Chapter 3 - Tuning the Solvation Term in the MM-PBSA/GBSA Binding Affinity Predictions

Abstract

Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) and Molecular Mechanics Generalized Born Surface Area (MM-GBSA) are widely used methods for the prediction of binding free energies in drug design/discovery. Indeed, their computational efficiency makes them applicable also within virtual screening protocols. Thus, in order to be useful for drug design/discovery purposes, MM-PBSA and MM-GBSA binding energy predictions have to correlate well with experimental activities. Nowadays the global effort to find a way to improve the predictivity of MM-PBSA/GBSA calculations is also focused on the solvation term by using various approaches. This chapter reports on the application of MM-PBSA/GBSA methods within the process of drug discovery and, in particular, on strategies that can be applied to improve the correlation between MM-PBSA/GBSA predicted binding affinities and experimental pharmacological activities by acting on the way the solvent is treated in such calculations. Indeed, in PB and GB models, the solvent is described as a continuous medium with a fixed dielectric constant (i.e. ε=80 for water), while a low internal dielectric constant is assigned to the solute (generally εin=1 or 2 for proteins). However, the default approach could in some cases lead to a weak correlation between predicted binding free energies and experimental data. The aim of this chapter is to present and exemplify the ways to improve the prediction of ligand binding affinity by acting on the solvation term. Different methods are observed in the literature, e.g. tuning the εin value depending on the features of the binding site, including a selection of explicit water molecules in order to better describe the solute-solvent interactions, tuning the grid size in PB calculations and/or using different PB solvers, or modifying the non-polar term of the solvation free energy. The pros and cons of the above mentioned methods will be critically discussed in order to help the reader in choosing the most performing protocol in terms of both calculation time and prediction quality, depending on the molecular system under evaluation.