From Structures to Affinities: Harnessing Hydration-centric Molecular Simulations for Modulating Protein Interactions

University Of Pennsylvania, Philadelphia PA

Investigators

Abstract

Project Summary The quest for modulators of protein interactions is undergoing a renaissance with advent of powerful computa- tional methods for the prediction and de novo design of protein structures. However, translating this static struc- tural information into an accurate prediction of the thermodynamics and kinetics of dynamic protein interactions remains a challenge, limiting our ability to robustly design tight-binding modulators of protein interactions. The limitations of existing structure-based drug design approaches can be traced back to their accounting of collective hydration effects using additive approximations, such as hydropathy scales, and their tendency to adopt a static perspective, assuming a relatively rigid protein. Thus, there is a critical need for developing novel computational methods capable of accurately accounting for the collective reorganization of the hydrogen bonding network of water as well as any conformational changes that accompany binding. To address this critical knowledge gap, the proposed program seeks to develop novel hydration-based, enhanced-sampling molecular simulation techniques for characterizing the reorganization of water structure in response to conformational change and binding, and to use such a characterization for informing the thermodynamics and kinetics of protein interactions as well as designing interaction modulators. These techniques will be developed within the context of three judiciously cho- sen projects spanning a diversity of protein interactions: (i) interactions between reader proteins and methylated histone tails, (ii) interactions between galectins and carbohydrates, and (iii) interactions between the p38! MAP kinase and single-domain antibodies. For each project, close interactions with experimental collaborators will facilitate the refinement of our methods and the validation of our predictions, enabling us to uncover the molec- ular underpinnings of specificity. By developing novel computational methods to bridge the gap from protein structures to affinities, and leveraging these methods to characterize the interactions of proteins with peptides, carbohydrates, and other proteins, the proposed program thus seeks to transform our ability to predict, under- stand and modulate protein interactions.