Validated virtual screening models for identifying allosteric inhibitors of ATP-citrate lyase: the role of docking scores and protein–ligand interaction similarity in hit selection
摘要
ATP-citrate lyase (ACLY) is an upstream enzyme involved in fatty acid synthesis, cholesterol metabolism, and histone acetylation. Therefore, selective inhibition of ACLY represents a promising strategy for the treatment of dyslipidemia and various cancers. Recently, the cryo-EM structure of the ACLY complex with the allosteric inhibitor NDI-091143 has been reported, providing an opportunity to discover new potent inhibitors of this emerging target. In this in silico study, we report structure-based models that were rigorously developed and evaluated using reported allosteric inhibitors of ACLY. The pharmacophore model (ROC–AUC = 0.85, GH = 0.78, and EF1% = 49.18) and the molecular docking model (RMSDredock = 0.884 Å and ROC–AUC = 0.95) were applied to virtual screening of the ZINC15 library. During hit selection for further evaluation by molecular dynamics simulations, post-docking analysis was performed based on docking scores (ΔGdock) alone and in combination with the Tanimoto similarity coefficient of protein–ligand interaction fingerprints (TcIFP). The combined ΔGdock and TcIFP approach enabled the identification of four out of five selected top hits with binding free energies more favorable than that of the reference compound NDI-091143, supporting their potential as allosteric ACLY inhibitors. These compounds may be subjected to further experimental evaluation to confirm their biological activity. In addition, the workflow developed in the present study may provide a basis for future discovery and optimization of allosteric ACLY inhibitors.
Graphical abstract