Statistical Characterization of Food-Derived α-Amylase Inhibitory Peptides: Computer Simulation and Partial Least Squares Regression Analysis

Abstract

α-Amylase inhibitory peptides are used to treat diabetes, but few studies have statistically characterized their interaction with α-amylase. This study performed the molecular docking of α-amylase with inhibitory peptides from published papers. The key sites, side chain chargeability, and hydrogen bond distribution characteristics were analyzed. Molecular dynamics simulated the role of key sites in complex stability. Moreover, partial least squares regression (PLSR) was used to analyze the contribution of different amino acids in the peptides to inhibition. The results showed that, for the α-amylase molecule, His201 and Gln63, with the highest interaction numbers (INs, 15, 15) and hydrogen bond values (HBVs, 11.50, 10.33), are the key sites on α-amylase, and amino acids with positively charged side chains were important for inhibitory activity. For the inhibitory peptides, Asp and Arg had the highest HBVs, and amino acids with charged side chains were more likely to form hydrogen bonds and exert inhibitory activity. In molecular dynamics simulations, peptides involving key binding sites formed more stable complexes with α-amylase than α-amylase alone, suggesting enhanced inhibitory effects. Further, PLSR results showed that amino acids close to the N-terminus of the inhibitory peptide, located in the third and fifth positions, were significantly correlated with its inhibitory activity. In conclusion, this study provides a new approach to developing and screening α-amylase inhibitors.