Leveraging Conformational Diversity for Enhanced Structure-Based Virtual Screening: Insights from Molecular Dynamics Simulations of HIV-1 Protease-Ligand Complexes

Abstract

Structure-based virtual screening aims to identify high-affinity ligands by estimating binding free energies between proteins and small molecules. However, the conformational flexibility of both proteins and ligands challenges conventional rigid docking methods that assume a fixed receptor structure. In this study, we examined the impact of conformational diversity on binding energy calculations using 79 HIV-1 protease-ligand complexes. Molecular dynamics simulations were employed to generate structural ensembles for both proteins and ligands in aqueous environments. RMSD-based clustering was applied to reduce redundancy while preserving structural diversity. Binding energies were computed using van der Waals and electrostatic interactions. The results demonstrated that native protein-ligand pairs consistently yielded favorable binding energies, whereas non-native pairings often failed to reproduce binding. Furthermore, clustering thresholds influenced the balance between computational cost and interaction accuracy. These findings underscore the importance of incorporating multiple protein and ligand conformations in SBVS protocols to improve prediction reliability and support more effective drug discovery strategies.