Accurate prediction of binding affinities of drug candidates to their targets remains challenging because of protein flexibility. When the binding free energy (delta-G) of twenty-four ATP-competitive inhibitors toward CDK2 and ERK2 kinases was assessed using an alchemical free energy perturbation method, MP-CAFEE, where the free energy simulation starts from a single equilibrated conformation, large calculation errors of 2-3 kcal/mol were observed. Here, we have developed a new delta-G computation method, where the starting structure was set to multiple conformations to cover conformational flexibility in the protein-ligand interaction, resulting in significant decreases in the calculation errors. MP-CAFEE and the newly-developed delta-G computation method have been applied to prediction of mutation-induced drug sensitivity changes in oncogenic kinases such as ALK, RET, and EGFR. In cases of drug resistance mutations in ALK and EGFR, most of which occur near the drug binding site (&lt; 15 angstroms), decreases in the efficacy tend to be successfully quantified by single free energy simulations based on short MDs (~50ns), characterizing conformational differences in the drug binding site between wild-type and mutant proteins. In contrast, in case of a RET mutation (S904F), which occurs on the activation loop away from the drug binding site (~26 angstroms), multiple free energy simulations based on microsecond MDs successfully characterize a drug resistant conformer induced allosterically by the mutation, supporting the results of crystallographic and thermal analyses of the RET mutant. Structure information on protein-drug interactions, captured by these simulations, is expected to be useful for drug design to overcome the resistance.

Mutation-induced drug sensitivity prediction using massively-parallel molecular dynamics free energy simulations

ACS Spring Meeting 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.