Applications Of QSAR/QSPR Methods In Environment、 Drug And Materials Chemistry

Quantitative Structure-activity/property relationship （QSAR/QSPR） study, as one of the most common method for theory prediction, is the combination of theoretical or semi-experimental calculation and statistical analysis to establish models between the molecular structure descriptors and physicochemical properties or biological activities of compounds. It is not only helpful to disclose the structure factors effecting the properties of compounds from the established QSAR/QSPR models, but also to understand and predict the influence of microstructure on some properties of substance. The methods have been widely used for prediction of various physicochemical properties and biological activities of organic compounds in environmental chemistry, pharmaceutical chemistry, material chemistry and other chemical related fields.In this dissertation, based on QSAR/QSPR method, some 2D and 3D structural descriptors having clear the physical and chemical meaning, containing large amount of information and more comprehensive characterization were adopted to describe structural characteristics of the compounds, combined with mathematical statistics, the following aspects were carried out.（1） QSPR studies of several important organic pollutants using MEDV.Contamination of organic pollutants is of important environmental issue, which is restricted to human survival and development. They enter the environment through a variety of ways, and have caused serious harm on human health and ecological balance. It has important significance to study some behaviors of organic pollutants for the governance and ecological risk assessment. The molecular electro-negativity distance vector （MEDV） based on atoms types and attributes was used to describe the chemical structures of Polychlorinated dibenzofurans（PCDFs）、Polychlorinated benzothiophene（PCDTs）、Polychlorinated Biphenyls （PCBs）、 Organophosphorus pesticides、Polycyclic aromatic hydrocarbons downstream products of coal and nitro energetic materials. Corresponding QSAR/QSPR models were established by multiple linear regression （MLR） and partial least square regression（PLS）. Satisfactory results were obtained. The models provided theoretical basis for chromatographic separation and environmental migration behavior of organic pollutants. The detailed results are the followings: ① SPR study on photodegradation behavior of PCDFs using MEDV Two QSPR models involving two parameters on photolysis half-lives （t1/2） of PCDFs on spruce needles and fly ash surfaces were developed. The QSPR models of spruce needles and fly ash surfaces are the followings respectively:lgt1/2=1.166+0.034×M13+0.012×M23 n=37, m=2, R=0.826, SD=0.052, Rcv=0.797, SDcv=0.056, Qext=0.939lgt1/2=1.770-0.024×M12+0.022×M33 n=42, m=2, R=0.824, SD=0.046, Rcv=0.778, SDcv=0.051, Qext=0.853The results show that the photolysis half-lives is concerned with the third type atoms. The more the third type atoms are, the longer photolysis half-lives are.② QSRR study on GC retention times of PCDTs using MEDVUsing multiple linear regression （MLR） and partial least square regression（PLS）, quantitative structure-retention relationship model of 37 PCDTs was developed.RI=-21.8121+5.8378×M13+2.3531×M23 R=0.9951, SD=2.6938, RCV=0.9942, SDCV=2.9218, Q2ext=0.9839Results showed that the model exhibited excellent estimate capability for internal sample set and good predictive capability for external sample set. For the compounds investigated, the third type atoms have significant influence on the retention indices. The more the number of the third type atoms, the longer the RI is.③ QSPR studies of PCBs using MEDVSix QSPR models of PCBs were developed with high correlation coefficients. They are the relationship between MEDV and gas-chromatographic retention indices（RRI）,aqueous solubility（LgSw）,soil absorption（LgKoc）,aqueous activity coefficient（LgYw）,total molecular surface area（TSA）, Enthalpy of vaporization（△Hvap） respectively. The linear correlation coefficients and leave-one-out （LOO） cross-validation （CV） are all about 0.90-0.99. The results show that the MEDV may be characteristic the structure information of PCBs and the QSPR models possess good stability and predictability. The results provide important information for the environmental assessment.④ SRR study on GC retention times of organophosphorus pesticideQuantitative structure-retention relationship （QSRR） between MEDV of 55 organophosphorus pesticide residues from cereals and oilseeds and their gas chromatography retention times was studied. The estimation stability and generalization ability of the models were strictly analyzed by both internal and external validation. The result showed that the MEDV can be used for the characterization of the structure information of organophosphorus pesticide residues compounds, the model is good of stability and predictability. The model may be not only to predict gas chromatographic retention values for the same pesticide and to achieve qualitative identification of pesticides in the absence of some pesticide standards. The results have great practical value.⑤ SRR study on LC retention index of PAHsThe quantitative structure-retension relationship （QSRR） model on the retention index of liquid chromatography for polycyclic aromatic hydrocarbons was established using MLR and PLS. The correlation coefficient R of established MLR,PLS models, leave-one-out （LOO） cross-validation （CV）,Qext were 0.9970,0.9950,0.9925 （MLR）; 0.9930, 0.9790,0.9917 （PLS）,respectively. The result shows that the MEDV may be characteristic the structure information of Polycyclic Aromatic Hydrocarbons and the stability of the model is good. And it provided valuable theoretical reference for the establishment of governance、separation、purification and detection methods of PAHs.⑥ SPR study on impact sensitivity of nitro energetic materialsThe relationship between MEDV of nitro energetic materials and impact sensitivity was studied. And QSPR model was built as follow:1gH50=1.582+0.030M33-0.184M24-0.482M34+0.041M13 N=34 m=4 R=0.882 SD=0.290 RLOO=0.836 SDLOO=0.337It is concluded that there is much larger influence of the forth type atom on impact sensitivity and it showed a negative correlation. Moreover, the third type atom has a certain influence and there was a positive correlation.（2） QSAR studies of several drugs3D holographic vector of atomic interaction field（3D-HoVAIF）based on electrostatic, steric and hydrophobic three kinds of non-key mode of interaction was used to describe the structures of triazine herbicides、 phenylsulfonyl acid esters、sulfa drugs、coumarin compounds and 5’-nucleoside inhibitors. Corresponding QSAR/QSPR models were established using MLR and PLS. The models provided theoretical basis for predicting the activity、toxicity and discussing the mechanism of action and computer-aided molecular design of pesticides. The detailed results are the followings:① SAR study on weeding activity of triazine herbicidesThe QSAR model between 3D-HoVAIF of 30 triazine herbicides and weeding activity was built.PI50=0.754S2-6-0.517E2-10+0.714E3-4+0.854S2-2 R=0.897, SD=0.4702, Rcv=0.8179, SDcv=0.613, Qext=0.8558The result demonstated that 3D-HoVAIF can better characterize the molecular structure information of triazine herbicides. On the other hand, it is concluded that the influences of steric interaction and electrostatic interaction are relatively much larger on weeding activity. They are the steric interaction between SP3 C and SP3 C, SP2 N and SP3 C, respectively. And the electrostatic interaction between SP2 C and SP3 N, SP3 C and halogen, respectively. The activity will be improved either the length of substituent group is properly long or the nitrogen atom is monosubstitution. And the same time, the electronegativity of substituent group is had to consideration. These findings provided important theoretical guidance to the synthesis of triazine herbicides.②QSAR study on acute toxicity of phenylsulfonyl acid estersPhenylsulfonyl acid esters are a class of sulfur compounds, widely used as insecticides, herbicides and insect repellent. Its unique biological activity makes a greater contribution to mankind, but it is also toxic to the environment and has caused a lot of negative impact.The QSAR model between 3D-HoVAIF of phenylsulfonyl acid esters and photobacterium acute toxicity was established. The correlation coefficient R of established model,leave-one-out（LOO）cross-validation（CV）are0.9470,0.9176,respectivel y. This method has the advantage of predicting the toxicity of the molecule without advance overlapping operations. The model showed there is a negative correlation between acute toxicity and steric interaction of sulfone and ester group. The conclusion provides an important theoretical basis for studying toxic effects、toxicity mechanism and acute toxicity evaluation of these compounds on the photobacterium.③ QSPR study on PKa of sulfonamidesExtensive use of sulfonamides caused a series of animal-derived drug residue problems, causing serious environmental pollution. So it is particularly important to study the adsorption and migration behavior of sulfonamides in the environment. Quantitative structure-property relationship （QSPR） between the 3D-HoVAIF parameters and the PKa values of sulfonamides was generated.pKa=-0.616E5.9+0.324E8-8+0.383E3-6+0.206E5-8 n=21, R=0.950, Rcv=O.926, RMSEF=0.547, RMSEP=1.18The result disclosed electrostatic interaction is the key factor of affecting Pka. There is a negative correlation of the electrostatic interaction between SP3 N and SP2 S in especial.This conclusion broadened an idea for researching the adsorption and migration behavior of sulfonamides in the environment.④ QSAR study on inhibitory activity of 5’-nucleoside inhibitorsQSAR model between the 3D-HoVAIF parameters of 5’-nucleoside inhibitors and activity was generated as below.LogKi=12.764-23.873×E2.6+6.608×E2-8-1.592×10-6×H2₈—0.565×S2-5 R=0.874, SD=0.688; Rcv=0.773, SDcv=0.900, Qext=0.953The result showed that the model is of good stability and preditivity. In addition, the result disclosed the electrostatic interactions between sp3 C and sp2N, sp3C and sp3O/S are important for inhibitory activity. And these results are expected to provide a reliable theoretical basis for forecasting inhibitory activity of other 5’-nucleoside inhibitors and synthetic analogues MTA.⑤ SPR study on the binding constants of coumarins and Human Serum AlbuminThe binding capacity between drugs and human serum albumin has a great influence on drug efficacy and metabolic effects. The QSPR model between the binding constants of coumarins and Human Serum Albumin was built as below:1gK=0.916V112+0.584 V27+0.233 V118 N=20 R2=0.898 SD=0.324 R2cv=0.854 SDcv=0.392The model pointed out hydrophobic and electrostatic interaction will promote the combination and the effect of hydrophobic interaction is greater than the electrostatic interaction.（3） Study on the modification of two types of multiwall carbon nanotubes using mesophase pitch fluorideThe surfaces of two types of Multiple Wall Carbon Nanotubes（MWCNTs） were successfully modified by mesophase pitch fluoride coating. The results showed their electrochemical properties were improved to different extent. This fact illustrated the prediction that mesophase pitch fluoride with a similar structure of graphite fluoride can enhance the electrochemical property of MWCNTs are correct, which open up a new surface modification method for carbon nanotubes.