Résumé
Proteus mirabilis lipase (PML) features tolerance to organic solvents and great potential for biodiesel synthesis. However, the thermal stability of the enzyme needs to be improved before it can be used industrially. Various computational design strategies are emerging methods for the modification of enzyme thermal stability. In this paper, the complementary algorithm-based ABACUS, PROSS, and FoldX were employed for positive selection of PML mutations, and their pairwise intersections were further subjected to negative selection by PSSM and GREMLIN to narrow the mutation library. Thereby, 18 potential single-point mutants were screened out. According to experimental verification, 7 mutants had melting temperature (Tm) improved, and the ΔTm of K208G and G206D was the highest, which was 3.75 ℃ and 3.21 ℃, respectively. Five mutants with activity higher than the wild type (WT) were selected for combination by greedy accumulation. Finally, the Tm of the five-point combination mutant M10 increased by 10.63 ℃, and the relative activity was 140% that of the WT. K208G and G206D exhibited certain epistasis during the combination, which made a major contribution to the improvement of the thermal stability of M10. Molecular dynamics simulation indicated that new forces were generated at and around the mutation sites, and the rearrangement of forces near G206D/K208G might stabilize the Ca2+ binding site which played a key role in the stabilization of PML. This study provides an efficient and user-friendly computational design scheme for the thermal stability modification of natural enzymes and lays a foundation for the modification of PML and the expansion of its industrial applications.