Molecular docking, DFT analysis, pharmacokinetic profiling and MD simulation of Ilex aquifolium L. flavonoids as potential α-glucosidase predicted inhibitors for diabetes

Abstract The growing diabetes trend worldwide demands the identification of new α-glucosidase inhibitors for delaying postprandial hyperglycemia. However, the use of existing drugs is often limited by adverse effects, which has increased interest in natural products as safer alternative sources of α-glucosidase inhibitors. I. aquifolium was selected based on reports of flavonoid-rich phytochemistry, leading to the hypothesis that its major flavonoids could act as potential α-glucosidase inhibitors for the management of diabetes. Herein, the virtual α-glucosidase inhibitory potential of eight flavonoid-based phytochemicals of I. aquifolium extracts was investigated through an integrated in silico workflow, which includes molecular docking, DFT analysis, and ADMET profiling. Docking analysis identified IA- 2 (Rutin) and IA- 5 (Kaempferol-3-O-rhamnoglucoside) as the most potent ligands with binding affinities of − 9.59 kcal/mol and − 9.18 kcal/mol, respectively, closely approaching the standard acarbose (–10.96 kcal/mol). The re-docking RMSD of 2.1375 Å fell within the acceptable validation benchmark, supporting the reliability of the docking protocol. The second and third positions were occupied by IA- 1 (Quercetin) and IA- 3 (Quercetin-3-O-hexoside) as potential inhibitors, as they formed hydrogen bonds with hydrophobic contacts. DFT study suggested that all the flavonoids exhibited moderate to high dipole moments (4.88–7.36 Debye) along with the high HOMO–LUMO gaps. Additionally, IA- 7 (Apigenin) showed the highest electrophilicity (3.313) among all the phytochemicals. The pharmacokinetic evaluation of ligands by SwissADME and PkCSM suggested IA- 1 as the most potential hit, and it occupied the main drug-likeness parameters with a bioavailability score of 0.55, whereas low bioavailability scores (0.17) for IA- 2 and IA- 5 indicated their poor permeability through the membrane. The potential toxicity of the ligands through SwissADME predicted that all the ligands have renal and respiratory issues, but IA- 1 showed the highest potential of acute dermal toxicity. MD simulations supported the docking results by showing protein backbone RMSD stabilization at 2.0–2.4 Å and revealing that IA-2 and IA-5 form the most dynamically stable complexes (lowest ligand RMSD, compact Rg, persistent H bonds), while IA-3 and IA-1 display greater ligand flexibility, thereby prioritizing these compounds for experimental validation in a hypothesis-generating context. However, these computational findings do not replace experimental validation and should be interpreted as hypothesis-generating predictions. Overall, the study provides a hypothesis-generating framework, prioritizing IA-2 and IA-5 as potential α-glucosidase inhibitory candidates, with IA-1 emerging as a top-ranked ligand through pharmacokinetic parameters, required further experimental validation.