Cross-region combinatorial mutagenesis reveals context-dependent enhancement of 3α-hydroxysteroid dehydrogenase catalytic efficiency.

This study focused on enhancing the catalytic efficiency of 3α-hydroxysteroid dehydrogenase (3α-HSD) from Comamonas testosteroni toward androsterone. By employing enzyme engineering strategy combining semi-rational design with the protein language model Saprot, a high-performance mutant, H119A/R201L/R216L/N152F/K208A, was obtained. First, ethanol was used as a cosolvent to stabilize the reaction system, and the optimal ratio of androsterone solution in the reaction system was determined, thereby optimizing the dissolution conditions for subsequent mutant screening. Based on Caver channel analysis and Saprot-PLM intelligent screening, the K208A and N152F mutants exhibited enzyme activities 1.27-fold and 1.23-fold higher than the wild-type, respectively. Strikingly, combining the positive-effect mutation N152F with the negative-effect scaffold H1 revealed context-dependent epistasis, enzyme activity was increased to 153% of the wild-type. Furthermore, the five-point combination mutant H119A/R201L/R216L/N152F/K208A, incorporating K208A and R201L, exhibited a specific activity of 176.3 U/mg, with kcat/Km 3.54-fold of the WT. Enzymatic characterization revealed that both the optimal temperature and pH increased, and thermal stability significantly enhanced. Hydrogen bonding analysis and MM/PBSA calculations revealed that N152F reconfigures the electrostatic microenvironment of the active site via the aromatic ring of phenylalanine, reversing the electrostatic interaction from a repulsive state to a strongly attractive state, while simultaneously enhancing protein conformational flexibility to achieve a highly efficient catalytic mechanism characterized by "moderate binding and dynamic precision." This strategy yields an efficient biocatalyst for steroid biomanufacturing and provides a new perspective for enzyme design.