Modeling of Biomolecular Systems Interactions, Dynamics, and Allostery: Bridging Experiments and Computations - September 10-14, 2014, Istanbul, Turkey

Modeling of Biomolecular Systems Interactions, Dynamics, and Allostery Poster Session II

59-POS Board 12 Are There any Structural Taboos in Protein-Protein Interactions? Insights from a Genome-Wide Study of Interacting and Non Interacting Protein Pairs. Juliette Martin , Nicoletta Ceres, Guillaume Launay, Richard Lavery. CNRS , Lyon, France. Functional interactions take place between specific proteins in a healthy cell. Yet, the density of a cell in macromolecules is such that proteins are constantly in contact or in close proximity with other proteins that are not functional partners. How is the specificity of functional interactions maintained? In particular, is the ability to interact encrypted in the protein's 3D structure? To address this question we have performed an in silico large-scale experiment to explore the structural repertoire of interacting and non-interacting proteins pairs, and to see whether they differ. We considered yeast cerevisiae , for which 3D structural models and protein-protein interaction data are available, as well as four different data sets of nominally “negative interactions”. Positive interactions, i.e. experimentally known interactions, were analyzed in parallel to serve as positive control. For each interaction (positive or negative), we considered the 3D structures of the proteins involved, and tried to determine whether the PDB contains a protein-protein complex where similar structures interact. When a complex was found in the PDB, we qualified the interaction as “supported” by the PDB. We find that negative interactions selected on the basis of different sub-cellular localizations are supported by PDB protein-protein complexes to the same extent as positive interactions. This finding stresses the importance of physical sequestration to maintain interaction specificity. Finally, we have analyzed the structural features of the resulting 3D models of binary protein complexes and also their stability using our coarse-grain model PaLaCe. We found that the more sophisticated PaLaCe energy evaluation was necessary to distinguish interacting and non- interacting protein pairs. Interestingly, some nominally non-interacting pairs are nevertheless predicted to form stable complexes.

114