Prediction of new C-terminal Hsp90 inhibitors based on deguelin scaffold: homology modeling, virtual screening, QM/MM docking, MM/GBSA, and molecular dynamics simulations.

Abstract

Background and purpose: The N-terminal Hsp90 inhibitors are promising targets for cancer treatment; however, inducing the heat shock response is one of the most significant limitations. A prominent way to overcome this limitation is by inhibiting the Hsp90 C-terminal domain.

Theoretical approach: In this study, a set of structure-based methods was engaged to predict the new C-terminal inhibitors. Since there was no human PDB structure of the Hsp90 C-terminal domain, homology modeling was done using the SWISS-MODEL server online. The 3D structure of the model was refined through energy minimization using molecular dynamics (MD) simulation for 10 ns. The active site of the created model was validated by novobiocin docking. Four steps of virtual screening, including HTVS, SP, XP, and QM/MM docking, were performed on the created library (151,332 compounds) based on 80% similarity to deguelin as the C-terminal inhibitor. The best-obtained compounds were introduced to MM-GBSA studies. Finally, the stability of the best compound was investigated using a 100 ns MD simulation.

Results/findings: Four steps of virtual screening were performed on the created library. The extracted 46 compounds with the XP GlideScore of < -4.164 kcal/mol were introduced to MM-GBSA studies, and rescoring was done. The stability of compound CID_14018348, the best compound (ΔG_binding = -80.45 kcal/mol), was investigated using MD simulation.

Conclusion and implications: The compound CID_14018348 was identified as the most promising candidate through computational techniques; therefore, the computational methods outlined can be applied in the development of potent anticancer agents.