Improvement of Sampling Efficiency through Combined use of Molecular Dynamics Simulations with Implicit and Explicit Solvent Models

Abstract

The molecular dynamics (MD) simulation is powerful tool for analyzing the dynamical or physicochemical properties of biomolecules that are not accessible experimentally. It is however that the time scale of the MD simulation is not enough for examining the biologically significant events such as protein folding, domain movement, and so on, due to the computational limitation. Therefore the enhanced sampling methods to overcome the computational limitation are so important.This study aims at the reduction of the computational cost for sampling through combined use of the simulations with the implicit and explicit solvent; the former is used for broad sampling of the conformational space and the latter for correcting the conformational space sampled by the former simulation. The system with the explicit solvent was coupled with the subsystem solvated with the implicit solvent, which accelerate the conformational sampling. The replica exchange molecular dynamics (REMD) was adopted for the subsystem in this study. For estimating the ability of the new method, chignolin, 10-residue mini protein, was chosen as a model system. The results show that the ability of conformational sampling by this method is higher than that of the conventional MD simulation and nearly corresponds to the temperature REMD simulation with the explicit solvent. Furthermore the computational cost was reduced by approximately six-fold as compared as the temperature REMD on this system.