On evolutionary conservation of thermodynamic coupling in proteins.

The inherent complexity of thermodynamic coupling in proteins presents a major challenge in understanding and engineering protein function. Recent work has argued that the study of proteins can be simplified by the use of correlated mutations in the evolutionary record to locate a small subset of thermodynamically coupled residues that participate in functionally important, evolutionarily conserved energetic pathways. To test this hypothesis, we examined the predictions of correlated mutation algorithms for a number of proteins for which coupling between residues has been determined by analysis of double mutant cycles. We find that correlated mutation algorithms can find residue pairs that are physically close and that physically close residue pairs tend to be thermodynamically coupled. We find little evidence, however, for the hypothesis that thermodynamic coupling is limited to the subset of evolutionarily constrained residue positions.