Selecting Better Biocatalysts by Complementing Recoded Bacteria**

Abstract

In vivo selections are powerful tools for the directed evolution of enzymes. However, the need to link enzymatic activity to cellular survival makes selections for enzymes that do not fulfill a metabolic function challenging. Here, we present an in vivo selection strategy that leverages recoded organisms addicted to non‐canonical amino acids (ncAAs) to evolve biocatalysts that can provide these building blocks from synthetic precursors. We exemplify our platform by engineering carbamoylases that display catalytic efficiencies more than five orders of magnitude higher than those observed for the wild‐type enzyme for ncAA‐precursors. As growth rates of bacteria under selective conditions correlate with enzymatic activities, we were able to elicit improved variants from populations by performing serial passaging. By requiring minimal human intervention and no specialized equipment, we surmise that our strategy will become a versatile tool for the in vivo directed evolution of diverse biocatalysts.

A Robust Growth-Based Selection Platform to Evolve an Enzyme via Dependency on Noncanonical Tyrosine Analogues

“…These are (1) cam-3nY, which was employed throughout the selection campaigns, (2) cam-3-L-iodo-tyrosine (cam-3iY), an alternative m-substituted tyrosine analogue, and (3) cam-L-tyrosine (cam-Y). Consistent with previous reports, 16,32 SmLcar displays poor activity for all three substrates, but has a strong preference for cam-Y (v obs = 0.71 min −1 ) over cam-3nY (0.28 h −1 ) and cam-3iY (0.12 h −1 ). Excitingly, all 10 SmLcar variants from our panel outperform the wild-type enzyme for all substrates�typically by 2−4 orders of magnitude depending on the carbamoylated precursor.…”

A robust life-or-death selection platform for enzyme evolution

“…These are (1) cam-3nY, which was employed throughout the selection campaigns, (2) cam-3-L-iodo-tyrosine (cam-3iY), an alternative m-substituted tyrosine analog, and (3) cam-L-tyrosine (cam-Y). Consistent with previous reports, 16,31 SmLcar displays poor activity for all three substrates, but has a strong preference for cam-Y (vobs = 0.71 min -1 ) over cam-3nY (0.28 h -1 ) and cam-3iY (0.12 h -1 ).…”

“…2A). 16 As proof-of-principle, we applied our platform to engineer a L-N-carbamoylase from Sinorhizobium meliloti strain CECT 4114, SmLcar, 31 to promote the hydrolysis of the N-carbamoylated ncAA, cam-L-3-nitro-tyrosine (cam-3nY, Fig. 2B).…”

“…This finding is consistent with our previous evolutionary campaign, where we demonstrated that enlarging the binding pocket via a L217G substitution is critical to accommodate cam-3nY in a productive conformation, in which the aromatic side chain is flipped by ~90°. 16 (3) The substitutions at Gly327 and Phe329 reveal two distinct solutions, which are able to drastically boost the growth rates of recoded bacteria (Fig. 4D).…”

“…1D). [13][14][15][16] Lastly, previously employed selections for biocatalysts that directly confer a growth advantage to a host organism (e.g. chorismate mutase or β-lactamase) could be employed to identify enzyme variants with a wide range of activities.…”