文摘
The directed evolution of enzymes promises to eliminate the long-standing limitations of biocatalysis in organic chemistry and biotechnology—the often-observed limited substrate scope, insufficient activity, and poor regioselectivity or stereoselectivity. Saturation mutagenesis at sites lining the binding pocket with formation of focused libraries has emerged as the technique of choice, but choosing the optimal size of the randomization site and reduced amino acid alphabet for minimizing the labor-determining screening effort remains a challenge. Here, we introduce structure-guided triple-code saturation mutagenesis (TCSM) by encoding three rationally chosen amino acids as building blocks in the randomization of large multiresidue sites. In contrast to conventional NNK codon degeneracy encoding all 20 canonical amino acids and requiring the screening of more than 1015 transformants for 95% library coverage, TCSM requires only small libraries not exceeding 200–800 transformants in one library. The triple code utilizes structural (X-ray) and consensus-derived sequence data, and is therefore designed to match the steric and electrostatic characteristics of the particular enzyme. Using this approach, limonene epoxide hydrolase has been successfully engineered as stereoselective catalysts in the hydrolytic desymmetrization of meso-type epoxides with formation of either (R,R)- or (S,S)-configurated diols on an optional basis and kinetic resolution of chiral substrates. Crystal structures and docking computations support the source of notably enhanced and inverted enantioselectivity.