Decrypting the allosteric network: multi-scale dynamics of SARS-CoV-2 main protease inhibition by Moringa oleifera...

The SARS-CoV-2 main protease (M^pro) is a prototypical enzyme whose activity is governed by long-range allosteric communication within its dynamic structural network. Flavonoids from Moringa oleifera have been reported to interact with M^pro, yet the dynamic mechanisms underlying their modulation remain unclear. Here, we employ an integrated computational biophysics framework to elucidate structure-dynamics-function relationships governing the interaction of 23 natural metabolites with M^pro. Our workflow combines ensemble docking using free-energy-landscape-derived conformations, microsecond-scale aggregate molecular dynamics simulations, and binding energy decomposition to characterize ligand-dependent interaction patterns. By integrating residue interaction network (RIN) analysis and dynamic cross-correlation matrices (DCCM), we demonstrate that specific flavonoids induce pronounced perturbations in a GLU166-centered interaction hub, reshaping long-range correlated motions essential for functional coupling within the protease. These allosteric effects are not captured by conventional affinity-based metrics alone. Explainable machine-learning interpretation and structure-activity relationship analysis further reveal how key physicochemical features modulate dynamic coupling rather than static binding. Collectively, this study provides mechanistic insight into allosteric regulation in M^pro and establishes a generalizable computational framework for probing ligand-induced perturbations of biomolecular dynamic networks.