In Silico Enzymolysis-Guided Mining of Aminopeptidases with Molecular Insights into Their Substrate Specificity Mechanism

The palatability of protein-based food heavily relies on aminopeptidases with the ability for bitter peptides degradation. However, there is a lack of methods for rational mining of aminopeptidases toward different proteins as well as the evolution direction for substrate specificity remaining unclear. In this study, an in silico simulated enzymolysis-based method for aminopeptidases mining was developed with Crassostrea gigas protein as a model. Results indicated that Ala and Ile were the most frequently exposed hydrophobic amino acids, causing a bitter taste in C. gigas hydrolysates. Furthermore, an aminopeptidase APs1 with putative Ala specificity was heterologously expressed and characterized with high enzyme activity toward Ala (4.92 ± 0.136 U/mg) and Arg (3.50 ± 0.178 U/mg). Site-saturation mutation and molecular docking results revealed that changes in steric hindrance and salt bridge formation within the active site contribute to enhanced catalytic efficiency. Among the mutants, APs1­(F316M) showed significantly improved activity toward tested hydrophobic amino acids, especially the activity toward Ala was increased to 14.50 ± 0.137 U/mg. This study presents a directional approach to aminopeptidase mining and evolution, contributing to the rapid selection and combination of protein-degrading enzymes for food quality improvement.