Mutagenesis and stress responses induced in Escherichia coli by hydrogen peroxide

Importance of anaerobic superoxide dismutase synthesis in facilitating outgrowth of Escherichia coli upon entry into an aerobic habitat

The manganese-containing isozyme of superoxide dismutase (MnSOD) is synthesized by Escherichia coli only during aerobiosis, in accordance with the fact that superoxide can be formed only in aerobic environments. In contrast, E. coli continues to synthesize the iron-containing isozyme (FeSOD) even in the absence of oxygen. A strain devoid of FeSOD exhibited no deficits during either anaerobic or continuously aerobic growth, but its growth lagged for 2 h during the transition from anaerobiosis to aerobiosis. Complementation of this defect with heterologous SODs established that anaerobic SOD synthesis per se is necessary to permit a smooth transition to aerobiosis. The growth deficit was eliminated by supplementation of the medium with branched-chain amino acids, indicating that the growth interruption was due to the established sensitivity of dihydroxyacid dehydratase to endogenous superoxide. Components of the anaerobic respiratory chain rapidly generated superoxide when exposed to oxygen in vitro, suggesting that this transition may be a period of acute oxidative stress. These results show that facultative bacteria must preemptively synthesize SOD during anaerobiosis in preparation for reaeration. The data suggest that evolution has chosen FeSOD for this function because of the relative availability of iron, in comparison to manganese, during anaerobiosis.

Comparative studies of S-layer proteins from Bacillus stearothermophilus strains expressed during growth in continuous culture under oxygen-limited and non-oxygen-limited conditions

The specific properties of S-layer proteins from three different Bacillus stearothermophilus strains revealing oblique, square, or hexagonal lattice symmetry were preserved during growth in continuous culture on complex medium only under oxygen-limited conditions in which glucose was used as the sole carbon source. When oxygen limitation was relieved, amino acids became metabolized, cell density increased, and different S-layer proteins from wild-type strains became rapidly replaced by a new common type of S-layer protein with an apparent subunit molecular weight of 97,000 which assembled into an identical oblique (p2) lattice type. During switching from wild-type strains to variants, patches of the S-layer lattices characteristics for wild-type strains, granular regions, and areas with oblique lattice symmetry could be observed on the surface of individual cells from all organisms. The granular regions apparently consisted of mixtures of the S-layer proteins from the wild-type strains and the newly synthesized p2 S-layer proteins from the variants. S-layer proteins from wild-type strains possessed identical N-terminal regions but led to quite different cleavage products upon peptide mapping, indicating that they are encoded by different genes. Chemical analysis including N-terminal sequencing and peptide mapping showed that the oblique S-layer lattices synthesized under increased oxygen supply were composed of identical protein species.

Involvement of Escherichia coli DNA polymerase II in response to oxidative damage and adaptive mutation

DNA polymerase II (Pol II) is regulated as part of the SOS response to DNA damage in Escherichia coli. We examined the participation of Pol II in the response to oxidative damage, adaptive mutation, and recombination. Cells lacking Pol II activity (polB delta 1 mutants) exhibited 5- to 10-fold-greater sensitivity to mode 1 killing by H2O2 compared with isogenic polB+ cells. Survival decreased by about 15-fold when polB mutants containing defective superoxide dismutase genes, sodA and sodB, were compared with polB+ sodA sodB mutants. Resistance to peroxide killing was restored following P1 transduction of polB cells to polB+ or by conjugation of polB cells with an F' plasmid carrying a copy of polB+. The rate at which Lac+ mutations arose in Lac- cells subjected to selection for lactose utilization, a phenomenon known as adaptive mutation, was increased threefold in polB backgrounds and returned to wild-type rates when polB cells were transduced to polB+. Following multiple passages of polB cells or prolonged starvation, a progressive loss of sensitivity to killing by peroxide was observed, suggesting that second-site suppressor mutations may be occurring with relatively high frequencies. The presence of suppressor mutations may account for the apparent lack of a mutant phenotype in earlier studies. A well-established polB strain, a dinA Mu d(Apr lac) fusion (GW1010), exhibited wild-type (Pol II+) sensitivity to killing by peroxide, consistent with the accumulation of second-site suppressor mutations. A high titer anti-Pol II polyclonal antibody was used to screen for the presence of Pol II in other bacteria and in the yeast Saccharomyces cerevisiae. Cross-reacting material was found in all gram-negative strains tested but was not detected in gram-positive strains or in S. cerevisiae. Induction of Pol II by nalidixic acid was observed in E. coli K-12, B, and C, in Shigella flexneri, and in Salmonella typhimurium.

Comparison of the sensitivities of Salmonella typhimurium oxyR and katG mutants to killing by human neutrophils

The respiratory burst of neutrophils is believed to kill bacteria by generating oxidative species, such as superoxide anion, hydrogen peroxide, and oxidized halogen species. The oxyR gene of Salmonella typhimurium controls a regulon induced by oxidative stress, such as exposure to hydrogen peroxide. Some researchers have suggested that oxyR may play a key role in bacterial survival following phagocytosis. We have tested this possibility by comparing the survival, following exposure to human neutrophils, of isogenic strains bearing different oxyR alleles. Neither inactivation of the oxyR gene nor constitutive overexpression of the oxyR-regulated proteins (oxyR1 allele) greatly alters bacterial resistance to neutrophils. The katG gene, encoding the oxyR-regulated enzyme hydroperoxidase I, was also without effect on survival following exposure to neutrophils. We conclude that the oxyR response does not play a significant role in the resistance of S. typhimurium to phagocytic killing in vitro.

Heat-shock-increased survival to far-UV radiation in Escherichia coli is wavelength dependent

Heat-shock-induced resistance to far-UV (FUV) radiation was studied in Escherichia coli. The induction of FUV resistance was shown to be dependent on the products of the genes uvrA and polA in bacteria irradiated at 254 nm. Heat shock increased the resistance to 280 nm radiation in a uvrA6 recA13 mutant. Heat shock lowered the mutation frequency (reversion to tryptophan proficiency) in wild-type or uvrA strains irradiated at 254 nm. When these strains were irradiated at 280 nm, heat shock did not interfere with the mutation frequency in the wild-type strain, but greatly enhanced mutations in the uvrA mutant. After heat-shock treatment, the wild-type strain irradiated at 254 nm showed increased DNA degradation, indicating enhanced repair activity. However, heat shock did not stimulate SOS repair triggered by FUV. An increased survival of bacteriophages irradiated with FUV and inoculated into heat-shock-treated bacteria was not detected. The possibility that heat shock enhances excision repair activity in a wavelength-dependent manner is discussed.

Effects of nerve growth factor on catalase and glutathione peroxidase in a hydrogen peroxide-resistant pheochromocytoma subclone

Stepwise selection in increasing H2O2 concentrations was used to obtain a PC12 cell variant designated HPR. This variant was stably resistant to H2O2 as compared with the parental PC12 cell line. HPR cells responded to nerve growth factor (NGF) by further enhancing H2O2 resistance. This variant was subcloned by limiting dilution to obtain the line referred to as HPR-C, which was stably resistant to H2O2 toxicity and retained NGF responses, including morphologic changes and further reduction of H2O2 toxicity. When compared with the parental PC12 line, the HPR-C subclone did not have higher levels of catalase or glutathione peroxidase (GSH Px) activity or mRNA expression (as assessed by PCR analysis of cDNA reverse transcribed from total cellular RNA). HPR-C cells retained the ability to respond to NGF treatment by increasing catalase and GSH Px activity and expression. These data suggest that the protective effects of conditioning lesions, unlike those of neurotrophins, are in part independent of changes in the activity or expression of antioxidant enzymes.

Differential effects of histidine on hydrogen peroxide-induced bacterial killing and DNA nicking in vitro

The hydrogen peroxide dose-response curves for Escherichia coli killing and DNA nicking in vitro display remarkably similar bimodal patterns. The concentrations of the oxidant resulting in maximum mode one killing, however, exceeds by two orders of magnitude those resulting in the mode one DNA nicking response. Addition of histidine differentially affects the experimental curves describing the dose-dependency for bacterial killing and DNA damage in vitro. Indeed, the lethal effect elicited by the oxidant in the presence of the amino acid is strictly concentration-dependent and thus the inactivation curve loses its bimodal character. In marked contrast, histidine abolishes DNA damage generated by low concentrations of hydrogen peroxide (< 100 microM) in the in vitro system (the mode one DNA nicking response) but greatly increases DNA damage produced by concentrations of the oxidant higher than 1 mM (the mode two DNA nicking response). Experimental results also suggest that treatment of covalently closed circular double-stranded super-coiled DNA with hydrogen peroxide, in the presence of both histidine and iron, may result in the formation of DNA double strand breakage, a type of lesion which is not efficiently produced by the oxidant in the absence of the amino acid. Taken together, the above results indicate that histidine differentially affects the in vitro DNA cleavage and E. coli lethality induced by hydrogen peroxide and suggest that different molecular events mediate mode one DNA nicking in vitro and mode one killing of bacterial cells.

Near-ultraviolet mutagenesis in superoxide dismutase-deficient strains of Escherichia coli

We compared mutagenic spectra induced by polychromatic near-ultraviolet radiation (near-UV; 300-400 nm) with superoxide anion (O2-) -dependent mutagenesis using a set of Escherichia coli tester strains. Near-UV radiation produced increased frequencies of G:C to A:T transitions, G:C to T:A and A:T to T:A transversions, and small increases in frameshift mutations in wild-type cells. Tester strains lacking superoxide dismutase (SOD) activity (sodAsodB double mutants) demonstrated high spontaneous mutation frequencies and increased near-UV sensitivity. The double mutants also showed increased mutations induced by near-UV compared to either isogenic wild type, sodA or sodB single mutants. Furthermore, these mutants had an unusual spontaneous mutation spectrum, with a predominance of A:T to T:A transversions, followed by G:C to T:A transversions and frameshifts generated in runs of adenines in both the +1 and -1 direction. Other frameshifts were detected to a lesser degree. The oxygen dependency and the type of mutations spontaneously induced in SOD-deficient cells indicated that this mutagenic spectrum was caused by oxidative DNA damage. However, no apparent synergistic action between near-UV radiation and an increased flux of O2- could be detected. From the frequency and types of mutations induced by the two agents, we speculate that near-UV-induced mutagenesis and O2--dependent mutagenesis involve, in part, different lesion(s) and/or mechanism(s). The nature and possible mutagenic pathways of each are discussed.

Escherichia coli produces linoleic acid during late stationary phase

Escherichia coli produces linoleic acid in the late stationary phase. This was the case whether the cultures were grown aerobically or anaerobically on a supplemented glucose-salts medium. The linoleic acid was detected by thin-layer chromatography and was measured as the methyl ester by gas chromatography. The linoleic acid methyl ester was identified by its mass spectrum. Lipids extracted from late-stationary-phase cells generated thiobarbituric acid-reactive carbonyl products when incubated with a free radical initiator. In contrast, extracts from log-phase or early-stationary-phase cells failed to do so, in accordance with the presence of polyunsaturated fatty acid only in the stationary-phase cells.

Characterization of a catalase-deficient strain of Neisseria gonorrhoeae: evidence for the significance of catalase in the biology of N. gonorrhoeae

We obtained a catalase-deficient (Kat-) strain of Neisseria gonorrhoeae isolated from a patient who had been unsuccessfully treated with penicillin. Quantitative enzyme assays and electrophoresis of cell extracts on native polyacrylamide gels subsequently stained for catalase and peroxidase activities failed to detect both enzymes. The strain exhibited no growth anomalies or unusual requirements when grown under ordinary laboratory conditions. However, the Kat- strain proved extremely sensitive to exogenous hydrogen peroxide, and analysis of the bacterial DNA after such exposure showed extensive single-strand breakage in both chromosomal and plasmid DNAs. Partial characterization of the gonococcal catalase from a Kat+ laboratory strain revealed that the enzyme had the physical and chemical properties of both catalase and peroxidase.

Bleaching: is it safe and effective?

It has been well documented that bleaching whitens teeth, but has its safety been documented? This paper reviews bleaching's predictability, esthetics, longevity, and side effects. A discussion of the bleaching reaction on teeth and soft tissue raises concerns over the safety of the procedure.

Multicellular oxidant defense in unicellular organisms

Although catalase is thought to be a major defense against hydrogen peroxide (H2O2), the catalase activity within individual Escherichia coli fails to protect against exogenous H2O2. Contrary to earlier reports, we find that dilute suspensions of wild-type and catalase-deficient E. coli are identical in their sensitivity to H2O2, perhaps because even wild-type, catalase-positive E. coli cannot maintain an internal/external concentration gradient of this highly diffusible oxidant. However, concentrated suspensions or colonies of catalase-positive E. coli do preferentially survive H2O2 challenge and can even cross-protect adjacent catalase-deficient organisms. Furthermore, high-density catalase-positive--but not catalase-negative--E. coli can survive and multiply in the presence of competitive, peroxide-generating streptococci. These observations support the concept that bacterial catalase may defend colonial, but not individual, E. coli against environmental H2O2. Group protection by the activity of enzymes that mitigate oxidative stress may have been a driving force in the evolution of multicellular organisms.

Regulation of katF and katE in Escherichia coli K-12 by weak acids

Chromosomal transcriptional and translational lacZ fusions to the katE (structural gene for the HPII hydroperoxidase) and katF (putative sigma factor required for katE expression) genes of Escherichia coli were isolated, and the regulation of these fusions was used to identify factors that control the expression of these two important antioxidant factors. While katE was found to be regulated primarily at the level of transcription (since induction patterns were similar for both transcriptional and translational fusions), katF expression was a function of both transcriptional and translational signals. The katE gene was induced 57-fold as cells entered the stationary phase, while katF was induced 23-fold. katF induction was coincident with katE induction and occurred at the onset of the stationary growth phase. Expression of both katE and katF could be induced by resuspending uninduced exponential-phase cells in spent culture supernatant recovered from stationary-phase cells. The component of stationary-phase culture supernatant responsible for induction of the katF regulon appeared to be acetate, since expression of both katE and katF fusions was induced when exponential-phase cells were exposed to this weak acid. Other weak acids, including propionate and benzoate, were also found to be effective inducers of expression of both katF and katE. Induction of katE and katF fusions was unaffected in merodiploid strains containing both mutant and wild-type alleles, indicating that expression of both genes is independent of the wild-type gene product. Examination of catalase zymograms prepared from cells exposed to various levels of acetate revealed that both HPI and HPII catalases are induced by this weak acid, suggesting that there is a common link in the regulation of these two enzymes.

Exogenous quinones directly inhibit the respiratory NADH dehydrogenase in Escherichia coli

The ability of naphthoquinones to generate reactive oxygen species has been widely exploited in studies of oxidative stress. However, excess superoxide dismutase and catalase failed to protect Escherichia coli in rich medium against growth inhibition by plumbagin, indicating that its toxic effect was not due to the production of partially reduced oxygen species. Respiration failed immediately upon the addition of growth-inhibitory levels of plumbagin. Studies in vitro showed that plumbagin and other redox-active quinones intercept electrons from NADH dehydrogenase, the primary respiratory dehydrogenase in glucose-containing media. An excess of oxidative substrate, such as plumbagin, inactivates this enzyme, which appears to be redox-regulated. The resultant respiratory arrest is a cautionary example of metabolic dysfunction from redox-cycling drugs that cannot be attributed to superoxide or hydrogen peroxide.

Prophage induction by DNA topoisomerase II poisons and reactive-oxygen species: role of DNA breaks

Various compounds were evaluated for their ability to induce prophage lambda in the Escherichia coli WP2s(lambda) microscreen assay. The inability of a DNA gyrase subunit B inhibitor (novobiocin) to induce prophage indicated that inhibition of the gyrase's ATPase was insufficient to elicit the SOS response. In contrast, poisons of DNA gyrase subunit A (nalidixic acid and oxolinic acid) were the most potent inducers of prophage among the agents examined here. This suggested that inhibition of the ligation function of subunit A, which also has a DNA nicking activity, likely resulted in DNA breaks that were available (as single-stranded DNA) to act as strong SOS-inducing signals, leading to prophage induction. Agents that both intercalated and produced reactive-oxygen species (the mammalian DNA topoisomerase II poisons, adriamycin, ellipticine, and m-AMSA) were the next most potent inducers of prophage. Agents that produced reactive-oxygen species only (hydrogen peroxide and paraquat) were less potent than adriamycin and ellipticine but more potent than m-AMSA. Agents that intercalated but did not generate reactive-oxygen species (actinomycin D) or that did neither (teniposide) were unable to induce prophage, suggesting that intercalation alone may be insufficient to induce prophage. These results illustrate the variety of mechanisms (and the relative effectiveness of these mechanisms) by which agents can induce prophage. Nonetheless, these agents may induce prophage by producing essentially the same type of DNA damage, i.e., DNA strand breaks. The potent genotoxicity of the DNA gyrase subunit A poisons illustrates the genotoxic consequences of perturbing an important DNA-protein complex such as that formed by DNA and DNA topoisomerase.

Construction of an Escherichia coli K-12 strain deleted for manganese and iron superoxide dismutase genes and its use in cloning the iron superoxide dismutase gene of Legionella pneumophila

An Escherichia coli K-12 strain deleted for sodA and sodB (manganese and iron superoxide dismutases) was constructed and characterized by Southern blotting, enzyme assays, and physiological analyses. The sod deletion strain was used to clone the iron superoxide dismutase gene of Legionella pneumophila by complementation to paraquat resistance.

Effects of overproduction of superoxide dismutases in Escherichia coli on inhibition of growth and on induction of glucose-6-phosphate dehydrogenase by paraquat

Stationary phase inocula were more susceptible to the growth inhibitory effect of paraquat than were log phase inocula and this difference was exacerbated in strains overproducing superoxide dismutases (SOD). Glucose-6-phosphate dehydrogenase (G-6-PD), a member of the soxR regulon, was induced by paraquat promptly in the case of log phase cells; but only after a lag in stationary phase cells and this difference was also exaggerated in strains overproducing SOD. The negative consequences of overproduction of SOD on the adaptation of stationary phase cells to paraquat may be attributed to competition for cellular resources with an attendant delay in biosynthesis of other components of soxR. Since overproduction of SOD did not prevent log phase cells from inducing G-6-PD in response to paraquat, it appears likely that soxR can respond to aspects of redox status other than O2-. This conclusion is in accord with data which is already in the literature.

Isolation and characterization of the Escherichia coli htrD gene, whose product is required for growth at high temperatures

Those genes in Escherichia coli defined by mutations which result in an inability to grow at high temperatures are designated htr, indicating a high temperature requirement. A new htr mutant of E. coli was isolated and characterized and is designated htrD. The htrD gene has been mapped to 19.3 min on the E. coli chromosome. Insertional inactivation of htrD with a mini-Tn10 element resulted in a pleiotropic phenotype characterized by a severe inhibition of growth at 42 degrees C and decreased survival at 50 degrees C in rich media. Furthermore, htrD cells were sensitive to H2O2. Growth rate analysis revealed that htrD cells grow very slowly in minimal media supplemented with amino acids. This inhibitory effect has been traced to the presence of cysteine in the growth medium. Further studies indicated that the rate of cysteine transport is higher in htrD cells relative to the wild type. All of these results, taken together, indicate that the htrD gene product may be required for proper regulation of intracellular cysteine levels and that an increased rate of cysteine transport greatly affects the growth characteristics of E. coli.

Suppression of oxidative envelope damage by pseudoreversion of a superoxide dismutase-deficient mutant of Escherichia coli

Mutants of Escherichia coli that are devoid of superoxide dismutase (SOD) fail to grow in aerobic minimal medium. This is largely because of the O2- sensitivities of several amino acid biosynthetic pathways, since amino acid supplements can restore growth, albeit at a slow rate. We now report that growth in amino acid-supplemented medium can be further stimulated by the presence of extracellular osmolytes. Osmolytes also partially suppress the amino acid requirements of the SOD mutant. These data suggest that the combination of oxidative injury and turgor pressure permeabilizes the cell envelope and that critical metabolites, including the limiting products of damaged biosynthetic pathways, escape from the cell. External osmolytes may offer protection by countervailing the usual turgor pressure and thus stabilizing the damaged envelope. This model is consistent with the previous observation that deficiency of cell wall components is lethal to SOD mutants. A pseudorevertant that can grow at a moderate rate in normosmotic medium without amino acid supplementation has been obtained (J. A. Imlay and I. Fridovich, Mol. Gen. Genet. 228:410-416, 1991). Analysis suggests that the suppressor mutation allows the envelope either to resist or to tolerate oxidative lesions. Study of the pseudorevertant may illuminate the molecular basis of this oxidative envelope injury.

The role of extracellular medium components and specific amino acids in the cytotoxic response of Escherichia coli and Chinese hamster ovary cells to hydrogen peroxide

A concentration of H2O2 resulting in mode one killing of Escherichia coli is more toxic when exposure to the oxidant is performed in complete medium (K medium), as compared to a saline (M9 salts). Inorganic salts (MgSO4 and CaCl2), thiamine or glucose, when added separately, or combined, to M9 salts had no effect on the cytotoxic response to H2O2. In contrast, the lethality of the oxidant was highly dependent on the presence of the amino acids in the incubation medium. The addition of glucose further enhanced this response. Among the seventeen amino acids which are present in the complete amino acid mixture, only two, i.e. L-histidine and L-cystine, were found to increase the toxicity of H2O2. Again, glucose augmented this response. The effect of these amino acids on the growth inhibitory action of hydrogen peroxide was also tested in Chinese Hamster Ovary cells. It was found that L-histidine was capable of increasing the toxicity of the oxidant whereas all the other amino acids did not affect the toxicity of the oxidant. Glucose only slightly augmented this effect of L-histidine. DNA single strand breakage produced by H2O2 was increased by L-histidine and was not significantly modified by the other amino acids. DNA double strand breakage was also shown to occur in cells exposed to H2O2-L-histidine, and this effect was independent on the presence of glucose. These results demonstrate that the cytotoxic response of bacterial and mammalian cells to challenge with H2O2 is highly dependent on the composition of the extracellular milieu.(ABSTRACT TRUNCATED AT 250 WORDS)

Microbial strategies to prevent oxygen-dependent killing by phagocytes

Microorganisms which are taken up by professional phagocytic cells of a host organism (e.g., by macrophages and polymorphonuclear leukocytes) encounter a series of antimicrobial events including confrontation with toxic oxygen species, derived mainly from the superoxide radical produced by phagocytic NADPH oxidase after uptake of the microorganism. Many microbes are susceptible to the oxygen-dependent phagocytic stress and are efficiently killed. The strategies of some microorganisms to bypass an encounter with the phagocytes' reactive oxygen species, and biochemical systems contributing to the microbes' resistance to killing by reactive oxygen species are outlined.

Oxidative stress responses in Escherichia coli and Salmonella typhimurium

Oxidative stress is strongly implicated in a number of diseases, such as rheumatoid arthritis, inflammatory bowel disorders, and atherosclerosis, and its emerging as one of the most important causative agents of mutagenesis, tumorigenesis, and aging. Recent progress on the genetics and molecular biology of the cellular responses to oxidative stress, primarily in Escherichia coli and Salmonella typhimurium, is summarized. Bacteria respond to oxidative stress by invoking two distinct stress responses, the peroxide stimulon and the superoxide stimulon, depending on whether the stress is mediated by peroxides or the superoxide anion. The two stimulons each contain a set of more than 30 genes. The expression of a subset of genes in each stimulon is under the control of a positive regulatory element; these genes constitute the OxyR and SoxRS regulons. The schemes of regulation of the two regulons by their respective regulators are reviewed in detail, and the overlaps of these regulons with other stress responses such as the heat shock and SOS responses are discussed. The products of Oxy-R- and SoxRS-regulated genes, such as catalases and superoxide dismutases, are involved in the prevention of oxidative damage, whereas others, such as endonuclease IV, play a role in the repair of oxidative damage. The potential roles of these and other gene products in the defense against oxidative damage in DNA, proteins, and membranes are discussed in detail. A brief discussion of the similarities and differences between oxidative stress responses in bacteria and eukaryotic organisms concludes this review.

Lethal and mutagenic action of hydrogen peroxide on Haemophilus influenzae

The lethal and mutagenic effects of H2O2 on wild-type Haemophilus influenzae Rd and on uvr1, uvr2, rec1, and rec2 mutant strains were studied. The first two mutants are sensitive to UV, and the second two are defective in recombination. Rd, urv1, and rec1 strains were more sensitive to the killing effect of H2O2 treatment than were uvr2 and rec2 strains. There were peaks of mutagenesis at two H2O2 concentrations over a range of 30 to 275 mM. Our results suggest a specific repair of H2O2 damage that is independent of the Uvr2 and Rec2 gene products. Sensitivity to the killing effect of H2O2 and to the lethal action of near-UV light were similar for Rd and uvr1 strains. This finding suggests that the mechanisms of killing by and repair of H2O2 damage may have some overlap with those of near-UV radiation.

Oxidative stress effects on conjugational recombination and mutation in catalase-deficient Escherichia coli

The objective of the present investigation was to determine the effects on genetic recombination and mutation in Escherichia coli of either endogenous increases in oxygen radicals resulting from catalase deficiencies, or exogenous increases resulting from H2O2 treatment. Using the classical paradigm of Escherichia coli bacterial conjugation, strains deficient in the production of hydroperoxidase I (HPI) and/or hydroperoxidase II (HPII) were used as recipients in Hfr x F- matings. 'Background' recombination rates, measured by the rate of appearance of threonine prototrophs, was similar to wild-type levels in the HPI-deficient (katG) strain, but were significantly decreased in HPII- (katE) mutants. The addition of relatively nontoxic H2O2 concentrations (0.25 mmoles dm-3) to the mating mixtures stimulated recombination rates in wild-type and katE strains, but decreased rates in katG and katEkatG strains. A 0.5 mmoles dm-3 concentration of H2O2 inhibited recombination rates in all strains. In order to gauge the level of recA-dependent 'SOS' processes occurring under the experimental conditions, 'background' mutation rates were determined in both fluctuation and forward mutation (thyA) assays. Mutation rates in aerobically-grown cultures were increased up to 2.2-fold in katG and katEkatG strains. Treatment with relatively nontoxic H2O2 concentrations elevated the thyA mutagenesis up to 8-fold in catalase-deficient cultures. Furthermore, these studies along with data presented elsewhere show that the SOS phenotype of katEkatG is more resistant than that of katG strains. These studies clearly show that cellular oxidative stress occurring from catalase deficiency interferes with normal DNA metabolism.

Isolation and genetic analysis of a mutation that suppresses the auxotrophies of superoxide dismutase-deficient Escherichia coli K12

The most striking phenotype associated with superoxide dismutase (SOD) deficiency in Escherichia coli is the inability to grow in aerobic minimal medium, which is due to the sensitivity of several amino acid biosynthetic pathways to superoxide. We have isolated two classes of pseudorevertants that grow on minimal medium at modest rates. Of these, the class that exhibited the faster growth carries mutations at a single locus, denoted ssa, which was mapped to 4 min on the E. coli chromosome. This class constituted the majority of the spontaneous pseudorevertants that were selected by the transfer of independent SOD-deficient cultures in minimal medium from anaerobic to aerobic growth conditions. Pseudoreversion at ssa suppressed requirements for a variety of unrelated amino acid supplements. Further, the SOD-deficient strains were unable to assimilate diaminopimelic acid from the growth medium, whereas the ssa pseudorevertants did so. The viability of these pseudorevertants indicates that superoxide-sensitive biosynthetic enzymes do retain some function in SOD-deficient cells during aerobic growth.

Penicillin-binding protein inactivation by human neutrophil myeloperoxidase

Myeloperoxidase (MPO), H2O2, and chloride comprise a potent antimicrobial system believed to contribute to the antimicrobial functions of neutrophils and monocytes. The mechanisms of microbicidal action are complex and not fully defined. This report describes the MPO-mediated inactivation, in Escherichia coli, Staphylococcus aureus, and Pseudomonas aeruginosa, of a class of cytoplasmic membrane enzymes (penicillin-binding proteins, PBPs) found in all eubacteria, that covalently bind beta-lactam antibiotics to their active sites with loss of enzymatic activity. Inactivation of "essential" PBPs, including PBP1-PBP3 of E. coli, leads to unbalanced bacterial growth and cell death. MPO treatment of bacteria was associated with loss of penicillin binding by PBPs, strongly suggesting PBP inactivation. In E. coli, PBP inactivation was most rapid with PBP3, where the rate of decline in binding activity approximated but did not equal loss of viability. Changes in E. coli morphology (elongation), observed just before bacteriolysis, were consistent with early predominant inactivation of PBP3. We conclude that inactivation of essential PBPs is sufficient to account for an important fraction of MPO-mediated bacterial action. This feature of MPO action interestingly recapitulates an antibacterial strategy evolved by beta-lactam-producing molds that must compete with bacteria for limited ecologic niches.

Hydrogen peroxide as an electron acceptor for mitochondrial respiration in the yeast Hansenula polymorpha

Chemostat cultures of a catalase-negative mutant of Hansenula polymorpha CBS 4732 were able to decompose hydrogen peroxide at a high rate. This was apparent from experiments in which the yeast was grown under carbon limitation in chemostat culture on mixtures of glucose and H2O2. The enzyme responsible for H2O2 degradation is probably the mitochondrial enzyme cytochrome c peroxidase (CCP), which was present at very high activities. This enzyme was partially purified and shown to be specific for reduced cytochrome c as an electron donor; no reaction was observed with NAD(P)H. Thus, reducing equivalents for H2O2 degradation by CCP must be provided by the respiratory chain. That H2O2 can act as an electron acceptor for reducing equivalents could be confirmed with experiments in which cells were incubated with ethanol and H2O2 in the absence of oxygen. This resulted in oxidation of ethanol to equimolar amounts of acetate. Energetic aspects of mitochondrial H2O2 decomposition via CCP and the physiological function of CCP in yeasts are discussed.

Quinolinate synthetase: the oxygen-sensitive site of de novo NAD(P)+ biosynthesis

The ability of niacin to relieve the growth-inhibiting effect of hyperoxia on Escherichia coli can be attributed to the dioxygen sensitivity of quinolinate synthetase. The activity of this enzyme within E. coli was diminished by exposure of the cells to 4.2 atm O2, while the activity in extracts was rapidly decreased by 0.2 atm O2. Neither catalase nor superoxide dismutase afforded detectable protection against the inactivating effect of O2, indicating that H2O2 and O2- were not significant intermediates in this process. Nevertheless, H2O2 at 1.0 mM did inactivate quinolinate synthetase, even under anaerobic conditions and in the absence of catalatic activity which might have generated O2. Addition of paraquat to aerobic cultures of E. coli caused an inactivation of quinolinate synthetase, which may be explained in terms of an increase in the production of H2O2. The O2-dependent inactivation of quinolinate synthetase in extracts was gradually reversed during anaerobic incubation and this reactivation was blocked by alpha, alpha'-dipyridyl or by 1,10-phenanthroline. The sequence of the quinolinate synthetase "A" protein contains a--cys-w-x-cys-y-z-cys--sequence, which is characteristic of (Fe-S)4-containing proteins. This sequence, together with the effect of the Fe(II)-chelating agents, suggests that the O2-sensitive site of quinolinate synthetase is an iron-sulfur cluster which is essential for the dehydration reaction catalyzed by the A protein.

Characterization of an inducible oxidative stress system in Bacillus subtilis

Exponentially growing cells of Bacillus subtilis demonstrated inducible protection against killing by hydrogen peroxide when prechallenged with a nonlethal dose of this oxidative agent. Cells deficient in a functional recE+ gene product were as much as 100 times more sensitive to the H2O2 but still exhibited an inducible protective response. Exposure to hydrogen peroxide also induced the recE(+)-dependent DNA damage-inducible (din) genes, the resident prophage, and the product of the recE+ gene itself. Thus hydrogen peroxide is capable of inducing the SOS-like or SOB system of B. subtilis. However, the induction of this DNA repair system by other DNA-damaging agents is not sufficient to activate the protective response to hydrogen peroxide. Therefore, at least one more regulatory network (besides the SOB system) that responds to oxidative stress must exist. Furthermore, the data presented indicate that a functional catalase gene is necessary for this protective response.

Influence of DNA adenine methylase on the sensitivity of Escherichia coli to near-ultraviolet radiation and hydrogen peroxide

Near-ultraviolet (NUV) radiation and hydrogen peroxide (H2O2) inactivation studies were performed on Escherichia coli K-12 DNA adenine methylation (dam) mutants and on cells that carry plasmids which overexpress Dam methylase. Lack of methylation resulted in increased sensitivity to NUV and H2O2 (a photoproduct of NUV). In a dam mutant carrying a dam plasmid, the levels of Dam enzyme and resistance to NUV and H2O2 were restored. However, using a multicopy dam+ plasmid strain, increasing the methylase above wildtype levels resulted in an increase in sensitivity of the cells rather than resistance.

Transcriptional regulator of oxidative stress-inducible genes: direct activation by oxidation

The oxyR gene positively regulates genes induced by oxidative stress in Salmonella typhimurium and Escherichia coli. Purification of the OxyR protein showed that oxidized but not reduced OxyR activates transcription of oxidative stress-inducible genes in vitro. Conversion between the two forms of OxyR is rapid and reversible. Both the oxidized and the reduced forms of the OxyR protein are capable of binding to three diverse sequences upstream of OxyR-regulated promoters, but the interactions of the two forms of OxyR with the promoter regions are different. The results suggest that direct oxidation of the OxyR protein brings about a conformational change by which OxyR transduces an oxidative stress signal to RNA polymerase.

Thiol and hydrogen peroxide modification of recA induction in UV-irradiated wild-type and catalase-deficient Escherichia coli K12

Induction of recA in Escherichia coli, monitored as beta-D-galactosidase (beta-Gal) activity in recA-lacZ fusion strains, was shown to be elevated and prolonged by dithiothreitol (DTT) treatment after far-UV radiation. Pretreatment of UV-irradiated cultures using DTT led to a shortened recA response and little increase of beta-Gal yield. Similar studies were performed using a catalase-deficient recA-lacZ strain in which the major feature was elevated levels of recA-lacZ induction. Catalase activity in UV-irradiated wild-type cells was reduced by DTT treatment to levels as low as in a katE mutant strain, leading to similar recA-lacZ induction patterns between the strains. Neither DTT nor H2O2 treatment of cells could induce significant recA transcription in the absence of UV-radiation, implying that both agents modify recA activity primarily by interfering with repair of recA-inducing DNA lesions. The results confirm previous studies suggesting that modification of DNA repair is probably a significant portion of thiol radiation protection.

Microbial stress proteins

There is general agreement that a function, perhaps the major function, of stress proteins under normal physiological conditions is to help assembly and disassembly of protein complexes and to catalyse protein-translocation processes. It remains unclear, however, as to what role these processes play in stressed cells. It could be that cells under stress produce abnormal, misfolded or otherwise damaged proteins and that increased synthesis of stress proteins is required to counter protein modifications. A role for stress proteins in recovery of cells from stress, as opposed to a role in helping cells to withstand a lethal stress, is thus suggested. The intracellular location of stress proteins, in the unstressed and stressed cell, is worthy of further studies. Members of the hsp70 family are associated with the cytosol, mitochondria and endoplasmic reticulum. There is evidence, particularly from studies on mammalian cells (Tanguay, 1985; Welch and Mizzen, 1988; Arrigo et al., 1988), that following stress hsps migrate to various cellular compartments and subsequently delocalize after stress. However, there is little comparable data from microbial systems for this phenomenon (e.g. Rossi and Lindquist, 1989). The question as to the role of stress proteins in the transient acquisition of thermotolerance remains to be answered. It is insufficient to equate the kinetics of stress-protein synthesis with acquisition of thermotolerance. Quantitative data on the amount of stress protein present at various times, including the recovery period, is required. The demonstration that microbial stress proteins are important antigenic determinants of micro-organisms causing major debilitating diseases in the world is an exciting observation. Studies on the interplay of pathogen and host, both carrying similar antigenic hsp determinants, will be a challenging area for future research. It is likely that E. coli and Sacch. cerevisiae, with their well-established biochemical and genetic properties, will continue to be the experimental systems of choice for studies on stress proteins. On the other hand, it is encouraging that studies on other micro-organisms have expanded in the past few years and have made substantial contributions towards our understanding of the stress response. The ubiquitous nature of the stress response and the remarkable evolutionary conservation of the stress proteins continue to be attractive areas for research.

Morphological changes in Escherichia coli cells exposed to low or high concentrations of hydrogen peroxide

Escherichia coli cells challenged with low or high concentrations of hydrogen peroxide are killed via two different mechanisms and respond with morphological changes which are also dependent on the extracellular concentration of the oxidant. Treatment with low concentrations (less than 2.5 mM) of H2O2 is followed by an extensive cell filamentation which is dependent on the level of H2O2 or the time of exposure. In particular, addition of 1.75 mM H2O2 results in a growth lag of approximately 90 min followed by partial increase in optical density, which was mainly due to the onset of the filamentous response. In fact, microscopic analysis of the samples obtained from cultures incubated with the oxidant for various time intervals has revealed that this change in morphology becomes apparent after 90 min of exposure to H2O2 and that the length of the filaments gradually increases following longer time intervals. Analysis of the ability of these cells to form colonies has indicated a loss in viability in the first 90 min of exposure followed by a gradual recovery in the number of cells capable of forming colonies. Measurement of lactate dehydrogenase in culture medium (as a marker for membrane damage) has revealed that a small amount of this enzyme was released from the cells at early times (less than 150 min) but not after longer incubation periods (300 min). Cells exposed to high concentrations of H2O2 (greater than 10 mM) do not filament and their loss of viability is associated with a marked reduction in cell volume. In fact, treatment with 17.5 mM H2O2 resulted in a time-dependent decrease of the optical density, clonogenicity, and cellular volume. In addition, these effects were paralleled by a significant release in the culture medium of lactate dehydrogenase thus suggesting that the reduced cell volume may be dependent on membrane damage followed by loss of intracellular material. This hypothesis is supported by preliminary results obtained in electron microscopy studies. In conclusion, this study further demonstrates that the response of E. coli to hydrogen peroxide is highly dependent on the concentration of H2O2 and further stresses the point that low or high concentrations of the oxidant result in the production of different species leading to cell death via two different mechanisms and/or capable of specifically affecting the cell shape.

Induction of the SOS response by hydrogen peroxide in various Escherichia coli mutants with altered protection against oxidative DNA damage

The induction of the SOS response by H2O2 was measured in Escherichia coli by means of a sfiA::lacZ operon fusion. The effects of mutations in genes involved in DNA repair or DNA metabolism on the SOS response were investigated. We found that in an uvrA mutant, H2O2 induced the SOS response at lower concentrations than in the uvr+ parent strain, indicating that some lesions induced by H2O2 may be repaired by the uvrABC-dependent excision repair system. A nth mutation, yielding deficiency in thymine glycol DNA glycosylase, had no detectable effect on SOS induction, indicating that thymine glycol, a DNA lesion expected to be induced by H2O2, does not participate detectably in the induction of the SOS response by this chemical under our conditions. H2O2 still induced the SOS response in a dnaC(Ts) uvrA double mutant under conditions in which no DNA replication proceeds, suggesting that this chemical induces DNA strand breaks. Induction of the SOS response by H2O2 was also assayed in various mutants affected in genes suspected to be important for protection against oxidative stress. Mutations in the catalase genes, katE and katG, had only minor effects. However, in an oxyR deletion mutant, in which the adaptative response to H2O2 does not occur, SOS induction occurred at much lower H2O2 concentrations than in the oxyR+ parent strain. These results indicate that some enzymes regulated by the oxyR gene are, under our conditions, more important than catalase for protection against the H2O2-induced DNA damages which trigger the SOS response.

The spectrum of mutations generated by passage of a hydrogen peroxide damaged shuttle vector plasmid through a mammalian host

Treatment of a plasmid shuttle vector (pZ189) with a combination of hydrogen peroxide and a ferric iron/EDTA complex prior to transfection and passage in simian (CV-1) cells increases the frequency of mutations at the supF locus by up to 60-fold over the spontaneous background. This increase in mutation frequency is abolished when the inhibitors desferrioxamine, superoxide dismutase, catalase or dimethyl sulfoxide are included in the initial reaction or when the iron/EDTA complex is omitted, a strong indication that the premutagenic damage arises as a result of direct attack by hydroxyl radical generated in a superoxide driven Fenton reaction. DNA sequence analysis of the mutated plasmids shows that 1) Deletions occuring in combination with base-substitutions arise in 22.5 percent of the induced mutants compared with only 3 percent of spontaneous mutants 2) Sixty percent of all induced deletion mutations involve the loss of a single base and 77 percent of these (20 out of 26) occur at two adenine-containing sites 3) The base-change spectrum of mutants arising in the treated plasmid population is marked by the predominance of mutants containing a single base-change and by an increase in changes at AT base pairs. These results provide direct information concerning the nature of mutations arising in mammalian cells as a result of hydroxyl radical mediated DNA damage.

Cloning, genetic characterization, and nucleotide sequence of the hemA-prfA operon of Salmonella typhimurium

The first step in heme biosynthesis is the formation of 5-aminolevulinic acid (ALA). Mutations in two genes, hemA and hemL, result in auxotrophy for ALA in Salmonella typhimurium, but the roles played by these genes and the mechanism of ALA synthesis are not understood. I have cloned and sequenced the S. typhimurium hemA gene. The predicted polypeptide sequence for the HemA protein shows no similarity to known ALA synthases, and no ALA synthase activity was detected in extracts prepared from strains carrying the cloned hemA gene. Genetic analysis, DNA sequencing of amber mutations, and maxicell studies proved that the open reading frame identified in the DNA sequence encodes HemA. Another surprising finding of this study is that hemA lies directly upstream of prfA, which encodes peptide chain release factor 1 (RF-1). A hemA::Kan insertion mutation, constructed in vitro, was transferred to the chromosome and used to show that these two genes form an operon. The hemA gene ends with an amber codon, recognized by RF-1. I suggest a model for autogenous control of prfA expression by translation reinitiation.

Escherichia coli proteins inducible by oxidative stress mediated by the superoxide radical

Two-dimensional gel analyses were made of proteins synthesized in Escherichia coli during various O2- -generating conditions. Nine proteins were constitutively synthesized over wild-type levels in superoxide dismutase (sodA sodB) double mutants. Addition of redox cycling agents such as paraquat and plumbagin at various concentrations induced up to 13 proteins in wild-type cells. Among these 13 were 5 of the 9 constitutively synthesized in the sodA sodB double mutants. Addition of these agents to the superoxide dismutase mutants in low micromolar concentrations induced an additional set of 14 proteins. The proteins induced included only five proteins that have been previously associated with stress responses, consisting of endonuclease IV (Nfo), three oxyR-regulated proteins, and one heat shock protein. O2- -mediated induction of the superoxide inducible (Soi) proteins in the wild type was independent of the oxyR+ gene for all but the three oxyR-regulated proteins. Analyses of proteins from three soi::lacZ gene fusions previously isolated (T. Kogoma, S. B. Farr, K. M. Joyce, and D. O. Natvig, Proc. Natl. Acad. Sci. USA 85:4799-4803, 1988) indicated the specific loss of one of these induced proteins in each fusion strain and the constitutive expression of some Soi proteins.

Multiple pathways for repair of hydrogen peroxide-induced DNA damage in Escherichia coli

The repair response of Escherichia coli to hydrogen peroxide has been examined in mutants which show increased sensitivity to this agent. Four mutants were found to show increased in vivo sensitivity to hydrogen peroxide compared with wild type. These mutants, in order of increasing sensitivity, were recA, polC, xthA, and polA. The polA mutants were the most sensitive, implying that DNA polymerase I is required for any repair of hydrogen peroxide damage. Measurement of repair synthesis after hydrogen peroxide treatment demonstrated normal levels for recA mutants, a small amount for xthA mutants, and none for polA mutants. This is consistent with exonuclease III being required for part of the repair synthesis seen, while DNA polymerase I is strictly required for all repair synthesis. Sedimentation analysis of cellular DNA after hydrogen peroxide treatment showed that reformation was absent in xthA, polA, and polC(Ts) strains but normal in a recA cell line. By use of a lambda phage carrying a recA-lacZ fusion, we found hydrogen peroxide does not induce the recA promoter. Our findings indicate two pathways of repair for hydrogen peroxide-induced DNA damage. One of these pathways would utilize exonuclease III, DNA polymerase III, and DNA polymerase I, while the other would be DNA polymerase I dependent. The RecA protein seems to have little or no direct function in either repair pathway.