Because enzymes are selective, efficient, and biodegradable, the ability to design enzymes to catalyze organic transformations has the potential to transform classical organic synthesis. However, even the most successful engineered enzymes lag far behind naturally evolved enzymes in their catalytic activity. In fact, enzymes designed in silico are typically subjected to directed evolution experiments to introduce beneficial mutations. While the in silico design process typically focuses on the region around the active site, many of the beneficial mutations introduced by directed evolution are far from the active site. Thus, the current in silico design process misses critical aspects of catalysis that are identified and exploited by directed evolution. The goal of this proposal is to develop an iterative in silico procedure to detect and evaluate mutations that enhance catalytic activity, whether they are near or far from the active site. First, molecular dynamics simulations will be developed as a tool to screen potential mutants, in particular by using Markov State Models to accelerate the simulations and make statistical predictions about whether the optimal catalytic geometry is likely to persist for long time periods. Second, bioinformatics tools will be incorporated into the design process to identify amino acid residues that are not highly conserved by evolution, which can then be targeted as mutation hot spots in the design procedure. Third, high-throughput molecular flexibility analysis will be incorporated into the design process to eliminate enzyme candidates with undesired flexibilities near the active site and to identify other flexible regions where mutations can enhance catalysis. The goal of these innovations is to transform the directed evolution process as much as possible from a random in vitro process into a rational in silico process that identifies beneficial mutations both near and far from the active site. Finally, these innovations will be applied to two related projects in the design of enzymes known as transaminases. First, these methods will be used to understand the famously successful engineering of a sitagliptinase enzyme that catalyzes the final step in the synthesis of the type 2 diabetes drug sitagliptin. The goal is to show that in silico methods can successfully explain the roles of mutations both near and far from the active site developed over many years by a tortuous directed evolution process. Second, these methods will be used to engineer a new transaminase enzyme to synthesize the anti-depression drug (R)-3-phenylGABA. The best natural enzyme for this synthesis is plagued by the fact that it produces significant amounts of the undesired (S)-enantiomer. The goal is to show that the in silico methods developed in this proposal can identify beneficial mutations both near and far from the active site to design a new enzyme that increases the enantiomeric excess of the desired (R)-enantiomer from a moderate 68% to a synthetically-useful level of 98%.
The efficient and selective synthesis of drugs is essential to further progress in public health. To meet this goal, one promising approach is to engineer enzymes that catalyze new reactions. This proposal outlines new methods to overcome key limitations in the current enzyme design procedure, and applies them to improve the design of enzymes used in the synthesis of a type 2 diabetes drug and an anti-depression drug.
| Saleem-Batcha, Raspudin; Stull, Frederick; Sanders, Jacob N et al. (2018) Enzymatic control of dioxygen binding and functionalization of the flavin cofactor. Proc Natl Acad Sci U S A 115:4909-4914 |
| Newmister, Sean A; Li, Shasha; Garcia-Borràs, Marc et al. (2018) Structural basis of the Cope rearrangement and cyclization in hapalindole biogenesis. Nat Chem Biol 14:345-351 |
