This paper presents a mechanistic analysis of aquatic toxicity data, quantified as pIGC(50) assessed in the 40 h Tetrahymena pyriformis population growth impairment assay, for 40 polyhydroxybenzene derivatives. The toxicity trends of these phenolic compounds have been shown to be consistent with mechanistic organic chemistry principles. Thus, it is shown that the compounds can be grouped into two chemical mechanism of action domains, according to whether they can be oxidized to electrophilic quinones or quinone methides. Compounds in which the hydroxy groups are oriented meta, but not ortho or para, to one another cannot be oxidized to electrophilic quinones or quinone methides and act as polar narcotics. Their toxicities are found to be well-correlated with hydrophobicity (modeled by log D): pIGC(50) = 0.83 (+/-0.04) log D - 1.27 (+/-0.09): n = 10, r(2) (adj) = 0.981, q(2) = 0.974, s = 0.15, and F = 460. Compounds with hydroxy groups oriented ortho or para to one another are more toxic than predicted by this equation, and the toxicity trends within this group of compounds are rationalized in terms of the electrophilic chemistry of their oxidation products. A quantitative correlation is demonstrated between toxicity and electrophilicity of the oxidation products, as modeled by the activation energy index (AEI), a new molecular orbital parameter derived from the computed highest occupied molecular orbital (HOMO) and HOMO-1 orbital energies of the electrophiles and the intermediates for Michael addition of n-butylamine: pIGC(50) (adj) = -0.49 (+/-0.06) AEI + 6.85 (+/-0.69): n = 18, r(2) (adj) = 0.810, q(2) = 0.774, s = 0.24, and F = 73. Outliers to these quantitative structure-activity relationships (QSARs) are easily rationalized in terms of their chemistry (tetrabromocatechol, 4,6-dinitro-1,2,3-trihydroxybenzene, and 2,3,4-trihydroxybenzophenone) or in a demonstrable deficiency in the descriptor (the methyl-substituted hydroquinones, for which the AEI parameter as defined here fails to model the electron donation effects of the methyl groups). The AEI parameter is a mechanism-based molecular orbital parameter new to QSAR and, on the basis of the present findings, it shows promise for further applications. However, some deficiencies have been identified with it, particularly with regard to modeling the electronic effects of methyl (and presumably other alkyl) groups, and there is scope to refine the concept so as to deal with these deficiencies.
Available at: http://works.bepress.com/terry_schultz/29/