Evolutionary Adaptability Coupled with Computation-Driven Engineering for Thermostability Enhancement of a Deoxynivalenol-Detoxifying Fusion Enzyme.

Deoxynivalenol (DON), a trichothecene mycotoxin commonly found in cereal grains and their derived products, poses significant risks to human and animal health. In previous work, a fusion enzyme composed of the dehydrogenase DADH and the aldo-keto reductase AKR13B3 was engineered to convert DON into the non-toxic 3-epi-DON in a single step. However, the poor thermal stability of this fusion enzyme limited its industrial application. In this study, EVcouplings and the GRAPE-WEB platform were utilized to identify key amino acid residues governing the thermal stability of the fusion enzyme AKR13B3-DADH. Through single-point mutation screening and the combination of beneficial mutation sites, a triple mutant M361L/T508Y/Y603F (M1) was obtained. The half-life of M1 at 50 °C reached about 500 min, representing a 16.4-fold increase compared with the wild type, while its catalytic activity increased by 2.7-fold. The apparent melting temperature increased by approximately 4 °C. Molecular dynamics simulations verified that the improved thermostability results from reduced conformational flexibility in key regions, enhanced structural packing, and a strengthened hydrogen bond network. These results demonstrate the successful development of a thermostable DON-detoxifying fusion enzyme and provide a practical basis for its industrial application.