Laboratory of Biotechnology, Chulabhorn Research Institute, Lak Si, Bangkok 10210, Thailand.
Department of Biotechnology, Faculty of Science, Mahidol University, Rama 6 Road, Bangkok, Thailand.
†The first two authors contributed equally to this work. ‡Present address: Section of Microbiology, Cornell University, Ithaca, NY 14853, USA. *For correspondence at the first address. E-mail firstname.lastname@example.org; Tel. (+66) 2 574 0622, ext. 1402; Fax (+66) 2 574 2027.
In Xanthomonas campestris pv. phaseoli, a gene for the alkyl hydroperoxide reductase subunit C (ahpC) had unique patterns of regulation by various forms of OxyR. Reduced OxyR repressed expression of the gene, whereas oxidized OxyR activated its expression. This dual regulation of ahpC is unique and unlike all other OxyR-regulated genes. The ahpC transcription start site was determined. Analysis of the region upstream of the site revealed promoter sequences that had high homology to the Xanthomonas consensus promoter sequence. Data from gel shift experiments indicated that both reduced and oxidized OxyR could bind to the ahpC regulatory region. Moreover, the reduced and the oxidized forms of OxyR gave different DNase I footprint patterns, indicating that they bound to different sites. The oxidized OxyR binding site overlapped the −35 region of the ahpC promoter by a few bases. This position is consistent with the role of the protein in activating transcription of the gene. Binding of reduced OxyR to the ahpC promoter showed an extended DNase I footprint and DNase I hypersensitive sites, suggesting that binding of the protein caused a shift in the binding site and bending of the target DNA. In addition, binding of reduced OxyR completely blocked the −35 region of the ahpC promoter and prevented binding of RNA polymerase, leading to repression of the gene. Monitoring of the ahpC promoter activity in vivo confirmed the location of the oxidized OxyR binding site required for activation of the promoter. A mutant that separated OxyR regulation from basal ahpC promoter activity was constructed. The mutant was unable to respond to oxidants by increasing ahpC expression. Physiologically, it had a slower aerobic growth rate and was more sensitive to organic peroxide killing. This indicated that oxidant induction of ahpC has important physiological roles in normal growth and during oxidative stress.
Xanthomonas campestris pv. phaseoli is an important aerobic bacterial plant pathogen. During plant–microbe interactions, bacteria are exposed to reactive oxygen species (ROS) including H2O2, organic peroxides and superoxide anions synthesized by plants as an active defence response to inhibit bacterial proliferation (Levine et al., 1994; Baker and Orlandi, 1995). Moreover, normal aerobic metabolism generates a considerable quantity of these ROS (Gonzalez-Flecha and Demple, 1997). Xanthomonas species have evolved complex strategies to protect themselves from ROS (Loprasert et al., 1996).
The ability to detoxify organic peroxides is important to cell survival. Organic peroxides participate in free radical reactions that result in increased production of the toxic radicals. Alkyl hydroperoxide reductase (AhpR) is the best-characterized organic peroxide detoxification enzyme in bacteria (Demple, 1991; Storz and Imlay, 1999). The enzyme consists of two subunits, the 22 kDa catalytic C subunit that reduces organic peroxides to corresponding alcohols (Poole, 1996), and the 56 kDa reductase F subunit that is involved in regeneration of oxidized AhpC using either NADPH or NADH as cofactors (Poole and Ellis, 1996). Purified AhpR can also use H2O2 as a substrate (Niimura et al., 1995; Poole, 1996). ahpC is highly conserved and can be found in organisms from bacteria to man (Chae et al., 1994). This suggests that it serves important physiological roles. Mutations in ahpC result in increased sensitivity to organic peroxide toxicity and in an altered overall oxidative stress response (Storz et al., 1989; Antelmann et al., 1996; Bsat et al., 1996; Rocha and Smith, 1999).
In many bacteria, understanding the regulation of ahpC has been of major interest, as alterations in ahpC expression have profound effects on bacterial physiology. For example, in Mycobacterium tuberculosis, ahpC promoter mutations confer resistance to the antibiotic isoniazid (Sherman et al., 1996; Wilson and Collins, 1996; Heym et al., 1997). In Xanthomonas, a multiple peroxide resistance mutant has an unregulated high-level expression of ahpC (Fuangthong and Mongkolsuk, 1997). In all bacterial systems studied thus far, ahpC expression is oxidative stress inducible in a manner dependent on either OxyR, a peroxide sensor and transcription regulator in Gram-negative bacteria, or PerR, a peroxide stress regulator in Gram-positive bacteria (Storz et al., 1990; Bsat et al., 1996; Dhandayuthapani et al., 1997; Mongkolsuk et al., 1998a; Storz and Imlay, 1999). OxyR can exist in two forms; in uninduced cells, OxyR exists in a reduced form. Upon exposure to peroxide, OxyR becomes oxidized by the formation of a disulphide bond at the highly conserved C199–C208 cysteine residues (Zheng et al., 1998). In general, oxidized OxyR activates ahpC expression, whereas reduced OxyR does not interact with the gene promoter region (Toledano et al., 1994). Detailed analysis of molecular interactions between OxyR and ahpC promoters have been performed in a few cases (Toledano et al., 1994; Hattman and Sun, 1997). Results from these studies have shown variation in OxyR binding sites and in their effect on gene expression.
In Xanthomonas, ahpC is transcribed as a monocistronic mRNA. The gene is located upstream of the ahpF–oxyR operon (Loprasert et al., 1997). ahpC expression is inducible by oxidants in an OxyR-dependent manner (Mongkolsuk et al., 1997; 1998a). Here, we describe the effects of different redox states of OxyR on ahpC expression. Molecular interactions between different forms of OxyR and the ahpC promoter were also characterized. In addition, we examined the physiological consequences of altered ahpC regulation during normal growth and under peroxide stress in a mutant in which OxyR regulation has been eliminated.
Results and discussion
OxyR regulation of ahpC expression
Xanthomonas ahpC has unique expression patterns compared with other OxyR-regulated genes in various bacteria. Oxidant induction of the gene is dependent on functional oxyR (Mongkolsuk et al., 1998a). In addition, we have observed that the uninduced AhpC levels in an oxyR mutant are higher than the uninduced levels in a parental strain (Mongkolsuk et al., 1998a). These results suggested that reduced and oxidized OxyR might act differently from regulate ahpC. Experiments were performed to test this hypothesis. It was evident that levels of ahpC mRNA in an oxyR mutant were much higher than uninduced levels in its parental X. campestris pv. phaseoli strain (Fig. 1A). This finding is in agreement with previous Western analysis of AhpC in an oxyR mutant (Mongkolsuk et al., 1998a). In normal growing cells, OxyR exists in reduced form (Zheng et al., 1998). The data implied that ahpC expression was repressed by reduced OxyR in the parental strain and that inactivation of oxyR led to relief of this repression. This observation was extended by determining ahpC mRNA levels in an oxyR mutant harbouring either poxyR (Mongkolsuk et al., 1998a), which produced high levels of the reduced OxyR, or poxyRC199S, which had the critical redox active C199 residue mutated to S199. This mutation locks OxyR in the reduced form (Kullik et al., 1995). The results shown in Fig. 1A showed a threefold reduction in the amount of ahpC mRNA detected in cells harbouring these plasmids when compared with the uninduced level in the X. campestris pv. phaseoli strain. These observations confirmed that reduced OxyR repressed ahpC expression.
The oxyR-dependent oxidant induction of ahpC suggests that oxidized OxyR activates expression of the gene. The idea was tested using an inducing concentration of menadione (MD) to treat X. campestris pv. phaseoli, the strain harbouring poxyR and the oxyR mutant. In oxidant treatment of these cells, MD was used instead of H2O2 because Xanthomonas harbouring various poxyR plasmids produced high levels of catalase that rapidly degraded H2O2. Northern analysis of ahpC mRNA levels of these cells showed that exposure to 100 µM MD induced ahpC expression in X. campestris pv. phaseoli and in the strain harbouring poxyR (Fig. 1B) to high levels. Expression of the ahpC mRNA was induced to similarly high levels in the two strains, even though the uninduced level of the gene was much lower in X. campestris pv. phaseoli poxyR. This suggested that ahpC was expressed close to the maximum level after MD treatment in both strains. As expected, MD induction was dependent on functional oxyR and was absent in the oxyR mutant (Mongkolsuk et al., 1998a). ahpC mRNA levels detected in the oxyR mutant represented unrepressed levels but, in wild-type cells treated with oxidant, ahpC expression was activated further. This indicates that ahpC could be regulated either negatively or positively depending on the redox state of OxyR. This is unique to Xanthomonas ahpC. In all other OxyR-regulated genes in various bacteria, different forms of OxyR either repressed or activated gene expression, but not both from the same promoter (Toledano et al., 1994; Dhandayuthapani et al., 1997; Hattman and Sun, 1997).
Identification of ahpC transcription start sites
To understand the mechanisms of OxyR regulation of ahpC, it was essential to determine the location of the ahpC promoter. Primer extension experiments were performed to locate ahpC transcription start sites. The results presented in Fig. 2 show two major primer extension products designated P1 and P2. These transcription start sites corresponded to positions 40 bp for P1 and 15 bp for P2 upstream of the ahpC translation initiation codon (ATG) (Fig. 5). The Xanthomonas consensus promoter sequences are TTGTNN at the −35 region and T/GATNAA/T at the −10 region, although the distance between these two sequence motifs varied from 16 bp to 24 bp (Katzen et al., 1996). Examination of the upstream region of the weaker transcription start site (P1) showed sequences TTGAGG and TACCAT at the −35 and −10 regions of P1, respectively, and they were separated by 17 bp. These two regions of the P1 promoter show five out of six nucleotides that matched the promoter consensus sequences and have perfect distance between the two conserved regions. The data suggest that P1 probably acts as a strong promoter in vivo. On the other hand, P1 accounted for less than 30% of the primer extension products. Inspection of the region upstream of the start of the major primer extension product (P2) revealed no promoter-like sequences that had homology to Xanthomonas promoter consensus sequences. We suggest that the P2 product probably corresponds to premature termination of reverse transcriptase caused by the presence of poly(C) residues (13 C out of 18 residues) (Fig. 5) upstream of the P2 transcription start site.
Next, we examined whether the increased amounts of ahpC mRNA in response to the oxidant treatments shown in Fig. 1B resulted from increased transcription initiation of the gene. Primer extension experiments were performed on RNA samples extracted from uninduced and MD-induced Xanthomonas cultures (Fig. 2). In the uninduced sample, there was much less primer extension product compared with the MD-induced sample (Fig. 2). This suggested that increased transcription initiation was responsible for the induction of ahpC expression. Moreover, the observation was consistent with the model in which oxidized OxyR activates gene transcription by recruiting RNA polymerase to the promoter region (Toledano et al., 1994).
Different forms of OxyR had differed binding affinity to the ahpC promoter region
To understand better the regulation of ahpC by various forms of OxyR, the protein–DNA binding gel shift technique was used to investigate binding of OxyR to the ahpC promoter. OxyR was purified as described in Experimental procedures. The results of a gel shift experiment are shown in Fig. 3A. The addition of increasing amounts of purified OxyR to end-labelled DNA fragments containing the ahpC promoter (from −112 to +93) resulted in the retardation of fragment migration in native polyacrylamide gels. Two slower migrating DNA bands designated S1 and S2 were detected when OxyR was added to the binding reactions (Fig. 3A). The faster migrating band S1 was detected when low concentrations of OxyR were added to binding reactions. As OxyR concentrations increased, a second slower migrating DNA fragment (S2) was apparent. The addition of a 100-fold greater concentration of a non-specific protein such as bovine serum albumin (BSA) did not produce any DNA fragment mobility shift (data not shown). However, adding an anti-OxyR polyclonal antibody to the binding reaction after the addition of OxyR resulted in supershifting of the protein–DNA complex (Fig. 3A). These data confirmed that the observed gel retardation of DNA fragments resulted from OxyR binding to the ahpC promoter. Also, at high concentrations, OxyR might bind to the ahpC promoter as an oligomer. This could be responsible for the observed S2 band in the gel shift experiment (Fig. 3A).
We wished to determine and to compare the affinity of the oxidized and the reduced forms of OxyR for the ahpC promoter using gel shift assays. Dithiothreitol (DTT; 200 mM) was added to the reaction to reduce OxyR. The results in Fig. 3B show that both oxidized and reduced OxyR could bind to the DNA fragment containing the ahpC promoter. As the binding reactions were carried out at low OxyR concentrations, the second (S2) species of retarded DNA fragment was not detected. Densitometer analysis of the ratio of bound versus unbound DNA fragments indicated that reduced OxyR had around 20% higher affinity for the ahpC promoter than oxidized OxyR. This small difference in affinity was observed consistently even when different batches of OxyR and labelled ahpC promoter fragments were used. This supports the observations from Fig. 1 that reduced OxyR in uninduced cells would bind to the ahpC promoter and inhibit its transcription. Unlike other bacteria, exposure of Xanthomonas to oxidants induces synthesis and accumulation of oxidized OxyR (Mongkolsuk et al., 1997). Although oxidized OxyR has lower affinity for the ahpC promoter, this does not affect the activation of the promoter by the protein caused by an increased concentration of oxidized OxyR as a result of oxidant treatment.
Determination of reduced and oxidized OxyR binding sites
Xanthomonas ahpC promoter is unique with regard to differential regulation by reduced and oxidized forms of OxyR. This allowed for an examination of how different forms of OxyR interacted with their binding sites to achieve differential effects on a single promoter. DNase I footprinting experiments were performed to locate OxyR binding sites (Fig. 4). Clearly, binding of the reduced and oxidized OxyR to both top and bottom strands gave distinct footprint patterns. On the bottom strand, binding of oxidized OxyR protected a region from −80C to −34G, whereas binding of reduced OxyR showed an extended footprint covering the region −67A to −24T. In addition, a partial DNase I protected region from −80C to −67A was also detected (Figs 4 and 5). Major and minor DNase I hypersensitive sites in the middle of the reduced OxyR protected region at −52A, −53T and −48C, respectively, were also observed (Fig. 4). On the top strands, binding of reduced or oxidized OxyR protected DNA from −78A to −28T and from −73A to −15C respectively.
The data are consistent with the idea that oxidized OxyR binds to four successive major grooves on one face of the helix, whereas reduced OxyR binds to a pair of major grooves separated by one helical turn (Toledano et al., 1994). Binding of reduced OxyR introduces bending in the target DNA that results in DNase I hypersensitive sites. These observations are similar to DNase I footprints of reduced OxyR high-affinity binding sites in two other promoters, Escherichia coli oxyR (Toledano et al., 1994) and the Mu phage mom operon (Sun and Hattman, 1996; Hattman and Sun, 1997), in which the protein represses expression from both promoters.
OxyR has extended binding sites, and the proposed consensus binding motif for E. coli has twofold dyad symmetry with the following sequence: ATAGntnnnanCTATnnnnnnnATAGntnnnanCTAT (Tartaglia et al., 1992; Toledano et al., 1994). Examination of the putative oxidized OxyR binding site (ATAGnxnnnanCTATnnnnnnnATxxnxnnnanCxAT) upstream of the Xanthomonas ahpC promoter showed a sequence motif that had 14 out of 20 bases matched to the consensus OxyR binding site (Fig. 5). In addition, it also contained the highly conserved LysR-binding motif, T-N11-A (Schell, 1993). The proposed location of the binding site is in agreement with the footprinting data. The degree of homology of the Xanthomonas ahpC promoter to the consensus sequence is similar to other binding sites of OxyR-regulated genes in E. coli (Toledano et al., 1994). Also, the Xanthomonas OxyR binding site has closer homology to the E. coli consensus sequence than to the proposed shorter binding site for Mycobacterium spp. (Dhandayuthapani et al., 1997). The E. coli reduced OxyR binding is extended by one helical turn, and this is responsible for extended footprints. This suggests that a fifth contact point is essential for the reduced OxyR binding (Toledano et al., 1994). We searched for a motif in the fifth region of the ahpC promoter and identified a CTAT motif 9 bases from the oxidized OxyR binding motif. Examination of two other high-affinity reduced OxyR binding sites showed that, 9 bases from the oxidized OxyR binding sites, there are motifs CxAT for mom and CxAx for oxyR. The location of this motif corresponds to the fifth motif of the OxyR binding site of the E. coli oxyR promoter. The deletion analysis of the OxyR binding site of the E. coli oxyR promoter has shown that removal of the fifth region affected reduced OxyR binding (Toledano et al., 1994).
The region of partial DNase I protection and two DNase I hypersensitive sites caused by binding of the reduced OxyR to the region are unusual. Closer examination of the protected region revealed an alternative OxyR binding site. The alternative binding site (ATxxnntnnnxnCxATnnnnnnnATxxnxnnnanCTAT) matched at 13 out of 20 bases of the consensus OxyR binding site (Fig. 5). This site was protected from DNase I digestion by reduced OxyR, but not by oxidized OxyR (Fig. 4). It is possible that in vitro reduced OxyR could bind to both sites, leading to two DNase I hypersensitive sites and an area of partial DNase I protection. We are examining whether the alternative OxyR binding site functions in vivo. Regardless, binding of reduced OxyR to either site completely blocked the −35 region of the ahpC promoter leading to repression of the gene.
In vivo promoter analysis
The ability of a DNA fragment containing the ahpC promoter and OxyR binding sites to direct oxidant-inducible expression of a reporter cat gene in vivo was tested. The experiment was performed to confirm data from primer extension and DNase I footprinting experiments, which located regions important for OxyR-dependent, oxidant-inducible expression of ahpC. DNA fragments containing the ahpC promoter either with (pUTTnCP1) or without (pUTTnCP2) OxyR binding sites were transcriptionally fused to a promoterless cat gene and subsequently cloned into a mini-Tn5 vector (De Lorenzo and Timmis, 1994). pUTTnCP1 and pUTTnCP2 had DNA fragments up to −209 and −39, respectively, plus the cat reporter gene and the rest of ahpC. These mini-Tn5 constructs were mobilized and transposed into the X. campestris pv. phaseoli chromosome. The ahpC promoter activity in X. campestris pv. phaseoli TnCP1 and TnCP2 was determined by Western analysis of Cat (Fig. 6). Densitometer analysis of the data showed that MD treatment induced more than twofold increased Cat levels in the strain containing TnCP1. Similar treatment did not result in increased Cat levels in the strain containing TnCP2. Oxidant-inducible cat expression in the strain containing TnCP1 and lack of induction in the strain containing TnCP2 confirmed the proposed binding site of the oxidized OxyR required for in vivo activation of the promoter. Unexpectedly, basal Cat levels in the strain containing TnCP2 were threefold less than those in the strain containing TnCP1. The ClaI deletion of the ahpC promoter removed most of the reduced OxyR binding sites, but left the promoter intact. This should relieve reduced OxyR repression of the promoter, leading to higher levels of gene expression from the ahpC promoter similar to the higher levels of ahpC expression in the oxyR mutant (Fig. 1). The contradictory results shown in Fig. 6 suggested that the ahpC promoter might require regions upstream of −35 for full uninduced activity. Interestingly, the ClaI deletion of the ahpC promoter (TnCP2) had half the extended reduced OxyR binding site that contained a CTAT palindrome sequence and a conserved LysR-binding motif (T-N11-A) (Schell, 1993) intact (Fig. 5). In Mycobacterium, a shorter region containing one palindrome sequence and a conserved LysR binding site is thought to be sufficient for OxyR binding (Dhandayuthapani et al., 1997; Pagan-Ramos et al., 1998). This shorter OxyR binding site has not been demonstrated to be functional in Gram-negative bacteria. In order to account for the Cat levels in the strain containing TnCP2 (Fig. 6), we proposed that, in the absence of normal OxyR binding sites, it is possible that OxyR could bind to the half-palindrome CTAT site, as exposure to an oxidant did not effect this repression, suggesting that both reduced and oxidized OxyR could bind to the half-site. The half OxyR binding site and the conserved LysR-binding motif are located between the −35 and −10 regions of the ahpC promoter. Binding of either reduced or oxidized OxyR to these sites prevents RNA polymerase binding to the promoter by blocking the −35 region. The results in Fig. 6 show that basal levels of Cat specified by TnCP2 in the oxyR mutant were higher than the Cat level attained in a parental strain. These results suggested that, in the absence of OxyR, no repression occurred at the ClaI-deleted ahpC promoter.
Characterization of a regulatory mutant of ahpC
Alterations in ahpC expression in various bacteria have important physiological consequences (Antelmann et al., 1996; Dhandayuthapani et al., 1996; Sherman et al., 1996; Fuangthong and Mongkolsuk, 1997; Rocha and Smith, 1999). The majority of known mutations are located in promoter or structural regions. No mutants have been constructed in genes for oxidative stress protection that separated oxidant regulation from basal expression levels without inactivation of either structural or regulatory genes. Such mutants would permit analysis of the role of oxidant-induced alteration in gene expression on bacterial physiological responses. We were interested in making an ahpC mutant that uncoupled gene regulation from the basal level of gene expression without knocking out either ahpC or oxyR. This would provide a unique opportunity to examine the physiological roles of oxidant induction of ahpC expression during normal growth and peroxide stress. An ahpC mutant was constructed by the insertion of an eryR at the ClaI site (Fig. 7A). This separated OxyR binding sites from the ahpC promoter. The mutated gene was transferred into X. campestris pv. phaseoli and marker exchanged into the chromosome, as described in Experimental procedures and Fig. 7A. This resulted in a X. campestris pv. phaseoli ahpCE mutant (Fig. 7). The integrity of the mutant was confirmed by Southern analysis (data not shown). Northern analysis revealed that ahpC was no longer inducible by menadione in the mutant (Fig. 7B). The data confirmed that insertion of an eryR gene between the OxyR binding site and the promoter separated OxyR activation of gene expression from normal promoter functions.
We have observed that mutations in genes involved in oxidative stress response often resulted in altered aerobic growth and sensitivity to oxidants (Mongkolsuk et al., 1996; 1998a). Mutations in ohr, a gene involved in organic peroxide protection in Xanthomonas, show increased sensitivity to organic peroxide killing (Mongkolsuk et al., 1998b). Next, we determined the physiological consequences of separation of OxyR regulation from basal ahpC promoter functions. First, the mutant aerobic growth rate was determined in a complex SB medium (Fig. 8A). Under these conditions, the mutant had a slower doubling time of 140 min, compared with 110 min for the parental strain. Nevertheless, the mutant reached a similar density to the parental strain by stationary phase. Next, we examined growth of the mutant in the presence of a low concentration (150 µM) of tBOOH. In the presence of tBOOH, the mutant had a doubling time of 170 min, compared with 110 min for the parental strain (Fig. 8A). We extended these observations by determining the mutant resistance levels to killing concentrations of tBOOH and H2O2. With killing concentrations of tBOOH and H2O2, respectively, the mutant showed a growth inhibition zone of 16 mm and 15 mm, compared with 10.5 mm and 19 mm in the parental strain (Fig. 8B). Normal aerobic metabolism generated H2O2 and organic peroxide, which could induce ahpC expression via OxyR (Gonzalez-Flecha and Demple, 1997). In the mutant, the inability to respond to increased levels of oxidant probably led to the accumulation of toxic organic peroxides that resulted in the observed slower growth rate (Fig. 8A). This deficiency was accentuated when the mutant was exposed to exogenous organic peroxide, as shown by severely reduced growth rate and resistance levels. Unexpectedly, the mutant had small increases in resistance to H2O2 killing, which correlated with a small increase in total catalase levels. The mutant had 3.5 U mg−1 compared with 5.6 U mg−1 protein in the parental strain. The small increase in catalase levels could result from a compensatory response in the mutant for its inability to increase ahpC expression in response to changes in the environment. A similar compensatory response has been observed in Bacillus subtilis (Antelmann et al., 1996; Bsat et al., 1996).
In the absence of exogenous oxidants or under growth conditions in which fewer oxidants were being generated intracellularly, ahpC expression would be repressed by reduced OxyR. However, when cells are exposed to oxidants up-expression of the gene can be achieved rapidly by activation of the promoter by oxidized OxyR. The dual regulation of the ahpC promoter by both reduced and oxidized forms of OxyR allows fine tuning of AhpC levels. Analysis of the mutant that lacks OxyR regulation of ahpC expression physiological responses suggested that fine tuning of gene regulation is important to the cells. In the mutant, the gene basal level of expression was not sufficient to protect cells from oxidants generated intracellularly, as reflected by a slower aerobic growth rate. This deficiency was more pronounced when the mutant was challenged with organic peroxide.
Growth, culture and oxidant killing conditions
All Xanthomonas strains were grown aerobically in SB medium at 28°C. E. coli strains were grown in LB at 37°C. For Xanthomonas, the following concentrations of antibiotics were used; kanamycin 30 µg ml−1; erythromycin 100 µg ml−1. To test the effects of oxidants on the growth rate of Xanthomonas, overnight cultures were diluted into fresh SB medium to give A600 of 0.1. The cultures were allowed to grow for 1 h before oxidants were added to give desired final concentrations. Growth was monitored spectrophotometrically at A600. The killing zone method for determining the sensitivity of a strain to oxidant killing was performed by adding 108 log phase cells to 3 ml of warm top SB agar. The mixture was poured onto an SB agar plate. Then, 6 mm paper discs containing 7 µl of desired concentrations of oxidants were placed on the cell lawn. It was essential to use cells from similar stages of growth, as levels of resistance to oxidant killing in Xanthomonas varies with stage of growth (Vattanaviboon et al., 1995). Zone of growth inhibition was measured after 24 h incubation.
Nucleic acids isolation and analysis
Genomic DNA extraction from Xanthomonas strains and Southern blot experiments were performed as described previously (Mongkolsuk et al., 1996). Total RNA was isolated using the modified hot acid phenol method (Mongkolsuk et al., 1996). For Northern blot experiments, RNA samples were separated on formaldehyde agarose gels and subsequently capillary transferred to nylon membranes. Radioactive labelled ahpC probes were prepared from polymerase chain reaction (PCR) fragments containing the ahpC coding region (Mongkolsuk et al., 1997). PCR fragments were purified from agarose gels and radioactively labelled using a commercial random prime kit with [α-32P]-dCTP. The probes were boiled and added to hybridization bags. Prehybridization, hybridization and washing were performed under high-stringency conditions as described previously (Mongkolsuk et al., 1996).
Primer extension experiments were carried out to determine ahpC transcription start sites (Storz and Altuvia, 1994). PE1 primer (5′-TTGCCGTTGTGGTACGCATT-3′) located at nucleotide position 80–99 of ahpC (Fig. 5) was labelled with T4 polynucleotide kinase and [γ-32P]-ATP. The labelled primer was annealed with 5 µg of total RNA and incubated further at 50°C for 30 min. Then, 200 units of Superscript II MMLV reverse transcriptase was added to the reaction, and incubation was continued at 42°C for 60 min. The extension products were analysed on sequencing gels next to sequence ladders.
Purification of OxyR.
Plasmid pOXX (Mongkolsuk et al., 1998a) containing oxyR was used as a template in PCR reactions with primer A (5′-CGTCTAGAAGGCTGCTGCATAT-3′) and primer B (5′-TTGTCGACAGCCGCAACCGCCTT-3′), which covered the 5′ and 3′ regions of the gene respectively. The 940 bp PCR products were digested with SalI and XbaI and cloned into pCYB4 (Biolabs) digested with XhoI and XbaI. The recombinant plasmid pINT-oxyR was transformed into an E. coli GSO8 (oxyR <kanR>). The OxyR fusion to intein was verified using both anti-intein and anti-OxyR antibodies. For purification of OxyR, GSO8 harbouring pINT-oxyR were grown in LB to log phase and induced with 2 mM IPTG for 6 h at 28°C. We found that induced expression of oxyR at 28°C gave better product yields. Cells were pelleted and washed once with Tris buffer, pH 7.8. The pellet was resuspended in binding buffer (20 mM Tris, pH 8.0, 500 mM NaCl and 0.1 mM EDTA) plus 0.1% Triton X-100 before being sonicated on ice. A clear lysate was obtained after centrifugation at 10 000 g for 10 min, and it was used to bind to chitin beads for 30 min before the mixture was loaded into a column and washed extensively. Bound fusion OxyR protein was cleaved by incubating the column content in a cleavage buffer (binding buffer plus 300 mM DTT) overnight at 4°C. Fusion protein was eluted by washing the column with cleavage buffer, and 0.5 ml fractions were collected. These fractions were analysed by SDS–PAGE gels, and fractions that contained a high concentration of purified OxyR protein were pooled and dialysed against the binding buffer at 4°C overnight. OxyR protein purity was greater than 90% judged by SDS–PAGE.
Gel shift and DNase I footprinting
Both gel shift and DNase I protection experiments were performed as described previously (Storz and Altuvia, 1994). End-labelled DNA fragments (205 bp) were used. Radioactively labelled DNA fragments were prepared by labelling primer PE1 with [γ-32P]-ATP and T4 kinase. The labelled primer was mixed with plasmid pKSahpC and a second unlabelled primer 127 (5′-TAGGATCCACTGCGACTG-3′) in PCR reactions performed for 25 cycles. The 205 bp labelled products were purified from agarose gel and used in gel shift and DNase I footprinting experiments. For DNase I footprinting of the top strand, primer 127 was end labelled with [γ-32P]-ATP and T4 kinase. The labelled primer was mixed with pahpC and primer 213 (5′-CCGACCTTGCGACGAA-3′) in PCR reactions. The 746 bp PCR products were then cleaved with NaeI, end labelled with a 270 bp fragment purified from agarose gels and used in DNase I footprinting experiments.
The gel shift reactions were performed by adding 3 fmol of labelled probe to TM buffer (50 mM Tris-HCl, pH 7.9, 12.5 mM MgCl2, 20% glycerol, 1 mM EDTA, pH 8.0, 0.1% Nonidet P-40 and 100 mM KCl). Purified OxyR was added to give a final concentration of 0.5 × TM in 25 µl. To assay OxyR binding under reducing conditions, 200 mM DTT was added to binding reactions (Storz and Altuvia, 1994). For DNase I protection assays, the same labelled fragment and OxyR binding conditions were used as described in the gel shift assay. After 10 min incubation, 25 µl of Mg2+ and Ca2+ and 0.5 U of DNase I were added. The reaction was continued for 1 min before 200 µl of a stop solution (20 mM EDTA, pH 8.0, 1.0% SDS, 0.2 M NaCl and 250 µg ml−1 tRNA) was added. The mixture was extracted with phenol–chloroform and ethanol precipitated. Dry pellets were mixed with a sequencing loading buffer and loaded onto a sequencing gel. The DNA sequence ladder was performed using fmolR sequencing kits (Promega) and radioactively labelled primer 192 on pKSahpC template.
Molecular cloning of ahpC and promoter analysis
Study of regulated promoters on a multiple-copy promoter probe vector often leads to deregulation of the promoter caused by titrating out a limited amount of regulatory factors. To avoid this problem, ahpC promoter activity was monitored in cells with transposons containing ahpC promoter fused to a reported gene integrated into the chromosome. PCR fragments (895 bp) containing Xp ahpC from −208 to +687 bp were cloned into pBluescript KS. The PCR primers corresponding to 5′ (5′-TAGGATCCGGCGAAGCAACTG-3′) and 3′ (5′-GTAAGCTTCCGGCACCGGCTC-3′) of ahpC were added to 0.5 µg of Xp genomic DNA, dNTP and buffer and amplified using the following conditions; 96°C denaturing, 60°C annealing of primers and 72°C extension of polymerase reaction. Taq polymerase (2 U) was added at the start of the reaction. PCR products were purified and cloned into pBluescript KS. The recombinant plasmid pKSahpC was transformed into dam−E. coli. The promoterless chloramphenicol acetyltransferase gene (cat) from pSM-cat1 (Mongkolsuk et al., 1993) digested with BamHI and BglII was gel purified and cloned into pKSahpC digested with BclI. This resulted in a new recombinant plasmid pKSahpCcat containing an ahpC–cat transcriptional fusion. The plasmid was then digested with XhoI–HindIII, and a 1.7 kb fragment containing the ahpC–cat fusion was cloned into SalI–HindIII-digested pUC18Sfi, resulting in pUC18Sfi-ahpCcat1 (De Lorenzo and Timmis, 1994). A second construct containing the ClaI deletion was made by digesting pUC18Sfi-ahpCcat1 with ClaI. It was then gap filled with DNA polymerase and SmaI, followed by blunt-ended DNA ligation. This resulted in pUC18Sfi-ahpCcat2. The ahpC–cat fusions in pUC18Sfi were cloned into a mini-Tn5 vector pUTTn5lacZ1 (De Lorenzo and Timmis, 1994). pUTTn5lacZ1 was then digested with SfiI. The vector portion containing a mini-Tn5 and a selectable marker was gel purified and ligated with the SfiI fragments containing ahpC–cat from SfiI-digested pUC18Sfi-ahpCcat1 or pUC18Sfi-ahpCcat2. This resulted in two new recombinant plasmids pUTahpCcat1 and pUTahpCcat2. These plasmids were transformed into E. coli S-17λpir (TpR SmRrecA, thi, pro, hsdR−M+ RP4:2-Tc:Mu:KmR Tn7, λpir) and subsequently conjugated into X. campestris pv. phaseoli. Conjugants were selected for RifR and KanR and scored for carbeniciline sensitivity. To determine the in vivo ahpC promoter activity, cells containing TnCP1 and TnCP2 were grown under uninduced and menadione-induced (100 µM) conditions. Cat levels were monitored using Western blot analysis with an anti-Cat polyclonal antibody. Experiments were repeated four times.
Construction of an ahpCE mutant
eryR derived from pE194 was digested with BamHI and EcoRV followed by gap filling with DNA polymerase I. The blunt-ended 0.9 kb fragment was gel purified and cloned into the ClaI site (Fig. 5) of pahpC4.1 (Loprasert et al., 1997), resulting in pahpCE, which conferred ApR and EryR. pahpCE was electroporated into X. campestris pv. phaseoli using previously described conditions (Mongkolsuk et al., 1997). EryR could arise from a single recombination of the plasmid with the chromosome. Transformants were selected for EryR and scored for ApS. Several putative ahpCE mutants were characterized at both Southern (data not shown) and Western levels. One of the mutants that showed correct restriction enzyme patterns was selected for physiological analysis.
We thank Tim Flegel for reviewing the manuscript, and G. Storz for useful comments and suggestions. P. S. Lovett has kindly provided derivatives of the pE194 plasmid. This research was supported by a grant to the Laboratory of Biotechnology from the Chulabhorn Research Institute, a Thailand Research Fund BRG 10-40 grant and a NSTDA Career Development award, RCF 01-40-005 to S.M.
†The first two authors contributed equally to this work.
‡Present address: Section of Microbiology, Cornell University, Ithaca, NY 14853, USA.