(507a) The Role of Complementary Shape in Protein Dimerization

Proteins bind to other proteins to form complexes with biological function. Predicting the structure of these complexes and understanding their assembly mechanisms are of fundamental importance for protein assembly, rational drug design, and the engineering of proteins as building blocks for new materials. Understanding the role of the various forces contributing to protein-protein interactions is particularly critical in this regard. Shape complementarity is one such contribution known to play a role in protein binding. Using molecular simulation, we investigated the binding of 50 protein dimers with atomic-level resolution of rigid molecular shape and a generic entropic depletion interaction to isolate the role of shape in the assembly of these dimers, and compared the assembled configurations with known native structures. Our simulations predict the yield of native contacts in thermodynamic equilibrium or equivalently, the dissociation constant. We find that shape complementarity is sufficient to predict native complexes as equilibrium assemblies in six cases. We use machine learning classifiers to assess the quality of the prediction by analyzing the importance of competing binding configurations and how they affect the assembly, resulting in a precision of 89.14% and recall of 77.11%. This suggests the power of shape complementarity to predict native protein interfaces in certain instances.