Identifying and Overcoming the Sampling Challenges in Relative Binding Free Energy Calculations of a Model Protein:Protein Complex

Relativealchemical binding free energy calculations are routinelyused in drug discovery projects to optimize the affinity of smallmolecules for their drug targets. Alchemical methods can also be usedto estimate the impact of amino acid mutations on protein:proteinbinding affinities, but these calculations can involve sampling challengesdue to the complex networks of protein and water interactions frequentlypresent in protein:protein interfaces. We investigate these challengesby extending a graphics processing unit (GPU)-accelerated open-sourcerelative free energy calculation package (Perses) to predict the impactof amino acid mutations on protein:protein binding. Using the well-characterizedmodel system barnase:barstar, we describe analyses for identifyingand characterizing sampling problems in protein:protein relative freeenergy calculations. We find that mutations with sampling problemsoften involve charge-changes, and inadequate sampling can be attributedto slow degrees of freedom that are mutation-specific. We also explorethe accuracy and efficiency of current state-of-the-art approaches Ralchemicalreplica exchange and alchemical replica exchange with solute tempering Rforovercoming relevant sampling problems. By employing sufficiently longsimulations, we achieve accurate predictions (RMSE 1.61, 95% CI: [1.12,2.11] kcal/mol), with 86% of estimates within 1 kcal/mol of the experimentallydetermined relative binding free energies and 100% of predictionscorrectly classifying the sign of the changes in binding free energies.Ultimately, we provide a model workflow for applying protein mutationfree energy calculations to protein:protein complexes, and importantly,catalog the sampling challenges associated with these types of alchemicaltransformations. Our free open-source package (Perses) is based onOpenMM and is available at https://github.com/choderalab/perses.