B M ] 2 8 Ja n 20 04 Stability and the Evolvability of Function in a Model Protein Running Title : Evolving Stability and Function

Functional proteins must fold with some minimal stability to a structure that can perform a biochemical task. Here we use a simple model to investigate the relationship between the stability requirement and the capacity of a protein to evolve the function of binding to a ligand. Although our model contains no built-in tradeoff between stability and function, proteins evolved function more efficiently when the stability requirement was relaxed. Proteins with both high stability and high function evolved more efficiently when the stability requirement was gradually increased than when there was constant selection for high stability. These results show that in our model, the evolution of function is enhanced by allowing proteins to explore sequences corresponding to marginally stable structures, and that it is easier to improve stability while maintaining high function than to improve function while maintaining high stability. Our model also demonstrates that even in the absence of a fundamental biophysical tradeoff between stability and function, the speed with which function can evolve is limited by the stability requirement imposed on the protein. For nearly all proteins found in nature, there is a unique mapping from the linear protein sequence to a thermodynamically stable three-dimensional native structure, with the mapping determined by the laws of physics (Anfinsen, 1973). However, this unique mapping from sequence to conformation is not a general property of polypeptide sequences, since most randomly generated sequences do not have stable folded structures (Keefe and Szostak, 2001; Davidson et al., 1995). In other words, natural protein sequences exist in the space of foldable sequences, which is but a small subset of the space of all possible sequences. Therefore, evolution must have acted heavily on natural proteins in order to select those with stable native structures. Although natural proteins possess stable native structures, the evolutionary fitness of a protein depends not on the stability of the native structure per se, but rather on the stability of this structure being appropriate to allow the protein to perform a function such as catalyzing a chemical reaction or binding to a ligand. Stability is therefore under selection only insofar as it is necessary for biochemical function, and most natural proteins are only marginally stable at their physiologically relevant temperatures (Fersht, 2002). In protein mutagenesis studies, stability and function can appear to be competing properties, with mutations that increase stability often reducing function (Shoichet et al., 1995; Schreiber et al., 1994), and mutations that improve or alter function often decreasing stability (Wang et al., 2002). However, several lines of evidence demonstrate that high stability and high functionality are not inherently incompatible. In Nature, there is a strong correlation between the temperature of an organism’s environment and the stability of its proteins, indicating that natural evolution is able to create functional and highly