Testing oxidative-stress hypotheses generated by systems-biology assays

Role of yaaA in hydrogen peroxide stress response in Escherichia coli Hydrogen peroxide (H2O2) is commonly formed in microbial habitats by either chemical oxidation processes or host defense responses. H2O2 can penetrate cell membranes and damage key intracellular biomolecules, including DNA and certain groups of iron-dependent enzymes. Bacteria defend themselves against this H2O2 stress by inducing a set of genes that engages multiple defensive strategies. In Escherichia coli, this defense is regulated by a transcriptional factor, OxyR. A previous microarray study suggested that yaaA, an uncharacterized gene found in many bacteria, was induced by H2O2 in Escherichia coli as part of its OxyR regulon. Here I confirm that yaaA is a key element of H2O2 stress response. In a catalase/peroxidase-deficient (Hpx-) background, yaaA deletion mutants grew poorly, filamented extensively and lost substantial viability when they were cultured in aerobic LB medium. The results from a thyA forward mutagenesis assay and the growth defect of the yaaA deletion in DNA repair-deficient background indicated that yaaA mutants accumulated high levels of DNA damage. The growth defect of yaaA mutants was suppressed by either the addition of iron chelators or by mutations that slow iron import, suggesting that the DNA damage was caused by the Fenton reaction. Spin-trapping experiments confirmed that Hpx- yaaA cells had a higher hydroxyl radical level than yaaA+ strains. EPR analysis showed that the proximate cause was an unusually high level of intracellular unincorporated iron. These results demonstrate that during periods of H2O2 stress the induction of YaaA is a critical device to reduce intracellular iron levels; it thereby attenuates the Fenton reaction and the DNA damage that would otherwise result. The molecular mechanism of how YaaA does it remains unknown. Do antibiotics cause oxidative stress? Since 2007, several studies have reported that classic bactericidal antibiotics triggered the oxidation of an intracellular fluorescein probe, and this effect was attributed to hydroxyl radicals. Therefore, investigators hypothesized that different classes of antibiotics share a common bactericidal mechanism: antibiotics stimulate the formation of superoxide and H2O2, which damage certain iron-dependent enzymes, releasing iron to the cytosol. As a consequence, the higher levels of H2O2 and free iron drive the Fenton reaction to generate hydroxyl radicals which cause lethal DNA damage. If this model is true, then it might guide new antibiotic therapies. Therefore, I evaluated its possibility by testing specific steps of this hypothesis in Escherichia coli. I used three different antibiotics--kanamycin, ampicillin and norfloxacin--all of which have been shown to increase fluorescein-probe oxidation. However, contrary to the hypothesis, these antibiotics neither induced the peroxide-responsive OxyR regulon, nor accelerated endogenous H2O2 production. Furthermore, the metallo-enzymes that are known to be sensitive to superoxide and H2O2 were not damaged, and the level of intracellular free iron did not increase. Finally, the lethal effect of antibiotic persisted in the absence of oxygen, whereas DNA repair mutants were not hypersensitive, challenging the idea that toxicity arose from oxidative DNA lesions. Taken together, I conclude that bactericidal antibiotics did not generate reactive oxygen species to kill bacteria