RESUMO
A phytotoxicity assay based on seed germination/root elongation has been optimized and used to evaluate the toxic effects of some phenolic herbicides. The method has been improved by investigating the influence of experimental conditions. Lepidium sativum was chosen as the most suitable species, showing high germinability, good repeatability of root length measurements, and low sensitivity to seed pretreatment. DMSO was the most appropriate solvent carrier for less water-soluble compounds. Three dinitrophenols and three hydroxybenzonitriles were tested: dinoterb, DNOC, 2,4-dinitrophenol, chloroxynil, bromoxynil, and ioxynil. Toxicity was also determined using the Vibrio fischeri Microtox® test, and a highly significant correlation was found between EC50 values obtained by the two assays. Dinoterb was the most toxic compound. The toxicity of hydroxybenzonitriles followed the order: ioxynil >bromoxynil >chloroxynil; L. sativum exhibited a slightly higher sensitivity than V. fischeri to these compounds. A QSAR analysis highlighted the importance of hydrophobic, electronic, and hydrogen-bonding interactions, in accordance with a mechanism of toxic action based on protonophoric uncoupling of oxidative phosphorylation. The results suggest that the seed germination/root elongation assay with L. sativum is a valid tool for the assessment of xenobiotic toxicity and can be recommended as part of a test battery.