Salmonella typhimurium's Adaptation: Catalase Induction & Stress Resistance, Lab Reports of Pharmacology

Download Salmonella typhimurium's Adaptation: Catalase Induction & Stress Resistance and more Lab Reports Pharmacology in PDF only on Docsity! Cell, Vol. 41, 753-762, July 1985, Copyright © 1985 by MIT 0092-8674./85/070753-10 $02.00/0 Positive Control of a Regulon for Defenses against Oxidative Stress and Some Heat-Shock Proteins in Salmonella typhimurium Michael F. Christman, Robin W. Morgan, Fredric S. Jacobson, and Bruce N. Ames Department of Biochemistry University of California Berkeley, California 94720 Summary S. typhimurium become resistant to killing by hydro- gen peroxide and other oxidants when pretreated with nonlethal levels of hydrogen peroxide. During adapta- tion to hydrogen peroxide, 30 proteins are induced. Nine are constitutively overexpressed in dominant hydrogen peroxide-resistant oxyR mutants. Mutant oxyR1 is resistant to a variety of oxidizing agents and overexpresses at least five enzyme activities involved in defenses against oxidative damage, Deletions of oxyR are recessive and uninducible by hydrogen peroxide for the nine proteins overexpressed in oxyR1, demonstrating that oxyR is a positive regulatory ele- ment. The oxyR1 mutant is also more resistant than the wild-type parent to killing by heat, and it constitu- tively overexpresses three heat-shock proteins. The oxyR regulatory network is a previously uncharacter- ized global regulatory system in enteric bacteria. levels of catalase activity have been described in E. coli (Loewen, 1984) and in S. typhimurium (Levine, 1977), none have been demonstrated to be regulatory. No mu- tants altered in superoxide dismutase activity have been described. In previous studies, we have shown a variety of oxida- tive stresses and heat-shock to induce rapidly the ac- cumulation of AppppA and of a series of related adeny- lylated nucleotides (Bochner et al., 1984; Lee et al., 1983). These dinucleotides may be alarmones (Stephens et al., 1975) the function of which is to redirect cellular metabo- lism to cope with oxidative stress. Hydrogen peroxide in- duces the synthesis of adenylylated nucleotides and is a useful compound for studying cellular responses to reac- tive oxygen species because it is more stable than su- peroxide anion, singlet oxygen, and the hydroxyl radical. E. coli is capable of adapting to hydrogen peroxide (Dem- pie and. Halbrook, 1983) and, recently, it has been reported that S. typhimurium also can adapt to hydrogen peroxide (Winquist et al., 1984). As a first step toward understanding the regulation of defenses against oxida- tive stress, we present here a genetic and biochemical analysis of the adaptation to hydrogen peroxide in S. typhimurium. Results Introduction All organisms that use molecular oxygen must defend themselves from the toxic byproducts of oxygen metabo- lism. In addition to the complete four electron reduction of molecular oxygen to water, reactive species such as hydrogen peroxide, superoxide anion, singlet oxygen, and the hydroxyl radical can be generated in vivo during respi- ration (Fridovich, 1978; Chance et al., 1979). These reac- tive oxygen species can oxidize membrane fatty acids ini- tiating lipid peroxidation (Mead, 1976), oxidize proteins (Brot et al., 1981), and damage DNA (Demple and Linn, 1982; Levin et al., 1982; Hollstein et al., 1984; Cathcart et al., 1984). Enteric bacteria have several enzyme activities that may protect the cells from oxidative damage. These in- clude superoxide dismutase and catalase (Fridovich, 1976). In addition, other enzymes such as exonuclease III (Demple et al., 1983) and recA protein (Carlsson and Car- penter, 1980) appear to be important in repairing DNA le- sions resulting from oxidative damage. The induction of catalase and superoxide dismutase in E. coli during anaerobic to aerobic shifts has been stud- ied extensively by Hassan and Fridovich (Hassan and Fridovich, 1977a; 1977b; 1977c; 1978). Treatment with hydrogen peroxide has been shown to induce catalase ac- tivity both in S. typhimurium (Finn and Condon, 1975) and in E. coli (Richter and Loewen, 1981). Despite the in- ducibility of these activities, little is known about the regu- latory mechanisms involved. Although mutants with altered Adaptation to Hydrogen Peroxide in S. typhimurium Exponentially growing cells of Salmonella typhimurium treated with 60/~M hydrogen peroxide for 60 min are resis- tant to killing by 10 mM hydrogen peroxide when com- pared with cells that have not been pretreated (Figure 1). Cells pretreated with 60/~M hydrogen peroxide in the pres- ence of chloramphenicol do not acquire the resistance to killing by 10 rnM hydrogen peroxide, indicating a require- ment for de novo protein synthesis for the adaptation. The adaptation to hydrogen peroxide in S. typhimurium occurs under conditions similar to those previously described for Escherichia coli (Demple and Halbrook, 1983). As in E. coil, the adaptation is recA-independent (data not shown). Hydrogen peroxide-damaged P22 bacteriophage were not reactivated in S. typhimurium cells adapted to hydrogen peroxide (data not shown), in contrast to the reported finding of reactivation of P1 bacteriophage in adapted E. coil (Demple and Halbrook, 1983). It has been observed previously that S. typhimurium cells induce catalase activity upon exposure to low levels of hydrogen peroxide (Finn and Condon, 1975) in agreement with our results, which show that catalase activity in exponential cultures is induced 4- to 5-fold (from 0.09 units/A650 to 0.44 units/A6~0) after 1 hr of exposure in minimal glucose medium to 60/~M hydrogen peroxide. Hydrogen Peroxide Adaptation Induces Resistance to Killing by Other Stresses S. typhimurium cells that have been exposed to the adap- tive dose of hydrogen peroxide (60/~M) for 1 hr are resis- Cell 754 treated ~ 0 ~ 1 retreated O ° '~ 1 0 / • .. pret er a t e d / ~ . ~ / + chloramphenicol ~ ! 1 I I 0 20 40 60 Minutes after 1OmM H 2 0 2 Challenge Figure 1. Adaptation to Hydrogen Peroxide in Salmonella typhi- murium Wild-type (S. typhimurium LT2) cells were grown in minimal glucose medium at 37°C with shaking to an Ass0 of 0.2 and were then treated with 60 ~M hydrogen peroxide for 60 rain in the presence or absence of 100/~g/ml chloramphenicol. After 60 rain, 10 mM hydrogen peroxide was added to all cultures. Aliquots were withdrawn, diluted in minimal salts, and plated on nutrient broth plates at 37°C to monitor cell viability. Cells pretreated with 60/~M hydrogen peroxide in the presence (rq), and absence (©), of 100 ~g/ml chloramphenicol. No treatment: (O). tant to killing both by a variety of chemical oxidants and by heating at 50°C (Figure 2). N-ethylmaleimide is a non- specific thiol inactivating reagent, chlorodinitrobenzene depletes glutathione, and quinones such as menadione cause redox cycling, which can generate oxygen radicals (Chesis et al., 1984). Heat shock and these chemical oxi- dants all elevate levels of AppppA and other dinucleotides in S. typhimurium (Bochner et al., 1984). Although the mechanism of bacterial heat killing is not known, our results suggest a connection between oxidative stresses and heat stress (Bochner et al., 1984). The responses to these stresses are not identical, however, as heat shock induces thermotolerance (Yamamori and Yura, 1982) but not resistance to hydrogen peroxide (data not shown). Proteins Induced by Hydrogen Peroxide Two-dimensional protein gel analysis of S. typhimurium cells treated with 60/~M hydrogen peroxide revealed that the rate of synthesis of 30 proteins is elevated in the 60 min following the addition of hydrogen peroxide (Morgan et al., unpublished data). This dramatic change in the pat- tern of protein synthesis has been consistently observed through many different experiments. Cells were pulse- labeled with L-[3sS]methionine for 10 min periods begin- ning at various times after the addition of hydrogen perox- ide. These time-course experiments revealed two major temporal classes of hydrogen peroxide-inducible proteins (Figure 3). The first class consists of the "early" proteins; those for which synthesis is maximal during the first 10 Heating at 50°C 100 o i 100 10 1 1 101 0.1 0.1 0.01 j 0.01 0 1 2 1 -chloro-2,4-dinitrobenzene 100 100 1 ~ 10 0. 0.1 0.01 I 0.01 10 20 t i m e (mJn) pretreated with 60~M H202 e - - e no pretreatment N-ethylmaleimide i 10 20 Menadione 1 ~ " ~ " ~ i 10 20 Figure 2. Hydrogen Peroxide-Induced Resistance to Other Stresses Wild-type cells were grown in VBC salts containing 0.4% glucose at 37°C with shaking to an A6~0 of 0.2 and were then treated with N-ethyl- maleimide (1 mM), 1-chloro-2,4-dinitrobenzene (10 mM), or menadione (50 mM). Aliquots were withdrawn at the indicated times, diluted in VBC salts, and plated on nutrient broth plates at 37°C to monitor cell viability. For heat killing, an aliquot (0.2 ml) was withdrawn and heated at 50°C in a heating block for the indicated times. Heated samples were withdrawn, cooled on ice, diluted and plated as described above. min interval following the addition of hydrogen peroxide. For eight of the early proteins the rate of synthesis has returned to normal within 30 min after hydrogen peroxide addition. A second class of hydrogen peroxide-inducible proteins, the "late" proteins, is synthesized at a maximal rate starting 10-30 min after hydrogen peroxide addition. A number of the late proteins are induced to some extent before 20 min, and most continue to be synthesized at an elevated rate between 50 and 60 min after hydrogen per- oxide addition. There are 12 early and 18 late proteins. The fact that 60/~M hydrogen peroxide induces the syn- thesis of 30 proteins illustrates that the adaptation re- sponse is complex, and that it involves more than the in- duction of catalase. A Hydrogen Peroxide Resistant Mutant Constitutively Expresses Nine of the Twelve Early Proteins We have isolated a large number of hydrogen peroxide- resistant mutants following mutagenesis with diethylsul- fate. The mutants arise at a frequency of about 1 x 10 -6 per survivor following mutagenesis. No spontaneous hydrogen peroxide-resistant mutants have been isolated, indicating a spontaneous frequency to hydrogen peroxide resistance of less than 1 x 10 -a. Only one major class of hydrogen peroxide-resistant mutants arise under the con- ditions used, as determined by two-dimensional protein gel analysis, enzyme activities (see below), and a variety of other phenotypes, such as colony morphology and re- sistance to alkyl hydroperoxides (data not shown). There- Regulation of Defenses against Oxidative Stress 757 H pt1T H PI -n ~" A. Cata lase B. S u p e r o x i d e D i s m u t a s e 1 2 3 4 5 6 1 2 i -.iP.~- - - Mn SOD Fe SOD Figure 5. Catalase and Superoxide Dismutase Activity Stains of LT2 and oxyR1 Extracts Cells were grown to stationary phase in Luria Broth, and extracts were prepared by sonication. For both catalase and superoxide dismutase activity stains, 9% acrylamide gels were run as described previously (Davis, 1964). Catalase activity staining gels were slightly modified, such that the running gel buffer was at pH 8.0. The activity stains are described in Experimental Procedures. (A) 1, 30/~g LT2 crude extract; 2, 30 ~g oxyR1 crude extract; 3, 4 ~g LT2 0-25% ammonium sulfate; 4, 4 #g oxyR1 0-25% ammonium sulfate; 5, 25 ~g LT2 25-55% ammo- nium sulfate; and 6, 25 Fg oxyR1 25-55% ammonium sulfate. (B) 1,100 /~g LT2 crude extract; and 2, 100 #g oxyR1 crude extract. duced after an anaerobic to aerobic shift (Gregory and Fridovich, 1973). Native gels stained for SOD activity on wild-type and oxyR1 extracts are shown in Figure 5B. In- tegrations of the gel densitometer tracings show that the activity band of slower mobility is increased about 2.2-fold in oxyR1, whereas the faster mobility band is half as abun- dant in oxyR1 as in the wild type. The induction of the slower SOD activity band and the repression of the faster activity band has been observed in E. coil B and K12 fol- lowing anaerobic to aerobic shifts (Hassan and Fridovich, 1977a). The slower band has been identified as the man- ganese containing SOD, and the faster form, as an iron- containing SOD in E. coil (Hassan and Fridovich, 1977a). In S. typhimurium and in E. coil, activity of the faster band is inhibited by hydrogen peroxide and is not produced dur- ing growth on minimal glucose. Therefore, we presume that the slower mobility band in S. typhimurium represents the manganese SOD, and the faster band, the iron SOD. Glutathione Reductase Glutathione is present at levels of 5-10 mM in E. coil (Meister and Anderson, 1983). It can be easily oxidized to the intermolecular disulfide, which subsequently be- comes reduced by glutathione reductase (Meister and An- derson, 1983). Extracts of oxyR1 contain 4-fold higher lev- els of glutathione reductase than do wild-type extracts (Table 2). Reduced glutathione is probably important for preventing excessive damage from toxic oxygen species (Penninckx and Jaspers, 1982; Morse and Dahl, 1978), and its coregulation with catalase and superoxide dismu- tase supports this proposed function. NAO(P)H.Dependent Alkyl Hydroperoxide Reductase oxyR1 is resistant to growth inhibition by curnene hydro- peroxide and t-butyl hydroperoxide (Table 1). In an attempt to determine the mechanism of resistance to these or- ganic hydroperoxides (which are not catalase substrates), we assayed the ability of extracts of oxyRl to degrade cu- mene hydroperoxide using a reverse phase HPLC assay (see Experimental Procedures). Either whole cells in the presence of glucose or extracts of oxyRl in the presence of NAD(P)H will reduce cumene hydroperoxide to the cor- responding alcohol at 5-20 times the rate that wild-type cells or extracts catalyze the reaction (Table 2). To our knowledge, this is a novel activity in enteric bacteria, al- though NADH-dependent hydrogen peroxide reductases have been described in Streptococcus faecalis (Dolin, 1957) and in Lactobacillus casei (Mizushima and Kita- hara, 1962). The activity has been purified to near homogeneity from oxyR1 extracts (Jacobson etal., unpub- lished data), and has been shown to consist of two pro- teins, both of which are required for activity. One compo- nent binds to 5'AMP nucleotide affinity columns, contains FAD, and has a 52 kd subunit molecular weight. The other component, which does not bind to a 5'AMP column, has a 22 kd subunit molecular weight. Two-dimensional elec- trophoresis of the purified components with wild-type cell extracts allows assignment of the proteins with those shown in Figure 4. The 52 kcl protein is oxyR1 protein F52a, and the 22 kd protein is oxyR1 protein C22 (see Fig- ure 4). This represents a fourth activity related to protec- tion against oxidative damage that is overexpressed in oxyR1 mutants. In addition to the activities described above, oxyR1 extracts produce 2-fold higher levels of glucose-6-phosphate dehydrogenase than do wild-type extracts. This enzyme, which produces NADPH, may serve to provide the reducing equivalents needed for glutathione reductase and the alkyl hydroperoxide reduc- tase. There is no difference between the levels of hexo- kinase and lactate dehydrogenase in oxyR1 and LT2 extracts. oxyR1 Mutants Overexpress Three Heat-Shock Proteins Both oxyR1 and hydrogen peroxide-treated wild-type cells are resistant to heat killing (Table 1 and Figure 2). Heat in- duces production of the family of dinucleotides that are also induced by oxidants (Bochner et al., 1984; Lee et al., 1983). This led us to investigate possible overlaps be- tween proteins induced by a temperature shift and overex- pressed proteins in oxyR1 mutants, oxyR1 proteins F52a, D64a, and E89 are also heat-shock proteins in wild-type S. typhimurium (Figure 6). Two of these three proteins (D64a and E89) can be induced further by heat shock in oxyRl, indicating the existence of at least two levels of control for these proteins in S. typhimurium. We do not know whether the high-level expression of these heat- shock proteins in oxyR1 is responsible for the resistance to heat killing. This establishes that a single regulatory lo- cus controls the expression of some heat-shock proteins and of proteins involved in defenses against oxidative damage. Deletions of oxyR Are Recessive and Uninducible by Hydrogen Peroxide for the Nine Proteins Overexpressed by oxyR1 Mutants Tnl0-mediated deletions have been obtained from both wild-type and oxyR1 strains carrying a TnlO in argH using a positive selection for tetracycline sensitivity (Bochner et al., 1980). Thirty tetracycline-sensitive isolates from TA4106 Cell 758 A _ 0 ~ Wild ~ ~ i ° ~_ ~+~ Type ~ ~ Controls H 2 02 , 8 ~ " i oxyA2 -'~ ©U~a. U F 5 2 ~ & . B Proteins Overproduced in oxyR 1 H e a t - S h o c k Type t o x y Wild &2 Inducing Condition H 2 0 2 H e a t - S h o c k G18 -f- -- C 2 2 + -- D29 + -- G35 + - F 5 2 a + + D 6 4 a + -I- D69 -t- -- D71 --I-- -- E89 + + H 2 0 2 m m m m m m Heat -Shock ( - ) ( - ) ( - ) ( - ) + ( - ) ( - ) + Figure 6. Hydrogen Peroxide and Heat-Shock Induction of Proteins in oxyA2 Cells were grown in VBC salts containing 0.4% glucose at 37°C with shaking. Labeling of pro- teins after hydrogen peroxide addition and heat shock were as described in Experimental Procedures. (1) A small region of the 2D gel containing the three heat-shock proteins that are overexpressed in oxyR1 and some of the hydrogen peroxide-inducible proteins. (2) Results for all of the oxyR1 proteins are summa- rized. Note that protein E89 is the smaller spot in the center of the circle labeled E89. (LT2/argH1823::Tn10) and 25 isolates from TA4105 (oxyR11argH1823::TnlO) were screened for sensitivity to hydrogen peroxide. Two isolates from each group were ex- tremely sensitive to hydrogen peroxide (diameter of killing zone with 300 vg of hydrogen peroxide: TA4106, 25mm; TA4105, 16mm; and hypersensitive deletions, 45mm). Two-dimensional protein gels of TA4105 and one of the hydrogen peroxide-hypersensitive deletions derived from it, oxyA1 (TA4107), indicate that the nine proteins overex- pressed in TA4105 are only present at wild-type levels in oxyA1. Furthermore, oxyA1 has roughly wild-type levels of catalase and of the alkyl hydroperoxide reductase. One of the hydrogen peroxide-hypersensitive deletions derived from TA4106, oxyA2 (TA4108), was tested for induc- tion of adaptation proteins by 60/~M hydrogen peroxide, and the results are shown in Figure 6. All nine of the oxyR1 proteins are uninducible in oxyA2 by 60 /~M hydrogen peroxide. Most of the 21 hydrogen peroxide-inducible pro- teins that are not overexpressed in oxyR1 are induced nor- mally in oxyA2 (data not shown). Most of the heat-shock proteins can be induced normally by a temperature shift in oxyA2, including oxyR1 proteins D64a and E89 (Figure 6). The only exception is protein F52a, which can no longer be heat-induced in oxyA2. However, the F52a pro- tein spot is visible on two-dimensional gels of oxyA2, and the alkyl hydroperoxide reductase activity, of which F52a is an essential component, is present at wild-type levels in oxyA2. Introduction of an E. coil episome covering the argH re- gion (F'14) into oxyA1 and oxyA2 restores their resistance to hydrogen peroxide to the wild-type level. The inducibil- ity of the nine proteins overexpressed in oxyR1 by 60 #M hydrogen peroxide is restored in oxyA2 strains containing F'14 (TA4109). Discussion We have observed that S. typhimurium is capable of adapting to hydrogen peroxide in a manner similar to that previously described for E. coil (Demple and Halbrook, 1983). Adapted cells are resistant to a variety of other agents causing oxidative damage, as well as to heat kill- ing. Two-dimensional protein gels show that S. typhimu- rium cells induce the synthesis of 30 proteins when ex- posed to 60/~M hydrogen peroxide. A class of hydrogen peroxide-resistant mutants, represented by oxyR1, has been isolated and has been shown to overexpress nine of the twelve most rapidly induced hydrogen peroxide adaptation proteins constitutively, oxyFtl mutants have elevated levels of catalase/peroxidases, Mn superoxide dismutase, glutathione reductase, and a novel alkyl hydroperoxide reductase. In addition, oxyR1 mutants con- stitutively overexpress three heat-shock proteins. A dele- tion generated from a TnlO that is closely linked to oxyR1 is thought to remove the oxyR gene, as it is hypersensitive to hydrogen peroxide, recessive, and uninducible by hydrogen peroxide for the nine oxyR1 proteins. Heat- shock protein induction is normal in the oxyA2 deletion for Regulation of Defenses against Oxidative Stress 759 Table 3. Bacterial Strains Strain Genotype Source TA4100 oxyR1 This study TA4101 oxyRllzhh116::Tn5 TA4102 oxyR*/zhhll6::Tn5 TA4103 HfrK4/zhh116::Tn5 TA4104 argE94/zhh116::Tn5 " TA4105 oxyRllargH1823::Tn10 " TA4106 LT21argH1823::TnlO TA4107 oxyA1 [oxyA(oxyRargH) 1 ] " TA4108 oxyA2 [oxyA(oxyRargH)2] " TA4109 oxyA2/F'l 4 TA4110 oxyR2 (E. coli) TA4112 oxyA3[oxyA(oxyRbt u B)3] " (E. coli, derived from RK4936) TA4113 katAl[katA(metBkatGargH)l] " TT137 argH1823::Tn10 John Roth TT2385 zii614::Tn10 RK4936 araD 139/(argF-lac)205/flbB5301/ R. Kadner non-9gyrA219/relA1/rpsL150/ rnetE7OIbtuB::TnlO (E. coli) most proteins with the exception of F52a (the alkyl hydro- peroxidase flavoprotein component), which is not heat in- ducible in oxyA2. We have isolated and characterized mu- tants similar to oxyR1 and oxyA2 in E. coil (Table 3). The oxyR Gene Product Is a Positive Regulator of Defenses Against Oxidative Stress and Some Heat-Shock Proteins Deletion oxyA2 is presumed to inactivate the oxyR gene, since it is uninducible by hydrogen peroxide for the nine oxyRl proteins, and an E. coli episome restores the ability of oxyA2 to induce these proteins. The most reasonable interpretation is that the oxyR gene product is a positive effector of gene expression, the activity or expression of which is hydrogen peroxide-inducible, oxyRl mutants may either overproduce the oxyR gene product or produce an inducer-independent product. Such mutants are pre- dicted to be dominant, as is observed for oxyRl. The map positions of three of the genes controlled by oxyR are known in E. coli. Since the genetic maps of these two closely related bacteria are almost identical, these genes probably map in the same region in S. typhimurium. They are sodA at 89,5 min. (Mn SOD, Tou- ati, 1983), gor at 77 min. (glutathione reductase, Davis et al., 1982), and katG at 88 min. (HPI-II catalases, Loewen et al., 1985). Since sodA and gor are unlinked to the regulatory gene, oxyR must act in trans to affect gene ex- pression. Deletions extending from argH to metB (e.g., TA4113) abolish HPI-II activity but do not delete oxyR. Therefore, oxyR maps to the opposite side of argH from katG, indicating that these two genes are not part of the same operon. The genes for proteins D64a and E89 are subject to multiple controls, since their hydrogen peroxide induction is oxyR-dependent but their basal level of expression and their heat induction are oxyR-independent. The expres- sion of protein F52a appears to be different. Its hydrogen peroxide and heat induction are oxyR-dependent, but its basal level of expression is oxyR-independent. The oxyR gene product could function in a manner analogous to other positive regulatory proteins in bacteria such as, an alternate sigma factor for RNA polymerase (Grossman et al., 1984), a nucleotide binding protein that interacts with RNA polymerase at certain promoters, or an indirect-acting positive regulator such as the recA protein. Substantial homology in primary amino acid sequence has been observed between the htpR sigma factor and the normal sigma protein (Landick et al., 1984) as well as be- tween positive regulatory proteins CRP and Fnr (Shaw et al., 1983). Thus, the DNA sequence of the oxyR gene may shed light on the mechanism of action of the gene product. Purpose of an Adaptive Response to Oxidative Stress Oxygen is highly toxic to anaerobic organisms, and its metabolic products are toxic to aerobes as well (Fridovich, 1978). The major toxic products that arise from oxygen me- tabolism are superoxide anion, hydrogen peroxide, singlet oxygen, and the hydroxyl radical (Fridovich, 1978). The ex- istence of an adaptive response to oxidative agents such as hydrogen peroxide, therefore, might be expected in fac- ultative anaerobes such as S. typhimurium and E. coil, which undergo shifts from anaerobic to aerobic environ- ments. Hydrogen peroxide adaptation may also be useful as a means for potential pathogens, such as S. typhimurium and E. coli, to survive respiratory bursts of hydrogen peroxide and superoxide anion associated with phagocy- tosis of bacteria by activated granulocytes. The respira- tory bursts are capable of generating millimolar concen- trations of hydrogen peroxide and are thought to facilitate killing of engulfed bacteria (Root and Cohen, 1981). In ad- dition, studies with catalase deficient mutants of Staphylo- coccus aureus have demonstrated a direct correlation be- tween catalase activity and pathogenicity (Mandell, 1975). oxyR Controls a Fourth Global Response to a DNA-Damaging Agent Hydrogen peroxide has been shown to damage DNA in vitro (Demple and Linn, 1982) and cause mutations in vivo (Levin et al., 1982). Thus, in addition to the SOS response (Walker, 1984), the adaptive response to alkylating agents (Samson and Cairns, 1977) and heat-shock (Krueger and Walker, 1984), the oxyR regulatory network is a fourth ma- jor regulon in enterics that can be induced by an agent that damages DNA. The SOS and adaptive responses in- volve the induction of DNA repair systems that repair DNA damage caused by the inducing agent. Hydrogen perox- ide adaptation in E. coli may involve the induction of DNA repair specific for oxidative damage (Demple and Hal- brook, 1983; Hollstein et al., 1984), although the gene products involved have not been identified. Each of these four systems is activated by a positive effector of gene ex- pression (Walker, 1984; Krueger and Walker, 1984). The Relationship between Dinucleotides, Heat Shock, and Oxidative Stress There is evidence that heat shock and oxidative stress may be related phenomena (Lee et al., 1983). The heat- shock response in eukaryotic cells can be induced by a variety of oxidizing agents, including hydrogen peroxide,