The use of the ionization constant (pKa) in selecting models of toxicity in phenols

Phenols elicit a toxic response by one of two mechanisms: polar narcosis or uncoupling of oxidative phosphorylation, and pKa values appear to be useful in predicting the mechanism of toxic action of a given phenol. The relative biological response (log BR) of 21 selected phenols in the static Tetrahymena pyriformis population growth test was determined. Seven derivatives, including dinitro- and polyhalogen-substituted phenols, were selected for testing because they were potential uncoupling agents. The other 14 derivatives were all suspected polar narcotics. 1-octanol/water partition coefficient (log Kow)-dependent regression analysis of the two subsets of derivatives results in two linear equations. The polar narcosis model is log BR = 0.6128 (log Kow) - 1.1297; r2 = 0.958, s = 0.187. The uncoupling of oxidative phosphorylation model is log BR = 0.4485 (log Kow) + 0.3007; r2 = 0.985, s = 0.209. Polar narcotic chemicals have pKa values greater than 8.00, while uncoupling agents have pKa values less than 6.50. Combining descriptors and modeling across mechanisms of toxicity in the model log BR = 0.5671 (log Kow) - 0.1885 pKa + 0.8190; r2 = 0.958, s = 0.225.