A Method for the Incorporation of Protein Dynamics into Computational Enzyme Design using the Rosetta Software Suite

Abstract

The ability to rationally design enzymes remains a challenge in the field of protein engineering. Despite years of concerted effort, the activities of designed enzymes are typically orders of magnitude less efficient than their naturally occurring counterparts. The reason is likely because the link between protein dynamics and both catalytic ability and substrate specificity is not integrated, even in state-of-the-art enzyme design methods. Because of the computational expense required to simulate protein dynamics, it is often precluded from inclusion in the enzyme design process. However, the Ozkan laboratory at Arizona State University recently developed computational methods that enable a protein's unique dynamic profile to be assessed with minimal computational expense. This suggests that, for the first time, protein dynamics can be considered during the enzyme design process, representing a potential solution to previous challenges. We are exploring this hypothesis using the model enzyme TEM-1 β-lactamase. The dynamics profiles of TEM-1 and one of its ancestral homologues were generated, allowing us to identify residues directly coupled to the active site through extended dynamic networks. Despite high structural similarity between TEM-1 and its ancestral homolog, their substrate profiles are vastly different. We hypothesize that differences in dynamics determine the substrate profiles of these proteins. To test this, the Rosetta computational protein design software was used to re-engineer the environments surrounding these residues and the dynamic profiles of the mutants were re-analyzed. Experimental characterization of five variants revealed two mutants whose substrate preferences more closely resembled those of the ancestral variant. In addition to providing a better understanding of the relationship between catalysis, specificity, and dynamics in enzymes, our novel computational methods could be used to design more efficient enzymes with far-reaching applications in numerous fields.