Computational modeling of substituent effects on phenol toxicity

Standard computational models of cytotoxicity of substituted phenols relate the toxicity to a set of quatitative structure-activity relationship (QSAR) descriptors such as log P, pK(a), OH bond dissociation enthalpy (BDE), etc. Implicit in this approach is the idea that the phenoxyl radical is disruptive to the cell and factors increasing its production rate will enhance the toxicity. To improve the QSAR correlations, substituents are usually divided into electron-donating groups (EDG) and electron-withdrawing groups (EWG), which are treated separately and thought to follow different mechanisms of toxicity. In this paper, we focus on one important aspect of toxicity, the rate constant for production of phenoxyl radical. Activation energies are obtained for the reaction of X-phenol with peroxyl radical by using the Evans-Polanyi principle, giving rate constants as a function of Delta BDE values for both EDG and EWG sets. We show that (i) a plot of log k for phenoxyl formation vs Delta BDE shows a double set of straight lines with different slopes, justifying the usual EDG and EWG separation but without requiring any change in mechanism; (ii) the same method can be effectively used for different target radicals (e.g., tert-butoxyl) or different sets of parent compounds (e.g., substituted catechols), thus giving a useful general approach to analysis of toxicity data; (iii) regions of constant toxicity in all cases are predicted; and (iv) we argue that competing parallel mechanisms of toxicity are likely to be dominant for EWG-substituted phenols.