Incorporation of Protein Flexibility in Structure-Based Drug Discovery of HIV-1 Protease and Bacterial Hsp70 Chaperone.

[摘要] Damm et al. utilized a multiple protein structure (MPS)-based method and discovery an allosteric site of HIV-1 protease (HIV-1p) that had not been described before. Based on this new site, a MPS-based pharmacophore model was built to screen for potential HIV-1p allosteric binder, in which several new scaffolds were identified. Markush searches based on successful inhibitors were used to extend the chemical space of lead compounds. Subsequently biochemical and kinetics experiments were carried out to fully describe the mode of action of the allosteric inhibitors.An alternative MPS-based method is mixed-solvent molecular dynamics (MixMD), which can be used to identify binding hot-spots on protein surfaces. MixMD fully samples the conformational space of the protein while uses small organic solvents as probes to compete with water on the protein surface. Using HIV-1p as a test case, MixMD successfully identifies the allosteric site described by Damm et al. Other known and proposed sites were also identified. Unusual ;;false positives” were actually identified as the common protein-protein interfaces utilized in crystal packing.MPS has also been applied to heat shock protein 70 (Hsp70), a chaperone that is highly conserved among different species. Utilizing the ensemble of conformations generated through molecular dynamics simulations, potential hinge residues of Hsp70 were identified. Mutagenesis of these hinge residues has been carried out and mutation of a key residue, Gly228, is shown to decouple the allosteric communication between its catalytic activity and substrate binding activity. Hence, this demonstrates the strength of MPS methods in identifying key structural elements in a protein.Lastly, efforts have been made to evaluate existing SBDD software through the Community Structure Activity Resource (CSAR). Many scoring approaches were assessed and their correlation to experimental binding affinity ranged R2 = 0.14-0.51, with no particular scoring approach having an overwhelming advantage over others. There is a need for CSAR to develop data sets of congeneric series with a range of different physical characteristics for improved training of computational methods.

[发布机构] University of Michigan

[效力级别] Computational Chemistry

[关键词] Medicinal Chemistry;Computational Chemistry;Protein Dynamics;Drug Discovery;HIV Protease;Heat Shock Protein 70 Hsp70 Chaperone;Biological Chemistry;Pharmacy and Pharmacology;Chemistry;Science (General);Health Sciences;Science;Medicinal Chemistry