DynamicBind and Molecular Dynamics Study of Diarylheptanoids from Curcuma comosa Roxb. as Potential Caspase-3 Inhibitors

The discovery of caspase-3 inhibitors is a significant area in drug development. Diarylheptanoids from Curcuma comosa rhizomes present a promising source of such inhibitors. This study explores the use of DynamicBind, a deep learning-based docking program, as a tool for identifying potential inhibitors. The optimal parameters for DynamicBind docking, including the number of samples, inference steps, and seed number, were validated. The accuracy was tested by redocking an X-ray ligand into caspase-3 (1NMS), achieving an RMSD of 0.536 Å. Six diarylheptanoids were docked, and compound 3 showed a low RMSD (0.827 Å) and a strong confidence score (cLDDT 0.625), interacting with four amino acids in the binding site, matching those in the X-ray ligand of 1NMS. Notably, GLY122 was unique to this test and may be key to its activity. These results are consistent with the reported anti-apoptotic activity. Molecular dynamics simulations with GROMACS confirm the binding results and reveal the complex's stability. In conclusion, DynamicBind is an evolved docking program for identifying protein-ligand pairs without prior knowledge of the binding site. It can isolate an active compound from inactive derivatives with greater confidentiality. This program has potential prominence in drug design and discovery.