Modeling and predicting pK(a) values of mono-hydroxylated polychlorinated biphenyls (HO-PCBs) and polybrominated diphenyl ethers (HO-PBDEs) by local molecular descriptors

摘要

Hydroxylated polychlorinated biphenyls (HO-PCBs) and polybrominated diphenyl ethers (HO-PBDEs) are attracting considerable concerns because of their multiple endocrine-disrupting effects and wide existence in environment and organisms. The hydroxyl groups enable these chemicals to be ionizable, and dissociation constant, pK(a), becomes an important parameter for investigating their environmental behavior and biological activities. In this study, a new pK(a) prediction model was developed using local molecular descriptors. The dataset contains 21 experimental pK(a) values of HO-PCBs and HO-PBDEs. The optimized geometries by ab initio HF/6-31G** algorithm were used to calculate the site-specific molecular readiness to accept or donate electron charges. The developed model obtained good statistical performance, which significantly outperformed commercial software ACD and SPARC. Mechanism analysis indicates that pK(a) values increase with the charge-limited donor energy EQ(occ) on hydroxyl oxygen atom and decrease with the energy-limited acceptor charge QE(vac) on hydroxyl hydrogen atom. The regression model was also applied to calculate pK(a) values for all 837 mono-hydroxylated PCBs and PBDEs in each class, aiming to provide basic data for the ecological risk assessment of these chemicals. (C) 2015 Elsevier Ltd. All rights reserved.