Structure-guided discovery of natural α-glucosidase inhibitors from food sources for blood sugar regulation

Type II diabetes mellitus (T2DM) has become a global health crisis, with prevalence rates escalating dramatically over the past three decades. α-glucosidase inhibitors have emerged as effective substances for glycemic control . In this study, a library of α-glucosidase inhibitors with definite inhibitory action was constructed, and models for the quantitative structure-activity relationship (QSAR) were developed by combining four descriptors—MOE, ChemoPy, Mordred, and RDKit, and three machine learning algorithms—RF, LDA, and SVM . Molecular docking revealed that hydrogen bonding and van der Waals interactions critically stabilize inhibitor-enzyme binding, while molecular dynamics (MD) simulations (100 ns) confirmed the structural stability of the complexes. From FooDB, amentoflavone emerged as a potent natural inhibitor with an IC 50 of 10.77 ± 1.03 μM. Notably, amentoflavone’s minimal side effects and natural origin highlight its potential as a safer therapeutic alternative. Our multi-faceted approach combining QSAR, molecular docking, MD dynamics, and in vitro validation, provides a rapid and cost-effective framework for discovering functional hypoglycemic compounds from food sources, advancing both mechanistic understanding and practical applications in diabetes management. • A strategy for the development of a novel food-derived hypoglycemic compound. • QSAR modeling of α-glucosidase inhibitors was developed. • Bioinformatics facilitates the discovery of new food-derived hypoglycemic compounds. • Molecular dynamics simulations of selected compounds.