Over the past few decades, various inhibitors of PRMT5 have been developed because of its involvement in a variety of tumor development processes. As of now, no drugs targeting PRMT5 have been approved, and multiple drugs entering clinical trials have proven to have side effects. In this study, PRMT5 was used to perform virtual screening of 52119 marine natural compounds by combining various methods. We constructed 20 pharmacophore models based on multiple ligands. The best pharmacophore model AARR_2 was selected by analyzing the statistical parameters of the pharmacophore model and the binding characteristics of the ligand active site, and then 3552 compounds were screened out. Compared with the positive compound, 46 compounds were selected based on the molecular docking fraction and docking mode analysis. Then, 3D-QSAR was used to analyze the relationship between structure and activity of the compounds. Then, in addition to marine compounds 36404, 36405 and 14436, we selected compound 46 (the positive control compound) and used the CLC-Pred online Web server to predict their cytotoxicity to human cell lines, making cell experiments possible. Finally, we conducted the prediction of ADMET in order to better promote clinical trials. After comprehensive judgment, we screened out the marine natural compounds 36404 and 36405 as candidates for PRMT5 substrate competitive inhibitors.Communicated by Ramaswamy H. Sarma.
Keywords: Marine natural compounds; molecular docking; molecular dynamics; multiligand-based pharmacophore model; virtual screening.