Computational Alanine Scanning with Interaction Entropy for Protein-Ligand Binding Free Energies

In protein ligand binding, only a few residues contribute significantly to the ligand binding. Quantitative characterization of binding free energies of specific residues in protein ligand binding is extremely useful in our understanding of drug resistance and rational drug design. In this paper, we present an alanine scanning approach combined with an efficient interaction entropy method to compute residue-specific protein ligand binding free energies in protein drug binding. In the current approach, the entropic components in the free energies of all residues binding to the ligand are explicitly computed from just a single trajectory MD simulation by using the interaction entropy method. In this approach the entropic contribution to binding free energy is determined from fluctuations of individual residue ligand interaction energies contained in the MD trajectory. The calculated residue-specific binding free energies give relative values between those for ligand binding to the wild type protein and those to the mutants when specific results mutated to alanine. Computational study for the binding of two classes of drugs (first and second generation drugs) to target protein ALK and its mutant was performed. Important or hot spot residues with large contributions to the total binding energy are quantitatively characterized and the mutation effect for the loss of binding affinity for the first generation drug is explained. Finally, it is very interesting to note that the sum of those individual residue-specific binding free energies are in quite good agreement with the experimentally measured total binding free energies for this protein ligand system.