Phenolic toxicity--detection and classification through the use of a recombinant bioluminescent Escherichia coli.

A genetically engineered Escherichia coli strain, DPD2540, containing a fabA::luxCDABE fusion that gives a bioluminescent output when membrane fatty acids are limited was used to determine the extent to which phenolics result in the limitation of membrane fatty acids. Tested phenolics were found to be classifiable into two groups according to the bioluminescent response they elicited and their pKa. A dose-dependent bioluminescent response, due to fatty acid limitation, was seen for phenolics with pKa values greater than seven, which exist mainly in the protonated form (HA), while no significant bioluminescent response was seen, compared with the control, for phenolics with pKa values lower than seven, which exist almost entirely as A-. A newly modified distribution model for phenolic compounds in the cellular membrane is proposed and used to predict the bioluminescent response induced by group I phenolics and the cellular toxicity for both groups. The [HA]*, obtained with this model, shows good correlation with the various bioluminescent responses produced by group I phenolics. It was also found that the distribution ratio between the medium and the cell membrane, K1, calculated as well using the proposed model, is a good representative parameter for the cellular toxicity of the phenolic compounds according to their substituted groups when compared with the conventional method of using the octanol-water partition coefficient, log Kow. As a new parameter, the critical concentration was also shown to be a good representative of the cellular toxicity for group I phenolics to the Escherichia coli cells.