Quantitative estimation of the protein-ligand absolute binding free energy (ABFE) <i>in silico</i> is a promising strategy to improve the efficiency of drug design. Alchemical free energy simulation methods such as free energy perturbation and thermodynamic integration (TI) based on all-atom molecular dynamics (MD) simulations in an explicit solvent were demonstrated to provide accurate ABFE predictions for diverse protein targets. Recent advances in graphic processing units (GPUs), force fields, and sampling algorithms have expanded the use of alchemical simulations<b> </b>in drug discovery projects. In this work, we developed an automated protocol for alchemical ABFE calculations utilizing AMBER biomolecular simulation package. The protocol includes protein and ligand parametrization, set-up of protein-ligand restraints using virtual bond approach, GPU-accelerated MD TI simulations, and statistical analysis of the MD output data. Here, we report on exploiting this protocol for large-scale ABFE calculations for two projects: 1) identifying binders of WD-repeat (WDR) domain of leucine-rich repeat kinase 2 (LRRK2) for CACHE (critical assessment of computational hit-finding experiments) challenge; 2) binding of AMPA receptor (AMPAR) to its non-competitive small-molecule antagonist GYKI 53655 (GYKI). Both systems are of interest for drug design: LRRK2 WDR is a potential target for Parkinson’s disease treatment, and AMPAR is a target for antiepileptic drugs. For LRRK2 WDR, we calculated ABFEs for approximately 1000 small molecules from Enamine REAL and Mcule databases which resulted in identifying 43 molecules with nanomolar predicted binding affinity. For AMPAR-GYKI system, our previous MD simulations demonstrated that protein linkers in the allosteric binding site are flexible while ligand binding mode remains stable. To account for linker flexibility, we computed ABFEs for 800 structures selected equidistantly from 500 ns AMPAR-GYKI MD trajectories. Results demonstrate the dependency of ABFE on protein conformation and reveal key protein-ligand interactions for the binding of GYKI to AMPAR.

Large-scale protein-ligand binding free energy calculations using automated GPU-accelerated molecular dynamics simulations

Where thousands of chemistry professionals meet to share ideas and advance scientific and technical knowledge. Spring 2023 abstracts, posters, and presentations will launch at the end of March.

ACS Spring 2023

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.