Rosetta-Driven Rapid Evolution of meso-Diaminopimelate Dehydrogenases for Structurally Diverse D-Amino Acid Production

meso-Diaminopimelate dehydrogenase (DAPDH) holds considerable promise for D-amino acid synthesis but is limited by its low catalytic efficiency toward non-natural substrates. To address this, we present a Rosetta-driven evolution framework that enables the rapid tuning of DAPDH substrate specificity. This approach leverages Rosetta design to guide the construction of a smart-small mutant library, facilitating evolution in just one or two rounds with minimal screening (100-1,000 transformants), thus offering superior efficiency and cost-effectiveness. Using this strategy, the activities of two DAPDHs toward model substrates─benzoylformic acid (1b) and 3-indolepyruvic acid (10b)─were enhanced from undetectable levels to 29.5 and 5.1 U/mg, respectively. The activity toward 1b was nearly ten times higher than the best-reported mutant. The evolved DAPDHs displayed an expanded substrate spectrum and high catalytic performance, enabling the efficient synthesis of structurally diverse bulky D-amino acids at both analytical and preparative scales. The highest concentration of 1b achieved was 0.5 M (100 mL scale), with a space-time yield of 128.46 g/L/day. This study underscores the power of Rosetta-based evolution in accelerating enzyme engineering, offering a versatile and efficient platform for optimizing DAPDHs and other industrially significant enzymes.