Co-Evolution of the Activity and Thermostability of (R)-Transaminase AcTA by Data-Driven Exploration of Combinatorial...

Optically active (R)-amines are pivotal building blocks of a variety of agrochemicals and food-related bioactive compounds. To construct a biocatalytic route for the efficient synthesis of (R)-amines, a high-performance (R)-TA, AcTA from Aspergillus caelatus, was identified through computation-guided virtual screening. A data-driven exploration of combinatorial mutagenesis (DECM) framework was developed by leveraging the advantages of directed evolution and machine learning. The DECM framework enables the exploration of the global combinatorial mutagenesis landscape for the co-evolution of activity and thermostability. A total of 25 combinatorial variants were experimentally tested, of which 18 were found to exhibit improved performance over the wild type. M3 (H113F/V145C/T272S) was identified with a 5.6-fold increase in k_cat/K_M and a 10.2-fold longer half-life at 40 °C, enabling >98.0% conversion at 45 °C. MD simulations revealed structural evidence for enhanced activity and thermostability of M3. These findings demonstrate the effectiveness of DECM in exploring combinatorial mutations and accelerating directed evolution.