Re-Design of Protein Interfaces: Co-Optimization of Packing and Electrostatic Interactions

The Problem: A number of powerful optimization techniques have been developed to deal with different aspects of molecular design problems that arise in biology and biological engineering, including the design of protein folds and protein binding interfaces. The two key examples are algorithms for optimizing packing interactions and a separate approach for optimizing electrostatic interactions. Packing algorithms select favorable arrangements of chemical groups whose shapes are complementary to one another; the approaches are generally founded on discrete-space sampling using algorithms such as dead-end elimination [2], the A* algorithm [5], self-consistent mean field theory [4], simulated annealing [6], genetic algorithms [1, 10], and combinatorial search [10]. Electrostatic optimization finds complementary distributions of polar and charged chemical groups that lead to tight binding based on a continuum electrostatic treatment of the effects of solvent and dielectric screening [3, 7]. While each of these approaches is extremely effective in its domain of applicability, packing optimization approaches are inconsistent with electrostatic optimization techniques. Thus, there is currently no good approach to co-optimize both sets of interactions simultaneously. Our goal is to merge these two powerful approaches through the development of a single methodology that performs simultaneous optimization of packing and electrostatic contributions.