Engineering a Non-Natural Photoenzyme for Improved Photon Efficiency.

Abstract

Photoenzymes are biological catalysts that use light to convert starting materials into products. These catalysts require photon absorption for each turnover, making quantum efficiency an important optimization parameter. Flavin-dependent "ene"-reductases (EREDs) display latent photoenzymatic activity for synthetically valuable hydroalkylations; however, protein engineering has not been used to optimize this non-natural function. We describe a protein engineering platform for the high throughput optimization of photoenzymes. A single round of engineering results in improved catalytic function toward the synthesis of γ, δ, ϵ-lactams, and acyclic amides. Mechanistic studies show that key mutations can alter the enzyme's excited state dynamics, enhance its photon efficiency, and ultimately increase catalyst performance. Transient absorption spectroscopy reveals that engineered variants display dramatically decreased radical lifetimes, indicating an evolution toward a concerted mechanism.