[QSPR study on the lipid-water partition coefficients of dioxins based on DFT].

With computational method of density functional theory (DFT), quantified model study of equilibrium partitioning properties of polychlorinated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PCDFs) between lipid phase of organism (Poecilia reticulata) and water phase was carried out based on quantum chemical and further calculated parameters, namely frontier orbital energies, entropies, traceless quadrupole moments as well as molecular absolute hardness, electronegativities and electrophilicity indices, which all were derived from full geometry optimization of PCDD/Fs. Through multiple linear regression (MLR) analyses, quantitative structure-property relationship (QSPR) was successfully proposed in the form of multi-parameter quadratic function: lgK(hw) = 5.343 - 0.001(S - 125.480)(2) - 0.355(omega - 3.239)(2) + 0.006( Q = - 2.950)(2) - 22.728(eta - 2.365)(2). It was shown that the obtained QSPR had higher goodness of fitting and robustness, determination coefficient and cross-validated correlation coefficient being 0.943 and 0.908 respectively, and it was also provided with ideal interior and exterior predictive abilities so that it could be used for prediction of unknown lipid-water partitioning properties. By comparison, QSPR in this research was superior to that from previous SOFA (solubility parameter for fate analysis) method on the whole. Lipid-water partitioning properties (coefficients) of PCDD/Fs should be mainly related to molecular volume and aryl hydrocarbon molecular interactions determined by charge distribution. To a certain degree, they also might be influenced by potential biotransformation and molecular reactivity.