Virtual Screening And Molecular Dynamic Simulation Of ?-Glucosidase Inhibitors

In this paper, the target protein is a-glucosidase which plays a key role in the biological glucose metabolism, mainly involved in sugar decompositions. The competitive inhibitor can help slow down the rate of glucose production, slow down the intestinal absorption of glucose, and maintain the blood glucose concentration at a normal level, so as to control the occurrence of diabetes and complications. This paper aims at the study of the structure-activity relationship using molecular docking, virtual screening and molecular dynamics simulation.(1) In order to make the virtual screening results more accurate, we have tested three kinds of docking tools including UCSF DOCK, LeDock, AutoDock Vina, comparing with the crystal complex structure, and found that LeDock is the best. Considering that UCSF DOCK can do massively parallel computation, we combined UCSF DOCK and LeDock into a comprehensive approach.(2) In the virtual screening stage, DOCK and LeDock were applied to 3.68 million compounds of ZINC Leads Now database, against the active site of a-glycosidase, using 11 crystal structure models, employing a multi-scoring treatment including consensus scoring, consensus docking, pose scaling, size normalization and clustering. The energy value of ZINC65748051 and ZINC71751922 are-9.71 kcal/mol and-8.73 kcal/mol respectively, both better than the energy-8.58 kcal/mol for acarbose.(3) The selected 20 molecules with better poses were subjected to 50 ns molecular dynamics simulations and binding free energy calculation. We found that amino acids Asp203, Tyr299, Asp327, Trp406, Trp441, Asp443, Met444, Glu446, Asp542 and Phe575 contribute more than other residues to the protein-ligand interaction. Analyzing their structural characteristics, we found that aromatic amino acids, acidic amino acids and sulfur-contained amino acid is conducive to enhancing ligand binding. Comparing the binding free energy values,13 compounds were selected, leading to a rule that the combination of three factors, i.e. a less branched linear structure, cyclic structure and more N, O, S, provides better interaction and inhibition.