Abstract / Summary
The continuous evolution of SARS-CoV-2 highlights the need for alternative antiviral discovery strategies that exploit structurally conserved regions within the spike glycoprotein. Phytochemical constituents of Xeromphis nilotica stem bark were screened using an integrated deep learning-guided QSAR approach, followed by physicochemical, pharmacokinetic, and drug-likeness filtration. From 170 GC–MS-identified compounds, eight fulfilled key bioavailability and medicinal chemistry criteria and were advanced to molecular docking against the receptor-binding domain (RBD) of the spike protein. Cepharanthine (reference ligand; -8.30 kcal/mol) showed the strongest interaction, whereas CID: 994 (-5.70 kcal/mol) and CID: 135,408,763 (-5.50 kcal/mol) emerged as the top-ranking X. nilotica-derived ligands with relevant contacts at ACE2-interacting or conformation-stabilising regions of the RBD. Although absolute QSAR predictive performance was limited due to dataset size, cross-validation and ranking consistency supported its application as a prioritisation tool rather than a quantitative predictor of IC₅₀ values. Molecular dynamics simulations spanning 500 ns confirmed the superior structural resilience of Cepharanthine, followed by CID: 994, which exhibited moderate RMSD and Rg values and lower SASA fluctuations compared to CID: 135,408,763. Clustering and ligand-protein distance analyses further supported the improved binding persistence of CID: 994. MM/GBSA binding free energies reinforced this hierarchy, with Cepharanthine (-45.11 kcal/mol) outperforming CID: 994 (-39.18 kcal/mol), which in turn exceeded CID: 135,408,763 (-37.22 kcal/mol). The integration of QSAR prioritisation, docking, and long-timescale MD simulation provides a consistent multi-level validation framework, reducing reliance on any single computational metric. Density functional theory data revealed a wider HOMO-LUMO gap and lower dipole moment for CID: 994, supporting improved electronic stability and reduced susceptibility to conformational disruption under dynamic conditions. Therefore, CID: 994 is identified as the most promising X. nilotica phytochemical for potential optimisation as an RBD-directed inhibitor after experimental evaluation as a potential inhibitor of SARS-CoV-2 spike-mediated host cell entry.