Multidimensional computational strategies enhance the thermostability of alpha-galactosidase

Alpha-Galactosidase has significant industrial application value in food processing, animal nutrition and medical applications. Microbial-derived α-galactosidases predominate industrial implementation due to high productivity, yet their inherent thermal instability necessitates systematic protein engineering. In this study, we established a dual-strategy protein engineering framework to enhance the thermostability of Aspergillus tubingensis α-galactosidase (AtWU_04653). Strategy I employed integrative computational design tools (ABACUS2/PROSS/DBD2) for mutational library construction, which yielded the dominant mutant A169P exhibiting remarkable performance: 78.52 % enhancement in thermal half-life at 55 °C (pH 4.0) and 52.04 % increase in catalytic efficiency (k_cat /K_m). Strategy II implemented a physics-based computational methodology combining GROMACS molecular dynamics simulations with Rosetta unfolding free energy calculations and SPIRED machine learning predictions, successfully deriving three stabilized variants (E429I, N380L, T64P) displaying 57.33 %, 67.17 %, and 41.34 % extended half-lives respectively. Notably, E429I and T64P demonstrated concurrent 85.25 % and 65.90 % catalytic activity augmentation (k_cat /K_m). Both strategies achieved substantial reduction in experimental screening workload while enabling synergistic thermostability-activity optimization. This study uses sequence conservation analysis, unfolding free energy calculation, molecular dynamics simulation, and innovative protein prediction models to establish multidimensional computational strategies for designing mutants, providing new and important technical references for computational design and functional optimization of enzymes.