A Reparametrized Implicit Solvent Model for Accurate Computation of Hydration Free Energies

Abstract

While solvation plays an important role in many chemical and biological processes, detailed solvent behavior is typically of much less interest than the effect of the solvent on the solute. Although explicit solvent treatment provides a very accurate description of the considered system, it is not necessary for many applications and can be replaced by implicit solvent models. A typical application is the computation of hydration free energies for small molecules. However a recent study [1] showed that there is still a gap between the accuracy of implicit models typically used in biomolecular simulations and explicit solvent treatment. To address this issue we have reparametrized an implicit solvent model consisting of a generalized Born term to model polar solvation and an extended nonpolar term that models cavity formation and Lennard-Jones interactions with water. When fitted to experimental hydration free energies of a large set of 499 small molecules, the model is able to reproduce hydration free energies with a similar error than those derived from explicit solvent computation, while still offering some degrees of freedom in the fit parameters. To test the transferability of the resulting model, we have applied it to the prediction of hydration free energies of Glycine-X-Glycine tri-peptides. This model will be part of the SIMONA [2] simulation package.[1] Knight, J. L. & Brooks III, C. L. Surveying implicit solvent models for estimating small molecule absolute hydration free energies. Journal of Computational Chemistry 32, 2909-2923 (2011).[2] Strunk, T. et al. SIMONA 1.0: An efficient and versatile framework for stochastic simulations of molecular and nanoscale systems. Journal of Computational Chemistry (2012).doi:10.1002/jcc.23089