| Polychlorinated biphenyls (PCBs) are important synthetic chlorinated organic compounds, which have high toxicity, persistence, easy to bio-accumulate, and could long-range transport in the environment. PCBs have some characteristics, such as high lipophilicity, low water solubility, and stability. Therefore PCBs are easy to bioaccumulate in the adipose tissue and biomagnified along the food chains, which would cause a range of adverse health effects to animals and human. As outstanding flame retardants, polybrominated diphenyl ethers (PBDEs) have been widely used in industrial products, especially electronic products as well as products such as foam, furniture and indoor decorations. For PBDEs are high lipophilicity and low water solubility, they are easy to bioaccumulated and biomagnified along the food chains. In addition, PBDEs have stable physical and chemical properties and resist degradation. Therefore the potential hazards of PBDEs to animals and human have attracted worldwide attentions.Experimental studies confirmed that a variety of PCBs and PBDEs and their hydroxylated derivatives have a variety of endocrine disrupting effects on estrogen, androgen and thyroid hormone, etc. However, given the number of PCBs and PBDEs and their hydroxylated derivatives is large, and either in vivo or in vitro bioassay test is time-consuming and expensive, it is impossible to determine the endocrine disrupting effects of each compound, so relatively fewer studies have addressed the mechanisms of endocrine disrupting effects and the ecological risk assessments. Hence, based on limited toxicity data, some fast and resource-effective in silico techniques, such as quantitative structure-activity relationship (QSAR) are considered as powerful alternatives. QSAR method could be used to predict the endocrine disrupting effects, provide a helpful reference for the mechanism of PCBs and PBDEs and their hydroxylated derivatives' endocrine disrupting effects and supply data support for further ecological risk assessments.PCBs and PBDEs and their hydroxylated derivatives (HO-PCBs and HO-PBDEs) are the target compounds in this thesis. A multistep modeling framework combining three-dimensional QSAR (3D-QSAR), molecular docking and molecular dynamics (MD) simulations was performed to establish the endocrine disrupting activity predict models, reveal the key structural features influencing the endocrine disrupting activities and investigate the detailed interaction mechanisms of ligands and receptors.In this thesis, a brief description of the experimental studies on the endocrine disrupting activities of PCBs and PBDEs and their hydroxylated derivatives were given in Chapter 1. Meanwhile, the basic theories, methods and applications of QSAR, molecular docking and MD simulations were also introduced.The Chapter 2 was composed of three parts. In Chapter 2.1, a combined computational approach by 3D-QSAR, molecular docking, and MD simulations was performed to reveal the key structural features influencing the anti-androgen activity of PCBs and investigate the detailed binding mode between PCBs and androgen receptor (AR). The probable bioactive pose of these compounds was generated by molecular docking, the developed receptor-based CoMSIA model yielded high q2 and lvalues (q2=0.665, r2=0.945). Moreover, detailed interaction mechanisms of two PCBs with many different activities were analyzed on the basis of the results of molecular docking and MD simulations. The results showed that there were strong hydrophobic interactions between the ligand-receptor. The results of binding free energy calculations indicated that van der Waals interaction energy was the predominant driving forces for the binding of PCBs and androgen receptor. Amino acid residues Leu704, Leu707, Met742, Met745 and Phe764 were the key residues related with the binding process.In Chapter 2.2,3D-QSAR was adopted to study the anti-androgen activity of HO-PCBs. A total of 82 HO-PCBs were collected to build 3D-QSAR models. The model was optimized through the region focusing and the resulted model yields a q2 of 0.554, an r1 of 0.918, proving a good stability and predictive ability. Molecular docking revealed that the hydrogen bond between HO-PCBs and residue Gln711 as well as hydrophobic interactions with some hydrophobic residues played important roles for the binding of HO-PCBs and androgen receptor. Furthermore, the average structures of MD simulations verified the reliability of docking.In Chapter 2.3, a multistep framework by combining molecular docking, MD simulations, and 3D-QSAR was performed to study the estrogenic activity of HO-PCBs. The influence of two different alignment rules (ligand-based alignment and receptor-based alignment) on estrogenic activity predictive model was also analyzed. The statistics of both models indicated that the receptor-based CoMSIA model (g2=0.648, r2=0.968) produced better results than the ligand-based CoMSIA model(g2=0.283, r2=0.648), which suggested that the bioactive pose generated by molecular docking could improve the quality of the model significantly. MD simulations revealed that the intermolecular van der Waals interactions and the electrostatics interactions were two predominant driving forces for the binding of HO-PCBs and estrogen receptor. The order of the calculated binding affinities of the compounds was in accordance with the order of experimental estrogenic activities.The Chapter 3 consisted of three parts. In Chapter 3.1, based on the homology modeled 3D structure of AhR, molecular docking was performed to identify the probable bioactive conformations of ligands. On the basis of the docking conformations of ligands, an optimal CoMSIA model of AhR binding affinity of PBDEs was developed to find the key structural features affecting the binding affinity. The statistical parameters indicated that the receptor-based CoMSIA model (q2=0.605, r2=0.996) exhibited better statistical results than the common scaffold-based model(q2=0.509,r2=0.780). Docking results showed that the hydrogen bond and hydrophobic interactions are the main factors influence the binding of ligands and receptor.In Chapter 3.2, the hormone activities of a series of HO-PBDEs to thyroid receptors ? were studied based on the combination of 3D-QSAR, molecular docking, and MD simulations methods. For CoMSIA analysis, the model based on steric, electrostatic, hydrogen-donor and hydrogen bond-acceptor fields was found to be the most accurate model. In addition, the model produced from the receptor-based alignment had better statistical results than the model from the ligand-based alignment. The final model was validated by a series of internal validation and external validation methods. The results further confirmed that the model was reliable and has high predictive ability. In addition, molecular docking elucidated the conformations of compounds and key amino acid residues at the docking pocket, MD simulations further determined the binding process and validated the rationality of docking results.In Chapter 3.3, molecular docking, MD simulations and binding free energy calculations and decomposition were performed to further understand the detailed interaction mechanism of HO-PBDEs with ERa. Molecular docking was employed to reveal the probable binding conformations of the compounds at the active site of ERa, and MD simulations were performed to determine the detailed binding process. The binding free energy calculations and energy decomposition were applied to reveal the dominant driving force for the binding and identify the contributions of the key residues related with the binding process.In Chapter 4, the wok of this thesis was summarized, the conclusion and innovation of this thesis were also introduced. Besides, a follow-up study was prospected. |
PDF Full Text Request