The repressor for an organic peroxide-inducible operon is uniquely regulated at multiple levels

Authors


  • *For correspondence. E-mail skorn@tubtim.cri.or.th; Tel. (+66) 2 574 0630; Fax (+66) 2 574 2027. Present addresses:

Summary

ohrR encodes a novel organic peroxide-inducible transcription repressor, and we have demonstrated that ohrR is regulated at the transcriptional and the post-transcriptional levels. Primer extension results show that ohrR transcription initiates at the A residue of the ATG translation initiation codon for the ohrR coding sequence. Thus, the gene has a leaderless mRNA. The ohrR promoter (P1) has high homology to the consensus sequence for Xanthomonas promoters, which is reflected in the high in vivo promoter activity of P1. Deletion of a 139 bp fragment containing the P1 promoter showed that the sequences upstream of –35 regions were required for neither the promoter activity nor OhrR autoregulation. In vitro, purified OhrR specifically binds to the P1 promoter. DNase I footprinting of OhrR binding to the P1 revealed a 44 bp region of protection on both DNA strands. The protected regions include the –35 and –10 regions of P1. We suggest that OhrR represses gene expression by blocking RNA polymerase binding to the promoter. There are two steps in the post-transcriptional regulation of ohrR, namely differential stability and inefficient translation of the mRNA. The bicistronic ohrR–ohr mRNA was highly labile and underwent rapid processing in vivo to give only stable monocistronic ohr mRNA and undetectable ohrR mRNA. Furthermore, the ohrR mRNA was inefficiently translated. We propose that, in uninduced cells, the concentration of OhrR is maintained at low levels by the autoregulation mechanism at the transcriptional levels and by the ohrR mRNA instability coupled with inefficient translation at the post-transcriptional level. Upon exposure to an organic peroxide, the compound probably interacts with OhrR and prevents it from repressing the P1 promoter, thus allowing high-level expression of the ohrR–ohr operon. The rapid processing of bicistronic mRNA gives highly stable ohr mRNA and corresponding high levels of Ohr, which remove an organic per-oxide. Once the peroxide has been removed, the autoregulation mechanism feeds back to inhibit the expression of the operon.

Introduction

Xanthomonas is a soil bacterium that causes diseases in plants. In the environment, Xanthomonas is exposed to reactive oxygen species (ROS) from a variety of sources (Baker and Orlandi, 1995; Gonzalez-Flecha and Demple, 1997). These ROS (e.g. superoxide, H2O2 and organic peroxide) are highly toxic to biological systems (Halliwell and Gutteridge, 1984). Bacteria have evolved multiple protective pathways to ensure their detoxification. Bac-terial defence against organic peroxide is a complex process involving several structural and regulatory genes (Storz and Imlay, 1999).

Alkyl hydroperoxide reductase is the best characterized organic peroxide detoxification enzyme in bacteria (Poole, 1996; Poole and Ellis, 1996). ahpC encodes the catalytic subunit of the enzyme and is regulated by OxyR, the global regulator of peroxide stress response (Storz and Imlay, 1999). ahpC expression is induced by exposure of bacteria to various oxidants (Bsat et al., 1997; Loprasert et al., 1997; Rocha and Smith, 1999; Storz and Imlay, 1999). Inactivation of ahpC results in pleiotropic changes in oxidative stress response, suggesting that it involves processes other than organic peroxide detoxification (Bsat et al., 1996; Rocha and Smith, 1999; Mongkolsuk et al., 2000; Seaver and Imlay, 2001).

We discovered a novel family of genes in Xanthomonas designated ohr that are involved in organic peroxide resistance (Mongkolsuk et al., 1998). ohr homologues have subsequently been found in both Gram-positive and Gram-negative bacteria (Atichartpongkul et al., 2001; Fuangthong et al., 2001; Ochsner et al., 2001; Rince et al., 2001), and inactivation of the homologues in these bacteria leads to reduced resistance to organic peroxide but not to other oxidants (Mongkolsuk et al., 1998; Atichartpongkul et al., 2001; Fuangthong et al., 2001; Ochsner et al., 2001). ohr has an unique expression pattern; its expression is induced only by exposure of bacteria to organic peroxide, a feature highly conserved among diverse bacteria. Moreover, ohr expression is independent of OxyR (Sukchawalit et al., 2001). Recently, we identified a regulatory gene ohrR that belongs to a family of organic peroxide-sensing transcription repressor genes in an operon with ohr (Fig. 1A; Fuangthong et al., 2001; Sukchawalit et al., 2001). In both Gram-positive and Gram-negative bacteria, organic peroxide-induced expression of ohr depends on a functional ohrR (Fuangthong et al., 2001; Sukchawalit et al., 2001). However, regulation of ohr by ohrR in Xanthomonas is complex. Current evidence suggests that induction of ohr expression by organic peroxide requires a functional OhrR, a strong upstream ohrR promoter (P1) and an RNA processing step (Sukchawalit et al., 2001). Here, we have characterized the ohrR promoter (P1) and show that the repressor, OhrR, interacts specifically with its promoter to autoregulate itself. Post-transcriptional control of ohrR was also investigated and found to occur at the level of differential mRNA stability and inefficient translation of the mRNA. These mechanisms maintain a low intracellular concentration of OhrR.

Figure 1.

ohrR–ohr operon and summary of various important transcriptional regulatory elements in P1.

A. ohrR–ohr operon. P1 and arrowhead indicate promoter and direction of transcription. B, BamHI; K, KpnI; N, NcoI; P, PstI; and S, SfiI.

B. The locations of various deletions from the 5′ of P1, the direct repeats D1 and D2, an inverted repeat I1, the –35 and –10 regions of the promoter are shown. The DNase I-protected regions resulting from OhrR binding to P1 of coding and non-coding strands are marked by brackets, and numbers at each end of the brackets indicate the location of the beginning and end of the protected region with respect to the transcription start site (+1). Bold ATG is the ohrR translation initiation codon.

Results and discussion

Primer extension and identification of the P1 promoter

We have identified a DNA fragment upstream of ohrR containing a promoter that is responsible for regulated ohrR and ohr expression (Sukchawalit et al., 2001). Primer extension was performed on RNA samples from uninduced and tert-butyl hydroperoxide (tBOOH)-induced samples to locate the ohrR transcription start site. The results show that the major primer extension products corresponded to the transcription start site located at the A residue of the translation initiation codon, ATG of ohrR (Fig. 2). Examination of ohrR sequence shows that the gene has no ribosome binding site (RBS) upstream of the translation initiation codon. Thus, transcription and translation of the gene initiate from the same site, resulting in a leaderless mRNA.

Figure 2.

Primer extension analysis of ohrR. Primer extension was done on RNA samples isolated from uninduced (U) and tBOOH (I)-induced cultures. The sequence ladder (G, A, T, C) was done according to Experimental procedures. The location of the ohrR transcription start site is shown by the arrow, and the bold ATG indicates the translation initiation codon of ohrR.

Sequence analysis of the region upstream of the ohrR transcription initiation site revealed sequence motifs with high homology to the consensus promoter sequence for Xanthomonas (Katzen et al., 1996). The sequences TTGCAA and TATAAA have five out of six and six out of six matches to the TTGTNN and G/TATNAA sequences for the –35 and –10 regions of the consensus promoter sequence for Xanthomonas respectively. The –35 and –10 regions of ohrR promoter are separated by 15 bp (Fig. 1B). Thus, the ohrR promoter should have a high promoter activity, an assumption supported by the in vivo analysis of P1 promoter activity (Fig. 3; Sukchawalit et al., 2001). No other transcription start sites or sequences resembling the consensus sequence for a Xanthomonas promoter were found in the 139 bp P1 fragment. The ohrR promoter is designated P1.

Figure 3.

Deletion analysis of a 139 bp DNA fragment containing P1 and the OhrR operator. Western analysis of Cat specified by pP1-cat and various deleted P1 fragments cloned in the pUFRcat2-Km promoter probe vector.

A. X. campestris pv. phaseoli harbouring various P1 deleted plasmids.

B. X. campestris pv. phaseoli harbouring pBBRohrR and various P1 deleted plasmids. The locations of various deletions are shown in Fig. 1B. Total protein (10 μg) was loaded into each lane. U, uninduced; I, tBOOH induced.

Analysis of primer extension products performed on RNA samples isolated from uninduced and tBOOH-induced Xanthomonas campestris pv.◊phaseoli cultures show that the tBOOH treatment clearly increased ohrR transcription severalfold over the uninduced levels (Fig. 2). This suggests that tBOOH-inducible expression of ohrR and ohr, detected by Northern and Western experiments (Mongkolsuk et al., 1998; Sukchawalit et al., 2001), results primarily from increased transcription ini-tiation from P1. These findings also suggest that OhrR represses gene expression by preventing RNA polymerase binding to the promoter.

In vivo deletion analysis and autoregulation of P1 by OhrR

We performed sequential deletions from the 5′ end of the 139 bp fragment containing P1 to identify regions important for the promoter activity. The locations of various deletions are summarized in Fig. 1B. The deleted fragments were cloned upstream of a promoterless chloramphenicol acetyltransferase gene (cat) in pUFRcat2-Km, a low-copy-number, broad-host-range promoter probe vector (Mongkolsuk et al., 1993), and transformed into X. campestris pv.◊phaseoli. Western analysis of Cat levels showed that deletions upstream of the proposed –35 region had no effect on P1 promoter activity, whereas a deletion that removed the –35 region abolished the promoter activity (Fig. 3A). The data are therefore consistent with our proposed locations for the –35 and –10 regions of P1.

Next, the OhrR autoregulation of P1 was investigated. pBBRohrR (ohrR in an expression vector; Sukchawalit et al., 2001) was transformed into cells harbouring various deletions of P1 fragments in the promoter probe vector, and Cat levels were determined by Western analysis. High-level expression of ohrR from an expression vector strongly repressed the P1 promoter activity (Fig. 3B). This confirms the role of OhrR as a negative autoregulator of P1. Furthermore, the Western results show that, in all strains harbouring plasmids with deletions upstream of the –35 region of P1, OhrR could strongly repress the promoter activity (Fig. 3B). Removal of half the direct repeat sequence ATAAATCGC (deletion no. 87) had no effect on OhrR repression of P1.

The autoregulation mechanism is a common feature among genes coding for transcriptional repressors including several marR family members (Martin et al., 1995; Poole et al., 1996; Xiong et al., 2000). The mechanism allows fine tuning of the repressor concentration and prevents excessive synthesis of the repressor. Nonetheless, ohrR autoregulation is not highly conserved in other bacteria. In Bacillus subtilis, OhrR regulates ohrA (an ohr homologue) but not itself; B. subtilis ohrR appears to be regulated by sigma factor A (SigA) (Fuangthong et al., 2001). This illustrates a different strategy that diverse bacteria can use to regulate an important transcriptional modulator. It remains to be seen how ohrR is regulated in other bacteria.

OhrR binding to P1

Next, we assessed the binding of purified OhrR to a 293 bp fragment containing the P1 in the DNA bandshift assay. The results of the DNA bandshift assay show that OhrR binds specifically to P1 (Fig. 4). The OhrR binding to P1 was abolished by unlabelled competing P1 fragment but not by unrelated DNA sequence (Fig. 4). Moreover, the substitution of a protein unrelated to OhrR did not produce mobility shift of P1 (Fig. 4). The addition of increasing concentrations of OhrR to the P1 fragment did not produce additional species of slower migrating bands, suggesting that there was no co-operative binding of OhrR to the operator, and probably only one binding site for the protein was present within the P1 fragment (Fig. 4).

Figure 4.

DNA bandshift assay for binding of OhrR to P1. Purified OhrR was added to 293 bp of radioactively labelled P1 fragment in the binding buffer and separated in a polyacrylamide gel performed as described in Experimental procedures. The binding reaction consisted of labelled P1 fragment and 400 ng of OhrR. UP is the addition of 2 μg of unrelated protein (BSA) to the binding reaction; F is free P1 probe; the addition of increasing concentrations of OhrR (100, 200, 400, 800 and 1200 ng) to labelled P1 probe; P1 is the addition of 2 μg of unlabelled P1 DNA to the binding reaction; UD, the addition of 3 μg of unrelated DNA (pUC18 plasmid) to the binding reaction. The positions of free (F) and bound (B) P1 probe are shown on the left.

The location of the OhrR binding site within the P1 promoter fragment was determined by DNase I protection assay (Fig. 5). Analysis of the footprint patterns shows that OhrR binding to P1 produced a DNase I-protected region of 44 bp extending from –1 to –45 on the coding strand and from –7 bp to –51 bp on the non-coding strand (Fig. 5). The DNase I-protected regions com-pletely overlap the –35 and –10 regions of P1. Thus, the binding of OhrR prevents RNA polymerase from binding to the promoter, resulting in repression of gene expression.

Figure 5.

DNase I protection assay to locate OhrR binding site to P1. The DNase I protection assay for the binding of OhrR to P1. P1 represents the DNA fragments treated with DNase I; P1 + OhrR represents the binding of OhrR to the DNA fragments before DNase I treatment. The labelling of non-coding and coding strands was done as described in Experimental procedures. The arrows indicate the size of the fragments in bp, and the numbers in the brackets indicate the position of the protected regions with respect to the +1 transcription initiation site. G, A, T and C are the sequence ladder. M is radioactively labelled φX174 HinfI molecular weight markers.

Analysis of the DNA sequence within the OhrR protected region reveals several features that could constitute the OhrR operator site. There are several atypical AT-rich regions of two direct repeats ATAAATCG (D1) separated by 22 bp, TTGCAA (D2) separated by 21 bp and an inverted repeat TTGCAATT-AATTGCAA (I1) separated by 17 bp surrounded by normal GC-rich Xanthomonas DNA (Fig. 1B). These elements are all located in close proximity to the –35 and –10 regions of P1, and binding of OhrR to these sites would prevent RNA polymerase from binding to the promoter (Fig. 1B). The analysis of P1 deletion showed that half the ATAAATCG (D1) direct repeat could be removed (deletion no. 87) without altering the ability of OhrR to repress cat expression (Fig. 3), indicating that this motif was not crucial for the binding of OhrR to P1. In the B. subtilis system, the proposed sequence for the putative binding site for OhrR located in front of the ohrA also has AT-rich regions of overlapping inverted and direct repeats (Fuangthong et al., 2001). The importance of these elements as putative OhrR binding sites is suggested by analysis of a non-inducible ohrA mutant with a deletion that removes half the inverted and direct repeats (Fuangthong et al., 2001). OhrR from Xanthomonas and B. subtilis also shares high levels of homology at the amino acid sequence level, suggesting that they might recognize similar DNA motifs as the binding site. Comparison of the AT-rich regions of P1 and the ohrA promoter revealed a region with a high degree of similarity. Based on the analysis of the sequence alignment of these regions (data not shown), we proposed the putative OhrR operator site to be an inverted repeat TTnCAATT-(16/17)-AATTGnAA. The site consists of AT-rich inverted repeats separated by a relatively long space of 16–17 bp. In X. campestris pv. phaseoli, the putative OhrR operator consisted of a perfect inverted repeat separated by 17 bp (I1) located between the –35 and 3′ of –10 regions of the P1 promoter (Fig. 1B). Results of both the DNA footprinting of OhrR binding to P1 and the in vivo deletion analysis suggest that the binding site of OhrR is located within the –35 and –10 regions of P1 (Figs 3 and 1B ). In B. subtilis, the OhrR operator has an inverted repeat containing three mismatches and separated by 16 bp, and the operator overlaps the –35 and –10 regions of the ohrA promoter (Fuangthong et al., 2001). OhrR appears to recognize an extended operator site of about 32–33 bp. This probably accounts for the observed long (44 bp) protected region resulting from OhrR binding to P1 in the DNase I footprinting experi-ment (Fig. 5). At present, we do not know whether the D2 direct repeats contribute to OhrR binding to P1. The importance of the inverted and direct repeats is being investigated.

Overlapping binding sites for OhrR and RNA polymerase suggest that the repressor binding to the operator prevents RNA polymerase from initiating transcription at P1. Generally, a repressor has a higher binding affinity for an operator site than does RNA polymerase for a promoter. In uninduced cells, most of the OhrR binds to the operator, resulting in repression of P1. Exposure to organic peroxide probably inactivates OhrR and prevents it from binding to the operator, which allows RNA polymerase to bind and initiate transcription. This assumption is supported by the primer extension data showing that the tBOOH treatment increased the transcription initiation at P1 (Fig. 2). Moreover, preliminary investigations suggested that, in vitro, organic peroxide might directly modify OhrR and prevent it from binding to P1 (data not shown).

The stability of ohrR–ohr and ohr mRNA

We have shown that the bicistronic ohrR–ohr mRNA is processed by cutting in the loop section of the stem–loop structure 3′ of ohrR possibly by an RNase III-like enzyme to give monocistronic ohr mRNA and rapidly degraded monocistronic ohrR mRNA (Sukchawalit et al., 2001). This reduces the functional level of ohrR mRNA and ultimately affects the level of OhrR. As transcription of ohrR and ohr is driven from the P1, the stability of these mRNAs would have important effects on their expression. Here, we determined the stability of ohrR–ohr and ohr mRNA by measuring the half-lives of these mRNAs in tBOOH-induced cells. The data show that ohr mRNA is highly stable and has a half-life of >15 min (Fig. 6). In contrast, the bicistronic ohrR–ohr mRNA was highly labile, and the unprocessed bicistronic mRNA could not be detected 6 min after the addition of rifampicin (Fig. 6). Consistent with previous observations, no monocistronic ohrR mRNA could be detected (Fig. 6; Sukchawalit et al., 2001). This differential stability of ohrR and ohr mRNA would result in high concentrations of functional ohr mRNA, giving correspondingly high Ohr levels, and would reduce the level of ohrR mRNA. As Ohr is responsible for the detoxification of organic peroxide, a high level of the protein would be beneficial to the bacteria during exposure to organic peroxide stress, although low levels of OhrR also prevent excessive repression of the operon by the autoregulatory process. Thus, the differential stability and rapid processing of the mRNA exerts an additional post-transcription step to regulate the in vivo concentration of OhrR. This mechanism is not unique to the regulation of ohrR. Post-transcriptional regulation at the level of mRNA stability has been observed in diverse bacteria (Takata et al., 1989; Nilsson et al., 1996; Hebermehl and Klug, 1998; Homuth et al., 1999).

Figure 6.

Analysis of bicistronic ohrR–ohr and ohr mRNA stability. Total RNA was isolated from culture treated with 100 μM tBOOH for 10 min before the addition of rifampicin. Time zero represents the steady-state level of mRNA before the addition of 150 μg ml−1 rifampicin. At the stated time (min) after rifampicin addition, samples were withdrawn and total RNA extracted. RNA (15 μg) from each time point was loaded into each well and separated in a formaldehyde gel. The samples were blotted to a nylon membrane and probed with radioactively labelled ohr in (A). The steady-state (0) RNA sample was probed with ohrR in (B).

ohrR is inefficiently translated

We have observed in both uninduced and tBOOH-induced cells that OhrR is barely detectable by Western analysis, despite the gene having a highly efficient promoter (data not shown). Although we identified a post-transcriptional regulatory step that involves the differential stability of the mRNA, lack of correlation between the promoter strength and the concentration of OhrR suggested further regulation, perhaps at the translational level. ohrR produced a leaderless mRNA (Fig. 2). An alternative interpretation that ohrR mRNA could be translated from other translation initiation codons located further downstream of the ATG is unlikely, as there are no other commonly used Gram-negative translation initiation codons (ATG to GTG) nearby (Fig. 1B). Furthermore, studies on the translation of leaderless mRNA indicate that the ATG codon is required (Winzeler and Shapiro, 1997; Wu and Janssen, 1997). Here, we investigated the role of the ATG codon in the transcription of the gene and the translation of ohrR mRNA. A site-directed mutagenesis of the gene was done to change the ‘ATG’ codon to a rarely used translation initiation codon, ‘CTG’ (Fig. 7). Subsequently, the levels of ohrR transcription and translation in vivo were determined by making transcription (cat2; Mongkolsuk et al., 1993) and translation (cat3; Mongkolsuk et al., 1993) fusions of cat reporter genes to the mutated CTG-ohrR at the PstI site. Cat levels specified by Xanthomonas harbouring pOPcat2 and pOPcat3 were determined and compared with the levels attained by the strains harbouring non-mutated ohrR–cat fusion plasmids (Fig. 7). The results show that changing the translation initiation codon from ATG to CTG did not affect the level of transcription (pOPcat2). In contrast, the translation fusion of CTG-ohrR to cat3 (pOPcat3) was abolished, and no Cat fusion protein was detected (Fig. 7). The evidence confirmed that the translation of ohrR mRNA occurred at the proposed ATG and that the codon was required. However, the ATG codon was not important to the transcription of the gene.

Figure 7.

ohrR mRNA is inefficiently translated.

A. Diagrammatic representation of various plasmids showing the site of either transcriptional or translational fusions between ohrR, mutated CTG-ohrR and cat. cat2 and cat3 fusions represent transcriptional and translational fusions respectively.

B. Total protein (10 μg) prepared from cultures of Xanthomonas harbouring various plasmids was loaded into each lane. Separated protein samples were transferred to membranes after gel electrophoresis. Cat was detected by Western blot. U and I are uninduced and 100 μM tBOOH-induced cultures respectively.

We also investigated further the level of transcription and translation of ohrR using the cat reporter gene fusions. Initially, the transcription and translation fusions were made at the KpnI site located in the middle of ohrR (Fig. 7). The results of densitometer analysis show that the Cat levels specified by transcriptionally fused pOKcat2 were at least fivefold higher than the levels specified by the translationally fused pOKcat3 (Fig. 7). To confirm these observations, additional transcriptional and translational cat fusions were made at the PstI site located closer to the translation initiation of ohrR (Fig. 7). Cat levels specified by pOPcat2 were at least fivefold higher than the levels attained by pOPcat3. These fusions gave similar patterns regardless of the locations of the gene fusions and indicated that ohrR mRNA is inefficiently translated and that the translational level of the ohrR mRNA occurs in only about 20% of the total ohrR mRNA.

High levels of OhrR reduce organic peroxide resistance

The existence of multiple regulatory mechanisms to ensure that OhrR is not produced at high levels implies that the tight regulation of the gene must have important physiological consequences on the cells’ ability to respond to oxidative stress. The assumption was tested by measuring the effects of killing concentrations of H2O2, a superoxide generator (menadione) and organic peroxides on X. campestris pv. phaseoli harbouring pBBRohrR. In the bacteria harbouring pBBRohrR, tBOOH and cumene hydroperoxide gave zones of growth inhibition of 32 mm and 31 mm, respectively, compared with 26 mm for the bacteria harbouring pBBRMCS-1. H2O2 and menadione produced similar sizes of zone of growth inhibition in both strains. The data indicate that high levels of OhrR have detrimental effects on oxidative stress response by decreasing the organic peroxide resistance level, most probably by repression of the ohr expression. Thus, it is not surprising that Xanthomonas has evolved multiple mechanisms to ensure proper regulation of ohrR. At present, we do not know other genes in the OhrR regulon; thus, it is possible that other stress responses could be affected by high levels of OhrR.

A model for transcriptional and post-transcriptional regulation of ohrR

The transcriptional regulation of ohrR involves the autoregulation of the gene at the P1 promoter by OhrR. The binding target for OhrR overlaps with the –35 and –10 sites, enabling repressor to block RNA polymerase from binding to the promoter. In uninduced cells, OhrR represses the expression of its own operon. The post-transcriptional regulation of ohrR occurs in two steps. First, the bicistronic ohrR–ohr mRNA is rapidly processed giving high levels of ohr mRNA and rapid degradation of ohrR mRNA. ohr mRNA is highly stable, in contrast to highly labile bicistronic ohrR–ohr. This greatly reduces the functional concentration of ohrR mRNA. Secondly, ohrR mRNA is inefficiently translated, and translation of the mRNA occurred at only 20% of the transcription level. Multiple regulatory mechanisms at transcriptional and post-transcriptional levels ensure that the intracellular level of OhrR remains low. When cells are exposed to organic peroxides, they presumably inactivate OhrR and prevent the repressor from binding to the operator, resulting in high levels of expression of ohrR and ohr. Once Ohr removes organic peroxide, OhrR then auto-regulates itself and represses the expression of the operon (Fig. 8).

Figure 8.

A model for ohrR regulation at transcriptional and post-transcriptional levels.

Experimental procedures

Bacterial culture conditions and transformation

All Xanthomonas strains were grown aerobically in SB (Silva–Buddenhagen medium: 0.5% peptone, 0.5% yeast extract, 0.5% sucrose and 0.1% glutamic acid, pH 7.0) at 28°C (Mongkolsuk et al., 1997). The oxidant induction experiments were performed on exponential phase cells by the addition of 100 μM tBOOH to cultures followed by incubation for an additional 15 min for Northern analysis and 30 min for Western analysis before cells were harvested for lysate preparation. Antibiotics were used at the following concentrations: for the selection of chromosomal integrated mutants, 15 μg ml−1 kanamycin; and for selection of plasmids, 15 μg ml−1 gentamicin and 30 μg ml−1 kanamycin. All plasmids were transformed into X. campestris pv. phaseoli by electroporation using previously described conditions (Mongkolsuk et al., 1997).

Western blots

Crude protein (10 μg) was loaded into each lane of a 10% SDS–PAGE gel. The separated proteins were electrophoretically transferred to a sheet of nylon membrane. Membrane blocking, primary antibody reaction, washing and subsequent detection of the immune reaction by alkaline phosphatase-conjugated secondary antibody were performed as described previously (Loprasert et al., 2000).

Nucleic acid purification and ohrR primer extension

Total RNA was isolated using the modified hot phenol method from uninduced and tBOOH-induced X. campestris pv. phaseoli cultures (Mongkolsuk et al., 1997). Primer extension experiments were carried out using 32P-labelled OR1 primer (5′-ATACAACGCAAAGCACAGCTG-3′), 5 μg of total RNA and 200 U of SuperScript II MMLV reverse transcriptase. The extension products were analysed on sequencing gels next to sequence ladders. The sequence ladders were done using a polymerase chain reaction (PCR) sequencing kit with labelled OR1 primer and pBBRohrR plasmid as the template.

Construction of ohrR fusions

The intein–OhrR fusion was made by cloning of 480 bp NcoI–XhoI-digested PCR products from pBBRohrR template and ohrRI1 (5′-GGCTCGAGTCCCGCGCCAAGG-3′) and ohrRI2 (5′-GGAAACAGCTATGACCATG-3′) into similarly digested pCYB4 (New England BioLabs) giving pINTohrR. The nucleotide sequences of fused genes were determined to confirm fusion in the correct reading frame and that no other mutations had occurred.

Purification of OhrR

The intein fusion system was used by making the fusion at the carboxy-terminus of OhrR. Escherichia coli harbouring pINTohrR was grown to mid-log phase before 1 mM IPTG was added and incubation continued for 3 h. The culture was harvested, and cell pellets were resuspended in the column buffer (20 mM Tris-HCl, pH 8.0, 500 mM NaCl, 0.1 mM EDTA, 0.1% Triton X-100) and sonicated. The lysate was spun at 10 000 g for 15 min, and lysate was loaded on to a chitin bead column, which was subsequently washed extensively with column buffer. The on-column cleavage of fusion protein was done by the addition of buffer C [20 mM Tris-HCl, pH 8.0, 50 mM NaCl, 0.1 mM EDTA and 30 mM dithiothreitol (DTT)] at 4°C overnight. The eluted fractions containing OhrR were pooled and dialysed against Z buffer (10 mM HEPES, pH 8.0, 1 mM EDTA, 20 mM MgCl2, 60 mM KCl and 20% glycerol).

DNA bandshift assay and DNase I footprinting

Strand-specific, radioactively labelled DNA fragments were prepared by PCR using labelled M13 reverse primer for the coding strand and BT18 for the non-coding strand with pohrRsfi as templates. The PCR products generated 293 bp fragments, which were used in footprinting the coding strand. For the non-coding strand, the PCR product was digested with EcoRI to give 180 bp that was used in the footprinting experiment. The DNA bandshift reactions were performed by adding 3 fmol of labelled probe to the buffer [20 mM Tris, pH 7.0, 50 mM KCl, 1 mM EDTA, 5% glycerol, 50 μg ml−1 BSA, 5 μg ml−1 calf thymus DNA, 0.5 mg ml−1 poly-(dI–dC)]. Purified OhrR (400 ng) was added, and the reaction was incubated at 25°C for 15 min. For DNase I footprinting, 25 μl of 0.2 mM Mg2+ and 0.1 mM Ca2+ and 0.5 U of DNase I were added to the binding reaction, and incubation was continued for 1 min before 200 μl of stop solution (20 mM EDTA, pH 8.0, 1.0% SDS and 0.2 M NaCl) was added. The mixture was extracted with phenol–chloroform and ethanol precipitated. The pellets were resuspended in sequencing buffer and loaded onto a sequencing gel. The DNA sequence ladder for the non-coding strand was performed with fmol sequencing kits (Promega) using labelled M13 primer and pahpC plasmid (Mongkolsuk et al., 1997). Sequence reactions were loaded on to a denatured DNA sequencing gel.

Determination of mRNA stability

Xanthomonas campestris pv. phaseoli strain 182 cultures were used in the determination of mRNA stability. Exponential phase cultures of the bacteria (OD600 of 0.7) were treated with 150 μg ml−1 rifampicin to inhibit new RNA synthesis. Then, at 1 min intervals, 10 ml of culture was withdrawn, pelleted rapidly and extracted for total RNA (Mongkolsuk et al., 1997). Subsequent steps in the RNA extraction, formaldehyde gel electrophoresis and construction of ohr and ohrR probes were performed as described previously (Sukchawalit et al., 2001).

Site-directed mutagenesis and gene fusions

To investigate the role of the ohrR translation initiation codon in transcription and translation of the gene, the codon was mutated from ATG to CTG using primers CT (5′-ATTGCAACCCTGGACACCACC-3′), CTR (5′-GGTGGT GTCCAGGGTTGCAAT-3′) and pBBRohrR as the template in a PCR mutagenesis reaction. The mutated gene CTG-ohrR was digested with SfiI–PstI, and the 139 bp fragment was cloned into similarly digested pUFRcat2-Km, resulting in a transcription fusion of the gene to the cat2 in pOPcat2. The translation fusion of the gene was constructed using a similar strategy to that described for the non-mutated ohrR, except that CTG-ohrR was used as the starting material. This gave pOPcat3. pOPcat2 and pOPcat3 were trans-formed into X. campestris pv. phaseoli, and Cat levels were determined.

Construction of ohrR transcriptional and translational fusions

The transcriptional fusions of ohrR to cat were made by cloning either the 610 bp BamHI–KpnI or the 139 bp SfiI–PstI fragments in front of a promoterless cat in the promoter probe vector pUFR027cat2-Km, giving pOKcat1 and pOPcat1 respectively. These recombinant plasmids were confirmed by restriction mapping. The translational fusions were made by cloning these fragments into the polylinker region of the plasmid pSM-cat3 (containing cat coding sequence minus the ribosome binding site; Mongkolsuk et al., 1993) giving pKcat3 and pPcat3. The cloning sites were chosen so that the ohrR fused in frame with the cat coding sequence. All fusion joints were sequenced. Moreover, these OhrR–Cat fusions should give proteins with a higher molecular weight than Cat. Western analysis was used to confirm the size of the recombinant translation fusions. pKcat3 and pPcat3 were digested with HindIII and EcoRI, and the DNA fragments containing the ohrR–cat fusion were cloned into pUFR027cat2-Km giving pOKcat3 and pOPcat3 respectively. All recombinant transcriptional and translational fusion plasmids on the broad-host-range replicons were electroporated into X. campestris pv. phaseoli.

Qualitative determination of oxidant resistance levels

The zone of growth inhibition method was used to measure the resistance level to various oxidants. Exponential phase cells (106) were mixed with semi-soft SB agar and poured on top of an SB plate. After the semi-soft agar had solidified, 6 mm 3M paper disks individually impregnated with 6 μl of 1 M tBOOH, cumene hydroperoxide, menadione and H2O2 were placed on top of the cell lawn. The zone of growth inhibition was measured after 24 h incubation at 28°C.

Acknowledgements

We thank P. S. Lovett for editorial assistance and useful suggestions. The research was supported by grants from Chulabhorn Research Institute to the Laboratory of Biotechnology, an NSTDA career development award RCF 01-40-005 and the Thailand Research Fund BRG 10-43 to S.M. P.V., W.P. and W.E. were supported by a postdoctoral (PDF/65/2543) and the Royal Golden Jubilee (PHD/0146/2542 and PHD/0200/2543) fellowships from the Thailand Research Fund respectively. Parts of this work are from W.P.’s PhD thesis at Mahidol University.

Ancillary