Protein design at room temperature: the role of side-chain conformational entropy

New detailed models of the protein structures by means of a physical description at the atomic level have improved the possibilities to treat de novo computational protein design. The existing methods mostly rely on combinatorial optimization using a scoring function that estimates the folding free energy of a protein sequence, in its optimal side-chain configuration, on a given main-chain structure. While the solvation entropy term is often taken implicitly, the conformational entropy stemming from alternative side-chain arrangements is usually omitted (or not properly evaluated) since its computation is generally intractable. A method recently proposed by the authors [1] incorporates such conformational entropy based on statistical mechanics principles. In this work we further test the protein design methodology, that we applied to the complete redesign of 27 proteins, and study how the entropy affects the ranking of the same sets of sequences at low and high temperatures. We also investigate how the new methodology affects the fraction of aminoacids of each kind that are found in solvent-protected positions. Our results indicate that accounting for entropic contribution in the score function affects the outcome in a highly non-trivial way, and might improve current computational design techniques based on protein stability. Indeed, ranking at low, and high temperatures are, in general, weakly correlated, pointing out the importance of accounting for the entropy. We also notice that the free-energy driven design yields sequences that differ in many positions from those obtained with the standard design, while the burial fraction for the aminoacids does not change much.