Computationally designed libraries expand the functional scope of enzymes

Conference Dates

September 24-28, 2017


The evolution of altered or improved function is constrained by epistatic relationships among the spatially close positions that make up an active site. A further constraint is due to stability-function tradeoffs, whereby mutations accumulated along an evolutionary trajectory increasingly reduce stability and functional expression. To address these problems, we have developed a new method that uses bioinformatics and Rosetta atomistic simulations to substantially stabilize enzymes without altering their functions1. Applying the algorithm to several well-known “tough-nut” cases, such as human acetylcholinesterase and a malaria vaccine immunogen, resulted in three orders of magnitude improvement in bacterial expression and 10-20oC higher thermal resistance with no change in activity. We have further expanded this methodology to introduce combinations of mutations at the active site without impacting the core catalytic residues. Applied to the enzyme phosphotriesterase (PTE), we designed 50 active-site variants, all of which expressed in E. coli as well as the wild-type, and were active. Furthermore, the activity profile across the designs was substantially altered relative to wild-type, with up to 1000-fold enhancement of promiscuous activities and expanded enantioselectivity. Our design algorithms can therefore address the most significant difficulty in the evolution of altered function, how to find tolerated mutations among epistatically related positions without succumbing to the stability-function tradeoffs. The methods are fully automated, requiring only a structure and a sequence alignment of homologues, and can be applied to proteins, whether or not they are amenable to high-throughput screening. References Goldenzweig A, Goldsmith M et al. Molec Cell 2016, 63:337-346.

This document is currently not available here.