Quantitative Structure-activity Relationship Study Of Some Environmental Hormones

It is estimated that 300 million tons compounds are produced per year and 100 thousands kinds of chemicals are used each day.These chemicals are released into the environment and go into human body through various exposure paths.Some of them are named after Environmental Hormones,i.e.EHs.Intensive research has revealed that these chemicals released into the environment have the potential to interrupt the normal functioning of the endocrine systems of humans and wildlife by mimicking or antagonizing natural hormones.Such EHs exposures may induce the reproductive,developmental toxicity,immunotoxicity,tumorigenesis and teratogenesis.In Vivo and in Vitro tests can be used to measure the estrogen activity of chemicals,but such tests are laborious,time-consuming,and expensive.Thus,it is practically impossible to carry out thorough toxicological tests on the more than 58000 potential xenoestrogens that may ultimately need to be evaluated.On being faced with the challenging task of screening large molecule libraries for biological activity,the benefits of virtual screening with quantitative structure-activity relationship(QSAR)techniques become immediately obvious when endeavoring to identify possible endocrine disruptors.Instead of arduous and expensive laboratory work,biological activity will be predicted solely on the basis of the molecular structure of the compound and will,thus,decrease the number of animal tests.The compounds,which are concerned for their potential threat to the health of human and wildlife,including steroids,androgens,phytoestrogens,environmental estrogens,Polychlorinated biphenyls(PCBs),Poly-and perfluorinated compounds(PFCs)were investigated in this thesis.Cell-based reporter gene assays are used to determine the activity of PFCs towards human pregnane X receptor(hPXR).Heuristic method,best subset modeling method and Polynomial Neural Networks(PNN)were combined to construct 2D-QSAR models.CoMFA and CoMISA were performed to develop 3D-QSAR.Molecular docking and site-directed mutagenesis were combined to investigate the interaction between the PFCs and hPXR,thus to provide insight on the moleculer mechanism related to PFCs activity towards hPXR.Molecular descriptors used to construct 2D-QSAR models includes:(a)OD-constitutional(atom and group counts);(b)ID-functional groups,1D-atom centered fragments;(c)2D-topological,2D-BCUTs,2D-walk and path counts,2D-autocorrelations,2D-connectivity indices,2D-information indices,2D-topological charge indices,and 2D-eigenvalue-based indices;and(d)3D-Randic molecular profiles from the geometry matrix,3D-geometrical,3D-WHIM,and 3D-GETAWAY descriptors.Variable selection was performed to develop the robust QSAR models with good predictive ability and feasible to explain.In this thesis,variable selection was performed based on heuristic method,best subset modeling method and Polynomial Neural Networks(PNN).The critical structural features related to activity were identified.3D-QSAR methods,such as CoMFA and CoMSIA,respesent the structural information using steric field,electrostatic field,hydrophobic field,hydrogen bond donor field and hydrogen bond acceptor field.These models consider the 3D bioactive conformation of a compound and the biological environment surrounding small molecules,thus more descripting the interaction between the small molecule and receptor.However,the bioactive conformations and alignment rules are very sensitive to the result of 3D-QSAR analysis,especially for flexible molecules with complex structures.In this thesis,three alignments,including atom-fit,field-fit and receptor-based alignments,were investigated.The effects of different alignments on models results were discussed.Molecular docking is performed to investigate the binding mode of compounds with receptors,which combines with coefficient contours from CoMFA and CoMSIA analysis to investigate the structural features and molecular mechanism related to bioactivities of compounds.In addition,docking analysis was performed to identify the key residues of binding pocket of hPXR,which was then tested and verified by using site-directed mutagenesis.Molecular mechanism of PFCs to activate hPXR seems to be revealed.The cell-based reporter gene assays and QSAR models developed in this thesis can provide a guidance for risk assessments of environmental hormones.