Application of the lambda-dynamics method to evaluate the relative binding free energies of inhibitors to HCV protease.

Abstract

The lambda-dynamics method was used to calculate the relative binding free energies of inhibitors to the hepatitis C virus (HCV) protease. A total of seven HCV protease p-side product inhibitors were used in this study. The inhibitors are 6-mer peptides spanning P6-P1 (Ac-Asp-d-Glu-Leu-Ile-Cha-P1-CO(2)H). For this protein, S1 is a major hydrophobic pocket for binding. Binding of various residues to this pocket was investigated through free energy simulations and experimental inhibition constants. Several 300 ps lambda-dynamics simulations in explicit solvent were performed. The relative binding free energy was estimated from these simulations. From a single simulation, the inhibitors can be correctly classified into highly potent and weakly potent groups. The multiple simulations give an accurate rank ordering of inhibitor potency; computed and experimental binding free energies agree with 0.6 kcal/mol for five of the seven inhibitors. In addition, free energy perturbation (FEP) calculations were carried out to validate the results from lambda-dynamics. A total of 6 ligand pairs were compared. For each pair, 5-11 windows were used to map one ligand to the other. The cumulative simulation time was over 2 ns for each ligand pair. For four of the six ligand pairs, the lambda-dynamics free energy difference fits better than the FEP difference to the experimental value. The fact that the lambda-dynamics method achieved similar results in only a fraction of the total simulation time for FEP further demonstrates the robustness of the lambda-dynamics method.