J Biol Chem, Vol. 275, Issue 18, 13398-13405, May 5, 2000

Transcriptional Regulation of Glutaredoxin and Thioredoxin Pathways and Related Enzymes in Response to Oxidative Stress^*

María-José Prieto-Álamo, Juan Jurado^§, Rafaela Gallardo-Madueño, Fernando Monje-Casas^¶, Arne Holmgren, and Carmen Pueyo^**

From the  Departamento de Bioquímica y Biología Molecular, Universidad de Córdoba, 14071-Córdoba, España and the  Medical Nobel Institute for Biochemistry, Department of Medical Biochemistry and Biophysics, Karolinska Institute, S-171 77, Stockholm, Sweden

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We examined the in vivo expression of up to 16 genes encoding for components of both glutaredoxin and thioredoxin systems and for members of the OxyR and SoxRS regulons. We demonstrated that grxA (Grx1) transcription is triggered in bacteria lacking Trx1 (trxA) and GSH (gshA) in an OxyR-dependent manner. We also indicated that, unlike OxyR, SoxR is not constitutively activated in the oxidizing environment of trxA gshA mutants. We discovered that the lack of Trx1 plus GSH increases the steady-state levels of Trx reductase (trxB) and Trx2 (trxC) transcripts. This increase and the trxB and trxC up-regulation caused by the constitutive oxyR2 allele indicate that OxyR also plays a role in the regulation of the thioredoxin pathway. On the contrary, no change in the expression of genes for Trx1, Grx2, and Grx3 was observed. Transcription of nrdAB (RRase) was not induced by oxidative stress yet was induced by hydroxyurea (RRase inhibitor). Induction level was as the enhanced nrdAB basal expression of trxA grxA mutants, indicating that RRase operation without Trx1 and Grx1 must lead to disturbances sensed as those caused by hydroxyurea. We also demonstrated an inverse relation between nrdAB expression and that of genes coding for components of both glutaredoxin (grxA, gorA) and thioredoxin (trxB, trxC) systems.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Thioredoxins (Trxs)¹ and glutaredoxins (Grxs) are known hydrogen donors for ribonucleotide reductase (RRase), the essential key enzyme for deoxyribonucleotide and DNA biosynthesis. In the reduced form, both Trxs and Grxs contain two redox-active cysteine thiols, which by dithiol-disulfide interchange reduce an acceptor disulfide in the active center of RRase. Trxs and Grxs differ in the manner they are reduced in the cell, although ultimately reducing equivalents come from NADPH. Thus, the thioredoxin system is composed of NADPH, the flavoprotein thioredoxin reductase, and Trx; and the glutaredoxin system of NADPH consists of the flavoprotein glutathione reductase, the ubiquitous tripeptide glutathione (GSH), and Grx (1) (Fig. 1).

View larger version (20K): [in this window] [in a new window]   Fig. 1.   Known components of the thioredoxin and glutaredoxin systems and related enzymes. Gene names are given in parentheses. OxyR* and SoxR* denote the oxidized states of these two regulatory proteins. OxyR-dependent regulation of grxA, trxB, and trxC is shown in the present study.

Escherichia coli contains genetic information for three different RRases (2). The NrdAB (encoded by the nrdAB operon) is active during aerobiosis, NrdDG is strictly anaerobic and uses formate as electron donor, and NrdEF is a cryptic enzyme of unknown physiological role. At present, two thioredoxins (Trx1 and Trx2) and three glutaredoxins (Grx1, Grx2, and Grx3) have been discovered in the bacterial cytoplasm (1, 3, 4). Grx1 (grxA) and Trx1 (trxA) are the two main hydrogen donors of E. coli RRase (5, 6). In comparison, Grx3 (grxC) is a very inefficient reductant (about 5% of the catalytic activity of Grx1), while Grx2 (grxB) lacks activity as hydrogen donor for RRase (3). Similar to Trx1, Trx2 (trxC) is a functional electron donor for NrdAB. However, compared with Trx1, Trx2 is a 5-fold less abundant protein and a 2.5-fold less efficient enzyme (4). A fifth potential reductant is the NrdH-redoxin, which has thioredoxin-like activity but a glutaredoxin-like amino acid sequence (7). NrdH is a functional hydrogen donor for RRase with higher specificity for the NrdEF than for the NrdAB enzyme. However, the physiological function of NrdH in E. coli is not well understood, since the nrdH gene is located in the same poorly transcribed operon as nrdEF (8).

Different inducible responses are critical in protecting E. coli from the damage caused by oxidative stress (9). Key regulators of adaptive responses to hydrogen peroxide (H₂O₂) and superoxide anion (O₂) are the OxyR and SoxRS transcription factors, respectively (recently reviewed in Ref. 10). Among the OxyR-regulated genes are those encoding for hydroperoxidase I (catalase, katG), glutathione reductase (gorA), Grx1 (grxA), and OxyS RNA (oxyS). Regulation of the soxRS regulon occurs by a two-stage process; the constitutively expressed SoxR protein is first converted to an active form, which stimulates soxS transcription, and the increased levels of SoxS in turn activate expression of the regulon. Among the SoxRS-regulated genes are those encoding for manganese superoxide dismutase (sodA) and glucose-6-phosphate dehydrogenase (zwf).

The OxyR protein is directly sensitive to oxidation, and only the oxidized OxyR is capable of activating transcription (11). Recent studies have revealed that oxidation of OxyR leads to the formation of an intramolecular disulfide bond and that OxyR is reduced and deactivated by enzymatic reduction with Grx1 at the expense of glutathione (12). SoxR is a homodimer containing two [2Fe-2S] centers essential for its transcriptional activity (13). The univalent oxidation of the iron-sulfur clusters appears to be the mechanism for activating SoxR as a transcriptional factor (14, 15). The mechanism of SoxR reduction/deactivation has not been elucidated, except for an NADPH-dependent SoxR reductase that has recently been purified from E. coli (16). Likewise, it has recently been suggested that cellular monothiols, like GSH, and low molecular weight dithiols and dithiols proteins, such as Trx1, may contribute to SoxR regulation by promoting the disassembly (14) and reassembly (17) of the [2Fe-2S] clusters.

Recently, we reported that the in vivo transcription of the nrdAB operon and of the grxA and fpg (coding for formamidopyrimidine-DNA glycosylase; Ref. 18) genes is triggered in E. coli lacking both Trx1 and Grx1 (trxA grxA mutant) or Trx1 and GSH (trxA gshA mutant) (19). The present study aimed to examine the role of OxyR and SoxRS in determining such increased basal levels of expression. To this end, trxA grxA and trxA gshA defective bacteria, additionally mutated in the oxyR or soxRS regulatory genes, have been isolated, and the in vivo expression of up to 16 different genes has been examined by reverse transcription/multiplex polymerase chain reaction (RT/MPCR). The genes studied include most components of both thioredoxin and glutaredoxin systems as well as known members of the OxyR and SoxRS regulons.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Materials-- Media and chemical reagents were prepared or purchased as described previously (19). Hydrogen peroxide, paraquat, and hydroxyurea were purchased from Sigma.

Media-- Bacteria were grown in Luria-Bertani (LB) nutrient broth or M9 minimal medium (20). The media were supplemented (when necessary) with chloramphenicol (34 µg/ml), kanamycin (50 µg/ml), or tetracycline (20 µg/ml). The minimal medium contained arginine (40 µg/ml), D-biotin (5 µg/ml), thiamine (5 µg/ml), glucose (2 g/liter), and casamino acids (2 g/liter).

Bacterial Strains-- Strains used are listed in Table I. All strains are E. coli K-12, UC5710 (21) being considered the parental wild type. UC844, UC827, and UC859 have been previously described (22, 23). Strains were constructed by P1 transduction (24). The mutant alleles of the genes used to construct the strains are as follows: sox-8::cat (25), oxyR::kan (26), btuB::Tn10 (obtained from M. Blanco), oxyR2 (27) linked to btuB::Tn10, and gshA linked to srl::Tn10 (28). The sox-8::cat deletion covers the region encoding both the SoxR and SoxS proteins (25). Strains with the sox-8::cat or oxyR::kan allele did not induce glucose-6-phosphate dehydrogenase (coded for by the zwf gene) or hydroperoxidase I catalase (coded for by the katG gene) activity upon exposure to 100 µM paraquat or 500 µM H₂O₂, respectively. Strains carrying the oxyR2 allele exhibited constitutive high levels of catalase activity. UC1351 was made because the generation of a tetracycline-sensitive clone leaves a genetic lesion at the original site of the insertion. Tc^S bacteria were selected as described (29). The genetic lesion carried by UC1351 extended outward from the Tn10 insertion site into the grxA locus, making the bacteria additionally sensitive to kanamycin (Kan^S). UC1351 and its parental UC827 showed identical growth rates in LB broth and the same basal levels of gene expression, as measured by the set A of primers. Strains with either the gshA::Tn10kan (UC859, UC1336, and UC614) or the gshA (UC1358, UC1369, UC1363, and UC1395) mutant allele exhibited identical sensitive to diamide and undetectable levels of glutathione, as determined by high pressure liquid chromatography (23). The second trxA gshA double mutant (UC1358) displayed a greater growth defect than the first construction (UC859) (doubling time in LB of 60 ± 1 and 37 ± 2 min, respectively). Nonetheless, suppressor mutations (no reversions of the gshA mutant allele) that allow UC1358 to grow faster arose at a high frequency (approximately 10⁴). UC1369 was constructed by selecting for spontaneous reversion to utilization of sorbitol (Srl⁺). This precise eductant became sensitive to tetracycline.

Treatments-- E. coli cells from an overnight culture in LB broth were diluted 100-fold into 50 ml of M9 minimal medium and incubated at 37 °C and 150 rpm to reach an A₆₀₀ of 0.2. At this stage, the bacteria were further grown in the absence or the presence of hydrogen peroxide, paraquat, hydroxyurea, or diethylmaleate for a fixed time period. The cells were then rapidly cooled to 0 °C for total RNA purification.

RNA Purification-- Total RNA was extracted using the hot phenol method previously described (30). The quality of the samples was checked electrophoretically, and quantification was done spectrophotometrically. At least two independent RNA preparations were isolated for each experimental condition.

Primers-- Primers used are listed in Table II. Set A has been previously described (19). Set B and set C were designed with the Primer Select 3.03/96 (DNA Star, Madison, WI) and Oligo 5.0/96 (National Biosciences, Plymouth, MN) programs, in order to obtain the highest specificity and performance in multiplexed PCRs. Target genes code for the two subunits of NrdAB (nrdA, nrdB); Trx1 and Trx2 (trxA, trxC); Trx reductase (trxB); Grx1, Grx2, and Grx3 (grxA, grxB, grxC); glutathione reductase (gorA); -glutamylcysteine synthetase (gshA); thioredoxin-linked Tpx (tpx); formamidopyrimidine-DNA glycosylase (fpg); hydroperoxidase I catalase (katG); OxyS RNA (oxyS); SoxS transcription factor (soxS); and glucose-6-phosphate dehydrogenase (zwf). As described previously (19, 31), the gapA gene, which codes for D-glyceraldehyde-3-phosphate dehydrogenase, was used as internal standard. An exogenous fragment (referred as external standard) of the gene coding for cytochrome P4501A1 from Liza aurata (32) was coamplified with the target genes and the internal standard of set C. Primers for katG and oxyS were alternately included in set B, since both were labeled with the same fluorophor and gave PCR products of identical length. As control, primers of new sets amplify some of the genes of set A. Thus, set A shares the nrdA, trxA, and grxA genes with set B and the gorA gene with set C. For those genes in common, data obtained with different primers were essentially identical; accordingly, only the results obtained with one of the sets are presented.

Reverse Transcription/Multiplex PCR-- Synthesis of cDNA was carried out with the GeneAmp RNA PCR kit, as described previously (19). In short, 1 µg of bacterial RNA (plus 30 pg of external standard RNA, in the case of set C) was retrotranscribed for 15 min at 42 °C with 2.5 units of murine leukemia virus reverse transcriptase, using random hexamers. The enzyme was inactivated by heating for 5 min at 99 °C. Each RNA sample was retrotranscribed on three separated occasions. PCR amplification of cDNA was carried out using the primer pair sets listed in Table II. As detailed by Gallardo-Madueño et al. (19) for set A, PCR conditions were optimized to produce fluorescence intensities only of the desired products and in the range of linearity. The primer concentrations were titrated so that the genes yielded quantifiable amounts of PCR products over a wide range of different expression levels when co-amplified. The multiplex PCR amplification was performed in a mixture (25 µl final volume) containing 1.5 units of DNA polymerase, 2.5 µl of MPCR buffer 3, a 1 mM concentration of each dNTP, and primers at the following amounts: (i) set A, 2.75 pmol (nrdA), 3 pmol (nrdB), 1.25 pmol (trxA), 1.25 pmol (grxA), 3 pmol (gorA), 2.5 pmol (fpg), and 2 pmol (gapA); (ii) set B, 0.8 pmol (nrdA), 3.8 pmol (trxA), 2.8 pmol (grxA), 0.76 pmol (grxB), 1.4 pmol (grxC), 0.36 pmol (katG) or 2.4 pmol (oxyS), 1.15 pmol (soxS), and 0.74 pmol (gapA); (iii) set C, 2.75 pmol (trxB), 3.25 pmol (trxC), 2.75 pmol (gorA), 3.0 pmol (gshA), 2.5 pmol (tpx), 3.5 pmol (zwf), 1.25 pmol (gapA), and 3.5 pmol (external standard). Twenty-seven cycles of PCR were performed with set A, 24 cycles with set B, and 25 cycles with set C. Each cycle consisted of 1 min of denaturation at 94 °C, 15 s of annealing at 70 °C, and 30 s for enzymatic primer extension at 72 °C. Multiplex PCR products were quantified as described previously (19). Differences among PCR outcomes were normalized by comparing the fluorescence intensity of each band to that resulting from gapA amplification (internal standard). Samples for comparison of different experimental conditions or different bacterial strains were handled in parallel. Data are the means ± S.E. from n independent multiplexed PCR amplifications. Comparison between groups was done by Student's t test.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We have recently devised a sensitive, rapid, nonradioactive, and semiquantitative procedure (RT/MPCR) for comparing the in vivo expression of metabolically related genes (19, 31). In this technique, the target genes and a constitutively expressed reporter gene are coamplified in the same reaction vessel. Specific fluorophor-labeled primers are used, and amplification products are subsequently analyzed with a DNA sequencer (GeneScan). The levels of expression of the genes of interest are presented in reference to the internal standard. Here we have used three sets of primers to examine the expression of genes coding for most components of both thioredoxin and glutaredoxin systems as well as pivotal members of the OxyR and SoxRS regulons. One of the sets includes an external standard to control the potential variability of the reporter gene.

Gene Expression Induction by Oxidative Stress-- Hydrogen peroxide activates the transcription factor OxyR through the oxidation of two cysteines and formation of an intramolecular disulfide bond (12). Superoxide-generating compounds, such as paraquat, activate the transcription factor SoxR by oxidizing the [2Fe-2S] clusters in the protein through an unknown mechanism (14-15, 33). By using the set A of primers (19), we examined the expression of genes coding for the NrdAB ribonucleotide reductase (nrdA and nrdB), its two main hydrogen donors Trx1 (trxA) and Grx1 (grxA), the glutaredoxin pathway enzyme glutathione reductase (gorA), and the DNA repair glycosylase Fpg (fpg), in response to increasing concentrations of H₂O₂ and paraquat (Fig. 2). Treatment with H₂O₂ stimulated the expression of the two genes (gorA and grxA), identified as being under the oxyR control (11, 31, 34), as well as that of the fpg gene. Maximal induction levels of ~3.5-, ~1.6-, and ~13.0-fold were observed for fpg, gorA, and grxA, respectively, after 10 min of treatment with concentrations of 100 µM H₂O₂. Paraquat also activated the expression of these three genes but at higher concentrations than did H₂O₂. On the contrary, expression of nrdA, nrdB, and trxA genes was not induced after oxidative stress by either H₂O₂ or paraquat.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Gene expression induction by hydrogen peroxide and paraquat. Wild-type cells were treated for 10 min with the hydrogen peroxide (H2O2) or paraquat concentration (in µM) indicated on the abscissa. cDNAs were amplified by using the set A primers. The fluorescence signal of each PCR product was referred to that of gapA (internal standard). Data are from an average of six independent multiplexed PCR amplifications. Values from treated samples were divided by those from the corresponding control and plotted as a function of H2O2 or paraquat concentration. All genes of set A were analyzed, but only those genes for which highly statistically significant (p  0.001) increases were observed at a given H2O2 or paraquat concentration are represented. Error bars were estimated from the corresponding S.E.

To examine the role of OxyR and SoxRS regulators in fpg, gorA, and grxA expression induction by H₂O₂ and paraquat, strains carrying the oxyR::kan (OxyR) or sox-8::cat (SoxRS) mutant allele were constructed and subjected to oxidative stress conditions in conjunction with wild-type bacteria (Fig. 3). H₂O₂ was used at a concentration of 100 µM, and paraquat was used at a concentration of 500 µM; gene expression was examined at 5 min upon exposure. The induction of fpg, gorA, and grxA expression by H₂O₂ or paraquat was abolished by the introduction of the oxyR::kan mutation, indicating that activation by both oxidants is OxyR-dependent. In contrast, this transcriptional up-regulation was preserved in the strain with the sox-8::cat mutation. It should be noted that induction factors for fpg (~3.0-fold), gorA (~3.5-fold), and particularly grxA (~45.5-fold) genes were higher at 5 min (Fig. 3) than at 10 min (Fig. 2) after the addition of H₂O₂ or paraquat, in agreement with previous observations showing that oxyR-regulated genes are most strongly induced within the first 5 min of H₂O₂ exposure (31). The paraquat activation of genes that are induced by H₂O₂ and controlled by OxyR can be explained by the spontaneous and superoxide dismutase-mediated conversion of O₂ to H₂O₂ (35). In agreement with this, the lower levels of paraquat induction in the SoxRS deletion mutant (Fig. 3), as compared with wild type, are expected for the inability of SoxRS bacteria to induce sodA (Mn-superoxide dismutase) transcription upon paraquat exposure (25).

View larger version (20K): [in this window] [in a new window]   Fig. 3.   OxyR and SoxRS regulation of gene expression induced by hydrogen peroxide and paraquat. UC5710 (wild type), UC1342 (oxyR::kan), and UC1333 (sox-8::cat) cells were treated for 5 min with 100 µM hydrogen peroxide (H2O2) or 500 µM paraquat. The fluorescence signal of each PCR product was referred to that of gapA (internal standard). Data were from an average of six independent multiplexed PCR amplifications. Values from treated samples were divided by those from the corresponding control. All genes of set A were analyzed, but only those genes for which highly statistically significant (p  0.001) increases were observed with a given bacterial strain are represented. Error bars were estimated from the corresponding S.E.

Construction of Mutants Lacking Components of the Thioredoxin and Glutaredoxin Systems and Additionally Mutated in the soxRS or oxyR Regulatory Genes-- We have recently demonstrated that untreated trxA grxA and trxA gshA mutant bacteria show increased basal levels of nrdAB and of grxA and fpg expression (19). Likewise, it has been recently reported that both trxA gorA and trxA gshA double mutants display partial induction of the OxyR-regulated oxyS gene (36) and that trxA gorA bacteria exhibit also slightly elevated basal expression of the SoxR-regulated soxS gene (17). To explore the role of OxyR and SoxRS regulators on basal levels of gene expression in bacteria lacking components of both thioredoxin and glutaredoxin systems, soxRS and oxyR mutant derivatives were isolated from trxA grxA and trxA gshA double mutant strains. Strains UC1337 (trxA grxA::kan zbi::Tn10 sox-8::cat), UC1336 (trxA gshA::Tn10kan sox-8::cat), and UC1356 (trxA grxA oxyR::kan) were readily constructed by transducing to Cam^R or Kan^R the corresponding isogenic parental strain (UC827, UC859, or UC1351). We were unable to construct an oxyR::kan derivative from UC859 (P1 (UC1381) × UC859; selection for Tc^R and screening for hydroperoxidase I catalase induction). However, the oxyR::kan allele was successfully moved into a second trxA gshA srl::Tn10 double mutant (UC1358). The oxyR2 mutation causes overexpression of oxyR-regulated proteins in the absence of oxidative stress (27). Unlike the oxyR::kan insertion mutation, oxyR2 was moved with similar efficiency into both UC859 and UC1358 genetic backgrounds.

Effects of Mutations in soxRS or oxyR on Basal Levels of Gene Expression of Mutants Lacking Components of the Thioredoxin and Glutaredoxin Systems-- The lack of SoxRS did not change significantly the steady-state levels of fpg, gorA, grxA, nrdA, and nrdB transcripts in both trxA grxA and trxA gshA double mutant strains (Fig. 4). Thus, UC1337 and UC1336, which lack the functional soxRS genes, did not appear to be hindered in their abilities to display large increments in nrdAB and grxA and fpg expression, respectively, as compared with the parental wild type. As indicated above, we failed in the isolation of an OxyR derivative from UC859. Unexpectedly, however, the oxyR::kan allele was readily transferred to the new trxA gshA double mutant (UC1358) isolated in this work. By using the set A of primers, we detected substantial differences in the basal levels of nrdAB and grxA expression between the two trxA gshA mutants UC859 and UC1358; of note was the additional observation that the transcriptional up-regulation of grxA in UC1358 was completely absent from its OxyR derivative. Encouraged by these results, we proceeded to investigate in more detail the effect of OxyR on the steady-state levels of gene expression exhibited by bacteria simultaneously defective in Trx1 and GSH and the reasons for differences between mutant strains. To this end, constitutive OxyR^c derivatives were constructed, and new sets of primers were designed in order to examine by RT/MPCR the expression of 10 additional genes. These genes code for the two new Grx2 (grxB) and Grx3 (grxC) glutaredoxins, the recently discovered Trx2 (trxC) thioredoxin, the thioredoxin pathway enzyme Trx reductase (trxB), the OxyR-regulated hydroperoxidase I catalase (katG) and OxyS RNA (oxyS), the SoxR-regulated SoxS transcriptional factor (soxS), the SoxRS-regulated NADPH supplier glucose-6-phosphate dehydrogenase (zwf), the Tpx enzyme linked to the thioredoxin system (tpx), and the GSH biosynthetic pathway enzyme -glutamylcysteine synthetase (gshA). The expression levels of 15 genes were examined in UC5710, UC859, UC1358, and their respective OxyR and/or OxyR^c mutant derivatives. The amounts of the trxA, grxB, grxC, soxS, zwf, tpx, and gshA transcripts remained basically unchanged among the strains. Genes trxA and gshA are not expressed in trxA gshA mutants. Putative differences in gapA (internal standard) expression among strains were ruled out by expressing the levels of gapA in reference to the heterologous external standard included in set C. Those genes for which significant differences with respect to basal levels of wild-type parent strain (UC5710) were found are shown in Fig. 5.

View larger version (12K): [in this window] [in a new window]   Fig. 4.   Basal levels of gene expression in SoxRS proficient and deficient bacteria. UC5710 (wild type), UC827 (trxA grxA::kan zbi::Tn10), UC859 (trxA gshA::Tn10kan), UC1333 (sox-8::cat), UC1337 (trxA grxA::kan zbi::Tn10 sox-8::cat), and UC1336 (trxA gshA::Tn10kan sox-8::cat) cells were grown in LB broth to reach an A600 of 0.7. cDNAs were amplified by using the set A primers. The fluorescence signal of each PCR product was referred to that of gapA (internal standard). Data are from an average of six independent multiplexed PCR amplifications. Values from UC827 and UC859 were divided by those from UC5710, and values from UC1337 and UC1336 were divided by those from UC1333. -Fold increases relative to UC5710 (SoxRS+) or UC1333 (SoxRS) were plotted for the different genes of set A. Bacteria carrying the null trxA allele had undetectable levels of the corresponding mRNA.

View larger version (24K): [in this window] [in a new window]   Fig. 5.   OxyR regulation of basal levels of gene expression. UC5710 (wild type), UC1342 (oxyR::kan), UC1394 (oxyR2 btuB::Tn10), UC859 (trxA gshA::Tn10kan), UC614 (trxA gshA::Tn10kan oxyR2 btuB::Tn10), UC1358 (trxA gshA srl::Tn10), UC1363 (trxA gshA srl::Tn10 oxyR::kan), and UC1395 (trxA gshA oxyR2 btuB::Tn10) cells were grown in LB broth to reach an A600 of 0.7. cDNAs were amplified by using the set A, set B (including katG), and set C primers. The fluorescence signal of each PCR product was referred to that of gapA (internal standard). Data were from an average of six independent multiplexed PCR amplifications. -Fold increases relative to UC5710 were plotted for the different bacterial strains. All genes were examined, but only those genes (gorA, fpg, nrdA, and nrdB from set A; grxA and katG from set B; trxB and trxC from set C) for which highly statistically significant (p  0.001) increases were observed with a given bacterial strain are represented. Error bars were estimated from the corresponding S.E.

Transcription of grxA in the oxyR::kan mutant (UC1342) was 2.9-fold lower than in UC5710. On the contrary, a dramatic increase of 33-fold in grxA expression was observed in the strain (UC1394) carrying the oxyR2 missense mutation, which is considered highly effective causing a constitutively active phenotype (26). The trxA gshA double mutant UC859 showed a basal level of grxA message as high as that of the constitutive mutant UC1394. This level was further elevated in its oxyR2 constitutive derivative (UC614), where a 62-fold increase in the amount of grxA transcript was observed. The second trxA gshA double mutant UC1358 showed lower increment in grxA expression (8.2-fold relative to wild type) than UC859. This basal level was further modulated by mutations in the oxyR regulatory gene. Thus, grxA expression was not elevated in the absence of OxyR (UC1363), while it was induced 45-fold in the isogenic constitutive derivative (UC1395). It should be noted that the three oxyR2 derivatives displayed different amounts of grxA message, which increased in the same order that the grxA basal expression of their respective parental strains (UC5710 < UC1358 < UC859). This finding seems to indicate that oxyR2 mutants are still redox-active, as previously demonstrated for other constitutive mutations (26). This was confirmed by demonstrating that UC1394 (oxyR2) displayed a 2-fold induction of grxA expression (over its high steady-state level) upon treatment with 100 µM H₂O₂ for 5 min (data not shown).

As shown in Fig. 5, the expression of katG (OxyR-regulated) and that of trxB and trxC genes followed basically the same pattern as the expression of grxA; the exception was that basal expression levels were not significantly altered upon the introduction of the oxyR::kan mutation in either UC5710 or UC1358 background. The expression pattern of gorA (OxyR-regulated) showed an additional exception with that of grxA gene, since both trxA gshA double mutants (UC859 and UC1358) displayed an identical 2.5-fold increase in the amount of gorA transcript. Expression of nrdA and nrdB genes followed a pattern opposite to that of the OxyR-regulated genes. Therefore, the second trxA gshA double mutant UC1358 showed a higher (not lower) increment in nrdAB expression than UC859. Likewise, the oxyR2 constitutive mutation caused a decrease (not an increase) in the steady-state levels of UC5710, UC1358, and UC859. The fpg expression pattern was unique in that basal levels were not apparently altered by mutations in the oxyR regulatory gene. This last finding was further substantiated by examining the effect of diethylmaleate (DEM) on gene expression of bacteria defective in Trx1 (UC844) compared with the isogenic oxyR::kan derivative (UC1343). DEM is an electrophilic compound that conjugates with GSH and thus is considered a very effective agent for in vivo depletion of glutathione (37). As previously reported (19), GSH depletion by DEM triggered both the grxA and fpg expression in UC844 (Fig. 6); induction ratios similar to the increased basal levels displayed by the second trxA gshA mutant UC1358 (Fig. 5) were detected after 10 min of treatment with 30 mM DEM. However, while the fpg induction ratio was identical in UC844 and UC1343, no significant increment in grxA expression was observed in the trxA oxyR double mutant strain.

View larger version (27K): [in this window] [in a new window]   Fig. 6.   Gene expression induction by diethylmaleate. UC844 (trxA) and UC1343 (trxA oxyR::kan) bacteria were treated for 10 min with 30 mM DEM. cDNAs were amplified by using the set A primers. The fluorescence signal of each PCR product was referred to that of gapA (internal standard). Data were from an average of six independent multiplexed PCR amplifications. Values from treated samples were divided by those from the corresponding control. All genes of set A were analyzed, but only those genes for which highly statistically significant (p  0.001) increases were observed upon diethylmaleate exposure of UC844 are represented. Error bars were estimated from the corresponding S.E.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Taking advantage of the RT/MPCR technique, we have monitored for the first time the simultaneous in vivo transcription of up to 16 different genes in response to oxidative stress by either exposure to chemical oxidants (H₂O₂ or paraquat) or by a decrease in the cellular thiol-disulfide ratio (Trx1 and Grx1 or Trx1 and GSH deficient bacteria). Genes under study code for most components of both glutaredoxin and thioredoxin pathways, some related enzymes, and several pivotal members of the OxyR and SoxRS regulons.

We have been able to show dramatic increases in the yield of grxA transcript (maximal induction ratio of about 50-fold) upon exposure to 100 µM H₂O₂ in an oxyR-dependent and soxRS-independent manner. The OxyR-mediated response to 100 µM H₂O₂ was previously reported (34), yet much lower induction ratios (from 1.3- to 3.0-fold) were quantified for -galactosidase activity by means of two grxA-lacZ operon fusions. We demonstrated that the redox-cycling agent paraquat induces also high levels of grxA expression. However, OxyR, but not SoxRS, was required for paraquat induction, thus indicating an indirect effect via the conversion of O₂ to H₂O₂. The OxyR regulation of grxA expression was further proved by showing that deletion of the oxyR locus from wild-type UC5710 decreases about 3-fold the basal level of grxA transcript, and that, conversely, the grxA message is overexpressed greater than 30-fold upon the introduction of the constitutive oxyR2 mutant allele.

It has been demonstrated that both the thioredoxin and glutaredoxin systems play a role in determining the thiol-disulfide balance in the E. coli cytoplasm (40). Therefore, disulfide bond formation was found to be 8.4 times more efficient in bacteria simultaneously lacking Trx1 and GSH (the physiological reductant of all Grxs) than in wild type. Likewise, we have shown recently that the in vivo transcription of grxA is triggered in trxA gshA double mutant cells (19). Since OxyR becomes active through the oxidative formation of a disulfide bond between two of its cysteine residues (12), it is feasible that trxA gshA mutant cells provide enough titer of oxidized OxyR to activate grxA expression. Here we proved this possibility by demonstrating that a triple trxA gshA oxyR mutant does not show increased basal levels of grxA mRNA and that a double trxA oxyR mutant fails to induce grxA transcription after depletion of GSH with diethylmaleate. In addition, we established that trxA gshA mutants show increased basal expression of other OxyR-regulated genes, like gorA and katG, and we confirmed (data not shown) the induction of OxyS RNA in these mutants in the absence of oxidative stress by H₂O₂ (36). Interestingly, the extent of activation of grxA transcription by either H₂O₂ or deficiencies in the cellular disulfide-reducing systems (trxA gshA double mutants) was much greater (by a factor of 10) than the extent of induction observed for gorA. This suggests the need for higher levels of Grx1 than for glutathione reductase (the enzyme that regenerates its reduced form) in the OxyR-mediated cellular defense against oxidative stress. The extent of katG induction was in between those of grxA and gorA, thus following the same order that the information content previously calculated for their respective OxyR binding sites (i.e. grxA > katG > gorA) (39).

The results presented here demonstrate that different trxA gshA double mutant strains can display different degree of OxyR activation, which seems to indicate differences in their respective ability to form cytoplasmic disulfide bonds. Interestingly, the expression level of the OxyR-regulated genes in these strains correlates with their growth rate (higher for the fastest-growing strain). Suppressor mutations, arising at relatively high frequency, that improve the growth of mutant strains defective in different components of both the thioredoxin and glutaredoxin systems, like trxA grxA, trxB gshA, or trxB gorA, have been previously reported (40, 22, 38). Suppressor mutations unwittingly selected during the manipulation of mutants missing parts of both systems might explain why the ability of proteins to form disulfide bonds in the cytoplasm of these strains is not always altered in the expected ways (38).

Trx2 accounts for the viability of trxA gshA bacteria, since the triple trxA trxC gshA mutant is nonviable in the absence of the disulfide reductant dithiothreitol (41). Our data demonstrate for the first time that the lack of Trx1 and the simultaneous blockage of the GSH/glutaredoxin pathway stimulate the thioredoxin pathway by increasing the steady-state level of trxB and trxC transcripts. Like in the case of the OxyR-regulated genes, the fastest growing trxA gshA bacteria displayed the highest induction ratios. In addition, we showed that transcription of trxB and trxC is up-regulated by the constitutive oxyR2 mutant allele. Taken together, these results indicate that OxyR, either directly or indirectly, play also a role in the regulation of the thioredoxin pathway by controlling Trx2 and the enzyme that regenerates its reduced form. In contrast, we did not observe any significant change in the expression of the gene (trxA) coding for Trx1 either in the expression of grxB or grxC coding for the two new glutaredoxins.

With regard to the nrdAB operon, the data presented show that its transcription is not induced in response to H₂O₂ or paraquat, indicating that nrdAB is not a member of the OxyR or SoxRS regulons. In contrast, nrdAB transcription was readily induced by the RRase inhibitor hydroxyurea (data not shown), at concentrations that block DNA synthesis without effect on other metabolic processes (42). Interestingly, maximal induction ratios by hydroxyurea (7.5- and 5.2-fold for nrdA and nrdB genes, respectively) were similar to the increments in basal level of nrdAB expression displayed by trxA grxA mutant bacteria. Therefore, operation of RRase in the absence of its two main reductans Trx1 and Grx1 must lead to disturbances in deoxyribonucleotide production equivalent to those caused by hydroxyurea treatment (42). Moreover, we demonstrated a tight and inverse relation between the level of nrdAB expression and those of genes coding for some components of both the glutaredoxin (grxA and gorA) and the thioredoxin (trxB and trxC) systems. This compensation between RRase and its reductants might explain why, ironically, the fastest growing trxA gshA mutant bacteria showed the lowest increments in nrdAB mRNA. In the absence of Trx1 and the physiological reductant of all Grxs, Trx2 is the alternate hydrogen donor for RRase in ribonucleotide reduction (38), but a relatively low level of Trx2 expression has been reported (4). If the Trx2 level is limiting, the induction of high levels of trxC message concomitantly to induction of trxB transcription would be clearly beneficial for trxA gshA mutants, thus explaining that the highest increments in trxC and trxB expression, compared with wild type, were observed in the trxA gshA mutant (UC859) that grows faster in both rich and minimal media. Interestingly, OxyR was found to be essential for the viability of UC859.

Like OxyR, the transcription factor SoxR has the unusual property of being inactive under reducing conditions but is activated when the redox potential of the environment becomes more oxidizing (43). Cells mutated in genes for both Trx1 and GSH provide a highly oxidizing cytoplasmic environment by negatively affecting the two major redox-balancing systems in the E. coli cytoplasm (38). The experiments presented here demonstrated that in vivo transcription of the SoxR-dependent soxS and zwf genes is not elevated in trxA gshA mutant cells, indicating that, unlike OxyR, SoxR is not constitutively activated in such a deficient genetic background. These in vivo results are not easily interpreted because while reduced Trx1 promotes in vitro the aerobic assembly of SoxR [2Fe-2S] clusters (17), GSH has an opposite disrupting effect (44). Considerable additional investigation will be required to unravel the in vivo role of thioredoxins and glutaredoxins on SoxR regulation. To this end, some of the mutant strains isolated in this work in combination with the RT/MPCR technique for in vivo analysis of gene expression can be of great utility.

Finally, data on fpg expression deserve some particular comments. Here we demonstrated that fpg transcription is induced in response to oxidative stress by H₂O₂ in an OxyR-dependent manner, therefore suggesting the possibility of the fpg gene being considered a new member of the oxyR regulon. A previous report (45) has excluded such a possibility based on the finding that H₂O₂ (even at concentrations of up to 1000 µM) failed to induce the 8-hydroxyguanine endonuclease activity of the Fpg protein. This apparent discrepancy may be explained if the putative increment in 8-hydroxyguanine endonuclease upon H₂O₂ exposure had fallen back to basal levels after 1 h of treatment (45), in agreement with the observation that OxyR-dependent genes exhibit a remarkably rapid and reversible induction in response to H₂O₂ (31). Besides, our data confirm at the transcriptional level that the induction of 8-hydroxyguanine endonuclease by paraquat is independent of SoxRS (45), but they further attribute such an induction to the conversion of O₂ into H₂O₂. Notwithstanding, in contrast to other OxyR-regulated genes, the fpg mRNA basal level of wild-type cells and the enhanced transcript level exhibited by trxA gshA double mutant strains were not increased to a greater level in a strain constitutive in OxyR. This difference might be a consequence of the complex regulation that apparently controls 8-hydroxyguanine endonuclease in E. coli, as suggested by a recent report (45), demonstrating that this DNA repair activity increases under anaerobic conditions in mutant strains deficient in Fnr in particular, as well as in Fur, ArcA, and combinations thereof.

	ACKNOWLEDGEMENTS

We are indebted to members of the laboratory for discussion and technical advice; to Dr. J. Alhama for GSH determination; and to Dr. B. Weiss, Dr. G. Storz, and Dr. M. Blanco for providing E. coli strains.

	FOOTNOTES

^* This work was supported by Junta de Andalucía (group CVI 0187) Grants PB95-0557-CO2-01 and PB98-1627 (to D. G. E. S.) and by the Swedish Cancer Society.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ Recipient of a postdoctoral contract from the Ministerio de Educación y Cultura.

^¶ Recipient of a predoctoral fellowship from Junta de Andalucía.

^** To whom correspondence and reprint requests should be addressed: C. Pueyo, Departamento de Bioquímica y Biología Molecular, Avda. de Medina Azahara s/n, Universidad de Córdoba, 14071 Córdoba, España. Tel.: 34-957-218695; Fax: 34-957-218688; E-mail: bb1pucuc@uco.es.

	ABBREVIATIONS

The abbreviations used are: Trx, thioredoxin; Grx, glutaredoxin; RRase, ribonucleotide reductase; RT/MPCR, reverse transcription/multiplex polymerase chain reaction; Tc, tetracycline; Kan, kanamycin; Cam, chloramphenicol; Tpx, thiol peroxidase; Fpg, formamidopyrimidine-DNA glycosylase; DEM, diethylmaleate; PCR, polymerase chain reaction.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES