Rapid and Accurate Binding Free Energy Prediction for Inhibitor-Bound HIV-1 Enzymes

Abstract

Medical practitioners have limited ways of matching a drug to the unique genetic profile of a virus population as it mutates within a patient under drug-related selective pressure. Currently, knowledge based decision support software based on existing clinical records and associated viral genotypic data is used to aid inhibitor selection. In the instance of the emergence of drug resistance and associated treatment failure, the ineffective treatment may be minimized by selection of the next most appropriate drug regimen. The latest generation of petascale computational resources offer the potential to enhance these systems by using predictive modelling to explain and quantify the effects of resistance mutations. We show here that it is possible to quantitatively predict the differences in strength of inhibitors binding to wildtype and mutant HIV-1 proteases using the established MM-PBSA free energy calculation methodology. Excellent agreement between simulation and experimental results have been achieved for both absolute and relative binding affinities in a series of resistant HIV-1 protease mutants bound to the inhibitor lopinavir using an ensembles of 50 simulations. By utilising ensembles of short simulations we achieve both efficient sampling of phase space and reduced turn around times. This combination allows simulations to be performed on a timescale relevant to medical practitioners. Preliminary results indicate that our methodology is also be applicable to other drug and enzyme combinations. Our studies are facilitated by the Binding Affinity Calculator (BAC), which performs the rapid and automated construction, deployment, implementation and post processing of simulations across multiple supercomputing grid-based resources. BAC has been integrated with the ViroLab Virtual Laboratory suite of decision support and research tools. This provides a user friendly interface designed to encourage users outside the existing academic research community to perform molecular level simulations.