The SARS-CoV-2 main protease (Mpro), a cysteine protease essential for viral polyprotein processing and replication, remains a validated molecular target for the development of COVID-19 therapeutics. This study employed an integrated in silico drug discovery workflow combining structure-based and ligand-based pharmacophore modeling, large-scale virtual screening, molecular docking, and ADME–Tox prediction to identify potential marine natural product inhibitors of Mpro. A dataset of over 30,000 compounds from the Comprehensive Marine Natural Products Database (CMNPD) was screened using a ligand-based pharmacophore derived from 30 reported Mpro inhibitors and a structure-based pharmacophore generated from the Mpro–boceprevir co-crystal complex. Initial screening yielded 900 ligand-based and 1,259 structure-based hits. Subsequent molecular docking analysis identified 88 and 98 compounds, respectively, exhibiting more favorable binding energies than the reference inhibitor boceprevir. Lead candidates, including versixanthone F and penicillixanthone B, demonstrated critical interactions with the catalytic residue Cys145 within the active site. ADME–Tox filtering further prioritized purealin D, pseudogymnoascin A, and pseudogymnoascin B, although pharmacokinetic limitations were noted. Additional non-hepatotoxic candidates such as homodolastatin 3 and dolastatin 3 emerged as promising scaffolds for structural optimization and future experimental validation.
