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 IC₅₀ 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.