for monomeric LHCII and a single LHCII polypeptide unfolded out of the trimer reveals selective stabilization in the trimer. In some contour length intervals unfolding forces in the trimer are similar to those in the monomer, whereas in other parts of the polypeptide -although most of the disparities are not significant unless stated otherwise-the forces are higher for trimers than for monomers (Fig. 2b top). While the N-terminal hydrophilic domain at contour lengths of l C ≈ 10 nm shows no detectable unfolding events for the monomer, there is a high probability to observe unfolding events for trimeric LHCII (Fig. 2b top, indicated by 1). Along those lines, unfolding forces of the trimer were generally higher from l C ≈ 27 nm to l C ≈ 34 nm (∆F ≈ 8 pN), at l C ≈ 47 nm (∆F ≈ 7 pN), l C ≈ 77 nm (∆F ≈ 12 pN), and for the C-terminus at l C > 82 nm (∆F ≈ 8 pN, same figure, indicated by 2, 3, 4 and 5, respectively). Interestingly, in two segments of the protein, unfolding forces were slightly higher for the monomer than for the trimer (from l C ≈ 15 nm to l C ≈ 25 nm and from l C ≈ 65 nm to l C ≈ 70 nm).</p><p>Stabilization of the trimer compared to the monomer, as indicated by higher unfolding forces, can be linked to interactions in distinct protein domains facilitating LHCII trimerization. Both the N-terminus and the C-terminus have previously been described to be important for trimerization based on mutation analysis 8,9 and structural data 2 , which is mirrored here in the significant stabilization at l C ≈ 10 nm (peak 1) and the stabilization at l C > 82 nm (peak 5). The key trimerization motif has been identified as WYGPDR at aa17-aa22 (l C = 6-8 nm) 8 . However, detaching the cantilever tip from the membrane surface causes adhesion peaks at the beginning of each force-distance curve that may obscure unfolding events in the very proximal segment of the N-terminus. Hence, the precision in determining contour lengths in this domain is not sufficient to be sure that the peak at l C ≈ 10 nm is in fact due to the trimerization motif.</p><p>A critical factor for trimerization is the binding of pigments and lipids. Liu et al. pointed out an important role of helix 3 by coordinating several Chl b molecules, mediating hydrophobic interactions between adjacent monomers 2 . Here, we may observe a contribution to trimer stability by one chlorophyll in particular: Chl8 (nomenclature according to Standfuß et al. 3 ) is bound to two different monomers via coordination of the central Mg 2+ by His212 (l C = 77 nm, corresponding to peak 4 within helix 5) of one monomer and hydrophobic interaction of the phytol tail with Trp128 (l C = 47 nm, corresponding to peak 3 at the beginning of helix 3) of the other monomer (Supplementary Figure 3). In addition, Trp222 (l C = 81 nm) has been described to be a sensitive site for trimer formation and stability 9,45 , which may explain an enhanced stabilizing effect at peak 4 compared to the other regions.</p><p>Liu et al. identified Chl4 and Chl5 bound via Glu65 (l C = 24 nm) and His68 (l C = 25 nm), respectively, to be crucial for the trimer structure 2 . However, unfolding forces of trimeric and monomeric LHCII show little differences at these contour lengths. Strong intramolecular interactions, like the salt bridge at the helix cross between helix 1 at Arg70 (l C = 25 nm) and helix 3 45 , may be dominating in this segment, obscuring the effect of neighboring monomers as discussed by Sapra et al. for bacteriorhodopsin assemblies 47 .</p><p>There is no straightforward assignment of the significant stabilization between l C ≈ 27 nm and l C ≈ 34 nm (region 2) to the structure of LHCII. This contour length interval corresponds to the lumenal end of the transmembrane helix 1 and the amphiphilic helix 2. However, Liu et al. found a highly reduced degree of trimerization after an exchange of W97 to alanine (l C = 35 nm) at the beginning of helix 2 48 , representing a binding site for lutein 2 45 . This pigment has been suggested to assume a trimer specific conformation 7 , indicating that lutein 2 and hence W97 might contribute to the stability of trimeric LHCII to a certain extent as reflected by our data.</p><p>It seems surprising that we observed one position at the N-terminal domain close to helix 1 and one at the beginning of helix 4 with higher unfolding forces for the monomer than for the trimer. The PG molecule bound to LHCII via Tyr44 (l C = 16 nm) and Lys182 (l C = 66 nm) 3 has been identified to be a major factor in trimerization 49,50 . Moreover, there are indications that the monomeric complex alone is not able to bind PG 50 . While PG facilitates trimer formation, electrostatic repulsion between the negatively charged head groups of the bound lipid and the membrane lipids is energetically unfavorable and might destabilize the trimeric protein at these positions in a POPG matrix.</p><p>Effect of lipid matrix. To assess the impact of the lipid matrix on the mechanical stability of LHCII we performed unfolding experiments with trimeric LHCII in three different lipid matrices: pure POPG, pure DGDG and a mixture of DGDG and MGDG (2:1). Lipid mixtures containing more than roughly 30% MGDG no longer form lipid vesicles 51 , therefore we abstained from using higher percentages of MGDG, expecting that if the lateral membrane pressure has an effect on LHCII, this is seen at 30% MGDG. Since neither of the bilayer lipids POPG and DGDG affects the lateral pressure profile of a lamellar membrane, differences in unfolding forces between these lipid environments can primarily be attributed to the charge of the lipid head groups (Fig. 1a,b). Alterations in protein stability upon addition of 30% non-bilayer lipid MGDG to DGDG membranes provide information about the impact of the modified lateral pressure profile 24 , while the charge of the membrane and the chemical properties of the head groups remain unaffected.</p><p>The distribution of unfolding forces varies only little between the two cylindrical lipids POPG and DGDG (Fig. 2a top left, bottom left; shown on the y-axes). When considering the most likely unfolding force for every contour length, the N-terminus around l C ≈ 10 nm and the region from the end of the stromal loop to the beginning of helix 4 around l C ≈ 60 nm are slightly more stable (∆F ≈ 10 pN) in DGDG compared to POPG (Fig. 2b bottom). The lumenal side of helix 4 from l C ≈ 70 nm to l C ≈ 75 nm is greatly stabilized (∆F ≈ 25 pN) in both glycolipid matrices (DGDG and DGDG + MGDG). This narrow peak points to a protein domain interacting with lipid head groups on the lumenal side of helix 4.</p><p>Adding MGDG to the DGDG membrane significantly raises the overall occurrence of unfolding forces above 70 pN (Fig. 2a bottom left, bottom right; shown on the y-axes) and strongly enhances the unfolding forces in several segments of the protein (Fig. 2b bottom). Unfolding forces are increased by up to 20 pN from l C ≈ 6 nm to l C ≈ 12 nm close to the N-terminus, from l C ≈ 24 nm to l C ≈ 32 nm within helix 1, from l C ≈ 34 nm to l C ≈ 46 nm comprising helix 2, the lumenal loop and part of helix 3, and from l C ≈ 50 nm to l C ≈ 54 nm at the stromal side of helix 3. In contrast to the protein stabilization due to trimer formation, the stabilized regions along the peptide chain due to the surrounding lipids are rather broad extending over up to 30 amino acids, indicating a regional stabilization rather than local interactions between individual amino acids.</p><!><p>Unfolding monomers and trimers reveals a connection between structure and stability of LHCII, despite the structural intricacy of the pigment-protein complex. Site-directed and complete unfolding of recombinant LHCII by SMFS was facilitated by linking the cantilever to an N-proximally introduced Cys 3 cluster 32 . In contrast to other transmembrane proteins studied by SMFS 32,36,37,[39][40][41][42] , unfolding of different LHCII peptides does not always follow the same unfolding trajectory as indicated by the notable absence of clear peaks in the unfolding contour length histograms (Fig. 2a; see also Supplementary Figure 2). Since the broad distribution of contour lengths could in principle also arise from careless data evaluation or experimental errors, a number of quality control measures were taken to minimize the impact of those effects. To this end, the protein was site-specifically linked to the cantilever at the amino-terminus, with the aim to reduce the impact of nonspecific adhesion and random pulling from the stromal or the lumenal side during the unfolding experiments. Notably we do not observe any symmetry in the most frequent unfolding forces along the LHCII polypeptide, which would indicate stochastic pulling from either the N-or the C-terminus (Fig. 2b). Nonspecific attachment of the cantilever to the stromal or lumenal loop leads to force-distance curves with unfolding events at tip-sample separations much shorter than 78 nm, which were excluded from analysis. Similarly, force-distance curves lacking the characteristic WLC shape of protein unfolding were not analyzed. Additionally, the AFM setup used here has low instrumental noise with a standard deviation of 4 pN due to rigorous screening against external acoustic sources.</p><p>Therefore, we ascribe our findings to the intricate structure of the protein. The two tilted helices 1 and 4 are interlocked by two salt bridges, forming a super-secondary structure 3,45 (see Fig. 1c). In addition, the high number of pigments coordinated by LHCII (18 per monomer) leads to added links between different segments of the polypeptide, each capable to stabilize the protein structure. Considering more than 20 known pigment binding sites 45 and the contour length of the LHCII monomer of 88 nm, the average distance along the peptide chain between two pigment binding sites is around 4 nm. Assuming an experimental error for the determination of the contour length of an unfolded segment of 2 nm (as a lower bound), it becomes clear that adjacent pigment binding sites cannot be resolved with the current experimental capabilities according to this simple calculation.</p><p>By analyzing the most likely unfolding force at every position along the polypeptide chain we avoided to focus on selected structural elements in the protein originating from the frequency of contour lengths only. We used this approach to analyze unfolding curves of the LHCII trimer and an LHCII point mutant incapable of trimerization, revealing localized differences in unfolding forces between monomeric and trimeric LHCII which can be directly linked to the protein structure. Although not at all positions along the contour of the protein statistically secured, we find this approach for data analysis reasonable as the disparities in unfolding forces due to trimerization can be largely correlated with previously published structural and biochemical data. We therefore relied on the method to further explore how the stability of LHCII is modulated by the lipid matrix surrounding the protein. In contrast to the rather subtle impact of oligomerization on the unfolding forces of LHCII we found that the lipid matrix has a substantial influence on protein stability as discussed below.</p><p>Lateral pressure by MGDG greatly stabilizes trimeric LHCII due to its hourglass shape. The regions of the LHCII trimer that are stabilized in DGDG compared to POPG membranes are mainly located in the extra-membrane domains of the protein in plane with the lipid head groups, suggesting a destabilizing impact of the negatively charged phosphate group and/or stabilizing effects by the galactose moieties (Fig. 2b bottom). The N-terminal region from l C ≈ 11 nm to l C ≈ 13 nm is rich in polar amino acids (Ser29, Ser32, Ser34, Tyr35 and Thr37) capable of forming hydrogen bonds with the sugar residues, which may lead to an increased stability by both glycolipids (MGDG and DGDG) in this segment of the polypeptide. Negatively charged amino acids (Asp162, Asp168, Asp169 and Glu171) in the stromal loop from l C ≈ 59 nm to l C ≈ 62 nm may generate repulsive ionic interactions with phosphate head groups and thus destabilize the protein at the end of the stromal hydrophilic domain in a POPG matrix. In addition, repulsive forces due to the bound PG-molecule at l C ≈ 66 nm may extend the destabilization of this segment to the stromal side of helix 4, similar to the statement made above for the comparison of the monomer and the trimer 3 . The greatly enhanced stabilization of LHCII against unfolding by both glycolipid mixtures at the lumenal side of helix 4 around l C ≈ 75 nm probably arises from hydrogen bonds between the corresponding polar amino acids (Gln197, Thr201 and Lys203) and the galactose rings of either DGDG or MGDG, as the presence of positively charged Lys203 renders a repulsive ionic interaction with the POPG head group highly unlikely. However, besides the vast chemical differences between POPG and DGDG, especially considering the different charges, disparities in unfolding force are minuscule.</p><p>By contrast, large differences are seen in unfolding forces between the DGDG/MGDG mixture and pure DGDG, despite the similarities between DGDG and MGDG with regard to charge and functional groups. This rules out specific or electrostatic interactions as a reason for the mechanical stabilization. The stabilizing effect of MGDG is not restricted across the membrane plane since it includes transmembrane as well as extra-membrane domains of LHCII (shown as schematic representation in Fig. 3a). A view on the stromal and lumenal faces of LHCII reveals that the stabilization is confined to the periphery of the trimer where it is in direct contact with the lipid matrix (Fig. 3b). We therefore attribute the stabilization to steric interactions between the hourglass shape of trimeric LHCII and the conical shape of MGDG (Fig. 3c), pointing to bulk physical properties of the lipid since it is not part of the LHCII structure 2 . Our results are in accordance with recent simulations demonstrating that concavely formed proteins are energetically favored in membrane domains exhibiting negative curvature stress 52 , which is a measure for alterations in the lateral membrane pressure profile induced by non-bilayer lipids such as MGDG 24,26 . Yang et al. monitored the dissociation of LHCII trimers into monomers as a function of thylakoid lipid composition, showing only a slight effect of MGDG 31 . This is in agreement with our data as the impact of MGDG is particularly intense in the periphery of LHCII (Fig. 3b), whereas trimerization is mainly based on pigment mediated interactions in the hydrophobic core 2 .</p><p>A lot of attention has been paid to how non-bilayer lipids, particularly phosphatidylethanolamine (PE) representing up to 38% of the inner mitochondrial membrane lipids 53 , modulate the folding, conformation and functionality of integral membrane proteins 26,27 (see also references therein). However, only little is known about their impact on the stability of transmembrane proteins, even though complementary shapes of transmembrane proteins and lipids appear to be a widespread phenomenon. The concavely formed ion channel gramicidin and the non-bilayer lipid lysophosphatidylcholine as well as the potassium channel KcsA and PE are prominent examples for the proposed steric interplay 54,55 . Unfolding of LHCII via SMFS allowed us to observe a stabilizing effect of complementary lipid and protein shapes (as illustrated in Fig. 3c), thereby displaying the impact of the lateral pressure profile over the whole membrane cross section (Fig. 3a). While LHCII is needed to stabilize MGDG in a lamellar phase, our data clearly show that vice versa MGDG increases the mechanical stability of LHCII. Our results will contribute to the ongoing discussion of the role on non-bilayer lipids in many biological membranes.</p><!><p>Preparation of recombinant LHCII. The LHCII variants used in this study were derivatives of Lhcb1*2 (AB80) gene from pea (Pisum sativum) 56 containing a C-terminal hexahistidine (His 6 ) tag and having the native cysteine at position 79 replaced with serine. In order to achieve site-specific unfolding a cysteine motif (Cys-Cys-Cys) was added at the N-terminus of the protein by using the Phusion site-directed mutagenesis kit (Thermo Scientific). This version also served as a template to construct the LHCII variant WY16,17AV 8 by using the QuikChange Lightning site-directed mutagenesis kit (Agilent Technologies), which was applied for the monomer studies. The genes were inserted in the pDS12-RBSII vector and overexpressed overnight in Escherichia coli strain JM101, followed by isolation of the apoprotein according to established protocols 57 . The apoprotein was reconstituted to monomeric LHCII with total pigment extract from pea thylakoids 58 by the detergent exchange method 59 . His 6 tag mediated trimerization of the monomers was performed on Ni 2+ -sepharose columns 60 , followed by purification via ultracentrifugation through a sucrose density gradient as described earlier 61 , but using a different buffer (500 mM sucrose, 10 mM Tris-HCl pH 7.5, 5 mM TCEP, 0.05% (w/v) Triton X-100). In case of the variant WY16,17AV the trimerization step was omitted and the reconstitution mix was directly loaded on the sucrose gradient containing a higher concentration of Triton X-100 (0.1%). The bands corresponding to LHCII trimers or monomers (in case of WY16,17AV) respectively were collected and checked for correct assembly via CD spectroscopy on a J-810 spectropolarimeter (Jasco) in the visible range, with trimeric LHCII exhibiting a characteristic negative peak at 474 nm 61 . For insertion into preformed liposomes the buffer was exchanged to AFM buffer (50 mM NaCl, 10 mM Tris-HCl pH 7.5, 5 mM TCEP) using Amicon 30-kDa centrifugal filters, while the suspension still contained Triton X-100.</p><!><p>Lipids), pure DGDG or a mixture of DGDG and MGDG (each lipid purchased from Lipid Products) at a molar ratio of 2:1 were dissolved in chloroform. The chloroform was slowly removed in a rotary evaporator by which a thin lipid film was formed on the inner wall of a glass flask. After complete removal of the solvent under high vacuum at 40 °C for 1 h, the lipid film was hydrated by rigorous mixing (3 × 30 s at 60 °C, MS2 minishaker, IKA) in AFM buffer at a total lipid concentration of 1.5 mM to produce preformed liposomes. For MGDG containing samples the mixing step was repeated several times until no more lipid traces could be detected by eye on the bottom of the flask. Sonication in a tip sonicator (Vibra cell, Sonics & Materials) for 4 min, followed by 3 freeze-thaw cycles, yielded large unilamellar vesicles. The vesicle suspension was then extruded 21 times (LipoFast-Basic, Avestin) through a polycarbonate membrane (pore diameter 100 nm) and mixed with Triton X-100 to a final concentration of 0.05% (w/v). Reconstitution of recombinant LHCII into preformed liposomes was performed according to established protocols 31 , but with some modifications: the protein suspension was added dropwise under continuous mixing at 4 °C in the dark to a molar lipid/protein ratio of 500-1000, and subsequently polystyrene beads (Bio-Beads SM-2; Bio-RAD) at 30 mg/ml were added. In order to remove Triton X-100 the mixture was incubated overnight under constant rotating at 4 °C in the dark, and the supernatant was removed to a new tube containing fresh Bio-Beads, followed by incubation for 1 h. The last step was repeated, and finally the supernatant was collected. The inserted LHCII trimers or monomers were checked for integrity on a FluoroMax-2 spectrometer (Horiba Scientific) as described earlier 60 .</p><!><p>Prior to the SMFS experiments, proteoliposomes were extruded 21 times (LipoFast-Basic, Avestin) through a polycarbonate membrane (pore diameter 100 nm). CaCl 2 solution (100 mM) was added to a final concentration of 5 mM and the suspension (60 µl) was incubated for 10 min on freshly cleaved mica, leading to surface supported bilayers containing LHCII proteins (Supplementary Figure 1). Nonadsorbed vesicles were removed by intensive rinsing (4 × 1 ml) with AFM buffer. SMFS measurements were conducted in AFM buffer at room temperature using a commercial atomic force microscope (MFP-3D Infinity, Oxford Instruments Asylum Research) and gold-coated silicon nitride cantilevers (Biolever, Olympus). Each cantilever was calibrated with the thermal noise method yielding spring constants around 6 pN/nm. The cantilever was moved toward the surface at 500 nm/s until a load force of 100 pN was reached. The tip was kept in contact with the membrane for 1 s to allow for binding of the Cys 3 -motif at the N-terminus of LHCII to the gold-coated cantilever tip. The cantilever was then retracted at 1000 nm/s to unfold the peptide, giving rise to a force-distance curve. For each sample only a single force-distance curve was collected within an area of 1 µm 2 and each area of the supported bilayer was only scanned once. Roughly 10% of the curves showed force peaks corresponding to unfolding events of the protein, of which ~7% exhibited a force peak corresponding to a contour length of 88 ± 10 nm indicating full unfolding of the LHCII polypeptide. Only the latter were selected for further analysis, thus more than 25,000 force-distance curves were recorded for each experimental condition to obtain a sufficient number of curves showing full unfolding (monomer in POPG: n = 193 out of ~30,000 curves; trimer in POPG: n = 328 out of ~50,000 curves; trimer in DGDG: n = 262 out of ~50,000 curves; trimer in DGDG/MGDG: n = 195 out of ~25,000 curves). The SMFS measurements were repeated at least 3 times for each experimental condition.</p><!><p>Only force-distance curves showing the saw tooth signature characteristic for protein unfolding with an unfolding peak corresponding to a contour length of 88 ± 10 nm were selected for further analysis. Each individual unfolding peak was fitted using a WLC model 43,44 :</p><p>2 with the force F x ( ) at tip-sample separation x, the Boltzmann constant k B , the persistence length = b 4 Å, the contour length of the unfolded peptide l C and the temperature = T 298 K. Additionally, the peak force F max of each unfolding peak was recorded. To estimate the distribution of the value pairs of l C and F max for all recorded unfolding events, we calculated a multivariate kernel density estimate using the approximated experimental uncertainty σ = l ( ) 2 nm C and σ = F ( ) 12 pN as kernel bandwidths, as estimated from the standard deviation of the measured force during the baseline of the force-distance curve, i.e. recorded far away from the surface. The uncertainty of the contour length was calculated from the uncertainty of the measured force and the spring constant. This gives us an unfolding force distribution at any given contour length. The maxima of these distributions give us the most likely unfolding force ⁎ F as a function of the contour length. The most likely unfolding force was evaluated at discrete contour lengths with a fixed distance between two points Δl C = 1.25 nm. As an estimate of the uncertainty of the most likely unfolding force we calculated the standard error of the most likely unfolding force σ ⁎ F ( ) for each contour length l C as σ = and the number of data points in that interval N .</p><!><p>The LHCII structure (PDB ID 2BHW) 3 representation was created using the UCSF chimera package 62 .</p>

Scientific Reports - Nature

Chemical Reaction Stoichiometry: A Key Link between Thermodynamics and Kinetics, and an Excel Implementation

Various special methods are generally taught for the construction of stoichiometric equations for simple reacting systems, including inspection, oxidation-reduction, and ion-electron approaches, which typically fall under the topic of "balancing a chemical equation". However, apart from the simplest cases described by single chemical equations, only matrix methods can express the Law of Conservation of Mass (LCM) by generating a set of multiple stoichiometric equations of the appropriate number. Chemical Reaction Stoichiometry (CRS) represents the Law of Conservation ofMass (LCM) in the context of chemical reaction, both in and out of its chemical equilibrium state; it may be expressed as a set of linear equations involving the conservation of mass, of chemical elements, or of another appropriate quantity, for example chemical moieties or genetic units. CRS expresses this information in terms of a set of stoichiometric equations, which are equivalent representations of the LCM and have the appearance of, but are distinct from, actual chemical reaction mechanisms. These equations may be determined by algorithms implemented in symbolic mathematics programs or by purpose-specific computer programs. In order to provide a more generally accessible means of implementing the matrix method, we provide a Microsoft Excel implementation, XSTOICH, as a Supplementary file that contains many examples.After first briefly reviewing the matrix method for the standard case based on a set of chemical species and their molecular formulae, we show how to incorporate complications arising from stoichiometric restrictions such as those imposed by a kinetic model of the system. The need for background mathematical knowledge is avoided by the provision of simple worked motivating examples in the text. The chemical motivation for more complex problems is made clear, and their solution is illustrated by means of the simple input procedures of XSTOICH.

chemical_reaction_stoichiometry:_a_key_link_between_thermodynamics_and_kinetics,_and_an_excel_implem

Introduction<!>The Matrix Method for a One-Reaction System<!>The Universal Matrix Method (UMM)<!>Stoichiometric Restrictions: Explicit and Implicit<!>Explicit restriction example

<p>The Law of Conservation of Mass (LCM) is fundamental to chemistry. It applies to all closed chemical systems as well as to open flow systems (those with both inlets and exits) at steady state; we refer to both situations as closed systems in the following. The LCM can be expressed most directly by means of a set of linear equations involving the species amounts, representing the conservation of the atom types in the system (including the conservation of electronic charge, which is incorporated by imposing a value of zero for its conserved value). This is equivalent to requiring that changes in the amounts of the conserved quantities which accompany arbitrary changes in the species amounts from a given composition must always be zero. The LCM may equivalently be expressed in the form of "chemical equations", which have the appearance of, but are distinct from, actual chemical reaction mechanisms.</p><p>Chemical Reaction Stoichiometry (CRS) is the study of the inter-relationships between the two expressions of the LCM. Although these inter-relationships can provide insights into the underlying chemical phenomena, the topic is not a component of traditional undergraduate chemistry pedagogy, in spite of the fact that it is based on the use of relatively simple mathematical manipulations.</p><p>In undergraduate chemistry education, reaction kinetics and equilibrium thermodynamics are two important topics, traditionally taught as separate subjects. However, they share the common link of the LCM. Not only must the overall reaction conserve mass, but each reaction that appears in a chemical kinetic description of a system must individually satisfy the LCM.</p><p>The topics are further connected by the fact that reaction processes in a closed system under a given set of overall thermodynamic conditions (for example, specified temperature and pressure) evolve with time towards a unique chemical equilibrium composition. Although the details of a system's reaction mechanism and its kinetic description may be exceedingly complex or, indeed, even unknown, the equilibrium composition itself is independent of such considerations, and is determined by the global minimum of the system's Gibbs function.</p><p>Thus, the calculation of chemical reaction equilibrium requires only species reference state Gibbs energy information in conjunction with the specification of the LCM for the system and an appropriate chemical potential model. This also means that the limiting compositional state of any chemical kinetic model for such systems must coincide with the equilibrium composition. In addition, a link may be made from a chemical kinetic model to its implied LCM conservation equations using the concepts of Chemical Reaction Stoichiometry (CRS).</p><p>For the aforementioned reasons we believe that a focus on the pedagogical treatment of the topic of CRS itself would serve to unify the treatment of chemical equilibrium and of chemical kinetics within the field of chemistry and, by avoiding its duplication in each, provide a more efficient pedagogical approach to both topics. We also note in passing the increasing importance of CRS concepts in systems biology, 1,2 providing an additional field that could benefit from a common treatment of the topic.</p><p>The commonest pedagogical application of the LCM in a system undergoing chemical change has been the calculation of the chemical equilibrium in closed systems, for which the LCM may be expressed in terms of sets of equations that have the superficial appearance of describing actual chemical reaction processes. This emphasis, and traditional applications to relatively simple systems that require only a single such equation, have both obscured their basic purpose and led to the teaching of a myriad of methods for "balancing a chemical equation", including "inspection", simple algebra, 3 the oxidation number method 4 and the half-reaction (or ion-electron) method. 5 Only the matrix method, 6 which we herein discuss, can be extended to the ubiquitous case of multi-reaction systems. This approach also addresses the core task of determining the appropriate number, R, of independent chemical equations that are required to implement the LCM; this number is a crucial aspect of chemical equilibrium problems.</p><p>The importance of chemical reaction stoichiometry (CRS) has been explored by one of us in many articles [6][7][8][9][10][11] and a website 12 that have advocated for an increased emphasis on the LCM itself and its application to both chemical reaction equilibrium and to chemical reaction kinetics. Software to implement the general matrix method for the LCM was first published in the 1982 textbook of Smith and Missen, 6 and later (1989) by Missen and Smith 8 in the form of a BASIC program to generate a maximal set of chemical equations from the system's formula matrix. To the best of our knowledge, no other generally available current software provides this universality. However, in order to run the BASIC program in current Windows, Mac and Linux computers, it is necessary to install an emulator 13 which makes the program unacceptable for general use. The method was later coded as a web-based Java Applet, JSTOICH. 14 However, current security considerations mean that Java Applets are no longer accessible through web browsers, except under special circumstances. 15,16 The purpose of this paper is to briefly describe the matrix method for CRS and to present its implementation in Microsoft VBA Excel, which is rather widely available and not subject to browser restrictions. The Excel program is made available as a supplement, XSTOICH, to this manuscript. Examples of its use are illustrated in the paper.</p><p>Although CRS is based upon the mathematics of sets of linear algebraic equations, the need for background knowledge in this area is avoided by the consideration of very simple worked examples, which are easily understood by first-year chemistry students. The chemical motivation for more complex problems is explained clearly, and their implementation is illustrated by simple input entries to XSTOICH.</p><p>The paper is organized as follows. In the next section, we briefly review the matrix method for the determination of a set of chemical equations consistent with the most general form of the LCM that arises when the set of species chemical formulae is specified. We proceed from the simple case of a one-equation system to the general case of multiple chemical equations, demonstrated in XSTOICH on examples. The subsequent section discusses the use of XSTOICH to consider less well-known LCM situations, in which restrictions are placed on its general form. These may arise either from explicit restrictions on the element conservation equations over and above the conservation of the elements, or from implicit restrictions arising from a specified set of chemical equations. In both cases, the restrictions must correspond to experimental observations.</p><!><p>CRS implements the law of the conservation of mass (LCM) in a closed chemical system undergoing chemical change. It expresses the conservation of appropriate entities such as atomic elements, electronic charge, quantities such as group moieties in biological processes being transferred between molecules, 17 and the transfer of genetic information, 1,2 all of which will henceforth be referred to as "elements". The LCM is expressed in terms of the set of homogeneous linear algebraic equations arising from the requirement that changes in the species amounts must result in a net zero change in the amount of each conserved element in the system. This may be represented by a system of homogenous linear equations involving the changes in the species amounts (thus having their right sides equal to zero) and treated by the methods of linear algebra, most conveniently by the use of simple matrix manipulations.</p><p>The matrix method can be most straightforwardly derived by stating that there is zero change in the amount of each chemical element for any set of changes in the molar amounts of the chemical species involved in a closed system. In the case of a single-reaction system, this is the principle underlying the familiar "ICE method" 18,19 used in solving a chemical equilibrium problem involving a single reaction. As an initial simple example, consider the system of coal (in the form of solid carbon), steam and synthesis gas (C(s), H 2 O, CO, H 2 ). The 3×4 system formula matrix, A, is shown in A maximal set of independent solutions of these equations can be found by a standard Gauss-Jordan elimination process 20 implemented in XSTOICH, of generating the diagonal Row-Reduced Echelon Form (RREF) of A, labelled as the matrix, A * , containing a leading 3×3 unit matrix, as depicted in Fig. 1(b). The RREF procedure consists of considering each row in turn, and adding (positive or negative) multiples of rows together, ("swapping" of rows is a special case of this), so as to produce a "1" in the column corresponding to the row and zeros in all other rows of that column. (This process may usefully be followed interactively and step-by-step on the RESHISH Matrix Calculator website 21 ) At the termination of this process, the final column of the RREF matrix in Fig. 1(b) is a solution stoichiometric vector, ν, which may be interpreted as the relationship of the formula vector of the single "noncomponent species" listed in the header to the formula vectors of each of the prior header species (termed the "component species") with their corresponding unit vectors beneath them, in sequence, thus:</p><p>or, replacing the formula vectors by their chemical names</p><p>which re-arranges to the more familiar chemical equation</p><p>In general, the A * matrix is comprised of a leading unit matrix of dimension C followed by a C ×R stoichiometric matrix (a 3×1 vector in this case), where C represents the number of "components species" (or "basis chemical species") in the system and R represents the number of noncomponent species which may be produced from the component species by the R reactions. This example is a very simple one, since it has R = 1; note also that C = M in this case. Finally, note that we could have arranged the species columns in any order and would have arrived at the same resulting Eq. (1) (but the sets of component species and non-component species would be different). The astute reader will also notice that R = N − C in this simple example, a result that carries over to the multiple reaction case. We shall see that the order of the columns affects the nature of the particular set of reactions that are generated when R is greater than 1.</p><!><p>Although simple chemical processes may be represented by a single chemical equation, most realistic cases result in multiple chemical equations. In elementary chemistry, these are sometimes characterised as being "difficult to solve" or having multiple solutions. 22 We believe that this characterization is pedagogically counterproductive since it emphasizes "singlereaction chemistry" (the notion that all reacting systems can be reduced to a single reaction);</p><p>in fact, every one of the possible "multiple solutions" can be written as a linear combination of a set of R independent chemical equations. The fundamental task of CRS is to find the values of C and R and to find a set of R independent chemical equations from the system formula matrix A. The value of R is unique, but the set of independent chemical equations is not, with the only requirement being that the equations are "independent", that is, any individual equation cannot be obtained by a linear combination of the other members of the set.</p><p>Symbolic algebra programs to perform the necessary matrix manipulations on the formula matrix are generally expensive and require some familiarity in use, while Microsoft Excel (and similar programs) are widely available (with some being free of charge) while still requiring some experience in data manipulation. We therefore introduce XSTOICH, a "universal" programmed method in Excel to perform the task. 6,7,9,12 In this method, the formula matrix A is transformed as in Fig. 1 by Gauss-Jordan elimination 21,23 to a "Row-Reduced Echelon form" (RREF), resulting in a matrix A * consisting of a unit matrix followed by stoichiometric coefficient vector(s) from which a set of independent stoichiometric equations may be directly written. Since A * is obtained from A by linear combinations of its rows, the matrices satisfy</p><p>where the rank of a matrix is its number of linearly independent rows or columns. (We remark that XSTOICH is configured to require the formula matrix A to be entered in its transposed form, that is, the columns contain the formula vectors.)</p><p>As a simple example of a multi-reaction system, consider the partial combustion of methane to produce synthesis gas, CO and H 2 , together with CO 2 and H 2 O. The 3 × 6 stoichiometric formula matrix for this system appears in Fig. 2(a). After Gauss-Jordan elimination, we observe in the RREF matrix A * (Fig. 2(b)) that there is a leading 3 × 3 unit matrix, followed by three additional columns; rank(A) is readily determined by counting the number of 1s in the unit matrix, three in this case. Each of the last three columns is a set of coefficient vectors (for the non-component species heading that column) corresponding to the reaction forming one mole of the non-component species from the prior component species indicated above the unit matrix. Thus, the chemical equation for the species</p><p>Multiplying by 2 and re-arranging, we obtain</p><p>The two remaining column vectors produce the equations</p><p>The set of three chemical equations that we have obtained are not unique, and can be replaced by other sets which are linear combinations of the three equations we have found;</p><p>however, the number of three independent stoichiometric equations for this system is unique.</p><p>(As an exercise, the reader may re-run this example in the supplied program XSTOICH with different orderings of species and see what happens.)</p><p>Finally, a consequence of the operations used by RREF on A to obtain A * is that the M × N formula matrix A and the N × R stoichiometric matrix ν derived from it by the RREF procedure satisfy</p><p>and the ranks of the matrices satisfy</p><p>If we let the species formulae stand for their formula vectors in Eq. ( 9), the equation can be expressed in the form:</p><p>where the matrix of numbers is the stoichiometric matrix, constructed from A * by appending the negative of a unit matrix below the right-most part of A * . The notation in Eq. (11) follows the usual matrix multiplication rules, where on the left side, the row of species names is a 1 × N (N = 6 in this case) row vector is multiplied into the columns of the N × R (R = 3 in this case) stoichiometric matrix, yielding a 1 × R row vector of linear combinations of species names, with each linear combination set equal to the corresponding zero in the row vector on the right side.</p><!><p>A complication arises when constraints are imposed on the reactions undergone by a chemical system over and above the conservation of the elements; these typically arise either from experimental observations or are imposed by a kinetic model. When these constraints are equivalent to specifications of linear of species mole numbers, we call these constraints "stoichiometric restrictions".</p><p>Stoichiometric restrictions may arise either explicitly or implicitly. In the former case, one or more additional independent conservation equations on the species mole numbers are imposed, and the goal is to derive a modified system formula matrix A that takes the restriction(s) into account. A stoichiometric matrix is then derived by the UMM in the usual way from A .</p><p>A general result from linear algebra 20 states that for an arbitrary matrix pair (A, ν ) satisfying Eq. ( 9), their ranks satisfy</p><p>where N is the number of columns of A. rank(ν ) is called the number of stoichiometric degrees of freedom, F s , which is the number of values of appropriate species mole numbers (or independent conditions to obtain them) that must be specified in order to determine the values of the remaining mole numbers (see also Smith and Missen, 6 Section 2.4.3). A matrix pair (A, ν ) satisfying Eq. ( 12) as an equality is said to be stoichiometrically compatible.</p><p>As noted, the RREF procedure in the implementation of the UMM always produces such a matrix pair (see Eq. ( 10)), in which case rank(ν ) is of maximal rank R for the given A matrix, where</p><p>.</p><p>In the case of stoichiometric restrictions, the left side of Eq. ( 12) is strictly less than the right side and the difference is defined to be the number of stoichiometric restrictions, r,</p><p>given by</p><!><p>An example of an explicit restriction 10 occurs for the reaction with potassium permanganate used in the analytical determination of hydrogen peroxide. Without restrictions, application of the UMM for this system with N = 7 shows that rank(A)=5, indicating that every possible reaction is described by a linear combination of the two (F s = N − rank(A) = 7 − 5 = 2) stoichiometric equations 6H However, it is known experimentally that peroxide and permanganate always react in the fixed ratio of 5/2. 24 This can be expressed as the following explicit restriction on their mole number changes, ∆n i :</p><p>This can be expressed by inserting an additional row into the system formula matrix (see This system thus exhibits one stoichiometric restriction, from Eq. ( 14):</p>

ChemRxiv

Evaluation of anticancer effects of carboplatin–gelatin nanoparticles in different sizes synthesized with newly self-assembly method by exposure to IR light

Carboplatin (CP), a platinum analog, is one of the most widely used chemotherapeutic agents in the treatment of colorectal cancer. Although platinum-based drugs are quite effective in anticancer treatments, their use in a wide spectrum and effective treatment possibilities are limited due to their systemic side effects and drug resistance development. In recent years, studies have focused on increasing the therapeutic efficacy of platinum-based drugs with drug delivery systems. Gelatin, a protein, obtained by the hydrolysis of collagen, is a biocompatible and biodegradable material that can be used in nano drug delivery systems. In this study, CP-loaded gelatin-based NPs (CP-NPs) were exposed to IR light in different temperatures at 30, 35, 40, 45, and 50 °C and characterized by FESEM-EDX, FTIR, UV-Vis, DLS. Accordingly, we synthesized gelatin-based CP-NPs of different sizes between 10-290 nm by exposure to IR. We found that CP-NPs-50, 16 nm nano-sized, obtained at 50 °C had the most cytotoxicity and was 2.2 times more effective than the free drug in HCT 116 colon cancer cells. Moreover, we showed that the cytotoxicity of CP-NPs-50 in normal HUVEC cells was lower. Additionally, we demonstrated that CP-NPs enhanced apoptotic activity while not developing MDR1-related resistance in colon cancer cells. In this study, for the first time drug loaded gelatin-based nanoparticles were synthesized in different sizes with a newly self-assembly method by exposing them to infrared light at different temperatures and their anticancer effects were evaluated subsequently.Colorectal cancer is the second most common cause of mortality from cancer among both men and women 1 . Despite various treatment options, it continues to be an important health problem worldwide. Although the clinical applications of platinum-based drugs are extremely effective, their toxicity profile restricts their extensive and effective application. Therefore, it is extremely important to develop new chemotherapy drug formulations that are more effective in cancer cells by reducing systemic side effects and drug resistance in the organism. The recent studies focus on developing new platinum-based drug formulations, expanding the therapeutic aspect. Carboplatin (CP), a platinum analog, is one of the most widely used chemotherapeutic agent in the treatment of colorectal cancer. CP was approved by the FDA in the 1980s and continues to be used in the treatment of various cancers. It has a broad spectrum of chemotherapy in various malignancies including ovarian cancer, small cell lung cancer, head and neck cancer, thoracic cancers, and bladder cancer 2 . Although CP, a derivative of

evaluation_of_anticancer_effects_of_carboplatin–gelatin_nanoparticles_in_different_sizes_synthesized

<!>Materials and methods<!>Preparation of nanoparticles.<!>UV-Vis spectroscopy. Absorbance measurements of the gelatin based CP-NPs were evaluated by Hach<!>Results and discussion<!>Conclusion

<p>cisplatin, has a similar therapeutic action, it differs in structure and toxicity from cisplatin. CP can create defects in DNA through platinum and cause cell death by inhibiting replication transcription. However, its side effects and development of drug resistance can significantly limit the full potential scope of the drug and cause serious problems 3 . Its common side effects such as nausea, vomiting, nephrotoxicity, neurotoxicity, and ototoxicity limit its use 4 . Additionally, carboplatin cannot be used at an effective dose due to the development of drug resistance such as cisplatin 5,6 . Therefore, it is imperative to develop new formulations in order to benefit from its therapeutic effect with the highest efficacy.</p><p>Gelatin is all-purposed biopolymer and is widely utilized in cosmetic, food, pharmaceutical, and medical area. Gelatin is formed by the hydrolysis of collagen, during which the collagen is denatured and its triple helix structure loosens and loses its secondary structure. Since collagen is insoluble in water, it hydrolyses to gelatin in acidic or basic solutions. During gelling, the chains of gelatin undergo a conformational rearrangement and the triple-helix structure gets partially renewed. Accordingly, two sort of gelatin are produced: Type A by acid hydrolysis and Type B by base hydrolysis 7,8 . In addition, gelatin has two different isoelectric points, such as type A and type B. While the isoelectric point of type A is between 7-9, the isoelectric point of type B is around 4-5. Therefore, different types of nanoparticles can be obtained from the nano-sized gelatins to be produced 9,10 . Being able to control the physical form of the gelatin molecule with temperature also increases the possibility of production at nanoscale.</p><p>Gelatin nanoparticles are natural polymers and widely used as drug carriers to target tumor tissues in diseases such as cancer due to their biocompatibility and biodegradability 11 . Furthermore, the major advantages of gelatin nanoparticles is not only very low toxicity, but also them having the opportunity for multiple modifications with excellent pharmaceutical excipient, thermos-reversible gel formation, high purity, low immunological and non-allergenic properties, steerable physical parameters. Moreover, it can also be produced easily at low cost and integrates easily with many molecules, and the fact that the gelatin matrix molecule has amino acid side chains allows for the formation of numerous other modifications. The surface of gelatin nanoparticles can be modified with site-specific ligands, cationized with amine derivatives, or coated with polyethyl glycols to achieve targeted and sustained release drug delivery. Compared with other colloidal carriers, gelatin nanoparticles are more stable in biological fluids to provide the desired controlled and sustained release of entrapped drug molecules 7,10 . Accordingly, gelatin is a unique nanocarrier to reduce the systemic side effects of carboplatin and increase its accumulation in cancer cells, including improved efficacy to overcome drug resistance. Although there are some studies on gelatin-based nanodrugs, which are usually related to paclitaxel [12][13][14][15][16][17][18] , there are few studies with the combination of carboplatin gelatin.Their nanoparticle size is very large and their in vitro anticancer activity has not been investigated 19,20 .</p><p>In order to produce nano gelatin, the literature so far has included (1) two-step desolvation 21,22 (includes centrifugation and lyophilization steps by adding crosslinker), (2) simple coacervation 23,24 (utilizing liquid-liquid extraction), (3) solvent evaporation 25,26 (evaporation of the solvent can also cause uneven particle size distribution in NP production), (4) microemulsion 27,28 (emulsification with surfactants). It is difficult to control the size with these methods except for microemulsion. In addition to all these, no other method that can provide self-assembly and size control while loading drugs into the carrier has been found in the literature. Therefore, in this study, we examined whether we can control the size of a macromolecule such as gelatin, whose physical form changes with temperature, by an IR light source.</p><p>In this study, for the first time in the literature, a drug (CP) was adsorbed to macromolecule gelatin induced by an IR light source and its anticancer effects were investigated. The originality of this work is not only to create a drug delivery system by stimulating it with an IR light source, but also to add carboplatin to this system. A literature review reveals that paclitaxel is used for bonding to gelatin-based drug carriers and there are very few studies on the combination of carboplatin and gelatin. As far as we know, there is no study in the literature realizing direct (self-assembled) gelatin macromolecule-nanodrug synthesis in different nanoscales when stimulated at different temperatures and investigating their anticancer effects. Obtaining nanoparticles with this method allows the drug to self-attach to the macromolecule, thus saving time, cost and effort. We believe that such a drug storage method will contribute to both nanotechnology and effective anticancer drug studies.</p><!><p>Colon cancer cells (HCT116) and human umbilical vein endothelial cells (HUVEC) were obtained from the American Type Culture Collection (ATCC). As anticancer drug Carboplatin was purchased from Koçak Farma and Gelatin A from Sigma (USA). Regarding chemicals, 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide kit (MTT), Dimethylsulfoxide (DMSO), Dulbecco's Modified Eagle's Medium (DMEM) for cell culture were obtained from Sigma (St. Louis, USA) Trypsin, penicillin, streptomycin from Gibco (UK). Phosphate buffer saline (PBS) tampon solution for release studies was acquired from Thermo Fisher (USA). AnnexinV/PI kits were procured from Becman Coulter (USA) while Anti-MDR1/ABCB1 Antibody (UIC2) for P-gp evaluation was obtained from Santa Cruz (USA).</p><!><p>As seen in Fig. 1, a setup was prepared to obtain carboplatin loaded gelatin-based nanoparticles. 20 mL of carboplatin (1 mg/mL, distilled water) and 20 mL of gelatin (10 mg/mL, distilled water) stock solutions were used. A 150-W IR light source (Rotlichtlambe IR 150) was placed approximately 20 cm from the sample and at a 45° angle to illuminate the entire sample (Fig. 1) and a thermometer was placed inside the sample container. Thus, self-assembly binding of the CP to the gelatin nano carrier was expected by exposure to IR light and vibrations. As the temperature of the solution in the sample container reached 30-35-40-45-50 °C, 2 mL of solution was taken at each specified temperature with a separate injector immersed in the sample container. After the final sampling at 50 °C, there was about 20 mL more of the mix-ture in the sample container. Although the protein structure of gelatin is denatured above 40 °C, it preserves its protein properties, but at temperatures above 50 °C, the protein structure may degrade and lose its protein properties. For this reason, we did not include temperatures above 50 °C into our study 8,29 . In addition, since the room temperature was also approximately 24 °C, it was appropriate to start taking samples of gelatin-based NPs at 30 °C. The experimental setup was protected with foams providing thermal insulation as shown in the figure . Particle size distribution and zeta potential distribution analysis. The size distribution of the gelatin based CP-NPs was determined by dynamic light scattering (DLS) analysis by Zetasizer Nano ZS (4 mW He-Ne laser operating) at room temperature., distilled water was used as reference liquid.</p><p>Fourier transform infrared spectroscopy analysis. Fourier Transform Infrared spectroscopy (FTIR) analysis were performed using the BRUKER ALPHA spectrometer with diffuse reflection mode at 4 cm −1 resolution. The measurements of each sample were recorded after 10 scans. Before the measurements, the correction was made by taking pure water as a reference.</p><p>EDX-FESEM analysis. First, each sample were individually dropped onto an amorphous glass lamella. The samples were then allowed to dry at room temperature under normal atmospheric conditions and the dried samples were used for FESEM-EDX analysis. Next, the surface characteristics of the gelatin-based CP-NPs were examined by a Gemini 500 digital transmission electron microscope (SEM). Quantitative elemental analyzes of each CP-NPs were determined by an EDX spectrometer attached to SEM.</p><!><p>Lange 500 Spectrophotometer at 25 °C and distilled water was used on the reference beam. The absorbance spectra of the gelatin-based CP-NPs were measured in the range of 200-1100 nm wavelength.</p><p>Drug release studies. To evaluate drug release profile of drug loaded gelatin based NPs, CP from gelatin based NPs were used in a dialysis membrane in pH 7.4 PBS tampon buffer, at 37 °C. Wavelengths of the carboplatin was measured at 235 nm 30 using the UV spectrophotometer. Then, 2 mL of gelatin based carboplatin loaded NPs were put in a dialysis membrane and placed in 50 mL of buffer solutions of pH 7.4 at 37 °C. Then, the sample was shaken for different periods and samples were taken at different times and measurements of the samples were taken in a UV spectrometer at 235 nm.</p><p>Cell culture. HCT116 colon cancer and HUVEC normal cell lines purchased were cultured in DMEM supplemented with 10% FBS, 100 µg/mL streptomycin and 100 units/mL penicillin in 5% CO 2 atmosphere in a humidified incubator. To procure enough cells for the cell studies, the medium of cells was changed every 2-3 days and passaged using trypsin when the cell confluent reached 70-80% confluence in the flask.</p><!><p>In this study, IR light with 150-W energy was used as a thermal and vibrational energy source to synthesize gelatine based CP-NPs in different sizes. First, 10 nm and 24 nm gelatin-based CP-NPs were obtained at 30 and 35 °C respectively, and afterwards, interestingly, the NPs sizes increased to 190 nm at 40 °C, and, then the size of gelatin-based CP-NPs decreased to 15 nm and 16 nm at 45 and 50 °C respectively (Fig. 2). According to Fig. 2, gelatin-based CP-NPs of different characters and sizes are formed at different temperatures when exposed to heat via IR light. CP-NPs-50 synthesized at 50 °C were more stable than the others (PDI = 0.607). These were also found to have the highest cytotoxicity on HCT116 colon cancer cells. CP-NPs-40, which was synthesized at 40 °C and had the largest nanoparticle size at 190 nm, had the highest PDI value of 0.849 and was least effective in HCT 116 colon cancer cells (Supplementary Fig. 1, Figs. 2, 6). According to Fig. 2, while 0.82 µg/mL CP was loaded on the gelatin nano-carrier at 30 °C, 0.738, 0.737, 0720 and 0.689 µg/mL CPs were loaded at 35, 40, 45 and 50 °C, respectively. Accordingly, as the temperature increased, the amount of drug that the gelatin nano-carrier could carry decreased (Supplementary Fig. 2 and Supplementary Table 1). At the same time, EDX analysis also shows the drug concentration by the relative % of platinum within (Supplementary Fig. 3). Accordingly, the amount of platinum in the structure of gelatin-bound CP-NPs was: Pt atomic 0.50% at 30 °C, Pt atomic 0.43% at 35 °C, Pt atomic 0.40% at 40 °C, Pt atomic 0.36% and 50% at 45 °C, Pt atomic was measured as 0.22% at 50 °C. These drug ratios were in good correlation with the results obtained from the UV-Vis spectrum.</p><p>In this study, the reason for the gelatin-based NPs we synthesized, having different NP sizes with the variable thermal effect of the IR lamp, is that gelatin has a thermo-reversible gel formation. It is reported that gelatin forms a tight gel at 26-30 °C, starts to melt after 33-34 °C and reaches a solubility above 40 °C7 . In this study, gelatin-based NPs were in tight gel form at 30 °C, holding less water and having a size of 10 nm, reaching 16 nm at 35 °C, again a small size, while at 40 °C, it had the largest size of 190 nm with the highest water content. Interestingly, with increasing solubility of gelatin at 45 and 50 °C, it is noteworthy that again 15 and 16 nm small size NPs were obtained. It is known that gelatin has a high water holding capacity at 40 °C32 . The large nanoparticle size (190 nm) of CP-NPs-40 synthesized at 40 °C in our study is supported by the literature 7 . It is clearly seen that gelatin-based CP-NPs of different characters and sizes are formed when heated to different temperatures via IR Light.</p><p>As seen in Fig. 3, the FESEM images are in good correlation with the nanosize measurements of DLS (Supplementary Fig. 2). According to the DLS results, the nanosizes of the gelatin-based CP-NPs obtained at 30 °C were between 7-15 nm, while they ranged between 15-32 nm at 35 °C and 141-255 nm at 40 °C, 11-24 nm at 45 °C and 11-24 nm at 50 °C. According to the FESEM images, after each sample was dried on the substrate, nanoparticles of exactly these sizes were detected very clearly, except for some agglomeration.</p><p>Figure 4A shows the infrared spectra of gelatin (Gel), free carboplatin (CP) and Gelatin-based CP-NPs (CP-NPs). Considering the IR spectrum of gelatin, vibration signals of 3263 cm −1 and 1637 cm −1 , respectively, originate from amide-I and amide-II bands. The reason why the peak at 3263 cm −1 is wide and flat in this way is due to the water in the environment. Amide-I, C=C bond stretch of amide proteins; amide II, N-H vibrational groups and stretching vibrations of C-N groups have contributed to these bands 33 . When we look at the IR spectrum of free carboplatin, the -NH vibration signals of the NH3 group bound to the platinum ligand at 3263-3123 cm −1 , the -CO and -C=O vibrational signals of the ester group at 1364 cm −1 and 1602 cm −1 , 2964 at cm −1 , -CH vibration signals in the ring structure in carboplatin were detected 34 . Complex vibration signals in the fingerprint region were not taken into account. Since the IR spectra of the gelatin-based CP-NPs we obtained at different temperatures were almost the identical and the spectrum of one of the CP-NPs samples is given as an example. According to the IR spectrum of gelatin-based CP-NPs, it is seen that some specific peaks of drug and macromolecule gelatin such as -CH aliphatic and N-H, C-N were shifted. In fact, some coalescence in the www.nature.com/scientificreports/ amide-I and amide-II bands is also striking. This spectrum is a serious proof that free carboplatin is adsorbed by the gelatin macromolecule. This proof is also supported by UV-Vis spectra and EDX analyses.</p><p>Cumulative drug release of gelatin-based CP-NPs synthesized at different temperatures was monitored for 50 h at physiological pH (pH 7.4) and 37 ± 0.5 °C. According to Fig. 5, 94% CP was released from CP-NPs-50 after 20 h, 93% CP was released from CPNPs-40 at the end of 20 h, and 81% from CP-NPs-30. According to Fig. 4B, it is seen that CP-NPs-50 has the highest release and CP-NPs-30 has the lowest release.</p><p>Table 1 and Fig. 5A show the cytotoxic effects of free CP and gelatin-loaded CP-NPs on HCT-116 colon cancer. According to Table 1, the IC 50 value of free CP on HCT-116 cells was 87.75, while CP-NPs-35, CP-NPs-40, CP-NPs-45 and CP-NPs-50 were 62.25, 60.45, 78.01, 60.21, 39.39 µM respectively. Accordingly, all CP-NPs were more effective than free CP on HCT116 cells. Moreover, gelatin-based CP-NPs-50 was by far the most effective in all samples. Moreover, it is noteworthy that the CP-NPs-50 (IC 50 : 39.39 µM) we synthesized is 2.22 times more cytotoxic than free CP on cancer cells. Furthermore, regarding normal HUVEC cells (Fig. 5B), the IC 50 values of all CP-NPs above 80 µM indicate that the nanoparticles we synthesize are specific to cancer cells. Thus, after treatment with CP-NPs, colon cancer cells would be killed with 2.22 times lower concentration of CP, while side effects would be significantly reduced. Our findings show that gelatin-based NP, which we synthesized at 50 °C (CP-NPs-50), is a promising agent for colon cancer treatment. www.nature.com/scientificreports/ Jahanshahi et al. synthesized the gelatin nanoparticles at temperatures of 40, 50, 55, and 60 °C and when they compared the NPs, the optimum temperature for gelatin nanoparticles was 50 °C35 . In another study, Jahanshahi et al. suggested that it is not possible to form nanoparticles at low temperature due to the highly viscous nature of gelatin, and that the particle size increases above 50 °C. This situation is thought to be related to the gelation property of the gelatin. When the viscosity decreases and the temperature increases, the triple helix structure dissolves at 50 °C35,36 . Although our study also supports that 50 °C is the best temperature for gelatin nanoparticles, the anticancer effects of drug-gelatin-based nanoparticles produced at different temperatures were compared for the first time in this study. In addition, in this study, small-sized nanoparticles (CP-NPs-30, CP-NPs-35) were obtained which can be loaded with more drugs at low temperatures. However, their drug release and cytotoxic effect are not better than CP-NPs-50.</p><p>To determine apoptotic activity (in Fig. 6), among gelatin-based CP-NPs, CP-NPs-50, which are the most effective in cancer cells were chosen and compared with free CP by flow cytometry analysis. Apoptotic activity was assessed using the concentration at IC 50 values of the compounds to match with cytotoxicity tests after 72 h. That is, a concentration of 39.39 µM of CP-NPs-50 was applied against 87.75 µM free CP. Free carboplatin caused 13.16% early apoptosis, CP-NPs-50 caused 15.59% early apoptosis, 32.33% late apoptosis and 9.56% necrosis were observed with CP, while 31.31% late apoptosis and 8.68% necrosis were observed with CP-NPs-50. Accordingly, it was determined that cell death by apoptosis occurred at a concentration less than half of CP-NPs-50 compared to free CP. According to these results, it is seen that more effective cancer cell death can be achieved with a lower CP concentration as a result of the accumulation of CP-NPs-50 in cells with the effect of CP-NPs on cancer cells. As a result of our findings, it is clear that CP-NPs-50 increases apoptotic activity at lower concentrations compared to free CP.</p><p>The effects of CP and CP-NPs-50 on drug resistance were evaluated in HCT116 cells by flow cytometry analysis by measuring Pgp expression (MDR1). Cells treated with CP and CP-NPs were compared with untreated cells for MDR1 activity. According to MDR1 measurements, free CP and CP-NPs-50 did not change Pgp expression www.nature.com/scientificreports/ in drug treated cells compared to untreated control cells (Fig. 7). This indicates that the nanoparticles did not develop MDR1-related resistance in colon cancer cells and can be safely used in therapy.</p><p>While CP-NPs-50 had the highest cytotoxic effect in cancer cells, CP-NPs-40 had the lowest toxicity on cancer cells. It is not a coincidence that CP-NPs-50 is around 16 nm and CPNPs-40 is around 190 nm, since it is the size of the NPs that significantly influences cytotoxicity.</p><p>It has been reported that the particle size, shape, surface charge and chemistry, stability of nanoparticles seriously affect the contact of NPs with the cell surface and signal transmission 37 . It is also notable that CP-NPs-40 has the highest PDI, while CP-NPs-50 has the lowest PDI (Fig. 2). This also shows that CP-NPs-50 is more stable.</p><p>Our findings show that gelatin-based NP, which we synthesized at 50 °C (CP-NPs-50), is a promising agent for colon cancer treatment.</p><!><p>For the first time in the literature, drug-macromolecule complexes in different nanosizes at different temperatures were successfully produced by the self-assembly method by means of an IR light source and were investigated for their anticancer activity. With this method, which we proved in our study, we achieved to produce different nano-sized drug-macromolecule complexes only by temperature difference via IR lamp. It is also important that the gelatin we choose as the drug delivery system is a cheap, biocompatible, and biodegradable material. This method also will reduce time and cost in production, both for researchers and also in terms of mass production. Subsequently, anticancer studies of the compounds synthesized in the study were carried out. It has been observed that all CP-NPs were more effective than free CP on HCT116 cancer cells. Moreover, gelatin-based CP-NPs-50 was by far the most effective in all samples. Moreover, it is noteworthy that the CP-NPs-50 (IC 50 : www.nature.com/scientificreports/ 39.39 µM) we synthesized is 2.22 times more cytotoxic than free CP on cancer cells. CP-NPs-50 was selective to cancer cells when compared with normal HUVEC cells. In addition, it would be a serious therapeutic advantage for a drug such as carboplatin, which has very high side effects, to be effective at 2.2 times lower concentration.</p><p>Our findings show that gelatin-based NP, which we synthesized at 50 °C (CP-NPs-50), is a promising agent for colon cancer treatment.</p>

Scientific Reports - Nature

Identification of the site of binding of sulfated, low molecular weight lignins on thrombin

Sulfated, low molecular weight lignins (LMWLs), designed recently as macromolecular mimetics of the low molecular weight heparins (LMWHs), were found to exhibit a novel allosteric mechanism of inhibition of human thrombin, factor Xa and plasmin, which translates into potent human blood anticoagulation potential. To identify the site of binding of sulfated LMWLs, a panel of site-directed thrombin mutants was studied. Substitution of alanine for Arg93 or Arg175 induced a 7\xe2\x80\x938-fold decrease in inhibition potency, while Arg165Ala, Lys169Ala, Arg173Ala and Arg233Ala thrombin mutants displayed a 2\xe2\x80\x934-fold decrease. Other exosite 2 residues including those that play an important role in heparin binding, such as Arg101, Lys235, Lys236 and Lys240, did not induce any deficiency in sulfated LMWL activity. Thrombin mutants with multiple alanine substitution of basic residues showed a progressively greater defect in inhibition potency. Comparison of thrombin, factor Xa, factor IXa and factor VIIa primary sequences reiterated Arg93 and Arg175 as residues likely to be targeted by sulfated LMWLs. The identification of a novel site on thrombin with capability of allosteric modulation is expected to greatly assist the design of new regulators based on the sulfated LMWL scaffold.

identification_of_the_site_of_binding_of_sulfated,_low_molecular_weight_lignins_on_thrombin

1. Introduction1<!>2.1 Sulfated LMWL, Chromogenic Substrate and Recombinant Thrombins<!>2.2 Quantitative Measurement of Thrombin Inhibition Potential of CDSO3<!>2.3 Analysis of Serine Protease Sequences and X-ray Crystal Structures<!>3.1 Specific Residues of Thrombin are Involved in Recognition of CDSO3<!>3.2 Three-dimensional structures reiterate Arg93 and Arg175 as key points for interaction with CDSO3<!>3.3 CDSO3-based allosteric inhibition of thrombin offers a powerful avenue to discover new anticoagulants

<p>Allosteric regulation of coagulation enzymes, especially thrombin and factors Xa, IXa, and XIa, is a fundamental property exploited by nature [1,2] to maintain homeostatic balance between coagulation and anticoagulation. The primary allosteric regulator of these enzymes is heparin, an animal-derived mixture of millions of polysaccharide chains, which binds in a site remote from the active site and enhances the inhibition of the enzymes by antithrombin, a plasma glycoprotein inhibitor [3].</p><p>Although heparins (unfractionated heparin or low molecular weight heparin (LMWH)) have been used as anticoagulants since a long time, the agents are beset with a number of adverse reactions including enhanced bleeding risk, immunological reaction, poor oral bioavailability, patient-to-patient response variability, narrow therapeutic index, possibility of contamination and others [2,4,5]. Yet, it has been difficult to replace these macromolecular entities because these are excellent anticoagulants and fairly inexpensive.</p><p>We recently designed sulfated LMWLs as macromolecular mimetics of LMWHs that exhibited potent inhibition of coagulation in vitro and ex vivo [6-8]. Sulfated LMWLs are sulfated oligomers of varying lengths and substitution pattern (Fig. 1) that attempt to mimic the structural diversity of LMWHs and are readily synthesized in a simple two-step chemical process. Structurally, sulfated LMWLs possess an aromatic backbone decorated with few sulfate and carboxylate groups, while LMWHs are highly anionic, carbohydrate-based molecules. The combination of a hydrophobic, aromatic scaffold and selected number of charged groups in sulfated LMWLs induces novel physicochemical and protein recognition properties [7-9]. In this respect, sulfated LMWLs are proving to be unlike any other class of anticoagulants being investigated to-date, including the heparins, the coumarins, the hirudins, the peptidomimetics and the small molecule direct inhibitors [2,10].</p><p>Functionally, sulfated LMWLs display plasma and blood anticoagulation profiles similar to that of LMWHs [8]. Enzyme inhibition studies have shown that these molecules inhibit thrombin, factor Xa and plasmin in an antithrombin–independent manner [6,11]. Interestingly, mechanistic studies have shown that sulfated LMWLs utilize exosite 2 of thrombin to induce inhibition [7]. This novel allosteric inhibition mechanism distinguishes sulfated LMWLs from the heparins, which do not directly inhibit these enzymes. In fact, sulfated LMWLs appear to be the only molecules that allosterically induce inhibition of thrombin through an exclusive exosite 2 interaction.</p><p>To elucidate the site of binding of sulfated LMWLs, we studied the inhibition properties of a panel of single, double and triple site-directed thrombin mutants. The results show that Arg93 and Arg175 are two key residues that recognize CDSO3, a specific, highly potent sulfated LMWL, while Arg165 and others are also important. Comparison of thrombin, factor Xa, factor IXa and factor VIIa primary sequences supports these conclusions. The identification of a novel site on thrombin with capability of allosteric modulation is expected to greatly assist the design of new regulators based on the sulfated LMWL scaffold.</p><!><p>CDSO3, a specific sulfated LMWL (Fig. 1), was synthesized in two steps from caffeic acid using chemo-enzymatic synthesis described by Monien et al [6]. The average molecular weight of CDSO3 was measured using a Shodex Asahipak GS-320 HQ size-exclusion (SEC) column (Showa Denko America, Inc, New York, NY) eluted with 0.1 M NaOH at 0.7 mL/min and detected at 280 nm. Polystyrene sulfonate (PSS) standards (4200–33000) from American Polymer Standards (Mentor, OH) were used as standards for calibration of the column. The average molecular weight of CDSO3 was calculated to be 3,320 Da. Chromogenic substrate S2238 (H-d-Phe-Pip-Arg–p-nitroanilide) was purchased from ANASPEC (Fremont, CA). All other chemicals were analytical reagent grade from either Sigma Chemicals (St. Louis, MO) or Fisher (Pittsburgh, PA) and used as such.</p><p>Recombinant wild-type and mutant thrombins were prepared in the Rezaie laboratory at St. Louis University School of Medicine, as described earlier [12,13]. Briefly, Arg93Ala, Arg97Ala, Arg101Ala, Arg165Ala, Lys169Ala, Arg173Ala, Arg175Ala, Arg233Ala, Lys235Ala, Lys236Ala, or Lys240Ala thrombin was prepared in prothrombin-1 form by PCR mutagenesis and expression in baby hamster kidney cells (BHK) using the pNUT-PL2 expression/purification vector system. The mutants were purified to homogeneity by immunoaffinity chromatography using the Ca2+-dependent monoclonal antibody, HPC4, and activated to thrombin. The active-site concentrations of thrombin mutants were determined by an amidolytic acivity assay and stoichiometric titrations with antithrombin [12,13]. These concentrations were within 90-100% of those expected on the basis of their absorbance at 280 nm. Stock solutions of these thrombins were prepared in 20 mM sodium phosphate buffer, pH 7.4, containing 100 mM NaCl and 2.5 mM CaCl2.</p><!><p>CDSO3 inhibition of recombinant wild-type and mutant thrombins was studied in a manner identical to that described earlier for human plasma thrombin [6,7]. The buffer used for these experiments was 20 mM Tris-HCl buffer, pH 7.4, containing 100 mM NaCl, 2.5 mM CaCl2 and 0.1 % polyethylene glycol (PEG) 8000 in PEG 20,000-coated acrylic cuvettes. S2238 was used as substrate and the residual thrombin activity was quantified by measuring the initial rate of hydrolysis from the linear increase in absorbance at 405 nm as a function of time under conditions wherein less than 10% substrate is consumed. Briefly, a solution of 10 μL CDSO3 at concentrations ranging from 0.1–8600 μM was diluted with 963 μL of buffer and 7 μL of 0.23–1.07 μM thrombin and mixed well. This was followed by addition of 20 μL S2238 (2 mM) and the initial rate was rapidly measured. Logistic equation 1 was used to fit the dose dependence of residual thrombin activity to obtain IC50 of inhibition.</p><p>In this equation, YM and YO are the maximum and minimum values of the fractional residual thrombin activity; IC50 is the concentration of the inhibitor that results in 50% inhibition of enzyme activity. Sigmaplot 8.0 (SPSS, Inc. Chicago, IL) was used to perform nonlinear curve fitting in which YM, YO, and IC50 were allowed to float.</p><!><p>The primary amino acid sequence of thrombin, factor Xa, factor IXa and factor VIIa was retrieved from the ExPASy web server (http://www.expasy.org). A multiple alignment was performed on the sequences using ClustalX version 2.0.11 with default parameters [14]. A representative crystal structure for each protease was retrieved from the RCSB Protein Data Bank (PDB; http://www.pdb.org). SYBYL 8.1 (Tripos, Inc., St. Louis, MO) was used to prepare the structures and figures derived from the parent PDB file. All redundant subunits, water, cofactors, inhibitors and other ligands were removed from the crystal structures prior to molecular surface calculations. In addition, the SYBYL Mutate Monomers function was used to assign coordinates to the incompletely resolved arginine and lysine side chain atoms. Hydrogen atoms were added using the Add Hydrogens feature. The serine proteases were then aligned spatially (Fit Monomers function) using the backbone atoms of the amino acid positions corresponding to the basic arginine and lysine residues of thrombin exosite 2, i.e., Arg93, Arg101, Arg165, Arg233, Lys236 and Lys240. Unless otherwise noted, the chymotrypsin numbering system [15] is used to facilitate identification of corresponding residues in the chymotrypsin-like serine protease family members.</p><!><p>Our earlier work led to the conclusion that CDSO3 does not bind in the active site or exosite 1 of thrombin [7]. Experiments using exosite 2 ligands such as unfractionated heparin and LMWH showed good competition with CDSO3 [7]. Likewise, sulfated LMWLs were also found to compete with heparin for binding to human plasmin also [8], which supported exosite 2-like site as the prime site of CDSO3 recognition. Yet, exosite 2 of thrombin is a rather large ellipsoidal domain spanning an area of approximately 20×30 Å2. It consists of several basic residues, e.g., Arg93, Arg101, Arg165, Arg233, Lys235, Lys236, and Lys240, some of which are known to play important roles in heparin binding. Interestingly, Arg93 and Arg101 of this group are adjacent to a hydrophobic patch, which we reasoned may be important for recognizing the aromatic rings of CDSO3. To identify the key residues of exosite 2 that are important for CDSO3 binding, we studied inhibition of a library of thrombin mutants. The library included single-site replacement of basic residues of exosite 2 (identified above) with alanine. In addition, additional residues near exosite 2 including Arg97, Lys169, Arg173, and Arg175 were also studied. The preparation and characterization of these thrombin mutant have been described earlier [12,13]. Direct inhibition of these mutants by CDSO3 was studied in a manner similar to that used for wild-type thrombin [7].</p><p>The dose–response profiles of eleven mutants studied here were essentially identical to the recombinant wild-type enzyme (Fig. 2). The measured IC50s for Arg93Ala and Arg175Ala thrombin mutants were 313 and 275 nM, which were 7–8-fold greater than that of recombinant wild-type thrombin (39 nM) (Table 1). In contrast, alanine replacement at Arg97, Arg101, Lys235, Lys236, or Lys240 positions did not affect the CDSO3 inhibition potency (0.9–1.5-fold effect), while substitutions at the 165, 169, 173 and 233 positions introduced a 2.2–3.7-fold decrease in inhibition potency (Fig. 3A).</p><p>The maximal inhibition efficacy of CDSO3 across these thrombin mutants ranged from 47–82% in comparison to 58% observed for the recombinant wild-type form. These efficacies are generally similar to that noted earlier for human plasma thrombin and human plasma plasmin (∼80%) [7,8].</p><p>To assess the effect of multiple replacements on CDSO3 inhibition potency, double and triple replacement at the 93, 97, and 101 positions were studied. Replacing arginines at both 93 and 97 positions with alanines shifted the dose-response profile significantly to the right in comparison to both the wild-type enzyme as well as either single-site mutant (Table 1, Fig. 3B). The measured IC50 for the 93,97-double mutant was 679 nM, which indicated an increase of ∼17-fold and ∼2-fold over wild-type and Arg93Ala enzymes, respectively. Introducing alanines for arginines at 93, 97, and 101 positions led to an increase in IC50 of ∼36-fold from the wild-type form (Fig. 3B). The efficacy of inhibition for the double and triple mutants was comparable to other thrombins studied in this work (∼80%, Table 1).</p><!><p>Our earlier work shows that CDSO3 potently inhibits thrombin and factor Xa, moderately inhibits plasmin, and does not inhibit factor VIIa and factor IXa [7]. Each of these enzymes is a trypsin-related serine protease with considerable three-dimensional structural similarity and we reasoned that CDSO3 could be reliably expected to bind in domains similar to exosite 2 of thrombin.</p><p>To evaluate the specificity of recognition of enzymes, the primary sequences of human thrombin, factor Xa, factor IXa, factor VIIa, and plasmin were aligned using ClustalX (see Figure S1 in Supplementary Material for full alignment). The enzymes display considerable similarity in length and sequence, as expected, especially in the exosite 2 region (Table 2). The alignment indicates that thrombin and factor Xa are the most closely related enzymes in terms of their exosite 2 regions with percent identity as high as 36%. Both enzymes display electropositive residues at 93, 165, 169, 236 and 240 positions (Fig. 4). Although factor Xa contains a neutral residue at 175, the side chain of Lys169 can effectively mimic the positive charge of Arg175 of thrombin (not shown). Likewise, human plasmin contains Lys645 (plasminogen numbering) that is equivalent to Arg93 of thrombin in three-dimension (not shown), while also containing basic residues at 101, 175 and 233 positions. In contrast, neither factor IXa nor factor VIIa possess an arginine at 93 or at 175, and display much poorer sequence similarity to thrombin (≤18%, Table 2).</p><p>Measurement of partition coefficient (logP), a simultaneous measure of solubility in aqueous and hydrophobic media, has suggested that sulfated LMWLs are considerably hydrophobic [9] in addition to being anionic. Analysis of hydrophobicity maps of the five enzymes being studied here shows that Arg93 of thrombin and factor Xa is located on the periphery of a fairly extensive hydrophobic patch made by side chains of Leu60, Pro60B, Ile88, Ile90, Tyr94 and Trp96 (Fig. 4). These residues can potentially interact with the aromatic rings of CDSO3. Such an extensive hydrophobic patch is almost non-existent in factors VIIa and IXa (not shown). Thus, it appears that CDSO3 recognizes a specific sub-site within exosite 2 that affords interaction with hydrophobic side-chains near appropriate basic residues (Fig. 4), which explains the observed selectivity of recognition by sulfated LMWLs.</p><!><p>Discovering allosteric modulators of protein function is challenging and yet highly sought after considering that this form of regulation is suggested to offer finer control necessary for therapeutic success [2,16]. Despite this expectation, allosteric modulators of cardiovascular enzymes, other than heparin and heparinoids, have not been vigorously pursued [17,18]. In this regard, sulfated LMWLs offer excellent promise. The identification of the site of binding of these novel molecules on thrombin implies by analogy a corresponding site on factor Xa and plasmin, which offer new avenues for discovering anti- as well as pro-coagulants [11].</p><p>The availability of the library of recombinant thrombin variants offered a unique tool in identifying the sub-domain consisting of Arg93, Arg175, and Arg165 as important for binding. The results also show that Lys235, Lys236 and Lys240 are not important for CDSO3 recognition. Heparin, on the other hand, recognizes Lys236, Lys240, Arg93, Arg101 and Arg233 [19] in that order of importance. Interestingly, the magnitude of binding defect introduced by these single mutations is significantly different for the two polyanions. Whereas Arg236 and Arg233 replacement introduced ∼43- and 8-fold defects for heparin binding [20], the maximal defect for CDSO3 binding was ∼8-fold for Arg93Ala thrombin. These results reveal interesting similarity as well as differences between the two polyanionic molecules and highlight the importance of aromatic backbone of CDSO3. More importantly, identification of this sub-domain implies that for the first time detailed molecular modeling studies can be initiated to design advanced molecules based on CDSO3 structure.</p><p>This work also explains the selectivity of enzyme recognition by CDSO3 and other sulfated LMWLs. Although each enzyme studied to date is a trypsin-based serine protease, the ability to selectively target thrombin and factor Xa is intriguing. Analysis of the three-dimensional structure of thrombin indicates that Arg93, the key residue for interaction with CDSO3, forms a 'ridges' adjacent to a well-defined hydrophobic 'valley' (Fig. 4). Such a well-defined hydrophobic region is present only in thrombin, factor Xa, and plasmin, and not in factors IXa and VIIa. We hypothesize that sulfated LMWLs, which contain multiple aromatic rings, utilize this hydrophobic region in addition to positive charge of Arg93 and other basic residues for interacting with select group of coagulation enzymes.</p><p>The reason why Lys235, Lys236 and Lys240 do not play a more important role for CDSO3 as compared to that for LMWHs is not clear. The hydrophobic 'valley' similar to that noted adjacent to Arg93 (described above) is not present near these contiguous residues, which is in line with our hypothesis that an optimal combination of charge and hydrophobic character, and not just polyanionic character, in CDSO3 induces recognition of coagulation enzymes. We believe that this hypothesis could be advantageously exploited for other coagulation enzymes so as to design highly specific CDSO3-based structures that target only one enzyme in an allosteric manner.</p><p>The observation that double and triple mutations progressive weaken the inhibition potential of CDSO3 is interesting considering that individual single mutations (Arg101Ala or Arg97Ala) do not appear affect potency to a significant degree. In either case, Arg93, the key residue for CDSO3 recognition, was intact, which possibly suppresses the loss.</p><p>Finally, sulfated LMWLs may offer a unique advantage with regard to the mechanism of action. The exosite 2-mediated allosteric regulation of thrombin offers a sound possibility of an antidote strategy with non-inhibitory sulfated carbohydrates. For example, small oligosaccharides that bind in exosite 2 without inducing much direct inhibition, such as sucrose octasulfate [21], may serve as effective antidotes of CDSO3-based anticoagulants. This would impart considerable safety to anticoagulant therapy based on sulfated LMWLs.</p>

PubMed Author Manuscript

Design and Synthesis of C-2 Substituted Thiazolo and Dihydrothiazolo Ring-Fused 2-Pyridones; Pilicides with Increased Antivirulence Activity

Pilicides block pili formation by binding to pilus chaperones and blocking their function in the chaperone/usher pathway in E. coli. Various C-2 substituents were introduced on the pilicide scaffold by design and synthetic method developments. Experimental evaluation showed that proper substitution of this position affected the biological activity of the compound. Aryl substituents resulted in pilicides with significantly increased potencies as measured in pili-dependent biofilm and hemagglutination assays. The structural basis of the PapD chaperone-pilicide interactions was determined by X ray crystallography.

design_and_synthesis_of_c-2_substituted_thiazolo_and_dihydrothiazolo_ring-fused_2-pyridones;_pilicid

Introduction<!>Results and discussion<!>Conclusion<!>General<!>General procedure for the preparation of 6g-m<!>General procedure for the preparation of 7a, 7c-d and 7f-n<!>Structure determination of pilicide 1 and 5d bound to PapD-PapH

<p>The increasing bacterial resistance to many antibiotics has resulted in a growing interest in new antibacterial agents with new modes of action directed towards Gram-negative bacteria.1 Contrary to broad-range bactericidal or bacteriostatic antibiotics, drugs that specifically target bacterial virulence factors would exert less selective pressure on bacteria and minimize the risk for horizontal spread of drug-resistance genes.2-4 It is estimated that approximately 11 million cases of urinary tract infection (UTI) occur in the U.S. each year, primarily in women. Treatment of UTI, like other microbial infections, is exacerbated by increasing antimicrobial resistance. This has been described as an impending "public health crisis". In addition, over 900,000 women and men in the U.S. experience three or more UTI episodes per year. Thus, antibiotics do not stop recurrence in this population, independent of antibiotic resistance, highlighting the need for alternative therapeutics.</p><p>Many Gram-negative bacteria, such as uropathogenic E. coli (UPEC), produce long filamentous structures called pili/fimbriae assembled by the chaperone/usher pathway (CUP pili)5, 6 that mediate colonization and invasion of host tissues.4, 7, 8 Compounds that target CUP pili have the potential to block virulence in a broad range of bacterial strains. The chaperones that are required for pilus assembly bind to pilus subunits and facilitate their folding by a mechanism termed donor strand complementation.9-11 Chaperone-subunit complexes are targeted to the outer membrane usher, which catalyzes pilus assembly by a process known as donor strand exchange.5, 12, 13 The chaperone is comprised of two immunoglobulin-like domains arranged in a boomerang shape. The usher has a 24 stranded beta barrel that forms a channel across the outer membrane. The channel is gated by a plug domain. Ushers also have two periplasmic domains. The N terminal periplasmic domain interacts, in part, with a hydrophobic patch on the N terminal domain of the chaperone exposed on incoming chaperone/subunit complexes. One class of compounds termed pilicides, consisting of a dihydrothiazolo ring-fused 2-pyridone scaffold, has been shown to prevent the formation of pili by blocking the interaction of chaperone/subunit complexes with the N terminal domain of the usher, thus preventing the donor strand exchange reaction and the formation of chaperone-usher assembled pili, exemplified in UPEC.14, 15 The X-ray crystal structure of pilicide 1 bound to the P pilus chaperone PapD revealed that compound 1 binds to the hydrophobic patch on the N terminal domain of the chaperone – formed by residues I93, L32, and V56 – which forms the well-conserved usher binding site (Figure 1).14, 16 Competition binding studies confirmed that the pilicide 1 prevents binding of chaperone-subunit complexes to the usher N-terminal domain. Pilicide 1, therefore functions, by blocking delivery of subunits to the usher thus preventing donor strand exchange and inhibiting pilus assembly (Figure 1). Besides being useful as potential antibacterial agents, pilicides have been used as chemical tools to study details of pilus assembly and their role in disease processes. 14, 17</p><p>In the crystal structure showing pilicide 1 bound to the chaperone-subunit complex there is unoccupied space near the thiazolo-part of the pilicide scaffold (Figure 1, iii). We hypothesized that introducing substituents in this part of the scaffold could thus lead to favourable interactions between the chaperone and the pilicide and thereby result in increased biological activity. We decided to decorate the C-2 position on the scaffold to create interactions with the unoccupied space in the crystal structure in an attempt to improve pilicide activity. The dihydrothiazolo ring-fused 2-pyridone scaffold (4), that constitutes the platform for design of new pilicides, can be synthesized via an acylketene-imin cyclo-condensation. 18, 19 This method introduces substituents directly in the C-7 and C-8 position on the scaffold leaving position C-6 and C-2 open for further improvements of the pilicides (Figure 2). Previous studies have also described a number of methods for the introduction of various substituents in position C-6 on the core structure.20-22 Although this did not lead to any substantially increased biological activity, substitutions at C-6 allow for the introduction of solubility enhancing substituents. The C-2 position on the scaffold had not been varied until recently, when we reported on the decoration of this position starting from the α,β-unsaturated methyl ester (6) (Figure 2).23 From this α,β-unsaturated methyl ester substituents were introduced by conjugate additions resulting in an addition with complete trans selectivity. In addition, substituents were introduced with a retained double bond by the use of Heck couplings or deprotonation using Lithium diisopropylamide (LDA) and subsequently reacted the lithiated scaffold with different electrophilic reagents. Together, these different methods opened up the possibility to introduce a wide range of substituents with varying size, electronic properties and spatial arrangement (Figure 2).</p><p>Here, we present biological and crystallographic data on second generation compounds that were synthesized via C-2 substitution of the thiazolo ring-fused 2-pyridone 6, based on the pilicide 1 – PapD complex crystal structure. The C-6 morpholinomethyl substituent in pilicide 1, earlier introduced for solubility reasons, was not included in this study as it is not vital for pilicide activity. The activities of the compounds were evaluated in pilus dependent biofilm and hemagglutination assays. Compared to the parent pilicide 5a, C-2 substituted compounds have significantly improved IC50's in inhibiting pilus formation. The molecular basis of the observed improved biological activity of the compounds was determined by elucidating the X-ray crystal structure of pilicides bound to the PapD-PapH chaperone-subunit complex (PapH is the P pilus termination subunit24). The pilicides bound to the previously identified binding site on the chaperone N-terminal domain near strands F1, C1 and D1. The introduced substituents were accommodated in a shallow hydrophobic pocket on the chaperone surface, formed by residues P30, L32, I93 and P95.</p><!><p>The developed methods for C-2 substitution were implemented to introduce a diverse set of substituents on a known pilicide (4)23 in an attempt to increase the potency of the compounds in inhibiting pilus assembly. The generated set of compounds (4a-d and 6a-f) were then subjected to hydrolysis, by three different methods depending on the nature of the substituents, to give the corresponding carboxylic acids (5a-d and 7a-f) in 80 to 95% yield (Table 1). The synthesized set of C-2 substituted thiazolo and dihydrothiazolo ring-fused 2-pyridones were biologically evaluated for their ability to block pilus formation as measured in a pili-dependent biofilm assay on polyvinylchloride plastic.25 In this assay, deletion of the type 1 pilus gene cluster in E. coli strain UTI89 completely destroys the ability of the bacteria to form biofilm. Thus, the amount of biofilm that is formed in UTI89 grown in the presence of pilicide, is related to the potency of the compound in blocking piliation. This whole bacteria assay is suitable for the screening of the synthesized compounds and provides more relevant biological information than elementary in vitro binding assays. The results are summarized in table 1.</p><p>These experiments revealed that the addition of substituents in the C-2 position of the pilicide scaffold significantly enhanced the potency of the pilicides resulting in increased inhibition of pili dependent biofilm formation. The phenyl substituted compounds 5d and 7c proved to be the best pilicides with significantly improved potency compared to their parent lead compound 5a. Also the methyl substituents proved to increase the potency. However, the influence of the spatial arrangement on activity is different between the phenyl and methyl substituents. In the phenyl case, the unsaturated 7c is more potent than the saturated counterpart 5d, whereas in the methyl case the situation is the opposite, saturated 5b is more potent than the unsaturated 7a. The remaining substituents resulted in 15-30 times lower potency, compared to the most active 7c (5c, 7f, and 7b), or were completely inactive (7d and 7e).</p><p>Based on the results with phenyl substituents, and the unsaturated analogue in particular, an additional investigation with C-2 aryl and heteroaryl substituents was performed. We decided to use the Suzuki–Miyaura cross-coupling to introduce these substituents.26 In order to implement these cross coupling reactions a bromo-substituted unsaturated ring-fused 2-pyridone 8 was desired. The synthesis of 8 had previously been accomplished by a two step process starting with oxidation of 4a, followed by lithiation and bromination.23 In order to avoid unnecessary purification and isolation of the oxidized ring-fused 2-pyridone 6, a one-pot procedure going directly from 4a to 8 by modifying the developed oxidation method was envisioned. Consequently, by adjusting the reaction conditions and treating 4a with 3 equivalents of NaH followed by 3 equivalents of BrCCl3 and finally 2 equivalents of MeOH the desired bromo derivative 8 was synthesized in 91% yield (Scheme 1).</p><p>After successfully established a robust one-pot method for the key compound 8, our focus was directed to the subsequent Suzuki–Miyaura cross-coupling. For this reaction we adapted a method previously used within our lab for Suzuki–Miyaura cross-couplings.27 Interestingly, we discovered that this reaction could be performed with no use of ligands and still give the same or increased yields without influencing reaction time and temperature. Hence, seven different aryl or heteroaryl boronic acids could be coupled to give C-2 substituted thiazolo ring-fused 2-pyridones 6g-m by running the reaction at 100 °C for 10 min by microwave irradiation (MWI) with 10 mol% Pd(OAc)2, 2 equivalents of boronic acid and 1.9 equivalents of KF in dry MeOH (Table 2). The reaction resulted in good to excellent yields (72-97%, entries 1-5, Table 2) with the furane boronic acids as the only exceptions resulting in thiazolo ring-fused 2-pyridones 6l and 6m in 46% and 34% isolated yield, respectively (entries 6 and 7, Table 2).</p><p>To expand this set of compounds, a benzyl substituted analogue 6n was synthesized in 75% yield via lithiation followed by addition of benzyl bromide using the previously described procedure.23 Before being evaluated the compounds were hydrolysed (KOH in THF/MeOH and microwave irradiation (MWI) 10-20 min at 90 °C, method 2). This rendered the corresponding carboxylic acids (7g-n) in 50-92% yield (Scheme 2).</p><p>In addition to this, it has been shown that the carboxylic acid is important for a retained biological activity28 and that the use of carboxylic acid isosteres (e.g. tetrazoles or acyl sulfonamides) could lead to increased pilicide activity.29 Hence, 7c was also transformed into its corresponding methyl acyl sulfonamide 9, which had proven to be one of the most promising isosteres, in 60% yield (Scheme 3).</p><p>Finally, the Suzuki coupled aryl and heteroaryl compounds, the benzyl derivative, and the methyl acyl sulfonamide acid isostere were all biologically evaluated for their ability to prevent pili dependent biofilm formation (Table 3).</p><p>We found that many of the aryl and heteroaryl-substituted compounds efficiently inhibited formation of biofilm. The benzyl substituted compound 7n had an even lower IC50 than the most active compound from the first biofilm evaluation (7c) (Table 3). The tolyl substituted compounds 7g and 7h were both highly active; 3-tolyl (7g) resulted in a similar activity as the phenyl substituted analogue 7c (Table 3). Of the smaller heteroaryls, the thiophene substituted 7k proved to be more potent than the furan substituted 7l and 7m (Table 3). Sterically more demanding substituents as in 7i and 7j were not tolerated and both of them turned out to be inactive (Table 3). Interestingly, the acid isosteric acyl sulfonamide in 9 did not give any additive effect but instead gave a 3-fold decreased potency, compared to the carboxylic acid 7c (Table 3). Next, the best compounds from the biofilm evaluations were further tested using a hemagglutination (HA) assay (Table 3).14 In this assay, the degree of piliation of the culture is related to the HA titer. Thus, after growth in the presence of pilicide, the HA titers of the cultures were determined to evaluate the relative potencies of the compounds in blocking pilus formation. Growth of UTI89 in the presence of all five tested compounds resulted in significantly lower HA titers. The benzyl substituted 7n and the 3-tolyl substituted 7g resulted in 16 times increased activity compared to the hydrogen substituted compound 5a (Table 3, entry 11, 4, and 1). Interestingly, 7n, 7g and the thiophene substituted compound 7k were all more active than the phenyl substituted compounds 5d and 7c (Table 3, entry 11, 4, 8 versus 2 and 3). Importantly, the compounds did not affect the bacterial growth (Figure S2 in supporting information).</p><p>The X-ray crystal structure of pilicide 1 bound to the P pilus chaperone PapD previously revealed that the pilicide binds a conserved hydrophobic patch on the chaperone N-terminal domain, formed by residues L32, V56 and I93, that coincides with the usher binding site. Subsequently, pilicide 1 was shown to block the binding of chaperone-subunit complexes to the usher in an in vitro binding assay. Thus, we hypothesized that pilicide 1 should be capable of binding to chaperone-subunit complexes to exert its mechanism in blocking binding of the complex to the usher. This hypothesis was tested by determining the X-ray crystal structure of pilicide 1 bound to the PapD-PapH chaperone-subunit complex. Pilicide 1 was found to bind to the same global binding site as seen in the PapD-pilicide 1 structure, however, substantial reorientation was observed, possibly as a result of the conformational changes in the loop between strands F1 and G1 in the subunit-bound and free chaperone state (Figure 3, i). In the pilicide 1-PapD structure, the pilicide carboxylic acid and carbonyl group make an electrostatic interaction with R96 and R58, respectively (Figure 1, i). In the PapD-PapH bound structure, these interactions are broken. We note that in the PapD-pilicide 1 complex, the morpholine ring of 1 is in contact with a neighboring molecule, such that the effect of crystal contacts in the observed reorientation in the PapD- and PapD-PapH-bound compound cannot be excluded. Finally, to investigate the structural basis for the observed improved IC50's after introduction of C-2 substituents, PapD-PapH crystals were soaked with the best compounds. Soaks with compound 5d generated a crystal showing the pilicide bound to the chaperone-subunit (PapD-PapH) complex at near the exact same position as the C-2 unsubstituted compound 1 (Figure 3, ii). Compared to the pilicide 1-bound form, L32 undergoes a conformational change when bound to 5d, which creates a shallow pocket that accommodates the phenyl substituent at the C2 position (Figure 3, ii & iii). The crystal structure of the N terminal domain of the FimD usher bound to the FimC-FimH158-279 16 chaperone-subunit complex is known. Modeling, based on this information, clearly shows the clash between pilicide and the usher binding site (Figure 3, iv).</p><!><p>A new route for fast and convenient synthesis of C-2 aryl substituted pilicides was developed and applied. Biofilm and HA assays revealed that proper decoration of this position significantly can improve the anti-virulence activity of the compounds. The most rewarding substituents proved to be phenyls in general and the benzyl and tolyl substituents in particular. Smaller heteroaryls (thiophene and furanes) also improved activity whereas sterically more demanding groups as indoles and 1,4-benzodioxanes abolished all activity. Combining a C-2 phenyl substituent with an acid isoster did not give any additive effect, instead a drop in activity was observed.</p><p>X ray crystallography of the pilicides bound to the PapD-PapH chaperone-subunit complex revealed that introduction of a C-2 phenyl substituent induced a reorientation of one of the chaperone (PapD) residues (L32), generating a shallow pocket that accommodates this substituent.</p><!><p>Flash column chromatography (eluents given in brackets) was performed on silica gel (Matrex, 60 Å, 35-70 μm, Grace Amicon). Parallel flash chromatography was performed on a Gradmaster parallel, Jones Chromatography. 1H and 13C NMR spectra were recorded on a Bruker DRX-400 in CDCl3 [residual CHCl3 (δH 7.26 ppm) or CDCl3 (δC 77.0 ppm) as internal standard] at 298 K. IR spectra were recorded on an ATI Mattson Genesis Series FTIR spectrometer. Microwave reactions were carried out using a monomode reactor (Smith Creator, Biotage AB) in Teflon septa capped 0.5-2 ml or 2-5 ml Smith TM process vials with stirring. Reaction times refer to irradiation time at target temperature, as measured by IR sensor. Purities of key compounds were >95% as determined by 1H NMR and HPLC.</p><!><p>The procedure for 8-Cyclopropyl-7-naphthalen-1-ylmethyl-5-oxo-2-m-tolyl-5H-thiazolo[3,2-a]pyridine-3-carboxylic acid methyl ester (6g) is representative. Dry MeOH (0.9. ml) was added to pyridone 8 (20mg, 0.043 mmol), Pd(OAc)2 (1 mg, 0.004 mmol), KF (5 mg, 0.81 mmol) and boronic acid (12 mg, 0.085 mmol). The reaction mixture was heated in a sealed tube by microwave irradiation (MWI) at 100 °C for 10 min. The resulting mixture was diluted with CH2Cl2, washed with sat. (aq.) NaHCO3 and the aqueous layer was extracted with CH2Cl2. The combined organic layers were concentrated under reduced pressure and purification by parallel flash chromatography [heptane:EtOAc 100:0-0:100] gave pyridone 6g (20 mg, 96% yield) as yellow foam. 1H NMR (400 MHz, CDCl3) δ 0.77-0.84 (m, 2H), 1.02-1.10 (m, 2H), 1.76-1.86 (m, 1H), 2.41 (s, 3H), 3.88 (s, 3H), 4.55 (s, 2H), 5.94 (s, 1H), 7.21-7.26 (m, 2H), 7.30-7.36 (m, 1H), 7.37-7.44 (m, 3H), 7.45-7.52 (m, 2H), 7.75-7.81 (m, 1H), 7.81-7.91 (m, 2H). 13C NMR (100 MHz, CDCl3) δ 7.9 (2C), 10.9, 21.4, 36.2, 53.3, 111.7, 111.9, 123.7, 125.4, 125.5(2C), 125.7, 126.2, 127.3, 127.6, 128.6, 128.8, 128.9, 129.0, 129.1, 130.8, 131.9, 133.9, 134.2, 139.0, 146.3, 153.4, 158.9, 161.8.</p><!><p>The procedure for 8-Cyclopropyl-7-naphthalen-1-ylmethyl-5-oxo-2-m-tolyl-5H-thiazolo[3,2-a]pyridine-3-carboxylic acid (7g) is representative. KOH(s) (20 equiv) was added to 6g (17 mg, 0.0355 mmol) dissolved in THF (1 mL) and MeOH (0.5 mL). The suspension was heated by microwave irradiation for 25 min at 90 °C before being cooled to rt. The reaction mixture was diluted in CH2Cl2 and washed twice with 1M HCl. The combined organic layers were concentrated and purified by column chromatography [CH2Cl2:MeOH 97:3→CH2Cl2:MeOH 95:5 and 1% AcOH] to give 7g (15.2 mg, 92 % yield) which was lyophilized from MeCN/H2O to give a light yellow solid. 1H NMR (400 MHz, CDCl3) δ 0.68-0.75 (m, 2H), 0.94-1.03 (m, 2H), 1.66-1.76 (m, 1H), 2.32 (s, 3H), 4.45 (s, 2H), 6.12 (s, 1H), 7.02-7.09 (m, 1H), 7.14-7.18 (m, 1H), 7.20-7.25 (m, 1H), 7.27-7.38 (m, 1H), 7.37-7.41 (m, 1H), 7.41-7.50 (m, 3H), 7.70-7.80 (m, 2H), 7.82-7.88 (m, 1H). 13C NMR (100 MHz, CDCl3) δ 8.3 (2C), 11.4, 21.5, 36.7, 111.3, 113.7, 124.2, 126.0, 126.1, 126.2, 126.7, 127.9, 128.1, 128.3, 128.7, 129.26, 129.38, 129.43, 129.6, 131.1, 132.4, 134.5, 134.8, 139.4, 148.0, 154.4, 160.0, 164.5.</p><!><p>The PapD-PapH complex (where for PapH, residues 1-22 of the mature protein are removed for stability) was purified and crystallized as previously described.31 Complexes with pilicide 1 and 5d were formed by a 24 hour incubation of PapD-PapH crystals in the crystallization solution (10 mM cobalt chloride, 100 mM MES pH 6.5 and 1.8 M ammonium sulphate) containing 2 mM compound.</p>

PubMed Author Manuscript

Radiosynthesis, ex Vivo Biodistribution, and in Vivo Positron Emission Tomography Imaging Evaluations of [11C]2-Pyridinealdoxime Methiodide ([11C]2-PAM): A First-In-Class Antidote Tracer for Organophosphate Intoxication

2-Pyridinealdoxime methiodide (2-PAM) is a widely used antidote for the treatment of organophosphorus (OP) exposure that reactivates the target protein acetylcholinesterase. Carbon-11 2-PAM was prepared to more fully understand the in vivo mode of action, distribution, and dynamic qualities of this important countermeasure. Alkylation of 2-pyridinealdoxime with [11C]CH3I provided the first-in-class [11C]2-PAM tracer in 3.5% decay corrected radiochemical yield from [11C]CH3I, >99% radiochemical purity, and 4831 Ci/mmol molar activity. [11C]2-PAM tracer distribution was evaluated by ex vivo biodistribution and in vivo dynamic positron emission tomography (PET) imaging in na\xc3\xafve (OP exposure deficient) rats. Tracer alone and tracer coinjected with a body mass-scaled human therapeutic dose of 30 mg/kg nonradioactive 2-PAM demonstrated statistically similar tissue and blood distribution profiles with the greatest uptake in kidney and significantly lower levels in liver, heart, and lung with lesser amounts in blood and brain. The imaging and biodistribution data show that radioactivity uptake in brain and peripheral organs is rapid and characterized by differential tissue radioactivity washout profiles. Analysis of arterial blood samples taken 5 min after injection showed ~82% parent [11C]2-PAM tracer. The imaging and biodistribution data are now established, enabling future comparisons to outcomes acquired in OP intoxicated rodent models.

radiosynthesis,_ex_vivo_biodistribution,_and_in_vivo_positron_emission_tomography_imaging_evaluation

INTRODUCTION<!>RESULTS AND DISCUSSION<!>CONCLUSIONS<!>General.<!>Synthesis of 2-PAM under Limiting Methyl Iodide Reagent Conditions.<!>Radiosynthesis of [11C]2-PAM.<!>Dose Formulation.<!>Rat Biodistribution Studies.<!>Calculated Estimates of Blood Activity in Biodistribution Rat Brain Tissues.<!>Rat Arterial Blood Metabolites.<!>Rat PET-CT and MR Imaging Studies.

<p>The therapeutic agent 2-pyridinealdoxime methiodide (2-PAM, pralidoxime; Figure 1) is widely accepted as an antidote therapy in the United States (US) for the reactivation of acetylcholinesterase (AChE) that has been inhibited by organophosphorus compounds such as chemical warfare agents and certain insecticide oxons (Scheme 1).1–3 The combination of 2-PAM with other centrally acting drugs such as a muscarinic blocker (e.g., atropine4) and antiseizure medication (midazolam5,6) along with other supportive regimens define the standard of care for OP exposures, including self-poisoning.7</p><p>The mechanism by which 2-PAM or P2S (the analogous methanesulfonate salt), in addition to other select oximes, regenerates acetylcholinesterase enzymatic activity occurs via nucleophilic displacement at the phosphorus atom of the serine-O-phosphoryl linkage (Scheme 1) by the oxime C=N−OH moiety. The charged pyridinium ring of 2-PAM plays a supportive role in this mechanism as it is believed to initially guide the antidote to the cation-attracting, peripheral binding domain of AChE and aids docking in the active site similar to that achieved by the natural substrate acetylcholine that also bears a quaternary amine. As an antidote therapeutic to OP intoxication, 2-PAM reacts selectively with OP-modified AChE and other OP-adducted esterases. In the absence of OP-AChE and esterase modifications, 2-PAM is without a specific high affinity in vivo target.</p><p>Despite the favorable role of the cation moiety in the enzyme activity restoration mechanism, it is believed that cationic 2-PAM is primarily distributed to peripheral tissues because the blood−brain barrier (BBB) impedes facile diffusion of the cationic antidote into the central nervous system (CNS).8,9 Diminished CNS access by charged therapeutic agents such as 2-PAM limits their usefulness. However, select studies reported that 2-PAM does reactivate OP-modified AChE in rat brain.10 In support of that work, Sakurada and colleagues reported that 2-PAM enters rat brain in a dose-dependent manner in the absence of OP treatment using microdialysis measures.11 Additionally, it has been shown that [14C]2-PAM enters brain more readily following exposure to the OP trichlofon (Dipterex; O,O-dimethyl, 2,2,2-trichloro-1-hydroxyethyl phosphonate).10,12 Although low concentrations of 2-PAM can enter the brain, it remains less clear how brain levels might vary over time within the first hour of 2-PAM administration, as a function of the absence or presence of OP exposure. In an effort to further interrogate 2-PAM brain penetration under these and related conditions, we sought to generate and initially evaluate carbon-11 (t½ = 20.4 min) radiolabeled 2-PAM ([11C]2-PAM) tracer that would enable facile ex vivo and in vivo CNS and peripheral tissue determinations.</p><p>The [11C]2-PAM tracer was considered ideal for preliminary assessments of pharmacokinetic (PK) properties in rat CNS and peripheral tissues by ex vivo biodistribution (bio-d) and in vivo positron emission tomography (PET) imaging evaluations. The acquired tracer data would ultimately enable correlation to previously established 2-PAM biochemical and toxicological study outcomes.1–3,7,11–14 In this study, we report the radiosynthesis of a first-in-class [11C]2-PAM tracer and initial evaluations of the tracer using naïve rats that have not undergone OP exposure. The assessments include ex vivo bio-d measures, blood metabolite assessment, and in vivo PET imaging determinations after administration of tracer alone (baseline profiles). Additionally, we describe select experiments in which a therapeutic dose level of nonradioactive (cold) 2-PAM was coadministered with the tracer. Together, the naïve rat [11C]2-PAM tracer profiles serve as fundamental data for comparisons to related future tracer measures in OP intoxicated rats.</p><!><p>The high molar activity radiosynthesis of [11C]2-PAM was developed based on literature precedent15 from the readily accessible radiolabeling synthon [11C]methyl iodide ([11C]-CH3I)16 and 2-pyridinealdoxime, such that the tracer could be generated by a one-step alkylation procedure (Scheme 2). Special considerations were taken using the [11C]CH3I approach because most cold preparations of N-methylpyridinium ions use a large surplus of the alkyl iodide and long reaction times, relying on precipitation of the salt for isolation. Preliminary cold chemistry studies using a large inverse ratio of 10- to 50-fold 2-pyridinealdoxime relative to methyl iodide revealed that a high conversion (>80% based on methyl iodide as the limiting reagent) to 2-PAM could be accomplished in less than 5 min at 120 °C in dimethylformamide (DMF). Acetonitrile (CH3CN) was also found as a suitable solvent and led to ~40% conversion to cold product under similar conditions. Methylation in other solvents such as DMSO, N-methylpyrrolidine, and dimethoxyethane (DME) were found to be less effective than DMF or CH3CN when a surplus of 2-pyridinealdoxime was used. The nonradioactive reaction yield in either DMSO or CH3CN was not improved at temperatures below 90 °C nor above 120 °C (sealed tube conditions).</p><p>Adaptation of the modified cold synthetic route to the corresponding radiosynthesis required [11C]CH3I, which was prepared by an established gas-phase automated method.16 In this approach, [11C]CH3I was bubbled into a solution of 2-pyridinealdoxime (1 mg) in CH3CN; the reaction vessel was sealed and heated at 120 °C for 8 min, then cooled briefly, diluted with saline, and the tracer was purified by reversed-phase high performance liquid chromatography (HPLC). The rapid formation of [11C]2-PAM occurred in a decay-corrected radiochemical yield of 3.5 ± 0.9% (n = 8), relative to [11C]CH3I with a measured molar activity of 4,831 ± 911 Ci/mmol (n = 8), >99% radiochemical purity, and in a total synthesis time of 40 min from delivery of [11C]CH3I. A similar protocol using DMF as the reaction solvent at 140 °C afforded a higher decay-corrected tracer yield (16.5 ± 6.5%, n = 3) relative to [11C]CH3I; yet, purification by HPLC was difficult when DMF was used as reaction solvent, resulting in a lower tracer radiochemical purity (85−90%, see: Supporting Information). A limited series of reactions was conducted in an attempt to increase the yield of [11C]2-PAM using CH3CN as the reaction solvent. Adjusting temperature (80−120°C) and mass of precursor (5−10 mg) did not afford an improved yield of [11C]2-PAM (per Supporting Information, Additional Labeling Studies for [11C]2-PAM Preparation, Table S1). Because the use of CH3CN solvent was superior for gaining higher final tracer purity, albeit affording tracer in lower radiochemical yield, the CH3CN protocol was used for the reported preclinical studies. Tracer doses, formulated in ~1 mL sterile isotonic saline, were found to be stable for >2 h at room temperature by HPLC analysis, and up to 10 mCi (nondecay corrected) of final formulated [11C]2-PAM tracer was typically achieved.</p><p>To initially appraise CNS and peripheral tissue tracer uptake profiles, bio-d determinations and PET imaging studies were performed using 250−400 g (mean body weight 305 g) male Sprague−Dawley rats (n = 3 per group) that were administered [11C]2-PAM by tail vein intravenous (i.v.) injection under light isoflurane (1−2%) anesthesia. Biodistribution experiments used 100−200 μCi tracer doses (1.0−1.5 mL) where the initial studies were carried out with tracer alone (baseline conditions). Baseline ex vivo biodistribution radio-activity profiles (blood and tissues) were sampled at select 2− 60 min time points after tracer injection (Figure 2) for blood, brain, liver, heart, kidney, and bone (femur). The samples were counted affording data as decay-corrected percent injected dose radioactivity per gram tissue values (% ID/g) ± standard error measures (SEM). The bio-d radioactivity profiles reveal highest uptake in kidney with significantly diminished amounts in liver, lower uptake in lung and heart, and the lowest radioactivity in blood, brain, and bone. For all samples, time-dependent radioactivity uptake and washout profiles were observed for the tissues and blood over 60 min. The rapid radioactivity uptake and washout seen within naïve rat tissues (i.e., lack of an OP-AChE modification) are consistent with the absence of a high affinity 2-PAM target within the tissues and blood.</p><p>The elevated radioactivity in kidney at 2 and 5 min was anticipated for a cationic, highly water-soluble compound and thought to be associated with clearance.17,18 Limited brain radioactivity uptake observed at all times is considered to be a function of the diminished ability of cationic [11C]2-PAM (and/or related species) to diffuse across the BBB.11,12 Applying the average-sized rat brain blood volume of ~30− 35 μL/g of tissue19 for the blood and brain % injected dose (ID) per gram profiles over time (Figure 2) indicates that ~8−15% of the brain radioactivity detected over the 2−60 min is contributed from the cerebral blood pool radioactivity within brain tissue. Thus, the majority of the brain radioactivity (Figure 2) is associated with tissue (see Methods section and Supporting Information).</p><p>Tracer metabolism events are thought to contribute to the liver radioactivity levels (Figure 2).12,20 Because the highest liver radioactivity was observed at 5 min, evaluation of blood nonprotein-bound fraction components was conducted to assess the amount of parent tracer activity relative to metabolites. Tail artery blood sampling at 5 min, blood sample processing and chromatographic profiling against nonradioactive (cold) chemical metabolite standards afforded a distribution of radioactivity components (Figure 3). At 5 min after tracer injection, the nonprotein-bound fraction showed 82.3% parent [11C]2-PAM tracer and a mixture of metabolites, including 4.4% N-methyl-2-pyridinecarboxaldehyde, 4.7% of a mixture of N-methyl-2-pyridinecarboxamide, N-methyl-2-pyridinecarboxylic acid, N-methyl-2-pyridinecarbonitrile, and 8.6% N-methyl-2-pyridone (Figure 3). The metabolite structures and relative distribution are consistent with 2-PAM metabolites reported previously,12,15,20 albeit with some minor differences that could result from dose sizes and rat strains. The Figure 3 metabolism results provide insight into the maximal 5 min liver bio-d profile. At present, it remains unclear whether uptake of the metabolites into brain occurs, if are produced in brain, or both.</p><p>We sought to further interrogate the 5 min bio-d time point because previous studies with 2-PAM have shown a dose-dependent increase in brain penetration.11 Our initial appraisal used tracer that was coinjected with a 30 mg/kg of cold 2-PAM dose that is akin to a human 30 mg/kg therapeutic 2-PAM dose,1,3,11 where the total mass of cold 2-PAM injected was not allometrically scaled. The experiment was designed to compare the presence versus absence of cold 2-PAM in relation to activity uptake and washout profiles in naïve rats in the absence of an OP-AChE modification. Further, differential tracer tissue interactions associated with metabolism, clearance or other processes might also be detected in the presence of cold therapeutic agent. The results are shown in Figure 4 with baseline (tracer alone) versus 30 mg/kg cold 2-PAM dose coinjected with tracer. One-way ANOVA analyses show no significant differences between the two tracer dose regimens, [11C]2-PAM only (baseline) versus tracer in the presence of 30 mg/kg cold 2-PAM, in most of tissue activity distributions at 5 min except the liver (P = 0.0085). The significant elevated liver activity in the presence of cold 2-PAM is thought to be a result of altered tracer metabolic processes. No significant difference in 2-PAM brain radioactivity uptake between tracer alone and tracer plus therapeutic cold 2-PAM dose experiments was found, suggesting that the uptake was not affected by mass dose level. With the exception of liver and in the absence of OP-AChE modification, the naïve rat activity biodistribution PK profiles resulting from [11C]2-PAM tracer low mass doses (Figure 4) are thought to be similar to those of higher mass doses of nonradioactive 2-PAM.17,18</p><p>To distinguish the in vivo radioactivity tissue distributions in naïve rats, 30 min dynamic in vivo PET scanning (5 min frames) was employed that used 0.5−1.0 mCi tracer i.v. doses. The 30 min scan time was based on the rapid radioactivity uptake and washout defined by the bio-d profiles (Figure 2). The PET scans were carried out to afford in vivo data for comparison to the ex vivo biodistribution profiles. The PET imaging was carried out in parallel with microcomputed axial tomography (CT) and magnetic resonance (MR) scanning for the sake of anatomical tissue information and data coregistration allowing for conservatively defined volumetric three-dimensional regions of interest (ROIs) relative to established rat brain and peripheral tissue atlas designations.21–23 Regional radioactivity signals were determined as standardized uptake values (SUV, see Methods section)24 that permitted imaging evaluations across rats for determinations of the mean (n = 3) SUV values ± SD per ROI vs time (min). The resultant tracer baseline and tracer in the presence of cold 2-PAM time−activity curves (Figure 5) and expanded the brain TAC data (Figure 6) were obtained. A typical sagittal PET view of summed radioactivity coregistered with CT data after baseline [11C]2-PAM tracer injection is depicted in Figure 7.</p><p>Baseline PET imaging TAC plots (Figure 5, Panel A) show high radioactivity in kidney, less in liver, low levels in heart and lung, and lower level activity in whole brain. These relative in vivo SUV profiles are similar to the ex vivo biodistribution radioactivity determinations (Figure 2). An analogous assessment was made when [11C]2-PAM was coinjected with 30 mg/kg of cold 2-PAM (Figure 5, Panel B) showing the similarity (Panel A vs Panel B) and significant amounts of radioactivity washout by 25 min after tracer injection. The relative distributed amounts of radioactivity detected (Figure 5, Panel A and B) are found similar to the ex vivo bio-d profiles of Figure 2; that is, radioactivity distribution profiles as kidney ≫ liver > heart, lung > brain. The elevated radioactivity found in kidney and liver as compared to heart and lung (Figure 5) likely represents greater tissue reservoirs for 2-PAM (and related radiolabeled species) associated with clearance and metabolism, respectively. The Figure 5 mean radioactivity liver values at 5 min after tracer injection do not clearly show a distinction between the Panel A (tracer alone) vs Panel B (tracer in the presence of 30 mg/kg cold 2-PAM), whereas differences are found from the similar liver bio-d determinations as shown in Figure 4. The reasons for the differences remain unclear. The expanded naïve rat brain TAC plots (Figure 6) for baseline and the coinjection of cold 2-PAM experiments reveal that the curves are not significantly different from each other. Radioactivity enters the brain rapidly, albeit at low levels, and reaches a steady concentration in the brain after 5 min. Low brain radioactivity uptake was anticipated based on previous biodistribution reports using naïve rats.11,12</p><p>The PET imaging results demonstrate that the use of the short carbon-11 half-life [11C]2-PAM tracer within naïve rats is suitable for quantitatively detecting radioactivity PK uptake and washout changes in tissues of interest. The baseline PET data reveals radioactivity distributions that are relatively uniform relative to the ex vivo time−radioactivity biodistribution determinations. Within naïve rat tissues, the PET TAC data from the coinjection of nonradioactive 2-PAM is similar to those from baseline profiles, which we consider consistent for the naïve rat condition that is devoid of an OP-AChE adduct (Scheme 1), and lacks a high affinity 2-PAM target.</p><!><p>To gain deeper insights into the pharmacokinetic properties of the therapeutic antidote 2-PAM, the first-in-class [11C]2-PAM tracer was selected for study initially within naïve (OP exposure deficient) rats that lack a high affinity in vivo target for the tracer. The tracer was prepared by a straightforward approach using a one-step alkylation reaction with [11C]CH3I followed by HPLC purification, providing stable (2 h) saline-formulated dose preparations that were administered intravenously. Initial uptake and washout radioactivity PK assessments within CNS and peripheral tissues, and also blood, were made by ex vivo biodistribution profiling over 60 min. Radioactivity distributions were realized for baseline (tracer alone) and at the selected 5 min time with tracer coinjected with cold 2-PAM (human therapeutic dose level, 30 mg/kg). A range of high to low radioactivity was found in blood and tissues; respectively, kidney ≫ liver > lung and heart ≫ blood, brain and bone. The high kidney radioactivity was anticipated for a water-soluble cationic compound, and similarly, the low brain activity was expected based on the diminished ability of cationic species to diffuse across the BBB. Realizing the liver activity maxima at 5 min could be related to previously described 2-PAM metabolic transformations and that this activity level was affected by cold 2-PAM coadministration, arterial blood collection under baseline conditions at 5 min and subsequent processing revealed plasma free-fraction activity components as 82.3% parent [11C]2-PAM tracer and just 17.7% as cognate metabolites previously described in the literature.</p><p>Dynamic in vivo PET imaging for 30 min using [11C]2-PAM tracer under baseline conditons and also tracer coinjected with 30 mg/kg cold 2-PAM revealed CNS and peripheral tissue radioactivity distributions that were relatively similar to the time−radioactivity profiles established by the ex vivo biodistribution experiments. The nonradioactive 2-PAM plus tracer conditions were found without significant differences on brain radioactivity profiles over 30 min, which was expected for naïve rats that lack a high-affinity OP-AChE adduct target. The PET imaging confirms low radioactivity uptake into brain as previously described in the literature for nonradioactive 2-PAM. In summary, the naïve rat ex vivo biodistribution and in vivo PET imaging time−radioactivity measures after [11C]2-PAM tracer i.v. injection serve as fundamental correlated data sets. This requisite data is now poised to enable meaningful comparisons relative to tracer determinations in OP-intoxicated rats in which the detection of [11C]2-PAM participation in the critical therapeutic regeneration of acetylcholinesterase activity will be probed.</p><!><p>Reagents (e.g., 2-pyridinealdoxime, iodomethane, etc.) and solvents such as anhydrous dimethylformamide (DMF), anhydrous acetonitrile (CH3CN), and 2-pyridinealdoxime methiodide were reagent grade or better, used without any additional purification, and were purchased from Sigma-Aldrich Chemical Co. (Milwaukee, WI). USP-grade sterile isotonic saline was also purchased from Sigma-Aldrich. 2-Pyridinecarboxaldehyde, 2-pyridinecarboxamide, 2-pyridi-necarboxylic acid, and 2-pyridinecarbonitrile were converted into the N-methyl pyridium metabolite standards per literature methods.25,26 Nuclear magnetic resonance (NMR) data were recorded in CDCl3 on a Varian Avance 400 MHz spectrometer. High performance liquid chromatography (HPLC) was performed with a Waters 590 system (Milford, MA) coupled to a Shimadzu SPD UV−visible detector (Columbia, MD) and a gamma counting in-line radiation flow detector (Model 105s, CRA; Berkeley, CA). The HPLC data was collected with a SRI Peaksimple, model 304, data system (Torrance, CA). Counting of tissue and blood samples utilized a Hidex automated gamma counter (Turku, Finland).</p><p>Male Sprague−Dawley rats (250−400 g; Charles River, Inc., Skokie, IL) were used for the [11C]2-PAM biodistribution, blood, and PET imaging studies. The animals were cared for and used at the University of California, San Francisco (UCSF) facilities that are accredited by the American Association for Accreditation of Laboratory Animal Care (AAALAC). The animal studies adhered to UCSF IACUC approved protocols that satisfied NIH guidelines and institutional regulations. Prior to injection of [11C]2-PAM, rats were lightly anesthetized (~1−2% isoflurane), and lateral tail vein catheters were installed in the lower tail portion. The catheter line was flushed with 200 μL of saline and capped. Thereafter, the catheter cap was removed, tracer was injected as a bolus, and then the catheter was flushed with 0.3 mL saline.</p><!><p>2-Pyridinealdoxime (10 mg; 0.082 mmol) was dissolved in 200 μL of DMF at 125 °C or CH3CN at 80 °C and added dropwise to a solution of methyl iodide (2.0 μL) in 50 μL of DMF or CH3CN over 5 min. After the addition, the reaction was kept at the indicated temperature for 5 min. The percent conversion to 2-PAM was 80−90% (DMF) and 35−40% (CH3CN) as determined by NMR integration of aldoxime CH=N peaks or by the ratio of the N−CH3 to aromatic region in d6-DMSO. Conversion to 2-PAM in CH3CN was 40 ± 4% (using 2-PAM as internal standard) as measured by reversed-phase HPLC (UV at 290 nm) following evaporation of the solvent in vacuo, using a Sonoma C18(2) 5 μ 100 Å 15 cm × 4.6 mm (ES Industries; West Berlin, NJ) column with a mobile phase of 10% C2H5OH/10 mM 2-(N-morpholino)ethanesulfonic acid in H2O at1.0 mL/min.</p><!><p>2-Pyridinealdoxime (1−10 mg) was dissolved in 300 μL of anhydrous CH3CN in a 4 mL 1-dram vial and sealed with a Teflon-lined silicone septum screw-cap. Utilizing a gas-phase method16 and a GE Tracerlab FX C Pro radiosynthesis box, [11C]CH3I was produced and then bubbled into the CH3CN solution, and the reactor vial was sealed. The mixture was heated at 80−140 °C for 8 min on an aluminum block heater. The reactor vial was removed from the heating source and allowed to cool for 60 s. The cap was removed; the CH3CN solution was diluted with 2.5 mL of saline, and the crude material was injected onto a 10 × 250 mm, 10 μm Hamilton PRP-1 HPLC column. The HPLC purification used saline as mobile phase at a flow rate of 3 mL/min with a retention time of [11C]2-PAM at 8.5 min (see Supporting Information, Chromatograms section for representative HPLC elution profiles, and the Additional Radiolabeling Studies for [11C]2-PAM Production section).</p><!><p>Doses of [11C]2PAM were formulated in sterile isotonic saline (pH ~ 7) as 100−200 μCi in 1.0−1.5 mL for the biodistribution and blood determinations, and also 0.5−1.0 mCi in 0.5−1.0 mL for the PET imaging evaluations; in which the average dose volume across both sets of experiments was ~1.0 mL. Before animal dosing, quality control of the dosing was accomplished by analytical HPLC (7 μm Hamilton PRP-1 4.6 × 250 mm flow rate 1 mL/min) demonstrating >99% radiochemical purity and >95% chemical purity. In separate experiments, doses were allowed to remain at room temperature for 2 h and then subjected to reversed-phase analytical HPLC analysis (see Supporting Information).</p><!><p>Rats received a bolus injection of 100−200 μCi of [11C]2-PAM via tail vein catheter port followed by a 0.3 mL saline flush. Rats were euthanized by lateral chest puncture under anesthesia and dissected at 2, 5, 10, 30, and 60 min after injection (n = 3 per time point). Blood was collected by cardiac puncture immediately prior to euthanasia at each time point. Whole organs of brain, liver, heart, kidneys, lungs, bone (femur), and blood were collected into 20 mL preweighed glass scintillation vials and capped, where they were placed into a Hidex automated gamma counter (Hidex AMG) to be counted and weighed. Decay corrected data was exported from the Hidex instrument to Excel (Microsoft Office) software and plotted as decay corrected percent injected dose per gram (% ID/g) ± SEM versus time (min). To compare the % ID/g values of the two tracer dose regimens (i.e., tracer alone and tracer in the presence of 30 mg/kg of cold 2-PAM) one-way ANOVA analyses were performed for each tissue and blood using R software (version 3.0 or later).</p><!><p>Rat blood and brain tissue biodistribution radioactivity data from 2 to 60 min were analyzed to estimate the amount of the percent contribution of blood signal in brain tissue. Everett19 previously reported that average-sized rat brain blood volume is ~30−35 μL/g of tissue. Using a midrange brain blood volume of 32.5 μL/g (equal to 0.0325 mL/g), and blood and brain % ID/g values per time point (Figure 2; per Supporting Information, Numerical Blood and Brain Data) with a blood density of 1 g/mL afforded the following relationships (eqs 1 and 2) and ratio (eq 3), where t = time point measured (for example at t1): (1)blood at t1(%ID/g)=blood %ID/g  at t1× 1 g/1 mL×0.0325 mL/g (2)brain at t1 (%ID/g)=barin  %ID/g at t1 (3)% of blood in brain tissue at t1(%ID/g)=blood at t1(%ID/g)/brain at t1 (%ID/g)×100</p><!><p>Rats were installed with a tail artery catheter at the upper tail region for blood sampling, and the catheter was flushed with saline heparin solution prior to blood collection. Tracer (1−3 mCi) was injected via a lower tail vein catheter, and then 5 min later arterial blood (100 μL) was collected by heparin treated syringe. Blood samples were placed in heparin treated plastic 1.5 mL snap-lid vials containing 10 μL of 10 mg/mL citric acid and centrifuged at 13 200 g for 1 min, and remaining supernatant was removed and placed in a second tube. The supernatant was treated with 100 μL of acetonitrile to precipitate protein; the protein pellet was removed after centrifugation, and the serum supernatant was collected. Chromatographic separation of serum radioactive components was conducted by preparative thin layer chromatography (TLC, 3.5 × 10 cm aluminum-backed silica with fluorescent indicator (254 nm), mobile phase = n-butanol:acetic acid:water (5:1:3)) to accommodate the carbon-11 half-life. The chromatography was run in parallel with known cold chemical metabolite standards12,15,20 that were prepared in-house by literature methods.25,26 Co-migration of standards with radioactive spots were excised with scissors, placed in 20 mL glass scintillation vials, and counted with the Hidex gamma counter. The data were exported from the Hidex instrument to Excel software and reported as relative % of total radioactivity, defined as the decay corrected ratio of component radioactivity over the sum of all radioactivity counted across the samples × 100.</p><!><p>Tracer [11C]2-PAM i.v. bolus doses of 0.5−2.0 mCi as formulated in 0.5−2.0 mL of saline were administered. The cold 2-PAM PET imaging experiments were accomplished by coinjecting 2-PAM at 30 mg/kg, saline (1−2 mL) with tracer. The PET, CT, and MR imaging were performed northermic (37 °C) under isoflurane anesthesia (1−1.5%). The PET and CT imaging data were acquired with a Siemens Inveon microPET/CT scanner system (ca. 1.5 mm PET imaging spatial resolution). Dynamic PET imaging data were acquired over 30 min beginning ~0.5 min after the time of injection of tracer. The PET data were reconstructed with using a Siemens Inveon reconstruction program suite, including OSEM2D; as 6 frames, 300 s per frame for the 30 min scans, decay time corrected, and quantified with a radiation phantom instrument calibration factor. A partial volume correction was not applied, and conservative ROI definitions were used as described below. The CT data were acquired in standard rat mode: 80 kVp, 225 mA; 400 ms exposure, 194 steps × 194 degrees, and 97-μm isotropic resolution. The MR data were acquired with a Bruker Biospin 7-T magnet using a multislice 2D FLASH protocol with the following parameters: T2*-weighted gradient recall echo, TR = 1528.3 ms, TE = 7 ms, and 256 × 256 × 50 voxels, affording 16 μm3 resolution.</p><p>The reconstructed MR, CT, and PET imaging data files were processed with AMIDE open source software27 (SourceForge), version 1.0.4 (or later versions). The MR and CT images were oriented as defined by Paxinos21,22 and Walker.23 Cranial landmarks of bregma and lambda were identified from the CT images. The x, y, and z coordinates of imaging views were centered at bregma, equivalent to the origin of the first scan, and then consistent landmark structures were iteratively coregistered and template fit against the cranial structures of the first scan landmarks. Subsequently, the landmarks were correlated with cerebral soft tissues from the MR scans.</p><p>The coregistered imaging data permitted ROIs (Figures 5–7) to be defined within their ROI volume size limits and locations and against established stereotaxic three-dimensional locations.21–23 The ROIs were defined as follows: whole brain, heart, lung, kidney, and liver. PET scan regional tissue radioactivity is reported as Standardized Uptake Value (SUV) defined as [activity concentration in the tissue region of interest (MBq/cc)/decay corrected injected dose at time = 0 (MBq)] × body weight of the rat as gram (g).24 ROI PET scan statistics (SUV ± SD) collected from time points at midframe were exported to Excel, and plots of SUV versus time were generated using either Excel or GraphPad Prism (La Jolla, CA) software.</p>

PubMed Author Manuscript

Nickel(0)-Catalyzed Linear-Selective Hydroarylation of Unactivated Alkenes and Styrenes with Arylboron Acids

Herein, we realized the first linear-selective hydroarylation of unactivated alkenes and styrenes with organoboronic acids by introducing directing group on alkenes. Our method is highly efficient and scalable, and provides a modular route to assemble structurally diverse alkylarenes, especially for γ-aryl butyric acid derivatives, which have been widely utilized as chemical feedstocks to access multiple marketed drugs, and biologically active compounds.Supporting information for this article is given via a link at the end of the document.((Please delete this text if not appropriate))

nickel(0)-catalyzed_linear-selective_hydroarylation_of_unactivated_alkenes_and_styrenes_with_arylbor

<!>Scheme 4. Four additional transformations with our product 3b

<p>Transition metal catalyzed alkene hydroarylation involving metal hydrides (M−H) [1] has become one of the most widely adopted strategies to alkylarenes. [2] Three types of different aryl sources have been used for this purpose: 1) simple arenes; 2) aryl halides; 3) arylmetals (i.e., aryl stannanes or boronic acids). Simple arenes as the aryl source in alkene hydroarylation are usually restricted to heteroarenes, and arenes with a directing group because of the challenge on the inert C−H bond activation of simple arenes. [3] On the other hand, the indispensable need of O2 or overstoichiometric amount of reductants such as silanes and alkyl halides somewhat also reduces the practicality in alkene hydroarylations with aryl halides or arylmetals (Scheme 1a). [4,5] A significant breakthrough was made by Zhou et al. for the highly selective Ni-catalyzed hydroarylation of styrenes and 1,3-dienes with arylboron compounds under redox-neutral conditions (Scheme 1b). [6] However, Zhou's system has still suffered from some limitations: 1) the system was only efficient with styrenes and 1,3-dienes; 2) only branched products can be achieved. Inspired by the recent work in olefin functionalziation with the control of the reactivity and regioselectivity via the introduction of an additional coordinating group on diverse alkenes, [7,8] herein, we disclose the first directing groupcontrolled linear-selective hydroarylation of diverse alkenes with organoboron compounds under redox-neutral conditions (Scheme 1c). Our method is highly efficient and scalable, and provides a modular route to assemble structurally diverse alkylarenes, especially for γ-aryl butyric acid derivatives, which have been widely utilized as chemical feedstocks to access multiple marketed drugs, and biologically active compounds (Scheme 1). γ-Aryl butyric acids are classically formed by Friedel-Crafts reaction with butyrolactones, [9] which suffers from several disadvantages such as harsh reaction conditions, limited substrate scope (only suitable for electron-rich arenes) and weak regioselectivity.</p><p>In our preliminary experiment, 3-butenoic acid derivative 1a bearing 8-aminoquinoline (AQ)-directing group (0.2 mmol) was treated with Ni(COD)2 (5 mol%), PCy3 (10 mol%) and 2naphthalenylboronic acid 2a (0.3 mmol) in MeOH (1 mL) at 70 °C for 24 h, which gave the linear product 3a in 26% yield (Table 1, entry 1). Change of the MeOH to t-AmylOH slightly improved the yield (Entry 2). Furthermore, when we added the 1.5 equiv. K2CO3 to the reaction, the yield was further improved to 38% (entry 3). Subsequently, the different bases were evaluated (entries 3 −7). CsOPiv was proved to be the best choice. We also found that the modification of the mole ratio of 1a and 2a to 1:2 increased the yield to 61% (entry 8). Finally, the effect of the various phosphine ligands was investigated (entries 9−12). When we switched the ligand from PCy3 to PPh3 the reaction dramatically improved (85%, entry 9). Control experiment confirmed that the reaction does not occur in the absence of Ni(COD)2 (entry 13).</p><p>Next, the scope of aryl boronic acids was investigated using 3-butenoic acid derivative 1a as the alkene (Scheme 2). To our delight, whether aryl boronic acids are electron-rich, electronpoor, or sterically hindered, all of them afforded good to excellent yields (3a-u). Heteroaryl boronic acids were also reactive (3v&w). Besides diverse (hetero)aryl boronic acids, alkenylboric acid also showed good reactivity in this reaction selectively delivering the (E)-6-phenylhexenoic acid derivative 3x under our optimized condition. In addition, high-yielding hydroarylation of 1a can also be obtained by using arylboroxine, arylboronic ester as the aryl source (3b). It is important to stress that these reaction conditions were remarkably compatible with a variety of functional groups such as halogens (F, Cl, and Br), acetyl, cyano, and methoxy groups on the aryl boronic acids, which could be subjected to further synthetic transformations (3is). In addition, we proved that the reaction can easily be employed on gram-scale without any decrease in yield (1.5 g of 3b, 86%). The structure of 3b was confirmed by single crystal xray diffraction. [10] We next examined the scope of unactivated alkenes (Scheme 3). The terminal alkenes bearing monosubstitution at the α-or βposition proceeded smoothly to afford the desired products in good to excellent yields (Scheme 3, 4a-b). The sterically congested α,α-disubstituted terminal alkene is also reactive to furnish the desired products 4c in moderate yield, although higher temperature (125 º C) and PhPMe2 (10 mol%) were required. Furthermore, a variety of internal alkenes bearing diverse substitution at γ-position could also be efficiently converted into the corresponding products with nearly neglectable steric hindrance and electronic effect of substituents (Scheme 3, 4d-l). Notably, both the trans-and cis-3-hexenoic acid derivatives produced the same product 4e with the same yields. In addition, α,γ-disubstituted alkene could also give the desired products (Scheme 3, 4i-j). Surprisely, the γ,γdisubstituted alkene can also be hydroarylated to produce compound 4l bearing a quaternary carbon center in 88% yield. Furthermore, we wondered if the introduction of the AQdirecting group at the ortho-position on styrene is also capable of switching the branched selectivity of Sigman & Zhou's system on hydroarylation of styrenes with aryl metals to linear selectivity. To our delight, after an extensive screening, we found that the reactions of styrene derivative 1m with various aryl boronic acids occurred smoothly with fully linear-selectivity in moderate yields by employment of the PhPCy2 as the ligand at 125 °C (Eq. 1), which clearly proved the importance of the directing group on the control of the regioselectivity of the hydrometalation.</p><p>Our products as the synthetic intermediates are capable of accessing structurally diverse important building blocks by additional transformations (Scheme 4). The AQ-directing group on compound 3b can be removed smoothly to provide 4phenylbutanoic acid 5b with 97% yield. The acid 5b could easily undergo the intramolecular cyclization to yield the tetralinderivative 6a. The carboxyl group could also be transformed to Iodine to yield compound 6b. The five-membererd lactone 6c can also be formed by treatment of the acid 5b with hypervalent iodine reagent. In addition, the acid 5b can be easily attacked by CH3NO2 to afford the corresponding ketone-derivative 6d. To further demonstrate the synthetic usefulness, we synthesized the marketed drug Sensipar from the hydroarylated product 3s (Scheme 4). The free acid 5s was first synthesized by removing the AQ-directing group. Then, it was transformed into the corresponding alkyl iodide 6s by treatment with NIS/I2 in DCE. A simple amination was then performed to introduce the amine moiety, giving Sensipar in 97% yield.</p><!><p>To probe the reaction mechanism, a series of control and deuteration experiments were carried out. First we conducted the reaction with N-(naphthalenyl)butenamide 7 and 2naphthalenylboronic acid 1a which is absolutely inactive (Scheme 5, eq.1). This result indicates that the attachment of the AQ-directing group on unactivated alkene is indispensable for this transformation. Treatment of PhB(OD)2 2a-D with the alkene 1a in MeOH or (PhBO)3 with 1a in CD3OH gives the undeuterated product 3b (Scheme 5, eq.2-3). However, the use of MeOD as the solvent led to the formation of β,γ-position Finally, reductive elimination of D gives the desired product and regenerates the Ni 0 catalyst. Alternatively, the intermediate B could also undergo a hydrometalation to form five-memebered nickelacyclic intermediate C, which would lead to the branched product. we have never detected the branched product, which might because the rigidity of the five-memebered nickelacyclic intermediate C is unfavourable for the subsequent transmetalation. [11] Scheme 6. Proposed mechanism of the nickel(0)-catalyzed directing grouphydroarylation of unactivated alkenes with organoboronic acids.</p><p>In conclusion, we have developed the first nickel(0)-catalyzed directing group-controlled linear-selective hydroarylation of unactivated alkenes and styrenes with organoboronic acids under redox-neutral conditions. Our method is highly efficient and scalable, and provides a general strategy to structurally diverse alkylarene derivatives, especially for γ-aryl butyric acid derivatives, which have been widely utilized as chemical feedstocks to access multiple marketed drugs, and biologically active compounds.</p>

ChemRxiv

MOLECULAR ANALYSIS OF JUVENILE HORMONE ANALOG ACTION IN CONTROLLING THE METAMORPHOSIS OF THE RED FLOUR BEETLE, Tribolium castaneum

The juvenile hormone analogs (JHA) are known to disrupt insect development but the molecular mechanisms of their action have been studied only in a few model insects belonging to orders Diptera and Lepidoptera. Here, we investigated the mechanisms of JHA action in red flour beetle, Tribolium castaneum, belonging to the order Coleoptera. Application of JHA during penultimate and final instar larval stages blocked larval-pupal metamorphosis and induced supernumerary larval molts. When compared to the control insects undergoing larval-pupal molt, down-regulation of expression of transcription factor, Broad, and up-regulation of other genes involved in 20-hydroxyecdysone (20E) action (FTZ-F1, E74) were observed in JHA-treated larvae undergoing supernumerary larval molts. The presence of JHA during the final instar larval stage blocked the midgut remodeling wherein programmed cell death (PCD) of larval cells and proliferation and differentiation of imaginal cells to pupal gut epithelium were impaired. The comparative analysis of 20E-induced gene expression in the midguts of JHA-treated and control insects revealed that JHA suppressed the expression of EcRA, EcRB, Broad, E74, E75A, and E75B, resulting in a block in PCD as well as proliferation and differentiation of imaginal cells.

molecular_analysis_of_juvenile_hormone_analog_action_in_controlling_the_metamorphosis_of_the_red_flo

INTRODUCTION<!>Rearing and Staging<!>Hormonal Treatment<!>cDNA Synthesis and Quantitative Real-Time Reverse-Transcriptase PCR (qRT-PCR)<!>Histology<!>Imaging and Documentation<!>Juvenile Hormone Analogs (JHA) Block Larval-Pupal Metamorphosis<!>Differences in Gene Expressions During Supernumerary Larval Molt and Larval-Pupal Molt<!>Window of Sensitivity to Hydroprene<!>JHA Blocks Midgut Remodeling and Interferes With 20E-Induced Gene Expression<!>DISCUSSION

<p>Insect development is regulated by several hormones including the steroid, 20-hydroxyecdysone (20E), and sesquiterpenoid, juvenile hormone (JH). During larval/nymphal development in lepidopteran and hemimetabolous insects, JH prevents metamorphosis (Williams, 1961; Zhou and Riddiford, 2002). At the end of larval development, JH titers decrease, enabling 20E to trigger metamorphosis (Riddiford, 1996). In most insects the addition of JH at this time causes the formation of a supernumerary larva (Truman and Riddiford, 2002). However, the scenario is different in dipteran insects such as Drosophila melanogaster and mosquitoes, wherein the application of JH analog (JHA) did not block puparium formation or pupation (Wilson, 2004; Wu et al., 2006). Though the molecular mechanisms by which JHA affects metamorphosis is well understood in D. melanogaster, Manduca sexta, and Aedes aegypti (Riddiford, 1996; Zhou et al., 1998a; Nishiura et al., 2005), the information gained may not be applicable to all insect species since the JH actions or sensitivity to JH varies among insect species.</p><p>We studied the molecular mechanisms of JHA during larval-pupal metamorphosis in the red flour beetle, Tribolium castaneum. T. castaneum is a coleopteran insect representing 25% of animal kingdom species. Like other holometabolous insects, it develops from egg to adult through the intermediate larval and pupal stages. Besides being a stored grain pest, T. castaneum is amenable to molecular genetic studies. Completion of whole genome sequencing (Tribolium Genome Sequencing Consortium, 2008) and functioning of systemic RNAi (Tomoyasu and Denell; 2004; Arakane et al., 2005; Parthasarathy et al., 2008b) make T. castaneum an ideal model insect. When compared to other model insects, detailed analysis of the development and molecular mechanisms of hormonal regulation on metamorphosis are not available for T. castaneum.</p><p>We investigated the role of JH on the metamorphosis of T. castaneum by using JHA to mimic JH action. Application of JHA blocked larval-pupal metamorphosis and prolonged larval life-span by inducing supernumerary larval molts. The gene expression varied significantly between insects undergoing supernumerary larval molt or larval-pupal molt. Based on the sensitivity to JHA, the critical period of pupal commitment most likely occurred between 72–96 h after ecdysis to the final instar larval stage. The presence of JHA during the final instar larvae blocked midgut remodeling and suppressed the expression of genes involved in 20E action in the midgut. Thus, this study provides a basis to understand the molecular mechanisms of hormonal regulation of metamorphosis in coleopteran insects.</p><!><p>The rearing and staging of GA-1 strain of T. castaneum was done as described in Parthasarathy et al. (2008a). The final instar larvae were identified as soon as they molted by untanned white cuticle, designated as 0 h AEFL (after ecdysis into the final instar larval stage). The following days of final instar larvae were designated as L24, L48, L72, and L96 h AEFL. The beginning of the quiescent stage was designated as Q0 and was determined based on cessation of feeding and movement. The following days were recognized by characteristic "C"-shaped larvae and were collected at 12-h intervals from Q0. White pupae were designated as 0 h AEPS (after ecdysis into the pupal stage) and staged at 24-h intervals. The supernumerary larval stage was designated as L'.</p><!><p>Methoprene (isopropyl (E,E)-RS)-11-methoxy-3,7,11-trimethyl-2,4-dodeca-dienoate) and Hydroprene (Ethyl (2E,4E,7S)-3,7,11-trimethyl-2,4-dodecadienoate) were a gift from Wellmark International (Dallas, TX). Technical grade compounds were dissolved in acetone and used at 0.1 ml/g of diet for all dosages in feeding bioassays. For topical application, 0.5 μl of Hydroprene (2 μg/μl) in acetone was applied on the dorsal side of the thorax and abdomen of final instar larvae prior to 24-h AEFL. All control larvae were treated with equivalent amounts of acetone alone.</p><!><p>Total RNA was extracted from whole body and midguts of staged larvae and pupae using TRI reagent (Molecular Research Center Inc., Cincinnati, OH). cDNA was synthesized using 2 μg of DNAse1 (Ambion, Austin, TX) -treated RNA and iScript cDNA synthesis kit (Biorad Laboratories, Hercules, CA) in a 20-μl reaction volume as per the manufacturer's instructions. Real-time quantitative reverse-transcriptase PCR was performed using MyiQ single color real-time PCR detection system (Biorad Laboratories). PCR reaction components were: 1 μl of cDNA, 1 μl each of forward and reverse sequence-specific primers, 7 μl of H2O, and 10 μl of supermix (Biorad Laboratories). The sequences of primers used here have been reported in Parthasarathy et al. (2008a,b) and Tan and Palli (2008a,b). PCR conditions were: 95°C for 3 min followed by 45 cycles of 95°C for 10 sec, 60°C for 20 sec, 72°C for 30 sec. Both the PCR efficiency and R2 (correlation coefficient) values were taken into account prior to estimating the relative quantities. Relative expression levels of each gene were quantified using ribosomal protein, rp49, expression levels as an internal control.</p><!><p>The midguts from staged larvae and pupae were dissected in 1 × PBS (phosphate buffered saline, Sigma) and fixed in 4% paraformaldehyde (Sigma). Sectioning was done as previously described (Parthasarathy and Palli, 2007). The sections were deparaffinized through successive baths of Xylene, rehydrated through serial grades of ethanol, water, and 1 × PBS. Nuclear staining was done with DAPI (4′, 6-Diamidino-2-phenyl indole, Sigma) at 1 μg/ml concentration for 10 min. The slides were washed with 1 × PBS twice and mounted in 50% glycerol.</p><!><p>For light microscopy, the modular zoom system (Leica Z16 APO, Germany) fitted with JVC 3CCD Digital Camera KY-F75U was used. The images were documented using Cartograph version 6.1.0 (GT Vision Demonstration). Image processing was done using Archimed version 5.2.2 (Micovision Instruments).</p><p>For fluorescent images, an Olympus FV1000 laser scanning confocal microscope was used. DAPI was excited using a 405-nm laser line. The primary objective used was an Olympus water immersion PLAPO40XWLSM-NA1.0. Image acquisition was conducted at a resolution of 512 × 512 pixels and a scan-rate of 10 ms/pixel. Control of the microscope, as well as image acquisition and exportation as TIFF files, was conducted using Olympus Fluoview software version 1.5. Figures of all micrographs were assembled using Photoshop 7.0.</p><!><p>The newly molted penultimate and final instar larvae were starved for 4 h and fed with diet containing different doses of methoprene and hydroprene continuously and the developmental events were recorded (Fig. 1A and B). At higher dosages of 5 and 10 ppm, methoprene blocked larval to pupal metamorphosis completely (100%). Most of the larvae molted into supernumerary larval instar (Fig. 1A). At 1 ppm, methoprene blocked larval to pupal metamorphosis in 85% of larvae treated during the penultimate larval stage and more than 95% of larvae treated during the final instar larval stage. The remaining 5–15% of larvae treated during both stages pupated and died subsequently and no adults emerged from these pupae. Fifty to sixty percent of insects treated with a 0.1-ppm dose of methoprene during the penultimate and final instar larval stage molted into the supernumerary larval stage. The remaining insects pupated and adults emerged from these pupae. All control larvae developed normally into adults (Fig. 1A).</p><p>The effect of hydroprene was similar to that of methoprene at a high dose of 10 ppm (Fig. 1B). At a 1-ppm dose, hydroprene blocked larval-pupal metamorphosis completely (100%) in larvae treated during the penultimate and final instar stages. Even at low dosages of 0.1 and 0.5 ppm, hydroprene blocked pupal metamorphosis in more than 80% of larvae irrespective of stage of treatment. The remaining 20% of insects that became pupae did not survive to adulthood (Fig. 1B). Hydroprene at 0.5 ppm blocked larval-pupal metamorphosis in more than 90% of final instar larvae. Hence, the above JHA, dose, and stage were used in subsequent experiments.</p><p>Continuous feeding of hydroprene at 0.5 ppm during the final instar larval stage prolonged the larval life-span by inducing supernumerary larval molts (Fig. 2). The final instar larvae molted into two subsequent larval stages at weekly intervals and died finally. The supernumerary larval molts produced giant larvae that showed a darker color integument than the integument of control final instar larvae. The control final instar larvae treated with acetone alone entered the quiescent stage and pupated within a week.</p><!><p>The relative mRNA levels were determined by real-time quantitative RT-PCR analysis (Fig. 3). Interestingly, the gene expression varied between the supernumerary larval molt and the larval-pupal molt. The mRNA levels of FTZ-F1 (Fuzhi-tarazu) were up-regulated significantly by 11.6-fold during the supernumerary larval molt when compared to the larval-pupal molt, while expression of EcRB (Ecdysone receptor), E75B (Ecdysone-inducible gene), E74, and HR3 (Hormone receptor) were up-regulated by 1.5–2.5-fold more in the supernumerary larval molt than in the larval to pupal molt. The mRNA levels of Broad were down-regulated by 2.3-fold in the supernumerary larval molt in comparison with the larval-pupal molt. However, the differences in gene expression of only Broad, FTZ-F1, and E74 alone were statistically significant and the expression of EcRA, E75A, and Met were not significantly different between the supernumerary larval molt and the larval-pupal molt.</p><!><p>To determine the critical period of JH sensitivity, 0.5 ppm hydroprene was administered with diet at different time points during the final instar larval stage and the effect on the development was recorded (Fig. 4A). The final instar larvae were sensitive to hydroprene until 60-h AEFL. More than 90% of the treated insects remained as larvae by molting into the supernumerary larval stage. Administration of hydroprene at 72- and 96-h AEFL did not block larval-pupal metamorphosis since more than 90% of the treated larvae became pupae but the pupae eventually died and no adults emerged from these pupae. All the control larvae pupated and emerged as adults.</p><p>The final instar larvae slowly reduced food consumption and stopped feeding between 72–96 h AEFL upon gaining the critical weight and also as a mark of prepupal behavior. The effect of hydroprene by oral feeding after 72-h AEFL, as observed above, may not represent the actual response of the final instar larvae. Hence, topical application on the integument was performed starting at 60-h AEFL at 12-h intervals. Topical application at 60-h AEFL served as a positive control (Fig. 4B, a). Surprisingly, the topical application of hydroprene showed various phenotypes. Hydroprene blocked larval-pupal metamorphosis when applied at 72-h AEFL and all the larvae died during the quiescent stage (Fig. 4B, b). Application of hydroprene at 84-h AEFL resulted in larval-pupal intermediaries. The ecdysis was completed and the resultant stage had underdeveloped wings (Fig. 4B, c). Hydroprene did not interfere with larval-pupal metamorphosis when applied at 96-h AEFL; however, the pupae were malformed (Fig. 4B, d) and died subsequently. All the control larvae pupated.</p><!><p>We examined the morphology of the gut epithelium by nuclear staining of the cross-sections of midgut dissected from hydroprene-treated and control insects (Fig. 5). Application of hydroprene at the beginning of the final instar larvae blocked midgut remodeling. In the treated insects, the midgut epithelium consisted of large larval cells with a few imaginal cells on the periphery at the end of the final instar larval stage (Fig. 5, a). The larval cells had intact nuclei with no signs of PCD indicated by fragmented nuclei. In contrast, in the control insects, the larval cells moved into the gut lumen with fragmented nuclei and the small imaginal cells differentiated to form the pupal midgut epithelium (Fig. 5, b). The cross-sections of midguts dissected from supernumerary larvae treated with hydroprene during the final instar larval stage resembled the morphology of midgut epithelium of the early stage final instar larvae except for the number of imaginal cells on the periphery (Fig. 5, c and d). The mRNA levels of genes such as EcRA, EcRB, Broad, E74, E75A, and E75B were compared in midguts dissected from hydroprene-treated and control insects using qRT-PCR (Fig. 6). The expression levels were monitored at different time points during the final instar larval stage including the quiescent stages and after pupation in the control larvae and the corresponding time points in the hydroprene-treated larvae. In the control midguts, the expression of EcRA, EcRB, Broad, E74, E75A, and E75B showed a peak at the end of the quiescent stage before pupation. The expression levels of all these genes were low after pupation. In hydroprene-treated larval midguts, the expression of all these genes was suppressed. At the end of the quiescent stage, the mRNA levels of EcRA were 6-fold less in midguts dissected from hydroprene-treated insects when compared to the levels in midguts dissected from acetone-treated insects. Similarly, at this stage the EcRB mRNA levels were 2.5-fold less in midguts dissected from hydroprene-treated insects when compared to the levels in midguts dissected from acetone-treated insects. Broad mRNA levels remained low throughout the period of observation. E74 had an expression pattern similar to EcRB. E75A expression levels were relatively low when compared to control. The mRNA levels of E75B were similar to control until 12 h after entering the quiescent stage, but the peak expression level at 36 h after entering the quiescent stage was only 1.5-fold less than in the control.</p><!><p>Numerous analogs of juvenile hormones have been synthesized and a few are being used for insect pest control (Dhadialla et al., 1998). The discovery that exogenously applied juvenile hormone could interfere with metamorphosis in insects (Wiggles-worth, 1965; Williams, 1961) has also been useful in understanding the mechanisms by which hormones control metamorphosis (Riddiford, 1996). There are several reports of JHA action in insects belonging to different orders (Zhou et al., 1998a; Wu et al., 2006; Parthasarathy and Palli, 2007). Though the action of JHA as a potential insecticide has been demonstrated in several coleopteran stored grain insect pests (Kostyukovsky et al., 2000; Arthur, 2001; Toews et al., 2005), the molecular mechanisms of JH action remains unknown. Here we used JHA in T. castaneum to mimic JH action and studied its role in metamorphosis.</p><p>In this study, the sensitivity of T. castaneum to JHA such as methoprene and hydroprene varied, hydroprene being more effective than methoprene. Riddiford and Ashburner (1991) showed that pyripoxyfen was a more powerful JH agonist than methoprene in D. melanogaster. The side chains (methyl, ethyl, isopropyl) of these compounds make them resistant to metabolism in insects and determine the half-life/persistence in the treated insects (Hammock and Quistad, 1981). It is also possible that the proteins involved in JH action recognize and respond to these molecules at different efficiencies resulting in differential action. Unlike in dipteran insects such as D. melanogaster and Ae. aegypti (Wilson, 2004; Wu et al., 2006), the presence of JHA during the penultimate or final instar larvae of T. castaneum blocked larval-pupal metamorphosis and induced supernumerary larval molts. This action of JHA in coleopteran insects is similar to that in lepidopteran insects (Zhou et al., 1998a; Parthasarathy and Palli, 2007).</p><p>The latest larval stage at which application of exogenous JH results in the delay or blockage of pupation is known as the JH-sensitive period and this varies among insects (Webb and Riddiford, 1988; Lan and Grier, 2004). In the case of T. castaneum, based on the topical application studies, the critical period of JH sensitivity appeared to occur between 72- and 96-h AEFL. Treatment of hydroprene until 60-h AEFL resulted in supernumerary larval molt. Application of hydroprene at 72-h AEFL resulted in a block in pupation and application of hydroprene at 84- and 96-h AEFL led to larval-pupal intermediaries and malformed pupa. There are small peaks of ecdysteroids in the final instar larvae of T. castaneum around 60- and 84-h AEFL (Parthasarathy et al., 2008a). Hence, it is likely that pupal commitment in T. castaneum occurs between 72- and 96-h AEFL.</p><p>To identify the molecular mechanisms that underlie the larval-pupal metamorphosis, we compared gene expression in hydroprene-treated larvae that undergo supernumerary larval molt and control larva that metamorphose into pupa. Interestingly, the pupal-specific gene, Broad, was down-regulated in the hydroprene-treated larvae undergoing larval-larval molt. Broad mRNA transcripts appeared in the epidermis of M. sexta and Bombyx mori larvae only during the later stages of the final instar larvae coinciding with the pupal stage (Zhou et al., 1998a; Ijiro et al., 2004). In this study, FTZ-F1 was up-regulated significantly in the larval-larval molt in the hydroprene-treated larvae when compared to the larval-pupal molt in the control larvae. Woodard et al. (1994) showed that FTZ-F1 repressed its own expression and expressed only for a brief period in mid pre-pupae of D. melanogaster. The expression of FTZ-F1 during the larval-larval molt has not been addressed so far. Further studies are needed to identify the specific role of FTZ-F1. In this study, E74, but not E75 isoforms, expression was up-regulated during the larval-larval molt in hydroprene-treated larvae. The expression of E74, E75A, and E75B was observed during both the larval-larval and larval-pupal molts in M. sexta (Zhou et al., 1998b; Stilwell et al., 2003). Beckstead et al. (2007) showed that E74B was induced by JH III while E74A and E75A showed no response to JH in larval organ cultures of Drosophila. In vitro experiments with wing discs of M. sexta revealed that E75A was not upregulated by pyripoxyfen treatment (Keshan et al., 2006). From these studies, it is clear that expression of isoforms of 20E-induced transcription factors and their response to JH varies between molts and also among insects. Recent studies showed that isoforms of EcR play distinct roles in T. castaneum development (Tan and Palli, 2008b), it is likely that JH is involved in isoform-specific action of EcR and other genes involved in 20E signal transduction.</p><p>Also, the presence of JHA during the final instar larvae blocked midgut remodeling. Midgut remodeling occurs during the larval-pupal metamorphosis wherein the larval midgut cells undergo PCD, move into the gut lumen, and get eliminated from the gut. Simultaneously, the imaginal cells adjoining the basement membrane proliferate and differentiate into the pupal gut epithelium and replace the larval gut. These events occur during the quiescent stage (Parthasarathy and Palli, 2008). In the present study, due to the presence of JHA during the later stages of the final instar larvae, the larval gut epithelium failed to undergo PCD and the proliferation and differentiation of the imaginal cells were impaired. The morphology of the gut epithelium is maintained in the supernumerary larval molt except for an increase in the number of imaginal cells, which appear to proliferate during the final instar larval stage. Similar observations were made when methoprene was administered in the final instar larval stages of mosquitoes (Nishiura et al., 2003; Wu et al., 2006) and Heliothis virescens (Parthasarathy and Palli, 2007). The comparison of gene expression in the midguts of hydroprene-treated and control larvae revealed that the presence of JHA suppressed 20E-induced gene expression as observed in other insects (Nishuira et al., 2005; Wu et al., 2006). Thus, the role of JHA in regulating midgut metamorphosis appears to be conserved in most of the holometabolous insects. Taken together, this study affords an opportunity to understand the molecular mechanisms of hormonal regulation of beetle metamorphosis.</p>

PubMed Author Manuscript

Kinetic and Mechanistic Characterization of the Glyceraldehyde 3-Phosphate Dehydrogenase from Mycobacterium tuberculosis

Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) is a glycolytic protein responsible for the conversion of glyceraldehyde 3-phosphate (G3P), inorganic phosphate and nicotinamide adenine dinucleotide (NAD+) to 1,3-bisphosphoglycerate (1,3-BPG) and the reduced form of nicotinamide adenine dinucleotide (NADH). Here we report the characterization of GAPDH from Mycobacterium tuberculosis (Mtb). This enzyme exhibits a kinetic mechanism in which first NAD+, then G3P bind to the active site resulting in the formation of a covalently bound thiohemiacetal intermediate. After oxidation of the thiohemiacetal and subsequent nucleotide exchange (NADH off, NAD+ on), the binding of inorganic phosphate and phosphorolysis yields the product 1,3-BPG. Mutagenesis and iodoacetamide (IAM) inactivation studies reveal the conserved C158 to be responsible for nucleophilic catalysis and that the conserved H185 to act as a catalytic base. Primary, solvent and multiple kinetic isotope effects revealed that the first half-reaction is rate limiting and utilizes a step-wise mechanism for thiohemiacetal oxidation via a transient alkoxide to promote hydride transfer and thioester formation.

kinetic_and_mechanistic_characterization_of_the_glyceraldehyde_3-phosphate_dehydrogenase_from_mycoba

Introduction<!>Materials<!>Cloning, Expression and Purification of Mtb-GAPDH<!>Construction and Expression of Mtb-GAPDH mutants C158A, C162A and H185A<!>Protein Concentration<!>Measurement of Enzymatic Activity<!>Synthesis of D-glyceraldehyde 3-phosphate and [1-2H]D-glyceraldehyde 3-phosphate<!>Initial Velocity Studies<!>pH Dependence Studies<!>Inactivation Studies<!>Kinetic Isotope Effects<!>Cloning, Expression and Purification<!>Measurement of Enzyme Activity<!>Kinetic Mechanism<!>pH Dependence Studies<!>Mutagenesis and Inactivation Studies<!>Kinetic Isotope Effects<!>Chemical Mechanism

<p>The etiological agent of tuberculosis, Mycobacterium tuberculosis (Mtb), has infected nearly one-third of the human population (1). Approximately 10% of TB-infections lead to an active, symptomatic infection that resulted in nearly 1.4 million deaths in 2011 (1). In addition, multi-drug resistant strains have been reported in every country surveyed by the World Health Organization (1). Yet some of the most basic metabolic enzymes of this bacterium have yet to be characterized.</p><p>Glyceraldehyde 3-phosphate dehydrogenase is a highly conserved enzyme that is utilized in central carbon metabolism by some of the most ancient forms of life (2). GAPDH is best known for its role in glycolysis, catalyzing the reversible conversion of glyceraldehyde 3-phosphate (G3P), inorganic phosphate and NAD+ to 1,3-bisphosphoglycerate (1,3-BPG) and NADH (3). This dehydrogenase is also unusual in that it utilizes a covalent thiohemiacetal intermediate to promote hydride transfer and catalysis (3). The reaction of GAPDH is essential for the regeneration of the two molecules of ATP used to phosphorylate the hexose carbon source, glucose. The cleavage of fructose-1,6-bisphosphate yields the two triose phosphates that are interconverted into G3P. The oxidation of the aldehyde and substrate-level phosphorylation catalyzed by GAPDH generate NADH and the high energy carboxy-phosphoric anhydride containing 1,3-bisphosphoglycerate (1,3-BPG) that is used in the subsequent reaction catalyzed by 3-phosphoglycerate kinase to regenerate the two molecules of ATP used earlier in the glycolytic sequence. The very reactive nature of the product of GAPDH, 1,3-BPG, has recently been shown to be capable of non-enzymatic modification of proteins, including GAPDH (4).</p><p>Recent studies have also found GAPDH to be involved in a variety of cellular processes in addition to its major role in glycolysis. GAPDH has been shown to play a role in transcription, assisting in the formation of both DNA and RNA binding complexes as well as acting as a transcription factor co-activator (5–7). Additionally, GAPDH has been identified as a microtubule-binding protein, a lactoferrin receptor, and as an apoptosis-inducer (8–11). More information on the extra-glycolytic roles of GAPDH can be found in the review by Nichollis et. al. (12).</p><p>Despite decades of work on GAPDH's from prokaryotic and eukaryotic sources, no work has been conducted on the GAPDH from Mycobacterium tuberculosis. It was discovered early on in our work that this enzyme had significant solubility issues. This obstacle was overcome by co-transforming the Mtb-GAPDH plasmid along with a plasmid expressing the chaperones GroEL/GroES (13). These chaperones are believed to create an environment suitable for proper folding yielding soluble and active Mtb-GAPDH (13). In this study, we report the first successful purification and mechanistic evaluation of Mtb-GAPDH using steady-state kinetics, pH-rate profiles, isotope effects and mutagenesis to elucidate both the kinetic and chemical mechanism of Mtb-GAPDH.</p><!><p>All chemicals were purchased from Sigma-Aldrich unless otherwise noted. The Mtb-GAPDH gene was cloned into the Novagen pET-28a(+) vector. The GroEL/GroES plasmid was a gift from the Shrader lab (13). Primers were purchased from Invitrogen. All cloning enzymes and T7 competent Escherichia coli were purchased from New England Biolabs. Complete EDTA-free protease inhibitor cocktail and DNase were purchased from Roche. 99.9% deuterated water was purchased from Cambridge Isotope Laboratories.</p><!><p>The M. tuberculosis gap gene (Rv1436) was PCR amplified from the Erdman strain with a forward primer 5′-GGAATTCCATATGGTGACGGTCCGAGTAGGC-3′ and a reverse primer 5′-GTCGGCAAGTCGCTCTAGAAGCTTGGG-3′. NdeI and HindIII restriction sites were used for forward and reverse primers, respectively. The PCR fragment was ligated into the pET28a(+) vector encoding for a N-terminal His6-tag. The plasmid was then sequenced and confirmed. The Mtb-GAPDH-containing plasmid was co-transformed along with the GroEL/GroES plasmid into T7 Express E. coli competent cells. Kanamycin (35 μg/mL) and tetracyclin (6 μg/mL) were used for selection. Cultures were grown in LB broth at 30°C and induced with 500 μM IPTG at an A600 of ~0.6–0.8 and then grown overnight at 18°C. Cells were harvested by centrifugation and stored at −20°C. The pellets were resuspended in 25 mM HEPES (pH 7.5) containing 300 mM NaCl, 10 mM imidazole, and 1 mM NAD+. Cells were lysed using an EmulsiFlex-C3 and centrifuged to remove cellular debris. The clear supernatant was then added to a Ni2+-NTA agarose column and eluted with a linear imidazole gradient (10 mM–250 mM). Fractions containing Mtb-GAPDH were pooled and dialyzed into 25 mM HEPES (pH 7.5) containing 300 mM NaCl, 1 mM NAD+ and 5% glycerol then concentrated and stored at −20°C in 12.5% glycerol.</p><!><p>Mtb-GAPDH/pET28a was used as a template to generate C158A, C162A and H185A mutants. The mutants were constructed by overlap mutagenesis (14). The mutation has been underlined in the following sequences. The forward primer for C158A was 5′-CTCCAATGCGTCGGCCACCACGAACTGCC-3′ and the reverse primer was 5′-GGCAGTTCGTGGTGGCCGACGCATTGGAG-3′. The forward primer for C162A was 5′-GTGCACCACGAACGCCCTTGCGCCGCTGG-3′ and the reverse primer was 5′-CCAGCGGCGCAAGGGCGTTCGTGGTGCAC-3′. The forward primer for H185A was 5′-GATGACCACCATCGCCGCCTACACTCAGG-3′ and the reverse primer was 5′-CCTGAGTGTAGGCGGCGATGGTGGTCATC-3′. DNA sequencing was used to confirm the described mutations. Each mutant was expressed and purified in the same manner as wild-type Mtb-GAPDH.</p><!><p>Protein concentration was determined using Bio-Rad Bradford protein assay using bovine serum albumin as a standard.</p><!><p>Enzymatic activity was measured spectrophotometrically by monitoring the conversion of NAD+ to NADH at 340 nm. Reactions were conducted in 50 mM HEPES, pH 8 including substrates NAD+, Na2AsO4, and DL-G3P in a total volume of 500 μL. Reactions were initiated by the addition of Mtb-GAPDH. Initial velocities were measured before 10% conversion of substrate to product had occurred and initial rates were calculated using the molar extinction coefficient of NADH (ε340 = 6,220 M−1 cm−1) and the total enzyme concentration.</p><p>The measurements of kinetic isotope effects were conducted in the same manner with a few differences. We synthesized and utilized the individual stereoisomers, [1-1H]D-glyceraldehyde 3-phosphate (1-[1H]D-G3P) and [1-2H]D-glyceraldehyde 3-phosphate (1-[2H]D-G3P) for these studies. Concentrations of 1-[1H]D-G3P and 1-[2H]D-G3P were determined enzymatically using excess NAD+ and arsenate.</p><!><p>Briefly, the known compound, methyl (R)-2-benzyloxy-3-trityl-oxypropanoate (1), was synthesized from commercially available methyl 2,3-O-isopropylidene-D-glycerate according to previously published procedures (15). The 3-O-trityl ether was removed to generate compound 2 bearing a free hydroxyl group at the C3 position (Scheme 1). This hydroxyl group was phosphorylated using a phosphoramidite coupling reagent to generate the common intermediate compound 3 that was used for the synthesis of both D-glyceraldehyde 3-phosphate and [1-2H]D-glyceraldehyde 3-phosphate. To generate the aldehyde at the C1 position, compound 3 was treated with DIBAL-H or DIBAL-D to generate compounds 4 and 5, respectively. D-glyceraldehyde 3-phosphate (6) and [1-2H]D-glyceraldehyde 3-phosphate (7) were then generated through hydrogenolysis (1 atm H2, Pd/C catalyst) of the benzyl-protected compounds. Further experimental detail and NMR spectra can be found in the Supplementary Information.</p><!><p>Data were fit using SigmaPlot 11.0. All points are the mean of experimental duplicates and the error bars are plus/minus one half of the difference between the experimental duplicates. Substrate inhibition curves were determined and fitted to eq 1 for linear substrate inhibition</p><p> [1]v=VS/(Km+S+S2/Ki) where v is the velocity, S is the concentration of substrate and Ki is the substrate inhibition constant. Initial kinetic parameters were estimated by saturating two substrates while varying the concentration of the third. These rates were fit to eq 2</p><p> [2]v=(VS)/(Km+S) where V is the maximal velocity and S is the concentration of the varied substrate. Initial velocity studies were conducted by saturating with one substrate, fixing the concentration of another, and varying the concentration of the third. The results were fitted for intersecting and parallel patterns using eq 3 and eq 4, respectively</p><p> [3]v=(VAB)/(KiaKb+KaB+KbA+AB) [4]v=(VAB)/(KaB+KbA+AB) where V is the maximal velocity, A and B are substrate concentrations, Ka and Kb are the respective Michaelis constants for each of the substrates, and Kia is the inhibition constant for substrate A. Product inhibition was performed using reduced nicotinamide adenine dinucleotide (NADH) as an inhibitor. Data was obtained at variable concentrations of NAD+, G3P or arsenate at several fixed concentrations of NADH. The data were fit to competitive, uncompetitive, and noncompetitive inhibition using eq 5, 6, and 7, respectively</p><p> [5]v=(VS)/[Km(1+I/Kis)+S] [6]v=(VS)/[Km+S(1+I/Kii)] [7]v=(VS)/[Km(1+I/Kis)+S(1+I/Kii)] where S is the varied substrate concentration, I is the inhibitor concentration, Kis is the inhibition constant for the slope, Kii is the inhibition constant for the intercept, and Km is the Michaelis constant for the substrate S.</p><!><p>The pH dependence of the kinetic parameters was determined in 50 mM concentrations of buffers at various pH values and using the following structurally related buffers at the pH values noted in parentheses: MES (pH 5–6.5), HEPES (pH 6.5–8), TAPS (pH 8–10). The pH profile was determined varying either G3P or Na2AsO4 with other substrates held saturating. The data were fit to eq 8</p><p> [8]logk=log[c/(1+Kb/10-pH+10-pH/Ka)] where c is the pH-independent plateau value and Ka and Kb are dissociation constants of the ionizing groups.</p><!><p>Inactivation studies were performed using iodoacetamide (IAM) as an inactivator. Enzyme activities were measured for native Mtb-GAPDH every 15-minutes and every 20-minutes for Mtb-GAPDH C162A at varying concentrations of IAM. Kinetic constants were calculated as described by Kitz and Wilson (16). Briefly, relative inhibition was determined by comparing the rates of time-matched samples exposed to varying concentrations of IAM to control samples lacking IAM. The results were plotted on a logarithmic scale versus time and the slopes were determined by linear regression. The values of the slopes represent the inactivation rate constants and were used to calculate the half-life of the inactivation (kinactivation = 0.693 / t1/2). A Kitz-Wilson plot was produced using the calculated half-lives as the y-axis and the reciprocal of the inhibitor concentration as the x-axis (16). Values for Ki and kinactivation were determined from the negative reciprocal of the x-intercept and reciprocal of the y-intercept, respectively.</p><!><p>Solvent kinetic isotope effects were determined at varying concentrations of either D-G3P or Na2AsO4 while the other substrates were kept at saturating concentrations in H2O or 90% D2O. A viscosity control of 9% glycerol was performed and revealed no effect on V or V/Km (17). The data were fit to eq 9</p><p> [9]v=(VS)/[KA(1+Fi(EV/K-1))+S(1+Fi(EV-1))] where V is the maximal velocity, S is the substrate concentration, EV/K is the effect on kcat/Km, EV is the effect of kcat and Fi is the fraction of the isotope (Fi = 0 for H2O and Fi = 0.9 for D2O). A proton inventory was then conducted using saturating conditions of substrates while varying D2O concentrations from 0 to 90%. The results were fit to a linear equation.</p><p>Primary and multiple kinetic isotope effects were determined at varying concentrations of either [1-1H]D-G3P or [1-2H]D-G3P while other substrates were kept at saturating conditions in either H2O or 90% D2O. The results were fitted to eq 9.</p><!><p>Mtb-GAPDH was PCR amplified then cloned into the pET28a expression vector encoding a N-terminal His6-tagged Mtb-GAPDH. Sequencing confirmed that no mutations were introduced during the cloning process. Initial expression studies revealed that Mtb-GAPDH was insoluble under normal conditions. Co-transformation of the Mtb-GAPDH plasmid along with a GroEL/GroES plasmid yielded soluble and active protein. After purification and dialysis, the final protein preparation was >95% pure as determined by SDS-PAGE. The addition of 1mM NAD+ to the sonication and purification buffers was essential to maintain the activity of GAPDH (data not shown).</p><!><p>The activity of Mtb-GAPDH was determined using a direct spectrophotometric assay measuring the conversion of NAD+ into NADH at 340 nm. NADP+ was screened as an alternate substrate for NAD+ but no activity could be demonstrated. Na2AsO4 was substituted for the normal Mtb-GAPDH endogenous substrate, Na2PO4, to prevent the reverse reaction. This substitution yields the product, 1-arseno-3-phosphoglycerate, which is rapidly hydrolyzed. Na2AsO4 exhibited kinetic constants (Km = 6.2 ± 0.6 mM, V = 1590 ± 40 min.−1) similar to Na2PO4 (Km = 6 ± 1 mM, V = 1450 ± 70 min.−1) (Fig. 1A&B). These studies also revealed that both Na2PO4 and Na2AsO4 cause linear substrate inhibition at concentrations higher than 30 mM (Fig. 1A&B, insets), presumably due to competition with the phosphate binding site for G3P. Fits of these data to eq 1 yielded Ki values of 60 ± 10 mM and 93 ± 7 mM for Na2PO4 and Na2AsO4, respectively.</p><!><p>Initial velocity studies were conducted to elucidate the kinetic mechanism of Mtb-GAPDH. Commercially available DL-G3P was used as a substrate for these studies due to the laborious procedures required to synthesize the single D-G3P stereoisiomer. Consequently, concentration values reported herein are for the DL-G3P mixture and presumed to be two times the value for D-G3P alone. Double-reciprocal plots yielded a series of parallel lines when varying G3P concentration at several fixed NAD+ concentrations and an optimal, but not saturating concentration of Na2AsO4 (Fig. 2A). Because the concentration of Na2AsO4 was chosen to avoid substrate inhibition, reported values of V may slightly underestimate the actual value. Fitting these data to eq 4 yielded V = 1670 ± 90 min.−1, KG3P = 280 ± 30 μM and KNAD+ = 40 ± 4 μM. Varying Na2AsO4 concentrations at several fixed NAD+ concentrations resulted in a series of intersecting lines that when fit to eq 3 yielded KNa2AsO4 = 3.9 ± 1.5 mM, and KNAD+ = 70 ± 20 μM (Fig. 2B). Varying G3P concentrations at several fixed concentrations of Na2AsO4 resulted in a series of parallel lines that when fit to eq 4 yielded V = 1710 ± 60 min.−1 and KNa2AsO4 = 5.1 ± 0.3 mM (Fig. 2C). To elucidate the binding order of substrates and product release, product inhibition studies were performed. NADH was determined to exhibit competitive inhibition versus both NAD+ and Na2AsO4 (Fig. 3A&B). Fits of these data to eq 5 yielded Ki values of 14 ± 1μM and 37 ± 1 μM versus NAD+ and Na2AsO4, respectively. NADH exhibited noncompetitive inhibition versus G3P and when fit to eq 7 yielded Kii = 24 ± 5 and Kis = 31 ± 7 μM (Fig. 3C).</p><p>Together, these data are consistent with an unusual kinetic mechanism in which the free enzyme is in reality the E-NAD+ complex. This is supported by both the parallel line initial velocity pattern observed when G3P and NAD+ are varied (requiring a product to be released between them) and the requirement of NAD+ for stability. G3P first binds to the E-NAD+ complex, and reacts with the active site C158 to generate the thiohemiacetal intermediate (E-NAD+-X in Scheme 2). Hydride transfer to NAD+ generates NADH and the enzyme thioester intermediate (E-Y-NADH in Scheme 2), which undergoes a "nucleotide exchange" reaction to generate the thioester-NAD+ complex (E-Y-NAD+). Binding of phosphate and phosphorolysis of the thioester generates the product, 1,3-bisphosphoglycerate, leaving the enzyme with NAD+ bound and ready to react with another G3P molecule. A similar "nucleotide exchange" kinetic scheme was reported for GAPDH from both rabbit and pig muscle (3, 18, 19).</p><!><p>To determine which enzyme residues are required for catalytic function, pH dependence studies were performed. Control experiments were conducted to determine the stability of Mtb-GAPDH. These studies limited our profile to the pH range of 5.5 – 8.5. Experiments varying G3P concentrations at saturating concentrations of NAD+ were used to evaluate the effect of ionizations on log kcat and log kcat /KG3P. This experiment yielded no detectable change in either kinetic parameter over this pH range (Fig. 4A), suggesting that the ionization of the active site cysteine and histidine residues (20) implicated in the first chemical reaction (thiohemiacetal formation and hydride transfer to generate the thioester) were outside of the experimental range. Unfortunately, due to enzyme instability outside of this pH range we were unable to probe the enzymatic reaction at pH's that were closer to the pK1 and pK2 values previously reported by Polgar for NAD+-bound GAPDH (5.2 and 8.9, respectively) (20). Similar experiments varying Na2AsO4 to probe for ionizations implicated in the phosphorolysis reaction also yielded no evidence for enzyme side chain participation (Fig. 4B).</p><p>Studies with GAPDH homologs have demonstrated the essential role of an active site cysteine responsible for the formation of the thioacyl intermediates (3, 18, 21). The active site cysteine has been reported to form a thiolate-imidazolium ion-pair with an active site histidine that increases the reactivity of the cysteine (3, 20, 22–24). These residues have been identified through sequence alignment to correspond to C158 and H185 in Mtb-GAPDH. The lack of any pH dependence of the kinetic parameters in the pH range tested for Mtb-GAPDH is likely due to the unusual kinetic mechanism and the requirement of bound NAD+ (or NADH) for stability and throughout the catalytic cycle. Crystal structures have shown the binding of NAD+ to induce a conformational change in the GAPDH of Bacillus stearothermophilus leading to a repositioning of active site residues favoring the stabilization of a thiolate-imidazolium pair (25). Polgár has theorized that the formation of the thiolate-imidazolium pair in GAPDH would lead to altered pK values that would appear outside the pH range tested (20). Given our inability to assess the required appropriate ionization state of residues implicated in catalysis by variations in external pH, mutagenesis studies were conducted to probe the roles of C158, C162 and H185.</p><!><p>In order to verify the essential role of C158 in catalysis, kinetic inactivation studies were performed. Iodoacetamide (IAM), an irreversible, cysteine-specific alkylator was used to probe the role of the active site cysteines C158 and C162A in catalysis. Fig. 5A illustrates the inhibition of Mtb-GAPDH with IAM in both a time- and concentration-dependent manner. A Kitz-Wilson re-plot gave a kinactivation and Ki of IAM of 0.16 min.−1 and 7.8 mM, respectively (Fig. 5B). This data supports the proposed function of an active site cysteine in catalysis, as previously reported in other GAPDH homologs (3, 19, 21).</p><p>The C158A, C162A and H185A mutant forms of Mtb-GAPDH were constructed and purified in order to determine their role in catalysis. The C158A mutant was inactive, but the C162A mutant exhibited a comparable V to native Mtb-GAPDH and only a 2-fold increase in the value of KG3P (data not shown). Time- and concentration-dependent inactivation studies with IAM were conducted with C162A to verify that acetamidation of C158, and not C162 was responsible for the loss of activity (Fig. 5C). A Kitz-Wilson replot gives a kinactivation and Ki of IAM of 0.12 min.−1 and 14 mM, respectively, comparable to the wild-type enzyme (Fig. 5D). The H185A mutant was also inactive suggesting that C158 is the active site nucleophile reacting with the aldehyde group of G3P to generate the thiohemiacetal and that H185 is additionally required to either stabilize thiolate anion formation or act as a catalytic acid/base group.</p><!><p>Studies of kinetic isotope effects were performed to identify the rate-limiting chemical steps. In order to measure the primary kinetic isotope effect on the Mtb-GAPDH catalyzed reaction, the synthesis of a single stereoisomer of glyceraldehyde 3-phosphate ([1-1H]D-G3P) and [1-2H]D-glyceraldehyde 3-phosphate ([1-2H]D-G3P) was performed. Using these compounds, only modest isotope effects were observed on V (DVG3P = 1.2 ± 0.1) and V/K (DV/KG3P = 1.5 ± 0.2, Fig. 6). The expression for DV/K includes rate constants from the binding of G3P to the E-NAD+ complex to the first irreversible step, which we suggest is NADH release (Scheme 2). The value of DV/KG3P is significantly smaller than one would expect for a hydride transfer reaction, and indicates that G3P is sticky. Certainly the formation of a covalent thiohemiacetal would contribute to a high commitment factor for this substrate. This chemistry also results in there being a small (ca. 1.2) inverse equilibrium effect on the formation of the thiohemiacetal when using [1-2H]D-G3P versus [1-1H]D-G3P due to the change in hybridization from sp2 to sp3. Together these data suggest that the hydride transfer is only partially rate-limiting in the chemical oxidation of the covalently-bound thiohemiacetal. The even lower value of DV may reflect additional rate limitation by product release. These results are quantitatively similar to those obtained under similar experimental conditions with rabbit muscle GAPDH (18).</p><p>In order to probe additional chemical steps in the oxidation reaction, solvent kinetic isotope effects (SKIE) were used to investigate the potential rate-limiting nature of proton transfer steps. A proton inventory was conducted under "pseudo-V" conditions (5–10 times Km values for G3P and NAD+ and maximal, but not substrate inhibitory concentrations of arsenate) in 10% increments of D2O from 0–90%. This experiment yielded a linear, normal dependence of the rate on solvent isotopic composition, indicative of a single solvent-derived proton being involved in the oxidative half-reaction (Fig. 7A). The extrapolated magnitude of "D2OV" was approximately 1.7.</p><p>Experiments performed with 1-[1H]D-G3P in H2O and 90% D2O at pH 8 yielded values of D2OV of 1.7 ± 0.1 and D2OV/KG3P of 2.5 ± 0.4, respectively, (Fig. 7B). As is the case for the primary KIE, D2OV/KG3P includes rate constants from the binding of G3P to the release of NADH. In the steps that represent the oxidation of the thiohemiacetal (Scheme 3), the formation of the thiohemiacetal is accompanied by protonation of the carbonyl oxygen. Oxidation and hydride transfer will occur only after deprotonation of the alcohol, and it is this step that is the most likely source of the single-proton solvent KIE. The magnitude of V/KG3P suggests that this step is slow, and that the relatively rate-limiting formation of the alkoxide by His185-assisted deprotonation promotes a more rapid hydride transfer to NAD+ from the thiohemiacetal alkoxide. This interpretation is further supported by the inactivity of H185A, as has been previously suggested (24). The smaller magnitude of D2OV suggests that steps after the oxidative half-reaction contribute to overall rate limitation, as was discussed for the primary KIE.</p><p>Oxidation of secondary alcohols can presumably be carried out in either a concerted (deprotonation-hydride transfer) or stepwise manner. Theory has been developed that allows the discrimination of these two mechanisms using multiple kinetic isotope effects (26). We elected to perform multiple kinetic isotope effect (MKIE) experiments examining the effect of substrate deuteration on the solvent KIE, since the solvent KIE's are larger and more likely to reveal statistically significant differences, if present. Using 1-[2H]D-G3P in H2O and 90% D2O at pH 8, the solvent kinetic isotope effects on D2OV1-[2H]D-G3P was 1.6 ± 0.1 and D2OV/K1-[2H]D-G3P was 1.4 ± 0.2 (Fig. 7C). D2OV using either 1-[1H]D-G3P or 1-[2H]D-G3P were equivalent (1.6–1.7) and equal to the extrapolated value from the proton inventory experiment. However, D2OV/K1-[2H]D-G3P was significantly smaller than when 1-[1H]D-G3P was used as substrate (1.4 vs. 2.5). This suggests that the two isotopic perturbants affect different steps (alkoxide formation or hydride transfer) and that as the hydride transfer transition state is raised by substrate deuteration that the proton transfer step becomes less rate-limiting and the magnitude of D2OV/KG3P is decreased. This argues persuasively that the primary and solvent kinetic isotope effects are reporting on separate steps, e.g., hydride transfer occurring after alkoxide formation. These data allow us to propose a detailed chemical mechanism for the reaction catalyzed by Mtb-GAPDH.</p><!><p>The chemical mechanism for Mtb-GAPDH shown in Scheme 3 is supported by our determination of the kinetic mechanism, mutagenesis and isotope effect studies. The "free enzyme" is actually present with bound NAD+, and the two important catalytic residues, Cys158 and His185, are present as a thiolate-imidazolium ion pair. Binding of glyceraldehyde-3-phosphate yields an initial Michaelis complex, and the Cys158 thiolate nucleophilicially attacks the aldehyde to generate the neutral thiohemiacetal after proton donation from His185. Solvent kinetic isotope effects support a mechanism for oxidation that requires an initial deprotonation of the alcohol to generate the alkoxide that is the largely rate-limiting step. Alkoxide formation promotes hydride transfer from C1 of the covalently-bound G3P thiohemiacetal to NAD+. This generates the covalently-bound thioester and NADH. Nucleotide exchange and inorganic phosphate binding set up the second half-reaction, the phosphorolysis of the thioester to generate the high energy mixed carboxy-phosphoric anhydride of the product 1,3-bisphosphoglycerate (1,3-BPG), which is released in the final step of the catalytic cycle. The thioester and carboxy-phosphoric anhydride are of roughly equal energy, and permit the substrate level phosphorylation reaction catalyzed by GAPDH to drive the subsequent step in glycolysis catalyzed by phosphoglycerate kinase: the formation of ATP and phosphoglycerate from ADP and 1,3-bisphosphoglycerate.</p>

PubMed Author Manuscript

Polyadenylation and nuclear export of mRNAs

In eukaryotes, the separation of translation from transcription by the nuclear envelope enables mRNA modifications such as capping, splicing, and polyadenylation. These modifications are mediated by a spectrum of ribonuclear proteins that associate with preRNA transcripts, coordinating the different steps and coupling them to nuclear export, ensuring that only mature transcripts reach the cytoplasmic translation machinery. Although the components of this machinery have been identified and considerable functional insight has been achieved, a number of questions remain outstanding about mRNA nuclear export and how it is integrated into the nuclear phase of the gene expression pathway. Nuclear export factors mediate mRNA transit through nuclear pores to the cytoplasm, after which these factors are removed from the mRNA, preventing transcripts from returning to the nucleus. However, as outlined in this review, several aspects of the mechanism by which transport factor binding and release are mediated remain unclear, as are the roles of accessory nuclear components in these processes. Moreover, the mechanisms by which completion of mRNA splicing and polyadenylation are recognized, together with how they are coordinated with nuclear export, also remain only partially characterized. One attractive hypothesis is that dissociating poly(A) polymerase from the cleavage and polyadenylation machinery could signal completion of mRNA maturation and thereby provide a mechanism for initiating nuclear export. The impressive array of genetic, molecular, cellular, and structural data that has been generated about these systems now provides many of the tools needed to define the precise mechanisms involved in these processes and how they are integrated.

polyadenylation_and_nuclear_export_of_mrnas

Introduction<!>Nuclear pore structure and function<!>Overview of nuclear transport<!><!>Nuclear export of mRNA<!><!>Nuclear export of mRNA<!><!>Nuclear export of mRNA<!>Checkpoints<!><!>Splicing<!>Polyadenylation<!><!>Polyadenylation<!>Questions outstanding<!>

<p>In eukaryotes, the nuclear envelope functions to separate transcription from translation, and this enables transcripts to be modified substantially in the nucleus before they are exported to the cytoplasm where they are translated. During these nuclear steps in the gene expression pathway, pre-RNA transcripts are modified by the addition of 5′ caps and 3′ poly(A) tails, and frequently introns are removed to generate mature mRNAs. A broad spectrum of ribonuclear proteins are bound to transcripts during the nuclear processing (for reviews, see Refs. 1 and 2) that ultimately generates mature export-competent messenger ribonuclear particles (mRNPs)2 that are exported to the cytoplasm through the nuclear pores that perforate the nuclear envelope and mediate the movement of macromolecules between the cytoplasmic and nuclear compartments. The nuclear processing steps in the gene expression pathway are crucial to ensuring that the coding sequence of the gene is delivered to ribosomes to generate the appropriate proteins, as well as regulating mRNA stability and mediating translation. Errors or defects in the nuclear transcript modification steps or in mRNA nuclear export frequently result in impaired grow or death of cells and are associated with a broad range of different disease conditions. Consequently, these processes are closely coordinated and are coupled to the nuclear export of mRNA to ensure that only mature and completely processed transcripts reach cytoplasmic ribosomes for translation into proteins.</p><p>The transport of macromolecules, such as proteins and RNAs, between the cytoplasmic and nuclear compartments is mediated by nuclear pores that are huge macromolecular assemblies that span the nuclear envelope. Nuclear transport is an active, energy-requiring process and is highly selective. Because the pores function as a barrier to the passive diffusion of molecules larger than ∼40 kDa, only macromolecules that are bound to specific carrier molecules ("transport factors") are able to move through them.</p><!><p>Considerable progress has been made on establishing the structure of nuclear pores and relating this to their function. Nuclear pores are constructed from multiple copies of upward of 30 different proteins that are arranged to generate a cylindrical structure that has 8-fold rotational symmetry and that has a central transport channel through which macromolecules move (for reviews, see Refs. 3–5). The individual proteins from which nuclear pores are constructed are collectively referred to nucleoporins, or "nups" for short, and are frequently distinguish by a number that reflects their Mr (such as, for example, Nup42 or Nup153). Although there is considerable sequence variation between species, the overall architecture of nuclear pores is generally conserved; albeit budding yeast nuclear pores are somewhat smaller than their metazoan counterparts (3–5). Crystal structures of many individual components and the subcomplexes they form have been obtained and, together with recent cryo-EM structures of intact pores, have provided a detailed model of how they are arranged to generate the cylindrical skeleton of the pore (for reviews, see Refs. 3–5). There are also fibrous assemblies protruding from both faces of the pore. In the nucleus, these fibers form a nuclear basket, whereas at the cytoplasmic face they form a series of filaments that extend into the cytoplasm.</p><p>The architecture of nuclear pores has provided a basis for understanding the mechanism by which nuclear pores generate a barrier to the passive diffusion of large macromolecules while facilitating the selective active transport of specific cargoes. In addition to the structured protein domains that constitute the nuclear pore skeleton, many nucleoporins also contain natively unfolded regions that lack defined secondary structure. These regions fill the central channel through which macromolecules are transported and generate a barrier that impairs movement through the pores of macromolecules larger than ∼40 kDa. These natively unfolded regions generally contain tandem sequence repeats that contain hydrophobic cores rich in phenylalanine and glycine, with sequences such as GLFG and FXFG, separated by linkers that are generally hydrophilic, and are referred to as FG-nucleoporins ("FG-nups"). The transport barrier generated by these densely packed regions of FG-nups in the transport channel can be overcome by a range of nuclear transport factors ("carriers") that mediate the import and export of specific macromolecular cargoes (6, 7). There is a range of different models for how the FG-nups generate barrier function, including entropic effects due to molecular crowding related to the formation of molecular brushes (8–10) and cohesion between nucleoporins (11). These mechanisms are not mutually exclusive, and all may contribute to impeding the movement of macromolecules through the pore channel. Transport factors are able to overcome the barrier function by transiently interacting with the hydrophobic cores of the FG-nups and so enabling cargo:carrier complexes to passively diffuse rapidly back and forth through the pores. Although this interaction enables movement back and forth through the pores, energy is required to impose directionality and generate vectorial transport. Therefore, nuclear transport is different from simple passive diffusion through a pore in which movement is dictated simply by the chemical potential. Instead, metabolic energy is used to rectify simple diffusion so that the direction of transport does not depend on the difference in chemical potential of the cargo in the donor and acceptor compartments. In this way, nuclear transport facilitates the generation of the distinctive compositions of the nuclear and cytoplasmic compartments.</p><!><p>Although nuclear transport is an active energy-requiring process, the movement of cargo:carrier complexes through the nuclear pore transport channel itself is mediated by simple passive diffusion-based Brownian motion. The energy that enables transport independent of the chemical potential of the cargoes in each compartment is provided indirectly through the assembly and disassembly of the cargo:carrier complexes, which for the transport of protein cargoes is mediated by the Ras-family GTPase, Ran (Gsp1 in budding yeast). For example, in nuclear protein import (for reviews, see Refs. 6 and 7), karyopherin-β–family carriers such as importin-β (Kap95 in yeast) bind cargo proteins in the cytoplasm, and after passage through the pores to the nucleus, the cargo:carrier complex is dissociated by RanGTP binding to the karyopherin carrier, releasing the cargo (Fig. 1A). The carrier:RanGTP complex then returns to the cytoplasm where the Ran GTPase-activating protein (RanGAP) catalyzes GTP hydrolysis, generating RanGDP, which dissociates from the karyopherin, freeing it for a further import cycle. The RanGDP then returns to the nucleus (using NTF2 as a carrier (12)) where the Ran guanine nucleotide exchange factor (RanGEF), RCC1 (Prp20 in yeast), recharges it with GTP. Analogous machinery is used for the nuclear export of proteins and small RNAs, although here cargo:carrier assembly in the nucleus requires β-karyopherins, such as CRM1 (Xpo1), to be bound to RanGTP, with the cargo being released following GTP hydrolysis in the cytoplasm (Fig. 1B). The net result of these cycles is that the energy derived from GTP hydrolysis is used to impose directionality on transport by rectifying the thermal diffusion (Brownian motion) of the components of the transport machinery by facilitating assembly of the cargo:carrier complex in the donor compartment and its disassembly in the target compartment and so is an example of a Brownian ratchet mechanism (13). This mechanism enables transport to be independent of the concentration or chemical potential of cargoes in each compartment and so facilitates the different macromolecular compositions of the nucleus and cytoplasm. Although they do not contribute directly to the energetics of the transport cycle and the relative concentrations of cargoes in each compartment, several nucleoporins do contribute to the kinetics of transport by concentrating material at one face of the nuclear pores. For example, Nup1 binding helps concentrate Kap95 complexes and RanGTP at the nuclear face of the pore (14), and Nup358 appears to perform an analogous function at the cytoplasmic face through binding RanGAP (15).</p><!><p>Overview of nuclear transport pathways for macromolecules. These pathways are all based on a thermal ratchet mechanism in which energy is used to rectify Brownian motion by mediating the assembly of cargo:carrier complexes in the donor compartment and their disassembly in the acceptor compartment. Proteins and small RNAs are transported using karyopherin-based carriers to mediate movement through nuclear pores. In nuclear protein import (A) karyopherins (Kap; yellow) such as importin-β bind their cargoes (green) in the cytoplasm (often employing an adapter such as importin-α), and then, when the cargo:carrier complex reaches the nucleus, RanGTP (red) binding to the karyopherin dissociates the cargo, after which the karyopherin:RanGTP complex returns to the cytoplasm where the RanGTPase is activated by RanGAP, generating RanGDP that dissociates from the karyopherin, freeing it for a further import cycle. The RanGDP is then recycled to the nucleus where RanGEF (RCC1, Prp20) recharges it with GTP. The export of proteins and small RNAs (B) is mediated by an analogous pathway, except that here the cargo binds to the karyopherin (such as Crm1) complexed with RanGTP in the nucleus and is released in the cytoplasm following GTP hydrolysis. The export of mRNAs (C) employs a different pathway that uses the Mex67:Mtr2 (NXF1:NXT1) complex (Fig. 3) as a carrier to which binding in the nucleus and release in the cytoplasm are mediated by DEAD-box helicases (Sub2 and Dbp5, respectively) that hydrolyze ATP to remodel the mRNP. An additional feature of mRNA export is that it is necessary for the nuclear steps of the gene expression pathway to have been completed so that only fully matured transcripts are exported for translation in the cytoplasm. The machinery involved in these steps is summarized in Fig. 2.</p><!><p>Although some small mRNAs and virus RNAs are exported using CRM1 (through binding of adapter proteins such as HIV REV), the overall mechanism of nuclear export of most mRNA is conserved within eukaryotes and employs an analogous thermal ratchet mechanism (Fig. 1C), although here the carriers are not members of the karyopherin-β family, and the energy used to rectify movement through the assembly and disassembly of the cargo:carrier complexes is provided by ATP hydrolysis on DEAD-box helicases (16–20). Transcripts are only exported when a complex series of steps in the nuclear segment of the gene expression pathway have been completed (Fig. 2), which entails a considerable level of coordination and relies on a complex signaling network to impair improperly or incompletely processed mRNAs reaching the translation machinery in the cytoplasm.</p><!><p>Outline of the gene expression pathway from transcription to translation for budding yeast. Transcripts are modified by the addition of 5′ caps and 3′ poly(A) tails and the splicing out of introns if present before a structural rearrangement, mediated by Sub and the TREX complex, removes Yra1 and attaches Mex67:Mtr2 to generate an export-competent mRNP to which Nab2 is also bound. The precise nature of the structural rearrangement remains to be established but may involve the generation of RNA hairpins that bind more strongly to the transport factor. The mRNP can then diffuse back and forth through nuclear pores as a result of interactions between Mex67:Mtr2 and FG-nucleoporins overcoming the barrier function of the pore. At the cytoplasmic face of the pore, the DEAD-box helicase Dbp5, working in conjunction with Nup42 and Gle1, remodels the RNA to release Mex67:Mtr2 and Nab2, thereby preventing the mRNA from returning to the nucleus. Pab1 also replaces Nab2 on the poly(A) tail. Although this sequence of processing steps tends to resemble a production line, the steps may not necessarily occur in a defined sequence, and nuclear export, the culmination of the nuclear phase of the gene expression pathway, appears to only require that all steps have been completed successfully. Metazoans have an analogous pathway but differ in some details, primarily in the addition of the EJC immediately after the 3′ splice site when exons are joined. In metazoans, the first EJC together with the 5′ cap facilitates the binding of the TREX complex and subsequent attachment of the Mex67:Mtr2 homologue NXF1:NXT1. IP6, inositol hexaphosphate; Pol II CTD, polymerase II C-terminal domain.</p><!><p>The budding yeast Mex67:Mtr2 complex (Fig. 3A) and the homologous metazoan NXF1:NXT1 (also called TAP:P15) complex function as general mRNP export factors to mediate the movement of mature mRNPs through nuclear pores to the cytoplasm (21, 22). However, Mex67:Mtr2 and NXF1:NXT1 bind RNA nonspecifically, and so additional factors are needed to mediate their attachment to mRNAs. The binding of Mex67:Mtr2 to mature mRNA in the nucleus is mediated by DEAD-box helicases such as Sub2 (UAP56 in metazoans) and possibly Dbp2 (23, 24), whereas disassembly of the complex when it reaches the cytoplasmic face of the pore is mediated by the DEAD-box helicase Dbp5 (vertebrate DDX19). Both the attachment and detachment of Mex67:Mtr2 to mRNAs are probably mediated through generating conformational changes in mRNA structure (17, 18, 20). Both of these remodeling processes also involve a number of accessory proteins, including the TREX complex and Yra1/ALYREF in the nucleus and Gle1 and Nup42 in the cytoplasm (25–38). After the transport factor has been dissociated from the mRNA, it is imported back through nuclear pores to participate in another mRNA export cycle. The actual transport of mRNPs through nuclear pores is rapid compared with the time taken at the nucleoplasmic and cytoplasmic faces to assemble and disassemble export-competent mRNPs and associated quality control processes (39, 40).</p><!><p>Mex67:Mtr2/NXF1:NXT1 structure and interaction with CTE-RNA. A, domain structure of the S. cerevisiae Mex67:Mtr2 complex. The metazoan homologue, NXF1:NXT1 (also called TAP:P15), has a similar structure but has an additional unstructured arginine-rich domain at its N terminus. Mex67 contains four structural modules (RRM domain, LRR domain, NTF2-like domain, and UBA domain) that are connected by flexible linkers. Mtr2 also has an NTF2-like fold and binds to the Mex67 NTF2-like domain to form a heterodimer. The RRM, LRR, and NTF2-like domains bind to RNA, whereas the NTF2-like and UBA domains interact with FG-nucleoporins (based on Protein Data Bank (PDB) codes 1OAI and 4WWU). B, complex formed between viral CTE-RNA and the RRM and LRR domains of TAP (the human homologue of Mex67) showing the secondary structure of the RNA (based on PDB code 3RW6) that forms a bent hairpin (blue). Similar RNA secondary structures may be generated in the nucleus by helicases to facilitate the binding of Mex67:Mtr2 to generate export-competent mRNPs.</p><!><p>Yeast Mex67 is constructed from four domains: an N-terminal RRM domain followed by a LRR domain, an NTF2-like domain that binds Mtr2 (which also has an NTF2 fold), and finally a C-terminal UBA-fold domain (Fig. 3). NXF1 has a similar architecture with an additional N-terminal arginine-rich RNA-binding domain (RBD) (41). The UBA domain and the NTF2 domain heterodimer bind to FG-nups, whereas the RBD, RNP, and LRR domains bind to RNA (20). The DEAD-box helicase Dbp5 removes Mex67:Mtr2 from transcripts at the cytoplasmic face of the pores and probably also removes other factors such as Nab2 (a factor that helps coordinate the nuclear steps as well as regulating poly(A) tail length) from the mRNPs (29, 31, 35, 42). The activity of Dbp5 is enhanced greatly by nuclear pore components Gle1 and Nup42 and by inositol hexaphosphate as well as Gfd1, a protein that also binds to Nab2 and that is thought to accompany it to the cytoplasm (25, 36). DBP5 and GLE1 function to remove NXF1:NXT1 from metazoan transcripts in an analogous manner.</p><p>Yra1 and Sub2, working in conjunction with the TREX complex, are necessary to generate mature, export-competent transcripts in the nucleus of budding yeast (20, 28, 32, 38). Yra1 acts as an adapter, facilitating Mex67 attachment to the transcript, but paradoxically is removed before export (29), probably as a result of the TREX complex (especially the DEAD-box helicase Sub2/UAP56) orchestrating a remodeling of the transcript that likely involves a change in mRNA secondary structure (18, 19, 43), together with Yra1 ubiquitinylation by the Tom1 E3 ligase (44). ALYREF functions as an analogous adapter for TAP:P15 in metazoans. In addition, RNA-binding proteins rich in Ser and Arg (SR proteins), such as Npl3, may also function as adapters for the Mex67:Mtr2 complex in some instances (45).</p><p>The TREX complex, which is conserved between yeast and metazoans, contributes to the integration of the nuclear steps of the gene expression pathway and nuclear export (32). The yeast TREX complex is primarily associated with the transcription machinery and is based on a THO complex core, consisting of Tho2, Hpr1, Mft1, and Thp2, to which Yra1 and Sub2 become attached. In metazoans, however, TREX appears instead to associate primarily with the splicing machinery through its binding to the exon-junction complex (EJC) that is deposited ∼20 nucleotides upstream of the most 5′ exon–exon junction (46). At a later stage, the TREX complex contributes to the generation of export-competent mRNPs; albeit this process appears to be more complex in metazoans and may also involve relieving an autoinhibition based on the arginine-rich N-terminal of NXF1 together with additional components of the TREX complex (41, 43).</p><p>A hypothesis for the function of these DEAD-box helicases in rectifying the Brownian motion of the Mex67:Mtr2:mRNA complex has envisaged two reciprocal remodeling steps (7, 19), perhaps in some ways analogous to the remodeling mediated by DEAD/H-box helicases during splicing (47, 48). In the first step of such a mechanism, Sub2/UAP56 hydrolyzes ATP to generate the mRNA conformation to which the NXF1:NXT1/Mex67:Mtr2 export factor binds in the nucleus, thereby generating an export-competent mRNP that can then diffuse back and forth through the nuclear pore transport channel. Once in the cytoplasm, the complex encounters Dbp5 that, in conjunction with Gle1 and Nup42, hydrolyzes ATP to again remodel the complex and revert to the initial RNA conformation, facilitating release of the export factor and thereby preventing return of the transcript to the nucleus. NXF1:NXT1/Mex67:Mtr2 is then recycled to the nucleus to participate in a further export cycle.</p><p>Little information is available about the precise nature of the remodeling of the RNA that mediates binding of the transport factor in the nucleus and its release in the cytoplasm, although the viral constitutive transport element (CTE) may give some clues to the RNA secondary structure elements that have greater affinity for NXF1:NXT1. The CTE that is found in some unspliced RNAs from simple retroviruses enables them to bypass the normal pathway for generating mature mRNPs and use the NXF1:NXT1 pathway more efficiently than host transcripts for nuclear export. The ∼130-nt CTE RNA has a 2-fold symmetric motif that enables it to bind to NXF1:NXT1 without a requirement for adapter proteins such as ALYREF (Fig. 2B). Each motif in the CTE RNA is predicted to form a distinctive L-shaped stem loop, and structural studies have indicated how this motif binds primarily to the RRM, LRR, and NTF2-like domains as well as suggested how NXF1:NXT1 could dimerize to form a platform with its binding sites arranged to bind the two motifs simultaneously and so facilitate rapid nuclear export of the viral RNA (49, 50). It is likely that a similar RNA secondary structure is generated in host transcripts by nuclear DEAD-box helicases to facilitate NXF1:NXT1 binding, and this is probably a metastable conformation that is easily reversed by the DBP5 helicase to release the transport factor in the cytoplasm.</p><p>In addition to facilitating processing and coordinating steps in the nuclear processing of transcripts, some of the proteins that become attached to transcripts in the nucleus result in compaction of the mRNP. As a result of this compaction, in electron micrographs of mature yeast mRNPs, the particles are considerably shorter than would be expected if their RNA had an extended conformation (51), which is also consistent with their observed diffusion properties in vivo (40). One well-characterized protein associated with compaction is Nab2 in budding yeast (51), which, in addition to binding to and regulating the length of poly(A) tails, also binds to A-rich regions in the coding region (51–53) and, through its potential to form dimers that are bound to different regions of the transcript, could facilitate compaction (54). SR proteins also appear to contribute to compaction of mRNPs (2).</p><p>Mature transcripts become concentrated at the nuclear face of the pores before they are exported, and indeed the time mRNPs reside at the nuclear face is considerably longer than that taken to pass through the nuclear pores to the cytoplasm (39, 55). It is thought that proteins located in the nuclear basket, such as Mlp1 in budding yeast and possibly TPR in metazoans, contribute to this process (56). The N-terminal domain of the poly(A)-binding protein Nab2, for example, binds to Mlp1 and so could facilitate the localization of mature transcripts to the nuclear face of the pores (56). In metazoans, the TREX-2 complex, which is based on a core of GANP to which THP1, DSS1, and ENY2 are bound (57), has been proposed to contribute to chaperoning mRNPs generated in processing centers deep in the nucleus to the nuclear pores to facilitate their transport (58, 59).</p><!><p>The export of mature transcripts to the cytoplasm for translation represents the culmination of the nuclear phase of the gene expression pathway (Fig. 2). It is important that export-competent mature mRNPs are only generated when splicing and polyadenylation have been completed, and so there are checkpoints that monitor completion of each process. Almost all genes in higher eukaryotes contain introns, and although introns are less common in budding yeast, many highly expressed genes contain introns so that roughly half of the transcripts generated require splicing (for a review, see Ref. 47). Polyadenylation is critical for the stability of most transcripts and so needs to be completed successfully. Although the checkpoints for both polyadenylation and splicing are similar between yeast and higher eukaryotes, there are some differences to accommodate differences in detailed machinery involved, and generally each process tends to be understood in greater detail in yeast. For example, in yeast the Zn-finger poly(A) RNA–binding protein Nab2 (Fig. 4) appears to play a central role in coordinating the nuclear processing and export machinery, but the role of its metazoan analogue ZC3H14 is less well defined (60, 61).</p><!><p>Nab2 structure and its binding to poly(A) RNA. A, schematic illustration of the domains present in Nab2. The N-terminal Nab2N domain is essential and interacts with Mlp1 and Gfd1. The RGG domain also contains the nuclear localization sequence that is recognized by transportin to return Nab2 to the nucleus after an export cycle. There are seven Zn fingers arranged in three groups (ZnF1+2, ZnF3+4, and ZnF5–7). The fingers within each group interact so that they have a defined orientation to one another. B, arrangement of Nab2 Zn fingers 5–7 that impairs binding of a single poly(A) RNA chain to all three simultaneously (based on PDB code 5L2L). C, dimerization of Nab2 Zn fingers 5–7 brought about by binding A11G RNA (54). The RNA binds to both Nab2 chains to generate the dimer, and it is likely that similar dimers can be formed with full-length Nab2 in vivo (based on PDB code 5L2L).</p><!><p>It is crucial that the intervening intron sequences that are present between the exon coding sequences of transcripts are removed by splicing before translation takes place. Most metazoan genes and many yeast genes contain introns, and these are removed in the nucleus by the spliceosome through a complex series of reactions that function to join the 3′ end of one exon to the 5′ end of the next exon (for a review, see Ref. 47). Although the overall splicing mechanism appears to be retained between different eukaryotes (for a review, see Ref. 47), there are some differences in the way in which completion is indicated and how splicing is integrated with mRNP maturation and nuclear export. In metazoans, when each splicing reaction is completed, an exon junction complex (EJC) is deposited close to where the two exons are joined, ∼20–24 nucleotides upstream from the 5′ end of the splice junction, and serves as a platform for interacting with a broad range of other components of the gene expression pathway, both in the nucleus and the cytoplasm. The most 5′ EJC has been proposed to function to recruit the TREX complex that, together with the 5′ cap component CBP80, facilitates binding of NXF1:NXT1 and nuclear export of the mature mRNPs (46). In addition, a broad spectrum of interactions have been demonstrated between the splicing and polyadenylation machineries in both yeast and vertebrate cells (62).</p><p>Impeding the nuclear processing and subsequent nuclear export of incompletely or incorrectly spliced transcripts has been proposed to facilitate their elimination by the exosome and has led to a kinetic model of surveillance (62). In yeast, one way in which the export of intron-containing transcripts is impeded involves the splicing and retention complex (63), together with contributions made by Mlp1 and by Nab2, a poly(A)-binding protein that is important in controlling the length of poly(A) tails in Saccharomyces cerevisiae (62) that also shows genetic interactions with splicing factors, most notably with the Mud2 component of the U1 small nuclear RNP (61). However, because they fail to copurify after RNase treatment, it is not clear whether Mud2 and Nab2 interact directly or instead may both bind to the same transcript (61). Nab2 is not required for efficient splicing in vitro, but nab2 mutants do show a mild increase in splicing defects in vivo, although these do not appear to involve Nab2 having a direct role in splicing itself but rather a quality control function to ensure that only completely spliced transcripts undergo further processing and export (62). Reciprocally, Mud2 deletion generates longer poly(A) tails, reinforcing the evidence for a functional interaction between the splicing and polyadenylation machineries. Moreover, the mammalian analogue of Nab2, ZC3H14, also interacts functionally with a number of spliceosome components (61). Mlp1 and Mlp2, components of the nuclear basket, also function to inhibit the nuclear export of unspliced transcripts, either by retaining them in the nucleus or, alternatively, by accelerating the export of spliced transcripts (64).</p><!><p>Although the lengths of poly(A) tails (∼250 nt in higher eukaryotes but only ∼60 nt in S. cerevisiae) and the detailed mechanisms by which length is controlled vary, similar large multisubunit complexes mediate cleavage in the 3′-UTR and subsequent generation of a poly(A) tail in eukaryote transcripts before they are exported to the cytoplasm (65–67). In S. cerevisiae, these processes are mediated by the cleavage and polyadenylation factor (CTF) that is organized into three structural modules that mediate key functions: a nuclease module that cleaves the transcript; a polymerase module in which Cft1 binds four other components, including poly(A) polymerase (Pap1) and Fip1; and a phosphatase/APT module that regulates 3′ end processing (65). Metazoans employ the analogous CPSF complex (66, 67). The polyadenylation signal, the sequence motif recognized by the RNA cleavage complex, varies among groups of eukaryotes. Most human polyadenylation sites contain the AAUAAA sequence (68), but this sequence is less common in plants and fungi.</p><p>Once the 3′-UTR RNA is cleaved, polyadenylation starts, catalyzed by poly(A) polymerase that binds the growing end of the poly(A) tail. Poly(A) polymerase is composed of three domains that encircle the active site (69), and in vivo, the poly(A) chain is prevented from leaving the enzyme by processivity factors, such as PABPN1 in mammals (70) or CF1 and CPF in yeast (69, 71), that prevent dissociation of the poly(A) tail from the enzyme following the addition of each new nucleotide. In S. cerevisiae, Fip1 regulates the activity of poly(A) polymerase (Pap1) through multiple interactions (72) and is thought to be flexible, which would help accommodate the growing poly(A) loop (60, 73, 74) that results from the poly(A) tail being held at both ends (Fig. 5). The processive action of Pap1 requires attachment to the CPF via Fip1 (75) and possibly other components. Polyadenylation is terminated by release of Pap1 from the CPF (67), after which the enzyme ceases to be processive and dissociates from the poly(A) tail. The length of the poly(A) tail is regulated by PABPN1 (for a review, see Ref. 76) and ZC3H14 (77) in higher eukaryotes and Nab2 in S. cerevisiae (60). In higher eukaryotes, each PABPN1 chain is thought to bind ∼12–15 adenines (78, 79), and when a sufficient number of PABPN1 molecules are bound, steric factors are proposed to result in the dissociation of poly(A) polymerase from CPSF (66, 67). In vitro, PABPN1 forms ∼21-nm-diameter roughly spherical particles when bound to poly(A) RNA that seem to contain about 200–300 nt and so would be attractive candidates for the oligomerization that results in dissociation of poly(A) polymerase from CPSF (78).</p><!><p>Model for the regulation of poly(A) tail length by Nab2 in yeast. Poly(A) tail length in S. cerevisiae appears to be regulated by a mechanism analogous to that proposed for the way in which PABPN1 functions in metazoans (66, 67) to terminate polyadenylation by dissociating poly(A) polymerase from the cleavage and polyadenylation machinery (CPSF) so that it ceases to be processive. Each PABPN1 binds to ∼12–15 adenines, and when a sufficient number have been added, they form an approximately spherical aggregate that is proposed to stiffen the poly(A) tail, forcing poly(A) polymerase to dissociate from the CPSF, after which it ceases to be processive and dissociates from the poly(A) tail, terminating polyadenylation. An analogous model for polyadenylation termination in S. cerevisiae envisages that following cleavage, poly(A) polymerase (Pap1) synthesizes the poly(A) tail processively but only while it remains attached to the polyadenylation factor (CPF) through binding to Fip1 and Cft1 (for simplicity, the cleavage module has been omitted from the figure). Pap1 holds the growing poly(A) tail at one end while Rna15 and Cft1 hold its other end so that it forms a loop that can be accommodated by the flexibility of both the poly(A) chain and Fip1 (A). However, when a sufficient number of adenosines (probably ∼60) have been added to facilitate Nab2 binding, the reduction in flexibility of the poly(A) tail produced by Nab2 dimerization introduces a stress that detaches Pap1 from Cft1 and Fip1 (B). Once detached from Fip1, the processive activity of Pap1 is impaired and it dissociates from the tail, thereby terminating polyadenylation. The termination of polyadenylation then appears to be signaled to the TREX complex via CF1A component Pcf11 to initiate the formation of an export-competent mRNP. The highly schematic representation of the CF1A and polymerase modules of CPF is based on the recent cryo-EM structure of this complex (65).</p><!><p>Termination of polyadenylation in S. cerevisiae is mediated by Nab2 (Fig. 4A), which contains seven Zn fingers (76), and could result from either inhibiting Pap1 (by detaching it from Flp1 or from Cft1 or by rephosphorylation of Pta1) or removing the poly(A) tail from Pap1. Nab2 Zn fingers 5–7, which are necessary and sufficient for high-affinity poly(A) RNA binding (80, 81), adopt a conformation in which finger 6 is oriented on the opposite side of the chain to fingers 5 and 7 (54, 80), impairing all three fingers binding simultaneously to a short stretch of a poly(A) chain (Fig. 4B). In vitro, binding poly(A) RNA induces Nab2 Zn fingers 5–7 to dimerize (Fig. 4C), and this feature could provide a mechanism (Fig. 5) analogous to that proposed for PABPN (68, 82) in metazoans for terminating polyadenylation (54). This would propose that, when the poly(A) tail had become sufficiently long to wrap around two Nab2 chains so that they form a dimer, the structure generated would constrain conformation of the poly(A) loop sufficiently to weaken the attachment of Pap1 to Fip1 and CPF (Fig. 5) so that it detaches, after which it ceases to be processive and dissociates from the poly(A) tail. Consistent with this hypothesis, cryo-EM studies (65) of S. cerevisiae CPF observed two different objects, either with or without Pap1, indicating that Pap1 can be dissociated from CPF. Once Pap1 has been dissociated from the CPF, it is likely that the transcript is able to dissociate from the complex, freeing the mRNP for nuclear export. In S. cerevisiae, it is likely that, following the termination of polyadenylation and dissociation of the CPF from the transcript, conformational changes in the CPF are transmitted to the TREX complex, possibly via CFP component Pcf11. This, in turn, may mediate the attachment and SUMOylation of Yra1 (44, 83) and activation of Sub2 and/or Dbp2, resulting in the attachment of Mex67:Mtr2 and the generation of an export-competent mRNP.</p><!><p>Overall, the mechanism by which transport factors mediate the transport of mature transcripts across the nuclear envelope through nuclear pores and the machinery by which transport factors are released in the cytoplasm are understood in some detail. Major questions outstanding are how transcripts are remodeled to bind transport factors and release adapters and how successful completion of the nuclear processing segment of the gene expression pathway is signaled to initiate generation of export-competent mRNPs.</p><p>To enable the export of mRNA through nuclear pores to the cytoplasm, it is necessary both to add nuclear export factors and to release the transcript from RNA polymerase and nuclear processing machinery. After cleavage in the 3′-UTR, the RNA polymerase is still probably attached to the transcript through interactions between its C-terminal domain and the cleavage and polyadenylation machinery that are likely retained until polyadenylation has been completed. Similarly, retention factors appear to impair export until splicing has been completed. Both splicing and polyadenylation appear to be coordinated with the attachment of nuclear export factors, but the temporal sequence of events remains unclear. Indeed, these steps may not necessarily have to occur in a defined sequence, and instead export may only occur when all have been completed. Although most of the factors involved in the generation of export-competent mRNPs have been identified, the precise mechanism by which they function to coordinate the different steps of the nuclear processing machinery is less clear, and a number of aspects remain to be clarified.</p><p>(i) Although attachment of nuclear export factors is a prerequisite for movement through nuclear pores to the cytoplasm, it remains unclear how many of these factors are attached to an individual transcript. In higher eukaryotes, the EJC and 5′ cap generate attachment mediated through TREX, whereas completion of polyadenylation, at least in budding yeast, is thought to generate an analogous attachment. It appears that the 5′ and 3′ ends of the transcript are often quite close (85), but it is also possible that a number of export factors are also added along the body of the complex. Although cross-linking and immunoprecipitation studies indicate that ALYREF binds to a region near the 5′ end of the mRNA in a CBP80-dependent manner, they also indicate that it binds near the 3′ end in a PABPN1-dependent manner (86), and other studies (53) indicate binding throughout the coding region. It is not clear how many export factors bind to a given transcript or what features their binding site (or the site to which adapters such as Yra1 bind) might have.</p><p>(ii) The termination of polyadenylation appears to be mediated by dissociation of poly(A) polymerase from the cleavage and polyadenylation machinery as a result of steric factors associated with the binding of PABPN1 in higher eukaryotes and Nab2 in budding yeast. In vitro, PABPN forms globular aggregates on poly(A) RNA (78) that are proposed to eventually force the dissociation of poly(A) polymerase (for a review, see Ref. 87), but the precise structure of these aggregates needs to be established to define how many PABPN1 chains each contains and how the chains are arranged to generate a defined particle size. In budding yeast, Nab2 clearly controls poly(A) tail length (60) and probably mediates this function through dimer formation (54), but the structure of the Nab2 dimers in vivo and whether there are one or two dimers per tail remain to be established.</p><p>(iii) Polyadenylation is terminated by the dissociation of poly(A) polymerase from cleavage and polyadenylation machinery, but how this machinery detaches from the 3′-UTR has not been established, although it appears that the generation of export-competent mRNPs by Mex67 binding contributes to this step (88). It may be that the attachment to the 3′-UTR is sufficiently weak to simply dissociate, but it seems more likely that dissociation of poly(A) polymerase generates some conformational change and/or post-transcriptional modification (such as phosphorylation or SUMOylation) in the complex that could also facilitate its release, although detailed evidence for this point is lacking.</p><p>(iv) In at least budding yeast, the termination of polyadenylation signals a TREX/Sub2-mediated remodeling of the transcript that results in Yra1 dissociation and Mex67:Mtr2 attachment (Fig. 3). Pcf11 appears to be important in this step (83), but precisely how poly(A) polymerase dissociation influences Pcf11 and how this in turn influences TREX/Sub2 remain obscure.</p><p>(v) In budding yeast, there are indications of Mud2 functioning to coordinate splicing and polyadenylation, but the precise mechanism by which this is mediated and the role of other factors that show genetic interactions between these functions are unclear. Similarly, intron-containing transcripts appear to be retained in the nucleus, either at the nuclear basket in budding yeast or in nuclear speckles in higher eukaryotes, and although several components of the machinery involved have been identified, the precise mechanism by which this is mediated remains obscure.</p><p>In summary, many of the components of the gene expression pathway that are involved in the nuclear processing and export of transcripts have been identified, but the precise mechanisms by which these steps are coordinated remain to be established. Although the thermal ratchet-based mechanism by which mRNPs are exported to the cytoplasm through nuclear pores has been established in considerable detail, many aspects of the generation of export-competent mRNPs in the nucleus remain unclear. Key checkpoints in the pathway leading to the generation of export-competent mRNPs, which entail establishing that both splicing and polyadenylation have been completed, need to be activated in a timely manner. Providing a comprehensive description of this complex series of integrated processes represents an important challenge and a fruitful area for future studies in this area.</p><!><p>This work was supported in part by Medical Research Council (MRC) Grants MC_U105178939, MC-A025-5PL41, and MC_UP_1201/6 and Leverhulme Emeritus Fellowship EM-2016-062 (to M. S.). The authors declare that they have no conflicts of interest with the contents of this article.</p><p>messenger ribonuclear particle</p><p>nucleoporin</p><p>GTPase-activating protein</p><p>guanine nucleotide exchange factor</p><p>RNA recognition motif</p><p>leucine-rich repeat</p><p>ubiquitin-associated</p><p>RNA-binding domain</p><p>exon-junction complex</p><p>constitutive transport element</p><p>RNA-binding protein rich in Ser and Arg</p><p>nucleotide(s)</p><p>cleavage and polyadenylation factor</p><p>cleavage and polyadenylation specificty factor</p><p>small ubiquitin-like modifier.</p>

PubMed Open Access

Fas Receptor-deficient lpr Mice are protected against Acetaminophen Hepatotoxicity due to Higher Glutathione Synthesis and Enhanced Detoxification of Oxidant Stress

Acetaminophen (APAP) overdose is a classical model of hepatocellular necrosis; however, the involvement of the Fas receptor in the pathophysiology remains controversial. Fas receptor-deficient (lpr) and C57BL/6 mice were treated with APAP to compare the mechanisms of hepatotoxicity. Lpr mice were partially protected against APAP hepatotoxicity as indicated by reduced plasma ALT and GDH levels and liver necrosis. Hepatic Cyp2e1 protein, adduct formation and hepatic glutathione (GSH) depletion were similar, demonstrating equivalent reactive metabolite generation. There was no difference in cytokine formation or hepatic neutrophil recruitment. Interestingly, hepatic GSH recovered faster in lpr mice than in wild type animals resulting in enhanced detoxification of reactive oxygen species. Driving the increased GSH levels, mRNA induction and protein expression of glutamate-cysteine ligase (gclc) were higher in lpr mice. Inducible nitric oxide synthase (iNOS) mRNA and protein levels at 6h were significantly lower in lpr mice, which correlated with reduced nitrotyrosine staining. Heat shock protein 70 (Hsp70) mRNA levels were substantially higher in lpr mice after APAP. Conclusion: Our data suggest that the faster recovery of hepatic GSH levels during oxidant stress and peroxynitrite formation, reduced iNOS expression and enhanced induction of Hsp70 attenuated the susceptibility to APAP-induced cell death in lpr mice.

fas_receptor-deficient_lpr_mice_are_protected_against_acetaminophen_hepatotoxicity_due_to_higher_glu

1. INTRODUCTION<!>2.1 Animals<!>2.2 Experimental design<!>2.3 Histology<!>2.4 Glutathione quantification<!>2.5 APAP-protein adducts<!>2.6 Western Blotting<!>2.7 Real-time PCR for mRNA quantification<!>2.8 Statistics<!>3.1 Liver injury in C57BL/6 and lpr mice<!>3.2 Inflammation and neutrophil recruitment during APAP overdose<!>3.3 Metabolism of APAP in C57BL/6 and lpr mice<!>3.4 Protection in lpr mice due in part to enhanced glutathione recovery<!>3.5 Attenuated injury in lpr mice correlates with decreased peroxynitrite formation<!>3.6 Increased expression of hepatoprotective genes in lpr mice<!>4. DISCUSSION<!>4.1 Inhibition of metabolic activation<!>4.2 Inhibition of Fas receptor-mediated apoptosis<!>4.3 Attenuation of an inflammatory response<!>4.4 Induction of GSH synthesis and hepatoprotective genes<!>4.5 Summary

<p>Acetaminophen (APAP) overdose is a major clinical problem in the US and Europe, often resulting in severe liver injury and potentially acute liver failure (Lee, 2004; Larson et al., 2005). Many studies have described the mechanisms of injury which are initiated by the metabolism of APAP to the reactive metabolite N-acetyl-p-benzoquinone imine (NAPQI) resulting in glutathione (GSH) depletion and adduction of cellular proteins (Cohen et al., 1997). In particular, it is thought that the binding of NAPQI to mitochondrial proteins results in mitochondrial dysfunction (Nelson, 1990), which propagates injury through a mitochondrial oxidant stress and peroxynitrite formation (Cover et al., 2005) resulting ultimately in the opening of the mitochondrial membrane permeability transition (MPT) pore with collapse of the membrane potential and ATP depletion (Kon et al., 2004; Masubuchi et al., 2005; Ramachandran et al., 2011). Early mitochondrial bax translocation and the later MPT lead to permeabilization of the outer membrane with release of the intermembrane proteins endonuclease G and apoptosis-inducing factor (AIF), which translocate to the nucleus and initiate DNA fragmentation (Bajt et al., 2006, 2008, 2011). Together, these events lead to hepatocyte necrosis (Gujral et al., 2002) and the release of damage associated molecular patterns (DAMPs), which trigger a sterile inflammatory response (Jaeschke et al., 2012b).</p><p>Although APAP-induced cell death has some overlap with features of apoptosis, such as mitochondrial bax translocation, cytochrome c release, and DNA fragmentation, the overwhelming evidence suggests a necrotic cell death. Most importantly, morphological characteristics include cell swelling, karyorrhexis and vacuolation (Gujral et al., 2002). In addition, there is no relevant caspase activation (Lawson et al., 1999; Adams et al., 2001, El-Hassan et al., 2003), and pancaspase inhibitors neither protect against cell death nor prevent DNA fragmentation (Lawson et al., 1999; Cover et al., 2005; Jaeschke et al., 2006; Williams et al., 2010b, 2011b). However, in the past decade several studies have associated the Fas receptor, which can trigger apoptosis, with APAP hepatotoxicity. First, it has been shown that serum levels of soluble Fas receptor are increased in patients with APAP-induced liver failure (Tagami et al., 2003). An experimental study used a small-interfering-RNA (siRNA) construct to target Fas and showed that knocking it down resulted in reduced APAP-induced injury (Zhang et al., 2000). Another study demonstrated that protection against APAP induced injury could be observed in lpr (Fas receptor-deficient) mice (Liu et al., 2004). However, it was unclear in these reports why preventing Fas receptor signaling could reduce APAP hepatotoxicity. The Fas receptor is expressed on many cell types, including hepatocytes, but the highest expression of Fas can be seen on immature lymphocytes (Nisihara et al., 2001); this is part of an essential mechanism to eliminate self-reactive lymphocytes and is the cause of the autoimmune phenotype observed in lpr mice. This phenotype includes high lymphocyte counts, greatly enlarged lymph nodes and spleen, circulating rheumatoid factor, and an overall autoimmune phenotype (Hutcheson et al., 2008). Interestingly, lpr mice were protected against bile duct ligation-induced liver injury due to the reduced inflammatory response in the absence of apoptosis (Gujral et al., 2004). This suggested that lpr mice can show reduced liver injury independent of Fas receptor-induced apoptosis. Thus, the objective of this investigation was to evaluate the mechanisms of protection against APAP hepatotoxicity in Fas receptor-deficient lpr mice by comparing APAP-induced toxicity between C57BL/6 and lpr mice in vivo.</p><!><p>Eight to twelve week old male C57BL/6J and age-matched B6.MRL-Faslpr/J (lpr) mice were purchased from Jackson Labs (Bar Harbor, ME) with an average weight of 19 to 24 g and maintained at the University of Kansas Medical Center. All animals were housed in environmentally controlled rooms with 12 h light/dark cycle and allowed free access to food and water. Experiments followed the criteria of the National Research Council for the care and use of laboratory animals in research and were approved by the University of Kansas Medical Center Institutional Animal Care and Use Committee.</p><!><p>All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) unless otherwise stated. Mice were intraperitoneally (i.p.) injected with 300 mg/kg APAP (dissolved in warm saline) or with an equivalent volume of saline after overnight fasting (~15 h). Animals were terminated at 0 h (n=4 per strain), 0.5 h (n=4 per strain), 6 h (n=7 per strain) or 24 h (n=7 per strain) after APAP. Blood was drawn into heparinized syringes for measurement of plasma alanine aminotransferase (ALT) activity (Pointe Scientific, Canton, MI) and glutamate dehydrogenase (GDH) activity as described (McGill et al., 2012). The liver was removed and was rinsed in cold saline; liver sections were fixed in 10% phosphate buffered formalin for histological analyses. The remaining liver lobes were snap-frozen in liquid nitrogen and stored at −80 °C.</p><!><p>Formalin-fixed tissue samples were embedded in paraffin and 5 μm sections were cut. Sections were stained with hematoxylin and eosin (H&E) for blinded evaluation of the areas of necrosis by the pathologist. The percent of necrosis was estimated by evaluating the number of microscopic fields with necrosis compared to the total cross sectional area. Additional sections were stained for neutrophils using the anti-mouse neutrophil allotypic marker antibody (AbD Serotec, Raleigh, NC) as previously described (Williams et al., 2010a). Positively stained neutrophils consistent with cellular morphology were quantified in 15 high power fields (HPF). Sections were also stained using an anti-nitrotyrosine antibody (Invitrogen, Carlsbad, CA) as previously described (Knight et al., 2002).</p><!><p>Glutathione (GSH) and glutathione disulfide (GSSG) were measured from liver homogenate using the modified Tietze method as previously described in detail (Jaeschke and Mitchell, 1990). Briefly, frozen tissue was homogenized in sulfosalicylic acid/EDTA. For total GSH determination samples were assayed using dithionitrobenzoic acid. Similarly, measurement of GSSG was performed using the same method after trapping and removal of GSH with N-ethylmaleimide.</p><!><p>Liver APAP protein adducts were measured based on a method developed by Muldrew et al. (2002) with minor modifications (Ni et al., 2012) and normalized to total protein (MicroBCA kit, Thermo Scientific, Rockford, IL).</p><!><p>Liver tissue was homogenized in 25 mM HEPES buffer (containing 5 mM EDTA, 2 mM DTT and 0.1% CHAPS) and diluted to uniform protein concentration. Tissue homogenate was used for β-actin (Santa Cruz Biotech, Santa Cruz, CA), cyp2e1 (Abcam, Cambridge, MA), iNOS (BD, Franklin Lakes, NJ) and gclc (GeneTex, Irvine, CA) western blotting as described in detail (Bajt et al., 2000).</p><!><p>mRNA expression of several genes was performed by real-time PCR (RT-PCR) analysis as previously described (Gujral et al., 2004). cDNA was generated by reverse transcription of total RNA by M-MLV reverse transcriptase with random primers (Invitrogen, Carlsbad, CA). Forward and reverse primers for the genes were designed using Primer Express software (Applied Biosystems, Foster City, CA) and listed in supplemental table 1. After normalization of cDNA concentration, SYBR green PCR Master Mix (Applied Biosystems) was used for analysis. The relative differences in expression between groups were expressed us ing cycle time (Ct) values generated by the ABI 7900 instrument (Applied Biosystems). All genes evaluated were first normalized to GAPDH and then expressed as a fold increase relative to control (arbitrarily set as 1.0). Calculations are made by assuming one cycle is equivalent to a two-fold difference in copy number which is the 2^(-ddCt) formula.</p><!><p>All results were expressed as mean ± SE. Comparisons between multiple groups were performed with one-way ANOVA or, where appropriate, by two-way ANOVA, followed by a post hoc Bonferroni test. If the data were not normally distributed, the Kruskal-Wallis test (nonparametric ANOVA) followed by Dunn's Multiple Comparisons Test was used. P < 0.05 was considered significant. All statistics were evaluated using SigmaStat (Systat Software, San Jose, CA).</p><!><p>To assess the differences in liver injury, lpr and wildtype C57BL/6 mice were treated with 300 mg/kg APAP for 6 or 24 h. APAP caused severe liver injury in wildtype animals as indicated by the massive increase in plasma ALT activities and the development of severe centrilobular necrosis (Fig. 1A,B,D; Supplemental Figure 1). To evaluate the degree of mitochondrial injury, the mitochondrial-specific enzyme glutamate dehydrogenase (GDH) was measured in plasma. GDH activities also increased substantially at 6 and especially at 24 h after APAP administration (Fig. 1C). All parameters indicated that APAP overdose caused significantly less injury in lpr mice at both time points (Fig. 1A-D), however the difference in injury appeared to be most pronounced at 6 h.</p><!><p>It was hypothesized that altered inflammatory response and neutrophil recruitment into the liver potentiates APAP induced injury in lpr mice (Liu et al. 2004). To further evaluate this hypothesis, hepatic cytokine and chemokine induction (Table 1) and hepatic neutrophil recruitment were quantified in lpr and C57BL/6 mice (Fig. 1E,F). Some variation in cytokine and chemokine induction was observed between genotypes with C57BL/6 mice having higher IL-6 and IL-10 induction at 6h while CXCL1 induction was higher in lpr mice at 6 and 24 h. Despite these differences, the overall pattern of cytokine and chemokine induction was similar between genotypes. To determine if these slight differences altered neutrophil recruitment, immunohistochemistry was performed (Fig. 1E,F). When quantified, the distribution and total number of hepatic neutrophils was equivalent between genotypes despite a reduced injury in lpr mice.</p><!><p>To determine if the protection seen in lpr mice was due to inhibition of reactive metabolite formation, we measured the time course of GSH depletion and recovery and APAP-protein adducts. Fasted control lpr mice had 26% more GSH than fasted control C57BL/6 mice (Fig. 2A). Thirty minutes after APAP injection liver GSH levels were depleted >80% in both genotypes (Fig. 2A). In addition, APAP-protein adducts were similar at 6 h and at 24 h (Fig. 2D). Furthermore, western blot analysis of cyp2e1, the dominant Cyp responsible for metabolism of APAP, showed no significant difference between wild type and lpr mice (Fig. 2E). Together, these data show that the metabolic activation of APAP is similar in both genotypes. Interestingly, during the recovery phase of the hepatic GSH levels, i.e. after APAP was metabolized, the liver GSH content recovered faster in lpr mice compared to the wildtype animals (Fig. 2A).</p><p>During APAP overdose, GSH is critical for detoxifying reactive oxygen species (ROS), particularly in the mitochondria, generating glutathione disulfide (GSSG) selectively in mitochondria as a result (Jaeschke, 1990). Quantification of GSSG showed increased detoxification of ROS in lpr mice compared to C57BL/6 at 6 h and an even more dramatic increase at 24 h (Fig. 2B). When GSSG is normalized to total GSH the difference between genotypes is lost at 6 h, but at 24 h lpr mice had a 3-fold higher GSSG-to-GSH ratio compared to C57BL/6 (Fig. 2C). These data show that lpr mice have higher liver GSH and enhanced capacity to detoxify reactive oxygen generated during mitochondrial dysfunction.</p><!><p>The rate limiting step in GSH synthesis is catalyzed by glutamate-cysteine ligase which is composed of the catalytic (gclc) and modulatory (gclm) subunits. To investigate why lpr mice have higher basal GSH levels and enhanced GSH recovery after APAP, we measured gclc mRNA and protein levels in the liver. Basal levels of gclc mRNA were slightly higher in lpr mice, however, the difference did not reach statistical significance. The same trend was observed 30 minutes after APAP (Fig. 3A). Western blotting showed no difference in gclc protein levels between genotypes in control or 30 minute APAP-treated mice (Fig. 3C). Six hours post-APAP lpr mice had substantially higher mRNA induction of gclc (Fig. 3A) and more liver gclc protein (Fig. 3C). By 24 h post-APAP, gclc mRNA levels declined in lpr mice but further increased in wild type animals (Fig. 3A). Protein levels of gclc were equivalent between genotypes at 24 h (Fig. 3C). These data show no quantifiable difference in gclc protein in untreated animals, however, the induction and translation of gclc occurs sooner in lpr mice allowing for increased GSH levels in these mice, and improving the scavenging capacity for reactive oxygen species.</p><!><p>One of the most powerful oxidants in vivo is peroxynitrite, which can be formed when superoxide generated from mitochondrial dysfunction reacts with nitric oxide produced in part by nitric oxide synthases (Pryor and Squadrito, 1995). During APAP overdose, the inflammatory response activates the inducible form of nitric oxide synthase (iNOS) thereby generating nitric oxide which is freely diffusible across cell membranes. Although lpr mice showed slightly higher baseline values of iNOS protein expression compared to wild type animals, there was only a mild induction after APAP treatment (Fig. 3C). In contrast, extensive iNOS induction was observed in wild type mice with peak levels at 6 h, and returned to baseline by 24 h (Fig. 3C). Consistent with the stronger iNOS expression at 6 h, wild type animals stained more extensively for nitrotyrosine adducts, which did not significantly change at 24 h (Fig. 3D). In contrast, there was very limited nitrotyrosine staining in the livers of lpr mice at 6 h but more extensive staining at 24 h (Fig. 3D). Together these data suggest that the more robust induction of iNOS, together with the ROS formation and slow recovery of GSH in wild type animals, contributes to the more severe liver injury in wild type compared to lpr mice.</p><!><p>The induction of a number of endogenous genes has been shown to limit APAP-induced liver injury. These include metallothionein (Mt) (Saito et al., 2010a), heme oxygenase-1 (Ho-1) (Chiu et al., 2002), and heat shock protein 70 (Hsp70) (Salminen et al., 1998). Hsp70 mRNA induction was substantially higher in lpr mice compared to wild type animals (Figure 4A). Although Mt-1, Mt-2 and Ho-1 mRNAs were upregulated after APAP treatment, there was no significant difference between wild type and lpr mice (Fig. 4B-D).</p><!><p>Lpr mice are a common model to study autoimmunity, in particular systemic lupus erythematosus (SLE) (Hutcheson et al., 2008). The observed autoimmunity of lpr mice results in various reported abnormal phenotypes which alone or in combination have the potential to modulate stress responses and influence drug toxicity testing. A recent example in which a gene knockout was shown to alter important off-target functions is the chronic stress observed in conditional Atg5-deficient mice. The long-term inhibition of autophagy triggers apoptotic cell death, regeneration and Nrf-2 activation (Ni et al., 2012). This compensatory response makes the animals resistant to APAP overdose (Ni et al., 2012). For these reasons it was important to evaluate the mechanism of protection against APAP-induced liver injury in lpr mice.</p><!><p>The first concern regarding any intervention that reduces APAP hepatotoxicity has to be whether formation of reactive metabolites and protein adducts is affected. Both wild type and lpr mice experienced a similar hepatic GSH depletion during the first 30 min, which is a sensitive indirect measure of NAPQI formation (Jaeschke et al., 2011). In addition, cysteine protein adducts, which represent >90% of all APAP protein adducts (Streeter et al., 1984; Muldrew et al., 2002), were quantitatively not different between the different mouse genotypes. Thus, it can be concluded that there was no significant difference in reactive metabolite formation between wild type and lpr mice.</p><!><p>It was previously concluded that APAP hepatotoxicity is a "mixed necrotic and apoptotic event" and for this reason lpr mice have attenuated progression of injury (Liu et al., 2004). However, overwhelming evidence in animals (Gujral et al., 2002) and humans (McGill et al., 2012; Antoine et al., 2012) after APAP overdose, and in murine hepatocytes (Bajt et al., 2004; Kon et al., 2004) and human HepaRG cells (McGill et al., 2011) indicates that APAP-induced cell death occurs by oncotic necrosis not apoptosis. APAP does not cause relevant caspase activation (Lawson et al., 1999; Adams et al., 2001; El-Hassan et al., 2003; Jaeschke et al., 2006; Williams et al., 2010b, 2011b) and pancaspase inhibitors do not protect (Lawson et al., 1999; Jaeschke et al., 2006; Williams et al., 2010b/2011b; Antoine et al., 2010). As a consequence, nuclear DNA fragmentation, a well-established phenomenon after APAP overdose (Ray et al., 1990; Lawson et al., 1999; Jahr et al., 2001) cannot be prevented by caspase inhibitors but only by inhibiting mitochondrial dysfunction (Cover et al., 2005). Nuclear DNA fragmentation occurs through mitochondria-derived endonuclease G and AIF (Bajt et al., 2006; 2008; 2011). Some of the confusion comes from the overlap in signaling mechanisms between apoptosis and necrosis and the fact that some assays, e.g. the TUNEL assay, DNA ladders, mitochondrial bax translocation and cytochrome c release, are not specific for apoptosis (Jaeschke and Lemasters, 2003). Nevertheless, if multiple parameters are measured and if a positive control for apoptosis, e.g. Fas receptor or TNF receptor-induced apoptosis, is used, there is little doubt that APAP-induced cell death is caused by necrosis and thus, the protection in lpr mice cannot be caused by inhibition of Fas receptor-mediated apoptosis.</p><!><p>Previous studies have shown a reduced inflammatory injury after bile duct ligation in lpr mice (Gujral et al., 2004). This opens up the possibility that a reduced inflammatory response might explain the protection against APAP toxicity in lpr mice. The extensive necrosis induced by APAP overdose leads to extensive release of damage associated molecular patterns, including high mobility group box-1 protein and DNA fragments in mice and humans (Martin-Murphy et al., 2010; Antoine et al., 2009, 2012; McGill et al., 2012). As a result, there is cytokine formation and hepatic neutrophil and monocyte recruitment (Lawson et al., 2000; James et al., 2005; Holt et al., 2008). However, whether or not these inflammatory cells actually aggravate the injury is controversial (Jaeschke et al., 2012b). Although there are papers that suggested that neutrophils may cause additional injury (Liu et al., 2006, Ishida et al., 2006), the effect can be explained by off-target effects of the neutropenia-inducing antibody (Jaeschke and Liu, 2007). In contrast, numerous interventions that prevent neutrophil cytotoxicity such as antibodies against CD18, inhibitors of NADPH oxidase as well as deficiency of various adhesion molecules or NADPH oxidase are all ineffective in attenuating APAP hepatotoxicity (Lawson et al., 2000; James et al., 2003a; Cover et al., 2006, Williams et al., 2010a,b; 2011a). Since pro-inflammatory cytokine formation and hepatic neutrophil recruitment was not significantly different between wild type and lpr mice, and given the extensive evidence against an aggravation of injury by neutrophils in this model, it is unlikely that the reduced injury in lpr mice was caused by a reduced inflammatory response.</p><!><p>Mitochondrial oxidant stress and peroxynitrite formation are critical for APAP-induced cell death (Jaeschke et al., 2012a). The extensive depletion of GSH, particularly in the mitochondria, severely impairs this endogenous defense mechanism and makes the cell highly susceptible to oxidant stress. Therefore, replenishing mitochondrial GSH levels by supply of sulfhydryl reagents is highly effective for scavenging ROS and peroxynitrite and attenuating APAP-induced liver injury, which promotes regeneration and recovery (Knight et al., 2002; Bajt et al., 2003; James et al., 2003b; Saito et al., 2010b). Our results demonstrate that hepatic GSH levels recover faster in lpr mice and this correlates with increased GSSG formation and less nitrotyrosine adducts. This effect may have been triggered in part by the more extensive induction of gclc, the rate-limiting enzyme of the GSH synthesis pathway (Lu, 2009). This suggest that the accelerated recovery of GSH concentrations in the hepatocytes, which translates in a faster uptake of GSH into mitochondria (Saito et al., 2010b), detoxified more effectively the oxidant stress and peroxynitrite resulting in less injury in lpr mice. A similar mechanism, i.e. enhanced GSH recovery, has been proposed to make female mice less susceptible to APAP-induced liver injury (Masubuchi et al., 2011). In addition to the effects on gclc, the attenuated induction of iNOS in lpr mice may have contributed to reduced peroxynitrite formation. Thus, at least during the early phase of APAP-induced injury, reduced peroxynitrite formation together with improved detoxification of oxidant stress appear to be the dominant mechanisms of protection in lpr mice. Similar to effects observed with GSH treatment (Bajt et al., 2003), the initial protection fades somewhat during prolonged oxidant stress.</p><p>In addition to GSH recovery, other enzymes, e.g. Ho-1, are known to generate antioxidants and protect against APAP overdose (Chiu et al., 2002). In addition, Mt-1/2 can scavenge NAPQI (Saito et al., 2010a) and heat shock proteins, e.g. Hsp70, can protect against APAP toxicity (Tolson et al., 2006). However, neither Mt1/2 nor Ho-1 mRNA was induced differently between wild type and lpr mice. In contrast, induction of Hsp70 mRNA was dramatically higher in lpr mice compared to wild type animals which may contribute to the reduced injury in lpr mice.</p><!><p>Our data confirmed the reduced susceptibility of Fas receptor-deficient lpr mice against APAP hepatotoxicity. There was no evidence that the protection was related to inhibition of metabolic activation, Fas receptor-mediated apoptosis or modulation of the hepatic inflammatory response. In contrast, the faster recovery of hepatic GSH levels during the mitochondrial oxidant stress and peroxynitrite formation, reduced iNOS induction and enhanced expression of Hsp70 attenuated the detrimental effects of the oxidant stress and thus reduced the susceptibility to APAP-induced cell death. Thus, when working with gene-deficient animals it is critical to consider and mechanistically evaluate potential off-target effects as the reason for the modified susceptibility against hepatotoxic agents. Often these gene deficient animals have multiple pathologies and compensatory responses that are not sufficiently considered in the interpretation of the data. This can lead to unjustified conclusions about therapeutic targets that are not correct for the experimental model and could not be translated into clinical practice.</p>

PubMed Author Manuscript

Are boat transition states likely to occur in Cope rearrangements? A DFT study of the biogenesis of germacranes

It has been proposed that elemanes are biogenetically formed from germacranes by Cope sigmatropic rearrangements. Normally, this reaction proceeds through a transition state with a chair conformation. However, the transformation of schkuhriolide (germacrane) into elemanschkuhriolide (elemane) may occur through a boat transition state due to the final configuration of the elemanschkuhriolide, but this transition state is questionable due to its high energy. The possible mechanisms of this transformation were studied in the density functional theory frame. The mechanistic differences between the transformation of (Z,E)-germacranes and (E,E)-germacranes were also studied. We found that (Z,E)-germacranolides are significantly more stable than (E,E)-germacranolides and elemanolides. In the specific case of schkuhriolide, even when the boat transition state is not energetically favored, a previous hemiacetalization lowers enough the energetic barrier to allow the formation of a very stable elemanolide that is even more stable than its (Z,E)-germacrane.

are_boat_transition_states_likely_to_occur_in_cope_rearrangements?_a_dft_study_of_the_biogenesis_of_

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

<p>Germacranes are biogenetic precursors of elemanes [1][2][3][4], because germacranes can be easily transformed into elemanes by heating through a Cope rearrangement. In some cases, these transformations are so favorable that it has been mentioned that the observed elemanes are only artifacts produced at the extraction [5][6][7][8]. It is known that 1,5-dienes suffer Cope rearrangements at temperatures between 200 and 300 °C, but some struc-tural changes in the diene, such as the anionic oxy-Cope transformation allows the reactions to happen at temperatures below 0 °C [9]. The Cope rearrangement is a [3,3]-sigmatropic reaction and in general, occurs through a single transition state (TS), which has, normally, a chair conformation due to the higher energy of the boat conformation [2,7,[10][11][12][13][14][15][16][17][18][19]. In this mechanism, the electron density of the TS is delocalized into the six carbon atoms [20][21][22]. However, if the diene contains free radical stabilizing groups, this mechanism could have significant contributions from other mechanisms that involve radical species [13,16,20,[23][24][25][26][27]. Detailed discussions about Cope rearrangements can be found in several studies and reviews that have been published previously [20, [28][29][30][31][32]. The configuration of elemanes formed via a Cope rearrangement from germacranolides only depends on the configuration of the most stable germacrane conformer since it is mainly a concerted reaction [15,18,33]. It is accepted that the conformers that normally carry out a Cope rearrangement are the ones that have crossed double bonds, as they can generate a chair TS. The configuration of the final elemanolide is also affected by the substituents in the germacranolides, the pseudo-equatorial position is preferred over the pseudo-axial position [5,34,35]. These are the factors that dictate that specific germancranes will only rearrange to yield one or potentially two elemanolide configurations.</p><p>The schkuriolide (1, Scheme 1) is a sesquiterpene lactone, specifically a (Z,E)-germacranolide, named melampolide, that coexists in the same natural source with the elemanschkuhriolide (3), which is an elemanolide with a stereochemistry structurally similar to 1 (C 14α H 5β ). In order to know if 1 and 3 have biogenetic relation, 1 was transformed into 3 by heating 1 for 10 minutes at 200 °C. This suggests that 1 is a biogenetic precursor of 3 [36]. It is important to mention that 1 suffers a hemiacetalization in addition to a Cope rearrangement to form 3. The non-hemiacetaled compound 3 was found in neither the natural source nor the products of the biomimetic transformation of 1 into 3. This transformation is very interesting since in order to explain the stereochemistry of elemane 3, a boat-like TS is necessary (path M, Scheme 1) [36,37]. This is one of the few reported cases of elemane's biogenetic formations where a boat TS can be proposed instead of the normal chair TS [34,[36][37][38][39][40][41]. In a second proposed mechanism for the transformation of 1 into 3, the (Z,E)-germacranolide isomerizes into (E,E)-germacranolide and in a second step a Cope rearrangement forms the elemane. In this case a normal chair TS is proposed to generate the correct elemane configuration (path N, Scheme 1) [37,[39][40][41]. It is possible that an enzyme is responsible to allow reactions that happen in the flask at very high temperatures in two ways, stabilizing the transition state, or destabilizing the ground states energy of the reactants. An antibody-catalyzed oxy-Cope reaction has already been described [42] as well as a proposed reaction mechanism [43]. In the study presented in this paper, we performed density functional theory (DFT) calculations of the possible mechanisms for the transformation of 1 into 3, to elucidate which mechanism is more likely and to determine if the Cope TS with a boat conformation during the transformation is energetically favorable. The study will also help to under-stand the structural factors that determine the energetic evolution of germacranolides' Cope transformations. Scheme 1: Biogenetic hypothesis for the transformation of schkuhriolide (1) into elemanschkuriolide (3).</p><!><p>DFT has been proved to be a good method for the study of reaction mechanisms of natural products' biogenesis and it has been used in many studies [44][45][46][47][48][49][50][51][52][53][54][55] and it is the method of choice for pericyclic reactions studies [20,56]. In particular, third generation hybrid functionals have improved the description of the potential energy surface and produce very reliable results [57][58][59][60]. Our studies in terpene biogenesis show that these hybrid functionals competes successfully with others in the determination of the energetic profile of reaction coordinates [52]. The third-generation hybrid functional improves the description of the energetic barriers with respect to the popular B3LYP functional [61]. Moreover, the B3LYP functional was used in a Cope rearrangement study of several germacranes and it was unable to obtain accurate results when the energy differences between germacranes and elemanes were small [62].</p><p>All calculations were performed with Gaussian 09 [63]. All the geometries were fully optimized using the DFT hybrid method M06x [57], a functional that is very reliability in calculations of activation energies [59,61]. The 6-31+G(d,p) basis set was used for all calculations. Diffuse functions in double split valence basis have shown to be more important than a triplet split of the valence basis when reaction energies and activation energies are calculated with DFT [64]. The stability of the wave functions of all the transition states was checked. An unrestricted wave function was used to calculate the activation energy of the cis/trans isomerization of the (Z,E)-germacranolide. All energies were reported with zero-point energy corrections and all TS geometries have only one imaginary frequency.</p><!><p>Besides the two previously proposed mechanisms (Scheme 1), there are two other possible mechanisms for the transformation of 1 into 3. It is also likely that the hemiacetalization occurs before the Cope rearrangement. Figure 1 shows the reaction coordinate of these four mechanisms. In the first proposed mechanism (path M, Figure 1) a conformational transformation of 1 must occur first. The most stable conformer has chair-boat conformation that according to Samek nomenclature is [ 15 D 5 , 1 D 14 ] (1a). This conformer is the one that is present in solution [37]. Nevertheless, conformer 1a does not have the proper geometry to directly generate the correct stereochemistry of 3. Both C-C bonds next to the C10-C1 double bond of conformer 1a have to rotate to generate the boat-boat conformer (1b, [ 15 D 5 , 1 D 14 ]), which is 3.5 kcal/mol less stable than 1a, but conformer 1b has the proper conformation to generate the configuration of 3 (ground state destabilization). The second step is the Cope rearrangement. The saddle point (TS1b-2) for this process has a high relative energy (47.0 kcal/mol). Thus, the transformation of 1 into 2 through TS1b-2 is unlikely at 200 °C (temperature at which the biogenetic transformation was performed) [36]. In case of path N, the activation energy of 4 to reach the Cope TS (TS4-2) is 25.9 kcal/mol, and the relative energy of TS4-2 is 35.0 kcal/mol. The chair TS was, as expected, less energetic than the boat TS. However, before the Cope rearrangement can proceed, the (Z,E)-germacranolide 1, must isomerize to the corresponding (E,E)-germacranolide 4. This process is highly unfavorable, its energetic barrier is about 55.7 kcal/mol, which is very close to the reported activation energies for the ethylene thermal isomerization (≈65 kcal/mol) [65][66][67][68]. Therefore, this high energy TS makes path N and path P unlikely. It is important to point out that in nature this isomerization of germacranes can be catalyzed by different mechanism. For example, other cis/trans transformations have been biomimetically cata-lyzed by SeO 2 [8,[69][70][71]. The only remaining route for the thermal transformation of 1 into 3 is path O. In this path, the hemiacetalization is the first step. We used a water molecule to facilitate the proton transfer in this stage. In the experiment, a hydroxylic group of other proximate germacranolide molecule or an actual water molecule could participate as donor and acceptor of the germacranolide proton. In fact, in the solid state 1 cocrystalizes with a water molecule [72]. The next step in this mechanism is the Cope rearrangement which have a TS (TS5-3) less energetic than the Cope TS without hemiacetal group (TS1b-2). This could be because the hemiacetalization reduces the transannular distance between C10 and C5 (3.20 Å, 2.91 Å, 2.72 Å and 2.66 Å for 1b, 4, 5 and 6, respectively) that facilitates orbital interactions and bond formation. The chair TS (TS6-3) is still more stable, the energy difference between TS5-3 and TS6-3 is almost the same than in TS1b-2 and TS4-2, but the energy of boat TS (TS5-3) decreases to 38.0 kcal/mol, which is small enough to be overcome at 200 °C. Thus, hemiacetalization lowers the activation energy of the boat Cope TS which allows the reaction to be completed at a temperature significantly lower than the temperature that a standard boat TS would need (≈260 °C) [10].</p><p>The hemiacetalization also allows the transformation of a (Z,E)germacranolide 1 into a elemanolide 3. This is an exception because all the biomimetical transformation of germacranolides into elemanolides reported until now are of (E,E)-germacranolides [41,[73][74][75]. Figure 1 shows that the elemanolide 2 is less stable than the (Z,E)-germacranolide 1, so it is not possible to obtain 2 from 1 without a transformation that stabilizes 2. In this special case, the hemiazetalization significantly lowers the energy of the elemanolide what makes the global process spontaneous. In fact, previous studies show that the transformation of (Z,E)-germacranolides with a blocked C6 hydroxy group do not produce the corresponding elemanolide (Scheme 2) [35]. Contrary, elemane (2) is more stable than the (E,E)-germacranolide. Therefore, a elemanolide can be formed directly from a (E,E)-germacranolide. Moreover, the (E,E)-germacranolide 4 is even less stable than 1 (9.1 kcal/mol), which explains the lack of published cases for a transformation of an (E,E)-germacranolide into a (Z,E)-germacranolide, but in the opposite direction there are some examples [37,69,76].</p><p>The hemiacetalization by itself does not guarantee the stabilization of an elemane. If compound 1 had to suffer a normal Cope (chair TS), it would generate the C5 epimer of 2 (2', Figure 2). This epimer is 2.6 kcal/mol less stable than 1a, so the formation of 2' from 1, as in case of 2, is thermodynamically forbidden. Epimer 2' can also produce a hemiacetal (3') but this compound has a higher energy than 3 by 4.2 kcal/mol; this is due to the C5 propenyl group in 3' is axially oriented instead of Scheme 2: Similar compounds to melampolide 1 unable to be hemiacetaled.</p><p>equatorially as in 3. In contrast to 3, the formation of epimer 3' from 1 is not thermodynamically highly favored. Therefore, the hemiacetal formation with the right orientation is fundamental to produce an elemanolide more stable than the (Z,E)-germacranolide. It has been proposed that the configuration of an elemane depends on the most stable conformation of the germacradiene from which it is derived [33,41], although this is not a general rule and, in some cases, a conformer with higher energy is the conformer that reacts. To explain this behavior, some authors have proposed that the least energetic conformation of the Cope TS is what determines the elemane configuration [2,17,77]. However, any of these arguments can explain the configuration of 3. Compound 3 has neither the configuration of the most stable conformer of 1 (1a) nor the configuration of the least energetic conformation of the Cope TS (a chair TS that would generate 2'). Compound 3 comes from conformer 1b which is not the most stable one and from a boat TS that is not the least energetic TS. Then, why does compound 3 have this configuration? The answer is simple, although not obvious; the elemanolide 2 has the right configuration to allow a hemiacetalization that reduced its energy via the formation of a significantly more stable hemiacetaled elemanolide 3. The configuration of 3 is the most stable among all other possible configurations. Therefore, the energy of the different elemane configurations and their possible subsequent rearrangement reactions should also be considered when a prediction of the configuration of an elemane is determined before a Cope rearrangement.</p><p>Finally, we studied the role the γ-lactone ring plays in the transformation of germacranes into elemanes. Takeda et al. carried out a series of experiments, which proved that γ-lactone rings prevent the Cope rearrangement of (Z,E)-germacranolides when the ring is closed but not when the ring is open [41]. A study of the Cope rearrangement in open ring (Z,E)-germacranolide 1 (7, Figure 3) and (E,E)-germacranolide 4 (8, Figure 3) was done in order to analyze the effect of the lactone ring. Figure 3 shows that energetic differences between cis and trans isomers do not vary significantly. When the lactone ring is closed, this differ-ence is 9.1 kcal/mol, and when it is open it is 8.9 kcal/mol, so the opening of the lactone ring does not affect in any way the relative stability of the isomers. Another possible explanation of the inhibition of Cope rearrangement by the lactone ring (proposed by Takeda) is that the lactone ring raises the Cope TS energy, as the lactone ring strains the germacrane ring. Contrary to what Takeda predicted, the relative energies of opened lactones, TS7-9 (50.8 kcal/mol) and TS8-9 (39.3 kcal/mol), are higher in comparison with their respectively closed lactone, TS1-2 (47.0 kcal/mol) and TS4-2 (35.0 kcal/mol). Elemanolide 9, product from Cope rearrangements of 7 and 8, has a closer energy to (Z,E)-germacranolide than its closed ring analog, 2. The difference between 1 and 2 is 4.5 kcal/mol while between 7 and 9 is only 2.1 kcal/mol. Thus, the lactone ring destabilizes the elemanolide, which explains why closed ring (Z,E)-germacranolides cannot carry out a Cope rearrangement contrary to the open ring (Z,E)-germacranolides. Again, the relative stability of an elemane versus that of a germacrane determines the likelihood of the transformation. The conclusion is that the γ-lactone ring in the Cope rearrangement destabilizes the corresponding elemane and it has a little or no effect in the Cope TS. This conclusion could be extrapolated to any 5-membered or smaller rings. The smaller a ring is the more susceptible it is to the strain generated by a second ring (5-members or smaller) fused to it. Therefore, the impact of the γ-lactone ring on the elemane ring (6-members) is significantly more than on the Cope TS ring (10-members).</p><!><p>The Cope rearrangement is commonly used to determine the germacrane conformation in solution, since the specialized literature establishes that the elemane configuration is due to the most stable conformer of germacrane. However, this is not always true as in the case studied here, where the product observed has neither the configuration of the most stable conformer nor the configuration of the least energetic conformation of the Cope TS. The configuration of elemane 3 is the most stable configuration of this compound. Therefore, it is also important to consider the energy of the different configurations of an elemane to correctly predict the conformation of a germacrene.</p><p>Interestingly, the (Z,E)-germacranes are significantly more stable than (E,E)-germacranes. Then, cis/trans isomerization can only happen in one way, (E,E)-germacranes → (Z,E)germacranes. Moreover, this isomerization cannot be thermally activated because of the high energy of the associated TS. (Z,E)-Germacranolides are also more stable than elemanolides, (Z,E)-germacranolides cannot transform into elemanolides unless there is a subsequent reaction that reduces the energy of the elemanolide, like the hemiacetalization does in the case of 3. The transformation studied herein is possible due to a previous hemiacetalization that reduces the energy of the boat transition state by enforcing a shorter distance between the atoms that form the new C-C bond. Moreover, the transformation (Z,E)germacranolide → elemanolide is only possible when the lactone ring is open, since in this case the elemanolide is more stable than the (Z,E)-germacranolide. Contrary to what other authors have proposed, the inability of (Z,E)-germacranes to transform into elemanes via a Cope rearrangement when (Z,E)germacranes have small rings (lactone) fused is not due to an increase of the activation energy of the Cope rearrangement. The activation energy does not change significantly when the fused ring is open or closed, but when the ring is open, the elemane is more stable. A fused small ring produces a lot of strain in the elemanes. In summary, fused small rings increase significantly the energy of elemanes, but those rings do not significantly modify the energy of germacrane's Cope TSs.</p>

Beilstein

3′-NADP and 3′-NAADP, Two Metabolites Formed by the Bacterial Type III Effector AvrRxo1*♦

An arsenal of effector proteins is injected by bacterial pathogens into the host cell or its vicinity to increase virulence. The commonly used top-down approaches inferring the toxic mechanism of individual effector proteins from the host's phenotype are often impeded by multiple targets of different effectors as well as by their pleiotropic effects. Here we describe our bottom-up approach, showing that the bacterial type III effector AvrRxo1 of plant pathogens is an authentic phosphotransferase that produces two novel metabolites by phosphorylating nicotinamide/nicotinic acid adenine dinucleotide at the adenosine 3′-hydroxyl group. Both products of AvrRxo1, 3′-NADP and 3′-nicotinic acid adenine dinucleotide phosphate (3′-NAADP), are substantially different from the ubiquitous co-enzyme 2′-NADP and the calcium mobilizer 2′-NAADP. Interestingly, 3′-NADP and 3′-NAADP have previously been used as inhibitors or signaling molecules but were regarded as “artificial” compounds so far. Our findings now necessitate a shift in thinking about the biological importance of 3′-phosphorylated NAD derivatives.

3′-nadp_and_3′-naadp,_two_metabolites_formed_by_the_bacterial_type_iii_effector_avrrxo1*♦

Introduction<!>AvrRxo1 Shows Phosphotransferase Activity<!><!>AvrRxo1 Shows Phosphotransferase Activity<!><!>AvrRxo1 Shows Phosphotransferase Activity<!>AvrRxo1 Expression Leads to Accumulation of Phosphorylated NAD in E. coli<!><!>AvrRxo1 Catalyzes the Phosphorylation of the 3′-Hydroxyl Group of the Adenosine Moiety of NAD<!>ADP Production by AvrRxo1 Is Stimulated by NAD and NADH<!><!>ADP Production by AvrRxo1 Is Stimulated by NAD and NADH<!>AvrRxo1 Phosphorylates NAD and NAAD with Similar Efficiency<!><!>AvrRxo1 Phosphorylates NAD and NAAD with Similar Efficiency<!>AvrRxo2 Is a Potent, Mixed Inhibitor of AvrRxo1<!><!>Discussion<!>Cloning, Expression, and Purification of AvrRxo1 and AvrRxo2 Constructs<!><!>Cloning, Expression, and Purification of AvrRxo1 and AvrRxo2 Constructs<!>Phosphotransferase Assays<!>Detection of AvrRxo1 Products in Small Metabolite Extracts<!>In Vitro Preparation and Purification of AvrRxo1 Products<!>Purification of 3′-NADP from Small Metabolite Extracts<!>Electrospray Ionization Mass Spectrometry of AvrRxo1 Products<!>d-Glucose-6-phosphate Dehydrogenase Assays of 3′-NADP<!>NMR Characterization of 3′-NADP<!>Spectrophotometric Substrate Comparison Assays<!>Michaelis-Menten Kinetic Measurements of AvrRxo1 and AvrRxo1/2<!>Author Contributions<!>

<p>Bacterial pathogens cause a multitude of severe human, animal, and plant diseases. The vast majority of these bacterial pathogens rely on sophisticated secretion systems by which they either secrete so-called effector molecules into the vicinity of the host cell or translocate them directly into the host cell cytoplasm. All effector molecules, particularly those that are translocated into the host cell cytoplasm, support bacterial invasion, colonization, and proliferation inside the host. They do so by interfering with or remodeling of important host cell processes as for example disguising the invader from the host's immune system (1, 2). For the latter, plant and animal pathogens face entirely different preconditions.</p><p>In contrast to animals, plants possess neither a somatic adaptive immune system nor mobile defender cells. Instead, each individual plant cell relies on its innate immunity, enabling it to initiate an appropriate response when infected (3–5). Defense responses of plants against invading microbes are induced at two levels and can eventually culminate in a hypersensitive response (HR)2 during which affected cells undergo programmed cell death to prevent systemic spread of the pathogen (6). On the first level, bacterial elicitor active epitopes are recognized by pattern recognition receptors. These elicitor molecules are thus termed pathogen/microbe-associated molecular patterns. Their recognition by specific receptors initializes pathogen-associated molecular pattern-triggered immunity (PTI) (4, 7). Consequently, pathogens evolved effector molecules able to suppress PTI by the host plant and thereby increase microbial pathogenicity (3, 8–10). This elaborate bacterial countermove leads to effector-triggered susceptibility (ETS) of the host and re-establishes microbe pathogenicity.</p><p>Certain plant cultivars are capable of overcoming ETS by the second level of defense in which plant resistance proteins specifically recognize one or multiple effectors. Thereby, plants circumvent the pathogen's abrogation of the PTI response and initiate a cellular defense program, leading to effector-triggered immunity (ETI) (10, 11). Effectors recognized by resistance proteins are thus termed avirulence (Avr) proteins.</p><p>Most microbial type III effectors (virulent and/or avirulent) are, however, challenging to study because they are unrelated at their sequence or structural level (12–14), making any a priori prediction of their functional mechanisms difficult. In addition, most pathogens secrete a number of different effectors, thereby targeting different pathways of their host cell simultaneously (15). Deducing a specific effector function from the observed infection phenotype is consequently almost impossible. Structural studies of these multifaceted effectors have therefore paved the way for follow-up studies focusing on the identification of their biological targets in the host. Such successful pioneering studies were for instance the identification of E3 ubiquitin ligase domains of AvrPtoB and XopL (16, 17) or the inhibitory effect of AvrPto on the Pto kinase activity deduced from the AvrPto-Pto complex structure (18). By a similar approach, the structural homology of the effector protein AvrRxo1 from the pathogen Xanthomonas oryzae pv. oryzicola to nucleotide kinases led to the recent proposal that AvrRxo1 contains a polynucleotide kinase domain with an unknown toxic mechanism in plants (19).</p><p>AvrRxo1 is a type III effector that is highly conserved in various Asian X. oryzae pv. oryzicola strains (20, 21). It was originally identified as a gene product of this particular pathogen that elicits a non-host HR in maize lines harboring the Rxo1 resistance gene (20). Interestingly, Rxo1 was also shown to act as a resistance protein in maize against infection by Paraburkholderia andropogonis (20), a pathogen encoding a protein highly homologous to AvrRxo1 (NCBI entry ALF40614). AvrRxo1 and Rxo1 are therefore most likely involved in a gene-for-gene relationship in certain maize cultivars in which the AvrRxo1 effector is recognized by an ETI program. In contrast, no single gene resistance against X. oryzae pv. oryzicola has been detected in rice (20). Hence, AvrRxo1 most likely constitutes a virulence factor that elicits ETS, and the rice Rxo1 homolog does not confer immunity (22). The role of AvrRxo1 as a virulence factor is further supported by the finding that it is toxic when expressed in tobacco and yeast cells (20, 21, 23). Furthermore, when ectopically expressed in Escherichia coli, AvrRxo1 was shown to inhibit cell growth, a phenotype that could be suppressed by co-expression of AvrRxo2 (19).</p><p>The AvrRxo2 open reading frame (ORF) is part of a bicistronic operon and is found downstream of the AvrRxo1 ORF in X. oryzae pv. oryzicola (20). Inhibition of the bacteriostatic phenotype is most likely accomplished by complex formation as inferred from the extensive interaction interface between AvrRxo1 and AvrRxo2 observed in the crystal structure (19). Given that X. oryzae pv. oryzicola has emerged as a prevalent pathogen that causes rice bacterial leaf streak disease, impairing the production of this staple crop in much of Asia, parts of Africa, and Australia (24), investigating the AvrRxo1/AvrRxo2 system as a major contributor to the pathogen's virulence is overdue.</p><p>Here, we describe our bottom-up approach used to identify the effector function of AvrRxo1. We show that AvrRxo1 is a hitherto unknown type of nucleotide kinase that catalyzes the formation of 3′-NADP and 3′-NAADP, two novel compounds that might interfere with conventional NAD(H)/2′-NADP(H)-dependent pathways and host cell Ca2+ signaling. In addition to revealing the enzymatic function of AvrRxo1, we show that the associated, chaperone-like protein AvrRxo2 acts as a highly potent inhibitor of AvrRxo1.</p><!><p>AvrRxo1 from X. oryzae pv. oryzicola is a multidomain protein consisting of a central potential kinase domain and an N-terminal domain that has been suggested to contain a thiol protease active site (20, 21). Whereas the kinase domain is conserved among AvrRxo1 homologs from different plant pathogens (Fig. 1), the N-terminal region is highly divergent, and the potential thiol protease active site is only found in a few Xanthomonas strains. We therefore exclusively used the truncated variant AvrRxo1ΔN88 that lacks the divergent N terminus in our experiments and will refer to this as AvrRxo1 throughout.</p><!><p>Structure-based sequence alignment of different AvrRxo1 homologs with zeta toxins. The C-terminal kinase domain of AvrRxo1 is conserved among X. oryzae pv. oryzicola (NCBI entry WP_014504815.1) and other plant pathogens including Acidovorax citrulli (NCBI entry AIE45656.1) and P. andropogonis (NCBI entry ALF40614.1). The primary structure of PezT from Streptococcus pneumoniae (NCBI entry WP_000405360.1), a representative UNAG-3P kinase belonging to the ζ toxin family, is given. Residues are colored according to their conservation from dark green (high conservation) to orange (low conservation). Secondary structure elements are depicted with cylinders (α-helices), arrows (β-strands), and gray lines (loop regions). Circles between the sequences indicate residues involved in ATP binding. Triangles mark residues of ζ kinases that coordinate UNAG. Other important structural features are highlighted in boxes. Note that ATP binding residues, but not UNAG binding residues, are conserved between AvrRxo1 and the ζ toxin PezT.</p><!><p>Apart from the structural homology of this central AvrRxo1 domain with polynucleotide kinases, the recently reported 3D structure of AvrRxo1 revealed a strong homology of AvrRxo1 with UDP-N-acetylglucosamine (UNAG) kinases of the ϵ-ζ toxin-antitoxin family (19). This raised the question whether AvrRxo1 might be an authentic UNAG kinase or not as the previously identified, characteristic UNAG binding motif of ζ kinases (25) cannot be identified in the effector's primary structure (Fig. 1). In fact, neither we nor others (19) could show that AvrRxo1 catalyzes the transfer of the γ-phosphate group of ATP onto UNAG to a significant extent because only minor traces of a potential UNAG-3P species were detected when a reaction mixture of AvrRxo1 incubated with UNAG and ATP for 5 h was analyzed using anion exchange chromatography (Fig. 2a).</p><!><p>AvrRxo1 phosphorylates adenine- but not guanine-containing nucleotides. Different nucleotides (each at 500 μm) were tested as potential donor and acceptor substrates (gray bars) of AvrRxo1. After 5 h of incubation at 25 °C, products were detected by separation on an anion exchange chromatography column. Products of the reactions are marked with green bars. Traces shown are of A260 (black), A280 (blue), A340 (red), and eluate conductivity (dotted gray line). A, incubation of AvrRxo1 with UNAG and ATP does not lead to significant amounts of phosphorylated UNAG but results in accumulation of ADP and two apparently phosphorylated ADP and ATP species (pADP and pATP). B, similarly, incubation of AvrRxo1 with ATP alone results in accumulation of ADP and phosphorylated ADP and ATP species. C, when incubated with GTP alone, neither GDP nor any other nucleotide species is formed. The detected GDP species was identified as an impurity of the GTP stock and is thus marked with an asterisk. D, in contrast, incubation of AvrRxo1 with equimolar amounts of GTP and ADP causes accumulation of GDP and phosphorylated ADP. AvrRxo1 thus utilizes GTP as a phosphate donor but accepts neither GDP nor GTP as phosphate acceptors in contrast to ADP. E, when incubated with ATP and NAD, nearly all NAD becomes phosphorylated (pNAD) by AvrRxo1, whereas phosphorylation of ADP and ATP is virtually absent. F, chasing equal amounts of NAD and NADH as phosphate acceptor candidates reveals that AvrRxo1 favors NAD under the experimental conditions. Note that any minor traces of ATP used in reaction F would co-elute with phosphorylated NADH (pNADH) and are therefore not indicated. AU, absorbance units.</p><!><p>In contrast, we observed a significant accumulation of ADP and two new, apparently phosphorylated species, indicating that AvrRxo1 is a nucleotide kinase with hitherto uncharacterized substrate specificity (Fig. 2a). Formation of these new species was not dependent on the presence of UNAG as incubation of AvrRxo1 with ATP alone resulted in the accumulation of products with identical retention times (Fig. 2b). Determination of their molecular masses suggested that AvrRxo1 catalyzed the phosphorylation of ATP (m/zobs = 586.0) and ADP (m/zobs = 506.0) in these long term incubation assays. This activity could be assigned to AvrRxo1 because neither of these products accumulated when the catalytically impaired AvrRxo1(D193N) variant was incubated with ATP (supplemental Fig. S1).</p><p>Because AvrRxo1 harbors the conventional P-loop motif (26), we wondered whether the enzyme also utilizes guanine nucleotides as substrates. In contrast to ATP, however, neither GDP nor any new, potentially phosphorylated GTP/GDP species accumulated when AvrRxo1 was incubated with GTP alone (Fig. 2c). We then scrutinized substrate binding site specificities by incubating AvrRxo1 with equimolar amounts of GTP and ADP and found that ADP was phosphorylated under the consumption of GTP (Fig. 2d). The phosphate donor site of AvrRxo1 thus binds ATP as well as GTP, whereas the phosphate acceptor binding site is selective for adenine-containing nucleotides rather than guanine nucleotides.</p><p>Because we observed significant amounts of residual ATP even after 5 h of incubation at 25 °C with ATP alone (Fig. 2b), we searched for better, adenine-containing substrates. We found that nicotinamide adenine dinucleotide (NAD) was quantitatively phosphorylated by AvrRxo1 after 5 h of incubation, whereas only minor traces of the phosphorylated ADP species were detected (Fig. 2e). NAD must consequently be the superior substrate when compared with ATP or ADP. We next tested whether the oxidation state of the nicotinamide moiety influences phosphorylation efficiency and incubated AvrRxo1 with equimolar amounts of NAD and NADH and reaction-limiting concentrations of ATP. Nearly all NAD was found to be phosphorylated by AvrRxo1, whereas only a minor fraction of NADH was phosphorylated (Fig. 2f). This apparent selection between NAD and NADH by the acceptor substrate binding site under the experimental conditions strongly argues for NAD as a substrate being phosphorylated in vivo.</p><!><p>The observed preferential phosphorylation of NAD in vitro prompted the question whether NAD is also the preferred substrate of AvrRxo1 in vivo. We therefore prepared small metabolite extracts (SMEs) from E. coli cells expressing the AvrRxo1 protein and separated them using anion exchange chromatography.</p><p>We found that cells expressing active AvrRxo1 contained a significant amount of a single nucleotide species with similar retention time as the phosphorylated NAD standard produced in vitro. In contrast, no such species could be detected in SMEs prepared from untransformed or AvrRxo1(D193N)-expressing cells (Fig. 3). To further characterize this nucleotide species, we isolated it in a semipreparative manner from large scale SMEs. Electrospray ionization mass spectrometry revealed identical masses for the accumulated small metabolite and the in vitro produced, phosphorylated NAD (m/z = 742.1) (supplemental Fig. S2), and collision-induced dissociation gave comparable fragmentation patterns (Fig. 4, a and b). Additionally, the obtained fragments allowed assigning the site of phosphorylation to the hydroxyl groups of the adenosine moiety.</p><!><p>E. coli cells expressing AvrRxo1 accumulate a phosphorylated NAD species co-eluting with the enzymatic product obtained from in vitro reactions. An anion exchange chromatogram shows A260 traces obtained by separation of small metabolite extracts prepared from E. coli cells. Traces shown are from extracts of untransformed cells (black), cells expressing AvrRxo1 (green), or cells expressing the catalytically impaired AvrRxo1(D193N) variant (gray). The phosphorylated NAD species (blue) prepared from in vitro reactions of AvrRxo1, ATP, and NAD was used as an authentic standard at 50 μm and co-elutes with the nucleotide species accumulating in vivo upon AvrRxo1 expression (inset). Note that at retention times 5.5 and 8 min two additional species accumulate in extracts from both the catalytically impaired AvrRxo1(D193N) and the wild type AvrRxo1 protein. Most likely, these are expression artifacts and not related to AvrRxo1 catalysis.</p><p>AvrRxo1 catalyzes the formation of 3′-phosphorylated NAD. Purified reaction products of AvrRxo1 in vitro reactions and small metabolite extracts were used for characterization by ESI MS-MS, a spectrophotometric assay using the 2′-NADP-dependent enzyme G6P-DH, and NMR spectroscopy. A–C, ESI MS-MS spectra were obtained by selecting the 742.1/743.1-Da heavy, phosphorylated NAD/NAAD species (blue diamonds) and fragmenting them at −45 and −40 eV, respectively. A, ESI MS-MS spectrum of phosphorylated NAD purified from in vitro reactions. The detected fragments correspond to those obtained by fragmenting phosphorylated NAD from SMEs (B) and in vitro phosphorylated NAAD (C). Fragments detected can be ascribed to phosphorylated ADP-ribose (620.0 Da), ADP-ribose (540.1 Da), ATP/phosphorylated ADP (506.0 Da), and ADP (408.0 Da). For a structural representation of the obtained fragments, see supplemental Fig. S3. D, using the 2′-NADP-dependent enzyme G6P-DH, the phosphorylated products of AvrRxo1 were shown to be differently phosphorylated than conventional 2′-NADP. Neither addition of 200 μm in vitro phosphorylated NAD (blue) nor addition of the in vivo product to 200 μm (green) could stimulate G6P-DH activity. In contrast, 200 μm 2′-NADP were rapidly reduced to 2′-NADPH by the enzyme as shown by the increase in absorbance at 340 nm. Hence, AvrRxo1 produces a compound different from 2′-NADP but with identical mass. E, NMR characterization of in vitro phosphorylated NAD showed that the phosphate group transferred by AvrRxo1 is situated at the 3′-hydroxyl group of adenosine (3′A). Although the spectrum for the 3′-proton signal of adenosine (3′A) showed increased peak multiplicity due to 1H-31P coupling in the 1H NMR experiment when compared with the 1H-31P decoupled spectra, the spectrum for the 2′-proton signal (2′A) remained unchanged in both experiments (5.0–4.8 ppm and highlighted in gray). The entire spectra are shown in supplemental Fig. S4.</p><!><p>Phosphorylation of NAD at the adenosine ribose moiety by AvrRxo1 could either lead to conventional 2′-NADP or a novel, small metabolite, 3′-NADP. We thus probed phosphorylated NAD purified from AvrRxo1-expressing E. coli cells and the in vitro counterpart for their capability to stimulate activity of the 2′-NADP-dependent enzyme d-glucose-6-phosphate dehydrogenase similarly as described before (27). To our surprise, neither of the two compounds could be reduced by the enzyme (Fig. 4d), thereby strongly suggesting that AvrRxo1 catalyzes the formation of 3′-NADP and not conventional 2′-NADP.</p><p>To directly prove that phosphorylation catalyzed by AvrRxo1 occurs at the 3′-hydroxyl group of the adenosine moiety, we isolated NAD phosphorylated by AvrRxo1 in vitro in a preparative setup and elucidated its structure by 1D and 2D NMR experiments. 31P experiments clearly confirmed the presence of a third phosphate group, but the recorded proton shifts differed significantly from those reported for 2′-NADP (28), implying that the 3′-hydroxyl group of the adenosine moiety is being phosphorylated by AvrRxo1 (Fig. 4e). Ultimately, 1H-31P decoupling experiments together with 1H-1H COSY (supplemental Fig. S4) confirmed that the phosphate group transferred onto the NAD molecule by AvrRxo1 is located at the 3′-hydroxyl group of the adenosine moiety.</p><!><p>To compare the kinetics of ADP production arising from 3′-NADP formation with those of ATP phosphorylation, we scrutinized AvrRxo1 activity using a quantitative, spectroscopic ATPase steady state assay (29). Briefly, ATP was regenerated from newly formed ADP by pyruvate kinase (PK) upon conversion of phosphoenolpyruvate (PEP) to pyruvate. The latter was converted to lactate by lactate dehydrogenase (LDH) upon oxidation of NADH. During all experiments, PK, LDH, and PEP were adjusted to concentrations that allow the background reaction of the detection system to proceed much faster than AvrRxo1, which was checked by ADP control titrations (Fig. 5). Because we knew from previous experiments that NADH, ADP, and ATP can be phosphorylated by AvrRxo1, the assay was performed as a burst/chase experiment. To this end, AvrRxo1 was incubated with either ATP and NAD or ATP alone for 15 min. NADH was then titrated into the reaction setup at equimolar concentrations, and the decrease in absorbance at 340 nm was monitored.</p><!><p>ADP formation by AvrRxo1 is stimulated by NAD and NAAD but not by the structurally similar dinucleotide UpA. A burst/chase experiment using the PK/LDH assay shows that conversion of ATP to ADP by AvrRxo1 is stimulated by NAD, NAAD, and NADH but not by the dinucleotide UpA. AvrRxo1 was added as indicated to a reaction mixture including PK, LDH, PEP, ATP, and a potential phosphate acceptor substrate or water. The reaction was incubated for 15 min to allow accumulation of pyruvate through ATP regeneration by PK from any newly formed ADP upon consumption of PEP. Subsequently, 200 μm NADH were titrated into the setup, and the absorbance at 340 nm was measured. Traces of A340 shown are of reactions in which the following compounds were present: H2O (black), NAD (green), NAAD (blue), UpA (dotted gray). The reaction setup with ATP, NAD, and no AvrRxo1 (dotted black) showed negligible activity. The maximum assay velocity was determined by incubation with 200 μm ADP (gray) for 15 min prior to NADH addition.</p><!><p>Because we identified NAD as a suitable substrate, we expected that incubation with ATP and NAD would lead to production of ADP prior to NADH addition. Hence, pyruvate would accumulate substantially and LDH would become the rate-limiting factor, causing a drastic decrease in NADH absorbance (burst). In contrast, incubation with ATP alone should yield minor traces of ADP, and significant amounts thereof are only formed after NADH addition to the assay. Here, AvrRxo1 becomes rate-limiting over LDH, and the decrease in absorbance is much slower.</p><p>As expected, this experimental setup also revealed that NAD is preferred over ATP (Fig. 5). Because AvrRxo1 was recently suggested to be a polynucleotide kinase based on structural homology (19), we used this assay to compare a short polynucleotide analog and NAD as potential substrates. To structurally be as homologous as possible to NAD, we tested uridine 5′ → 3′ adenine dinucleotide (UpA) but found that incubation with ATP and UpA did not lead to production of any detectable amounts of ADP prior to NADH addition (Fig. 5). Hence, polynucleotides are unlikely the authentic substrates of AvrRxo1.</p><!><p>Having confirmed that AvrRxo1 is a novel type of kinase that phosphorylates NAD in vitro and in vivo, we wondered whether it also accepts the biochemical precursor of NAD, NAAD, as an acceptor substrate. We thus tested NAAD as described in the previous section and found that it stimulates ADP production by AvrRxo1 similarly to NAD (Fig. 5). We therefore performed steady state ATP hydrolysis assays to characterize AvrRxo1 with regard to its kinetic parameters Km and kcat for ATP, NAD, and NAAD (Fig. 6 and Table 1).</p><!><p>AvrRxo1 is a highly efficient NAD/NAAD kinase. Steady state Michaelis-Menten kinetics using the PK/LDH assay revealed that the Km values of AvrRxo1 for NAD (green) and NAAD (blue) are identical at 1.2 ± 0.1 mm at saturating Mg2+-ATP concentrations (3 mm). Turnover by the enzyme is rapid with kcat values of 430 ± 10 s−1 for NAD and 270 ± 10 s−1 for NAAD. A comparable kcat for ATP (black) at saturating NAD concentrations (4 mm) of 460 ± 10 s−1 was determined. The Km for ATP is slightly lower as for NAD with a value of 1.0 ± 0.1 mm. S.E. of triplicates are given as bars for each point measured. The 95% confidence interval for each fit is indicated by dotted lines.</p><p>Kinetic properties of AvrRxo1</p><p>The values are derived from the fit of the arithmetic mean of independent triplicates, and their respective error within the 95% confidence interval is given. Kinetic parameters of ATP were determined at 4 mm NAD, and those of NAD and NAAD were determined at 3 mm Mg2+-ATP. The individual substrates for which the kinetic parameters were determined are given in bold characters, whereas the subscript indicates the substrate that was added at saturating conditions. App., apparent.</p><p>a These values were derived from single measurements.</p><!><p>For the phosphate donor ATP, we determined the Km under saturating NAD concentrations to be 1.0 ± 0.1 mm, whereas the apparent turnover rate (kobs) is 460 ± 10 s−1. Strikingly, the Km for NAD and NAAD at saturating ATP concentration was determined to be identical at 1.2 ± 0.1 mm, and the kcat values were 430 ± 10 s−1 for NAD and 270 ± 10 s−1 for NAAD. Hence, the specificity constants of AvrRxo1 for NAD and NAAD are comparable with values of 358 and 225 s−1 mm−1, respectively. AvrRxo1 therefore phosphorylates NAD and NAAD with similar efficiency. Which of the two compounds is phosphorylated in vivo consequently depends on the respective subcellular concentrations of these dinucleotides and the localization of the effector.</p><p>Furthermore, this reveals that AvrRxo1 is a highly potent NAD/NAAD kinase, especially when compared with the NAD kinase from E. coli that catalyzes the phosphorylation of NAD to 2′-NADP with a kcat of 125 s−1 and a Km of 2 mm (kcat/Km = 62.5 s−1 mm−1) (30). It is important to note that the NAD concentrations used in NAD titration assays had to be corrected by a constant amount of NAD produced by the PK/LDH system due to an ADP impurity of the ATP stock (1.9%). The increase in NAD and concomitant decrease in NADH concentrations were identical for all NAD titration experiments and were determined to be 56 μm. Intriguingly, the PK/LDH system ensured constant NAD and ATP concentrations throughout each measurement, i.e. as long as NADH was present, representing ideal steady state conditions. In addition, NADH contributed to the observed initial velocities of each measurement although it is phosphorylated to a much lesser extent than NAD (Fig. 2f). In a first approximation, we assumed this to be constant in all measurements and therefore included an additive, basal background reaction, kbasal, in the Michaelis-Menten equation (see "Experimental Procedures"). This basal background activity causes the fit not to intersect with the origin for NAD and NAAD titrations as would be the case for an unmodified Menten curve (Fig. 6).</p><!><p>Having revealed that AvrRxo1 is a potent 3′-NAD/NAAD kinase, we decided to investigate the effect of the second polypeptide encoded by the AvrRxo1/2 bicistron, AvrRxo2, on the kinase activity of AvrRxo1. Co-expression with small polypeptides from a bicistron is a common hallmark of bacterial type III effectors (31). The majority of these small proteins interact with a single, cognate effector and are involved in prevention of detrimental interactions between the effector and other proteins, effector stabilization, and regulation of effector secretion (31). Similarly, AvrRxo2 has been proposed to act as such a chaperone-like protein (20). Recent findings, however, are somewhat contradictory to a pure chaperone function of AvrRxo2 in that co-expression of AvrRxo2 and AvrRxo1 suppressed the bacteriostatic phenotype observed for E. coli cultures expressing only AvrRxo1 (19).</p><p>Using the spectroscopic ATPase assay, we found that the apparent kcat of AvrRxo1 for NAD was already strongly affected at equimolar concentrations of AvrRxo1 and AvrRxo2 (Fig. 7). Furthermore, also the Km was found to increase proportionally with AvrRxo2 concentration, showing that AvrRxo2 interferes as a mixed inhibitor with AvrRxo1 kinase activity. Both proteins might therefore constitute an effector/immunity pair (32) in Xanthomonas.</p><!><p>AvrRxo2 is a potent inhibitor of AvrRxo1. A reaction setup similar to that described for Fig. 6 was used to determine the effect of AvrRxo2 on AvrRxo1 kinase activity. A constant 2.5 nm AvrRxo1 together with 3 mm Mg2+-ATP was incubated with different concentrations of AvrRxo2: 0 (dark green diamonds), 2.5 (light green circles), 5.0 (light blue squares), 7.5 (dark blue triangles), and 12.5 nm (black crosses). NAD was titrated into the reactions, and the initial velocities of each reaction were determined and plotted. Equimolar concentrations of AvrRxo2 already have a dramatic effect on AvrRxo1 kinase activity, revealing AvrRxo2 as a potent inhibitor of the effector. Because both kcat and Km of AvrRxo1 are affected, AvrRxo2 acts as a mixed inhibitor.</p><!><p>The bacterial type III effector protein AvrRxo1 has recently been identified as an important virulence factor of the rice pathogen X. oryzae pv. oryzicola (20). Strikingly, it is also found in the genome of other pathogenic bacteria that infect crops from different plant families (Fig. 1). Despite recent insights into the role of AvrRxo1 in plant-pathogen interactions (20, 21), its enzymatic properties remained enigmatic so far. The recent structure-based speculation that AvrRxo1 might be a polynucleotide kinase (19) disagrees with our finding that no ADP formation by AvrRxo1 could be induced even in the presence of the short dinucleotide UpA. Instead, we showed that AvrRxo1 is a novel type of kinase that modifies the coenzyme NAD and its biochemical precursor NAAD (Figs. 4 and 5), whereas NADH is phosphorylated with much lower efficiency (Fig. 2f).</p><p>We verified that phosphorylation occurs at the adenosine 3′-hydroxyl group of NAD by NMR characterization of the enzymatic product (Fig. 4e). AvrRxo1 thus catalyzes the formation of two previously unknown natural compounds, 3′-NADP and 3′-NAADP, that are different from conventional 2′-NADP. We additionally provide strong evidence that AvrRxo1 also produces significant amounts of 3′-NADP in vivo.</p><p>The first reported chemical synthesis of 3′-NADP dates back to the mid-1950s when it was produced by acidic isomerization of 2′-NADP (27). Since then, 3′-NADP was tested as a potential synthetic inhibitor for a variety of NAD(P)-dependent enzymes (27, 33–35). In the last two decades, however, 3′-NADP was orphaned, probably due to its surmised artificial character. We now show that AvrRxo1 effectors secreted by plant pathogens produce high levels of this compound in living organisms, and thus 3′-NADP needs to be included in the league of natural compounds. Our studies open the possibility for further elaborated in vivo studies by enzymatically producing this compound upon AvrRxo1 expression.</p><p>Interestingly, 3′-NADP was shown to be a potent, non-competitive inhibitor of the maize NADP-dependent malic enzyme with a Kis of 56.7 μm (36). NADP-dependent malic enzymes play a central role in the carbon fixation process of globally important C4 plants by catalyzing the decarboxylation of malate to pyruvate and CO2 in bundle sheath cells (37). It is tempting to speculate that, after being injected into the host cell, AvrRxo1 might interfere with carbon fixation pathways in C4 plants and thereby establishes cytotoxicity. Consequently, certain cultivars of the C4 plant maize might have evolved an Rxo1 resistance protein that establishes ETI when exposed to AvrRxo1 (38). The finding that the maize pathogen P. andropogonis encodes an AvrRxo1 homolog suggests that this ETI program has evolved against infections with P. andropogonis rather than X. oryzae pv. oryzicola.</p><p>In contrast, the C3 plant rice suffers from leaf streak and apparently does not elicit an ETI once infected with X. oryzae pv. oryzicola, although it encodes an Rxo1 protein highly similar to that of maize (39). This might reflect that the selective pressure on C3 plants by inhibition of NADP malic enzymes is lower than on C4 plants.</p><p>Furthermore, our kinetic studies showed that the second potential product of AvrRxo1 catalysis is 3′-NAADP because phosphorylation of NAD and NAAD has similar in vitro kinetic properties. NAAD is a low abundance metabolite that is converted to NAD by NAD synthetase. This occurs in E. coli with high efficiency, causing extremely low cellular levels of NAAD when compared with NAD concentrations (40). Thus, we did not expect to detect significant amounts of 3′-NAADP in our SME analysis. Similar to 3′-NADP, 3′-NAADP is known since the mid-1970s as a synthetic compound produced by acidic isomerization of 2′-NAADP (41). The latter is an established and highly potent Ca2+-mobilizing second messenger in animals and plants (42–44). Interestingly, 3′-NAADP also is highly potent in mobilizing Ca2+ ions (45, 46). Similar to 3′-NADP, 3′-NAADP was orphaned due to its synthetic character and has not been further investigated as a potential Ca2+-mobilizing second messenger. Based on our findings, however, this might need some reconsideration.</p><p>Nevertheless, to what extent 3′-NAADP is formed after translocation of AvrRxo1 into the plant cell cytosol needs to be determined as the cellular NAAD concentrations are low. However, even small amounts of 3′-NAADP should suffice to interfere with the intracellular "Ca2+ signature" because the half-maximal effective concentration of 3′-NAADP for triggering a Ca2+ release in sea urchin eggs was shown to be approximately 300 nm (46). The finding that AvrRxo1 catalyzes the formation of the Ca2+ releaser 3′-NAADP is particularly intriguing because Ca2+ signaling is a highly sensitive, universal signal transduction mechanism in eukaryotes (47, 48), and Ca2+ influx from the apoplast is a characteristic process in plant cells initiating an HR against invading pathogens (49, 50).</p><p>It thus seems plausible that 3′-NAADP produced by AvrRxo1 interferes with host cell immune responses by manipulating the intracellular Ca2+ signature. Further substantiating this hypothesis, the non-host HR of tobacco against X. oryzae pv. oryzae could be suppressed by transforming the bacteria with an AvrRxo1-encoding plasmid (21). A curious side note is that structural studies on the type III effector XopQ from X. oryzae pv. oryzae suggested a nucleoside hydrolase activity on the Ca2+-mobilizing second messenger cyclic adenosine diphosphate ribose (51). Hence, manipulation of the host cell Ca2+ signature might not only be performed by AvrRxo1 but could be a common approach of plant pathogens.</p><p>We have exemplarily discussed potential scenarios how 3′-NADP and 3′-NAADP production by AvrRxo1 could enhance virulence of plant pathogens. However, effector molecules do not necessarily promote pathogenicity by targeting a single cellular pathway but frequently do so by simultaneously interfering with multiple cellular processes (1). Both products of AvrRxo1 might thus also interfere with a vast array of other, pivotal biological processes requiring NAD and its derivatives as co-enzymes. Such processes are for instance anabolic and catabolic pathways as well as regulation of gene expression by posttranslational modifications of proteins including mono/poly(ADP-ribosyl)ation (52) and histone deacetylation (53). Interestingly, NAD is also required for the synthesis of cyclic ADP-ribose, another highly potent Ca2+ releaser, by ADP-ribosylcyclases (54).</p><p>Revealing the enzymatic properties of AvrRxo1 enabled us to further characterize the function of AvrRxo2. It has previously been shown that co-expression of AvrRxo2 and AvrRxo1 cures the bacteriostatic phenotype observed when cells express AvrRxo1 alone (19). We showed that AvrRxo2 strongly inhibits AvrRxo1 nucleotide kinase activity, thereby most likely protecting the pathogen from cytotoxic effects due to AvrRxo1 expression, similar to common effector/immunity systems. We therefore assume that AvrRxo1 becomes liberated from AvrRxo2 just before translocation via the type III secretion system and consequently exerts its detrimental effects exclusively within the plant host cell.</p><p>In conclusion, we provide experimental evidence that AvrRxo1 is a novel type of kinase that phosphorylates NAD and NAAD at the 3′-hydroxyl groups of their adenosines. Both products of the effector, 3′-NADP and 3′-NAADP, which were so far assumed to be purely synthetic, are thus also found as natural compounds produced by plant pathogens. As bacterial effector proteins like AvrRxo1 have evolved to subversively manipulate critical host cell processes, the identification of their mechanism provides insights into conserved eukaryotic pathways and greatly contributes to advances in the field investigating host-pathogen interactions.</p><!><p>The construct pGEX4T-1_AvrRxo1ΔN88 encoding a truncated AvrRxo1 variant (henceforth referred to as AvrRxo1) from X. oryzae pv. oryzicola was generated by introduction of a BamHI restriction site into pGEX4T-1_AvrRxo1(ΔN65) (19) (kindly provided by B. Zhao, Virginia Tech) before the Thr-89 codon following the QuikChange protocol (Agilent Technologies, Santa Barbara, CA) with primer pair Avr1Bam_f/r (see Table 2 for all primers). The coding region for residues 66–88 was removed by religating the BamHI-digested plasmid. The BamHI restriction site was then removed by site-directed mutagenesis using primer pair Avr1BamX_f/r. The plasmid pGEX4T-1_AvrRxo1ΔN88 (D193N) encoding the catalytically impaired AvrRxo1 variant AvrRxo1 (D193N) was obtained by site-directed mutagenesis of pGEX4T-1_AvrRxo1ΔN88 using primer pair Avr1QC_D193N_f/r.</p><!><p>Primers used for cloning of AvrRxo1 and AvrRxo2 constructs</p><!><p>Two different translational start sites have been predicted for the ORF of AvrRxo2 from X. oryzae pv. oryzicola, giving two hypothetical AvroRxo2 polypeptide chains with either 117 amino acids (AA) (NCBI entry AEQ98134) or 98 AA (NCBI entry WP_041183485) in length. We purchased plasmid pMA(AEQ98134) containing a synthetic gene for the ORF of the 117-AA AvrRxo2 variant flanked at the 5′-end with a start codon overlapping the NcoI restriction site, a 3′-terminal coding sequence for a hexahistidine tag just before the stop codon, and a BamHI restriction site downstream thereof (Life TechnologiesTM, Thermo Fisher Scientific, Waltham, MA). The synthetic gene was excised by a restriction digest with NcoI and BamHI and ligated into pET28b (pET28b_AEQ98134). Expression and purification using pET28b_AEQ98134 revealed that both AvrRxo2 variants are equally expressed from this construct, indicating that a functional Shine-Dalgarno sequence and a start codon for the 98-AA AvrRxo2 variant are present (data not shown). We thus decided to exclusively use the 98 AA-encoding ORF (WP_041183485) in all our experiments and will refer to its product as AvrRxo2 similarly as described before (19). To this end, DNA for the AvrRxo2 coding sequence was PCR-amplified from pMA(AEQ98134) using primers Avr2_NdeI and M13_rev. Following an NdeI and BamHI restriction digest, the fragment was ligated into a pET24b vector (pET24b_AvrRxo2). This plasmid was used to generate pET24b_AvrRxo2(W97A) by site-directed mutagenesis using primer pair Avr2QC_W97A_f/r. The expression construct pET24b_AvrRxo2ΔN25 encoding an AvrRxo2 variant truncated by the first 25 residues and lacking the C-terminal His tag was obtained by PCR amplification of the encoding region from pET24b_AvrRxo2 plasmid DNA using primers Avr2_dN25_NdeI and Avr2_STOP_rev. Subsequently to a restriction digest with NdeI and NotI, the PCR product was ligated into an equivalently cut pET24b expression vector. The identity and correctness of all expression vectors were verified by sequencing of purified plasmid DNA.</p><p>For expression and purification of AvrRxo1 and AvrRxo2 constructs, E. coli BL21(DE3)-RIL cells were either transformed with vector pET24b_AvrRxo2 or pGEX4T-1_AvrRxo1ΔN88(D193N) alone or with vector pGEX4T-1_AvrRxo1(ΔN88) together with either pET24b_AvrRxo2(ΔN25) or pET24b_AvrRxo2(W97A). Cells were grown at 37 °C overnight on Luria-Broth (LB) medium agar plates containing 0.1 mg/ml ampicillin, 0.05 mg/ml kanamycin, or both in the case of co-expression. A liquid preculture of LB medium containing 0.1 mg/ml ampicillin and/or 0.05 mg/ml kanamycin was grown to exponential phase at 37 °C and used as inoculum for expression cultures. Final cell cultures for expression were grown to an A600 of 0.3–0.4 at 37 °C, cooled to 16 °C, and further incubated until reaching an A600 of 0.6. Protein expression was induced by addition of 0.5 mm isopropyl β-d-1-thiogalactopyranoside. Cells were harvested by centrifugation after overnight expression, and pellets of all expression constructs were suspended in cold buffer A (50 mm Tris-HCl, pH 8.5, 150 mm NaCl, 50 mm (NH4)2SO4, 5 mm β-mercaptoethanol (β-me)). All subsequent purification steps were performed at 4 °C or on ice. Cell walls were broken by sonication, and the lysate was cleared by centrifugation. The supernatant was further filtered through a 0.45-μm-cutoff syringe filter before being loaded onto the first column.</p><p>Because AvrRxo1 proved to be poorly soluble when expressed alone in E. coli, we co-expressed AvrRxo1 together with the N-terminally truncated AvrRxo2ΔN25 variant. This substantially improved AvrRxo1 solubility during expression and allowed removal of AvrRxo2ΔN25 due to its reduced affinity for AvrRxo1 compared with full-length AvrRxo2. To purify AvrRxo1, the filtered supernatant was loaded onto a 1-ml-column volume (CV) HisTrap Ni-NTA column (GE Healthcare) equilibrated with buffer A. The column was washed with 10 CVs of buffer B (50 mm Tris-HCl, pH 8.5, 150 mm NaCl, 50 mm (NH4)2SO4, 30 mm imidazole, 2 mm dithioerythritol (DTE)). The majority of AvrRxo2ΔN25 was in the flow-through. Bound protein was eluted in 5 CVs of buffer C (50 mm Tris-HCl, pH 8.5, 150 mm NaCl, 50 mm (NH4)2SO4, 200 mm imidazole, 2 mm DTE). The eluate was supplemented with 30 μg/ml TEV protease to remove the N-terminal GST and the C-terminal hexahistidine tags of AvrRxo1. The eluate was then dialyzed overnight against buffer D (50 mm Tris-HCl, pH 8.5, 100 mm NaCl, 10 mm (NH4)2SO4, 2 mm DTE). Precipitated protein was removed by centrifugation, and the supernatant was applied onto a 0.5-ml-CV Ni-NTA-agarose (Qiagen, Hilden, Germany) gravity flow column equilibrated with buffer D to remove the hexahistidine-tagged TEV protease and residual, still tagged AvrRxo1 protein. The flow-through was diluted to a conductivity of 10–11 mS/cm with buffer E (50 mm Tris-HCl, pH 8.5, 2 mm DTE) and loaded onto a 1-ml-CV HiTrap Heparin HP column (GE Healthcare) equilibrated with buffer E. Protein was eluted with a linear gradient of 20 CVs to buffer F (50 mm Tris-HCl, pH 8.5, 400 mm NaCl, 2 mm DTE). Pure protein fractions were pooled, concentrated in a 10,000-molecular weight-cutoff centrifugal filter and loaded onto a Superdex 75 10/300 GL column (GE Healthcare) equilibrated with buffer G (50 mm HEPES-NaOH, pH 7.5, 200 mm NaCl, 2 mm DTE) for final polishing and buffer exchange. The catalytically impaired, non-toxic AvrRxo1(D193N) variant could be expressed alone and was purified following the same protocol.</p><p>Similarly to AvrRxo1, the cleared supernatant from AvrRxo2 expression was loaded onto a 1-ml-CV Ni-NTA-agarose column (Qiagen) gravity flow column equilibrated with buffer A. A wash step of 10 CVs of buffer H (50 mm Tris-HCl, pH 8.5, 150 mm NaCl, 50 mm (NH4)2SO4, 5 mm imidazole, 5 mm β-me) was performed. Bound protein was eluted in 7 CVs of buffer C. To improve the yield, the flow-through of the first Ni-NTA purification step was rechromatographed. Both eluates were pooled and, the conductivity was adjusted to 10–11 mS/cm with buffer I (50 mm Tris-HCl, pH 8.0, 20 mm (NH4)2SO4, 2 mm DTE). To remove anionic species and contaminating oligonucleotides, AvrRxo2 was passed through a Mono Q 10/100 GL column (GE Healthcare) equilibrated with buffer I. The flow-through was concentrated by binding to a 1-ml-CV HisTrap Ni-NTA column (GE Healthcare) equilibrated with buffer A and eluted from the column with buffer J (50 mm Tris-HCl, pH 8.5, 150 mm NaCl, 50 mm (NH4)2SO4, 250 mm imidazole, 2 mm DTE). The AvrRxo2-containing eluate was dialyzed overnight against buffer K (50 mm Tris-HCl, pH 8.0, 150 mm NaCl, 50 mm (NH4)2SO4, 5 mm DTE). Finally, the protein was concentrated in a 3,000-molecular weight-cutoff centrifugal filter and loaded onto a Superdex 75 10/300 GL column (GE Healthcare) equilibrated with buffer G.</p><p>The cleared supernatant of AvrRxo1/AvrRxo2(W97A) co-expression was loaded onto a 0.8-ml-CV Ni-NTA-agarose (Qiagen) gravity flow column equilibrated with buffer A at pH 8.0. The column was washed with 12 CVs of buffer A at pH 8.0 supplemented with 2 mm DTE instead of β-me, and bound proteins were subsequently eluted with 6 CVs of buffer C at pH 8.0. The eluate was dialyzed overnight against buffer A at pH 8.0 and applied onto a 0.5-ml-CV glutathione-Sepharose 4 FF (GE Healthcare) gravity flow column equilibrated with buffer A at pH 8.0. A wash with 10 CVs of buffer L (50 mm Tris-HCl, pH 8.0, 150 mm NaCl, 50 mm (NH4)2SO4, 0.5 mm EDTA, 0.5 mm MgCl2, 2 mm DTE) was performed. Subsequently, the column material was suspended in 8 ml of buffer L supplemented with 1 mg of TEV protease and incubated at room temperature for 2.5 h under constant, mild agitation. The supernatant containing the complex of untagged AvrRxo1 and AvrRxo2(W97A) was collected, concentrated in a 10,000-molecular weight-cutoff centrifugal filter, and loaded onto a Superose 12 10/300 GL column (GE Healthcare) equilibrated with buffer G.</p><p>All purification procedures were monitored by Coomassie-stained 20% (w/v) SDS-PAGE, and purity of final protein batches was judged to be better than 98%. Protein concentrations of batches were determined on a Nanodrop (Thermo Fisher Scientific) using the following extinction coefficients at 280 nm: ϵ(AvrRxo1/AvrRxo1(D193N)) = 32,890 m−1 cm−1, ϵ(AvrRxo2) = 11,460 m−1 cm−1, and ϵ(AvrRxo2(W97A)) = 5,960 m−1 cm−1.</p><!><p>AvrRxo1 at 120 nm was probed for phosphotransferase activity in 200 μl of buffer M (50 mm HEPES-NaOH, pH 7.5, 200 mm NaCl, 15 mm KCl, 3 mm MgCl2, 1 mm EDTA, 1 mm TCEP supplemented with different nucleotides at 500 μm (ADP (Pharma Waldhof GmbH, Düsseldorf, Germany), GDP, UNAG (both from Sigma-Aldrich), UpA (IBA GmbH, Göttingen, Germany), NAD or NADH (both from Roche Diagnostics GmbH)), and equimolar amounts of ATP (Pharma Waldhof GmbH) or GTP (Sigma-Aldrich). Reactions were incubated for 5 h under mild agitation at 25 °C, quenched by freezing in liquid nitrogen, and stored at −80 °C. For analysis, samples were diluted with 2 ml of deionized water. The total sample was loaded onto a Mono Q 5/50 GL column (GE Healthcare) equilibrated with deionized water. Bound compounds were eluted in a linear gradient of 30 CVs to an aqueous solution of 1.5 m NH4+CH3COO− at pH 8.0. Eluting species were detected by measuring absorbance at 260, 280, and 340 nm.</p><!><p>Precultures of 50 ml of LB medium supplemented with 0.1 mg/ml ampicillin were inoculated with E. coli BL21(DE3)-RIL cells transformed with either pGEX4T-1_AvrRxo1ΔN88 or pGEX4T-1_AvrRxo1ΔN88(D193N) as a negative control and incubated at 37 °C. Cultures of untransformed cells were performed identically in LB medium without ampicillin. From these, final 500-ml expression cultures were inoculated and incubated at 37 °C until the A600 reached 0.2. The temperature was set to 16 °C, and protein expression was induced at an A600 of 0.3 by adding 0.5 mm isopropyl β-d-1-thiogalactopyranoside. Cells were harvested 90 min postinduction by centrifugation at 4 °C. Small metabolite extraction and HPLC analysis were performed as described previously (55).</p><!><p>Co-purified AvrRxo1-AvrRxo2(W97A) complex at 21 nm was incubated in a preparative enzymatic setup together with 6.8 mm NAD and 7 mm Mg2+-ATP in 10 ml of buffer N (50 mm HEPES-NaOH, pH 7.5, 200 mm NaCl, 15 mm KCl, 7 mm phosphoenolpyruvate, 1 mm EDTA). An additional 1.4 units/ml PK and 2 units/ml LDH (Sigma-Aldrich) were added. The reaction was incubated at room temperature for 19 h in the dark. Proteins were removed by filtration through a 10,000-molecular weight-cutoff centrifugal filter. The filtrate was diluted with deionized water to a conductivity of 13 mS/cm, and 40-ml aliquots were loaded onto a Mono Q 10/100 GL column equilibrated with deionized water. Phosphorylated NAD was separated from ATP by a step elution with 400 mm ammonium acetate buffer at pH 8.0. For polishing, the eluate was diluted with the same volume of deionized water and reloaded onto the Mono Q 10/100 GL column equilibrated with water. Most impurities were removed in a first step gradient to 240 mm ammonium acetate buffer at pH 8.0. The phosphorylated NAD product was then eluted from the column with 350 mm ammonium acetate buffer at pH 8.0. Subsequently, the buffer was exchanged to the more volatile ammonium carbonate buffer at pH 8.0 by reloading phosphorylated NAD diluted in deionized water onto the Mono Q 10/100 GL column followed by a step elution with 150 mm ammonium carbonate buffer at pH 8.0. The purified product was concentrated under vacuum, and a brownish contaminant could be removed by 70% (v/v) ethanol precipitation, whereas the phosphorylated NAD remained in solution. Last traces of contaminants were removed by reloading the highly concentrated product onto a Mono Q 5/50 GL column and elution with 85 mm ammonium carbonate buffer at pH 8.0. After freeze-drying, the pure product was obtained as a colorless solid phase and dissolved in 600 μl of D2O for NMR characterization.</p><p>Phosphorylated NAAD was purified from reactions of Michaelis-Menten kinetic experiments. After measurements were taken, all reactions were pooled, frozen in liquid nitrogen, and stored at −80 °C. Purification of phosphorylated NAAD followed the same protocol as was used for purification of phosphorylated NAD with the only difference being that 600 mm NH4HCO3 at pH 8.0 was used for initial elutions.</p><!><p>Purification of 3′-NADP from SMEs was performed similarly as described for the in vitro product. Extracts were obtained from 3.5 liters of AvrRxo1-expressing E. coli culture identically as described above. 3′-NADP was purified by suspending the dried extracts in 10 ml of H2O and loading it onto a Mono Q 10/100 GL column equilibrated with H2O. The flow-through containing species less negatively charged than 3′-NADP was discarded, and bound species were eluted with a linear gradient of 0–1 m ammonium acetate buffer at pH 8.0 over 15 CVs. Eluate fractions were pooled and applied onto a Mono Q 5/50 GL column equilibrated with H2O. 3′-NADP was separated from other bound species by washing the column with 5 CVs of 300 mm ammonium acetate buffer at pH 8.0 and then eluting 3′-NADP with a linear gradient over 4 CVs of the buffer from 300 to 500 mm. Fractions containing 3′-NADP were again loaded onto a Mono Q 10/100 GL column equilibrated with H2O for buffer exchange to the more volatile ammonium carbonate. 3′-NADP was eluted with 350 mm ammonium carbonate buffer at pH 8.0. The obtained fractions were lyophilized and suspended in H2O until the sample no longer showed any traces of ammonia.</p><!><p>Molecular masses of AvrRxo1products purified from in vitro reactions and small metabolite extracts were determined on a Bruker maXis II mass spectrometer (Bruker Corp., Billerica, MA) following dilution in deionized water. Compounds were fragmented by collision-induced dissociation at −45 eV for phosphorylated NAD and −40 eV for phosphorylated NAAD.</p><!><p>3′-NADP purified from in vitro reactions or small metabolite extracts was assayed for its capability to induce activity of d-glucose-6-phosphate dehydrogenase from Leuconostoc mesenteroides (Sigma-Aldrich). Reactions were performed similarly to spectroscopic assays using the PK/LDH system in 10-mm-path length quartz cuvettes in buffer O (50 mm HEPES-NaOH, pH 7.5, 200 mm NaCl). Final concentrations of d-glucose-6-phosphate dehydrogenase (G6P-DH) were 14.7 units/ml. 200 μm 3′-NADP from in vitro reactions or SMEs was added to the reaction after 2 min and then incubated for 10 min. To verify G6P-DH activity, 200 μm 2′-NADP (Roche Diagnostics GmbH) was added after the 10-min incubation period, and the reaction was allowed to proceed until all 2′-NADP was reduced to 2′-NADPH.</p><!><p>The concentration in our NMR experiments was 3 mg/ml. 3′-NADP purified from in vitro reactions was dissolved in 99.9% D2O and spectra were recorded at 25 °C using a Varian 500 NMR system spectrometer (Agilent). One dimensional NMR experiments were processed and analyzed using the TopSpin 3.2 software (Bruker BioSpin), heteronuclear single quantum coherence experiments were analyzed using the MestReNova 10.0 (MestreLab Research) software. Signal assignment of 1H NMR experiments (500 MHz, 1H-1H COSY, D2O) is: δ = 9.33 (s, 1H, H-N2), 9.15 (d, 1H, 3JHH = 6.2 Hz, H-N6), 8.86 (d, 1H, 3JHH = 8.2 Hz, H-N4), 8.47 (s, 1H, H-A8), 8.21 (m, 1H, H-N5), 8.16 (s, 1H, H-A2), 6.07 (m, 2H, N1′, H-A1′), 4.87 (m, 1H, H-A2′), 4.84–4.81 (m, 1H, H-A2′), 4.58 (m, 1H, H-A4′), 4.55 (m, 1H, H-N4′), 4.50 ("t", 1H, 3JHH ∼ 5.3 Hz, H-N2′), 4.44 (dd, 3JHH = 5.0 Hz, H-N3′), 4.39–4.35 (m, 1H, H-N5′a), 4.31–4.22 (m, 3H, H-N5′b, H-A5′a,b) ppm.</p><p>Assignment of 13C NMR experiments (152.7 MHz, APT, HSQC, D2O) is: δ = 165.12 (C = O), 155.07 (C-A6), 152.36 (C-A2), 149.02 (Cq), 145.69 (C-N4), 142.29 (C-N6), 139.82 (C-N2), 133.59 (Cq), 128.57 (C-N5), 118.36 (Cq), 99.97 (C-N1′), 87.00 (JCP = 8.9 Hz, C-N4′), 86.18 (C-A1′), 83.21 (JCP = 3.9 Hz, JCP = 3.7 Hz, C-A4′), 77.51 (C-N2′), 73.90 (JCP = 5.0 Hz, C-A3′), 73.33 (JCP = 5.0 Hz, C-A2′), 70.61 (C-N3′), 65.34 (JCP = 5.4 Hz, C-A5′), 64.89 (JCP = 5.2 Hz, C-N5′) ppm.</p><p>Assignment of 31P NMR decoupling experiments (202.4 MHz, D2O) is: δ = 0.15 (s, P-A3′), −11.37 (d, JPP = 20.3 Hz, P-5′), 11.66 (d, JPP = 20.3 Hz, P-5′) ppm.</p><!><p>Substrate comparisons were performed with the PK/LDH coupled spectrophotometric assay similarly as described (29). Final assay conditions were 100 nm AvrRxo1 or AvrRxo1(D193N) (diluted in buffer G supplemented with 1 mg/ml BSA and 1 mm TCEP), 2.7 units/ml PK, and 4.0 units/ml LDH (Sigma-Aldrich) in 100 μl of buffer M containing 200 μm ATP and additional 400 μm PEP. Acceptor substrate candidates (NAD, NAAD, and UpA) or H2O as a negative control was added to 200 μm prior to starting the measurements. After 2 min, either AvrRxo1 or buffer G supplemented with 1 mg/ml BSA and 1 mm TCEP was added to the 10-mm-path length cuvette, and the reaction was incubated for 15 min. Subsequently, 200 μm NADH was added as substrate for LDH. The absorbance at 340 nm was constantly measured with a Jasco V-650 spectrophotometer.</p><!><p>Steady state kinetic measurements were performed with the PK/LDH coupled spectrophotometric assay similarly as described (29). Final assay conditions were 2.5 nm AvrRxo1, 2.7 units/ml PK, and 4.0 units/ml LDH (Sigma-Aldrich) in 550 μl of buffer M supplemented with 400 μm NADH and 1 mg/ml BSA. Note that NAD concentrations during titrations thereof had to be corrected by 56 μm due to a 1.8–1.9% ADP impurity of the ATP stock. For Michaelis-Menten kinetics in which Mg2+-ATP was titrated, a constant concentration of 4 mm NAD was included in the assay; when NAD or NAAD was titrated, a constant concentration of 3 mm Mg2+-ATP was used. Each NAD titration series of AvrRxo2 inhibition kinetics was performed in the absence or presence of 2.5, 5, 7.5, or 12.5 nm AvrRxo2. Here, Mg2+-ATP was also used at 3 mm. All reactions were carefully mixed in Eppendorf tubes and immediately transferred into quartz cuvettes (5-mm path length). Apparent reaction velocities were determined by monitoring the decrease in absorbance at 340 nm at 25 °C in a Jasco V-650 spectrophotometer using an automated, thermostated cell changer. Individual reaction velocities (ΔA/min) were determined using the kinetics analysis tool of the Spectra Manager software supplied by the manufacturer in time intervals of 145–756 s containing 5–36 data points. Those were converted to their apparent kobs (s−1) values and fitted to a modified Michaelis-Menten equation, kobs = ((kcat × [S])/(Km + [S])) + kbasal, using GraphPad Prism v.6.05 (GraphPad, La Jolla, CA). Note that a basal rate of ADP formation (kbasal) independent from the titrated substrate was evident and thus included in the fit during NAD and NAAD titration experiments. This rate was determined to be 59 s−1.</p><!><p>F. S. and A. M. conceived and coordinated the study and wrote the paper. F. S., A. R., and D. E. designed, performed, and analyzed the experiments shown in Figs. 1, 2, and 4. F. S. and C. B. designed, performed, and analyzed the experiments shown in Fig. 3. F. S. and J. S. designed, performed, and analyzed the experiments shown in Figs. 5 and 6. All authors reviewed the results and approved the final version of the manuscript.</p><!><p>The authors declare that they have no conflicts of interest with the contents of this article.</p><p>This article was selected as a Paper of the Week.</p><p>This article contains supplemental Figs. S1–S4.</p><p>hypersensitive response</p><p>pathogen-associated molecular pattern-triggered immunity</p><p>effector-triggered susceptibility</p><p>nicotinic acid adenine dinucleotide phosphate</p><p>effector-triggered immunity</p><p>avirulence</p><p>UDP-N-acetylglucosamine</p><p>small metabolite extract</p><p>pyruvate kinase</p><p>phosphoenolpyruvate</p><p>lactate dehydrogenase</p><p>uridine 5′ → 3′ adenine dinucleotide</p><p>amino acid(s)</p><p>β-mercaptoethanol</p><p>column volume</p><p>nickel-nitrilotriacetic acid</p><p>dithioerythritol</p><p>tobacco etch virus</p><p>millisiemens</p><p>d-glucose-6-phosphate dehydrogenase</p><p>tris(2-carboxyethyl)phosphine)</p><p>electrospray ionization.</p>

PubMed Open Access

Design, synthesis, and anticonvulsant effects evaluation of nonimidazole histamine H3 receptor antagonists/inverse agonists containing triazole moiety

AbstractHistamine H3 receptors (H3R) antagonists/inverse agonists are becoming a promising therapeutic approach for epilepsy. In this article, novel nonimidazole H3R antagonists/inverse agonists have been designed and synthesised via hybriding the H3R pharmacophore (aliphatic amine with propyloxy chain) with the 1,2,4-triazole moiety as anticonvulsant drugs. The majority of antagonists/inverse agonists prepared here exerted moderate to robust activities in cAMP-response element (CRE) luciferase screening assay. 1-(3-(4-(3-Phenyl-4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine (3l) and 1-(3-(4-(3-(4-chlorophenyl)-4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine (3m) displayed the highest H3R antagonistic activities, with IC50 values of 7.81 and 5.92 nM, respectively. Meanwhile, the compounds with higher H3R antagonistic activities exhibited protection for mice in maximal electroshock seizure (MES)-induced convulsant model. Moreover, the protection of 3m against the MES induced seizures was fully abrogated when mice were co-treated with RAMH, a CNS-penetrant H3R agonist, which suggested that the potential therapeutic effect of 3m was through H3R. These results indicate that the attempt to find new anticonvulsant among H3R antagonists/inverse agonists is practicable.

design,_synthesis,_and_anticonvulsant_effects_evaluation_of_nonimidazole_histamine_h3_receptor_antag

Introduction<!><!>Introduction<!>Chemistry<!><!>Evaluation of H3R antagonistic activity<!><!>Evaluation of H3R antagonistic activity<!><!>Evaluation of H3R antagonistic activity<!><!>Evaluation of H3R antagonistic activity<!><!>Evaluation of H3R antagonistic activity<!>Anticonvulsant activity evaluation<!>Protective effects of H3R antagonists/inverse agonists 3a-3q on MES-induced convulsions<!><!>Protective effects of H3R antagonists/inverse agonists 3a-3q on PTZ-induced convulsions<!><!>Effects of compound 3m on MES-induced convulsions in dose dependent manner<!><!>Effects of RAMH pre-treatment on the compound 3m-provided protection in MES-induced seizure model<!><!>Conclusion<!>Synthesis<!>Synthesis of compounds 1a-1e<!>Synthesis of compounds 2a-2h<!>Synthesis of compounds 3a-3q<!>3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)-N,N-diethylpropan-1-amine hydrochloride (3a)<!>3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)-N,N-dipropylpropan-1-amine hydrochloride (3b)<!>4-(4-(3-(pyrrolidin-1-yl)propoxy)phenyl)-4H-1,2,4-triazole hydrochloride (3c)<!>1-(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperazine dihydrochloride (3d)<!>1-(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)-4-methylpiperazine hydrochloride (3e)<!>1-(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)-4-phenylpiperazine (3f)<!>4-(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)morpholine hydrochloride (3g)<!>1–(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine (3h)<!>1-(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)-4-phenylpiperidine (3i)<!>1-(3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine-4-carboxamide (3j)<!>1-(3-(4–(3-methyl-4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine hydrochloride (3k)<!>1-(3-(4-(3-phenyl-4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine (3l)<!>1-(3-(4-(3-(4-chlorophenyl)-4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine (3m)<!>1-(3-(4-(3-([1,1’-biphenyl]-4-yl)-4H-1,2,4-triazol-4-yl)phenoxy)propyl)piperidine (3n)<!>1-(2-(4–(4H-1,2,4-triazol-4-yl)phenoxy)ethyl)piperidine hydrochloride (3o)<!>1-(4–(4-(4H-1,2,4-triazol-4-yl)phenoxy)butyl)piperidine hydrochloride (3p)<!>1-(5-(4-(4H-1,2,4-triazol-4-yl)phenoxy)pentyl)piperidine (3q)<!>Cell culture and transfection<!>CRE-driven reporter gene assay<!>Drugs and animals<!>MES-induced seizure<!>PTZ-induced seizures<!>Statistics<!>Homology modelling<!>Molecular docking

<p>Epilepsy, a very common neurologic disorder, affects about around 1% of world population1,2. Presently, antiepileptic drugs (AEDs) are the main strategy of therapy. However, the AEDs available in the clinic such as phenytoin, carbamazepine, sodium valproate, topiramate, and oxcarbamazepine are only effective in approximately 70% of the patients with epilepsy. Moreover, their use is long-term and often accompanied with severely side effects, including naupathia, headache, and ataxia, even threaten the life of patients3–5. Investigations for more effective and safer AEDs are still a formidable and urgent task for medicinal chemists.</p><p>The role of central histaminergic system being concerned in epilepsy have been demonstrated in many experimental and epidemiological studies, in which histamine regulated seizure susceptibility as an anticonvulsant neurotransmitter6–8. For example, H1-antagonists such as pyrilamine, ketotifen that decrease brain histamine levels increased the duration of convulsive phase in electrically-induced convulsions model9. Histidine, as the precursor of histamine, showed protection against chemically-induced convulsions in rats, via activating the histamine H1 receptors10.</p><p>Histamine H3 receptors (H3R) as a G-protein coupled receptor (GPCR) binding to histamine like other histamine receptors, is expressed mainly in the central nervous system, where it acts as an auto-receptor in histaminergic neurons, and negatively regulates the synthesis and release of histamine11. What is more, as a inhibitory heteroreceptor, H3R also regulates the release of other neurotransmitters including dopamine, acetylcholine, serotonin, norepinephrine, γ-aminobutyric acid, and glutamate. These neurotransmitters, especially γ-aminobutyric acid and glutamate, are related to epilepsy inextricably12,13. Therefore, more attention has been focussed on H3R as an attractive therapeutic target for epilepsy treatment14.</p><p>A large number of experimental studies involved in acute and chronic models of epilepsy confirmed the anticonvulsive potential of H3R antagonists/inverse agonists. They showed the protection against experimental seizures by feedback increase of histamine release and binding with H1 receptors15,16. Besides, other mechanisms might be involved in their anticonvulsive action, such as facilitating of GABA release17–19, increasing histidine decarboxylase (HDC) activity20,21 and synergism with AEDs17,18,22.</p><p>Early, anticonvulsant activity of some imidazole H3R antagonists such as thioperamide and clobenpropit was confirmed in models of epilepsy (Figure 1)16,19,20,23. Recently, a large number of non-imidazole H3R antagonists such as DL77 (Figure 1) prepared by a group/team of Kiec-Kononowicz exhibited excellent anticonvulsant activity in the electrically-induced seizures model and subcutaneously pentylenetetrazole (PTZ)-induced seizure model at dose-dependent, and the therapeutic action was proved through H3R24–28. Sadek et al. synthesised some histamine H3R ligands (Figure 1, I) incorporating different antiepileptic structural motifs to investigate if the H3R pharmacophore could be combined to some antiepileptic molecules, and give some new anticonvulsants by the multiple-target approaches. The results were encouraging, which indicated that the H3R pharmacophore successfully combined to the antiepileptic molecules, maintaining the H3R affinity and anticonvulsant activity, although the anticonvulsant activity decreased compared to the prototypal antiepileptic molecules (Figure 1)29,30.</p><!><p>Structures of histamine H3 receptor ligands with anticonvulsant activity, triazole derivatives with anticonvulsant activity and target compounds 3a-3q designed.</p><!><p>Pitolisant (PIT), a H3R antagonist/inverse agonist, has been subjected into clinical Phase III for the treatment of epilepsy31. When used alone or in combination with other AEDs in the human photosensitivity model at dose ranges of 30–60 mg, PIT showed a favourable EEG profile in a dose-dependent manner32.</p><p>Supported by the above results, in this work, we designed and synthesised some novel H3R antagonists/inverse agonists by hybriding the H3R pharmacophore (aliphatic amine with propyloxy chain) with the 1,2,4-triazole, the latter have been identified as an important and effective anticonvulsive fragment in recent years (Figure 1, II and III)33–37. According to Quan's reports, the 1,2,4-triazole derivatives were likely to have several mechanisms of action such as inhibiting voltage-gated sodium ions channel and modulating GABAergic activity38–40. And a group of Plech illustrated the anticonvulsive effects of 4-alkyl-5-aryl-1,2,4-triazole-3-thione derivatives and suggested that the influence on the voltage-gated Na+ channels was involved in them at least41,42. Therefore, in this work, our strategy was to design molecules combining pharmacophores of H3R antagonists and another anticonvulsant active pharmacophore (e.g. 1,2,4-triazole moiety) into one skeleton, and then produced a synergism for anticonvulsant active.</p><!><p>According to Schemes 1 and 2, the target compounds (3a-3q) were synthesised smoothly. In brief, formyl hydrazine reacted with 4-aminophenol in dimethoxyl-N,N-dimethyl formamide (DMF-DMA) to give the 4-(4H-1,2,4-triazol-4-yl)phenol (1a). Compound 1a underwent a nucleophilic substitution with 1-bromo-3-chloropropane to get 4-(4-(3-chloropropoxy)phenyl)-4H-1,2,4-triazole (2a). The reaction was conducted in the presence of potassium hydroxide in dimethyl sulfoxide (DMSO) at room temperature to ensure the formation of single-substituted derivatives. Finally, proper amines reacted with compound 2a in the presence of K2CO3 and KI in the solvent of CH3CN to give the desired compounds 3a-3j. To enrich the structure–activity relationship, we also prepared the derivatives of 3h via introducing the substituents at the triazole ring and adjusting the length of the link. The reaction conditions used to prepare these compounds (3k-3q) were the same as above. Compounds (3a, 3b, 3c, 3d, 3e, 3 g, 3k, 3o, and 3p) obtained as the form of oil were transformed to hydrochlorate. Their structures were characterised and confirmed by1H-NMR, 13 C-NMR, and HR-MS.</p><!><p>The synthesis route and conditions for the preparation of compounds 3a-3j.</p><p>The synthesis route and conditions for the preparation of compounds 3k-3q.</p><!><p>cAMP-response element (CRE) reporter gene assay has been extensively used to evaluate the efficacy of GPCR antagonists or agonists. In this work, the H3R antagonistic activities of the prepared 3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)-propylamine derivatives have been screened by CRE-driven luciferase assay, in which the HEK-293 cells expressing the human H3R and a reporter gene consisting of the firefly luciferase coding region were used43,44. Ciproxifan (CXP) and Pitolisant (PIT) were employed as the positive controls. Initially, compounds and positive controls were tested at two concentrations (100 nM and 1 µM) to obtain the preliminary investigation of their H3R antagonistic activities. In the assays, the antagonistic activities was positively correlated with the rise of the fluorescence value and indicated by the % antagonism. For the prominent compounds IC50 values were determined at additional assays.</p><p>As seen in Table 1, majority of the synthesised compounds displayed gentle to robust H3R antagonistic activities and eight of them exhibited micromolar inhibitory activity. The antagonistic activities of all compounds depended on the concentration treated. It is worth mentioning that compounds 3l (IC50 = 7.81 nM) and 3m (IC50 = 5.92 nM) displayed the most potent H3R antagonistic activities, with the much stronger potency than that of CXP (IC50 = 0.082 µM) and PIT (IC50 = 0.5 µM) in the CRE reporter gene assay.</p><!><p>H3R antagonistic activity of compounds 3a-3q.</p><p>a% Antagonism, value represented as mean ± standard deviation of three independent experiments.</p><p>bNT, IC50 was not tested.</p><p>cCPX, an antagonist of H3R ciproxifan.</p><p>dPIT, an antagonist/inverse agonist of H3R pitolisant.</p><!><p>Surprisingly, antagonism percent of some compounds as well PIT were above 100%. It is well known that H3R is a GPCR coupled with Gαi. When the ligand (histamine) binds to H3R, the dissociated Gαi inhibits the activity of adenylate cyclase (AC) and down-regulates the level of intracellular cAMP. In this CREs driven luciferase assay, when cells were pre-treated by H3R antagonists, the downregulation of cAMP would be inhibited, and the level of intracellular cAMP would regain to the initial level. However, in some cases, the cAMP levels raised above the initial level, giving the % antagonism greater than 100%. Our first speculation was that these compounds might stimulate the AC directly and up-regulate the level of cAMP. However, an additional assay indicated that these compounds didn't have effects on the level of cAMP when pre-treated alone. As shown in Figure 2, forskolin (2 µM) treated group gave more than 200 times rise for the cAMP level when compared to the control group. While Histamine, ciproxifan, and compounds 3a, 3c, 3h, 3k, 3l, and 3m have no significant effects on the level of cAMP when carried out comparisons by ANOVA followed by Dunnett's test.</p><!><p>Effects of compounds 3a, 3c, 3h, 3k, 3l, and 3m on the level of intracellular cAMP administrated alone.</p><!><p>Another explanation is that these compounds may be inverse agonists when binding to H3R, which not only antagonise the function of histamine, but also give the inverse agonistic performance. Actually, PIT is a well-known H3R antagonist and reverse agonist. So, the above results make sense. Based on the above, we further assessed the H3R inverse agonist activity of compound 3m and PIT by using CRE-luciferase assay. In this experiment, transfected HEK-293 cells were stimulated with 10 µM forskolin or 10 µM forskolin plus different concentrations of compound 3m. The raise of luciferase activity after adding compound represented the inverse agonistic activity. The EC50 was calculated by seven concentrations. Cytotoxicity appeared at 100 µM. As shown in Figure 3, PIT and compound 3m showed effective H3R inverse agonistic activity with an EC50 value of 403 and 129 nM, respectively.</p><!><p>The H3R inverse agonistic activity (EC50, μM) of Pitolisant and compound 3m.</p><!><p>Simple structure–activity relationships (SARs) could be obtained from Table 1. In the series of 3a-3j, the different tertiary amines significantly influenced the H3R antagonistic activities. The N-ethyl derivative 3a showed an IC50 of 2.9 µM, while the activity declined sharply for the N-propyl derivative 3b. Interestingly, compounds containing piperazine or morpholine (3d-3g) exhibited weaker activities than those with piperidine or pyrrolidine (3c and 3h). This probably attributed to the increase of the molecular polarity. The introduction of phenyl or amide group on the piperidine ring of compound 3h, gave the compounds 3i and 3j, which also decreased the H3R antagonistic activities when compared to compound 3h. Based on the facts above, it could be concluded that the N,N-diethyl group, pyrrolidine and piperidine were more of benefit to the H3R antagonistic activities of the 3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)-propylamine skeleton, and piperidine derivative (3h) was the best one with the IC50 of 0.127 µM.</p><p>To enrich the structure–activity relationships, we prepared the derivatives of 3h via introducing the substituents at the triazole ring and adjusting the length of the link.</p><p>Compounds 3k, 3l, 3m, and 3n were substituted on 3-position of 1,2,4-triazole ring with methyl, phenyl, para-chlorophenyl, and biphenyl, respectively. Encouragingly, the introduction of methyl, phenyl, and para-chlorophenyl groups significantly increased the H3R antagonistic activities, giving the two prominent compounds 3l and 3m with nanomolar IC50 values. While the biphenyl substituted compound 3n showed weaker activity when compared to 3h. Replacing the three-carbon link in the compound 3h with two-carbon, four-carbon and five-carbon links, gave the compounds 3o, 3p, and 3q, respectively. It could be seen that the length of the link had a direct impact on H3 receptor antagonistic activities of the 3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)-propylamine derivatives. The activity order of the link length of carbon was 3 > 2 > 4 ≫ 5.</p><p>To investigate the molecular determinants that manage the antagonistic activities of the tested compounds, molecular docking studies of PIT, 3h, and 3m with the H3R homology model were carried out. The homology model was constructed from the crystal structure of the H1 receptor (PDB ID: 3RZE)45. The docking results are shown in Figure 4.</p><!><p>The predicted configurations for PIT (A), 3h (B) and 3m (C) binding with H3R, and their overlying pattern (D).</p><!><p>As shown in Figure 4(A), PIT bound to H3R through two critical H-bond interactions with Tyr115 and Glu206, and other interactions with amino acid residues Arg381, Phe193, Met378 and so on. Figure 4(B) revealed that compound 3h had a similar binding pattern to PIT, interacting with the same amino acid residues Glu206, Tyr115, Arg381, and Met378. Surprisingly, the compound 3m with the highest H3R antagonistic activity showed a different binding pattern to PIT (as seen in Figure 4(C)). The overlying pattern of PIT, 3h, and 3m was shown in Figure 4(D). The piperidine group of 3m was not involved in the formation of the salt bridge or hydrogen-bond interactions with Glu206, which was generally considered as the critical residue of H3R46,47. The unexpected binding pattern of 3m might be due to the phenyl group on the triazole ring, which did not fit into the hydrophobic cavity in TMs 3-5-6 region of H3R, even though the compound 3m showed a forceful binding with H3R via another mode. The triazole nitrogen established an ionic bond with Glu206, and a hydrogen bond was observed between piperidine nitrogen and Tyr115. π-π shaped, and alkyl interactions with Trp371, Tyr343, Arg381, His187, Leu199, and ALA202 were observed to support the forceful binding with H3R.</p><!><p>To investigate the anticonvulsive effects, all the target compounds (3a-3q) were screened in the MES-induced and PTZ-induced convulsion models in mice. Compounds were administered intraperitoneally (i.p.) to mice at dosage of 10 mg/kg in the both models. PIT and valproic sodium (VPA) were used as positive controls in the tests.</p><!><p>Protection for the mice was defined as the reduction or abolition of the tonic hind limb extension (THLE) in the MES model in mice. As seen in Figure 5, compounds 3a, 3 g, 3h, 3l, 3m, and 3o showed moderate protection for the electro-stimulated mice with significant difference from that of the control group (p < 0.05, p < 0.01, or p < 0.001). Mice pre-treated with PIT (10 mg/kg, i.p.) and VPA (300 mg/kg, i.p.) were moderately or potently protected, respectively. Generally, the anticonvulsant activities of these compounds in MES model correlated directly to their H3R antagonistic activities. For example, the antiepileptic activity obtained of compound 3m was the highest, and in vitro H3R antagonistic activity measured for 3m with IC50 of 5.92 nM was also the highest. Compounds 3a, 3h, 3l, and 3o, showing anticonvulsant activity in the MES model, also showed good H3R antagonistic activities. Compound 3c, 3k, and 3p reduced the average duration of THLE, although they did not achieve a significant difference from the control group.</p><!><p>Effects of H3R antagonists/inverse agonists 3a-3q (10 mg/kg, i.p.), PIT (10 mg/kg, i.p.) and anticonvulsant drug VPA (300 mg/kg, i.p.) against MES-induced convulsions. Protection for mice was defined as the reduction or abolition of the tonic hind limb extension (THLE) in MES model. Results are showed as mean ± SEM with seven animals in each group. Values are considered significant at *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 as compared to saline-treated group.</p><!><p>Some experiments indicated that H3R antagonists/inverse agonists could protect animals in PTZ-induced convulsions model26,29. So the compounds 3a-3q, PIT, and VPA were also screened in the PTZ model in mice. Unfortunately, all compounds tested at the dose of 10 mg/kg (i.p.) did not show any protection against the seizures induced by PTZ. PIT also failed to protect the PTZ-treated mice as well at the same conditions. By contrast, anticonvulsant agent VPA showed full protection against the PTZ-induced convulsions (Figure 6).</p><!><p>Effects of compounds 3a-3q (10 mg/kg, i.p.), and reference drug PIT (10 mg/kg, i.p.) and VPA (300 mg/kg, i.p.) against PTZ-induced convulsions. Results are showed as mean ± SEM of seven mice in each group. & represent full protection.</p><!><p>In a further experiment, compound 3m, as the most active one in the MES-induced seizure model, was chosen to verify its protective effect in different doses. Encouragingly, the 3m-provided protections were observed and were dose dependent. The standard antagonist PIT also displayed anticonvulsive activity dose-dependently at the same condition. Notably, when pre-treated with 20 mg/kg dose, PIT could fully abrogate the tonic hind limb extension induced by electro-stimulation, showing its potential anticonvulsant activity (Figure 7). To exclude the possibility that the anticonvulsant activity of 3m was connected with sedative effect, we carried out a rotarod test for 3m. The result showed that compound 3m had no neurotoxicity at the maximum dose of 10 and 20 mg/kg (the details could be seen in Support Table 1).</p><!><p>Protective effects of compound 3m and reference drug PIT against MES-induced convulsions in different doses. Protection in the test was defined as the reduction or abolition of the THLE in mice. Results were showed as mean ± SEM with seven animals in each group. Values are considered significant at *p < 0.05, **p < 0.01, ***p < 0.001 when compared to saline-treated group. &PIT, at 20 mg/kg dose, fully abrogate the THLE for all the tested mice.</p><!><p>To investigate the correlation between the anticonvulsant activity and H3R antagonistic activity of compound 3m, the protection provided by compound 3m against MES-induced seizure was reassessed after the administration of RAMH (10 mg/kg, i.p.), a CNS penetrant histamine H3R agonist. The results indicated that when co-administration with RAMH, compound 3m lost its original protective effect (Figure 8). Administration of RAMH alone also did not affect the duration of THLE of mice with p > 0.05 for saline versus RAMH. The above findings suggested that H3R antagonism was the main contributor for the anticonvulsant activity of compound 3m in MES model. When the H3R was blocked by H3R antagonist 3m, histamine or other neurotransmitter such as GABA in the CNS increased, finally leading to anticonvulsive effects.</p><!><p>Protective effects of compound 3m (10 mg/kg, i.p.) against MES-induced convulsions when pre-treatment of RAMH (10 mg/kg, i.p.). Protection in the test was defined as the reduction or abolition of the THLE in mice. Results are showed as mean ± SEM with seven animals in each group. Values are considered significant at *p < 0.01 as compared to saline-treated group, and #p < 0.01 as compared to 3m + RAMH treated group.</p><!><p>To identify novel H3R antagonists/inverse agonists with potential anticonvulsant activities, a series of 3-(4-(4H-1,2,4-triazol-4-yl)phenoxy)-propylamine derivatives were designed through combining pharmacophore of H3R antagonists and another anticonvulsant active pharmacophore (1,2,4-triazole moiety) into one molecule. The majority of those prepared compounds displayed moderate to robust H3R antagonistic activities. The SAR analysis revealed that piperidine and triazolephenol linked by three-carbon chain was benefit for the H3R antagonistic activity, and substitution by aromatic nucleus on the 3-position of 1,2,4-triazole further increased the H3R antagonistic activities. The most potent H3R antagonists/inverse agonists 3l and 3m exhibited nanomolar H3R antagonistic activities with IC50 of 7.81 nM and 5.92 nM, respectively. Molecular docking analysis demonstrated that 3m strongly bound to H3R via interactions with Tyr115, Glu206, Trp371, Tyr343, and so on, although its binding mode was not similar to PIT. The anticonvulsive screens in vivo indicated that compounds with higher H3R antagonistic activities showed more protection in the MES-induced convulsant model in mice, while no one was observed protective effect in PTZ-induced convulsant model. Moreover, the protection of 3m in the seizure model was fully abrogated when mice were co-treated with a H3R agonist RAMH, which suggested that its potential therapeutic effect was through H3R.</p><!><p>All the chemical solvents and reagents were purchased from supplier and used as received. Unless otherwise specified, reactions were monitored by thin-layer chromatography (TLC). All NMR spectrum was carried out on an AV-300 spectrometer with 300 MHz. High resolution mass spectra were measured on an MALDI-TOF/TOF mass spectrometer.</p><!><p>Taking compound 1a as an example: dimethoxyl-N, N-dimethyl formamide (DMF-DMA, 1.31 g, 11 mmol) and formyl hydrazine (0.65 g, 11 mmol) were added into a flask containing 30 ml of acetonitrile. The mixture was heated up to 60 °C for 30 min, then 4-aminophenol (0.60 g, 5.5 mmol) and acetic acid (3 mL) were added and heated up to 120 °C for 9 h. The mixture was cooled, filtered and washed by acetonitrile to give the product 1a. Chemical formula: C8H7N3O (MW = 161.16). m.p. 270–272 °C, yield 73%. 1H-NMR (300 MHz, DMSO-d6): δ 6.90 (d, 2H, J = 8.9 Hz, Ph-H), 7.46 (d, 2H, J = 8.9 Hz, Ph-H), 8.94 (s, 2H, N = CH), 9.88 (s, 1H, OH). 13 C-NMR (DMSO-d6, 75 MHz): 157.75, 142.14, 126.18, 123.65, 116.60. The compounds 1 b-1e were obtained according to the above method using the other hydrazides.</p><!><p>Taking compound 2a as an example: compound 1a (0.50 g, 3.1 mmol) and KOH (0.35 g, 6.2 mmol) were put into a flask with 5 mL of DMSO. The mixture was stirred for 5 min at 20 °C. Then added 1-bromo-3-chloropropane (0.98 g, 6.2 mmol) into the mixture and continued the reaction. After completion of the reaction indicated by the TLC (developing agent ratio: CH2Cl2/CH3OH = 15/1), the mixture was poured into 30 mL of water. The solution was extracted with dichloromethane three times. The organic layers were combined, washed with saturated salt water, dried over MgSO4, filtered, and concentrated. Purification by column chromatography (silica gel, 0-5% methanol in CH2Cl2) gave the compound 2a. Chemical formula: C11H12ClN3O (MW = 237.69). m.p. 102–104 °C, yield 81%. 1H-NMR (300 MHz, DMSO-d6): δ 2.15–2.23 (m, 2H, OCH2CH2), 3.80 (t, 2H, J = 6.5 Hz, ClCH2), 4.16 (t, 2H, J = 6.1 Hz, OCH2), 7.13 (d, 2H, J = 9.0 Hz, Ph-H), 7.60 (d, 2H, J = 9.0 Hz, Ph-H), 9.00 (s, 2H, N=CH). 13 C-NMR (75 MHz, DMSO-d6): δ 158.40, 142.03, 127.75, 123.40, 116.03, 65.22, 42.35, 32.05. As mentioned above, replaced the reactant 1a by the alternative 4-(3-substituted-4H-1,2,4-triazol-4-yl)phenols (1b-1e) to give the compounds 2b-2e. Compounds 2f, 2g, 2h were obtained by the same method as above just replacing 1-bromo-3-chloropropane by 1-bromo-2-chloroethane, 1-bromo-4-chlorobutane, 1-bromo-2-chloropentane, respectively.</p><!><p>Taking compound 3a as an example: in a 100 mL round-bottom flask with 15 mL of acetonitrile, compound 2a (0.40 g, 1.68 mmol), diethylamine (0.245 g, 3.36 mmol), K2CO3 (0.46 g, 3.36 mmol) and potassium iodide (0.56 g, 3.36 mmol) were added one by one. The mixture was heated up to 110 °C for 12–16 h. After cooing the mixture to 40 °C, it was filtered and dried by vacuum to obtain a residue. Purification by column chromatography (silica gel, 0–20% methanol in CH2Cl2) gave the compound 3a. The same conditions were used to prepare the compounds 3b-3q. Compounds 3a, 3c, 3d, 3e, 3 g, 3k, 3o, and 3p obtained as oils were transformed into the corresponding hydrochloride by hydrogen chloride in CH2Cl2.</p><!><p>Chemical formula: C15H22N4O × HCl (MW = 310.83). m.p. 105–106 °C, yield 82%. 1H-NMR (300 MHz, DMSO-d6): δ 1.25 (t, 6H, J = 7.2 Hz, CH3), 2.18 (t, 2H, J = 9.0 Hz, OCH2CH2), 3.08–3.16 (m, 6H, N(CH2)3), 4.17 (t, 2H, J = 6.0 Hz, OCH2), 7.19 (d, 2H, J = 8.8 Hz, Ph-H), 7.75 (d, 2H, J = 8.8 Hz, Ph-H), 9.84 (s, 2H, N=CH), 10.96 (s, 1H, HCl). 13 C-NMR (75 MHz, DMSO-d6): δ 159.34, 142.39, 126.44, 124.35, 116.08, 66.00, 48.04, 46.59, 23.40, 8.87. ESI-HRMS calculated for C15H23N4O+ ([M-Cl]+): 275.1866; found: 275.1860.</p><!><p>Chemical formula: C17H26N4O × HCl (MW = 338.88). m.p. 180–183 °C, yield 62%. 1H-NMR (300 MHz, CDCl3,): δ 1.05 (t, 6H, J = 7.3 Hz, CH3), 1.77–1.90 (m, 4H, N(CH2CH2)2), 2.27–2.37 (m, 2H, OCH2CH2), 3.15 (t, 4H, J = 8.0 Hz, N(CH2)2), 3.40 (t, 2H, J = 8.3 Hz, NCH2), 4.18 (t, 2H, J = 5.6 Hz, OCH2), 7.05 (d, 2H, J = 8.9 Hz, Ph-H), 7.43 (d, 2H, J = 8.9 Hz, Ph-H), 8.62 (s, 2H, CH=N)., 9.51 (s, 1H, HCl). 13 C-NMR (75 MHz, CDCl3): δ 163.08, 146.48, 132.18, 128.41, 120.58, 69.93, 59.42, 55.13, 28.54, 21.98, 15.92. ESI-HRMS calculated for C17H27N4O+ ([M-Cl]+): 303.2179; found: 303.2178.</p><!><p>Chemical formula: C15H20N4O × HCl (MW = 308.81). m.p. 120–122 °C, yield 56%. 1H-NMR (300 MHz, DMSO-d6): δ 1.92–2.00 (m, 4H, NCH2CH2), 2.19–2.22 (m, 2H, OCH2CH2), 3.00–3.53 (m, 6H, N(CH2)3), 4.17 (t, 2H, J = 5.7 Hz, OCH2), 7.17 (d, 2H, J = 8.7 Hz, Ph-H), 7.72 (d, 2H, J = 8.7 Hz, Ph-H), 9.75 (s, 2H, N=CH), 11.45 (s, 1H, HCl). 13 C-NMR (75 MHz, DMSO-d6): δ 159.25, 142.36, 126.57, 124.25, 116.07, 66.03, 53.21, 51.46, 25.53, 23.24. ESI-HRMS calculated for C15H21ClN4O+ ([M-Cl]+): 273.1710; found: 273.1711.</p><!><p>Chemical formula: C15H21N5O × 2HCl (MW = 323.83). m.p. 146–148 °C, yield 52%. 1H-NMR (300 MHz, CDCl3): δ 2.20–2.30 (m, 2H, J = 7.0 Hz, OCH2CH2), 3.32 (t, 2H, J = 7.8 Hz, NCH2), 3.37–3.57 (m, 8H, Piperazine-H), 4.17 (t, 2H, J = 5.8 Hz, OCH2), 7.16 (d, 2H, J = 8.9 Hz, Ph-H), 7.68 (d, 2H, J = 8.9 Hz, Ph-H), 9.44 (s, 2H, CH=N), 10.04 (s, 2H, HCl). 13 C-NMR (75 MHz, DMSO-d6): δ 158.91, 142.29, 127.01, 123.96, 116.10, 65.89, 53.41, 48.17, 45.78, 23.50. ESI-HRMS calculated for C15H22N5O+ ([M-2HCl + H]+): 288.1819; found: 288.1820.</p><!><p>Chemical formula: C16H23N5O × HCl (MW = 337.85). m.p. 220–223 °C, yield 65%. 1H-NMR (300 MHz, CDCl3): δ 2.26 (s, 2H, OCH2CH2), 2.84 (s, 3H, CH3), 3.35-3.66 (m, 10H, N(CH2)5), 4.17 (t, 2H, J = 5.2 Hz, OCH2), 7.16 (d, 2H, J = 8.7 Hz, Ph-H), 7.68 (d, 2H, J = 8.7 Hz, Ph-H), 9.46 (s, 2H, CH=N), 10.57 (s, 1H, HCl). 13 C-NMR (75 MHz, DMSO-d6): δ 158.93, 142.30, 126.96, 123.99, 116.10, 65.90, 54.53, 48.75, 45.75, 23.60, 15.53. ESI-HRMS calculated for C16H24N5O+ ([M-Cl]+): 302.1975; found: 302.1976.</p><!><p>Chemical formula: C21H25N5O (MW = 363.47). m.p. 172–174 °C, yield 64%. 1H-NMR (300 MHz, DMSO-d6): δ 1.91–1.97 (m, 2H, OCH2CH2), 2.47–2.55 (m, 6H, N(CH2)3), 3.12–3.15 (m, 4H, N(CH2)2), 4.09 (t, 2H, J = 6.2 Hz, OCH2), 6.76 (t, 1H, J = 7.2 Hz, Ph-H), 6.92 (d, 2H, J = 8.0 Hz, Ph-H), 7.10 (d, 2H, J = 8.9 Hz, Ph-H), 7.2 (t, 2H, J = 7.9 Hz, Ph-H), 7.58 (d, 2H, J = 8.9 Hz, Ph-H), 9.00 (s, 2H, CH=N). 13 C-NMR (75 MHz, DMSO-d6): δ 158.70, 151.50, 142.03, 129.36, 127.52, 123.37, 119.21, 116.00, 115.77, 66.81, 54.82, 53.26, 48.68, 26.60. ESI-HRMS calculated for C21H26N5O+ ([M + H]+): 364.2132; found: 364.2133.</p><!><p>Chemical formula: C15H20N4O2 × HCl (MW = 324.81). m.p. 237–239 °C, yield 51%. 1H-NMR (300 MHz, CDCl3): δ 2.27 (q, 2H, J = 6.0 Hz, OCH2CH2), 3.12-3.48 (m, 6H, N(CH2)3), 3.92 (t, 4H, J = 7.7 Hz, OCH2), 4.16 (t, 2H, J = 6.0 Hz, OCH2), 7.14 (d, 2H, J = 8.9 Hz, Ph-H), 7.67 (d, 2H, J = 8.9 Hz, Ph-H), 9.52 (s, 2H, N=CH), 11.44 (s, 1H, HCl). 13 C-NMR (75 MHz, DMSO-d6): δ 159.02, 142.22, 126.81, 123.93, 116.01, 65.94, 63.63, 53.95, 51.55, 23.37. ESI-HRMS calculated for C15H21N4O2+ ([M - Cl]+): 289.1659; found: 289.1658.</p><!><p>Chemical formula: C16H22N4O (MW = 286.38). m.p. 249–252 °C, yield 67%. 1H-NMR (300 MHz, DMSO-d6): δ 1.38–1.91 (m, 6H, NCH2CH2(CH2)2), 2.20–2.30 (m, 2H, OCH2CH2), 2.83–3.44 (m, 6H, N(CH2)3), 4.15 (t, 2H, J = 6.0 Hz, OCH2), 7.17 (d, 2H, J = 8.9 Hz, Ph-H), 7.70 (d, 2H, J = 8.9 Hz, Ph-H), 9.62 (s, 2H, CH=N). 13 C-NMR (75 MHz, DMSO-d6): δ 159.09, 142.30, 126.78, 124.09, 116.07, 66.20, 53.67, 52.40, 23.59, 22.71, 21.90. ESI-HRMS calculated for C16H23N4O+ ([M + H]+): 287.1866; found: 287.1867.</p><!><p>Chemical formula: C22H26N4O (MW = 362.48). m.p. 124–126 °C, yield 50%. 1H-NMR (300 MHz, DMSO-d6): δ 1.88–2.10 (m, 6H, NCH2CH2), 2.17–2.22 (m, 2H, NCH2), 2.56 (t, 1H, J = 6.0 Hz, NCH2CH2CH), 2.64 (t, 2H, J = 7.0 Hz, NCH2), 3.14 (t, 2H, J = 6.4 Hz, NCH2), 4.11 (t, 2H, J = 6.0 Hz, OCH2), 7.03-7.32 (m, 9H, Ph-H), 8.47 (s, 2H, CH=N). 13 C-NMR (75 MHz, DMSO-d6): δ 159.49, 146.04, 141.92, 128.46, 126.83, 126.71, 126.24, 123.98, 115.85, 66.88, 55.36, 54.42, 42.51, 33.23, 26.68. ESI-HRMS calculated for C22H27N4O+ ([M + H]+): 363.2179; found: 363.2178.</p><!><p>Chemical formula: C17H23N5O2 (MW = 329.40). m.p. 196–198 °C, yield 68%. 1H-NMR (300 MHz, CDCl3): δ 1.74–2.06 (m, 7H, NCH2(CH2)3CH), 2.14–3.00 (m, 6H, N(CH2)3), 4.07 (d, 2H, J = 6.5 Hz, OCH2), 5.70 (s, 2H, NH2), 7.02 (d, 2H, J = 8.9 Hz, Ph-H), 7.28 (d, 2H, J = 8.9 Hz, Ph-H), 8.39 (s, 2H, CH=N). 13 C-NMR (75 MHz, CDCl3): δ 177.53, 159.53, 141.87,126.30, 123.98, 115.81, 66.76, 54.96, 53.23, 42.71, 28.94, 26.71. ESI-HRMS calculated for C17H24N5O2+ ([M + H]+): 330.1925; found: 330.1926.</p><!><p>Chemical formula: C17H24ClN4O × HCl (MW = 336.86). m.p. 240–242 °C, yield 61%. 1H-NMR (300 MHz, CDCl3 + DMSO-d6): δ 1.86–2.43 (m, 8H, NCH2CH2CH2), 2.64 (s, 3H, N=CCH3), 2.94–3.55 (m, 6H, N(CH2)3), 4.22 (s, 2H, OCH2), 7.13 (d, 2H, J = 5.8 Hz, Ph-H), 7.55 (d, 2H, J = 5.8 Hz, Ph-H), 9.33 (s, 1H, CH=N), 11.31 (s, 1H, HCl). 13 C- NMR (75 MHz, CDCl3 + DMSO-d6): δ 164.95, 156.81, 149.03, 132.21, 129.15, 120.75, 70.66, 59.08, 57.74, 28.43, 27.48, 26.65, 15.04. ESI-HRMS calculated for C17H25N4O+ ([M-Cl]+): 301.2023; found: 301.2025.</p><!><p>Chemical formula: C22H26N4O (MW = 362.48). m.p. 88–90 °C, yield 59%. 1H-NMR (300 MHz, DMSO-d6): δ 1.68 (s, 2H, NCH2CH2CH2), 1.84–1.92 (m, 4H, NCH2(CH2)2, 2.18–2.30 (m, 2H, OCH2CH2), 3.18-3.25 (m, 6H, N(CH2)3), 4.13 (t, 2H, J = 6.0 Hz, OCH2), 7.00 (d, 2H, J = 8.9 Hz, Ph-H), 7.23 (d, 2H, J = 8.9 Hz, Ph-H), 7.32-7.46 (m, 5H, Ph-H), 8.45 (s, 1H, CH=N). 13 C-NMR (75 MHz, CDCl3): δ 163.64, 157.72, 150.07, 134.62, 133.40, 133.26, 132.35, 132.05, 131.39, 120.31, 70.19, 59.33, 58.11, 28.99, 28.09, 26.66. ESI-HRMS calculated for C22H27N4O+ ([M + H]+): 363.2179; found: 363.2178.</p><!><p>Chemical formula: C22H25ClN4O (MW = 396.92). m.p. 165–167 °C, yield 75%. 1H-NMR (300 MHz, CDCl3): δ 1.54 (s, 2H, NCH2CH2CH2), 1.72 (s, 4H, NCH2(CH2)2), 2.09–2.14 (m, 2H, OCH2CH2), 3.08–3.33 (m, 6H, N(CH2)3), 4.10 (t, 2H, J = 6.0 Hz, OCH2), 7.06 (d, 2H, J = 8.8 Hz, Ph-H), 7.34 (d, 2H, J = 8.8 Hz, Ph-H), 7.41 (d, 2H, J = 8.6 Hz, Ph-H), 7.48 (d, 2H, J = 8.6 Hz, Ph-H), 8.79 (s, 1H, CH=N). 13 C-NMR (75 MHz, DMSO-d6): δ 159.14, 151.96, 146.43, 135.01, 130.54, 129.21, 128.04, 127.56, 126.12, 115.85, 65.98, 54.08, 52.99, 29.45, 24.31, 23.56. ESI-HRMS calculated for C22H26ClN4O+ ([M + H]+): 397.1790; found: 397.1791.</p><!><p>Chemical formula: C28H30N4O (MW = 438.58). m.p. 66–68 °C, yield 58%. 1H-NMR (300 MHz, DMSO-d6): δ 1.73 (s, 2H, NCH2CH2CH2), 2.01–2.09 (m, 4H, NCH2CH2), 2.39–2.48 (m, 2H, OCH2CH2), 3.12–3.20 (m, 6H, N(CH2)3), 4.14 (t, 2H, J = 6.0 Hz, OCH2), 6.96 (d, 2H, J = 8.8 Hz, Ph-H), 7.20 (d, 2H, J = 8.8 Hz, Ph-H), 7.35-7.58 (m, 9H, Ph-H), 8.34 (s, 1H, CH=N). 13 C-NMR (75 MHz, DMSO-d6): δ 158.97, 153.04, 145.16, 142.55, 139.81, 128.89, 128.86, 127.91, 127.63, 127.63, 127.22, 127.01, 125.05, 115.61, 65.70, 55.17, 53.69, 24.32, 23.08, 22.30. ESI-HRMS calculated for C28H31N4O+ ([M + H]+): 439.2492; found: 439.2494.</p><!><p>Chemical formula: C15H20N4O × HCl (MW = 308.81). m.p. 85–87 °C, yield 72%. 1H-NMR (300 MHz, CDCl3): δ 1.63–1.96 (m, 6H, NCH2(CH2)2CH2), 2.95–3.07 (m, 2H, OCH2CH2), 3.48 (t, 4H, J = 5.5 Hz, N(CH2)2), 4.55 (t, 2H, J = 5.1 Hz, OCH2), 7.23 (d, 2H, J = 8.8 Hz, Ph-H), 7.75 (d, 2H, J = 8.8 Hz, Ph-H), 9.69 (s, 2H, CH=N), 11.21 (s, 1H, HCl). 13 C-NMR (75 MHz, CDCl3): δ 158.38, 142.33, 127.14, 124.20, 116.33, 63.37, 54.89, 53.01, 22.70, 21.68. ESI-HRMS calculated for C15H21N4O+ ([M-Cl]+): 273.1710; 273.1711.</p><!><p>Chemical formula: C17H24N4O × HCl (MW = 336.86). m.p. 80–82 °C, yield 69%. 1H-NMR (300 MHz, CDCl3): δ 1.88–2.21 (m, 8H, NCH2(CH2)3CH2), 2.22–2.36 (m, 2H, OCH2CH2), 2.82–3.73 (m, 6H, N(CH2)3), 4.08 (t, 2H, J = 5.8 Hz, OCH2), 7.00 (d, 2H, J = 8.9 Hz, Ph-H), 7.30 (d, 2H, J = 8.9 Hz, Ph-H), 8.49 (s, 2H, CH=N). 10.22 (s, 1H, HCl). 13 C-NMR (75 MHz, CDCl3): δ 158.93, 141.86, 126.87, 123.86, 115.89, 67.27, 56.93, 53.26, 26.37, 22.50, 21.88, 20.76. ESI-HRMS calculated for C17H25N4O+ ([M-Cl]+): 301.2023; found: 301.2021.</p><!><p>Chemical formula: C18H26N4O × HCl (MW = 314.43). m.p. 70–74 °C, yield 77%. 1H-NMR (300 MHz, CDCl3): δ 1.56–2.09 (m, 14H, NCH2(CH2)3(CH2)2CH2), 3.04–3.09 (m, 4H, N(CH2)2), 4.01 (t, 2H, J = 6.0 Hz, OCH2), 7.00 (d, 2H, J = 8.8 Hz, Ph-H), 7.30 (d, 2H, J = 8.8 Hz, Ph-H), 8.44 (s, 2H, CH=N). 13 C-NMR (75 MHz, CDCl3): δ 159.25, 141.88, 126.69, 123.91, 115.86, 67.76, 57.25, 53.28, 28.32, 23.37, 22.49, 21.92, 21.88. ESI-HRMS calculated for C18H27N4O+ ([M + H]+): 315.2179; found: 315.2180.</p><!><p>Human thalamus poly-A RNA (Clontech, Palo Alto, CA, USA) was used to clone the hH3R gene by RT-PCR. DNA PCR primers were designed in the light of the reported human histamine receptor gene sequences (GenBank accession no.AF140538). HEK-293 cells were cultured and transfected for the luciferase assay. The detailed procedures were described in the previous publication40.</p><!><p>Stable HEK-293 cells, which had been co-transfected with H3R and pCRE-Luc, were seeded in a 96-well plate overnight, and were grown to 90–95% confluence43,44. Then the cells were treated with various concentrations of tested compounds in serum-free DMEM and incubated for 20 min. Cells were then stimulated with 100 nM Histamine in serum-free DMEM containing 2 µM Forskolin and incubated for 4 h at 37 °C. Firefly luciferase assay kits (Ken-real, Shanghai, China) were used to determine the luciferase activity.</p><!><p>Valproic acid (VPA) was obtained from melongpharma, Dalian, China. Ciproxifan maleate was purchased from Shanghai Hanxiang Biotechnology Co., Ltd, China. R-(α)-methyl-histamine (RAMH), Pitolisant (PIT), and Pentylenetetrazol (PTZ) were bought from Macklin Co. KunMing mice were purchased from Changsha tianqin Biotechnology Co., Ltd, China and used with the body weight 20–25 g. Procedures involving animals were performed according to the Guide for the Care and Use of Laboratory Animals (8th Edition, National Academies Press, Washington, DC), and was approved by the local animal ethics committee (Institutional Animal Ethics Committee of Jinggangshan University, approval number: 201906018).</p><!><p>Ear stimulation with alternating current (0.2 s, 60 Hz, 50 mA) was used to induce the seizures in mice. The reduction or abolition of the hind limb tonic extension (THLE) of mice was considered protective against the MES-induced seizures29,48. Test compounds and positive drugs VPA (300 mg/kg) and PIT (10 mg/kg) were i.p. administrated half an hour prior to the electric stimulation. To investigate the mechanism of action, the most promising one 3m was chose for a further test. In one group of animals of seven mice, compound 3m (10 mg/kg) co-injected with RAMH 10 mg/kg with 5 min interval. The animals in other three groups were single treated with RAMH 10 mg/kg, compound 3m 10 mg/kg, and vehicle, respectively.</p><!><p>PTZ (85 mg/kg) was injected subcutaneously to induce seizures. Firstly, vehicle, tested compounds 3a-3q (10 mg/kg), and positive controls (VPA 300 mg/kg, PIT 10 mg/kg) were administered i.p. After 30 min, PTZ was injected to all the animals. Animals were observed for 30 min (experiment period) for any convulsion signs, and graded scores were used to assess the seizures severity. The following are the specific meaning of graded scores: score 0 = normal, score 1 = eyelids or facial twitches, score 2 = agitation with body twisting, score 3 = myo-clonic jerks or rearing, score 4 = turn over into one side position or running violently, and score 5 = turn over into back position, limbs tonic extension, or die during the experiment period30.</p><!><p>GraphPad Prism was used for statistical analysis. All data in vivo were presented as the mean ± standard error of mean (SEM). One-way analysis of variance (ANOVA), followed by the Dunnett's post-test were conducted for multiple comparisons. The statistical significance was defined as p values <0.05.</p><!><p>Crystal structure of the histamine H1 receptor (PDB ID: 3RZE) was used to construct the H3 receptor homology model45. The hH3R primary sequence was downloaded from the Universal Protein Resource (UniProt ID: Q9Y5N1). DS MODELLER (Discovery Studio 2019) was used to construct a 3D model of the H3R. Then the model was assessed in accordance with the PDF Total Energy and the Profile-3D procedure. The 3D model of H3R with the lowest PDF total energy was selected for the next docking test.</p><!><p>To consider the conformation of the protein and its ligands, Flexible Docking (Discovery Studio 2019) was used for the docking procedure. The initial ligand and water were removed, and hydrogen atoms were added. Three-dimensional structures of compounds 3h, 3m as well as PIT were generated and then placed into the protein structure during the molecular docking procedure. Interactions of the protein with 3h, 3m, and PIT were analysed.</p>

PubMed Open Access

The identification of cutin synthase: formation of the plant polyester cutin

A hydrophobic cuticle consisting of waxes and the polyester cutin covers the aerial epidermis of all land plants, providing essential protection from desiccation and other stresses. We have determined the enzymatic basis of cutin polymerization through characterization of a tomato extracellular acyltransferase, CD1, and its substrate, 2-mono(10,16-dihydroxyhexadecanoyl)glycerol (2-MHG). CD1 has in vitro polyester synthesis activity and is required for cutin accumulation in vivo, indicating that it is a cutin synthase.

the_identification_of_cutin_synthase:_formation_of_the_plant_polyester_cutin

<p>Fossil evidence suggests that evolution of a hydrophobic cuticle was essential for terrestrial colonization by plants, ~400 million years ago1. As the primary interface between the plants and their above-ground environment, the cuticle is critically important in limiting water loss and plays additional key roles in defense against pests and pathogens, as well as establishing organ boundaries during development2. The cuticle consists of an insoluble polyester of hydroxy fatty acids, known as cutin, which is covered and infiltrated with a variety of waxes. While the generic composition of the cutin polymer is known, the mechanism and site of cutin polymerization have remained long standing questions3,4.</p><p>Cutin is typically exceptionally abundant in the fruit cuticles of tomato (Solanum lycopersicum); however we previously identified several tomato mutants with dramatic deficiencies in cutin5. One of these, cutin deficient 1 (cd1), has approximately 5–10% levels of fruit cutin compared with the wild type (M82) genotype, an extremely thin cuticle and increased sensitivity to water loss and pathogen susceptibility (Fig. 1a, b; ref. 5). Fine mapping of the cd1 mutation revealed it to lie within a five-exon gene (CD1; Supplementary Results, Supplementary Fig. 1) that is predicted to encode a member of the GDSL motif lipase/hydrolase (GDSL) family of proteins (Supplementary Fig. 2). GDSLs collectively exhibit diverse functions and substrate specificities6 and a broad taxonomic distribution, including prokaryotes and eukaryotes. In plants they are present as large gene families7 and, based on their expression patterns, it has been speculated that GDSLs may play a role in cuticle biosynthesis8–12.</p><p>The cd1 mutant has a point mutation introducing a stop codon upstream of two of the three predicted catalytic amino acid residues (Supplementary Fig. 1b, Supplementary Fig. 2). In the mutant, cd1 transcript levels are reduced (Supplementary Fig. 3a) but the CD1 protein was not detected (Supplementary Fig. 3c,d), indicating that it is a null mutant. Complementation of the cd1 mutant with the wild type gene driven by the constitutive Cauliflower Mosaic Virus 35S promoter rescued the phenotype (Supplementary Fig. 4a,b), confirming that the mutation in CD1 is responsible for the cutin deficiency.</p><p>An analysis of the spatial distribution of CD1 proteins or transcripts showed that expression is highest in expanding organs, which require rapid cuticle synthesis to accommodate growth13, but is undetectable in roots, which have no cuticle (Supplementary Fig. 3b–d). Additionally, we used laser-capture microdissection of various pericarp tissues from young fruit to show that CD1 transcript levels are highest in the outer and inner epidermal cell layers (Supplementary Fig. 5; ref. 9), both of which are responsible for cuticle synthesis9. Thus, CD1 expression parallels spatial and temporal patterns of cuticle deposition at several levels.</p><p>Immunolocalization of CD1 in M82 fruits indicates that the protein is almost exclusively localized in the cuticle (Fig. 1c, Supplementary Fig. 6). More specifically, labeling density follows the contour of the cuticle over both the periclinal and anticlinal cell walls (Fig. 1d). This localization pattern suggests a role for CD1 late in the cutin biosynthetic pathway, leading us to investigate whether CD1 is directly involved in cutin polymerization.</p><p>Several enzymes have been shown through analysis of Arabidopsis thaliana mutants to be required for formation of the cutin polymer, including glycerol phosphate acyltransferase enzymes (GPATs)4. Recently, biochemical characterization of GPAT4 and GPAT6 showed them to possess both glycerol-3-phosphate acyltransferase activity specific to the sn-2 position, and phosphatase activity14. This may indicate a structural role for 2-monoacylglyceryl esters (2-MAGs) in the cutin polymer, as these were identified in small quantities in the products of partially depolymerized cutin15. Alternatively, the 2-MAG products of GPAT4 and GPAT6 may act primarily as acyl donors for the polymerization reaction. If this is true, and CD1 is indeed a cutin polymerase, we would expect that 2-MAGs would accumulate as free lipids in the surface tissues of the cd1 mutant fruit, but not in the M82 wild type genotype.</p><p>Soluble surface lipids, collectively termed cuticular waxes, can readily be extracted from plants by brief immersion of intact organs in organic solvents16. In tomato fruits, these waxes consist primarily of a mixture of high melting-point alkanes and triterpenoids, while the cutin, a polyester of principally 10,16-dihydroxyhexadecanoic acid, is insoluble under these conditions. Although soluble 2-MAGs can be found in the waxes associated with suberin, they are not observed in cuticular waxes17. GC-MS analysis identified the 2-MAG species 2-mono(10,16-dihydroxyhexadecanoyl)glycerol (2-MHG, 1) in soluble surface lipids from cd1 fruits at the rapidly expanding stage, when CD1 is normally most highly expressed (Supplementary Fig. 3c), but not in equivalent extracts from M82 fruit (Fig. 2). While chromatographic resolution was incomplete, the coincident single ion chromatograms of diagnostic fragments clearly show the specific accumulation of 2-MHG in the mutant (Fig. 2a). An additional, later-eluting trace peak of these ions likely corresponds to the thermodynamically favored 1-mono(10,16-dihydroxyhexadecanoyl)glycerol (1-MHG) isomer. The identity of the larger of the two peaks as representing the 2-isomer is confirmed by its earlier elution and the absence of the M-103 = 547 ion produced by α-cleavage between the 2- and 3- carbon in 1-MHG (Fig. 2b; ref. 15). Despite the clear accumulation of 2-MHG in the cd1 mutant and not M82, the amount detected was relatively low (on the order of 0.1 µg·cm−2, based on comparison to wax compound abundance), possibly due to feedback regulation of the upstream biosynthetic pathway or the relatively polar nature of 2-MHG compared with other surface soluble lipids.</p><p>We propose a model for cutin polymerization wherein CD1 transfers the hydroxyacyl group of 2-MHG to either another molecule of 2-MHG, or the growing cutin polymer itself (Fig. 3a). Experiments involving partial depolymerization of tomato cutin have identified oligomers primarily consisting of directly coupled 10,16-dihydroxyhexadecanoic acid monomers18,19. This, combined with the observation that glycerol is quantitatively a minor component of tomato cutin15, suggests that the principal linkage in tomato cutin is between the carboxylic acid and hydroxyl groups of 10,16-dihydroxyhexadecanoic acid. The detection of small amounts of 2-MHG in the cutin polymer15 may therefore reflect the presence of 2-MHG 'primers' remaining in the polymer. The presence of polymerized 1-MHG could be a consequence of spontaneous acyl migration accelerated by the alkaline conditions used for in vitro depolymerization. To test our hypothesis that CD1 acts as an acyltransferase, we purified recombinant tomato CD1 protein following expression in Nicotiana benthamiana (Supplementary Fig. 7). Racemic 2-MHG was synthesized in six steps from monobenzyl-protected decane-1,10-diol (Supplementary Methods, Supplementary Scheme 1), and used as a substrate for in vitro polymerization assays. Lipid products of the assay were extracted with ethyl acetate and analyzed by MALDI-TOF mass spectrometry. A major series of ions separated by m/z = 270.2 was observed, consistent with the expected masses of sodium and potassium adducts of polyester oligomers with a glycerol end group and up to seven 10,16-dihydroxyhexadecanoyl monomers (Fig. 3b). A control assay was performed using the S32A variant of CD1, prepared by site-directed mutagenesis. As expected, mutation of the conserved catalytic serine of CD1 to an alanine eliminated acyltransferase activity (Supplementary Fig. 8).</p><p>In vivo ester synthesis via transesterification of acyl glycerol by a lipase-like enzyme is not without precedent. For example, in animals, the extracellular acylation of cholesterol by transesterification of lecithin is catalyzed by lecithin cholesterol acyltransferase (LCAT). In the absence of cholesterol as an acyl acceptor, LCAT has acyl esterase activity20. The unique feature necessary for this transesterification reaction is the action of the enzyme at the lipid-aqueous interface of high density lipoproteins. Here, cholesterol concentrations are high enough to favor the resolution of the acyl-enzyme intermediate by transesterification rather than hydrolysis. We propose that CD1 acts through a similar mechanism at the interface between the aqueous environment of the plant cell wall and the lipid phase of the nascent cuticle. Thermodynamically, the aqueous solubility of 2-MHG and insolubility of the polyester product would further drive the reaction towards polyester synthesis.</p><p>In vitro incorporation of fatty acids into the cutin polymer by crude plant enzyme preparations was first reported more than thirty years ago21. Moreover, recent molecular genetic characterization of cutin polymer synthesis has identified several intracellular acyltransferases that are involved in biosynthesis of presumed cutin precursors4,22. However, the molecular basis of cutin polymerization following secretion of the precursors into the cell wall has remained a mystery. Here we show that CD1 is an extracellular enzyme that localizes in the developing cuticle and is required for cutin biosynthesis. We detected accumulation of 2-MHG, the corresponding 2-MAG of the major cutin monomer of tomato, in the cutin deficient cd1 mutant and showed that recombinant CD1 catalyzes the successive transesterification of 2-MHG to yield polyester oligomers in vitro. Taken together, these results lead us to propose that CD1 is the principal catalyst of cutin polymerization and that the polymerization process is extracellular, at the site of cuticle deposition. Furthermore, a survey of protein sequences reveals CD1 homologs in widely diverse species (Supplementary Fig. 9), and it has been reported that silencing the expression of two Arabidopsis thaliana homologs of CD1 resulted in phenotypes similar to other cutin deficient mutants23, suggesting an evolutionarily conserved and ubiquitous mechanism of cutin biosynthesis in land plants.</p>

PubMed Author Manuscript

A structural investigation of heteroleptic lanthanide substituted cyclopentadienyl complexes

The substituted cyclopentadienyl group 1 transfer agents KCp 00 , KCp 0 0 0 and KCp tt (Cp 00 = {C 5 H 3 (SiMe 3 ) 2 -1,3} À ;were prepared by modification of established procedures and the structure of [K(Cp 00 )(THF)] N ÁTHF (1) was obtained. KCp 00 and KCp tt were reacted variously with [Ln(I) 3 (THF) 4 ] (Ln = La, Ce) in 2 : 1 stoichiometries to afford monomeric [La(Cp 00 ) 2 (I)(THF)] (2aÁTHF) and the dimeric complexes [La(Cp 00 ) 2 (m-I)] 2 (2a), [Ce(Cp 00 ) 2 (m-I)] 2 (2b) and [Ce(Cp tt ) 2 (m-I)] 2 (3). KCp 0 0 0 was reacted with [Ce(I) 3 (THF) 4 ] to afford the mono-ring complex [Ce(Cp 0 0 0 )(I) 2 (THF) 2 ] (4), regardless of the stoichiometric ratio of the reagents. Complex 4 was reacted with [KN(SiMe 3 ) 2 ] to yield [Ce(Cp 0 0 0 ) 2 (I)(THF)] (5), [Ce(Cp 0 0 0 ){N(SiMe 3 ) 2 } 2 ] (6) and [Ce{N(SiMe 3 ) 2 } 3 ] by ligand scrambling. Complexes 1-6 have all been structurally authenticated and are variously characterised by other physical methods.

a_structural_investigation_of_heteroleptic_lanthanide_substituted_cyclopentadienyl_complexes

Introduction<!>Results and discussion<!>Conclusions<!>Materials and methods<!>Synthetic procedures

<p>Since the first reported synthesis of lanthanide (Ln) cyclopentadienyl (Cp) complexes in the 1950s, 1 substituted Cp ligands, Cp R , in which up to five ring protons have been replaced by various R-groups, have been employed ubiquitously in f-element organometallic chemistry. 2 Cp R ligands typically occupy three coordination sites and their steric demands are readily tunable. Bulky R-groups may be used to block undesired ligand scrambling and oligomerisation decomposition pathways. This, together with the electronically stabilising multihapto-donor properties of Cp R ligands, has been exploited in the stabilisation of Ln Cp R complexes that exhibit unusual Ln oxidation states and bonding modes. 3 Cp R ligands have been shown to be particularly effective in stabilising heterobimetallic systems that contain Ln-transition metal (TM) bonds, in which the two fragments are supported by the metalmetal interaction. 4,5 There are currently very few examples of structurally characterised Ln-TM bonds, 6 and given that recent landmark metal-metal bonds have provided step changes in our understanding of chemical bonding 7 it would be considered prudent to explore novel Ln-TM systems.</p><p>A well-established synthetic route for the synthesis of heterobimetallic Ln-TM complexes is alkane elimination. 4c However, alkali metal salt elimination pathways provide a useful synthetic alternative that can be less sluggish and produce fewer byproducts over alkane elimination in some cases. 4b,5b It follows that novel heteroleptic [Ln(Cp R ) 2 (X)] (X = halide) complexes that offer significant kinetic stabilisation and are robust with respect to ligand exchange could be useful precursors for supporting novel Ln-element bonding motifs in the future. For the larger lanthanides (La, Ce, Pr, Nd) there is a surprising lack of reports on structurally authenticated [Ln(Cp R ) 2 (X)] complexes, 4b,8 and there are very few examples that contain bromide or iodide. 9 The employment of the heavier halides is particularly advantageous in salt elimination reactions, where they are less prone to ate complex formation and can offer facile reaction workups due to the insolubility of salts such as KI in most organic solvents. 5a,b To target the synthesis of novel [Ln(Cp R ) 2 (I)] complexes of the larger Ln, we focused our attention on a family of established Cp R ligands: 1,3-bis(trimethylsilyl)cyclopentadienyl (Cp 00 , {C 5 H 3 (SiMe 3 ) 2 -1,3} À ), 1,2,4-tris(trimethylsilyl)cyclopentadienyl (Cp 0 0 0 , {C 5 H 2 (SiMe 3 ) 3 -1,2,4} À ) and 1,3-bis(tert-butyl)cyclopentadienyl (Cp tt , {C 5 H 3 ( t Bu) 2 -1,3} À ). These ligands have been previously utilised to prepare a variety of homoleptic and heteroleptic La and Ce complexes. 3d,8b-d,10 Herein we report the reactions of the group 1 ligand transfer agents KCp 00 , KCp 0 0 0 and KCp tt in salt metathesis reactions with selected Ln triiodides, leading to the stabilisation and structural authentication of novel monomeric and dimeric [Ln(Cp R ) 2 (I)] (Ln = La, Ce) complexes that are robust towards ligand exchange side-reactions over typical experimental timescales.</p><!><p>KCp 00 , 11 KCp 0 0 0 11 and KCp tt 12 (Cp 00 = {C 5 H 3 (SiMe 3 ) 2 -1,3} À ; Cp 0 0 0 = {C 5 H 2 (SiMe 3 ) 3 -1,2,4} À ; Cp tt = {C 5 H 3 ( t Bu) 2 -1,3} À ) were prepared by slight modifications of published procedures by deprotonation of the pro-ligands with KH or [KN(SiMe 3 ) 2 ]. On one occasion, crystals of [K(Cp 00 )(THF)] N ÁTHF (1) were obtained, and these were subjected to a single crystal XRD study. The crystals were of poor quality and weakly diffracting, leading to a poor quality dataset. As a result no discussion is given of the metrical parameters although the connectivity is clear-cut (see ESI ‡). KCp 00 and KCp tt were reacted separately in 2 : 1 stoichiometric ratios with [La(I) 3 (THF) Crystals suitable for X-ray diffraction studies were obtained for 2a, 2b and 3, and their solid state structures were determined (Fig. 1 space required by co-ligands. This behaviour was reported for a series of heteroleptic Cp tt cerium complexes, which exhibited three different Cp tt ring orientations in the solid state (Fig. 3A-C). 8d,10i Complexes 2a and 2b adopt orientation C in the solid state and the presence of two SiMe 3 resonances in the 1 H NMR spectra of 2a and 2b (see above) indicates that a low-symmetry orientation is maintained in solution.</p><p>On one occasion in an attempted synthesis of 2a we were able to determine the molecular structure of the monomeric THF adduct [La(Cp 00 ) 2 (I)(THF)], 2aÁTHF (Fig. 4; see ESI ‡), which resulted from incomplete removal of THF in vacuo prior to recrystallization from toluene. Complex 2aÁTHF displays similar LaÁ Á ÁCp centroid distances to its dimeric counterpart [La(1)Á Á ÁCp centroid = 2.539(4) Å mean], however the Cp centroid -La-Cp centroid angle in 2aÁTHF [127.97( 12)1] is smaller. This is due to the absence of a relatively rigid La 2 I 2 rhomboid in 2aÁTHF, which enables the two Cp 00 rings to interlock in a staggered conformation (Fig. 3D). To the best of our knowledge, this has not previously been observed for heteroleptic Cp 00 and Cp tt lanthanum and cerium complexes, but it has been reported in Cp 00 and Cp tt U(IV) chemistry. 13 In the solid state 3 exhibits a dimeric structure, with CeÁ Á ÁCp centroid distances ranging between 2.510(2) and 2.528(2) Å. These distances are comparable to those observed for [Ce(Cp tt ) 2 -(m-Cl)] 2 [2.523(10) Å mean] and [Ce(Cp tt ) 2 (m-H)] 2 [2.538(4) Å mean] 8d,10i but are longer than those observed in [Ce(Cp t ) 2 (m-I)] 2 [range CeÁ Á ÁCp centroid 2.492(5)-2.510(4) Å]. 9a Furthermore, the Cp centroid -Ce-Cp centroid angles in 3 are relatively small [124.49(7)1 and 125.29( 7)1], bringing the Cp tt rings closer to each other than the Cp 00 rings in 2a and 2b. These angles are even smaller than those observed in 2aÁTHF, in which a staggered conformation of the Cp tt rings was observed in the solid state, yet 3 exhibits a pseudoeclipsed configuration of the Cp tt rings (Fig. 3B). It is noteworthy that the 1 H NMR spectrum of 3 exhibits only one signal for the tert-butyl protons (see above), indicating that this high symmetry environment is maintained in solution. In [Ce(Cp tt ) 2 (m-Cl)] 2 the two Cp tt rings deviate even further from the ideal parallel sandwich configuration [Cp centroid -Ce-Cp centroid = 121.0(2)1], 8d highlighting the potentially large effect of crystal packing interactions on these configurations. 14 An increase in the number of sterically demanding R-groups around a Cp R ring can have significant effects on its coordination to metal centres. It has previously been shown that salt metathesis reactions between group 1 salts of the more sterically encumbered Cp 0 0 0 ligand with LaI 3 afford the mono-ring complex [La(Cp 0 0 0 )(I) 2 (py) 3 ] as the main product, 15 with the bis-Cp 0 0 0 derivative [La(Cp 0 0 0 ) 2 (I)(py)] isolated in poor yields. 12) and 3.1733( 12) Å], although this complex contains an additional molecule of THF. 16 In an attempt to prepare a heteroleptic Cp 0 0 0 -silylamide cerium complex, ''[Ce(Cp 0 0 0 )(I){N(SiMe 3 ) 2 }]'', complex 4 was reacted with one equivalent of [KN(SiMe 3 ) 2 ] (Scheme 2). An orange crystalline product was obtained from the reaction mixture, allowing the structural identification of [Ce(Cp 0 0 0 ) 2 (I)(THF)] ( 5), which has formed by ligand scrambling (Fig. 6; see ESI ‡). Complex 5 is 7)1], 8b though it is noteworthy that [Ce(Cp R ) 2 (X)] n (X = anionic ligand) complexes with larger Cp centroid -Ce-Cp centroid angles have been previously reported in the literature. 8a,9b,c Further evidence of the steric repulsion in 5 is given by the SiMe 3 substituents, which are pushed away from the cerium centre in comparison to 4 [range of Cp centroid -C-Si angles, 5: 157.4(3)-169.4(4)1; 4: 168.1(6)-168.9( 5 In order to better understand the ligand scrambling process which led to the formation of 5, several crystals obtained from the third crop of 5 were screened by single crystal XRD. Two additional crystal types were observed: orange needles and dark orange blocks. The needles were identified as [Ce{N(SiMe 3 ) 2 } 3 ] and the blocks were found to be the heteroleptic complex [Ce(Cp 0 0 0 ){N(SiMe 3 ) 2 } 2 ] (6) (Fig. 7; see ESI ‡). The structure of 6 is based on a two-legged piano stool motif and is comparable to [Ce(C 5 Me 5 ){N(SiMe 3 ) 2 } 2 ], 17 which exhibits similar Ce-N distances [2.353 (7) Å mean] to 6 [2.353(8) Å mean]. The CeÁ Á ÁCp centroid distance in 6 [2.551( 5) Å] is comparable to that observed for 4 and all other metrical parameters of 6 are unremarkable. It was not possible to fractionally crystallise analytically pure 6 under the conditions employed due to the similar solubility of [Ce{N(SiMe 3 ) 2 } 3 ] in hexanes/toluene, therefore no further analytical data were obtained. The global yield of 5 from this reaction was 50%, however, reaction conditions were not optimised as such ligand scrambling mechanisms can be difficult to control. It has been reported that mono-ring C 5 Me 5 cerium complexes can undergo ligand scrambling in solution (Scheme 3), 17 and if similar processes are occurring during the synthesis of 5 this would affect the yields of all the products.</p><p>The Evans method was employed to determine the room temperature solution magnetic moments of all isolated cerium complexes. 18 The values obtained for 2b (m eff = 2.79 m B ), 3 (2.23 m B ), 4 (2.41 m B ) and 5 (2.22 m B ) are in good agreement with the predicted magnetic moment for a Ce(III) 4f 1 2 F 5/2 ground state (2.54 m B ) 19 and previously reported magnetic moments of Ce(III) complexes in the literature (range 1.88-2.75 m B ). 20</p><!><p>We have reported the synthesis and structural characterisation of a novel series of heteroleptic substituted Cp Ln complexes using salt metathesis methodologies. These include bis-substituted Cp R Ln dimeric complexes with bridging iodides (2-3) and for the bulky Cp 0 0 0 ligand we were able to isolate a mono-ring Ce complex 4, as disubstitution was not possible for this ligand under the conditions employed. The monomeric Ce complex 5 was prepared by the reaction of 4 with [KN(SiMe 3 ) 2 ], triggering a ligand scrambling process which gave 5 in fair yield together with the mixed substituted Cp 0 0 0 silylamide Ce complex 6 and [Ce{N(SiMe 3 ) 2 } 3 ]. Although the yield of 5 is only moderate, it is considerable when compared to the previously reported low-yielding synthesis of the closely related La complex, [La(Cp 0 0 0 ) 2 (I)(py)]. 8b Solid state characterisation of 2-6 provides in-depth knowledge of the structural features of these systems and this will help us to predict their reactivity profile. In our hands, complexes 2a-b and 5 have not shown any tendency to ligand scramble in both coordinating and non-coordinating solvents over experimental timescales. As such, we envisage that they will be useful precursors for the preparation and stabilisation of novel La/Ce-element bonding motifs, including the preparation of heterobimetallic species, by salt metathesis methodologies.</p><!><p>All manipulations were carried out using standard Schlenk and glove box techniques under an atmosphere of dry argon. Solvents were dried by refluxing over potassium and were degassed before use. All solvents were stored over potassium mirrors (with the exception of THF, which was stored over activated 4 Å molecular sieves). Deuterated solvents were distilled from potassium, degassed by three freeze-pump-thaw cycles, and stored under argon. KCp 00 , 11 KCp 0 0 0 , 11 KCp tt , 12 [Ln(I) 3 (THF) 4 ] (Ln = La, Ce) 21 and [KN(SiMe 3 ) 2 ] 22 were prepared according to published procedures. KH was obtained as a suspension in mineral oil and was washed three times with hexane and dried in vacuo. 1 H, 13 C{ 1 H} and 29 Si{ 1 H} NMR spectra were recorded on a spectrometer operating at 400.2, 100.6 and 79.5 MHz respectively; chemical shifts are quoted in ppm and are relative to TMS. FTIR spectra were recorded as Nujol mulls in KBr discs on a PerkinElmer Spectrum RX1 spectrometer. Elemental microanalyses were carried out by Stephen Boyer at the Microanalysis Service, London Metropolitan University, UK.</p><!><p>[La(Cp 00 ) 2 (l-I)] 2 (2a). A Schlenk flask was charged with KCp 00 (1.74 g, 7 mmol) and [La(I) 3 (THF) 4 ] (2.83 g, 3.5 mmol). The flask was cooled to À78 1C and THF (15 ml) was added dropwise with stirring. The yellow reaction mixture was allowed to warm slowly and stirred for 72 h, during which time a suspension formed. The mixture was allowed to settle for 2 hours and the suspension was filtered. The solvent was removed in vacuo (10 À2 mbar) and the solid residue extracted with toluene (10 ml). The solution was concentrated to approximately 3 ml and stored at À30 1C to afford 2a as colourless crystals (1.16 g, 48%). On one occasion crystals of 2aÁTHF were found. Anal. calcd (%) for C 44 H 84 La 2 I 2 Si 8 : C, 38.59; H, 6.18. Found: C, 38.47; H, 6.28 [Ce(Cp 00 ) 2 (l-I)] 2 (2b). A Schlenk flask was charged with KCp 00 (4.97 g, 20 mmol) and [Ce(I) 3 (THF) 4 ] (8.01 g, 10 mmol). The flask was cooled to À30 1C and THF (20 ml) was added dropwise with stirring. The reaction mixture was allowed to slowly warm to room temperature and then stirred for a further 16 hours. The mixture was allowed to settle for 2 hours and the suspension was filtered, giving a clear yellow solution. Volatiles were removed in vacuo (10 À2 mbar), affording a bright pink solid. Recrystallisation from toluene (30 ml) afforded 2b as a crystalline product (2.85 g, 42%). [Ce(Cp tt ) 2 (l-I)] 2 (3). A Schlenk flask was charged with KCp tt (0.82 g, 4 mmol) and [Ce(I) 3 (THF) 4 ] (1.62 g, 2 mmol). The flask was cooled to À78 1C and THF (15 ml) was added dropwise with stirring. The yellow reaction mixture was allowed to slowly warm to room temperature and then stirred for a further 16 hours. The mixture was allowed to settle for 2 hours and the suspension was filtered. Volatiles were removed in vacuo (10 À2 mbar), affording a bright orange solid. The solid residue was extracted with toluene (8 ml) and stored at room temperature, affording 3 as an orange crystalline product (0.30 g, 24%). 1 [Ce(Cp 0 0 0 )(I) 2 (THF) 2 ] (4). A THF (20 ml) solution of KCp 0 0 0 (1.60 g, 5 mmol) was added dropwise to a THF (15 ml) slurry of [Ce(I) 3 (THF) 4 ] (4.05 g, 5 mmol) at À4 1C with stirring. A colour change to yellow with a white precipitate was observed, and the mixture was slowly warmed to room temperature and stirred for 72 hours. Volatiles were removed in vacuo, yielding a bright yellow powder. The solid was extracted with hexane (20 ml) and volatiles were removed in vacuo (10 À2 mbar), affording 4 as a yellow powder (2.39 g, 58%). A small crop of crystals were grown from hexanes (2 ml) at À25 1C. [Ce(Cp 0 0 0 ) 2 (I)(THF)] ( 5) and [Ce(Cp 0 0 0 ){N(SiMe 3 ) 2 } 2 ] (6). A toluene (20 ml) solution of [KN(SiMe 3 ) 2 ] (1.00 g, 5 mmol) was added dropwise to a toluene (20 ml) solution of 4 (4.10 g, 5 mmol) at À20 1C with stirring. The reaction mixture was warmed to room temperature and stirred for 16 hours, forming a dark orange mixture with a white precipitate. The suspension was filtered, the solution concentrated to 5 ml and cooled to 4 1C, yielding orange crystals of 5. More crops of 5 were obtained from the supernatant liquid of the first crystallisation (2.36 g, 50%). From the third recrystallization several crystals with different morphologies were observed. These were identified as [Ce(N{SiMe 3 } 2 ) 3 ] (orange needles) and 6 (orange blocks) via single crystal X-ray studies. Data for 5: 1 H NMR (d 6 -benzene, 298 K) d = À5.75 (s, 2H, Cp 0 0 0 -CH), À5.34 (br s, 2H, Cp 0 0 0 -CH), 0.30 (s, 36H, Si(CH 3 ) 3 ), 0.90 (s, 18H, Si(CH 3 ) 3 ). m eff (Evans method, 298 K, d C, 42.64; H, 7.27. Found: C, 42.72; H, 7.38 (%). FTIR (Nujol, cm À1 ): n = 1249 (s), 1090 (s), 986 (s), 935 (m), 837 (br s), 753 (m).</p>

Royal Society of Chemistry (RSC)

REINVENT 2.0 -an AI tool for de novo drug design

In the past few years, we have witnessed a renaissance of the field of molecular de novo drug design. The advancements in deep learning and artificial intelligence (AI) have triggered an avalanche of ideas how to translate such techniques to a variety of domains including the field of drug design. A range of architectures have been devised to find the optimal way of generating chemical compounds by using either graph or string (SMILES) based representations. With this application note we aim to offer the community a production-ready tool for de novo design, called REINVENT. It can be effectively applied on drug discovery projects that are striving to resolve either exploration or exploitation problems while navigating the chemical space. It can facilitate the idea generation process by bringing to the researcher's attention the most promising compounds. REINVENT's code is publicly available at https://github.com/MolecularAI/Reinvent

reinvent_2.0_-an_ai_tool_for_de_novo_drug_design

Introduction<!>Application Overview<!>Diversity Filters<!>Reinforcement Learning<!>Scoring Functions<!>Transfer Learning (TL)<!>Logging<!>Implementation<!>Conclusion

<p>The main goal of de novo drug design is to identify novel active compounds that can simultaneously satisfy a constellation of essential optimization goals such as activity, selectivity, physico-chemical and ADMET properties. Because of the sheer number of possible solutions, it is a non-trivial task to optimally satisfy such a multitude of requirements which makes the search process slow and costly even when it is only conducted in silico. Therefore, having an efficient solution which enables the navigation of chemical space and generation of relevant ideas is essential. To address such needs the research community has recently turned its focus towards artificial intelligence (AI) based generative models that are capable of proposing promising small molecules. The potential of generative models for chemical space exploration has been demonstrated in numerous studies [1][2][3][4][5][6][7][8][9][10][11][12][13] . Various neural network architectures have been engineered and a plethora of AI training strategies have been employed in the race to device more efficient methods for the generation of compounds. A number of architectures, such as Variational Autoencoders (VAEs) 7,14 , Recurrent Neural Networks (RNNs) with Long Short-Term Memory (LSTM) cells 15 , Conditional RNNs or Generative Adversarial Networks have been proven successful in generating molecules by using data representation of molecules either as molecular graphs or SMILES 8,[16][17][18] .</p><p>Most tools for de novo drug design, regardless of the specifics of their implementation, can be generalized to three main components: search space (SS), search algorithm and a search objective 19 . In this context we can refer to the generative models as the search space. We also observe two main trends of using generative models for de novo design: distribution-learning and goal-directed generation. Distribution-learning efforts are mostly focused on generating ideas that resemble a particular set of molecules. Goal-directed generation methods are typically using search algorithms while aiming to suggest molecules that satisfy the given objective (or objectives) without having to sample the entire search space. In both cases results are ultimately filtered by user defined scoring function (search objective) either during the generation in the goal-driven case or after sampling the entire set of solutions in the distribution-learning scenario.</p><p>While a common issue of using goal-directed approach is the narrow set of solutions the opposite approach of using distribution learning leads to screening through a vast variety of irrelevant suggestions. These two extreme scenarios represent the attempt to achieve either exploration or exploitation of the search space.</p><p>There is an increasing variety of open-source solutions based on generative models aiming to address these two aspects of de novo design separately 2,9,20,21 . Ideally, users should be allowed to navigate the chemical space efficiently in both exploration and exploitation mode while using the same de novo design tool. For exploitation, users define an area of interest and focus on generating compounds that share similar structural features. In contrast, the exploration mode enables them to obtain compounds that share less structural similarity but still satisfy other desired features. To achieve this in a fashion different from plain distribution-learning approach we need the goal-directed learning to store in memory and adapt to the suggested solutions that have been produced in the course of a single search run. This implies the necessity to utilize not only predictive models and structure similarity/dissimilarity but also various rule-based scoring components to push towards or pull away from specific areas of the chemical space. Moreover, to be able to adapt appropriately to any given drug discovery project at hand, the ability to finetune each of these potential scoring function components is paramount.</p><p>In this application note we are describing REINVENT 2.0 which is a tool for de novo design of small molecules. REINVENT As a de novo design application REINVENT 2.0 covers both distribution-learning and goal-directed scenarios. The goal-directed use case uses a generative model as a SS, RL as a search algorithm and flexible scoring function that can combine the scores from different components to form a reward as a score objective. The calculations in the individual components can be run in parallel.</p><p>Scores can be also modulated by a diversity filter which penalizes redundancy and rewards diversity in the found solutions thus stimulating exploration. Comprehensive logging is implemented for each use case. Additionally, the option to send logs to a remote REST endpoint is also available which allows to put the application behind a web interface. REINVENT 2.0 can be also used to build generative models from scratch. For details on the generative model, please refer to Arus-Pous et al. 23 More details on these features are provided in the sections below.</p><!><p>In its core, REINVENT is using a generative model. The generative model has an architecture derived from the work of Arus-Pous et al 23 which in turn is inspired by Segler et al 6 REINVENT provides different running modes listed in table S1. Different combinations of the running modes allow the users to achieve either exploitation or exploration of the chemical space, see table S2. Further discussion on the general use cases can be found in the supporting materials.</p><p>One of the key features that allow achieving an exploratory behavior are the Diversity Filters.</p><!><p>DF can be regarded as a collection of buckets that are used for keeping track of all generated scaffolds and the compounds that share those scaffolds. A bucket is a collection of compounds that share the same scaffold. Obviously, not all generated compounds are of interest and only those that are ranked by the multi-parameter objective (MPO) score above a certain user-defined threshold will enter the scaffold buckets. When the average score settles above this threshold we have reached state of productivity. This means that the majority of the compounds from each step (the ones that score above threshold) will be collected and stored in the memory. Once a compound with a score above the threshold has been generated, its scaffold is extracted and stored in a scaffold registry and the compound enters the corresponding bucket. The buckets have limited capacity and once the limit of compounds in a given bucket has reached the allowed threshold, any subsequent bucket affiliation will be penalized. Every new compound that enters a full bucket will be assigned a score of zero thus informing the agent that this area of chemical space has become unfavorable. It is important to note that compounds will be added to the bucket even if the bucket limit has been exceeded. The only impact will be on the agent, since it will be constantly discouraged from producing similar compounds that share a given scaffold. This will enforce the agent to seek alternative solutions thus achieving in effect chemical space exploration and will prevent the agent from becoming stuck in local minima and generating the same compounds repeatedly. All collected compounds are kept and stored until the end of the RL run and become available as a csv formatted file.</p><p>Users can select their diversity strategy by using Topological DF 28 , Identical Murcko DF or a Scaffold Similarity DF 29 . The Topological DF is the most restrictive since it is agnostic of the atom types. It is created by removing all side chains and subsequently converting all atoms in the structure to sp3 carbons. The other two DF also remove all side chains but retain the atom types.</p><p>Identical Murcko DF only checks if there is a bucket with exactly the same scaffold while Scaffold</p><p>Similarity is more permissive and can include compounds into the bucket if they satisfy a certain threshold of scaffold similarity. The threshold is user defined and is sensitive to the discrete definition of the scoring function. Setting it to higher values would clearly result in less compounds passing the threshold.</p><!><p>It is often necessary to direct the generative model towards relevant areas in the chemical space that contain compounds of interest. We achieve this by subjecting it to a RL 25 scenario while aiming to satisfy a set of user-defined requirements that reflect the most important features of the desired compounds. In other words, the generative model will try to maximize the outcome of a scoring function that contains multiple components/parameters, thus computing an MPO score 30 . To generate compounds from a specific part of the chemical space, REINVENT employs a composite scoring function consisting of different user-defined components. Each component is responsible for a simple target property. The feedback from the scoring function is used in a RL loop with a policy iteration as described by Olivecrona et al. 22 .</p><p>The components of the RL loop used in REINVENT are shown in figure S2. Commonly the RL setup consist of an actor and an environment in which the actor takes a set of actions and receives a reward. The reward reflects how well the actor solved the problem at hand. The set of actions is referred to as policy, and the reward after completing the policy is known as a policy iteration. In our case, the actions are the individual steps necessary for building sequences of tokens which translate into SMILES. The role of the environment is played by the score modulating block in figure S2 and the actor is denoted as an "agent". After the agent samples a batch of smiles the reward is influenced by several components: scoring function, "prior" and a diversity filter.</p><p>The "prior" is a generative model which shares identical architecture and vocabulary with the agent. It possesses a great generative capacity and the potential to sample compounds from a comparably vast area of the chemical space. Essentially, the prior is the same as the agent at the beginning of the RL. There is, however, a use case where the agent might be subjected to initial transfer learning in which case the models will have different weights. Further details are described in workflow "E" table S2. The role of the prior is to serve as a reference point for the likelihood of sampling a given SMILES. For every batch of SMILES generated by the agent, the prior calculates the negative log-likelihood denoted as NLL (eq. 1). NLL reflects how likely it is to</p><p>Analogously the NLL(S) for the given string S is also calculated by the agent. The SMILES string is also evaluated by the scoring function which we denote as a multi-parameter objective (MPO).</p><p>MPO is a value in the range [0,1]. At this step the DF is used to evaluate whether the SMILES string has been sampled before or whether it satisfies the DF policy. The MPO score will be set to 0 if the DF filters determine that the provided compound already exists or if there are too many compounds of the same scaffold and their number exceeds the user defined threshold. For more details on the types of DF please consult with the supporting materials table S4. The resulting MPO score is combined with the prior's likelihood and used to form the augmented likelihood (eq 2). The MPO score is multiplied by σ which is a scalar value used for scaling up the scoring function output to the same order of magnitude as the NLL. Otherwise, the low MPO score ranging between [0,1] will have no impact whatsoever. The higher MPO score translates into higher augmented likelihood values. Ultimately, the loss is calculated as the squared difference between the agent's likelihood and the augmented likelihood (eq 3).</p><p>𝑁𝐿𝐿(𝑺) 𝐴𝑢𝑔𝑚𝑒𝑛𝑡𝑒𝑑 = 𝑁𝐿𝐿(𝑺) 𝑃𝑟𝑖𝑜𝑟 − 𝛔 * 𝑀𝑃𝑂(𝑺) 𝑠𝑐𝑜𝑟𝑒</p><p>(2)</p><p>The final component of the RL loop as shown on figure S2 is inception. The purpose of inception is to keep track of previously well scored compounds and to randomly expose a subset of them to the agent thus helping to direct the learning. More details about inception are provided in the supporting information. Finally, after including the compounds from inception's memory to the batch the loss is propagated back and only the agent is updated thus receiving the feedback from its interaction with the environment. The environment is represented by the score modulating block on figure S2. The prior on the other hand does not undergo any changes.</p><p>The duration of RL is pre-defined by the user in terms of number of RL steps to be performed. This is very case specific and is normally determined by the complexity of the problem at hand. In terms of computational cost the scoring function is the costliest element of the RL loop since it may contain a variable number of components including slow predictive models, docking and/or pharmacophore similarity (the latter two are not included in the current release).</p><!><p>REINVENT offers two general scoring function formulations (eqs 4 and 5). The individual components of the scoring function can be either combined as a weighted sum or as a weighted product 24 . The individual score components can have different weight coefficients reflecting their importance in the overall score. Score contribution from each component can vary in a [0,1] range. As a result, the overall score is also within the same [0,1] range. In the equations below the score for sequence x is denoted S and is either a weighted product (eq 5) or a weighted sum (eq 4). The user-selected components are denoted as p in both equations and the corresponding weights are denoted as w. Weights can vary in the range of [1, +∞).</p><p>Both formulations are provided for user convenience and flexibility. More details on the scoring functions and a full list of components included in this release is provided in table S3 and in the supporting materials.</p><!><p>As an alternative to the goal-directed generation, distribution-learning is also supported in REINVENT. This approach requires a pre-trained generative model with the generative capacity and the potential to sample compounds from a rather vast area of the chemical space. We refer to this generative model as the prior. This prior is subjected to transfer learning with a smaller set of compounds which are relevant for a given project. For example, if we aim to maximize a predictive model among the other components we would use all the compounds that are considered as active by this model. If we aim towards certain subseries of compounds we would only use those that share the series-specific features (for example scaffold). This will result in a model that produces compounds similar to the target dataset with a higher probability. We refer to that model to as "focused prior". The user can subsequently sample this model and score the generated compounds by using the scoring mode in REINVENT.</p><p>As an alternative we could also use the resulting "focused prior" as an agent in the RL loop. The resulting generative model from distribution-learning is a suitable starting point for goal-directed generation 31 . This pre-focusing of the prior can speed up the overall RL process since the chance of producing compounds of relevance will be much higher compared to using a general, unfocused prior as an agent. Once focused, the agent will have an increased probability of sampling a chemical subspace of interest thus reaching a state of productivity sooner. Both use cases are further illustrated in figure S3 and table S2 within the supporting information.</p><!><p>Essential for monitoring of the learning process is the availability of a comprehensive logging system. In REINVENT we utilize Tensorboard 32 to provide information about the evolution of the agent during TL by sampling after each step and displaying the likelihood distribution for the sampled data. Stats on validity of the smiles and the most frequently encountered molecules are also shown. For RL we are plotting the evolution of the scoring function and the individual scoring component contributions to the overall score. We are also displaying the highest scoring compounds after each RL step. As an alternative, we also provide the implementation used by us for remote logging which can be set up to post the logging results to a custom REST endpoint.</p><!><p>REINVENT is an open-source Python application. It uses PyTorch 1.</p><!><p>We have described a production-ready, open-source application for de generation of small molecules. It can be used to address both exploration and exploitation type of problems while allowing a flexible formulation of complex MPO scores. Examples of various use cases are provided with the code repository and in https://github.com/MolecularAI/ReinventCommunity.</p><p>Apart from providing a ready-to-use solution, with releasing the code, we are hoping to facilitate the research on using generative methods for drug discovery. We also hope that it can be used as an interaction point for future scientific collaborations.</p>

ChemRxiv

New functional aspects of the atypical protein tyrosine phosphatase VHZ

LDP3 (VHZ) is the smallest classical protein tyrosine phosphatase (PTP) known to date, and was originally misclassified as an atypical dual specificity phosphatase (DSP). Kinetic isotope effects with steady state and pre-steady state kinetics of VHZ and mutants with para-nitrophenol phosphate (pNPP) have revealed several unusual properties. VHZ is significantly more active than previously reported, but remains one of the least active PTPs. Highly unusual for a PTP, VHZ possesses two acidic residues (E134 and D65) in the active site. D65 occupies the position corresponding to the typical general acid in the PTP family. However, VHZ primarily utilizes E134 as the general acid, with D65 taking over this role when E134 is mutated. This unusual behavior is facilitated by two coexisting, but unequally populated, substrate binding modes. Unlike most classical PTPs, VHZ exhibits phosphotransferase activity. Despite the presence of the Q-loop that normally prevents alcoholysis of the phosphoenzyme intermediate in other classical PTPs, VHZ readily phosphorylates ethylene glycol. Although mutations to Q-loop residues affect this phosphotransferase activity, mutations on the IPD-loop that contains the general acid exert more control over this process. A single P68V substitution on this loop completely abolishes phosphotransferase activity. The ability of native VHZ to catalyze transphosphorylation may lead to an imbalance of intracellular phosphorylation, which could explain the correlation of its overexpression with several types of cancer.

new_functional_aspects_of_the_atypical_protein_tyrosine_phosphatase_vhz

<!>Protein cloning, expression and purification<!>Cloning and purification of SsoPTP<!>Preparation of p-nitrophenyl phosphate (pNPP)<!>Quantification of inorganic phosphate<!>Determination of inhibition constants<!>Steady-state kinetic analysis<!>Isotope effect measurements<!>Detection of phosphotransferase activity in the presence of ethylene glycol<!>Pre-steady state kinetics<!>Computational analysis of pNPP binding to VHZ<!>Buffer and Oxyanion Inhibition<!>Kinetic Isotope Effects<!>Steady-state kinetics analysis<!>pH dependency for the hydrolysis of pNPP by VHZ<!>Pre-steady state kinetics<!>Reaction in the presence of ethylene glycol<!>Discussion<!>VHZ is inhibited by commonly used buffers, and by inorganic phosphate more strongly, than typical PTPs<!>The leaving group is not fully neutralized in the first step of the VHZ reaction<!>Two potential and functional general acids in the active site of VHZ<!>D65 has a role primarily in the second step, and affects substrate binding<!>VHZ has a low catalytic efficiency and different rate-limiting step compared to most PTPs<!>VHZ has two substrate binding modes for pNPP<!>VHZ exhibits phosphotransferase activity controlled by a single residue in an unexpected location<!>Conclusions<!>

<p>The protein tyrosine phosphatases (PTPs) are a large family of enzymes responsible for intracellular dephosphorylation. Together with protein tyrosine kinases (PTKs), PTPs control the level of protein phosphorylation, which modulates numerous aspects of cell life, such as growth, proliferation, metabolism, intercellular interaction, immune responses, and gene transcription (1). PTPs contain a highly conserved HCXXGXXRS/T signature sequence motif but share very little sequence similarities outside of the conserved regions, which are comprised of the phosphate binding loop (P-loop); the general acid loop, often referred to as the WPD-loop; and the Q-loop that bears conserved glutamine residues that orient the water nucleophile in classical PTPs and prevent phosphotransferase activity to other potential nucleophiles.</p><p>All PTPs utilize a two-step double-displacement mechanism of phosphate monoester hydrolysis (Scheme 1) mediated by an invariant cysteine-arginine-aspartic acid triad of catalytic residues (2). The mechanism proceeds through a phosphoenzyme intermediate where the second chemical step is often rate limiting (3). In the first step the P-loop orients the substrate as the nucleophilic cysteine attacks phosphorus with simultaneous expulsion of the leaving group protonated by the catalytic general acid. In the second step a water molecule, directed by the aspartic acid residue that served as the general acid in the first step and Q-loop glutamine residues, attacks the phosphoenzyme intermediate.</p><p>The PTP family is subdivided into several groups based on substrate specificity, subcellular localization, and size. The classical PTP family selectively hydrolyzes phosphotyrosine containing peptides, and includes the well-studied bacterial effector protein YopH, responsible for the virulence of notorious Y.pestis, and human PTP1B, which plays an important role in insulin signaling (4). Classical PTPs have a modular organization and, in addition to the catalytic phosphatase domain, contain non-catalytic domains that control subcellular localization and protein-substrate interactions. All classical PTPs are tyrosine specific enzymes. The members of the dual-specificity phosphatases (DSPs) subfamily hydrolyze phosphoserine and phosphothreonine in addition to phosphotyrosine containing target sites. Within the DSP subfamily, the atypical DSPs are smaller and contain only a catalytic domain (5). The classical PTPs and DSPs also differ in their phosphotransferase ability. In classical PTPs the phosphoenzyme intermediate is attacked only by water due to the shielding effect of conserved Q-loop residues, named for the presence of conserved glutamines (6). In contrast, DSPs such as VHR, and the low-molecular weight LMW-Ltp1, both of which lack the Q-loop, display significant phosphotransferase ability (7). On this basis, it has been concluded that the presence of the Q-loop prevents phosphotransferase activity.</p><p>VHZ, and the closely related phosphatase S.solfataricus PTP (SsoPTP), are among the smallest classical PTPs known to date (Figure 1). The SsoPTP (161 amino acids) is similar to VHZ (150 amino acids) in size and catalytic activity. Both VHZ and SsoPTP consist of a single, catalytic domain that is more similar to classical PTPs than DSPs (8, 9) and contain identical secondary structural elements, but, unlike most classical PTPs, lack an N-terminal extension that forms a substrate recognition/binding loop. Like VHZ, the general acid in SsoPTP resides on a rigid IPD-loop, which, unlike the flexible WPD-loop in classical PTPs, permanently occupies a closed conformation. Unlike VHZ, and like classical PTPs, SsoPTP/WT contains no additional general acid in its Q-loop region. VHZ was originally classified as an atypical DSP and named after its prototypical member as VH1-related protein member Z. In previous work, we presented results indicating that VHZ should be classified as a PTP rather than a DSP, on the basis of a structural analysis and results of a phosphopeptide substrate screen in which VHZ showed activity against pY–containing peptides but not toward pS- or pT-peptides (8).</p><p>In the present work, we show that the catalytic activity of VHZ was significantly underestimated in previous reports, as a result of pronounced product inhibition, and the inhibitory effect of certain buffers. Despite much in common with classical PTPs, VHZ is highly unusual in possessing two acidic residues in the active site, D65 and E134. Our results indicate that under certain circumstances either of these residues can serve as the general acid in the first step of the reaction. We also present results demonstrating that VHZ, despite the presence of a Q-loop, catalyzes phosphoryl transfer to alcohols (alcoholysis) in addition to water (hydrolysis) (Scheme 2).</p><p>The mutagenesis of several residues in VHZ in parallel with SsoPTP has revealed that, in addition to the Q-loop, particular residues in the general acid IPD-loop play a crucial role in nucleophilic selectivity. A combination of kinetics and mutagenesis experiments have revealed unusual aspects of the kinetic behavior of VHZ and given insights into factors that control the phosphotransferase activity of VHZ, and possibly in other PTPs.</p><!><p>VHZ mutants were made using the Qiagen QuikChange Lightning Site-Directed Mutagenesis Kit. VHZ and mutants were purified as previously described (8). A His-tagged version of VHR was prepared as follows. In the first step the gene of VHR was amplified from pT7-7 plasmid using the following primers: Fwd1: GAA AAC CTG TAT TTT CAG GGC ATGTCGGGCT CGTTCGAGCT, Rev1: GGA GAG CTC CTA GGG TTT CAA CTT CCC CTC CTT GGC TAG to incorporate TEV protease cleavage site (Fwd1, in bold) immediately upstream of the protein gene, and Sac-I restriction site (Rev1, in bold) was added at the end of the gene sequence. In the second step, the product of the first PCR step was used as a template and a KpnI restriction site was added upstream of the TEV protease cleavage sequence using the following set of primers: Fwd2: CGGGGTACCGAAAACCTGTAT, Rev1: GGA GAG CTC CTA GGG TTT CAA CTT CCC CTC CTT GGC. The resulting PCR product was digested with Kpn-I and Sac-I (Fermentas) and ligated into the pet-45(B+) vector (Novagen) pre-digested with the same set of restriction enzymes. The E. coli DH5α competent cells were transformed with 5 µL of the ligation mixture and plated on an ampicillin-containing agar plate. DNA sequencing confirmed the presence of the desired gene. BL-21 DE-3 (codon+) E. coli competent cells were transformed with the peT-45(B+) -VHR vector. 10 mL of LB media were inoculated with a single colony and incubated at 37°C on a shaker overnight. 1L of 2xYT media containing ampicillin and chloramphenicol were inoculated with 10 mL of overnight cell growth. The cells were grown at 37°C until OD600nm reached 1.2–1.5 a.u. 100 mg of IPTG were added (final concentration 100 mg/L), the flask was transferred to a room temperature shaker and incubated for 18–20 hours. The cells were harvested by centrifugation and resuspended in Ni loading buffer containing 50 mM Tris, 500 mM NaCl, 20 mM imidazole, 5 mM 2-mercaptoethanol, 5 % glycerol, pH 8.0. Cells were sonicated on ice, and after centrifugation the supernatant was decanted and filtered through 0.45 micron Millipore syringe filter. Ni-fast flow High Affinity Resin (GE Healthcare) was washed 3 times with Ni loading buffer and incubated with supernatant for 30 min at 4° C with gentle shaking. The resin slurry was transferred to a 15 mL glass chromatography column. The column was extensively washed with Ni loading buffer to reach baseline absorbance and protein was eluted with a 120 mL linear gradient with buffer containing 50 mM Tris, 500 mM NaCl, 20 mM imidazole, 2-mercaptoethanol, 5 % glycerol, pH 8.0 at 2 mL/min. The fusion VHR protein obtained was dialyzed stepwise in TEV-cleavage buffers containing 50 mM Tris-Base, 0.5 mM EDTA, 3 mM DTT, 5% glycerol and reducing concentration of NaCl (300 mM, 150 mM, 0 mM), at 1.5 hour per step followed by addition of 2 mg of TEV protease. The cleavage of VHR was rapid and complete in two hours at 4 ° C. VHR was dialyzed in the original Ni-loading buffer to remove DTT and EDTA for two hours, and the solution was passed through Ni-Fast flow High Affinity Resin to remove cleaved polyhistidine tag and TEV-protease. Collected supernatant was concentrated to 3–5 mL using Millipore Centrifuge Concentrator tube and loaded on Superdex 75 26/60 gel filtration column (GE-Healthcare) pre-equilibrated with buffer containing 50 mM HEPES, 150 mM NaCl, 50 mM Imidazole, 5 mM DTT, 10 % glycerol. A single peak corresponding to the VHR was collected, concentrated to 6 mg/mL, flash-frozen in the liquid nitrogen and stored at −80 °C.</p><!><p>The S.solfataricus PTP gene sequence from Uniprot (ID Q97VZ7) was optimized for the E. coli expression system and synthesized by GenScript delivered in a shuttle vector. The gene was ordered with a KpnI restriction site followed by TEV cleavage site directly upstream of the SsoPTP sequence and a SacI restriction site directly downstream of the stop codon. The shuttle vector containing the described gene was amplified in E. coli DH5α, isolated, and digested with KpnI –SacI restriction enzymes in the Green Buffer (Fermentas). The digestion products were separated by electrophoresis using 1% agarose gel. The band correspondent to the desired gene was extracted using Qiagene Gel Extraction Kit and ligated into pet45 (B+) plasmid pre-digested with the same set of restriction enzymes. Unfortunately, unlike VHR (and similar to VHZ(8)), the resulting fusion protein was resistant to TEV cleavage, so the poly-His tag and TEV-cleavage site was removed from the pet45B(+)-SsoPTP vector by one-step overlapping PCR using the following primers: Fwd: TATACCATGTACTGGGTCCGTCGCAAAACG; Rev: GACCCAGTACATGGTATATCTCCTTCTTAAAGTAAACAAAATTATTCTAG. Tagless SsoPTP was expressed in E. coli BL-21 DE-3 (codon+) cells. The growth and expression conditions were analogous to VHZ and VHR. The cells were harvested by centrifugation and resuspended in the buffer containing 50mM Tris, 1mM EDTA, 10 mM DTT, 5% glycerol pH 7.4 (4°C). Cells were disrupted by sonication on ice, and the pellet was separated by centrifugation. The supernatant was treated with a 10 % w/v solution of polyethyleneimine (PEI, Mw 50000, pH 8.0 at 4 °C) added dropwise at 4°C to reach a final PEI concentration of 0.5% (w/v). After centrifugation the supernatant was loaded at 1mL/min on a Q-HiTrap (GE Healthcare) column pre-equilibrated with buffer containing 50 mM sodium acetate, 1 mM EDTA, 3 mM DTT, pH 5.5. After washing the protein was eluted with a 120 mL linear gradient (2 mL/min) buffer containing 50 mM sodium acetate, 1 mM EDTA, 600 mM NaCl, 3 mM DTT, pH 5.5. A single isolated peak eluted at 200–250 mM NaCl corresponding to SsoPTP, concentrated to 3–5 mL and loaded on a Superdex 75 26/60 gel filtration column (GE-Healthcare) pre-equilibrated with buffer containing 25 mM Tris, 50 mM imidazole, 150 mM NaCl, 1 mM EDTA, 5 mM DTT, pH 7.5 (at 4°C). A single peak was collected, concentrated to 10 mg/mL (A280 of a 1 mg/mL solution = 2.36) and flash frozen in liquid nitrogen in the same buffer with 25% of glycerol added.</p><p>YopH, PTP1B and their mutants were expressed and purified as previously described (10, 11).</p><p>All enzymes were purified to 99% + homogeneity based upon SDS PAGE analysis (data not shown). Concentration of each enzyme was determined by measuring absorbance at 280 nm using calculated extinction coefficients.</p><!><p>The dicyclohexylammonium salt of pNPP was synthesized as previously described (12). A phosphate free solution of pNPP was prepared in two separate steps. The crude product was dissolved in a minimal volume of 0.8 M NaOH. Neutral cyclohexylamine was extracted with five equal portions of chloroform, accompanied by reduction of pH to 8.0–8.5. Inorganic phosphate was precipitated by addition of a solution containing 1M MgCl and 5 M NH4Cl (1:5 ratio of Mg2+/NH4+) to reach 0.1 M final magnesium concentration. The pH was adjusted to 9.0–9.3 with ammonium hydroxide and stirred for 10 minutes until the solution turned cloudy due to the precipitation of MgNH4PO4. The concentration of pNPP in solution remains essentially unchanged, in contrast to calcium or magnesium chloride precipitation alone, which mostly precipitates pNPP. The precipitate was removed by filtration using a fritted glass funnel, and 10 g of pre-activated Chelex100 resin suspension in water (pH 9.0) was added to scavenge the remaining magnesium. The resin was removed by filtration, and cyclohexylammonium hydrochloride (Fisher) solution (pH 8–9) was added. The dicyclohexylammonium salt of pNPP precipitated, collected by filtration, washed with cold absolute ethanol, and dried overnight under vacuum to yield phosphate-free dicyclohexylammonium salt of pNPP as a white solid, which was stored at −20 °C under nitrogen.</p><p>The sodium salt of pNPP allows preparation of stock solutions of higher concentrations, and was produced by dissolving the dicyclohexylammonium salt, obtained above, in 0.8 M NaOH solution followed by removal of cyclohexylamine by chloroform extractions. The pH of the aqueous layer was adjusted to pH 12 with 0.1M NaOH, as necessary. A small amount (0.5–1 g) of Amberlite IR120 H resin, extensively pre-washed with deionized water, was added to the solution and stirred for 1 minute. When pH reached 6.0–6.5 the resin beads were removed by filtration. The solution was adjusted to pH 8.5 with dilute sodium hydroxide, and stirred on ice under reduced pressure overnight to remove any traces of organic solvent. The concentration of the stock solution was assayed by adding 10 uL of the pNPP solution to 1 mL of 100mM Tris buffer (pH 10) followed by complete hydrolysis with alkaline phosphatase. The final concentration of liberated p-nitrophenol was found using the value of λ400= 18300 M−1cm−1.</p><!><p>A malachite green assay was used to determine the concentration of inorganic phosphate in the pNPP substrate, and the concentration of inorganic phosphate in inhibition studies. Briefly, 1.5 g of ammonium hexamolibdate were dissolved in 85 mL of deionized water. The volume was adjusted to 100 mL with 15 mL of concentrated (60.5%) perchloric acid yielding Solution A. The use of perchloric acid was found to improve sensitivity and rate of the color development. Malachite green hydrochloride (0.2 g) were added to 50 mL of solution A with stirring. After 30 min the dark orange solution was centrifuged to separate any undissolved particles yielding Solution B. Solution A was used to adjust Solution B to a final A450 = 15 a.u. (measured by assaying diluted aliquots and calculating back to the original concentrate). In a separate 250 ml Erlenmeyer flask 2.5 g of hydrolyzed PVA (Acros) were stirred in 100 mL of deionized water under gentle heating avoiding boiling for several hours to yield clear colorless Solution C (2.5% w/v). Prior to use 2 ml of solution C were added to 10 mL of adjusted Solution B and mixed to yield a dark brown working solution. In a 96 well plate 280 µL of working solution were mixed with 30 µL of diluted phosphate standards in the 0–200 µM range. The color was fully developed in 3–5 minutes and absorbance measured at 625 nM was plotted versus inorganic phosphate concentration to obtain a calibration curve.</p><!><p>The effect of buffers and other inhibitors on the activity of VHZ was tested with pNPP (sodium salt) in 50 mM sodium acetate buffer (pH 5.5), which showed no inhibitory effect, with inhibitor concentrations in the range 25–200 mM. The data was fitted to several inhibition models using non-linear least squares fit (Origin 8.5.1) with the competitive model yielding the best results in all cases.</p><!><p>All reactions were performed in non-inhibitory buffers in the presence of 1mM DTT and constant ionic strength adjusted to 150 mM with NaCl at 25 °C. The following buffers, which showed no inhibition against VHZ, were used: pH 4.75–5.5, sodium acetate; 5.75–6.5, sodium succinate; 6.75–7.25, 3,3-dimethylglutarate. All buffers were 50 mM. At least eight substrate concentrations ranging from 0.5–4 x KM were used for each enzyme to obtain Michaelis-Menten curves. Reaction progress was followed continuously by the change in absorbance at 400 nm at each pH using a VersaMax plate reader (Molecular Devices). Partial extinction coefficients obtained as described in the SI were used to convert the measured absorbance values into amount of p-nitrophenol released. VHZ was susceptible to product inhibition, so linear initial velocities were used at each substrate concentration. Control reactions in the absence of enzyme verified that non-enzymatic hydrolysis could not be detected during the time spans used for enzyme kinetics. The initial rates were plotted vs. substrate concentrations and fitted to the Michaelis-Menten equation using Origin 8.5.1 to obtain kinetic parameters kcat and KM. The values of kcat and kcat/KM were plotted versus pH and fitted to the equations shown in the SI to obtain the pKa values of catalytically important ionizable residues, presented in the Table S1.</p><!><p>Kinetic isotope effects (KIEs) on the VHZ-catalyzed reaction with pNPP were measured using the competitive method, and thus are isotope effects on V/K. Figure 2 shows the positions where KIEs were measured in the substrate and the designations used. Natural abundance pNPP was used for measurements of 15(V/K). The 18O KIEs 18(V/K)bridge and 18(V/K)nonbridge were measured by the remote label method, using the nitrogen atom in p-nitrophenol as a reporter for isotopic fractionation in the labeled oxygen positions (13). The isotopic isomers used are shown in Figure 2, and their synthesis was as previously described (14). KIEs were measured at pH 5.5 using the buffer 100 mM acetate, 1 mM DTT, 0.5 mM EDTA. The reactions were conducted at 25°C and [pNPP] = 25 mM. All reactions for bridged, non-bridged and natural abundance isotope effects were performed in triplicate and the progress was monitored continuously by measuring absorbance at 400 nM. Reactions were stopped at 40, 50 and 60 % completion by acidification to pH 3.0 with HCl, which caused some precipitation of protein, which was separated by centrifugation. To remove remaining enzyme, the reaction solution was centrifuged in a Millipore Amicon-Ultra 3kDa protein concentrator. The resulting solution was extracted with three 50 mL portions of diethyl ether. Ether fractions were collected, dried with anhydrous magnesium sulfate, filtered, and ether was removed by rotary evaporation. The unhydrolyzed pNPP remaining in the aqueous layer was hydrolyzed using alkaline phosphatase after the pH was adjusted to 10 by addition of 1 M Tris to pH 11. After 8 hours the pNP product was extracted and collected as previously described. After sublimation the p-nitrophenol samples were analyzed by isotope ratio mass spectrometry. The isotope effect was calculated from the nitrogen isotopic ratios in the p-nitrophenol product at partial reaction, in the residual substrate, and the starting material, as described in Supporting Information. The 18(V/K)nonbridge values were corrected for the dianionic fraction of pNPP at pH 5.5.</p><!><p>The hydrolysis of pNPP by VHZ and VHZ/P68V was monitored using a JEOL 300 MHz NMR spectrometer at the 121.5 MHz 31P resonance frequency. Reaction was started by addition of VHZ to a solution containing 20 mM pNPP in 50 mM sodium acetate (pH 5.5), 0.5 mM EDTA, 3mM DTT and 1M ethylene glycol. The instrument was locked using D2O in a coaxial tube. The instrument was set to collect 64 scans, relaxation delay 4 sec, sweep 50 ppm.</p><p>The dependence of kcat was measured as a function of ethylene glycol concentration for VHR, VHZ, VHZ/E134Q, VHZ/E134A, VHZ/D65A, VHZ/P68V, SsoPTP, and SsoPTP/V72P by monitoring release of p-nitrophenol. Reactions were performed in 50 mM sodium acetate buffer (pH 5.5) and concentration of ethylene glycol varied in the range of 0–3 M.</p><!><p>Measurement of pNPP hydrolysis catalyzed by VHZ, VHR, as and VHZ/D65A mutant were performed at 25 °C using a stopped-flow spectrophotometer (KinTek). Release of p-nitrophenol was monitored by increase in absorbance at 400 nM in 100mM succinate buffer (pH 5.85), using an extinction coefficient corrected for this pH (see SI). Enzyme concentration varied in the range 20–65 µM. The pNPP concentration was 20 mM for VHZ/D65A, and 50 mM for VHR/WT and VHZ. Absorbance traces of at least five separate experiments at each substrate and enzyme concentration were averaged. The data was fitted to the equation [p-nitrophenol] = At + B (1-e-kt). At saturating concentrations of substrate, k = k3 + k5. The linear steady-state phase A = k3k5 / (k3 + k5). The magnitude of the burst B = Eo [k3 / (k3 + k5)]2 / (1 + KM / So)2.</p><!><p>The computational packages AutoDock (15, 16) and FlexX (17) were used to model the docking of pNPP in the active site of VHZ. The 1.15 Å resolution crystal structure of the VHZ-metavanadate complex (PDBID: 4ERC) was used as the starting model. The VO3 ligand at the active site was removed, and the following charges/protonation states were assigned to the residues: C95 (−1); H94 and R101,(+1); D65 and E134, neutral. The side chains of D65, F66, L97 and E134 residues were treated as flexible, the rest of the protein was rigid. The pNPP ligand coordinates were extracted from the crystal structure of SsoPTP (PDBID: 2I6P) and the di-anionic form of the phosphate ester was used for docking. Free rotation was allowed around the phosphate monoester bond, while the p-nitrophenol ring was treated as planar and rigid. The Lamarckian Genetic algorithm was selected for its reliability and ability to calculate "deeper minima" (18) and the search was conducted within a 15 Å grid region centered at the active site.</p><!><p>The effects of buffers and oxyanions on the catalytic activity of VHZ are summarized in Table 1. No inhibitory effects were observed for acetate, succinate, 3,3-dimethylglutarate, or diglycine. VHZ is weakly inhibited by triethanolamine hydrochloride and cyclohexylamonium hydrochloride, while N-ethyl morpholine (NEM) hydrochloride displayed no inhibitory effect. As expected, the sulfonate containing buffer HEPES showed relatively strong inhibition. Cyclohexylammonium is a common counterion for the commercially available substrate pNPP, and although its effect was relatively weak, the sodium salt was used in all experiments.</p><!><p>The KIEs in the positions shown in Figure 2 for the VHZ-catalyzed hydrolysis of pNPP are presented in Table 2, together with previously reported KIEs for PTP1B and VHR. The 18(V/K)nonbridge KIEs for VHZ are very similar to these precedents (10, 19), suggesting that, like these related enzymes, the VHZ-catalyzed reaction proceeds via a loose transition state. The 18(V/K)bridge KIE is higher in VHZ and a significant magnitude for the 15(V/K) isotope effect contrasts with the absence of a measureable KIE in this position in the other two phosphatases. Both observations are consistent with incomplete neutralization of the leaving group by the general acid in the transition state.</p><!><p>Table 3 shows the kinetic parameters, kcat, KM, and kcat/KM for VHZ, VHR, YopH, PTP1B, SsoPTP, and several VHZ mutants, measured at pH 5.5 at 23 °C. This corresponds to the pH optimum of many PTPs, including VHZ. The kinetic parameters for VHR, YopH, and PTP1B measured by continuous assay at 400 nm are in good agreement with the literature values. We obtained values of kcat = 3.9 s−1 and KM = 8.3 mM for pNPP hydrolysis by VHZ at the pH optimum of 5.5, compared with previously reported kcat = 0.009 s−1 and KM = 1.5 mM (20). The revised kcat is comparable to those of native VHR and SsoPTP. The KM is significantly higher than previously reported, resulting in lower overall enzymatic efficiency (kcat/KM).</p><!><p>The pH-rate profiles for native VHZ and the mutants D65A and E134Q are shown in Figure 3. The pKa values of ionizable residues were determined by non-linear least squares fitting of experimental data to the appropriate equations (see SI), for kcat and kcat/KM, and shown in Table S-1. The second ionization constant for pNPP was set to the literature value of 4.96 (12). The pKa values obtained are in good agreement with those previously obtained for other PTPs, and the bell-shaped profiles are typical of the PTP family. The acid limb is ascribed to the deprotonated C95 nucleophile, which has an unusually low pKa due to stabilizing influences of the neighboring H94 in VHZ. The basic limb is ascribed to general acid catalysis in the first step consistent with the PTP mechanism (Scheme 1).</p><p>The IPD-loop in VHZ bears the putative D65 general acid. This loop and the position of the acid correspond to the conserved WPD-loop in classical PTPs (Figure 4). D65 was proposed to be the general acid catalyst by structural analogy, and based on the observation of lost hydrolysis of pNPP by the D65A mutant (21, 22). In our hands, the D65A mutation resulted in a 5-fold reduction of kcat and a 40-fold reduction of KM. The pH-rate profile remains bell-shaped (Figure 3). In contrast, the analogous mutation of the general acid in other PTPs to alanine or to asparagine results in a 100 to 1000-fold reduction in kcat, and the loss of the basic limb in the kinetic pH-rate profile, indicating loss of general acid catalysis (23, 24). In contrast, the kinetic constants for the D65N mutant are similar to those of native VHZ (Table 3).</p><p>The active site of VHZ possesses another acid (E134 from the Q-loop) in position to potentially protonate the leaving group (Figure 4). The E134Q and E134A mutations both reduced kcat by about an order of magnitude (Table 3) with no significant effect on KM. The pH profile of the E134Q mutant remained bell-shaped. Only the simultaneous removal of both general acids (D65N/E134Q) led to a significant reduction of kcat and loss of the basic limb of the pH-rate profile (Figure 3). The double mutants precipitated below pH 5.25, precluding measurement of kinetic data below this pH.</p><!><p>Many PTPs, including YopH, PTP1, and VHR, exhibit burst kinetics, indicating that the second step is rate-limiting (3, 25). When native VHR was rapidly mixed with pNPP a pre-steady state burst of p-nitrophenol release was observed (Figure 5). The values of k3 and k5 obtained for VHR are close to those previously reported (25). Under the same conditions native VHZ and its E134Q mutant revealed no pre-steady state burst. In contrast, the VHZ/D65A mutant displayed a pre-steady state burst stoichiometric with the amount of enzyme. Values of k3, k5 and KS (the true substrate binding constant) were calculated for native VHR and VHZ/D65A and are presented in Table 4. Although the decrease in the ratio (k5/k3) in VHZ/D65A reduces KM (see SI), it does not fully explain the magnitude of the reduction, indicating that KS is reduced as well. Despite the absence of a burst, the value of KS for native VHZ can be calculated if we assume that the primary role of D65 is orienting the nucleophilic water in the second step. This assumption is supported by the negligible effect on kcat of the D65N substitution. By setting k3 for native VHZ equal to that of D65A, the other elementary rate constants for the reaction of the native enzyme were calculated as shown in Table 4. The VHZ/D65A substitution decreases the magnitude of KS 14-fold. Because substitution of D65 with A (but not N) increases enzyme substrate affinity, we conclude the existence of an unfavorable steric, rather than electrostatic, interaction between D65 and the aromatic ring of the substrate. The Ki for inorganic phosphate is lower than KS for pNPP in both native VHZ and VHZ/D65A. However, whereas KS for pNPP is reduced 14-fold by the removal of D65 side chain (Table 4), the Ki for inorganic phosphate is reduced only ~ 1.5-fold (Table 1).</p><!><p>The dependence of kcat on ethylene glycol concentration was measured for native VHZ and several mutants. It should be noted (discussed in the SI) that the slope of such a plot where kcat = k5 + k7 ×[ROH] yields the second order rate constant for ethylene glycol phosphorylation (k7, Scheme 2) only when k3>>k5. In the more general case, when no distinct rate-limiting step exists and no pre-steady state burst is present, the expression kcat=k3×k5k5+k3+k3×k7k5+k3×[ROH] holds and interpretation of the calculated slope is more complex. However, the nucleophilic Slope selectivity S=SlopeIntercept=k3×k7k5+k3k3×k5k5+k3=k7k5 (Table 5, column 3) is independent of k3 and reflects the relative preference for alcoholysis over hydrolysis.</p><p>The E134Q mutation results in no significant change in phosphotransferase ability. The E134A mutation significantly increases the alcoholysis to hydrolysis ratio, indicating that E134 participates in the second step of the reaction, like Q446 in YopH. Although Q-loop residues in PTPs are generally thought to control access to the phosphoenzyme intermediate, the D65A mutation on the IPD-loop has a more pronounced effect than E134A. Because the closely related SsoPTP lacks phosphotransferase activity, we compared the effects of mutations on the IPD-loops in VHZ and SsoPTP. The SsoPTP contains only two proline residues on its IPD-loop, P68 and P73, corresponding to P64 and P69 in VHZ. These residues are highly conserved in classical PTPs (see the sequence alignment in Figure 8) and mutations result in similar, adverse catalytic effects in VHZ (Table 3). VHZ possesses another proline, P68, which is occupied by V72 in SsoPTP. The VHZ/P68V mutation abolished phosphotransferase activity (Figure 6). This effect was further supported by 31P NMR spectroscopy shown in Figure 7. The orthogonal SsoPTP/V72P mutation conferred phosphotransferase ability comparable to that of native VHZ.</p><!><p>Although VHZ is related to classical PTPs and utilizes the same mechanism (8) it differs in several important ways from most PTP family members.</p><!><p>Previous kinetic investigations of VHZ utilized single time point assays after 20 or 30 minutes, and buffers containing Tris (20, 22) or Bis-Tris, which are weak competitive inhibitors (Table 1). Although the inhibition constants are in the high millimolar range, buffers are typically used at such concentrations. While buffers containing the sulfonate functional group, such as HEPES, are recognized as inhibitors of PTPs, and have been observed in some PTP crystal structures, the inhibition of VHZ by buffers such as Tris and triethanolamine was unexpected. This behavior may result from the combined effect of several anionic residues near the VHZ active site (D65, E44, E134 and E137) but the molecular origin of buffer inhibition was not tested. The inhibition constant for inorganic phosphate is lower than in classical PTPs, and is significantly lower than the KS for pNPP (Table 4). Probable structural origins for this difference are discussed below. The net effect of these properties led to a significant underestimation of VHZ activity in previous reports.</p><p>In addition to avoiding inhibitory buffers, we validated a method for the continuous collection of rate data using the substrate pNPP monitoring reaction progress at 400 nm using extinction coefficients measured at the experimental conditions (see SI). Continuous monitoring of the reaction showed that VHZ becomes inhibited by product after several minutes. Finally, commercial pNPP, which often contains small but experimentally significant amounts of inorganic phosphate, was used in previous studies. We developed a purification strategy to minimize contamination by inorganic phosphate, that also offers a simple means to convert the di-cyclohexylammonium salt to the more soluble disodium salt.</p><p>Using non-inhibitory buffers, a continuous assay, and phosphate-free substrate, VHZ proved to be significantly more active than previously reported (kcat = 3.9 s−1 versus 0.009 s−1). We also developed a process for the expression and purification of the native enzyme with no tags. Although previous reports used tagged versions of VHZ, the difference in activity is too significant to be explained by the absence of tags. Furthermore, discontinuous assays of tagless VHZ using the previously reported methods yielded results similar to the previous reports (20, 22).</p><!><p>The KIEs for the hydrolysis of pNPP have been reported for a number of PTPs including YopH (26), PTP1B (27), VHR(19), and the LMW PTP Stp1(28). Because these are measured by the competitive method, the KIEs reflect the portion of the kinetic mechanism up to the first irreversible step, cleavage of pNPP. The mechanistic origins of the KIEs for pNPP hydrolysis have been described elsewhere (13). In brief, a normal 15(V/K) up to a maximum of 1.003 arises from negative charge development on the leaving group. Efficient general acid catalysis by PTPs abolishes this effect, resulting in a 15(V/K) of unity. Fission of the P-O bond produces a normal 18(V/K)bridge effect. Simultaneous leaving group protonation produces an inverse effect that partially reduces the normal effect from P-O cleavage. Thus, in general acid mutants, 18(V/K)bridge is typically ~ 1.03, compared to ~1.015 in native PTPs. The 18(V/K)nonbridge KIE responds to change of hybridization state of the phosphoryl group. This KIE is slightly inverse to near unity in PTPs, reflecting the loose metaphosphate-like transition state; associative transition states result in normal values.</p><p>Both of the oxygen isotope effects for the VHZ reaction are within experimental error of previous data with PTP1B and VHR (Table 4). In contrast, the 15(V/K) differs from the other PTPs and is slightly normal, indicating that the leaving group is not completely neutralized in the transition state. The magnitude suggests approximately 1/3 of a negative charge, from protonation that is not fully synchronous with P-O bond fission. This has been observed in one previous PTP family member, the LMW PTP Stp1(28). Incomplete protonation of the leaving group in the first step may contribute to a reduction of k3 and explain the absence of a burst in the VHZ catalyzed reaction of pNPP.</p><!><p>The D65 residue resides on the IPD-loop, a structure analogous to the WPD-loop in classical PTPs that bears the conserved general acid (Figure 8). The E134 residue in VHZ is superimposable with a conserved Q residue located on the Q-loop in classical PTPs that orients the nucleophilic water in the second step (29). The pH rate profiles of the D65 and E134 mutants both retain their basic limbs. The D65N mutation has no significant effect on catalysis, and the D65A and E134Q mutations have kcat values that are reduced by only an order of magnitude. Only in the double mutant is catalysis reduced to the extent seen in general acid mutants of other PTPs, consistent with complete elimination of general acid catalysis. We conclude the D65 and E134 single mutants both retain general acid function. We conclude that native VHZ utilizes E134 as the primary general acid, with a minor contribution from D65, which becomes the major general acid when E134 is mutated. The data suggest that, unlike any other known PTP or DSP, VHZ contains two acidic residues in the active site, either of which can protonate the leaving group in the absence of the other.</p><p>One might consider whether these results arise from the good leaving group in pNPP that may not need an enzymatic general acid to protonate it at the acidic pH optimum. However, several lines of evidence show that general acid catalysis is part of the mechanism of pNPP hydrolysis by PTPs, and with other substrates. KIE results at the acidic pH optima across the PTP family show that the leaving group leaves as the neutral phenol in native enzymes, but is charged when the enzymatic general acid is mutated. In such mutants the rates of pNPP hydrolysis are significantly reduced, and pH-rate profiles lose their basic limbs (13).</p><!><p>The fact that the D65N mutation does not significantly affect activity, while the D65A mutation does, suggests that the reduced catalysis in the D65A mutant arises due to inability of the alanine side chain to participate in phosphoenzyme hydrolysis by orientation of the nucleophilic water. In this sense, the roles of D65 and E134 are reversed from classical PTPs, in which the conserved glutamine corresponding to E134 positions the nucleophilic water, and the acid corresponding to D65 is the general acid in the first step and a general base in the second step.</p><p>The D65A mutation significantly reduces KM and KS. This effect is less pronounced for mutations of the corresponding residue in other PTPs. This may be explained by fact that in classical PTPs this residue resides on the mobile WPD-loop, which is primarily in an open conformation in the free enzyme. Because the IPD-loop in VHZ is permanently closed (8), the D65 side chain is fixed in a position that restricts access to the deep and narrow VHZ active site. DSPs, such as VHR, also have a non-movable general acid loop; however, in these enzymes the general acid is positioned on another side of the active site (Figure 4), and presents less steric hindrance to incoming ligands. This explains the higher KM for pNPP in VHZ compared to other PTPs. The structurally analogous SsoPTP, which also contains a rigid IPD-loop, has a KM value approximately 3-fold lower than VHZ and comparable to that of VHR. However, unlike VHZ, which has a narrow and deep active site pocket (8) the active site of SsoPTP is broad and shallow due to the presence of multiple surrounding glycine residues in the P-loop and IPD-loop (Figure 8).</p><p>Neither D65A nor D65N substitution significantly affects the Ki for inorganic phosphate (Table 1). It also confirms that the mutation of D65 side chain side chain does not disrupt the P-loop, which serves as the dominant binding element to the anionic phosphoryl group. In contrast, the D65A mutation lowers the Ks for pNPP 14-fold. This suggests the effect of D65 substitutions on binding is primarily steric rather than electrostatic, involving the ester group of the substrate more than the phosphoryl group. This would permit the biological activity of VHZ to be more regulated by levels of intracellular phosphate than most classical PTPs. The intracellular regulation of phosphatases by phosphate has been recently discussed (30). Unlike classical PTPs with much higher Ki values, VHZ, SsoPTP, and VHR have inhibition constants similar to the average physiological concentration of inorganic phosphate (1–1.3 mM) (31).</p><!><p>Despite the fact that VHZ is more active than previously thought, it remains one of the least active PTPs. A significant part of its reduced catalytic efficiency arises from its high KS. Indeed, for the PTP catalyzed reaction (Scheme 1) kcatKM=k3KS (see SI for derivation), which means that VHZ requires a 5–8 fold higher substrate concentration to achieve its limiting velocity. In addition to high Ks values, the kcat for VHZ is lower than classical PTPs. Our data suggest that both steps of the VHZ catalyzed reaction are slower than in classical PTPs, and both contribute to the overall rate (kcat=k5×k3k3+k5) under steady state conditions. According to the KIE results the neutralization of the leaving group in the first step is incomplete, which reduces k3, but cannot explain the reduction of k5. It was previously shown that mutations of Q-loop residues in classical PTPs reduces k5 by 2 orders of magnitude, an effect that was used to trap the phosphoenzyme intermediate of PTP1B (24, 29). The E134Q mutation makes VHZ structurally more similar to classical PTPs and affects only the first step, resulting in the 8-fold reduction of activity. The kcat value of the E134A mutant is 70% that of the E134Q mutant. Such an insignificant reduction from complete removal of the functional group indicates that the residue does not function in the second step like the Q in this position in classical PTPs, explaining the lower k5 in VHZ compared to other PTPs. The k5 value of native VHZ is more similar that of VHR (Table 4) which, in place of the Q-loop, (Figure 4) has an N-loop region (Figure 8) that is highly conserved among atypical DSPs.</p><!><p>The expressions for kcat and KM for the PTP-catalyzed reaction (Scheme 1), kcat=k5×k3k3+k5 and KM=KS×k5k3+k5 contain the same elementary constants, and thus, any change in kcat is reflected in KM and vice versa. A reduction in k3 results in increased KM as is seen in the VHZ/E134Q/D65N double mutant, in which both general acids important in k3 are mutated, but k5 remains unaffected. As k3 becomes smaller with respect to k5, KM approaches KS. Because of this, the KM for the VHZ/D65N/E134Q double mutant (Table 3) is close to the KS for native VHZ determined from pre-steady state kinetics (Table 4). The same effect has been previously observed in the general acid D92N mutant of VHR (23). In contrast, mutations of Q loop residues that do not affect k3 but decrease k5, have the opposite effect on KM. The same trend was previously observed in the Q556M Q446A mutants of YopH (6). The reduction of k5 reduces the rate of formation of the free enzyme form E, which lowers the KM parameter in the steady-state experiment. This mutual dependence of kcat and KM becomes less obvious when k3 >>k5 but, because in native VHZ the two constants are relatively similar, (consistent with the absence of a burst) even small changes should be easily detected. The E134Q mutation in VHZ eliminates the major general acid but results in only a 10-fold reduction of kcat and has no effect on KM. The D65N substitution results in a modest rate reduction, consistent with its role as the minor general acid, but also has no significant effect on KM. The pH rate profiles of both mutants remain bell-shaped, indicating that when one general acid is eliminated, the other one takes over, maintaining general acid catalysis. The kinetic behavior of the mutants suggests the presence of two catalytically equivalent, but differently populated, forms of the Michaelis complex. Similar to the mode of action of a noncompetitive inhibitor on the native enzyme, the E134Q or D65N mutation renders one of the two conformations catalytically unproductive, which reduces the rate but has no effect on KM.</p><p>In order to obtain insights into substrate binding modes in VHZ that might explain how either D65 or E134 can act as a general acid, the programs AutoDock (15, 16, 32, 33) and FlexX (17) were used to predict the orientation of pNPP in the active site. Both programs predicted two possible conformations, presented in Figure 9. In each, the phosphate moiety of the substrate is coordinated by the P- loop and the side chain of R101, and is properly positioned for nucleophilic attack by the negatively charged cysteine at the bottom of the active site. In conformation A the scissile oxygen is oriented towards E134, consistent with its assignment as the primary general acid. The D65 side chain turns away to avoid a steric clash with the p-nitrophenyl ring. We conclude that this substrate conformation is the predominant one, consistent with the kinetic effects of the E134Q mutation and steric relief observed in the D65A mutant. The substrate conformation in panel 9B is the one commonly observed in classical PTPs, except for the position of the glutamine residue analogous to E134, which occupies a different conformation to avoid a steric clash with the substrate phenyl ring (29). However, it resembles the position of the corresponding Q135 residue in several structures of SsoPTP in complex with peptide substrates (9). In this conformation the phenolate ring is in close proximity to the L97 side chain, and D65 is oriented in position to donate its proton to the scissile oxygen of the leaving group. This substrate orientation is consistent with the kinetic behavior of the E134Q mutant, which utilizes D65 as its general acid.</p><p>The presence of two binding conformations explains why the E134Q and D65N mutants have a reduced kcat with no significant effect on KM. If conformation B, in which the E134Q mutant can utilize the D65 general acid, presents a minor fraction of the overall enzyme-substrate complex, then it would kinetically appear as a reduction of kcat because the predominant conformation A would be catalytically unproductive in this mutant. The different effects of the D65N and E134Q mutations on kcat with no significant effect on KM suggest that while both conformations A and B coexist, conformation A dominates. The E134Q or D65N substitution, which results in conformation A or B becoming non-productive, respectively, is similar to the effect from noncompetitive inhibition. Noncompetitive inhibition results from the formation of a non-productive enzyme-substrate-inhibitor complex that reduces Vmax but has no effect on KM. The reaction catalyzed by the E134Q mutant where conformation A is unproductive and conformation B leads to product formation according to the equation: VmaxA=VmaxWT(1+1KAB), where VmaxWT is the maximal rate of the reaction by the native VHZ, and VmaxA is the maximal velocity for the E134Q mutant. The calculated value of KAB=[B][A]=0.12 indicates that about 89% of the VHZ [ES] complex exists in conformation A, and 11% in form B. This ratio is consistent with the modest reduction of kcat in the D65N mutant. The equilibrium constant of 0.12 implies a modest energetic difference of slightly more than 1 kcal/mol, consistent with the nearly equal scoring function obtained from the docking programs for the two conformations.</p><p>The predominance of conformation A in which the phenolate ring interacts with the D65 side chain explains the steric relief produced by the D65A mutation. The presence of the E134 general acid and its use in conformation A may be a reasonable evolutionary solution to permit a bulky substrate to enter the narrow and sterically demanding active site pocket. Conformation B is the one the most commonly observed in the classical PTPs, which possess a movable WPD-loop. This conformation allows VHZ to utilize its D65 general acid. However, under the condition of a non-movable general acid-bearing loop, conformation B is sterically disfavored, making it less populated than conformation A. The presence of the two coexisting but unequally populated conformations explains why both the E134Q and D65N mutants retain their basic limbs in the pH-rate profile, but produce proportionally different effect on kcat without affecting KM.</p><p>A similar phenomenon may explain anomalous behavior reported in the Tk-PTP, a protein tyrosine phosphatase isolated from the Thermococcus kodakaraensis KOD1. On the basis of structural comparisons with known PTPs, D63 was assigned as the putative general acid. Unexpectedly, the D63A mutant was found to be more active, not less, than the native enzyme (34). No structure of Tk-PTP has been published; however, the protein sequence (Figure S-2) suggests the presence of a glutamic acid residue in a position corresponding to that of E134 in VHZ. Like VHZ, Tk-PTP may utilize a different general acid than the one implicated by structural comparisons with the PTP family.</p><p>The finding of two alternate binding modes and two general acids may be physiologically relevant, but the results presented here do not address whether the same flexibility pertains to peptide substrates. Physiological substrates for VHZ have not been confirmed. Recently, the existence of two alternate peptide substrate binding modes, depending on sequence, has been reported for the related enzyme VHR (35). In that case, only a single general acid is in the active site and is used in both conformations.</p><!><p>In classical PTPs, Q-loop residues such as Q446 and Q450 in YopH (Figure 4), in locations analogous to E134 and Q138 in VHZ, position the nucleophilic water for the second step and shield the phosphoenzyme from larger nucleophiles (6). In classical PTPs catalysis is unaffected by added ethylene glycol. In contrast, YopH mutants in which these glutamines are mutated show significant phosphoryl transfer to ethylene glycol (6). This has provided a rationalization for why the native VHR and LMW-Ltp1, which lack an analogous Q-loop, display phosphotransferase activity (7). Thus, the observation of phosphotransferase activity by VHZ was unexpected, since its E134 and Q138 residues are superimposable with Q446 and Q450 in YopH. The ratio S of the second-order rate constant for alcoholysis by ethylene glycol (kt) to the hydrolysis of the phosphoenzyme intermediate (kcat') is 14.4 (Table 5). This ratio was not significantly affected by the E134Q mutation, indicating that the E and Q residues, as was previously shown with YopH, are interchangeable (6). The higher transphosphorylation by the E134A mutant indicates that that E134 provides some shielding of the phosphoenzyme. However, the analogous YopH/Q446A mutation results in a greater increase, from which we conclude that E134 in VHZ does not function as effectively in the second step as Q446 in YopH. The VHZ/D65A substitution results in the most pronounced increase in the S ratio, which, together with the reduced kcat, indicates that D65 is important in the second step.</p><p>Interestingly, unlike VHZ, there is no effect of ethylene glycol concentration on catalysis by SsoPTP. A sequence alignment of the IPD-loops of VHZ and SsoPTP revealed that while SsoPTP contains two proline residues in this region, a pattern that is highly conserved in the PTP family, VHZ has a third proline, P68. In SsoPTP, V72 is found in the corresponding position. The VHZ/P68V mutation results in no significant change in kcat; however, phosphotransferase activity is lost and there is no dependence of rate on ethylene glycol concentration. The orthogonal SsoPTP/V72P mutation confers phosphotransferase ability similar to that of native VHZ. We thus conclude that the single point mutation of proline at the P68 position in VHZ, and the analogous V72 position in SsoPTP, controls the ability of these enzymes to phosphorylate alcohols. Because the IPD-loops in VHZ and SsoPTP are not mobile this mutation has no effect on kcat. Further studies, including structural comparisons, are underway to seek an explanation of how this residue controls phosphotransferase activity.</p><!><p>Although VHZ is more closely related to classical PTPs than to DSPs, it is unique and has a number of unusual properties. VHZ has two functional general acids, and, for the small molecule substrate pNPP, two substrate binding modes. Both binding modes are catalytically equivalent but unequally populated. Each binding mode utilizes a different general acid.</p><p>Despite the fact that VHZ shares many active site characteristics of classical PTPs, it is a significantly less efficient catalyst. Both catalytic steps k3 and k5 are slower. The reduced k3 is consistent with incomplete neutralization of the leaving group revealed by KIEs. Reduction of k5 is due to the less efficient involvement of the E134 residue in the second step, as revealed by the E134Q and E134A mutations.</p><p>The results also provide a rationale for the advantage of mobility of the general acid-bearing WPD-loop in classical PTPs. The availability of an uncatalytic, open conformation allows the general acid to avoid sterically unfavorable interactions with substrate binding, and to alleviate product release and reduce product inhibition.</p><!><p>The fact that the reaction rate increases linearly with ethylene glycol concentration indicates the second step is at least partially rate limiting. Together with the absence of a burst in the pre-steady state, we conclude that both steps contribute to the overall rate.</p><p>The Q-loop does not fully participate in protecting the active site from incoming alcohol nucleophiles. Mutations in the IPD-loop of VHZ have more effect on phosphotransferase ability. It is logical to conclude that the phosphotransferase activity observed in VH1-related DSPs and Low-molecular weight PTPs can no longer be explained solely by their absence of a Q-loop. The significant difference in position of the general acid loop between classical PTPs, LMW-PTPs and VH1-related DSPs undoubtedly contributes as well.</p><p>The presence of phosphotransferase activity in some phosphatases, but not others, may have biological consequences. It may not be coincidental that some proteins known to be associated with cancer, such as VH1-related DSPs and LMW-PTPs, and, recently, VHZ(36), reveal a high level of phosphotransferase ability. Whether this process is random and nonspecific, or transphosphorylation is selective for some protein target, remains to be discovered. However, because VHZ has a stringent phosphotyrosine specificity because of its deep and narrow active site, we can conclude that its transphosphorylation target would likely be limited to tyrosine protein sites, or to small molecule nucleophile acceptors capable of entering the active site.</p>

PubMed Author Manuscript

Effects of external field wavelength and solvation on the Photophysical Property and Optical Nonlinearity of 1,3-Thiazolium-5-Thiolates Mesoionic Compound

The photophysical property and optical nonlinearity of an electronic push-pull mesoionic compond, 2-(4-trifluoromethophenyl)-3-methyl-4-(4-methoxyphenyl)-1,3thiazole-5-thiolate were theoretically investigated with a reliable computing strategy.The essence of the optical properties were then explored through a variety of wave function analysis methods, such as the natural transition orbital analysis, hole-electron analysis, (hyper)polarizability density analysis, decomposition of the (hyper)polarizability contribution by numerical integration, and (hyper)polarizability tensor analysis, at the level of electronic structures. The influence of the electric field and solvation on the electron absorption spectra and (hyper)polarizabilities of the molecule are highlighted and clarified. This work will help people to understand the influence of external field wavelength and solvent on the optical properties of mesoionic-based molecules, and provide a theoretical reference for the rational design of chromophores with adjustable properties in the future.

effects_of_external_field_wavelength_and_solvation_on_the_photophysical_property_and_optical_nonline

Introduction<!>Computational Details<!>Results and Discussion<!>Electron Absorption Spectrum and the Nature of Electron Transition<!>Molecular (Hyper)Polarizability and Its Essential Feature<!>Contributions to the Second-Order NLO Activity<!>EFISHG-Derived (Hyper)Polarizability and Its Essential Feature<!>Components of molecular (hyper)polarizability<!>Conclusion<!>Supporting Information Available:

<p>Organic optoelectronic materials, which can be adopted in various optical devices such as photoelectric sensors, optical information processing and storage devices, and optical communication elements, have always been one of the hot spots of people's attention and research. [1][2][3] The excellent optical response characteristics of material molecules are the premise for them to be used to construct advanced photoelectric functional materials. Therefore, the field of optical materials has been exploring how to develop new compounds with tunable photophysical properties or enhanced (hyper)polarizabilities for a long time. [4][5][6] So far, researchers have successfully recognized a variety of molecules with exciting optical properties, including inorganic complexes, [7][8][9] organic molecules, [10][11][12][13] metal clusters, 14,15 doped polarizable electrides, 16,17 and even macrocyclic compounds with special topology, 6,18,19 and applied some of them into practice. Among them, the conjugated organic chromophores with electron donating and electron accepting groups at both ends were one of the most popular optical units, partially due to their convenient structural tailoring. Generally speaking, the photophysical and nonlinear optical (NLO) properties of these chromophores can be controlled by adjusting their structural modules: electron donor, π-conjugated linker, or electron acceptor. [20][21][22][23][24] Since Baker and coworkers put forward the concept of "mesoionic" in 1949, 25,26 mesoionic rings defined by planar five-or six-membered heterocyclic betaines with at least one side-chain whose -atom in the ring plane have been recognized as a promising candidate for conjugated bridges of optical materials because of their electronic mobility. 27 Numerous mesoionic compounds with different electron donors and electron acceptors have been studied both theoretically and experimentally. [28][29][30][31] Recently, for example, Barbosa-Silva et al. prepared three mesoionic compounds and measured the optical properties of them by Hyper-Rayleigh scattering (HRS) experiment. 32 Our theoretical calculation for the same systems not only perfectly reproduced the electronic absorption spectra and first-order hyperpolarizabilities observed in experiment, but more importantly, it deeply revealed the essences of the optical properties of these mesoionic molecules from the electronic structure level. 33 At the same time, we also found that Lyra and collaborators had theoretically predicted the first-order static hyperpolarizability on a variety of mesoionic compounds including the above-mentioned three ones by employing the semi-empirical time-dependent Hartree-Fock (AM1-TDHF) method. 34 However, due to the constraints of the level of calculation adopted, the response properties of these molecules obtained by them are obviously different from the measurements of the experiment and those calculated by us with a more reliable strategy. 32,33 In view of the fact that the influence of external field wavelength and solvent usually used to adjust the optical properties of organic materials has not been reported yet for mesoionic systems, in this work, we selects an electronic push-pull mesoionic compound, 2-(4-trifluoromethophenyl)-3-methyl-4-(4-methoxyphenyl)-1,3-thiazole-5-thiolate (hereinafter referred to as MIC), that possessing substituents with strong electron-donating/accepting capacity and also involving in previous calculation studies 34 as the research object to explore the influence of environments on the optical properties of this kind of organic molecules.</p><!><p>The structure of the MIC was optimized at the PBE0 35 /def-TZVP 36 level in dimethylsulfoxide (C2H6OS) solution, which is the actual environment of the similar experimental studies. 32 The optimized geometry is characterized to be stable point on potential energy surface with no imaginary frequency. Then, the electron absorption based on density (SMD) was adopted to consider solvation effects on molecular structure and properties. 37 The excitation energy and oscillator strength of the lowlying singlet states of the MIC were studied with the time-dependent density functional theory (TD-DFT) 38,39 at the PBE0/def2-TZVP 40 level. The molecular (hyper)polarizability and (hyper)polarizability density were evaluated using analytic derivatives of the system energy (namely, coupled-perturbed Kohn-Sham method, CPKS) 41 and finite difference of electron density, respectively, at the CAM-B3LYP 42 /aug-cc-pVTZ(-f,-d) level, where aug-cc-pVTZ(-f,-d) is a reduced version of the aug-cc-Pvtz 43 basis set with the removal of f-type polarization functions of nonhydrogen atoms and d-type polarization functions of hydrogen atoms. The selections of PBE0 functional in calculating the excited state and CAM-B3LYP functional in evaluating molecular nonlinearity have been proved to be reasonable by many theoretical works. 33,[44][45][46] The decompositions of the (hyper)polarizability components into each structural unit were obtained by numerical integrations of the (hyper)polarizability density in the space of every atom partitioned by Beckes method. 47 For the formulas of the (hyper)polarizability, (hyper)polarizability density, and (hyper)polarizability decompositions, see the descriptions in the Supporting Information.</p><p>All (TD-)DFT calculations were performed with the Gaussian 16 program. 48 The wave function analyses were implemented in Multiwfn 3.7 code, 49,50 and the visualizations of the isosurface maps were realized with VMD software. 51</p><!><p>The Cartesian coordinates of the ground-state MIC in C2H6OS solution are available in Table S1, and the molecular structure is illustrated in Scheme 1. Scheme 1. Structure of the MIC studied in this work a a Also shown is the Cartesian axis. Atom color code: white, H; grey, C; blue, N; red, O; cyan, F; yellow, S.</p><!><p>The selected data related to electron transitions for the MIC are listed in Table S2.</p><p>The calculated absorption band of the molecule in vacuum almost covers the whole visible region of 420-800 nm, as shown in Figure 1. The strongest absorption is located at 548 nm followed by a peak at ultraviolet range of 333 nm with medium intensity. As also can be seen from the figure, solvation effect makes the two absorption bands shift significantly towards the short-wave direction regardless of the polarity of the solvent, accompanied by the enhancement and attenuation in the absorption intensity of the maximum absorption peak and its companion, respectively. However, there is no remarkable distinction of the absorption spectra in different solvents, and specifically speaking, with the increase of the solvent polarity, the absorption peak of the MIC shows a slight blue-shift from 481 nm in CH2Cl2 and seems to converge to a fixed value of 466 nm, but the absorption intensity of the spectrum does not show any regular change. For each maximum absorption, we performed the natural transition orbital (NTO) 52 analysis to visualize the orbital characteristics in the process of electron transition.</p><p>The inset in Figure 1 shows the isosurfaces of critical NTO hole/particle pairs of the MIC in various environments. It can be seen that the NTO pairs (mainly in the NTO hole) of the molecule in vacuum are only slightly different from those in solvents, while those in various solvents are indistinguishable. Therefore, it is not difficult to understand why the MIC displays very similar UV-Vis absorption spectra in solvents while subtly different one in vacuum. Further analysis shows that all maximum absorptions of the MIC are attributed to the electron transitions from lone-pair orbital, n, of S atoms and π-bonding orbital at mesoionic ring (B M ) to π*-antibonding orbital distributed at the B M and benzene unit (B A ) connected to the electron acceptor (R A = -CF3), namely, n→π* and π→π* electron excitations on the molecular skeleton, and the former is more significant.</p><p>Figure S1 shows the highest occupied molecular orbitals (HOMOs) and the lowest unoccupied molecular orbitals (LUMOs) of the MIC in various environments. A comparison between Figures 1 and S1 shows that the distribution of NTO hole/particle related to the maximum absorption of the MIC is very similar to its HOMO/LUMO, which suggests that the HOMO and LUMO are the crucial orbitals during the maximum absorption. Indeed, for the electron excitation in the maximum absorption process, the calculated contribution of HOMO-LUMO transition in vacuum is more than 92%, and the proportion of this transition in various solvents is nearly 99%. The blue-shift of the maximum absorption peak caused by solvation can therefore be reasonably explained by the fact that the HOMO-LUMO gap difference of MIC in solutions is significantly larger than that in vacuum, as shown in Figure S1.</p><p>Oscillator strength (f) of an excited state is proportional to the square of transition dipole moment (μ T ) corresponding to the state. 53 From the μ T of the molecule during the maximum absorption listed in Table S2, one can see that their changing trends in different environments are completely consistent with those of corresponding f value.</p><p>The hole-electron analysis 53</p><!><p>As in our previous paper about mesoionic compounds, 33 we found that when the fundamental wavelength of external field reaches a certain size, the molecular dipole moment and (hyper)polarizabilities will converge to a specific value. In solution, the convergence value is different from the calculated static one, but the two values are the same in vacuum. Considering the comparison in Ref. 33 between our calculated values and the corresponding experimental results, we believe that the convergence value at infinite wavelength rather than response value calculated under static conditions is a credible quantity that can be compared with the experimental measurement. Therefore, the response properties described below with "zerofrequency limit" is the calculated value at 99999 nm incident light, which is the maximum incident wavelength setting that Gaussian program accept in this kind of calculation. The molecular nonlinearity at other five different frequencies of incident light (λ = 1907, 1460, 1340, 1180, and 1064 nm, respectively) are also calculated for comparison.</p><!><p>In order to provide predictions for HRS experiment, we calculated the HRS-derived first-order hyperpolarizabilities HRS 2; S3.</p><p>Our HRS ()  value of MIC at the zero-frequency limit in vacuum is 27.9×10 -30 esu, which is smaller than the calculated result of 41.7×10 -30 esu by Lyra et al. 34 This discrepancy is believed to be due to the relatively rough AM1-TDHF method used in their calculations. Similar situations have been observed in calculating the hyperpolarizabilities of other mesoionic compounds. 33 The HRS-derived first-order hyperpolarizability of the MIC in this work enlarges with the increase of the external field frequency, that is, as shown in Figure 3</p><!><p>In general, the prerequisite for a bulk material to exhibit excellent NLO properties is that its constituent molecules have a large (hyper)polarizability. In order to estimate the potential application of the studied molecule in nonlinear optics, we calculated its</p><p>, and     , respectively, where λ denotes the incident wavelength],</p><p>related to the electric-field-induced second-harmonic generation (EFISHG) technique.</p><p>Then, the nonlinear essences of the optical response of the MIC were further investigated by the (hyper)polarizability density analysis, the decomposition of the (hyper)polarizability contribution by numerical integration, and the (hyper)polarizability tensor analysis.</p><!><p>The selected components of (hyper)polarizabilities of the MIC at different external fields are arranged in Table S4.</p><p>As can be seen, the x-component of the (hyper)polarizability accounts for a considerable proportion, and the contribution of the z-component to the (hyper)polarizability is not ignorable although it is small. By comparison, the ycomponent of molecular response properties, especially for the first-order and secondorder hyperpolarizabilities, is relatively trivial.</p><p>It is helpful to study the structural origin of the response property by decomposing the overall axial (hyper)polarizability component of molecules into the contributions of their constituent units, which can provide a reference for exploring how to modify and tailor the structure of organic molecules to improve their optical nonlinearity. The integrand functions of the (hyper)polarizability densities, that is, the local contributions of the (hyper)polarizability tensor, in each Cartesian axis [ We also integrated the local contributions of the (hyper)polarizability tensor of each atomic region to quantify the contribution of molecular fragments to the (hyper)polarizability components. The decomposed (hyper)polarizability values of each structural unit are listed in Table S5 and plotted in Figure 5. By comparing Tables S5 and S4, we can see that the overall components of the (hyper)polarizability of the MIC estimated by numerical integration is very consistent with those calculated by the analytic derivative method, thus ensuring the reliability of our (hyper)polarizability decomposition research. Figure 5(a) shows that the components of molecular polarizability in all three Cartesian directions contribute to each structural unit in a certain amount, which eventually leads to the insignificant difference of polarizabilities in three coordinates. In contrast, it can be seen from S4 and illustrated in Figure 6.</p><p>For a particular solvent in Figure 6(a), the iso ()   under different external fields is almost the same, that is to say, there is no obvious resonance of the molecular polarizability under the applied electric field. Generally speaking, the polarizability of molecules is indeed insensitive to the external field, which has been observed in our previous studies on optical nonlinearity of other systems. 6,54 Under the same electric field, on the other side, the iso ()   of the molecule calculated in solvent is slightly higher than that in vacuum, but there is little difference of it in different sorts of solvents. By comparing Figure 6(b) and (c), it can be seen that the external field resonance or solvation effect has the similar effect on vec ()  and || () , which trend similar to that of the HRS ()  in response to these influence discussed above. To be exact, the hyperpolarizability of the MIC in various environments increases with the increase of frequency in external field. However, only under the action of higher-frequency external field, the difference of hyperpolarizability in different solvents becomes significant. The molecular hyperpolarizability in vacuum is most sensitive to frequency of incident light, and it changes from indistinguishable with that in solvents at zero-frequency limit to absolute superiority at 1064 nm fundamental wavelength.</p><p>This trend can be well explained by the conclusion deduced from the two-state model 55 that the dynamic perturbation to the hyperpolarizability of chromophores is expected to reach the maximum by excitation using photons with twice the wavelength of one-photon transition, because the incident wavelength at 1064 nm is exactly twice the maximum absorption wavelength of the studied MIC in vacuum.</p><p>The (hyper)polarizability tensors in Figure 7</p><!><p>The effects of external field wavelength and solvation on the optical properties of a mesoionic compound, MIC, with push-pull electronic structure were studied in detail by means of quantum chemistry calculations. The solvation effect leads to a significant blue-shift of the electron absorption spectra of the MIC in the solution compared with that in vacuum, but no obvious difference is observed in the spectra in different solvents. The NTO analysis identified the contribution to the maximum absorption of the MIC originating from the n→π* and π→π* transitions at the mesoionic ring and benzene unit adjacent to the electron acceptor, and the conclusion from frontier molecular orbital analysis and hole-electron analysis are very consistent with it. The HRS-derived first-order hyperpolarizability and the polarization scan of the HRS intensity enlarge with the increase of the external field frequency but reduce as the polarity of the environment increases, which reveals the induced strengthening and weakening effects on the response properties of electric field and solvation to the studied system, respectively. In the coordinate system of the molecule, the (hyper)polarizability component on the x-axis makes a great contribution to EFISHGderived response properties. In contrast, the contribution of the z-axis component can not be ignored, while y-component is almost zero. Detailed (hyper)polarizability density and (hyper)polarizability decomposition analysis reveal the contribution of each structural unit of the studied molecule to its (hyper)polarizability component in terms of visualization and quantification analyses, respectively. In addition, the molecular (hyper)polarizabilities which can be compared with the EFISHG experiments were predicted, and it is found that their response to the environment is exactly the same as the trend of HRS-derived hyperpolarizability discussed. This work provides a physical perspective for understanding the photophysical property and optical nonlinearity for mesoionic compounds in different external field and solvent environments, which will be helpful to design high-performance optoelectronic materials with stimulus response.</p><!><p>Detailed formulas for calculating molecular response properties; optimized Cartesian coordinates for ground-state MIC; selected data related to electron transitions for the MIC; HOMOs and LUMOs for MIC; selected parameters related to the HRS study of the MIC; harmonic light intensity ( 2 V I   ) for the MIC as a function of the polarization angle (Ψ) by polar representation; selected parameters related to the EFISHG technique of the studied MIC; plots of (hyper)polarizability density functions of the MIC; components of the (hyper)polarizability of the constituent units of the MIC.</p>

ChemRxiv

High-Affinity Binding of Remyelinating Natural Autoantibodies to Myelin-Mimicking Lipid Bilayers Revealed by Nanohole Surface Plasmon Resonance

Multiple sclerosis is a progressive neurological disorder that results in the degradation of myelin sheaths that insulate axons in the central nervous system. Therefore promotion of myelin repair is a major thrust of multiple sclerosis treatment research. Two mouse monoclonal natural autoantibodies, O1 and O4, promote myelin repair in several mouse models of multiple sclerosis. Natural autoantibodies are generally polyreactive and predominantly of the IgM isotype. The prevailing paradigm is that because they are polyreactive, these antibodies bind antigens with low affinities. Despite their wide use in neuroscience and glial cell research, however, the affinities and kinetic constants of O1 and O4 antibodies have not been measured to date. In this work, we developed a membrane biosensing platform based on surface plasmon resonance in gold nanohole arrays with a series of surface modification techniques to form myelin-mimicking lipid bilayer membranes to measure both the association and dissociation rate constants for O1 and O4 antibodies binding to their myelin lipid antigens. The ratio of rate constants shows that O1 and O4 bind to galactocerebroside and sulfated galactocerebroside, respectively, with unusually small apparent dissociation constants (KD ~0.9 nM) for natural autoantibodies. This is approximately one to two orders of magnitude lower than typically observed for the highest affinity natural autoantibodies. We propose that the unusually high affinity of O1 and O4 to their targets in myelin contributes to the mechanism by which they signal oligodendrocytes and induce central nervous system repair.

high-affinity_binding_of_remyelinating_natural_autoantibodies_to_myelin-mimicking_lipid_bilayers_rev

INTRODUCTION<!>Preparation of antibodies<!>O1 and O4 binding to myelin by ELISA<!>Immunocytochemistry using O1 and O4 on live oligodendrocytes<!>Nanohole array fabrication<!>Formation of vesicles and supported lipid bilayers<!>SPR biosensing<!>O1 and O4 binding to oligodendrocytes and myelin<!>Supported lipid bilayers on nanohole arrays<!>SPR sensing of O1 and O4 binding to SLBs<!>Affinity vs. Avidity<!>CONCLUSION

<p>Multiple sclerosis (MS) is a neurologic disorder characterized by the loss of myelin sheaths in the central nervous system. Myelin loss impairs neuronal signal transduction and manifests itself in a number of symptoms including sensory and motor deficits.1 Approximately 1 out of every 1000 Americans has MS,2 but unfortunately there is no cure. Treatment options generally decrease exacerbations, but do not stop the long-term progression of the disease. Therefore, a major thrust in MS treatment research is to develop strategies to promote remyelination and protect vulnerable axons as a means to improve neurologic function. In earlier work, Rodriguez and coworkers showed that the IgM-isotype mouse monoclonal natural autoantibodies O1 and O4 promote remyelination in mouse models of MS.3 Furthermore, O1 and O4 are widely used to define oligodendrocyte lineage and monitor oligodendrocyte differentiation. O4 binds to sulfated galactocerebroside (Sulf) expressed on progenitor glial cells beginning at the time of commitment to the oligodendrocyte lineage through terminal differentiation of the cell and mature myelin membrane formation. O1 recognizes the presence of galactocerebroside (GalC) which is inserted into the oligodendrocyte membrane after the appearance of Sulf signifying later stages of maturation.4 Both O1 and O4 are polyreactive; O4 reacts with Sulf along with seminolipid and cholesterol, while O1 reacts with GalC as well as monogalactosyl-diglyceride and psychosine.5 Asakura et al. showed that both O1 and O4 have very few mutations from germline gene sequences, which indicates that they are natural autoantibodies.6 While these antibodies are useful in evaluating cell lineage and development, they also can initiate cell signaling and differentiation. When treated with O4, both astrocytes and oligodendrocytes display intracellular Ca2+ transients resulting from Ca2+ release from intracellular stores (astrocytes) or influx of extracellular Ca2+ (oligodendrocytes).7 Also, Bansal et al. showed that growing the cells in media containing O4 promoted oligodendrocyte differentiation whereas treating cultures with O1 did not.8</p><p>A number of lipid cell-surface antigens can be shown to bind O1 and O4 antibodies, but despite their wide use as detailed above, little is known about the affinity of the interactions or the kinetics by which these antibodies interact with their primary antigens, Sulf and GalC. Indeed, quantifying antibody binding kinetics to lipid bilayers mimicking cellular membranes is a very challenging task that requires label-free biosensors that can incorporate fluid lipid bilayer membranes. Commercial surface plasmon resonance (SPR) instruments such as Biacore have been widely used to measure receptor-ligand binding kinetics using ligands that are covalently attached on functionalized gold sensing surfaces.9,10 However, adapting conventional SPR instruments for sensing with fluid lipid membranes presents a series of challenges. Functionalized gold surfaces used in commercial SPR instruments can incorporate membrane-bound receptors on captured liposomes, however the morphology of the captured lipid membranes is not precisely known and appears to be dominated by intact liposomes rather than a planar lipid bilayer.11,12 Moreover, the capture efficiency of liposomes can be modulated by their charge state,12 making direct comparison between multiple liposome compositions more complex. Alternatively, lipid-antibody binding kinetics can be measured using planar supported lipid bilayer (SLB) membranes, which can be readily integrated with surface-sensitive techniques such as SPR. The majority of conventional SPR sensors utilize gold surfaces modified with self-assembled monolayers, whereas the formation of SLB is much easier on glass or silica surfaces.13 As Dahlin et al. have shown, the formation of SLBs on gold surfaces can be considerably simplified by patterning nanoholes in the gold film to expose the underlying silica surface.14 Thus we have employed the variant of SPR based on nanohole arrays in a gold film as the sensor chip and developed a new surface modification technique to conformally coat the entire sensor surface with an ultra-thin silica (SiO2) shell using atomic layer deposition (ALD) to facilitate the SLB formation without sacrificing the signal-to-noise ratios. This type of silica-based SPR chip functionalization is not available commercially. Furthermore, on the instrumentation side, the nanohole geometry facilitates simple optical measurements of molecular binding kinetics via recording changes in the transmission of light through nanoholes in a gold film, which is governed by a process known as extraordinary optical transmission (EOT).15,16,17 When light strikes a periodic array of nanoholes in a metallic film, surface plasmons, which are surface electromagnetic waves propagating at the metal interface, are launched. In EOT, these surface plasmons funnel through the nanoholes and enhance optical transmission at characteristic wavelengths. The wavelengths at which peaks in the transmission spectrum occur are related to the periodicity of the nanoholes, the dielectric function of the metal film and the refractive index of the overlying media. In our case, the overlying media includes the silica shell and the SLB. Binding of analytes to ligands immobilized in the SLB over the sensing surface increases the local refractive index and shifts the characteristic transmission wavelengths, analogous to the angular shift monitored in conventional SPR instruments. The positions of peaks and minima in the transmission spectrum are measured continually in real-time for label-free measurements of binding kinetics.18 While nanohole SPR sensors have been used for many proof-of-concept experiments,14,19-23 they have not been used for quantitative measurements between therapeutic antibodies and fluid SLBs. In this work, SLBs containing Sulf and GalC were formed on the silica-coated gold nanohole array, and antibody binding kinetics to these lipids was quantified in a label-free manner. Furthermore, the lipid-containing sample is injected through disposable microfluidic channels to avoid clogging, which can be problematic in permanent fluidic systems in commercial SPR instruments. The use of optimized nanohole SPR instruments tailored for SLBs has led to a high signal-to-noise ratio of ~800.</p><p>Current dogma states that natural IgM autoantibodies bind to targets with low affinity. The equilibrium-dissociation constants for natural IgM autoantibodies typically range from mM to 10 nM.24-26 In this work, however, we show that O1 and O4 bind to Sulf and GalC with significantly greater affinity; i.e. apparent dissociation constants in the sub-1 nM regime, which is at least an order of magnitude greater affinity than other natural autoantibodies. Furthermore, the nanohole SPR instrument can detect O4 binding to Sulf-containing lipid bilayers at concentrations as low as 310 pM. These high-affinity interactions may partially explain how therapeutic antibodies signal cells in culture and promote remyelination in vivo. In fact, we have shown that a natural IgM autoantibody derived from human serum with character similar to O4 and O1 can promote remyelination with a single 500-ng dose in a mouse model of MS.27 The high affinity of this class of IgMs may facilitate targeting in vivo at very low local tissue concentrations. These observations contrast the observations that, in general, the low affinity and varied pharmacokinetics of IgMs make them poor drug candidates.28</p><!><p>The mouse monoclonal IgMs O4 and O1 were purified from hybridoma supernatant using PEG precipitation.29 The antibodies were further purified by gel filtration chromatography and the purity was verified by PAGE. Antibody concentration was determined using capture ELISA compared to commercial IgMs.</p><!><p>Myelin was isolated from SJL/J mouse whole brain according to established procedures.30 Myelin quality was determined by Western blotting for the presence of myelin associated glycoprotein, myelin oligodendrocyte glycoprotein, proteolipid protein, 2′, 3′-cyclic nucleotide 3′-phosphodiesterase, and myelin basic protein, via the binding of well characterized anti-myelin lipid antibodies by direct ELISA. For antibody binding to myelin by ELISA, mouse myelin (10 μg/well diluted in PBS) was first dried overnight onto poly-D-lysine coated 96-well flat bottom plates (Nunc Immunosorp). Wells were washed twice with PBS and blocked for 2 hr with 3 % BSA/PBS, before antibodies diluted as indicated in 1 % BSA/PBS were added and incubated overnight at 4 °C. Following washing with PBS, bound mouse antibodies were detected using a goat anti-mouse μ-chain-specific alkaline phosphatase conjugated secondary antibody (Sigma) diluted in 1 % BSA/PBS for 4 hr followed by Sigma 104 phosphatase substrate (1 mg/mL) in Tris/Glycine buffer. Optical density was read at 405 nm (OD405) using a SpectraMax M2 micro plate reader (Molecular Devices, Sunnyvale, CA). Mean OD405 of triplicate values were calculated.</p><!><p>Immunocytochemistry surface staining was performed in the cold as previously described on unpermeabilized unfixed cells using 10 μg/mL of primary antibody.31 Briefly, mixed primary glial cells were prepared from Sprague-Dawley neonatal rats. Cells were plated at low density on poly-D-ornithine-coated glass cover slips (Fisher) and cells were grown for 14 days in DMEM 10 % fetal calf serum. After blocking for 10 min with 5 % BSA in HEPES buffered Earle's balanced salt solution antibodies at 10 μg/mL in 1 % BSA were added for 15 minutes. After washing and fixation for 10 min with 4 % paraformaldehyde, bound primary antibodies were detected using fluorophore-conjugated goat anti-mouse μ-chain specific Fab'2 fragments (Jackson Immunoresearch). Coverslips were mounted in using Vectashield (Vector Labs) and viewed using an epi-fluorescence microscope.</p><!><p>Standard microscope slides were cleaned by acetone, methanol, isopropyl alcohol, and water, and dried by a stream of nitrogen gas, then a 200 nm-thick gold film with a 5 nm-thick Cr adhesion layer was deposited by evaporation. Then, a 30 nm-thick Al2O3 masking layer was deposited on the gold-coated slides using ALD at 250 °C. After spin-coating a thermal nanoimprint resist (NXR-1025), a silicon imprint stamp (Lightsmyth) with a two-dimensional array of circular posts with 210 nm in diameter, 350 nm in depth, and 500 nm in periodicity was imprinted onto the resist with a pressure of 300 psi for 2 min at 130 °C. After partially removing residual resist with an oxygen dry etching, the Al2O3 was patterned using a dry etcher. Then the underlying gold film was etched with the patterned Al2O3 layer as a mask using an ion mill and the Al2O3 layer was removed. After making the nanohole array, the gold surface was uniformly covered by an 11 nm-thick SiO2 layer using ALD at 250 °C. The deposition rate was about 12 Å per cycle.</p><!><p>Vesicles were composed of egg phosphatidylcholine (Avanti Polar Lipids) and GalC and Sulf (Sigma-Aldrich). To render the vesicles fluorescent, 1,2-dimyristoyl-sn-glycero-3-phosphatidylethanolamine-N-(lissamine rhodamine sulfonyl) was included. To form the vesicles, we dissolved the appropriate relative amounts of lipids in chloroform. All lipid compositions were calculated as mole percent. The chloroform was evaporated under vacuum for at least 6 hr to form a dry lipid film in a glass vial. The dry film was then rehydrated with Tris buffer (100 mM NaCl, 10 mM Tris, 1 mM EDTA) to a total lipid concentration of 2 mg/mL and allowed to rest overnight to form a vesicle suspension. Vesicles containing 6 % Sulf and 16 % GalC were rehydrated at 55 °C. The next day, the suspension was vortex mixed then sonicated for 10 min in a bath sonicator at room temperature. The vesicle suspension was then extruded through polycarbonate filters with pore diameter of 200 nm with an Avanti Mini-Extruder. The vesicles were then diluted to 1 mg/mL in Tris buffer containing calcium (100 mM NaCl, 10 mM Tris, 10 mM CaCl2). For FRAP studies, extruded vesicles were exposed to SiO2-coated gold nanohole arrays for 1 hr, then vigorously rinsed with buffer to remove excess unruptured vesicles. The FRAP experiments were carried out with an Olympus FV1000 BX2 confocal microscope equipped with a 60× water immersion objective with NA = 0.90. The diffusion coefficients for lipids in the SLBs were calculated by fitting the normalized fluorescence recovery curves to a single exponential function. The time at which half-maximum recovery occurred (t1/2) was determined and used to determine the diffusion coefficient using the equation D = r2/4t1/2, where r is the radius of the photobleached spot. At least 3 photobleached spots were used to calculate the average D values.</p><!><p>For SPR kinetic experiments, a poly(dimethylsiloxane) (PDMS) microfluidic flow-cell was assembled onto the silica-coated nanohole array chip. The microfluidic channels were 300 μm wide and 50 μm high. The assembled sensor chip was mounted on an upright microscope stage and illuminated with a tungsten-halogen light source. The optical transmission spectra were monitored with an Ocean Optics fiber-optic spectrometer and normalized to the spectrum of the light source. The flow of all solutions was controlled with a Harvard Apparatus syringe pump. After flowing a PBS solution through the microfluidic chip for 5 min, an extruded vesicle solution in Tris buffer containing calcium was injected and incubated for 1 hr at a flow rate of 3 μL/min to form a SLB on the silica-coated nanohole sensor surface. The channel was then washed by PBS at 30 μL/min for 30 min to remove excessive vesicles and lipid overlayers. Then the surface was blocked by injecting a 1 mg/mL BSA solution at 30 μL/min for 30 min followed by PBS washing for another 30 min. For kinetic measurement, baseline transmission spectra were measured with PBS at 30 μL/min for 3 min, and then an association curve was measured by flowing an antibody solution at 30 μL/min for 5 min. The injection flow rate was tuned to minimize mass transport limitations. After 5 min, PBS was injected for 15 min to measure a dissociation curve. Between each kinetic measurement, the surface was regenerated by dissociating all antibodies bound on the SLB. This was done by flowing a 2.7 M MgCl2 solution at 30 μL/min for 20 sec followed by PBS washing for 3 min. Instrumental noise was measured by calculating the standard deviation of the peak position when PBS was flowing through the microfluidic chip with a SLB on the sensing surface.</p><!><p>The monoclonal autoantibodies O1 and O4 bind to oligodendrocytes in vitro in a developmentally dependent manner.4 Figure 1a shows a scheme illustrating the stages of oligodendrocyte development as well as which antibodies bind to the cell surface during each developmental stage. When the cells are in immature, proliferative stages O4 will bind, as will A2B5, an antibody that binds several complex gangliosides present on early oligodendrocytes.32 As the oligodendrocytes mature, they insert more Sulf into their plasma membrane and, thus, strongly bind O4. At the preoligodendrocyte stage, the cells do not express GalC or other lipids associated with O1 binding, as such O1 will not bind. When the oligodendrocytes begin the process of myelinogenesis, they express Sulf and GalC, as well as myelin basic protein (MBP) and myelin oligodendrocyte glycoprotein (MOG). Therefore, antibodies O1 and O4 will bind to the cell surface as will antibodies directed toward MBP and MOG. Figure 1b shows an immature oligodendrocyte that strongly binds O4, while Figure 1c shows a mature oligodendrocyte that strongly binds O1. Compared to the cell in Figure 1b, it is clear that the cell in Figure 1c has more fully developed processes and larger soma associated with oligodendrocyte maturity.</p><p>We also examined how O1 and O4 bind to myelin extracted from mouse brain. Figure 1d shows results from ELISA where varying concentrations of O1, O4 and an isotype control IgM (MMEN-OE5 antibacterial lipopolysaccharide IgM κ) were exposed to myelin. Assuming a molecular weight for an IgM pentamer of 900 kDa,33 the mass concentration range shown in Figure 1d corresponds to molar concentrations ranging from 69 pM to 71 nM. At the low end of the concentration scale, there is little difference in optical density between O1, O4 and control samples. However, as the concentrations of antibodies increase, O1 and O4 bind myelin in a concentration-dependent fashion. At ~0.4 μg/mL (0.44 nM), O1 and O4 binding is clearly greater than the control. Furthermore, as the concentrations increase, both O1 and O4 bind myelin with similar intensity.</p><p>Myelin is a complex mixture of lipids and proteins; about 70 % and 30 % of the dry weight respectively,34 whereas most other cell membranes are roughly 50 % lipid and 50 % protein. Combined Sulf and GalC make up over 20 % of the lipid weight of myelin.35 Because oligodendrocyte membranes and myelin are complex mixtures of lipids and proteins, we used myelin-mimicking SLBs on a nanohole SPR sensor to quantify the specific interactions between O1/GalC and O4/Sulf.</p><!><p>Periodic nanohole arrays with 200 nm hole diameter and 500 nm periodicity were made in a cm2 area of a gold film by nanoimprint lithography. After fabrication, a thin conformal layer (~11 nm) of SiO2 was deposited on the nanohole array by ALD. Figure 2a shows a schematic of the fabrication process, while Figure 2b shows a scanning electron micrograph (SEM) of a nanohole array in a gold film. The size of patterned area is scalable, similar to the nanoimprint stamp. Nanoimprint lithography patterned the nanoholes over a large area (Figure 2b inset) to allow integration with a multi-channel microfluidic flow cell.</p><p>SLBs have been employed widely for studying membrane biochemistry and biophysics.36-38 SLBs were formed by spontaneous rupture of phospholipid vesicles on SiO2, which promotes the rupture of phospholipid vesicles by a well characterized mechanism.39,40 The vesicles used consisted of egg phosphatidylcholine (egg PC) and 2 % GalC or 2 % Sulf. When necessary, 1 % of a fluorescent lipid, 1,2-dimyristoyl-sn-glycero-3-phosphatidylethanolamine-N-(lissamine rhodamine sulfonyl) (Rho-PE), was included for imaging. Figure 2c shows a schematic of a SLB on a nanohole array.</p><p>The formation of SLBs on nanohole arrays was confirmed by fluorescence recovery after photobleaching (FRAP).41 For FRAP, small circular areas of the SLBs were photobleached, and the recovery of fluorescence in the photobleached area was monitored as a function of time to generate recovery curves. Figures 3a shows four frames from a photobleached SLB on a nanohole array. Figure 3b shows representative recovery curves obtained from FRAP experiments on SLBs with 2 % GalC, 2 % Sulf and a control SLB free of GalC or Sulf. The recovery curves show that the lipid mixtures do indeed form continuous fluid SLBs on SiO2-coated nanohole arrays. The diffusion coefficients (D) for lipids in the SLBs containing Sulf and GalC SLBs were not significantly different from the Egg PC control, as determined by unpaired t-test. The D for the control was calculated to be 1.84 ± 0.14 μm2/s, while for the 2 % Sulf SLB, D = 1.60 ± 0.14 μm2/s, and for the 2 % GalC SLB, D = 1.95 ± 0.16 μm2/s (mean ± standard deviation). The D values observed here are somewhat higher than those obtained for SLBs on random nanohole arrays in SiO2-coated gold,22 but lower than those observed on nonporous flat SiO2.42 The maximum fluorescence recovery for the SLBs on nanohole arrays is 75 – 80 %, which suggests an immobile lipid fraction and/or unruptured vesicles on the surface. Because the SLBs conform to the nanohole topography,43 there may be pinning of lipids at tightly curved surfaces at the rim of the nanoholes, around the interior of the nanohole surface and around the base of the nanohole which hinders diffusion and leads to an apparent immobile fraction.</p><!><p>For SPR sensing a poly(dimethylsiloxane) (PDMS) microfluidic chip prepared by soft lithography44 was attached to the nanohole SPR sensor. Figure 4a shows transmission spectra for a SiO2-coated gold nanohole array with hole diameter of 200 nm and periodicity of 500 nm, the same nanohole array after formation of a SLB containing 2 % Sulf and after O4 binds to Sulf in the SLB. When the SLB is formed over the nanohole array the optical transmission spectrum shifts toward longer wavelengths by approximately 2 nm. This shift was consistent between different channels on the chip and from experiment to experiment, which confirmed that the SLBs formed on the nanohole arrays without fluorescence imaging. Yang and Cremer previously showed that SLBs will readily form in microfluidic channels.45 The injection of 44 nM O4 red-shifts the spectrum by 0.8 nm. While the magnitudes of the shifts seem small, the standard deviation of the noise in our instrument is extremely small, typically in the range of 10−3 nm, which means that the limit of detection (defined as a signal to noise (S/N) ratio of 3) is a shift of 0.003 nm and the S/N for a 0.8 nm spectral shift is on the order 800.</p><p>In nanohole-based SPR sensors, kinetic sensorgrams can be recorded by tracking either the transmission maximum (peak) or minimum (dip) in real time.46 In our system, binding kinetics extracted from the transmission minimum showed a slightly better signal-to-noise ratio, thus we tracked the position of the local minimum in the EOT spectrum around 700 nm (Figure 4a) as a function of time. Figure 4b shows a sensorgram encompassing an entire experiment from SLB formation to multiple antibody injections and surface regeneration. The course of the experiment was as follows. A SLB containing Sulf was formed on the nanohole sensor, resulting in a spectral shift of approximately 2 nm after washing excess vesicles from the sensor. Then the SLB was blocked with BSA to minimize nonspecific adsorption. Following blocking, 50 nM O1 was injected, and no significant shift was observed, indicating that O1 does not bind to a Sulf-containing egg PC membrane, consistent with previous reports.5 After O1, 44 nM O4 was injected and the spectrum appreciably red-shifts to a maximum of about 0.8 nm. After O4 association, the injection solution was switched to antibody-free phosphate buffered saline (PBS) to monitor O4 dissociation. After O4 dissociation, the remaining O4 was removed from the surface with a high ionic strength regeneration solution containing 2.7 M MgCl2.47 The large fluctuations in position seen in the sensorgram during this step are due to the vastly different refractive index of the regeneration solution. After the regeneration wash, the peak position nearly returned to the pre-injection position, indicating that the regeneration wash removed the bound O4, but did not disrupt the SLB on the nanohole sensor. After regenerating the surface, repeated injections of 50 nM O1 and 44 nM O4 confirm that the SLB is intact and that serial measurements could be carried out on our nanohole-SLB platform.</p><p>Figures 5a and 5b show binding kinetic curves for multiple concentrations of O4 and O1 binding to SLBs containing 2 % Sulf and 2 % GalC, respectively. The curves in Figures 5a and 5b show that the spectral shift increases with increasing concentrations of antibody injected, as expected. After injection of the antibodies, the injection solution was switched to antibody-free PBS. Upon switching the injection solution, the antibodies dissociate from the SLB surface, and spectral shift decay is observed. The association and dissociation curves were fit to exponential functions to determine the association and dissociation rate constants (kon and koff).48 The kon and koff rate constants for O1 binding to GalC were (6.98 ± 4.92) × 105 M−1 s−1 and (1.10 ± 0.51) × 10−3 s−1 (mean ± standard error of the mean (s.e.m.)), respectively. For O4 binding to Sulf kon and koff were (2.65 ± 1.71) × 105 M−1 s−1 and (3.73 ± 1.15) × 10−4 s −1, respectively. The ratio of the rate constants calculated for each individual experiment were used to determine the apparent dissociation constant (KD) with the equation KD = koff / kon. The KD values from individual experiments (N = 3) were averaged to calculate the mean KD value. The KD values for GalC/O1 and Sulf/O4 were (2.37 ± 0.56) nM and (2.19 ± 0.61) nM, respectively (mean ± s.e.m.). Table 1 shows the KD values obtained from these studies compared to KD values determined for other IgM natural autoantibodies. The KD values for O1/GalC and O4/Sulf interactions are nearly 20-fold smaller than previously reported values, indicating that these antibodies have unexpectedly high affinities for their targets compared to other IgM autoantibodies.</p><p>We conducted a number of negative control experiments to confirm that the binding of antibodies was due to specific interactions. (Figure 5c) With both 20 nM and 50 nM O1 and O4, no binding was observed to SLBs composed solely of Egg PC. Also, when a nanohole array without a SLB was exposed to 1 mg/mL BSA, similar to the blocking step mentioned above, and then 20 nM O1 or O4 were injected, no spectrum shifts were observed. This indicates that neither O1 nor O4 nonspecifically adsorb to the egg PC matrix of the SLB, or the BSA used for blocking.</p><p>Sulf and GalC make up 6 mole % and 16 mole % of the myelin lipid composition, respectively.53 To better mimic the lipid composition of myelin, a SLB was formed on the nanohole sensor containing 6 % Sulf and 16 % GalC in a matrix of egg PC. With this SLB on the sensor, lower concentrations of O4 were injected to approach the detection limit for this system. We injected 1.25 nM and 310 pM O4 and both concentrations were readily detectable. (Figure 5d) The kinetic rate constants obtained for O4 binding to these myelin-mimicking SLBs were 1.26 × 105 M−1 s−1 (kon) and 1.16 × 10−4 s−1 (koff) and the apparent KD was calculated to be 920 pM. This indicates that O4 binds myelin-like membranes with unusually high affinity for a natural autoantibody.</p><p>Because little is known about the affinities of O1 and O4 for GalC and Sulf, it is difficult to compare the results obtained with nanohole SPR sensors to results obtained with commercial SPR instruments or other alternative methods. However, in previous work, our group compared kinetic rate constants and KD values obtained from a nanohole SPR instrument to those obtained from a commercial SPR instrument.16 Antibodies binding to non-lipid antigens with known KD values ranging from 200 pM to 40 nM were examined. For all binding interactions we studied, the rate constants and KD values obtained with nanohole SPR were quite similar to those measured with a commercial Biacore 3000 system. This confirms that the nanohole SPR platform is capable of accurate quantification of binding kinetics.</p><p>The KD values obtained in this study were compared to those of other antibodies from previously published studies. (Table 1) The antibodies listed in Table 1 are all IgM natural autoantibodies, and the KD values determined in previous studies range from 100 μM to 40 nM. The smallest KD value is almost 20-fold larger than the values determined in this study. In the previous examples the KD values were determined by SPR or equilibrium assays, such as ELISA. However, because traditional equilibrium assays only offer a static snapshot of the situation at equilibrium, they cannot determine the rate constants kon and koff. Furthermore, the dissociation constant can be calculated from the ratio of the dissociation and association rate constants. Thus, it is possible for two given antibody-antigen interactions to have the same KD value, but drastically different kon and koff values. For any type of potential therapeutic molecule, including remyelinating IgMs, the koff can be a more powerful predictor of potential efficacy than KD.54 Regardless of this fact, in vitro measurements of KD have become acceptable stand-ins for in vivo efficacy in the drug discovery process.55 There are examples, however, where KD and efficacy are well correlated.56 Because drugs are only effective when bound to their receptor, the residence time of a drug at its receptor is a better measure of drug-receptor interactions than the affinity. Residence time is inversely proportional to koff and correlates better with in vivo efficacy and duration of action than does KD.54 Therefore, kinetic-sensing schemes that determine rate constants are preferable when investigating potential therapeutic molecules.</p><!><p>IgM antibodies can bind antigens with high avidity because a single IgM is capable of binding 10 antigens. Therefore it is likely that many of the binding events on our SPR sensors are polyvalent. When an antibody encounters the SLB surface, only one antibody/antigen binding event is required for association and a measured shift in the SPR signal. Subsequent binding events for a single antibody do not contribute to major shifts in the SPR signal because the antibody has already associated with the SLB. On the flip side, for an IgM to completely dissociate from the sensor surface, all of the antibody/antigen interactions must dissociate. Thus, the increased binding avidity of IgMs could contribute to the relatively low apparent KD values observed in this study. We are primarily concerned with comparing O1 and O4 to other IgM natural autoantibodies, which also can display high binding avidity. A particularly apt comparison is to look at kon and koff rate constants for other IgM natural autoantibodies studied by SPR and compare those to the rate constants obtained for O4/Sulf and O1/GalC binding using myelin-mimicking lipid membranes. For example, the F5-2 natural autoantibody studied by Diaw et al. had kon rate constants that were between 9 and 42-times slower (depending on antigen) than those measured for O4/Sulf and O1/GalC.49 (Table 1) The koff rate contants for F5-2 were a maximum of 3-times faster than that observed for O4/Sulf and O1/GalC binding. Because KD is a ratio of these rate constants, the lower KD values observed in our study are largely due to the increased kon of O4 binding to Sulf and O1 binding to GalC. This suggests that the initial recognition process that leads to antibody association, i.e. the affinity between a single antibody binding site and an antigen, is the primary cause of the smaller apparent KD (higher apparent affinity) of O1 and O4 antibodies, rather than increased avidity of O1 and O4 for GalC and Sulf, respectively. An alternative, though unlikely, explanation for the increased kon rates is that the sensor nanohole architecture, to which the SLB with antigens conforms, may artificially increase association rates and/or slow dissociation rates by making lipid antigens more available for initial binding or multivalent binding due to the curvature of the SLB. However, because the nanoholes themselves comprise only 12.5 % of the sensor area, this is unlikely to have a pronounced effect.</p><!><p>The present experiments have great significance in defining the mechanism of action by which mouse and human monoclonal antibodies promote remyelination.57 The human antibodies that promote remyelination have also been found to be of the IgM subclass and have the characteristic features of natural autoantibodies. No human IgG antibodies that promote tissue repair in vivo have been identified. Efforts to convert the IgMs to an IgG form, either by spontaneous mutation or by molecular construction, have resulted in the absence of Ca2+ responses in cultured glial cells but, most importantly, an absence of in vivo remyelinating activity.58 Therefore, the high-affinity, high-avidity pentameric structure is essential for the therapeutic response of those antibodies. Traditionally, IgM antibodies, specifically natural autoantibodies, have been considered to have low affinity binding.24-26 The fact that our present experiments show the opposite result, that is, small apparent KD values, may explain why these antibodies have therapeutic efficacy in vivo. In vivo experiments using magnetic resonance imaging show human natural autoantibodies cross the blood brain barrier and indicate that if the antibody has a clear target to bind within the central nervous system (CNS), then the therapeutic antibody can be detected for long as one week in the CNS.59 Similar results have been obtained with 35S-labeled antibodies using EM autoradiography.60 The fact that O1 and O4, despite being IgMs, bind tightly to their antigens is further confirmation of their strong affinity and being IgMs they have high avidity for the myelin target. On the technology side, the combination of nanohole SPR sensor chip with biomimetic SLBs shows great promise for studying these important lipid-antibody reactions in a quantitative manner to reveal not only equilibrium constants but also on-off rates that can help understand the mechanisms of molecular recognition as well as stability of binding. Finally, as human natural autoantibodies begin entering clinical trial, these results may help to determine the dosing schedule for optimum therapeutic efficacy.</p>

PubMed Author Manuscript

The diverse chemistry of cytochrome P450 17A1 (P450c17, CYP17A1)

The steroid hydroxylation and carbon-carbon bond cleavage activities of cytochrome P450 17A1 (CYP17A1) are responsible for the production of glucocorticoids and androgens, respectively. The inhibition of androgen synthesis is an important strategy to treat androgen-dependent prostate cancer. We discuss the different enzymatic activities towards the various substrates of CYP17A1, demonstrating its promiscuity. Additionally, a novel interhelical interaction is proposed between the F-G loop and the B\xe2\x80\xb2-helix to explain the 16\xce\xb1-hydroxylase activity of human CYP17A1 with progesterone as the substrate. The techniques used by biochemists to study this important enzyme are also summarized.

the_diverse_chemistry_of_cytochrome_p450_17a1_(p450c17,_cyp17a1)

1. Cytochrome P450 enzymes<!>2. CYP17A1 gene, physiology, and disease<!>3.1. Steroid substrates for CYP17A1<!>3.2. Steroid products of CYP17A1<!>3.3. Experimental systems and approaches<!>4.1. Synthesis of isotopically labeled substrates<!>4.2. Metabolic switching<!>4.3.1. KIE theory<!>4.3.2. Intermolecular KIE<!>4.3.3. Intramolecular KIE<!>5.1. Progesterone 16\xce\xb1-Hydroxylation and alanine-105<!>5.2. Progesterone 21-Hydroxylation<!>5.3. Inversion of configuration at C-16<!>6. CYP17A1-catalyzed carbon-carbon bond cleavage reactions<!>7. Conclusions and future directions

<p>Microsomal human cytochrome P450 17A1 (CYP17A1, 17α-hydroxylase, 17,20-lyase) belongs to the cytochrome P450 super family that contains a conserved cysteine residue, which provides the axial sulfur ligand attached to a heme prosthetic group. The P450 class of enzymes conducts a myriad of chemical transformations including oxidation, reduction, and non-redox reactions [1]. The prototypical reaction catalyzed by this family of enzymes is the oxygenation of C-H bonds to afford alcohol products. This process involves the two-electron reduction of a molecule of oxygen where one of the oxygen atoms goes to water and the other is incorporated into the alkane substrate [2]. The oxoiron active intermediate (compound I) responsible for C-H activation has been characterized by Mossbauer and UV-Vis spectroscopy [3]. The compound I intermediate was isolated through stopped flow by mixing CYP119 in one syringe and m-chloroperoxybenzoic acid in the other syringe (2:1 ratio) followed by freeze quenching with liquid ethane. Most P450 enzymes share a common 3-dimensional structure. Deisenhoefer and co-workers compared the crystal structures of the soluble bacterial enzymes P450cam (CYP101), P450terp and P450BM-3 and observed structural similarities in the three different proteins: 13 α-helices (A, B, B′, C-L), and 5 β-sheets (β1-β5) [4]. The x-ray crystal structures of several eukaryotic, membrane-bound P450 enzymes reveal the same basic fold with some variable areas. There are 57 human cytochrome P450 genes (that metabolize small molecules including fatty acids, xenobiotics, and steroids. Six of these (CYP11A, CYP11B1, CYP11B2, CYP17A1, CYP19A1, CYP21A2) are found in the steroid hormone biosynthesis pathway [5]. Crystal structures of CYP17A1 with azole inhibitors bound have been recently elucidated, and these structures can be used to increase our understanding of this enzyme [6]. This report will focus on the hydroxylation activity of CYP17A1, although the carbon-carbon (C-C) bond cleavage activity is also important and a topic of intense study [7]. The use of steroid analogs to show the diverse reactivity of this enzyme can enhance our understanding of the function of this important enzyme in human physiology and disease. The goal of this review is to educate the reader of the different approaches to study the enzymology of CYP17A1. The function of this enzyme may be elucidated with the information gained from biochemical studies.</p><!><p>The human CYP17A1 gene is localized to chromosome 10q24.3 [8]. This gene is expressed in the adrenals and gonads, with minor amounts in the brain, placenta, and heart [9]. The same mRNA and protein is expressed in the adrenal and gonad. The 17-hydroxylase activity of CYP17A1 is required for the production of the glucocorticoid cortisol, whereas the 17,20-lyase activity leads to androgens (and in turn estrogens). Most clinical mutations in CYP17A1 that have been reported result in the loss of both 17-hydroxylase and 17,20-lyase activities. The loss of all or most CYP17A1 activities leads to 17-hydroxylase/17,20-lyase deficiency (17OHD), a form of congenital adrenal hyperplasia. The phenotype of 17OHD consists primarily of hypertension and hypokalemia—from the accumulation of the mineralocorticoids 11-deoxycorticosterone (DOC) and corticosterone upstream of cortisol, plus sexual infantilism—from a lack of androgen and estrogen synthesis.</p><p>An interesting variant of 17OHD has been reported in a few patients, which results in the loss of solely the 17,20-lyase activity [10]. This loss of the C-C bond cleavage activity results in the loss of the production of 19-carbon androgens, which leads to undervirilization of 46,XY infants and pubertal failure in 46,XX children.</p><p>A more common condition involving CYP17A1 is not enzyme deficiency but rather androgen-dependent cancers and hyperplasias. Prostate cancer is typically androgen-dependent, and medical or surgical castration induces remission for most patients early in the course of the disease. Later, the disease progresses despite low testosterone (T) synthesis, a condition called castration-resistant prostate cancer (CRPC). CYP17A1 inhibitors ketoconazole, abiraterone acetate, and newer agents have been developed to treat these diseases [11, 12]. The androgen dependent-cancer patient usually takes the CYP17A1 inhibitor along with a glucocorticoid, prednisone, to lower adrenocorticotropin (ACTH) and prevent the accumulation of cortisol precursors that cause hypertension and hypokalemia [13].</p><!><p>The first of two major physiologic steroid substrates of CYP17A1 contains the pregnenolone backbone (1) in the A/B-rings of the steroid, which consists of a 3β-hydroxy-Δ5,6 moiety. The other major steroid substrates contain the progesterone backbone (6), which is composed of a 3-keto-Δ4,5 enone group in the A/B-rings. In vivo, the 3-keto-Δ4,5 steroid backbone arises from the 3β-hydroxy-Δ5,6 moiety by the activity of 3β-hydroxysteroid dehydrogenase/isomerase enzymes (3βHSD, two isoenzymes in humans). In the crystal structure of abiraterone complexed with CYP17A1, the polar side chain of asparagine-202 hydrogen bonds to the 3β-hydroxyl group of abiraterone, which mimics the 3β-hydroxyl group in pregnenolone [6].</p><p>Although these two backbones are common to most steroidogenic pathways and lead to most active steroid hormones, the active site of CYP17A1 accommodates additional similar configurations of the steroid A/B-rings. The 3-keto, 5α-reduced steroid backbone (5α-dihydroprogesterone, 12), which arises from the action of 5α-reductases (2 major isoenzymes in humans) on the 3-keto-Δ4,5 steroids, is a viable substrate class for CYP17A1 [14]. This reduced steroid indicates that a double bond in the A/B ring system is not a structural requirement for substrates. More importantly, 5α-pregnan-3α-ol-20-one (allopregnanolone, 15), which is derived from 5α-dihydroprogesterone from the action of the reductive 3αHSD enzymes (aldo-keto reductase types 1C, AKR1C), is also a substrate for CYP17A1 [14]. The product of the 17,20-lyase reaction in this series is androsterone (17), which can be converted in two steps to dihydrotestosterone (DHT), the most potent endogenous androgen, without the intermediacy of T. This alternate or "backdoor" pathway to DHT significantly contributes to androgen production in certain pathologic states and explains some discrepancies between serum T concentrations and clinical parameters of androgen action. The positioning of the axial 3α-hydroxyl of allopregnanolone (15) within a trans-decalin system is vastly different than the equatorial 3β-hydroxyl of pregnenolone (1) in a twist-boat ring, yet both are excellent substrates for CYP17A1. This striking structural diversity demonstrates that the enzyme shows considerable tolerance for the orientation of the 3-hydroxy (or 3-keto) group, despite the hydrogen bonding seen in the crystal structures with bound 3β-hydroxy inhibitors.</p><!><p>Pregnenolone (1) is the first 21-carbon steroid precursor found in the steroid biosynthesis pathway and is derived from cholesterol by the action of CYP11A1. CYP17A1 converts pregnenolone (1) exclusively to 17-hydroxypregnenolone (17Preg, 2) and also cleaves the C17–C20 carbon-carbon bond of 17Preg to dehydroepiandrosterone (DHEA, 3) (Figure 1). The presence of cytochrome b5 (b5) enhances the rate of the 17,20-lyase reaction to form DHEA from 17Preg by an order of magnitude. With pregnenolone substrate, the presence of b5 diverts 1–10% of the product to androsta-5,16-dien-3β-ol (5) (Fig. 1), depending on species and incubation conditions. With allopregnanolone substrate (15), CYP17A1 catalyzes the same reactions: 17-hydroxylation, 17,20-lyase, and andiene formation. While b5 stimulates the same two activities, androsterone (17) formation is very efficient and stimulated only 3-fold in the presence of b5. CYP17A1 primarily 17-hydroxylates progesterone (6) to 17-hydroxyprogesterone (17OHP, 7), but 20% of the product formed is 16α-hydroxyprogesterone (16OHP, 8), and a trace of DOC, which is 21-hydroxyprogesterone (9). The biological significance of 16OHP (8) has been studied [15], but its physiological role remains unclear. Although not discussed in the original report of 5α-dihydroprogesterone as a CYP17A1 substrate [14], the 16α-hydroxylated steroid appears to be a minor product, similar to the metabolism of progesterone. For both 3-ketosteroid substrate series, the 17,20-lyase reaction is very slow relative to the rate with 3-hydroxysteroids. Consequently, CYP17A1 shows considerable promiscuity: several substrates, each with different products; stimulation and product diversion via b5; and both hydroxylation and carbon-carbon bond cleavage chemistries. Some of these phenomena are very important in human physiology, such as b5 stimulation of androgen synthesis in specific steroidogenic cell types. Other processes might be important in disease states, such as the ordinarily minor 21-hydroxylation reaction in genetic 21-hydroxylase deficiency and a source of adrenal-derived mineralocorticoid.</p><!><p>To understand the current state of knowledge, we provide a brief overview of the experimental systems and paradigms that enzymologists use to study CYP17A1 activities in vitro, in order to understanding the physiological roles of the enzyme and its variants. For example, how do we know that clinical mutations R358Q or R347H, which cause isolated 17,20-lyase deficiency, disrupt positive charges on the enzyme surface and impair interactions with POR and b5 [16]. The reactivity of CYP17A1 can be studied in vitro by expressing the native enzyme (or mutated variants) in mammalian cells or in yeast together with its redox partners human (or yeast) P450-oxidoreductase (POR) and/or b5 and isolating microsomes with the enzyme system [17]. Alternatively, modified and C-terminal polyhistidine tagged CYP17A1 enzyme is expressed in E. coli, purified using a Ni-NTA column [18], and studied in reconstituted assays with phospholipid and purified POR [18], with or without b5. The steroid substrate, which may be labeled with a stable isotope or radioisotope, is added in a small volume of water-miscible solvent. As the source of electrons, these incubations use NADPH or an NADPH generating system containing glucose-6-phosphate, glucose-6-phosphate dehydrogenase, and NADP+ [19]. For more detailed descriptions of experimental protocols, refer to the references cited within this section.</p><p>After extraction with an organic solvent, the products are resolved using high performance liquid chromatography (HPLC, typically reverse phase) and quantitated with UV-detection at 254 nm and/or scintillation counting when radiolabeled steroid is used [20]. Alternatively, liquid chromatography-mass spectrometry (LC-MS) may be employed, and this technique can assess for changes in isotopic labeling. For example, if the labeled carbon site is a methylene (-CH2-) and the labeling is stereospecific (-CHαDβ- or -CHβDα-), UV detection cannot distinguish between hydroxylated products with and without deuterium incorporated at that same carbon site. Thin-layer chromatography (TLC) analysis is another option: radiolabeled steroids should be used ([14C]- or [3H]- labeled) to detect steroids by phosphorimaging [16] or by scraping the sections containing the steroids from the TLC plates for quantifyication by liquid scintillation counting.</p><p>Each method has advantages and disadvantages. Although mass spectrometry is useful in measuring both the quantities and isotopic composition of products, some compounds ionize poorly and thus require derivatization, and the standardization is necessary to measure multiple compounds simultaneously with different ionization efficiencies. UV-detection is generally less sensitive than mass spectrometry and more vulnerable to interfering substances, and the compound must have a chromophore to be UV-active. For CYP17A1, the 3-keto-Δ4,5-series of steroids are amenable to UV-detection at 254 nm (ε~16,000 M−1cm−1) [21]. Steroids containing the 3β-hydroxy-Δ5,6-functional groups, in contrast, must be converted to the enone to enable UV-detection, and cholesterol oxidase is often used for this purpose. It is possible to detect the pregnenolone metabolites with UV-absorption in the far-UV such as 216 nm [22], but the absorbance is very weak. Radioactive substrates with [14C]- or [3H]-labels sites remote from the reacting carbon(s) offer high sensitivity and minimal interference but require exposure precautions and scintillation counting [20]. These radiolabeled compounds are expensive, and not all are commercially available. Liquid chromatography methods are time consuming [20], while the TLC method [16] allows one to simultaneously run multiple samples on each plate.</p><!><p>To add to the arsenal of the enzymologist, kinetic isotope effects have been used to determine the bond-breaking step(s) [20]. Deuterium-labeled substrates have been used for decades to study mechanistic and energetic aspects of enzymatic reactions involving C-H bond-breaking chemistry, such as cytochrome P450-catalyzed hydroxylations [23]. These experiments require access to the substrate bearing deuterium label at the bonds where chemistry occurs. Methods to regioselectively and stereoselectively introduce deuterium only at the desired positions are preferable to base-catalyzed exchange (i.e. KOD in CH3OD). For example, in order to study the 17-hydroxylation of progesterone, these conditions would not only exchange the C17-proton but all C21-protons, as well as the C2-, C4-, and C6-protons. We have found that convenient access to the deuterated compounds is achieved with oxidative addition of zinc onto an alkyl bromide precursor, followed by quenching with a deuteron source in situ (Fig. 2) [20]. In most instances, we treat the brominated steroid precursor with zinc dust in deuterated acetic acid, the latter being generated inexpensively with D2O and acetic anhydride in diethyl ether or other aprotic solvent. The desired deuterium labeling is then confirmed with 1H NMR (Fig. 3) and MS. As shown in Fig. 3B, the deuterium incorporation might not reach 100% (presence of minor amounts of H-17 at a chemical shift of ~2.6 ppm) for various reasons, primarily: (i) the H-17 of progesterone was exchanged during the purification or handling process, or (ii) in the deuteration source (Fig. 2) is not 100% deuterated, from traces of atmospheric moisture in the solvent, reagents, and glassware. The isotopic abundance is measured by integrating the peaks of the protons in the NMR spectra and/or by measuring the exact masses of the desired deuterium-enriched compounds by MS.</p><p>This zinc-mediated chemistry does not require the bromine to be in the α-position relative to the C20-carbonyl (i.e. 17- or 21-position). Thus, 16α-bromopregnenolone acetate also undergoes regioselective and stereospecific deuteration at H-16α under these conditions, even though the bromine substituent is in the non-exchangeable β-position relative to the C20-carbonyl [20]. The hydroxylation activities of CYP17A1 were confirmed by incubation of regioselectively deuterated (either 16α-, 17-, or 21,21,21- positions) steroid substrates with purified enzyme from E. Coli in reconstituted assays and with microsomes from yeast expressing CYP17A1 and POR [20].</p><!><p>Metabolic switching occurs when the activated P450 enzyme (Fe=O) forms a complex with its substrate at the preferred reaction site, CA (H-CA), and the complex (i.e. Fe=O---H-CA) "switches" to an alternative C-H site, CB (i.e. Fe=O---H-CB). The most common approach to experimentally induce metabolic switching is substitution of deuterium or tritium for hydrogen on CA. Metabolic switching may be observed for any enzyme that hydroxylates more than one carbon position on a substrate to give rise to two or more possible products. Selective deuteration at the major reaction site decreases the rate of hydroxylation at that site (see section 4.3 below). While the overall substrate conversion to product remains unaffected, the proportion of the minor product increases [24]. Metabolic switching is usually not absolute in that the reaction at the primary site still occurs, but the proportion of products is variably changed. Deuteration will not induce metabolic switching if the energy barriers to reach alternative reaction sites are too high, rendering other sites of reaction inaccessible.</p><p>For CYP17A1, the metabolic switching phenomenon was nicely illustrated for [2H]-labeled progesterone substrates. Using 17-[2H]-progesterone as substrate (Fig. 3B, 20), the product distribution of 17OHP:16OHP shifted from 4:1 to 1:1, without significantly affecting the conversion from starting material to hydroxylated products. Conversely, 16α-[2H]-progesterone (Fig. 3A, 19) afforded a 9:1 ratio of 17OHP:16OHP products. In contrast, for 17-[2H]-pregnenolone substrate, the only product remained 17Preg [20]. Thus, CYP17A1 exhibits metabolic switching for some substrates but not others. Additionally, deuteration at the C17-position of progesterone increased CYP17A1-catalyzed hydroxylation at a previously unidentified position, C21, which yields deoxycorticosterone as the product.</p><!><p>The substitution of deuterium or tritium for hydrogen in the position of hydroxylation on the steroid ring might not only induce metabolic switching but also change the reaction kinetics, resulting in a KIE. Figure 4A illustrates the origin of the KIE and its use in studying CYP17A1 enzymology. The zero-point energy of the C-H bond is dependent on the reduced mass, based on Hooke's law and the harmonic oscillator model. The substitution of deuterium for hydrogen in a C-H bond lowers the zero point energy by about 1.15 kcal/mol. In a symmetric transition state, the distances between the hydrogen atom and FeO (acceptor) and substrate (donor) are the same. Since this vibrational energy is excluded in a symmetrical transition state, the activation energy (Ea or ΔG‡) is greater for a C-D bond, and thus the reaction is slowed by a factor equal to the KIE. Metabolic switching is observed in the HPLC-UV analysis of the incubation extracts (Fig. 4B).</p><!><p>The change in steady-state kinetics with labeled substrate (KIE) reflects the contribution of C-H bond breaking to the overall reaction rate. The magnitude of a primary KIE (KIE on the bond to be broken) varies from a maximum of ~7 if the transition state is symmetrical and C-H(D) bond-breaking step is fully rate limiting to the overall reaction rate to near 1 if the transition state is very asymmetrical and/or the C-H(D) bond-breaking step is much faster than one or more other microscopic steps, either chemical or physical [25]. A secondary KIE occurs when deuterium substitution changes the rate of the reaction, but the C-D bond is not broken. Most commonly, secondary KIEs result from changes in carbon hybridization, such as sp2 to sp3. However, the magnitude of the secondary KIE is not nearly as high as the primary KIE, in the range of 0.7–1.3 [26].</p><p>These experiments comparing the overall rate of the specific reaction involving C-H or C-D bond breaking are used to determine what is termed an intermolecular KIE. The experiment can be conducted by incubating the two different substrates (labeled and unlabeled) with the enzyme simultaneously in the same reaction tube, and the resulting products are measured using one or more techniques, which discriminate the contribution from two isotopically distinct substrates—such as UV detection, radiochemical detection, or mass spectrometry. For CYP17A1, we employed progesterone with a [3H]- or [14C]-label remote from the reaction site for the "unlabeled" (C-H) substrate plus the second progesterone substrate, which contained a deuterium at the reaction site and was used in >1,000-fold excess. The radiolabeled steroids were measured with scintillation counting, while the products from the deuterium-labeled progesterone were detected by UV-absorption at 254 nm [20]. An alternative approach is to label both substrates differently at remote site(s) with stable isotopes and to measure the products, using mass spectrometry to distinguish the products derived from each substrate.</p><p>Another approach for measuring KIE values is to separately determine the kinetic constants kcat, Vmax, and KM at steady state conditions, and this measurement is termed non-competitive kinetic isotope effect. These conditions require: (i) multiple incubations with varying substrate concentrations (several points ranging from 100 nM to 500 μM), (ii) uniform incubation times with adequate turnover at all concentrations to accurately measure products (i.e. ideally, each point should not surpass 20% conversion of starting material to product due to competing steps that may underestimate the rate such as product release), and (iii) use of a curve-fitting software such as GraphPad Prism to fit the v ((moles of product formed)*(moles of enzyme) −1*(time) −1) vs. [S] data to the Michaelis-Menten or other appropriate equation, to obtain these constants. Each approach has advantages and disadvantages related to cost, ease of synthesis, accuracy of determination, and confounding factors. Ideally, the data are obtained from more than one method to ensure internal consistency.</p><!><p>If C-H(D) bond-breaking is not significantly rate limiting, deuterium labeled substrates might still provide information on reaction energetics if more than one chemically equivalent hydrogen atom is subject to abstraction or if metabolic switching is involved, via the intramolecular KIE [27–29]. In the former case with more than one equivalent C-H bond (i.e. a -CH3 group, with 3 hydrogen atoms that are chemically equivalent due to free rotation), the site can be isotopically labeled with a single deuterium or tritium atom and subjected to hydroxylation chemistry. The intramolecular KIE is calculated from the isotopic composition of the products using mass spectrometry. For a methylene (-CH2-) group, similar calculations of intramolecular KIE apply, with the caveat that for an enzymatic reaction, the two hydrogen atoms might not be chemically equivalent if they are also diastereotopic. In the latter case in which metabolic switching occurs, the KIE is directly proportional to the change in product distribution with the deuterated substrate. For example, the 17OHP/16OHP product ratio for CYP17A1-catalyzed progesterone hydroxylation falls from 4:1 with natural abundance substrate to 1:1 with 17-[2H]-progesterone substrate (Fig. 3B, 20), yielding an intramolecular KIE of ~4 (Fig. 4B(ii), Fig. 5).</p><p>The C-H abstraction step is only one of several steps in the P450 reaction cycle and the C-H hydroxylation process, which includes substrate binding, electron transfers, oxygen activation, C-H abstraction, radical recombination, product release, redox partner binding, and any other enzyme conformational changes (Fig. 5). A comparison of the intermolecular and intramolecular KIE values reveals any "masking" of the intermolecular KIE, meaning a reduction in magnitude. The intramolecular KIE measurements are always higher than the intermolecular KIE values, since the intramolecular KIE values directly reflect the energetics of the C-H vs. C-D abstraction step without any masking. The intermolecular KIE value is masked or lowered when the transition state is asymmetrical or when C-H(D) bond breaking is not significantly rate-limiting. On the other hand, larger KIE values (~80) have been observed in the case where hydrogen atom tunneling is involved [30]. Hydrogen atom tunneling, or proton-coupled electron transfer, is invoked to explain anomalously large intramolecular KIE values, because in these cases the rate of the reaction is exquisitely sensitive to bond length.</p><!><p>Baboon CYP17A1 has only 10% 16α-hydroxylase activity with progesterone substrate, compared to ~20% for human CYP17A1. This observation prompted the Swart laboratory to probe which residue(s) might be responsible for this difference in 16α-hydroxylase activity between the two species. In the original report, these investigators mutated a region in the F-G loop of human CYP17A1 to the sequence found in baboon CYP17A1 (AVIE to KIVH), but this change did not affect the 16α-hydroxylase activity. Instead, mutation of A105 to leucine, as found in the baboon enzyme, did not change 17-hydroxylase activity but raised the 17OHP:16OHP product ratio from 4:1 to 9:1, similar to baboon CYP17A1. Thus, residue 105 is the most important residue yet identified for maintaining the progesterone 16α-hydroxylase activity of human CYP17A1 [31]. The contributions of these differences in 16α-hydroxylase chemistry to adrenal physiologies among primates remains to be explored.</p><p>X-ray crystallography is another method employed by enzymologists to gain a structural understanding to rationalize enzymatic activity. Two crystal structures of CYP17A1 bound to either abiraterone or galeterone (TOK-001) were reported by Devore et al. [6], which provided insight to some key amino acid residues in the active site of CYP17A1 and the origins of its promiscuity for steroid hydroxylation. Alanine 105 is located on B′ helix of human CYP17A1. This amino acid residue does not seem to directly interact with substrate, because the structure of CYP17A1 with bound abiraterone shows that the alanine-105 side chain is 6.4 Å away from the C-4 position of abiraterone. Leucine-209 on the F-helix, however, is 3.7 Å away from the methyl group of alanine-105, suggesting a direct hydrophobic interaction (Fig. 6). The omnipresent F-G loop in P450 enzymes is viewed as the "lid" that opens and closes upon substrate entry and binding. These van der Waals forces between alanine-105 and leucine-209 might maintain the close interaction between the F-G loop and the steroid, which in turn pushes the steroid "down" so that C-16 of progesterone is exposed to the heme moiety. When leucine is substituted for alanine-105, the increased steric clash between leucine-105 and leucine-209 brings the F-G loop to a more "open" position, which in turn raises the steroid and C-16, while C-17 and C-21 swing closer to the heme center (Fig. 7). If this analysis is correct, double mutations of residues 105 and 209 might be used to validate the importance of this probable interaction. Other possible amino acid residues on the F-G loop that might directly interact with the side chain of leucine-105 are isoleucine-205 on the F-helix and leucine-232 on the G-helix, which are 5.2 and 5.4 Å away from alanine-105, respectively.</p><!><p>The use of selectively deuterated steroids not only allowed for calculation of KIEs for CYP17A1-catalyzed hydroxylations at these labeled positions but also confirmed a minor 21-hydroxylation activity with progesterone. When [17-2H]- or [16α-2H]-progesterone substrates (Fig. 3, 20 or 19) were incubated with CYP17A1, the 21-hydroxylation product (DOC) increased relative to 17OHP and 16OHP (Fig. 6). Conversely, when the [21,21,21-2H3]-progesterone (Fig. 3C, 21) was used as the substrate, the 21- hydroxylation product decreased relative to 17OHP and 16OHP. In other words, increasing the activation energy for one of the major products via deuterium labeling at that C-H bond favored formation of the minor product DOC, and vice versa. This 21-hydroxylase activity of CYP17A1 was further verified with mutation A105L [20], which has enhanced 21- hydroxylase activity and decreased 16α-hydroxylase activity (Fig. 8).</p><!><p>Experiments with CYP17A1 and [16α-2H]-progesterone (Fig. 3A, 19) showed that the 16β-hydrogen atom is also accessible for abstraction (Fig. 8, pathway 3d). If the 16α-hydrogen of this substrate is always abstracted, the 16OHP product of these incubations should lose all deuterium enrichment. In contrast, mass spectrometry demonstrated retention of the deuterium in 33–40% of the 16OHP product (Fig. 8, 8B) [20]. This type of switching at the same methylene site does not always occur for P450 enzymes. In the case of CYP3A4-catalyzed testosterone 6β-hydroxylation, incubations with [6β-2H]-testosterone substrate afforded only unenriched 6β-hydroxytestosterone, and no [6α-2H]-6β-hydroxytestosterone was detected [32].</p><p>Thus, human CYP17A1 hydroxylates the 16α-, 17-, and 21-positions of progesterone and also abstracts the 16β-hydrogen atom. In contrast, we could only demonstrate 17-hydroxylation of pregnenolone when [17-2H]-pregnenolone was used as the substrate. This strict regioselectivity might be due to the hydrogen-bonding located between the 3β-hydroxy group of the steroid ring and asparagine-202, as observed in the CYP17A1 crystal structure [6].</p><!><p>CYP17A1 inhibitors are used to treat androgen dependent cancers, because the 17,20-lyase activity is required for all androgen biosynthesis. Although all 17-hydroxysteroid intermediates are substrates for the 17,20-lyase reaction, the preferred substrates are 17Preg and 17-hydroxyallopregnanolone (5α-pregnan-3α,17-diol-20-one) over 17OHP and its 5α-reduced homolog. The 17,20-lyase reaction, as well as other P450-catalyzed C-C bond cleavage reactions, has been studied intensively, yet the mechanism of this reaction remains controversial. Akhtar and colleagues proposed a mechanism equivalent to an enzymatic Baeyer-Villiger reaction using a ferric peroxide intermediate as a nucleophile onto the C20-carbonyl of 17Preg (Figure 9A) [33, 34]. Evidence in favor of this mechanism derived from studies with the incubation of [16α,17,21,21,21-2H5]-pregnenolone with pig microsomes, where an 18O atom from molecular oxygen (18O2) was incorporated into the acetic acid product, which was derivatized as benzyl acetate and detected by gas-chromatography mass spectrometry [33]. The fact that an 18O-atom was incorporated into the acyl cleavage product, however, does not rule out a mechanism that includes compound I (Fig. 9B). In these studies, the desaturase products were also observed (Fig. 10A).</p><p>These data were, however, derived from low-resolution mass spectrometry, with resolution (ΔM/M) of 1,000 full-width at half maximum (FWHM) to analyze the cleaved products. These experiments were reproduced with the use of purified human enzyme [35] and a 17-carboxaldehyde analog of pregnenolone; however, low-resolution mass spectrometry was again used to analyze the incubation products [34, 35]. Particularly for the experiments using a mixture of deuterated 17-carboxaldehydes, low-resolution mass spectrometry cannot reliably distinguish endogenous formic acid from the non-18O-incorporated formic acid product of the enzyme reaction: the [13C]-benzyl formate isotopomer (m/z 137.0552, background source) from the [2H]-benzyl formate (m/z 137.0582, incubation source) due to the small mass difference (17 ppm) [35]. Moreover, if there is any other compound with a similar mass that co-elutes with the target mass range (i.e. within 50 ppm), then low-resolution mass spectrometry would not be able to distinguish the impurity from the target mass. Consequently, product percentages derived from these studies should be interpreted with caution. High-resolution mass spectrometry, which is available today, can provide a resolving power of >100,000 FWHM [36, 37] and reliably distinguish these compounds. Alternatively, fragmentation data of the isolated benzyl formate masses, using tandem mass spectrometry, can verify the identities of the observed molecular ions. On the other hand, retention of deuterium and incorporation of 18O in the benzyl acetate derived from the natural substrate pregnenolone was nicely shown in spectra from the published work [33].</p><p>In contrast, hydrogen peroxide does not support enzyme-mediated conversion of 17Preg to DHEA, which would exclude a ferric-peroxide mediated mechanism if the ferric peroxide is formed under the experimental conditions [38]. Radical-based mechanisms remain plausible as well. Currently, no definitive evidence is available to refute either proposal.</p><p>When the enzyme was first purified from porcine testis, the 17,20-lyase activity was lost disproportionate to the 17-hydroxylase activity [39]. Addition of b5 to the reconstituted assay restored most of the 17,20-lyase activity [34, 40]. Recent descriptions of undervirilized males with low testosterone production from b5 deficiency have confirmed the critical importance of b5 for enhancing the 17,20-lyase activity of CYP17A1 [41, 42]. The action of b5 appears to be an allosteric function rather than electron transfer [43, 44] and involves CYP17A1 residues R347 and R358 [16] as well as b5 residues E48 and E49 [45, 46].</p><p>In the presence of b5, porcine CYP17A1 catalyzes two other C-C bond cleavage reactions with pregnenolone (not 17Preg) substrate. The first is andien-β-synthase activity, yielding androsta-5,16-dien-3β-ol (Fig. 10, 5 from 1) [35], which is a precursor to an active pheromone in boar taint. The second trace activity is formation of androsta-5-en-3β,17α-diol (Fig. 11, 4), which was first observed from crude microsomal enzyme sources [33, 47–50], and later confirmed with purified recombinant human CYP17A1 from both pregnenolone and its 17-carboxaldehyde analog in the presence of b5 [35].</p><p>Moreover, in the Swinney study, 17-O-acetyltestosterone (Fig. 1, 11) product was observed upon incubation of purified CYP17A1 from neonatal pig testes with progesterone [51]. A Baeyer-Villiger mechanism was proposed for the formation of this product. The assignment of 17-O-acetyltestosterone as product (17β-hydroxy, 11) was based on its comigration with authentic standard by HPLC and GC-MS. Curiously, this reaction yields the opposite stereochemistry (17α-hydroxy) when the pregnenolone substrate is used. Furthermore, the acetyl group appears to be retained in the C-C bond cleavage product with progesterone substrate [51] but is lost as acetic acid [35] for pregnenolone substrate. The reasons for the differences in stereochemistry at C17 in the 17-hydroxy androgen products and the fates of the acetate moieties derived from C20–C21 in these reactions are not clear. One possible explanation is the existence of a 17-radical intermediate in the case of the pregnenolone substrate (Figure 11, 1-r), which undergoes inversion of configuration and oxygen rebound on the α-face of the steroid, yielding the 17α-hydroxyandrogen (4). While the 17α-hydroxyandrogens are not androgen receptor agonists, these alternative 19-carbon steroid compounds might serve as biomarkers of health and disease.</p><p>The microsomal b5 (type A form) stimulates both andien-β-synthase and 17,20-lyase activities [52], while a second form of b5, the type B form located on the outer mitochondrial membrane, was shown to only stimulate the 17,20-lyase activity to form DHEA. The presence of cytochrome b5 reductase promoted the andien-β-synthase activity of both porcine and bovine CYP17A1 [52, 53]. Human CYP17A1 catalyzes the same andien-β-synthase activity with b5 using either pregnenolone (1) or allopregnanolone (15) as substrate [14], but the functions of these products in human physiology, if any, remain obscure.</p><!><p>Evidence derived from experiments using site-directed mutagenesis, substrate analogs, structural biology, and isotopic labeling have begun to yield the biophysical principles that yield the catalytic promiscuity of this versatile and physiologically critical enzyme. Human CYP17A1 is known to catalyze at least 13 different reactions with endogenous substrate, and even some minor activities are physiologically important. The physiologic functions of human CYP17A1 in the adrenal and gonads demands a high degree of promiscuity, to both synthesize cortisol via the Δ4-pathway and to synthesize sex steroids via the Δ5-pathway. The activities in the backdoor pathway and other minor reactions add to the intrigue, leading to androgen and mineralocorticoid synthesis in disease states, pheromone production, and salvage pathways.</p><p>The study of additional mutations to alter the enzyme activity of CYP17A1 holds the potential for additional discovery. For example, if the alanine-105 interaction with leucine-209 is crucial for regiochemistry, then substitution of other bulky amino acids at position 105 or smaller residues at 209 might increase progesterone 21-hydroxylase activity further. If small molecules could be found to allosterically force the enzyme to favor these conformations leading to DOC formation, then this strategy could coopt a minor activity to treat the mineralocorticoid deficiency of 21-hydroxylase deficiency. Additionally, since P450 enzymes are found in all kingdoms of life and are studied in other research areas (i.e. biotechnology, drug metabolism and pharmacokinetics, reaction development, etc.), knowledge in the reactivity of CYP17A1 can benefit other fields of research. In conclusion, the KIE and related studies with CYP17A1 have revealed that this enzyme is not meant to have strict fidelity. The promiscuity of biological catalysts as illustrated with CYP17A1 depicts the complexity of the big picture and suggests that there may be hidden features that still need to be discovered, even for systems that might seem to be fully characterized.</p>

PubMed Author Manuscript

Nickel Catalyzed Cross-Coupling of Vinyl-Dioxanones to Form Enantiomerically Enriched Cyclopropanes

Under the conditions of nickel(0) catalysis, enantiomerically enriched vinyl-dioxanones engage boroxines or B2(pin)2 in stereospecific cross-coupling to form diverse tetrasubstituted cyclopropanes bearing all-carbon quaternary stereocenters. The collective data corroborate a mechanism involving nickel(0)-mediated benzylic oxidative addition with inversion of stereochemistry followed by reversible olefin insertion to form a (cyclopropylcarbinyl)nickel complex, which upon reductive elimination releases the cyclopropane.

nickel_catalyzed_cross-coupling_of_vinyl-dioxanones_to_form_enantiomerically_enriched_cyclopropanes

<p>Cyclopropanes appear as substructures across diverse secondary metabolites1 and are frequently found in commercial medicines, agrochemicals and fragrances.2 Hence, the development of methods for cyclopropane formation represents a persistent challenge in chemical research.3 Among the most effective methods for the preparation of enantiomerically enriched cyclopropanes is the reaction of olefins with metal carbenoids.3 Here, we report a strategy for the asymmetric synthesis of cyclopropanes under the conditions of metal catalyzed cross-coupling. Specifically, nickel(0) catalysts4,5 react with enantiomerically enriched 4-aryl-5-vinyl-1,3-dioxanones to form (cyclopropylcarbinyl)nickel(II) species, which, in the presence of organoboron reagents or B2(pin)2deliver cyclopropanes in a stereospecific manner. Thus, the enantioselective synthesis of tetra-substituted cyclopropanes bearing all-carbon quaternary stereocenters is achieved (Scheme 1).</p><p>In connection with ongoing investigations into the formation of C-C bonds via hydrogenation and transfer hydrogenation,6 we recently reported an iridium catalyzed coupling of primary alcohols with isoprene oxide to form products of tert-(hydroxy)-prenylation – a byproduct-free transformation that occurs with exceptional control of anti-diastereo- and enantioselectivity.7 It was posited that cyclic carbonates derived from these reaction products should be predisposed toward cyclopropane formation under cross-coupling conditions, as geminal substitution of the neopentyl glycol precludes competing β-hydride elimination of the σ-benzylmetal intermediate and should conformationally bias the system toward olefin insertion8 via Thorpe-Ingold effect.9 However, the facility of conventional benzylic cross-coupling rendered the feasibility of the proposed cyclopropane formation uncertain.10</p><p>In an initial experiment, vinyl-dioxanone 1a was exposed to the catalyst derived from Ni(cod)2 (10 mol%) and PCy3 (20 mol%) in the presence of tri(p-tolyl)boroxine 2a and K3PO4 (200 mol%) in toluene (0.1 M) at 60 °C. To our delight, cyclopropane 3a was formed in 36% yield as a single diastereomer. Conversion was found to be sensitive to concentration and temperature. At 45 °C under otherwise identical conditions, a 53% yield of cyclopropane 3a was obtained. Using the nickel catalyst modified by PCy2Ph (20 mol%), cyclopropane 3a was obtained in 77% yield. Finally, at slightly higher concentration (toluene, 0.2 M), an 85% yield of cyclopropane 3a was achieved (eq. 1). Stereospecificity was corroborated by chiral stationary phase HPLC analysis of cyclopropane 3a. Relative stereochemistry of cyclopropane 3a was confirmed by single crystal X-ray diffraction analysis. p-Tolylboronic acid also delivers cyclopropane 3a (eq. 1), but in slightly lower yield. Application of these optimal conditions to unsubstituted methyl carbonate model-1a did not result in cyclopropane formation; rather, the indicated product obtained through β-hydride elimination of the σ-benzyl intermediate was formed (eq. 2). Cyclic carbonate 1a reacted more efficiently than related acyclic carbonates, suggesting the internal alkoxide generated upon ionization-decarboxylation facilitates group transfer from boron to nickel through an internal boron ate-complex. (eq. 1) (eq. 2)</p><p>Optimal conditions utilizing tri(p-tolyl)boroxine 2a were applied to a structurally diverse set of enantiomerically enriched vinyl-dioxanones 1a–1i (Table 1). Vinyl dioxanones bearing a variety of substituted aromatic (1a–1d) and heteroaromatic (1e–1i) rings were converted to cyclopropanes 3a–3i in good yield with complete levels of diastereoselectivity. Relative stereochemistry was assigned in analogy to that determined for 3a. Although the preexisting non-epimerizable quaternary stereocenter serves as an "internal standard," stereospecificity was spot-checked for compounds 3a, 3b, 3d and 3h. Notably, unlike prior work involving nickel catalyzed benzylic substitution, extended aromatic systems are not required.11 Standard conditions also were applied to the coupling of vinyl-dioxanones 1a and 1h with boroxines 2b–2d, which incorporate p-CF3-phenyl, p-methoxyphenyl and (E)-styryl moieties, respectively (Table 2). The resulting cyclopropanes 3j–3o were formed in good yield in a completely stereoselective fashion. The coupling of vinyl-dioxanones 1a, 1b, 1d, 1h and 1f with B2(pin)2 under standard conditions delivers the cyclopropylcarbinyl boronates 3p–3t in good yield with complete stereocontrol (Table 3).12 To briefly illustrate the utility of coupling products, the neopentyl alcohol 3a was subjected to Jones oxidation to provide the cyclopropyl carboxylic acid 4a in good yield (eq. 3). Additionally, the cyclopropylcarbinyl alcohol 3h was exposed to Mitsunobu conditions in the presence of phthalimide to furnish 4b in excellent yield (eq. 4). (eq. 3) (eq. 4)</p><p>A general mechanism for stereospecific cyclopropane formation under the conditions of nickel catalyzed cross-coupling has been proposed (Scheme 2). Stereospecific oxidative addition of a nickel(0) species to the benzylic C-O bond occurs with inversion to furnish the indicated σ-benzylnickel(II) complex.10 Decarboxylation and transmetalation delivers the indicated alkene complex, which upon reversible migratory insertion8 provides a (cyclopropylcarbinyl)nickel(II) complex. Regardless of the kinetic diastereoselectivity of olefin insertion, reductive elimination occurs exclusively from a single stereoisomer of the (cyclopropylcarbinyl)nickel(II) species to release the cyclopropane and regenerate the zero-valent nickel catalyst.</p><p>In summary, we report a new method for the preparation of enantiomerically enriched cyclopropanes via stereospecific nickel catalyzed cross-coupling of vinyl-dioxanones with boroxines or B2(pin)2. The collective data are consistent with a catalytic mechanism involving nickel(0)-mediated benzylic oxidative addition with inversion of stereochemistry followed by reversible olefin insertion to form a (cyclopropyl-carbinyl)nickel complex, which upon reductive elimination delivers the cyclopropane. The novel reactivity embodied by this process should serve as the basis for the syntheses of diverse enantiomerically enriched cyclopropanes.</p><p>Supporting Information Available: Experimental procedures and spectral data. HPLC traces corresponding to racemic and enantiomerically enriched samples. Single crystal X-ray diffraction data for 3a. 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

Diastereoselective synthesis of 3,4-dihydro-2<i>H</i>-pyran-4-carboxamides through an unusual regiospecific quasi-hydrolysis of a cyano group

An efficient diastereoselective approach for the synthesis of functionalized 3,4-dihydro-2H-pyran-4-carboxamides with variable frame was developed based on the reaction of available 4-oxoalkane-1,1,2,2-tetracarbonitriles (adducts of TCNE and ketones) with aldehydes in an acidic media. An unusual process of quasi hydrolysis of the cyano group was observed in the course of the described regio-and diastereoselective transformation.2093

diastereoselective_synthesis_of_3,4-dihydro-2<i>h</i>-pyran-4-carboxamides_through_an_unusual_regios

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

<p>Dihydro-and tetrahydropyran moieties are very important structural fragments in organic synthesis. They are part of many natural compounds, biologically active substances and drugs [1][2][3][4]. For instance, Zanamivir and Laninamivir which are recommended for the treatment and prophylaxis of influenza caused by influenza A and B viruses contain a 3,4-dihydro-2Hpyran fragment [5]. Moreover, dihydropyrans have been proven to be particularly useful in the preparation of cyclic components of macrocyclic antibiotics [6,7] and as precursors in the synthesis of C-glycosides [8].</p><p>Modern and convenient methods for the construction of 3,4dihydro-2H-pyrans are based on the interaction between ylidene derivatives of methylene-active compounds and β-oxo derivatives of acids or ketones [9][10][11], and also on the reactions of phenols with carbonyl compounds [12]. An interesting example that had been recently described is the involvement of diazolactones in an inverse electron-demand Diels-Alder reaction [13]. At the same time the synthesis of 3,4-dihydro-2H-pyrans with a carboxamide group is a not sufficiently explored area. There is only one way to produce 3,4-dihydro-2H-pyran-4-carbox-Scheme 2: Known approach to pyran derivatives based on ketonitriles 1.</p><p>amides from substrates containing no pyran ring described in the literature (Scheme 1) [14].</p><p>Scheme 1: An exclusive approach to 3,4-dihydro-2H-pyran-4-carboxamides from non-pyran sources.</p><p>In view of a directed synthesis of inaccessible heterocyclic molecules, 4-oxoalkane-1,1,2,2-tetracarbonitriles 1 are very promising substrates. They can be easily prepared from an appropriate ketone and TCNE (Scheme 2). Previously we reported about several ways of heterocyclization thereof [15][16][17][18][19][20][21][22][23], including the formation of pyran derivatives [20][21][22][23][24].</p><p>Our recent work should be noted individually, we were able to prepare various functionalized pyrano [3,4-c]pyrrole derivatives via diastereoselective cascade reaction [23]. The crucial stage of the described transformation is the formation of a pyran-4carboxamide intermediate. A trace amount of it was isolated accidentally and we could not repeat this procedure and characterize the compound by spectra.</p><!><p>In continuation of our interest in this area, we focused our attention on the extention of the existing methods for the synthesis of a series of pyran heterocycles. Therefore, we have studied more thoroughly the transformation of ketonitriles 1 in acidic media in the presences of aldehydes. We had found that 3,3,4tricyano-3,4-dihydro-2H-pyran-4-carboxamides 2 nevertheless could be obtained in good yields (57-69%) by the action of hydrochloric acid. A prominent feature of the reaction that has been developed is the possibility to vary substituents at the 2-, 5-and 6-positions of the pyran cycle, using alkyl and aryl moieties to design the rare pyran-4-carboxamide molecules. Moreover, during the regio-and diastereoselective transformation a quasi-hydrolysis of only one of the cyano groups had occurred. As a result only one diastereomer of 3,4-dihydro-2Hpyran-4-carboxamides 2 was obtained (Table 1).</p><p>The structure of pyrans 2 as well as the trans configuration of the substituents at the asymmetric atoms C2 and C4 were established by X-ray diffraction (Figure 1 a Reaction conditions: 3h (0.5 mmol), acid (0.75 mL), solvent (0.75 mL, if marked); b Isolated yield; c The reaction does not stop at the amide formation, further transformations proceeds. d The results were published previously [23].</p><p>reaction was separately described by us previously and it is the crucial reason for the diastereoselectivity of the whole transformation [21,22]. Further acid-catalyzed addition of water to the imino (B) is probably accompanied with decyclization (C) and dehydratation processes forming compound 2 (Scheme 3, path I). However, according to the literature, in most cases, the addition of water to the similar cyclic iminoethers leads to hydrolysis and formation of lactones [26,27]. Another probable pathway involves the protonation of the imino group nitrogen atom to form an iminium salt (D) and subsequent decyclization of iminolactone ring (E) without participation of water (Scheme 3, path II).</p><p>To establish the actual mechanism and to prove the intermediate formation of bicyclo[3.2.1]octane derivatives 3 we carried out the reaction of specially prepared 6-imino-2,7-dioxabicyclo[3.2.1]octane-3,3,4-tricarbonitrile 3h [21] with dry trifluoroacetic acid under anhydrous conditions (Table 2, entry 6).</p><p>The successful implementation of this process and isolation of product 2h with good yield indicate the path of transformations without the participation of water (Scheme 3, path II) as actual.</p><p>Such an abnormal resistance of the imine moiety to be hydrolyzed in acidic media is very exciting. Therefore we attempted to use other acids to investigate the behavior of the imine moiety in presence thereof, but all acids promote only the iminolactone decyclization process without traces of lactone derivative 4 (Table 2).</p><p>It is also important to note that during the reaction pathway (Scheme 3) a carbonyl-assisted carbonitrile hydration effect (CACHE) was occured. The carbonyl group, through the forma-tion of hemiketal (A), became a source of a hydroxy group that cyclized regiospecifically to the spatially proximate cyano group and caused the carboxamide formation. CACHE processes are essential for the chemistry of oxonitriles and sometimes can be the reason for unusual quasi-hydrolysis of the cyano group under mild conditions [15,19,23], but it was not mentioned by many authors [28][29][30][31].</p><!><p>In conclusion, we have developed a new efficient approach to a rare group of heterocycles, namely, functionalized 3,4-dihydro-2H-pyran-4-carboxamides with variable frame and exceptional diastereoselectivity based on the reaction of available 4-oxoalkane-1,1,2,2-tetracarbonitriles (adducts of TCNE and ketones) with aldehydes in acidic media. Unusual processes of regiospecific quasi-hydrolysis of a cyano group and abnormal resistance of the imine moiety to be hydrolyzed in aqueous acidic media were observed in the course of the described transformation. Moreover, the intermediate 6-imino-2,7-dioxabicyclo[3.2.1]octane-4,4,5-tricarbonitriles 3 had demonstrated a cytotoxic activity in various cancer cell lines [32], therefore the derived 3,4-dihydro-2H-pyrans 2 are very promising for biological studies.</p><!><p>Supporting Information File 1</p><p>Experimental data and characterization of all new compounds.</p><p>[http://www.beilstein-journals.org/bjoc/content/ supplementary/1860-5397-12-198-S1.pdf]</p>

Beilstein

Elucidating the Binding Mode between Heparin and Inflammatory Cytokines by Molecular Modeling

Heparan sulfate (HS) interacts with a broad spectrum of inflammatory cytokines, thereby modulating their biological activities. It is believed that there is a structural-functional correlation between each protein and sugar sequences in the HS polysaccharides, however, the information in this regard is limited. In this study, we compared the binding of four inflammatory cytokines (CCL8, IL-1beta, IL-2 and IL-6) to immobilized heparin by an SPR analysis. To define the molecular base of the binding, we used a heparin pentasaccharide as representative structure to dock into the 3D-molecular structure of the cytokines. The results show a discrepancy in K D values obtained by SPR analysis and theoretical calculation, pointing to the importance to apply more than one method when describing affinity between proteins and HS. By cluster analysis of the complex formed between the pentasaccharide and cytokines, we have identified several groups in heparin forming strong hydrogen bonds with all four cytokines, which is a significant finding. This molecular and conformational information should be valuable for rational design of HS/ heparin-mimetics to interfere cytokine-HS interactions.

elucidating_the_binding_mode_between_heparin_and_inflammatory_cytokines_by_molecular_modeling

Introduction<!>Surface Plasmon Resonance (SPR) Analysis of the Cytokines Binding to Immobilized Heparin<!>Analysis of Heparin-Cytokine Interactions by Molecular Modeling<!>Structural Dependence of IL-6 Interaction with Sulfated Polysaccharides<!>Common Features of Heparin in Interaction with Cytokines<!>Conclusion<!>Reagents<!>Preparation of Biotinylated Heparin and Immobilization<!>Immobilization of IL-6 onto the Chips<!>Surface Plasmon Resonance (SPR) Analysis<!>Molecular Modeling of Heparin-Protein Interaction

<p>Heparan sulfate denotes a group of negatively charged polysaccharides ubiquitously expressed on the cell surface and the extracellular matrix (ECM), [1] regulating a broad spectrum of biological and pathophysiological activities. The functions of HS are transmitted, mainly, through interaction with proteins, for example, cytokines and chemokines. [2] The multi-functional properties of HS owe to its structural diversity, such as distinctive molecular structures which are expressed on the same cell surface, binding to different protein ligands for distinct functions. The complexity and variety of HS structures are generated through a complex biosynthesis process dictated by a remarkable regulatory machinery, enabling a stringent structure-function correlation for a given circumstance. [3] Inflammatory cytokines are signaling molecules secreted predominantly from immune cells, for example, T helper cells and macrophages, promoting inflammatory reactions. The cytokines act on the target cells through binding to receptors on the cell surface. An array of inflammatory cytokines can be activated by HS, [4] such as microphage inflammatory protein (MIP-1α), monocyte chemoattractant protein 1 (MCP1/CCL2), [5] monocyte chemoattractant protein 2 (MCP2/CCL8) [6] and several interleukins. [7] On the target cell surface, HS binds to the cytokines, serving as a co-receptor. In the extracellular matrix, HS functions as a storage for secreted cytokines and modulates their activities. A HS polysaccharide chain is generally composed of about 100 sugar units that are variably sulfated and may harbor several binding sequences for different cytokines. Thus, to illustrate the molecular interaction mode between a given protein and HS, it is of importance to understand the functional implications. [8] Heparin is an anticoagulant drug widely used for prevention and treatment of thrombosis. Clinical applications observed, apart from its anticoagulation activity, beneficial 'side effects' of heparin, one of which is attenuation of inflammation. Indeed, our retrospective study found that application of low molecular weight heparin (LMWH) significantly reduced the IL-6 level in the plasma of severely ill COVID-19 patients. [9] These functions of heparin are most likely due to interferences arising from the interactions between HS and inflammatory cytokines, because heparin and HS polysaccharides are synthesized by the same process and share high structural similarity. Based on the structural similarity, heparin is generally capable of binding to all HS-binding proteins, disrupting their functions. Such effects of heparin are of potential value for expanded applications. To avoid the potential risk of bleeding, heparin has been chemically modified to abolish its anticoagulation activity, resulting in non-anticoagulant heparin. [10] An alternative approach is to produce heparin/HS-mimetics. A few species of such heparin/ HS-mimetics have been explored for anticancer activity. [11] To develop heparin/HS-mimetics for interfering with the HSprotein interaction, one challenge is to elucidate the HS sequences specifically binding to the target proteins. In this work, we aimed at defining the structure-function correlation of heparin (as well as heparin/HS-mimetics) with four inflammatory cytokines, CCL8, IL-1beta, IL-2 and IL-6. Surface Plasmon Resonance (SPR) analysis showed a differential affinity of the cytokines in binding to immobilized heparin. The interaction mode was analyzed by molecular modeling. The results demonstrate that the combination of biochemical and modeling analysis could contribute to rational design of heparin/HSmimetics for the purpose of modulating the activities of cytokines as well as other HS-binding proteins.</p><!><p>To confirm the immobilization of the biotinylated heparin onto the SA gold sensor chip surface, we utilized FGF2 as a well characterized heparin-binding protein to test the binding capacity. The serial diluted FGF2 samples (from high to low concentration) were injected into the channel at a constant flow of 30 μL min À 1 . The changes in refractive index caused by molecular interactions at the sensor surface were monitored and recorded as response unit. Calculation of the dissociation constant showed a K D value of 5.79 nM (Figure S2, Supporting Information), which is in good agreement with previous report, [12] indicating a good reactivity of the immobilized heparin. Using the same setting, IL-1beta, IL-2, IL-6 and CCL8 were sequentially analyzed. The chip was washed with HBS-P for 360 s and regenerated with 50 mm NaOH for 10 s between each run. The sensorgrams of the analysis are shown in Figure 1. The collected data are calculated by the software (GraphPad, La Jolla, CA) the SPR instrument is equipped with. Comparison of the dissociation constants revealed a dramatic difference between the cytokines in binding to heparin. IL-1beta did not bind to the immobilized heparin, while CCL8 displayed an affinity as high as FGF2.</p><!><p>To explore the mechanistic basis for the differential affinity between the cytokines and heparin, we analyzed the interactions by molecular modeling using the AutoDock VINA and molecular dynamics (MD) in the YASARA program [13] and APBS (Adaptive Poisson-Boltzmann Solver) Electrostatics [14] Plugin in PyMOL (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC). The structural information of the cytokines was obtained from the PDB as described in the Experimental Section. To reduce the complexity of the macromolecular heparin, we used a pentasaccharide sequence (Figure 2) that was docked into the molecular structures of the four cytokines, respectively. The results show formation of an interaction complex between each of the cytokines with the pentasaccharide (Figure 3). Calculation of binding free energy and predicted K D values (Table 1) show a higher affinity between CCL8 and the pentasaccharide, which is in agreement with the result obtained by SPR. This is likely due to the basic amino acids at the surface (blue), which provide a strong positive electrostatic environment, forming salt bridge interactions with the negatively charged groups in heparin. In contrast, the pentasaccharide was docked into a patch of acidic amino acids in IL-2, indicating that, while salt-bridge interactions do not contribute to the interaction between IL-2 and the pentasaccharide, weaker hydrogen-bond interactions do.</p><p>To elucidate the binding modes, we examined the docking complexes through a 2D diagram (Figure 4). The amino acid residues (Thr10, Ile9, Asn33, Thr32, Arg30, Ile31, Tyr28, Asp68) on the surface of CCL8 (Figure 4D and Table S4) interact with both N-and 6-O-sulfo groups of GlcNS6S, 2-O-sulfo, as well as the hydroxy groups of IdoA2S, GlcA and GlcNAc by hydrogen bonds. In comparison, the negatively charged amino acid residues (Asp142 and Asp145) of IL-1beta form hydrogen bonds with hydroxy groups of GlcNAc (Figure 4B), while the negatively charged amino acids of Glu55, Glu56, Glu57 and Glu63 on IL-2 enabled formation of hydrogen bonds with N-and 6-O-sulfo groups of GlcNS6S and hydroxy groups of IdoA2S and GlcNAc (Figure 4A). Notably, the negatively charged groups on amino acid residues of Ser13 (IL-1beta), Glu55, Glu56 and Glu57 (IL-2) and Glu93 (IL-6) form negative-negative ionic interactions with hydroxy groups of GlcA, 6-O-sulfo groups of GlcNS6S, which may inactivate the anionic attractivity of the sulfo-and carboxyl groups in the pentasaccharide.</p><!><p>To further probe the selectivity of cytokine binding to polysaccharides, we applied a different approach by immobilization of a protein (IL-6) on the Series S Sensor Chip CM5 to which binding of the four polysaccharides were analyzed. Serial dilutions of heparin at the concentration of 500-15.625 μm, OD-heparin (O-desulfated heparin) at the concentration of 100-12.5 μm, NS-K5 (N-sulfated K5 polysaccharide) at the concentration of 20-1.25 μm and HS (heparan sulfate) at the concentration of 200-6.25 μm were injected at a constant flow of 30 μL min À 1 . SPR analysis resulted in a low RU recording for all samples, indicating an overall low binding of the polysaccharides to the immobilized IL-6. Nevertheless, calculation of K D values revealed differences in binding affinity of the polysaccharides. OD-heparin and NS-K5 had a slightly higher affinity than heparin in binding to IL-6, while the native HS displayed the lowest affinity to IL-6 (Figure S3).</p><p>To find out whether there is indeed a difference in binding of the polysaccharides to IL-6, we analyzed the interaction mode by molecular modeling using a pentasaccharide structure of each polysaccharide (Figure 5) including heparin (Figure 2). Docking of the oligosaccharides into the IL-6 3D molecular structure resulted in complex formation (Figure 6), indicating binding of each pentasaccharide to IL-6. Molecular docking 2).</p><p>Overlay of the complexes shows that the pentasaccharides of heparin/NS-K5/HS (green/cyan/light blue) docked into the domain of B and D helixes in IL-6, while OD-heparin (pink) docked into the domain of B and C helixes in IL-6 (Figure 6, panel B), shifted by � 16 Å away from the site where the other three oligosaccharides bound. From an overhead view of the overlapping complexes (Figure 6, panel C), HS appeared having a loose reducing end (shown in red circle) on the outer side of electron density map of IL-6, having no interaction with IL-6, which may lead to its weaker interaction with IL-6. The 2D diagrams further illustrate the different mode of interactions between IL-6 and the pentasaccharides (Figure 7). The distinctive binding site of OD-heparin from the other three pentasaccharides may lie beside the middle domain of B and head domain of C helixes in IL-6, where the binding site might be the closest position of these pentasaccharides binding to IL-6.</p><!><p>Since all four cytokines showed interaction with the heparin pentasaccharide, we wanted to examine whether there are specific groups of heparin contributing to the interactions with the cytokines. By cluster analysis, we identified the key chemical groups on heparin that form strong hydrogen bonds with amino acid residues of all four proteins (Figure 8). To justify whether the models constructed from molecular docking are reliable, the molecular dynamics (MD) simulations of each cytokine with a pentasaccharide heparin bound were performed. The RMSD values of the structures aligned by PyMOL including the last structure of IL-1beta + Heparin/IL-2 + Heparin/IL-6 + Heparin/IL-6 + HS/CCL8 + Heparin/IL-6 + NS-K5/ IL-6 + OD-heparin obtained from the MD and the complex of IL-1beta + Heparin/IL-2 + Heparin/IL-6 + Heparin/IL-6 + HS/CCL8 + Heparin/IL-6 + NS-K5/IL-6 + OD-heparin obtained from molecular docking complexes are small (Figure S4), which means that the structural variance is minor and the structures of complexes obtained from molecular docking are reliable.</p><!><p>Heparan sulfate plays vital roles in animal development and homeostasis through modulating biological functions of a large number of growth factors and cytokines. [15] Biochemical analysis identified hundreds of proteins interacting with HS (and heparin due to structural similarity); [16] however, information on the binding mode between the proteins and HS/heparin is barely available. In this study, we examined the molecular interactions of heparin with four inflammatory cytokines (CCL8, IL-1beta, IL-2 and IL-6) that are known binding to HS by molecular docking [17] in combination with the conventional SPR analysis.</p><p>Comparing the K D values calculated from the molecular docking and SPR analysis finds a dramatic difference, indicating the two methods are not directly comparable for assessment of apparent affinity. This discrepancy between the two methods may be caused by several factors. The SPR used full-length heparin that is immobilized on a surface, while the molecular docking used a pentasaccharide that is freely movable in the system. Nevertheless, CCL8 displayed the highest affinity and IL-1beta had the lowest affinity by both methods. The dramatic discrepancy in the binding affinity between IL-6 and heparin obtained by immobilization of heparin (K D = 681.5 nm) and immobilization of IL-6 (K D = 696 μm) is of worth noting. Most of the reported biochemical analysis for protein-sugar binding (including heparin, heparin-mimetics and glycosaminoglycans) uses the method to immobilize sugars, including glycanarrays [18] and SPR analysis. [19] Since the immobilization may affect the reactivity and interaction mode, different methods, including molecular modelling, [20] should be applied when define binding affinity between a given protein and heparin/HS or glycosaminoglycans. A significant point revealed by molec- ular modeling is that the positively charged amino acids (Lys66, Lys70 and Lys86) on the surface of IL-6 formed hydrogen bonds primarily with the N-and 6-O-sulfo groups of GlcNS6S and 2-Osulfo group of IdoA2S in heparin, instead of salt bridges as observed for CCL8. This is due to the negative-negative ionic interactions between Glu93 in IL-6 and 6-O-sulfo group of GlcNS6S in the pentasaccharide, leading to losing or weakening of the binding between positively charged amino acids and the negative groups in heparin.</p><p>One unexpected finding by overlay of the IL-6 complexes is that the pentasaccharide of HS, the theoretical endogenous molecule, displayed the lowest affinity, while the pentasaccharide of OD-heparin showed the highest affinity, though it binds on a negative electrostatic environment of IL-6 surface (Figure S5). OD-heparin and NS-K5 carry the same number of Osulfo groups, but NS-K5 has two glucuronic acids (GlcA), while OD-heparin has two iduronic acids (IdoA). Most likely, the conformation of IdoAs renders the sugar fitting well with the shape complementarity of IL-6 on a different site from where the other three pentasaccharides bind. Regardless, none of the pentasaccharides binds to the receptor binding site on IL-6, [21] indicating a potential risk that these structures of heparin mimetics may not interfere with the interaction of IL-6 with its receptor (Figure 6; panel B).</p><p>A last important finding consists of the identification of active groups in heparin. By cluster analysis to compare the binding between the heparin pentasaccharide with the four cytokines, we localized the chemical groups in heparin that have a high frequency to interact with the proteins by forming hydrogen bonds. This information is valuable for rational design of HS/heparin-mimetics to interfere with the HS-protein interactions. As an implication for the chemical synthesis of nonsugar-based HS/heparin-mimetics, these reactive groups might be included in a given organic molecule.</p><p>In consideration of the important pathophysiological functions of HS in diverse diseases, for example, inflammationincluding COVID-19 suffering from cytokine storms [22] -and lupus nephritis, [23] exploring the potential of HS/heparin-mimetics for pharmaceutical applications is raising attention. To specifically target a HS-protein interaction, it is essential to know the molecular structures of the target protein and the binding sequence of HS. With the increasing availability of protein structures in the PDB database discovery and rational design of specific HS-sequence through molecular modeling became more convenient. The combination of virtual screening and biochemical analysis can provide powerful tools to discover novel pharmaceutical compounds.</p><!><p>The cytokines used in this study include IL-2 (PeproTech, USA), IL-1beta, IL-6 and CCL8 (Sino Biological, China or PeproTach, USA). Heparin (MW � 12 kDa) was obtained from SPH No.1 Biochemical Pharmaceutical CO., LTD, China. OD-heparin (O-desulfated heparin; MW � 12 kDa), NS-K5 (N-sulfated K5 polysaccharide; MW � 60 kDa) were prepared as described, [24] HS (heparan sulfate; MW � 30 kDa) was purified from bovine lung. [25]</p><!><p>For biotinylation of heparin at the reducing end, 10 mg heparin (MW � 15 kDa, SPH No.1 Biochemical Pharmaceutical CO., LTD, #C-HEPPIM) and aniline (11 μL) in NaOAc buffer (100 mm, pH 6.0, 1.08 mL) were mixed with 120 μL of EZ-Link Alkoxyamine-PEG4-Biotin in DMSO (50 mm, Thermo Scientific #26137) and incubated at 37 °C for 48 hr. The product was purified in a 2 mL-DEAE SEPHACEL (Cytiva) column, and concentrated by ultrafiltration with a 3 kD cutoff Millipore Amicon ultrafiltration tube, followed by desalting with PD-10. Then, the lyophilized biotinylated heparin was used for the following experiment.</p><p>The streptavidin gold sensor chip was plasma cleaned prior to immobilization. The biotinylated heparin (5 μL, 2 mg ml À 1 ) was dissolved in 200 μL of HBS-P (10 mm HEPES, 150 mm NaCl, 0.005 % (v/v) surfactant P20) and immobilized onto the sensor chip with the ligand in ~800 response unit and HBS-P was used as running buffer. A multi-channel Surface Plasmon Resonance (SPR) instrument (Biacore S200, GE Healthcare) was used for the analysis. Equilibration of the baseline was performed by a continuous flow of HBS-P buffer through the chip surface for 2-4 hr between application of each protein ligand. The data were collected at 25 °C with HBS-P as running buffer at a constant flow of 30 μL min À 1 .</p><!><p>The Series S Sensor Chip CM5 was prepared by mixing 400 mm EDC and 100 mm NHS (GE Healthcare) at a flow rate of 10 μL min À 1 immediately prior to immobilization of the protein. IL-6 was dissolved in 10 mm NaAc (pH 4.0) to the concentration of 20 μg mL À 1 and immobilized onto the activated sensor chip at a flow rate of 10 μL min À 1 . The ligand density was 1050 response unit. The chip was deactivated by 1 m ethanolamine hydrochloride-NaOH at a flow rate of 10 μL min À 1 for 420 s. A multi-channel Surface Plasmon Resonance (SPR) instrument (Biacore 8 K, GE Healthcare) was used for the analysis with 1 × PBS (2 mm KH 2 PO 4 , 10 mm Na 2 HPO 4 , 137 mm NaCl, 2.7 mm KCl, pH = 7.4) as running buffer at a constant flow of 30 μL min À 1 at 25 °C.</p><!><p>For the experiment of protein binding to immobilized heparin, the cytokines were dissolved in HBS-P buffer at the concentration of 10 mm and serially diluted with running buffer to the concentrations of 10.000, 3.333, 1.111, 0.370, 0.123, 0.041, 0.014 and 0.005 mm. The diluted samples were injected into the channel for 120 s, followed by washing of more than 360 s with the HBS-P running buffer. For the experiment of polysaccharides binding to immobilized protein, heparin, OD-heparin, NS-K5 and HS were dissolved in 1 × PBS buffer to 500 μm as start concentration that was serially diluted with running buffer to the concentrations indicated in Figure S3. The diluted samples were injected into the channel for 60 s, followed by washing more than 90 s with 1 × PBS running buffer. RU was recorded to determine the binding activity of analyte. The equilibrium dissociation constant (K D ), the association {k on } and dissociation {k off } rate constants were determined using Equations ( 1) and (2) as follows.</p><p>Here R represents the response unit and C is the concentration of the analytes.</p><!><p>The three-dimensional structures of IL-1beta (PDB ID: 1ITB), IL-2 (PDB ID: 6YE3), IL-6 (PDB ID: 1ALU) and CCL8 (PDB ID: 1ESR) were obtained from RCSB-PDB (www.rcsb.org ). For the interaction studies, the PDB files were applied with monomers retained and cleaned with the heteroatoms (HETATM) of the receptor and ligand removed by the online CHARMM-GUI program (http:// www.charmm-gui.org). [26] The PDB files of heparin and its mimetics were prepared by Chem3D 17.0. [27] The molecular dockings of the proteins and heparin were performed by AutoDock VINA in YASARA to evaluate the binding sites and binding strength of the ligands. In brief, the proteins were maintained rigid and heparin (as ligand) was fully flexible. To remove bumps and ascertain the covalent geometry of the ligand, a pentasaccharide heparin structure was energy-minimized with the NOVA force field, [28] using the Particle Mesh Ewald algorithm [29] to treat long-range electrostatic interactions. After removal of conformational stress by a short steepest descent minimization, the procedure was continued by simulated annealing (time step 2 fs, atom velocities scaled down by 0.9 every 10 th step) until convergence was reached. The blind dockings were undertaken in two steps, the former step was setting boxes of sizes 54.2 Å × 54.2 Å × 54.2 Å for IL-1beta, 52.16 Å × 52.16 Å × 52.16 Å for IL-2, 59.5 Å × 59.5 Å × 59.5 Å for IL-6 and 49.09 Å × 49.09 Å × 49.09 Å for CCL8. The latter step was docking of heparin onto each receptor protein, leading to 25 poses and 9 clusters for each situation. The binding affinity (dissociation constants, K D values) was predicted by the calculation of free binding energy in the docking experiments.</p><p>The PQR file of IL-1beta, IL-2, IL-6 and CCL8 were generated by the online PDB2PQR server (https://server.poissonboltzmann.org/ pdb2pqr) based on the molecular docking results of proteins and heparin. [30] The electrostatic potential maps of the IL-1beta, IL-2, IL-6 and CCL8 were visualized using PyMOL (version 2.3.4 by Schrödinger, LLC).</p><p>To justify whether the models constructed from molecular docking are reliable, the molecular dynamics (MD) simulations of each cytokines with a pentasaccharide heparin bound were performed. The AMBER14 force field was used for the MD simulation as implemented in the YASARA program. The MD simulation employed periodic boundary conditions, the particle-mesh Ewald method for the treatment of the long-range coulomb forces beyond a 8 Å cutoff. 0.9 % NaCl (a mass fraction) was used. The cell was rescaled such that residues named HOH reach a density of 0.997 g ml À 1 . No restraints were applied during the MD simulation using the settings employed in the second equilibration dynamics.</p><p>The energies and coordinates every 100 ps were saved with a total simulation length of 400 ns at constant temperature (298 K) and pressure uncontrolled in NVT ensemble. Structural stability of the receptor-ligand complex was examined by analyzing the average values of potential energy with root mean square deviation (RMSD) throughout the trajectory. The RMSD profiles of all MD structures (Figure S1) show that the variation of the RMSD values tends to be stable (< 1 Å) after 300 ns, which means that the equilibrium structures have been obtained and the last MD structures can be chosen as representative ones from the most populated cluster.</p>

Chemistry Open

Theoretical Assessment of Dinitrogen Fixation on Carbon Atom

Dinitrogen activation in non-metallic systems has received considerable attention in recent years. Herein, we report the theoretical feasibility of N 2 fixation using aminocarbenes (L) or their anionic derivatives. The molecular descriptors of L and anionic L‾, which affect the interaction of L and anionic L‾ with N 2 , were identified through multiple linear regression analysis. Additionally, the electron flow during C-N bond formation was confirmed by performing intrinsic reaction coordination calculations with intrinsic bond orbital analysis for the reaction of anionic L‾ with N 2 .

theoretical_assessment_of_dinitrogen_fixation_on_carbon_atom

<!>G (kcal/mol)

<p>Dinitrogen (N 2 ) is the most abundant and easily accessible nitrogen source in nature. The synthesis of useful chemicals using N 2 , from NH 3 [1] to N-containing organic molecules [2] , has a long history and is still considered most challenging in chemistry. The development of systems for metal-free N 2 fixation is particularly difficult. Recently, boron-containing organic materials have attracted attention as materials for metal-free N 2 fixation. The fixation and electrocatalytic reduction of N 2 using graphene doped with boron were reported by Zheng et al. [3] The Braunschweig group also reported the fixation and reduction of N 2 using organic borylene. [4] Zhu et al. reported the theoretical systematic design of frustrated Lewis pairs using a highly Lewis acidic borole as the active site, with N-heterocyclic carbenes (NHCs). [5] However, to date, there are no reported examples of N 2 fixation using only carbon active sites.</p><p>Persistent aminocarbenes, represented by NHCs, have been utilized for metal-free small-molecule activation. Persistent aminocarbenes successfully react with CO 2 , SO 2 , NO, and N 2 O as well as inert molecules such as H 2 and CO. [6] Furthermore, the metal-free catalytic conversion of CO [7] over an aminocarbene catalyst was also reported. Considering that CO is isoelectronic with N 2 , aminocarbene can be effectively utilized as the carbon active site for N 2 fixation.</p><p>Herein, we report theoretical studies on N 2 fixation on carbon atoms. We explored 20 typical examples of aminocarbene (L) [8] in five binding modes with N 2 (Figure 1). We confirmed that although thermodynamically favored products can be obtained from the reaction of L and N 2 , the reaction is not kinetically favored.</p><p>The thermodynamic and kinetic effects on the interaction between L and N 2 were investigated using the molecular descriptor L. We also reported that the anionic L‾ radical can overcome the high activation barrier of L in the reaction with N 2 . First, we attempted to establish possible candidates that could be obtained from the reaction of L with N 2 . With the proposal that the reactivity of NHCs can mimic the reactivity of transition metals, [6a] we proposed five structures of possible products based on previously reported transition metal-N 2 complexes (Figure 1). [9] The ligand hapticity of N 2 has been reported for both η 1 and η 2 for transition metal-N 2 complexes. Thus, L(η 1 -N 2 ) and L(η 2 -N 2 ) were chosen as mononuclear L(N 2 ) complexes. Furthermore, referring to the examples of the reported bimetallic N 2 complexes, [9] L 2 (μη 1 :η 1 -N 2 ), L 2 (μ-η 2 :η 2 -N 2 ), and L 2 (μ-η 1 :η 2 -N 2 ) were also selected as binuclear L 2 (N 2 ) complexes. L(η 1 -N 2 ) and L(η 2 -N 2 ) can be considered as diazoalkane and diazirine derivatives, which have been actively used as precursors of transient carbenes. L 2 (μη 1 :η 1 -N 2 ) can also be considered an azoalkane derivative. Each structure was optimized by Gaussian 16 using the B3LYP functional with the Def2-SVP basis set. As a result, the formation of L(η 1 -N 2 ), L(η 2 -N 2 ), and L 2 (μ-η 2 :η 2 -N 2 ) was thermodynamically unfavorable for all L (Figure 2a).</p><p>Unfortunately, the structure optimization of L 2 (μ-η 1 :η 2 -N 2 ) for some L failed; however, the formation of other successfully optimized L 2 (μ-η 1 :η 2 -N 2 ) was thermodynamically unfavorable, except for L13 and L14. The formation of L 2 (μ-η 1 :η 1 -N 2 ) was thermodynamically preferred for most L, except for L1 to L3 and L8. Therefore, we chose L 2 (μ-η 1 :η 1 -N 2 ) as a suitable model for N 2 fixation on carbon atoms. Despite the high thermodynamic stability of L 2 (μ-η 1 :η 1 -N 2 ), the real challenge for dinitrogen fixation on carbon atoms is that the formation of intermediate L(η 1 -N 2 ) is kinetically and thermodynamically unfavorable. We confirmed the high activation barrier for the formation of L(η 1 -N 2 ) (∆G ‡ (L(η 1 -N 2 )) = 27-55 kcal/mol) between L4 and L20 (Figure 2a). The activation barrier for the formation of L 2 (μ-η 1 :η 1 -N 2 ) from L(η 1 -N 2 ) was relatively low (∆G ‡ (L 2 (μ-η 1 :η 1 -N 2 )) = 10-27 kcal/mol). We identified the correlation of ∆G and ∆G ‡ with the normalized molecular descriptors for L through multiple linear regression (MLR) analysis. [10] Considering the small size of dinitrogen, steric descriptors, such as V bur and Sterimol parameters, were omitted from the selection of the molecular descriptor of L. We selected the molecular descriptor for MLR analysis using a correlation map to remove multicollinearity of the selected molecular descriptors (Figures S1 and S2).</p><p>Because of the strong correlation between ∆G(L(η 1 -N 2 )) and ∆G ‡ (L(η 1 -N 2 )) (R 2 = 0.93), similar models were produced by performing MLR on selected molecular descriptors with the robustness of the correlations (R 2 = 0.92 for Figure 2b and R 2 = 0.98 for Figure 2c). ∆G(L(η 1 -N 2 )) strongly depends on E HOMO and E LUMO and weakly on d C-N and φ ∠NCA . Similarly, ∆G ‡ (L(η 1 -N 2 )) depends on E HOMO , E LUMO , q NBO N , d C-N , and φ and weakly on d C-A . Commonly, ∆G(L(η 1 -N 2 )) and ∆G ‡ (L(η 1 -N 2 )) are more dependent on E LUMO than on E HOMO , indicating that the reactivity of L with N 2 was more influenced by the Lewis acidity of L.</p><p>For N 2 fixation using L without additional reagents, it is necessary to design an ambiphilic L with high HOMO and low LUMO energy levels. However, the low thermal stability [8e, 11] of previously reported ambiphilic aminocarbenes is not suitable for overcoming the high activation barrier (∆G ‡ (L(η 1 -N 2 )) > 27 kcal/mol) for the formation of L(η 1 -N 2 ). Instead of designing a new ambiphilic aminocarbene with high thermal stability, we proposed another strategy for N 2 fixation using previously reported L to achieve a low activation barrier for N 2 fixation. Low-valent transition metal complexes have received continuous interest as promising catalyst candidates for N 2 fixation and catalytic reduction. [12] The high electron density of the low-valence transition metal allows the formation of a stable N 2 complex through strong π-backdonation on the N 2 ligand. We proposed that the aminocarbene derivatives L, mimicking the reactivity of transition metals, can form stable L(η 1 -N 2 ) radical anions via one-electron reduction. The structures of anionic L‾, L(η 1 -N 2 )‾ and L 2 (μ-η 1 :η 1 -N 2 )‾ radicals and their transition states were optimized at the B3LYP/Def2-SVP level. The results indicated that the formation of anionic L(η 1 -N 2 )‾ and L 2 (μ-η 1 :η 1 -N 2 )‾ radicals is thermodynamically more favored than that of neutral L(η 1 -N 2 ) (Figure 3a). The formation of L(η 1 -N 2 ) radical anions was slightly exothermic for L10, L15, and L16 (−2.2, −9.1, and −0.5 kcal/mol). Furthermore, we confirmed that the activation barrier of the reaction between the anionic radical derivatives of L and N 2 decreased drastically. For example, the free energy of L9(η 1 -N 2 ) formation decreased from 33.6 kcal/mol to 4.7 kcal/mol, and its activation barrier also significantly decreased from 37.0 kcal/mol to 12.2 kcal/mol (Figure 4). The activation barrier of L9 2 (μ-η 1 :η 1 -N 2 ) formation was also decreased from 11.6 kcal/mol to 6.8 kcal/mol. However, the activation barrier of L(η 1 -N 2 ) formation for L6, L11, L12, and L14 radical anions, which are highly ambiphilic aminocarbenes, was not significantly decreased. (ΔΔG ‡ < 10 kcal/mol).</p><!><p>Figure 4. Free energy profile of the reaction of L9 and L9 -1 with N2 as calculated using Gaussian 16 at the B3LYP/Def2-SVP level.</p><p>We also performed MLR analysis to determine how the molecular descriptors of L affect the free energy and activation barrier of anionic L(η 1 -N 2 )‾ formation. We considered both the molecular descriptors of neutral and anionic L‾. In addition, the free energy and activation barrier of neutral L(η 1 -N 2 ) formation were considered as additional molecular descriptors. In the removal of multicollinearity using the correlation maps of selected parameters, the molecular descriptors of neutral L were selected prior to the molecular descriptors of anionic L‾ (Figure S1).</p><p>Interestingly, there is a relatively low correlation between the free energy of anionic L(η 1 -N 2 )‾ formation and its activation barrier (R 2 = 0.52), unlike neutral L(η 1 -N 2 ) formation (R 2 = 0.93) (Figures S1 and S2). The free energy and activation barrier of the L(η 1 -N 2 ) radical anion were significantly dependent on the free energy of neutral L(η 1 -N 2 ) (Figure 3b and 3c). In addition, the free energy of the L(η 1 -N 2 ) radical anion decreased with decreasing E HOMO and increasing E LUMO of neutral L. However, the activation barrier of the formation of the L(η 1 -N 2 ) radical anion decreased significantly with increasing E SOMO(α) of anionic L‾. This result indicates that the SOMO(α) of anionic L‾ largely contributes to the reaction of L with N 2 . Intrinsic reaction coordination calculations and intrinsic bond orbital (IBO) analysis [13] were performed to confirm the electron flow during C-N bond formation between anionic L9 and N 2 (Figure 5). In the optimized transition state, N 2 approached the vertical direction of the carbene. Thus, σ-bond formation between the π* orbital of N 2 and the SOMO(α) of anionic L‾ occurred preferentially. Thereafter, the new π-bond derived from the HOMO was formed when the other π* orbital of N 2 and the HOMO of anionic L‾ were sufficiently close. In summary, we confirmed that the reactivity of L with N 2 was influenced by their HOMO and LUMO energy levels, C-N and C-A bond lengths, bond angles of N-C-A, and nitrogen atom charges. From these results, we discovered that N 2 fixation using the anionic L‾ radical significantly lowers the activation energy of the reaction of L with N 2 via the interaction between the SOMO(α) of L‾ and the π* orbital of N 2 . This study suggested that N 2 fixation on carbon atoms can be achieved through the design of (1) thermodynamically highly stable ambiphilic aminocarbenes or (2) aminocarbenes capable of one-electron reduction, which can serve as an important route for the entry of non-metallic systems.</p>

ChemRxiv

Exploring the pH-dependent structure-dynamics-function relationship of human renin

Renin is a pepsin-like aspartyl protease and an important drug target for the treatment of hypertension; despite three decades\xe2\x80\x99 research, its pH-dependent structure-function relationship remains poorly understood. Here we employed the continuous constant pH molecular dynamics (CpHMD) simulations to decipher the acid/base roles of renin\xe2\x80\x99s catalytic dyad and the conformational dynamics of the flap, which is a common structural feature among aspartyl proteases. The calculated pKa\xe2\x80\x99s suggest that the catalytic Asp38 and Asp226 serve as the general base and acid, respectively, in agreement with experiment and supporting the hypothesis that renin\xe2\x80\x99s neutral optimum pH is due to the substrate-induced pKa shifts of the aspartic dyad. The CpHMD data confirmed our previous hypothesis that hydrogen bond formation is the major determinant of the dyad pKa order. Additionally, our simulations showed that renin\xe2\x80\x99s flap remains open regardless of pH, although a Tyr-inhibited state is occasionally formed above pH 5. These findings are discussed in comparison to the related aspartyl proteases, including \xce\xb2-secretases 1 and 2, capthepsin D, and plasmepsin II. Our work represents a first step towards a systematic understanding of the pH-dependent structure-dynamics-function relationships of pepsin-like aspartyl proteases that play important roles in biology and human disease states.

exploring_the_ph-dependent_structure-dynamics-function_relationship_of_human_renin

<!>Asp38 is the general base and Asp226 is the general acid.<!>Number of hydrogen bonds is the major determinant of the dyad pKa order.<!>Detailed hydrogen bond environment of the aspartyl dyad.<!>Comparison to the pH-activity measurement.<!>Conformational dynamics of the flap.<!>Comparison of the flap conformation with that observed in the X-ray crystal structures.<!>Comparison of the dyad protonation states with other pepsin-like aspartyl proteases.<!>System preparation.<!>Replica-exchange CpHMD simulations.<!>pKa calculations.

<p>Renin-angiotensin-aldosterone system (RAAS) is a critical regulator for blood pressure and systemic vascular resistance.1,2 As a part of the RAAS system, the aspartic protease renin cleaves a protein called angiotensinogen to generate an inactive decapeptide, angiotensin I, which is further cleaved by the angiotensin converting enzyme (ACE) to produce shorter peptides, including the octapeptide angiotensin II which binds and activates the angiotensin II type 1 (AT1) and type 2 (AT2) receptors. The primary effects of AT1 receptor activation include vasoconstriction and stimulation of aldosterone synthesis and release with subsequent sodium and fluid retention, which tend to elevate blood pressure.1 Because of its essential role in the function of RAAS and the high specificity for its only known substrate angiotensinogen, renin is an attractive drug target for the treatment of hypertension. Despite the considerable efforts,3–8 Aliskiren is the only renin inhibitor that has reached the market for the treatment of hypertension. Interestingly, angiotensin-converting enzyme 2 (ACE2), which functions as a receptor for Severe acute respiratory syndrome (SARS) coronaviruses,9 counters RAAS activation by processing angiotensin II10. Thus, the use of Aliskiren and other RAAS inhibitors in COVID-19 patients with hypertension is currently under debate.11. Select preclinical studies have suggested that RAAS inhibitors may increase ACE2 expression, thus raising concerns regarding their safety for Covid-19 patients who have hypertension; however, it was also argued that abrupt withdrawal of RAAS inhibitors in high-risk patients may result in clinical instability and adverse health outcomes.11</p><p>Renin belongs to the A1 family of aspartic peptidases, which are also known as pepsin-like enzymes due to their structural similarity to pepsin.12,13 This subfamily also includes other monomeric aspartic proteases such as cathepsin D (CatD), β-secretases 1 and 2 (BACE1 and BACE2), and plasmepsin II (PlmII). Renin shares a sequence identity level of 46% with CatD, 33% with PlmII, 25% with BACE2, and 24% with BACE1, and the respective sequence similarity levels are 66%, 53%, 47%, and 43% (Figure 1). Similar to these other aspartic proteases, the active form of renin is a 340-residue monomer, which folds to a predominantly β-sheet structure, whereby a β-hairpin loop known as the flap (Thr80 to Gly90) is located over the active site comprised of an aspartic dyad (Asp38 and Asp226)14 (Figure 2).</p><p>The catalytic activity of renin is pH dependent; however, unlike a typical pepsin-like aspartyl protease which has a bell-shaped pH profile with an acidic optimum pH, renin's catalysis occurs in the pH range 5.5–8, with a peak near neutral pH.15–19 Interestingly, several experimental studies found that the peak of renin's pH-activity profile shifts depending on the substrate.15–21 Based on the pH-activity measurements of renin with wild-type and mutant angiotensinogen, it was hypothesized that the active-site residues interact with the His residues at the P2 and P3′ positions, resulting in the unusual optimum pH of renin.17,18,20,22</p><p>The catalytic function of an aspartyl protease is carried out by the catalytic aspartic acids, whereby the general base or nucleophile is deprotonated (lower pKa) while the general acid or proton donor is protonated (higher pKa) in the enzyme active pH range. Thus, a mechanistic understanding of the catalytic function starts with the knowledge of the dyad protonation states.23 However, existing theoretical and computational studies of renin either do not distinguish the dyad residues or assign protonation states to them in an ad hoc fashion.24–26 For example, Calixto et al.26 employed a quantum mechanical/molecular mechanical (QM/MM) method to study the catalytic mechanism of renin, but the protonation states of the dyad were assigned (protonated Asp38 and unprotonated Asp226) based on a proposed mechanism of HIV-1 protease. Note, renin and HIV-1 protease belong to different subfamilies of aspartyl proteases, and there is no experimental or theoretical evidence to support the same pKa order for the aspartic dyad.</p><p>In this work, we employed the hybrid-solvent continuous constant pH molecular dynamics (CpHMD) simulations with pH replica-exchange27,28 to determine the catalytic dyad protonation states of renin and characterize the possibly pH-dependent conformational dynamics of the flap. The replica-exchange hybrid-solvent CpHMD method has been previously applied to study several pepsin-like aspartyl proteases BACE I,29,30 BACE II,31 and CatD30 and obtained dyad pKa order in agreement with experiments. These studies and another work that examined also other enzymes,23 demonstrated that the general base or nucleophile forms more hydrogen bonds than the general acid or proton donor and some of the hydrogen bonds are absent in the crystal structure and emerge during proton-coupled conformational sampling. The present CpHMD simulations of renin confirmed this hypothesis. Interestingly, we found that the aspartic dyad residues in renin have opposite catalytic roles as the analogous residues in BACE1, BACE2, and CatD, despite the sequence and structural similarities. To corroborate the finding, we also performed a simulation for PlmII, which belongs to the same pepsin-like protease family. We found that the aspartic dyad residues in PlmII have the same acid/base roles as the analogous ones in renin. The simulation results allowed us to test the experimental hypothesis and understand why renin's catalytic function occurs in a neutral pH despite the low pKa's of the dyad residues. Finally, we found that renin's flap is open regardless of pH; however, a Tyr-inhibited state due to the formation of a hydrogen bond between Tyr and Asp analogous to BACE1 is formed above pH 5. These findings enable us to systematically understand the pH-dependent structure-activity relationships of pepsin-like aspartyl proteases which play important roles in biology and human disease states.</p><p>We performed the pH replica-exchange hybrid-solvent CpHMD simulations starting from the X-ray crystal structure of the apo human renin (PDB: 2ren14). 24 independent pH replicas in the pH range 1–9 underwent constant NPT simulations, with a total sampling time of 672 ns. To substantiate the statistical significance of our results, another set of replica-exchange CpHMD simulations was performed starting from the X-ray structure of renin in complex with an inhibitor (PDB: 3sfc32) but the inhibitor was removed. The total sampling time was 648 ns. The pKa's of the dyad residues in both sets of simulations were well converged (Figure S1). The data from the first 4 ns (per replica) was discarded in the analysis.</p><!><p>We first examine the pH titration of the aspartic dyad Asp38 and Asp226 in renin. By fitting the residue-specific unprotonated fractions as a function of pH to the generalized Henderson-Hasselbalch equation, we obtained the microscopic pKa's of 3.7 and 5.2 for Asp38 and Asp226, respectively (Figure 3a, Table 1). Like many other catalytic aspartyl dyads,23 the carboxylate sidechains of Asp38 and Asp226 are hydrogen bonded to each other in the crystal structure (a minimum oxygen-oxygen distance of 2.9 Å in PDB: 2ren14). Therefore, to characterize the coupled titration, we calculated the macroscopic stepwise pKa's by fitting the total number of protons of the dyad as a function of pH to a coupled two-proton model (Eq. 2), which resulted in the pKa's of 3.2 and 5.3 (Table 1, Figure S2). Compared to the microscopic pKa's, the splitting between the stepwise pKa's is increased by 0.6 pH units, indicating a small degree of coupling between the titration of Asp38 and Asp226.</p><p>In order to understand the sequence of titration events, we calculated the pH-dependent probabilities of the four possible protonation states of the dyad. As pH decreases from 7 to 4, the probability of the doubly deprotonated state D38−/D226− decreases to nearly zero, while that of the singly protonated state D38−/D226H increases to a maximum of about 75% and that of the alternative singly protonated state D38H/D226− increases to about 10% (Figure 3b). As pH further decreases to 2, the probabilities of both singly protonated states decrease to nearly zero, while that of the doubly protonated state D38H/D226H increases to almost one (Figure 3b). This data indicates that Asp226 accepts a proton first as pH decreases, which is in agreement with its macroscopic pKa (5.3) being higher than Asp38 (3.2).</p><p>To corroborate the results, we performed another set of replica-exchange CpHMD simulations starting from a different crystal structure (PDB ID: 3sfc,32 inhibitor removed). These simulations gave similar results. The residue-specific pKa's are 3.5 and 5.8 for Asp38 and Asp226, respectively, and the stepwise pKa's are 3.3 and 5.9 attributable to Asp38 and Asp226, respectively (Table 1, Figure S2). Thus, our data suggest that Asp38 is the general base and Asp226 is the general acid in renin, which is opposite to what was suggested by Calixto et al. based on the analogy to HIV-1 protease.26</p><!><p>Solvent accessibility, hydrogen bonding, and Coulomb interactions are three major factors contributing to the pKa shift of an amino acid sidechain relative to the solution or model value. Catalytic residues are typically buried in the protein interior which lacks other charged/titratable residues. Our previous work based on CpHMD simulations of several enzymes, BACE1/BACE2, capthepsin D, hen egg lysozyme, and Staphylococcal nuclease showed that solvent accessibility and hydrogen bonding are the determinants of the pKa order for the catalytic dyad.23,30,31 We found that both dyad residues are acceptors to a number of hydrogen bonds formed during the MD simulations, and the general base (or nucleophile) which has a lower pKa forms a significantly larger number of hydrogen bonds than the general acid (or proton donor).23,30,31 We also found that the nucleophile is more solvent exposed as compared to the proton donor.23,30,31 These two observations can be rationalized, as the formation of hydrogen bonds stabilizes the deprotonated aspartate, while solvent exclusion stabilizes the protonated form.</p><p>To test the hypothesis that the general base Asp38 is more solvent accessible than Asp226, we calculated the number of water (hydration number) surrounding each dyad residue as a function of pH. As expected, the hydration number for both Asp38 and Asp226 increases with pH, indicating that water enters the active site as the degree of deprotonation increases (Figure 3c). Closer examination revealed that the hydration increase starts at around pH 4 and plateaus at pH 6–7, which coincides with the decrease in the fraction of the singly protonated state D38−/D226H and with the increase in the fraction of the doubly deprotonated state. In roughly the same pH range, the number of bridging water between the two residues increases from one and then plateaus at about 1.5 (Figure 3c). In the entire pH range, the hydration number for Asp226 is slightly larger than Asp38, although the difference is very small. Thus, the data of renin does not support the hypothesis that the general base is more hydrated.</p><p>Next, we tested the hypothesis that the deprotonated Asp38 forms a larger number of hydrogen bonds than the proton donor Asp226. Below pH 7, Asp38 forms up to one more hydrogen bond than Asp226, while above pH 7 when both residues are fully deprotonated, the number of hydrogen bonds is comparable (Figure 3d). Importantly, the pH profile of the hydrogen bond count for each residue matches the corresponding titration curve. As Asp38 deprotonates in the pH range 2–5, its hydrogen bond count increases from 1 to 1.5; as Asp226 deprotonates in the pH range 4–8, its hydrogen bond count increases from 0.5 to 1.5. This data supports the hypothesis that the Asp38 forms more hydrogen bonds than Asp226, suggesting that it is the driving force for the lower pKa of Asp38 relative to Asp226.</p><!><p>To further understand the physical origin of the different proton affinities of the dyad residues, we examined the specific hydrogen bonds (Figure 4). The carboxylate groups of Asp38 and Asp226 are within hydrogen bonding distance of each other in the crystal structure (PDB: 2ren14). Simulations showed that the D38–D226 hydrogen bonds are formed in a pH-dependent manner in the pH conditions where at least one of the two residues samples the protonated state (pH < 6), indicating that it serves as the proton donor while the other serves as the proton acceptor.</p><p>In addition to the interaction with Asp226, the carboxylate of Asp38 accepts hydrogen bonds from several other residues, including Gly40, Ser41, Tyr83, and Gly228 (Figure 4a and b). Asp38 maintains a stable hydrogen bond with the backbone amide of Gly40 in the entire simulation pH range, and forms hydrogen bonds with the backbone amide and sidechain hydroxyl of Ser41. Interestingly, while the hydrogen bond with the backbone Ser41 becomes more stable with pH (probability increases), the hydrogen bond with the sidechain Ser41 is most stable between pH 4 and 7 and the probability reaches a maximum around pH 5.5.</p><p>In addition to the interaction with Asp38, the carboxylate of Asp226 is a hydrogen bond acceptor to Gly228 and Ala229 (Figure 4c and d). Interestingly, Gly228 can form hydrogen bonds with both dyad residues. While the pH profile for the Gly228–Asp38 hydrogen bond is bell shaped with a maximum round pH 4, the pH profile for the Gly228–Asp226 hydrogen bond has an inverted bell shape with a minimum at the pH 4. This behavior can be readily understood from the difference in the titration pH range for the two residues. In the pH range 2–4, the probability for the Gly228–Asp38 hydrogen bond increases with pH due to the increasing deprotonation of Asp38 while Asp226 remains fully protonated. Above pH 4, the probability for the Gly228–Asp226 hydrogen bond increases due to the increasing deprotonation of Asp226 while Asp38 remains fully deprotonated. Above pH 4, Asp226 also accepts a hydrogen bond from the backbone amide of Ala229, and the probability increase with pH.</p><!><p>The experimental pKa's of the apo renin dyad are unknown. However, the order of the calculated macroscopic pKa's of 3.2(3.3) and 5.3(5.9) attributable to Asp38 and Asp226 in the apo renin, respectively, is in agreement with the pH-dependent activity measurements of renin cleavage reaction with two substrate peptides, the porcine tetradecapep-tide PTDP and the mutant HP2Q.20 PTDP which represents the amino terminus of the wild-type porcine angiotensinogen (P1–P4 sequence identical to the human form) has the sequence Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-His-Leu*-Leu-Val-Tyr-Ser, whereby asterisk denotes the scissile peptide bond, while the mutant HP2Q contains a His-to-Gln substitution at the P2 position. The macroscopic pKa's of 5.3 and 6.3 were obtained in the presence of PTDP, and the pKa's of 4.4 and 7.4 were obtained in the presence of HP2Q.20 The authors suggested20 that both Asp38 and Asp226 in the apo renin are deprotonated at physiological pH, and during catalysis unprotonated Asp38 is the base that abstracts a proton from the nearby catalytic (nucleophilic) water, while Asp226 forms a charged hydrogen bond (salt bridge) with the P2 His, allowing the pair to form the "surrogate" acid (proton donor) at physiological pH.</p><p>The simulation data of the apo renin are consistent with the above experiment. In the simulations, a bridging water between the dyad residues is always present (Figure 3c), which is consistent with presence a catalytic water as proposed by experiment.20 The lower pKa of Asp38 relative to Asp226 is in agreement with the respective roles of general base and acid as suggested by the experiment.20 Our calculated dyad pKa's of 3.2(3.3)/5.3(5.9) are lower than the experimental estimate of 4.4/7.4 (without interaction with the P2 His) for the apo renin. This discrepancy is consistent with our previous CpHMD titration studies of aspartyl proteases which showed a systematic underestimation of 1–1.5 pH units,23,29 likely due to the limitation of the hybrid-solvent method. The lower pKa's compared to experiment may also be attributed to the fact that the active site in the simulations is more hydrated than the substrate-bound state observed in experiment.</p><!><p>A common structural feature of aspartyl proteases is a β-hairpin loop that covers the active site, commonly known as the flap.33,34 Our previous CpHMD simulations showed that the flap in BACE1 is flexible and can open and close in a pH-dependent manner.29 At low and high pH, the flap in BACE1 is closed, while at intermediate pH where the enzyme is active (pH 3.5–5.5), the flap is open.29 In contrast, CpHMD simulations showed that the flap in CatD is rigid and remains open in a broad pH range 2.5–6, consistent with the pH activity profile.30 Thus, we were curious about the conformational dynamics of renin's flap and whether it is pH dependent. To examine the position of renin's flap relative to the dyad, we used two distances, the distance between Tyr83:OH and Asp38:CG (R1) and the distance between Ser84:CB and Asp226:CG (R2, Figure 5a). Tyr83 is a conserved residue among all pepsin-like aspartyl proteases16 and is capable of forming a hydrogen bond with the dyad in BACE1 to prevent substrate access.29,35 Ser84 is located near the tip of the flap and other aspartyl proteases have a similar amino acid at the same location (Figure 2).</p><p>The free energy surface as a function of R1 and R2 was calculated for different pH conditions (Figure 5b–d and Figure S3). At all pH conditions, renin's flap samples a broad free energy minimum region with R1 between 5 and 7.5 Å and R2 between 7 and 12 Å, which will be referred to as the most probable state (black box in Figure 5c). At pH greater than 5.5, however, two additional minima appear with R1 located at 3.5 Å and R2 located at 8 or 13 Å. Close examination revealed that these two minima represent the conformations in which the hydroxyl group of Tyr83 donates a hydrogen bond to the carboxylate of Asp38. An analogous hydrogen bond was found for BACE129,35 and other aspartyl proteases,37,38 and its formation is linked to the flap closure, leading to the so-called Tyr-inhibited state. Our analysis shows that the occupancy of the Tyr83–Asp38 hydrogen bond in renin increases between pH 5 and 7.5 and plateaus at about 20% (Figure S4a), which is significantly lower than the analogous Tyr-Asp hydrogen bond in BACE1.29 It is worth noting that, despite the Tyr-Asp hydrogen bond, the tip of renin's flap can sample two positions, with the R2 value of 8 and 13 Å, respectively, suggesting that the flap remains flexible. This is in stark contrast to the flap of BACE1, as the analogous Tyr is near the tip of the flap, and its hydrogen bonding with the catalytic Asp forces the flap into a closed state. We also note that the pH-dependent formation of the Trp45-Tyr83 hydrogen bond which is conserved in pepsin-like proteases12 correlates well with the pH-dependent probability of the most probable state of renin (Figure S4); a similar observation was made for BACE1.29</p><!><p>We compare renin's flap dynamics observed in the simulations with the available crystal structures of renin in complex with substrate or inhibitor. A scatter plot of 86 PDB entries (165 subunits) of human renin structures (pH conditions vary from 3.0 to 8.5) shows that in a majority of structures, the R1 distance falls into the range of 5.8–7.5 Å and the R2 distance falls into the range of 7.4–12 Å (Figure S5). Thus, the free energy minimum region from the simulations is in agreement with the majority of the X-ray structures (Figure 5c, black box). Curiously, some crystal structures show a widely open flap, with the R1 distance above 8.9 and the R2 distance above 13 Å (Figure S5). This "diffuse" state29 is often found among X-ray structures bound with larger inhibitors, and is correlated with the absence of the Trp45-Tyr83 interaction (Figure S6). We hypothesize that the latter may increase the flexibility of the flap, allowing it to open further. Interestingly, this flap conformation is occasionally sampled at pH 4 (Figure 5b) but missing at higher pH (Figure 5c–d). It is worth noting that the Tyr-inhibited state observed in the simulations is not represented by any crystal structures (Figure 5c and Figure S5), which is perhaps due to the fact that most crystal structures except for two are inhibitor/substrate-bound complexes.</p><!><p>We compare the calculated dyad pKa's of renin with those of the homologous pepsin-like aspartic proteases BACE1,29 BACE2,23,31 and CatD,30 which share a sequence similarity level of 43%, 47%, and 66%, respectively, with renin (Figure 1). Surprisingly, the pKa order of renin's catalytic dyad is opposite to that of BACE1, BACE2, and CatD. According to the CpHMD simulations and experiment,23,29–31 the residue homologous to Asp38 in renin, Asp32 in BACE1, Asp48 in BACE2, or Asp33 in CatD, has a higher pKa than the residue homologous to Asp226 in renin, Asp228 in BACE1, Asp241 in BACE2, or Asp231 in CatD (Table 1). The relative pKa's suggest that Asp32, Asp48, and Asp33 are protonated serving the role of a general acid, while Asp228, Asp241, and Asp331 are deprotonated serving the role of a general base in the catalytic reactions of BACE1, BACE2, and CatD, respectively.</p><p>Interestingly, the relative pKa's of renin is identical to those of the pepsin-like protease PlmII which shares 53% sequence similarity with renin (Table 1). The CpHMD titration showed that Asp34 and Asp214 in PlmII, which are analogous to Asp38 and Asp226 in renin, have the stepwise pKa's of 3.4 and 4.3, and thus serving the roles of general base and general acid, respectively. We note, a detailed study of PlmII will be published in a future work. Taken together, the comparison shows that the catalytic roles of the dyad are not conserved among the pepsin-like proteases.</p><p>CpHMD simulations have been performed on the apo renin to understand the roles of the catalytic dyad and conformational dynamics of the flap. The calculated macroscopic pKa's attributable to Asp38 and Asp226 from two sets of CpHMD simulations are 3.2(3.3) and 5.3(5.9), respectively, suggesting that Asp38 serves as a general base and Asp226 serves as a general acid during renin catalysis. Interestingly, the relative pKa order of the catalytic dyad in renin is opposite to BACE1, BACE2, CatD, but identical to PlmII, suggesting that the acid/base roles are not conserved among pepsin-like aspartyl proteases. We also note that the acid/base roles of the aspartyl dyad in HIV-1 protease are opposite to those in renin. Thus, assigning protonation states for renin based on HIV-1 protease26 is incorrect. The analysis of CpHMD trajectories shows that the deprotonated carboxylate of Asp38 forms more hydrogen bonds than Asp226, consistent with our previous finding that the general base (or nucleophile) of a catalytic dyad accepts more hydrogen bonds than the general acid (or proton donor) and that some of the hydrogen bonds are absent in the X-ray crystal structure but emerge during the proton-coupled conformational sampling in the simulation.23</p><p>The catalytic roles of the renin dyad residues assigned by the CpHMD simulations are in agreement with those inferred from the pH-activity measurement of the renin cleavage reaction with two peptide substrates.20 The latter yielded Asp38/Asp226 pKa's of 5.3/6.3 in the presence of the wild-type substrate and 4.4/7.4 in the presence of a mutant substrate, in which the P2 His was substituted by Gln. The differences in the pKa's were hypothesized to originate from a charged hydrogen bond between Asp226 and the P2 His, which allows the Asp226–His ion-pair to act as a general acid in catalysis at neutral pH.20 The Asp226–His interaction was also hypothesized to raise the pKa of Asp38, thus shifting the enzyme optimum pH higher to the neutral range.20 Considering the known systematic underestimation of 1–1.5 pH units for the aspartyl protease dyad pKa's by hybrid-solvent CpHMD,23,29 our calculated pKa's of 3.2(3.3)/5.3(5.9) for Asp38/Asp226 are consistent with the experimental estimate of 4.4/7.4 for the apo renin, thus supporting the hypothesis that renin's higher optimum pH relative to other pepsin-like proteases is due to the effect of substrate interactions. To directly test the hypothesis of substrate-directed catalysis, future simulations will be carried out using the substrate-bound renin structure.</p><p>The CpHMD simulations also revealed the distinctive flap dynamics in renin. Unlike the flap in BACE1 which displays pH-dependent dynamics, renin's flap is open regardless of pH, similar to BACE2 and CatD. However, the Tyr-inhibited state, in which the conserved Tyr83 forms a hydrogen bond with Asp38, is occasionally sampled above pH 5 and the population increases to 20% at pH 7.5; this state is not sampled by BACE2 and CatD. Interestingly, while the analogous hydrogen bond in BACE1 prevents the flap from opening, the flap in renin remains open, as Tyr83 is positioned lower on the flap and not on the tip as in BACE1. Aspartyl proteases are an important class of enzymes; our work demonstrates that CpHMD simulations is a powerful tool for advancing the detailed knowledge of their pH-dependent structure-function relationships which remain poorly understood.</p><!><p>The coordinates of the X-ray crystal structures of human renin (PDB: 2ren,14 apo) and an inhibitor-bound complex (PDB: 3sfc,32 subunit B) were retrieved from the PDB. The positions of the missing residues 53–55, 166–170, and 287–295 in the apo renin structure (2ren) were built by superimposing the backbone atoms onto those of the holo structure (3sfc) which has coordinates for all residues. The root-mean-square deviation (RMSD) of the backbone atoms between the two structures was 0.84 Å. For PlmII, the X-ray crystal structure (PDB: 1sme39) was used with the ligand being removed. The hydrogen atoms were added using the HBUILD facility in CHARMM.40 To remove unfavorable contacts, the apo renin structure was first minimized 100 steps using the Adopted Basis Newton-Raphson (ABNR) method, and the PlmII structure was energy minimized using 10 steps of steepest descent (SD) and 10 steps of Adopted Basis Newton-Raphson (ABNR) method. The protein was then solvated in an octahedral water box with a heavy-atom distance of at least 10 Å between the protein and the edge of the water box. Following solvation, the water positions were energy minimized in several stages. First, 50 steps of SD followed by 50 steps of ABNR minimization was performed, with the protein heavy atoms fixed. Next, a five-stage restrained minimization was performed, where the harmonic force constant on the backbone heavy atoms was 100, 50, 25, 5, and 0 kcal·mol−1·Å−2. Each stage included 50 steps of SD and 100 steps of ABNR, except for the first stage which included 50 steps of SD and 10 steps of ABNR minimization.</p><p>Two simulations of renin were performed; run 1 started from the apo structure and run 2 started from the holo structure with the inhibitor removed. Only one simulation was performed for PlmII. All simulations were performed with the CHARMM program c36a2.40 The protein was represented by the CHARMM22/CMAP all-atom force field,41,42 and water was represented by the modified TIP3P water model.40 The system was gradually heated from 100 K to 300 K over the course of 120 ps. The system was subsequently equilibrated for 280 ps in four stages, where the harmonic force constant for the protein heavy atoms was 5 (40 ps), 1 (40 ps), 0.1 (100 ps), and 0 (100 ps) kcal·mol−1·Å−2 (100 ps). The system was further equilibrated for 580 ps without any restraint. In the heating and equilibration stages, constant pH functionality (PHMD module in CHARMM) was turned on and pH was set to the crystal pH conditions (pH 4.7 for the renin simulation run 1, pH 4.5 for the renin simulation run 2, and pH 6.5 for the plmII simulation).</p><!><p>Following equilibration, hybrid-solvent CpHMD simulations with the pH replica-exchange protocol were performed. Detailed methodology can be found in the original work27 and a review.28 24 replicas were placed in the pH range 1–9 (1–8.5 for plm II). Each replica was simulated in the NPT ensemble at 300 K and 1 atm. The particle mesh Ewald method43 was used to calculate long-range electrostatic interactions, with a real space cutoff of 12 Å and a sixth-order interpolation with a 1-Å grid spacing. The SHAKE algorithm was used to constrain bonds involving hydrogen atoms to enable a 2-fs timestep. A Generalized Born (GB) calculation was invoked every 10 molecular dynamics (MD) steps to update the titration coordinates. Every 500 MD steps (1 ps), the adjacent pH replicas attempted to swap conformational states based on the Metropolis criterion.27 All sidechains of Asp, Glu, and His residues were allowed to titrate. Each replica ran for 28 and 27 ns in the renin simulation run 1 and 2, with the aggregate sampling time of 672 ns and 648 ns, respectively, In the PlmII simulation, each replica ran for 29 ns, with the aggregate time of 696 ns. To verify convergence, the time series of the cumulatively calculated pKa values of the catalytic dyad were examined (Figures S1). For analysis, the data from the first 4 ns per replica was discarded for renin and the first 9 ns was discarded for PlmII.</p><!><p>In CpHMD simulations,28 the continuous variables λ and x are used to represent the titration coordinates and tautomer interconversion, respectively. A protonated state was defined as those with λ < 0.2, and x < 0.2 or x > 0.8, while an unprotonated state was defined as those with λ > 0.8, and x < 0.2 or x > 0.8. Accordingly, the fraction of unprotonated state (S) was calculated for each titratable site at each simulation pH. The microscopic residue-specific pKa was calculated by fitting S at different pH to the generalized Henderson-Hasselbalch (modified Hill) equation, (1)S=11+10n(pKa−pH), where n is the Hill coefficient that represents the slope of the transition region in the titration curve. To compare with experiment, we also calculated the macroscopic stepwise pKa's by fitting the total number of bound protons to the catalytic dyad (Nprot) to the following statistical mechanics based two-proton model:44,45 (2)Nprot =10pK2−pH+2×10pK1+pK2−2pH1+10pK2−pH+10pK1+pK2−2pH where pK1 and pK2 are the macroscopic pKa's of the dyad, and the denominator represents the partition function.</p>

PubMed Author Manuscript

Iodide-enhanced palladium catalysis via formation of iodide-bridged binuclear palladium complex

The prevalence of metalloenzymes with multinuclear metal complexes in their active sites inspires chemists' interest in the development of multinuclear catalysts. Studies in this area commonly focus on binuclear catalysts containing either metal-metal bond or electronically discrete, conformationally advantageous metal centres connected by multidentate ligands, while in many multinuclear metalloenzymes the metal centres are bridged through μ2-ligands without a metal-metal bond. We report herein a μ2-iodide-bridged binuclear palladium catalyst which accelerates the C-H nitrosation/annulation reaction and significantly enhances its yield compared with palladium acetate catalyst. The superior activity of this binuclear palladium catalyst is attributed to the trans effect-relay through the iodide bridge from one palladium sphere to the other palladium sphere, which facilitates dissociation of the stable six-membered chelating ring in palladium intermediate and accelerates the catalytic cycle. Such a trans effect-relay represents a bimetallic cooperation mode and may open an avenue to design and develop multinuclear catalysts.

iodide-enhanced_palladium_catalysis_via_formation_of_iodide-bridged_binuclear_palladium_complex

<!>Results and discussion<!>Methods

<p>T he prevalence of the metalloenzymes containing binuclear or polynuclear transition metal complexes in their active sites has motivated the interest of chemists in the binuclear or polynuclear homogeneous catalysts for organic transformations [1][2][3][4][5][6][7][8][9][10] . A number of well-defined binuclear catalysts have been reported with unique activity and selectivity accessible through the cooperation between two metal centres 1,[4][5][6][8][9][10][11][12] . Two types of bimetallic cooperation are commonly encountered in the catalytic processes of these binuclear catalysts: one is the simultaneous activation of two reaction partners by the binuclear catalysts in which multidentate ligands hold two metal centres in close proximity to create two electronically discrete, conformationally advantageous active sites for binding the corresponding substrate molecules 1,4,5,13,14 , as illustrated by olefin polymerization 13 or enantioselective binuclear catalysts 5,14 ; the other is that two metal centres of the metal-metal bond-containing binuclear catalyst synergistically participate in elemental redox steps in catalytic pathways 6,9,10,[15][16][17][18][19][20][21][22][23][24][25][26][27][28][29][30][31][32] , as exemplified by Pd(III) dimer-catalysed oxidative C-H functionalization [17][18][19] and dirhodium-catalysed reactions for carbene [20][21][22] or nitrene insertion 23 into C-H bond. In many of the metalloenzymes, on the other hand, metal centres at the active site are held together by μ2bridging ligands, with the metal-metal bond absent, implicating that the catalysis therein may invoke different bimetallic or polymetallic cooperation in which μ2-bridging ligands likely play key roles. Disclosing the bimetallic or polymetallic cooperation mechanism for catalysis would provide platforms to rationally design the binuclear or polynuclear metal catalysts with the unique activity and selectivity. Due to the rapid equilibration of the multinuclear complexes with other metal species in the catalytic conditions, however, it has been highly challenging to identify the active species and reveal the responsible cooperation mechanism in a definitive manner 7 , which is the reason why the promising binuclear or multinuclear metal catalysts are relatively underdeveloped compared with the mononuclear metal catalysts that dominate in homogenous catalysis.</p><p>Here, we report an iodide-bridged binuclear palladium catalyst generated in situ from palladium acetate, azobenzene and tetra-nbutyl ammonium iodide (TBAI), which accelerates C-H nitrosation/annulation reaction and significantly enhances its yields compared with palladium acetate alone as a catalyst (Fig. 1). This binuclear palladium species, according to the kinetic studies, retains the integrity of its iodide-bridged binuclear core structure during the catalytic cycle. Computational studies further reveal that a strongly σ-donating η 1 phenyl ligand around one palladium centre of this binuclear complex is able to exert a trans effect 33 , through the bridging iodide ligand, on the ligand at the other palladium centre and therefore labilize the coordination bond trans to this bridging iodide 34,35 . Consequently, two palladium centres within the binuclear cluster cooperatively decrease the activation barriers of dissociation of the chelating product fragment from catalyst and accelerate the whole catalytic cycle. The binuclear metal catalyst that features the trans effect-relay through bridging ligand may provide a solution to metalcatalysed efficient syntheses of the chelating compounds that often impede catalytic cycle.</p><!><p>Discovery of beneficial effect of iodide. Our interest in the binuclear palladium catalyst stemmed from a discovery during exploration of palladium-catalysed ortho C-H bond nitrosation reaction of azobenzenes. The development of this Pd-catalysed aryl C-H nitrosation was aimed at achieving a straightforward approach to nitrosoarenes that are a class of versatile synthetic intermediates utilized in a variety of transformations 36 . Importantly, the metal-catalyzed aryl C-H nitrosation method has a potential to get over substrate limitation, poor regioselectivity 37 and substrate pre-activation 36 problems encountered in the existing methods for syntheses of nitrosoarenes. To this end, we started our investigation by examining the reaction of azobenzene (1a) with two equivalents of nitrosonium tetrafluoroborate ([NO] [BF 4 ]) (2) conducted in 1,2-dicholobenzene (DCB) of 1.5 mL at 80 °C for 24 h in the presence of 1 mol% Pd(OAc) 2 as a catalyst (Table 1). The initial reaction conditions enabled the desired ortho C-H nitrosation but generated 2H-benzotriazole N-oxide product (2-phenyl-2H-benzo[d][1,2,3]triazole 1-oxide 3a) as a final product in only 7% yield via a C-H nitrosation/annulation sequence (entry 1 in Table 1).</p><p>The fact that 2H-benzotriazole heterocyclic N-oxides are the privileged structural motifs in biologically active compounds, pharmaceuticals, and functional materials [38][39][40] prompted us to identify the optimal conditions for this Pd-catalysed reaction of azobenzene with [NO][BF 4 ]. Increasing catalyst loading to 10 mol % offered a 46% yield of the desired 3a with azobenzene recovered (entry 2 in Table 1). Screening a variety of additives revealed that 10 mol% tetra-n-butyl ammonium chloride (TBAC), in combination with 10 mol% Pd(OAc) 2 , enhanced the yield of 3a to 76% (entry 3 in Table 1). Interestingly, TBAI exhibited a more beneficial effect on the reaction than TBAC. 1 mol% TBAI allowed using 1 mol% Pd(OAc) 2 to achieve a 90% yield (entry 4 in Table 1). Considering that p-toluenesulfonic acid (TsOH) may facilitate cyclopalladation of azobenzene, 2 mol% TsOH was introduced into the reaction system with 1 mol% TBAI and 1 mol % Pd(OAc) 2 , but still afforded a 90% yield (entry 5 in Table 1). Other protonic acids did not show beneficial effect on the reaction yield as well (see Supplementary Table 1). Control experiment showed that Pd(OAc) 2 was indispensable for this reaction (entry 6 in Table 1). The beneficial effect of iodide on the Pd-catalysed C-H nitrosation/annulation sequence is to some extent surprising because it is at variance of the previous findings that iodide anion tends to retard late transition metal-catalysed C-H functionalization reactions by weakening electrophilicity of transition metal ions 41,42 . Although halide anions including iodide enhance Pd-catalysed traditional cross-coupling reactions of aryl halides with nucleophilic organometallics 43 , these Pdcatalysed traditional cross-coupling reactions via Pd(0)/Pd(II) cycle are distinctly different from the Pd-catalysed oxidative C-H nitrosation/annulation reaction that would begin with the electrophilic cyclopalladation of Pd(II) species and involves Pd (II)/Pd(III) [17][18][19] or Pd(II)/Pd(IV) 18,44,45 cycle.</p><p>Identification of iodide-bridged binuclear palladium complex as a catalytically active intermediate. Intrigued by this beneficial effect of iodide, we performed mechanistic investigation to identify its origin. Initially, we tried to isolate a palladium intermediate lying on catalytic cycle. Under the conditions mimicking the aforementioned Pd(OAc) 2 /TBAI-catalysed reaction, the reaction of 0.2 mmol Pd(OAc) 2 with one equivalent of TBAI, 100 equivalents of azobenzene and 200 equivalents of [NO][BF 4 ] afforded an iodide-bridged binuclear palladium complex bearing cyclopalladated azobenzene ligand as dark red solid (4a) (Fig. 2a). Single-crystal X-ray diffraction analysis revealed that the molecule of 4a comprises a twofold iodide-bridged binuclear palladium core with each palladium atom chelated by an azobenzene ligand through nitrogen and carbon atoms (Fig. 2a) 46 , and square planar coordination environments around these two palladium atoms are identical and lie in the same plane because they are related by centrosymmetry. Within 4a, the I-Pd bond trans to Pd-C bond is 0.12 Å longer than the other I-Pd bond between the iodide atom in question and the other palladium centre owing to strong trans influence of phenyl ligand, leading to unsymmetrical iodide bridges. Because the poor solubility of crystalline 4a in 1,2-dichlorobenzene impeded the investigations of reactivity of 4a, we turned our attention to synthesis of highly soluble congeners of 4a. The treatment of Pd(OAc) 2 with one equivalent of TBAI and 12 equivalents of 4,4′-di-n-butyl-azobenzene (1b) in DCB at 80 °C for 12 h afforded a soluble analogue of 4a, an iodide-bridged binuclear palladium complex of cyclopalladated 4,4′-di-n-butyl-azobenzene ligand (4b) in 65% yield with concomitant formation of [Pd 2 I 6 ](n-Bu 4 N) 2 47 in 11% yield (Fig. 2b). The cyclopalladated azobenzene ligands in 4a and 4b indicate that Pd-introduced C-H bond cleavage can occur in the presence of iodide anion, though iodide anion was thought to tend to form stable Pd-I bond and thus weaken the Pd electrophilicity that is required for C-H metallation 41 .</p><p>Using soluble 4b as a model compound, reactivity and kinetics of the iodide-bridged binuclear palladium complexes were then examined to identify their roles in the catalytic process. The reaction of 4b with 3 equivalents of [NO][BF 4 ] at room temperature for 24 h showed that cyclopalladated 4,4′-di-nbutyl-azobenzene ligand in 4b underwent nitrosation and subsequent annulation to form 2H-benzotriazole N-oxide derivative bearing two n-butyl groups (3b) in 38% yield (Fig. 2c), which raised the possibility that such an iodide-bridged binuclear palladium complex was a reaction intermediate in the catalysis cycle. Furthermore, 1.5 mol% 4b, together with 3 mol% tetra-nbutyl ammonium acetate (n-Bu 4 NOAc) as an additive, served as a catalyst to effect the nitrosylation/annulation reactions of both 4,4′-di-n-butyl-azobenzene (1b) and 4,4′-dimethyl-azobenzene (1c) in the yields (Fig. 2d) comparable with those obtained with the combination of 3 mol% Pd(OAc) 2 and 3 mol% TBAI (see below). The use of 3 mol% n-Bu 4 NOAc as an additive in these 4b-catalysed reactions is because the reaction of azobenzene catalysed by Pd(OAc) 2 and TBAI might generate in situ the iodide-bridged binuclear palladium complex along with 2 equivalents of n-Bu 4 NOAc and 2 equivalents of acetic acid. In the 4b-catalysed transformation of azobenzenes to the corresponding 2H-benzotriazole N-oxide products, 1.5 mol% 4b provided slightly higher initial reaction rate compared with the combination of 3 mol% Pd(OAc) 2 , 3 mol% TBAI and 6 mol% TsOH, indicating that this binuclear palladium complex 4b is a kinetically competent catalyst (Fig. 2e). In the experiments determining the initial rate for reaction with 3 mol% Pd(OAc) 2 and 3 mol% TBAI, 6 mol% TsOH was introduced as an additive into the reaction system to neutralize OAc − anion from Pd (OAc) 2 and rule out the influence of OAc − when comparing with the reaction with 4b. The initial rates of the nitrosylation/ annulation of 1c exhibited a first-order dependence on the concentration of 4b (Fig. 2f), supporting that the iodide-bridged binuclear palladium complex 4b and its congeners are catalytically active species, and retain integrity of its [Pd 2 I 2 ] core structure during catalysis. The first-order dependence of the initial reaction rate on the concentration of 4b may also account for the observation that the reaction catalysed by 1.5 mol% 4b was slightly faster than the reaction catalysed by the combination of 3 mol% Pd(OAc) 2 and 3 mol% TBAI since the concentration of the binuclear palladium complex generated in situ from self-assembly of 3 mol% Pd(OAc) 2 , 3 mol% TBAI and azobenzene is likely less than 1.5 mol%.</p><p>To confirm that 4b and its congeners are the catalytically active intermediates lying on catalytic cycle, we need to preclude the possibility that 4b and its congeners are the off-cycle precatalysts. The acetate-bridged binuclear palladium complex bearing a cyclopalladated 4,4′-di-n-butyl-azobenzene ligand on each palladium centre (5b) 48 was observed to react with 2 equivalents of TBAI to lead to replacement of acetate bridge by iodide bridge and afford 4b in 74% yield (Fig. 3a). The conversion of 5b to 4b in turn implicated that the iodide-bridged binuclear palladium complexes were stable against the Pd-I bond cleavage caused through the ligand substitution, which is consistent with the fact that 4b was generated from the reaction of Pd(OAc) 2 with excessive coordinating azobenzenes. Moreover, treatment of 0.1 mmol of 4b with 3 equivalents of [NO][BF 4 ] and 6 equivalents of 4,4′-di-n-butyl-azobenzene in 5 mL DCB at 90 °C for 4 h gave rise to formation of 2H-benzotriazole N-oxide derivative 3b in 133% yield relative to 4b with 21% of 4b recovered (Fig. 3b), which supported that 4b could be re-generated during catalysis and that [NO][BF 4 ] did not fragment this binuclear palladium complex via oxidation of bridging iodide. [Pd 2 I 6 ](n-Bu 4 N) 2 (1.5 mol%), a sideproduct in the preparation of 4b (Fig. 2b), was observed to catalyse the reaction of 1a with [NO][BF 4 ] to afford 3a in 82% yield (Fig. 3c). However, in the stoichiometric reaction of [Pd 2 I 6 ] As shown in Fig. 2e, the reaction catalysed by the combination of 3 mol% Pd(OAc) 2 and 3 mol% TBAI was two times as fast as the reaction catalysed by 3 mol% Pd(OAc) 2 , revealing that 4a as a catalyst is more active than Pd(OAc) 2 . 1.5 mol% 5b provided the similar initial rate to 3 mol% Pd(OAc) 2 , supporting the reaction catalyzed by Pd(OAc) 2 proceeds via cyclopalladation of azobenzenes to form analogues of 5b. In line with this surmise, an acetate-bridged binuclear palladium complex bearing cyclopalladated unsubstituted azobenzene ligand (5a) 48 , an analogue of 5b, reacted with [NO][BF 4 ], in the presence of one equivalent of deuterated azobenzene (1a-D 10 ) at 80 °C for h, to produce the expected 3a in 40% yield (Fig. 3e). In contrast, the stoichiometric reaction of 5a with [NO][BF 4 ] gave none of 3a in the absence of 1a-D 10 , which may implicate the reason why the reaction of 1a with [NO][BF 4 ] catalyzed by 10 mol% Pd(OAc) 2 alone did not go to completion.</p><p>Computational studies on the origin of the high activity of iodide-bridged binuclear palladium catalyst. The density functional theory (DFT) calculations on the mechanism for the nitrosylation/annulation reaction of azobenzene with [NO][BF 4 ] catalysed by 4a. For comparison, DFT studies of this process with Pd(OAc) 2 alone as the precatalyst were also carried out, in which acetate-bridged binuclear palladium complex 5a is assumed to be the catalyst generated in situ on the basis of aforementioned investigations.</p><p>Figure 4a presents the computed reaction pathways with catalyst 4a (blue line) and catalyst 5a (red line). These two pathways comprise similar elemental step. Reactions begin with oxidative addition of nitrosonium to one of two palladium atoms in 4a and 5a to produce higher oxidation state nitrosyl-palladium intermediates (LM1′ and LM1) (see Fig. 16), which are followed by reductive elimination step to construct C-N bond and form six-membered chelation ring complexes (LM2′ and LM2). Then, assisted by the coordination of azobenzene ligand to palladium centre, dechelation of such sixmembered chelation rings via TS2′ and TS2, and subsequent N-N bond formation for annulation proceed to release the target product 2H-benzotriazole N-oxide, respectively. Finally, with azobenzene as a base, azobenzene coordinating to palladium centre undergoes C-H palladation via a concerted metalation-deprotonation pathway to re-generate catalysts (4a and 5a) from the corresponding σ-complexes LM4′ and LM4. In the pathway of 4a, redox steps involve the change of the formal oxidation state of only one palladium atom from II to IV 18 , while in the pathway of 5a, both palladium atoms synergistically participate in redox steps via switch between Pd(II)-Pd(II) and Pd (III)-Pd(III) oxidation states [17][18][19] . The rate-determining step in the pathway of 4a is the C-H activation step with the activation barrier of 24.8 kcal/mol, in accordance with the experimentally observed intermolecular kinetic isotope effect (KIE) value of 4.4 (see Supplementary Fig. 14). In contrast, the step for dechelation of the palladium intermediate containing six-membered chelating ring is the rate-determining step in the pathway of 5a with the activation barrier of 27.5 kcal/mol. The lower activation barrier in the pathway of 4a is consistent with the observation that the 4bcatalysed reaction is faster than 5b-catalysed reaction (Fig. 2e).</p><p>Figure 4b shows the calculated structures of two palladium intermediates containing six-membered chelating rings (LM2 and LM2′). As reflected by bond lengths of the structure of LM2′, phenyl part of cyclopalladated azobenzene weakens the trans Pd2-I2 bond due to its strong trans influence, and therefore strengthens the Pd1-I2 bond, which enhances the trans influence of I2 on Pd1-N1 bond to labilize Pd1-N1 bond for dechelation. As a result, the calculated bond length for Pd1-N1 bond of iodide-bridged binuclear palladium species is longer by 0.045 Å than the corresponding bond of acetate-bridged species. In line with the bond lengths, the calculated Pd1-N1 bond strength in 4a-derived LM2′ is weaker than that of the 5a-derived LM2 (Fig. 4c). The further calculations (Fig. 4b) on the dechelation step without aid of azobenzene illustrate that the dechelation/ N-N bond formation step of 5a-derived LM2 is infeasible in kinetics with a high free energy barrier of 34.3 kcal/mol (TSA), while the same step of 4a-derived LM2′ only needs to experience a lower barrier of 25.5 kcal/mol (TSA′). Such a conclusion is supported by the experimental observations that in the absence of free deuterated azobenzene, the reaction of 5a with [NO][BF 4 ] at 80 °C did not produce 3a (Fig. 3e), while without any free azobenzene, 4b reacted with [NO][BF 4 ] at room temperature to give the expected product (Fig. 2c), and that the 10 mol% Pd (OAc) 2 -catalysed reaction of 1a with [NO][BF 4 ] did not go completion (entry 2, Table 1). As such, the trans effect-relay through bridging iodide in the 4a-derived binuclear palladium intermediate make the dechelation step easier and accelerates the reaction.</p><p>Substrate scope. This Pd-catalyzed C-H nitrosylation/annulation reaction of azobenzenes with [NO][BF 4 ] is quite general. To achieve the satisfied yields, some of reactions were conducted at 90 °C for 48 h with 3 mol% Pd(OAc) 2 , 3 mol% TBAI and 6 mol% TsOH. In some cases, the combination of 0.5 mol% Pd(OAc) 2 and 0.5 mol% TBAI afforded good yields. As shown in Fig. 5a, a spectrum of 4,4′-di-substituted symmetrical azoarenes bearing nbutyl, methyl, cyclohexyl, chloro, bromo, iodo and ester substituents could participate in the reaction to generate the corresponding products in good-to-excellent yields (3b-k) with the exception of 3e and 3f. The low yield of 3e and 3f may result from the steric hindrance from ortho-substitutes, which impeded the ligation of azobenzene nitrogen atoms to palladium centre. As shown in Fig. 5b, electronically unsymmetric azobenzenes containing an array of functional groups preferentially underwent reactions at more electron-rich benzene moiety with good-toexcellent yields obtained (3l-3s). For unsymmetric azobenzenes containing ortho-substituted benzene moieties, ortho-substituents led reactions to selectively occur at ortho-substituted benzene rings in good yields (3t-3z, 3aa-3ae). Ortho-chloro group, an electronwithdrawing group, favoured reaction at chloro-containing The reaction was run at 100 °C for 24 h. ii The ratio of product isomers was determined by NMR. iii The reaction was run in nitrobenzene (1.5 mL) at 100 °C for 24 h. iv 0.6 mmol scale reaction was run at 80 °C for 24 h, with [NO][BF 4 ] (2 equiv), Pd(OAc) 2 (0.5 mol%), TBAI (0.5 mol%). v The reaction was run in DCB (1.0 mL) and nitrobenzene (0.5 mL) at 90 °C for 48 h.</p><p>benzene moiety (3ab), which illustrated that the effect overrode the electronic effect in the control of reaction selectivity. The regioselectivity controlled by steric factor was also observed in 3,5-dimethyl azobenzene that favoured the reaction at less sterically hindered moiety with 18.5:1 selectivity (3ae). In summary, an iodide-bridged binuclear palladium complex has been identified as an efficient catalyst to enhance both yield and rate of the nitrosylation/annulation reaction of azobenzenes with [NO][BF 4 ]. The good performance of this binuclear catalyst arises from the bimetallic cooperation by which the strongly σdonating η 1 phenyl ligand on the "spectator" metal centre exerts a trans on the chelating fragment on the catalytic metal centre through a bridging iodide ligand, and facilitates product release to re-generate catalytically active species. The trans effect-relay through bridging ligand within a binuclear complex represents a new bimetallic cooperation mode for catalysis and opens an avenue to and develop multinuclear catalysts, especially for syntheses of chelating products that often impede re-generation of active metal catalysts therefore retard cycles.</p><!><p>General procedure for the reaction of azobenzene with [NO][BF 4 ]. In a glove box, a mL of the Schlenk tube equipped with a stir bar was charged with Pd (OAc) 2 (3 mol%, 0.0014 g), TBAI (3 mol%, 0.0022 g), p-toluenesulfonic (6 mol%, 0.0021 g), azobenzene (0.2 mmol), [NO][BF 4 ] (0.6 mmol, g) and DCB (1.5 mL). The tube was sealed removed out of the glove box. The reaction mixture was stirred at 90 °C for 48 h. Upon completion, the reaction mixture was diluted with 10 mL of ethyl acetate, filtered through a pad of silica gel, followed by washing the pad of the silica gel with the ethyl acetate (20 mL). The filtrate was concentrated under reduced pressure. The residue was then purified by chromatography on silica gel to provide the corresponding product.</p><p>Optimization studies. See Supplementary Methods and Supplementary Table 1.</p><p>Identification of iodide-bridged binuclear palladium complex as a catalytically active species. See Supplementary Methods for details.</p>

Nature Communications Chemistry

Epigenetic Therapy of Leukemia: an update

Carcinogenesis is classically thought to result from genetic alterations in DNA sequence such as deletions, mutations, or chromosomal translocations. These in turn may lead to the activation of oncogenes, inactivation of tumor suppressor genes or formation of chimeric oncoproteins. Epigenetics, in contrast, refers to a number of biochemical modifications of chromatin, either to DNA directly or to its associated protein complexes, that affect gene expression without altering the primary sequence of DNA1,2. A fundamental difference between genetic and epigenetic alterations is the irreversible nature of genetic lesions whereas epigenetic ones are potentially reversible, allowing for therapeutic intervention. In the last decade, it has become apparent that epigenetic changes play an important role in cancer, particularly in leukemia. Significant advances have been made in the elucidation of these processes as well as in translating this knowledge to the clinic, as in the development of new prognostic biomarkers or targeted therapies. In this review, we will focus on recent advances in epigenetic therapy in leukemia.

epigenetic_therapy_of_leukemia:_an_update

TARGETING ABERRANT DNA METHYLATION<!>5-AZACITIDINE<!>Impact on survival of 5-azacitidine in MDS<!>Risk of AML transformation with 5-azacitidine in patients with MDS<!>Impact on quality of life with 5-azacitidine<!>Transfusion independence in CALGB studies<!>Time to response and duration of therapy with 5-azacitidine<!>Route and schedule of administration of 5-azacitidine<!>Use of growth factors<!>5-AZA-2\xe2\x80\x99-DEOXYCYTIDINE (DECITABINE)<!>DECITABINE VERSUS 5-AZACITIDINE<!>TARGETING HISTONE ACETYLATION<!>Sodium phenylbutyrate<!>Valproic acid<!>Depsipeptide (FK228)<!>MS-275<!>LBH589<!>MGCD0103<!>Vorinostat (Suberoylanilide hydroxamic acid, SAHA)<!>COMBINATION EPIGENETIC THERAPY<!>Decitabine and VPA<!>5-azacitidine, VPA and ATRA combination<!>5-azacitidine and PB<!>5-azacitidine and MGCD0103<!>5-azacitidine and MS-275<!>CONCLUSIONS AND NEEDS<!>

<p>DNA methylation refers to the addition of a methyl group to the C5 position of the pyrimidine ring of cytosine (C).1 This reaction occurs only when the (C) is followed by guanine (G) (so called CpG dinucleotides), and is mediated by enzymes called DNA methyltransferases (DNMTs) where S-adenosyl-methionine serves as the methyl donor. CpG dinucleotides are found with a lower frequency in the human genome than expected assuming a statistically random distribution, but they are clustered. CpG islands are defined as regions of greater than 500 base pairs with a CG content greater than 55%. They may be located near the promoter and exonic regions of approximately 40% of the mammalian genes (promoter associated CpG islands) where they are generally not methylated. In contrast, CpG islands in intergenic regions are normally methylated. This may contribute to the transcriptionally inert state of non-coding DNA, while the unmethylated configuration of the promoter regions of CpG islands allows for gene expression. In cancer cells, these promoter associated CpG islands can be hypermethylated. The resulting aberrant gene silencing is functionally similar to inactivation by genomic alterations such as mutations or deletions.2,3 Induction of hypomethylation of these promoter associated CpG islands can lead to gene re-expression and therefore be of great clinical value. Currently, two hypomethylating agents are in clinical use: 5-azacitidine and 5-aza-2`-deoxycytidine (decitabine) (Figure 1).</p><!><p>5-azacitidine is a nucleoside analogue with the capacity to induce DNA hypomethylation in vitro and in vivo. It was initially investigated at high doses (600–1500 mg/m2 per course) as a cytotoxic agent for hematological malignancies in 1970s and 80s. These studies found high rates of toxicity, including severe and prolonged myelosuppression and thus high-dose 5-azacitidine was not pursued further.4–9 Lower doses of 5-azacitidine (75 mg/m2 per day for 7 days every 28 days) were investigated by the Cancer and Leukemia Group B (CALGB) in patients with MDS (myelodysplastic syndromes). Two phase II studies were reported, one using intravenous (IV) 5-azacitidine (CALGB 8421)10 and another using a subcutaneous (SC) formulation (CALGB 8921).11 Also, one phase III randomized study was conducted comparing SC 5-azacitidine with best supportive care (CALGB 9221).12 The results of these 3 trials have been updated using the World Health Organization (WHO) classification of myelodysplastic syndrome (MDS)/acute myeloid leukemia (AML) as well as the International Working Group (IWG) criteria for disease response (Table 1).13 In CALGB 8421, 48 patients (52% of whom were reclassified with a diagnosis of AML by WHO classification) were treated with IV 5-azacitidine. A complete response (CR), partial response (PR) and hematological improvement (HI) of 15%, 2% and 21%, respectively, were reported.10 The overall response rate (ORR) was 44%. Similar response rates were observed in CALGB 8921 in which 5-azacitidine was administered by SC route.14 </p><p>CALGB 9221 was a phase III randomized study of 191 patients with MDS. Patients were randomized to SC 5-azacitidine (n=99) or best supportive care (n=92).12 After 4 months, patients in the best supportive care arm were allowed to cross over to the 5-azacitidine arm following evidence of disease progression. In the 5-azacitidine arm, CR, PR and HI were documented in 10%, 1% and 36% of patients compared to no CR or PR in the supportive care only arm (n=41). For patients (n=51) who crossed over to the 5-azacitidine arm, CR, PR and HI were documented in 6%, 4% and 25% of patients.13 Combining the data from both SC CALGB trials (8921 and 9221), 169 patients were treated with 5-azacitidine with an ORR of 44% (CR 13%, PR 1%, HI 31%).13 5-azacitidine has also been evaluated by French investigators in patients with high risk MDS and secondary AML with an ORR of 62% (CR 16%, PR 25%, HI 21%).15 </p><!><p>A trend towards improved survival was noted in the CALGB 9221 trial for patients treated with 5-azacitidine but statistical significance could not be achieved due to the cross-over design of the study. In a landmark analysis at 6 months, median survival for the 5-azacitidine arm was 18 months compared to 11 months for the supportive care arm.12 This led to clinical trials focusing on the impact on survival of hypomethylating agents in MDS. At the 2007 American Society of Hematology (ASH) meeting, Fenaux and colleagues presented results of a phase III international multicenter randomized study comparing 5-azacitidine (n=179) to one of the three conventional care regimens (n=179) [(i)best supportive care (use of transfusions, antibiotics and G-CSF for neutropenic infections); (ii) low dose cytarabine (20mg/m2/day for 14 days every 28 days); (iii) conventional induction/consolidation chemotherapy] in patients with intermediate-2 or higher risk MDS.16 This trial did not allow erythropoietin. The use of 5-azacitidine led to significantly improved overall survival compared to the conventional care regimens (24.4 months vs. 15 months, p=0.0001). This is the first therapy to ever show an improvement in overall survival in MDS.16 The survival advantage seen with 5-azacitidine was irrespective of age, sex, WHO classification, karyotype, or international prognostic scoring system (IPSS) group. At 2 years, 51% of 5-azacitidine treated patients were alive, compared to 26% in the conventional care regimens.16 </p><!><p>5-azacitidine has been shown to reduce the risk of AML transformation in patients with MDS. In CALGB 9221, median time to AML or death was 21 months for 5-azacitidine group compared to 12 months for best supportive care (p=0.007).12 Similarly the preliminary results by Fenaux et al, in intermediate-2 and high risk MDS patients, showed a median time to AML or death of 13 months in the 5-azacitidine group vs. 7.6 months in the conventional care regimens.16 </p><!><p>Quality of life in patient with MDS is improved by treatment with 5-azacitidine. In CALGB 9221, patients on the 5-azacitidine arm experienced significantly greater improvement in fatigue, dyspnea, physical functioning, and psychological distress over the course of the study period than those in the supportive care arm.17 </p><!><p>In CALGB 9221, 65 of the 99 patients treated with 5-azacitidine were transfusion dependent at baseline. Of these, 29 patients (45%) became transfusion independent during the course of treatment for a median of 9 months, a significant improvement over the supportive care group.13,18 </p><!><p>Time to response is of particular importance given that several courses of 5-azacitidine may be required before any evidence of response is appreciated. Silverman et al investigated this issue in the CALGB studies. They found the median number of cycles with 5-azacitidine needed for any response was 3, with 90% of responders doing so by cycle 6 (range, 1–17 cycles).10,12–14 This supports the theory that the clinical effect of 5-azacitidine in this patient population is largely epigenetic rather than cytotoxic. It also underscores the importance of making patients aware of the relatively slow clinical outcome expected with this type of treatment. In addition, neutropenia should not delay treatment with hypomethylating agents, especially for the first 3 cycles.</p><p>Another important issue is the duration of treatment. While this has yet to be systematically studied, multiple investigators agree that therapy cessation is associated with earlier relapse. Indefinite treatment is therefore recommended.</p><!><p>Based on results from CALGB studies 8921 and 9221, 5-azacitidine was approved by the FDA on May 2004 for the treatment of MDS, in both low-risk and high-risk patients.19 The approved dose is 75 mg/m2/day SC for 7 days every 28 days. In January 2007, an IV formulation of 5-azacitidine was also approved by the FDA.20 The IV formulation boasts the advantage of being devoid of skin reactions often observed with the SC administration. Conventional dosing for 5-azacitidine is for 7 consecutive days. Because of the inconvenience of the weekend injections, a shorter 5-day schedule of 5-azacitidine has been studied. In the preliminary results by Lyons21, patients were randomized to 1 of 3 schedules of 5-azacitidine administered every 4 weeks: 5-2-2 (75 mg/m2/day × 5 days, followed by 2 days off and two additional days); 5-2-5 (50 mg/m2/day × 5 days, followed by 2 days off and 5 additional days); or 5 days only (75 mg/m2/day × 5 days).21 All three treatment regimens showed similar outcomes in terms of hematologic improvement, red blood cell transfusion independence, and safety profile. These results indicate that the 5 day schedule may be an alternative to 7 or 10 days schedules, although survival data is lacking with the 5-day schedule. In addition, an oral formulation of 5-azacitidine is being developed which, if found clinically effective, would provide a convenient alternative to the SC or IV routes of administration.22 </p><!><p>CALGB studies prohibited the use of growth factors such as G-CSF and erythropoietin in order to avoid confounding variables in the assessment of 5-azacitidine activity. Rossetti et al reported excellent response rates in a retrospective study of 5-azacitidine with concurrent use of G-CSF and erythropoietin.23 A prospective evaluation of this approach is warranted.</p><!><p>Decitabine is another nucleoside analogue, structurally related to 5-azacitidine. Decitabine is a derivative of 5-azacitidine and a more potent hypomethylating agent on a molar basis. Like 5-azacitidine, decitabine was initially explored as a cytotoxic agent at doses of 1500–2500 mg/m2 per course and was not pursued further due to prolonged myelosuppression.24,25 Zagonel et al explored two low-dose schedules of decitabine (45 mg/m2/day over 4 hr for 3 days and 50 mg/m2/day continuous daily infusion for 3 days) in 10 patients with MDS. ORR was 50% with 4 patients achieving CR. A phase II trial was initiated in Europe in the early 1990s exploring this low dose schedule of decitabine. Twenty-nine patients (MDS 20, AML 9) were treated with decitabine (50 mg/m2/d continuous intravenous infusion for 3 days every 6 weeks).26 Overall response was observed in 54% of patients (CR 29%, PR 18%, HI 7%) with a median response duration of 31 weeks. This trial led to a larger phase II trial by Wijermans and colleagues in which a slightly different dosing schedule was explored.27 Decitabine was given at 15 mg/m2 every 8hr (45 mg/m2/day) IV daily for 3 days every 6 weeks. ORR of 49% was observed with 20% CR. A pooled analysis of the phase II studies from Europe has also been published.28 In this analysis, a total of 177 patients received decitabine 40–50 mg/m2/day for 3 days every 6 weeks and a 49% ORR was reported (CR 24%, PR 10%, HI 14%). Median response duration was 36 weeks and median survival for the whole group was 15 months. These encouraging results led to a phase III randomized trial in the United States where patients were randomized to decitabine (15 mg/m2 every 8hr (n=89) IV daily for 3 days every 6 weeks) plus supportive care vs. supportive care alone (n=81).29 Of the patients on the decitabine arm, 69% had IPSS intermediate-2 or higher and 74% were transfusion dependent. ORR was 30% (CR 9%, PR 8%, HI 13%). Median time to response was 3.3 months. Patients treated with decitabine had a trend toward a longer median time to AML or death compared with patients treated with supportive care alone (12.1 mo vs. 7.8 mo, p=0.16). In the subgroup analysis, patients on the decitabine arm experienced a longer median time to AML or death than those who received supportive care, if they were treatment-naive (12.3 mo vs. 7.3 mo, p=0.08), had an IPSS score of intermediate-2/high risk (12.0 mo vs. 6.8 mo, p=0.03), were classified based on IPSS score as high risk (9.3 mo vs. 2.8 mo, p=0.01), or had de novo MDS (12.6 mo vs. 9.4 mo, p=0.04).29 A total of 43 patients (48%) received ≤2 cycles of decitabine which might have contributed to the low response rate in this phase III trial compared to previous reported phase II studies with decitabine.26–28 Based on this trial, the FDA approved decitabine for MDS (Intermediate-1 or higher IPSS class) in May, 2006.</p><p>Even lower doses of decitabine have been explored in the clinical setting, based on a number of observations. First, a low dose of decitabine (0.015 mg/kg daily for 10 days) was found to have biologic efficacy in reactivating hemoglobin F in patients with sickle cell disease.30 Also, the drug has a short half-life, and an absolute requirement for DNA synthesis for activity. In addition, a phase I trial was reported showing efficacy of lower doses of decitabine in hematopoietic malignancies.31 Kantarjian et al reported results of a phase II study in which the total dose of decitabine was decreased to 100 mg/m2/course (from 135 mg/m2/course previously reported), and the schedule was increased to every 4 weeks instead of every 6 weeks.32 Patients were randomized following a Bayesian adaptive design to one of three arms: (1) 20 mg/m2 IV over 1 hour daily for 5 days; (2) 20 mg/m2 daily given in two SC doses for 5 days, or (3) 10 mg/m2 IV daily for 10 days. Cycles were repeated every 4 weeks as long as there was evidence of residual marrow disease and no life-threatening complications. The ORR by the modified IWG criteria was 73%, with 32 patients (34%) achieving CR. Patients randomized to 20 mg/m2 daily for 5 days (the most dose-intensive arm) had the best response rate with 39% CR.32 The median overall survival time for the entire group was 19 months. Based on these results, a dose/schedule of 20mg/m2 IV daily for 5 days every 4 weeks is now considered the optimal schedule of decitabine. A phase II single arm study of decitabine (the ADOPT trial) was presented at the ASH 2007 conference, confirming the safety and activity of decitabine when used in this fashion.</p><!><p>An important clinical question is whether one of the two FDA approved hypomethylating agents for MDS is better than the other. There has been no head to head comparison of the two agents. The phase III studies of the two agents comparing each to supportive care showed an overall response rate of 47% for the 5-azacitidine trial compared to 30% for the decitabine trial, whereas the complete response rate was similar (10% vs. 9%).13,29 One important difference between the two phase III trials is that in the 5-azacitidine trial median time to response was 3 treatment courses,13 while in the decitabine trial, 48% of patients had ≤2 cycles of treatment.29 This might explain the low response rate with the latter. In addition, the optimal dosing of the two agents is still being established. Two survival studies, one with 5-azacitidine, the other with decitabine, have been completed. The first was presented at ASH 200716 (described above) and the second's results are expected at ASH 2008. As previously discussed, 5-azacitidine use is associated with improved survival. A comparative study of 5-azacitidine versus optimal dose/schedule of decitabine is necessary to establish any superiority of one agent over the other.</p><!><p>Like DNA methylation, acetylation of histone proteins plays an important role in gene transcription regulation. Two units each of the 4 core histones, H2A, H2B, H3 and H4 form the nucleosome around which DNA is wrapped.33 Core histones have NH2-terminal tails, which are lysine-rich and can undergo post-translational covalent modifications to effect gene expression. One such important modification is acetylation and deacetylation of key lysine residues of histone H3 and H4.34,35 Acetylation of histones leads to open chromatin configuration and gene transcription. On the other hand, deacetylation leads to a repressive state. These changes are mediated by histone acetyltransferases (HAT) and histone deacetylases (HDAC). The use of histone deacetylase inhibitors leads to a more permissive state and allows for gene expression. Several such agents are being explored in clinical studies (Table 3).</p><!><p>Sodium phenylbutyrate (PB) is an aromatic fatty acid compound which was initially developed for treatment of urea cycle disorders and thalassemia.36,37 Based on their preclinical study indicating a role of PB in differentiation and inhibition of growth of primary leukemia cells,38,39 Gore et al. explored the use of single agent PB in patients with MDS and AML. In the phase I trial, continuous 7 day infusion of PB was used (7/28 schedule: 7 days on, 21 days off schedule) in 11 patients with MDS and 16 patients with AML and a dose of 375 mg/kg/day was established as the maximally tolerated dose (MTD).40 Dose limiting toxicities (DLT) were neurological complications (lethargy and confusion) and were reversible within 24–48 hours of stopping the infusion. In a follow-up trial, PB at the MTD dose was used in 2 more frequent dose schedules (7/14 schedule: continuous 7 day infusion with 7 day rest, 21/28: continuous 21 day infusion with one week rest) in 23 patients with MDS and AML.41 In both these studies with single agent PB, no CR or PR was observed and less than 10% had HI. HDAC inhibition was not measured in these trials as these were initiated prior to the knowledge of HDAC inhibition by PB.38,42 </p><!><p>Valproic acid (VPA) is a short chain fatty acid used as an antiepileptic and mood stabilizer.43 VPA has been shown to affect the growth of malignant cells in vitro,44,45 to prolong the G1 phase of the cell cycle,46 and to have antiangiogenic activity in vitro.47 It is possible that the antineoplastic effects of VPA are related to HDAC inhibitory activity.48–50 </p><p>In their first trial, Kuendgen et al used VPA alone and in combination with all-trans retinoic acid (ATRA) in patients with MDS and secondary AML.51 Use of ATRA was based on in vitro studies indicating synergism between VPA and ATRA in cellular differentiation and apoptosis induction in leukemia cell lines.48,52,53 VPA was administered to reach target serum concentrations between 50–100 µg/ml and ATRA was given at 80 mg/m2 daily on a one-week-on one-week-off schedule. ATRA was planned for patients who did not respond to VPA or had relapsed after it. A total of 23 patients were treated (18 with VPA monotherapy, 5 with combination). In the VPA monotherapy group, 7 patients had hematological improvement and one patient had PR, for a response rate of 44% (8/18 patients) by IWG criteria. None of the 5 patients on combination therapy responded. Overall response rate was 35% for the entire study population and no patient achieved CR. All 3 patients with IPSS low risk MDS had a major hematological response whereas only 1 out of 4 patients with high risk MDS had a response (minor erythroid response). In this study, 3 of the 9 patients with elevated blast count had significant reduction in peripheral and bone marrow blast counts. This observation led to the expansion on this study to include more high risk MDS and AML patients. In the second report by Kuendgen, 75 patients were treated with a response rate of 30% in the MDS cohort (n=43) and 16% in the AML cohort (n=32).54 However, most of the responses were HI and only one CR (high risk MDS patient) and one PR (intermediate-1 MDS patient) were noted. Patients with low risk MDS had the best response rate at 70%. There was no correlation between dose of VPA, VPA blood levels, and response. The benefit of ATRA was seen in patients who had relapsed on VPA. Of the 10 relapsed patients treated with ATRA, 4 had second remission with a median duration of 21 months (better than the median first remission of 4 months, =0.01). In another report from the same trial, the authors concluded that VPA monotherapy is of minimal activity in AML and that the addition of ATRA to VPA was not of benefit.55 Several other groups have used the VPA and ATRA combinations either concomitantly or sequentially, in smaller number of patients with limited activity.56–59 </p><!><p>This agent has been shown to have significant clinical activity in cutaneous lymphoma60. Byrd et al. reported on the use of this agent in patients with AML (n=10) or CLL (n=10).61 Depsipeptide was given as an IV infusion at a dose of 13mg/m2 on days 1, 8, and 15 of a 28 day cycle. Nausea, fatigue and other constitutional symptoms were seen in the majority of patients. No CR or PR was observed in either CLL or AML cohorts and further development of use of this agent in AML and CLL was stopped by this group. This agent was tested at MD Anderson Cancer Center in combination with decitabine but significant cardiac toxicity was observed and the study was halted (Issa JP, personal communication).</p><!><p>MS-275 has been shown to exhibit time and dose dependent growth inhibition of leukemia cell lines as well as primary leukemia blasts.62–64 Gojo et al. conducted a phase 1 trial of MS-275 (a synthetic benzamide derivative) in patients with AML and MDS.65 Based on the study of MS-275 in solid tumors,66 a starting dose of 4 mg/m2 weekly for 2 or 4 consecutive weeks followed by 2-week washout was established. A total of 38 patients received the study drug at doses ranging from 4–8mg/m2 weekly. Fatigue, nausea and vomiting were frequent non-hematological side effects noted. The MTD (maximally tolerated dose) of MS-275 was established as 8 mg/m2 administered weekly for 4 weeks with 2 weeks washout. Responses were minimal with no patients achieving CR or PR. Three of the 34 evaluable patients were noted to have >50% reduction in bone marrow blasts. Histone acetylation was documented in all patients in either peripheral blood or bone marrow mononuclear cells. In addition, a synergistic action with fludarabine has been described in both AML and ALL cells67.</p><!><p>LBH 589 is a cinnamic hydroxamic acid analogue which has been shown to induce apoptosis and histone acetylation in acute leukemia cells. It is one of the most potent HDAC inhibitors currently available. Giles et al. treated 15 patients (13 AML, 1 MDS, 1 ALL) with LBH589 at doses ranging from 4.8 to 14.0 mg/m2 IV daily for 7 days every 21 days68. Reversible QTc prolongation was the DLT (dose limiting toxicity). The MTD (maximum tolerated dose) was not established, as 14mg/m2 exceeded it and the lower dose cohort of 11.5 mg/m2 could not be expanded given the concern for QTc prolongation. 27% of patients were also noted to have grade 3–4 hypokalemia, however no relation was noted between QTc prolongation and hypokalemia. Response was limited with no CR or PR. Reduction in peripheral blood blasts was seen in 8 of 11 patients whereas only 1 patient had reduction in bone marrow blasts.</p><!><p>MGCD0103 is an isotype specific aminophenylbenzamide and has been shown to inhibit HDAC isotypes 1,2,3, and 11.69 A dose escalation phase 1 study of oral MGCD0103 given 3 times a week in patients with AML and MDS has been conducted.70 Doses of MGCD0103 ranged from 20–80 mg/m2 orally. Twenty nine patients were treated (22 AML, 5 MDS, 1 ALL, 1 CML). Median age was 65 years and 83% of patients had received prior chemotherapy. Fatigue, nausea, diarrhea, vomiting were the most frequently reported adverse events. MTD was established at 60 mg/m2 orally three times a week. No CR or PR was observed. Bone marrow response was seen in 3 of the 23 evaluable patients. One-third of the patients at MTD had induction of histone H3 acetylation. This agent is now being studied in combination trials with 5-azacitidine71.</p><!><p>Vorinostat is a small molecule HDAC inhibitor, which has been approved by the FDA for treatment of skin manifestations of cutaneous T-cell lymphoma.72–74 Vorinostat has been shown to promote cell cycle arrest and apoptosis and has shown in vitro activity against leukemia cells.75–78 A phase 1 study of single agent vorinostat in patients with advanced leukemia and MDS was recently conducted.79 A total of 41 patients were treated (31 AML, 4 CLL, 3 MDS, 2 ALL, 1 CML) with a classical 3 + 3 dose escalation design. The starting dose was 100 mg orally three times daily for 2 weeks with 1 week washout. Both twice daily and three times daily regimens were tested. DLTs included fatigue, nausea, vomiting, and diarrhea. The MTD was established as 200mg BID or 250mg TID daily for 14 days every 21 days. Of the 41 patients, 2 patients achieved CR and 2 CRi. Additionally, 7 patients had HI (>50% decrease in blast count). Median number of cycles to response/improvement was 2 (range, 1–8) and median response duration was 6 weeks. Transient acetylation of histone H3 was noted in all patients, irrespective of the dose level or response.</p><!><p>Despite activity in MDS and AML, use of hypomethylating agents is associated with relatively low CR and PR rates (20–35%) in this patient population when used as monotherapy. In recently reported studies of single-agent HDAC inhibitors, even lower response rates have been observed. Thus there is a need to develop effective combination regimens aimed at improving response rates, duration and eventually survival in patients with AML and MDS. One such approach is the use of a hypomethylating agent with an HDAC inhibitor. The rationale for this combination emerged from studies demonstrating synergistic effects in the reactivation of epigenetically silenced genes.80–83 Gene silencing associated with methylation of promoters is in part due to recruitment of transcriptional repression complexes such HDACs. These results also indicated that DNA methylation was the dominant alteration and DNA methylation followed by histone deacetylation resulted in maximal gene reactivation. These studies are summarized in table 4.</p><!><p>Based on a preclinical study evaluating the antileukemia effect of decitabine and VPA,84 we evaluated the safety and activity of this combination in a phase 1/2 study.85 The preclinical evaluation showed the growth inhibitory activity of these drugs to be independent of the sequence used.84 For the clinical trials we selected a concomitant dosing schedule. Fifty four patients were treated (89% AML, 11% MDS)85 : 20% of the patients were previously untreated and all were older than 60 years of age. In the phase 1 portion of the study, 22 patients were treated. Three dose levels of VPA were studied (20, 35 and 50 mg/kg orally daily for 10days). Decitabine was given as 15 mg/m2 IV as a 1 hour infusion daily for 10 days. A dose of 50mg/kg daily of VPA was established as MTD and a total of 32 patients were treated in the phase 2 portion of the study. Nine of the 32 patients developed grade ≥3 non-hematological toxicity, mainly reversible neurotoxicity. CR was documented in 22% of patients (12 of 53 evaluable patients), including 4% with CRp (complete response with incomplete platelet recovery). Median time to response was 60 days (range 29–138 days) and median response duration was 5.6 months. Of note, induction mortality was extremely low at only 2%. Median survival was 15.3 months in responders and 4.9 months in non-responders. For previously untreated patients (n=10), 50% achieved CR + CRp. Higher VPA levels were associated with higher non-hematological toxicity, however no correlation between VPA levels and response was observed in whole group analysis. Interestingly, when only previously untreated patients were analyzed, responders did have significantly higher free VPA levels at day 10 compared to non-responders (32.4 mg/L vs. 14.0 mg/L, p= 0.03). Histone H3 and H4 acetylation was documented in 35% of patients at 50 mg/kg VPA dose level. No correlation was observed between histone acetylation or induction of global hypomethylation and response. Currently, a randomized trial comparing the use decitabine with and without VPA is underway.</p><!><p>Given the results of the above mentioned trials of VPA plus decitabine and VPA plus ATRA, the combination of VPA, 5-azacitidine and ATRA was investigated.86 The dose of 5-azacitidine was 75mg/m2 SC daily for 7 days (days 1–7) and ATRA was 45mg/m2 orally daily for 5 days (days 3–7). In the phase 1 portion of the study, 3 dose levels of VPA were tested (50, 62.5, 75 mg/kg orally daily for days 1–7). Nineteen patients were treated and a dose of 50 mg/kg orally daily for 7 days each course was established to be MTD for VPA. In the phase 2 part, 34 patients were treated at the MTD of VPA of which 8 had ≥3 grade nonhematologic toxicity (mainly reversible confusion and somnolence). Induction death was seen in 5% of patients. Overall 27% CR/CRp was observed (CR 22%, CRp 5%). In this study, bone marrow response (blasts <5% without meeting peripheral blood criteria for CR or CRp) was also documented and was seen in an additional 13% of patients. In the subgroup of previously untreated patients (n=33), 33% CR, 9% CRp and 9% bone marrow responses were noted. There was also no induction death observed in this subgroup. These were especially encouraging results, as all these patients were >60 years of age. In addition, 62% of the poor risk cytogenetics patients (-5,-7 chromosomal abnormality) responded (5/8 patients). Patients with increased VPA levels had a significantly increased rate of response (median bound VPA level on day 5 in responders was 132 µg/ml compared to 104 µg/ml in non responders, p<0.005), though there was no relation between VPA dose and response. Histone H3 and H4 acetylation was noted in 54% of patients, and a transient, global DNA hypomethylation was also seen. Again, there was no correlation between histone acetylation or DNA hypomethylation and clinical response.</p><p>An important observation from the 2 combination trials of VPA with 5-azacitidine or decitabine is that higher VPA levels are associated with higher response rates.85,86 This data has significant implications. First, if clinical responses are related to the HDAC inhibitory activity of this class of agents, substituting VPA with more potent HDAC inhibitors, such as vorinostat, MGCD0103 or LBH589, may result in effective combinations (below). Second, if clinicians are to use VPA, then higher doses (50 mg/kg/day, as done in these 2 studies) should be recommended.</p><!><p>Two studies have been reported so far combining 5-azacitidie and PB in patients with AML and MDS. In the first, Gore et al. used varying doses of 5-azacitidine (25–75 mg/m2/day SC) for 5–14 days followed by 7 days continuous infusion of PB at 375mg/kg/day.87 Of the 29 evaluable patients, 11 responded (5 CR, 1 PR, 5 HI). Reversible neurotoxicity secondary to PB occurred in 9 patients. In the second study Maslak et al. examined a fixed dose of 75 mg/m2 SC daily of 5-azacitidine for 7 days, followed by PB at 200 mg/kg/day in 1–2 hour infusion for 5 days.88 Of the 10 evaluable patients, 3 achieved PR. No CR was observed. Median time to response was 31 days and median duration of response was 45 days. Neurotoxicity related to PB (disorientation, somnolence) was seen in 80% of patients and was reversible upon cessation of therapy. Histone H4 acetylation was seen in all patients but there was no relation to clinical response. Neurotoxicity and the need for continuous infusion have hampered the development of PB as a therapeutic agent in myeloid malignancies.</p><!><p>Based on the single agent activity of both 5-azacitidine and MGCD0103, the combination has been used in 37 patients in a phase1/2 study.71 5-azacitidine was administered at 75mg/m2 SC daily for 7 days every 28 days. MGCD0103 was used in a dose escalation design with doses of 35, 60, 90, 110, and 135mg orally three times a week starting on day 5 of every cycle. In the preliminary analysis, ORR was 30% (CR 11%, CRi 14%, PR 5%). No induction mortality was observed. A randomized study is ready to be initiated comparing 5-azacitidine with or without MGCD0103.</p><!><p>Gore et al. recently reported preliminary results of a phase1/2 study of 5-azacitidine and MS-275.89 31 patients were treated with the combination of 5-azacitidine (dose 30, 40 or 50 mg/m2 SC daily for 10 days every 28 days) and MS-275 (dose 2, 4, 6 or 8 mg/m2 on days 3 and 10, every 28 days). Of the 27 evaluable patients, 2 achieved CR, 4 PR and 6 bilineage hematological improvements. An intergroup trial is currently comparing the combination of 5-azacitidine plus MS275 with 5-azacitidine alone.</p><p>Other combinations reported in preliminary fashion include vorinostat and decitabine90,91 and vorinostat and idarubicin92,93 in patients with MDS and AML.</p><!><p>The field of epigenetic therapy is expanding rapidly. A large number of drugs with the capacity to alter the epigenetic structure of cancer cells are being used in clinical trials either as single agents or in combinations. Hypomethylating agents such as 5-azacitidine and decitabine are now established, approved therapies for MDS. Several issues related to hypomethylating treatments such as the optimal dose, schedule or route of administration are still being elucidated in clinical trials. The use of HDAC inhibitors as single agents has proved to be of limited clinical efficacy. Ongoing randomized trials exploring the use of HDAC inhibitors in combination with hypomethylation as well as other therapeutic agents will help elucidate the role of these agents in the treatment of MDS and leukemia.</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

The Balance of Beneficial and Deleterious Health Effects of Quinones: A Case Study of the Chemical Properties of Genistein and Estrone Quinones

Substances containing a phenolic moiety are often metabolized to quinones whose high reactivity makes them difficult to study. Some of these precursors have clear health benefits, and some quinones themselves are used in cancer therapy, whereas others are deleterious. For example, dietary intake of phytoestrogen, genistein (Gen), seems to play a preventive role in breast cancer (BC) whereas prolonged exposure to chemically similar mammalian estrogens is clearly associated with elevated incidence of BC. Although both can be metabolized to reactive quinones, the catechol estrogen quinones (CEQs) modify DNA by redox cycling and/or depurination via a Michael addition. Here, we report an investigation of the chemical reactivity of Gen and estrone quinones to determine the chemical differences in of these two biologically important molecules. The catechol genistein quinone (CGenQ), has a half life of 4 \xc2\xb1 1 s under physiological condition, as determined by glutathione trapping. It disappears by reacting with H2O to give a dihydrate, CGenQ\xc2\xb7(H2O)2, whose structure was proved by NMR. Under reductive conditions, CGenQ\xc2\xb7(H2O)2 is readily reduced to reform the catechol genistein (CGen). This reversible oxidation of CGen to CGenQ and the prompt moderation of its reactivity by hydration to CGenQ\xc2\xb7(H2O)2 effectively moderates any redox cycling or depurination reaction of CGenQ with DNA. Catechol estrogen quinones, on the other hand, are sufficiently long-lived that they can damage DNA via a Michael addition or by redox cycling. Although the reactivity of CEQ in a nonaqueous solvent is similar to that of CGenQ, its reactivity in aqueous media with the free Ade base is more than 600 times that of CGenQ. These results suggest that rapid hydration of a quinone can moderate its reactivity toward biomolecules, allowing them to express, for example, estrogen-like properties without exhibiting the deleterious properties of redox cycling or directly damaging DNA via depurination reactions.

the_balance_of_beneficial_and_deleterious_health_effects_of_quinones:_a_case_study_of_the_chemical_p

Introduction<!>Materials<!>Preparation of Catechol Quinones<!>Preparation of Catechol Adenine Adducts<!>Kinetic Experiments with GSH Trapping<!>Preparation of Catechol from Catechol Quinones<!>Liquid Chromatographic Mass Spectrometric (LC/MS) Analysis<!>NMR Analysis<!>(1) Formation and Stability of CGenQ or CE1Q<!>(2) Half Life of Catechol Genistein Quinone in water<!>(3) Products of the reaction of Quinones in H2O<!>(4\xe2\x80\x931) Formation of Ade Adducts in Aqueous solution<!>(4\xe2\x80\x932) Formation of Ade Adducts in DMF<!>(4\xe2\x80\x933) Formation of DNA Adducts with Calf Thymus DNA<!>(5) Metabolism of Genistein and its Quinones<!>Conclusions<!>Supporting Information Available<!>

<p>Quinones are commonly produced by oxidation of many natural products (e.g., estrogens, phytoestrogens) and related compounds. The usual substrates, an aromatic compound containing oxygen substituents (e.g. phenols, hydroquinones, catechols), become quinones by the action of monooxygenase or peroxidase.1 Among the quinones to which humans are exposed are those from (1) benzene and poly-aromatic hydrocarbons (PAH),2–9 (2) antioxidants from dietary intake of soy foods, red wine, etc,10–14 (3) xenobiotic drugs for anti-tumor and anti-biotic therapy15–17, and (4) xenoestrogens and endogenous estrogens used in hormone replacement therapy.18–20 Quinones are often toxic, eliciting a variety of cytotoxic and genotoxic effects in vivo.21 For example, quinones are redox-active, forming reactive oxygen species (ROS) via redox cycling and causing oxidative damage to DNA and proteins.22–24 Moreover, quinones are electrophiles and can effect serious cell damage via direct covalent modification of DNA and/or proteins.12, 24–26 To illustrate ROS generation and direct DNA modification, we show in Scheme 1 example reactions involving phenol, catechol, and catechol quinone.</p><p>Owing to widespread human exposure to quinones and their important implications in carcinogenesis, we hypothesize that the chemical properties of quinone species are a starting point to understand their biological properties. Our assessment makes use of two model quinones, one generated from the phytoestrogen, genistein (Gen), and the other from the mammalian estrogen, estrone (E1). Estrogens, E1 and 17β-estradiol (E2), are potent sex hormones in mammals, whereas Gen is an important member of the isoflavone family and a phytoestrogen, for which there is a some evidence that dietary intake moderates the risk of hormone-related cancers.27 These materials exert biological effects via binding to estrogen receptors (ERs).28–30 Although both have a phenol substructure that can be oxidized to a catechol and ultimately to a potentially toxic catechol quinone31–35, they do exhibit some significant differences in their biological properties. Both Gen and estrogens can serve as antioxidants that scavenge free radicals36–45 presumably by donating hydrogen atoms from the phenolic hydroxyl groups, forming ArO• radicals.46 Although the catechols are good antioxidants47, owing to stabilization of the catecholic ArO• radical by hydrogen bonding with the adjacent hydroxyl group46, the ArO• radical is also a semiquinone radical (see Scheme 1) that can be reduced back to catechol quinones, producing ROS and causing oxidative damage to DNA and proteins.</p><p>From the perspective of human health, the balance of chemopreventive and chemopromotive effects of quinones is highly variable, depending on chemical structure. The quinones of Gen and estrogens are appropriate examples. Epidemiologic and experimental evidence reveals that prolonged exposure to estrogens is associated with elevated incidence of breast cancer (BC), leading to significant concerns about hormone replacement therapy.48–50 On the other hand, dietary intake of Gen isoflavones from soy is a useful source of antioxidants and may be linked with the lower BC incidence in Asian women. Nevertheless, its protective effects have not been conclusively established.51–56</p><p>Although there is a clear association between estrogen exposure and BC risk, the mechanism whereby estrogens contribute to BC remains to be firmly established. One view is that ER-mediated cell proliferation may account for the correlation.57–62 Another is that a competitive mechanism focused on the genotoxicity of estrogen quinones may actually initiate BC. Extensive cell-culture and animal studies support estrogen-induced mutations through either removal of DNA bases25, 63–68 or generation of ROS via their endogenous catechol estrogen quinone metabolites,21, 26, 58, 64, 69, 70. According to the genotoxic mechanism, endogenous estrogens are oxidized to catechol estrogens (CE) to form either 2-hydroxyestrogen (2-OH-E) or 4-hydroxyestrogen (4-OH-E) via phase I metabolism mediated by P450 enzymes. 2-OH-E and 4-OH-E are subsequently oxidized to catechol estrogen-2,3-quinone (CE-2,3-Q) and 3,4-quinone (CE-3,4-Q), respectively. These CEQs react with nucleophilic adenine (Ade) and/or guanine (Gua) bases by a 1,4-Michael addition, often leading to rapid loss of the purine by glycosyl-bond cleavage and leaving an abasic site in the DNA. Although these sites can be repaired enzymatically, an overabundance or error-prone repair may lead to DNA mutations and ultimately cancer. Do similar oxidative processes occur with Gen, forming catechol genistein (CGen) and ultimately genistein quinones (CGenQ)? As of yet there are no answers to this question.</p><p>The stability of the various quinones may explain the balance of salutary and deleterious effects of the oxidation products of Gen and estrone. Our specific hypothesis is that longer-lived, more stable quinones have more time to participate in both direct and indirect DNA damage whereas more reactive quinones react rapidly with solvent. Supporting this idea is the greater half life of the more genotoxic CE-3,4-Q compared to the 2,3-isomer under physiological conditions.71 To test further the hypothesis, we investigated the reactivity of both CGenQ and CE1Qs, employing nuclear magnetic resonance (NMR) and liquid-chromatography-electrospray ionization mass spectrometry (LC/MS). Using the approach of glutathione trapping, first proposed by Bolton and coworkers71, we determined the half life of the quinone from genistein, characterized the products of both its reaction in H2O and its subsequent reduction, and conducted analogous analyses for the quinones of estrone. We also obtained data that allow comparisons of the reactivity of both quinones in reaction with the Ade free base, under aqueous and nonaqueous conditions, and with calf thymus DNA. We are intrigued by the prospect that the deleterious effect of certain quinones are counterbalanced by the salutary effects of others, possibly promoting a "fight fire with fire" strategy whereby one quinone precursor may serve as a sacrificial agent to lower the toxicity of quinones from another.</p><!><p>Oxone, 2-iodobenzoic acid (IBA), adenine (Ade), disodium hydrogen phosphate (Na2HPO4), glutathione (GSH), dimethyl formamide (DMF), trifluoacetic acid (TFA, LC/MS grade), formic acid (FA, LC/MS grade), ascorbic acid, Omnisolv acetonitrile (ACN), Omnisolv methanol (MeOH) and ethanol (EtOH) were purchased from Sigma-Aldrich (St. Louis, MO). E1 was purchased from Steraloids (Newport, RI). Gen was purchased from Acros (Morris Plains, NJ). D2O, deuterated d7-DMF and uniformly 15N-labeled (U-15N) adenine 5′-monophosphate (AMP) were from Cambridge Isotope Laboratory (Andover, MA). Calf thymus DNA was purchased from USB Corp (Cleveland, OH). Focus solid-phase extraction (SPE) cartridges (50 mg sorbent, 6 mL load capacity) were from Varian Inc (Palo Alto, CA). Milli-Q water was used for LC/MS and SPE.</p><p>The oxidant, 2-iodoxybenzoic acid (IBX), was synthesized from IBA and oxone, as previously described.72 The U-15N-labelled internal standard (IS), 4-hydroxyestrone-1-N3(U-15N)Ade (4-OH-E1-1-N3(U-15N)Ade), was synthesized and characterized, as reported recently.73 Deuterated d7-DMF was used in NMR experiments.</p><!><p>To form the catechol quinone product of Gen, IBX (1 mg) and Gen (1 mg) were mixed in 1 mL of DMF and allowed to react for 30 min. Catechol genistein-3′,4′-quinone (CGen-3′,4′-Q) was obtained and was used as such. An identical procedure was used to prepare simultaneously two catechol estrone quinones: catechol estrone-2,3-quinone (CE1-2,3-Q) and catechol estrone-3,4-quinone (CE1-3,4-Q) (see Scheme 2 for the structures of Gen, E1, CGen-3′,4′-Q, CE1-2,3-Q and CE1-3,4-Q). Given that CGen-3′,4′-Q is the only quinone species derived from Gen, it will be abbreviated simply as CGenQ in this paper.</p><!><p>To explore the quinone chemistry, a general procedure of synthesizing the quinones and studying their reactivity with adenine was adopted (Scheme 2). (1) To prepare catechol adenine adducts in the absence of H2O, freshly prepared quinones (1 mg Gen or E1 + 1 mg IBX in 1 mL DMF) were mixed with Ade free base (1 mg) and reacted for 3 h. (2) To prepare adducts in predominantly aqueous media, 50 μL of DMF containing freshly prepared quinones was replaced with 100 μL of ACN. Five μL of the quinone(s) in ACN were added to 1 mL of a Na2HPO4 (50 mM, pH 7.4) solution containing Ade free base (1 mg) and incubated at 37 °C overnight. (3) To measure the reactivity with double-stranded DNA, the Ade free base in procedure (2) was replaced with calf thymus DNA (1 mg).</p><p>The products obtained from the reaction with the Ade free base were diluted by 500 fold with H2O/ACN/FA (93.5/6.5/0.1%) in preparation for LC/MS analysis. Cleanup steps were carried out for products obtained from the reaction with calf thymus DNA. To precipitate DNA, 2.5 volumes of EtOH were added, and the solution kept overnight at −20 °C. The suspension was centrifuged at 14 k/rpm, 4 °C for 10 min, the supernatant isolated, and the solvent removed under reduced pressure. The analyte was reconstituted in 3 mL of H2O/MeOH (2/1, v/v). The U-15N-labeled internal standard, 4-OH-E1-1-N3(U-15N)Ade (1 pmole), was added, and the solution submitted to SPE cleanup. The SPE cartridge was conditioned with 5 mL MeOH followed by 5 mL H2O. The analyte was loaded and washed with 5 mL H2O followed by 5 mL 10% ACN. The DNA adducts were eluted with 4 × 1 mL MeOH/ACN/H2O/TFA (30/60/10/0.1%). The eluent solvent was removed under reduced pressure, reconstituted in 20 μL of H2O/ACN/FA (93.5/6.5/0.1%) in preparation for LC/MS analysis. Samples in 5 μL were loaded for LC/MS analysis for analysis of the various products.</p><!><p>Freshly prepared quinones from IBX + Gen (or E1) (500 μg in 0.50 mL DMF for 30 min) were added to 9.5 mL Na2HPO4 buffer at 37 °C. Aliquots of 200 μL was removed at various intervals and combined with 200 μL GSH (10 mM in 50 mM Na2HPO4 buffer) at 37 °C. The aliquots were diluted by 100 fold with H2O/ACN/FA (93.5/6.5/0.1%) and analyzed by LC/MS. Gen was used as is an internal standard to quantify the levels of catechol quinones trapped with GSH.</p><!><p>To obtain catechols from corresponding catechol quinones, five stoichiometric amounts of ascorbic acid were added to freshly prepared quinones in DMF.</p><!><p>Microflow LC/MS and LC tandem MS (LC/MS/MS) were conducted on an LCQ Deca quadrupole ion-trap mass spectrometer (Thermo-Fisher, San Jose, CA) coupled to a capillary LC (Waters Corp., Milford, MA). An uncoated 360/75 μm OD/ID fused-silica capillary column with 15 μm PicoFrit™ nano tip (New Objective Inc., Woburn, MA) was custom-packed with C18(2) particles (Luna 3 μm 100 Å, Phenomenex, Torrance, CA) to afford a column that was ~ 12 cm in length. The sample solution was loaded and analyzed with solvent A (97% H2O/3% ACN/0.1% FA) and B (3% H2O/97% ACN/0.1% FA) starting with 97% A for 8 min and then a 37 min linear gradient to 100% B. The eluent with a total flow rate of 8 μL/min was split before the column to achieve a flow rate of 270 nL/min at the tip.</p><p>Sample ions were introduced with electrospray ionization (ESI) into the mass spectrometer operated in the positive-ion mode. The spray voltage was 1.8–2.5 kV, and no nitrogen sheath gas was used. The maximum time for injection of ions into the trap was 150 ms, and the number of microscans was one. The mass range for the mass spectrometer scan was m/z 100 to 1000 unless otherwise noted. For MS2 experiments, the collision energy was optimized at 40% of the maximum energy available (~ 5 eV).</p><p>When accurate mass measurements were needed, LC/MS was carried out with an ion-trap/FT-ICR mass spectrometer (LTQ-FT, Thermo Fisher, San Jose, CA). A resolving power of the FT component was 100,000 (at m/z 400), and the mass range for scanning was from m/z 100–1000.</p><!><p>1H NMR measurements were performed on an INOVA-500 (Varian, Inc., Palo Alto, CA) at 25 °C. The 1H NMR assignments were established by Correlation Spectroscopy (COSY), gradient Heteronuclear Multiple Quantum Coherence (gHMQC) and gradient Heteronuclear Multiple Bond Coherence (gHMBC). 1H and 13C results were referenced to d7-DMF. The NMR characterization of CGenQ was performed after 30 min of reaction of 1 mg Gen and 1 mg IBX in 0.5 mL d7-DMF. The two quinone products (CE1-2,3-Q and CE1-3,4-Q) were prepared similarly. For the characterization of CGenQ·(H2O)2 adduct, 1.6 mg Gen and 3.2 mg IBX were reacted in 0.4 mL d7-DMF for 3 h, 0.2 mL D2O was added, and the solution vortexed. For the kinetic measurements for the formation of CGenQ or CE1Q, 1H NMR spectra were collected under quantitative conditions by using 8 transients, a 70° flip angle pulse width and a spectral width of 4820 Hz. The acquisition started after mixing 1 mg Gen (or E1) and 1 mg IBX in 1 mL d7-DMF and was repeated at various intervals at constant temperature (25 °C).</p><!><p>The product formed in the IBX oxidation of genistein (Gen) in DMF is strikingly simple (Table 1 summarizes the 1H and 13C chemical shifts for the product). The oxidation occurs at the 3′ and 4′ positions of ring B, resulting in the formation of CGen-3′,4′-Q (Scheme 2A). No detectable oxidation of the A ring occurs. Given the deleterious effects of long-term exposure to estrogens, we reinvestigated the formation of CE1Qs, using IBX oxidation to prepare the materials.74 The products of the reaction are CE1-2,3-Q and CE1-3,4-Q at a ratio of 60:40, according to NMR (Scheme 2B, see Figure S1 in Supporting Information for the NMR data). Both classes of quinones, CGenQ and CE1Qs, are relatively stable in DMF. The disappearance of both Gen and E1 quinones follow 2nd-order kinetics (Figure S2 in Supporting Information) with rate constants of 3.3×10−3 and 6.4×10−3 M−1 s−1, respectively. The smaller rate constant for the Gen-quinone reaction accounts for the lower amount of CGenQ, approximately 40% of the total CE1Qs (CE1-2,3-Q + CE1-3,4-Q), after 30 min of reaction in DMF.</p><!><p>Unlike in DMF, quinones are labile in water under physiological conditions. Bolton and coworkers75 demonstrated that glutathione (GSH) trapping can be used for measuring the half lives of quinones; specifically, CE-2,3-Q, CE-3,4-Q71 and raloxifene, in H2O. GSH is a strong nucleophile that reacts with electrophilic catechol quinones via a Michael addition to form covalently bound adduct(s). We adopted this GSH-trapping protocol to determine the half life of CGenQ; GSH binds to CGenQ to form predominantly two mono GSH adducts. Analogous to the reaction of estrogens, the binding sites are the 2′ and 5′ position of CGenQ, giving 3′-hydroxygenistein-2′-SG (3′-OH-Gen-2′-SG) and 3′-hydroxygenistein-5′-SG (3′-OH-Gen-5′-SG). In addition to the mono GSH adducts, a small fraction of CGenQ reacts to add two GSH molecules whereby the second GSH molecule binds to the 2′ or 5′ position of CGenQ to afford 3′-hydroxygenistein-2′,5′-diSG (3′-OH-Gen-2′,5′-diSG). Simple addition occurs to conjugate the quinones in the first step of reaction with GSH, yielding presumably 3′-OH-Gen-2′-SG and 3′-OH-Gen-5′-SG of M.W. 591, as determined by ESI MS. The 3′-OH-Gen-2′,5′-diSG conjugate has a molecular weight of 896, two mass units less than would expected from a simple addition of a second GSH. To add a second GSH, we propose that 3′-OH-Gen-2′-SG and 3′-OH-Gen-5′-SG are re-oxidized to catechol quinones, CGen-3′,4′-Q-2′-SG and CGen-3′,4′-Q-5′-SG (M.W. 589), respectively. We confirmed the formation of these MW-589 species with LC/MS analyses, supporting that conjugation of these intermediate quinones with a second GSH to afford 3′-OH-Gen-2′,5′-diSG.</p><p>To measure the half life of CGen-3′,4′-Q, we followed the time course for 3′-OH-Gen-2′-SG and 3′-OH-Gen-5′-SG by using their mass spectral signal intensities, summing and normalizing them to 100% at t = 0. Diluting the final CGenQ/GSH mixture by 100 fold prior to analysis prevented additional oxidation of residual Gen by IBX during the LC/MS analysis. To quantify the GSH adducts, the signal for unreacted Gen was used as an IS (see Figure 1 for the reaction of CGenQ in H2O trapped by GSH). The data, when fit with a single exponential function, reveal that the half life of CGenQ is 4 ± 1 s at 37 °C, pH 7.4. In contrast, the half lives of CE1-2,3-Q and CE1-3,4-Q are considerably longer at 42 s and 730 s (12.2 min), respectively.71</p><!><p>To characterize the decay product(s) of CGenQ in H2O, we mixed d7-DMF containing freshly prepared CGenQ with D2O (v/v, 2:1) and monitored the products by NMR (for a summary of the 1H and 13C chemical shifts for the product of CGenQ reacting with water, see Table 2). The NMR results are consistent with CGenQ being converted to a dihydrate CGenQ·(H2O)2 (see Scheme 3 for structures). CGenQ·(H2O)2 is stable in H2O, and no additional reaction of the dihydrate occurs, even for a period of 48 h. A mechanism for converting CGenQ to CGenQ·(H2O)2 (Scheme 3) shows that the O1 ether and the C2-C3 double bond in the C ring of genistein provide a driving force for facile protonation of CGenQ, whose conjugate acid is stabilized as an intermediate oxonium ion with a high degree of conjugation. The oxonium ion undergoes subsequent hydration to CGenQ·H2O. This hydration reaction and the overall stabilization it provides may be characteristics of many quinones that are formed from isoflavones and other health-promoting materials containing phenolic moieties.</p><p>We also investigated the hydration of CGenQ by using ESI and accurate-mass analysis (LTQ-FT) (see Figure 2 for the LC trace of the products generated from the reaction of CGenQ in H2O). CGenQ reacts rapidly in H2O, making the detection of unhydrated CGenQ difficult by LC/MS. CGenQ·(H2O)2 is expected to lose water molecule(s) readily after protonation as the dihydrate carries several aliphatic hydroxyl groups. The dominant product is seen, however, as protonated CGen of m/z 287, as confirmed by accurate mass measurements. At first thought, this is inconsistent with the NMR results, which indicate formation of the dihydrate (expected m/z 321 for the protonated species, [CGenQ·(H2O)2 + H]+). To characterize further the species of m/z 287 seen by MS, we prepared 3′-OH-Gen by the reduction of CGen-3′,4′-Q in DMF. Mixing an equal amount of the synthetic 3′-OH-Gen with the unknown and analyzing again by LC/MS show that the unknown co-elutes with synthetic 3′-OH-Gen, and both are seen as ions of m/z 287 in the mass spectrum. The product-ion mass spectra (MS/MS) of the [M + H]+ of the two compounds are nearly identical (see Table S1 in Supporting Information), establishing that the material arising from the dihydrate during the LC/MS analysis is 3′-OH-Gen.</p><p>We can rationalize the results from MS and NMR by invoking reduction of the CGenQ·(H2O)2 to CGen in the LC/MS solution. The CGenQ·(H2O)2 presumably has a high reduction potential, allowing ready conversion back to CGen with H2O release. The likely reducing agent is formic acid (FA), which we add as a modifier for the LC/MS analysis. We are unable to test easily this proposal because, without the addition of acid in the LC experiment, we are unable to detect any products. The presence of FA, even an unintended amount, leads to rapid reduction of CGenQ·(H2O)2 to CGen.</p><p>During the trapping experiment, GSH reacts with CGenQ to form exclusively GenQ-GSH conjugates at t = 0; no CGen is detectable. The observation reveals that, in reaction with GSH, conjugation is preferred over reduction of the GenQ. CGen, however, forms later, as seen in the LC/MS of GenQ trapped by GSH after incubating GenQ in H2O. We propose that GSH cannot form a conjugate with CGenQ·(H2O)2; rather the reduction by GSH to form CGen becomes the dominant path. This facile conversion obviates detection of any major amount of CGenQ·(H2O)2 in the LC/MS analysis. We do clearly see, however, a residual amount of mono-hydrated CGenQ, CGenQ·H2O, as shown in the reconstructed ion chromatogram in the inset of Figure 2. The window for mass selection was centered at the theoretical mass of [CGenQ·H2O + H]+ with a width of ± 0.001 mass units (m/z 303.0499 ± 0.0010). The reconstructed ion chromatogram shows that the monohydrate, CGenQ·H2O, elutes at 24.99 min whereas CGen elutes at 25.14 min. Although the retention time difference is small, it is reproducible in replicates done even over many days, indicating that that the hydration is not caused by the ESI process. The difference indicates that a more hydrophilic species is formed by the addition of H2O to CGenQ. The mono-hydrate, CGenQ·H2O, may be a product of in-source fragmentation of CGenQ·(H2O)2 or an intermediate in the di-hydration reaction (Scheme 3). In negative-ion ESI, where formic acid can be avoided and the water lose attenuated, accurate mass measurements show that both the dihydrate CGenQ·(H2O)2 and monohydrate CGenQ·H2O can be observed at m/z 319.0458 (~ 3 ppm) and 301.0352 (~ 3 ppm), respectively. The abundance ratio of the two ions is 1:4, respectively.</p><p>For the estrone quinones, NMR analysis indicates that CE1-2,3-Q is formed in higher abundance than CE1-3,4-Q. Owing to its faster reaction in H2O, CE1-2,3-Q is of lower abundance as seen by LC/MS analysis (Figure 3). In addition to the two quinones, there are two other species whose [M + H]+ ions are of m/z 285 and that elute at 22.04 and 23.04 min, respectively, both of which have the formulae of CE1Q as established by accurate mass measurements. These new species are probably quinone methides (QMs); a previous study revealed that CE1Qs isomerize to their QMs via a process assisted by H2O.26, 71, 76 Interestingly, we found no QMs in the reactions of CGenQ. A plausible explanation is that the presence of O1 and the absence of OH at C2 in CGenQ prevent the formation of a QM.</p><p>Analogous to the products from Gen, we also see two species of m/z 287 in the LC/MS analysis of the reaction products of CE1Q. They coelute with authentic 2,3-OH-E1 and 3,4-OH-E1, supporting their assignments as catechol estrones. The product pattern that results from the reaction of CE1Qs is more complicated than for the genistein quinone. Analogous to the reactions of the latter species, estrone quinones also produce a monohydrate, CE1Q·H2O (see inset in Figure 3 for the accurate-mass selective ion chromatogram for CE1Q·H2O). To achieve high specificity, we set the window for mass selection at the theoretical mass of [CE1Q·H2O + H]+ with a variability of ± 0.001 mass units (i.e., m/z 303.1591 ± 0.0010). Two different hydrates elute, presumably CE1-2,3-Q·H2O and CE1-3,4-Q·H2O.71 Furthermore, two new species elute at 24.70 and 25.16 min, respectively. Accurate mass measurements reveal that each has the formula [CE1Q + O] (m/z 301 for a protonated species). Given that the further oxidation of CE1Qs is not of major interest here, we did not investigate them further.</p><!><p>The rapid reaction of the genistein quinone in water suggests that the probability for the quinone to react with DNA bases is low. To test, we compared the ratio of modified Ade species formed from the genistein quinone reacting with Ade to those of estrone quinones also reacting with Ade. We added CGenQ or CE1Q in ACN (5 μL) to 1 mL of Na2HPO4 solution (50 mM, pH 7.4) containing the Ade free base and incubated the solution overnight at 37 °C. Any Ade adducts, formed under these identical conditions, were combined in a 50:1 (Gen:E1) ratio, in anticipation that the diminished reactivity of CGenQ with respect to CE1Q would make difficult a quantitative measurement over a large concentration range, and the solution was diluted with 50 mM Na2HPO4. This strategy allowed us to measure accurately the relative amount of Ade modified by catechol genistein quinone (CGen-Ade) compared to the amount of Ade modified by the estrone quinones (CE1-Ade). An LC/MS trace (Figure 4) shows that the Gen-modified Ade species elute from 16.8–18.2 min, whereas CE1-Ade species elute from 18.5–21.2 min. A reconstructed ion chromatogram shows the elution of CGen-Ade (m/z 420.1) and CE1-Ade adducts (m/z 420.2); we confirmed the assignments by accurate mass measurements. Product-ion mass spectra (MS/MS) provide additional support for these adduct assignments. One of the characteristic fragmentations of the CGen-Ade adducts is a through-ring cleavage in ring C to generate product ions of m/z 153 and/or a complementary ion of m/z 268 (see Figure S3 in Supporting Information). Putative structures of the CGen-Ade adducts, along with their major cleavages induced by MS/MS are given in Figure S4 in Supporting Information.</p><p>In accord with the half lives of quinones, the high propensity of CGenQ to undergo hydration diminishes its reactivity with the nucleobase such that only trace amounts of CGen-Ade adducts can form compared to CE1-Ade adducts. Considering that 50 times more CGenQ-Ade adducts were injected, we found that integration of peak area in LC shows that the total CE1-Ade adducts are 620 ± 60 times more abundant than the CGen-Ade adducts.</p><p>The most abundant CE1-Ade adduct, eluting at 18.95 min, is probably weakly bound and dissociates rapidly during electrospray to form the fragment ion of m/z 136, [Ade + H]+, and its complement, the ion of m/z 285, [CE1Q + H]+. In addition to the weakly-bound species, there are two major and more stable CE1-Ade adducts that elute at 18.51 min and 21.16 min. The early eluter is the well characterized 4-OH-E1-1-N3Ade (see Figure 5, bottom panel, for the product-ion mass spectrum taken in an MS/MS experiment), which is a putative biomarker for breast cancer.77, 78 The nearly equally abundant CE1-Ade eluting at 21.16 min gives an unusual product-ion mass spectrum (top panel in Figure 5). Missing in this spectrum are ions formed by the losses of NH3 and CH2NH from [4-OH-E1-1-N3Ade + H]+. These ions are formed by migration of H-atom(s) from the estrone to the Ade moiety73 and are characteristic of a nucleobase that adds to the C1 position of the A ring of the steroid. The unknown CE1-Ade adduct likely has a structure in which the migration of H atom(s) to Ade is prohibited. A full structural study is beyond the scope of this article.</p><!><p>Both CGenQ and CE1Qs, however, are relatively stable in DMF. In this nonaqueous solvent, the reactivities of the quinones are more "intrinsic" (not significantly determined by the properties of the solvent). In this medium, the ratio of CGen-Ade adducts to CE1-Ade adducts increases remarkably to 1:3 (Figure S5 in Supporting Information, also see Table S2 for accurate mass measurements for CGen-Ade adducts). Given that the amount of CGenQ is ~ 40% of the total CE1Qs, we conclude that the "intrinsic reactivities" to modify Ade are similar for the two quinones. The significantly shorter half life of CGenQ in H2O accounts for its greatly reduced reactivity with Ade in aqueous solution (i.e., under physiological conditions).</p><p>MS/MS product-ion analyses reveal that the two CGen-Ade adducts formed in DMF and eluting at 17.95 and 19.38 min are the same species formed in aqueous solution and eluting at 16.82 and 18.21 min, respectively, as determined at a different time. Another product eluting at 18.38 min in the reaction with DMF, however, is barely detectable in the reaction in H2O. Clearly, the solvent affects the formation not only of the CGen-Ade adducts but also the CE1-Ade adducts. The species eluting at 21.22 and 21.32 min are 4-OH-E1-1-N3Ade and the weakly-bound CE1-Ade, respectively. Unlike in H2O, where the weakly bound species is ~ 4 times as abundant as 4-OH-E1-1-N3Ade, these two species are produced in DMF in nearly equal amounts. The species produced in DMF and eluting at 23.67 min corresponds to that eluting at 21.16 min from the reaction in water, but its abundance is considerably less than 4-OH-E1-1-N3Ade when formed in DMF.</p><!><p>We extended the reactivity comparison of the two quinones by carrying out a competitive reaction with calf thymus DNA. The reaction involved approximately 60 nmole of CE1-2,3-Q and 40 nmole of CE1-3,4-Q in 5 μL ACN, to which we added 1 mg calf thymus DNA in 1 mL of H2O (50 mM Na2HPO4, pH 7.4). After overnight incubation at 37 °C, we added the internal standard, U-15N labeled IS, 4-OH-E1-1-N3(U-15N)Ade (1 pmole), to assure good quantification and submitted the mixture to SPE cleanup. The total ion chromatogram (TIC) obtained by monitoring the elution of the IS, 4-OH-E1-1-N3(U-15N)Ade (tR = 19.57 min), which was obtained in a LC/MS analysis (see the LC traces in Figure 6), allows the identification of one of the CE1-Ade adducts and allows for quantification of all of them. A TIC produced by integrating the product-ion intensities generated by MS/MS of the m/z 420.2 [M + H]+ ions show two species that elute at tR of 19.55 and 21.68 min. One coelutes with the U-15N labeled 4-OH-E1-1-N3Ade, indicating that the species at tR of 19.55 min is 4-OH-E1-1-N3Ade. Comparing the MS/MS product-ion spectra (see Figure S6 in Supporting Information) with those of the CE1-Ade adducts formed from the reaction with Ade free base in H2O, we conclude that the two species formed from the reaction with calf thymus DNA correspond to those eluting at tR 18.51 and 21.16 min, respectively, formed in the reaction in water (Figure 5). (Retention times cannot be precisely compared because the analyses were done at different times and under slightly different conditions.)</p><p>The product-ion spectra of 4-OH-E1-1-N3Ade with U-15N labeled IS are characteristic of the assigned structure, as was discussed in a recent publication.73 The identities of these CE1-Ade adducts are also consistent with accurate mass measurements (top panel in Figure 6). The mass tolerance is ±0.0015 mass units (±3.6 ppm) from the theoretical mass, and all signals within that tolerance and produced within the mass range of 200–2000 comprise the chromatograms in the figure. Two CE1-Ade adducts are clearly characterized and a third eluting at 19.93 min, is missed in the TIC generated from the product ions (middle panel, Figure 6). The shoulder species may correspond to the weakly bound adduct (species of tR 18.95 in Figure 4 and tR 21.32 min in Figure S5), which dissociates promptly during the precursor-ion isolation step of the MS/MS sequence, and thus cannot be seen in the MS/MS mode as displayed in the middle panel of Figure 6. The use of U-15N labeled IS allows us to quantify the amount of 4-OH-E1-1-N3Ade formed by depurinating of the double-stranded DNA to be 190 ± 40 fmol (formed in the incubation of 60 nmole of CE1-2,3-Q and 40 nmole of CE1-3,4-Q with 1 mg calf thymus DNA under physiological conditions). In addition, there is a second CE1-Ade adduct of approximately equal abundance. We were unable to detect any of modified guanine bases, which is in accord with previous observations that catecholestrogen-modified Gua depurinates from DNA at a slower rate than does modified Ade. The amount of a modified quinone is less than 30% of the adenine. The more facile release of CE-modified Ade suggests that there will be a higher probability in identifying them in tissue extracts or biological fluids for biomarker discovery. In contrast, CE-Gua adducts may be a more appropriate target for exploring biomarkers that remain in the double-stranded DNA.</p><p>The same procedure with CGenQ at either the same level or 20 times of the amount used for CE1Qs produced, not surprisingly, no detectable CGen-modified Ade species (detection limit is 50 fmol). The lower reactivity is in accord with more rapid reaction of CGenQ in H2O.</p><!><p>The observations of the rapid conversion of CGenQ to its H2O conjugate and the facile reduction of CGenQ·(H2O)2 to CGen allow us to propose a mechanism whereby the ability of certain quinones (e.g., that from genistein) to damage macromolecules is considerably diminished by a fast reaction with H2O (Scheme 4). Thus, although Gen and related quinones can be metabolized to catechols and to quinones, the quinones do not participate significantly in redox cycling or add directly to purine bases of DNA because they are rapidly and covalently hydrated. The structure of CGenQ·(H2O)2 is one that would be expected to be incapable of generating ROS and/or attacking nucleophilic sites in DNA. This stable dihydrate can be reduced back to CGen, 3′-OH-Gen. The short "dwell time" of Gen in its quinone state should lead to low cytotoxicity and/or genotoxicity, as must be proved in other studies. Instead, we expect that Gen would act primarily as an antioxidant in vivo.</p><p>In comparison, CE1Qs, CE1-3,4-Q in particular, react with H2O at substantially slower rates. These longer-lived quinones are available to participate in redox cycling and/or to damage DNA by direct reaction followed by depurination. Once CE1Q hydrates are formed, they can be reduced back to their catechols, CE1s and then reoxidized generating again relatively long-lived quinones available again for redox cycling and depurinating reactions. If Gen and similar materials are present along with estrogens, however, oxidative phase I metabolism of Gen may absorb the oxidizing equivalents in vivo, preventing oxidation of estrogens and preventing its untoward biological effects.</p><!><p>The quinone from genistein is much less reactive with Ade and with DNA, suggesting that it is less biologically deleterious than the CEQs. The diminished reactivity of CGenQ is due to its rapid reaction with water, converting it to a dihydrated species, CGenQ·(H2O)2, under physiological conditions. Under mild reductive conditions, the dihydrate is reduced back to CGen. The cycling of CGenQ, CGenQ·(H2O)2 and CGen in water is an effective means to diminish any potential toxicity. The rapid reaction with water exhibited by the Gen quinone suggests to us that Gen exerts primarily anti-oxidation activities as radical scavenger. In addition, by competing for quinone formation, the oxidative metabolism of Gen may reduce ROS generation by other more deleterious quinone species. Simple chemical reactivities, of course, are not the only predictors of genotoxicity; other factors (e.g., bioavailability) may also play a role. Nevertheless, the chemical properties of estrogen and estrogen-like materials are so compellingly different that we expect those differences to have an impact on biological and toxicological properties.</p><!><p>LC/MS, MS/MS, accurate mass, NMR and additional mass spectra. This material is available free of charge via the Internet at http://pubs.acs.org.</p><!><p>Measurement of the half lives of CGen-3′,4′-Q with GSH trapping. Quinones were deactivated in H2O at 37 °C, pH 7.4. Aliquots were removed and combined with GSH at various time. Data were fitted with an exponential function.</p><p>LC/MS analysis of CGen-3′,4′-Q. The identity of [CGen-3′,4′-Q + H2O] conjugate (tR = 24.99 min) is consistent with an accurate mass measurement.</p><p>LC/MS analysis of CE1Qs. The identities of two [CE1Q + H2O] conjugates (tR = 24.76 and 25.44 min) are consistent with accurate mass measurement.</p><p>Reconstructed ion chromatogram showing the formation of CGen-Ade adducts and CE1-Ade adducts. Analysis is of a 50:1 mixture of CGenQ + Ade and CE1Q + Ade (37 °C, pH 7.4). The observed ratio of CGen-Ade adducts to CE1-Ade adducts is 1:620.</p><p>MS/MS product-ion spectra of the two CE1-Ade adducts, 4-OH-E1-1-N3Ade at tR = 18.51 min (bottom panel) and unknown at tR = 21.16 min (top panel) from Figure 4.</p><p>Product mixture from reaction of CE1Q and calf thymus DNA at 37 °C, pH 7.4, as analyzed with the ion-trap/FT-ICR mass spectrometer. Accurate mass measurement in top panel shows two and possibly three major CE1-Ade adducts. The middle panel show the corresponding TIC measured in ion trap. Bottom panel shows the TIC of internal standard, 4-OH-E1-1-N3 (U-15N)Ade in ion trap.</p><p>Mechanisms for quinone-induced carcinogenesis, either indirectly via ROS or directly via Michael addition and depurination.</p><p>A) Synthesis of CGen-3′,4′-Q and its CGen-Ade adducts from Gen. IBX oxidizes Gen at ring B, forming a single CGenQ product, CGen-3′,4′-Q. CGen-3′,4′-Q reacts with Ade and forms several CGen-Ade adducts. B) Synthesis of CE1-2,3-Q, CE1-3,4-Q and their CE1-Ade adducts from E1. IBX oxidizes E1, forming two CE1Qs: CE1-2,3-Q and CE1-3,4-Q. CE1Qs react with Ade to form 4-OH-E1-1-N3Ade and other CE1-Ade adducts, whose structures are not shown.</p><p>Proposed mechanism for the reaction of CGenQ, Gen-3′,4′-Q, with H2O to form CGenQ·(H2O)2.</p><p>Proposed mechanism for the antioxidant activities of Gen and its catechol and the "detoxification" of its quinone metabolite.</p><p>1H and 13C NMR assignment for CGen-3′,4′-Q (structure in Scheme 3).</p><p>1H peak splitting (s, singlet; d, doublet; dd: doublet of doublet).</p><p>Determined by gradient Heteronuclear Multiple Quantum Coherence (gHMQC).</p><p>Determined by gradient Heteronuclear Multiple Bond Coherence (gHMBC).</p><p>1H, 13C NMR assignment for CGenQ·(H2O)2 (structure in Scheme 3).</p><p>1H peak splitting (s, singlet; d, doublet; dd: doublet of doublet).</p><p>Determined by gradient Heteronuclear Multiple Quantum Coherence (gHMQC).</p><p>Determined by gradient Heteronuclear Multiple Bond Coherence (gHMBC).</p>

PubMed Author Manuscript

DiPODS: A Reagent for Site-Specific Bioconjugation via the Irreversible Rebridging of Disulfide Linkages

Chemoselective reactions with thiols have long held promise for the site-specific bioconjugation of antibodies and antibody fragments. Yet bifunctional probes bearing monovalent maleimides \xe2\x80\x94 long the \xe2\x80\x98gold standard\xe2\x80\x99 for thiol-based ligations \xe2\x80\x94 are hampered by two intrinsic issues: the in vivo instability of the maleimide-thiol bond and the need to permanently disrupt disulfide linkages in order to facilitate bioconjugation. Herein, we present the synthesis, characterization, and validation of DiPODS, a novel bioconjugation reagent containing a pair of oxadiazolyl methyl sulfone moieties capable of irreversibly forming covalent bonds with two thiolate groups while simultaneously re-bridging disulfide linkages. The reagent was synthesized from commercially available starting materials in 8 steps, during which rotamers were encountered and investigated both experimentally and computationally. DiPODS is designed to be modular and can thus be conjugated to any payload through a pendant terminal primary amine (DiPODS\xe2\x80\x93PEG4-NH2). Subsequently, the modification of a HER2-targeting Fab with a fluorescein-conjugated variant of DiPODS (DiPODS-PEG4-FITC) reinforced the site-specificity of the reagent, illustrated its ability to rebridge disulfide linkages, and produced an immunoconjugate with in vitro properties superior to those of an analogous construct created using traditional stochastic bioconjugation techniques. Ultimately, we believe that this work has particularly important implications for the synthesis of immunoconjugates, specifically for ensuring that the attachment of cargoes to immunoglobulins is robust, irreversible, and biologically and structurally benign.

dipods:_a_reagent_for_site-specific_bioconjugation_via_the_irreversible_rebridging_of_disulfide_link

INTRODUCTION<!>Synthesis and Characterization<!>Variable Temperature NMR<!>Computational Studies<!>Synthesis of a Fluorophore-Bearing Variant<!>Reactivity with a Model Thiol<!>Bioconjugation and Characterization<!>In Vitro Evaluation<!>CONCLUSION<!>Reagents<!>Characterization Methods<!>Synthesis of 5-[[(1,1-dimethylethoxy)carbonyl]amino]-1,3-dimethyl ester (1)<!>Synthesis of Compound 2<!>Synthesis of Compound 3<!>Synthesis of Compound 4<!>Synthesis of Compound 5<!>Synthesis of Compound 6<!>Synthesis of Compound 7<!>Synthesis of Compound 8<!>Synthesis of Compound 9<!>Synthesis of Compound 10<!>Synthesis of Compound 11<!>Synthesis of Compound 12<!>Synthesis of DiPODS<!>Synthesis of DiPODS-FITC<!>Variable Temperature NMR Experiments<!>Computational Studies<!>Preparation of Reduced FabHER2<!>Preparation of Reduced Fabns<!>Preparation of FabHER2-DiPODS-FITC<!>Preparation of Fabns-DiPODS-FITC<!>Preparation of FabHER2-Lys-FITC Conjugate<!>Procedure for Quantitating Fluorescein Degree of Labeling<!>Procedure for Quantitating Sulfhydryl Groups<!>SDS-PAGE Analysis of Conjugates<!>Analytical Size-Exclusion Chromatography<!>Stability Study in Human Serum Albumin<!>MALDI-ToF Mass Spectrometry<!>Circular Dichroism Spectroscopy<!>Flow Cytometry

<p>Over the last two decades, immunoconjugates have emerged as vitally important therapeutic and diagnostic tools. However, the imprecise synthetic methods used to create many antibody-drug conjugates (ADCs) and radioimmunoconjugates remains an impediment to their widespread success.2 Traditional approaches to bioconjugation are predicated on the indiscriminate attachment of payloads — e.g. chelators, fluorophores, or toxins — to lysine residues within antibodies. Yet these non-site-specific synthetic strategies inevitably lead to heterogeneous product mixtures and can produce constructs with suboptimal immunoreactivity and in vivo performance.</p><p>In light of these issues, the development of 'site-specific' bioconjugation methods designed to append cargoes only at well-defined sites within an antibody's macromolecular structure has become an area of intensive research.3-8 A wide variety of these approaches have been devised, including variants based on the manipulation of the heavy chain glycans, the use of peptide tags, and the genetic incorporation of unnatural amino acids. Far and away the most popular methods, however, rely upon the reaction between maleimide-based bifunctional probes and cysteine residues within the biomolecule (Figure 1A). While maleimide-based bioconjugation strategies are undeniably facile, rapid, and modular, they nonetheless suffer from a critical flaw: the inherent instability of the thioether bond between the maleimide and cysteine. The Michael addition reaction that forms this linkage is reversible in vivo both spontaneously (retro-Michael) and in the presence of competing thiols.9, 10 This, of course, can be a significant problem. In the context of radioimmunoconjugates, for example, this process can result in the in vivo release of radionuclides, reducing target-to-background activity concentration ratios and increasing radiation doses to healthy tissues.11-15</p><p>Two years ago, in an effort to circumvent the inherent limitations of maleimides, we reported the synthesis, characterization, and in vivo validation of an alternative: PODS.3 Inspired by the work of the late Carlos Barbas III, PODS is an easily synthesized phenyloxadiazolyl methyl sulfone-based reagent capable of rapidly and irreversibly forming covalent linkages with thiols (Figure 1B).16-19 This work clearly illustrated that a 89Zr-DFO-labeled variant of the huA33 antibody synthesized using a PODS-based bifunctional chelator exhibited superior in vitro stability and — even more importantly — in vivo performance compared to an analogous radioimmunoconjugate synthesized using a traditional, maleimide-based probe.3 Furthermore, the innate modularity of PODS enabled the creation of PODS-CHX-A″-DTPA and PODS-DOTA bifunctional chelators for the synthesis of radioimmunoconjugates labeled with lutetium-177 and actinium-225.</p><p>While PODS-based reagents represent a distinct improvement compared to their maleimide-based forerunners, neither tool can avoid an intrinsic problem common to the overwhelming majority of thiol-targeted bioconjugations. In the absence of free cysteine residues incorporated via genetic engineering, all of the cysteines within an antibody are paired to form 8 intrachain and 8 interchain disulfide bridges. As a result, thiol-based bioconjugation strategies require the reduction of these disulfide bridges to generate free thiols, with the slightly-easier-to-reduce interchain linkages often the target of selective scission.8 While the subsequent reaction of these free cysteines with thiol-selective probes enables the site-specific attachment of cargoes to the immunoglobulin, it simultaneously seals the fate of the broken disulfide bridges, potentially reducing the stability of the macromolecule and attenuating effector functions.20 A handful of reagents capable of reacting with two thiols and thus re-forming the covalent bridge between the reduced cysteine residues have been developed.20-33 But immunoconjugates synthesized using the most widely studied of these tools — dibromo- and dithiophenolmaleimides — are still prone to instability in vivo. While the developers of this 'next generation maleimide' technology tout this reversibility as an advantage in the context of ADCs, it nonetheless remains an obstacle for radioimmunoconjugates.</p><p>Herein, we present the development and evaluation of DiPODS, a novel reagent bearing two oxadiazolyl methyl sulfone moieties designed to provide a modular platform for irreversible bioconjugations while simultaneously re-bridging disulfide linkages (Figure 1C). Following the synthesis and chemical characterization of the DiPODS scaffold — during which rotamers were discovered and investigated both experimentally and computationally — a fluorescein-labeled variant of the reagent (DiPODS-FITC) was created for proof-of-concept bioconjugation experiments. More specifically, the reaction conditions for DiPODS-FITC were optimized using both isotype-control and HER2-targeting Fab fragments, and the FITC-bearing immunoconjugates were characterized via gel electrophoresis, size exclusion HPLC, and circular dichroism spectroscopy. Finally, the cell binding behavior of the HER2-targeting Fab-DiPODS-FITC was interrogated via flow cytometry and compared to that of an analogous Fab-FITC immunoconjugate created via a traditional, stochastic lysine-based approach to bioconjugation.</p><!><p>DiPODS was prepared in 8 synthetic steps with good to high yield at each step, but the synthetic journey was tortuous (Scheme 1). The synthesis began with the Boc-protection of aminoisophthalate, which followed a published procedure with some minor alteration.34 The Boc-protection was performed under nitrogen atmosphere overnight and produced compound 1 with 74% yield after purification. Surprisingly, the 1H NMR spectrum of the crude mixture of compound 1 revealed three sets of signals for all functional groups except for the proton of the secondary amine, which was represented by a single broad peak in the 1H-NMR spectra (vide infra). While a combination of normal-phase chromatography and precipitation facilitated the partial separation of these products, all three revealed the same molecular weight by mass spectrometry, suggesting that they are conformers of 1 (for further exploration of this phenomenon, see below). The crude mixture of 1 was then treated with hydrazine hydrate and, somewhat surprisingly, produced 5-amino isophthalic dihydrazide 2 in quantitative yield. This intermediate was subsequently treated with ethanol, KOH, and carbon disulfide to create phenyl-bis(oxadiazole thiol) 3 in 91% yield. Next, the methylation of 3 using methyl iodide generated the bis(methyl thioether) 4 in near-quantitative yield.</p><p>The first challenges in the synthesis emerged as we progressed beyond compound 4. In our first attempt, bis(methyl thioether) 4 was directly oxidized via meta-chloroperoxybenzoic acid (mCPBA) to form the bis(methyl sulfonyl) 5 followed by Boc-deprotection to form compound 6 (Scheme 2A). The plan was to use compound 6 in a coupling reaction with a carboxylic acid-bearing polyethylene glycol (PEG) chain. However, several attempts at this peptide coupling reaction failed or resulted in unacceptably poor yields. Aryl amine groups are notoriously poor nucleophiles, and we believe that the reactivity of the aryl-amine in compound 6 is reduced even further by the electron-withdrawing methyl sulfonyl substituents.</p><p>Methyl thioether groups are less electron-withdrawing than the methyl sulfonyl substituents. Following this logic, we decided to perform the coupling reaction prior to the formation of the methyl sulfonyl moieties with the hope that this version of the aryl-amine had enhanced nucleophilicity. To this end, compound 4 was first deprotected in quantitative yield to produce 7. Subsequently, in the first attempt at coupling, 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5b]pyridinium 3-oxide (HATU) was used alongside N,N-diisopropylethylamine (DIEA). These conditions yielded <15% of the desired product, a result that mass spectrometry analysis suggested is related to the degradation of the starting materials. In response, DIEA was then swapped for a milder base — 2,4,6-trimethylpyridine (TMP) — and the reaction was attempted at room temperature as well as 50 °C, yet both attempts proved unsuccessful. Next, we decided to change our synthetic strategy and reverse the coupling chemistry by transforming the aryl-amine into a carboxylic acid via the reaction of 7 with succinic anhydride to form 8 (Scheme 2B).</p><p>With compound 8 now containing a carboxylic acid, we attempted a peptide coupling reaction with a mono-Boc-protected bisamino-PEG chain. But the use of HATU and DIEA at both room temperature and 50 °C resulted in an unwanted cyclization and the formation of compound 9 — a clear dead end — as the major product. This same transformation was then attempted using N-ethoxycarbonyl-2-ethoxy-1,2-dihydroquinoline (EEDQ) as an alternative coupling reagent (Scheme 3C). Disappointingly, this reaction resulted in the recovery of nearly 33% starting material as well as two products: the cyclized phenyl succinimide 9 (9% yield) and the desired product 10 (<18% yield).</p><p>To continue our efforts to search for a higher yielding route forward, we returned to bis(methyl thioether) 7 as a starting point to test a new set of peptide coupling conditions: oxyma with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) (Scheme 1). At room temperature, this reaction yielded <10% of the desired product (compound 11), with mass spectrometry revealing the presence of unreacted starting material 7 as well as the O-acylisourea-activated EDC intermediate. Suspecting that the EDC intermediate was trapped in a step with an energy barrier that was impassable at room temperature, we repeated the same reaction at 50 °C. This time, the PEGylated product 11 was obtained in 55% yield. To complete the sequence, 11 was then oxidized with mCPBA to create bis(methylsulfonyl) 12 in 73% yield, and — finally — compound 12 was deprotected to provide DiPODS in ~90% yield. We concluded this synthetic journey with an 8-step synthetic route to produce DiPODS with a cumulative yield of ~15%. As synthesized, DiPODS is modular and can be coupled to any number of different bifunctional chelators, dyes, or other payloads. The primary amine of DiPODS can be reacted with a number of electrophilic bioconjugation reagents such as activated esters or phenyl-isothiocyanates.</p><!><p>Before interrogating the reactivity of DiPODS, we thought it interesting and important to take a more detailed look at the intriguing structural data collected for compound 1. As we noted above, the 1H-NMR of the crude mixture of 1 revealed a mixture presumed to be composed of conformers (Figures 2A and 2B). The 1H-NMR spectrum contained three sets of signals — sets A, B, and C — for each functional group, with the exception of the Boc-protected amine, which produced a single broad peak (Figure 2B, Figure S1). Curiously, the use of this mixture — without any separation — resulted in the formation of compound 2 in quantitative yield (Scheme 1).</p><p>In an attempt to separate and identify the components of the crude product mixture, it was dissolved in warm DCM and stored at −20 °C overnight. The first attempt at precipitation produced a shiny white precipitate that was separated from the mother liquor via filtration and dried under high vacuum. The 1H-NMR of this white precipitate displayed only one set of signals — set A — for all the functional groups, including the proton of the secondary amine (Figures 2C and S8). The isolation of pure set A, and the following investigation of the remaining mother liquor mixture by VT-NMR strongly suggests that compound 1 — like many other carbamate-bearing molecules — exists as syn- and anti-rotamers.35, 36 The anti-rotamers of compound 1 are more energetically favored due to less steric hindrance between the tert-butyl group and the ester group (Figures 2A and 3; for a more detailed discussion see Computational Studies). Therefore, the signals of set A were assigned to a mixture of the anti-rotamer of compound 1. To be more specific, while the anti-rotamer configuration of the Boc group remains constant, the two methyl ester groups can rotate freely, creating a subset of conformers for each of the syn- and anti-rotamers (i.e. sub-conformers).</p><p>After isolating the precipitate from the crude product mixture, the solvent was removed from the mother liquor, and the solid residue was subjected to several more rounds of precipitation. After each round, the precipitate was isolated, and each time it was found via 1H-NMR to be predominantly composed of the anti-rotamer (set A). Following several cycles, the aggregate mother liquor was concentrated under vacuum and found via 1H-NMR to contain both sets B and C as well as a small amount of set A (Figure S31 and 2D). In order to better understand the NMR spectrum of the product mixture of compound 1, a series of NMR spectra were collected at different temperatures. Two NMR samples were prepared from the components of crude compound 1. The first contained only the precipitate, i.e. the anti-rotamers (Figure 2C, set A). The second contained the mixture isolated from the mother liquor following precipitation (Figure 2D). The latter is composed mostly of the compounds responsible for sets B and C but also some of the anti-rotamer (set A). A more detailed explanation of the VT-NMR experiments and assignments can be found in the Experimental section and the Supporting Information (Figures S36 and S37), with the results summarized in Figures 2 and 3.</p><p>Ultimately, set B was attributed to a doubly Boc-protected version of compound 1 based on the integration ratio between the methyl ester (6) and tert-butyl (18) protons as well as the presence of a tertiary amine group with no proton signal. High resolution mass spectrometry subsequently confirmed this assignment. As removing the first of two Boc protecting groups is easier than the second, the doubly protected compound (set B) appears to be converted to compound 1 at elevated temperatures (VT NMR, Figure 2E, Table S1, Figures S36 and S37). Set C, in contrast, has an integration ratio of 6:9 between the methyl ester (6) and tert-butyl (9) protons, confirming that the compound responsible for these peaks has a single Boc group. However, no proton associated with the amine was observed. Furthermore, upon heating to 90 °C set C disappeared almost entirely. We propose that this phenomenon can be explained by a tautomerization reaction involving the transfer of a proton from the amine to the neighboring oxygen (Figure 3). The assignment of set C as a tautomeric form of set A would explain why the integration of the former matches that of the latter except for the absence of the proton from the amine group. In the end, these NMR data permit us to deconvolute the constituents of the original compound 1 product mixture: an anti-rotamer of compound 1 (anti-1, set A), a doubly Boc-protected variant of compound 1 [(Boc)2-1, set B], and an imidic acid tautomer of compound 1 (tatomer-1, set C) (Figure 3). These findings also explain how a crude mixture of 1 containing all of these components was reacted with hydrazine hydrate and produced 5-amino isophthalic dihydrazide 2 in quantitative yield.</p><!><p>Our computational investigation of the isomers of compound 1 supports the assignments made based on the VT-NMR data. The calculated Gibbs free energies of the rotamers of compound 1 revealed that the anti-rotamers are favored by ~2.0 kcal/mol (Figure 4). Figure 4 shows the calculated structures of two groups of rotamers and a tautomer of compound 1. The first group of rotamers includes four configurations of anti-rotamers (anti-1a, anti-1b, anti-1c, and anti-1c′) with energies similar to each other and to tautomer-1. The second group — which consists of three configurations of syn-rotamers (syn-1a, syn-1b, and syn-1c′) with similar energies — lies ~2.0 kcal/mol higher than the anti-rotamers and the tautomer. The energy difference between the four anti-rotamers is very small (~0.5 kcal/mol), suggesting that they can interconvert at room temperature. This explains why they all manifest as a single set of signals (set A) in the 1H NMR spectrum of compound 1 despite having different point group symmetries. Despite the calculated similarity in energy between the set A anti-rotamers and the imidic acid tautomer-1 (Set C), they do not appear to interconvert at ambient temperature (Figure 2B). This suggests that a higher energy transition state must be passed for conversion, which is supported by the disappearance of the tautomer (Set C) at elevated temperatures.</p><p>The interconversion between the anti- (set A) and syn- (set D) rotamers occurs via the rotation of the Boc group attached to the amine. In order to further understand this process, we calculated the transition state for one such rotation between anti-1b and syn-1b (Figure 5). To identify the transition state (1b*), we started with the anti-1b rotamer, and the dihedral angle of interest was varied in a stepwise fashion towards that of the syn-1b rotamer using Spartan 14 software. The structure with the highest energy was carried forward for optimization as the transition state in Gaussian 16.37 The harmonic vibrational frequencies showed only one imaginary frequency, corresponding to the desired transition. The energy difference between the transition state and the anti-1b rotamer is substantial (~15 kcal/mol) and thus might not be overcome at room temperature, depending on other conditions such as solvent (Figure 5). One way to overcome this large energy barrier, however, is via heating, which could explain the formation of a separate set of 1H NMR signals (set D) at elevated temperatures. It is important to note that the energy difference between the syn-rotamers is also small (~0.3 kcal/mol), suggesting that they can interconvert easily at elevated temperatures and thus explaining their appearance as a single set of peaks in the 1H-NMR spectra.</p><p>Taken together, the aforementioned NMR and computational studies helped deconvolute the mixture of components formed when synthesizing compound 1. Furthermore, these data help explain how this mixture of anti-rotamers, tautomer-1, and a doubly Boc-protected variant of compound 1 can react together to form compound 2: the elevated temperature of the reaction — 90 °C for 3 days — would overcome any rotational energy barriers and allow for the production of compound 2 in quantitative yield.</p><p>Although these detailed VT-NMR and computational studies might appear to be more detail than needed for characterizing what could be viewed as a mere synthetic intermediate, we thought it prudent to better understand the behaviors of this molecule in terms of its observed conformers and the energetic barriers to its observed isomerization. The desired applications we are pursuing with DiPODS require it to react in a predictable and reproducible manner with two thiols. For example, if the reactivity were to be different for two rotamers/isomers of DiPODS, this could become an important physical property to understand. From our investigations, we believe the rotamer behavior appears to be largely the result of the Boc-protected amine and therefore not likely to be an issue for the final DiPODS compounds. Further, the imidic acid tautomer that we propose as Set C is not likely to form in DiPODS itself or its derivatives, as the pKa of the amide in these final conjugates is higher than that of the Boc-protected carbamate (which forms Set C).</p><p>We also turned to computational methods to compare the thermodynamic stability of the conjugation product formed by DiPODS to those formed by a bivalent maleimide, a monovalent maleimide, and a monovalent PODS. To this end, ethanethiol was employed as a simple surrogate substrate, and the total energy of the final product(s) was compared to the total energy of the starting materials using the UAHF model for improved solvent modeling (Figures 6 and S1). Since all of the reactions were modeled as isodesmic, we were able to calculate the difference in total energy — i.e. Gibbs free energy — between the reactants and products in each case, thereby enabling a comparison between the net change in thermodynamic stability of each transformation. The ligation between ethanethiol and the monovalent maleimide result in a net Gibbs free energy change of −5.3 kcal/mol, while that between the same substrate and the monovalent PODS is slightly more stabilizing, with a net change of −5.6 kcal/mol (Figure S2). Not surprisingly, the divalent reagents created larger changes in free energy. More specifically, the reaction of the bivalent maleimide resulted in a change in Gibbs free energy of −10.3 kcal/mol, while the ligation between DiPODS and a pair of ethanethiols provided an even greater gain in stability: −12.4 kcal/mol. While an extra ~2.1 kcal/mol of stabilization does not represent a dramatic improvement, it — combined with the irreversibility of the DiPODS-based conjugation — certainly suggests that DiPODS-based conjugates will be more stable than their bismaleimide-based analogues both in vitro and in vivo.</p><!><p>DiPODS was designed to be modular, as its reactive primary amine facilitates the coupling of cargoes such as chelators, dyes, and toxins. In order to facilitate proof-of-concept reactivity and bioconjugation experiments, a fluorescein-bearing variant of DiPODS — DiPODS-FITC — was prepared via the reaction of DiPODS with fluorescein isothiocyanate in the presence of DIEA (Scheme 3).</p><!><p>N-acetyl-L-cysteine methyl ester was used as a model thiol to evaluate the reactivity of DiPODS-FITC (Scheme S1). To this end, DIPODS-FITC was incubated at room temperature with 10 equivalents of N-acetyl-L-cysteine methyl ester and 5 equivalent of a mild reducing agent, tris(2-carboxyethyl)-phosphine (TCEP). The progress of the reaction was interrogated via LC-MS 5 minutes after mixing, and quantitative conversion to DiPODS-FITC-Cys2 was observed. (Figures S4 and S5).</p><!><p>Fab fragments — rather than full-length IgGs — were selected for our proof-of-concept bioconjugation experiments with DiPODS-FITC because of the presence of only a single interchain disulfide linkage (rather than 8) dramatically simplifies the analysis of the products. In practice, two Fabs were employed: a commercially-available, non-specific Fab based on human plasma IgG (Fabns) and a HER2-targeting Fab created via the enzymatic digestion of trastuzumab (FabHER2). In each case, the Fab was first treated with TCEP to reduce the interchain disulfide bridge and then incubated with DiPODS-FITC (Figure 7A). Ultimately, the following optimal reaction conditions were identified: 2 h at 37 °C with 20 equivalents of TCEP followed by 16 h with 15 equivalents of DiPODS-FITC at the same temperature. Subsequently, UV-Vis spectrophotometry was used to measure the degree of labeling (DOL) of each immunoconjugate, revealing that Fabns-DiPODS-FITC and FabHER2-DiPODS-FITC were modified with 0.86 ± 0.02 and 0.95 ± 0.01 FITC/Fab, respectively (Table 1). MALDI-TOF mass spectrometry confirmed a degree of labeling of ~1 for each fluorophore-modified Fab (Figure S3).</p><p>The stepwise progress of the bioconjugation procedure was monitored using both gel electrophoresis and Ellman's reagent, a chemical tool for the detection of free thiols. In the case of FabHER2, for example, the former illustrates the decoupling of the intact fragment's VHCH1 and VLCL chains upon reduction with TCEP (Figure 7B, lanes 1 and 2,) and the subsequent reunification of the two domains after treatment with DiPODS-FITC (Figure 7B, lane 3). The analysis of the gel using a fluorescence imager reveals only a single fluorescent band corresponding to an intact, 40-50 kDa Fab and does not show any multimeric cross-bridged species (i.e. Fab-DiPODS-Fab) (Figure 7C). The use of Ellman's reagent to assess the number of free thiols present at different points of the procedure reinforced the quantitative nature of the approach. The purified FabHER2 starting material contains no detectable free thiols. Reduction with TCEP creates the expected maximum of 1.94 ± 0.11 thiols/Fab, a value which went back to effectively zero again upon crossbridging with DiPODS-FITC (Table 1). Importantly, both analytical techniques provided similar results for the bioconjugation of Fabns.</p><p>Next, circular dichroism (CD) spectroscopy was employed to interrogate the structure and melting point of FabHER2, reduced FabHER2, and FabHER2-DiPODS-FITC. Generally speaking, the spectra — which exhibit a positive peak around 205 nm and shallow negative peak around 217 nm — are characteristic of a protein rich in β-sheet content, consistent with the known secondary structure of Fab fragments (Figure S40). The data suggest that the trio of constructs have similar overall structures: the far-UV CD spectra of all three samples have the same shape profile, with only minor differences in ellipticity values which may reflect local conformational adjustments due to the reduction or re-bridging of the disulfide bonds. Importantly, the CD data also indicate that the three fragments also share similar thermal stability, as the melting temperatures for FabHER2, reduced FabHER2, and FabHER2-DiPODS-FITC are 65.5 °C, 66.8 °C and 64.5 °C, respectively, when monitored at 205 nm (Figure S41).</p><p>Finally, in order to assess the serum stability of Fabns-DiPODS-FITC and FabHER2-DiPODS-FITC, the fragments were incubated in 50% human serum albumin (HSA) for 7 days at 37 °C. Size exclusion HPLC of each fluorophore-bearing fragment after 7 days yielded a single, unchanged peak (Figure S6). Neither aggregates nor separate VHCH1/VLCL chains nor free fluorophores could be observed, underscoring the stability of the FITC-modified immunoconjugates and the irreversibility of the DiPODS linkage.</p><!><p>With the chemical characterization of FabHER2-DiPODS-FITC complete, the next step was to ensure that the immunoconjugate retained its ability to bind its molecular target. To this end, we performed flow cytometry experiments using two human breast cancer cell lines: HER2-positive BT474 cells and HER2-negative MDA-MB-235 cells. As a point of comparison, a non-site-specifically modified, HER2-targeting immunoconjugate (FabHER2-Lys-FITC) was synthesized using a traditional lysine-based approach to bioconjugation and used alongside FabHER2-DiPODS-FITC in all cell cytometry experiments. The in vitro experiments clearly confirm the specificity of both immunoconjugates, as binding was observed with HER2-positive BT474 cells but not HER2-negative MDA-MB-231 cells. Just as important, however, are the differences between the behavior of the two FITC-modified Fabs and HER2-positive BT474 cells. Under identical conditions — i.e. concentration of cells, concentration of fragments, incubation time — only a single population of fluorophore-positive cells were detected after incubation with FabHER2-DiPODS-FITC, but both fluorophore-positive and fluorophore-negative cells were observed after incubation with FabHER2-Lys-FITC (Figure 8).</p><p>These data indicate that the immunoroeactivity of FabHER2-DiPODS-FITC is higher than that of FabHER2-Lys-FITC, most likely because the heterogeneous mixture of products that comprises the latter includes immunoconjugates in which fluorophores have been inadvertently appended to the antigen-biding domain of the fragment. These data serve as a reminder that the benefits of site-specific bioconjugation extend beyond simply producing better-defined and more homogeneous immunoconjugates.</p><!><p>In the preceding pages, we have described the synthesis, chemical characterization, computational investigation, and biological evaluation of DiPODS, a reagent for site-specific bioconjugation bearing two thiol-reactive oxadiazolyl methyl sulfones. Where maleimide-thiol conjugations form reversible and labile linkages, DiPODS-thiol conjugations form strong and irreversible linkages. Proof-of-concept bioconjugation experiments with a pair of Fab fragments demonstrate that DiPODS reliably facilitates the construction of stable and homogeneous immunoconjugates via the rebridging of interchain disulfide bonds. Moreover, flow cytometry experiments with human breast cancer cells illustrate that a fluorescent, HER2-targeting immunoconjugate synthesized using DiPODS exhibits superior in vitro performance compared to an analogous construct synthesized via a non-site-specific, lysine-based approach to bioconjugation. Efforts to leverage DiPODS for the construction of immunoconjugates based on full-length IgGs are already underway, and in the near future, we plan to exploit the modularity of DiPODS to create derivatives bearing bifunctional chelators for the construction of immunoconjugates for PET, SPECT, and targeted endoradiotherapy.</p><!><p>Unless otherwise stated, all chemicals and reagents were obtained commercially and used without further purification. 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC), carbon disulfide, peptide synthesis-grade N,N-diisopropylethylamine (DIEA), triethylamine, sodium hydroxide (NaOH), and trifluoroacetic acid (TFA) were purchased from Fisher Scientific. 4-(Dimethylamino)pyridine, hydrazine hydrate, iodomethane, and oxyma pure were purchased from Sigma-Aldrich. Dimethyl 5-aminoisophthalate was purchased from Alpha-Aesar. Di-tert-butyl decarbonate and 3-chloroperoxybenzoic acid (mCPBA) were purchased from AK Scientific. Fluorescein isothiocyanate isomer I and t-Boc-N-amido-PEG4-acid were purchased from BroadPharm. All chemicals for the in vitro and in vivo experiments, unless otherwise noted, were acquired from Sigma-Aldrich and used as received without further purification. All water used was ultra-pure (>18.2 MΩcm−1), and dimethylsulfoxide was of molecular biology grade (>99.9%). The unconjugated FabHER2 and was provided by Rockland Immunochemicals, Inc. (Pottstown, PA).</p><!><p>1H- and 13C-NMR spectra were recorded on a 500 MHz Bruker Avance NMR spectrometer at 25 °C in [D6]DMSO. Variable Temperature 1H NMR were recorded on a Bruker Avance III HD 600 MHz spectrometer. 1H chemical shifts were referenced to the residual protons of the deuterated [D6]DMSO solvent at δ = 2.50 ppm;38 13C chemical shifts were referenced to the [D6]DMSO signal at δ = 39.52 ppm.39 Coupling constants are reported to the nearest 0.5 Hz (1H NMR spectroscopy) or rounded to integer values in Hz (13C NMR spectroscopy). Assignments were supported by additional NMR experiments (DEPT135, HMQC, COSY). High resolution mass spectra were measured on a JEOL AccuTOF GCv 4G using field desorption ionization (FDI). For the isotopic pattern only, the mass peak of the isotopologue or isotope with the highest natural abundance is given. Low resolution mass spectrometry and LCMS was performed using an Advion Expression-L system (mass range <2000 amu). FTIR spectroscopy was performed using a Bruker Tensor 27 FT-IR spectrometer equipped with ATR attachment and OPUS data collection program. HPLC purifications were performed using a Vanquish HPLC equipped with C18 reversed-phase column (Spursil Semipreparative DIKMA; 5μm, 21.2×250mm), a VF-D40-A UV detector, two VF-P10-A pumps, a Chromeleon 7 communication software, and a DIONEX UltiMate 3000 fraction collector, using a flow rate of 8 mL/min and a gradient of MeCN:H2O (both with 0.1% TFA). Compounds 5, 6, 8, 9, and 10 were only characterized using 1H NMR spectroscopy and low-resolution mass spectrometry, as they were only used toward developing an optimized procedure.</p><!><p>Compound 1 was prepared according to a published procedure with some alteration.34 Using Schlenk techniques, portions of commercially procured dimethyl 5-aminoisophthalate (3.00 g, 14.34 mmol) and 4-(dimethylamino)pyridine (2.24 g, 18.36 mmol) were purged with N2 gas, transferred to a reaction vessel under N2 gas, and dissolved in anhydrous THF (50.0 mL). The clear yellow mixture was cooled to 0 °C prior to the addition of di-tert-butyl dicarbonate (4.00 mL, 17.50 mmol); upon which a thick, white precipitate formed inside the reaction vessel. The reaction mixture was kept under N2 flow as it initially stirred at 0 °C (ice bath) and gradually warmed to room temperature as the ice bath melted and the reaction proceeded over 24 h. Volatiles were removed by rotary evaporation under vacuum before reconstituting the mixture with EtOAc (100 mL). The organic solution was washed twice with 0.5 N HCl (2×100 mL), three times with brine solution (3×100 mL), and deionized water (100 mL) to achieve a neutral pH. The organic layer was dried over Na2SO4 before removing the solvent under reduced pressure to yield a mixture of the product as an off-white solid with negligible amounts of the Boc-deprotected starting material observed by 1H NMR (4.72 g, 106%). The crude mixture was purified by column chromatography to isolate the major component from the mixture (3.28 g, 74%). However, it was later determined that the crude product could be used without further purification for the added benefit of increased yields for the production of compound 2. 1H NMR of the crude mixture (500 MHz, [D6]DMSO, 25 °C, TMS):set A (anti-rotamers, anti-1) δ = 1.49 [s, 9H; NHCO2C(CH3)3], 3.88 (s, 6H; CO2CH3), 8.08 (t, J = 1.4 Hz, 1H; Ar-CH), 8.36 (m, 2H; Ar-CH), 9.89 (br, 1H; NH) ppm; set B [doubly Boc-protected derivative of compound 1, (Boc)2-1]: δ = 1.39 {s, 18H; N[CO2C(CH3)3]2}, 3.90 [s, 6H; CO2CH3], 8.01 (d, J = 1.5 Hz, 2H; Ar-CH), 8.41 (t, J = 1.5 Hz, 1H; Ar-CH) ppm; set C (tautomers of compound 1, tautomer-1): δ = 1.42 [s, 9H; NC(OH)OC(CH3)3], 3.91 (s, 6H; CO2CH3), 8.17 (d, J = 1.5 Hz, 1H; Ar-CH), 8.45 (t, J = 1.5 Hz, 2H; Ar-CH) ppm; 13C{1H} NMR of the crude mixture (126 MHz, [D6]DMSO, 25 °C, TMS): set A (anti rotamers, anti-1):δ = 28.1 [NHCO2C(CH3)3], 52.5 (CO2CH3), 79.9 [NHCO2C(CH3)3], 122.5, 123.0 (Ar-CH), 130.6 (Ar-C attached to CO2CH3), 140.7 [Ar-C attached to NHCO2C(CH3)3], 152.7 [NHCO2C(CH3)3], 165.4 (CO2CH3) ppm; Signals associated with set B [doubly Boc-protected derivative of compound 1, (Boc)2-1], and set C (imidic acid tautomers of compound 1, tautomer-1) of the crude mixture could not be distinguished from one another by 13C NMR techniques and are reported together herein. Crude mixture sets B and C: δ = 27.5, 27.5 {N[CO2C(CH3)3]2} and NC(OH)OC(CH3)3], 52.8, 52,8 (CO2CH3), 83.1, 84.4 {N[CO2C(CH3)3]2} and [NC(OH)OC(CH3)3], 128.3, 128.8 (Ar-CH), 131.0, 131.2 (Ar-C attached to CO2CH3), 133.0, 133.1 (Ar-CH), 139.5, 140.0 [Ar-C attached to {N[CO2C(CH3)3]2} and NC(OH)OC(CH3)3], 150.6, 150.8 {N[CO2C(CH3)3]2} and [NC(OH)OC(CH3)3], 164.7, 164.8 [(CO2CH32] ppm; HRMS (FDI): m/z calcd. for C15H19NO6: 309.11977[M]+; found: 309.11989; m/z calcd. for C20H27NO8+Boc: 409.17815[M+Boc]+; found: 409.17820; IR (FTIR): v~=3364.40 (w) [NHCO2C(CH3)3, N-H], 2980.61 (w), 2954.57 (w) [{N[CO2C(CH3)3]2} & NC(OH)OC(CH3)3, C-H], 2360.57 (w), 2339.30 (w) [NC(OH)OC(CH3)3, N = C], 1741.50 (m), 1726.07 (s), 1704.85 (s) [CO2CH3, C=O] 1604.57 (w), 1549.61 (m) cm−1 [{N[CO2C(CH3)3]2}, C=O].</p><p>The major component of the compound 1 product mixture, anti-rotamers (anti-1, set A), was isolated from the crude mixture for further analysis via precipitation. The crude product (2.51 g, 8.12 mmol) was dissolved completely in a minimal amount of warm DCM and stored at −20 °C overnight. The anti-rotamers of compound 1 (anti-1, set A) were precipitated under the aforementioned storage conditions and were isolated from the mother liquor as a shiny white precipitate after vacuum filtration. The mother liquor was collected after vacuum filtration and subjected to rotary evaporation to remove solvent residues. The residual solids were re-dissolved in minimal amounts of warm DCM to repeat the precipitation process. Precipitation of the anti-rotamers of compound 1 (anti-1, set A) was repeated several times until a pure product could no longer be isolated from the mother liquor solution (anti-1, set A: 1.56 g, 66%). 1H NMR of anti-1, set A (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.49 [s, 9H; NHCO2C(CH3)3], 3.88 (s, 6H; CO2CH3), 8.08 (t, J = 1.4Hz, 1H; Ar-CH), .8.35 (d, J = 0.86Hz, 2H; Ar-CH), 9.87 (br, 1H; NH) ppm; 13C{1H} NMR of anti-1, set A (126MHz, [D6]DMSO, 25 °C, TMS): δ = 28.0 [NHCO2C(CH3)3], 52.5 (CO2CH3), 79.9 [NHCO2C(CH3)3], 122.4, 123.0 (Ar-CH), 130.6 (Ar-C attached to CO2CH3), 140.6 [Ar-C attached to NHCO2C(CH3)3], 152.7 [NHCO2C(CH3)3], 165.4 (CO2CH3) ppm; HRMS (FDI): m/z calcd. for C15H19NO6: 309.11977[M]+; found: 309.12124; IR (FTIR): v~=3363.43 (m) [NHCO2C(CH3)3, N-H], 2952.65 (b) [NHCO2C(CH3)3, C-H], 1718.36 (m), 1703.89 (s) (CO2CH3, C=O), 1608.43 (m), 1545.75 (s) cm−1(Ar-CH).</p><!><p>Hydrazine hydrate (28.0 mL, 451 mmol) was added to a clear colorless solution of compound 1 (3.49 g, 11.28 mmol) in EtOH (150 mL) at room temperature. The color of the solution turned a pale-yellow. The mixture was refluxed at 90 °C for three days. All volatiles were removed under reduced pressure to yield quantitative amounts of the product, compound 2, in the form of a fine, matte-white powder (3.49 g, quantitative yield). The product material was used without further purification to proceed forward with the synthesis of compound 3. 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.48 [s, 9H; NHCO2C(CH3)3], 4.49 (br, 4H; CONHNH2), 7.74 (t, J = 1.4Hz, 1H; Ar-CH), 7.96 (d, J = 1.0Hz, 2H; Ar-CH), 9.64 (br, 3H; NH) ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 28.1 [NHCO2C(CH3)3], 79.5 [NHCO2C(CH3)3], 118.8, 119.6 (Ar-CH), 134.3 (Ar-C attached to CONHNH2), 139.8 [Ar-C attached to NHCO2C(CH3)3], 152.8 [NHCO2C(CH3)3], 165.8 (CONHNH2) ppm; HRMS (FDI): m/z calcd. for C13H19N5O4: 309.14491[M]+; found: 309.14370.</p><!><p>Potassium hydroxide (1.686 g, 29.16 mmol) was added to a suspension of compound 2 (4.14 g, 13.25 mmol) in EtOH (130 mL) and stirred for 10 minutes. Carbon disulfide (17.5 mL, 291.5 mmol) was added dropwise to this emulsion. The reaction mixture turned yellow followed by formation of a large amount of precipitate. The reaction mixture was heated to reflux at 90 °C for 16 h. Upon refluxing the reaction mixture became a clear pale-yellow solution with small amount of white precipitate. After cooling the reaction mixture to room temperature, EtOAc (320 mL) was added until complete dissolution of the precipitate material was achieved. The resulting clear yellow mixture was washed two times with 1 M HCl (2 × 320 mL) followed by washing 3 times with deionized water (3 × 320 mL) to achieve a neutral pH. The yellow organic layer was washed with brine (320 mL), then dried over Na2SO4. All volatiles were removed under reduced pressure; yielding compound 3 as shiny white solid (4.73 g, 91% yield). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.51 [s, 9H; NHCO2C(CH3)3], 7.82 (t, J = 1.5Hz, 1H; Ar-CH), 8.22 (d, J = 1.0Hz, 2H; Ar-CH), 10.05 (br, 1H; NH), 14.75 (br, 2H; SH) ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 28.0 [NHCO2C(CH3)3], 80.3 [NHCO2C(CH3)3], 116.2, 117.5 (Ar-CH), 124.2 (Ar-C attached to C2N2O), 141.5 [Ar-C attached to NHCO2C(CH3)3], 152.7 [NHCO2C(CH3)3], 159.4 (C2N2O attached to Ar), 177.5 (C2N2O attached to SH) ppm; HRMS (FDI): m/z calcd. for C15H15N5O4S2: 393.05703[M]+; found: 393.05654.</p><!><p>Triethylamine (3.39 mL, 24.2 mmol) was added to a clear yellow solution of compound 3 (2.63 g, 6.68 mmol) in dry THF (63.5 mL). After 10 minutes of stirring at room temperature, the color of the reaction mixture changed to a peach hue. To prevent light sensitive reagents from degrading, the reaction vessel was covered with aluminum foil. Iodomethane (1.11 mL, 17.8 mmol) was slowly added to the reaction mixture and reacted for 3 h at room temperature. During the first 1-2 minutes of stirring, the clear peach solution quickly became opaque with white precipitate. Once the reaction was completed, the THF solvent was removed in vacuo to afford a mixture of white and tan colored solids. A crude form of the product material was extracted from the mixture with EtOAc (500 mL). The solution volume was reduced to 1/3 by evaporating volatiles under reduced pressure. This mixture was washed with a 0.1 M aqueous solution of Na2CO3 (2 × 100 mL); (pH = 11). The deep-yellow organic phase was washed with brine (150 mL) and deionized water until a neutral pH was achieved. As the organic layer was neutralized, its coloring changed from deep-yellow to a clear pale yellow. The organic phase was dried over Na2SO4 and filtered before removing solvent under reduced pressure. Product 4 was obtained as an off-white powder (2.61 g, 93% yield). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.51 [s, 9H; NHCO2C(CH3)3], 2.79 (s, 6H; SCH3), 8.00 (t, J = 1.43 Hz, 1H; Ar-CH), 8.30 (m, 2H; Ar-CH), 9.99 (br, 1H; NH) ppm; 13C(1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 14.4 (SCH3), 28.0 [NHCO2C(CH3)3], 80.2 [NHCO2C(CH3)3], 116.9, 117.7 (Ar-CH), 124.7 (Ar-C attached to C2N2O), 141.5 [Ar-C attached to NHCO2C(CH3)3], 152.7 [NHCO2C(CH3)3], 164.2 (C2N2O attached to Ar), 165.3 (C2N2O attached to SCH3) ppm; HRMS (FDI): m/z calcd. for C17H19N5O4S2: 421.08627[M]+; found: 421.08784.</p><!><p>At 0 °C, mCPBA (365.0 mg, 1.6 mmol) was added slowly into a clear yellow solution of compound 4 (101.1 mg, 0.240 mmol) in dry DCM (6.0 mL). The reaction mixture stirred overnight at room temperature. It formed a suspension with a lot of precipitate. The mixture was quenched with an aqueous solution of NaHCO3. The aqueous phase was washed several times with DCM. The combined organic phase was washed with brine, dried over Na2SO4 and filtered before removing all volatiles under reduced pressure. The crude product was dissolved in a solvent mixture (17.0 mL) of MeCN:H2O (3:1) for HPLC purification. Compound 5 was isolated by semi-preparative RP-HPLC purification as a white solid (88.3 mg, 77%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.52 [s, 9H; C(CH3)3], 3.74 (s, 6H; SO2CH3), 8.26 (m, 1H; Ar-CH), 8.53 (m, 2H; Ar-CH), 10.16 (br, 1H; NHBoc) ppm. As compound 5 was only used for developing the optimized procedure, it was not fully characterized. The identity of compound 5 was confirmed with LR-mass spectrometry and 1H NMR spectroscopy.</p><!><p>Compound 5 (15.6 mg, 0.032 mmol) was dissolved in a 1:1 mixture of TFA:DCM (2 mL). It was stirred for 3 hours at room temperature. The reaction solvent was removed under reduced pressure while the majority of TFA was removed azeotropically by washing the reaction mixture several times with toluene. Compound 6 was obtained as a yellowish white solid (9.2 mg, 74%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 3.72 (s, 6H; SO2CH3), 6.19 (br, 2H; NH2), 7.57 (m, 2H; Ar-CH), 7.80 (m, 1H; Ar-CH) ppm. As compound 6 was only used for developing the optimized procedure, it was not fully characterized. The identity of compound 6 was confirmed with LR-mass spectrometry and 1H NMR spectroscopy.</p><!><p>Compound 4 (0.310 g, 0.712 mmol) was dissolved in a 1:1 mixture of TFA:DCM (16.0 mL). The mixture was stirred for 3 hours at room temperature. The reaction solvent was removed under reduced pressure while the majority of TFA was removed azeotropically by washing the reaction mixture several times with toluene. The crude product was washed with a 5:2 mixture of MeCN: H2O and centrifuged at 4700 rpm for 9 minutes; yielding an off-white pellet and clear yellow solution. The pellet was washed twice with deionized water (2 × 20 mL) and freeze-dried for 24 hours to afford compound 7 as a fine, off white powder (0.154 g, 67%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 2.77 (s, 6H; SCH3), 5.96 (br, 2H; NH2), 7.36 (d, J = 1.4 Hz, 1H; Ar-CH), 7.56 (d, J = 1.4 Hz, 2H; Ar-CH) ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 14.3 (SCH3), 110.5, 113.5 (Ar-CH), 124.7 (Ar-C attached to C2N2O), 150.3 (Ar-C attached to NH2), 164.8 (C2N2O attached to Ar), 164.8 (C2N2O attached to SCH3) ppm; HRMS (FDI): m/z calcd. for C12H11N5O2S2: 321.03545[M]+; found: 321.03542.</p><!><p>Compound 7 (49.2 mg, 0.153 mmol) and succinic anhydride (46.6 mg, 0.466 mmol) were dissolved in dry THF (5.0 mL). To this clear yellow solution was added triethylamine (44 μL, 0.315 mmol) at room temperature. The reaction mixture was submerged into an oil bath and heated to 45 °C with stirring for 72 hours at 45 °C. All volatiles were removed under reduced pressure. Deionized water (20.0 mL) was added to this solid residue and let stir for 2 minutes. It formed a suspension with white precipitate and colorless mother liquor. The white precipitate was further washed with deionized water until the mother liquor's pH = 7. The final precipitate was pelleted from the water via centrifugation, the water decanted, and the wet solid was freeze-dried for 24 hours to afford compound 8 as a white solid (48.7mg, 76%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 2.56 [m, 2H; NHCO(CH2)2COOH], 2.63 [m, 2H; NHCO(CH2)2COOH], 2.78 (s, 6H; SCH3), 8.09 (m, 1H; Ar-CH), 8.45 (m, 2H; Ar-CH), 10.60 [br, 1H; NHCO(CH2)2COOH], 12.21 [br, 1H; NHCO(CH2)2COOH] ppm. As compound 8 was only used for developing the optimized procedure, it was not fully characterized. The identity of compound 8 was confirmed with LR-mass spectrometry and 1H NMR spectroscopy.</p><!><p>A solution of HATU (45.7 mg, 0.120 mmol) and DIEA (0.05 mL, 0.287 mmol) in a 1:2 solvent mixture of DMF:DCM (3 mL) was added to a solution of compound 8 (43.3 mg, 0.103 mmol) in a 1:1 solvent mixture of DMF:DCM (2 mL). This solution was added to a solution of BocNH(PEG)3NH2 (32.6 mg, 0.111 mmol) in DCM (2.2 mL). The reaction mixture was stirred at room temperature for 18 hours followed by stirring at 50 °C for 3 hours. All volatiles were removed under reduced pressure to form a brownish residue. To this residue was added deionized water (1 mL) and MeCN (5 mL) followed by vigorous stirring for 2 minutes. A large quantity of precipitate was formed with a brown mother liquor remaining. The mother liquor was decanted, and the precipitate was washed several times with MeCN until the mother liquor remained colorless. The final precipitate was dried under high vacuum, which afford compound 9 (13.0 mg) as a white solid. All volatiles were removed from mother liquor mixture. It was re-dissolved in a solvent mixture (13.0 mL) of MeCN:H2O (5:8) for a HPLC purification. More of compound 9 (10.1 mg) by semipreparative RP-HPLC purification was isolated as the major component of the mother liquor mixture by HPLC purification. 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 2.80 (s, 6H; SCH3), 2.82 [s, 4H; N(CO)2(CH2)2], 8.14 (m, 2H; Ar-CH), 8.44 (m, 1H; Ar-CH) ppm. As compound 9 was only used for developing the optimized procedure, it was not fully characterized. The identity of compound 9 was confirmed with LR-mass spectrometry and 1H NMR spectroscopy.</p><!><p>EEDQ (19.1 mg, 0.077 mmol) was added to a solution of compound 8 (27.7 mg, 0.066 mmol) in a 1:1 solvent mixture of DMF:DCM (3 mL). BocNH(PEG)3NH2 (16.1 mg, 0.055 mmol) was added to this mixture after stirring for 50 minutes at room temperature. The reaction mixture was stirred at room temperature for 96 hours. All volatiles were removed under reduced pressure. The solid residue was dissolved in DCM (1.0 mL) with 10% MeOH (0.1 mL) for automated flash column chromatography [Biotage column with silica; 100%MeCN→DCM+10%MeOH→DCM:MeOH (1:1)]. After column chromatography, three compounds were separated. The first collected fraction gave compound 9 (2.3 mg, 9%), the second collected fraction afforded the desired product, compound 10 (6.8 mg, <18%; note: product contained quinoline, a side product, as an impurity), and the third collected fraction was starting material, compound 8 (9.0 mg, 33%). 1H NMR of compound 10 (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.36 [s, 9H; C(CH3)3], 2.45 [m, 2H, NH(CO)(CH2)2CONH(OC2H4)3NHBoc], 2.60 [m, 2H, NH(CO)(CH2)2CONH(OC2H4)3NHBoc], 2.79 (s, 6H; SCH3), 3.05 (m, 2H; CH2 of PEG), 3.20 (m, 2H; CH2 of PEG), 3.36 (m, 2H; CH2 of PEG), 3.40 (m, 2H; CH2 of PEG), 3.46-3.52 (m, 8H; CH2 of PEG), 6.76 [br, 1H, NH(CO)(CH2)2CONH(OC2H4)3NHBoc], 8.09 (m, 1H; Ar-CH), 8.45 (m, 2H; Ar-CH), 8.91 [br, 1H, NH(CO)(CH2)2CONH(OC2H4)3NHBoc], 10.52 [br, 1H, NH(CO)(CH2)2CONH(OC2H4)3NHBoc] ppm. As compound 10 was used only for developing the optimized procedure, it was not fully characterized. The identity of compound 10 was confirmed with LR-mass spectrometry and 1H NMR spectroscopy.</p><!><p>In a 50 mL Schlenck flask, compound 7 (105.7 mg, 0.329 mmol), EDC (86.0 mg, 0.45 mmol), and oxyma (70.6 mg, 0.497 mmol) were placed under vacuum for 30 minutes. The vessel was subsequently purged with N2 gas and the reagent mixture was dissolved with anhydrous DMF (10.0 mL) to form a clear light-yellow solution. BocNH(PEG)4COOH (108.6 mg, 0.297 mmol) was added to the reaction mixture at room temperature followed by addition of triethylamine (0.125 mL), which led to a change in color as the solution changed from yellow to orange. The mixture was heated at 50 °C for 48 h under N2 flow. All volatiles were removed under high vacuum following the reaction and the resulting orange residue was dissolved in a solvent mixture (30.0 mL) of MeCN:H2O (1:1) for a semi-preparative RP-HPLC purification. Compound 11 was isolated by HPLC purification as a sticky pale yellow solid (110.7 mg, 55%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.35 [s, 9H; C(CH3)3], 2.62 (t, J = 5.9 Hz, 2H; CH2 of PEG), 2.79 [s, 6H; S(CH3)], 3.03 (q, J = 5.8 Hz, 2H; CH2 of PEG), 3.31-3.35 (m, 1H; CH2 of PEG), 3.42-3.55 (m, 13H; CH2 of PEG), 3.73 (t, J = 5.9Hz, 2H; CH2 of PEG), 6.73 (m, 1H; NHBoc), 8.09 (t, J = 1.5 Hz, 1H; Ar-CH), 8.46 (d, J = 1.5 Hz, 2H; Ar-CH), 10.52 (s, 1H; Ar-NH) ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 14.4 [S(CH3)], 28.2 [C(CH3)3], 37.3, 40.0, 66.4, 69.1, 69.5, 69.67, 69.70, 69.73, 69.8 (CH2 of PEG), 77.6 [C(CH3)3], 117.9, 118.7 (Ar-CH), 124.8 (Ar-C attached to C2N2O), 140.9 [Ar-C attached to NH(CO)PEG4], 155.6 [PEG4NH(CO)OC(CH3)3], 164.2 (C2N2O attached to Ar), 165.4 [C2N2O attached to S(CH3)], 170.2 [NH(CO)PEG4NH(CO)OC(CH3)3] ppm; HRMS (TOF): m/z calcd. for C28H40N6O9S2+Na+: 691.2217 [M+Na]+; found: 691.2190.</p><!><p>At 0 °C, mCPBA (114.8 mg, 0.670 mmol) was added slowly into a clear yellow solution of compound 11 (101.8 mg, 0.150 mmol) in dry DCM(4.0 mL). The reaction mixture stirred overnight at room temperature. It formed a clear colorless solution. The mixture was washed with 0.1 M aqueous solution of NaOH (6.0 mL). The organic phase was washed several times with deionized water until a neutral pH was achieved. The organic phase was dried over Na2SO4 and filtered before removing all volatiles under reduced pressure. Compound 12 was obtained as a glassy colorless solid (79.9 mg, 73% yield). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 1.36 [s, 9H; C(CH3)3], 2.65 (t, J = 6.0 Hz, 2H; CH2 of PEG), 3.03 (q, J = 6.0 Hz, 2H; CH2 of PEG), 3.30 (s, 3H; SO2CH3), 3.31-3.37 (m, 1H; CH2 of PEG), 3.43-3.56 (m, 13H; CH2 of PEG), 3.74 (s, 3H; SO2CH3), 3.75 (m, 2H; CH2 of PEG), 6.74 [t, J = 5.2 Hz, 1H; ArNH(CO)PEG4NHBoc], 8.34 (m, 1H; Ar-CH), 8.66 (m, 1H; Ar-CH), 8.68 (m, 1H; Ar-CH), 10.66 [m, 1H; ArNH(CO)PEG4NHBoc] ppm; note: 1H NMR sample contained small amount of mCPBA with signals at 7.54, 7.70, 7.87-7.91, 13.34 ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 28.2 [C(CH3)3], 37.3 (CH2 of PEG), 43.1 (SO2CH3), 54.9, 66.3, 69.1, 69.5, 69.68, 69.72, 69.8 (CH2 of PEG), 77.6 [C(CH3)3], 119.6, 119.8 (d, Ar-CH, J = 32.0 Hz), 120.5, 120.7 (d, Ar-CH, J = 32.0Hz), 124.20 (Ar-C attached to C2N2O), 124.5 (Ar-CH), 141.2 (Ar-C attached to NH(CO)PEG4NHBoc ), 155.6 [ArNH(CO)PEG4NH(CO)OC(CH3)3], 162.4 (C2N2O attached to Ar), 164.7, 164.8 (d, C2N2O attached to SO2CH3, J = 7.0Hz), 164.9, 165.0 (d, C2N2O attached to SO2CH3, J = 6.0Hz), 170.4 [ArNH(CO)PEG4NH(CO)OC(CH3)3] ppm; note: 13C NMR sample contained small amount of mCPBA with signals at 120.2, 127.9, 128.8, 130.7, 167.7 ppm; 13C NMR sample contained small amount of mCPBA with signals at 127.9, 128.8, 130.7, 167.7 ppm; HRMS (TOF): m/z calcd. for C28H40N6O13S2+Na+: 755.1960 [M+Na]+; found: 755.1987.</p><!><p>Compound 12 (64.9 mg, 0.089 mmol) was dissolved in a 1:1 mixture of TFA:DCM (6 mL). It was stirred for 3 hours at room temperature to facilitate Boc deprotection. The reaction solvent was removed under reduced pressure while the majority of TFA was removed azeotropically by washing the reaction mixture several times with toluene. The crude product was dissolved in deionized water (7.0 mL) and EtOAc (4.0 mL). The aqueous phase was washed two times with EtOAc (2×4.0 mL). The aqueous phase was freeze dried for 24 hours to afford DiPODS as a fine, off white solid (50.6 mg, 90%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 2.65 (t, J = 6.0 Hz, 2H; CH2 of PEG), 2.96 (q, J = 5.2 Hz, 2H; CH2 of PEG), 3.27-3.31 (m, 1H; CH2 of PEG), 3.30 (s, 3H; SO2CH3), 3.46-3.58 (m, 14H; CH2 of PEG), 3.75 (s, 3H; SO2CH3), 3.73-3.77 (m, 1H; CH2 of PEG), 7.72 (br, 2H; ArNH(CO)PEG4NHBoc), 8.34 (m, 1H; Ar-CH), 8.66 (m, 1H; Ar-CH), 8.68 (m, 1H; Ar-CH), 10.68 (m, 1H; ArNH(CO)PEG4NH2) ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 27.4, 37.3, 38.6 (CH2 of PEG), 43.1 (SO2CH3), 56.0, 66.3, 66.7, 69.59, 69.66, 69.69, 69.73 (CH2 of PEG), 119.6, 119.8 (d, Ar-CH, J = 32.5 Hz), 120.5, 120.7 (d, Ar-CH, J = 32.5 Hz), 124.2 (Ar-CH), 124.5 (Ar-C attached to C2N2O), 140.9, 141.2 [d, Ar-C attached to NH(CO)PEG4NH2, J = 27.0 Hz], 162.4 (C2N2O attached to Ar), 164.7 (C2N2O attached to SO2CH3), 170.4 [ArNH(CO)PEG4NH2] ppm; note: 13C NMR sample contained small amount of mCPBA with signals at 127.9, 128.8, 130.7, 132.7, 166.1, 167.8 ppm; HRMS (TOF): m/z calcd. for C23H32N6O11S2+H+: 633.1646 [M+H]+; found: 633.1643.</p><!><p>Diisopropylethylamine (0.11mL, 0.63mmol) was added into a clear colorless solution of DiPODS (72.0 mg, 0.114 mmol) in dry DMF (7.0 mL). This solution was stirred for 10 minutes before adding it dropwise into a solution of dye (48.6 mg, 0.125 mmol) in dry DMF (1.3 mL). The reaction vessel was covered with aluminum foil and it was stirred overnight at room temperature. All volatiles were removed under high vacuum. The obtained orange residue was re-dissolved in 40% MeCN and water (46.0 mL) for semi-preparative RP-HPLC purification. The product, DiPODS-FITC, was obtained as an orange fluffy solid (50.9 mg, 44%). 1H NMR (500 MHz, [D6]DMSO, 25 °C, TMS): δ = 2.63 (t, J = 6.4 Hz, 2H; CH2 of PEG), 3.29 (s, 6H; SO2CH3), 3.48-3.54 (12H; CH2 of PEG), 3.56 (m, 2H; CH2 of PEG), 3.65 (br, 2H; CH2 of PEG), 3.73 (t, J = 6.4 Hz, 2H; CH2 of PEG), 6.48-6.66 (br, 8H; from FITC dye), 7.14 (d, J = 8.5Hz, 1H; FL-dye), 7.72 [br, 1H; ArNH(CO)PEG4NH(CS)NH attached to FL-dye], 8.23 (br, 2H; OH of FL-dye), 8.32 (t, J = 1.4Hz, 1H; Ar-CH), 8.65 (d, J = 1.4Hz, 2H; Ar-CH), 10.18 [s, 1H; ArNH(CO)PEG4NH(CS)NH attached to FL-dye], 10.76 [s, 1H; ArNH(CO)PEG4NH(CS)NH attached to FL-dye] ppm; 13C{1H} NMR (126 MHz, [D6]DMSO, 25 °C, TMS): δ = 20.7, 37.3, 42.6 (CH2 of PEG), 43.7 (SO2CH3), 66.4, 68.4, 69.6-69.8 (CH2 of PEG), 102.1 (Ar-CH of FL-dye), 102.3 (Ar-CH of FL-dye), 119.3 (Ar-CH of FL-dye), 120.2 (Ar-CH of FL-dye), 124.4 (Ar-C of DiPODS attached to C2N2O), 129.2 (Ar-CH of FL-dye) and (Ar-C of FL-dye), 141.2 (Ar-C of FL-dye) and [Ar-C attached to NH(CO)PEG4NH-FL-dye], 152.4 (Ar-C of FL-dye), 165.0 (C2N2O attached to Ar), 167.7 (C2N2O attached to SO2CH3), 168.6 (CO of FL-dye), 170.4 [ArNH(CO)PEG4NH-FL-dye], 180.5 [ArNH(CO)PEG4NH(CS)NH attached to FL-dye] ppm; HRMS (TOF): m/z calcd. for C44H43N7O16S3-2O+N+: 1012.1951 [M-2O+Na]+; found: 1012.1914.</p><!><p>Two NMR samples were prepared in [D6]DMSO in 5 mm Wilmad High throughput borosilicate NMR tubes. As described above, compound 1 contained a mixture of components, which were partially separated by successive precipitations out of cold DCM. The crude compound 1 mixture was separated into 1) solid precipitate containing set A (anti-rotamers), and 2) the components remaining in the mother liquor, which were dried under vacuum. The first NMR sample contained 5.21 mg of the residue isolated from the mother liquor, which was composed of a mixture of set A (anti-rotamers of compound 1, anti-1), set B [doubly Boc-protected derivative of compound 1, (Boc)2-1], and set C (tautomers of compound 1, tautomer-1). The second NMR sample contained the precipitate, anti-rotamers of compound 1 (set A, anti-1) (5.21 mg, 00168 mmol). Both samples were dissolved in [D6]DMSO (0.41 mL and 0.40 mL respectively) immediately before subjecting them to 1H NMR analysis at 25 °C (298.15 K) on a Bruker Avance III HD 600 MHz spectrometer, with 16 scans collected every 10 minutes for 1 hour. Once the last data point was obtained at 25 °C (298.15 K), the temperature was increased to 75 °C (348.15 K) and stabilized for 10 minutes before re-tuning, re-shimming, and re-locking the system onto the deuterated solvent signal for the acquisition of 16 scans every 10 minutes for 1 hour. This process was repeated for each sample at 80 °C (353.15 K), 85 °C (358.15 K), and 90 °C (363.15 K). Calculations implementing the experimental data were based on the integrated values of the peaks located in the designated regions of the 1H NMR spectra. The integration values of peaks located at 8.09 (t, J = 1.4Hz, 1H; Ar-CH), 8.41 (t, J = 1.4Hz, 1H; Ar-CH), and 8.45 (t, J = 2.9Hz, 1H; Ar-CH) ppm were used for assessing the proportions of set A (anti-rotamers of compound 1, anti-1), set B [doubly Boc-protected derivative of compound 1, (Boc)2-1], and set C (imidic acid tautomer of compound 1, tautomer-1), respectively. Integration values of peaks located at 7.88 (t, J = 1.5Hz, 1H; Ar-CH), 8.10 (t, J = 1.5Hz, 1H; Ar-CH), 8.14 (t, J = 1.5Hz, 1H; Ar-CH), 8.42 (t, J = 1.5Hz, 1H; Ar-CH), and 8.46 (t, J = 2.8Hz, 1H; Ar-CH) ppm were used for assessing the proportions of the Boc-deprotected derivative of compound 1, set A (anti-rotamers of compound 1, anti-1), set D (syn-rotamers of compound 1, syn-1), set B [doubly Boc-protected derivative of compound 1, (Boc)2-1], and set C (tautomer of compound 1, tautomer-1), respectively. The data acquired from these experiments are summarized in Table S1.</p><p>In order to better understand the NMR spectrum of the product mixture of compound 1, a series of NMR spectra were collected at different temperatures. Two NMR samples were prepared from the components of crude compound 1. The first contained only the precipitate — i.e. the anti-rotamers (Figure 2C, set A). The second contained the mixture isolated from the mother liquor following precipitation (Figure 2D). The latter is composed mostly of the compounds responsible for sets B and C but also some of the anti-rotamer (set A). While the initial 1H-NMR of the sample of the anti-rotamers displayed only the peaks of set A, that of the mother liquor mixture showed the peaks of sets A, B, and C in a ratio of 1.0:1.5:1.6, respectively. The 1H-NMR spectra for both samples were collected again after 24 h at room temperature where no new peaks were observed in either sample and no change in the ratio between the peaks was observed in the spectrum of the mother liquor mixture.</p><p>Subsequently, variable temperature 1H-NMR spectra of the sample containing the mother liquor mixture were collected every 10 minutes for 1 hour at 25, 75, 80, 85, and 90 °C. Upon heating the mother liquor mixture sample to 90 °C, the proportion of peaks from set A increased while those of sets B and C decreased (Figure 2E, Table S1). Indeed, the proportion of the peaks in set C shrank to nearly zero (Table S1 and Figures S35 and S36). We also observed the appearance of two new sets of peaks in the sample at this elevated temperature. The first new set — set D — contained signals for all of the functional groups of compound 1 and was assigned to the syn-rotamers. The second new set was assigned to the Boc-deprotected variant of compound 1. Importantly, however, the signals of set D were not significant, suggesting that sets B and C were converted to set A at the elevated temperature (Table S1).</p><p>Ultimately, set B was attributed to a doubly Boc-protected version of compound 1 based on the integration ratio between the methyl ester (6) and tert-butyl (18) protons as well as the presence of a tertiary amine group with no proton. High resolution mass spectrometry subsequently confirmed this assignment. As removing the first of two Boc protecting groups is easier than the second, it is likely that the doubly protected compound is converted to compound 1 at elevated temperatures. Set C, in contrast, has an integration ratio of 6:9 between the methyl ester (6) and tert-butyl (9) protons, confirming that the compound responsible for these peaks has a single Boc group. However, no proton associated with the amine was observed. Furthermore, upon heating at 90 °C set C disappeared almost entirely. We propose that this phenomenon can be explained by a tautomerization reaction involving the transfer of a proton from the amine to the neighboring oxygen (Figure 3). The assignment of set C as a tautomeric form of set A would explain why the integration of the former matches that of the latter except for the absence of the proton from the amine group.</p><!><p>All calculations were performed with the Gaussian 16 software package revision B.01.37 The B3LYP exchange-correlation functional with the polarized diffuse split-valence 6-311+G(d,p) basis set was used for all geometry optimizations, transition state calculations, and infrared frequency calculations. For calculating the solvent effects the integral equation formalism model (IEFPCM)40 (DMSO) was used.41 For the molecular cavities the united atom topological model for Hartree-Fock (UAHF) was used.42 Before performing calculations at a high level of theory with the Gaussian 16 software, two calculations were performed for each compound using the Spartan 14 software package. First a large set of conformers were generated using the Merck Molecular Force Fields (MMFF). Then structures with the lowest energies (within 3.0 kcal/mol energy difference) were optimized at the HF/3-21G level of theory. Then the lowest energy conformers (within 3 kcal/mol energy difference) were submitted to Gaussian 16 for full optimization at the level of theory described above. The frequency calculations were used to evaluate conformers' zero-point vibrational energy (ZPVE) and thermal corrections at 298.15 K. For all ground state conformers all harmonic vibrational frequencies were positive, confirming that they were local minima. For thermodynamic evaluation of ethanethiol conjugations, all reactions were computed with UAHF for improved solvent modeling, and all reactions were performed as isodesmic so total energy differences could be compared. For P2, the most stable optimized geometry from a separate geometry optimization calculation without UAHF was used in a single point energy calculation with UAHF. All graphical depictions of the structures were generated using CYLview software43 using coordinates from the optimized log file.</p><!><p>To a suspension of 150 μg of FabHER2 in PBS pH 7.4 (1.19 mg/mL) was added the appropriate volume of a fresh TCEP solution (10 mM in water) to reduce the interchain disulfide bridge. The reaction mixture was stirred on a thermomixer (25 °C or 37 °C) for 2 hours. The reductant was then removed using centrifugal filtration units with a 3,000 Da molecular weight cut off (Amicon™ Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA) at 4 °C into fresh PBS pH = 7.4. The reaction mixture was immediately stored at 4 °C to prevent the re-oxidation of the interchain disulfide bond.</p><!><p>To a suspension of 150 μg of Fabns in PBS pH 7.4 (14.64 mg/mL) was added 5.99 μL of a fresh TCEP solution (10 mM in water). The reaction mixture was stirred on a thermomixer (25 °C or 37 °C) for 2 hours. The reductant was then removed using centrifugal filtration units with a 3,000 Da molecular weight cut off (Amicon™ Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA) at 4 °C into fresh PBS pH = 7.4. The reaction mixture was immediately stored at 4 °C.</p><!><p>To a suspension of 150 μg of FabHER2 in PBS pH 7.4 (1.19 mg/mL) was added 5.99 μL of a fresh TCEP solution (10 mM in water, 20 eq.).The reaction mixture was stirred on a thermomixer (37 °C) for 2 hours. The reductant was then removed using centrifugal filtration units with a 3,000 Da molecular weight cut off (AmiconTM Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA) at 4 °C into fresh PBS pH = 7.4. The appropriate volume of a DiPODS-FITC solution (10 mM in DMSO) was immediately added to the reduced FabHER2. The reaction mixture was stirred on a thermomixer (25 °C or 37 °C) for 2, 4, 8, 16, and 24 h. The conjugate was then purified on a size exclusion column (Sephadex G-25 M, PD-10 column, GE Healthcare; dead volume = 2.5 mL, eluted with 2 mL of PBS, pH 7.4) and concentrated using centrifugal filtration units with a 30,000 Da molecular weight cut off (Amicon™ Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA).</p><!><p>To a suspension of 150 μg of Fabns in PBS pH 7.4 (14.64 mg/mL) was added 5.99 μL of a fresh TCEP solution (10 mM in water, 20 eq.). The reaction mixture was stirred on a thermomixer (37 °C) for 2 hours. The reductant was then removed using centrifugal filtration units with a 3,000 Da molecular weight cut off (AmiconTM Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA) at 4 °C into fresh PBS pH = 7.4. 4.50 μL of a DiPODS-FITC solution (10 mM in DMSO) was immediately added to the reduced Fabns. The reaction mixture was stirred on a thermomixer (37 °C) for 16 hours. The conjugate was then purified on a size exclusion column (Sephadex G-25 M, PD-10 column, GE Healthcare; dead volume = 2.5 mL, eluted with 2 mL of PBS, pH 7.4) and concentrated using centrifugal filtration units with a 30,000 Da molecular weight cut off (Amicon™ Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA).</p><!><p>Bioconjugation conditions were adjusted in order to obtain a DOL of ~1 for the final conjugate. To a suspension of 150 μg of FabHER2 in PBS pH 7.4 (1.19 mg/mL) was added 12.5 μL of a fresh Na2CO3 solution (0.1 mM in water) to adjust pH to 9.0. 1.13 μL of a NCS-FITC solution (10 mM in DMSO, 3.75 eq.) was then added. The reaction mixture was stirred on a thermomixer (37 °C) for 1 h. The conjugate was then purified on a size exclusion column (Sephadex G-25 M, PD-10 column, GE Healthcare; dead volume = 2.5 mL, eluted with 2 mL of PBS, pH 7.4) and concentrated using centrifugal filtration units with a 30,000 Da molecular weight cut off (Amicon™ Ultra 0.5 Centrifugal Filtration Units, Millipore Corp. Billerica, MA).</p><!><p>UV-Vis measurements were taken on a Shimadzu BioSpec-Nano Micro-Volume UV-Vis Spectrophotometer (Shimadzu, Kyoto, Japan). The fluorescein to Fab ratio was determined via UV-Vis spectrophotometry of the conjugates at 280 nm and 495 nm followed by calculation using the following equation: AbsFab=Abs280−(Abs495∗CF)DOL=[Absmax∗MWFab]∕[[Fab]∗εDye495] in which the correction factor (CF) for DiPODS-FITC was 0.34 based on the absorbance spectrum of DiPODS-FITC in PBS, MWFab = 50,000, εDye495 = 75,000, and ε280, mAb = 69,000.</p><!><p>To a suspension of 150 μg of Fab in PBS pH 7.4 was added 50 μL of a fresh solution of the Ellman reagent44 (10 mM in DMSO/water, 25% v/v) and the appropriate volume of PBS to obtain a final volume of reaction mixture of 250 μL. The reaction mixture was stirred on a thermomixer protected from light for 30 min. The sulfhydryl to Fab ratio was determined via UV-vis spectrophotometry of the Fab mixture at 412 nm followed by calculation using the following equation: Ratio -SH/Fab={[(Abs412∕εDTNB412)∗Vreaction]∕VFab}∗106∕[Fab] in which εDTNB412 = 14,150 M−1cm−1, Vreaction (L) = 250 (μL)*10−6 and VFab is the volume for 150 μg of Fab in liters.</p><!><p>5 μg of Fab (5 μL of a 1.0 mg/mL stock) was combined with 18.5 μL H2O, and 7.5 μL 4X electrophoresis buffer (NuPAGE® LDS Sample Buffer, Thermo Fisher, Eugene, OR). This mixture was then denatured by heating to 60 °C for 5 min using an agitating thermomixer. Subsequently, 20 μL of each sample was then loaded alongside an appropriate molecular weight marker (Novex® Sharp Pre- Stained Protein Ladder, Life Technologies) onto a 1 mm, 10 well 4-12% Bis-Tris protein gel (Life Technologies) and run for ~ 4 h at 80 V in MOPS buffer. The completed gel was washed 3 times with H2O, stained using SimplyBlueTM SafeStain (Life Technologies) for 1 h, then destained overnight in H2O. The gel was then analyzed using an Odyssey CLx Imaging system (LI-COR Biosciences, Lincoln, NE). Fluorescence signal was analyzed using a Typhoon™ FLA 7000 Imaging system (GE Healthcare, USA).</p><!><p>Analytical size-exclusion chromatography (A-SEC) of the FabHER2-DiPODS-FITC was conducted on a Shimadzu UFLC System with a SPD-20A UV-Vis Detector, CBM-20A system controller, DGU-20A3R degassing unit and LC-20AB binary pump using a Superdex™ 200 Increase 10/300 GL column (GE Healthcare, USA) with a flow rate of 0.75 mL/min. Elution of conjugates have been achieved during a 45 minute isocratic gradient using a phosphate buffer at pH 7.4 as a mobile phase. UV chromatograms were recorded at 280 nm.</p><!><p>In an Eppendorf-tube, 150 μL human serum albumin (HSA) were mixed with 50 μL FabHER2-DiPODS-FITC (2.0 mg/ml) for each sample individually to give a final solution of 0.5 mg/ml conjugates in 75% HSA serum. Samples were incubated at 37 °C for 0, 1, 2, 3, 4, 5, 6 and 7 days. Samples (50 μL) were quenched with extraction buffer (100 μL) consisting of 50% ACN/H2O, 0.1 M NaCl, and 1% TFA, chilled on ice for 5 min, centrifuged (14000 rpm, 10 min) and analyzed (2 x 25 μL injection) by SEC-HPLC. Samples were prepared and analyzed in triplicate.</p><!><p>As a means by which to quantify the number of cargoes per antibody, the Fab conjugates were analyzed by MALDI-ToF MS/MS using a Bruker Ultraflex MALDI-ToF/ToF (Bruker Daltonic GmbH). 1 μL of each sample (1 mg/mL) was mixed with 1 μL of sinapic acid (10 mg/ml in 50% acetonitrile:water and 0.1% trifluoroacetic acid). 1 μL of the sample/matrix solution was spotted onto a stainless steel target plate and allowed to air dry. All mass spectra were obtained using a Bruker Ultraflex MALDI-ToF/ToF (Bruker Daltonic GmbH). Ions were analyzed in positive mode and external calibration was performed by use of a standard protein mixture (Bovine Serum Albumin). Samples were prepared and analyzed in triplicate. The degree of labeling (DOL) of FabHER2-DiPODS-FITC was calculated from MALDI-ToF spectra analysis using the following equation: DOL=(m/zFabHER2-DiPODS-FITC - m/zFabHER2)∕(m/zDiPODS-FITC) in which m/z DiPODS-FITC is ~ 927. After calculation, a DOL of ~ 1 was obtained.</p><!><p>All circular dichroism (CD) spectroscopy was performed using a Chirascan™ V100 (Applied Photophysics Ltd, UK). Temperature was controlled via the Pro-Data Chirascan™ program and monitored with the Chirascan™ CS/PSM Turret T1 temperature probe. The far UV CD spectra of FabHER2, reduced FabHER2, and FabHER2-DiPODS-FITC (0.12 mg/mL in PBS Buffer, pH 7.4) were taken from 200-280 nm at 5 °C in 1 mm optical path quartz cuvettes. Spectra values were corrected by subtracting buffer baselines determined in the same cuvette, and the adjusted values were then converted to mean residue molar ellipticity (MRE).</p><p>Thermal stability experiments were performed by taking CD spectra from 200-280 nm at 0.5 degrees increments from 5-80 °C with an equilibration time of 2 min at each temperature. The data were subsequently analyzed using SciDAVis software in which thermal unfolding transition profiles for each fragment were obtained using CD signals at 205 nm. To this end, MRE values were converted to fraction folded, plotted against temperature, and fitted to a Boltzmann (Sigmoidal) curve. The melting temperatures — i.e. the temperature at which the faction folded is 0.5 — were then extrapolated from the curves.</p><!><p>Flow cytometry experiments were performed with HER2-positive BT474 cells. Modified Fab conjugates — FabHER2-DiPODS-FITC and FabHER2-Lys-FITC — were incubated at 0.8 mg/ml in suspension with 106 cells/ml, for 30 min on ice. Cells were washed by pelleting and resuspension three times, then analyzed on a BD LSR-II (BD Biosciences). Samples were prepared and analyzed in triplicate.</p>

PubMed Author Manuscript

The Kinetic Mechanism for DNA Unwinding by Multiple Molecules of Dda Helicase Aligned on DNA\xe2\x80\xa0

Helicases catalyze the separation of double-stranded nucleic acids to form single-stranded intermediates. Using transient state kinetic methods we have determined the kinetic properties of DNA unwinding under conditions that favor a monomeric form of the Dda helicase as well as conditions that allow multiple molecules to function on the same substrate. Multiple helicase molecules can align like a train on the DNA track. The number of base pairs unwound in a single binding event for Dda is increased from ~19 bp for the monomeric form to ~64 bp when as many as four Dda molecules are aligned on the same substrate, while the kinetic step-size (3.2 \xc2\xb1 0.7 bp) and unwinding rate (242 \xc2\xb1 25 bp s\xe2\x88\x921) appear to be independent of the number of Dda molecules present on a given substrate. The data support a model in which the helicase molecules bound to the same substrate move along the DNA track independently during DNA unwinding. The observed increase in processivity arises from the increased probability that at least one of the helicases will completely unwind the DNA prior to dissociation. These results are in contrast to previous reports in which multiple Dda molecules on the same track greatly enhanced the rate and amplitude for displacement of protein blocks on the track. Therefore, only when the progress of the lead molecule in the train is impeded by some type of block, such as a protein bound to DNA, do the trailing molecules interact with the lead molecule in order to overcome the block. The fact that trailing helicase molecules have little impact on the lead molecule in the train during routine DNA unwinding suggests that the trailing molecules are moving at similar rates as the lead molecule. This result implicates a step in the translocation mechanism as contributing greatly to the overall rate-limiting step for unwinding of duplex DNA.

the_kinetic_mechanism_for_dna_unwinding_by_multiple_molecules_of_dda_helicase_aligned_on_dna\xe2\x80

<!>Materials<!>Helicase Substrates<!>Single-Turnover RQF Helicase Unwinding Experiments<!>Non-linear Least Squares Analysis of Unwinding Data<!>Data Analysis by Kintek Global Explorer<!>DNA unwinding by monomeric Dda helicase<!>Protein-protein interactions are not required to account for increased processivity during DNA unwinding by Dda<!>The kinetic step size for multiple Dda molecules is similar to that of monomeric Dda<!>Dda-catalyzed unwinding of DNA substrates with \xe2\x89\xa5 21 nt ssDNA overhang lengths<!>The stepwise kinetic mechanism for multiple Dda molecules proceeds through a non-uniform \xe2\x80\x98pause\xe2\x80\x99 characterized by an intermediate species<!>Functional cooperativity without protein-protein interactions can account for the increased processivity exhibited by multiple Dda molecules<!>A non-uniform step in the kinetic mechanism is observed for multiple Dda molecules<!>The alignment of Dda molecules on DNA has a greater impact when translocation or DNA unwinding is impeded when compared to routine unwinding of dsDNA<!>

<p>The nature of nucleic acid metabolism requires enzymes that are able to catalyze the separation of double stranded helices to transiently form single stranded intermediates for the purpose of cellular events such as replication, recombination, repair, transcription, translation, and splicing. Helicases fulfill such a requirement by coupling energy associated with NTP hydrolysis to the manipulation of nucleic acid structure (1-5). They are ubiquitous in nature and function in coordination with highly regulated macromolecular complexes (1). The manner in which helicases achieve strand separation appears to be a variation upon a common theme, in which molecular motors power translocation and separation in a directionally biased manner. Different helicases are likely to have different mechanistic features. For some helicases, translocation has been described in terms of an inch-worm mechanism, in which strand-separation may take a form analogous to a 'snow-plow' or 'wire-stripper' (6). Other helicases are proposed to actively interact with the duplex region at a single-strand/double-strand junction to melt the DNA (7). At least one helicase, RecBCD, utilizes two molecular motors to translocate along each strand of the duplex during DNA unwinding (8-10). One motor moves 5′-to-3′ while the second motor moves 3′-to-5′, thereby ensuring very robust separation of the DNA. Variation in helicase mechanisms are also revealed by the oligomeric state of the functional enzyme, which can include monomers (11;12), dimers (13;14), and hexamers (15-19).</p><p>Unwinding of duplexes of varying length has led to several descriptors of the kinetic and physical constants associated with helicases. One of the most confusing values relates to the 'step size'. The kinetic step size refers to the number of base pairs unwound prior to a rate limiting kinetic step. The physical step size refers to the number of base pairs that are unwound simultaneously. A helicase might unwind one base pair at a time (physical step of one), but then proceed through a slow conformational change that occurs every three base pairs, resulting in a kinetic step size of 3 bp. The chemical step size refers to the number of base pairs unwound per ATP hydrolyzed. In the simplest case, all of these values are equal to one.</p><p>A fundamental component of the kinetic mechanism describing helicase action is the number of base pairs that are unwound during a rate-limiting kinetic step. A method to measure this number, termed the kinetic step size, was developed by the Lohman laboratory, which has reported kinetic step sizes of 4 bp for the UvrD helicase and 6 bp for the RecBCD helicase (20-22). Recently, the large kinetic step size for UvrD was proposed to be made of smaller kinetic steps of ~ 1 bp, at least in regards to translocation on ssDNA (23;24). The kinetic step size of the hexameric helicase DnaB was reported as 1 bp (25), whereas the gene4 helicase from bacteriophage T7 exhibited a kinetic step size of 1.4 bp (26). Studies of the NS3 helicase indicated a large kinetic step size of 18 bp (27). More recent single molecule experiments have shown that the large kinetic step size consists of smaller steps of 3-4 bp (28), which are further made up of even smaller physical steps of one bp (29). Indeed, physical models for PcrA and UvrD helicases have been proposed based on multiple x-ray crystal structures in which the number of base pairs unwound per ATP hydrolysis event is one (7;30). The Type I restriction modification enzymes contain a helicase-like motor which moves along dsDNA with one bp step size with one ATP consumed per step (31;32). Hence, it appears that the fundamental, physical step size of many helicases and translocases is one bp.</p><p>Dda is one of three helicases encoded by the bacteriophage T4 genome and is believed to play a role in the initiation of DNA replication forks at a DNA origin of replication (33) as well as during replication fork progression (34;35). Biochemical characterization of Dda indicates that it can function as a monomeric molecular motor that does not readily form higher-order oligomeric species in solution (11;36). Dda translocates with a 5′-to-3′ directional bias and can remove several different types of protein blocks from its path (37-39). A 'cooperative inchworm' model was proposed to explain how multiple Dda molecules line up along ssDNA and function together to enhance displacement of streptavidin from biotin-labeled oligonucleotides (40). The enhanced activity of multiple Dda molecules does not appear to result from specific protein-protein interactions. Rather, the presence of multiple motors moving in the same direction on ssDNA is thought to increase force production and prevent movement backwards on the ssDNA, thereby driving streptavidin displacement. Multiple Dda molecules align along the DNA to enhance many enzymatic activities including DNA unwinding (41), displacement of tryp repressor from dsDNA (42), and translocation of Dda past chemically modified DNA (43). For some activities, such as displacement of streptavidin, it is clear that one Dda molecule 'pushes' another to produce greater force in the direction of translocation. However, it is not clear whether this mechanism applies to unwinding of dsDNA. A model whereby multiple helicase molecules can enhance DNA unwinding by simply increasing the probability that unwinding will occur may also account for the observed increase in activity. Such a model, termed functional cooperativity, has been proposed for the helicase domain of the Hepatitis C viral helicase, NS3 (44). The monomeric form of Dda exhibits a kinetic step size of ~ 3 bp per step and unwinds DNA at a rate of ~250 bp/s, albeit with low processivity (45). The objective of the current study was to determine how the processivity of Dda is increased when multiple molecules bind to the same substrate. The unwinding rate and step size were measured under conditions that favor binding of more than one helicase molecule per DNA substrate and models that allow for functional cooperativity were explicitly considered to analyze the data.</p><!><p>ATP (disodium salt) and Sephadex (G-25) were obtained from Sigma. HEPES, Na4EDTA, BME, BSA, Mg(OAc)2, KOAc, SDS, xylene cyanol, bromophenol blue, NaCl, glycerol, and KOH were obtained from Fisher. T4 polynucleotide kinase was purchased from New England Biolabs. [γ32P]ATP was purchased from New England Nuclear. DNA oligonucleotides (IDT) were purified by preparative polyacrylamide gel electrophoresis and stored in 10 mM HEPES (pH 7.5) and 1 mM EDTA. Recombinant Dda was overexpressed and purified from E. coli as previously described (36).</p><!><p>Purified oligonucleotides were 5′-radiolabeled with T4 polynucleotide kinase at 37 °C for 1 hour. The kinase was inactivated by heating to 70 °C for 10 minutes, and unincorporated [γ32P]ATP was removed by passing the reaction mixture over two Sephadex G-25 columns. Helicase substrates were prepared by adding 1.2 equivalents of complement to the 5′-radiolabeled oligos, followed by heating to 95 °C for 5 minutes, and then slow cooling to room temperature.</p><!><p>Unwinding assays were performed with a Kintek rapid chemical quench-flow instrument (Kintek, Austin, TX) maintained at 25 °C with a circulating water bath. All concentrations listed are after mixing, unless otherwise stated. The helicase reaction buffer consisted of 25 mM HEPES pH 7.5, 0.1 mM Na4EDTA, 0.1 mg/ml BSA, 2 mM BME. Dda was diluted into 25 mM HEPES pH 7.5, 1 mM Na4EDTA, 0.1 mg/ml BSA, 2 mM BME, 50 mM NaCl, and 20% glycerol prior to performing the unwinding assays. For the enzyme limiting pre-steady-state experiments, Dda (final concentration of 25 nM) was incubated for 2-5 min with 100 nM radiolabeled DNA substrate and reaction buffer at 25 °C. For the excess enzyme experiments, Dda (final concentration of 100 nM) was incubated for 2-5 min with 10 nM radiolabeled DNA substrate and reaction buffer at 25 °C. Unless otherwise stated, the reaction was initiated by adding 5 mM ATP, 10 mM Mg(OAc)2. In order to prevent ssDNA product from re-forming dsDNA substrate, 300 nM re-annealing trap was placed in the receiving vial for each sample. In the enzyme limiting experiments, which contained 100 nM DNA substrate, 500 nM re-annealing trap was placed in the receiving vial, which was sufficient because only ~ 25 nM ssDNA product was produced under these conditions. The sequence of the annealing trap is complementary to the displaced strand of the substrate so that re-annealing of the displaced strand to the radiolabeled loading strand was prevented. For single-turnover conditions, 5 μM polydT was included in order to prevent Dda from re-binding to the radiolabeled substrate following the first catalytic turnover of substrate. The reaction mixture was rapidly mixed with 400 mM EDTA to quench the reaction following the allotted timeframe. 25 μl of the quenched solution was then added to 5 μl of non-denaturing gel loading buffer (0.1% bromophenol blue, 0.1% xylene cyanol in 6% glycerol). Finally, the dsDNA substrate was separated from ssDNA product on a 20% native polyacrylamide gel. Radiolabeled substrate and product were visualized by using a Molecular Dynamics Phosphorimager system and ImageQuant software. The quantity of radioactivity was used to determine the ratio of double stranded oligonucleotide substrate to single stranded oligonucleotide product as a function of time.</p><!><p>Data fitting was performed using the program Scientist (Micromath, St. Louis, MO). The function fss(t) that describes the formation of ssDNA product as a function of n-steps for scheme 1 has been defined previously as equation 1 (21;46). equation 1fss(t)=Pn(1−∑γ=1n((kobs)t)γ−1(γ−1)!e−(kobs)t) where kobs can be defined as the sum of the forward and dissociative rate constants, ku and kd (equation 2). equation 2kobs=ku+kd Processivity, P, is defined by equation 3, where m is the DNA unwinding step size and N is the average number of bp unwound. equation 3Pn=(kuku+kd)n=e−(m∕N) The method of Laplace transforms has been used previously to solve the system differential equations for reaction schemes similar to those utilized in this study (21;46). The resulting expression Fss(s) is the Laplace transform of equation 1, describing the minimal reaction scheme that describes unwinding for a helicase that dissociates readily from the DNA substrate lengths used for in vitro unwinding assays. equation 4Fss(s)=kuns(ku+kd+s)n where s is the Laplace variable of the fraction of ssDNA product formed over time, fss(t) (equation 1). The inverse Laplace transform, L −1, can be obtained using the numerical integration capabilities of Scientist to obtain fss(t) as shown in equation 5. equation 5fss(t)=L(Fss(s))−1=(kuns(ku+kd+s)n) Scheme 1 assumes that each step in the series along the unwinding pathway is identical. The kinetic step-size (m) is then defined with equation 6. equation 6m=LT−L0n where LT equals the total length of dsDNA in bp, and L0 equals the minimal length of dsDNA that is stable in the presence of an active helicase. The resulting individual kinetic step-size estimates may then be used to obtain an average kinetic step-size for the helicase under investigation.</p><p>Scheme 2 proposes that a second step occurs h number of times during strand separation as a function of the forward rate constant kc. The inverse Laplace transform fss(t) for scheme 2 was obtained using equation 7, as described previously (21;46). equation 7fss(t)=L−1(Fss(s))=(kunkchs(ku+kd+s)n(kc+kd+s)h) The value h may or may not be involved with strand separation and/or translocation. If it is assumed that h is not involved in either of these two properties then the kinetic step-size is defined only by the number of intermediates, n, that define the step-size (m) using equation 6.</p><!><p>The model for functional cooperativity was analyzed by fitting data using Kintek Global Kinetic Explorer (Kintek Corporation, Austin, TX) (47). The scripts used for the data analysis are shown in the supplemental data.</p><!><p>Previous work has evaluated DNA unwinding by monomeric Dda under conditions in which a DNA substrate was provided that was long enough to accommodate only one molecule of Dda (45). However, results indicated that binding of Dda to the duplex portion of the substrate may impede or reduce DNA unwinding under conditions where the helicase concentration exceeds the DNA substrate concentration (41). Here we have performed a kinetic analysis of DNA unwinding under conditions in which the concentration of DNA substrate is four-fold greater than Dda helicase in order to ensure that only one molecule of enzyme is bound to the substrate. Initial experiments were performed with substrate set 1 (Table 1). Previously, a 12 nt ssDNA overhang provided the best results for unwinding of a single length of dsDNA (41), however, a detailed kinetic analysis was not performed, and the effect of increasing duplex length was not evaluated. The DNA substrates used throughout this work are named based on the length of ssDNA overhang and the length of the duplex. For example, the DNA substrate with a 12 nt ssDNA overhang and a 16 bp dsDNA region is labeled as 12T16bp. Four substrates were examined in which the duplex contained 16, 20, 24, or 28 bp in order to measure the kinetic step-size and processivity of monomeric Dda. All four progress curves were fit simultaneously to equation 5 representing Scheme 1 using Scientist (Fig. 1, Table 2).</p><p>DNA unwinding produced steadily less product as the duplex length increased from 16bp up to 28 bp (Fig. 1). The kd values were allowed to float for each substrate. The rationale for this was based on the idea that longer duplexes may actually present a greater opposing force than shorter duplexes as previously suggested (26). Hence, longer duplexes might exhibit a different dissociation rate than shorter duplexes. All kinetic parameters that use the length of dsDNA in the calculation assumed that the final 8 bp melt spontaneously. For example, in order to calculate the step-size for the 12T16bp substrate, we assume that Dda is catalyzing the separation 8 bp of the initial 16 bp. The 8 bp value is based on previous results with methylphosphonate modified substrates, which indicated that the last few bp melt spontaneously during Dda-catalyzed DNA unwinding (45). The value for the unwinding rate, ku, is the rate constant for each kinetic step, which can be multiplied by the number of base pairs unwound per step to obtain an overall unwinding rate constant ku,bp. The step-size, 2.4 ± 0.7 bp, and the overall unwinding rate constant, 256 ± 74 bp s−1, are similar to values obtained previously with a substrate that can accommodate one molecule of Dda on the ssDNA overhang (45). The small increase in the kd values obtained here for longer duplexes (Table 2) may reflect the greater opposing force presented by the longer regions of duplex DNA (26). We conclude that DNA unwinding by monomeric Dda under conditions in which substrate concentration exceeds enzyme concentration proceeds with similar kinetic constants, ku, kd and kinetic step size (m) as were observed under conditions in which the enzyme concentration exceeded the substrate concentration (45).</p><!><p>It is known that multiple molecules of Dda bound to the same DNA substrate give rise to increased processivity for DNA unwinding (41), however, the mechanism for the increased processivity is not clear. Processivity (P) is generally controlled by the rate for DNA unwinding, ku, and the rate for dissociation from the DNA substrate, kd according to the equation P = ku/(kd + ku). The increase in processivity could arise from an increase in unwinding, decrease in dissociation, or both. It is also possible that the increase in processivity is due to functional cooperativity, whereby the probability of one molecule of Dda successfully unwinding a DNA substrate is increased when multiple molecules start on the same substrate (44). In this case, processivity is increased, but neither ku nor kd are necessarily altered. Figure 2 illustrates how functional cooperativity can allow multiple opportunities for unwinding to occur, thereby increasing processivity without changing rate constants. The model shows two molecules of helicase bound to the ssDNA overhang. Three kinetic steps lead to unwinding of sufficient base pairs to allow spontaneous melting of the remaining base pairs (25;45). The model allows the leading or trailing molecule to dissociate independently of one another at any step during the reaction. If the leading molecule dissociates, then the trailing molecule must translocate to the ss/dsDNA junction for unwinding to continue. Translocation steps may not occur at the same rate as the steps for unwinding (23;24), but in the case of Dda, translocation rates are very similar to unwinding rates (Byrd, Matlock and Raney, in preparation). Therefore, each stepping rate, unwinding (ku) and translocation (kt), was made equal in this model.</p><p>DNA unwinding was conducted with substrates containing varying length ssDNA overhangs of 12, 14, 21, and 28 nt and duplex lengths of 16 and 20 bp (Figure 3). Longer ssDNA overhang lengths were largely based on fluorescence titration data suggesting that a single Dda molecule occupies 6-7 nt (40). Also, substrates with an overhang of less than 6 nt do not exhibit significant unwinding in vitro (41). Kinetic mechanisms were defined in which one, two, three, or four molecules of Dda were bound to the substrates containing 12, 14, 21, and 28 nt overhangs, respectively. Three or four unwinding steps were designated for the 16 bp or 20 bp substrates, respectively based on the average kinetic step size of 3.2 bp and the number of base pairs that spontaneously melt. For example, 8 base pairs melt spontaneously, so only 8 bp from the 16 bp substrate and 12 bp from the 20 bp substrate are considered in the mechanism. Progress curves for each substrate were fit individually to their respective kinetic models by using Kintek Global Kinetic Explorer (47), allowing the rate constants ku and kd to float. The resulting kinetic parameters are listed in Table 3. The dissociation rate constants vary by four-fold (5.6 ± 1.2 s−1 to 25 ± 3.4 s−1). It has been suggested that longer duplexes may present a greater force opposing the movement of a helicase (26), which might account for the small variation observed here when comparing a 16bp to a 20bp duplex. The rate constant for each kinetic step, ku, ranged from 65 ± 5 s−1 to 116 ± 6 s−1. The small increase in ku may relate to the fact that the exact number of kinetic steps may be slightly less than 3 for the 16 bp substrate and slightly more than 4 steps for the 20 bp substrate. However, the overall consistency in the unwinding rate constants within each substrate set (16 bp duplexes and 20 bp duplexes) indicates that ku does not increase as more molecules of Dda are added to the substrate.</p><p>We conclude that the model depicted in Figure 2 for functional cooperativity readily accounts for the increased processivity observed when multiple molecules of Dda unwind the same DNA substrate. The significance of this model is in the fact that protein-protein interactions are not required to account for the observed increase in product formation. This result is in contrast to previous results for streptavidin-displacement activity exhibited by Dda, which clearly requires trailing helicase molecules to 'push' the lead molecule in order to push streptavidin from biotin-labeled DNA (40).</p><!><p>In addition to the ku and kd values, the kinetic step size, m, also defines the kinetic mechanism for helicase-catalyzed unwinding. The kinetic step size can be determined by measuring DNA unwinding with DNA substrates of increasing duplex length (20;21). In previous work, we found that the kinetic step size varied from 3.5 to 7.0 bp unless one accounted for spontaneous melting of the last few base pairs of the substrate (45). The final 8 bp of the DNA were observed to melt spontaneous, and when this was taken into account, the step size varied from 2.5 to 3.5 bp for different substrate lengths. Therefore, spontaneous melting of the last 8 bp was included in the analysis of all substrates in this work. Single-turnover RQF experiments were performed with substrates possessing a 14 nt ssDNA overhang (substrate set 2). Enzyme concentration greatly exceeded the substrate concentration so that more than one molecule of Dda binds to the substrate prior to initiation of the reaction. Two molecules of Dda should be able to bind to the 14nt overhang based on the binding site size of 6-7 nt. Four substrates containing a 14 nt ssDNA overhang and increasing lengths of dsDNA were examined (Fig. 4A). All four data sets were fit simultaneously to equation 5 describing reaction Scheme 1. The rate constant ku was constrained to be identical for all data sets, while the kd value and number of steps, n, were allowed to float. The resulting kinetic parameters are listed in Table 4. The step-size, 3.1 ± 0.9 bp, and the value for ku, 66 ± 7, were similar to those obtained under enzyme-limiting conditions (45). Small variations in the kd values were also observed, although no clear trend was evident with the different substrates (Table 4).</p><!><p>Next, single-turnover RQF experiments were performed with substrates possessing 21 nt and 28 nt ssDNA overhangs (substrate sets 3 and 4, respectively) under excess-enzyme conditions (Fig. 4B, 4C). These conditions should favor binding of 3 or 4 molecules of Dda to the 21nt and 28nt substrates, respectively. One of the most obvious differences between the data for longer ssDNA overhangs and those obtained under conditions that favor a monomer is the increased amplitude of product formation, especially for the longer duplexes (compare unwinding for the 28 bp substrate in Figure 4C to the same length in Figure 4A).</p><p>Another intriguing observation resides in the fact that product curves for substrates containing long ssDNA overhangs exhibit a discontinuous change in the lag phases of different lengths of dsDNA. For example, there is a similar lag phase for the 21T16 bp and 21T20 bp substrates, but the lag phase for the 21T24 bp and 21T28 bp substrates is much longer (Fig. 4B and 4C). Hence, the four substrates containing the 21nt overhang appear to fall into two distinct populations, as do the four substrates containing the 28nt overhang. If the helicase reaction scheme is defined by a series of 'fast' steps followed by a rate-limiting 'slow' step, then as the rate of each 'fast' step approaches infinity, unwinding curves for substrates with the same number of 'slow' steps will become super-imposable (21). Such a phenomenon has been described for the HCV NS3 helicase using a combinatorial, time-resolved unwinding assay with single bp resolution (48), and may explain the data here for Dda. Substrates that can accommodate 3-4 Dda molecules (i.e. 21 nt and 28 nt ssDNA overhang) clearly exhibit a trend in which the unwinding curves for substrates of different physical lengths become super-imposable as the formation of ssDNA product becomes limited by a 'slow' step. Consequently, the reaction scheme defining DNA unwinding by multiple Dda molecules was expanded to include a non-identical step along the reaction pathway (Scheme 2), where the "slow" step is defined by the rate constant kc. The unwinding results for the 21T and 28T ssDNA overhang substrates were fit to equation 7 representing Scheme 2 and the resulting kinetic parameters are listed in Tables 5 and 6, respectively.</p><p>The amplitudes for product formation are plotted for each of the substrate sets in Figure 5A. The trend in the data clearly shows how the substrates with longer ssDNA overhangs which can accommodate more Dda molecules give rise to more product. The number of base pairs unwound per binding event can be estimated by fitting the data to a linear function and then solving for y = 0. Unwinding per binding event is increased from ~30 bp for the monomeric enzyme to ~110 bp for multiple-molecules bound to the 28 nt ssDNA overhang (calculated based on the 28 bp duplex).</p><p>Figure 5B shows the unwinding rate constants and step-size values plotted as a function of the number of Dda molecules bound to substrate. The unwinding rate constants remain similar, regardless of the number of Dda molecules that are initially bound to the substrate. No clear trend is observed in the kinetic step-size values as a function of the number of Dda molecules bound. Hence, the major role of multiple Dda molecules in DNA unwinding is to provide increased processivity, by increasing the probability for the helicase to complete unwinding prior to dissociation.</p><!><p>A single-stranded oligonucleotide that is complementary to the strand that is displaced by the helicase must be introduced into the standard in vitro unwinding assay in order to prevent spontaneous re-annealing of the ssDNA products. This oligonucleotide is referred to as a re-annealing trap. By increasing the amount of re-annealing trap present in solution, we were previously able to observe the transient appearance of an intermediate, which was correlated with a 'pause' in Dda-catalyzed unwinding occurring as a function of dsDNA length (45). This intermediate species was observed with 24 bp and 28 bp DNA substrates, but not with shorter 16 bp and 20 bp substrates. Therefore, the non-identical step, or pause, takes place after unwinding ≥ 20 bp. Since Dda only catalytically separates ~12 bp of a 20 bp substrate (taking into account spontaneous melting of the final 8 bp), 'pausing' was proposed to occur once every ~10 bp.</p><p>The appearance of the non-uniform step, as shown by the intermediate, was investigated by using the 12T substrates under conditions that favor monomeric Dda (substrate concentration > enzyme concentration). The concentration of annealing trap was increased from 300 nM to 5 μM and the results clearly show the appearance of the intermediate, but only with the duplex substrates of 24 and 28 bp (Fig. 6A). Since the data with the 12T substrates were obtained under conditions in which the substrate concentration exceeds the enzyme concentration, the appearance of the intermediate clearly shows that monomeric Dda undergoes the non-uniform step in the kinetic mechanism.</p><p>In order to determine whether the intermediate appeared under conditions in which multiple Dda molecules bind to the same substrate, DNA unwinding assays were performed with the substrates containing longer ssDNA overhangs under conditions that favor binding of multiple Dda molecules per substrate (enzyme concentration > substrate concentration). A large excess of annealing trap was included (500-fold excess) in these experiments and single-turnover RQF experiments were performed. The appearance of the intermediate was observed for all ssDNA overhang lengths tested, but only for dsDNA lengths of 24 bp and 28 bp (Fig. 6B-6D). Thus, multiple Dda molecules aligned on the same substrate exhibit a similar non-uniform step in the unwinding mechanism as monomeric Dda.</p><p>The intermediate appears if duplex lengths are sufficiently long, indicating that Dda must unwind several base pairs prior to undergoing the slower step. If this intermediate is related directly to DNA unwinding, then its appearance should be delayed under conditions in which unwinding is slowed. To determine whether the intermediate appears later in the progress curve under conditions of slower unwinding, the reaction was performed at 100 μM ATP. Unwinding by multiple Dda molecules is slowed under these conditions, but the processivity remains similar based on the amplitude of product formation (Fig. 7A). However, the appearance of the intermediate is delayed when the concentration of ATP is reduced, indicating that the intermediate is directly related to DNA unwinding and to the number of base pairs unwound (Fig. 7B, open circles). The presence of the intermediate provides evidence that the non-uniform step occurs even when multiple Dda molecules are involved in DNA unwinding.</p><!><p>Previous work in which multiple Dda molecules were placed on the same substrate molecule indicated that Dda molecules "pushed" one another upon encountering a blockade along the DNA. For example, when biotin/streptavidin blocks are placed in the path of Dda, the rate of streptavidin displacement is increased by ~ one million fold when 5-6 molecules of Dda push together compared to monomeric Dda (38). Similarly, monomeric Dda was unable to displace trp repressor protein bound to DNA under single cycle conditions, whereas two or more Dda molecules readily pushed the protein off of the DNA (42). Perturbations in the DNA structure such as abasic sites blocked DNA unwinding by monomeric Dda, but were overcome by the presence of two or more molecules bound to the same substrate (43). Hence, one molecule of Dda can push another molecule through a block in the path of the helicase.</p><p>Adding additional molecules of Dda to the same DNA substrate results in increased product formation, however the kinetic mechanism for this increase was not known until now. Multiple molecules may lead to increased unwinding rates, decreased dissociation rates, or both. It is also possible that the observed increase in product formation results from functional cooperativity (44), whereby the unwinding rates and dissociation rates need not change, but unwinding is increased by the increase in the probability that at least one helicase molecule completes the unwinding process. Data in this report favor a mechanism in which the increase in product is accounted for simply due to the increased probability that at least one molecule completes the unwinding process. This phenomenon has been referred to as 'functional cooperativity' (44). Cooperativity is generally associated with protein-protein interactions that give rise to enhanced activity within an enzyme active site. In functional cooperativity, no protein-protein interactions are required because of the nature of the DNA substrate. Helicases can simply align on the same substrate molecule, thereby increasing the likelihood of that DNA duplex being unwound by mass action of the bound helicases. However, protein-protein interactions may contribute to functional cooperativity whereby a trailing helicase molecule engages a leading molecule on the same DNA track. Whether the trailing helicase pushes the lead helicase, or simply prevents the lead helicase from slipping back has not been completely addressed. However, loading of an ATPase-deficient form of Dda on the same track as active molecules failed to enhance the rate of streptavidin displacement (40).</p><p>The rate constants that control DNA unwinding processivity, as well as the kinetic step size, were determined under conditions in which only one molecule of Dda could bind to the DNA substrate (Fig. 1, Table 2). The kinetic values were then used to construct a model for functional cooperativity in which one, two, three, or four molecules of Dda were bound to the same DNA substrate. The data obtained from these experiments were readily fit by the model for functional cooperativity (Figs. 2 and 3, Table 3). No trends were observed in resulting unwinding rate constants as a function of increasing numbers of Dda molecules on the same DNA substrate. Hence, for regular DNA unwinding, the inherent unwinding activity of Dda is not activated by the presence of multiple Dda molecules. This result is in sharp contrast to the results obtained for protein displacement or for unwinding of DNA that contains chemical perturbations. We conclude that Dda molecules move along the DNA independently, until the lead molecule encounters a block to translocation. At that point, multiple molecules 'pile up' and begin to push one another, resulting in displacement of the block or movement passed the block.</p><p>The data are also in contrast to helicases such as E. coli Rep (49), and E. coli UvrD (13), and B. stearothermophilus PcrA (50;51) which are highly activated for DNA unwinding upon formation of dimeric or higher order oligomers. Recent work indicates that PcrA helicase can be activated by the presence of a DNA binding protein as well. Processivity for DNA unwinding by PcrA increases dramatically in the presence of the initiator protein for plasmid replication, RepD (52).</p><p>We found that small variations in the dissociation constant were observed for monomeric Dda as the length of the duplex increases (Table 1). This may be due to the increased stability exhibited by longer duplexes, which has been described in terms of an opposing force to the helicase (26). Dda is especially sensitive to opposing forces encountered during translocation on DNA as shown by the previously mentioned results for displacement of the trp repressor bound to DNA in which monomeric Dda could not displace the protein. Here we observed a small increase in dissociation constants when the length of the duplex was increased from 16 bp to 28 bp, which may be due to an increased opposing force caused by greater strand rigidity in the longer duplexes (53).</p><!><p>A feature of the kinetic mechanism for unwinding by Dda is the appearance of an intermediate species that can be observed by increasing the concentration of annealing trap in the reaction (45). Analysis of our data provides two lines of evidence for the existence of a non-uniform step in the kinetic mechanism of Dda, even when multiple molecules are present on the same substrate. First, using a ssDNA overhang of ≥21 nt results in an unusual trend in the amount of time required to unwind substrates (i.e. the lag phase) of different dsDNA lengths. For example, the 21T16 and 21T20 substrates are unwound in a similar amount of time and to a similar extent (Fig. 4B). The lag phases for Dda-catalyzed unwinding of the 21T24 and 21T28 substrates are also very similar, but Dda requires a much longer period of time for product to form relative to the 21T16 and 21T20 substrates. Such a large difference in lag phases for 20 bp and 24 bp substrates strongly suggests that a slower, rate-limiting event takes place somewhere between separation of 20 bp and 24 bp lengths. Second, an intermediate species is clearly observed for Dda-catalyzed separation of ≥24 bp of dsDNA (Fig. 6). Taken together, the two results reported here support a mechanism in which Dda exhibits a non-uniform step that occurs as a function of dsDNA length (Scheme 2).</p><p>Others have described non-uniform stepping in the case of DNA unwinding by RecBCD (46), RNA unwinding by NS3 (48), and translocation on ssDNA for UvrD (24). Previously, we suggested that this step for Dda involved a "hand-off" of the displaced strand during DNA unwinding (24;45). Dda may interact with both strands of the duplex during the unwinding reaction, and after unwinding of around one turn of the duplex, the displaced strand is released. Reports on NS3-catalyzed RNA unwinding have proposed a similar mechanism by which the displaced strand remains bound to the enzyme during several sub-steps of unwinding (27). Release of the displaced strand is proposed to limit the overall unwinding cycle, resulting in a slow kinetic step that gives rise to a large kinetic step size. In the case of UvrD, the pause in the kinetic mechanism was observed during translocation on ssDNA, so a different mechanism must be operating to produce the non-uniform step. Movement of UvrD helicase domains during ATP hydrolysis was proposed to "scrunch" the ssDNA between the domains, thereby leading to a slow step during which this ssDNA was released (24). It appears that non-uniform steps in the kinetic mechanisms of helicases occur within several different sub-families of helicases, but different mechanisms may give rise to these steps.</p><!><p>The rate constant for streptavidin displacement increases by 75-fold when comparing one molecule vs. 2 molecules of Dda aligned on the same substrate (40). Therefore, when translocation is impeded, the activity of two or more molecules is clearly enhanced due to the trailing molecule being able to push the lead molecule. Such a synergistic effect is not observed during routine DNA unwinding in going from one molecule to 2 molecules of Dda. The major gain of function for routine DNA unwinding when multiple Dda molecules align is in terms of processivity, which increases by less than four-fold in going from one molecule of Dda to 4 molecules of Dda on the same substrate. Therefore, multiple molecules of Dda aligned on the same substrate impart a much greater affect on displacement of protein blocks than on DNA unwinding. Protein displacement is clearly an important function for some helicases, as illustrated by the role of Srs2 in disassembly of Rad51 protein filaments to regulate homologous recombination (54). It remains to be determined whether helicase-mediated protein displacement in the cell is regulated by the number of helicase molecules that are aligned along the DNA.</p><p>DNA unwinding processes in the cell clearly require greatly varying degrees of processivity. Replication requires unwinding of thousands of base pairs per single binding event, whereas some DNA repair processes require unwinding of only a few base pair. Modulation of processivity is likely to be strictly controlled in the cellular environment. This control occurs through the inherent activity of a particular helicase or through protein-protein interactions that can greatly alter helicase processivity. Highly processive helicases typically have multiple sites of interaction with DNA, as with RecBCD helicase, or encircle the DNA, as with replicative hexameric helicases. Translocases such as the Type 1 restriction-modification enzymes contain "helicase" motors that can move along dsDNA with very high processivity, as a result of multiple contacts between protein subunits and DNA (31;32). Helicases with inherently low processivity, such as PcrA helicase, can exhibit very high processivity when bound to accessory factors (52). The role of accessory factors for helicases may be similar as with processivity factors and DNA polymerases. The processivity factor essentially provides additional contacts between the protein complex and the DNA, thereby holding the complex onto the DNA during translocation. Dda can interact with other T4 proteins, such as gp32 and UvsX, which are likely to modulate its processivity during viral replication.</p><p>DNA unwinding can be broken down into steps for melting of the duplex and for translocation along the strand. The fact that trailing molecules of Dda do not push the lead molecule during DNA unwinding indicates that the lead molecule is moving along the DNA at similar rates as the trailing molecules. Therefore, the lead molecule is able to unwind duplex DNA and move forward at similar rates as the trailing molecules which only need to translocate on ssDNA. This implies that the rate-limiting step in moving along DNA is the same for DNA unwinding as it is for translocation on ssDNA. Therefore, the likely rate-limiting step for the activity of Dda helicase is contained in the translocation step, and DNA melting does not limit the overall rate of DNA unwinding.</p><!><p>This work was supported by the National Institutes of Health grant R01 GM059400 (K.D.R) and the UAMS Committee for Allocation of Graduate Student Research Funds (R.L.E).</p><p>SUPPORTING INFORMATION AVAILABLE</p><p>Scripts describing the data analysis using Kintek Global Kinetic Explorer (Kintek Corporation, Austin, TX) are available free of charge via the internet at http://pubs.acs.org.</p><p>Rapid Quench Flow</p><p>streptavidin</p><p>non-linear least squares</p><p>DNA unwinding by monomeric Dda. Unwinding was measured under conditions in which substrate concentration (100 nM) exceeded the helicase concentration (25 nM) to ensure that only one molecule of Dda was bound per substrate molecule. DNA substrates contained a 12nt overhang of ssDNA and varying lengths of duplex including 12T16bp (), 12T20bp (), 12T24bp (), and 12T28bp (). The experiments were performed in the presence of 5 μM polydT to create single-turnover conditions. The data were fit to equation 5 using Scientist and the resulting kinetic parameters are listed in Table 2.</p><p>Model for functional cooperativity for Dda helicase. Two Dda molecules are shown bound to the 14T16bp substrate (14 nt overhang and 16 bp of duplex DNA). Upon addition of ATP and Mg+2, DNA unwinding occurs according to rate constant ku (green arrows). Three kinetic steps are shown for unwinding of the 16 bp substrate (11). Either enzyme molecule can dissociate from the DNA substrate according to rate constant kd (red arrows). If the leading enzyme molecule dissociates, then the trailing molecule can translocate to the ss/ds junction according to rate constant kt. The blue dotted lines and arrows indicate two kinetic steps that are required for the trailing enzyme to move to the ss/ds DNA junction. The final base pairs can melt spontaneously giving rise to ssDNA product.</p><p>Dda helicase-catalyzed unwinding of 16 bp (A) and 20 bp (B) substrates containing varying length ssDNA overhang. For all substrates containing 12 nt of ssDNA, the DNA substrate concentration (100 nM) exceeded the concentration of Dda (25 nM) and the quantity of product was divided by the enzyme concentration to obtain the fraction unwound. These conditions were chosen to ensure that only one molecule of Dda was bound to the substrate. For all other substrates, the concentration of Dda (100 nM) exceeded the DNA substrate concentration (10 nM) in order to ensure that multiple molecules were bound to the ssDNA overhangs. All experiments were performed in the presence of 5 μM poly dT to create single-turnover conditions with respect to the DNA substrate. The lengths of the ssDNA overhangs are listed in the figure. Data were fit to a model as depicted in Figure 2 by using the program Kintek Global Kinetic Explorer (Kintek Corp.). Three or four kinetic steps for unwinding were used for the unwinding step with the 16 bp substrate and 20 bp substrate, respectively. One, two, three, or four molecules of Dda were pre-bound to the DNA substrate for the 12 nt, 14 nt, 21 nt, and 28 nt substrates, respectively. The rate constant for translocation was set equal to that for DNA unwinding. The rate constants for DNA unwinding, ku, and dissociation, kd, were allowed to float for each substrate to obtain the best fit for each mechanism. The resulting rate constants are shown in Table 3.</p><p>Single-turnover results for Dda-catalyzed unwinding of DNA substrates possessing 14, 21, and 28 nt ssDNA overhangs. A. Results for 100 nM Dda-catalyzed unwinding of 10 nM 14T16bp (), 14T20bp (), 14T24bp (), and 14T28bp () DNA substrates, respectively. The data were fit to equation 5 using Scientist and the resulting kinetic parameters are listed in Table 4. B. Results for 100 nM Dda-catalyzed unwinding of 10 nM 21T16bp (), 21T20bp (), 21T24bp (), and 21T28bp () DNA substrates, respectively. The experiments were performed in the presence of 5 μM polydT to create single-turnover conditions. The data were fit to equation 7 using Scientist and the resulting kinetic parameters are listed in Table 5. C. Results for 100 nM Dda-catalyzed unwinding of 10 nM 28T16bp (), 28T20bp (), 28T24bp (), and 28T28bp () DNA substrates, respectively. The experiments were performed in the presence of 5 μM polydT to create single-turnover conditions. The data were fit to equation 7 using Scientist and the resulting kinetic parameters are listed in Table 6.</p><p>Multiple Dda molecules increase the amplitude of ssDNA formed, but do not change the kinetic step size, or the rate for unwinding. A. Re-plot of product amplitudes as a function of dsDNA length for the 12T (), 14T (), 21T (), and 28T () substrate sets. B. Re-plot of the unwinding rate constant () and the kinetic step-size () as a function of the estimated number of Dda molecules bound to the different substrates at the start of the reaction. One, two, three, or four Dda molecules were bound to substrate 12T, 14T, 21T, and 28T, respectively. Error bars are the errors in the fits to the data.</p><p>A transient increase in the amount of ssDNA product in the presence of excess re-annealing trap suggests a slow step for unwinding of longer duplexes. All of the experiments shown were performed with 5 μM re-annealing trap in the reaction. The transient peak at around 0.035 s is observed for all of the 24 and 28 bp dsDNA lengths tested. Panel A shows the results of enzyme-limiting conditions (25 nM Dda and 100 nM DNA); all of the other experiments (panels B-D) were performed under excess enzyme conditions (100 nM Dda and 10 nM DNA). All experiments were initiated with 5 mM ATP and 10 mM Mg(OAc)2. Each panel uses the same symbol for each length of dsDNA: 16 bp (), 20 bp (), 24 bp (), and 28 bp ().</p><p>Dda-catalyzed unwinding of a 28T28bp substrate in the presence of 500-fold excess re-annealing trap results in the appearance of an intermediate species. A. Results for 100 nM Dda-catalyzed unwinding of 10 nM 28T28bp substrate in the presence of 500-fold excess re-annealing trap and ATP concentrations of either 5 mM (●) or 100 μM (○). Appearance of the transient peak in product formation is delayed when ATP concentration is reduced. B. Results in panel 7A are shown to 0.1 seconds.</p><p>The "14T" ssDNA overhang substrates contain 14 thymidine residues with the 16 bp, 20 bp, 24 bp, and 28 bp duplex sequences shown above.</p><p>The "21T" ssDNA overhang substrates contain 21 thymidine residues with the 16 bp, 20 bp, 24 bp, and 28 bp duplex sequences shown above</p><p>The "28T" ssDNA overhang substrates contain 28 thymidine residues with the 16 bp, 20 bp, 24 bp, and 28 bp duplex sequences shown above.</p><p>N (the number of bp unwound per binding event) was estimated according to Eq. 3 by using the step-size, m, the ku value and the average of the kd values (21).</p><p>Data were fit by using Kintek Global Kinetic Explorer (Kintek Corp.). Kinetic schemes were based on the model from Figure 2, which depicts two molecules bound to the substrate. Three or four unwinding steps were used for the 16bp and 20bp substrates, respectively. More complex kinetic mechanisms were used for substrates containing three or four Dda molecules (Supplementary data). Errors are standard errors obtained from the best fit of the data.</p><p>The number of Dda molecules bound per substrate is estimated from the binding site size and the length of the ssDNA overhang. In the case of the 12T substrates, the concentration of DNA substrate was in four-fold excess over the concentration of enzyme to ensure binding of one molecule of Dda per substrate molecule.</p><p>N (the number of bp unwound per binding event) was estimated according to Eq. 3 by using the step-size, m, the ku value and the average of the kd values (21).</p><p>N (the number of bp unwound per binding event) was estimated according to Eq. 3 by using the step-size, m, the ku value and the average of the kd values (21).</p><p>N (the number of bp unwound per binding event) was estimated according to Eq. 3 by using the step-size, m, the ku value and the average of the kd values (21).</p>

PubMed Author Manuscript

Tuning interactions between zeolite and supported metal by physical-sputtering to achieve higher catalytic performances

To substitute for petroleum, Fischer-Tropsch synthesis (FTS) is an environmentally benign process to produce synthetic diesel (n-paraffin) from syngas. Industrially, the synthetic gasoline (iso-paraffin) can be produced with a FTS process followed by isomerization and hydrocracking processes over solid-acid catalysts. Herein, we demonstrate a cobalt nano-catalyst synthesized by physical-sputtering method that the metallic cobalt nano-particles homogeneously disperse on the H-ZSM5 zeolite support with weak Metal-Support Interactions (MSI). This catalyst performed the high gasoline-range iso-paraffin productivity through the combined FTS, isomerization and hydrocracking reactions. The weak MSI results in the easy reducibility of the cobalt nano-particles; the high cobalt dispersion accelerates n-paraffin diffusion to the neighboring acidic sites on the H-ZSM5 support for isomerization and hydrocracking. Both factors guarantee its high CO conversion and iso-paraffin selectivity. This physical-sputtering technique to synthesize the supported metallic nano-catalyst is a promising way to solve the critical problems caused by strong MSI for various processes. There is a recognized need to synthesize clean liquid fuel, including gasoline and diesel, from renewable biomass to solve the global petroleum shortage crisis and also realize a green process by decreasing CO 2 emissions. Fischer-Tropsch synthesis (FTS) 1 is a promising way to meet this demand using syngas (CO 1 H 2 ), which is easily derived from biomass, garbage, and shale gas 2 , to produce environmentally benign sulfur-free liquid fuels 3 . It not only avoids the acid rain and photochemical smog but also solves the sulfur-poison problem for catalytic aftertreatment systems of automobiles. For example, the conventional Lean NO x Trap catalysts, which are readily poisoned by sulfur, can only be commercialized on lean-burn engines using sulfur-free fuels 4 . The FTS reaction usually uses a Co, Ru or Fe based catalyst 5 . The synthetic diesel, i.e. long chain normal paraffin (n-paraffin), is the dominant FTS product, and the carbon number distribution is controlled by the Anderson-Schulz-Flory (ASF) law 5 . It is difficult to disobey this law and achieve selective and specified carbon numbers, e.g. gasoline range (C 5 -C 11 ) iso-paraffin products.Many efforts have been made to obtain these gasoline range iso-paraffin products from FTS. For example, the Shell SMDS plant in Malaysia used a separate reactor containing solid acid catalysts to isomerize and hydrocrack the long chain n-paraffin produced from FTS 6 . Theoretically, if the FTS and the subsequent isomerization/ hydrocracking processes occur on a combined catalyst including both FTS and acidic catalysts under the same reaction conditions, only one reactor is needed. This design will decrease economic cost significantly. For such a consecutive reaction, its catalytic performance can be remarkably enhanced by keeping these two kinds of sites close to each other to improve their dispersion 7 . Here, we expect that the produced long n-paraffin products at the FTS sites can be shortened and branched to iso-paraffin at the adjacent acidic sites. Accordingly, the iso-paraffin products (premier gasoline) are efficiently synthesized by one step.Zeolite, such as H-ZSM5, is one of the conventional solid acid catalysts possessing good hydrocracking and isomerization performances 8 . Therefore, some efforts have been done to synthesize iso-paraffin from syngas using

tuning_interactions_between_zeolite_and_supported_metal_by_physical-sputtering_to_achieve_higher_cat

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

<p>the physical-mixed FTS and zeolite catalysts [9][10][11] . However, a lot of the linear hydrocarbons directly escaped from the FTS sites due to the uneven distribution of the FTS sites and acidic sites. As a result, the formed FTS wax was easily deposited on the FTS active sites, which seriously caused catalytic deactivations. Therefore, this methodology is insufficient for the synthesis of the commercial synthetic gasoline. Directly loading a FTS catalyst onto an acidic zeolite support by the impregnation method was also reported, but only a low FTS activity was achieved 12,13 . The strong MSI between the FTS catalyst and the zeolite support severely inhibited the reduction of the FTS catalyst precursor, i.e. the metal oxides. Depositing cobalt cations into micropores of zeolite by the ion-exchange method 14 is also inconvenient because the strong interactions between the cobalt cations and zeolites would require a high temperature to reduce them to the metals, e.g. above 700uC 15 . Such a high temperature will further destroy the zeolite frameworks in the presence of water.</p><p>Recently, we proposed one kind of the encapsulated catalysts by coating an acidic zeolite onto a FTS catalyst pellet for the direct synthesis of iso-paraffin from syngas [16][17][18] . These tailor-made encapsulated catalysts can break the ASF law, reduce the heavy hydrocarbon products, and enhance the selectivity of the gasoline range paraffin, especially iso-paraffin. However, it hardly avoids the formation of the cracks and defects on the zeolite membrane during the reactions due to the fraction between the catalyst pellets and the thermal expansion caused by heat.</p><p>To overcome these problems, herein, we demonstrate a facile drymethod to synthesize H-ZSM5 supported metallic cobalt nanoparticles using a hexahedral-barrel sputtering system 19 . This system has been employed to deposit nano-particles, such as Pt, Ru, Pt-Ru alloy and TiO 2 , onto various inert powder supports acting as an electrode catalyst for fuel cell or a catalyst for CO 2 methanation [20][21][22][23] . Unlike the common physical-sputtering devices, which sputter atomic clusters from an aimed target onto a glass or ceramic disc, our rotated polygonal-barrel sputtering system can deposit nanoparticles onto the powders homogeneously 20,21 ; the other advantage is the narrow and controllable size distribution of the nano-particles through varying its ac power levels 21 . This technique provides a new possibility to synthesize nano-catalysts with weak MSI, since the nano-particles are wedged onto the support by physical forces. The possible reaction routes of this catalyst during the FTS, isomerization and hydrocracking reactions are proposed in Fig. 1.</p><!><p>Preparation and characterizations. Differently from catalytically inert support powders until now, here, cobalt nano-particles were directly sputtered from the metallic Co target onto the acidic H-ZSM5 zeolite powders, and the obtained catalyst was denoted as Co/H-ZSM5-S. For comparison, the other H-ZSM5 supported cobalt catalyst was also synthesized by a conventional wet impregnation method, and was denoted as Co/H-ZSM5-I. The Co/ H-ZSM5-S and Co/H-ZSM5-I catalysts had the same cobalt weight loading, i.e. 7.0 wt.%, and presented metallic luster and green-yellow appearances, respectively (see Supplementary Fig. S1 online). It implies the presence of the metallic cobalt species on the Co/H-ZSM5-S catalyst, whereas no cobalt-related phase was detected from the XRD pattern (see Supplementary Fig. S2 online). Fig. 2 shows their High Resolution Transmission Electron Microscopy (HR-TEM) images. For the Co/H-ZSM5-I catalyst, the cobalt www.nature.com/scientificreports particles seriously aggregated, and randomly deposited on the H-ZSM5 support with the size range of 10-30 nm, whereas the cobalt particles of the Co/H-ZSM5-S catalyst homogeneously distributed on the H-ZSM5 surface with the narrow size range of 2-4 nm.</p><p>The H 2 Temperature-Programmed Reduction (H 2 -TPR) profiles of the catalysts are presented as Supplementary Fig. S3 online. Herein, the H 2 -TPR profiles of the both cobalt catalysts presented two reduction peaks due to Co 3 O 4 R CoO and CoO R Co 0 24,25 . It is worth to note that the initial reduction temperature of the cobalt oxides on the sputtered catalyst was about 80uC lower than that on the impregnated one indicating that the former had a better reducibility.</p><p>The results of the X-ray Absorption Near Edge Structures (XANES) show that for the unreduced Co/H-ZSM5-I catalyst the Co species existed in the form of Co 3 O 4 , while for the unreduced H-ZSM5-S catalyst the Co species was composed of metallic Co, CoO and probable Co 3 O 4 (see Supplementary Fig. S4 online). Fig. 3 shows the Radial Structure Functions (RSFs) determined by Extended Xray Absorption Fine Structures (EXAFS) of the catalysts with the various pretreatments and the standard cobalt samples as well. As shown in Fig. 3A, for the fresh Co/H-ZSM5-I catalyst, the cobalt species existed as Co 3 O 4 ; after it was reduced in H 2 at 260uC for 1 h (catalyst denoted as Co/H-ZSM5-I-260), the cobalt species was mainly in the form of CoO; after increasing the reduction temperature to 400uC and prolonging for 10 h (catalyst denoted as Co/H-ZSM5-I-400), the metallic cobalt became the dominant species. In Fig. 3B, the presence of the metallic species in the fresh Co/H-ZSM5-S catalyst was clearly observed. After the catalyst was reduced in H 2 at 260uC for 1 h (denoted as Co/H-ZSM5-S-260), the metallic cobalt and CoO were the major species. Furthermore, the magnitude of the coordination peak of the Co/H-ZSM5-S catalyst is much weaker than that of the Co/H-ZSM5-I catalyst, demonstrating that the former has a smaller coordination number and a higher disorder degree. Here, all of the EXAFS signals were ex situ recorded in air. The metallic Co nanoparticles on the Co/H-ZSM5-S and Co/H-ZSM5-S-260 catalysts were readily oxidized in air due to their small sizes and crystal defects (low coordination number). Consequently, the cobalt oxides were identified in Fig. 3B.</p><p>The qualitative contents of the cobalt compounds in the catalysts were achieved by simulation of the XANES spectra and the simulated results can be found as Supplementary Table S1 The elemental surface analysis of the catalysts without any reduction treatment was determined by X-ray Photoelectron Spectroscopy (XPS), and the details can be found as Supplementary Table S2 online. About 8.0 atm.% cobalt, i.e. 21.0 wt.% cobalt elements was detected on the surface of the Co/H-ZSM5-S catalyst, while only 7.1 wt.% of cobalt was detected on the surface of the Co/H-ZSM5-I catalyst. The cobalt content on the Co/H-ZSM5-S catalyst greatly exceeds the theoretical value, i.e. 7.0 wt.%. This enrichment is due to the physical anchor of the cobalt nano-particles on the external surface of the zeolite by sputtering with weak MSI. It was reported that only the easily reducible cobalt particles on the external surface of the zeolite support were the active sites for FTS 27 . Herein, the readily reducible cobalt nano-particles on the Co/H-ZSM5-S catalyst were homogeneously wedged on the external surface of the H-ZSM5 with more exposed edges, corners and faces as compared with the Co/H-ZSM5-I catalyst. Thus, these differences probably resulted in a better FTS performance. Moreover, the high dispersion and enrichment of the cobalt nano-particles on the surface of the H-ZSM5 would increase the chance for the produced long chain n-paraffin products therein to be captured by the adjacent acidic sites on the H-ZSM5 support. This is beneficial to the production of iso-paraffin.</p><p>Catalytic performance. To certify our assumption, the catalytic activities of the Co/H-ZSM5-I and Co/H-ZSM5-S catalysts were compared for the direct synthesis of iso-paraffin from syngas. The obtained catalytic performances are listed in Table 1, and the product distributions are given in Fig. 4. It shows clearly that the CH 4 selectivity is in the sequence of Co/H-ZSM5-I-260 . Co/H-ZSM5-I-400 . Co/H-ZSM5-S-260, while the tendency of the CO conversion is on the contrary. Moreover, the Co/H-ZSM5-S-260 catalyst provides the highest iso-paraffin selectivity among these catalysts, and its molar ratio of iso-to n-paraffin in the products ($C 4 ) reaches 2.2. The similar trend was observed over the Co/H-ZSM5 catalysts with the 2.2 wt% cobalt loading (see Supplementary Table . S3 online).</p><p>The pyridine FT-IR spectra of the H-ZSM5 support, Co/H-ZSM5-S-260 and Co/H-ZSM5-I-400 catalysts see Supplementary Fig. S6 online. The amount of the Brønsted acid sites on both cobalt supported catalyst is smaller than that of the H-ZSM5 support. The Co/ H-ZSM5-S-260 catalyst had the similar amount of Brønsted acid sites with the Co/H-ZSM5-I-400 catalyst, but the cobalt nano-particles were highly dispersed on the former one, which increased the chance for the formed n-paraffin to be captured by the adjacent acidic sites. Thus, it resulted in a much better isomerization and hydrocracking performances. Additionally, the much higher ratio of the Brønsted to Lewis acid sites of the Co/H-ZSM5-S-260 catalyst, as compared with the Co/H-ZSM5-I-400 here, might be beneficial to the isomerization and hydrocracking of n-paraffin on the H-ZSM-5 support 8 , as well.</p><p>Usually, the unreduced cobalt oxides will cause the high CH 4 selectivity and the low CO conversion 3 . The large-size metal catalysts have weak MSI and excellent reducibilities compared with the small ones 28 . Here, the Co/H-ZSM5-S-260 catalyst has the higher CO conversion and the lower CH 4 selectivity than the Co/H-ZSM5-I-400 catalyst. It indicates that the Co/H-ZSM5-S-260 catalyst possesses much more active cobalt FTS sites. It is in good agreement with the in situ DRIFTS results (see Supplementary Fig. S5 online). Moreover, the selectivity of the C 5 -C 11 products of the Co/H-ZSM5-S-260 catalyst is around 67%, much higher than that is predicted by the ASF law (about 45% 3 ), and the proportion of the C 5 -C 11 range iso-paraffin products is about 60%. Undoubtedly, this increases the octane number of the oil products greatly. The high selectivity of the gasoline range iso-paraffin here is due to the in situ isomerization and hydrocracking of the long chain n-paraffin at the acidic sites of the H-ZSM5 surrounding the cobalt. These findings demonstrate that the utilization of the combined cobalt and H-ZSM5 bifunctional catalyst synthesized by the physical-sputtering method is an effective way to produce synthetic premier gasoline from syngas directly by a synergistic way. After the reaction, little wax was deposited on the Co/H-ZSM5-S-260 catalyst (see Supplementary Fig. S7, S8 online). Thus, little change on the acidic property of the fresh and used Co/H-ZSM5-S-260 catalysts was observed from the FT-IR spectra for pyridine adsorption (see Supplementary Fig. S9 online). On the contrary, the intensity of the IR bands for both Brønsted and Lewis acid sites seriously dropped for the used Co/H-ZSM5-I-400, probably due to the formation of waxy hydrocarbons therein, considering weaker hydrocracking ability of the Co/H-ZSM5-I-400.</p><!><p>Directly depositing cobalt on the acidic catalysts, such as zeolite, typically has the lower activity for the straightforward synthesis of the gasoline range iso-paraffin than using the separately loaded FTS and acidic catalysts 12,13,29 . The hindrance is mainly due to the low reduction degree of the supported metal catalyst determined by the strong MSI and the spatial arrangement of the FTS-active sites and acidic sites. The bifunctional H-ZSM5-supported cobalt nano-catalyst synthesized by the dry physical-sputtering method introduced herein showed the much higher catalytic performances than the conventional catalyst. This catalyst possesses the homogeneous distribution of the cobalt nano-particles with the narrow size range and high dispersion. Furthermore, the cobalt nano-particles are readily reduced to the metallic state due to the weak MSI. This methodology is not only beneficial to synthesis of the gasoline range iso-paraffin herein, but also provides a new strategy to avoid a negative effect on other catalytic processes resulting from the strong MSI.</p><!><p>Catalyst preparation. The H-ZSM5 support (Su ¨d-Chemie catalyst Co. Ltd., H-MFI-90, 362.3 m 2 g 21 , SiO 2 /Al 2 O 3 5 83.7 in molar ratio) was calcined at 400uC for 2 h before utilization. For the Co/H-ZSM5-S catalyst, the metallic cobalt nano-particles were sputtered onto the pretreated H-ZSM5 powder with a metallic Co sputtering target (purity 99.9%, 5 cm 3 10 cm) in a polygonal barrel-sputtering equipment 17 . The vacuum chamber was carefully evacuated to 8.0 3 10 24 Pa, followed by introducing a pure Ar flow (purity: 99.995%, flow rate 5 29 mL min 21 ) into the chamber until the pressure reached 2.0 Pa. A generated Ar plasma was used to attack the Co target and to sputter Co clusters onto the H-ZSM5 surface (input power 400 W, frequency 13.56 MHz 6 5 KHz, rotating rate 3.5 rpm). The whole experiment took 170 min, and 7.0 wt% of cobalt was loaded onto the H-ZSM5 powder. Thereafter, a 1.0% O 2 /N 2 flow (flow rate 5 29 mL min 21 ) was gradually www.nature.com/scientificreports introduced into the vacuum chamber to reach the atmosphere pressure, and kept for 1 h to stabilize the metallic cobalt supported catalyst before exposure in air.</p><p>For the Co/H-ZSM5-I catalyst, a certain amount of Co(NO 3 ) 2 ?6H 2 O was dissolved in distilled water, followed by impregnation of the aqueous solution onto 5.0 g of the H-ZSM5 powder by a conventional incipient wetness impregnation method. This precursor was then dried overnight at 70uC under vacuum, and calcined at 400uC for 2 h to decompose nitrate completely. The weight percentage of the loaded cobalt was 7.0 wt.%. Before reaction, the catalysts were in situ reduced by hydrogen at the specific temperature (260uC for 1 h or 400uC for 10 h) inside the high-pressure flow-type reactor.</p><p>Catalyst characterization. The HR-TEM images were taken with a Philips Tecnai G2F20 instrument operating at 200 kV. XAFS signals of the Co K-edge were collected using a fluorescence mode at 14W1 beamline of the Shanghai Synchrotron Radiation Facility (SSRF). The storage ring was operated at 3.5 GeV with a current of ,250 mA. A Si (111) double-crystal monochromator was used to reduce the harmonic content in the source beam. The RSFs was achieved by Fourier transforming of the k 3weighted EXAFS data in the range of k 5 3.1-13.1 A ˚21 using a Bessel function window.</p><p>Catalytic testing. The FTS reaction was carried out with a continuous flowing fixedbed reactor from syngas (Ar: 3.0%, CO: 32.3%, balance with H 2 ). Behind the reactor, an ice trap with the solvent and inner standard was equipped to capture the heavy hydrocarbons in the effluent. 0.5 g of the catalyst was used here. Reaction conditions were 260uC, 1.0 MPa, and W catalyst /F 5 10 g h mol 21 . Briefly, the effluent gas released from the reactor was analyzed by an on-line gas chromatograph (Shimadzu, GC-8A) using an active charcoal column equipped with a TCD. The products of light hydrocarbons (C 1 -C 10 ) were also analyzed by an online gas chromatograph (GC-FID, Shimadzu, GC-14B) with a capillary column (J&W Scientific GS-Alumina, i.d. 0.53 mm, length 5 30 m) to separate iso-paraffin and n-paraffin. The products with a carbon number higher than 10 were analyzed by a high temperature distillation-type gas chromatograph (HP-6890).</p><p>Each reaction was continuously implemented for 20 h time-on-stream and the activity generally reached maximum during 1 st -5 th h and then became stable.</p>

Scientific Reports - Nature

Predicting effects on oxaliplatin clearance: in vitro, kinetic and clinical studies of calcium- and magnesium-mediated oxaliplatin degradation

This study evaluated the impact of calcium and magnesium on the in vitro degradation and in vivo clearance of oxaliplatin. Intact oxaliplatin and Pt(DACH)Cl 2 were measured in incubation solutions by HPLC-UV. A clinical study determined changes in plasma concentrations of calcium and magnesium in cancer patients and their impact on oxaliplatin clearance. Kinetic analyses modelled oxaliplatin degradation reactions in vitro and contributions to oxaliplatin clearance in vivo. Calcium and magnesium accelerated oxaliplatin degradation to Pt(DACH)Cl 2 in chloride-containing solutions in vitro. Kinetic models based on calcium and magnesium binding to a monochloro-monooxalato ring-opened anionic oxaliplatin intermediate fitted the in vitro degradation time-course data. In cancer patients, calcium and magnesium plasma concentrations varied and were increased by giving calcium gluconate and magnesium sulfate infusions, but did not alter or correlate with oxaliplatin clearance. The intrinsic in vitro clearance of oxaliplatin attributed to chloride-, calcium-and magnesium-mediated degradation predicted contributions of <2.5% to the total in vivo clearance of oxaliplatin. In conclusion, calcium and magnesium accelerate the in vitro degradation of oxaliplatin by binding to a monochloro-monooxalato ring-opened anionic intermediate. Kinetic analysis of in vitro oxaliplatin stability data can be used for in vitro prediction of potential effects on oxaliplatin clearance in vivo.

predicting_effects_on_oxaliplatin_clearance:_in_vitro,_kinetic_and_clinical_studies_of_calcium-_and_

<!>Results<!>Kinetic analysis of in vitro degradation.<!>Plasma calcium and magnesium concentrations after CaGluc/MgSO 4 infusions in patients.<!>In vitro Prediction of calcium-, magnesium-and chloride-mediated clearance of oxaliplatin.<!>Discussion<!>Materials and Methods<!>and Supplementary Figure).

<p>also been suggested that oxaliplatin and/or its metabolites may increase calcium influx into the cytosol of peripheral neurons, which in turn may enhance the activation of sodium and potassium channels, calcium regulated transcription factors and intracellular signalling involving calcium-dependent protein kinases 20,21 . On the basis of a retrospective clinical study suggesting neuroprotective effects of calcium gluconate and magnesium sulfate (CaGluc/MgSO 4 ) infusions given concurrently with oxaliplatin therapy 22 and supporting preclinical data 23 , these infusions became routinely used without further confirmatory prospective randomised trials or evaluation of any potential pharmacokinetic interactions with oxaliplatin. Since 2013, after two prospective randomised trials definitively showing a lack of neuroprotective effects of these infusions against both acute and chronic forms of oxaliplatin neurotoxicity 24,25 , they are now commonly omitted from the routine clinical practice 26 .</p><p>The potential impact of calcium and magnesium on the in vitro degradation and in vivo clearance of oxaliplatin was previously unknown. Once administered oxaliplatin has been proposed to undergo rapid non-enzymatic biotransformation reactions with water and nucleophiles such as chloride, methionine and glutathione [27][28][29] . It is not known to be subject to CYP-P450 mediated metabolism. The initial reaction in oxaliplatin biotransformation can involve displacement of the oxalate group by water and chloride, resulting in formation of reactive species such as monochloro-, dichloro-, and diaquo-DACH platinum, where DACH is the common abbreviation for cyclohexane-1,2-diamine 29 . Initial degradation of oxaliplatin in the presence of chloride in vitro was previously suggested to lead to formation of an intermediate species, monochloro-monooxalato ring-opened complex, [Pt(DACH)oxCl] −30 (Fig. 1). This complex can convert back to oxaliplatin or transform into the final product, Pt(DACH)Cl 2 , resulting in overall slow degradation of oxaliplatin. The contributions made by these different ligand displacement reactions to the in vivo clearance of oxaliplatin are not well understood. To date, there have been few previous attempts to apply in vitro -in vivo extrapolation methods to predicting effects on the in vivo clearance of oxaliplatin. In drug discovery, techniques for predicting in vivo clearance of drugs using in vitro drug metabolism kinetic data has been used, particularly in vitro hepatocyte or microsomal metabolic stability kinetic data for predicting hepatic clearance of drugs in vivo 31,32 . For drugs, such as oxaliplatin, that are not metabolised by the liver, there is little literature available about the prediction of in vivo clearance using in vitro kinetic data.</p><p>With this background, we sought to evaluate the potential impact of calcium and magnesium on the in vitro degradation and the in vivo clearance of oxaliplatin. Drug stability studies and kinetic modelling were used to explore reactions of oxaliplatin and its degradation products with calcium and magnesium in vitro. A clinical study was undertaken to determine changes in plasma concentrations of calcium and magnesium in cancer patients given oxaliplatin with or without infusions of calcium gluconate and magnesium sulfate, and their impact on oxaliplatin clearance. These in vitro and clinical datasets provided an opportunity to develop and exemplify experimental approaches for prediction of effects on oxaliplatin clearance in vivo from oxaliplatin stability data generated in vitro.</p><!><p>In vitro degradation of oxaliplatin. To evaluate the potential impact of calcium and magnesium on the in vitro degradation of oxaliplatin, oxaliplatin was incubated in water and solutions containing NaCl, CaCl 2 and MgCl 2 at 37 °C for 8 hours. Incubation samples were collected and analysed using a validated HPLC-UV method 33 to determine oxaliplatin and Pt(DACH)Cl 2 concentrations at pre-defined time points over the incubation period. Data were presented as plots of concentration versus time and the mass balance of the reaction (Fig. 2). Oxaliplatin was unstable in the presence of chloride and degraded to Pt(DACH)Cl 2 via an intermediate species (Fig. 1), as previously shown by Jerremalm et al. 30 who had identified the intermediate to be [Pt(DACH) oxCl] − . In contrast, oxaliplatin remained stable in water (Fig. 2). The rate of oxaliplatin degradation to Pt(DACH) Cl 2 was accelerated in the presence of calcium and magnesium ions in association with higher transient formation of the intermediate species than was observed in NaCl. 1.</p><!><p>The kinetic model proposed by Jerremalm et al. 29,30 for the stepwise reaction of oxaliplatin with chloride, comprising of a reversible first step producing [Pt(DACH)oxCl] − followed by an irreversible second step producing Pt(DACH)Cl 2 (Fig. 1), was fitted to the mean measured concentration values obtained at time points during the in vitro degradation studies. Rate constants (Table 1) were obtained by least-squares fitting to the analytical expressions shown below (Equations (1-5) for the reaction scheme in Fig. 1 to the three species simultaneously 34 . Standard errors for the rate constants were derived from a Jack-knife procedure 35 .</p><p>2</p><p>(1)</p><p>The parameters λ 1 and λ 2 are the observed biexponential decay constants, and are combinations of the rate constant for oxalate ring opening k 1 , the rate constant for ring closing k −1 , and the rate constant for oxalate loss k 2 . The kinetic models fitted the experimental data for all conditions (Fig. 2). Addition of CaCl 2 or MgCl 2 increased the observed rate of degradation of oxaliplatin and caused increased amounts of the intermediate species to form. However, analysis of the data showed that the k 1 rate constants for the oxalate ring opening step in the oxaliplatin degradation reaction were similar in 150 mM NaCl, 75 mM CaCl 2 and 75 mM MgCl 2 , while the k −1 rate constants for the back reaction corresponding to oxalate chelate ring closing and reformation of oxaliplatin from the [Pt(DACH)oxCl] − intermediate were decreased in 75 mM CaCl 2 and 75 mM MgCl 2 as compared to 150 mM NaCl. Finally, the values for the k 2 rate constant for the loss of the monodentate oxalate ligand from [Pt(DACH) oxCl] − leading to formation of Pt(DACH)Cl 2 were increased in 75 mM CaCl 2 and 75 mM MgCl 2 as compared to 150 mM NaCl. Thus, the observed increase in degradation is not due to an increased rate of the initial oxalate ring opening, but rather is due to a combination of a decrease in the reverse rate of oxalate ring closure combined with an increase in the rate of loss of oxalate.</p><p>These findings suggested that calcium and magnesium did not interact significantly with oxaliplatin at the concentrations studied, but bound to the monodentate oxalate-containing complex [Pt(DACH)oxCl] − , and in doing so, decreased the rate of oxalate ring closure and increased the rate of loss of oxalate from the ring-opened intermediate. The kinetics of the reaction mechanism were therefore modelled further by including an additional metal bonded intermediate species as shown in the shaded area of Fig. 1. With the assumption that the rate constant for the reformation of oxaliplatin from the metal bonded intermediate (k −1,M ) was zero, then the calculated values of the equilibrium constant (K) for the reversible conversion of [Pt(DACH)oxCl] − to the metal bonded intermediate were 20 M −1 for Ca 2+ and 30 M −1 for Mg 2+ . If k −1,M was not zero then the equilibrium constants would be higher. These equilibrium constant values indicate that in 75 mM CaCl 2 or MgCl 2 60% or 70% respectively of the [Pt(DACH)oxCl] − is ion-paired with the metal. Using these equilibrium constant values, the values of the rate constant for the conversion of the metal bonded intermediate to Pt(DACH)Cl 2 (k 2,M ) were calculated to be 3.3 hr −1 and 1.5 hr −1 for calcium and magnesium, respectively.</p><!><p>The in vitro findings described above suggested that calcium and magnesium may alter the in vivo clearance of oxaliplatin by accelerating its degradation via binding to [Pt(DACH)oxCl] − . Plasma calcium and magnesium concentrations vary between patients, on different occasions, and are altered in the presence of cancer, associated disease processes and treatments 36,37 . Until recently, CaGluc/MgSO 4 infusions had been routinely given concurrently with oxaliplatin for the purpose of limiting neurotoxicity 26 but their effects on oxaliplatin clearance or plasma calcium and magnesium levels had not been well studied. In a randomised placebo-controlled crossover clinical study, we sought to understand how variation in plasma calcium and magnesium concentrations, and the administration of CaGluc/MgSO 4 infusions, influenced oxaliplatin clearance in cancer patients receiving chemotherapy, by measuring plasma concentrations of calcium and magnesium before and after CaGluc/MgSO 4 and placebo infusions given with oxaliplatin. We previously reported that oxaliplatin clearance was not altered by giving CaGluc/MgSO 4 infusions 24 . The mean oxaliplatin clearance was 35.3 L/hr (SD 9.8) with CaGluc/MgSO 4 infusions and 33.6 L/hr (7.7) with placebo infusions (p = 0.17). Here we report that plasma concentrations of calcium and magnesium were significantly higher after the first CaGluc/MgSO 4 infusions than at baseline (calcium 1.04-fold increase, p < 0.001; magnesium 1.2-fold increase, p < 0.001) (Fig. 3). Plasma calcium concentration had returned to near baseline levels immediately prior to the second CaGluc/MgSO 4 infusion two hours later but increased again after the second infusion (1.09-fold increase, p < 0.001). Plasma magnesium concentration, however, remained significantly higher (1.13-fold increase, p < 0.001) than baseline when the second infusion was due and were increased further following the second infusion (1.5-fold increase, p < 0.001). There were no significant changes in plasma calcium and magnesium levels after the placebo infusions. Oxaliplatin clearance values measured on treatment cycles given with CaGluc/MgSO 4 or placebo infusions were plotted against their corresponding maximal plasma calcium and magnesium levels for each study patient (Fig. 4). Maximum plasma concentrations of calcium and magnesium were within normal limits after placebo infusions. After CaGluc/MgSO 4 infusions, however, maximum plasma levels exceeded the upper limit of normal in 4 (21%) and 19 (100%) patients for calcium and magnesium, respectively. According to Common Toxicity Criteria Adverse Effect version 4.0 grading, elevated plasma calcium were severity grade 1 in 3 patients (16%) and grade 2 in 1 patient (5%), while elevated plasma magnesium was severity grade 1 in 12 patients (63%) and grade 3 in 7 patients (37%). Pearson co-efficient correlation analysis showed that oxaliplatin clearance did not correlate with plasma calcium (p = 0.33) or magnesium (p = 0.60) concentrations.</p><!><p>To estimate potential contributions of calcium-, magnesium-and chloride-mediated degradation to the in vivo clearance of oxaliplatin in patients, an in vitro -in vivo extrapolation method was developed. First, the in vitro intrinsic clearances of oxaliplatin attributable to calcium-, magnesium-and chloride-mediated degradation were calculated. Non-compartmental analysis was used to calculate the AUC 0−infinity for oxaliplatin concentration versus time for each experimental condition. Then the total in vitro clearance of oxaliplatin for each condition was obtained by dividing the amount of oxaliplatin added to the solution per unit volume by the AUC 0−infinity . Oxaliplatin clearance attributable to calcium-or magnesium-mediated degradation was then calculated by subtracting the clearance attributable to chloride from the total in vitro clearance value for the relevant experimental condition. Calculated values for oxaliplatin clearance attributable to chloride, calcium and magnesium were then plotted against the concentrations of the relevant ion in the incubation solution and analysed by linear regression. The in vitro intrinsic clearance attributable to chloride, calcium and magnesium was taken from the slope of the linear regression fit to its data as shown in Table 2 and Supplementary Figure. These data were then scaled to the in vivo setting by calculating the total extracellular fluid (ECF) content of calcium, magnesium and chloride from estimates of ECF volume and measured plasma concentrations of these ions in patients from our clinical study. In vivo clearance predictions were then made by multiplying the in vitro intrinsic clearance value by the calculated ECF content value of each ion.</p><p>The in vitro intrinsic clearance of oxaliplatin attributable to calcium-, magnesium-and chloride-mediated degradation was calculated to be 0.0039, 0.0023 and 0.00037 L/hr/mmol, respectively (Table 3). The predicted in vivo oxaliplatin clearances attributable to calcium, magnesium and chloride were 0.15 L/hr, 0.03 L/hr and 0.68 L/ hr, respectively, before CaGluc/MgSO 4 infusions. After the CaGluc/MgSO 4 infusion, these predicted in vivo oxaliplatin clearance values increased slightly to 0.16 L/hr for calcium and 0.05 L/hr for magnesium but remained unchanged for chloride. The combined oxaliplatin clearance attributable to the sum of calcium-, magnesium-and chloride-mediated degradation was predicted to be 0.86 and 0.89 L/hr before and after CaGluc/MgSO 4 infusions, respectively. Considering the total in vivo clearance of oxaliplatin measured in our clinical study (35.3 L/hr), calcium-, magnesium-and chloride-mediated degradation was predicted to account for less than 2.5% of the total clearance of oxaliplatin in vivo.</p><!><p>This study showed that calcium and magnesium accelerate oxaliplatin degradation by binding to the monochloro-monooxalato ring-opened anionic oxaliplatin intermediate, [Pt(DACH)oxCl] − , that forms in chloride-containing physiological solutions in vitro. To our knowledge, calcium and magnesium have not been previously reported to alter oxaliplatin degradation or bind to its degradation intermediates. We showed that the in vitro degradation of oxaliplatin to Pt(DACH)Cl 2 was faster in the presence of calcium and magnesium by directly quantifying both intact oxaliplatin and Pt(DACH)Cl 2 using HPLC-UV. Mass balance analysis revealed a deficit after accounting for these two compounds consistent with the transient formation of an intermediate species. Further evidence came from kinetic analyses and rate constants calculated for the reaction scheme in Fig. 1 suggesting that calcium and magnesium interacted with the oxalate ring-opened anionic oxaliplatin intermediate, [Pt(DACH)oxCl] − , resulting in a decreased rate of oxalate ring closure, and an increased loss of monodentate oxalate leading to formation of Pt(DACH)Cl 2 . Oxaliplatin was unstable in the presence of chloride as previous reported by Jerremalm et al. 30 but our findings also provide a mechanism of how calcium and magnesium influence this process of oxaliplatin degradation. This new information may have important implications for understanding how oxaliplatin behaves under in vivo conditions when other reactants or cations are present. This current study provides a new experimental approach for predicting effects on the in vivo clearance of oxaliplatin from in vitro studies of oxaliplatin degradation. It is, to our best knowledge, the first time that in vitro -in vivo extrapolation methods have been applied to predicting effects on oxaliplatin clearance. The concept was based on methods which are already established for in vitro -in vivo prediction of drug clearance mediated by hepatic microsomal metabolism by CYP-P450 and other enzymes, and which are commonly used in pre-clinical studies of the metabolic stability of new chemical entities 32 . However, oxaliplatin is not subject to hepatic metabolism, but degrades via leaving group displacement reactions, although little is known about their contributions to the in vivo clearance of oxaliplatin. To estimate effects on the in vivo clearance of oxaliplatin, we first obtained experimentally determined values for the in vitro intrinsic clearance of oxaliplatin attributable to calcium-, magnesium-and chloride-mediated oxaliplatin degradation, calculated from kinetic analyses of in vitro oxaliplatin stability data. Then, scaling factors were used to estimate the contributions of these processes to the in vivo clearance of oxaliplatin. This experimental approach has potential for providing new insights into mechanisms of oxaliplatin clearance. Previously, for example, it was suggested that oxaliplatin may initially react with water or chloride during its in vivo biotransformation 28,29 . However, we found oxaliplatin to be stable in water in vitro, and that chloride-mediated oxaliplatin degradation was predicted to contribute only 1.9% to the total in vivo clearance of oxaliplatin. This stability of oxaliplatin in pure water or in chloride may be due to slow oxalate ring opening and/or fast oxalate ring closing. These findings suggest that oxaliplatin degradation mediated only by water or chloride may contribute less than previously thought to the total clearance of oxaliplatin in vivo, as suggested by a low formation of Pt(DACH)Cl 2 from oxaliplatin under in vivo conditions 24,38 .</p><p>In cancer patients receiving oxaliplatin chemotherapy, we found that plasma calcium and magnesium concentrations were increased significantly after the infusions of calcium gluconate and magnesium sulfate given immediately before and after oxaliplatin chemotherapy. These infusions were, until recently, routinely used in the clinic for the purpose of reducing neurotoxicity. Statistically significant increases in plasma calcium and magnesium occurred following each of the two CaGluc/MgSO 4 infusions. Post-infusion elevations in plasma calcium were CTCAE severity grade 1 in three patients and grade 2 in one patient, while elevations in plasma magnesium were CTCAE grade 1 in twelve patients and grade 3 in seven patients. There has not been a previous study, which we are aware of, showing these effects on plasma calcium and magnesium levels in patients immediately or soon after these infusions prior to this current study. In our study, plasma calcium and magnesium levels were measured only at baseline, 20 minutes after the first CaGluc/MgSO 4 infusions and immediately after the second CaGluc/ MgSO 4 infusion. We did not fully evaluate the time-course, duration, extent or clinical safety concerns related to these treatment-associated elevations in plasma calcium and magnesium levels. Previously, Gamelin et al. 39 reported no difference in plasma calcium and magnesium levels after the end of oxaliplatin infusion given without CaGluc/MgSO 4 infusions. Their reported values were very similar to those found in the current study at baseline or after placebo infusions. This clinical study also provided an opportunity to explore potential relationships between plasma calcium and magnesium levels and the clearance of oxaliplatin. Plasma calcium and magnesium varied widely between different patients and treatment cycles but did not correlate with oxaliplatin clearance. This finding was in keeping with the in vitro prediction that calcium-and magnesium-mediated degradation contributed a relative small amount (<2.5%) to the total in vivo clearance of oxaliplatin.</p><p>The method we describe for the in vitro prediction of effects on the in vivo clearance of oxaliplatin could assist with translating new treatments to clinical evaluation for preventing oxaliplatin neurotoxicity. Many new candidate treatments for preventing neurotoxicity of oxaliplatin are being identified by preclinical studies, for example, L-type calcium channel blockers 20 and calcium/calmodulin-dependent kinase inhibitors 21 . Our method could identify the potential for deleterious pharmacokinetic interactions with oxaliplatin prior to clinical studies. In this way, this in vitro assessment will complement trial design and endpoints we recently proposed for the early clinical evaluation of investigational treatments for preventing oxaliplatin neurotoxicity 40 .</p><p>In conclusion, calcium and magnesium accelerate the in vitro degradation of oxaliplatin by binding to a monochloro-monooxalato ring-opened anionic intermediate. Kinetic analyses of in vitro oxaliplatin stability data predicted the contributions of calcium-, magnesium-and chloride-mediated degradation to the total in vivo clearance of oxaliplatin in patients. In vitro-in vivo extrapolation methods can be used in future studies for the in vitro prediction of potential effects on oxaliplatin clearance in vivo.</p><!><p>In vitro incubation studies. Oxaliplatin and Pt(DACH)Cl 2 were obtained from Sigma-Aldrich (St Louis, MO, USA). Hydroxyethylpiperazine-N'-2-ethane sulfonic acid ('HEPES') was obtained from Gibco-BRL Life Technologies (Grand Island, NY, USA). Powder sodium chloride (NaCl) and magnesium chloride (MgCl 2 ) anhydrous were obtained from Sigma Aldrich (St Louis, MO, USA). Calcium chloride (CaCl 2 ) was obtained from Riedel-de Haen AG (Germany). Methanol of chromatographic grade, triflic acid (98% reagent grade) was obtained from Sigma Aldrich (St Louis, MO, USA). All solutions and mobile phase were prepared using Milli-Q grade water (Millipore, Bedford, USA).</p><p>To study the stability of oxaliplatin in the presence of chloride, calcium and magnesium under physiological conditions, oxaliplatin (100 μM prepared in Milli-Q water) was incubated in water alone, NaCl (15, 50, and 150 mM), CaCl 2 (1.8, 37.5 and 75 mM), and MgCl 2 (1.8, 37.5 and 75 mM) in HEPES buffer at pH 7.3 and 37 °C. A temperature regulated water-bath (Julabo TW12, John Morris Scientific Ltd., Auckland) was used to maintain temperature during incubation. Samples for analysis were taken at 0, 5, 10, 20, 30, and 60 minutes then hourly thereafter until 8 hours. To detect oxaliplatin and its degradation product, Pt(DACH)Cl 2 , these incubation samples were analysed using a Hewlett Packard HP1200 HPLC online system. This included a binary pump, a degasser and an autosampler (Wilmington, DE, USA), a Waters µBondapak C 18 3.9 × 300 mm column (Waters, Massachusetts, USA) with a guard column (Phenomenex, Torrance, LA, USA). The UV detector used was Millipore Waters Lambda-max model 480 LC Spectrophometer (Millipore, Land Cove, Australia). The HPLC separations were performed at room temperature using a mobile phase containing 3% methanol in Milli-Q water (adjusted to pH 2.5 with triflic acid). The flow rate was 0.5 mL/min. The UV wavelength monitored was 210 nm with a reference of 550 nm. The injection volume of all samples was 50 µL. Data acquisition and processing were performed using HP4500 ChemStation and HP1200 Agilent ChemStation offline software B.04.01 (Agilent Technologies, Avodale, USA). The samples were analysed immediately after they were taken from the incubation solutions whenever possible, otherwise they were snap frozen using liquid nitrogen and thawed within a minute before being injected onto the HPLC for analysis to avoid continuation of degradation of oxaliplatin at higher temperatures. All samples from each incubation study were analysed on the same day using the same mobile phase. Concentrations of oxaliplatin and Pt(DACH)Cl 2 were determined by using their respective calibration curves to convert the areas under the chromatographic peaks to concentration values. The concentrations of the intermediate species were estimated from a mass balance after accounting for oxaliplatin and Pt(DACH)Cl 2 . The kinetics of degradation reactions were modelled as described below.</p><p>Clinical study. In our previously reported clinical study 24 , a randomised double-blind placebo-controlled design was used to evaluate the effects of calcium gluconate (1 g) and magnesium sulfate (1 g) (CaGluc/MgSO 4 ) infusions on the pharmacokinetics and acute neurotoxicity of oxaliplatin. Each patient undergoing oxaliplatin chemotherapy was given either CaGluc/MgSO 4 or placebo infusions immediately before and after oxaliplatin infusion on cycle 1 then the opposite study infusion on cycle 2. This study was approved by the Northern Y Regional Ethics Committee (approval number NTY/11/01/005) and was conducted in accordance with its guidelines and regulations. Informed consent was obtained from all study participants. The plasma pharmacokinetic samples were collected at 13 predefined times including at baseline, during and 3 hours post oxaliplatin infusion. The samples used for this current study were taken at baseline, 20 minutes post first CaGluc/MgSO 4 or placebo infusion, immediately prior and post second CaGluc/MgSO 4 or placebo infusions. The plasma was immediately prepared by centrifugation at 4 °C and 5000 G for 5 minutes, then snap-frozen using liquid nitrogen. The samples were stored at −80 °C until analysis. Plasma calcium (albumin adjusted) and magnesium concentrations were measured using Cobas 8000 modular analyser (Roche Diagnostics Ltd., Switzerland) at LabPlus, Auckland City Hospital (Auckland, New Zealand). Data were presented as measured values and the levels of calcium and magnesium were graded using Common Toxicity Criteria Adverse Effect (CTCAE) version 4. Plasma calcium and magnesium concentrations were compared between baseline and post CaGluc/MgSO 4 or placebo infusions by repeated measures one-way ANOVA analysis. Correlation between calcium and magnesium concentrations and observed oxaliplatin clearance for each patient obtained from our previous study 24 was assessed by Pearson correlation analysis. All statistical analyses were performed using PRISM 6, GraphPad, San Diego, USA. Kinetic modelling. Kinetic analysis of oxaliplatin reactivity in the presence of calcium and magnesium was done using the mean concentration values for oxaliplatin and Pt(DACH)Cl 2 at each sampling time. Rate constants were obtained by least-square fitting the analytical expressions for the reaction scheme to the three species simultaneously (Fig. 1 -unshaded area), using the Solver function in Microsoft Excel with the three rate constants k 1 , k −1 , and k 2 as adjustable parameters. Estimates of the standard errors for these parameters were then obtained using a Jack-knife procedure.</p><p>To determine in vitro intrinsic clearance of oxaliplatin attributable to chloride-, calcium-and magnesium-mediated degradation, the following steps were undertaken: (1) AUC 0−infinity for oxaliplatin concentration versus time for each in vitro incubation experimental condition was calculated using non-compartmental analysis; (2) total in vitro clearance of oxaliplatin for each condition was calculated by dividing the amount of oxaliplatin in the incubation solution per unit volume by the corresponding AUC 0−infinity ; (3) oxaliplatin clearance attributable to calcium-and magnesium-mediated degradation was calculated by subtracting the clearance attributable to chloride from the total in vitro clearance value for each condition; (4) calculated values for oxaliplatin clearance attributable to these ions were then plotted against the concentrations of the corresponding ions in the incubation solution and analysed by linear regression; and (5) finally, the in vitro intrinsic clearance attributable to chloride, calcium and magnesium was taken from the slope of the linear regression fit to its data (Table 2</p><!><p>To extrapolate in vitro kinetic data from the incubation studies to in vivo oxaliplatin clearance attributable to plasma chloride, calcium and magnesium, a scaling factor was applied to the calculated intrinsic clearance for each of chloride, calcium and magnesium ions. The approach to scaling used in our study was a modified version of those used in the microsomal in vitro-in vivo extrapolation methods 32 . We used the estimated extracellular fluid (ECF) contents of chloride, calcium and magnesium in each patient. First, total body water was estimated as 60% of body weight in men and 50% in women, and the ECF volume was estimated as 40% of total body water. The ECF content of chloride, calcium and magnesium was then calculated by multiplying the measured plasma concentrations of these ions by the estimated ECF volume in each patient. Finally, in vivo clearance of oxaliplatin mediated by each of chloride, calcium and magnesium was calculated using the following formula: in vivo CL = in vitro CL int x ECF content, where units for CL = L/h; CL int = L/h/mmol; and ECF content = mmol).</p>

Scientific Reports - Nature

Synthesis of (\xc2\xb1)-7-Hydroxylycopodine

A seven-step synthesis of (\xc2\xb1)-7-hydroxylycopodine that proceeds in 5% overall yield has been achieved. The key step is a Prins reaction in 60% sulfuric acid that gave the key tricyclic intermediate with complete control of the ring fusion stereochemistry. A one-pot procedure orthogonally protected the primary alcohol as an acetate and the tertiary alcohol as a methylthiomethyl ether. The resulting product was converted to 7-hydroxydehydrolycopodine by heating with KO-t-Bu and benzophenone in benzene followed by acidic workup. During unsuccessful attempts to make optically pure starting material, we observed the selective Pt-catalyzed hydrogenation of the 5-phenyl group of a 4,5-diphenyloxazolidine under acidic conditions and the Pt-catalyzed isomerization of the oxazolidine to an amide under neutral conditions. In attempts to hydroxylate the starting material so that we could adapt this synthesis to the preparation of (\xc2\xb1)-7,8-dihydroxylycopodine (sauroine) we observed the novel oxidation of a bicyclic vinylogous amide to a keto pyridine with Mn(OAc)3 and to an amino phenol with KHMDS and oxygen.

synthesis_of_(\xc2\xb1)-7-hydroxylycopodine

Introduction<!>Synthesis of tricyclic model 26<!>Synthesis of 7-hydroxylycopodine (5)<!>Biological Activity<!>Approaches to the Synthesis of Optically Pure 12<!>Approaches to introduce the additional hydroxy group of sauroine<!>General Experimental Methods<!>2,3,4,6,7,8-Hexahydro-7-methyl-5(1H)-quinolinone (12)7b,9<!>2,3,4,6,7,8-Hexahydro-1,7-dimethyl-5(1H)-quinolinone (13)7b<!>trans-1,2,3,4,6,7,8,8a-Octahydro-1,7-dimethyl-5-(1-methylethoxy)-8a-(2-propynyl)quinoline (16)<!>Propargylmagnesium bromide<!>1-(2,3,4,6,7,8-Hexahydro-1-methyl-5(1H)-quinolinylidene)-2-propanone (20)<!>(\xc2\xb1)- (4aS,5R,7S,8aS)-Octahydro-5-hydroxy-1,7-dimethyl-1H-5,8a-propanoquinolin-10-one (26)<!>1-Hydroxybicyclo[3.3.1]nonan-3-one (28)<!>3-Chloro-1-ethoxybicyclo[3.3.1]non-2-ene (31)<!>1-(3-(Dimethylethyl)dimethylsiloxypropyl)-2,3,4,6,7,8-hexahydro-7-methyl-5(1H)-quinolinone (34)<!>trans-1,2,3,4,6,7,8,8a-Octahydro-1-(3-(dimethylethyl)dimethylsiloxypropyl)-7-methyl-5-(1-methylethoxy)-8a-(2-propynyl)quinoline (36a)<!>trans-1,2,3,4,6,7,8,8a-Octahydro-1-(3-(dimethylethyl)dimethylsiloxypropyl)-7-methyl-5-(1-methylethoxy)-8a-(3-trimethylsilyl-2-propynyl)quinoline (36b)<!>(Trimethylsilyl)propargyllithium<!>(\xc2\xb1)-(4aS,5R,7S,8aS)-Octahydro-5-hydroxy-1-(3-hydroxypropyl)-7-methyl-1H-5,8a-propanoquinolin-10-one (6)<!>(\xc2\xb1)-(4aS,5R,7S,8aS)-1-(3-Acetoxypropyl)-octahydro-5-(methylthio)methoxy-7-methyl-1H-5,8a-propanoquinolin-10-one (40)<!>(\xc2\xb1)-(4aS,5R,7S,8aS)-Octahydro-1-(3-hydroxypropyl)-5-(methylthio)methoxy-7-methyl-1H-5,8a-propanoquinolin-10-one (41)<!>(\xc2\xb1)-(8aS,9R,11S,12aR)- 3,4,6,7,8,8a,9,10,11,12-Decahydro-9-hydroxy-11-methyl-1,9-ethanobenzo[i]quinolizin-14-one (42)<!>(\xc2\xb1)-(1S,8aS,9R,11S,12aR)-Dodecahydro-9-hydroxy-11-methyl-1,9-ethanobenzo[i]quinolizin-14-one (7-hydroxylycopodine, 5)<!>(1S,2R,3aS,5aS,6S,8S,9aR)-2-Cyclohexyl-decahydro-8-methyl-1-phenyl-5H-oxazolo[3,2-a]quinolin-6-ol (47)<!>(4aS,5S,7S,8aR)-5-Hydroxy-7-methyl-octahydro-1-[(1R)-1,2-diphenylethyl]-2(1H)-quinolinone (50)<!>1-Benzyl-2,3,4,6,7,8-hexahydro-7-methyl-5(1H)-quinolinone (55) and [(1E)-2-(Methylsulfinyl)ethenyl]-benzene (56)<!>1-Benzyl-2,3,4,6,7,8-hexahydro-7-methyl-5(1H)-quinolinone (55)<!>7,8-Dihydro-7-methyl-5(6H)-quinolinone (58)<!>1,2,3,4-Tetrahydro-1,7-dimethyl-5-quinolinol (59)<!>5-(Dimethylethyl)dimethylsiloxy-1,2,3,4-tetrahydro-1,7-dimethylquinoline (60), (\xc2\xb1)-(4aR,7S)-5-(dimethylethyl)dimethylsiloxy-4a-((dimethylethyl)dimethylsilyl)dioxy-1,2,3,4,4a,7-hexahydro-1,7-dimethylquinoline (61), and (\xc2\xb1)-(4aS,7S)-5-(dimethylethyl)dimethylsiloxy-4a-((dimethylethyl)dimethylsilyl)dioxy-1,2,3,4,4a,7-hexahydro-1,7-dimethylquinoline (62)<!>6-(Dimethylethyl)dimethylsiloxy-1,3,4,8,9,10-hexahydro-1,4-dimethyl-2,7-azecinedione (65)<!>Assay for Acetylcholinesterase Activity

<p>The lycopodium alkaloids are a large, extensively studied alkaloid family that have been of interest to synthetic chemists for more than 50 years.1 Huperzine A (1), which is probably the medicinally most significant member of this family, is an acetylcholinesterase inhibitor that may have other important roles including treatment of Alzheimer's disease as discussed in recent reviews (see Figure 1).2 We were therefore intrigued by the 2009 report3a that sauroine (4, 7,8-dihydroxylycopodine), which was first reported in 2004,3b improves memory retention in the step-down test in male Wistar rats and significantly improves hippocampal plasticity. The isolation of 7-hydroxylycopodine (5) was also reported in 2004 without any indication of biological activity.4</p><p>Racemic lycopodine (2) was first synthesized by Stork5a and Ayer5b in 1968, with additional syntheses by Kim,5c Heathcock,5d Schumann,5e Wenkert,6a Kraus,5f Grieco,5g Padwa5h, and Mori5i over the next 30 years.5l The first synthesis of lycopodine in optically pure form was recently reported by Carter in 2008.5j,k 8-Hydroxylycopodine (3, clavolonine) was synthesized in racemic form in 1984 by Wenkert6a and more recently in optically pure form by Evans,6b Breit,6c and Fujikoka.6d</p><p>The bridgehead hydroxy group of sauroine (4) and 7-hydroxylycopodine (5) is not compatible with most of the approaches that have been used for the synthesis of lycopodine (2) and clavolonine (3). Fortunately, the presence of a 3-hydroxycyclohexanone suggests that this ring might be formed by an aldol or Prins reaction. We decided to start by synthesizing 7-hydroxylycopodine (5) and then to adapt the route to the synthesis of sauroine (4) that contains the additional 8-hydroxy group so that these molecules will be available for study of their biological properties.</p><p>We envisioned that the fourth and final ring of 5 could be constructed from dihydroxy ketone 6 by oxidation of the primary alcohol to an aldehyde, intramolecular aldol reaction and dehydration and then hydrogenation as developed by Heathcock5d in his lycopodine synthesis and later used by Kraus5f and Carter (see Scheme 1).5j,k We hoped that we could prepare 6 from diketone 7 with a suitable protecting group on the primary alcohol by an intramolecular aldol reaction. We planned to introduce the side chain ketone of diketone 7 by a Wacker oxidation of the allyl group of ketone 8, which will be prepared by hydrolysis of enol ether 9. This approach is intriguing because Wiesner and co-workers developed an efficient route to 15, the analogue of 9 with an N‐methyl rather than a protected N‐hydroxypropyl group in a very early unsuccessful approach to lycopodine synthesis.7b This route to 7-hydroxylycopodine is very short and attractive. Our one major concern was the stereochemistry of protonation of the enol ether of 9 which could give both the desired ketone 8 with a trans ring fusion and the undesired stereoisomer with a cis ring fusion that would lead to 12-epi-7-hydroxylycopodine. This was of particular concern because related syntheses by Wiesner led to 12-epi-lycopodine, rather than the desired target lycopodine.7c,d The brevity of this route is appealing and the stereochemical question is addressed early in the synthesis so we chose to explore this route.8</p><!><p>We started with a model study designed to prepare the tricyclic skeleton of 6 with an N-methyl, rather than an N-hydroxypropyl, group because the nitrogen substituent should have no effect on the stereochemistry of the enol ether protonation, Wacker oxidation or aldol cyclization. Weisner7b prepared 129 in unspecified yield by heating 5-methylcyclohexane-1,3-dione (11) with 3-amino-1-propanol in benzene at reflux and then heating the resulting vinylogous amide with 1 equivalent of pyridine hydroiodide neat at 140 °C for 2 h (see Scheme 2). We prepared 12 by microwave heating diketone 11 with 3-bromopropylammonium bromide (10) and 2,6-lutidine in ethanol at 130 °C for 20 min in a modification of a literature procedure for related compounds that heated for 1 h in butanol at reflux (see Scheme 2).10a Methylation of vinylogous amide 12 with NaH and iodomethane in THF gave 137b in 48% overall yield from 11. Using Wiesner's protocol we converted 13 to 15 by heating 13 in excess 2-iodopropane for 42 h at 55 °C and concentration to give the unstable salt 14, which was immediately taken up in THF and treated with allylmagnesium chloride in THF for 2 h at 0 °C to give enol ether 15 as a single stereoisomer in 55% yield from 13.</p><p>Stirring enol ether 15 in 1 M hydrochloric acid for 5 h at 25 °C afforded a ∼3:1 mixture of stereoisomeric ketones 17 in 77% yield. The major stereoisomer was easily isolated in pure form, but the stereochemistry of the ring fusion could not be easily assigned by analysis of the NMR spectral data. We therefore used mixture 17 for our initial explorations of the Wacker oxidation. Unfortunately, all attempts to carry out the Wacker oxidation of 17 to give diketone 18 were unsuccessful. We were somewhat concerned that the amino group might be interfering, although Wacker oxidations of amino alkenes to amino methyl ketones have been reported under acidic conditions in which the amine is protonated.11</p><p>We then attempted to prepare diketone 18 by hydrolysis of enol ether alkyne 16. Addition of propargylmagnesium bromide to a solution of salt 14 in THF provided enol ether alkyne 16 as a single stereoisomer in 42% yield from vinylogous amide 13. To our surprise hydration of the triple bond of 16 with HgO in 1:1 MeOH/1 M aqueous sulfuric acid at 65 °C provided dienone 20 in 39% yield rather than the desired diketone 18. Presumably hydrolysis of the alkyne affords the desired dione 18 with a protonated amino group. Elimination of the β-amino group leads to monocyclic amino dione 19, which condenses with the cyclohexanone to form the fully conjugated δ-amino dienone 20. In retrospect, the failure of the Wacker oxidation of 17 might be due to the decomposition of the desired diketone 18 under the acidic reaction conditions to give compounds that are further oxidized by Pd(II).</p><p>Although enol ether alkyne 16 was initially prepared as a possible precursor to diketone 18, it fortuitously allowed us to develop a one-step route to the desired model tricyclic hydroxy ketone 26 that does not proceed through diketone 18 (see Scheme 3). In 1965, Wiesner and coworkers reported that treatment of enol ether alkene 15 in 75% sulfuric acid for 12 h at 25 °C afforded tricyclic alkene 24 in 70% yield as a single isomer with unknown stereochemistry at the ring fusion and unknown double bond position.7b Acidic hydrolysis of the enol ether afforded ketone 21 (protonated form of ketone 17), which underwent a Prins cyclization to give secondary cation 22. A facile transannular 1,5-hydride shift then took place to form the more stable tertiary cation 23. Loss of a proton then formed the observed product 24. The desired ring fusion stereochemistry with a β-hydrogen, which would result from the trans-fused stereoisomer of bicyclic ketone 21 was suggested on the basis of kinetic and thermodynamic conformational analysis. However later work from this group leading to two syntheses of 12-epi-lycopodine7c,d using related chemistry and our observation that hydrolysis of 15 with hydrochloric acid afforded 17 as a ∼3:1 mixture of stereoisomers indicate that this stereochemical assignment is far from secure. Unfortunately, the functionality is now in the wrong ring of 24, which makes it useless for lycopodine synthesis regardless of the stereochemistry at C12.</p><p>Enol ether alkyne 16 was treated with aqueous sulfuric acid with the expectation that enol ether hydrolysis and Prins cyclization would occur analogously to give alkenyl cation 25. A 1,5-hydride shift should not occur because the hydrogen is too far from the vacant sp2 orbital on the alkenyl cation. Instead, the alkenyl cation should react with water to give an enol that will tautomerize to the desired tricyclic hydroxy ketone 26. We were pleased to find that reaction of 16 in 60% sulfuric acid for 8 h at 25 °C afforded 26 in 70% yield as a single stereoisomer. The yield was slightly lower in 75% sulfuric acid. In 50% sulfuric acid, the enol ether hydrolyzed to give the ketone, which did not add to the alkyne. The stereochemistry of 26 was assigned from a series of 1D NOESY experiments as shown in Figure 2. The stereochemistry at C12 can be unambiguously assigned from the NOEs between H12 and H14 and between H6 and H11.</p><p>The stereoisomer derived from the cis-fused bicyclic ketone was not observed. The origin of this stereoselectivity is not obvious. The trans-fused bicyclic ketone is calculated by molecular mechanics to be 1.4 kcal/mol more stable than the cis isomer, but bicyclic ketone 17 was formed as a ∼3:1 mixture of stereoisomers by hydrolysis of enol ether alkene 15.</p><p>Prins reactions with alkynes are uncommon, but known.12 The isolation of β-hydroxy ketones from Prins reactions of keto alkynes is rarely observed because these initial products will usually undergo dehydration to give conjugated ketones. In this case, ring strain prevents dehydration to form a conjugated ketone, which would have a bridgehead double bond.</p><p>We briefly explored the scope of this alkyne ketone Prins reaction as a route to bicyclic hydroxy ketones. Stirring 3-propargylcyclohexanone (27)13 in 60% sulfuric acid for 16 h at 25 °C provided 1-hydroxybicyclo[3.3.1]nonan-3-one (28)14 in 76% yield indicating that the protonated nitrogen of 16 is not necessary for the reaction (see Scheme 4). Similar treatment of 3-propargylcyclopentanone (29)13 gave a complex mixture and N-propargylpiperidine-4-one (30)15 was recovered unchanged from 60%, 80%, or concentrated sulfuric acid. These initial results suggest that this Prins reaction will be best suited for the synthesis of 1-hydroxybicyclo[3.3.1]nonan-3-ones.</p><p>We wanted to explore the trapping of the vinyl cation intermediate with nucleophiles other than water. We therefore explored the reaction of 27 with EtAlCl2 in CH2Cl2 (see Scheme 5).16 Initial reactions in which oxygen was not carefully excluded gave chloroalkenyl ether 31 in about 50% yield. However, when we carefully excluded oxygen, we obtained a complex mixture of products. We hypothesized that EtAlCl2 reacted with oxygen to give EtO2AlCl2, which reacted with second molecule of EtAlCl2 to give two molecules of EtOAlCl2, which was responsible for the formation of 31.17,18 We therefore treated EtAlCl2 with 1 equiv of EtOH in CH2Cl2 for 1 h to give EtOAlCl2 and ethane. Addition of 27 and stirring for 20 h afforded 31 in 71% yield. Presumably EtOAlCl2 reacts with 27 to give 32, which reacts further to give alkoxonium ion 33, which cyclizes to give 31 after trapping of the vinyl cation with chloride.</p><!><p>For the synthesis of 7-hydroxylycopodine precursor 6 we needed to replace the methyl group of 13 with a protected hydroxypropyl group. We chose to use the TBS group because it should be stable to the conditions needed to add the propargyl group, but then be hydrolyzed without an additional step during the sulfuric acid catalyzed cyclization. Crude vinylogous amide 12 was treated with NaH and then I(CH2)3OTBS19 in 1:1 THF/DMF for 18 h at 25 °C to afford 34 in 50% overall yield from 5-methylcyclohexane-1,3-dione (11) (see Scheme 6). Heating a solution of 34 in 2-iodopropane for 42 h at 55°C afforded the intermediate cation 35, which was treated with propargylmagnesium bromide in THF to yield 36a in 40% yield from 34. Stirring a solution of 36a in 60% aqueous sulfuric acid for 8 h at 25 °C afforded the requisite tricyclic ketone 6 with the required hydroxypropyl side chain in 50% yield. As in the hydrolysis and Prins cyclization of 16, the hydrolysis and cyclization of 36a affords 6 as a single stereoisomer whose stereochemistry was established by a series of 1D NOESY experiments that are summarized in Figure 3. The stereochemistry at C12 can be unambiguously assigned from the NOEs between H12 and H14, between H6 and H11, and between H4 and both H9 and H11.</p><p>Propargylmagnesium bromide is best used immediately and its preparation requires the use of toxic HgCl2 in addition to magnesium. We therefore explored the use of other propargyl nucleophiles. Trimethylsilylpropargyllithium, which is easily prepared by deprotonation of 1-trimethylsilylpropyne with n-BuLi, added to cation 35 to give 36b in 51% yield. The preparation of 36b is operationally simpler, proceeds in higher yield and does not produce mercury containing waste. The hydrolysis and cyclization of 36b in 60% sulfuric acid was slightly slower (10 h) than that of 36a but afforded 6 in comparable yield (50%). During the conversion of 36b to 6, the enol ether is hydrolyzed, the TMS alkyne and TBS ether are deprotected and the keto alkyne cyclizes to form the tricyclic hydroxy ketone.</p><p>Heathcock was unable to synthesize lycopodine by an intramolecular alkylation of the analogue of 6 lacking the tertiary hydroxy group20 and therefore used a modified Oppenauer oxidation-aldol reaction with benzophenone and KO-t-Bu to form dehydrolycopodine.5d, 21 Unfortunately, with 6 we obtained a complex mixture containing at most 5% of the desired enone 42 using a variety of bases (KO-t-Bu, NaO-t-Bu, or NaH) and hydride acceptors (benzophenone or fluorenone). We suspect that a retro aldol reaction of the β-hydroxy ketone took place under the very basic reaction conditions. We therefore explored other oxidation conditions that might oxidize the primary alcohol to the aldehyde in the presence of an amine and tertiary alcohol. Decomposition occurs with PCC, PDC, or excess TPAP. No reaction occurred under Swern or Parikh-Doering conditions. Dess-Martin oxidation formed the β-amino aldehyde which appeared to be oxidized further to give the β-aminoenal R2NCH=CHCHO based on NMR peaks at δ 9.16 (d, 1, J = 7.6 Hz), 7.39 (br d, 1, J = 12.4 Hz), and 5.33 (dd, 1, J =12.4, 7.6 Hz).22 Oxidation with IBX in DMSO gave mainly recovered 6 containing a trace of β-aminoenal. Eventually we found that the primary alcohol of 6 could be selectively oxidized to unstable amino aldehyde 37 under Narasaka-Mukaiyama conditions with t-BuOMgCl and azodicarbonylpiperidine in THF (See Scheme 7).23 Unfortunately, all attempts to carry out an intramolecular aldol reaction with 37 to give 42 using either acid (TsOH or AcOH), base (KO-t-Bu, K2CO3, or pyrrolidine) or salt (piperidinium acetate, dibenzylammonium trifluoroacetate, or piperidinium trifluoroacetate) catalysis resulted in loss of acrolein to form the unstable tricyclic secondary amine 38.</p><p>We therefore decided to protect the tertiary alcohol of 6 and to then reinvestigate the modified Oppenauer oxidation-aldol reaction sequence. The strongly basic and high temperature conditions of the Oppenauer oxidation preclude the use of an ester or silyl ether protecting group. We initially protected the primary alcohol with pyridine and acetic anhydride to give 39 in quantitative yield (see Scheme 8). Attempted benzylation or allylation of the tertiary alcohol of 39 under a variety of conditions led to complex mixtures, recovered starting material or hydrolysis of the acetate. We were also unable to form the tertiary ethoxyethyl or THP ether from the acid catalyzed reaction of 39 with ethyl vinyl ether or dihydropyran. Eventually we found that treatment of 39 with acetic anhydride in DMSO generated MeS=CH2+ by a Pummerer rearrangement which reacted with the tertiary alcohol to give methylthiomethyl ether 40 in 90% yield.24 Hydrolysis of the acetate of 40 with K2CO3 in MeOH afforded primary alcohol 41 in 92% yield. This sequence was improved by carrying out the first two steps in a single pot. A solution of 6 in Ac2O was stirred for 12 h at 25 °C to form the primary acetate. DMSO was added and stirring was continued for 72 h to form 40 in 90% yield.</p><p>We were pleased to find that treatment of 41 with KO-t-Bu (5 equiv) and benzophenone (16 equiv) in carefully degassed benzene in a sealed tube at 110 °C for 1 h followed by workup with 3 M hydrochloric acid to cleave the methylthiomethyl protecting group gave 7-hydroxydehydrolycopodine (42) in 50% yield. This reaction was carried out with an excess of strong base suggesting that prior deprotection of the primary acetate of 41 was unnecessary. As expected similar treatment of 40 using an extra equiv of KO-t-Bu afforded 42 in 47% yield. In this step, the primary acetate is cleaved, the resulting alcohol is oxidized to the aldehyde, the aldol reaction closes the final ring, dehydration leads to the enone, and the protecting group is cleaved during workup. Hydrogenation of enone 42 over PtO2 with 1 atm of H2 for 10 h afforded 7-hydroxylycopodine (5) in 95% yield. The spectral data for the hydrochloride salt of 5 in CD3OD are identical to those reported for the natural product,4 thereby confirming the stereochemical assignment of 6 made on the basis of NOE experiments.</p><!><p>The crude alkaloid extract of Huperzia Saururus, which contained no huperzine A (2), inhibited human acetylcholinesterase (AChE) with an IC50 of 0.58 μg/mL.3 However, the major alkaloid sauroine (1), which constitutes 57-67% of the alkaloid mixture depending on the harvest season, did not inhibit AChE below 10 μg/mL indicating that there are other potent AChE inhibitors in the crude mixture.3a Although sauroine is a weak AChE inhibitor it did demonstrate a significant increase of hippocampal plasticity and memory retention in rats.3c,d We found that 7-hydroxylycopodine (5) is a weak inhibitor of eel AChE and human erythrocyte cholinesterases with an IC50 of 1.67 and 2.08 mM (439 and 547 μg/mL) respectively, (see Figure 4). At this concentration, 7-hydroxylycopodine does not inhibit human butyrylcholinesterase (data not shown), indicating that the observed IC50 of 2.08 mM is due to human erythrocyte AChE inhibition. In contrast, huperzine A inhibits human erythrocyte acetylcholinesterase with an IC50 of 0.09 μM.2a</p><!><p>We have developed a short, practical, six step synthesis of (±)-7-hydroxylcyopodine (5) from racemic 12, which is prepared from prochiral 5-methylcyclohexane-1,3-dione (11). We briefly explored approaches for the formation of optically pure 12 from 11. Rychnovsky found that Cu(OTf)2 catalyzed reaction of racemic 43 with amino alcohol 44 afforded 45 in 95% yield as a 20:1 mixture of diastereomers which he used for a synthesis of (−)-lycoperine (see Scheme 9).25 Unfortunately, as suggested by Rychnovsky's work, we were unable to remove the chiral auxiliary from 45 by hydrogenation (1-3 atm H2) over Pd/C, Pd(OH)2/C, or PtO2. We obtained either recovered starting material or complex mixtures, whose spectra suggested that the tetrasubstituted double bond had been partially reduced. We observed an apparent quartet at δ 0.43 (J = 12 Hz) similar to that observed in 46 at δ 0.47, whereas the most upfield ring proton of 45 is at δ 1.55.</p><p>Rychnovsky reduced 45 to 46 in 67% yield with sodium in THF/isopropanol and found that reductive removal of the auxiliary from 46 required forcing conditions (500 psi H2, 20% Pd(OH)2 on carbon) to give a lycoperine precursor.25 Although the preparation of 12 from 46, which lacks both the carbonyl group and tetrasubstituted double bond of 45, would not be straightforward, we briefly examined hydrogenolysis of 46 under milder conditions.</p><p>We observed two interesting reactions on attempted hydrogenolysis of 46 over PtO2 under more moderate H2 pressures. Stirring a solution of 46 under H2 (50 psi, 3.3 atm) with PtO2 in 10:1 MeOH/conc HCl for 40 h selectively hydrogenated one of the two phenyl rings to give 47 in 69% yield. This selectivity is precedented in the hydrogenation of 51 to give 52 over Rh/Al2O3 reported by Nugent.26 Under neutral conditions and 1 atm H2 with PtO2 we observed the formation of amide 50 in 73% yield after 40 h. Presumably, platinum inserts in the benzylic carbon-oxygen bond to give 48, which undergoes a β-hydride elimination to form amide 49. Reductive elimination would then form 50. The acidic solution is important for the selective hydrogenation of one benzene ring to give 47, because hydrogenation of 46 under H2 (3 atm) and PtO2 in MeOH for only 10 h afforded only 5-10% of cyclohexane 47, about 50% of amide 50, and 40-45% unreacted 46. The facile isomerization of aminal 46 to amide 50 at 1-3 atm of H2 may be the reason that forcing conditions (500 psi H2) were needed for the reductive removal of the protecting group in Rychnovsky's lycoperine synthesis.25 High H2 pressure should accelerate hydrogenolytic cleavage of intermediate 48 and thereby prevent the formation of amide 50.</p><p>We then turned to the resolution of racemic 12. The anion of 12 was acylated with α-acetoxy and α-methoxyphenylacetyl chloride to give diastereomeric vinylogous imides, which were inseparable by TLC. We then decided to react the anion of 12 with trans-stilbene oxide in the hope that the two diastereomers of 53 with the stereocenters in closer proximity might be separable. To our surprise, treatment of 12 with NaH and trans-stilbene oxide in DMSO at 50 °C for 12 hours gave only the N-benzyl product 55 (12%) and unsaturated sulfoxide 56 (12%) (see Scheme 10). The structure of 55 was confirmed by its preparation from 12, sodium hydride and benzyl bromide. Presumably, the anion of 12 undergoes the desired SN2 reaction with trans-stilbene oxide to give alkoxide 53, which fragments to give benzaldehyde and carbanion 54, which is protonated to give 55. Condensation of benzaldehyde with DMSO is known to give 56.27</p><p>Unfortunately, we were unable to develop an efficient synthesis of optically pure 12, but our attempts led to the observation of the three unusual reactions, namely, the selective hydrogenolysis of one of two benzene rings of aminal 46 to give 47, the Pt-catalyzed isomerization of 46 to amide 50, which may explain why high pressure conditions are needed for the removal of this chiral auxiliary, and the unusual fragmentation of alkoxide 53 to give benzaldehyde and 55.</p><!><p>We then investigated the hydroxylation28 of 34 to give 57 (see equation 1), which could be elaborated to sauroine (4) by a route similar to that used for the synthesis of 7-hydroxylycopodine (5). Our initial studies were carried out with model 13 with an N-methyl group (see Scheme 11). We initially examined the reaction of 13 with Mn(OAc)3 in benzene at elevated temperatures because Watt and Demir had shown that these conditions introduce an α′-acetoxy group onto an α,β-unsaturated ketones or β-alkoxy- α,β-unsaturated ketones.29 Unfortunately, the vinylogous amide 13 reacted differently. Microwave heating a solution of 13 with Mn(OAc)3 in benzene at 150 °C for 20 min afforded a mixture of recovered 13 (35%), demethylated vinylogous amide 12 (20%), and pyridine 5830 (35%). We suspected that the slow step was the demethylation to give 12 and confirmed this by the oxidation of 12 with Mn(OAc)3 in benzene which proceeded efficiently to give 58 in 65% yield with microwave heating for 20 min at only 90 °C. This facile oxidation with Mn(OAc)3 to generate the pyridine is synthetically useful, but does not provide a procedure for the desired α′-hydroxylation.</p><p>We then attempted to form the α′-enolate and trap it with either oxygen, MoO5·pyr·HMPA, or a sulfonyloxaziridine to introduce the α′-hydroxy group.31 Treatment of 13 with 1.5 equiv of KHMDS in THF at −78 °C and then stirring under an oxygen atmosphere led to phenol 59 in 20% yield rather than the desired hydroxy compound (see Scheme 12). Similar results were obtained with NaHMDS. Only starting material was recovered when we added MoO5·pyr·HMPA to the enolate. Adding (8,8-dichlorocamphorylsulfonyl)oxaziridine gave recovered starting material at -78 °C and a complex mixture on warming to 0 °C. Silylation of the α′-enolate of 13 to give a silyl ether that could then be hydroxylated was not practical since the required silyl ether would be a very hydrolytically unstable 1-amino-3-silyloxy-1,3-diene. The formation of phenol 59 from 13 could involve trapping of the enolate with oxygen and a fragmentation process or more likely an electron transfer reaction of the enolate to give an α-keto radical which could then be oxidized further to give the phenol. A related oxidation of a cyclohexenone to a phenol with KO-t-Bu and oxygen that does not proceed through a hydroperoxide has been reported.32</p><p>We then tried to prevent formation of phenol 59 by trapping the intermediates with TBSOTf. Vinylogous amide 13 was treated with NaHMDS (2.2 equiv) in THF at −78 °C, oxygen was bubbled through the solution for 50 min, and excess TBSOTf was added. The NMR spectrum of the crude product indicated the presence of a 1:2:2 mixture of 60, 61, and 62 in about 50% yield. Flash chromatography provided phenyl silyl ether 60 (10%), which was hydrolyzed with 3 M HCl to give 59 in 90% yield, and 61 (22% yield). The other diastereomer 62 decomposed on chromatography. Presumably, deprotonation occurs at the γ-position to give the thermodynamic enolate, which is hydroxylated at the α-position to give the α-hydroperoxy-β,γ-unsaturated ketone which is enolized by excess base at the α′-position. Trapping with TBSOTf will give bis silyl ethers 61 and 62.</p><p>MMX calculations with conformational searching were carried out on the bis TMS ethers corresponding to bis TBS ethers 62 and 61 using PCMODEL 9.3. The Boltzmann averaged coupling constants between the doubly allylic methine hydrogen and the alkene hydrogens were calculated to be 3.1 and 3.1 Hz in the bis TMS ether corresponding to 62, in which the methyl and peroxy groups are trans, and 4.3 and 4.3 Hz in the bis TMS ether corresponding to 61, in which the methyl and peroxy groups are cis. The stereochemistry of 62 and 61 is tentatively assigned based on the observed coupling constants of 2 and 2 Hz in 62 and 4.4 and 4.4 Hz in 61.</p><p>We examined the deprotection of 61 as a means of confirming the structure assignment. Treatment of 61 with pyridine (2.5 equiv) and pyrdine·(HF)x (3 equiv) in CDCl3 for 30 min afforded a compound in 40% yield whose structure was eventually established as 65 by careful analysis of 1D and 2D NMR spectra (see Scheme 13). MMX calculations with conformational searching were carried out on the analogue of 65 with TMS rather than TBS ethers using PCMODEL 9.3. The s-trans amide conformer is calculated to be 5.4 kcal/mol more stable than the s-cis amide conformer. The observed and (calculated) coupling constants between H3a and H4 [13.2 (12.1) Hz], H3b and H4 [4.4 (4.1) Hz] and H4 and H5 [10.0, (8.6)] correspond closely. A plausible pathway for the conversion of 61 to 65 involves the hydrolysis of the peroxy silyl ether and protonation of the enamine to give cation 63, which cyclizes to give dioxetane 64, which fragments33 to give hexahydro-2,7-azacinedione 65. A related oxidative cleavage of the ring fusion double bond of 1-acetyl-1,2,3,4,5,6,7,8-octahydro-6-methylquinoline with RuO4 afforded 1-acetyl(octahydro)-5-methyl-2,7-azacinedione.34</p><p>Although we haven't achieved the desired α′-hydroxylation of 34 to give sauroine precursor 57, we have found novel procedures that oxidize vinylogous amide 12 to keto pyridine 58 and N-methyl vinylogous amide 13 to phenol 59 and that oxygenate 13 at the α-, rather than α′-position to give 61 and 62. Other approaches for the synthesis of 57 are currently being explored.</p><p>In conclusion, we have completed a seven-step synthesis of (±)-7-hydroxylycopodine (5) from 12 that proceeds in 5% overall yield making it readily available for biological evaluation. The key step in the synthesis is the Prins reaction of 36a or 36b in 60% sulfuric acid to give the key tricyclic intermediate 6 with complete control of the ring fusion stereochemistry. We also developed a one-pot procedure to orthogonally protect the primary alcohol of 6 as an acetate and the tertiary alcohol as a methylthiomethyl ether giving 40, which was converted to 7-hydroxydehydrolycopodine 42 by heating with KO-t-Bu and benzophenone in benzene followed by acidic workup. 7-Hydroxylycopodine inhibits eel and human acetylcholinesterase with an IC50 of 1.67 and 2.21 mM, respectively.</p><p>We were unable to develop an efficient synthesis of optically pure 12, but our attempts led to the observation of the three unusual reactions, namely, the selective hydrogenolysis of one of two benzene rings of aminal 46 to give 47, the Pt-catalyzed isomerization of aminal 46 to amide 50, and the unusual fragmentation of alkoxide 53 to give benzaldehyde and 55. Although we haven't achieved the desired α′-hydroxylation of 34 to give sauroine precursor 57, we have found novel procedures that oxidize vinylogous amide 12 to keto pyridine 58 and N-methyl vinylogous amide 13 to phenol 59 and that oxygenate 13 at the α-, rather than α′-position to give 61 and 62.</p><!><p>Reactions were conducted in flame- or oven-dried glassware under a nitrogen atmosphere and were stirred magnetically. The phrase "concentrated" refers to removal of solvents by means of a rotary-evaporator attached to a diaphragm pump (15-60 Torr) followed by removal of residual solvents at < 1 Torr with an vacuum pump. Flash chromatography was performed on silica gel 60 (230-400 mesh). Analytical thin layer chromatography (TLC) was performed using silica gel 60 F-254 pre-coated glass plates (0.25 mm). TLC Plates were analyzed by short wave UV illumination, or by spraying with permanganate solution (5 g KMnO4 in 495 mL water). THF and ether were dried and purified by distillation from sodium/benzophenone. Et3N, pyridine, acetonitrile and benzene were distilled from CaH2. 1H and 13C NMR spectra were obtained on a 400 MHz spectrometer in CDCl3 with tetramethylsilane as internal standard unless specifically indicated. Chemical shifts are reported in δ (ppm downfield from tetramethylsilane). Coupling constants are reported in Hz with multiplicities denoted as s (singlet), d (doublet), t (triplet), q (quartet), p (pentet), m (multiplet) and br (broad). IR spectra were acquired on an FT-IR spectrometer and are reported in wave numbers (cm−1). High resolution mass spectra were obtained using electrospray ionization (ESI).</p><!><p>A solution of 5-methylcyclohexane-1,3-dione (11) (1.20 g, 9.5 mmol), 3-bromopropylammonium bromide (10) (2.17 g, 9.9 mmol, 1.04 equiv) and 2,6-lutidine (3.2 mL, 2.94 g, 27.5 mmol, 3.0 equiv) in 6 mL of EtOH was heated at 130 °C for 20 min in a microwave reactor. The reaction mixture was quenched by addition of 50 mL of 1 M NaOH and extracted with CH2Cl2 (50 mL × 3). The combined organic layers were washed with brine, dried over Na2SO4 and concentrated to about 5 mL. Acetonitrile (30 mL × 2) was added and the resulting solution was concentrated again to give 1.50 g of crude 12 as a brown powder, which was directly used in next step without further purification. A small portion was purified by flash chromatography on silica gel (100:1:1 EtOAc/MeOH/NEt3) to give 12 as a white solid: mp 173-174 °C; 1H NMR 4.56 (br s, 1, NH), 3.32-3.20 (m, 2), 2.41 (dd, 1, J = 16.8, 3.0), 2.34 (apparent t, 2, J = 6.4), 2.22-2.07 (m, 3), 2.02 (dd, 1, J = 16.8, 11.0), 1.86-1.73 (m, 2), 1.04 (d, 3, J = 6.1); 13C NMR 194.2, 158.6, 104.2, 44.8, 41.5, 37.5, 29.0, 21.2, 21.1, 18.9; IR 3239, 3081, 1573, 1526 (strong).</p><!><p>To a suspension of NaH (60% in mineral oil, 454 mg, 11.3 mmol) in 2 mL of THF, a solution of 750 mg of crude 12 in 10 mL of THF was added dropwise at 0 °C. The mixture was stirred at 0 °C for 10 min and 0.71 mL (1.62 g, 11.4 mmol) of MeI was added dropwise. The reaction was warmed to room temperature and stirred for 4 h. The reaction was quenched by addition of water (30 mL) at 0 °C. The mixture was extracted with EtOAc (20 mL × 3). The combined organic layers were washed with brine (60 mL), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 409 mg (48% from 11) of 13 as a white solid: mp 46-47 °C; 1H NMR 3.24-3.17 (m, 2), 2.99 (s, 3), 2.55 (dd, 1, J = 15.9, 4.3), 2.46-2.23 (m, 3), 2.16-1.94 (m, 3), 1.80-1.71 (m, 2), 1.06 (d, 3, J = 6.1); 13C NMR 193.4, 159.6, 105.6, 51.3, 43.9, 38.5, 35.1, 28.8, 21.5, 21.1, 19.5; IR 1611 (weak), 1551 (strong).</p><!><p>To a resealable tube was added 409 mg (2.3 mmol) of 13 and 3 mL (5.1 g, 30 mmol, 13 equiv) of 2-iodopropane. The reaction mixture was sealed and heated at 55 °C for 42 h, diluted with 9 mL of benzene, and concentrated to remove excess 2-iodopropane to give crude cation 14.</p><p>To a solution of the residue (cation 14) in 9 mL of THF, 4.1 mL (1.8 equiv) of propargylmagnesium bromide solution (see below for preparation) was added at 0 °C. The mixture was kept at 0 °C for 1 h. The reaction was quenched by addition of water (1 mL) at −78 °C and filtered through a pad of Celite. The filtrate was diluted with saturated ammonium chloride solution (20 mL) and extracted with EtOAc (30 mL × 3). The combined organic layers were washed with brine (100 mL), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (3:2 hexanes/EtOAc) gave 251 mg (42% from 13) of 16 as a pale yellow oil: 1H NMR 4.08 (heptet, 1, J = 6.1), 3.01-2.93 (m, 1), 2.74 (dd, 1, J = 17.2, 2.4), 2.69-2.57 (m, 2), 2.43 (dd, 1, J = 17.2, 2.4), 2.38 (s, 3), 2.39-2.35 (m, 1), 2.12-2.03 (m, 2), 2.03 (dd, 1, J = 2.4), 1.87-1.77 (m, 1), 1.68-1.46 (m, 3), 1.28-1.22 (m, 1), 1.20 (d, 3, J = 6.1), 1.14 (d, 3, J = 6.1), 1.01 (d, 3, J = 6.7); 13C NMR 145.3, 121.7, 82.7, 71.5, 68.7, 59.3, 50.8, 43.9, 38.7, 35.0, 25.8, 25.7, 22.9, 22.3, 22.1, 20.8, 19.5; IR 3308, 2112 (weak), 1674 (weak), 1115, 1070; HRMS (ESI) calcd for C17H28NO (MH+) 262.2165, found 262.2168.</p><!><p>To a flame-dried flask was added 1.0 g of Mg, 24 mg of HgCl2, and 4 mL of ether. 0.1 mL of propargyl bromide (80% in toluene) was added and the reaction was initiated by heating with a heat gun. The mixture was cooled to 0 °C and a solution of 1.4 mL of propargyl bromide (80% in toluene) in 8 mL of ether was slowly added over 1 h. The reaction was stirred at 0 °C for 0.5 h and allowed to settle at 0 °C for 0.5 h to give a ∼1 M solution.</p><!><p>To a resealable tube was added a solution of 88 mg (0.34 mmol) of 16 in 1 mL of MeOH and a solution of 12 mg (0.055 mmol, 0.16 equiv) of HgO in 1 mL of 1 M H2SO4. The reaction was sealed and heated at 65 °C for 10 h. The reaction was cooled to room temperature and diluted with saturated NaHCO3 solution (20 mL). The mixture was extracted with EtOAc (20 mL × 3). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:0.5 EtOAc/MeOH/NEt3) gave 29 mg (39% from 16) of 20 as a yellow oil: 1H NMR 5.64 (s, 1), 3.78 (dd, 1, J = 17.2, 2.4), 3.28-3.12 (m, 2), 2.93 (s, 3), 2.44 (dd, 1, J = 12.0, 2.9), 2.28 (ddd, 1, J = 14.0, 4.8, 4.8), 2.20-2.10 (m, 1), 2.13 (s, 3), 2.05 (dd, 1, J = 17.2, 11.6), 1.98-1.70 (m, 4), 1.05 (d, 3, J = 6.7); 13C NMR 196.4, 157.9, 153.8, 107.5, 103.3, 50.8, 38.6, 35.6, 35.1, 32.0, 28.5, 22.7, 21.7, 21.6; IR 1637 (weak), 1484 (strong), 1400 (strong); HRMS (ESI) calcd for C14H22NO (MH+) 220.1696, found 220.1694.</p><!><p>A solution of 163 mg (0.62 mmol) of 16 in 5 mL of 3:2 conc H2SO4/H2O was stirred at room temperature for 8 h. The reaction mixture was quenched by addition of water (5 mL) at 0 °C and neutralized with 6 M NaOH until the pH reached 11. The mixture was saturated with NaCl and extracted with CH2Cl2 (30 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 104 mg (70% from 16) of 26 as a white solid: mp 124-125 °C; 1H NMR 2.69 (br d, 1, J = 12.8, H9eq), 2.66 (br d, 1, J = 17.1, H4ax), 2.57 (dd, 1, J = 17.1, 1.4, H6ax), 2.46 (ddd, 1, J = 12.8, 12.8, 3.1, H9ax), 2.40 (d, 1, J = 17.1, H6eq), 2.24 (s, 3, Me1), 2.18 (d, 1, J = 17.1, H4eq), 2.10 (br d, 1, J = 12.8, H11eq), 2.09 (br d, 1, J =12.8, H14eq), 1.85 (br dd, 1, J = 12.8, 4.4, H8eq), 1.82 (br ddd, 1, J = 12.8, 3.2, 3.2, H10eq), 1.70 (ddddd, 1, J = 12.8, 12.8, 12.8, 4.0, 4.0, H10ax), 1.61 (br d, 1, J = 12.8, H12, showed "W" couplings to H4eq and H6eq in COSY), 1.54-1.42 (m, 1, H15), 1.41 (br, 1, w1/2 = 5 Hz, OH), 1.34-1.22 (m, 2, H8ax, H11ax), 1.08 (ddd, 1, J = 12.8, 12.8, 1.6, H14ax), 0.92 (d, 3, J = 6.7, Me16); 13C NMR 210.0, 72.2, 58.7, 50.6, 50.5, 50.3, 50.2, 46.7, 37.22, 37.19, 25.8, 25.2, 22.4, 20.0; IR 3433 (broad), 1703 (strong), 1032; HRMS (ESI) calcd for C14H24NO2 (MH+) 238.1807, found 238.1803.</p><p>Lycopodine numbering is used so the methyl group is carbon 1 and there are no carbons 2 and 3. A 1D NOESY experiment with irradiation of H6ax at δ 2.57 showed a strong NOE to H6eq at δ 2.40 and weak NOEs to H4ax at δ 2.66 and H11ax at δ 1.28. A 1D NOESY experiment with irradiation of H4eq at δ 2.18 showed a strong NOE to H4ax at δ 2.66 and a weak NOE to H14eq at δ 2.09. A 1D NOESY experiment with irradiation of H14ax at δ 1.08 showed a strong NOE to H14eq at δ 2.09 and weak NOEs to H12eq at δ 1.61, Me16 at δ 0.92, Me1 at δ 2.24, and H8ax at δ 1.28.</p><!><p>A solution of propargyl ketone 2713 (98 mg, 0.72 mmol) in 2 mL of 3:2 conc H2SO4/H2O was stirred at room temperature for 16 h. The reaction mixture was quenched by addition of water (10 mL) at 0 °C and neutralized with 6 M NaOH until the pH reached 11. The mixture was saturated with NaCl and extracted with EtOAc (20 mL × 3). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (1:3 hexanes/EtOAc) gave 72 mg (76% from 27) of 28 as a white solid: mp 236-238 °C (lit.14a 233-240 °C; lit.14b 228-231 °C; lit.14c 220-223 °C); 1H NMR 2.60 (ddd, 1, J = 16.0, 2.0, 2.0), 2.50 (br s, 1, OH), 2.47 (d, 1, J = 16.0), 2.42 (dd, 1, J = 16.4, 6.0), 2.32 (dd, 1, J = 16.4, 1.8), 1.98 (dddd, 1, J = 12.0, 2.8, 2.6, 1.8), 1.87-1.76 (m, 3), 1.75-1.46 (m, 4), 1.34 (ddddd, 1, J = 13.6, 13.6, 13.6, 4.4, 4.4); 13C NMR 211.0, 71.1, 55.2, 45.6, 41.3, 40.6, 30.64, 30.58, 20.1; IR 3421 (broad), 1693, 1229, 1083, 991, 932, 732. The 1H NMR, 13C NMR, and IR spectral data are identical to those previously reported.14a</p><!><p>To 1 mL of CH2Cl2 was added EtAlCl2 (0.2 mL, 0.9 M in heptane, 0.18 mmol, 1.2 equiv) and EtOH (10.5 μL, 0.18 mmol, 1.2 equiv) successively. The mixture was stirred at room temperature for 1 h. A solution of 27 (20 mg, 0.15 mmol) in 1 mL of CH2Cl2 was added dropwise. The resulting mixture was stirred at room temperature for 20 h. The reaction was quenched by addition of 1 M HCl solution (5 mL) and extracted with CH2Cl2 (5 mL × 3). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (25:1 hexanes/EtOAc) gave 21 mg (71% from 27) of 31 as a colorless oil: 1H NMR 5.72 (s, 1), 3.55 (dq, 1, J = 9, 7), 3.45 (dq, 1, J = 9, 7), 2.60 (br dd, 1, J = 18.4, 6.8), 2.40-2.32 (m, 1), 2.02 (d, 1, J = 18.4), 1.96 (br d, 1, J = 12.0), 1.69-1.40 (m, 7), 1.19 (t, 3, J = 7.2); 13C NMR 133.7, 130.4, 76.2, 57.5, 39.0, 35.5, 35.1, 32.7, 30.4, 20.6, 16.2; IR 1648, 1103, 1087, 1052, 945, 820, 745, 661; HRMS (ESI) calcd for C11H18ClO (MH+) 201.1041, found 201.1039.</p><!><p>To a suspension of NaH (60% in mineral oil, 650 mg, 16 mmol) in 12 mL of 1:1 THF/DMF at 0 °C, a solution of crude 12 (1.50 g, from 9.5 mmol of 11) in 12 mL of 1:1 THF/DMF was added dropwise. The mixture was stirred at 0 °C for 10 min and 1-((dimethylethyl)dimethylsiloxy)-3-iodopropane19 (4.50 g, 15 mmol) was added dropwise. The reaction mixture was warmed to room temperature and stirred for 18 h. The reaction was quenched by addition of water (20 mL) at 0 °C. The mixture was extracted with EtOAc (50 mL × 3). The combined organic layers were washed with brine (150 mL × 5), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 1.60 g (50% from 11) of 34 as a pale yellow oil: 1H NMR 3.62 (t, 2, J = 5.1), 3.40 (dt, 1, J = 14.6, 7.0), 3.30 (dt, 1, J = 14.6, 7.0), 3.27-3.12 (m, 2), 2.59 (br d, 1, J = 12.8), 2.42-2.24 (m, 3), 2.12-1.96 (m, 3), 1.86-1.72 (m, 4), 1.03 (d, 3, J =6.1), 0.89 (s, 9), 0.05 (s, 6); 13C NMR 193.6, 159.3, 105.5, 59.6, 49.0, 47.6, 44.1, 34.5, 31.5, 28.9, 25.8 (3 C), 21.4, 21.2, 20.0, 18.1, ‐5.4 (2 C); IR 1613, 1551 (strong); HRMS (ESI) calcd for C19H36NO2Si (MH+) 338.2515, found 338.2519.</p><!><p>To a resealable tube was added 1.60 g (4.7 mmol) of 34 and 9 mL (15.3 g, 90 mmol, 19 equiv) of 2-iodopropane. The reaction mixture was sealed and heated at 55 °C for 42 h, diluted with 27 mL of benzene, and concentrated to remove excess 2-iodopropane to give cation 35.</p><p>To a solution of the residue (cation 35) in 24 mL of THF, 8.5 mL (1.8 equiv) of propargylmagnesium bromide solution (for the preparation of propargylmagnesium bromide see the preparation of 16) was added at 0 °C. The mixture was kept at 0 °C for 1 h. The reaction was quenched by addition of water (1 mL) at −78 °C and filtered through a pad of Celite. The filtrate was diluted with saturated ammonium chloride solution (30 mL) and extracted with EtOAc (40 mL × 3). The combined organic layers were washed with brine (100 mL), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (8:1 hexanes/EtOAc) gave 798 mg (40% from 34) of 36a as a pale yellow oil: 1H NMR 4.08 (heptet, 1, J = 6.1), 3.68-3.56 (m, 2), 3.11 (ddd, 1, J = 12.8, 7.9, 7.9), 3.02-2.93 (m, 1), 2.81 (br d, 1, J = 12.2), 2.68 (dd, 1, J = 17.1, 2.4), 2.47 (dd, 1, J = 17.1, 2.4), 2.42 (ddd, 1, J = 12.2, 12.2, 2.4), 2.34 (br d, 1, J = 12.5), 2.16-2.00 (m, 3), 1.97 (dd, 1, J = 2.4, 2.4), 1.82-1.72 (m, 1), 1.67-1.50 (m, 4), 1.48-1.37 (m, 1), 1.21 (dd, 1, J = 12.5, 12.5), 1.18 (d, 3, J = 6.1), 1.16 (d, 3, J =6.1), 1.00 (d, 3, J = 6.1), 0.90 (s, 9), 0.05 (s, 6); 13C NMR 145.1, 121.3, 83.1, 70.9, 68.6, 61.2, 59.8, 47.0, 46.4, 44.0, 34.7, 32.7, 26.0, 25.9 (3 C), 25.6, 22.8, 22.4, 22.3, 21.2, 20.4, 18.3, ‐5.26, ‐5.29; IR 3311, 2113 (weak), 1673 (weak), 1253, 1097 (strong); HRMS (ESI) calcd for C25H46NO2Si (MH+) 420.3298, found 420.3298.</p><!><p>To a resealable tube was added 1.60 g (4.7 mmol) of 34 and 9 mL (15.3 g, 90 mmol, 19 equiv) of 2-iodopropane. The reaction mixture was sealed and heated at 55 °C for 42 h, diluted with 27 mL of benzene, and concentrated to remove excess 2-iodopropane to give cation 35.</p><p>The residue (cation 35) was taken up in 18 mL of THF and the resulting solution was added slowly to a solution of (trimethylsilyl)propargyllithium (see below) at −78 °C. The mixture was kept at −78 °C for 2 h and slowly warmed to room temperature over 4 h. The reaction was quenched by addition of saturated ammonium chloride solution (20 mL) and water (10 mL). The mixture was extracted with EtOAc (50 mL × 3). The combined organic layers were washed with brine (100 mL), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (12:1 hexanes/EtOAc) gave 1.19 g (51% from 34) of 36b as a pale yellow oil: 1H NMR 4.07 (heptet, 1, J = 6.1), 3.68-3.56 (m, 2), 3.10 (ddd, 1, J = 13.4, 7.9, 7.9), 2.92-3.00 (m, 1), 2.83-2.76 (m, 1), 2.75 (d, 1, J = 17.4), 2.46 (d, 1, J = 17.4), 2.40 (ddd, 1, J = 12.4, 12.4, 2.4), 2.36 (br d, 1, J = 12.2), 2.16-2.02 (m, 3), 1.82-1.71 (m, 1), 1.68-1.50 (m, 4), 1.46-1.35 (m, 1), 1.17 (d, 3, J = 6.1), 1.16 (dd, 1, J = 12.2, 12.2), 1.16 (d, 3, J = 6.1), 0.99 (d, 3, J = 6.1), 0.90 (s, 9), 0.12 (s, 9), 0.04 (s, 6); 13C NMR 144.9, 121.6, 106.3, 87.3, 68.6, 61.2, 59.9, 47.0, 46.3, 43.8, 34.6, 32.7, 26.1, 26.0 (3 C), 25.6, 22.9, 22.4, 22.3, 21.7, 21.2, 18.3, ‐0.1 (3 C), ‐5.23, ‐5.25; IR 2954, 2856, 2171, 1676 (weak), 1250, 1099; HRMS (ESI) calcd for C28H54NO2Si2 (MH+) 492.3693, found 492.3692.</p><!><p>To 1-trimethylsilylpropyne (1.10 mL, 7.1 mmol, 1.5 equiv) in 7 mL of THF, n-BuLi (1.6 M, 4.70 mL, 1.6 equiv) was slowly added at −78 °C. The reaction was kept at −78 °C for 5 min, warmed to −30 °C over 15 min and kept at 0 °C for 1 h.</p><!><p>A solution of 798 mg (1.9 mmol) of 36a in 27 mL of 3:2 conc H2SO4/H2O was stirred at room temperature for 8 h. The reaction mixture was quenched by addition of water (30 mL) at 0 °C and neutralized with 6 M NaOH until the pH reached 11. The mixture was saturated with NaCl and extracted with CH2Cl2 (90 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 256 mg (48% from 36a) of 6 as a white solid.</p><p>A solution of 1.17 g (2.4 mmol) of 36b in 40 mL of 3:2 conc H2SO4/H2O was stirred at room temperature for 10 h. The reaction mixture was quenched by addition of water (40 mL) at 0 °C and neutralized with 6 M NaOH until the pH reached 11. The mixture was saturated with NaCl and extracted with CH2Cl2 (120 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 334 mg (50% from 36b) of 6 as a white solid: mp 157-158 °C; 1H NMR 5.92 (br, 1, w1/2 = 16 Hz, OH), 3.86-3.75 (m, 2, 2 H3), 3.16-3.04 (m, 2, H1, H9eq), 2.66 (d, 1, J = 17.1, H4ax), 2.56 (d, 1, J = 16.8, H6ax), 2.49 (br, 1, w1/2 = 11 Hz, OH), 2.39 (d, 1, J = 16.8, H6eq), 2.29-2.14 (m, 4, H1, H11eq, H9ax, H14eq), 2.10 (d, 1, J = 17.1, H4eq), 2.02-1.82 (m, 3, H2, H8eq, H10eq), 1.62-1.40 (m, 4, H2, H10ax, H12, H15), 1.35-1.22 (m, 2, H8ax, H11ax), 1.17 (dd, 1, J = 12.8, 12.8, H14ax), 0.95 (d, 3, J = 6.1, Me16); 13C NMR 209.6, 71.9, 64.3, 59.3, 51.0, 50.4, 49.9, 49.4, 46.5, 46.4, 38.2, 27.6, 25.8, 25.0, 22.4, 19.9; IR 3372 (broad), 1702 (strong), 1055, 1032; HRMS (ESI) calcd for C16H28NO3 (MH+) 282.2069, found 282.2067.</p><p>Lycopodine numbering is used so the hydroxypropyl group contains carbon 1-3. A 1D NOESY experiment with irradiation of H4ax at δ 2.66 showed a strong NOE to H4eq at δ 2.10 and weak NOEs to H9ax at δ 2.24 and H11ax at δ 1.28. A 1D NOESY experiment with irradiation of H6ax at δ 2.56 showed a strong NOE to H6eq at δ 2.39 and a weak NOE to H11ax at δ 1.28. A 1D NOESY experiment with irradiation of H6eq at δ 2.39 showed a strong NOE to H6ax at δ 2.56 and a weak NOE to H8eq at δ 1.87. A 1D NOESY experiment with irradiation of H14ax at δ 1.17 showed a strong NOE to H14eq at δ 2.19 and weak NOEs to H12eq at δ 1.58 and Me16 at δ 0.94.</p><!><p>To a solution of 100 mg (0.36 mmol) of 6 in 3 mL (2.93 g, 37 mmol, 103 equiv) of pyridine was added 1.2 mL (1.30 g, 12.7 mmol, 35 equiv) of acetic anhydride. The mixture was stirred at room temperature for 10 h. The reaction was quenched by addition of 20 mL of saturated NaHCO3 solution. The mixture was extracted with CH2Cl2 (20 mL × 3). The combined organic layers were dried over Na2SO4 and concentrated to about 5 mL. Toluene (6 mL) was added and the resulting solution was concentrated again to give 115 mg of crude (±)-(4aS,5R,7S,8aS)-1-(3-acetoxypropyl)-octahydro-5-hydroxy-7-methyl-1H-5,8a-propanoquinolin-10-one (39) as a yellow sticky oil, which was directly used in next step without further purification.</p><p>To a solution of 115 mg of the crude primary acetate 39 in 3 mL (3.30 g, 42.2 mmol) of DMSO was added 3 mL (3.24 g, 31.8 mmol) of acetic anhydride. The mixture was stirred at room temperature for 72 h. The reaction was quenched by addition of saturated NaHCO3 solution (30 mL) and extracted with extracted with CH2Cl2 (30 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Most of the remaining DMSO was removed by blowing air on the compound for 6 h. Flash chromatography of the residue on silica gel (1:1 hexanes/EtOAc) gave 136 mg (90% from 6) of 40 as a pale yellow oil.</p><p>Alternatively, a solution of 334 mg (1.2 mmol) of 6 in 8 mL (8.64 g, 84.7 mmol, 70 equiv) of acetic anhydride was stirred overnight under nitrogen. 8 mL (8.80 g, 113 mmol, 94 equiv) of DMSO was added and the reaction mixture was stirred for 72 h. The reaction was quenched by addition of saturated NaHCO3 solution (80 mL) and extracted with CH2Cl2 (100 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Most of the remaining DMSO was removed by blowing air on the compound for 6 h. Flash chromatography of the residue on silica gel (1:1 hexanes/EtOAc) gave 409 mg (90% from 6) of 40 as a pale yellow oil: 1H NMR 4.52 (s, 2), 4.19-4.06 (m, 2), 2.88 (ddd, 1, J = 13.4, 7.9, 7.9), 2.82 (br d, 1, J = 12.3), 2.75 (d, 1, J = 17.0), 2.57 (d, 1, J = 16.9), 2.54 (d, 1, J = 16.9), 2.26 (ddd, 1, J = 12.3, 12.3, 2.8), 2.19 (s, 3), 2.08 (d, 1, J = 17.0), 2.04 (s, 3), 2.05-1.86 (m, 4), 1.81-1.64 (m, 4), 1.56 (ddddd, 1, J = 12.8, 12.8, 12.8, 4.3, 4.3), 1.49-1.33 (m, 2), 1.24 (dddd, 1, J =13.1, 13.1, 13.1, 3.7), 0.96 (dd, 1, J = 12.2, 12.2), 0.92 (d, 3, J = 5.5); 13C NMR 210.0, 171.1, 77.7, 66.0, 62.4, 58.9, 48.3, 47.8, 47.1, 46.6, 45.0, 44.6, 38.8, 27.8, 25.4, 25.2, 22.7, 21.0, 20.3, 14.6; IR 1735 (strong), 1705 (strong), 1553, 1241 (strong), 1043 (strong); HRMS (ESI) calcd for C20H34NO4S (MH+) 384.2209, found 384.2209.</p><!><p>To a solution of 136 mg (0.36 mmol) of 40 in 6 mL of MeOH at 0 °C was added 400 mg (2.9 mmol, 8.0 equiv) of K2CO3. The reaction was kept at 0 °C for 2 h and filtered. The filtrate was concentrated and diluted with 10 mL of saturated NaHCO3 solution. The mixture was extracted with CH2Cl2 (10 mL × 4). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 111 mg (92%) of the primary alcohol 41 as a pale yellow sticky oil: 1H NMR 5.47 (br, 1, w1/2 = 12 Hz, OH), 4.47 (s, 2), 3.81-3.70 (m, 2), 3.11-2.96 (m, 2), 2.66 (d, 1, J = 17.1), 2.53 (s, 2), 2.23-2.10 (m, 3), 2.14 (s, 3), 2.06 (d, 1, J = 17.1), 2.02 (dd, 1, J = 12.0, 2.8), 1.97-1.89 (m, 2), 1.79 (ddd, 1, J = 10.4, 2.7, 2.7), 1.70 (dd, 1, J = 12.5, 3.2), 1.52 (ddddd, 1, J = 13.2, 13.2, 13.2, 4.0, 4.0), 1.44-1.31 (m, 3), 1.22 (dddd, 1, J = 12.8, 12.8, 12.8, 3.6), 1.12 (dd, 1, J = 11.6), 0.92, (d, 3, J = 5.5); 13C NMR 209.0, 77.5, 66.0, 64.3, 59.0, 49.2, 48.2, 48.1, 46.5, 46.1, 44.5, 38.4, 27.9, 25.5, 24.9, 22.5, 20.0, 14.5; IR 3395 (broad), 1703 (strong), 1041 (strong); HRMS (ESI) calcd for C18H32NO3S (MH+) 342.2103, found 342.2104.</p><!><p>To a resealable tube was added 100 mg (0.26 mmol) of 40, 180 mg (1.60 mmol, 6.0 equiv) of KO-t-Bu, 780 mg (4.3 mmol, 16 equiv) of benzophenone, and 3 mL of dry benzene. The mixture was subjected to three cycles of freeze-pump-thaw degas protocol and was sealed and heated at 110 °C for 1 h. After cooling to room temperature, the reaction was quenched by addition of 10 mL of 3 M HCl solution. The mixture was extracted with ether (10 mL × 2). The aqueous layer was neutralized with Na2CO3 powder until pH 11 was reached. The solution was extracted with CH2Cl2 (15 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 32 mg (47% from 40) of 42 as a white solid.</p><p>Alternatively, to a resealable tube was added 111 mg (0.32 mmol) of 41, 182 mg (1.63 mmol, 5.0 equiv) of KO-t-Bu, 948 mg (5.2 mmol, 16 equiv) of benzophenone, and 3 mL of dry benzene. The mixture was subjected to three cycles of freeze-pump-thaw degas protocol and was sealed and heated at 110 °C for 50 min. After cooling to room temperature, the reaction was quenched by addition of 12 mL of 3 M HCl solution. The mixture was extracted with ether (12 mL × 2). The aqueous layer was neutralized with Na2CO3 powder until pH 11 was reached. The solution was extracted with CH2Cl2 (20 mL × 5). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 42 mg (50% from 41) of 42 as a white solid: mp >172 °C (decomposition); 1H NMR 7.06 (dd, 1, J = 3.7, 3.7), 3.55 (ddd, 1, J = 14.7, 11.0, 6.1), 2.75 (dd, 1, J = 14.7, 7.3), 2.70-2.50 (m, 5), 2.14 (dd, 1, J = 12.8, 2.6), 2.10-1.99 (m, 2), 1.89 (br d, 1, J = 12.2), 1.79 (ddd, 1, J = 13.2, 2.8, 2.8), 1.65 (ddddd, 1, J = 12.8, 12.8, 12.8, 4.0, 4.0), 1.64 (dd, 1, J = 12.8, 3.5), 1.49-1.36 (m, 1), 1.35 (ddd, 1, J = 12.2, 12.2, 1.6), 1.22 (dd, 1, J = 12.0, 12.0), 1.02 (dddd, 1, J = 12.8, 12.8, 12.8, 3.7), 0.93 (d, 3, J = 6.1);13C NMR 197.5, 136.9, 136.3, 70.9, 58.4, 50.8, 50.2, 49.01, 49.00, 48.1, 43.9, 25.6, 25.1, 22.2, 21.04, 20.97; IR 3389 (broad), 1680 (strong), 1612 (strong), 1244, 1030; HRMS (ESI) calcd for C16H24NO2 (MH+) 262.1807, found 262.1805.</p><!><p>To a solution of 42 mg (0.16 mmol) of 42 in 3 mL of MeOH was added 5 mg of PtO2. The mixture was stirred under 1 atm of H2 (balloon) at room temperature for 10 h. The mixture was filtered through a pad of Celite and the filtrate was concentrated. Flash chromatography of the residue on silica gel (40:1:1 EtOAc/MeOH/NEt3) gave 40 mg (95%) of 5 as a white solid: mp 201-202 °C; 1H NMR (CDCl3) 3.37 (ddd, 1, J =14.3, 14.3, 3.7, H1ax), 3.12 (ddd, 1, J = 12.2, 12.2, 2.8, H9ax), 2.83 (br dd, 1, J = 11.6, 2.4, H4), 2.65 (br d, 1, J = 12.2, H9eq), 2.62 (br d, 1, J = 12.8, H14eq), 2.61 (br d, 1, J = 15.6, H6ax), (The peaks at δ 2.62 ppm and δ 2.61 ppm overlap and are assigned from analysis of the COSY spectra.), 2.56 (dd, 1, J = 14.3, 4.9, H1eq), 2.38 (dd, 1, J = 15.6, 1.5, H6eq), 2.11-2.01 (m, 2, H11eq, H3eq), 1.89 (ddd, 1, J = 12.0, 3.2, 3.2, H10eq), 1.85 (dd, 1, J = 12.0, 3.8, H8eq), 1.84 (ddddd, 1, J = 13.6, 13.6, 13.6, 4.8, 4.8, H2ax), 1.69 (ddddd, 1, J = 12.8, 12.8, 12.8, 3.8, 3.8, H10ax), 1.67 (br dd, 1, J = 12.8, 3.1, H12), 1.54 (dddd, 1, J = 13.1, 13.1, 13.1, 4.8, H3ax), 1.44 (dddd, 1, J = 12.8, 12.8, 12.8, 3.8, H11ax), 1.38 (br d, 1, J = 13.6, H2eq), 1.41-1.33 (m, 1, H15), 1.28 (ddd, 1, J = 12.0, 12.0, 2.2, H8ax), 0.91 (dd, 1, J = 12.8, 12.8, H14ax), 0.89 (d, 3, J = 6.1, Me16); (CD3OD) 3.35 (ddd, 1, J = 14.0, 14.0, 3.5, H1ax), 3.23 (ddd, 1, J = 12.4, 12.4, 2.3, H9ax), 3.03 (br dd, 1, J = 11.6, 2.8, H4), 2.69 (br d, 1, J = 15.6, H6ax), 2.63 (br d, 1, J = 12.4, H14eq), 2.58 (br d, 1, J = 12.4, H9eq), 2.52 (dd, 1, J = 14.0, 4.8, H1eq), 2.25 (dd, 1, J = 15.6, 1.2, H6eq), 2.06-1.98 (m, 2, H3eq, H11eq), 1.95-1.85 (m, 2, H2ax, H10eq), 1.79 (br d, 1, J = 12.4, H8eq), 1.73-1.46 (m, 4, H10ax, H12, H11ax, H3ax), 1.42 (br d, 1, J = 13.6, H2eq), 1.37-1.26 (m, 1, H15), 1.28 (ddd, 1, J = 12.4, 12.4, 1.7, H8ax), 0.91 (d, 3, J = 5.6, Me16), 0.90 (dd, 1, J = 12.6, 12.6, H14ax); (DCl salt in CD3OD) 3.82 (ddd, 1, J = 13.2, 13.2, 2.0), 3.73 (ddd, 1, J = 13.6, 13.6, 4.4), 3.30 (m, 1, obscured by the residual solvent peak, assigned from analysis of the COSY spectra), 3.18 (br d, 1, J = 13.2), 3.03 (dd, 1, J = 13.6, 4.8), 2.75 (br d, 1, J = 16.0), 2.71 (dd, 1, J = 12.4, 3.0), 2.40 (dd, 1, J = 16.0, 1.2), 2.20-2.10 (m, 3), 2.06-1.80 (m, 5), 1.81 (dddd, 1, J = 12.8, 12.8, 12.8, 3.2), 1.70-1.58 (m, 1), 1.47-1.33 (m, 2), 1.33 (dd, 1, J = 12.4, 12.4), 0.98 (d, 3, J = 5.6); 13C NMR (CDCl3) 210.3, 72.5, 59.5, 51.4, 50.9, 50.4, 47.3, 47.1, 42.5, 41.8, 25.4, 25.2, 22.6, 19.6, 19.5, 18.5; (CD3OD) 212.6, 73.1, 61.2, 52.6, 51.7, 51.1, 48.4, 48.0, 43.3, 43.1, 26.6, 26.3, 23.0, 20.6, 20.4, 19.7; (DCl salt in CD3OD) 206.6, 72.5, 66.2, 51.9, 51.0, 50.8, 44.1, 40.7, 26.6, 24.2, 22.6, 19.1, 18.9, 18.7 (carbons 1 and 9 are obscured by the residual solvent peak); IR 3458 (broad), 1702 (strong), 1312; HRMS (ESI) calcd for C16H26NO2 (MH+) 264.1964, found 264.1960.</p><p>The data for 5 in CD3OD are referenced to the residual solvent peaks at δ 3.31 and δ 49.15. The 1H NMR data for 5·DCl in CD3OD (also referenced to δ 3.31 for residual CD2HOD) are identical to those reported for 5·DCl except that all absorptions are 0.03 ppm downfield. The 13C NMR data for 5·DCl are referenced to δ 49.3 so that the data match those reported in the literature.4 The 1H and 13C NMR spectra of natural and synthetic 5·DCl are shown in Tables S1 and S2 in the Supporting Information.</p><!><p>To a solution of 4625 (30 mg, 0.082 mmol) in 2 mL of 1:10 conc HCl/MeOH was added 5 mg of PtO2. The mixture was shaken in a Parr apparatus at an initial pressure of 50 psi for 40 h. The mixture was filtered through a pad of Celite and the filtrate was concentrated. Flash chromatography of the residue on silica gel (2:1 hexanes/EtOAc) gave 21 mg (69%) of 47 as a colorless sticky oil: [α]D22 = −5.1 ° (c 0.25, CH2Cl2); 1H NMR 7.60-7.53 (br, 1), 7.34-7.25 (br, 1), 7.22-7.15 (br, 2), 7.12-7.05 (br, 1), 3.85 (d, 1, J = 7.9, H-1), 3.80 (dd, 1, J = 10.0, 2.5, H-3a), 3.63 (dd, 1, J = 9.2, 7.9, H-2), 3.25 (ddd, 1, J = 10.4, 10.4, 4.0, H-6), 2.33-2.22 (m, 2), 2.14 (dddd, 1, J = 11.6, 3.0, 2.9, 2.5), 1.92-1.80 (m, 2), 1.72-1.55 (m, 2), 1.52-1.24 (m, 6), 1.19-0.90 (m, 6), 0.82 (ddd, 1, J = 12.0, 12.0, 12.0), 0.72 (d, 3, J = 6.8), 0.70-0.62 (m, 2), 0.33 (ddd, 1, J = 12.0, 12.0, 12.0); 13C NMR 142.9, 130.2 (br), 128.6 (br), 127.6 (br), 126.9 (br), 126.6, 93.4, 86.3, 72.8, 65.5, 62.5, 49.5, 43.8, 39.8, 37.7, 29.5, 29.4, 28.9, 28.8, 26.4, 25.5, 25.4, 25.2, 21.8; IR 1453, 1096, 1029, 1009, 735, 702; HRMS (ESI) calcd for C24H36NO2 (MH+) 370.2746, found 370.2743.</p><p>A COSY experiment showed a cross peak between H1 at δ 3.85 and H2 at δ 3.63. A COSY experiment showed cross peaks between H3a at δ 3.80 and the hydrogens at δ 2.14, at δ 1.64. A COSY experiment showed cross peaks between H2 at δ 3.63 and H1 at δ 3.85, the hydrogen at δ 1.00. A COSY experiment showed cross peaks between H6 at δ 3.25 and the hydrogens at δ 1.82, at δ 1.11, and at δ 0.82.</p><!><p>To a solution of 4625 (30 mg, 0.082 mmol) in 2 mL of MeOH was added 5 mg of PtO2. The mixture was stirred under 1 atm of H2 (balloon) at room temperature for 40 h. The mixture was filtered through a pad of Celite and the filtrate was concentrated. Flash chromatography of the residue on silica gel (1:1 hexanes/EtOAc) gave 22 mg (73%) of 50 as a sticky colorless oil: [α]D22 = −55 ° (c 0.11, CH2Cl2); 1H NMR 7.43-7.19 (m, 10), 4.75 (br, 1, PhCH), 3.87 (dd, 1, J = 13.4, 10.8, PhCH2), 3.30 (dd, 1, J = 13.4, 5.6, PhCH2), 3.02 (ddd, 1, J = 10.4, 10.4, 4.4, H-5), 2.46 (ddd, 1, J = 18.0, 4.8, 1.4, H-3), 2.29 (ddd, 1, 18.0, 12.8, 6.0, H-3), 2.11 (br dd, 1, J = 10.8, 10.8), 1.99 (dddd, 1, J = 12.8, 6.0, 2.4, 2.4), 1.93 (br d, 1, 13.6), 1.88 (ddd, 1, J = 12.8, 4.0), 1.79-1.67 (m, 1), 1.33-1.16 (m, 1), 1.19 (dddd, 1, J = 12.4, 10.0, 10.0, 2.4), 0.92 (ddd, 1, J = 11.6, 11.6, 11.6), 0.87 (d, 3, J = 6.4), 0.83 (ddd, 1, J = 11.6, 11.6, 11.6), 0.72 (dddd, 1, J = 12.8, 12.8, 12.8, 5.2);13C NMR 171.1, 141.0 (br), 139.1, 129.4 (2 C), 128.3 (2 C), 128.1(2 C), 126.8 (2 C), 126.5, 126.4, 71.6, 63.4 (br), 62.7 (br), 48.2, 43.5, 40.0, 37.7, 33.9, 28.7, 21.8, 21.4; IR 1623, 1455, 1275, 1261, 750, 701; HRMS (ESI) calcd for C24H30NO2 (MH+) 364.2277, found 364.2275.</p><p>A COSY experiment showed cross peaks between the hydrogen at δ 4.75 (PhCH) and the hydrogens at δ 3.87 (PhCH2), at δ 3.30 (PhCH2). A COSY experiment showed cross peaks between H5 at δ 3.02 and the hydrogens at δ 1.88, at δ 1.19, and at δ 0.92. A COSY experiment showed a cross peak between H3 at δ 2.46 and H3 at δ 2.29</p><!><p>To a suspension of NaH (60% in mineral oil, 20 mg, 0.50 mmol, 2.3 equiv) in 1 mL of DMSO at room temperature under N2, a solution of 12 (36 mg, 0.22 mmol) in 2 mL of DMSO was added dropwise. The mixture was stirred at room temperature for 20 min and a solution of stilbene oxide (72 mg, 0.37 mmol, 1.7 equiv) in 1 mL of DMSO was added dropwise. The resulting mixture was heated to 50 °C and stirred for 12 h. The reaction was quenched by addition of water (5 mL) at room temperature. The mixture was extracted with EtOAc (10 mL × 3). The combined organic layers were washed with brine (20 mL × 4), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 4.3 mg (12% from 12) of 56 as a colorless oil followed by 6.7 mg (12% from 12) of 55 as a colorless sticky oil.</p><p>Data for 55: 1H NMR 7.38 (dd, 2, J = 7.3, 7.3), 7.31 (dd, 1, J = 7.3, 7.3), 7.16 (d, 2, J = 7.3), 4.58 (d, 1, J = 17.1), 4.43 (d, 1, J = 17.1), 3.26-3.21 (m, 2), 2.56-2.30 (m, 4), 2.20-1.98 (m, 3), 1.92-1.74 (m, 2), 1.01 (d, 3, J = 6.1); 13C NMR 194.0, 159.0, 137.1, 128.9 (2 C), 127.4, 125.9 (2 C), 106.0, 54.0, 49.5, 44.0, 34.8, 29.0, 21.4, 21.1, 19.8; IR 1674, 1607, 1550, 1275, 1263, 749, 699; HRMS (ESI) calcd for C17H22NO (MH+) 256.1696, found 256.1702.</p><p>Data for 56: 1H NMR 7.50-7.22 (m, 5), 7.25 (d, 1, J = 15.6), 6.90 (d, 1, J = 15.6), 2.71 (s, 3). These data are identical to those previously reported.27</p><!><p>To a suspension of NaH (60% in mineral oil, 65 mg, 1.6 mmol, 2.0 equiv) in 1 mL of 1:1 THF/DMF, a solution of 12 (131 mg, 0.80 mmol) in 1 mL of 1:1 THF/DMF was added dropwise at 0 °C. The mixture was stirred at 0 °C for 20 min and 0.15 mL (210 mg, 1.2 mmol, 1.5 equiv) of BnBr was added dropwise. The reaction was warmed to room temperature and stirred for 18 h. The reaction was quenched by addition of water (5 mL) at 0 °C. The mixture was extracted with EtOAc (5 mL × 3). The combined organic layers were washed with brine (20 mL), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 142 mg (70% from 12) of 55 as a colorless sticky oil with 1H NMR, 13C NMR and IR spectral data identical to those of 55 prepared from 12 and stilbene oxide.</p><!><p>To a solution of 13 (15 mg, 0.084 mmol) in 1 mL of benzene, Mn(OAc)3 (86 mg, 0.37 mmol, 4.4 equiv) was added. The resulting mixture was heated at 150 °C for 20 min in a microwave reactor. The reaction mixture was quenched by addition of saturated NaHSO3 solution (10 mL) and extracted with EtOAc (10 mL × 3). The combined organic layers were washed with brine, dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 4.7 mg (35%) of 58 as a colorless oil, followed by 5.3 mg (35% recovered) of 13 as a pale yellow oil, and then 2.6 mg (20%) of 12 as a white solid.</p><p>To a solution of 12 (10 mg, 0.061 mmol) in 1 mL of benzene, Mn(OAc)3 (60 mg, 0.26 mmol, 4.3 equiv) was added. The resulting mixture was heated at 90 °C for 20 min in a microwave reactor. The reaction mixture was quenched by addition of saturated NaHSO3 solution (10 mL) and extracted with EtOAc (10 mL × 3). The combined organic layers were washed with brine, dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (100:1:1 EtOAc/MeOH/NEt3) gave 6.5 mg (67%) of 58 as a colorless oil, followed by 1.5 mg (15%) of recovered 12 as a white solid.</p><p>The data for 58: 1H NMR 8.69 (dd, 1, J = 4.8, 1.9), 8.27 (dd, 1, J = 8.0, 1.9), 7.29 (dd, 1, J = 8.0, 4.8), 3.23 (dd, 1, J = 17.0, 3.6), 2.86 (dd, 1, J = 17.0, 10.4), 2.78 (dd, 1, J = 12.4, 1.9), 2.48-2.36 (m, 1), 2.37 (dd, 1, J = 12.4, 12.4), 1.20 (d, 3, J = 6.4); 13C NMR 198.0, 163.0, 153.6, 134.8, 127.6, 122.2, 46.6, 40.8, 29.2, 21.2; IR 1689, 1584, 1457, 1293, 912, 731. The 1H NMR spectral data are identical to those previously reported.30</p><!><p>To a solution of 13 (8.0 mg, 0.048 mmol) in 1.5 mL of THF was added a solution of KHMDS (0.5 M in toluene, 0.15 mL, 1.5 equiv) at −78 °C under N2. The mixture was stirred at the same temperature for 20 min and O2 was introduced to the reaction container from a balloon via a syringe. The reaction was slowly warmed to 0 °C over 50 min and stirred at 0 °C for 30 min. The reaction was quenched by addition of saturated Na2SO3 solution (5 mL). The mixture was extracted with EtOAc (5 mL × 3). The combined organic layers were washed with brine (20 mL), dried over Na2SO4, and concentrated. Flash chromatography of the residue on silica gel (5:1 hexanes/EtOAc) gave 1.6 mg (20%) of 59 as a white solid: mp 95-96 °C; 1H NMR 6.08 (s, 1), 6.01 (s, 1), 4.53 (br s, 1, OH), 3.16 (t, 2, J = 5.6), 2.87 (s, 3), 2.63 (t, 2, J = 6.8), 2.22 (s, 3), 1.98 (tt, 2, J = 6.8, 5.6); 13C NMR 153.2, 147.9, 136.9, 105.8, 105.0, 104.6, 50.9, 39.8, 21.9, 21.6, 20.6; IR 1618, 1580, 1517, 1273, 1124, 915, 807, 748; HRMS (ESI) calcd for C11H16NO (MH+) 178.1232, found 178.1232. Similar results were obtained using NaHDMS.</p><!><p>To a solution of 13 (180 mg, 1.0 mmol) in 8 mL of THF at -78 °C, a solution of NaHMDS (1.0 M in THF, 2.2 mL, 2.2 equiv) was added and the mixture was stirred at the same temperature for 50 min. Oxygen (dried over CaSO4) was bubbled into the solution for 50 min at the same temperature. To the solution was added TBSOTf (0.60 mL, 2.6 mmol, 2.6 equiv) and the resulting mixture was stirred at the same temperature for 50 min. The reaction was quenched by addition of saturated NaHCO3 solution (30 mL) and extracted with Et2O (30 mL × 3). The combined organic layers were dried over MgSO4 and concentrated. Flash chromatography of the residue on MeOH-deactivated silica gel (50:1 hexanes/EtOAc) gave 30 mg (10%) of 60 as a colorless liquid, 10 mg of a mixture of 60, 61 and 62 (1:4:10) as a pale yellow liquid, and 98 mg (22%) of 61 as a pale yellow liquid.</p><p>The data for 60: 1H NMR 6.10 (s, 1), 6.01 (s, 1), 3.13 (t, 2, J = 5.6), 2.86 (s, 3), 2.63 (t, 2, J = 6.8), 2.22 (s, 3), 1.93 (tt, 2, J = 6.8, 5.6), 0.99 (s, 9), 0.21 (s, 6); 13C NMR 153.2, 147.9, 136.1, 110.7, 108.1, 105.4, 51.1, 39.7, 25.8 (3 C), 22.2, 21.8, 21.7, 18.3, -4.1 (2 C); IR 1606, 1576, 1269, 1191, 1135, 836, 778.</p><p>The structure of 60 was confirmed by hydrolysis of 60 (20 mg, 0.068 mmol) in 2 mL of 3 M HCl at room temperature for 2 h. The reaction was quenched by addition of saturated NaHCO3 solution (20 mL) and extracted with EtOAc (20 mL × 3). The combined organic layers were dried over Na2SO4 and concentrated. Flash chromatography of the residue on silica gel (5:1 hexanes/EtOAc) gave 11 mg (90%) of 59 as a white solid with mp, 1H NMR, 13C NMR and IR spectral data identical to those of 59 prepared from 13, KHMDS and oxygen.</p><p>The data for 61: 1H NMR 5.01 (br d, 1, J = 4.4), 4.72 (br d, 1, J = 4.4), 2.92 (br dd, 1, J = 10.8, 5.2), 2.86-2.77 (m, 1), 2.43 (ddd, 1, J = 12.4, 10.8, 3.2), 2.43 (s, 3), 2.32-2.24 (m, 1), 2.12-1.96 (m, 1), 1.52-1.42 (m, 2), 1.09 (d, 3, J = 7.2), 0.94 (s, 9), 0.87 (s, 9), 0.19 (s, 3), 0.18 (s, 3), 0.10 (s, 6); 13C NMR 146.8, 142.5, 112.1, 108.5, 78.4, 53.9, 40.5, 30.9, 30.8, 26.2 (3 C), 25.8 (3 C), 22.4, 20.4, 18.20, 18.18, -4.4, -5.1, -5.5, -5.8; IR 1686, 1647, 1178, 913, 835, 745; HRMS (ESI) calcd for C23H46NO3Si2 (MH+) 440.3016, found 440.3021.</p><p>A COSY experiment showed large cross peaks between the hydrogen at δ 2.86-2.77 and the hydrogens at δ 5.01, δ 4.72, and δ 1.09. A COSY experiment showed a small cross peak for W coupling between the hydrogens at δ 5.01 and at δ 4.72.</p><p>The data for 62 were obtained by comparing the NMR spectra of 60, 61, and the mixture of all three: 1H NMR 4.97 (dd, 1, J = 2, 2), 4.65 (dd, 1, J =2, 2), 2.97-2.88 (m, 2), 2.50-2.42 (m, 1), 2.45 (s, 3), 2.32-2.24 (m, 1), 2.12-1.96 (m, 1), 1.52-1.40 (m, 2), 1.02 (d, 3, J = 7.2), 0.94 (s, 9), 0.84 (s, 9), 0.20 (s, 3), 0.18 (s, 3), 0.10 (s, 6); 13C NMR 147.0, 142.3, 113.4, 109.5, 78.1, 53.8, 40.5, 31.0, 30.8, 26.1(3 C), 25.8 (3 C), 23.5, 20.4, 18.22, 18.20, −4.4, −4.9, −5.4, −5.8.</p><p>A COSY experiment showed large cross peaks between the hydrogen at δ 2.97-2.88 and the hydrogens at δ 4.97, δ 4.65, and δ 1.02. A COSY experiment showed a small cross peak for W coupling between the hydrogens at δ 4.97 and at δ 4.65.</p><!><p>To a solution of 61 (22 mg, 0.050 mmol) in 1 mL of CDCl3, pyridine (10 μL, 0.012 mmol, 2.5 equiv) and pyr·(HF)x (4 μL, 0.15 mmol, 3 equiv) were added successively. The reaction was stirred at RT for 30 min. The solvent was evaporated and flash chromatography on silica gel (3:1 hexanes/EtOAc) gave 6.5 mg (40%) of 65 as a colorless oil: 1H NMR 4.66 (d, 1, J = 10.0), 4.41 (ddd, 1, J = 13.6, 10.4, 3.6), 3.05-2.94 (m, 1), 2.82 (ddd, 1, J = 15.2, 12.0, 2.0), 2.80 (s, 3), 2.69 (dd, 1, J = 13.2, 4.4), 2.43 (ddd, 1, J = 13.6, 4.0, 4.0), 2.23-2.11 (m, 1), 2.08 (ddd, 1, J = 15.2, 8.0, 1.9), 1.94 (dd, 1, J = 13.2, 13.2), 1.94-1.84 (m, 1), 1.09 (d, 3, J = 6.8), 0.90 (s, 9), 0.21 (s, 3), 0.14 (s, 3); 13C NMR 203.2, 172.3, 150.5, 115.1, 47.0, 45.0, 37.8, 36.0, 26.2, 25.5 (3 C), 23.6, 21.0, 18.0, −4.7, −4.8; IR 1696, 1636, 1258, 1225, 1084, 1067, 913, 837, 745; HRMS (ESI) calcd for C17H32NO3Si (MH+) 326.2151, found 326.2150. The fully assigned spectral data as determined by analysis of COSY, HSQC, and HMBC 2D NMR spectra are shown in Table S3 in the Supporting Information.</p><!><p>Inhibition of acetylcholinesterase activity was determined with a modified micro-Ellman assay.35a,b 7-Hydroxylycopodine was initially dissolved in 0.12 M HCl to a concentration of 0.076 M and then diluted in saline (0.15 M sodium chloride) as needed. Electric eel (Electrophorus electricus) acetylcholinesterase (AChE) (15 U/mL) was pre-incubated with 7-hydroxylycopodine for 15 min and then 0.075 U of pre-incubated enzyme was added to the reaction mixture containing 7-hydroxylycopodine at the pre-incubation concentration, 1 mM acetylthiocholine (ATC), 0.1 mM 4,4′-dithiopyridine, and 50 mM pH 8.0 sodium phosphate buffer. The reaction was monitored for 5 min at 325 nm in a SpectraMax Plus 96 well plate reader at 24 °C. Human whole blood was diluted initially 1 to 10 in saline and then 1 to 2 with 7-hydroxylycopodine solution of the appropriate concentration. To distinguish ATC hydrolysis by human blood AChE and butyrylcholinesterase (BChE), blood samples were first treated with 3.33 mM tetramonoisopropyl pyrophosphortetramide (Iso-OMPA) prior to 7-hydroxylycopodine. BChE activities were determined with 1.0 mM butyrylthiocholine (BTC) as the substrate. IC50 values were calculated from non-linear regression analysis of the plotted data using GraphPad Prism Ver. 4.0.</p>

PubMed Author Manuscript

A Humanin Analog Decreases Oxidative Stress and Preserves Mitochondrial Integrity in Cardiac Myoblasts

A potent analog (HNG) of the endogenous peptide humanin protects against myocardial ischemia-reperfusion (MI-R) injury in vivodecreasing infarct size and improving cardiac function. Since oxidative stress contributes to the damage from MI-R we tested the hypotheses that: 1. HNG offers cardioprotection through activation of antioxidant defense mechanisms leading to preservation of mitochondrial structure and that, 2. the activity of either of a pair of non-receptor tyrosine kinases, c-Abl and Arg is required for this protection. Rat cardiac myoblasts (H9C2 cells) were exposed to nanomolar concentrations of HNG and to hydrogen peroxide (H2O2). Cells treated with HNG in the presence of H2O2 demonstrated reduced intracellular reactive oxygen species (ROS), preserved mitochondrial membrane potential, ATP levels and mitochondrial structure. HNG induced activation of catalase and glutathione peroxidase (GPx) within 5 minutes and decreased the ratio of oxidized to reduced glutathione within 30 minutes. siRNA knockdown of both Abl and Arg, but neither alone, abolished the HNG-mediated reduction of ROS in myoblasts exposed to H2O2. These findings demonstrate an HNG-mediated, Abl- and Arg-dependent, rapid and sustained activation of critical cellular defense systems and attenuation of oxidative stress, providing mechanistic insights into the observed HNG-mediated cardioprotection in vivo.

a_humanin_analog_decreases_oxidative_stress_and_preserves_mitochondrial_integrity_in_cardiac_myoblas

1. INTRODUCTION<!>2.1. Cell culture<!>2.2. ROS determination<!>2.3. Measurement of Mitochondrial transmembrane potential (\xce\x94\xcf\x88m)<!>2.4. Assessment of mitochondrial function via determination of intracellular ATP<!>2.5. Immunofluorescence<!>2.6. Assays for non-enzymatic and enzymatic antioxidant activity<!>2.7. Abl and Arg knockdowns<!>2.8. qRT-PCR<!>2.9. Immunoblots<!>2.10. Statistics<!>3.1. HNG maintains low intracellular ROS levels<!>3.2. HNG preserves mitochondrial membrane potential (\xce\x94\xcf\x88m)<!>3.3. HNG increases ATP production<!>3.4. HNG preserves mitochondrial structural integrity<!>3.5. HNG increases cellular antioxidant activities<!>3.6. A double knockdown of both Abl and Arg, but neither knocked down alone, eliminated the HNG-mediated reduction in oxidative stress<!>DISCUSSION<!>

<p>Coronary heart disease is the major cause of heart disease resulting in about 1 out of every 5 deaths [1]. Besides immediate risks, around 20% of those experiencing myocardial infarct will also develop heart failure [2]. Though reperfusion is critical for cardiac myocyte survival, minimizing the ventricular wall stress that favors remodeling, enlargement and heart failure [3], most cell death and long-term damage to cardiac function also occur at this stage.</p><p>Influx of oxygen and recovery of mitochondrial respiration during the reperfusion phase increase reactive oxygen species (ROS), levels of which are known to increase following many pathologic situations including MI-R [4, 5, 6]. Calcium uptake into the mitochondria via the calcium uniporter and loss of mitochondrial membrane potential (MMP) were recently shown to occur during reperfusion [7]. Increased ROS overwhelms normal antioxidant defenses and favors the opening of the mitochondrial permeability transition pore, leading to further membrane depolarization, ATP hydrolysis and mitochondrial swelling leading to rupture of the outer membrane and release of pro-death proteins into the cytosol [8]. Since ROS are important eukaryote cell signals, the generation of hydrogen peroxide is tightly regulated as are localization, expression and activation of antioxidant enzymes [9, 10]. Major players in cell defense against excessive ROS include non-enzymatic small molecule antioxidants such as glutathione and enzymes that include superoxide dismutase (SOD), catalase and glutathione peroxidase (GPx).</p><p>Our lab has been investigating a highly potent analog of the endogenous, 24-aa peptide Humanin (HN), HNG (HN in which the serine 14 is replaced by glycine). Originally isolated from a protected lobe of a brain from an Alzheimer's disease patient, HN has been shown to be neuroprotective and also protects other cell types from a variety of insults. A role for HN in cardiovascular diseases has been shown by us and others [11, 12, 13, 14, 15]. Circulating humanin levels in humans were recently found to be associated with impaired microvascular coronary endothelial function [16]. HN is highly expressed in unstable carotid plaques in atherosclerotic patients, with higher levels found in symptomatic patients relative to the asymptomatic group [17]. We previously showed that HNG administration protected against MI-R injury in a mouse model, decreasing infarct size and improving LV function [18].</p><p>The translational promise shown by HN in cardiovascular disorders underscores the need for a more mechanistic understanding of its action in the heart. The involvement of oxidative stress in the injury produced by MI-R led us to hypothesize that the mechanism of the cardioprotection provided by HNG may involve mitigation of oxidative stress and activation of antioxidant defense mechanisms.</p><!><p>Low passage (< 15) H9C2 cells (ATCC, CRL-1446) were grown at 37°C in 10% CO2 in DMEM with D-glucose (1000 mg/L), L-glutamine (584 mg/L), sodium pyruvate (1 mM) and 10% FBS (Biowest). For ROS determination and JC-10 staining, the same growth medium was used without phenol red.</p><!><p>H9C2 cells were grown to 70% confluence in black, 96-well plates. Four of the eight wells in each column were loaded with a 10µM solution of a thiol reactive, non-fluorescent chloromethyl derivative of DCF (Molecular Probes, C6827), the other four with only PBS as controls for auto-fluorescence. After loading, the plate was reacted for 40 minutes at 37°C in growth medium containing control (saline) or HNG with or without H2O2 (100µM). Fluorescence measurements were obtained at 485 excitation/535 emission using a Molecular Devices SpectraMax M5e scanning spectrophotometer.</p><!><p>HNG effects on MMP changes were assessed using the fluorochrome dye JC-10 (Enzo Life Sciences). H9C2 cells seeded in black, 96-well, at 70% confluence were treated with growth media containing either the saline or HNG (10nM) and H2O2 (40 µM) for 30 minutes at 37°C. This media was aspirated and replaced with 1X Hanks buffered salt solution containing 20 mM Hepes and 20µM JC-10 and the plate was incubated for an additional 30 minutes at 37°C. After 3 rinses with wash buffer (1X HBSS, 10 mM Hepes and 0.02mg/100ml D-glucose), the fluorescent signal was read on a Molecular Devices SpectraMax M5e scanning spectrophotometer using 490 and 525 nm as green excitation and emission wavelengths, respectively (cutoff 515 nm) and 490 and 590 nm as red excitation and emission wavelengths, respectively (cutoff 570 nm). The ratio of the reading at 590 nm to that at 525 nm was considered as the relative Δψm value.</p><!><p>H9C2 cells were treated for the indicated time with saline or HNG (10 nM). ATP was extracted as previously described [19]. Briefly, cells were dispersed into 85 mM sodium citrate, extracted with a final concentration of 2.3% TCA, neutralized with Tris-acetate EDTA buffer (0.1M Tris, pH 7.75 with acetic acid, 2mM EDTA) and boiled for 3 minutes. ATP levels were measured with a Bioluminescent Assay Kit (Sigma-Aldrich, St. Louis, MO, USA, FL-AA) and readings taken by a BMG LABTECH FLUOstar OPTIMA multimode microplate reader (Ortenberg, Germany).</p><!><p>H9C2 cells growing on polylysine-coated coverslips were pretreated for 10 minutes with saline or HNG and then exposed to 100 µM H2O2 for 3 hours in the presence or absence of HNG. Cells were loaded with Mitotracker Red CMXROS (Molecular Probes, M-7512), fixed with 3% formaldehyde and 0.02% glutaraldehyde for 30 minutes at room temperature, permeablized with 0.1% Triton X-100 for 5 minutes and labeled with anticytochrome-c followed by an Alexa Fluor-488, green fluorescent secondary antibody (Jackson ImmunoRearch Laboratories, Inc.) and dapi prior to fluorescence photography. Photographs were prepared with a Zeiss Axioskop II microscope with Zeiss Axiovision software and fitted with fluorescence filters for dapi, FITC and Rhodamine, at 63× magnification.</p><!><p>The level of the oxidized disulfide dimer (GSSG) and the reduced form of glutathione (GSH) was determined in lysates from H9C2 cells treated for various time points with HNG (10 nM). These values were obtained by carrying out a series of reactions as described by the Glutathione assay kit (Cayman Chemical Company, catalog # 703002). Assays for determination of enzyme activities—total superoxide dismutase (SOD), catalase and glutathione peroxidase (GPx) (BioVision Research Products, catalog #s K335, K773 and K762, respectively) were performed, as instructed, on H9C2 cell lysates following treatment with either saline or HNG (10nM) at the time points indicated. These values were normalized to total protein in the lysates.</p><!><p>Synthetic siRNAs for Abl and for control transfections were obtained from Qiagen. The sequence used to knockdown Abl (RN_Abl1_4) was catalog number SI01484308. The siRNA sequence used to knockdown Arg was synthesized and purified by Integrated DNA Technologies, Inc. and the sequence was as follows: sense: 5'GGAAAUCAAGCAUCCUAAUUUAG3', antisense: 5"UACUAAAUUAGGAUGCUUGAUUUCCUU3'. BLASTn searches were carried out to detect possible off targets for both siRNAs. AllStars_3 (catalog number SI04939025) siRNA was used as the positive control and AllStars Negative Control siRNA (catalog number 1027280) as the negative control in this optimization procedure. Transfections with siRNAs were performed in 6-well plates with 1.25×105 cells per well in 2.3 ml of growth medium for 72h before analyzing the degree of target knockdown. For immunoblots and Abl/Arg knockdowns prior to seeding in 96-well black plates for ROS determinations, these conditions were scaled up to 100 mm dishes as described in the HiPerFect directions.</p><!><p>Total RNA was extracted from H9C2 cell cultures using RNAseasy Mini Kits µg of total RNA was reverse transcribed into cDNA using the Superscript III First-Strand Sysnthesis System (Invitrogen). The cDNA was then subjected to real-time PCR amplification using Light Cycler 480 SYBR Green I Master kit in a LightCycler® 480 Real –Time PCR System (Roche). The forward and reverse primers for each gene were as follows: Abl-cttgcaggaaaaccacctct and tctgggacagtttgtgagca; Arg-aaggcaaggagaggaatggt and ctggcactttgtggttgtg. The data were then normalized to the housekeeping gene RPL19-gaagaggaagggtactgccaac and tttttgaacacattccctttga.</p><!><p>Cell lysates were prepared in RIPA buffer, resolved on SDS-PAGE gradient gels, transferred to PVDF membranes and probed with relevant antibodies (Abl: Abcam Inc, #10528, Arg: Epitomics, #5431-1, β-actin: Cell Signaling, #4967). Signals for Abl and Arg were normalized to the signal for the loading control, β-actin.</p><!><p>All values result from a minimum of 3 experiments and are presented as means ± se. When appropriate, data were evaluated by the 2-sample Student's t test. A value of p < 0.05 was considered significant. To take into consideration false-positive associations resulting from multiple comparisons, the stringent Bonferroni correction was used; the value of p for inclusion was set to 0.05/5 = 0.01.</p><!><p>Chloromethyl-DCF-loaded cells challenged with H2O2 showed an 86% increase in intracellular ROS compared with controls (1.7228 ± 0.1469 vs.0.9287± 0.1113). There was a 48% decrease in ROS, in HNG-treated cells relative to controls when both were challenged with 100 µM H2O2 (0.8998 ± 0.1650 vs.1.7228 ± 0.1469) (Fig. 1A). Levels of cellular ROS with HNG are comparable to the baseline (0.9287 ± 0.1113) and following treatment with ROS scavenger N-acetyl cysteine (NAC)(0.6687 ± 0.2090).</p><!><p>HNG produced a significant 11% increase in baseline Δψm (0.5226 ± 0.009 vs 0.5817 ± 0.007) (Fig. 1B). H2O2 (40 µM) decreased the membrane potential by 6% relative to the vehicle control alone (0.4901 ± 0.006 vs 0.5226 ± 0.009). In the presence of HNG (10 nM), the H2O2-induced decrease in the MMP was significantly attenuated, remaining at levels seen in the control group in the absence of H2O2 (0.5226 ± 0.009 vs 0.5128 ± 0.014).</p><!><p>There was a 92% increase in ATP levels at 30 minutes (95.41 ± 7.52 vs. 182.81 ± 18.60 pM/cell) and a 134% increase in 120 minutes (70.85 ± 5.58 vs. 165.89 ± 25.75 pM/cell) (Fig. 1C) in cells treated with HNG.</p><!><p>Immunofluorescence was used to visualize the mitochondria and a major mitochondrial component, cytochrome-c. An overlay photograph of saline control cells (Fig. 2A) displays a gold-colored, tubular network of intact, fused mitochondria, produced by the overlay of red (mitotracker) and green (cytochrome-c localized to mitochondria) fluorescence signal. In contrast, the cells pretreated with saline for 10 minutes and then exposed to H2O2 (100 µM) for 3 hours contain fragmented, punctate, red mitochondria with the green cytochrome-c dispersed throughout the cytosol (Fig. 2B). When the cells were preincubated with 10 nM HNG for 10 minutes prior to addition of the H2O2despite some punctate mitochondria and free cytochrome-c, much of the intact mitochondrial network was observed (Fig. 2C).</p><!><p>Levels of reduced glutathione (GSH) and the oxidized form (GSSG) were determined at various time points in the presence of either the control or HNG. The ratio of oxidized to reduced glutathione, a measure of oxidative stress, was decreased by 47% within 30 minutes of HNG treatment (0.2130 ± 0.0114 vs. 0.1131 ± 0.0179) (Fig. 3A).</p><p>The reduced form of glutathione is required to activate the critical antioxidant enzyme, glutathione peroxidase (GPx). This enzyme activity showed a 635% increase relative to control within the first 5 minutes of HNG treatment (98.26 ± 12.69 vs. 722.04 ± 253.09 mU/mg protein) (Fig. 3B). The activation decreased but remained significant throughout the 120 minutes investigated. Catalase, the other major antioxidant enzyme responsible for the removal of H2O2 was also activated within 5 minutes of HNG addition, producing a 213% increase in activity relative to controls (5.04 ± 0.46 vs. 15.78 ± 3.22 mU/mg protein) and was maintained at 30 minutes (Fig. 3C). Assay of total superoxide dismutase (SOD) activity demonstrated a 71% HNG-mediated enzyme activation by 120 minutes of treatment (1.63 ± 0.1483 vs. 2.78 ± 0.458 U/mg protein) (Fig. 3D).</p><!><p>A pair of non-receptor tyrosine kinases, c-Abl and Arg have been demonstrated to participate in the post-translational activation of both catalase and GPx [20, 21]. We carried out transient, siRNA, Abl and Arg knockdowns in H9C2 cells and then assessed the ability of HNG to maintain the low intracellular ROS levels observed in HNG-treated WT cells. Results from siRNA knockdowns of c-Abl and Arg were verified at 72h post-transfection with qRT-PCR (Figure 4A). siRNA knockdown of Abl alone produced an 78% decrease in Abl mRNA (0.998 ± 0.001 vs. 0.218 ± 0.046). A double transfection with siRNAs for both Abl and Arg resulted in an 81% reduction in the Abl mRNA level (0.998 ± 0.001 vs. 0.191± 0.046). Arg knockdown produced a 79% loss of Arg mRNA relative to negative control knockdowns (0.998 ± 0.001 vs. 0.210 ± 0.024). The double knockdown decreased the Arg message level by 76% (.998 ± 0.001 vs. 0.242 ± 0.025). Immunoblots (Fig. 4B) provide evidence that the Abl and Arg knockdowns were also successful at the level of protein expression. We observed that negative control knockdown cells reiterate the data seen in Fig. 1A with WT cells, with H2O2 producing an increase in relative ROS in cells treated with the vehicle control (1.223 ± 0.335 vs. 2.347 ± 0.246). This increase was prevented by co-treatment with 10 nM HNG, with the DCF fluorescence value unchanged relative to the control and 94% lower than the control + H2O2 value (1.209 ± 0.254). This pattern was also observed following knockdown of either Abl or Arg, as loss of either of the kinases alone had no effect on the ability of HNG to maintain a low, control intracellular ROS level after addition of 100 µM H2O2 to the media. When Abl knockdowns were treated with H2O2HNG produced a 35% decrease in ROS relative to saline (2.442 ± 0.240 vs. 1.582 ± 0.241, for saline vs. HNG). For Arg knockdowns exposed to H2O2HNG lowered the cellular ROS by 60% when compared to the effect of saline (2.252 ± 0.300 vs. 0.901 ± 0.339, for saline vs. HNG). In contrast, the cells with both Abl and Arg mRNA levels knocked down were no longer protected from oxidative stress by treatment with HNG and the relative ROS resulting from the H2O2 challenge matched that seen in the vehicle control (1.662 ± 0.366 vs 1.635 ± 0.325, for saline vs. HNG).</p><!><p>These findings demonstrate that HNG mitigates oxidative stress in a rat myoblast cell-line through rapid and sustained activation of several antioxidant enzymes that require the activity of one of a pair of non-receptor tyrosine kinases. The attenuation of oxidative stress results in preservation of mitochondrial membrane integrity and function.</p><p>Our findings are of significant clinical relevance as cardiac risk factors and resulting diseases that contribute to significant mortality and morbidity such as MI, cardiac failure, diabetes, atherosclerosis are all associated with ROS-induced cardiac damage. Free radical oxidants are also believed to contribute to the production of mitochondrial dysfunction and damage produced by the normal aging process [22]. Tissues such as cardiac muscle, with a high energy expenditure and abundant mitochondria, are particularly subject to mitochondrial oxidative damage, additional generation of ROS and further damage to critical molecular structures. ROS also stimulate the production of inflammatory cytokines from cardiac fibroblasts favoring a rise in collagen deposition and impaired cardiac function [23]. The ability of HNG to decrease cellular ROS and maintain mitochondrial integrity/function despite oxidative stress could explain the enhanced myocyte survival and the maintenance of cardiac function observed in vivo in MI-R models [18].</p><p>Both non-enzymatic and enzymatic antioxidant mechanisms appear to contribute to the observed oxidative stress reduction by HNG. The ratio of oxidized to reduced glutathione, a measure of oxidative stress, decreases within 30 minutes and the additional stores of reduced glutathione allows GPx to continue removing H2O2 from cells for the two hours observed in this study. Since catalase accounts for almost 80% of cardiomyocyte peroxidase activity [24], activation of both catalase and GPx within 5 minutes of HNG treatment and their sustained increases in activity can account for the rapid and persistent protection of the heart when HNG is provided just prior to reperfusion in the in vivo model of ischemia-reperfusion. The observed activation by HNG of total SOD in cardiac myoblasts supports recent similar observations in cortical neurons [25].</p><p>Since the rapid activation of catalase and GPx by HNG precludes de novo synthesis of these large enzymes, we focused on mechanisms that would produce their rapid, allosteric activation. The mammalian non-receptor, tyrosine kinase isoforms, Abl and Arg are constitutively repressed in the cytosol, requiring displacement of N-terminal and Src-homology regions from their kinase regions [26]. This displacement may be produced with physical interference from many different stimuli, allowing the enzyme to activate itself via autophosphorylation. Our results are consistent with the observations by Cao et al [20, 21] and show that either one of these isoforms is sufficient to retain the biologic effects of HNG in mitigating oxidative stress. Cells with an Abl knockdown but WT Arg levels could still use HNG to prevent an increase in ROS upon addition of H2O2. Despite a higher target protein expression in the Arg knockdowns, the phenotypic pattern seen in the Abl knockdowns was repeated. Both single knockdowns demonstrated a generalized increase in intracellular ROS levels at baseline relative to the negative knockdowns, while retaining the same pattern of protection from an additional oxidative challenge in the presence of HNG. In the presence of simultaneous knockdown of Abl and only a partial loss of Arg protein expression, HNG could no longer prevent the rise in intracellular ROS upon challenge with H2O2. These studies provide evidence that HNG prevents oxidative stress via a mechanism requiring a critical level of activity from either Abl or Arg. The question of whether HNG itself physically interferes with the inactive conformations of Abl and Arg or whether it activates another required kinase suggests future areas of investigation.</p><p>Therapies that mitigate oxidative stress prevent the onset or progression of cardiac dysfunction through decreased apoptosis and increased cardiomyocyte survival. However, antioxidants alone have not proven to be clinically effective. Through its unique roles in substrate metaboilsm, decreased apoptosis and decreased ROS, HNG may offer translational promise as a cardioprotective factor in the clinical setting.</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

A Metabolic Study on Colon Cancer Using 1H Nuclear Magnetic Resonance Spectroscopy

Background. Colorectal carcinoma is the third cause of cancer deaths in the world. For diagnosis, invasive methods like colonoscopy and sigmoidoscopy are used, and noninvasive screening tests are not very accurate. We decided to study the potential of 1HNMR spectroscopy with metabolomics and chemometrics as a preliminary noninvasive test. We obtained a distinguishing pattern of metabolites and metabolic pathways between colon cancer patient and normal. Methods. Sera were obtained from confirmed colon cancer patients and the same number of healthy controls. Samples were sent for 1HNMR spectroscopy and analysis was carried out Chenomex and MATLAB software. Metabolites were identified using Human Metabolic Data Base (HDMB) and the main metabolic cycles were identified using Metaboanalyst software. Results. 15 metabolites were identified such as pyridoxine, orotidine, and taurocholic acid. Main metabolic cycles involved were the bile acid biosynthesis, vitamin B6 metabolism, methane metabolism, and glutathione metabolism. Discussion. The main detected metabolic cycles were also reported earlier in different cancers. Our observations corroborated earlier studies that suggest the importance of lowering serum LCA/DCA and increasing vitamin B6 intake to help prevent colon cancer. This work can be looked upon as a preliminary step in using 1HNMR analysis as a screening test before invasive procedures.

a_metabolic_study_on_colon_cancer_using_1h_nuclear_magnetic_resonance_spectroscopy

1. Introduction<!>2.1. Sample Collection<!>2.2. Sample Preparation for 1HNMR Spectroscopy<!>2.3.1. Principle Component Analysis (PCA)<!>2.3.2. Partial Linear Square (PLS)<!>2.3.3. Human Metabolome Database (HDMB)<!>2.3.4. Metabolic Pathway Analysis<!>3. Results<!>4. Discussion<!>5. Conclusions

<p>Colorectal carcinoma (CRC) ranked third as the cause of cancer death in the world. It is estimated that 142,820 people will be diagnosed with CRC and 50,830 men and women will die of it in 2014. In the US the death rate has dropped due to screening and the age-adjusted incidence rate is 45.0 per 100,000 [1]. The mortality rate in the US has dipped sharply due to public awareness and insurance support for screening tests after the age of fifty. In Iran the case of CRC is on the rise from 6 to 7.9 per 100,000 in 2005 to 38.0 per 100,000 in 2012 and is the fourth common cancer [2].</p><p>CRC screening is carried out by different procedures such as fecal occult blood test (FOBT), sigmoidoscopy, colonoscopy, virtual colonoscopy, and double contrast barium enema (DCBE). Each has its own advantages and disadvantages. A digital rectal exam (DRE) during routine physical examinations is performed by some physicians and they may use this procedure to check the lower part of the rectum [3].</p><p>Noninvasive methods do not seem to be very accurate but are more economical and easier to perform. There are two FOBT tests to detect the presence of hemoglobin in stool. One uses a dye for detection, guaiac FOBT, and the other fecal immunochemical testing uses specific immunoglobulin. This method is performed every two or three years for individuals above fifty years and is effective in reducing the cases of CRC by 15 to 33 percent [4]. A digital rectal exam and contrast barium enema are also used for screening but cannot detect about 50% of polyps identified by colonoscopy [5]. Sigmoidoscopy and colonoscopy are invasive but seem to be the most effective tools of diagnosis of CRC [6, 7]. A combination of methods used depends on many factors such as patient history, age, and insurance coverage in many countries [8].</p><p>The genetic changes in CRC have been studied extensively. Metabolomics represents one of the new omics sciences which takes advantage of the unique presence and concentration of small molecules in tissues and body fluids to make a fingerprint that can be unique to the individual. Metabolomics has the potential to serve an important role in diagnosis and management of many human disorders such as CRC. More investigations are required as small molecules like the metabolites are difficult to characterize and require high throughput technology [9]. Recently with the advent of new technology like mass spectrometry and 1H nuclear magnetic resonance (1HNMR), the provision for analyzing small molecules are carried out [10]. After obtaining the spectral pattern for the required samples, chemometrics is done using different mathematical modeling like principle component analysis (PCA), partial linear square (PLS), or PLS-DA (discriminate analysis) [11, 12]. This gives us a pattern to distinguish the metabolites between normal and abnormal samples. Using the pattern obtained by the differentiating chemical shifts and the Human Metabolome Database (HDMB), the metabolites are identified. Using other software like Metaboanalyst and KEGG pathway analysis, the main pathways involved are obtained [13].</p><p>Since most CRC screening methods such as colonoscopy and sigmoidoscopy are invasive, we decided to use high throughput technology and find out a metabolite pattern for cancer patients which would differentiate them from normal.</p><!><p>5 mL of blood was collected from people who were on a liquid diet for at least 48 hours and were referred for colonoscopy to the Gastroenterology Department at Amir Alam Hospital, Tehran. One group comprised of 33 patients diagnosed with cancer by colonoscopy and biopsy and the second group of individuals without colon cancer. Sera were separated and stored at −80°C. All groups were made to fill a consent form as per the requirements of Pasteur Institute Ethics Committee before serum collection.</p><!><p>600 μL of serum with 70 μL D2O and 1 mM sodium 2-trimethylsilylpropionate (TMSP) was used as internal reference in a 5 mm NMR tube at room temperature and data acquisition was carried out.</p><p>For NMR data collection, one-dimensional 1HNMR spectra were acquired on a Bruker DRX-500 NMR spectrometer operating at 500.13 MHZ and Carr-Purcell-Meiboom-Gill (CPMG) 90-(t-180-tn-acqusition) (τ = 200, n = 100) pulse sequence as described earlier.</p><p>Analysis of data and pattern recognition were performed using Chenomix 6.4 software.</p><!><p>Initially the NMR variables were mean centered; then PCA was used to detect the outliers data [11] for detecting strong outliers and by Q residuals for detecting modest outliers with 95% confidence level.</p><!><p>PLS is a supervised method that uses multivariate regression technique to extract via linear combination of original variables (X) the information that can predict the class membership (Y). PLS was applied after OSC using the Y matrix including 0 for normal and 1 for abnormal for all the data set. PLS was performed with and without OSC and results were obtained with more than 95% confidence [12].</p><p>OSC filters were developed to remove unwanted variation from spectral data. PLS was applied after OSC using the Y matrix including 0 for normal and 1 for abnormal for all the data set described formerly [14]. Matrix X comprises of 1HNMR data of samples from normals and matrix Y samples from cancer patients. OSC subtracts from X, factors that account for as much as possible the variance in X and are orthogonal to Y. It is important to avoid overfitting after OSC treatment, so as to prevent poor predictive performance. Hence, precise determination of the number of removed OSC factors is very important. Only one factor was removed.</p><!><p>The NMR search link of HDMB was used to detect the metabolites of certain chemical shifts which contains information about metabolites found in the human body [15].</p><!><p>It was performed with Metaboanalyst 2.0 for pathway analysis and visualization. These pathways were affected by the metabolites in colon cancer patients.</p><!><p>PLS was applied after OSC using the Y matrix including 0 for normal and 1 for abnormal for all the data set. PLS was performed with and without OSC and results were obtained with more than 95% confidence level. Figures 1 and 2 show a complete separation pattern between the colon cancer and the normal groups. Figure 3 shows loading plot of the samples which is an indicator of ascending and descending level of metabolites. With the help of the numbers and chemical shifts, 13 metabolites were identified as shown in Table 1. 15 metabolic pathways were detected from the above-differentiating metabolites after an enrichment analysis was carried out in Figure 4. Overrepresentation analysis, as shown in Table 2, was done to detect the impact of pathways, depending on the number of changed metabolites and to test if a particular group of compounds is represented more than expected by chance within the user uploaded compound list in the Metaboanalyst software. In the context of pathway analysis, compounds involved in a particular pathway are enriched and compared by random hits as tested. The detailed results from the pathway analysis are depicted in Table 2, and, since many pathways are tested at the same time, the statistical P values from enrichment analysis are further adjusted for multiple tests.</p><!><p>In this study a number of metabolites and their pathways were detected which have an important impact on colon cancer. Of all the cycles involved, primary bile acid biosynthesis and degradation of ketone bodies and cyanoamino acid metabolism were the major ones (Table 2). Of all the metabolites involved, the main ones were cholesterol, glycine, glycocholic acid, and taurocholic acid which are involved in primary bile acid biosynthesis. Secondary bile acids are formed by enzymatic deconjugation and dehydroxylation of the primary bile acids by anaerobic bacteria in the large intestine. Reports show the presence of higher deoxycholic acids in serum and bile of the patients with colonic adenomas than in the healthy controls [16]. Studies have shown that these secondary bile acids exhibit tumour-promoting ability in animals and they could trigger apoptosis and may also act as regulatory molecules involved in different cell signaling pathways in colon cells [17]. It is interesting that the ratio of LCA (lithocholic acid)/DCA (deoxycholic acid) may be an important factor to distinguish tendency to colon cancer [18]. Reduction of LCA/DCA is reported on experiments carried out on rats fed on a high diet supplemented with vitamin B6 [19]. The second metabolic cycle involved in colon cancer is vitamin B6. There are several theories for the role of this vitamin. It upregulates a protective factor, insulin like growth factor binding protein 1 (IGFBP1); its mRNA is upregulated in HT29 colon carcinoma cells exposed to pyridoxal (a form of vitamin B6). IGFBP1 is secreted from the liver and is hypothesized to exert a protective role in the development of cancer and cardiovascular diseases [20]. ELISA indicated analysis showed that supplemental vitamin B6 significantly lowered levels of colonic HSP70, heme-oxygenase-1, and HSP32 which increase cell proliferation and colonic damage. Heat shock proteins (HSPs, molecular chaperones) have been suggested to be associated with colon carcinogenesis [21]. In mice receiving the colonic carcinogen azoxymethane, the development of colonic aberrant crypt foci, precursor lesions of colon cancer, and cell proliferation are suppressed by vitamin B6 supplementation [22].</p><p>The next effective cycle is the synthesis and degradation of ketone bodies with 3-hydroxybutyric acid as the metabolite involved. It is interesting to note that recent studies suggest that a ketogenic diet helps overcome different kinds of cancers. This is attributed to the fact that most malignant cells depend on glucose as fuel and cannot metabolize fatty acids easily due to dysfunction of the mitochondria. Malignant cells grown in vitro are negatively affected by low glucose and a similar antitumorigenic property of low carbohydrate diets is shown in mice in vivo experiments [23]. 3-Hydroxybutyric acid is also a biomarker detected by GC-MS from sera of colorectal patients [24].</p><p>Glycine is a very important metabolite as it takes part in the next four cycles of cyanoamino metabolism, thiamine metabolism, methane metabolism, and glutathione metabolism. Association between increase in serum glycine and colon cancer has been shown in many studies [25]. Glycine is differentiating metabolites for colon cancer as seen in metabolomic studies by time of flight mass spectrometry (TOFMS) [26]. An in vitro study carried out on cancer cells by metabolite profiling indicates key role of glycine in cancer cell proliferation [27].</p><p>Thiamine or vitamin B1 has an important role in cancer cells as shown by investigations using the thiamine-degrading enzyme, thiaminase. Liu et al. showed that the addition of thiaminase into cell culture media containing thiamine had a significant inhibitory effect on growth of breast cancer cells [28]. Also a pegylated version of thiaminase was capable in delaying tumor growth and prolonging survival in an RS4 leukemia xenograft model [29, 30] and it is seen that the increased glycine and thiamine are linked.</p><p>Methane metabolism by bacteria in the large intestine has been reported as early as 1977. This report showed that excretion of methane in breath occurred twice as frequent in patients with colonic cancer as normal individuals [31]. This suggests the difference between the anaerobic intestinal flora in patients and normal subjects and implies that colorectal cancer may be caused by carcinogens formed by nuclear bile acid dehydrogenation in the large intestine by anaerobic bacteria [32]. The glutathione pathway is seen after increasing glycine. Glutathione levels in primary colorectal cancer tissues were significantly higher than in the corresponding normal tissues. Reports show that elevated glutathione levels had a significant negative effect on survival rate in patients with colorectal cancer [33].</p><p>Reports have shown that in colon cancer the metabolism and catabolism of amino acid increase. Glycine also is seen to participate in glycine, serine, and threonine metabolism which increases in colon cancer and is part of nitrogen metabolism [34].</p><p>(R)-3-Hydroxybutyric acid is very important in butanoate metabolism. The antitumor effects of butyrate were described in studies using colorectal cancer cell lines in which butyrate inhibits growth and induces differentiation and apoptosis [35]. In other studies butyrate was able to inhibit tumor growth in vivo in murine models [36]. But there are conflicting reports about the protective role of butyrate as seen that colorectal cells still increase and grow even though there are high concentrations of butyrate in the colon [37].</p><p>Fucose and mannose metabolism are affected by L-fucose which is an important posttranslational modification in cancer and inflammation. Sera and total cellular proteins of cancer patients showed increase in fucosylation levels and recently some fucosylated proteins have been identified as novel cancer biomarkers in glycoproteomic analyses [38].</p><!><p>Using 1HNMR analysis and chemometrics, a differentiation pattern was obtained between the metabolites in the sera of colon cancer patients and normals. Using HDMB 15 main metabolites were identified and Metaboanalyst software detected 13 metabolic cycles which had been reported as playing an important part in cancers and tumor progression. The main pathways were bile acid biosynthesis and vitamin B6 biosynthesis, and our study corroborates early findings and suggests the importance of lowering serum LCA/DCA and increasing vitamin B6 intake to help prevention of colon cancer.</p>

PubMed Open Access

A Terminal Imido Complex of an Iron\xe2\x80\x93Sulfur Cluster

We report the synthesis and characterization of the first terminal imido complex of an Fe\xe2\x80\x93S cluster, (IMes)3Fe4S4=NDipp (2; IMes = 1,3-dimesitylimidazol-2-ylidene, Dipp = 2,6-diisopropylphenyl), which is generated by oxidative group transfer from DippN3 to the all-ferrous cluster (IMes)3Fe4S4(PPh3). This two-electron process is achieved by formal one-electron oxidation of the imido-bound Fe site and one-electron oxidation of two IMes-bound Fe sites. Structural, spectroscopic, and computational studies establish that the Fe\xe2\x80\x93imido site is best described as a high-spin Fe3+ center, which is manifested in its long Fe\xe2\x80\x93N(imido) distance of 1.763(2) \xc3\x85. Cluster 2 abstracts hydrogen atoms from 1,4-cyclohexadiene to yield the corresponding anilido complex, demonstrating competency for C\xe2\x80\x93H activation.

a_terminal_imido_complex_of_an_iron\xe2\x80\x93sulfur_cluster

<p>Studies of terminal imido complexes of the transition metals continue to push the frontiers of metal–ligand multiple bonding and present new opportunities in catalysis.[1–5] In particular, terminal Fe–imidos have attracted interest in these regards and as models for plausible intermediates in biological nitrogen fixation.[6–12] However, it is unclear in what ways the findings from mononuclear Fe–imido chemistry would translate to the chemistry of Fe–S clusters (such as the catalytic cofactors of nitrogenases). This is in part because mononuclear Fe–imido complexes are themselves electronically diverse and have been characterized across a wide range of oxidation and spin states.[10,13,14] Moreover, the ability of Fe–S clusters to support Fe–N multiple bonding is not well-established, and it is not known how the quintessential electronic features of Fe–S clusters—weak ligand fields and significant Fe–Fe exchange coupling—would impact the Fe–N bonding in an Fe–imido group. Characterization of a synthetic Fe–S–imido cluster could address these questions, and although progress has been made in preparing terminal Fe–imidos in weak ligand fields,[15–20] polynuclear terminal Fe–imidos remain rare,[15,21] and no terminal imido complex of an Fe–S cluster has been reported. As such, we have begun investigating the chemistry of Fe–S–imido clusters, and we herein report the synthesis and characterization of a terminal imido complex of an [Fe4S4] cluster.</p><p>A common route to metal–imido complexes entails two-electron oxidation with a nitrene group transfer reagent such as an organic azide. To extend this methodology to an Fe–S cluster, we used the (IMes)3Fe4S4 platform (IMes = 1,3-dimesitylimidazol-2-ylidene), which features a cuboidal [Fe4S4] cluster in which three Fe sites are supported by IMes ligands and the remaining Fe site is available for reaction chemistry.[22] Mild, one-electron reduction of the previously reported cluster (IMes)3Fe4S4Cl was achieved via Cl-atom abstraction using Cummins's reagent Ti(N[tBu]Ar)3 (Ar = 3,5-dimethylphenyl).[23] The resulting all-ferrous, [Fe4S4]0 cluster was trapped using PPh3 to produce (IMes)3Fe4S4(PPh3) (1) in 77% yield (Scheme 1). Room-temperature addition of 1 equiv 2,6-diisopropylphenyl azide (DippN3) to 1 in C6D6 resulted in an immediate color change from red-brown to green-brown. 1H NMR analysis of the reaction mixture revealed production of 1 equiv PPh3 as well as the terminal imido complex (IMes)3Fe4S4=NDipp (2; Scheme 1), whose structure was confirmed by single-crystal X-ray diffraction (vide infra). Although 2 is sufficiently stable for most characterization techniques, it slowly degrades in solution to give a mixture of species including the anilido complex (IMes)3Fe4S4–N(H)Dipp (3), which was independently synthesized and characterized (see SI); 3 can also be cleanly generated from 2 via H-atom abstraction from weak C–H bonds (vide infra).</p><p>Cluster 2 exhibits C3v symmetry in solution with contact-shifted resonances between +16 and −4.5 ppm in its room-temperature 1H NMR spectrum (Figure S2). The modest shifting of the 1H resonances suggests that 2, like nearly all other [Fe4S4]2+ clusters,[24,25] features an S = 0 ground state with some thermal population of paramagnetic excited states. The diamagnetic ground state of 2 was confirmed by variable-temperature SQUID magnetometry measurements, which exhibit χMT values near 0 cm3 K mol−1 between 2 and 80 K (Figure S12).</p><p>The solid-state structure of 2 (Figure 1)[26] reveals an Fe–N bond length of 1.763(2) Å, which is among the longest Fe–N distances reported for crystallographically characterized Fe–NR complexes (R = aryl, alkyl).[27] The average Fe–N distance in four-coordinate Fe–NR complexes is 1.66 Å (Figure S19); such short bond lengths reflect low- or intermediate-spin electronic configurations with unoccupied Fe–N π* orbitals and Fe–N triple bond character (see Table S10 for representative examples). Notable exceptions include a pair of high-spin Fe3+–aryliminyl ([NAr]•−) complexes with Fe–N bond lengths of 1.766(4) and 1.768(2) Å, respectively, and an Fe3+–alkyliminyl complex with an Fe–N bond length of 1.761(7) Å; all three feature Fe–N bond lengths nearly identical to that observed for 2.[19,28,29] The long Fe–N distance in 2 therefore indicates a high-spin configuration at the unique Fe site, attenuated Fe–N multiple bonding, and perhaps contribution from an iminyl-type electronic configuration (vide infra). However, we note that the computed Fe–N bond lengths in the hypothetical, tetrahedral model complexes [Cl3Fe(NPh)]2– (high-spin Fe3+–imido) and [Cl3Fe(NPh)]– (high-spin Fe3+–iminyl) are identical (1.76 Å; see Table S8), and as such, the redox state of the [NAr] fragment in 2 cannot be unambiguously inferred from the Fe–N bond length.</p><p>Because of the propensity for terminal Fe–imido complexes to exhibit high Fe–N covalency, we anticipated that the Fe–S cluster core of 2 could be geometrically and/or electronically distinct from those of typical [Fe4S4] clusters. Interestingly, however, the cluster's Fe–Fe and Fe–S distances fall within typical ranges for [Fe4S4]n clusters (n = 1+, 2+, 3+)[30] including other structurally characterized members of the (IMes)3Fe4S4 family.[22,31] The [Fe4S4] core of 2 exhibits a characteristic tetragonal compression from idealized Td symmetry, with eight long Fe–S bonds (avg. 2.316(2) Å) defining two opposing [Fe2S2] rhombs, and four short Fe–S bonds (avg. 2.241(1) Å) that connect the two rhombs (Figure 1B).[32] The average Fe–S bond length for the imido-bound Fe (2.310(1) Å) is ca. 0.04 Å longer than those of the IMes-bound sites (vide infra). The modestly elongated—rather than significantly shortened—Fe–S distances at the imido-bound Fe site suggest that it does not adopt a locally high-valent state (e.g., Fe4+, as was observed for the structurally related cluster [Fe4(μ3-NtBu)4(NtBu)Cl3][15]) and is instead consistent with a local valence in the Fe2+–Fe3+ range as is typically observed in Fe–S clusters. Taken together, the unusually long Fe–N bond length and unexceptional cluster core metrics of 2 support a high-spin, mid-valent (Fe2+ or Fe3+) configuration for the imido-bound Fe.</p><p>To further investigate the electronic structure of 2, we undertook a combined Mössbauer spectroscopic and computational study. The 90 K 57Fe Mössbauer spectrum of 2 is relatively sharp (Figure 2) and several simulations fit the data satisfactorily (Figures 2 and S13 and Tables 1 and S1). Before further discussing the simulation, we note some observations based on the spectrum itself. The absence of a quadrupole doublet centered at or below ca. 0 mm s−1 precludes the presence of either a valence-localized Fe4+ ion or a low-spin Fe3+ imido site (see Table S10 for comparisons); this interpretation is consistent with the structural observations discussed above. In addition, the spectral centroid of the Mössbauer spectrum of 2 (0.43 mm s−1) is significantly lower than that of [Fe4S4]+–anilido cluster 3 (0.53 mm s−1; Figure S14); this difference is similar to that observed for other [Fe4S4]2+/+ clusters[33] and is indicative of substantial charge depletion of the Fe sites in 2 relative to those in 3. On this basis, as well as the computational studies described below, we conclude that the contribution of an [Fe4S4]+–iminyl configuration to the electronic structure of 2 is likely minor and that 2 is best described as an [Fe4S4]2+ cluster with a closed-shell imido ([NAr]2−) ligand.</p><p>Broken-symmetry density functional theory (BS DFT) calculations were carried out on a truncated model of 2 to aid in interpretation of its electronic structure and Mössbauer spectrum. Several BS determinants were evaluated, and in all cases the [NAr] fragment converged to a closed-shell, [NAr]2– configuration (see SI). The lowest-energy determinant is characterized by an electronic configuration in which the Fe spins are coaligned within each crystallographically observed [Fe2S2] rhomb, with the two rhombs antiferromagnetically coupled to one another; this picture corresponds to the usual coupling pattern for an S = 0 [Fe4S4]2+ cluster.[24,25] The calculated Mössbauer parameters (Table 1) include isomer shifts (δ) of 0.40 and 0.50 mm s−1 for the imido-bound Fe and the spin-aligned IMes-bound Fe, respectively, with the remaining two IMes-bound Fe sites in the opposing [Fe2S2] rhomb taking on isomer shifts similar to that of the imido site (0.42 and 0.40 mm s−1) but with significantly smaller quadrupole splittings (|ΔEQ|). This approximate 1:2:1 pattern of quadrupole doublets is well-accommodated by the experimental Mössbauer spectrum (Figure 2) and is reflected in our preferred simulation, though we again caution that the data can be fit by several parameter sets (see Table S1).</p><p>To facilitate interpretation of the bonding in 2, the BS DFT-derived canonical orbitals were localized,[34] and Löwdin population analyses were used to derive formal oxidation states (FOSs) of each site (Table 1 and Figure 3; see SI for details). The picture that emerges is a perturbation of the canonical Heisenberg double-exchange model for [Fe4S4]2+ clusters.[24,25] Each Fe site has five singly occupied 3d orbitals in the α (or β) manifold, with an additional electron delocalized between pairs of Fe centers in the β (or α) manifold (Figure 3A). As expected, the two double-exchange-coupled, IMes-ligated Fe sites evenly share their itinerant electron (Figure 3B, left), leading to FOSs of 2.52+ and 2.55+, respectively (consistent with their similar Mössbauer isomer shifts in Table 1). In contrast, the double-exchange interaction between the imido-bound Fe site and its spin-aligned coupling partner is significantly polarized (Figure 3B, right), with FOSs approaching complete valence trapping (2.91+ and 2.16+, respectively). This electronic buffering is reminiscent of that observed for alkylated [Fe4S4]2+ clusters in which the strongly donating alkyl group induces partial localization of Fe3+ character at the alkyl-bound Fe with concomitant localization of Fe2+ character at its spin-aligned partner.[35,36]</p><p>Although we conclude that the imido-bound Fe site is best described as Fe3+, it is noteworthy that its Mössbauer isomer shift is essentially identical to those observed and calculated for the formally more reduced Fe2.5+–IMes sites (Table 1). The relatively high isomer shift for the Fe–imido site can be rationalized by considering the impact of local symmetry on Fe–N and Fe–S covalency: the singly occupied orbitals with Fe–N π* character are also antibonding with respect to the Fe–S bonds, and therefore destabilization of these orbitals due to Fe–N π bonding is compensated by lengthening of the Fe–S bonds, as reflected in the bond metrics (Figure 1B). In this manner, the Fe–N covalency, which lowers the Mössbauer isomer shift, is in part offset by a symmetry-enforced decrease in Fe–S covalency. Overall, these findings demonstrate a dynamic interplay between Fe–N, Fe–S, and Fe–Fe bonding in Fe–S clusters featuring Fe–N multiple bonding.</p><p>Preliminary reactivity studies of 2 demonstrate its proclivity for H-atom abstraction. Exposure of a solution of 2 in Et2O to excess 1,4-cyclohexadiene (C–H bond dissociation free energy (BDFE) in Et2O of 72.2 kcal mol–1)[37] results in clean conversion to anilido complex 3 over 1 h with concomitant formation of 0.5 equiv benzene (Scheme 2 and Figure S6). Cluster 2 does not abstract H-atoms from toluene (C–H BDFE in benzene of 86.5 kcal mol–1)[37] at room temperature or at 50 °C, with no 1,2-diphenylethane detected by 1H NMR spectroscopy under these conditions (Figures S7 and S8). Taken together, these observations suggest that 2, like other high-spin[19,28,29,38] as well as some transiently observed or inferred intermediate-[16,39–41] and low-spin[42,43] Fe–NR complexes, is competent for activation of weak C–H bonds, and that the N–H BDFE of 3 is between ca. 72 and 87 kcal mol–1.</p><p>In conclusion, an [Fe4S4] cluster with a terminal imido ligand was prepared by oxidative group transfer between an organic azide and an all-ferrous [Fe4S4] cluster; the cluster's reducing equivalents are formally derived from the imido-bound Fe site as well as a pair of IMes-ligated Fe sites. The local high-spin state of the Fe3+–imido unit results in attenuated Fe–N multiple bonding and imbues H-atom abstraction reactivity. Overall, this study establishes the ability of Fe–S clusters to support Fe–ligand multiple bonding, demonstrates how the plasticity of Fe–S and Fe–Fe interactions accommodates covalent Fe–N bonding, and links the rich chemistry of Fe–imidos with Fe–S cluster model chemistry. Further studies on the reactivity of 2 and related complexes are underway.</p>

PubMed Author Manuscript

A Third Generation Potentially Bifunctional Trithiol Chelate, Its nat,1XXSb(III) Complex, and Selective Chelation of Radioantimony (119Sb) from Its Sn Target

The therapeutic potential of the Meitner-Auger- and conversion-electron emitting radionuclide 119Sb remains unexplored because of the difficulty of incorporating it into biologically targeted compounds. To address this challenge, we report the development of 119Sb production from electroplated tin cyclotron targets and its complexation by a novel trithiol chelate. The chelation reaction occurs in harsh solvent conditions even in the presence of large quantities of tin, which are necessary for production on small, low energy (16 MeV) cyclotrons. The 119Sb-trithiol complex has high stability and can be purified by HPLC. The third generation trithiol chelate and the analogous stable natSb-trithiol compound were synthesized and characterized, including by single-crystal X-ray diffraction analyses.

a_third_generation_potentially_bifunctional_trithiol_chelate,_its_nat,1xxsb(iii)_complex,_and_select

INTRODUCTION<!>Materials and General Methods.<!>Target Production.<!>Target Irradiation.<!>Target Yield Measurements.<!>Synthesis of Dimethyl 5-(3-Bromo-2,2-bis-(bromomethyl)propoxy)isophthalate [C15H17Br3O5], 1.<!>Synthesis of 5-(3-Bromo-2,2-bis(bromomethyl)propoxy)-isophthalic Acid [C13H13Br3O5], 2.<!>Synthesis of 5-(3-Thiocyanate-2,2-bis(thiocyanatomethyl)-propoxy)isophthalic Acid, [C16H13N3O5S3], 3.<!>Synthesis of 5-(3-Mercapto-2,2-bis(mercaptomethyl)propoxy)-isophthalic Acid [C13H16O5S3], 4.<!>Synthesis of 5-((2,6,7-Trithia-1-stibabicyclo[2.2.2]octan-4-yl)-methoxy)isothalic Acid, 5.<!>Method b:<!>Direct Radioantimony Labeling and Purification.<!>Stability and Cysteine Challenges.<!>Single-Crystal X-ray Diffraction Analysis.<!>RESULTS AND DISCUSSION<!>Trithiol Ligand Synthesis.<!>Sb-Trithiol Complex.<!>X-ray Diffraction Studies.<!>Target Irradiation and Chelator-Based Separation.<!>Radiolabeling.<!>CONCLUSIONS

<p>Targeted radionuclide therapy (TRT) has gained momentum as an attractive cancer treatment method, as seen in recent clinical trials.1,2 Among many TRT radionuclide choices, Meitner-Auger Electron (MAE) emitting radionuclides have unique properties to be advantageously leveraged in cancer TRT.3–7 Antimony-119 (119Sb, t1/2 = 38.19 h, EC = 100% (MAE))8 is considered one of the most promising MAE emitting radionuclides for TRT applications9,10 and has potential application in theranostic nuclear medicine with its imaging radioisotope congener antimony-117 (117Sb, t1/2 = 2.8 h, γ = 85.9%, 158.56 keV, EC = 97.3%, β+ = 1.8%).11</p><p>Antimony-119 is a MAE emitting radionuclide that emits 23–24 conversion electrons and MAEs with energies up to 28 keV and no dosimetrically problematic gammas.12 Electrons with these energies have high linear energy transfer (LET) and consequently high relative biological effectiveness (RBE) in a short range (nm-μm), capable of delivering lethal radiation doses in subcellular ranges while having little off-target toxicity.13 The densely ionizing pathlengths up to 15 μm length in tissue12 reach a cell nucleus from the cell surface.9,14–17 Its lighter radioisotope 117Sb emits a 158.6 keV gamma ray that is ideal for single photon emission computed tomography (SPECT).14 Together, 119Sb and 117Sb form a theranostic (therapy and diagnostic) radionuclide pair, facilitating precise measurement of patient specific biodistribution and calculations of in vivo dosimetry. 119Sb can be produced indirectly via a tellurium generator.18–22 However, direct, no-carrier-added (n.c.a.) production routes of irradiating tin with protons or deuterons avoid medically incompatible eluent systems and dosimetrically problematic tellurium radioisotopes, making the generators difficult to shield.23–25</p><p>Antimony complexes are used for treatment of parasitic infections Schistosomiasis and Leishmaniasis. These complexes deliver pentavalent antimony to the parasites, and it is believed that reduction to Sb(III) interferes with the parasites' thiol redox metabolism.26 Clinical treatments for antimony and fellow pnictogen arsenic poisoning employ two chelators: meso-2,3-dimercaptosuccinic acid (DMSA) and 2,3-dimercaptopropane-1-sulfonic acid (DMPS).27 Antimony is considered a large,28,29 soft atom,30,31 and similar to arsenic, antimony is thiophilic and forms complexes with thiol donor atoms.32–35 The greatest barrier to medical application of 119Sb is the lack of stable complexation with a bifunctional chelator that has the ability to chelate Sb and also covalently bind to a targeting moiety. The only reported chelation of radioantimony was a 117Sb Potassium Antimonyl Tartrate (PAT) complex published in 1969.36 No complex stability measurements are reported, and the PAT complex is incapable of functionalization to molecular targeting agents.</p><p>This work explores stable, selective chelation of radioantimony by a potentially bifunctional trithiol chelator originally developed for the complexation of arsenic.37–40 The trithiol chelator design includes three thiol functional groups to complex radioantimony and two carboxylic acid arms for potential linker incorporation. The selectivity of the novel trithiol for antimony circumvented usual and cumbersome radiochemical isolation of n.c.a. radioantimony from the tin accelerator target material. We describe a novel, economical production route to small quantities of 119Sb, the synthesis and characterization of a novel trithiol chelate, its chemistry with natSb(III), and its radiochemistry with 1XXSb. Tin target recycling, facilitating isotopically enriched 119Sn targetry and production of radioisotopically pure 119Sb, will be reported in a future article.</p><!><p>All solutions were prepared with 18 MΩ·cm deionized water and optima grade H2SO4, HCl, ethanol, and acetonitrile (MeCN) from Fisher Chemical (Hampton, NH). Dimethyl 5-hydroxyisophthalate, pentaerythritol tetrabromide, tris(2-carboxyethyl)phosphine hydrochloride (TCEP), silica gel 60 Å, potassium thiocyanate, antimony trichloride, and potassium carbonate were purchased from Fisher Scientific (Waltham, MA) or Sigma-Aldrich (St. Louis, MO). Silica gel w/UV 254 TLC plates were purchased from Sorbtech Technologies (Norcross, GA). All solvents and reagent grade acids and bases were purchased from Fisher Scientific or Sigma-Aldrich and used without further purification. Unless otherwise noted, only 18 MΩ·cm water was used.</p><p>Microwave reactions were performed using a CEM Discover SP microwave reactor (CEM Corp., Matthews, NC). The 1H and 13C NMR spectra were obtained in CDCl3 or d6-DMSO on a Bruker ARX-500 or 600 MHz spectrometer and calibrated with the respective residual solvent. Infrared (IR) spectra were obtained on a Thermo Nicolet Nexus 670 Fourier transform infrared (FT-IR) spectrophotometer. Elemental analyses were performed by Atlantic Microlabs, Inc. (Norcross, GA). High-resolution mass spectral (HRMS) analyses were performed at the University of Missouri Charles W Gehrke Proteomics Center; briefly, the sample was loaded by an EASY-nLC system with methanol solvent and analyzed by nanoelectrospray ionization in positive-ion/negative-ion mode on a ThermoScientific LTQ Orbitrap XL mass spectrometer. The methanol solvent flow rate was set to 600 nL/min. HRMS data were acquired for 5 min per sample (30,000 resolving power, 120–1000 m/z, 1 microscan, maximum inject time of 500 ms, automatic gain control = 5 × 105).</p><p>Caution! Antimony and Sn radioisotopes are produced by deuteron irradiation of Sn targets as listed in Table 1. All are gamma emitters. Proper radiation safety procedures and shielding were used when handling these radionuclides in laboratories approved for radioisotope use.</p><!><p>Metallic tin was electroplated as previously reported41,42 from a stannous sulfate based electrolytic solution43 onto 2 mm thick silver disks (∅24 mm). Briefly, stannous sulfate (90 mg; Strem Chemicals, Newburyport, MA) was dissolved in 100 μL of concentrated H2SO4. After addition of phenol sulfonic acid (45 μL; Sigma-Aldrich), gelatin (2 mg; Dot scientific, Burton, MI), and 2-naftol (1 mg; Chem Cruz, Dallas, TX), the solution was gently heated (50 °C) and vortexed. Additional DI water was added to a total volume of 1 mL before adding it into the electrodeposition chamber. Voltage control with a constant 3.0 V applied between the platinum wire anode and silver disk cathode resulted in approximately 50 mg of tin deposited onto an 8 mm diameter exposed surface of the Ag disk over a time period of 48 h.</p><!><p>Using a PETtrace (General Electric, Sweden) cyclotron, antimony radioisotopes (Table 1) were produced by irradiating electroplated tin targets with deuterons at an energy of 8 MeV and beam intensities between 20 and 40 μA, inducing natSn-(d,n)1XXSb and few natSn(d,2n)1XXSb nuclear reactions. An ARTMS Quantm Irradiation System (Richmond, BC) automated retrieval of the irradiated target from the cyclotron, reducing operators' radiation exposure. For yield measurements and radiolabeling experiments, targets received 10 and 30 μA·h of integrated beam fluence over 0.5 and 1 h, respectively. The radionuclides produced and their decay characteristics are listed in Table 1.44</p><!><p>Two hours after the end of bombardment (EOB), radionuclide activities were quantified via High Purity Germanium (HPGe) gamma spectrometry using an aluminium-windowed detector (Ortec, Knoxville, Tennessee) coupled to a Canberra (Concord, Ontario) Model 2025 research amplifier and multichannel analyzer calibrated for energy and efficiency using 241Am, 133Ba, 152Eu, 137Cs, and 60Co sources (Amersham PLC, Little Chalfont, U.K.). The system's full width at half-maximum resolution at 1333 keV is 1.8 keV. A second HPGe spectrum collection after all 117Sb decayed (>30 h) allowed quantification of 117mSn activity and subsequent correction for 117mSn contribution to 158.56 keV signal. Target yields were measured and decay corrected to EOB. Using a saturation factor dependent upon radionuclide decay properties and length of irradiation, EOB yield can be converted to yield at End of Saturated Bombardment (EOSB).</p><!><p>Synthesis of compound 1 was accomplished following a modified literature procedure.45 Pentaerythritol tetrabromide (13.8 g; 35.7 mmol), potassium carbonate (9.83 g; 71.1 mmol), and dimethyl 5hydroxyisophthalate (4.98 g; 23.7 mmol) were dissolved in 100 mL of anhydrous dimethylformamide (DMF) under a nitrogen atmosphere in a 500 mL round-bottom flask. The reaction mixture was stirred and heated in a 70 °C oil bath for 24 h. After cooling to room temperature, DMF was removed under vacuum. Deionized water (500 mL) was added to the reaction, and the mixture was extracted with dichloromethane (DCM; 3 × 300 mL). The organic layers were collected and combined, dried over anhydrous sodium sulfate, filtered, and taken to dryness to afford the crude product. The crude product was purified via silica gel column chromatography using hexanes:DCM (3:1) as the mobile phase (dimethyl 5-hydroxyisophthalate, Rf ≈ 0; pentaerythritol tetrabromide; Rf ≈ 0.5; 1, Rf ≈ 0.15). The product was eluted with DCM, and the fractions were collected, combined, and taken to dryness to afford the pure product as a white solid. Yield: 67%, 8.2 g. 1H NMR (CDCl3; 600 MHz) δ ppm: 3.680 (s, 6H, CH2Br), 3.956 (s, 6H, OCH3), 4.147 (s, 2H, OCH2), 7.779 (d, 2H, CH), 8.333 (t, 1H, CH). 13C NMR (CDCl3; 125 MHz) δ ppm: 34.22 (CH2Br), 43.83 (C(CH2)4), 52.89 (CH3), 67.89 (CH2O), 120.09 (OC=C), 124.05 (O=CC=C), 132.14 (O=CC), 158.24 (OC=C), 166.02 (COO). HRMS (m/z): 514.86899 (514.8704 calc'd for [M + H]+ of [C15H17Br3O5]). Elemental Anal. Calc'd (found) for C15H17Br3O5: C, 34.85 (35.03); H, 3.31 (3.27). (See Supplemental Figures S1–S3.)</p><!><p>Synthesis of compound 2 was achieved by hydrolysis of 1. Compound 1 (2.6 g, 5 mmol) and NaOH (2.016 g, 50 mmol) were dissolved in 30 mL of MeCN in a 100 mL round-bottom flask. The reaction mixture was stirred and refluxed at 94 °C for 18 h. Silica gel TLC was used to monitor the reaction progress (ethyl acetate (EtOAc):DCM:glacial acetic acid (20:80:0.5 v/v/v); 1, Rf ≈ 0.95; 2, Rf ≈ 0.4). The reaction was cooled to room temperature, and the MeCN was removed under vacuum. DI water (50 mL) was added to the reaction flask, the mixture was extracted with EtOAc (3 × 50 mL), and the combined organic fractions were back-extracted with DI water (2 × 50 mL). The combined aqueous phase was adjusted to pH 2 with 3 M HCl. A white precipitate formed, and it was collected by vacuum filtration and dried under vacuum. Yield: 1.222 g; 50%. The purity was over 95% by 1H NMR, and no further purification was performed. 1H NMR (DMSO; 600 MHz) δ ppm: 3.701 (s, 6H, CH2Br), 4.113 (s, 2H, OCH2), 7.753 (d, 2H, CH), 8.114 (t, 1H, CH). 13C NMR (CDCl3; 150 MHz) δ ppm: 34.66 (CH2Br), 43.36 (C(CH2)4), 67.38 (CH2O), 119.36 (OC=C), 122.98 (O=CC=C), 132.80 (O=CC), 158.21 (OC=C), 166.30 (COO). HRMS (m/z): 484.82146 (484.822938 calc'd for [M –H]− of [C13H13Br3O5]). Elemental Anal. Calc'd (found) for C13H13Br3O5: C, 31.93 (32.80); H, 2.68 (2.76). (See SupplementalFigures S4–S6.)</p><!><p>Synthesis of compound 3 was performed by thiocyanation of 2. Compound 2 (350 mg, 0.83 mmol) and KSCN (725.2 mg, 8.3 mmol) were suspended in 4 mL of DMF in a 10 mL microwave reaction vessel outfitted with a stir bar. The reaction vessel was placed into the microwave at a fixed power of 1 kW, temperature of 120 °C, and reaction time of 20 min. The reaction mixture was cooled to room temperature, poured into ice water, and then placed in the freezer overnight. The solid was collected by vacuum filtration, dissolved in EtOAc, and then dried over anhydrous sodium sulfate. The solvent was removed under vacuum to yield the product as a white solid. Yield: 140 mg, 40%. X-ray quality crystals were grown by dissolving 3 in a 70/30 (v/v) MeCN/H2O mix at 70 °C and cooling to room temperature. 1H NMR (DMSO; 600 MHz) δ ppm: 3.626 (s, 6H, CH2SCN), 4.288 (s, 2H, OCH2), 7.733 (d, 2H, CH), 8.117 (t, 1H, CH), 13.326 (s, COOH). 13C NMR (DMSO; 150 MHz) δ ppm: 37.34 (CH2SCN), 44.49 (C(CH2)4), 68.76 (CH2O), 113.20 (CH2SCN), 119.36 (OC=C), 122.97 (O=CC=C), 132.55 (O=CC), 157.40 (OC=C), 166.30 (COO). HRMS (m/z): 445.98983 (445.99095 calc'd for [M + Na]+ of [C16H13N3O5S3]). Elemental Anal. Calc'd (found) for C16H13N3O5S3: C, 45.38 (45.38); H, 3.09 (3.17); N, 9.92 (9.49); S, 22.71 (21.82). (See Supplemental Figures S7–S9.)</p><!><p>Compound 4 was prepared by reductive deprotection of 3. Compound 3 (80 mg, 0.189 mmol) was dissolved in 7 mL of 90% MeCN in water in a round-bottom flask and placed in a 55 °C water bath. TCEP (538 mg, 1.89 mmol) dissolved in 70% MeCN in water (7 mL) was added. Ethanol can be used instead of acetonitrile in the deprotection of 3. After 2 h, the reaction was cooled to room temperature, and the solvent was removed under vacuum. The crude product was extracted with EtOAc (3 × 20 mL), which was collected and back extracted with water (3 × 20 mL). The organic layer was collected and dried over anhydrous sodium sulfate. The solvent was removed under vacuum to yield the product as a white powder. Yield: 26.3 mg; 40%. The trithiol product is easily oxidized, and thus the crude reaction was used in situ in the complexation with antimony. 1H NMR (DMSO; 600 MHz) δ ppm: 2.439 (t, 3H, CH2SH), 2.680 (d, 6H, CH2SH), 3.992 (s, 2H, OCH2), 7.705 (d, 2H, CH), 8.084 (t, 1H, CH), 13.294 (s, COOH). (See Supplemental Figure S10.)</p><!><p>Compound 5 was synthesized by two methods. Method a: Compound 3 (15 mg, 0.0354 mmol) and TCEP (101.4 mg, 0.3538 mmol) were dissolved in 5 mL of 70% ethanol in water and stirred on a hot plate at 55 °C for 2 h. After the reductive deprotection reaction was complete, SbCl3 (11.81 mg, 0.0425 mmol) dissolved in 1 mL of ethanol was added, and the reaction mixture was kept at 55 °C for another 45 min. A white precipitate formed during this time. The reaction mixture was cooled, the white precipitate was collected by filtration, washed with water, followed by diethyl ether, and dried in vacuo. Yield: 14.08 mg; 85%. X-ray quality crystals were grown by dissolving 5 (5 mg) in DMSO (2 mL) containing H2O (200 mL) at 70 °C and allowing the mixture to sit at room temperature for 2 weeks. 1H NMR (DMSO; 600 MHz) δ ppm: 3.193 (s, 6H, CH2S), 3.828 (s, 2H, OCH2), 7.649 (d, 2H, CH), 8.093 (t, 1H, CH). 13C NMR (DMSO; 150 MHz) δ ppm: 28.66 (CH2S), 56.02 (C(CH2)3), 76.46 (OCH2), 119.28 (CH), 122.56 (CH), 132.60 (CCO), 158.58 (COCH2), 166.32 (COOH). HRMS (m/z): 464.88606 (464.88795 calc'd for [M –H]− of [C13H13O5S3Sb]). Elemental Anal. Calc'd (found) for C13H13O5S3Sb: C, 33.42 (30.49); H, 2.80 (2.74); S, 20.59 (17.32). FT-IR (cm−1): 1711 (C=O), 1197 (C–O). Note, 1H NMR shows excess residual water at 3.30 ppm. Elemental Anal. Calc'd (found) for C13H13O5S3Sb·2.5H2O: C, 30.48(30.49); H, 3.54 (2.74); S, 18.78 (17.32). (See Supplemental Figures S11–S14.)</p><!><p>Compound 3 (15 mg, 0.0354 mmol) and TCEP (101.4 mg, 0.3538 mmol) were dissolved in 2 mL of 70% ethanol in water in a 10 mL microwave vessel. The vial was capped and placed into the microwave unit and set at a fixed power of 15 kW and a temperature of 70 °C for 5 min to generate 4. Following the reduction reaction, SbCl3 (11.81 mg, 0.0425 mmol) dissolved in 1 mL of ethanol was transferred into the reaction vessel via syringe. The reaction vessel was then microwaved at a fixed power of 15 kW and a control temperature of 70 °C for 5 min. After cooling, the reaction mixture was centrifuged, and the white precipitate was filtered and washed three times with water and three times with ether. Yield: 14.88 mg; 90%.</p><!><p>The trithiol chelator (3, 10 mM in MeCN) was deprotected in 1:1 MeCN:H2O using TCEP (100 mM in H2O) to yield 4. After irradiation, the tin/radioantimony target (∼50 mg) was dissolved in 3 mL of concentrated HCl (3 h, 90 °C), and without purification, the 1XXSb reacted in situ (30 min, 25 °C) with 0.01–1 mM 4. A C18 Sep-Pak (55–105 μm particle size, 125 Å pore size; Waters Corporation, Milford, MA) was preconditioned with 5 mL of ethanol (Fisher Chemical) and 5 mL of H2O. The labeled target solution was diluted (1/20) with H2O and passed through the preconditioned C18 cartridge. Five milliliters of H2O was passed through the cartridge to remove unchelated tin target material, and [1XXSb]5 was eluted in 3 mL of MeCN and dried under nitrogen. Next, 20 sequential 30 min HPGe activity assays quantified 118mSb, 120mSb, 122Sb, and 124Sb, and, by fitting the decay of 117Sb and 117mSn's shared 158.56 keV gamma to double exponential decay equations, determined 117Sb and 117mSn activities in the final purified fraction.</p><!><p>The stability of the [1XXSb] 5 complex was determined by challenging with various chelating agents endogenous to biological systems. After drying, the purified [1XXSb]5 was resuspended in either phosphate buffered saline (PBS; Thermo Scientific), 25 mM cysteine (Thermo Scientific) in PBS, or fetal bovine serum (FBS; ATCC, Manassas, VA) and allowed to sit at room temperature. At various time points (0, 24, 72 h), aliquots were analyzed by reversed phase high performance liquid chromatography (RP-HPLC) using a C18 Acclaim column (4.6 mm I.D. × 250 mm, 5 μm particle size, 120 Å pore size; DIONEX, Sunnyvale, CA) and 1 mL/min flow rate with the following H2O:MeCN gradient: 25% MeCN 0–3.5 min; 25–50% MeCN 3.5–23.5 min; 50–90% MeCN 23.5–24 min; 90% MeCN 24–29 min; 90–25% MeCN 29–30 min; 25% MeCN 30–35 min. For the FBS challenge solutions, an equal volume of MeCN was added to the aliquot to precipitate large serum proteins, which were removed by centrifugation (12,000 rpm, 5 min; Beckman Coulter Microfuge 22R Centrifuge, Brea, CA) prior to HPLC analysis.</p><!><p>Single-crystal X-ray diffraction data were collected on a Bruker X8 Prospector diffractometer (Bruker-AXS, Inc., Madison, WI, USA) using Cu Kα radiation (l = 1.54178 Å) from a microfocus source. The crystals were cooled to 100 K during collection using a Cryostream 700 cryostat (Oxford Cryosystems, Oxford, UK). Hemispheres of data were collected out to resolutions of at least 0.81 Å using strategies of scans about the phi and omega axes. Unit cell determinations, data reduction, absorption corrections, and scaling were performed using the Bruker Apex3 software suite.46 The crystal structure of 3 was solved by an iterative dual space approach as implemented in SHELXT,47 and 5 was solved by direct methods.48 Both structures were refined by full-matrix least-squares refinement using SHELXL49 implemented via Olex2.50 Non-hydrogen atoms were located from the difference maps and refined anisotropically. Hydrogen atoms were placed in calculated positions, and their coordinates and thermal parameters were constrained to ride on the carrier atoms. The crystal structure of 3 was found to contain regions of disordered solvent that could not be accurately modeled; these were treated by applying a solvent mask as implemented in PLATON SQUEEZE.51 Six hundred thirty-seven electrons were removed from a total void volume of 2014 Å3 per unit cell, equivalent to 1.5 acetonitrile molecules per formula unit. Crystal data, structure refinements, and bond distances and angles are reported in the Supporting Information (Tables S1–S4).</p><!><p>In targeted radionuclide therapy contexts, 119Sb has potential to effectively eliminate disease when targeting cellular locations outside of the cell nucleus14 while mitigating off target radiation toxicities.9,15 However, exploration of this potential is limited to in silico studies due to a lack of stable complexing agents capable of functionalization to a targeting moiety. No literature reports of stable radioantimony complexation by a bifunctional chelator exist.52 Complexation of radioantimony builds upon previous work complexing fellow pnictogen arsenic. Antimony is the heavier congener of arsenic. Both arsenic and antimony are thiophilic, generally found in nature as their sulfides (M2S3) and dithiolates (e.g., DMSA, DMPS, British anti-Lewisite (BAL)) are used for treating arsenic and antimony poisoning.53 Thiolate compounds of Sb(III) have been reported, including a recent trithiolate complex.54–57 Previously, various trithiol ligands and their natAs and 77As complexes were reported.37,39,40 This is the third iteration in our development of a hydrophilic bifunctional trithiol chelate.</p><!><p>A trithiocyanate protected trithiol ligand 3 was synthesized as shown in Scheme 1 and deprotected to 4 just prior to use as free thiols (−SH) are readily oxidized. Since alkyl bromides are good leaving groups and are readily substituted with the thiocyanate thiol-protecting group, pentaerythritol tetrabromide was used as the starting material. The two ester groups were then hydrolyzed to their carboxylic acids with one carboxylic acid group available for bioconjugation with a targeting moiety (e.g., peptide, antibody) and the other carboxylic acid improving hydrophilicity. Thiocyanate is a very stable thiol-protecting group and is easily reduced to the free thiol with the mild reducing reagent TCEP.</p><p>Compound 3 was reduced with TCEP in MeCN or ethanol with 20–50% water to form the reduced form of the free thiol −SH (compound 4) just prior to complexation with Sb(III) to form compound 5. Free thiol groups are readily oxidized and must be used in situ. The advantage of this synthetic route is that, following reduction to the free-thiol, no separation was needed for the metal complexation reaction with Sb(III), either at the macroscopic or radiotracer (nM or less) level, similar to observations with As(III).37,39</p><p>Compounds 1–3 were characterized by their elemental analyses, 1H NMR and 13C NMR spectra and HRMS mass spectra. Trithiol 4 was only characterized by its 1H NMR spectra due to its ease of oxidation. Compound 3 was also characterized by single-crystal X-ray diffraction (Figure 1, Supplemental Figure S15). This trithiol chelator is expected to further improve the hydrophilicity of the resultant radiometal bioconjugate compared to its previous trithiol analogues.37,39,40 The previous two trithiol bioconjugates were quite lipophilic based on the high hepatobiliary clearance of their 77As labeled trithiol-bioconjugates in mice.37,39 The presence of two carboxylic acid groups in 3 (Figure 1) allows one to be conjugated to a targeting moiety while leaving the second one to increase hydrophilicity relative to the previous analogues.</p><!><p>The antimony trithiol 5 was successfully synthesized according to reaction Scheme 2. Precursor 3 (1 equiv) was reduced to trithiol 4 with TCEP (10 equiv) after which SbCl3 (1.2 equiv) in ethanol solution was added. The product immediately precipitated from solution. The synthesis was evaluated by both conventional and microwave heating methods. Both reaction methods provided high yields over 85% of the product; however, the microwave reaction was more time efficient (10 min vs 45 min). The antimony trithiol (5) was characterized by 1H NMR and 13C NMR spectroscopy, FT-IR spectroscopy, HRMS, and elemental analysis. X-ray quality crystals of 5 were obtained from a DMSO/H2O mix (Figure 2; Supplemental Figures S16–S17). The 1H NMR spectrum shows the expected disappearance of the −SH protons and a downfield shift of the −CH2S protons of 4 on coordination to Sb. The FT-IR shows the expected stretches from the −COOH groups at 1711 cm−1 (C=O) and 1197 cm−1 (C–O).</p><!><p>Compound 3 crystallized from MeCN and water in the rhombohedral R3¯ space group. It packed forming discrete hydrogen bonded rings containing six molecules, all interacting through the carboxylic acid groups. The six-molecule rings stacked above and below each other and interfaced through the thiocyanate groups, which interact with each other through electric dipole interactions and with the phenyl π system. The disordered solvent is modeled as 1.5 acetonitrile molecules per formula unit. Bond lengths and angles are very similar to the previously characterized trithiocyanate protected trithiol ligands.40 Selected bond distances and angles are listed in Table 2.</p><p>Compound 5 crystallized from DMSO and water over 2 weeks. The crystal structure of compound 5 (Figure 2) confirms the expected molecular structure, which is consistent with its solution structure based on NMR and HRMS. The molecule crystallized with 2 equiv of DMSO in the monoclinic space group P21/c. Both acid groups are protonated, and the three thiols are coordinated trigonally to the Sb(III), making it a discrete, neutral molecule. The geometry about the Sb atom is trigonal pyramidal, but the S–Sb–S bond angles are all closer to 90° (an octahedron missing 3 vertices) than they are to 109.5° (a tetrahedron missing one vertex). The S–Sb–S bond angles in the structure are in agreement with those previously reported.54,58,59 Previously characterized As-trithiol complexes have bond angles about arsenic close to 97°.40 The average Sb–S bond length is 2.440(8) Å, which is in agreement with the previously reported average Sb–S (2.447(7) Å) bond distance.54,58,59 Selected bond distances and angles are listed in Table 2.</p><p>Several different interactions influence the crystal packing in this structure. An important interaction appears to be between the Sb atom and the S atoms of adjacent molecules, which organize the molecules into chains along a (Supplemental Figure S17). Each Sb atom interacts with 2 S atoms from one neighboring molecule (Sb·····S2 = 3.6987(6) Å, Sb·····S3 = 3.5370(6) Å) and a third S atom from a second neighboring molecule (Sb·····S1 = 3.2529(5) Å), although they are not octahedrally arranged about the Sb. This seems to be an interaction of the Lewis acidic Sb and the Lewis basic nonbonding S lone pair and has been observed in previously reported structures.54,58,59 There are also two short S·····S distances between neighboring molecules (S1·····S2 = 3.3212(7) Å, S2·····S3 = 3.3829(7) Å). These interactions polymerize the structure into an infinite chain parallel to the a axis (Supplemental Figure S17). Additionally, the acids groups each donate a hydrogen bond to the oxygen of a DMSO molecule (Table 2) and accept a hydrogen bond from the methyl group of a different DMSO molecule. These hydrogen bonds cross-link the chains into a network.</p><!><p>A choice cyclotron target provides strong thermal adherence and electrical conductance, dissipating kW/cm2 beam thermal power, conducting tens of microamperes of electrical current, and withstanding transitions from microtorr pressures to standard atmosphere. Electrodeposition uses electrical reduction to transition metal ions in solution to a solid phase, forming a metallic material that is electrically and thermally well-joined to a durable, rigid backing. Electroplated natSn targets with lineal mass density 88–128 mg/cm2 withstood 8 MeV energy deuterons at a beam current of 40 μA. End of saturated bombardment (EOSB) theoretical, estimated, and measured yields and EOB measured yields are reported in Table 3. The yield at the end of bombardment, AEOB, is described mathematically according to eq 1, where N is the target atom density, x is the thickness of material slice under consideration, I is the accelerated particle current, σ is the nuclear reaction cross section at specified particle energy, λ is the decay factor, and t is the time duration of irradiation.</p><p>A column was added to Table 3 listing EOB values for comparison.</p><p>The yield at EOSB, or "end of saturated bombardment", AEOSB, describes the yield when the rate of production is equal to the rate of radioactive decay. It is convenient for comparing yields of short-lived radionuclides. EOSB yield is described by eq 2.</p><p>Conversion between EOB and EOSB is described in eq 3 and is dependent only upon the half-life of the radionuclide and the length of time duration of the irradiation.</p><p>We report a measured EOSB yield as opposed to a theoretical number predicted from a modeled nuclear excitation function.</p><p>Proton irradiation of natSn produces larger quantities of long-lived radioantimony impurities: 124Sb (t1/2 = 60.2 d), 122Sb (t1/2 = 2.72 d), and 120mSb (t1/2 = 5.76 d). Deuteron irradiation of natSn produces only small quantities of 120mSb, 122Sb, 124Sb, the relatively short-lived 117Sb (t1/2 = 2.8 h), and 118mSb (t1/2 = 5.0 h) in addition to desired 119Sb, dramatically reducing troublesome longer-lived contaminants. For example, deuteron irradiation of the most abundant tin isotope, 120Sn (32.58%), predominantly forms stable 121Sb via the 120Sn(d,n)121Sb reaction. With a reaction threshold of 4.7 MeV, the 122Sn-(d,2n)122Sb nuclear reaction is initiated when bombarding natSn, producing less long-lived 122Sb than proton bombardment of natSn. Using TALYS theoretical excitation functions,60 a typical 1 h, 40 μA 8 MeV deuteron irradiation of a 100 mg natSn target theoretically produces approximately 47 MBq of 119Sb, 350 MBq of 117Sb, 22 MBq of 118mSb, 1.6 MBq of 120mSb, 2.8 MBq of 122Sb, 0.1 MBq of 124Sb, and 2.6 kBq of 125Sb at EOB. A similar length and intensity 12 MeV proton irradiation would produce a factor of 6 more long-lived 120mSb (10.4 MBq), 14 times more 122Sb (40 mBq), and over 6 times the activity of very-long-lived 124Sb (0.9 MBq). Although less suitable than proton irradiation for producing large quantities of 119Sb, deuteron irradiation produces relatively low 122Sb and 124Sb activities, the longest lived and most problematic antimony radioisotopes in a working laboratory environment, while also producing relatively larger amounts of 119Sb. The deuteron targets also require less natSn material because deuterons have a shorter range than protons, reducing possible contaminant mass in processing. Thus, deuteron irradiation of natSn is an economical route to developmental, preclinical quantities of 119Sb and 120mSb.</p><p>Due to the inability to resolve the low characteristic photon emission of 119Sb (23.87 keV), the EOSB yield of 119Sb was not measured directly. Instead, 119Sb EOSB yield was estimated using the ratio of 117Sb to 119Sb TALYS theoretical excitation functions60 and the measured 117Sb EOSB yield. Because 117Sb and 117mSn share a 158.56 keV characteristic emission, 117mSn activities within irradiated targets were quantified after 117Sb decayed away (>30 h), and measured 117Sb EOSB yields were corrected for 117mSn 158.56 keV signal contribution.</p><!><p>Trithiol 4 rapidly complexes n.c.a. pmol quantities of antimony in 30 min at room temperature in the presence of mmol quantities of the tin target (108-fold excess) at ligand concentrations down to 0.01 mM. At low radiolabeling concentrations of 10 μM, 30 nmol of ligand was used to complex 5.8 pmol of various antimony isotopes—calculated from HPGe gamma spectroscopy measured radioantimony activities, theoretical accelerator produced stable antimony isotopes, and trace metal contaminant limits of detection from reagents within the dissolved target solution. After an irradiation of 30 μAh 8 MeV deuterons, the calculated molar activity for 49 MBq of 117Sb is 8.4 MBq/pmol, and the calculated molar activity for an estimated 240 kBq 119Sb is 42 kBq/pmol. HPLC traces of radioantimony unbound and bound by 4 are shown (Figure 3). The nonradioactive [natSb]5 standard coelutes with radioactive [1XXSb]5 with a retention time of 23.7 min (Figure 4). The trithiol chelator 4 has a strong selectivity for antimony over tin as seen by the full chelation of radioantimony (<nM) in the presence of macroscopic quantities of tin and harsh (∼6 M HCl) solvent conditions. The ability to directly radiolabel radioantimony quantitatively from unseparated target material greatly simplifies radiopharmaceutical production. Greater than 99% of [1XXSb]5 activity was trapped onto the C18 cartridge and eluted with a radiochemical yield of 65% ± 20% (N = 3).</p><p>[1XXSb]5 was stable over 72 h when challenged with biologically relevant complexing agents. HPLC analyses showed [1XXSb]5 to be 91% ± 9% (N = 3) intact in 25 mM cysteine at 72 h and 97.5% ± 1.6% (N = 3) intact in FBS at 72 h. HPLC traces of [1XXSb]5 challenged with 25 mM cysteine at 72 h and FBS at 24 h are reported in the Supporting Information (Supplemental Figure S18). [1XXSb]5 is thus expected to be stable in vivo. Unchelated compound 4 has a retention time of 24.5 min (Figure 4), which is approximately 0.5 min longer than [1XXSb]5, and the two peaks would be distinguishable and separable. The molar activity of the final solution can be increased by removing nonradiolabeled chelator from [1XXSb]5. The reaction of trithiol chelator 4 with antimony is the first reported complexation of radioantimony with a bifunctional chelator capable of both stably retaining the radiometal and providing a linker group that can be conjugated to disease targeting moieties.</p><p>Fractions isolated from the HPLC [1XXSb]5 peak, Rt = 24.0 min, were assayed by HPGe, and the spectra are shown in Figure 5 left. To determine the degree to which 4 complexes solely antimony and not tin target material, 117mSn activity within a HPLC purified fraction was measured. Antimony-117 decays to ground state 117Sn.11 The fitted, logged decay of the sample enabled a half-life measurement that distinguishes 117Sb (2.80 h) from 117mSn (14 d). Fitting the 158.56 keV photopeak over a 30-h span constructed the decay curve (Figure 5 right) and measured a half-life of 2.86 ± 0.02 h, 2.3% larger than the true 117Sb half-life, 2.80 h.11 A measured half-life larger than the true half-life indicates the presence of a longer-lived radionuclide, in this case, 117mSn.</p><p>Double exponential decay equations describe and quantify relative activities of mixed radionuclide samples in time; using this fit, the 117mSn activity coeluting with [1XXSb]5 at HPLC separation was calculated to be 11.29 ± 0.12 Bq, and comparing initial reaction to final purified 117mSn activity provides a tin decontamination factor of 1.41 × 103. This decontamination factor, describing the high level at which tin target material was removed from the radiopharmaceutical, is impressive for a nontraditional radiochemical production and resulted in a final formulation of tin mass that is orders of magnitude below the estimated daily intake of 4.003 mg inorganic tin for an adult in the United States.62 For a target with lineal mass density of 120 mg/cm2, the final sample holds an estimated 45 μg of tin. The Agency for Toxic Substances and Disease Registry (ATSDR) reports no evidence that inorganic tin is a neurotoxin, mutagen, carcinogen, or immunotoxin or affects reproduction or development in humans.62 No radionuclidic impurities were observed besides the various radioantimony isotopes (useful for radioantimony activity quantification) and 117mSn (Figure 5 left).</p><!><p>Deuteron irradiation of inexpensive natSn targets produces a profile of radioantimony isotopes with far shorter half-lives compared to proton bombardment while still producing preclinical quantities of 119Sb, making it a better production route for research and development purposes. The various radioantimony isotopes allow useful tracking and quantification of radioantimony activity. No unexpected radionuclide impurities were observed in the final product. We report a method for radioantimony chelation using a functionalizable trithiol ligand, circumventing usual radionuclide isolation from dissolved accelerator targets. This is the first report of radioantimony complexation with a chelator capable of bifunctionalization—an essential step toward exploration of 117Sb and 119Sb in theranostic applications of targeted radionuclide therapeutic contexts.</p>

PubMed Author Manuscript

Functional Similarities between the Protein O-Mannosyltransferases Pmt4 from Bakers' Yeast and Human POMT1*

Protein O-mannosylation is an essential post-translational modification. It is initiated in the endoplasmic reticulum by a family of protein O-mannosyltransferases that are conserved from yeast (PMTs) to human (POMTs). The degree of functional conservation between yeast and human protein O-mannosyltransferases is uncharacterized. In bakers' yeast, the main in vivo activities are due to heteromeric Pmt1-Pmt2 and homomeric Pmt4 complexes. Here we describe an enzymatic assay that allowed us to monitor Pmt4 activity in vitro. We demonstrate that detergent requirements and acceptor substrates of yeast Pmt4 are different from Pmt1-Pmt2, but resemble that of human POMTs. Furthermore, we mimicked two POMT1 amino acid exchanges (G76R and V428D) that result in severe congenital muscular dystrophies in humans, in yeast Pmt4 (I112R and I435D). In vivo and in vitro analyses showed that general features such as protein stability of the Pmt4 variants were not significantly affected, however, the mutants proved largely enzymatically inactive. Our results demonstrate functional and biochemical similarities between POMT1 and its orthologue from bakers' yeast Pmt4.

functional_similarities_between_the_protein_o-mannosyltransferases_pmt4_from_bakers'_yeast_and_human

Introduction<!>S. cerevisiae Pmt4 but not Pmt1-Pmt2 Complexes Mannosylate the Human POMT Substrate α-Dystroglycan in Vitro<!><!>S. cerevisiae Pmt4 but not Pmt1-Pmt2 Complexes Mannosylate the Human POMT Substrate α-Dystroglycan in Vitro<!><!>Pmt4 and Pmt1-Pmt2 Have Distinct Detergent and Acceptor Substrate Requirements in Vitro<!><!>Pmt4 and Pmt1-Pmt2 Have Distinct Detergent and Acceptor Substrate Requirements in Vitro<!>Yeast Pmt4 and Human POMTs Have Similar Preferences toward Mannosyl Acceptor Peptides in Vitro<!><!>Impact of POMT1 Mutations on Pmt4 in Vitro Mannosyltransferase Activity<!><!>Impact of POMT1 Mutations on Pmt4 in Vivo Mannosyltransferase Activity<!><!>Impact of POMT1 Mutations on Pmt4 in Vivo Mannosyltransferase Activity<!>Discussion<!>Yeast Strains and Plasmids<!>Immunoprecipitation<!>Preparation of Crude Membranes and Cell Wall Extracts from S. cerevisiae<!>Preparation of GST-tagged αDG Mucin Domain<!>Biotinylated Peptide Acceptors<!>Pmt4 Specific in Vitro Mannosyltransferase Activity Assay<!>In Vitro Mannosyltransferase Assay Using Biotinylated Peptide Acceptors<!>Western Blotting Analysis<!>Author Contributions<!>

<p>O-Mannosylation of secretory and membrane proteins is a conserved, essential modification in eukaryotes. In yeast, this post-translational modification is important for the biosynthesis and maintenance of the cell wall, and even affects quality control of proteins in the endoplasmic reticulum (ER)6 (reviewed in Refs. 1 and 2). In humans, a heterogeneous group of congenital muscular dystrophies, collectively referred to as secondary α-dystroglycanopathies, is connected to the reduced O-mannosylation of the cell surface-associated basement membrane receptor α-dystroglycan (αDG) (reviewed in Ref. 3). Furthermore, O-mannosyl glycans are also commonly present among members of the cadherin and plexin families and influence cadherin-mediated cell-cell adhesion (4–6).</p><p>Biosynthesis of O-mannosyl glycans is initiated by a family of protein O-mannosyltransferases (PMTs), which is highly conserved among eukaryotes except plants and nematodes (reviewed in Refs. 1 and 7). PMTs catalyze the transfer of mannose from dolichol-phosphate mannose (Dol-P-Man) to the hydroxyl group of serine or threonine residues of proteins in the lumen of the ER. These enzymes are essential for the viability of yeasts, filamentous fungi, and animals (5, 8–12). Numerous mutations in the human PMTs (POMT1, POMT2) have been identified that cause various forms of α-dystroglycanopathies, with Walker-Warburg syndrome (WWS) being the most severe form of these disorders (13–16).</p><p>In bakers' yeast, the redundant PMT family is grouped into three subfamilies: PMT1 (Pmt1 and Pmt5), PMT2 (Pmt2, Pmt3, and Pmt6), and PMT4 (Pmt4). The main mannosyltransferase activities are due to Pmt1, Pmt2, and Pmt4 (8, 17). Today, the best studied family member is Pmt1. A topology model comprising seven transmembrane spans and two prominent luminal loops has been established for Saccharomyces cerevisiae Pmt1 and is likely conserved in other eukaryotic PMTs (18). An acceptor substrate binding domain has been mapped to Pmt1-loop1 using photoreactive peptides, but amino acids crucial for enzyme function are spread over the entire protein, which hampered attempts to pinpoint the catalytic center via targeted mutagenesis (19, 20). The large hydrophilic Pmt1-loop5 domain is crucial for enzyme function although it is most probably not involved in the basic catalytic mechanism (20). Database mining has revealed that this loop contains three conserved, so called MIR motifs (21), but the function of those, as well as the entire loop5 is undefined.</p><p>The yeast PMT1 and PMT2 family members and Pmt4 differ in several aspects. PMT1 and PMT2 mannosyltransferases form distinct heteromeric complexes, whereas, Pmt4 acts as a homomeric complex (22). Mutations of a conserved DE-motif in the loop1 domain that influence protein substrate binding of Pmt1, differentially affect mannosyltransferase activities of Pmt1-Pmt2 and Pmt4 (19). Pmt1-Pmt2 and Pmt4 act on distinct protein substrates in vivo, and Pmt4 preferentially modifies membrane-anchored proteins (23, 24). Furthermore, in vitro assays suitable to measure the enzymatic activity of Pmt1-Pmt2 did not monitor Pmt4 mannosyltransferase activity, further pointing to distinct acceptor requirements (8). How PMTs recognize their acceptor substrates is still enigmatic although computational and experimental approaches have been conducted to define consensus mannosylation motifs (reviewed in Ref. 1; 25).</p><p>In contrast to yeast, in mammals only two PMTs, namely POMT1 and POMT2 are present (reviewed in Ref. 26). However, although Pmt2 is the closest homolog of mammalian POMT2, POMT1 is a homolog of yeast Pmt4, and the PMT1 subfamily is absent in higher eukaryotes. POMT1 and POMT2 have been demonstrated to act as a heteromeric complex, however, when compared with yeast different amino acid residues might govern complex formation (27, 28). Furthermore, it was suggested that the two mammalian proteins contribute differentially to mannosyltransferase activity (28). POMT1-POMT2 act on αDG in vivo and in vitro (29, 30). But, just like for yeast PMTs, mannosylation motifs are poorly defined.</p><p>The molecular analysis of eukaryotic polytopic transmembrane protein O-mannosyltransferases is still a challenge to which heteromeric complex formation adds a further level of complexity (19). Thus, homomeric Pmt4 appears to be a promising model to further characterize this essential class of enzymes. In this study we describe in vitro properties of Pmt4 from bakers' yeast and show its functional relationship with human POMT1.</p><!><p>The PMT4 family mannosyltransferases from bakers' yeast (Pmt4) and human (POMT1) show a high degree of conservation (Fig. 1) (26). To establish an in vitro assay to monitor Pmt4-mediated mannosyl transfer, we thus tested conditions previously used for in vitro activity measurements of the mammalian POMTs (31). Indeed, the use of GST-tagged α-dystroglycan mucin domain (GST-αDG) as acceptor substrate and β-octylthioglucoside (β-OTG) as detergent enabled detection of yeast Pmt4 activity in vitro (Fig. 2A) (19). In a reaction mixture containing [3H]mannose-labeled Dol-P-Man as donor substrate and crude membranes isolated from wild-type yeast strain SEY6210 as enzyme source, typically 10 to 15% of the tritiated mannose were transferred to GST-αDG, but not to GST alone (Fig. 2A). Remarkably, this assay exclusively monitored Pmt4 activity because in total membranes from a pmt4Δ null mutant strain mannosyltransferase activity was around background level, although all PMT1 and PMT2 family members were present (Fig. 2A). Expression of a FLAG-tagged version of Pmt4 (Pmt4FLAG) from a multicopy plasmid restored in vitro mannosyltransferase activity in the latter strain. The roughly 2-fold increase in mannosyl transfer (Fig. 2A) correlated well with 2.1-fold higher enzyme content when compared with wild-type membrane preparations (Fig. 2B).</p><!><p>Conservation of PMT4 family members from fungi and mammals. Alignment of the regions surrounding the characterized WWS-associated mutants (highlighted in red) of PMT4 family proteins from S. cerevisiae (ScPmt4), S. pombe (SpOma4), Candida albicans (CaPmt4), human (HsPOMT1), mouse (MmPOMT1), and rat (RnPOMT1) are shown on the right. The location of the amino acid exchanges G76R/I112R and V428D/I435D (red stars) is illustrated within a topology model of PMT/POMTs (left). The conserved DE motif and the three MIR domains are marked.</p><p>Pmt4 in vitro mannosyltransferase activity. A, crude membranes from wild-type strain SEY6210 (WT) and pmt4 null mutant transformed with pRS423 (vector) or pJK4-B1 (Pmt4FLAG) were tested for in vitro mannosyltransferase activity. The standard assay containing 29,067 dpm of Dol-P-[3H]Man per reaction was performed as described under "Experimental Procedures." Mean ± S.D. values of at least three independent experiments are shown. B, quantification of the Pmt4 content of the membranes used in A. Western blottings were probed with anti-Pmt4 and anti-Sec61 antibodies. Signal intensities were quantified as outlined under "Experimental Procedures." The indicated numbers represent the relative Pmt4 content (ratio Pmt4:Sec61) with reference to WT, which was set as 1. C–H, characterization of the Pmt4 in vitro mannosyltransferase activity. Standard assays containing between 20,000 and 30,000 dpm of Dol-P-[3H]Man per reaction were performed. Mean ± S.D. values of independent experiments are shown (n = 2 to 4). Dependence of the in vitro mannosyltransfer reaction on the concentration of the mannose acceptor GST-αDG (C), the reaction time (D), the membrane protein input (E), the reaction temperature (F), the pH (G), and the presence of divalent cations and EDTA at the depicted concentrations (H) was determined using SEY6210 membranes as an enzyme source.</p><!><p>Based on these observations, a standardized Pmt4 in vitro mannosyltransferase activity assay was elaborated (see "Experimental Procedures") using membrane preparations from wild-type yeast, and various parameters were characterized and optimized. Variation of the GST-αDG input yielded a plateau at around 0.1 μg/μl validating that the acceptor substrate was not limiting at a concentration of 0.2 μg/μl, which was routinely used in standard reactions (Fig. 2C). Time course experiments revealed a plateau of the reaction after ∼15 min, although a significant amount of the [3H]mannose was not detected on GST-αDG at that time (Fig. 2D). For up to 2 μg/μl of total membrane protein the incorporation of [3H]mannose increased proportional to the enzyme input (Fig. 2E). Temperatures above 25 °C inactivated Pmt4, which showed an optimal activity between 20 and 25 °C (Fig. 2F). The optimal pH was determined to be around 7.5 (Fig. 2G). At concentrations between 1 and 10 mm, the divalent cations Mg2+, Ca2+, and Mn2+ had a beneficial effect on Pmt4 activity, but EDTA did not significantly impact on the mannosyl transfer (Fig. 2H).</p><p>Increasing Dol-P-Man concentrations enhanced the in vitro transfer of [3H]mannose almost linearly (Fig. 3A). At all concentrations the amount of the [3H]mannose transferred to GST-αDG was proportional to the amount of the Pmt4 input (Fig. 3A, WT: 0.5 and 0.25 μg/μl of membranes) although relative incorporation levels did not increase to more than 25% (Figs. 2A and 3A). Likely explanations of this limited incorporation are the transfer of [3H]mannose to endogenous mannosyl acceptors by PMTs and other Dol-P-Man utilizing enzymes present in the crude membranes. This view is supported by the fact that overexpression of Pmt4FLAG steepened the level of [3H]mannose transferred to GST-αDG (Fig. 3A, Pmt4FLAG: 0.5 μg/μl of membranes), and that additions of fresh Dol-P-Man to the standard assay in 15-min intervals linearly increased incorporation of [3H]mannose into GST-αDG (Fig. 3B).</p><!><p>Dependence of the Pmt4 in vitro mannosyltransferase activity on the Dol-P-[3H]Man input. A and B, unless otherwise stated, in vitro mannosyltransferase activity was determined under standard conditions including 4 μg of the mannosyl acceptor GST-αDG as detailed under "Experimental Procedures." A, crude membranes from strains SEY6210 (WT; squares, 0.5 μg/μl; diamonds, 0.25 μg/μl) and pmt4 expressing pRS423 (pmt4; triangles, 0.5 μg/μl) or pJK4-B1 (overexpression of Pmt4FLAG; circles, 0.5 μg/μl; see also Fig. 2B), were used as enzyme source and in vitro mannosyltransfer at the indicated donor substrate concentrations was determined. Results of independent experiments are shown in different shades. B, SEY6210 (WT) membranes were used and in vitro mannosyltransfer was determined at indicated time points for the standard reaction and for fresh Dol-P-[3H]Man addition every 15 min (indicated by arrows) (n = 2–5). The average Dol-P-[3H]Man input was between 18,000 and 32,000 dpm in the standard reaction and between 23,000 to 34,000 dpm for all consecutive additions.</p><!><p>Our analysis revealed that in the presence of β-OTG as a detergent yeast Pmt4 can mannosylate the mammalian POMT substrate GST-αDG (Fig. 2A). In a pmt4Δ mutant, other O-mannosyltransferase activities were below the detection limits of the assay (Fig. 2A), although the Pmt1-Pmt2 complex is fully active in vitro in the absence of Pmt4 (with Triton X-100 as detergent) (8). In contrast, our previous studies showed in vitro mannosylation of the acceptor peptide bio-YATAV by Pmt1-Pmt2, but not by Pmt4, in the presence of the detergent Triton X-100 (17, 32). To further address in vitro mannosyl acceptor specificities of the yeast PMT family members, we first analyzed the detergent requirements in more detail. To individually record endogenous Pmt4 and Pmt1-Pmt2 activities, membranes from pmt1 and pmt4 deletion mutants, respectively, were used as an enzyme source. For vivid depiction, membranes from these strains are identified as pmt1/Pmt4 and pmt4/Pmt1/2 in Table 1 and Fig. 4. Under the conditions applied, in mutant pmt1Δ in vitro enzymatic activity of Pmt2 and other PMT1 and PMT2 family members is negligibly small (17). Even in the presence of Triton X-100, GST-αDG did not serve as acceptor substrate of Pmt1-Pmt2 (data not shown). But, a proven POMT in vitro substrate, the αDG-derived synthetic peptide including amino acids 401 to 420 (29) (Fig. 4A) qualified as a more general mannosyl acceptor. Following the transfer of [3H]mannose from Dol-P-Man to peptide 401–420-bio showed that Pmt4 and Pmt1-Pmt2 complexes were both active with β-OTG as detergent, and showed similar β-OTG optima at ∼1.5 × critical micelle concentration (∼0.42%). Although protein levels of Pmt4 are at least 2–3 times lower when compared with Pmt1 and Pmt2 (data not shown), Pmt4 activity was significantly higher (Fig. 4B). In contrast, Pmt4 was almost inactive with Triton X-100 as detergent, whereas Pmt1-Pmt2 activity was characterized by an optimal curve with the highest [3H]mannose transfer at ∼0.12% Triton X-100 (∼10 × critical micelle concentration; Fig. 4C).</p><!><p>Substrate preferences of Pmt4 and Pmt1-Pmt2 complexes</p><p>In vitro mannosyltransferase activity was determined as detailed under "Experimental Procedures." Mean ± S.D. values of three replicates are shown. Substrate specificities of endogenous Pmt4 and Pmt1-Pmt2 complexes in comparison to SEY6210 (WT) were determined using membranes from pmt1 (pmt1/Pmt4) and pmt4 (pmt4/Pmt1/2) deletion strains, respectively. Assays were performed using different biotinylated acceptor peptides (bio-YATAV, 401–420-bio, and 418–440-bio) and standard reaction conditions with β-OTG as detergent. The average Dol-P-[3H]Man input was 49,936 dpm per reaction for 401–420-bio and 418–440-bio peptides and 39,323 dpm per reaction for the bio-YATAV peptide.</p><p>Detergent requirements of Pmt4 and Pmt1–2 complexes. In vitro O-mannosyltransferase activity was determined as detailed under "Experimental Procedures." Mean ± S.D. values of three replicates are shown. A, amino acid sequence of peptide 401–420-bio. B and C, dependence of mannosyltransferase activity on detergent concentrations for β-OTG (B) or Triton X-100 (C) is shown for Pmt4 (enzyme source: membranes from a pmt1 deletion strain; pmt1/Pmt4) and for Pmt1-Pmt2 complexes (enzyme source: membranes from a pmt4 deletion strain; pmt4/Pmt1/2) using the 401–420-bio peptide. Detergent concentrations are indicated as a function of the critical micelle concentration (cmc). The average Dol-P-[3H]Man input was 23,357 (B) and 33,225 dpm (C) per reaction.</p><!><p>Because both Pmt4 and Pmt1-Pmt2 activities could be monitored in the presence of β-OTG (Fig. 4B), in vitro substrate preferences were further determined in standard reactions including this detergent at 1.5 × critical micelle concentration (detailed under "Experimental Procedures"). Even with β-OTG, peptide bio-YATAV was only mannosylated by Pmt1-Pmt2 (Table 1). In line with previous observations, Pmt1-Pmt2 in vitro activity was enhanced ∼1.5-fold in the absence of Pmt4 (Table 1) (32). In contrast to bio-YATAV, the αDG-derived acceptor peptides 401–420-bio and 418–440-bio (29) served as mannosyl acceptors for Pmt4 and Pmt1-Pmt2, although to differing degrees (Table 1). Peptide 401–420-bio (four putative O-mannosyl acceptor sites) was preferentially mannosylated by Pmt4 (∼70% of the mannosyl transfer activity in wild-type membranes; Table 1), whereas the poor POMT substrate 418–440-bio (ten putative O-mannosyl acceptor sites; 29), which was also a weak substrate for yeast PMTs, was favored by Pmt1-Pmt2 (∼57% of the mannosyl transfer activity in wild-type membranes; Table 1).</p><!><p>Our analysis revealed that in vitro properties of Pmt4 resemble those from human POMTs, but show some distinct features when compared with yeast Pmt1-Pmt2. To further explore similarities in acceptor preferences between POMTs and yeast Pmt4, in vitro O-mannosylation of peptide 401–420-bio-derived variants in which the four putative mannosylation sites were individually exchanged to alanine were analyzed (Fig. 5A). It was previously reported for the human POMT complex that changing Thr-404, Thr-406, and Thr-414 to Ala greatly reduced acceptor efficiency of peptide 401–420 (29). As shown in Fig. 5B, Pmt4-based transfer of [3H]mannose to peptides T404A, T406A, T414A, and T418A was significantly decreased with T414A showing the most pronounced effect (∼2% acceptor efficiency) (Fig. 5B).</p><!><p>Mannosyl acceptor peptide preferences of Pmt4. In vitro O-mannosyltransferase activity was determined as described under "Experimental Procedures." Mean ± S.D. values of three replicates are shown. A, peptide sequences of the mannosyl acceptors 418–440-bio, 401–420-bio, and Thr to Ala mutations thereof. B, mannosyltransferase activity of endogenous Pmt4 was measured using crude membranes from a pmt1 deletion mutant (pmt1/Pmt4). Standard reaction conditions were applied for the indicated acceptor peptides. Relative activities are displayed with respect to the peptide 401–420-bio for which activity was set to 100%. The average Dol-P-[3H]Man input was 33,314 dpm/reaction.</p><!><p>Our experiments showed that in vitro, Pmt4 and POMTs recognize the same αDG-derived acceptor substrates and have similar detergent requirements. These findings prompted us to analyze whether dystroglycanopathy-associated mutations in human POMT1 also affect enzymatic activity of yeast Pmt4. We chose the point mutations G76R and V428D that had originally been detected in POMT1 of WWS patients (13), and created the corresponding Pmt4 mutants I112R and I435D. The exchanged amino acids are not highly conserved between POMT1 and Pmt4, however, that position is never occupied by a charged residue (Fig. 1). According to the topology model of PMTs (18), mutation G76R/I112R locates to the conserved loop1 region (Fig. 1). In Pmt1, this loop has been recently implicated in acceptor substrate binding (19). Mutation V428D/I435D is situated within a moderately conserved stretch of the second MIR motif within loop5 (Fig. 1). As a control, Pmt4 Ile-435 was changed to valine (I435V), which mimics the wild-type POMT1 allele (Fig. 1).</p><p>Under standard conditions with GST-αDG as mannosyl acceptor substrate, mannosyltransferase activity of the different enzyme variants was analyzed. Crude membranes were hence prepared from a pmt4 null mutant expressing FLAG-tagged versions of either wild-type or mutant Pmt4. Western blotting analysis revealed that the steady-state levels of Pmt4FLAG and mutants I112R, I435D, and I435V did not significantly differ from each other (Fig. 6B). In these preparations, mannosyltransferase activity of mutant I112R was highly reduced (by ∼97.5% when compared with the quantifiable activity of the wild-type enzyme). Enzymatic activity of mutant I435D could not be detected within the limits of the assay, whereas mutant I435V was at wild-type level (Fig. 6A).</p><!><p>In vitro mannosyltransferase activity of Pmt4 mutant proteins. A, in vitro activities of Pmt4 and the variants thereof. In vitro mannosyltransferase activity was determined as specified under "Experimental Procedures." Mean ± S.D. values of at least three independent experiments are shown as relative activities referring to Pmt4FLAG for which activity was set to 100%. Crude membranes from pmt4 null mutants expressing pJK4-B1 (Pmt4), pMS1 (I112R), pMS2 (I435D), or pDB6 (I435V) were used as enzyme source. The average Dol-P-[3H]Man input was 33,942 dpm/reaction for Pmt4, I112R, and I435D or 29,758 dpm/reaction for I435V. Maximal activities of about 7,000 dpm could be measured for wild-type Pmt4 with background values of around 100 dpm when no acceptor was added. Thus, activities below 1.5% of WT could not be evaluated. B, Western blotting analysis based quantification of Pmt4 variants in the membrane preparations which were used in A. Blots were probed with anti-Pmt4 and anti-Sec61 antibodies. Signal intensities were detected and analyzed as outlined under "Experimental Procedures." The indicated numbers represent the relative Pmt4 content (ratio Pmt4:Sec61) with reference to wild-type Pmt4, which was set as 1.</p><!><p>We further examined whether the observed loss of the in vitro activity of mutants I112R and I435D also reflects the in vivo situation. An indicator of yeast Pmt4 functionality in vivo is the synthetic temperature sensitivity of the double deletion strain pmt1pmt4. This strain fails to grow at 37 °C unless a functional variant of Pmt4 is expressed (8). As shown in Fig. 7A, pmt1pmt4 transformed with either the empty vector or plasmids expressing FLAG-tagged Pmt4 and variants thereof (the inactive Pmt4 mutant R142E (22) served as negative control) were viable at 25 °C but only wild-type Pmt4 restored thermotolerance of the pmt1pmt4 strain (Fig. 7A), indicating that Pmt4 mutants I112R and I435D have indeed very low or no mannosyltransferase activity.</p><!><p>Functionality of Pmt4 mutants in vivo. A, complementation of the temperature-sensitive phenotype of strain pmt1pmt4. The double deletion strain pmt1pmt4 was transformed with plasmids pRS423 (vector), pJK4-B1 (Pmt4), pMS1 (I112R), pMS2 (I435D), and pVG45 (R142E). Strains were grown at 25 °C (left) and 37 °C (right) for 3 days. B, Pmt4-dependent glycosylation of the cell wall protein Ccw5. Yeast mutant pmt4 expressing HA-tagged Ccw5 from plasmid pCCW5-HA was transformed with the plasmids pRS423 (vector), pJK4-B1 (Pmt4), pMS1 (I112R), or pMS2 (I435D). Ccw5HA was isolated from the indicated strains as described under "Experimental Procedures." O-Mannosylated (dark gray arrowhead) and N-glycosylated (white arrowhead) isoforms were monitored based on their different electrophoretic mobility during SDS-PAGE. Western blotting which was probed with anti-HA antibodies. C, formation of homomeric Pmt4 complexes. Protein extracts from wild-type strain SEY6210 transformed with pRS423 (vector) or plasmids encoding FLAG-tagged Pmt4 variants (pJK4-B1 (Pmt4), pMS1 (I112R), and pMS2 (I435D)) were prepared and immunoprecipitation using anti-FLAG antibodies was performed as outlined under "Experimental Procedures." Immunoprecipitates (IP) and aliquots of the corresponding input material were analyzed by SDS-PAGE and Western blotting analysis using anti-Pmt4 antibodies. FLAG-tagged Pmt4 variants and endogenous Pmt4 are highlighted with light gray and black arrowheads, respectively.</p><!><p>In addition, we took advantage of the fact that the pmt4 null mutant strain displays distinct molecular features, e.g. aberrant glycosylation of the cell wall protein Ccw5 (33). This protein harbors an N-glycosylation sequon (114NAT116) situated in a region that is O-mannosylated by Pmt4 in the wild-type, whereby N-glycosylation of Asn-114 is prevented. Loss of O-mannosylation by Pmt4, however, results in the addition of an N-linked glycan at that position and, as a consequence, a shift of the apparent molecular mass of the mature Ccw5 protein from 40 to ∼100–250 kDa (33) (Fig. 7B). Ccw5 glycosylation as reflected by its electrophoretic mobility can therefore serve as an indicator for Pmt4 functionality in vivo. We hence analyzed the glycosylation status of Ccw5 isolated from cell walls of a pmt4 deletion strain expressing wild-type or mutant Pmt4 by Western blotting. As evident from Fig. 7B, the glycosylation pattern of Ccw5 isolated from mutants I112R and I435D and from the pmt4 null mutant are highly similar (∼100–250 kDa) demonstrating that enzyme function in vivo is largely lost in both mutants. In agreement with the in vitro data, a small amount of the 40-kDa form of Ccw5 could be detected in the I112R mutant, pointing to a minor residual enzymatic activity of this Pmt4 variant.</p><p>Assembly into homomeric complexes is a prerequisite for Pmt4 activity (22). To test whether the loss of the enzymatic activity of mutants I112R and I435D resulted from deficient complex formation, we performed co-immunoprecipitation experiments. For that purpose, FLAG-tagged Pmt4 variants were expressed in wild-type yeast containing endogenous Pmt4. Proteins were co-precipitated by anti-FLAG antibodies and analyzed by Western blottings which were probed with polyclonal anti-Pmt4 antibodies (Fig. 7C). Endogenous Pmt4 co-purified with mutant proteins I112R and I435D showing that complex formation is not grossly affected by these amino acid substitutions even though minor effects cannot be ruled out entirely.</p><!><p>Although assays for the monitoring of yeast PMT activity in vitro have been described already in the early 1970s (34), it became clear, along with the discovery and characterization of the corresponding enzymes, that these assays, which use small peptide acceptors, detect only a specific subset of the PMT enzymes (32). The reason why Pmt1-Pmt2 complexes act on a variety of short peptides such as the pentapeptide YATAV, whereas Pmt4 as well as the mammalian POMT1-POMT2 complex do not, is unclear to date. The lack of an adequate substrate has hindered attempts to develop an in vitro assay system for Pmt4 activity. Here we show that the mammalian O-mannosylation substrate αDG is suitable for the specific detection of yeast Pmt4 activity. This finding allows monitoring enzymatic activity of homomeric Pmt4 complexes in vitro.</p><p>Our analyses revealed that yeast Pmt1-Pmt2, Pmt4, and mammalian POMT1-POMT2 enzymes show various similarities in vitro such as pH and temperature optima, stimulation by divalent cations, and resistance to EDTA (Fig. 2) (17, 31, 35). However, their detergent requirements differ substantially. Pmt1-Pmt2 complexes are active in vitro in a wide range of detergents (Fig. 4) (17, 35), whereas Pmt4 and mammalian POMT in vitro activities largely depend on β-OTG even though numerous alternative detergents have been tested (Fig. 4 and data not shown) (27, 31). In addition, unlike Pmt1-Pmt2, Pmt4 and POMTs mannosylate GST-αDG in vitro (Fig. 2A) (31). Why the mucin domain of αDG does not serve as an acceptor substrate of Pmt1-Pmt2 remains unclear. This issue is especially puzzling because αDG-derived peptides are in vitro mannosyl acceptors for all PMTs, albeit with different preferences (Table 1). One possible explanation could be that in the 171-amino acid long αDG domain, which is rich in proline residues (roughly 20% overall content) and harbors a Ser/Thr content of nearly 30%, Pmt1-Pmt2 acceptor sites are masked. A putatively unstructured polypeptide containing numerous acceptor sites, however, might be sufficient to trigger O-mannosylation by Pmt4, at least in vitro. Recently, the yeast O-mannose glycoproteome revealed general characteristics of O-mannosylation sites but sequence features suggestive of a glycosylation motif did not become evident. Yeast O-mannosyl glycans are enriched in unstructured regions and β-strand folds that might be attributed to the discrete substrate and/or glycosylation site specificities of the different yeast PMT family members (25).</p><p>Yeast Pmt4 and POMTs do not only act on the same αDG-derived protein substrate, they also show similar mannosyl acceptor preferences. A previous study by Manya and co-workers (29) identified the αDG-derived peptide 401–420 as a POMT in vitro substrate, which is most frequently mannosylated at Thr-414 (T414A substitution decreased the mannosyl transfer by 93%), and mannosylation of this Thr residue most likely facilitates subsequent modification by the POMT complex. In good agreement with these data, peptide 401–420-bio is a preferred in vitro substrate of yeast Pmt4 (Table 1, Figs. 4 and 5). Although O-mannosylation site occupancy has not been directly analyzed by mass spectrometry, individual substitutions of the Thr residues in this peptide for Ala indicate that all four hydroxy amino acids serve as acceptor sites of Pmt4, and that Thr-414 is especially crucial for acceptor efficiency (decrease of mannosyl transfer by 98%, Fig. 5B). Furthermore, mannosylation effectiveness of Thr-404 and Thr-406 is comparable between Pmt4 (Fig. 5B) and POMTs (∼20–30%; 29). Nevertheless, Thr-418 was only poorly mannosylated by POMTs (29) but serves as a mannosyl acceptor of Pmt4 (Fig. 5B), indicating also distinct differences in acceptor selection. In vivo Pmt4 favors membrane-bound substrates (24). Although in contrast to the Pmt4 properties in vitro, this feature is in agreement with the notion that virtually all described in vivo substrate proteins of the mammalian POMTs are membrane-associated (7 and references therein). Consistently, the acceptor selection of the mammalian POMT complex differs in vivo and in vitro. Although the αDG-derived peptide 401–420 serves as a POMT acceptor substrate in vitro, in vivo domains upstream of the actual acceptor sites are additionally required for O-mannosylation (29, 30).</p><p>Taken together, our findings demonstrate functional similarities between Pmt4 and the mammalian POMTs and distinguish them from the fungal Pmt1-Pmt2 family members that have distinct detergent requirements and acceptor substrates in vitro (Fig. 4, Table 1), and mannosylate both, soluble and membrane proteins in vivo (reviewed in Ref. 1). Growing evidence suggests the involvement of yeast Pmt1-Pmt2, but not Pmt4, in a novel ER quality control system (reviewed in Ref. 2), further emphasizing their differences.</p><p>Pmt4 is the closest phylogenetic relative of mammalian POMT1, which acts in a heteromeric complex with POMT2 (27). Akasaka-Manya and co-workers (28) recently suggested that POMT1 and POMT2 might fulfill discrete functions, because mutations of conserved amino acids differentially affect enzymatic activity of the complex. Exchanging a single amino acid of the conserved Asp-Glu (DE) motif in loop1 (Fig. 1) with Ala in POMT1 (E44A) resulted in the loss of in vitro activity of the complex, whereas the corresponding mutation in POMT2 (E86A) only slightly affected mannosyltransferase activity. Intriguingly, the corresponding mutation also rendered yeast Pmt4 (E81A) inactive but only moderately affected Pmt1 (E78A) (19). Although functional similarities of mammalian and yeast PMT4 family members exist, heterologous expression of human POMT1 and POMT2 separately or in combination did not result in the complementation of the temperature-sensitive phenotype of S. cerevisiae double mutant pmt1pmt4 (data not shown). Furthermore, POMT1 and POMT2 did not rescue the lethality of Schizosaccharomyces pombe mutant oma2Δ (9).7 This may be for several reasons including number and nature of the substrate proteins and/or impaired association of human POMTs with the yeast Sec61 translocon, which has been recently demonstrated for S. cerevisiae PMTs (36). The findings are in line with our previous observations that even between S. pombe and S. cerevisiae PMTs are only partially functional interchangeably (9).7</p><p>POMT mutations are frequently associated with congenital muscular dystrophies with widely varying clinical phenotype (13, 37–39). We and others showed that the degree of severity of the disease of patients with POMT1 mutations is inversely proportional to the POMT in vivo and in vitro activity (16, 40). Here, we mimicked two WWS-associated POMT1 amino acid exchanges in yeast Pmt4. The Pmt4 mutants I112R and I435D presented to a large extent proteolytically stable and properly folded as judged from complex formation (Figs. 6B and 7C). Yeast mutants proved catalytically inactive in vitro and in vivo (Figs. 6A and 7), which is consistent with the severe phenotypes of the patients in which the corresponding POMT1 mutations had been discovered and the highly reduced POMT in vitro activity of fibroblasts derived from a WWS patient carrying the homozygous mutation G76R (13, 16).</p><p>Without structural models it is difficult to judge how the analyzed mutations affect PMT/POMT activity. The loop5 domain is highly homologous to the ER-resident soluble stromal cell-derived factor 2 (SDF2) (21). Recently, we resolved the three-dimensional crystal structure of Arabidopsis thaliana SDF2 at 1.95-Å resolution that revealed the typical β-trefoil fold and consists of 12 β-strands and three 310 helices forming a globular barrel (41). Conserved leucine and valine residues form hydrophobic layers of the barrel with a crucial role in maintaining the β-trefoil structure (41). To further address the impact of the analyzed Pmt4 I435D mutation, based on AtSDF2 we generated structural models of the yeast Pmt4-, Pmt1-, and human POMT1-loop5 domains (supplemental Fig. S1). Like in AtSDF2, the β-trefoil-fold of loop5 is made up of three structural repeats that correspond to the three MIR motifs, giving rise to a pseudo 3-fold symmetry. Pmt4 Ile-435 and POMT1 Val-428 are part of a hydrophobic layer at the top of the bottom layer of the barrel (supplemental Fig. S1). Thus, mutations Pmt4 I435D and POMT1 V428D almost certainly interfere with the general structure of the respective loop5 domain. Similarly, exchange of a conserved leucine residue of the hydrophobic layer of Pmt1-loop5 (L408A) also affects enzymatic function (supplemental Fig. S1) (20), suggesting that the β-trefoil-fold of loop5 is a key feature of all PMTs/POMTs.</p><p>In summary, we set up a robust in vitro assay that allows the quantitative determination of Pmt4 activity. With this tool in hand, studies elucidating the fundamentals of protein O-mannosyltransferases are now simplified due to the amenability of a homomeric protein complex. In particular, using Pmt4 as a model will greatly facilitate structural analyses such as three-dimensional crystallization.</p><!><p>The S. cerevisiae disruptants pmt4 (pmt4::TRP1) (32), pmt1 (pmt1::HIS3) (42), and pmt1pmt4 (pmt1::URA3, pmt4::TRP1) (8) are descendants of the reference strain SEY6210 (MATα, his3-Δ200, leu2–3, −112, lys2–801, trp1-Δ901, ura3–52, suc2-Δ9) (43). Yeasts were grown under standard conditions and transformed according to Hill et al. (44) with plasmids pJK4-B1 (PMT4FLAG), pVG45 (PMT4-R142EFLAG) (22), and pCCW5-HA(33) and the plasmids described below. PMT4 point mutations were introduced into pJK4-B1 via site-directed mutagenesis using recombinant PCR (45). Sequences of the oligonucleotides used in this study are available upon request. DNA constructs were processed using standard procedures and routinely verified by sequence analysis.</p><p>To create plasmid pMS1 (PMT4-I112RFLAG), a PCR fragment generated with the mutagenic primer pair 510/511 in combination with the outer primers vg28 and 512, was subcloned into pJK4-B1 via PaeI and Van91I. Plasmid pMS2 (PMT4-I435DFLAG) was generated with the mutagenic primer pair 508/509 and outer primers vg28 and vg27. The resulting fragment was subcloned into pJK4-B1 via Van91I and KspAI. Plasmid pDB6 (PMT4-I435VFLAG) was assembled by generating a PCR fragment with the mutagenic primer pair 2485/2486 and the outer primers vg28 and 512 and subcloning the resulting fragment into pJK4-B1 linearized with NcoI and EcoNI via homologous recombination in yeast.</p><!><p>Immunoprecipitation experiments were performed as previously described (22).</p><!><p>Crude yeast membranes were prepared essentially as previously described (20) with minor modifications. Briefly, exponentially growing yeast cells were harvested by centrifugation, washed once with water, and once with 50 mm Tris-HCl, pH 7.4, 5 mm MgCl2. The pellet was resuspended in the same buffer plus 1 mm PMSF, 1 mm benzamidine, 0.25 mm 1-chloro-3-tosylamido-7-amino-2-heptanone, 50 μg/ml of l-1-tosylamido-2-phenylethyl chloromethyl ketone, 10 μg/ml of antipain, 1 μg/ml of leupeptin, and 1 μg/ml of pepstatin. An equal volume of glass beads was added and cells were lysed in a Hybaid RiboLyser (4 × 25 s at level 4.5 with 1-min intervals on ice) at 4 °C. The bottom of the tube was punctured, and the lysate was collected into a new tube. Cell debris was removed by two successive centrifugations (5 min at 1,500 × g, 4 °C). The supernatant was then centrifuged 1 h at 20,000 × g at 4 °C. Pelleted membranes were resuspended in 20 mm Tris-HCl, pH 8.0, 10 mm EDTA, 15% (v/v) glycerol plus protease inhibitors (see above), frozen in liquid nitrogen, and stored at −80 °C. Cell wall extracts were prepared as described in Ref. 33.</p><!><p>GST-αDG and GST were purified from Escherichia coli BL21(DE3) cells. The corresponding expression plasmids and the purification procedure are described in Ref. 46.</p><!><p>401–420-bio and bio-YATAV were purchased from Biopolymers Thermo Scientific, 418–440-bio and Thr to Ala substitutions of 401–420-bio from Intavis AG.</p><!><p>Pmt4 in vitro activity measurements were based on the previously published protocol by Manya and co-workers (31) with minor modifications. Standard reactions contained 20,000–40,000 dpm (150–300 fmol) of [3H]mannose-labeled Dol-P-Man (mannosylphosphoryldolichol-95, American Radiolabled Chemicals, 60 Ci/mmol), 4 μg of acceptor protein (GST-αDG or GST), 30 μg of total protein of crude membrane preparations, 0.45% (w/v) β-OTG, 2 mm β-mercaptoethanol, 10 mm EDTA, and 20 mm Tris-HCl, pH 7.5, in a total reaction volume of 20 μl. Prior to the reaction, Dol-P-Man was dried in a glass vial under a stream of nitrogen and resuspended in the reaction mixture (containing everything but the proteins) by extensive vortexing and ultrasonication. After addition of the acceptor protein, the reaction was started by the addition of membranes, incubated 15 min at 20 °C, and then stopped with 200 μl of ice-cold PBS plus 1% Triton X-100. After centrifugation for 10 min at 20,000 × g at 4 °C, the supernatant was incubated with GSH-Sepharose beads (GE Healthcare) for 90 min at 4 °C. The beads were washed two times with PBS plus 1% Triton X-100 and two times with PBS. Incorporated radioactivity was measured by liquid scintillation counting.</p><!><p>Standard reactions were performed as described above. Instead of GST-αDG, synthetic peptides (200 μm) carrying a C-terminal Biotin tag were included as mannosyl acceptor. After stop of the reaction, mixtures were centrifuged for 10 min at 20,000 × g at 4 °C. The supernatant was incubated with 20 μl of slurry of High Capacity Neutravidin®-agarose (Thermo Scientific) for 1 h at 4 °C. Beads were washed two times with PBS plus 1% Triton X-100 and two times with PBS. Incorporated radioactivity was measured by liquid scintillation counting.</p><!><p>Protein samples were resolved by SDS-PAGE on 8% polyacrylamide gels and transferred to nitrocellulose. Monoclonal mouse anti-FLAG (M2, Sigma) antibodies were used at a dilution of 1:5,000. Polyclonal rabbit Pmt4- (22) and Sec61-directed (47) antibodies were used at a dilution of 1:2,500 and 1:1,000, respectively. Blots were incubated with horseradish peroxidase-conjugated anti-mouse or anti-rabbit secondary antibodies (Sigma). Protein-antibody complexes were visualized by enhanced chemiluminescence and quantified with the ImageQuant LAS 4000 imaging system (GE Healthcare).</p><!><p>D. B. designed, performed, and analyzed the experiments shown in Figs. 4–6, supplemental Fig. S1, and Table 1. J. E. and T. J. designed, performed, and analyzed the experiments shown in Figs. 2, 3, and 6. M. S. performed and analyzed the experiments shown in Fig. 7. J. E. designed Fig. 1. S. S., J. E., and D. B. wrote the manuscript. All authors reviewed the results and approved the final version of the manuscript. S. S. conceived and coordinated the study.</p><!><p>This work was supported by the Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 638, project A18. The authors declare that they have no conflicts of interest with the contents of this article.</p><p>This article contains supplemental Fig. S1.</p><p>T. Willer and S. Strahl, unpublished data.</p><p>endoplasmic reticulum</p><p>dolichol monophosphate-activated mannose</p><p>α-dystroglycan</p><p>protein O-mannosyltransferase</p><p>Walker-Warburg syndrome</p><p>β-octylthioglucoside.</p>

PubMed Open Access

In silico screening of naturally occurring coumarin derivatives for the inhibition of the main protease of SARS-CoV-2

The dissemination of a novel corona virus, SARS-CoV-2, through rapid human to human transmission has led to a global health emergency. The lack of a vaccine or medication for effective treatment of this disease has made it imperative for developing novel drug discovery approaches. Repurposing of drugs is one such method currently being used to tackle the viral infection. The genome of SARS-CoV-2 replicates due to the functioning of a main protease called M pro . By targeting the active site of M pro with potential inhibitors, this could prevent viral replication from taking place. Blind docking technique was used to investigate the interactions between 29 naturally occurring coumarin compounds and SARS-CoV-2 main protease, M pro , out of which 17 coumarin compounds were seen to bind to the active site through the interaction with the catalytic dyad, His41 and Cys145, along with other neighbouring residues. On comparing the ΔG values of the coumarins bound to the active site of M pro , corymbocoumarin belonging to the class pyranocoumarins, methylgalbanate belonging to the class simple coumarins and heraclenol belonging to the class furanocoumarins, displayed best binding efficiency and could be considered as potential M pro protease inhibitors. Preliminary screening of these naturally occurring coumarin compounds as potential SARS-CoV-2 replication inhibitors acts as a stepping stone for further in vitro and in vivo experimental investigation and analytical validation.

in_silico_screening_of_naturally_occurring_coumarin_derivatives_for_the_inhibition_of_the_main_prote

Introduction<!>Materials and Methods<!>Preparation of M pro for docking studies<!>Preparation of ligands<!>Molecular docking and visualization<!>Drug-likeliness studies<!>Results and Discussion<!>Docking studies for Furanocoumarins with SARS-CoV-2 M pro<!>Docking studies for Pyranocoumarins with SARS-CoV-2 M pro<!>Conclusion

<p>The year 2020 is continuing to be a year where extreme pressure is being put on the health care system of nations all over with world with the outbreak of a novel corona virus, SARS-CoV-2. This novel corona virus causing SARS-like acute respiratory syndrome was first reported in late December 2019 in the Chinese city of Wuhan, Hubei Province (Wang et al., 2020). Because this disease caused by SARS-CoV-2 was discovered in the end of 2019, the World Health Organisation (WHO) on 12 th February 2020 named this disease as COVID-19 which stands for coronavirus disease 2019 (https://www.who.int/emergencies/diseases/novelcoronavirus-2019/situation-reports). With the rapid spread of the COVID-19 disease beyond the borders of China, leading to a sustained risk of further global spread, on 11 th March 2020 the WHO declared COVID-19 as a global pandemic (Cucinotta & Vanelli, 2020).</p><p>Despite various actions being taken by global governments to contain this virus, the total number of global COVID-19 positive cases stands at 31,75,207 including 2,24,172 deaths as of 1 st May 2020, reported by WHO (https://www.who.int/emergencies/diseases/novelcoronavirus-2019/situation-reports).</p><p>The SARS-CoV-2 still remains a mystery in terms of its origin, its transmission to human beings and most importantly, its treatment. The ease at which human to human transmission of this virus takes place has put global nations to an economic and social halt. Despite the collaborative efforts of researchers all over the world, a suitable and effective drug candidate or vaccine to treat patients or to mitigate virus transmission is still unavailable. With a time constraint to combat the SARS-CoV-2 virus, efforts are now being put into repurposing of existing drugs (Muralidharan et al., 2020). Anti-malarial drugs such as hydroxychloroquine and chloroquine have been approved by the FDA to be tested against COVID-19 (Dong et al., 2020). To name a few, anti-viral drug such as flavilavir is being used in China to treat patients with COVID-19 symptoms (Li et al., 2020). Darunavir an anti-retroviral protease inhibitor is reported to inhibit SARS-CoV-2 replication (Dong et al., 2020). Infected patients treated with anti-retroviral protease inhibitors lopinavir/ritonavir have been associated with improvements (Chan et al., 2003;Chu et al., 2004). Now many drugs are being repurposed, through preliminary screening, in order to study their anti-SARS-CoV-2 replication efficiency.</p><p>Replication in SARS-CoV-2 is aided by a chymotrypsin-fold proteinase called main protease (M pro or 3CL pro ) (Anand et al., 2003;Boopathi et al., 2020). M pro plays an indispensible role in processing polyproteins at 11 cleavage sites in order to generate critical non-structural proteins such as helicase, methyl transferase and RNA dependent RNA polymerase, all important for viral replication (Cui et al., 2019). Thus M pro is considered an attractive drug target to block its proteolytic action which will ultimately lead to an inhibition of viral replication (Khan et al., 2020;Morse et al., 2020;Wu Y-S, 2006). The crystal structure of M pro has been published and made available since February 2020. His41 and Cys145, the catalytic dyad, is found at the junction of domain I and Domain II of M pro and is responsible for its catalytic activity (Zhang et al., 2020). The crystal structure of M pro used in this study (PDB ID. 6Y84) is depicted in Figure 1 in which active site residues have been highlighted. Preliminary screening of drugs using in silico methods have proved to be useful in meeting challenges of discovering potential candidates. Along with synthetic and semi synthetic drugs being used to target many viral proteins to inhibit viral spread, natural compounds are also being screened in a similar manner. Natural compounds are structurally and chemically diverse, with least side effects and toxicity, making them valuable contenders in the fight against COVID-19. Reports suggest that 45% of the best selling drugs have been derived from natural products or their derivatives (Lahlou, 2013). With this knowledge, naturally occurring coumarin compounds were selected and screened out as potential M pro inhibitors using molecular docking studies. Various substituents and conjugates present on the central bicyclic structure of coumarins bestow on it many biological effects and potential therapeutic and pharmacological properties. In addition, their stability, low side effects and toxicity, oral bioavailability and broad spectrum have made coumarins find a place in the medicinal field (Wang H, 2009). Numerous studies have reported anti-viral activity of naturally occurring coumarins inhibiting the functioning of viral proteins such as proteases, integrase, reverse transcriptase, DNA polymerase and also in preventing viral entry (Hassan et al., 2016;Mishra et al., 2020). For example, coumarins such as oxypeucedanin, heraclenol, pranferol, and xanthotoxin has been known to exhibit anti-HIV activity (Zhou et al., 2000). Structurally similar coumarin compounds, psoralen, saxalin and bergapten have been reported to prevent HIV replication (Shikishima et al., 2001). Also, mesuol and isomesuol have been found to suppress HIV replication in Jurkat T cells (Márquez et al., 2005). Rutamarin, a naturally occurring furanocoumarin, and kellerin, a sesquiterpene coumarin, were reported to be anti-HSV agents (Ghannadi et al., 2014;Xu et al., 2014) In this study, molecular docking method was used to study the interactions between naturally occurring coumarin compounds and the SARS-CoV-2 main protease, M pro . We reported that a total of 17 coumarin compounds could interact with the catalaytic dyad in the active site of M pro , namely, mesuol (4441487), pabulenol (446229), pranferol (140461), saxalin (158533), seselin (61531) xanthotoxin (3971), heraclenol (66005), kellerin (40580807), oxypeucedanin (141075), rutamarin (25094), anomalin (4477595), esculin (4444765), methylgalbanate (5428912), isofraxidin (4477107), osthole (9811) sphondin (97199) and corymbocoumarin (969471). The numbers indicated in bracket for the first sixteen coumarin compounds depicts their ChemSpider ID and the number for the last coumarin, corymbocoumarin, signifies its PubChem ID.</p><!><p>Finding potential lead compounds using in silico methods can help to reduce the time and resources involved in the drug development process. By conducting molecular docking studies, potential drug candidates can be screen out from a large number of compounds by studying their interactions with biological molecules.</p><p>In this study, blind docking analysis was carried out between some naturally occurring coumarins and the main protease, M pro , of SARS-CoV-2.</p><!><p>The crystal structure of the main protease M pro (PDB ID. 6Y84) of SARS-CoV-2 was retrieved from RCSB Protein Data Bank (PDB) (Owen, 2020). The main protease obtained consisted of a dimer of two homologous amino acid chain, chain A and chain B, in which chain A was used as a target for molecular docking. Before docking studies was carried out the target molecule was prepared which involved the removal of water molecules using PyMOL (Schrodinger, 2017).</p><!><p>Sixteen of the coumarin ligands were downloaded in the form of 3D coordinates from Chemspider (www.chemspider.com) and one coumarin ligand from PubChem (www.pubchem.ncbi.nlm.nih.gov) in .mol format. ArgusLab using PM3 methods was then used to conduct energy optimization of the ligands.</p><!><p>The final M pro PBD file and the geometry optimised ligands structures were uploaded onto the SWISSDOCK server (http://www.swissdock.ch/docking) for molecular docking analysis.</p><p>SwissDock is based on the docking software EADock DSS, whose algorithm consists of generating a number of binding modes either through local docking or blind docking, estimates CHARM (Chemistry at HARvard Macromolecular Mechanics)energy on a grid and producing clustered binding modes with most favourable binding energies (Grosdidier et al., 2011). Analysis is carried out on the docked pose which has the minimum fullfitness score.</p><p>Visualization of the results obtained from SWISSDOCK was done in a molecular visualization tool, UCFS Chimera (Pettersen et al., 2004). The dock poses and 2D interaction plots were then prepared using Discovery Studio Visualizer (BIOVIA., 2019). An online server (http://cib.cf.ocha.ac.jp/bitool/ASA/) was used to determine the changes in the accessible surface area of the main protease when bound to the ligand. Finally, APBS plugin of PyMOL was used to determine the electrostatic surface potential (Schrodinger, 2017).</p><!><p>Lipinski's rule of five dictates four parameters (molecular weight <500 Da, no of hydrogen bond donors should be less than 5, no. of hydrogen bond acceptors should be less than 10 and log P should not be greater than 5) in order for a compound to be considered a potential drug.</p><p>The coumarin ligands were uploaded on to SWISSADME (www.swissadme.ch/index.php) server to obtain the above mentioned five parameters of Lipinski's rule. The chemical structures, chemical formula and the Lipinski's rule of five parameters of the ligands have been provided in Table S1 (Supplementary Information).</p><!><p>In order to tackle the spread of COVID-19 infection, a promising mechanism would be to inhibit SARS-CoV-2 main protease, M pro , which is critical for viral replication. Keeping this in mind, SWISSDOCK, an online molecular docking tool, was used to study the molecular interactions between naturally occurring coumarin compounds and the catalytic residues of the active site of SARS-CoV-2's main protease, M pro . The SWISSDOCK server also provided the binding modes of these ligands (coumarin compounds) which were listed according to their best fullfitness score. The minimum fullfitness score, for each ligand, with its corresponding minimum fullfitness score has been tabulated in Table 1. Mesuol, isofraxidin, methylgalbanate, esculin, osthole and kellerin, falling under the class of simple coumarins, were seen to bind to the active site of M pro as illustrated by their respective docked poses along with their corresponding 2D interaction plots (Figure 2). Mesuol, a reported anti-viral agent known to inhibit HIV replication (Marquez et al., 2005), interacted with one of the catalytic dyad residue, His41, in the active site of M pro , through π-alkyl interaction. Similar type of interaction was also seen between Cys44, Met49 of M pro and mesuol. Formation of a hydrogen bond between Glu166 and mesuol along with van der Waals interactions with other residues, including Cys145, was also depicted in the 2D interaction plot (Figure 2a). Isofraxidin was seen to interact with the active site of M pro through hydrogen bonding with His164, π-alkyl interaction with Met165, carbon hydrogen bond with Gln189, π-donor hydrogen bond with Glu166 and non-covalent van der Waals interaction with other surrounding residues (Figure 2b). A stable interaction was also noted through the formation of a hydrogen bond between the naturally occurring coumarin, osthole, and Gly143 of M pro . The 2D interaction plot also displayed other interactions such as alkyl hydrophobic interactions with Cys145 and His163 and π-alkyl interactions with Cys145 and Met49 (Figure 2c). Methylgalbanate formed numerous non covalent interactions with residues present in the active site of M pro such as alkyl hydrophobic interaction with Cys145, His41, His163, Met165 and Leu167, π-alkyl interaction with Cys145, carbon hydrogen bond with Asn142 and Glu166 and finally few van der Waal interactions with surrounding residues (Figure 2d). The catalytic dyad, His 41 and Cys145, of M pro interacts with esculin through carbon hydrogen bonding and π-sulphur interaction respectively. Esculin also formed hydrogen bonds with Glu166 and Gly143, carbon hydrogen bonds with Met165 and Asn142 and non-covalent van der Waals interaction with nearby residues of M pro (Figure 2e).</p><p>Kellerin, a coumarin with reported anti-viral activity, stabilised the active site of M pro through non covalent interactions such as carbon hydrogen bond with Asn142 and Met165, alkyl hydrophobic interactions with His41 and Cys145, π-alkyl interaction with Cys145 and van der Waals interaction with corresponding amino acids (Figure 2f). The estimated free energy of binding, represented by ΔG, which signifies the binding affinity of the above mentioned naturally occurring coumarin compounds towards M pro , was arranged in ascending order starting with methylgalbanate (-8.30 kcal/mol), kellerin (-8.18kcal/mol), esculin (-7.74 kcal/mol), mesuol (-7.38 kcal/mol), osthole (-7.24 kcal/mol) and finally isofraxidin (-7.00 kcal/mol). Figure 5a represents the electrostatic surface potential of the binding site of M pro along with all the six simple coumarin compounds.</p><!><p>The docked poses of minimum energy of eight furanocoumarin compounds used in this study along with their 2D interaction plots are represented in Figure 3. Rutamarin, with reported anti-Herpes Simplex Virus (HSV) properties (Xu et al., 2014), interacted with the a number of residues present in the active site of M pro through hydrogen bond formation with Ser46 and Glu166, alkyl hydrophobic interactions with His41, Met165 and Met49, π-alkyl interactions with Cys145 and His41, carbon hydrogen bond formation with Thr45 and van der Waals interaction with neighbouring residues (Figure 3a). Heraclenol formed hydrogen bonds with His41 and Ser144, carbon hydrogen bonds with Asn142, π-alkyl interactions with Cys145,</p><p>Met145 and Met49, π-donor hydrogen bond with Glu166 and van der Waals interaction with corresponding residues in order to stabilise the active site of M pro (Figure 3b). Non covalent interactions was also seen between oxypeucedanin and M pro active site residues through the formation of hydrogen bond with Gly143, carbon hydrogen bonds with His41, Cys44 and Asn142, alkyl hydrophobic interactions with Cys145, His41and Leu27, π-alkyl interaction with Met49 and van der Waals interaction with other residues as depicted in Figure 3c.</p><p>Saxalin stabilized the active site of M pro through carbon hydrogen bond formation with</p><p>Leu141 and His41, π-donor hydrogen bond formation with Glu166, π-sigma interaction with His41, alkyl hydrophobic interaction with His41 and Met49, π-alkyl interaction with Met165, π-sulphur interaction with Cys145 and van der Waals interaction with nearby residues (Figure 3d). A number of interactions were witnessed between pabulenol and the residues of the active site of M pro (Figure 3e). Hydrogen bonding of pabulenol with Leu141, π-sulphur interaction with Cys145, π-alkyl interaction with His163 and Met49 and carbon hydrogen bonding with Asn142. Van der Waals interactions with adjoining residues also helped to stabilise the interaction between pabulenol and M pro . Another furanocoumarin, pranferol interacted with active site M pro through the formation of a hydrogen bond with Glu166, πalkyl interactions with Pro168 and Met165, alkyl hydrophobic interaction with Pro168, carbon hydrogen bonding with Thr190 and Gln189 and van der Waals forces was also a very dominant interaction present (Figure 3f). Xanthotoxin underwent a number of non-covalent interactions to stabilize the active site of M pro . Residues located in the active site such as</p><p>Gly143 and Glu166 formed hydrogen bonds with xanthotoxin, Cys145 formed π-sulphur type of interaction, Met165 formed π-alkyl interaction, Glu166 formed a carbon hydrogen bond with xanthotoxin and van der Waals interactions among corresponding residues was also present (Figure 3g). On the other hand, sphondin stabilised the active site of M pro through the formation of carbon hydrogen bonds with His163, Leu141 and Asn142, π-alkyl interactions with Cys145, Met165 and Met49 and also through van der Waals interactions with adjoining amino acid residues located in the active site (Figure 3h). The ΔG values, from Table 1, of the above mentioned furanocoumarins were compared and it was seen that heraclenol (-8.20 kcal/mol) had the most negative ΔG value followed by rutamarin (-7.63 kcal/mol), pabulenol (-7.42 kcal/mol), peucedanin (-7.26 kcal/mol), pranferol (-7.16 kcal/mol), saxalin (-7.14 kcal/mol), sphondin (-6.94 kcal/mol) and lastly followed by xanthotoxin (-6.80 kcal/mol).</p><p>Finally, Figure 5b depicts the electrostatic surface potential of the binding site of M pro with the synchronous presence of the eight furanocoumarin compounds.</p><!><p>Three pyranocoumarins, seselin, anomalin and corymbocoumarin, were docked on the active site of SARS-CoV-2 M pro protease as illustrated in Figure 4 along with their 2-D interaction plot. Seselin interacted with the amino acid residues found in the active site of M pro through the formation of a hydrogen bond with Gly143, π-alkyl interaction with Cys145, His41 and Met165, alkyl hydrophobic interactions with His41, Cys145and Leu27 and finally van der Waals interactions were also seen to interact with corresponding residues (Figure 4a). Many con-covalent interactions between anomalin and M pro helped to stabilise this interaction.</p><p>Alkyl hydrophobic interactions with His41, Met49, Cys145, Met165 and Pro168, π-alkyl interaction with His41, π-anion interaction with Glu166 and van der Waals interaction with many surrounding residues were among such stabilising forces (Figure 4b).</p><p>Corymbocoumarin however formed stabilizing forces with M pro through carbon hydrogen bond formation with Met165, alkyl hydrophobic interaction with His41, Cys145, Met49, Pro168 and Met165, π-anion interaction with Glu166, π-lone pair interaction with Asn142, πalkyl interaction with His41 and van der Waals interaction with other residues as shown in Figure 4c. For the above three pyranocoumarins the ΔG value from table 1 were ranked with corymbocoumarin (-8.57 kcal/mol) displaying least ΔG value, followed by anomalin (-8.18 kcal/mol) and finally seselin with a ΔG value of -7.00 kcal/mol. The electrostatic surface potential of the binding site of M pro bound with all the three pyranocoumarins synchronously has been depicted in Figure 5c. Repurposing or repositioning of many therapeutic drugs is being carried out by many researchers and scientists all over the world as an effective means of treatment against COVID-19. The main protease of SARS-CoV-2, M pro , is a popular target for such studies as inhibition of this enzyme would lead to viral replication coming to a halt. In order to achieve this, blocking the amino acid residues located in the active site of M pro would be required. A recent study on α-ketoamide inhibitors used against SARS-CoV-2 main protease, M pro , revealed that His41, Met49, Gly143, Cys145, His163, His164, Glu166, Pro168, and Gln189 amino acids are the inhibitor binding site residues (Zhang et al., 2020) out of which His41 and Cys145 are known as the catalytic dyad (Jin et al., 2020).</p><p>Blind docking studies carried out on three classes of naturally occurring coumarin compounds against the M pro target protein revealed that all the seventeen coumarins which bound to the active site interacted in some manner or another with the above mentioned amino acid residues which suggests that they could impede the proper functioning of the protease. The distance of His41 and Cys145 of each coumarin along with the change in accessible area from the residues located in the active site of M pro has also been represented</p><p>in Table 2.</p><p>Table 2. The distance of the coumarin derivatives from the catalytic site residues His41 and Cys145 along with their ∆ASA. In the structural class of simple coumarins, methylgalbanate was seen to have the lowest energy score, -8.30kcal/mol, compared to the other coumarin compounds, falling within the same class. Under the structural class furanocoumarin, heraclenol had the most negative ΔG value, -8.20 kcal/mol. This is significant due to the fact that anti-viral properties of herclenol, particularly anti-HIV activity, have been reported in previous studies (Shikishima et al., 2001;Wu et al., 2001). And finally, corymbocoumarin (-8.57 kcal/mol) was ranked with the lowest energy score, among the pyranocoumarin class and also among the other two classes of coumarin compounds. Despite the ΔG values of control drugs such as ritonavir (-9.52 kcal/mol), and lopinavir (-9.00 kcal/mol) when docked with M pro (Das et al., 2020) was more negative than the above mentioned coumarins, however when compared with the ΔG of hydroxychloroquine (-7.75 kcal/mol) (Das et al., 2020), a potential drug undergoing assessment for COVID-19 treatment in patients, the coumarin compounds displayed better binding affinity due to their higher negative ΔG values. Results from ADME studies (Table S1) also revealed that majority of naturally occurring coumarins, including corymbocoumarin and heraclenol, were highly drug likely due to their non-violation of Lipinski's rule of five as compared to ritonavir (2 violations) and lopinavir (1 violation). Coumarin compounds, due to their natural origin, have many benefits as potential drug candidates due to their low toxicity and least side effects, problems that are faced on administration of most synthetic and semisynthetic drugs. With this, we can conclude by saying that heraclenol coumarin, in addition to being of natural origin, drug-likely and most importantly, having anti-viral properties, it also displayed a comparable binding energy value with that of methylgalbanate and corymbocoumarin. It could thus be considered a potential SARS-CoV-2 M pro inhibitor but further wet lab studies and investigation needs to be carried out in order to establish it protease inhibiting properties.</p><!><p>The rapid spread and transmission of COVID-19 infection across oceans and continents has created a worldwide scare with positive cases and deaths being reported everyday, globally.</p><p>The lack of effective and fool proof medication as well as broad spectrum inhibitors to control the viral infection through human transmission remains a challenge for researchers and scientists. At the same time, the development of new anti-viral agents followed by their clinical trials and approval is a time taking process and time is what we do not have. Without a vaccine to prevent COVID-19 infections, the fastest method to treat infected patients has led to repurposing of drugs which are being used for treatments of other existing ailments. As</p><p>Hippocrates said "Nature itself is the best physician", most drugs being developed today is mainly derived from plant sources. Naturally occurring compounds found in plants are also being screened as an effective treatment against SARS-CoV-2 virus. In this study, naturally occurring coumarin compounds were screened using molecular docking studies against SARS-CoV-2 main protease M pro out of which 17 coumarins could interact with the residues present in the active site of M pro . This interaction could suggest blockage of the active site residues of M pro which are required for proteolytic enzymatic activity to enable viral replication. These potential coumarin protease inhibitors displayed a binding energy score in the range of -6.80 kcal/mol to -8.57 kcal/mol out of which corymbocoumarin (-8.57 kcal/mol), methylgalbanate (8.30 kcal/mol) and hercalenol (8.20 kcal/mol) displayed best negative energy scores from their respective structural classes. The preliminary investigation of these coumarin compounds as potential viral protease inhibitors were carried out using computational methods and further in vitro and in vivo research work needs to be undertaken to corroborate our findings.</p>

ChemRxiv

Organ-on-chip applications in drug discovery: an end user perspective

Organ-on-chip (OoC) systems are in vitro microfluidic models that mimic the microstructures, functions and physiochemical environments of whole living organs more accurately than two-dimensional models. While still in their infancy, OoCs are expected to bring ground-breaking benefits to a myriad of applications, enabling more human-relevant candidate drug efficacy and toxicity studies, and providing greater insights into mechanisms of human disease. Here, we explore a selection of applications of OoC systems. The future directions and scope of implementing OoCs across the drug discovery process are also discussed.

organ-on-chip_applications_in_drug_discovery:_an_end_user_perspective

Introduction<!>Schematic diagram depicting an example of a basic liver-on-chip model as a representative organ-on-chip (OoC).<!>Examples of commercialised organ-on-chip (OoC) systems.<!>Introduction<!>Timeline for the evolution of organ-on-chip (OoC).<!>Introduction<!>Replicating in vivo conditions<!>Evaluating drug safety and efficacy<!>Incorporating immune cells<!>Materials and scaffolding<!>Increasing general uptake<!>Multi-organ combination in vitro<!>Conclusions<!>Competing Interests<!>

<p>Attrition rates have long been considered as the main cause of costs in drug development, which has been estimated to approximate $1 billion per new medicine entity between 2009 and 2018 [1]. Roughly half of the clinical trial terminations are attributed to a lack of efficacy and a further quarter are related to safety concerns. The lack of clinical translation of preclinical models used for assessing drug efficacy or toxicity is one of the major causes behind the high attrition rates [2]. Clearly, there is an unmet need to update the current preclinical testing paradigm.</p><p>Motivation to address these limitations has given rise to the organ-on-chip (OoC) technology: microfluidic, chip-based, three-dimensional (3D) cell culture models with an active flow (Figure 1). In recent years, commercially available OoCs have been increasingly integrated within the drug development phase, to replace more traditional preclinical models (Figure 2; [3–6]).</p><!><p>The active flow within the channel of a chip enables cells to be perfused with media containing oxygen and nutrients (inlet) and the removal of waste products and the sampling of metabolites for assessment of cell function (outlet). This schematic is for illustrative purposes only and does not represent any particular existing OoC model or all possible types of OoC models. It is intended to exemplify some of the basic bioengineering principles required for the development of an OoC unit.</p><!><p>(A) The HUMINIC chip 4 manufactured by Tissuse. (B) The OrganoPlate® 2-lane 96 manufactured by MIMETAS. (C) The PhysioMimix™ OoC manufactured by CN-Bio. The pictures are reprinted courtesy of the manufacturers and with their permission.</p><!><p>OoCs originated from the development of miniature devices compatible with cell culture and imaging [7] with the earliest OoC developed in 2010 by Donald Ingber's group [8]. The concept evolved from a research interest in the mechanical control of tissue and organ development [9] and led to the development of a lung-on-chip device; this integrated different tissues on the same chip to replicate the alveolar–capillary interface and recreated a functional, structural and mechanical representation of a lung alveolus. The temporal evolution of OoCs from precursor devices to the current state-of-the-art is depicted in Figure 3 [8,10–17].</p><!><p>The first OoC model to accomplish organ-level functionality, tissue–tissue interactions and a physiologically relevant organ microenvironment with vascular perfusion came in 2007 as a lung-on-chip by Huh et al. Since then, funding from organisations such as DARPA and NIH, and the founding of Wyss Institute, has propelled OoC technology from a nascent idea to a rapidly growing area of research with potential for utility across the drug discovery process. DARPA, Defence Advanced Research Projects Agency; FDA, US Food and Drug Administration; IO, immuno-oncology; MEM, micro electromechanical system; microTAS; miniaturised total chemical analysis system; NIH, National Institute of Health; PDMS, polydimethylsiloxane; TCTCs, Tissue Chip Testing Centers.</p><!><p>These models were initially referred to as microphysiological systems (MPS) and the denomination has since evolved with the technology itself. However, the aim has remained constant: creating tissue or organ functionality, beyond the capabilities of standard 2D or static 3D cell culture systems [18]. While 'MPS' and 'OoC' are often used interchangeably, the FDA define OoCs as part of a sub-category of MPS; they classify an MPS as an in vitro system, with cells isolated from tissues/organs or from organoids, that aims to recreate the physiological microenvironment. In contrast, an OoC is defined as a miniaturised MPS that is 'engineered to yield and/or analyse functional tissue units capable of modelling specified/targeted organ-level responses' [19], which is similar to the ORCHID definition [20]. Examples of OoC models replicating the microstructure, mechanical properties and functionalities of living organs have been reviewed [18,21]. Within the chip, one or more cell types are usually cultured within a 3D scaffold where cells can attach to an extracellular matrix (ECM) or porous membranes. The cells are continuously perfused with media containing oxygen and nutrients to maintain their function and survival, and the application of physical cues, such as shear stress, is possible. More recent advances, have enabled structures with higher aspect ratios and sensors, to monitor cellular and environmental changes [22,23].</p><p>The initial focus of OoCs was on single organs relevant to toxicology studies such as the kidney [13,24], liver [25,26], lung [8] or heart [27,28] It is now widely accepted that more accurate toxicology readouts may be achieved by incorporating multiple organs, to enable their interaction and more accurate evaluation of downstream metabolites. Efforts to integrate such models to form organ combinations are emerging, with the eventual goal of creating a 'body-on-chip' [29–31]. This provides greater scope and offers hope of gaining insight into human disease mechanisms, predicting the safety and efficacy of novel therapies and reducing the use of laboratory animals.</p><p>This review discusses the current applications of OoC technology and its scope for implementation in the drug discovery process, as well as its future directions from a drug discovery focussed end user perspective.</p><!><p>The primary aim of developing any OoC model for drug discovery purposes is to replicate in vivo physiology and provide a translational in vitro model. Despite efforts, to attain true clinical translatability remains a challenge and the technology is limited to a few examples.</p><p>One path to overcoming the translational hurdle is through the inclusion in microfluidic systems of tissue-specific environmental cues, such as flow and mechanical stress, recreating the tissue microenvironment in which cells would normally reside in vivo. Karalis and co-workers [32] recently showed how cyclic stretch could lead to the development of a gut-on-chip model that is physiologically more relevant than the corresponding static system; the transcriptomic profile of this model more closely resembled that of a gut tissue biopsy than the corresponding non-chip-based organoid.</p><p>Maintaining long-term cell viability and functionality in vitro has been a major challenge, but continuous perfusion of oxygen and nutrients, and removal of metabolic waste products in OoC models, enable longer-term culture and relevant cell analysis [33,34]. The use of primary cells obtained from multiple donors can capture donor-to-donor variation and provide an important insight into how in vitro results may translate to the clinic. In a liver-on-chip model, inter donor variability in metabolic clearance was predicted with albumin, urea, lactate dehydrogenase and cytochrome P450 mRNA levels. The correlation of predicted clearance with in vivo values enabled the development of an in silico drug metabolism model predicting pharmacokinetic variability [35]. The study provided insight into donor variability; however, the donor number was low and the model could not accurately reproduce the variability of some of the drug clearance parameters measured. It is likely that such models will be refined with the advent of higher throughput OoC models that enable testing of a greater number of donors tested at statistical powers equivalent to clinical trial.</p><p>The use of primary cells, where possible, is an important consideration when replicating in vivo conditions in OoC models. Obtaining good quality primary cells from tissues such as the central nervous system or a lung alveolus, remains a challenge. Further complexity arises when multiple cell types from a single donor are required to reconstitute organ functionalities. For example, a liver-on-chip model may require combinations of cell types including stellate cells, Kupffer cells, hepatocytes and liver sinusoidal endothelial cells. Induced pluripotent stem cells (iPSCs) are considered as an alternative to primary cells for autologous systems due to their wider availability and amenability to genetic modification [36,37]. Their ability to differentiate into different cell types, retaining single donor characteristics once differentiated, could enable personalised medicine development. There are disadvantages of iPSC technology: dependent on the cell type, the differentiation protocols can be complex and lengthy; iPSC derived cells are often structurally and functionally immature, as exemplified by iPSC derived cardiomyocytes that resemble foetal cardiomyocytes [38]. The development of more refined iPSC differentiation protocols leading to mature cell phenotypes would facilitate the establishment of autologous OoC models.</p><!><p>One of the main potential applications of OoCs is to assess the safety of drugs prior to entering clinical trials. Only 48% of adverse drug reactions in humans are predicted in preclinical testing [39]. This is in part due to the limited ability of commonly used preclinical species to capture human drug toxicities, as shown for the liver-induced toxicity of diclofenac [40].</p><p>A standard approach for the validation of a novel safety model, is to determine the sensitivity and specificity of the model with a set of compounds that have a large body of preclinical and clinical data available [41,42]. Once the dynamic range of the model is established, clinical safety outcomes are then replicated in an acute or chronic state. An immunocompetent liver-on-chip model described by Sarkar et al. [40] incorporated hepatocytes and nonparenchymal cells such as Kupffer cells to assess the acute form of drug-induced liver injury resulting from diclofenac secondary metabolite formation. As these cells typically lack functional longevity in standard static cultures, controlled perfusion via a microfluidic pump and a 3D scaffold, enabled the formation of tissue-like structures, which are critical for maintaining functionality [43]. In the model, diclofenac produced a metabolism and toxicity profile comparable to that observed in humans, but not in animal models. This liver-on-chip model subsequently enabled the development of a viral infection model by virtue of maintaining 40-day functional stability of the system for chronic safety testing [44].</p><p>The anti-proliferative effects of cancer drugs on haematopoietic stem cells are also not well-predicted in preclinical animal models. Bone marrow-on-chip models are being developed to examine potential toxicity screening of various compounds, including small molecules and large biologics. The bone marrow is a highly specialised niche composed of an array of cell types with complex phenotypes and these models must mimic the in vivo microenvironment and enable the growth of mesenchymal stem cells [45]. In collaboration with AstraZeneca, researchers from the Wyss Institute developed a vascularised bone marrow-on-chip model that successfully captured the dynamic physiology of a healthy bone marrow; the model supported the differentiation of multiple hematopoietic cell lineages over the course of 4 weeks, which reproduced the clinically observed toxicity profile of the inhibitor of aurora B kinase AZD2811. This enabled AstraZeneca to investigate better-tolerated dosing regimens for this compound in the clinic [46]. This model appears promising for evaluating long-term bone marrow toxicities of new chemotherapies and drug combinations. Furthermore, Cohen et al. [47] were able to study the mechanism of the drug-induced nephrotoxic side effects of cisplatin and cyclosporine with a novel kidney-on-chip model. The authors hypothesised that the clinical adverse effects of the two drugs were linked to glucose accumulation. This was validated by retrospectively analysing the clinical data of 247 patients that received cyclosporine or cisplatin in combination with the glucose reabsorption inhibitor empagliflozin, which was found to significantly reduce the incidence of kidney damage when compared with control groups. This work provides a perfect example of the potential for unravelling drug-induced toxicities with OoCs.</p><p>Predicting drug efficacy in preclinical models is another major challenge in drug discovery because current animal models do not always replicate well the pathophysiology of human disease. Recreating disease pathology affecting the pulmonary vasculature, for example, is highly challenging. A lung-on-chip model that successfully recreated a functional alveolar–capillary interface paved the way for various applications in the respiratory disease area [8]. The architecture of the lung in idiopathic pulmonary fibrosis, with its patchy stiff fibrotic tissue, has also been modelled by a lung-on-chip model that enables the study of anti-fibrotic drugs in greater detail than possible in vivo, such as the inhibitory effect of nintedanib on neo-vascularisation [48].</p><p>There has been significant interest in building human in vitro OoC models of disease in cases where animal models do not exist. One such example is Hepatitis B virus (HBV) infection; associated with liver cirrhosis and hepatocellular carcinoma, it is a global health concern, with over 240 million people infected worldwide [49]. Our understanding of the host–pathogen interactions is limited by the complexity of establishing a relevant model: patient-derived isolates of primary human hepatocytes (PHH) must be permissive to HBV infection and infection throughout all stages of the viral life cycle must be maintained. To overcome this hurdle, a liver-on-chip was developed to mimic HBV infection [44]. PHH cells were seeded onto the platform; continuous circulation of nutrients and oxygenated media led to the formation of hepatocyte microtissues which could be maintained in culture for at least 40 days, enabling the completion of the viral life cycle. Importantly, HBV infection resulted in innate immune responses, which replicated the clinical outcome in infected patients and supported the use of clinically relevant low viral titres. This aspect demonstrates the potential of OoC models to enable the investigation of immune evasion pathways of viruses, the modelling of drug treatment and the identification of novel clinical biomarkers.</p><p>In vitro human disease models are now proving suitable for efficacy testing and improving our understanding of molecular mechanisms of disease. Non-alcoholic steatohepatitis (NASH), the most severe form of non-alcoholic fatty liver disease (NAFLD), is a prime example. With the increasing prevalence of diabetes and obesity, NAFLD has become the most common chronic liver disease in developed countries with no specific pharmacological therapeutic options [50,51]. Previously, the mechanism for the association of a genetic variant of the lipase PNPLA3 in NAFLD was not well understood [52]. PHH and Kupffer cells were cultured on chip with wild-type or PNPLA3 I148M mutant hepatic stellate cells in the presence of free fatty acids to induce a NASH-like phenotype. In the model, hepatic stellate cells carrying the mutation, potentiated the disease state. The addition of the anti-NASH compound obeticholic acid reduced inflammatory mediators, as observed in clinical trials [53,54]. Those in vitro observations from the liver-on-chip model would have not been possible in a static 3D culture model, such as a PHH spheroid model, for several reasons: static models are incompatible with maintaining the physiological function of hepatocytes for prolonged periods and adding Kupffer and stellate cells to hepatocytes without the disruption of the spheroid structure is extremely challenging.</p><p>In conclusion, whilst OoCs appear to have tremendous potential for the assessment of drug safety and efficacy, systematic studies comparing the predictive power of available OoCs to those of current methodologies are lacking. It is, therefore, too early to provide a definitive view on the translational relevance of the OoC technology and how it compares to the current approaches in drug discovery.</p><!><p>The immune system is influential in the progression of many diseases including cancer, neurodegenerative diseases, chronic infections and autoimmunity. Considering the substantial differences between the human and animal immune system, the ability to incorporate immune cells into OoC systems to model human-specific immune responses to treatments targeting the immune system, such as biologics or cell therapies, will enable immunotoxicity assessments that are otherwise missed in in vivo models. The challenge is recapitulating the structural and functional complexity of the human immune system in a relevant manner [55].</p><p>The addition of immune cells to tumour-on-chip models enabled the study of the migratory phenotype of activated natural killer cells and their ability to penetrate into a glioblastoma tumour using time-lapse microscopy [56]. In another example, an immune-competent tumour-on-chip model, was used to track interactions between tumour fragments and autologous tumour infiltrating lymphocytes (TILs). Using automated quantitative image analysis, the fraction of cell death attributable to TILs was found to respond to the immune checkpoint inhibitor anti-PD-1 [16]. A hepatocellular carcinoma-on-chip model was used to evaluate the impact of tumour microenvironment conditions on the cytotoxicity of different T cells engineered to express a tumour-specific T cell receptor (TCR). Using this tumour-on-chip, it was possible to detect how T cell-mediated cytotoxicity was influenced by the tumour relevant hypoxic and inflammatory conditions. Optimal T cell-mediated cytotoxicity was observed under normoxic and inflammatory conditions, whereas hypoxia reduced their functionality. This difference in avidity of TCR–T cells was beyond the sensitivity of the 2D well plate-based assay that was run in parallel, showing the potential of OoC in immuno-oncology applications. A vascularised breast cancer-on-chip also enabled the identification of the immunomodulatory effect of cancer-associated fibroblasts on the anti-HER2 antibody, trastuzumab, mediated cytotoxicity [57].</p><p>Emulating the immune system in a standalone tissue-on-chip, for example a lymph node-on-chip, would further advance this technology. A simple design has recently been developed, where T cells and dendritic cells can interact with each other in a single channel, on the same plane. This interaction aimed to recapitulate the T cell and antigen-presenting cell cross-talk in the lymph node, which marks the initiation of the adaptive immune response through the appropriate antigen interrogation [58]. Ingber and co-workers [17] further developed a model and used it to evaluate a vaccine response; cell types self-organised into lymphoid follicles, the building blocks of germinal centres where B cells differentiate into antibody-secreting plasma cells. While these studies are promising, the qualification of these lymph node-on-chip models in other settings is important, such as testing the ability to generate antigen-specific neutralising antibodies during pathogen infections.</p><!><p>Polydimethylsiloxane (PDMS) is used extensively to manufacture chips and membranes and is advantageous for several reasons [59]. Its gas permeability enables oxygen supply to cells in microchannels, which is particularly beneficial for cultures of primary cells with high metabolic demands such as hepatocytes [60]. Its flexibility enables dynamic forces to be applied to cells: the mechanical stress of respiratory movements can be replicated in lung-on-chip models, where cyclic strain is applied to a PDMS membrane to act as a micro-diaphragm, replicating in vivo conditions [32,61]. The optical clarity of PDMS facilitates on-chip immunohistochemical staining and imaging, enabling easy characterisation of microtissues. Despite its versatility, the material properties of PDMS also present challenges. Its high hydrophobicity precludes cell adhesion and chemical or biological modifications are necessary to enable cells adhering to its surface [62,63]. Of importance for pharmaceutical applications, the hydrophobic surface also encourages the non-specific, unpredictable binding of small molecules to its surface, which reduces free drug concentration [64,65]. Another challenge with the use of PDMS in microfluidic systems is the leaching of remaining uncured oligomers into the culture medium and cells which can interfere with biological processes and lead to spurious experimental outcomes [66–68] The field has been looking into alternative materials that are non-absorbent, gas permeable, biocompatible, optically clear and amenable to mass manufacturing. Thermoplastics, hydrogels, glass and biocompatible materials with a long history of being used for tissue engineering, such as polylactic acid (PLA), as well as the combinations of these materials, are investigated to support the next generation of microfluidic chip manufacturing [69–71].</p><!><p>For a wider adoption of OoCs by the pharmaceutical industry, increased throughput and integration with existing laboratory equipment need to be addressed [72]. In lead optimisation, there is an increasing desire to evaluate potency in more complex models. This is in contrast with the late preclinical phase of clinical candidate identification where numbers of compounds tested are typically in the single-digit range and where in vivo relevance becomes a greater priority. The lead optimisation phase is where OoCs could gain traction, if throughput and cost could be improved. The OrganoPlate® platform with a microtitre plate footprint and 96-well capabilities exemplifies more recent trends fulfilling these conditions [73]. Its thin glass-bottom facilitates microscopic imaging and the microtitre plate format is compatible with automated readers and robot handling. Further development of OoC systems with the standard footprint of 24-, 96- or 384-well microtitre plates would ensure their seamless integration within the current robotic solutions and automation equipment already in place in the pharmaceutical research and development environment.</p><p>Achieving regulatory acceptance is also key to the mainstream adoption of the technology by end users. Currently, the use of OoC models is limited to preclinical drug discovery, with the data restricted to internal study reports. Given the potential power of OoC platforms in their clinical predictability, it is regrettable that those data are not included in regulatory documents as part of investigational new drug (IND) submission to regulatory bodies. To gain regulatory acceptance, models should have a defined test methodology, proven relevance, qualification and evidence of reliability in their context of use. A list of reference compounds with known toxicology and metabolism data has been compiled by the industry, to aid the qualification of liver-on-chip models [41]. A dialogue is encouraged by regulatory authorities both in Europe and the United States (US); by the European Medicines Agency Innovation Task Force and are also by the US Food and Drug Administration via the Center for Drug Evaluation and Research and the Alternative Methods Working Group. There are also ongoing efforts to qualify OoCs through the National Institute of Health's National Center for Advancing Translational Sciences. Ultimately, these efforts must be replicated globally to accelerate the wider acceptance of OoCs.</p><!><p>Although standalone OoC models can improve the predictability of preclinical testing, their full potential will be unleashed through multi-organ combinations. Single-organ models are not always fully predictive of the pharmacokinetic and safety properties of drugs, a multi-organ model has enormous potential in reducing or replacing animal usage. One of the first liver-on-chip designs developed at the Massachusetts Institute of Technology by Linda Griffith, has been adopted to create 4-, 7- and 10-way organ models that have demonstrated organ functionality for up to 4 weeks [74]. A similar approach with a 4-organ chip that connects the gut, liver, kidney and bone marrow was used to predict the pharmacokinetic properties of cisplatin [75]. Maschmeyer et al. [76] developed both a 2-organ and 4-organ chip, and used the 4-organ chip to interconnect models of human intestine, liver, brain and kidney derived from iPSCs, using a universal medium to enable the maintenance of the different cell types used for 2 weeks. Multiple examples of interconnected OoC models quickly followed, with an integrated gut–liver [77], female reproductive tract [78] and an 8-OoC combination coupled to an automated culture and sampling system [79]. Although these initial multi-organ models represent promising proof of concept studies, their day-to-day relevance to drug discovery remains to be confirmed. More targeted approaches with 2- to 4-organ combinations enabling specific pharmacokinetic/pharmacodynamic (PK/PD) or absorption, distribution, metabolism and excretion (ADME) modelling are more likely to be initially adopted.</p><p>The main bottleneck to developing multi-organ chips lies in chip design and tissue maturation. A potential barrier to interconnecting multiple OoC models is the wide variety of platforms being designed, pointing out a clear need for standardisation. For maximum usability, each module should have connector valves and similar scaling to accurately mimic in vivo vascular blood supply and flow rate. The tissue culture media must also be compatible with cell types across all connected models; a challenge, considering the specific requirements of cell types to specialised culture media [45]. Advances in the development of common media for multiple organ-derived cell types have been reported, but these only support short-term co-cultures [30]. Modular 'plug and play' systems are being developed to provide flexible connections without leakage. For example, the μOrgano system is a Lego®-like plug and play system which enables the initial culturing of independent OoC systems and subsequent connection to create an integrated multi-organ system [80]. The flexibility to integrate models at different timepoints, opens the opportunity for tissue-specific maturation processes and truly multi-organ chip models. There are multiple engineering challenges, such as different chip design requirements per organ, integrating a vascular connection to each organ and avoiding the introduction of air bubbles or infection risks at connection points.</p><!><p>OoC technology is still in its early stages of uptake in drug discovery, however, its potential to predict drug safety and efficacy could have significant impact throughout the drug discovery process. The potential of OoC models, to recapitulate the physiological, mechanical and biochemical complexity of human tissues may enable the assessment of in-depth functional parameters beyond the scope of current in vitro cultures. The qualification and regulatory acceptance of these models will help to derisk their uptake by the pharmaceutical industry and should facilitate wider adoption. Plug and play immune-competent multi-organ systems manufactured in the right materials, compatible with existing laboratory equipment and meeting the throughput needs of the industry is key for their overall success. Ultimately, adopting a collaborative approach, with academia and industry working in close partnership will be required in order to meet those objectives. New disruptive drug discovery paradigms are on the horizon.</p><!><p>The authors declare that there are no competing interests associated with the manuscript.</p><!><p>hepatitis B virus</p><p>induced pluripotent stem cells</p><p>microphysiological systems</p><p>non-alcoholic fatty liver disease</p><p>non-alcoholic steatohepatitis</p><p>polydimethylsiloxane</p><p>primary human hepatocytes</p><p>T cell receptor</p><p>tumour infiltrating lymphocytes</p>

PubMed Open Access

Identification of novel small molecule inhibitors for solute carrier SGLT1 using proteochemometric modeling

Sodium-dependent glucose co-transporter 1 (SGLT1) is a solute carrier responsible for active glucose absorption. SGLT1 is present in both the renal tubules and small intestine. In contrast, the closely related sodium-dependent glucose co-transporter 2 (SGLT2), a protein that is targeted in the treatment of diabetes type II, is only expressed in the renal tubules. Although dual inhibitors for both SGLT1 and SGLT2 have been developed, no drugs on the market are targeted at decreasing dietary glucose uptake by SGLT1 in the gastrointestinal tract. Here we aim at identifying SGLT1 inhibitors in silico by applying a machine learning approach that does not require structural information, which is absent for SGLT1. We applied proteochemometrics by implementation of compound- and protein-based information into random forest models. We obtained a predictive model with a sensitivity of 0.64 ± 0.06, specificity of 0.93 ± 0.01, positive predictive value of 0.47 ± 0.07, negative predictive value of 0.96 ± 0.01, and Matthews correlation coefficient of 0.49 ± 0.05. Subsequent to model training, we applied our model in virtual screening to identify novel SGLT1 inhibitors. Of the 77 tested compounds, 30 were experimentally confirmed for SGLT1-inhibiting activity in vitro, leading to a hit rate of 39% with activities in the low micromolar range. Moreover, the hit compounds included novel molecules, which is reflected by the low similarity of these compounds with the training set (< 0.3). Conclusively, proteochemometric modeling of SGLT1 is a viable strategy for identifying active small molecules. Therefore, this method may also be applied in detection of novel small molecules for other transporter proteins.Electronic supplementary materialThe online version of this article (10.1186/s13321-019-0337-8) contains supplementary material, which is available to authorized users.

identification_of_novel_small_molecule_inhibitors_for_solute_carrier_sglt1_using_proteochemometric_m

Introduction<!><!>Merging different datasets<!><!>Proteochemometric modeling of hSGLT1<!><!>Screening for hSGLT1 actives in a commercially available compound library<!><!>Screening for hSGLT1 actives in a commercially available compound library<!>Cytotoxicity of hSGLT1 actives<!>Compound activity for hSGLT2<!>Hit compound analysis<!>Aiming for specific targeting at the gastrointestinal tract<!><!>Conclusions<!>Compounds and assay materials<!>Assay procedure<!>Cytotoxicity assay<!>Dataset<!>Compound descriptors<!>Protein descriptors<!>Machine learning<!>T-distributed stochastic neighbor embedding<!>Clustering of hSGLT1 actives to explore binding modes<!>Computational hardware<!>

<p>Sodium-dependent glucose co-transporters, or sodium-glucose linked transporters (SGLTs), are solute carriers (SLCs) that are responsible for glucose (re)absorption. SGLTs are members of the sodium-dependent transporters and are encoded by the SLC5A genes [1]. SGLTs are interesting targets in the treatment of diabetes mellitus, as their inhibition reduces the risk of hyperglycemia by decreasing glucose (re-)uptake [2]. In the human body two SGLT isoforms are involved in glucose transport: SGLT1 and SGLT2 [3]. Both SGLT1 and SGLT2 are expressed in the kidney, whereas SGLT1 is also expressed in the small intestine [4]. SGLT2 is a high capacity transporter responsible for 90% of glucose reuptake in the renal tubules and multiple compounds have been developed that inhibit this solute carrier [5, 6]. Furthermore, SGLT2 inhibition has been shown to decrease blood glucose levels in diabetes type 2 patients [7]. In contrast to SGLT2, SGLT1 is a low-capacity glucose transporter [1]. However, SGLT1 has a higher glucose affinity than SGLT2 and is additionally capable of transporting galactose [1]. Dual inhibitors blocking both SGLT1 and SGLT2 are currently in clinical development [8, 9]. In line with previous evidence we suggest that SGLT1 inhibition in the intestine will lower blood glucose levels as well [10, 11]. Compounds that do not penetrate the intestinal wall can achieve selective targeting of SGLT1 in the intestine, as they would not reach the renal tubules [12].</p><p>The complexity and the hydrophobic nature of transporter proteins make them challenging to crystalize. Crystal structures of transporters are scarce and binding locations of small molecules to these transporters are often unknown. For human SGLTs no protein structures are available negating the use of structure-based modeling techniques. However, the publicly available compound database ChEMBL includes ligand–protein binding information for multiple SGLTs [13–15], allowing the use of statistical modeling techniques such as quantitative structure–activity relationship analysis (QSAR) and proteochemometrics (PCM) [16]. These techniques, which make use of machine learning, do not require protein structural information and can therefore be applied in the context of SLCs. Although ligand-based pharmacophore modeling, QSAR, and PCM have only been applied to a few SLCs [17, 18], these techniques are well established on other drug targets including membrane proteins such as G protein-coupled receptors [19–21].</p><p>Unfortunately, the publicly available compound interaction data for SGLTs is limited from the point of chemical diversity as the major share of ligands are glycoside-like compounds and oxopyrrolidine-carboxamides. This limited chemical space hence restricts the applicability domain of QSAR and PCM models [22]. The applicability domain of computational models can be interpreted as the theoretical ensemble of molecular structures to which a model can be applied accurately. This domain is dependent on the model input and can therefore be quantified by similarity with the training molecules.</p><p>In the current work we show how we expanded the chemical space of SGLT inhibitors (using an in-house dataset [Oranje et al. manuscript in preparation]), and with that the applicability domain of our SGLT models. We constructed PCM models based on SGLT1 and its closest family members to predict compound activity for SGLT1. We successfully identified novel SGLT1 inhibitors that display low similarity towards the training set.</p><!><p>Chemical space of the public and in-house datasets. a The t-SNE shows molecular structure and affinity (pKi for public data and % of (negative) control for in-house data) for representative hSGLT1 compounds. b Molecular weight and ALogP distribution of compounds in the training sets</p><!><p>To merge the public and in-house dataset the difference in activity units for both sets had to be resolved. The public dataset contains pChEMBL values, representing a standardized unit for affinity and potency values such as Ki, IC50, EC50, and Kd [26]. The potency values in the in-house dataset were available as percentage activity compared to (negative) control at a concentration of 50 μM, which could not be converted into a pChEMBL value. Hence, binary classification models were chosen over regression.</p><!><p>Activity threshold grid search. Searching the activity threshold grid for in-house (activity percentage compared to negative control) and public data (pChEMBL value). Model performance was measured using Matthews Correlation Coefficient (MCC), which was 0.48 for the final selected thresholds of < 70% for in-house data and pChEMBL > 8.5 for public data</p><p>Model performance depends on datasets that are used in training</p><p>PD public data, IH in-house data, EV external validation on 30% of data, CV fivefold cross validation on 20% of the data per iteration</p><!><p>Next, a PCM model was constructed based on the combined full data set consisting of all public and in-house data. To validate the performance of this model, fivefold cross-validation was applied with the same test sets as applied in validation of performance of the public data model: rotationally 20% of the in-house hSGLT1 data was used as holdout test set; the remaining 80% was used in training. In each case the test set contained compounds not available for training. This resulted in the following performance: sensitivity 0.64 ± 0.06, specificity 0.93 ± 0.01, PPV 0.47 ± 0.07, NPV 0.96 ± 0.01, and MCC 0.49 ± 0.05. Overall performance of this PCM model was regarded satisfactory for predictions of new compounds and was comparable with the QSAR benchmark model used for activity threshold determination previously.</p><p>Additionally the performance of models trained on in-house data only was tested to assess the effect of addition of public data. Public domain compounds contributed slightly to the predictive performance of the model in specificity, PPV, and MCC. This was observed by a minor decrease in performance upon removal of the public data from the training set: sensitivity 0.69 ± 0.07, specificity 0.89 ± 0.02, PPV 0.38 ± 0.06, NPV 0.97 ± 0.01, and MCC 0.45 ± 0.05. Although the difference in performances is not significant, it is remarkable that the number of false positives decreases considerably when public data is included in training, whereas the number of true positives is only slightly negatively affected: false positives 28 ± 6 versus 43 ± 6, true positives 24 ± 4 versus 26 ± 4 (with and without public data, respectively). Apparently, the public data by itself is not sufficient in predicting hSGLT1 activity in the chemical space of the in-house compounds but does add favorably to model performance when supplemented to the in-house dataset.</p><!><p>Chemical space of the selected compounds compared to the training and screening datasets. a The Diverse set (yellow) and Cluster set (green) are displayed compared to the training (orange and red) and Enamine screening set (blue). The Enamine set is represented by a random selection of 20,000 out of the total of 1,815,674 compounds (~ 1%) in the screening set to limit t-SNE calculation time. b The molecular weight and ALogP of the Diverse and Cluster set compared to the training and screening sets</p><!><p>To increase confidence in the activity of compounds the screened set was pre-filtered by selecting compounds with a predicted class probability of ≥ 0.8 on a scale from 0 to 1. Here, a resulting score of 1 represents compounds predicted to be in the 'active' class, a score of 0 indicates that the compounds are predicted 'inactive'; ascending scores indicate higher certainty of compounds belonging to the 'active' class. Additionally, compounds with molecular weight ≤ 300 were removed to exclude fragment-like compounds. The final filtered set contained 672 compounds.</p><!><p>Reference hSGLT1 inhibitors for Cluster set and their inhibitory activity. Inhibitory activities (compared to negative control, where 100% is no inhibition) and chemical structures of four recently identified novel hSGLT1 inhibitors: bepridil, bupivacaine, cloperastine, and trihexyphenidyl</p><!><p>The total selection of 77 unique compounds was tested in vitro in cells expressing hSGLT1 in a single point measurement at a concentration of 50 μM. From the 40 diverse predicted hits that were assessed, 15 compounds were defined active as they displayed hSGLT1 inhibition in vitro with an activity reaching values below 70% compared to the negative control (100%: no inhibition) (Additional file 5: Data S5). From the 37 Cluster set compounds, an additional 15 compounds were confirmed to be active (Additional file 6: Data S6).</p><!><p>The potential cytotoxicity of the screening compounds (Diverse set and Cluster set) was investigated by analysis of secreted adenylate kinase (AK), a marker of cell wall integrity loss. Most compounds did not show any indication of cyotoxicity, however one active from the Diverse set displayed moderate impairment of the cell wall (Z1416510792: activity 43 ± 9%, cytotoxicity 25%). The cytotoxicity assay was limited by the available supernatant from the activity screen. Therefore not all compounds were measured in duplicate and cytotoxicity of one active from the Cluster set could not be determined (Z817504494: activity 45 ± 3%).</p><!><p>Both the Diverse set and Cluster set compounds were additionally measured for hSGLT2 inhibitory activity to assess their selectivity between the two transporters. The same cellular screening assay was performed as was used for hSGLT1 (single point measurement at a concentration of 50 μM). More actives were defined for hSGLT2 compared to hSGLT1 using the same activity threshold of 70% activity relative to negative control (100%: no inhibition): 22 actives in the Diverse set and 19 in the Cluster set. Almost all hSGLT1 actives showed activity for hSGLT2 with the possible exception of Z105569118, which only marginally surpassed the activity threshold for hSGLT2 (activity of hSGLT1 64 ± 4% and hSGLT2 76 ± 5%). No selective compounds were identified for hSGLT1, with 14% being the highest observed difference in inhibition (Z46160496: hSGLT1 41 ± 4% and hSGLT2 55 ± 2%). For hSGLT2 the biggest difference in inhibition was found for Z1318177320 that showed a difference of 39% (hSGLT1 93 ± 20% and hSGLT2 54 ± 0%).</p><!><p>The activities of the hit compounds of the Diverse and Cluster set were analyzed. The strongest inhibitors, Z163972344 and Z915954934, were derived from the Diverse set with activities of 24 ± 1% and 28 ± 4% (100%: no inhibition), respectively. Z163972344 has low similarity (0.27 based on Tanimoto FCFP6) with the training set, indicating that this is a truly novel inhibitor for hSGLT1. The average similarity of actives in the Diverse set compared to training was 0.33, with Z1416510792 being the active that is most similar to the compounds in the training set with a similarity score of 0.61 (this compound showed moderate AK secretion in the cytotoxicity assay).</p><p>For the Cluster set a total of 15 actives were validated for the four different clusters. The cloperastine cluster encompassed the most actives (60% actives), whereas the trihexyphenidyl and bepridil clusters contained the least actives with 29% and 30% actives, respectively. The bupivacaine cluster had an intermediate hit rate of 40%, which is comparable with the overall hit rate of the total Cluster set (41%). The variance in hit rates between the four clusters is also reflected in the similarity of compounds toward their cluster reference: the cloperastine and bupivacaine clusters contained the most similar compounds (average similarities towards cluster reference compound were 0.43 and 0.42, respectively); the trihexyphenidyl and bepridil clusters contained less similar compounds (0.35 and 0.31, respectively).</p><p>Although the cloperastine and bupivacaine clusters contained the most similar cluster members, no conclusive SAR could be determined. The cluster members displayed variations in methyl substituents, which showed an effect for two compounds in the bupivacaine cluster [Z46224544 (45 ± 10%) and Z2217101732 (74 ± 8%)]. This was however not observed for compounds in the cloperastine cluster: Z31367782 (36 ± 4%), Z31371621 (37 ± 3%), Z31367784 (43 ± 7%), and Z31370217 (45 ± 10%). The positions of the methyl substituents were too distinct to make solid conclusions on their relationship with compound activity.</p><p>In general, the novel active entities contain at least one aromatic ring and two hydrogen bond acceptors. Only two of the 30 actives did not adhere to Lipinski's rule of five, with an ALogP of 5.2 and 6.2 for Z1844922248 (activity 49 ± 7%) and Z56906862 (activity 38 ± 5%), respectively.</p><!><p>As mentioned in the Introduction, hSGLT1 inhibition at the intestinal wall is desired. Based on chemical structure and physicochemical properties the identified hit compounds will most likely be absorbed. However, it is suggested that modifications can be introduced to improve specific intestinal targeting. These alterations, such as a higher molecular weight, can prevent compounds from being absorbed or transported by the intestinal wall [28]. Intestinal SGLT1 blockers are expected to display less renal damage, which is an adverse effect observed for SGLT2 inhibitors [6]. Moreover, drug action restricted to the gastrointestinal tract also limits other off-target interactions, which were observed for the marketed SGLT2 inhibitor canagliflozin [29]. An example of a compound that was optimized for specific targeting at the gastrointestinal tract is LX2761, an inhibitor aimed at intestinal SGLT1 that decreased glucose uptake in mice [30, 31]. Although SGLT1 inhibition at the intestine may not compromise renal function, other adverse effects that can result from intestinal targeting need to be considered [32, 33].</p><!><p>Clustering of hSGLT1 actives. Active hSGLT1 compounds in the training set clustered into ten chemical clusters (Tanimoto, FCFP6). Molecular structure and affinity (pKi for public data and  % of (negative) control for in-house data) for representative cluster compounds are shown. In-house compounds with activity < 70% of (negative) control and public compounds with pChEMBL ≥ 6.5 were used in clustering. a t-SNE plot of the chemical clusters. b The molecular weight and ALogP distribution of compounds in the chemical clusters</p><!><p>We have demonstrated that PCM modeling is a viable method to identify novel inhibitors for solute carrier hSGLT1 and hence likely any solute carrier protein. A predictive SGLT model was built with a MCC value of 0.49 ± 0.05, estimated with fivefold cross-validation. With the optimized model a hit rate of 38% was achieved when it was applied to screen for diverse molecules (Diverse set). In parallel, the model was used to boost identification of actives with a given chemotype (Cluster set). Although additional active compounds were identified, the data was too ambiguous to gain insight into the SAR of hSGLT1 inhibitors.</p><p>Diversity was found within the in-house dataset and differences were observed between the in-house chemical space and that of the public dataset. Furthermore, the intrinsic variety in chemical structure of active compounds implies that there may be multiple binding sites at the transporter protein.</p><p>The novel identified inhibitors showed low similarity towards the training set and belong to the same chemical space of the in-house dataset, in contrast to the public dataset. Although the inhibitors were not optimized for specific drug delivery to the gastrointestinal tract, it is suggested that alterations (such as an increase in molecular weight and size) can make these inhibitors selective for intestinal hSGLT1.</p><!><p>DMEM-F12 (Biowest, Cat. No. L0092-500), DMEM (Lonza, BE12-604F/U1), Heat Inactivated Foetal Bovine Serum (HI-FBS, Biowest, Cat. No. S181H-500) and HBSS without Ca and Mg (HyClone, Cat. No. SH30588.01), DPBS (HyClone, Cat. No. SH30028.02), isopropanol (20,842.312), clear-bottom black 96 well plates (Greiner, Cat. No. 655090) and polypropylene 96-well plates (Nunc, Cat. No. 151193) were all obtained from VWR (Amsterdam, the Netherlands). TrypLE Express (Gibco, Cat. No. 12605010), geneticin (Gibco, Cat. No. 10131027), d-glucose free DMEM (Gibco, Cat. No. 11966025), water soluble probenecid (Invitrogen, Cat. No. P36400), 5000 U/mL penicillin–streptomycin (Gibco, Cat. No. 15070063) were all ordered from Thermo Fisher Scientific (Breda, the Netherlands). 1-NBD-Glucose was custom synthesized by Mercachem (Nijmegen, the Netherlands). Bovine serum albumin (Cat. No. A8806), poly-l-lysine hydrobromide mol. wt. 30,000–70,000 (Cat. No. P2636), cell culture grade DMSO (Cat. No. D2650) were all acquired from Sigma-Aldrich Chemie (Zwijndrecht, the Netherlands). The hSGLT1 cDNA cloned in the pCMV6-neo vector was purchased from Origene Technologies (Rockville, USA, Cat. No. SC119918). The hSGLT2 cDNA was custom synthesized and cloned into the pcDNA3.1 vector by Thermo Fisher Scientific (Breda, the Netherlands). The experimentally tested Enamine screening compounds were acquired from Enamine (Kyiv, Ukraine).</p><!><p>Two days in advance, CHO-hSGLT1 or CHO-hSGLT2 cells were seeded in maintenance medium (DMEM-F12 supplemented with 10% HI-FBS and 400 μg/mL geneticin) at 60,000 cells/well in clear-bottom black 96 well plates, pre-coated with 100 μg/mL poly-lysine. Cells were washed with 240 μL/well d-glucose free DMEM. Dilutions of test compounds and controls prepared in d-glucose free DMEM with 350 μM 1-NBd-Glucose, 0.3% BSA, and 2 mM probenecid were added at 90 μL/well and placed in a humidified incubator at 37 °C with 5% CO2 for 30 min. Subsequently cells were washed once with ice-cold DMEM-F12 and once with ice-cold HBSS, both at 240 μL/well. Finally, 1-NBd-Glucose was extracted from the cells with 100 μL/well isopropanol for 10 min at 600 rpm on an orbital shaker. Fluorescence was measured on a Flexstation 3 (Molecular Devices, San Jose, USA) with excitation at 445 nm, emission at 525 nm and cut off 515 nm. The uptake of 1-NBD-Glucose was normalized to the dynamic range between minimal inhibition (0.2% DMSO vehicle control) and maximal inhibition (100 μM phloridzin, > 100 × SGLT1/2 IC50). Phloridzin is a strong inhibitor of SGLT1 and SGLT2 and was used as 0% reference, with 100% being no inhibition. A concentration of 100 μM phloridzin was used to ensure full SGLT1/2 inhibition. The Z-factor for the controls was determined and only data with Z > 0.4 (average Z SGLT1 assays: 0.8 ± 0.1, average Z SGLT2 assays: 0.6 ± 0.1) was used [37].</p><!><p>The cytotoxicity of compounds was tested with the ToxiLight bioassay kit (Lonza, obtained from VWR, Amsterdam, The Netherlands) according to the supplier's instructions. This non-destructive assay measures leakage of the enzyme AK from damaged cells into the CHO-hSGLT1/2 inhibition assay media, i.e. the degree of cytolysis. AK converts ADP into ATP and the enzyme luciferase subsequently catalyzes the formation of light from ATP and luciferin. Briefly, 20 mL of CHO-SGLT1/2 inhibition assay medium was added to 100 mL reconstituted AK detection reagent in white 96 wells Cellstar plates (Greiner bio-one, obtained from VWR, Amsterdam, The Netherlands) and incubated for 5 min at room temperature. Next, bioluminescence was measured on a FlexStation 3 Multi-Mode Microplate Reader (Molecular Devices, San Jose, USA) by 1 s integrated reading. Cytotoxicity was expressed as the percentage of bioluminescence of the 0.5% DMSO vehicle control which was set at 0%. The average cytotoxicity was calculated from biological replicates as indicated and average values > 20% were considered toxic (arbitrary threshold).</p><!><p>Publicly available data from ChEMBL (version 23) was extracted for human SGLT1 (accession: P13866), human SGLT2 (P31639), and related proteins human SGLT3 (Q9NY91), rat SGLT1 (P53790), rat SGLT2 (P53792), mouse SGLT1 (Q9QXI6), mouse SGLT2 (Q923I7), and mouse SGLT3 (Q8R479). The retrieved compounds were standardized by removing salts, keeping the largest fragment, standardizing stereoisomers, standardizing charges, deprotonating bases, protonating acids, and optimizing the 2D structure by correcting bond lengths and angles. Activity values with confidence score 7 and 9 were kept and duplicate activity values were discarded based on activity standard unit ranking: Ki > IC50 > EC50 > Kd. For duplicate compounds with similar activity standard units (e.g. a compound with two Ki values), the average pChEMBL value was calculated.</p><p>An additional in-house dataset was provided by Unilever, Vlaardingen [Oranje et al., manuscript in preparation]. This dataset was based on the Spectrum Collection compound library (MicroSource Discovery Systems) extended with additional compounds that were similar to primary bioassay screening hits. This dataset consisted of compound activity data for hSGLT1 and hSGLT2. The activity was expressed as percentage 1-NBD-Glucose uptake compared to control at 50 μM, with control being the absence of inhibitor (= 100%). Molecular structures were standardized in the same manner as the public data. The final dataset (public and in-house datasets combined, no duplicates) encompassed 3686 unique compounds with 4208 derived activities, of which 2888 for hSGLT1.</p><!><p>Compounds were described using 512 FCFP6 fingerprint bits and the following physicochemical properties: molecular weight, ALogP, number of hydrogen bond acceptors, number of hydrogen bond donors, number of rotatable bonds, number of bridge bonds, and number of aromatic rings. Fingerprints and physicochemical descriptors were calculated in Pipeline Pilot (version 16.1.0) [38].</p><!><p>Protein sequences were aligned using whole sequence alignment in Clustal Omega (version 1.2.2) [39]. Subsequently the sequences were converted to protein descriptors using Z-scales [40]. The first three Z-scales were implemented as protein descriptor as these were shown to perform well in previous work [41]. These three Z-scales include information on residue lipophilicity, size, and polarity.</p><!><p>Models were trained using the Random Forest R component in Pipeline Pilot (version 16.1.0). The number of trees was 500 and number of variables tried at each split was 38 (square root of the number of descriptors). Remaining settings were kept default.</p><!><p>T-SNE was calculated on FCFP6 fingerprint descriptors that were converted to 2024 bits. The t-SNE component in Pipeline Pilot (version 18.1.0) was used to perform tSNE. The derived t-SNE values are represented by two components: CSNE1 and CSNE2.</p><!><p>hSGLT1 active compounds in the training set were clustered into ten clusters using the cluster molecules component in Pipeline Pilot (version 16.1.0). Compounds from the in-house set were included as 'active' when percentage of (negative) control was < 70%. Compounds from the public data set were termed 'active' when pChEMBL value ≥ 6.5.</p><!><p>Experiments were performed on a server running CentOS 6.9 equipped with a dual Xeon E-5 2630 v2 processor and 128 GB of RAM.</p><!><p>Additional file 1. T-SNE representation of the chemical space of the public and in-house datasets colored by species.</p><p>Additional file 2. Schematic overview of the experimental workflow of this study.</p><p>Additional file 3. Random Forest SGLT PCM model used for final predictions.</p><p>Additional file 4. T-SNE representation of actives and inactives of selected compounds compared to the training set.</p><p>Additional file 5. Bioactivities, cytotoxicities and Tanimoto similarities of the Diverse set.</p><p>Additional file 6. Bioactivities, cytotoxicities and Tanimoto similarities of the Cluster set.</p><p>Additional file 7. Cluster centers and distribution of physicochemical properties of hSGLT1 active compound clusters.</p><p>adenylate kinase</p><p>high-throughput screening</p><p>Matthews correlation coefficient</p><p>negative predicted value</p><p>proteochemometrics</p><p>positive predicted value</p><p>quantitative structure–activity relationship</p><p>sodium-dependent glucose co-transporter 1/2</p><p>t-distributed stochastic neighbor embedding</p>

PubMed Open Access

Using the SARS infection transcriptional signature to identify potential treatments for Covid-19

In the light of the SARS-CoV-2 pandemic there is a pressing need to trial as wide a range of drugs as possible as there is at present no effective treatments. There is limited time to develop drugs according to the standard pipeline. This makes repurposing, where existing therapeutics with known side-effect profiles and pharmacokinetic data are deployed for diseases other than those they were originally developed to ameliorate, an attractive endeavour. Here, publicly available gene expression data associated with SARS infection of human epithelial cell lines from multiple independent studies has been harnessed to delimit a SARS signature profile. The SARS signature exhibits a strong upregulation of gene sets involved in viral defence mechanisms. The SARS signature has then been queried against gene expression data bases of cellular transcriptional response to approved drugs and candidate therapeutics hypothesised on the basis of their ability to bolster the host cell response to SARS infection. Candidate drugs thus identified, show a marked enrichment for reported anti-viral activity, with a number also reported to exhibit potent activity against SARS-CoV-2. It is hoped that this approach may lead to an effective and rapid repurposing strategy for treating the Covid-19 pandemic.

using_the_sars_infection_transcriptional_signature_to_identify_potential_treatments_for_covid-19

Introduction<!>Methods<!>The SARS infection signature profile (SISP)<!>SISP in a wider context<!>Drug repurposing with the SISP<!>Conclusion

<p>The rapid evolution rate of certain coronavirus strains has, over the recent past, led to increasingly severe zoonotic human epidemics. In 2002 a severe acute respiratory syndrome (SARS) emerged in China resulting in 8,000 deaths (10% of infections) across 29 countries. Ten years later the more deadly (36% of infections) but less infectious middle east respiratory syndrome (MERS) developed in Saudi Arabia spreading to 27 countries with total infections currently at 2494 with 858 deaths. SARS-CoV-2, which emerged in China in 2019, appears to be less deadly than SARS-CoV or MERS-CoV, with current estimates of between 1 and 2%, but it has proved to be highly infectious with current infections standing at nearly four and half million and as many as 314,000 fatalities as of May 2020.</p><p>Efforts are underway to develop a SARS-CoV-2 vaccine (Amanat and Krammer, 2020), while inhibiting viral proteases (Jin et al., 2020) and the recent solution of the structure of the SARS-CoV-2 spike protein bound with its accessory protein ACE2 (Lan et al., 2020) offers the possibility of developing agents to block viral entry. However, the current pandemic trajectory calls for a more rapid intervention deployment. To this end the broad spectrum anti-viral Remdesivir targeting RNA polymerisation has been approved for emergency use in the USA and for severe cases in Japan (Ko et al., 2020). Recently, a more ambitious initiative has sought to repurpose existing therapeutics, not necessarily of an anti-viral class, for SARS-CoV-2 based on compiling evidence for 332 separate interactions between viral and host cell proteins (Gordon et al., 2020). This database was then used to identify existing drugs for potential repurposing based on their targeting the human proteins that interact with the viral proteins.</p><p>Drug repurposing is an attractive strategy because it circumvents the formidable hurdles that are inevitable with the exploration of novel chemical entities that result from high throughput screening of compound libraries against defined targets. With repurposing the target is rather a phenotype as there is little hope of finding a specific high affinity compound against a specified target from a limited set of a few 1000 drugs. Phenotype-based screening requires the development of high content quantitative descriptor of the underlying biological disease state and one such is transcriptional profiling. In addition to the 3.5 million samples curated by NCBI GEO (Barrett et al., 2007) there are large collections of publicly available compound associated expression databases developed as part of the connectivity map (CMAP) project (Lamb et al., 2006). The CMAP2.0 database contains the transcriptional profiles of 1,309 drug-like compounds and therapeutics profiled across four human cancer cell lines (www.broadinstitute.org/connectivity-map-cmap). The CMAP database has recently been greatly expanded to include profiles for over 15,000 compounds defined across almost 100 different cell lines (Subramanian et al., 2017). The latter LINCS database is compiled from the expression levels for 1000 landmark genes from which the values for the entire genome are imputed with linear modelling. In this sense the two datasets are complimentary.</p><p>Transcription-based drug repurposing is founded upon the observation that transcriptional changes seen in disease states are largely recapitulated across independent samples and thereby can serve as effective disease descriptors or quantitative phenotypes (Golub et al., 1999;Lee and Young, 2013). This led the way for the hypothesis that drugs tending to reverse disease associated gene expression changes may thereby function as disease modulators and potential therapeutics (Hughes et al., 2000;Marton et al., 1998;Walf-Vorderwulbecke et al., 2018;Wei et al., 2006;Williams, 2012;Zhang et al., 2012). However, in some cases the expression changes seen in disease states, at least in the initial response phase, are beneficial and it could therefore be counterproductive for these changes to be reversed. In the particular case of viral infection, there is an established cellular response that may benefit from being enhanced by therapeutic intervention rather than inhibited.</p><p>The transcriptional consequences of viral infection have been assayed in multiple cellular contexts and animal models (Mitchell et al., 2013;Webster et al., 2009;Yoshikawa et al., 2010). The motivation for the present work is to harness this rich data source and; firstly to assess the extent to which gene expression changes are robust, being recapitulated across multiple studies; secondly, to inspect the class of genes being perturbed by viral infection and assess whether this response is a defence mechanism or a facilitating alteration of the cellular genome. Once the nature of the cellular response to the virus has been established compound libraries can be queried for candidate therapeutics on the basis of either reversing the virus induced changes or enhancing them.</p><p>The analysis presented below establishes a clear conservation of gene expression changes across independent SARS infection data sets for multiple human epithelial cell types. This leads on to the development of a SARS signature profile (SISP) of 192 up-regulated genes and 136 down-regulated genes. The biological nature of this response is investigated through an enrichment analysis against established pathway gene sets and gene ontology classes. It is shown that the SISP is dominated by a viral defence mechanism and it is therefore concluded that a drug repositioning endeavour should be based on recapitulating this response. To this end the SISP is queried against the CMAP and LINCS databases and a set of candidate therapeutics proposed. It is strikingly clear that drugs positively correlating with the viral response tend to have established anti-viral activities, with a number already shown to protect against SARS-CoV-2 infection in the low micromolar range in in vitro cell-based assays (Choy et al., 2020;Jeon et al., 2020).</p><!><p>Transcriptional data for SARS infection of human cells was obtained by querying NCBI GEO with relevant key words. In total 17 SARS-associated profiles have been generated sourced from two RNAseq series (GSE14507 (Webster et al., 2009) and GSE148729) and four microarray series (GSE17400 (Yoshikawa et al., 2010), GSE37827, GSE47963 (Mitchell et al., 2013), GSE48142). Profiles were generated based on grouping treatment and control samples and performing a student's t-test p < 0.05 filter. Profile correlation scores were generated by simple linear regression and measured with a Pearson correlation coefficient and significance given by the corresponding Z-score.</p><p>The relationship of the SISP to deposited transcriptional data was established through SPIED searches (www.spied.org.uk) of human and mouse samples. Pathway analysis was based on calculating the exact Fisher test enrichment of the up and down-regulated SISP genes for the Reactome dataset (Fabregat et al., 2017). Gene ontology (Carbon et al., 2009) enrichment analysis was performed in the same manner and based on gene sets downloaded from geneontology.org/.</p><p>CMAP profiles were generated as previously described (Williams et al., 2019). LINCS profiles were generated based on the deposited dataset series GSE92742 and GSE70138. Data is available in the form of Z scores for gene expression relative to plate average. The LINCS portal of SPIED hosts profiles in the form of categorical calls on the up/down status of genes generated based on combining expression data for drug/cell replicates. In the present work the LINCS profiles for different cell types are combined for each drug based on a given gene being assigned an up or down regulated status based on a majority vote across the cell types. Correlations of drug profiles against SISP were defined based on an exact Fisher test for the up/down category tables.</p><!><p>There is a high degree of correlation between the 17 profiles demonstrated by the significant positive Pearson correlations and the associated Z scores, see Table 1. The high degree of mutual similarity facilitates the definition of a composite signature for SARS infection based on at least 2/3 of genes being modulated in the same direction. This consists of 192 up regulated and 136 down regulated genes. The spectrum of expression changes for the genes across the profiles is shown in Table 2. The pathways and gene ontology classes associated with the up and down regulated genes are shown in Tables 3 and 4. As expected, the up regulated set is dominated by an immune response. In particular, pathways associated with interferon signalling, cytokine response and anti-viral mechanisms dominate the significantly enriched pathways. This is largely recapitulated with gene ontology class set enrichment analysis. In contrast, the down regulated gene set harbours fewer significantly enriched pathways and ontology classes, with the cell cycle set enriched as a pathway and ontology class.</p><!><p>The SISP profile can be examined in the wider context of deposited transcriptional data through search engines like SPIED (Williams, 2013). The results are shown in Table 5 and highlight a similarity to other virus induced gene expression changes. It is important to note here that this diversity in virus type with similar transcriptional effects upon diverse cell types is a reflection of the broad-brush nature of the signature. When it comes to delimiting candidate therapeutics based on the SISP it is to be expected that drugs with general, as well as specific, anti-viral properties will be returned. The repurposing methodology does not have the aim to discover drugs with specificity against novel targets, which is rather the aim of the traditional and protracted target-based novel chemical entity approach. In contrast, repurposing aims to harness the global phenotypic effects of drugs with the hope that these will tend to shift the underlying biology, in the present case by enhancing the cells natural defence response.</p><!><p>The top correlating drugs from CMAP and LINCS searches are listed in Table 6. There is considerable overlap in the CMAP and LINCS outputs, see column on the right of Table 6. As expected, there is a conspicuous enrichment for drugs with established anti-viral activities, these are highlighted in yellow.</p><p>In particular, there are 20 compounds with published antiviral activity in the top 45 CMAP hits and 6 of these compounds are also in the top 45 LINCS hits, which has an additional 16 compounds with established antiviral activity, see Table 6 for references. Furthermore, four compounds that are hits for both CMAP and LINCS, Emetine, Ouabain, Digoxin and Niclosamide, have each been found to display potent anti-SARS-CoV-2 activities. In the report from Choy et al., 2020, Emetine effectively inhibited SARS-CoV-2 replication in Vero cells at 0.5 μM, while in that same study the top hit in the LINCS set, homoharringtonine, inhibited the virus with an EC 50 of 2.10 μM (Choy et al., 2020). In a second report, (Jeon et al., 2020), Ouabain, Digoxin and Niclosamide each potently inhibited SARS-CoV-2 replication, again in Vero cells, with IC50's of 0.097 µM, 0.19 µM and 0.28 µM, respectively (Jeon et al., 2020). Also tested by Jeon et al., were two further drugs present at 4 th and 9 th position in the CMAP data set, Terfenadine and Mefloquine, which exhibited IC50's of 3.0 µM and 4.33 µM, respectively.</p><!><p>The SARS-CoV-2 pandemic is the most serious medical emergency since the Spanish flu of over 100 years ago. At present there are ongoing efforts to develop a vaccine and harness the increasing amount known about the mechanism of SARS-CoV-2 infection in terms of structure and protein targets to embark on a drug discovery pipeline. In the interim researchers have started to look at repurposing existing therapeutics, whether these be broad spectrum anti-viral drugs or more general therapeutics with indirect beneficial effects. The present paper seeks to highlight the effectiveness of transcriptional profiling to firstly define the activated pathways upon viral infection and secondly to use the infection associated expression changes as a quantitative phenotype to query drug databases for possible therapeutics. This approach has identified a broad panel of therapeutic entities ranging from plant extracts; the expectorant (and at higher doses, emetic) emantine from the rhizome of Carapichea ipecacuanha and the foxglove phytosteroid, digoxin, used for atrial fibrillation; through antibiotics such as puromycin, an aminonucleoside drived from the bacterium Streptomyces albiniger, through to compounds such as thioridazine, a first generation anti-psychotic and the synthetic thiazine dye, methylene-blue.</p><p>The cellular transcriptional response to viral infection has been assembled from multiple independent publicly available data sets and shown to harbour a consistent set of viral response genes that are transcribed in response to insult. This together with the extensive coverage of the drug associated transcriptome suggests a repurposing strategy where drugs tending to potentiate the cellular viral response are hypothetical candidate therapeutics. The results presented here largely validate this hypothesis. Two large drug driven transcription data sets were queried with the SISP viral response signature with the results that many drugs with established antiviral activities were high scoring candidates. This observation on the one hand offers a novel approach to identify drugs with no previous experimental data on their antiviral potential for trial, but also, reveals an intriguing unifying principle or at least effective biomarker underlying a disparate set of established antiviral agents.</p><p>That among the drugs identified by this approach six have already been shown, by others, to have potent anti-SARS-CoV-2 activities (Choy et al., 2020;Jeon et al., 2020), surely lends some validity of using transctiptional profiling as an useful way to rapidly identify effective treatment for the current pandemic and potentially for other viral pandemics that will inflict mankind in the future. While some of these "actives" may not readily lend themselves as treatments for Covid-19, such as the emetic, ementine, and ouabain, others, niclosamide and digoxin may. Given this high "hit rate" of true actives from the small number of our data set tested, it now warrants the rapid assessment of other candidates by those with access to SARS-CoV-2 assays with the view to testing the most suitable in clinical trials as a matter of urgency.</p>

ChemRxiv

Characterizing the Structure and Oligomerization of Major Royal Jelly Protein 1 (MRJP1) by Mass Spectrometry and Complementary Biophysical Tools

Royal jelly (RJ) triggers the development of female honeybee larvae into queens. This effect has been attributed to the presence of major royal jelly protein 1 (MRJP1) in RJ. MRJP1 isolated from royal jelly is tightly associated with apisimin, a 54-residue \xce\xb1-helical peptide that promotes the noncovalent assembly of MRJP1 into multimers. No high-resolution structural data are available for these complexes, and their binding stoichiometry remains uncertain. We examined MRJP1/apisimin using a range of biophysical techniques. We also investigated the behavior of deglycosylated samples, as well as samples with reduced apisimin content. Our mass spectrometry (MS) data demonstrate that the native complexes predominantly exist in a (MRJP14 apisimin4) stoichiometry. Hydrogen/deuterium exchange MS reveals that MRJP1 within these complexes is extensively disordered in the range of residues 20\xe2\x80\x93265. Marginally stable secondary structure (likely antiparallel \xce\xb2-sheet) exists around residues 266\xe2\x80\x93432. These weakly structured regions interchange with conformers that are extensively unfolded, giving rise to bimodal (EX1) isotope distributions. We propose that the native complexes have a \xe2\x80\x9cdimer of dimers\xe2\x80\x9d quaternary structure in which MRJP1 chains are bridged by apisimin. Specifically, our data suggest that apisimin acts as a linker that forms hydrophobic contacts involving the MRJP1 segment 316VLFFGLV322. Deglycosylation produces large soluble aggregates, highlighting the role of glycans as aggregation inhibitors. Samples with reduced apisimin content form dimeric complexes with a (MRJP12 apisimin1) stoichiometry. The information uncovered in this work will help pave the way toward a better understanding of the unique physiological role played by MRJP1 during queen differentiation.

characterizing_the_structure_and_oligomerization_of_major_royal_jelly_protein_1_(mrjp1)_by_mass_spec

<!>Sample Preparation<!>Optical Spectroscopy<!>Blue Native Polyacrylamide Gel Electrophoresis (BN-PAGE)<!>Native Top-Down Mass Spectrometry<!>Hydrogen/Deuterium Exchange Mass Spectrometry<!>Stoichiometry of Native MRJP1/Apisimin Complexes<!>Effects of Glycan and Apisimin on the Quaternary Structure of MRJP1<!>Characterization by Optical Spectroscopy<!>Hydrogen/Deuterium Exchange<!>Hydropathy Analysis<!>Implications for the Native MRJP1/Apisimin Complex<!>CONCLUSIONS

<p>Adult honeybees (Apis mellifera) form three castes. Females can develop into queen or worker bees, while males are known as drones.1,2 Each colony has only a single queen that mates with drones and lays eggs. Worker bees are sterile but perform numerous tasks, including foraging for food and feeding of larvae.3 The dimorphism of female bees is governed by the nutritional environment experienced by the larvae.4–7 Future queens are fed royal jelly (RJ) throughout their development. Worker larvae receive RJ only for an initial period of 3 days, after which their food is switched to worker jelly.1,4 Male larvae receive drone jelly. The various jellies differ in their protein composition, as well as their fructose:glucose ratio and vitamin content.8–10</p><p>RJ is secreted by the hypopharyngeal and mandibular glands of worker bees at an early stage of their life cycle, during which they are members of the nurse subcaste.10 RJ contains proteins, carbohydrates, lipids, vitamins, salts, and free amino acids.11 The so-called major royal jelly proteins (MRJPs) constitute roughly 90% of the total RJ protein.9 This family comprises nine homologous members, designated MRJP1–9.10,12,13 The high percentage of essential amino acids in MRJPs underscores their nutritional role in bee larvae.9,14</p><p>The ability of RJ to modulate the development of female larvae may be partially related to the presence of histone deacetylase inhibitors,4 microRNAs,3 and other factors.10 However, the key ingredient that drives queen development appears to be MRJP1, the most abundant protein in RJ.15 Although the exact role of MRJP1 remains under dispute,16 there is strong evidence that this protein triggers queen differentiation via an epidermal growth factor receptor (EGFR)-mediated signaling pathway.15,17 MRJP1 also exhibits antibacterial effects,18 as well as antihypertension19 and growth factor-like activity in mammalian cells.10</p><p>MRJP1 is expressed as a chain consisting of 432 amino acids. Subsequent cleavage by a signal peptidase removes an N-terminal 19-residue segment,9 resulting in a chain with an expected molecular weight (MW) of 46861 Da. During maturation, MRJP1 undergoes additional post-translational modifications.20 Isoelectric focusing reveals the presence of nine MRJP1 isoforms that share a similar MW but have slightly different pIs between 4.7 and 5.2.21,22 Differences in the nature and extent of post-translational modifications are thought to be chiefly responsible for this heterogeneity.9</p><p>Mature MRJP1 is a glycoprotein, and its glycans have been analyzed in great detail.23,24 Sugars are bound mainly at N144 and N177, although other attachment sites also exist.24 The glycans were shown to include a unique Galβ1–3GalNAc unit.23 Deglycosylation by PNGase F causes a mass shift on sodium dodecyl sulfate (SDS) gels from an apparent MW of 56 kDa to an apparent MW of 47 kDa,25 the latter being consistent with the amino acid sequence of the mature protein.9</p><p>When isolated from royal jelly, MRJP1 is copurified with the 54-residue (5.54 kDa) α-helical peptide apisimin.26 Apisimin promotes the association of MRJP1 into higher-order structures,11,26 producing MRJP1/apisimin complexes that exhibit considerable thermal stability.22,27 The composition of these complexes remains uncertain.10 Electrophoretic densitometry suggested a 5:1 MRJP1:apisimin stoichiometry,11 but other binding ratios could not be ruled out. Chromatographic, electrophoretic, and light scattering studies yielded size estimates of 280–420 kDa for these MRJP1/apisimin assemblies.11,22,26 Binding is mediated solely by noncovalant interactions, as intermolecular disulfide bridges are absent.28</p><p>Structural investigations of MRJP1/apisimin complexes are a prerequisite for gaining a better understanding of the unique role that MRJP1 plays during queen development.15 Unfortunately, it has not been possible thus far to generate high-resolution conformational data for this system.10 In the work described here, we applied different mass spectrometry (MS) techniques and various other biophysical tools to close some of the existing knowledge gaps. Our primary goal was to elucidate aspects of the MRJP1/apisimin structure and dynamics under native conditions. In addition, we examined how the formation of higher-order complexes depends on the presence of apisimin (A) and glycan chains (G). We compared the behavior of native MRJP1 in the presence of apisimin ("A+") and with glycans attached ("G+") to that of samples that had undergone deglycosylation ("G−") as well as partial apisimin removal (denoted as "A−"). The four types of samples generated in this way are denoted A+G+, A−G+, A−G−, and A+G−. Our data allow us to propose simple structural models for the complexes encountered under these different conditions.</p><!><p>Protein samples were prepared in 50 mM HEPES (pH 7.5), unless otherwise noted. The MRJP1/apisimin complex was purified as described previously,22 resulting in stock solutions containing 50 μM protein. A subset of these A+G+ samples was subjected to apisimin depletion. Size exclusion and ion exchange methods have previously been shown to be ineffective for separating apisimin from MRJP1, pointing to high-affinity noncovalent interactions.26 The presence of intermolecular disulfide bonds can be ruled out because apisimin does not contain Cys residues.26 We thus attempted to separate apisimin from MRJP1 by dialysis. Initial tests during which the MRJP1/apisimin complex was dialyzed against native buffer or 7 M urea solutions did not result in any appreciable change in the apisimin:MRJP1 ratio, as judged by ESI-MS signal intensities after separation on a reversed phase column (BEH300 C4, 1.7 μm, 2.1 mm × 50 mm, Waters, Milford, MA). A more effective strategy for apisimin depletion was dialysis against dilute aqueous formic acid solutions at pH 2 for 14 days on ice [using molecular weight cutoff (MWCO) 20000 cassettes from Thermo, Waltham, MA]. This procedure reduced the apisimin:MRJP1 ratio by approximately 50% relative to that of the original A+G+ samples. The apisimin-depleted protein was exchanged back into nondenaturing HEPES at pH 7.5 and equilibrated for 24 h on ice prior to further analyses. The samples obtained in this way are denoted "A−G+". MRJP1 deglycosylation (with or without apisimin depletion) was performed using PNGase F (Promega, Madison, WI),25 producing deglycosylated samples A−G− and A+G−. Quantitative glycan removal was confirmed using SDS–polyacrylamide gel electrophoresis (PAGE).25</p><!><p>CD data were recorded on a Jasco (Easton, MD) J-810 spectropolarimeter using a 1 mm optical path length. Protein-free blanks were subtracted from the reported data. Secondary structure analysis was performed using the Spectra Manager software package supplied by the instrument's manufacturer. Fluorescence data were acquired using a Jasco FP-6500 spectrofluorimeter with an excitation wavelength of 285 nm.</p><!><p>For BN-PAGE,29,30 protein samples were transferred into 50 mM bis-Tris-HCl buffer containing 15% (w/v) glycerol. Separation was performed on 5 to 18% (w/v) polyacrylamide gradient gels that were run at 15 mA and 4 °C for 4 h in an SE 600 electrophoresis system (Hoefer, San Francisco, CA). The anode buffer consisted of 50 mM bis-Tris HCl, while the cathode buffer consisted of 50 mM Tricine, 15 mM bis-Tris, and 0.02% (w/v) Coomassie brilliant blue G-250 (Bio-Rad, Hercules, CA). Thyroglobulin (669 kDa), ferritin (440 kDa), catalase (232 kDa), lactate dehydrogenase (158 kDa), and bovine serum albumin (66 kDa) were used as calibrants. Apparent molecular weights (MWapp) were determined from using distance of migration versus log(MW) plots.29</p><!><p>Native ESI-MS was performed as described previously31 using a modified Orbitrap Q-Exactive instrument (Thermo Fisher, Bremen, Germany). Fragment ion spectra were calibrated internally and matched manually with a tolerance of 10 ppm using mMass.32 MRJP1 samples were washed three times in a 100 mM neutral ammonium acetate solution, using Amicon filters with a MWCO 30000 membrane (Millipore, Darmstadt, Germany) prior to infusion. ESI was performed using a sheath-flow capillary electrophoresis device operated at a sprayer voltage of 1.5 kV.31,33 Prosight PC 3.0 (Thermo Fisher) and Prosight Lite were used for data analysis.</p><!><p>MRJP1 samples (5 μM) were mixed with D2O-based labeling buffer in a 1:9 volume ratio at 22 °C. Aliquots of 200 μL were removed after 10 s, 1 min, 10 min, and 100 min and reactions quenched by adding an equal volume of an ice-cold solution containing 0.675 M guanidinium chloride, 1% formic acid, and 15 μL of a 3 mg mL−1 pepsin solution (final pH of 2.5). Offine pepsin digestion took place on ice for 1 min. The samples were flash-frozen in liquid nitrogen and stored at −80 °C prior to analysis. The aliquots were then rapidly thawed to ~0 °C and manually injected into a nanoACQUITY UPLC instrument with HDX technology (Waters). Desalting and peptide separation were performed at 0 °C within 25 min on an equilibrated reversed phase column (BEH C8, 1.7 μm particle size, 1 mm × 100 mm) using a water/acetonitrile gradient with 0.1% formic acid at a rate of 40 μL min−1. Analysis was performed on a Waters Synapt G2 mass spectrometer. Zero-time controls (m0) for the correction of in-exchange were performed by exposing MRJP1 to quenching buffer, followed by D2O exposure, resulting in the same final solution composition that was seen for all other samples. Controls for fully exchanged MRJP1 (m100, for the correction of back exchange) were prepared by incubating 5 μM MRJP1 in a labeling solution at pH 2.0 and 70 °C for 12 h. Normalized deuteration levels are reported as (mt − m0)/(m100 − m0) × 100%, where mt represents the centroid mass of the peptide of interest after HDX for time t. Extensive washing and blank injections were used between individual runs to prevent carryover, thereby eliminating false EX1 artifacts.34 Peptide identification was performed using MSE and PLGS version 2.5.3 (Waters). HDX data were analyzed using DynamX version 3.0 (Waters). All experiments were conducted in triplicate</p><p>Thirty-four peptic peptides could be consistently observed with adequate signal/noise (S/N) ratios across all four types of samples. The HDX/MS peptic digestion map is depicted in Figure S1, using the standard residue numbering of full-length MRJP1.9 Only peptides that were observed under all four conditions are shown, providing a sequence coverage of 55%. This relatively low coverage is partially attributed to the presence of glycans in G+ samples, which precluded the detection of peptides in the vicinity of glycosylation sites N144 and N177.24 Control experiments using trypsin yielded a sequence coverage of 64%, only slightly higher than that for pepsin (Figure S2). The sequence stretches covered by trypsin mapping resemble those detected after pepsin digestion. The limited sequence coverage seen with both proteases is consistent with the fact that MRJP1 experiences various post-translational modifications9,20 in addition to glycosylation.24 Any such covalent modifications will interfere with peptide matching, which relies on comparisons with the cDNA-derived MRJP1 sequence.9 Analyses of the apisimin HDX behavior were not possible because of the low signal intensity of the corresponding peptides.</p><!><p>As a first step toward a comprehensive characterization of the MRJP1/apisimin system, we determined the composition of unmodified (A+G+) samples. Native ESI-MS35–38 of A+G+ yielded a series of peaks, corresponding to MRJP1/apisimin complex ions with an average mass of 231.88 kDa and charge states of 27+ to 32+ (Figure 1a). Collisional activation of 30+ precursor ions released monomeric species with an average mass of 51.97 kDa and charge states of 12+ to 20+ (Figure 1b). This dissociation behavior is similar to asymmetric charge partitioning events seen for other complexes.35 The spectrum in Figure 1b also shows an intense apisimin2+ signal, consistent with the expected 5.54 kDa mass of the peptide. Activation of the intact complex (m/z 7000–8000) in the HCD39 cell yielded a remarkably clean spectrum with three dominant peaks corresponding to apisimin 1+ to 3+ (Figure 1b, inset). Pseudo-MS3 was implemented by subjecting monomeric MRJP112+ to HCD, generating fragment ions that were matched to the expected sequence9 (Figure 1c). The matched fragments correspond to cleavages within the 35 N-terminal and 16 C-terminal residues, consistent with reports that glycosylation24 and other post-translational modifications9 affect only residues closer to the center of the sequence.</p><p>The data depicted in Figure 1 unambiguously reveal the binding stoichiometry of the MRJP1/apisimin complexes in A +G+ samples. The measured mass (231.88 kDa) represents 4 times the measured MRJP1 mass plus 4 times the mass of apisimin (4 × 51.97 kDa + 4 × 5.54 kDa = 230.04 kDa). The slight difference between the two values (231.88 kDa vs 230.04 kDa) likely reflects the loss of weakly bound nonspecific adducts during collisional activation, as previously reported for other protein complexes.35 Overall, we conclude that native A +G+ samples contain a substantial fraction of complexes with a (MRJP14 apisimin4) stoichiometry. The ESI-MS-derived mass is in reasonable agreement with estimates of ~290 kDa that had been obtained using light scattering.22 We do not rule out the possibility that larger aggregates may form under mildly denaturing conditions (e.g., at pH 9), as reported previously.22,26</p><p>The measured monomer mass (51.97 kDa) significantly exceeds that expected from the MRJP1 amino acid sequence (46.86 kDa).9 This 5.11 kDa difference is attributed largely to MRJP1 glycosylation.23–25 Glycosylation usually causes mass heterogeneity,40,41 consistent with the substantial width (full width at half-maximum of ≈1 kDa) of the spectral signals in Figure 1b. Other post-translational modifications may contribute to MRJP1 mass heterogeneity, as well.9,20</p><p>In addition to (MRJP14 apisimin4), native ESI-MS reveals the presence of monomeric MRJP1 in A+G+ samples (Figure S3). Unfortunately, detection biases and differences in ionization efficiency make it difficult to estimate the molar ratio of monomeric MRJP1 to (MRJP14 apisimin4) in solution from these ESI-MS data.38,42</p><!><p>BN-PAGE is a complementary approach for monitoring biomolecular interactions. The association with Coomassie dye imparts negative charge to quasi-native proteins within the gel, allowing their electrophoretic separation while preserving interactions with binding partners.29,30 BN-PAGE was particularly useful for interrogating the properties of MRJP1 samples after apisimin depletion (A−) and after glycosylation (G−), because it was difficult to generate high-quality native ESI mass spectra (not shown) under those conditions. Hence, we applied BN-PAGE to MRJP1 samples of the type A+G+, A−G+, A−G−, as well as A+G− (Figure 2).</p><p>A+G+ samples exhibit an intense band at a MWapp of ≈287 kDa corresponding to the native (MRJP14 apisimin4) complex. In addition, these samples show monomeric protein at a MWapp of ≈55 kDa. These data are in agreement with the native ESI-MS data depicted in Figure 1.</p><p>A−G+ and A−G− samples both exhibit BN-PAGE bands at a MWapp of ≈86 kDa. To interpret these bands, we recall two points that became apparent during sample preparation (as noted above). (i) Apisimin exhibits a a very high binding affinity for MRJP1.26 (ii) As a result of this high affinity, apisimin depletion in our "A−" samples was incomplete. The apisimin:MRJP1 ratio of these preparations was reduced by ~50% compared to that of the the original A+G+ samples. Accordingly, it seems likely that the MWapp ≈ 86 kDa bands predominantly correspond to (MRJP12 apisimin1) assemblies. The measured MWapp is lower than expected for such 2:1 complexes (theoretical MWs of 109.5 kDa for A−G+ and 99.3 kDa for A−G−). However, such deviations are quite common in BN-PAGE, where exposed hydrophobic areas can favor excessive Coomassie binding. Such conditions cause elevated electrophoretic mobilities that give rise to abnormally low MWapp values.43 A+G− samples form large assemblies that cannot migrate into the gel (last lane in Figure 2).</p><p>In summary, the combination of native ESI-MS and BN-PAGE yields a consistent picture of the binding stoichiometries for the different types of samples studied here. A+G+ samples contain (MRJP14 apisimin4) complexes, as well as MRJP1 monomers. A−G+ and A−G− conditions give rise to the formation of (MRJP12 apisimin1) complexes. The prevalence of dimeric MRJP1 assemblies in these perturbed samples suggests that native (MRJP14 apisimin4) possesses a "dimer of dimers" architecture. A similar situation is encountered for many other tetrameric proteins, where two pairs of tightly bound chains come together to form the overall complex.44–46 Large assemblies of the type (MRJP1m apisiminn), where m ≫ 4 and n ≫ 4, are encountered for A+G− samples.</p><p>Our data support the view11,26 that apisimin promotes the noncovalent association of MRJP1. Apisimin depletion shifts the binding stoichiometry from species containing four MRJP1 chains to assemblies that contain only two MRJP1 chains (Figure 2). After deglycosylation, the presence of apisimin triggers the formation of large aggregates [A+G− (Figure 2)]. The fact that these large assemblies form only after glycan removal points to the role of glycans as aggregation inhibitors, in line with reports for other proteins.47–49</p><!><p>Far-UV CD spectra provide insights into secondary structure50 (Figure 3a). A+G+ samples exhibited a main minimum at 208 nm. Deconvolution of this spectrum suggests 47% antiparallel β-sheet and 28% random coil, with the remainder being due to α-helices, parallel β-sheet, and β-turns. Apisimin depletion and/or deglycosylation caused only minor changes in the CD spectra. A+G− samples displayed a small shift in the position of the CD minimum from 208 to 207 nm, suggesting a 5% reduction in antiparallel β-sheet content and a subtle (2%) increase in random coil character. While these percentages represent only semiquantitative estimates,51,52 Figure 3a nonetheless indicates that MRJP1 exhibits a relatively high degree of disorder under all conditions studied.</p><p>MRJP1 possesses five Trp residues. The Trp emission maximum for A+G+ was at 335 nm (Figure 3b). The other three types of samples showed spectra that were slightly red-shifted, with maxima of 338 nm (A−G+), 341 nm (A−G−), and 340 nm (A+G−). These UV emission properties report on the environment of Trp side chains.53 The observed spectral shifts relative to A+G+ indicate structural changes from a hydrophobically buried environment to a conformation in which Trp side chains become slightly more solvent accessible.54,55 Thus, some hydrophobic contacts that exist in native MRJP1 (A+G+) become disrupted after apisimin depletion and/or deglycosylation.</p><p>The spectroscopic data of Figure 3 also provide insights into the nature of the large (MRJP1m apisiminn) assemblies formed under A+G− conditions. The CD and fluorescence signatures of these species remain quite similar to those seen for the other sample types. This behavior reflects the fact that the A+G− assemblies do not precipitate. Insoluble precipitates would show spectra with greatly reduced signal amplitudes and diminished S/N ratios.56,57 In other words, our data imply that A+G− conditions produce soluble aggregates, resembling the behavior seen for several other proteins under mildly denaturing conditions.58,59</p><!><p>HDX/MS is a sensitive tool for examining how protein structure and dynamics respond to changes in biomolecular interactions or other external factors.45,60–62 HDX/MS measures the mass increase that results from the deuteration of backbone NH groups in D2O-based labeling buffer. At near-neutral pH, disordered segments undergo exchange with rate constants (kch) on the order of ≈1 s−1.63–65 HDX rates are greatly reduced in well-structured regions that are stabilized by backbone hydrogen bonds. These folded regions usually undergo deuteration in the EX2 regime, which is characterized by isotope envelopes that gradually shift toward higher masses. EX2 behavior is indicative of opening/closing transitions that take place on a time scale much faster than 1 s−1.63 Conversely, correlated opening/closing events much slower than 1 s−1 give rise to bimodal isotope distributions (commonly termed EX1 exchange).61,66,67</p><p>Representative HDX kinetic plots are depicted in Figure 4 (see Figure S4 for the complete data set). Most peptides exhibit deuteration levels greater than 50% already at the earliest labeling time point of 10 s. After 100 min, all peptides approach complete deuteration. This is in contrast to the behavior of many other proteins, which possess well-folded regions that are highly protected against HDX.45,68,69 The across-the-board rapid deuteration seen here for MRJP1 under the various experimental conditions is reminiscent of data previously reported for intrinsically disordered proteins.64,65,70,71</p><p>To compare the properties of A+G+, A−G+, A−G−, and A +G− in a comprehensive fashion, we will focus on deuteration levels for t = 1 min (Figure 5a–d). The deuteration patterns of the four samples show many similarities. With a fully deuterated N-terminal region as a starting point, the HDX levels decline to values of ~60% for segment 69–77, suggesting some weak hydrogen bonding in this region, possibly in combination with sequestration of NH groups in the protein interior. Between residues 91 and 265, MRJP1 is highly disordered with deuteration values close to 100% for A+G+, A−G−, and A +G−, while slight protection is seen in this range for A−G+ (Figure 5b). Protection is most pronounced for all four samples beyond residue 265, with many deuteration values between 25 and 75%. We conclude that the C-terminal region comprising residues 266–432, while still being quite dynamic, is the most structured part of MRJP1.</p><p>The subtle HDX changes caused by apisimin depletion and deglycosylation are best visualized by resorting to difference plots, using the A+G+ data as a reference. Negative values for A−G+ in the range of residues 202–302 reveal that A−G+ is slightly more protected in this region than A+G+ is (Figure 5e). The HDX properties of A−G− are quite similar to those of A+G+, which is evident from difference values close to zero throughout the entire sequence range (Figure 5f). The largest changes are seen for A+G−, where the C-terminal region of residues 266–369 exhibits deuteration levels significantly elevated compared to those of A+G+ (Figure 5g).</p><p>To gain additional insights into the weak HDX protection observed under the four conditions, it is instructive to look at unprocessed mass spectra. Remarkably, all MRJP1 regions that showed incomplete deuteration exhibited bimodal isotope distributions, implying that HDX proceeds in the EX1 regime for all four conditions (Figure 6).61,66,67 Already after 10 s, the EX1 distributions showed a well-developed high-mass component, with an amplitude on the order of 30–60% for most peptides. This behavior reveals that all of the MRJP1/apisimin samples exist as partially structured species that are in equilibrium with conformers that are more unfolded. In addition, the reversible dissociation of higher-order structures into smaller building blocks (e.g., tetramer ↔ monomer transitions in the case of A+G+) could contribute to the observed EX1 behavior. The level of EX1 low-mass components diminished over time but remained detectable even after 100 min. These slow kinetics imply that the interconversion of partially structured species with more unfolded conformers takes place on a time scale of several hours.61,66,67</p><!><p>Most water-soluble proteins fold into structures in which nonpolar side chains are buried in the core, while hydrophilic residues remain solvent accessible.72,73 Our data demonstrate that MRJP1/apisimin complexes are quite disordered, without a well-developed core. However, it is known that even for such disordered proteins some clustering of nonpolar residues can take place, specifically in regions that are involved in intermolecular contacts.74,75</p><p>To identify possible MRJP1 regions that might show nonpolar clustering, we conducted a Kyte–Doolittle analysis,76 using the standard scale that ranges from −4.5 for Arg as the most hydrophilic residue to +4.5 for Ile as the most hydrophobic. The average hydropathy of MRJP1 is −0.44 ± 1, reflecting the low percentage of nonpolar residues in this protein. The 316VLFFGLV322 segment stands out as the most hydrophobic region, which is evident from the prominent spike centered at residue 319 in Figure 7a. Interestingly, this region is most protected against deuteration in unmodified MRJP1 (Figures 4d and 5a). As in the case of other disordered proteins,74,75 MRJP1 therefore contains at least one hydrophobic region that is relatively structured.</p><p>In contrast to MRJP1, apisimin is quite hydrophobic with an average hydropathy of 0.95 ± 0.9 (Figure 7b). This nonpolar character is particularly pronounced for the C-terminal half of the peptide (25IVS…VFA54) that has a hydropathy of 1.4 ± 0.5.</p><!><p>The establishment of nonpolar contacts generally represents the dominant driving force for the formation of protein–protein interactions.77,78 This is particularly the case for intrinsically disordered proteins, in which the corresponding nonpolar regions tend to be among the most structured elements.74,75 Accordingly, the MRJP1 316VLFFGLV322 region is a prime candidate for the formation of intermolecular contacts because this segment exhibits the highest hydrophobicity (Figure 7a). The 316VLFFGLV322 region was not completely covered in our HDX experiments. However, the partially overlapping segment 303–318 showed the strongest protection in MRJP1 (Figures 4d and 5a), supporting the view that this region is a key binding element.</p><p>Our data (Figure 2) as well as earlier work11,26 demonstrate that the association of MRJP1 complexes is mediated by apisimin. The native ESI-MS data of Figure 1 uncovered the fact that these complexes possess a (MRJP14 apisimin4) stoichiometry, suggesting that apisimin acts as linker that binds MRJP1 monomers together. The hydrophobic nature of apisimin (Figure 7b) implies that these linkages are dominated by nonpolar contacts. It seems likely that linkages within the native (MRJP14 apisimin4) complex involve nonpolar contacts between the C-terminal half of apisimin (25IVS…VFA54) and the MRJP1 316VLFFGLV322 segment (Figure 7).</p><!><p>Previous efforts to generate high-resolution structural data for MRJP1 were unsuccessful. In the study presented here, we applied a range of biophysical techniques, each of which provides insights into specific aspects of protein conformations and interactions. The information obtained allows us to propose simple structural models of MRJP1/apisimin complexes that are encountered for the four types of samples studied here. All of the species (A+G+, A−G+, A−G−, and A+G−) are relatively disordered, and they undergo slow (EX1) interconversion with coexisting species that are even more unfolded.</p><p>Native A+G+ conditions favor the presence of (MRJP14 apisimin4) complexes (Figure 8a). The MRJP1 chains within these complexes are largely unstructured in the range of residues 20–265; this includes glycan attachment sites N144 and N177. Residues 266–432 adopt a marginally stable fold that may be organized in an antiparallel β-sheet conformation, as suggested by the CD data of Figure 3. The 316VLFFGLV322 hydrophobic segment is among the most structured segments. Apisimin acts as an α-helical linker that binds the MRJP1 chains together. A key element of these interactions are nonpolar contacts between 316VLFFGLV322 and the hydrophobic C-terminal portion of apisimin (25IVS…VFA54). In addition, direct contacts between adjacent MRJP1 chains cannot be ruled out. The four MRJP1 molecules are assembled in a "dimer of dimers" architecture, implying the presence of two tight subunit–subunit interfaces and two loose interfaces (corresponding to the horizontal and vertical contacts in Figure 8a, respectively). The relatively low ESI charge states formed in the native solution (Figure 1a) indicate that despite their disorder, the overall structure of the (MRJP14 apisimin4) complexes is quite compact.79</p><p>In A−G+ and A−G− samples, the apisimin concentration is reduced by roughly 50% relative to those of the A+G+ preparations. Linker depletion causes dissociation of (MRJP14 apisimin4) into (MRJP12 apisimin1) subcomplexes, as indicated in panels b and c of Figure 8. This dissociation causes only minor changes in the overall conformational properties of the MRJP1 chains. However, Trp side chains become more solvent accessible, and newly exposed interfaces cause excessive Coomassie dye binding, consistent with the disruption of protein–protein contacts.</p><p>The glycan chains at residues N144 and N177 play a role in controlling the degree to which apisimin can cause association of MRJP1 into larger aggregates. For other proteins, it has been demonstrated that glycans can inhibit aggregation, because steric clashes of the heavily hydrated sugar chains interfere with self-association of the protein.47–49 In the case of MRJP1, deglycosylation results in apisimin-mediated association into large soluble assemblies of the type (MRJP1m apisiminn) where m ≫ 4 and n ≫ 4. These A+G− conditions destabilize previously existing backbone hydrogen bonds in the range of residues 266–432 (Figure 8d).</p><p>From the proposed structural models of Figure 8, uncovering the exact mechanism by which MRJP1 might tap into EGFR-mediated events that trigger the development of bee larvae into queens still requires a great deal of work.15–17 With the nutritional role of MRJP1 in mind,9,14 the semiunfolded nature of (MRJP14 apisimin4) complexes in RJ may be important for ensuring efficient hydrolysis by proteases in the larval digestive tract.80 Interestingly, proteomic analyses have uncovered significant differences in the makeup of the gastrointestinal enzymes of queens relative to that of worker bees and drones.81 It seems likely that queen development is not triggered by the intact (MRJP14 apisimin4) complexes per se, but by proteolytic products that are generated in the digestive tract and absorbed into the hemolymph.</p><p>Considering the disordered nature of native MRJP1/apisimin complexes, it is not surprising that crystallization efforts (in M.V.d.S.'s laboratory, unpublished observations) were unsuccessful. In previous work on other proteins, HDX/MS was applied to pinpoint regions of disorder, thereby allowing the design of "disorder-depleted" constructs that then yielded high-quality crystals.82 Our HDX/MS experiments identify residues 20–265 as the most disordered region within native MRJP1/apisimin assemblies. In future studies, it will be interesting to perform crystallization trials on constructs where this N-terminal part of MRJP1 has been deleted. Our data predict that such constructs should still be able to form tetrameric assemblies with apisimin (Figure 8a), because we believe the apisimin binding site to be in the vicinity of residue 320.</p>

PubMed Author Manuscript

Formins Regulate Actin Filament Flexibility through Long Range Allosteric Interactions*

The members of the formin family nucleate actin polymerization and play essential roles in the regulation of the actin cytoskeleton during a wide range of cellular and developmental processes. In the present work, we describe the effects of mDia1-FH2 on the conformation of actin filaments by using a temperature-dependent fluorescence resonance energy transfer method. Our results revealed that actin filaments were more flexible in the presence than in the absence of formin. The effect strongly depends on the mDia1-FH2 concentration in a way that indicates that more than one mechanism is responsible for the formin effect. In accordance with the more flexible filament structure, the thermal stability of actin decreased and the rate of phosphate dissociation from actin filaments increased in the presence of formin. The interpretation of the results supports a model in which formin binding to barbed ends makes filaments more flexible through long range allosteric interactions, whereas binding of formin to the sides of the filaments stabilizes the protomer-protomer interactions. These results suggest that formins can regulate the conformation of actin filaments and may thus also modulate the affinity of actin-binding proteins to filaments nucleated/capped by formins.

formins_regulate_actin_filament_flexibility_through_long_range_allosteric_interactions*

<!>Protein Preparations and Modifications<!>Fluorescence Experiments<!>Differential Scanning Calorimetry (DSC)<!>Determination of the Phosphate Release Rate<!>Co-sedimentation Assays<!>RESULTS<!>Temperature-dependent FRET Can Be Applied for Examining the Effects of Formins on Actin Filaments<!>mDia1-FH2dimer Induces Conformational Changes in Actin Filaments<!>Formin Decreases the Thermal Stability of Actin<!>Formin Accelerates the Dissociation of Phosphate from Actin<!>Cofilin Depolymerizes Actin More Effectively in the Presence of Formin<!>The Interpretation of the FRET Results<!>The Effect of mDia1-FH2dimer on Actin Filaments<!>A Model for the Interaction of Actin and mDia1-FH2dimer<!>A Possible New Mechanism to Regulate the Formation of Cytoskeletal Complexes<!>Conclusions

<p>Formins are evolutionarily conserved proteins (1, 2) that activate signaling pathways and nucleate actin filaments independently of the Arp2/3 complex (3-6). In mammalian cells, formins play a role in the formation of stress fibers, cell motility, signaling, gene transcription, and embryonic development (7-13). In yeast, formins organize cytoplasmic actin cables and the contractile ring (1, 3, 14-17). Formins are composed of multiple domains (2), which can include formin homology domains (18) (FH1, FH2, FH3), N-terminal GTPase-binding domain (GBD),3 and C-terminal diaphanous-autoregulatory domain (DAD). FH1 and FH2 domains are present in all formins (15). The proline-rich FH1 can interact with profilin, with factors involving the SH3 domain and the Src family kinases (9, 14, 17, 19, 20). The FH2 domain is required for the interaction with actin, for the stabilization of microtubules, and for serum response factor activation (5, 9, 12, 21). Diaphanous-related formins involve GBD and DAD domains (22). In some diaphanous-related formins, binding of activated Rho relieves intramolecular interactions between the DAD and N-terminal sequences (19, 23).</p><p>Biophysical characterization of formin fragments from mammalian sources (from mouse, mDia1 (4, 24-26) and mDia3 (25)), from Saccharomyces cerevisiae (Bni1p and Bnr1p) (3, 24), and from Schizosaccharomyces pombe (Cdc12p) (27) established that they were potent actin nucleators in vitro and that the FH2 domain was essential for the nucleation. Recent structural studies have given insights into the molecular mechanisms responsible for the formin functions. The structures of the FH2 domains from mDia1 (25), from Bni1p (28), from the complex of actin with Bni1-FH2 (29), and from the complex of the GBD from mDia1 and Rho A (30) have been determined so far.</p><p>Despite the numerous studies characterizing the effect of formin fragments on the polymerization properties of actin, very little is known about the effect of formins on the conformation of actin. One of the few related observations was that the affinity of CapG protein for actin filaments decreased by a factor of ~100 in the presence of mDia1-FH2 fragments (26), and the weaker affinity was not due to the direct competition of the two proteins. This observation indicates that mDia1-FH2 (and also the FH1-FH2 fragment) may induce conformational changes at the barbed end of actin filaments. Supporting the existence of the formin effect on actin, the recently solved structure of the complex of Bni1p-FH2 and actin showed that a special conformational strain was generated when formin bound to the actin structure mimicking the barbed end (29).</p><p>In this work, we studied the effects of mDia1-FH2 fragments on the conformational properties of actin filaments by using temperature-dependent fluorescence resonance energy transfer (FRET) measurements. The results revealed that mDia1-FH2 fragments increased the flexibility of actin filaments. The magnitude of the effect of formin depended on the formin:actin concentration ratio. The characteristics of the formin concentration dependence indicated that more than one mechanism was involved in the formin effects. The results were interpreted assuming that formin binding to the barbed end made the actin filaments more flexible. This formin effect propagated through allosteric interaction to actin protomers far from the barbed end. At greater formin concentrations, the binding of formin fragments to the sides of the actin filaments could stiffen the interaction between neighboring actin protomers along the filaments. Based on these observations, we hypothesize that actin filaments serve as information channels in vivo. The formin-induced conformational changes in actin filaments can play a role in modifying and regulating the affinity of various actin-binding proteins for actin. This possible new mechanism could operate in the case of other actin nucleating factors as well and may provide a novel mechanism to regulate the formation of cytoskeletal protein complexes.</p><!><p>A core formin mDia1-FH2 fragment (amino acids 826–1163) and a longer mDia1-FH2 fragment (amino acids 752–1163) involving the linker region between the FH1 and FH2 domains was prepared as described in Ref. 25. The former is a monomeric protein (25) and assigned in this study as mDia1-FH2monomer, whereas the latter is a dimer and indicated as mDia1-FH2dimer throughout the study. After the preparation, the formin fragments were kept in −80 °C in storing buffer (50 mm Tris/HCl, pH 7.6, 50 mm NaCl, 5 mm dithiothreitol, 5% glycerol). Mouse cofilin-1 was purified as described in Ref. 31. Actin from rabbit skeletal muscle was prepared as described previously (32, 33). The results presented in this work were obtained using five independent formin and more than 10 independent actin preparations. Further purification of actin by a Sephacryl 300 column did not affect the results of our experiments (data not shown). The concentrations of actin and formin are given as monomer concentrations throughout the study. The concentration of actin was determined by using the extinction coefficient of 0.63 mg−1 ml cm−1 at 290 nm (34). The mDia1-FH2monomer and mDia1-FH2dimer concentrations were determined by measuring the absorption at 280 nm in 6 m GuHCl using the extinction coefficients of 7,680 m−1 cm−1 and 20,580 m−1 cm−1, which were estimated with ProtParam from the sequences (ExPASy Proteomics tools). Actin was labeled with pyrene (35), IAEDANS, or IAF (36, 37) on the Cys374 according to standard procedures. The concentrations of the probes were determined using the absorption coefficients of 22,000 m−1cm−1 at 344 nm for pyrene, 6,100 m−1 cm−1 at 336 nm for IAEDANS, and 60,000 m−1 cm−1 at 495 nm for IAF. The labeling ratio, i.e. the molar ratio of the bound probe to the actin concentration, was 0.9–1.0 for pyrene, 0.8–0.9 for IAEDANS, and 0.6–0.7 for IAF.</p><!><p>Monomeric calcium-actin was in 4 mm Tris/HCl, pH 7.3, 0.2 mm ATP, 0.5 mm dithiothreitol, and 0.1 mm CaCl2 after preparation. The bound calcium was replaced with magnesium by adding 200 μm EGTA and 50 μm MgCl2 and incubating the samples for 5–10 min. The magnesium-actin was polymerized with 0.5 mm MgCl2 and 10 mm KCl in either the presence or the absence of formin fragments. The actin concentration was 5 μm unless stated otherwise.</p><p>IAEDANS served as donor and IAF served as acceptor in the FRET experiments. Actin monomers separately labeled with either donor or acceptor were mixed in a 1:9 = donor-labeled:acceptor-labeled actin ratio. For FRET, reference samples were also prepared with 10% donor-labeled and 90% unlabeled actin. Then MgCl2 and KCl were added to the monomer actin to initiate polymerization, and the samples were incubated overnight at 4 °C before the fluorescence experiments. The volume of the storing buffer, regardless of whether it was added with or without formin, was kept constant (5%) in the samples.</p><p>To determine the efficiency (E) of the FRET, the fluorescence emission of the donor was measured in the absence (FD) and presence (FDA) of acceptor, and the following equation was applied. (Eq.1)E=1−FDA∕FD</p><p>Fluorescence experiments were carried out with a PerkinElmer Life Sciences LS50B spectrofluorometer. To determine FD and FDA, the spectra of the samples were recorded using the excitation wavelength of 350 nm. The measured intensities were corrected for the inner filter effect using the absorption spectra of the samples (38), and the integrated intensities of the emission spectra between 440 and 460 nm in the absence and presence of acceptor were used as FD and FDA, respectively.</p><p>To obtain information regarding the dynamic and conformational properties of actin filaments, the value of E was determined at different temperatures between 6 and 30 °C, and a special FRET parameter, the normalized FRET efficiency (f'), was calculated using (Eq. 2)f′=E∕FDA The temperature dependence of the f' is informative regarding the flexibility of the investigated protein (39, 40) and was used previously to characterize the effect of cations (41, 42), nucleotides (43), and pH (37) on actin filaments. For the interpretation of the FRET results, the temperature dependence of the relative f', defined as the value of f' at the given temperature divided by the value at the lowest temperature (6 °C), is presented in this work.</p><p>The intensity of actin-bound pyrene increases by a factor of ~25 upon polymerization (44). Pyrene actin was used here to monitor the time dependence of the formation of actin filaments. In these experiments, 5% of the actin was labeled with pyrene. The excitation and emission wavelengths were 365 and 407 nm, respectively, whereas the optical slits were set to 5 nm in both the excitation and the emission paths. The elongation rate was obtained from the slopes of the pyrene intensity versus time curves at 50% completion of the polymerization. When actin (5% pyrene-labeled) was polymerized overnight at different concentrations, the actin concentration dependence of the pyrene intensity was used to determine the critical concentration of actin filament assembly.</p><!><p>The thermal denaturation of actin filaments was monitored between 0 and 100 °C with a SETARAM Micro DSC-II calorimeter. The heating rate was 0.3 K/min, and the actin concentration was 60 μm. The experiments were carried out in the absence or presence of 3 μm mDia1-FH2dimer. Calorimetric enthalpy change (ΔH) of the endothermic transitions was calculated from the area under the heat absorption curve. Transition entropy change (ΔS) was calculated for the peak transition temperature (Tm) from the following equation. (Eq. 3)ΔS=ΔH∕Tm The Gibbs free enthalpy change was calculated for 22 °C from the following equation. (Eq. 4)ΔG=ΔH−TΔS</p><!><p>The rate of dissociation of phosphate from actin filaments was measured based on the method originally described by Webb (45) using the EnzChek phosphate assay kit (Molecular Probes). In the presence of Pi, the substrate 2-amino-6-mercapto-7-methylpurine riboside is converted enzymatically by purine nucleoside phosphorylase to ribose 1-phosphate and 2-amino-6-mercapto-7-methyl-purine. The enzymatic conversion of 2-amino-6-mercapto-7-methylpurine riboside was followed by measuring the absorption at 360 nm. The experimental strategy was similar to that applied earlier by Carlier (46). First the bound calcium was replaced by magnesium in actin monomers, and then the actin at 100 μm was polymerized by the addition of 0.5 mm MgCl2 and 10 mm KCl in the absence or presence of formin. The polymerization was carried out for 2 min in the absence and for 1 min in the presence of formin. Then actin was added to the reaction buffer to a final concentration of 10 μm, and the absorbance at 360 nm was monitored. The reaction medium contained the standard buffer from the manufacturer (100 mm Tris-HCl, pH 7.5, 2 mm MgCl2, and 0.2 mm sodium azide) supplemented with 10 mm KCl.</p><!><p>To characterize the binding of formin to actin filaments, actin (5 μm; 200 μl) was polymerized overnight at 4 °C as in the FRET assays in the absence or presence of various mDia1-FH2dimer concentrations ranging from 0 to 5 μm. The samples were then centrifuged at 400,000 × g for 30 min at 20 °C with a Beckman Optima MAX benchtop ultracentrifuge and a TLA-100 rotor. The supernatants were separated from the pellets, and the pellets were resuspended in 200 μl of buffer. All supernatants and pellets were applied to a 12% SDS gel, and the gels were stained with Coomassie Blue. The actin band intensities were determined with a Syngene bio-imaging system. The affinity (K) of formin to the sides of actin filaments was estimated by analyzing the [bound formin] versus [total formin] plots using the following equation (47). (Eq. 5)[A]0D2−([A]0+[D]0+K)D+[D]0=0 where [D]0 and [A]0 are the total formin and actin concentrations, respectively, and D is the fraction of bound formin. The fraction of bound formin was calculated as the ratio of the intensity of formin bands to that of the actin bands in the pellets.</p><p>When the effect of formin on the interaction of actin filaments with cofilin was tested, preformed actin filaments (5 μm) polymerized either in the absence or in the presence of 0.5 μm mDia1-FH2dimer were incubated overnight at 4 °C with various concentrations of cofilin ranging from 0 to 5 μm. The samples were centrifuged at 400,000 × g for 30 min at 20 °C, and the supernatants were applied to 12% SDS gels to quantify the concentration of the depolymerized actin. The actin band intensities were determined using a Syngene bio-imaging system. The actin concentrations in the supernatants were determined as (Eq. 6)cactinsn=[Bsn∕(Bsn+Bpel)]×cactintotal where Bsn and Bpel are the actin band intensities in the supernatant and pellet, respectively, and cactintotal is the total actin concentration (5 μm).</p><!><p>Our initial assays demonstrated that the fluorescence emission of IAEDANS-actin filaments (10% of the total 5 μm actin) decreased in the presence of acceptor due to the effect of FRET (Fig. 1A). In test experiments, the spectra were also recorded in the presence of 500 nm mDia1-FH2dimer (Fig. 1B). The results showed that the mechanism of FRET was effective to decrease the donor intensity in the presence of formin as well, which allowed the use of FRET method to examine the effects of formins on actin filaments.</p><!><p>A characteristic feature of formins is that they accelerate the polymerization of actin solutions in vitro by enhancing the rate of the formation of actin nuclei and by promoting barbed end assembly. By following the time kinetics of the pyrene signal, we observed that 64 nm mDia1-FH2dimer accelerated the actin polymerization ~8-fold (Fig. 2A). Control experiments showed that mDia1-FH2dimer was effective in enhancing the polymerization rate at 6, 22, and 30 °C as well (Fig. 2A, inset), indicating that the interaction between the actin and formin was maintained between 6 and 30 °C.</p><p>In some cases, fluorescent labeling of macromolecules can severely affect their behavior. We tested whether the formin enhanced the polymerization of donor- and acceptor-labeled actin samples. The fluorescence intensity of IAEDANS in the absence of the acceptor increased by ~10% upon polymerization. The kinetics of the IAEDANS intensity change (Fig. 2B) showed that 64 nm mDia1-FH2dimer increased the polymerization rate by ~8-fold, to an extent similar to that observed with pyrene actin. When the IAEDANS emission was recorded in the presence of acceptor, the intensity decreased upon polymerization due to the appearance of FRET (Fig. 2B). In this case, the kinetics observed in the absence of formin was ~2-fold slower than that measured with pyrene actin, suggesting that the rate of association of acceptor-labeled actin monomers to the filaments was slower than that of the unlabeled ones. Formin (64 nm mDia1-FH2dimer) accelerated the polymerization of the double-labeled actin samples ~10-fold, which is close to the value measured with pyrene actin (~8-fold). These observations demonstrated that mDia1-FH2dimer retained its ability to accelerate the polymerization of actin after the actin was labeled with the fluorescent probes. Similar experiments at various formin concentrations (from 50 nm to 1.5 μm) corroborated this conclusion (data are not shown).</p><p>In this study, we applied temperature-dependent fluorescence experiments. Under any conditions, a fraction of the actin population, characterized by the critical concentration, is in monomeric form. Actin in its monomeric form cannot contribute to the measured transfer efficiency, and the temperature-induced changes in the critical concentration could insert errors to the interpretation of the FRET data. It was therefore important to test whether the temperature change between 6 and 30 °C shifted the equilibrium between the monomeric and filamentous forms of actin.</p><p>The critical concentration of actin was measured using the pyrene assay at 6, 22, and 30 °C (data are not shown). The critical concentration was 250–300 nm, in agreement with our previous observations (25), and proved to be temperature-independent. The experiments were repeated in the presence of 250 nm mDia1-FH2dimer (Fig. 2C). Neither the formin nor the temperature changed the critical concentration for actin assembly. The formin independence of the critical concentration is in agreement with previous observations (4, 26).</p><p>To test the effect of the applied donor and acceptor on the temperature dependence of the critical concentration, the FRET efficiency (E) was determined as a function of actin concentration. The method is based on the consideration that FRET cannot effectively occur between unpolymerized actin monomers, and thus the FRET efficiency can be used as the measure of actin filament concentration. Fig. 2D shows the data obtained at 22 °C. Fit to the actin concentration dependence of the FRET efficiency gave critical concentration of 421 ± 42 nm in the absence and 359 ± 67 nm in the presence of 250 nm mDia1-FH2dimer. These values are somewhat greater than the one obtained with the pyrene assay (250–300 nm). The higher critical concentration is most likely a consequence of the association rate of donor- and acceptor-labeled actin, which was slower than that of pyrene-labeled (5%) monomers (Fig. 2B). Using the FRET method, the critical concentration of the actin was found to be 397 ± 73 nm at 6 °C and 420 ± 48 nm at 30 °C in the absence of formin. In the presence of 250 nm mDia1-FH2dimer, the critical concentration was 457 ± 60 nm at 6 °C and 420 ± 55 nm at 30 °C. These results indicate that the critical concentration of the actin labeled with the applied fluorophores remained temperature- and formin-independent.</p><p>In general, temperature-induced changes in the mean donor-acceptor distance could affect the temperature dependence of the relative f'. In the actin filament, one donor is interacting with more than one (maximum four) acceptor on neighboring protomers. Due to the helical symmetry of the actin filaments, any change in the position of the labeled residue (Cys374 here) parallel to the longitudinal filament axis will not cause changes of the mean donor-acceptor distance. It is only the radial coordinate (the distance between the residue and the longitudinal filament axis) that can modify this distance. We tested the value of radial coordinates using the method of Taylor et al. (48) by measuring the FRET efficiency in the absence and presence of mDia1-FH2dimer ([formin]:[actin] = 1:10 ratio) at seven acceptor molar labeling ratios between 0 and 0.7 at 10, 22, and 30 °C. The results showed that the radial coordinate of the Cys374 residue was not affected by the temperature in either the absence or the presence of formin (data are not shown), indicating that the changes of this parameter were not responsible for the temperature dependence of the relative f'.</p><!><p>To characterize the formin-induced conformational changes in actin filaments, we carried out temperature-dependent FRET experiments. The efficiency (E) of the FRET (Equation 1) and the normalized transfer efficiency (f'; Equation 2) were determined over the temperature range from 6 to 30 °C. In the absence of formin, the temperature dependence of the relative f' measured at 5 μm actin showed monotonic increase to ~150% (Fig. 3A). In the presence of equimolar mDia1-FH2dimer, the temperature dependence of the relative f' was still monotonic but increased more steeply (to 175% at 30 °C) than in the absence of formin. This observation indicated that the mDia1-FH2dimer fragment modified the conformation of actin filaments by making the filaments more flexible. When the experiments were repeated by keeping the 1:1 formin:actin ratio and increasing the concentration of both proteins to 10 μm, the FRET results were essentially identical to those obtained at 5 μm actin and formin (Fig. 3A, inset). The actin concentration independence of the FRET results corroborated the conclusion from critical concentration experiments that the temperature cannot induce substantial shift in the monomer-filament equilibrium, and thus the temperature-dependent FRET data were not biased by this undesired temperature effect.</p><p>In control measurements, the FRET experiments were carried out under various salt conditions. At 1 mm MgCl2 and 50 mm KCl, the effect of mDia1-FH2dimer was smaller on the temperature dependence of the relative f' by about a factor of 2 than measured under low salt conditions (0.5 mm MgCl2 and 10 mm KCl). When the ionic strength was further increased by using 2 mm MgCl2 and 100 mm KCl, the formin effect further decreased to ~30% of the effect observed under low salt conditions. Considering these observations, we performed the formin concentration-dependent FRET experiments under low salt conditions.</p><p>The temperature-dependent FRET experiments were also carried out by keeping the actin concentration constant (5 μm) and varying the mDia1-FH2dimer concentration between 5 nm and 5 μm. Fig. 3B shows the results obtained at 500 nm mDia1-FH2dimer. At this formin concentration, the relative f' showed a steep increase to ~270% (Fig. 3B). Repeated experiments at 500 nm mDia1-FH2dimer (n = 3) gave relative f' increases between 250 and 290%. The change of the f' was greater at 500 nm mDia1-FH2dimer than at 5 μm, indicating that the effect of mDia1-FH2dimer depended on the formin concentration.</p><p>To interpret the formin concentration dependence of the FRET results, we choose the value of the relative f' at 30 °C as the measure of the actin filament flexibility. This parameter increases monotonically with the increasing flexibility but is not linearly proportional to that (39, 40). The dependence of this flexibility parameter on the formin:actin ratio is presented in Fig. 4. The flexibility of actin filaments was increased by the formin fragment at all formin:actin ratios. The effect of mDia1-FH2dimer increased with the increasing formin concentration up to ~750 nm, which corresponded to a 0.15:1 = formin:actin ratio. Above 750 nm, the effect of formin became smaller with increasing formin concentration.</p><p>For the interpretation of the FRET results, we considered the ability of formin fragments to bind to the side of actin filaments (see "Discussion"). It was recently reported that mDia1-FH2 fragments bind to the sides of actin filaments with an affinity of ~3 μm (4). To test whether our mDia1-FH2dimer construct behaves similarly under the conditions applied here, co-sedimentation assays were carried out (Fig. 4B). At 5 μm mDia1-FH2dimer, the sum of formin band intensities in the pellet and supernatant was identical to the sum of actin (5 μm) band intensities after correcting for the difference in the molecular weights (data are not shown), indicating that Coomassie Blue stained equal masses of the two proteins identically. Although mDia1-FH2dimer did not pellet in the absence of actin (Fig. 4B, lane i), in the presence of actin, formin was found in the pellets, indicating that mDia1-FH2dimer bound to the sides of actin filaments (Fig. 4B). The concentration of formin in the pellets increased with increasing total formin concentrations (Fig. 4B). Although this tendency did not perfectly define a hyperbola, the data could be fitted with Equation 5 assuming that formin binds to the sides in a 1:1 stoichiometry (Fig. 4B, solid line). The hyperbola fit gave a binding affinity of 2.0 ± 0.2 μm. This affinity is tighter than that (3 μm) reported by Li and Higgs (4). Considering that their study was carried out at greater salt concentrations (1 mm MgCl2 and 50 mm KCl) than applied here (0.5 mm MgCl2 and 10 mm KCl), the difference between the two estimates suggests that the affinity of mDia1-FH2 for the sides of actin filaments is salt-dependent.</p><p>The effect of mDia1-FH2monomer on actin filaments was also described by the FRET method. Fig. 4C shows the dependence of actin filament flexibility on the mDia1-FH2monomer concentration. The actin filaments were more flexible in the presence of mDia1-FH2monomer than in the absence of it, although the transient formin concentration dependence found with mDia1-FH2dimer (Fig. 4A) was not observed in experiments with the mDia1-FH2monomer.</p><!><p>In previous studies, the dynamic behavior of actin resolved by spectroscopic assays often correlated with its thermal stability (e.g. Ref. 49). To test whether it was the case here, we measured the heat denaturation of actin filaments (60 μm) using differential scanning calorimetry. The experiments were carried out in the absence or presence of 3 μm mDia1-FH2dimer, i.e. at formin:actin ratio of 1:20. The results showed that the peak temperature of the endothermic melting (Tm) decreased from 61.9 °C in the absence of formin to 60.4 °C in the presence of formin (Fig. 5A). The relatively small formin-induced decrease of Tm was reproducible in repeated experiments (n = 3) and indicated that the thermal stability of actin filaments decreased in the presence of formin. The calorimetric enthalpy change (ΔH), the entropy change (ΔS), and the Gibbs free enthalpy (ΔG) values were smaller when the experiments were carried out in the presence of formin (Table 1), supporting the conclusion that the filaments were thermodynamically less stable in the presence of mDia1-FH2dimer than in the absence of it.</p><!><p>The dynamic properties of actin filaments were reported previously to correlate with the rate of phosphate dissociation from them (e.g. Ref. 50). Here we tested the effect of formin on the rate of phosphate dissociation from actin filaments by using the method of Webb (45). A prerequisite for these experiments is that the formation of actin filaments is much faster than the phosphate release rate. To achieve this aim, the actin was first polymerized at relatively high concentration (100 μm). Under the applied conditions, more than 90% of the actin was in polymer form within the first 2 min as indicated by pyrene fluorescence (data are not shown). In the presence of formin, more than 90% of the actin polymerized in ~1 min. The filaments formed during this polymerization phase contained more Pi than in a dynamic equilibrium, which resulted in a burst phase of phosphate release during the first 10–20 min of the experiments (Fig. 5B, inset). A subsequent phase was also observed that followed linear tendency and was attributed to the phosphate release during the slow turnover of actin filaments. The absorption transients were analyzed by fitting the sum of a single exponential and a linear function to the experimental data (Fig. 5B, inset). In the presence of formin, the linear component of the fit appeared to be slower by approximately a factor of 2. This finding is in accordance with previous observations that mDia1-FH2 is a barbed end capper and slows down the association and dissociation of actin monomers to or from filaments. The amplitude of the exponential component was in the range of 0.05–0.07, which corresponded to 6–8 μm Pi according to the calibration curve obtained with standard phosphate solutions. The rate determined from the exponential component corresponded to the first order phosphate release rate from actin filaments. In the absence of formin, the phosphate release rate was 1.85 × 10−3 s−1. This value is in close agreement with previous estimates for this parameter (e.g. 1.98 × 10−3 s−1 (51)). The phosphate release rate from actin filaments was increased by formin (Fig. 5B) and followed a formin concentration dependence similar to that observed for the filament flexibility in the FRET assays (Fig. 4A). These observations indicated that there was a direct correlation between the flexibility of actin filaments and the rate of phosphate dissociation from actin filaments.</p><!><p>We attempted to test whether the change of the actin conformation was accompanied by modifications in its interaction with cofilin. Actin (5 μm) was incubated and co-sedimented with various cofilin concentrations (0–5 μm) either in the absence or in the presence of 0.5 μm mDia1-FH2dimer. Fig. 5C shows the actin band intensities in the supernatants. In the presence of formin, greater actin concentrations were found in the supernatants, showing that cofilin was more effective in increasing the amount of monomeric actin in the presence of formin than in the absence of it. However, at this point, it is not possible to distinguish whether the increased actin disassembly results from formin-induced changes in the filament structure or from the direct effects of mDia1-FH2 to actin assembly at the filament barbed end.</p><!><p>To characterize the effect of mDia1-FH2 on the dynamic properties of actin filaments, temperature-dependent FRET experiments were carried out. The basis of the application of this method is that the value of f' is sensitive to the average distance between the donor and acceptor and also to the amplitude of the relative motion of these two probes (39, 40). Increasing temperature increases the rate and amplitude of the relative fluctuations of the donor and acceptor. Due to the inverse 6th power dependence of the transfer efficiency on the donor-acceptor distance, these effects of the increasing temperature result in an increase of f' (39, 40). When the probes are located in a more flexible protein environment, the increase of the amplitude of the donor-acceptor fluctuations is greater, and thus the increase in the value of f' is greater as well. Accordingly, the steeper temperature dependence of f' reflects the more flexible form of the protein. Note that due to the nature of this method, the relative f' is sensitive to all kinds of motions at any time scale that can influence the donor-acceptor distance. Thus the term flexibility does not exclusively correspond to either the torsional or the bending or any other specific filament flexibilities here. The separation of the effects of these different modes of motions on the temperature dependence of f' is very difficult. This FRET method was previously applied in numerous cases to describe conformational changes in actin (37, 42, 43, 49, 52). In the present study, due to the applied experimental strategy, there was no actin monomer containing both donor and acceptor, and thus FRET always occurred between probes on different actin protomers in the filaments. This strategy assured that the inter-protomer dynamics of the filaments were described.</p><p>We found in control experiments that the mDia1-FH2dimer fragment studied here interacted properly with actin over the investigated temperature range. Donor or acceptor labeling of actin did not diminish the formin-actin interactions. Neither formin nor the change in temperature caused substantial shift in the monomer-filament equilibrium. We concluded that the observed formin-induced changes in the temperature dependence of the f' were attributed to the conformational changes within the actin filaments. The exact nature of these conformational changes is, however, not yet known, and we cannot exclude at this stage that apart from intramolecular conformational transitions, the binding of formin alters the supra-molecular filament structures.</p><!><p>The temperature dependence of the f' was steeper in the presence of mDia1-FH2dimer than in the absence of it at all investigated formin concentrations (Fig. 4), indicating that mDia1-FH2dimer increased the flexibility of actin filaments. We found that the effect of mDia1-FH2dimer on the flexibility of actin filaments was more pronounced at 0.5 mm MgCl2 and 10 mm KCl than at 1 mm MgCl2 and 50 mm KCl. This observation suggests that the formin effect may be smaller at physiological ionic strength. The increase in the flexibility detected by the FRET assay was accompanied by the decrease of the thermal stability of the filaments (Fig. 5A). The formin-induced flexibility change strongly depended on the applied formin concentration (Fig. 4A). The largest formin effect was observed at ~750 nm mDia1-FH2dimer, which corresponds to a 0.15:1 formin:actin protomer ratio. The formin concentration dependence of the actin flexibility presented in Fig. 4 suggests that the effect of mDia1-FH2dimer on actin involves at least two mechanisms.</p><!><p>It was established previously (4, 24, 25) that mDia1 fragments could bind to actin in two distinct ways, either at the barbed end or on the side of the filaments. We interpret our results considering these two binding mechanisms. The affinity for the binding of dimeric mDia1 fragments to the barbed end falls into the 20–50 nm range (4, 26), whereas side binding is characterized with a much weaker affinity of ~2-3 μm (4) (Fig. 4B). A scheme in Fig. 6 outlines the binding geometries involved in the interpretation.</p><p>At low mDia1-FH2dimer concentrations (<750 nm), the binding of mDia1-FH2dimer to the barbed end is dominating over the side binding due to two reasons: 1) the affinity is much tighter for the barbed end than for the side binding and 2) the formin:actin stoichiometry allows the saturation of only the minority of the filament side-binding sites. Increasing the formin concentration within the 0–750 nm range results in a saturation of the filament ends. The formin dependence up to 750 nm (Fig. 4) can be explained by assuming that formin binding to the barbed end increases the actin filament flexibility. The mDia1-FH2 used here is dimeric, which means that for example, at a 1:10 = formin:actin ratio, there are at least 20 actin protomers for each formin dimer. However, this estimate puts only a lower limit to the number of actin monomers in a polymer capped by one single formin dimer.</p><p>Due to the actin nucleation activity of formins, the number of actin filaments formed initially in the presence of formins exceeds substantially the number of filaments generated in the absence of formin. A previous study from the Pollard group showed that end-to-end annealing was effective to produce long filaments from fragmented short actin filaments (53). The rate of annealing depends on the number of filaments. In our experiments, the number of shorter filaments was relatively high immediately after the initiation of polymerization by formins, suggesting that the annealing progressed rapidly. Andrianantoandro et al. (53) also showed that end-to-end annealing occurred, although at a lower rate, even in the presence of capping protein, which binds the barbed end of filaments tightly. They observed that in the presence of capping protein 24 h after fragmentation of actin filaments, the filament length distribution was re-established and was similar to that observed 1 h after fragmentation in the absence of capping protein. This observation was explained by the gated reaction theory based on the slow diffusion of actin filaments and the dynamic equilibrium between the capping protein and the barbed end of the filaments (53). Because the mDia1-FH2 fragments were reported to be in a rapid equilibrium with the filament ends (26), and because in this study, we incubated the formin-actin samples overnight before the experiments, it is likely that at the time of the FRET measurements, the length distribution of actin filaments was similar in the absence and presence of formin.</p><p>According to these considerations, there were hundreds of actin protomers for each formin dimer bound to a barbed end of a filament. The fact that the change of the flexibility could be observed in our assays indicates that one dimer must have changed the conformation of many actin protomers, i.e. formin binding to the barbed end produced long range allosteric conformational changes along the actin filaments. The current set of data does not allow the determination of the ratio of the concentration of actin protomers influenced by formin to the total actin concentration, i.e. the range of the allosteric interactions cannot be determined.</p><p>It was reported previously in numerous cases that effects of ligands and actin-binding proteins propagated along actin filaments through long range allosteric interactions (e.g. Refs. 54-59). Related to the present study, a long range allosteric effect of gelsolin, another barbed end-binding protein, was observed. The Egelman group (57) showed that when actin filaments were nucleated in the presence of gelsolin, the altered conformational state of the filament could be detected throughout the whole filament. In accordance with this observation, the Thomas laboratory (58) reported gelsolin-induced long range cooper-ativity in actin filaments where a conformational change induced by the binding of a single gelsolin molecule to the barbed end is propagated throughout the actin filament. These observations and our present findings indicate that it is probably the inherent property of actin filaments to be able to transport the conformational changes initiated at the barbed end through many protomer-protomer interactions to distant locations along the actin polymers.</p><p>It was also suggested (e.g. Ref. 59) that the regulatory effect of actinbinding proteins on the polymerization state of actin is tightly correlated with the various, often co-existing, conformations of actin protomers and with the modifications in their interactions. Recently, Orlova et al. (59) suggested that actin-depolymerizing proteins such as ADF/cofilin exert their effect on actin filaments by inducing conformational transitions, which result in a less stable state of the filaments. These less stable states are proposed to exist in earlier phases of polymerization as well. In the light of our data, it seems possible that cofilin is more effective (Fig. 5C) at depolymerizing the more flexible (Fig. 3) and less stable (e.g. thermodynamically, Fig. 5A) actin filaments generated in the presence of formin because the activation energy required to shift the conformations to that established by cofilin was lower in the presence of formin than in the absence of it.</p><p>The altered interaction of actin filaments with cofilin in the presence of formin indicates that the formin-induced conformational changes can have functional consequences. In accordance with this conclusion, we also showed in this study that another functional feature of actin, one of the steps of the ATPase cycle, the phosphate release, was modified by the binding of formins. The close correlation between the formin dependence of the actin flexibility (Fig. 4A) and the phosphate release rate (Fig. 5B) suggests that the dissociation of Pi from the looser actin filaments formed in the presence of formin is thermodynamically more favorable than that in the absence of formin.</p><p>The changes in the actin flexibility above 750 nm cannot be explained by further barbed end binding of the formin fragment as most of the barbed ends are formin-bound at 750 nm. On the other hand, the flexibility in the 750–5000 nm range follows a tendency opposite to that observed below 750 nm mDia1-FH2dimer. The simplest way to interpret the FRET results above 750 nm formin is to assume that binding of the dimeric mDia1-FH2dimer to the sides of the actin filaments stabilizes the protomer-protomer interactions. The structural details of the interaction between the sides of the actin filaments and formin fragments are unknown. Our results suggest that in these interactions, one formin dimer is interacting with at least two actin protomers, which could form the bases of the stabilization of the filaments.</p><p>The results of the FRET experiments can therefore be explained by assuming that a formin "cap" at the barbed end increases the flexibility of the filament, whereas formin "cramps" linking actin protomers along the sides of actin filaments can stabilize the molecular interactions between neighboring protomers and make the filaments stiffer. The superposition of these two effects was observed in the FRET measurements.</p><p>We attempted to test this hypothesis by using the mDia1-FH2monomer fragment (Fig. 4C), which was characterized previously (25). The hypothesis predicted that if the binding of mDia1-FH2monomer can induce the loosening of the actin filaments, the mDia1-FH2monomer concentration dependence of the effect should correlate with the affinity of this mDia1-FH2monomer fragment to the filament barbed ends. The affinity of mDia1-FH2monomer to the barbed end is ~2 μm, and it binds the sides of actin filaments with an ~3 μm affinity (25). The filament flexibility increased over the investigated concentration range up to 12.5 μm mDia1-FH2monomer (Fig. 4C). Although the fact that the affinities of mDia1-FH2monomer for the filament end- and side-binding sites are similar complicates the interpretation of the FRET data, the observation that high mDia1-FH2monomer concentrations were required to increase the flexibility of the actin filaments is in agreement with our model. The lack of the decrease of the formin effect at high formin concentrations indicates that the mDia1-FH2monomer fragment bound to the sides of actin filaments was not able to stabilize the interaction between neighboring protomers, suggesting that this property is only characteristic for the formin dimers.</p><!><p>The function of the actin cytoskeleton requires the coordinated work of many regulatory and structural proteins. In many cases, the proteins involved in these complexes can be attributed to specific actin nucleation factors. It is important to note that in cells, a group of proteins localizes to formin-nucleated actin structures, whereas other proteins are typically associated with actin filaments nucleated by the Arp2/3 complex (for a summary, see Ref. 60). However, the mechanism by which, e.g. ADF/cofilins, preferentially localize to Arp2/3-nucleated actin structures and tropomyosins localize to formin-nucleated structures is not understood. One possible mechanism of regulation would be based on the effect of actin nucleation factors on the conformation of actin filaments. One can envisage that these nucleation factors can change the affinity of actin-binding proteins to actin filaments and thus determine which proteins are involved in the corresponding complexes. Due to the nature of their function, actin nucleation factors are the first to bind to the newly generated filaments, and the problem of timing the sequence of the binding of the actin-binding proteins would be overcame by the regulatory role of these nucleation factors. A prerequisite for this mechanism is that the nucleation factors can change the conformation of actin filaments. These effects should be long range allosteric effects as the nucleation factors bind to the ends of filaments. Our FRET results provided evidence that formin fragments can have substantial and allosteric effects on the conformation of actin filaments by binding to the barbed end. This conclusion is independent of the validity of proposed explanation (Fig. 6). Based on this observation, we propose that a special mechanism is involved in the regulation of the formation of cytoskeletal protein complexes in which the conformational changes induced by the nucleation factors in actin filaments play an essential role in determining the affinity of actin-binding proteins for actin filaments. Although further experiments are required to test this hypothesis, this novel mechanism would be one of the first examples of how actin filaments can serve as regulatory information channels in living cells.</p><!><p>We showed here that the binding of mDia1-FH2dimer to actin increased the flexibility of actin filaments by long range allosteric interactions. In accordance with this observation, the filaments became less resistant to heat denaturation in the presence of formin. These conformational changes were accompanied by the modifications of functional properties of actin as its interaction with cofilin and the rate at which phosphate dissociates from actin filaments were also influenced by the binding of formin.</p><p>The present results, on the other hand, raised and left many questions unanswered. Further experiments will be required to determine the distance to which formin at the barbed end can allosterically modify the conformation of actin protomers. Also to be tested is how abundant this formin effect is among different members of the formin family. One would wonder whether other actin nucleation factors, such as the Arp2/3 complex, can also modify the conformation of actin filaments. Also, finally, the confirmation of our hypothesis describing the new mechanism, which can serve to regulate the formation of cytoskeletal complexes, requires further investigations.</p>

PubMed Author Manuscript

The maximum chemiluminescence intensity predicts severe neutropenia in gemcitabine-treated patients with pancreatic or biliary tract cancer

PurposeTo assess the predictive ability of the maximum chemiluminescence intensity (CImax) for severe neutropenia (SN) during neoadjuvant chemo(radio)therapy [NAC(RT)] in patients with advanced pancreatic or biliary tract cancer.MethodsClinicopathological variables and blood test data before NAC(RT) were evaluated in 64 patients with advanced pancreatic or biliary tract cancer who received gemcitabine plus tegafur/gimeracil/oteracil as NAC(RT).ResultsThirty-nine patients (60.9%) developed Grade 3–4 SN. The median time between commencing NAC(RT) and the onset of SN was 15 (range 10–36) days. SN occurred during the NAC period, not the RT period. The CImax, neutrophil count, serum interleukin-6 level, C-reactive protein level, complement C3 titer, serum complement titer, and 50.0% hemolytic unit of complement before NAC(RT) were significantly lower in patients with SN than in those without SN (P < 0.05). Multivariate analysis confirmed the CImax to be the sole independent predictor of SN (P < 0.05). The optimal threshold for the CImax was 46,000 RLU/s. The sensitivity and specificity were 46.2% and 80.0%, respectively. Majority of the patients (81.8%) with a low CImax before NAC(RT) experienced SN during NAC(RT).ConclusionsCImax before NAC(RT) predicts SN during NAC(RT) in patients with advanced pancreatic or biliary tract cancer.

the_maximum_chemiluminescence_intensity_predicts_severe_neutropenia_in_gemcitabine-treated_patients_

Introduction<!>Study population<!>Neoadjuvant chemo(radio)therapy<!>Evaluated factors<!>Maximum chemiluminescence intensity<!>Endpoints<!>Statistical analyses<!>Patient characteristics<!><!>Univariate analysis<!><!>Multivariate analysis<!><!>Prediction ability<!><!>Relative dose intensity of gemcitabine<!>Discussion<!><!>Discussion<!>Funding<!>Conflict of interest<!>Ethical approval<!>Informed consent

<p>Majority of the patients with advanced pancreatic or biliary tract cancer have a poor prognosis. Complete surgical resection is currently the only potentially curative treatment for long-term survival. However, majority of the patients considered to have localized cancer by radiographic examination actually have undetected systemic disease and are unlikely to benefit from surgical treatment alone [1, 2]. However, with rapid developments in chemotherapeutic regimens, both adjuvant chemotherapy and neoadjuvant chemo(radio)therapy (NAC(RT)) have been shown to be beneficial for patients with borderline resectable pancreatic cancer [3] or advanced biliary tract cancer [4]. While gemcitabine plus tegafur/gimeracil/oteracil (GS) as NAC(RT) has been reported to be safe and effective for patients with borderline resectable pancreatic cancer [5], GS as adjuvant chemotherapy [6] and gemcitabine as NAC [7] have been reported to be safe and effective for patients with advanced biliary tract cancer.</p><p>We have performed NAC(RT)-GS in patients with advanced pancreatic or biliary tract cancer. Bone marrow suppression, an adverse effect of anticancer drugs, was frequently observed in patients treated with NAC(RT)-GS. The incidence of severe neutropenia (SN) [Grade ≥ 3 according to the Common Terminology Criteria for Adverse Events (version 4.0) [8]] was particularly high and has been reported in approximately 62.2% of patients [9]. SN is a major toxicity that forces a reduction in the relative dose intensity (RDI) of the anticancer drugs used in NAC(RT). SN during NAC(RT) can also complicate tumor resection. However, the risk of SN during gemcitabine-based therapy has not been extensively studied.</p><p>Previous studies have demonstrated that risk factors for SN and febrile neutropenia include old age [10, 11], female sex [12], a poor Eastern Cooperative Oncology Group performance status [13], a low body mass index [14], a small body surface area [15], a history of cardiovascular disease [10], diabetes mellitus [16], a poor nutritional status, inflammation [11, 13], and a low baseline absolute neutrophil count (ANC) [17]. Kiguchi et al. [18] reported that the maximum chemiluminescence intensity (CImax), as assessed by an in vitro reaction between peripheral neutrophils and endotoxin, is indicative of the maximum neutrophil activity in whole blood. A low CImax is also associated with the exhaustion of peripheral polymorphonuclear leukocytes. CImax has been suggested to be predictive of mortality in patients with sepsis. Therefore, in addition to the baseline ANC, we also focused on neutrophil activity for predicting the onset of SN.</p><p>The aim of this study was to investigate potential markers of SN in patients with advanced pancreatic or biliary tract cancer who received NAC(RT)-GS by evaluating clinicopathological variables and nutritional and immune markers (including CImax) before NAC(RT).</p><!><p>We conducted a retrospective observational study in the Department of Gastroenterological Surgery at Yokohama City University Graduate School of Medicine (Yokohama, Japan). The study protocol was approved by the Ethical Review Board of Yokohama City University Hospital (Yokohama, Japan) (approval number: 121101023). Sixty-four chemo-naïve patients with histologically proven advanced pancreatic or biliary tract cancer who were treated with NAC(RT)-GS between June 2013 and December 2015 were analyzed. Patients with multiple primary cancers and a history of prior chemotherapy, as well as those who did not complete NAC(RT)-GS, were excluded.</p><p>Pancreatic and biliary tract cancer was diagnosed and staged on the basis of ultrasonography, abdominal computed tomography, magnetic resonance imaging, ultrasound endoscopy, endoscopic retrograde cholangiopancreatography, positron emission tomography, cytological or histological examinations, and explorative laparotomy. NAC(RT) was administered to patients with borderline resectable pancreatic cancer as defined by National Comprehensive Cancer Network Guidelines Version 1/2012 [19]. NAC was administered to patients with biliary tract (hilar cholangiocarcinoma with arterial invasion, metastatic lymph nodes, or Bismuth type IV) or gallbladder cancer (plural metastatic lymph nodes or clinical T3–4 disease according to the tumor-node-metastasis classification of the International Union Against Cancer, 7th edition [20]).</p><!><p>The NAC(RT) regimen for patients with pancreatic or biliary tract cancer consisted of gemcitabine (1000 mg/m2 administered intravenously on days 8 and 15) plus tegafur/gimeracil/oteracil (60 mg/m2 administered orally on days 1–14). Patients with pancreatic cancer (n = 50) received two courses of GS followed by 30 Gy of radiation therapy. Patients with biliary tract cancer (n = 14) received three courses of GS.</p><!><p>Age, sex, body mass index, body surface area, Eastern Cooperative Oncology Group performance status, and a history of smoking, cardiovascular disease, and diabetes mellitus before receiving NAC(RT)-GS were evaluated from clinical records. CImax, white blood cell count, ANC, lymphocyte count, platelet count, serum interleukin 6 level, C-reactive protein level, complement C3 titer, complement C4 titer, 50.0% hemolytic unit of complement, albumin level, prognostic nutritional index, serum carcinoembryonic antigen level, carbohydrate antigen 19-9 level, pancreatic cancer-associated antigen level, and s-pancreas-1 antigen level were also evaluated.</p><!><p>CImax was assessed based on an in vitro reaction between peripheral neutrophils and endotoxin. An endotoxin activity assay was performed as described previously [21]. Fifty microliter samples of whole blood and appropriate controls were incubated in duplicate with saturating concentrations of an anti-lipid A immunoglobulin M antibody, and then stimulated with opsonized zymosan. The resulting respiratory burst was detected by a chemiluminometer (Autolumat LB953; Berthold Technologies GmbH & Co. KG, Bad Wildbad, Germany) as light released from the lumiphore luminol. The maximum stimulated response (termed as CImax by Kiguchi et al. [18]) was measured using lipopolysaccharide (4.6 ng/mL) as the stimulant.</p><!><p>The primary endpoint was the incidence of Grade ≥ 3 SN (ANC < 1000/mL) during NAC(RT)-GS. The observation period was from the first to the last day of NAC(RT)-GS. The secondary endpoint was the RDI of gemcitabine.</p><!><p>Data are expressed as the median and range or number and percentage. Continuous variables were analyzed using the Mann–Whitney U test and categorical variables were analyzed using Chi square or Fisher's exact test. Univariate and multivariate logistic regression analyses were performed to determine independent predictors of the incidence of SN during NAC(RT)-GS. Odds ratios and their 95.0% confidence intervals were calculated. Continuous variables were adjusted by dividing by the standard deviation and comparing to the odds ratio. Statistical analyses were conducted using Statistical Package for the Social Sciences for Windows (software version 23.0; SPSS Japan Inc., Tokyo, Japan). A P < 0.05 was considered statistically significant.</p><!><p>The patient characteristics are summarized in Table 1. The median age was 71 (range 39–85) years. Thirty-six patients (56.3%) were male. All patients had an Eastern Cooperative Oncology Group performance status of 0–1. The general condition of the patients was satisfactory. Fifty patients (78.1%) had advanced pancreatic cancer. Fourteen patients (21.9%) had advanced biliary tract cancer. The median serum albumin levels were slightly lower, while the carbohydrate antigen 19-9 levels, pancreatic cancer-associated antigen levels, and s-pancreas-1 antigen levels were higher, than their normal ranges. SN was detected in 39 patients (60.9%), within a median of 15 (range 10–36) days from commencing NAC(RT)-GS. All instances of SN occurred during the chemotherapy period and not the radiotherapy period.</p><!><p>Patient characteristics</p><p>ANC absolute neutrophil count, BMI body mass index, BSA body surface area, CA19-9 carbohydrate antigen 19-9, CEA carcinoembryonic antigen, CImax maximum chemiluminescence intensity, CRP c-reactive protein, CVD cardiovascular disease, DM diabetes mellitus, DUPAN-2 pancreatic cancer-associated antigen, ECOG Eastern Cooperative Oncology Group, F female, IL-6 interleukin 6, LC lymphocyte count, M male, O-PNI Onodera's prognostic nutritional index, PC platelet count, PS performance status, RLU relative light unit, SPan-1 s-pancreas-1 antigen, WBC whole blood count</p><!><p>The results of the univariate analysis are summarized in Table 2. CImax, ANC, interleukin 6 level, C-reactive protein level, complement C3 titer, and 50.0% hemolytic unit of complement before NAC(RT) were identified as significant factors. Conversely, no epidemiological, tumor-related, or nutritional factors were found to be significant.</p><!><p>Univariate analysis of factors predicting severe neutropenia (SN) in patients with advanced pancreatic or biliary tract cancer</p><p>(+) positive, (−) negative, ANC absolute neutrophil count, BMI body mass index, BSA body surface area, C3 complement C3, C4 complement C4, CA19-9 carbohydrate antigen 19-9, CEA carcinoembryonic antigen, CH50 50.0% hemolytic unit of complement, CImax maximum chemiluminescence intensity, CVD cardiovascular disease, DM diabetes mellitus, DUPAN-2 pancreatic cancer-associated antigen, F female, GEM gemcitabine, IL-6 interleukin 6, LC lymphocyte count, M male, NACRT neoadjuvant chemoradiotherapy, O-PNI Onodera's prognostic nutritional index, PC platelet count, RDI relative dose intensity, RLU relative light unit, S1 tegafur/gimeracil/oteracil, SPan-1 s-pancreas-1 antigen, WBC white blood cell count</p><p>*P < 0.05</p><!><p>The results of the multivariate analysis, which was performed using the six significant factors identified in the univariate analysis, are summarized in Table 3. Independent variables were selected using the simultaneous method for CImax and stepwise methods for the remaining five factors (the criterion for adding a new variable was P < 0.05). CImax was identified as a significant independent predictor of SN during NAC(RT)-GS (odds ratio: 0.248, 95.0% confidence interval 0.073–0.850; P = 0.026).</p><!><p>Multivariate analysis of factors predicting severe neutropenia in patients with advanced pancreatic or biliary tract cancer</p><p>ANC absolute neutrophil count, C3 complement C3, CH50 50.0% hemolytic unit of complement, CI confidence interval, CImax maximum chemiluminescence intensity, CRP c-reactive protein, IL-6 interleukin 6, OR odds ratio, SD standard deviation</p><p>*P < 0.05</p><!><p>The area under the receiver operating characteristic curve of the incidence of SN predicted by CImax was 0.704 (Fig. 1). The optimal threshold for the CImax was 46,000 RLU/s. Applying this cutoff, the sensitivity and specificity were 46.2% and 80.0%, respectively. The majority of patients (n = 18; 81.8%) with a low CImax before NAC(RT) experienced SN during NAC(RT)-GS.</p><!><p>Receiver operating characteristic curve of the maximum chemiluminescence intensity for predicting severe neutropenia in patients with advanced pancreatic or biliary tract cancer</p><!><p>The RDI of gemcitabine in patients with SN was lower than in those without SN (Table 2). At CImax cutoff of 46,000 RLU/s, the median RDI of gemcitabine was significantly lower in the low CImax group than in the high CImax group (65.0% vs. 75.0%, respectively; P = 0.014).</p><!><p>Gemcitabine has a wide spectrum of anticancer activities with few non-hematological adverse events [22]. However, the GEST study [9] showed that SN had occurred in 62.2% of patients with locally advanced and metastatic pancreatic cancer in Japan and Taiwan. Consistent with the GEST study [9], the incidence of SN in our study was 60.9%. Patients who develop SN during NAC require serial dose reductions until a tolerable dose is reached. In this study, the RDI of gemcitabine was significantly lower in patients who developed SN.</p><p>Studies [10–16] have shown epidemiological (age, sex, or a prior history), tumor-related, nutritional, and inflammatory statuses to be risk factors for SN. However, in relation to gemcitabine-based chemotherapy, risk factors for SN include a low ANC, a low white blood cell count, a low carbohydrate antigen 19-9 level, and no prior history of smoking [17, 23, 24]. In this study, no association was identified between epidemiological, tumor-related, or nutritional factors and the incidence of SN during NAC(RT)-GS.</p><p>In the present study, logistic regression analysis identified CImax before NAC(RT) as an independent predictor of SN during NAC(RT)-GS. At an optimal cutoff value of 46,000 RLU/s, the specificity of the CImax for predicting SN was 80.0%. The positive predictive value was 81.8%.</p><p>In patients with lipopolysaccharide-induced sepsis, the oxidative burst is significantly diminished in non-survivors compared to survivors [25]. Reduced oxidative activity may be associated with immune dysfunction and high mortality [25, 26]. Kiguchi et al. [18] reported that CImax before commencing treatment is indicative of the maximum neutrophil activity in whole blood and is highly predictive of mortality in patients with sepsis. CImax reflects neutrophil vitality or fatigue. Therefore, we hypothesized that CImax may be a predictor of SN during NAC.</p><p>In this study, univariate analysis identified a low CImax, a low ANC, a low interleukin 6 level, a low C-reactive protein level, a low complement C3 titer, and a low 50.0% hemolytic unit of complement as risk factors for SN. In advanced cancer, chronic inflammation caused by an elevation in inflammatory cytokines leads to tumor progression [27], and activates neutrophil functions, such as the expression of adhesion molecules, phagocytosis, and the production of reactive oxygen species [28]. CImax reflects the reactive oxygen species production of neutrophils. Therefore, a low CImax indicates the inhibition of these inflammatory reactions. Recently, it has been reported [29, 30] that single nucleotide polymorphisms (SNPs) determine the individual components and/or the total white blood cell count. Furthermore, SNPs may correlate with the incidence of chemotherapy-induced neutropenia. Patients with SN may have a low baseline CImax, resulting in the overall inhibition of inflammatory responses due to unknown SNPs, which result in lower neutrophil counts and reduced neutrophil activity. CImax may be a surrogate marker of not only predicting the incidence of SN, but also the patient's overall immune function.</p><p>Measurement of CImax is quick and convenient for clinical use. If CImax could be adopted in daily clinical practice then patients at risk of developing SN may be given prompt and appropriate treatment. Classifying patients into different risk groups based on their CImax may also help physicians choose the optimal treatment strategy. For instance, patients at a high risk of SN can be started with a reduced dose of GS and have their ANC strictly monitored by frequent blood tests. If patients are informed of their risk of neutropenia in advance, they can pay more attention to their own condition during treatment.</p><p>There is the potential that SN may arise exclusively from a low baseline ANC due to the inhibition of inflammation or neutrophil lowering SNPs. In other words, a low baseline ANC alone may be sufficient to cause to SN. However, we determined that the correlation coefficient between CImax and ANC was 0.462 (Fig. 2). Hence, CImax and ANC are independent factors. The multivariate analysis showed CImax to be a better predictive factor, representing the conclusive status of neutrophils.</p><!><p>Correlation between the maximum chemiluminescence intensity (CImax) and absolute neutrophil count (ANC)</p><!><p>SN is associated with the long-term prognosis of patients with various types of cancer, including breast [31, 32], small cell lung [33, 34], gastric [35], colorectal [36, 37], ovarian [38], and pancreatic [39, 40]. The long-term prognosis may be associated with SN and CImax, though this was not evaluated in this study. We are currently conducting a prospective cohort study focusing on the long-term outcomes of patients with advanced pancreatic cancer.</p><p>There are two limitations of this study. The first is that only a limited number of cases from a single institution were available at the time of conducting the study and the second is that only NAC(RT)-GS cases of advanced pancreatic or biliary tract cancer were analyzed. This paper only reports preliminary results. However, we believe that our findings are of interest. Multicenter studies involving greater number of patients with varying types of cancer and NAC regimens are needed to evaluate whether a low CImax is a useful marker for predicting SN.</p><p>In conclusion, our findings suggest that patients with advanced pancreatic or biliary tract cancer with a low CImax who are receiving NAC(RT) are at a high risk of developing SN during NAC(RT)-GS. Further studies with larger sample sizes are needed to validate our findings with other regimens (e.g., FOLFIRINOX and gemcitabine plus nanoparticle albumin-bound paclitaxel) and to confirm the associations between CImax, SN, SNPs, and long-term survival.</p><!><p>This study was funded by Chugai Pharmaceutical Co., Ltd. and Yakult Honsha Co., Ltd.</p><!><p>The authors declare that they have no conflict of interest.</p><!><p>All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.</p><!><p>Informed consent was obtained from all individual participants included in the study.</p>

PubMed Open Access

A novel chemopreventive mechanism for a traditional medicine: East Indian sandalwood oil induces autophagy and cell death in proliferating keratinocytes

One of the primary components of the East Indian sandalwood oil (EISO) is \xce\xb1-santalol, a molecule that has been investigated for its potential use as a chemopreventive agent in skin cancer. Although there is some evidence that \xce\xb1-santalol could be an effective chemopreventive agent, to date, purified EISO has not been extensively investigated even though it is widely used in cultures around the world for its health benefits as well as for its fragrance and as a cosmetic. In the current study, we show for the first time that EISO-treatment of HaCaT keratinocytes results in a blockade of cell cycle progression as well as a concentration-dependent inhibition of UV-induced AP-1 activity, two major cellular effects known to drive skin carcinogenesis. Unlike many chemopreventive agents, these effects were not mediated through an inhibition of signaling upstream of AP-1, as EISO treatment did not inhibit UV-induced Akt, or MAPK activity. Low concentrations of EISO were found to induce HaCaT cell death, although not through apoptosis as annexin V and PARP cleavage were not found to increase with EISO treatment. However, plasma membrane integrity was severely compromised in EISO-treated cells, which may have led to cleavage of LC3 and the induction of autophagy. These effects were more pronounced in cells stimulated to proliferate with bovine pituitary extract and EGF prior to receiving EISO. Together, these effects suggest that EISO may exert beneficial effects upon skin, reducing the likelihood of promotion of pre-cancerous cells to actinic keratosis (AK) and skin cancer.

a_novel_chemopreventive_mechanism_for_a_traditional_medicine:_east_indian_sandalwood_oil_induces_aut

Introduction<!>Materials<!>Cells<!>EISO treatment<!>MTS assay<!>Apoptotic analysis<!>Cell cycle analysis<!>Western blotting<!>Luciferase reporter assays<!>Fluorescein diacetate hydrolysis<!>Immunocytochemistry<!>EISO is cytotoxic in cultured HaCaT keratinocytes at low concentrations<!>EISO induces cell cycle arrest<!>AP-1 activity is inhibited by EISO, but not through blocking of UV-induced MAPK or PI3-K activation<!>EISO treatment compromises plasma membrane integrity<!>LC3 expression and cleavage increases as a result of EISO treatment<!>Discussion<!>Conclusion<!><!>Conclusion

<p>Ultraviolet light induces tumor promoting events in keratinocyte cells that, if allowed to proceed, will ultimately lead to the development of pre-cancerous conditions such as actinic keratosis (AK) and skin tumors, including papillomas and squamous cell carcinoma (SCC). In humans, these promotion events result from constant and unavoidable exposure to the sun over many decades, a fact which highlights the need to identify new chemopreventive strategies that can be applied over a lifetime. One strategy includes topical formulations of novel natural products that are non-toxic to normal cells, but which prevent the outgrowth of tumors through a mechanism that targets UV-induced promotion events.</p><p>Sandalwood oils are essential oils commonly used as fragrances for body oils and incense as well as in medicines and cosmetics. Sandalwood oils have many well-known health benefits due to their antiinflammatory and anti-septic properties, among others [1]. The principal commercial oils are steam distilled from the wood of two species of sandalwood trees: East Indian sandalwood (Santalum album) and West Australian sandalwood (Santalum spicatum). The compositions of the oils from the two types of trees are remarkably different with the fragrance and quality of East Indian sandalwood oil (EISO) considered superior to Western Australian sandalwood oil (WASO). One clear difference between the two oils is the content of α-santalol, which can vary from ~20% of the total oil content in WASO to more than 50% in EISO.</p><p>α-Santalol is one of the primary components of sandalwood oil and has been recently investigated for its chemopreventive properties. In fact, purified α-santalol, as well as sandalwood oil, has previously been demonstrated to prevent skin tumor development in mice [2-7]. Cell-based studies have found that α-santalol activates proapoptotic caspases, induces G2/M cell cycle arrest and blocks inflammation, which may be responsible for the prevention of tumor development after UV exposure [8, 9]. Although these studies identified some of the chemopreventive properties of α-santalol, little is known about the essential oil from which it was extracted and its potential value in preventing UV-induced skin cancer.</p><p>In the current study, we investigated the effects of EISO on cultured HaCaT keratinocytes. HaCaT cells were established from adult sun-damaged skin and are well characterized as representing an initiated human keratinocyte cell line expressing mutant dysfunctional p53 and a defective NF-kB signaling pathway, both of which are commonly found in UV-initiated keratinocytes in human skin [10-12]. We investigated the use of EISO on cells irradiated with UVB light. Although UVB light comprises only 1-10% of solar UV light, UVB acts as a complete carcinogen capable of activating signaling pathways in HaCaT cells known to stimulate cell proliferation and survival, including p38, JNK, ERK and PI3-K upstream of activator protein-1 (AP-1) transcription factor activation. UVB-induced AP-1 activity has been linked to cellular proliferation and survival, and in mouse skin AP-1 activation has been demonstrated to be a major cause of skin cancer [13]. In addition, we investigated the effects of EISO on cell cycle progression and cell membrane integrity. We determined that the effects of EISO are more pronounced in proliferating cells than in quiescent cells. Loss of cell membrane integrity and expression of a prominent marker of autophagy were both clearly more prominent in cells stimulated to grow than in serum-starved quiescent cells. These findings suggest that EISO may be valuable as a topical nondestructive chemopreventive agent through selective targeting of proliferative cancer and pre-cancerous cells.</p><!><p>EISO was provided by Santalis Pharmaceuticals, Inc. (San Antonio, TX). Annexin V antibodies, bovine pituitary extract (BPE) and epidermal growth factor (EGF) were obtained from Life Technologies/Invitrogen (Grand Island, NY). Propidium Iodide was purchased from Sigma-Aldrich (St. Louis, MO). Antibodies used for Western blot analysis of signaling proteins (phospho-ERK, phospho-p38, phospho-JNK, phospho-Akt) were all purchased from Cell Signaling Technology (Danvers, MA). All other reagents were purchased from Sigma-Aldrich.</p><!><p>The human keratinocyte cell line, HaCaT, was established from cells obtained from adult sun damaged skin and have been described previously[10-12]. HaCaT cells contain UV-signature mutations and express mutant dysfunctional p53 and a defective NF-kB signaling pathway, which are common findings in UV-initiated keratinocytes in human skin. However, these cells maintain normal signaling in response to UV irradiation compared to normal human keratinocytes. HCL-14 and FL-30 cells are HaCaT-derived cell lines that stably express a luciferase reporter gene driven by a portion of the human collagenase I promoter containing a single activator protein-1 (AP-1) binding site, or a full length human c-fos promoter, respectively [14, 15]. The cells were cultured in Dulbecco's modified Eagle's Medium (DMEM) with 10% fetal bovine serum (FBS) and 100 units/ml penicillin/streptomycin at 37°C and in 5% CO2. In experiments using growth arrested cells, the cells were cultured to 80–90% confluence and maintained in serum-free DMEM (SFM) for 24 h prior to UVB exposure. In experiments using proliferating cells, 25 μg/ml (0.25% v/v) BPE and 0.2 ng/ml EGF was added back to cells after serum starvation for 3 hr prior to subsequent treatment.</p><!><p>HaCaT cells were treated with EISO diluted to a 1000× stock concentration in DMSO. Control cells were treated in the same manner with DMSO alone. Cells were treated for 1 hr prior to irradiation with ultraviolet B light. HaCaT cells were washed once in PBS and irradiated with a dose of 250 J/m2 UVB using a bank of two SF20 UVB lamps (National Biological Corp., Beachwood, OH) providing a peak emission of 313 nm. Control cells were treated in the same manner and mock irradiated. Following irradiation, HaCaT cells were again washed with PBS and returned to DMEM. During the post-irradiation incubation, cells were again treated with the indicated dose of EISO.</p><!><p>An MTS assay to assess cell viability after EISO treatment was performed using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (Promega). HaCaT cells were plated in 96 well plates and cultured and treated with EISO as indicated in triplicate. At the end of the treatment period, DMEM containing EISO was removed and replaced with 100 μ1 fresh SFM and 20 μ1 MTS substrate and incubated at 37°C for 30 minutes for color development. Absorbance was read at 490 nm and triplicate values averaged. Results representative of n = 3.</p><!><p>HaCaT keratinocytes were serum-starved for 24 hr before pretreatment with EISO (0.0005% or 0.001% in DMSO) for 1 hr prior UVB irradiation. Cells were then washed with PBS and irradiated with 250 J/m2 of UVB. Then the cells were incubated with EISO (0.0005% or 0.001%) in serum-free DMEM media for an additional 6.5 hours. Floating cells were collected and pooled with adherent cells removed with trypsin. Cells were spun briefly and cell pellets incubated with Alexa 488-conjugated anti-annexin V (AnnV) antibodies and propidium iodide for 30 minutes at room temperature in the dark. Labeled cells were counted by flow cytometry, collecting a total of 10,000 data points per treatment condition using a FACSCanto II device (BD Biosciences).</p><!><p>HaCaT cells were starved in SFM for 24 hr. Cell culture medium was then changed to SFM or to SFM supplemented with 25 μg/ml bovine pituitary extract and 0.2 ng/ml EGF to stimulate cellular proliferation. After 3 hr incubation, HaCaT cell were treated with EISO for 24 hr. At the end of the treatment period, adherent cells were detached with trypsin and combined with floating cells, pelleted by centrifugation and incubated in 70% EtOH overnight at -20°C. The following day, cells were resuspended in PBS and incubated with 40 μg/ml propidium iodide (PI) and 0.5 mg/ml RNase A for 15 minutes. Single-color flow cytometric-based cell cycle analysis was performed using a BD FACScan (BD Biosciences) equipped with an air-cooled 15-mW argon ion laser tuned to 488 nm. Propidium iodide was detected in the FL2 detector through a 585/42 bandpass filter. List mode data files were acquired and analyzed using CellQuest PRO software (BD Biosciences). A total of 10,000 events were recorded and analyzed. Analysis of DNA content yielded individual populations of cells in G1 or G2/M phases, with S phase cells having intermediate DNA content. Data are expressed as % cells in each cell cycle phase. Three independent experiments were normalized so that the DMSO control equals 1 and the results were pooled for statistical analysis.</p><!><p>Serum-starved HaCaT cells were pretreated with BPE/EGF and/or EISO as described above, irradiated with 250 J/m2 UVB, and incubated in SFM supplemented with the same pretreatment compounds. HaCaT cells were lysed in RIPA buffer containing protease and phosphatase inhibitors as described previously [16]. Membranes were blocked in Tris buffered saline containing 0.1% Tween-20 (TBST) and either 5% nonfat dry milk or 5% BSA (for phospho-specific antibodies). Primary antibodies in 5% BSA were incubated overnight at 4°C. After washing in TBST, membranes were incubated with HRP-conjugated secondary antibodies and then washed extensively in TBST. Antigen-antibody complexes were detected using Amersham ECL Detection Reagent (GE Healthcare, Buckinghamshire, UK). Results representative of n = 3.</p><!><p>HCL-14 and FL-30 cells were treated and UVB-irradiated in triplicate for each independent experiment. Twelve hours after exposure to UVB, cells were lysed and a total of 20 μg protein per sample replicate were assayed for luciferase activity according to the manufacturer's instructions for the Luciferase Assay System (Promega, Madison, WI) using a TD 20/20 luminometer (Turner Designs, Sunnyvale, CA). The experiment triplicates were averaged and the means from each independent experiment were subsequently averaged and analyzed by Student's t test for statistical significance. Results representative of n = 3.</p><!><p>Cultured HaCaT cells were treated with BPE/EGF and/or EISO as described above and then adherent cells were detached by trypsinization and pooled with floating cells. Cells were pelleted, resuspended in PBS, and incubated with 2 μg/ml non-fluorescent fluorescein diacetate (FDA) for 15 min in the dark prior to analysis of FDA hydrolysis to fluorescein and retention of the fluorescent signal by flow cytometry. The emission fluorescence of cellular fluorescein was detected and recorded through a 530/30 bandpass filter in the FL1 detector. List mode data files were acquired and analyzed using CellQuest PRO software. Results representative of n = 2.</p><!><p>HaCaT cells plated on 8-well chamber slides were treated with EISO as indicated for 24 hr. At the end of the treatment period, cells were fixed in 2% paraformaldehyde in PBS for 30 min, permeabilized with 1% Triton X-100/PBS for 30 min, blocked in 2% goat serum for 1 hr, and incubated with rabbit anti-LC3 A/B primary antibodies (Cell Signaling Technologies, Danvers, MA) for 1 hr. Cells were then washed and incubated with Alexa Fluor 488 conjugated anti-rabbit secondary antibodies. Excess antibody was washed five times, and cells were mounted in Vectashield mounting media containing DAPI, a nuclear stain (Vector Laboratories, Burlingame, CA).</p><!><p>A major active component of sandalwood oil, α-santalol, has been reported to induce apoptosis in skin and cancer cell lines [3, 4, 16-22]. However, little is known about the cellular response to pure sandalwood oil. Understanding this response is crucial if essential oils are to be used safely in cosmetic formulations. For initial testing, an MTS assay was performed to assess the general cellular effect of EISO on cell proliferation and/or viability in cultured HaCaT cells (Figure 1A). MTS conversion to a soluble formazan product occurs in the presence of live cells and the extent of conversion is dependent on cell number. HaCaT cells were serum starved for 24 hr and then treated with EISO. Cells treated with EISO up to 0.0005% were not adversely affected, but instead showed an increase or no effect in proliferation and/or viability compared to DMSO-treated controls. HaCaTs began to show reduced conversion of the MTS substrate at concentrations above 0.0005% EISO. Percent of MTS conversion was 53% of the DMSO control at 0.001% EISO and approximately 3% at concentrations of 0.002% and higher, suggesting a very narrow therapeutic index in cultured HaCaT cells. Cell density also impacted the effectiveness of EISO treatment (Figure 1B). Increased cell densities reduced the effectiveness of EISO treatment at both 0.001% and 0.002% final EISO concentration. This factor was taken into account in all subsequent experiments.</p><!><p>To determine whether the result of reduced MTS conversion was due to inhibition of proliferation or induction of cell death, cell cycle analysis and assessment of apoptosis by annexin V staining were both performed on serum-starved EISO-treated HaCaT cells. We found that EISO did not induce apoptotic death of HaCaT cells, (Figure 2A). EISO treatment (0.0005% and 0.001%) resulted in no significant increase in annexin V staining nor was there an increase in propidium iodide staining. Moreover, there was no significant difference observed between EISO treated cells compared to control cells after UVB-induced apoptosis. Cell cycle analysis revealed that in serum-starved cells, EISO induced S-phase arrest in a concentration-dependent manner, suggesting that EISO was stimulating quiescent G0/G1 phase cells to enter S phase while simultaneously preventing them from progressing completely through S phase (Figure 2B). There was also a mild dose-dependent rise in cells in the G2/M phase. To better understand this effect of EISO, a second experiment was performed in which cells were first stimulated to proliferate prior to EISO treatment. Serum-starved HaCaT cells were treated with a cocktail of bovine pituitary extract (BPE) and epidermal growth factor (EGF) for 3 hr before treatment with EISO. Cell cycle analysis revealed a trend toward G2/M block observed with the highest dose of 0.002% EISO, but experimental variability precluded statistical significance for these findings (data not shown). These data indicate an anti-proliferative, yet not apoptotic effect of EISO on cultured HaCaT cells.</p><!><p>Activation of AP-1 is critically important in UVB-induced skin carcinogenesis [13]. Treatment of HaCaT cells with EISO up to a concentration of 0.0005% resulted in a concentration-dependent reduction in AP-1 activity measured through the use of a luciferase reporter cell line, HCL-14 cells (Figure 3A). Concentrations of 0.001% EISO and above resulted in significant detachment of HCL-14 cells from the culture dish, causing AP-1 activity to be unmeasurable. PI3-K and p38 MAPK are known to be activated by UVB and coordinate upstream signaling that results in c-fos gene expression leading directly to increased AP-1 activity in HaCaT cells [23]. HaCaT cells were treated with 0.0005% EISO to determine the effect of the compound on signaling upstream of AP-1. Figure 3B shows that EISO treatment slightly potentiated UVB-induced phosphorylation of p38, ERK and Akt, instead of causing the expected response of inhibition of activity. EISO greatly potentiated phosphorylation of JNK by UVB. Therefore, the signaling through PI3-K, p38, ERK and JNK were intact in the presence of EISO, suggesting that AP-1 inhibition was through an alternate mechanism. To confirm this finding, we measured UVB-induced c-fos promoter activity in FL-30 cells and observed no significant decrease in c-fos-induced luciferase expression, consistent with the lack of inhibition of upstream signaling (Figure 3C). These findings suggest that the effects of EISO are not mediated though inhibition of UVB-induced signaling.</p><!><p>Serum-starved HaCaT cells were treated with EISO and/or BPE/EGF and loaded with cell-permeable, non-fluorescent fluorescein diacetate (FDA). Conversion of FDA to fluorescein and retention of the signal was measured over a period of 12 hrs in 3 hr increments. We discovered that in cells stimulated with BPE/EGF to induce proliferation, sensitivity to EISO treatment was much higher than in quiescent serum-starved cells, resulting in a more rapid loss of fluorescein signal (Figure 4 – compare 4C R7 population to that of 4D R7). Cells were also labeled with propidium iodide to confirm that the plasma and nuclear membranes were compromised during the EISO treatment. Through flow cytometric analysis, we identified a population of cells with intact DNA which confirmed that early in the treatment period (6 hr), cell membranes were being compromised by EISO but biochemical markers of cell death were not apparent.</p><!><p>We were interested in determining whether the blockade of cell proliferation, AP-1 activation, and compromised membrane integrity involved the induction of autophagy of EISO-treated HaCaT cells. To assess this, we treated serum-starved and BPE/EGF-stimulated cells as indicated and harvested lysates to measure the expression and cleavage of LC3 by Western blotting and immunocytochemistry (ICC), since LC3 is a key biomarker in the autophagic process [24]. ICC was performed on serum-starved cells only since EISO treatment of BPE/EGF resulted in a significant loss of cells due to detachment. ICC revealed that expression of LC3 was increased by very low doses of EISO (0.0002%) and was predominantly nuclear and perinuclear (Figure 5). Western blot analysis of both serum-starved and BPE/EGF stimulated cells revealed that 0.001% EISO induced accumulation of LC3 II in both populations, although this was less pronounced in quiescent cells (Figure 6). Irradiation with UVB also induced LC3 II, and this induction was again less dramatic in serum-starved cells. The effect of EISO and UV irradiation on LC3 processing was additive, resulting in the LC3 II band being largest in BPE/EGF+EISO+UVB treated HaCaT cells. Poly ADP ribose polymerase (PARP) cleavage, a marker of apoptosis, was induced as expected in control cells treated with UVB both in serum starved and BPE/EGF-stimulated conditions. Exposure to 0.001% EISO + DMSO resulted in minor PARP cleavage in both treatment conditions. However, this cleavage was not notably increased when EISO-treated cells were exposed to UVB, although LC3 II clavage was highest in these samples (lanes 4 and 8 in Figure 6).</p><!><p>In the current study, our goal was to evaluate EISO as a novel chemopreventive agent, particularly because it is already in widespread use as a cosmetic, a fragrance and traditional medicine but also since the properties of one of the primary components of the oil, α-santalol, has been investigated as a preventive agent against skin cancer. It is interesting that although sandalwood oil has been utilized by cultures across the world for its health benefits for centuries, very few studies on the chemopreventive properties of the oil have been published [2, 5]. We propose that there are several potential mechanisms by which EISO may act as a chemopreventive agent, and in the current study we investigated the effect of EISO on apoptosis and cell death, cell proliferation, and inhibition of key signaling pathways that mediate the cellular response to UV. Recently, a study using purified α-santalol in vivo indicated that UVB-induced apoptosis, inflammation, proliferation and cell cycle control were all being affected by treatment with this compound, the net effect being significant reduction in UV-induced tumorigenesis in SKH-1 mice [6]. The EISO used in this study contains 45-50% α-santalol. We were interested in utilizing the extract instead of purified α-santalol because many cosmetics and natural remedies use the full extract, suggesting that the presence of other components may affect functionality. To our knowledge, this is the first time that purified EISO has been evaluated as an agent suitable for use as a chemopreventive substance against skin carcinogenesis.</p><p>We determined that treatment of cultured HaCaT keratinocytes with EISO alone does not induce apoptotic cellular responses, contrary to what has been previously reported for treatment with purified α-santalol [19]. However, EISO did induce growth arrest in an interesting manner that was dependent on the proliferative state of the cells. In quiescent (serum and hormone-starved) cells primarily in the G1/G0 phase, EISO-treated cells entered into S-phase but then primarily failed to progress into the G2 or M phase except at the highest EISO doses 24 hr post treatment. In proliferating HaCaT cells (serum-starved cells stimulated with BPE and EGF for 3 hr prior to treatment), EISO treatment resulted in a trend toward cell cycle blockade in the G2/M phase, although sample variability precluded finding significance in this experiment (data not shown). G2/M phase blockage has previously been reported in skin cells and in prostate cancer cells treated with α-santalol [21]. One possible explanation for this noted difference of the effect of EISO in quiescent versus proliferating cells is that the S-phase checkpoint through which the quiescent cells failed to progress was already passed by the proliferating cells. This suggests that there are at least two points in the cell cycle at which cell proliferation is inhibited by treatment with EISO. Since skin cells are largely quiescent in vivo, this finding supports the hypothesis that EISO has chemopreventive properties against the development of skin cancer.</p><p>We next investigated signaling responses commonly activated in keratinocytes by UV light to identify a possible mechanism by which EISO was inhibiting cell growth and proliferation. Information on the effects of sandalwood oil or α-santalol in this context is unavailable, as previous studies have either not investigated the effect or any findings from such studies have not been reported. To our surprise, unlike many other agents being investigated for chemopreventive activities, EISO had no inhibitory effect on the UV-stimulated PI3-K/Akt signaling pathway or on MAPK signaling pathways, instead slightly stimulating activation of these pathways even in control conditions. Interestingly, we discovered that UV-induced AP-1 signaling was significantly inhibited by EISO treatment and that the inhibition occurred in a dose-dependent manner. However, consistent with our finding that signaling pathways upstream of AP-1 activity were not affected by EISO treatment, c-Fos promoter activity was not inhibited by EISO. These findings argue that EISO may also elicit chemopreventive action by direct inhibition of AP-1 activity, a major known causative factor in UV-induced skin cancer [13]. There is precedence for direct inhibition of UV-stimulated AP-1 by other natural products in the literature [25].</p><p>We were interested in assessing the effects of EISO chemistry on HaCaT cells, specifically on plasma membrane integrity. The hydrophobic properties of the oil could potentially compromise plasma membrane integrity and cause cell death through a mechanism that is independent of the induction of apoptotic signaling or inhibition of cell survival mechanisms, as was seen with Annexin V staining, PARP cleavage and Akt activation. We found that cellular retention of a cell permeable, hydrolyzable fluorescein molecule, fluorescein diacetate, was compromised by treatment of cells with EISO in a dose-dependent manner, although the dose response occurred in a very narrow treatment window arguing that there is a maximum tolerance threshold for EISO in HaCaT cells. We monitored the effect of EISO on plasma membrane integrity over a period of 12 hr in both quiescent and in hormone-stimulated proliferating HaCaT cells and found that proliferating cells were more sensitive to EISO treatment than quiescent cells, although plasma membrane integrity was compromised in both. The reason for this differential effect is unknown, but it could be postulated that the membrane is more fluid in cells undergoing cell division, which may prime the cell membranes for more effective disruption by EISO treatment. Regardless of the cause of this observation, this finding highlights the fact that proliferative cells within an organ containing predominantly quiescent cells, like those found in a neoplasm of the skin, may be more responsive to treatment with EISO. In addition to this finding, we observed an increase in the apparent expression and processing of LC3, indicating an autophagic response to EISO treatment, and the significance of this response is unknown. Autophagy is a complex mechanism with a number of potential cellular effects, including the removal of organelles damaged by cytotoxic treatment as a way for tumor cells to escape chemotherapeutic treatment and as an alternative method to initiate cell death when apoptotic mechanisms are inactive or inhibited. However, it is now appreciated that cell death occurs both through, and in spite of, autophagy [26-29]. Our findings regarding LC3 induction and processing in response to EISO treatment along with a severe impairment of plasma membrane integrity and the lack of an apoptotic response suggests that EISO treatment induces organelle damage that, if sufficiently severe, ultimately leads to cell death.</p><p>α-Santalol, and its isomer, β-santalol, are the primary components of EISO, and although some of the cellular effects of treatment with the α-isomer have been investigated, it is likely that there are multiple targets for these two compounds. Additionally, there are numerous other ingredients in EISO that likely induce cellular responses, and it is therefore to be expected that treatment with the essential oil mixture will have multiple specific cellular effects.</p><!><p>In summary, we determined that EISO treatment of HaCaT cells resulted in:</p><!><p>a concentration-dependent inhibition of AP-1 activity through a mechanism independent of upstream signaling pathway inhibition,</p><p>an inhibition of cellular proliferation, as evidenced by a blockade of cell cycle progression,</p><p>a severely compromised plasma membrane resulting in an impaired ability to retain an intracellular fluorophore, and</p><p>induction of LC3 processing in the absence of any evidence of apoptosis, suggesting an initiation of autophagy.</p><!><p>Due to the complex nature of EISO, it is unclear as to whether these cellular effects are dependent on one another or if they occur independently. However, each of these effects suggests the need for continued study of EISO for use in protective and beneficial skin treatments. For example, the inhibition of AP-1 signaling and cell cycle progression may have occurred as a direct result of EISO treatment or since AP-1 activity drives cellular proliferation. However, it is also conceivable that cell proliferation was inhibited due to the direct effects of EISO on cellular membranes as a mechanism by which the cells conserve energy and first repair the membrane damage before resuming normal proliferation. Regardless of the precise mechanism, the blockade of cellular growth and the induction of cell death, to a greater extent in cells stimulated to divide, both support the hypothesis that EISO acts as a chemopreventive agent in skin cells. These studies are the first of their kind to investigate potential chemopreventive properties of purified East Indian sandalwood oil and support the hypothesis that EISO would be an effective treatment for pre-cancerous conditions such as AK and the prevention of skin cancer.</p>

PubMed Author Manuscript

Photophysical Properties of Benzophenone-Based TADF Emitters in Relation to Their Molecular Structure

Thermally activated delayed fluorescence (TADF) materials are commonly used in various apparatus, including organic light-emitting device-based displays, as they remarkably improve the internal quantum efficiencies. Although there is a wide range of donor–acceptor-based compounds possessing TADF properties, in this computational study, we investigated TADF and some non-TADF chromophores, containing benzophenone or its structural derivatives as the acceptor core, together with various donor moieties. Following the computational modeling of the emitters, several excited state properties, such as the absorption spectra, singlet–triplet energy gaps (ΔEST), natural transition orbitals, and the topological ΦS indices, have been computed. Along with the donor–acceptor torsion angles and spin-orbit coupling values, these descriptors have been utilized to investigate potential TADF efficiency. Our study has shown that on the one hand, our photophysical/structural descriptors and computational methodologies predict the experimental results quite well, and on the other hand, our extensive benchmark can be useful to pinpoint the most promising functionals and descriptors for the study of benzophenone-based TADF emitters.

photophysical_properties_of_benzophenone-based_tadf_emitters_in_relation_to_their_molecular_structur

Introduction<!><!>Introduction<!><!>Introduction<!>Methodology<!>Results and Discussion<!><!>Results and Discussion<!><!>Benchmark Calculations and UV–vis Absorption Spectra<!>D–A Torsion Angles<!><!>D–A Torsion Angles<!>NTOs and ΦS Indices<!><!>NTOs and ΦS Indices<!><!>Singlet–Triplet Energy Gaps (ΔEST)<!><!>Singlet–Triplet Energy Gaps (ΔEST)<!>Spin-Orbit Couplings<!><!>Spin-Orbit Couplings<!>Conclusions<!>

<p>Organic light-emitting diodes (OLEDs) have attracted widespread attention since the invention of the first organic electroluminescent device in 1987.1 As compared to conventional LEDs and liquid crystal display systems, OLEDs do not require a backlight unit, as they are self-illuminating. Given this distinct feature, OLEDs offer several advantages, such as flexible device structures, decreased panel thickness, improved brightness, and reduced power consumption, rendering them most suitable for various devices.2 Nevertheless, OLEDs also display serious efficiency drawbacks, which are grounded in the fundamental spin statistics rule (Figure 1) because the population of the nonemissive triplet state comprises 75% of the generated excitons in the device, hence leading to the loss of two thirds of the applied energy.3 Phosphorescent organic light-emitting diodes (PhOLEDs), which contain heavy atoms, can improve the efficiencies of OLEDs with the help of enhanced intersystem crossing (ISC) and improved phosphorescence rates as a result of increased spin-orbit coupling (SOC).4,5 However, heavy metals instigate environmental issues while significantly increasing the cost of the device, thus limiting the commercial availability of PhOLEDs.6</p><!><p>Schematic representation of ISC, RISC, TADF, and phosphorescence processes in a Jablonski diagram along with the possible design strategies for TADF emitters.</p><!><p>Thermally activated delayed fluorescence (TADF) materials, first proposed by Adachi et al. in 2011, represent a most suitable alternative to overcome spin statistics burdens and achieve high internal quantum efficiencies (IQEs) in OLED displays.7 In TADF processes, the population of triplet excitons undergoes a slow reverse intersystem crossing (RISC), re-populating the emissive first excited singlet state (S1) and hence avoiding the inclusion of heavy metals. To assure efficiency, RISC or the upconversion of the triplet states [up-intersystem crossing (UISC)] require a small S1-T1 energy gap (ΔEST). UISC can be achieved by promoting the population of intramolecular charge-transfer (ICT) states and is usually, albeit not necessarily, correlated with a small gap between the highest occupied molecular orbital and lowest unoccupied molecular orbital. To achieve efficient ICT, the molecular architecture requires the presence of a donor (D) moiety bridged to an acceptor (A) (Figure 1). The D and A groups can be separated by using bulky substituents to increase steric effects and maintain orthogonal molecular structures, hence minimizing conjugation and delocalization, or alternatively by π-bridges.8,9</p><p>Benzophenone is a widely used building block in the design of OLED devices, in part due to its strong and efficient electron accepting and efficient ultraviolet (UV) absorbing abilities, as well as its high ISC efficiency.10−13 Adachi and co-workers showed in 2014 that efficient deep blue TADF could be achieved by using benzophenone-based D-A-D frameworks.12 Thus far, many benzophenone-based TADF emitters possessing a small ΔEST have been reported to feature full-color delayed fluorescence emission, usually in a range from deep blue to green and with external quantum efficiencies up to 14.3%.14 Benzophenone-based luminogens can also be used to induce aggregation-induced delayed fluorescence.15,16</p><p>Nonetheless, the rather floppy phenyl moieties of benzophenone itself may induce intramolecular rotations, which enhance nonradiative decay and lead to relatively low reverse ISC rates (kRISC).17 Hence, more compact and rigid benzophenone derivatives, such as anthraquinone,18,19 xanthone,20−23 dibenzothiophene-benzoyl24,25 and phenylcarbazole-benzoyl,26 are deemed more promising in terms of their TADF efficiencies. Figure 2 depicts the molecular structures of benzophenone derivatives commonly used in TADF materials, together with some of the most widely used D groups.</p><!><p>Electron-donating and -accepting moieties frequently used in the design of benzophenone-based TADF emitters.</p><!><p>We used molecular modeling to elucidate the relationship between the molecular structure and optimal photophysical properties that are essential for effective TADF emission. Various descriptors are used to define TADF properties, which include ΔEST, SOC, natural transition orbitals (NTOs), the topological ΦS index quantifying the charge-transfer amount, and the torsion angles between D and A moieties.27−33 The oscillator strength (f) is also a critical parameter to attain a reasonable radiative decay rate (kr) from the S1 state to the ground state. The oscillator strength is usually closer to 0 for orthogonal donor–acceptor compounds, where CT is prominent.29 Negligible oscillator strengths represent a challenge in TADF emitters because it is essential to balance a small ΔEST and a sufficiently large kr to reach high IQE in OLED devices.</p><p>In this study, a series of experimentally studied TADF emitters, employing benzophenone and its derivatives, have been investigated by quantum chemical calculations. Several non-TADF benzophenone chromophores have also been modeled for comparison. Absorption spectra of the selected compounds have been generated, including a sampling of the Franck–Condon region by Wigner distribution, to include dynamic effects. To determine the degree of charge separation in the excited state, NTOs and ΦS indices have been computed, as well as the ΔEST values. NTO and ΔEST calculations have been performed on both the ground state (S0) and the lowest triplet state (T1) equilibrium geometries because geometrical reorganization can also play a role in photophysical processes. Furthermore, the SOCs between singlet and triplet excited states have been computed to better estimate RISC probability.</p><p>The mentioned descriptors have proven to be insightful in determining potential TADF efficiency, as well as shedding light on the correlation between molecular structure and TADF performance. Indeed, the investigation of the photophysical characteristics of both the excited and ground state geometries, and also the assessment of the DFT functionals for the excited state calculations of TADF emitters, may help provide a better understanding for the photophysical processes related to TADF at a molecular level.</p><!><p>All ground state geometry optimizations have been performed using the Gaussian 16 program package,34 and a comprehensive conformational search has been carried out. UV–vis absorption spectra and Boltzmann-weighted ΔEST calculations (Tables S1 and S2) showed that different conformations did not exhibit significant differences in their excitation energies and on the ordering of the excited states. Hence, T1 geometry optimizations and related excited state calculations were carried out solely on the most stable conformer. The M06-2X functional35 has been used together with the 6–31 + G(d,p) basis set for S0 and T1 geometry optimizations. This choice is justified because M06-2X is well known to correctly reproduce medium-range interactions, electronic energies, and equilibrium geometries of compounds with aromatic ring systems.35 However, a dispersion-corrected functional, B3LYP-D3, was also tested for the geometry optimization of rather larger compounds.36 In order to increase the accuracy, 6–311++G(3df,3pd)37 and 6–311++G(2d,2p)38 basis sets have been used for compounds including sulfur and phosphorus, respectively. Calculations have been performed taking into account the experimentally employed solvents by the polarizable continuum model in the integral equation formalism (IEF-PCM). CYLview software package has been used for visualization purposes.39</p><p>Similar to ground state calculations, excited state calculations have been carried out by using Gaussian 16 software package. Tamm–Dancoff approximation (TDA) has been used, as this approach avoids unphysically stable triplet states in CT molecules while maintaining a good description of the singlet excited states, hence yielding more accurate and balanced results for ΔEST calculations.40,41 The absorption spectra as well as the energies of the lowest lying singlet and triplet excited states have been computed with different functionals (B3LYP,42−45 BLYP,42,45,46 PBE0,43,45,47,48 M06-2X,42,45,49 CAM-B3LYP,42,45,50,51 and LC-ωPBE52−54) and 6–31 + G(d,p) basis set, and the performances of these functionals have been evaluated with respect to experimental data. Absorption spectra have been modeled, including the effects of vibrational and thermal motion via Wigner distribution sampling of the equilibrium region on the potential energy surface by generating 40 conformations via the Newton-X program55 on the equilibrium region of the PES.</p><p>NTOs and ΦS indices have been calculated for the S1 state by using the Gaussian 16 and Nancy_EX program packages,56 while hole and electron NTOs have been visualized with the Avogadro program package.57 ΦS index can be defined as the spatial overlap between attachment and detachment densities. Values closer to 1 indicate the presence of local excitation character, whereas values approaching 0 imply that the CT character is dominant.56 SOC values between S1 and T1, and in some cases between S1 and T2, have been calculated using the Amsterdam Density Functional software package by utilizing a DZP basis set.58</p><!><p>The benzophenone emitters investigated in this study have been grouped according to the type of A moieties used. As depicted in Figure 3, Group 1 emitters are in the form of D-A-D, and they contain benzophenone A cores (in red) bridged with various electron-donating groups. They can be further subdivided into two subgroups: symmetric and asymmetric emitters. Symmetric emitters Px2BP,12 DMAC-BP,59 Cz2BP,12 and CC2BP12 contain phenoxazine (PXZ), dimethylacridine (DMAC), and carbazole (Cz) donors (in blue), respectively. Asymmetric emitters, A-BP-TA60 and OPDPO,61 include D groups such as thianthrene, phenothiazene (PTZ), and diphenylphosphineoxide.</p><!><p>Classification of TADF emitters modeled in this study.</p><!><p>Group 2 emitters have fused D-A ring structures where DBT-BZ-PXZ,25 DBT-BZ-PTZ,25 and DBT-BZ-DMAC24 bear the A unit dibenzothiophene-benzoyl (in red), while CP-BP-PXZ26 and CP-BP-DMAC26 have the A unit of phenylcarbazole-benzoyl (in red). D groups such as PXZ, PTZ, and DMAC have been employed in these emitters as well.</p><p>Group 3 emitters have para-substituted structures in which the A unit is anthraquinone.18 These include D-A-D type emitters a1-a4, in which the D units are diphenylamine (DPA), bis(4-biphenyl)amine, 3,6-di-tert-butylcarbazole, and DMAC, respectively. Similarly, in the D-π-A-π-D structures, b1 and b4, the D groups are DPA and DMAC.</p><p>Group 4 emitters have D-A type of structures in which the A unit is xanthone. DMAC, PXZ, and PTZ donors have been used in ACRXTN,23 3-PXZ-XO,20 and PTZ-XT,62 respectively. In MCz-XT,21 1,3,6,8-tetramethylcarbazole is present as the D unit.</p><p>Finally, seven non-TADF compounds have been selected from literature in an attempt to elucidate the main structural differences leading to TADF emission (Figure 4). While some of these non-TADF emitters have D-A type structures, including MC2,63 OPM,64 and p-Cz,65 others (ODFRCZ,66 ODBTCZ,67 C1,15 and C215) possess D-A-D type of structures.</p><!><p>Non-TADF benzophenone emitters.</p><!><p>M06-2X/6–31 + G(d,p) and B3LYP-D3/6–31 + G(d,p) methods have been tested for the ground state geometry optimizations of a series of compounds selected from each group. Table S1 depicts the optimized geometries at both levels of theory, and Table S2 includes the D-A torsion angles for the optimized geometries. Tables S1 and S2 suggest that there is no significant change in the geometries optimized at both levels as both functionals yielded similar structures with very close D-A torsion angles. Because the compounds investigated in this study are relatively small in which long-range interactions are not so dominant, we decided to continue with M06-2X/6–31 + G(d,p) for the geometry optimizations because the latter is computationally more affordable.</p><p>A series of widely used DFT functionals have been chosen for excited state calculations, BLYP, B3LYP, PBE0, and M06-2X. More specifically, UV–vis absorption spectra have been generated for a group of emitters, and the results have been compared with the available experimental spectra.</p><p>The conformational space has been taken into account in the modeling of four compounds (Px2BP, Cz2BP, DBT-BZ-PXZ, and DBT-BZ-DMAC) (Table S1). Their absorption spectra have been obtained as the union of the spectra of all the conformers according to their Boltzmann weights. The merged spectra have then been compared to those obtained from the most stable conformation only. Table S3 clearly demonstrates that weighted conformations exhibit similar photophysical features and spectral shapes as the most stable conformation, the maximum deviation being around 3 nm for B3LYP (DBT-BZ-PXZ), 4 nm for PBE0 (DBT-BZ-PXZ), and 2 nm for M06-2X (Cz2BP). In addition to the UV–vis absorption spectra, different conformations of Cz2BP and CC2BP have very similar ΔEST values, and the Boltzmann-weighted ΔEST values of Cz2BP and CC2BP were found to be in good agreement with the experimental findings (Table S4). Thus, henceforth, the most stable conformer will solely be considered. In Tables S5–S7, the absorption spectra for Group 1 emitters and three emitters from Group 2 have been computed by using BLYP, B3LYP, PBE0, and M06-2X. The results suggest that the spectra obtained with M06-2X exhibit hypsochromic shift as compared to the experimental spectrum.</p><p>As opposed to M06-2X, BLYP produced bathochromic shifts and lower ΔEST values (energies are given in Table S8; histogram charts are depicted in Figures S1 and S2) when compared with the experimental data. This behavior is most probably due to well-known unphysical stabilization of CT states by LDA functionals.68 In fact, hybrid B3LYP and PBE0 functionals yield the best agreement with experimental spectra and ΔEST. In order to further confirm this finding, two long-range corrected functionals, CAM-B3LYP and LC-ωPBE, were tested for selected compounds. Table S9 depicts the ΔES1-T1 values (eV) calculated with BLYP, B3LYP, PBE0, CAM-B3LYP, and LC-ωPBE for the S0 and T1 optimized geometries of the compounds. Similar to M06-2X, the ΔEST values calculated with the long-range corrected CAM-B3LYP and LC-ωPBE functionals are significantly higher as compared with experiment and the other functionals. It is also noteworthy that BLYP gave results consistent with the experiment for a few molecules (Px2BP and DMAC-BP). This is probably due to the fact that, due to the relatively small size of the compounds, the CT states are not long-range and hence can be correctly reproduced by hybrid functionals as already observed for instance in some organometallic compounds, hence outperforming medium- or long-range corrected functionals.69</p><p>In an attempt to take into account the role of excitation on the molecular geometry, the ΔEST values have been computed from the S0, S1, and T1 geometries with B3LYP, PBE0, and BLYP functionals for a series of compounds, and the results have been reported in Table S10. The results clearly indicate that the ΔEST values computed from the S0 and S1 geometries are usually very similar. This can be explained by the similar molecular structures obtained for S0 and S1 geometries (Figure S3), as TADF efficiency relies on the molecular geometry to a great extent. Also, considering the computational cost of S1 geometry optimizations, for the rest of the calculations, the photophysical properties have been calculated solely with the S0 and T1 geometries.</p><p>The absorption spectra of all other molecules have hence been calculated with BLYP, B3LYP, and PBE0 functionals (Figures S4–S8). From the results, it can be deduced that compounds with more rigid electron-donating groups such as PXZ, PTZ, and DMAC possess a broadened band (∼350 to 500 nm) appearing after the high intensity local excitation band (∼300 nm), which can be ascribed to the presence of CT states.70−72 Nevertheless, the reasonably rigid emitters of Group 4 also have blue-shifted absorption spectra, which is most likely due to the less efficient nature of the specific D-A type architecture. Indeed, the presence of a unique D unit might decrease the CT character of the electronic transitions, and coherently reduced T1-S1 upconversion efficiency has been previously reported for some D-A type emitters.73</p><p>Oscillator strengths calculated with B3LYP, PBE0, and BLYP functionals for the transitions from S0 to S1, and the reorganization energies between S0 and S1 geometries have also been reported in Tables S11 and S12, respectively. While the latter remain low to moderate and are between around 15 and 6 kcal/mol, no systematic behavior between the different classes of molecules can be underlined, thus hampering the use of this parameter to preview TADF efficiency. Low to moderate geometrical distortion and reorganization energies may also be another reason for the very close ΔEST values calculated from the S0 and S1 geometries. The oscillator strengths were calculated to be near-zero for most of the TADF emitters with very low ΔEST values and ΦS indices, for which we assume that the CT characteristics are more pronounced. As is the case in this study, it was previously reported that the oscillator strengths computed at the TDA or full TD-DFT level might be erroneous, especially for the transitions involving CT states.74</p><!><p>Several computational TADF descriptors were assessed in order to analyze and investigate the structural and photophysical phenomena related to TADF emissions. One of the most critical TADF descriptors is the torsion angle between the D and A units. The torsion angle must be close to 90° in an ideal TADF emitter because an effective CT configuration can be achieved by spatially separating the hole and electron densities and by breaking the conjugation pattern.75</p><p>The S0 and T1 optimized geometries of all emitters and the conformation energies (M06-2X/6–31 + G(d,p)) are given in Supporting Information (Tables S13–S17). Figure 5 demonstrates the torsion angles measured for all emitters investigated in this study.</p><!><p>Donor–acceptor torsion angles for the S0 and T1 optimized geometries of the investigated emitters shown with stacked histograms (M06-2X/6–31 + G(d,p)).</p><!><p>Because symmetrical emitters have similar torsion angles for each D–A bond, only one of the torsion angles has been shown in Figure 5, while numerical values for torsion angles depicted in Figures S9–S13 are reported in Table S18. For asymmetrical emitters, the torsion angle of the electron-donating unit has been ascribed based on the NTO analysis discussed in the following sections. Similarly, if the electron–hole density is localized on a different electron-donating unit in the T1 equilibrium geometry, the corresponding torsion angle has been considered for the analysis of the T1 equilibrium geometries.</p><p>It is evident that the emitters containing rigid electron-donating moieties, including DMAC, PXZ, and PTZ, have significantly higher D–A torsion angles. While reasonably high torsion angles have been observed for Group 1, Group 2, and Group 4, Group 3 emitters bear relatively low torsion angles approximately in the range of 30–35°. This feature can be due to the presence of less rigid DPA donors whose low steric hindrance may lead to highly planar structure with extended π-conjugation patterns. The highest torsion angles belong to a4 and b4 among Group 3 emitters due to the presence of DMAC electron donors in these groups. Indeed, the additional dimethyl units in the acridine structure impose sterical constraints that restrict the free rotation of the D. Carbazole (Cz)-containing compounds also exhibit low torsion angles, around 55°, which are slightly higher than that for the DPA-bearing compounds. Interestingly, the majority of the non-TADF compounds, where the D unit is Cz, have notably lower torsion angles, along with two compounds from Group 1 emitters (Cz2BP and CC2BP). The torsion angles measured for Group 4 emitters are satisfactorily high and close to 90° regardless of the electron-donating unit. This situation can be attributed to the rigid xanthone skeleton that fixes the D–A torsion angle at a desirable position. This observation is in line with the study of Kreiza et al.17 Nonetheless, the obtained results suggest that such a descriptor can only provide a partial understanding of TADF efficiency because Group 3 compounds, which are known as TADF emitters, possess very low torsion angles, despite the presence of the rigid anthraquinone A. Hence, a finer analysis of the property and reorganization of the electronic density should be considered to improve TADF efficiency predictions.</p><p>While in most cases the torsion angles are similar for both T1 and S0 equilibrium geometries, there is a sharp decrease in the torsion angle for T1 equilibrium geometry in A-BP-TA. Indeed, the hole density in the triplet state is localized on the less rigid electron D thianthrene, causing a rather important planarization of the global structure. This will be better understood in the following section.</p><!><p>NTOs and ΦS indices have been calculated for all emitters in an attempt to determine the electron–hole density reorganization. Accordingly, electron and hole NTOs describing electronic density reorganization in the S1 state are shown in Tables S19–S28. Only the NTOs generated from B3LYP/6–31 + G(d,p) densities are given because no remarkable difference has been observed with BLYP and PBE0. ΦS values describing the spatial overlap between attachment and detachment densities76 are also reported in Tables S29–S33.</p><p>In general, the overlap between electron and hole densities has been observed to be high for compounds bearing low D-A torsion angles. This is expected because low torsion angles cause high degrees of π-delocalization and a corresponding planarization of the molecular core. As a result, compounds bearing Cz and DPA donors, including Cz2BP, CC2BP, a1, a2, a3, b1, and most of the non-TADF compounds, show highly overlapping hole and electron NTOs. Interestingly, in non-TADF emitters ODBTCZ and ODFRCZ, the highly flexible dibenzothiophene and dibenzofuran units did not show electron-donating ability at S0 or T1 equilibrium geometries. Obviously, because high ΦS indices are indicative of spatially overlapping densities, the corresponding excited states have a more prominent local excitation character. On the other hand, emitters with more rigid D units (DMAC, PXZ, and PTZ) usually have low amounts of overlap between their NTOs, and consequently small ΦS values, even in the presence of small torsion angles. Hence, even if more computationally expensive, the explicit analysis of the excited state density is much more informative in predicting TADF potential.</p><p>It is also noteworthy that the structural reorganization may have an undeniable impact on the amount of CT from the accessible excited states, as in the case of excited state twisting.77 Indeed, some compounds, particularly those containing butterfly-shaped PXZ, PTZ, or thianthrene units, underwent drastic changes; more specifically, T1 equilibrium geometries exhibited planarization in the butterfly-shaped moieties (Table 1). This, in turn, is related to a change in the shape of the NTOs obtained from the T1 equilibrium geometries. The change in the NTO distribution is indicative of the fact that geometric relaxation leads, for some compounds, such as Px2BP, A-BP-TA, and b1, to the population of a different diabatic state. The most obvious change in the NTO localization pattern is seen in A-BP-TA (Figure 6), where at T1 equilibrium geometry, the hole NTO is localized mostly on the now planar thianthrene unit instead of the rigid DMAC unit. This effect also produces an increase in the ΦS indices at the T1 equilibrium geometry for Px2BP and its meta-substituted analogue C2 (Tables S19 and S28). The electron NTO localizes only on one of the PXZ units, which becomes planar at T1 equilibrium geometry; this feature is also indicative of a consistent pattern, which is observed in two other similar cases. In the π-bridged anthraquinone-based b1 and b4, similar changes have been observed in the transition from S0 to T1 equilibrium geometry (Table S24). These consist of slight rotations in the bridging phenyl links, which cause disruptions in orthogonality and give rise to an increased electron delocalization. For a better visualization, the hole and electron NTOs and ΦS indices, calculated at T1 and S0 equilibrium geometries for two TADF and one non-TADF compounds, are shown in Figure 6.</p><!><p>Hole–Electron levels and ΦS indices calculated with B3LYP/6–31 + G(d,p) for the molecules (a) ACRXTN, (b) MC2, and (c) A-BP-TA.</p><!><p>As expected, non-TADF compounds exhibit higher ΦS, which are due to the presence of weaker electron donors like Cz, dibenzofuran, and dibenzothiophene. However, despite the fact that the only noticeable structural difference is related to the position of the PXZ substituents, the meta-substituted non-TADF C2 also exhibits higher ΦS indices compared to its para-substituted analogue, Px2BP, in line with its classification as a non-TADF compound. ΦS indices calculated with B3LYP for all compounds are depicted in Figure 7, and the ΦS indices calculated with PBE0 and BLYP are given in Figures S14–S15. TADF behavior is clearly inferred from these values, hence making this descriptor quite promising for the discrimination and prediction of optical properties with TADF potential.</p><!><p>ΦS indices for the TADF emitters calculated from S0 and T1 geometries with B3LYP/6–31 + G(d,p).</p><!><p>The energy difference between the first singlet and triplet excited states is an important descriptor for predicting the possibility of TADF emission because the upconversion of T1 to S1 can be achieved only if the energy gap is low enough to be overcome by thermal energy. According to previous studies, ΔEST values above 0.3 eV usually decrease the likelihood of the RISC process, whereas the upconversion of the triplet becomes more likely if the ΔEST value is below 0.1 eV.5</p><p>Even though the RISC process usually takes place from the T1 state, upper lying triplet excited states may also influence the efficiency of RISC if they are in close proximity to T1. Spin-vibronic coupling may cause state mixing between triplet energy levels, giving rise to strongly coupled S1 and T1 states. Therefore, the strong impact of internal conversion on ISC and RISC processes cannot be underestimated.78 Thus, the energy differences between S1-T2 (ΔES1-T2) states have also been reported along with the ΔES1-T1 values when the T2 state lies below S1 for any given compound. The calculated values together with the experimental findings are reported in Tables S34–S38 and in Figure 8.</p><!><p>ΔEST values for TADF emitters calculated from S0 and T1 geometries with B3LYP/6–31 + G(d,p). (Experimental ΔEST values also included).</p><!><p>Figure 8 depicts the ΔES1-T1 values calculated with B3LYP for almost all groups except for Group 3 because ΔES1-T2 values calculated for them are in a better agreement with the experimental findings. It is evident that the lowest ΔEST values have been obtained from the BLYP calculations, which is consistent with the lowest ΦS indices computed with this functional. On the other hand, ΔEST values calculated with B3LYP and PBE0 are closer to the experimental ones. Overall, the calculated ΔEST values are usually consistent with the experimental observation, especially for the S0 geometries. The ΔEST values calculated with PBE0 and BLYP are also given in Figures S16 and S17.</p><p>It is noteworthy that ΔEST values are closely correlated with torsion angles and ΦS indices because the compounds possessing low torsion angles and high ΦS indices, such as Cz2BP, CC2BP, most of the Group 3 compounds, and non-TADF compounds, also have relatively higher ΔES1-T2 and in some cases even high ΔES1-T1 values. Freely rotating donors DPA and sterically less hindered donors Cz lead to higher ΔEST values because of enhanced geometrical relaxation. Despite having DPA donors, b1 has been observed to possess relatively low ΔEST values, which may have been acquired through the phenyl bridges that lead to a higher spatial separation of electron and hole densities. Also, as expected, emitters with DMAC, PXZ, or PTZ donors have satisfying ΔEST values usually at both S0 and T1 equilibrium geometries due to the induced orthogonality. There is a sharp increase in the ΔEST values obtained from the T1 equilibrium geometry of A-BP-TA, where the electron-donating group is the less rigid thianthrene moiety. This is also consistent with the NTO calculations previously mentioned, pointing out the close relationship between ΔEST and electron–hole separation. This result for A-BP-TA also further proves the reliability of NTO calculations and ΦS index in elucidating the excited state properties by relating them to the molecular structure.</p><p>Almost all of the non-TADF compounds have high ΔEST values due to higher amounts of density overlap introduced by freely rotating D–A moieties, which planarize to enhance conjugation. However, C2 is an exception as it exhibits notably lower ΔEST values in spite of its high ΦS index and overlapping NTOs. The calculated ΔEST value is also lower as compared with its experimental ΔEST value, clearly identifying this compound as an outlier. Still, out of 28 benzophenone derivatives investigated in this study, C2 is the only emitter in which an inverse relationship between ΔEST values and ΦS indices was observed, and this can also be resulting from the insufficiency of our methodology in reproducing excited state energies for this compound.</p><!><p>The last descriptor analyzed in this study is the SOC. The mixing of the wavefunctions of singlet and triplet energy levels are indeed strongly dictated by SOC.79 Therefore, by performing SOC calculations, crucial information regarding the feasibility of ISC or RISC processes is obtained. Indeed, SOC matrix elements significantly different from zero are necessary to attain RISC and, hence, TADF emission. However, SOC usually increases with atomic number (Z); thus small organic chromophores are usually less prone to ISC and RISC as compared to heavy-metal containing complexes.80 Additionally, the coupling of CT states produce negligible SOC values as a consequence of El-Sayed's rule, which highlights the need for local triplet excited states for an effective RISC process.79 Efficient TADF emitters usually have lower SOC values due to their orthogonal D–A type molecular backbones, where CT character is more pronounced; this unfavorable factor is, however, compensated by small singlet–triplet energy gaps, which are inversely proportional to the (R)ISC probability. Hence, SOC between the S1 and T1 states of TADF emitters could also indicate the presence of CT states. Because we are dealing with RISC, calculations were solely performed on equilibrium T1 geometries. SOC between S1 and T2 states have also been computed from the T1 geometries if the T2 lies below the S1 state, in a similar approach adopted for ΔEST calculations. The calculated SOC values have been reported in Tables S39–S43. For a better comprehension, the SOC values calculated with B3LYP, BLYP, and PBE0 functionals are shown in Figure 9.</p><!><p>SOC values for TADF emitters calculated with B3LYP, BLYP, and PBE0 using the T1 geometries.</p><!><p>The results indicate that the compounds possessing lower ΔEST values and ΦS indices, such as most of Group 1 emitters (Px2BP, DMAC-BP, and OPDPO), Group 2 emitters, a4 and b4, and Group 4 emitters, exhibit extremely low SOC values as a result of the higher CT character of their S1 and T1 geometries.</p><p>Higher SOC values have been obtained for compounds with less steric hindrance, including Cz2BP, CC2BP, most of Group 3 emitters, and all non-TADF emitters, because local excitations become more favorable in these emitters. A-BP-TA also exhibits relatively high SOC values because its hole density becomes localized on the flexible thianthrene unit in its T1 geometry, reducing the CT character in a similar manner observed in its ΔEST values and ΦS index. Additionally, higher SOC constants obtained for S1-T2 interactions may indicate that the triplet upconversion might be taking place from the T2 rather than the T1 state.</p><p>Table 2 summarizes the D–A torsion angle, ΦS index, ΔEST, and SOC results of the TADF and non-TADF emitters investigated, showing a consistent behavior for all descriptors studied, pointing toward the successful accuracy of NTOs and ΦS in inferring the TADF behavior. More importantly, in most cases, the TADF capability may be correctly predicted by calculating the excited state indicators at the Franck–Condon geometry, hence avoiding rather expensive excited state optimizations.</p><!><p>In this study, the photophysical and structural properties of 21 benzophenone-based TADF emitters and 7 benzophenone-based non-TADF emitters have been investigated by using computational descriptors with the aim to elucidate the factors causing RISC. The main descriptors have been identified as the torsion angle between the D and A, NTOs, ΦS indices, ΔEST values, and SOC constants.</p><p>It was observed that the orthogonality of the molecular backbone is an important factor for achieving a successful RISC process. The compounds adopting a rigid molecular structure where free rotations are more restricted tend to have lower ΔEST values and ΦS indices due to increased ICT through well separated electron and hole densities. Conversely, the emitters possessing less rigid and more freely rotating D-A arrangements were observed to have higher ΔEST values and ΦS indices due to increased local excitation character induced by high amounts of overlap between the hole and electron and enhanced π-delocalization. Although molecular rigidity relies on both the molecular structures of the D and A, it was shown that the role of D units is more dominant, as freely rotating electron donors such as Cz, DPA, or thianthrene disrupt the orthogonality of the compounds, which leads to, in most cases, lower torsion angles (∼30 to 55°). The more rigid electron donors, such as DMAC, PXZ, or PTZ, enhance TADF efficiency through increased orthogonality and electron/hole separation. However, the A rigidity must also be taken into account. The compounds presenting more rigid and planar benzophenone derivatives, such as xanthone, as the A moiety, exhibit satisfying TADF characteristics, regardless of the electron donors. On the contrary, the emitters bearing the benzophenone A, in which the loose phenyl rings have more free rotation, exhibit less satisfying TADF properties. The presence of a π-bridge may also favorably affect the TADF properties owing to the increased spatial separation of D and A units as observed in compounds b1 and b4. Moreover, SOC values were calculated to be lower for the TADF compounds when the presence of highly twisted D-A frameworks induces lower ΔEST values due to enhanced electron/hole separation.</p><p>Excited state calculations were carried out using both S0 and T1 geometries because excited state processes, such as RISC, should take place from the T1 equilibrium geometry. In some cases, a significant geometrical relaxation was observed, in particular for the butterfly-shaped PXZ, PTZ, or thianthrene units, which become planar in T1 geometry. In these cases, different indicators were obtained for calculations performed at S0 or T1 geometries. However, in general, the outcome of the photophysical properties can be correctly inferred by solely considering the Franck–Condon geometry. The benchmark studies revealed that B3LYP and PBE0 can reproduce the excited state energies and UV–vis absorption spectra correctly for almost all emitters, while BLYP yielded satisfying results for several compounds only.</p><p>In summary, this study gives a comprehensive outlook on the relationship between molecular structure and photophysical features of a series of benzophenone-based TADF and non-TADF compounds, employing computational tools and several descriptors. It can be concluded that these descriptors, especially the topological ΦS index, can be safely used to investigate TADF properties of different classes of compounds, as they predict the experimental findings quite accurately.</p><!><p>3D representations of the optimized geometries at M06-2X/6–31 + G(d,p) and B3LYP-D3/6–31 + G(d,p) levels, investigation of the conformational effects, benchmark calculations, UV–vis absorption spectra, investigated donor–acceptor torsion angles, hole and electron NTOs, ΦS indices, low lying singlet–triplet energy gaps, spin-orbit coupling values, and S0, T1, and S1 cartesian coordinates of the most stable conformations optimized at the M06-2X/6–31 + G(d,p) level of theory (PDF)</p><p>jp1c08320_si_001.pdf</p><p>The authors declare no competing financial interest.</p>

PubMed Open Access

Effect of the Nature of Donor Atoms on the Thermodynamic, Kinetic and Relaxation Properties of Mn(II) Complexes Formed With Some Trisubstituted 12-Membered Macrocyclic Ligands

During the past few years increasing attention has been devoted to Mn(II) complexes as possible substitutes for Gd(III) complexes as contrast agents in MRI. Equilibrium (log KMnL or pMn value), kinetic parameters (rates and half-lives of dissociation) and relaxivity of the Mn(II) complexes formed with 12-membered macrocyclic ligands were studied. The ligands were selected in a way to gain information on how the ligand rigidity, the nature of the donor atoms in the macrocycle (pyridine N, amine N, and etheric O atom), the nature of the pendant arms (carboxylates, phosphonates, primary, secondary and tertiary amides) affect the physicochemical parameters of the Mn(II) complexes. As expected, decreasing the denticity of DOTA (to afford DO3A) resulted in a drop in the stability and inertness of [Mn(DO3A)]− compared to [Mn(DOTA)]2−. This decrease can be compensated partially by incorporating the fourth nitrogen atom into a pyridine ring (e.g., PCTA) or by replacement with an etheric oxygen atom (ODO3A). Moreover, the substitution of primary amides for acetates resulted in a noticeable drop in the stability constant (PC3AMH), but it increased as the primary amides (PC3AMH) were replaced by secondary (PC3AMGly) or tertiary amide (PC3AMPip) pendants. The inertness of the Mn(II) complexes behaved alike as the rates of acid catalyzed dissociation increased going from DOTA (k1 = 0.040 M−1s−1) to DO3A (k1 = 0.45 M−1s−1). However, the rates of acid catalyzed dissociation decreased from 0.112 M−1s−1 observed for the anionic Mn(II) complex of PCTA to 0.0107 M−1s−1 and 0.00458 M−1s−1 for the cationic Mn(II) complexes of PC3AMH and PC3AMPip ligands, respectively. In spite of its lower denticity (as compared to DOTA) the sterically more hindered amide complex ([Mn(PC3AMPip)]2+) displays surprisingly high conditional stability (pMn = 8.86 vs. pMn = 9.74 for [Mn(PCTA)]−) and excellent kinetic inertness. The substitution of phosphonates for the acetate pendant arms (DOTP and DO3P), however, resulted in a noticeable drop in the conditional stability as well as dissociation kinetic parameters of the corresponding Mn(II) complexes ([Mn(DOTP)]6− and [Mn(DO3P)]4−) underlining that the phosphonate pedant should not be considered as a suitable building block for further ligand design while the tertiary amide moiety will likely have some implications in this respect in the future.

effect_of_the_nature_of_donor_atoms_on_the_thermodynamic,_kinetic_and_relaxation_properties_of_mn(ii

Introduction<!><!>Introduction<!>Materials<!>Equilibrium measurements<!>Kinetic measurements<!>Relaxivity measurements<!>Stability of the Mn(II) complexes<!><!>Stability of the Mn(II) complexes<!><!>Stability of the Mn(II) complexes<!><!>Demetallation of the Mn(II) chealtes<!><!>Demetallation of the Mn(II) chealtes<!><!>Demetallation of the Mn(II) chealtes<!><!>Relaxivity of the Mn(II) complexes<!><!>Conclusions<!>Author contributions<!>Conflict of interest statement

<p>In the recent years, the research in the field of Mn(II) coordination chemistry has been intensified aiming to develop Mn(II) complexes suitable for in vivo application as magnetic resonance imaging (MRI) contrast agents (CA) (Drahos et al., 2012; Gale et al., 2015; Forgács et al., 2016; Garda et al., 2016). Gd(III) complexes are used as CAs in millions of doses. These agents had been assumed to be safe, however, nephrogenic systemic fibrosis (NSF), a devastating disease discovered in the late 90s has pointed out that they can cause serious health problems in patients with severe renal impairment (Idee et al., 2006). Thus, the design of safer CAs for MRI might be achieved by replacing the paramagnetic metal center with one that is better tolerated by the living systems (e.g., essential metal ions like Mn(II) or Fe(II)). The Mn(II) complexes are considered to be safe alternatives to Gd(III) in MRI as Mn(II) is an essential metal ion and therefore, biological systems can efficiently control its homeostasis (Murakami et al., 1996; Aime et al., 2002; Balogh et al., 2007; Drahos et al., 2011; Kálmán and Tircsó, 2012; Gale et al., 2015; Garda et al., 2016). However, the lack of ligand-field stabilization in high spin Mn(II) complexes results in lower thermodynamic stability compared to other divalent essential metal and the trivalent Gd(III) complexes. In addition, most Mn(II) chelates are kinetically labile. Due to these factors, the development of safe Mn(II) MRI CAs for in vivo applications remains challenging. We have shown 6 years ago that only trans-CDTA (trans-CDTA = trans-1,2-diaminocyclohexane-N,N,N′,N′-tetraacetic acid) forms thermodynamically stable and kinetically inert Mn(II) chelate with acceptable relaxation enhancement out of several Mn(II) complexes of open-chain and AAZTA (AAZTA = 6-amino-6-methylperhydro-1,4-diazepine tetraacetic acid) ligands (Kálmán and Tircsó, 2012). This observation inspired the development of new open-chain ligands for Mn(II) complexation such as PyC3A, PhDTA, BEDIK (PyC3A = N-picolyl-N,N′,N′-trans-1,2-cyclohexylenediamine-triacetate, PhDTA = ortho-phenylenediamine-N,N,N′,N′-tetraacetic acid, BEDIK = 2-(aminomethyl)aniline-N,N,N′,N′-tetraacetic acid), derivatives of 2,6-bis-aminometyl pyridine and various open-chain ligands incorporating the picolinate moiety (Su et al., 2012; Forgács et al., 2015, 2017; Gale et al., 2015; Phukan et al., 2015; Wu et al., 2015; Póta et al., 2018) as well as redox activated Mn(II)-based MRI CA candidates (Loving et al., 2013). The design and synthesis of new bifunctional chelators (BFCs) derived from trans-1,2-diaminocyclohexane for targeted imaging applications has also been reported (Gale et al., 2015; Vanasschen et al., 2017).</p><p>Macrocyclic ligands have also been screened with the aim of finding a suitable macrocyclic platform for Mn(II) complexation. The investigated macrocyclic ligands were mostly the acetate, rarely the phosphonate and phosphinate derivatives of tacn (tacn = 1,4,7,-triazacyclononane), cyclen and pyclen; efforts were also devoted to the investigation of some rigid pyridine-based 15-membered and other macrocyclic complexes formed with 9-, 12-, 14-, and 15-membered macrocyclic ligands (Dees et al., 2007; Drahos et al., 2010, 2011, 2012; Molnar et al., 2014; Forgács et al., 2016, 2017; Garda et al., 2016). The main goal of these studies was the development of a ligand platform that would allow a solvent molecule to be coordinated in the inner coordination sphere of the Mn(II) complexes, which is necessary to achieve appropriate relaxation enhancement. This was achieved in most cases; actually, some of the chelates, such as Mn(II) complexes of the rigid 15-PyaneN5 (3,6,9,12,18-pentaazabicyclo[12.3.1]octadeca-1(18),14,16-triene) and 15-PyaneN3O2 (3,12,18-Triaaza-6,9-dioxabicyclo[12.3.1]octadeca-1(18),14,16-triene) was found to have not just one, but two bound water molecules. However, most of these complexes were kinetically too labile for in vivo applications. The only structures that represented an acceptable compromise between the apparently contradictory requirements (thermodynamic and redox stability/inertness/relaxivity) were the Mn(II) chelates of the cis-DO2A and the 15-PyaneN3O2 ligands (Dees et al., 2007; Drahos et al., 2010; Garda et al., 2016). The kinetic properties improved by the replacement of two carboxylate moieties with dimethylamide metal binding units (Forgács et al., 2016). The results obtained in our laboratory by studying monopicolinates of 9-, 12- and 14-membered macrocyclic ligands have also indicated that the best kinetic data were obtained for the Mn(II) complex of a 12-membered macrocyclic derivative, but further ligand optimization is required to identify the best candidates for in vivo applications (Molnar et al., 2014).</p><p>Since we have access to several 12-membered macrocyclic heptadentate ligands (DO3A, ODO3A, PCTA, DO3P, DO3AM, PC3AMH, PC3AMGly PC3AMPip) from our previous studies performed with their Ln(III) complexes, we decided to study the stability and dissociation kinetic properties of the Mn(II) complexes formed with these chelators. We expected to gain more information on how the rigidity (DO3A vs. PCTA, or DO3AMH vs. PC3AMH), the nature of the donor atoms in the macrocycle (DO3A vs. ODO3A) and the nature of the pendant arms (DO3A vs. DO3P, i.e., replacement of acetate pendants by phosphonates), DO3A vs. DO3AM (replacement of acetate pendants by amides), or PCTA vs. PCTAM (replacement of acetate pendants by amides for rigidified macrocycle) and the nature of amide pendants (PC3AMH vs. PC3AMGly or PC3AMPip (replacement of primary amides by secondary and tertiary amides, respectively) affect the physicochemical properties of Mn(II) complexes. The goal of this project was to better understand how these (functional) modifications in the ligand structure affect equilibrium and kinetic behaviors of the corresponding Mn(II) complexes. Such data can help us to design ligands for improved Mn(II) complexation. It should be kept in mind, however, that some of these Mn(II) complexes would not be very efficient relaxation agents because of the lack of inner sphere water molecule. The formulae of the studied ligands are shown in Figure 1.</p><!><p>Structures of hexa-, hepta- and octadente ligands studied and compared in this work.</p><!><p>The stability constants of the Mn(II) complexes were determined by pH-potentiometry and/or 1H relaxometry (measuring T1 relaxation times and plotting 1/T1 values normalized to 1 mM paramagnetic substance as a function of pH). The kinetic inertness of the Mn(II) complexes have been evaluated by studying the acid catalyzed dissociation or metal/ligand exchange reactions occurring with Cu2+ (for [Mn(DOTAM)]2+, [Mn(DO3AMH)]2+, [Mn(DO3A)]−, [Mn(PCTA)]− and [Mn(ODO3A)]− complexes), or transCDTA (as in the case of [Mn(DO3A)]− chelate), or DTPA (as in the case of [Mn(DOTP)]6− and [Mn(DO3P)]4− chelates). Based on these data the half-lives were calculated at pH = 1 and 7.4 and compared to those of DO2A and DOTA derivatives published in the literature. The relaxivity values of some complexes were also determined in order to confirm that the relaxivity of the Mn(II) complexes formed with 12-membered heptadentate macrocyclic chelators is purely of outer-sphere in origin since the Mn(II) complexes of macrocyclic ligands with seven or more donor atoms are not expected to have an inner-sphere water molecule (Rocklage et al., 1989).</p><!><p>The chemicals used in the experiments were of the highest analytical grade. The concentration of the MnCl2, CuCl2 and ZnCl2 solutions was determined by using standardized Na2H2EDTA solution and eriochrome black T (Mn(II)), murexide (Cu(II) and xylenolorange (Zn(II)) as indicator. The ligands were either prepared by following literature procedures (DOTAM, DO3A, DO3P, PCTA, ODO3A, PC3AMGly) (Amorim et al., 1988; Maumela et al., 1995; Aime et al., 1997; Siaugue et al., 2000; Sun et al., 2003; Rojas-Quijano et al., 2009; Nithyakumar and Alexander, 2016) or obtained from commercial sources (DO3AMH - CheMatech, Dijon (France)). The PC3AMH and PC3AMPip ligands were prepared by alkylating the pyclen macrocycle prepared according to literature description (Stetter et al., 1981) with a suitable bromoacetamide derivative (see the Supplementary Material for the detailed synthesis and analytical data) available from commercial sources or prepared by following literature synthesis (Kaupang and Bonge-Hansen, 2013).</p><!><p>The concentration of the ligand solutions was determined by pH-potentiometric titration in the presence and absence of an excess (5–10 fold) of MnCl2. For determining the protonation constants of the investigated ligands, pH-potentiometric titrations were made by means of 0.2 M standardized NaOH in 2 and 3 mM ligand solutions in the pH-range of 1.75–11.85. All equilibrium measurements were performed at a constant 0.15 M NaCl ionic strength at 25 °C. The protonation and the stability constants of the Mn(II) complexes were determined by pH-potentiometric titrations (DOTP, DO3A, PCTA, ODO3A, DO3P) while owing to the slow formation rates of Mn(II) complexes formed with amide type ligands, the stability constants of DOTAM, DOTMA, DO3AMH, PC3AMH, PC3AMGly, PC3AMPip complexes were determined by out-of-cell pH-potentiomertry in combination with 1H relaxometry. The metal-to-ligand concentration ratio in the solutions was 1:1. For the calculation of protonation constants of the ligands and the log KMnL and log KMnLHi values of the complexes, mL–pH data pairs (50–180), obtained in the pH range of 1.75–11.85 or r1obs-pH data pairs (10–15 batch samples) obtained in the pH range of 2–4.5 were used (the samples were equilibrated for 4–7 days). The equilibrium constants characterizing the deprotonation that occurs at basic pH were determined from the data obtained via direct pH-potentiometric titrations performed on the pre-formed complexes.</p><p>The pH-potentiometric titrations were carried out with a Methrohm 888 Titrando titration workstation using a Metrohm-6.0233.100 combined electrode. The samples (6.00 mL) were thermostated at 25 °C. The samples were stirred and kept under inert gas (N2) to avoid the effect of CO2. KH-phthalate (pH = 4.005) and borax (pH = 9.177) buffers were used for the pH-calibration. The method proposed by Irving et al. was used for the calculation of H+ concentrations from the measured pH values (Irving et al., 1967). A 0.01 M HCl (I = 0.15 M set with NaCl) solution was titrated with standardized NaOH solution of known concentration (approx. 0.2 M). The correction factor obtained as difference of calculated pH and measured pH values was used to calculate the [H+] in the samples. The ionic product of water was determined from the same titrations (HCl/NaOH) from the data collected in the pH range of 11.20–11.85.</p><p>The relaxometric data were collected using a Bruker Minispec MQ20 NMR relaxometer. The batch samples were equilibrated for 4–7 days (until no further change in the relaxivity was observed for the samples prepared in duplicates) and the T1 values were measured multiple times. A datapoint plotted on figures represents an average of 5–6 T1 measurements (Figures S3–S9 in Supplementary Material). The data were fitted by using the molar r1p relaxivity value of the MnCl2 (7.92 M−1s−1 at 0.49 T and 25 °C). The equilibrium constants were calculated using the PSEQUAD program (Zékány and Nagypál, 1985).</p><!><p>The dissociation reactions were followed by UV-vis spectrophotometry at 269 nm for the pyclen derivatives or at 300 nm for other complexes in the pH range of 3.6–5.0 for the exchange reactions with Cu(II) and up to pH = 6.0 when Zn(II) was used as a ligand scavenger. The acid catalyzed dissociation reactions were performed in the HCl concentration range of 0.05–1.0 M. The concentration of the complexes in the samples was 0.25 mM, while the concentration of the exchanging metal ion in the metal initiated transchelation reactions was 10–40 times higher to ensure pseudo-first-order conditions. Relaxometry was used to follow the transchelation reactions of [Mn(DOTAM)]2+, [Mn(DO3AMH)]2+, [Mn(DO3A)]−, [Mn(PCTA)]− and [Mn(ODO3A)]− with Cu(II)/Zn(II) ions in the pH range of 3.6–5.5 (Figures S17–S19 in Supplementary Material). The concentration of the complexes in these samples was set to 1.0–2.0 mM to achieve relatively large change during the dissociation reaction. The relaxivity of the complexes is in the range of 1.00–1.44 mM−1s−1, which increases to 7.92 mM−1s−1 upon Zn(II) mediated decomplexation. The temperature was maintained at 25 °C and the ionic strength of the solutions was kept constant at 0.15 M with NaCl. N,N′-dimethyl- (DMP) and N-methyl-piperazine (NMP) buffers (log K2H = 4.18 and 4.90 at 25 °C and I = 0.15 M NaCl, respectively) were used at 0.05 M concentration to keep the pH constant. The pseudo-first-order rate constants (kobs) were calculated by fitting the absorbance or relaxivity-time data pairs to Equation (1) (1)Xt=(X0-Xe)e-kobst+Xe where Xt, X0, and Xe are the absorbance or relaxivity at time t, at the start and at equilibrium, respectively. The calculations were performed with the computer program Micromath Scientist, version 2.0 (Salt Lake City, UT, USA) using a standard least-squares procedure.</p><!><p>The longitudinal water proton relaxation rates (r1 = 1/T1-1/Tw) were measured at 20 MHz with a Bruker Minispec MQ-20 relaxometer (Bruker Biospin, Germany). Samples were thermostated by using a circulating water bath at 25.0 ± 0.2 °C. The longitudinal relaxation times (T1) were measured by the inversion-recovery method (180° − τ − 90°), averaging 5–6 data points collected for each concentration point obtained from 14 different τ values (τ values ranging between 0 to at least 6 times the expected T1). The relaxivity of the complexes was determined by titrating a nearly 1.0 mM ligand solution with a Mn(II) stock solution at pH = 7.22–7.43 (50 mM HEPES buffer, I = 0.15 M NaCl, 25 °C). Under these conditions, the only Mn(II) ion containing species present in solution is the [Mn(L)], which was supported by the r1 vs. pH profiles. The relaxivity of the complex was determined as the slope of the strait line obtained by plotting 1/T1p values as a function of Mn(II) concentration for samples with [Ligand] > [Mn(II)]. The relaxivity of [Mn(DOTP)]6− and [Mn(DO3P)]4− was determined in the pH range of 10.0–11.0 and 8.5–9.3, respectively where mainly the [Mn(DO3P)]4− and [Mn(DOTP)]6− complex exists in solution (Figures S4 and S9 in Supplementary Material).</p><!><p>The protonation and complexation equilibria of the macrocyclic ligands and their Mn(II) complexes have been studied by pH-potentiometric and relaxometric methods in the presence of 0.15 M NaCl whose ionic strength mimics the physiological conditions. The protonation constants of the ligands were evaluated by fitting the pH-V(mL) base data pairs collected in the pH-potentiometric titrations. The protonation equilibria of the ligands can be described by Equation (2). The log KH values are listed in Table 1. (a comparison of previously published log KH data available in the Supplementary Material) together with the constants of H4DOTA, cis- and trans-H2DO2A derivatives given for comparison. (2)KiH=[HiL][Hi-1L][H+],i=1,2,…</p><p>As it has been stated before in several publications (Desreux et al., 1981), the first two protonations of these tetraazacyclododecane derivatives occur at the opposite ring nitrogen donor atoms followed by further protonation steps depending on the nature of donor groups incorporated into the pendant arms. The protonation sequence of PCTA and ODO3A and their derivatives differ slightly from that of DOTA and its derivatives owing to the asymmetric nature of the macrocyclic ligands (Aime et al., 1997). The first protonation in these ligands occurs at the nitrogen atom situated trans to the pyridine N or etheric O atom. The protonation of a cis nitrogen atom in the second step forces the first proton to shift to the other cis nitrogen of the macrocycle leaving the trans N atom unprotonated (Aime et al., 1997). The substitution of amides for the carboxylate groups (DO3AMH, PC3AMH, etc.) results in a significant decrease, while the introduction of phosphonate groups (e.g., DO3P) causes a notable increase in the total basicity of the ligands. Not only the value but also the number of the protonation constants is affected by these structural alterations. As it is seen from the data shown in Table 1, the sum of the first two protonation constants (Σ log β2H = log K1H+log K2H) is significantly higher for the acetate derivatives than that for the amide derivative ligands because the number of the negatively charged sidearms decreases, which results in a weaker hydrogen bonding existing between the protonated donor atoms of the macrocycle and the amide group. On the other hand, the basicity of the macrocyclic nitrogen atoms in DO3P is higher than that in DO3A resulting in a noticeably higher ligand basicity. The ligand ODO3A, derived from DO3A by substituting an etheric O-atom for the macrocyclic NH, also displays a drop in the Σ log β2H value due to a change in the protonation sequence (Amorim et al., 1988).</p><!><p>Protonation constants of the hexa-, hepta- and octadentate ligands (I = 0.15 M NaCl, 25 °C).</p><p>1H- and 31P-NMR;</p><p>I = 0.1 M KCl (Aime et al., 2011);</p><p>0.1 M KCl (Takács et al., 2014);</p><p>Garda et al. (2016);</p><p>I = 0.1 M KCl (Forgács et al., 2016);</p><p>I = 0.1 M KCl (Forgács et al., 2017).</p><!><p>In order to gain information on the thermodynamic stability of the Mn(II) complexes, samples containing the metal and ligand in 1 to 1 molar ratio were studied by pH potentiometry or 1H-relaxometry and in some cases by a combination of the two methods. The titration data were fitted by assuming the existence of [M(L)] and some protonated species ([M(HiL)]) at lower pH values. In some cases to fit the data obtained at higher pH, the formation of ternary hydroxido complexes ([M(L)(OH)]) were also included in the equilibrium model. The stability constants (KML) and the various protonation forms (KMHiL) of ML metal chelates are defined by Equations (3)–(5). However, a comparison of the stabilities of different complexes solely on the basic of their stability constants might be misleading because of the differences in ligand basicities. Therefore, we have calculated the pMn values at pH 7.4 defined as pMn = –log [Mn2+]free with cMn = cL = 1 × 10−5 M as suggested by Tóth and co-workers (Drahos et al., 2010). The pMn value accounts for the effect of proton competition (ligand basicity and complex protonation) on the stability constant essentially offering the same information as the conditional stability constant. Higher pMn values indicate a higher chelate stability at specified conditions. Considering the above mentioned conditions (cMn = cL = 1 × 10−5 M, pH = 7.4), the minimum value of the pMn is 5, which corresponds to 0% complexation. (3)KML=[M(L)][M][L] (4)KMHiL=[M(HiL)][M(Hi-1L)][H+],  where i=1,2,… (5)KMLOH=[M(L)][M(L)(OH)][H+]</p><p>The ligand DOTA is generally considered as the "gold standard" for metal based systems. The [Mn(DOTA)]2− complex has a pMn value of 9.02 (cMn = cL = 1 × 10−5 M at pH = 7.4) corresponding to log KMnL of 19.44. As seen from the data shown in Table 2, the stability constants obtained for the complexes with phosphonate pendant arms ([Mn(DOTP)]6− and [Mn(DO3P)]4−) are close to that of [Mn(DOTA)]2−; they are the highest among the complexes we studied. The pMn values however, indicate that the conditional stability of these complexes is the lowest among the studied systems. By analyzing the data shown in Table 2 one can conclude that among the studied heptadentate ligands the Mn(II) complex formed with the rigid pyridine macrocyclic PCTA ligand has the highest conditional stability near to physiological conditions. It is higher than that of the [Mn(DO3A)]− complex, although the number and nature of the donor atoms coordinating the Mn(II) ion are nearly the same (4N and 3O) in these ligands. The pMn value characterizing [Mn(PCTA)]− is higher than that calculated for [Mn(DOTA)]2− by 0.7 pM unit while the pMn values of the complexes formed with DO3A, ODO3A and amide derivatives of the rigidified pyclen macrocycles PC3AMGly and PC3AMPip ligands are slightly lower than that of [Mn(DOTA)]2− (with a perceptible increase from primary to secondary and to tertiary amide sidearms). These differences can be explained in terms of ligand protonation constants, as it is stated above. The first two protonation constants (log K1H and log K2H) of DOTA and its derivatives (except for the DOTAM and DO3AMH chelators) are greater than pH 7.4 meaning that the ligands are diprotonated above this pH value, which used in pMn calculations. In contrast, for PCTA and its derivatives only the log K1H value is higher than 7.4. In case of ODO3A the log K2H is slightly higher than 7.4 (Table 1) whereas for the DOTA and its derivatives the second protonation occurs above pH = 9.0 (25 °C and I = 0.15 M NaCl). Consequently, these ligands exist in their diprotonated form near pH = 7.4 and so the proton competition is more significant in these systems.</p><!><p>Stability and protonation constants as well as pMn values for the Mn(II) complexes (I = 0.15 M NaCl, 25 °C).</p><p>1H-relaxometry;</p><p>I = 0.1 M KCl (Takács et al., 2014);</p><p>Tircsó and Woods (in preparation);</p><p>the 4th protonation constant of the complex is (log KMH3LH = 5.02 (1);]</p><p>Garda et al. (2016);</p><p>I = 0.1 M KCl (Forgács et al., 2016);</p><p>I = 0.1 M KCl (Forgács et al., 2017);</p><p>The pMn values were calculated at pH = 7.4, cM = cL = 10−5 M by using the conditions as suggested by Drahos et al. (2010).</p><!><p>Linear relationships between experimentally measured log KM(L) values and ligand basicities were reported for more than 60 years ago (Martell and Calvin, 1952). Later Choppin proposed the inclusion of linear polyamino polycarboxylate systems and demonstrated a single linear correlation between log KM(L) values and Σ log KiH for polydentate ligands that form five-membered chelate rings with Ln(III) cations (Choppin, 1985). More recently, such empirical relationships have been found for macrocyclic polyamino polycarboxylates as well (Kumar et al., 1995; Huskens et al., 1997). As the diversity of ligands has grown over the years, it has become apparent that a major uncertainty in establishing such relationships is the number of ligand protonation steps that should be included in the calculation of Σ log KiH (the basicity of amides, phosphinates, carboxylates and phosphonates differ considerably i.e., these correlations are expected to exist for structurally similar ligands). A similar empirical relationship between the stability (log K[Cu(L)]) and the sum of the first two log KiH values corresponding to the protonation of N-atoms has been shown to hold for Cu(II) complexes of several linear and macrocyclic ligands (Lukes et al., 2001). Even though this relationship ignores the basicity of the side-chain coordinating groups, an analogous approach for Gd(III) complexes with macrocyclic ligands (including 1,4,7-triazacyclonanane, 1,4,7,10-tetraazacyclododecane and 1,4,8,11-tetraazacyclotetradecane with carboxamide, carboxylate and methylenephosphonate pendant arms) gives an acceptable linear correlation with R2 = 0.86 (log KGd(L) = 1.5 (log K1H + log K2H) − 9.7) (Brücher et al., 2013). We applied a similar approach to the Mn(II) complexes studied in the current work (Figure 2) This resulted in an unacceptable correlation (R2 = 0.63) (Figure 2, red points: log KMnL = 0.50 (log K1H + log K2H) + 6.22). However, omission of the data corresponding to PCTA (positive deviation) and DO3AMH (negative deviation) considerably improved the correlation (log KMnL = 0.47 (log K1H + log K2H) + 6.68 with R2 = 0.89) while the numbers describing the correlation did not change significantly. Figure 2 shows how the weakness of one structural motif can be compensated in part by the strength of another when the effects of various structural features are combined. [Mn(PCTA)]− has higher, whereas [Mn(DO3AMH)]2+ has lower stability than expected based on the ligand basicity. Likewise, the Mn(II) complex of PCTA-tris(amide), (PC3AMH) is less stable than expected while [Mn(PC3AMGly)]− and [Mn(PC3AMPip)]2+ have stabilities higher than expected. Our data also indicate that Mn(II) complexes of all the bisamides of cis-DO2A reported in the literature have lower stability than expected on the basis of ligand basicity, highlighting the importance of further ligand design.</p><!><p>Plot of stability constants (I = 0.15 M NaCl and 25 °C) of Mn(II) complexes of 12-membered macrocyclic ligands vs. basicity of the macrocyclic nitrogen atoms (log K1H + log K2H) of the ligands (green: disubstituted, red: trisubstituted and blue: tetrasubstituted derivatives).</p><!><p>The kinetic inertness of the Mn(II) complexes is the other important parameter for the complexes considered for use in vivo. The inertness of Mn(II) complexes formed with trisubstituted 12-memberd macrocyclic ligands were characterized either by studying the rates of acid catalyzed dissociation or by studying transmetallation reactions occurring between the complexes and a suitable exchanging metal ion such as Zn(II) or Cu(II). The demetallation reactions of labile complexes were performed at high pH by following the ligand exchange reactions with trans-CDTA and DTPA ligands. The demetallation reactions were studied in the presence of a large Cu(II)/Zn(II)/trans-CDTA/DTPA excess or by using high acid concentration in order to ensure pseudo-first-order conditions (Equation 6). Scheme 1 outlines all the possible dissociation pathways for the Mn(II) complexes. Scheme 1 shows that the dissociation may occur through exchange reactions with other M' metal ion or L' ligand by the assistance of protons. The reactions of the released Mn(II) and L ligand with the exchanging L' and M' reactants are assumed to be very rapid.</p><!><p>Assumed reaction pathways for the dissociation of Mn(II) complexes in the presence of metal ions and ligands available for exchange reactions (M = Cu(II) or Zn(II), L' = trans-CDTA or DTPA, the charges are omitted for clarity).</p><!><p>The rate constants k0, kH, kHH, kM, kMH, kL, and kLH characterize the rate of the spontaneous, proton-assisted, metal-assisted and proton-metal-assisted (when the exchanging metal attacks the protonated or the proton attacks the dinuclear complexes) ligand and proton-ligand reaction pathways, respectively. K1H, K2H, KMH, and KLH are the protonation constants of the [Mn(L)], [Mn(HL)], [Mn(L)M], [Mn(L)(L')], and the stability constant of the dinuclear (KM) [Mn(L)M] and ternary (KL) [Mn(L)(L')] complexes, respectively.</p><p>The rate of dissociation can be expressed by the Equation (6), where kobs is the pseudo-first-order rate constant and [Mn(L)]t is the total concentration of the species containing [Mn(L)] complex. The [Mn(L)]t will differ slightly depending on the experimental conditions (in the presence of a large metal or ligand excess). As [Mn(L)]t = [Mn(L)] + [Mn(HL)] + [Mn(H2L)] + [Mn(L)M] + [Mn(L)(L')] + [Mn(HL)(L')] for reactions performed at high pH (in the pH range of 3.0–5.0 often applied in Cu(II) exchange reactions) will involve fewer protonated species while the acid catalyzed dissociation reactions run in the strongly acidic media (0.1–1.0 M acid concentration range) predominantly contain protonated species ([Mn(L)]t = [Mn(L)]+[Mn(HL)]+[Mn(H2L)]). (6)-d[MnL]tdt=kobs[MnL]t</p><p>By taking into account the various complex species and the possible reaction pathways as well as the protonation and stability constants of the reactive intermediates, the following equation (Equation 7) can be derived to describe the rates of dissociation. (7)kobs=k0+k1[H+]+k2[H+]2+k3[M2+]+k4[M2+][H+]+k5[L′]+k6[L′][H+]1+K1H[H+]+K1HK2H[H+]2+KM[M2+]+KMH[M2+][H+]+KL′[L′]+KL′H[L′][H+] where k1 = kH·K1H, k2 = kHH·K1H·K2H, k3 = kM·KM, k4 = kMH·KMH, k5 = kL·KL′ and k6 = kLH·KL′H. By fitting the experimental kobs data to Equation (7) resulted in the rate constants characterizing spontaneous (k0, s−1), acid catalyzed (k1, M−1s−1 and k2, M−2s−1) and metal-assisted (k3, M−1s−1) dissociation for the Mn(II) complexes that are listed in Table 3. For the calculation we had to have some information about the importance of different reaction pathways. The role of the metal-proton-assisted (k4, M−2s−1), ligand (k5, M−1s−1) and ligand-proton-assisted (k6, M−2s−1) pathways in these demetallation reactions were negligible and in most cases the protonation and stability constants of the intermediates (K1H, K2H, KM, and KL′H) could not be determined because of their very low values.</p><p>As an example, Figure 3 shows a representative plot of kobs values as a function of acid concentration obtained for [Mn(PCTA)]− during the Zn(II) induced transmetallation reactions in the pH range of 3.09–5.88 (some more data are included in the ESI). The rates of acid catalyzed dissociation reactions obtained for PCTA and its amide derivatives are shown in Figure 4.</p><!><p>Rate constants of spontaneous (k0), proton-assisted (k1 and k2), metal-ion-assisted (k3) pathways and half-lives characterizing the dissociation of Mn(II) complexes (I = 0.15 M NaCl, 25 °C).</p><p>relaxometry, the stability constant of the dinuclear intermediate (KM) was found to be 38 ± 6 M−1;</p><p>stopped-flow in highly acidic condition where the k0 is the spontaneous and k1 is the proton-assisted dissociation of the [Mn(HL)] complex;</p><p>the k0 is the spontaneous and k1 is the proton-assisted dissociation of the [Mn(H3L)] complex;</p><p>the K1H was fixed to 600 based determined by pH-potentiometry;</p><p>ligand exchange reaction with trans-CDTA4− or DTPA5− in the pH range 8-9.5, the stability constant of the protonated intermediate (K1H) was found to be (2.0 ± 0.4) × 108, the KH was fixed to 108.06;</p><p>Tircsó and Woods (in preparation) in 1.0 M KCl, where the k0 is the spontaneous and k1 is the proton-assisted dissociation of the [Mn(H2L)] complex;</p><p>the stability constant of the [Mn(DOTA)Zn] dinuclear intermediate (KM) was found to be 68 M−1 (Drahos et al., 2011);</p><p>Garda et al. (2016);</p><p>I = 0.1 M KCl (Forgács et al., 2016);</p><p>I = 0.1 M KCl (Forgács et al., 2017);</p><p>the half-life (t1/2) corresponding to 0.1 M acid concentration were not calculated since the dissociation reactions were carried out at high pH (where [Mn(DO3P)]4− and [Mn(DOTP)]6− exist in multiple protonated forms up to pH = 9.0 and their dissociation is extremely fast), i.e., the rate constants were determined only for the fully deprotonated species;</p><p>the half-life (t1/2) was not calculated since the dissociation reactions were carried out under acidic conditions where a protonated complex exists.</p><p>Plot of the pseudo-first-order rate constants (kobs) as a function of Zn(II) and H+ ion concentration for [Mn(PCTA)]− (blue: 10-fold Zn(II) excess, green: 20-fold Zn(II) excess, red: 30-fold Zn(II) excess and black: 40-fold Zn(II) excess).</p><p>Plot of the pseudo-first-order rate constants (kobs) as a function of H+ ion concentration for [Mn(PCTA)]− (black), [Mn(PC3AMH)]2+ (green), [Mn(PC3AMGly)]− (blue) and [Mn(PC3AMPip)]2+ (red).</p><!><p>The kobs values increase with increasing H+ ion concentration in all cases (k1, k2) and either increase (k3) or remain unaffected by increasing the exchanging metal ion concentration. It is difficult to directly compare the rate constants that characterize the different reaction pathways. Therefore, the half-lives (t1/2) of the dissociation reactions of Mn(II) complexes were calculated under physiological conditions (pH = 7.4, 10 μM concentration of the exchanging metal ion) and also for 0.1 M H+ concentration since the most inert Mn(II) complexes were studied in acidic milieu (Table 3).</p><p>The comparison of the k1 (rate constant characterizing the acid catalyzed dissociation) or half-live (t1/2) values shows that the dissociation kinetics of the Mn(II) complexes of 12-membered macrocyclic ligands strongly depends on the rigidity of the macrocycles, the nature of donor atoms present in the macrocycle, as well as on the metal binding moieties attached to the nitrogen atoms of the macrocycles. For a class of very similar ligands (i.e., 12-membered macrocycles), the rate of acid catalyzed dissociation (k1) was found to differ as much as eight orders of magnitude (2.4 × 105 M−1s−1 for DO3P vs. 4.64 × 10−3 M−1s−1 for PC3AMPip), which underlines the importance of both the macrocyclic backbone and the metal binding sidearms in ligand design. As seen from the data shown in Table 3, the replacement of the secondary amine (>NH in DO3A) in the macrocycle by etheric oxygen atom (-O- as in ODO3A) resulted in a 50 fold increase in the k1 value of the Mn(II) complexes, which indicates that the given structural modification should not be considered for Mn(II) chelators. On the other hand, the incorporation of the secondary amine moiety (>NH) into a pyridine ring has a beneficial effect on the kinetic parameters as evidenced by the 5-fold decrease of the k1 value of its Mn(II) complex. When analyzing the effect of the nature of pendant arms on the dissociation kinetic properties, it can be concluded that the inertness of Mn(II) complexes of amide derivatives is higher than that of Mn(II) complexes formed with ligands having acetate or especially, phosphonate sidearms. This behavior is clearly related to the decreased basicity of the amide based ligands as well as the hindered proton transfer from the protonation site to the ring nitrogen donor atom(s), which is essential for the acid-assisted path. It was previously proved that the substitution of a phosphonate for the carboxylate group can facilitate the proton transfer to the macrocyclic nitrogen since the phosphonate groups (in acidic solutions) can be protonated even when they are metal-bound (Kálmán et al., 2008). The replacement of the acetates by primary amides resulted in noticeable increase in the kinetic inertness going from PCTA to its trisamide derivatives (PC3AMH). Further improvement was observed going from the primary (PC3AMH) toward the tertiary amide (PC3AMPip) complexes (the k1 values decrease from 0.109 M−1s−1 observed for the anionic PCTA complex to 0.0107 M−1s−1 and 0.00464 M−1s−1 for the cationic Mn(II) complexes of PC3AMH and PC3AMPip, respectively). This tendency is analogous to the kinetic behavior of Ln(III) complexes of DOTA-tetraamides (Aime et al., 1999; Pasha et al., 2007). A break in this trend is apparent for the [Mn(PC3AMGly)]− complex, which is due to the presence of charged (and protonatable) groups in this ligand. The sterically more hindered amide [Mn(PC3AMPip)]2+ complex displays surprisingly high kinetic inertness as evidenced by its k1 rate constant, which is significantly lower than that of [Mn(DOTA)]2−) in spite of its lower denticity.</p><p>We have attempted to correlate the dissociation kinetic data and the stability or pMn values of the Mn(II) complexes by plotting the –log k1 values against the basicity of the nitrogen atoms of the ligands (log K1H, log K2H, and log K1H + log K2H). Plots of the –log k1 values against the first (log K1H) and second (log K2H) protonation constants are included in the supporting information (Figures S11–S14). Although no direct correlation is expected to exist between thermodynamic and kinetic parameters, we could observe a linear correlation when the –log k1 values were plotted as a function of pMn values (Figure 5). However, the acetate (blue line) and amide (green line) systems had to be considered separately, although it should be emphasized that we had very limited data to work with. The basicity of amide ligands weaker than that of the acetate derivatives, thus the Mn(II) complexes of the amide derivatives tend to dissociate more slowly via the proton assisted dissociation path. This is reflected by the fact that the correlation curve for the amide complexes runs above the line of the acetate derivatives and it has slightly larger slope. This observation strongly implies if one could design amide ligands that would form complexes with higher thermodynamic (and conditional) stability, then the kinetic inertness of such complexes is also expected to increase.</p><!><p>Plot of –log k1 values (k1 is the rate constant of the acid-assisted dissociation) as a function of pMn for acetate (light blue) and amide (light green) derivatives of 12-membered macrocycles (I = 0.15 M NaCl and 25 °C). Disubstituted ligands are shown in green, trisubstituted derivatives are in red and tetrasubstituted derivatives are in blue.</p><!><p>The relaxivity (r1p, mM−1s−1) is defined as the longitudinal water proton relaxation rate in a solution containing 1.0 mM concentration of the paramagnetic species. It characterizes the efficiency of a paramagnetic metal chelate to enhance the solvent proton relaxation rate at a given magnetic field (usually 20 MHz) and temperature. The inner-sphere contribution to relaxivity is directly proportional to the number of metal bound water molecules (q), and therefore, relaxivity values may provide some useful information on the hydration state of paramagnetic complexes of similar molecular weights and electronic relaxation times. As published in the literature, the Mn(II) complex formed with the heptadentate ligand DO3A does not possess a metal bound water molecule (q = 0), and therefore, it has low relaxivity (1.30 mM−1s−1) originating solely from second and outer sphere contributions (Rolla et al., 2013). The r1p values of the Mn(II) complexes formed with heptadentate 12-membered macrocyclic ligands were determined at 0.5 T field strength and 25 °C and reported in Table 4 along with some relaxivity data published for the complexes of hexa-, hepta- and octadentate ligands. The relaxivity of the [Mn(PCTA)]− and [Mn(ODO3A)]− complexes are indeed very similar to that of the [Mn(DO3A)]− suggesting the absence of metal bound water molecules in these complexes as well. In fact, the X-ray crystal structure of the Mn(II) chelate formed with tris(ethyl ester) of PCTA confirms the absence of the metal bound solvent molecule in the [Mn(PCTA-OEt3)]2+ complex (Wen et al., 2012) suggesting that this is also true for the PCTA complex. The Mn(II) complexes of amide derivatives show even lower relaxivity, which can be explained by weaker second and outer sphere contributions (for example, X-ray crystallography has shown that [Mn(DO3AMH)]2+ does not have an inner sphere water molecule) (Wang and Westmoreland, 2009) whereas increased second and outer sphere effects are likely responsible for the slight increase in the relaxivities observed for [Mn(DOTP)]6− and [Mn(DO3P)]4− in the pH ranges of 10.0–11.0 and 8.5–9.3, respectively, where the deprotonated complex exists in solution. Below these pH ranges the phosphonate moieties undergo protonation, which further increases the relaxivity (2.60–2.90 mM−1s−1 values were observed near pH = 7.4, see Supplementary Material). Similar increase in the relaxivity (to 3.20 mM−1s−1 at 20 MHz, 37 °C) is observed for the [Mn(DOTA)]2− complex upon its protonation and formation of the thermodynamically stable diprotonated [Mn(H2DOTA)] complex that exists as a non-hydrated chelate both in solution and in the solid state (Wang and Westmoreland, 2009). The relaxivity of the diprotonated species at t → 0 time point was determined from the extrapolation of the relaxivity vs. time data collected after acidification of the samples containing the complexes. The X-ray crystallographic structure of [Mn(H2DOTA)] indicates the protonation of two trans carboxylate moieties (e.g., the coordination around the metal ion is similar to that observed in [Mn(trans-DO2A)] as determined by DFT calculations). Thus, the relaxivity of the diprotonated [Mn(H2DOTA)] and [Mn(H2DOTMA)] complexes is expected to be similar to that of [Mn(trans-DO2A)] (when q = 0). The value obtained by us is considerably higher, which highlights the importance of prototropic exchange as a useful tool to improve the relaxivity of the complexes (however, it has to be underlined that the protonation of the complex generally leads to a decrease in kinetic inertness). A similar relaxivity enhancing role of prototropic exchange was demonstrated for Gd(III) complexes of phosphonate derivatives such as ([Gd(DOTP)]5− and [Gd(DOTA-4Amp)]5−) (Avecilla et al., 2003; Kálmán et al., 2007).</p><!><p>Relaxivities (r1p) determined for selected Mn(II) complexes (at 0.5 T field strength, 25 °C and pH = 7.4).</p><p>Rolla et al. (2013);</p><p>determined in the pH range of 8.5–9.3 where mainly the [Mn(DO3P)]4− complex exists in solution;</p><p>AAZTA = 6-amino-6-methylperhydro-1,4-diazepine tetraacetic acid;</p><p>Tei et al. (2011);</p><p>determined in the pH range of 10.0–11.0 where the deprotoanted complex exists in solution;</p><p>Wang and Westmoreland (2009) measured at 0.5 T field strength and 37 °C;</p><p>Tircsó and Woods (in preparation);</p><p>Forgács et al. (2016);</p><p>Forgács et al. (2017).</p><!><p>In this work, we summarized the results of our studies on the equilibrium and kinetic properties of Mn(II) complexes of several trisubstituted 12-membered macrocyclic ligands tested in our laboratory during the past 5–6 years. Originally, these ligands were intended for Ln(III) complexation. In this project, these structurally extremely diverse ligands were investigated for their Mn(II) chelating ability and we believe that the results presented here will help us to better understand the relationship that exists between the ligand structures and the thermodynamic/kinetic/relaxometric properties of their Mn(II) complexes. The stability constants of the complexes were determined by pH-potentiometry and often supported by 1H-relaxometry. As expected, decreasing the denticity and basicity of the parent ligand DOTA (i.e., removing one metal binding pendant arm to afford DO3A) resulted in a drop in the stability of the Mn(II) complex. This decrease can be compensated by incorporating the fourth (secondary) nitrogen atom into a pyridine ring (e.g., PCTA) or replacing the secondary amine with an etheric oxygen atom. The substitution of primary amides for the acetates also resulted in a drop in the stability constant (DO3AMH or PC3AMH), but the stability increased as the primary amides (PC3AMH) were replaced by secondary (PC3AMGly) or tertiary amide (PC3AMPip) moieties. The ligands incorporating phosphonate pendant arms were found to form the most stable Mn(II) complexes but their conditional stability at pH = 7.4 was the lowest. Very similar conclusion was derived by analyzing the dissociation kinetics data. Among the studied complexes, [Mn(DO3P)]4− was found to be the most labile while the Mn(II) complex of the rigid macrocycle (pyclen) based tertiary amide chelator PC3AMPip has the highest kinetic inertness whose acid catalyzed dissociation rate is nearly 8 orders of magnitude lower than that of [Mn(DO3P)]4−. By considering the structure of PC3AMPip it can be concluded that the rigidity of the macrocycle is one of the key structural features that determine the dissociation kinetic properties of Mn(II) complexes. Further improvements can be achieved by proper selection of the donor atoms in the metal binding sidearms attached to the nitrogen atoms of the macrocycles. The tertiary amide coordinating sidearm appears to be one of the best candidates in this respect. The Mn(II) complexes of the studied heptadentate ligands are poor T1 relaxation agents because they do not have a metal bound water molecule. However, the trends observed for Gd(III) complexes seems to hold for the Mn(II) chelates as evidenced by the low relaxivity of the amide based systems or by the increase in the relaxivity due to accelerated prototropic exchange in [Mn(DOTP)]6− and [Mn(DO3P)]4−.</p><!><p>The ligand synthesis was accomplished by ZK and GT. Equilibrium and relaxometric studies were performed by ZG, EM, FKK, RB, ZB, and GT while were analyzed with the help of IT and EB. Kinetic studies were performed by ZG, EM, FKK, and GT and the data were evaluated with the help of IT and EB. The manuscript was written through contributions of all authors. All authors have given approval to the final version of 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

Comment on Enantioselective total synthesis of (−)-colchicine, (+)-demecolcinone and metacolchicine: determination of the absolute configurations of the latter two alkaloids by B. Chen, X. Liu, Y.-J. Hu, D.-M. Zhang, L. Deng, J. Lu, L. Min, W.-C. Ye and C.-C. Li, <i>Chem. Sci.</i>, 2017, <b>8</b>, 4961–4966

We note that some features of the NMR spectra deposited for the purported isomeric metacolchicines (Chem. Sci. 8, 4961 (2017)) are not compatible with the assignment of those isomers as being atrop-diastereomers. We suggest that there are no such isomeric metacolchicines as reported. The differences in the spectra could rather be a consequence of a (precedented) monomer/dimer equilibrium of metacolchicine in solution.Metacolchicine has been isolated and characterized in 2011. 1 Its structure assignment rests on 1 H and 13 C NMR spectra including 1 H-1 H-COSY, HMBC und NOE investigations. The authors fully assigned the spectra.Recently approaches to the synthesis of metacolchicine were published by Li. 2 The Li group used a complex reaction sequence to arrive at compound 26 which did not fully match the data reported for metacolchicine. This non-identity apparently was reason to prepare the bis-4nitrobenzoate of 26 in order to subject it to X-ray crystallographic analysis. This revealed that compound 26 possesses the structure assigned to metacolchicine.

comment_on_enantioselective_total_synthesis_of_(−)-colchicine,_(+)-demecolcinone_and_metacolchicine:

<p>In the course of their efforts to synthesize metacolchicine the Li group explored a second route originating from colchiceine which furnished compound 4. The spectral data of the latter were claimed to be identical to those of natural metacolchicine. 3 These ndings suggested that there are two different compounds with the constitution of metacolchicine. Li and his coauthors address these, more or less by default, as being stereoisomers, i.e. atrop-diastereomers, 26 ¼ (aR,7S) and 4 ¼ (aS,7S). Unfortunately they did not provide any evidence (such as CD-spectra cf. 4 ) for this interpretation.</p><p>This interpretation appears daring, as the rotational barrier at the axis of colchicine and colchicine-derivatives (which should include metacolchicine) is #90 kJ mol À1 (ref. 5) rendering atrop-diastereomers short-lived. Unfortunately Li and coworkers do not report information on any thermal interconversion of 26 and 4. Atrop-diastereomers can be isolated and studied, though, in the case of iso-colchicinederivatives. 4 In the search for features which are characteristic of individual atropisomers we noted the 1 H NMR signal position of H-7 in the isocolchicines: 4,6 In the (aR,7S)-isomer it appears at 4.6 ppm, whereas in the (aS,7S)-isomers it resonates at $5.0 ppm. This difference is caused by the position of H-7 over the aromatic ring in the (aR,7S)-isomer, whereas aer rotation at the axis and concomitant inversion of the cycloheptane ring H-7 enjoys a position remote from the aromatic ring in the (aS,7S)-isomers. This should apply as well to colchicinederivatives, where the same conformation-based determining factors prevail. Li and coworkers report that the signal of H-7 resonates in both compound 26 and 4 at 4.6 ppm. Hence, both compounds belong to the same (aR,7S)-family. This nding then renders the thesis that 26 and 4 are atrop-diastereomers untenable.</p><p>By the same token, the claim 2 that colchiceine would exist in solution as the (aS,7S)-atropisomer is contrasted by the chemical shi of H-7 in colchiceine which resonates at d ¼ 4.7 ppm.</p><p>When the two isomeric metacolchicines are not atropdiastereomers, what else are they? Any consideration along these lines has to start from the differences in the NMR spectra, which are summarized in Table 1:</p><p>The pattern of the 13 C NMR differences turned out to be not compatible with the assumption, that compound 4 might be an iso-colchicine type compound, that could arise by the synthetic route employed to generate 4 (Nor was this pattern in line with the differences displayed by atrop-diasteromeric iso-colchicide derivatives 7 ).</p><p>In view of the fact that the spectral differences between 26 and 4 are rather small, we found ourselves eventually confronted with the question: are they real? Could it be that the small differences in the NMR spectra are simply caused by solvent or concentration effects? This appeared the more probable, given the solvent and concentration effects reported for the 1 H-and 13 C NMR spectra of colchicine. 8 These effects are due to a monomer/hydrogen-bonded dimer equilibrium. This dimer formation should be even more facile in the case of metacolchicine, due to two additional hydrogen bonds:</p><p>Looking at the 13 C NMR spectra, we compared the differences recorded for 26 vs. 4 with the differences reported for colchicine in CDCl 3 (high dimer content) and DMSO-D 6 (low dimer content).</p>

Royal Society of Chemistry (RSC)

Nitrite reduction by copper through ligand-mediated proton and electron transfer

Nitrite reduction by a copper complex featuring a proton-responsive tripodal ligand is demonstrated. Gaseous nitric oxide was confirmed as the sole NO X by-product in quantitative yield. DFT calculations predict that nitrite reduction occurs via a proton and electron transfer process mediated by the ligand.The reported mechanism parallels nitrite reduction by copper nitrite reductase.

nitrite_reduction_by_copper_through_ligand-mediated_proton_and_electron_transfer

Introduction<!>Results and discussion<!>Conclusions

<p>Nitrite reductases (NiRs) are enzymes found in prokaryotic organisms which catalyze the one-electron (e À ) reduction of nitrite to nitric oxide (NO). 1 Copper nitrite reductases (CuNiRs) are homotrimeric enzymes with each monomer containing two copper centers: a T1 site for electron transfer, and a catalyticallyactive T2 site. 2 Although X-ray crystallographic studies have provided structural snapshots of intermediates along the reduction pathway, 3 the precise mechanism by which CuNiR catalyzes nitrite reduction has been disputed. 1c In one proposed pathway, nitrite coordinates to the reduced T2 center, then following two proton (H + ) transfer events, water is released and a copper-nitrosyl species is generated (described as either Cu(I)-NO + or Cu(II)-NO); Fig. 1A. 3c,4 In support of this mechanism, a crystal structure of CuNiR with NO bound to the reduced T2 site was reported with an unusual side-on binding mode of NO, 5 which suggests that NO coordination to copper is at least possible under reducing conditions. 6 However, an oxidized Cu(NO) unit (i.e. Fig. 1A) would be capable of nitrosylating nearby amino acid residues, 7 and thus is unlikely to be formed under catalytic conditions. An alternative pathway for nitrite reduction catalyzed by CuNiR has been described wherein nitrite rst coordinates to the oxidized T2 center followed by H + transfer from a nearby aspartic acid residue to protonate the coordinated nitrite. 8 In this case, protonation of nitrite triggers e À transfer from the T1 center to the T2 center, with release of NO to form a copper-hydroxide (Fig. 1B). This mechanism is consistent with isolated crystal structures of CuNiR with nitrite bound to the oxidized T2 center, 3c steady-state kinetics and pulsed radiolysis experiments, 9 and computational modeling. 10 Although the intimate pathway of nitrite reduction may be disputed, the network of hydrogen bonds (H-bonds) provided by nearby amino acid residues is widely accepted to play a key role in positioning substrate and facilitating e À transfer. 1c With the goal of clarifying the fundamental pathways of nitrite reduction, the reactivity of synthetic copper complexes toward nitrite has been extensively studied, 1c although limited examples have been reported that demonstrate secondary sphere interactions with a copper nitrite complex. 11 Prior reports have largely focused on the preparation of copper(I) nitrite adducts and subsequent reactivity with exogenous H + sources to release NO. One critical distinction between synthetic copper(I) nitrite adducts and CuNiR is the observed mode of nitrite coordination: there are no reported synthetic copper(I) complexes supported by biologically-relevant ligands that feature the kO-nitrite coordination observed in CuNiR, and instead exclusively feature kN-coordination. 12 While select systems have been shown to produce NO from a copper(I) nitrite complex, 1c,13 the mechanism by which these reactions proceed are oen not fully resolved, thus precluding direct mechanistic comparisons to CuNiR. One reason for limited mechanistic insight has been the isolation of terminal copper(II) complexes which do not contain the inorganic products of nitrite reduction (i.e. H 2 O or NO). 13b,13d,13f Although copper(II)-nitrosyls have been implicated in many synthetic nitrite reduction pathways, their isolation has remained elusive, 7b,14,15 thus calling into question their role in synthetic and biological copper nitrite reduction schemes. A nitrite reduction pathway that circumvents the formation of a highly reactive copper-nitrosyl species is likely operative in CuNiR (Fig. 1B). In this Edge Article, we demonstrate that nitrite reduction by a copper complex supported by a proton-responsive ligand proceeds through a parallel pathway. A mechanism is described whereby a H + /e À transfer pathway to nitrite releases NO, which bypasses the formation of an unstable copper(II)nitrosyl species, and provides a synthetic mechanistic analogue for nitrite reduction in CuNiR.</p><!><p>Recently we described copper complexes supported by H 3 thpa, a tripodal ligand featuring pendent hydroxyl groups capable of engaging in H-bonding interactions with metal bound substrates. 16 We previously showed that the H-bonding manifold presented by H 3 thpa allows access to a unique copper(I) uoride complex (CuF(H 3 thpa), 1) that is best described as containing a 'captured' uoride anion in the secondary coordination sphere. 16b We envisioned that anions capped by SiR 3 + units would react with 1 via metathesis to provide new copper complexes in which we could interrogate H-bonding and H + /e À transfer reactivity toward reducible substrates. Specically, we sought to exploit this methodology to examine the reactivity of 1 with nitrite in an effort to experimentally distinguish between two mechanistic pathways of nitrite reduction. Given the capability of the H 3 thpa ligand scaffold to deliver H + and subsequently provide H-bond donors and acceptors, we hypothesized that the inorganic products of nitrite reduction would be captured within a hydrogen bonding network surrounding the copper center.</p><p>In order to assess the ability of 1 to engage in metathesis reactivity with silyl-anions, we rst examined the substitution of the uoride anion in 1 for chloride. When 1 was treated with an equivalent of Ph 3 SiCl at room temperature, the previously described CuCl(H 3 thpa) 16a complex and Ph 3 SiF were generated quantitatively; both of which were conrmed by 1 H and 19 F NMR spectroscopy (Fig. S1 and S2 †). Based on the clean reactivity observed with Ph 3 SiCl, we synthesized a reagent capable of transferring nitrite to 1 by preparing Ph 3 Si(ONO) via salt metathesis of Ph 3 SiCl and AgNO 2 in benzene solvent. 17 The reaction of 1 with Ph 3 Si(ONO) occurs immediately in dichloromethane solvent: when Ph 3 Si(ONO) is added to a yellow solution of 1, a rapid color change to green occurs. 19 F NMR spectra of the resulting solution conrm the quantitative formation of Ph 3 SiF, indicating metathesis of uoride. The color change is indicative of oxidation from copper(I) to copper(II), a supposition conrmed by EPR and UV-vis spectra collected of reaction solutions. 18 Solid state IR spectra of the isolated green material from this reaction do not show bands associated with nitrite or a metal-nitrosyl species, however a new ligand C]O stretch at 1658 cm À1 was visualized (consistent with a change in protonation state of the ligand), along with broadened ÀOH bands. Based on the above ndings, we hypothesized that the terminal copper-containing product in this reaction was a copper(II)-aquo complex (Cu(OH 2 )Hthpa, 2, Fig. 2).</p><p>We sought an alternative preparation of 2 to conrm its formation during the reaction of 1 with Ph 3 Si(ONO). Authentic samples of complex 2 were prepared by allowing equimolar amounts of the ligand, H 3 thpa, and Cu(OH) 2 to react in dichloromethane/ethanol solution. The solution characterization data (UV-vis and EPR) for 2 prepared in this manner were identical to those from the reaction of 1 with Ph 3 Si(ONO). 18 The crystal structure of 2 contains two independent molecules along with an ethanol solvate which is engaged in H-bonding interactions with only one of the molecules. The solid state structure of one of the independent molecules of 2 is presented in Fig. 2 and reveals a trigonal bipyramidal coordination geometry at copper, consistent with the solution state coordination geometry as determined by EPR and UV-vis spectroscopy. This network of intermolecular H-bonding, which forms a 1-D chain in the extended structure, is presumably a manifestation of crystal packing and unlikely to persist in solution under dilute conditions. 19 In addition to the formation of 2 from 1 and Ph 3 Si(ONO), NO was also conrmed as a reaction product by gas-phase IR spectroscopy, as well as by trapping experiments with CoTPP (TPP ¼ tetraphenylporphyrin) and quantication using UV-vis spectroscopy. 20 Headspace analysis of reaction mixtures revealed two broad bands at 1904 and 1844 cm À1 in the IR spectrum, consistent with NO, as well as bands associated with dichloromethane solvent vapors (Fig. S12 †). To further support the assignment of gaseous NO as a by-product, we prepared the 15 N isotopologue, Ph 3 Si(O 15 NO), and allowed it to react with 1.</p><p>Headspace analysis of the reaction mixture by IR spectroscopy showed two broad bands at 1868 and 1817 cm À1 , consistent with 15 NO, further substantiating NO formation during the reaction of 1 and Ph 3 Si(ONO). The formation of NO was quantitative, as revealed by CoTPP trapping experiments.</p><p>Control reactions conrmed that the quantitative generation of NO was unique to 1. For instance, to examine whether a direct reaction of the silyl-reagent with a metal uoride induced NO extrusion, we performed a control experiment where the known copper(I) uoride CuF(PPh 3 ) 3 (3) 21 was allowed to react with Ph 3 Si(ONO). 22 Headspace analysis of this reaction mixture by IR spectroscopy revealed negligible bands associated with NO (Fig. S12 †). Trapping experiments with CoTPP did not reveal any signicant formation of NO above the background during the reaction of 3 and Ph 3 Si(ONO).</p><p>The quantitative yield of NO from complex 1 cannot be attributed solely to a disproportionation reaction of nitrite. Under sufficiently acidic conditions, nitrite disproportionates to form NO, along with other NO x species. 23 To examine a possible acid-promoted nitrite disproportionation pathway that produces NO, we probed the reactivity of Ph 3 Si(ONO) with CuF(H 3 thpa)BF 4 (4). 16b Complex 4 serves as an ideal platform to test the ability of the ligand framework to deliver H + equivalents to nitrite since (1) e À transfer is not possible and (2) the -OH groups within the putative Cu(H 3 thpa) 2+ dication are expected to be more acidic than in the analogous Cu(H 3 thpa) + monocation. 24 Although NO was detected by IR spectroscopy in the headspace of reaction samples containing 4 and Ph 3 Si(ONO), quantication using CoTPP revealed formation of NO in only ca. 10% yield above the background decomposition. 18 Taken together, these results conrm the necessity of the H 3 thpa ligand framework on copper(I) to mediate nitrite reduction in the absence of any exogenous H + source. The quantitative formation of NO from 1 and Ph 3 Si(ONO) conrms that 1 serves to deliver H + and e À , as opposed to initiating acid-mediated disproportionation.</p><p>The generation of a copper(II)-(OH x ) species from copper(I) and nitrite is reminiscent of a proposed pathway for biological nitrite reduction in CuNiR (Fig. 1B). We sought to investigate key intermediates along the nitrite reduction pathway in our system to provide insight into the reaction sequence that may be applied to CuNiR. However, no intermediates were detected by 1 H NMR spectroscopy during the reaction of 1 with Ph 3 -Si(ONO) at or below À50 C, which precluded mechanistic analysis in solution. 18 Accordingly, we examined the reaction prole using density functional theory (DFT) calculations. 18 The initial nitrite binding step was rst interrogated for a series of possible Cu(H 3 thpa)-nitrite adducts. Three nearly isoenergetic (<0.5 kcal mol À1 difference) structures were computationally identied that feature an interaction of nitrite with the putative Cu(H 3 thpa) + cation (Fig. S16 †). 25 These included an 'encounter' complex (I, Fig. 3A) where, reminiscent of the F À binding in 1, nitrite is positioned in the second coordination sphere by Hbonding interactions. Additionally, two other isomers were located containing two nitrite coordination modes; an h 1 -kO isomer (I 0 ) and an h 1 -kN isomer (I 00 , Fig. S16 †).</p><p>The nitrite reduction sequence subsequent to the initial binding event was evaluated. In analogy to Fig. 1A, nitrite reduction via water elimination from the h 1 -kN isomer (I 00 ) to provide a copper-nitrosyl (III, Scheme S2 †) was found to be an endothermic process by 17.9 kcal mol À1 , presumably due to the formation of a high energy copper(II)-nitrosyl (Fig. S17 †). In contrast, and in analogy to Fig. 1B, reduction via NO elimination from the h 1 -kO isomer (I 0 ) was found to be an exothermic process by 3.8 kcal mol À1 to generate the experimentally observed copper(II)-aquo (II) species (Fig. 3A).</p><p>The transition state for the critical N-O bond breaking step (TS I ) was located computationally and is 21.6 kcal mol À1 above the starting 'encounter' complex I. In the calculated transition state TS I , H + transfer from the H 3 thpa scaffold to the coordi- transfer, accompanied by NO ejection, is followed by an additional H + transfer from the H 2 thpa ligand to copper-bound hydroxide to form the nal copper(II)-aquo product II.</p><p>To understand the impact of the binding mode of nitrite in our system as compared to CuNiR we analyzed the frontier molecular orbitals of structures along the O-NO bond cleavage reaction coordinate. The empty s * O-NO orbital overlaps with the lled Cu d z 2 orbital (Fig. 3B). This is in contrast to CuNiR, where the highest occupied molecular orbital (HOMO) is primarily d x 2 Ày 2 character and subsequently nitrite coordinates in a bidentate manner to maximize orbital overlap between d x 2 Ày 2 and s * O-NO for electron transfer. 10 For Cu(H 3 thpa) + , the HOMO is d z 2 and favors an alternative mode of nitrite binding. The binding of nitrite as a monodentate ligand kO maximizes overlap between d z 2 and s * O-NO for electron transfer (Fig. 3C). The d z 2 ground state predicted for II, subsequent to electron transfer and NO release, is in agreement with the experimental EPR data for 2. 16a,b,18 The difference in mode of nitrite binding between our synthetic system and CuNiR has ramications on the calculated barrier height for O-NO bond cleavage. In CuNiR, bidentate coordination of nitrite to copper maximizes back-bonding interactions and signicantly lowers the calculated barrier to O-NO bond cleavage (16 kcal mol À1 ). 10 When a monodentate coordination was examined in silico, the extent of back-bonding between copper and nitrite was signicantly lower and consequently, a higher barrier to O-NO bond cleavage was observed (26 kcal mol À1 ). 10 The barrier in the present system (21 kcal mol À1 ) is intermediate of these regimes in CuNiR. The Hbonding interactions in TS I between the distal oxygen of the nitrite anion and the ÀOH groups of the H 3 thpa ligand serves to lower the s * O-NO orbital energy, similar in fashion to bidentate coordination to copper. In contrast to energetic consequences on orbital energies imposed by geometric constraints, the hydrogen bonding interactions in H 3 thpa provide an alternative means by which to lower the s * O-NO orbital energy, and this approach may well be exploited as a general strategy to minimize energetic costs associated with substrate reduction. Such strategies have been thoroughly described for biological systems. 26</p><!><p>In conclusion, we have demonstrated nitrite reduction by a copper complex featuring a proton-responsive tripodal ligand. The formation of NO was conrmed by gas-phase IR spectroscopy, isotope labelling, and trapping experiments. DFT calculations predict that nitrite binds in an h 1 -kO fashion to the copper center prior to reduction via a H + /e transfer. This reaction, facilitated by the secondary sphere environment, parallels the crucial role of H-bonding residues near the active site of CuNiR, which serve to position nitrite and facilitate electron transfer. This is, to our knowledge, the rst synthetic copper system to reduce nitrite in such a fashion. 27 Ongoing efforts on this system are focused at utilizing the H 3 thpa scaffold to facilitate multiple H + /e À transfer events to coordinated substrates.</p>

Royal Society of Chemistry (RSC)

Boosting Immunity to Small Tumor-Associated Carbohydrates with Bacteriophage Q\xce\xb2 Capsids

The development of an effective immunotherapy is an attractive strategy towards cancer treatment. Tumor associated carbohydrate antigens (TACAs) are overexpressed on a variety of cancer cell surfaces, which present tempting targets for anti-cancer vaccine development. However, such carbohydrates are often poorly immunogenic. To overcome this challenge, we show here that the display of a very weak TACA, the monomeric Tn antigen, on bacteriophage Q\xce\xb2 virus-like particles elicits powerful humoral responses to the carbohydrate. The effects of adjuvants, antigen display pattern and vaccine dose on the strength and subclasses of antibody responses were established. The local density of antigen rather than the total amount of antigen administered was found to be crucial for induction of high Tn-specific IgG titers. The ability to display antigens in an organized and high density manner is a key advantage of virus like particles such as Q\xce\xb2 as vaccine carriers. Glycan microarray analysis showed that the antibodies generated were highly selective towards Tn antigens. Furthermore, Q\xce\xb2 elicited much higher levels of IgG antibodies than other types of virus like particles and the IgG antibodies produced reacted strongly with the native Tn antigens on human leukemia cells. Thus, Q\xce\xb2 presents a highly attractive platform for the development of carbohydrate based anti-cancer vaccines.

boosting_immunity_to_small_tumor-associated_carbohydrates_with_bacteriophage_q\xce\xb2_capsids

INTRODUCTION<!>Syntheses of Q\xce\xb2\xe2\x80\x93Tn conjugates and control particles<!>Establishment of enzyme-linked immunosorbent assay (ELISA) protocol<!>Effects of immuno-adjuvants and the nature of the particle on anti-Tn antibody responses<!>The effect of dose on anti-Tn immune response and antibody subtype<!>Anti-Tn responses do not correlate with antibody levels towards competing epitopes<!>High local antigen density is critical for high antibody responses<!>Glyco-microarray results confirmed the Tn selectivity in antibody recognition<!>The induced anti-Tn antibodies reacted with \xe2\x80\x9cnative\xe2\x80\x9d Tn epitopes on tumor cells<!>CONCLUSION<!>General procedures for mouse immunization<!>Procedures for ELISA

<p>Vaccine strategies aiming to stimulate the immune system to eradicate tumor cells are highly attractive for tumor prevention and therapy.1 One such approach is to target tumor associated carbohydrate antigens (TACAs), which are overexpressed on the surfaces of many types of malignant cells. Some cancer patients can naturally develop antibodies against TACAs and high levels of anti-TACA antibodies have been correlated with improved prognosis and survival.2 However, the induction of highly potent immune responses against TACAs is challenging.</p><p>The major difficulty in generating effective anti-TACA immunity is that TACAs are typically T cell independent B cell antigens.3 TACAs can directly interact with B cells leading to the secretion of low titers of IgM antibodies, which generally last only a short time. To elicit strong and long lasting immunity, B cell stimulation by T helper cells is required, leading to a switch of antibody isotype from IgM to IgG and the induction of immunological memory.4 In this way, when transformed cells bearing TACAs emerge in immunized hosts, the immune system can recognize the disease related carbohydrate epitopes, quickly reactivate and remove the malignant cells.</p><p>Innovative studies have been carried out to develop TACA based anti-cancer vaccines,5–7 which include glyco-engineering to improve immunogenicity,7–10 synthesis of novel adjuvants,11 enhancement of immune responses through pre-existing antibodies,12–13 as well as the development of novel carriers.14–20 We focus on the last of these approaches, since carriers can provide the necessary epitopes to promote T helper cell activation in a general way, applicable to a variety of antigens.</p><p>Immunogenic proteins such as keyhole limpet hemocyanin (KLH) and tetanus toxoid (TT) are popular choices as carriers, having been conjugated with TACAs to induce strong IgM and IgG responses in various preclinical and clinical evaluations.15, 21 These carrier proteins contain multiple T helper cell epitopes, which are important in the potentiation of antibody responses.21 This mechanistic viewpoint has inspired several elegant reductionist approaches as well, featuring synthetic multi-component vaccines bearing well defined T helper cell epitopes, TACA antigens, and adjuvants with promising results.22–23 In addition, a variety of non-protein carriers have been investigated to improve the quality of anti-TACA immune responses, including dendrimers,14, 20 polysaccharides,16 gold nanoparticles,17–18 and very small proteoliposomes.19</p><p>We have become interested in exploring virus like particles (VLPs) as TACA delivery platforms.24–26 VLPs, which are composed of subunit proteins that self-assemble in a highly ordered manner, emerge as a promising direction in vaccine development.27 Non-infectious to humans or animal hosts, VLPs are safe and highly immunogenic because of several properties,28 including their sizes (<100 nm diameter, promoting both uptake by antigen presenting cells and trafficking to lymph nodes29), repetitive structure promoting B cell recognition,30 ability to crosslink B cell receptors,31 and presentation of immunostimulatory motifs such as RNA ligands of Toll-like receptors.32</p><p>One potential limitation of using immunogenic carrier proteins is the possible diminution of the desired immune response due to pre-existing anti-carrier immunity established by prior exposure to the carrier. Such carrier induced epitopic suppression has been reported for both KLH and TT.33–34 To overcome this challenge, structurally distinct vaccine carriers are needed. This prompted us to investigate the utility of a VLP, i.e., bacteriophage Qβ as a new TACA carrier. This particle is composed of 180 copies of a 14.3-kD subunit, arranged in an icosahedral structure approximately 28 nm in diameter,35 and is very stable toward chemical manipulation. Therapeutic vaccine candidates based on Qβ have demonstrated preclinical or clinical efficacy in a variety of diseases. Herein we report Qβ as a powerful antigen delivery platform for anti-cancer vaccine development by boosting the immune response to a TACA, the Tn antigen.</p><p>The Tn antigen (GalNAc-α-O-Ser/Thr), is overexpressed on the surface of a variety of cancer cells including breast, colon, and prostate cancer,36–37 and is involved in aggressive growth and lymphatic metastasis of cancers.38 The expression level of Tn was found to correlate strongly with overall survival, nodal status and tumor size in breast cancer patients.36 However, the development of anti-Tn vaccines has been quite challenging. Presumably due to its small size, the monomeric Tn antigen did not generate strong responses in immunized mice, even after conjugation to an immunogenic carrier.20, 39–40 To overcome this, a common approach is to cluster multiple consecutive Tn together rendering it more prominent to the immunological system.20, 40 This can be synthetically challenging and time-consuming. Rather than constructing Tn clusters, we attached multiple copies of the monomeric glycan to the icosahedral Qβ scaffold, a one-step operation that can conveniently provide high-density organized displays of the desired epitope. As a weak TACA, Tn serves as a useful model and the results obtained here can provide critical insights for the rational design of effective TACA based anti-cancer vaccines.</p><!><p>Tn ligation onto the Qβ VLP was performed with the copper-catalyzed azide-alkyne cycloaddition (CuAAC) reaction, optimized for bioconjugative operations.41 This process allows high densities of decoration to be achieved on the desired biomolecular scaffold with a minimum of high-value reactants.42 Qβ capsid was first treated with a large excess of the pent-4-ynoic acid N-hydroxysuccinimide ester 1, to place alkyne groups at 540 amino groups (lysine side chains and subunit N-termini) accessible on the outer surface of the capsid (Figure 1). Following removal of the excess reagent, azide modified Tn 3 was added to the alkyne Qβ 2 in the presence of the copper catalyst using ligand 6 (THPTA).41 Depending on the reaction time and reagent concentrations, the number of Tn units attached could be varied. A "high loading" of 340 copies of Tn per particle was produced using only 5 equiv. of azide 3 per subunit (1.5–2 equiv. per alkyne), and the excess Tn could be recovered from the reaction mixture. The remaining unreacted alkyne groups on Qβ virion were capped in a second CuAAC reaction using a large excess of 3-azido 1-propanol 4 producing Qβ-Tn 7. Control particles bearing n-propanol (Qβ-propanol 8) and glucose (Qβ-glucose 9) were prepared similarly, each displaying an average of 450 ± 50 attached units per capsid (Figure 1). These controls were designed to probe the effects of capsid and the carbohydrate structural specificity of immune responses.</p><p>The reactive amino groups on the Qβ exterior are generally well dispersed over the protein surface (Figure 1). While K16 (green residue in figure 1) is most exposed, there are no great differences in accessibility to solution-phase species among the amine groups, except when extensive substitution is desired at the K2 residues clustered tightly around the 3-fold symmetry axis (blue residue). Thus, acylation and subsequent triazole formation should occur evenly at the exterior amine positions.43 (Interior-surface amine groups are blocked by the bacterial and plasmid RNA that is sequestered inside the particle.44) When the overall number of attached Tn moieties per particle is increased, they are expected to be arrayed in larger numbers of repetitive clustered patterns over the icosahedral capsid.</p><!><p>With Qβ-Tn and the control particles in hand, we examined the humoral responses to these particles in the presence of three different adjuvants – complete Freund's adjuvant (CFA), alum, or TiterMax Gold. Three groups of five C57BL/6 mice were vaccinated three times at two-week intervals with Qbeta-Tn 7, Qbeta-propanol 8, or Qbeta-glucose 9 in the presence of the respective adjuvant. An additional group of mice was inoculated using Qbeta-Tn 7 without any exogenous adjuvant (PBS group). One week after the final boost, serum samples were taken and analyzed for anti-Tn antibody responses.</p><p>For ELISA analysis of the resulting sera, a bovine serum albumin (BSA) amide-linked conjugate of Tn was prepared with 30–35 copies of Tn units attached per protein (BSA-Tn 12). The amide connection between Tn and BSA in 12 was used to focus the assay on the Tn moiety rather than to the triazole present in the linker of the administered antigens. Sera from mice receiving Qβ-Tn 7 bound substantially better to BSA-Tn 12 than sera from mice receiving controls 8 or 9, and very little binding was observed to unmodified BSA, indicating that anti- Tn responses were obtained using this protocol (Figure 2). In contrast, no selective binding was observed to a biotinated Tn antigen (13) immobilized onto neutravidin-coated plates, which is a commonly used ELISA protocol (Figure S1).24–25, 45 These divergent ELISA results suggest that polyvalent antibody binding is required for detection since neutravidin-based display (four biotin-binding sites per neutravidin) is likely to be of substantially lower density than the BSA-Tn conjugate resulting in weak interactions and low avidities with anti-Tn antibodies.46</p><p> </p><!><p>Although some intragroup variation was observed, Qβ-Tn immunized mice receiving the CFA adjuvant produced significantly more IgG antibodies than those receiving TiterMax Gold or alum (Figure 3a). Interestingly, mice immunized with Qβ-Tn in PBS (without external adjuvant) also responded. There were no statistically significant differences in IgG levels between the PBS groups and CFA groups, although the CFA groups gave higher mean values. In contrast to the IgG antibodies, the IgM responses from all the groups were similar (Figure 3b). Based on these results, CFA was selected as the adjuvant for further studies.</p><p>For CFA immunization, CFA was only applied in the primary injection. In the two boost injections, incomplete Freund's adjuvant (IFA), which is just mineral oil without mycobacteria components, was used to avoid over-stimulation of the immune system and toxicity of CFA. When IFA was applied as the adjuvant in both prime and boost injections with Qβ-Tn 7, a much poorer IgG response than the CFA protocol was obtained (Figure S2), suggesting the bacterial components in CFA were important in eliciting high IgG responses. The success of Qβ-Tn 7 in PBS without exogenous adjuvant to generate IgG responses, is consistent with the expectation that Qβ is of has favorable characteristics of size, structure, and costimulatory abilities.</p><p>Besides Qβ, other VLPs such as cowpea mosaic virus capsid (CPMV) and tobacco mosaic virus capsid (TMV) could be used as Tn carriers as well.24–25 Tn was linked onto CPMV and TMV with on average 120 copies of Tn on CPMV (Supporting Info, Figures S5 and S6) and greater than 2000 copies of Tn installed on TMV.24 The abilities of CPMV-Tn and TMV-Tn to potentiate anti-Tn IgG responses were compared with that of Qβ-Tn 7 under the identical immunization protocol (CFA adjuvant, 4 μg Tn). As shown in Figure S3, the IgG titers induced in mice following Qβ-Tn immunization were significantly higher compared to the levels induced by either CPMV-Tn or TMV-Tn. The underlying mechanisms behind the superior immunogenicity of Qβ remain to be fully elucidated, which is the subject of further investigation.</p><!><p>The dose of immunogen can significantly impact the immune responses. To establish the optimal dose, mice were immunized with Qβ-Tn 7 in amounts to provide 1, 4 and 20 μg Tn per injection following the aforementioned protocol. Determination of antibody titers by ELISA against BSA-Tn 12 showed high anti-Tn IgG titers for all Qβ-Tn immunized groups (Figure 4a). The highest average titer (263,000) was obtained from the 4 μg group although the differences between groups did not reach statistical significance. The mean IgG titers of the 4 μg group were two orders of magnitude higher than those from the control groups (Qβ-propanol and Qβ-glucose, p < 0.001). Substantial anti-Tn IgM antibodies were elicited with highest average titer (13,400) from the 20 μg group (Figure 4b). As discussed in the introduction, it has been challenging to elicit high titers of IgG antibodies using monomeric form of Tn as the antigen.39–40 The high titers obtained with Qβ-Tn highlights the advantage of Qβ as a TACA carrier.</p><p>As discussed above, a prerequisite for potent IgG responses is the stimulation of B cells by T helper cells.4 The high IgG titers obtained following Qβ-Tn immunization indicate that Qβ contains the necessary T helper epitopes for T cell potentiation and Tn functionalization of Qβ does not significantly disrupt the functions of these epitopes. We suggest that the covalent linkage between Tn and Qβ ensures that Tn is processed and presented together with Qβ by the same B cells, thus presumably allowing potent B cell stimulation by matched activated T helper cells. The subclasses of IgG were also analyzed (Figure S4). Substantial amounts of all IgG subclasses were generated with Qβ-Tn 7, indicating the induction of a balanced IgG immune response involving a broad range of cytokines,47 as well as T helper cell independent IgG3response.48</p><!><p>For each group of Qβ-Tn immunized mice, variations were observed in anti-Tn IgG titers. Immune responses to the capsid or the triazole linker may be competitive, suppressing anti-glycan immunity in some cases. To test this possibility, we analyzed the levels of antibodies against unmodified Qβ particle and the triazole-containing linker, the latter displayed on BSA as conjugate 14.</p><p>All sera, including those from mice immunized with control particles 8 and 9, showed responses to both capsid and the triazole containing linker with no correlations to the levels of anti-Tn antibodies (Figure 5 and Figure S5). This suggests that the low magnitudes of the anti-Tn IgG responses observed for some Qβ-Tn 7 immunized mice were not due to immune impairment. The anti-Qβ virion and anti-triazole antibody titers did not significantly vary with the dose or structures of Qβ constructs administered (Figure S5). The compatibility of the triazole linker with TACA based vaccine constructs is consistent with several prior reports.42, 49–51 It is known that rigid structures such as triazole can induce antibodies against themselves, which may consume the Th cells and reduce anti-glycan responses.52 Thus, it is possible that a flexible linker between Tn and Qβ can reduce the amount of anti-linker antibodies thus further improving the anti-Tn immunity, which will be explored in the future.</p><!><p>To better understand the origin of immune potentiating effect of Qβ, two additional Qβ-Tn constructs (Qβ-Tn 10 and 11, Figure 6a) were prepared with lower Tn densities by decreasing the amount of Tn-azide 3 used in the CuAAC conjugation reaction. Relative to particle 7 (high loading, designated "H", 340 Tn per particle), particle 10 bears a medium Tn density (150 copies per particle, "M"), and 11 is lightly loaded (78 Tn molecules per particle, "L"). As each Qβ capsid contains 180 subunits, in the high density Qβ-Tn 7, there were on average close to 2 Tn per capsid with each pair of the carbohydrates expected within 5 nm of each other. Lower Tn loading on Qβ-Tn 10 and 11 provides more dispersed presentations of Tn.</p><p>Two sets of immunization experiments were performed. First, three groups of mice (groups 1–3) were immunized with Qβ-Tn 11, 10 and 7 respectively, with the amounts of Qβ capsid kept constant within all three groups (Figure 6a). Due to the differential Tn loading on the capsid, these groups received increasing amounts Tn. ELISA analysis showed that group 3 mice generated the highest IgG antibody responses (Figure 6b). This suggested that although equal amounts of T helper epitopes were presented by the capsid, higher levels of Tn (~ 1 μg) were needed to generate strong IgG titers. Interestingly, the Tn specific IgM responses, indicative of early B cell activation, were comparable from all groups (Figure 6c). This is consistent with results by Bachmann and co-workers,53 who reported that the density of peptide antigens on Qβ did not affect IgM titers but had great influence on the T helper cell dependent IgG responses. In another study, Schiller and co-workers demonstrated that to elicit strong antibodies against tumor necrosis factor-α using papillomavirus VLP as the carrier, a high density of antigen was crucial to break B cell tolerance against this self antigen.54</p><p>In the second experiment, the amount of Tn was kept constant, and unmodified Qβ particles (designed "N" in Figure 6a) were added when necessary to equalize the total amount of particle administered. Thus, group 4 received 75 μg of particle 11 in each injection, which delivered 1 μg of attached Tn. Groups 5 and 6 received 33 μg and 17 μg of particles 10 and 7, respectively, to carry the same total amount of Tn, along with the unmodified particle to bring the total virion to 75 μg. In group 4, Tn antigen was randomly distributed on all Qβ capsids. In contrast, Tn was present only on 23% of the capsid in group 6, but with much higher local organization. If antigen display patterns were not important, groups 4, 5, and 6 should give similar anti-Tn responses. However, ELISA analysis demonstrated that only group 6 generated significant IgG responses (Figure 6b). As before, IgM response was independent of these variations (Figure 6c).</p><p>These results suggest that the organized epitope display is crucial for induction of antibody isotype switching from B cells, but not for early B cell activation. This is consistent with the fact that the spacing between many neighboring Tn molecules on Qβ-Tn 7 is approximately around 5 nm, which is suitable for strong BCR crosslinking, leading to potent B cell activation and isotype switching producing IgG antibodies.3 BCR crosslinking rather than simple binding has been proposed as a requisite signal to induce IgG production.31 Highly organized antigen geometry cannot be achieved with the traditional amorphous carriers such as KLH and tetanus toxoid (TT) highlighting a key advantage of Qβ as antigen carriers.</p><p>Groups 3 and 6 mice received the same amount of Tn antigen displayed in the same organized manner, while group 6 had four times the amount of Qβ capsids. The IgG levels generated from these two groups were comparable (Figure 6b). In addition, comparison of groups 1 and 2 with 4 and 5 demonstrated that increasing the dose of Qβ and thereby the amount of T helper epitopes available did not compensate for the inefficient IgG response resulting from poorly organized antigen display. These results together suggest excess Qβ did not significantly suppress55 or improve humoral responses under the conditions evaluated.</p><!><p>Our ELISA results demonstrated that the post-immune mouse sera contained antibodies capable of recognizing the Tn antigen. To establish the binding specificities of the antibodies, the sera from mice immunized with Qβ-Tn 7 and control particles 8 and 9 were screened against a carbohydrate microarray.56 This glycan array contained 329 members, which were prepared by attaching a variety of O-linked glycopeptides, N-linked glycopeptides, glycolipids, and glycoproteins to BSA, followed by printing the BSA conjugates on the slides (A full list of the library components is given in Table S1 in Supporting Information). After incubation with each mouse serum sample and washing, a fluorescently labeled secondary antibody was used to quantify the relative amounts of serum antibody bound to individual array components. For clarity, the average intensities from each group of mice were presented (Figure 7). Consistent with the ELISA results, sera from all mice immunized with Qβ-Tn 7 reacted well with the immunizing antigen (R-Tn(Ser)-hydroxyethylamide in Figure 7) compared to the control sera and the pre-vaccination sera. A variety of Tn containing glycopeptides and GalNAc terminated glycans were also recognized suggesting the GalNAc moiety was the major element in recognition. Both monomeric Tn and Tn clusters reacted with the Qβ-Tn post-immune sera. The microarray also contained spots with BSA bearing different densities of terminal GalNAc: GalNAc-α-04 (4 GalNAc per BSA) vs. GalNAc-α-22 (22 GalNAc per BSA) and GalNAcα1-6Galβ-04 (4 GalNAcα1-6Gal per BSA) vs. GalNAcα1-6Galβ-22 (22 GalNAcα1-6Gal per BSA) (see GalNAc terminal glycans group in Figure 7). In both cases, significantly greater binding was observed to the higher-density display, much like the aforementioned ELISA observations using BSA-Tn 12 and Tn-biotin 13.</p><p>Although the dose of Qβ-Tn did not significantly impact the titers against the immunizing Tn antigen 3, interesting dose dependent effects were observed on antibody selectivity and specificity. Both IgG and IgM antibodies from the 20 μg group bound strongest with Tn containing array components compared to antibodies from the groups receiving lower doses (Figure 7). Furthermore, antibodies from the 20 μg group showed more specificity towards GalNAc and less cross-reactivities with other glycans. In other words, the higher glycopeptide antigen dose induced a more focused immune response toward the glycan antigen, suggesting the potential advantage of using high antigen dose for immunization.</p><!><p>For an effective immunotherapy, it is important that anti-TACA antibodies generated by the synthetic vaccine can recognize the antigens present in the native environments such as the surfaces of cancer cells. To test this, we analyzed the binding of sera showing high IgG titers by ELISA to Jurkat cells, a human leukemia cell line expressing Tn antigens on the surface (Figure 8).57 Significant cell recognition by these serum antibodies was observed using flow cytometry, whereas serum from mice given control particle 8 or 9 showed no binding (Figure 8d). Consistent with ELISA data, the percentages of cells exhibited positive reactivities with Qβ-Tn in FACS analysis were higher than those from TMV-Tn (Figure S6).</p><!><p>We demonstrate here that the bacteriophage Qβ capsid is a powerful antigen delivery platform capable of boosting the humoral immune responses to a very weak TACA, the Tn antigen. The addition of external Freund's adjuvant to Qβ-Tn conjugates produced higher IgG titers, although strong immune responses were also obtained without exogenous adjuvants. The IgG antibodies produced were highly selective towards Tn antigen binding and reacted strongly with native Tn antigens on human leukemia cells demonstrating the physiological relevance of the IgG antibodies induced by the Qβ-Tn vaccine construct. Once the density of Tn antigen administered passed a threshold level, the pattern of Tn antigen display, rather than the total amount of Tn antigen administered appeared to be crucial for the induction of high anti-Tn IgG titers. The knowledge gained from the current study will facilitate the rational design and optimization of carbohydrate based anti-cancer vaccines using virus-like particles such as Qβ as a powerful antigen delivery system.</p><!><p>Pathogen-free C57BL/6 female mice age 6–10 weeks were obtained from Charles River and maintained in the University Laboratory Animal Resources facility of Michigan State University. All animal care procedures and experimental protocols have been approved by the Institutional Animal Care and Use Committee (IACUC) of Michigan State University. Groups of five C57BL/6 mice were injected subcutaneously under the scruff on day 0 with 0.1 mL various Qβ constructs as emulsions in complete Freund's adjuvant (Fisher), TiterMax Gold adjuvant (Sigma Aldrich), Imject Alum (Thermo Scientific), or Incomplete Freund's adjuvant (Fisher) according to the manufacturer's instructions. Boosters were given subcutaneously under the scruff on days 14 and 28 (for complete Freund's adjuvant group, incomplete Freund's adjuvant was used). Serum samples were collected on days 0 (before immunization), 7 and 35. The final bleeding was done by cardiac bleed.</p><!><p>A 96-well microtiter plate was first coated with a solution of BSA-Tn, BSA-triazole, or wide type Qβ viron in PBS buffer (10 μg mL−1) and then incubated at 4 °C overnight. The plate was then washed four times with PBS/0.5% Tween-20 (PBST), followed by the addition of 1% (w/v) BSA in PBS to each well and incubation at room temperature for one hour. The plate was washed again with PBST and mice sera were added in 0.1% (w/v) BSA/PBS. The plate was incubated for two hours at 37 °C and washed. A 1:2000 dilution of horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG, IgM, IgG1, IgG2b, IgG2c, or IgG3 respectively (Jackson ImmunoResearch Laboratory) in 0.1% BSA/PBS was added to each well. The plate was incubated for one hour at 37 °C, washed and a solution of 3,3,5′,5′-tetramethylbenzidine (TMB) was added. Color was allowed to develop for 15 min and then a solution of 0.5 M H2SO4 was added to quench the reaction. The optical density was then measured at 450 nm. The titer was determined by regression analysis with log10 dilution plotted with optical density. The titer was calculated as the highest dilution that gave three times the absorbance of normal mouse sera diluted at 1:1600 (about 0.1 for all sera).</p>

PubMed Author Manuscript

An investigation of the reactions between azido alcohols and phosphoramidites

The reactions of several \xce\xb2-, \xce\xb3-, and \xce\xb4-azido alcohols with dibenzyl and dimethyl N,N-diisopropylphosphoramidites were examined. Detailed analysis of the intermediates and products formed from the reactions under different conditions provided useful information to gain insights into their mechanisms involving intramolecular Staudinger reaction, as well as the structure-reactivity relationships of both substrates. The reactions of \xce\xb3- and \xce\xb4-azido alcohols with dibenzyl N,N-diisopropylphosphoramidite could produce 6- and 7-membered cyclic phosphoramidates, thereby providing a new synthetic method for these biologically important molecules.

an_investigation_of_the_reactions_between_azido_alcohols_and_phosphoramidites

<p>The Staudinger reaction is valuable for creating phosphorous-nitrogen bonds via the reaction between phosphites and organic azides (Scheme 1).1 This reaction forms first a phosphazide ylide intermediate, which can convert into a phosphazine intermediate by losing a molecule of nitrogen. Next, the resultant phosphazine is transformed into phosphoramidate as the end-product either spontaneously or under mild acidic conditions.</p><p>Previously, we observed the undesirable formation of a cyclic phosphoramidate (1) in the synthesis of a CD52 GPI anchor when an azido alcohol (2) was reacting with a phosphoramidite.2 We hypothesized that the reaction first formed an azido phosphite intermediate 3, which underwent an intramolecular Staudinger reaction to result in 1 (Scheme 2). According to this mechanism, it was expected that other azido alcohols might behave similarly, thereby forming a novel synthesis of cyclic phosphoramidates.</p><p>To the best of our knowledge, the Staudinger reaction has not been used to synthesize cyclic phosphoramidates, which are useful compounds in the design and development of pesticides, cancer chemotherapeutics and enzyme inhibitors.3–7 Current synthetic routes to these molecules include the reactions of phosphoryl trichloride with amino alcohols or other methods using relatively harsh reaction conditions.3–7 One exception to this is the recent development by the Johnston group of a new method for the synthesis of C- and P-chiral cyclic phosphoramidates employing Brønsted acid-catalyzed intramolecular phosphoramidic acid addition to alkene.8</p><p>To investigate the above hypothesis, we prepared several azido alcohols and probed their reactions with phosphoramidites. The reaction of dibenzyl N,N-diisopropylphosphoramidite 4 with a δ-azido alcohol 5 in the presence of tetrazole was very fast (finished in about 10 min at room temperature) to produce a less polar product, which remained unchanged in the mixture for hours. However, our attempt to isolate this product by silica gel column chromatography failed. We suspected that it might be the azido phosphite intermediate that was oxidized in air on the column. To examine this assumption, we treated the reaction mixture with m-chloroperbenzoic acid (m-CPBA) before workup, in an effort to transform the azido phosphite into the corresponding phosphate. Indeed, under revised conditions, azido phosphate 6 was isolated as the major product in a 75% yield (Table 1, entry 1) after silica gel column chromatography. Similarly, the reaction of 4 with γ-azido alcohol 7 or β-azido alcohols 9 and 11 afforded azido phosphates 8, 10 and 12, respectively, in good to excellent isolated yields (Table 1, entries 2–4). These results have proved unambiguously that while the reactions between 4 and alcohols are easy and fast to afford azido phosphites that are stable in the reaction mixture at room temperature, under these conditions inter- and intramolecular Staudinger reactions were very slow. Buoyed by our previous discovery of cyclic phosphoramidate formation,2 we began to explore different conditions to promote the intramolecular Staudinger reaction of the azido phosphite intermediates formed from the above reactions. We found that in the presence of tetrazole, the reactions of 4 with 5 and 7 at 80 °C gave cyclic phosphoramidates 13 and 14, respectively, in good to moderate yields (Table 2, entries 1 and 2). The reaction products were conveniently isolated by column chromatography and fully characterized with HR-MS and NMR. To our surprise, however, the reaction between 4 and cis-β-azido alcohol 9 under the same conditions resulted in a complex mixture but did not produce a significant amount of the corresponding cyclic phosphoramidate, even though the azido and the hydroxyl groups in 9 were at cis positions, which are favorable for intramolecular reactions. Not surprisingly, the reaction between 4 and unfavorable trans-azido alcohol 11 did not afford the cyclic phosphoramidate either. The results suggested that the ring size of the reaction intermediate or product might play a critical role in intramolecular Staudinger reactions of azido phosphites.</p><p>In view of the potential influence of electronic and steric effects of the phosphite group on its Staudinger reactions,9 we studied subsequently the reactions of the same set of azido alcohols with phosphoramidite 15 bearing methyl instead of benzyl esters. We anticipated that this would facilitate the intramolecular Staudinger reaction of the resultant azido phosphite intermediates. Indeed, even at room temperature the reaction of 15 with 5 produced a modest yield (40%) of the desired cyclic phosphoramidate 16 (Table 3, entry 1). More interestingly, a substantial quantity (10%) of cyclic phosphorazide 17 was also obtained. Clearly, 17 was stable to column chromatography. The reactions of 15 with azido alcohols 7 and 9 were quite complex and the isolated major products were phosphorazides 18 and 19 (Table 3, entries 2, 3), respectively. For azido alcohol 11, the reaction with 15 was clean but the product was isolable only after oxidation with m-CPBA to afford azido phosphate 20 in an excellent yield (91%), suggesting the formation of a stable azido phosphite intermediate. These results revealed that the azido phosphite intermediates formed from the reactions of dimethyl phosphoramidite 15 with azido alcohols could undergo intramolecular Staudinger reactions even at room temperature, except where the cyclization was hindered by unfavorable trans configurations in substrate 11. Evidently, these reaction conditions were much milder than those required for the intramolecular Staudinger reaction to occur with dibenzyl phosphoramidite 4.</p><p>The results shown in Table 2 and Table 3 clearly demonstrated that the reaction yield and stability of the cyclic phosphoramidate products followed an order of: 7-membered ring > 6-membered ring > 5-membered ring. In agreement with this trend, the results in Table 3 also suggested that the reaction yield of phosphorazide intermediates was in a reversed order, namely, 19 > 18 > 17. For the reaction of 15 with 7 or 9, no cyclic phosphoramidate product was observed, whilst the latter reaction gave the highest isolated yield of phosphorazide 19. Both the cyclic phosphoramidate and the cyclic phosphorazide products should have been formed via intramolecular Staudinger reaction (see Scheme 2). However, the observation of mainly cyclic phosphorazide for the reactions of 15 with 7 and 9, as well as a mixture of cyclic phosphoramidate and phosphorazide for the reaction of 15 with 5, also indicated that intramolecular Staudinger reactions of the azido phosphite intermediates were not the limiting factor for the transformation into phosphoramidate products. Yet, another interesting finding was that at room temperature 17 could not be converted into 16 on treatment with tetrazole and diisopropanylamine, indicating that 16 obtained from the reaction of 15 with 5 might be formed through a more reactive intermediate rather than through cyclic phosphorazide 17. Furthermore, heating β-azido phosphites at 80 oC gave complex results that were different from the reactions of β-azido alcohols with phosphites or phosphines.10,11</p><p>To explain the above results and observations, we propose that the phosphazide ylide intermediate generated by intramolecular Staudinger reaction of azido phosphite endured two competitive pathways as shown in Scheme 3. One of these pathways involved ring contraction via automatic elimination of a nitrogen molecule to produce phosphazine 21 that could undergo protonation and nucleophilic de-O-alkylation to afford cyclic phosphoramidate 1 (Scheme 3A). Another pathway involved direct protonation and nucleophilic dealkylation of the phosphazide ylide intermediate to give cyclic phosphorazide 22. It seemed that once phosphazide 22 was formed it could not be converted into phosphoramidate under the probed conditions. However, as some of the reactions were very complex and gave low yields of identified products, we could not completely eliminate the possibility of intermolecular Staudinger reaction of azido phosphites to form oligo-/polymers, which are difficult to purify and characterize.</p><p>Moreover, the results indicated that the ring contraction reaction was highly dependent on the reaction condition and the substrate structure. It was favorable at higher temperature and with larger cyclic phosphazide ylides (≥ 8-membered ring, producing ≥ 6-membered ring product). At 80 °C, cyclic phosphoramidates were obtained as the major products and no phosphazide was isolated in a significant yield from any of the reactions, probably because, even if phosphazides were generated, they would decompose at a high temperature. In addition to the 7- and 6-membered ring products 13 and 14 that were readily formed from the reactions of 4 with 5 and 7 (Table 2), the by-product observed in our synthesis of a CD52 GPI anchor was an 8-membered ring cyclic phosphoramidate, which could be formed at room or lower temperature.2 Therefore, it seemed that the 8-membered cyclic phosphoramidate product was even more readily generated from intramolecular Staudinger reaction than the 7-membered cyclic phosphoramidate, such as 13. Continuing this trend, 5-membered cyclic phosphoramidate products were not observed from the reaction of 9 with either 4 or 15.</p><p>To further verify the conclusion that it would be difficult to form 5-membered cyclic phosphoramidates through intramolecular Staudinger reaction, we prepared two more β-azido alcohols, 23 and 24, and probed their reactions with 4 and 15 under above-described conditions (Scheme 4). The results were in agreement with our prediction. As shown, the reactions of 23 and 24 with dibenzyl phosphoramidite 4 at room temperature produced the corresponding azido phosphites as major products, which were converted into azido phosphates 25 and 28, respectively, after m-CPBA oxidation; however, at 80 °C, these reactions were complex and no major products could be isolated and identified. For the reactions of 23 and 24 with dimethyl phosphoramidite 15 at room temperature, a significant amount of phosphazides 26 and 29 were isolated. The results were similar to that of the reactions between related 1,2-azido alcohol 9 and phosphoramidites 4 and 15. In addition, aziridine phosphoramidate 27 was also observed in the reaction of 23 with 15, and similar products were reported previously.12–14</p><p>In summary, the reactions between phosphoramidites and azido alcohols were examined in detail. It was disclosed that the results of these reactions were dependent on the substrate structure and reaction conditions. The capture of azido phosphite intermediates or isolation of cyclic phosphorazide products from the reactions carried out at room temperature enabled us to gain more insights into the mechanisms of these reactions and the structure-reactivity relationships of both substrates. Significantly, it was observed that the reactions of γ- and δ-azido alcohols and their larger homologs with dibenzyl N,N-diisopropylphosphoramidite could produce corresponding cyclic phosphoramidates in good yields. These yields may be improved further via optimization of reaction conditions, paving the foundation for a new and facile method to access these important molecules under neutral and mild conditions. This method could be especially useful for synthesizing large cyclic phosphoramidates, such as macrocyclic phosphoramidates.15</p>

PubMed Author Manuscript

A microscopic description of SARS-CoV-2 main protease inhibition with Michael acceptors. Strategies for improving inhibitor design

A microscopic description of SARS-CoV-2 main protease inhibition with Michael acceptors. Strategies for improving inhibitor design Classical and QM/MM Molecular Dynamics simulations are used to unveil the reaction mechanism for the SARS-CoV-2 3CL protease inhibition with a Michael acceptor. The on-the-fl y string method is used to explore the multidimensional free energy landscape associated with the reaction showing that the process requires the activation of the catalytic dyad (Cys-His), followed by the nucleophilic attack and a water-mediated proton transfer. These fi ndings could guide the design of new inhibitors of the protease to be employed as therapeutic agents against COVID-19. rsc.li/chemical-scienceA microscopic description of SARS-CoV-2 main protease inhibition with Michael acceptors. Strategies for improving inhibitor design † Carlos A. Ramos-Guzm án, J. Javier Ruiz-Pern ía * and I ñaki Tu ñ ón *The irreversible inhibition of the main protease of SARS-CoV-2 by a Michael acceptor known as N3 has been investigated using multiscale methods. The noncovalent enzyme-inhibitor complex was simulated using classical molecular dynamics techniques and the pose of the inhibitor in the active site was compared to that of the natural substrate, a peptide containing the Gln-Ser scissile bond. The formation of the covalent enzyme-inhibitor complex was then simulated using hybrid QM/MM free energy methods.After binding, the reaction mechanism was found to be composed of two steps: (i) the activation of the catalytic dyad (Cys145 and His41) to form an ion pair and (ii) a Michael addition where the attack of the Sg atom of Cys145 to the Cb atom of the inhibitor precedes the water-mediated proton transfer from His41 to the Ca atom. The microscopic description of protease inhibition by N3 obtained from our simulations is strongly supported by the excellent agreement between the estimated activation free energy and the value derived from kinetic experiments. Comparison with the acylation reaction of a peptide substrate suggests that N3-based inhibitors could be improved by adding chemical modifications that could facilitate the formation of the catalytic dyad ion pair.

a_microscopic_description_of_sars-cov-2_main_protease_inhibition_with_michael_acceptors._strategies_

Introduction<!>Results and discussion<!>Formation of the covalent E-I complex<!>Conclusions<!>Conflicts of interest

<p>A powerful strategy to ght against infectious diseases is the development of drugs to inhibit the activity of one of those enzymes that are crucial in the life cycle of the pathogenic agents. This is the case of the main protease, or 3CL protease (3CL pro ), of coronaviruses in general and of SARS-CoV-2 in particular. The 3CL pro cleaves the polyproteins translated into the infected cells to produce functional proteins for the coronavirus. 1 As in other cysteine proteases, the proteolysis is performed in the active site of 3CL pro by a Cys/His catalytic dyad, the substrate cleavage taking place between Gln at the P1 position of the peptide chain and a Gly/Ala/Ser at the P1 0 one. 2 This enzyme plays an essential role during the replication of the virus and has no closely related homologues in human cells, making it in an attractive drug target. 3 Several lead compounds have already been demonstrated to be effective at inhibiting the activity of SARS-CoV-2 3CL pro , including Michael acceptors, 4 a-ketoamides, 5 carbamoyl derivatives 6 and aldehydes. 7 N3 is a Michael acceptor, an a,b-unsaturated carbonyl compound, that was designed as an inhibitor of the 3CL pro of several coronaviruses, including SARS-CoV and MERS-CoV 8 and that has been demonstrated to have inhibitory activity against the ortholog enzyme of SARS-CoV-2. 4 This compound has a chemical structure similar to that of a peptide, the natural substrate of the enzyme (see Fig. 1). However, the microscopic details of 3CL pro inhibition by N3 are still unclear.</p><p>Kinetic experiments showed that N3 is a potent timedependent irreversible inhibitor of SARS-CoV-2 3CL pro that follows the next kinetic scheme:</p><p>In a rst stage, the inhibitor reversibly binds into the active site of the enzyme forming a noncovalent complex (EI) with a dissociation constant (K I ¼ k 2 /k 1 ). Aerwards, the inhibitor irreversibly reacts with the enzyme, with a rate constant k 3 , to give a stable acylenzyme (E-I). This acylenzyme is characterized by the formation of a covalent bond between the Sg atom of Cys145 and the Cb atom of the inhibitor, as observed in the Xray structure of the inhibited enzyme. 4 N3, or any of the other 3CL pro inhibitors characterized until now, can be used as a starting point for the development of an efficient drug for the treatment of COVID-19. One of the steps in this development is the optimization of the thermodynamic (binding) and kinetic properties of the inhibitor. This improvement should be based in the microscopic knowledge of the inhibition process, which in part relies on the details provided by simulations of the enzyme and the complex formed with the inhibitor. The analysis of the reaction step requires of the use of QM/MM potentials, which are adequate to describe bond forming and breaking processes. QM/MM techniques have been already employed to study, at microscopic level, the SARS-CoV-2 3CL pro hydrolysis mechanism with a natural peptide 9 and a modied peptide having a uorescent tag as leaving group 10 as substrates. We here use these methods to investigate the inhibition process of this enzyme by N3. The atomistic details provide here could be applied to improve the design of future drugs based in this compound.</p><!><p>As detailed in the Methods section (see ESI †), we carried out classical Molecular Dynamics (MD) simulations of the noncovalent enzyme-inhibitor (EI) complex built from the 7BQY PDB structure. 4 A total of 4.0 ms (2 replicas) of classical MD simulation were run using the AMBER19 GPU version of pmemd. 11,12 We then explored the reaction mechanism for the formation of the covalent acylenzyme complex (E-I, see eqn (1)) using QM/MM simulation methods at the hybrid B3LYP/MM level, 13,14 including D3 dispersion corrections, 15 with the 6-31+G* basis set, as explained in Methods section. The stringmethod 16,17 was employed to nd the reaction minimum free energy paths (MFEP) on multidimensional free energy surfaces dened by a set of Collective Variables (CVs) in which we included those geometrical parameters (bond lengths) suffering noticeable changes during the process. A path-CV (s) that measures the advance along the MFEP was dened to trace the corresponding free energy proles. Umbrella sampling 18 along a distinguished coordinate was used to obtain the free energy difference between the neutral and ion pair (IP) forms of the catalytic dyad (Cys145/His41) within the same QM/MM approach. This methodological combination was previously used in the study of the acylation and de-acylation steps of a natural peptide substrate by SARS-CoV-2 3CL pro with results in excellent agreement with experiments. 9 In that work we used the string method to explore the reaction mechanism, with different starting points and initial guesses for the path. In the string method, the initial and nal nodes are allowed to evolve until they reach free energy minima while the rest of nodes trace a MFEP between them. Aer all our attempts, we found two kinds of mechanisms: either the reaction goes through a metastable IP intermediate, where the Cys-His proton transfer precedes the dyad attack on the substrate or the reaction proceeds without formation of the ion pair by means of a direct proton transfer from the catalytic cysteine to the substrate; presenting this last mechanism a signicantly higher activation free energy. 9 The existence of an IP dyad is compatible with the experimental observations made during the kinetic characterization of the homologue protease of SARS-CoV (96% identical), in which a proteolysis mechanism involving the IP formation was proposed on the basis of the pH-inactivation prole and the analysis of solvent isotope effects. 19 We have then here explored similar reaction mechanisms for the SARS-CoV-2 3CL pro protease inactivation with N3. It must be also noticed that, in a recent work published during the revision of this manuscript, Moliner and coworkers explored SARS-CoV-2 inhibition by N3 and related inhibitors nding also a reaction mechanism that involves rst the formation of the IP followed by the formation of the enzyme-inhibitor covalent bond. 20 The noncovalent EI complex N3 has a chemical structure that resembles that of a peptide substrate (see Fig. 1). As such, the pose found for N3 in the active site of SARS-CoV-2 3CL pro during our MD simulations is also quite similar to that described for the peptide. 9 Fig. 2a shows the N3 inhibitor in the active site of one of the protomers of the dimeric enzyme. Analysis of root-mean-square-deviation (RMSD) of the protein and the inhibitor shows that this conguration of the system was stable during the simulated time (see Fig. S1 †). In this complex, the Sg atom of Cys145 remains close to the Cb atom of N3. According to the probability distribution shown in Fig. 2b, the most probable distance between these two atoms is 3.3 Å (there is a small fraction of congurations with larger distances, about 5 Å, that corresponds to a congurational change of the side chain of Cys145 from trans to gauche conformation). In the enzyme-inhibitor complex the catalytic dyad remains hydrogen bonded with the sulydryl proton pointing towards the N3 atom of His41 (the most probable donor-acceptor distance is 3.3 Å, see Fig. 2b). This conguration suggests a mechanism for the formation of the acylenzyme (E-I) in which the catalytic dyad could be activated by means of a proton transfer from Cys145 to His41 to form an ion pair (IP). This activation mechanism of the catalytic dyad was found to be the rst step in the acylation of the natural substrate 9 and it is compatible with experimental kinetic observations on the ortholog enzyme of SARS-CoV, which is highly similar to the 3CL pro of SARS-CoV-2. 19 Fig. 2c displays the fraction of hydrogen bond interactions established between protein residues and the different groups of the inhibitor and the peptide substrate. 3CL pro presents an absolute requirement for Gln at P1 position. 21 As seen in Fig. 2c the P1 residue of the peptide substrate is the one establishing more hydrogen bond interactions with the enzyme. 9 In particular, main chain atoms of Gln-P1 form hydrogen bonds with residues Gly143, Ser144 and His164, while the side chain is accommodated through hydrogen bond contacts with Phe140, Leu141, His163 and Glu166. In N3, the Gln residue is substituted by a g-lactam ring at P1 position, which essentially reproduces the same hydrogen bond interactions. The interaction pattern of the P2-P5 groups is also quite similar in the inhibitor and the peptide substrate, which could explain the affinity between the enzyme and the inhibitor. Important differences appear at the P1 0 site, where the serine residue of the peptide substrate is substituted in the inhibitor by a benzyl ester group. While the main chain O atom of Ser-P1 0 forms hydrogen bonds with the amide group of Gly143 and the side chain of Asn142, the hydroxyl group of its side chain can contact the catalytic dyad (Cys145 and His41). In the inhibitor, the carbonyl O atom of the P1 0 group can form a hydrogen bond contact with the main chain NH of Gly143. The terminal benzyl group can establish a CH/p interaction with the methyl group of Asn142 side chain, while the other side of the ring remains solvent exposed. Because of its mobility, this benzyl ester group can also establish interactions with threonine residues placed at positions 24-26 and nearby residues. It is also noticeable to remark that the strong interactions established by Gly-P2 0 of the peptide substrate with Thr25 and Thr26 are signicantly weakened or absent in the inhibitor, opening a way to improve the binding affinity between the protease and N3.</p><!><p>According to the analysis performed on the noncovalent EI complex, we investigated a possible activation of the catalytic dyad via IP formation. With this purpose, we obtained the B3LYPD3/6-31+G*/MM free energy prole associated to the proton transfer from the Sg atom of Cys145 to the N3 atom of His41 using an antisymmetric transfer coordinate (d(Sg-H)d(N3-H)). Fig. 3a shows the prole obtained with N3 present in the active site; as well as the free energy proles corresponding to the same proton transfer in the apo enzyme and when a peptide substrate is present in the active site. 9 The free energy cost to form the IP from the neutral catalytic dyad is the lowest (2.9 kcal mol À1 ) for the apo enzyme, because the charged residues (CysS À and HisH + ) can be stabilized by solvent molecules (see Fig. 3c). The free energy cost is increased up to 4.8 kcal mol À1 when the peptide substrate is present in the active site. In this case the hydroxyl group of Ser-P1 0 can contribute to stabilize the negative charge on Cys145, but the accessibility of water molecules to the ion pair dyad is signicantly reduced when compared to the apo form. In the case of the N3 inhibitor the free energy cost of forming the IP is increased up to 10.7 kcal mol À1 , a value obtained as the average between the forward (11.1 kcal mol À1 ) and backward (10.3 kcal mol À1 ) proles. The ionized catalytic dyad is stabilized by a single water molecule that enters into the active site and is placed in between the inhibitor and His41, being hydrogen bonded to the two residues of the dyad (see Fig. 3b). It must be noticed that the barrier for the proton transfer back from His41 to Cys145 from the IP is very small, which suggest that this protonation state is not very stable and that could only appear as a transient species during the acylation process. 9 Recently, Warshel and coworkers, using an Empirical Valence Bond method, reported an identical value, 2.9 kcal mol À1 , for the formation of the IP in the apo enzyme and a similar increase when an a-ketoamide inhibitor is bound in the active site, 7.3 kcal mol À1 . 22 Altogether, these values indicate that desolvation of the active site upon ligand binding can destabilize the IP form. In fact, formation of the IP, where two charged residues are found at short distance, results in a large dipole moment (about 14 D at the B3LYP/6-31+G* level) that can be stabilized by solvent molecules. This result also suggests that binding of the ligand into the active site aer IP formation could have associated a large energy penalty due to the need to remove water molecules from the active site. Classical MD simulations of the noncovalent complex with the catalytic dyad in the IP form show a trend for the ligand to slowly depart from the active site aer several hundred nanoseconds (see Fig. S2 †), being the active site then occupied by water molecules. In the recent work of Moliner and coworkers, in which a combination of AM1 dynamics and M06-2X energies are used to describe the QM region and the Amber ff03 force eld for the MM region, the IP was found to be only 1.3 kcal mol À1 above the neutral dyad when N3 is present in the active site. 20 As discussed below, this value leads to an estimated activation energy for the covalent inhibition of 3CL pro by N3 which seems to be too small when compared to the reference values derived from the inhibition rate constant of closely related enzymes. 8 Because of the comparatively larger free energy cost of forming the IP from the noncovalent EI complex with the N3 inhibitor found in our study, we investigated rst the possibility of a reaction mechanism for the formation of the covalent E-I complex that does not involve a proton transfer from Cys145 to His41. In such a mechanism the sulydryl proton is directly transferred to the Ca atom of N3 while the Sg atom attacks on the Cb one (see Fig. S3 †). However, the activation free energy found for this mechanism (about 50 kcal mol À1 ) is too high and incompatible with the observed inhibition rates (see below).</p><p>We thus explored a reaction mechanism for the formation of the E-I acylenzyme from the IP form. This mechanism implies the proton transfer from the N3 atom of His41 to the Ca atom of the inhibitor, mediated by the water molecule placed in between, and the nucleophilic attach of the Sg atom of Cys145 to the Cb atom of N3 (see Fig. 4a). The results obtained for the MFEP corresponding to this mechanism at the B3LYPD3/6-31+G*/MM level are shown in Fig. 4b and c. According to the free energy prole the reaction proceeds via two Transition States (TS1 and TS2) separated by a shallow intermediate (see Fig. 4b). TS1 is the rate-limiting one with a free energy 10.6 kcal mol À1 higher than the IP, while the free energy difference corresponding to TS2 is 9.4 kcal mol À1 . The evolution of the CVs used to dene the multidimensional free energy surface (Fig. 4c) shows that TS1 is associated to the nucleophilic attack of the Sg atom to the Cb atom and the change of the bond between Ca and Cb atoms from double to single. The Sg-Cb distance at TS1 has been reduced from 3.3 to 2.33 Å, while the Cb-Ca distance has been slightly increased from 1.34 to 1.41 Å (see Fig. 4d). TS2 corresponds to the proton transfer from His41 to the neighbor water molecule and from this to the Ca atom, being the rst proton transfer more advanced than the second one (see Fig. 4e). At TS2 the Sg-Cb bond is signicantly shorter (1.91 Å) while the Cb-Ca distance has been elongated up to a value close to that of a single bond (1.50 Å). Note that in this mechanism the sequence of nucleophilic attack and proton transfer is just the reverse of that observed for the acylation mechanism of the peptide substrate, where the proton transfer to the N atom of the scissile bond precedes the nucleophilic attack on the carbonyl carbon atom. 9 Finally, the reaction product, where a proton has been transferred to the Ca atom, is shown in Fig. 4f. The Sg-Cb bond distance found at the E-I complex (1.85 Å) is close to the value found in the X-ray structure of the inhibited enzyme (1.77 Å). 4 The overlap between the QM/MM and X-ray structures is shown in Fig. S4. † In order to check the robustness of our mechanistic proposal, the string calculation was repeated using the M06-2X functional with the same basis set and D3 corrections. The resulting free energy prole was almost identical to the B3LYP one, both from the energetic and structural points of view: the geometries and energies of the transition states were very similar at both theoretical levels, as can be seen in Fig. S5. † This result conrms the adequacy of the B3LYP functional for the present Michael addition, in spite of the reported limitations of this functional to correctly describe some enolate or carbanion intermediates. 23,24 Note that these species are not strictly found in the proposed mechanism because of the proton transfer from His41 to the substrate. It must be also stressed that at both theoretical levels, the string converges to a mechanism evolving from the IP to the covalent product, conrming that, in agreement with our previous work, 10 the IP is a metastable species from which the most favorable mechanism may proceed.</p><p>A complete representation of the free energy path from the noncovalent complex (EI) to the covalent one (E-I) is provided in Fig. 5. According to this free energy prole, resulting from the combination of those presented in Fig. 3 and 4, the transformation from the noncovalent EI complex to the covalent one (E-I) is an exothermic process. The free energy difference between E-I and IP (Fig. 4b) is À25.7 kcal mol À1 , while the free energy difference between IP and EI is 10.7 kcal mol À1 (Fig. 3a). Combining these two values, our simulations predict that the covalently bonded E-I complex is À15.0 kcal mol À1 more stable than the noncovalent EI complex, which agrees with the observed irreversibility of the inhibition process of 3CL pro by N3. 4 Regarding the inactivation rate (k 3 in eqn (1)), our simulations predict that the associated activation free energy results also from the sum of two contributions: the free energy cost of creating the IP form from EI (10.7 kcal mol À1 , Fig. 3a) plus the activation free energy of TS1 relative to IP (10.6 kcal mol À1 , Fig. 4b). This gives in a total activation free energy of 21.3 kcal mol À1 , which according to Transition State Theory (see eqn (S1) in SI †) corresponds to a rate constant of 1.9 Â 10 À3 s À1 at 300 K. Unfortunately, only the second-order rate constant (k 3 / K I ) and not the inactivation rate constant of SARS-CoV-2 3CL pro by N3 has been estimated. 4 However, the inactivation rate constant by N3 (k 3 ) was determined for the highly similar ortholog protease of SARS-CoV. 8 In this case the activation free energy derived from the rate constant measured at 303 K (3.1 Â 10 À3 s À1 ) is 21.2 kcal mol À1 . Comparison between SARS-CoV and SARS-CoV-2 main proteases seems appropriate considering that they present identical active sites, the same substrate specicity and very similar reaction rate constants for the hydrolysis of peptides. 25 The order of magnitude predicted for the k 3 rate constant seems also correct when compared to the values determined for the main proteases of other coronaviruses. 8 Even if the rate constant for SARS-CoV-2 inhibition would be one order of magnitude faster, which could account for the rapid inhibition reported experimentally, 4 the resulting activation free energy for the SARS-CoV-2 enzyme would only be 1.4 kcal mol À1 smaller than the reported value for the SARS-CoV one. Lastly, our prediction for the rate constant is compatible with the experimental estimation of k 3 /K I for the SARS-CoV-2 enzyme (11 300 M À1 s À1 ). 4 Combination of our predicted k 3 with the reported estimation for k 3 /K I gives a K I of $0.2 mM, a value similar to those determined for the inhibition of other coronaviruses's proteases with N3. 8 The excellent agreement between the experimental values and our theoretical estimation strongly supports our mechanistic proposal for the inhibition of SARS-CoV-2 3CL pro by a Michael acceptor.</p><p>The QM/MM study of Moliner and coworkers found an activation free energy of 11.2 kcal mol À1 and a reaction free energy of À17.9 kcal mol À1 . 20 While the latter value is close to our ndings (À15.0 kcal mol À1 ), the former departs signicantly from our estimation (21.3 kcal mol À1 ). Their activation free energy provides a rate constant of $10 4 s À1 , signicantly larger than the aforementioned value measured for the homologous SARS-CoV protease. In their mechanistic proposal the proton transfer from His41 to the inhibitor is direct and not water-mediated. However, this mechanistic difference does not explain the gap between their and ours calculated activation free energies. In their simulations the rate-limiting TS corresponds, as in our case, to the Sg-Cb bond formation and the free energy difference with the IP is 9.9 kcal mol À1 , very close to our value of 10.6 kcal mol À1 (see Fig. 4b). The main difference between our results and those reported by Moliner and coworkers is found in the rst part of the process, the free energy cost of forming the IP from the EI complex, 10.7 and 1.3 kcal mol À1 , respectively. Differences between the two works may arise from the different QM levels of theory, MM forceelds or to the sampling of different enzymatic congurations (their exploration of the mechanism started from the E-I complex while we started from the noncovalent EI complex); factors that could affect the relative stability of the neutral and ionic forms of the catalytic dyad. According to our previous discussion, we think that our simulations provides a general picture (see Fig. 5) in better agreement with current experimental results.</p><p>Roughly speaking, in our simulations the two steps presented in Fig. 3 and 4, IP formation and Michael addition, contribute similarly to the activation free energy of the inhibition process (about 10 kcal mol À1 each of them). This suggests that the kinetic properties of inhibitors can be improved also by stabilizing the ligand-bound ion pair state. It has been already suggested for other related 3CL proteases (from MERS and SARS-CoV) that stabilization of a charged catalytic dyad could promote catalysis. 26 For the SARS-CoV-2 protease it has been shown that inhibitors can shi the protonation state of some residues, not only the catalytic dyad but also other residues found in the vicinity of the active site. 27 In principle, a possible strategy is the introduction of chemical groups in the inhibitor structure that imitate the role played by Ser-P1 0 in the natural substrate. The hydroxyl group of this residue can make contacts with the catalytic dyad that, together with the presence of solvent molecules, can contribute to stabilize the IP. 9 Interestingly, the position of the water molecule in the rate limiting TS structure found in this work (TS1, see Fig. 4d) could be useful to assist in the design of inhibitors that favor this stabilization process. In this sense, a recently reported potent inhibitor of SARS-CoV-2 3CL pro (PF-00835231) 28 presents a hydroxyl group that matches the position of the water molecule in TS1 (see Fig. S6 † for an overlap of the X-ray structure of the enzyme inhibited by PF-00835231and TS1). This observation illustrates the insights offered by mechanistic studies for the design of new inhibitors.</p><!><p>We have here presented the results of microscopic simulations of SARS-CoV-2 3CL protease inhibition by N3, a Michael acceptor. Classical and hybrid QM/MM simulations were performed to investigate the noncovalent and the covalent enzyme-inhibitor complexes, respectively.</p><p>Molecular dynamics simulations of the noncovalent EI complex show that the inhibitor mimics the interactions established by the P1-P5 residues of the natural substrate. Our analysis also shows that an interesting strategy to improve a potential inhibitor based in N3 could be the introduction of chemical changes in the benzyl ester group in such a way that could restore the interactions that the P2 0 group of the peptide substrate establishes with Thr25, Thr26 and Gly143. This change could increase the affinity between the inhibitor and the protein, reducing the dissociation constant K I .</p><p>Regarding the formation process of the covalent E-I complex, our simulations show that the inhibition mechanism of SARS-CoV-2 3CL pro by a Michael acceptor involves two steps aer binding the inhibitor: (i) the activation of the catalytic dyad by means of the formation of an ion pair and (ii) a Michael addition process where Cys145 attacks to the Cb atom of the Michael acceptor and a proton is transferred, water mediated, from His41 to the Ca atom of the inhibitor. The contribution of each of the two steps to the activation free energy of the inhibition process is roughly the same (about 10 kcal mol À1 ) and thus inhibition kinetics can be favored by reducing either of the two contributions. The free energy cost to form the IP is substantially smaller in the enzymatic complex with the peptide substrate than when N3 is present in the active site (about 5.9 kcal mol À1 , according to Fig. 3a) and the activation free energy for the acylation of a peptide substrate is also signicantly smaller than for the N3 inhibitor (by about the same quantity reected in Fig. 3a). 9 This clearly suggests that, in order to improve the kinetic behavior of newly designed inhibitors (increasing k 3 ), attention must be paid to the formation of the IP. In this sense, the structures found along our reaction path, an in particular the rate-limiting transition state, could be useful to guide that design.</p><!><p>The are no conicts to declare.</p>

Royal Society of Chemistry (RSC)

Cerebral Blood Flow in Normal Aging Adults: Cardiovascular Determinants, Clinical Implications, and Aerobic Fitness

Senescence is a leading cause of mortality, disability, and non-communicable chronic diseases in older adults. Mounting evidence indicates that the presence of cardiovascular disease and risk factors elevates the incidence of both vascular cognitive impairment and Alzheimer\xe2\x80\x99s disease (AD). Age-related declines in cardiovascular function may impair cerebral blood flow (CBF) regulation, leading to the disruption of neuronal micro-environmental homeostasis. The brain is the most metabolically active organ with limited intracellular energy storage and critically depends on CBF to sustain neuronal metabolism. In patients with AD, cerebral hypoperfusion, increased CBF pulsatility, and impaired blood pressure control during orthostatic stress have been reported, indicating exaggerated, age-related decline in both cerebro- and cardiovascular function. Currently, AD lacks effective treatments; therefore, development of preventive strategy is urgently needed. Regular aerobic exercise improves cardiovascular function, which in turn may lead to a better CBF regulation, thus reducing the dementia risk. In this review, we discuss the effects of aging on cardiovascular regulation of CBF and provide new insights into the vascular mechanisms of cognitive impairment and potential effects of aerobic exercise training on CBF regulation.

cerebral_blood_flow_in_normal_aging_adults:_cardiovascular_determinants,_clinical_implications,_and_

Background<!>Main Determinants of CBF<!>Neurovascular coupling<!>Arterial BP<!>Carbon dioxide (CO2)<!>Cardiac output (CO)<!>Autonomic neural activity<!>Age and Steady-State CBF<!>Age and CBF Oscillation<!>Age and CBF Pulsatility<!>Age and CVMR<!>Cerebral perfusion<!>CA and cardiovagal baroreflex function<!>CBF pulsatility<!>CVMR<!>Methodological Considerations<!>Summary

<p>We are facing the unprecedented aging of population. According to a recent report from the United Nations, a global share of older adults 60 years and over are expected to more than double from 2013 to 2050. In addition, the older population itself is aging; persons aged 80 years and over (the "oldest old") will more than triple by 2050 (United_Nations 2013). Senescence is a leading cause of non-communicable chronic disease; therefore, population aging is expected to impose a substantial burden on our healthcare system.</p><p>Dementia represents a leading cause of death, disability, and loss of autonomy among older adults (Kochanek 2016). Alzheimer's disease (AD) is the dominant type of dementia (Querfurth & LaFerla 2010), and its incidence doubles every 5 years after age of 65 and afflicts ~50% of adults aged >85 years (Prince et al. 2013). With the population aging, patients with AD are expected to triple from 24.3 million in 2001 to 81.1 million in 2040 (Ferri et al. 2005), if no effective therapy or preventive measures are developed in near future.</p><p>The proposed etiology of AD has been centered on the amyloid hypothesis over the last 3 decades, which states that cerebral accumulations of amyloid-β and hyper-phosphorylated tau peptides lead to neuronal dysfunction and cognitive impairment (Hardy & Selkoe 2002). However, mounting evidence indicates multifactorial nature of AD and that cerebrovascular pathology coexists in most of AD patients (Viswanathan et al. 2009). In support of vascular contributions to AD, the presence of midlife cardiovascular risk factors has been shown to accelerate age-related cognitive decline and increase the risk of AD in later life (Whitmer et al. 2005).</p><p>The brain is limited by intracellular energy substrates to sustain neuronal metabolism and critically depends on the cardiovascular supply of cerebral blood flow (CBF). Therefore, age-related impairment of cardiovascular function may impair CBF regulation and disrupt neuronal homeostasis (de la Torre 2004). In contrast, pharmaceutical and non-pharmaceutical interventions that can ameliorate age-related cardiovascular dysfunction may improve CBF supply, thereby decreasing the risk of cognitive impairment. In this regard, previous studies have demonstrated the potential benefits of aerobic exercise training on cognitive function (Smith et al. 2010). In this brief review, we will discuss 1) the effects of aging on cardiovascular regulation of CBF and 2) the association between regular aerobic exercise and CBF regulation. This review summarizes the major findings from recent studies performed in humans and discusses their clinical implications with a particular focus on cognitive impairment.</p><!><p>The brain weighs only ~2% of the body mass while its metabolic rate accounts for ~20% of the whole body (Elia 1992). Despite a high metabolic rate, the brain contains only 3–4 umol/g of the intracellular glycogen, when compared for example to the liver that contains 200–400 umol/g (Oz et al. 2007). In addition, the rate of glycogen turnover is slow in the brain and only provides glucose under chronic hypoglycemia (Oz et al. 2009). Therefore, a stable supply of CBF is critical for normal brain function (Williams & Leggett 1989) which is regulated by both the local and systemic mechanisms.</p><!><p>Regional CBF is tightly coupled to neuronal metabolism which is heterogeneous in space and time. During neuronal activation, synaptic release of neurotransmitters (e.g. glutamate) leads to an elevation of regional CBF (functional hyperemia) through vasodilation (Attwell et al. 2010). With advent of neuroimaging technology, functional hyperemia can be assessed by the blood-oxygen-level-dependent (BOLD) signal using functional magnetic resonance imaging (MRI) (Moseley & Glover 1995). The BOLD signal is based on T2*-weighted signal which depends on local changes in the concentration of deoxygenated hemoglobin via neurovascular coupling (Ogawa et al. 1992). In a previous study, we measured brain resting-state BOLD signal while participants quietly laid on the scanner table (Zhu et al. 2015). We found that under resting conditions, regional BOLD spontaneous fluctuations at the very low frequency (<0.05 Hz) exhibit substantial overlap with the oscillations of systemic arterial blood pressure (BP) and global CBF measured from the middle cerebral artery (Figure 1). These observations suggest that regional dynamic changes in CBF may be influenced by the upper-stream changes in cardiovascular function.</p><!><p>Arterial BP is the determinant of cerebral perfusion pressure (CPP), which is a pressure gradient between arterial BP and intracranial pressure and represents the driving force for CBF. While intracranial pressure is largely influenced by body posture or hydrostatic gradient between the head and heart positions (Chapman et al. 1990), arterial BP is regulated by an integrated mechanisms of autonomic, humoral, and vascular factors.</p><p>Arterial BP fluctuates spontaneously at rest and is influenced by extrinsic stimuli (e.g. postural changes) (Zhang et al. 2000, Claassen et al. 2009). In response to dynamic changes in BP, cerebrovascular bed behaves as a "high-pass filter" system, which buffers the effects of BP fluctuations on CBF via cerebral autoregulation (CA) (Lassen 1964, Rickards & Tzeng 2014). The CA is more or less a frequency dependent phenomenon that operates more efficiently at very low frequency (<0.05Hz) (Giller 1990, Aaslid et al. 1989, Zhang et al. 1998, Panerai et al. 1999) and controlled by myogenic, neurogenic, and metabolic mechanisms (McHedlishvili 1980). Recently, CA has been shown to interact with cardiovagal baroreflex function in healthy young adults for maintaining CBF and brain homeostasis (Tzeng et al. 2010). The baroreceptor reflex controls short-term changes in BP through the distortion of barosensory arteries (input) and the modulation of autonomic neural activity of the heart and vascular system (output) (Guo et al. 1982).</p><p>Pulse pressure (PP), which represents another dynamic component of arterial BP, is directly correlated with the amplitude of CBF pulsatility (Tarumi et al. 2014). In cardiovascular system, elevated PP represents a hallmark of vascular aging and results from stiffening of the central large arteries (e.g. aorta, carotid artery) (Nichols 2005). PP is likely situated outside the operating frequency range of CA and its transmission is most likely determined by the impendent property of cerebrovascular bed (Windkessel effect) (Zhu et al. 2011). With high elasticity of large cerebral arterial wall, PP may be dampened by the expansion and recoiling of cerebral arteries during each cardiac cycle.</p><!><p>The partial pressure of carbon dioxide in the arterial blood (PaCO2) has potent effects on cerebral vasomotor tone. Elevated PaCO2 (hypercapnia) dilates cerebral arteries leading to increases in CBF whereas reduced PaCO2 (hypocapnia) decreases CBF via vasoconstriction. These CBF responses to changes in PaCO2, termed cerebral vasomotor reactivity (CVMR), are likely to represent a vital homeostatic function that regulates the brain pH level and affects respiratory drive via central chemoreceptors (Chesler 2003). Cerebral vasodilation during hypercapnia increases CBF, washes out the excess CO2 from the blood, and maintains the brain pH level. The mechanism underling CVMR is not fully understood, but it is likely to involve the release of multiple vasoactive agents such as prostaglandins (Barnes et al. 2012, Fan et al. 2010) and nitric oxide (Schmetterer et al. 1997). The reduced production of prostaglandins via indomethacin ingestion has been shown to decrease basal CBF and significantly attenuate CVMR during hypo- and hypercapnia (Barnes et al. 2012, Fan et al. 2010). The inhibition of nitric oxide synthase has also shown the attenuation of hypercapnic CVMR that was reversed by administration of nitric oxide donors (Schmetterer et al. 1997). Finally, elevations of CBF during hypercapnia may be facilitated by the dilation of upper-stream extracranial arteries via shear-stress mediated release of endothelium-derived vasodilatory agents (Hoiland et al. 2017).</p><!><p>The brain continuously receives ~15% of CO to meet the metabolic demand of neuronal activity. Although CO is a key determinant of arterial BP when coupled with total peripheral resistance, alterations in CO per se may influence CBF (Meng et al. 2015). In healthy adults, reduction of CO using lower body negative pressure and elevation of CO via albumin infusion demonstrated a linear correlation between changes in CO and CBF at rest and during exercise, independent of changes in arterial BP or PaCO2 (Ogoh et al. 2005). On the other hand, heart failure patients demonstrated a non-linear relationship between acute changes in CO and CBF while changing the posture from the supine to sitting (Fraser et al. 2015). Furthermore, heart failure patients with depressed CO showed lower CBF than normal control subjects, but heart transplantation restored their CBF to the similar level observed in the control subjects (Gruhn et al. 2001).</p><!><p>The autonomic neural activity is likely to have profound impact on dynamic CBF regulation as cerebral arteries are richly innervated by the adrenergic and cholinergic fibers (Edvinsson 1975). In healthy humans, complete autonomic blockade using trimethaphan impaired CA at the very low frequencies (Zhang et al. 2002a, Zhang et al. 2002b). Furthermore, recent studies demonstrated that sympathetic blockade using α-adrenergic antagonist (Hamner et al. 2010) and cholinergic blockade using muscarinic receptor antagonist (Hamner et al. 2012) impaired CA and enhanced the amplitude of CBF oscillations at the very low frequencies. Collectively, these findings provide strong evidence that autonomic nervous system contributes to dynamic CA in healthy adults.</p><!><p>The steady-state level of CBF progressively decreases in normal aging men and women while women tend to have higher levels of CBF than men (Lu et al. 2011). This age-related reduction of CBF may reflect decreased cerebral metabolic rate (Marchal et al. 1992) and cerebrovascular dysfunction (Zhu et al. 2011). Across the adult lifespan, age decreases cerebral metabolic rates for oxygen and glucose by ~5% per decade, and these reductions of metabolic rate are coupled to the concurrent decrease in CBF (Leenders et al. 1990, Petit-Taboue et al. 1998). Mechanistically, age may impair neuronal and glial mitochondrial metabolism. The in vivo MR spectroscopy study demonstrated that metabolic rates for neuronal tricarboxylic acid and glutamate-glutamine cycles are reduced in older adults (Boumezbeur et al. 2010). Aside from these aging effects, it has been shown that women have higher levels of cerebral metabolic rate for glucose (Willis et al. 2002) and oxygen (Lu et al. 2011) which may explain why women have higher levels of CBF than men.</p><p>Normal aging is associated with gradual increase in mean arterial pressure (Franklin et al. 1997). Mechanistically, heightened sympathetic neural activity and impaired peripheral vasodilatory function (e.g. endothelial dysfunction) are likely to increase total peripheral resistance and therefore mean arterial pressure in older adults (Hart et al. 2012). On the other hand, studies with direct intracranial pressure monitoring using an intra-parenchymal probe demonstrated a negative correlation between age and intracranial pressure in patients with head injury (Czosnyka et al. 2005). If this observation can be extrapolated to healthy aging adults, age may increase CPP due to the effects of both increased mean arterial pressure and decreased intracranial pressure. In the face of elevated CPP, cerebrovascular bed may undergo compensatory remodeling by increasing resistance in order to protect the delicate brain tissues from overperfusion.</p><p>Age-related reduction of CBF may also be related to concurrent changes in CO (Brandfonbrener et al. 1955). The widely accepted dogma suggests that ~15% of CO is distributed to the brain in healthy adults (Williams & Leggett 1989); however, it is not well understood whether this proportion can change with the alteration of CO and/or CBF in aging adults. Therefore, we studied healthy aging adults (20–80 years) who do not have a history of neurological, cardio- or cerebrovascular disease. CBF was measured from the bilateral internal carotid and vertebral arteries using phase-contrast MRI and CO was measured by echocardiography. We found that advancing age is associated with the decreasing proportion of CO distributed to the brain; however, CO was maintained and only CBF was decreased in older adults (Xing et al. 2016) (Figure 2). These findings suggest that age-related reduction of CBF may not be attributed to the reduction of CO.</p><p>With regard to clinical perspective, cerebral hypoperfusion may be linked to the pathological onset of AD. The data-driven analysis of CSF, neuroimaging, and plasma biomarkers showed that cerebral hypoperfusion, as measured by arterial spin labeling using MRI, is the earliest event of late-onset AD before the manifestation of traditional biomarker abnormalities (e.g. CSF amyloid, cerebral hypometabolism) (Iturria-Medina et al. 2016). These findings are also supported by animal studies which demonstrated that mild to moderate cerebral hypoperfusion impairs neuronal protein synthesis which can subsequently lead to learning and memory dysfunction; ischemia may further impair action potential generations which can increase cerebral glutamate concentrations and promote the accumulations of neuronal toxins such as amyloid-β proteins (Zlokovic 2011).</p><!><p>To understand the aging effects on cardio- and cerebrovascular variability, we measured beat-by-beat changes in heart rate, arterial BP, CBF at rest and during repeated sit-stand maneuvers in healthy adults (21–80 years) (Xing et al. 2017). The repeated sit-stand maneuvers were performed at 0.05 Hz (cycles of 10-second sit and 10-second stand) to augment BP variability at the auto-regulatory frequency. Changes in CBF were measured by transcranial Doppler (TCD) from the middle cerebral artery. To characterize cardio- and cerebrovascular hemodynamics, we used spectral and transfer function analysis that estimated gain, phase, and coherence of dynamic CA (mean arterial pressure → CBF velocity) and cardiovagal baroreflex (systolic BP → R-R interval). In short, transfer function gain reflects a magnitude relation (slope) between input and output signals while the phase represents their temporal association. The coherence function reflects a strength of their linear correlation and provides the validity of gain and phase estimation (Zhang et al. 1998).</p><p>This study demonstrated that under resting conditions, low frequency oscillations (0.07–0.20 Hz) of mean arterial pressure and CBF variability were reduced in older adults compared with young and middle-aged adults. In contrast, repeated sit-stand maneuvers (0.05 Hz) augmented their oscillations to a greater extent in older adults than in younger and middle-aged adults (Figure 3). With regard to dynamic CA, older adults showed the elevations of low and high frequency gain (0.07–0.35 Hz) under resting conditions, suggesting impaired CA. However, these differences were abolished during repeated sit-stand maneuvers. In both conditions, heart rate variability was substantially reduced and cardiovagal baroreflex sensitivity was significantly attenuated in older adults compared with young adults (Xing et al. 2017). We also observed that low frequency CA gain is inversely correlated with the baroreflex gain in young subjects, as reported from a previous study (Tzeng et al. 2010).</p><p>In summary, these findings collectively demonstrate the presence of age-related reductions in BP, CBF, and heart rate variability in the low frequency range and impaired cardiovagal baroreflex and dynamic CA in older adults at rest. Furthermore, augmented BP and CBF variability during repeated sit-stand maneuvers indicate diminished cardiovascular regulatory capability in older adults and increased hemodynamic stress on the cerebral circulation with advanced aging (Xing et al. 2017).</p><p>Clinically, these observations suggest that postural changes may cause transient cerebral hypoperfusion and increase the risk of falls and syncope in older adults. Also, chronic intermittent cerebral hypoperfusion or ischemia is associated with white matter lesions and cognitive decline (O'Sullivan et al. 2002). In AD patients, a higher prevalence of orthostatic hypotension has been reported when compared with cognitively normal adults, and is associated with worse performance on cognitive assessment (Mehrabian et al. 2010). Mechanistically, central arterial stiffening may represent a common mechanism underling short-term BP dysregulation and cognitive decline. The stiffening of barosensory arteries, such as the aorta and carotid artery, can blunt the sensitivity of baroreceptor function (Monahan et al. 2001a, Okada et al. 2012), which in turn may contribute to the dysregulation of arterial BP, CPP, thus CBF. Consistently, our recent study also demonstrated that aortic stiffening and blunted cardiovagal baroreflex sensitivity are associated with reductions of brain neuronal fiber integrity and executive function performance in older adults (Tarumi et al. 2015). Nevertheless, the potential causal effect of arterial stiffening or blunted baroreflex sensitivity on cognitive impairment requires further research, and it should also be kept in mind that brain neurodegenerative disease per se may impair the baroreflex function.</p><!><p>Advancing age progressively stiffens the proximal aorta and central large elastic arteries (e.g. carotid artery) via increasing the wall contents of collagen relative to elastin (Zieman et al. 2005). The aortic stiffening elevates left ventricular afterload, as accompanied by an earlier return of arterial pressure wave reflection (Nichols 2005). Consequently, systolic BP and PP progressively increase during adult lifespan while diastolic BP gradually decreases after middle age (Franklin et al. 1997).</p><p>Elevated PP is a strong risk factor for mortality and cerebrovascular disease (Staessen et al. 2000); however, existing data are limited as to the effect of PP on CBF in normal aging adults. Therefore, we studied healthy subjects (22–80 years) who were rigorously screened for neurological and vascular disease as well as cardiovascular risk factors, including hypertension, obesity, diabetes, and smoking. To assess steady-state and pulsatile CBF, we used phase-contrast MRI and TCD respectively while measuring aortic stiffness, carotid PP, and arterial pressure wave reflection. In addition, white matter hyperintensity volume, which reflects the severity of cerebral small vessel disease (Benjamin et al. 2016), was measured by T2-weighted fluid-attenuation-inversion-recovery imaging (Tarumi et al. 2014).</p><p>This study showed several key findings. First, advancing age is associated with an accelerated increase in CBF pulsatility after midlife (Figure 4) while steady-state CBF linearly decreases across the adult lifespan. We also observed that diastolic CBF is lower but steady-state CBF is higher in women than in men of the similar age. Second, the age- and sex-related differences in CBF pulsatility are independently associated with carotid PP. Third, higher CBF pulsatility is correlated with the greater volume of white matter hyperintensities in older adults. Collectively, these findings demonstrated a close link between cardio- and cerebrovascular hemodynamics in healthy aging adults and suggested potential clinical implications to structural brain damage (Tarumi et al. 2014).</p><p>In response to elevated PP, cerebrovascular bed may undergo a compensatory remodeling to protect the underling brain tissues from hemodynamic insults. To study this hypothesis, we tested a group of young (28±4 years) and older (70±6 years) healthy adults by simultaneously recording carotid PP and CBF velocity from the middle cerebral artery using TCD. Transfer function analysis was used to quantify their magnitude (modulus) and temporal (phase) relation, which represent the cerebrovascular impedance. This study demonstrated that older adults have an increased modulus of cerebrovascular impedance compared with younger adults while the phase being similar between the groups (Figure 5). Furthermore, the elevated impedance modulus was correlated with the reduced systolic and diastolic CBF velocity, suggesting that the elevation of cerebrovascular impedance may attenuate the transmission of PP into the cerebral microcirculation (Zhu et al. 2011).</p><p>Clinically, aortic stiffening and augmentations of carotid PP and flow pulsatility have been shown to correlate with cerebral small vessel disease (Mitchell et al. 2011). Consistent with our observations, CBF pulsatility measured from the middle cerebral artery using TCD was positively correlated with the volume of brain whiter matter hyperintensities (Webb et al. 2012). Moreover, CBF pulsatility measured from the large intracranial arteries was elevated in AD patients and exhibited the high sensitivity and specificity for detecting AD patients compared with non-demented control subjects (Roher et al. 2011).</p><!><p>Previous research of the aging effects on CVMR generated inconsistent findings which may be explained by the lack of a standardized protocol for CVMR assessment. We recently studied young control and older healthy participants using 2 different CVMR protocols: 1) TCD measurement during hypoventilation and rebreathing (Zhu et al. 2013) and 2) BOLD measurement during steady-state hypercapnia (Thomas et al. 2013). In the first study, participants hyperventilated to induce a brief period of hypocapnia. Following the recovery of baseline hemodynamics, modified rebreathing method was used to induce a progressive increase in PaCO2. During the entire protocol, we measured breath-by-breath changes in end-tidal CO2 and beat-by-beat changes in arterial BP and CBF velocity.</p><p>This study demonstrated several key findings. First, older participants had lower CBF velocity and higher cerebrovascular resistance index at rest than young control participants. Second, compared with young control participants, hypocapnic CVMR (vasoconstriction) was significantly attenuated whereas hypercapnic CVMR (vasodilation) was elevated in older participants. Third, hypocapnic CVMR was inversely correlated with hypercapnic CVMR across all participants. Collectively, these observations suggest that advancing age is associated with increased cerebral vasoconstrictor tone at rest which in turn limits the hypocapnic vasoconstrictor capacity while increasing the relative hypercapnic vasodilator capacity.</p><p>In the second study, we tested a similar group of young and older participants using functional MRI during steady-state hypercapnia (Thomas et al. 2013). In this study, alternating blocks of room air (1 min) and hypercapnia (1 min) was inhaled by each participant while BOLD images were acquired continuously for 7 minutes. Hypercapnia was induced by having participants to inhale 5% CO2 balanced with 21% oxygen and 74% nitrogen. During the entire protocol, end-tidal CO2, arterial blood oxygen saturation, heart rate, and breathing rate were monitored. In contrast to the former study, we found no group difference in hypercapnic CVMR between young and older participants at both global and regional levels (Thomas et al. 2013).</p><p>The inconsistent results from these studies may be explained by different protocols used for CVMR assessment. First, neither BOLD nor TCD measures CBF per se (please read the Methodological Considerations section). Second, the steady-state and rebreathing methods may elicit different autonomic and cardiovascular responses. The rebreathing technique has been shown to elicit greater chemoreflex sensitivity and sympathetic neural response compared with the steady-state method (Mohan et al. 1999, Shoemaker et al. 2002). This greater level of sympathetic neural response may affect the vasodilatory effect of CO2 on cerebral arteries and may result in different CVMR between the protocols (Claassen et al. 2007). Third, the magnitude of hypercapnia induced by steady-state (~10 mmHg) and rebreathing (~16 mmHg) methods was different and the relationship between PaCO2 and CBF may not be linear within these ranges.</p><p>In clinical setting, CVMR assessment may help identify individuals who have elevated risks for cerebrovascular disease. According to a meta-analysis of patients with carotid arterial stenosis or occlusion, attenuation of hypercapnic CVMR was associated with the increased incidence of future stroke or transient ischemic attack (Gupta et al. 2012). Mechanistically, carotid arterial stenosis or occlusion may decrease CPP distal to the lesions and exhaust the autoregulatory vasodilatory reserve; therefore, additional stimuli such as hypercapnia may not further vasodilate the cerebral arteries (Gupta et al. 2012). In addition to stroke, attenuated hypercapnic CVMR has been shown to correlate with cognitive decline in AD patients (Silvestrini et al. 2006), microstructural damage of the cerebral white matter (Sam et al. 2016), and increased mortality (Portegies et al. 2014).</p><!><p>Regular aerobic exercise may attenuate age-related reductions of CBF. To test this hypothesis, we used arterial spin labeling technique to measure cerebral perfusion in Masters Athletes (MA) who have participated in lifelong aerobic exercise training and regularly competed in endurance events. Our analysis showed that compared with young control subjects, MAs and sedentary elderly adults have similar reductions of global CBF. However, when examining regional perfusion normalized against global CBF, MAs had higher perfusion in the posterior cingulate and precuneus than young control and sedentary elderly participants (Figure 6) (Thomas et al. 2013). Consistent with these observations, another study of middle-aged MAs showed the higher occipitoparietal perfusion compared with age-matched sedentary subjects (Tarumi et al. 2013).</p><p>Furthermore, short-term aerobic exercise training may alter regional cerebral perfusion. An intervention study of 3 months aerobic exercise training reported the elevation of anterior cingulate perfusion in the previously sedentary older adults (Chapman et al. 2013). On the other hand, a 10-day cessation of aerobic exercise training in MAs decreased regional cerebral perfusion, including the hippocampus (Alfini et al. 2016). Taken together, these findings suggest that aerobic exercise training may increase regional cerebral perfusion, particularly at the area of default mode network that is known to be affected by the process of normal aging and AD (Buckner et al. 2008).</p><!><p>Regular aerobic exercise may not alter CA in older adults. We studied endurance MAs by measuring their beat-by-beat changes in CBF velocity from the middle cerebral artery, arterial BP, and heart rate at rest and during repeated sit-stand maneuvers. The transfer function analysis of dynamic CA showed that compared with age-matched sedentary subjects, MAs have an attenuation of the very low frequency gain under resting conditions; however, repeated sit-stand maneuvers abolished this group difference (Aengevaeren et al. 2013). Consistent with our observations, dynamic CA measured by thigh cuff technique also showed no difference between young athletes and age-matched sedentary adults (Lind-Holst et al. 2011, Ichikawa et al. 2013).</p><p>On the other hand, aerobic exercise training has been shown to increase cardiovagal baroreflex sensitivity in older adults (Aengevaeren et al. 2013, Monahan et al. 2001b, Deley et al. 2009). This may improve arterial BP regulation and decrease the contributions of CA to buffering CBF fluctuations. Mechanistically, exercise-related improvement of the baroreflex sensitivity may be linked to elevated tonic vagal activity that can decrease heart rate (Shi et al. 1995) but increase R-R interval variability (Raczak et al. 2006). Aerobic exercise training can also increase stroke volume and the elasticity of barosensory arteries, which together may improve the transduction of mechanical stimuli to the baroreceptors (Monahan et al. 2001b). Finally, endurance training may increase cardiac cholinergic responsiveness (Poller et al. 1997).</p><!><p>The question of whether aerobic exercise training alters CBF pulsatility in older adults currently remains unknown. To our knowledge, there is only one single study that addressed this question in healthy young adults (Tomoto et al. 2015). In this study, collegiate tennis players underwent 16 weeks of the combined moderate-intensity continuous aerobic exercise training and high-intensity interval training. In addition to CBF pulsatility assessment using TCD, they measured carotid arterial stiffness and left ventricular systolic function. After exercise training, CBF pulsatility did not change; however, the reductions of carotid artery stiffness were individually associated with the attenuations of CBF pulsatility. Therefore, if these findings can be extrapolated to older adults, aerobic exercise training that can reduce carotid arterial stiffness may decrease CBF pulsatility in older adults.</p><!><p>The previous studies investigating the effect of aerobic exercise training on CVMR showed mixed results. Using TCD during hyperventilation and modified rebreathing, we observed that both hypo- and hypercapnic CVMRs were similar between endurance MAs and age-matched sedentary older adults (Zhu et al. 2013). In a similar group of study participants, we also measured CVMR using functional MRI during steady-state hypercapnia and observed that MAs have lower CVMR than the sedentary older adults (Thomas et al. 2013). On the other hand, using TCD during steady-state hypercapnia, aerobic exercise training studies showed improvements in CVMR in healthy young and older adults (Murrell et al. 2013) as well as in stoke survivors (Ivey et al. 2011). Taken together, these studies made inconsistent observations over the effect of aerobic exercise training on CVMR and suggest that methods used to measure CBF (TCD, BOLD) or to elicit CO2 stimulus (rebreathing, steady-state hypercapnia) may alter CVMR quantifications.</p><!><p>CBF is highly variable in space and time, and currently there is no single method that can measure CBF at the sufficiently high level of spatial and temporal resolutions. For example, numerous studies investigating CBF regulation have used TCD that is non-invasive, relatively inexpensive, and readily accessible. TCD has a major strength of recording CBF velocity at high temporal resolution from the large intracranial arteries, but it can only assess changes in CBF under the assumption of constant insonated arterial diameter. Although this assumption may hold true under the relatively mild stimulus (Giller et al. 1993), recent studies using high-resolution MRI revealed that moderate to severe hypo- and hypercapnia can alter the diameter of middle cerebral artery which may have significant impact on CBF quantification (Verbree et al. 2014).</p><p>Recently, color-coded duplex ultrasound imaging of the internal carotid and vertebral arteries is gaining popularity in CBF research. This technique is strengthened by the ability to measure global and regional CBF in the anterior and posterior circulations at high temporal resolution (Liu et al. 2016). This imaging method is easily accessible at bedside and can be performed on patients who have contraindications to MRI (e.g. metal implants or claustrophobia). On the other hand, technical aspect of this method can sometimes be challenging depending on the individual differences in vascular anatomy. For example, the bifurcation site of the common carotid artery may be located close to the jaw and make the placement of an ultrasound probe difficult to accurately measure the diameter and blood flow velocity in the internal carotid artery. Also, vertebral arteries are relatively small compared with the carotid arteries and located deep in the neck. In such cases, an alternative method may be phase-contrast MRI combined with the time-of-flight angiography. In our previous study, we found that color-coded duplex ultrasonography and phase-contrast MRI have the similar estimations of volumetric CBF measured from the internal and vertebral arteries (Khan et al. 2017).</p><p>Neuroimaging techniques such as positron emission tomography (PET), single-photon emission computed tomography (SPECT), and arterial spin labeling have higher spatial resolution of CBF measurement than TCD, color-coded duplex ultrasonography, or phase-contrast MRI; however, these methods are limited by lower temporal resolution. In addition, these imaging modalities are more expensive and require the infusion of radioactive tracer to blood circulation (PET and SPECT). To avoid tracer infusions, arterial spin labeling technique using MRI allows non-invasive assessment of regional cerebral perfusion, but this method is also limited by a low signal-to-noise ratio and requires multiple image acquisitions for averaging to obtain reliable results. In addition, BOLD signal acquired by functional MRI has been used to assess regional CVMR; however, this technology depends on changes in the local concentration of deoxyhemoglobin which is influenced by altered neuronal activity and/or changes in the total concentration of hemoglobin by altered CBF (Halani et al. 2015). Thus, a multi-modal approach may complement each method of CBF measurement and provides a better understanding of cerebral hemodynamics.</p><!><p>Age profoundly alters CBF and its regulatory mechanisms (Figure 7). Specifically, steady-state CBF progressively decreases across the adult lifespan while CBF pulsatility increases after midlife (Tarumi et al. 2014). The fluctuations of CBF during postural changes are also augmented in older adults compared with younger adults (Xing et al. 2016). These age-related changes in CBF are, at least in part, explained by the concurrent alterations of arterial BP, primarily the elevations of systolic BP and PP. In addition, although CA may not be affected by age, older adults have marked reduction of cardiovagal baroreflex sensitivity which likely augments BP and CBF fluctuations during postural changes. In patients with cerebrovascular disease and AD, these age-related alterations of CBF are exaggerated and may contribute to the disease onset and progression. In particular, cerebral hypoperfusion (Iturria-Medina et al. 2016), augmented CBF pulsatility (Roher et al. 2011), and orthostatic hypotension (Mehrabian et al. 2010) all have been reported in AD patients compared with aged-matched cognitively normal adults. In contrast, regular aerobic exercise may attenuate the age-related reduction of CBF, especially in the default mode neural network. Taken together, gaining the knowledge of normal age-related changes in CBF may help us identify the individuals at risk for cerebrovascular disease and dementia, including AD. We now understand that dementia is a multifactorial disease. Therefore, understanding of vascular mechanisms, identification of vascular biomarkers, and development of vascular-based interventions may pave promising avenues to maintain brain health in older adults.</p>

PubMed Author Manuscript

Time-resolved analysis of photoluminescence at a single wavelength for ratiometric and multiplex biosensing and bioimaging

Simultaneous analysis of luminescence signals of multiple probes can improve the accuracy and efficiency of biosensing and bioimaging. Analysis of multiple signals at different wavelengths usually suffers from spectral overlap, possible energy transfer, and difference in detection efficiency. Herein, we reported a polymeric luminescent probe, which was composed of a phenothiazine-based fluorescent compound and a phosphorescent iridium(III) complex. Both luminophores emitted at around 600 nm but their luminescence lifetimes are 160 times different, allowing time-resolved independent analysis. As the fluorescence was enhanced in response to oxidation by hypochlorite and the phosphorescence was sensitive toward oxygen quenching, a four-dimensional relationship between luminescence intensity, fluorescence/phosphorescence ratio, hypochlorite concentration, and oxygen content was established.In cellular imaging, time-resolved photoluminescence imaging microscopy clearly showed the independent fluorescence response toward hypochlorite and phosphorescence response toward oxygen in separated time intervals. This work opens up a new idea for the development of multiplex biosensing and bioimaging.

time-resolved_analysis_of_photoluminescence_at_a_single_wavelength_for_ratiometric_and_multiplex_bio

Introduction<!>Results and discussion<!>Conclusions

<p>Photoluminescence bioimaging tracks emissive probes in biological environments to create visual images by analyzing and processing luminescence signals, giving structural and functional information of biological molecules, [1][2][3][4][5][6][7] organelles, [8][9][10][11] living cells, [12][13][14][15][16] tissues, [17][18][19][20] and organs. 21,22 Laser scanning confocal microscopy has been widely used for analyzing luminescence intensity of a probe to construct images, which simply and directly reect the spatial distribution and local concentration of the probe at the subcellular level. 23,24 Once the luminescence intensity of the probe is correlated with its specic interaction with functional biomarkers, luminescence imaging can locate and quantify the marker while continuously monitoring the process of life activities involved in the cell. [25][26][27][28][29] In order to improve the sensing and imaging accuracy and efficiency, it is usually necessary to employ multiple luminophores and simultaneously analyze two or more luminescence signals, thereby allowing one signal to be used as an internal standard, [30][31][32][33] or multiple signals to respond toward different targets. [34][35][36] In practical applications, the luminophores simultaneously used are designed to have different emission wavelengths, so that they can be readily distinguishable by placing optical lters before the detector. 37,38 However, there are some limitations seriously interfering with the sensing and imaging during wavelength-resolved analysis. Firstly, many organic uorophores and phosphorescent transition metal complexes exhibited broad emission spectra. It is difficult to completely avoid spectral overlap, so it is usually necessary to sacrice the brightness to avoid signal leakage and loss of sensitivity. Secondly, as the luminophores are separated in wavelength, energy transfer from the higher energy luminophore to the lower energy one is possible, resulting in undetectable highenergy signals. 39,40 Thirdly, the refractive index, reectance, and absorbance of photons of different wavelengths in biological samples are different. Hence, the loss of photons when passing through biological samples and the number of photons reaching the detector strongly depends on the wavelength. 41 As shown in Fig. 1a, when dealing with wavelength-resolved two luminescence signals in cellular imaging, increasing intralipid in the culture medium caused reduced detection efficiency of photons at both 500 and 600 nm. The intensity ratio of the two wavelengths is also inuenced by the amount of the intralipid due to their different transmittance, indicating that luminophores emitting at different wavelengths are not suitable for ratiometric imaging.</p><p>To address these concerns, in this work, we proposed to analyze multiple photoluminescence signals at the same wavelength via a time-resolved technique. A new dual-emissive polymeric probe com-posed of two luminophores was designed and synthesized. The two luminophores emitted at almost the same wavelength to ensure limited energy transfer and consistent photon-detection efficiency, but exhibited different luminescence lifetimes, enabling independent analysis of each spectrum via a time-resolved technique 42 (Fig. 1b). Although the amount of the intralipid in the culture medium caused photon loss during imaging, the ratio of short-lived luminescence signals over long-lived ones remained unchanged. The two luminophores were designed to respond toward hypochlorite and oxygen, respectively. Hypochlorite plays an important role in the destruction of pathogens in the immune system 43 and unregulated cellular production of hypochlorite may be associated with various diseases. 44 Cellular oxygen contents are kept in a certain range and hypoxia is an important feature of many diseases including tumors. 45 Therefore, sensing and imaging cellular hypochlorite and oxygen are of great importance and have attracted much attention. 31,38,46,47 Under constant oxygen conditions, the polymeric probe was used for time-resolved ratiometric sensing and imaging of hypochlorite by using the oxygensensitive luminescence as an internal standard. When the oxygen content changed, the probe was used for simultaneously sensing and imaging of hypochlorite and oxygen via time-resolved luminescence analysis.</p><!><p>The dual-emissive polymeric probe P1 was designed by incorporation of a phenothiazine-based uorescent compound 1, a phosphorescent iridium(III) complex 2, and polyethylene glycol (PEG) into the side chains on the polyethylene backbone (Fig. 2a and S1 †). Compound 1 was weakly emissive owing to the photoinduced electron transfer (PET) from the electron-rich sulfur atom to the electron-decient pyridinium. 48 In response to hypochlorite, which oxidizes the thioether to sulfoxide, compound 1 exhibited signicant uorescence enhancement at around 600 nm with a lifetime of 2.5 ns (Fig. S2 †). Complex 2 emitted at the same wavelength with a much longer lifetime of about 0.4 ms owing to its phosphorescence nature. Meanwhile, the phosphorescence of complex 2 was readily quenched by molecular oxygen via triplet-triplet energy transfer (Fig. S2 †). Both compound 1 and complex 2 were excitable at 405 nm, which is the most widely used excitation laser source in confocal microscopy. PEG was used to improve the water solubility and biocompatibility. Polymeric luminescent probes P2 and P3 (Fig. 2a), where complex 2 and compound 1 was absent, respectively, were also synthesized for comparison studies. All the small molecular intermediates and monomers were characterized via 1 H and 13 C nuclear magnetic resonance (NMR), matrix assisted laser desorption ionization time-of-ight (MALDI-TOF) mass spectrometry (MS) and the polymers were synthesized via radical polymerization and characterized via gel permeation chromatography (GPC). The contents of compound 1 and complex 2 in P1 were about 6% and 1% (molar ratio) calculated according to the absorption spectra. The averaged molecular weight (M w ) of P1 was about 133 000 g mol À1 with a polydispersity index (PDI) of 1.13. P1 was completely soluble in aqueous solution and suitable for bioimaging.</p><p>The luminescence response of P1 toward hypochlorite was investigated in phosphate buffer saline (PBS) under ambient conditions. Upon photoexcitation at 405 nm, P1 showed a mixed emission spectrum at 500-700 nm which was composed of the uorescence of compound 1 and the phosphorescence of complex 2. Addition of hypochlorite led to luminescence enhancement of P1 (Fig. 2b) because of oxidation-induced uorescence turn-on of compound 1. Similar luminescence enhancement was also observed in the polymeric probe P2 but not in P3 in which compound 1 was absent (Fig. S3 †). The remarkable luminescence response was specic toward hypochlorite in preference to other reactive oxygen species and biothiols (Fig. S4 †). Luminescence lifetime analysis revealed biexponential decay of P1 with lifetimes of about 2.7 and 500.0 ns corresponding to compound 1 and complex 2, respectively (Fig. 2c). The proportion of short-lived uorescence (I f ) at 600 nm in the spectrum increased from 11.7% to 78.0% upon addition of hypochlorite (Fig. 2d). The proportion of longlived phosphorescence (I p ) decreased from 88.3% to 22.0% when the uorescence was turned on by hypochlorite. According to these ratios, the uorescence and phosphorescence were isolated from the total luminescence via computational calculation. The isolated spectra showed that the uorescence was signicantly enhanced while the phosphorescence was insensitive to hypochlorite (Fig. 2e). Using the phosphorescence signal as an internal standard, the intensity ratio (I f /I p ) of uorescence over phosphorescence increased from 0.13 to 3.5 (about 27 fold) when the concentration of hypochlorite was increased from 0 to 10 mM. Further addition of hypochlorite did not change the total luminescence spectrum as well as the intensity ratio.</p><p>The luminescence response of P1 toward oxygen was recorded in PBS in the absence of hypochlorite under an O 2 /N 2 mixed atmosphere with different O 2 contents. When increasing and reducing the O 2 content, the emission at about 600 nm was quenched and enhanced, respectively (Fig. 2f), owing to the O 2induced phosphorescence quenching of complex 2. Luminescence lifetime analysis showed that while the uorescence lifetime remained unchanged, the phosphorescence lifetime was shortened in pure O 2 and elongated in pure N 2 . Meanwhile, the phosphorescence proportion (88.3% in air) in the total emission spectra was reduced to 84.5% and increased to 93.0%, respectively (Fig. 2g). Similar response toward oxygen was also observed for P3 but not for hypochlorite-oxidized P2 in which the phosphorescent iridium(III) complex was absent (Fig. S5 †).</p><p>To use P1 for simultaneous detection of hypochlorite and oxygen content, a functional relationship between the luminescence of P1 and the two variables was established. According to the luminescence titration curves of P1 upon addition of hypochlorite, the uorescence intensity showed a linear relationship with the concentration of hypochlorite in the range of 0-10 mM (Fig. 3a). As the phosphorescence was insensitive to hypochlorite, the I f /I p ratio also exhibited a linear relationship with the concentration of hypochlorite (Fig. 3b). In the absence of hypochlorite, the phosphorescence response toward oxygen followed Stern-Volmer equation 49</p><p>), giving a hyperbola relationship between the phosphorescence intensity and the oxygen content</p><p>), where I p0 and I p are phosphorescence intensities in the absence and presence of O 2 , respectively, and K SV is the Stern-Volmer constant (Fig. 3c). During simultaneous detection of hypochlorite and oxygen, the total luminescence spectrum gave the sum of uorescence and phosphorescence, and time resolved analysis distinguished between uorescence and phosphorescence based on their different decay rates in time domain and indicated their proportions in the total spectrum. The data obtained from the luminescence titrations were tted and followed the following two equations:</p><p>A four-dimensional relationship between luminescence intensity, uorescence/phosphorescence ratio, hypochlorite concentration, and oxygen content was established (Fig. 3d). The projections of the three-dimensional surface onto the three perpendicular two-dimensional coordinate systems were shown in Fig. 3e-g. Fig. 3e showed the linear relationship between the luminescence intensity and the hypochlorite concentration in pure N 2 and O 2 . Fig. 3f showed the luminescence response toward oxygen in the presence and absence of hypochlorite (10 mM). Fig. 3g illustrated the relationship between the uorescence/phosphorescence ratio and hypochlorite and oxygen. To evaluate the accuracy of the 4D relationship, luminescence titration against hypochlorite was performed in pure N 2 , and the luminescence response toward oxygen was investigated in the presence of 10 mM of hypochlorite. As shown in Fig. 3h and i, the experimental data were highly consistent with the calculated data based on the 4D relationship graph. Additionally, luminescence decay curves of P1 in the presence and absence of hypochlorite (10 mM) were measured under N 2 and O 2 atmospheres, respectively (Fig. 3j). As shown in Fig. 3k, the uorescence and phosphorescence proportions were well in line with the calculated ones. The difference between the experimental and calculated values was slightly larger in the absence of hypochlorite compared with that in the presence of hypochlorite, because the experimental error increased when the uorescence was too weak in the absence of hypochlorite.</p><p>The cellular imaging and sensing properties of P1 was investigated using human cervix epithelioid carcinoma (HeLa) cells as a mode cell line. The cytotoxicity of P1 toward HeLa cells was evaluated via the MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay. HeLa cells maintained a viability >90% aer incubation with P1 at a concentration of 300 mg mL À1 for 24 h (Fig. S6 †), conrming negligible cytotoxic effect of P1 under the following imaging conditions. The microscopy imaging was performed under excitation at 405 nm, and the luminescence signals at 600 AE 25 nm were collected and analyzed. The microscopy luminesce images of HeLa cells treated with P1 are illustrated in Fig. 4a. HeLa cells incubated with P1 (200 mg mL À1 , 37 C, 1 h) showed cytoplasmic staining. Further incubation of the cells with hypochlorite for 30 min led to signicant enhancement of intracellular luminescence, which has been attributed to the oxidation-induced uorescence turn-on. However, it is difficult to quantitatively analyze the luminescence intensity against the hypochlorite concentration in the culture medium, because when the excitation power and slit width were xed, the intracellular luminescence signals were either too weak to create clear images in the absence of hypochlorite or too strong that exceeded the detection limit of the detector when the uorescence was turned on. Both the luminescence lifetime and the uorescence/phosphorescence ratio were independent of the excitation power and thus high quality images were taken under suitable but different excitation powers. 50 Time-resolved analysis of the images was performed via photoluminescence lifetime imaging microscopy (PLIM). The PLIM images showed that the luminescence of intracellular P1 exhibited a unique lifetime and distributed evenly in the cytoplasm (Fig. 4a). Statistical analysis revealed that the lifetimes of the intracytoplasmic P1 were normally distributed in the range of 300-600 ns with a mean of about 450 ns, which is fully in line with the phosphorescence lifetime of P1 (Fig. 4b). Further incubation with hypochlorite (10 mM, 30 min) gave rise to some short-lived pixels in the image and an additional distribution at around 150 ns appeared. The lifetime value was much longer than the uorescence lifetime and has tentatively been assigned to the averaged uorescence and phosphorescence lifetime of the hypochlorite-oxidized P1. The signals at about 150 ns became dominant when the hypochlorite concentration in the medium increased to 30 mM (Fig. 4a and b). The relative occurrence of the short-lived (<300 ns) and long-lived (>300 ns) signals linearly increased from 0.22 to 3.32 as the hypochlorite concentration in the medium increased from 0 to 30 mM (Fig. 4c). In another experiment, HeLa cells were pretreated with elesclomol (125 nM, 37 C, 2 h) to induce production of endogenous hypochlorite. 51 The cells were then incubated with P1 (200 mg mL À1 , 37 C, 1 h). The PLIM image showed substantial short-lived signals <300 ns (Fig. 4d and e), suggestive of oxidation of intracellular P1 by endogenous hypochlorite. The occurrence ratio of the shortlived signals over the long-lived signals (>300 ns) was about 1.29, which indicated that the amount of endogenously produced hypochlorite was almost the same as internalized amount during incubation of the cells with about 12 mM of hypochlorite for 30 min (Fig. 4c).</p><p>The utilization of P1 for analysis of intracellular oxygen content was then demonstrated. HeLa cells were preloaded with P1 (200 mg mL À1 , 37 C, 1 h), and then cultured under 2% O 2 atmosphere for 30 min. Although luminescence images of the cells did not show remarkable changes, the luminescence lifetimes were extended from 450 ns to 480 ns (Fig. 5a and b). The oxygen content in the culture atmosphere was then increased to 95%. The intracellular luminescence became very dark owing to the quenching of phosphorescence. Clear image was obtained by turning up the laser power and increasing the slit width (Fig. 5a). PLIM analysis showed that the luminescence lifetimes were shortened to about 390 ns (Fig. 5a and b). The less occurrence indicated that the phosphorescence was quenched by 66%. Aer that, hypochlorite (5 mM) was added to the medium and the cells were further cultured under ambient condition (37 C, air) for 30 min. As expected, the phosphorescence lifetime was restored to about 450 ns and additional lifetime distribution at about 150 ns appeared (Fig. 5c and d). The occurrence ratio of the short-lived signals over the longlived signals was about 0.53, which was in line with the linear relationship in Fig. 4c. Further changing the oxygen content in the culture atmosphere to 2% or 95% O 2 caused the corresponding phosphorescence response without interfering with the short-lived signals <300 ns (Fig. 5c and d). These results showed that the single-wavelength dual-emissive polymeric probe P1 be used for simultaneously and independently imaging and analyzing multiple analytes in living cells via timeresolved photoluminescence analysis.</p><!><p>Simultaneously analyzing signals of multiple luminescent probes effectively improves the accuracy and efficiency of luminescence biological detection and disease diagnosis. During signal processing, it is particularly important to avoid mutual interference of the signals from different probes. In the traditional method, identication of luminescence signals is usually based on their different wavelengths. In this work, we reported a novel method where identication of luminescence signals is based on their different decay rates. In the wavelength-resolved luminescence processing, luminescence intensity at different wavelengths is analyzed independently. The wavelength range covers the visible region and extends to the near-infrared region, usually being 400-900 nm with l max / l min to be 2.25. However, it is difficult to avoid spectral overlap and possible energy transfer between probes. Additionally, the difference in the penetration depth of light at different wavelengths in biological samples causes their detection efficiency to be different. In the time-resolved luminescence analysis, probes are allowed to emit at the same wavelength but different excitedstate decay rates are required. The luminescence with different lifetimes is analyzed independently. We selected a phenothiazine-based uorescent compound and a phosphorescent iridium(III) complex to construct a single-wavelength dualemissive polymeric probe. The two luminophores emitted at almost the same wavelength but their luminescence lifetimes are 160 times different. The uorescence was signicantly enhanced upon selectively oxidation by hypochlorite and the phosphorescence was efficiently quenched by molecular oxygen. The luminescence spectrum gave the sum of uorescence and phosphorescence, and analysis of the biexponential decay rates showed their proportions in the total spectrum. A four-dimensional relationship between luminescence intensity, uorescence/phosphorescence ratio, hypochlorite concentration, and oxygen content was established, which allowed simultaneous quantitative determination of hypochlorite concentration and oxygen content by measuring the luminescence intensity and the uorescence/phosphorescence ratio at a single wavelength. Owing to the excellent water solubility and biocompatibility, the single-wavelength dual-emissive polymeric probe was used for living cell imaging. Under ambient conditions, the phosphorescence remained unchanged. Timeresolved ratiometric sensing and imaging of intracellular hypochlorite was demonstrated using the phosphorescence signal as an internal standard. When the intracellular oxygen content changed, time-resolved photoluminescence analysis clearly showed the independent uorescence response toward hypochlorite and phosphorescence response toward oxygen in separated time intervals. Using time-resolved photoluminescence imaging to analyze three or more analytes is in progress in our laboratory. The luminescence lifetimes of available probes varied from nanoseconds of uorescent dyes to microseconds of transition metal complexes, to milliseconds of lanthanide chelates, to seconds of aerglow probes, giving s max / s min to be 109. The probes are designed to have the same emission wavelength in the red or near-infrared region to ensure the deep tissue penetration and the same photon detection efficiency.</p>

Royal Society of Chemistry (RSC)

Superelectrophilic carbocations: preparation and reactions of a substrate with six ionizable groups

A substrate has been prepared having two triarylmethanol centers and four pyridine-type substituent groups. Upon ionization in the Brønsted superacid CF 3 SO 3 H, the substrate undergoes two types of reactions. In the presence of only the superacid, the highly ionized intermediate(s) provide a double cyclization product having two pyrido[1,2-a]indole rings. With added benzene, an arylation product is obtained. A mechanism is proposed involving tetra-, penta-, or hexacationic species.

superelectrophilic_carbocations:_preparation_and_reactions_of_a_substrate_with_six_ionizable_groups

Introduction<!>Results and Discussion<!>Conclusion<!>Experimental<!>Supporting Information

<p>During the 1970s and 80s, Olah and co-workers described the novel chemistry of highly-charged organic cationic species. This work lead to the concept of superelectrophilic reactivity [1][2][3][4][5]. Examples of superelectrophiles include the nitronium dication (1) and the acetylium dication (2, Scheme 1). Both of these species have been proposed as superelectrophilic intermediates in the reactions of nitronium (NO 2 + ) and acetylium (CH 3 CO + ) salts in superacids. In sufficiently acidic media, cationic electrophiles such as the nitronium ion may undergo protonation, leading to the nitronium dication (1), and a greatly enhanced electrophilic reactivity. In superacidic solutions, nitronium salts have been shown to react with deactivated arenes and saturated hydrocarbons (including methane) [6][7][8][9].</p><p>Numerous studies from our group and others have shown that relatively stable cationic centers -such as ammonium, phosphonium, and pyridinium groups -may also be part of superelectrophilic systems [10]. Recently, we described the chemistry of tri-, tetra-, and pentacationic electrophiles based on the triarylmethyl cation scaffold (3-5, Scheme 1) [11,12]. These systems utilized pyridyl rings to produce increasing amounts of positive charge adjacent to the carbocation center. Both theoretical calculations and experimental results indicated that the carbocation center undergoes a high degree of delocalization into the neighboring phenyl group. Both the trication 3 and the tetracation 4 were directly observed from FSO 3 H-SbF 5 solution using low temperature NMR. Experimental observations also revealed an exceptionally high acidity of the pyridinium N-H bonds. Here, we describe the preparation and chemistry of a substrate with six ionizable groups -four pyridyl rings and two carbinol centers.</p><!><p>The desired substrate was prepared in four steps from 2,6-dibromopyridine (Scheme 2). Utilizing 2-lithio-6-bromopyridne, product 6 is formed in modest yield by reaction with benzaldehyde. A nickel-catalyzed procedure gives the dipyridyl intermediate 7 [13]. This is easily oxidized to the diketone 8 and reaction of this substance with 2-lithiopyridine gives the precursor 9. The diol 9 is a substrate with six ionizable groups.</p><p>Upon ionization in superacidic CF 3 SO 3 H (triflic acid), compound 9 undergoes two types of reactions. When the substrate is ionized in the presence of benzene and CF 3 SO 3 H, the arylation product 10 is formed as the major product (Scheme 3). Presumably, compound 10 is formed as a mixture of meso and dl stereoisomers. Similar reaction products were observed in our studies of tri-, tetra-, and pentacationic systems [11,12]. This product, 10, is the result of charge migration involving the carbocation center and the phenyl group (vide infra). When compound 9 is treated with only superacid, the bis(pyrido[1,2a]indole) 11 is formed as the major product. Likewise, the tri-, tetra-, and pentacationic systems 3-5 provide the pyrido[1,2a]indole ring system. Interestingly, there was no evidence of the cyclization product 11 (NMR analysis) when compound 9 reacts with superacid in the presence of benzene -product 10 is formed as the major product.</p><p>These products may be explained by mechanisms involving highly ionized intermediates. It is proposed that compound 9 initially reacts in excess superacid to give the tetracationic species 12 (Scheme 4). Protonation of the hydroxy groups leads to immediate ionization and formation of the carbocation centers. For related conversions, computational and experimental data indicated that the protonated hydroxy groups (oxonium ions) are not persistent intermediates, but rather cleavage of the carbon-oxygen bond is almost instantaneous [12]. It is assumed that ionization to the carbocations occurs in a stepwise process, first providing pentacation 13 then the hexacation 14. The product forming steps occur through either 13 or 14. For the arylation product 10, charge delocalization at the carbocation leads to nucleophilic attack at the para-position of the phenyl group. This S E Ar step is followed by protonation at the methine posi-Scheme 3: Isolated yields of products from diol 9.</p><p>Scheme Proposed mechanisms leading to products 10 and 11.</p><p>Scheme 5: Products and relative yields from the reaction of alcohol 18 with CF 3 SO 3 H and C 6 H 6 [12]. tion and deprotonation of the para-carbon to complete the arylation step. For the cyclization product 11, theoretical calculations indicate that cyclization requires deprotonation at the pyridinium ring [11]. Thus, either the tetracation 15 or the pentacation 16 is the likely precursor to the pyrido[1,2-a]indole ring system. The conversion to product 11 also involves a stepwise process -initially forming compound 17 then a second cyclization gives the final product 11.</p><p>As noted above, no cyclization product 11 was observed when the chemistry was done in the presence of benzene. Only product 10 was obtained. This observation is in stark contrast to our earlier results involving both the tetracation 4 and pentacation 5. When these highly charged ions are generated in CF 3 SO 3 H and C 6 H 6 , significant quantities of the cyclization products are formed with the respective arylation products. For example, alcohol 18 provides a mixture of products 19 and 20 in a 32:68 ratio, presumably through the pentacation 5 (Scheme 5). Even with the use of the stronger Brønsted acid, CF 3 SO 3 H-SbF 5 , significant quantities of the cyclization product 20 are observed. This raises an obvious question: why does substrate 9 provide exclusively the arylated product 10 in the presence of benzene, while substrate 18 leads to a significant proportion of cyclization product 20?</p><p>Our proposed mechanism of cyclization involves deprotonation of an N-H bond at the pyridinium ring. Although pyridinium deprotonation is not generally expected in a Brønsted superacid, it can occur in these systems because of the large amount of cationic charge on these structures. In comparing the chemistry of compounds 9 and 18, compound 9 can lead to the hexacationic intermediate 14 while compound 18 can lead to the pentacationic intermediate 5. Examination of these two structures suggests that the phenyl groups have a profound effect on stabilizing the ions. Thus, the hexacationic system 14 is stabilized by two phenyl groups, while the pentacationic system 5 has one stabilizing phenyl group (Scheme 6). The ratio of charges is 2:1 (pyridinium/benzylic carbocation) for the hexacation 14, while the ratio of charges is 4:1 (pyridinium/benzylic carbocation) for the pentacation 5. Similarly, we previously reported good chemoselectivity for trication 3 -no cyclization product observed in the presence of benzene -but poor chemoselectivity for the tetracation 4. The ratio of charges in these systems are consistent: a charge ratio of 2:1 shows good chemoselectivity (trication 3), while a charge ratio of 3:1 shows poor chemoselectivity (tetracation 4). As we previously reported, a reaction of tetracation 4 with in CF 3 SO 3 H leads to a mixture of arylated product (52%) and cyclization product (48%). Thus, we observed clean arylation chemistry when the carbocation sites are flanked by not more than two pyridinium groups. This also means that increasing the number of adjacent pyridinium groups destabilizes the system as a whole and leads to greater N-H deprotonation. Tetracation 4 and pentacation 5 tend to undergo N-H deprotonation more readily, and consequently, this leads to rapid cyclization reactions.</p><p>Regarding the site of deprotonation, hexacation 14 could potentially undergo N-H deprotonation at the inside pyridinium ring (16) or the outside pyridinium ring (21, Scheme 7). While inside deprotonation should give the observed cyclization product 11, outside deprotonation would give an entirely different product, one having the pyrido[1,2-a]indole ring at the end of the structure. Compound 11 is the only major product observed from the superacid-promoted reaction of diol 9. This suggests outside deprotonation -and formation of ion 21 -does not occur.</p><p>The preference for inside deprotonation may be understood to be a consequence of charge-charge repulsive effects. In the case of 21, the five cationic charges are on neighboring positions, while in the case of 16, the five cationic charges are separated into groups of three and two charges. The increased stability of the separated cationic charge is evident in the DFT calculated energies of the ions. At the B3LYP 6-311G (d,p) level, ion 16 is calculated to be 32.7 kcal•mol −1 more stable than ion 21 [8]. Thus, highly charged organic ions may benefit by having groups of charges separated into smaller clusters rather than having all of the charges grouped together.</p><!><p>We have prepared a substrate with six ionizable sites. Reaction of the substrate in superacidic CF 3 SO 3 H leads to cyclization or arylation products, depending on the presence or absence of benzene. A mechanism is proposed involving tetra-, penta-, and hexacationic reactive intermediates. Most notably, this system shows remarkably good chemoselectivity in its reaction with benzene (only arylation product is observed). This is attributed to the presence of two carbocationic sites stabilized by benzylic-type resonance. Thus, molecular structures having a very large overall charge may be viable if stabilizing groups are incorporated into the structure.</p><!><p>General. All reactions were performed using oven-dried glassware under an argon atmosphere. Trifluoromethanesulfonic acid (triflic acid) was freshly distilled prior to use. All commercially available compounds and solvents were used as received. 1</p><!><p>Supporting Information File 1</p><p>Experimental procedures, compounds characterization, and NMR spectra; computational methods and results.</p><p>[https://www.beilstein-journals.org/bjoc/content/ supplementary/1860-5397-15-153-S1.pdf]</p>

Beilstein

Electro-catalytic amplified sensor for determination of N-acetylcysteine in the presence of theophylline confirmed by experimental coupled theoretical investigation

The 1,l/-bis(2-phenylethan-1-ol)ferrocene, 1-butyl-3-methylimidazolium hexafluoro phosphate (BMPF6) and NiO-SWCNTs were used to modify carbon paste electrode (BPOFc/BMPF6/NiO-SWCNTs/ CPE), which could act as an electro-catalytic tool for the analysis of N-acetylcysteine in this work. The BPOFc/BMPF6/NiO-SWCNTs/CPE with high electrical conductivity showed two completely separate signals with oxidation potentials of 432 and 970 mV for the first time that is sufficient for the determination of N-acetylcysteine in the presence of theophylline. The BPOFc/BMPF6/NiO-SWCNTs/ CPE showed linear dynamic ranges of 0.02-300.0 μM and 1.0-350.0 μM with the detection limit of ~ 8.0 nM and 0.6 μM for the measurement of N-acetylcysteine and theophylline, respectively. In the second part, understanding the nature of interaction, quantum conductance modulation, electronic properties, charge density, and adsorption behavior of N-acetylcysteine on NiO-SWCNTs surface from first-principle studies through the use of theoretical investigation is vital for designing highperformance sensor materials. The N-acetylcysteine molecule was chemisorbed on the NiO-SWCNTs surface by suitable adsorption energies (− 1.102 to − 5.042 eV) and reasonable charge transfer between N-acetylcysteine and NiO-SWCNTs.

electro-catalytic_amplified_sensor_for_determination_of_n-acetylcysteine_in_the_presence_of_theophyl

<!>Results and discussion<!>Conclusion

<p>The thiolic biological compounds such as cysteine, N-acetylcysteine, homocysteine, glutathione, captopril etc. play important roles in human health [1][2][3][4] . N-acetylcysteine as thiol drugs showed much application in the treatment of chest pain, Alzheimer disease, and overdose with acetaminophen [5][6][7][8][9][10] . In addition, N-acetylcysteine is used for the control of lipoprotein and homocysteine levels that are harmful to the human body at high levels 11 . On the other hand, the high consumption of N-acetylcysteine can increase risk of vomiting, nausea, and constipation which is very significant for the investigation of N-acetylcysteine in real samples [12][13][14][15] .</p><p>On the other hand, theophylline is a methylxanthine drug with a wide range of application for the treatment of severe asthma, prevention of contrast-induced nephropathy, chronic bronchitis, chest tightness, and wheezing. According to the scientific report, the combination of theophylline and N-acetylcysteine is useful for treating Real sample preparation. The water and pharmaceutical serum samples were purchase from the local market and pharmacy and directly used for electrochemical analysis. Tablet samples were purchased from local pharmacy and then were completely ground and homogenized. Next, their calculated values were weighed and then dissolved in 50 mL of water/ethanol solution and the mixture was filtered for real sample analysis.</p><p>Computational details. The electronic and structural properties of N-acetylcysteine adsorbed onto a NiO-SWCNTs surface was investigated using the plane-wave DFT calculations as implemented in the Cambridge Serial Total Energy Package code 50 . The generalized gradient approximation (GGA) of the Perdew-Burke-Ernzerhof (PBE) functional 51 and ultrasoft pseudopotentials 52 were used to describe the exchange-correlation and core-valence electron. The adsorption energies were calculated using the dispersion correction by Grimme 53 , since van der Waals interactions were anticipated to affect the adsorption energies. A vacuum gap of 20 Å was used to prevent the interactions between the periodic slabs perpendicular to the surface, resulting in a simulation supercell of 10.393 × 8.520 × 35.073 Å 3 . The Monkhorst-Pack 54 with k-mesh of 4 × 1 × 1 was used to sample the Brillouin zone. The wave functions of the valence electron were described using a plane-wave basis set with a cut-off energy of 400 eV. To account for the metallic behavior of NiO 2 (the oxidation states of each element in NiO 2 are + 4 (Ni) and -2 (O)), the atomic positions were optimized via the BroydeN-Fletcher-Goldfarb-Shanno scheme 55 with an energy convergence criterion, force, and displacement of less than 10 −6 eV/atom, 0.3 eV/Å, and 0.01 Å, respectively. However, all other atoms and lattice vectors on the top layer of the slab were allowed to relax, since surface adsorption occurred on the topmost layer. The Hirshfeld's analysis 56 was used to evaluate the charge transfer between N-acetylcysteine and NiO-SWCNTs.</p><p>The stability of N-acetylcysteine adsorption on the NiO-SWCNTs surface was evaluated by calculating the adsorption energy (E ads ):</p><p>(1)</p><p>where E N−acetylcysteine@NiO−SWCNTs , E N−acetylcysteine@NiO−SWCNTs , E N−acetylcysteine , E NiO and E SWCNTs are the sum of the energies of N-acetylcysteine which are adsorbed on NiO-SWCNTs surface, monolayers of N-acetylcysteine, NiO and SWCNTs, respectively. Generally, a negative E ads signifies that the adsorption process was exothermic and energetically stable 57 .</p><p>The highest occupied molecular orbital (HOMO) − lowest unoccupied molecular orbital (LUMO) gap (HLG) was evaluated following Eq. ( 2):</p><p>where E LUMO and E HOMO are the energies of the LUMO and HOMO, respectively. The electronic sensitivity of the NiO-SWCNTs towards the adsorption of N-acetylcysteine was assessed by calculating the change in the HLG 58 :</p><p>where E g1 and E g2 represent the HLG before and after adsorption.</p><!><p>NiO/SWCNTs characterization. The elemental analysis of NiO-SWCNTs is shown in Fig. 1. The existence of C, Ni and O elements confirm the purity of synthesized NiO-SWCNTs nano-composites by the recommended procedure. The decoration of NiO/NPs at functional SWCNTs was characterized by TEM method (Fig. 2A). The presence of nickel oxide nanoparticles on the single-wall carbon nanotubes surface is well represented in Fig. 2A. In contrast, the XRD patterns of NiO-SWCNTs confirm the occurrence of (002) at 2θ ~ 26° plane relative to carbon nanotubes and other planes, i.e. (111), (200), (220), (311) and (222) at positions of 37.171°, 43.231°, 62.791°, 75.321° and 79.191° relative to NiO nanoparticle with FCC structure (Fig. 2).</p><p>(2) it is observed that maximum electro-catalytic interaction could occur at pH = 7.0 and this pH was chosen as the best condition.</p><p>The signal for the oxidation of 1.0 mM N-acetylcysteine was recorded at BPOFc/NiO-SWCNTs/CPE (Fig. 3, curve b), BPOFc/BMPF6/CPE (Fig. 3, curve c), BPOFc/BMPF6/NiO-SWCNTs/ CPE (Fig. 3, curve d), BMPF6/ NiO-SWCNTs/CPE (Fig. 3, curve e), BMPF6/CPE (Fig. 3, curve f), NiO-SWCNTs/CPE (Fig. 3 curve g) and CPE (Fig. 3, curve h). On the other hand, BPOFc/BMPF6/NiO-SWCNTs/CPE exhibited an oxidation/reduction signal with ∆Ep = 130 mV that confirms quasi-reversible behavior of BPOFc/BMPF6/NiO-SWCNTs/CPE in the aqueous solution (curve a). The increasing oxidation signal of BPOFc/BMPF6/NiO-SWCNTs/CPE and simultaneous decrease in reduction signal of mediator after addition of 1.0 mM N-acetylcysteine, confirms an EC / interaction 59,60 between mediator and N-acetylcysteine on the surface of BPOFc/BMPF6/NiO-SWCNTs/ CPE. The comparison of the electro-catalytic oxidation signal of N-acetylcysteine at the surface of BPOFc/ BMPF6/NiO-SWCNTs/CPE with its signal at BPOFc/NiO-SWCNTs/CPE and BPOFc/BMPF6/CPE confirmed that the conductivity of electrode surface could be enhanced by the existence of NiO-SWCNTs and BMPF6. This increase in conductivity helps to improve oxidation current and decrease oxidation potential. In addition, the comparison of the electro-catalytic oxidation signal of N-acetylcysteine at the surface of BPOFc/NiO-SWCNTs/ CPE with its signal at BPOFc/BMPF6/CPE shows that reduction in oxidation potential at the surface of BPOFc/ NiO-SWCNTs/CPE is more than its reduction at the surface of BPOFc/BMPF6/CPE. This point can be related to the high viscosity of BPOF, which makes it difficult to access the electrode surface. On the other hand, oxidation signal of N-acetylcysteine showed a weak signal at the surface of CPE. After modification of CPE with NiO-SWCNTs or BPOFc, the oxidation signal of N-acetylcysteine was improved, that could be related to high conductivity of mediators at the surface of CPE. In addition, after modification of CPE with NiO-SWCNTs and www.nature.com/scientificreports/ (sensitivity 0.1643 μA/μM and R2 = 0.9975), respectively. The detection limit (3σ) was set at ~ 8.0 nM and 0.6 μM for N-acetylcysteine and theophylline at the surface of BPOFc/BMPF6/NiO-SWCNTs/CPE as a novel electrochemical sensor using (LOD = 3S b /m) equation. The BPOFc/BMPF6/NiO-SWCNTs/CPE displayed better dynamic range or the limit of detection for determination of N-acetylcysteine compared to another electrochemical methods suggested (Table 1).</p><p>The differential pulse voltammograms of different concentration of N-acetylcysteine and theophylline were measured at the BPOFc/BMPF6/NiO-SWCNTs/CPE surface (Fig. 6A). The voltammograms showed two oxidation peaks separated at potentials of ~ 432 mV and 970 mV that is relative to oxidation of N-acetylcysteine and theophylline, respectively. Figure 6B,C showed the plots of oxidation current vs. concentration of drugs. As can be seen, the sensitivity for the simultaneous investigation of N-acetylcysteine and theophylline is equal to 0.2078 and 0.1627 μA/μM, which are comparable with sensitivity obtained for the two drugs in linear dynamic range determination. This study revealed that a concurrent determination of N-acetylcysteine and theophylline is possible at BPOFc/BMPF6/NiO-SWCNTs/CPE surface with no interference.</p><p>The stability of BPOFc/BMPF6/NiO-SWCNTs/CPE was also checked in the presence of 10.0 μM N-acetylcysteine + theophylline solution. BPOFc/BMPF6/NiO-SWCNTs/CPE was stored at the laboratory temperature, and the electro-catalytic signal of BPOFc/BMPF6/NiO-SWCNTs/CPE had no apparent decrease in the first fifteen days. Compared with its first electro-catalytic signal, the response sensitivity remained at 96% after 50 days. The obtained results confirmed good stability of BPOFc/BMPF6/NiO-SWCNTs/CPE as a new electrochemical sensor.</p><p>To check the selectivity of BPOFc/BMPF6/NiO-SWCNTs/CPE, the interference effects of some usual biological, cationic, and anionic compounds are investigated in the solution containing 10.0 μM N-acetylcysteine + theophylline. The results indicated that 1000-fold of K + , F -, Na + , Br − and Ca 2+ and 600-fold of glucose, phenylalanine, and urea have no major influence on the investigation of 20.0 μM N-acetylcysteine.</p><p>The ability of the BPOFc/BMPF6/NiO-SWCNTs/CPE was investigated for the study of N-acetylcysteine and theophylline in the tablet samples by standard addition technique. The obtained data are shown in Table 2. The Theoretical studies. Electronic properties and structural stability of NiO and SWCNTs. Before exploring the adsorption properties of N-acetylcysteine using NiO-SWCNTs, the organizational constancy of NiO and SWCNTs was evaluated using Eq. ( 4) and ( 5) 61 :</p><p>where E(NiO) , E(SWCNTs) , E(Ni) , E(O) , and E(C) are the total energies of NiO, SWCNTs, isolated Ni, O, and C atoms, respectively. Moreover, x, y, and z are the number of Ni, O and C atoms, respectively. The formation energy of NiO and SWCNTs was calculated as − 9.83 and − 8.71 eV, respectively, confirming their stable structure. The (4) 7).</p><p>Several types of C-C bonds were observed in the SWCNTs with different bond lengths of 1.40-2.44 Å (bonds shared between two hexagons) and 1.44 Å (the bond shared between a hexagon and pentagon), which were comparable with other studies 62 . The Ni-O bond length of 2.10 Å was in agreement with the earlier results (2.08 Å) 63 . The electronic properties of NiO and SWCNTs were described based on the HLG. The band structure of NiO in Fig. 8 revealed that the SWCNT is a semiconductor with an E g of 0.67 eV. The LUMO and HOMO of the SWCNT were − 5.01 and − 5.81 eV, respectively.</p><p>From the DOS plot of NiO, an HLG of 0.80 eV was revealed. The obtained PDOS results showed that the 3d orbitals of the surface Ni were mainly located at the HOMO, while at the LUMO, the hybridization was mostly contributed by Ni 4 s orbitals. The PDOS results suggested that the Ni 3d orbitals play a key influence on the adsorption process.</p><p>Several configurations were explored to find the most feasible adsorption sites where four local minima were obtained (Fig. 9).</p><p>Based on the E ads calculations, the four configurations of N-acetylcysteine adsorption onto NiO-SWCNTs were exothermic processes with negative adsorption energies ranging between − 1.102 and − 5.042 eV (Table 3). Moreover, the adsorption energy varies owing to the interactions of N-acetylcysteine molecule with several adsorption sites with the NiO-SWCNTs. As presented in Table 3, the four relaxed configurations with more negative adsorption energy values and small interaction distances (ranging from 1.689 to 1.980 Å) between the N-acetylcysteine and NiO-SWCNTs, signify strong interactions and stability. This strong interaction indicates that the NiO-SWCNTs is a prominent sensor for the adsorption of N-acetylcysteine with good response to all the adsorption sites considered. Moreover, the more negative adsorption energy value suggests that the reaction will release more energy. Among these configurations, the most stable (SNA3) is where the acidic end is bonded strongly with the interfacial Ni atoms of the substrate [64][65][66] .</p><p>The interaction between the N-acetylcysteine molecule and NiO-SWCNTs was anticipated to alter the electronic property of N-acetylcysteine, which could be understood by the variation in its energy band gap [67][68][69] . The electronic property of N-acetylcysteine molecule was studied based on the HLG and density of states (DOS) spectrum (Fig. 10). The DOS of N-acetylcysteine molecule possesses a broad HOMO and LUMO separated by a wide HLG (Fig. 10). After adsorption, the N-acetylcysteine molecule introduced sharp occupied bands in the HLG of all the configurations. The TDOS results revealed similar changes, which indicated that NiO-SWCNTs might be an effective sensor towards the N-acetylcysteine molecule. The adsorption of N-acetylcysteine molecule shifted the HOMO levels to a higher energy, whereas the LUMO levels remained unaffected. Thus, the HLG value of N-acetylcysteine molecule was significantly reduced compared to its isolated molecule. The average HLG variation (|ΔHLG| (%)) upon adsorption of N-acetylcysteine molecule onto NiO-SWCNTs surface is connected with the sensitivity of adsorbent, as well as modifying its electrical conductivity. The |ΔHLG| (%) of 63.09, 68.33, 69.83, and 65.84% for configurations SNA1, SNA2, SNA3, and SNA4, respectively (see Table 3), signified high sensitivity of NiO-SWCNTs towards the adsorption of N-acetylcysteine molecule on its surface. From the HLG variation result, it was established that the sensing response of NiO-SWCNTs towards N-acetylcysteine molecule was observed to be rather higher for SNA3 configuration. Further understanding into the bonding mechanisms between the N-acetylcysteine molecule and NiO-SWCNTs was obtained by analyzing the DOS of N-acetylcysteine molecule before and after adsorption onto the NiO-SWCNTs surface (Fig. 10). After adsorption, the DOS of N-acetylcysteine molecule was broadened owing to the strong hybridization with the adsorbed Ni ion. This showed a chemisorption state of N-acetylcysteine molecule. where σ , k, E g and T are the electrical conductivity, Boltzmann's constant, bandgap energy and thermodynamic temperature, respectively. According to this equation, smaller HLG values lead to a larger electrical conductivity. The electrical conductivity before adsorption was 2.86 × 10 -9 . Therefore, the electrical conductivity of SNA1, SNA2, SNA3 and SNA4 configurations was higher after adsorption.</p><p>To evaluate the interactions between the N-acetylcysteine and NiO-SWCNTs, the three-dimensional (3D) charge density difference was calculated, as given in Fig. 11.</p><p>In this 3D charge density difference plot, the electron enrichment and depletion are shown as blue and yellow isosurfaces, respectively. The electronic interaction largely occurred at the top of Ni atoms of the NiO-SWCNTs nanocomposite, which was in direct contact with the N-acetylcysteine molecule. The electrons transfering from the N-acetylcysteine molecule to NiO-SWCNTs indicated that the Ni atoms were oxidized during the adsorption process. Since the electron accumulation sites were mostly located at the interface, they confirmed that the bond between N-acetylcysteine molecule and NiO-SWCNTs was of a covalent nature. However, less electron density was observed at the NiO-SWCNTs interface, signifying that the SWCNTs was less influenced electronically by the interaction with NiO nanoparticle. The interactions between the N-acetylcysteine molecule and NiO-SWCNTs indicates a substantial charge transfer, which was evaluated grounded on the Hirshfeld charge analysis. The charge migration analysis of SNA1, SNA2, SNA3, and SNA4 configurations was found to be 1.11, 1.18, 1.21 and 1.14 |e|, respectively. A positive value of Hirshfeld charge analysis was observed for the four interaction sites considered in this study. This further confirmed electrons transfer from the N-acetylcysteine molecule to NiO-SWCNTs. Table 3. The adsorption energy (E ads ), adsorption distance (D), LUMO energy (E LUMO ) , HOMO energy (E HOMO ) , HOMO-LUMO gap (HLG), change of HLG (|ΔHLG|), charge transfer (QT) and conductivity (σ ) for the adsorption of N-acetylcysteine molecule onto NiO-SWCNTs surface.</p><!><p>The electro-catalytic interaction between BPOFc and N-acetylcysteine was studied at the BPOFc/BMPF6/NiO-SWCNTs/CPE surface. The cyclic voltammograms data confirms the good selectivity and high sensitivity of BPOFc/BMPF6/NiO-SWCNTs/CPE for determination of N-acetylcysteine. Moreover, the most prominent adsorption site, sensitivity, conductivity, charge transfer, electronic and structural properties of N-acetylcysteine molecule adsorption onto NiO-SWCNTs surface was studied using DFT studies. The negative adsorption energies in the range of -1.102 to -5.042 eV and suitable charge transfer confirmed the stability of N-acetylcysteine adsorption at NiO-SWCNTs surface. In addition, the adsorption of N-acetylcysteine molecule was chemisorption. Therefore, the most prominent adsorption site of N-acetylcysteine molecule at NiO-SWCNTs surface</p>

Scientific Reports - Nature

The structure of the bacterial iron–catecholate transporter Fiu suggests that it imports substrates via a two-step mechanism

The ferric iron uptake (Fiu) transporter from Escherichia coli functions in the transport of iron–catecholate complexes across the bacterial outer membrane, providing the bacterium with iron, which is essential for growth. Recently it has become clear that Fiu also represents a liability for E. coli because its activity allows import of antimicrobial compounds that mimic catecholate. This inadvertent import suggests the potential utility of antimicrobial catechol siderophore mimetics in managing bacterial infections. However, to fully exploit these compounds, a detailed understanding of the mechanism of transport through Fiu and related transporters is required. To address this question, we determined the crystal structure of Fiu at 2.1–2.9 Å and analyzed its function in E. coli. Through analysis of the Fiuo crystal structure, in combination with in silico docking and mutagenesis, we provide insight into how Fiu and related transporters bind catecholate in a surface-exposed cavity. Moreover, through determination of the structure of Fiu in multiple crystal states, we revealed the presence of a large, selectively gated cavity in the interior of this transporter. This chamber is large enough to accommodate the Fiu substrate and may allow import of substrates via a two-step mechanism. This would avoid channel formation through the transporter and inadvertent import of toxic molecules. As Fiu and its homologs are the targets of substrate-mimicking antibiotics, these results may assist in the development of these compounds.

the_structure_of_the_bacterial_iron–catecholate_transporter_fiu_suggests_that_it_imports_substrates_

Introduction<!>Fiu is a member of a distinct clade of iron–catecholate transporters<!><!>Fiu is a member of a distinct clade of iron–catecholate transporters<!>The crystal structures of Fiu reveal a large, gated internal chamber<!><!>The crystal structures of Fiu reveal a large, gated internal chamber<!>In silico docking suggests that Fiu possesses multiple substrate-binding sites<!><!>The location of the putative external Fiu substrate-binding site is conserved among TBDTs<!><!>The amino acids at the Fiu external binding site are important for iron acquisition in vivo<!><!>The amino acids at the Fiu external binding site are important for iron acquisition in vivo<!>Discussion<!>TBDT phylogeny construction, Fiu homolog search, and clustering analysis<!>Construction of a multiple TBDT knockout E. coli BW25113 strain<!>Testing the growth of E. coli BW25113 TBDT deletion strains under iron-limiting conditions<!>Complementation of TBDT receptor mutants with WT and mutant Fiu<!>Protein expression and purification<!>Protein crystallization, data collection, and structure solution<!>In silico docking of Fe–DHB with the Fiu crystal structure<!>Analysis of TBDT ligand binding sites<!>Author contributions<!>

<p>The outer membrane of Gram-negative bacteria provides a selective permeability barrier to molecules with a molecular mass greater than ∼600 Da (1). The barrier provides superb protection against antimicrobials and toxic compounds (2). The selective permeability of the outer membrane also restricts the uptake of nutrients, including iron, which, although abundant on Earth, is often growth-limiting because of its insolubility under the oxidizing conditions of the terrestrial atmosphere (3, 4). Aerobic organisms solubilize iron through the formation of iron-chelating chemicals (siderophores) or incorporate it into other organic structures, such as the porphyrin ring of heme or iron-binding proteins (5–7). These iron-containing complexes are larger than the diffusion limit of the bacterial outer membrane, and so, to obtain the iron required for growth, bacteria have evolved transporters capable of selectively binding and importing iron-containing complexes (8, 9). Members of the TonB-dependent transporter (TBDT)3 family drive transport of their substrates through interaction with the energy-transducing protein TonB (10). TBDTs are highly divergent in sequence but share a common structural architecture, consisting of a 22-stranded transmembrane β-barrel, the lumen of which is selectively occluded by a globular plug domain (8).</p><p>The ability of TBDTs to import large substrates comes at a cos, an evolutionary arms race exists between TBDT-producing bacteria and organisms seeking to kill them by hijacking these transporters (11–15). Both small-molecule and protein antibiotics mimic TBDT substrates, leading to their inadvertent import into the bacterial cell (12, 16–21). Catecholates are one of the four recognized classes of siderophores (22); they have a strong affinity for iron and are abundant secretion products of both bacteria and fungi (23, 24). The ferric iron uptake (Fiu) transporter is a TBDT responsible for import of molecules containing the catecholate functional group (2, 25). In addition to its role in iron uptake, Fiu has also been shown to be important for sensitivity to antimicrobials that share the common feature of a catecholate functional group or the analogous dihydroxypyridine moiety (21, 25, 26). It is thought that the presence of these chemical mimetics leads to their inadvertent import into the bacterial cell via Fiu (27). This Fiu-mediated sensitivity is observed even in the absence of iron because Fiu functions in import of catecholate-containing molecules, seemingly independent of their size or Fe coordination state (18). The potential of antimicrobial catechol siderophore mimetics as therapeutic agents is demonstrated by development of the 3,4-dihydroxypyridine–containing sulbactam BAL30072 by Basilea Pharmaceutica and the catecholate-containing cephalosporin cefiderocol by Shionogi Inc. BAL30072 and cefiderocol have entered clinical trials for treatment of infections by Gram-negative bacteria in the human lung and urinary tract, respectively (28, 29).</p><p>In this study, we show that, although Fiu and its homologs PiuA and PiuD function in import of catecholate siderophores (25, 30), they are evolutionarily distinct from the well-studied catecholate siderophore transporters FepA/PfeA and Cir. To investigate the substrate import mechanism of the Fiu/Piu TBDT subgroup, we solved the crystal structure of Fiu. Analysis of this structure in combination with in silico docking and mutagenesis identified an external substrate-binding site in Fiu, which is conserved among diverse TBDTs. In addition, the presence of a large selectively gated internal chamber in Fiu, capable of accommodating a Fe–siderophore complex, suggests that these transporters may function via a two-step gating mechanism.</p><!><p>It has been demonstrated previously that the archetypical Escherichia coli strain BW25113 possesses three TBDT transporters that function in the uptake of catecholate siderophores: FepA, Cir, and Fiu (25). FepA imports the endogenously produced siderophore enterobactin with high affinity (31), whereas Cir and Fiu have been shown to transport monomeric catecholate compounds, either alone or in complex with iron (32). Although these transporters recognize a common functional group, they share limited amino acid sequence identity, and their evolutionary relationship remained undetermined. To resolve this question, we performed phylogenetic analysis of these transporters in the context of a panel of diverse TBDTs of known structure and/or function. This analysis revealed that, although Cir and FepA belong to the same clade of the TBDT phylogram, Fiu belongs to a distal clade with the TBDTs PiuA and PiuD that also mediate catecholate transport (30, 33) (Fig. 1A). A wider analysis of this Fiu/Piu clade, identified by HMMER search revealed that related sequences are widespread in proteobacteria (Fig. S1 and Tables S1 and S2). Based on clustering analysis of these sequences, all members of this expanded group are more closely related to Fiu than to either Cir or FepA (Fig. S1). Furthermore, consistent with our phylogram, these transporters are more closely related to the hydroxamate–siderophore transporter FhuA than either Cir or FepA (Fig. 1A and Fig. S1). These data demonstrate that, although Fiu, Cir, and FepA all transport catecholate-containing substrates, Fiu is evolutionarily distinct from Cir and FepA and may have arrived at its substrate specificity because of convergent evolution between these transporters.</p><!><p>Fiu belongs to a distinct group of catecholate siderophore transporters. A, a phylogenetic tree of diverse, functionally characterized TBDTs, showing that Fiu forms a clade with PiuA/D that is distant from the catecholate siderophore transporters FepA and Cir. Catecholate-transporting TBDT subgroups are color coded: blue, enterobactin-transporting; green, nonenterobactin-transporting FepA-related; red, nonenterobactin-transporting Fiu-related. Circles represent TBDTs present in E. coli BW25113. B, a scheme showing the sequential deletion of TBDTs in E. coli BW25113 utilized in this study, colored as in A. C, strains from B grown on LB agar in the presence of 0–150 μm 2,2′-bipyridine (BP). Sequential loss of FepA and Fiu leads to defects in the ability of strains to grow under iron-limiting conditions.</p><!><p>TBDT mediated iron uptake-systems are generally redundant to provide the means to obtain iron under a variety of environmental conditions (23, 34, 35). Thus, to dissect the specific role of Fiu, each of the six TBDTs known to be involved in iron acquisition were sequentially deleted from E. coli BW25113 in the following order: ΔfhuA (ferrichrome transporter), ΔfecA (ferric citrate transporter), ΔcirA (Fe–catecholate siderophore transporter), ΔfepA (enterobactin transporter), ΔfhuE (rhodotorulic acid transporter), and Δfiu (Fig. 1B). The phenotypes of these mutants were assessed by growth on LB agar containing the iron chelator 2,2′-bipyridine (Fig. 1C). The first three receptors (FhuA, FecA, and CirA) were dispensable for growth in this assay, but subsequent loss of the enterobactin receptor FepA affected the growth of the mutant strain at 2,2′-bipyridine concentrations of more than 50 μm (Fig. 1C). There was no further phenotype from loss of the coprogen receptor FhuE under our assay conditions (Δ5, Fig. 1C). Subsequent loss of Fiu led to impaired growth on LB agar and completely prevented growth at 2,2′-bipyridine concentrations of 50 μm or higher (Δ6, Fig. 1C). This growth defect was restored either by in trans complementation with a plasmid encoding Fiu (Fig. S2) or supplementation of the growth medium with Fe(II)SO4.</p><p>These data show that Fiu is able to provide iron to the cell when present as the sole outer-membrane iron transporter. As Fiu is unable to transport endogenously produced enterobactin (31, 36), in this context, it most likely functions to transport enterobactin breakdown products (i.e. 2,3-dihydroxybenzoyl-l-serine (DHBS)) in complex with iron. The inability of Fiu to support growth in the presence of high concentrations of 2,2′-bipyridine may be due to the lower affinity of the monomeric catecholates for Fe3+ or a low affinity of Fiu for the Fe–DHBS complex.</p><!><p>To obtain insight into the structural basis of substrate binding and import by Fiu, we determined the structure of Fiu by X-ray crystallography (Table S3). The structure of Fiu consists of a 22-stranded transmembrane β-barrel characteristic of the TBDT superfamily, with a number of extended extracellular loops that might serve in the initial steps of substrate binding (Fig. 2A). In agreement with our phylogenetic analysis (Fig. 1A), the DALI web server (37) identified PiuA from Acinetobacter baumannii (PDB code 5FP1) as the closest structural homolog to Fiu in the PDB (Dali server Z-score = 45, backbone atom RMSD of 6.182 Å, 33% amino acid identity). The structure of Fiu was solved in three different crystal forms, revealing Fiu in two distinct states (Table S3). In crystal state 1, extracellular loops 7–9 of the β-barrel were disordered, as was the extended extracellular loop of the N-terminal plug domain, which occluded the lumen of the Fiu β-barrel (Fig. 3A). In contrast, in crystal state 2, the entire polypeptide chain (amino acids 50–760) C-terminal of the TonB box (which is disordered in both crystal forms) could be modeled into the available electron density (Figs. 2A and 3A). The disorder of the plug domain loop in crystal state 1 opens a large cavity in the interior of Fiu to the external environment, whereas, in crystal state 2, this cavity is present but occluded in the lumen of the Fiu barrel (Fig. 2B). Interestingly, analysis of the structures of PiuA from A. baumannii and PiuA and PiuD from Pseudomonas aeruginosa revealed crystal states analogous to Fiu (Fig. 3, B and C) (30, 33). In PiuA from A. baumannii, all extracellular loops are ordered, with the plug loop occluding the entrance to an internal cavity (Fig. 3, B and C). In PiuA and PiuD from P. aeruginosa, the extracellular and plug loops are disordered, with the external cavity of PiuA exposed to the external environment (Fig. 3, B and C).</p><!><p>The crystal structure of Fiu reveals a large gated internal cavity. A, the crystal structure of fully ordered Fiu (crystal state 2), shown as a cartoon representation with rainbow colors running from N-terminal (blue) to C-terminal (red). B, cutaway outline representation of Fiu crystal structures, showing the internal cavity selectively occluded in crystal state 2 as well as the effect of removal of the labile subdomain of the TBDT plug on Fiu channel formation through the membrane. C, composite cutaway view of Fiu, showing the N-terminal plug domain as a cartoon, with the variably ordered plug loop (blue) and labile plug subdomain (red) highlighted. Outer membrane is abbreviated to OM in this figure.</p><p>Extracellular loop stability in crystal states of Fiu and PiuA/PiuD. A, stereo cartoon view of Fiu in crystal states 1 and 2, illustrating the variably ordered extracellular barrel loops 7, 8, and 9 (L7–L9) and the plug domain loop (PL). Loops are color-coded as follows: L7, salmon; L8, red; L9, brick red; plug domain loop, blue. Loop termini are shown as spheres where the loop is disordered. B, PiuA and PiuD crystal structures from P. aeruginosa and A. baumannii presented as for Fiu in A. The crystal structures of PiuA and PiuD display an analogous pattern of loop order/disorder to that observed for Fiu. C, cutaway outline representation of PiuA and PiuD, showing the presence of a selectively gated internal cavity. Outer membrane is abbreviated to OM in this figure.</p><!><p>Utilizing force spectroscopy, Hickman et al. (37) showed that the N-terminal plug domain of TBDTs consists of labile and nonlabile subdomains. Upon substrate binding, TonB is recruited to the TonB box at the N terminus of the TBDT and facilitates reversible displacement of the labile subdomain of the TBDT plug via mechanical energy provided by the proton-motive force (37, 38). In support of this study, it has been shown that a number of charged residues at the interface between the labile and nonlabile subdomains in TBDTs are important for substrate transport but not binding (39). Based on these studies, we identified that the labile subdomain of the Fiu plug extends from the N terminus of the protein to the start of the extracellular plug domain loop, which is selectively ordered in our crystal structures (Fig. 2C). In crystal state 2, removal of the labile subdomain opens the internal cavity of Fiu to the periplasm, but because of the presence of the plug loop, this does not create a membrane-spanning channel. In crystal state 1, removal of this subdomain opens a large channel between the periplasm and the external environment (Fig. 2B).</p><p>TBDTs selectively transport their substrate across the outer membrane while preventing antibiotics and other deleterious molecules from entering the cell (8). Therefore, it is likely to be undesirable for Fiu to exist in the open-channel state, which would result from simultaneous displacement of the plug subdomain and disorder of the plug domain loop. The internal cavity we observed in our structure, gated by the selectively ordered plug loop, may provide a solution to this problem. The internal cavity is large enough (∼3200 Å3) to accommodate a Fe–siderophore complex. If a siderophore entered this chamber prior to removal of the labile plug subdomain, then it could enter the periplasm without formation of a membrane-spanning channel through the pore of Fiu.</p><!><p>To determine the substrate-binding site of Fiu, we attempted cocrystallization and soaking of Fiu crystals in the presence of the monomeric catecholate compounds DHBS and 2,3-dihydroxybenzoic acid (DHB) in complex with Fe3+. These monomeric catecholates form a 3:1 complex with a single Fe3+ ion at the center. Despite the presence of DHBS at a high concentration (100–1000 μm) during crystallization screening and soaking and the resulting Fiu crystals diffracting well (2.8–2.0 Å), no electron density corresponding to DHBS was observed in the resulting density maps. Some crystals of Fiu grown in the presence of high concentrations of Fe–DHB (1 mm) exhibited the characteristic purple color of the Fe–catecholate complex (Fig. S3A). Although these crystals only diffracted to low resolution (anisotropic diffraction, 3.2–5.9 Å) (Table S4) Fo-Fc densities attributable to two Fe–DHB complexes were observed (Fig. S3B and Data S1). However, these Fe–DHB complexes were located on the side of the Fiu barrel, distal from the extracellular binding pocket, and were involved in crystal packing, suggesting that they are bound nonspecifically (Fig. S3C). Although it has been demonstrated previously that Fiu is capable of transporting DHB and DHBS in vivo (32), our inability to obtain a legitimate cocrystal structure suggests that Fiu has a low affinity for these compounds. As TBDTs generally bind their ligands with very high affinity (8), this suggests that monomeric catecholate compounds may not be the preferred substrate for Fiu, and its target siderophore remains unidentified.</p><p>To determine potential substrate-binding sites in Fiu, we performed in silico docking between Fiu and Fe–DHB using Autodock Vina (40). The rationale for this experiment is that, although Fe–DHB appears to be a low-affinity ligand, the ability of Fiu to transport it suggests that the high-affinity substrate for this transporter is likely to be a catecholate siderophore, which would contain a Fe3+–catecholate complex analogous to DHB. Thus, although the results should be interpreted cautiously, docking with Fe–DHB provides an indication of substrate-binding sites in Fiu. Two docking runs were performed (Data S1). For the first run, the entire extracellular portion of closed Fiu (crystal state 2) was defined as the search area. In this experiment, the majority of the solutions placed Fe–DHB in the internal chamber of Fiu, with the third most favored solution positioning Fe–DHB in the extracellular cavity of the protein (Fig. 4, A and B, and Fig. S4 and Table S5). Although the internal cavity would be inaccessible to Fe–DHB in the closed state, because of the fully ordered plug loop, it would be accessible in state 1, as this loop is disordered. In the second docking experiment, the internal cavity was excluded from the search area. In this experiment, all solutions placed Fe–DHB in the Fiu extracellular cavity, with the majority of solutions clustered at a single location (Fig. 4, A and C, and Fig. S4). Suggestively, the top-ranking solution from this docking run was identical to the third top solution from the first experiment. Taken together, these results suggest that Fiu possesses a binding site capable of accommodating a Fe–catecholate complex in its extracellular cavity. In addition, these data show that the internal cavity of Fiu is capable of accommodating the Fe–DHB complex.</p><!><p>Top-ranked docking solutions between Fiu and the Fe–DHB substrate. A, the location of the top-ranked docking modes in a cartoon representation of Fiu, for docking run 1 (R1M1), which includes the entire extracellular portion of the transporter, and docking run 2 (R2M1), in which the internal cavity is excluded from the search area. B, the same docking solutions as in A, with a cutaway composite view of Fiu.</p><!><p>To validate the external substrate-binding site identified in our docking analysis, we compared its location with the substrates of other TBDTs that have been structurally characterized. For this analysis, we selected 12 nonredundant TBDT–ligand structures (Table S6) (38, 41–49) and superimposed them with the structure of Fiu. In 11 of these 12 structures, the substrate bound in an analogous location to our Fe–DHB docking solution, with the metal ion of the respective ligands located between 2.8 and 9.5 Å from the Fe of Fe–DHB in our docked complex (Fig. 5). The 12th structure, the whose ligand did not colocalize with Fe–DHB, is the enterobactin transporter PfeA from P. aeruginosa (Fig. 5). In this structure, the binding of enterobactin in PfeA occurs on the external face of the extracellular loops of the transporter, which entirely encloses the entrance to the lumen of the barrel (42). It has been demonstrated that PfeA binds enterobactin via two-site binding and that the site observed in this crystal structure represents the initial substrate-binding site, with the second binding site located deeper in the transporter barrel (42). This two-site binding model is further supported by the structural analysis of FepA, a close homolog of PfeA, which shows that it binds enterobactin at two locations, one of which is analogous to the PfeA-binding site (50). Fiu and the other TBDTs analyzed lack the extracellular loop structure required to bind their substrates in this external location, and so it is likely that the binding site observed in PfeA and FepA is distinct from other TBDTs. These data suggest that TBDTs share a common substrate-binding site that is analogous to the Fiu external binding site identified by the docking analysis, providing validation of this result.</p><!><p>The location of the Fiu putative external Fe–DHB binding site compared with other TBDT substrate complexes. A, the location of the Fe–DHB complex docked with Fiu compared with that of other substrates bound to in the crystal structures of superimposed TBDTs. Fiu is shown as a cartoon rainbow and in the same view as a white surface representation below. The location of the metal centers of different TBDT substrates are shown as colored spheres and labeled. B, the Fiu Fe-DHB docked complex shown as in A but in a different orientation. Representative distances between Fe–DHB and the TBDT substrate metal ions are shown, colored as in A.</p><!><p>To determine the role of the amino acids that define the putative Fiu external substrate binding site, we assessed the functionality of variants of Fiu with mutations in this region (Fig. 6 and Fig. S5). Small side chains around the cavity (contributed by alanine, serine, and threonine) were mutated to the bulky amino acid tryptophan to sterically occlude the binding pocket, whereas larger side chains defining the pocket were mutated to alanine. Plasmid-borne constructs of these mutant Fiu proteins were transformed into the E. coli BW25113 strain Δ6 (Fig. 1, B and C). To test for the restoration of iron transport activity, the transformed strains were streaked onto LB agar with 2,2′-bipyridine and scored for growth (Fig. 6 and Fig. S5). Mutation of phenylalanine 105 (F105A), glutamate 108 (E108A), and arginine 142 (R142A) grossly affected the function of Fiu (Fig. 6 and Fig. S5). Fiu(E108A) was nonfunctional, displaying growth identical to the negative control. Fiu(F105A) and Fiu(R142A) displayed minimal complementation (Fig. 6 and Fig. S5). Two other mutations, threonine 113 to tryptophan (T113W) and serine 139 to tryptophan (S139W), also exhibited some defect in function compared with WT Fiu (Fig. 6 and Fig. S5).</p><!><p>The effect of mutations in the Fiu external substrate-binding site on the function of Fiu in vivo. A, bar graph indicating the effect of Fiu binding site mutations of the ability of pBAD24Fiu to complement E. coli BW25113 Δ6. The length of the bar graph indicates the maximum concentration of 2,2′-bipyridine at which growth of the complemented strain was observed on solid LB agar; experimental data are shown in Fig. S5. B, stick and cartoon representation of the Fiu extracellular loops, showing the location of the residues in the putative substrate-binding site subjected to mutagenesis. C, magnified view of the Fiu substrate-binding site shown in B, rendered as a surface model with mutated residues labeled (left) and Fe–DHB shown (right). D, stereo image showing the residues mutated in the Fiu external binding site as a stick representation. The Fe–DHB complex is shown as a line and sphere representation for DHB and Fe3+, respectively. Colors are consistent through all panels and indicate the effect of mutagenesis on Fiu function: brick red, inactivation or significant defect in Fiu function; pink, minor defection in Fiu function; green, no defect in Fiu function.</p><!><p>It is possible the Fiu mutant variants generated in this experiment may have reduced expression or stability, leading at least partially to the observed phenotypic differences in this assay. However, none of the amino acids mutated play a structural role in Fiu, and the TBDT fold is highly stable, making it relatively unlikely that these mutations would have a major effect on Fiu expression or stability. Based on this assumption, we offer the following analysis. Residues Phe-105 and Glu-108 are located on plug domain loop 1 within the predicted Fe–DHB binding pocket (Fig. 6, B and C). Phe-105 is ordered in both crystal states and forms a hydrophobic pocket that shelters the aromatic carbons of one of the DHB monomers of the Fe–DHB complex in our docked structure. Glu-108 is disordered in the open state of Fiu and is not within bonding distance of docked Fe–DHB (5.3 Å) but may form interactions with polar substrate groups in vivo or generally stabilize the binding pocket (Fig. 6, B and D). Arg-142 is minimally surface-exposed and unlikely to interact directly with catecholate substrates (Fig. 6, B and C). However, it is located in plug domain loop 2 and forms hydrogen bonds with the carbonyl groups of amino acids Asp-136 and Gly-138. These interactions stabilize this loop, which defines the lower section of the putative substrate-binding pocket, and this may account for the phenotypic effect of the mutation. Mutation of Asn-111, Tyr-337, and Lys-368, which directly interact with Fe–DHB in our docked structure, did not affect Fiu function in our assay (Fig. 6). This may reflect a lack of precision of our docked model or the fact that single point mutations of substrate-interacting residues may be insufficient to affect substrate import, as reported for other TBDTs (41). Given that several single mutations in this region dramatically effect Fiu function, these data provide evidence that the external binding site identified in our docking is important for Fiu function.</p><!><p>E. coli BW25115 possesses three outer-membrane TBDTs known to be responsible for the transport of Fe–catecholate complexes: Fiu, FepA, and Cir (32, 51). In this work, we show that, despite sharing substrates related by a catecholate functional group, Fiu is only distantly related to FepA and Cir. This demonstrates that TBDTs from different lineages have the ability to transport similar substrates, and they may have arrived at this specificity via convergent evolution.</p><p>FepA is a high-affinity transporter for enterobactin, a siderophore endogenously produced by E. coli (31, 36). Although FepA transports enterobactin, like Fiu and CirA, it is also able to mediate the import of iron in complex with monomeric catecholates, albeit with a lower affinity (32). As Fiu does not transport enterobactin, it is unclear whether its major physiological role is transport of iron in complex with monomeric catecholate molecules or import of a so far unidentified xenosiderophore. In this work, we show that Fiu supports growth of E. coli in the absence of an exogenously supplied substrate, likely through import of Fe–catecholate compounds generated through breakdown of enterobactin. However, the ability of Fiu to transport iron under these conditions is inferior to that of FepA and does not support growth under stringent iron limitation. The relatively poor ability of Fiu to transport iron under these conditions suggests that transport of enterobactin breakdown products is a secondary function for this transporter, leaving its high-affinity substrate to be identified. This hypothesis is further supported by our inability to obtain a bona fide complex between Fiu and either Fe–DHB or Fe–DHBS, despite identifying conditions that were clearly compatible with substrate crystallization.</p><p>The crystal structure and associated analysis we present further reinforce the differences between Fiu and FepA. FepA and its closely related homolog PfeA from P. aeruginosa initially bind enterobactin at an external binding site formed by extracellular loops that entirely enclose the entrance to the lumen of the transporter's β-barrel (42, 50). Fiu lacks the extracellular loop structure required for this external binding site, and our docking analysis suggests that it binds its substrate deep in its internal cavity, at a location shared with other TBDTs (38, 41–49).</p><p>Structural analysis of Fiu reveals that a number of extracellular loops are selectively ordered and that this selective order is shared among the homologous receptors PiuA and PiuD (30, 33). The observed crystal states may represent conformational transitions the receptor undergoes during substrate binding and transport. Specifically, the external plug loop of Fiu in our structures selectively gates a large cavity in the lumen of the transporter. Our docking shows that this cavity is large enough to accommodate a Fe–catecholate siderophore complex. Additionally, based on predictions of the labile subdomain of the N-terminal plug (37, 39), this cavity is in the path of substrate transport (Fig. 2C). This selectively ordered Fiu plug loop may allow its substrate to enter the internal cavity prior to removal of the labile subdomain. If the external plug loop then adopted an ordered state during subdomain removal by TonB, it would prevent channel formation through Fiu during import and prevent nonspecific import of deleterious substances. Despite their significant structural differences and distinct initial binding sites, the essence of this mechanism may be shared with FepA, which possesses a second enclosed binding site deeper in the barrel of the transporter (42, 50).</p><p>Gram-negative bacteria can exhibit formidable resistance to antibiotics, largely because of the protective semipermeability of the outer membrane (2, 52, 53). Catecholate-containing antibiotics use molecular mimicry to facilitate their import into bacteria via Fiu and related transporters (25). As these transporters are present in diverse bacterial groups, siderophore-mimicking antibiotics are attractive lead compounds for the development of new therapeutics to treat Gram-negative bacterial infections, representing an area of considerable recent interest for antibiotic development (28, 29, 54). In A. baumannii and P. aeruginosa, PiuA has been shown to greatly promote susceptibility to the 3,4-dihydroxypyridine–containing sulbactam BAL30072 (30, 55). By investigating the structural basis of substrate binding and import by Fiu, this work will assist with the exploitation of Fiu and related transporters as a conduit for catecholate-containing antibiotics into the bacterial cell.</p><!><p>To determine the phylogenetic relationship between Fiu and TBDTs of know structure and/or function, a subset of TBDT sequences were selected, and these sequences were obtained from the NCBI database. Sequences were aligned using the Clustal algorithm, and the alignment was utilized to build a bootstrapped phylogenetic tree (100 repetitions), which was visualized using the FigTree software (56).</p><p>A HMMER search using the Fiu sequence from E. coli BW25113 as the search query was established using the stable and unbiased proteome dataset RP55 (57, 58). The search was restricted to sequences with an E-value of less than 1e−75, limiting the outcome to 502 sequences (Table S1). For sequence clustering, classification used all-against-all BLAST clustering based on pairwise similarities and visualized with CLANS (59), with an E-value cutoff of 1 × 10−120. To assess the similarity of sequences identified in this search to other TBDTs present in E. coli BW25113, sequences for FhuA, FecA, CirA, FepA, BtuB, YddB, and YncD were added to this dataset, and clustering was performed with an E-value cutoff of 1 × 10−120.</p><!><p>E. coli BW25113 mutants were created using the λ-red system (60). Kanamycin resistance cassettes flanked by 300 bp of genomic DNA either side of the genes encoding TBDTs of interest were amplified using specific mutants from the E. coli mutant Keio collection (61) as templates. Primers utilized are summarized in Table S7.</p><p>The host strain E. coli BW25113 was transformed with the λ-red recombinase plasmid pKD46 (60), grown at 30 °C (LB broth and 100 μg·ml−1 ampicillin) to an A600 nm of 0.1 before λ recombinase was induced by addition of 0.2% l-arabinose. Thereafter, cultures were grown at 30 °C until A600 nm 0.6–0.8 and transformed using the room temperature electroporation method (62). Briefly, bacterial cells were isolated by centrifugation at 3000 × g for 3 min and washed twice with a volume of sterile 10% glycerol equal to the volume of the culture used. Cells were then resuspended in 10% glycerol to a volume of 1:15 that of the original culture. 100–500 ng of PCR-amplified KanR KO cassette for the gene of interest was then added to 100 μl of the resuspended bacteria, and the mixture was electroporated. 1 ml of LB broth was added to the cells after electroporation, and the culture was recovered at 37 °C for 1 h before plating onto LB agar and 30 μg·ml−1 kanamycin. PCR was used to validate that colonies did indeed have the KanR cassette in place of the gene of interest.</p><p>To remove the KanR cassette, deletion mutant strains were transformed with the plasmid pCP20 (63) containing the "flippase cassette." Cells were grown under either ampicillin (100 μg·ml−1) or chloramphenicol (30 μg·ml−1) selection to maintain the plasmid. For removal of the KanR cassette, a single colony of the mutant strain was used to inoculate 1 ml of LB broth (no selection). The culture was grown overnight at 43 °C to activate expression of the flippase gene. This culture was then subject to 10-fold serial dilution in sterile LB and plated onto LB agar with no selection. The resulting colonies were patched onto LB agar containing kanamycin, chloramphenicol, or no selection. PCR was used to validate colonies unable to grow on kanamycin or chloramphenicol; those able to grow in the absence of selection did indeed represent successful removal of the KanR cassette. This process was repeated sequentially to derive strains multiply defective in up to six TBDT receptors. The order of deletion was ΔFhuA, ΔFecA, ΔCirA, ΔFepA, ΔFhuE, and then ΔFiu. Mutant strains created in the process were designated TBDT Δ1, Δ2, Δ3, Δ4, Δ5, and Δ6, based on the number of receptors deleted. Strains created are listed in Table S8.</p><!><p>E. coli BW25113 deletion strains (Δ4, Δ5, and Δ6) grew poorly on LB agar. To ameliorate this phenotype, all mutant strains were maintained on LB agar and 250 μm Fe(II)SO4. All deletion strains grew well under these conditions. To test the ability of mutant strains to grow under iron-limiting conditions, strains were grown in LB broth until stationary phase. Cells were harvested from 0.5 ml of this stationary-phase culture, and the supernatant was removed. Cells were resuspended in 0.5 ml of 1× M9 salts, and a minimal quantity of this suspension was streaked onto LB agar containing 0–150 μm 2′2-bipyridine. Plates were incubated at 37 °C overnight, and growth was observed and scored.</p><!><p>The ORF for Fiu, including the sequence encoding the signal peptide, was amplified from E. coli BW25113 by PCR (Table S7) and cloned into the pBAD24 plasmid at EcoRI and HindIII restriction sites. The resulting vector, designated pBAD24Fiu, was then transformed into E. coli BW25113 Δ6 and maintained using 100 μg·ml−1 ampicillin. To test for complementation, E. coli BW25113 TBDTΔ6 pBAD24Fiu was streaked onto LB agar and 0.2% arabinose, 100 μg·ml−1 ampicillin, and 0–120 μm 2,2′-bipyridine. Growth under these conditions was compared with that of E. coli BW25113 Δ6 containing pBAD24 as a vector control. Mutations of the putative Fiu substrate-binding site were created via whole-plasmid mutagenesis using pBADFiuCom as the starting vector (64). The mutations were introduced using the primer sequences provided in Table S7. The sequence of the pBAD24Fiu template and introduction of the specified mutations in the resultant plasmids were confirmed by Sanger sequencing. Mutant plasmids were transformed into E. coli BW25113 TBDTΔ6, maintained, and tested for function as described above for pBAD24Fiu.</p><!><p>DNA encoding the mature form of Fiu lacking the signal peptide was amplified from E. coli BW25113 using the primers shown in Table S7. NcoI and XhoI restriction sites incorporated into the primers were used to clone the DNA fragment into a modified pET20b vector with a 10× N-terminal His tag followed by a TEV cleavage site. The resulting plasmid was transformed into E. coli BL21 (DE3) C41 cells, and protein expression was induced in cultures grown in terrific broth (12 g of tryptone, 24 g of yeast extract, 61.3 g of K2HPO4, 11.55 g of KH2PO4, and 10 g of glycerol) with 100 mg·ml−1 ampicillin for selection. Cultures were grown at 37 °C until A600 of 1.0, induced with 0.3 mm isopropyl 1-thio-β-d-galactopyranoside, and grown for a further 14 h at 25 °C. Cells were harvested by centrifugation, lysed with a cell disruptor (Emulseflex) in lysis buffer (50 mm Tris, 200 mm NaCl, and 2 mm MgCl2) plus 0.1 mg·ml−1 lysozyme, 0.05 mg·ml−1 DNase1, and Complete protease mixture inhibitor tablets (Roche). The resulting lysate was clarified by centrifugation at 20,000 × g for 10 min, the supernatant was then centrifuged for a further 1 h at 160,000 × g to isolate membranes. The resultant supernatant was decanted, and the membrane pellet was suspended in lysis buffer using a tight-fitting homogenizer. When homogenized, the membrane fraction was solubilized by addition of 10% Elugent (Santa Cruz Biotechnology) and incubated with gentle stirring at room temperature for 20 min. Solubilized membrane proteins were clarified by centrifugation at 20,000 × g for 10 min. The supernatant was applied to Ni-agarose resin equilibrated in Ni binding buffer (50 mm Tris, 500 mm NaCl, 20 mm imidazole, and 0.03% dodecyl maltoside (DDM) (pH 7.9)). The resin was washed with 10–20 column volumes of Ni binding buffer before elution of the protein with a step gradient of 10%, 25%, 50%, and 100% Ni gradient buffer (50 mm Tris, 500 mm NaCl, 1 m imidazole, and 0.03% DDM pH 7.9)). Fiu eluted at 50% and 100% gradient steps. Eluted fractions containing Fiu were pooled and applied to a 26/600 S200 Superdex size exclusion column equilibrated in SEC buffer (50 mm Tris, 200 mm NaCl, and 0.03% DDM pH 7.9)). To exchange Fiu into octyl β-d-glucopyranoside (βOG) for crystallographic analysis, fractions from SEC containing Fiu were pooled and applied to Ni-agarose resin equilibrated in βOG buffer (50 mm Tris, 200 mm NaCl, and 0.8% octyl β-d-glucopyranoside (pH 7.9)). The resin was washed with 10 column volumes of βOG buffer before elution with βOG buffer and 250 mm imidazole. Fractions containing Fiu were pooled, and 1 mg·ml−1 His6-tagged TEV protease and 1 mm DTT were added. This solution was then dialyzed against βOG buffer at 4–6 h at 20 °C to allow TEV cleavage of the His10 tag and removal of excess imidazole. The solution was then applied to Ni-agarose resin to remove TEV protease and the cleaved polyhistidine peptide. The flow-through containing Fiu from this step was collected concentrated to 14 mg·ml−1 in a 30-kDa cutoff centrifugal concentrator, snap-frozen, and stored at −80 °C.</p><!><p>Purified Fiu in βOG buffer was screened using commercially available crystallization screens (∼600 conditions). Hexagonal crystals formed in the JCSG screen in 1 m LiCl, 20% PEG 6000, and 0.1 m trisodium citrate (pH 4.0) (65). These crystals were looped and the drop solution was removed, and they were flash-cooled and stored in liquid N2. These crystals diffracted poorly (>3.3 Å) and suffered from considerable anisotropy. To improve diffraction, Fiu under the above condition was subjected to an additive screen (Hampton Research). Hexagonal crystals grew with many additives; however, in the presence of 5% polypropylene glycol P400 (PPG 400), flat, diamond-shaped plates formed. These crystals were looped, the mother liquor was removed by wicking, and they were flash-cooled and stored in liquid N2 at 100 K. Data were collected at the Australian Synchrotron, with crystals diffracting to 2.1 Å in the space group C2221. Despite relatively low sequence identity (33%) between Fiu and the catecholate receptor PiuA from A. baumannii (PDB code 5FP1), a molecular replacement solution was obtained using Phaser, with the crystal structure of the PiuA receptor as a starting model (30, 66). The model was built and refined using the Phenix package and Coot (67, 68). The majority of the Fiu polypeptide chain could be modeled into the available density; however, loops 7–9 and the plug domain loop were disordered in this structure. The Fiu model from these crystals was designated crystal state 1.</p><p>To obtain Fiu in additional crystal states, purified Fiu in βOG buffer with 5% PPG 400 was rescreened for crystallization (∼600 conditions). Fiu crystallized under multiple addition conditions in the presence of PPG 400. Using crystals looped and frozen directly from these screens, the structure of Fiu was solved in two further crystal forms: the P1 form at 2.9 Å in 20% PEG 3350, 0.2 m Na2 malonate, and 0.1 m BisTris propane (pH 8.5) and the C21 form at 2.5 Å in 0.1 m Tris, 20% PEG 6000, and 0.2 m NaCl (pH 8.0). In these crystal forms, all loops of Fiu were ordered and conformationally analogous, allowing complete tracking of the Fiu sequence amino acids 50–760. Fiu modeled from these crystals was designated crystal state 2. Cocrystallization between Fiu and Fe–DHB or Fe–DHBS was performed by adding 300 μm to 1 mm of these compounds to purified Fiu in βOG buffer with or without 5% PPG 400 to at a final concentration of ∼100 μm (8 mg/ml). Crystal screening was performed as above. Crystals were harvested directly from screening trays, mother liquor was removed by wicking, and crystals were cryocooled in liquid N2 at 100 K. For soaking experiments, Fiu crystals from crystal states 1 and 2 were transferred to crystallization solution containing 1 mm Fe–DHB or Fe–DHBS and incubated in this solution for 1–5 min prior to cryocooling. Crystallization of Fiu, 5% PPG 400, and 1 mm Fe–DHB yielded purple crystals in the JCSG screen (65) condition (0.1 m Tris (pH 8.5), 20% (w/v) PEG 8000, and 0.2 m MgCl2). These crystals were looped, wicked to remove mother liquor, and flash frozen as above. These crystals diffracted modestly, with the best crystal diffracting anisotropically to 3.2–5.9 Å. Molecular replacement with refined Fiu was performed on this dataset to identify the location of the Fe–DHB complex. RMSD calculations were performed using the RMSD tool in the Align command in PyMOL.</p><!><p>To determine potential ligand bindings sites in the Fiu crystal structure, an in silico docking approach was applied using Autodock Vina in the Chimera software package (40, 69). Coordinates for iron in complex with three molecules of 2,3-dihydroxybenzoic acid (Fe–DHB) was obtained from PDB code 3U0D, the structure of human Siderocalin bound to the bacterial siderophore 2,3-DHBA. Coordinates for Fiu were taken from molecule A of the C21 crystal form in crystal state 2. Ligand and Fiu coordinates were imported into Chimera and optimized using the Dock Prep utility. Docking was performed using the Autodock Vina dialog, and two box sizes were utilized for docking (box 1 = 46.5, 56.0, 45.5 Å; box 2 = 46.5, 35.2, 45.5 Å). Box 1 encompassed the entire extracellular portion of Fiu and all cavities accessible from the extracellular environment. Box 2 excluded the large cavity in Fiu, gated by the extracellular plug domain. A total of nine binding modes were sought for each docking run, with search exhaustiveness of between eight and 300 and a maximum energy difference of 3 kcal/mol. Docking solutions did not differ significantly as a result of changes to search exhaustiveness. Docking solutions were inspected visually, and the highest-rated solution was used for the main figures and for "Discussion."</p><!><p>The PDB was searched manually for structural coordinates of TBDT in complex with substrate compounds. TBDT receptor complexes were aligned to the crystal structure of Fiu (Table S6) based on the TBDT chain using the Super command in PyMOL. The location bound substrates were determined by manual inspection in PyMOL. The location and volume of Fiu substrate binding cavities were estimated using CASTp (70).</p><!><p>R. G. and T. L. conceptualization; R. G. and T. L. resources; R. G. data curation; R. G. formal analysis; R. G. and T. L. funding acquisition; R. G. investigation; R. G. visualization; R. G. methodology; R. G. writing-original draft; R. G. project administration; R. G. and T. L. writing-review and editing; T. L. supervision.</p><!><p>This work was supported by the Australian Research Council (FL130100038) and the National Health and Medical Research Council (NHMRC Program in Cellular Microbiology, 1092262). The authors declare that they have no conflicts of interest with the contents of this article.</p><p>This article contains Figs. S1–S5, Tables S1–S8, Data S1, and references.</p><p>TonB-dependent transporter</p><p>ferric iron uptake</p><p>2,3-dihydroxybenzoyl-l-serine</p><p>root mean square deviation</p><p>2,3-dihydroxybenzoic acid</p><p>tobacco etch virus</p><p>dodecyl maltoside</p><p>octyl β-d-glucopyranoside</p><p>polypropylene glycol</p><p>lysogeny broth.</p>

PubMed Open Access

Independent Generation and Time-Resolved Detection of 2\xe2\x80\x99-Deoxyguanosin-N2-yl Radical

Guanine radicals are important reactive intermediates in DNA damage. Hydroxyl radical (HO\xe2\x80\xa2) has long been believed to react with 2\xe2\x80\x99-deoxyguanosine (dG) generating 2\xe2\x80\x99-deoxyguanosin-N1-yl radical (dG(N1-H)\xe2\x80\xa2) via addition to the nucleobase \xcf\x80-system and subsequent dehydration. This basic tenet was challenged by an alternative mechanism, in which the major reaction of HO\xe2\x80\xa2 with dG was proposed to involve hydrogen atom abstraction from the N2-amine. The 2\xe2\x80\x99-deoxyguanosin-N2-yl radical (dG(N2-H)\xe2\x80\xa2) formed was proposed to rapidly tautomerize to dG(N1-H)\xe2\x80\xa2. We report the first independent generation of dG(N2-H)\xe2\x80\xa2 in high yield via photolysis of 1. dG(N2-H)\xe2\x80\xa2 is directly observed upon nanosecond laser flash photolysis (LFP) of 1. The absorption spectrum of dG(N2-H)\xe2\x80\xa2 is corroborated by DFT studies, and anti- and syn-dG(N2-H)\xe2\x80\xa2 are resolved for the first time. The LFP experiments showed no evidence for tautomerization of dG(N2-H)\xe2\x80\xa2 to dG(N1-H)\xe2\x80\xa2 within hundreds of microseconds. This observation suggests that the generation of dG(N1-H)\xe2\x80\xa2 via dG(N2-H)\xe2\x80\xa2 following hydrogen atom abstraction from dG is unlikely to be a major pathway when HO\xe2\x80\xa2 reacts with dG.

independent_generation_and_time-resolved_detection_of_2\xe2\x80\x99-deoxyguanosin-n2-yl_radical

Introduction<!>Design and synthesis of a photochemical precursor (1) for dG(N2-H)\xe2\x80\xa2.<!>Photochemical generation of dG(N2-H)\xe2\x80\xa2 and product studies.<!>dG(N2-H)\xe2\x80\xa2 characterization by laser flash photolysis and DFT studies.<!>Conclusion

<p>Nucleic acid oxidation is central to human health. For instance, it is a factor in aging and in the development of cancer.[1] DNA is also the molecular target for a variety of cancer treatments. Many of these methods involve one electron oxidation of DNA, with ionizing radiation being the most common.[2] Ionizing radiation damages DNA directly by ionizing DNA and indirectly by ionizing water, which generates hydroxyl radical (HO•), a highly reactive DNA damaging species.[3] 2'-Deoxyguanosine (dG) is the most readily oxidized of the 4 native nucleosides and is also a primary contributor to electron transfer in one-electron oxidized DNA.[4] Consequently, the reactive intermediates produced upon dG oxidation, and their reactivity, have been the focus of important theoretical studies and experimental investigations for the past 30 years.[5] Pulse radiolysis has been used extensively to characterize the early, rapid dG oxidation events that are complete on the sub-millisecond timescale.[5c] dG(N1-H)• is the major and thermodynamically most stable intermediate generated by reaction of dG with HO• (Scheme 1).[6] However, the mechanism(s) by which dG(N1-H)• is formed is controversial and was recently proposed to arise by tautomerization of dG(N2-H)• (Scheme 1).[7] We have resolved this problem by using near-UV photolysis of a synthetic precursor to dG(N2-H)• in conjunction with time-resolved spectroscopy, time-dependent DFT calculations and product studies.</p><p>It is widely accepted that HO• reacts with dG yielding dG(N1-H)• via an addition-elimination mechanism. HO• adds to C4, C5 and C8 atoms of guanine. C4-OH (Scheme 1) is proposed to be the major product, accounting for 60–70% of the reactions.[6a] Computational studies indicate that upon barrierless addition of HO•, C4-OH produces dG(N1-H)• via loss of hydroxide to form an ion pair, followed by N1-deprotonation.[5b] Calculations predict that ion pair formation encounters ~6.5 kcal/mol barrier and is the rate determining step. N1-deprotonation within the ion pair is kinetically and thermodynamically favored (>7 kcal/mol) over the N2-position. The calculated energy difference between the radicals is basis set dependent, but dG(N1-H)• is 2–4 kcal/mol more stable than dG(N2-H)•.[8]</p><p>The HO• pathway to dG(N1-H)• through C4-OH was challenged by a series of pulse radiolysis experiments carried out on various guanine derivatives.[7] The authors posited that HO• preferentially abstracts the N2-hydrogen atom, instead of adding to the p-bond to generate C4-OH (Scheme 1). Initially formed dG(N2-H)• was proposed to rearrange to the more stable dG(N1-H)• with a < 30 μs half-life at 298 K and an activation energy of ~5.5 kcal/mol. The feasibility of the hydrogen atom abstraction of this alternative mechanism was corroborated by DFT calculations.[5b] Pulse radiolysis, in conjunction with spectroscopic detection (and other measurements), is a powerful approach for studying reactive intermediate chemistry. However, one limitation is that multiple reactive intermediates can be produced. This can potentially complicate analysis, particularly if reactive intermediates have overlapping spectral properties. To simplify the examination of purine radicals and resolve mechanistic conflicts in the aforementioned studies, we designed a photochemical precursor (1) that generates a single purine intermediate, dG(N2-H)•. We thoroughly examined the reactivity of this radical by product analysis following UV-photolysis of 1 and time-resolved LFP experiments that corroborate each other. In addition to refuting the sub-millisecond tautomerization of dG(N2-H)•, we also distinguished the anti- and syn-conformers of dG(N2-H)• for the first time.</p><!><p>We previously generated dG(N2-H)• and 2'-deoxyadenosin-N6-yl radical (2) from the corresponding diphenyl hydrazines (e.g. 3, Scheme 2).[9] However, photochemical conversion of 3 to dG(N2-H)• is too inefficient for laser flash photolysis examination of dG(N2-H)•. More recently, we reported on a method for generating 2'-deoxyadenosin-N6-yl radical (2) from a ketone precursor (4, Scheme 2).[10] Upon photolysis, 4 undergoes Norrish Type I photocleavage followed by rapid b-fragmentation. Using acetone as triplet photosensitizer greatly accelerated the conversion of 4 and allowed us to obtain the spectrum of 2. However, photosensitization of 3 by ketones was not attempted, because photodissociation of tetraphenylhydrazine occurs from the excited singlet.[11] Furthermore, we anticipated that the ketones, which photo oxidize 8-oxodGuo, would do the same to the more readily oxidizable 3.[9, 12] We rationalized that 1 would yield dG(N2-H)• via the analogous cascade of reactions that 4 undergoes upon photolysis, and could also be sensitized by ketones (Scheme 2). The synthetic approach to.1 was strongly influenced by that of 4 and started from previously reported 5 (Scheme 3).[10a, 13] Substitution of the bromide in 5 by hydroxylamine (6) was slower than the analogous reaction in the synthesis of 4 due to the increased electron density of guanine. The formation of 6 required higher temperature than that for the dA analogue, and dioxane was substituted for THF, because of its higher boiling point. Introduction of the hindered ketone 8 using previously reported conditions, in which NaH was used as base led to the formation of an undesired product. Subsequently, the substitution was successfully carried out using Cs2CO3 as base. The desired ketone (1) was then obtained via standard debenzylation and desilylation conditions.</p><p> </p><!><p>With an eye on utilizing 1 as a source of dG(N2-H)• in DNA, we carried out photolyses in Pyrex vessels using lamps whose maximum output is at 350 nm. Although the λmax for 1 occurs in a far shorter region (λmax = 260 nm, ε = 1.22 × 104 M−1s−1 in H2O) than where these lamps emit, the absorption band tails above 300 nm (Figure S5). The quantum yield for disappearance of 1 under these conditions, measured using 2-hydroxy-2-methylpropiophenone as an actinometer, was similar (F = 1.8 ´ 10−3) to that of 4 (F = 1.5 ´ 10−3).[10a, 14] However, the weaker absorbance of 1 above 300 nm resulted in less efficient photochemical conversion than 4, such that only ~10% of the ketone was consumed following 8 h direct irradiation. Inspired by the sensitization used during the photolysis of 4, we used acetone (2% v/v,150 mM) to sensitize the reaction and achieved increased conversion to 62.4 ± 0.9% after 8 h in the presence of PhSH as reducing agent. The yield of dG and mass balance were high when Fe2+ or PhSH were used as a reducing agent (Table 1). In contrast to 2, bmercaptoethanol (BME) also effectively trapped dG(N2-H)• producing dG.[10a] We attribute this difference to the fact that guanine is more electron-rich than adenine. The corresponding nitrogen radical (dG(N2-H)•) is less electrophilic than 2, and encounters lower energy barriers when reacting with electronegative thiol hydrogen atom donors. These observations are consistent with our hypothesis regarding the polarity matching between hydrogen atom donors and neutral purine radicals.[10a] Purine electronic properties also manifest themselves when photolysis is carried out in the absence of a reducing agent. Under these conditions the mass balance and dG yield decrease almost two-fold but the conversion rate of 1 is approximately twice as high as when a reducing agent added (Figure S6). This is consistent with the proposal that the radical precursor can serve as a reducing reagent for dG(N2-H)•, a pathway that will be significant in laser flash photolysis (LFP) experiments described below.</p><p>Although acetone photosensitizes 1, comparison to the sensitizer's effect on 4 indicated that photochemical conversion would be too low for LFP studies.[10a] Consequently, we considered photolysis in the presence of other photosensitizers. Acetophenone was a promising candidate due to its efficient intersystem crossing, long triplet lifetime, and relatively high triplet energy.[15] Anaerobic photolyses were carried out in the presence of acetophenone (1% v/v, 86 mM), and the consumption of the precursor was significantly accelerated. The sensitization efficiency is proportional to the percentage of acetonitrile in phosphate buffer (Table 2). The source of this solvent effect is uncertain but it is unlikely that it is due to the decreased energy of π, π* excited triplet state of acetophenone in more polar solvent mixtures.[16] In the presence of 100 mM BME, the photolysis of 1 quantitatively yielded dG, indicating that sensitization was not detrimental to the fidelity of the photochemistry.</p><!><p>In light of the observation that photosensitized photolysis of 1 is a high fidelity source of dG(N2-H)•, this system was used to directly observe the latter via transient absorption spectroscopy. Rich transient features are observed upon nanosecond pulses (355 nm) of solutions of 1 (1 mM) and acetophenone (30 mM) in aqueous buffer (pH 7.0)/acetonitrile (1:1, v:v). Following 355 nm excitation, the triplet acetophenone absorption band at ~340 nm is immediately observed.[17] This transient decays within 4 μs (Figure 1a, S7), and is accompanied by a build-up of a strong absorption band with maxima at 610 nm and 650 nm. We attribute these observations to photosensitization of 1 by triplet acetophenone and the resulting reactions that give rise to dG(N2-H)•. The timescale for the growth of dG(N2-H)• is consistent with laser flash photolysis and computational studies on the formation of 2 via b-fragmentation following Norrish Type I photocleavage of 4 (Scheme 2).[10a] The transient feature centered at 610 nm is consistent with the previously reported absorption of N1-MedG(N2-H)• from N1-MedG.[5c, 7d, 18] The red-shifted peak at 650 nm has not been reported in relevant studies.</p><p>A variety of possible molecules responsible for the peak at 650 nm were considered. The assignment of this peak to the triplet excited state of 1, or the aminoxy alkyl radical intermediate 10 resulting from Norrish Type I cleavage analogous to that produced from 4 were ruled out by calculations. DFT calculations on 10 and the triplet excited state of 9 (the analogue of 1 lacking a 2'-deoxyribose, Figure 2) indicate that neither absorbs above 600 nm (Table S2). The observation that the growth and decay kinetics of the 650 nm peak are essentially identical to that of the 610 nm peak (Figure 1), suggested that they belong to very similar species. Given that product studies indicate that 1 produces dG(N2-H)• with high fidelity, we postulated that the peaks at 610 and 650 nm belong to different conformational isomers of dG(N2-H)• in which the remaining N2 hydrogen is either syn or anti with respect to the guanine N3 atom (Figure 2). (The naming convention is that utilized by Sevilla.[19]) The optical spectra (Figure 3a) for these two conformers in analogues lacking the deoxyribose ring (syn-, anti-Gua(N2-H)•) were calculated by TDDFT-B3LYP/6–311++G(d,p) in vacuum and under PCM. To compare with experiments, the TDDFT calculated absorption maxima for guanine radicals usually require being red-shifted by 40 – 70 nm.[6b] The calculated spectra for each Gua(N2-H)• conformer features an intense absorption band above 600 nm and weaker band in the UV (Figure 3a). The calculations reproduce the main feature of the experimental spectra for dG(N2-H)•. In addition, it is found that the λmax of the main absorption band > 600 nm is different for these two conformers (Table S2). Under PCM solvation model, the calculated lmax (after adding 60 nm) for the syn-conformer (632 nm) is red-shifted relative to the anti-conformer (616 nm). In vacuum, the calculated λmax (after adding 40 nm) for syn-Gua(N2-H)• is 652 nm and that for the anti-conformer is 612 nm. The theoretically predicted absorption wavelengths for the two conformers match the experimental spectrum obtained from the photolysis of 1 (Figure 1b), indicating that the peak with λmax = 650 nm is ascribable to syn-dG(N2-H)•, which is partially resolved from anti-dG(N2-H)• λmax = 610 nm. Experiments in G-quadruplex DNA corroborate the predicted spectral dependence upon dG(N2-H)• conformation.[20] When dG(N2-H)• is produced from the deprotonation of dG•+ in G-quadruplex structure, the observed spectrum centered at ~600 nm is consistent with anti-dG(N2-H)•. Selective formation of anti-dG(N2-H)• in the G-quadruplex is attributed to preferential syn-N2 deprotonation, which does not disrupt Hoogsteen hydrogen bonding.[20] The predicted conformational dependence of guanine radical λmax is also consistent with calculations on N9-methyl guanine radicals.[20–21]</p><p> </p><p>Our calculations based on the DFT/B3LYP/6–311++G(d,p)//PCM method show that syn-Gua(N2-H)• is ~ 0.1 eV (~ 2.3 kcal/mol) lower in energy than the anti-conformer, which is in general agreement with the literature (3.0 kcal/mol at the level of B3LYP/6–31G(d)//PCM).[19] In the presence of five explicit waters and under PCM solvation model (Figure S8), the B3LYP/6–311++G(d,p) calculated energy difference between the syn- and anti-Gua(N2-H)• is further reduced to 0.064 eV (~ 1.5 kcal/mol). The small energy difference means that the two conformers of dG(N2-H)• may coexist in aqueous solution. Previously, N1-MedG(N2-H)• was produced by deprotonation of the corresponding radical cation.[5c, 7d, 18] Given the small energy difference of the two conformers and low energy barrier for deprotonation (~ 4.75 kcal/mol), the anti-and syn-conformational isomers are expected to form in approximately equal amounts from this intermediate.[22] The dominant absorption peaks (> 600 nm) of anti-dG(N2-H)• and syn-dG(N2-H)• are predicted to overlap significantly with comparable intensity (Figure 3a). Consequently, it is expected that the two peaks of the anti- and syn- dG(N2-H)• may not be resolved in the spectrum containing both conformers in approximately equal amounts. The spectra of N1-MedG(N2-H)• produced by one-electron oxidation and deprotonation, where only a single broad peak centered at ~ 630 nm are consistent with this.[5c, 7d, 18]</p><p>In contrast to the generation of N1-MedG(N2-H)• via deprotonation of N1-MedG•+, the conformation dG(N2-H)• generated via sensitized photolysis of 1 is controlled by the conformation of the precursor, which also exists in anti- and syn-conformations (Figure 2). B3LYP/6–311++G(d,p)//PCM calculations of the analogue lacking a 2'-deoxyribose indicate that syn-8 is more stable than anti-8 by 0.21 eV (~4.83 kcal/mol). The more abundant syn-1 is expected to result in a greater amount of syn-dG(N2-H)• than anti-dG(N2-H)•. Moreover, DFT calculations at DFT/B3LYP/6–311++G(d,p)//PCM level showed that the energy barrier for the transition between syn-dG(N2-H)• and anti-dG(N2-H)• is 0.89 eV (20.5 kcal/mol) (Figure S9). This significant energy barrier suggests that the interconverison between the two conformers is kinetically infeasible on the LFP experiment timescale. These results are consistent with a recent report by Sevilla in which rotation of the N2-H group in N1-MedG(N2-H)• from 0 to 60° with respect to the purine ring also indicated a high rotational barrier.[6b] Consequently, the significant rotational barrier and the preference for generating syn-dG(N2-H)• by 1 results in a partially resolved spectrum of the two conformers. (Figure 1). The transient spectrum for dG(N2-H)• with maximum signal (4 μs) is reproduced by combining two broad peaks with λmax at ~ 610 and 650 nm (Figure 4). Dividing these two peak areas by their respective molar absorption coefficients (estimated using the calculated oscillator strengths in Table S2) suggests that the observed spectrum is the result of an ~ 2:1 mixture of syn-dG(N2-H)• relative to anti-dG(N2-H)•.</p><p>The decays of the 610 and 650 nm bands produced upon sensitized photolysis of 1 (1 mM) in the absence of additional reducing agent were fitted to double-exponentials and exhibited comparable lifetimes (Figure 5, Table 3), corroborating the proposal that the bands are attributable to syn- and anti-dG(N2-H)•. The shorter lifetime decay constants change little with the addition of reducing agent, but are dependent on the concentration of the radical precursor. We attribute this decay pathway to the reduction of dG(N2-H)• by its precursor (1), a process that was detected in product studies (Table 1). This bimolecular process is more prominent in the laser flash photolysis experiments, which are carried out at significantly higher concentrations of 1 and greater photon fluxes. In contrast, the slower decay process is affected by the addition of glutathione (GSH) or BME. Attributing the change in the lifetimes of the slower decay constants for the transient at 610 nm and 650 nm to reduction of dG(N2-H)• by the thiols indicates that BME (~2.7 – 2.9 × 104 M−1s−1) reacts approximately 10-fold more slowly with the nitrogen-centered radical than does GSH (~3.0 – 3.2 × 105 M−1s−1). These rate constants are considerably slower than what would be expected for reaction with a carbon-centered radical.[23] However, they are consistent with reactions of other nitrogen-centered radicals that are conjugated to electron accepting substituents, such as a purine ring (e.g. 2).[10a, 24] These radicals and dG(N2-H)• are electron deficient and kinetically mismatched for reaction with thiols.</p><p>dG(N2-H)• also features a less intense absorption band at ~ 370 nm, which is evident upon diminution of the transient absorption of triplet acetophenone after 4 μs. Subsequently, the band at 370 nm decays at comparable rates as those at 610 nm and 650 nm (Figure S10).[7] This observation is inconsistent with the proposed tautomerization of dG(N2-H)• to dG(N1-H)•. Pulse radiolysis experiments indicate that dG(N1-H)• absorbs strongly in this region.[5c] TDDFT-B3LYP/6–31++G(d,p)//PCM calculations indicate that dG(N1-H)• should absorb more strongly in this region than dG(N2-H)• (Figure 3, Table S2). These spectral features for the two radicals are affirmed by other computational studies.[6b] Furthermore, radiolysis studies showed that dG(N1-H)• decays relatively slowly with a lifetime of ~0.07 s.[25] Consequently, if dG(N2-H)• tautomerized to dG(N1-H)• with the reported first-order rate constant of 2.3 × 104 s−1 (t½ < 30 μs), the absorption in the 370 nm region would have maintained its intensity or even increased slightly during the time that the longer wavelength bands for dG(N2-H)• decay (Figure 1, 3).[7d].[6b] Finally, the observation that the intense absorption features of dG(N2-H)• at 610 nm and 650 nm exist for hundreds of microseconds, provide additional evidence against rapid tautomerization (Figure S11).</p><!><p>On account of their being common intermediates in DNA damage, the structure and reactivity of purine radicals have garnered significant interest. UV-photolysis of appropriately designed precursors is a common approach for generating homogeneous solutions of these and other DNA radicals.[26] The generation and reactivity of dG(N2-H)• has been a contentious issue, in part because it has only been produced using radiolysis, which may not produce homogeneous solutions of the radical. To address this issue, we developed a photochemical precursor (1) that produces dG(N2-H)• in high yield, as evidenced by product studies. Photosensitization by acetophenone enabled using 1 as a high fidelity source of dG(N2-H)• in laser flash photolysis experiments, where the distinct spectral features of anti- and syn-dG(N2-H)• are resolved . LFP affirmed product studies, showing that dG(N2-H)• is reduced by precursor 1. dG(N2-H)• also reacts with thiols, albeit significantly more slowly than carbon-centered radicals.</p><p>Importantly, photochemical generation of dG(N2-H)• from 1 enabled us to address the proposed microsecond timescale tautomerization of dG(N2-H)• to dG(N1-H)•. Chatgilialoglu, Steenken and Sevilla had independently reported spectra that they attributed to dG(N2-H)• or N1-MedG(N2-H)• upon radiolysis of a variety of guanosine derivatives.[5c, 7, 18] The observed absorption spectra following irradiation of N1-methylated substrates under different conditions were in good agreement, exhibiting defined absorption bands with λmax between 610 and 630 nm.[5c, 7b, 18] These spectra were very different than that observed following generation of dG•+ near neutral pH where a transient exhibiting λmax ~370 nm and a weaker absorption band at ~500 nm were attributed to dG(N1-H)•.[5c] These well-defined spectra were also in contrast to those reported following reaction of 8-bromoguanosine with solvated electron, and either dG or guanosine (G) with HO•.[7b–d] The authors ascribed the broad transients that extend from ~500 – 650 nm to dG(N2-H)•. In addition, the authors attributed the first order decay (t½ <30 μs) of absorption at 620 nm to tautomerization of dG(N2-H)• to dG(N1-H)•, despite the lack of sufficient spectroscopic evidence for the growth of the latter[7a],[7c] and the contradiction with the high barrier of 18.68 kcal/mol for the tautomerization.[8] LFP generation of dG(N2-H)• from 1 unambiguously shows that the radical does not tautomerize to dG(N1-H)• on even the hundreds of microseconds timescale, an observation that is consistent with recent experiments in G-quadruplexes.[20] Given these two independent reports that refute purine radical tautomerization, one must also question whether hydroxyl radical generates substantial quantities of dG(N1-H)• by abstracting a hydrogen atom from the N2-amino group of 2'-deoxyguanosine and subsequent tautomerization.[7c, 7d] The differing conclusions drawn from various radiolysis experiments may be attributable to differences in precursors, concentrations, doses and dose rates, as well as the inherent lack of chemical specificity when high energy species such as hydroxyl radical are used to generate reactive intermediates. UV-photolysis of a designed precursor (1) to dG(N2-H)• is not limited in this way. We conclude that it is unlikely that hydroxyl radical reacts directly with 2'-deoxyguanosine to yield dG(N2-H)•, and that this radical does not readily tautomerize to the more stable dG(N1-H)•.</p>

PubMed Author Manuscript

Glycopolymer Induction of Mouse Sperm Acrosomal Exocytosis Shows Highly Cooperative Self-Antagonism

Identifying inducers of sperm acrosomal exocytosis (AE) to understand sperm functionality is important for both mechanistic and clinical studies in mammalian fertilization. Epifluorescence microscopy methods, while reproducible, are laborious and incompatible for high throughput screening. Flow cytometry methods are ideal for quantitative measurements on large numbers of samples, yet typically rely on the use of lectins that can interfere with physiologic AE-inducers. Here, we present an optimized triple stain flow cytometric method that is suitable for high-throughput screening of AE activation by glycopolymers. SYTO-17 and propidium iodide (PI) were used to differentiate cells based on their membrane integrity or viability, and membrane impermeable soybean trypsin inhibitor (SBTI) was used to monitor acrosome exocytosis. The SBTI/PI/SYTO-17 combination provides a positive screen for viability and AE of live sperm cells with minimal noise or false positives. A scattering gate enables the use of samples that may be contaminated with non-cellular aggregates, e.g., cryopreservation agents. This assay format enabled detailed analysis of glycopolymer dose response curves. We found that fucose polymer has a narrow effective dose range (EC50 = 1.6 \xc2\xb5M; IC50 = 13.5 \xc2\xb5M); whereas mannose polymer and \xce\xb2-N-acetylglucosamine polymer have broader effective dose ranges (EC50 = 1.2 \xc2\xb5M and 3.4 \xc2\xb5M, respectively). These results highlight the importance of testing inducers over a large concentration range in small increments for accurate comparison.

glycopolymer_induction_of_mouse_sperm_acrosomal_exocytosis_shows_highly_cooperative_self-antagonism

INTRODUCTION<!>MATERIALS AND METHODS<!>Fluorescent Stains<!>Glycopolymer solutions<!>Sperm Treatment<!>Flow Cytometry<!>Statistical Analysis<!>Gating analysis of triply stained sperm<!>Induction of AE by glycopolymers<!>DISCUSSION<!>

<p>Sperm acrosomal exocytosis (AE1) is a vital event for fertilization to occur [1]. Various physiological and nonphysiological agonists have been shown to activate mammalian AE including zona pellucida (ZP) and cumulus terminal monosaccharides [2–7]. Multiple sperm–carbohydrate binding events must occur simultaneously to stimulate AE and identification of the associated receptors is complex [8–11]. Thus, synthetic biomaterials to assess sperm AE induction reproducibly are required to clearly identify receptors responsible, to develop simple semen analysis assays, and to identify new targets for contraceptives [11]. Several groups have used neoglycoconjugates to explore the involvement of individual types of saccharides in AE [12–14]. Previously, we found that a multivalent display of mannose, GlcNAc or fucose residues linked to a norbornene-derived polymer backbone triggers AE in a concentration dependent manner, and the signaling induced by these polymers converges onto the same intracellular signaling pathways [14]. The chemotypes suggest that single ligand type-receptor interactions are equivalent, but redundant, and further study was warranted. In order to further develop these glycopolymers as probes of AE, we developed the flow cytometry method described in this work as a robust and reliable assay that can be used to analyze larger numbers of samples that utilize a neoglycopolymer as the AE-inducer.</p><p>Our original studies, like many others, were performed utilizing Peanut agglutinin (PNA), lectin staining and a combined epifluorescence and DIC microscopy assay. Although highly reproducible, assessment of AE by combined fluorescence and DIC microscopy is laborious and incompatible with high-throughput analysis. In addition, it is difficult to differentiate healthy acrosome exocytosed sperm from damaged sperm after fixing and lectin staining. In order to analyze the activities of a large set of related glycopolymers, we set out to explore an alternative method to assess simultaneously AE and sperm viability.</p><p>Triple-stain flow cytometry methods to evaluate sperm viability and acrosome integrity have been developed [15, 16]. Using a combination of membrane permeable and impermeable DNA binding dyes such as SYTO-17 and propidium iodide (PI), respectively, with an acrosomespecific lectin provides a method of distinguishing viable from damaged cells by flow cytometry, eliminates non-cellular aggregates that can be present in samples, and identifies sperm that have undergone AE [17–19]. PNA, often used in AE assessment, is a membrane impermeable lectin that binds to soluble acrosomal molecules that contain a Gal-β(1–3)-GalNAc sugar sequence [18, 20]. Membrane impermeable PNA can only bind to its ligand when cells are at the early stage of acrosome exocytosis [21]. As exocytosis progresses, ligand bound-PNA diffuses out of the cell and the fluorescent signal is lost. The kinetics of AE have been previously shown to vary across individuals [22], therefore utilizing the time-sensitive lectin PNA on live cells is unsuitable for high-throughput analysis. Membrane fixation and permeabilization is necessary in order for PNA to bind to acrosomal ligands of intact cells [7, 21, 23], and thus, viability is not evaluated. The combination of Pisum sativum agglutinin (PSL) and PI was used to assess AE in live human sperm.[24] In contrast to use of PSL with fixed sperm, PSL diffuses into the acrosome after exocytosis begins and pores open, and upon binding, stabilizes the acrosome matrix. Thus PSL provides a positive fluorescence signal for AE in live sperm that is persistent for greater than ten minutes, thereby simplifying detection. However, PSL works through its high affinity interaction with mannose and glucose. The use of lectins in general is problematic with saccharide-based AE inducers in a live monitoring assay. Because the recognition elements for lectins are sugars, AE-inducers that utilize glycoproteins from ZP or glycopolymers can interfere with lectin recognition of sperm AE. Thus, the currently established staining systems were insufficient for our purposes.</p><p>A number of proteases have been identified in the acrosome of mouse sperm including acrosin, an endoprotease with trypsin-like specificity located on the inner acrosomal membrane (IAM) [25, 26]. Acrosin is derived from the enzymatically inactive zymogen, proacrosin, which is converted to the active form as a result of AE [27]. Soybean trypsin inhibitor (SBTI) inhibits the catalytic activity of serine proteases, including acrosin, by directly binding to the protease [28]. Localization of Alexa Fluor® 488 SBTI revealed that binding increases rapidly as the IAM is exposed during AE [29]. This increase is followed by a slow decline in fluorescence representing the release of non-membrane associated acrosin into the surrounding medium. Ultimately the fluorescence signal from remaining IAM-bound acrosin stabilizes [30, 31]. The stable SBTI fluorescence at a raised plateau provides a larger window of time for measuring AE reproducibly. Therefore, we set out to develop and optimize a triple stain flow cytometric method using SBTI, SYTO-17 and PI that is suitable for high-throughput screening and that would produce results comparable to our previously published microscopy data. This method was used to assess the AE activity of glycopolymers with varying structures and lays the groundwork for AE activation assays with clinical samples.</p><!><p>All experiments performed on mice were approved by the Stony Brook University IACUC (Protocol 0616) and were conducted in accordance with the National Institute of Health and the United States Department of Agriculture guidelines. DMSO and A23187 calcium ionophore were purchased from Sigma-Aldrich. A23187 stock solution (1 mg/ml) was prepared by dilution with DMSO. Propidium iodide was purchased from Fisher Scientific. Alexa Fluor® 488 soybean trypsin inhibitor (SBTI) and SYTO-17 were purchased from Life Technologies. All chemicals for assay buffers were purchased from Sigma-Aldrich, Fisher Scientific and VWR.</p><!><p>Propidium iodide (PI) was dissolved in water (2.4 mM) and stored at −20 °C. Alexa Fluor® 488 soybean trypsin inhibitor (SBTI) was dissolved in water (1 mg/ml) and stored at −20 °C in 40 µl aliquots. SYTO-17 was diluted in DMSO (1 mM) and stored at −20 °C.</p><!><p>Deprotected polymers were synthesized and characterized as reported [14]. Stock solutions of poly(Glu)100, poly(Gal)100, poly(GalNAc)100, poly(Man)100, poly(Fuc)100 and poly(GlcNAc)100 were stored in DDI water at −20 °C at a polymer concentration of 100 µM.</p><!><p>Sperm were isolated by force from the cauda epididymis of two 9- to 12-week-old CD-1 male breeders (Charles River) in a phenol red-free M16 medium (2.57 mM CaCl2) (6 mL) supplemented with 0.3% BSA (w/v). The sperm suspension was then gently pipetted into a polypropylene culture tube (12 mm × 75 mm) and incubated for 10 min at 25 °C. Then the concentration of sperm concentration was assessed by haemocytometer and motility was examined by phase-contrast microscopy (20X). Aliquots (125 µl) containing about 2.5× 106 sperm were subsequently diluted to a final volume of 250 µl (1.25×107 sperm/ml). The sperm suspension was incubated for another 15 min to get additional capacitation at 25 °C. A 250 µl solution containing SYTO 17 (2.5 µl), SBTI (1 µl) and either A23187 (5 µM, positive control), glycopolymer (0.5 µM – 30 µM, sample), or Dulbecco's Phosphate-Buffered Saline (PBS, negative control) was added to the 250 µl sperm suspension to a final volume of 500 µl. The mixture was kept at 0 °C in the dark for 15 min and then 20 min in the dark at rt. Samples were centrifuged for 5 min at 500g, and the resulting pellet was re-dissolved with 500 µl of a PBS solution containing PI (24 µM), SBTI (2 µg/ml) and SYTO-17 (5 µM). Samples were re-suspended at 1 min intervals and allowed to incubate for 20 min at 25 °C before flow cytometric analysis.</p><!><p>Analyses were performed on a Becton Dickinson LSR Fortessa system in combination with the BD-FACSDiva™ Software Version 8. SYTO-17 stained all cellular events and was excited at 640 nm and detected at 670/30 nm. PI stained all nonviable cells and was excited at 561 nm and detected at 582/15 nm. Alexa Fluor® 488 SBTI stained all cells that have undergone acrosome exocytosis and was excited at 488 nm and detected at 530/30 nm. These three stains have minimal emission overlap. Utilizing the forward and sideways scatter, small debris was gated out of analyses. As previously reported, additional gating was necessary to remove noncellular events which displayed similar scatter characteristics to sperm [16]. Events which exhibited weak SYTO-17 and PI staining (weak DNA content) were gated out before acrosome integrity and viability could be determined. Viability and acrosome integrity was measured at a flow rate of 600 cells/sec for 1 min. On average, 20% of sperm were PI negative (alive). 9% of live sperm cells undergo spontaneous acrosome exocytosis based on the negative control. In the presence of 5 µM A23187, approximately 21% of live cells undergo acrosome exocytosis. AE induced by A23187 was concentration dependent. However, at high concentrations, a large amount of cell death was observed. In addition, excess cellular debris complicates gating at high concentrations. Therefore, 5 µM A23187 was selected for use as a positive control.</p><!><p>The acrosome integrity of live sperm was normalized to PBS (negative control) and 5 µM A23187 (positive control). Only live sperm were included in the analysis and each experiment was conducted in biological triplicate. Normalized AE% was calculated using [AE% of glycopolymers − AE% of negative control)]/[AE% positive control − AE% negative control]. Data represent mean ± SEM of at least three independent experiments when compared to the negative control. EC50 was obtained by fitting equation (1) and IC50 was obtained by fitting equation (2) using GraphPad or Kaleidagraph to the data: (1)AE%=AEmin+(AEmax−AEmin)/{1+10^((logEC50)−log([polymer])*s)} (2)AE%=AEmin+(AEmax−AEmin)/{1+10^((logIC50)+log([polymer])*s)}</p><p>Where AEmax is the maximum percentage of AE observed, AEmin is the lowest percentage of AE observed, EC50 is the concentration at which 50% AE is observed, and s is the slope of the response. For the poly(Fuc)100 data, the EC50 and IC50 response curves were fit independently to the corresponding subset of data that encompassed the activation or inhibition arm of the curve. The average AEmax and AEmin from these two curve fits was then used to refit the data to equations (3) and (4) to determine EC50 and IC50, respectively. Data represent mean ± SEM of between two and four independent experiments. (3)AE%=5.3+(14.2)/{1+10∧((logEC50)−log([polymer])*s)} (4)AE%=5.3+(14.2)/{1+10∧((logIC50)+log([polymer])*s)} </p><!><p>After AE induction, sperm were labeled with SYTO-17, PI and Alexa Fluor 488 SBTI and analyzed by flow cytometry for viability and acrosome integrity. Gating was conducted in a similar manner to previously published methods [16], in which a two-dimensional dot plot of the side versus the forward scatter was used to remove small debris from the analysis, and a subsequent two-dimensional dot plot of PI vs SYTO-17 was used to remove non-cellular events from the analysis (Fig. 1, A and B, respectively). The number of live-acrosome-reacted sperm was determined from the two dimensional dot plot of SBTI versus PI and the resulting two dimensional histogram for SBTI staining of PI negative (live) cells (Fig. 1, C, D, and E respectively).</p><p>Live acrosome-exocytosed cells exhibited high SBTI fluorescence emission resulting in the appearance of a new distinct peak on the SBTI histogram (Fig. 1, C and E). In some instances, this peak was flat and broad. However, acrosome-intact and acrosome-exocytosed sperm were distinguished by comparison with the control samples.</p><!><p>AE induction by six glycopolymers was measured at 5 µM, 10 µM and 20 µM by flow cytometry. The levels of AE induced were equivalent to those measured by fluorescence microscopy (Fig. 2). The high AE inducers poly(Man)100, poly(GlcNAc)100 and poly(Fuc)100 show small differences in the maximal induction achieved. With the higher throughput of the flow cytometry assay, we undertook investigation of the activity of these AE-inducing glycopolymers across a larger range of concentrations in smaller steps (Fig. 3, Table 1). Polymers were not tested above 20 µM as sperm viability was reduced upon addition of high concentrations of polymer.</p><!><p>Our requirements for a flow cytometry based analysis of acrosomal exocytosis were two-fold. First, we required a gain of signal indicator of AE. Second, we required an AE probe that would not succumb to interference by a variety of glycopolymers. Previous methods have described the evaluation of acrosome integrity with SBTI and PI [32]. Because the SBTI ligand is a protein, we reasoned that glycopolymer AE inducers would not interfere with a live sperm labelling assay. We explored dual staining with SBTI and PI. However, in the absence of a DNA content gating protocol, i.e., SYTO-17/PI (Fig. 1B), a significant amount of sperm-sized debris or non-cellular events was included in the final analysis. In the absence of the SYTO-17 gating, up to 70% of events tagged as "live sperm" were counted as acrosome exocytosed in the positive control (5 µM A23187). In contrast, analysis of equivalent samples by microscopy identified approximately 25% of fixed sperm as having undergone AE. With the SYTO-17 gating step, noncellular events were eliminated from the analysis. This method identified a similar fraction of cellular events to that observed using PNA/SYBR-14/PI [17].</p><p>We selected an SBTI staining time of 20 min in order to balance the stability of the signal and to minimize the level of spontaneous AE observed in negative controls. Live sperm percentage was the same at 25 °C in ambient air and at 37 °C with 5% CO2 consistent with previous reports.[33] We selected 25 °C/ambient air as the incubation temperature to minimize spontaneous AE and to facilitate handling of samples.</p><p>Wu and Sampson [14] previously demonstrated using inverse fluorescence and DIC microscopy that homoglycopolymers displaying 100 copies of mannose, GlcNAc or fucose along a norbornyl backbone induce mouse sperm AE in a concentration dependent manner. The same polymer batches used by Wu and Sampson [14] were retested with the flow cytometry assay described here. The normalized AE percentages measured for these polymers are equivalent to our previously published fluorescence microscopy data.</p><p>Analysis by flow cytometry allowed us to measure the activity of these polymers with smaller concentration increments across a wide range of concentrations. The concentration studies provided full dose response AE curves that revealed poly(NBFuc)100 is the most effective AE inducer (Figure 3A). The differences in the shape of these three dose-response curves highlight the importance of measuring concentration dependence for polyvalent inducers in small step sizes in order to avoid missing the effective dose range of inducers. At concentrations above 10 µM poly(NBFuc)100, highly cooperative self-antagonism is observed. This inhibition is characteristic of a receptor that requires multivalent engagement of receptor [34–37]. The inhibition arm may result from competition of the multivalent interaction with monovalent binding (K1 vs K2, Figure 3B). In addition, steric occlusion may prevent engagement of every ligand on the polymer chain. However, the inhibition response is steeper than expected for either or both cases [37]. We hypothesize that the high cooperativity of inhibition results from the formation of a larger number of receptor-polymer multivalent complexes that donot activate signaling (K2') than of complexes that do activate signaling (K2) (Figure 3B). The consequence is that the effective dose range is quite narrow.</p><p>The EC50 for poly(NBMan)100 is identical to that of poly(NBFuc)100 and no inhibition is observed. Despite a broader effective dose range, the maximal AE induction observed is lower than for fucose, indicating that not all complexes formed initiate AE signaling. poly(NBGlcNAc)10 was less effective than fucose both in EC50 and maximal response, and poly(NBGlcNAc)100 shows the beginning of antagonism at the highest concentrations measurable.</p><p>The differences in the shape of these three dose-response curves for these polymers highlight the importance of measuring concentration dependence for polyvalent inducers in small step sizes in order to avoid missing the effective dose range of inducers/inhibitors. The dose responses are consistent with our previous conclusion that the three polymers activate AE through different receptors that converge into a single signaling pathway [14]. Further interpretation of the physiologic significance awaits identification of the receptors that are engaged and the structures of the receptor-polymer complexes [38].</p><p>The triple stain flow cytometric assay discussed in this study provides a quantitative method to inform on sperm viability and acrosome integrity in the presence of exocytosis inducers and/or inhibitors. The rapid throughput format opens the possibility of testing libraries of compounds to both investigate reasons for infertility and ways to block acrosome exocytosis. Ultimately, these types of probes will serve as platforms for infertility testing.</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>Abbreviations: AE: acrosomal exocytosis; ZP, zona pellucida; GlcNAc: N-acetylglucosamine; DIC, differential interference contrast; Gal, galactose; PNA: peanut agglutinin; IAM, inner acrosomal membrane: SBTI: soybean trypsin inhibitor; PI, propidium iodide; Glu, glucose.</p>

PubMed Author Manuscript

Computer-aided design of multi-target ligands at A1R, A2AR and PDE10A, key proteins in neurodegenerative diseases

Compounds designed to display polypharmacology may have utility in treating complex diseases, where activity at multiple targets is required to produce a clinical effect. In particular, suitable compounds may be useful in treating neurodegenerative diseases by promoting neuronal survival in a synergistic manner via their multi-target activity at the adenosine A1 and A2A receptors (A1R and A2AR) and phosphodiesterase 10A (PDE10A), which modulate intracellular cAMP levels. Hence, in this work we describe a computational method for the design of synthetically feasible ligands that bind to A1 and A2A receptors and inhibit phosphodiesterase 10A (PDE10A), involving a retrosynthetic approach employing in silico target prediction and docking, which may be generally applicable to multi-target compound design at several target classes. This approach has identified 2-aminopyridine-3-carbonitriles as the first multi-target ligands at A1R, A2AR and PDE10A, by showing agreement between the ligand and structure based predictions at these targets. The series were synthesized via an efficient one-pot scheme and validated pharmacologically as A1R/A2AR–PDE10A ligands, with IC50 values of 2.4–10.0 μM at PDE10A and Ki values of 34–294 nM at A1R and/or A2AR. Furthermore, selectivity profiling of the synthesized 2-amino-pyridin-3-carbonitriles against other subtypes of both protein families showed that the multi-target ligand 8 exhibited a minimum of twofold selectivity over all tested off-targets. In addition, both compounds 8 and 16 exhibited the desired multi-target profile, which could be considered for further functional efficacy assessment, analog modification for the improvement of selectivity towards A1R, A2AR and PDE10A collectively, and evaluation of their potential synergy in modulating cAMP levels.Electronic supplementary materialThe online version of this article (10.1186/s13321-017-0249-4) contains supplementary material, which is available to authorized users.

computer-aided_design_of_multi-target_ligands_at_a1r,_a2ar_and_pde10a,_key_proteins_in_neurodegenera

Background<!><!>Background<!>Design of synthetically feasible A1R/A2AR–PDE10A multi-target ligands<!>Target prediction of the designed RECAP library<!>Docking of the compounds predicted as A1R/A2AR–PDE10A multi-target ligands<!><!>Substructure analysis of the compounds predicted as A1R/A2AR–PDE10A multi-target ligands<!><!>Pharmacological evaluation of novel 2-aminopyridine-3-carbonitriles<!>(SAR) structure–activity relationship analysis<!>Compound selectivity assessment<!><!>Analysis of the molecular docking studies of the synthesized 2-aminopyridine-3-carbonitriles<!>Computational assessment of CNS permeability<!>Conclusions<!>Selecting reference molecules for the design of multi-target ligands<!>Designing new multi-target ligands<!>Target prediction<!>Receptor preparation<!>Ligand preparation<!>Cut-off generation for compound selection from docking models<!>Docking<!>Substructural analysis<!>Synthesis of novel 4,6-substituted 2-amino-pyridin-3-carbonitriles<!>Chemistry<!>Synthetic procedure<!>2-amino-6-(4-fluorophenyl)-4-phenylpyridine-3-carbonitrile (1)<!>2-amino-6-(4-hydroxyphenyl)-4-phenylpyridine-3-carbonitrile (2)<!>2-amino-4-phenyl-6-(1,3-thiazol-2-yl)pyridine-3-carbonitrile (3)<!>2-amino-6-(1-methyl-1H-pyrrol-2-yl)-4-phenylpyridine-3-carbonitrile (4)<!>2-amino-4-(2-methoxyphenyl)-6-phenylpyridine-3-carbonitrile (5)<!>2-amino-4-(2,4-dimethoxyphenyl)-6-phenylpyridine-3-carbonitrile (6)<!>2-amino-4-(2H-1,3-benzodioxol-5-yl)-6-phenylpyridine-3-carbonitrile (7)<!>2-amino-4-cyclohexyl-6-phenylpyridine-3-carbonitrile (8)<!>2-amino-4-cyclohexyl-6-(2-fluorophenyl)pyridine-3-carbonitrile (9)<!>2-amino-4-cyclohexyl-6-(2-methylphenyl)pyridine-3-carbonitrile (10)<!>2-amino-4-cyclohexyl-6-(thiophen-2-yl)pyridine-3-carbonitrile (11)<!>2-amino-4-cyclohexyl-6-(thiophen-3-yl)pyridine-3-carbonitrile (12)<!>2-amino-4-cyclohexyl-6-(furan-2-yl)pyridine-3-carbonitrile (13)<!>2-amino-6-(2-fluorophenyl)-4-(4-methoxyphenyl)pyridine-3-carbonitrile (14)<!>2-amino-4-(4-methoxyphenyl)-6-(2-methylphenyl)pyridine-3-carbonitrile (15)<!>2-amino-6-(furan-2-yl)-4-(4-methoxyphenyl)pyridine-3-carbonitrile (16)<!>2-amino-6-(4-hydroxyphenyl)-4-(4-methoxyphenyl)pyridine-3-carbonitrile (17)<!>2-amino-4,6-bis(2-fluorophenyl)pyridine-3-carbonitrile (18)<!>2-amino-6-(2-fluorophenyl)-4-(2-methoxyphenyl)pyridine-3-carbonitrile (19)<!>2-amino-4-(2-methoxyphenyl)-6-(2-methylphenyl)pyridine-3-carbonitrile (20)<!>2-amino-6-(furan-2-yl)-4-(2-methoxyphenyl)pyridine-3-carbonitrile (21)<!>2-amino-6-(4-hydroxyphenyl)-4-(2-methoxyphenyl)pyridine-3-carbonitrile (22)<!>2-amino-4-(2-chlorophenyl)-6-(4-hydroxyphenyl)pyridine-3-carbonitrile (23)<!>2-amino-4,6-bis(4-hydroxyphenyl)pyridine-3-carbo-nitrile (24)<!>2-amino-4-(furan-2-yl)-6-(thiophen-3-yl)pyridine-3-carbonitrile (25)<!>Pharmacological evaluation of novel 4,6-substituted 2-amino-pyridin-3-carbonitriles<!>

<p>Neurodegeneration involves the progressive loss of the structure and function of neurons, which is common in Parkinson's, Huntington's disease and schizophrenia [1]. Recently, there has been substantial interest in the search for alternative non-dopamine (non-DA) based approaches for the treatment of neurodegenerative diseases, as the classical DA-based approaches have long been associated with many undesirable side effects such as dyskinesia, hallucinations, and on/off effects [2]. Given that the adenosine neuromodulation system (via the adenosine A1 and A2A receptors) has been identified as a key target for the management of neurodegenerative diseases, this qualifies its targeting as a potential promising non-DA based treatment approach [3, 4]. Indeed, modulation of cAMP levels has proven to have benefits in neuronal survival in an adenosine receptor-dependent manner [5]. In addition, recent findings suggest that phosphodiesterase 10A (PDE10A) also plays a role in neurodegenerative diseases such as Parkinson's, Huntington's disease, and schizophrenia [6–8]. Inhibition of PDE10A resulting in maintenance of elevated intracellular cAMP concentrations, has been suggested to be effective in the treatment of these diseases. Thus multi-target ligands that bind to different adenosine receptors subtypes (A1 and A2A receptors) while simultaneously inhibit PDE10A might be synergistic in modulating cAMP levels, which is of therapeutic potential for neurodegenerative diseases [9–11].</p><p>Conceptually, multi-target drugs work by creating a combination effect on multiple targets in the biological network simultaneously, which may (through e.g. synergistic effects) decrease the therapeutic dose required, thus increasing therapeutic efficacy, preventing drug resistance, and reducing target-related adverse effects [12–14]. Also, another advantage of multi-target drugs over other types of treatments such as combination therapies, is a reduced likelihood of drug–drug interactions [15, 16].</p><p>However, it remains a challenging task for medicinal chemists to design drugs with a specific multi-target profile and to achieve selectivity for specific targets over off-target effects with suitable pharmacokinetic properties [17, 18]. In fact, the field of multi-target drug design has recently become an active field of research in the pharmaceutical industry, where around 20 designed multi-target drugs have either reached advanced development stages or are already approved [14, 19, 20].</p><p>In particular, for Central Nervous System (CNS) diseases, there has been growing interest in exploiting the multi-target profiles of existing compounds to investigate their potential applicability as drugs. For example, multi-target profiles of drugs and drug candidates affecting the dopaminergic system have been investigated. Examples include Aripiprazole, Amitriptyline, Chlorpromazine, and Clozapine [21]. In addition, various multi-target based virtual screening protocols for multi-target drug design have been developed [13, 22–24]. Examples of ligand-based protocols include in silico target prediction and Chemogenomic and pharmacophore-based approaches, which resulted in the discovery of CNS drugs with multi-target combinations such as MAO-A/MAO-B/AChE/BuChE, AChE/BuChE, and H3-R/HMT/AChE/BuChE [21–24]. Structure-based approaches such as docking and molecular dynamics calculations have also been employed for the discovery of new multi-target ligands such as BuChE inhibitors/hCB2R and MAO-A/MAO-B/AChE/BuChE ligands to treat neurodegenerative diseases [25].</p><p>In this work, we offer a computational strategy for designing synthetically feasible ligands that bind to A1R and A2AR, and inhibit PDE10A—a novel multi-target combination of G protein-coupled receptors (GPCRs) and an enzyme, which has not, to our knowledge, been previously exploited. The designed ligands with this multi-target combination are intended as starting points for future development of multi-target drugs treating neurodegenerative diseases. It should be noted here that in the current study we only consider affinity of ligands to the above receptors, which we also experimentally validate as outlined below. However, for therapeutically relevant purposes also functional effects and optimization of selectivity towards A1R, A2AR and PDE10A need to be considered, which will be the area of a future study.</p><!><p>The computational strategy for rational design of A1R/A2AR–PDE10A multi-target ligands started with a focused chemical space consisting of known actives of A1R, A2AR and PDE10A, and formed new synthetically feasible compounds which were subjected to target prediction and docking for synthesis and pharmacological evaluation</p><!><p>A series of 2-aminopyridine-3-carbonitriles were selected for prospective validation of the pipeline, a series which was synthetically accessible via a one pot synthetic scheme i.e. providing products with the desired properties: cost-effective, synthetically efficient and available in a timely fashion [28, 29].</p><p>Subsequently the synthesized compounds were experimentally tested and confirmed as A1R/A2AR–PDE10A multi-target ligands. Selectivity against other subtypes of both protein families confirmed the pharmacological profile of the compound series, and structure activity relationships (SAR) were also deduced. Hence, in this work we report a successful computational strategy, which allowed the discovery of the first A1R/A2AR–PDE10A multi-target ligands. The novel A1R/A2AR–PDE10A ligands are sought to display a combination effect in modulating the A1R, A2AR, and PDE10A targets simultaneously similar to that of combination compounds of Adenosine receptors and PDEs, reported by Rickles et al., which were synergistic in modulating cAMP levels [10].</p><!><p>Human enzyme and receptor data were extracted from ChEMBL 20 [30]. Substructure analysis of A1R, A2AR ligands and PDE10A inhibitors with Ki and IC50 values less than or equal to 1 µM revealed that the most frequently occurring common heterocycles among the actives against the three target classes were pyridine, pyrimidine, piperazine, and 1H-pyrazole (Additional file 1: Figure S1). Subsequently, A1R (2104), A2AR (2489) and PDE10A inhibitors (679) containing those frequent heterocycles were subjected to RECAP analysis/synthesis in MOE (see Methods for details) [26]. As a result, 458,839 (potentially) synthetically accessible ligands were formed in silico. This list of candidates was filtered to those retaining the common heterocycles (listed above), in order to create a focused chemical space characteristic of A1R, A2AR and PDE10A (with the simultaneous trade-off of reduced novelty), giving rise to 22,233 compounds.</p><!><p>To assess the likelihood of active compounds against A1R, A2AR and PDE10A, PIDGIN 1.0 (Prediction including Inactivity), a tool which uses ECFP 4 circular Morgan fingerprints and trained on ChEMBL actives and PubChem inactives, was used to perform in silico target prediction for the focused RECAP library (22,233 compounds) [24]. Subsequent enrichment analysis of the predictions was done using an estimation score, average ratio as developed by Liggi et al. [31] and via Chi square test [32]. For targets to be considered as enriched according to these methods, the estimation score and the Chi square test p value should be less than or equal to 0.01 and 0.05, respectively. Hence, upon analyzing the enrichment parameters for the A1R, A2AR and PDE10A targets that were predicted for the focused RECAP library (Additional file 1: Figure S2), the three targets were predicted with an estimation score equal to 0 (enriched) as well as average ratios less than 0.1 (enriched) with Chi squared p values < 0.005. The percentage of RECAP compounds of the focused library that were predicted as actives against the A1R, A2AR and PDE10A targets were 51.1, 52.8, and 24.5% respectively. These numbers are relatively high, which however is understandable given that the input to the RECAP analysis consisted of experimentally established known ligands of the above protein targets.</p><!><p>In the next step docking and further substructure analysis were performed on compounds of the focused RECAP library, which were predicted as A1R/A2AR–PDE10A multi-target ligands from the ligand-based side in the previous step. 2563 compounds were predicted as actives against the three desired targets, and they were subsequently docked against a high resolution (1.8 Å) A2AR protein crystal structure (PDB ID: 4EIY) [33] its corresponding A1R homology model (see Methods for details), and PDE10A (PDB ID: 4DDL) [34].</p><p>Compounds which were carried forward to substructural analysis were selected when their docking score gave a value less than a pre-determined cut-off value computed from the docking scores. This cut-off value was evaluated as the docking score with the best F measure statistic obtained by docking a set of known actives and inactives against the protein crystal structures and the homology model (see Methods for details).</p><p>As a result, a distribution of RECAP compounds that were favorable as multi-target ligands by target prediction and docking was obtained, where 62.47% of the RECAP compounds that were predicted as A1R/A2AR–PDE10A multi-target ligands and docked against PDE10A exhibited docking scores lower than − 6.49 (the threshold of the best F measure discriminating between actives and inactives for known ligands). Out of the RECAP compounds which displayed docking scores lower than − 6.49 against PDE10A, 48.89 and 35.23% displayed docking scores lower than − 7.26 and − 8.49 against A1R and A2AR (the thresholds of the best F measures).</p><!><p>2563 compounds of the focused RECAP library were predicted as A1R/A2AR–PDE10A multi-target ligands, and docked against the A2AR protein crystal structure (PDB ID: 4EIY), A1R homology model, and the PDE10A protein crystal structure (PDB IB: 4DDL), the RECAP series which showed an agreement between the ligand-based and structure-based predictions were mainly a 6,7-alkoxyisoquinolines b [1,2,4] triazolo[1,5-c]quinazolines c 2-aminopyridine-3-carbonitriles d imidazo[1,5-a]quinoxalines</p><!><p>The chemical series were identified as [1,2,4]triazolo[1,5-c]quinazolines (50.4% of all positively predicted multi-target ligands by in silico target prediction as well as docking), imidazo[1,5-a]quinoxalines (14.4%), 6,7-alkoxyisoquinolines (10.6%), and 2-aminopyridine-3-carbonitriles (9.2%). These were in addition to various compounds containing the common and frequent heterocycles identified earlier (15.4%). Each series identified could be considered for synthesis, SAR studies and validation as A1R/A2AR–PDE10A multi-target ligands.</p><!><p>The one-pot synthetic route followed for the synthesis of novel 4,6-substituted 2-amino-pyridin-3-carbonitriles</p><p>Percent inhibition of the synthesized 4,6-substituted 2-amino-pyridin-3-carbonitriles at 10 µM (PDE10A) or IC50 (µM) and percentage displacement at 0.1 µM (A1R and A2AR), or Ki</p><p>IC50 values of the 2-aminopyridines-3-carbonitriles were measured for the four phosphodiesterases PDE7A, PDE7B, PDE9A and PDE10A at 10 μM concentration. For those compounds that showed percentage inhibition greater than 70% and selectivity against other measured isoenzymes, IC50 were determined. Calculation of the Ki values at A1R, A2AR, A2BR and A3R was approximated using the Cheng-Prusoff equation: Ki = IC50/[1 + (C/KD)], where IC50 is the concentration of compound that displaces the binding of the radioligand by 50%, C is the concentration of radioligand, and KD is the dissociation constant of each radioligand</p><!><p>Given that the objective of this work is to find compounds displaying specific multi-target activity, compounds 8, 16, 21, and 25 were identified as A1R/A2AR–PDE10A multi-target ligands, inhibiting PDE10A with IC50 values of 2.4, 3.2, 10.0, and 5.1 µM respectively, and binding to A1R with Ki values of 294 and 34 nM (compounds 8 and 16, respectively), and to A2AR with Ki values of 41, 95, and 55 nM (compounds 16, 21, and 25, respectively). Notably, compound 16 exhibited the desired multi-target profile as a PDE10A inhibitor and a dual binder to A2AR and A1R.</p><p>It was previously reported that substituted pyridines exhibited PDE inhibitory activity [41, 42], and 2-amino-pyridin-3-carbonitriles are adenosine receptor ligands [36]. In this study we have now identified suitable compounds matching both criteria as A1R/A2AR–PDE10A multi-target ligands, satisfying the original compound design objective.</p><!><p>The purpose of the SAR analysis was to rationalize the variation in activity of the newly discovered A1R/A2AR–PDE10A multi-target ligands against PDE10A, given that 2-amino-pyridin-3-carbonitriles have been discovered as a novel class of PDE10A inhibitors. Also due to the fact that compounds of this substructural class were documented as adenosine receptor ligands [36], computational SAR studies were focused on the PDE10A data, where the variation in potency was rationalized in relation to the physicochemical properties of the compounds (which were computed by FAFDrug3, Additional file 1: Table S1) [40].</p><p>A trend observed repeatedly in several cases was that when logP decreased, associated with an increase in tPSA, then this led to an improvement in the activity against PDE10A. Initial analysis concentrated on compounds 1–4, which have a phenyl substituent at position 4 of the pyridine ring. Compound 3 was the most potent PDE10A inhibitor with an IC50 of 2.0 µM, and a computed logP of 3.1 and tPSA of 103.9 Å2. Similarly, for compounds 5–7 having a phenyl substituent at position 6 of the pyridine ring, compound 6 was the most potent against PDE10A with an IC50 of 5.7 µM and a computed logP of 4.0 and tPSA of 81.2 Å2. For compounds 8–13, which have a cyclohexyl ring at position 4 of the pyridine ring, compound 12 displayed the most potent PDE10A inhibitory activity with an IC50 of 0.9 µM and a computed logP of 4.7 and tPSA of 90.9 Å2. For compounds 14–17, with a p-methoxyphenyl substituent at position 4 of the pyridine ring, compound 16 with the smallest predicted lipophilicity of 3.1 and tPSA of 85.1 Å2 displayed a good PDE10A inhibitory activity with an IC50 value equal to 3.2 µM, yet the most potent compound was 15 with an IC50 value of 1.5 µM and a computed logP of 4.4 and tPSA of 71.9 Å2. For compounds 19–22, with an o-methoxyphenyl substituent at position 4 of the pyridine ring, compound 22 displayed PDE10A inhibitory activity with the highest potency (IC50 value of 5.6 µM), and a computed logP of 3.7 and tPSA of 92.2 Å2. Finally a similar general trend is observed for the compounds 23 and 24 with a 4-hydroxyphenyl substituent at position 6 of the pyridine ring, where compound 24 was a more potent PDE10A inhibitor with an IC50 of 3.1 µM and computed logP of 3.4 and tPSA of 103.2 Å2. Hence, it could be deduced that in the majority of the series considered, where the substituents on a single position is varied, a decrease in computed lipophilicity associated with an increase in polarity generally improved the activity of compounds against PDE10A. This general trend can be attributed to the hydrophilic nature of the pocket, which favours the interactions between the ligand and the PDE10A protein by compounds exhibiting these properties.</p><!><p>The selectivity of compounds 1–25 against the selected major off-targets A2BR, A3R, PDE7A, PDE7B, and PDE9A, was predicted using PIDGIN at a threshold for binding greater than or equal to 0.8, and subsequently tested experimentally. It is noted here that the IC50 values were determined for compounds with % inhibition at phosphodiesterases greater than 70%. As shown in Additional file 1: Table S2, the synthesized compounds are mostly inactive against those off-targets except for compounds 16, 17, 21, and 23 that exhibited IC50 values of 3.4, 3.5, 15.1 and 1.8 µM against PDE7A, and compounds 23 and 25, which exhibited IC50 values of 7.3 and 4.7 µM against PDE7B. Remarkably, compound 8 was found to exhibit selectivity over all tested off-targets using the above criterion, with the lowest selectivity measured for PDE7B (of 55% inhibition at 10 µM ligand concentration). This can be compared to the IC50 value of 8 at PDE10A, which is 2.4 μM (indicating approximately twofold selectivity for 8).</p><p>In general, the experimental results on off-target prediction for the synthesised 4,6-substituted 2-amino-pyridin-3-carbonitriles 1–25 agree with the predictions generated using PIDGIN utilised to bias the compound design towards selective compounds such as 8 (Additional file 1: Table S2). This compound would serve as a good starting point for analog modification to improve the selectivity of the synthesized ligands towards PDE10A.</p><!><p>Docking studies predicted molecular interactions characteristic of the 4,6-substituted 2-amino-pyridin-3-carbonitriles with the A2AR protein crystal structure (PDB ID: 4EIY), A1R homology model, and PDE10A protein crystal structure (PDB ID: 4DDL), which are displayed for representative multi-target ligands with the following combinations: compound 8 (A1R–PDE10A), 18 (A1R–A2AR), and 25 (A2AR–PDE10A): a interactions with A2AR: the overlaid compounds 18 and 25 exhibit H-bonds via amino and carbonitrile groups with Asn253, and the pyridine rings are π-stacked with Phe168 b interactions with A1R: the overlaid compounds 8 and 18 exhibit H-bonds via amino and carbonitrile groups with Asn254, and the pyridine rings are π-stacked with Phe171 c interactions with PDE10A: the overlaid compounds 8 and 25 have the pyridine rings π-stacked with Phe686 and Phe719. The molecular interactions predicted for the active molecules are consistent with observed interactions between co-crystallised ligands and their corresponding protein crystal structures (PDB ID: 4EIY and 4DDL) [33, 34] and the interactions with the A1R homology model reported in the literature [51, 52]</p><!><p>It can be seen that compounds 8 and 25, with IC50 values of 2.4 and 5.1 µM respectively, share similarities in predicted binding modes, since their pyridine rings display π-stacking with Phe686 and Phe719 of PDE10A (Fig. 3). These are the type of interactions predicted to be exhibited by the majority of the synthesized ligands from this work, as well as the only existing interactions between co-crystallised PDE10A inhibitors discovered by fragment screening (PDB ID: 5C2E, 5C1W, 5C29, 5C2A ligands with Ki values of 2, 8, 700, 880, and 4.8 nM, respectively) [43]. It is noted that the ligand of 5C2A exhibits a considerable selectivity towards PDE10A over all the other PDEs (in the range of 100–1000 fold and greater over the majority of PDEs, with the least selectivity observed being in the range of 25–100 fold). This ligand exhibits only π-stacking interactions with Phe686 and Phe719, similar to the mode of interactions of compound 8 with PDE10A, which is relatively selective over all tested PDEs, with the lowest selectivity being measured for PDE7B (of 55% inhibition at 10 µM ligand concentration) and compound 25, which is selective against all tested PDEs except PDE7B (Table 1 and Additional file 1: Table S2). Additional interactions were seen in analogs discovered by fragment screening, namely hydrogen bonding with Gln716 and Tyr683 in the PDE10A selectivity pocket (PDB ID: 5C28 and 5C2H with Ki values of 2200 and 0.0082 nM respectively). [43] The ligand of 5C2H exhibits π-stacking with Phe686 and Phe719 and hydrogen bonding with Tyr683 in the PDE10A selectivity pocket. The 5C2H ligand showed a very high selectivity towards PDE10A, greater than 5000 fold, which emphasizes the consideration of compound 8 for analog modification to target the selectivity pocket in order to improve the folds of selectivity towards PDE10A. In addition, hydrogen bonding with Tyr683 in the PDE10A selectivity pocket is also seen in many other highly selective PDE10A inhibitors reported in the literature [44] (PDB ID: 5DH5, [45] 5B4L, [46] with Ki = 0.23 nM, and IC50 = 0.76 nM respectively), which further highlights the importance of analog modification to target the PDE10A selectivity pocket.</p><p>Moreover, it is noted that compounds 16 and 21 with IC50 values of 3.2 and 10.0 µM respectively (which are selective against all tested PDEs except PDE7A, Table 1 and Additional file 1: Table S2) were predicted to exhibit an additional type of interaction, H-bonding with Gln716 via their overlaid furan rings at position 6 of the pyridine ring (Additional file 1: Figure S3). In fact H-bonding with Gln716 was the only interaction, besides π-stacking with Phe686 and Phe719, which has been observed in many of the highly selective PDE10A ligands reported in the literature (PDB ID: 4DDL, [34] 3SN7, 3SNL, and 3SNI, [47] 5DH4 and 5DH5, [45] with IC50 values of 4.9, 0.7, 0.7, 11 nM and Ki = 0.23 nM respectively). As for other type of interactions generally exhibited by known PDE10A inhibitors such as hydrogen bonding with Gln726 and π-stacking with Phe729 (PDB ID: 5EDE) [48], none has been predicted for any of the compounds presented in this work.</p><p>Common predicted binding modes can also be observed for the synthesized compounds against the adenosine receptors A2AR and A1R. Figure 3 displays the interactions of two representative compounds 18 and 25, which exhibit Ki values of 948 and 55 nM respectively, and these are H-bonding of their pyridine rings with Asn253 and π-stacking of their amino and carbonitrile groups with Phe168 of A2AR. As for A1R, the overlaid compounds 8 and 18, with Ki values of 294 and 78 nM respectively, H-bond via their amino and carbonitrile groups with Asn254, and their pyridine rings are π-stacked against Phe171. It can be observed that the ligand/protein interactions predicted for the active compounds against the A2AR are also those seen in the co-crystallised ligand/protein crystal structures (PDB ID: 4EIY, [33] 3EML, [49] 5IU4, [50] with a Ki value of 0.8 nM for ZM241385, which is the common ligand for the three PDB IDs). Similar was the case for the reported interactions with the A1R homology model in the literature (with IC50 values of 2.9 and 6.2 nM for the reported ligands predicted to bind to the homology model of A1R) [51, 52].</p><p>Generally the compounds exhibited good selectivity towards A1R and A2AR (Table 1 and Additional file 1: Table S2) with a nanomolar range of binding affinities. As for the selectivity towards PDE10A, it could be improved by analog modification of compound 8, which favors the hydrogen bonding with Tyr683 in the PDE10A selectivity pocket. In addition, the potency of compounds against PDE10A could be optimized in itself, in order to achieve therapeutically relevant efficacy.</p><!><p>Compounds 8 and 16 exhibited the desired multi-target profile by inhibiting PDE10A and binding to A2AR and/or A1R. The physicochemical properties of these compounds were calculated by FAFDrug3 [40], and both compounds passed the Lipinski rule of 5 and the CNS filter, which takes into consideration the assessment of their ability to pass the blood brain barrier (Additional file 1: Figure S4) [53]. Hence, while further experimental work would be needed to establish the validity of those predictions, compounds 8 and 16 may serve as good starting points for further functional efficacy assessment and selectivity optimization towards PDE10A, A2AR and/or A1R for the subsequent consideration of multi-target drug development for the treatment of neurodegenerative diseases.</p><!><p>Here we report a successful computational strategy for designing the first A1R/A2AR–PDE10A multi-target ligands as a therapeutic prospect for neurodegenerative diseases. A retrosynthetic approach was employed using MOE/RECAP, followed by target prediction and docking of the resulting library against the desired targets. We have identified 2-aminopyridine-3-carbonitriles as a series that showed agreement between both the ligand- and structure-based predictions of activity against A1R, A2AR and PDE10A. The synthesis of this series via a one-pot synthetic scheme was pursued experimentally. As a result, compounds 8, 16, 21, and 25 were validated as A1R/A2AR–PDE10A multi-target ligands with IC50 values of 2.4, 3.2, 10.0, and 5.1 µM against PDE10A, and binding to A1R with Ki values of 294 and 34 nM (8 and 16 respectively), and to A2AR with Ki values of 41, 95, and 55 nM (16, 21, and 25 respectively). Furthermore, selectivity profiling of the synthesized 4,6-substituted 2-amino-pyridin-3-carbonitriles against other subtypes of both protein families showed that the multi-target ligand 8 exhibited a minimum of twofold selectivity over all tested off-targets. In addition, compounds 8 and 16 exhibited the desired multi-target profile against A1R, A2AR and PDE10A, which would serve as good starting points for further functional efficacy assessment and analog modification for the improvement of selectivity. In particular, this comprises investigating the signal transduction profiles of these compounds using techniques some of the authors have described before [51], as well as evaluating functional effects in cAMP assays to determine if these compounds do provide synergistic elevations in intracellular cAMP. One specific functional profile that would be of high interest and which is likely to elevate cAMP levels synergistically via the combination effect on multiple targets simultaneously, is the A1R antagonist/A2AR agonist, and PDE10A inhibitor.</p><p>In summary we have investigated a computational approach for the design of multi-target ligands that was validated experimentally via synthesis and pharmacological evaluation of 2-aminopyridine-3-carbonitriles as A1R/A2AR–PDE10A ligands. This approach is generally applicable to a wide range of multi-target ligand design problems, across disease areas and target families.</p><!><p>Using SQL (script provided in Additional file 1), human A1R (2860), A2AR (3566) ligands and PDE10A inhibitors (843) were extracted from the ChEMBL 20 database with Ki and IC50 values less than or equal to 1 μM respectively, and confidence scores of 8 or 9 [30]. Following extraction, the most frequent and common heterocycles between A1, A2A receptor ligands and PDE10A inhibitors were found by performing substructure analysis on each structure using the "Chemistry-> Analyze scaffolds" function in DataWarrior 4.2.2 [54]. Analysis of A1R, A2AR ligands and PDE10A inhibitors identified common and frequent heterocycles (pyridine, 1H-pyrazole, pyrimidine and 9H-purine for A1R and A2AR), and these were extracted from each set using RDKit, 9.1, Python [55]. It should be noted that compounds containing 9H-purine were also extracted from the original set even though this substructure is characteristic of A1R and A2AR only, since it is structurally similar to the common and frequent heterocycles identified (pyridine, 1H-pyrazole, and pyrimidine). Additional file 1: Figure S1 shows the most frequent heterocycles for the A1R, A2AR ligands, and PDE10A inhibitors and their relative frequencies in each set. It was found that they are furan, pyridine, xanthine, 1H-pyrazole, pyrimidine, piperazine, and 9H-purine. All of these heterocycles ranked among the top 30 for A1R, A2AR ligands and PDE10A inhibitors. This indicated their suitability for designing multi-target ligands at these protein targets, given the overlap in chemical (heterocyclic) space. In the case where no percentage is displayed for a particular target, this means that the heterocycle does not appear among the top 30 for the set of compounds involved.</p><!><p>A1R (2104), A2AR (2489) and PDE10A inhibitors (679) consisting of the common and frequent heterocycles, were subjected to RECAP analysis/synthesis in MOE [26]. The RECAP function electronically fragments and recombines molecules based on chemical knowledge of 11 chemical bond types derived from common chemical reactions [27]. As a result, 458,839 novel RECAP-derived compounds were formed. Finally the designed RECAP library was filtered using RDKit, Python according to the common and frequent heterocycles identified, which narrowed the list down to 22,233 compounds.</p><!><p>The SMILES of the designed RECAP library were standardized using the ChemAxon Command-Line Standardizer where the following options were selected: "Remove Fragment" (keep largest), "Neutralize", "RemoveExplicitH", "Clean2D", "Mesomerize" and "Tautomerize" [56]. The standardized canonical SMILES were exported to CSV files, and subjected to enriched target prediction using PIDGIN 1.0 implementing the method developed by Liggi et al. [24, 31]. The target prediction for the designed RECAP library was performed using a recall probability threshold of 0.01 (which is a value consistent with greater confidence in the more positive predictions).</p><p>Enrichment calculations for the predicted targets of the designed RECAP library were performed to assess the likelihood of the active compounds against the targets of interest. In this procedure, the frequency of predicting A1R, A2AR and PDE10A targets for the designed RECAP library was compared with a background distribution of a diverse library covering a large chemical space and was assessed by two parameters: the estimation score and the average ratio. The cutoff selected for considering a target as sufficiently enriched required an estimation score less than or equal to 0.01 [31]. The statistical relevance of the prediction was assessed via a Chi squared test with yates correction in Scipy [32], using the contingency table of the RECAP library and background of randomly sampled PubChem compounds (Additional file 1: Figure S2).</p><!><p>Docking with Glide [57] was performed against the human A2AR protein crystal structure (PDB ID: 4EIY) bound to the antagonist ZM241385 and the PDE10A crystal structure (PDB ID: 4DDL) complexed with an inhibitor [33, 34]. Protein structures were prepared using the protein preparation wizard of maestro 9.3 [58], following the default protocol which accounts for energy refinement, hydrogen addition, pKa assignment, and side-chain rotational isomer refinement. Resolved water molecules were discarded, and the structure was centered using the co-crystallized ligand as the center of the receptor grid generated for each protein structure. The co-crystal structures of A2AR with 4-{2-[(7-amino-2-furan-2-yl[1, 2, 4]triazolo[1,5-a][1, 3, 5]triazin-5-yl)amino]ethyl}phenol (PDB ID: 4EIY), and PDE10A with 2-{1-[5-(6,7-dimethoxycinnolin-4-yl)-3-methylpyridin- 2-yl]piperidin-4-yl}propan-2-ol (PDB ID: 4DDL), were selected as target structures.</p><p>The A1R homology model (Additional file 2) was constructed according to the method reported by Yaziji et al. [59–61], where the protein sequence of the human A1R (accession number P30542) was aligned with the A2AR template of PDB ID: 4EIY.</p><!><p>The entire set of 2563 ligands was prepared for docking with LigPrep 2.5 [62] using the default settings and the Epik option which introduces energy penalties associated with ionization and tautomerization [63].</p><!><p>In an attempt to validate the constructed A2AR, A1R, and PDE10A docking models, a set of known actives and inactives were docked against each target to ensure that they enriched actives. 81 A2AR receptor ligands reported in the literature were docked against the A2AR model [64, 65]. For consistency 81 ChEMBL actives were also selected (for each of the A1R and PDE10A proteins whose Ki and IC50 values are less than 10 µM), and these were docked against their respective target class. In addition, PubChem inactives (200 compounds) of each target class were docked.</p><p>A good separation was obtained for the medians of docking score distribution for actives versus inactives confirming that the actives are enriched. Additional file 1: Figure S5 shows the separation of the medians for the three docking models, − 6.93 (actives) versus − 5.64 (inactives) for the PDE10A docking model, − 7.66 (actives) versus − 6.01 (inactives) for the A2AR docking model, and − 7.60 (actives) versus − 5.66 (inactives) for the A1R docking model. Statistical analysis was performed with R using a Mann–Whitney test [66] on the active and inactive docking score distributions of each target. The differences in medians were significant with p values < 0.05 (script provided in Additional file 1).</p><p>The F1 score which is the harmonic mean of precision and recall, was computed (using a Python script, see Additional file 1) for all the docking scores of the ChEMBL actives and PubChem inactives for each model. A search was performed for a docking score threshold that gave the highest F1 score, in order to perform substructure analysis on compounds that were predicted as A1R/A2AR–PDE10A multi-target ligands by target prediction, and displayed docking scores that are lower than or equal to those with the highest F1 score for each of the three docking models (A1R, A2AR, and PDE10A, see Additional files 3, 4, and 5). Furthermore, the thresholds found are intended to serve as reference scores for any structure-based design problem at these target classes.</p><!><p>The RECAP compounds that were predicted as A1R/A2AR–PDE10A multi-target ligands were docked against the A2AR protein crystal structure (PDB ID: 4EIY) [33], the A1R homology model and the PDE10A protein crystal structure (PDB ID: 4DDL) [34] to investigate the molecular interactions. The Glide docking parameters used here are given in Additional file 1: Table S3. The parameters were deduced from docking experiments using known actives and inactives against each protein model.</p><!><p>Subsequently, substructure analysis was performed using DataWarrior 4.2.2, on the proposed A1R/A2AR–PDE10A multi-target ligands predicted by both ligand-based and structure-based techniques (considering docking scores less than or equal to the threshold of the best F measure for each docking model). The chemical series found were [1,2,4] triazolo[1,5-c]quinazolines (50.4%), imidazo[1,5-a]quinoxalines (14.4%), 6,7-alkoxyisoquinolines (10.6%), and 2-aminopyridine-3-carbonitriles (9.2%), in addition to various compounds consisting of the common and frequent heterocycles identified originally in the substructural analysis of the extracted ChEMBL compounds.</p><!><p>Due to both ease of the reaction and yield, a one-pot synthetic scheme was optimized for the purpose of synthesizing 2-aminopyridine-3-carbonitriles. For the other series, the synthetic routes were multi-step reactions, which due to synthetic complexity are not reported here.</p><p>The synthetic routes reported in the literature for the formation of derivatives of 6,7-alkoxyisoquinolines as selective PDE10A inhibitors involved multi-step reactions ranging from 3 to 13 steps [67, 68]. Whereas, the procedures for the synthesis of the imidazo[1,5-a]quinoxalines, known PDE10A inhibitors, consisted of 3–7 step reactions [69–72]. The [1,2,4]triazolo[1,5-c]quinazolines have been reported as potent and selective A2AR antagonists and PDE10A inhibitors, and their synthesis involved 4–7 step reactions [73–75].</p><p>Hence, given the fact that the 2-aminopyridine-3-carbonitriles were the only RECAP series that could be synthesized via a one-pot synthetic scheme [37, 76, 77], we have selected these for synthesis and subsequent validation as multi-target ligands. In particular, we selected compounds, which did not exihibit any potential PAINs liability upon screening with the FAFDrug3 ADME-Tox Filtering Tool [40].</p><!><p>Unless otherwise indicated, all starting materials, reagents and solvents were purchased and used without further purification. After extraction from aqueous phases, the organic solvents were dried over anhydrous sodium sulfate. The reactions were monitored by thin-layer chromatography (TLC) on 2.5 mm Merck silica gel GF 254 strips, and each of the purified compounds showed a single spot; unless stated otherwise, UV light and/or iodine vapor were used to detect compounds. The synthesis of the target compounds was performed in coated Kimble vials on a PLS (6 × 4) Organic Synthesizer with orbital stirring. Filtration and washing protocols for supported reagents were performed in a 12-channel vacuum manifold. The purity and identity of all tested compounds were established by a combination of HPLC, elemental analysis, mass spectrometry and NMR spectroscopy as described below. Purification of isolated products was carried out by column chromatography (Kieselgel 0.040–0.063 mm, E. Merck) or medium pressure liquid chromatography (MPLC) on a CombiFlash Companion (Teledyne ISCO) with RediSep pre-packed normal-phase silica gel (35–60 µm) columns followed by recrystallization. Melting points were determined on a Gallenkamp melting point apparatus and are uncorrected. The NMR spectra were recorded on Bruker AM300 and XM500 spectrometers. Chemical shifts are given as δ values against tetramethylsilane as internal standard and J values are given in Hz. Mass spectra were obtained on a Varian MAT-711 instrument. Analytical HPLC was performed on an Agilent 1100 system using an Agilent Zorbax SB-Phenyl, 2.1 mm × 150 mm, 5 µm column with gradient elution using the mobile phases (A) H2O containing 0.1% CF3COOH and (B) MeCN and a flow rate of 1 mL/min. The purity of all tested compounds was determined to be greater than or equal to 95%.</p><p>The synthesis of the 4,6-substituted 2-amino-pyridin-3-carbonitriles 1–25 was done via the one-pot synthetic route shown in Scheme 1. Varying both substituents on the ylidene malononitrile and the ketone reagents resulted in a variation of the substituents on positions 4 and 6 of the pyridine ring.</p><!><p>Substituted ylidene malononitrile (1.0 mmol), ketone (1.0 mmol) and ammonium acetate (5.0 mmol) in a 1:1 toluene/EtOH mixture (7 mL) were stirred in a coated Kimble vial at 120 °C for 12–24 h. After reaction completion (TLC control), distilled water was added and the mixture was extracted with ethyl acetate (3 × 10 mL). The organic phase was dried (Na2SO4) and evaporated under reduced pressure to afford an oily residue that was purified by column chromatography using n-hexane-ethyl acetate in 2:1 mixture.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.246 g, 85% yield (97% purity by HPLC). MP 226–228 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 8.08–7.95 (m, 2H), 7.69–7.58 (m, 2H), 7.60–7.47 (m, 3H), 7.23–7.09 (m, 3H), 5.34 (s, 2H). MS (EI) m/z (%): 289.07 (M+, 100), 262.07 (7). Analysis calculated for C18H12FN3: C, 74.73; H, 4.18; F, 6.57; N 14.52. Found: C, 74.70; H, 4.19; F, 6.55; N, 14.54.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.227 g, 79% yield (96% purity by HPLC). MP 241–243 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 9.92 (s, 1H), 7.99 (d, J = 8.6 Hz, 2H), 7.78–7.59 (m, 2H), 7.58–7.47 (m, 3H), 7.15 (s, 1H), 6.88 (s, 2H), 6.83 (d, J = 8.7 Hz, 2H). MS (EI) m/z (%): 287.04 (M+, 100), 259.89 (10). Analysis calculated for C18H13N3O: C, 75.25; H, 4.56; N, 14.63; O, 5.57. Found: C, 75.27; H, 4.54; N, 14.62; O, 5.59.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.172 g, 62% yield (95% purity by HPLC). MP 154–156 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.95 (d, J = 3.0 Hz, 1H), 7.72 (s, 1H), 7.66–7.65 (m, 2H), 7.52–7.50 (m, 4H), 5.30 (s, 2H). MS (EI) m/z (%): 278.03 (M+, 100), 276.97 (45). Analysis calculated for C15H10N4S: C, 64.73; H, 3.62; N, 20.13; S, 11.52. Found: C, 64.85; H, 3.48; N, 20.25; S, 11.42.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.189 g, 69% yield (98% purity by HPLC). MP 152–153 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.67–7.54 (m, 2H), 7.56–7.42 (m, 3H), 7.30 (s, 1H), 6.91 (s, 1H), 6.66–6.59 (m, 2H), 5.23 (s, 2H), 3.70 (s, 3H). MS (EI) m/z (%): 274.14 (M+, 100). Analysis calculated for C17H14N4: C, 74.43; H, 5.14; N, 20.42. Found: C, 74.57; H, 5.12; N, 20.30.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.238 g, 79% yield (97% purity by HPLC). MP 199–200 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 8.03–7.93 (m, 2H), 7.52–7.41 (m, 4H), 7.31 (dd, J1 = 7.5 Hz, J2 = 1.8 Hz, 1H), 7.17 (s, 1H), 7.11–7.02 (m, 2H), 5.27 (s, 2H), 3.88 (s, 3H). MS (EI) m/z (%): 301.16 (M+, 100), 270.12 (7), 120.10 (16.3). Analysis calculated for C19H15N3O: C, 75.73; H, 5.02; N, 13.94; O, 5.31. Found: C, 75.76; H, 5.04; N, 13.92; O, 5.33.</p><!><p>Purified by column chromatography (n-hexane–ethyl acetate 2:1) and then recrystallized from EtOH to give 0.238 g, 72% yield (99% purity by HPLC). MP 155–157 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 8.02-7.90 (m, 2H), 7.52–7.38 (m, 3H), 7.32–7.22 (m, 1H), 7.16 (s, 1H), 6.69–6.55 (m, 2H), 5.25 (s, 2H), 3.88 (s, 3H), 3.86 (s, 3H). MS (EI) m/z (%): 331.14 (M+, 100), 165.51 (9), 120.16 (11.3). Analysis calculated for C20H17N3O2: C, 72.49; H, 5.17; N, 12.68; O, 9.66. Found: C, 72.50; H, 5.19; N, 12.71; O, 9.70.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.236 g, 75% yield (96% purity by HPLC). MP 220–221 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 8.12–7.86 (m, 2H), 7.56–7.38 (m, 3H), 7.20–7.08 (m, 3H), 6.95 (d, J = 8.0 Hz, 1H), 6.06 (s, 2H), 5.33 (s, 2H). MS (EI) m/z (%): 315.11 (M+, 100), 157.52 (5). Analysis calculated for C19H13N3O2: C, 72.37; H, 4.16; N, 13.33; O, 10.15. Found: C, 72.45; H, 4.06; N, 13.49; O, 10.00.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.216 g, 78% yield (98% purity by HPLC). MP 125–126 °C. 1H NMR (300 MHz, CDCl3) δ (ppm): 7.95–7.92 (m, 1H), 7.53–7.43 (m, 3H), 7.05 (s, 1H), 6.73 (s, 1H), 5.22 (s, 2H), 2.90–2.85 (m, 2H), 1.90–1.78 (m, 4H), 1.52–1.39 (m, 4H), 1.33–1.25 (m, 1H). MS (EI) m/z (%): 277.25 (M+, 74), 246.15 (56), 222.15 (100). Analysis calculated for C18H19N3: C, 77.95; H, 6.90; N, 15.15. Found: C, 78.03; H, 6.96; N, 15.01.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.186 g, 63% yield (95% purity by HPLC). MP 126–127 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.89 (td, J = 7.8, 1.9 Hz, 1H), 7.47–7.31 (m, 1H), 7.25–7.03 (m, 3H), 5.18 (s, 2H), 2.98–2.67 (m, 1H), 1.99–1.73 (m, 5H), 1.53–1.16 (m, 5H). MS (EI) m/z (%): 295.15 (M+, 98.05), 263.05 (23.28), 251.00 (12), 240.00 (100). Analysis calculated for C18H18FN3: C, 73.20; H, 6.14; F, 6.43; N, 14.23. Found: C, 73.22; H, 6.17; F, 6.44; N, 14.25.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.236 g, 81% yield (97% purity by HPLC). MP 120–121 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.73–7.10 (m, 4H), 6.71 (s, 1H), 5.20 (s, 2H), 2.95–2.77 (m, 1H), 2.35 (s, 3H), 2.01–1.69 (m, 5H), 1.56–1.34 (m, 4H), 1.34–1.18 (m, 1H). MS (EI) m/z (%): 291.14 (M+, 100), 236.12 (48), 208.10 (91.7). Analysis calculated for C19H21N3: C, 78.32; H, 7.26; N, 14.42. Found: C, 78.48; H, 7.18; N, 14.34.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.167 g, 59% yield (98% purity by HPLC). MP 160–162 °C. 1H NMR (300 MHz, CDCl3), δ(ppm) 7.63–7.62(m, 1H), 7.44 (d, J = 4.5 Hz, 1H), 7.12–7.09 (m, 1H), 6.96 (s, 1H), 5.14 (s, 2H), 2.82–2.79 (m, 1H), 1.90–1.78 (m, 5H), 1.55–1.43 (m, 4H), 1.30–1.19 (m, 1H). MS (EI) m/z (%): 283.04 (M+, 100), 251.99 (19), 228.02 (92). Analysis calculated for C16H17N3S: C, 67.81; H, 6.05; N, 14.83; S, 11.31. Found: C, 67.89; H, 6.13; N, 14.77; S, 11.21.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.147 g, 52% yield (96% purity by HPLC). MP 145–146 °C. 1H NMR (300 MHz, CDCl3), δ(ppm) 7.94 (dd, J = 3.0, 1.3 Hz, 1H), 7.59 (dd, J = 5.1, 1.3 Hz, 1H), 7.38 (dd, J = 5.1, 3.0 Hz, 1H), 6.93 (s, 1H), 5.14 (s, 2H), 2.95–2.73 (m, 1H), 2.06–1.73 (m, 5H), 1.56–1.37 (m, 4H), 1.38–1.19 (m, 1H). MS (EI) m/z (%):(%): 283.07 (M+, 100), 228.04 (93), 214.96 (52).Analysis calculated for C16H17N3S: C, 67.81; H, 6.05; N, 14.83; S, 11.31. Found: C, 67.91; H, 6.09; N, 14.67; S, 11.33.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.174 g, 65% yield (98% purity by HPLC). MP 177–178 °C. 1H NMR (300 MHz, CDCl3), δ(ppm) 7.55 (dd, J = 1.7, 0.8 Hz, 1H), 7.06 (dd, J = 3.4, 0.8 Hz, 1H), 7.03 (s, 1H), 6.54 (dd, J = 3.5, 1.8 Hz, 1H), 5.15 (s, 2H), 3.01–2.68 (m, 1H), 2.04–1.74 (m, 5H), 1.55–1.39 (m, 4H), 1.34–1.20 (m, 1H). MS (EI) m/z (%): 267.11 (M+, 100), 212.02 (69). Analysis calculated for C16H17N3O: C, 71.89; H, 6.41; N, 15.72; O, 5.98. Found: C, 71.91; H, 6.43; N, 15.71.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.188 g, 59% yield (97% purity by HPLC). MP 180–181 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.96 (td, J1 = 7.8, J2 = 1.9 Hz, 1H), 7.65–7.58 (m, 2H), 7.47–7.37 (m, 1H), 7.31–7.23 (m, 2H), 7.23–7.09 (m, 1H), 7.09–6.98 (m, 2H), 5.32 (s, 2H), 3.88 (s, 3H). MS (EI) m/z (%): 319.12 (M+, 100), 304.18 (12), 249.13 (16). Analysis calculated for C19H14FN3O: C, 71.46; H, 4.42; F, 5.95; N, 13.16; O, 5.01. Found: C, 71.48; H, 4.44; F, 5.97; O, 5.05.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.205 g, 65% yield (95% purity by HPLC). MP 151–152 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.61 (d, J = 8.3 Hz, 2H), 7.40 (d, J = 7.3 Hz, 1H), 7.37–7.27 (m, 3H), 7.03 (d, J = 8.2 Hz, 2H), 6.86 (s, 1H), 5.32 (s, 2H), 3.87 (s, 3H), 2.42 (s, 3H). MS (EI) m/z (%): 314.10 (M+, 100), 271.06 (7), 208.11 (52). Analysis calculated for C20H17N3O: C, 76.17; H, 5.43; N, 13.32; O, 5.07. Found: C, 76.31; H, 5.33; N, 13.52; O, 4.84.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.198 g, 68% yield (99% purity by HPLC). MP 205–207 °C. 1H NMR (300 MHz, CDCl3) δ (ppm): 7.65–7.54 (m, 3H), 7.16 (s, 1H), 7.11 (d, J = 3.5 Hz, 1H), 7.03 (d, J = 8.8 Hz, 2H), 6.62–6.51 (m, 1H), 5.30 (s, 2H), 3.88 (s, 3H). MS (EI) m/z (%): 291.12 (M+, 100), 145.63 (5). Analysis calculated for C17H13N3O2: C, 70.09; H, 4.50; N, 14.42; O, 10.98. Found: C, 70.21; H, 4.38; N, 14.68, O, 10.73.</p><!><p>Purified by column chromatography (n-hexane- ethyl acetate 2:1) and then recrystallized from EtOH to give 0.222 g, 70% yield (99% purity by HPLC). MP 248–250 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 9.89 (s, 1H), 7.98 (d, J = 8.7 Hz, 2H), 7.61 (d, J = 8.7 Hz, 2H), 7.11–7.06 (m, 3H), 6.84–6.81 (m, 4H), 3.82 (s, 3H). MS (EI) m/z (%): 317.17 (M+, 100), 302.04 (6), 158.50 (14). Analysis calculated for C19H15N3O2: C, 71.91; H, 4.76; N, 13.24; O, 10.08. Found: C, 71.94; H, 4.79; N, 13.25; O, 10.11.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.219 g, 73% yield (98% purity by HPLC). MP 180–181 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 8.05–7.90 (m, 1H), 7.56–7.41 (m, 2H), 7.33–7.06 (m, 6H), 5.34 (s, 2H). MS (EI) m/z (%): 307.06 (M+, 100), 279.99 (8). Analysis calculated for C18H11F2N3: C, 70.35; H, 3.61; F, 12.36, N, 13.67. Found: C, 70.37; H, 3.63; F, 12.33; N, 13.66.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.245 g, 78% yield (97% purity by HPLC). MP 187–188 °C. 1H NMR (300 MHz, CDCl3), δ (ppm) 7.97 (td, J = 7.8, 1.9 Hz, 1H), 7.52–7.35 (m, 2H), 7.31 (td, J = 7.2, 1.5 Hz, 1H), 7.26–7.19 (m, 2H), 7.17–6.95 (m, 3H), 5.27 (s, 2H), 3.88 (s, 3H). MS (EI) m/z (%): 319.12 (M+, 100), 290.14 (7), 138.01 (14). Analysis calculated for C19H14N3FO: C, 71.46; H, 4.42; F, 5.95; N, 13.16; O, 5.01. Found: C, 71.44; H, 4.43; F, 5.92; O, 5.04.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.186 g, 64% yield (98% purity by HPLC). MP 181–183 °C. 1H NMR (300 MHz, CDCl3) δ (ppm): 7.47–7.40 (m, 2H), 7.32–7.28 (m, 4H), 7.09–7.02 (m, 2H), 6.86 (s, 1H), 5.29 (s, 2H), 3.88 (s, 3H), 2.43 (s, 3H). MS (EI) m/z (%): 315.13 (M+, 100), 298.16 (12), 284.09 (18), 208.10 (81.6). Analysis calculated for C20H17N3O: C, 76.17; H, 5.43; N, 13.32; O, 5.07. Found: C, 76.19; H, 5.41; N, 13.36; O, 5.03.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.244 g, 77% yield (96% purity by HPLC). MP 187–188 °C. 1H NMR (300 MHz, CDCl3), δ (ppm): 7.55 (s, 1H), 7.44 (t, J = 8.1 Hz, 1H), 7.30 (dd, J = 7.4, 1.7 Hz, 1H), 7.15–6.98 (m, 4H), 6.54 (dd, J = 3.3, 1.7 Hz, 1H), 5.24 (s, 2H), 3.87 (s, 3H). MS (EI) m/z (%): 291.10 (M+, 100), 262.14 (10). Analysis calculated for C17H13N3O2: C, 70.09; H, 4.50; N, 14.42; O, 10.98. Found: C, 70.11; H, 4.51; N, 14.41; O, 11.01.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.193 g, 60% yield (96% purity by HPLC). MP 210–212 °C. 1H NMR (300 MHz, DMSO-d6), δ (ppm): 9.91 (s, 1H), 7.93 (d, J = 9.0 Hz, 2H), 7.45 (t, J = 7.8 Hz, 1H), 7.29 (dd, J = 7.4, 1.7 Hz, 1H), 7.16 (d, J = 8.3 Hz, 1H), 7.07 (d, J = 7.5 Hz, 1H), 7.03 (s, 1H), 6.82 (d, J = 8.9 Hz, 2H), 6.77 (s, 2H), 3.77 (s, 3H). MS (EI) m/z (%): 317.13 (M+, 100), 300.09 (8), 286.11 (6).Analysis calculated for C19H15N3O2: C, 71.91; H, 4.76; Cl, 13.24; O, 10.08. Found: C, 71.92; H, 4.74; Cl, 13.27; O, 10.05.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.179 g, 59% yield (98% purity by HPLC). MP 215–217 °C. 1H NMR (300 MHz, DMSO-d6), δ (ppm): 9.90 (s, 1H), 8.16–7.22 (m, 2H), 7.69–7.30 (m, 4H), 7.16–6.50 (m, 5H). MS (EI) m/z (%): 320.99 (M+, 100), 286.04 (5). Analysis calculated for C18H12ClN3O: C, 67.19; H, 3.76; Cl, 11.02; N, 13.06; O, 4.97. Found: C, 67.37; H, 3.94; Cl, 11.18; N, 12.88; O, 4.63.</p><!><p>Purified by column chromatography (n-hexane–ethyl acetate 2:1) and then recrystallized from EtOH to give 0.151 g, 53% yield (97% purity by HPLC). MP 299–300 °C. 1H NMR (300 MHz, DMSO-d6), δ (ppm) 9.92 (s, 2H), 8.19–7.79 (m, 2H), 7.68–7.37 (m, 2H), 7.42–6.99 (m, 1H), 7.01–6.62 (m, 6H). MS (EI) m/z (%): 303.06 (M+, 100), 184.01 (6). Analysis calculated for C18H13N3O2: C, 71.28; H, 4.32; N, 13.85; O, 10.55. Found: C, 71.40; H, 4.54; N, 13.75; O, 10.31.</p><!><p>Purified by column chromatography (n-hexane-ethyl acetate 2:1) and then recrystallized from EtOH to give 0.123 g, 46% yield (95% purity by HPLC). MP 156–157 °C. 1H NMR (300 MHz, CDCl3) δ(ppm): 8.01 (dd, J = 3.0, 1.2 Hz, 1H), 7.66 (dd, J = 5.1, 1.2 Hz, 1H), 7.62 (dd, J = 1.8, 0.6 Hz, 1H), 7.48 (dd, J = 3.6, 0.6 Hz, 1H), 7.45 (s, 1H), 7.40 (dd, J = 5.1, 3.0 Hz, 1H), 7.40 (dd, J = 5.1, 3.0 Hz, 1H), 6.61 (dd, J = 3.6, 1.8 Hz, 1H), 5.26 (s, 2H). MS (EI) m/z (%): 267.06 (M+, 100), 237.98 (6), 210.99 (7). Analysis calculated for C14H9N3OS: C, 62.91; H, 3.39; N, 15.72; O, 5.99; S, 11.99. Found: C, 63.11; H, 3.47; N, 15.58; O, 5.97; S, 11.87.</p><!><p>Pharmacological evaluation was performed in a radioligand binding competition assay, using A1, A2A, A2B, and A3 human receptors expressed in transfected CHO (A1), HeLa (A2A and A3), and HEK-293 (A2B) according to the procedure reported by Bosch et al. [78].</p><p>The activity measurements against the phosphodiesterases PDE7A, PDE7B, PDE9A and PDE10A were performed using AD293 cells that were transiently and separately transfected with human PDE7A, PDE7B, PDE9A, and PDE10A following the procedure described by Shipe et al. [43]. The IC50 values were obtained by fitting the data with non-linear regression using Prism 2.1 software (GraphPad, San Diego, CA) [79], and the reported results are the mean of 3 experiments (n = 3) each performed in duplicate.</p><!><p>Additional file 1. Supplementary data describing substructural analysis of extracted ChEMBL compounds, statistical analysis of enriched target prediction of RECAP compounds, separation in medians of active/inactive docking score distributions for the docking models, computed logP and tPSA values and selectivity profiling data for compounds 1–25, docking parameters used, scripts for compound extraction from the ChEMBL database, computation of Mann–Whitney test and F1 scores.</p><p>Additional file 2. Coordinates of the A1R homology model.</p><p>Additional file 3. CSV file of computed F1 scores of the A1R docking model.</p><p>Additional file 4. CSV file of computed F1 scores of the A2AR docking model.</p><p>Additional file 5. CSV file of computed F1 scores of the PDE10A docking model.</p><p>adenosine receptor</p><p>A1 adenosine receptor</p><p>A2A adenosine receptor</p><p>A2B adenosine receptor</p><p>A3 adenosine receptor</p><p>cAMP and cAMP-inhibited cGMP 3′,5′-cyclic phosphodiesterase 10A</p><p>cyclic adenosine monophosphate</p><p>high affinity cAMP-specific 3′,5′-cyclic phosphodiesterase 7A</p><p>cAMP-specific 3′,5′-cyclic phosphodiesterase 7B</p><p>high affinity cGMP-specific 3′,5′-cyclic phosphodiesterase 9A</p><p>platelet-derived growth factor receptor</p><p>vascular endothelial growth factor receptor</p><p>tyrosine-protein kinase ABL1</p><p>proto-oncogene tyrosine-protein kinase Src</p><p>epidermal growth factor receptor</p><p>receptor tyrosine-protein kinase erbB-2</p><p>amine oxidase [flavin-containing] A</p><p>amine oxidase [flavin-containing] B</p><p>acetylcholinesterase</p><p>butyrylcholinesterase</p><p>histamine H3 receptor</p><p>histone methyltransferases</p><p>human cannabinoid receptor 2</p><p>ethanol</p><p>Electronic supplementary material</p><p>The online version of this article (10.1186/s13321-017-0249-4) contains supplementary material, which is available to authorized users.</p>

PubMed Open Access

High-throughput secretomic analysis of single cells to assess functional cellular heterogeneity

Secreted proteins dictate a range of cellular functions in human health and disease. Due to the high degree of cellular heterogeneity and, more importantly, polyfunctionality of individual cells, there is an unmet need to simultaneously measure an array of proteins from single cells and to rapidly assay a large number of single cells (more than 1000) in parallel. We describe a simple bioanalytical assay platform consisting of a large array of sub-nanoliter microchambers integrated with high-density antibody barcode microarrays for highly multiplexed protein detection from over a thousand single cells in parallel. This platform has been tested for both cell lines and complex biological samples such as primary cells from patients. We observed distinct heterogeneity among the single cell secretomic signatures that, for the first time, can be directly correlated to the cells\xe2\x80\x99 physical behavior such as migration. Compared to the state-of-the-art protein secretion assay such as ELISpot and emerging microtechnology-enabled assays, our approach offers both high throughput and high multiplicity. It also has a number of clinician-friendly features such as ease of operation, low sample consumption and standardized data analysis, representing a potentially transformative tool for informative monitoring of cellular function and immunity in patients.

high-throughput_secretomic_analysis_of_single_cells_to_assess_functional_cellular_heterogeneity

INTRODUCTION<!>Design, fabrication and assembly of a single-cell secretomic analysis chip<!>Protein panel and validation<!>Single-cell protein secretomic analysis on cell lines<!>Correlation between secretomic signature and migratory property<!>Secretomic profiling of single tumor cells from clinical patient specimens<!>CONCLUSIONS

<p>Secreted proteins including cytokines, chemokines and growth factors represent important functional regulators mediating a range of cellular behavior and cell-cell paracrine/autocrine signaling, e.g. in the immunological system1, tumor microenvironment2 or stem cell niche3. Detection of these proteins is of great value not only in basic cell biology but also for disease diagnosis and therapeutic monitoring. However, due to co-production of multiple effector proteins from a single cell, referred to as polyfunctionality, it is biologically informative to measure a panel of secreted proteins, or secretomic signature, at the level of single cells. Recent evidence further indicates that a genetically-identical cell population can give rise to diverse phenotypic differences4. Non-genetic heterogeneity is also emerging as a potential barrier to accurate monitoring of cellular immunity and effective pharmacological therapies5,6, suggesting the need for practical tools for single cell analysis of proteomic signatures.</p><p>Fluorescence-activated cell sorting (FACS) represents the state-of-the-art for single cell analysis7. FACS is typically used to detect and sort cell phenotypes by their surface markers. It has been extended to the detection of intracellular proteins7–9, including cytokines within the cytoplasm, by blocking vesicle transport10. However, intracellular cytokine staining is not a true secretion analysis, and it also requires cell fixing, which means the cells are no longer alive after flow cytometric analysis and cannot be recovered for further studies. The mainstay of real single cell secretion analysis to date is a simple approach called ELISpot that detects the secretion footprint of individual cells using an immunosandwich-based assay11. Immune cells are loaded into a microtiter plate that has been pre-coated with a layer of primary antibody. After incubation, secreted proteins are captured by the antibodies located proximal to the cells, giving rise to spots indicative of a single cell secretion footprint 12. Recently, a variant of ELISpot, called FLUOROSpot, which exploits two fluorescent dyes to visualize protein secretion footprints, enabled a simultaneous dual function analysis. Highly multiplexed measurements of proteins secreted from a population of cells can be done using an encoded bead assay such as the Illumina VeraCode system13 or antibody microarrays manufactured using a pin-spotting technique 14,15. However, these highly multiplexed technologies cannot perform single cell measurements. Microfabricated chips have emerged as a new category of single cell analytic technologies16–21. A prototype microchip has demonstrated the feasibility of multiplexed protein secretion assay and revealed significant polyfunctional heterogeneity in phenotypically similar immune cells from patients22,23, pointing to the urgent need for single cell secretion profiling in clinical diagnosis and therapeutic monitoring. However, these microchips either lack sufficient throughput or multiplicity, or require sophisticated operation, precluding widespread application in cell biology and clinical evaluation of cellular functions.</p><p>Herein we describe a high-throughput single-cell secretomic analysis platform that integrates a sub-nanoliter microchamber array and high-density antibody barcodes for simultaneous detection of 14 cytokines from more than a thousand single cells in parallel. The chip can be executed in a simple assay "kit" with no need of sophisticated fluid handling or bulky equipment. We demonstrate the utility of this device for analyzing the secretion of human cell lines and primary cell samples dissociated from fresh tumor of patients. The results reveal that there is distinct heterogeneity among the single cell secretomic signatures of a population and that the correlations obtained between the various proteins studied are in agreement with their functional classifications. This technology builds upon prior successes in antibody barcode-based protein secretion measurement technique22, but uses simplified schemes of cell capture24, quantification, automated data analysis, and eliminates bulky fluid handling systems, resulting in a truly practical and informative tool that may find immediate use in both laboratory research and clinical cellular diagnosis.</p><!><p>Our single-cell secretomic analysis device consists of two separate parts (Fig. 1a): a high-density antibody barcode encoded glass substrate for surface-bound immunoassay and a sub-nanoliter microchamber array for capture of single cells. The antibody barcode array slide comprises 30 repeats of barcodes, each of which contains up to 20 stripes of different antibodies, immobilized on a poly-L-lysine-coated surface. The antibody stripes are 20μm in width and the pitch size of a full barcode is 1mm. The microchamber array is a one-layer microchip fabricated by soft lithography25 from polydimethylsiloxane (PDMS)25, an optically transparent silicone elastomer widely used for biological microfluidics. It contains 5440 rectangular microchambers, each of which is 1.8mm, 20μm and 15μm, in length, width and depth, respectively. These two parts were manufactured independently and combined during the assay such that the barcode slide acts as a disposable test strip and the microchamber array as a reusable device. To use this platform, a drop of single cell suspension (~106cells/ml) is directly pipetted onto the surface of the microchamber array chip. The cells fall into the microchambers by gravity, and then the aforementioned antibody barcode array slide is placed antibody-side down on top of the microchambers such that the antibody barcodes are perpendicular to the length of the microchambers. The microchamber is designed to be sufficiently long as to contain at least a full set of barcodes, thereby eliminating the need for precise alignment. Finally this assembly is fixed by two transparent plastic plates with four spring-adjusted screws (Supporting Fig. S1) and placed in a conventional tissue incubator for single-cell secretion measurement. Proteins secreted from individual cells are captured by the antibody barcodes and read out by incubating with biotinylated detection antibodies and then streptavidin conjugated with a fluorescence probe (e.g. Cy5 in our experiments). As compared to the prototype single cell proteomic chip22, this setup does not require a sophisticated microfluidic control system or any bulky equipment to operate and thus is more amenable to widespread use by researchers and clinicians with minimal engineering background.</p><p>The high-density antibody barcode array is fabricated using a microchannel-guided flow patterning technique (Supporting Methods) modified from the approach reported for patterning DNA barcodes26. The flow-patterning chip is a separate PDMS slab that has inlets leading to 20 individual serpentine microchannels in which individual antibody solutions (1μL each) are precisely metered, added and flowed through all microchannels in parallel to ensure uniform loading of antibodies on the surface. Fluorescein isothiocyanate labeled bovine serum albumin (FITC-BSA) solution was used to evaluate the patterning quality. The result shows successful fabrication of high-density protein array across a large area (1in × 2in) and excellent uniformity (< 5% in fluorescent intensity) as revealed by the fluorescence intensity line profile (Fig. 1b and Supporting Fig. S2). This ensures the observed protein signal variations (> 10%) from the following single cell secretomic assays are attributed to cellular heterogeneity, rather than the non-uniformity of the starting antibody barcode array.</p><p>A motorized phase-contrast imaging system has been developed to image all cells in the cell capture chip within 10 minutes (Fig. 1c) and an image analysis algorithm allows for identification of individual cells and their x/y coordinates, and counting of cells in each microchamber. The simple microchamber array chip format was chosen because it is easy to operate, but as a consequence it is not possible to ensure that one cell is captured per chamber. However, optimization of cell density in the stock solution readily gives rise to more than 1000 single cell chambers in a microchip (Supporting Fig. S3), permitting high-throughput analysis of single cells.</p><!><p>The proteins assayed by the antibody barcodes are listed in Figure 2a. Assessment of these particular proteins secreted from single cells is of particular importance due to their functions in a range of cellular processes27–30. They include cytokines, chemokines and growth factors involved in a wide range of immunological or pathophysiological processes. Assessment of these proteins secreted from single cells is of importance in the study of cellular immunity and cell-cell signaling networks. In order to simultaneously measure these proteins from single cells, capture antibodies are immobilized on the substrate as a high-density barcode array. Prior to performing single-cell analysis, we validated the assay using recombinant proteins. Individual recombinant protein was spiked into fresh cell culture medium over a 4-log range of concentrations and exposed to the full panel of antibodies in order to assess cross-reactivity, the limit of detection (LOD) and dynamic range. The antibodies with cross-reactivity over 5% (at 5ng/mL of protein concentration) is eliminated or replaced. Ultimately we obtained a panel of antibody pairs as summarized in Supporting Table S1. The titration curves (Fig. 2b) demonstrate the feasibility of quantitative measurements of these proteins in the multiplexed array, with a typical measurement range of 3 orders of magnitude. The LOD ranged from 400pg/ml to below 10pg/ml depending on the affinity of antibody pairs. Based on the volume of a microchamber (~0.54 nl) and the representative detection sensitivity (~10 pg/ml), the amount of protein that can be detected by antibody barcodes in a microchamber is on the order of 5.4 attograms, which is approximately equal to ~160 molecules. Thus, our platform has the sensitivity to detect proteins secreted from a single cell (typical copy number ~102–5).</p><!><p>We first used the single-cell secretomic analysis chip to measure 14 proteins from a human glioblastoma multiforme cell line (U87). In this experiment, up to 10 cells were captured in each microchamber, with 1278 of the microchambers capturing single cells. During the flow patterning of antibody barcodes, FITC-BSA (0.5mg/ml) was always flowed in Channel 1 to form a continuous line of fluorescence signal serving as both a position reference and an internal quality/uniformity control. As shown in a representative region of the scanned fluorescence image (Fig. 3a and Supporting Fig. S4a), both the blue FITC-BSA reference line and the red patterned signals corresponding to protein secretion levels are readily visible. Shown in the same figure are a bright field image of 14 microchambers with cells loaded, the corresponding fluorescent barcode image, and an overlay of the two. The major proteins observed after 24 hours of incubation (FGF, VEGF, MIF, IL-6, IL-8, and MCP-1) are mainly pro-inflammatory cytokines or chemoattractant proteins.</p><p>We conducted automated quantitation of the fluorescence intensity of each protein in a microchamber using the image analysis software Genepix 6.0. We extracted the secretomic profile for only those microchambers containing single cells and a heat map of the resulting secretion profiles (Fig. 3b and Supporting Fig. S5) indicates the existence of cell-cell variation. While the majority of cells produce IL-6, and IL8, the level of these proteins varies among individual cells and the secretion of other lower abundance proteins such as MCP-1 and FGF apparently exhibit heterogeneous signatures – only a small fraction of cells express these proteins at high levels. To verify the single cell measurements, a kinetic bulk population secretion measurement was performed in parallel on the supernatant collected from the same cells, incubated over the same time, and measured using a conventional pin-spotted microarray. The result (Fig. 3d and Supporting Fig. S6) also reveals FGF, VEGF, IL-6, IL-8, and MCP-1 as the top five proteins that are all consistent with single-cell analysis although the relative levels are different. However, MIF did not show up in the population assay. Interestingly, we observed that the protein level as measured by fluorescence intensity is not always proportional to the number of cells and sometimes cannot be interpreted by an additive effect (Fig. 3c and Supporting Fig. S7). A secretomic analysis chip was loaded with many more cells, MIF signal decreases with increasing number of cells in the capture chambers, revealing possibility of paracrine signaling and the regulation of MIF with increasing number of cells (Supporting Fig. S7). This small level of discrepancy is expected as the two assays are not biologically identical (for example, the bulk assay detects the end point protein profile while the single cell assay measures accumulated signals over the period of incubation; the population arrays are subjected to paracrine signaling while single cell measurements are not). Overall, these comparative studies are in good agreement with each other and demonstrated the validity of the single cell secretomic analysis microchip. Another advantage of our platform is that it also measures proteins secreted from multiple cells at the same time. While IL-6 and IL-8 secretion increases with increasing number of cells in a microchamber, the amount of MCP-1 or MIF increase does not change significantly when cell number exceeds 2, suggesting the existence of a possible mechanism similar as 'quorum sensing' in which the paracrine mechanism in the multi-cellular system controls homeostasis.</p><p>The single cell secretomic analysis chip was also used to measure two additional cell lines in order to assess the broad applicability of this platform. The first is an immune cell line (U937). The cells are human monocytes, which can be stimulated with phorbol myristate acetate (PMA) to differentiate into functional macrophage cells and then challenged by cytotoxin lipopolysaccharide (LPS) to stimulate cytokine production. This process emulates the inflammatory immune response of human macrophages to Gram-negative bacteria31. The major proteins observed are RANTES, TNFα, MCP-1, IL6 and IL-8 (Supporting Fig. S4b). While the majority of cells produce RANTES, IL8 and TNFα, the level of these proteins varies among individual cells and the secretion of other lower abundance proteins such as MCP-1 and IL-6 exhibit heterogeneous signatures. A bulk population secretion measurement was performed in parallel on the supernatant collected from the same cells to verify the single cell experiments (Supporting Fig. S8). The result also reveals RANTES, IL8, MCP-1, IL-6 and TNFα as the top five proteins, consistent with the single cell analysis. IL-8 and RANTES secretion increases with increasing number of cells in a microchamber, the amount of MCP-1 or TNFα increases does not change significantly when cell number exceeds 4 (Supporting Fig. S4c). The second is a lung carcinoma cell line (A549) that constitutively produces cytokines or growth factors. Therefore, we measured the basal level secretion from these cells with no stimulation. The major proteins observed are MCP-1, IL-6, IL-8, VEGF and FGF (Supporting Fig. S9) that were also validated by the bulk population assay using standard pin-spotted antibody microarray assays (Supporting Fig. S10). The proteins secreted from A549 cells include both pro-inflammatory cytokines and growth factors, in agreement with the role of lung tumor cells in both maintaining tumor growth and promoting an inflammatory microenvironment. Altogether, the cell line studies demonstrated that our platform is capable of rapid, quantitative and high throughput analysis of protein secretion profiles in single cells compared to current conventional methods such as ELISpot. Validating our platform with cell lines allow us to expand our sample repertoire to include more complex samples such as tissue specimens from patients.</p><!><p>Although flow cytometry-based single cell analysis allows for multiplexed protein measure, the measured protein profile cannot be directly correlated to the cell's behavior and activity such as migratory property. Our platform utilizes live cell imaging to count captured cells, thus permitting simultaneous measurement of cellular behavior and subsequent correlation to the corresponding protein profile of the same cell. Herein we measure the migration of lung cancer cells (A549) loaded in the single cell secretomic analysis chip by measuring the distance of movement before incubation and after 24 hours of incubation (Fig. 4a). These cells were seen to adhere to the channel wall and migrate at varying speeds. The results are summarized in a heatmap showing single-cell secretomic profiles sorted by increasing cell migration distance(Supporting Fig. S11). We performed p-value analysis of cytokine levels in high motility (top 20% over the observed range) vs low motility (bottom 20% of the same range) cells. While the majority of the cells do not migrate, the highly migratory cells are statistically associated with high expression of IL-8 (P < 0.01) (Fig. 4b). the correlation between the secretion of MCP-1 and cell migration was less significant (Fig. 4c, d). While IL-6 appears to be negatively associated with cell motility in the scatter plots, but does not show statistical correlation using the aforementioned test. These proteins have been linked to the increase of motility and metastatic potential in different cancers32–34, and through the investigation of single cell IL-8 secretion, it may be possible to study the secretomic signatures of individual cells linked to metastasis. In brief, our platform for the first time shows simultaneous measurement of protein secretomic signature and phenotypic properties (e.g. migration) of single live cells that can lead to improved understanding of cellular functions and the underlying molecular mechanisms.</p><!><p>To expand the utility of our platform to measuring multiplexed secretion in cells derived from complex biospecimens, we also applied our device to the measurement of fresh primary tumor tissue from three patients (Supporting Table S2) with malignant brain tumor, glioblastoma multiforme (Patients 1&2), or meningioma (Patient 3). A portion (<0.2g) of the surgically-resected tumor tissue is washed with ice cold phosphate-buffered saline, minced into smaller fractions and then dissociated into a single cell suspension using collagenase (Fig. 5a and Supporting Methods). The cells were spun down and re-suspended in medium at a density of ~106 cells/mL. Within 1 hour of tissue procurement, the single cell suspension is loaded onto the single-cell secretomic analysis device via pipette. After allowing the cells to secrete cytokines for 12 hr, the pattern on the barcode array is developed with detection antibodies and scanned. A raw fluorescent image (Fig. 5b, Patient 1) shows excellent protein signals and similar background compared to the scanned image from cell lines. The antibody barcode array includes 14 proteins as shown in Figure 5b. In this experiment, between 0 to 22 cells were captured within a microchamber, with 1058 of the microchambers capturing single cells. We quantified the fluorescence intensities of each secreted cytokine from each individual channel, and then generated a heat map of the single cell secretion profiles (Fig 5c). Unsupervised hierarchical clustering of the single cell secretion profiles resolved three separate populations of cells with varying activity. One cluster of cells (Fig. 5c, blue cluster) was generally more active, secreting a wider range of proteins presumably corresponding to more aggressive phenotype, while the cells indicated by green exhibit the lowest level of cytokine production and may represent more quiescent phenotypes such as tumor stem/progenitor cells35. The large fraction indicated by orange are a variety of functional phenotypes. The result from the Patient 2 (Fig. 5d) shows similarities to the results from Patient 1, such as MIF and IL-8 as major proteins, but different pattern in that it has much reduced production of inflammatory cytokines and higher level of EGF. The second tier proteins all show distinct cellular heterogeneity. Supporting Figure S12 and Supporting Figure S13 presents histograms and scatter plots of individual proteins, which show both the relative levels of proteins and the distributions amongst the cell population.</p><p>We compiled pseudo-three-dimensional scatter plots of the single cell cytokine measurements for the patient primary tumors in the format of flow cytometric plots and formed a 14×14 mosaic matrix (Fig. 5e). The proteins are shown at the diagonal line and each panel is a pair-wise correlation plot, for each of which we performed a linear regression analysis to yield the R value. Then the whole matrix is color-coded by red (positive correlation) and blue (negative correlation), and the color intensity is proportional to R. In the Patient 1 matrix, all the inflammatory cytokines are apparently associated within one cluster and several growth factors are grouped in a separate cluster, reflecting their functional difference. In the result for Patient 2, the pro-inflammatory cytokines, although generally expressed at low levels, also show inter-correlation. Interesting, the secretion of EGF is negatively correlated to pro-inflammatory and chemoattactant proteins (MCP-1, GMCSF and IL-8). We also analyzed a third sample from a patient (Patient 3) with transitional meningioma, which is considered a more homogeneous and less inflammatory tumor. Indeed our results for Patient 3 (Supporting Fig. S14 and Supporting Fig. S15) show reduced pro-inflammatory cytokine signals. These studies, while still very preliminary, imply the relevance of our results to these cells' physiological functions or pathological condition. Currently surgical treatment remains the most effective therapy of human glioblastoma. Afterwards, chemotherapy might be carried out systemically or by putting drug-containing wafers to the surgical cavity to further eradicate invasive tumor cells that have diffused to normal brain tissue36. Our platform potentially can distinguish and quantitate invasive cell phenotypes as they generally produce more cytokines as well as different profile of cytokines, which has the clinical value to determine tumor invasiveness and tailor the chemotherapeutic strategy for individual patients. In addition, these proteins that act as the soluble signals to mediate cell-cell communication in tumor microenvironment may be identified as new therapeutic targets for personalized treatment37–39.</p><!><p>Single cell proteomic analysis has generally been much more challenging than genetic analysis from single cells, due to the lack of equivalent amplification methods for proteins such as polymerase chain reaction (PCR) for nucleic acids. Recent advance in flow cytometric analysis allows for 34-plexed measurement of protein markers from single cells, but most proteins are either surface receptors or cytoplasmic proteins40. Intracellular cytokine staining (ICS) enables indirect assessment of "secreted" proteins, but currently the number of cytokines that can be measured is practically limited to below 10, presumably due to increased non-specific binding from a large number of antibodies in the limited volume of a single cell. Moreover, unlike protein secretion, it is not a direct measurement of cell function. Thus, multiplexed protein secretion measurement is a missing piece of functional characterization of single cells. It has become increasingly evident that even genetically homogeneous cells can be extremely heterogeneous, leading to many unanswered questions in studying their biology41. Studying the secretion profile of single cells can reveal much more about tumor heterogeneity than studying the signaling patterns of cells in population wherein the signals become averaged out and all defining information is lost, emphasizing the need for studying single cell secretion42,43.</p><p>We have described a sub-nanoliter multiplexed immunoassay chip that enables high throughput, simultaneous detection of a panel of 14 cytokines secreted from over a thousand single cells in parallel. This platform provides significant advantages specific to the detection of secreted proteins and offers information complementary to that obtained through flow cytometry. An example scenario where this device would offer unique advantages is that when a cell separation tool is used to sort out a phenotypically identical cell population using specific surface markers, these cells can then be placed in our device to further reveal cellular heterogeneity at the functional level. For instance, human T cell lineages often display a number of functions and the complex combinations of multiple functions in a single T cells dictates the "quality" of this cell in response to a specific antigen. Recent studies showed that multifunctional T cells often exhibit greater potency and durability.44 The latest HIV vaccine trials employed the ELISpot technique to count interferon-γ(IFN-γ)-secreting T cells as a means to assess the efficacy of vaccination, but it turns out that most IFN-γ-secreting cells are terminally differentiated effector T cells and have minimal protective effect against viral infection. Our platform represents a promising tool to perform polyfunctional analysis on the cells isolated from flow cytometer or other separation techniques, e.g. magnetically assisted cell sorting45, to bring single-cell protein assay to another level of functional analysis. A potential concern of our platform is that cells are isolated in the sealed micro-chamber and may experience a condition that affects the normal functioning of primary cells ex vivo. As a bioanalytical tool, our microchip was not intended to perform long-term culture of cells and the typical assay time is a few hours to 1 day. It has been reported that ex vivo assay of primary cells in sealed and isolated environment do produce proteins over a long time as anticipated for their intrinsic physiological activity 22,46 and interestingly the cells could gain greater viability in a sealed nL-chamber because it recapitulates the in vivo crowdedness in primary tissue and retains sufficient concentrations of cytokines for more effective autocrine signaling. Thus, our microchip is a promising platform for high-throughput analysis of protein secretion profiles from single primary cells and may assist in differential diagnosis and monitoring of cellular functions in patients.</p>

PubMed Author Manuscript

Mechanism of Error-Free DNA Replication Past Lucidin-Derived DNA Damage by Human DNA Polymerase \xce\xba

DNA damage impinges on genetic information flow and has significant implications in human disease and aging. Lucidin-3-O-primeveroside (LuP) is an anthraquinone derivative present in madder root, which has been used as a coloring agent and food additive. LuP can be metabolically converted to genotoxic compound lucidin, which subsequently forms lucidin-specific N2-2\xe2\x80\xb2-deoxyguanosine (N2-dG) and N6-2\xe2\x80\xb2-deoxyadenosine (N6-dA) DNA adducts. Lucidin is mutagenic and carcinogenic in rodents but has low carcinogenic risks in humans. To understand the molecular mechanism of low carcinogenicity of lucidin in humans, we performed DNA replication assays using site-specifically modified oligodeoxynucleotides containing a structural analogue (LdG) of lucidin-N2-dG DNA adduct and determined the crystal structures of DNA polymerase (pol) \xce\xba in complex with LdG-bearing DNA and an incoming nucleotide. We examined four human pols (pol \xce\xb7, pol \xce\xb9, pol \xce\xba, and Rev1) in their efficiency and accuracy during DNA replication with LdG; these pols are key players in translesion DNA synthesis. Our results demonstrate that pol \xce\xba efficiently and accurately replicates past the LdG adduct, whereas DNA replication by pol \xce\xb7, pol \xce\xb9 is compromised to different extents. Rev1 retains its ability to incorporate dCTP opposite the lesion albeit with decreased efficiency. Two ternary crystal structures of pol \xce\xba illustrate that the LdG adduct is accommodated by pol \xce\xba at the enzyme active site during insertion and postlesion-extension steps. The unique open active site of pol \xce\xba allows the adducted DNA to adopt a standard B-form for accurate DNA replication. Collectively, these biochemical and structural data provide mechanistic insights into the low carcinogenic risk of lucidin in humans.

mechanism_of_error-free_dna_replication_past_lucidin-derived_dna_damage_by_human_dna_polymerase_\xce

INTRODUCTION<!>Materials<!>Primer Extension Assays<!>NanoLC\xe2\x80\x93MS/MS Analysis of Primer-Extension Products by Pol \xce\xba<!>Crystallization of DNA:Pol \xce\xba:dNTP Ternary Complexes<!>Pol \xce\xba Efficiently Replicates Past LdG Adduct<!>dNTP Incorporation Opposite LdG Adduct Is Mainly Error-Free<!>Postlesion DNA Synthesis Is Mostly Error-Free<!>Pol \xce\xba-Catalyzed Bypass of LdG Is Error-Free As Revealed by LC\xe2\x80\x93MS-Based Oligodeoxynucleotide Sequencing<!>Structural Basis of Pol \xce\xba-Catalyzed LdG Bypass<!>CONCLUSION

<p>The genetic material DNA is susceptible to numerous endogenous and exogenous chemicals. The chemically modified DNA, often referred to as DNA adduct or DNA lesion, has important biological implications and has been used as biomarkers in molecular epidemiology studies.1–4 Unrepaired DNA lesions can block DNA replication, cause mutations and cell death, and contribute to pathogenesis.5 To ensure continuous DNA duplication, specialized DNA polymerases (pols) are utilized to incorporate one or a few nucleotides past a noncanonical DNA structure.6,7 This DNA damage tolerance mechanism, known as translesion synthesis (TLS), is conserved from bacteria to humans.8 Notably, TLS is error-prone and plays a vital role in DNA damage-induced mutations and disease etiology.5,9 In humans, pol η, pol ι, pol κ, Rev1, and pol ζ are major TLS pols, of which each has a unique DNA damage bypass and fidelity profile.6,7 For example, human pol κ is known for its capability of performing error-free DNA replication to bypass the N2-dG DNA adduct formed by carcinogen benzo[a]pyrene-7,8-dihydrodiol-9,10-epoxide (BPDE) in vitro and inside cells.10,11</p><p>Madder color, extracted from Rubia tinctrum L. (madder root), has been used as a dyeing agent since ancient times. Madder root has also been used as an herbal medicine for treating kidney and bladder stones12,13 and urinary disorders.14 In Japan and Korea, madder color had been used as a natural food colorant and additive, and such use was banned in 2004 in both countries due to its toxicity in experimental animals.15–17 In rats, madder root exerts carcinogenicity;17,18 however, in humans madder root is classified as Group 3 (i.e., not classifiable as to its carcinogenicity to humans) by the International Agency for Research on Cancer.19 The reason for the low carcinogenic risk of lucidin in humans remains unclear.</p><p>Madder root contains various hydroxyanthraquinones, such as lucidin, alizarin, and their glycoside derivatives such as lucidin-O-primeveroside (LuP).19 LuP and lucidin can be metabolically activated to form quinone methide intermediates that are reactive toward DNA (Figure 1). Lucidin is a potent carcinogen that affects the kidney and liver in rats.17 Lucidin-specific N2-2′-deoxyguanosine (N2-dG) and N6-2′-deoxyadenosine (N6-dA) DNA adducts have been identified in vitro and in LuP-treated rats.20,21 In vivo mutation assays have shown that lucidin induces G to T and A to T transversion and A to G transition mutations in rats.18,22 Therefore, LuP is considered as a major contributor to the madder root-induced carcinogenicity in rats.</p><p>The biological significance and potential association with animal carcinogenicity of lucidin-N2-dG and lucidin-N6-dA adducts remain elusive. Because these DNA modifications can interfere with the Watson–Crick hydrogen bonding, they are likely to be replication-blocking or mutagenic. To better understand the miscoding property of lucidin-derived DNA damage in humans, we developed a method to site-specifically incorporate a structural analogue N2-methyl-(1,3-dimethoxyanthraquinone)- deoxyguanosine (LdG) of lucidin-N2-dG (Figure 1) into DNA oligomers.23 The structural analogue was used due to unsuccessful attempts to synthesize 2-(aminomethyl)-1,3-diacetoxyanthraquinone, an acetyl protected lucidin amine required to assemble the N2-dG nucleoside via Buchwald-Hartwig coupling.23 In vitro DNA replication assays using model Escherichia coli DNA polymerase I Klenow fragment revealed that the LdG lesion inhibited primer extension and induced misincorporations.23 Herein, we unravel the miscoding property of LdG with key human TLS DNA polymerases and describe the mechanism of error-free DNA bypass by pol κ using biochemical and structural approaches. We discover that DNA replication past LdG is mainly error-free; in particular, human pol κ bypasses LdG efficiently and accurately. Three other TLS pols (pol η, pol ι, and Rev1) showed different degrees of attenuation in DNA replication fidelity and efficiency when bypassing LdG. To determine the structural basis of pol κ-catalyzed bypass, we solved two high-resolution X-ray crystal structures of pol κ:LdG-DNA:dNTP, correlating to the insertion (opposite the lesion) and extension steps during lesion bypass. Pol κ accommodates LdG in both structures and maintains productive conformations during TLS. Taken together, this work provides the first mechanistic explanation concerning the low carcinogenic risk of lucidin in humans and sheds light on the versatility of pol κ in replicating past the N2-dG DNA damage.</p><!><p>Unlabeled dNTPs, T4 polynucleotide kinase, and uracil DNA glycosylase (UDG) were from New England Biolabs (Ipswich, MA). [γ-32P]ATP (specific activity of 3000 Ci/mmol) was from PerkinElmer (Waltham, MA). All other commercial chemicals were from Sigma-Aldrich (St. Louis, MO) or Research Products International (Mt Prospect, IL) and were of the highest quality available. Unmodified DNA oligodeoxynucleotides were synthesized and PAGE-purified by Integrated DNA Technologies (Coralville, IA). DNA oligodeoxynucleotides containing a site-specifically modified LdG adduct were synthesized and purified as described previously (see Supporting Information for details).23 The catalytic fragments of human Y-family DNA polymerases pol ι (1–420),24 pol η (1–432),25 pol κ (19–526),26 and REV1 (330–833)27 were purified as described previously.</p><!><p>A 15- or 16-mer primer was 5′ [γ-32P]ATP end-labeled and annealed to a 50-mer unmodified or LdG-bearing oligomer; sequences are shown in Table S1 of the Supporting Information. Primer-extension assays were performed and analyzed as previously described.28 Reaction conditions are specified as follows. Full-length extension assays were performed at 37 °C using 80 nM primer-template DNA, 80 nM DNA polymerase, four dNTPs at their physiological concentrations29 (i.e., 10 μM for dGTP and 40 μM for dATP, dCTP, and dTTP), 4% (v/v) glycerol, 5 mM DTT, 50 mM NaCl, 5 mM MgCl2, and 100 μg/mL bovine serum albumin (BSA) in 50 mM Tris-HCl (pH 7.4 at 25 °C). Single-nucleotide incorporations were conducted under the same condition except that 80–400 nM primer-template DNA, 0.5–20 nM DNA polymerase, and varying concentrations of a single dNTP were used to keep the reaction under steady-state kinetic conditions. Gels were imaged using a phosphor-imaging system (GE Healthcare, Typhoon FLA7000) and analyzed with ImageQuant software. Data were fit to the Michaelis–Menten equation using Prism software (GraphPad, San Diego, CA).</p><!><p>A 15-mer primer 5′-TAATGGCTAACGC(dU)T-3′ was annealed to an unmodified or an LdG-containing 29-mer oligomer at a 1:1.2 molar ratio (template sequence shown in Table S1). The pol κ-catalyzed polymerization was allowed for 3 h in the presence of 2 μM DNA complex and 0.3 μM pol κ in a total volume of 50 μL (other reaction components are the same as those used in full-length primer extension assays) at 37 °C. Reactions were quenched with 10 mM EDTA (final concentration) followed by UDG digestion overnight at 37 °C. Phenol/chloroform extraction was used to purify DNA oligomers, which were subsequently cleaved with hot piperidine (0.25 M piperidine at 90 °C for 1 h). Reaction products were purified with a C18 SampliQ solid-phase extraction cartridge (Agilent Technologies). The fractions containing oligomer were dried under vacuum and suspended in 30 μL of chromatography mobile phase A vide infra.</p><p>NanoLC–MS/MS analysis was performed on a nanoAcquity ultraperformance liquid chromatography system (Waters Corp.) connected to a Finnigan LTQ XL mass spectrometer (Thermo Scientific Corp.). Data were collected under negative ionization mode. The column used was a PicoChip column (75 μm ID, 105 mm bed length and a 15 μm tip, New Objective, Woburn, MA) packed with Reprosil-PUR C18 (3 μm, 120 Å) chromatography media. Chromatography mobile phase A was 400 mM 1,1,1,3,3,3-hexafluoro-2-propanol in water (pH adjusted to 7.0 with triethylamine) and mobile phase B was methanol. The following gradient program was used at a flow rate of 250 nL/min: 0–5 min, maintained at 95% A/5% B (v/v); 5–45 min, linear gradient to 30% B (v/v); 45–55 min, linear gradient to 50% B; 55–60 min, hold at 50% B; 60–105 min, linear gradient to 5% B; 105–130 min, hold at 5% B to re-equilibrium the column. A 5 μL aliquot was injected onto the column. Nanoelectrospray conditions were as follows: ionization voltage 3 kV, capillary temperature 300 °C, capillary voltage −45 V, tube lens voltage −110 V. MS/MS conditions were as follows: normalized collision energy 40%, activation Q 0.250, and activation time 30 ms. Product ion spectra were acquired over the range m/z 300–1800. The most abundant species was used for CID analysis. Theoretical fragmentated ions were calculated using the Mongo Oligo mass calculator version 2.06 (hosted by the State University of New York at Albany).30 The relative yield of extension products was determined based on the integrated peak areas in extracted ion chromatograms, which were set to extract multiple species with different charged states. The sum of peak area ratios of all products/residual primer was set to 100% for each reaction.</p><!><p>Human pol κ protein containing residues 1–526 with 6xHis-tag on the N-terminal was overexpressed and purified as described previously.31 The 22-mer template 5′-ATGG*CTGATCCGCGCGGATCAG-3′ was used to cocrystallize the insertion complex, and 5′-CTATG*TCGATCCGCGGATCGAC-3′ was used to cocrystallize the extension complex (G* denotes LdG). These DNA substrates are self-annealing DNA oligomers; their sequences are designed to avoid hairpin formation and maximize the stability of primer-template DNA. The self annealing DNA was incubated with pol κ in a 1.2:1 ratio in presence of 10 mM MgCl2. The ternary complexes of pol κ with DNA containing LdG and an incoming nucleotide were prepared as previously described.31,32 Briefly, pol κ was mixed with primer-template DNA and 2 mM of nonhydrolyzable nucleotide dCMPNPP (2′-deoxycytidine-5′-[(α,β)-imido] triphosphate, hereafter termed dCTP*) for the insertion complex or dAMPNPP (2′-deoxyadenosine-5′-[(α,β)-imido] triphosphate, hereafter termed dATP*) for the extension complex. The complex mixture was incubated at room temperature for 30 min before setting up crystallization. The ternary complex cocrystals were grown by hanging drop vapor diffusion method at 22 °C with a well solution containing 25–30% PEG400 and 0.1–0.2 M ammonium nitrate. To obtain good quality crystals, drops were streak seeded several times. The pol κ:DNA cocrystals were picked from the drop, cryoprotected with 35% PEG400, 0.2 M ammonium nitrate and 20% ethylene glycol and then flash frozen in liquid nitrogen for data collection. X-ray diffraction data were collected at the beamline 24-ID-E operated by the Northeastern Collaborative Access Team (NE-CAT) at the Advanced Photon Source. Diffraction data were processed using iMolsfm and Scala in the CCP4 suite.33 Structure refinement was performed using RE-FMAC5.34 Model building and inspection was performed in Coot35 and the crystallographic figures were generated using the molecular graphics program PyMOL.36</p><!><p>The lucidin-N2-dG adduct is a major adduct observed in reactions between the reactive metabolites of LuP and DNA in vitro and in the kidneys and livers of LuP-treated rats.20,21 The objectives of this study are (1) to decipher the miscoding pattern of lucidin-N2-dG lesion during human TLS and (2) to gain mechanistic insight into such processes. To this end, we used our recently developed method to synthesize site-specifically modified oligodeoxynucleotides containing an LdG analogue.23 Using steady-state kinetic analysis, we determined the bypass capability, dNTP incorporation efficiency, and DNA replication fidelity of translesion and postlesion syntheses for human TLS pols, η, ι, κ, and Rev1. These pols are Y-family enzymes in DNA damage bypass. The DNA damage bypass occurs when the DNA replication is blocked by a DNA adduct or a noncanonical DNA structure, and involves several polymerase- switching processes to unload a replicative DNA polymerase and recruit one or more TLS pols to rescue a stalled replication fork.37 In addition, translesion synthesis can also be used to fill postreplicative gaps.38–40 To first qualitatively assess the DNA bypass abilities of different TLS DNA polymerases, a reconstituted system was used containing an LdG-harboring primer-template DNA (or an unmodified DNA), a TLS pol, and four dNTPs at physiological concentrations. The extent of primer-extension is an indication of bypass capability of a particular pol. As shown in Figure 2A, relative to an undamaged DNA, LdG had the least impact on the DNA replication activity of pol κ. Pol κ was capable of bypassing LdG with high enzymatic activity, resulting in approximately 71% of the products extended to full-length in a 5 min reaction, which is similar to the yield obtained with the undamaged substrate. This observation is consistent with the known property of pol κ in tolerating N2-dG DNA adducts.10,31,41–44 On the other hand, pol η and pol ι primarily produced one-nucleotide-extension products with the LdG-containing substrate (Figure 2A). Pol η weakly bypassed LdG and generated 16% full-length products in a 30 min reaction. Pol ι barely produced any products other than the products with one-nucleotide extension, in keeping with the low processivity of this enzyme.45 Rev1, a dCTP transferase, is able to perform dNTP incorporation with the LdG-substrate, albeit with lower activity relative to the unmodified DNA. Together, these results demonstrate that the bulky LdG exhibits a minimal impact on the dNTP-incorporation activity of pol κ and a minor impact on that of Rev1, whereas the same lesion compromises the enzyme activities of pol η and pol ι.</p><!><p>To understand the specificity of dNTP incorporation opposite LdG, we determined enzyme kinetic parameters, such as Michaelis constant (Km), turnover number (kcat), catalytic efficiency (kcat/Km), and misincorporation frequency using steady-state kinetic assays (Figure 2B and the Table S2 of the Supporting Information). These parameters allow quantitative assessment of the bypass efficiency and miscoding properties of different enzymes with different DNA and dNTP substrates. As summarized in Figure 2B, the catalytic efficiency of the correct dCTP opposite LdG decreased 10-fold for pol κ, 200-fold for pol η, 70-fold for pol ι, and four-fold for Rev1. Despite the moderate decrease in replication efficiency, pol κ and Rev1 maintained the overall replication fidelity with the LdG-containing DNA substrate. On the contrary, pol η-mediated LdG bypass was error-prone, whereby misincorporations occur more readily (approximately 30–50-fold higher in misincorporation efficiencies) relative to the control DNA. Similarly, the fidelity of DNA replication by pol ι was compromised when replicating past LdG, with dTTP:LdG mispairing occurring at a higher catalytic efficiency than the correct pair, suggesting that pol ι can potentially induce G to A mutations during LdG bypass. The ability of Rev1 to perform dCTP insertion is not significantly perturbed by LdG, judging from the moderate decrease (four-fold) from the catalytic efficiency, which suggests the potential for Rev1 to function redundantly with pol κ for correct dNTP incorporation opposite LdG. Together, these data are consistent with the extension abilities of four polymerases observed in Figure 2A, and demonstrate that dNTP incorporation opposite LdG by major human TLS pols seems to be error-free, especially with pol κ and Rev1.</p><!><p>The complete bypass of bulky DNA lesions is thought to involve multiple steps by a single or multiple DNA polymerases. The first step is to insert one dNTP opposite the DNA lesion (insertion stage), and the second is to further extend the primer by a few additional dNTPs before a replicative DNA polymerase resumes its function (extension stage).46,47 On the basis of this two-step bypass model, we further examined the DNA replication fidelity and efficiency at the extension stage. Steady-state kinetic assays were performed with a 16-mer primer annealed to a dG or LdG-containing substrate. The 16-mer primer was designed to have a dCMP at the 3′-end because dCTP was major nucleotide inserted opposite LdG by four TLS pols (Figure 3 and Table S3 of the Supporting Information). We tested the dNTP-incorporation specificity of three pols (pol κ, pol η, and pol ι) that have shown replication activity past the lesion. As shown in Figure 3, in terms of catalytic efficiency of correct nucleotide (dATP), a moderate decrease was observed with all three pols (six-fold for pol κ, six-fold for pol η, and three-fold for pol ι), indicating that LdG does not significantly alter the enzymatic activity at the extension stage. All three polymerase retained the overall fidelity during postlesion synthesis. Our results demonstrate that LdG does not significantly affect the postlesion DNA synthesis in vitro, suggesting that three pols tested here can potentially serve as "extender" polymerases. Taking into account these kinetic data and the "two-step, two-polymerase" bypass model, it is likely that both pol κ and Rev1 can insert a dC opposite LdG, and that pol κ, pol η, and pol ι can act redundantly during postlesion extension to achieve error-free bypass.</p><!><p>Pol κ is proficient in DNA synthesis at both insertion and extension stages with LdG adduct as observed in gel electrophoretic analysis (Figure 2A); however, pol κ-mediated TLS is known to generate products containing deletion or frameshift mutations,48 which cannot be directly observed in the electrophoretic analysis. To detect the potential formation of such erroneous replication products, we used LC–MS/MS-based oligodeoxynucleotide sequencing to determine the sequence and relative yield of products during pol κ-catalyzed DNA synthesis. This method is complementary to the kinetic analysis and is particularly powerful for detecting insertion and deletion mutations.49,50 As shown in Table 1, translesion synthesis by pol κ past LdG is error-free in the presence of physiological concentrations of nucleotides. The sequence of the oligodeoxynucleotide was obtained from the fragmentation pattern in tandem mass spectrometry analysis (Figures S5 and S6 of the Supporting Information). Major products observed include the cleaved primer (from UDG digestion and hot piperidine treatment) and the polymerization products. The extended products from an LdG-bearing primer-template DNA are essentially error-free except the relative yields of two products vary slightly from those obtained with a native DNA. The error-free extension by pol κ is consistent with its preference for the correct dNTP incorporation observed in the steady-state kinetic analysis (Figures 2 and 3). Together with the steady-state kinetic analysis, these data demonstrate that pol κ is likely to be an important polymerase for error-free DNA replication and extension beyond lucidin-N2-dG DNA lesion, although we cannot exclude the potential involvement of other DNA polymerases. Misincorporations caused by translesion synthesis across LdG by major TLS DNA polymerases tested here are T or G by pol η and T by pol ι (Figure 2B and Table S2 of the Supporting Information). These mispairs can potentially lead to G to A transitions and G to C transversions, which have been observed in LuP-treated rats,16 although they are considered as minor mutations relative to the three major mutations noted earlier.</p><!><p>To gain mechanistic insight into the pol κ-catalyzed LdG bypass, we cocrystallized pol κ with the LdG-adducted DNA substrate and an incoming nucleotide. The first structure (Figure 4A), refined to 2.90 Å resolution, captures pol κ at the insertion stage with LdG pairing with a nonhydrolyzable dCMPNPP (dCTP*). The second structure (Figure 4B), refined to 2.50 Å resolution, represents the extension stage with an incoming dAMPNPP (dATP*) pairing with the 5′-adjacent (of LdG) dTMP. Both ternary complexes are in the same crystal form (P212121) with two molecules in the asymmetric unit (Table 2), which are isomorphous to the crystals of the pol κ ternary complexes with undamaged DNA (PDB ID: 4U6P) or the BPDE-adducted DNA (PDB IDs: 4U7C, 5T14).31,32 In the insertion complex, the incoming dCTP* forms a Watson–Crick pair with the template adducted dG (Figure 4A, bottom panel). The distance between the α phosphate of dCTP* and the O atom of 3′–OH of the primer terminal dG is 4.3 Å, indicating a productive conformation for DNA replication at the insertion stage. The lucidin ring of LdG is positioned toward the 3′-end of the template DNA, stabilized by van der Waals interactions with a polyethylene glycol molecule along with several amino acid residues (Tyr 112, Phe 171, Pro169, and Arg175). These van der Waals interactions together shield the hydrophobic lucidin ring from the aqueous solvent. The base pair at the −1 position is tilted due to the presence of LdG at the 0 position, which compromises the stability of DNA at the active site and potentially accounts for the decreased replication activity relative to an unmodified substrate observed in Figure 2B.</p><p>In the extension complex, the incoming dATP* pairs with the template dT via canonical Watson–Crick base pairing (Figure 4B, bottom panel). A distance of 4.5 Å between the α phosphate of dATP* and the O atom of 3′–OH of the primer terminal dC suggests a productive conformation for DNA replication at the extension stage. The lucidin ring, a glycerol molecule, and Phe 171 form van der Waals contacts. The presence of LdG at the −1 position does not seem to affect the base pairing at −1 and 0 positions, except that the bulky LdG causes a tilted base pair at the −2 position (Figure 5B, tilted template base in blue relative to the unmodified DNA substrate in gray). Compared to the insertion complex, the replicating base pair (at the 0 position) and its base pair beneath (at the −1 position) form better stacking interactions at the extension stage (Figure 5), potentially alleviating the disrupting effect for DNA replication. This latter observation is consistent with a moderate decrease in catalytic efficiency for pol κ at the extension stage (Figure 3).</p><p>The key structural component of pol κ is the open active-site cleft at the minor groove side of substrate DNA (Figure 4), which accommodates the lucidin ring (or the BPDE ring as observed previously31,32). This structural feature of pol κ keeps the DNA substrate in the standard B-form (Figure 5, right panels), judging from the superimposed structures of LdG-DNA and the unmodified, B-form DNA from another pol κ structure.44 Indeed, pol κ maintains the same conformation in a number of solved ternary complexes regardless of different DNA substrates; structural comparisons using root-mean-square deviations are shown in Table S4 of the Supporting Information. The presence of the open active-site cleft is consistent with the ability of pol κ to accommodate an array of DNA lesions.31,41–44 The lucidin ring of LdG is positioned toward the 3′-end of the template DNA, consistent with the orientation predicted using the molecular dynamic simulations. 23 Collectively, these structural data shed light on the mechanism of pol κ-mediated LdG bypass.</p><p>The replication-blocking and mutagenic properties of bulky DNA lesions are attributed in part to the difficulty for DNA polymerases to accommodate these modifications at the enzyme active site.51 For replicative DNA polymerases, the DNA minor groove is enclosed by surrounding amino acid residues at the active site,52–54 which explains why LdG is a potent inhibitor for DNA synthesis by Klenow fragment.23 On the other hand, most Y-family DNA polymerases, such as pol η and pol ι, have a narrow opening at the substrate DNA minor groove, likely to be insufficient to house the lucidin ring. Our observation that LdG drastically reduces the replication efficiency of pol η and pol ι is consistent with this notion (Figure 2B). Modeling BPDE-N2-dG adduct (also a bulky N2-dG adduct) to the active sites of pol η and pol ι results in steric clash, which further confirms the incompetency of these pols in bypass bulky N2-dG adducts.31,37 Another model Y-family DNA polymerase, Sulfolobus solfataricus P2 DNA polymerase IV (Dpo4), excludes the bulky BPDE ring out of the active site due to the lack of space.55 However, such relocation of bulky lesions destabilizes the double helical DNA structure and introduces misinsertions during replication.55</p><!><p>In conclusion, we have obtained the first set of biochemical and structural data to elucidate the miscoding property of lucidin-N2-dG DNA lesion with human enzymes. Our results demonstrate that DNA replication past the LdG is largely error-free with four human TLS pols examined. In addition, two X-ray crystal structures of pol κ provide mechanistic insights into the error-free LdG bypass catalyzed by pol κ. Together, these data support the low carcinogenic risk of madder root in humans. We reason that lucidin-N2-dG DNA lesion can be repaired efficiently or bypassed largely error-free by pol κ in humans. Future biological studies are warranted to further illuminate these points.</p>

PubMed Author Manuscript

Translating Drug-Induced Hibernation to Therapeutic Hypothermia

Therapeutic hypothermia (TH) improves prognosis after cardiac arrest; however, thermoregulatory responses such as shivering complicate cooling. Hibernators exhibit a profound and safe reversible hypothermia without any cardiovascular side effects by lowering the shivering threshold at low ambient temperatures (Ta). Activation of adenosine A1 receptors (A1ARs) in the central nervous system (CNS) induces hibernation in hibernating species and a hibernation-like state in rats, principally by attenuating thermogenesis. Thus, we tested the hypothesis that targeted activation of the central A1AR combined with a lower Ta would provide a means of managing core body temperature (Tb) below 37 \xc2\xb0C for therapeutic purposes. We targeted the A1AR within the CNS by combining systemic delivery of the A1AR agonist 6N-cyclohexyladenosine (CHA) with 8-(p-sulfophenyl) theophylline (8-SPT), a nonspecific adenosine receptor antagonist that does not readily cross the blood\xe2\x80\x93brain barrier. Results show that CHA (1 mg/kg) and 8-SPT (25 mg/kg), administered intraperitoneally every 4 h for 20 h at a Ta of 16 \xc2\xb0C, induce and maintain the Tb between 29 and 31 \xc2\xb0C for 24 h in both na\xc3\xafve rats and rats subjected to asphyxial cardiac arrest for 8 min. Faster and more stable hypothermia was achieved by continuous infusion of CHA delivered subcutaneously via minipumps. Animals subjected to cardiac arrest and cooled by CHA survived better and showed less neuronal cell death than normothermic control animals. Central A1AR activation in combination with a thermal gradient shows promise as a novel and effective pharmacological adjunct for inducing safe and reversible targeted temperature management.

translating_drug-induced_hibernation_to_therapeutic_hypothermia

INTRODUCTION<!>CHA Induces a Decrease in Tb and Heart Rate<!>The CHA-Induced Decrease in Tb Depends on Ta<!>A1AR Antagonist 8-SPT Reverses Bradycardia during Therapeutic Hypothermia without Affecting Tb<!>CHA-Induced Therapeutic Hypothermia following Cardiac Arrest Decreases the Extent of Brain Damage<!>Sustained Hypothermia in Conscious Rats<!>Implantation of Data Loggers and a Programmable Pump<!>Drugs<!>Drug Delivery via Sequential IP Injections<!>Drug Delivery through iPRECIO Pumps<!>Therapeutic Benefit of Sustained Hypothermia in Conscious Rats Subjected to Asphyxial Cardiac Arrest<!>Statistics

<p>Mild therapeutic hypothermia, in which the core body temperature (Tb) is reduced to 32–34 °C for ≥24 h, is becoming the standard of care for cardiac arrest patients.1 However, technical challenges may limit the use of therapeutic hypothermia. Shivering is one of the most problematic issues in targeted temperature management (TTM) and is controlled with pharmacological adjuncts, such as paralytics, narcotics, sedatives, or a combination of these such as meperidine and buspirone.2,3</p><p>Here we study 6N-cyclohexyladenosine (CHA), an A1 adenosine receptor (A1AR) agonist found to induce hibernation,4 as a novel pharmacological adjunct, to facilitate effective techniques for TTM. Tb in hibernating ground squirrels can fall to as low as −3 °C,5 through a process regulated by A1AR signaling within the central nervous system (CNS),4,6 a mechanism common to other types of torpor.7,8 Activation of the CNS A1AR suppresses shivering and nonshivering thermogenesis. 9–12 A1AR agonists protect against ischemic injury and seizure but have not been used clinically because of side effects, principally hypothermia, bradycardia, and hypotension.13 Given the central site of action for A1AR-mediated inhibition of thermogenesis, this study tests the hypotheses that (i) A1AR agonist-induced cooling can be modulated by ambient temperature (Ta), (ii) A1AR-mediated bradycardia can be managed by co-administration of an adenosine receptor antagonist that does not penetrate the blood–brain barrier, and (iii) reversal of bradycardia will occur without interfering with the cooling effects of CHA. Finally, we use a rat model of cardiac arrest to provide a proof of concept that this approach to TTM will improve survival and decrease the extent of brain injury following cardiac arrest.</p><!><p>Overcoming thermoregulation is one of the most problematic issues in targeted temperature management (TTM). Shivering is controlled with pharmacological adjuncts, such as paralytics, narcotics, sedatives, or a combination of these such as meperidine and buspirone,2,3 with little regard to mechanism of action. Here we tested the concept that hypothermia in rats could be induced by mimicking the central A1AR mechanism used by hibernating species to enter hibernation. Drugs were administered systemically because this is the most feasible route of administration in a clinical setting. We report that CHA [1.0 mg/kg, intraperitoneally (IP)] decreased Tb to 33 °C within 1 h (Figure 1A). Data show that repeated injection of CHA and 8-SPT at 4 h intervals (which approximates the half-life of CHA) induces a steady minimum in Tb after the fourth injection (Figure 1A). When cold, animals show neurological deficits consistent with low Tb values (Figure 1B). All animals were rewarmed after the drug was discontinued and they were moved to a Ta of 20 °C. No adverse events were noted after rewarming from an assessment of neurological deficit (Figure 1B). At 24 h, there was evidence of a slight but statistically significant bradycardia in CHA-treated rats (Figure 1C). The heart rate remained significantly elevated 5.5 h after the last CHA and 8-SPT injections, consistent with enhanced thermogenesis following cessation of CHA administration (Figure 1C) [p < 0.0001; two-way analysis of variance (ANOVA); time × treatment; n = 6].</p><!><p>If CHA-induced cooling is due to inhibition of thermogenesis, we predicted that the magnitude of cooling should depend on the thermal gradient, i.e., the difference between Tb and Ta, as seen during hibernation. To address this question, we employed programmable minipumps to deliver CHA continuously for 24 h to rats housed at a Ta of 16 or 25 °C. Results show that continuous administration of CHA at Ta values of 16 and 25 °C decreases Tb (p < 0.0001) and heart rate {p < 0.01; three-way ANOVAs, main effects of group [CD vs CHA (Figures 2 and 3)]} with no effect on hemoglobin oxygen saturation (sO2) (not shown). Moreover, Tb was lowest at a Ta of 16 °C. At a Ta of 16 °C, the mean minimum Tb was 29.3 ± 0.3 °C (Figure 2), and at a Ta of 25 °C, the mean minimum Tb was 35.6 ± 0.1 °C. CHA is thought to facilitate the onset of torpor by suppressing thermogenesis via activation of the A1AR within the CNS.11,14 Core heat then dissipates at rates governed by Ta and thermal conductance. These results are consistent with this mechanism of action because the magnitude of CHA-induced cooling increased with a decrease in Ta. Although systemic administration is more likely to translate to a clinical scenario than ICV administration, direct stimulation of the A1AR on the heart produces profound bradycardia.11,15</p><!><p>Knowing that CHA induces torpor via effects on the brain suggested that an A1AR antagonist with poor blood–brain barrier permeability might reverse bradycardia while sparing the decrease in Tb. CHA lowered Tb when 8-SPT was administered 15 min prior to CHA in the preceding experiment. However, the short half-life of 8-SPT (45 min in rabbit16) suggested that heart rate measured 4 h after the last 8-SPT injection under-reported the full effect of 8-SPT because the drug would have been cleared by this time. Thus, during the 24 h period of continuous CHA administration, 8-SPT (or saline vehicle) was delivered, and heart rate and sO2 were monitored at 10 min intervals for 60 min. At a Ta of 16 °C, in control (CD-treated) animals, both 8-SPT and saline vehicle increased heart rate. The increase in HR in both of these groups was interpreted as an effect of the injection because the effect of 8-SPT did not differ from the effect of saline [p = 0.0434; two-way ANOVA; main effect of time (Figure 2A,B)]. By contrast, in animals treated with CHA, 8-SPT (and not the vehicle saline) produced a 2-fold increase in heart rate (p < 0.0001; two-way ANOVA; time × treatment) with no influence on Tb (Figure 2C,D). As expected because of the short half-life of 8-SPT, the effects on heart rate subsided within 60 min of drug administration. Bradycardia at −1 h is not apparent in Figure 2C most likely because of random error in heart rate measurement. Bradycardia is evident in Figure 1C after the subjects had been cooled for 24 h and in Figure 2D during continuous CHA administration.</p><p>At a Ta of 25 °C, 8-SPT and saline vehicle delivered to control (CD treated) animals had no effect on heart rate or Tb (Figure 3A,B). However, when 8-SPT was delivered to CHA-treated animals, this adenosine receptor antagonist produced a small but significant increase in heart rate (p < 0.0001; two-way ANOVA; time × treatment) with no influence on Tb (Figure 3C,D). Under these conditions, CHA induced only a slight decrease in Tb and no significant decrease in HR, showing that cooling is more effective at producing bradycardia than this dose of CHA. Importantly, the magnitude of the effect of 8-SPT on heart rate was greater in animals housed at a Ta of 16 °C than on those housed at a Ta of 25 °C (Figures 2 and 3). These results show that both hypothermia and CHA contribute to an adenosine receptor-mediated bradycardia, both of which can be managed with 8-SPT without compromising hypothermia.</p><p>The time to target temperature may influence outcome. Thus, control over time to target Tb is desired even though optimal timing remains an area of active research.17 Here we show that pharmacological intervention with adjustments in the temperature differential can be used to manage the rate of cooling and maintenance of hypothermia; animals were cooled faster and to a lower Tb at the colder Ta. In addition, continuous administration of CHA produced a faster decline and a steadier minimum Tb when compared to intermittent injections. A1AR agonists applied too soon after the ischemic event, however, could be detrimental. An increased level of adenosine signaling immediately after traumatic brain injury contributes to respiratory depression and death such that caffeine, a nonselective adenosine receptor antagonist, prevents acute mortality when administered immediately after traumatic brain injury.18 Other studies show that a longer time lag between the return of spontaneous circulation (ROSC) and target Tb is associated with a more favorable neurologic outcome in patients after cardiac arrest when compared to patients with a shorter time lag between ROSC and target Tb.17 Nonetheless, because the rate of CHA-induced cooling and minimum Tb depend on the temperature differential, one is able to adjust the timing and depth of hypothermia through control of the temperature differential, i.e., the heat sink. The temperature differential may be achieved through surface cooling, intravascular cooling,19,20 or Ta. Moreover, the ability to control the rate and depth of cooling with Ta is consistent with other evidence that CHA produces a decrease in Tb by inhibiting thermogenesis.9</p><!><p>Hypothermia is well-known to enhance survival after cardiac arrest.21 To ensure that CHA or 8-SPT did not interfere with the documented benefits of TH, we next sought to test if CHA-induced TH in conscious rats would improve survival and decrease the extent of brain damage following global cerebral ischemia using a model of asphyxial cardiac arrest (CA). With this model, in our hands, asphyxia for 6 and 8 min produces a similar loss of CA1 neurons 8 days after ROSC.16</p><p>Rats subjected to CA for 8 min and treated with 8-SPT and CHA at a Ta of 16 °C survived better than the normothermic control (NC) group. At the onset of treatment, Tb was 33.5 ± 0.1 and 33.5 ± 0.1 °C in the NC and TH groups, respectively. Tb in all three NC rats increased to 36.5–36.8 °C within 15 min of placement at 29 °C and remained between 36.2 and 37.3 °C until death. Only one rat in the NC group survived to 8 days. The two remaining rats died between 13 and 18 h after ROSC (Figure 4). Tb in the three rats treated with TH decreased to 31.0–31.6 °C within 3 h of CHA injection and remained between 31.8 and 29.2 °C for 24 h before the rats were rewarmed; the mean of individual Tb minima was 29.7 ± 0.3 °C. Rats were rewarmed without intervention within 5 h of transfer to a Ta of 20 °C. Transfer occurred 4 h after the last injection of CHA. All three rats treated with TH survived for 8 days despite a CA-induced decrease in MABP similar to that of the NC group (Figure 4). Histopathology showed ischemia-induced cell death of CA1 neurons in the hippocampus of the one normothermic control animal that survived for 8 days (Figure 5). The number of healthy neurons per millimeter of CA1 (mean ± standard error of the mean) was 135.5 ± 4.6 for the TH group (n = 3) and 47.8 for the normothermic control group (n = 1). This compares to the value of 180.8 ± 12.3 for naïve rats (n = 5) reported previously.16 The hypothermic group showed better survival but was also exposed to a lower Ta. A lowered Ta is unlikely to have improved survival unless it also affected core Tb. This is because, without CHA, a lowered Ta stimulates thermogenesis, and the greater metabolic demand would be expected to worsen the outcome.</p><p>In addition to effects on thermogenesis, A1AR agonists are neuroprotective in animal models of cardiac arrest22 and attenuate seizure activity.23 These direct effects within the CNS may have contributed to enhanced survival in this study and contribute to the therapeutic efficacy of these drugs if used as pharmacological adjuncts for TH. Importantly in this study, enhanced survival was observed relative to a NC group where Tb varied between 33.5 and 37.1 °C. In no case did Tb exceed 37.1 °C in the NC group at any of the times measured. Thus, enhanced survival is not due to the absence of hyperthermia, an issue emphasized in a recent clinical study that questions the value of cooling.24</p><p>These results show that A1AR agonist CHA, combined with a decrease in Ta, is an effective adjunct for inducing TH. Decreasing Ta to 16 °C and administering CHA intermittently or continuously produces a rapid and sustained decrease in core Tb with minimal effects on heart rate. Moreover, slight bradycardia is readily reversed with 8-SPT, which has no effect on Tb. We, and others, have shown that activation of CNS adenosine A1 receptors (A1ARs) is necessary and sufficient to induce hibernation and torpor in mice.4,7,8 These results suggest that some of the torpor-inducing effects of CHA may translate to refined methods of facilitating hypothermia for therapeutic purposes. Pharmacological management of shivering and nonshivering thermogenesis is necessary to counteract normal thermoregulatory mechanisms, especially shivering.2 Cold infusions alone do not keep patients cool.25 In comatose cardiac arrest patients, shivering may be controlled by a number of pharmacological adjuncts, including sedatives, narcotics, and paralytics.2 Surface cooling can cause skin lesions when shivering is not adequately controlled.26 More importantly, shivering can prevent the attainment of the target temperature and contribute to adverse effects of TH. Shivering is especially difficult to manage in conscious patients,3 which may limit the benefit of TH in stroke patients.27 A recent study in rats9 shows that CHA acts, in part, within the nucleus of the solitary tract (NTS) to produce effects that closely resemble the suite of thermoregulatory and autonomic nervous system changes that accompany the onset of hibernation.14,28,29 These data suggest that A1AR agonists are promising pharmacological adjuncts for TH for the treatment of stroke, cardiac arrest, and other conditions.30 The data shown here with CHA complement prior work using 5′AMP as an A1AR agonist to induce TH, an approach found to reduce infarct size following middle cerebral artery occlusion in rats.12</p><!><p>Experiments were performed in accordance with the Guide for the Care and Use of Laboratory Animals, 8th Edition (National Research Council, National Academies Press, 2010), and protocols were approved by the Institutional Animal Care and Use Committee of the University of Alaska Fairbanks. Male Sprague-Dawley rats (2–3 months old, 375–400 g; obtained directly or derived from breeders obtained from Simonson Laboratories, Gilroy, CA) were housed in pairs at 20 °C on a 12L:12D photoperiod, fed ad libitum, and allowed at least 2 weeks to acclimate before being used.</p><!><p>Baytril [5 mg/kg, subcutaneously (SC)] began 24 h before surgery and continued BID for 3 days. On the day of surgery, isoflurane anesthesia mixed with medical grade O2, delivered at a rate of 1.5 L/min, was induced at 5% and maintained at ≤2% depending on respiratory frequency. Surgery was followed by 10–14 days of postoperative recovery. iButton data loggers (Maxim Integrated, San Jose, CA) were implanted IP alone or in combination with an iPRECIO pump (Data Sciences International, St. Paul, MN). Pumps were implanted SC at the back of the neck. The outlet tube on the pump was tunneled subcutaneously from the dorsal pocket to the incision site on the neck and secured. Prior to implantation, the pumps were programmed to deliver at a rate of 1.0 µL/h except on the two experimental days, during which the flow rate increased to 30.0 µL/h for 24 h. CHA or vehicle delivery commenced just after residual saline was withdrawn and the 900 µL reservoir filled with CHA or vehicle. Delivery ended after 24 h when residual CHA or vehicle was withdrawn and the reservoir refilled with saline.</p><!><p>N6-Cyclohexyladenosine (CHA) and 8-(p-sulfophenyl)-theophylline hydrate (8-SPT) were purchased from Sigma-Aldrich (St. Louis, MO). Hydroxypropyl-β-cyclodextrin (CD) was purchased from TCI America (Portland, OR).</p><!><p>Animals were instrumented with IP iButton data loggers (Maxim Integrated) programmed to record body temperature (Tb) every 10 min and allowed 14 days postoperative recovery prior to drug testing. 6N-Cyclohexyladenosine (CHA) (A1AR agonist) was dissolved in 0.01 M phosphate buffer (PB); 8-(p-sulfophenyl)theophylline (8-SPT) (non-selective adenosine receptor antagonist) was dissolved in 0.9% saline and filter sterilized on the day of administration. PB for CHA and saline for 8-SPT were administered as vehicle controls where indicated. The day before the experiment, animals in both treatment and control groups were moved to a Ta of 16 °C and remained at this Ta until they were returned to a Ta of 20 °C, 4 h after the last injection. Animals in the treatment group received a total of six injections of CHA (1.0 mg/kg, IP) every 4 h and a total of six injections of 8-SPT (25 mg/kg, IP), administered 15 min prior to each CHA injection. The control group received the same number of injections, with 8-SPT replaced by saline and CHA replaced by PB. Treatment and control conditions were tested in all animals with at least 1 week between experiments using a balanced crossover design such that one-half of the animals received CHA and 8-SPT during the first experiment and the other half received CHA and 8-SPT during the second experiment. Except for moving rats to a Ta of 20 °C 4 h after the last injection, no other means were used to facilitate rewarming. Neurological deficits, heart rate, and sO2 were measured 2 h and immediately before injection, at 24 h, after the rats had been rewarmed and daily for the next 3 days using a pulse oximeter applied to the hind paw (Vet/Ox TM 4402L, Sensor Devices, Waukesha, WI). The total neurological deficit score (NDS) consists of five components: consciousness and respiration, cranial nerve function, motor function, sensory function and coordination (leg/tail movement, cleaning, depth perception, and righting reflex), and motor and sensory function as described previously.31 NDS ranges between 0 (no neurological deficiency, normal function) and 100 (maximal neurological deficiency). Healthy neurons in CA1 were counted as described previously.16</p><!><p>To determine if the effects of CHA on heart rate were due to direct effects of CHA on the heart or the effects of tissue temperature, we employed programmable peristaltic minipumps to deliver CHA continuously for 24 h at Ta values of 25 and 16 °C. During constant delivery of agonist, we also tested the effects of 8-SPT on heart rate and hemoglobin saturation (sO2) at 10 min intervals appropriate for the short half-life of the drug. For these experiments, rats were instrumented with programmable iPRECIO pumps (Data Sciences International). CHA was dissolved in 25% (w/v) hydroxypropyl-β-cyclodextrin (CD) in sterile water, and 8-SPT was dissolved in 0.9% saline. Pumps delivered the same mass of CHA as in the first experiment (6 mg/kg over a 24 h period); however, in this case, CHA was delivered at a constant rate of 30 µL/h. HR and sO2 were monitored using a pulse oximeter every 10 min for 1 h following a single injection of 8-SPT (25 mg/kg, IP) or vehicle as indicated. As before, animals in both treatment and control groups were moved to a Ta of 16 or 25 °C the day before the experiment and remained at this Ta until they were returned to a Ta of 20 °C, 4 h after the end of drug delivery. Drug and vehicle treatments were administered to all animals using a balanced crossover design. At least 1 week separated CHA and vehicle (CD) treatment, and 1 h separated 8-SPT and vehicle (saline) injections.</p><!><p>Rats 68–75 days of age were subjected to asphyxial cardiac arrest for 8 min, and animals that were resuscitated within 120 s and met additional inclusion criteria 60 min after ROSC (Table 1) were randomly allocated to a therapeutic hypothermia (TH) or a normothermic control (NC) group using a computer-generated randomization schedule (http://www.jerrydallal.com). Treatment commenced 70 min after ROSC. Animals assigned to the TH group were moved to 16 °C and CHA and 8-SPT delivered as described above for Drug Delivery via Sequential IP Injections. Animals assigned to the NC group were moved to a neonatal incubator set to 29 °C and vehicles (PB and saline) delivered as described above for the control group (see Drug Delivery via Sequential IP Injections). At the end of 24 h, all rats were moved to and housed at an Ta of 20 °C for 7 days until they were euthanized for tissue collection. Body temperature was monitored prior to each injection throughout the treatment and daily thereafter using SC IPTT-300 transponders (BioMedic Data Systems, Inc., Seaford, DE).</p><!><p>Data are reported as means ± the standard error of the mean unless otherwise indicated. Data were analyzed by two-way ANOVA with repeated measures over time and Tukey post hoc comparisons (SAS, version 9.1.3) or a t test (Excel 2010) where indicated.</p>

PubMed Author Manuscript

Sampling of fluid through skin with magnetohydrodynamics for noninvasive glucose monitoring

Out of 463 million people currently with diabetes, 232 million remain undiagnosed. Diabetes is a threat to human health, which could be mitigated via continuous self-monitoring of glucose. In addition to blood, interstitial fluid is considered to be a representative sample for glucose monitoring, which makes it highly attractive for wearable on-body sensing. However, new technologies are needed for efficient and noninvasive sampling of interstitial fluid through the skin. In this report, we introduce the use of Lorentz force and magnetohydrodynamics to noninvasively extract dermal interstitial fluid. Using porcine skin as an ex-vivo model, we demonstrate that the extraction rate of magnetohydrodynamics is superior to that of reverse iontophoresis. This work seeks to provide a safe, effective, and noninvasive sampling method to unlock the potential of wearable sensors in needle-free continuous glucose monitoring devices that can benefit people living with diabetes.Diabetes causes 4 million deaths and costs 800 billion USD per year, affecting 463 million people globally. Additionally, 374 million people have impaired glucose tolerance (IGT), a high-risk state for diabetes 1 . By 2030, the number of people with diabetes and IGT is projected to increase to 578 million and 454 million, respectively 2 . Despite the high human, social, and economic cost that diabetes imposes on nations, diabetes awareness remains low and nearly 50% of all people with diabetes remain undiagnosed. Wearable devices for continuous selfmonitoring of glucose can play a crucial role in the fight against diabetes by promoting early detection, adequate management, and increased awareness of diabetes. However, widespread adoption of continuous glucose selfmonitoring requires further technological developments that improve user-friendliness and performance 3 .Blood glucose monitoring is essential for the diagnosis and management of diabetes [3][4][5][6][7][8] . The main drawback of existing technology is the non-continuous mode and discomfort related to invasive blood sampling. Therefore, extreme glucose levels may remain unnoticed, which can compromise the health of diabetic persons. Therefore, much research is currently directed towards the development of noninvasive and wearable glucose sensors and monitoring devices [9][10][11] .Lately, there has been significant progress in the development of wearable sensors for continuous monitoring of glucose and other biomarkers in biofluids such as interstitial fluid, sweat, tears, and saliva [12][13][14][15][16][17][18] . Among these biofluids, interstitial fluid (ISF) has a similar glucose concentration to blood plasma, while the concentration of glucose in saliva and sweat is much lower 14,54 . From an analytical point of view, ISF is therefore an attractive sample for noninvasive glucose monitoring. GlucoWatch ® (Cygnus, Inc.) 19 , which was commercialized as a wearable glucose monitoring device, used reverse iontophoresis to extract ISF through skin. However, the device was withdrawn from the market in the late 2000s, indicating the great scientific and technological challenges related to noninvasive glucose monitoring [12][13][14][15][16][17][18] . Despite intensive research and development, the market is still void of a noninvasive glucose monitor 12 .Measuring glucose concentrations for clinical purposes via noninvasive glucose monitoring has proven to be challenging. In recent years, extraction of ISF by reverse iontophoresis has received relatively little attention for glucose monitoring 20,21 compared to glucose measurements in sweat [22][23][24][25][26][27][28] . Sweat is easily accessible but the sweat

sampling_of_fluid_through_skin_with_magnetohydrodynamics_for_noninvasive_glucose_monitoring

<!>Principle and experimental setup for MHD-based extraction of interstitial fluid.<!>Extraction time optimization.<!>Discussion<!>Synthesis of gelatin methacryloyl (GelMA).

<p>rate and chemical composition of sweat varies greatly as a function of physical activity of the person 28 . During physical exercise, sweat production is sufficiently high to allow for microfluidic sampling and monitoring [23][24][25][26] . At rest, a sufficiently high rate of perspiration can be attained by locally activating sweat production through iontophoretic delivery of substances such as pilocarpine, acetylcholine, or methacholine into the skin 28 . Other noninvasive approaches include glucose measurements in tears 29,30 , while microneedles represent a minimally invasive option for measurements in ISF 31,32,33,34 . All commercially available devices for continuous glucose monitoring, i.e. Guardian Connect CGM (Medtronic Inc.), Dexcom G6 (Dexcom Inc.), FreeStyle Libre (Abbott Diabetes Care Inc.), and Eversense CGM System (Senseonics Inc.) rely on invasive sampling of ISF using microneedles or subcutaneous implants 25,26,32,35 . This indicates that, in addition to blood, ISF is a relevant biological fluid for glucose measurements.</p><p>Here, we present a novel method based on magnetohydrodynamics (MHD) for extraction of interstitial fluid. Magnetohydrodynamics is a physical phenomenon where fluid flow is induced by the Lorentz force generated by external magnetic and electric fields. The same physical mechanism has been used in other biomedical applications, e.g. in micropumps [36][37][38] and jet injectors 30,39 . However, to our knowledge MHD is yet to be investigated as a physical mechanism to extract interstitial fluid from the skin.</p><p>Using porcine skin as an ex-vivo model, we show that MHD allows faster extraction compared to reverse iontophoresis. Furthermore, we show that the extracted glucose exhibits a linear relationship to the glucose concentration in a hydrogel in contact with the porcine skin, indicating the feasibility to use MHD as a quantitative tool. MHD increases the total amount of extracted glucose by a factor of two and the active extraction by a factor of 13 when compared to reverse iontophoresis. Hence, MHD reduces the amount of energy applied to the skin required for dermal interstitial fluid sampling and therefore potentially reduces the risk of skin reactions at the extraction site. In this instance, this extraction method is not specific to glucose. It could be applied to extract or deliver other diagnostically or therapeutically valuable molecules through the skin.</p><!><p>To demonstrate the efficacy of MHD, we performed experiments using an extraction cell (Fig. 1a) inspired by the work of Ching et al. 40 . We chose to use porcine skin, which is widely employed as a model of human skin for in vitro studies 41,42 . Especially, stratum corneum from porcine ear has shown significant correlation to its human counterpart in biophysical properties 42 . To model the deeper skin layers, we used gelatin methacryloyl (GelMA) hydrogel because it mimics the collagen-rich extracellular matrix 43 . We used an extraction cell featuring a lower and an upper chamber. We used an extraction cell featuring a lower and an upper chamber. The lower chamber (Fig. S2a) houses a two-layer skin model constructed from GelMA hydrogel, saturated with a solution of glucose of known concentration (Fig. S2b), and porcine ear skin (Fig. S2c). The upper part of the diffusion chamber, depicted in Figs. 1a and S2s, has three functions. Firstly, it creates a tight seal on the skin model, preventing any leakages. Secondly, it protects the hydrogel and skin from drying. And thirdly, provides a robust experimental setup for glucose extraction depicted in Figs. 1a and S2d. The upper part of the diffusion chamber features two cylindrical electrode wells (⌀ = 6 mm; height = 24 mm; separation = 4 mm) with square openings (l = 7 mm) facing the porcine skin. One electrode well was used as cathodic electrode (−) and one as anodic (+). This was determined by the direction of the applied current. Each well was filled with 400 µl of phosphate buffered saline (PBS, 10 mM, pH 7.4) (Fig. S2e) to conduct the electric current from the anodic Ag/AgCl wire through the skin and to the cathodic wire.</p><p>This extraction cell was positioned between two magnets to create a homogenous magnetic field across the cell (Fig. 1b). This magnetic field was combined with electric field to apply MHD extraction through our skin model. The extraction process and forces acting on the interstitial fluid in porcine skin graft are schematically presented in Fig. 1c. The electric field induces electro-osmotic flow from the anode, through the skin, and towards the cathode. The magnetic field orthogonal to the electric field induces a Lorentz force on the interstitial fluid and hence magnetohydrodynamic fluid flow from the deeper skin layers towards the outer skin surface. The extracted fluid accumulates in the electrode wells, and the glucose concentration was determined after each experiment. The experiments were carried out at room temperature and no self heating was detected.</p><p>Comparison of glucose extraction methods. We compared the amount of glucose actively extracted by MHD and reverse iontophoresis against passive diffusion (Fig. 2). The amount of glucose diffusing passively through the skin was measured using the same set-up that was used in the MHD and reverse iontophoresis experiments but without applying electric current or magnetic field. We chose reverse iontophoresis as a reference for active extraction because reverse iontophoresis is the most studied noninvasive method for extraction of glucose and other analytes from the skin 40,[44][45][46][47][48] . The amount of glucose extracted and collected in the cathodic electrode well was quantified using a colorimetric assay. We found that a large part of the measured glucose diffused passively to the electrode wells, and no clear difference between reverse iontophoresis and passive diffusion was observed (Fig. 2a). However, there was a substantial increase in the amount of extracted glucose when MHD was applied. In order to determine the contribution of the active transport of glucose by MHD and reverse iontophoresis, the effect of diffusion was subtracted from the extraction data for both reverse iontophoresis and MHD. Accordingly, the estimated amount of glucose extracted actively with MHD (2.28 ± 0.32 µg) was 13 times higher than the amount of glucose extracted actively with reverse iontophoresis (0.17 ± 0.36 µg), when using 10 min extraction time, 300 µA extraction current, and 300 mT for MHD.</p><p>The distribution of the current density in the MHD and reverse iontophoresis experiments depends on the size and morphology of the electrode areas, the electrode separation, and the impedance profile of the skin. In the case of MHD, the distribution of the electric current further depends on the strength and distribution of the www.nature.com/scientificreports/ magnetic field. At constant electric energy deposition and constant electrode contact area, the distribution of the current density in the skin does not significantly affect the extraction efficiency according to the literature of reverse iontophoresis 45,49 . For MHD, the distribution of the current density in the skin determines the distribution of the Lorentz force according to F = J × B. This may allow maximizing the energy efficiency of the extraction by optimizing the electrode morphology and the distribution of the magnetic field. Since the magnetic field was measured inside the empty cathodic electrode well, the exact distribution of the magnetic field across the GelMA hydrogel and the skin was unknown. However, the extraction cell was positioned at the center and in between the magnets, where the magnetic field is substantially homogenous. Furthermore, skin, water, air, and the materials of the set up (i.e. PMMA/Acrylic) feature similar relative magnetic permeabilities (μ r ≈ 1). Hence, little distortion of the magnetic field induced by the different materials was expected.</p><p>During each extraction, we monitored the voltage between the extraction electrodes to investigate potential discrepancies in electric impedance in the skin samples. Figure 2c shows the measured voltage responses for reverse iontophoresis and MHD with different magnetic fields. For each condition, the voltage stayed relatively constant and the average voltage for each extraction was 0.24 V (Fig. S3). After each experiment, we counted the number of hair follicles inside the area of the electrode well (Fig. 2d) to investigate potential correlation between the number of hair follicles and the amount of extracted glucose. The number of hair follicles was between 5 and 26. However, most skin samples featured 5 to 15 hair follicles. Even though the spread in the number of hair follicles per well area was relatively large, we observed no obvious correlation between the number of hair follicles and extracted glucose (Fig. 2e). Furthermore, we studied whether small differences in skin thickness affect the extraction (Fig. 2f). We chose skin thicknesses between 600 and 700 µm and observed no trends in the amount of glucose extracted with increasing skin thickness. However, the amount of glucose extracted during 10 min depended on the glucose concentration in the GelMA hydrogel (Fig. 2g). This finding indicates that www.nature.com/scientificreports/ MHD could potentially be used in a quantitative manner with a suitable sensor to measure glucose levels in human interstitial fluid.</p><!><p>To further investigate the glucose extraction from our porcine skin/ hydrogel skin model, we varied the extraction time while keeping the extraction current at 300 µA (Fig. 3a,b). As expected, the amount of extracted glucose increased with increasing extraction time for both extraction methods (Fig. 3a). However, after the effect of passive glucose diffusion was subtracted (Fig. 3b), the actively extracted amount of glucose remained relatively constant at 0.31 ± 1.1 µg, 0.54 ± 0.86 µg, and 0.17 ± 0.36 µg when using iontophoresis for 1 min, 5 min, and 10 min, respectively. In contrast, MHD achieved a significant increase in the amount of actively extracted glucose from 0.52 ± 0.88 µg to 2.50 ± 0.37 µg when increasing the extraction time from 1 to 5 min. Interestingly, the amount of actively extracted glucose remained relatively constant when further increasing the extraction time from 5 to 10 min. Hence, most of the glucose extracted with active methods (reverse iontophoresis and MHD) occurred during the first 5 min. This may be related to time-dependent changes in concentration gradients causing back-diffusion of glucose towards the hydrogel during prolonged extraction (Fig. 3b). Similar effects have been observed by others with urea, potassium, and cysteine as analytes using reverse iontophoresis 44,50,51 , however, no comprehensive explanation of this effect currently exists. www.nature.com/scientificreports/ Skin damage assessment. Potential damage to the skin induced by either MHD or reverse iontophoresis was investigated with trans-epidermal water loss measurements (TEWL) 52,53 and visual inspection performed before and after the extraction (Fig. 4a). Average TEWL measurements after the extraction were slightly higher than before the extraction for MHD, reserve iontophoresis, and passive diffusion (Fig. 4b). Reverse iontophoresis showed higher average ΔTEWL when compared to MHD and passive diffusion. However, there was no statistical difference between the extraction and reference experiments, which indicates that the increase was caused by the highly hydrated conditions in the extraction cell. This implies that neither of the extraction methods affect the skin barrier function.</p><!><p>We presented a novel extraction method for interstitial fluid sampling that relies on MHD and the Lorentz force. Utilizing glucose as the analyte and a well-established substitute for human skin, the porcine ex vivo skin model, we compared the efficiency of MHD and reverse iontophoresis to passive diffusion under well-defined experimental conditions. The results showed that the efficiency of the MHD extraction method is superior to that of reverse iontophoresis. Using a 300 mT magnetic field we achieved a 13-fold increase in active glucose extraction. Furthermore, the amount of extracted glucose was proportional to the glucose concentration in the sample, which is important when considering potential applications in glucose sensing where the extracted amount of glucose in ISF is expected to correlate with the concentration of glucose in blood. Transepidermal water loss measurements before and after extraction showed no significant differences, implying that the extraction does not damage the skin permeability barrier. In this paper, we reported relatively high values of diffusion when compared to the active extraction. This indicates that the porcine skin grafts were probably more preamble than intact human skin. Thus, this study encourages more research into ex vivo models which could better represent the effect of microcirculation and barrier function of living skin.</p><p>The proposed technology can be optimized by exploring different electrode shapes, sizes, and distances, in addition to extraction current waveforms, to increase the glucose extraction rate obtained with MHD. Moreover, the magnet arrangements can be decreased in size if the magnets are moved closer to the skin surface. For example, when using a Neodymium magnet as small as 5 mm × 5 mm, a magnetic field of 265 mT is present at the surface of the magnet (Fig. S7). Thus, the MHD technology has the potential to be applied in wearable devices for noninvasive glucose monitoring. These small magnets could be installed into a portable device such as a wrist band or sport watch with glucose sensitive electrodes.</p><p>Since the MHD technology is non-selective, these findings imply that the MHD method could potentially be used to extract other analytes present in ISF (e.g. Na + , K + , and lactate) which could be of interest in biomarker sensing. Furthermore, the extraction rate achieved with MHD could enable detection of sparse analytes, such as cortisol, that exist in interstitial fluid at concentrations below 1 µM 54 . Consequently, the presented MHD extraction method could be valuable to the development of wearable chemical sensors and biosensors utilizing interstitial fluid as the sample. Furthermore, by switching either the poles of the magnet or the direction of the www.nature.com/scientificreports/ current, the direction of the magnetohydrodynamic force can be reversed. Thus, this technology could potentially also be used to deliver molecules through the skin and into the human body.</p><p>In conclusion, we anticipate that the results presented in this work may encourage researchers to revisit the utilization of ISF as a sample for noninvasive on-body chemical sensing.</p><!><p>GelMA was synthetized by functionalising gelatin from bovine skin (type B, Sigma Aldrich) with methacrylic anhydride (MEA, Sigma Aldrich) using a protocol adapted from ref. 55 . Briefly, 10% gelatin solution was prepared by dissolving gelatin powder into PBS (pH 7.4) at 50 °C. While vigorously stirring, 0.6 g of MEA per 1 g of gelatin was added dropwise into the gelatin solution and the reaction was stirred at 50 °C for 3 h. Unreacted MEA was removed by 5 min centrifugation at 2180 G followed by 5-day dialysis (10 kDa cut off) of the supernatant. Finally, GelMA was lyophilized for one week and stored at www.nature.com/scientificreports/ − 20 °C. The degree of functionalization (DOF) in GelMA synthesis was measured using an OPA (o-phthalaldehyde) assay 55 and 86% DOF was achieved.</p><p>Preparation of GelMA hydrogel. The protocol for GelMA hydrogel preparation was adapted from ref. 56 .</p><p>GelMA hydrogel was prepared from 10% GelMA solution (in PBS, pH 7.4) with 0.15% of lithium phenyl-2,4,6trimethylbenzoylphosphinate photoinitiatior (LAP). The mixture was placed in a hand casting mold (Mini-PROTEAN ® Tetra Handcast System, Biorad) with a 1.5 mm spacer and was polymerized with UV-C light (Carbon XM-120V, Green UV) for 5 min. The crosslinked GelMA was placed into PBS to remove residues of LAP.</p><p>For doping the gel with glucose, the GelMA powder was dissolved into PBS with glucose and the hydrogels were made as previously described. Crosslinked GelMA hydrogels were then stored in PBS with glucose until used.</p><p>Preparation of porcine skin. We purchased the porcine ears from a local eco farm (Kiven säästöpossu, Karkkila, Finland). We collected the ears from the slaughterhouse, where the ears remained at room temperature for 2 h after sacrificing the animals. Afterwards, the ears were transported on ice for 1 h. Then, the ears were washed with cold running water. From the posterior surface of each ear, 2 to 3 round samples of 4 cm diameter and between 600 and 700 µm thickness (Fig. S4) were dissected using a dermatome (Nouvag, Switzerland). Individual skin samples were wrapped in parafilm and stored at -20 °C and were used within the next two weeks.</p><p>Trans-epidermal water loss was measured before freezing and after defrosting to track damage of the skin during skin sample storage (Fig. S5). The measurements were done with a Tewameter ® (TM 300, Courage + Khazaka electronic GmbH, Germany) using the software (MPA WL version 1.3.1) according to the manufacturer's instructions. An example of a Tewameter ® reading is shown in Fig. S5. After defrosting, the skin samples were equilibrated at room temperature (22 ± 1 °C) in PBS for 1.5 h with changing the buffer every 30 min. Extra buffer on the skin surface was removed by absorbing it with paper tissue before each experiment.</p><p>Silver wire chlorination. Silver wires (diameter = 1.0 mm, Sigma Aldrich) were cleaned with 120-grit sandpaper and rinsed with water and 2-propanol. The clean wires were placed in a 1 M KCl bath and 300 µA current was passed through them for 3 h. After chlorination, the wires were rinsed with deionized water, dried gently with pressurized air, and stored protected from light.</p><p>Experiments with extraction cell. All measurements and experiments were done at 22 ± 1 °C. A twolayer hydrogel skin model was constructed in an extraction chamber (Fig. S1) by carefully placing a piece of GelMA hydrogel 1.5 mm thick and 3 cm in diameter with a specific glucose concentration into the extraction chamber. A porcine skin sample (thickness: 600 to 700 µm; ø = 4 cm) was placed on top of the GelMA (care was taken to avoid trapping air between the layers). Then, the diffusion chamber was tightly sealed with plastic screws. Subsequently, the electrode wells were filled with 400 µl of PBS (pH 7.4, Sigma Aldrich) and the Ag/ AgCl electrode wires were immersed into the buffer solution avoiding contact with the walls of the well. A current of 100, 200 or 300 µA was passed through the electrodes using a current source (Model 6220, Keithley Instruments). For MHD extraction, the set up was placed at the center and in between two or more magnets (neodymium magnets, Goliath, Supermagnete) which established a homogeneous magnetic field at the extraction site. The set of magnets were positioned carefully so that the magnetic field was perpendicular to the current field thereby causing the Lorenz force predominantly orthogonal to and pointing away from the skin surface. The two highest magnetic fields were obtained by placing the extraction chamber in the middle of an ensemble of four magnets (Fig. S1). The magnetic field strength was measured inside an empty electrode well with an AC/ DC magnetic meter (PCE-MFM 3000). Trans-epidermal water loss was measured before and after extraction to track potential damage of the skin during extraction experiments. The water-loss measurements were conducted with a Tewameter ® (TM 300, Courage + Khazaka electronic GmbH, Germany).</p><p>Determination of glucose concentration. The glucose concentration was measured using a protocol adapted from ref. 57 . Accordingly, 125 µl samples were taken from the electrode wells and mixed with 250 µl of reaction mixture containing 12.8 U/ml of glucose oxidase, 2.6 U/ml of horse radish peroxidase, and 0.13 mg/ml of o-dianisidine in PBS (pH 7.4). This mixture was incubated at 37 °C for exactly 30 min. Immediately thereafter, 250 µl of 12% sulfuric acid was added to the mixture. The light absorbance of the reaction product, oxidized o-dianisidine, was measured at 540 nm using a UV-VIS spectrophotometer (UV-1900, Shimadzu) and was compared to the standard curve (Fig. S6).</p><p>Received: 8 November 2020; Accepted: 18 March 2021</p>

Scientific Reports - Nature

Copper-catalyzed selective hydroamination reactions of alkynes

The development of selective reactions that utilize easily available and abundant precursors for the efficient synthesis of amines is a longstanding goal of chemical research. Despite the centrality of amines in a number of important research areas, including medicinal chemistry, total synthesis and materials science, a general, selective, and step-efficient synthesis of amines is still needed. In this work we describe a set of mild catalytic conditions utilizing a single copper-based catalyst that enables the direct preparation of three distinct and important amine classes (enamines, \xce\xb1-chiral branched alkylamines, and linear alkylamines) from readily available alkyne starting materials with high levels of chemo-, regio-, and stereoselectivity. This methodology was applied to the asymmetric synthesis of rivastigmine and the formal synthesis of several other pharmaceutical agents, including duloxetine, atomoxetine, fluoxetine, and tolterodine.

copper-catalyzed_selective_hydroamination_reactions_of_alkynes

<!>Direct hydroamination: development and scope<!>Reductive hydroamination: development and scope<!>Application to drug synthesis<!>Mechanistic discussion<!>Conclusion<!>Methods

<p>Complex organic molecules play a crucial role in the study and treatment of disease. The extent to which they can be utilized in these endeavors depends on the efficient and selective chemical methods for their construction1. Amines are widely represented in biologically active natural products and medicines2 (a small selection of which are shown in Fig. 1a). Consequently, the selective assembly of amines from readily available precursors is a prominent objective in chemical research3. There are a number of powerful methods that address this challenge including metal-catalyzed cross-coupling4,5, nucleophilic addition to imines6, C–H nitrogen insertion7, and enzymatic methods8,9. However, the direct production of amines from simple olefins or alkynes represents a highly attractive alternative, given the abundance and accessibility of these starting materials. For this reason, the addition of nitrogen and hydrogen across carbon–carbon multiple bonds (hydroamination) has long been pursued as a means to access amines10-12. While much progress has been made, a generally effective strategy to achieve chemo-, regio-, and enantioselective hydroamination of simple alkenes or alkynes remains elusive.</p><p>We became interested in developing hydroamination reactions of alkynes as a convenient and powerful means of accessing aminated products (Fig. 1b). Reactions that employ alkynes as starting materials are synthetically versatile, since alkynes can be prepared by a variety of strategies, including Sonogashira coupling13, nucleophilic addition of metal acetylides14, and homologation of carbonyl groups15. In addition, one or both π-bonds of alkynes may be utilized, further increasing their flexibility as starting materials. For these reasons, the hydrofunctionalization of alkynes has recently become an active area of research16-22. We23 as well as Hirano and Miura24 recently detailed catalyst systems for the asymmetric hydroamination of styrenes that operate through addition of a catalytic copper hydride species25-32 to a carbon–carbon double bond followed by carbon-nitrogen bond formation using an electrophilic nitrogen source33-36. We surmised that we could apply this approach to the selective functionalization of alkynes wherein alkyne hydrocupration would give a stereodefined vinylcopper intermediate. We anticipated that, in analogy to our previous work, direct interception of this intermediate would potentially enable the stereoselective formation of enamines (Fig. 1b, A). Enamines are versatile intermediates in organic synthesis and while catalytic methods have been developed for their synthesis by alkyne hydroamination, control of the regio- and stereochemistry of enamine formation is nontrivial16.</p><p>In addition to enamine synthesis, we speculated on the possibility that conditions could be developed to convert alkynes to enantioenriched α-branched alkylamines and/or linear alkylamines in one synthetic operation (Fig. 1b, B). Such cascade processes are highly desirable in organic synthesis, since potentially difficult workup and isolation steps can be avoided, and the generation of chemical waste is minimized37. In particular, we envisioned a scenario in which the starting alkyne is initially reduced to a transiently-formed alkene, which would then undergo hydroamination to afford the alkylamine product. If successful, this approach would be particularly attractive due to the ease and low cost of the Sonogashira process for the preparation of alkyne starting materials relative to the cross coupling of stereodefined vinylmetal reagents or other routes used to access geometrically pure alkenes for hydroamination. We were aware that, mechanistically, the vinyl- and alkylcopper intermediates in the proposed process are required to react in a highly chemoselective manner (Fig. 1c). Specifically, the vinylcopper species formed upon hydrocupration of the alkyne would need to be selectively intercepted by the proton source in the presence of the aminating reagent to furnish the intermediate alkene while the alkylcopper species formed upon hydrocupration of this alkene would need to selectively engage the electrophilic nitrogen source in the presence of a proton donor to ultimately furnish the desired alkylamine product. While both steps (i.e., alkyne semireduction38-40 and alkene hydroamination23) are well precedented, the ability to achieve the desired selectivity in one pot via a cascade sequence has never been demonstrated41,42. Herein, we report mild and scalable conditions for the highly chemo-, regio-, and stereoselective synthesis of enamines ("direct hydroamination") and alkylamines ("reductive hydroamination") products from alkynes, employing a single copper catalyst system.</p><!><p>To assess the feasibility of the outlined alkyne hydroamination (Fig. 1b, A), we treated 1,2-diphenylacetylene (1a) with N,N-dibenzyl-O-benzoylhydroxylamine (2a, 1.2 equiv.) and an excess of diethoxymethylsilane (3) in the presence of 2 mol % copper acetate and a range of phosphine ligands. A number of ligands could be used to perform the direct hydroamination reaction, and the resulting enamine 4a was efficiently produced as a single geometric isomer, as determined by 1H NMR analysis (Table 1, entries 1–4). Although copper catalysts based on 2,2′-bis(diphenylphosphino)-1,1′-binaphthalene (BINAP, L1), 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene (Xantphos, L2) or 4,4′-bi-1,3-benzodioxole-5,5′-diylbis(diphenylphosphane) (SEGPHOS, L3) were effective in this context, the catalyst based on 5,5′-bis[di(3,5-di-tert-butyl-4-methoxyphenyl)phosphino]-4,4′-bi-1,3-benzodioxole (DTBM-SEGPHOS, L4) was found to be the most efficient and generally applicable. We then evaluated the substrate scope of this reaction and, as shown in entries 5–9, a diverse range of aryl-substituted internal alkyne substrates could be converted to the corresponding (E)-enamines 4 with complete stereoselectivity (4b–4e; 80–99%). Notably, sterically hindered amines, which were problematic substrates for previously reported hydroamination reactions of alkynes43, could be successfully transformed using the current conditions (4b and 4d). More importantly, direct hydroamination of unsymmetrical internal alkynes occurred with excellent regioselectivity (4c–4e; >19:1). In addition, we found that a 1,2-dialkylacetylene was left intact under these conditions (4e) and pharmaceutically important heterocycles, including morpholine (4c), thiophene (4d), piperidine (4e), and pyrimidine (4e) were well-tolerated. While the direct hydroamination of terminal alkynes to construct monosubstituted enamines was not successful, the current method represents a rare example of a highly regio- and stereoselective hydroamination of internal alkynes for the construction of dialkyl enamines43.</p><!><p>As previously described, we were hopeful that the addition of a protic additive could divert this reaction from direct alkyne hydroamination to the outlined reductive hydroamination by selective protonation of the formed vinylcopper intermediate (Fig. 1c). Indeed, inclusion of methanol as an additive under the reaction conditions in Table 1 resulted in formation of the desired reductive hydroamination product 5a in moderate yield, along with a significant amount of enamine 4a (18%) and stilbene (17%) as side products (Table 2, entry 1). Fortunately, an evaluation of alcohol additives revealed that ethanol was a suitable proton source, which minimized the formation of these side products to afford benzylamine 5a in excellent yield and high enantioselectivity (entry 2, 92% yield, 89% e.e.). Interestingly, in contrast to the direct alkyne hydroamination protocol for enamine formation, L4 was uniquely able to perform the reductive hydroamination cascade reaction: reaction utilizing copper catalysts based on L1, L2 or L3 provided only enamine 4a in high yields even in the presence of ethanol (entries 4–6). We attribute the success of the catalyst system based on L4 to the ability of the CuH species to hydrocuprate alkynes and alkenes more rapidly. In contrast, the hydrocupration of alkynes occurred less efficiently when L1-L3 were employed, resulting in the consumption of the alcohol additive by the CuH before alkyne hydrocupration could take place. Therefore, only the enamine product was obtained in these cases. In addition, we found that arylacetylenes could also undergo reductive hydroamination, although in the case of these substrates, isopropanol was a superior protic additive (entry 8).</p><p>Under the optimized set of reaction conditions, a range of chiral benzylamine derivatives could be prepared in moderate to high yield (61–85%) with very high levels of enantioselectivity (≥97% e.e., Table 3). These mild catalytic conditions tolerated a range of common functional groups including ethers (5c, 5h), alcohols (5i), aryl halides (5e, 5f), pyridines (5d), indoles (5g), acetals (5j), and ketals (5m, 5n). Moreover, a reaction conducted on 10 mmol scale proceeded efficiently in the presence of 1 mol % catalyst, furnishing the product in undiminished yield and enantioselectivity (5j). The applicability of new synthetic methods to the late-stage modification of complex natural products is a highly desirable feature, as analogs of bioactive molecules can be prepared without the need for de novo synthesis. Accordingly, readily available alkynes derived from the natural products δ-tocopherol and estrone were subjected to asymmetric reductive hydroamination conditions to afford aminated products with good yields and excellent, catalyst-controlled diastereoselectivities (d.r.: 99:1, 5k–5n). It is noteworthy that in all reductive hydroamination reactions employing aryl-substituted alkynes, the amination products were delivered with exclusive Markovnikov regioselectivity, with C–N bond formation occurring adjacent to the aryl group.</p><p>In addition to aryl-substituted alkynes, we found that terminal aliphatic alkynes readily participate in catalytic reductive hydroamination to deliver alkylamines (Table 4). In contrast to aryl-substituted alkynes, anti-Markovnikov regioselectivity was observed when simple alkylacetylene substrates were used, giving rise to linear tertiary amines in high yields (71–88% yield). We note that it was crucial to use a slight excess of isopropanol compared to the alkylacetylene substrate in the case of terminal alkyne substrates, probably due to deactivation of the catalyst through formation of a copper acetylide species when the amount of isopropanol was insufficient40. It is noteworthy that this methodology could be applied to a substrate bearing an unprotected secondary amine to provide 1,3-diamine 6a in high yield. Furthermore, alkynol silyl ethers were suitable substrates for the current method. Upon reductive hydroamination and silyl deprotection, 1,3-amino alcohol products were prepared in good yields (6f, 6g). An enantioenriched 1,3-amino alcohol could be generated from the optically active alkynol silyl ether (98% e.e.) without erosion of enantiomeric excess (6g, 98% e.e.). The use of the current reductive hydroamination methodology, in conjunction with well-developed methods for the asymmetric synthesis of alkynols14, represents an attractive approach for the synthesis of these biologically relevant compounds.</p><!><p>The functionalized amines obtained by the cascade hydroamination developed in this study have broad utility. To illustrate this point, 1,3-amino alcohol 6f can be readily transformed to the antidepressant drug duloxetine following the literature procedure44 (Fig. 2a). Moreover, 1,3-amino alcohol 6g is a key intermediate for the synthesis of the attention deficit hyperactivity disorder (ADHD) drug atomoxetine45, as well as the antidepressant drug fluoxetine46 (Fig. 2b). Furthermore, diisopropylamine 8 could be prepared by reductive hydroamination of readily synthesized alkyne 7 (Fig. 2c). Tolterodine, a drug used for symptomatic treatment of urinary incontinence, could be prepared by demethylation of 847. Our method was also applied towards a new two-step synthesis of rivastigmine, a drug that is used to treat dementia (Fig. 2d). Condensation of commercially available reagents gave carbamate 9, which upon asymmetric reductive hydroamination furnished the target in enantiopure form and good chemical yield. These syntheses exemplify the potential utility of the current hydroamination method for the rapid and efficient construction of medicinally relevant molecules.</p><!><p>The proposed mechanisms for the formation of enamines and alkylamines (shown in Fig. 3a) both commence with syn-selective Cu–H addition to the alkyne substrate to give vinylcopper species 11. In the absence of a proton source (alcohol), direct interception by electrophilic amine 2 (likely via oxidative addition/reductive elimination) would produce the (E)-enamine product 4 and copper benzoate complex 13 that, upon transmetalation with hydrosilane, would regenerate the active CuH species 10. On the other hand, in the presence of alcohol, direct protonation of the vinylcopper intermediate 11 could afford the cis disubstituted alkene 14. A hydroamination similar to that we have previously reported via alkylcopper species 15 and 16 would deliver the alkylamine 5. As previously detailed, these amination protocols both provide products in excellent regioselectivity; aryl-substituted internal alkynes (and alkenes) proceed through a Markovnikov selective hydrocupration event and terminal aliphatic alkynes (and alkenes) undergo anti-Markovinikov hydrometalation. This regioselectivity is in accord with the results that were previously reported for copper-catalyzed alkyne semireduction39,40 and hydroamination23. We rationalize the Markovnikov regioselectivity observed for aryl alkynes as arising from the electronic stabilization of the vinyl- or alkylcopper species by an adjacent aryl group. In contrast, we attribute the anti-Markovnikov selectivity observed in the case of terminal aliphatic alkynes to steric effects (i.e., formation of the less substituted vinyl- or alkylcopper species).</p><p>We conducted mechanistic experiments to gain a better understanding of the factors that result in the high selectivities observed in our system. Subjecting enamine 4a to the conditions developed for either alkyne hydroamination or reductive hydroamination resulted in no observed reaction (Fig. 3b). The inertness of the enamine under these conditions accounts for the exclusive formation of monoamination product in the case of alkyne hydroamination. In addition, these experiments suggest that alkyne hydroamination followed by enamine reduction is not occurring in the case of reductive hydroamination. Furthermore, we subjected cis-stilbene (18) to the hydroamination conditions in the presence of 1.5 equiv ethanol (Fig. 3c). Although a small amount of 1,2-diphenylethane (19, 3% yield) was formed, presumably as a result of protonation of the alkylcopper intermediate48, hydroamination adduct 5a was generated as the predominant product (97% yield). This result suggests that amination of the alkylcopper species 15 occurs selectively in the presence of a proton source. Combined, the results of these experiments are in agreement with our original hypothesis that vinylcopper species 11 and alkylcopper species 15 undergo selective protonation and amination respectively, thereby allowing the desired cascade reaction to proceed as designed.</p><!><p>In conclusion, we have developed catalytic conditions that allow for the controlled construction of enamines or alkylamines from alkynes and electrophilic amine sources. The products from these complementary systems were obtained with uniformly high levels of regio- and stereocontrol. Both catalytic processes operate through the formation of a vinylcopper intermediate, the product being determined by the presence or absence of an alcohol additive. The development of a protocol for the direct conversion of alkynes to alkylamines is especially notable, given the ease of access to requisite substrates and the demonstrable applicability of this method to the rapid synthesis of a number of pharmaceutical agents. Beyond the broad utility of this new protocol, we anticipate that this cascade strategy will motivate the design of other cascade processes for the more efficient synthesis of valuable targets.</p><!><p>A typical procedure for the copper-catalyzed reductive hydroamination of alkynes 1 is as follows (all reactions were set up on the benchtop using standard Schlenk technique). An oven-dried screw-top reaction tube equipped with a magnetic stir bar was charged with Cu(OAc)2 (3.6 mg, 0.02 mmol, 2 mol %) and (R)-L4 (26 mg, 0.022 mmol, 2.2 mol %). The reaction tube was sealed with a screw-cap septum, then evacuated and backfilled with argon (this process was repeated a total of three times). Anhydrous THF (0.5 mL) and hydrosilane 3 (0.64 mL, 4.0 mmol, 4.0 equiv.) were added sequentially via syringe. The resulting mixture was stirred at room temperature (rt) for 15 min and the color of the mixture changed from blue to orange. A second oven-dried screw-top reaction tube equipped with a stir bar was charged with alkyne substrate 1a (178 mg, 1.0 mmol, 1.0 equiv.) and hydroxylamine ester 2a (381 mg, 1.2 mmol, 1.2 equiv.). The reaction tube was sealed with a screw-cap septum, and then evacuated and backfilled with argon (this process was repeated a total of three times). Anhydrous THF (0.5 mL) and EtOH (88 μL, 1.5 mmol, 1.5 equiv.) were added, followed by dropwise addition of the catalyst solution from the first vial to the stirred reaction mixture at rt. The reaction mixture was then heated at 40 °C for 18 h. After cooling to rt, the reaction was quenched by addition of EtOAc and a saturated aqueous solution of Na2CO3. The phases were separated, the organic phase was concentrated, and product 5a was purified by flash column chromatography. The enantiomeric excesses (e.e.) of the products were determined by HPLC analysis using chiral stationary phases. All new compounds were fully characterized (see Supplementary Information).</p>

PubMed Author Manuscript

Molecular dynamics studies unravel role of conserved residues responsible for movement of ions into active site of DHBPS

3,4-dihydroxy-2-butanone-4-phosphate synthase (DHBPS) catalyzes the conversion of D-ribulose 5-phosphate (Ru5P) to L-3,4-dihydroxy-2-butanone-4-phosphate in the presence of Mg 2+ . Although crystal structures of DHBPS in complex with Ru5P and non-catalytic metal ions have been reported, structure with Ru5P along with Mg 2+ is still elusive. Therefore, mechanistic role played by Mg 2+ in the structure of DHBPS is poorly understood. In this study, molecular dynamics simulations of DHBPS-Ru5P complex along with Mg 2+ have shown entry of Mg 2+ from bulk solvent into active site. Presence of Mg 2+ in active site has constrained conformations of Ru5P and has reduced flexibility of loop-2. Formation of hydrogen bonds among Thr-108 and residues -Gly-109, Val-110, Ser-111, and Asp-114 are found to be critical for entry of Mg 2+ into active site. Subsequent in silico mutations of residues, Thr-108 and Asp-114 have substantiated the importance of these interactions. Loop-4 of one monomer is being proposed to act as a "lid" covering the active site of other monomer. Further, the conserved nature of residues taking part in the transfer of Mg 2+ suggests the same mechanism being present in DHBPS of other microorganisms. Thus, this study provides insights into the functioning of DHBPS that can be used for the designing of inhibitors.In bacteria, riboflavin (Vitamin B 2 ) is produced via riboflavin biosynthesis pathway. Riboflavin is the universal precursor of flavocoenzymes -flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD). It has been estimated that up to 3.5% of bacterial proteins use flavocoenzymes 1 . These flavocoenzymes are involved in various biochemical reactions such as oxidation, epoxidation, sulfoxidation, amine oxidation, selenide oxidation, Baeyer-Villiger oxidation, phosphite ester oxidation, hydroxylation, halogenation, and dehydrogenation, which are part of different metabolic pathways such as the citric acid cycle, β -oxidation, degradation of amino acids, and detoxification of a vast spectrum of xenobiotics 2 . FMN and FAD are also involved in the biosynthesis of steroids, thyroxin, coenzyme A, coenzyme Q, heme, and pyridoxal 5′ -phosphate 3,4 . Moreover, these are essential in numerous physiological processes such as light sensing 5 , light driven DNA repair 6 , bacterial bioluminescence 7,8 , and regulation of biological clock 9 . These roles of FMN and FAD make riboflavin as an essential element for microorganisms. The riboflavin biosynthesis pathway is present in plants and many pathogens. Interestingly, it is absent in animals, and they obtain riboflavin from the nutritional sources. This makes the riboflavin biosynthesis pathway a rich source of targets to design selective anti-infective agents. Additionally, it provides an alternate source of novel targets urgently needed to tackle the problem of antibiotic resistance 10,11 .Seven enzymes take part in the bacterial riboflavin biosynthesis pathway. These are GTP cyclohydrolase II, pyrimidine deaminase, pyrimidine reductase, putative pyrimidine phosphatase, 3,4-dihydroxy-2-butanone-4-phosphate synthase (DHBPS), lumazine synthase, and riboflavin synthase. DHBPS converts the substrate D-ribulose 5-phosphate (Ru5P) into L-3,4-dihydroxy-2-butanone-4-phosphate (DHBP) and formate using Mg 2+ as co-factor 4 . According to the proposed reaction mechanism, Mg 2+ makes coordinate bonds with Ru5P, water, and enzyme residues, initiates reaction by proton abstraction, advances it through enolization, protonation, dehydration, and skeletal rearrangement to release the products, DHBP and formate [12][13][14][15] . The structures of DHBPS in its apo, Ru5P or Ru5P-ion bound form have been solved for many organisms viz. Escherichia coli 16 ,

molecular_dynamics_studies_unravel_role_of_conserved_residues_responsible_for_movement_of_ions_into_

<!>Results<!>Interactions of DHBPS with Mg<!>Entry of Mg<!>Molecular dynamics of mutant DHBPS-Ru5P complexes. MD simulations of DHBPS-Ru5P com-<!>Asp114Leu mutant DHBPS-Ru5P complex.<!>Discussion<!>Material and Methods<!>Mutation of residues.

<p>Magnaporthe grisea 17,18 , Methanocaldococcus jannaschii 14,19 , Yersinia pestis, Candida albicans 20,21 , Streptococcus pneumoniae 22 , Mycobacterium tuberculosis 23,24 , Salmonella typhimurium 25 , and Vibrio cholerae 26 . In these structures, DHBPS mostly exists as homo-dimer (monomer-A and monomer-B), and Ru5P binds in the active site of each monomer. An active site cavity is formed mostly by loops of one monomer with a side being lined by loops of another monomer. Accordingly, in Vibrio cholera DHBPS, monomer-A active site is surrounded by loop-1 (34-41), loop-2 (82-98), and loop-3 (175-185) of the monomer-A, and loop-4 (103′-111′ ) and loop-5 (132′ -138′ ) of the monomer-B 26 . A similar arrangement of loops is found in the active site of monomer-B, as shown in Supplementary Fig. S1. DHBPS-Ru5P structures have been solved for C. albicans 20,21 , S. typhimurium 25 , and V. cholerae 26 whereas DHBPS-Ru5P-ion complexes are solved for M. jannaschii 14 and V. cholerae 26 in the presence of inactive ions such as Zn 2+ and Ca 2+ . Numerous structural studies have speculated that inactive ions occupy the same position as that of active Mg 2+ , and mimics native like binding 14,18,25,26 . Failure of inactive ions to initiate the reaction, even if these bind at the same position, forming the same number of co-ordinate bonds, having similar charge and ionic radii to that of Mg 2+ , is attributed to their lower Lewis acid strength 14 .</p><p>Earlier, our group has solved the structures of Vibrio cholerae DHBPS in complex with Ru5P as well as in complex with Ru5P and Zn 2+ 26 . DHBPS-Ru5P-Zn 2+ structure shows two Zn 2+ along with Ru5P bound in each active site 26 . These ions occupy M1 and M2 positions forming coordinate bonds with the surrounding water, Ru5P, and DHBPS residues as shown in Supplementary Fig. S2. This assembly of the substrate-dimetal center is well established by several ion bound structures of DHBPS 14,18,25,26 . M1 and M2 positions of the ions coincide in all the structures of DHBPS that have complexed with metal ions. However, for M. jannaschii DHBPS 14 , the M2 position is slightly changed. Ca 2+ at this position forms interactions with nearby ligands with longer bond lengths than Zn 2+ at the M1. It has been speculated that the metal ion at the M1 position is sufficient to initiate the catalytic reaction, and the second ion may be involved in avoiding unproductive side reactions of highly reactive intermediates 14 . It has also been put forward that more DHBPS structures may validate the exact positions of two metal ions, especially in the presence of Mg 2+ 26 . Zn 2+ bound structure of DHBPS-Ru5P complex is obtained by soaking the crystals of DHBPS-Ru5P complex into the solution of ZnCl 2 26 . This experiment suggests that ions enter the active site from the bulk solvent. However, the mechanism behind entry and positioning of ions into the active site is still unknown. Comparison of DHBPS-Ru5P and DHBPS-Ru5P-Zn 2+ structures show that the loop-2 is partially ordered in the absence of ions, but becomes fully ordered in the presence of Zn 2+ . A similar change is also observed in conformation of Ru5P where flexible Ru5P becomes more rigid in the presence of Zn 2+ . Overall, it reveals that the presence of ions brings noticeable changes in DHBPS-Ru5P complex.</p><p>So far, most of the crystallographic studies that have been performed in the presence of inactive metal ions display interactions presumably formed by the Mg 2+ . Despite the availability of the crystal structures of DHBPS, and its complexes with metal ions and the substrate, the molecular mechanistic insights such as reordering of loop-2 in the presence of the metal ions, events marking the entry of metal ions, subsequent conformational changes in substrate (Ru5P) etc. are still elusive. Therefore, molecular dynamics (MD) simulations of DHBPS-Ru5P complex have been carried out in the presence of Mg 2+ for the first time. Simulations on wild type DHBPS and mutant complexes have depicted a possible path traveled by ions into the active site, besides the support provided by conserved residues for the positioning of metal ions.</p><!><p>Molecular dynamics of DHBPS-Ru5P complex. Multiple MD simulations (in triplicate) are carried out on a system, where the crystal structure of DHBPS-Ru5P complex is present along with explicit water molecules and ions (Mg 2+ , Na + , and Cl − ), by changing the initial velocities. In all of these simulations, Mg 2+ has entered the active site. In two simulations, it has entered the active site of one monomer while, in the third simulation, it has entered the active sites of both monomers. Results of the simulations, in which Mg 2+ has entered the active site of a monomer, are described here as these encompass the changes associated with the presence and the absence of ions in the active sites. During simulation, we have monitored parameters like volume, pressure, temperature, density, and energy to ensure the stability of the system. We have compared the conformations of DHBPS sampled during 50 ns of MD simulation by superimposing these on the crystal structure. Root mean square deviation (RMSD) of the conformations sampled during production simulation has been found to be fluctuating around an average of 1.7 Å within a narrow range (Fig. 1). Radius of gyration values are also stabilized suggesting that the dimer complex has preserved its folded form over the course of 50 ns (Supplementary Fig. S3). To understand the dynamics of C α atoms of the protein, root mean square fluctuations (RMSF) are calculated, and are presented in Supplementary Fig. S4. The results indicate that the structure of DHBPS-Ru5P complex is stable, and can be used for further analysis. The plots of RMSD and radius of gyration versus time, for other two simulations are provided in Supplementary Fig. S5.</p><p>Additional MD simulations of DHBPS-Ru5P complex with neutral His-154 show that an ionic interaction between the phosphate group of Ru5P and the guanidine group of Arg-38 is lost after 15 ns of simulation in both the monomers (Supplementary Fig. S6). Loss of this interaction increases the distance between center of mass of loop-1 residues and Ru5P atoms from 8.5 Å to 13 Å (Supplementary Fig. S7), leading to a partial opening of the loop-1. Further, Mg 2+ could not enter the active site of any of the monomers of DHBPS-Ru5P complex, suggesting the existence of His-154 in the protonated form. 2+ . In a system of DHBPS-Ru5P complex prepared for MD simulations, every Mg 2+ has six water molecules in its first water shell forming co-ordinate bonds. At the beginning of MD simulations, all of the Mg 2+ are more than 10 Å away from the protein. However, at the end of 50 ns simulation one Mg 2+ has entered the active site of monomer-A, and has stayed at a distance of 5-6 Å from O2 and O3 of Ru5P. This Mg 2+ has been in an octahedral coordination shell with surrounding water molecules during the entire period of simulation (Supplementary Fig. S8). To understand the conformational changes associated with the movement of Mg 2+ in the active site, we have superimposed the last structure of 50 ns simulations with the crystal structure (Fig. 2).</p><!><p>Analysis of the active site of monomer-A reveals that at 50 ns, C α atoms of loop-1 residues are deviated sideways by a distance of 1.5 Å to 3.2 Å from that of the crystal structure. C α atoms of Loop-3 residues are also moved sideways by a distance of 1.5 Å to 3.6 Å. These loop movements are away from Ru5P. In line with above loops, the loop-4 of monomer-B also moves away from Ru5P. In contrast to loop-1 and loop-3, loop-4 does not retain its original conformation as it varies in the side chain as well as backbone conformation. Thus, all the three loopsloop-1, loop-3, and loop-4 that are protruded inside the active site of crystal structure, are retracted backwards in the 50 th ns structure of simulations. Unlike these loops, loop-2 has shown movement into the active site. In monomer-A, C α atoms of Ala-91 and Asn-92 of this loop have shifted into the active site by a distance of 1.6 Å and 1.7 Å, respectively from their initial position (Fig. 3a). Asn-92 of this loop interacts, through its sidechain oxygen, with Mg 2+ entering the active site, as well as through its sidechain nitrogen, with one of the oxygens of a phosphate group of Ru5P. Once formed, these interactions are continued throughout the simulation period (Supplementary Fig. S9). In addition to these interactions, sidechain oxygen of Ser-90 and backbone nitrogen of Ala-91 form transient polar interactions with the water molecules that are surrounding the Mg 2+ . However, in monomer-B, loop-2 is shifted away from the active site. C α atoms of Ala-91 and Asn-92 have moved away from the active site by a distance of 2.3 Å and 1.8 Å, respectively from their original position (Fig. 3b). The different behavior of loop-2 in monomer-A and monomer-B may be attributed to the presence of Mg 2+ in the vicinity. Unlike in monomer-A, Asn-92 in monomer-B is unable to form hydrogen bond with the phosphate oxygen for the longer period, as observed in multiple simulations.</p><p>We have analyzed backbone flexibility of loop-2 residues during 50 ns of simulation by calculating B-factors. It has been observed that residues 89-93 are less flexible in monomer-A compared to that in monomer-B (Fig. 4). Substrate conformations are also observed to be compact in monomer-A while these are variable in monomer-B, as observed in multiple MD simulations. In the active site of monomer-A, O3 of Ru5P shows hydrogen bond interaction with the side chain of Asp-43. This interaction has started after the entry of ions into the active site, and continued throughout the simulation. However, it is not observed in monomer-B (Supplementary Fig. S10). 2+ into active site: Loop-4 as lid. During simulations, conformational changes in loop-1 and loop-4 are associated with the entry of Mg 2+ into the active site of monomer-A. In the initial phase of simulation, the water molecule, W6, becomes unstable, and moves away from its position (Fig. 5). It leads to loss of interactions of Glu-39 and Glu-41 with Ru5P, as these are formed via the W6. Due to the absence of strong interactions with other residues and Ru5P, Glu-39 moves away from Ru5P and becomes exposed to the solvent. Glu-41 also loses its interaction with Thr-108′ of loop-4. This event is then followed by the movement of Mg 2+ towards monomer-A. Positively charged Mg 2+ is attracted by the negatively charged residues of loop-1. Asp-37 being exposed to the solvent, initiates electrostatic interactions with the incoming Mg 2+ . Transient and subsequent interactions with Asp-37, Glu-39, and Glu-41 guides Mg 2+ towards the active site. Before entering the active site, Mg 2+ is stabilized by both the exposed residues, Glu-39 and Glu-41.</p><!><p>Entry of this ion into the active site of monomer-A is blocked by loop-4 of monomer-B, especially by Thr-107′ and Thr-108′ . It has been noted that O2 and O3 of Ru5P act as points of contact for Mg 2+ to enter the active site. However, movement of loop-4 residues is essential to make O2 and O3 atoms of Ru5P accessible to Mg 2+ . This is achieved by the formation of hydrogen bond between the side chain oxygen of Thr-108′ and the side chain oxygen of Asp-114′ (Fig. 2). The hydrogen bond between side chains of Thr-108′ and Asp-114′ occurs at the expense of conserved water molecule, W1 (Fig. 5). In the first structure of simulation, both the residues are in contact with each other by forming water (W1) mediated hydrogen bond. During MD simulations, the water molecule, W1, becomes unstable, and pulls the interacting hydroxyl group of Thr-108′ to take its position, and subsequently leaves its place. The hydroxyl group of Thr-108′ then initiates hydrogen bond interactions with backbone nitrogen atoms of Gly-109′ , Val-110′ , and Ser-111′ and side chain oxygen of Asp-114′ (Supplementary Fig. S11) leading to reorientation and opening of loop-4, thereby creating space for the entry of Mg 2+ . Mg 2+ then moves by 3 Å towards the Ru5P from a distance of 8-9 Å, and gets stabilized at 5-6 Å by interacting with the O2 and O3 atoms of Ru5P (Fig. 6). However, above mentioned interactions are not observed in the crystal structure of DHBPS-Ru5P complex, and point of entry of metal ion is covered by residues of loop-4 (Fig. 7). Thus, it appears that this loop-4 of one monomer acts as "lid" over the active site of another monomer.</p><!><p>plex suggest that movement of Mg 2+ in the active site has been channeled by the formation of hydrogen bond between Thr-108 and Asp-114. To substantiate the observed set of events, we have mutated these residues in both monomers of DHBPS-Ru5P complex, and have observed the movement of Mg 2+ during MD simulations. RMSD analysis of these mutant DHBPS-Ru5P complexes shows that these complexes are stable during MD simulations (Supplementary Fig. S12).</p><p>Thr108Ser mutant DHBPS-Ru5P complex. At the end of 50 ns MD simulation of Thr108Ser mutant DHBPS-Ru5P complex, one Mg 2+ is stabilized in the active site of monomer-A. Ser-108′ forms hydrogen bond with the carboxylic group of Asp-114′ through the sidechain hydroxyl group during the simulation. This hydrogen bond is formed in the initial few picoseconds, and continued till 50 ns (Supplementary Fig. S13), resulting in the formation of a necessary cavity for Mg 2+ to enter the active site. After entering the active site, Mg 2+ is stabilized at its position throughout 50 ns simulation. However, in monomer B, one of the Mg 2+ is attracted towards the acidic active site loop, and is stabilized at a distance of ~ 10 Å from Ru5P without entering the active site. The hydrogen bond between Ser-108′ and Asp-114′ is not observed, keeping loop-4 at its position, and forbidding the entry of ion (Supplementary Fig. S14). This simulation supports that formation of hydrogen bond between Thr-108 and Asp-114 is essential for the entry of ions into the active sites of DHBPS.</p><p>Thr108Val mutant DHBPS-Ru5P complex. In this study, we have mutated Thr-108 to valine. Valine has the same number of heteroatoms in the sidechain as that of threonine, but it lacks the side chain hydroxyl group. It is expected that Mg 2+ may not enter the active site as the absence of hydrogen bond between Val-108 and Asp-114 shall not reorient loop-4. MD simulations depict that one Mg 2+ has entered the active site of monomer-A, and is stabilized, contrary to our expectations. The backbone nitrogens of Val-108′ and Thr-107′ form strong hydrogen bonds with the side chain of Asp-114′ . Before the entry of Mg 2+ into the active site, the hydroxyl group of Thr-107′ also interacts with the carboxyl group of Asp-114′ . These interactions have caused flipping of dimethyl group of valine as well as reorientation of loop-4, opening a space for metal ions to enter. Mg 2+ enters the active site, and is stabilized by the formation of interactions with O3 and O4 of Ru5P, instead of O2 and O3 of Ru5P as observed in the crystal structure. Further, the conformation of Ru5P in this mutant complex is unlike that of the crystal structure. These observations suggest that mutation of Thr-108 to valine allows entry of Mg 2+ into the active site, but it may not result in the proper orientation of Mg 2+ and Ru5P, as required for the catalysis.</p><p>Asp114Ser mutant DHBPS-Ru5P complex. In another mutational study, we have replaced Asp-114 with serine. The side chain of serine is shorter in length by one carbon atom, and it may not form interactions like aspartic acid as observed in the wild type. During MD simulations of this complex, Mg 2+ does not enter the active site of any of the monomers. In monomer-A, one of the ions gets stabilized at 11 Å from the substrate without entering active site throughout simulations as the hydrogen bond between Thr-108′ and Ser-114′ is not formed (Supplementary Fig. S15).</p><!><p>In this mutational study, Asp-114 in DHBPS-Ru5P complex has been mutated to leucine. MD simulations show that none of the Mg 2+ interacting with the acidic active site loop has moved into the active site of both the monomers. As Leu-114′ lacks sidechain carboxylic group, it cannot form the hydrogen bond with the loop-4 residues viz. Thr-107′ , Thr-108′ , and Gly-109′ . It supports that these interactions are essential for the opening of loop-4, suggesting loop-4 as a "lid" over the active site.</p><p>Glu39Ala, Glu41Ala, and His154Ala mutant DHBPS-Ru5P complexes. The root mean square deviations of these mutant complexes have been shown to be stable during MD simulations (Supplementary Fig. S16). The analysis of results shows that a Mg 2+ enters the active site in a few of the MD simulations. However, the conformations of Ru5P, and the positions of stabilized Mg 2+ does not superpose with those of wild type (Supplementary Fig. S17). Further, the formation of hydrogen bond between Thr-108 and Asp-114 has been observed before the entry of Mg 2+ .</p><!><p>DHBPS is vital to the survival of many pathogens and thus considered as a potential drug target to design antibacterial agents. Mechanistic and dynamic details of the binding of substrate, ions, and the role played by residues are required to understand the functioning of DHBPS, as well as for the effective design of novel inhibiting agents. In a first of its kind, we have performed conformational analysis of DHBPS and Ru5P in the presence of Mg 2+ ions using MD simulations. During 50 ns of simulation of DHBPS-Ru5P complex, loop-2 in the active site of monomer-A has been found to be ordered and stable. It forms interactions with entered Mg 2+ and phosphate oxygen atoms of Ru5P. However, loop-2 in the active site of monomer-B is ordered but less stable, and does not form interactions with the substrate for most of the time of simulation. Similar behavior of the loop-2 is reported in Ru5P complexed crystal structures. In DHBPS-Ru5P complex of V. cholera 26 and C. albicans 20 , this loop is partially disordered whereas, in DHBPS-Ru5P-Zn 2+ complex of V. cholera 26 and M. jannaschii 14 , it is in ordered conformation, forming weak water mediated interactions with the phosphate group of Ru5P. However, in MD simulations this loop forms strong and direct interactions with the phosphate group via Asn-92. This study highlights the role of loop-2 in positioning and stabilization of Mg 2+ at M2 position along with the stabilization of phosphate group of Ru5P. In the absence of ions, this loop does not significantly contribute to the stabilization of substrate as its residues remain flexible and interact transiently.</p><p>MD simulations show substrate, Ru5P to be flexible in monomer-B but adopts a compact conformation in monomer-A. Interaction analysis of Ru5P reveals that Ru5P undergoes Mg 2+ induced conformational change such that its O2 and O3 hydroxyl groups reorient to interact with the entered Mg 2+ . This reorientation is stabilized by the formation of hydrogen bond between O3 of Ru5P and sidechain oxygen of Asp-43. These events are supported by the observation that O3 of Ru5P moves by ~1 Å towards the sidechain oxygen of Asp-43 in the presence of ions upon comparison of DHBPS-Ru5P and DHBPS-Ru5P-Zn 2+ complexes. These two oxygen atoms then become a part of strongly bonded coordinate geometry of Mg 2+ providing conformational rigidity to Ru5P in monomer-A. The high flexibility of Ru5P in the monomer-B can be attributed to the absence of ion in the active site, as well as weaker binding of its hydroxyl groups with surrounding residues such as Asp-43. In the absence of Ru5P (i.e. O2 and O3 atoms), either phosphate or sulphate is present along with water molecules for necessary interactions with incoming Mg 2+ 14,17,18,25,26 . Even in the apo structure of V. cholera, water molecules present in the active site superpose well with hydroxyl oxygen atoms of Ru5P 26 .</p><p>In this study, we have performed MD simulations of DHBPS-Ru5P complexes using protonated His-154, and have found that His-154 forms stable interactions with a phosphate group of Ru5P, similar to that in the crystal structure. In another simulation of DHBPS-Ru5P complex, having neutral His-154, these interactions are lost resulting in destabilization of Ru5P. Interactions of metal ion at position M1 with residues like His-154 are found to be conserved in separate MD simulations of metal bound DHBPS-Ru5P complex having neutral His-154. However, these interactions are lost in the presence protonated His-154, destabilizing ion at M1 as well as Ru5P. In addition, pKa prediction results show that His154 is protonated in DHBPS-Ru5P complex while it is neutral in DHBPS-Ru5P-Zn 2+ complex. Our study finds that only one Mg 2+ gets recruited in each of the active sites of DHBPS monomers, and it occupies the position close to M2 of DHBPS-Ru5P-Zn 2+ complex. The second Mg 2+ that occupies the M1 position, coordinating with His-154, does not enter the active sites as protonation states of His-154 are not changed on the fly during the MD simulation. It is likely that first ion that enters the active site binds transiently at M2 position before moving to M1. This transfer, however, may require changes in the protonation state of His-154 from protonated to a neutral state, leading to the stabilization of ion at the position M1. Subsequently entering ion is then positioned at M2 site. Entry and stabilization of Mg 2+ at a position close to M2 can be attributed to the same protonation state of O2 and O3 atoms of Ru5P in DHBPS-Ru5P and DHBPS-Ru5P-Zn 2+ complexes. This is because of the fact that pKa of the hydroxyl groups of Ru5P is much higher (~12), suggesting that these groups are less likely to be deprotonated even if coordination to magnesium decreases their pKa values.</p><p>Formation of the hydrogen bond between the side chains of Thr-108 and Asp-114 is related to the transfer of ion into the active site, as observed during MD simulations. This interaction subsequently creates other hydrogen bond interactions between Thr-108 and loop-4 residues viz. Gly-109, Val-110, and Ser-111 providing stability to the flipped Thr-108. These residues are also involved in the reorientation of loop-4 along with the transfer of ion. This loop-4 acts as a "lid" that opens up and allows Mg 2+ to enter the active site of another monomer. Apo as well as Ru5P complexed structures of DHBPS, do not show the formation of this hydrogen bond, thus closing the entry point for metal ions. Thus, the "lid" is closed in DHBPS-Ru5P-Zn 2+ complexes, trapping ions in the closed state. We have analyzed the conserved nature of residues involved in ion transfer viz. Asp-114, Thr-108, Gly-109, Val-110, and Ser-111 across DHBPS of other microorganisms by sequence alignment (Fig. 8). The sequence identity of DHBPS between these microbial species and Vibrio cholerae ranges from 30% to 67%. Asp-114 is found to be mostly conserved. Only in S. pneumoniae, Asp-114 is replaced with another amino acid, i.e. glutamic acid (Fig. 8), which, however, conserves the side chain carboxylic oxygen atoms required for necessary interactions with Thr-108. MD simulations of Asp114Ser mutant complex show that ions do not enter the active site of both monomers, and may result in a reduction of catalytic activity of DHBPS, as seen in the same mutant of M. jannaschii 27 . It may be stated that DHBPS of Vibrio cholarae shows 30% amino acid identity with that of M. jannaschii, and both enzymes require Mg 2+ as a cofactor, and follow the same mechanism of catalysis 14,26 . In another MD study, results of Asp114Leu mutation show that entry of ion into the active site is prohibited. These two mutational studies establish the role of Asp-114 in ion transfer, and thus explain its conservation among DHBP synthases. Thr-108, another residue, is strictly conserved among DHBP synthases (Fig. 8). In MD study of Thr108Ser mutant complex, results have shown that an ion has entered the active site of DHBPS, similar to that in wild type. However, same mutation has resulted in the loss of activity of DHBPS in M. jannaschii 27 . This contrary result may be explained by the fact that the sidechain methyl group of Thr-108 stabilizes the sidechain of Glu-39 by hydrophobic interactions 14,26 , which may not be achieved by serine mutant due to absence of methyl group. Sidechain carboxylic oxygen atoms of Glu-39 have been shown to interact with metal ions at positions M1 and M2 14,26 and mutation of Glu-39 have resulted in the loss of activity of DHBPS in M. jannaschii 14 as well as in V. cholera 26 . In another MD study of Thr108Val mutant DHBPS-Ru5P complex, Mg 2+ enters the active site but it is stabilized away from M2 position, unlike Mg 2+ in wild type complex, explaining the loss of activity and importance of conservation of Thr-108. However, the catalytic activity of Thr-108 and Asp-114 mutants in V. cholerae needs to be validated experimentally. Nevertheless, results of MD studies on the mutant complexes involving mutations viz. Glu39Ala, Glu41Ala, and His154Ala suggest varied conformations of Ru5P and positions of Mg 2+ , explaining the loss of the activity of DHBPS, as observed in experimental studies 26 . Other residues such as Gly-109, Val-110, and Ser-111 that are involved in hydrogen bonding with Thr-108 are also either strictly or mostly conserved (Fig. 8). Glycine (− 109), is also proposed to play a structural role in E. coli DHBPS, and is predicted to have strained conformation for a non-glycine residue as there is no room for C β atom 16 . Thus, it is expected that Gly-109 allows Thr-108 to change its conformation and form the hydrogen bond with Asp-114. It is likely that the replacement of Gly-109 by any other residue may forbid the formation of this bond due to steric hindrance. Similar amino acids at positions of Val-110 and Ser-111 may participate in a reorientation of loop-4 as MD studies have shown that these residues interact with Thr-108 through backbone nitrogen atoms. Thus, conservation of these residues and interactions substantiates mechanism of ion transfer and puts forward that it may exist in other microorganisms as well.</p><!><p>Preparation of system for molecular dynamics. In this study, we have used DHBPS-Ru5P complex of Vibrio cholerae 26 . Structural coordinates of this complex have been obtained from protein data bank (PDB ID 4P77, resolution 2.04 Å). This structure is the homodimer of 431 residues (216 residues in chain A and 215 residues in chain B) with Ru5P bound in each of the active sites. The coordinates of residues-90-91 of A-chain and residues 87-93 of B chain are missing in this structure. These residues have been modeled from the Vibrio cholera DHBPS-Ru5P-Zn 2+ structure by direct transfer of coordinates (PDB ID 4P8E, resolution 2.04 Å) 26 . The complex has been prepared for MD simulations using various modules of AMBER14 28,29 . Initially, the protonation states of amino acids are determined in the presence of co-crystallized water molecules at pH 7.0 using PROPKA module and then verified manually 30,31 . Four of the residues, Glu-41, Asp-43, His-137, and His-154 are predicted to be protonated. Among these, visual inspection reveals that Glu-41 is exposed to the solvent, and Asp-43 forms anion-pi interactions (3.5 Å) with positively charged His-154, depicting these to be deprotonated. Then, according to the protonation states, hydrogen atoms have been added to all residues using LEaP module of AMBER14.</p><p>In Antechamber, for Ru5P, the AM1-BCC method has been utilized to calculate partial atomic charges 32,33 , and the general amber force field (GAFF) is used to assign the parameters 34 . Protein topology has been prepared with ff14SB force field 35 . Protein complex with its crystal waters is then solvated with TIP3P water model 36 in a truncated octahedral box of a volume of 579921.85 Å 3 , and the edge of the box is 8.5 Å from the protein. The system is neutralized by adding a requisite number of Mg 2+ and Na + ions. Subsequently, Mg 2+ and Cl − ions have been added to make 0.1 M concentration of the system.</p><p>Ions have been added in such as way that they are no closer than 10 Å from the complex, and are apart by 5 Å from each other. It avoids artifacts whereby ions may enter the active site because of their placement nearby the complex. A compromise (CM) set of ion parameters that reproduce the experimental and relative hydration free energy (HFE), and coordination number (CN) values have been used to define the divalent ions 37 . Force field parameters, fitting solvation free energies, radial distribution functions, ion-water interaction energies, and crystal lattice energies and lattice constants for non-polarizable spherical ions, developed by Joung and Cheatham are used for the monovalent ions 38,39 . A topology file containing force field parameters, and a co-ordinate file containing structural information of the complex, are then used as input to perform the MD simulations using pmemd.cuda module of AMBER14 29 .</p><p>At first, a two-stage energy minimization process has been used to allow the system to reach an energetically favorable conformation. In the first stage, positions of protein, Ru5P, and ions are restrained with a harmonic force constant of 10 kcal/mol/Å 2 while, explicitly added water molecules are allowed to move, and adjust their orientations. In the second stage, the force constant is reduced to zero, allowing the system to accommodate protein, Ru5P, water, and ions. The temperature of the system is then increased to 300 K using NVT ensemble over the period of 100 ps. In this step protein, Ru5P, and ions are restrained by a force constant of 0.5 kcal/ mol/Å 2 . Subsequently, the system is equilibrated such that it has a uniform density around 1 g/cm 3 , and has achieved stable RMSD of protein during a period of 10 ns. The sampling MD simulations are then performed in an isothermal-isobaric ensemble (NPT, P = 1 atm, and T = 300 K) for 40 ns. The time step is kept at 2 fs, while the trajectory is recorded every 4 ps. Simulation is repeated three times by changing the initial velocity. Langevin thermostat and Berendsen barostat are used for temperature and pressure coupling, respectively 40,41 . SHAKE algorithm is applied to constrain all the bonds containing hydrogen atoms 42 . A non-bonded cutoff is kept at 8 Å, and long range van der Waals interactions have been treated by Particle Mesh Ewald (PME) method 43 . Discovery Studio and Pymol have been used to prepare files, visualize 3D structures, and create graphics 44,45 while, VMD and Ptraj program are used to visualize trajectories, compare residues and ion movements, and compute interactions 28,46 .</p><!><p>During the simulations, amino acids Thr-108 and Asp-114 are observed to be critical for the movement of the metal ions in the active site. In DHBPS-Ru5P complex, these residues are mutated to serine (Thr108Ser and Asp114Ser) or valine (Thr108Val) or leucine (Asp114Leu) in LEaP module of AMBER14. As both residues do not form direct interaction with Ru5P, it is believed that their mutations do not affect the binding of substrate in the active site of DHBPS. In addition, catalytic residues viz. Glu39, Glu41, and His154 of DHBPS, have been mutated to alanine to understand the structural changes leading to loss of catalytic activity of the DHBPS. For MD simulations, the backbone of the mutated residues has been kept unchanged while side-chains are built into the structures with default geometry. The same protocol is followed to prepare systems for each mutation, and then subjected to MD simulations for 50 ns, in duplicate.</p><p>Sequence Alignment. Multiple sequence alignment is performed by aligning V. cholerae DHBPS with sequences from other species using default values in Clustal Omega 47 . DHBPS-Ru5P complex (PDB ID 4P77) is used to display the secondary structure on the top of the alignment. The alignment figure has been generated through ESPript 3 web server 48 , using BLOSUM 62 score matrix 49 .</p>

Scientific Reports - Nature

To Stay or to Leave: Stem Cells and Progenitor Cells Navigating the S1P Gradient

Most hematopoietic stem progenitor cells (HSPCs) reside in bone marrow, but a small amount of hematopoietic stem progenitor cells have been found to circulate between bone marrow and tissues through blood and lymph. Several lines of evidence suggest that sphingosine-1-phosphate (S1P) gradient triggers HSPC egression to blood circulation after mobilization from bone marrow stem cell niches. Stem cells also visit certain tissues. After a temporary 36 hours short stay in local tissues, HSPCs go to lymph in response to S1P gradient between lymph and tissue and eventually enter the blood circulation. S1P also has a role in the guidance of the primitive HSPCs homing to BM in vivo, as S1P analogue FTY720 treatment can improve HSPC BM homing and engraftment. In stress conditions, various stem cells or progenitor cells can be attracted to local injured tissues and participate in local tissue cell differentiation and tissue rebuilding through modulation the expression level of S1P1, S1P2 or S1P3 receptors. Hence, S1P is important for stem cells circulation in blood system to accomplish its role in body surveillance and injury recovery.

to_stay_or_to_leave:_stem_cells_and_progenitor_cells_navigating_the_s1p_gradient

Pathophysiological functions of sphingosine-1-phospahte<!>Hematopoietic stem progenitor cell trafficking<!>S1P gradient triggers cells trafficking<!>Bone marrow microenvironments: S1P gradient is essential for bone remodeling<!>The role of S1P in HSPC egression from BM<!>S1P1 Regulates the Egression of HSPCs from Tissues into Lymphatics<!>HSPC homing from the blood to the BM<!>Stem cell trafficking to injured tissues<!>1. Infection<!>2. Bone marrow MSC migration to injured Liver<!>3. Role of S1P in the migration of neural progenitor cells (NPCs) during brain infarction<!>4. Ischemia/Reperfusion (IR)<!>Discussions and Concluding Remarks<!>Future directions

<p>Sphingosine-1-phosphate (S1P), a serum-borne bioactive sphingolipid, regulates a variety of biological activities by acting either as an extracellular ligand or intracellular stimulus. As an extracellular ligand, S1P functions are mediated by the S1P family of G-protein-coupled receptors (GPCRs), identified as S1P1, S1P2, S1P3, S1P4, and S1P5 receptors (old nomenclature: EDG-1, EDG-5, EDG-3, EDG-6, EDG-8 respectively). Previously, it was showed that S1P1, a Gi-coupled plasma membrane GPCR abundantly expressed in endothelial cells (ECs), transduces S1P signaling to regulate endothelial survival, adherens junction formation, cytoskeleton architecture, and chemotactic response 1-3. Also, compelling evidences show that S1P can also function as an intracellular stimulus. Three intracellular targets of S1P have recently been identified as HDAC 4, TRAF2 5, and Prohibitin 2 6. The identification of intracellular targets of S1P suggests that S1P may directly regulate gene expression, protein turn-over, and cellular respiratory response 4-6. Moreover, S1P has been elegantly demonstrated to play critical roles in a wide array of pathophysiological functions, including angiogenesis and vasculogenesis, immune modulation, tumorigenesis, rheumatoid arthritis, asthma, inflammation, retinopathy, cardiovascular protection etc 1-6. Several of these S1P-regulated pathophysiological functions are extensively discussed in this review series, entitled "Sphingosine-1-phosphate in health and disease". Therefore, this review will focus on one emerging area of the S1P-regulated biological activities, i.e. the trafficking of hematopoietic stem progenitor cells (HSPCs).</p><!><p>Stem cells, the hope of the future tissue regeneration, reside in specific bone marrow (BM) niches for the long term hematopoietic reconstitution. Bone marrow contains two major populations of stem cells: hematopoietic stem progenitor cells (HSPCs) and mesenchymal stem cells (MSCs). HSPCs can remain in the stem cell niches within the bone marrow for a long time to constantly replenish the various short lived differentiated blood cells such as red blood cell and white blood cells. MSCs, the stroma in BM, provide a supporting role in HSPCs maintenance, survival and differentiation in BM. MSCs also play a key role in the regenerative medicine because its mighty capacities to differentiate into a variety of cell types such as myoblast, fibroblast etc. BM is not the only residential place for the stem cells. Recently, tissue specific stem cells have also been identified in various tissues to support local tissue renewal. Moreover, circulatory stem cells have been found to travel in blood circulation. The continuous trafficking of stem cells between BM and blood is not only to fill the empty distal bone marrow niches, but it may also keep the tissues in surveillance and provide effector cells to foster local tissue regeneration during injury 7-10. Thus, the understanding of the mechanism behind stem cell trafficking is fundamental to both stem cell biology and stem cell transplantation.</p><p>The bone marrow microenvironment provides hematopoietic stem cells with an unique capacity for self-renewal, multilineage differentiation and long-term survival 11. Stem cells homing to the bone marrow niches is mediated by a complex array of molecular interactions that include adhesion molecules, proteolytic enzymes, and cytokines, especially chemokines such as stromal cell-derived factor 1 (SDF-1, CXCL12) 12-13. SDF-1 is a key HSPCs chemotactic factor particularly expressed by endothelial cells along the endosteum region in the bone marrow niches. Binding of SDF-1 to its receptor CXCR4 induces migration and activation of human HSPCs, CD34–expressing cells 14-17. Furthermore, in SDF-1–null mice (embryonic lethal), hematopoietic progenitor cells fail to migrate from the fetal liver to the bone marrow 18. Although there is a very significant correlation between migration capacity and the expression of CXCR4, the homing capacity of CXCR4null stem cells are not completely lost 19-20. Recently, Ratajczak et al. found that hematopoietic stem progenitor cells (HSPC) egression from the bone marrow may occur in an SDF independent process 21. These findings suggest that stem cell homing and egression does not appear to depend exclusively on the interaction of CXCR4 and SDF-1; other chemotactic receptors and their ligands also play a role in stem cell migration.</p><p>Among the chemokines and inflammatory mediators known to exert potent cellular chemotactic effects, sphingosine-1-phosphate (S1P) is a good candidate for the induction of stem cell trafficking. S1P is an important bioactive lysophospholipid secreted in the blood plasma upon platelet activation. S1P has been demonstrated to induce the chemotaxis of human natural killer cells, immature dendritic cells, and endothelial cells 22,23. Several S1P receptors (S1PRs) appear to be expressed on murine hematopoietic progenitor cells, suggesting a role of the lipid mediator S1P in the hematopoietic microenvironment 24,25. The continuous presence of S1P in the hematopoietic microenvironment and the expression of S1P receptors on the hematopoietic progenitors raised the possibility that S1P might modulate the SDF-1/CXCR4-dependent HSPC homing and lodgment. A clearer understanding of S1P in the physiological regulation of this process may aid in improving the efficiency of stem cell bone marrow transplantation.</p><!><p>S1P is generated in the process of normal sphingolipid turnover and it is a critical signaling molecule in normal development 26-27 (also see review of Binks W. Wattenberg in this hot-topics review series: The Role of Sphingosine Kinase Localization in Sphingolipid Signaling). It is well known that S1P induces numerous biological responses such as cell proliferation, differentiation, apoptosis, cytoskeletal remodeling and migration 1-6, 23,27,28. The effects of S1P are mediated by the S1P family of G protein coupling receptors, the sphingosine-1-phosphate receptors (S1P1- S1P5) formerly termed endothelial differentiation gene (EDG) 1,28. S1P stimulates distinct pathways such as the Rho, phospholipase C, Ras, MAP kinase and PI3K pathways dependant on the expression patterns of tissue- and cell type-specific S1P receptors 1-3, 23,27-30.</p><p>S1P is a circulating bioactive lipid mediator that is abundant in the lymph and blood, largely in forms bound to plasma proteins including HDL and albumin (see review of Koichi Sato and Fumikazu Okajima in this hot-topics review series: Role of sphingosine 1-phosphate in anti-atherogenic actions of high-density lipoprotein). Plasma contains low micromolar concentrations of S1P which is mainly generated by radiation-sensitive hematopoietic cells such as erythrocytes and platelets. Lymph contains S1P in the nano-molar range, which is derived from a radiation-resistant source, possibly endothelial cells 31. S1P lyase, which is abundant in many tissues but absent in circulation, causes rapid interstitial S1P degradation, leading to the establishment of the S1P gradient 32. The S1P gradient is critical for the migration of heart progenitors and directed migration of prechordal plate progenitor cells during zebrafish development 33,34. Recently, the S1P gradient between lymphoid tissue and lymph fluid was reported to be necessary for lymphocyte egression from lymph nodes 32. There is an emerging concept of specific cell type population migrations being triggered by distinct S1P receptors. For example, S1P1 is involved in T cell egression from the thymus, and T and B cell egression from the peripheral lymphoid organs 31,32. However, S1P5 was identified to be responsible for nature killer (NK) cells egression from BM and lymph nodes, and the egression of CD8+ T cells from the lymph nodes 35. Although some S1P receptors promote cell migration, others may play an inhibitory role. In particular, S1P1 and S1P2 are homeostatically involved in S1P-induced chemotaxis 36,37. Expression of high levels of S1P1 promotes chemotaxis 3,23,28,36,37, whereas S1P2 has been shown to inhibit migration 36,37 (also see review of Yoh Takuwa et al. in the series: Roles of sphingosine-1-phosphate signaling in angiogenesis). Therefore, regulation of S1P receptors expression is essential for the trafficking of variety of immune cells during development, in normal physiological conditions and during inflammation.</p><!><p>The bone marrow microenvironment provides hematopoietic stem cells with an unique capacity for self-renewal, multilineage differentiation and long-term survival 11. Osteoclasts (OC) and osteoblasts (OB) are derived from HSPC myeloid lineage differentiation and MSC differentiation respectively. Several lines of evidence support that S1P controls the migration and differentiation of osteoclast and osteoblast precursors to dynamically regulate bone mineral homeostasis and bone marrow stem cell mobilization 38,39. During RANKL (Receptor Activator for Nuclear Factor κ B Ligand) stimulated osteoclast differentiation, the expression of S1P1 receptors is suppressed (Figure 1) 38. Osteoclast precursors express functional S1P1 receptors and exhibit positive chemotaxis along a S1P gradient in vitro. Osteoclast /monocyte specific S1P1-deficient (S1P1 −/−) mice were found to be osteoporotic with lower bone tissue density, lower trabecular thickness and density, and higher osteoclast attachment to the bone surface 38. Using intravital two-photon imaging of bone tissues, it was shown that the selective S1P1 agonist SEW2871 (which functionally activates S1P1) stimulated the motility of osteoclast precursor-containing monocytoid populations in vivo. Furthermore, SEW2871 treatment caused a decrease of monocytoids in the bone marrow and an increase of these cells in the peripheral blood circulation. Lastly, S1P analogue FTY720 treatment significantly relieved ovariectomy-induced osteoporosis in mice 38. These results strongly support that S1P gradient plays a critical role in directing the correct localization of the maturing OC cells to the bone surface. Also, the concentration of S1P in the blood is higher than that in tissues. This S1P gradient favors a recirculation of osteoclast precursor monocytes from bone tissues to systemic blood circulation, which could prevent excessive mature osteoclasts from causing bone destruction 38 (Figure 1).</p><p>Contrary to osteoclast differention, it was shown that the response to S1P during osteoblast differentiation is controlled by the developmental-stage specific expression of the S1P2 receptors 39. Migration of osteoblast precursor mesenchymal stem cells is controlled by a number of growth factors and cytokines such as Platelet Derived Growth Factor (PDGF-BB). The chemotaxis towards PDGF was inhibited by S1P in pre-osteoblasts so S1P acts as a chemorepellent in this scenario. Treatment with a highly selective S1P2 antagonist JTE-013 or ablation of S1P2 expression by RNA interference blocked the inhibitory effect of S1P on PDGF-induced chemotaxis, which suggests that S1P2 is responsible for mediating the inhibitory effect of S1P 39. Treatment of bone morphogenetic protein 2 (BMP2) induces the differentiation of pre-osteoblasts to alkaline phosphatase positive, mature osteoblasts. Strikingly, the inhibitory effect of S1P2 disappeared concomitantly with the down-modulation of S1P2 expression in the mature differentiated osteoblasts. Constitutive S1P2 expression confirmed that the presence or absence of S1P2 receptors is the sole determinant accounting to the change in S1P response during the BMP2-induced pre-osteoblast to osteoblast conversion 39 (Figure 1). Thus, upon conversion to the osteoblast phenotype, S1P2 receptor expression is repressed to favor chemotaxis induced by PDGF in bone tissue for new bone formation; whereas migration of immature pre-osteoblasts is restricted due to high S1P2 expression. This may reflect a mechanism that stem cells use to preserve the progenitor pool, only allowing the more differentiated cells to travel to sites of bone formation.</p><p>In summary, the migration of osteoclast precursors from the bone to the circulation is induced by S1P1expression, whereas osteoblast precursors stay in bone for bone remodeling via S1P2 expression. Furthermore, undifferentiated and differentiated OB and OC cells respond differently to the S1P gradient (Figure 1). The differential expression of specific S1P receptor subtypes during bone remodeling may be essential for bone marrow microenvironments, as it reflect a finely-tuned dynamic control of stem cell/progenitor cells trafficking during health and in various physiological conditions.</p><!><p>Understanding the process of HSPC mobilization will help a significant number of patients such as those who are poor HSPC mobilizers for the bone marrow transplantation 40,41. It has been found that small amount of HSPCs circulate in the peripheral blood circulation (PB) under steady-state conditions. Significant amounts of HSPCs can be mobilized from the BM into the PB during infection, tissue injury, and after administration of some pharmacological agents 42,43. Mobilization of hematopoietic progenitor cells using granulocyte colony-stimulating factor (G-CSF) is a multifactorial process caused by modulating the activity of granulocytes and the release of proteolytic enzymes to interfere with the major retention signals for HSPC in bone marrow such as SDF–CXCR4, VLA-4–VCAM-1, and cKit ligand–c-Kit receptor axes 42-44.</p><p>SDF-1 is essential for HSPC anchorage to the stem cell niches in the BM. The plasma concentrations of SDF-1 in either normal or chemical-induced mobilization individuals are low and should be insufficient to chemoattract murine BM HSPCs into circulation 21. Also, plasma-stimulated HSPC chemotactic activity was almost completely abolished after charcoal stripping the plasma, suggesting that bioactive lipids present in the plasma is required to mobilize HSPCs. S1P is a major chemoattractant that is several magnitudes higher in concentration than SDF-1 in normal plasma under steady-state conditions. Thus, S1P at physiologically relevant concentrations may already create a S1P gradient that continuously chemoattracts BM-residing HSPCs (Figure 2A). Erythrocytes are a major source/reservoir of S1P in the PB and form a buffer system that controls S1P levels in the PB as seen during hemolysis 45-48. It has been reported that the complement complex (CC) is activated in the BM during mobilization of HSPCs and erythrocyte lysates resulting from complement activation have a strong chemotactic effect on HSPCs. The S1P gradient is maintained by co-enzyme vitamin B6 dependent S1P lyase. DOP, a vitamin B6 antagonist, decreases S1P lyase activity in tissues. DOP-treated mice are poor mobilizers of HSPCs 21. Stem cells from DOP-treated BM do not respond to a S1P gradient, possibly because of exposure to oversaturation of S1P in the BM environment due to lack of S1P lyase activity (Figure 2A). Interruption of active anchorage of HSPCs in the BM might shift the BM-retention signal towards a plasma S1P gradient that directs the egression of HSPCs into the PB. These results suggest that the S1P gradient is important to stem cell mobilization from BM to peripheral blood circulation and failure in creating a S1P gradient from the BM to the PB greatly affects the HSPC mobilization (Figure 2A). However, it should be noted that the retention of HSPCs in the BM may be primarily regulated by the SDF-1/CXCR4 signaling. The S1P signaling might function in regulating the BM retention of HSPCs only when the SDF-1/CXCR4 signaling is abrupted.</p><!><p>Increasing evidence supports that circulating HSPCs also visit extramedullary tissues such as the liver 49 and spleen 10. An elegant experiment was performed in GFP and non-GFP parabiotic mice 50. Three days after crosscirculation was established, strong colony formation units (CFUs) chimerism was found in the blood and lymph, indicating that some HSPCs recirculate freely between the lymph and blood. Spleen, lung, liver and kidneys had the highest level of chimerism in the extramedullary tissues of parabiotic mice. The mean time of HSPCs that homed to the peripheral tissues was at least 36 hrs. It was estimated that about 200 clonogenic HSPCs passed through the lymph of the mice every day and at least twice as much HSPCs residing in extramedullary nonlymphoidal tissues 10.</p><p>Consistent with the parabiotic experiment, clonogenic progenitors were detected in many tissues, including the lung, liver, kidney, and blood 50. In addition, these tissue-derived clonogenic cells possess a capacity for multilineage reconstitution. Those tissue-derived clonogenic HSPCs were identified as BM derived using chimeric wild-type recipients of GFP+ BM. Lymph-borne HSPCs are also BM derived, and can recirculate back into the BM to maintain blood homeostasis. The primitive clonogenic HSPCs found in the efferent lymphatics possess the capacity for long-term multigenerational reconstitution, which meets the phenotypic and functional criteria for true HSPCs 50. Unlike lymphocytes, HSPCs travel directly from the extramedullary nonlymphoidal tissues to the lymph and do not necessarily need secondary lymphoid organs. Because the mammalian BM lacks lymphatic drainage, BM HSPCs are thought to egress directly into the blood. After BM stem cells traffick out of the BM directly into the blood, they travel constitutively to multiple extramedullary nonlymphoidal tissues, where they reside for at least 36 hrs until entering the draining lymphatics to await return to the bloodstream 50. Hematopoietic progenitors constitutively circulating in extramedullary tissues might provide a role in constitutively replenishing the different populations of old or damaged cells in the tissue microenvironment. For instance, after deposition into the kidney, injected GFP+ HSPCs can locally differentiate into various myeloid lineages 50.</p><p>Not much is known about the mechanism of HSPC trafficking from the blood into extramedullary tissues. However, there is evidence suggesting that S1P may play a function in tissue-residing HSPC egression to the lymphatics 50. Lymph-borne HSPCs were markedly reduced in S1P lyase blocker 2-acetyl-4-tetrahydroxybutylimidazole (THI)-treated mice, suggesting that the S1P gradient is required for HSPC egression from tissues and migration to lymphatics. Treatment with Gαi inhibitor pertussis toxin (PTX) caused a dramatic decrease in the number of lymph-borne CFU-Cs in 2–12 hrs, which implies an essential role for Gαi-mediated signaling to direct tissue-residing HSPCs into the draining lymphatics. Both treatment with FTY720 or selective S1P1 agonist SEW2871 depleted HSPCs from the lymph within 6 hrs, indicating S1P1 control over HSPCs exiting from nonlymphoid tissues into the draining lymph vessels 50,51. The treatment of mice with FTY720 over 7 days resulted in a significant increase in the number of HSPCs residing within extramedullary tissues. Hence, the S1P/ S1P1 signaling as well as S1P gradient might serve as a mechanism guiding stem cells and various progenitor cell populations egression from tissues into lymphatics (Figure 2B).</p><!><p>The critical aspects of primitive hematopoietic cell homing to the bone marrow after transplantation are fundamentally important for the self-renewal and development of hematopoietic stem cell. Characterization of the mechanism behind S1P-mediated HSPC migration is of great relevance to the understanding of stem cell transplantation. Nobuaki et al. first found that S1P triggers an invasion of the primitive hematopoietic Lin /Sca-1+/c-Kit+ expressing cell line (THS119) into the stromal cell layers in vitro to form cobblestone areas 24, which reflect the proliferation of primitive hematopoietic cells in the hematopoietic microenvironment. They also found that PTX, an inhibitor of trimeric Gi proteins, partially inhibited THS119 invasive activity 24,52. Therefore, the migratory ability of HSPCs was mediated at least in part by signals from GPCRs eliciting intracellular events that control proliferation and motility. The invasive ability of HSPCs was mediated by the lysophospholipids (LPL) S1P and LPA. LPL receptors are known to activate small-GTPase proteins Rac/Rho/Cdc42 to mediate cell migration. Indeed, C3 exotoxin, an inhibitor of Rho, partially inhibited THS119 invasive activity, which suggests that this is a Rho dependent signaling pathway. Both S1P and LPA induced THS119 invasion may be similar to the homing of hematopoietic stem cells 24. Indeed, Whetton found that LPL synergistically promotes SDF-1 mediated primitive hematopoietic cell chemotaxis 25. Studies with Rac/Rho/Cdc42 inhibitor Clostridium difficile B toxin, Rho GTPase activated kinase inhibitor Y27632, and Rac/Rho/Cdc42 guanyl nucleotide exchange factor Vav1-null mice, indicate a role of these G proteins in LPL and SDF-1–induced migration 25. PI3K inhibitors almost completely inhibit SDF-1 and/or LPL induced migration in Lin−Sca+Kit+ cells. Thus, it has been proposed that S1P receptors expressed in primitive hematopoietic cells bind cognate ligands to activate PI3K and thereby Vav-1, which in turn affect the Rho small-GTPases to control pluripotent cell motility 25.</p><p>Supporting the involvement of S1P in stem cell migration, S1P receptors were found in both murine and human HSPCs. S1P1, as well as S1P2, S1P3, and S1P4 have been found to be expressed on murine HSPCs. Among different donors, S1P1 mRNA was consistently expressed in human CD34+ stem cells 53. In another study, S1P5 was found in the more primitive human progenitor CD34+/CD38− cells and S1P2 in the more committed human progenitor CD34+/CD38+ cells 25. S1P analogue FTY720 enhanced SDF-1 mediated transmigration of both primitive and committed progenitor cells, which could be blocked completely by the addition of CXCR4 blocking antibody or pertussis toxin (PTX) 53. An in vivo study showed that S1P receptor agonist FTY720 could increase both short-term homing and long-term engraftment of HSPCs in the xenogeneic NOD/SCID mouse model 53. Pretreatment with FTY720 caused a rapid and significant increase of more primitive, CD34+/CD38− cells homing to the bone marrow 53. This is possibly due to the expression of S1P1 and S1P5 in the more primitive CD34+/CD38− progenitors, which can bind FTY720 (and FTY720-P) most avidly to stimulate migration (Figure 2C). Intriguingly, FTY720 did not affect the expression of CXCR4 or various integrins. Although FTY720 was shown to induce sustained calcium mobilization and actin reorganization which are critical for cellular locomotion, the molecular details for S1P/S1P receptor signaling in the regulation of HSPC homing remain to be elucidated.</p><p>To analyze the influence of S1P1 on stem cell chemotaxis and trafficking, S1P1 was over-expressed in mobilized CD34+ peripheral blood progenitor cells (PBPCs) 54. The result showed that S1P1 over-expression sensitized the transfected cells to S1P-mediated migration with the most effective dose being around 10 nM, instead of 100 nM in non-transfected control cells. However, incubation with S1P in CD34+ PBPCs over-expressing S1P1 significantly inhibited in vitro SDF-1-dependent migration. In addition, over-expression of S1P1 receptors caused a significant reduction of HSPCs homing potential to the bone marrow and spleen in vivo. S1P1 over-expression caused a significant reduction of surface CXCR4 expression and completely blocked SDF-1-induced ERK1/2 activation and calcium flux. Inhibition of SDF-1 mediated migration through S1P1 over-expression also occurred in Jurkat cells 54. It has been reported that S1P signaling could transactivate CXCR4. S1P and its synthetic analog FTY720 induced phosphorylation of CXCR4 through the S1P3 receptor to improve blood flow recovery and augment revascularization after hind limb ischemia 55. Thus, signaling cascades mediated by different subtype of S1P receptors might have distinct impact on the homing of hematopoietic progenitors through the stimulation or inhibition of the central SDF-1/CXCR4 axis. However, how various physiological and pathological situations regulating the differential expression of S1P receptor subtypes remains to be determined. These data suggest that the expression of S1P receptor subtypes could be used as a strategy for modulating the SDF-1/CXCR4 axis to regulate HSPC homing and engraftment in the hematopoietic microenvironment.</p><!><p>The fact that S1P is a multifunctional mediator released by many different cells during inflammation and injury implies that S1P may act as a direct chemoattractant for HSPCs under specific circumstances such as tissue damage and infection.</p><!><p>Interestingly, lymph-derived HSPCs have been found to express bacteria recognition pattern receptors TLR2 and TLR4 50. It has been reported that GPCRs play a role in Lipopolysaccharides (LPS) signaling and TLR4 signaling has been shown to cooperate with S1P1 and/or S1P3 to increase cytokine production of IL-6 and IL-8 in oral mucosal epithelial cells 56,57. LPS recognition of TLR4 caused HSPC retention within extramedullary tissues, as TLR stimulation blocked HSPC egression from inflamed tissues and abolished HSPC chemotaxis toward S1P both in vitro and in vivo. A possible explanation for this is that LPS/TLR4 signaling might interfere with S1P/ S1P1-regulated signaling. Indeed, the failure of LPS-treated HSPCs to migrate toward the S1P gradient was caused by the down-regulation of S1P1 in response to LPS treatment 50. The danger signal also induced HSPC proliferation and differentiation to rapidly produce large numbers of innate immune cells in response to tissue damage and infection. In vitro, TLR signaling can trigger HSPC proliferation and drive the differentiation of HSPCs into the myeloid lineage. When LPS-incubated HSPCs were implanted underneath the kidney capsule, clusters of local proliferating GFP+ HSPCs expressing myeloid lineage markers were found in kidney 50. Therefore, the circulation of HSPCs not only replenishes tissue residing hematopoietic cells in the absence of infection, but might also act as an immediate and highly adaptive source of progenitor cells that proliferate locally and generate innate immune effector cells to boost innate immunity to fight off life-threatening infections (Figure 3A).</p><!><p>MSCs are a natural regenerative source for damaged tissues in the adult organism, as they can trans-differentiate into various cell types such as myoblasts, hepatocytes, and even neuronal cells 58-61. S1P was found to be the most potent chemoattractant among serum-derived growth factors in inducing MSC mobilization in vitro. S1P is increased after chronic liver injury in both CCl4 treatment and the BDL mice model of liver injury 62,63. The level of S1P in the liver and serum increased significantly by approximately 1.5- and 2-folds, possibly due to the up-regulation of sphingosine kinase expression in liver tissue. The S1P gradient between the damaged liver and the BM is thus established to facilitate the recruitment of MSCs from the BM into circulation and then into the liver. GFP-positive cells of BM origin are positive for the myofibroblast marker α-SMA (α-smooth muscle actin), which is correlated with the progression of liver fibrosis. S1P may mediate the homing of MSCs towards the damaged liver and the differentiation of myofibroblasts to trigger matrix remodeling during acute and chronic liver injury.</p><p>To confirm the effect of S1P on MSC migration, a transwell migration assay was performed. S1P induced the migration of MSCs in a dose-dependent manner 62,63. MSCs express three S1P receptor subtypes: S1P1, S1P2 and S1P3; however, only S1P3 was consistently reported to be markedly up-regulated after liver damage. Nonspecific S1P3 receptor antagonist suramin and specific S1P3 siRNA could significantly inhibit in vitro migration of MSCs in a dose-dependent manner. Suramin also markedly blocked the in vivo liver migration of GFP-positive MSCs in a transplantation experiment. Therefore, S1P/ S1P3 signals might play a critical role in mediating the trafficking of MSCs toward the injured liver both in vitro and in vivo 63 (Figure 3B). Also, it was shown that S1P mediates MSCs migration through important signaling pathways such as cytoskeleton remodeling, disassembly of the FAK (focal adhesion kinase)-Paxillin complex, and linkage of the extracellular matrix to the actin cytoskeleton 63. Moreover, it has been found that RhoA/ ROCK and the catalytic activity of matrix metalloproteinases (MMPs) that is involved in the S1P-induced regulation of ERK activation and Paxillin redistribution and FAK phosphorylation 64. The study of the mechanism underlying MSC migration in response to S1P will help optimize the use of bioengineered MSCs as a potent cellular therapeutic tool.</p><!><p>NPCs are self-renewing stem cells that are important in neurogenesis. Migration of NPCs is important not only for development of the embryonic nervous system, but also for the repair of the nervous system after injury 65. S1P was shown to be a critical mediator for the injury-mediated NPC migration 66,67. It was shown that the neural stem cells of the subventricular zone (SVZ) migrate laterally to sites of brain injury for region-specific neurogenesis. When experimental brain infarction was induced, most glial cells and neurons in the affected brain region were destroyed. Subsequently, microglia and myeloid lineage cells expressing CD11b accumulated 67. S1P was shown to be gradually increased at the site of ischemia at 3 days after insult and peaked at 14 days later. The high S1P level found in the region of microglia accumulation in the infarcted area, suggesting that the local elevation of S1P might be from the release of S1P from the microglia and could be a physiological chemoattractant to enhance the migration of NPCs to induce the subsequent neuroprotective regeneration after a central nervous system injury.</p><p>S1P induced NPC migration maximally at 100 nM and it has been previously shown that S1P1 contributed to NPC migration toward areas of high S1P concentration in the injured central nervous system (CNS) 66. NPCs expressed all known S1P receptor subtypes, with S1P1 and S1P2 being the most highly expressed 67. The S1P1-specific agonist failed to enhance NPC migration in the presence of S1P, suggesting that the activation of S1P1 itself could not overcome the inhibitory effect of S1P2 in NPCs 67. Indeed, specific S1P2 antagonist JTE-013 and short interfering RNA against S1P2 significantly enhanced the migration of NPCs induced by S1P in vitro. In vivo, ventricular infusion of JTE-013 promoted dramatic NPC migration towards the ischemic area where S1P increased. Modulation of S1P2, instead of S1P1, could be a more practical strategy to mobilize NPCs to migrate after a brain ischemia (Figure 3C). However, to rule out the effects of JTE-013 unrelated to S1P2 antagonism, the S1P2 gene-deficient mice would be required to confirm the full effects of S1P2 inhibition in the NPC migration.</p><!><p>In areas of vascular injury, platelet aggregation and activation causes the local release of S1P, which results in high concentrations of S1P as well as SDF-1 to enhance the mobilization of HSPCs to sites of vascular injury for myocardial remodeling. It has been reported that exogenous HDL and its lipid component S1P attenuated the infarction size dramatically and that the level of S1P in the serum is the most reliable marker for predicting the cardiovascular events to follow 68,69 (also see review of Koichi Sato and Fumikazu Okajima in this hot-topics: Role of sphingosine-1-phosphate in anti-atherogenic actions of high-density lipoprotein). S1P-induced recovery of blood flow is due to neovessel formation. The mouse hindlimb model is one of the well-established animal models for ischemia-induced angiogenesis in vivo for evaluating the potential of angiogenic factors as therapeutic agents. The protective effects of HDL and S1P seem to be mediated through S1P3 27,70. S1P may recruit bone marrow-derived circulating endothelial precursor cells (EPCs) to the ischemic tissues 27. EPCs express the S1P3 receptors and stimulation with S1P or FTY720 activates the CXCR4 chemokine receptor which is essential for the EPC mediated angiogenesis 55. Therefore, S1P may stimulate angiogenesis through recruitment of circulating endothelial precursor cells (Figure 3D).</p><!><p>There are increasing lines of evidence which show the importance of the previously unrecognized roles of the S1P gradient in the guidance of stem cell trafficking in homeostasis and stressed conditions. However, some discrepancies exist regarding the effect and optimal dosage of S1P on HSPC migration in vitro. Some studies showed that the physiological 100 nanomolar concentration of S1P was enough to simulate HSPC chemotaxis in vitro while others report that micromolar concentrations of S1P are required. These differences might be due to different isolation methods or the status and populations of cells isolated. As shown by Ratajczack's group, the chemotactic responsiveness of HSPCs to S1P depends on the source of the cells 21. S1P strongly chemoattracted the BM-residing clonogeneic progenitors, but this effect was significantly weaker for those pre-exposed with S1P 21. This finding that the chemotactic responsiveness to S1P is affected by previous exposure may help clear up some previous inconsistent observations. For an instance, S1P agonist had no effect on the spontaneous migration of G-CSF-mobilized human peripheral blood progenitor cells across BMEC cells in vitro, possibly due to the fact that these cells have been exposed to S1P in the peripheral blood before mobilization.</p><p>S1P's function in vivo seems complex as it is related to both HSPC homing and mobilization. The S1P gradient established between bone parachyma and circulation seems to attract stem cells towards the circulation, as indicated by the significant spontaneous HSPC mobilization after AMD3100 blockage or using G-CSF and zymason. Therefore, to stay in specific stem cell niches, stem cells first must overcome this effect which is mediated by various retention signals. The requirement for S1P in the efficient egression of HSPCs from the BM parenchyma into the sinusoids reflects a role for S1P gradient, possibly through S1P1, in spontaneously attracting HSPCs after the retention signaling is overcome. The specific function of S1P in HSPCs traveling to and from the BM are poorly understood, as it still needs to be dissected in the various stages, such as passing in and out of the trans-endothelial barrier. In addition, the differential stages of HSPCs might employ distinct S1P receptor subtype to regulate retention or chemotaxis. There also might be a threshold regulation on the expression level of S1P receptor subtypes in HSPCs in a different tissue microenvironment which helps regulates cellular functions. For example, diminished S1P1 by TLR signaling retains HSPCs to certain extramedullary tissues to differentiate into specific population of cells to fight off the infection 50.</p><p>In stressed conditions, S1P as well as SDF-1 released from damaged local tissues might re-establish a new S1P gradient so that injured tissue might attract progenitor cells from circulation. This could be achieved by the up-regulation of sphingosine kinase which would generate more S1P. Alternatively, S1P lyase in tissue could be inactivated to maintain the high S1P concentration compared to the concentration in circulation. If this is the case, S1P may be able to direct chemotaxis to foster the local renewal of damaged cells or the production of tissue-residing innate immune cells in response to stress. It is also possible that under stress conditions, S1P may have a synergic effect with other chemokines to regulate HSPC egression from the BM to migrate to the injured tissues. It was found that the migratory ability of the less motile primitive hematopoietic cells were greatly enhanced (12 fold) when SDF-1 was combined with LPA and S1P 25. Therefore, S1P may play a role in stressed conditions where there is high cytokine production to function with other signals in mobilizing hematopoietic cells. Cytokines such as IL-8 are known to mobilize stem cells into the peripheral blood 67,71. During stressed conditions, bone marrow residing cells, such as osteoclasts, may release lipid mediators like S1P locally to trigger stem cell mobilization from specific niches towards the injury.</p><p>The SDF-1/CXCR4 axis is the key retention signal in HSPC anchorage in the bone marrow stem cell niches, and the expression of CXCR4 has been shown to be correlated with success of recovery from bone marrow transplantation. Modulation of the SDF-1/CXCR4 axis through S1P signaling may have therapeutic usage in bone marrow transplantation. It has been shown that CXCR4 inhibition can block FTY720 mediated lymphocyte homing and FTY720 inhibited endothelial cell sprouting 72. Furthermore, S1P3 could transactivate CXCR4 and overexpression of S1P1 reduced CXCR4 expression. Therefore, S1P family of receptors may play an important role in the regulation of the SDF-1/CXCR4-mediated HSPC lodgment to the stem cell niches. Kimura et al. showed that preincubation of CD34+ PBPCs with FTY720 increased SDF-1-dependent in vitro transendothelial migration and in vivo stem cell homing after transplantation 53. Because S1P1 overexpression strongly inhibits expression of CXCR4, CXCR-4 mediated signaling and chemotaxis in human CD34+ PBPCs, FTY720 might be beneficial for bone marrow transplantation as FTY720 treatment leads to the internalization and degradation of S1P1 receptors 73. Moreover, S1P chemical agonists/antagonists can activate or antagonize with various S1P receptor subtypes, and have been widely used to study the specific S1P receptor subtypes involved in cell trafficking. However, caution must be taken when analyzing the effects of S1P agonists/antagonists in a physiologically relevant condition, as small molecules may cause S1P receptors to transduce inappropriate intracellular signaling, and mask the true functions mediated by S1P receptors. Also, undesired non-specific effects are frequently associated with the utilization of pharmacological reagents. Nevertheless, it remains to be determined how FTY720 affects these S1P receptors, from full agonism to functional antagonism, and which receptors are affected. Another issue needed to be noted is that S1P may affect many different steps of HSPC recirculation. For example, FTY720 could induce the disappearance of HSPCs from blood due to the inhibition of HSPC recirculation from the extramedullary tissues 50. Although S1P and its receptors may represent one of several mechanisms that modulate CXCR4-dependent migration in vivo, it should be noted that modulation of S1P signaling in stem cell transplantation might interfere with stem cell activity and long-term engraftment.</p><p>It has been reported that the migration of malignant cells underneath the stromal layer depends on the SDF-1/CXCR4 system. Consistent with the notion that S1P increases the migratory capacity of murine and human HSPCs 24,25,74,75, it is very likely that cancer stem cells or cancer initiating cells adopt the same strategy using S1P signaling to transactivate SDF-1/CXCR4 to facilitate their growth and migration. Indeed, S1P has been found at high concentrations in many tumors microenvironments. In addition, locally produced S1P by cancer cells could recruit the MSCs with multi-lineage differentiation potential to promote cancer growth and invasion 76,77. Moreover, S1P and its receptor S1P1 are essential for the recruitment of pericytes and smooth muscle cells to the nascent capillaries, and thus facilitate angiogenesis for building the tumor's blood supply. Therefore, it has become an urgent need to study how S1P signaling regulates stem cell trafficking and this will provide many insights into the cancer stem cell biology.</p><!><p>Mounting evidence suggests that S1P plays a critical role in stem cell development and maintenance. The retaining of primitive cells in the stem cell niches and the release of more mature stem cell into circulation is in part regulated by S1P. Under normal circumstances, the primitive c-Kit+ cells are less motile than the committed Kit− cells, and respond less to SDF-1, LPA or S1P in order to keep the primitive hematopoietic cells dormant. This is supported by an in vitro chemotactic experiment, which showed that S1P does not regulate immature hematopoietic cell migration during steady-state hematopoiesis 24. More light could be shed on this process if the regulation of the expression of S1P receptor subtypes during HSPC development is studied. Certain transcriptional factors are known to regulate the expression of S1P receptors in a different environmental milieu, such as T-bet regulated NK maturation and egression from the BM 35. How and which transcription factors regulate the distinct S1P receptor subtypes in HSPC differentiation and egression from BM remain to be explored.</p><p>During HSPC circulation in the peripheral blood stream, while some organs can actively recruit HSPCs, other certain organs lack these circulating HSPCs. For example, donor-derived CFUs are rarely found in the brain. This might be due to the secured blood brain barrier. Perhaps only locally differentiated neural stem cells have the ability to replace the damaged cells. It would be important to investigate the different strategies employed by the different organs in the recruitment of stem cells. S1P1 was first identified in endothelial cells and is critical to endothelial cell function 2,3,26,27. S1P may alter the permissiveness of the endothelium at the egression sites. This is supported by a recent finding that bone marrow progenitor cells could enhance endothelial adherens junction integrity by paracrine S1P release and the following Rac1 and Cdc42 signaling 78. To precisely dissect the roles of S1P receptor subtypes at the site of egression, distinct S1P receptors must be selectively knocked out or blocked in the endothelial cells of the bone marrow.</p><p>S1P regulates a wide array of biological activities and physiological functions by functioning as an extracellular ligand or intracellular mediator. In this review, we have summarized recent evidence which strongly suggests the notion that S1P gradient plays a critical role in the mobilization, trafficking and homing of HSPCs. This notion needs to be further confirmed by utilizing sphingosine kinases, rate-limiting enzymes for S1P synthesis, null mice. There are two isoforms of sphingosine kinases, SphK1 and SphK2. No phenotypic alterations have been observed in either SphK1 or SphK2 knockout mice, and SphK1/SphK2 double null mice are embryonic lethal (see reviews: "The Role of Sphingosine Kinase Localization in Sphingolipid Signaling" and "Regulation of Cancer Cell Migration and Invasion by Sphingosine-1-Phosphate" in this hot-topics). Thus, the development of the tet-on/tet-off SphK1/SphK2 conditional double null mice is a need to precisely determine the role of S1P in HSPC mobilization. Also, most of the knowledge for the role of S1P receptor subtypes in stem cell mobilization was obtained by employing pharmacological agonists/antagonists. Similarly, more detailed studies utilizing knockout mice of S1P receptor subtypes are needed. Finally, S1P has been shown to be an important intracellular mediator. Several intracellular targets of S1P have recently been identified, including HDAC 4, TRAP2 5, and Prohibitin 2 6. Therefore, the determination of the intracellular roles of S1P in stem cell mobilization remains to be further explored in the future.</p>

PubMed Author Manuscript

Metal-Free Reductive Cleavage of C–N and S–N Bonds by Photoactivated Electron Transfer from a Neutral Organic Donor**

A photoactivated neutral organic super electron donor cleaves challenging arenesulfonamides derived from dialkylamines at room temperature. It also cleaves a) ArC–NR and b) ArN–C bonds. This study also highlights the assistance given to these cleavage reactions by the groups attached to N in (a) and to C in (b), by lowering LUMO energies and by stabilizing the products of fragmentation.

metal-free_reductive_cleavage_of_c–n_and_s–n_bonds_by_photoactivated_electron_transfer_from_a_neutra

<!>Supporting Information<!>

<p>Recently, we have developed a range of highly reactive organic electron donors, including 1–3 (Scheme 1). These compounds undergo oxidation to radical cations and dications after loss of one and two electrons, respectively, and the aromaticity of these products contributes to the driving force for the oxidations.1 The radical cation 4 and dication 5 derived from 3 are shown in Scheme 1. The donor 1 was the first neutral organic compound to reductively cleave aryl iodides to aryl radicals,1a while the stronger donors 21b and 31d converted aryl iodides into aryl anions. The compounds 2 and 3 also reduced Weinreb amides,1f acyloin derivatives,1h and some sulfones.1c</p><p>Protected amines and neutral organic electron donors. DMF=N,N-dimethylformamide, Ts=4-toluenesulfonyl.</p><p>The donor 2 also cleaved arenesulfonamides, but not N,N-dialkyl arenesulfonamides.1c For the sulfonamides 6, 8, and 10, the electron is transferred from the donor to the arenesulfonyl unit, where the LUMO is located. The substrates 6 and 8 underwent efficient reductive cleavage of the N–S bond to afford the corresponding amines 7 and 9 [using donor 2 (6 equiv), DMF, 110 °C, 18 h]. In these cases, the nitrogen leaving groups are stabilized by resonance. However, the sulfonamide 10, which, after fragmentation, would produce a nitrogen-centred leaving group which is unstabilized by resonance, remained completely unchanged.</p><p>Most recently, our donors 2 and 3, vivid yellow and purple solids, respectively, were tested under photoactivation conditions and proved even more powerful than in the ground state, in that they were now able to reductively cleave the Ar–Cl bond in chlorobenzenes, a reaction which had never been seen with our ground-state electron donors.1q In addition, they were able to donate an electron to the cis-diphenylcyclopropane 11, ultimately affording the 1,3-diphenylpropane 12 as a product.1q</p><p>These advances encouraged us to test the photoactivated 3 in other very challenging transformations, that is, the reductive cleavage of 1) difficult arenesulfonamides like 10,2 and 2) N-benzyl groups,3 and we report herein our results. The donor 3 was selected since it is as strong as 2, but is much more conveniently prepared.</p><p>The sulfonamides 10, 14, and 16 were chosen as substrates for reduction (Scheme 2). Fragmentation of their radical anions would give rise to a nitrogen-centred leaving group which would not be stabilized by resonance. Under photoactivated conditions (λ=365 nm, 2×100 W) at room temperature, each of the substrates underwent cleavage to afford the parent amine in good yield after work-up. The λ=365 nm irradiation does not overlap with the chromophore of the sulfonamides, and hence it is the highly colored 3 which undergoes photoactivation. This result marks the first time that dialkyl arenesulfonamides have been reductively cleaved by a neutral organic electron donor.</p><p>Reactions of sulfonamides with the photoactivated donor 3.</p><p>To verify the nature of the activation, two types of blank experiments were also conducted for the substrates 10, 14, and 16. These blank reactions were conducted a) without 3, but in the presence of photoactivation, and b) with 3, but in the absence of photoactivation. In all cases, no product was formed and the starting substrate was recovered in excellent yield (see the Supporting Information). This reinforced the message that photoactivation of the donor (or the donor–substrate complex; see the Supporting Information), significantly enhances the driving force for electron transfer.</p><p>For completeness, the more reactive substrates, 6 and 8, were also treated with 3 under photoactivated conditions (Scheme 2). They underwent efficient cleavage, as expected, with the products 7 and 9, respectively, being isolated in excellent yield after shorter reaction times in the presence of three equivalents of 3.</p><p>For the radical anion 19, formed at the sulfonamide group in 10 after electron transfer (Scheme 2), we calculated whether fragmentation to either the dialkylamide anion 22 and a sulfonyl radical 23, or the dialkylaminyl radical 20 and a sulfinate anion 21 would be preferred.4 Density functional theory5, 6 (DFT) calculations were employed to optimize all structures with the gradient corrected B97-D functional with a long-range dispersion correction.7 All atoms were described with the 6-311++G(d,p) basis set.8, 9 Subsequent single-point energy calculation of the optimized geometry was performed at the same level of theory within a polarizable continuum model (CPCM)10 with the dielectric constant of N,N-dimethylformamide (DMF, ε=37.219; see the Supporting Information for details). These studies indicated that fragmentation of the radical anion 19 into 20 and 21 is thermodynamically preferred, both when 19 alone is considered as an entity, and also when a complex between the radical cation of the donor and 19 is considered. Experimental evidence in favor of the formation of an aminyl radical following fragmentation was also seen when substrate 24 was subjected to cleavage. Here fragmentation into the toluenesulfinyl radical 23 and dialkylamide anion 28 should lead to the isolation of the corresponding amine, N-cyclopropyl-N-tetradecylamine, upon work-up. In contrast, cleavage to 21 and the aminyl radical 29 should lead to rapid opening of the cyclopropane ring to afford the imine 30, which could undergo quenching of its radical in a number of ways1k to form the imine product 31, and thus undergo hydrolysis under mild work-up conditions to afford the tetradecylamine 26. In fact, the experiment afforded 26 (85 %) and 25 (78 %), thus supporting 29 as an intermediate.</p><p>The next task was to investigate whether N-benzyl groups could be cleaved. The outcome from the substrate 8 deserves comment because it shows no cleavage of the benzyl group. This is in accord with expectations in this case, as the LUMO is localized on the electron-poor p-toluenesulfonyl (tosyl) group, rather than on the benzyl group. When the S–N bond is cleaved, the aniline anion 18 results, and the arene ring is too electron-rich to receive another electron. Either during the reaction or upon work-up, 18 undergoes protonation to 9.</p><p>To modify the structure of 8 so that cleavage of an N-benzyl bond might occur, we designed substrates such that the benzyl group was the site of the LUMO within the substrate. To this end, benzyl alkyl methanesulfonamide derivatives (33 a–i) were chosen (Table 1). DFT studies (B3LYP / 6-31G*), taking 33 c and 33 g as examples, in a DMF solvent continuum indicated that the LUMO of these compounds was located on the benzyl group. If electron transfer occurred to the benzyl group, then the leaving group would likely be the sulfonamide anion, and this would be protonated to form 34 upon work-up. Upon trying the reactions, very good yields of N-benzyl bond cleavage were seen in each case (Table 1). The substrate 33 e features a benzyl group and a dimethoxybenzyl group. The outcome shows competitive cleavage of these two benzyl groups, with marginal selectivity for the formation of 34 e*, which is consistent with a very slightly preferential electron transfer to the less electron-rich aryl ring, that is, the C6H5 ring. To show that photoactivation was required for these reactions, 33 g was subjected to a parallel reaction in which photoactivation was omitted. This reaction afforded an excellent recovery of the unchanged 33 g (94 %).</p><p>Reductive deprotection of benzyl methanesulfonamides (33) with 3</p><p>[a] Recovered starting material. [b] Yield of isolated product. Ms=methanesulfonyl.</p><p>Since benzyl methanesulfonamides had worked so well we next investigated the more challenging allyl methanesulfonamides. Because these compounds have less extensive π systems, their LUMO energies are expected to be higher than their benzyl counterparts. In support of this, the mixed allyl benzyl substrate 35 a showed selectivity for the benzyl cleavage to 37 (62 %; Table 2). This outcome was in line with expectations since the LUMO of this substrate (and the SOMO of its radical anion) were associated with the arene ring, rather than with the allyl group or with the sulfonyl group. However, the presence of some product resulting from allyl cleavage, that is, 36 a (10 %), encouraged us to think that substrates which did not feature an N-benzyl group might undergo cleavage of the allyl group. This selectivity in 35 a for benzyl cleavage over allyl cleavage contrasts with that seen in palladium-induced reduction of benzyl allylamines where the affinity of Pd0 for olefins dominates the reactivity.11 It also surprisingly contrasts with the selectivity in favor of deallylation seen in the reductive deprotection of sugars with SmI2 reported by Hilmersson et al.12</p><p>Reductive deprotection of allyl methanesulfonamides 35</p><p>[a] Recovered starting material. [b] Yield of isolated product. We recognize that 36 a=34 e, 36 d=34 d, and 36 e=34 b. [c] When additional donor 3 (6 equiv) was added after 72 h, and the reaction continued for a further 72 h, 36 d (81 %) was isolated.</p><p>When the allyl alkyl methanesulfonamides 35 b–e were treated under photoactivation conditions with 3, cleavage of the allyl group was exclusively seen, with moderate to good yields of the products 36 being isolated. Taking the substrates 35 d and 35 e as examples, the LUMO of the substrates (and the SOMO of their radical anions) are sited on the allyl groups, thus allowing the selectivity of the observed reactions to be easily understood. In 35 b and 35 c, the LUMO lies on the aryl ring, however, the radical anion shows spontaneous cleavage of the allyl group. In this case, electron transfer to the arene should occur preferentially. There is no driving force for fragmentation of the arene radical anion in these two cases, since that would give an alkyl leaving group unstabilized by resonance, so intramolecular electron transfer to the allyl group can occur, thus leading to the observed fragmentation. To explore whether photoactivation was needed to trigger these reactions, the substrate 35 b was subjected to the same reaction conditions, except that no photoactivation was provided. In this case, no deprotection occurred and 35 b was recovered in quantitative yield.</p><p>This ability to transfer an electron to an N-allylsulfonamide takes the photoactivated electron donors into new territory, as no previous deallylation reaction has been reported. To check if the allyl group was really needed, or if N,N-dialkyl methanesulfonamides would undergo reaction by electron transfer to the sulfonyl group, the N,N-dioctyl methanesulfonamide 38 was subjected to reaction with the photoactivated 3. In this case, no new product was detected and the starting material 38 (92 %) was recovered unchanged.</p><p>The ability to transfer an electron to an ArC–N ring group is evident in the above results with the substrates 33 a–i, and this led us to investigate what happens in the transposed case, that is, ArN–C. An amine nitrogen atom directly attached to the arene should make electron transfer to the arene more difficult, but the accessibility of the LUMO for electron transfer should depend upon the third group attached to the nitrogen atom. With the simple N-methyl-N-allylaniline 39 a, very little cleavage occurred, but the product that was isolated, 40 a (6 %), showed cleavage of the N-allyl bond (Table 3). To better facilitate the cleavage reaction, the N-Me group was replaced by an N-acyl group. The electron-withdrawing acyl group can lower the LUMO energy and hence make electron transfer to the LUMO easier. In the event, protection of the nitrogen atom as an acetamide (39 c), a pivalamide (39 d), and a urethane (39 e), all enhanced the cleavage of the allyl group.13 A blank experiment was also conducted on 39 c (in the absence of photoactivation) and this showed no conversion into the product, but rather quantitative recovery of 39 c, thus illustrating the essential role of the photoactivation of the donor. The pivalamide was most successful, thus affording the product 40 d in 83 % yield. The significant difference in efficiency between 39 c and 39 d led us to investigate whether deprotonation of the acetyl group by the basic donors might be occurring. When a repeat of the experiment with 39 c was subjected to addition to D2O, as opposed to H2O, prior to acidification and extraction, both the product 40 c and the recovered starting material 39 c showed incorporation of a single deuterium atom by mass spectrometry. For further thoughts on the role of deprotonation, see discussion of reactivity of substrate 44.</p><p>Reductive deprotection of allylanilines with electron donor 3</p><p>[a] Recovered starting material. [b] Yield of isolated product.</p><p>Since the ease of bond cleavage in the radical anion seems to correlate with the stabilization given to the radical and anion products, then replacing the alkene of the allyl group in 39 by a carbonyl group, as in 41, (Table 4) might additionally facilitate the cleavage reactions, since the anionic leaving group would be an enolate, in place of an allyl anion. Accordingly, the substrates 41 a–c were prepared. Encouragingly, the N-methyl substrate 41 a showed a higher yield of cleavage product [here 40 a (34 %)] than had been seen for the corresponding allyl case, 39 a (6 %). The N-acetyl and the N-carbethoxy cases, 41 b and 41 c, respectively, underwent very efficient reaction (74 % and 92 % yield of products respectively) with loss of the CH2CO2Et side chain. This outcome shows that ArN–C bonds are also subject to reductive cleavage, and that the efficiency of cleavage correlates with stabilization of the radical and anion produced. A repeat reaction was carried out for 41 b, but in the absence of photoactivation. This reaction gave no 40 c, but gave recovered starting material (41 b, 91 %). A final example of this series, 41 d, was reacted to give a comparison with other sulfonamide substrates reported herein, and this afforded 42 (89 %), the expected product of fragmentation of the arenesulfonamide radical anion.</p><p>Reductive deprotection of N-(acylmethyl)anilines with electron donor 3</p><p>[a] Recovered starting material. [b] Yield of isolated product.</p><p>To verify the importance of the aryl group, the substrate 43 was next prepared (Scheme 3). If electron transfer to the ester group occurred, then cleavage of the N–CH2CO2Et bond might have been expected, but none was seen. Accordingly, the N-aryl group is crucial for the N–C cleavage to occur.</p><p>Substrates for ArN–C cleavage.</p><p>Finally, we prepared the modified ArN–C substrate 44 where cleavage of the ArN–C bond at the radical anion stage (46) would leave the radical and anion tethered together in 47 (Scheme 3). In this case, an intriguing rearrangement of the pyrrolidine into a piperidone product 45 (30 %) occurred. Efforts to improve the conversion by adding more equivalents of 3 were not successful, and this is consistent with the representation in Scheme 3. The initial radical anion 46 undergoes fragmentation to 47. In the presence of excess 3, further reduction to the amidyl anion 48 should occur rapidly. The dianion 48 is unlikely to cyclize, but cyclization could occur after proton transfer from another molecule of 44, thereby forming the enolate 49 which will not undergo any reduction. Finally, cyclization of the anion 50 would afford the piperidone 45. If this proposal is correct, it would also be relevant for the closest analogue of 44, that is, 41 a. The lower yield in these two substrates could therefore be explained both by this proton transfer from substrate and by the inherent difficulty of electron transfer to an N,N-dialkylaniline.</p><p>To conclude, electron transfer from the photoactivated neutral electron donor 3 delivers high yields of S–N and C–N cleavage products for a range of nitrogen-containing species including anilines, sulfonamides, and amides. These reactions proceed at room temperature and under mild reaction conditions in the absence of any metal reagents, thus illustrating challenging reactions which can be achieved by photoactivated neutral organic electron donors.</p><!><p>As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.</p><!><p>miscellaneous_information</p>

PubMed Open Access

Exploiting the versatile alkyne-based chemistry for expanding the applications of a stable triphenylmethyl organic radical on surfaces

The incorporation of terminal alkynes into the chemical structure of persistent organic perchlorotriphenylmethyl (PTM) radicals provides new chemical tools to expand their potential applications. In this work, this is demonstrated by the chemical functionalization of two types of substrates, hydrogenated SiO 2 -free silicon (Si-H) and gold, and, by exploiting the click chemistry, scarcely used with organic radicals, to synthesise multifunctional systems. On one hand, the one-step functionalization of Si-H allows a light-triggered capacitance switch to be successfully achieved under electrochemical conditions. On the other hand, the click reaction between the alkyne-terminated PTM radical and a ferrocene azide derivative, used here as a model azide system, leads to a multistate electrochemical switch. The successful post-surface modification makes the self-assembled monolayers reported here an appealing platform to synthesise multifunctional systems grafted on surfaces.

exploiting_the_versatile_alkyne-based_chemistry_for_expanding_the_applications_of_a_stable_triphenyl

Introduction<!>Functionalization of Si-H surfaces with PTM radicals<!>À2<!>Functionalization of PTM radical molecules via click chemistry, in solution and on the gold surface<!>Charge transport measurements across PTM monolayers bound to gold<!>Conclusions<!>Conflicts of interest

<p>The impact of (bio)molecular surface modication in different elds of research is supported by the wide variety of applications that derive from it. The goal of the functionalization can be either transferring specic molecular properties to the substrate, i.e. by graing (multi)functional molecular systems, or generating novel interfaces, for example to promote cell-adhesion, (bio)sensing or to inhibit undesired reactions, among others. [1][2][3][4][5][6][7] The molecular modication of surfaces has also been a determining factor in the emerging eld of molecular (spin)electronic devices. 8 In this context, among the extensive library of functional molecules investigated, optically-, redox-and magnetically-active stable organic radicals are very attractive. 9 Nowadays, these materials have attracted much interest and their exploitation as components of molecular functional materials is boosting a new generation of devices for applications in OLEDs, 10,11 energy storage and conversion, 12,13 molecular spintronics, 14 imaging, 15,16 sensors 17 and memory devices. [18][19][20] It has been shown that for the device implementation, in many cases their immobilization onto surfaces is needed. 21 To this end, the formation of stable self-assembled monolayers (SAMs) on substrates typically employed for the fabrication of electronic components, such as silicon or gold, is a wellestablished strategy. 22,23 The hybrid system robustness, and hence its potential applicability, is equally dependent on the stability of the molecular entity as well as on the stability of the molecule-substrate interface. For this, beyond the search for transferring a specic molecular functionality to the surface (mainly magnetically-, redox-or optically-active), great efforts have been focused on: (i) the search for new robust anchoring groups and (ii) the post-graing modication of SAMs. Terminal alkynes have appeared as an appealing alternative to the thiol group as they spontaneously and covalently react with gold. 24,25 Besides the stability of the C-Au bond, the high directionality of the s-bonded C-Au has been demonstrated to provide higher conductance at the single-molecule level, 26 which makes terminal alkynes very attractive for molecular electronics. Interestingly, these functional groups can also be used to covalently react with hydrogen-terminated silicon (Si-H) surfaces. [27][28][29] Concerning post-graing modication, preparing monolayers with exposed functional end groups which can react with other (bio)molecular systems is of great importance to expand the applicability of these surfaces. Among the several chemical strategies reported to perform interfacial reactions on SAMs, 30 the so-called "click chemistry" has been one of the most employed, in particular the 1,3-dipolar cycloaddition of azides and alkynes to form 1,2,3 triazoles. [31][32][33][34] Hence, alkynes can be considered as very versatile groups to be employed for the modication and post-modication of substrates.</p><p>The functionalization of SiO 2 -free silicon and gold substrates with electrical and/or light triggered molecules has been pursued with the aim of obtaining switchable surfaces. In the case of silicon, unlike metals, its electronic properties can be nely tuned by modifying the density and the nature of the charge carriers (electrons and holes) under light illumination, which can be used as a second gate for the tuning of the properties of the modied surface. 29 In this work, by exploiting the rich chemistry offered by alkynes and the high chemical and thermal stability of perchlorotriphenylmethyl (PTM) radicals, 35 the redox properties of a PTM radical bearing one and two terminal alkyne groups (Fig. 1) have been dually exploited as: (i) a capacitance switch on Si-H triggered by light and, (ii) as an organic radical-based platform to be further modied by click chemistry giving rise to a multistate electrochemical switch. In both cases, a particular focus has been placed on the optimization of the experimental conditions to ensure the integrity of the radical and the terminal alkyne.</p><!><p>Interfacing technologically important semiconducting surfaces, such as silicon, with high-quality, and stable redox-active lms has appeared as a promising strategy toward functional devices for charge storage. Remarkably, by taking advantage of the fact that the charge transfer characteristics of silicon can be either inhibited or activated upon light illumination, this substrate constitutes a relevant platform for the development of photochemically switchable systems. 29,36,37 Then, silicon itself can be used as a light-controlled gate to turn ON/OFF the electronic communication with the graed electroactive centers, in this case the PTM radicals. With this aim, we exploit herein the reactivity of the alkyne terminated PTM (1-Rad) to chemically modify SiO 2 -free p-type Si-H substrates, forming robust Si-C]C bonds. To date, the functionalization of Si-H surfaces has been achieved with numerous electroactive molecules, among them are ferrocene, quinones, tetrathiafulvalenes (TTF) and metallic complexes. 29 However, to the best of our knowledge, before this work, only one example of an organic radical graed on Si-H has been reported. In that case, the graing was done through a two-step approach, to avoid the reactivity of the unpaired electron with the silicon surface bonds, thus leading to a loss of the radical character. 38 Herein, the Si-H functionalization was carried out in a single step by the hydrosilylation reaction of molecule 1-Rad through the terminal acetylene group leading to the covalently bound PTMterminated monolayer (SAM-1-Rad-Si) (Fig. 2a).</p><p>Although the hydrosilylation reaction between alkynes and Si-H can be performed by either thermal 39,40 or photochemical 41 activation, the instability in solution under light of the radical PTM moiety excluded the possibility of using in this work such a photochemical route. Several graing attempts were carried out by varying the concentration of the PTM radical solution, solvent, temperature and immersion time (see the experimental details in the ESI and Table S1 †). PTM radical monolayers with the highest surface coverage of PTM (estimated from the electrochemical measurements, vide infra) and the lowest oxidation level of underlying silicon were obtained at 145 C for 20 h using 1,2-dichlorobenzene (DCB) as a high-boiling solvent and a ca. 7-10 mM concentration of 1-Rad.</p><p>The SAM-1-Rad-Si monolayer was characterized by X-ray photoelectron spectroscopy, which, additionally to the chemical composition, provides information about the oxidation state of the underlying silicon (Fig. S1 †). For the C 1s, two main peaks at, 285.0 and 286.6 eV, assigned to the C]C and the C-Cl bonds, respectively, were observed. The Cl 2p spectrum displayed a typical doublet with both components at 200.8 (2p 3/2 ) and 202.5 (2p 1/2 ) eV, arising from the chlorinated phenyl rings of the PTM unit. Besides, the Si 2p spectrum was deconvoluted into two contributions at 99.2 and 99.8 eV assigned to bulk and interfacial silicon, respectively, in crystalline Si(111). 42,43 Additionally, a low intense peak appeared at 103 eV, which is attributed to the unavoidable oxidation of a certain content of remaining Si-H sites by water and atmospheric oxygen, indicating that, as expected, the surface is not fully passivated by the bulky radical 1-Rad. Nevertheless, the low contribution of this peak was positively surprising considering the bulkiness of the PTM moiety.</p><p>SAM-1-Rad-Si was electrochemically characterized in the dark and under illumination through a red lter (l > 600 nm) to avoid the possible degradation of the graed PTM radical. As can be seen in Fig. 2b, the cyclic voltammograms (CVs) in the dark showed negligible oxidation and reduction currents (lower than 10 mA cm À2 at 10 V s À1 ), as expected for a semiconductor under depletion conditions, 44 i.e., when only a few majority charge carriers (holes in the case of p-type Si) are available for charge transfer (vide infra).</p><p>Upon illumination of SAM-1-Rad-Si, an intense reversible redox wave was observed at E 0 ¼ 0.21 V vs. Ag/AgCl, KCl 3 M (average of the anodic E pa and cathodic E pc peak potentials) corresponding to the PTM(radical) 4 PTM(anion) process promoted by captured photogenerated electrons (Fig. 2b). The redox response observed indicates that the radical character of the PTM remains unaltered upon graing. Additionally, both the anodic and cathodic peak photocurrents I pa and I pc , corresponding to the reversible redox couple, were found to be proportional to the potential scan rate v, as expected for a surface-conned reversible redox species (Fig. S2a-c †). 45 The variation of E pa and E pc with v (Fig. S2b †) enabled us to determine the apparent rate constant for electron transfer at the bound PTM center, k et,ap , using the recent theoretical model developed by Vogel et al. accounting for semiconductor diode effects. 46 A value of 90 AE 20 s À1 was estimated, in accordance with the literature data reported for other silicon electrodes modied with electrochemically reversible systems. 29,46 The surface coverage of attached PTM moieties was electrochemically estimated from CVs of illuminated SAM-1-Rad-Si (eqn S1 †). Indeed, anodic charge integration at several scan rates (between 0.4 and 1 V s À1 ) resulted in an average value of (8.5 AE 0.3) Â 10 À11 mol cm À2 , very close to the analogous SAMs on Au (see Section 2.2.2).</p><p>To obtain further insights into the light dependence of the redox process in SAM-1-Rad-Si, differential capacitance measurements were performed in the same electrolytic medium (i.e. CH 3 CN/Bu 4 NClO 4 ). First, the atband potential (E  ), i.e. the electrode potential for which there is no space-charge region in the semiconductor, was estimated from the commonly used Mott-Schottky plot (C sc À2 vs. E, eqn S(2) †) that gives the spacecharge capacitance C sc as a function of the electrode potential E under depletion conditions, i.e., the depletion of valence band holes in the space charge region of the p-type surface. 47 In the dark, a linear C sc</p><!><p>-E plot was obtained for potentials below 0.25 V vs. Ag/AgCl, KCl 3 M, in which the intercept and the slope of the curve enable E  and the dopant density N D to be determined, respectively (Fig. S3 †). The calculated N D value (1.2 Â 10 15 boron atoms cm À3 ) was consistent with the dopant density derived from the four-probe resistivity measurements of silicon samples, between 5 and 10 U cm. The extracted value of E  was 0.25 AE 0.02 V. Based on this parameter, it can be concluded that the PTM(radical) 4 PTM(anion) redox process occurs in a potential range wherein the semiconductor is in weak depletion. The small potential difference between E 0 and E  (around 50 mV) explains, however, why low (but not zero) oxidation currents are observed at SAM-1-Rad-Si in the dark (inset in Fig. 2b). Much higher currents were observed under illumination because the redox process could now occur with high rate thanks to the high concentration of photogenerated minority charge carriers (i.e. electrons).</p><p>In the dark, consistent with the current response, the measured capacitance values were small and did not exceed 1 mF cm À2 . In contrast, the capacitance curve under red light illumination was characterized by a much more intense capacitance peak (enhancement by a factor of $10) at 0.18 V, close to the formal potential of bound PTM determined by CV (Fig. 3a). This capacitance peak was clearly attributed to the charging/discharging currents associated with the oxidation/ reduction of the bound PTM centers, 48,49 in perfect line with previous reports on ferrocene-modied Si surfaces. 36,50 This signicant contrast between the two states (ON and OFF) permitted us to exploit this system as a capacitance switch and its cyclability was investigated by carrying out consecutive ON/OFF switching cycles by turning on and off the light along time (Fig. 3b). A $30% decrease in the maximum photocapacitance was observed along the rst 30 minutes before reaching a situation of higher stability. This loss in the switching ability is believed to be caused by the gradual degradation of the electrical properties of the interface due to the oxidation of the underlying silicon. Indeed, owing to the moderately dense packing of the PTM monolayer (due to the bulky nature of the PTM head groups), traces of water and/or oxidizing species present in the electrolytic medium unavoidably penetrate through the molecular layer via defects or pinholes to react with remaining Si-H sites. Despite this, the functionalization was found to be remarkably stable.</p><!><p>In solution. The click chemistry is used in different elds of research, such as biomedical science, chemistry and materials science. 51,52 Remarkably, the functionalization of stable organic free radicals via this reaction has been only marginally explored. This is basically attributed to the low stability of the radical character under the most standard reductive "click" experimental conditions. In this work, we report for the rst time the experimental conditions to achieve a click reaction between an alkyne-terminated PTM radical and an azide derivative (Scheme 1). The experimental conditions were rst optimized in solution and then were used to engineer a radical-PTM-based SAM which acted as a platform to elaborate more complex multifunctional surfaces. In this work, the focus was placed on the reaction between the organic radical (2-Rad, Fig. 1), previously synthesized, 26 and an azidomethyl ferrocene (Fc-N 3 ) derivative leading to a donor-acceptor (D-A) SAM.</p><p>For the synthesis of the target radical PTM-Fc dyad (3-Rad) a copper-catalyzed azide-alkyne cycloaddition (CuAAC) click chemistry strategy was envisaged. [53][54][55] The standard conditions for the click reaction that generates in situ the copper(I) catalyst by reduction of CuSO 4 using ascorbic acid or ascorbate 56 could not be used here because the ascorbic acid reduces the radical to the anion that in turn is irreversibly protonated to the aH form. 57 Therefore, as mentioned above, before performing the reaction on the surface, the synthetic procedure was optimized in solution. The successful synthesis of the target molecule 3-Rad with the highest reaction yield was obtained by using copper(I) iodide as a catalyst, and N-ethyldiisopropylamine (DIPEA) and triphenylphosphine as ligands (see ESI † for further details on the synthesis). As shown in Scheme 1, two routes were followed. First, the CuAAC reaction conditions starting with the non-radical precursor, aH-PTM molecules (2-H), were optimized. Both mono-and di-cycloaddition compounds, 3-H and 4-H, were obtained as products of the click reaction between the azidomethyl ferrocene and the PTM derivative with two acetylene-terminated groups (2-H) even when a 1 : 1 molar ratio of the azide and 2-H was used. Next, the corresponding radicals 3-Rad and 4-Rad were prepared through treatment of the aH adducts, 3-H and 4-H, with tetra-n-butylammonium hydroxide (TBAOH), to generate the corresponding anions, and then the oxidation of these anions to the PTM radicals with p-chloranil. The second route consisted of directly coupling the open-shell compound, 2-Rad, with the azidomethyl ferrocene to obtain also the two cycloaddition compounds 3-Rad and 4-Rad with better yield than in the case of the non-radical derivative. All the compounds were fully characterized by spectroscopic techniques (NMR, FT-IR, and UV/vis), cyclic voltammetry and electron paramagnetic resonance for the radicals (see ESI †).</p><p>On-surface PTM radical functionalization. PTM SAMs were prepared using a freshly template-stripped Au (Au TS ) surface as the substrate. 58 Its ultraat topography makes it an ideal bottom electrode for the charge transport measurements, as well as for promoting the formation of higher quality monolayers. 59 Compounds 2-Rad, 3-Rad, 2-H and 3-H were used to generate SAM-2-Rad, SAM-3-Rad, SAM-2-H and SAM-3-H, respectively. These SAMs were prepared following a similar previously reported methodology, 26 working under inert conditions to avoid the oxidation of the alkyne (see ESI † for further details on the SAM preparation). All the SAMs were characterized by different electrochemical techniques: CV, square wave voltammetry (SWV) and electrochemical impedance spectroscopy (EIS). This allowed us to obtain information about the faradaic transfer process of the anchored molecules, the surface reaction yield and the capacitive behavior at different potentials.</p><p>SAM-5-Rad was obtained by the on-surface post-modication of SAM-2-Rad. The determination of the Fc/PTM ratio for the SAM-5-Rad by SWV and its control were essential to (i) estimate the yield of the on-surface click reaction, (ii) improve the faradaic-background current ratio, and (iii) decrease both capacitive currents and the double layer impact. It is important to mention that there is an intrinsic error in the estimation of the Fc/PTM ratio due to the different electron transfer rates of the electrochemical processes of both redox components but the values were validated by comparison with those of SAM-3-Rad (Fig. S8 †) in which the Fc/PTM ratio should be 1 : 1. Following the conditions used in solution, an $30% reaction yield (Fc vs. PTM radical) was obtained (Fig. S4 and S5 †). This value was enhanced to approximately 50% when a toluene-THF mixture was changed for acetonitrile as the solvent (Fig. S6 and S7 †). This gain was attributed to both the better solubility of CuI and stability of Cu(I) in solution. Other parameters such as temperature and/or reaction time did not lead to a higher Fc surface coverage. The bulkiness of the PTM moieties, which can induce certain disorder of the layer, as was observed before by STM for a previously reported thiolated PTM SAM, 60 would be responsible for the yield decrease. It is worth noting that the redox potential of the PTM radical bound to gold was ca. 400 mV lower than that observed for SAM-1-Rad-Si under illumination. This trend is not surprising and is usually observed for electroactive molecule-modied photoelectrodes. 29 In the case of silicon, the reduction of the PTM radical is easier owing to the photogenerated electron-induced activation of the redox process. 61 As expected, for SAM-3-H and SAM-2-Rad, only a single redox process attributed to either Fc or the PTM radical, respectively, was observed. In the case of SAM-3-Rad and SAM-5-Rad, two redox processes were present owing to the double functionality of the graed monolayer.</p><p>Interestingly, the PTM radical surface coverage estimated for the SAM-3-Rad, which was obtained by direct Au functionalization with the PTM-Fc dyad 3-Rad, was very similar to that of the SAM-2-Rad (7.9 and 8.9 Â 10 À11 mol cm À2 , respectively) (Table S2 †). Such a result denotes that the self-organization and the reactivity of the molecules for both SAMs were quite similar. These values further support that only one of the terminal alkynes of the 2-Rad reacted with the Au surface.</p><p>To back up these ndings, we investigated SAM-2-Rad and SAM-3-Rad by XPS 62 which is an effective tool for studying organic radical thin lms. 63 It was found that the ratio of the integrated signal intensities of the different lines of the XPS curves agrees to a good extent with the theoretical percentages indicating that SAM-2-Rad and SAM-3-Rad were effectively obtained. The carbon concentration slightly exceeds the theoretical values (see Tables S3 and S4 †). This is the usual case in samples prepared ex situ with wet-environment techniques. Molecular orbitals are highly directional and, consequently, they are perturbed if a different/new chemical interaction occurs. In this case, no deviation from the expected shapes was observed from the XPS main lines. Furthermore, a best t procedure applied to the XPS spectra allows identifying the contributions from atomic sites having slightly different binding energies due to variations in the chemical environment. 62 Several constraints based on electronegativity and bond strength were applied 64,65 (see ESI † for details). The comparison of the tted results with the theoretical stoichiometry of the carbon atoms supports the successful preparation of SAM-3-Rad (Fig. S18 † and Table S3 †). In fact, any perturbation of the chemical environment of the carbon atoms that would hint at different chemical congurations would deliver a different best t curve.</p><p>As depicted in the CVs shown in Fig. 4c and d, due to the presence of two electroactive moieties, an enhanced number of stable redox states can be realized, which is of interest for molecular memory devices. In particular, the PTM-Fc dyad displays two redox processes and hence, three distinct redox states, [PTManion-Fc] À , PTMradical-Fc and [PTMradical-Fc] + , in the À0.6 V to 0.8 V potential range. Such an electrochemical response prompted us to characterize SAM-3-Rad by EIS, using the applied potential as a perturbation signal and the capacitive real component as an output, characteristic of the interface at each redox state. This methodology was employed before in Fcbased and tetrathiafulvalene (TTF)-based SAMs by some of us. 66,67 It is important to emphasize that the three redox states are well differentiated in a narrow and stable potential window which makes the system very appealing. 68 The measurements were acquired in the frequency range from 100 kHz to 0.1 Hz with a 20 mV amplitude and the applied potentials were À0.5, 0.3 and 0.7 V. These values were chosen from the CV results to eliminate the contribution of the faradaic processes and thus to only consider the capacitive behaviour of the double layer. The Nyquist plots showed a common shape for a SAM-based interface with one semicircle without the diffusional process (Fig. S17 †). The difference between the [PTManion-Fc] À , PTMradical-Fc and [PTMradical-Fc] + species was evidenced in the Cole-Cole plot, where the imaginary capacitance (C im ) is plotted vs. the real capacitance(C re ) 48 (Fig. 5a). The obtained C re values (indicated in the plot) were: C re1 (PTManion-Fc) ¼ 6.5 mF cm À2 , C re2 (PTMradical-Fc) ¼ 8.0 mF cm À2 and C re3 (PTMradical-Fc + ) ¼ 12.8 mF cm À2 .</p><p>Clearly, the higher C re value for the positively charged SAM (C re3 ) is attributed to a well-dened charged double layer at the interface. The charged ferrocenium species are exposed to the top solid-electrolyte interface favoring the formation of a tight ion pair with the electrolyte anion (Fig. 5b, C re3 ). In the case of the PTManion-Fc SAM, the negative charge of the PTM anion is located in the central carbon atom, which is shielded by the bulky chlorinated triphenyl rings. So, it forms a loose ion pair with the electrolyte cation. 69 Therefore, this ill-dened double layer generated at the PTManion-Fc SAM/electrolyte interface is expected to be very similar to that of the neutral PTM-Fc SAM (C re2 ).</p><!><p>Charge transport measurements across the Au TS /SAM-N-rad/ liquid metal were performed in order to evaluate the SAMs as molecular wires, and to examine the inuence of the D-A dyad. Moreover, the electrical output of the junction could be used to evaluate the second moiety introduced through the click chemistry on the PTM radicals (Fig. 6). The eutectic gallium indium alloy (EGaIn) was chosen as the top electrode, 70,71 since it has been successfully used previously to investigate different PTM-based SAMs and lms. [72][73][74] The spontaneously formed oxide skin (mainly Ga 2 O 3 ) allows the electrode to be shaped as a cone. In this work, the area of the GaO x /EGaIn tip that was brought into contact with the sample surfaces was around 1000-2500 mm 2 in order to obtain stable measurements (see ESI † for details on the experimental setup). 75 Fig. 6 shows that the three monolayers display a different molecular wire behaviour, in terms of the measured current through the junction, showing a current density (J) (SAM-2-Rad) > J (SAM-5-Rad) > J (SAM-3-Rad) (see ESI † for details on the statistical analysis). The differences between the three layers were mainly attributed to the different tunneling distances between the two electrodes arising from the thickness of the molecular layer. Taking into account that an approximately 50% of Fc incorporation was achieved to form SAM-5-Rad through the click reaction on SAM-2-Rad, the electrical response measured for this SAM was considered to be an average contribution of SAM-2-Rad and SAM-3-Rad. Current rectication was previously reported for SAMs incorporating Fc as the redox-active moiety. 73,76 Here, in the AE1 V range, we did not observe such a behaviour (inset in Fig. 6). This could be attributed to a poor molecular order within the layer and hence a poor directionality of the Fc moiety towards the top electrode, compared to well-ordered Fc-thiolated SAMs presenting a well-dened interface with the liquid-metal electrode. 76 XPS and Near Edge X-ray Absorption Fine Structure (NEX-AFS) measurements helped to conrm this interpretation. NEXAFS measurements of SAM-5-Rad show a clear Fe signal (Fig. 7d), however, the survey XPS and the Fe 2p core level spectra at 800 eV photon energy (Fi. S18 in ESI †) are characterized by a signal that is proportional to a lower concentration of iron atoms in comparison with the theoretical stoichiometry. These ndings can be taken as an indication that SAM-5-Rad was obtained, however, the monolayer has a mixed nature: also molecules that do not carry the ferrocene unit are present, i.e., SAM-2-Rad, and they contribute to the signal. NEXAFS measurements also allow determining the average molecular arrangement of the molecules in the lms. 77 The C-K edge of SAM-5-Rad shows typical features due to transitions from the C 1s core levels to the unoccupied states, p* and s* (Fig. 7b). In analogy with previous NEXAFS measurements of derivatives of the PTM radical, two main regions can be identied in the spectra in Fig. 7: the p* region up to around 290 eV and the s* region in the photon energy range above 290 eV. 73,78,79 We can assign the feature at around 285.4 eV to contributions due to transitions from the C 1s core levels located in the carbon atoms of the aromatic rings and of the perchlorinated benzene rings to p* orbitals. The feature at 287.2 eV (see Fig. 7b, grazing incidence) is typical of ferrocene C-K NEXAFS spectra and is due to the C 1s to Fe 3d/p* transitions. [80][81][82] We observe a shoulder at 285.9 eV that is also assigned to contributions belonging to the ferrocene moiety. [80][81][82] The intensity at around 282.2 eV corresponds to transitions from the C 1s core levels to the singly unoccupied molecular orbital (SUMO). 60 The determination of the average molecular orientation of SAM-5-Rad is very challenging, the Cl-L edge signal shows a very weak dichroism (Fig. 7a), and the resonances in the C-K edge spectra due to transitions from aromatic carbon atoms are all very close in energy (Fig. 7b). Thus, we have focused on the single building blocks. We have calculated an average molecular orientation of the ferrocene moiety of 22 with respect to the substrate using the intensities for the two polarizations at 285.9 eV and cross-checking with the intensities at 287.2 eV (21 ). 77,83,84 The relative torsion of the component blocks in SAM-5-Rad together with the presence of SAM-2-Rad might hinder a good interaction with the top contact and, thus, contribute to the absence of a current rectication behaviour.</p><!><p>In summary, we have demonstrated that the functionalization of electro-and magnetically-active perchlorotriphenylmethyl radicals with terminal alkynes clearly permits us to expand their applicability. It has been shown that using the alkyne group as the graing unit, chemically bonded self-assembled monolayers can be successfully prepared both on hydrogenated silicon and on gold in one step and preserving the radical character. The electronic properties of Si have allowed us to fabricate a light-triggered capacitance switch exploiting the redox properties of the graed radicals. In the case of gold, not only the formation of the SAM but also its post-modication with other functional molecules through click chemistry has demonstrated the potential of such a versatile platform for achieving multifunctional layers displaying the fascinating properties of the attached organic radicals. In particular, here a radical donor-acceptor dyad has been on-surface synthesized following optimized conditions that enable the spin to be unaltered. For all this, we believe that the ndings reported in this manuscript are a signicant step forward in the implementation of organic radicals in molecular based devices with different properties and applications in elds such as energy storage and conversion, sensing, imaging, memory devices, and spintronics.</p><!><p>There are no conicts to declare.</p>

Royal Society of Chemistry (RSC)

Solid State and Solution Dynamics of Pyridine Based Tetraaza-Macrocyclic Lanthanide Chelates Possessing Phosphonate Ligating Functionality (Ln-PCTMB): Effect on Relaxometry and Optical Properties

The macrocyclic ligand 3,6,9-tris(methylenebutyl phosphonic acid)-3,6,9-15-tetraazabicyclo [9.3.1]pentadeca-1(15),11,13-triene (PCTMB) was synthesized and complexes of Eu3+, Tb3+, and Gd3+ studied by X-ray crystallography, luminescence, and relaxometry. In the crystal these complexes are dimeric and possess 8-coordinate Ln3+ centers that are linked by bridging phosphonates. The rigidity introduced by the pyridyl nucleus forces the EuPCTMB and TbPCTMB to adopt a twisted snub disphenoid (TSD) coordination geometry. Examination of the 5D0 \xe2\x86\x92 7F0 luminescent transition of EuPCTMB in the solid state confirmed the existence of a single distinct Eu3+ coordination environment, whereas two Eu3+ coordination environments were observed in aqueous solution. Lifetime analysis of aqueous TbPCTMB solutions determined that q = 0.1 and q = 1.0 for the two coordination environments and Stern-Volmer quenching constants (KSV\xcf\x84 = 1101 M\xe2\x88\x921, KSV\xce\xa6 = 40780 M\xe2\x88\x921) support the presence of a monomer/dimer equilibrium. Relaxivity studies of GdPCTMB in H2O/CH3OH exhibited a concentration dependency (0.02 mM \xe2\x80\x9310.00 mM) ranging from r1 = 7.0 mM\xe2\x88\x921s\xe2\x88\x921 to 4.0 mM\xe2\x88\x921s\xe2\x88\x921 consistent with the trend observed by luminescence.

solid_state_and_solution_dynamics_of_pyridine_based_tetraaza-macrocyclic_lanthanide_chelates_possess

Introduction<!>General Remarks<!>3,6,9-Tris[methylene(di-butylphosphonate)]-3,6,9,15-tetraazabicyclo[9.3.1] pentadeca-1(15), 11, 13-triene (1)<!>3,6,9-Tris[methylene(butylphosphonate)]-3,6,9-15-tetraazabicyclo [9.3.1]pentadeca-1(15),11,13-triene potassium salt (K3PCTMB)<!>General procedure for the preparations of LnPCTMB Chelates<!>Eurpoium (III) 3,6,9-Tris[methylene(butylphosphonate)]-3,6,9-15-tetraazabicyclo [9.3.1]pentadeca-1(15),11,13-triene (EuPCTMB)<!>Gadolinium(III) 3,6,9-Tris[methylene(butylphosphonate)]-3,6,9-15-tetraazabicyclo [9.3.1]pentadeca-1(15),11,13-triene (GdPCTMB)<!>Terbium(III) 3,6,9-Tris[methylene(butylphosphonate)]-3,6,9-15-tetraazabicyclo [9.3.1]pentadeca-1(15),11,13-triene (TbPCTMB)<!>X-Ray Crystallography Data Collection: [EuPCTMB]2\xe2\x80\xa29H2O and [TbPCTMB]2\xe2\x80\xa2925H2O<!>Luminescent Measurements<!>Synthesis<!>Crystallographic Studies<!>Solution State Studies<!>Relaxometric Studies<!>Luminescence Studies<!>Conclusions<!>

<p>The lanthanide series of metal ions possess a diverse array of physical properties that have been the source of many technological innovations over the years.1 In particular, lanthanides have become invaluable components for contrast enhancement media used in nuclear medicine, magnetic resonance imaging and as optical probes. Equally important for advancing their use in medicine has been the molecular architecture of tailored ligand systems that can provide an optimal coordination environment for maximum stability; a requirement for rendering the metal ion inert and non-toxic for in vivo applications. Furthermore, the organic ligand framework can function to enhance photo-physical properties2–5 and modulation of water exchange kinetics in the chelate structure.6–8 For biological applications, the cumulative body of knowledge has clearly demonstrated that the high thermodynamic and kinetic inertness of chelates formed between Ln3+ ions and polyaza-macrocyclic ligand systems is a paramount feature for safeguarding long term chelate integrity.9 Of considerable interest in this regard is the family of chelates derived from 1,4,7,10-tetraazacyclododecane (cyclen) which is a familiar signature of many MRI and radiopharmaceutical contrast agents in clinical use and under development.</p><p>The versatility of cyclen-based ligands for lanthanide coordination is evidenced by the many related anologs present in the literature that are designed for enhancing various aspects of performance through structural "fine tuning". A noteworthy example reported over 25 years ago is the 12-membered tetraaza-macrocyclic ligand which incorporates a pyridine nucleus within the macrocyclic ring, pyclen.10 The aminocarboxylic acid derivative of pyclen, PCTA, is known to form stable lanthanide chelates (log KLnL = 20.39),11 its Gd3+ chelate was also considered as a neutrally charged, general perfusion MRI contrast agent.12 Surprisingly, however, these fascinating derivatives have received far less attention than their cyclen-based counterparts. Recently pyclen-based ligands have been rediscovered through several reports that survey the effect of structural morphology upon water exchange kinetics in lanthanide chelates; a potentially useful tool for optimizing the performance of MR contrast media.13–15 In addition, these same ligands have been found to possess rapid chelation kinetics under very mild conditions thus stimulating renewed interest for nuclear medicine applications where the time required for complexation to occur is an important concern.11</p><p>Our interest in the pyclen family of chelates stems from their unique multi-modal imaging potential and the ability to control in vivo tissue targeting through alterations of ligating functionality. In particular it has been demonstrated that incorporation of a phosphonate ester ligating functionality provides an efficient means of altering biodistribution properties and for selective targeting of cancer.16 In the case of highly lipophilic phosphonate chelates the formation of non-covalent dimers has been inferred to be an important aspect of the cellular targeting process. To gain a better understanding of the chemistry responsible these observations we undertook a study of the solid state and solution behaviour of a pyclen-phosphonate chelate which has been suggested to dimerize in aqueous media. Herein is presented a detailed analysis of the solid state crystal data and solution dynamics that are relevant to the design of targeted diagnostic and therapeutic agents structured around the lanthanide chelate.</p><!><p>All solvents and reagents were purchased from commercial sources and used as received unless otherwise stated. 1H, 13C and 31P NMR spectra were recorded on a Varian Mercury or Bruker Avance spectrometer operating at 299.99, 75.43 and 121.44 MHz, respectively. Infrared spectra were recorded on a Perkin Elmer 1600 Series FTIR. 3, 6, 9, 15-tetraazabicyclo[9.3.1] pentadeca-1(15), 11, 13-triene (pyclen) was prepared using previously published methods.10</p><!><p>Paraformaldehyde (1.2 g, 38.2 mmol) was added to a solution of 3, 6, 9, 15-tetraazabicyclo[9.3.1]pentadeca-1(15), 11, 13-triene (2.5 g, 12.1 mmol) in THF (30 mL) and the resulting slurry stirred at ambient temperature for 30 minutes. Tri-n-butyl phosphite (9.6 g, 38.2 mmol) was then added and the turbid reaction mixture stirred for an additional 72 hours at ambient temperature. The resulting homogeneous reaction mixture was concentrated in vacuo to give a pale yellow, viscous oil (10.0 g). The crude product was purified by column chromatography on basic alumina (5 × 26 cm) eluting with chloroform. Two fractions were collected (2 column volumes) beginning with the second column volume of eluent. After concentration in vacuo the product was isolated as colorless oil (9.06 g, 92%). 1H NMR (300 MHz, CDCl3) δ = 0.87 (m, CH3, 18H), 1.35 (m, CH2, 12H), 1.59 (m, CH2, 12H), 2.62 (m, CH2, 4H), 2.74 (m, CH2, 10H), 3.05 (d, CH2, 2JPH 10 Hz, 4H), 4.01 (m, OCH2, 12H), 7.17 (d, 3-Ar, 3JH-H 8 Hz, 2H), 7.58 (t, 4-Ar, 3JH-H 8 Hz, 1H); 13C NMR (75 MHz, CDCl3) δ = 13.5 (CH3), 13.52 (CH3), 18.6 (CH2), 18.7 (CH2), 32.2 (d, CH2, 3JPC 6 Hz), 32.5 (d, CH2, 3JPC 6 Hz), 50.1 (d, CH2, 2JPC 8 Hz), 51.2 (d, CH2, 1JPC 157 Hz), 51.2 (d, CH2, 2JPC 10 Hz), 51.9 (d, CH2, 1JPC 157 Hz), 60.8 (CH2), 65.6 (d, CH2, 3JPC 6 Hz), 67.2 (d, 2JPC 6 Hz), 123.0 (3-Ar), 137.0 (2-Ar), 157.2 (4-Ar); 31P[1H] NMR (121.44 MHz, CDCl3) δ = 24.94 (2P), 24.96 (1P).</p><p>m/z: (ESI+); 826 (100% [M + H]+.</p><!><p>A solution of the hexa-ester 1 (9.06 g, 11.36 mmol) and KOH (7.65 g, 136.3 mmol) in water (100 mL) and 1,4-dioxane (30 mL) was heated to reflux for 18 hours. The reaction mixture was then cooled and filtered while warm. The filtrate was concentrated to give a solid which was suspended in 5:1 v/v CHCl3/MeOH (200 mL) and heated to reflux. The hot solution was filtered and the solvent removed from the filtrate in vacuo to afford an off-white solid. The residue was taken up into CHCl3 (60 mL) with stirring followed by the addition of CH3CN in small portions until the solution became slightly turbid. Upon cooling with continued stirring in an ice bath a white precipitate was observed. The precipitate was isolated by filtration and washed with CH3CN. After drying in vacuo the title compound was isolated as a colorless solid (4.14 grams, 49%). mp: 150–160 °C, dec.; 1H NMR (300 MHz, D2O) δ = 0.78 (m, CH3, 9H), 1.22 (m, CH2, 6H), 1.43 (m, CH2, 6H), 2.47 (m, CH2, 6H), 2.81 (m, CH2, 10H), 3.65 (d, CH2, 2JPH 10 Hz, 4H), 4.84 (m, OCH2, 6H), 7.22 (d, 3-Ar, 3JH-H 8 Hz, 2H), 7.63 (t, 4-Ar, 3JH-H 8 Hz, 1H); 13C NMR (75 MHz, D2O) δ = 15.9 (CH3), 21.3 (CH2), 35.2 (CH2), 35.2 (CH2), 52.0 (CH2), 52.9 (CH2), 55.9 (d, 1JP-C 144 Hz), 62.5 (CH2), 67.2 (d, -OCH2, 2JP-C 6 Hz), 126.3 (3-Ar), 141.7 (2-Ar), 159.7 (4-Ar); 31P[1H] (121.44 MHz, D2O) δ = 20.59; IR υmax/cm−1: 1211 (P=O), 1068 (P-O-C); m/z: (ESI+); 695 (100% [M + H3 + K]+; Anal. Found C 35.4 H 5.9 N 5.7, C26H48K3N4O9P3•2.5CH3OH•2KOH requires C. 35.5 H 6.3 N. 5.8.</p><!><p>K3PCTMB (352 mg, 0.4 mmol) was dissolved in water (5 mL) and the strongly basic solution adjusted to pH 5 via drop-wise addition of 6N HCl. An aqueous LnCl3 solution (0.4 mmol in 1.5 mL) was then added to the ligand in 200 μL aliquots. The reaction was stirred throughout and the pH maintained close to 5 by addition of a 1N KOH solution. Following the addition of each LnCl3 aliquot, chelation progress was monitored by RP- HPLC (Phenomenex PRP-1 C18 column (4.6 × 250 mm), 1 mL/min, 80/20 v/v methanol:water, λ = 266 nm; ligand tR = 6 min., chelate tR = 10 min.). When all the free ligand was found to be consumed the aqueous solution was filtered through a 0.2 μm syringe filter and lyophilized to give the chelate as a flocculent white solid. Each chelate was then dissolved in a minimum of boiling water and, upon cooling was found to crystallize as a colorless solid.</p><!><p>m/z: (ESI+); 807 (100%, [EuLH]+, an appropriate isotope pattern was observed); IR υmax/cm−1: 1234 (P=O), 1054 (P-O-C); Anal. Found: C 34.9 H 6.1 N 6.2 C26H48N4O9 KP3Eu•3H2O requires C 34.8 H 6.1 N 6.2.</p><!><p>m/z: (ESI+); 812 (100% [GdLH]+, an appropriate isotope pattern was observed); IR υmax/cm−1: 1235 (P=O), 1054 (P-O-C); Anal. Found: C 35.2 H 6.2 N 6.3 C26H48N4O9 KP3Gd•2H2O requires C 35.3 H 6.0 N 6.3.</p><!><p>m/z: (ESI+); 813 (100% [TbLH]+, an appropriate isotope pattern was observed); IR υmax/cm−1: 1234 (P=O), 1054 (P-O-C); Anal. Found: C 34.5 H 6.0 N 6.1 C26H48N4O9 KP3Tb•3H2O requires C 34.5 H 6.0 N 6.2.</p><!><p>A Leica Z microscope was used to identify a suitable colorless needle 0.3mm × 0.08mm × 0.03mm from a representative sample of crystals of the same habit. The crystal was coated in a cryogenic protectant (paratpne), and was then fixed to a loop which in turn was fashioned to a copper mounting pin. The mounted crystal was then placed in a cold nitrogen stream (Oxford) maintained at 110K.</p><p>A BRUKER SMART 1000 X-ray three-circle diffractometer was employed for crystal screening, unit cell determination and data collection. The goniometer was controlled using the Smart software suite (Microsoft operating system). The sample was optically centered with the aid of a video camera such that no translations were observed as the crystal was rotated through all positions. The detector was set at 5.0 cm from the crystal sample (CCD-, 512×512 pixel). The X-ray radiation employed was generated from a Mo sealed X-ray tube (Kα = 0.70173Å with a potential of 50 kV and a current of 40 mA) and filtered with a graphite monochromator in the parallel mode (175 mm collimator with 0.8 mm pinholes).</p><p>Dark currents were obtained for the appropriate exposure time of 10 sec and a rotation exposure was taken to determine crystal quality and the X-ray beam intersection with the detector. The beam intersection coordinates were compared to the configured coordinates and changes were made accordingly. The rotation exposure indicated acceptable crystal quality and the unit cell determination was undertaken. Forty data frames were taken at widths of 0.3° with an exposure time of 10 seconds. Over 200 reflections were centered and their positions were determined. These reflections were used in the auto-indexing procedure to determine the unit cell. A suitable cell was found and refined by nonlinear least squares and Bravais lattice procedures and reported. The unit cell was verified by examination of the hkl overlays on several frames of data including zone photographs. No super-cell or erroneous reflections were observed.</p><p>After careful examination of the unit cell a standard data collection procedure was initiated. This procedure consists of collection of one hemisphere of data collected using omega scans, involving the collection over 2400 0.3° frames at fixed angles for φ, 2θ, and χ (2θ = −28°, χ = 54.73°), while varying omega. Each frame was exposed for 20 sec and contrasted against a 20 sec. dark current exposure. The total data collection was performed for duration of approximately 15 hours at 110 K. No significant intensity fluctuations of equivalent reflections were observed. After data collection the crystal was measured carefully for size, morphology and color.</p><!><p>Steady state and lifetime measurements were performed on an Edinburgh Instruments FL/FS900CDT Fluorometer equipped with a 450W xenon arc lamp and a 100W μF 920H flash lamp. High resolution emission spectra (ΔJ = 0) were recorded from 578.5–582.0 nm using a 0.05 nm emission monochromator step size (λex= 270 nm). Full emission spectra were recorded from 525–725 nm using a 0.25 nm step size. For Stern-Volmer static and dynamic quenching constant measurements λex= 270 nm, λem= 616 nm. Solutions of Eu-PCTMB were prepared at 0.01 mM with varying quencher concentration (Nd-PCTMB) ranging from 0.0 mM – 0.04 mM. All solutions were stirred for one hour at 60 °C then equilibrated for 4 hours at room temperature prior to measurements.</p><!><p>Pyclen was synthesized according to procedure described by Stetter et al.10 Subsequently, pyclen was reacted with paraformaldehyde and tributyl phosphite in THF to afford the hexa-butyl phosphonate 1. Selective hydrolysis of the ester intermediate to the mono ester was achieved under basic hydrolysis conditions using 1 equivalent of KOH per phosphonate in dioxane and water. The hydrolysis reaction was followed using 31P NMR by monitoring the disappearance di-alkyl phosphonate ester resonances at 25 ppm and the appearance of the mono-ester resonances at approximately 20 ppm. Upon completion the reaction mixture was concentrated and the crude product crystallized as the potassium salt from MeOH/CH3CN. Lanthanide chelates were then prepared by acidifying an aqueous solution of K3PCTMB (pH 5) followed by aqueous LnCl3 in small aliquots. The pH was maintained between 5 and 6 by addition of small amounts of a KOH solution and the chelation reaction was monitored by RP-HPLC. When all the free ligand was consumed the aqueous solution was freeze dried and crystallized from hot water to afford X-ray quality crystals.</p><!><p>X-ray diffraction of single crystals of Eu1 and Tb1 revealed that both chelates crystallized in the Pı̄ space group (Table 1). The two chelates are isostructural and have 2 chelate molecules and 9 waters of crystallization in the unit cell. The PCTMB ligand occupies 7 of the 8 coordination sites of each metal ion with water excluded from the inner coordination sphere by a fourth coordinating phosphonate monoester. The source of this eighth ligand is the neighbouring chelate molecule of the unit cell in much the same way as is observed for the analogous cyclen-based systems 2 and 317, 18 wherein the metal ions of Eu2 and Tb3 are sandwiched between 4 nitrogen and 4 oxygen donor atoms (Figure 1). Although superficially this appears similar to many macrocyclic octa-coordinate Ln3+ chelates of cyclen-based phosphonate (DOTP)19 and phosphinate ligands, Ln2-6,20–22 there are significant differences between the coordination environment in those chelates and the chelates of PCTMB. These differences stem primarily from the nature of the macrocyclic ring. Cyclen adopts a square conformation, defined by Dale's nomenclature23 as [3,3,3,3], in which the nitrogen atoms are located on the sides of the ring and each ethylene bridge adopts a gauche conformation.24–26 It has been observed that if the conformation of cyclen is distorted into a another conformation significant loss of chelate stability can sometimes result.27 In the case of pyclen, however, the rigidity of the pyridine group dictates that one side of the macrocycle must incorporate 4 bonds and so a [3,3,3,3] conformation is impossible. Instead the macrocyclic ring adopts a [4,2,4,2] ring conformation with the nitrogen atoms located in the centre of each side (Figure 2). The structure of the pyclen ring in Eu1 (Figure 1) closely resembles that reported previously for the macrocycle alone.28</p><p>The incorporation of the pyridyl group into the macrocycle also has one further consequence for the chirality of the macrocyclic ring. Cyclen may adopt one of two conformations, each of which is chiral; either (δδδδ) or (λλλλ) conformation may be adopted according to the helicity of its ethylene bridges. Each ethylene bridge in cyclen adopts the same helicity as the others within the cyclen ring. These two conformations of cyclen may interconvert through a ring flipping motion. In contrast, the conformation of each N-C-C-N bridge in pyclen alternates around the ring, such that pyclen adopts a (δλδλ) conformation. Because (δλδλ) pyclen is the mirror image of (λδλδ) pylcen (Figure 2), pyclen is achiral. This means that the chelates of pyclen have one fewer elements of chirality than their related cyclen-based chelates. This is significant because pyclen, once it has become part of a chelate, is unable to flip its conformation, as cyclen can, owing to the presence of the fused pyridyl ring. This means that despite being structurally rigid in a chelate, pyclen, unlike rigid cyclen derivatives,29–31 adopts one major low energy conformation.</p><p>It is informative to compare the structures of EuPCTMB and TbPCTMB with those of the cyclen-based chelates structures Eu2 and Tb3; some selected geometrical parameters of these chelates are presented in Table 2. Comparing these structures it can be seen that the two ethylene bridges of pyclen adopt gauche conformations, comparable with the cyclen derivatives. On the opposite side of the ring however, pyclen is much more strained, a consequence of including pyridine in the system. Because of the planar arrangement of the pyridine ring and its substituents in the 2- and 6- positions, the N-C-C-N torsion angles on this side of the macrocycle are approximately half that of a gauche conformation. The result is that the nitrogen atoms in the 4- and 10- positions are oriented below the level of those in the 1- and 7- positions. In contrast to cyclen, the nitrogen atoms of pyclen are not co-planar, lying about 0.4 Å above or below the mean plane (Table 2). Not surprisingly the nitrogen of the pyridine ring (N-1) lies furthest from this plane. This in turn distorts the donor oxygen atom plane of EuPCTMB and TbPCTMB by a similar amount. As a consequence, the chelates of PCTMB are unable to assume the twisted square antiprismatic (TSAP) coordination geometry adopted by the chelates of Eu2 and Tb3. Instead the EuPCTMB and TbPCTMB are found to adopt a twisted snub dispheniod (TSD) coordination geometry (Figure 3). In so doing, the chelates of PCTMB are able to maintain metal-donor atom bond distances comparable to those observed in the chelates Eu2 and Tb3.</p><p>Although the coordination geometry of EuPCTMB and TbPCTMB are different from those of the cyclen based phosphinate and phosphonate derivatives 2-6 the phosphonate groups remain successful at excluding water from the inner coordination sphere. Lukeš and co-workers32 have suggested that a critical parameter in achieving this goal is the O-M-O bond angle, β (Table 3). If this angle becomes tighter than 136° then the vacant coordination site on the metal ion becomes too sterically encumbered to accommodate a water molecule. The parameters collected in Table 3 seem to confirm this observation with angles (β) significantly smaller than 136° observed for all q = 0 chelates; a category that includes all the phosphonate and phosphinate derivatives. Indeed the narrowest O-M-O angles (β) of all are observed in the cases of EuPCTMB and TbPCTMB and arise from the different coordination geometry observed in these chelates. Although the LnPCTMB chelates have the smallest β angles, of all the chelates collected in Table 3 they also have the largest β′ angles, notably the O-M-O angles are much larger than the critical 136° required by Lukes et al. for water coordination, and yet these chelate remain q = 0. These large differences in O-M-O angles are partly the result of the [4,2,4,2] ring conformation observed in these chelates that also results in a very narrow N-M-N bite angle, which, it should be noted, is not entirely symmetrical. The presence of the pyridine ring results in one side of the macrocycle being brought closer to the metal ion, with the other, more flexible, side bowed out slightly. This is sharp contrast to the near perfect symmetry observed in the structure of GdDOTP.19</p><p>In chelates of DOTA and DOTAM, both of which possess one bound water molecule (q = 1), the Ln3+ ion lies much closer to the 4 oxygen donor atoms than it does to the 4 nitrogen donors. For these chelates the metal ion is typically found to lie approximately 7/10 of the distance (d/c, Table 3) to the mean oxygen atom plane, irrespective of which coordination geometry (SAP or TSAP) is observed.33 In contrast the d/c ratios for the q = 0 cyclen-based phosphonate and phosphinate chelates (Table 3) are much smaller (typically ≈ 0.62), indicating that the Ln3+ ion lies significantly closer to the mean nitrogen plane. Significantly, the d/c ratios of EuPCTMB and TbPCTMB are in line with those observed for these cyclen-based phosphonate and phosphinate systems. Although the distance between the mean nitrogen and oxygen atom planes is about 0.15 Å greater in the phosphonate and phosphinate systems compared with the acetate and amide systems, the metal also lies closer to the nitrogen plane in absolute terms, by about 0.1 Å. The crystal structures of the Ln3+ chelates of ligands 3 and 4 and their d/c ratios provides some insight into the relationship between the position of the Ln3+ ion and the hydration state (q value) of the chelate. The crystal structures of Ln3 chelates across the lanthanide series reveal that the d/c ratio falls across the series, reaching a minimum of about 0.61 at europium and remaining more or less constant thereafter. Thus, the chelates of the smaller heavy lanthanides Eu3, Tb3, Er3 Yb3, Eu4 and Y4, all of which are q = 0, have small d/c ratios, between 0.61 and 0.63.18 Nd3 has slightly larger d/c ratio, 0.65, but a small O-M-O bond angle, 129°, and is also q = 0.18 However, chelates of larger Ln3+ ions La4, La3 and Ce3 are q = 1 and have d/c ratios comparable to those of DOTA and DOTAM chelates, d/c = 0.68 – 0.70.18, 21 The smaller of the two O-M-O bond angles (β) in the chelates of La3 and Ce3 are 135°,18 a shade narrower than the cut-off proposed by Lukeš and co-workers,32 and yet a water molecule is still able to coordinate with the metal center. The crystal structure of Pr3, presented subsequently,34 provides reason for pause at this point; it is presented as a q = 1 chelate and yet has β = 129.7° and β′ = 136.4° and a relatively small d/c ratio of 0.65. It is not until one considers that the reported water bond distance, 2.820 Å, is longer than is normally considered a bonding interaction that we can really understand this chelate. Pr3 is, in reality, probably q = 0 but is able to maintain an interaction with a water molecule as a result of its d/c ratio. This evidence suggests that when the Ln3+ ion in aza-crown based chelates such as these is octa-coordinate and a water molecule is absent from its coordination sphere, the Ln3+ ion moves towards the amines of the macrocycle in search of increased electron density. The question as to which comes first, movement of the Ln3+ ion or departure of the water molecule, is something of a 'chicken and the egg problem'. However it seems clear that this has implications for understanding dissociative water exchange processes during which it now seems likely that the position of the Ln3+ ion fluctuates according to the hydration state of the chelate. It also implies that the d/c ratio may provide a better delineation between chelates that can be hydrated (q = 1) and those that cannot (q = 0) given that the O-M-O angle which may vary significantly, even within the same chelate. It seems that one may conclude from this that a d/c ratio much smaller than 0.68 will result in water being excluded from the inner coordination sphere, but that until this ratio is smaller than 0.65 non-bonding interactions between metal and water are still possible.</p><p>Owing to the absence of chirality in the pyclen ring, the chirality of an LnPCTMB chelate is determined by the orientation of the pendant arms (Δ or Λ) and the configuration at phosphorus (R- or S-). The prochiral phosphorus atom of the ligand becomes chiral upon coordination with the metal ion, so either configuration may be result from the synthesis of the chelate. Studies into the solution and solid state structures of Ln4 and Ln5 chelates have shown an interdependence of the orientation of the pendant arms and the configuration at phosphorous.20, 37 A single C4-symmetric coordination isomer predominates in solution for Ln4 and Ln5 chelates, while crystal structure data reveals that a Λ orientation of the pendant arms is associated with an R- configuration at phosphorous and a Δ orientation with an S- configuration at phosphorus.20, 21 This observation is reminiscent of that observed in α-substituted acetate derivatives of cyclen.29, 38–40 In sharp contrast, these two elements of chirality do not exhibit any such interdependence in the Ln6 chelates, despite the apparent structural similarity of the ligand systems.22 Multiple isomers of Ln6 chelates are observed in solution and, at least in the crystal, the pendant arms bind cooperatively (i.e. with the same helicity) even though the configuration at phosphorous alternates RSRS- around the ring.</p><p>The structures of LnPCTMB, Ln2 and Ln3 chelates in the crystal share some common features (Figure 4 and supplementary figures S1 and S2). Each chelate is present in a dimer. The dimer is associated by two bridges in which one phosphonate of each chelate coordinates to both Ln3+ ions of the dimer. For each of these chelates the dimer is made up of two chelate molecules that are enantiomers of one another. The pendant arms bind cooperatively, with one molecule of the dimer exhibiting a Δ orientation and its partner a Λ orientation. However, when it comes to the configuration at phosphorous these chelates resemble Ln6 chelates more than either the Ln4 or Ln5 chelates. Clearly, the configuration at phosphorous is strongly influenced by the nature of the phosphorous R-substituent. In the structures of LnPCTMB, Ln2, and Ln4 the preferred conformation of the pendant arm is with the R-substituent of phosphorous in a gauche position with respect to the nitrogen of the macrocycle (Figure 5). However, in the case of Ln3 chelates the R-substituent exhibits a preference for a position anti- to the nitrogen of the macrocycle, with only the middle arm adopting the gauche conformation preferred in the chelates of PCTMB, 2 and 4.</p><p>In both LnPCTMB and Ln3 chelates phosphorus adopts RSR-/SRS- configurations going around the macrocyclic ring. In contrast, the configuration at phosphorous in Ln2 chelates is RRS-/SSR-, this despite the interdependence with pendant arm orientation observed in the analogous tetra-phosphinate chelates Ln4 and Ln5. Notably, it is the pendant arm that bridges the two chelate molecules that has the inverted configuration at phosphorous and it is only through this inversion of configuration that the pendant arm is able to make a second oxygen donor atom available for coordination in the dimeric structure (Supplementary Figure S1). It appears that in the case of Ln2, the energy penalty incurred by placing the phosphorous R-substituent into an anti- position is more than compensated for by the drop in energy, and increase in enthalpy, associated with dimerization of the chelate.</p><p>It is less clear why the central pendant arm of LnPCTMB and Ln3 chelates should adopt an anti- conformation. One explanation may lie in a phenomenon exhibited by both chelates and clearly visible by inspection of Figure 2; by inverting the configuration of the middle pendant arm these chelates are able to bind a second sphere water molecule in a pincer action. Hydrogen bonded by both non-bridging phosphonates this water molecule lies closest to the metal ion, 5.574 Å from the Eu3+ ion and 5.476 Å from Tb3+ and may have a residence lifetime on the chelate long enough to have significance in relaxometric studies of the Gd3+ chelate. Notably this pincer binding action used to hold water molecule is absent from the structures of Ln2 chelates. As a result the 4 second sphere water molecules of Ln2 chelates are hydrogen bonded to just one phosphinate each and lie further from the metal ion, over 6.5 Å away.</p><!><p>The dimeric forms of Ln2 and Ln3 chelates that are observed in the crystal structure have also been shown to persist in the solution state.17, 18 The presence of dimeric structures in solution may be beneficial, as in the case of luminescent probes, or detrimental, as in the case of an MRI contrast agent. For these reasons it is important to understand the behaviour of the dimeric structure of PCTMB chelates in solution. LnPCTMB chelates are minimally soluble in water at room temperature. However, upon heating to reflux it is possible to obtain a chelate solution that is 1 – 2 mM, which persists after cooling. Solutions of EuPCTMB and TbPCTMB are brilliantly luminescent when irradiated with common UV sources which provide a convenient tool for probing their sensitized photo-physical properties.</p><!><p>Rapidly exchanging water molecules in the inner-coordination sphere of a gadolinium ion contribute significantly to its overall relaxivity. Clearly then a chelate that excludes all water molecules from the inner-coordination sphere is likely to be ineffective as an MRI contrast agent. Relaxivity (r1) is the measure of how effective a contrast agent and is defined as the increase in longitudinal relaxation rate per unit concentration of contrast agent. Relaxivity is usually determined by measuring the longitudinal relaxation rate (R1 which = 1/T1) of solutions of the contrast agent at different concentrations. A linear regression analysis then affords the relaxivity as the slope of the line. Accordingly, the longitudinal relaxation rates of solutions of GdPCTMB were measured over the concentration range 0.06 – 6.0 mM. Measurements were performed in 2:1 v/v water/methanol solution, owing to the poor solubility of the chelate in water at concentrations as high as 6.0 mM. The results, shown in Figure 6 on a logarithmic axis for greater clarity, reveal that R1 is not linearly dependent upon the concentration. This non-linearity is most usually observed if a structural change, that alters relaxivity, occurs as a result of a change in concentration. Fitting data in the concentration ranges 0.06 – 0.2 mM and 2.0 – 6.0 mM confirms this observation. The data in each of these two concentration ranges fits well to a straight line (Figure 6). Data fitting in the high concentration range affords a value of r1 = 4.1 mM−1s−1. In contrast, the low concentration range affords a value of r1 = 7.0 mM−1s−1.</p><p>These observations are consistent with dissociation of the dimeric structure of GdPCTMB, observed in the crystal structures of EuPCTMB and TbPCTMB, as the concentration of the chelate is decreased. The relaxivity of GdPCTMB obtained at higher concentrations (r1 = 4.1 mM−1s−1 at >2.0 mM) is significantly higher than that observed for Gd2 (r1 = 1.9 mM−1s−1 at 0.1 mM) which was shown to exist solely as a dimer across the entire concentration range studied.17 Parker and co-workers described Gd2 as an entirely outer-sphere chelate owing to this dimerization phenomenon in solution.17 It seems reasonable to believe that at higher concentration, above 2.0 mM, GdPCTMB exists solely in a dimeric form in solution; the enhanced relaxivity of this chelate can then be ascribed to the effect of water molecules in the second hydration sphere. As described earlier LnPCTMB chelates appear to bind water molecules in the second-hydration sphere using two phopshonate mono-esters in a pincer action; this not only decreases the distance of closest approach of molecules in the second hydration sphere but may reasonably be expected to increase their residence lifetime on the chelate. It is now well established that a long lived second-hydration sphere in Gd3+ complexes can lead to substantial relaxivities,41 even in the absence of an inner-hydration sphere; GdDOTP is just one example of this.42 It then appears that as the concentration of GdPCTMB decreases the chelate begins to dissociate, permitting water into the inner coordination sphere. By allowing water into the inner coordination sphere relaxivity the relaxivity of the chelate is increased to 7.0 mM−1s−1. It seems apparent that LnPCTMB chelates, unlike the analogous Ln2 and Ln3 chelates, do not persist as doubly bridged dimers in solution over a wide concentration range. Rather, the extent of dimerization in LnPCTMB chelates is highly concentration dependent, a conclusion supported by the results of luminescent studies on EuPCTMB.</p><!><p>The luminescent properties of Eu3+ can provide valuable insight into the speciation of its chelates in solution. In addition to allowing the hydration state (q value) to be determined using Horrocks' method,43, 44 later modified by Parker and co-workers,45 the 5D0 → 7F0 transition of the Eu3+ emission spectrum provides a single line for each Eu3+ species present in solution, the result of the non-degeneracy of both the 5D0 and 7F0 states. Providing that the energy of this transition is different for each species in solution it is possible to examine each species individually by following this transition. Examining this transition has been used with considerable success in examining hydration equilibria in Eu3+ chelates.46 However, in order to examine the 5D0 → 7F0 transition it is necessary to acquire the emission spectrum with high spectral resolution, 0.1 nm or better. The high resolution emission spectrum of EuPCTMB in the crystalline phase (Figure 7, bottom) is characteristic of a single Eu3+ coordination environment. A single 5D0 → 7F0 transition is observed and the 3 transitions expected for the 5D0 → 7F1 transition of a chelate of this type39 are visible and clearly spaced. Dissolution in water at 0.4 mM significantly changes the nature of the emission spectrum (Figure 7 top); most clearly the two lines observed for the 5D0 → 7F0 transition indicate that two Eu3+ coordination environments are now present. The line at 580.4 nm corresponds closely to the single line observed for the dimer in the crystalline phase at 580.2 nm. The new line, observed at 579.8 nm, may be ascribed to the presence of a hydrated form of the Eu3+ chelate in solution.</p><p>Hydration of a Ln3+ center in this new species was confirmed by determining the q values using an adaptation of Horrocks' method.45 Owing to the poor solubility of the chelates in aqueous solution the hydration state determination was performed on the more emissive Tb3+ chelate. Following excitation at 280 nm and monitoring emission at 545 nm the decay of Tb3+-based luminescence from a 0.15 mM solution of TbPCTMB was monitored. Double exponential decay of the Tb3+ excited state was observed for TbPCTMB. For chelates with a single hydration state a single exponential decay is expected; a double exponential decay is indicative of the presence of two different Tb3+ hydration states. Fitting the luminescent decay curve to a double exponential model (Table 4 and Supplementary Figure S3) revealed that two species were present, with the more prevalent species a q = 0 and the less prevalent q = 1.</p><p>The simplest explanation for these results is that dissolution in water leads to dissociation of the dimeric structure observed in the crystal and that in solution a mixture of discrete monomer and dimer are present. However, it is important to keep in mind that the emission spectrum of Eu3+ affords information limited to the immediate coordination environment of the Eu3+ ion. In the event that a bridging phosphonate were replaced by a water molecule this change would most likely be reflected by a change in the emission spectrum of Eu3+ to a change in the spectrum and in particular the ΔJ = 0 transition. A simple monomer/dimer equilibrium would be expected to be characterized by a change in the relative intensities of the two ΔJ = 0 transitions that was directly dependent upon the chelate concentration.</p><p>The Eu3+ ΔJ = 0 transitions in the emission spectra of EuPCTMB were examined as a function of the concentration of the chelate in solution (Figure 8). Over the concentration range 4.3 × 10−1 M to 4.3 × 10−4 M two lines were observed for this transition. Of these, the line at 580.2 nm (ascribed to the dimeric structure, vide infra) is more intense at higher concentrations and, as the solution becomes more dilute, this line becomes relatively less intense. The relative intensity of the line at 579.8 nm (ascribed to a hydrated Eu3+ ion, vide infra) is found to increase as the solution becomes more dilute. However, the trend of these changes is non-linear with respect to chelate concentration which would seem to indicate that a complicated set of equilibria exist in solutions of LnPCTMB chelates. It is also worth noting that given the millisecond timescale of the 5D0-7F0 measurements, all observable species must be exchanging at rate slower than 100Hz.</p><p>Examination of the 5D0 → 7F0 transition in Eu3+ chelates is unlikely to be able to distinguish between the presence of a simple monomer/doubly bridged dimer equilibrium and a more complex equilibrium that also involves a singly bridged dimer (Chart 2). This latter scenario would involve a species that had two Eu3+ coordination environments neither of which could be readily distinguished from a monomer/doubly bridged dimer equilibrium. To probe the extent of dimerization in aqueous solution the Stern-Volmer quenching of EuPCTMB by NdPCTMB was assessed using a similar procedure to that reported by Morrow et. al.47 Stern-Volmer quenching constants of KSVτ = 1101 ± 61 M−1 and KSVΦ = 40780 ± 2531 M−1 were thus obtained. Given that the KSVτ is predominantly determined by dynamic, or collisional, quenching it is to be expected that this value would remain largely constant regardless of the extent of dimerization. KSVΦ is a reflection both static and dynamic quenching, the static component of which will only be present when dimerization causes an increase in the rate of excited state deactivation. Thus, in a purely monomeric system KSVΦ should have the same value as KSVτ; however, in this case the large disparity between the two values is a clear indicator of dimerization. From the extremely high KSVΦ value we can conclude that the extent of dimerization far exceeds the levels of doubly bridged dimer that one may expect from Figure 8 and that substantial quantities of singly bridged dimer are present in aqueous solution.</p><p>This conclusion is consistent with the relaxometric results. Care should be taken not to draw parallels that are too close between these two sets of experiments because different solvent systems were employed in each case; Senanayake et al. observed substantially different speciation of Ln2 upon addition of methanol to an aqueous solution of the chelate.17 Nonetheless, if a decrease in chelate concentration leads to progressively more singly bridged dimer, in which one Gd3+ ion is hydrated, then this would reasonably be expected to increase relaxivity. The relatively large hydrodynamic volume of this species would enhance τR, while allowing water access to the inner hydration sphere of one Gd3+ ion of the dimer would introduce an effectively large inner-sphere contribution to relaxivity. We may postulate that this is the origin of the concentration dependent relaxivity of GdPCTMB.</p><!><p>The Ln3+ chelates of PCTMB provide a fascinating insight into the behaviour of macrocyclic Ln3+ chelates and in particular those of triphosphonate-based ligands. In the crystal the chelates exist as doubly bridged dimers in which all water is excluded from the inner coordination sphere of the Ln3+ ion. This behaviour closely parallels that of related cyclen-based phosphinate systems, Ln2 and Ln3. However, in solution the behaviour is substantially different from the cyclen-based systems. Although Ln2 and Ln3 chelates are reported to exist exclusively as dimers in aqueous solution the speciation of solutions of LnPCTMB chelates is not only complex but highly concentration dependent. There can be little doubt that a certain amount of dimer dissociation occurs in solution, this gives rise to observation of a q = 1 species in the Tb3+ chelate, an additional line in the 5D0 → 7F0 transition of the Eu3+ emission spectrum and enhanced relaxivity of the Gd3+ chelate at low concentrations. The results of Stern-Volmer experiments show that in addition to mixture of monomer and doubly-bridged dimer, a substantial proportion of singly-bridged dimer (Chart 2) must also be present. Ultimately the considerable relaxivity afforded by the Gd3+ chelate and strong emissive properties of the Eu3+ and Tb3+ chelates of PCTMB, along with the known tumor targeting properties of these chelates,16 affords a system of considerable interest for further study in vivo imaging and therapy applications.</p><!><p>An ORTEP rendering of the crystal structure of EuPCTMB (50% ellipsoids, dimer partner omitted) viewed from the top (above) and the side (below). Hydrogen atoms and all but the closest water molecule of crystallization have been omitted for clarity. The hydrogen bonding interaction between the closest outer sphere water molecule and two phosphonate mono-esters are shown by dotted green lines. Considerable disorder may be observed in the butyl chains.</p><p>The conformations of cyclen (top) and pyclen (bottom) according to Dale's nomenclature.</p><p>The coordination geometry of a twisted snub dispheniod (TSD) adopted by the pyclen derivatives EuPCTMB and TbPCTMB (left) and the twisted square antiprism (TSAP) commonly adopted by methylene phosphonate and phosphinate derivatives of cyclen (right).</p><p>An ORTEP rendering of the crystal structure of Tb1 (50% ellipsoids) showing the dimeric nature of the chelate. Hydrogens have been omitted for clarity.</p><p>The conformation of the pendant arms in the Ln3+ chelates of PCTMB, 2 and 3, with the substituent gauche to the coordinating amine (left) and anti to the coordinating amine (right).</p><p>The dependence of the paramagnetic contribution to R1 upon the concentration of GdPCTMB is non-linear (20 MHz, 298 K).</p><p>The emission spectra of EuPCTMB (λex = 280 nm) in solution at 0.43 mM (top) and in the crystal (bottom). The inset in each spectrum shows the non-degenerate 5D0 → 7F0 transition, highlighting the number of species present (λex = 280 nm).</p><p>The 5D0 → 7F0 transition of the Eu3+ emission spectra (λex = 280 nm) of EuPCTMB in aqueous solution at: a) 4.3.×10−3M; b) 4.3×10−4M; c) 4.3×10−5M; d) 4.3×10−6M</p><p>The preparation of PCTMB from pyclen. Reagents and conditions: i. (CH2O)n/(nBuO)3P/THF; ii. KOH/1,4-dioxane/H2O; iii. LnCl3/H2O/pH 5.5.</p><p>Structures of the three components thought to make up the solution state equilibrium in aqueous EuPCTMB samples</p><p>A summary of crystallographic data for EuPCTMB and TbPCTMB</p><p>Selected bond lengths [Å], bond angles [°] and distances [Å] from the crystal structures of EuPCTMB and TbPCTMB. For comparative purposes the values for Eu2 and Tb3 have also been included from references 17 and 18, respectively. The N4- and O4-planes are the mean planes of the four nitrogen and four oxygen atoms, respectively.</p><p>Selected geometrical parameters of the structures of EuPCTMB, TbPCTMB and some related cyclen derived chelates. Atom labels are given according to the numbering scheme used for the structure of EuPCTMB</p><p>Data from reference 17</p><p>Data from reference 18</p><p>Data from reference 19</p><p>Data from reference 35; DOTA is 1,4,7,10-tetraazacyclododecane tetraacetic acid,</p><p>Data from reference 36; DOTAM is 14,7,10-tetraazacyclododecane tetraacetamide,</p><p>X = P, except for DOTA and DOTAM where X = C.</p><p>The hydration state determination of TbPCTMB in aqueous solution at 0.15 mM using an adaptation45 of Horrocks' method.43, 44 Data were collected using λex = 280 nm λem = 545 nm and fitted to double exponential model of decay to provide lifetime for each Tb3+ coordination environment.</p>

PubMed Author Manuscript

Comparison of Nine Programs Predicting pKa Values of Pharmaceutical Substances

Knowledge of the possible ionization states of a pharmaceutical substance, embodied in the pKa values (logarithm of the acid dissociation constant), is vital for understanding many properties essential to drug development. We compare nine commercially available or free programs for predicting ionization constants. Eight of these programs are based on empirical methods: ACD/pKa DB 12.0, ADME Boxes 4.9, ADMET Predictor 3.0, Epik 1.6, Marvin 5.1.4, Pallas pKalc Net 2.0, Pipeline Pilot 5.0, and SPARC 4.2; one program is based on a quantum chemical method: Jaguar 7.5. We compared their performances by applying them to 197 pharmaceutical substances with 261 carefully determined and highly reliable experimental pKa values from a literature source. The programs ADME Boxes 4.9, ACD/pKa DB 12.0, and SPARC 4.2 ranked as the top three with mean absolute deviations of 0.389, 0.478, and 0.651 and r2 values of 0.944, 0.908, and 0.894, respectively. ACD/pKa DB 12.0 predicted all sites, whereas ADME Boxes 4.9 and SPARC 4.2 failed to predict 5 and 18 sites, respectively. The performance of the quantum chemical-based program Jaguar 7.5 was not as expected, with a mean absolute deviation of 1.283 and an r2 value of 0.579, indicating the potential for further development of this type of approach to pKa prediction.

comparison_of_nine_programs_predicting_pka_values_of_pharmaceutical_substances

INTRODUCTION<!>DATA SET<!>COMPUTATIONAL METHODS<!>Execution Speeds.<!>Prediction Results.<!>ACD/pKa DB 12.0.<!>ADME Boxes 4.9.<!>ADMET Predictor 3.0.<!>Epik 1.6.<!>Jaguar 7.5.<!>Marvin 5.1.4.<!>Pallas pKalc Net 2.0.<!>Pipeline Pilot 5.0.<!>SPARC 4.2.<!>CONCLUSION

<p>Most pharmaceutical substances will be protonated or deprotonated in aqueous solution; for example, data from the 1999 World Drug Index suggest that 63% of the 51600 listed drugs are ionizable, of which 15% are acids, 67% bases, and 18% ampholytes.1 The ionization ability is quantified by a parameter, the (logarithm of the) acid ionization constant (pKa), which is also called the protonation constant, equilibrium constant, or (acid) dissociation constant in the literature. Along with the partition coefficient, solubility, and reaction rate, pKa is the most important physicochemical property of a substance to be formulated into a useful medicine. As a function of its intrinsic pKa value(s) and the pH value of the solution, the extent of ionization of a drug controls its solubility, dissolution rate, and consequently has great impact on gastrointestinal uptake into the bloodstream, distribution, cell permeability, drug–receptor binding, reaction kinetics, metabolism, elimination, etc.2,3 In the preformulation stage, knowledge of ionization constants is useful when trying to form a salt in order to obtain biopharmaceutical properties and solid-state characteristics that may be lacking in the free form of the compound.</p><p>pKa values can be either measured or calculated. There are a number of methods to use for the experimental determination of pKa values and closely related quantities such as pH values.4–6 These methods have been used extensively in drug discovery and various development stages in the pharmaceutical industry. The highly accurate measurement methods include conductance methods (reliable to ±0.0001 pK units or better) and electrochemical cells without liquid junction potentials (reliable to ±0.001 pK units or better), both of which, unfortunately, have been rarely used to measure ionization data for pharmaceutically relevant organic acids and bases. For this class of substances, most pKa measurements are based on relationships between the measured solution pH and a measured physicochemical quantity such as added titrant concentration, solubility, etc., which limits the expected accuracy. Measurement of pKa values has become easier and more convenient over recent years. Nevertheless, in early drug discovery, measuring millions of compounds in large screening libraries is costly and simply not practical, and outright impossible for virtual libraries, which makes in silico prediction of pKa values vital in modern drug discovery.</p><p>At present, standard methods for pKa prediction for pharmaceutical substances can be classified into two major groups: empirical methods and quantum chemical methods. On the basis of the detailed approach used, the empirical methods can be further divided into three groups: (1) linear free-energy relationships (LFER), methods utilizing the empirical relations of Hammett and Taft, (2) quantitative structure–property relationships (QSPR), methods correlating calculated structural descriptors with pKa values, and (3) database lookup, i.e., methods searching of similar structures in a predetermined database of molecules with known measured pKa values.7 One of the strengths of the empirical methods is their high speed, useful when processing large databases of drug-like molecules. The alternative to empirical methods, quantum chemical methods, are supposed to have higher accuracy because they are based on, or closer to, first principles when calculating quantum mechanical descriptors. However, these methods are much more time-consuming than empirical methods.</p><p>While a significant number of recent publications can be found reporting on new and better programs and methods for pKa prediction (see refs 8–16 and 17–21 for additional approaches using empirical or quantum chemical methods, respectively), albeit mostly for specific classes of compounds, much less prior work exists comparing such approaches. Melun et al. used the REGDIA regression diagnostics algorithm in S-Plus to examine the accuracy of pKa predictions of four programs: ACD/pKa, Marvin, PALLAS, and SPARC.22 Three different validation data sets were taken from literature, including 64 pKa values for different drugs. We therefore felt there was the need to not only compare a larger number of available programs but also use more pKa values for pharmaceutical substances in order to cover as much as possible of the drug-like chemical space. In this article, we use eight empirical programs and one quantum chemical program (Table 1) to predict 261 carefully determined pKa values of 197 pharmaceutical substances in order to compare the predictive power of the nine programs.</p><p>While we have tried to bring together as many of the existing programs that have some usage, we do not claim that we have tested each and every code that predicts pKa values. One program, for example, that we are aware of but which was not included is MoKa,13 marketed by Molecular Discovery, Ltd.35 Unfortunately, a mutually satisfactory agreement to gain access to the program for the purpose of this study could not be reached with the company.</p><!><p>Significant numbers of experimental pKa data on aqueous ionization chemicals have been collected over the decades. Some of them have found their way into literature compilations36–40 which, however, are not focused on pharmaceutical substances, although they do include some such compounds. In 2007, a book titled Profiles of Drug Substances, Excipients, and Related Methodology: Critical Compilation of pKa Values for Pharmaceutical Substances was published.41 The author of this book systematically collected nearly 3500 reported pKa values for drugs and related compounds from the pertinent primary and secondary literature and then, using the IUPAC classification and guidelines given below, assessed the reliability of these reported pKa values. On the basis of the aspects of a pKa measurement, such as the experimental method, mathematical definition selected to calculate the value from the raw data, and degree to which technical refinements have been applied, the IUPAC established in the 1960s the criteria for its compilations of dissociation constants for weak organic acids and bases: very reliable (VR; pKa error < ±0.005), reliable (R; pKa error ±0.005 to ±0.02), approximate (A; pKa error ±0.02 to ±0.04), and uncertain (U; pKa, error > ±0.04).36,37 The author of this book set the cutoff for uncertain to > ±0.06 pKa unit. On the basis of this modified IUPAC criteria, about 74% of the collected pKa, values in this book were found to be of uncertain quality, whereas only 0.1% qualified as very reliable and 0.33% qualified as reliable, which left ~25% being classified as approximate. The compilation has two sections: Appendix A comprises pKa values for which the measurements were sufficiently well described for the data to be assessed for reliability, and Appendix B comprises those pKa values for which little reliability data could be assessed, which were mostly from the secondary literature.</p><p>For this project, the pKa values used to compare the abilities of the nine programs to predict pKa values for the nine programs were chosen according to the following criteria: (1) Only pKa values that came from Appendix A of the above-mentioned reference were chosen. (2) The data qualities had to be VR, R, or A. (3) The solvent contained only water except some possible inorganic ions. (4) The temperature at which the measurement was taken was in the range of 25 ± 2 °C (except for compound 134, measured at 20 °C, which was included because otherwise we would have lost the only thiol in the final set). These quite rigorous criteria finally left us with 261 pKa values measured for 197 compounds whose structures are shown in Figure 1. The molecular weight and pKa value distributions and numbers of acidic and basic sites of these compounds are shown in Figure 2 and Table 2, respectively, both of which demonstrate that, as far as chemical species, number, and value distribution are concerned, this data set can be considered high-quality for the purpose of this comparison. A cautionary note is warranted for compound 99, which may be capable of forming aggregates42,43 at high concentration because this compound likely has low solubility in water as a monomer. This might then affect the measured pKa values. Because firm evidence exists neither for nor against aggregation, however, we decided to keep it in our compound set (hoping that the authors of the original paper were aware of this issue and performed the necessary procedures to rule it out or prevent it).</p><p>It is worthwhile to point out that although this data set has dozens of multiprotic molecules, not all of the acidic and basic sites of every multiprotic molecule were predicted in this comparison. The main reason is that the book does not note every pKa value of every multiprotic molecule or that the data quality of some of these sites is below A.</p><p>It should also be pointed out and kept in mind by the reader, that these programs tend to be trained with as much of the data available in the literature as possible. Testing these programs using exclusively information from the literature, as we unavoidably had to do for this study, therefore entails the risk that the programs may effectively "look up" known pKa values rather than predict them. It would therefore not be a wrong strategy for the serious user to test any of these programs themselves using unpublished or private pKa data if possible.</p><!><p>All eight programs based on empirical methods were executed one compound at a time or in batch mode. Some programs such as ACD/pKa, DB, ADME Boxes, Epik, Marvin, and SPARC can calculate different kinds of pKa values. For this study, only the pKa types that were designated as predicting experimental pKa values were calculated. ACD/pKa DB 12.00, ADME Boxes 4.9, ADMET Predictor 3.0, Marvin 5.1.4, Pallas pKalc Net 2.0, and Pipeline Pilot 5.0 were run using the default options in their respective graphical user interfaces on a Windows XP computer. SPARC 4.2 was run via a web-based interface (IE 7.0) on a Windows XP machine. Epik 1.6 was run via command line on a Linux system.</p><p>pKa values are in direct proportion to ΔG°, the free energy change for transition from the protonated state to the deprotonated state. A small calculation error of ΔG° (on the order of a few kcal/mol) can therefore lead to a significant prediction error for pKa for programs that are based on quantum–chemical methods. To correct for this deficiency that affects the quantum chemistry-based program Jaguar, this program employs two additional empirical parameters, scaling and additive factors.</p><p>pKa predictions of Jaguar 7.5 consist of a series of calculations on the protonated and deprotonated forms of the target molecule, followed by the aforementioned empirical correction. Because the calculated results partly depend on the conformation of the target molecule, first a conformational search was performed with MacroModel.44 As recoimnended by Schrödinger, the lowest-energy conformers found were used for further pKa calculations. Jaguar calculates microscopic (atomic) pKa values but not macroscopic (experimental) pKa values. If two or more microscopic pKa values lie within one pKa unit of each other, the macroscopic pKa values can markedly differ from the corresponding microscopic values. Therefore, for such a multiprotic molecule, in order to obtain the macroscopic pKa values, Schrödinger suggests in the Jaguar manual to run 2n states (n being the number of close pKa values in a multiprotic molecule) and then to assemble the titration curve. Given that each QM computation takes already orders of magnitude longer than the corresponding empirical calculation, this would have heavily increased the amount of calculation necessary for the project while creating hard-to-assess additional sources of potential error. We therefore decided to drop the calculation of such protonation sites. We set the cutoff here for two experimental pKa values in a multiprotic molecule being too close to each other if their difference was less than or equal to 2.5 pKa units. This led to a reduction of the number of calculated sites from 261 to 204. For some kinds of acidic or basic sites, Jaguar 7.5 does not have parameters for an aqueous solution, which led to another 11 sites being dropped to yield the final number of 193 sites included in the calculations. For those multiprotic molecules whose experimental pKa values are well separated (by >2.5 pKa units), when calculating the lowest pKa value, the sites with higher pKa values were in the protonated states; when calculating the middle pKa value, the sites with higher and lower pKa values were in the protonated and deprotonated states, respectively; and when calculating the highest pKa value, the sites with lower pKa values were in the deprotonated states.</p><p>Among the 197 compounds, some molecules, for example, compounds 79, 128, and 164, have two equivalent sites for protonation or deprotonation. In this situation, the need for a statistical correction factor arises from the increased entropy of the appropriate species. A correction of +0.60 (log1022) or −0.60 was added by hand to the result obtained from running the pKa prediction module on the basis of whether the calculated molecule has two equivalent acidic sites or basic sites because Jaguar 7.5 does not automatically recognize equivalent sites.</p><!><p>Program execution was very fast for all programs based on empirical methods except SPARC 4.2. For example, ADMET Predictor 3.0 and Pipeline Pilot 5.0 finished the calculation of predicting the pKa values of these 197 pharmaceutical substances in less than 1 s on a Windows computer. Epik 1.6 took 119 s on one CPU (AMD 64, dual core FX-61) of our Linux cluster. It is therefore possible to predict pKa values of millions of compounds in a tractable time by using these seven programs. The calculation speed of SPARC 4.2, because of its use of the PMO theory, was much slower than that of the other seven programs. For example, submission of compound 74, which has five protonation sites, resulted in the following message: "This will result in ~5120 calculations and may take as long as 51.87 minutes". In fact, the server did not produce any result for this compound but an error message: "The request has exceeded the allowable time limit Tag". Yet for simpler compounds with one or two protonation sites, the calculation time was acceptable. One-by-one submission and longer execution times make SPARC 4.2 not suitable for predicting pKa values of large number of compounds.</p><p>The pKa prediction of Jaguar 7.5 was very time-consuming, and it was strongly dependent on the size and flexibility of the molecule. For examples, on the above-mentioned AMD 64, dual core FX-61 CPU, it took about 4, 23, 96, 202, 1568, and 3831 min, respectively, to predict pKa values of the carboxylic acid sites in compounds 3 and 148, the barbituric acid site in compound 36, the pyridine site in compound 145, the carboxylic acid site in compound 104, and the tertiary amine site in compound 116. The molecular weights of these six compounds in the same order are 46.03, 123.11, 226.28, 282.22, 573.67, and 592.69, respectively. After more than three days of computation, Jaguar 7.5 failed for an unknown reason (but presumably due to resource exhaustion) in the prediction of the primary amine site of compound 99, which is the largest and most flexible molecule in the test set with a molecular weight of 744.05. With such slow calculation speeds, we can conclude that Jaguar 7.5 is not a practical solution to handle many compounds, especially when they are large, flexible, and/or multiprotic with close pKa values.</p><!><p>The prediction results for the 261 pKa protonation sites are shown in Table S1 of the Supporting Information and Figure 3 and summarized in Table 3 and Figure 4. Only ACD/pKa DB 12.0, ADMET Predictor 3.0, and Marvin 5.1.4 predicted all 261 protonation sites; ADME Boxes 4.9, Epik 1.6, Pallas pKalc Net 2.0, Pipeline Pilot 5.0 and SPARC 4.2 failed for 5, 4, 2, 9, and 18 sites, respectively. Jaguar 7.5 failed for 11 sites with the error message "No pKa functional group with parameters for water could be identified".</p><p>When based on mean absolute deviation (MAD), the rank order of this comparison is ADME Boxes 4.9, ACD/pKa DB 12.0, SPARC 4.2, ADMET Predictor 3.0, Pipeline Pilot 5.0, Pallas pKalc Net 2.0, Marvin 5.1.4, Epik 1.6, and Jaguar 7.5. In terms of r2, the situation changes a little, but the top three are still ADME Boxes 4.9, ACD/pKa DB 12.0, and SPARC 4.2. It might be argued that this criteria is actually unfair for the three programs that predicted all protonation sites.</p><p>When predicting pKa values as part of the drug discovery process, researchers may typically be interested mostly in the protonation state at physiological pH, i.e., 7.4. Table 4 shows the performance of these nine programs, specifically for those 116 sites whose measured pKa values are in the range of 5.4–9.4. In general, on the basis of the mean absolute deviation, most programs performed worse in this range than for the full set of sites, except Jaguar 7.5 and Pallas pKalc Net 2.0. The top three performers in terms of MAD were ADME Boxes 4.9, ACD/pKa DB 12.00, and Pallas pKalc Net 2.0; and ADME Boxes 4.9, ACD/pKa DB 12.00, and SPARC 4.2 in terms of r2.</p><p>About thirty of the compounds were consistently predicted poorly, i.e., more than six of the programs predicted them with an deviation of at least 0.5 log units or even failed to predict one or more sites of them altogether: 6, 9, 52, 53, 60, 62, 67–69, 74–76, 79, 85, 99, 102, 110, 119, 121, 128, 129, 134, 144, 155, 156, 174, 175, 178, and 188–190. It is not clear whether this points to a general weakness in the understanding and/or algorithms in the field for these molecules or if this may indicate potential problems with the experimental results.</p><!><p>AC D/pKa DB 12.0 ranks second in this comparison with an MAD of 0.478 and an r2 of 0.908. A strong point of ACD/pKa DB was that it calculated all 261 protonation sites. It produced 64 (24.52%) and 185 (70.88%) predicted values with accuracies of ±0.1 and ±0.5 log unit, respectively. This makes it a well-built and robust program for the prediction of ionization states of pharmaceutical substances. Nevertheless, there are still 10 sites (3.83%) whose predicted accuracies were more than 2.0. The two largest deviations are 7.07 and −4.05 log units, which occurred on the general C substituted amide site of compound 6 and the substituted aniline site of compound 93, respectively. For the tertiary amine sites of the three tetracycline antibiotics (compounds 74–76), the prediction errors are significant: all of them are around +3.50 log units. The other very poorly predicted sites include the primary amine sites of three penicilloic acids (compounds 67–69), the phenol site of ebifuramin (compound 102), and the benzodiazepine site of compound 156.</p><!><p>This program ranks first if one only focuses on the MAD (0.389) or r2 (0.944) values. However, it failed to predict five protonation sites: one enol site of compound 23 (another enol site was successfully predicted), one acid site of compound 60 (which actually is an inorganic acid), the substituted aniline site of compound 93, and the two heterocycle acid sites of compounds 118 and 119. Apart from these cases, ADME Box 4.9 predicted normal organic compounds very well: it produced 95 (37.11%) and 197 (76.95%) predicted values with accuracies of ±0.1 and ±0.5 log unit, respectively. Each of these two values is at the top of its list for the nine programs. There are seven sites (2.73%) whose prediction accuracies were worse than 2.0 log units, which also represents the top spot among the nine compared programs. The largest deviation was −3.5, which occurred for one of the two acid sites of carbonic acid (compound 62). The other very poorly predicted six sites include the three phenol sites of the three tetracycline antibiotic compounds 74 – 76, which all were predicted higher than experimentally measured, one tertiary amine site of the antipsychotic compound 85, the phenol site of compound 102, and one carboxylic acid site of compound 128.</p><!><p>This program ranks fourth with a MAD of 0.659 and an r2 of 0.837. It also is one of the three programs that did not fail to predict even one site. It gave 40 (15.33%) and 159 (60.92%) predicted values with accuracies of ±0.1 and ±0.5 log unit, respectively. There were 18 sites (6.90%) whose predicted accuracies were more than 2.0 log units. The calculated largest deviation is −7.86, which occurred for compound 60. Besides this site, there are 17 additional rather poorly predicted sites: both of the two acid sites of carbonic acid, compound 62; the three primary amine sites of the three penicilloic acids (compounds 67–69); the three enol sites, and the three tertiary amine sites of compounds 74–76 (whereas the three phenol sites of these three compounds were predicted quite well, with absolute deviations between 0.24 and 0.79); the substituted aniline site of compound 93; the thiol site of compound 134; the pyridine site of compound 145; the primary amine sites of compounds 155 and the 175; and the heterocycle acid site of compound 190.</p><!><p>The r2 and MAD values of this program were 0.802 and 0.893, respectively, which represents the sixth and last rank, respectively, among the eight programs utilizing empirical methods. The prediction of two protonation sites, the acid site of the inorganic acid compound 60, and the thiol site of compound 134 were not completed by this program. The enol sites of compounds 75 and 102 were tautomerized into carbonyl groups, which are not protonation sites (it is interesting to note that the enol sites of compounds 74 and 76 were not tautomerized in the pKa prediction process by the program, although the structures of the compounds 74–76 are almost identical to each other). All in all, Epik predicted 257 sites and failed for four sites. It produced 30 (11.67%) and 117 (45.52%) predicted values with accuracies of ±0.1 and ±0.5 log unit, respectively. There were 23 sites (8.95%) whose predicted accuracies were off by more than 2.0 log units. The largest deviation was 6.26, which occurred for the CH acid site of compound 157. The other very poorly predicted protonation sites comprised both of the two acid sites of carbonic acid, compound 62; the three primary amine sites of the three penicilloic acids (compounds 67–69); the three guanidine sites of compounds 70–72; the two tertiary amine sites of compounds 74 and 76; the phenol sites of compounds 81, 144, and 176; the tertiary amine site of compound 88; the substituted aniline site of compound 93; the primary amine sites of compounds 9 and 99; the secondary amine site of compound 103; the carboxylic acid site of compound 109; the imine site of compound 151; the heterocycle site of compound 156; and the heterocycle acid site of compound 190.</p><!><p>Even though based on a more sophisticated fundamental approach than the other eight programs, which are based on empirical methods, Jaguar 7.5 did not pull ahead of any one of them on the basis of our analysis. Only 193 protonation sites were predicted. One reason is the lack of parameters for some less frequently occurring sites, the other is the closeness of the pKa values for some sites in a multiprotic molecule. The r2 value for Jaguar 7.5 was 0.579, 0.178 lower than the worst-performing one of the other eight programs. Likewise, its MAD was 1.283, 0.390 higher than the next-ranked program. Jaguar 7.5 produced 24 (12.44%) and 62 (32.12%) predicted values with accuracies of ±0.1 and ±0.5 log units, respectively. The 24 sites whose predicted values were very close to the best literature values include the barbituric acid site of compound 37, the carboxylic acid site of compound 77, the phenol site of compound 81, the secondary amine sites of compounds 87, 105, and 131, the tertiary amine site of compound 114, the heterocycle site and the primary amine site of compound 118, one carboxylic acid site of compound 130, and several others. One strong point of Jaguar 7.5 over the other eight programs is that it can distinguish diastereomers. For example, compounds 152 and 153 are diastereomers of each other. The experimental and predicted pKa values for the two primary amine sites of compound 152 are 9.05 and 9.0, respectively, and for compound 153, they are 9.19 and 9.3, respectively. Nevertheless, this version of Jaguar would not appear to be an ideal tool for predicting pKa values of pharmaceutical substances. There were 40 sites (20.73%) whose predicted accuracies were off by more than 2.0 log units. Frequently, Jaguar 7.5 worked incorrectly when there were one or more charged groups in the molecule, even though this molecule is not multiprotic. For example, for the three antibiotic compounds 15, 18, and 20, which are monoprotic, the predictions for the three carboxylic acid sites were acceptable: The calculated deviations were 0.4, 0.7, and 0.9 respectively. However, the deviations for the other three antibiotic compounds 16, 17, 19 were 4.5, 5.7, and 6.3, which are clearly not acceptable. Each of these three compounds has an −NH2 group, which we set as protonated and then maintained positively charged for the prediction of the pKa value of the carboxylic acid site. When calculating the pKa value of the second acid group of carbonic acid (compound 62), the first carboxylic acid was deprotonated. This led to the pKa value being predicted as 1.6, which is even lower than the first calculated one and then brings the deviation to −8.8. The compounds 89–92 are acids with quaternary ammonium groups; however, they are monoprotic, not diprotic. The positive quaternary ammonium groups induced big deviations for every carboxylic acid site: 3.3, 4.2, 5.7, and 4.4, respectively. In fact, among the 40 very poorly predicted sites (>2.0 log units), 29 were calculated when a charge existed in the molecule. Schrödinger's explanation in the Jaguar manual acknowledges this shortcoming: "When the ionziable groups are close together in the molecule, the calculated pKa may not be as accurate because the two groups could interact in ways that the existing parameterization cannot handle."</p><!><p>With no sites with failed prediction attempts, the r2 and MAD values for the program were 0.763 and 0.872, respectively, which makes this program rank seventh. It produced 39 (14.94%) and 130 (49.81%) predicted values with accuracies of ±0.1 and ±0.5 log units, respectively. Marvin calculated barbituric acid sites relatively poorly. The average deviation for the 29 barbituric acid sites was 1.254; two deviations were more than 2.0 log units. There were 22 sites (8.43%) whose predicted accuracies were more than 2.0 log units. The worst five predicted deviations were −9.88, 6.93, −5.88, 5.34, and −4.08, which happened on the substituted aniline site of compound 93, the primary amine site of compound 26, the pyridine site of compound 159, the carboxylic acid site of compound 26, and the tertiary amine site of compound 10, respectively. The other very poorly estimated sites include the four guanidine sites of the four biguanidine compounds 70–73, the primary amine site of compound 9, one enol site of compound 23, the phenol sites of compounds 75 and 102, one tertiary amine site of compound 85, the tertiary amine site of compound 115, the pyridine site of compound 149, the benzodiazepine site of compound 156, the enol site of compound 166, the imide site of compound 179, and the heterocycle site of compound 190.</p><!><p>This program failed to predict two sites: the acid site of compound 60 and the enol site of compound 166. With this, it garnered an r2 value of 0.803 and an MAD value 0.787, which helped it rank fifth and sixth, respectively, among the nine programs. It produced 48 (18.53%) and 153 (59.07%) predicted values with accuracies of ±0.1 and ±0.5 log units, respectively. There were 26 sites (10.04%) whose calculated deviations were more than 2.0 log units, among which the worst three predicted deviations occurred on the enol sites of the antibiotic compounds 74–76. The other 23 very poorly predicted protonation sites included the alcohol sites of compound 5, the general C substituted amide site and the heterocycle site of compound 6, the primary amine site of compound 9, one acid site of compound 62, the tertiary amine sites of compounds 65 and 66, the guanidine site of compound 70, the phenol sites of compound 81, 102, and 176, the substituted aniline site of compound 93, the secondary amine sites of compounds 103, 123, and 188, one carboxylic acid site of compound 128, the thiol site of compound 134, the pyridine site of compound 145, the quinolin-2-one site of compound 151, the pyridine sites of compounds 174 and 175, the imide site of compound 179, and the heterocycle site of compound 190.</p><!><p>With nine failed sites, the r2 and MAD values of this program were 0.757 and 0.769, respectively, which made it rank eighth and fifth in these two categories. The failed nine sites were the carboxylic acid site of compound 3, the alcohol sites of compound 5, the two sites of compound 6, the heterocycle sites of compounds 27, 118, 119, and 190, and the quinolin-2-one site of compound 151. Pipeline Pilot 5.0 produced 77 (29.62%) and 151 (57.85%) predicted values with accuracies of ±0.1 and ±0.5 log units, respectively. The quality of the predictions of Pipeline Pilot for the pKa values of the barbituric acids varied strongly. Some compounds, such as 36 and 56, were predicted well, with deviations close or equal to zero, whereas other barbituric acids were predicted very poorly. For example, the deviations for compounds 30 and 52 were −4.62 and −4.96, respectively. The predicted values of the two enol sites of compound 23 were very close to the experimental values; unfortunately, Pipeline Pilot assigned them to the wrong sites, which led to the two largest deviations, 7.43 and −7.22. Besides these four sites, there were 20 sites whose predicted accuracies were off by more than 2.0 log units. These were the carboxylic acid sites of compounds 9, 79 and 149, the barbituric acid sites of compounds 53 and 54, the tertiary amine sites of compounds 61, 74, and 183, the two acid sites of compound 62, the primary amine site of compound 99, the phenol sites of compound 102, 106–108, one tertiary amine site of compound 120, the thiol site of compound 134, the pyridine site of compound 145, the heterocycle site of compound 156, and the sulfonamide site of compound 185.</p><!><p>This program gave up on the prediction of 18 protonation sites, mostly because it exceeded its time limit. It ranks third with an r2 value of 0.894 and an MAD value of 0.651 in this comparison. This program produced 30 (12.35%) and 126 (51.85%) predicted values with accuracies of ±0.1 and ±0.5 log units, respectively. These two values are actually the second worst ones in each category among the eight empirical methods; nevertheless, this program did not produce too exorbitantly bad deviations. Only 12 deviations (4.94%) were larger than 2.0 log units. The worst two were −3.11 and 3.10, which happened for one acid site of compound 62 and the benzodiazepine site of compound 156. This would seem to confirm that the PMO method is effective in preventing big deviations when it is used to predict pKa values of pharmaceutical substances, albeit at the cost of prolonging calculation times. The other 10 poorly predicted sites involve the primary amine sites of compounds 9 and 99, the guanidine site of compound 82, one tertiary amine site of compound 85, one carboxylic acid site of compound 128, the phenol site of compound 129, the tertiary amine sites of compounds 139 and 140, the enol site of compound 166, and the heterocycle site of compound 190.</p><!><p>Predicting pKa values of pharmaceutical substances is both challenging and important in drug development and thus is an intriguing task in computational chemistry. We have compared nine currently available programs, including eight based on empirical methods and one based on a quantum chemical approach as to their ability to accurately predict 261 carefully experimentally measured pKa values of 197 pharmaceutical compounds. It had been suggested that an approach based on a quantum chemical method would have led to better predictions, but our study did not bear this out. On the contrary, the only program in this comparison based on a higher level of theory than empirical methods ranked last in essentially all respects, indicating that the more recently introduced quantum chemical approach for predicting pKa values has not yet reached the maturity level of the empirical methods.</p><p>Among the eight programs that are based on empirical methods, we found several that performed very well with this test set, with two predicting all or nearly all protonation sites, and doing so with an r2 value of better than 0.9 and a mean absolute deviation of less than half a log unit. While extrapolation to any possible compound of interest in drug development is obviously risky, we believe that the best pKa predicting programs currently available are useful tools in the arsenal of the drug developer.</p>

PubMed Author Manuscript

Generation of Nucleophilic Chromium Acetylides from gem-Trichloroalkanes and Chromium Chloride: Synthesis of Propargyl Alcohols

Nucleophilic mixed chromium(II) and chromium(III) acetylides are generated from the smooth reduction of primary 1,1,1-trichloroalkanes with chromium(II) chloride in the presence of an excess amount of triethylamine at room temperature. These species arise from chromium(III) vinylidene carbenoids. It has been demonstrated that uncommon low-valent CrII acetylides are formed by C\xe2\x80\x93H insertion of CrIICl2 into terminal alkynes, formed in situ through the Fritsch\xe2\x80\x93 Buttenberg\xe2\x80\x93Wiechell (FBW) rearrangement, whereas CrIII acetylides are concomitantly generated by HCl elimination from the chromium(III) vinylidene carbenoid. Both divergent pathways result, overall, in the formation of nucleophilic acetylides. In situ trapping with electrophilic aldehydes afforded propargyl alcohols. Furthermore, deuteration experiments and the use of deuterium labeled 1,1,1-trichloroalkane substrates demonstrated the prevalence of low-valent CrII acetylides, potentially useful, yet highly elusive synthetic intermediates.

generation_of_nucleophilic_chromium_acetylides_from_gem-trichloroalkanes_and_chromium_chloride:_synt

Introduction<!>Results and Discussion<!>Conclusions<!>General<!>General Procedure for the Generation of Chromium Acetylides and Their Reaction with Aldehydes<!>1-[4-(1-hydroxy-3-phenylprop-2-ynyl)phenyl]ethanone (3d)<!>1-(4-Methoxyphenyl)-3-(p-tolyl)prop-2-yn-1-ol (3g)<!>1-{4-[1-hydroxy-3-(4-methylphenyl)prop-2-ynyl]phenyl}ethanone (3h)<!>1-(4-Fluorophenyl)-3-(p-tolyl)prop-2-yn-1-ol (3i)<!>

<p>Since their discovery in 1957, organochromium reagents have been the focus of constant development and innovations.[1] Due to their unique combination of chemical features and remarkable compatibility with a wide range of functional groups, these reagents have become indispensable tools for advanced organic synthesis and for natural product synthesis.[2] The last 10 years, in particular, have witnessed an enormous growth in terms of new reagents and reaction modalities.[3–11] For instance, our laboratories and others have described several new chromium intermediates including chromium vinylidene carbenoids,[4] halogenated chromium enolates,[6] chromium Fischer halocarbenes,[5,7,11] and carbynes.[9] In continuation of our harvest of novel intermediates with unusual reactivities formed through the chromium-mediated reduction of 1,1,1-trihaloalkanes, we report herein that the reduction of 1,1,1-trichloroalkanes by chromium(II) chloride in the presence of triethylamine (TEA) induces the smooth formation of alkynylchromium reagents in high yields under exceptionally mild reaction conditions. Interception of the alkynylchromium intermediates with electrophilic aldehydes provides convenient access to functionalized propargyl alcohols (Scheme 1).[12]</p><p>Heretofore, the generation of metal acetylides under mild conditions compatible with various functional groups has long been an unresolved synthetic challenge.[13] Classical methods have mainly exploited the relatively high acidity of terminal acetylenic C–H bonds to form metal alkynylides, either by direct metalation using strong bases, such as n-butyllithium or lithium diisopropylamide at low temperature (−100 to −80 °C),[14] or upon treatment with tertiary amines in the presence of a stoichiometric or catalytic amount of the metal salt of interest (Figure 1, route a).[15] Lithium and silver acetylides prepared by this approach are also utilized for the preparation of other acetylides by transmetalation with magnesium, zinc, cerium, and other metals (Figure 1, route b).[16–20] Alternatively, lithium acetylides can also be prepared through the Fritsch– Buttenberg–Wiechell (FBW) rearrangement/metalation of 1,1-dibromoolefins when treated with an excess amount of n-butyllithium (Figure 1, route d).[20,21] An in situ metalation/desilylation strategy has recently been applied successfully to the preparation of highly stable ruthenium acetylides of interest for their electronic properties (Figure 1, route c).[22]</p><p>Chromium(III) acetylides, that are mainly generated by reduction of alkynyl halides with chromium(II) chloride (Figure 1, route e),[23] or more recently by transmetalation of lithium acetylides (Figure 1, route b),[24] have received scant attention, in spite of their demonstrated synthetic utility in a great number of natural product total syntheses[25] and for their interesting electronic properties.[26]</p><!><p>The generation of alkynylchromium(III) reagents by the most widely used reductive route (Figure 1, route e) is, however, plagued by drawbacks, inter alia, dependency upon nickel(II) additives, polar solvents, and/or high reaction temperatures.[1c] In sharp contrast, we have discovered that alkynylchromiums can be easily synthesized from the reduction of gem-1,1,1-trihaloalkanes by using 6 equiv. of CrCl2 and 10 equiv. of TEA (Figure 1, route f; Scheme 2).</p><p>The postulated mechanism for the formation of chromium acetylides likely proceeds through a comparatively stable chromium(III) vinylidene carbenoid 6, generated through the syn-β-elimination of chromium hydride from the unstable 1-chloro-1,1-bis-chromium alkane carbenoid 5, initially formed by the reduction of two C-Cl bonds in 1,1,1-trichloralkane 4 (Scheme 2).[4a] Subsequent β-elimination of hydrogen chloride induced by Et3N abstraction of the vinylic proton of 6 gives rise to chromium(III) acetylide 7 (Scheme 2, path a).</p><p>Surprisingly, whereas DBU, pyridine, 1,5,7-triazabicyclo-[4.4.0]dec-5-ene (TBD), and DABCO were completely inactive and even inhibited the reduction of trichloroalkane 4, TEA was unique amongst the common organic bases and did not interfere with the overall transformations of gem-1,1,1-trichloroalkanes. The role of TEA is not known yet, although it is assumed that the dramatic decrease in the pKa of the carbenoid vinylic proton can be ascribed to the concerted metal-assisted ionization phenomenon (MAI),[27] which triggers the formation of chromium(III) acetylide 7 as shown in postulated binuclear complex 13 (Scheme 3).</p><p>Hydrogen abstraction from 13 (Scheme 2, path a) is assumed to be kinetically competitive with the FBW rearrangement that leads to terminal alkyne 9 (Scheme 2, path b).[28] We postulate that this unprecedented reactivity of chromium vinylidene carbenoids is a result of coordination of the σ-donor lone pair of the basic nitrogen to chromium(III), yielding chromium(III) acetylides 7. This mechanistic pathway was partially corroborated by the fact that (Z)-2-chloroalk-2-en-l-ols 14 were formed in low amounts (4–10%) as byproducts when the reaction was performed under Barbier conditions in the presence of various aldehydes (Scheme 4).</p><p>The high yield of propargyl alcohol 3, combined with the ready availability of 1,1,1-trichloroalkanes la–d, make this methodology very attractive for the preparation of a large panel of propargyl alcohols (Table 1).[29] Aromatic aldehydes 12a–d and 12g–i bearing diverse, sensitive functional groups such as bromo, cyano, acetoxy, methoxy, and fluoro, were well tolerated and afforded expected adducts 3a– d[30,31] and 3g–i[32] in good to excellent yields (Table 1, Entries 1–4, 7–9). Similarly, an α,β-unsaturated aldehyde like (E)-cinnamaldehyde (12f) reacted smoothly and delivered cleanly the corresponding propargyl allyl alcohol 3f[30] in a good isolated yield of 63% (Table 1, Entry 6). In sharp contrast, the use of an enolizable aliphatic aldehyde like dihydrocinnamaldehyde (12e) under the same conditions was problematic, and the yield of addition product 3e[33] was drastically decreased to 22% because of competitive cross-aldol as well as elimination. (Table 1, Entry 5).[34] Primary 1,1,1-trichloroalkanes lc reacted moderately and resulted in the formation of 3j[35] in reasonable isolated yield (56%) when benzaldehyde (12j) was used as the electrophile (Table 1, Entry 10). It is worth mentioning that the use of allylic 1,1,1-trichloroalkane 1d offered the opportunity to extend the scope of this transformation in generating useful enyn–alcohol 3k[36] in excellent yield (Table 1, Entry 11). Attempts to reduce the amount of chromium reagent by using multicomponent redox system for chromium recycling [CrCl2 (10 mol-%)/Mn0/TMSCl][37] and TEA (10 equiv.) in THF at room temperature for 12 h were not satisfactory as a result of limited conversion of the starting 1,1,1-trichloroalkanes (<10%). This result suggests, most likely, that the initial generation of key vinylidene carbenoid 6 (Scheme 2) might be the rate-limiting step of the process under these specific conditions.</p><p>Mechanistically, to determine whether terminal alkyne 9, generated by FBW rearrangement from 6, could eventually be an intermediate in the overall transformation (Scheme 2, path b), we examined the reactivity of terminal alkynes with chromium(III) and chromium(II) chloride in the presence or absence of TEA. Notably, as a control experiment, 1-phenylacetylene (11) does not react with chromium(III) chloride in the presence or in the absence of TEA in THF under the same conditions as 1, thus excluding pathway a in the formation of chromium(III) acetylide 7a (Scheme 5).</p><p>However, we were surprised by the fact that CrIICl2 reacted smoothly with 11 in the presence of TEA (Scheme 5, path b), as evidenced by the formation of a minimum of 60% of adduct 16 when the organometallic species was trapped with benzaldehyde under Barbier conditions. Because the direct insertion of chromium(III) into the C–H bond of terminal alkynes is excluded for the formation of chromium(III) acetylide 7a, these results suggest strongly that the nucleophilic metalated acetylide is the uncommon low-valent chromium(II) acetylide 15. This result might be explained by ligand exchange of CrII, allowing nucleophilic substitution of labile ligands (e.g., Cl) to give nucleophilic chromium(II) acetylide 15. Indeed, like ZnII, CuI, or AuI acetylides that are generated in situ from terminal alkynes at room temperature upon treatment with an organic base (TEA, iPr2NnPr, or NH4OH)[13,15c,15d,29] by ligand exchange, this substitution reaction occurs for CrII. This mechanism is supported by kinetic studies reported by Merbach, who showed that this ligand exchange is kinetically very fast and favored for CrII, whereas CrIII is known to be extremely resistant to this process.[38] Indeed, the exchange ligand rates are ca. 15 orders of magnitude higher for CrII than CrIII.[39] The synthesis of end-bound acetylide ligands with monovalent and divalent octahedral chromium(II) has been recently reported; however, to the best of our knowledge, there is no report that accounts for their reactivity towards C–C bond-forming reactions.[26] At this point in our investigation, we provided evidence that the reduction of 1,1,1-trichloroalkanes with chromium(II) chloride in the presence of TEA affords unprecedented mixed chromium(II) and chromium(III) acetylides through two divergent pathways from chromium(III) vinylidene carbenoid 6 (Scheme 2). To further distinguish the prevalence of one pathway with respect to the other, we examined the reduction of deuterated substrate 17 and performed additional deuteration experiments (Scheme 6). Substrate 17 was first treated with 4 equiv. of CrCl2 under the same experimental conditions as those outlined for compounds 1. Quenching the reaction with H2O afforded a 2:3 mixture of protonated terminal alkyne 18 and its deuterated analogue 19 (FBW product) as determined by quantitative GC–MS analysis.</p><p>Interestingly, this result was corroborated by the reduction of 1a with 4 equiv. of CrCl2,[4a] in the presence of an excess amount of TEA, and subsequent deuteration with DCl in D2O, which yielded 30–35% of [D1]phenylacetylene (20), along with 70–65% of 11. These observations demonstrated unambiguously the propensity of vinylidene carbenoids 6, prepared in the presence of TEA, to undergo FBW rearrangement predominantly and, therefore, the prevalence of pathway b that leads to CrII acetylides 10 (Scheme 2). The smooth generation of low-valent chromium(II) acetylides from 1,1,1-trihaloalkanes and CrCl2 as well as the reactivity of such organometallic species as new alkynylating agents have not been studied earlier. Although their nucleophilic behavior has been shown by trapping with electrophiles such as aldehydes, these reagents could be eventually used either in one-pot, metal (Pd, Ni, Fe) cross-coupling reactions or engaged in situ in [3+2]-Huisgen dipolar cycloaddition with azides.[40] The optimization as well as the scope and the potential applications of these synthetically useful reagents are underway in our laboratories and will be disclose elsewhere.</p><!><p>In summary, we have found that the reduction of 1,1,1-trichloroalkanes with an excess amount of chromium(II) provides direct access to mixed nucleophilic chromium(II) and chromium(III) acetylides. Both species are formed through divergent pathways, from chromium(III) vinylidene carbenoids, which favor the formation of uncommon low-valent chromium(II) acetylides, generated through the FBW route. We have demonstrated that in situ generated mixed chromium acetylides react smoothly with electrophilic aldehydes, providing access to propargyl alcohols, complementing other known strategies. Further promising developments using this new reductive/oxidation reaction of 1,1,1-trichloroalkanes by CrCl2 are expected to emerge and will be disseminated shortly.</p><!><p>All reactions were performed under an argon atmosphere. The solvent (THF) was distilled from Na and benzophenone. All commercially available reagents were used without further purification. Analytical thin-layer chromatography (TLC) was performed on glass-backed silica gel plates. Visualization of the developed chromatogram was performed by using UV absorbance and staining with a vanillin, phosphomolybdic acid, or cerium sulfate solution. Flash column chromatography was performed with silica gel (40–63 µm) according to a standard technique. Nuclear magnetic resonance spectra (1H, 13C, and 19F) were recorded with a Bruker 400 MHz spectrometer equipped with a BBI or a DUAL probe. Chemical shifts for 1H and 13C NMR spectra are recorded in parts per million by using the residual chloroform as an internal standard (1H, δ = 7.26 ppm; 13C, δ = 77.16 ppm). Multiplicities are indicated by s (singlet), br. s (broad singlet), d (doublet), t (triplet), and m (multiplet). Mass spectra were recorded with a MS–MS high-resolution Micromass ZABSpecTOF spectrometry. Infrared spectra were recorded with an FTIR spectrometer equipped with KRS-5.</p><!><p>All the reactions were performed on 1-mmol scale of trichloroalkanes 1a–d. To a solution of trichloroalkane 1 (1 equiv.) in THF (15 mL) was added aldehyde 12 (1 equiv.), CrCl2 (6 equiv.), and TEA (10 equiv.) under an inert atmosphere. The whole mixture was allowed to stir overnight at room temperature (10 h). After completion of the reaction (TLC analysis), the mixture was quenched with 1 n HCl (5 mL) and extracted with EtOAc (2 × 10 mL). The organic layer was washed with water and brine and dried with Na2SO4. Evaporation of the solvent under reduced pressure gave the crude product, which was purified by silica gel column chromatography. Elution with EtOAc/cyclohexane gave desired propargyl alcohol 3 and (Z)-2-chloroalk-2-en-1-ol (14) as an inseparable mixture. [32]</p><!><p>Yield: 192 mg (72%), colorless sticky solid. 1H NMR (400 MHz, CDCl3): δ = 7.54 (d, J = 8.4 Hz, 2 H), 7.37–7.39 (m, 2 H), 7.22–7.24 (m, 3 H), 7.03 (d, J = 8.8 Hz, 2 H), 5.5 (br. s, 1 H), 2.41 (br. s, 1 H), 2.20 (s, 3 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 169.5, 169.4, 150.6, 138.2, 131.7, 128.3, 127.9, 122.3, 121.7, 88.5, 86.8, 64.5, 21.1 ppm. IR (film): ṽ = 3414, 3056, 2183, 1753, 1504, 1195, 1164, 1064, 733 cm−1. HRMS (EI): calcd. for C17H14NaO3 [M + Na]+ 289.0835; found 289.0845.</p><!><p>Yield: 164 mg (65%), light-yellow-colored viscous solid. 1H NMR (400 MHz, CDCl3): δ = 7.48–7.45 (m, 2 H), 7.29 (d, J = 8.4 Hz, 2 H), 7.05 (d, J = 8 Hz, 2 H), 6.84–6.86 (m, 2 H), 5.56 (d, J = 5.6 Hz, 1 H), 3.75 (s, 3 H), 2.27 (s, 3 H), 2.12 (d, J = 6 Hz, 1 H) ppm. 13CNMR (100 MHz, CDCl3): δ = 159.7, 138.7, 131.6, 133.1, 129.2, 129.0, 128.9, 128.1, 128.0, 119.4, 114.0, 88.2, 86.6, 64.8, 55.3, 21.5 ppm. IR (film): ṽ= 3400, 2921, 2835, 2197, 1608, 1508, 1245, 1170, 1029, 814 cm−1. HRMS (EI): calcd. for C17H16NaO2 [M + Na]+ 275.1042; found 275.1053.</p><!><p>Yield: 239 mg (79%), colorless sticky solid. 1H NMR (400 MHz, CDCl3): δ = 7.53–7.51 (m, 2 H), 7.27–7.25 (m, 2 H), 7.03–7.00 (m, 4 H), 5.56 (br. s, 1 H), 2.51 (br. s, 1 H), 2.25 (s, 3 H), 2.20 (s, 3 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 169.5, 150.5, 138.8, 138.4, 131.6, 129.1, 127.9, 121.7, 119.2, 87.9, 86.9, 64.5, 21.5, 21.3, 21.1 ppm. IR (film): ṽ = 3432, 3043, 2931, 2195, 1754, 1505, 1194, 1163, 1012, 815, 734 cm−1. HRMS(EI): calcd. for C18H16NaO3 [M + Na]+ 303.0991; found 303.1000.</p><!><p>Yield: 195 mg (81%), colorless sticky solid. 1H NMR (400 MHz, CDCl3): δ = 7.52–7.48 (m, 2 H), 7.28–7.26 (m, 2 H), 7.05–6.96 (m, 2 H), 7.04 (d, J = 8 Hz, 2 H), 5.57 (d, J = 5.2 Hz, 1 H), 2.30 (d, J = 6 Hz, 1 H), 2.26 (s, 3 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 163.9 (d, 1 J= 246 Hz, C), 161.4, 138.9, 136.6 (d, 4J= 3 Hz, C), 131.6, 129.1, 128.6 (d, 3J = 8 Hz, C), 128.5, 125.5, 119.1, 115.5 (d, 2J = 22 Hz, C), 115.3, 115.2, 87.8, 87.0, 64.4, 21.5, 21.3 ppm. 19F NMR (376.49 MHz, CDCl3): δ = −113.8 (s, 1 F) ppm. IR (film): ṽ = 3338, 2921, 2229, 1604, 1506, 1221, 1156, 1013, 836, 814 cm−1. HRMS (EI): calcd. for C16H14FO [M + H]+ 241.1023; found 241.1036.</p><!><p>Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/ejoc.200901476.</p><p>Supporting Information (see footnote on the first page of this article): 1H, 13C, and 19F (for 3i) NMR spectra of the original compounds.</p>

PubMed Author Manuscript

Unified Synthesis of Polycyclic Alkaloids via Complementary Carbonyl Activation

A complementary dual carbonyl activation strategy for the synthesis of polycyclic alkaloids has been developed. Successful applications include the synthesis of tetracyclic alkaloids harmalanine, harmalacinine, pentacyclic indoloquinolizidine alkaloid nortetoyobyrine, and octacyclic β-carboline alkaloid peganumine A. The latter synthesis features a protecting-group-free assembly and an asymmetric disulfonimide catalyzed cyclization. Furthermore, formal syntheses of hirsutine, deplancheine, 10-desbromoarborescidine A, and oxindole alkaloids rhynchophylline and isorhynchophylline have been achieved. Finally, a concise synthesis of berberine alkaloid ilicifoline B was completed.G. He thanks the China Scholarship Council and Dahlem Research School for doctoral scholarships. We thank members and technician team in AK List (Max-Planck-Institut für Kohlenforschung) for the support of the chiral catalysts.

unified_synthesis_of_polycyclic_alkaloids_via_complementary_carbonyl_activation

<!>Entry for the Table of Contents<!>Carbonyl Activation<!>General information<!>Chemicals<!>Mass Spectrometry<!>IR Spectroscopy<!>Specific Rotations<!>General procedure A<!>General procedure B<!>4-Nitrophenyl 5-((2-(1H-indol-3-yl)ethyl)amino)-5-oxopentanoate (8b).<!>Tert-butyl 5-((2-(1H-indol-3-yl)ethyl)amino)-5-oxopentanoate (8l).<!>1,1,1,3,3,3-Hexafluoropropan-2-yl 2-(2-((2-(1H-indol-3-yl)ethyl)amino)-2-oxoethyl)benzoate (11j).<!>N-(2-(1H-indol-3-yl)ethyl)-8-((1,1,1,3,3,3-hexafluoropropan-2-yl)oxy)-7-oxooctanamide (11l).<!>S25<!>S34<!>1,1,1,3,3,3-Hexafluoropropan-2-yl 5-oxo-5-(phenethylamino)pentanoate (S5).<!>General procedure C<!>S44<!>8,13-Dihydroindolo[2',3':3,4]pyrido[1,2-b]isoquinolin-5(7H)-one (12j).<!>FT-IR<!>S53<!>FT-IR<!>2,3-Dimethoxy-5,6-dihydro-8H-isoquinolino[3,2-a]isoquinolin-8-one (12s).<!>5,6-Dihydro-8H-isoquinolino[3,2-a]isoquinolin-8-one (12t).<!>S59<!>Detail and consideration of the annulation of the additional substrates<!>Scheme S-1: Group I examples<!>2)<!>4(3H)-one (S15).<!>Procedure for the scale-up reaction:

<p>Despite the advancement of combinatorial strategies, natural products remain an indispensable source for the discovery of new molecular entities. [1] Their diverse scaffolds with hydrogen bond donor and acceptor groups positioned in a well-defined spatial arrangement make them attractive starting points and inspiration for drug development. [2] Bioactive polycyclic alkaloids, such as yohimbine (1), hirsutine (2), deplancheine (3), eburnamonine (4), ilicifoline B (5), peganumine A (6), and reserpine (7), contain the common quinolizidine core I fused to different heterocyclic rings (Figure 1). We reasoned that developing a straightforward annulation method for efficient construction of these scaffolds is beneficial for the total synthesis of polycyclic natural products and their analogs. Since the indole substructure is a privileged [3] and very common motif in these polycyclic natural products, we started our synthetic journey with the quinolizidine-fused indole core. We strategized that incorporating an enamide motif into the A ring would provide a flexible handle for subsequent transformations. Therefore, intermediate III was considered the central linchpin for a divergent synthesis of polycyclic alkaloids. It was envisioned to be derived from IV by an annulation sequence involving an electrophilic cyclization followed by lactamization. Toward this goal we identified two major challenges: 1) selective activation of the amide carbonyl group to participate in the electrophilic cyclization; [4] 2) subsequent selective activation of the second carbonyl group to achieve lactamization. conditions afforded tricyclic imine 9 as the major product. Unfortunately, the subsequent imine acylation to give tetracyclic product turned out to be challenging. With POCl 3 , 10 was isolated in 6% yield along with 84% of imine 9 (Table 1, entry 1). This result indicated that the reaction had stopped after the first cyclization. We hypothesized that imine-enamine tautomerization during second cyclization could also be a critical prerequisite for the second cyclization. [9] After a screening of bases (see the SI), we achieved a slight improvement to 10% yield of 10 using K 2 CO 3 (Table 1, entry 8). With nBu 4 NBr as phase transfer catalyst and methanol, the yield of 10 was further increased to 18% (Table 1, entry 9). Inspired by active ester activation strategies used in peptide synthesis, [10] we tested a variety of ester derivatives (see the SI). Satisfyingly, with 1,1,1,3,3,3-hexafluoro-2-propoxy ester 8d, we achieved a 90% yield of 10 (Table 1, entry 12). Entry [a] Substrate Amide activation reagent With optimized conditions in hand, we explored the scope of the reaction for the synthesis of diverse polycyclic scaffolds (Table 2). Substitutions at the indole ring with electron donating groups (12a and 12b) and electron withdrawing groups (12c and 12d) were well tolerated, providing the corresponding tetracyclic scaffolds in good yields (76-86%). [a] Reactions were performed with substrate (0.10-4.7 mmol) using the standard procedure, isolated yields. See the SI for details.</p><p>Encouraged by these results, we investigated additional substitution patterns and ring systems. Substituting the quinolizidine core afforded the tetracycles 12e-12j in good yields (64-83%) thus providing access to the indoloquinolizidinetype alkaloid nortetoyobyrine (12j) [11] in an additional step. The 7/6, 8/6, 6/7 and 6/8 fused ring systems were obtained in moderate yield (12k-12n, 26%-52%). Finally, we successfully expanded our strategy to benzene derivatives and hetero-aromatic compounds, such as furan, thiophene, and benzothiophene (12o-12x, 33%-91%)</p><p>We next turned our attention to manipulations in the A ring in order to fully exploit our scaffold for natural product synthesis. Through oxidation, a second double bond could be easily introduced to the 3,4-position (VII). Reduction of the double bond in the 1,2-position (VIII) could be achieved with or without concomitant reduction of the lactam. Moreover, introduction of a carbonyl group in 4-position (IX) was key to the synthesis of more complex natural product.</p><p>Scheme 2. Diversification strategy for the tetracyclic scaffold.</p><p>Starting with the dehydrogenation, we tested selenium-and sulfur-based reagents, such as PhSeCl, PhSeBr, PhSSPh and N-tert-butyl phenylsulfinimidoyl chloride (see the SI). [12] Among standard protocols, only N-tert-butyl phenylsulfinimidoyl chloride afforded traces of the desired product. Gratifyingly, using the palladium-catalyzed amide dehydrogenation protocol developed by Newhouse, [13] demethoxyharmalanine (14a), harmalanine (14b), demethoxyharmalacinine (14c) and harmalacinine (14d), [14] were successfully obtained in an excellent yield (60-77%). Racemic 15 can be obtained through selective catalytic hydrogenation of 10 using palladium on carbon. From this intermediate, selenoxide elimination affords 16, a key intermediate in the total synthesis of hirsutine (2), rhynchophylline (17) and isorhynchophylline (18). [15] An asymmetric reduction of the C-C double bond was realized using chiral phosphoric acid (CPA) 19 and Hantzsch ester ( 20) system [16] to give 15 in 80% ee and 61% yield. This material can be converted into (S)-deplancheine (3) and (S)-10desbromoarborescidine A (21) as previously reported. [17] Scheme 4. Reductive diversification.</p><p>To further demonstrate the synthetic potential of this method, we envisioned to use our annulation in a protecting-group-and transition-metal-free asymmetric total synthesis of peganumine A (6). [18] Scheme 5. Protecting-group-free synthesis of ()-peganumine A.</p><p>Following the established protocol, we successfully prepared the tetracyclic intermediate 25 in 85% yield. Subsequently, ketoenamide 26 was obtained in 50% yield through a two-step α-oxidation sequence. The Boc-derivative of 26 constitutes an intermediate in Zhu's elegant total synthesis of peganumine A (6). [19] At this point, we contemplated the possibility of a protecting-group-free synthesis. The key cascade cyclization was achieved using 0.2 equiv. of TFA in toluene to complete a protecting-group-free synthesis of ()-6 in 42% yield.</p><p>Encouraged by the success of the previous cascade cyclization, we initiated investigations toward an asymmetric total synthesis. First, we tested the chiral thiourea (CTU) and PhCO 2 H system developed by Jacobsen, [20] which afforded 92% ee in Zhu's synthesis [19] for the Boc-protected substrate. In our protectinggroup-free substrate, with 27 and PhCO 2 H, the enantioselectivity was 9% ee. We speculated that the remarkable difference in enantioselectivity could be attributed to an impaired recognition between substrate and catalyst. It is possible that either the Boc group is crucial for the recognition, or that the free αketoenamide 26 interrupted the substrate binding. Based on these considerations, we proposed to either apply a multibinding-site catalyst to rigidify the transition state, or to use asymmetric counteranion directed catalysis (ACDC) [21] as stronger chiral acid to activate the imine more efficiently. First, we tested the conjugate-base-stabilized Brønsted acid (CBSBA) 28 developed by Seidel, [22] and 1,2,3,4,5pentacarboxycyclopentadiene (PCCP) derived pentamenthyl ester 29, a novel C-H acid discovered by Lambert, [23] which are all multi-binding-site catalysts. However, no improvement of the enantioselectivity could be achieved with our substrate. Moving to the ACDC using CPA-2 (30), a significant improvement of the enantioselectivity (31% ee) was observed. Expanding on this idea, we applied the stronger chiral Brønsted acid disulfonimide (DSI) [24] to the reaction and obtained 79% ee with DSI-1 (31).</p><p>Encouraged by this promising result, and after intensive screening of DSIs (see the SI), we finally discovered that using DSI-2 (32) could achieve 97% ee and 81% yield. Inspired by the great potential of total synthesis of indole alkaloids, this annulation was further extended to the synthesis of dimeric berberine alkaloid ilicifoline B (5). [25] Using our standard reaction sequence, 8-oxopseudopalmatine (36) was obtained in 95% yield for the annulation. Using Opatz's dimerization procedure, [26] racemic ilicifoline B was synthesized. Moreover, 8-oxopseudopalmatine (36) can be transformed into the tetracyclic protoberberine alkaloid xylopinine (37) according to the reported method. [27] Scheme 6. Synthesis of berberine alkaloids.</p><p>In summary, we have developed an efficient method that is enabling to the rapid assembly of polycyclic scaffolds of bioactive alkaloids, through a straightforward annulation reaction featuring a complementary carbonyl activation strategy. Diverse polycyclic ring systems were accessed in good yields, enabling the total synthesis of different types of alkaloids and their analogs. Through diverging pathways, the total synthesis of five alkaloids and formal total synthesis of six alkaloids were completed. Among them, a synthesis of (+)-and (-)-peganumine A (6) was achieved in a protecting-group-free sequence using a DSI catalyzed Pictet-Spengler reaction as the key step. Finally, we also applied this method to a synthesis of dimeric berberine alkaloid ilicifoline B (5).</p><!><p>An efficient construction of polycyclic scaffolds of bioactive alkaloids via a novel annulation strategy has been developed. Among several natural products prepared, a synthesis of the dimeric berberine alkaloid ilicifoline B and a protecting-group-free synthesis of (+)-and (-)-peganumine A have been achieved. The latter features an asymmetric Brønsted acid catalyzed cyclization cascade.</p><p>Institute and/or researcher Twitter usernames: @ChristmannGroup @FU_Berlin @ListLaboratory @maxplanckpress</p><!><p>Guoli He, Benjamin List* and Mathias Christmann* S2</p><!><p>Unless otherwise stated, all reactions were magnetically stirred and conducted in anhydrous solvents under argon, applying standard Schlenk techniques. Solvents and liquid reagents, as well as solutions of solid or liquid reagents were added via syringes, stainless steel cannulas through rubber septa or micropipettes in a weak stream of argon.</p><p>Solid reagents were added through a weak argon counter-flow. Cooling baths were prepared in Dewar vessels, filled with ice/water (0 °C), or dry ice/acetone (-78 °C).</p><p>Heated oil baths were used for reactions requiring elevated temperatures. Solvents were removed under reduced pressure at 40 °C using a rotary evaporator. All given yields are isolated yields of NMR spectroscopically pure materials, unless otherwise stated. Fractions containing a desired substance were combined and concentrated in vacuo.</p><!><p>Solvent mixtures (mobile phase) are reported in terms of volume ratios (v/v).</p><!><p>High resolution mass spectrometry (HRMS) was performed on a Finnigan MAT 95 (EI), Bruker APEX III FTMS (ESI) or Agilent 6210 (ESI). The ionization method and mode of detection employed is indicated for the respective experiment and all masses are reported in atomic units per elementary charge (m/z).</p><!><p>IR Spectra were recorded on a JASCO FT/IR-4100 spectrometer. Characteristic absorption bands are reported in wavelengths v in cm -1 and were analyzed with the software Spectral Manager from JASCO.</p><!><p>Specific</p><!><p>A mixture of the amine (1.0 equiv.) and the anhydride (1.0 equiv.) in DMF, was stirred overnight at room temperature.</p><p>Then N-(3-dimethylaminopropyl)-N′ethylcarbodiimide hydrochloride (EDCI) (2.0 equiv.), 4-dimethylaminopyridine (DMAP) (0.2 equiv.) and corresponding alcohol (5.0 equiv.) were added to the reaction S6 mixture subsequently. The resulting mixture was stirred for 24 h at room temperature, before the reaction mixture was diluted with H2O and EtOAc. After washing by brine for three times, the organic phase was dried over MgSO4, and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography.</p><p>(Note: It is not necessary to perform the reaction under argon.)</p><!><p>A mixture of the amine (1.0 equiv.) and the anhydride (1.0 equiv.) in corresponding alcohol was stirred overnight at room temperature. The mixture was then cooled to 0 ºC and SOCl2 (1.0 equiv.) was added dropwise. Immediately after the addition, the reaction mixture was allowed to warm to room temperature and stirred for 2 h. The reaction was quenched with NaHCO3 (sat. aq.) and extracted with DCM. The combined organic phases were washed with brine, dried over MgSO4 and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography.</p><p>(Note: It is not necessary to perform the reaction under argon.)</p><!><p>8b was prepared according to General procedure A, starting from tryptamine (300 mg, 1.87 mmol, 1.0 equiv.) and glutaric anhydride (214 mg, 1.87 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (718 mg, 3.74 mmol, 2.0 equiv.), DMAP (45.8 mg, 0.374 mmol, 0.2 equiv.) and p-nitrophenol</p><p>(1.30 g, 9.36 mmol, 5.0 equiv.). Purification by silica gel column chromatography (DCM/MeOH = 10:1) afforded 8b (318 mg, 0.806 mmol, 43%) as a yellow solid. 8c was prepared according to General procedure B, starting from tryptamine (500 mg, 3.12 mmol, 1.0 equiv.) and glutaric anhydride (356 mg, 3.12 mmol, 1.0 equiv.) in iPrOH (10.0 mL). The reaction mixture was subsequently treated with SOCl2 (221 µL, 3.12 mmol, 1.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 8c (524 mg, 1.66 mmol, 53%) as a white solid.</p><!><p>8l was prepared according to General procedure A, starting from tryptamine (150 mg, 0.938 mmol, 1.0 equiv.) and glutaric anhydride (107 mg, 0.938 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (358 mg, 1.88 mmol, 2.0 equiv.), DMAP (22.9 mg, 0.188 mmol, 0.2 equiv.) and tBuOH (445 µL, 4.69 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 8l (21.0 mg, 636 µmol, 7%) as a white solid.</p><p>S15 11a was prepared according to General procedure A, starting from 5methoxytryptamine (200 mg, 1.05 mmol, 1.0 equiv.) and glutaric anhydride (120 mg, 1.05 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (402 mg, 2.10 mmol, 2.0 equiv.), DMAP (25.7 mg, 0.211 mmol, 0.2 equiv.) and HFIP (556 µL, 5.26 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11a (270 mg, 0.595 mmol, 56%) as a white solid. 11b was prepared according to General procedure A, starting from 6methoxytryptamine (100 mg, 0.526 mmol, 1.0 equiv.) and glutaric anhydride (60.0 mg, 0.526 mmol, 1.0 equiv.) in DMF (3.0 mL). The reaction mixture was subsequently treated with EDCI (201 mg, 1.05 mmol, 2.0 equiv.), DMAP (12.9 mg, 0.105 mmol, 0.2 equiv.) and HFIP (278 µL, 2.63 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11b (100 mg, 0.220 mmol, 42%) as a white solid. 11c was prepared according to General procedure A, starting from 6-bromotryptamine (100 mg, 0.420 mmol, 1.0 equiv.) and glutaric anhydride (47.9 mg, 0.420 mmol, 1.0 equiv.) in DMF (3.0 mL). The reaction mixture was subsequently treated with EDCI (160 mg, 0.840 mmol, 2.0 equiv.), DMAP (10.3 mg, 84.0 µmol, 0.2 equiv.) and HFIP (222 µL, 2.10 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11c (120 mg, 0.239 mmol, 57%) as a white solid. 11d was prepared according to General procedure A, starting from 6-fluorotryptamine S18 (260 mg, 1.46 mmol, 1.0 equiv.) and glutaric anhydride (166 mg, 1.46 mmol, 1.0 equiv.)</p><p>in DMF (6.0 mL). The reaction mixture was subsequently treated with EDCI (558 mg, 2.92 mmol, 2.0 equiv.), DMAP (35.7 mg, 0.292 mmol, 0.2 equiv.) and HFIP (772 µL, 7.30 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11d (350 mg, 0.792 mmol, 54%) as a white solid. 11e was prepared according to General procedure A, starting from tryptamine (2.00 g, 12.5 mmol, 1.0 equiv.) and 3-methylglutaric anhydride (1.60 g, 12.5 mmol, 1.0 equiv.)</p><p>in DMF (50 mL). The reaction mixture was subsequently treated with EDCI (4.78 g, 25.0 mmol, 2.0 equiv.), DMAP (305 mg, 2.50 mmol, 0.2 equiv.) and HFIP (6.60 mL, 62.5 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11e (3.21 g, 7.33 mmol, 58%) as a white solid. 11f and 11g were prepared according to General procedure A, starting from tryptamine (500 mg, 3.12 mmol, 1.0 equiv.) and 2-methylglutaric anhydride (400 mg, 3.12 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (1.19 g, 6.25 mmol, 2.0 equiv.), DMAP (76.4 mg, 0.625 mmol, 0.2 equiv.) and HFIP (1.65 mL, 15.6 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11f (190 mg, 0.434 mmol, 14%) as a light brown solid and 11g (135 mg, 0.308 mmol, 10%) as a light brown solid. 11h was prepared according to General procedure A, starting from tryptamine (400 mg, 2.50 mmol, 1.0 equiv.) and 3,3-dimethylglutaric anhydride (355 mg, 2.50 S21 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (955 mg, 5.00 mmol, 2.0 equiv.), DMAP (61.1 mg, 500 µmol, 0.2 equiv.)</p><p>and HFIP (1.32 mL, 12.5 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11h (697 mg, 1.54 mmol, 62%) as a white solid. 11i was prepared according to General procedure A, starting from (S)-tryptophan methyl ester (300 mg, 1.38 mmol, 1.0 equiv.) and glutaric anhydride (157 mg, 1.38 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (526 mg, 2.75 mmol, 2.0 equiv.), DMAP (33.6 mg, 0.275 mmol, 0.2 equiv.) and HFIP (727 µL, 6.88 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11i (150 mg, 0.311 mmol, 23%) as a white solid. [𝛂] 𝐃 𝟐𝟔 = +9.2(c = 0.7, CHCl3).</p><!><p>11j was prepared according to General procedure A, starting from tryptamine (150 mg, 0.938 mmol, 1.0 equiv.) and homophthalic anhydride (152 mg, 0.938 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (197 mg, 1.03 mmol, 2.0 equiv.), DMAP (22.9 mg, 0.188 mmol, 0.2 equiv.) and HFIP (495 µL, 4.69 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 5:2) afforded 11j (32.0 mg, 678 µmol, 7%) as a white solid. 11k was prepared according to General procedure A, starting from tryptamine (400 mg, 2.50 mmol, 1.0 equiv.) and oxepane-2,7-dione (320 mg, 2.50 mmol, 1.0 equiv.)</p><p>in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (955 mg, 5.00 mmol, 2.0 equiv.), DMAP (61.1 mg, 0.500 mmol, 0.2 equiv.) and HFIP (1.32 mL, 12.5 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11k (289 mg, 0.660 mmol, 26%) as a white solid.</p><!><p>11l was prepared according to General procedure A, starting from tryptamine (400 mg, 2.50 mmol, 1.0 equiv.) and oxocane-2,8-dione (355 mg, 2.50 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (955 mg, 5.00 mmol, 2.0 equiv.), DMAP (61.1 mg, 0.500 mmol, 0.2 equiv.) and HFIP (1.32 mL, 12.5 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11l (300 mg, 0.644 mmol, 26%) as a white solid.</p><!><p>11m was prepared according to General procedure A, starting from 1H-indole-3propanamine (85.0 mg, 0.489 mmol, 1.0 equiv.) and glutaric anhydride (55.7 mg, 0.489 mmol, 1.0 equiv.) in DMF (2.0 mL). The reaction mixture was subsequently treated with EDCI (187 mg, 0.977 mmol, 2.0 equiv.), DMAP (11.9 mg, 98.0 µmol, 0.2 equiv.)</p><p>and HFIP (258 µL, 2.44 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11m (120 mg, 0.274 mmol, 56%) as a white solid. 11r was prepared according to General procedure A, starting from 3,4dimethoxyphenethylamine (500 mg, 2.76 mmol, 1.0 equiv.) and glutaric anhydride (315 mg, 2.76 mmol, 1.0 equiv.) in DMF (15.0 mL). The reaction mixture was subsequently treated with EDCI (1.05 g, 5.52 mmol, 2.0 equiv.), DMAP (67.5 mg, 0.552 mmol, 0.2 equiv.) and HFIP (1.46 mL, 13.8 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 5:2) afforded 11r (670 mg, 1.42 mmol, 55%) as a white solid. 11v was prepared according to General procedure A, starting from 2-(furan-2yl)ethan-1-amine (180 mg, 1.62 mmol, 1.0 equiv.) and glutaric anhydride (185 mg, 1.62 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (620 mg, 3.24 mmol, 2.0 equiv.), DMAP (39.6 mg, 0.324 mmol, 0.2 equiv.) and HFIP (857 µL, 8.11 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11v (250 mg, 0.667 mmol, 41%) as a light yellow solid.</p><p>S32 11w was prepared according to General procedure A, starting from 2-(thiophen-3yl)ethan-1-amine (150 mg, 1.18 mmol, 1.0 equiv.) and glutaric anhydride (135 mg, 1.18 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (451 mg, 2.36 mmol, 2.0 equiv.), DMAP (28.9 mg, 0.236 mmol, 0.2 equiv.) and HFIP (624 µL, 5.91 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11w (200 mg, 0.512 mmol, 43%) as a white solid.</p><!><p>11y was prepared according to General procedure A, starting from 6methoxytrytamine (100 mg, 0.526 mmol, 1.0 equiv.) and glutaric anhydride (67.4 mg, 0.526 mmol, 1.0 equiv.) in DMF (3.0 mL). The reaction mixture was subsequently treated with EDCI (201 mg, 1.05 mmol, 2.0 equiv.), DMAP (12.9 mg, 0.105 mmol, 0.2 equiv.) and HFIP (278 µL, 2.63 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 11y (150 mg, 0.320 mmol, 61%) as a white solid. S1 was prepared according to General procedure A, starting from tryptamine (250 mg, 1.56 mmol, 1.0 equiv.) and succinic anhydride (156 mg, 1.56 mmol, 1.0 equiv.) in DMF S35</p><p>(5.0 mL). The reaction mixture was subsequently treated with EDCI (328 mg, 1.72 mmol, 2.0 equiv.), DMAP (38.2 mg, 0.313 mmol, 0.2 equiv.) and HFIP (825 µL, 7.81 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded S1 (286 mg, 0.698 mmol, 45%) as a white solid. 3.60 (q, J = 6.5 Hz, 2H), 2.96 (td, J = 6.8, 0.9 Hz, 2H), 2.84 (t, J = 6.9 Hz, 2H), 2.44 (t, J = 6.9 Hz, 2H). S2 was prepared according to General procedure A, starting from 1H-indole-3methanamine (310 mg, 2.12 mmol, 1.0 equiv.) and glutaric anhydride (242 mg, 2.12 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (811 mg, 4.25 mmol, 2.0 equiv.), DMAP (51.9 mg, 0.425 mmol, 0.2 equiv.) and HFIP (1.12 mL, 10.6 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded S2 (300 mg, 0.732 mmol, 34%) as a white solid. S3 was prepared according to General procedure A, starting from tryptamine (200 mg, 1.25 mmol, 1.0 equiv.) and oxacycloundecane-2,11-dione (230 mg, 1.25 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (477 mg, 2.50 mmol, 2.0 equiv.), DMAP (30.5 mg, 0.250 mmol, 0.2 equiv.) and HFIP (660 µL, 6.25 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded S3 (150 mg, 0.304 mmol, 24%) as a white solid. S4 was prepared according to General procedure A, starting from (S)-tyrosine methyl ester (300 mg, 1.54 mmol, 1.0 equiv.) and glutaric anhydride (175 mg, 1.54 mmol, 1.0 equiv.) in DMF (10.0 mL). The reaction mixture was subsequently treated with EDCI (588 mg, 3.01 mmol, 2.0 equiv.), DMAP (37.6 mg, 0.308 mmol, 0.2 equiv.) and HFIP (813 µL, 7.69 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded S4 (170 mg, 0.370 mmol, 24%) as a white solid. [𝛂] 𝐃 𝟐𝟔 = +19.0 (c = 0.4, CHCl3).</p><!><p>S5 was prepared according to General procedure A, starting from phenethylamine S38 (200 mg, 1.65 mmol, 1.0 equiv.) and glutaric anhydride (188 mg, 1.65 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (631 mg, 3.31 mmol, 2.0 equiv.), DMAP (40.4 mg, 0.331 mmol, 0.2 equiv.) and HFIP (873 µL, 8.26 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 5:2) afforded S5 (200 mg, 0.519 mmol, 31%) as a white solid.</p><!><p>To a solution of the substrate 8 or 11a-x (1.0 equiv.) in toluene was added POCl3 To a solution of the substrate 8a (100 mg, 0.347 mmol, 1.0 equiv.) in toluene (3.0 mL), was added POCl3 (31.7 µL, 0.347 mmol, 1.0 equiv.). The mixture was heated to 115 ºC (oil bath temperature) until the starting material was fully consumed. Then the mixture was concentrated under reduced pressure and purified by silica gel flash column chromatography (DCM/MeOH= 10:1). 9 (79.0 mg, 0.293 mmol, 84%) was afforded as a dark yellow solid. (304 mg, 2.20 mmol, 10 equiv.) and nBu4NBr (7.1 mg, 22.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 12a (45.0 mg, 0.168 mmol, 76%) as a yellow solid.</p><!><p>12c was prepared according to General procedure C, starting from 11c (37.0 mg, 74.0 µmol, 1.0 equiv.) and POCl3 (6.7 µL, 74.0 µmol, 1.0 equiv.) in toluene (2.0 mL).</p><p>The reaction mixture was subsequently treated with MeOH (2.0 mL), K2CO3 (99.0 mg, 740 µmol, 10 equiv.) and nBu4NBr (2.4 mg, 7.0 µmol, 0.1 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 12c (20.0 mg, 63.0 µmol, 86%) as a yellow solid. (2.0 mL). The reaction mixture was subsequently treated with MeOH (2.0 mL), K2CO3</p><p>(198 mg, 1.43 mmol, 10 equiv.) and nBu4NBr (4.6 mg, 14.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 1:1) afforded 12i (30.0 mg, 0.101 mmol, 71%) as a yellow solid. [𝛂] 𝐃 𝟐𝟔 = +4.6 (c = 0.2, CHCl3).</p><!><p>12j was prepared according to General procedure C, starting from 11j (50.0 mg, S49 0.106 mmol, 1.0 equiv.) and POCl3 (9.7 µL, 0.106 mmol, 1.0 equiv.) in toluene (1.5 mL). The reaction mixture was subsequently treated with MeOH (1.5 mL), K2CO3</p><p>(146 mg, 1.06 mmol, 10 equiv.) and nBu4NBr (3.4 mg, 11.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 3:1) afforded 12j (25.0 mg, 87.4 µmol, 82%) as a light yellow solid. 12k was prepared according to General procedure C, starting from 11k (200 mg, 0.457 mmol, 1.0 equiv.) and POCl3 (41.7 µL, 0.457 mmol, 1.0 equiv.) in toluene (5.0 mL). The reaction mixture was subsequently treated with MeOH (5.0 mL), K2CO3</p><p>(631 mg, 4.57 mmol, 10 equiv.) and nBu4NBr (14.7 mg, 46.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 12k (30.0 mg, 0.119 mmol, 26%) as a yellow solid. 12m was prepared according to General procedure C, starting from 11m (50.0 mg, S51 0.114 mmol, 1.0 equiv.) and POCl3 (10.4 µL, 0.114 mmol, 1.0 equiv.) in toluene (1.5 mL). The reaction mixture was subsequently treated with MeOH (1.5 mL), K2CO3</p><p>(158 mg, 1.14 mmol, 10 equiv.) and nBu4NBr (3.7 mg, 11.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 12m (15.0 mg, 59.5 µmol, 52%) as a yellow solid.</p><!><p>12n was prepared according to General procedure C, starting from 11n (39.0 mg, 86.0 µmol, 1.0 equiv.) and POCl3 (7.9 µL, 86.0 µmol, 1.0 equiv.) in toluene (1.5 mL).</p><p>The reaction mixture was subsequently treated with MeOH (1.5 mL), K2CO3 (119 mg, 863 µmol, 10 equiv.) and nBu4NBr (2.8 mg, 9.0 µmol, 0.1 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 12n (10.0 mg, 37.6 µmol, 43%) as a yellow solid.</p><p>1 H NMR (700 MHz, CDCl3) δ [ppm] = 7.83 (s, 1H), 7.55 (dt, J = 7.9, 0.9 Hz, 1H), 7.33 (dt, J = 8.1, 0.9 Hz, 1H), 7.22 (ddd, J = 8.1, 7.0, 1.1 Hz, 1H), 7.14 (ddd, J = 8.0, 7.0, 2,2-Dimethyl-2,3,6,7-tetrahydro-4H-pyrido[2,1-a]isoquinolin-4-one (12o).</p><p>12o was prepared according to General procedure C, starting from 11o (138 mg, 0.334 mmol, 1.0 equiv.) and POCl3 (76.2 µL, 0.835 mmol, 2.5 equiv.) in toluene (4.0 mL). The reaction mixture was subsequently treated with MeOH (4.0 mL), K2CO3</p><p>(462 mg, 3.34 mmol, 10 equiv.) and nBu4NBr (10.8 mg, 33.0 µmol, 0.1 equiv.).</p><p>Purification by neutral aluminium oxide column chromatography (pentane/DCM = 1:1) afforded 12o (25.0 mg, 0.110 mmol, 33%) as a white solid. (td, J = 5.9, 0.9 Hz, 2H), 0.92 (s, 6H).</p><!><p>12p was prepared according to General procedure C, starting from 11p (80.0 mg, 0.181 mmol, 1.0 equiv.) and POCl3 (16.5 µL, 0.181 mmol, 1.0 equiv.) in toluene (3.0 mL). The reaction mixture was subsequently treated with MeOH (3.0 mL), K2CO3</p><p>(250 mg, 1.81 mmol, 10 equiv.) and nBu4NBr (5.8 mg, 18 µmol, 0.1 equiv.).</p><p>Purification by neutral aluminium oxide column chromatography (pentane/DCM = 1:1) afforded 12p (20.0 mg, 77.8 µmol, 43%) as a white solid. 12q was prepared according to General procedure C, starting from 11q (150 mg, 0.317 mmol, 1.0 equiv.) and POCl3 (28.9 µL, 0.317 mmol, 1.0 equiv.) in toluene (5.0 mL). The reaction mixture was subsequently treated with MeOH (5.0 mL), K2CO3</p><!><p>(438 mg, 3.17 mmol, 10 equiv.) and nBu4NBr (10.2 mg, 32.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 3:1) afforded 12q (80.5 mg, 0.280 mmol, 88%) as a yellow solid. (311 mg, 2.25 mmol, 10 equiv.) and nBu4NBr (7.2 mg, 22 µmol, 0.1 equiv.).</p><p>Purification by neutral aluminium oxide column chromatography (pentane/DCM = 1:1) afforded 12r (44.0 mg, 0.170 mmol, 76%) as a yellow solid.</p><p>1 H NMR (500 MHz, C6D6) δ [ppm] = 6.92 (s, 1H), 6.20 (s, 1H), 5.31 (td, J = 5.0, 1.9</p><p>Hz, 1H), 3.93 (td, J = 5.9, 1.8 Hz, 2H), 3.46 (s, 3H), 3.36 (s, 3H), 2.41 (td, J = 7.9, 1.8</p><!><p>12s was prepared according to General procedure C, starting from 11s (193 mg, 0.391 mmol, 1.0 equiv.) and POCl3 (35.7 µL, 0.391 mmol, 1.0 equiv.) in toluene S55</p><p>(5.0 mL). The reaction mixture was subsequently treated with MeOH (5.0 mL), K2CO3</p><p>(541 mg, 3.92 mmol, 10 equiv.) and nBu4NBr (12.6 mg, 39.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 3:1) afforded 12s</p><p>(100 mg, 0.326 mmol, 83%) as a yellow solid.</p><!><p>12t was prepared according to General procedure C, starting from 11t (217 mg, 0.501 mmol, 1.0 equiv.) and POCl3 (45.7 µL, 0.501 mmol, 1.0 equiv.) in toluene (6.0 mL). The reaction mixture was subsequently treated with MeOH (6.0 mL), K2CO3</p><p>(693 mg, 5.01 mmol, 10 equiv.) and nBu4NBr (16.2 mg, 50.0 µmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 5:1) afforded 12t</p><p>(60 mg, 0.243 mmol, 48%) as a yellow solid. 12u was prepared according to General procedure C, starting from 11u (43.0 mg, 0.115 mmol, 1.0 equiv.) and POCl3 (10.5 µL, 0.115 mmol, 1.0 equiv.) in toluene (1.5 mL). The reaction mixture was subsequently treated with MeOH (1.5 mL), K2CO3</p><p>(158 mg, 1.15 mmol, 10 equiv.) and nBu4NBr (3.7 mg, 11 µmol, 0.1 equiv.).</p><p>Purification by neutral aluminium oxide column chromatography (pentane/DCM = 1:1) afforded 12u (18.0 mg, 95.2 µmol, 83%) as a yellow solid.</p><p>1 H NMR (700 MHz, C6D6) δ [ppm] = 6.90 (t, J = 1.7 Hz, 1H), 5.80 (t, J = 1.7 Hz, 1H), 5.45 (t, J = 5.0 Hz, 1H), 3.82 (t, J = 6.0 Hz, 2H), 2.30 (t, J = 7.7 Hz, 2H), 2.06 (t, J = 6.0 Hz, 2H), 1.85 (td, J = 7.7, 5.0 Hz, 2H). 12v was prepared according to General procedure C, starting from 11v (40.0 mg, 0.107 mmol, 1.0 equiv.) and POCl3 (9.7 µL, 0.107 mmol, 1.0 equiv.) in toluene (1.5 mL). The reaction mixture was subsequently treated with MeOH (1.5 mL), K2CO3 S57 (147 mg, 1.07 mmol, 10 equiv.) and nBu4NBr (3.4 mg, 11 µmol, 0.1 equiv.).</p><p>Purification by neutral aluminium oxide column chromatography (pentane/DCM = 1:1) afforded 12v (10.0 mg, 54.0 µmol, 51%) as a yellow solid.</p><p>1 H NMR (700 MHz, C6D6) δ [ppm] = 6.90 (dd, J = 2.0, 0.7 Hz, 1H), 6.11 (d, J = 2.0 Hz, 1H), 4.93 (t, J = 4.9 Hz, 1H), 3.81 (t, J = 6.1 Hz, 2H), 2.33 (t, J = 7.7 Hz, 2H), 2.22-2.18 (m, 2H), 1.87-1.83 (m, 2H). 12w was prepared according to General procedure C, starting from 11w (48.0 mg, 0.123 mmol, 1.0 equiv.) and POCl3 (11.2 µL, 0.123 mmol, 1.0 equiv.) in toluene (1.5 mL). The reaction mixture was subsequently treated with MeOH (1.5 mL), K2CO3 (170 mg, 1.23 mmol, 10 equiv.) and nBu4NBr (4.0 mg, 12 µmol, 0.1 equiv.).</p><p>Purification by neutral aluminium oxide column chromatography (pentane/DCM = 1:1) afforded 12w (23.0 mg, 0.112 mmol, 91%) as a yellow solid.</p><p>1 H NMR (500 MHz, C6D6) δ [ppm] = 6.61 (dd, J = 5.1, 2.0 Hz, 1H), 6.31 (dd, J = 5.1, 2.0 Hz, 1H), 5.20 (td, J = 5.0, 2.0 Hz, 1H), 3.83 (td, J = 6.0, 2.0 Hz, 2H), 2.29 (td, J = 7.8, 1.9 Hz, 2H), 2.17 (td, J = 6.0, 2.0 Hz, 2H), 1.79 (tdd, J = 7.5, 4.9, 1.9 Hz, 2H). 13 was prepared according to General procedure C, starting from 11y (100 mg, 0.214 mmol, 1.0 equiv.) and POCl3 (19.5 µL, 0.214 mmol, 1.0 equiv.) in toluene (3.0 mL). The reaction mixture was subsequently treated with MeOH (3.0 mL), K2CO3</p><p>(295 mg, 2.14 mmol, 10 equiv.) and nBu4NBr (6.9 mg, 21 µmol, 0.1 equiv.).</p><!><p>Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 13</p><p>(47.0 mg, 0.167 mmol, 78%) as a yellow solid.</p><!><p>The annulation of the additional substrates can be divided into two parts. One is no observation of product by the NMR (group I), the other is NMR showed the product, but liable to decompose (group II).</p><!><p>There was no proof that the annulation products of S1-3 were formed. We proposed the The reaction of S5 and S6 showed the identical peak of this type of compound on the 1 H NMR, but there are some impurities which might be the side products formed during the reaction period or in the isolation. We hypothesized that the electrophilicity of the benzene ring influenced the reaction rate of the endo-dig cyclization. If the rate of the endo-dig cyclization is slow, some competition reactions could happen, leading to side products. In the exo-trig cyclization stage, the Thorpe-Ingold effect could increase the rate and drive the whole reaction equilibrium to form the desired product. (Scheme S-</p><!><p>In general, these compounds containing the [2,3-a]quinolizinone motif are sensitive to acid, in some case even the CDCl3 can lead to partly decomposition. The stability is dependent on the fused ring system. According to our experience, the indole fused compounds are stable. But the larger ring system is relatively unstable. To a 0 ºC solution of diisopropylamine (10.0 equiv.) in THF was added nBuLi (2.5 M in hexane, 10.0 equiv.). The reaction mixture was stirred for 1 h. A solution of enamide (1.0 equiv.) in THF was added dropwise into the resulting mixture. The reaction mixture was stirred for 1 h. ZnCl2 (1.9 M in 2-Me-THF, 3.5 equiv.) was added and stirred for with argon (this process was repeated 3 times). Allyl acetate (0.39 mL, 0.36 g, 3.6 mmol)</p><p>and THF (2.6 mL) were added sequentially and the solution was stirred for 0.5 h.)</p><p>[1] Y. Chen, A. Turlik, T. R. Newhouse, J. Am. Chem. Soc.2016 14c was prepared according to General procedure D, starting from diisopropylamine (0.17 mL, 1.2 mmol, 10.0 equiv.) in THF (3.0 mL) was added nBuLi (2.5 M in hexane, 0.48 mL, 1.2 mmol, 10.0 equiv.). A solution of 12e (30 mg, 0.12 mmol, 1.0 equiv.) in THF (3.0 mL) was added dropwise into the resulting mixture. Then the reaction mixture was treated with ZnCl2 (1.9 M in 2-Me-THF, 0.22 mL, 0.42 mmol, 3.5 equiv.). After that, the stock solution of [Pd(allyl)Cl]2 (1.1 mg, 3.0 µmol, 2.5 mol%) and allyl acetate (15 µL, 0.14 mmol, 1.2 equiv.) in THF (0.5 mL) was next added. Purification by silica gel column chromatography (pentane/EtOAc = 1:2) afforded 14c (21 mg, 84 µmol, 71%) as a yellow solid. With the classic dehydrogenation methods, the reactions turned out to be unsatisfying.</p><p>The best result was obtained with LDA, sulfinimidoyl chloride system. We consiedered that the enolation was a reasonable strategy to afford the regioselectivity. Finally, LDA was the best base for the reaction affording 76% yield. 16). [3] Diisopropylamine (234 µL, 1.67 mmol, 10.0 equiv.) was added to dry THF (5.0 mL) at 0 °C. nBuLi (2.5 M in hexane, 667 µL, 1.67 mmol, 10.0 equiv.) was added and the mixture was stirred for 30 min before cooling to -78 °C. A solution of rac-15 (40.0 mg, 0.167 mmol, 1.0 equiv.) in THF (2.0 mL) was added dropwise. The reaction was stirred at -78 °C for 1 h. A solution of PhSeBr (59.0 mg, 250 µmol, 1.5 equiv.) in THF (2.0 mL) was the added dropwise and the reaction mixture was stirred for 24 h. The reaction was quenched by NH4Cl (sat. aq.) and extracted with DCM. The combined organic phases were washed with brine, dried over MgSO4, filtered and concentrated under reduced pressure. The crude selenide was dissolved in methanol (5.0 mL) and water (1.0 mL).</p><p>NaIO4 (71.0 mg, 333 µmol, 2.0 equiv.) and NaHCO3 (16.8 mg, 200 µmol, 1.2 equiv.)</p><p>were added and the reaction was heated at 60 °C for 24 h. The reaction was quenched by NaHCO3 (sat. aq.) and extracted with DCM. The combined organic phases were washed with brine, dried over MgSO4 and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography (pentane/EtOAc = 1:1) to give 16 (16.0 mg, 67.2 µmol, 40%) as light yellow powder. 24 was prepared according to General procedure A, starting from 6methoxytryptamine (200 mg, 1.05 mmol, 1.0 equiv.) and 3,3-dimethylglutaric anhydride (150 mg, 1.05 mmol, 1.0 equiv.) in DMF (5.0 mL). The reaction mixture was subsequently treated with EDCI (402 mg, 2.11 mmol, 2.0 equiv.), DMAP (25.7 mg, 0.211 mmol, 0.2 equiv.) and HFIP (556 µL, 5.26 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 24 (370 mg, 0.768 mmol, 73%) as a yellow solid.</p><p>. 25 was prepared according to General procedure C, starting from 24 (100 mg, 0.207 mmol, 1.0 equiv.) and POCl3 (18.9 µL, 0.207 mmol, 1.0 equiv.) in toluene (3.0 mL). The reaction mixture was subsequently treated with MeOH (3.0 mL), K2CO3</p><p>(287 mg, 2.08 mmol, 10 equiv.) and nBu4NBr (6.7 mg, 0.021 mmol, 0.1 equiv.).</p><p>Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 25 (49.0 mg, 0.165 mmol, 85%) as a yellow solid.</p><!><p>To a solution of tetracyclic lactam 25 (100 mg, 0.34 mmol, 1.0 equiv.) in THF, was added LiTMP (1.0 M solution in THF, 3.40 mL, 3.38 mmol, 10.0 equiv.) at -78 °C. The resulting mixture was stirred at that temperature for 1 h before P(OMe)3 (250 µL, 1.01 mmol, 2.5 equiv.) was added. O2 was bubbled through the mixture for 20 min, and the mixture was stirred for additional 2 h under O2 atmosphere. Then the reaction mixture was quenched with saturated NaHCO3 (sat. aq.). The aqueous phase was extracted with DCM. The combined organic phases were washed with brine, dried with MgSO4, filtered and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography (pentane/EtOAc = 3:2) afforded S15</p><p>(75.0 mg, 0.240 mmol, 71%) as yellow solid.</p><p>). [4] S75</p><p>To a solution of NCS (171 mg, 1.28 mmol, 5.0 equiv.) in DCM (5.0 mL) at 0 °C was added Me2S (471 µL, 6.41 mmol, 25 equiv.) and the mixture was stirred at -78 °C for 1 h. The S15 (80 mg, 0.256 mmol, 1.0 equiv.) in DCM (2.0 mL) was added dropwise.</p><p>The resulting mixture was stirred at -78 °C for another two hours before triethylamine (712 µL, 5.13 mmol, 20 equiv.) was added. Then the mixture was stirred at -78 °C for 2 h. The mixture was quenched with NH4Cl (sat. aq.) and extracted with DCM. The combined organic phases were washed with brine, dried over MgSO4, filtered and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 26 (57.0 mg, 0.184 mmol, 71%) as an orange solid. S83 32 was prepared according to General procedure E, starting from S20 (300 mg, 0.464 mmol, 1.0 equiv.), (3,5-difluoro-4-(trifluoromethyl)phenyl)boronic acid (315 mg, 1.39 mmol, 3.0 equiv.) and Pd(OAc)2 (10.4 mg, 46.4 µmol, 0.1 equiv.) in THF (15 mL).</p><p>After degassing, the mixture was treated with Pt-Bu3 (1.0 M in toluene, 46.4 µL, 46.4 µmol, 0.1 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 2:1) afforded 32 (320 mg, 0.424 mmol, 91%) as a white solid. General Procedure for the small scale (0.50 mg or 1.00 mg) screening reactions A mixture of ketoamide 25 (1.00 mg, 3.00 µmol, 1.0 equiv.), 6-methoxytryptamine (22) (0.75 mg, 3.60 µmol, 1.2 equiv.) and 4 Å MS in corresponding solvent (0.5 mL) was stirred at room temperature under argon in sealed vial. A solution of corresponding catalyst in the same solvent was added to the reaction mixture and the reaction mixture was heated up to the corresponding temperature for the set duration. The reaction S84 mixture was cooled to room temperature and purified by silica gel column chromatography (pentane/EtOAc = 3:2) to yield the pure product as solid.</p><p>(Note: since the ketoamide is not a fine power and cannot dissolve in toluene (the frequently used solvent), the practical way to prepare the 1.00 mg or 0.50 mg ketoamide was that, using fine balance to take 10.00 mg ketoamide which was then dissolved in 5.0 mL DCM (10.0 mL for 0.50 mg preparation), then adding 0.5 mL of such solvent to the 1.5 mL glass reaction vials, finally removing the DCM by rotary evaporator. And the catalyst was prepared in the corresponding solvent to load into the reaction.)</p><!><p>A mixture of ketoamide 25 (66.0 mg, 0.213 mmol, 1.0 equiv.), 6-methoxytryptamine (48.5 mg, 0.255 mmol, 1.2 equiv.) and 4 Å molecular sieves (50.0 mg) in toluene (50 mL) was stirred at room temperature. A solution of DSI-2 (16.1 mg, 21.3 µmol, 0.1 equiv.) in toluene (5.0 mL) was added to the reaction mixture and the reaction mixture was stirred at 60 °C for 60 h. After the reaction mixture was cooled to room temperature, filtered through Celite ® , washed by NaHCO3 (sat. aq.), dried over MgSO4 and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 6 (83.3 mg, 0.173 mmol, 81%) as a white solid.</p><p>S87 35 was prepared according to General procedure A, starting from 33 (100 mg, 0.552 mmol, 1.0 equiv.) and 34 (123 mg, 0.552 mmol, 1.0 equiv.) in DCM (5.0 mL).</p><p>The reaction mixture was subsequently treated with EDCI (211 mg, 1.10 mmol, 2.0 equiv.), DMAP (13.5 mg, 0.110 mmol, 0.2 equiv.) and HFIP (292 µL, 2.76 mmol, 5.0 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 3:2) afforded 35 (190 mg, 0.342 mmol, 62%) as a light yellow solid. 36 was prepared according to General procedure C, starting from 35 (60 mg, 0.11 mmol, 1.0 equiv.) and POCl3 (9.9 µL, 0.11 mmol, 1.0 equiv.) in toluene (3.0 mL).</p><p>The reaction mixture was subsequently treated with MeOH (3.0 mL), K2CO3 (150 mg, 1.1 mmol, 10 equiv.) and nBu4NBr (3.5 mg, 0.022 mmol, 0.1 equiv.). Purification by silica gel column chromatography (pentane/EtOAc = 1:1) afforded 36 (38 mg, 0.104 mmol, 95%) as a yellow solid. A mixture of 8-oxypseudopalmatine (18 mg, 50 µmol, 1.0 equiv.) and PIFA (16 mg, 0.05 mmol, 1.0 equiv.) in DCM (2.0 mL), was added BF3•Et2O (9.7 µL, 0.10 mmol, 2.0 equiv.) at -78 °C. The mixture was stirred for 3 h, then quenched with NaHCO3 (sat.</p><p>aq.) and diluted with DCM, and the organic phase was separated. The aqueous phase was extracted with DCM and the combined organic layers were washed with brine, dried over MgSO4, filtered, and concentrated under reduced pressure. Purification by silica gel column chromatography (pentane/EtOAc = 1:1 to EtOAc) afforded 5 (11 mg, 15 µmol, 60%) as a yellow solid.</p>

ChemRxiv

A simple setup miniaturization with multiple benefits for Green Chemistry in nanoparticle synthesis

The development of nanomaterials often relies on wet-chemical syntheses performed in refluxsetups using round-bottom-flasks. An alternative approach to synthesize nanomaterials is here presented that uses glass tubes designed for NMR analysis as reactors. This approach uses less solvent, uses less energy, generates less waste, provides safer conditions, is less prone to contamination and is compatible with high throughput screening. The benefits of this approach are illustrated by an in breadth study with the synthesis of gold, iridium, osmium and copper sulfide nanoparticles.Introduction. Nanomaterials (NMs) are used in multiple applications ranging from catalysis, optics, medicine to water/air treatments. [1][2][3] Due to strong structure-property relations on the nanoscale, the careful and rational synthesis of NMs is important. Consequently, the controlled syntheses of NMs has been increasingly addressed as a key component of Green Chemistry. [4][5][6][7][8] The need for protocols generating less waste and increasing safety has been stressed. 6,7 A focus is often given to the solvents and reactants to select, 9 and the use of microwaves or ultrasound, considered more energy efficient. 8 The development of setups such as flow (micro)reactors is also useful to minimize waste and simplify NM synthesis. [10][11][12] However, ''despite these advantages, microfluidic systems have yet to be extensively adopted by the colloidal nanomaterial community''. 13 The above strategies often require specific equipment and expertise, which may account for their relatively limited implementation. Simpler miniaturized systems with high throughput potential would gain to be proposed. 3Wet-chemical methods often show promising scalability while being easily implemented in most modern laboratories. [14][15][16] NMs with various compositions, sizes and structures can be prepared by wet-chemical syntheses, which results in tuned NM properties to best match specific requirements for various applications. [16][17][18] The control over NM features is achieved by tuning experimental parameters, e.g. temperature, concentration of reactants, type of reactants or solvent composition.However, studying and understanding how synthetic parameters influence the NMs produced is often limited by the time and resources required to make a single batch of NMs. In a textbook approach using reflux-setups, 10-100 mL of solvent is typically required and only one experiment can be performed every few hours and per setup, see Scheme 1. To understand how NM structures change with synthetic parameters, alternative high throughput strategies are needed.As an alternative to the conventional reflux-setup, we here investigate the use of glass tubes designed for nuclear magnetic resonance (NMR) measurements as miniaturized reactors, 19,20 Scheme 1 and details in the Methods section. The multiple advantages of this simple alternative in Green Chemistry for NM synthesis 6,8 are illustrated by an in breadth study with the preparation of four different nanoparticles (NPs): gold (Au), iridium (Ir), osmium (Os) and copper sulfide (Cu2-xS). Although precious metal (PM) based NPs are made of non-renewable resources, they remain key materials in multiple applications, 7 e.g. to develop fossil-fuel-free technologies for energy conversion, 17 and as models to understand NM formation. 21 Even minor improvements in PM NP syntheses can have significant scientific, economic and ecologic impact.

a_simple_setup_miniaturization_with_multiple_benefits_for_green_chemistry_in_nanoparticle_synthesis

Results and Discussion<!>Conclusions.<!>Chemicals<!>Syntheses.<!>Cu2-xS NP synthesis.<!>Supporting Information.<!>Corresponding Author

<p>Au NPs. Au NPs have been a case study for Green Chemistry in NM synthesis. 7 In particular, the Turkevich synthesis is widely reported, where typically 0.1-0.5 mM HAuCl4 is reduced in water close to the boiling point by trisodium tricitrate. 22,23 The investigation of the many parameters controlling the properties of the Au NPs 22 would gain from high throughput approaches using less solvent than the typical 200-500 mL required. The Au NPs obtained performing the Turkevich synthesis in 500 mL (reflux-setup) or 0.2 mL (NMR tube) lead to the same NP size ca. 13 nm within the error of measurement by TEM, Figure 1. A range of alternative Au NP syntheses use organic solvents such as oleylamine and relatively high temperature. 24 Syntheses at higher temperature, e.g. 200 °C, are possible since nut and ferrule can be used to close the NMR tube, Ir NPs. We reported a surfactant-free colloidal synthesis of PM NPs directly relevant for the industry, 17,25 performed at low temperature (< 80 °C) in alkaline mono-alcohols like methanol or ethanol, leading to high catalytic activity for the oxygen evolution reaction, 17,26 without intensive washing or purification steps, thus already addressing few Green Chemistry principles for NP synthesis. 6,7,27 The Ir NPs are in the range 1-2 nm across a wide range of experimental parameters. 28 To possibly achieve size control, investigating high concentrations of precursor and long synthesis time are reasonable options. Using a microwave or even a classical reflux-setup, it is arguably challenging to investigate these hypotheses due to safety concerns, higher precursor cost and pressure on lab space and equipment access.</p><p>Using NMR tubes, safe and simple time resolved studies over weeks of synthesis using high precursor concentrations (100 mM IrCl3) are easily performed, see Figure 2. In line with previous reports using microwave synthesis 17 or UV-vis induced synthesis 28 it is found that Ir NPs with an average size of 1.6 nm are obtained. A significant achievement here is that the NPs are obtained at high concentration (100 mM) of precursor. This result underlines the relevance of this recently reported Ir NP synthesis to scale up the synthesis of extremely active catalysts for the oxygen reduction reaction. 17 Os NPs. Os NP synthesis has received less attention than other PMs. [30][31][32] Cu2-xS NPs. Cu2-xS NPs have applications in batteries, 36,37 sensors 38 and as an oxygen evolution reaction catalyst. 39 The analysis with a lab-source XRD instrument of the NPs synthesized in the NMR tubes suggests that small Cu2-XS NPs were formed, Figure 4A. TEM analysis shows a size of 4.5 ± 0.9 nm, Figure 4B and Figure S5. While some features are observed in the XRD pattern in Figure 4A, the broadening of Bragg peaks from these small NPs makes any structural investigation challenging. In contrast, the use of the NMR tube makes further characterization straightforward, e.g. using synchrotron scattering techniques, as many measurements can be done directly on the sample in the NMR tube. For instance, the as-prepared samples in closed NMR tubes can be sent to synchrotron for X-ray total scattering with pair distribution function (PDF)</p><p>analysis. The PDF can be understood as a histogram of atom-atom distances in the material. In Benefits. The use of NMR tubes readily lead to several advantages in terms of synthesis strategy.</p><p>Using 3 mm diameter NMR tubes, allows to perform syntheses with volume as low as 0.1 mL,</p><p>. This low volume reduces the amount of waste generated and allows investigating the effect of high concentration solutions towards scaling-up. 7,16 Sample holders can be designed to perform several experiments at a time (up to 9 experiments per holder in the current design, Figure S7) allowing high throughput screening while requiring minimal lab space. Small volume also allows safer operating conditions. As a result, both long(er) and an increased number of experiments can be performed with little equipment required, while optimizing the energy needed to heat up the solutions. Temperature control is achieved by either using the temperature control of a heating plate or using dedicated heating cartridge with a temperature controller to control the heating rate, Figures 5 in the Methods section. Septa can be used to close the NMR tubes to control the atmosphere, alternatively, nut and ferrules can be used for higher pressure experiments. The length of the NMR tube provides an area of contact with cool air to function under reflux conditions. This area can also be cooled down, e.g. with a fan or dedicated water-cooling devices.</p><p>It is worth stressing that the synthesis can be performed with stirring by simply using a commercially available stirrer bar (e.g. 8 mm x 1.5 mm) in the NMR tube, Figure 2.</p><p>Further considerations. NMR tubes have previously been used as reactors for in situ studies at synchrotrons or for NMR characterization. 20,41 However, they have often been used not as a preferred reaction vessel but to make the experiment compatible with a given type of measurement, e.g. NMR. The experimental conditions selected to perform these specific experiments are often the results of screening performed on larger scale syntheses. A drawback of screening studies developed and optimized in laboratories is thus that they are often challenging to directly adapt to the requirements and limitations of synchrotron setups. For example, PDF measurements are best performed on samples with high concentration of the material of interest, especially when considering poorly scattering elements like transition metal sulfides. Here, the NMR tube setup offers the option to readily investigate high concentration of materials during parametric studies performed in home laboratories (outside synchrotrons facilities). Ultimately, this allows comparing more directly selected measurements performed at synchrotron facilities with larger screening experiments done in a chemistry lab. This further enables researchers to design experiments for even more complex in situ analysis of various reactions. 19,42,43 NMR tubes make it possible to use a clean vessel for each experiment. This alleviates the need to clean glassware, thus limiting the chemical waste generation related to cleaning steps, e.g. using aqua regia for Au NPs. 44 Importantly, it alleviates the question of cross contamination, thus addressing the well-known issue of reproducibility in NM science, often ascribed to chemical impurities from various sources. 45 NM synthesis can be sensitive to variations of room temperature or room light, and the stability overtime of the chemicals and/or precursors solutions can be an issue. 46 These variables may be challenging to control across long periods of time when a study would last for instance several weeks of months. This drawback is alleviated with NMR tubes as reactors since several experiments including controls can be performed at the same time. Energy is saved because several experiments are heated all at once. The ease of simply performing relatively long experiments is also a positive feature towards improved yield of the precursor conversion to NMs.</p><p>It could be argued that due to the small diameter of the NMR tubes, capillary effects might come into play and the actual temperature-pressure during reaction might not be well-known. This issue of variability in physical parameters of reactors, e.g. heat transfer properties, stirring etc., is the same for most scaling up to date, e.g. when moving from reflux-setups to larger scale reactors. In this respect, the use of NMR tubes reactors remains a convenient and green approach to (pre)screen the influence of experimental parameters.</p><p>A final practical consideration is storage and reuse. The small NMR tubes are easily stored due to their small diameter and length. The 18 cm long NMR tube are easily cut with a commercially available glasscutter to save more space. The open end of the glass can be sealed by melting with a torch (e.g. butane torch), for instance when toxic reagents and/or air sensitive compounds are involved. The samples can be safely and space-efficiently stored for further analysis. Additionally, the cut section of the NMR tube can be cleaned and sealed by melting one extremity to be re-used as a new miniaturized vessel.</p><!><p>A simple alternative to the classical round-bottom flask synthesis approach to prepare NMs is presented. An in breadth study of four different nanomaterials is presented covering gold, iridium, osmium and copper sulfide nanomaterials. The general approach is relevant for the synthesis of other nanomaterials like palladium, platinum or silver that would be performed in similar solvent and/or for the same temperature range. 17,47 By using commercially available NMR tubes, a miniaturized vessel suitable for the synthesis of various NMs is readily obtained.</p><p>This approach complies with several of the principles of Green Chemistry for NM synthesis. 6,7 12 This approach is therefore directly relevant for academic research and research and development but also educational purposes.</p><p>With the increasing interest in machine learning and artificial intelligence, large datasets are needed to feed algorithms. 48 The present approach allows for high throughput (pre)screening.</p><p>Multi-technique characterization remains a general limiting factor in nanoscience 49 but the expected improvements in this area of research are promising. 6,19,41,50 If characterization was kept minimal in this in breadth approach, enough material is still obtained for characterization by TEM, XRD or PDF and naturally NMR, high resolution TEM or UV-vis characterization with an appropriate setup. It is expected that the simple alternative proposed here will also be relevant for various syntheses of molecules and other chemical reactions.</p><!><p>All chemicals were used as received: HAuCl4</p><!><p>Note: Some reactions can be dangerous, here in particular the synthesis of Os nanoparticles (NPs), since OsO4 can easily form and is highly toxic. The synthesis must be performed in a fumecupboard. Conveniently, the synthesis can be performed in NMR tubes closed with a cap (red cap Precision seal rubber setpa cap (for 3 mm OD tubes and ampoules) pack of 100 ea) and further sealed with Parafilm®. Alternatively, the NMR tube can be melted and sealed for the experiment or at the end of the experiments, e.g. with a butane torch. Otherwise nuts and ferrules (Swagelok) can be used. These opportunities to seal the vessel are also relevant for further storage.</p><p>Au NP synthesis. The Au NPs were obtained following the general procedure for the Turkevitch synthesis. 22,23 The Au NPs were produced from a mixture of 0.125 mM HAuCl4 with 2.2 mM trisodium citrate in water, for a total volume of 500 mL in a reflux setup or 0.2 mL in a 3 mm diameter NMR tube. The solution containing trisodium citrate was pre-heated at 100 °C before the gold-containing solution was added (the volume before gold injection was 499 mL or 0.1 mL). In both cases, a hot plate was used to heat up the solution to 100 °C and the reaction after adding gold was left to pursue for 1 hour at this temperature of 100 °C with a stirring rate set at 1000 rotation per minutes.</p><p>Alternatively, the Au NPs were obtained using organic solvents. 24,51 To a glass vial, 9 mL oleylamine (OLA) and 25 mM Au(ac)3 were added. The precursor solution was heated to 50 °C under air, while stirring at 400 rotation per minutes (rpms) until the metal ion precursor was dissolved. 4.5 mL of the precursor solution was taken to a separate vial, into which 25 μL dodecanethiol (DDT) was added and further heated to 50 °C under air, while stirring at 400 rpms. 0.3 mL of the solution was added to an NMR tube and closed by a Swagelok nuts and ferrules.</p><p>The NMR tube was placed in the heating block, where the solution was heated to 200 °C at a heating rate of 7 °C min -1 and maintained at this temperature for 1 hour. The solution was cooled to room temperature. The resulting solution was added to a centrifuge tube with 20 mL hexane and was spun at 10 000 rpms for 10 minutes to precipitate the NPs. The particles were washed with a 3:1 ratio of ethanol: hexane three times. The particles were then suspended in 5 mL of hexane for future use.</p><p>Ir NP synthesis. The Ir NPs were obtained following the general procedure previously reported. 17,28,52 A mixture of 100 mM IrCl3 in 1 M NaOH in methanol for a total volume of 0.2 mL was placed in a 3 mm diameter NMR tubes with a magnetic stirrer bar. The filled NMR tube was placed in a dedicated holder as presented in Scheme 1 and Figure 5. A hot plate was used to heat up the miniaturized vessel to 60 °C for up to 1 week with a stirring set at 1000 rotation per minutes. The solution, which is initially light brown, turns black over time, see Figure 2.</p><p>Note: We previously used 4.4 mM in 10 mL 17 or 27 mL 52 to get enough mass of iridium, 8.5 mg and 23 mg respectively, for electrochemical testing. Using here ca. 20 times less solvent, the same mass of Ir NPs can be obtained due to the high concentration used.</p><p>Os NP synthesis. The Os NPs were obtained from a mixture of 100 mM OsCl3 in methanol:water (1:2; v:v) for a total volume of 0.2 mL placed in a 3 mm diameter NMR tube with a magnetic stirrer bar. The filled NMR tube was placed in a dedicated holder as presented in Scheme 1 and Figure 5. A hot plate was used to heat up the miniaturized vessel to 85 °C for 1 week with a stirring set at 1000 rotation per minutes.</p><!><p>To form copper sulfide NPs, a synthesis reported in the literature was modified. 53 A copper oleate precursor was obtained by heating 1 mmol of Cu(II) acetate with 2 mL of oleic acid in an oil bath at 150 °C for 1 hour. To 2 mL of copper oleate, 0.2 mL DDT was added, resulting in a viscous white precipitate. The mixture was mixed using a table-top vortex to ensure solution homogeneity. The copper oleate and DDT mixture was then added to 9.5 cm x 3 16 mm NMR tubes (obtained after cutting the initially 18 cm long tubes with a glasscutter) with approximately 0.25 mL using a needle and luer lock syringe. The NMR tubes were filled with N2 and capped with a 3 mm silicone septa. The bottom of the septa was secured with epoxy and then wrapped in Teflon tape. The NMR tubes were then put into the heating setup equipped with a heating cartridge and a type J thermocouple that were plugged into a Cole parmer Digitroll II temperature controller. The samples were heated with a ramp rate of 185 °C h -1 to a final temperature of 200 °C and held at 200 °C for 3 hours before being cooled to room temperature. Two identical samples were prepared in NMR tubes so that three different analyses could be done: lab-source XRD, TEM and PDF analysis of X-ray total scattering data collected at 11-ID-B at Argonne National lab. Therefore, the NPs in one of the NMR tubes were washed by transferring the reaction liquid using a needle and a luer lock syringe to a 50 mL centrifuge tube. Approximately 5 mL of hexanes was used to rinse the NMR tube to ensure that as much of the reaction liquid as possible was transferred for washing. Acetone was then added to the NMR tube and it was centrifuged for 10 min at 9000 rpm. Particles were redispersed in hexanes and then washed one more time with acetone for 5 min and again with ethanol for 10 minutes before lab-source XRD and TEM analysis. solvent sample and an empty capillary were placed in an aluminum block for data collection at room temperature. The X-ray wavelength was 0.2115 Å and data were collected using a Perkin Elmer detector (200 um x 200 um pixel size) with a detector distance of 183.5 mm. Data were collected on the empty capillary, solvent and sample for 5 min. The data were integrated using Fit2D. 54 The data obtained from the glass capillary were subtracted from the data obtained from the reaction solution and solvent samples before using xPDFsuite 55,56 to subtract the signal from the solvent from the reaction solution and generate the PDF. A Qmax instrument of 18.5 Å -1 , Qmax 18 Å -1 , Qmin 1.5 Å -1 , and rpoly 1.1 were used to generate the PDF.</p><p>PDF refinements. CeO2 was used for the calibration for detector distance as well as to determine the Qbroad and Qdamp parameters of the instrument. The PDF from the CeO2 standard was modelled in the range from 1.3 Å to 60 Å with a Qbroad value of 0.0385 Å -1 and Qdamp value of 0.0373 Å -1 using the CeO2 crystal structure model. Modelling of the sample PDF was done using PDFgui using the β-chalcocite structural model in space group P63/mmc. Below are Tables S1-S2 of refined parameters. Two refinements were done at two different r-ranges. The first was done using the chalcocite structural model refined to the PDF in the range 2 Å to 30 Å, Table S1. The second was done using the chalcocite structural model refined to the PDF in the range 5.6 to 30 Å. The delta2 parameter (describing correlated motion) and spdiameter were fixed at values refined in the first refinement, Table S2.</p><!><p>The following files are available free of charge. TEM and PDF characterization, Sample holder specification, Figure S1-S7, Further discussion (PDF).</p><!><p>* kirsten@chem.ku.dk, jonathan.quinson@chem.ku.dk</p>

ChemRxiv

Studies of Nature of Uncommon Bifurcated I–I···(I–M) Metal-Involving Noncovalent Interaction in Palladium(II) and Platinum(II) Isocyanide Cocrystals

Two isostructural trans-[MI2(CNXyl)2]·I2 (M = Pd or Pt; CNXyl = 2,6-dimethylphenyl isocyanide) metallopolymeric cocrystals containing uncommon bifurcated iodine···(metal–iodide) contact were obtained. In addition to classical halogen bonding, single-crystal X-ray diffraction analysis revealed a rare type of metal-involved stabilizing contact in both cocrystals. The nature of the noncovalent contact was studied computationally (via DFT, electrostatic surface potential, electron localization function, quantum theory of atoms in molecules, and noncovalent interactions plot methods). Studies confirmed that the I···I halogen bond is the strongest noncovalent interaction in the systems, followed by weaker I···M interaction. The electrophilic and nucleophilic nature of atoms participating in I···M interaction was studied with ED/ESP minima analysis. In trans-[PtI2(CNXyl)2]·I2 cocrystal, Pt atoms act as weak nucleophiles in I···Pt interaction. In the case of trans-[PdI2(CNXyl)2]·I2 cocrystal, electrophilic/nucleophilic roles of Pd and I are not clear, and thus the quasimetallophilic nature of the I···Pd interaction was suggested.

studies_of_nature_of_uncommon_bifurcated_i–i···(i–m)_metal-involving_noncovalent_interaction_in_pall

Introduction<!><!>Introduction<!><!>Complexes 1 and 2 and Their Cocrystals<!>Single-Crystal X-ray Diffraction (SCXRD) Analysis<!><!>Single-Crystal X-ray Diffraction (SCXRD) Analysis<!>Theoretical Studies of Noncovalent Interactions<!>ESP Analysis<!><!>ESP Analysis<!>NCI-plot Analysis<!><!>NCI-plot Analysis<!>Philicity Definition: Analysis of ELF and ED/ESP Minima<!><!>Philicity Definition: Analysis of ELF and ED/ESP Minima<!><!>Philicity Definition: Analysis of ELF and ED/ESP Minima<!><!>Philicity Definition: Analysis of ELF and ED/ESP Minima<!><!>Philicity Definition: Analysis of ELF and ED/ESP Minima<!>LED Analysis<!><!>Conclusion<!>General Computational Details<!>Materials and SCXRD Details<!>Synthesis of trans-[PdI2(CNXyl)2]<!>Synthesis of I2 Cocrystal of trans-[PdI2(CNXyl)2]<!>Synthesis of trans-[PtI2(CNXyl)2]<!>Synthesis of I2 Cocrystal of trans-[PtI2(CNXyl)2]<!><!>Accession Codes<!>Author Contributions<!>

<p>Noncovalent interactions (NCIs) are a powerful instrument applied in such fields as synthesis,1 catalysis,2,3 design of photoactive materials,4−6 and biochemistry.7,8 Halogen bonding (XB), in particular, has been found to be a very useful NCI, for example, in the synthesis of self-assembled polymers,9,10 due to its high directionality and possibilities for fine-tuning. Recently, XB has been utilized in our research to create metallopolymeric systems.11 Known types of metal–halide interactions involved in the self-assembly of metallopolymers include classical XB12 (Figure 1A) and semicoordination bond via electron belt (Figure 1C).</p><!><p>Types of NCIs involving metal centers (M) and halogen atoms (drawn as black ovals), where electrophilic regions are colored as red and nucleophilic ones are colored as blue: metal-involving halogen bond (A); intermediate metal–halogen interaction (B); semicoordination bond (C).</p><!><p>In cocrystals of metal complexes, classical XB is represented by donor/acceptor interaction of an electron-deficient area (σ-hole) located on a XB donor (XBD) and an electron-rich area located either on a ligand or on the metal center itself. In the case of an interaction with square planar d8, linear d10 transition metal complexes, or metal surface, an electron lone pair on the d orbital acts as the nucleophile, while a σ-hole of a halogen atom acts as the electrophile. The first examples of the possible metal-involving XBs were represented by van Koten et al.13−16 for the I–I···PtII bonds between diiodine and NCN pincer PtII complexes. Theoretical investigations of these interactions showed that they are rather strong and comparable with coordinative bonds.17,18 Further works of van Koten et al. showed that the analogous palladium and nickel NCN pincer complexes interact with diiodine in other ways.19−21 Nevertheless, later works represented metal-involving XB not only with PtII22−28 and PdII23,27,29,30 but also with NiII,27,31 RhI,32,33 AuI,34−36 and Au0 centers37−41 as nucleophiles.</p><p>In contrast to XB, a semicoordination bond31 occurs when an electrophilic region of a metal center is interacting with the electron belt of a halogen atom or a nucleophilic halide anion. Particularly, examples of PdII···I31,42−47 and PtII···I48−50 semicoordination bonds have been described in the literature.</p><p>Both types of discussed noncovalent interactions between metal centers and halogen atoms can be considered polar NCIs (with clear electrophilic or nucleophilic51 roles assignable to interacting atoms). In this connection, it is worth noting the well-defined nonpolar NCIs (with unclear electro- or nucleophilic roles) between the halogen atoms (type-I halogen···halogen interactions caused by dispersive forces)3 and metallophilic interactions (closed-shell (d10, s2) or pseudoclosed shell (d8) weak attractive metal···metal contacts presumably dominated by electrostatic and dispersion forces).52 Although possible a nonpolar NCI involving metal center and halogen atoms is mentioned for the so-called C–I···Ni boundary case,31 the nature of nonpolar interaction (such as philicity of interacting centers and energy components) between halogen and metal atoms has never been studied thoroughly prior to this work.</p><p>As a continuation of our studies of metal-involving interactions53 and halogen bonding,26 especially between molecular iodine and iodide isocyanide complexes,53,54 the association of molecular iodine with trans-[MI2(CNXyl))2] (M = Pd (1) or Pt (2); CNXyl = 2,6-dimethylphenyl isocyanide) species was studied. The simple structure of molecular iodine allows a high level of control in the self-assembly of noncovalently bound metallopolymers. Bearing both an electron-deficient region of a σ-hole and an electron-rich area of an electron belt, I2 is prone to interact with both electrophilic and nucleophilic regions of other molecules.55 In addition, the relatively small size of the I2 molecule allows it to overcome steric constraints. Furthermore, square planar trans-[MI2(CNXyl)2] are promising building blocks in organometallic chemistry due to stabilizing metal–carbon π interactions.56,57 An occupied dz2 orbital of these Pd and Pt complexes is accessible for interaction, which opens up the opportunity for the generation of metal–involving XB systems.</p><p>In the current work, molecular iodine forms isostructural metallopolymeric cocrystals, trans-[MI2(CNXyl)2]·I2 (M = Pd or Pt), where complex units are noncovalently linked via I2 molecules (Figure 2). Careful analysis of experimental and theoretical data along with a literature search revealed atypical I–I···(I–M) bifurcated noncovalent bonds, in which classical halogen bond is additionally stabilized by an uncommon type of an I···M contact between a metal center and halogen atom. In this contact, the halogen atom is neither interacting via a σ-hole (Figure 1A) nor via an electron belt (Figure 1C), but presumably via a transitional area (Figure 1B). To understand the nature of this intermediate contact, it was comprehensively studied with various bond analysis methods such as electrostatic surface potential (ESP) analysis (to discover the angle limits of a σ-hole), NCIs plot (NCI-plot) analysis (to reveal the relative strength of the interaction), electron density (ED)/ESP analysis (to assign philicity of the interacting atoms), and local energy decomposition (LED) analysis (to indicate which interaction type best describes the contact).</p><!><p>Representations of trans-[MI2(CNXyl)2]·I2 (M = Pd or Pt) polymeric crystal structures visualizing π–π stacking (left) and noncovalent bifurcated I–I···(I–M) contact (right) in both cocrystals along the a-axis.</p><!><p>Syntheses of complexes trans-[PdI2(CNXyl)2]58 (1) and trans-[PtI2(CNXyl)2]59 (2) are presented in Scheme 1. A similar trans-[PdBr2(CNXyl)2] complex60 has been described previously. Cocrystals 1·I2 and 2·I2 were grown from 1:1 CH2Cl2/CHCl3 and CHCl3 solutions of a 1:1 mixture of the corresponding complex and I2, respectively.</p><!><p>Although syntheses of 1(58) and 2(61) are known, no SCXRD data was found for these complexes in the Cambridge Structural Database (CSD). Isostructural complexes 1 and 2 exhibit the same monoclinic lattice of the P21/c space group (for details, see the Supporting Information, "Single crystal X-ray Diffraction Data Analysis (SCXRD)" section). The complexes have square-planar structures with iodide ligands in the trans position to each other. Cocrystals of 1·I2 and 2·I2 have both a 1:1 molar composition; exhibit the same triclinic lattice P1̅ space group, and similar unit cell parameters, being isostructural to the original complexes. The fragments C–N–C–M are almost linear in both complexes (∠(Pd–C–N) = 179.2(4)° and ∠(Pt–C–N) = 178.6(4)°) and corresponding cocrystals (∠(Pd–C–N) = 179.4(10)° and ∠(Pt–C–N) = 179.0(12)°). The most notable difference between original 1 and 2 complexes and the corresponding cocrystals is in the position of the iodide ligands: While in 1 and 2 the iodide ligands are located in the same plane as the xylene rings of the CNXyl ligands, in cocrystals the iodide ligands are tilted away from the plane allowing interaction of I2 molecule with the complex. Additionally, in cocrystals CNXyl ligands are arranged in π–π stacks with centroid–centroid distances of 3.86–3.88 Å in 1·I2 and 3.87–3.91 Å in 2·I2 (Figure 2), whereas in original 1 and 2 complexes, this type of stacking is not observed.</p><p>The relative strengths of NCIs can be approximated by a comparison of the experimentally obtained distances between noncovalently interacting atoms and the sum of corresponding van der Waals radii (vdW).62 The distance reduction ratio of NCI (RIX) can be calculated as RIX = d(I···X)/(RIvdW + RXvdW), where I (iodine) represents a halogen bond donor (XBD) atom, X is a halogen bond acceptor (XBA) atom, d(I···X) is the distance between I and X in Å, and RIvdW and RXvdW are the vdW radii by Bondi63 of I and X in Å, respectively. The comparison shows that the characteristic parameters of the interactions correlate closely with each other emphasizing the isostructural nature of cocrystals (Figure 2, Table 1).</p><!><p>RIX = d(I···X)/(RIvdW + RXvdW), where RIX is distance reduction ratio, I is a donor atom, X is an acceptor atom (I, Pt, Pd), and d(I···X) is the distance between I and X in Å; RIvdW and RXvdW are the vdW radii of I and X correspondingly determined by Bondi.63</p><!><p>The uncommon bifurcated I–I···(I–M) contact can be subdivided into two types of NCIs: I···I and M···I. Within the cocrystals, the relative strength of the XB is rather similar: For I···I XB, RIX is 0.88 for 1·I2 and 0.89 for 2·I2; for M···I interaction, RIX is 0.94 for 1·I2 and 0.93 for 2·I2. Hence, in both cocrystals I···I XB is slightly stronger than M···I interaction. This might indicate the main role of I···I XB in the interaction (which is further confirmed in theoretical analysis of the structures, vide infra). Another parameter attracting attention is the ∠I–I···M angle in both cocrystals, which is significantly more acute (about 128°) than that of classical XB (180°)12 (Table 1). Variation of this parameter brings up a question on the nature of I···M interaction; thus, it was further studied computationally.</p><p>The I···Pd distances in 1·I2 are shorter by ∼0.06 Å than the same I···Pt distances in 2·I2 (Table 1). At the same time, the electron density values in the corresponding I···M bond critical points (BCPs) are similar (the difference is less than 0.001 a.u., see Table 3 in section 2.3.2). According to these observations, the vdW Pd radius may be similar or only slightly shorter than the Pt radius. This hypothesis is in disagreement with Bondi's vdW radii (1.63 Å for Pd vs 1.75 Å for Pt),62 and further detailed studies should be carried out in this direction.</p><p>Comparing M–I bond length in the cocrystals and in the corresponding complexes (Table 2), we observed elongation of the M–I bond in cocrystals, presumably due to a strong influence of the halogen bonding with I2 in the cocrystals. This elongation was also found in a few other cocrystals of square planar Pt complexes having different XBDs,22,64 and the M–I bond elongation is likely to be found in similar systems of square planar transition metal complexes interacting with XBD.</p><p>The I–I bond length is elongated in 1·I2 and 2·I2 cocrystals in comparison to that in the solid-state structure of I265 (Table 2). This elongation of a covalent bond in XBD is typical for a XB according to IUPAC XB definition.12</p><p>Although according to RIX value halogen bonding seems to be the strongest NCI, it is important to take into consideration the combination of all the involved NCIs like I···M interaction and π–π stacking. The significance of the discussed interactions for the structure arrangement was further elucidated by various computational methods (see the "Theoretical Studies" section and the "QTAIM Analysis" section of the Supporting Information, where QTAIM = quantum theory of atoms in molecules).</p><!><p>With the help of computational chemistry, the nature and relative strength of NCIs discovered by SCXRD can be thoroughly studied. Careful analysis of the calculated electron density distribution can reveal NCIs and their properties. A combination of several approaches such as analysis of ESP,66,67 NCI-plot68 analysis, combined electron localization function (ELF)69 and Bader's quantum theory of atoms in molecules (QTAIM) analysis,70 and LED71 analysis gives a broad look on NCIs. To support the idea that observed interactions are not only caused by packing effects, the data of single-point (SP) structures and optimized (OPT) ones were compared (for more details, see the Experimental Section).</p><!><p>Observed NCIs can be clarified by analysis of anisotropic charge distribution, which is visualized by ESP.3,67,72−76 ESP visualizes electron-rich and -deficient areas of the molecule that are likely to participate in electrostatic intermolecular interactions. This helps to estimate the geometries and expected strengths of the XB interactions. The strength of the XB formed by the XBD is related to the magnitude of σ-hole77 on the XBD atom that can be described by the maximum of ESP (Vs,max). The influence of the XBA on the XB can be estimated using the minimum of ESP (Vs,min) on the XBA atom electron density surface. ESP analysis was carried out on the 0.001 a.u. contour of molecule's electron density (that encompasses 96% of the molecular charge).78</p><p>Anisotropic charge distributions of I2, trans-[PdI2(CNXyl)2] (1), and trans-[PtI2(CNXyl)2] (2) were analyzed, and the corresponding ESPs are represented in Figure 3. An electron-deficient area corresponding to the σ-hole of the I2 molecule was calculated with Vs,max = 139 kJ mol–1 which is reasonably close to the Vs,max value (127 kJ mol–1) reported by Kolář et al.79 at much higher ab initio QCISD/def2-QZVP level of theory. As was suggested by the X-ray diffraction analysis, the I2 molecule is expected to behave in the cocrystals as a XBD, interacting with complexes 1 or 2 that act as XBAs. To participate in XB, complexes 1 and 2 are required to bear an electron-rich area around the I or M (M = Pd or Pt) atom. Indeed, ESP studies of 1 and 2 confirm the electron-rich areas (Vs,min) for iodine atoms (Vs,min = −102 kJ mol–1) and for the Pd (Vs,min = −81 kJ mol–1) and Pt centers (Vs,min = −89 kJ mol–1). These local nucleophilic areas roughly correlate with the regions of I···I and I···M interactions.</p><!><p>Electrostatic potential calculated at the M06L/def2TZVP/def2TZV computational level on the 0.001 a.u. molecular surface of I2, 1, and 2 with the color scale from −102 kJ mol–1 to 102 kJ mol–1.</p><!><p>To analyze if the uncommon I···M contact could be caused by XB-type interactions, we determined the σ-hole limiting angle80 [∠(I–I···XBA)] for the I2 molecule. The σ-hole limiting angle helps to estimate the angle range where nucleophilic atom can approach the electron-deficient area of I atom with a favorable electrostatic attraction. The limits of the interaction with the σ-hole were found to be 115–180° (see Figure S4). In the case of 1·I2 and 2·I2 cocrystals, ∠(I–I···M) is around 128° allowing M atoms to interact with the σ-holes of I2. While this provides evidence of the likely existence of I···M interaction, it does not give direct information on the nature of the interaction and further computational analyses were carried out to achieve this.</p><!><p>NCI-plot analysis is a powerful method to reveal the repulsive or attractive nature of the interaction and to describe the relative strength of noncovalent bonding.81,82 2D and 3D NCI-plots visualizing all the interactions in 1·I2 and 2·I2 cocrystals can be found in Figures S8–S13. Here only the interactions involved in the bifurcated contact are discussed.</p><p>For (1)4·I2 and (2)4·I2 clusters 2D plots of (s) against sign(λ2)ρ have a similar shape (see Figures S8 and S9). Two types of attractive NCIs (I···I and I···M XBs, where M = Pt or Pd) were found in the [−0.02, −0.008] a.u. range of sign(λ2)ρ and one type of repulsive interaction found in the [0.009, 0.018] a.u. range of sign(λ2)ρ (Figure 4). The repulsion areas for the I–I···(I–M) interactions can be explained by the repulsion of lone pairs of the metal center and iodide ligand in 1 and 2; the same areas can be found in isolated 1 and 2 (see the sign(λ2)ρ projections in Figure S14). Expectedly, the strength of XB in both cocrystals was found to be very similar. This observation correlates with data obtained experimentally (see Table 1).</p><!><p>NCI visualizations of the bifurcated contact for SP (left) and OPT (right) trans-[MI2(CNXyl)2]4·I2 clusters (where M = Pd or Pt). Corresponding 2D graphs, as well as full 3D visualizations containing all the interactions, can be found in Figures S8–S13.</p><!><p>Especially intriguing is the difference between SP and OPT structures. As expected, the strength of all interactions weakens in the optimized structures (Figure 4, Table 3). In the case of SP structures, I···I and I···M contacts have very similar interaction strengths, while in the OPT structures I···M contact is weaker. In the case of (2)4·I2, the change is more noticeable (i.e., interactions are more weakened) than that in the case of (1)4·I2 (Table 3).</p><!><p>ELF is useful in the investigation of XBs and related interactions.24,83−89 As a derivative of electron density ELF69,90−92 allows to locate areas of shared and unshared electron pairs. A combination of ELF and QTAIM70 methods visualizes bond paths at the interaction areas and facilitates defining the philicity of interacting atoms.26,93</p><p>Combined ELF and QTAIM analysis information for SP and OPT (1)4·I2 and (2)4·I2 model structures is presented in Figure 5 as projections on a plane formed by metal atoms, iodide ligands, and iodine molecules. In all four analyzed structures, the I–I···(I–M) bond paths go through the increased ELF areas on the iodides (i.e., through the lone pairs) and through decreased ELF regions on the diiodine I atoms (i.e., through the σ-holes). These observations support the XB nature of the I–I···(I–M) contacts where iodide ligands behave as nucleophiles toward electrophilic diiodine molecules. Similar behavior was observed in the case of the I–I···(I–Pt) XBs in [PtI2(1,5-cyclooctadiene)]·0.5I2 in our previous work,26 where the I···I bond paths go through the σ-hole (iodine in I2) and the lone electron pair (iodide ligand) ELF regions.</p><!><p>ELF projections with plotted contour lines (black, step is 0.05), bond paths (white lines), BCPs (blue dots), nuclear critical points (NCPs, brown dots), ring critical points (RCPs, orange dots), and cage critical points (green dots) for the I–I···I and I–I···M interactions in the SP (1)4·I2 (upper left), OPT (1)4·I2 (upper right), SP (2)4·I2 (lower left), and OPT (2)4·I2 (lower right) model clusters.</p><!><p>ELF projections show increased ELF areas around Pd and Pt atoms above and below the bond paths connecting metal centers and iodide ligands that can be interpreted as filled dz2 orbitals. The M···I bond paths that connect metal centers and I2 molecules go through these dz2 orbitals. However, the ELF values suggest only relatively weak concentrations of electron pairs in areas occupied by dz2 orbitals and the areas lack directional dependence outside the plane formed by metal centers and iodide ligands suggesting that metal centers are likely to act as weak nucleophiles at most. Actually, any bond path corresponding to NCI with the metal center would cross the area of dz2 orbitals, because any d8-metal-involving interaction is required to stay away from the ligands in the complex plane, and the nature of the NCI will depend more on the atom of the interacting partner. At the same time, in all four clusters the I···M bond paths connect to I2 iodine atoms through areas with intermediate ELF values which could indicate that the interactions are either weakly polar or nonpolar. Since the I···M bond paths connect atoms in each structure through areas described by intermediate ELF values, combined ELF and QTAIM analysis does not provide conclusive evidence on the philicity of atoms in these interactions.</p><p>An alternative way to assign philicity of noncovalently interacting atoms is to compare the minima of the electron density (ED) and ESP along the bond path.24,87,88,94−97 In polar NCIs the minimum of ESP is shifted toward the nucleophilic atom, while the ED minimum is shifted toward the electrophilic atom. In the case of the I–I···(I–M) interactions (M = Pd or Pt), 1D profiles of the ED and ESP functions along the I–I···I–M bond paths confirm the iodide nucleophilicity toward diiodine in all four clusters as shown in Figures S15 and S16.</p><p>In the case of I···Pd bond paths in (1)4·I2, the 1D profiles of the ED and ESP functions (Figure 6) show that their minima overlap both in the SP and OPT structures. Together with the combined ELF and QTAIM analysis information, this suggests that I···Pd interactions are best described as nonpolar with Pd and I atoms having similar roles. It is noteworthy that similar interactions, which can be also called intermediate between semicoordination (electrophilic metal center) and metal-involving halogen bonding (nucleophilic metal center), have been previously reported for a Ni(II) complex.31</p><!><p>1D profiles of the ED (black) and ESP (red) functions along the I···Pd bond paths in (1)4·I2 for SP (upper graphs) and OPT (lower graphs) structures.</p><!><p>The nonpolar noncovalent I···Pd interactions in (1)4·I2 are reminiscent of the noncovalent metal center involving interactions in related palladium and platinum chloride isocyanide complexes,98−101 where metal centers participate in nonpolar metallophilic Pd···Pd and Pt···Pt bonds. To compare these interactions DFT SP calculations (M06-L/def2-TZVP) were carried out for two model clusters (cis-[PdCl2(CNPh)2])2 and (cis-[PtCl2(CNPh)2])2, based on the experimental X-ray data from the structures COYBOI01 and CPICPT12,99 respectively. Combined ELF and QTAIM analysis of the (cis-[MCl2(CNPh)2])2 (M = Pd or Pt) clusters indicated the expected existence and nonpolar noncovalent nature of the Pd···Pd and Pt···Pt interactions (see Figure S18). Further confirmation of the nonpolar nature of the metallophilic interactions is provided by the 1D profiles of the ED and ESP functions along the M···M bond paths in (cis-[MCl2(CNPh)2])2 (M = Pd or Pt) clusters (Figure 7) where ED and ESP minima overlap in both cases.</p><!><p>1D profiles of the ED (black) and ESP (red) functions along the Pd···Pd bond path in (cis-[PdCl2(CNPh)2])2 (left) and the Pt···Pt bond path in (cis-[PtCl2(CNPh)2])2 (right).</p><!><p>The ED/Laplacian of ED values in Pd···Pd (0.012/0.030 a.u.) and Pt···Pt (0.016/0.038 a.u.) BCPs in (cis-[MCl2(CNPh)2])2 clusters are similar to the values in the I···Pd BCPs in (1)4·I2 SP (0.016/0.037–0.038 a.u.) and (0.012/0.027 a.u.) OPT structures. The nonpolarity of the I···Pd interactions in (1)4·I2 and the similarity of their strength to metallophilic interactions leads us to designate them as quasimetallophilic interactions.</p><p>Comparison of the 1D profiles of ED and ESP along the I···Pt bond paths in (2)4·I2 SP and OPT structures (Figure 8) shows that the ESP minima are slightly shifted toward the Pt atoms. The shift indicates that Pt atoms act as weak nucleophiles toward I2 iodine atoms, and the I–I···Pt can be treated as metal-involving halogen bonding.22,23,25 Interestingly the more nucleophilic character of PtII compared to PdII in isostructural cocrystals was previously observed for the X2CH–X···M (X = Br and I; M = Pd or Pt) halogen bonding.23</p><!><p>1D profiles of the ED (black) and ESP (red) functions along the I···Pt bond paths in (2)4·I2 SP (upper graphs) and OPT (lower graphs) structures.</p><!><p>However, the small values of the shifts between ED and ESP minima and the ∠(I–I···Pt) angles (128.1–128.3° in SP and 126.5° in OPT structures) that are far from linear leave open the possibility of considering the I···Pt interactions as having intermediate31 philicity i.e. treating them as nonpolar interactions.</p><!><p>The relative energy contributions different types of interactions have to the total interaction can be estimated with local energy decomposition (LED) analysis.102 To elucidate the nature of the I···M and I···I interactions, further local energy decomposition (LED) analysis71 on DLPNO–CCSD(T)103−105/def2-TZVPP106,107 wave function for fragments depicted in Figure 9 was carried out. The LED analysis results are given in Table 4. Comparison of the interaction energies between fragments 6 (I2) and 7 (trans-[MI2(CNXyl)2], M = Pd or Pt) with energies between fragments 2 (I) and 6 suggest that the interaction between I2 and trans-[MI2(CNXyl)2] in both cocrystals of Pd and Pt complexes is almost solely due to the XB between I2 and iodide coordinated to M. The interaction between M atom and I2 appears to have only a minor supporting role to the total interaction between I2 and the complex. This conclusion is in accordance with SCXRD analysis data (based on RIX index, I···I XB is slightly stronger than I···M, Table 1). The I···I2 XB interaction is classified as mainly electrostatic by the LED analysis with small covalent and dispersion contributions. The weaker I···M interaction has higher contributions from covalent and dispersion terms than does the stronger I···I XB interaction in line with the other analyses that described the I···M interaction as weakly polar or nonpolar.</p><!><p>Fragments of (trans-[MI2(CNXyl)2])2·I2 structures (M = Pd or Pt) used in the local energy decomposition analysis.</p><p>Exchange interaction, Eexch; electrostatic and polarization energy, Eelstat; dispersion interaction, EDISP; and contribution from triples correction, E(T). Only interactions of interest are represented in this table, detailed information on all the interactions can be found in Supporting Information (Tables S8–S9). Electronic preparation energies resulting from intrafragment changes in electron density and deformation energies due to geometrical differences of fragments in interacting structure compared to their separated equilibrium geometries that are required to derive the dissociation energies corresponding to the analyzed interactions have not been included in the analysis.</p><!><p>Two novel cocrystals of trans-[MI2(CNXyl)2]·I2 (where M = Pd or Pt) representing noncovalently linked metallopolymeric structures were synthesized and characterized. Analysis of crystal structure showed that trans-[MI2(CNXyl)2] units are interlinked via an uncommon I–I···(I–M) bifurcated contact with the I2 molecule. Bifurcated contact, in turn, can be subdivided into a I···I halogen bond and a I···M metal-involving interaction. To reveal the nature of the contact, it was studied with various computational methods such as NCI-plot, QTAIM and LED analyses, and ED/ESP minima comparisons. It was shown that the I···I halogen bond is the strongest NCI stabilizing the system, supported by a weaker I···M metal-involving interaction. ED/ESP minima comparisons showed the nonpolarity of I···M contact in the [PdI2(CNXyl)2]·I2 cocrystal; therefore, this interaction was suggested to be called quasimetallophilic. In the case of the [PtI2(CNXyl)2]·I2 cocrystal, similar studies showed the weakly nucleophilic nature of Pt center, which makes the I···Pt interaction polar and is best described as metal-involving halogen bonding. However, the differences between the I···Pd and I···Pt interactions are not crucial for directed crystal engineering, and the Pd/Pt isostructural exchange can be further used in the design of similar Pd- and Pt-containing cocrystals.</p><!><p>All the studied structures were optimized and analyzed using DFT theory. To achieve a good compromise between accuracy and computational demand for calculating systems containing NCIs M06-L functional108 combined with triple-ζ def2-TZVP106 basis sets was chosen as the calculation method. To further reduce computational time resolution of identity approximation109 together with def2-TZV density fitting basis sets106 was employed in the calculations. DFT calculations were carried out with Gaussian16 (revision C.01) program package.110 Complexes 1 and 2 and I2 were subjected to full energy minimization. Models for solid-state clusters (1)4·I2 and (2)4·I2 were directly cut from the corresponding experimental crystal structures. Bonding analyses of NCIs in model structures (1)4·I2 and (2)4·I2 were carried out on both optimized (OPT) and crystal structure derived SP structures (where only positions of H-atoms were optimized). SP calculations (M06-L/def2-TZVP) were also carried out for two model clusters, (cis-[PdCl2(CNPh)2])2 and (cis-[PdCl2(CNPh)2])2, based on the experimental X-ray data from the structures COYBOI01 and CPICPT12,99 respectively. The strength and topology of the interactions were studied with the NCI-plot program68 implemented in Critic2 software,111 and 2D and 3D visualizations were carried out in Gnuplot112 and VMD programs113 respectively. ESP surfaces of 1, 2, and I2 molecules were calculated and visualized using AIMALL software114 at 0.001 a.u. surfaces. ELF projections and QTAIM analyses were carried out in Multiwfn 3.7.115 DLPNO–CCSD(T)103−105 wave functions for the LED analyses,71 and the analyses themselves were calculated with ORCA 4.2 program116 using def2-TZVPP106 orbital and def2-TZVPP/C107 and def2/JK117 auxiliary basis sets.</p><!><p>All chemicals and solvents such as CHCl3 (VWR BDH Chemicals), CH2Cl2 (VWR BDH Chemicals), acetone (Fisher Scientific), KI (≥99.0%, Fisher Scientific), I2 (Mallinckrodt), 2,6-dimethylphenyl isocyanide (further CNXyl, ≥98.0 GC%, Aldrich), and [PdCl2(CH3CN)2] (99%, Aldrich) were used without additional purification. [PtI2COD] was synthesized according to the procedure reported by Rigamonti et al.61 The crystal data and details of data processing for the obtained cocrystals are summarized in the Supporting Information ("Single Crystal X-ray Diffraction data analysis (SCXRD)" section).</p><p>Caution! CNXyl is hazardous to health and should be handled with care.</p><!><p>Synthesis was adapted from a procedure presented by Crociani et al.58 Solid CNXyl (26.2 mg, 0.2 mmol) was added to the suspension of [PdCl2(CH3CN)2] (25.9 mg, 0.1 mmol) in 5 mL of CHCl3. The reaction mixture was refluxed with stirring for 3 h and then cooled to room temperature (RT), and the solvent was evaporated at a rotary evaporator to give cis-[PdCl2(CNXyl)2] as a white solid. Then solid KI (166 mg, 1 mmol) was added to cis-[PdCl2(CNXyl)2] (43.7 mg, 0.10 mmol), and acetone (20 mL) was added to the resulting mixture. The resultant yellow suspension was stirred at RT for 2 days. The solvent was then fully evaporated on a rotary evaporator at 50 °C, and the orange product was suspended in H2O. The product was extracted with CH2Cl2. The organic fraction was subjected to full solvent evaporation on a rotary evaporator, and the resulted orange solid was dissolved in CHCl3. Some white insoluble material was filtered off from solution; the filtrate was left for recrystallization at RT in darkness (from CHCl3). The yield of orange crystalline product was 55.6 mg (0.09 mmol, 89%). Elemental analysis (EA) CHN mode: Found: C 35.72; H 3.22; N 4.44. Calcd: C 34.73; H 2.91; N 4.50. 1H NMR (300 MHz, CDCl3, δ ppm): 2.55 (s, 12H), 7.09–7.14 (m, 4 H), 7.21–7.28 (m, 2H).</p><!><p>trans-[PdI2(CNXyl)2] (24.9 mg, 0.04 mmol) and I2 (15.2 mg, 0.06 mmol) were dissolved in CH2Cl2/CHCl3 (50:50 mixture, 8 mL). The solution was stirred at 50 °C (to dissolve iodine fully) until the mixture became homogeneous and then left for crystallization in dark at RT. The phase purity of the bulk material was confirmed by powder X-ray diffraction (PXRD, see the Supporting Information).</p><!><p>Synthesis was adapted from a procedure presented by Kaharu et al.59 [PtI2COD] (83 mg, 0.15 mmol) was added to a 5 mL of CH2Cl2 solution of a CNXyl (39.4 mg, 0.3 mmol), and the mixture was stirred for 3 days at RT in darkness. The solvent was evaporated, and obtained solid was crystallized from CH2Cl2. TLC (silica gel 60 plate + CHCl3) revealed byproducts. The product was purified by column chromatography (silica gel 60 + CHCl3) and recrystallized from CHCl3. The yield of yellow crystalline product was 99.8 mg (0.14 mmol, 93%). EA CHN mode: Found: C 31.83; H 2.81; N 4.11. Calcd: C 30.40; H 2.55; N 3.94. 1H NMR (300 MHz, CDCl3, δ ppm): 2.58 (s, 12H), 7.14–7.18 (m, 4 H), 7.26–7.32 (m, 2H).</p><!><p>trans-[PtI2(CNXyl)2] (21.3 mg, 0.03 mmol) and I2 (15.2 mg, 0.06 mmol) were dissolved in CHCl3, and the resulting dark brown mixture was left in an aluminum foil covered vial for slow evaporation at ambient conditions to give dark brown crystals of the desired product. The phase purity of the bulk material was confirmed by PXRD analysis (see Supporting Information).</p><!><p>Single-crystal X-ray diffraction data analysis (experimental procedures and crystallographic details); PXRD of 1·I2 and 2·I2 cocrystals (experimental procedures and detailed results of PXRD analysis); summary of computational studies on NCIs in 1·I2 and 2·I2 cocrystals (general computational details, ESP, QTAIM, LED, and NCI-plot analyses, ED/ESP minima criterion for I···I interactions); view of (cis-[MCl2(CNPh)2])2 (M = Pd or Pt) dimeric clusters and ELF projections for them (PDF)</p><p>ic1c01591_si_001.pdf</p><!><p>CCDC 2054859–2054862 contain the supplementary crystallographic data for this paper. These data can be obtained free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by emailing data_request@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44 1223 336033.</p><!><p>M.B. carried out synthesis, crystallizations, part of the SCXRD studies, part of the wave function calculations, NCI-plot analysis, ESP visualizations, and manuscript preparation. D.M.I. carried out a combination of ELF and QTAIM analyses as well as the ED/ESP minima analysis. Both M.B. and D.M.I. contributed to data interpretation and analysis. J.M.R. carried out calculations of wave functions, QTAIM, and LED analysis. K.-N.T. carried out part of the SCXRD studies. M.L. carried out PXRD studies of the bulk materials. M.A.K. contributed to the manuscript preparation and literature search. M.H. guided the research and experimental design. The manuscript was written through the contributions of all authors. All authors have approved the final version of the manuscript.</p><!><p>This work was supported by the Academy of Finland (project no. 295581). Part of the theoretical investigations (ELF and QTAIM analysis, ED/ESP minima analysis) were supported by the Russian Science Foundation (project no. 19–73–10016 for D.M.I.).</p><p>The authors declare no competing financial interest.</p>

PubMed Open Access

Designed to Dissolve: Suppression of Colloidal Aggregation of Cu(I)-Selective Fluorescent Probes in Aqueous Buffer and In-Gel Detection of a Metallochaperone

Due to the lipophilicity of the metal-ion receptor, previously reported Cu(I)-selective fluorescent probes form colloidal aggregates as revealed by dynamic light scattering. To address this problem, we have developed a hydrophilic triarylpyrazoline-based fluorescent probe, CTAP-2, that dissolves directly in water and shows a rapid, reversible, and highly selective 65-fold fluorescence turn-on response to Cu(I) in aqueous solution. CTAP-2 proved to be sufficiently sensitive for direct in-gel detection of Cu(I) bound to the metallochaperone Atox1, thus demonstrating the potential for cation selective fluorescent probes to serve as tools in metalloproteomics for identifying proteins with readily accessible metal-binding sites.

designed_to_dissolve:_suppression_of_colloidal_aggregation_of_cu(i)-selective_fluorescent_probes_in_

<p>Cation selective fluorescent probes have become increasingly important analytical tools for the detection of metal ions in environmental samples, for visualizing metal ions in cells and tissues,1 or as reagents for measuring metal affinities of biomolecules.2 Such probes are typically comprised of a chelator for selective recognition of the target ion and a fluorophore to optically transduce binding of the analyte.3 For passive diffusion across cellular membranes, the probes must be sufficiently lipophilic; however, in aqueous buffer the associated hydrophobic character might also lead to aggregate formation, which in turn may dramatically alter the photophysical properties, notably the brightness and emission wavelength of the probes. Because the majority of fluorescent probes used in biological research are lipophilic and often poorly water-soluble, incubation buffers are typically prepared starting from a stock solution in an organic solvent such as DMSO, which is then diluted into the buffer to a final probe concentration in the low micromolar range. Although this approach usually yields optically clear solutions, the absence of turbidity does not exclude the formation of a colloid composed of nanoparticles with sizes below the diffraction limit. In the course of our efforts in developing Cu(I)-selective fluorescent probes, we realized that the metal ion recognition site, typically composed of thioether donors,4–8 further increases the lipophilicity and thus the propensity towards aggregation. Accordingly, we found that the Cu(I)-responsive probe 1 (Chart 1), which we previously characterized in methanol,7 forms a clear homogenous solution when diluted from a 1 mM DMSO stock into aqueous buffer; however, dynamic light scattering measurements revealed the presence of colloidal aggregates with an average hydrodynamic radius of around 100 nm (Table 1, Figure 1). This observation prompted us to also test other Cu(I)-responsive probes previously characterized in aqueous buffer for their ability to form colloidal aggregates. While the BODIPY-based copper sensors CS19 and CS310 both yielded optically clear solutions in aqueous buffer at a concentration of 5μM, the autocorrelation curves obtained from dynamic light scattering measurements indicated the formation of nanoparticles with average sizes of 49 and 67 nm, respectively (Figure 1, Table 1). Surprisingly, our first generation probe CTAP-1,5 which is functionalized with a charged carboxylate group, also showed formation of colloidal aggregates under the same conditions (Table 1).</p><p>It has been recently recognized that many classes of bioactive organic molecules spontaneously form colloidal aggregates at micromolar concentrations, a problem noted to affect the reliability of high-throughput screening in early drug discovery.11 Similar to drugs, fluorescent probes must reach their cellular targets by crossing lipid bilayers, and therefore tend to be considerably lipophilic. The formation of colloidal aggregates might not necessarily jeopardize their utility in biological studies; however, the photophysical properties of the colloid may be dramatically different compared to the monomeric form. Therefore, great caution is advised when using fluorescent probes in a mixed polarity environment as found in cells, which is likely to shift the equilibrium between the aggregated and monomeric forms.</p><p>To address the problem of colloid formation, we designed a series of new water-soluble Cu(I)-selective probes 2a–c in which the thiocrown receptor was modified with four hydroxymethyl groups and combined with triarylpyrazoline fluorophores. To balance the hydrophilicity between receptor and fluorescent reporter and to further increase the overall water solubility, the pyrazoline moiety was functionalized with a sulfonate group, an established approach to solubilize organic fluorophores.12 Binding of the analyte is translated into a fluorescence increase through a photoinduced electron transfer (PET) switching mechanism as shown for a range of other pyrazoline-based fluorescent probes.13,14 The variable number of fluoro substituents served to adjust the PET driving force and thus to optimize the fluorescence enhancement factor upon saturation with Cu(I) as previously demonstrated.6,15</p><p>For the synthesis of 2a–c, we devised a modular approach in which the same cation receptor can be readily combined with various functionalized pyrazolines (Scheme 1). To this end, we sought a protective group strategy that would allow for masking of both the hydroxyl groups, which needed to be introduced early in the synthesis, and the sulfonate moiety, which could not be incorporated in the final step due to the presence of sensitive functional groups. Since the four hydroxyl groups could be efficiently protected pairwise as acetonides, we contrived an acetonide-based protective group for sulfonic acids. This was accomplished by combining a neopentyl sulfonate ester, which is sterically protected against external nucleophiles, with an acetonide moiety, which, upon hydrolysis, would provide hydroxyl groups that can intramolecularly displace the sulfonate under basic conditions. Compared to previously described neopentyl sulfonate ester protective groups,17 this acetonide derivative can be synthesized in fewer steps and is sufficiently robust to allow preparation of aryl-hydrazines from the corresponding fluorinated sulfonate esters by nucleophilic aromatic substitution with hydrazine (see supporting information). Following this approach, we successfully prepared the protected pyrazolines 12a–c, which were then converted to the desired products 2a–c.</p><p>Probes 2a–c readily dissolved in aqueous buffer and responded with strong fluorescence enhancements upon saturation with Cu(I), supplied either from a 2.5 mM stock solution of [Cu(I)(CH3CN)4]PF6 in CH3CN or by in situ reduction of CuSO4 with ascorbate (Table 2). The absorption and emission bands shifted to shorter wavelengths with increasing electron withdrawing ability of the 1-aryl ring (Figure S1), corresponding to a stepwise increase of the excited state energies ΔE00 from 2.79 to 3.06 eV (Table 2). Consistent with a PET quenching mechanism, the quantum yields in the presence and absence of Cu(I) decreased with increasing ΔE00, and the corresponding fluorescence enhancement fe decreased from 65 to 9. Because the latter is expected to follow a bell-shaped distribution,8 it was unclear whether 2a already embodied the maximum obtainable contrast or whether a derivative with lower ΔE00 might yield an even better performance. To address this question, we synthesized probe 13 in which the electron withdrawing sulfonate group is electronically separated from the fluorophore π-system through a methylene group (Supporting Information). As evident from Table 2, the lower excited state energy of 13 (2.70 eV) produced an increased quantum yield in neutral buffer but not in the presence of Cu(I), thus resulting in a lower contrast than 2a. Examination of the quantum yields under acidic conditions, where the arylamine is protonated and rendered inert towards oxidation, suggests that a quenching pathway other than acceptor-excited PET is responsible for the lower than expected quantum yield of 13-Cu(I); anomalously low quantum yields in polar solvents have been previously reported for 1,3,5-triaryl pyrazolines bearing electron rich 1-aryl rings.14,18</p><p>Given the superior contrast of 2a over the other probes, the remaining characterization focused exclusively on this compound, which we also named CTAP-2 as an identifier for these and other future studies. As expected for a high-affinity ligand, a fluorescence titration of CTAP-2 with Cu(I) showed a linear emission increase with sharp saturation at 1 molar equivalent (Figure 2A).</p><p>The UV-vis absorbance of CTAP-2 scaled linearly with concentrations from 0 to 5 μM, yielding a molar absorptivity of 2.9.104 M−1cm−1 at 396 nm (Figure S3, inset). Similarly, the absorption and emission intensity of Cu(I)-bound CTAP-2 increased linearly in the same concentration range (Figure S2). At values above 10μM CTAP-2 the absorbance vs concentration plot deviated slightly from linearity (Figure S3), indicating the presence of weak self-association. Non-linear least squares fitting of the experimental data assuming a simple dimerization equilibrium yielded an equilibrium constant of logK = 3.98 ± 0.06 and a dimer molar absorptivity of 4.3.104 M−1cm−1. Based on these data we estimated that at concentration of 5 μM approximately 4% of the probe is present as dimer. While not negligible, dimer formation was experimentally only evident at concentrations that substantially exceeded the working concentration typically used for fluorescence measurements. Similar dimerization constants in aqueous solution have been reported for xanthene dyes such as the cationic Rhodamine 6G.19 Most importantly, dynamic light scattering experiments with a 5 μM solution of CTAP-2 in MOPS buffer (10 mM, pH 7.2) gave count rates that were no higher than the background of the buffer alone, thus confirming the absence of colloidal aggregates.</p><p>To determine the Cu(I) affinity we used the formal potentials of the free and bound CuII/I couples and the Cu(II) affinity of CTAP-2 under mildly acidic conditions.20 In the presence of Cu(II), CTAP-2 showed a new quasi-reversible 1-electron process with a half-wave potential of 0.226 V vs. Fc+/0, corresponding to 0.626 V vs. SHE21 (10 mM PIPBS, 0.1 M KClO4, Figure S6). Under the same conditions, UV-vis titrations revealed a logKCu(II) = 2.97 ± 0.07 (Figure S5). Based on these data, we obtained logKCu(I) = 11.4 ± 0.1 or Kd = 4 ± 1 pM for CTAP-2 at pH 5.0 (I = 0.1 M), which compares well with the affinity of structurally related probes.5,9 Given the low pKa of 3.97 ± 0.03 for protonation of the thiazacrown receptor (Figure S4), the apparent Cu(I) affinity at pH 7.2 remains unchanged within experimental error. Furthermore, the fluorescence response of CTAP-2 proved to be very selective towards Cu(I) and unaffected by other biologically relevant ions (Figure 2B).</p><p>Encouraged by the high fluorescence contrast and selectivity towards Cu(I), we decided to explore the utility of CTAP-2 as a reagent for the in-gel detection of proteins containing a readily accessible Cu(I)-binding site as present in copper metallochaperones such as Atox1.22 For this purpose we purified recombinant hAtox1 and subjected the protein to native gel electrophoresis.</p><p>As illustrated in Figure 3, incubation of the gel with CTAP-2 revealed the presence of Atox1 in a copper-dependent manner. While untreated Atox1 gave rise to a fluorescence signal (lane 1), preincubation with KCN to remove Cu(I) from Atox1 abolished the response (lane 2). Furthermore, lanes 3 and 4 demonstrate reversible copper binding as expected for a metallochaperone. In contrast, no staining was observed for carbonic anhydrase (CA), a ZnII-containing enzyme, or superoxide dismutase (SOD1), an enzyme in which the copper-site is sterically inaccessible. Post-staining with coomassie blue revealed the presence of the proteins in each lane, and confirmed the removal of Cu(I) from hAtox1 based on the different mobilities of the apo- and holo-forms.23 Given the high Cu(I) affinity of Atox1 (logK = 17.4),24 it is probable that CTAP-2 associates with the protein in a Cu(I)-dependent manner without actually removing the metal ion from the binding site. Such ternary complexes might also be formed in a biological environment, as recently suggested in case of the zinc-responsive fluorescent probes FluoZin-3 and TSQ.25 Despite its net anionic charge at neutral pH, CTAP-2 proved to be cell permeant and produced in live NIH 3T3 cells a perinuclear staining pattern (Figure S7), reminiscent of the subcellular copper distribution previously reported;5,26 however, in view of above findings, the interpretation of the observed cellular staining is nontrivial and will require further detailed studies.</p><p>In conclusion, we have developed a fluorescent probe CTAP-2 that selectively responds to Cu(I) in aqueous buffer with a 65-fold fluorescence enhancement. The response of CTAP-2 is rapid and reversible, making it suitable as an indicator for titrations with Cu(I) or for monitoring equilibrium concentrations of Cu(I). While previously described fluorescent probes for Cu(I) have not been reported to dissolve directly in water,4,5,9,10,27 salts of CTAP-2 quickly dissolve in pure water up to millimolar concentrations, circumventing the possibility of colloidal aggregate formation that exists when organic stock solutions of poorly soluble dyes are diluted into aqueous buffer. Because the majority of fluorescent probes utilized in biology are considerably lipophilic, the formation of colloids is likely not limited to the probes investigated here but a rather widespread phenomenon that deserves particular attention when interpreting fluorescence microscopy data. Adding to the previously described applications of Cu(I)-responsive fluorescent probes, CTAP-2 was able to detect copper bound to a metallochaperone. As only proteins with accessible metal sites can give rise to a fluorescence response, the detection of metalloproteins with fluorescent indicators such as CTAP-2 nicely complements the currently available techniques for in-gel metal profiling, namely laser ablation-inductively coupled plasma-mass spectrometry (LA-ICP-MS)28 and synchrotron-based x-ray fluorescence mapping,29 both of which measure the total metal content regardless of its accessibility, thus further expanding the metalloproteomics toolbox.30</p>

PubMed Author Manuscript

Moisture-Responsive Graphene Actuators Prepared by Two-Beam Laser Interference of Graphene Oxide Paper

Here, we reported an ingenious fabrication of moisture responsive graphene-based actuator via unilateral two-beam laser interference (TBLI) treatment of graphene oxide (GO) papers. TBLI technique has been recognized as a representative photoreduction and patterning strategy for hierarchical structuring of GO. The GO paper can be reduced and cut into grating-like periodic reduced graphene oxide (RGO) microstructures due to laser ablation effect. However, the lower light transmittance of the thick GO paper and the corresponding thermal relaxation phenomenon make it impossible to trigger complete reduction, leading to the formation of the anisotropic GO/reduced GO (RGO) bilayer structure. Interestingly, the RGO side that feature lower OCGs and higher roughness shows strong water adsorption due to the formation of micronanostructures. Due to the different water adsorption capacities of the two sides, a flower moisture-responsive actuator has been fabricated, which exhibits “opening” and “closing” behavior under different humidity conditions.

moisture-responsive_graphene_actuators_prepared_by_two-beam_laser_interference_of_graphene_oxide_pap

Introduction<!>Preparation of GO Paper<!>TBLI Reduction of GO Paper<!>Characterization<!>Moisture Response of GO Paper<!><!>Moisture Response of GO Paper<!><!>Moisture Response of GO Paper<!><!>Moisture Response of GO Paper<!><!>Moisture Response of GO Paper<!>Conclusions<!>Data Availability<!>Author Contributions<!>Conflict of Interest Statement<!><!>Supplementary Material<!>

<p>The traditional actuators, such as energy conversion devices, sensors, robotics, and micro-electromechanical systems, are driving devices which have been extensively studied because of their irreplaceable application in basic scientific research and engineering technologies (Azizi and Khorasani, 2011; Petit et al., 2012; Tottori et al., 2012). However, with the emergence and development of technology industries, next-generation intelligent products without additional connections, and energy-supply equipment, such as wearable electro-skins (Ho, 2002), artificial muscles (Han et al., 2019), microrobots in tissue engineering (Kim et al., 2012), and lab-on-a-chip (LoC) systems (Kokalj et al., 2014), are in increasing demand. For this current situation and future prospects, stimuli-responsive materials (SRMs) that can rapidly and reversibly convert their structure/morphology/volume variation into mechanical deformation under certain environmental stimulation (Deng et al., 2018; Han B. et al., 2018) are ideal choices. External stimuli including light, temperature, electromagnetism, solvents, and humidity can be utilized for realizing intelligent actuators (Tian et al., 2011; Kim et al., 2018; Li and Yin, 2019; Zhu et al., 2019). As a typical example, an anisotropic bilayer structure can express the mechanical work generated in different material layers by external stimulus more efficiently and intuitively (Ma et al., 2018; Ohuchi et al., 2018). Many research groups have made successful progress in the study of bilayer actuators. For instance, Zhang et al. investigated the controlling helicity angle, chirality, diameter, and pitch of multilayered SiGe/Si/Cr nanobelts with widths <400 nm (Zhang et al., 2006). Kelby et al. presented an Au-polymer bilayer brush example for the controlled folding of 3D micro-structures (Kelby et al., 2011). However, as moving deformation occurs step by step in a motion device, poor adhesion between the different material layers directly affects the mechanical strength of the device during movement (Xin et al., 2016; Rodrigo et al., 2017; Padhi et al., 2018). In this regard, it is already a challenge to exploit new monomer material to design and fabricate smart bilayer-structure actuator through the precise control of the unilateral morphology and composition, which could realize the anisotropic property.</p><p>Graphene oxides (GOs), as derivatives of graphene materials (Zhu et al., 2010), possess immense potential for a broad range of graphene-device applications owning to several advantages, such as solubility, batch preparation, ease of fabrication of large films, eco-friendliness, and stable physicochemical properties (Becerril et al., 2008; Eda et al., 2008; Kudin et al., 2008). As the surface contains numerous hydrophilic oxygen-containing groups (OCGs), desiccative GO films are extremely sensitive to ambient humidity (Han D. et al., 2018). In spite of surface defects of GO, suitable chemical, thermal and photoreduction method have been successfully developed to realize some sort of OCGs removal, which leads to the formation of RGO structure with different components from GO (Stankovich et al., 2007; Li et al., 2011). In view of this principle, graphene smart actuators based on GO material have been successfully developed. Qu et al. successfully prepared the fiber-type smart robots through controllable laser reduction of GO fibers, which have been realized a series of one-dimensional mechanical deformations, such as folding, bending, and S-shaping (Cheng et al., 2013). Zhang et al. presented a facile preparation of moisture responsive graphene actuators by unilateral UV irradiation of graphene oxide (GO) papers, which have been mimicked the cilia of respiratory tract and tendril climber plant to transport objects (Han et al., 2015a). However, although the partly photoreduction treatment endow stimuli-responsive properties to the GO/RGO structure, the above-mentioned works with respect to regulation of the chemical compositions of the materials weaken the structure/morphology influence. It is blank to fabricate such graphene bilayer structure. The unique micronanostructures of material surfaces have attracted interest in a wide range of scientific fields, such as electronic devices, energy storage devices and biomimetic surfaces (Bi et al., 2013; Kumar et al., 2014; Wang et al., 2018), which will hold great potential in the field of smart actuators.</p><p>Herein, inspired by biomimetic structural surfaces, a micronano-patterning photoreduction method has been applied to prepare moisture responsive GO/RGO bilayer structure through the laser holography technique treatment of GO paper. The chemical properties of the GO surface can be considerably altered after two-beam laser interference (TBLI) treatment by the photothermal effects, causing the removal of most of the OCGs on and between the GO sheets. Simultaneously, abundant RGO micronanostructures including periodic micro-gratings and nanoscale sheets are formed, increasing the richness of the surface morphology, this changes the wettability of the resultant graphene films, causing hydrophobicity. Laser processing technology is an effective means to realize controlling of surface wettability by fabricating micro/nano structures on the surface of materials (Yin et al., 2017b, 2018; Duan et al., 2018), which own great potential in application of superhydrophilicity (Yin et al., 2017a; Yang et al., 2018). Considering the limited light transmittance of GO paper and thermal relaxation could not realize the totally reduction of GO paper, a gradient photoreduction process along the sectional direction of GO paper occurs spontaneously. Consequently, the anisotropic GO/RGO bilayer structure has been fabricated during the self-acting photoreduction. Taking advantage of the different water-absorbing capacities of the two sides of bilayer structure, the resultant GO/RGO film could be acted as a moisture-responsive actuator. We further demonstrated the graphene actuator to mimic opening and closing behavior of a flower.</p><!><p>The graphene oxide material we used here was prepared by Hummer's method. After high strong oxidant (such as H2SO4, KMnO4 and so on) treatment, the graphite powder was oxidized into single-layer and few-layer graphene nanosheets containing a large number of hydroxyl and carboxyl groups. Due to the presence of a mass of OCGs, the obtained GO material is easily soluble in water and becomes a water-soluble suspension. In order to remove residual sulfate and chloride ions, GO aqueous solution should be washed several times with deionized water. The synthesized GO solution was collected by centrifugation and then dispersed in distilled water at a concentration of 3 mg/mL under ultrasonic treatment. Subsequently, graphene oxide paper has been fabricated by ultrafiltration of the prepared GO aqueous solution with the help of the filtration membrane with pore diameter of 0.22 μm. And drying the GO/membrane in the ambient air, and then peeling off the GO paper for next treatment.</p><!><p>A standard single-mode Nd:YAG laser (Spectra-Physics, 355 nm, 10 Hz, and 10 ns pulse duration) was used as the light source for the GO reduction process. The laser beam with the diameter size of 8 mm was split into two beams, which own the same optical path lengths to the sample surface. The RGO with multilevel structures were fabricated by exposing the GO papers to the interfered laser region. Taking advantage of laser holography technique, the periodicity of the grating could be precisely controlled by changing the angle of two laser beams. By adjusting the appropriate laser power (~0.4 mW), exposure time (30 s) and laser light path, here we fabricated the graphene grating with 2 μm. The laser power is 0.4 mw, and the exposure time is 30 s, which is the relatively suitable laser processing parameter. When the laser power is too large and the exposure time is too long, the surface morphology of the processed sample will be ablated seriously. On the contrary, when the laser power is too low and the exposure time is too short, it is difficult to form a complete continuous periodic micro-nano grating structure (Supplementary Figure S3). Taking advantage of laser holography technique, the periodicity of the grating could be precisely controlled by changing the angle of two laser beams, as shown in the following equation:</p><p>where Λ is the period, λF is the laser wavelength and θ is the angle between two beams.</p><!><p>SEM images were obtained using a field-emission scanning electron microscope (JSM-7500, JEOL, Japan), which referred to as FE-SEM. X-ray photoelectron spectroscopy (XPS) was performed using an ESCALAB 250 spectrometer for chemical bond analysis. Raman spectroscopy was recorded on a Jobin-Yvon T64000 Raman spectrometer equipped with a liquid-nitrogen-cooled argon ion laser at 514.5 nm (Spectra-Physics Stabilite 2017) as the excitation source; the laser power used was ~10 mW with an average spot size of 1 μm in diameter. The contact angle (CA) measurements were performed using the Contact Angle System OCA 20 (DataPhysics Instruments GmbH, Germany) at ambient temperature. The CA was measured using a 4 μL water droplet. The humidity response cycle test is carried out in an environment with 100% humidity. The controlled humidity environments were achieved using saturated aqueous solutions of CH3COOK, MgCl2, K2CO3, NaBr, NaCl, and KCl in a closed glass vessel, which yielded ~23, 33, 44, 57, 75, and 86% RH, respectively. All of the measurements were conducted in air at room temperature (25°C).</p><!><p>As an efficient tool for large-area micronano-patterning without any shadow masks and chemicals, laser holography technique is undoubtedly a preferred choice for fabrication of micronanostructures (Wang et al., 2011; Sun et al., 2018). As a typical example, TBLI has already proved its value in the photoreduction of GO and fabrication of graphene micronanostructures synchronously in our previous works (Jiang et al., 2014). Hence, in this study, we applied the TBLI as the irradiation source to prepare RGO structure and fabricate the GO/RGO bilayer structure, for realizing humidity response. Figure 1 displays a schematic illustration of the design philosophy and manufacturing flow of moisture-responsive graphene actuator. We use the GO papers as experimental materials, which have been fabricated by vacuum filtration of GO aqueous through a filter membrane and naturally dried in air. The yellow-brown dried GO papers are showed in Figure 1A. Subsequently, after the GO paper was exposed to the laser interference region, the RGO side turned black and it may result from the removing of OCGs (Figure 1B). During the laser machining process, the periodic micro-gratings with nanoscale roughness have been produced along the direction of laser intensity distribution. In the laser interference region, the laser intensity distribution is constant along the y-axis, and sinusoidal along x-axis. The highest is four times of each laser beam, and the lowest is zero. At the high laser intensity region, the OCGs were thoroughly removed; whereas at the low intensity region, the GO survived, and was reduced partially. However, due to the limited light transmission of GO paper and delayed thermal transmission, the reduction reaction on the GO paper surface was incomplete. In this way, anisotropic GO/RGO bilayer structure has been realized along the sectional direction of GO paper, which showed in Figure 1C. For plane RGO surface, the hydrophilic GO surface has a stronger water absorbing capacity compared to the RGO because of the formation of hydrogen bonds on the GO sheets in the presence of moisture, as reported previously (Han et al., 2015b). But TBLI photoreduction method induced increased interlayer space and the new products of graphene micronanostructures on the RGO surface, such morphological changes would provide more space for water molecules to interact with RGO sheets through Van der Waals force. On account of this situation, when the GO/RGO bilayer film was exposed to humid environment, it would bend toward the GO side due to the asymmetric adsorption, as shown in Figure 1D.</p><!><p>The schematic illustration of the fabrication process of a moisture responsive GO/RGO bilayer film. (A) GO film. (B) TBLI treatment of GO film. (C) GO/RGO film. (D) Curving GO/RGO film.</p><!><p>As observed in the optical micrographs of bilayer structure (Figures 2A,B), the resultant RGO paper presents black in comparison with the dark brown of GO, proving the effective photoreduction of GO. However, the different color behaviors of two sides of the sample reveal the gradient change of the photoreduction process. The maximum diameter of light spot is about 1 cm, which should be taken 30 s for the fabrication of each spot with the grating structures. Large area of RGO structure surface can be prepared by splicing each spot together. The area has no upper limit. In order to investigate the TBLI photoreduction process, scanning electron microscopy (SEM) images of the GO and RGO sides were obtained. Figure 2C shows the SEM image of relatively smooth GO surface with rich wrinkles owned by unique graphene materials. Due to the laser ablation, ordered grating structures with a period of 2 μm were fabricated. Despite these periodic micro-gratings, the folded structures of graphene could be observed on the RGO side. To characterize the surface wettability of the two sides of bilayer structure, static water-droplet contact angle (CA) measurements have been applied, which could be observed in the insets of Figures 2C,D. The CA value of the GO side was ~45°, whereas the RGO side gives a significantly increased CA of ~128°, which is a significant improvement compared to the UV and other reduction methods. Generally, the obvious hydrophobicity on RGO surface could be mainly put down to the drastic removal of OCGs and preparation of graphene micronanostructures during photoreduction. It is to be noted that the CA of the RGO surface would show a certain descending over time because of the permeation of water droplets on the surface of GO paper.</p><!><p>(A,B) Photographs of the resultant GO/RGO bilayer structure viewed from the front side (A) and the reverse side (B), respectively. (C–G) SEM images of GO/RGO bilayer structure. (C,D) SEM images of GO side and RGO side. The insets show water droplet contact angles of both sides. (E) SEM image of the sectional view of GO/RGO bilayer structure. (F,G) High-resolution SEM images of RGO side.</p><!><p>Next, a section-view SEM image of the TBLI treated GO paper has been presented. In Figure 2E, the fluffy structures with larger layer-gaps (thickness is ~12.5 μm) in the up-side and the relatively dense structures with thinner thickness of ~5.8 μm in the bottom-side could be observed. Such formation of asymmetric structure is ascribed to the result of gas release, such as CO2, CO, H2O in the reduction layer, which have been produced by carbon components in the photothermal reaction. In the case of this situation, the thicknesses of the two layers structure show certain dependence on sectional dimension of GO paper and laser intensity of the TBLI treatment. Thus, we could observe a layer of micronanostructures on the top of up-side. To get further insight into the detailed morphology of micro-gratings, we measured the magnified SEM images of RGO surface (Figures 2F,G). Laser cutting effect not only prompts the formation of periodic micro-scale grating-like structures but also in the meantime induces the graphene nano-layer structures. The special attention should be paid to the laser parameters and optical path design. They will affect the structural morphology of the RGO surface. Adjusting the intensity and exposure time with high level, the GO paper could realize violent reduction or even combustion.</p><p>To evaluate the chemical composition change, we measured the X-ray photoelectron spectroscopy (XPS) of the two sides of the RGO paper irradiated by TBLI light. As shown in Figure 3A, the carbon and oxygen content comparisons between GO and RGO side are surveyed. The C/O ratio of the RGO side shows an increased value of 35.76 and the GO is about 2.52. This obvious change reveals the photoreduction gradient of GO/RGO bilayer structure. Besides, the C1s spectra of GO and RGO sides could be split into three peaks that correspond to C-C (284.6 eV, non-oxygen ring), C-O (286.8 eV, hydroxyl and epoxy carbon), and C = O (288.4 eV, carbonyl), as shown in Figure 3B. Obviously, OCGs are very rich on the GO surface, the C-O and C = O peaks of GO side show an higher value than that of RGO side irradiated after TBLI laser. And the C content increased from 71.58 to 97.28%, indicating the removal of OCGs. X-ray diffraction (XRD) of GO and RGO side have also been tested (Supplementary Figure S1). The XRD pattern of the GO sheets shows a typical diffraction peaks at 2θ = 11.55°, which indicates an ordered layered structure of GO. However, after TBLI treatment, the diffraction peak disappeared, which suggests that the RGO layers become disorder after the removal of OCGs.</p><!><p>(A) Survey XPS spectra and (B) C1s spectra of GO/RGO film measured on the GO side and RGO side. (C) Raman spectra of GO/RGO structure measured on the GO side and RGO side. (D) Optical image of RGO structure. (E) Raman mapping of (D), corresponding to wave number from 1,300–1,500, 1,500–1,800, 2,200–2,800 cm−1, respectively.</p><!><p>Furthermore, another characterization method, Raman spectroscopy has been used to test the degree of hybridization of GO and RGO, as shown in Figure 3C. Two characteristic peaks at 1,343 cm−1 (D band) and 1,588 cm−1 (G band) are observed. The G peak is used to characterize the degree of graphitization of carbon materials, and the D peak represents the defects, caused by dislocation, grain boundary, fold, and so on. In general, after the removal of OCGs by appropriate reduction method, the C-C hybridization on the surface would restore on a certain extent, and ID/IG shows an obvious decreased. However, the ID/IG ratio of RGO in our work slight increased from 0.97 (GO) to 0.98, which means that more new defects have been brought during the laser ablation. To further investigate the uniformity of laser interfered region of RGO, we also measure the Raman maps in an area of 5 × 10 μm2, as shown in Figure 3D. The spatial distribution of D, G, and 2D peak values are presented as red, green and blue maps, respectively (Figure 3E). It could be clearly observed that the D, G band intensity distribution almost keep the same as its morphology in optical microscope, which indicates chemical composition of the structures obtained by TBLI treatment was also relatively homogeneous.</p><p>After TBLI treatment, the GO/RGO bilayer structure has been formed, which would curve to the absorbent layer in moisture. Figure 4A shows the mechanism of deformation under humidity control. In general, the GO surface is hydrophilic due to lots of OCGs in the interlayer. Since the presence of OCGs accounts for the formation of hydrogen bonding to adsorb water. However, abundant micronanostructures (micro-gratings, nano-layers) have been made during the process of photoreduction on the RGO surface. Even though RGO layer has fewer hydrophilic groups, unique structures provide more space for water molecules to enter and adsorb. Considering the anisotropic water adsorption between two sides, the layer spacing of RGO side would become larger than GO, leading to GO/RGO ribbon bending to the GO side under high humidity. We further test the dependence of the bending curvature and humidity (RH) of the two layers, as shown in Figure 4B. As the humidity gradually increases, the bending degree of the bilayer structure becomes more obvious. When the humidity is 86%, the GO/RGO bilayer paper owns maximum bending angle and the curvature reaches 0.175. The original bilayer structure also has a certain degree of curvature (−0.047). By contrast, the pure GO paper is hard to bend in moisture. When the humidity is "on" and "off," the bilayer structure exhibits a reversible bending deformation. Figure 4C shows the highly repeatable cycling behavior upon seven cycles. For the GO/RGO bilayer structure, the response and recovery times are ~55 and ~45 s, respectively. Moreover, the GO/RGO bilayer also shows excellent stability during frequent bending–unbending actuation, and the curvature almost keeps a consistent value, indicating the good reproducibility, and precision of the actuation function (Supplementary Figure S2).</p><!><p>(A) Schematic illustration of the bending behavior of the GO/RGO bilayer structure. Water molecules have been demonstrated in blue color in GO/RGO structure for easy to identify. (B) Bending curvature for GO paper and GO/RGO bilayer structures. (C) Reversible moisture response of GO and GO/RGO. (D) The moisture-responsive flower robot made of GO/RGO structure.</p><!><p>The humidity response characteristics of the GO/RGO bilayer structure provide possibilities for the design and fabrication of intelligent graphene actuators. In our work, inspired from opening and closing behaviors of flowers, the unique flower-robot has been demonstrated. We cut the GO/RGO bilayer papers into six pieces to simulate the shape of the petals; and then put them together and glue to a branch directly, as shown in Figure 4D. When the environmental humidity is "on," the flower robot would exhibit the "closing" behavior. The bottom petals of flower robot were observed not moving in Figure 4D because of video recording angle. It took only 10 s for the "flower" to go from full bloom state to final closure. For comparison, we mark the original state of flowers in yellow. When the moisture is replaced by dry air, the "flower" slowly opens again from the closed state. The recovery process without any external energy supports lasts 64 s. The changes in humidity conditions can operate the GO/RGO bilayer structure, which also can be made into other forms of robots, such as claw, crawler and tendril.</p><!><p>In conclusion, TBLI treatment of GO papers has been evolved for successful fabrication of GO/RGO bilayer structure toward preparation of moisture-responsive graphene-based actuators. TBLI treatment was performed to remove the OCGs on the GO sheets and form graphene micronanostructures. Due to the light gradient transmittance and delayed thermal transmission of GO paper, the gradual reduction along its lateral direction occurred, resulting in differences in the reduction degree as well as surface morphology on both sides. In spite of hydrophilicity of GO, the RGO side has been observed stronger absorption ability of water molecules than GO side because of auxetic material porosity induced under laser ablation, which rendered the anisotropic GO/RGO structure impressible to environmental moisture. Interestingly, the graphene micronanostructures formed during the reduction process simultaneously increased the roughness of the RGO surface, leading to hydrophobicity on the RGO side. A stripe GO/RGO bilayer structure has been fabricated to demonstrate the reversible bending behaviors. A curvature of −0.047 to 0.175 was obtained, when the RH was adjusted between 23 and 86%. Inspired by the floral behavior, the fabricated humidity-driven graphene actuators were assembled to mimic the opening and closing behavior of flower. We deem that this laser lithography technology may enable biomimetic design and fabrication of intelligent graphene-devices.</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>H-BJ, YL, and L-QR conceived the idea and designed the experiments. H-BJ, S-YL, and Y-YS made the characterizations. H-BJ and JL performed the detections. H-BJ, D-DH, and YL contributed to data analysis and interpretation. H-BJ and D-DH wrote the paper. All authors discussed the results and commented on 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. This work was supported by National Natural Science Foundation of China (NSFC) under Grant #51705192. The project was funded by the China Postdoctoral Science Foundation, 2017M611325, and the National Postdoctoral Program for Innovative Talents, BX201600064.</p><!><p>The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fchem.2019.00464/full#supplementary-material</p><!><p>Click here for additional data file.</p>

PubMed Open Access

Cerium–quinone redox couples put under scrutiny†

Homoleptic cerous complexes Ce[N(SiMe3)2]3, [Ce{OSi(OtBu)3}3]2 and [Ce{OSiiPr3}3]2 were employed as thermally robust, weakly nucleophilic precursors to assess their reactivity towards 1,4-quinones in non-aqueous solution. The strongly oxidizing quinones 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) or tetrachloro-1,4-benzoquinone (Cl4BQ) readily form hydroquinolato-bridged ceric complexes of the composition [(CeIVL3)2(μ2-O2C6R4)]. Less oxidising quinones like 2,5-di-tert-butyl-1,4-benzoquinone (tBu2BQ) tend to engage in redox equilibria with the ceric hydroquinolato-bridged form being stable only in the solid state. Even less oxidising quinones such as tetramethyl-1,4-benzoquinone (Me4BQ) afford cerous semiquinolates of the type [(CeIIIL2(thf)2)(μ2-O2C6Me4)]2. All complexes were characterised by X-ray diffraction, 1H, 13C{1H} and 29Si NMR spectroscopy, DRIFT spectroscopy, UV-Vis spectroscopy and CV measurements. The species putatively formed during the electrochemical reduction of [CeIV{N(SiMe3)2}3]2(μ2-O2C6H4) could be mimicked by chemical reduction with CoIICp2 yielding [(CeIII{N(SiMe3)2}3)2(μ2-O2C6H4)][CoIIICp2]2.

cerium–quinone_redox_couples_put_under_scrutiny†

Introduction<!>Molecular redox precursors<!>Quinone oxidation of Ce[N(SiMe3)2]3 (1)<!><!>Quinone oxidation of Ce[N(SiMe3)2]3 (1)<!>Quinone oxidation of siloxides [Ce{OSi(OtBu)3}3]2 (2) and [Ce{OSiiPr3}3]2 (3)<!><!>Electrochemical investigation of complexes 4xhq, 5xhq and 6xhq<!>Reduction of silylamide 4hq with cobaltocene<!>Cerium semiquinolates<!>Conclusions<!>Conflicts of interest

<p>Quinones are multifunctional organic molecules exhibiting intriguing redox behaviour.1,2 Of particular note is their importance in biological electron–transfer processes (photosynthesis, respiration)3 and in industrial catalysis (anthraquinone process for hydrogen peroxide production).4 Quinones can engage in one or two electron redox processes involving the formation of either semiquinolates or hydroquinolates.5 Strikingly, the reduction potential of 1,4-benzoquinones (para-benzoquinones) can easily be modified by introducing electron-withdrawing or donating substituents into the benzene ring.5,6 As a consequence, tetrachloro-1,4-benzoquinone (chloranil, Cl4BQ) and even more so 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ) emerged as efficient oxidants in organic synthesis.7 DDQ has been further successfully applied in photoredox catalysis.8 Moreover, anionic η4-1,4-benzoquinone manganese tricarbonyl features a quinoid π-complex, broadly used for the fabrication of supramolecular metal–organometallic coordination networks.9 Relatedly, deprotonated variants of 2,5-dihydroxy-1,4-benzoquinone (DHBQ) were shown to act as rigid ditopic linkers,10e.g., to support the formation of pentagonal dodecahedral Ce2(H2O)18 cages or in permanently porous aluminium frameworks.11 DHBQ was also probed as a bridging redox-active ligand in bimetallic [LnCl2(thf)3]2(μ-bobq) (Ln = Y, Dy; bobq = 2,5-bisoxide-1,4-benzoquinolato) to build single-molecule magnets.12 More recently, the related semiquinolato radical-bridged dimeric complexes [LnCl2(thf)3(μ-Me4sq)2]2 (Ln = Y, Gd) were obtained by oxidation of the corresponding in situ formed hydroquinolate complexes with FeCl3.13 Semiquinolato-bridged scandium(III) species were reported to promote self-organised electron transfer from d-transition metals (Ir, Fe) to 1,4-quinones.14,15</p><p>Targeted metal-redox chemistry with quinones has been a recurring issue for the rare-earth-metal couples Ln(ii)/Ln(iii)16 and Ce(iii)/Ce(iv).17 Especially in the case of molecular cerium chemistry,17 its unique single-electron-transfer (SET) pathway has recently been extended beyond the traditional application of ceric ammonium nitrate (CAN; redox potential of 1.61 V vs. NHE) in organic synthesis18 to photoredox catalysis.19 On the other hand, redox protocols are known to provide efficient access to metalorganic CeIV complexes. Typically, such CeIII → CeIV transformations are promoted by halogenating oxidants (e.g. C2Cl6, Ph3CCl, PhICl2, TeCl4, FcPF6, FcBF4, Ph3CBF4, Ph3CPF6, I2),20 silver salts (AgX, X = F, I, BF4, OTf)21 or dioxygen.20b,22</p><p>Archetypical 1,4-benzoquinone (BQ) has been established as a versatile oxidant for the synthesis of homoleptic ceric complexes CeL4 from cerous ate complexes [CeL4M(do)x] via tandem oxidation-ligand redistribution protocols (L = monoanionic ligand, M = alkali metal and do = donor solvent; separation of an alkali-metal hydro-/semiquinolate).23 In the presence of sterically demanding ligands L, BQ was also shown to form hydroquinolato (hq)–bridged ceric complexes of the general composition [L3Ce–OC6H4O–CeL3].20g,24 This very CeIII → CeIV transformation was pioneered by Sen et al. in 1992, resulting in the isolation of [(tBu3CO)3Ce(OC6H4O)Ce(OCtBu3)3] (Chart 1, I).24a In the same paper, the oxidation of Ce(OCtBu3)3 with 2,6-di-tert-butyl-1,4-benzoquinone to the terminal CeIV-semiquinolate radical (tBu3CO)3Ce(O2C6H2tBu2) was described as evidenced by 1H NMR and EPR spectroscopic measurements.24a More recently, Schelter et al. reported on hq-bridged complex II resulting from the oxidation of cerous Ce(BINOlate)3(thf)Li3(thf)4 with 0.5 equivalents of BQ.24b Similarly, our group synthesized [Ce{N(SiMe3)2}3]2(μ2-O2C6H4)24c (III) and (CeCpR3)2(μ2-O2C6H4) (CpR = C5H4Me (IV) and C5H4(SiMe3) (V)).20g In contrast, the reaction of BQ with [Ce(Me2pz)3]x featuring the sterically less demanding and increasingly nucleophilic 3,5-dimethylpyrazolato ligand (Me2pz) led in fact to a transient CeIV hydroquinolate species (as indicated by the characteristic colour change), which, however, at ambient temperature was converted into the isolable trimetallic CeIII complex Ce3(pchd)2(Me2pz)5(thf)2 (pchd = 1,4-bis(3,5-dimethylpyrazol-1-yl)cyclohex-2,5-diene-1,4-diolato).23c Apparently, the new pchd ligand formed via 1,4-nucleophilic attack at bq by two adjacent Me2pz ligands. This nucleophilic reaction pathway could be prevented by using bulky tBu groups on the pz ligand, but homolpetic Ce(tBu2pz)4 was formed as the main ceric product via irreversible ligand rearrangement.23c</p><p>As such, cerium–1,4-benzoquinone couples have revealed distinct redox chemistry, we became curious about as to what extent such redox transformations are affected by both the type of 1,4-benzoquinone oxidant and the molecular CeIII precursor. The present study uncovers some unexpected correlation between CeIV–hydroquinolato stabilisation and quinone oxidant strength, as well as a new path to p-semiquinolato–radical-bridged rare-earth-metal complexes.</p><!><p>The quinones used in this study comprise 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ), tetrachloro-1,4-benzoquinone (Cl4BQ), 1,4-benzoquinone (BQ), tetramethyl-1,4-benzoquinone (Me4BQ), 2,5-di-tert-butyl-1,4-benzoquinone (tBu2BQ), 1,4-naphthoquinone (NQ), and 9,10-anthraquinone (AQ). All are commercially available and were selected according to their reduction potentials spanning a E0 range of 89 to 887 mV (2e−/2H+, vs. NHE, cf., Scheme 1).5,25 The cerous precursors were chosen according to the criteria solubility, weak nucleophilicity, proven access to the tetravalent state, and a stabilizing effect on the latter. Furthermore, the use of sterically bulky ligands was assumed to minimise the occurrence of ligand redistribution reactions. Accordingly, homoleptic Ce[N(SiMe3)2]3 (1) appeared to be an ideal benchmark system.24c After additional investigations into the respective pyrazolate chemistry, the abovementioned [Ce(R2pz)3] (R = Me, tBu) were discarded because of persisting alternative reaction pathways like 1,4-nucleophilic attack of BQ by Me2pz and ligand redistribution (formation of Ce(tBu2pz)4).23c The new pyrazolate studies clearly confirmed that steric hindrance of both the pyrazolato ligand and the 1,4-benzoquinone can minimise/counteract such undesired reactions, but the formation of product mixtures seems inevitable. Products crystallised from these reactions include minor amounts of ceric [Ce(tBu2pz)3(thf)]2(Me4hq) or a cerous product of partial pyrazolyl-promoted nucleophilic attack Ce3(bpad)(pasq)(Me2pz)6(thf) (bpad = 1,4-bis(3,5-dimethylpyrazol-1-yl)anthra-1,4-diolato; pasq = 1-(3,5-dimethylpyrazol-1-yl)anthra-1,4-semiquinolato) (84%) mixed with semiquinolate [Ce(Me2pz)2(thf)2(asq)]2 (asq = anthra-semiquinolato; cf. ESI† for structural details). The use of CeIII halides was discarded mainly for solubility issues.</p><p>In addition to silylamide 1, the siloxide derivatives [Ce{OSi(OtBu)3}3]2 (2)21d,26 and [Ce(OSiiPr3)3]2 (3) were assessed as suitable cerous precursors. Complexes 2 and 3, with and without intramolecular donor site, respectively, were readily obtained in pure form via protonolysis of 1 with the corresponding silanol.26 The crystal structure of the new complex 3 revealed a dimeric arrangement with two μ2-bridging and four terminal siloxy groups (Fig. 1), similar to that found for tris(tert-butoxy)siloxy congener 2 or [Ce(OSiPh3)3]2 27 or [Ce(OCHtBu2)3]2.28 The Ce–Oterminal (2.1659(14) and 2.1671(14) Å) and the Ce–Oμ2 distances (2.3951(12) and 2.4030(12) Å) of 3 are slightly shorter than those in 2 (Ce–Oterminal 2.202(3), 2.186(3) Å; Ce–Oμ2 2.532(2) Å) and [Ce(OSiPh3)3]2 (Ce–Oterminal 2.141(7), 2.185(6) Å; Ce–Oμ2 2.345(6), 2.583(5) Å) reflecting the lower coordination number (CN 4 vs. 5), but slightly longer than in [Ce(OCHtBu2)3]2 (Ce–Oterminal 2.142(2), 2.152(3) Å; Ce–Oμ2 2.363(3) Å).28 The 1H NMR spectrum of 3 in C6D6 shows two singlets at −28.82 and −17.23 ppm for the μ2-OSiiPr3 groups and two singlets at 6.46 and 9.09 ppm for the terminal siloxy ligands indicating a non-fluxional dimeric species in non-coordinating solvents. When recorded in THF-d8, only two signals for the OSiiPr3 groups appeared, in accordance with the formation of a monomeric adduct [Ce{OSiiPr3}3(thf-d8)x].</p><!><p>Treatment of Ce[N(SiMe3)2]3 (1) with each 0.5 equivalents of DDQ, Cl4BQ, Me4BQ, tBu2BQ and NQ, in a mixture of toluene and n-hexane, immediately led to a colour change from yellow to dark brown. Upon recrystallisation from toluene/n-hexane mixtures it was possible to isolate the hydroquinolato-bridged complexes [Ce{N(SiMe3)2}3]2(μ2-O2C6Cl4) (4Cl4hq), [Ce{N(SiMe3)2}3]2(μ2-O2C6Cl2(CN)2) (4ddhq), [Ce{N(SiMe3)2}3]2(μ2-O2C6Me4) (4Me4hq), [Ce{N(SiMe3)2}3]2(μ2-O2C6tBu2H2) (4tBu2hq) and [Ce{N(SiMe3)2}3]2(μ2-O2C10H6) (4nhq) in very good crystalline yields of 71 to 90% (Scheme 1). The crystal structures of the new complexes 4xhq are isostructural to the previously reported derivative 4hq24c and only differ in the bridging hq linker (Fig. 2). The Ce1–O1 distances of 2.084(6) to 2.173(2) Å (for a full list of interatomic distances see Table 1) are in the same range as found for other hq-bridged cerium complexes (2.086(10)–2.143(5) Å).20g,24 Likewise, the Ce1–N bond lengths compare well to other CeIV silylamides like [Ce{N(SiMe3)2}3]2(μ2-O2C6H4) (2.2388(14)–2.2487(14) Å),24c Ce[N(SiMe3)2]3Cl (2.217(3) Å),20a and Ce[N(SiHMe2)2]4 (2.2378(11)–2.2574(11) Å).20d Also, the C–C distances of the hq linker converge as the expected aromatic ring is formed and the C–O distances of 1.318(3) to 1.378(3) Å corroborate the formation of C–O single bonds.</p><!><p>NMR spectra recorded in C6D6.</p><p>UV-Vis spectra recorded in toluene.</p><p>Determined in THF using c(analyte) = 2 mM and c(electrolyte) = 0.1 M and a scan rate of 50 mV s−1.</p><p>Determined in toluene-d8 at 0 °C.</p><!><p>The 1H NMR spectra of compounds 4xhq in C6D6 show singlets for the trimethylsilyl (TMS) groups at 0.43 to 0.45 ppm along with signals for the bridging hydroquinolato moieties. Further, the 13C{1H} NMR spectra of 4ddhq and 4nq display a singlet for the TMS groups at 5.6 ppm and signals in the aromatic region for the different hydroquinolates, indicative of a successful reduction of the respective benzoquinone derivatives. The characterisation of 4tBu2hq in solution (C6D6) was not feasible, due to the prevailing equilibrium shown in Scheme 2, and ready back-formation of 1 and 2,5-di-tert-butyl-1,4-benzoquinone.</p><p>While the 1H NMR spectrum of 4tBu2hq primarily shows signals for the starting materials and only minor product signals, its DRIFT spectrum indicated the absence of any strong C <svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="13.200000pt" height="16.000000pt" viewBox="0 0 13.200000 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.017500,-0.017500)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z M0 280 l0 -40 320 0 320 0 0 40 0 40 -320 0 -320 0 0 -40z"/></g></svg> O absorption band, and therefore the stability of 4tBu2hq in the solid state (see Fig. S10 and S45 in ESI†). In contrast, complexes 4hq, 4Cl4hq and 4ddhq derived from the stronger oxidizing quinones are very stable in the solid state and in solution. This fits again well with the already pronounced instability of 4Me4hq and 4nhq which slowly decompose in n-hexane and toluene at ambient temperature and rapidly undergo decomposition in THF. Tracking of the progress of the decomposition by 1H NMR spectroscopy revealed the formation of 1 and other paramagnetic CeIII species which, however, could not be identified. The progressing decomposition can also be seen in the ligand-to-metal charge transfers observed in the UV-Vis spectra (Fig. S68, ESI†). As the spectra of 4hq, 4Cl4hq and 4ddhq show mainly one strong absorption band at around 500 nm (ε > 5060 L mol−1 cm−1), the spectra of 4Me4hq and 4nhq show several absorption bands with significantly lower intensities (ε < 4400 L mol−1 cm−1) indicative of CeIII species and therefore redox decomposition of the compounds.</p><p>All attempts to isolate putative 4ahq, derived from the weakest oxidising quinone under study, namely 9,10-anthraquinone (E0 = 89 mV; 2e−/2H+, vs. NHE),5 were unsuccessful with the reaction mixtures showing no colour change immediately after addition of the anthraquinone. However, a colour change from yellow to green occurred after two days and the respective 1H NMR spectrum showed multiple paramagnetic signals.</p><!><p>Reacting cerous siloxides 2 and 3 with the selected quinones in THF immediately gave a colour change of the reaction mixtures (from colourless to: dark purple (BQ), dark red (Cl4BQ), dark yellow/orange (DDQ), pale purple (tBu2BQ), pale blue (Me4BQ), pale green (NQ)). The ceric compounds [CeL3(thf)]2(μ2-O2C6H4) (5hq, 6hq), [CeL3(thf)]2(μ2-O2C6Cl4) (5Cl4hq, 6Cl4hq), [CeL3(thf)]2(μ2-O2C6Cl2(CN)2) (5ddq, 6ddq), with L = OSi(OtBu)3 or OSiiPr3 derived from quinones with a relatively strong oxidising effect were successfully isolated from these reactions (Scheme 1).</p><p>However, the weakly oxidizing quinones Me4BQ and NQ did not lead to tetravalent cerium species, as indicated by the detection of only paramagnetic signals in the 1H NMR spectra (for an example of such a 1H NMR spectrum, see Fig. S34 in the ESI;† formation of semiquinolates, vide infra). The accessible complexes 5 and 6 were obtained in moderate to good crystalline yields of 42 to 71% upon recrystallisation from THF or THF/Et2O mixtures. Crystals suitable for XRD analysis were obtained for complexes 5hq, 5Cl4hq, 6Cl4hq, 6ddhq and 6tBu2hq, revealing the same structural motif as complexes 4, that is two CeL3 moieties connected via a hydroquinolato linker (Fig. 2).</p><p>Strikingly, the 1H NMR spectrum of 6tBu2hq indicated the existence of an equilibrium similar to that of ceric 4tBu2hq (cf.Scheme 2). However, along with the reactants 3 and tBu2BQ additional signals assignable to distinct dia- and paramagnetic decomposition products were detected. Further, the crystal structures of complexes 5 and 6 show that the cerium atoms are additionally coordinated by THF donor molecules. The Ce1–Osiloxide distances of 2.066(2) to 2.1534(10) (see Table 2 for a complete list of interatomic distances) compare well to other ceric siloxides like Ce{OSi(OtBu)3}4 (2.089(2)–2.157(2) Å 26 and 2.084–2.160 Å 21d) or Ce{OSiPh3}4(dme) (2.098(1)–2.133(1) Å).29 Also, as seen for the silylamides 4, the Ce1–Ohq distances of 2.1244(10) to 2.2325(16), as well as the C–C and C–O distances underline the formation of an aromatic hq linker.20g,241H NMR spectroscopic measurements also validate the formation of CeIV species, showing a sharp singlet for the tert-butyl groups and a doublet plus a septet for the iso-propyl groups depending on the siloxy co-ligand.</p><!><p>NMR spectra recorded in THF-d8.</p><p>Spectra recorded in toluene.</p><p>Spectra recorded in THF.</p><p>Determined in THF using c(analyte) = 2 mM and c(electrolyte) = 0.1 M and a scan rate of 50 mV s−1.</p><!><p>Cyclic voltammetry (CV) measurements of complexes 4xhq, 5xhq and 6xhq have been conducted at ambient temperature in 0.2 mM solutions in THF and 0.1 M [nPr4N][B(C6H3(CF3)2-3,5)4] as a support electrolyte, and referenced vs. Fc/Fc+. Due to the low stability of compounds 4Me4hq, 4tBu2hq and 4nhq in polar solvents CV measurements of these complexes were not feasible. Most of the CV measurements revealed successive quasireversible (4) or irreversible (5/6) CeIV → CeIII reduction steps, but badly resolved (for Epc values see Tables 1 and 2). The detection of two closely adjacent redox events (Epc values) in some cyclic voltammograms may correspond to a successive reduction/oxidation of the cerium centres. Similar features were also described for the hq-bridged Ce(iv)–BINOLate complex II.24b All complexes under study display redox processes with a large separation of Epc and Epa (ΔE ≈ 0.6 V for 4; 1.5 V for 5 and 1.0 V for 6).</p><p>Representatively, the cyclic voltammograms of the DDQ-functionalized CeIII/CeIV redox couples are depicted in Fig. 3 (top graphic). The silylamide complexes 4 gave reduction potentials similar to those reported for halogenido-functionalised ceric complexes Ce[N(SiMe3)2]3X (E1/2 = −0.56 (X = F), −0.30 (X = Cl), −0.31 (X = Br)) with E1/2 values of −0.46 V for 4Cl4hq and −0.36 V for 4ddhq.30 Only 4hq with E1/2 = −0.76 V gave a significantly higher stabilisation by 0.20 V. The extra-large separation of the reduction/oxidation events observed for the siloxide complexes 5 and 6 had been noticed previously for rare-earth-metal siloxides and was assigned to oxidation-state-dependent ligand reorganisation processes.31</p><p>Stabilisation of the tetravalent oxidation state of cerium in complexes 4, 5, and 6 increases in the order of N(SiMe3)2 < OSi(OtBu3)3 < OSiiPr3 as co-ligand (Fig. 3/bottom) which is in accordance with previous findings.20f,26,30,31c Surprisingly, the stabilisation of CeIV proceeds in reverse order of the oxidation potential of the 1,4-quinones under study giving the most stable complexes for the hydroquinolato-bridged complexes and the least stable compounds for its 2,3-dichloro-5,6-dicyano-hydroquinolato congeners. A reason for this trend could be the increasingly electron-deficient nature of the aromatic hydroquinolato linkers due to the large −I effect of the substituents. The CeIV oxidation state can be stabilised by increasing donor strength of the ligands.32 Based on this, it seems surprising that isolable complexes 4Me4hq and 4nhq, derived from weakly oxidizing quinones, are not stable in solution at ambient temperature. This might be a result of another reaction pathway preferred after formation of the hydroquinolato-bridged CeIV complexes (like following up redox processes and the formation of CeIII semiquinolates, cf. vide infra).</p><!><p>Having investigated the electrochemical reduction of compounds 4, 5 and 6, the chemical reduction with cobaltocene (CoCp2) (−1.31 V vs. Fc/Fc+ in DME)2a was attempted, as it has already been shown to engage in such reductions.31a,33 Accordingly, treatment of a solution of 4hq in THF with two equivalents of CoCp2 resulted in a colour change from dark brown to pale yellow (Scheme 3). The 1H NMR spectrum of the reaction mixture showed complete consumption of CoCp2 and only broadened signals indicating the formation of a paramagnetic CeIII species. Crystallisation from a concentrated THF-d8 solution at −40 °C gave light brown crystals of the composition [(Ce{N(SiMe3)2}3)2(μ2-O2C6H4)][CoCp2]2 (7) (Fig. 4).</p><p>Complex 7 shows the same structural motif as 4hq but is flanked by two cobaltocenium cations. Compared to 4hq, the Ce–N and Ce1–O1 distances are elongated by approximately 0.19 Å as expected for the larger CeIII ion size.34 On the contrary, the bonding parameters within the bridging hydroquinolato linker did not change, further corroborating a cerium-borne redox chemistry. Reacting 4hq with one equivalent of CoCp2 did not lead to a mixed CeIII/IV complex but gave a mixture of 50% of 7 and 50% of unreacted starting material.</p><p>The reactions of CoCp2 with other complexes 4 to 6 in THF-d8 showed immediate decolourisation of the solution while the 1H NMR spectra of the reaction mixtures displayed only paramagnetic signals (for an example, see Fig. S33 in the ESI†). However, the isolation of additional reduced species similar to 7 was not successful.</p><!><p>A closer look at the reactions of cerous siloxides 2 and 3 with the weakly oxidising quinone Me4BQ (which did not produce any tetravalent cerium species; vide supra) revealed another important detail of the cerium–quinone redox system. Treatment of 2 or 3 with 0.5 equivalents of Me4BQ led to a colour change from colourless to light blue. Upon recrystallisation from THF dark blue crystals suitable for X-ray diffraction could be grown and were identified as cerous semiquinolates [CeL2(thf)2]2(μ2-O2C6Me4)2 (with L = OSi(OtBu)3 (8) or OSiiPr3 (9)) (Scheme 4).</p><p>Examining the reaction mixtures by 1H NMR spectroscopy in THF-d8 showed, besides paramagnetic signals for 8 and 9, a sharp singlet at 1.39 ppm (for 8) or a doublet plus a septet at 1.13 and 1.04 ppm (for 9), indicating the formation of homoleptic Ce[OSi(OtBu)3]4 or [Ce(OSiiPr3)4], respectively, as a result of the one-electron reduction of Me4bq followed by ligand redistribution. Crucially, such a reaction pathway seems unfeasible for complexes 4, since putative homoleptic "Ce[N(SiMe3)2]4" is unknown.20a Emergent kinetic constraints in the case of ceric complexes 4 were also suggested by the redox behaviour of [Ce{N(SiHMe2)2}3]2 derived from a less bulky silylamido ligand. Accordingly, the cerous bis(dimethylsilyl)amide complex was treated with one equivalent of both BQ and Me4BQ in THF-d8 and C6D6 (see Fig. S38–S41, ESI†). The 1H NMR spectra of these reactions suggest the formation of a tetravalent species of the composition "[Ce{N(SiHMe2)2}3]2(μ2-O2C6R4)". However, the ceric products appear to be unstable in solution at ambient temperature. In C6D6, the formation of Ce[N(SiHMe2)2]4 was observed in the reaction with BQ as well as other insoluble products. In THF-d8, the resulting product seemed more stable but after 24 h in solution also traces of decomposition products were found. The Me4BQ reaction in C6D6 also indicated successful oxidation, however, after 24 h the 1H NMR spectrum revealed signals for trivalent decomposition products as well as traces of Ce[N(SiHMe2)2]4. In THF-d8, the putatively formed hydroquinolate complex was even less stable, showing signals for trivalent by-products directly after addition of Me4BQ. In addition, the stability of 4bq was investigated in THF-d8 showing small amounts of decomposition products like Ce[N(SiMe3)2]3 after 24 h (Fig. S42, ESI†).</p><p>Regrettably, purification of complexes 8 and 9 was impeded by co-crystallisation with the ceric by-products CeL4. The crystal structures of 8 and 9 revealed two six-coordinate cerium atoms surrounded by two siloxy ligands, two THF donor molecules and two bridging tetramethyl semiquinolato moieties (Fig. 5). The Ce1–Osilanolato distances are elongated by about 0.1 Å compared to the respective tetravalent compounds 5 and 6 and in accordance with other CeIII siloxides like 3, Ce{OSi(OtBu)3]3(thf)3 (2.243(2)–2.249(2) Å), [Ce{OSi(OtBu)3}3]2 (Ce–Oterm 2.186(3)–2.202(3) Å),26 Ce{OSiPh3}3(thf)3 (Ce–Oavg 2.222(4) Å)35 and [Ce{OSiPh3}3]2 (Ce–Oterm 2.141(7)–2.184(6) Å).27 As expected for semiquinolato ligands the six-membered rings display two shortened C–C and four elongated C–C bonds. Additionally, the six-membered rings are slightly bent in comparison to the flat aromatic hydroquinolato linkers in complexes 4, 5 and 6 resulting in an angle of 170.34° for 8 and 170.29° for 9, respectively (see Fig. 5, bottom). Notwithstanding, the bridging radicals engage in significant π-stacking as indicated by close semiquinolato–semiquinolato distances of 3.112 Å for 8 and 3.156 Å for 9. Overall, complexes 8 and 9 display the same arrangement of the semiquinolato radical bridges as observed in complexes [LnCl2(THF)3(μ-Me4sq)2]2 (Ln = Y, Gd) (Ct⋯Ct 3.097 Å; Ct = centroid of benzene rings).13</p><p>To investigate the electronic behaviour of the bridging semiquinolates, X-band EPR spectra of compounds 8 and 9 were recorded from a crystal powder sample at 123 K (Fig. 6). For both complexes cw-EPR spectra are composed of two distinct sets of resonances, one that results from the transition within the Kramers doublet corresponding to mj = ±1/2 and one at half-field, H ≈ 160 mT. The transition for the mj = ±1/2 state associates with an axial g tensor with principal components gII = 2.094 and g⊥ = 2.032 for 8 and gII = 2.088 and g⊥ = 2.032 for 9, respectively. The transition locating at half-field gives rise to a very broad dispersion line with gII ≈ 4.359 for 8 and gII ≈ 4.351 for 9, respectively. Notably, an identical line pattern derives from a frozen 2-Me-THF solution at 77 K; cf. Fig. S67, ESI,† for pertinent details. The cw-EPR spectra corroborate the presence of Ce3+, and agree with early work on mononuclear complexes of Ce3+.36 This indicates a radical–radical π-bonding, as it was recently shown for yttrium and gadolinium semiquinolates [LnCl2(thf)2]2(μ2-O2C6Me4)2 (Ln = Y or Gd).13 Additionally and also similar to the yttrium semiquinolate, complex 9 shows a signal with a g value of 1.999, which most likely results from a non-coupled radical impurity.</p><p>Note that the reactions of cerium siloxides 2 and 3 with 1,4-naphthoquinone resulted also in the formation of homoleptic ceric siloxides as well as paramagnetic by-products (cf. Fig. S36, ESI†) indicating a similar reactivity as observed for Me4BQ. Unfortunately, any putative semiquinolate complexes could not be isolated.</p><!><p>Cerium(iii) silylamides and siloxides are suitable reagents for assessing the oxidising power/reducibility of differently substituted 1,4-quinones in non-aqueous solutions. The cerium–quinone redox matching is revealed by the ease of formation of CeIV hydroquinolates [CeL3]2(μ2-O2C6R4), in the case of the parent 1,4-benzoquinone (BQ) or when R represents electron-withdrawing groups (Cl, CN). Depending on their reduction potential, alkyl-substituted BQs engage in redox equilibria, with the CeIV hydroquinolate species being preferentially stable in the solid state, but also afford semiquinolates via redox ligand redistribution. The structurally characterized siloxide semiquinolate complexes [(CeL2(thf)2)(μ2-O2C6Me4)]2 (L = OSi(OtBu)3, OSiiPr3) exhibit a molecular arrangement, recently detected for [LnCl2(THF)3(μ-Me4sq)2]2 (Ln = Y, Gd).13 The stabilisation of the tetravalent oxidation state in hydroquinolato-bridged complexes [CeIVL3]2(μ2-O2C6R4) was examined by electrochemical measurements, as well as NMR and UV/Vis spectroscopies. In accordance with previous findings,16,18–20 the stability of the ceric complexes increases in the order of N(SiMe3)2 < OSi(OtBu3)3 < OSiiPr3 as supporting ligand, but surprisingly drops in reverse order of the oxidation potential of the 1,4-quinones, being the least stable for the 2,3-dichloro-5,6-dicyano-hydroquinolato congener. The preferred formation of hydroquinolato-bridged silylamides [Ce{N(SiMe3)2}3]2(μ2-O2C6R4) seems kinetically favoured. Finally, the electrochemical reduction of the hydroquinolato-bridged ceric complexes [CeIVL3]2(μ2-O2C6R4) can be mimicked by chemical reduction with cobaltocene, as shown for the isolation and structural characterisation of cerous [(Ce{N(SiMe3)2}3)2(μ2-O2C6H4)][CoCp2]2. The cerium–quinone redox matching and tuning might be used as a role-model in tetravalent praseodymium and terbium chemistry.</p><!><p>There are no conflicts to declare.</p>

PubMed Open Access