Computational Alchemy Using Accelerated GPU Calculations: Fine Structural Tuning Inhibitors of L,D-transpeptidase 2 from Mycobacterium tuberculosisстатья
Статья опубликована в журнале из списка RSCI Web of Science
Статья опубликована в журнале из списка Web of Science и/или Scopus
Аннотация:The computational alchemy is based on equilibrium methods of molecular dynamics (MD),including thermodynamic integration (TI) or free energy perturbation (FEP). It makes it possible to evaluate the energy characteristics of a subtile alchemical transformation of the prototypecompound structures in order to modulate their functional properties, what can be used searching for new analogues of enzyme inhibitors. In recent years, the method is gaining new opportunities due to high-performance computing as different MD engines got the ability to perform TI or FEP calculations using GPU acceleration. We have tested the method on two experimentally studied enzyme-inhibitor systems: penicillin acylase from E. coli (PaEc) and Influenza Avirus neuraminidase (NaIAv) and performed computational alchemy calculations for pairs of their known inhibitors. Thus, the validated methodology was then applied to estimate binding of L,D-transpeptidase 2 from M. tuberculosis (LdtMt2) to the structural analogs of inhibitors previously discovered in our group. We have found that the performance of computational alchemy predictions is highly dependent on properly chosen simulation parameters and at optimal performancethe difference between the calculated and experimentally determined relative binding energies is less than 1 kcal/mol. Thus, for optimal application to novel enzyme—inhibitor or receptor—ligand systems, we recommend prior adaptation of this method, based on validation of force field combinations as well as water models, to experimentally determined binding data of structurally related inhibitors/ligands to the molecular target of interest, if available, and then to proceed to the study of new compounds. High-performance computational alchemy can be used as a useful integrative part of the methodology searching for new enzyme inhibitors or receptor ligands and optimizing their structure at drug design.