Constrained versus unconstrained folding free-energy landscapes

Abstract

Protein folding can be described as a downhill process that brings the configuration of a chain of amino acids down to the bottom of a smooth free-energy funnel. Here, we use a recently developed coarse-grained protein model to assess the importance of frustration in the folding free-energy landscape. We compare the landscapes of natural proteins, computationally designed sequences, and structure-based potentials that force the contacts between the amino acids to adopt the native structure. Our results show that the structure-based potentials give a poor representation of the folding free-energy landscape, and that frustration is not just a perturbation over an otherwise perfect downhill folding.