Scoring binding affinity of multiple ligands using implicit solvent and a single molecular dynamics trajectory: application to influenza neuraminidase.

We explore a perturbative approach to calculation of binding free energy of multiple ligands, based on a single molecular dynamics simulation of a reference ligand-receptor complex and analysis via a hybrid force field/continuum model potential. The methodology is applied to prediction of relative binding free energies of 10 Influenza neuraminidase inhibitors, using Poisson-Boltzmann and generalised Born models of implicit solvent. These single-step MM-PB/SA and MM-GB/SA approaches predict the experimentally most potent ligand as first- or second-ranked according to total binding free energy. Ranking of inhibitors displays only moderate sensitivity to the choice of reference trajectory and ligand partial charge scheme. When ranked according to total electrostatic binding free energy, correlation with experiment improves (r(2) of 0.72); this may be related to underestimated first solvation shell effects by the implicit water models. Therefore, to increase the generality of this single-step approach as part of a potential computational compound optimisation strategy, further development of the treatment of short-range solvent interactions is warranted.