Dissecting the Contribution of Kinase Conformational Reorganization Energies to Inhibitor Selectivity

Abstract

Kinases are an important target for many cancer therapies. However, they can be difficult to selectively target without hitting one of the other ∼500 kinases in the human genome. Recent experimental and computational evidence suggests that the cost of confining the kinase to the binding-competent conformation could play a large role in determining inhibitor selectivity, specifically in the case of Src and Abl kinase and the ligand imatinib (also known as the paradigm shifting targeted cancer drug ‘Gleevec’). To test this hypothesis on a larger scale, and to dissect the contribution of this reorganization energy to kinase inhibitor affinity and selectivity, we have developed an approach that combines massively distributed molecular simulations with automated experimental biophysical measurements. Atomistic molecular simulations are run on the Folding@home worldwide distributed computing platform to collect ∼ms of simulation data. Markov state models are then constructed of this data to determine the accessible conformations and relative free energies. Alchemical binding free energy calculations are used to assess the ligand binding free energies of selective kinase inhibitors to individual kinetically metastable conformations, and automated biophysical experiments using a novel fluorescence-based assay are used to directly measure binding affinities of kinase inhibitors to a variety of kinases. By analyzing this data together, we believe this work provides a more global picture of the role that kinase conformational reorganization energies play in selectivity, pushing us toward a better understanding of the requirements for selective inhibitor design.