Bioconcentration Of Organic Pollutants And QSPR

The migration, transformation and behavior of organic pollutants have been of concern by environmental scientists. One of the main issues most scientists have paid close attention is the persistent organic pollutants with great bioconcentration factors and toxicity. Bioconcentration factor (BCF) is an important ecotoxicological parameter which is used to describe the accumulation of chemicals in organisms, primarily aquatic organisms living in contaminated environments. Fish are food for many species, including humans. Organic pollutants can be accumulated in human body through food chain and result in harmful effect to human beings.BCF factor can be obtained by using experimental method. However, it is very expensive and time-consuming to obtain the BCF from the traditional experimental methods. Sometimes, the BCF obtained from the experimental methods is not reliable. Quantitative structure–property relationship (QSPR) provides a promising method for the estimation of compound's BCF based on the descriptors derived solely from the molecular structure to fit experimental data. The advantage of this approach over other methods lies in the fact that it requires only the knowledge of chemical structure and is not dependent on the experiment properties.In this paper, 697 organic compounds, experimentally determined BCF were collected from two literature reviews, were used for the QSPR analysis. QSPR was carried out between a number of physicochemical parameters or molecular descriptors and bioconcentration factors. The results showed that the relationship between bioconcentration factors of organic pollutants and the traditional octanlo/water partition coefficient is not ideal. Introducing other parameters of physical and chemical properties can't improve the results. In order to explain the mechanism of bioconcentration and investigate the influence of molecular structure to bioconcentration, the compounds of data set were classed to 21 groups based on the structure and functional groups. The available data set of chemicals were split into a training set (Group1-18), 391 compounds, a testing set (Group19), 260 compounds, and ionic compounds (Group 20,21,46 compounds) according to the classification. The correlation analysis was performed between molecular descriptors and bioconcentration factors according to linear regression, nonlinear equations and linear free energy relationship (LFER). The results showed that the fitting results of models based on octanol/water partition coefficient are good within a certain range of log P (log P=1～6). However, big prediction errors were found for highly hydrophobic and hydrophilic compounds from linear log P models. The results also showed that predicted results of non-linear Gaussian model is better than that of octanol/water partition coefficient model for high hydrophobic compounds(log P>6)and high hydrophilic compounds (log P<0). For high hydrophobic compounds, there is a large deviation between measured values and predicted values. The reason is due to the low water solubility of compounds. A significant BCF model was developed by using log D instead of log P for ionic compounds in data set. At the same time, the LFER model was developed according to Abraham parameters, with square of the correlation coefficient (R2) 0.84. The mechanism of bioconcentration was discussed according to the equation coefficients.