Experimental Determination of Octanol–Water Partition Coefficients of Selected Natural Toxins

Natural toxins are widely occurring, diverse organic compounds by, for example, plants or fungi. In predictive environmental and risk assessment of organic chemicals for regulatory purposes, octanol–water partition coefficient (Kow) remains one of the key parameters. However, experimental for natural toxins are largely missing, and the current estimation for Kow show limited applicability multifunctional, ionizable compounds. Thus, log Kow data were first experimentally derived for a diverse of 45 largely ionizable natural toxins and then compared with predicted values from three different models (KOWWIN, ACD/Percepta, Chemicalize). Both approaches were critically evaluated with regard their applicability for multifunctional, ionizable compounds. The shake-flask approach allowed reliable quantification pH-dependent partitioning behavior for neutral, acidic, and basic natural toxins. All the analyzed toxins are rather polar an average log Kow Kow of 2. 7. Furthermore, comparison of experimental data for the neutral form of ionizable with those of commonly used prediction models showed that the match the former with only slightly increased errors. The Chemicalize gave the best overall predictions for the dataset generated here, a mean absolute error of 0. 49.