Final Report Summary - PROTINT (Towards a quantitative framework for understanding protein-protein interactions: from specific effects to protein ecology) Protein molecules are some of the most important biomolecules: they build up our tissues and organs, catalyze important reactions in our cells and transmit signals at different levels. However, to be able to carry out their functions, proteins need to constantly interact with other proteins as well as other biomolecules in their environment. The main theme of the project has been analysis of the physicochemical basis of protein interactions in biologically realistic environments with particular focus on: 1) dynamical aspects of how proteins recognize their targets, and 2) general, non-specific aspects of protein interactions and localization in the cell. In the course of the project, we have made significant progress in addressing these questions using computational and theoretical biophysical techniques in close collaboration with experimentalists. In particular, by focusing on an important signaling protein ubiquitin, we have analyzed the role of correlated motions in protein recognition events and showed that they behave in a remarkably predictable way. In the course of this analysis, we have developed a powerful new software for the analysis of entropy in protein binding. Moreover, we have provided essential contributions to studying the mechanism of binding of several specific proteins involved in cell signaling, muscle contraction and processing of cellular waste. Second, we have analyzed and characterized important parallels in the general physicochemical properties of proteins and their binding partners, including other proteins and RNA molecules. This has further allowed us to make predictions about potential, still-to-be-discovered interaction partners for a number of proteins, including remarkably their own mRNAs. The latter finding has unexpectedly provided a potentially important element for explaining the origin of the universal genetic code. Furthermore, we have analyzed the behavior of proteins in crowded environments and have discovered an unexpected deficiency in the modern-day computational models used for simulating proteins in such environments. Finally, we have developed and verified a large set of computational parameters for treating post-translational modifications (PTMs) of proteins and have developed a public server for automated modification of known protein structures using PTMs of choice. Moreover, we have extended our analysis to modifications of nucleobases and characterized their impact on protein interactions. Overall, our efforts have provided several important conceptual and methodological advances for understanding protein interactions in realistic environments, with implications extending to both fundamental and applied biological science.
