SMBp Group Meeting: Doug Renfrew (CCB)

Thursday Jan 13, 2022, 10:00 AM → 11:00 AM America/New_York

3rd Floor Conference Room/3-Flatiron Institute (162 5th Avenue)

Title: Incorporating Molecular-Dynamics-Guided Explicit Negative-Design into a Multistate Computational Enzyme Design Protocol

Abstract: I plan to present a short talk for a proposal for a collaboration between our group with SMBp that centers around some of our complementary expertise in protein modeling that I hope will generate a lot of useful discussion.

Several enzymes have been designed fully in silico. The computationally designed sequences have lower catalytic efficiency compared to many naturally occurring enzymes and their experimental validation involved testing thousands of potential designs. The catalytic efficiency of these enzymes have been improved by experimental directed evolution. It has also been observed that for some enzyme design tasks, comparable results can be achieved with directed evolution alone, and consequently less effort has been put toward improving computational techniques. Directed evolution is applicable when your reactants and products are well behaved and their degradation or formation can be easily observed (or linked to an observable) which limits the reactions and catalysts it can be used on. Improvements to computational design are needed to address the yield and catalytic efficiency of designed sequences.

I hypothesize that computational techniques could be improved by explicitly taking into account protein dynamics. I propose using advanced sampling molecular dynamics to generate (A) clustered ensembles of conformations that span the full breadth of conformations accessible to the protein, and (B) clustered ensembles of conformations from simulations that have been driven along the reaction coordinate. Using a multistate sequence design protocol that chooses sequences that favor conformations on the reaction coordinate (B) and disfavor conformations that are off the reaction coordinate (A-B), I believe that we can improve yields and catalytic efficiency of designed sequences.

If successful, these techniques could be applied to enzymes or reactions that are not amenable to directed evolution including non-biological polymer catalysts.