Toward Automated Free Energy Calculation with Accelerated Enveloping Distribution Sampling (A-EDS).

Jan Walther Perthold,Chris Oostenbrink,Dražen Petrov

doi:10.1021/acs.jcim.0c00456

Abstract

Free-energy perturbation (FEP) methods are commonly used in drug design to calculate relative binding free energies of different ligands to a common host protein. Alchemical ligand transformations are usually performed in multiple steps which need to be chosen carefully to ensure sufficient phase-space overlap between neighboring states. With one-step or single-step FEP techniques, a single reference state is designed that samples phase-space not only representative of a full transformation but also ideally resembles multiple ligand end states and hence allows for efficient multistate perturbations. Enveloping distribution sampling (EDS) is one example for such a method in which the reference state is created by a mathematical combination of the different ligand end states based on solid statistical mechanics. We have recently proposed a novel approach to EDS which enables efficient barrier crossing between the different end states, termed accelerated EDS (A-EDS). In this work, we further simplify the parametrization of the A-EDS reference state and demonstrate the automated calculation of multiple free-energy differences between different ligands from a single simulation in three different well-described drug design model systems.

Highlights

With the paradigm shift from the rigid lock-and-key hypothesis of receptor−ligand recognition to more dynamic conceptions of molecular interactions involving structural ensembles rather than single conformations and recent advances in computer technology, molecular dynamics (MD) simulations are being used more and more in the field of drug design
To determine the A-Enveloping distribution sampling (EDS) acceleration parameters Emax and Emin, and the freeenergy offset parameters ΔFiR, parameter-search simulations were performed for 10 ns for the ligands solvated in water
To calculate the free-energy offset parameters ΔFiR for the end states of the ligands bound to the proteins, accelerated EDS (A-EDS) acceleration parameters Emax and Emin were held fixed during the parametersearch simulation in the protein systems, as convergence of these parameters was very slow in preliminary simulations, given the generally complicated and possibly non-Gaussian energy landscapes of host−guest systems, where the EDS region is not surrounded by a homogeneous and fast-relaxing medium, but by the protein environment

Summary

■ INTRODUCTION

With the paradigm shift from the rigid lock-and-key hypothesis of receptor−ligand recognition to more dynamic conceptions of molecular interactions involving structural ensembles rather than single conformations and recent advances in computer technology, molecular dynamics (MD) simulations are being used more and more in the field of drug design. To calculate the free-energy offset parameters ΔFiR for the end states of the ligands bound to the proteins, A-EDS acceleration parameters Emax and Emin were held fixed during the parametersearch simulation in the protein systems, as convergence of these parameters was very slow in preliminary simulations, given the generally complicated and possibly non-Gaussian energy landscapes of host−guest systems, where the EDS region is not surrounded by a homogeneous and fast-relaxing medium, but by the protein environment. To calculate relative binding free energies for the ligands, A-EDS equilibrium simulations were performed both in water and in the proteinbound state for all systems with the A-EDS parameters determined for each acceleration level in the parameter-search simulations. For the TRP systems, calculated binding free energies were in better agreement with experimental data for acceleration levels of 1σ and 2σ than with an acceleration level of 3σ (Table S4 in the Supporting Information), possibly caused by insufficient state transitions for the given simulation time with 3σ. Acceleration level of 3σ might perform generally better if enough simulation time is generated, as the energy landscape is less flattened and important energy minima are more pronounced

■ CONCLUSION

■ ACKNOWLEDGMENTS

■ REFERENCES