Abstract:To enhance the catalytic activity and thermal stability of D-allulose 3-epimerase (DAEase) for industrial applications, high-performance mutants were engineered through integrating gene mining with semi-rational design. In this study, the DAEase gene from Microbacterium luteolum (MlDAEase) was heterologously expressed in Escherichia coli with soluble production, exhibiting optimal activity at 60 ℃ and pH 8.0. To address the poor thermal stability of MlDAEase that limits industrial applications, semi-rational design was performed guided by three-dimensional structural modeling and high-throughput screening, which identified the key mutants G65M, G65E, and M110W. The combinatorial mutant G65E/M110W, constructed via the CAST strategy, exhibited a 2.41-fold increase in relative activity and a 5.09-fold enhancement in catalytic efficiency compared to the wildtype. In industrial biocatalytic processes, the engineered mutant converted 500 g•L−1 D-fructose to 153 g•L−1 D-allulose, achieving a conversion rate of 30.60%, which represented a 2.09-fold improvement over the wild-type enzyme. The results demonstrated MlDAEase as a novel D-allulose 3-epimerase with great potential for engineering. Its catalytic performance and thermal stability were significantly improved through structure-guided semi-rational design. This study provides both a theoretical basis and a candidate biocatalyst for the efficient enzymatic production of D-allulose, offering promising prospects for industrial application.