Progesterone receptor ligand binding domain : from conformational dynamics to drug discovery
Date of Issue2019-04-26
School of Biological Sciences
Hormones are vital molecules for human differentiation, development, and health. Among them, the female reproduction, development, and cancer-related hormone progesterone has been cast interests for its involvement in nearly all aspects of female health. Recently researches have demonstrated that progesterone, in alliance with estrogen, would suppress breast cancer cell proliferation in cell lines and mouse models. Its receptor, the progesterone receptor (PR), which is a member of the nuclear receptor superfamily, has gradually been recognized as a potential drug target. The dysfunction of this PR would contribute to different types of cancers, including breast cancer. Although there have been several drug molecules designed or discovered against this PR, the side-effects of these drugs are not to be ignored. The necessity of synthesizing new types of drugs against PR, especially its ligand binding domain (LBD), is never decreased. Current studies about PR mainly focus on the biological functions, signaling pathway, its involvement in cancer models. The structural and dynamics of PR, which lay the foundation of the biological understandings, however, are limited, partially due to the difficulty of performing X-ray crystallography or nuclear magnetic resonance spectroscopy (NMR) of the PR, or its domains. We hope to incorporate the structure and dynamics information of PR, or more specifically, the LBD of PR, to identify high potential PR lead molecules for pharmaceutical purposes. Therefore, starting from the available experimental solved structures, we applied biophysics methods to 1) explore the dynamics of PR LBD, 2) and understand the interactions between PR LBD and its agonists and antagonists, 3) and screen large compound libraries to identify potential lead-like molecules. Molecular dynamics (MD) simulations and advanced sampling methods generate good models to estimate the energy landscape of the apo-form PR LBD. The results suggest that the agonistic PR LBD conformation is a meta-stable state. Several other meta-stable states have also been identified and they could be adopted for further virtual screening studies. And the ligand binding (either an agonist or an antagonist) would induce the LBD conformational adaptations more towards the “closed” agonistic state, though different ligands may induce different PR LBD dynamics. We also found that the co-repressor peptide is a necessary component for maintaining the antagonistic conformation. In meanwhile, the ligand-induced conformational adaptations result from both hydrophobic and electrostatic forces. Subsequently, large-scale virtual screening (VS), as well as machine learning-based rescoring calculations, were performed. Towards PR LBD VS, we constructed highly accurate and low false positive rate models and identified 21 potential lead-like molecules, which would be further tested in future. The computational methods or models we used here to fill in the void in structural and dynamics knowledge of PR. The identified molecules could be potential lead molecules for further analysis. The drug discovery strategy we applied would also be useful and could be applied to other targets.