Assessment of Virtual Screening Methods and the Discovery of Inhibitors Targeting HIV-1Integrase—LEDGF/p75Interaction
|School||East China University of Science and Technology|
|Keywords||Ligand-based Virtual Screening Structure-based Virtual Screening ScoringFunction Induced-fit HIV-1Integrase LEDGF/p75|
Virtual screening （VS） is an important component of Computer-aided Drug Design （CADD） and plays an essential role in the process of lead finding and lead optimization. This thesis consists of two parts. The first one is an assessment of VS methods, and the second one is the discovery of inhibitors targeting HIV-1integrase （IN）—LEDGF/p75interaction using VS methods.VS is now a routine technology in the process of drug discovery today. VS can be accomplished in either ligand-or structure-based methods. To evaluate the performance of the ligand-based methods, retrospective VS was performed on a tailored directory of useful decoys （DUD）. The VS performances of142D fingerprints and43D shape similarity methods were compared. The results revealed that2D fingerprints ECFP2and FCFP4could yield better performance than the3D Phase Shape methods.For structure-based VS, this study mainly focus on the evaluation of scoring functions, which can be categorized in three types including empirical-based, force field-based and knowledge-based scoring functions. The average docking results of40targets of DUD indicated that empirical-based scoring function could yield the highest early enrichment followed by force field-based scoring function, and then knowledge-based scoring function. The ligand-based methods were further compared with structure-based methods, which demonstrated that the superiority of ligand-based methods over the docking-based screening in terms of both speed and hit enrichment. Therefore, considering ligand-based methods first in VS workflow would be a wise option.Furthermore, VS performance of induced-fit model was compared with that of crystal structure based on DUD. Induced-fit structure was generated using the cognate ligand to induce each crystal structure. The structure with top IFD score was adopted as induced-fit model. In most targets, VS performance of induced-fit model declined. However, VS performance was improved in some targets, especially for those have larger rmsd value of ligand. The results suggest that for those targets with higher rmsd value after IFD, induced-fit model might be used to improve VS enrichment.The second part of this thesis is the discovery of inhibitors targeting the interaction between HIV-1integrase and human protein LEDGF/p75. Integration of viral-DNA into host chromosome mediated by the viral protein HIV-1IN is an essential step in the HIV-1life cycle. In this process, human protein Lens epithelium-derived growth factor （LEDGF/p75） is discovered to function as a cellular cofactor for the integration. In this study, several screening models were built up and two of them could yield the higher ROC enrichament, which were then adopted in virtural screen precess. Both ligand-based and structure-based VS were conducted to search multi-sources database. The database include DrugBank, National Compound Resource Center （NCRC） natural product library, commercial compounds library SPECS, ENAMINE, in-house library of natural products and their derivertives, in-house FXR antagonists library. In adddition, lead optimization was conducted based on virtual screening results.Totally,85compounds were identified as inhibitors targeting IN-LEDGF/p75interaction through VS, lead optimization and AlphaScreen bioassay. Among them,16ones showed IC50value as low as sub-micromole. Furthermore,5highly potent compounds were selected to test their antiviral activities. The results indicated that all5compounds could inhibit HIV-1（IIIB） replication in C8166cell and the most potent compound showed EC50value at0.57μg/mL. Effect of the5inhibitors on EGFP-IN intracellular distribution were assayed on293T cells, and the results confirmed the antiviral mechanism through disturbing IN-LEDGF/p75interaction to disrupt IN nuclear distribution. This study laid a solid foundation for lead optimization and pushed forward the drug discovery process targeting IN-LEDGF/p75interaction.