Mutagenesis and stress responses induced in Escherichia coli by hydrogen peroxide.

Killing of Escherichia coli by hydrogen peroxide proceeds by two modes. Mode one killing appears to be due to DNA damage, has a maximum near 1 to 3 mM H2O2, and requires active metabolism during exposure. Mode two killing is due to uncharacterized damage, occurs in the absence of metabolism, and exhibits a classical multiple-order dose-response curve up to at least 50 mM H2O2 (J. A. Imlay and S. Linn, J. Bacteriol. 166:519-527, 1986). H2O2 induces the SOS response in proportion to the degree of killing by the mode one pathway, i.e., induction is maximal after exposure to 1 to 3 mM H2O2. Mutant strains that cannot induce the SOS regulon are hypersensitive to peroxide. Analysis of the sensitivities of mutants that are deficient in individual SOS-regulated functions suggested that the SOS-mediated protection is due to the enhanced synthesis of recA protein, which is rate limiting for recombinational DNA repair. Specifically, strains wholly blocked in both SOS induction and DNA recombination were no more sensitive than mutants that are blocked in only one of these two functions, and strains carrying mutations in uvrA, -B, -C, or -D, sfiA, umuC or -D, ssb, or dinA, -B, -D, -F, -G, -H, -I, or -J were not abnormally sensitive to killing by H2O2. After exposure to H2O2, mutagenesis and filamentation also occurred with the dose response characteristic of SOS induction and mode one killing, but these responses were not dependent on the lexA-regulated umuC mutagenesis or sfiA filamentation functions, respectively. Exposure of E. coli to H2O2 also resulted in the induction of functions under control of the oxyR regulon that enhance the scavenging of active oxygen species, thereby reducing the sensitivity to H2O2. Catalase levels increased 10-fold during this induction, and katE katG mutants, which totally lack catalase, while not abnormally sensitive to killing by H2O2 in the naive state, did not exhibit the induced protective response. Protection equal to that observed during oxyR induction could be achieved by the addition of catalase to cultures of naive cells in an amount equivalent to that induced by the oxyR response. Thus, the induction of catalase is necessary and sufficient for the observed oxyR-directed resistance to killing by H2O2. Although superoxide dismutase appeared to be uninvolved in this enhanced protective response, sodA sodB mutants, which totally lack superoxide dismutase, were especially sensitive to mode one killing by H2O2 in the naive state. gshB mutants, which lack glutathione, were not abnormally sensitive to killing by H2O2.