【作者】 李斐；

【导师】 陈景文；

【作者基本信息】 大连理工大学 ， 环境科学， 2010， 博士

【摘要】 许多化学品具有潜在的内分泌干扰效应,会对人类健康和生态系统造成危害。仅仅通过实验方法筛选内分泌干扰物,成本昂贵、费时费力,且部分化合物尚缺乏标准样品,因此,有必要建立有机污染物内分泌干扰效应的高通量(虚拟)筛选模型和定量结构-活性相关(QSAR)模型。本研究结合分子对接和实验检测方法,建立了有机污染物内分泌干扰效应的QSAR模型；研究了蒽醌类(AQs)化合物和羟基多溴联苯醚(HO-PBDEs)与激素受体间的相互作用,揭示了上述化合物的拟激素效应与其分子结构间的内在联系。基于7类化合物的雌激素效应相对效应势(logRP)和DRAGON次件计算得到的分子结构参数,采用偏最小二乘(PLS)回归,建立了化合物logRP值的QSAR模型。利用重组基因酵母法测定了外部独立验证集的logRP值,对建立的QSAR模型进行了性能评价。研究结果表明：建立的QSAR模型具有较好的拟合优度、稳健性和预测能力,相关系数(R2)、交叉验证系数(Q2LOO)、外部可解释方差(Q2EXT)分别为0.89、0.90、0.78。化合物的雌激素效应不仅与分子的大小、形状和极化率有关,还与分子的立体结构特征及分子间的相互作用有关。通过分子对接剖析了AQs分子与雌激素受体(ERα)间的相互作用,采用重组基因酵母法检测了AQs化合物的雌激素效应,并建立了AQs化合物雌激素效应的QSAR模型。研究结果表明：AQs分子与ERα间存在氢键、疏水和静电相互作用。AQs化合物的雌激素效应主要与其分子的分配能力和静电相互作用有关。极化率项(π1)、结合能(Ebinding)、平均分子极化率(α)、分子中原子的最负电荷(q-)和分子表面负静电势的平均值(Vs-)是影响AQs化合物雌激素效应的重要因素。AQs分子的a越大,雌激素效应越强。Vs-较负的AQs分子更易于与受体ERα结合,显示较强的雌激素效应；建立的QSAR模型稳健性和预测能力较高,具有较强的机理可解释性,能够用于预测其他AQs化合物的雌激素效应。通过分子对接研究了HO-PBDEs与甲状腺激素受体(TRβ)间的相互作用,采用双杂交酵母检测了HO-PBDEs的甲状腺激素效应,建立了HO-PBDEs甲状腺激素效应的QSAR模型。研究结果表明：与TRβ结合是HO-PBDEs发挥甲状腺激素效应的关键步骤,HO-PBDEs分子与TRβ间存在氢键、疏水和π-π相互作用。与PBDEs相比,HO-PBDEs具有较强的甲状腺激素效应(p<0.05)。溴原子数目(nBr)、正辛醇-水分配系数(logΚow)、芳香性指数(IA)、分子最低未占据轨道能(ELUMO)、亲电性指数(ω)和偶极矩(μ)与HO-PBDEs的甲状腺激素效应密切相关。HO-PBDEs分子在水相和生物相间的分配能力是控制其甲状腺激素效应的重要因素。HO-PBDEs的非平面性和得电子能力也会影响其甲状腺激素效应的大小。得电子能力较高的HO-PBDEs分子与受体间呈现较弱的氢键作用,相应的显示较弱的甲状腺激素效应。建立的QSAR模型具有较好的稳健性、预测能力和机理可解释性,能够用于预测其他HO-PBDEs化合物的甲状腺激素效应。更多还原

【Abstract】 The studies on the eco-toxicological effects of the environmental organic pollutants are of great importance to their ecological risk assessment. A number of organic compounds are released into the environment, most of which are suspicious endocrine disrupting chemicals (EDCs). Experimental determination of EDCs is costly, time-consuming, and restricted by lack of sufficiently pure chemicals, thus there is a need to develop high throughput screening method to screen suspicious EDCs. The purpose of the present study was to develop quantitative structure-activity relationship (QSAR) models for endocrine disrupting effects, to clarify the mechanism of the endocrine disrupting of exogenous compounds and to establish the relationships between the hormone-like effect and the molecular structure parameters by an integrated in silico and in vitro approach.A QSAR model for a series of xenoestrogens was developed based on the determined estrogenic activities (expressed by logRP) by partial least squares (PLS) regression. An independent external validation set was employed to evaluate the performance of the developed QSAR model. The estabilshed QSAR model had goodness-of-fit, robustness and predictivity. The leave-one-out cross-validation squared correlation coefficient (Q2L∞) was 0.90. The predicted logRP values were consistent with the observed values, with a root mean square error (RMSE) of 0.74 log units and the squared correlation coefficient (Q2EXT) was 0.78. The current QSAR model could be used to high-throughput screen xenoestrogens. The logRP value was related to polarizability, molecular size, shape profiles and intermolecular interactions.To further understand the mechanism of action, we explored the estrogenic activities of 20 anthraquinone derivatives (AQs). Molecular docking analysis showed that hydrogen bonding, hydrophobic and electrostatic interactions between the AQs molecules and estrogen receptor alpha (ERa) governed the estrogenic activities of the AQs. The recombinant yeast-based assay was employed to determine the estrogenic activities of the 20 AQs. Based on the observed interactions between the AQs and ERa, the polarizability term (πl), the binding energy (Ebinding), the average molecular polarizability (a), the most negative formal charge in the molecule (q) and the average of the negative potentials on the molecular surface (Vs-) were adopted to develop a QSAR model, which had good robustness, predictive ability and mechanism interpretability. The average molecular polarizability (a) is the main molecular structural factor. The AQs molecule with greaterαvalue exerted higher estrogenic activity. Moreover, q and Vs-also affect the estrogenic activity. The results indicated that the estrogenic activity of the AQs depended on the partition ability of compounds into the bio-phase and the electrostatic interaction with the receptor.Molecular docking was adopted to simulate the interactions between 18 hydroxylated polybrominated diphenyl ethers (HO-PBDEs) and thyroid hormone receptors beta (TRβ). The results showed that the bindings to TRβare the key steps for the HO-PBDEs to exert their thyroid hormone activity, and hydrogen bonding, hydrophobic andπ-πinteractions were also found between the HO-PBDEs and TRβ. The thyroid hormone activities of the 18 HO-PBDEs were determined by the yeast two-hybrid assay. The 18 tested HO-PBDEs exhibit significant higher thyroid hormone activities than PBDEs (p< 0.05). Based on the observed mechanism of interactions, appropriate molecular structural parameters were adopted to develop a QSAR model. The number of bromine atoms (nBr) octanol/water partition coefficient (logΚow), aromaticity index (IA), energy of the highest occupied molecular orbital (EHomo), electrophilicity index (ω) and dipole moment (μ) were significant parameters explaining the thyroid hormone activity. The developed QSAR model had good robustness, predictive ability and mechanism interpretability. The partition ability of the HO-PBDEs into the bio-phase was an important factor governing their thyroid hormone activities. Non-planarity of HO-PBDEs facilitates the binding with TRβ. The HO-PBDEs with higher ability to accept proton (as indicated by ELUMO andω) tend to have weak H-bonding with the receptor, and lower thyroid hormone activities.更多还原